6?l, i A.b&B ENT 1 The Journal of VOLUME 23 1995 NUMBER 1 THE JOURNAL OF ARACHNOLOGY EDITOR: James W. Berry, Butler University ASSOCIATE EDITOR: Petra Sierwald, Field Museum EDITORIAL BOARD: A. Cady, Miami (Ohio) Univ. at Middletown; J. E. Carrel, Univ. Missouri; J. A. Coddington, National Mus. Natural Hist.; J. C. Cokendolpher, Lubbock, Texas; F. A. Coyle, Western Carolina Univ.; C. D. Dondale, Agriculture Canada; W. G. Eberhard, Univ. Costa Rica; M. E. Galia- no, Mus. Argentino de Ciencias Naturales; M. H. Greenstone, BCIRL, Columbia, Missouri; C. Griswold, Calif. Acad. Sci.; N. V. Horner, Midwestern State Univ.; D. T. Jennings, Garland, Maine; V. F. Lee, California Acad. Sci.; H. W. Levi, Harvard Univ.; E. A. Maury, Mus. Argentino de Ciencias Naturales; N. I. Plat- nick, American Mus. Natural Hist.; G. A. Polis, Vanderbilt Univ.; S. E. Riechert, Univ. Tennessee; A. L. Rypstra, Miami Univ., Ohio; M. H. Robinson, U.S. National Zool. Park; W. A. Shear, Hampden-Sydney Coll.; G. W. Uetz, Univ. Cincinnati; C. E. Valerio, Univ. Costa Rica. The Journal of Arachnology (ISSN 0160-8202), a publication devoted to the study of Arachnida, is published three times each year by The American Arach- nological Society. Memberships (yearly): Membership is open to all those in- terested in Arachnida. Subscriptions to The Journal of Arachnology and American Arachnology (the newsletter), and annual meeting notices, are included with mem- bership in the Society. Regular, $30; Students, $20; Institutional, $80 (USA) or $90 (all other countries). Inquiries should be directed to the Membership Secretary (see below). Back Issues: Patricia Miller, PO. Box 5354, Northwest Mississippi Community College, Senatobia, Mississippi 38668 USA. Telephone: (601) 562- 3382. Undelivered Issues: Allen Press, Inc., 1041 New Hampshire Street, PO. Box 368, Lawrence, Kansas 66044 USA. THE AMERICAN ARACHNOLOGICAL SOCIETY PRESIDENT: James E. Carico (1993-1995), Dept, of Biology, Lynchburg, Vir- ginia, 24501 USA. PRESIDENT-ELECT: Matthew H. Greenstone (1993-1995), USDA-BCIRL, Columbia, Missouri 65203 USA. MEMBERSHIP SECRETARY: Norman I. Platnick (appointed), American Museum of Natural History, Central Park West at 79th St., New York, New York 10024 USA. TREASURER: Gail E. Stratton (1993-1995), Department of Biology, Albion College, Albion, Michigan 49224 USA. BUSINESS MANAGER: Robert Suter, Dept, of Biology, Vassar College, Pough- keepsie, New York 12601 USA. SECRETARY: Alan Cady (1993-1995), Dept, of Zoology, Miami Univ., Mid- dleton, Ohio 45042 USA. ARCHIVIST: Vincent D. Roth, Box 136, Portal, Arizona 85632 USA. DIRECTORS: Allen R. Brady (1993-1995), Pat Miller (1993-1996), Ann Ryp- stra (1993-1995). HONORARY MEMBERS: C. D. Dondale, W. J. Gertsch, H. W. Levi, A. F. Millidge, W. Whitcomb. Cover illustration: Scanning electron microscope photograph of the abdomen of a mature female Philoponella vicina (Uloboridae). Photograph by Flory Pereira and William G. Eberhard. Publication date: 9 August 1995 ® This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 1995. The Journal of Arachnology 23:1-8 PHILOPONELLA REPUBLICANA (ARANEAE, ULOBORIDAE) AS A COMMENSAL IN THE WEBS OF OTHER SPIDERS Ann L. Rypstra and Greta J. Rinford1: Department of Zoology, Miami University, 1601 Peck Blvd., Hamilton, Ohio 45011 USA ABSTRACT. Juvenile individuals of the spider species, Philoponella republicana, were common in the webs of the social spider, Anelosimus eximius, and the solitary spider, Architis sp., in the forest habitats of the SE Peru. The abundance, size and location of P. republicana individuals were surveyed in each host web. Although the host webs were similar in size and conformation, more P. republicana individuals were found in the social spider webs than in the solitary host webs. Likewise, the number of P. republicana in the social spider webs was correlated with host web size. The mean size of prey captured by P. republicana was 2.1 mm, which was significantly smaller than the prey taken by the social spider, and, in feeding trials, Architis sp. individuals reacted only infrequently to prey of that size. This separation in the size of prey taken caused us to conclude that P. republicana acted as a commensal for the most part; however, they were observed to prey on the social spiders occasionally. Small P. republicana were the most common in both host webs and tended to be high in the barrier webbing. The largest individuals in the social host webs were located under the sheet area, and these individuals were observed to feed more frequently than spiders in other size classes and in other areas of the host webs. We conclude that juvenile P. republicana are commensals in both host webs but that they benefit more from the greater amount of activity in webs Mounting evidence from phylogenetically di- verse species shows that grouping behavior may simultaneously reduce individual risk of preda- tion and enhance feeding efficiency (Pulliam & Caraco 1984; Uetz 1988; Uetz & Heiber 1994). Heterospecific interactions within social groups can bring advantages to individuals in those groups that do not accrue to individuals in single species aggregations (Morse 1970; Barnard & Thompson 1985). Slightly different foraging modes and food preferences may lead to more efficient resource usage by mixed species groups which can simultaneously take advantage of oth- er kinds of advantages of being in a group. A wide variety of heterospecific relationships have been reported among spider species ranging from predation (Larcher & Wise 1985; Jackson & Whitehouse 1986; Jackson 1990) to klepto- parasitism (Vollrath 1987; Cangialosi 1990), to commensalism (Rypstra 1979; Bradoo 1986, 1989). In many of these instances, host spiders have large complex webs that can provide a liv- ing space with some support and protection for the second spider species (Rypstra 1979; Bradoo 1986, 1989; Hodge & Uetz 1992). In particular, the webs of communal or social spiders tend to provide habitat for other spider species who in- 1 Present address: Dept, of Ecology and Evolutionary 1 USA social spiders. teract with the host in both positive and negative ways (Rypstra 1979; Bradoo 1989; Cangialosi 1991; Hodge & Uetz 1992). A commensal association occurs when one species reaps some benefits by association with a host species but the host species is essentially unaffected, positively or negatively, by the as- sociation. Commensalism has been reported with some frequency among spider species in the fam- ily Uloboridae (Struhsaker 1969; Opell 1979; Bradoo 1986; 1989). Bradoo (1989) concludes that Uloborus ferokus Bradoo (Araneae, Ulobor- idae), living in the webs of the social spider Ste- godyphus sarasinorum Karsch (Araneae, Erisi- dae), receives protection, support and increased prey capture which increases its lifespan and fit- ness. The spider species, Philoponella republi- cana (Simon) (Araneae, Uloboridae), is frequent- ly found in single species aggregations (Smith 1985; Binford & Rypstra 1992); but, in addition, we have found immature individuals of the spe- cies in the interstices of the webs of almost all complex, semi-permanent spider webs at our study area in SE Peru. P. republicana were par- ticularly common in the webs of Anelosimus ex- imius Simon (Araneae, Theridiidae), a cooper- atively social species in this area. The goal of this logy, University of Arizona, Tucson, Arizona 85721 1 2 THE JOURNAL OF ARACHNOLOGY study is to describe the abundance and distri- bution of P. republicana in the large webs of this social spider in comparison with its distribution in the webs of a solitary species, Architis sp. (Ara- neae, Pisauridae), whose web is of similar size and structure (Nentwig 1985). METHODS Data were collected on spider populations in- habiting the subtropical moist forest of the Tam- bopata Reserved Zone, 35 km southwest of Puer- to Maldonado in Madre de Dios, Peru. Data were collected in the dry season: July and early August of 1987, 1988 and 1989 (see Erwin 1985 for complete description of the habitats). The webs of both host spiders were very sim- ilar in overall appearance. They consisted of a dense sheet of webbing subtended by a maze of barrier webbing encompassing neighboring veg- etation (Brach 1975; Christenson 1984; Nentwig 1985). A. eximius is a cooperatively social species so each web contained several hundred to several thousand individuals that worked together to capture prey (Brach 1975; Christenson 1984). Architis sp. is a solitary spider and a single in- dividual monitors insects arriving in the web from a funnel-shaped retreat at one end of the sheet area of the web (Nentwig 1985). Adults of A. eximius are 4-6 mm in length, which is substan- tially smaller than Architis adults which are 8- 12 mm in length. Surveys were conducted of all A. eximius webs found, a total of 46 webs, between 4 July and 4 August in 1987 (18 webs), 1988 (16 webs) and 1989 (12 webs). To avoid the confounding factor of repeated measures only one survey per social spider web was included in the data set. In order to standardize for season and temperature across the years we selected the first survey conducted on a web after 4 July on a dry day on which the temperature was between 24-28 °C. A total of 12 Architis sp. webs were surveyed a single time and under similar weather circumstances in July of 1 989. During each survey, P. republicana were classified into three size categories: large (4-6 mm in length), medium (2-4 mm in length), and small (less than 2 mm in length). P. republicana were also categorized by position in the host web. That categorization included spiders located un- der the sheet, just above the sheet (within 2 cm), in the low barrier of the web (2-20 cm above the sheet) and in the high barrier (20 cm or more above the sheet). In order to obtain one measure of site quality within the host web, we also attempted to de- termine the feeding frequency of P. republicana spiders located in different positions. A spherical bundle in the chelicerae of the spider was evi- dence that it had captured a prey item recently. One complication that arises in determining the likelihood of feeding is that the spider will feed longer on large prey than on small prey so a survey sampling technique would have biased the results toward large prey. In the case of A. eximius webs, we typically spent two or more hours observing so, for this study, we only count- ed the prey items that were captured during our observation times. For Architis webs, we sur- veyed a second time 2-3 hours after the first observation to estimate a feeding rate in a similar fashion. To determine whether the two species were actively competing for prey that entered the web or if there was a division of resources based on prey size, we needed to determine the range of prey sizes taken by each of the host species. The distribution and frequency of prey capture were obtained for A. eximius in the course of a si- multaneous study (Rypstra 1990; Rypstra & Tir- ey 1991). In order to determine whether the sol- itary Architis sp. actively preys on insects in the size class that P. republicana handles, we con- ducted a feeding experiment. Field-caught fruit flies (. Drosophila spp. 1. 5-2.0 mm in length, the mean size of prey taken by P. republicana) were gently blown into each of ten webs of Architis. In all cases the Architis individual was at the opening of its retreat in a feeding position at the time the prey were introduced. If the Architis spider retreated before the prey was in the web or if it was apparent that we had disturbed her in the process, no data were taken. If we suc- cessfully introduced the fly without disturbing the host spider and we were able to detect that the fly contacted the sheet in a way sufficient to vibrate the threads, we recorded the reactions of the Architis. Between 8-12 flies were tested in each of ten Architis webs. After each trial, a larger fly or grasshopper was introduced into the web to see if the host spider was receptive to any prey. If we could not get the spider to respond, the results of the trial were excluded from the anal- ysis. RESULTS All of the webs that we found in all three years had some P. republicana in them. On average, there were 8.4 ±3.3 (mean ± standard deviation) RYPSTRA & BINFORD-SPIDER COMMENSALS 3 1987 1988 1989 ARCHITIS WEB DIMENSION (CM) Figure I.—- The number of commensal Philoponella republicana individuals vs. the longest horizontal dimen- sion of the host web. Data points indicated for the three years (1987, 1988, and 1989) are all for the social host, Anelosimus eximius. Data for the solitary host, Architis sp., were all gathered in 1989. The correlation between web size and number of commensals is significant for the social Anelosimus eximius but is not significant for the solitary Architis sp. P. republicana individuals in the 46 A. eximius webs we surveyed over three years. There were no differences among the years (Kruskal- Wallis Multiple Comparisons, P > 0.05). There were significant positive relationships between social spider web size and the number of P. republicana in the web both within each year and when the data for all years were pooled. The strongest re- lationship was between the longest horizontal di- mension and number of spiders (for all years together: Spearman’s r = 0.85, P < 0.05) (Fig. 1). We found a mean of 4. 1 7 ± 2.6 P. republicana individuals in the 12 webs of Architis sp. that we surveyed in 1 989. This was significantly less than the numbers we found in the social spider webs (Mann-Whitney (7-Test, P < 0.05). The webs of Architis and A. eximius were similar in all the dimensions we measured: longest horizontal, perpendicular or web height, and height of the sheet above ground (Mann-Whitney U- Test for all, P > 0.25). However, there was no relation- ship between web size and number of P. repub- licana individuals in Architis sp. webs (longest horizontal web dimension and spider number: r = 0.4, P > 0.2) (Fig. 1). The distribution of P. republicana individuals in the various size classes we identified was not even within either host web (x2 Test, P < 0.05). In both web types, small spiders were most abun- dant and large spiders the least abundant (Fig. 2) . The size distributions of P. republicana in the two host types were significantly different from one another (%2 Test, P < 0.05). Most noticeably, there were more large individuals in A. eximius webs than there were in the webs of Architis sp. (Fig. 2). The distribution of P. republicana webs across the four positions we identified in the host webs was also skewed (x2 Test, P < 0.05). In Architis sp. webs, most of the P. republicana (40 of 63 total) were located in the barrier area (Fig. 3) . However, in A. eximius webs, P. republicana were evenly distributed between areas close to the sheet and barrier areas (Fig. 3). Specifically, the P. republicana in the social spider webs were most abundant under the sheet and in the high barrier; and they were least abundant just above the sheet and in the low barrier (Fig. 3). The distributions of P. republicana webs in the two host species we observed in 1989 were signifi- cantly different from one another (x2 Test, P < 0.05). 4 THE JOURNAL OF ARACHNOLOGY SMALL MEDIUM LARGE SIZE ARCHITIS A. EXIMIUS Figure 2. —The size distribution of Phiioponella repuhlicana individuals in the webs of the two host species, Architis sp. and Anelosimus eximus. “Small” spiders were all less than 2 mm in length; “Medium” spiders ranged from 2=4 mm in length, and “Large” spiders were between 4-6 mm in length. We had sufficient data to look specifically at the distribution of the different size classes and their feeding frequencies for P. repuhlicana in the social spider webs. Small P. repuhlicana were abundant in the high barrier but a very few were located just above the sheet (Fig. 4). Only 15 of the 172 small spiders we censused captured any prey item and fed during our observations, and the distribution of those individuals was not sig- nificantly different from the distribution of all small spiders within the host webs (x2 Test, P > 0.3) (Fig. 4). Medium-sized P. repuhlicana were evenly distributed across the positions within the social spider webs (x2 Test, P > 0.3); however, those located close to the sheet were more likely to be observed feeding than those in the barrier areas (25 of the 30 spiders that captured prey) (Fig. 4). Forty-seven of the 77 large spiders we observed were located under the sheet of the host web so the distribution of individuals in this size class was not even across the positions (x2 Test, P < 0.05) (Fig. 4). Twenty-four of the 77 large spiders we censused fed during our observations and 80% of those were located under the sheet (x2 Test, P < 0.05) (Fig. 4). Our sample size of P. repuhlicana in Architis sp. webs was not large enough to make the com- parisons of position, size and feeding that we were able to make in the social spider host. Only 5 of the 63 spiders censused in Architis sp. webs captured prey, and all of those were spiders in the large size category located in the barrier web- bing. The prey captured by P. repuhlicana in these host webs was 2.1 ± 1.2 mm in length which is much smaller than the mean prey size captured by A . eximius (5.9 ± 2.1 mm) (Mann-Whitney (7-Test, P < 0.05) (Rypstra 1990). In our prey introduction trials with Architis sp., spiders re- acted to only 1 7 of the 1 1 2 insects in this size class that we introduced into 10 different webs and the fruit flies were captured on only six oc- casions. In most of the introductions, the Architis web resident did not move at all when the prey were introduced but then would respond to the larger prey item at the end of the trial. The cribellar silk of the P. repuhlicana was able to detain A. eximius quite effectively if they hap- pened into one of the webs. On three occasions, a P. repuhlicana individual, located in the bar- rier, successfully captured and killed a penulti- mate or adult A. eximius female. In no case did we observe A. eximius capture a P. repuhlicana. At dusk, A. eximius has a period of web cleaning and maintenance; and, at that time, they would cut out and remove many of the webs of P. re- puhlicana that were located above the sheet and in the barrier webbing. When they did this, the RYPSTRA & BINFORD — SPIDER COMMENSALS 5 CD UJ <5 H (/) O UJ o cc LU a UNDER ABOVE LOW HIGH POSITION ARCHITIS A. EXIMIUS Figure 3.— The distribution of Philoponella republicana in the various positions in the host webs. The identified positions were “Under” the sheet, just “Above” the sheet (within 2 cm), in the “Low” barrier webbing (between 2-20 cm above the sheet) and in the “High” barrier (greater than 20 cm above the sheet). P. republicana would evacuate their web and hang motionless near the location. In only one in- stance did we observe A. eximius destroying a P. republicana web located below the sheet during this activity. DISCUSSION Our observations suggest that P. republicana is a commensal in the webs of A. eximius and Architis sp. since they capture prey much smaller than those captured by the host species. They appear to use the webs of other species as sup- port, perhaps to enable them to locate their small orb webs in areas otherwise unavailable to them. It is also possible that the webs of the host species enlarge the effective size of their own web allow- ing them to detect insects sooner and at a greater distance. Perhaps there is some ricochet effect as small insects are deflected and detained by strands of the large host web which could increase the rate of capture by the commensal spider (Uetz 1989). Bradoo (1986) observed U. ferokus mov- ing out of their orbs to capture prey on the surface of the host web but we never observed this sort of behavior by P. republicana . Bradoo’s (1986) descriptions suggest that the relationship be- tween that commensal and its social spider host is much more interdependent than that which we observed between P. republicana and A. ex- imius. There are more commensals located in the so- cial spider webs than in the solitary spider’s web. Likewise, as the social spider web becomes larger more potential web sites are formed and more P. republicana colonize them resulting in a cor- relation between their number and web size. However, even though more web sites would presumably be available as the Architis webs in- creased in size as well, no additional commensal spiders colonized them. We suspect that the in- creased activity in the social webs deflects more prey into the commensal’s webs, which would make those sites preferable. Unfortunately, prey capture by P. republicana in Architis sp. webs was sufficiently uncommon in our observations that we cannot verify that difference statistically. The fact that there were more large individuals in the social spider webs suggests that they feed more successfully there. In addition, the fact that they are occasional predators on the host in the social spider’s web indicates that more potential food is available there. On the other hand, the density of commensals in the webs of the social spider may not be related to prey capture at all. At some point in the evo- lution of sociality, spiders must become more tolerant of other spiders (Kullmann 1972; Wil- 6 THE JOURNAL OF ARACHNOLOGY SMALL SPIDERS SO 1 — — — — - — LU o oc UJ a. UNDER ABOVE LOW HIGH POSITION □ W/PREY ■ NO PREY MEDIUM SPIDERS SO T~ — h- Z LU o cc LU a. 60- 40- UNDER ABOVE LOW HIGH POSITION □ W/ PREY NO PREY LARGE SPIDERS POSITION W/ PREY NO PREY Figure 4.— The distribution of various size classes in the various positions in the social spider webs and the proportion of individuals which were observed to capture prey within a three-hour time period. KYP8TRA & BINF0RD — SPIDER COMMENSALS 7 son 1975). It may be that A . eximius webs are easier to colonize because of this factor. The so- cial spider regularly associates with lots of other spiders and therefore must have relaxed its ag- gressive tendencies toward them, whereas the solitary species can afford to rid aggressively its web of all other spiders indiscriminately except during the restricted circumstances when it Is being courted by a male. The fact that the social species is predisposed to tolerate other spiders may mean that it is easier for heterospecific re- lationships to perpetuate in their company. Locations under the sheet in the social spider’s webs seem to attract the largest P. republicana , and those locations also seem to be the areas of highest prey capture (Fig. 4). It could be that juveniles that happen to select web sites in that location capture more prey and thereby grow large more quickly and/or remain longer in the host web that those in other areas. It is also possible that, as the spiders grow, they are more able to compete for these high quality sites. Our data suggest that sites under the sheet afforded both higher prey capture rates and more protection from disturbance in the host webs of the social spider. One disadvantage to these sites may be that P. republicana that located there would en- counter fewer of the social spider individuals and therefore be less successful as a predator. It is curious to note that we never found an adult in the interstices of these complex host webs, even though adults were present in at the study site at the time these data were collected. Interestingly, in our study area we found adults only In aggregations composed exclusively of P. republicana individuals. Lubin (1980) suggested that the single species aggregations of P. repub- licana were sibling groups arising from a single egg case, and our surveys of P. republicana col- onies over the years support that idea (unpub. data). However, we have also observed that a large number of juveniles are commensals for some portion of their life. If this commensal state is a phase that many juveniles pass through, then it would be interesting to discover what cues they use to reaggregate with conspecifics in this com- plicated habitat. Nyffleler & Benz (1980) report that juvenile stages of several other species of orb weavers act as kleptoparasites and commen- sals in the webs of other spiders. They consider the commensal relationship as a transition to a more invasive kleptoparasitism. We have al- ready mentioned the more active prey capture out of the host web that Bradoo (1986) reports for U. ferokus which suggests that there is a con- tinuum of dependence on other spider species within the family Uloboridae. It is important to conduct more detailed studies of these relation- ships in order to understand more fully the evo- lution of these various behavior patterns. ACKNOWLEDGMENTS We thank K. Cangialosi, A. McCrate, S. Tirey, J. Whitis and the staff of the Explorer’s Inn for assistance in the field. We are grateful for the cooperation of the staff at the Ministerio de Agri- cultura de Peru in issuing authorization to work with the spiders of Tambopata. Voucher speci- mens were placed at the Museo de Historia Nat- ural, Lima, Peru and at the Smithsonian Insti- tution, Washington, D. C. M. A. Hodge and G. W. Uetz made many useful suggestions on an earlier draft of this manuscript. Financial sup- port for this research came from the Society of Sigma Xi, the FrizzelLExline Fund for Arach- nological Research, a Miami University Under- graduate Research Award (all to G. I. B.) and an NSF grant BSR86-04782 (to A. L. R.). LITERATURE CITED Barnard, C. J. & D. B. A. Thompson. 1985. Gulls and Plovers: the Ecology and Behaviour of Mixed- Species Feeding Groups. Columbia Univ. Press, New York, New York. Binford, G. J. & A. L. Rypstra. 1992. Foraging be- havior of the communal spider, Philoponella re- publicana (Araneae: Uloboridae). J. Ins. Behav., 5:321—335. Brach, V. 1975. The biology of the social spider A ne- losimus eximius (Araneae, Theridiidae). Bull. So. California Acad. ScL, 74:37-41. Bradoo, B. L. 1986. Feeding behaviour of a non- poisonous spider Uloborus ferokus Bradoo (Arane- ae: Uloboridae). Zool. Anz., 217:75-88. Bradoo, B. L. 1989. Advantages of commensalism in Uloborus ferokus Bradoo (Araneae: Uloboridae). J. Bombay Nat. Hist. Soc., 86:232-328. Cangialosi, K. R. 1990. Kleptoparasitism in colonies of the social spider Anelosimus eximius (Araneae, Theridiidae). Acta Zool. Fennica, 190:51-54. Cangialosi, K. R. 1991. Attack strategies of a spider kleptoparasite: effects of prey availability and host colony size. Anim. Behav., 41:639-647. Christenson, T. E. 1984. Behaviour of colonial and solitary spiders of the theridiid species Anelosimus eximius. Anim. Behav., 32:725-734. Erwin, T. L. 1984. Tambopata Reserved Zone, Ma- dre de Dios, Peru: history and description of the reserve. Rev. Peruana Ent., 27:1-8. Hodge, M. A. & G. W. Uetz. 1992. Antipredator 8 THE JOURNAL OF ARACHNOLOGY benefits of single- and mixed-species grouping by Nephila clavipes (L ) (Araneae, Tetragnathidae). J. Arachnol, 20:212-216. Jackson, R. R. 1990. Predator-prey interactions be- tween jumping spiders (Araneae, Salticidae) and Pholcus phalangioides (Araneae, Pholcidae). J. Zook, London, 220:553-559. Jackson, R. R. & M. E. A. Whitehouse. 1986. The biology of New Zealand and Queensland pirate spi- ders (Araneae, Mimetidae): aggressive mimicry, ar aneophagy and prey specialization. J. ZooL, London (A), 210:279-303. Kullmann, E. J. 1972. Evolution of social behavior in spiders (Araneae; Eresidae and Theridiidae). American ZooL, 12:419-426 Pa re he r , S. R. & D. H. Wise. 1985. Experimental studies of the interactions between a web-invading spider and two host species. J. Arachnol, 1 3:43-59. Lubin, Y. D. 1980. Population studies of two colonial orb-weaving spiders. ZooL J. Linn. Soc., 70:265- 287. Morse, D. H. 1 970. Ecological aspects of some mixed- species foraging flocks of birds. Ecol. Monogr., 40: 119-168. Nentwig, W. 1985. Architis nitidopilosa as a neo- tropical pisaurid with a permanent catching web. Bull. British Arachnol. Soc., 6:297-303. Nyffeler, M. & G. Benz. 1980. Kleptoparasitismus con juvenilen Kreuzspinnen and Skorpionsfliegen in den Netzen adulter Spinnen. Rev. Suisse ZooL, 87:907-918. Opell, B. P. 1979. Revision of the genera of tropical American species of the spider family Uloboridae. Bull. Mus. Comp. ZooL, 148:443-549. Pulliam, H. R., & T. Caraco. 1984. Living in groups: Is there an optimal group size? Pp. 122-147, In Be- havioural Ecology: An Evolutionary Approach. 2nd ed. (J. R. Krebs & N. B. Davies, eds.). Sinauer As- sociates, Sunderland, Massachusetts. Rypstra, A. L. 1979. Foraging flocks of spiders: a study of aggregate behavior in Cyrtophora citricola Forskal (Araneae: Araneidae) in west Africa. Behav. Ecol. Sociobiol., 5:291-300. Rypstra, A. L. 1990. Prey capture and feeding effi- ciency of social and solitary spiders: a comparison. Acta ZooL Fennica, 190:339-343. Rypstra, A. L. & R. S. Tirey. 1991. Prey size, prey perishability and group foraging in a social spider. Oecologia, 86:25-30. Smith, D. R. R. 1985. Habitat use by colonies of Philoponella republicana (Araneae, Uloboridae). J. Arachnol, 13:363-373. Struhsaker, T. T. 1 969. Notes on the spiders Uloborus mundior and Nephila clavipes in Panama. American Midi Nat., 82:611-613. Uetz, G. W. 1988. Risk sensitivity and foraging in colonial spiders. Pp. 353-377, In The Ecology of Social Behavior. (C. N. Slobodchikoff, ed.), Aca- demic Press, San Diego, California. Uetz, G. W. 1989. The “ricochet effect” and prey capture in colonial spiders. Oecologia, 81:154-159. Uetz, G. W. & C. S. Heiber. 1994. Group size and predation risk in colonial web-building spiders: analysis of attack abatement mechanisms. Behav. Ecol, 5:326-332. Vollrath, F. 1987. Kleptobiosis in spiders. Pp. 274- 286, In Ecophysiology of Spiders. (W. Nentwig, ed.) Springer- Verlag, Berlin. Wilson, E. O. 1975. Sociobiology - The new synthe- sis. Harvard Univ. Press, Cambridge, Massachu- setts. 697 pp. Manuscript received 26 November 1994, revised 1 De- cember 1994. 1995. The Journal of Aradmology 23:9-12 NEW SPECIES AND RECORDS OF THE GROUND SPIDER FAMILY GALLIENIELLIDAE (ARANEAE, GNAPHOSOIDEA) FROM MADAGASCAR Norman I. Platnick: Department of Entomology, American Museum of Natural History; New York, New York 10024 USA ABSTRACT. The females of Gailieniella bland Platnick and Legendrena perinet Platnick are described for the first time, as are two new species: Legendrena rothi and Legendrena spiralis. Recent fieldwork in Madagascar by several colleagues has resulted in much new information on the unusual ground spiders of the family Gal- lieniellidae, once thought to be endemic to that island but now known also from the Comoro Islands, Africa, and Australia (Platnick 1990a). As the Malagasy members of the family have been treated by Platnick (1984, 1990b, 1993), it seemed best to update that coverage by supplying new records and describing newly discovered taxa. Material was kindly made available from the collections of the California Academy of Sci- ences, San Francisco (CAS), by Charles Griswold and Darrell Ubick, the Museum of Comparative Zoology, Harvard University (MCZ), by Herbert Levi and Laura Leibensperger, and the National Museum of Natural History, Smithsonian Insti- tution (USNM) by Jonathan Coddington and Scott Larcher. The illustrations are by Moham- mad Shadab of the American Museum of Nat- ural History. The format of the descriptions fol- lows that of the original revision (Platnick 1 984); all measurements are in mm. Gailieniella mygaloides Millot New records. —MADAGASCAR. Fianarant - soa: Maharira summit, Ranomafana National Park, Apr. 9, 1992 (Albert, MCZ), 19; Rano- mafana National Park, Apr. 1992, pitfall trap (V., B. Roth, CAS), 161$. Gailieniella bland Platnick (Figs. 1, 2) Note.— -The female here assigned to this spe- cies was not taken with a male, but is paired on the basis of genitalic similarities of both males and females to those of G. mygaloides (Figs. 1 , 2). Diagnosis.— The newly described female can easily be separated from those of the other known species by the relatively large spermathecae, which approach the coiled anterior ducts in size. Female.— Total length, not including chelic- erae, 4.57. Carapace 2.00 long, 1.92 wide, dark chestnut brown; pars thoracica with recumbent white scales. From above, anterior eye row re- curved, posterior row slightly recurved; from front, both rows very slightly procurved; eye sizes and interdistances: AME 0.06, ALE 0.08, PME 0.06, PLE 0.05; AME-AME 0.13, AME-ALE 0.03, PME-PME 0. 1 5, PME-PLE 0.06, ALE-PLE 0.07; MOQ length 0.20, front width 0.25, back width 0.27. Clypeal height at AME about 1.6 times their diameter. Chelicerae extending for- ward distance about one-third of carapace length, bearing long fang without distinct ventral tuber- cle but abruptly narrowed at about one-third its length, narrow portion much paler than thicker portion; most distal promarginal tooth widely separated from other two subequal teeth, all pro- marginal teeth larger than two widely separated retromarginal teeth. Leg spination: femora II I V d 1 -0-0; tibiae: III p0-l-0, v2-2-0, r0-l-0; IV pO- 0-1, v4-2-2, r0-0- 1 ; metatarsus IV vl p-0-0. Legs light brown except sides of femora I, II, IV light yellow. Abdomen dark gray, dorsum with two longitudinal white stripes at sides, stripes con- nected posteriorly by seven chevrons; venter pal- er than sides. Palpal femur and more distal seg- ments with dorsal spines grading into bristles, tibia and tarsus each with proximal prolateral spine as well. Spermathecae almost as large as anterior, coiled epigynal ducts (Figs. 1 , 2). New record. —MADAGASCAR. Tohara: Ma- hafaly, nr. Eleotse, by Lac Tsimanampetsoa, 24°10'S, 43°45'E, Sept. 15-16, 1992 (V.,B. Roth, CAS), 1$. 9 10 THE JOURNAL OF ARACHNOLOGY Figures 1-4.- 1, 2. Gallieniella hlanci Platnick, epigynum: 1, ventral view; 2, dorsal view. 3, 4. Legendrena perinet Platnick, epigynum: 3, ventral view; 4, dorsal view. Legendrena perinet Platnick (Figs. 3, 4) Diagnosis.™ Females can be recognized easily by the anteriorly curled epigynal ducts (Figs. 3, 4). Female.— Total length, not including chelic- erae, 3.50. Carapace 1.84 long, 1.52 wide, light brown, without scales. From above, anterior eye row recurved, posterior row slightly recurved; from front, anterior row very slightly recurved, posterior row slightly procurved; eye sizes and interdistances: AME 0.06, ALE 0.09, PME 0.08, PLE 0.09; AME- AME 0.06, AME- ALE 0.01, PME-PME 0.11, PME-PLE 0.04, ALE-PLE 0.03; MOQ length 0.19, front width 0.18, back width 0.27. Clypeal height at AME only slightly greater than their diameter. Chelicerae extending for- ward distance greater than one-third of carapace length, bearing relatively short fang without ven- tral tubercle but slightly narrowed, lightened at about one-third its length; three promarginal teeth closely spaced, middle one largest; two retro- marginal teeth enlarged, widely separated. Leg spination: femora I-IV d 1-0-0. Femora, patellae, and tibiae light brown except distal half of tibia I and distal tip of tibiae III, IV lightened; meta- tarsi and tarsi light orange. Abdomen dark gray, dorsum unmarked, venter with two pale longi- tudinal stripes near sides. Palpal femur and more distal segments with dorsal spines grading into bristles, tibia and tarsus each with proximal pro- lateral spine as well. Epigynum with pair of an- terolateral pockets (Fig. 3), ducts curling ante- riorly (Fig. 4). New records. —MADAGASCAR. Fianarant- soa : Ranomafana National Park, Apr. 1992, pit- fall trap (V., B. Roth, CAS), 363$; 200 m N re- search cabin, Ranomafana National Park, Mar. 25, 1 992, in leaf litter (S. Kariko, V. Roth, MCZ), 161$. Legendrena steineri Platnick New records.— MADAGASCAR. Fianarant- soa : 7 km W Ranomafana, Feb. 23-28, 1990, elev. 900 m, flight intercept-yellow pan trap in malaise trap in small clearing, montane rain for- est (W. E. Steiner, USNM), 16; Ranomafana Na- tional Park, Apr. 1992, pitfall trap (V., B. Roth, PLATNICK-MALAGASY GALLIENIELLIDAE 11 Figures 5-7. Legend rena rothi new species, left male palp: 5, prolateral view; 6, ventral view; 7, retrolateral view. CAS), 1612, May 1992 (S. Kariko; V., B. Roth, CAS), 12; Ranomafana National Park, ca. 21°12'S, 47°27'E, Mar.-Apr. 1992, forest foliage (V., B. Roth, MCZ), 12, Apr. 1992, forest (V., B. Roth; S. Kariko, CAS), 12. Legendrena rothi new species (Figs. 5-7) Type.— Male holotype from Claire’s camp, Vatoaranana, Ranomafana National Park, Fian- arantsoa, Madagascar (May 9, 1992; V. Roth), deposited in CAS. Etymology. —The specific name is a patronym in honor of the collector. Diagnosis.— The basally incrassate tibia I, pal- pal tibial apophysis bearing cusps, and elongate embolus indicate that this is the sister species of L. steineri, from which it differs in having the embolus originating distally, rather than proxi- mally, on the palpal bulb (Figs. 5-7). Male.— Total length, not including chelicerae, 5.19. Carapace 2.14 long, 1.62 wide, dark chest- nut brown, without scales. From above, anterior eye row recurved, posterior row slightly re- curved; from front, both rows very slightly pro- curved; eye sizes and interdistances: AME 0.07, ALE 0.09, PME 0.08, PLE 0.07; AME-AME 0.08, AME- ALE 0.03, PME-PME 0.14, PME-PLE 0.04, ALE-PLE 0.04; MOQ length 0.18, front width 0.22, back width 0.30. Clypeal height at AME only slightly greater than their diameter. Chelicerae extending forward distance less than one-third of carapace length, bearing relatively short fang without ventral tubercle but slightly narrowed, lightened at about half its length; three promarginal teeth closely spaced, middle one largest; two retromarginal teeth enlarged, widely separated. Leg spination: femora III, IV d 1-0-0. Coxae and trochanters yellow; femora dark gray, I— III with dorsal yellow longitudinal stripe; pa- tella I dark gray, II, III yellow with lateral dark stripes; IV yellow; tibia I incrassate, dark gray, with ventral fringe of long setae proximally, yel- low distally, other tibiae yellow with dark lateral longitudinal stripes; anterior metatarsi and tarsi light orange, posteriors yellow. Abdomen dark gray, dorsum without distinct scutum, with vaguely indicated transverse white stripe at about one-third its length. Palpal tibia with elongate dorsal apophysis bearing retrolateral row of cusps; embolus originating distally, extending almost to tip of elongated cymbium (Figs. 5-7). Female. —Unknown. Distribution.— Known only from southeastern Madagascar. Legendrena spiralis new species (Figs. 8-10) Type.— Male holotype taken on foliage along Namorona River, Ranomafana National Park, 12 THE JOURNAL OF ARACHNOLOGY Figures 8=10 .—Legendrena spiralis new species, left male palp: 8, prolateral view; 9, ventral view; 10, retro- lateral view. Fianarantsoa, Madagascar (May 18, 1992; B. Roth), deposited in CAS. Etymology.— The specific name refers to the coiled embolus. Diagnosis. —Males can easily be separated from those of the other known species by the distally coiled embolus (Figs. 8=10). Male.— Total length, not including chelicerae, 3.89. Carapace 1.73 long, 1.46 wide, dark chest- nut brown, without scales. From above, anterior eye row recurved, posterior row slightly re- curved; from front, anterior row slightly re- curved, posterior row almost straight; eye sizes and interdistances: AME 0.06, ALE 0.06, PME 0.08, PLE 0.07; AME-AME 0.05, AME-ALE 0.02, PME-PME 0. 1 5, PME-PLE 0.06, ALE-PLE 0.07; MOQ length 0.19, front width 0.23, back width 0.25. Clypeal height at AME only slightly greater than their diameter. Chelicerae extending forward distance less than one-third of carapace length, bearing relatively short fang without ven- tral tubercle but slightly narrowed, lightened at about half its length; three promarginal teeth closely spaced, middle one largest; two retro- marginal teeth widely separated but not enlarged. Leg spination: femora I-IV d 1-0-0. Coxae and trochanters yellow, I— III with dark lateral stripes; femora dark gray, I with lateral light stripes along ventral half of sides, II, III with dorsal yellow longitudinal stripe; patella I dark gray, II, III yellow with lateral dark stripes, IV yellow prox- imally, dark gray distally; tibia I not incrassate or fringed, dark gray proximally, yellow distally, other tibiae dark orange with dark lateral lon- gitudinal stripes; anterior metatarsi and tarsi light orange, posteriors yellow. Abdomen dark gray, dorsum without distinct scutum, with vaguely indicated transverse white stripe at about one- third its length. Palpal tibia with short, hook- shaped retrolateral tibial apophysis; embolus originating distally, broadened proximally, coil- ing at tip of cymbium (Figs. 8-10). Female. —Unknown. Distribution.— Known only from southeastern Madagascar. LITERATURE CITED Platnick, N. I. 1984. Studies on Malagasy spiders, 1 . The family Gallieniellidae (Araneae, Gnaphoso- idea). American Mus. Novit, 2801:1-17. Platnick, N. I. 1990a. Spinneret morphology and the phylogeny of ground spiders (Araneae, Gnaphoso- idea). American Mus. Novit., 2978:1-42. Platnick, N. I. 1990b. A new species of Legendrena (Araneae: Gallieniellidae) from Madagascar. J. New York Ent. Soc., 98:499=501. Platnick, N. I. 1993. The female of Gallieniella be - troka (Araneae, Gallieniellidae). J. Arachnol., 21: 152. Manuscript received 10 November 1994, revised 12 De- cember 1994. 1995. The Journal of Arachnology 23:13-16 ON THE SPIDER GENU'S HEBRITHELE (ARANEAE, MITURGIDAE) Norman I, Plat nick: Department of Entomology, American Museum of Natural History; New York, New York 10024 USA Alexandre B* Bonaldo: Museu de Ciencias Naturais, Funda^ao Zoobotanica do Rio Grande do Sul, C. Postal 1188, CEP 90001-970, Porto Alegre RS, Brazil ABSTRACT. The genus He'd ri.ih.de has been known only from the female holotype of its type species, Hebritheie iongicauda Berland. The male of the species was apparently described earlier but was misplaced in the family Gnaphosidae. Hypodrassodes insulanus (Rainbow) is transferred to the Miturgidae and considered a senior synonym of Hebritheie Iongicauda . The spider genus Hebritheie was established by Berland (1938) on the basis of a single female from the New Hebrides. Berland considered the spider of ambiguous relationship, noting that the widely separated anterior lateral spinnerets sug- gested a placement in the Gnaphosidae, where the enlarged anterior median eyes suggested pos- sible relationships to such genera as Leptodrassus Simon or Anzacia Dalmas. However, the pres- ence of a long distal segment on the posterior lateral spinnerets made Berland hesitant to as- sign the genus to the Gnaphosidae, as gnaphosids do not share that feature. He therefore assigned the genus to the Clubionidae, suggesting a close relationship to such New World genera as Eu- tichurus Simon and Strotarchus Simon, both of which are currently assigned to the Miturgidae. Hebritheie was itself transferred to the Miturgi- dae by Lehtieen (1967), who did not assign the genus to any of the six miturgid subfamilies that he recognized (although he suggested a possible placement in the otherwise New World Eutichu- rinae). During a recent visit to the B. P. Bishop Mu- seum in Honolulu, the first author encountered, among Pacific spiders identified as gnaphosids, a male from the New Hebrides that seemed to belong to Hebritheie. A subsequent literature search indicated that this male was probably first described by Rainbow (1901) as the gnaphosid Leptodrassus insulanus . The holotype of that species, unfortunately, appears to be lost; it is not in the collection of the Australian Museum, Sydney, even though the remaining types from that paper are housed there (Dr. M. R. Gray, in litt.). Nevertheless, Rainbow’s figure of the male palp is detailed enough to leave few doubts about its identity with the Bishop Museum specimen. Rainbow’s species was later transferred to the gnaphosid genus Anzacia Dalmas by Dalmas (1919). For unknown reasons, the species was transferred by Roewer (1955) to the gnaphosid genus Hypodrassodes Dalmas. Examination of both sexes leads us to concur with Lehtinen on the placement of Hebritheie within the Miturgidae. We do not concur with his tentative placement of the genus in the sub- family Eutichurinae, however, as the thoracic groove is well demarcated (rather than reduced or absent) and the retrolateral margin of the cym- bium has an incision, rather than the projection typical of eutichurines (see Bonaldo 1994). Both the incised cymbium and the widely separated anterior lateral spinnerets suggest a placement in the Miturginae, closer to the American genera Teminius Keyserling (see Platnick & Shadab 1989) and Strotarchus Simon (transferred from the Eutichurinae to the Miturginae by Bonaldo 1994), and the Australasian genus Miturga Tbo rell, rather than in the Eutichurinae. All measurements are in millimeters. Hebritheie Berland Hebritheie Berland 1938: 137 (type species by original designation Hebritheie Iongicauda Berland). Diagnosis. — The combined presence of an elongated distal segment on the posterior lateral spinnerets, widely separated anterior lateral spinnerets, a well-marked thoracic groove (Figs. 13 Figures 1, 2 .—Hebrithele insulana (Rainbow), carapace, dorsal view. 1, male; 2, female. 1, 2), an incised cymbial margin (Fig. 4), a palpal bulb lacking a median apophysis, and large, ovoid spermathecae (Fig. 6) is diagnostic of the genus. Description.— Araneomorph, ecribellate, en- telegyne spiders. Carapace widest between coxae II and III, narrowed opposite palpal insertion, light brownish orange; cephalic area flattened, thoracic groove longitudinal, long, occupying over one-sixth of carapace length; ocular area and clypeus with numerous weak, white setae and several strong bristles. From above, anterior eye row recurved, posterior row procurved; from front, both rows procurved; all eyes circular, AME much larger than others; PME and PLE sub- equal, smaller than ALE; AME separated by less than their radius, by less than their radius from ALE; PME separated by about twice their di- ameter, by more than their diameter from PLE; ALE and PLE almost contiguous; MOQ about as wide in front as in back, wider than long; clypeal height slightly greater than AME diam- eter; chilum present as distinct triangular sclerite. Chelicerae usually with three promarginal teeth, median one largest, situated near tip of fang fur- row; three smaller retromarginal teeth situated closer to base of fang, with tooth closest to fang smaller than others (or missing). Mouthparts and sternum light brown, darkest at base of labium and endites; endites distally squared in males (distal margin of female endites more rounded), without oblique depressions, with strong serrula, extending far beyond labium; labium only slight- ly longer than wide, invaginated at posterolateral comers; sternum shield-shaped, not rebordered, with sclerotized extensions to each coxa and be- tween coxae I and II. Leg formula 4123; legs light brownish orange; tarsi with two dentate claws and conspicuous claw tufts; trochanters deeply notched; trichobothria present on tibiae, meta- tarsi, and tarsi. Abdomen brownish gray, coated with strong, dark setae; males without dorsal scu- tum; anterior lateral spinnerets elongated, sep- arated at base by almost their diameter, with distinct distal segment bearing one major am- pullate gland spigot and several piriform gland spigots not enlarged in either sex; posterior me- dian spinnerets short, tubular in both sexes; pos- terior lateral spinnerets with two long segments; colulus represented by wide setose area of cuticle. Male palp with femur and patella unmodified; tibia with retrolaterally directed retrolateral apophysis; bulb with strong embolus and mem- branous conductor, without median apophysis. Epigynum wide, heavily sclerotized posteriorly. Hebrithele insulana (Rainbow) new combination (Figs. 1-6) Leptodrassus insulanus Rainbow 1901: 523, pi. XXVIII, figs. 1, la (male holotype from Malekula Is., New PLATNICK & BONALDO -HEBRITHELE 15 Figures 3-6 .—Hebrithele insulana (Rainbow). 3, left male palp, ventral view; 4, same, retrolateral view; 5, epigynum, ventral view; 6, same, dorsal view. Hebrides, should be in the Australian Museum, Syd- ney, lost). Anzacia insulana : Dalmas 1919: 249. Hebrithele longicauda Berland 1938: 137, figs. 21-25 (female holotype from Malekula Is., New Hebrides, in MNHN, examined). NEW SYNONYMY. Hypodrassodes insulanus: Roewer 1955: 404. Diagnosis.— The laterally directed retrolateral tibial apophysis of males (Fig. 3) and the arched anterior epigynal margin situated at only half the length of the spermathecae of females (Fig. 5) are presumably diagnostic. Male.— Total length 7.60. Carapace 3.55 long, 2.75 wide. Eye diameters and interdistances: AME 0.25, ALE 0.15, PME 0.15, PLE 0.17; AME-AME 0.07, AME-ALE 0.05, PME-PME 0.25, PME-PLE 0.22, ALE-PLE 0.05; MOQ length 0.45, front width 0.55, back width 0.52. Chelicerae with three promarginal and three ret- romarginal teeth. Abdomen 4.05 long, 2. 1 5 wide; posterior lateral spinneret proximal segment 0.75 long, distal segment 0.52 long. Leg measure- ments (femur, patella, tibia, metatarsus, tarsus, total): I 3.10, 1.50, 2.65, 2.50, 1.25, 11.00; II 2.80, 1.45, 2.30, 2.30, 1.15, 10.00; III 2.50, 1.25, 1.80, 2.25, 1.05, 8.85; IV 3.30, 1.40, 2.70, 3.20, 1.25, 1 1.85. Leg spination (only surfaces bearing spines listed): femora: I dl-1-1, p0-0-2, r0-0-l; II dl-1-1, pO-1-2, rl-1-1; III, IV dl-1-1, pl-1- Ll, rl- 1-1-1; tibiae: I v2-2-lp; II v2-2-2; III, IV dlr-0-0, pl-1, v2-2-2, r 1 - 1 ; metatarsi: I v2-lr- lr; II v2-2-0; III, IV pl-2-2, v2-2-2, rl-2-2. Ret- rolateral tibial apophysis with tiny, triangular cusp on medial margin near tip (Fig. 3); embolus strong, arched, supported by membranous con- ductor (Fig. 4). Female.— Total length 7.90. Carapace 4.10 long, 3.05 wide. Eye diameters and interdist- ances: AME 0.22, ALE 0.20, PME 0.15, PLE 0.20; AME-AME 0.15, AME-ALE 0.07, PME- PME 0.32, PME-PLE 0.30, ALE-PLE 0.05; MOQ length 0.52, front width 0.60, back width 0.57. Chelicerae with two (right) or three (left) pro- marginal and three retromarginal teeth. Abdo- men 4.00 long, 2.30 wide; posterior lateral spin- neret proximal segment 0.80 long, distal segment 0.60 long. Leg measurements: I 3.20, 1.80, 2.75, metatarsi and tarsi missing; II 3.15, 1.65, 2.40, 2.25, 1.25, 10.70; III 2.80, 1.30, 1.85, 2.25, 0.95, 9.15; IV 3.50, 1.65, 2.85, 3.30, 1.25, 12.55. Leg 16 THE JOURNAL OF ARACHNOLOGY spination: femora: I, II dl-1-0; III dl-1-0, pi-1- 1, rl-1-1-1; IV dl-1-1, pl-1-1, rl-1-1-1; tibiae: I v2-2-0; II vlp-lp-0; III, IV dlr-0-0, pl-1, v2- 2-2, rl-1; metatarsi: I missing; II v2-0-0; III, IV pi -2-2, v2-2-2. r 1-2-2. Anterior epigynal margin bipartite, arched (Fig. 5), spermathecae large, oval (Fig. 6). Material examined.— NEW HEBRIDES. Epi Island : Lowekewou, Aug. 31, 1979, elev. 0-100 m (Barnes, Nishida, Gagne, Samuelson, BPBM), IS. Malekula Island : no specific locality, May 1934 (A. de la Rue, MNHN), 1$ (holotype). Distribution.— Known only from the New Hebrides. Synonymy . — The male and female differ slightly in endite shape, the relative width of the pars cephalica, and coloration; we attribute the first two differences to sexual dimorphism and the latter one to the holotype being a freshly molted specimen. Until additional species of the genus are found, these specimens are most par- simoniously considered conspecific. ACKNOWLEDGMENTS We thank Sabina Swift and Scott Miller of the Bishop Museum (BPBM) for access to the male, and Christine Rollard of the Museum National d’Histoire Naturelle, Paris (MNHN), for access to the type. LITERATURE CITED Berland, L. 1938. Araignees des Nouvelles-Hebrides. Ann. Soc. Ent. France, 107:121-190. Bonaldo, A. B. 1994. A subfamilia Eutichurinae na regiao neotropical, com a revisao do genero Euti- churus Simon, 1896 (Araneae, Miturgidae). Iher- ingia (Zool.), 76:101-159. Dalmas, R. de. 1919. Catalogue des araignees du gen- re Leptodrassus (Gnaphosidae), d’aprds les mater- iaux de la collection E. Simon au Museum National d’Histoire Naturelle. Bull. Mus. Natl. Hist. Nat., Paris, 1919:243-250. Lehtinen, P. T. 1967. Classification of the cribellate spiders and some allied families, with notes on the evolution of the suborder Araneomorpha. Ann. Zool. Fennici, 4:199-468. Platnick, N. I., & M. U. Shadab. 1989. A review of the spider genus Teminius (Araneae, Miturgidae). American Mus. Novit., 2963:1-12. Rainbow, W. J. 1901. Arachnida from the South Seas. Proc. Linn. Soc. New South Wales, 26:521-532. Roewer, C. F. 1955. Katalog der Araneae von 1758 bis 1940, bzw. 1954. Brussels, 2a-b: 1751 pp. Manuscript received 10 December 1 994, revised 10 Jan- uary 1995. 1995, The Journal of Arachnology 23:17-24 DISCRIMINACION POR METEPEIRA SEDITIOSA (KEYSERLING) (ARANEAE, ARANEIDAE) EN CONDICIONES EXPERIMENTALES SOBRE DOS PRESAS FRECUENTES EN EL MEDIO Carmen Viera: Division Zoologia Experimental, Institute de Investigaciones Biologicas Clemente Estable, Av. Italia 3318, Montevideo, Uruguay; Section Entomologia, Departamento de Biologia Animal, Facultad de Ciencias, Tristan Narvaja 1674, Montevideo, Uruguay ABSTRACT. The predatory behavior of Metepeira seditiosa on two prey organisms, Musca sp. and Acro- myrmex sp., was compared under experimental conditions. Frequency diagrams for various behaviors were constructed, and the stereotypy and relationship among the units in the sucession were established. Discrimi- nation between the prey organisms occurred in the detection and immobilization phases. Metepeira seditiosa has a vast repertory of behavior. The tactics used for capturing Musca sp. and Acromyrmex sp. were 100% successful. RESUMEN. Se describieron y analizaron las secuencias de unidades de comportamiento de captura de Me- tepeira seditiosa frente a dos presas {Musca sp. y Acromyrmex sp.) en condiciones experimentales. Se realizaron diagramas de frecuencias, estableciendose las relaciones entre unidades y la presencia de estereotipia en las sucesiones. Se comprobo la discrimination entre las presas mediante la comparacion de las Fases de Detection e Inmovilizacion. Metepeira seditiosa posee una amplia gama de unidades de comportamiento con las cuales selecciona la tactica depredadora adecuada ante Musca sp. y Acromyrmex sp., obteniendo un 100% de exito en las capturas. Estudios realizados por Riechert y Luczak (1982), Stowe (1986) y Nentwig (1987), entre otros, han demostrado que las aranas orbitelares exhiben una especializacion considerable en la dieta. Peters (1931, 1933) realize estudios sobre la capacidad que poseee las aranas de distinguir entre diferentes categorias de insectos: Observe que Araneus diadematus Clerck atacaba con di- ferente tactica a moscas vibrantes y moscas in- moviles. Robinson y Robinson (1976) observa ron, en por lo menos doce especies de araneidos, discrimination entre lepidopteros y otros insec- tos; Viera (1981) observe en tres especies dife- rentes de insectos discrimination en la tactica de ataque y en la eficiencia en la captura de Alpaida alticeps (Araneidae). En cuanto a lo que se sabe sobre Metepeira Burgess y Witt (1976) analizaron el diseno de las redes de Metepeira spinipes F. Cambridge y Me- tepeira labyrinthea (Hentz) que son similares a las telas de M, seditiosa. Viera (1986, 1989) des- cribio cualitativamente y cuantitativamente la red de M. seditiosa. Viera y Costa (1 985) y Viera (1986, 1994) hicieron aportes sobre el compor- tamiento de captura. Aim no se han realizado estudios sobre el comportamiento discrimina- torio de presas. Acorde con observaciones realizadas en el campo, por la aurora, fueron seleccionados dos representantes de los ordenes Diptera e Hyme- noptera como las presas mas frecuentes obser- vadas en las redes y capturadas por M. seditiosa para ser entregadas en condiciones experimen- tales. METODOS Para este estudio se recolectaron 53 ejemplares en Punta Espinillo (Montevideo, Uruguay) en abri! de 1990. Las aranas fueron criadas en el laboratorio en recipientes individuales de vidrio transparente de 14 cm de altura y 9 cm de dia- metro, cubierto con una malla de nailon, con un recipiente con agua y un bastidor de madera para soporte de la tela. En los periodos interexperi- mentales los ejemplares fueron alimentados con trozos de larva s de Tenebrio sp. (Coleoptera). Durante el periodo de cria y experimentation la temperatura media diaria fue 22.5 ± 2.67 °C 17 18 THE JOURNAL OF ARACHNOLOGY y el fotoperiodo de 12 h luz y 12 h oscuridad. Las presas utilizadas fueron Musca domestica (Diptera) y Acromyrmex sp. (Hymenoptera). El tamano de todas las presas fue similar o ligera- mente inferior al de la arana. Para las experiencias se utilizaron solo indi- viduos juveniles grandes y hembras adultas; los machos adultos se descartaron por su incapaci- dad para construir telas orbiculares (Viera y Cos- ta 1985). Se formaron dos grupos de 7-8 aranas cada uno, utilizandose una unica vez frente a cada presa. Las aranas se trasladaron, cinco dias antes de cada experiencia, a recipientes de vidrio transparente de 30 cm de altura, 30 cm de ancho y 10 cm de fondo, con una cara movil y un bas- tidor con un soporte central para sostener la tela y un recipiente con agua. Las observaciones se hicieron desde la entrega de la presa hasta la ingestion, abandono de la misma, o inmovilidad total de la arana por mas de 30 min. Las presas fueron entregadas siempre en el mismo lugar aproximado de la tela (zona inferior, ligeramente a la derecha), segun el siguiente cronograma: En la primera semana de experimentation se le en- trego al grupo A la presa Musca sp. En la segunda semana se le entrego al grupo B la presa Acro- myrmex sp. En la tercera semana se le entrego al grupo B la presa Musca sp. En la cuarta semana se le entrego al grupo A la presa Acromyrmex sp. En las observaciones, se coloco una cartulina negra detras del recipiente y se ilumino lateral- mente. Las aranas permanecieron dos dias des- pues de la observation, en los recipientes de ex- perimentacion para controlar la elimination de restos de presas y/o reparation de la tela. Todas las experiencias fueron relatadas y registradas en un grabador magnetofonico, midiendose la du- ration de las unidades de comportamiento. El comportamiento de captura se dividio en 1 6 uni- dades de comportamiento agrupadas en tres fa- ses: a) de Detection, b) Inmovilizacion, y c) Ter- minal. Se utilizaron las unidades Tensamiento, Envoi vimiento, Corte de Hilos, Transporte y Manipulation Preingestiva descriptas por Rob- inson y Olazarri (1971). Las unidades Despla- zamiento I, Desplazamiento II, Quietud, Aci- calamiento, Toqueteo, Mordeduras Cortas y Mordedura Prolongada fueron descriptas por Viera (1983, 1986). Las unidades Fijacion y Giro, Fijacion de Hilos, Otros Desplazamientos y Re- cuperation de la Presa fueron descriptas por Viera (1994). En el analisis estadistico se utilizo el test de diferencias de medias ( t de Student), con restric- ciones para la varianza (F de Snedecor) y el pa- quete estadistico PRESTA. Los ejemplares estudiados se depositaron en la coleccion aracnologica del Departamento de Entomologia de la Facultad de Ciencias, Mon- tevideo. RESULTADOS En la captura frente a Acromyrmex sp. (Fig. 1) la Fase de Detection se initio mayoritariamente con Tensamiento (14 en 15) que realizaron en el refugio y luego en el centre de la tela. La Fase de Inmovilizacion se initio exclusivamente con la unidad Envolvimiento que fue sucedida fre- cuentemente por Mordeduras Cortas. Envolvi- miento y Mordeduras Cortas se vincularon en menor medida con Mordedura Prolongada. La Fase de Inmovilizacion se vinculo con la Fase Final por medio de la sucesion Envolvimiento - Corte de Hilos. La unidad Acicalamiento tuvo una frecuencia alta (9 en 1 5), relacidnandose con las Fases de Inmovilizacion y Final. La unidad Otros Desplazamientos tuvo tambien una fre- cuencia alta (12 en 15). La unidad Quietud fue la mas frecuente (27 veces). En la captura frente a Musca sp. (Fig. 2) rea- lizaron Tensamiento en el refugio ( 1 3 en 1 5) como primera unidad de la Fase de Detection y luego en el centre de la tela. Se observo una alteration a esa sucesion en cinco individuos que no rea- lizaron Tensamiento en el centre de la tela y luego de Desplazamiento II, realizaron Toqueteo. El pasaje de la Fase de Detection a la Fase de In- movilizacion se realizo desde Toqueteo. La pri- mera unidad de esta Fase fue mayoritariamente Mordeduras Cortas. Se observo una vinculacion menor entre la dupla Envolvimiento - Morde- duras Cortas que entre Envolvimiento - Mor- dedura Prolongada solo en esa direction. Corte de hilos se vinculo principalmente con Envolvimiento. Dentro de la Fase Final se ob- serve una marcada estereotipia en la sucesion de unidades. La unidad Quietud se vinculo princi- palmente con Envolvimiento. Aunque se obser- vo 23 veces, las sucesiones fueron multiples, pero con frecuencias menores al 1% del total de las unidades, no apareciendo por ese motive en el diagrama. La unidad Acicalamiento se observo ocho veces y se vinculo a unidades de la Fase de Inmovilizacion. Analisis comparativo del comportamiento de captura: (Figs. 1, 2). Se compararon los com- portamientos de captura ante la misma presa VIERA — DISCRIMINACION DE PRESAS POR METEPEIRA SEDITIOSA 19 r DESPLAZAMIENTO 1 14 DESPLAZAMIENTO 2 12 13 TOQUETEO TENSAMIENTO 5 V GIRO- RECUPERA FIJACIONES PRESA t MANIPULACION Figura 1 . — Diagrama de frecuencias de la captura frente a Acromyrmex sp n= 15, numero total de unidades = 430. Se excluyeron de la figura las frecuencias inferiores o iguales al 1% del total de las unidades. entre los dos grupos experimentales utilizando las siguientes variables: a) Tiempo de latencia (periodo que va desde la entrega de la presa hasta la primera respuesta de la arana); b) Duracion total del comportamiento de captura; c) Numero total de unidades; d) Duracion de la Fase de Deteccion; e) Numero de unidades que compo- nen la Fase de Deteccion; f) Duracion de la Fase de Inmovilizacion; g) Numero de unidades que componen la Fase de Inmovilizacion; h) Dura- cion de las unidades que componen la Fase Final; i) Numero de las unidades que componen la Fase Final. En las Tabla 1 se pueden observar diferencias estadisticamente significativas entre las dura- ciones totales de la captura frente a Musca sp. y Acromyrmex sp. Asimismo se observaron diferencias estadis- ticamente significativas en la duracion de la Fase de Inmovilizacion entre las capturas de Musca sp. y Acromyrmex sp. Para comprobar que la entrega aleatoria de 20 THE JOURNAL OF ARACHNOLOGY Figura 2.— Diagrama de frecuencias de la captura frente a Musca sp.; n = 15, numero total de unidades = 340= Se excluyeron de la figura las frecuencias inferiores o iguales al 1% del total de las unidades. presas a los grupos de individuos A y B descar- tarian la posible influencia del aprendizaje en el comportamiento de captura se realizaron test de diferencias de medias con restricciones para la varianza en los dos grupos de experimentacion frente a cada tipo de presa. Las variables utili- zadas en la comparacion fueron: tiempo de la- tencia, duracion total de la captura y numero total de unidades. El test no refleja diferencias estadisticamente significativas (Tabla 2). Se observe) un 100% de exito en la captura de ambas presas. DISCUSION Viera (1983) y Viera y Costa (1985) compro- baron experimentalmente que el comportamien- VIERA — DISCRIMIHACIOM DE PRESAS POR METEPEIRA SEDITIOSA 21 Tabla 1.— Valores para nueve variables del comportamient© deltotal de los individuos frente a Acmmyrmex sp. y Musca sp. (n = 1 5). Los tiempos estan dados en segundos, Valores de t de Student y probabilidades en la computation. Acmmyrmex Musca X DT X DT t P 1. Tiempo de lateneia 2. Duration total 98.53 149.41 212.27 325.74 1.187 0.248 de la captura 3. Numero total 633.00 421.17 345.00 243.83 2.214 0.035 de unidades 4. Duration de 28.67 11.88 22.86 9.16 1.446 1.557 detection 5. Numero de unidades 33.13 32.47 57.93 99.37 0.888 0.609 de detection 6. Duration de 3.87 1.63 2.27 1.61 2.614 0.014 inmovilizacion 7. Numero de unidades 436.60 351.20 188.80 118.30 2.502 0.021 de inmovilizacion 8. Duration de fase 13.13 6.60 7.33 6.29 2.816 0.009 final 9. Numero de unidades 89.53 81.18 102.13 69.23 0.463 0.651 de fase final 1.67 1.01 2.13 0.88 1.300 0.202 to de captura en hembras a dull as y juveniles de Metepeira seditiosa era similar. Esto justified la utilization de ambos conjuntos como un grupo homogeneo. Robinson y Robinson (1976) demostraron la influencia del apreedizaje en un segundo en- cuentro con la misma presa. La ausencia de di ferencias estadisticamente significativas en la captura sobre la misma presa en Metepeira se- ditiosa pet ini t io descartar la posibilidad de que, las capturas previas de cada arafia, modificaran la siguiente; asi como el aprendizaje en indivi- duos que capturaron el mismo tipo de presa. El aprendizaje previo en el laboratorio, fue descar- tado, ya que se utilize cada arafia una unica vez frente a cada presa. Sin embargo, las arafias po- drfan haber tenido experiencias con estas presas en el campo. Considerando las diferencias en la captura de presas con las diferencias de tamafio observadas por Robinson y Mirick (197 1) y Japyassu y Ades (1990) y Uetz (1990) quien formula la hipotesis que, la discrimination en algunas arafias de la familia Araneidae aetua solo sobre el tamano de la presa y no sobre la naturaleza de la misma, se descartd la influencia del factor tamano, selec- tion an do presas de tamano y peso homogeneo en relation a las arafias. Bristowe (1941) use tres especies de hormigas del genero Acanthomyrmex como presas frente a 52 especies de arafias y comprobo que solo diez especies las capturaron. Debido a sus caracteres defensives (espinas, mandibulas y acido formi- co) la convierten en una presa peligrosa para un acercamiento directo. La peligrosidad de las hor- migas condicionaria el no ser aceptadas facil- mente como presas para algunas arafias. M. se- ditiosa obtuvo un 100% de exito en su captura, Tabla 2. —Las variables utilizadas en la comparacion fueron: lateneia, duration total de la captura y numero total de unidades. El test no refleja diferencias estadisticamente significativas. Tiempo de lateneia Duration total No unidades T P T P T P Acmmyrmex 1.068 0.366 0.255 0.800 0.175 0.859 Musca 0.230 0.818 0.483 0.645 0.767 0.530 22 THE JOURNAL OF ARACHNOLOGY pero le implied ataques de mas duration y de mayor cantidad de componentes, demostrando una mayor cautela que ante Musca sp. Eisner y Dean (1976) observaron que Nephila y Argiope frente a los escarabajos bombarderos realizaron Envolvimiento como prim era unidad de la Ease de Inmovilizaeion. Esta tactica de ataque coincidio con lo observado en Metepeira seditiosa frente a Acmmyrmex sp. La election de utilizar Envolvimiento como primera unidad en la tactica de ataque le permitiria a Metepeira seditiosa inmovilizar su presa y emponzonarla sin riesgos, debido a que evita un contact© es- treefao. Dicfaa tactica es usada sobre todas las presas que muerden, picae o esparcen fluidos nocivos cuando son atacadas (Robinson y Rob- inson 1981). La unidad con la que se inicia la Ease de Inmovilizaeion de la presa (Mordeduras Corns o Envolvimiento) indicaria la eficacia en la discrimination y la consiguiente tactica a uti- lizar. Hays (1985) observe que la presa Musca sp. utiliza como tactica antipredadora permanecer en quietud, mientras la arana realiza la Ease de Detection y cuando el acercamieeto de la arana es ieminente, se debate violentamente, intentan- do soltarse de la red. Nuestras observaciones coincidieron con Hays (1985). El mecanismo de- fensive utilizado por la mosca le permitiria con- fundirse con un objeto inmovil. Robinson y Robinson ( 1 975) y Harwood (1974) en Araneidae y Ades et a.L (1990) y Yoshida (1990) en Tetragnathidae ban observado la uti- lization de la unidad Mordeduras Cortas como inicio del ataque. Metepeira seditiosa utilize esta tactica frente a Musca sp.,. ademas de una mayor velocidad de respuesta observada en la captura. Este comportamiento resultaria adaptativo, de- bido al menor grade de adherencia de Musca sp. a la red, lo cual fue analizado por Eisner et al. (1964). Mediante la Ease de Detection la arana recibe information acerca del tipo de presa ra- pid© e inofensivo, lo que precipitaria el ataque que se initio con Mordeduras Cortas. Robinson et al. (1969) observaron que Argiope argentata y A. aemula discriminaron entre varias presas utilizando la sucesion Mordedura - Envolvi- miento para el ataque de presas rdativamente inocuas pero rapidas para escapar, y la sucesion Envolvimiento - Mordeduras para presas peli- grosas y fuertes. Dicha tactica, que favorece la discrimination de presas, fue observada asi- mismo en Metepeira seditiosa. Suter (1978) con state que Cyclosa turbinata utiliza principal mente la sucesion Mordeduras - Envolvimiento. Posiblemente esto se deba a la abundancia de presas rapidas e inofensivas en el habitat de estas aranas, lo que determina que esta tactica resulte eficiente. El aspecto funcional de la unidad Envolvimiento fue tratado extensa- mente por Robinson et al. (1969) y Robinson y Robinson (1975), quienes afirmaron que la ini- tiation con Envolvimiento de la Ease de In- movilizacion serfa un caracter altamente evol ucionado, presente en gran parte de la familia Araneidae. Las especies estudiadas cumplen una estrategia mas especialista que Metepeira sedi- tiosa. M. seditiosa es capaz de aceptar diferentes tipos de presas y capturarlas con exito. La alta correlation positiva entre el numero de unidades de la Ease de Inmovilizaeion y la duration de la misma, permitiria afirmar que el modelo de comportamiento de Metepeira sedi- tiosa tiene un repertorio limitado de unidades, y que la complejidad de dieho comportamiento resulta del aumento en la frecuencia de las uni- dades y no de la duration de las mismas. La duration y la complejidad de la Ease de Inmov- ilizacion nos informa acerca de las dificultades que la arana tiene para capturar, dependiendo de las caracteristicas de la presa. La Ease Final fue la mas estereotipada de las tres fases. Esta fase es menos importante para la discrimination, ya que las presas inmovilizadas no presentaron diferencias sustanciales entre si. No se observe Transporte de las presas en los queliceros y suponemos, de acuerdo con Rob- inson y Olazarri (1971), que se debio al tamano rdativamente grande de las presas. El Transporte en una pata serviria para mantener la presa ale- jada de la red cuando la arana se desplaza por ella (Robinson y Robinson 1970). En trabajos de campo, Pasquet (1984) y Pas- quet y Leborgne (1986, 1988) encontraron se- lection de presas en cinco especies de Araneidae. No pudimos comprobar en M. seditiosa si existio preferencia por alguna de las presas, debido a que el consume de alimento en el laboratorio es frecuentemente mas elevado que en el campo (Hagstram 1970; Breymeyer y Jowik 1975). El presente estudio mostro que Metepeira se- ditiosa posee una ampia gama de unidades de captura seleccionando la tactica depredadora adecuada para las diferentes tacticas defensivas ante Musca sp. y Acmmyrmex sp., obteniendo un 100% de exito de captura. Esta potencialidad depredadora de Metepeira seditiosa la senala como una especie potencialmente util para el VIERA — DISCRIMINACION DE PRESAS POR METEPEIRA SEDITIOSA 23 control biologico de insectos de interes econ- omic© como dipteros vectores y hormigas cor- tadoras, estas ultimas plagas de la agricultura. LITERATURA CITADA Ades, C.» C. Viera y H. Japyassu. 1990. Estrategias de caza na aranha Nephilengys cruentata (Araneae, Tetragnathidae) diante de presas diferentes. Anais 42a Reunion Soc. Bras. Prog. Ci., pp. 473-474. Breymeyer, A. & Jozwik, J. 1975. Consumption of wandering spiders (Lycosidae, Araneae) estimated in laboratory conditions. Bull. Acad. Polonaise Sci. Cl., II ser., Sci. Biol., 23:93-99. Bristowe, W. S. 1941. The Comity of Spiders. II. The Ray Society, London. 560 pp. Burgess, J. & P. N. Witt. 1976. Spider webs: design and engineering. Interdiscip. Sci. Rev., 1:322-335. Eisner, T., R. Alsop & G. Ettershank. 1964. Adhe- siveness of spider silk. Science, 146:1058-1061. Eisner, T. & J. Dean. 1976. Ploy and counterploy in predatory-prey interactions: orb-weaving spider versus bombardier beetles. Proc. Nat. Acad. Sc. USA, 73:1365-1367. Hagstrum, D. M. 1 970. Physiology of food utilization by the spider Tarentula kochi (Araneae, Lycosidae). Ann. Ent. Soc. America, 63:1305-1308. Harwood, R. H. 1 974. Predatory behavior of Argiope aurantia Lucas (Araneidae). American Midi. Nat., 90:47-55. Hays, H. E. 1985. Predator-prey interaction: garden spiders and house flies. Proc. Pennsylvania Acad. Sci., 59:29-32. Japyassu, H. F. & C. Ades. 1990. Influencia do ta- manho da presa na sequencia predatoria de Nephi- lengys cruentata (Araneidae). Res. Congr. Bras. Zool., Londrina, Brasil :27. Nentwig, W. 1 987. The prey of spiders. Pp. 249-263, In Ecophysiology of Spiders (W. Nentwig, ed.). Springer- Verlag. Berlin. 448 pp. Pasquet, A. 1984. Prey and predatory strategies of two orb-weaving spiders: Argiope bruennichi and Araneus marmoreus. Entomol. Exp. Appl., 36:177- 184. Pasquet, A. & R. Leborgne. 1986. Etude preliminaire des relations predateur-proies chez Zygiella x-no- tata (Araneae, Argiopidae). C. R. Soc. Biol., 180: 347-353. Pasquet, A. & R. Leborgne. 1988. Interception et capture des proies che 4 especes d’araignees orbi- teles. C. R. Xeme Coll. Europ. Arachnol. Bull. Soc. Sci. Bretagne, 59:175-178. Peters, H. M. 1931. Die Fanghandlung der Kreus- pinne (Epeira diademata Cl.): Experimentelle An- alysen der Verhaltens. Z. Vgl. Physiol., 1 5:693-748. Peters, H. M. 1933. Experimenteuber die Orienti- erung der Kreuspinne Epeira diademata Cl. Net. Zool. Jahrb. Abt. Allg. Zool. Physiol. Tiere, 5 1 :693- 748. Riechert, S. & J. Luczak. 1982. Spider foraging: be- havioral responses to prey. In Spider Communi- cation. (P. Witt & J. Rovner, eds.). Princeton Univ. Press. Princeton, New Jersey. 440 pp. Robinson, B. & M. H. Robinson. 1981. Ecologia y comportamiento de algunas aranas fabricadoras de redes en Panama: Argiope argentata, A. savignyi, Nephila clavipes y Eriophora fuliginea (Araneae: Ar- aneidae). Acad. Panamena de Med. y Cirugia, 6:90- 117. Robinson, M. H. & J. Olazarri. 1971. Units of be- havior and complex sequences in the predatory be- havior of Argiope argentata (Fabricius) (Araneae: Araneidae). Smithsonian Contr. Zool., 65:1-36. Robinson, M. H., H. Mirick & O. Turner. 1969. The predatory behavior of some araneid spiders and the origin of immobilization wrapping. Psyche, 76:486- 501. Robinson, M. H. & H. Mirick. 1971. The predatory behavior of the golden-web spider Nephila clavipes. Psyche, 78:123-139. Robinson, M. H. & B. Robinson. 1970. Prey caught by a sample population of the spider Argiope ar- gentata (Araneae: Araneidae) in Panama: a year’s census data. J. Linn. Soc. Zool., 49:345-358. Robinson, M. H. & B. Robinson. 1975. Evolution beyond the orb web: the web of the araneid spiders Pasilobus sp., its structure, operation and construc- tion. J. Linn. Soc. Zool., 56:301-314. Robinson, M. H. & B. Robinson. 1976. Discrimi- nation between prey types: animate component of the predatory behaviour of araneid spiders. Z. Tyerpsychol., 41:266-276. Stowe, M. K. 1986. Prey specialization in the Ara- neidae. Pp. 101-131, In Spiders: Webs, Behavior and Evolution. (W. A. Shear, ed.). Stanford Uni- versity Press. Suter, R. B. 1978. Cyclosa turbinata (Araneae: Ara- neidae): prey discrimination via web-borne vibra- tions. Behav. Ecol. Sociobiol., 3:283-296. Uetz, G. W. 1990. Prey selection in web-building spiders and evolution of prey defenses. Pp. 4:93- 128, In Insect Defenses. (D. L. Evans & J. O. Schmidt, eds.). State Univ. of New York Press. Viera, C. 1981. Discriminacion de Alpaida alticeps (Araneae, Araneidae) sobre tres ordenes de insectos. Trabajo de pasaje de curso de Profundizacion en Entomologia. Biblioteca de Fac. Ciencias. Uruguay. Viera, C. 1983. Comportamiento de captura de Al- paida alticeps (Keyserling 1879) (Araneae, Aranei- dae) sobre Acromyrmex sp. (Hymenoptera, For- micidae). Res. Com. Ill Jom. Cs. Naturales, Uru- guay, pp. 112-114. Viera, C. 1986. Comportamiento de captura de Me- tepeira sp. A (Araneae, Araneidae) sobre Acromyr- mex sp. (Hymenoptera, Formicidae) en condiciones experimentales. Aracnologia, 6:1-8. Viera, C. 1989. Caracteristicas de la tela orbicular de Metepeira sp. A (Araneae, Araneidae). Bol. Soc. Zool. Uruguay (2a epoca), 5:5-6. 24 THE JOURNAL OF ARACHNOLOGY Viera, C. 1994. Analisis del comportamiento depre- dador de Metepeira seditiosa (Keyserling) (Araneae, Araneidae) en condiciones experimentales. Arac- nologia (supl.), 8:1-9. Viera, C. y F. G. Costa. 1985. Captura de presas por machos adultos de Metepeira sp. A (Araneae, Ar- aneidae). Actas Jom. Zool. Uruguay, pp. 5-7. Yoshida, M. 1990. Predatory behavior of Meta re - ticuloides Yaginuma (Araneae: Tetragnathidae). Acta Aracnologica 39:27-39. AGRADECIMIENT OS A1 Prof. Roberto M. Capocasale por la lectura critica del manuscrito y su permanente apoyo y estimulo. A1 Dr. Cesar Ades por permitirme real- izar una pasantia en su laboratorio y vincularme con el Dr. William Eberhard (Smithsonian In- stitution y Univ. de Costa Rica). Agradezco sug- erencias a ambos sobre la tesis de Maestria que origino el presente trabajo. Manuscript received 1 January 1993, revised 9 January 1995. 1995. The Journal of Araefanology 23:25-30 THE WEB AND BUILDING BEHAVIOR OF SYNOTAXUS ECUADORENSIS (ARANEAE, SYNOTAXIDAE) William G. Eberhard: Smithsonian Tropical Research Institute and Escuela de Biologia, Universidad de Costa Rica, Ciudad Universitaria, Costa Rica ABSTRACT. Webs and building behavior of Synotaxm ecuadoremis are highly ordered and complex. Their webs differ from those of other Synotaxus species, but there are several apparent homologies in building behavior. The overall construction sequence differs from that of many other spiders in not being organized around a central portion or retreat. Instead, lines are added to one leading edge in a crochet-like fashion. Comparison with other Synotaxus species suggests how building behavior is organized within the spider. RESUMEN . La tela y el proceso de construction de la tela de Synotaxus ecuadorensis son complej as y altamente organizadas. La tela de esta especie difiere de las de otras especies de Synotaxus , pero el com porta mien to de construction muestra varias posibles homologias. La secuencia de pasos en la construction difiere de muchas otras aranas en no estar organizada alrededor de un punto o area central. A1 contrario, la arana agrega hilos al horde de la tela, en un proceso semejante a lo de la crochet. Se sugiere, a base de una comparicion entre los comportamientos de las diferentes especies de Synotaxus, como esta organizada el comportamiento de con- struction dentro de la arana. Synotaxus is a small neotropical genus con- taining five described and at least one unde- scribed species (H. W. Levi, pers. comm.). A recent study suggests that this genus, which has traditionally been placed in the family Theridi- idae, is part of a small group of genera (nearly all from New Zealand and Australia) that is the sister group of Nesticidae plus Theridiidae (For- ster et al. 1990). The webs of two species, S. turbinatus Simon, and Synotaxus sp., include highly regular arrays of approximately vertical and horizontal sticky and non-sticky lines (Eber- hard 1977). Webs are built as a series of ap- proximately rectangular modules or “unit webs"’. Each unit begins with a pair of more or less par- allel vertical, non-sticky lines. These are then joined by a series of more or less horizontal non- sticky lines, which are laid along with one to three zig-zag, sticky vertical lines in a complex series of events (Eberhard 1977). Both the geometric design of the web and the construction behavior of the species of this study, Synotaxus ecuadorensis Exline, are simpler. The construction behavior is of special interest be- cause it illustrates an overall building tactic (and thus a possible evolutionary route) which differs from that of many (perhaps most) other spiders, including those which make orbs (e. g., Foelix 1982; Eberhard 1990), those which make sheet webs, such as the theridiids Latrodectus spp. (Szlep 1965; Lamoral 1968), the psechrid Psech- rus sp. (Eberhard 1987), and the pholcid Modi- simus sp. (Eberhard 1992; Eberhard & Briceno 1985), and those which make other centrally or- ganized webs such as Titanoeca albomaculata (Szlep 1966) and Filistata spp. (Comstock 1948; Eberhard 1987, 1988). Instead of returning re- peatedly to a central point during construction, S. ecuadorensis adds to its web by moving back and forth along one edge, gradually extending it in a manner analogous to crocheting. METHODS Observations were made between 25 June and 3 July 1992 in the Reserva Natural La Planada (elev. 1800 m), 8 km S. of Chucones, Narino, Colombia, in an area classified as montane wet forest in the Holdridgean system (Espinal & Montenegro 1963). Webs were in grassy second growth and early secondary forest. Twelve webs of eight different spiders (four of which were adult females) were observed. Most of the construction of two of these webs (both of adult females) was observed using the white light of a headlamp. Web initiation appeared to be inhibited by il- luminating the spider, but once construction had begun the spider was apparently undisturbed by bright light. The use of bright light, plus the fact that spiders moved relatively slowly and used stereotyped movements which were frequently 25 26 THE JOURNAL OF ARACHNOLOGY repeated, made it possible to understand and rec- ord their actions in detail. Voucher specimens of the spiders observed (numbers 3638, 3645, 3646) and mature males are deposited in the Museum of Comparative Zoology, Cambridge, Massachusetts 02138, USA. RESULTS Webs were built under long (>15 cm), more or less horizontal leaves. The upper portion of the web was a tangle of non-sticky lines attached to the underside of the leaf. The lines of the mesh were more closely spaced in the area where the spider rested during the day against the underside of the leaf. The spider’s pale green color and its elongate abdomen, which it laid flat against the leaf, made it extremely cryptic. Egg sacs had thin walls with projecting processes, and the sphere of pale green eggs was plainly visible inside. The sacs (up to four per female) were suspended in the mesh under the leaf. The approximately planar prey capture web, strung vertically below the mesh, varied to some extent (Fig. 1). The lateral edges of the capture web were formed by two long, more or less ver- tical, non-sticky “frame” lines. The interior por- tions contained more or less regularly spaced lines, most of which had many short (0.2-0. 5 cm) seg- ments of adhesive on them. Initiation of capture web construction was not observed. Judging both by the lines present in the webs when first observed, and by the order in which subsequent lines were laid, it is probable that the first lines laid were the approximately vertical frame lines. One spider with only a mesh descended twice at the end of her dragline as if to begin construction early in the evening, but failed to contact a substrate below and climbed back up without making an attachment (and later abandoned the website). Subsequent lines were added in a highly ste- reotyped order (Fig. 2). The spider began by at- taching a dry line at the top of the capture web, usually near one edge. She then walked down- ward along the innermost line already present (this was the frame line in the first descent, and a line with adhesive segments in subsequent de- scents), attaching the non-sticky dragline she was laying periodically to the line along which she was walking (Fig. 2A). Immediately after each of these attachments, the spider backed up a short distance along the dragline and attached her dragline to it (Fig. 2A), thus making a short, more or less horizontal line (a “rung”), and then con- tinued her descent. In one case the spider broke and replaced the distal portion of the frame line along which she was walking as she neared the bottom of the capture web. After making the lowermost attachment to the line along which she was descending, the spider turned and began to climb the line she had just laid. She broke this line soon after she turned, and began reeling it up, replacing it with a new “sticky” line which consisted of a non-sticky line with evenly spaced short segments of sticky ma- terial. Each sticky segment was produced as both legs IV held the dragline and appeared to pull a short length from the spinnerets; the spider took one step forward with each leg IV (thus drawing out further silk), and then laid another sticky segment. Each time she reached a rung line, the spider broke it and performed a quick series of movements which I was unable to decipher, and then continued her upward climb. Judging by the pattern of lines when she was finished (Fig. 2B), probably the spider attached her dragline to the broken end of the rung line, paid out a short length of silk, and then attached her dragline to this line. A short segment of doubled line may have thus been produced, in a manner similar to the doubled lines laid during the descent (steps 2 and 3, and 4 and 5 in Fig. 2 A). Each finished rung had a single spot of white near the middle, apparently corresponding to the broken end of the line to which the spider had attached her dragline. In one web the spider alternated descents on the right and then the left side of the web. In another web she made several descents on one side (the larger of the two) before making any on the other. In both webs later lines with sticky material were progressively less vertical, as the spider filled in the central portion of the web (Fig. 2D). The final line was made following construc- tion of the lowermost sticky line. The spider moved more or less directly upward through the middle of the capture web to the mesh above, laying a sticky line as she went. In one web the spider clearly broke all of the lines she encoun- tered as she climbed, reattaching each to the sticky line she was laying. A short length of silk was paid out just before each reattachment, thus low- ering the tension on these lines. DISCUSSION Although the prey capture webs of S. ecu- adorensis are different from those of S. turbinatus and S. sp. (compare Figs. 1 and 2 with Fig. 3), EBERH ARD — WEB OF SYNOTAXUS ECUADORENSIS 27 Figure 1.— Two newly-built capture webs of mature female Synotaxus ecuadorensis coated with cornstarch. The spider is just visible at the top of the upper web. In the lower web the lines at the right and the curved line from the tip of the leaf at the left are out of the plane of the capture web (scale lines = 5.0 and 6.0 cm for upper and lower webs). Figure 2. —Diagrammatic representation of the probable order of operations by Synotaxus ecuadorensis build- ing a prey capture web. Thicker lines and large dots indicate the lines and attachments made during the period represented by each drawing; wiggly lines represent patches of adhesive; and numbers refer to the order of attachments. A. The spider starts from the mesh under the leaf along one of the two long non-sticky, more or less vertical lines that form the lateral borders of the web, laying a non-sticky dragline. Periodically she attaches the dragline to the frame (e. g., 2, 4), and backs up slightly and attaches to the line just laid (e. g., 3, 5) forming a “rung”. She then continues downward to make a final attachment to the frame (e. g., 6). B. Turning back immediately, the spider breaks the line she has just laid, attaches her trail line to the broken end (7), and begins laying another line with sticky patches on it as she climbs back up along the line she just laid. She breaks each rung and attaches her dragline to the broken end (e. g., 8). Backing up slightly, she attaches to the line she just laid (e. g., 9) and continues upward, finally attaching the sticky line to the mesh near where she started (12). C. Subsequent lines are laid on the same or the opposite side of the web with a similar series of movements. D. Sticky lines laid later are progressively less vertical. The attachments 9 and 12 were deduced from the positions of lines in finished webs, while all others, and the breaking of lines at 7, 8 and 10 were confirmed by direct observations. several details indicate that the prey capture web of S. ecuadorensis is homologous with a single “unit” of the web design of the others (Eberhard 1977) (Fig. 3). Both types of web are initiated with a pair of long, more or less vertical, straight, non-sticky lines which form their lateral margins. A complex sequence of behavior follows, in which construction of non-sticky and sticky lines alter- nate, with the sticky lines bearing widely spaced segments or dots of adhesive. One detail of this process in all three species is apparently unique to Synotaxus among all araneoid web builders studied to date: after attaching its dragline to another line, the spider backs up a short distance and makes another attachment to the line it just laid, then continues onward (e. g., Fig. 2 A, B). A further similarity is that sticky lines are laid as the spider climbs upward, each replacing a non-sticky line laid during an immediately pre- ceding descent. Construction ends with place- ment of a central sticky line laid as the spider ascends. Lines already present are apparently EBERHARD- WEB OF SYNOTAXUS ECUADORENSIS 29 Figure 3. —Tentative order of operations in the construction of a unit web of Synotaxus turbinatus (after Eberhard 1977). The spider moves from side to side as she descends, laying both sticky and non-sticky lines (A, B, C). After reaching the bottom (D), she climbs up the middle of the web, replacing the non-sticky line with a sticky line (E). broken and reattached to this central line. Both types of webs are vertical, more or less planar, and relatively fragile arrays that are rebuilt daily, and are located immediately below a more per- manent mesh of non-sticky lines near the un- derside of a large leaf where the spider rests. These proposed homologies must remain ten- tative, however, until further data on Synotaxus and related genera become available. The ap- parent homology of the S. ecuadorensis web to a unit of the webs of other Synotaxus species indicates that the “units” of these species are not simply abstractions, but that web construction may be also organized in the spider’s nervous system as units. Differences between the webs and construction behavior of S. ecuadorensis and that of other Synotaxus are also substantial. They include the following: the web is constructed as a single unit, with only a single pair of vertical frame lines rather than as a series of modules; a long, un- interrupted non-sticky line is laid during each descent and is nearly completely removed during the ensuing ascent; placement of adhesive ma- terial is in short segments rather than single balls on the sticky lines; and there is no non-sticky “frame” line at the bottom of the web (also some- times lacking in other Synotaxus). The construction behavior of S. ecuadorensis is to my knowledge the clearest described ex- 30 THE JOURNAL OF ARACHNOLOGY ample in which a spider does not organize its activities around a central area. Instead, after establishing three sides of the planar web (the mesh above and the two lateral frames), the spi- der moves back and forth across the fourth side, gradually extending the web in a process analo- gous to crocheting. I have seen a similar process only one other spider (a species of the theridiid Chrosiothes which repairs holes in its sheet in this manner) (Eberhard, pers. obs.). ACKNOWLEDGMENTS My trip to La Planada was financed by the Fundacion para la Educacion Superior; the staff at La Planada helped make my stay pleasant and productive. Carlos Valderrama kindly collected voucher males. Dr. H. W. Levi kindly identified spiders, and R. Gillespie and B. D. Opell made helpful comments on a preliminary manuscript. My research was supported by the Smithsonian Tropical Research Institute and the Vicerrectoria de Investigation of the Universidad de Costa Rica. I thank all for their help. LITERATURE CITED Comstock, J. 1948. The Spider Book (revised and edited by W. J. Gertsch). Cornell Univ. Press, Ith- aca. Eberhard, W. G. 1977. “Rectangular orb” webs of Synotaxus (Araneae: Theridiidae). J. Nat. Hist., 11: 501-507. Eberhard, W. G. 1987. Construction behavior of non-orb weaving cribellate spiders and the evolu- tionary origin of orb webs. Bull. British Arachnol. Soc., 7:175-178. Eberhard, W. G. 1988. Combing and sticky silk at- tachment behaviour by cribellate spiders and its tax- onomic implications. Bull. British Arachnol. Soc., 7:247-251. Eberhard, W. G. 1 990. Early stages of orb construc- tion by Philoponella vicina, Leucauge mariana, and Nephila clavipes (Araneae, Uloboridae and Tetrag- nathidae), and their phylogenetic implications. J. Arachnol., 18:205-234. Eberhard, W. G. 1992. Web construction by Mod- isimus sp. (Araneae, Pholcidae). J. Arachnol., 20: 25-34. Eberhard, W. G. & R. D. Briceno. 1985. Behavior and ecology of four species of Modisimus and Ble- chroscelis (Pholcidae). Rev. Arachnol., 6:29-36. Espinal, L. S. & Montenegro, E. 1 963. Formaciones vegetales de Colombia. Republica de Colombia, In- stitute Augustin Codazzi, Bogota. Forster, R. R., N. I. Platnick, & J. Codding- ton. 1990. A proposal and review of the spider family Synotaxidae (Araneae, Araneoidea), with notes on theridiid interrelationships. Bull. Ameri- can Mus. Nat. Hist. 193:1-116. Lamoral, B. H. 1 968. On the nest and web structure of Latrodectus in South Africa, and some observa- tions on body colouration of L. geometricus (Ara- neae: Theridiidae). Ann. Natal Mus., 20:1-14. Szlep, R. 1 965. The web-spinning process and web- structure of Latrodectus trecinguttatus, L. pallidus and L. revivensis. Proc. Zool. Soc. London, 145:75- 89. Szlep, R. 1966. Evolution of the web spinning ac- tivities: the web spinning in Titanoeca albomaculata Luc. (Araneae, Amaurobiidae). Israel J. Zool. 15: 83-88. Manuscript received 25 July 1994, revised 14 November 1994. 1995. The Journal of Arachnology 23:31-36 LOS NERVIOS OPTICOS EN CUATRO ESPECIES DE LACTRODECTUS (ARANEAE, THERIDIIDAE) Carmen J. de la Serna de Esteban y C. Monica Spinel!!: Facultad de Ciencias Exactas y Maturates, Universidad de Buenos Aires, Departamento de Ciencias Biologicas. Ciudad Universitaria Pabellon II 4 Piso Lab, 22- (1428) Buenos Aires, Argentina ABSTRACT. The pathway of the optic nerves in the studied species of Latrodectus shows intraspecific vari- ation. Dilatations, empty or containing a pigment of unknown function, can be seen in the nerves. Curiously, this pigment originates in the retinal cells. The optic nerves ran through the prosoma and merge, forming two or four optic centers, which are finally joined into a single one. RRSUMEN. La trayectoria de los nervios opticos en las especies de Latrodectus estudiadas muestra variation intraespecifica. En ellos existen dilataciones que pueden hallarse vacias o conteniendo un pigmento originado en las celulas retinianas, cuya funcion es desconocida. Estos nervios forman dos o cuatro centros opticos que luego se fusionan en un centro unico. El trayecto de los nervios opticos en el pro- soma de varias especies de Tegenaria Latreille 1804 fue estudiado por Legendre (1959), quien sin determinar el ordenamiento de los mismos confirm© que este genero caretia de quiasma op- tico. Este autor dice que los nervios opticos sur- gen de la superficie anterior del ganglio cere- broide, hallandose constituido cada uno de ellos por tres haces de fibras, los que se dirigen a los ojos laterales, y por encima de los cuales se en- cuentra un pequeno haz impar. El mismo autor, hallo estas mismas particularidades con mmimas variaciones en especies que no cita en el trabajo, asi como tampoco las familias sobre las que rea- lize el estudio. Homann (1947) considera que en la arana se- dentaria Araneus sexpunctatus , los nervios op- ticos penetran en conjunto en el ganglio cere- broide, originandose asi un centro optico unico. Baccetti y Bedini (1964) realizaron estudios mediante microscopia optica y electronica en los ojos de Arctosa variana (Lycosidae), en los que senalan la presencia de un pigmento claro en los nervios opticos, sin especificar origen ni natur- aleza quimica del mismo. En el presente trabajo se estudio el recorrido de los nervios opticos en el prosoma, en cuatro especies del genero Latrodectus (Araneae, Ther- idiidae) obteniendose resultados que no coinci- den con las observaeiones realizadas para otras especies, por los mencionados autores. Por otra parte el llamativo recorrido de los nervios opticos, con variaciones intraespecificas condujo a realizar un estudio comparative de los mismos. METODOS Se emplearon 1 1 ejemplares del genero Latro- dectus: dos individuos de L. geometricus Koch 1841, dos de L. mirahilis Holmberg 1 876, cuatro de L. antheratus Badcock 1932 y tres de L. cor- allinus Abalos 1953. Los fij adores utilizados fueron: formol 10%, Helly y Bouin y la inclusion se realizo en para fina. Las coloraciones histologicas de rutina fu- eron: Hematoxilina de Carazzi - Ponceau de Xil- idina - Azul de Anilina, Mallory - Heidenhain (Azan), fucsina paraldehida segun Gabe (Martoja y Martoja Pierson). Las coloraciones histoqmmi- cas fueron: Periodic Acid SchifF (P. A. S), color- ation para prottinas segun Martoja y Alcian Blue a diferentes pH. Para la indentificacion de los nervios opticos correspondientes a cada ojo, en su trayectoria, se confeccionaron esquemas rotulandose los ner- vios con las siguientes abreviaturas: OMA: ojo medio anterior; OLA: ojo lateral anterior; OMP: ojo medio posterior; y OLP: ojo lateral posterior. Dado que los nervios se fusionan de a pares a corta distancia de su emergencia del ojo, se ha empleado la siguiente denomination, nj = OMA derecho + OLA dereeho; n2 = OMA izquierdo 31 32 THE JOURNAL OF ARACHNOLOGY Figuras 1-5.— Nervios en especies de Latrodectus. I, Latrodectus mirabilis - fusion del nervio del ojo medio posterior con el del ojo lateral anterior (f); 2, Latrodectus geometricus - fusion del nervio del ojo medio posterior con el del ojo lateral posterior, nomp = nervio del ojo medio posterior, nolp = nervio del ojo lateral posterior, omp = ojo medio posterior; 3, Latrodectus mirabilis - Corte de la zona anterior del prosoma con los cuatro nervios opticos en la linea media, en n 1 se observan axones dilatados y vacios, except© uno superior que contiene restos de la secretion originada en el ojo, r = dilatation del nervio, s = restos de secretion; 4, Latrodectus mirabilis - Principio de fusion de n2 + n4, y + n3 ya fusionados, r = dilatation del nervio o reservorio vacio, s = restos de secretion; 5, Latrodectus corallinus - nt + n2 y n3 + n4 antes del punto de fusion + OLA izquierdo; n3 = OMP derecho + OLP derecho; n4 = OMP izquierdo + OLP izquierdo. Para apreciar las variaciones en la ordenacion de estos nervios fusionados, se ban considerado en el Cuadro I tres regiones representativas del prosoma: una zona anterior, una zona media y una posterior. En la zona anterior, donde los nervios se disponen en forma vertical, se ban asignado las letras A, B, C y D a los diferentes tipos de disposition, siendo: Tipo A = nl5 n2, n3, DE LA SERNA Y SFINELLI-LOS NERVIOS OPTICOS DE LATRODECTUS 33 Figuras 6-8. —Los centres opticos de especies de Latrodectus. 6, Latrodectus corallinus- Dos centros opticos (co) en el ganglio cerebroide; 7, Latrodectus geometricus - Cuatro centros opticos en el ganglio cerebroide, ca = centro optico de los ojos anteriores, op = centro optico de los ojos posteriores; 8, Latrodectus geometricus - En un corte transversal en vista anterior se observan dos centros opticos del lado derecho. En el lado izquierdo se observa solo uno, debido a la fusion del centro optico de los ojos anteriores, con el de los posteriores del mismo lado, ca = centro optico de los ojos anteriores, op = centro optico de los ojos posteriores. n4; Tipo B = n„ n2, n4, n3; Tipo C = n2, n1? n3, n4; Tipo D = n2, n1? n4, n3. OBSER V ACIONES En el genero Latrodectus se observaron solo cuatro nervios opticos en la linea media del pro- soma (Figs. 3, 5). Esto se debe a que el nervio de cada ojo medio anterior se fusiona con el nervio del ojo lateral anterior del mismo lado en un punto cercano a la salida del ojo. Identica disposition se observa en los ojos posteriores (Figs. 1,2). De los resultados volcados en el Cuadro I surge que en la zona anterior del prosoma, hay una ordenacion variable en la trayectoria de los ner- vios en cada especie estudiada y aun entre in- dividuos de la misma especie. En el trayecto prosomatico (zona media), se produce una nueva ordenacion y fusion de los nervios opticos uniendose de la siguiente forma: a) Todos los nervios correspondientes a los ojos anteriores (medios y laterales) del lado derecho, con todos los nervios de los ojos posteriores (me- dios y laterales) del mismo lado. b) Todos los nervios de los ojos anteriores y posteriores del lado derecho y a su vez, todos los nervios an- teriores y posteriores del lado izquierdo. En el caso a) penetran en primer termino los nervios correspondientes a los ojos posteriores y luego los correspondientes a los ojos anteriores. Dando origen a cuatro centros opticos, dos en cada lado (Figs. 7, 8). En el caso b) se forman solo dos centros opticos, uno de cada lado (Fig. 6). En ambos casos los centros opticos se fusionan entre 34 THE JOURNAL OF ARACHNOLOGY Figuras 9-10 .—Latrodectus geometricus - Neurosecrecion retinal en el nervio optico del ojo lateral posterior, pm = pigmento melanico, nr = neurosecrecion retinal; 10, Latrodectus corallinus- omp = ojo medio posterior, olp = ojo lateral posterior, nr = nervios con neurosecrecion retinal. si con posterioridad, formando un centre optico unico. Estudiando diversos preparados coloreados con las tecnicas habituates, se observe el pigmento amarillo - anaranjado detectado ya por Bacetti y Bedini (1964). Este pigmento es muy evidente dentro de los axones que constituyen los nervios opticos (Figs. 3, 4, 9, 10) pero nunca se presenta en los centres opticos del ganglio cerebroide. An- tes de alcanzar el centre optico el pigmento ha desaparecido, quedando en los axones de los ner- vios optico dilataciones muy llamativas, vacias o con restos de secrecion (Figs. 3, 4). DISCUSION Las observaciones realizadas coinciden con la afirmacion de Legendre (1959) sobre la ausencia de quiasma optico. Este autor senala la presencia de ocho nervios opticos en la linea media del prosoma en dos especies del genero Tegenaria, y sostiene que su origen se halla en el ganglio cerebroide, lo cual no ocurre en las especies es- tudiadas del genero Latrodectus. Las fibras o axones que constituyen los nervios opticos son puramente sensoriales y forman un neuropilo optico donde segun Trujillo-Cenoz (1965) sinaptan celulas visuales ganglionares. Por otra parte, en las especies estudiadas, la precoz fusion de los nervios opticos reduce su numero a cuatro, sin que ello signifique por ejemplo, que exista alguna relacion entre esa reduccion y el tipo de tela que la arana construye. Homann ( 1 947) halla en la familia Theriididae un centre optico unico en el ganglio cerebroide. El genero Latrodectus pertenece a la misma fam- ilia, y en las especies de este genero estudiadas en el presente trabajo, se comprobo que en la zona anterior del ganglio cerebroide, se pueden encontrar dos o cuatro centres opticos segun el tipo de fusion de los nervios opticos, los que a su vez se fusionan formando un centro optico unico. Heinrichs y Fleissner (1987) observan la pre- sencia de neurosecrecion llamada precerebral en la retina de los ojos medios, proveniente de neu- ronas centrales y relacionada con information circadiana, en el escorpion Androctonus austral- is. En este caso la neurosecrecion tiene su origen en neuronas cerebrales ubicadas en el ganglio cerebroide, y es llamada precerebral por estar dirigida hacia los ojos. No es este el caso de Latrodectus, pues la neurosecrecion se origina en neuronas sensoriales (celulas retinales) situadas por supuesto por delante del ganglio cerebroide. Con el fin de no confundir conceptos se ha optado en este caso por denominarla “neurose- crecion retinal”. Las dilataciones axonales corresponderian a sitios de almacenamiento de la secrecion prove- niente de las celulas retinales, y la naturaleza quimica de la misma no ha podido determinarse con exactitud mediante las tecnicas histoquim- icas empleadas aunque puede sospecharse que se trata de una sustancia proteica dado el color amarillo que presenta en preparados coloreados con la tecnica para proteinas de Martoja-Mar- toja. Estos resultados estarian apoyados por el hecho de que las hormonas elaboradas por neuronas, que son celulas de origen ectodermico, son po- lipeptidos. Esta reflexion conduce a desechar la posibilidad de que se trate de esteroides, puesto que los mismos son producidos por tejidos de origen mesodermico. DE LA SERNA Y SPINELLI-LOS NERVIOS OPTICOS DE LATRODECTUS 35 Cuadro 1.— Esquema de las diversas disposiciones observadas en corte transversal de los nervios opticos en cuat.ro especies de Latrodectus, en la region cercana a la Hnea media o mediana del prosoma, y considerando las tres zonas antes indicadas. Las posiciones relativas de los numeros dentro de cada casillero, indican la ordenacion espacial de los nervios opticos. (El tipo de disposicion A puede verse en la Fig. 5, y el tipo B en la Fig. 3.) Tipo de disposicion Zona anterior Zona intermedia Zona proxima a los centres opticos Latrodectus geometricus A 1 2 1 2 2 1 4 4 3 4 3 3 B 2 2 1 2 1 1 4 4 3 4 3 3 Latrodectus mirabilis C 2 2 1 2 1 1 3 3 3 4 4 4 B 2 2 1 2 1 1 4 4 3 4 3 3 Latrodectus antheratus D 1 2 1 2 2 1 3 4 3 4 4 3 A 1 1 1 2 2 2 4 4 3 4 3 3 C 2 1 1 2 1 2 3 4 3 4 4 3 B 2 2 1 2 1 1 4 4 3 4 3 3 Latrodectus corallinus D 1 1 1 2 2 2 3 3 3 4 4 4 A 1 1 1 2 2 2 4 4 3 4 3 3 B 2 2 1 2 1 1 4 4 3 4 3 3 36 THE JOURNAL OF ARACHNOLOGY CON CLU SIONES Resumiendo lo observado en los preparados y los datos consignados en el Cuadro I, se puede afirmar que: 1. No existe quiasma optico. 2. En las cuatro especies estudiadas (Latrodectus mi - rabilis, L. antheratus , L. corallinus y L. geome - tricus ) el nervio optico de cada ojo medio se fusiona con su lateral correspondiente en la zona anterior del prosoma. 3. Dentro de cada una de las especies estudiadas, los nervios no presentan una disposition constante en su trayecto a traves del prosoma. 4. En la zona media del prosoma, donde se produce la segunda fusion de nervios se mantiene el siguiente esquema, a) si la fusion se realiza entre los nervios de los ojos medios y laterales anteriores de ambos lados y los de los ojos medios y laterales posteriores tambien de ambos lados dan origen a cuatro centres opticos (dos de cada lado) ubicados en la region anterior del ganglio cerebroide, y b) si la fusion se produce entre los nervios de los ojos anteriores (medios y laterales) con los de los ojos posteriores (medios y laterales) del mismo lado originan solo dos centres opticos, ubicados tambien en la region anterior del ganglio cerebroide. 5. En la region posterior del ganglio cerebroide, los centres op- ticos se fusionan, cualquiera sea su numero (dos o cuatro) en un centre optico unico. 6. En el presente trabajo se considera que los nervios op- ticos estan constituidos por axones de las celulas retinales y no de las neuronas que constituyen el ganglio cerebroide. 7. En las celulas retinales del ojo se produce una neurosecrecion que se des- plaza por dentro de los axones de los nervios opticos y es eliminada antes de ingresar en los centres opticos del ganglio cerebroide. A dicha secretion se la ha denominado “neurosecrecion retinal”. AGRADECIMIENTOS Agradecemos al Lie. Pablo Winitzky su cola- boracion en la transcription del manuscrito. LITERATURA CITADA Baccetti, B. & C. Bedini. 1 964. Research on the struc- ture and physiology of the eyes of a Lycosid spider. Arch. Italiennes Biol, 102:97-121. Gabe, M. 1947. Donnees histologiques sur la neu- rosecretion chez les Arachnides. Arch. Anat. Micr. Morph. Exp., 44:351. Heinrichs, S. & G. Fleissner. 1987. Neural compo- nents of the circadial clock in the scorpion Androc - tonus australis. Central origin of the efferent neu- rosecretory elements projecting to the median eyes. Cell and Tissue Res., 250:277. Homann, H. 1947. Die Nebenaugen der Spinnen (Araneae). Z. Naturforsch., 2b: 16 1-1 67. Legendre, R. 1959. Contribution a l’etude du systeme nerveux des Araneides. Ann. des Sc. Nat. Zoologie (12 ser.):400. Martoja, R. y M. Martoja Pierson. 1970. Tecnicas de Histologia Animal. Toray-Masson S. A. Bar- celona. Trujillo-Cenoz, O. 1965. Some aspects of the struc- tural organization of the arthropod eye. Cold Spring Harbor Symp. Quant. Biol. 30:371. Manuscript received 25 October 1993, revised 1 Decem- ber 1994. 1995. The Journal of Arachnology 23:37-43 DIRECT EVIDENCE FOR TRADE-OFFS BETWEEN FORAGING AND GROWTH IN A JUVENILE SPIDER Linden E. Higgins: Dept, of Zoology, University of Texas at Austin, Austin, Texas 78712 USA ABSTRACT. A simple modification of classical optimal foraging models yields explicit predictions of the allocation of resources to orb-web synthesis and weight gain in a spider. Because the spiders may be trying to avoid weight loss, weight gain and web size are predicted to be be negatively correlated at higher food levels and positively correlated at lower food levels. These predictions were upheld qualitatively by experiments involving juvenile Nephila clavipes and N. maculata. RESUMEN. Una sincilla modificacion a los modelos clasicos del forrajeo optimo da predicciones explicitas sobre la repartition de recursos entre el sintesis de la tela orbicular y el aumento de peso en una arana. Dado que aranas posiblemente estan ententando evitar la perdida de peso, la interaction entre el aumento de peso y el tamano de la tela deberia estar negativamente correlacionada bajo majores niveles de alimentacion, y posi- tivamente correlacionada bajo de condiciones de alimentacion reducida. Estas predicciones fueron appoyadas experimentalmente utilizando juveniles de Nephila clavipes y N. maculata. Classical foraging models are based on the as- sumption that optimal allocation of limiting re- sources among various expenditures increases the fitness of the organism either directly (through increased reproductive success) or indirectly (through increased growth) (e. g., Schoener 1971; Pyke et al. 1977; Mangel & Clark 1986). Despite the key role of this assumption, few studies have described the interaction between foraging in- vestment and growth at different levels of food availability. This appears in part due to the dif- ficulty of finding common units of measure (Pianka 1981; Stephens & Krebs 1986). If com- mon resources are used for both growth and for- aging, then explicit predictions can be made con- cerning their allocation at different levels of prey availability. The study of foraging in orb-web building spi- ders allows observation of resource allocation because foraging investment is measurable as material investment into the web, and the orb is composed of physiologically important com- pounds. The orb web is renewed regularly, and changes in orb-web size reflect the response of the spider to foraging conditions (Buskirk 1975; Gillespie 1987; Higgins & Buskirk 1992). The viscid silk that forms the spiral of the orb is a protein thread coated with a mixture of organic compounds that also have physiological func- tions (Tillinghast 1984; Dadd 1985; Townley et al. 1991). Thus, construction of each orb requires decisions concerning the allocation of metabol- ically important compounds; such decisions could directly affect growth. In some large araneoid spiders, fitness is in- creased by rapid development and weight loss is predicted to reduce the fitness of the females (Higgins & Rankin, in press). The existence of resources common to both foraging and growth and the fitness penalties associated with weight loss imply that when resources are limiting, for- aging behavior might be different from that pre- dicted by foraging models calculating maximi- zation of the rate of energy intake. Here, I present a simple model that predicts patterns of alloca- tion of resources between foraging and weight gain at food levels varying from very low to very high. I experimentally tested this model using juvenile Nephila clavipes (L.) and N. maculata (Fabr.) (Araneae: Tetragnathidae) that were fed at very low, intermediate and high levels. THE MODEL If it is true that the functions of weight gain and web building are competing for common resources and if it is important to the individual to avoid weight loss, a simple modification of the classic foraging investment models (Schoener 1971) is necessary to predict behavior at very low levels of prey availability. The model is pre- 37 38 THE JOURNAL OF ARACHNOLOGY low high food level ro 0 1 0 ■O Figure L— The model of resource allocation by an orb-weaving spider. A. As food resources increase for a given individual, orb size increases then declines and weight gain increases. B. Among spiders, the residuals of orb size and weight gain regressed against spider size are compared (eliminating the variation due to spider size). At low food levels (square), orb size and weight gain are lower than average; at medium food levels (circle, diamond), orb size is greater and weight gain is lower than average; and at high food levels (triangle), orb size is lower and weight gain is higher than average. seated graphically in Fig. 1. At high to inter- mediate levels of food, the spiders have sufficient resources to avoid weight loss and increase in- vestment into foraging as food levels decrease. This is consistent with classical foraging models: as average foraging success decreases, the in- vestment into hunting (e. g., searching time per patch) is predicted to increase (maximizing en- ergy gain/effort; Stephens & Krebs 1 986). If com- mon resources are being allocated into growth and foraging then the rate of weight gain will be slower, due to both decreased prey capture and increased foraging investment. However, this prediction will not hold at lower food levels if the spiders must avoid weight loss. When re- sources are so limiting that weight loss would be the price of continued increases in foraging in- residual of orb size vestment, the spiders should reduce their for- aging effort. Thus, the relationship between weight gain and web investment in spiders is predicted to be negatively correlated at higher food levels, and positively correlated lower food levels. METHODS In order to test the prediction that the relative investment into the orb and weight gain will dif- fer between the shift from very low to interme- diate foraging success and the shift from inter- mediate to high foraging success, I controlled prey availability in juveniles of two species of Nephila : lab-reared N. davipes in Panama and field-collected N . maculata in Madang, Papua New Guinea. In both cases the investment into weight gain and orb area were recorded as func- tions of prey availability, measured as the daily prey weight/initial spider weight. The basic methods were the same in the two experiments. To control the structural environ- ment used in web construction, I placed indi- viduals on 30 cm diameter spherical frames con- sisting of two hoops of 3 mm x 6 mm fiberglass, fixed at right angles and suspended from the ceil- ing or braces in open-air insectaries. I fed spiders locally available stingless bees ( Trigona sp.), a common prey item of these spiders (Higgins & Buskirk 1992; Higgins, pers. obs.). Bees were col- lected from the same nests throughout the ex- periment. For each animal, I recorded daily orb-web area and rate of weight gain. (Mesh size, measured as the number of spiral strands per 2 cm (Higgins & Buskirk 1992), did not vary with food level in either species. N . davipes: F(1j34) = 3.16, ns; N . maculata : F(2> 13) = 0.34, ns). Orbs built by these spiders are asymmetrical, and orb-web area was estimated as a half ellipse using measurements of the vertical and horizontal radii. While many orb weavers replace the orb each day, these spi- ders sometimes only partially renew the orb (Lu- bin 1973). When a web was partially new, the area was estimated by calculating the area of the orb then reducing to the estimated area that was new (15) - 7.398, P < 0.025; (web/hunt)* - 8.506 + 10.67 (TPL), R2 = 0.34, F{1 15) = 8.26, P < 0.025). RESULTS Nephila clavipes.— Both rate of weight gain (5wt) and mean orb size varied with food level (Fig. 2). From 10% to 20% relative food (second trial), dwt and orb size increased, while from 20% to 35% relative food (first trial), 5wt increased and orb size decreased. Three MANOVA com- pared the responses of the spiders to food level within each trial, and compared between trials at the 20% food level; the significance levels were adjusted for multiple, non-independent tests. In the second trial, there was a significant effect of food level on <5wt but not on web size (Wilks’ X = 0.356, P < 0.01; ANOVA: 6wt: F(1> 13) = 23.5, P < 0.001; web size: F(1 13) = 0.20, ns). There was a significant effect in the first trial on both web size and <5wt (Wilks’ X = 0.568, P < 0.01; ANOVA: §wt: F(1j19) = 8.09, P = 0.01; web size: F(h 19) = 5.41, P = 0.03). Between trials, there was a significant difference in <5wt but not web size (Wilks’ X = 0.514, P < 0.01; ANOVA 5wt: F(1 17) = 1 6.05, P < 0.01; web size: F(1 17) = 0.03, ns). The effect of trial in the medium food level A spider size, TPL 1 , cm Figure 4.— The relationship between mean orb- web size and mean daily weight gain in individual Nephila maculata under three different food levels (% initial spider weight). Square - <17%, triangle - <22%, di- amond- <45%. A. There is no effect of the sex of the individual on resource allocation. Male - filled sym- bols, female = open symbols. B. There is no effect of initial spider size on resource allocation. Filled symbols = residual of orb-web size, open symbols = residual of 5wl. is probably due to the larger mean initial spider size in trial 2. Reduced rate of weight gain had a significant negative effect on the development rate of the spiders, observed as an increase in the number of days between molts. Eighteen spiders from trial 1 were held on the assigned diet until they molted and the number of days between molts was correlated with the rate of weight gain and initial spider size (In (days between molts) = 1 .96 - 37.7 (5wt) + 4.81 (TPL); R2 = 0.46, F(2> 15) = 6.25, P = 0.01 1; only 5 spiders molted in trial 2). 42 THE JOURNAL OF ARACHNOLOGY Nephila maculata . — Although the pattern of variation in orb investment and rate of weight gain among the three groups is similar to that observed for N. clavipes, the differences are not significant (Fig. 3; Wilks’ X - 0.78, P = 0.5). What is most striking is the variation among individ- uals in response to the same level of food intake (Fig. 4). Often, different individuals made op- posite responses: one spider building larger- than- average orbs and gaining little weight, another building smaller- than-average orbs and gaining weight rapidly. The variation was not a function of the sex nor of the initial size (TPL) of the individual (Fig. 4). CONCLUSIONS The two experiments uphold the predictions of the model. Juvenile Nephila of both species tended to increase orb- web size and decrease the rate of weight gain with slightly reduced prey capture; they decreased both orb-web size and rate of weight gain in response to greater reduc- tions in prey capture. This response resulted in the counter-intuitive pattern of decreased for- aging investment with low food availability, up- holding the initial assumptions that resource al- location is constrained by a need to avoid weight loss and that common resources are used in for- aging and growth in these spiders. At very low food levels, webs were small and few spiders gained weight but none lost weight. This food level is within the range captured in Panama, but much lower than median prey capture rates ob- served in Madang (4 prey/ 12 h diurnal obser- vation) (Higgins & Buskirk 1992; Higgins, pers. obs.). Nephila maculata juveniles varied greatly in their response to the experimental treatment. Such differences were not found in the experi- ment with N. clavipes, when the spiders were reared under homogeneous conditions. This dif- ference might represent some intrinsic difference between species, but probably reflects differences in experience prior to the experiment. N. ma- culata juveniles were collected in the field, and undoubtedly differed in their foraging histories. The foraging history of an individual may cause it to respond in a particular fashion to changes in prey capture rates. This implies differences in the response to short vs. long term variation in prey capture, which cannot be examined by the short-term experiments presented here. There is evidence that many spiders evolved under conditions of food limitation (Anderson 1974; Wise 1975, 1979). In web-building spiders, the web must be built before the foraging quality of a site can be assessed. Perhaps due to this initial investment, or due to temporal variation in prey availability, many web-building spiders are highly site-tenacious (Eberhard 1971; Enders 1976; Tanaka 1989; however, see Turnbull 1964). Rather than abandoning poor patches, a strategy often predicted by foraging theory (e. g., Chamov 1971; Pyke et al. 1977), spiders show other be- havioral and physiological responses to nutri- tional stress (Riechert & Luczak 1982; Sherman 1 994). However, these responses can slow growth and affect both survival and female fecundity (Higgins 1992a, 1992b, pers. obs.). The results of these studies support the hypothesis that there are potentially limiting resources necessary for both orb silk synthesis and growth, predicted by the chemical analyses of orb-silk by Tillinghast and Townley (Tillinghast 1984; Townley et al. 1991). The spiders must invest in the orb web in order to capture prey, but unless critical re- sources are in abundance, such investment nec- essarily delays growth and development and po- tentially compromises individual fitness. Pre- dation rate declines with increasing spider size and fecundity increases with early maturation (Higgins 1 992b, pers. obs.), so lengthening time between molts can have a detrimental effect on fitness. To avoid reduction in weight during pe- riods of low food availability, the interactions between foraging investment and growth may be more complex than predicted by classical opti- mal foraging models. ACKNOWLEDGMENTS J. Wright’s comments and logistical support from the Smithsonian Tropical Research Insti- tute staff were essential in the development of this project. M. Townley and E. Tillinghast pro- vided valuable information for the interpretation of the experiment and they, R. Buskirk, W. Eber- hard, M. A. Rankin, C. Macias, G. Miller, C. Pease, G. Perry, M. Singer, D. Wise and anon- ymous reviewers commented upon earlier ver- sions of the manuscript. This work was sup- ported by Sigma-Xi, the UT Graduate Fellow- ship program, an NSF Dissertation Improve- ment Grant (BSR - 8413831), a STRI short-term fellowship, and the Christensen Fund Grant. This is contribution No. 138 from the Christensen Research Institute (P.O. Box 305, Madang, Pa- pua New Guinea). HIGGINS TRADE OFFS BETWEEN FORAGING AND GROWTH 43 LITERATURE CITED Anderson, J. F. 1974. Responses to starvation in the spiders Lycosa lenta Hentz and Filistata hib- ernalis (Hentz). Ecology, 55:576-585. Buskirk, R. E. 1975. Coloniality, activity patterns and feeding in a tropical orb-weaving spider. Ecol- ogy, 56:1314-1328. Chamov, E. L. 1976. Optimal foraging, the mar- ginal value theorem. Theor. Pop. Biol., 9:129-136. Dadd, R. H. 1985. Nutrition: organisms. Compre- hensive Insect Physiology, Biochemistry and Phar- macology., Vol. 4, pp. 313-390. Eberhard, W. G. 1971. The ecology of the web of Uloborus diver sus (Araneae: Uloboridae). Oecologia (Berlin), 6:328-342. Enders, F. 1976. Effects of prey capture, web de- struction and habitat physiogamy on web-site te- nacity of Argiope spiders. J. Arachnol., 3:75-82. Gillespie, R. G. 1987. The role of prey availability in aggregative behavior of the orb weaving spider Tetragnatha elongata. Anim. Behav., 35:675-681. Higgins, L. E. 1990. Variation in foraging invest- ment during the in term oil and before egg laying in the spider Nephila clavipes. J. Insect Behavior, 3:773- 783. Higgins, L. E. 1992a. Developmental plasticity and fecundity in the orb-weaving spider Nephila cla- vipes. J. Arachnol., 20:94-106. Higgins, L. E. 1 992b. Developmental changes in the barrier web under different levels of predation risk. J. Insect Behavior, 5:635-655. Higgins, L. E. 1993. Constraints and plasticity in the development of juvenile Nephila clavipes in Mexico. J. Arachnol., 21:107-1 19. Higgins, L. E. & R. Buskirk. 1992. A trap-building predator exhibits different tactics for different as- pects of foraging behavior. Anim. Behav., 44:485- 499. Higgins, L. E. & M. A. Rankin. In press. Different pathways in arthropod post-embryonic develop- ment. Evolution. Lubin, Y. D. 1973. Web structure and function: the nonadhesive orb-web of Cyrtophora moluccensis (Araneae: Araneidae). Forma et Functio, 6:337-358. Mangel, M. & C. Clark. 1986. Towards a unified foraging theory. Ecology, 67:1 127-1 138. Pianka, E. R. 1981. Resource acquisition and al- location among animals. Pp. 300-314, In Physio- logical Ecology: An evolutionary approach to re- source use (Townsend, C.R. & P. Calow, eds.). Blackwell Scientific Press, Boston. Pyke, G. H., H. R. Pulliam, & E. L. Chamov. 1977. Optimal foraging: A selective review of theory and tests. Quart. Rev. Biol., 52:137-154. Riechert, S. E. & J. Luczak. 1982. Spider foraging: Behavioral responses to prey. Pp. 353-386, In Spi- der Communication: Mechanisms and ecological significance (P. N. Witt & J. S. Rovner, eds.). Prince- ton University Press, Princeton. Schoener, T. W. 1971. Theory of feeding strategies. Ann. Rev. Ecol. Syst., 11:369-404. Sherman, P. M. 1994. The orb-web: an energetic and behavioural estimator of a spider’s dynamic foraging and reproductive strategies. Anim. Behav., 48:19-34. Sokal, R. R. & F. J. Rohlf. 1981. Biometry. The Principles and Practice of Statistics in Biological Research, 2nd ed., W. H. Freeman & Co., San Fran- cisco. Stephens, D. W. & J. R. Krebs. 1 986. Foraging The- ory. Princeton University Press, Princeton. 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- 464. Tillinghast, E. K. 1984. The chemical fractionation of the orb web of Argiope spiders. Insect Biochem., 14:1 15-120. Townley, M. A., D. T. Bernstein, K. S. Gallagher & E. K. Tillinghast. 1991. Comparative study of orb web hygroscopicity and adhesive spiral composition in three araneid spiders. J. Exp. Zool., 259: 1 54-165. Turnbull, A. L. 1 964. The search for prey by a web- building spider. Canadian Entomol., 96:568-597. Wise, D. H. 1975. Food limitation of the spider Linyphia marginata : Experimental field studies. Ecology, 56:637-646. Wise, D. H. 1979. Effects of an experimental in- crease in prey abundance upon reproductive rates of two orb-weaving spider species. Oecologia, 41: 289-300. Manuscript received 16 August 1994, revised 25 October 1994. 1995. The Journal of Arachnology 23:44-45 RESEARCH NOTES NESTS OF HIBANA GRACILIS ARE REUSED BY PHIDIPPUS CLARUS IN WETLANDS OF NORTHEASTERN KANSAS Typical nests of wandering spiders consist of small, tightly packed hollow bundles. In these nests spiders molt, retire for part of the day and often lay their eggs (Jackson 1979; Foelix 1982). Nests may be constructed on the ground in nat- ural shelters such as leaf litter or woody debris or within a plant canopy among easily folded leaves or other structures (Johnson 1992). In wetlands of northeastern Kansas in early May of 1993 and 1994, I found many mating pairs of Hibana gracilis (Hentz) (Araneae, An- yphaenidae) within nests made in the small up- Figure l.— Mating pair of Hibana gracilis (Araneae: Anyphaenidae) in the uppermost expanded leaves of the milkweed plant. 44 RESEARCH NOTES 45 permost expanded leaves of common milkweed ( Asclepias syriaca) and sullivant’s milkweed (A. sullivantii ) (Fig. 1). Because these milkweed spe- cies may be more common in annually burned wetlands (Johnson & Knapp, in press), I quan- tified the density and diversity of H. gracilis nests in three annually burned wetlands. Each wetland was approximately 350-500 m2 in area and con- tained approximately 5.4 A. syriaca or A. sulli- vantii stems/m2 (Johnson, in press). Hibana gracilis nests were found on 72% of milkweed stems in 40 plots, making nest density among milkweed in these wetlands 3.88 nests/m2. Nests were observed in this density on either A syriaca or A. sullivantii depending on which species was more centrally located within wetlands. Other- wise, H. gracilis nests were found within basal leaf sheaths of prairie cordgrass {Spartina pec- tinata ) in a density of 0.4 nests/m2 (n = 30 m2 plots of S. pectinata). By late May, H. gracilis had abandoned most nests; but the majority of these nests were reoc- cupied by Phidippus clarus Keyserling (Araneae, Salticidae). In the same three wetlands described above, 64% ( n = 40) of former H. gracilis nests in milkweed stems were occupied by a female or a pair of P. clarus. Furthermore, these nests did not appear to have been altered in any way from the original structures made by H. gracilis. By mid-June 1994, P. clarus had also abandoned these milkweed leaf nesting sites. This pattern of nest reuse represents an inter- esting interaction between wandering arachnids and may be linked to 1) shortage of suitable nest- ing sites, 2) reoccupation of optimal nest sites, or 3) silk conservation. Young milkweed plants may represent ideal nesting sites which are shared by sympatric arachnids with similar hunting methods and prey choices, but allochronic mat- ing periods. ACKNOWLEDGMENTS I thank Dr. Bruce Cutler for his contributions in species identification and helpful advice. I also thank the Konza Prairie LTER Staff. LITERATURE CITED Foelix, R. F. 1 982. Biology of Spiders. Harvard Uni- versity Press, Cambridge. Jackson, R. R. 1979. Nests of Phidippus johnsoni (Araneae, Salticidae): characteristics, patterns of oc- cupation, and function. J. Arachnol., 7:47-58. Johnson, S. R. 1993. Jumping spiders nest in red plastic area marker flags in prairies of northeastern Kansas. Prairie Nat., 25:275. Johnson, S. R. In press. Wind-induced leaf binding by Spartina pectinata onto Asclepias syriaca in northeastern Kansas. Prairie Nat. Johnson, S. R. & A. K. Knapp. In press. The influence of fire on Spartina pectinata wetland communities in a northeastern Kansas tallgrass prairie. Canadian J. Bot. Stephen R. Johnson: Division of Biology, Ack- ert Hall, Kansas State University, Manhattan, Kansas 66506 USA. Manuscript received 31 October 1994, revised 22 Jan- uary 1995. 1995. The Journal of Arachnology 23:46-47 FLEXIBILITY IN FORAGING TACTICS OF BUTHUS OCCITANUS SCORPIONS AS A RESPONSE TO ABOVE-GROUND ACTIVITY OF TERMITES Foraging modes are known to be generally fixed within a species, and flexibility of foraging tactics is often restricted by physiological and morpho- logical traits, or by predation pressure (Huey & Fianka 1981). However, models and empirical evidence suggest that individuals may change foraging tactics as a response to variations in prey abundance (Schoener 1971; Stephens & Krebs 1986; Formanowicz & Bradley 1987). During research in the Negev desert of southern Israel in summer 1993, I observed scorpions of the species Buthus occitanus Israelis switching for- aging tactics as a response to sporadic above- ground activity of termites. Buthus occitanus Israelis is a burrowing scor- pion found In arid zones of Israel and Sinai (adult body length 5-7 cm) (Levy & Amitai 1 980). Like many other species of scorpions (Polls 1990), these are nocturnal, sit and- wait predators that stand In ambush position some distance away from their burrows and locate prey by sensing air and soil vibrations. However, active foraging was observed as well and may constitute an al- ternative foraging tactic. Of all the individuals that were observed foraging on 30 nights (120 hours of observation), only 20% of the males and 12.5% of the females were moving, while most were motionless in ambush positions. When ap- proached, these cryptic scorpions usually re- mained motionless, attempting to escape only if they were within 10 cm of their burrows. Buthus occitanus scorpions are opportunistic, generalist foragers. Foraging scorpions in the field and in the laboratory accept various types of prey - including crickets, Neuroptera, moths, bugs and various arachnids. However, in field observa- tions, 70% of the scorpions consuming prey that could still be identified in = 27), ate the wingless worker and soldier castes of harvester termites, Anacanthotermes sp. (specimens have been de- posited at the National Collection in Tel- Aviv University). These termites live underground and forage on the surface only in the vicinity of their burrows (usually within 70 cm of the entrance). When they encounter a predator they retreat into their nest within seconds. Observations revealed that widely-foraging B. occitanus scorpions approach the active termites and swiftly sting as many as possible before the termites disappear into the shelter of their burrow. The scorpion then moves around the area and collects the dead and dying termites with its pedipalps and chelicera. This foraging tactic was used by scorpions of all ages. On three occasions young scorpions were seen catching 1-2 termites at once, while on three other occasions large adult scorpions succeeded in collecting up to eight termites in a single at- tack. In this unique foraging tactic B. occitanus scor- pions exhibit a number of specialized behaviors that enhance the profitability of hunting for ter- mites: a) since the termites forage within a small range around their burrow, the scorpion must actively forage for them rather than employ the usual sit-and-wait tactic; and b) termites are small and possess an effective alarm system. To use the short time available before the termites dis- appear, the scorpion must skillfully sting as many individuals as possible, without wasting time on handling and collecting. A comparable termite hunting tactic was observed in the web-building spider Chrosiothes tonala that specializes on ter- mites (Eberhard 1991). Collecting dead, motionless termites requires use of sense organs other than the mechanical receptors that are used by most scorpions to de- tect movement of live prey. Krapf (1986) showed in laboratory conditions that B, occitanus scor- pions used contact chemoreception to detect mo- tionless prey. Chemoreceptors may play an im- portant role in hunting of termites, by allowing the scorpions to separate the short time available for attacking and subduing the prey from the time-consuming process of collecting and han- dling the dead prey. Sit-and-wait foraging has been described as a low-cost, low-profitability strategy that may have evolved in species under predation pressure (Huey & Pianka 1981). By switching from their normal mode of ambush foraging into a widely- foraging mode of termite hunting, B occitanus 46 RESEARCH NOTES 47 scorpions may enjoy increased foraging success. However, a widely-foraging scorpion might also have to endure greater metabolic costs of for- aging and increased risk of predation. ACKNOWLEDGMENTS I wish to thank Yad Lubin and Yoni Brandt for making helpful comments on the manuscript, and Danny Simon for identifying the termites and commenting on their behavior. The research was sponsored by the Explorers Club Grant-in- aid of Research, and by the Inter-University Ecology Fund of the Keren Kayemet in Israel. This is publication number 191 of the Mitrani Center for Desert Ecology. LITERATURE CITED Eberhard, W. G., 1991. Chrosiothes tonala (Ara- neae, Theridiidae): a web-building spider special- izing on termites. Psyche, 98:7-19. Formanowicz, D. R. & P. I. Bradley. 1987. Fluc- tuations in prey density: effects on the foraging tac- tics of scolopendrid centipedes. Anim. Behav., 35: 453-461. Huey, R. B. & E. R. Pianka. 1981. Ecological con- sequences of foraging mode. Ecology, 62:991-999. Krapf, D. 1986. Contact chemoreception of prey in hunting scorpions (Arachnida: Scorpiones). Zool. Anz., 217:1 19-129. Levy, G. & P Amitai. 1980. Fauna Palestine Arachnida I: Scorpiones, 1st Edition. Israel Acad. Sci. and Human., Jerusalem. Polis, G. A. 1990. Ecology. Pp. 247-294, In The Biology of Scorpions, 1st Ed. (G. A. Polis, ed.). Stan- ford Univ. Press, Stanford, California. Schoener, T. 1971. Theory of feeding strategies. Ann. Rev. Ecol. Syst, 2:369-404. Stephens, D. W, & J. R. Krebs. 1 986. Foraging the- ory. Princeton Univ. Press, Princeton. Orit Skutelsky: Mitrani Center for Desert Ecology, Blaustein Institute for Desert Re- search, Sde-Boqer Campus, Israel 84990. Manuscript received 12 June 1994, revised 31 October 1994 . 1995. The Journal of Arachnology 23:48-50 BOOK REVIEW Wise, D. H. 1993, Spiders in Ecological Webs. Cambridge University Press. ISBN 0-521-32547-1 (Price $79.95) In this 328 page book, David Wise offers a comprehensive discussion of the literature on the use of spiders to test ecological models and the- ories. Until now, most books on spiders have been edited volumes that bring together the ideas of many others (e. g,, Nentwig 1987; Shear 1986; Witt & Rovner 1982); equally valuable are the more general discussion of spiders (Foelix 1982) and taxonomic catalogs (Platnick 1989, 1993). David Wise’s book is a welcome addition to this growing literature because he brings the focus of a single author while summarizing and providing a critique of the work of many researchers. His emphasis is on describing and critiquing field experimentation (especially in the context of studying competition), and on studying spiders both as predators and as models of generalist predators in terrestrial systems. He takes a look at spiders in agricultural systems as well. Wise brings his considerable experience as a com- munity spider ecologist to his discussion and re- view of this literature. There are nine chapters in this book, starting with one entitled “The Spider in the Ecological Play”, which sets spiders as terrestrial, generalist predators onto the center stage of the ecological and evolutionary drama. This is followed by a progression of chapters first on hungry spiders (the importance of food limitation) and then se- ries of chapters on competition: competitionist views of spider communities, failure of the com- petitionist paradigm and how (some) spiders may avoid competition. Following this explicit look at competition, there are chapters on the impact of spiders on insect populations, “anchoring the ecological web” (subtitled “refining the meta- phor-the web’s non-trophic threads” which is a look particularly at the architecture of the veg- etation and leaf litter and how these factors affect the abundance of spiders), “untangling a tangled web” (in which Wise addresses “indirect factors” that help to structure complex communities, in- cluding intraguild predation), and finally, one en- titled “spinning a stronger story” where Wise closely scrutinizes his own biases in the book. Wise recognizes the limitation of the metaphor of spinning and webs that he uses throughout the book even as he attempts to use the metaphor to tease apart the threads of community struc- ture. Each chapter begins with a 1-2 page intro- duction and concludes with a synopsis of the main ideas. The format makes the information in the book easily accessible. And while the book “hangs together” nicely as a whole unit, it is also easy to focus on one chapter at a time. I was able easily to go back to different chapters and find specific information. The reference section includes 462 works. The literature spans from 1 8 1 5 to 1992 with the bulk of the citations (249 entries) from the 1980’s. Also included is a name index (for authors) and a subject index (with spider names and topics interspersed). The indexes are useful: the subject index includes family names, genus-species names and common names for the spiders. The name index works for second and third authors as well as for first authors. Spiders in Ecological Webs will be of interest to a wide variety of ecologists, arachnologists and other biologists. I will use examples from this book in my 200-level Ecology class and will buy it for my school’s (undergraduate) library. It will be useful to all researchers interested in spiders and/or interested in field experimentation. Wis- e’s critiques of methods and statistics will be helpful to graduate students, professional ecol- ogists and anyone planning field experimenta- tion. For a variety of reasons, many researchers (including Wise, in his earlier days) have fallen into the trap of “pseudoreplication” in planning and carrying out field studies - Wise provides a particularly useful and extensive discussion of this problem. Because of this critique, Spiders in Ecological Webs will be useful in Biostatistics courses or in Research Methodology courses as well. This book has several strengths. Clearly one of these is the detailed attention to studies of spiders in competition and studies of competi- tion in which spiders play a big role. Wise clearly 48 BOOK REVIEW 49 shows through this discussion how ideas about competition have changed dramatically in the last few decades. He very capably discusses the question “Is there evidence that spiders compete or that competition is important in structuring ecological communities?” A second major strength of the book is Wise’s attention to field studies and field manipulation and experimen- tation. This emphasis makes this work useful to a wide range of ecologists, even those without an interest in arachnids. However, the central strength of the book lies in Wise’s willingness to present a wide variety of studies, critique them closely, reanalyze data if need be and show how (even with a particular study’s weakness), in- sights can be gleaned from it. For example, the study by Clarke & Grant (1968) (discussed on pp. 147ff) is a classic study in leaf litter ecology and probably introduced the idea of field ma- nipulation to many ecologists. This study has been cited to support the idea that spiders have a significant impact on insect populations in leaf litter ecosystems. While this is an important and classic paper, Wise explains its flaws in terms of experimental design (in this case, primarily a lack of replication). This study has been followed by a variety of manipulative field experiments by Bultman & Uetz, Wise & Wagner and Kajak & Jakubczyk, among others. I think some of the manipulations done by later researchers would have amazed Clarke and Grant. Throughout reading the book, I felt a sense that there really is progress in science. You can see through the reviewed literature that studies have been done more and more carefully and researchers have learned what is necessary to tackle a problem appropriately and in many cases have designed and executed experiments that ecologists didn’t even imagine 20 years ago as in the case with Clarke and Grant. To his credit, Wise does not hesitate to criticize his own research. In Chapter 5, “How spiders avoid competition,” he reworks much of his own data and in some cases, comes to different con- clusions. For example, following his reanalysis, he finds that contrary to his earlier assertions, there is no evidence that intraspecific competi- tion affects fecundity in spring maturing filmy dome spiders. To me, this is the epitome of sci- ence - when an individual synthesizes from his/ her own work and from the work of others and moves forward in understanding a particular sys- tem. Wise does not cover the extensive and growing literature on spider behavioral ecology, nor does he attempt to cover the realms of spider evolu- tionary ecology or sensory ecology. Tackling these broad topics would have completely changed the nature of the book and probably would have made it prohibitively large. Other books on these topics remain to be written. Although there is some discussion of cursorial spiders, Wise emphasizes web building spiders in the northern hemisphere probably because many ecological studies have been on web build- ers in the northern hemisphere. Something Wise does not do in this book is to remind the readers of the number of terrestrial ecosystems that have not been studied as extensively by field experi- mentalists. I wonder what a “model spider” would be in a tropical system where the vast majority of collected spiders are “singletons”? Such sys- tems are much more difficult to manipulate, yet because of the vastness of the tropical biome, generalities without considering it are problem- atic. I think it is always useful to put our knowl- edge in context - to remind ourselves of the big- ger picture and what and how much we do not know. It is also worth mentioning that we prob- ably still know only a fraction of the all of the spider species. What is a “general spider” (a term Wise uses), and can we know if, as Coddington & Levi (1991) suggest, only 20% of the world’s spider fauna is known? There are more species of Salticidae than any other family, and among the six most speciose families, there are more species of non-web builders (8800 species) than there are of web building spiders (8500 species) (Coddington & Levi 1991). Perhaps a “general spider” should be a salticid spider from a tropical biome. Some of Wise’s terminology is problematic. For example, his use of the group “cribellate spi- ders” (p. 6) is unfortunate as this no longer a valid grouping (Coddington 1990). Also unfor- tunate is Wise’s groupings of “closely related spe- cies” -such as the Agelenidae and the Lycosidae (p. 8) - two families that are not particularly close (Coddington & Levy 1991). Wise only briefly touches on the very interesting phenomenon of sociality in spiders, particularly in tropical areas. Perhaps that will be covered in a another book on evolutionary ecology in spiders. While I was bothered at first by Wise’s attempt to generalize a single “spider persona”, I found myself enjoying the book more and more as I read through it. I particularly enjoyed being able to trace a series of papers done by one person 50 THE JOURNAL OF ARACHNOLOGY (or group of persons) and be able to see the evo- lution of thinking in various spider ecologists. It was also very helpful to see how frequently many researchers have tackled a particular problem; and taken together, we really do have a lot of insight on particular questions. I got a sense of continuity from this and satisfaction that I could go to one source and find such useful summaries and find so many references. On the whole this is a welcome addition to both the spider literature and the ecological lit- erature. I think Wise’s goal of introducing spiders to a wider ecological audience will be met in this book. It will also make accessible a large ecolog- ical literature to other arachnologists. The book left me hopeful on several counts - I hope that the momentum of research of the 1980’s contin- ues through the 1 990’s and into the next century. I hope future editions will take into account more from tropical ecosystems - and that there will be more research to report and summarize from other regions. And finally, I hope there will be other books on spider behavioral ecology and evolutionary ecology. It is an exciting time to be working on spiders. LITERATURE CITED Clarke, R. D. & P. R. Grant. 1968. An experimental study of the role of spiders as predators in a forest litter community. Part 1. Ecology, 49:1 152-1 154. Coddington, J. A. 1990. Cladistics and spider clas- sification: Araneomorph phylogeny and the mono- phyly of orb weavers (Araneae: Araneomorphae: Orbiculariae). Acta Zool. Fennica, 190:75-87. Coddington, J. A. & H. W. Levi. 1991. Systematics and evolution of spiders (Araneae). Annu. Rev. Ecol. Syst., 22:565-592. Foelix, R. 1982. Biology of Spiders. Harvard Uni- versity Press. Cambridge, Massachusetts. Nentwig, W. 1987. Ecophysiology of Spiders. Berlin, New York, London, Paris and Tokyo. Springer. Platnick, N. I. 1989. Advances in Spider Taxonomy 1981-1987. Manchester Univ. Press, Manchester and New York. 673 pp. Platnick, N. I. 1993. Advances in Spider Taxonomy 1 988-199 1 . New York Entomol. Soc. and American Mus. Nat. Hist., New York. 846 pp. Shear, W. A. 1986. Spiders - Webs, Behavior, and Evolution. Stanford Univ. Press, Stanford, Califor- nia. Witt, P. N. & J. S. Rovner. 1982. Spider Commu- nication. Princeton Univ. Press, Princeton. Gail E. Stratton. Dept, of Biology, Albion Col- lege, Albion, Michigan 49224 USA Manuscript received 20 January 1995. 1995. The Journal of Arachnology 23:51-54 BOOK REVIEW Platnick, N. I. 1993. Advances in Spider Tax- onomy 1988-1991 With Synonymies and Trans- fers 1940-1980 (edited by P. Merrett). The New York Entomological Society, New York. Ask any reasonably competent biologist and you will be told that evolution is change or ad- vance but that change does not necessarily equal progress. Advances may be made to the rear. Science has evolved (advanced) by leaps and bounds sideways, backwards, tangentially, and occasionally (but generally, I believe) forward over time. This evolution has been guided in part by the blinders of religious and political ideol- ogies and the vagaries and frustrations of fund- ing, peer review, turf wars, hidden agendas, and a bewildering array of ‘‘refereed” publications. Thus have spider taxonomy and systematics evolved over the last two centuries: a generally progressive trend emerging from apparent chaos as myriads of independent workers have pursued their independent taxonomic agendas under the one common (but vague) guiding principle of discovering (real or imagined) order in the con- fusing profusion of life forms. Battles rage, die, and then flare up again over species concepts and systematic methodologies but some constants re- main. First a species is whatever a researcher can convince his or her colleagues it is. Second, and more important, at the roots of all good biology are good phylogenies. Third, and most impor- tant, good phylogenies are rooted in good tax- onomy. The end result of all this is names and publications. Many, many names in many, many publications. That advances (progress) have been made at all in spider taxonomy and systematics is, at least in part, due to the unique and increas- ingly useful and user-friendly series of spider tax- onomic catalogues culminating in the most re- cent volume by Norman Platnick. Others have discussed the merits and short- comings of the araneological catalogues of Bon- net, Roewer, and Brignoli. In spite of the serious problems caused by Brignoli’s omission of syn- onymies and transfers of pre-Roewer names these works have been of immense value. Platnick’ s two volumes have continued and improved upon the works of his predecessors. The first, Advances in Spider Taxonomy 1981-1987 (1989), kept abreast of post-Brignoli developments listing ci- tations of illustrated taxonomic works and cat- aloguing new names, synonymies, and transfers published from 1981 to 1987. The second (cur- rent) volume does the same for the period 1988 to 1991 with the added bonus of listing all the synonymies and transfers omitted by Brignoli (1940-1980). It is now possible, as Platnick (with some modesty) points out in his introduction, “. . . using Roewer and the three supplemental volumes [of Brignoli and Platnick], to determine what genera belong to each family (along with their synonyms) and what species belong to each genus (along with their synonyms)”. This has been no mean feat but the catch-up job remains incomplete in one area. Still lacking are the ci- tations for the post- 1940 literature containing illustrated redescriptions not involving transfers or synonymies. It is hoped that this data {circa 1.5 megabytes) will be presented in the next sup- plement (Platnick, pers. comm.), pending ade- quate funding. Stet fortuna domus. The current catalogue is a weighty tome (846 pages) and much less visually exciting than the most recent catalogue from L. L. Bean. However, unlike the latter, the usefulness of Advances in Spider Taxonomy 1988-1991 is not time con- strained and I noted no apparent errors within it, typological or otherwise. A short tribute to the long standing, competent, and quiet technical support of Louis Sorkin is an appropriate open- ing to the volume. Following this, the Introduc- tion outlines Platnick’s rationale for deciding what to include in, what to omit from, and the pre- sentation format of his treatment. Platnick notes that the material presented is split roughly in half between new literature published from 1988 through 1991 and the synonymies and transfers missed in the previous catalogues. An effort has been made to avoid repetition of material suf- ficiently treated in the previous works. This has kept the published bulk to a manageable size but also means that researchers still need a complete set of Roewer (and/or Bonnet), Brignoli, Plat- nick, and Platnick to have as thorough a treat- ment of the taxonomic picture as is probably possible. Formatting of 1988-1991 follows the style oil 98 1-1 987. 51 52 THE JOURNAL OF ARACHNOLOGY A list of families shows we are status quo with respect to the number (105) listed in 1981-1987. A comparison of the two lists reveals minor changes including three new families (Synotax- idae, Trechaleidae, Lamponidae) and two res- urrections (Zoropsidae, Prodidomidae) balanced by five sinkings (Loxoscelidae, Hadrotarsidae, Dolomedidae, Platoridae, and Aphantochilidae absorbed by Sicariidae, Theridiidae, Pisauridae, Trochanteriidae, and Thomisidae, respectively). Families are listed in a one-dimensional reflec- tion of the current consensus of opinion on clas- sification. The family lists of Brignoli and 1981- 1987 reflected the major upheavals in spider classification resulting from the 1967 trashing of Cribellatae by Lehtinen and the general accep- tance through the 70’s and 80’s of cladistic meth- odology as the best way to divorce art from sci- ence in classification. It is somewhat of a relief to see no evidence in 1988-1991 of new major changes. Advance (progress) in spider higher classification during this short period has been limited to the support by Platnick et al. of the monophyly of Haplogynae (16 families from Fil- istatidae to Orsolobidae). This leaves only the question of the monophyly of the major (and troublesome) “RTA clade” (43 families from Ly- cosidae to Salticidae) unanswered (in spite of some excellent work within this group by Gris- wold, Platnick, Sierwald, and others on “dictyn- oids”, gnaphosoids, “amaurobioids”, and lyco- soids). Life continues. The Bibliography lists by year approximately 1700 publications referenced in the text. A small handful (38) covers works from 1867 to 1939 missed or with material excluded from the pre- vious catalogues (or otherwise needing repeti- tion). Over 900 entries relate to the synonymies and transfers from 1 940 to 1 980 omitted by Brig- noli. Some 73 references from 1981 to 1987 cov- er mostly material missed by 1981-1987. The remaining 653 references (averaging a fairly con- sistent 1 60 per year) are the new taxonomic pub- lications appearing around the world from 1988 to 1991. From there we go into the meat of the matter. Sanity constraints dictated that I concentrate my review efforts upon a particular aspect of the approximately 750 pages of the Catalog of Gen- era and Species. Synonymies of genus names and transfers and synonymies of species names are adequately cross-referenced under the appropri- ate family and genus headings. Every currently valid genus name is presented. All this is im- mediately evident from a casual inspection of the listings and is also explained in the Intro- duction. In my review of this section I did not make an attempt to keep track of the new sink- ings and other changes but concentrated on tal- lying up new species and genus names appearing from 1988 to 1991. This process was logistically simple, was an adequate indication of in what groups research is most active, and gave me an idea of how close we have come to the mythical figure of 40,000 described spider species. Roughly 1 80 new genera and 2070 new species were described during the four year period 1 988— 1991. In comparison to Coddington’s counts of 230 and 2581 for new genera and species de- scribed in the preceding seven year period there appears to be some consistency in a new species to new genera ratio of about 11:1 coupled with a substantial upswing in the curve of descriptive activity. Obviously there are still lots of spider species out there to be described and the reduced cadre of professional taxonomists is working harder than ever on the task. Further comparison of these figures with Platnick’s earlier estimate of roughly 3000 and 34,000 described genera and species in total shows the 11:1 ratio also is con- sistent with the historical trend in spider tax- onomy. Without considering new synonyms about 36,000 spider species have now been de- scribed. Coddington and Levi recently presented statistics in support of an estimate of approxi- mately 170,000 extant spider species. In a world reluctantly coming to acknowledge a) the car- dinal importance of arthropods in the faunal component of all ecosystems and b) the historical “megafaunal” bias in zoological inquiry (dis- cussed by Platnick elsewhere), perhaps there is cause for renewed hope for the future of spider taxonomy. What could we do with some serious funding for baseline “biodiversity” inventories? All the following species and genus numbers above 50 are approximate and rounded to the nearest 1 0. Not surprisingly, new genera and spe- cies in the suborder Opisthothelae ( 1 04 families, 180 new genera, 2060 new species) vastly out- number those in Mesothelae (1 family, no new genera, 1 4 new species). Similarly the bulk of new descriptions in Opisthothelae are in Araneo- morphae (89 families, 1 80 new genera, 1 960 new species). Mygalomorphae ( 1 5 families) has 5 new genera and 90 new species. Further down the araneomorph classificatory trail it is hardly sur- prising there has been no activity among the pa- leocribellates (I mean, how many new hypochil- BOOK REVIEW 53 ids do you expect there are left to find?) or the austrochiloid neocribellates (ditto). Within Araneomorphae (and Araneae in gen- eral) Araneoclada is where the action is hottest and some observations of descriptive trends within this grouping (note that four genera ac- count for well over 1 0% of all new spider species described) have prompted my proposal for a se- ries of awards suggested below. The newly ver- ified group Haplogynae (16 families) shows five new genera and 120 new species (nearly 90 of which are in the two dysderid mega-genera Dys- dera and Harpactea). Entelegynes, of course, ac- count for all the rest. Most activity was registered in Araneoidea (1 1 families) with nearly 130 new genera and 1 060 new species. Paramount in this superfamily are the linyphiids with over 80 new genera (many of which are monotypic) and near- ly 560 new species. This accounts for close to half of all new genera and over one quarter of all new species in the entire order Araneae. Other notables within Araneoidea include the synotax- ids (10 new genera, 50 new species), theridiids (3 and 52), anapids (21 and 60) and araneids (6 and 300). Fully two-thirds of the araneid new species are in the two mega-genera Araneus and Alpaida. One can only hope that with this level of activity we must be getting close to resolving the familial relationships within Araneoclada. The “RTA clade” has been relatively quiet (44 new genera, 740 new species) given its size (43 families). Salticid descriptions account for about a quarter of the new names with 1 1 new genera and 1 80 new species. Zodariids have 14 new gen- era and nearly 90 new species; amaurobiids have 7 and 1 10 (with over 60 new species in Coelotes). Clubionids and gnaphosids each have one new genus and over 50 new species (49 in Clubiona ); thomisids have over 50 new species as well. “Also ran” RTA’s include the lycosids (1 new genus, 39 new species) and heteropodids (3 new genera, 30 new species). For some years the araneological world has been poised, somewhat breathlessly, awaiting new revelations into the nature of the RTA clade (is it real or just an infatuation?) and its putative major sub-groupings Dionycha (22 families) and the “dictynoids” (7? families), “amaurobioids” (4? families), and lycosoids (10? families). Dic- tynoids and amaurobioids, with their tattered and tom remnants of many of the old, pre-Leh- tinen cribellate groupings, pose the biggest prob- lems for systematists (just what is a cybaeid, a dictynid, a hahniid, an agelenid, or even an amaurobiid anyway?) and yet 1987-1991 shows very little work (with the exception of Amau- robiidae) in these groups. The calm before the storm. . .? Concluding the Catalog is a short listing of all new nomenclatorial changes to be found in the previous 750 pages. Surprisingly there are very few and wherever possible Platnick has used new names provided by the original authors. Thus there are nine and 1 1 new synonyms, replace- ment names, or transfers of genera and species respectively. All the specific synonymies are from Platnick’s own work with gnaphosid type ma- terial. At this point, with tongue firmly planted in cheek (and hoping none take offense), I would like to propose the creation of the R. V. Cham- berlin Araneological Olympics with awards pre- sented to recognize conspicuous advances in ar- aneological taxonomy. For the first such Olym- pics I have identified a small number of com- petitive categories and chosen award winners from the listings in 1988-1991. Certainly some- one with a more active imagination and better database skills than I could expand upon the following. For “Most Species Described in One Genus” the gold medal goes to the United States for H. W. Levi’s work with Alpaida (94 new spe- cies), the silver goes to England (A. F. Millidge, 79 new species in Dubiaranea ), and the bronze to the United States (H. W. Levi again, 7 1 new species in Araneus). For “Most Monotypic New Genera in One Family” the gold medal is award- ed to England for A. F. Millidge’s 25 new genera in Linyphiidae (no silver or bronze medals were awarded in this category). For “Most Species Transferred to Other Genera” the clear winner of the gold medal is Araneus (363 transfers). In a distant second place the silver medal winner is Dendryphantes (210 transfers). Finally, there is one award to be presented in the N. I. Platnick Araneological Special Olympics for “Best New Name.” This category has been inspired gener- ally by the evidence of subtle taxonomic hu- morists among our ranks and specifically by the spooneristic, cinematically inspired name Apo- pyllus now (“an arbitrary combination of letters” indeed). Undoubtedly the following presentation is debatable, based as it is upon the decision of one judge, but for the period 1988 to 1991 the winner is Poland for W. Zabka’s new genus and species Abracadabrella birdsville. Congratula- tions to all. In closing I do not apologize for quoting from 54 THE JOURNAL OF ARACHNOLOGY Coddington’s review of the previous volume. . . Advances in Spider Taxonomy [1988-1991] is a splendid volume. I do not have to recommend that you buy it, because you already know that it is indispensable. Arachnologists and beyond owe Platnick a fervent thanks, because few works are as critical to good biology as nomenclatorial catalogs. If taxonomy is the sina qua non of all biological science, it is because of works such as this.” Here is an advance that definitely equates with progress. Robert Bennett: British Columbia Ministry of Forests, 7 3 80 Puckle Road, Saanichton, Brit- ish Columbia V8M 1W4 Canada Manuscript received 24 October 1994. INSTRUCTIONS TO AUTHORS (revised July 1995) Manuscripts are preferred in English but may be ac- cepted in Spanish, French or Portuguese subject to availability of appropriate reviewers. Authors whose pri- mary language is not English may consult the Associate Editor for assistance in obtaining help with English manuscript preparation. Ah manuscripts should be pre- pared in general accordance with the current edition of the Council of Biological Editors Style Manual unless instructed otherwise below. Authors are advised to con- sult a recent issue of the Journal of Arachnology for additional points of style. Manuscripts longer than 1 500 words should be prepared as Feature Articles, shorter papers as Research Notes. Send four identical copies of the typed material together with copies of illustrations to the Associate Editor of the Journal of Arachnology: Petra Sierwald, Associate Editor Division of Insects Dept, of Zoology Field Mluseum Roosevelt Road at Lakeshore Drive Chicago, IL 60605 USA [Telephone: (312) 922-9410, ext. 841; FAX: (312) 663-5397; Electronic mail: SIERWALD@FMNH.ORG], Correspondence relating to manuscripts should also be directed to the Associate Editor. After the manu- script has been accepted, the author will be asked to submit the manuscript on a computer disc, preferably in MS DOS WordPerfect. FEATURE ARTICLES Title page. — The title page will include the com- plete name, address, and telephone number of the au- thor with whom proofs and correspondence should be exchanged, a FAX number 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 capital letters should be placed at the beginning of the first paragraph set off by a period. In articles written in English, a second abstract in an acceptable language may be included pertinent to the nationality of the author(s) or geo- graphic region(s) emphasized in the article. A second abstract in English must be included in articles not written in the latter language. 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 head 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 underlined and sep- arated 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 period regardless of language of the text. Citation of references in the text: Cite only papers already published or in press. Include within parenthe- ses the surname of the author followed by the date of publication. A comma separates multiple citations by the same author(s) and a semicolon separates citations by different authors, e.g., (Smith 1970), (Jones 1988; Smith 1993), (Smith 1986, 1987; Smith & Jones 1989; Jones et al. 1990). Literature cited section. — Use the following style: Lombardi, S. J. & D. L. Kaplan. 1990. The amino acid composition of major ampullate gland silk (dragline) of Nephila clavipes (Araneae, Tetragnath- idae). J. Arachnol., 18:297-306. Krafft, B. 1982. The significance and complexity of communication in spiders. Pp. 15-66, In Spider Communications: Mechanisms and Ecological Sig- nificance. (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. Running head. — The author surname(s) and an ab- breviated 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 con- tact the Associate Editor. Tables. — Each table, with the legend above, should be placed on a separate manuscript page. Only hori- zontal lines (usually three) should be included. Use no footnotes; instead, include all information in the leg- end. Make notations in the text margins to indicate the preferred location of tables in the printed text. Illustrations. — Address all questions concerning il- lustrations to the Editor of the Journal of Arachnology: James W. Berry, Editor Dept, of Biological Sciences Butler University Indianapolis, Indiana 46208 USA [Telephone (317) 283-9344; FAX (317)-283-9519; Electronic mail: BERRY@BUTLER.EDU] All art work must be camera-ready for reproduction. In line drawings, pay particular attention to width of lines and size of lettering when reductions are to be made by the printer. Multiple photos assembled on a single plate should be mounted with only a minimum of space separating them. In the case of multiple illus- trations 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. The author should indicate whether the illustrations should be one column or two columns in width. The overall dimen- sions should be no more than 11 inches (28 cm) X 14 inches (36 cm). Larger drawings present greater diffi- culty in shipping and greater risks of damage for which the JOA assumes no responsibility. Make notations in the text margins to indicate the preferred position of illustrations in the printed text. 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-us, 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, holotype male; 33, 34, A-us y-us, male. Scale = 1.0 mm. Assemble manuscript for mailing. — Assemble the separate sections or pages in the following sequence: title page, abstract, text, figure legends, footnotes, ta- bles with legends, figures. Page charges and reprints. — The current charge per journal page will be assessed as follows: Manu- scripts in the appropriate word processing program $ 10/page; others $35/page. Reprints are available only from the printer and should be ordered when the author receives the galley proof pages. Waiver of page charg- es: In exceptional cases, the Editor will consider re- quests for a waiver of page charges for manuscripts which have been accepted for publication. Requests should be made in writing to the Editor. The author must certify that neither grant nor institutional funds are available to cover all of the publication costs. The Society will normally waive no more than 80% of the page charges. RESEARCH NOTES Instructions above pertaining to feature articles ap- ply also to research notes, except that abstracts and most headings are not used and the author’s name and address follow the Literature Cited section. CONTENTS THE JOURNAL OF ARACHNOLOGY VOLUME 23 Feature Articles NUMBER 1 Philoponella republicana (Araneae, Uloboridae) as a Commensal in the Webs of Other Spiders Ann L. Rypstra and Greta J. Binford 1 New Species and Records of the Ground Spider Family Gallieniellidae (Ara- neae, Gnaphosoidea) from Madagascar Norman I. Platnick 9 On the Spider Genus Hebrithele (Araneae, Miturgidae) Norman I. Platnick and Alexandre B. Bonaldo 13 Discriminacion por Metepeira seditiosa (Keyserling) (Araneae, Araneidae) en Condiciones Experimentales Sobre dos Presas Frecuentes en el Me- dio Carmen Viera 17 The Web and Building Behavior of Synotaxus ecuadorensis (Araneae, Syn- otaxidae) William G. Eberhard 25 Los Nervios Opticos en Cuatro Especies de Lactrodectus (Araneae, Theri- diidae) Carmen J. de la Serna de Esteban y C. Monica Spinelli .... 31 Direct Evidence for Trade-offs Between Foraging and Growth in a Juvenile Spider Linden E. Higgins 37 Research Notes Nests of Hibana gracilis are Reused by Phidippus clarus in Wetlands of Northeastern Kansas Stephen R. Johnson 44 Flexibility in Foraging Tactics of Buthus occitanus Scorpions as a Response to Above-ground Activity of Termites Orit Skutelsky 46 Book Reviews Spiders in Ecological Webs (by D. H. Wise) Gail E. Stratton 48 Advances in Spider Taxonomy 1988-1991 with Synonymies and Transfers 1940-1980 (by N. I. Platnick) Robert Bennett 51 & L 1 m The Journal of VOLUME 23 1995 NUMBER 2 THE JOURNAL OF ARACHNOLOGY EDITOR: James W. Berry, Butler University ASSOCIATE EDITOR: Petra Sierwald, Field Museum EDITORIAL BOARD: A. Cady, Miami (Ohio) Univ. at Middletown; J. E. Carrel, Univ. Missouri; J. A. Coddington, National Mus. Natural Hist.; J. C. Cokendolpher, Lubbock, Texas; F. A. Coyle, Western Carolina Univ.; C. D. Dondale, Agriculture Canada; W. G. Eberhard, Univ. Costa Rica; M. E. Galia- no, Mus. Argentino de Ciencias Naturales; M. H. Greenstone, BCIRL, Columbia, Missouri; C. Griswold, Calif. Acad. Sci.; N. V. Horner, Midwestern State Univ.; D. T. Jennings, Garland, Maine; V. F. Lee, California Acad. Sci.; H. W. Levi, Harvard Univ.; E. A. Maury, Mus. Argentino de Ciencias Naturales; N. I. Plat- nick, American Mus. Natural Hist.; G. A. Polis, Vanderbilt Unfv.; S. E. Riechert, Univ. Tennessee; A. L. Rypstra, Miami Univ., Ohio; M. H. Robinson, U.S. National Zool. Park; W. A. Shear, Hampden-Sydney Coll.; G. W. Uetz, Univ. Cincinnati; C. E. Valerio, Univ. Costa Rica. The Journal of Arachnology (ISSN 0160-8202), a publication devoted to the study of Arachnida, is published three times each year by The American Arach- nological Society. Memberships (yearly): Membership is open to all those in- terested in Arachnida. Subscriptions to The Journal of Arachnology and American Arachnology (the newsletter), and annual meeting notices, are included with mem- bership in the Society. Regular, $30; Students, $20; Institutional, $80 (USA) or $90 (all other countries). Inquiries should be directed to the Membership Secretary (see below). Back Issues: Patricia Miller, P.O. Box 5354, Northwest Mississippi Community 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: Matthew H. Greenstone (1995-1997), Plant Science & Water Conservation Laboratory, USDA; Stillwater, Oklahoma 74075 USA. PRESIDENT-ELECT: Ann L. Rypstra (1995-1997), Dept, of Zoology, Miami University, Hamilton, Ohio 45011 USA. MEMBERSHIP SECRETARY: Norman I. Platnick (appointed), American Museum of Natural History, Central Park West at 79th St., New York, New York 10024 USA. TREASURER: Gail E. Stratton (1993-1995), Department of Biology, Rhoades College, Memphis, Tennessee 38112-1690 USA. BUSINESS MANAGER: Robert Suter, Dept, of Biology, Vassar College, Pough- keepsie, New York 12601 USA. SECRETARY: Alan Cady (1993-1995), Dept, of Zoology, Miami Univ., Mid- dleton, Ohio 45042 USA. ARCHIVIST: Vincent D. Roth, Box 136, Portal, Arizona 85632 USA. DIRECTORS: James Carico (1995-1997), Pat Miller (1993-1996), Robert Su- ter (1995-1997). HONORARY MEMBERS: C. D. Dondale, W. J. Gertsch, H. W. Levi, A. F. Millidge, W. Whitcomb. Cover illustration: Scanning electron microscope photograph of the abdomen of a mature female Philoponella vicina (Uloboridae). Photograph by Flory Pereira and William G. Eberhard. Publication date: 14 November 1995 ® This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 1995. The Journal of Arachnology 23:55-59 DESCRIPTION OF THE SPIDER MASONCUS POGONOPHILUS (ARANEAE, LINYPHIIDAE), A HARVESTER ANT MYRMECOPHILE Paula E. Cushing: Department of Zoology, University of Florida, Gainesville, Florida 32611 USA ABSTRACT. One species of the genus Masoncus (Araneae, Linyphiidae) is described and illustrated. Masoncus pogonophilus new species has been collected exclusively inside the nests of the Florida harvester ant, Pogono- myrmex badius (Latreille) (Hymenoptera, Formicidae) and is, therefore, considered a myrmecophile, or obligate ant associate. Morphological characters separating this new species from two of the three described congeners (M. arienus and M. conspectus ) are noted. Three species are included in the genus Ma- soncus Chamberlin 1948: M. arienus Chamber- lin 1948, M. dux Chamberlin 1948, and M. con- spectus (Gertsch & Davis 1936) (synonymized with M. nogales Chamberlin 1948 by Ivie 1967). The female holotype of M. dux has been lost and I was unable to locate any specimens of this spe- cies. The female holotype, male allotype and paratypes of M. nogales designated by Cham- berlin (1948) have also been lost. However, the holotype of Tapinocyba conspecta is housed at the American Museum of Natural History in New York City, New York (AMNH) as are other rep- resentatives of this species. The holotype and paratypes of M. arienus designated by Cham- berlin (1948) are also at AMNH. One male rep- resentative of M. arienus is housed at the Cali- fornia Academy of Sciences in San Francisco, California (CAS). No information was recorded either in the original species descriptions or on the collecting labels of the existing specimens regarding the nat- ural history of the described species. M. dux was described from a single female collected in north- ern Manitoba, Canada. All specimens of M. ar- ienus were collected in Arizona. M. conspectus was described from the male holotype and two male paratypes collected in Texas. Other records of this species include Arizona and Florida (the latter collected by the shores of Newnan’s Lake in Alachua County). Masoncus pogonophilus new species was orig- inally collected by Sanford Porter from the nests of the Florida harvester ant, Pogonomyrmex badius (Latreille) (Hymenoptera, Formicidae) (Porter 1985). It is included in the genus Ma- soncus due to the presence of distinct cephalic pits and a straight, distally bifid embolic division in the males (see genus description below). In the species description that follows, I use primarily carapace, genitalic, chaetotaxic, nu- meric, and palpal characters deemed most useful by Millidge (1980) for erigonine spiders. These characters include: 1) the overall conformation of the male palpal organ, 2) the shape of the embolic division, 3) the external appearance of the epigynum, 4) the number of dorsal tricho- bothria present on the palpal tibia of both sexes, 5) the number of dorsal tibial spines present (ex- pressed by the formula a:b:c:d), 6) the number of dorsal metatarsal trichobothria present (ex- pressed by the formula I:II:III:IV), 7) the relative position of the dorsal metatarsal trichobothrium on leg I (expressed by the formula Tml = dis- tance from tibia-metatarsus joint to trichoboth- rium/distance from tibia-metatarsus joint to metatarsus-tarsus joint), and 8) the relative stoutness of tibia I (expressed by the formula Tibi = length of tibia/ width of tibia viewed lat- erally). Overall body size, body color, and num- ber of setae on the carapace are also given. Cer- tain of these characters as well as others used in Chamberlin’s (1948) descriptions or obvious on the existing specimens are of particular value in separating M. arienus, M. conspectus , and M. pogonophilus (Table 1). All measurements were taken directly from the specimens using an ocular micrometer in a dissecting microscope. Mea- surements were rounded to the nearest 0.1 mm. Masoncus Chamberlin 1948 The type species of the genus is M. arienus. The genus Masoncus is characterized by both cephalic pits in the males and a straight, distally 55 56 THE JOURNAL OF ARACHNOLOGY Table 1.— Morphological characters most useful in separating three of four Masoncus species. (All specimens of M. dux are lost, and the species description is based solely on the female holotype.) pme = posterior median eyes. Characters M. arienus M. conspectus M. pogonophilus Location of cephalic pits Pit opens back of posted- Pit opens and extends be- Pit opens and extends be- in males or eyes; not extending under pme neath pme neath pme Cheliceral spurs towards Present on males and fe- Present on males; reduced No cheliceral spurs on distal end males to very small black spurs on females males or females Setigerous nodule or spur Nodule present on males Spur on males; lacking on No setigerous nodule or anterior to fang groove and females females spur on males or fe- males Endites with small spur Present on males and fe- Present on males and fe- No spurs on endites of on ectal side of tip males males although less dis- tinct on latter males or females Shape of palpal tibia Widely spaced black- tipped processes on dis- tal edge flush w / surface of tibia; long setal fringe on lateral edge Closely spaced black- tipped processes on dis- tal edge extending slightly away from sur- face of tibia; long setal fringe on lateral edge (fig. 97 in Chamberlin 1948) Moderately spaced black- tipped processes on dis- tal edge flush w/ surface of tibia; long setal fringe on lateral edge (see fig. 4) Embolic division Bifurcation begins close to tail-piece; each segment of bifurcation coiled (fig. 101 in Chamberlin 1948) Distally bifid w/ proximal part of bifurcation bent forward and extending over most distal part which is, itself, squared off (fig. 98 in Chamber- lin 1948 shows it point- ed) Distally bifid w/ proximal part of bifurcation bent forward and extending over most distal part which is, itself, bifur- cated (see Fig. 3) bifid embolic division (Chamberlin 1948) (dia- gram of linyphiid palpal structures in Millidge 1980). Masoncus pogonophilus new species (Figs. 1-5) Type. —The male holotype was collected 23 cm below ground inside a nest chamber of the Florida harvester ant, Pogonomyrmex badius in Archer Sandhills, 1 .4 km west of the Levy Coun- ty line off of State Road 24. The female allotype was collected from the same P. badius nest. She was found in a nest chamber 46.5 cm below ground. Both were collected on 25 September 1994 and both will be deposited in the arach- nological collection at CAS. The holotype, 1 1 male paratypes, the allotype, and 1 2 female paratypes were used in this species description. The collecting information as well as the future museum destination for these para- types are presented in Table 2. Etymology.— The specific epithet is derived from the generic name of the host ant with which the spider is found. Holotype.— Total body length: 1.7 mm. Car- apace length: 0.9 mm. Carapace width: 0.7 mm. Colors: carapace orange; abdomen grey; legs or- ange; sternum orange. Number of setae along midline of carapace: three. Palp as in Fig. 3. Em- bolic division as in Fig. 4. Number of tricho- bothria on palpal tibia: two (Fig. 2). Number of dorsal tibial spines: 1:1: 1:1. Number of dorsal metatarsal trichobothria: 1:1: 1:0. Tml: 0.82. Tibi: 7.0. Males (general).— {n — 12). Total body length: 1. 6-2.1 mm (x = 1.8 ± 0.14). Carapace length: 0. 8=0.9 mm (x = 0.9 ± 0.04). Carapace width: CUSHING - DESCRIPTION OF MASONCUS POGONOPHILUS (LINYPHIIDAE) 57 Figures 1 -5 . - Masoncus pogonophilus new species. 1, male carapace, dorsal view (scale = 0.4 mm); 2, tibia and patella of left male palpus, dorsal view, trichobothria in circular pits (scale = 0.2 mm); 3, male palpus, prolateral view (bifurcation of embolic division just visible distally) (scale = 0.2 mm); 4, embolic division of left male palpus, meso ventral view (scale = 0.1 mm); 5, epigynum, ventral view (scale = 0.1 mm). 0.6-0. 8 mm (x - 0.7 ± 0.05). Colors: carapace yellow-orange to orange; abdomen grey; legs yel- low-orange to orange; sternum yellow-orange to orange. The color seems to fade severely when specimens are kept in isopropanol rather than ethanol. Number of setae along midline of car- apace (Fig. 1): variable, 2-4 (setae easily broken in preservation). Palp as in Fig. 3. Embolic di- vision as in Fig. 4. Number of trichobothria on palpal tibia: generally two (Fig. 2), however one male had two on the left palpal tibia and three on the right and another had three on the left and two on the right. Number of dorsal tibial spines: 1 : 1 : 1 : 1 . Number of dorsal metatarsal tri- chobothria: 1:1: 1:0. Tml: 0.82-0.88 (x = 0.84 ± 0.02). Tibi: 6.5-7.? (x = 7.0 ± 0.35). Females.— (n = 13). Total body length: 1.5- 1.9 mm (x = 1.8 ± 0.13). Carapace length: 0.8- 1.2 mm (x = 0.9 ± 0.1 1). Carapace width: 0.6- 0.9 mm (x = 0.7 ± 0.09). Colors: same as males. Number of setae along midline of carapace: vari- able, 2-5; females also had smaller setae scat- tered on either side of midline. Epigynum as in Fig. 5. Number of trichobothria on palpal tibia: generally three, however one female had two on both palps, three other females had three tricho- bothria on the left palpal tibia and two on the right. Number of dorsal metatarsal trichoboth- 58 THE JOURNAL OF ARACHNOLOGY Table 2. —Collection information and museum destination for the 23 paratypes. All were collected from the nests of the Florida harvester ant, P. badius. MCZ = Museum of Comparative Zoology, Cambridge, Massa- chusetts; DPI = Division of Plant Industry, Gainesville, Florida; AMNH = American Museum of Natural History, New York, New York; CAS = California Academy of Sciences, San Francisco, California. Collection date Florida county Collector Number of spec- imens Museum Males 10- XI- 1982 Leon S. D. Porter 1 MCZ 9- XI- 1984 Leon S. D. Porter 1 MCZ 13- 1-1990 Walton Skelley, Tumbow & Thomas 2 DPI 14- 1-1990 Okaloosa Skelley, Tumbow & Thomas 1 DPI 9- ¥-1992 Leon P. E. Cushing 1 DPI 26- ¥-1992 Levy P. E. Cushing 2 AMNH 8- X-1992 Levy P. E. Cushing 1 AMNH 20- 11-1993 Levy P. E. Cushing 1 CAS 25- IX- 1994 Levy P. E Cushing 1 CAS Females 3- 1-1990 Okaloosa P. Skelley 1 DPI 13- 1-1990 Walton Skelley, Tumbow & Thomas 1 DPI 9- ¥-1992 Leon P. E. Cushing 1 DPI 28- 11-1993 Putnam P. E. Cushing 4 AMNH 25- VII..j.993 Putnam P. E. Cushing 1 MCZ 25- IX- 1994 Levy P. E. Cushing 4 CAS ria: 1:1:1:0. Tml: 0.58-0.87 (x = 0.81 ± 0.09). Tibi: 6. 5-7.9 (X = 7.1 ± 0.38). Diagnosis.— The carapace of male M. pogon- ophiius most resembles that of M. conspectus (fig. 93 in Chamberlin 1948 and Fig. 1). In both spe- cies, the cephalic pits extend beneath the pos- terior median eyes (pme) whereas in M. arienus the cephalic pits open behind the pme. The em- bolic division of male M. pogonophilus new spe- cies most resembles M. conspectus (fig. 98 in Chamberlin 1948 and Fig. 4) in that both are Figure 6.— Sticky silk from adult male Masoncus pogonophilus new species web. Magnification 400 x . distally bifid with the proximal part of the bi- furcation bent forward and extending over the most distal part of the bifurcation. However, in M. pogonophilus the most distal part of the bi- furcation is, itself, bifurcated, whereas in M. con- spectus it is flattened (although fig. 98 in Cham- berlin 1 948 shows it to be pointed). In M. arienus the embolic division is also bifid, but the bifur- cation begins very close to the tailpiece and each segment of the bifurcation is coiled (see fig. 101 in Chamberlin 1948). The male palpal tibia of the new species, as with M. conspectus and M. arienus, is fringed laterally with long setae (Fig. 2). Chamberlin 1948 (fig. 102) does not show this fringe of setae on his drawing of M. arienus but it is evident on the preserved specimens. All three species have two black-tipped processes on the distal edge of the palpal tibia (Fig. 2). These processes are more widely spaced in M. arienus than in either M. conspectus or in M. pogono- philus. The black-tipped process in M. conspec- tus is found on a slight ridge that extends away from the surface of the tibia (fig. 97 in Cham- berlin 1948). Interestingly, M. conspectus is the only one of the three previously described con- geners whose known distribution extends into northern Florida. The new species can be sepa- CUSHING - DESCRIPTION OF MASONCUS POGONOPHILUS (LINYPHIIDAE) 59 rated from the congeners based primarily upon characters described in Table 1 as well as upon overall size; the new species being somewhat smaller than M. dux, M. arienus and M. con- spectus, which are all between 2.10-2.65 mm in length according to Chamberlin (1948) and Gertsch & Davis (1936). Natural History. — Masoncus pogonophilus new species lives within the nest chambers of the Florida harvester ant, P. badius. It is about lA the size of its 7-9 mm long host and feeds on collembolans found throughout the 1-3 m deep subterranean nests (Porter 1985). The ant nest provides a stable microclimate as well as an abundant food source for the spider. The spiders have never been collected outside the ant nests and are extremely susceptible to desiccation when removed from the nests. They appear, therefore, to be obligate ant symbionts, or myrmecophiles. Immigration to new nest sites is common in P. badius (Gentry & Stiritz 1972; Golley & Gen- try 1964; Gordon 1992). While observing three such colony migrations, each occurring either just after a summer shower or in the early morning when the surface temperature was cool and the humidity high, I saw spiders and collembolans moving from the old colony site to the new amidst their host ants within the emigration trails. Using a PCR (polymerase chain reaction)-based mo- lecular technique, I have also found evidence that spiders disperse between neighboring ant nests (pers. obs.). Both sexes of M. pogonophilus build prey cap- ture webs in the lab, and I have seen webs inside the ant nest chambers. Both males and females produce sticky silk (Fig. 6). Therefore, males pre- sumably retain the aggregate and flagelliform glands into adulthood; most adult male spiders lose these glands during the terminal molt and cannot subsequently produce sticky silk (Kovoor 1987). Maintaining the ability to produce sticky silk as adults may be common among male er- igonine Linyphiids as I have observed such be- haviors among other (unidentified) male erigon- ines. Female M. pogonophilus lay 1-6 eggs in a disk- shaped eggsac deposited in a depression in the wall of a nest chamber (n = 9 eggsacs, x = 2.9 ± 1.5 eggs/eggsac). The eggsac is flush against the surface of the chamber walls. Juvenile spiders molt once inside the eggsac and pass through three additional molts before reaching maturity. Juveniles are present inside the ant nests during all months of the year (Porter 1985; pers. obs.). Porter reported a 4:1 female-biased sex ratio among the spiders, while I have found an even more extreme 7.5:1 female-biased ratio. Due to the scarcity of eggsacs and the small number of eggs per eggsac, it has not been possible to de- termine whether this is a primary sex ratio bias. ACKNOWLEDGMENTS I am grateful to Mark Stowe, Susan Moegen- burg, John Arnett, Joyce Thomas, Barbara C. Dixon, Michael E. Morgan, and Rebecca Forsyth for their help in collecting specimens. Thanks also to Drs. Jonathan Reiskind, Sanford Porter, Charles Dondale, and Charles Griswold for their helpful comments on drafts of this manuscript. Thanks also to CAS, AMNH, DPI and MCZ for loaning specimens. This work was supported by a Grinter Fellowship from the University of Florida and a Theodore Roosevelt Grant from the American Museum of Natural History. LITERATURE CITED Chamberlin, R. V. 1948. On some American spiders of the family Erigonidae. Ann. Entomol. Soc. Amer- ica, 41:483-562. Gentry, J. B. & K. L. Stiritz. 1972. The role of the Florida harvester ant, Pogonomyrmex badius in old field mineral nutrient relationships. Environ. En- tomol., 1:39-41. Gertsch, W. J. & L. I. Davis. 1936. New spiders from Texas. American Mus. Novit., 881:1-21. Golley, F. B. & J. B. Gentry. 1964. Bioenergetics of the southern harvester ant, Pogonomyrmex badius. Ecology, 45:217-225. Gordon, D. M. 1992. Nest relocation in harvester ants. Ann. Entomol. Soc. America, 85:44-47. Ivie, W. 1967. Some synonyms in American spiders. J. New York Entomol. Soc., 75:126-131. Kovoor, J. 1987. Comparative structure and histo- chemistry of silk-producing organs in arachnids. Pp. 1 60-186, In Ecophysiology of Spiders. (W. Nentwig, ed.). Springer- Verlag, Berlin. Millidge, A. F. 1980. The erigonine spiders of North America. Part 1 . Introduction and taxonomic back- ground (Araneae: Linyphiidae). J. Arachnol., 8:97- 107. Porter, S. D. 1985. Masoncus spider: a miniature predator of Collembola in harvester-ant colonies. Psyche, 92:145-150. Manuscript received 4 December 1994, revised 6 March 1995. 1995. The Journal of Arachnology 23:60-64 A TEST OF THE CENTRAL-MARGINAL MODEL USING SAND SCORPION POPULATIONS {PARUROCTONUS MESAENSIS, VAEJOVIDAE) Tsunemi Yamashita1 and Gary A. Polis: Department of Biology, Box 1812 Station B, Vanderbilt University, Nashville, Tennessee 37235 USA ABSTRACT. The central-marginal model proposes marginal populations contain a lower density of individ- uals, lower levels of genetic variation, and are more isolated than populations in the center of a species range. Previous tests of the model used Drosophila, organisms capable of extended dispersal. We test the central- marginal model with scorpions, organisms with restricted dispersal abilities. We measured genetic variation through allozyme analysis of eight loci (five polymorphic, three monomorphic) to obtain estimates of hetero- zygosity. We compared differences between the two types of populations with a split-plot ANOVA. We also compared central and marginal populations using standard parametric tests. We found marginal populations contain lower genetic variation than central populations. These populations may be important as models in conservation to study the effects of fragmentation. Peripheral or marginal populations are those on the boundaries of a species’ geographic range. They exhibit unique properties not evident in populations in the center (Brussard 1984). Gen- erally, as one moves outward from the center of a species’ range, populations are hypothesized to become less dense, more isolated, and less vari- able genetically within populations (da Cunha et al. 1959; da Cunha & Dobzhansky 1954; Carson 1959; Soule 1973; Brussard 1984). These trends are embodied in the central-marginal model (Le- wontin 1974). Several explanations exist. One hypothesis (da Cunha & Dobzhansky 1954) states genetic polymorphism is positively correlated with the number of niches an organism occupies — more niches are available at the center of a species range. Yet Brussard’s (1984) free recom- bination hypothesis proposes linkage disequilib- rium is favored in central populations because extreme phenotypes are selected against, i. e., stabilizing selection occurs. Here, favored genes are linked together to create a stable phenotype with maximal fitness. In theory, at the range mar- gins, linkage equilibrium is selected in stressful environments to create novel phenotypes better able to survive sub-optimal conditions (Brussard 1984). Research with Drosophila suggests the predic- 1 Present address: Department of Biology, Northeast Louisiana University, Monroe, Louisiana 71209 USA tion that allozyme heterozygosity is reduced at range margins is not valid (Brussard 1984). How- ever, Drosophila probably disperse relatively great distances compared to most non-flying taxa. Non- flying taxa (vertebrates) show a decline in allelic diversity in marginal populations (Soule’ 1973). We propose scorpions, unable to disperse great distances, also may illustrate the predictions of Lewontin’s (1974) central-marginal model. METHODS Relevant scorpion biology. —The scorpion Pa- ruroctonus mesaensis Stahnke is restricted to sand dunes and sandy substrates scattered throughout the southwestern United States and northern Mexico. It occurs in dense populations (range = 1600-5000/ha) (Polis & Yamashita 1991) and it is an ecologically important species, as a gener- alist predator in desert food webs (Polis 1979; Polis & McCormick 1986, 1987). This scorpion is dispersal limited because specialized morpho- logical features adapt and restrict it to sand. The species possess numerous modified setae on their tarsi (sand shoes) to facilitate sand movement and burrow construction (Polis et al. 1986). Fur- ther, their ability to detect substrate vibrations to localize prey only functions well on sand (Brownell & Farley 1979). These specializations reduce the likelihood of extended migration. However, reproductive males move extensively when searching for mates (Polis & Farley 1979, 1980). 60 YAMASHITA & POLIS-SAND SCORPION CENTRAL-MARGINAL POPULATIONS 61 Collection sites: Central Populations: A. Glamis B. Yuma Desert C. Yuma Army D. Ogilby E. N Fontana F. Glamis East Marginal Populations: G. Windy Point H. Kelso I. Needles J. Bouse K. Phoenix L. Agua Caliente 0 40 Kilometers Figure 1 .— A partial map of the desert areas of California, Arizona, and Baja California Norte where populations of Paruroctonus mesaensis were collected. A proposed range of P. mesaensis is indicated by the dashed line. Collection and electrophoresis procedures.— We collected the scorpions over three years, 1 989— 1991. Paruroctonus mesaensis is easy to collect as it (and all scorpions) fluoresces when illumi- nated with a ultraviolet light (Sylvania F8T5/ BLB). We transported the scorpions alive to the laboratory, froze them with liquid nitrogen, and stored them at - 70 °C until electrophoretic anal- ysis. We electrophoresed five polymorphic and three monomorphic loci from 28 populations. This re- search was the first allozyme analysis of any scor- pion populations, and these eight loci were the only resolvable loci from a screening of 25 en- zyme loci on 1 1 different buffer systems. We as- sayed a mean of 29 individuals (± 5.7) from each population for each locus. The specific protocols and other pertinent methodology are described in Yamashita (1993). Central and marginal population determina- tion.—We determined the geographical center of P. mesaensis populations from collection data (Haradon 1983; D. Gaffin pers. comm.; Yama- shita 1 993) (Fig. 1 ). We determined the following range extremes: in the north (Death Valley, Cal- ifornia); in the south, Cabo Lobos (Sonora, Mex- ico); in the east (Phoenix, Arizona); and in the west (Windy Point in the Coachella Valley, Cal- ifornia). We determined the range center to be located 65 km north of Yuma, Arizona by lo- cating the midpoint between the extreme north- ern and southern and the western and eastern populations. We designated six populations within a 32 km radius of the range center as 62 THE JOURNAL OF ARACHNOLOGY Central populations: Glamis, Yuma desert, Yuma Army, Ogilby, N. Fontana, and Glamis East. We delineated peripheral populations as those near- est to edges of the range. These include Windy Point, Kelso, Needles, Bouse, Phoenix, and Agua Caliente. Statistical analyses. —We performed two types of analyses to determine if differences between central and marginal populations exist. The first, a method outlined by Weir (1990), tested if het- erozygosity differences exist between population types. This design, similar to a split-plot ANO- VA, considered variation from five sources: pop- ulations, individuals within populations, loci, loci by populations, and loci by individuals within populations. Heterozygotes are entered as 1’s, homozygotes, as 0’s and the data from each cen- tral or marginal populations were pooled. We used electrophoretic data from 20 individuals from each of six central and six marginal pop- ulations ( 1 20 individuals in each population type) and five loci. We used Weir’s analysis because estimates of heterozygosity often exhibit large interlocus variances and non-normal distribu- tions; therefore, many standard parametric tests may be inappropriate (Archie 1985). Second, we performed standard parametric statistics to determine differences between cen- tral and marginal populations. The variables ex- amined were observed average heterozygosity, mean allele number, and percent polymorphism of each population. We calculated these variables using BIOSYS-1 (Swofford & Selander 1989). Observed heterozygosity per locus is the fraction of heterozygous individuals from a given sample for a particular locus (Ferguson 1 980; Weir 1 990). Observed average heterozygosity is the mean value from all loci. Although heterozygosity val- ues commonly undergo an arcsine transforma- tion, our data did not require such a procedure because most heterozygosity values fell between 0.30 and 0.70, a range that does not require trans- formation (Sokal & Rohlf 1981). We used two other indices of genetic vari- ability. Percent polymorphism is the mean num- ber of polymorphic loci in a population. Here, a locus is polymorphic if the frequency of the most common allele is 0.95 or less. The mean number of alleles per locus is the number of alleles at each locus averaged across all loci. RESULTS The split-plot ANOVA and standard para- metric tests showed the mean heterozygosities Table 1.— Split plot ANOVA analysis of two pop- ulation categories (Central and Marginal). This design is taken after Weir 1990. df MS F-value P-value Category Individuals within 1 3.54 21.60 <0.001 categories 222 0.166 1.01 >0.437 Loci 4 4.29 26.17 <0.001 Category x loci 4 1.29 7.86 <0.001 Error 888 0.164 from the central and marginal populations were significantly different (Tables 1, 2). The results of the split-plot ANOVA (Table 1) show values from all levels of analysis (populations; individ- uals within populations; loci; population x loci) were significant ( P < 0.001) except individuals within populations ( P > 0.437). These results establish that central populations are signifi- cantly different from marginal ones in hetero- zygosity. Furthermore, the significant among loci effects suggests that each locus expressed a dif- ferent pattern of heterozygosity from other loci. Loci within marginal populations were signifi- cantly different from loci within central popu- lations, which suggests within each population type (central or marginal), the same locus ex- pressed significantly different heterozygosities. The mean genetic variability (observed het- erozygosity) in marginal populations (0.106 ± 0.025, n = 6) was significantly less than central populations (0.164 ± 0.018, n = 6; t = 3.80, 0.05 > P > 0.01) supporting the central-marginal model. The mean allele number for the central populations (1.73 ± 0.1 44) was marginally great- er (t = 2.21, 0.1 > P > 0.05) than that of mar- ginal populations (1.52 ± 0.095). The mean per- cent polymorphism for central populations (45.83 ± 6.45) was also marginally greater (35.42 ± 9.41, t = 2.71, 0.05 > P> 0.01). DISCUSSION The significant differences between central and marginal populations for allozyme heterozygos- ity, mean allele number, and percent polymor- phism are consistent with the central-marginal model (Brussard 1984). Paruroctonus mesaensis is one of the species that fits the predictions of this model; tests of the model using Drosophila allozymes failed to exhibit similar patterns (Brussard 1984). In our study, the decrease in genetic variability in marginal populations prob- YAMASHITA & POLIS-SAND SCORPION CENTRAL-MARGINAL POPULATIONS 63 Table 2.— A comparison of genetic variability between central and marginal populations. See text for discus- sion. Mean heterozygosity Mean allele number % Polymorphism Central populations Glamis 0.153 1.75 37.5 Yuma Desert 0.176 2.00 50.0 Yuma Army 0.136 1.63 50.0 Ogilby 0.167 1.63 37.5 N. Fontana 0.166 1.63 50.0 Glamis East 0.187 1.75 50.0 Mean 0.164 1.73 45.8 SD 0.018 0.144 6.5 Marginal populations Windy Point 0.103 1.50 25.0 Phoenix 0.078 1.38 25.0 Bouse 0.131 1.63 50.0 Agua Caliente 0.140 1.50 37.5 Kelso 0.084 1.63 37.5 Needles 0.097 1.50 37.5 Mean 0.106 1.52 35.4 SD 0.025 0.095 9.4 Central vs marginal ^-statistic 3.80 2.21 2.71 P values 0.05 > P > 0.01 0.1 > P> 0.05 0.05 > P > 0.01 ably stems from reduced gene flow or smaller overall population size. Because scorpion dis- persal is local, populations at the range margin are less likely to receive migrants from other pop- ulations compared to more central populations. Central populations exhibit the highest allele number and percent polymorphism. These pop- ulations may maintain higher genetic variability because exchange with other nearby populations is more frequent and population size is generally larger in the center of the range. However, mod- els suggest that a very small effective population size (ne < 10 individuals) is required to reduce significantly the number of alleles per locus with- in a population (Nei et al. 1975; Rice & Mack 1991). Some marginal populations (Needles, Bouse, and Phoenix) are geographically isolated from other populations. Needles, north of a mountain range present on either side of the Colorado Riv- er, is effectively isolated. The Bouse population exists on the eastern edge of the Cactus Plain, a large sandy region in western Arizona. It is sur- rounded by rocky habitat and isolated from the nearest population by 40 km. Although Bouse is not separated by a large distance, the interme- diate rocky substrate effectively curtails dispers- al. No P. mesaensis were observed in > 30 hours of searching on rocky habitats adjacent to sandy areas (Polis, unpubl. data). The Phoenix population is the most eastern and one of the most genetically depauperate pop- ulations. Emigration into this area probably oc- curred along the dry river beds of the Salt and Gila rivers. This population relies on unidirec- tional gene flow since the substrate outside the river bed is a dispersal barrier to the psammo- philic scorpion and no populations exist to the east. The low values of genetic variability may be a result of two primary factors: it is a periph- eral population with a low population size and receives little gene flow from other populations. Peripheral isolate formation may have been enhanced by large scale floods in the Holocene (Ely et al. 1993). The Salt and Gila rivers ex- perienced large-scale floods in the last 5600 years (Ely et al. 1993). Periods of minor and major floods were interspersed. These extreme floods may have created expansion corridors for scor- pion movement and isolated populations by fragmenting previous habitat. Although several studies have compared cen- tral and marginal populations, a clear trend is not evident (Hoffman & Parsons 1991). Some 64 THE JOURNAL OF ARACHNOLOGY report no decrease in genetic variability within marginal populations, e. g., Drosophila alio- zymes (Brussard 1984) and an annual grass (Rice & Mack 1991). Analysis of dispersal limited an- imals (frogs) report a decrease in genetic vari- ability among marginal populations (Sjogren 1991). Further research into the properties of marginal populations is warranted because mar- ginal populations, with their insular or penin- sular properties, are similar to populations frag- mented through man’s encroachment upon the environment. ACKNOWLEDGMENTS We are grateful to M. Hedin and M. Rosen for field collection. We are also grateful to C. Bas kauf, D. McCauley, and W. Eickmeier for dis- cussion and comments. Financial support was provided by Vanderbilt University graduate travel and research awards. LITERATURE CITED Archie, J.W. 1985. Statistical analysis of heterozy- gosity data: Independent sample comparisons. Evo- lution, 39:623-637. Brownell, P. H. & R. D. Farley. 1979. Prey-localizing behaviour of the nocturnal desert sand scorpion, Paruroctonus mesaensis : Orientation to substrate vibrations. Anim. Behav., 27:185-193. Brussard, P. F. 1984. Geographic patterns and en- vironmental gradients: The central-marginal model in Drosophila revisited. Ann. Rev. Ecol. Syst., 15: 25-64. Carson, H L 1959. Genetic conditions that promote or retard the formation of species. Cold Spring Har- bor Symp. Quant. Biol, 24:87-103. da Cunha, A. B., & T. Dobzhansky, 1954. A further study of chromosomal polymorphism in Drosophila willistoni in relation to environment. Evolution, 8:119-134. da Cunha, A. B , T. Dobzhansky, O. Pavlovsky, & B. Spassky. 1959. Genetics of natural populations. XXVIII. Supplementary data on the chromosomal polymorphism in Drosophila willistoni in relation to the environment. Evolution, 13:389-404. Ely, L. L, Y. Enzel, V. R. Baker, & D. R. Cayan. 1993. A 5000-year record of extreme floods and climatic change in the southwestern United States. Science, 262:410=412. Ferguson, A. 1980. Biochemical Systematics and Evolution. Wiley & Sons, New York, Mew York. Haradon, R. M. 1983. Smeringurus, a new species of Paruroctonus Werner (Seorpionies, Vaejovidae). J. Arachnoh, 11:251-270. Hoffman, A. A. & P. A. Parsons. 1991. Evolutionary Genetics and Environmental Stress. Oxford Univ. Press, Oxford. Lewontin, R. C. 1974. The genetic basis of evolu- tionary change. Columbia Univ. Press. New York, New York. Nei, M, T. Maruyama, & R. Chakraborty. 1975. The bottleneck effect and genetic variability in popula- tions. Evolution, 29:1-10. Polis, G. A. 1979. Prey and feeding phenology of the desert sand scorpion Paruroctonus mesaensis (Scor- pionidae, Vaejovidae). J. Zool., London, 188:333— 346. Polis, G. A. & R. D. Farley. 1979. Characteristics and environmental determinants of natality, growth, and maturity in a natural population of the desert scorpion, Paruroctonus mesaensis (Scorpionidae: Vaejovidae). J. Zool., London, 187:517-542. Polis, G. A. & R. D. Farley. 1 980. Population biology of a desert sand scorpion: Survivorship, microhab- itat, and the evolution of life history strategy. Ecol- ogy, 61:620-629. Polis, G. A. & S. J. McCormick. 1986. Patterns of resource use and age structure among a guild of desert scorpions. J. Animal Ecol., 55:59-73. Polis, G. A. & S. J. McCormick. 1987. Competition and predation among species of desert scorpions. Ecology, 68:332-343. Polis, G. A., C. A. Myers & M. A. Quinlan. 1986. Burrowing biology and spatial distribution of desert sand scorpions. J. Arid Environ., 10:137-146. Polis, G. A. & T. Yamashita. 1991. The ecology and importance of predaceous arthropods in desert com- munities. Pp. 180-222. In The Ecology of Desert Communities. (G.A. Polis, ed.) Univ. of Arizona Press. Tucson, Arizona. Rice, K. J. & R. N. Mack. 1991. Ecological genetics of Bromus tectorum. I. A hierarchical analysis of phenotypic variation. Oecologia, 88:77-83. Sjogren, P. 1991. Genetic variation in relation to demography of peripheral pool frog populations (Rana lessonae). EvoL Ecol, 5:248-271. Sokal, R. R. & F. J. Rohlf. 1981. Biometry, 2nd ed. W. H. Freeman & Co., San Francisco. Soule’, M. 1973. The epistasis cycle: A theory of marginal populations. Ann. Rev. Ecol. Syst., 4:165- 187. Swofford, D. L. & R. B. Selander. 1989. BIOSYS-1: A computer program for the analysis of allelic vari- ation in population genetics and biochemical sys- tematics. User’s Manual. Illinois Nat. Hist. Survey. Weir, B. S. 1990. Genetic Data Analysis. Sinauer, Sunderland, Massachusetts. Yamashita, T. 1993. Genetic and Morphological Variation in Scorpion Populations on Habitat Is- lands. Ph. D. dissertation. Vanderbilt University, Nashville, Tennessee. Manuscript received 5 January 1995 , revised 21 March 1995 . 1995. The Journal of Arachnology 23:65-70 NATURAL HISTORY, ACTIVITY PATTERNS, AND RELOCATION RATES OF A BURROWING WOLF SPIDER: GEOLYCOSA XERA ARCHBOLDI (ARANEAE, LYCOSIDAE) Samuel D. Marshall1: Department of Zoology & Graduate Program in Ethology, University of Tennessee, Knoxville, Tennessee 37996 USA ABSTRACT. Wolf spiders in the genus Geolycosa are obligate burro wers and sit-and-wait predators which typically retain their first burrow throughout life. In the present study I document the activity patterns, burrow closure, and relocation rates of an exception to this pattern: G. xera archboldi McCrone, endemic to the scrub of central Florida. In census studies of five field plots I document mean relocation rates of up to 3.2% per day. The smaller size/age classes made up the majority of the relocating spiders. I found that individual spiders had their burrows closed on 26 ± 7% of census days. These burrow closures lasted an average of 6.8 ± 0.3 days, with 10% of burrow closures lasting longer than 14 days. Geolycosa wolf spiders are sessile members of a largely vagrant family. In spite of having evolved the burrowing habit and restricting all activity to the vicinity of the burrow mouth they have retained behavioral features associated with their vagrant ancestors. These traits include vagility and the attachment of the egg sac to the spin- nerets. Published accounts of Geolycosa indicate that these spiders generally remain with one bur- row throughout their lives and relocate seldom (Wallace 1942; McQueen 1983; Conley 1985) or never (McCrone 1963; Miller 1989; Richardson 1990). The only exception to this has been re- ported by Richardson (1990) for a population of G. wrightii in the lakeshore dunes of Michigan which relocates more than once every two weeks. Studies of Geolycosa ecology have assumed that relocation is a rare enough event that burrow abandonment can be equated with mortality (McQueen 1983; Conley 1985). However, there have been no studies explicitly testing this as- sumption. The purpose of this study is to doc- ument the rates of relocation and activity of an apparent exception to the rule of extreme site tenacity in the genus Geolycosa. I have also quan- tified burrow closure across the population. In addition, I will present data on the natural his- tory of a species of special concern in Florida, a state undergoing vertiginous rates of develop- ment (Edwards 1994). •Present Address: Department of Zoology, Miami Uni- versity, Oxford, Ohio 45056 USA NATURAL HISTORY OF THE MODEL SYSTEM Study area.— The research presented here was performed at Archbold Biological Station, a pri- vate research facility 10 km south of Lake Placid in Highlands County, Florida. The communities represented at Archbold are unique to the south- ern Lake Wales ridge and belong to a class of habitats referred to collectively as ‘scrub’ (Abra- ham son et al. 1984). In the scrub G. xera is lim- ited to areas of open sand. This includes unpaved roads, the open sand of rosemary balds, and patches of open sand in the oak scrub. The nat- urally occurring areas are created by fire (which consumes the leaf litter) and wind (which moves the accumulated leaf fall). In these patches of open sand G. xera may be found in densities higher than any recorded for other Geolycosa species (Table 1). Study organism. — Geolycosa xera xera Mc- Crone and G. x. archboldi McCrone are endemic to the scrub and sandhill communities of the dry uplands of Highlands, Polk, Lake, Orange, and southern Volusia counties in central Florida (McCrone 1963). The subspecies in the present study, G. x. archboldi , is restricted to Highlands County. Geolycosa xera archboldi digs a distinctive ver- tical burrow 16.6 cm deep (± 3.2 cm standard deviation, n — 25). Because little silk is used in their construction, these burrows need to be ac- tively maintained in order to persist. A brief rain- fall is sufficient to wash burrow mouths closed. 65 66 THE JOURNAL OF ARACHNOLOGY Table 1.— Summary of Geolycosa population density estimates for adult females. Adult females are used as this is more widely reported than total population density. Number of samples indicates the number of inde- pendent areas or populations studied (modified from Richardson 1990, p. 44). Species Spiders/m2 Number of samples Locale Source G. domifex 0.03 1 Ontario McQueen 1978 G. rafaelana 0.11 ± 0.07 5 New Mexico Conley 1984 G. wrightii 0.07 1 Michigan Richardson 1990 G. missouriensis 0.01 ± 0.02 14 Oklahoma Richardson 1990 G. xera archboldi 0.49 ± 0.25 8 Florida This study I saw spiders reopen burrows within minutes of closure due to rainfall or my inadvertently step- ping on the burrow mouth. I found that burrow diameter is closely correlated with body size (P < 0.0001, r — 0.92, n = 161) as has been recorded for other Geolycosa species (McQueen 1983; Miller & Miller 1984). The foregoing discussion shows how detailed information on the size structure, distribution, and abundance of G. xera may be gathered on the basis of burrow char- acteristics alone. Geolycosa xera is active year-round. I ob- served adult females in all seasons, mature males in the fall through spring (see also McCrone 1963) and hatchlings in March through July. Clutch sizes are small (mean ± standard deviation: 24.0 ± 9.0, n = 5) compared to an average of 203 for G. domifex (McQueen 1978) and 179 for G. mis- souriensis (Richardson 1990). I saw in two suc- cessive years (1992, 1993) that the first annual hatchling dispersal from the maternal burrow oc- curred in mid-to-late March. Hatches were ob- served throughout the following spring and sum- mer months, but never with the synchrony of the first hatch of the year. I saw no obvious clima- tological correlates of the dispersal event to ex- plain this synchrony. March is during the dry season in Florida, and the weather is fairly con- stant. The long breeding season, coupled with year-round activity, the size-class distributions noted above, and observations of captives I have held in the lab for extended periods lead me to believe that G. xera matures in 1 8-24 months. This is similar to published accounts for other Geolycosa species (Wallace 1942; McCrone 1963; McQueen 1978; Miller & Miller 1987). METHODS Activity and relocation rates. — I established five 2.0 x 2.0 m unenclosed census plots in patches of open sand in scrubby flatwoods. These census plots were all at least several meters apart in distinct and separate sand patches. I censused these plots every other day from 6 March until 3 May 1991. Using dial vernier calipers, at each census I measured the burrow mouth diameter of new burrows to the nearest 0. 1 mm and marked them with a numbered surveyor’s flag. I also noted whether previously flagged burrows were open or closed. I estimated the mean length of burrow closure periods from those burrow clo- sures which were initiated and terminated within the census period. Extended burrow closure was assumed to be initiated and terminated by the spider. I calculated the percent of censuses in which the individual burrows were closed by di- viding the number of censuses the burrow was closed by the total number of censuses for those individual burrows which were both active at the end of the census period and censused at least 10 times. As these data were taken from five defined census plots I report the means and stan- dard deviations as calculated from census plot means. For the estimates of relocation rates I only considered those burrows found open after the first 14 days of the census period (this minimized the counting of reopened burrows as new, see results below). I also did not count hatchlings in order to avoid inflating relocation rates by in- cluding recruitment. Characterization of relocating individuals.— I enclosed a naturally-occurring habitat patch ap- proximately 12.5 m2 with sheet metal flashing. This site was chosen for the study of size-classes and movement as it had a larger population of G. xera (approximately 100 individuals) than the census patches used above. In order to examine the relative sizes of relocating versus resident in- dividuals, I arbitrarily divided the total number of new burrows for the period (10-18 July 1990) into four size classes and compared these data MARSHALL— GEOLYCOSA ACTIVITY PATTERNS 67 DAYS CLOSED Figure 1 .—Duration of burrow closures for a population of marked Geolycosa xera archboldi burrows censused every other day in five 4.0 m2 open census plots at Archbold Biological Station, Highlands County, Florida ( n = 250 burrow closures). census days. The proportion of new burrows found at each census was 2.0 ± 0.9% per day {n = 5, Table 2). This represents my estimate of the relocation rate. Characterization of relocating individuals.— New burrows belonged predominately to the smaller size classes (Fig. 2). Larger, and thus old- er, spiders are less likely to change burrow sites than the smaller, younger individuals. DISCUSSION Geolycosa xera exhibits unexpectedly high re- location rates for a fossorial spider. Whether the rates I measured in the spring remain as high throughout the year is unknown. However, these results indicate that Geolycosa wolf spiders may not all be as sedentary as previously thought. Given the energetic cost implicit in burrow construction and the risk of predation involved in leaving the security of a burrow, it would be predicted that these spiders would only move in extreme circumstances. Dispersal is assumed to to the four size classes of the long-term residents active on 14 July. RESULTS Activity and relocation rates. —I found that 90% of burrow closures lasted 14 days or less (Fig. 1). The duration of burrow closures was 6.8 ± 0.3 days (mean ± SD, n = 5). Individually, spider burrows were closed 26.0 ± 7.0% (n = 5) of Table 2.— Summary of relocation rates of Geolycosa xera archboldi at Archbold Biological Station in five 4.0 m2 census plots, 22 March to 19 April 1991. Site Density (spiders/m2) Mean daily relocation rate (%) 1 7.20 1.0 2 4.75 1.4 3 2.54 3.2 4 2.04 1.8 5 1.41 2.5 68 THE JOURNAL OF ARACHNOLOGY UPPER BOUND OF BURROW MOUTH DIAMETER SIZE CLASS Figure 2.— Composition of a population of Geolycosa xera archboldi at Archbold Biological Station. ‘Relo- cating’ (n = 36) are from new burrows found at daily censuses 10-18 July 1990. ‘Residents’ (n = 73) denotes burrows open on 14 July 1990 (exclusive of ‘Relocating’). be risky; and therefore, many studies of dispersal and migration have looked for an adaptive ex- planation for animal movement (Southwood 1962; Gaines & McClenaghan 1980; Johnson & Gaines 1990). There are no data on the cost of burrow construction in Geolycosa ; however, there are data which can give us an indication. Culik & McQueen (1985) studied activity patterns and metabolic rates in G. domifex and found that movement on the surface elevated metabolic rates 220%, and moving up and down the burrow el- evated rates 1780%. As burrow construction in- volves moving sand up the burrow to the surface, burrow construction should be even more costly than moving up and down the burrow unbur- dened. Janetos (1987) has documented that there is an inverse correlation between relocation rates and web cost in web spiders. While the evolu- tionary scenario for web-spiders may not extend to burrow-building spiders, a burrow is likely to represent a long-term investment in a site. The higher rates of relocation for the younger spiders I found are consistent with patterns re- ported by other workers (Conley 1985). Why younger spiders are more likely to relocate than older individuals is unknown. I do not feel that relocation to find a better microhabitat site is responsible given that all size/age classes of G. xera may be found in close proximity to each other. It may be that stochastic events (e. g., burrow collapse) or territorial interactions with neighbors may be responsible. I have no data on foraging and relocation rates for G. xera, but have observed several possible stochastic mechanisms to explain burrow aban- donment. I have seen excavations at the burrow mouth which I attributed to predation attempts which may have been the cause of subsequent MARSHALL— GEOLYCOSA ACTIVITY PATTERNS 69 burrow abandonment due to damage to the frag- ile burrow. I have twice observed that ants in- vading the burrow to pirate prey elicit a spec- tacular and immediate response: the spider bolts from the burrow and jumps up into the nearest vegetation. I have three times observed attacks by neighboring conspecific burrow holders on smaller individuals engaging in burrow mainte- nance behaviors. I have also seen burrow aban- donment correlated with encroaching leaf litter. The burrow of Geolycosa wolf spiders may represent both a prison and refuge. Given the cost of construction and maintenance, these spi- ders may have evolved a foraging strategy and life history centered on the burrow more like those other obligate fossorial spiders, the my- galomorphs, and very unlike their peripatetic confamilials. It seems probable that the obligate fossorial habit of Geolycosa evolved in abiotic- ally stringent habitats such as sand dunes. How- ever, Geolycosa can make up for this restriction to relatively barren and prey-depauperate sandy habitats by foraging for a longer period each day (vagrant lycosids in the scrub are nocturnal). The burrow permits individuals to shuttle between the thermally extreme conditions at the surface while foraging and the more moderate thermal environment of the burrow. The evolution of the fossorial habit also allowed an eresid to invade barren sand dunes in the Namib desert (Lubin & Henschel 1990). Geolycosa xera archboldi has the most restrict- ed range of any known Geolycosa wolf spider. It lives in a specific microhabitat within an endan- gered ecosystem (Edwards 1994; Marshall 1994). While the densities it achieves at suitable sites can be quite high, unless the habitat is periodi- cally burned, populations decline as patches of open sand are covered with leaf litter. Geolycosa x. archboldi cannot tolerate any leaf litter cov- ering the burrow mouth, and will abandon any burrow covered with leaf fall. Geolycosa xera is an excellent indicator species of the quality and health of patches of scrub in Florida’s belea- guered uplands habitats: it is active year-round, sensitive to bum frequency, and identifiable on the basis of burrow characteristics and locale alone. Populations persist at far smaller patches of scrub than do endemic vertebrates (e. g., Flor- ida scrub jays). ACKNOWLEDGMENTS I thank those who helped in the design and interpretation of my dissertation research: my dissertation committee, S. Riechert, C. Boake, G. Burghardt, A. Echtemacht, and D. Etnier; the members of the spider groups at the University of Tennessee and Miami University; and my col- league and wife M. Hodge. My field seasons were made all the more enjoyable and productive by the staff and regulars at Archbold Biological Sta- tion: particularly M. Deymp, B. Ferster, and W. Meshaka. Voucher specimens have been depos- ited at the American Museum of Natural History in New York. This research was funded by NICHD Training Grant (T32-HD-07303), and by grants from the Theodore Roosevelt Me- morial Fund, Sigma Xi, and Archbold Expedi- tions. LITERATURE CITED Abrahamson, W. G., A. F. Johnson, J. N. Layne, & P. A. Peroni. 1984. Vegetation of the Archbold Bi- ological Station: an example of the southern Lake Wales Ridge. Florida Scient, 47:209-250. Conley, M. R. 1984. Population regulation and prey community impact of Geolycosa rafaelana (Cham- berlin) (Araneae: Lycosidae). Dissertation, New Mexico State Univ., Las Cruces, New Mexico. Conley, M. R. 1985. Predation versus resource lim- itation in survival of adult burrowing wolf spiders (Araneae: Lycosidae). Oecologia, 67:71-75. Culik, B. M. & D. J. McQueen. 1985. Monitoring respiration and activity of the spider Geolycosa domifex (Hancock) using time-lapse television and C02 gas analysis. Canadian J. Zool. 63:843-846. Edwards, G. B. 1 994. McCrone’s burrowing wolf spi- der. Pp. 232-233, In: Rare and Endangered Biota of Florida, Vol. IV., Invertebrates (M. Deyrup & R. Franz, eds.). Univ. Florida Press, Gainesville, Flor- ida. Gaines, M. S. & L. R. McClenaghan, Jr. 1980. Dis- persal in small mammals. Ann. Rev. Ecol. Syst., 1 1 : 163-196. Janetos, A. C. 1 987. Web-site selection: are we asking the right questions? Pp 9-22, In: Spiders: Webs, behavior, and evolution. (W. A. Shear, ed.). Stan- ford Univ. Press, Stanford, California. Johnson, M. L. & M. S. Gaines. 1990. Evolution of dispersal: theoretical models and empirical tests us- ing birds and mammals. Ann. Rev. Ecol. Syst., 21: 449-480. Lubin, Y. & J. Henschel. 1990. Foraging at the ther- mal limit: burrowing spiders ( Seothyra , Eresidae) in the Namib desert dunes. Oecologia, 84:461-467. Marshall, S. D. 1994. Territorial aggregation in the burrowing wolf spider Geolycosa xera archboldi McCrone: Formation, maintenance, and conse- quences. Dissertation, Univ. Tennessee, Knoxville. McCrone, J. D. 1963. Taxonomic status and evo- lutionary history of the Geolycosa pikei complex in 70 THE JOURNAL OF ARACHNOLOGY the South-eastern United States (Araneae, Lycosi- dae). American Mid. Nat., 70:47-73. McQueen, D. J. 1978. Field studies of growth, re- production, and mortality in the burrowing wolf spider Geolycosa domifex (Hancock). Canadian J. Zool. 56:2037-2049. McQueen, D. J. 1983. Mortality patterns for a pop- ulation of burrowing wolf spiders, Geolycosa dom- ifex (Hancock), living in southern Ontario. Cana- dian J. Zool., 61:1263-1271. Miller, G. L. 1989. Subsocial organization and be- havior in broods of the obligate burrowing wolf spi- der Geolycosa turricola (Treat). Canadian J. Zool, 67:819-824. Miller, G. L. & P. R. Miller. 1984. Correlations of burrow characteristics and body size in burrowing wolf spiders (Araneae: Lycosidae). Florida Ent, 67: 314-317. Miller, G. L. & P. R. Miller. 1987. Life cycle and courtship behavior of the burrowing wolf spider Geolycosa turricola (Treat) (Araneae, Lycosidae). J. ArachnoL, 15:385-394. Richardson, R. K. 1990. Life history, soil associa- tions, and contests for burrows in a burrowing wolf spider, Geolycosa missouriensis Banks. Disserta- tion, Univ. Oklahoma, Norman, Oklahoma. Southwood, T. R. E. 1962. Migration of terrestrial arthropods in relation to habitat. Biol. Rev., 37: 171-214. Wallace, H. K. 1942. A revision of the burrowing spiders of the genus Geolycosa (Araneae, Lycosidae). American Mid. Nat., 27:1-62. Manuscript received 24 October 1 994, revised 31 March 1995 . 1995. The Journal of Arachnology 23:7 1-74 OBSERVATIONS OF HABITAT USE BY SARINDA HENTZI (ARANEAE, SALTICIDAE) IN NORTHEASTERN KANSAS Stephen R. Johnson1: Division of Biology, Ackert Hall, Kansas State University, Manhattan, Kansas 66506 USA ABSTRACT. During the mid-autumn of 1 993 and spring through summer of 1 994, 1 sampled tallgrass prairie in northeastern Kansas to determine habitat use, distribution and population dynamics of the ant-like jumping spider Sarinda hentzi (Banks). This species was not detected in any censused area of tallgrass prairie until early October 1993, when large juvenile specimens were found in a density of 0.54 individuals/m2 on inflorescences of Indian grass ( Sorghastrum nutans (L.) Nash.) in a biennially burned watershed. During the following spring, 0.6 adults/m2 were found on Indian grass panicles in a four-year burned watershed and 0.4 adults/m2 were located on Indian grass inflorescences in another biennially burned watershed. By mid-summer 1 994, this species was no longer found on Indian grass and only very small juveniles were found on various Car ex and Juncus species at the margins of wetlands. Spiders which mimic ants have a large suite of morphological and behavioral adaptations (Mclver & Stonedahl 1993). Many of the ant- mimicking spiders belonging to the families Clubionidae and Salticidae avoid predation by their mimicking deception (Cutler 1991). The feeding behaviors, food preferences and general biology of selected ant-like salticids have been well documented by Jackson (1982, 1986) and Wing (1983); however, there is little available information on habitat use by North American ant-like salticids (Duffy 1978; Cutler, pers. comm.). In northeastern Kansas, the family Salticidae is most commonly represented by the genera Me- taphidippus and Phidippus (Fitch 1963). Two species of ant-like salticids, Sarinda hentzi (Banks) and Synomosyna formica Hentz, may also be common (Fitch 1963; S. Johnson, pers. obs.); however, virtually nothing is known about the natural history of these species in this region (B. Cutler, pers. comm.). The purpose of this study was to 1) investigate habitat preferences, 2) estimate population den- sities and 3) record seasonal population dynam- ics of Sarinda hentzi in the tallgrass prairie of northeastern Kansas because S hentzi may be the most frequently encountered ant-like salticid 1 Present address: Department of Biology, Virginia Commonwealth University, Richmond, Virginia 23284-2012. of the tallgrass prairie in northeastern Kansas (S. Johnson, pers. obs.). METHODS Research was conducted at the Konza Prairie Research Natural Area (KPRNA) located ap- proximately 10 km south of Manhattan, Kansas (39®Q8'N, 96°35'W). This site is a member of the Long Term Ecological Research Network estab- lished in 1981 (Callahan 1984). Fire is an im- portant component in the maintenance of tall- grass prairie where, in the past, natural fires were common events (Pyne 1982; Abrams etal. 1986). Today, fire is a frequent management practice in tallgrass prairie (Anderson 1972; Blankespoor 1987). All prescribed fires occurred on KPRNA in mid-Spring (late March through April). In ear- ly June and mid- August through mid-October 1993, I conducted a survey to ascertain distri- bution and habitat preferences of Sarinda hentzi. Samples were collected by taking 50 sweeps while walking along parallel transects in hilltop pla- teau, midslope and wetland margin sectors of tallgrass prairie (Cutler 1971). Three separate samples were collected for each distinct zone by sweeping three parallel transects in each of the three distinct zones (Fig. 1). Samples were taken in three annually burned, two biennially burned, two 1 0-year and two 20-year burned watersheds. The number of spiders captured per sample was converted to an estimated density of spiders/m2 by taking the number spiders/sample and divid- 71 72 THE JOURNAL OF ARACHNOLOGY Figure 1.— Positions of sweep samples taken in hilltop, midslope and wetland margin zones of tallgrass prairie. Each replicate consisted of 50 sweeps at the inflorescence level. Duplicate samples in the hilltop and midslope zones were taken at the level of basal leaves approximately 0.3 m above the ground. ing it by the number of inflorescences/m2 in the habitat where the sample was collected. The den- sity of Indian grass and big bluestem ( Andropo - gon gerardii Vitman) culms and inflorescences was determined by taking 50 0.1 m2 quadrat samples in each location where sweep samples were taken. Where S. hentzi was frequently en- countered, sweeps of single infloresences were made to determine the spider density per indi- vidual inflorescence. This single inflorescence data was also used as an index of the estimated density of individuals/m2. To determine if S. hentzi was present elsewhere in the habitat, sep- arate sweep samples were taken among basal leaves of the grasses. Captured spiders were counted after each individual sweep and subse- quently released back onto grass infloresences. In 1 994, sampling began in early May and con- tinued until early August. Samples were taken in the same manner as described for 1993; however, the sampling procedure was expanded to include two additional four-year burned watersheds. RESULTS AND DISCUSSION In 1993, a year with rainfall amounts 33 cm above a 30-year mean (KPRNA weather data, unpubl.) no S. hentzi were collected in any wa- tershed until early October. Otherwise, from June to early October, the most commonly collected spiders on inflorescences of Indian grass (Sor- ghastrum nutans) in annually and biennially burned (1992 burned) watersheds were Marpissa pikei (G. & E. Peckham) (Salticidae) (0.45 ± 0.18 individuals/m2 in June, 0.27 ±0.15 ind./m2 in October), Thiodina puerpura Hentz (Salticidae) (0.32 ±0.19 ind./m2 in June, 0.18 ± 0.09 ind./ m2 in October) and Tibellus duttoni (Hentz) (Philodromidae) (0.56 ±0.12 ind./m2 in June, 0.32 ± 0.2 ind./m2 in October). In these water- sheds, Indian grass inflorescences occurred in an average density of 40.3 ± 8.2/m2 and the total density of Indian grass culms was 59.6 ± 12.2/ m2. Where big bluestem was dominant, it oc- curred in a density of 73.6 ± 14.3 culms/m2 with 68.3 ± 8.4 inflorescences/m2. In adjacent 10 and 20 year burned watersheds both Indian grass and big bluestem were more patchily distributed and occurred in less dense and often mixed stands. There, Indian grass inflorescences occurred in a density of 12.5 ± 2.2/m2 and big bluestem in- florescence densities were comparable. Also, broad leaved forbs were more common. In early June, Metaphidippus galathea (Walckenaer) and Phidippus clarus Keyserling were the most com- mon species collected. In early October, inflo- rescences of big bluestem were being used as perches by Phidippus audax (Hentz) (0.3 ± 0.19 ind./m2) and P. apacheanus Chamberlin & Gertsch (0.12 ± 0.08 ind./m2). No S. hentzi were found in annually, 10- or 20-year burned water- sheds. Beginning in early October 1993, large juvenile S. hentzi (~ 10 mm in length) were collected in a density of 0.54 ±0.12 individuals/m2 on in- florescences of Indian grass in the mid-slope transects of a biennially burned watershed (Fig. JOHNSON —HABITAT USE BY SARINDA HENTZI 73 Figure 2.— The estimated density of Sarinda hentzi individuals/m2 in early October 1993 and from 8 May through 15 August 1994. Vertical bars represent one standard error of the mean. 2). These juveniles were occasionally found in groups of two or three per inflorescence in pure stands of Indian grass. By mid-October, no S. hentzi juveniles were found in this area. In early May 1994, a year with a more normal level of rainfall, S. hentzi adults were found in a density of 0.6 ± 0.15/m2 on old inflorescences of Indian grass in a four-year burned watershed (Fig. 1). By mid-May, in the same watershed, the number of adults in Indian grass inflorescences had dropped to 0.25 ± 0. 14 individuals/m2 (Fig. 2). Sweeps of grass blades around these inhabited Indian grass canes failed to capture any S. hentzi. By the end of May, adult S. hentzi were no longer found on Indian grass inflorescences in this wa- tershed. In a biennially burned watershed located sev- eral kilometers south of the four-year burned wa- tershed, 0.35 ± 0.19 adult S. hentzi/m2 were found in Indian grass inflorescences on the hill- top sampling position in early June (Fig. 2). These individuals were sympatric with the same three grass-inhabiting species: M. pikei, M. puerpura and T. duttoni. Twice in the early June sampling I captured M. pikei being eaten by T duttoni ; however, I never observed S. hentzi being preyed upon by any of the other mentioned species. By late June all species were in reduced den- sities which were concentrated farther downslope from the sites sampled in early June. By mid- July, these grass-inhabiting species were most concentrated at the margins of wetlands among Carex annectens Bicknell, C. hystericina MuhL, Eleocharis erythropoda Steudel and Juncus tor- reyi Coville where the combined density of plant culms was 86.8 ± 23.5/m2. Here, I found the M. pikei and T. duttoni in densities of more than 0.7 individuals/m2 each. Conversely, I found only one adult S. hentzi in 1 00 sweeps. By mid- August I found no adult S. hentzi but did capture tiny juveniles (1.5-2 mm in length) in a density of 0.25 ± 0.2 ind./m2 (Fig. 2). A similar density of juvenile S. hentzi was found in the margin of the wetland basin of the four-year burned watershed where the high density of adults had been found during the preceding May. From these observations, I hypothesize that S. hentzi is a grass inhabiting species somewhat like Ma.rpi.ssa pikei in northeastern Kansas but may be more restricted in habitat. Because S. hentzi was only found in intermediate bum frequencies and generally absent from annually-burned and long-term unbumed areas, it may be avoiding potentially greater competition from higher den- sity spider populations in annually-burned prai- rie, following the pattern of many other grass- inhabiting spiders in long-term unbumed prairie (Weaver 1987; Usher 1988; S. Johnson, unpubl. data). With this habitat specificity, S. hentzi may also be an indicator of environmental quality (Clausen 1986). Also, it seems either to be prone to large fluctuations in population density or is sensitive in some way to sweep net sampling. Some evidence for this such sensitivity comes from population density dynamics of sympatric spiders. For example, while numbers of adult S. 74 THE JOURNAL OF ARACHNOLOGY hentzi dropped precipitously from early May to early June, numbers of Metaphidippus galathea and Tihellus duttoni in the same area remained fairly constant over the same period of time (0.27 ±0.18 and 0.66 ± 0.21 ind./m2 respectively, in early May; 0.21 ± 0.06 and 0.59 ± 0.20, re- spectively, in early June). Furthermore, while ex- perimental manipulation and statistical analysis were not part of this descriptive study, these ob- servations may help formulate appropriate ques- tions to design experimental work on arthropod habitat preferences in the tallgrass prairie eco- system. ACKNOWLEDGMENTS I thank Drs. Bruce Cutler and G. B. Edwards for very helpful advice and species identifica- tions. I also thank Drs. David B. Richman and Jan Weaver for helpful comments on an earlier version of this manuscript. This research was supported by the Konza Prairie Long Term Eco- logical Research Program (NSF grant DEB 9011662), the NSF Conservation and Restora- tion Program (DEB 9100164) and the Kansas Agricultural Experiment Station, LITERATURE CITED Abrams, M. D., A, K. Knapp & L. C. Hulbert. 1986. A ten year record of aboveground biomass in a Kan- sas tallgrass prairie: effects of fire and topographic position. American J. Bot., 73:1509-1515. Anderson, R. C. 1972. The use of fire as a manage- ment tool on the Curtis Prairie. Proceedings, Tall Timbers Fire Ecol. Conf., 12:23-35. Blankespoor, G. W. 1987. The effect of prescribed burning on a tallgrass prairie remnant in eastern North Dakota. Prairie Nat., 19:177-188. Clausen, I. H. S. 1986. The use of spiders (Araneae) as ecological indicators. Bull. British Arachnol Soc., 7:83-86. Cutler, B. 1971. Spiders from heading bluegrass (Poa pratensis L.) in Roseau County, Minnesota. Mich- igan Entomol., 4:123-127. Cutler, B. 1991. Reduced predation on the antlike jumping spider Synageles occidentalis (Araneae: Salticidae). J. Insect Behav., 4:401-4-07. Duffey, E. 1978. Ecological strategies of spiders in- cluding some characteristics of species in pioneer and mature habitats. Symp. ZooL Soc. London, 42: 109-123. Fitch, H. S. 1 963. Spiders of the University of Kansas Natural History Reservation and Rockefeller Ex- perimental Tract. Univ. Kansas Mus. Nat. Hist., Miscell. Publ. No. 33. 202 pp. Jackson, R. R. 1982. The biology of ant-like jumping spiders: intraspecific interactions of Myrmarachne lupata (Araneae, Salticidae). ZooL J. Linn. Soc., 76: 293-319. Jackson, R. R. 1986. The biology of ant-like jumping spiders (Araneae, Salticidae): prey and predatory be- haviour of Myrmarachne with particular attention to M. lupata from Queensland. ZooL J. Linn. Soc., 80:179-190. Mclver, J. D. & G. Stonedahl. 1993. Myrmecomor- phy: morphological and behavioral mimicry of ants. Ann. Rev. Entomol., 38:351-379. Pyne, S. J. 1982. Fire in America. Princeton Univ. Press, Princeton. Usher, M. B. & L. M. Smart. 1988. Recolonization of burnt and cut heathland in the North York moors by arachnids. Naturalist, 1 13:103-1 11. Weaver, J. C. 1987. The effect of fire on the spider community of a native tallgrass prairie. Ph. D. dis- sertation, Univ. Missouri, Columbia. Wing, K. 1 983. Tutelina similis (Araneae: Salticidae): an ant mimic that feeds on ants. J. Kansas Entomol. Soc., 56:55-58. Manuscript received 19 December 1994, revised 24 April 1994 . 1995. The Journal of Arachnology 23:75-84 LABORATORY STUDIES OF THE FACTORS STIMULATING BALLOONING BEHAVIOR BY LINYPHIID SPIDERS (ARANEAE, LINYPHIIDAE) Gabriel S. Weyman: School of Biological Sciences, University of Southampton, Bassett Crescent East, Southampton, S09 3TU, UK ABSTRACT. Linyphiid spiders were tested individually in the laboratory in order to assess possible factors stimulating the onset of ballooning behavior. An air low chamber was used for many of the tests. Air movement was found to be an important stimulus for initiating both the climb to a prominent position and subsequent take-off attempts. After instigation of pre-light behavior by an initial air low stimulus, climbing continued in still air, though take-off attempts generally ceased in the absence of further air movement. Neither circadian rhythmicity nor darkness were found to prevent exhibition of ballooning behavior at night. Length of time spent attempting to take-off appeared to be a factor in reducing a spider’s response to the stimuli causing ballooning behavior. Ballooning is the term commonly used to de- scribe aeronautic dispersal using wind drag on threads of silk for lift, as exhibited by several families of spiders. At certain times of the year these spiders take to the air en masse, though there are lower numbers ballooning throughout the year (Freeman 1 946; Sunderland 1991; Wey- man et al. 1995). There must be innate factors governing which spiders will balloon and at what stage during their lifetime, possibly with envi- ronmental factors adjusting the likelihood of its occurrence at a particular time (see Weyman 1 993 for a review). However, occurrence of this critical component of spider population dynamics is fi- nally driven by individual behavior in response to immediate stimuli. Although several studies have been carried out to investigate when, and under what meteoro- logical conditions, spiders take to the air (e.g., Greenstone 1990; Yugts & van Wingerden 1976; Thomas 1993), very little work has been done on the immediate factors that stimulate spiders to initiate and cease the behaviors that result in flight. The current investigation explored some of the factors that may be responsible by ob- serving individual spiders subjected to a variety of stimuli in the laboratory. This set of basic experiments assessed the importance of each stimulus, to detect those which are worthy of more intensive investigation. Air flow is an important factor in ballooning because it provides the source of power for spider flight (Humphrey 1987; Suter 1991, 1992). It would seem likely, therefore, that air movement might be the stimulus that elicits ballooning be- havior (taken here to include the pre-flight be- haviors). It is well documented that spiders will not take off in wind speeds over 3 m/s (Richter 1970; van Wingerden & Vugts 1974; Vugts & van Wingerden 1976; Greenstone 1990). There is also limited evidence that spiders do not bal- loon at night (Bishop 1990). Suction trap data (Sunderland unpubl.) show extremely low num- bers of spiders caught at night, even between high day catches (maximum night catch of 5, com- pared with a maximum day catch of 178). There are two possible explanations for this: 1) that meteorological conditions at night are not suit- able for ballooning; 2) that pre-flight behavior (climbing to a high point and attempting to take- off) ceases at a certain time in response to the light-dark cycle or an endogenous circadian rhythm. Laboratory experiments were carried out in an air flow chamber, similar to one used by Legel & van Wingerden (1980), to determine whether air movement was a sufficient stimulus to elicit climbing and take-off behavior. Further experi- ments were carried out to determine whether climbing and take-off behavior require contin- uous stimuli, or whether they continue after a short initial stimulus, in the manner of a fixed action pattern (FAP) (Manning 1979). The hypothesis that ballooning does not occur 75 76 THE JOURNAL OF ARACHNOLOGY 120 cm 60 cm Figure L— Schematic diagram of the laboratory air flow chamber. at night for behavioral reasons was tested in the laboratory. Evidence for an endogenous circa- dian rhythm was sought by subjecting spiders to ballooning stimuli during the night. Other spi- ders were tested for ballooning behavior in ef- fective darkness. Duration of the ballooning be- havior was also assessed as a possible factor lim- iting a spider’s responsiveness to the stimuli pre- sented. METHODS Collection of spiders.— Erigone spp. were col- lected as required from unsprayed grassland on a farm 20 miles north of Southampton, UK, by Bietriek vacuum insect net (D-vac). Samples were sorted on the day of collection and spiders were placed individually in 3 cm diameter Petri dishes with a small piece of moistened filter paper to avoid desiccation. Erigone spp. were identified live under a binocular microscope before testing, or preserved in 70% alcohol after testing awaiting identification. The spiders were not identified to species, but only £, atra (Wid.) and E. dentipalpis (BL) were found at the collection site in popu- lation samples during 1991 and 1992. These two species are expected to exhibit very similar be- haviors as they have virtually identical life-cy- cles, niches and habits (De Keer & Maelfait 1 988). Storage and testing of spiders were carried out in a controlled environment room at 1 8 °C ± 2, relative humidity 70% ± 10, L16:B8 cycle (ap- proximately in phase with external conditions), except where stated otherwise. The air flow chamber. —The laboratory air flow chamber used here was an acrylic sheeting (“per- spex” or “Plexiglas”) box 120 cm high, with a 55 x 60 cm base (Fig. 1). A suction fan in the top (Rotron Inc. Whisper Fan WR3A1, 230Y AC, 1 2 W) drew air through a 1 2 x 1 2 cm square opening in the center of the base, covered with metal gauze as a support. White netting covered WEYMAN — LABORATORY TESTS OF FACTORS STIMULATING BALLOONING 77 the whole of the base to aid viewing and a wood- en climbing frame with 1 2 cm uprights was placed over the opening. The updraft at 1 cm above base level in the frame was 0.57 ms-1 ± 0.01 (measured with a Solomat MPM 500e hotwire anemometer, 950 readings logged at 5 s inter- vals). Take-off within the confines of the cham- ber was not normally possible but pre-flight be- havior could be observed and recorded easily. This chamber was also used by Weyman et ah 1994, 1995). Erigone spp. are ground-dwelling spiders which do not normally climb (C. J. Topping pers. comm.), except when motivated to balloon. Having climbed to a prominent position, prior to take-off, a spider assumes the position known as “tip-toe” (Richter 1970), with legs extended and abdomen raised, and releases silk until the drag on the length of silk is sufficient to lift the spider from the substrate, or possibly until some higher optimal threshold is reached. In an alter- native take-off behavior, referred to here as “dropping” (after Jones 1994), the spider drops a short way on a thread, while a second thread provides wind-drag for lift. Factors eliciting ballooning behavior —Air- flow as the stimulus for climbing in the laboratory: The individual Petri dishes containing spiders were maintained in a large plastic container with high r.h. (« 100%) achieved by putting a small amount of water in the base. Testing for climbing behavior was then carried out in the laboratory air flow chamber. To observe the effect of airflow on the behavior of spiders previously in calm conditions, individual spiders were first ob- served in still conditions, then with air move- ment over them. Individual Petri dishes con- taining spiders were placed in the center of the climbing frame, with the fan off. A piece of plas- ticine (“modeling clay”) was attached to the lid of the Petri dish, attached to a cotton thread running out through a hole in the top of the chamber. The door of the chamber was then closed and the lid gently lifted off the Petri dish by pulling the cotton thread. The spider was ob- served for three minutes before the fan was switched on. Observations then continued for a further three minutes. Wind speed just above the open Petri dish was found to be 0.0 m/s with the fan off, and fluctuated between 0.0-0. 1 m/s with the fan switched on (measured with a Solomat MPM500e hotwire anemometer). Three separate experiments were carried out, with groups of 1 0, 29, and 29 spiders respectively. Air flow as the stimulus for take-off behavior in the laboratory: To determine the necessity of continuous air flow as a stimulus for ballooning behavior, 16 spiders were individually subjected to a short stimulation/disturbance by collection in a hand aspirator, and placed into the climbing frame in the chamber with the fan off. Each spi- der was observed for three minutes. Climbing and “tip-toe” behaviors were recorded. After a short rest period the same spiders were re-tested in the same way, but with the fan turned on. To further examine the effect of changing air flow stimulus, individual spiders were placed into the climbing frame, with the fan off, and ob- served for 44 minutes with the fan alternately turned on or off at ten minute intervals, starting with a still period. The occurrence of several be- havioral categories was recorded: moving or sta- tionary at base level (assigned the term “ground”); climbing the uprights (“climbing”); activity at the top of the uprights (“top active”); inactivity whilst at the top of the uprights (“top inactive”); take-off attempts by the dropping method (“dropping”); take-off attempts by the tip-toe method (“tip-toe”). The experimental time was divided into 30 s periods, and assessment of be- havior was by a standard presence or absence (1/ 0) method during each 30 s. The most advanced behavior towards take-off was noted for each 30 seconds. This method gave four interfaces of changing conditions for each spider tested and would show any possible effects of previous con- ditions on subsequent ballooning behavior. Ten minute intervals were used to allow valid as- sessment of individual behavior types over a large number of 30 s periods between each interface, with the final four minutes to allow any change after the final interface to become clearly appar- ent. Sample intervals of 30 s were chosen by experience, for logistic reasons. Four individuals were tested in this way, each constituting an in- dependent repetition. Factors limiting response to initiation stimu- li. — Testing for endogenous rhythms: Testing was carried out in a portable ballooning chamber, similar to the one described above but modified with a plywood base and a DC fan (Fapst mul- tifan 4 1 32, 9 cm, 1 2 Y DC, 5 W), giving an equiv- alent air flow. The chamber was situated in a room which did not have environmental control facilities, but variations in ambient humidity and temperature over the experimental period were not expected to affect the results. The room was illuminated naturally during the day and by a 78 THE JOURNAL OF ARACHNOLOGY exp 1 exp 2 exp 3 experiment climb with fan off Q] climb when fan on Figure 2. —The percentage of spiders climbing during the test period with the fan off, and the additional number climbing after the fan was switched on. 100 W incandescent source during dark hours only while testing was in progress. The results gained could be used to assess the necessity for more controlled conditions. Thirty-two spiders were individually tested for pre-aeronautic behavior within three minutes of being placed via hand aspirator onto the base of the wooden frame. Climbing and “tip-toe” be- haviors were recorded over a three minute pe- riod. The spiders were tested once during the daytime, once during darkness, and once again in the following light period. This was repeated for each spider on the subsequent day. Testing for limitation of ballooning response in darkness: To test whether darkness limits re- sponse by spiders to a ballooning stimulus, lin- yphiids were collected by hand aspirator as they attempted to take-off from fences and grass at the field collection site. The spiders were placed individually in Petri dishes then placed into an opaque box. The spiders were kept in the box for approximately three weeks before being as- sessed for ballooning behavior. Testing was carried out in an open arena with a fan blowing across it, where infra-red lighting was available, and where take-off was possible. Spiders are not thought to be sensitive to infra- red light (M. F. Land pers. comm.). Light inten- sity at floor level, where the spiders were placed, was 0.05 m Einsteins/m2/s (readings taken with a Li-Cor model Li-185B photometer). Ten spi- ders from the opaque box were individually test- ed for three minutes. Climbing, “tip-toe”, and take-off behaviors were recorded. Testing for limits to duration of time spent at- tempting take-off: To assess the duration of bal- looning behavior, individual spiders were ini- tially given six min in the laboratory ballooning chamber, with the fan on, to initiate ballooning behavior. Presence or absence of “ground”, climbing, “dropping” and “tip-toe” behaviors was determined over one min periods. If a spider exhibited ballooning behavior then it was al- lowed to continue until it descended to ground level and remained there for six consecutive min, with no indication of further climbing. The test WEYMAN LABORATORY TESTS OF FACTORS STIMULATING BALLOONING 79 behaviour fan off | | fan on Figure 3.— The behaviors exhibited in the ballooning chamber within three minutes of release with the fan either on or off, after the initial stimulation by hand aspiration. was then terminated. Individuals were re-tested to determine whether cessation was permanent or temporary. Six min was allowed to give added assurance of a spiders intent to start or stop bal- looning, compared to the usual three min period of other experiments. This was important be- cause the experiment was designed to measure accurately time spent showing ballooning be- havior for individual spiders, rather than being a qualitative comparison between different groups or conditions, as made in experiments with the shorter time period. A total of 45 tests was car- ried out on 2 1 spiders. RESULTS Factors eliciting ballooning behavior. —Air flow as the stimulus for climbing: A small proportion of spiders climbed with the chamber fan switched off. They may have responded to air movement lower than the threshold of the measuring equip- ment, or to disturbance as the Petri dish lid was removed (Fig. 2). The onset of air flow elicited climbing for most of the spiders, however. Air flow as the stimulus for take-off' behavior: After aspiration, as many of the 16 spiders climbed with the fan off as when it was on (Fig. 3), with no significant difference between the tests {G = 0.1, df= 1, P > 0.05). Although climbing behavior often ensued following the initial stim- ulus, “tip-toe” behavior was rare and only in- creased after further stimulation (Fig. 3), though the observed difference was not significant ( G = 2.3, df = 1, P > 0.05). The limitations of the G-test in this context, where repeated measures are made on the same subjects, are recognized. It is used here as a simple indicator of the level of differences between tests. In the second set of experiments, where four spiders were tested individually for 44 min each, pre-flight climbing again continued in the ab- sence of the air stimulus (Figs. 4-7). This may suggest the existence of a FAP for ballooning behavior. “Tip-toe” behavior was again largely 80 THE JOURNAL OF ARACHNOLOGY Figure 4.— The effect of air movement on pre-ballooning behavior for spider A. Behavior recorded in 30 second periods with fan on or off. dependent on air flow, but did occasionally ap- pear to be exhibited in its absence. This was difficult to assess, however, because spiders did not seem to fully extend into the “tip-toe” po- sition unless the silk thread released was being pulled by the updraft. All four spiders tested showed marked differences in their behavior when the fan was switched on or off. Most no- ticeably, a much higher proportion of time was spent tip-toeing with the fan on, and more time was spent climbing with the fan off (Table 1). Only spider B showed signs of ending ballooning behavior during the 44 min period (Fig. 5). Factors limiting response to initiation stimu- li . — Endogenous rhythms: Spiders showed climbing and “tip-toe” behaviors at all test times Figure 5.— The effect of air movement on pre-ballooning behavior for spider B. Behavior recorded in 30 second periods with fan on or off. WEYM AN — LABORATORY TESTS OF FACTORS STIMULATING BALLOONING 81 Table 1.— The proportion of the total assessment points at which each behavior was exhibited by each of four spiders. Spider A Spider B Spider C Spider D Behavior Fan on Fan off Fan on Fan off Fan on Fan off Fan on Fan off Tip-toe 0.575 0.000 0.100 0.000 0.500 0.083 0.550 0.021 Dropping 0.025 0.042 0.050 0.021 0.000 0.021 0.150 0,042 Top active 0.150 0.396 0.375 0.167 0.400 0.271 0.225 0.354 Top inactive 0.000 0.021 0.075 0.063 0.000 0.000 0.000 0.021 Climb 0.250 0.458 0.100 0.438 0.100 0.542 0.075 0.563 Ground 0.000 0.083 0.300 0.313 0.000 0.083 0.000 0.000 40 48 40 48 40 48 40 48 (Table 2), with no significant differences between times (G-test, P > 0,05, df = 2), suggesting that there is no endogenous circadian rhythm limiting ballooning behavior at night. Statistical limita- tions of the G- test apply, as stated above. Effect of darkness: Ten spiders were tested un- der infra-red light. Six of these showed balloon- ing behavior: three of the six showed climbing and “tip-toe” behavior, and three climbing and “dropping”. One of the tip-toeing spiders be- came airborne. Duration of time spent attempting to take-off: Of the 21 spiders tested, ten showed ballooning behavior on at least one occasion. A total of 1 6 ballooning periods was observed. Time to ces- sation of ballooning behavior ranged from 3-267 min. The mean duration ± SD was 48 min ± 73. Six of the 16 ballooning periods were in re- tests of spiders that had shown ballooning be- havior on a previous occasion, suggesting that respites are only temporary. There was no in Table 2.— Percentages of spiders showing pre-bal- looning behavior during the day and at night in the laboratory ballooning chamber, (ns = not significant) (dark at 1930 h). Testing time % climbing % tip-toeing Number tested Replicate 1 1 800™! 930 h 53% 44% 32 0000-0140 h 72% 44% 32 1 100-1300 h 56% 32% 32 G-statistic 2.6 (ns) 1.4 (ns) Replicate 2 1400-1600 h 63% 25% 32 2220-2400 h 63% 41% 32 0940-1 140 h 48% 23% 31 G-statistic 1.8 (ns) 2.8 (ns) dication of individuals having set durations for ballooning behavior periods, one spider showing a range of 8-129 min to cessation. DISCUSSION Air flow was indicated as an important initial stimulus in the initiation of ballooning behavior. In the experiment to determine whether air flow is the stimulus for climbing it could be argued that the extra time allowed more spiders to ex- hibit climbing during the second three min pe- riod, with the fan on. However, many spiders were observed to have an immediate and strong climbing response as soon as the fan was turned on, indicating the importance of air flow as a stimulus. When spiders were tested under infra-red lighting, to which they are thought to be insen- sitive, ballooning behavior could still be insti- gated. Thus darkness was not found to limit bal- looning. Spiders tested for ballooning tendencies at different times showed no indication of an endogenous circadian rhythm limiting balloon- ing after dark. These results do not necessarily mean that spiders regularly balloon at night, but they certainly suggest that there is no endogenous reason why they should not, due to circadian rhythms or photo-responsiveness. Meteorology is implicated as the major factor limiting bal- looning at night. Conditions may simply not be suitable for take-off, even if spiders are attempt- ing to fly at night Farrow (1982, 1986) found spiders flying in large numbers at night in the upper air. However, this was facilitated by a noc- turnal temperature inversion which actually pre- vents take-off as the surface air cools. The spi- ders, therefore, took off during the daylight hours, and remained aloft until re-inversion of the tem- perature gradient the next day. Observations might reveal spiders attempting to take-off at 82 THE JOURNAL OF ARACHNOLOGY Figure 6. —The effect of air movement on pre-ballooning behavior for spider C Behavior recorded in 30 second periods with fan on or off. night in greater numbers than previously record- ed, but not succeeding. It appears that, once initiated, ballooning be- havior is carried to completion (take-off, when possible) or until some limit is reached (indicated here as temporal or energetic), even if the stim- ulus is removed. This suggests that some sort of fixed action pattern (FAP) acts to maintain the behavior for some time after initiation, in the absence of further stimulus. An alternative ex- planation could be that a lack of air movement provided the stimulus for continuation of the behavior when the fan was switched off. How- ever, lift from an updraft is necessary for take- off so it would be expected that spiders should stop the pre-ballooning behavior and descend, or at least become less active to conserve energy and lower their visibility to predators such as Figure 7. —The effect of air movement on pre-ballooning behavior for spider D. Behavior recorded in 30 second periods with fan on or off. WEYMAN — LABORATORY TESTS OF FACTORS STIMULATING BALLOONING 83 birds, if ballooning attempts became unsuccess- ful The reason for the FAP type activity observed here is unclear. As mentioned above, spiders nor- mally occupying a niche on or near the ground under dense vegetation must be risking much higher chances of predation by climbing to a prominent position and moving around. It is possible that a FAP to continue the behavior allows for natural changes in the airflow above the vegetation, so the spiders do not continually climb and descend as windspeed rises and falls, thereby possibly missing suitable conditions for take-off attempts. The temporal shut-off suggested here, in the experiment to determine the duration of bal- looning behavior, is not representative of the field, where conditions are not continually suitable or stimulating and there may be enforced breaks in ballooning, as well as the voluntary respites sug- gested here. Also, in the laboratory flight was not possible, only constant attempts. A period of just 10 minutes of take-off attempts may equate to several successful flights in the field. Thomas (1993) found wide variation in the time between flights in spiders observed ballooning in the field, ranging from 1-66 min, with a modal interval of 1 min before flight was resumed. These times are taken between successful flights and, of course, are subject to changes in conditions for stimulus and flight. Results here suggest that spiders show highly variable duration of ballooning bouts, inter- spersed with temporary rest periods which, in combination with variable numbers of flights per day and variable flight distance (Thomas 1993) means that individuals will travel extremely variable distances on a day when conditions are suitable for ballooning. Dispersal will be very high, as will the variety of habitats sampled, compared to a group all travelling the same dis- tance in the same direction and landing more or less in a group downwind from the start point. It would be extremely useful, in terms of mod- elling spider spatial dynamics, to observe the full temporal and spatial range of different balloon- ing behaviors, and distances travelled in flight (and, indeed, cursorily) for a large number of individual spiders in the field. However, it is virtually impossible to follow a spider in flight for more than a short distance, or to follow one through the vegetation. It would also be valuable to discover the full range and nature of the stim- uli involved in ballooning in the field, though expensive equipment and many observation hours are required for accurate meteorological readings to be taken within and above a crop in combination with behavioral observations. Some work of this nature has been undertaken, by Tho- mas (1993), van Wingerden and Vugts (1974), and by the current author (unpubl), for example, though the complex interactions between mete- orological factors have tended to obscure results pertaining to stimulation of spider ballooning be- havior. ACKNOWLEDGEMENTS Thanks to Drs. Paul C. Jepson, Keith D. Sun- derland, Chris J. Topping, and C. F. George Tho- mas for expert advice and help in the preparation of this manuscript. The work was funded by a SERC CASE award at the University of South- ampton, in collaboration with Horticulture Re- search International, Worthing Road, Little- hampton, UK. LITERATURE CITED Bishop, L. 1990. Meteorological aspects of spider ballooning. Env. Entomol, 19:1382-1387. De Keer, R. & J. P. Maelfait. 1988. Laboratory ob- servations on the development and reproduction of Erigone atra Blackwall, 1833 (Araneae, Linyphi- idae). Bull British Arachnol. Soc., 7:237-242. Farrow, R. A. 1982. Aerial dispersal of microinsects. Pp. 51-66, In Proc. 3rd Australian Conf. Grass!. Invert. Ecol.. (K. E. Lee, ed.). S. A. Govt, printer, Adelaide. Farrow, R. A. 1986. Interactions between synoptic scale and boundary-layer meteorology on micro-in- sect migration. Pp. 185-195, In Insect flight: dis- persal and migration. (W. Danthanarayana, ed.). Springer, Berlin. Freeman, J. A. 1946. The distribution of spiders and mites up to 300 feet in the air. J. Anim. Ecol., 15: 69-74. Greenstone, M. H. 1990. Meteorological determi- nants of spider ballooning: the roles of thermals vs the vertical windspeed gradient in becoming air- borne. Oecologia, 84:164-168. Humphrey, J. A. C. 1987. Fluid mechanic constraints on spider ballooning. Oecologia, 73:469-477. Jones, D. 1 994. How ballooning spiders become air- borne. Newsletter British Arachnol. Soc., 69:5-6. Legel, G. J. & W. K. R. E. van Wingerden. 1980. Experiments on the influence of food and crowding on the aeronautic dispersal of Erigone arctica (White, 1852) (Araneae, Linyphiidae). Pp. 97-102, In Proc. 8th Int. Arachnol. Cong. (J. Gruber, ed.). Egermann, Vienna. Manning, A. 1979. An introduction to animal be- haviour, 3rd ed.. Edward Arnold. 84 THE JOURNAL OF ARACHNOLOGY Richter, C. J. J. 1970. Aerial dispersal in relation to habitat in eight wolf-spider species. Oecologia, 5:200- 214. Sunderland, K D 1991. The ecology of spiders in cereals. Pp. (1):269-28Q. In Proc. 6th Int. Symp. Pests and Diseases of Small Grain Cereals and Maize (T. Wetzel & W. Heyer, eds.)- Board of Plant Pro- tection Halle, Halle/Salle, Germany. Martin Luther Universitat, Halle Wittenberg. Suter, R. B. 1991. Ballooning in spiders: results of wind tunnel experiments. EthoL Ecol. Evol., 3: IS- IS. Suter, R. B. 1992. Ballooning: data from spiders in freefall indicate the importance of posture. I. Ar- achnol., 20:107-113. Thomas, C. F. G. 1993. The spatial dynamics of spiders in farmland. Ph. D. thesis, University of Southampton. Vugts, H. F. & W. K. R E, van Wingerden. 1976. Meteorological aspects of aeronautic behaviour of spiders. Oikos, 27:433-444. Weyman, G. S. 1993. The possible causative factors and significance of ballooning in spiders. EthoL Ecol. Evol., 5:279-291. Weyman, G. S., P. C. Jepson & K. D. Sunderland. 1 994. The effect of food deprivation on aeronautic dispersal behaviour (ballooning) in Erigone spp. spi- ders. Entomol. Exp. et App., 73:121-126. Weyman, G. S., P. C. Jepson & K. D. Sunderland. 1995. Do seasonal changes in numbers of aerially dispersing spiders reflect population density on the ground or variation in ballooning motivation? Oec- ologia, 101:487-493. Wingerden, W. K. R. E. van & H. F. Vugts. 1974. Factors influencing aeronautic behaviour of spiders. Bull British Arachnol Soc., 3:6-10. Manuscript received 27 October 1994, revised 27 Feb- ruary 1995. 1995. The Journal of Arachnology 23:85-90 CHANGES IN BIOMASS OF PENULTIMATE-INSTAR CRAB SPIDERS MISUMENA VATIA (ARANEAE, THOMISIDAE) HUNTING ON FLOWERS LATE IN THE SUMMER Douglass H. Morse: Department of Ecology and Evolutionary Biology, Box G-W, Brown University, Providence, Rhode Island 02912 USA ABSTRACT. Penultimate-instar crab spiders Misumena vatia foraging on goldenrod Solidago spp. in late summer gained mass slowly, averaging 0.8 mg/day, and usually did not molt into the adult stage before the end of the summer. In contrast, adults gained mass rapidly at this site, averaging 8.8 mg/day, taking larger prey and taking them more often than penultimates. The penultimates’ prey consisted primarily of small flies, bees and wasps, and overlapped broadly with those of the adults, but did not include the most important resource of the adults, bumble bees, upon which the adults registered most of their gain. Penultimates captured prey biomass at only one-tenth the rate of the adults, and one-third the rate of adults per unit body mass. Most studies of foraging animals have focused on adults (Morse 1980; Pyke 1984; Stephens & Krebs 1986), even though preadult stages may occupy the majority of many species’ lifetimes. Proponents argue that this approach addresses the stage that contributes directly to reproduc- tion, permitting the most ready estimation of fitness (see Lewontin 1978; Maynard Smith 1978). However, foraging events earlier in the life-cycle may affect one’s success as an adult (Morse 1980; Skelly & Werner 1990; Fraser & Gilliam 1992), and lack of early success may result in many individuals not even surviving to the adult stage (e. g., Forrest 1987). Thus, infor- mation is needed on immature stages to compare with adult success, to determine how well adult foraging success describes foraging success in general (see Stein & Magnuson 1976; Sih 1982), and to establish how foraging repertoires vary over a range of sizes and ages. Different life-cycle stages may vary in their proficiency at capturing a given kind of prey. This pattern is clearly illustrated by the success of the semelparous crab spider Misumena vatia Clerck (Thomisidae) at capturing bumble bees (Bombas spp.), which only adult females can sub- due (Morse, unpubl. data). This prey is critical to fitness - without bumble bees, adult females seldom if ever gain enough mass to lay their single clutch of eggs (Fritz & Morse 1985). Stages lacking similarly profitable prey will grow slowly, and molting and maturation may be greatly extended (Levy 1970). In addition to dangers resulting from increased time spent as a juvenile, slow growth may make such individ- uals vulnerable to critical seasons that can act as bottlenecks. For instance, many species of spi- ders do not overwinter as adults (e. g., Schaefer 1977), including Misumena (Morse unpubl. data). To evaluate growth patterns of a stage without access to extremely profitable prey, I gathered data on the activity patterns, prey capture, and rate of gain in mass of penultimate female Mis- umena, which cannot capture bumble bees. The results permitted me to compare the success of these penultimates with adult female Misumena, whose foraging and growth patterns my col- leagues and I have studied in detail (Morse 1979, 1981, 1986, 1988; Morse & Fritz 1982, 1987; Fritz & Morse 1985; Kareiva, Morse & Eccleston 1989). For direct comparison, I also gathered similar data on the small number of prereprod- uctive adult females present in the study area at the same time. Most adult females had already laid their eggs and were guarding them at that time (Morse 1987). I conducted this study during late summer on Misumena that frequented goldenrods Solidago spp., the dominant native flowering plants in eastern North America at this time of year. Gold- enrods are the principal foraging sites for Mis- umena at this season (Morse 1981, 1993). In the absence of an “ideal” prey for the penultimates, comparable to bumble bees for the adults, one may predict a relatively slower growth rate from the penultimates than from adults. CHARACTERISTICS OF MISUMENA The semelparous crab spider Misumena vatia (Thomisidae) has a life cycle of one year or more. 85 86 THE JOURNAL OF ARACHNOLOGY These spiders frequent flowers that attract large numbers of their insect prey. Typically, adult females lay their single clutch of 75-350 eggs during mid- or late summer, with extremes of early July to early September (Morse & Fritz 1982; Fritz & Morse 1985). Females guard their eggs in nests constructed from leaves, often until the young emerge nearly a month later (Morse 1987). Young overwinter as immatures in the litter, usually as third to sixth instars, but with less frequent penultimates and second instars. Some individuals probably overwinter more than once, but I have no evidence that they overwinter as adults. To date I have not recovered in the following spring any of the more than 300 marked adults (over a period of 1 5 field seasons) that did not lay eggs in the fall (Morse unpubl. data). [Neither have I recovered in the following spring any of the 1400+ egg-layers I have studied (Morse 1994) during this time.] Following wintering, these spiders hunt on flowers for insect visitors. Females subsequently grow rapidly if they find hunting sites that attract many large insect prey, sometimes increasing as young adults by as much as an order of magnitude (40 mg to 400 mg) within as little as two-three weeks (Fritz & Morse 1985). Penultimate females range from about 18 mg to 75 mg or more (Holdsworth & Morse in press), and can be readily separated from adults by the absence of red dorsolateral stripes on their abdomens (Gertsch 1939). Males are strikingly smaller than females, seldom exceeding 7 mg, and averaging 3-5 mg (Floldsworth & Morse in press). METHODS I carried out this study in a one ha field in Bremen, Lincoln County, Maine. The field is covered by a variety of grass species, with con- siderable numbers of forbs scattered throughout. During late July and August 1981, when this study was conducted, goldenrod was the domi- nant flowering forb. The principal species of goldenrod, blooming in sequence, were Solidago juncea, S. canadensis , and S. rugosa. They will be considered collectively for the purposes of this study. Clonal in nature, these goldenrods typi- cally exhibit a clumped distribution, with 1 0-7 5 flowering stems per site in the study area. This site is described in further detail elsewhere (Morse 1979, 1981). I used 45 immature female Misumena in this study, which, at 1 8-34 mg, were almost certainly all in the penultimate instar. [In another study, all 49 females of this size range captured in the field and subsequently reared through to their next molt assumed adult condition at 3 1 mg or more, with a mean of 54.7 mg ± 12.5 SD (Hold- sworth & Morse unpubl. data).] These penulti- mates were added to the study as found, thus Day 1 of observations was not the same for all individuals. I gathered data simultaneously on a sample of eight prereproductive adult females, all of these individuals I could find in the study area at this time. The small sample size of adults deserves note. Numbers of these individuals were ex- tremely low in the study area, and the failure of additional small, prereproductive adult females to appear during the study period strongly sug- gested that very little, if any, recruitment oc- curred from the pool of penultimate females into the adult population. Spiders were initially given a unique number, placed on the abdomen, using indelible ink. They were then weighed on a Kahn Electrobalance, returned to their sites on goldenrod and censused for presence, location, and prey capture at least every two hours during the daytime for periods of up to two weeks. Repeat weighings were made for 1 3 of the penultimate individuals and six of the adults at intervals averaging three-four days. Visits at two-hour intervals ensured that I re- corded nearly all of the diurnal prey that the spiders captured (Morse 1979, 1981). Any noc- turnal captures, which are infrequent on gold- enrod and confined to moths taken by adults (Morse 1981), could be detected by the presence on the following morning of carcasses still being fed on or recently discarded (Morse 1981). The penultimate spiders were divided into three groups on the basis of the information gathered upon them: 1) weighed two (3), three (7) or four (3) times and observed for several days (5-12 days); 2) weighed once and observed more than one day (3-1 1 days); and 3) weighed once and observed for a single day. As a result, some of these individuals provide more information than others. Six of the adult females met the criteria of Category 1 (3-13 days) and the other two of Category 2 (2-6 days). Adults in Category 1 were weighed two (1), three (4), or four (1) times. Insect prey were recorded and assigned to group (usually order, but species in case of principal prey) and size category. Routine weighings of fresh specimens of these species (Morse unpubl. data) permitted rough estimates of prey biomass at time of capture. Although ingested biomass MORSE- CRAB SPIDER BIOMASS 87 Table 1.— Characteristics of penultimate and adult female spiders. First two variables drawn from entire sample of individuals; last three variables drawn from sample that was weighed two or more times. Penultimates Adults X ± SD N Range Jc ± SD N Range Time between first two moves (days) 4.0 ± 3.0 24 1-11 5.3 ± 3.6 8 1-13 Distance of moves (cm) 50.0 ± 52.4 24 10-200 40.8 ± 15.8 4 25-60 Gain in mass (mg/day) 0.8 ± 0.5 13 -0.2-24 8.8 ± 6.2 6 -0.8-22.3 Gain in mass (%/day) 3.0 ± 2.4 13 -0.9-10.3 9.1 ± 5.5 6 -1.0-16.8 Time observed (days) 9.0 ± 2.0 13 7-12 6.5 ± 3.6 6 3-13 would have permitted a more precise measure of prey intake, it was impossible to capture and weigh the free-ranging prey before the spiders caught them. However, an earlier study (Morse 1979) demonstrated that adult female Misumena extracted very similar proportions of the smallest and largest prey items exploited by spiders in this study (57.5% of the ca. 4 mg syrphid fly Toxomerus marginatus and 57. 1% of ca. 150 mg bumble bees Bombus spp.). Therefore, use of the wet masses of prey species should provide an acceptable estimate for comparisons. Biomass captured (estimated cumulative wet mass of all prey) was divided by total days of observation to determine the rate at which spiders obtained resources. I determined gains in mass from the spiders I weighed more than once. Rates of gain were established by dividing the total change in mass between the first and last weighings by the days elapsed. RESULTS Penultimate spiders weighed twice or more and observed several days initially weighed 25.4 mg ±6.1 SD; those weighed once and observed more than one day weighed 24.9 mg ± 5.3 SD; and those weighed and observed one day weighed 23.2 mg ± 4.5 SD. These three groups did not differ in mass ( H = 0.891, df = 2, n = 45, P > 0.5 in a Kruskal- Wallis one-way ANOVA). Therefore, I treated them as members of a single group for each variable about which they yielded data. Gains of penultimates were slow, though vari- able, averaging 3% of their body mass per day (Table 1). In these individuals the mean increase amounted to 0.8 mg/day. At this rate they would gain an average of 24 mg/month, one month being the maximum possible time remaining with substantial numbers of flowers in bloom as hunt- ing sites and temperatures favoring prey activity. Two of the 13 individuals even lost a small amount of mass (0.4, 0.9% of initial mass/day), falling from 23 to 22 mg over nine days, and from 24 to 22 mg over 1 1 days, respectively. The remainder gained from 0.5 to 10.3%/day, with only three of the individuals increasing over 1 .0 mg/day. The two fastest gainers increased from 23 to 42 mg in eight days, and 23 to 40 mg in 1 1 days, respectively. Original mass and subse- quent gain of this group of 1 3 spiders were not significantly correlated (rs= -0.3 16 in two-tailed Spearman Rank Correlation Coefficient, n = 13, P > 0.2). The penultimates’ gains in mass can be com- pared with those of the adult females occupying the same flowers (Table 1). The adults, which weighed nearly four times as much as the penulti- mates (94.2 mg ± 29.8 SD, n = 8), gained mass 1 0 times more rapidly than penultimates and, as a function of gain per unit mass, three times more rapidly than penultimates (Table 1). These dif- ferences were both significant (P = 0.01 in two- tailed Mann-Whitney GTests). Two of these adult females more than doubled their original mass, in 7 and 1 3 days, respectively; none of the penul- timates exhibited such large relative gains. [Orig- inal mass and subsequent gain were not corre- lated in this sample ( rs = -0.236 in a two-tailed Spearman Rank Correlation Coefficient, n = 6, P > 0.2).] This difference in uptake of mass mirrored the adults’ superiority in total numbers of prey cap- tured per unit time (G = 6.89, df = 1, P < 0.01 in a G-Test: Table 2) and size of prey captured ( P < 0.01 in a Mann-Whitney C/Test: Table 2). Notwithstanding the pronounced difference in gain of mass by adults and penultimates, they did not differ significantly in either the frequency with which they changed goldenrod inflores- cences (Table 1; P > 0.7 in a two-tailed Mann- 88 THE JOURNAL OF ARACHNOLOGY Table 2. —Prey captured on goldenrod by penultimate and adult female spiders. Penultimate {n = 45) Adult (n = 8) Prey Mass (mg) Number caught Total (mg) Number caught Total (mg) Toxomerus marginatus (Syrphidae) 4 13 52 1 4 Other small Diptera 8 6 48 2 16 Medium Diptera 15 1 15 2 30 Small bee 10 4 40 3 30 Small wasp 10 4 40 3 30 Butterfly 15 1 15 0 0 Moth 125 0 0 1 125 Bumble bee ( Bombus spp.) 150 0 0 4 600 Spider (Salticidae) 10 0 0 1 10 Total - 29 210 17 845 Total spider days 228 53 Capture/day (mg) 0.92 15.9 Size of item (mg) 7.2 49.7 Whitney U Test) or in the distance that they moved at such times (Table 1 ;P— > 0.5, same test). Prey species taken by penultimates and adults overlapped broadly, but the penultimates did not capture any bumble bees, the major source of prey gain for the adults. DISCUSSION Although the critical data set is small, adult and penultimate female spiders differed mark- edly in their energetics during the late summer. The low rate of gain in mass suggests that many of the penultimate instars (probable minimum of 9 out of 1 3) would not reach the size at which they would likely molt into the adult stage (mean = 55 mg, minimum = 3 1 mg: see Methods) while foraging on goldenrod. Even those that grew large enough to molt to the adult stage would be very unlikely to grow large enough to lay a clutch of eggs. I have never found Misumena laying a clutch in the field if they weighed less than 1 1 4 mg (Fritz & Morse 1985; Morse unpubl. data). That few penultimates did molt into adults during this pe- riod is further suggested by the small numbers of prereproductive adult females present at that time ( n = 8) and the extremely low rate at which additional small prereproductive adults were found over the study period. This growth pattern of penultimates contrasts with the small cohort of adults observed at this time, which gained considerable mass as a result of capturing bumble bees, a prey that lies beyond the size range available to the penultimates. Adults in another study on goldenrod also gained mass rapidly (Morse 1981), a trait observed ear- lier in the summer on other species of flowers as well (Morse 1981; Fritz & Morse 1985). The pe- nultimates’ slow growth also contrasts with the success observed for early-instar Misumena on goldenrod, which recruit regularly onto this flow- er from their nests (Morse 1993), and frequently increase in mass several fold as a result of feeding on the small Diptera that frequent goldenrod flowers. In this northerly clime, penultimates have at most a month of feeding time left before tem- peratures decline abruptly (early to mid-Septem- ber), and an even shorter time before the last goldenrods senesce (late August-early Septem- ber). Following the disappearance of the last goldenrod flowers in early September, only scat- tered asters ( Aster spp.) remain, such that the opportunities for successful hunting decline markedly for these flower-frequenting individ- uals. These factors further decrease the possibil- ity that many of them will molt into the adult stage before the end of the season. Their slow rate of growth will therefore relegate them to a second overwintering, which potentially entails further mortality. These spiders will, however, presumably be among the first to molt into the adult stage during the following spring and early summer and the first to lay their eggs in mid- summer. Had they realized a rate of gain in body mass proportional to that of the adults on gold- enrod (also see Morse 1981 for comparable sue- MORSE CRAB SPIDER BIOMASS 89 cess rates of adults on goldenrod), many would probably have molted into the adult stage at this time. Even this increase would, however, have consigned them to a close race to produce a clutch of eggs by the end of the season. Since I have, to date, no evidence of adult Misumena surviving over winter, increase in size and subsequent molt would have likely obviated an opportunity to reproduce. If they did succeed in laying a clutch, it, too, might be vulnerable to cold damage. If such clutches do reach the hatching stage, the success of the earliest instars is open to question. Second-instar young are extremely vulnerable to starvation, at least earlier in the season (Yogelei & Greissl 1989; Morse 1993). Further, the num- bers of adult males at this season become ex- tremely low (Holdsworth & Morse in press), so the possibility of an unfertilized dutch among late-developing adults at this season is high. Thus, in spite of the dangers of the impending winter, the penultimates5 slow gains in mass on gold- en rod may be more advantageous than they first appear. The penultimates’ slow rate of gain may even allow them to accommodate for a natural “bottleneck” of the seasons. Their failure to change sites more rapidly or to move farther than adults at these times (Table 1) does not suggest that they are responding at this time to a per- ceived food shortage in a way characteristic of most animals, including adult Misumena (Morse & Fritz 1982). Although the exact basis for molt into the adult stage is not known for Misumena , nutrition, size and growth rate are very likely to be critical en- vironmental factors in terrestrial arthropods (Blakely 1981; Forrest 1987; Nijhout 1994). The marginal size (for molt) and slow growth rate characterizing these penultimates should be strong factors prolonging the onset of molt. The results do not justify arguing at this time that the penultimates deliberately ration their rate of intake. However, as suggested by both Miyashita (1969) and Wise (1976), low avail- ability of food late in the season may facilitate the delay of maturity in species that do not over- winter as adults. Some arthropods are intrinsically pro- grammed to suspend maturation in the latter part of summer. For example, the pitcher-plant mos- quito Wyeomia smithii does not proceed into the last instar under decreasing photoperiods, which signify the return of winter conditions before the mosquito can complete its reproductive period and its offspring can reach an overwintering stage (Istock 1981). ACKNOWLEDGMENTS I thank C. Harley for reading a draft of this manuscript. Supported by the National Science Foundation (DEB80-08502-A01, IBN93- 17652). I thank E. K. Morse for assistance in the field and E. B. Noyce for permitting use of the study site. LITERATURE CITED Blakely, N. 1981. Life history significance of size- triggered metamorphosis in milkweed bugs (Onco- peltus). Ecology, 62:57-64. Forrest, T. G. 1987. Insect size tactics and devel- opmental strategies. Oecologia, 73:178-184. Fraser, D. F. & J. F Gilliam. 1992. Nonlethal im- pacts of predator invasion: facultative suppression of growth and reproduction. Ecology, 73:959-970. Fritz, R. S. & D. H. Morse. 1985. Reproductive suc- cess, growth rate and foraging decisions of the crab spider Misumena vatia. Oecologia, 65:194-200. Gertsch, W. J. 1939. A revision of the typical crab- spiders (Misumeninae) of America north of Mexico. Bull. American Mus. Nat. Hist., 76:277-442. Istock, C. A. 1981. Natural selection and life history variation: theory plus lessons from a mosquito. Pp. 1 1 3-1 27, In Insect life history patterns (R. F. Denno & H. Dingle, eds.). Springer- Verlag, New York. Kareiva, P., D. H. Morse & J. Eccleston. 1989. Sto- chastic prey arrivals and crab spider giving-up times: simulations of spider performance using two simple “rules of thumb”. Oecologia, 78:542-549. Levy, G. 1970. The life cycle of Thomisus onustus (Thomisidae: Araneae) and outlines for the classi- fication of the life histories of spiders. J. ZooL, Lon- don, 160:523-536. Lewontin, R. C. 1978. Fitness, survival, and opti- mality. Pp. 3-21, In Analysis of ecological systems (D. J. Horn, G. R. Stairs & R. D. Mitchell, eds.). Ohio State Univ. Press, Columbus. Maynard Smith, J. 1978. Optimization theory in evo- lution. Ann. Rev. Ecol. Syst, 9:31-56. Miyashita, K. 1969. Seasonal changes of population density and some characteristics of overwintering nymph of Lycosa T-insignita BOES. et STR. (Ara- neae: Lycosidae). Appl. Entomol. ZooL, 4:1-8. Morse, D. H. 1979. Prey capture by the crab spider Misumena calycina (Araneae: Thomisidae). Oec- ologia, 39:309-319. Morse, D. H. 1980. Behavioral mechanisms in ecol- ogy. Harvard Univ. Press, Cambridge, Massachu- setts. Morse, D. H. 1981. Prey capture by the crab spider Misumena vatia (L.) (Thomisidae) on three com- 90 THE JOURNAL OF ARACHNOLOGY mon native flowers. American Midi. Natur., 105: 358-367. Morse, D. H. 1986. Foraging decisions of crab spiders {Misumena vatia) hunting on inflorescences of dif- ferent quality. American Midi. Natur., 1 16:341-347. Morse, D. H. 1987. Attendance patterns, prey cap- ture, changes in mass, and survival of crab spiders Misumena vatia (Araneae, Thomisidae) guarding their nests. J. Arachnol., 15:193-204. Morse, D. H. 1988. Cues associated with patch-choice decisions by foraging crab spiders Misumena vatia. Behaviour, 107:297-313. Morse, D. H. 1993. Some determinants of dispersal by crab spiderlings. Ecology, 74:427-432. Morse, D. H. 1994. Numbers of broods produced by the crab spider Misumena vatia (Araneae, Thom- isidae). J. Arachnol, 22:195-199. Morse, D. H. & R. S. Fritz. 1982. Experimental and observational studies of patch-choice at different scales by the crab spider Misumena vatia . Ecology, 63:172-182. Morse, D. H. & R. S. Fritz. 1987. The consequences of foraging for reproductive success. Pp. 443-455, In Foraging behavior (A. C. Kamil, J. R. Krebs & H. R. Pulliam, eds.). Plenum, New York. Nijhout, H. F. 1994. Insect hormones. Princeton Univ. Press, Princeton, New Jersey. Pyke, G. H. 1984. Optimal foraging theory: a critical review. Ann. Rev. Ecol. Syst., 15:523-575. Schaefer, M. 1977. Winter ecology of spiders (Ara- neida). Z. Angew. Entomol., 83:113-144. Sih, A. 1982. Optimal patch use: variation in selec- tive pressure for efficient foraging. American Natur., 120:666-685. Skelly, D. K. & E. E. Werner. 1990. Behavioral and life-historical responses of larval American toads to an odonate predator. Ecology, 71:2313-2322. Stein, R. A. & J. J. Magnuson. 1976. Behavioral re- sponse of crayfish to a fish predator. Ecology, 57: 571-581. Stephens, D. W. & J. R. Krebs. 1986. Foraging the- ory. Princeton Univ. Press, Princeton, New Jersey. Vogelei, A. & R. Greissl. 1989. Survival strategies of the crab spider Thomisus onustus (Chelicerata, Arachnida, Thomisidae). Oecologia, 80:513-515. Wise, D. H. 1976. Variable rates of maturation of the spider, Neriene radiata ( Linyphia marginata). American Midi. Natur., 96:66-75. Manuscript received 13 February 1995, revised 31 May 1995. 1995. The Journal of Arachnology 23:91-99 REDESCRIPTION OF THE SCORPION CENTRUROIDES THORELLI KRAEPELIN (BUTHIDAE) AND DESCRIPTION OF TWO NEW SPECIES W. David Sissom: Department of Biology & Geosciences, WTAMU Box 808, West Texas A & M University, Canyon, Texas 79016 USA ABSTRACT. The Central American scorpion Centruroides thorelli Kraepelin 1891 is redescribed, based on examination of the type material and additional specimens now available. It is readily diagnosed by its smaller body size, mottled color pattern, the shape of the female pectinal basal piece, and the shape of the male telson and subaculear tubercle. Two new species related to C. thorelli are also described, one from the lowlands of Guatemala and the other from mountainous areas in southern Tamaulipas and northern San Luis Potosi in Mexico. Centruroides thorelli was described on the ba- sis of six specimens collected by Stoll in Gua- temala (Kraepelin 1891). The species was ac- cepted as valid by Pocock (1902), who men- tioned several additional specimens in the Brit- ish Museum, also originating from Guatemala. In Hoffmann’s (1932) monograph on the scor- pions of Mexico, a new series of specimens was reported from Chiapas, Mexico. The thorough description given by Hoffmann and study of the types of C. thorelli provided enough evidence to determine that his specimens are not referable to this species. Other researchers have also mis- identified the species (Ocaranza 1926; Diaz Na- jera 1966); in particular, Moreno (1939, 1940) described two subspecies from Cuba, C. t. cub- ensis and C. t. aguayoi , which are now considered synonyms of C. guanensis Franganillo. Francke & Stockwell (1987) studied specimens from Guatemala, Belize, and Costa Rica and sug- gested that at least two forms were present among their samples, but opted to place their Costa Ri- can specimens under C. thorelli by giving a very general description. The present study confirms their original suspicions by recognizing two dis- tinct species, C. thorelli from high elevations in Belize and Guatemala and a new species from lowlands in Guatemala. The Costa Rican ma- terial, which was not available, awaits further study. A second new species is described from the central Mexican states of Tamaulipas and San Luis Potosi, representing a form that is wide- ly separated geographically from other members of the complex. Whether this disjunction is real, or is the result of inadequate collecting in the intervening areas, remains to be seen. Centruroides thorelli Kraepelin Figs. 1-9 Centrums thorelli Kraepelin 1891: 124; 1899: 89-90. Centruroides thorelli Pocock 1902: 22, pi. 5, figs. 2, 2a-c; nec Hoffmann 1932: 304-307, figs. 77, 78 (misidentification); 1938: 319 (misidentification); nec Ocaranza 1926: 77 (misidentification); nec Moreno 1939: 73-74 (misidentification); nec Diaz Najera 1966: 111, 113 (misidentification); 1975: 4, 18 (mis- identification); Stahnke & Calos 1977: 1 12, 1 14 (part); Moritz & Fischer 1980: 324; Francke & Stockwell 1987: 17-18, figs. 48-56, 100 (part); nec Armas, et al. 1992a: 6 (misidentification); nec Armas 1992b: 131-132, fig. 4 (misidentification). Rhopalurus testaceus thorelli Meise 1934: 32, 34. Type data.— Adult male lectotype, 1<3 paralec- totype, 49 paralectotypes (designated by Francke & Stockwell 1986) from Guatemala, Stoll (leg.); housed in the Zoologisches Museum, Berlin (Cat. No. ZMB 7633); examined. Distribution.— Known only from Guatemala. Comparative diagnosis. — Centruroides thorelli has been a rather distinctive, but poorly under- stood, member of the genus. The original de- scription (Kraepelin 1891) and the treatment by Pocock (1902) point out its uniqueness among Centruroides by being the only species with eight rows of denticles on the cutting edge of the chela fingers that has the dorsum mottled, rather than striped. In fact, most other cuticular surfaces are also mottled. Its small size (ca. 35-40 mm), low pectinal tooth counts (less than 1 7), and the bi- lobed telson in the male have also been cited in comparing it with other Centruroides (Pocock 1902). Unfortunately, subsequent researchers have used the name inappropriately because un- 91 92 THE JOURNAL OF ARACHNOLOGY Figures 1-9.— Morphology of Centruroides thorelli Kraepelin 1891. 1-6, Lectotype male. 1, Left lateral aspect of metasomal segments III-V and telson; 2, Telson, enlarged view of left lateral aspect; 3, Telson, ventral aspect; 4, Dorsal aspect of pedipalp femur; 5, Dorsal aspect of pedipalp patella; 6, External (lateral) aspect of pedipalp chela. 7-9, Paralectotype female. 7, Left lateral aspect of metasomal segments III-V and telson; 8, Telson, enlarged view of left lateral aspect; 9, Ventral aspect of sternum, genital opercula, and pectines. described mottled forms exist in southern Mex- ico and Central America. Two of these are de- scribed below and compared to the true C. tho- relli. Description of lectotype male. — Coloration: Base color yellow to light yellow brown. Cara- pace with anterior margin infuscate and distinct dusky marbling throughout. Tergites with dusky band along posterior margins; anterior portions of tergites with diffuse dusky markings; each ter- gite bearing a narrow yellow median longitudinal line. Metasomal segments I-IV light yellow; V SISSOM - CENTR UROIDES THORELLI AND RELATIVES 93 and telson slightly darker. Cheliceral manus with dusky marbling. Pedipalps and legs pale yellow, with faint to moderate dusky markings. Venter uniformly yellowish. Prosoma: Carapace mod- erately coarsely granular; anterior median furrow deep, rounded; posterior median furrow deep, narrow near ocular tubercle and broadening pos- teriorly; carapacial carinae inconspicuous, indi- cated by medium-sized rounded granules. Me- sosoma: Median carina on tergites I— II moderate, granular; on III-VI stronger, granular to crenu- late. Pretergites minutely granular; post-tergites with large, smooth patches anterolaterally; pos- terior third of each tergite moderately coarsely granular. Tergite VII with strong, granular me- dian keel and two pairs of strong, irregularly crenulate lateral keels. Pectines: Basal piece about 2.4 times wider than long with straight posterior margin; pectinal tooth count 16-16. Stemites III- VI smooth; VII with submedian and lateral ca- rinae moderate, crenulate. Metasoma: (Fig. 1). Segments X-IV: Dorsolateral carinae on I strong, serratocrenulate; on II— III moderate, crenulate; on IV strong, crenulate. Lateral supramedian carinae on I strong, serratocrenulate; on II-IV strong, crenulate. Lateral inframedian carinae on I strong, serratocrenulate; on II-IV absent. Ven- trolateral carinae on I moderate, granular to ir- regularly crenulate; on II-IV strong, crenulate. Ventral submedian carinae on I weak to mod- erate, granular; on II-IV strong, irregularly cren- ulate. Segment V: Dorsolateral carina moderate and crenulate anteriorly, weak and granular pos- teriorly; lateromedian carina essentially obso- lete; ventrolateral and ventromedian carinae moderate, feebly crenulate. Intercarinal spaces on all segments sparsely granular. Telson: (Figs. 2, 3). Vesicle distinctly bilobed ventrodistally; aculeus downwardly deflected at junction with vesicle. Subaculear tooth strong, spinoid; its point directed towards tip of aculeus. Ventral aspect of vesicle irregularly granular. Pedipalps: Ortho- bothriotaxia A (Vachon 1974); femur with al- pha-configuration of dorsal trichobothria (Va- chon 1975). Femur: (Fig. 4). Dorsointemal and dorsoextemal carinae strong, serratocrenulate; ventrointemal carina strong, serrate; ventroex- temal carina moderate, granular anteriorly and strong, serrate posteriorly; internal and external intercarinal spaces with large coarse granules. Pa- tella: (Fig. 5). Dorsointemal carina moderate, finely serratocrenulate; dorsomedian and dor- soextemal carinae weak, granular; external ca- rina weak, more or less smooth; ventroextemal carina weak, smooth; ventrointemal carina mod- erate, irregularly serratocrenulate; internal face with weak basal tubercle and several large gran- ules. Chela: (Fig. 6). Dorsomarginal, digital, and ventroextemal carinae weak, smooth; dorsoin- temal carina obsolete except for a few coarse distal granules; other carinae essentially obsolete. Fixed finger with eight oblique rows of granules (the two basal rows are virtually fused, separated only by a tiny gap), flanked by supernumerary granules. Fixed finger trichobothrium db posi- tioned just proximal to et. Movable finger with short apical row of four granules followed by seven rows of oblique granules (actually eight rows with two basal rows fused); granular rows flanked by supernumerary granules. Morphometries: See Table 1. Measurements of lectotype male: (in mm; L = length, W = width, D = depth). Total L, 36.80; carapace L, 3.55; mesosoma L, 8.90; metasoma L, 20.95; telson L, 3.40. Metasomal segments: I LAV, 3.10/1.65; II L/W, 3.85/1 .60; III L/W, 4.25/ 1.55; IV L/W, 4.70/1.55; V L/W, 5.05/1.50. Tel- son: vesicle L/W/D, 2.45/1.30/1.20; aculeus L, 0.95. Pedipalps: femur L/W, 3.90/0.85; patella L/W, 4.05/1.20; chela L/W/D, 6.20/1.25/1.30; fixed finger L, 3.35; movable finger L, 3.90; palm (underhand) L, 2.50. Measurements of paralectotype female: Total L, 38.55; carapace L, 4.30; mesosoma L, 1 1.60; metasoma L, 18.85; telson L, 3.80. Metasomal segments: I L/W, 2.90/2.30; II L/W, 3.55/2.20; III L/W, 3.65/2.05; IV L/W, 4.15/2.10; V L/W, 4.60/2.05. Telson: vesicle L/W/D, 2.50/1.65/ 1.45; aculeus L, 1.30. Pedipalps: femur L/W, 4.10/1.20; patella L/W, 4.35/1.60; chela L/W/ D, 7.15/1.45/1.65; fixed finger L, 4.00; movable finger L, 4.65; palm (underhand) L, 2.75. Variation.— Female morphology differs from that of the male as follows: the metasomal seg- ments are proportionately shorter (Fig. 7, Table 1), but the body excluding the metasoma is larg- er, the telson vesicle is more bulbous; pectinal tooth counts are slightly lower (see below); and the metasoma is more granular. The telson is more evenly ovoid and is not distally bilobed (Fig. 8). The basal piece of the female pectine is either straight or slightly convex (Fig. 9), similar to that of the male. Male pectinal tooth counts varied as follows: 3 combs with 17 teeth, 4 combs with 16 teeth, and 9 combs with 1 5 teeth; female counts varied as follows: 4 combs with 1 5 teeth, 9 combs with 14 teeth, and 2 combs with 13 teeth (one comb 94 THE JOURNAL OF ARACHNOLOGY Table 1.— Morphometric comparisons between Centruroides thorelli Kraepelin, C, schmidti new species and C. rileyi new species. Ratios are as follows: 1 = carapace length/metasoma V length; 2 = pedipalp femur length/ width; 3 = pedipalp patella length/width; 4 = metasoma III length/width; 5 = metasoma V length/width; 6 = metasoma V length/depth; 7 = pedipalp chela width/pedipalp patella width; 8 = pectinal basal piece width/ length. Males Females thorelli schmidti rileyi thorelli schmidti rileyi Ratio (n= 7) (n= 1) (n= 1) (n = 8) (n= 1) (n = 3) 1 min 0.612 0.651 0.833 0.920 0.809 0.872 max 0.725 0.967 0.882 2 min 4.059 4.000 3.438 3.417 3.600 3.056 max 5.158 3.857 3.444 3 min 3.000 2.818 2.652 2.719 2.645 2.345 max 3.821 3.000 2.520 4 min 2.621 3.450 2.609 1.610 2.552 2.250 max 3.471 1.780 2.375 5 min 3.241 4.300 3.429 2.190 3.241 2.833 max 4.344 2.361 3.000 6 min 3.370 3.909 3.130 2.359 2.849 2.615 max 4.633 2.528 2.786 7 min 0.917 0.818 0.783 0.893 0.774 0.759 max 1.107 0.968 0.840 8 min 2.100 1.625 1.556 1.733 1.250 1.333 max 2.625 2.364 1.429 was damaged and its teeth could not be counted). In addition, there were eight unsexed first instar specimens; among these juveniles, there were 2 combs with 1 6 teeth, 5 combs with 1 5 teeth, and 9 combs with 14 teeth. Color varied somewhat among the specimens examined, with some individuals being lighter in base color with less intense fuscosity. The dif- ferences may be due to preservation, although Centruroides spp. are known to exhibit consid- erable variation in these characters. Comments. —Hoffmann’s (1932) specimens of C. thorelli from Tuxtla Gutierrez, Chiapas and Francke & Stockwell’s (1987) specimens from Costa Rica are not referable to C. thorelli. The latter authors noted significant differences be- tween their specimens and the type series but opted not to describe their form as new. These specimens were not available for study. Although Armas (1992) reported only a small juvenile from Quintana Roo, I have been able to examine a mature female collected recently; as Annas has suggested (pers. comm.), it does not appear to be referable to C. thorelli. Additional specimens examined. —GUATEMALA: Chichivac, near Tecpan (P. J. W. Schimdt, Leon Man- del Guatemala Expedition), 1 Feb 1934, 1 <3, 1$, 8 first instars (FMNH); Finca San Rafael, Sacatepequez, elev. 6900 ft., under bark (R. D. Mitchell), 24 June 1948, IS (FMNH), 25 June 1948, IS, 19, 1 juv male (WDS), 29 June 1948, 2s, 2$ (FMNH). Centruroides schmidti new species (Figs. 10-18) Type data.— Adult holotype male “found on bones of crocodile skull” at Lake Tickamaya, Honduras on 26 April 1923 by K. Schmidt and L. Walters (Capt. Field Mus. Exped.); perma- nently deposited in the Field Museum of Natural History, Chicago. Etymology.— The specific epithet is a patron- ym honoring Dr. K. P. Schmidt, collector of the type specimens, for his many years of service to arthropod systematics at the Field Museum of Natural History. Distribution.— Known only from Honduras and eastern Guatemala. Comparative diagnosis. — Centruroides schmidti is similar to C. thorelli , but differs in the follow- ing characters: the mottling of the carapace, ter- gites, and legs is weaker; the distal segments of the metasoma and telson are much darker than the preceding segments; the male telson is not distally bilobed as in C thorelli ; the male sub- SISSOM— CENTR UROIDES THORELLI AND RELATIVES 95 Figures 10-1 8. — Morphology of Centruroides schmidti new species. 10-15, Holotype male. 10, Left lateral aspect of metasomal segments III-V and telson; 11, Telson, enlarged view of left lateral aspect; 12, Telson, ventral aspect; 13, dorsal aspect of pedipalp femur; 14, Dorsal aspect of pedipalp patella; 15, External (lateral) aspect of pedipalp chela. 1 6-18, Paratype female. 1 6, Left lateral aspect of metasomal segments III-V and telson; 17, Telson, enlarged view of left lateral aspect; 18, Ventral aspect of sternum, genital opercula, and pectines. aculear tubercle arises from a crenulated mid- ventral carina and is angular, rather than a single spinoid tooth; metasomal segment V in the male is essentially acarinate, rather with the keels well developed; chela fixed finger trichobothrium db is slightly distal to et (rather than proximal to it); the basal piece of the female pectines is pro- duced distally into a rounded lobe, rather than being straight or slightly convex; and there are several distinct morphometric differences (see Table 1). Description of holotype male. — Coloration: Base color light yellow brown above with faint to moderate dusky markings on dorsum, chelic- erae, pedipalps, legs, and stemites. Coloration fairly uniform except as follows: coxostemal re- gion light yellow, pectines very pale yellow, metasomal segment V and telson dark orange to reddish brown; cheliceral teeth and tip of aculeus dark reddish brown. Prosoma : Carapace mod- erately coarsely granular; anterior median furrow moderately deep; posterior median furrow shal- 96 THE JOURNAL OF ARACHNOLOGY low anteriorly, deeper posteriorly; carapacial ca- rinae inconspicuous, indicated by lines of small granules. Mesosoma: Median carina on I-YI moderate, granular. Pretergites minutely granu- lar, post-tergites moderately, coarsely granular throughout. Tergite YII with moderate, granular median keel and two pairs moderate, finely ser- rated lateral keels. Pectinal basal piece 1.8 times wider than long; posterior margin distinctly rounded; pectinal tooth count 15-15. Stemites III- VI essentially smooth, with some fine gran- ulation on VI; VII with submedian and lateral carinae moderate, finely serrate. Metasoma : (Fig. 10). Segments I-IV: Dorsolateral carinae on I- II moderate, finely serrate; on III moderate, cren- ulate; on IV weak, smooth. Lateral supramedian carinae on I— III moderate, finely serrate; on IV weak, smooth. Lateral inframedian carina on I moderate, serrate; on II-IV absent. Ventrolateral carinae on I— III moderate, finely serrate; on IV moderate, feebly granular. Ventral submedian carinae on I— III weak, finely serrate; on IV weak, feebly serrate. Intercarinal spaces shagreened. Segment V: Acarinate, intercarinal spaces sha- greened. Telson: (Figs. 11, 12). Vesicle elongate oval in shape with gently rounded dorsal margin; ventral aspect with row of small granules leading to subaculear tubercle; subaculear tubercle nar- row, but angular in lateral view, its point directed towards middle of aculeus. Ventral aspect of ves- icle shagreened. Pedipalps: Orthobothriotaxia A (Vachon 1974); femur with alpha-configuration of dorsal trichobothria (Vachon 1975). Femur: (Fig. 13). Dorsointemal, dorsoextemal, and ven- trointemal carinae strong, serrate; ventroexter- nal carina weak, smooth basally, crenulate dis- tally; internal face with serrated keel flanked by accessory granules; dorsal face moderately gran- ular. Patella: (Fig. 14). Ventrointemal carina strong, serrate; dorsointemal, dorsomedian, dor- soextemal, and ventroextemal carinae moderate, finely serrate; external carina moderate, feebly serrate. Inner face with seven larger, sharp, sub- conical granules. Chela: (Fig. 1 5). Dorsomarginal carina moderate, finely serrate; dorsal secondary carina moderate, granular to feebly crenulate; digital carina weak, smooth; external secondary carina weak, smooth; ventroextemal carina moderate, feebly granular; dorsointemal and ventrointemal carinae with relatively large, ser- rate granules. Fixed finger with eight oblique rows of granules, these flanked by supernumerary granules. Fixed finger trichobothrium db posi- tioned just distal to et. Movable finger with short row of four apical granules followed by eight oblique rows of granules; granular rows flanked by supernumerary granules. Morphometries, —See Table 1. Measurements of holotype male.— (in mm, L = length, W = width, D = depth). Total L ap- proximately 32 mm (tip of aculeus of telson bro- ken, rendering total length an estimate); carapace L, 2.80; mesosoma L, 8.10; metasoma L, 17.20; telson L, ? Metasomal segments: I L/W, 2.50/ 1.15; II L/W, 3.05/1.05; III L/W, 3.45/1.00; IV L/W, 3.90/1.00; V L/W, 4.30/1.00. Telson: ves- icle L/W/D, 1.75/080/0.85. Pedipalps: femur L/W, 2.80/0.70; patella L/W, 3.10/1.10; chela L/W/D, 4.70/0.90/0.95; fixed finger L, 2.85; movable finger L, 3.20; palm (underhand) L, 1 .65. Measurements of paratype female. —Total L, 37.35; carapace L, 3.80; mesosoma L, 1 1.20; me- tasoma L, 18.95; telson L, 3.40. Metasomal seg- ments: I L/W, 2.90/1.55; II L/W, 3.45/1.45; III L/W, 3.70/1.45; IV L/W, 4.20/1.45; V L/W, 4.70/ 1.45. Telson: vesicle L/W/D, 1.90/1.10/1.25; aculeus L, 1.50. Pedipalps: femur L/W, 3.60/1.00; patella L/W, 4.10/1.55; chela L/W/D, 6.30/1.20/ 1 .35; fixed finger L, 3.90; movable finger L, 4.50; palm (underhand) L, 2.10. Variation.— Only the holotype male and para- type female were available for study. The female differs from the male in the following characters: the metasomal segments are not as elongate (Fig. 1 6), and metasoma V bears well developed cren- ulated ventrolateral and ventromedian carinae; the telson is slightly more globose (Fig. 1 7); and the basal piece of the pectines is produced distally into a large, rounded lobe (Fig. 1 8). Pectinal tooth counts in the male and female were similar: the male count was 15-15 and the female count 1 5- 14. Paratypes.— GUATEMALA: Escobas , Iza- bal, 27 November 1 933 (K. P. & P. J. W. Schmidt, Leon Mandel Guatemala Exped.), Iv (FMNH). Centruroides rileyi new species Figs. 19-27 Type data.— Adult holotype male and adult paratype female from Bocatoma (= 7 km SSE Gomez Farias), Tamaulipas, Mexico on 25-30 March 1978 by E. G. Riley; permanently de- posited in the collection of the United States Na- tional Museum (Smithsonian), Washington, D. C. Etymology.— The specific name is a patronym honoring Dr. Edward Riley of Texas A & M University, the collector of the holotype. SISSOM — CENTR UROIDES THORELLI AND RELATIVES 97 Figures 19-27. —Morphology of Centruroides rileyi new species. 19-24, Holotype male. 19, Left lateral aspect of metasomal segments III-V and telson; 20, Telson, enlarged view of left lateral aspect; 21, Telson, ventral aspect; 22, Dorsal aspect of pedipalp femur; 23, Dorsal aspect of pedipalp patella; 24, External (lateral) aspect of pedipalp chela. 25-25, Paratype female. 25, Left lateral aspect of metasomal segments III-V and telson; 26, Telson, enlarged view of left lateral aspect; 27, Ventral aspect of sternum, genital opercula, and pectines. Distribution.— Known from several localities in southern Tamaulipas and northern San Luis Potosi, Mexico. Comparative diagnosis. — Centruroides rileyi is most similar to C. schmidti , but is clearly related to C. thorelli as well. Unlike the other two spe- cies, sexual dimorphism in the length of the metasomal segments is not as pronounced; male segments are more slender but not noticeably longer than those of the female. Based on the material at hand, C. rileyi exhibits smaller body size than either C. thorelli or C. schmidti. The shape of the subaculear tubercle is similar to that of C. schmidti, developed from a distinct ven- tromedian keel. Likewise the basal piece of the female pectines is produced distally into a round- ed lobe, a feature seen in C. schmidti but not in C. thorelli. Among specimens available, female pectinal tooth counts were slightly lower than in C. schmidti, with a range of 1 1-13 (13-15 in C. thorelli and 14-15 in C. schmidti). Metasomal segment V in the male of C. rileyi bears distinct 98 THE JOURNAL OF ARACHNOLOGY (but weak) crenulate carinae; these carinae are developed in C. thorelli, but not in C. schmidti. The dorsal carinae of the pedipalp chelae are moderately to strongly developed and distinctly crenulate, but are feeble and smooth to granular in C. thorelli and only the dorsal marginal carina is noticeably crenulated in C. schmidti. Mor- phometric comparisons are provided in Table 1 . Description of holotype male. — Coloration: Base color yellow to light yellow brown. Cara- pace with distinct dusky marbling concentrated mostly in median area. Tergites with distinct, regular pattern of blackish spots. Metasomal seg- ments I-IV light yellow, moderately infuscate; V and telson more heavily infuscate, appearing darker than preceding segments. Cheliceral ma- nus with strong dusky marbling. Pedipalps and legs yellow, with distinct dusky markings. Venter uniformly yellowish anteriorly; stemites lightly infuscate. Prosoma : Carapace moderately coarsely granular; anterior median furrow mod- erately deep; posterior median furrow shallow anteriorly, deeper posteriorly; carapacial carinae weak, indicated by lines of small granules. Me sosoma: Median carina on tergites I-IV mod- erate, granular; on V-VI moderate, granular to crenulate. Pretergites minutely granular, post- tergites with large, smooth patches anterolater- ally; posterior third of each tergite moderately, coarsely granular. Tergite VII with moderate, granular median keel and two pairs strong, finely serrated lateral keels. Pectinal basal piece 1.6 times wider than long; posterior margin distinct- ly rounded; pectinal tooth count 14-14. Stemites III-VI essentially smooth; VII with submedian carinae weak, granular and lateral carinae mod- erate, finely serrate. Metasoma: (Fig. 19). Seg- ments I-IV: Dorsolateral carinae on I— II strong, finely serrate; on III strong, irregularly crenulate; on IV moderate, crenulate. Lateral supramedian carinae on I— III strong, finely serrate; on IV strong, feebly serrate. Lateral inframedian carina on I strong, finely serrate; on II-IV absent. Ventro- lateral carinae on I— III strong, finely serrate; on IV strong, feebly serrate. Ventral submedian ca- rinae moderate, feebly serrate; on IV weak, feebly serrate. Segment V: Dorsolateral carina moder- ate, feebly granular; lateromedian carina obso- lete; ventrolateral and ventromedian carinae moderate, feebly serrate. All metasomal inter- carinal spaces sparsely granular. Telson: (Fig. 20- 21). Vesicle elongate oval in shape with aculeus moderately deflected downward; ventral aspect with median longitudinal row of small granules leading to subaculear tubercle; subaculear tuber- cle narrow, but angular in lateral view, its point directed towards middle of aculeus. Ventral as- pect of vesicle lightly granular. Pedipalps: Or- thobothriotaxia A (Vachon 1974); femur with alpha-configuration of dorsal trichobothria (Va- chon 1975). Femur: (Fig. 22). All carinae strong, serrate; internal face with series of large serrate granules; dorsal face moderately granular. Pa- tella: (Fig. 23). Dorsointemal carina moderate, serrate; dorsomedian and dorsoextemal carinae strong, serratocrenulate; external carina strong, feebly crenulate; ventroextemal carina moder- ate, crenulate; ventrointemal carina moderate, irregularly serrate. Inner face with eight to ten larger, sharp, subconical granules. Chela: (Fig. 24). Dorsomarginal carina strong, coarsely cren- ulate; dorsal secondary carina strong, finely ser- rate; digital and external secondary carinae mod- erate, finely crenulate; ventroextemal carina strong, finely crenulate; ventrointemal carina moderate with a few rounded granules; dorsoin- temal carina strong, coarsely serrate. Fixed finger with eight oblique rows of granules flanked by supernumerary granules. Fixed finger tricho- bothrium db positioned just distal to et. Movable finger with short apical row of four granules fol- lowed by eight oblique rows of granules; granular rows flanked by supernumerary granules. Morphometries.— See Table 1. Measurements of holotype male.— (in mm; L = length, W = width, D = depth). Total L, 29.75; carapace L, 3.00; mesosoma L, 9.10; metasoma L, 15.15; telson L, 2.50. Metasomal segments: I L/W, 2.35/1.30; II L/W, 2.80/1.20; III L/W, 3.00/ 1 . 1 5; IV L/W, 3.40/ 1 .05; V L/W, 3.60/1 .05. Teh son: vesicle L/W/D, 1.50/0.80/0.90; aculeus L, 1.00. Pedipalps: femur L/W, 2.75/0.80; patella L/W, 3.05/1.15; chela L/W/D, 4.75/0.90/1.05; fixed finger L, 2.85; movable finger L, 3.25; palm (underhand) L, 1.65. Measurements of female paratype — Total L, 29.85; carapace L, 3.15; mesosoma L, 9.40; me- tasoma L, 14.65; telson L, 2.65. Metasomal seg- ments: I L/W, 2.25/1.40; II L/W, 2.65/1.25; III L/W, 2.85/1.20; IV L/W, 3.30/1.20; V L/W, 3.60/ 1.20. Telson: vesicle L/W/D, 1.45/0.85/0.95; aculeus L, 1 .20. Pedipalps: femur L/W, 2.75/0.85; patella L/W, 3.15/1.25; chela L/W/D, 4.90/1 .05/ 1.10; fixed finger L, 3.00; movable finger L, 3.45; palm (underhand) L, 1.60. Variation. —Only a single adult male is known; however, there are three adult female specimens and a juvenile. Females differ from the males SISSOM - CENTR UROIDES THORELLI AND RELATIVES 99 only slightly in metasomal morphometries with the male metasoma being scarcely longer, but noticeably thinner (Fig. 25). There is also little difference in the shape of the telson (Fig. 26). Finally, the female pectinal basal piece is pro- duced distally into a distinct, rounded lobe (Fig. 27). Female pectinal tooth counts were as fol- lows: there were one comb with 1 1 teeth, four combs with 12 teeth, and one combs with 13 teeth. Paratypes.— MEXICO: San Luis Potosi, 5 km N Tamazunchale off Hwy 85, 1 August 1987 (J. A. Nilsson), 19 (JAN). Tamaulipas , Gomez Far- ias, 16 March 1977 (R. Schmidt), 19 (FSCA). ACKNOWLEDGMENTS I am extremely grateful to Manfred Moritz of the Zoologisches Museum in Berlin for allowing me to examine the type series of Centruroides thorelli. Additional material used in the descrip- tions was made available through the courtesy of Dan Summers of the Field Museum of Natural History, Chicago (FMNH); Jan Nilsson of Northern Arizona University (JAN); Edward Riley of Texas A & M University; and Douglas Rossman of Louisiana State University, whose specimens are now deposited in the Florida State Collection of Arthropods (FSCA). Luis F. de Ar- mas (Instituto de Ecologia y Sistematica, Aca- demia de Ciencias de Cuba) allowed me to ex- amine his specimen from Quintana Roo as a loan arranged through Norm Platnick of the Ameri- can Museum of Natural History, New York; I am grateful to them both for this courtesy. Lastly, page charges were paid by the Department of Biology & Geosciences of West Texas A & M University, and I am ever grateful for the sup- port. LITERATURE CITED Armas, L. F. de. 1992. Scorpiones y solpugida (Arachnida) de los Reserva de la Biosfera de Sian Ka’an, Quintana Roo. Pp. 129-137 In Nararro L., D. & E. S. Morales (eds.), Diversidad Biologica en la Reserva de la Biosfera de Sian Ka’an Quintana Roo, Vol. II. Centro de Investigaciones de Quintana Roo, Chetumal. Armas, L. F. de, D. Navarro L., & R. M. Medrano. 1992. Apuntes para el estudio de los alacranes (Arachnida: Scorpiones) de Quintana Roo. Ava- Cient. 3:3-7. Diaz Najera, A. 1966. Alacranes de la Republica Mexicana. Clave para identificar especies de Cen- trums (Scorpionida, Buthidae). Rev. Inst. Salubr. Enferm. Trop., Mexico, 26:109-129. Diaz Najera, A. 1975. Listas y datos de distribution geografica de los alacranes de Mexico (Scorpionida). Rev. Inst. Salubr. Enferm. Trop., Mexico, 35:1-36. Francke, O. F. & S. A. Stockwell. 1987. Scorpions (Arachnida) from Costa Rica. Spec. Publ. Mus. Tex- as Tech Univ., 25:1-64. Hoffmann, C. C. 1932. Los Scorpiones de Mexico. Segunda parte. Buthidae. An. Inst. Biol., Mexico, 2:243-361. Hoffmann, C. C. 1938. Nuevas consideraciones ac- erca de los alacranes de Mexico. An. Inst. Biol, Mexico, 9:317-337. Kraepelin, K. 1891. Revision der Skorpione. I. Die Familie der Androctonidae. Jahrb. Hamburg Wiss. Anst., 8:1-144. Meise, W. 1934. Scorpiones. Nytt. Mag. Naturvi- densk., Oslo, 72:25-43. Moreno, A. 1939. Contribution al estudio de los es- corpionidos cubanos. II. Superfamilia Buthoidea. Mem. Soc. Cubana Hist. Nat., 13:63-75. Moreno, A. 1940. Scorpiologia cubana. Rev. Univ. Havana, Cuba, 26/27:91-113. Moritz, M. & S.-C. Fischer. 1980. Die Typen der Arachniden-Sammlung des Zoologischen Museums Berlin. III. Scorpiones. Mitt. Zool. Mus. Berlin., 56: 309-326. Ocaranza, F. 1926. Estudio experimental acerca de la ponzona de los alacranes en Mexico. Cuarta Me- moria - Alacran de Sonora ( Centrums thorelli) rata blanca. Rev. Mexicana Biol., 6:77-80. Pocock, R. 1902. Arachnida: Scorpiones, Pedipalpi, and Solifugae. Pp. 1-71 In Biologia Centrali-Amer- icana, London. Stahnke, H. L. & M. Calos. 1977. A key to the species of the genus Centruroides Marx (Scorpionida: Buth- idae). Ent. News, 88:1 1 1-120. Vachon, M. 1974. Etude des caractties utilises pour classer les families et les genres de Scorpions (Ar- achnides). Bull. Mus. Natn. d’Hist. Nat. (Paris), ser. 3, 104:857-958. Vachon, M. 1975. Sur l’utilisation de la trichoboth- riotaxie du bras des pedipalpes des scorpions (Ar- achnides) dans le classement des genres de la familie des Buthidae Simon. C. R. Acad. Sci. (Paris), ser. D, 281:1597-1599. Manuscript received 15 February 1994, revised 9 May 1995. 1995. The Journal of Arachnology 23:100-1 10 DISTRIBUTIONS OF THE SCORPIONS CENTRUROIDES VITTATUS (SAY) AND CENTRUROIDES HENTZI (BANKS) IN THE UNITED STATES AND MEXICO (SCORPIONES, BUTHIDAE) Rowland M. Shelley: North Carolina State Museum of Natural Sciences, P. O. Box 29555, Raleigh, North Carolina 27626-0555 USA W. David Sissom: Department of Biology & Geosciences, West Texas A & M University, P. O. Box 808, Canyon, Texas 79016-0001 USA ABSTRACT. Specific locality records are presented to define the distributions of the scorpions Centruroides vittatus (Say) and C. hentzi (Banks) in North America. The former occurs in the Central Plains as far north as Thayer County, Nebraska; the Rio Grande and Sangre de Cristo Mountains form the western distributional boundary, and the Missouri and Mississippi Rivers essentially do likewise on the east. Centruroides vittatus occurs just across the latter watercourses in Holt County, Missouri, and Monroe and Randolph counties, Illinois, range extensions that probably can be attributed to rafting or natural alterations in the rivers’ courses. Other occurrences east of the Mississippi River, in northern Illinois, Kentucky, Tennessee, Louisiana, Mississippi and North Carolina, are associated with cities and are mostly far outside what we consider the natural range; such records are regarded as human introductions. One of these apparently represents a viable reproducing population in Rutherford County, Tennessee. Likewise, records far west of the Rio Grande, in Arizona and California, are interpreted as introductions. Centruroides vittatus traverses the Rio Grande south of Texas and occurs in Chihuahua, Coahuila, Nuevo Leon, and Tamaulipas, Mexico. Centruroides hentzi, previously known only from Florida in the United States, occurs in Mobile and Baldwin counties, Alabama, and in the southern tier of counties in Georgia. Occurrences of C. hentzi in Durham, Carteret, and Brunswick counties, North Carolina, Charleston County, South Carolina, and Harris and Muscogee counties, Georgia, are considered to represent accidental human importations, although it is also possible that the more proximal ones are peripheral isolates. The scorpion fauna of the United States east of the Mississippi River is depauperate in com- parison to that of the southwest. According to Muma (1967), five species - Tityus floridanus Banks, Isometrus maculatus (DeGeer), Centru- roides gracilis (Latreille), C. hentzi (Banks), and C. keysi Muma - occur in Florida. Presently, T. floridanus is a synonym of T. dasyurus Pocock, from Puerto Rico and the Virgin Islands (Lour- engo & Francke 1984), and C. keysi is considered a synonym of C. guanensis Frangillo, from Cuba, Hispaniola, and the Bahamas (de Armas 1981). The first two species are known in Florida only from single individuals ostensibly collected at Key West. These records are questionable, and C. gracilis , hentzi and guanensis are the only scorpions that will be encountered frequently in the state, if not the only ones actually occurring there. Elsewhere in the East, the only known na- tive scorpion is Vaejovis carolinianus (Beauvois), an upland species occurring primarily north and west of the Fall Zone as far north as the Ohio River in central Kentucky (Shelley 1994a). Say (1821) described “ Buthus vittatus ” from the “sea islands” of Georgia, but his type spec- imen^) are lost, and this name, long associated with the common midwestem species of Centru- roides, was formally assigned to it (Opinion 1680, 1992) in response to the petition by Stockwell & Levi (1989), as subsequently modified by re- spondent comments (Gentry et al. 1 99 1). As part of this opinion, a neotype of B. vittatus was des- ignated from Kinney County, Texas, instead of Georgia. Say’s locality record plausibly refers to C. hentzi, which occurs statewide in Florida (Muma 1967), but up to now is not supported by preserved specimens. Centruroides hentzi and vittatus have been introduced into North Caro- lina (Shelley 1 994b), and newspaper articles have reported scorpions from Kiawah Island and Isle of Palms, near Charleston, South Carolina (Langley 1991, 1994). These two scorpions are 100 SHELLEY & SISSOM -CENTRUROIDES VITTATUS & C HENTZ1 101 Figures 1-3.— Comparisons between Centruroides vittatus and C. hentzi in color patttem. 1, 2, Dorsal views of carapace and chelicerae of C. vittatus ; 3, Same for C hentzi. readily distinguished at any life stage by the char- acters in Table 1, which also serve to distinguish C. vittatus from C. guanensis (= C. key si). Dif- ferences in the pigmentations of the chelicerae and carapaces, and in the configurations of the telsons, are shown in Figs. 1-5. While recently examining museum specimens, we encountered samples of C. hentzi from south- ern Alabama and Georgia, thus establishing its occurrence north and west of Florida. The mu- seum holdings also included numerous new re- cords of C. vittatus that enable a detailed de- scription of its distribution. This distribution has been generally described as Louisiana west of the Mississippi River to New Mexico east of the Rio Grande, and from the Central Plains of the Unit- ed States to northern Mexico (Stahnke & Calos 1977; Stockwell & Levi 1989; Shelley 1994b). However, it is striking to note that, aside from Las Vegas, San Miguel County, New Mexico (Banks 1901); Cleveland, Garvin, and Seminole counties, Oklahoma (cited as Centrums caroli- nianus by Banks et al. 1932); the Wichita Moun- tains, Comanche County, Oklahoma (Coken- dolpher & Bryce 1980); and Thayer County, Ne- braska (Rapp 1987), few definite United States records exist outside of Texas, where C vittatus can be anticipated statewide with perhaps the exception of several southeastern coastal coun- ties. Published Mexican records (Hoffmann 1932; Diaz Najera 1975) are as follows: Cd. Juarez in Chihuahua; Cd. Acuna, Allende, Cuatro Ciene- gas, Lamadrid, and Sacramento in Coahuila; Hi- dalgo in Nuevo Leon; and Barrotal, Cd. Aleman, Guerrero, Matamoros, and San Fernando in Ta- maulipas. It is our purpose here to place the new samples on record, update the known distributions of these scorpions in the United States and Mexico, and provide additional habitat information based on notations on vial labels. Acronyms for sources of preserved material are as follows: AMNH - American Museum of Natural History, New York, New York; ANSP - Academy of Natural Sciences, Philadelphia, Pennsylvania; CAS - Cal- ifornia Academy of Sciences, San Francisco; CC - Biology Department, Columbus College, Co- lumbus, Georgia; CIM - Cumberland Island Mu- seum, St. Marys, Georgia; FMNH - Field Mu- seum of Natural History, Chicago, Illinois; FSCA - Florida State Collection of Arthropods, Gaines- ville; INKS - Illinois Natural History Survey, Champaign; LSU - Entomology Department, Louisiana State University, Baton Rouge; MCZ - Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts; MEM - Mississippi Entomological Museum, Mississippi State University, Starkville; MMNS - Mississip- pi Museum of Natural Science, Jackson; MPM - Milwaukee Public Museum, Milwaukee, Wis- consin; MWSU - Midwestern State Univ., Wich- ita Falls, TX; NCSM - North Carolina State Mu- seum of Natural Sciences, Raleigh; NCSU - En- tomology Department, North Carolina State University, Raleigh; NMNH - National Museum of Natural History, Smithsonian Institution, Washington, DC; OKSU - Emerson Entomolog- ical Museum, Oklahoma State University, Still- 102 THE JOURNAL OF ARACHNOLOGY Figures 4-5.— Differences in telson morphology between Centruroides vittatus and C. hentzi. 4, Lateral view of telson of C. vittatus; 5, Same for C hentzi. water; OMNH - Oklahoma Museum of Natural History, University of Oklahoma, Norman; PMNH - Peabody Museum of Natural History, Yale University, New Haven, Connecticut; RNH - Private collection of R. N. Henson, Boone, North Carolina; SEM - Snow Entomological Mu- seum, University of Kansas, Lawrence; SFASU - Biology Department, Stephen F. Austin Uni- versity, Nacogdoches, Texas; TAMU - Texas A & M University, College Station; TMM - Texas Memorial Museum, University of Texas at Aus- tin; TY - private collection of T. Yamashita; UCO - University of Colorado Museum, Boulder; UGA - University of Georgia Museum of Natural His- tor, Athens; UMN - Entomology Department, University of Minnesota, St. Paul; UMO - Enns Entomological Museum, University of Missouri, Columbia; UTEP - Biology Department, Uni- versity of Texas at El Paso; WDS - Private col- lection of W. D. Sissom, Canyon, Texas; WFR - Private collection of W. F. Rapp, Crete, Ne- braska; WTAMU - Department of Biology and Geosciences, West Texas A&M University, Can- yon, Texas. Centruroides vittatus Figs. 1, 2, 4, 6, 7 Habitat.— As reported by Shelley (1994b) and recorded through personal observations, C. vit- tatus occupies a variety of microhabitats in de- serts, deciduous and pine forests, and grasslands, inhabiting crevices of rocky outcrops, canyon walls, and volcanic hills, climbing into vegeta- tion, seeking refuge beneath yuccas and in trash dumps, and commonly entering houses. It has been collected from sea level to elevations of over 1 800 m in the Guadalupe and Chisos Mountains, Texas and 2340 m in mountains of Coahuila, Mexico. Additional microhabitats cited on labels with the present samples include under palm branches, rocks, bark and logs in a pine forest, cow dung, and old rags and debris at an aban- doned campsite; in a sabal palmetto grove; in the nest of a cactus rat; and in a molasses trap left overnight. Specimens were found in homes, motels, dormitories, and office buildings in Cole, St. Louis, and Taney counties, Missouri; Orleans Parish, Louisiana; Alfalfa, Kay, Marshall, Mus- kogee, Pawnee, Payne, and Stephens counties, Oklahoma; and DeBaca and Eddy counties, New Mexico. An individual from San Miguel County, New Mexico, was encountered inside the Las Vegas hospital. Distribution. — Centruroides vittatus has offi- cially been recorded from only 1 6 Texas counties - Andrews, Brewster, DeWitt, Edwards, Erath, Garza, Hall, Kinney, Lubbock, Mason, Parker, Travis, Uvalde, Val Verde, Williamson, and Wise (Reddell 1965, 1970; Rowland & Reddell 1976; Stockwell & Levi 1989; Formanowicz & Shaffer 1993). Although overlooking a number of sig- nificant collections, Stockwell (1986) reported it from 50 additional counties in an unpublished Master’s Thesis: Archer, Bexar, Blanco, Cam- eron, Clay, Coke, Crockett, Crosby, Culberson, Dallas, Foard, Gillespie, Gonzales, Grayson, Hi- dalgo, Jeff Davis, Johnson, Kaufman, Kent, Kerr, Kimble, King, Knox, LaSalle, Maverick, McMullen, Medina, Menard, Motley, Navarro, Oldham, Pecos, Presidio, Real, Reeves, San Pa- tricio, San Saba, Schleicher, Starr, Sutton, Tay- lor, Terrell, Tom Green, Victoria, Ward, Webb, Wichita, Winkler, Zapata, and Zavala. These re- cords are scattered across the state, and the scor- pion is now known from bordering states in the United States and Mexico in all directions. We therefore believe that C. vittatus can be antici- pated in every Texas county except perhaps Or- ange, Jefferson and Chambers, along the Gulf Coast east of Galveston Bay. Present records from Louisiana do not support its occurrence in this comer of Texas, and field collecting is needed in these counties and in Calcasieu and Cameron Parishes, Louisiana, to confirm or refute this SHELLEY & SISSOM— CENTR UROIDES VITTATUS & C HENTZI 103 Figure 6. —Distributions of C. vittatus (closed circles) and C. hentzi (open circles) in the United States and Mexico. Florida records for C. hentzi are for counties only and are partially based on Muma (1967). The approximate courses of the Missouri River and the Rio Grande are indicated in Missouri and New Mexico/ Colorado, respectively. The “?” in Iowa and Louisiana denotes the records from unknown counties along the Missouri and Mississippi Rivers, respectively. The “?” in Chihuahua, Mexico reflects our lack of knowledge on the extent of the distribution of C. vittatus in this region, which has been so poorly sampled. finding. Large numbers of samples from Texas exist in many museum collections, too many to be shipped for examination. We therefore list only additional Texas county records that were intermingled with material from other states, but plot all known localities accurately on Fig. 6. Determining the natural distribution of C. vit- tatus, and to a lesser extent C hentzi , is ham- pered by the number of specimens that man has accidentally transported into new areas, which tend to mask the indigenous range. Samples from distant states like California and North Carolina clearly represent human introductions, but ones from proximate sites like Memphis, Tennessee and Baton Rouge, Louisiana could plausibly re- flect peripheral native populations. Mapping of all the samples, however, reveals clusters of re- cords that we believe represent natural occur- rence; we use them as the basis for determining indigenous distributions, particularly when de- tached records are from urban environments and are consistent with human activities. Thus, as shown in Fig. 7, the eastern border formed by clustered records angles southwestward through southeastern Missouri into Arkansas, and then runs through central Arkansas and Louisiana be- fore turning westward into Texas, omitting the adjoining coastal comers of Louisiana and Tex- as. The only outlying records along this boundary are from the urban environments of Baton Rouge and Memphis, and are therefore treated as hu- man introductions. The overall distribution of C. vittatus (Fig. 6) extends southward from Thayer County, Ne- braska, and expands longitudinally to encompass all of Oklahoma, Arkansas, and Missouri south of the Missouri River. The Rio Grande in south- ern New Mexico and the Sangre de Cristo Moun- tains in northern New Mexico and south central Colorado form the western geographical bound- ary, and present Colorado records suggest west- ward expansion through the Arkansas River Val- ley. The only available records from the western half of Kansas are sight records from Clark and Trego counties, so collecting is needed to deter- mine the distribution in this part of the state. The Missouri and Mississippi Rivers essentially form distributional boundaries, as the only nat- ural occurrences to the north/east of the former, in Holt County, Missouri, and to the east of the 104 THE JOURNAL OF ARACHNOLOGY latter, in Monroe and Randolph counties, Illi- nois, are in bordering counties that could have resulted from rafting or natural alterations of the rivers’ courses; the Illinois records are directly across the Mississippi from an area where the scorpion is common in Missouri. However, oth- er sites east of the Mississippi, mostly urban ar- eas, represent obvious adventives. The range tra- verses the Rio Grande south of Texas, and C vittatus is known from over half the lengths of Tamaulipas, Nuevo Leon, and Coahuila, and the northern periphery of Chihuahua. NEW RECORDS Specimens that are considered to represent the native distribution were examined from the fol- lowing localities. Missing data (exact locality, date of collection, and collector(s)) are not reported. Sight records deemed reliable from additional counties are presented separately for each state after the locality listings but are plotted in Fig. 6. USA: ARKANSAS: Benton County , Pea Ridge Natl Battlefield, 23 April 1965, J D. Unzicker (INHS). Boone County , Harrison, Conard Fissure, 8 June 1932, F. D. Wood (PMNH). Cleburne Co., 9.6 km SSW Drasco, September 1979, D. Pearson (MPM). Crawford Coun- ty, 14.4 km N Mountainsburg, Boston Mts., 30 June 1 955, T. J. Cohn (AMNH); and Lee Creek, 9 July 1 968, R. & A. Graves (FSCA). Faulkner County , 12.8 km N Camp Robinson, 2 1 April 1 943, D. D. Davis (FMNH); Camp Robinson, 24 March 1943, R. C Ellis (FMNH). Franklin County , 6.4 km N Ozark, along AR hwy, 23, 12 October 1963 and 15 September 1964, Unzicker, Yamamoto, Rotramel (INHS). Garland County , 17.6 km W Hot Springs, 21 March 1958, K. P Schmidt. Izard County , May 1954, H. M. Bevel (MCZ). Logan County , Mt. Magazine, 21 June 1938, J. M. Schmidt (FMNH), 13 August 1966 and 13 April 1976, L D. Newsome (LSU), and 15 July 1949, M. W. Sanderson, Stannard (INHS); 1.6 km S lone, along AR hwy. 23, 15 September 1964, J. D. Unzicker (INHS); and 16 km S Booneville, along AR hwy. 23, 12 October 1963, Unzicker, Yamamoto (INHS). Marion County , 6.4 km SW Oakland, Ozark Isle Park, Bull Shoals L., 28 July 1970, P. J. Clausen (FSCA). Nevada County, 30 De- cember 1954, N. B. Causey (MCZ). Newton County , Buffalo National River, near Pruitt, no date, T. Ya- mashita (TV). Perry County , Williams Jet., Ouachita Mts., 19 July 1968, E. N. K. Waering (FSCA). Pike County , Kirby, 9 June 1970 (WTAMU); and Glen- wood, 23 July 1973 (WTAMU). Polk County , Mena, 10 March 1956, N. B. Causey (MCZ). Pulaski County, North Little Rock, 26 April 1962, J. Ball (MCZ); Little Rock, summer 1943, E. M. Nelson (FMNH); Pinnacle Mt. State Park, no date, T. Yamashita (TY); Camp Robinson, 9 May 1944, L. Hook (AMNH); and 12.9 Table 1.— Morphological differences between C. vit- tatus and C. hentzi and C. guanensis (=C key si). C. vittatus C. hentzi/ C. guanensis 1 . Carapace with black inverted triangle cov- ering ocular tubercle (Figs. 1-2) 2. Dorsal surface of chelicerae usually uniformly yellowish (Fig. 1), occasionally with trace of reticula- tion (Fig. 2) 3. Pedipalp chela fingers and manus uniformly yellowish; all pedi- palpal segments uni- formly yellowish 4. Median yellow stripe of dorsum as wide or wider than black stripes (when stripes are present) 5. Higher pectinal tooth counts Male: 21-30 Female: 20-27 6. Metasomal segments with a single solid dusky midventral stripe lying between ventral submedian carinae 7. Legs usually immac- ulate yellow, rarely faintly infuscate 8. Telson suboval when viewed from ventral aspect 9. Telson midventrally with very weak smooth, longitudinal carina leading into subaculear tubercle (Fig. 4) 10. Subaculear tooth small, spinoid (Fig. 4) Carapace uniformly light mottled brown (Fig. 3) Dorsal surface of chelic- erae with distinct brown reticulation (Fig. 3) Pedipalp chela fingers in- fuscate; pedipalpal seg- ments with dusky mar- bling Median yellow stripe usu- ally about half as wide as black stripes Lower pectinal tooth counts Male: 17-19 Female: 16-18 Metasomal segments with entire ventral aspect of metasomal segments infuscate with pale spots marking positions of setae Legs moderately to heavi- ly infuscate Telson broadest apically, with subtle “shoulders” Telson ventrally with moderate, crenulate ca- rina leading into suba- culear tubercle (Fig. 5) Subaculear tooth larger, angular (Fig. 5) km N Camp Robinson, 21 April 1943, D. D. Davis (FMNH). Scott County, nr. Boles, 9 September 1967, D. M. Smith (INHS). Sharp County , 0.4 km SW Ash Flat, 16 September 1964 (INHS). Washington County , 24 km W Prairie Grove, Cove Cr. Valley, Boston Mtns., February 1956, M. Hite (MCZ); Prairie Grove, 20 Oc- tober 1955, M. Hite (MCZ); Fayetteville, no date, T. SHELLEY & SISSOM— CENTR UROIDES VITTATUS & C. HENTZI 105 Figure 7.— Comparisons of native distributions of scorpions in the United States and northern Mexico, emphasizing those species occurring in eastern and southcentral United States. 1, C. vittatus\ 2, C hentzi\ 3, C. gracilis ; 4, C. guanensis (= C. keysi); 5, Vaejovis carolinianus. The dotted line in western New Mexico and Arizona indicates the eastern boundary of the range of C. exilicauda (Wood); the boundary in New Mexico is based on the unpublished data of the junior author. Yamashita (TY); Fayetteville, 4 October 1938, L. G. Hembest (NMNH) and 19 July 1953, N. B. Causey (MCZ); and 40 km W Fayetteville, 20 July 1969, A. Graves (FSCA). Sight records (T. Yamashita, in litt. to second author): Lake Ouachita (several sites, all rocky areas) in Hot Springs Co. COLORADO: Baca County, Regnier, ca. 36.8 km S Pritchett, Comanchee Nat. Grassland (AMNH). Fremont County, 0.8 km W Can- on City, 5 September 1958, A. W. Spencer (CAS); and Canon City, June 1968, R. L. Kaesler (SEM), and out- side Fly Cv., 28 August 1961, W. J. Gertsch, W. Ivie (AMNH). Las Animas County, along CO hwy. 109, 2 1 June-1 1 July 1966 (AMNH). Otero County, along CO hwy. 109, 6 July 1967 (AMNH). Prowers County, Two Buttes Reservoir, 3-19 July 1966 (AMNH). Pueblo County, Boone, 6 July-1 August 1967 (AMNH); along CO hwy. 78, 16 March 1963-15 August 1964 (AMNH); and Lime, ca. 16 km S, 3.2 km E Pueblo, 12 May year unknown, Brookhart (AMNH). ILLINOIS: Monroe County, 6.4 km N Fults, 9 April 1949, D. M. Smith (INHS); 4 km N Fults, 13 July 1949, A. G. Wright (INHS); Fults, 9 October 1948 and 16 June 1949, P. W. Smith (CAS, INHS), 15 July 1953, Hensley & Smith (INHS), and spring 1971, D. Daleske (FMNH); and 6.4 km S Valmeyer, 1 May 1956, P. W. Smith (INHS). Randolph County, 3. 2-4. 8 km N Prairie du Rocher, October 1980, R. W. Sites (UMO); and Prairie du Rocher, 28 June 1949, Smith & Stannard (INHS) and 29 September 1982, J. H. Gerrard (AMNH). KAN- SAS: Allen County, Humboldt, 24 August 1944 (CAS). Chase County, 2.4 and 4.8 km S Saffordville, 24 June 1964 and 10 May 1965, R. F. Clarke (CAS), and 1 1.2 km S Saffordville, 20 June 1965, R. Zwiefel (AMNH); 8 km S Strong City, 4 June 1957, C. E. Goulden (CAS); and east edge L. Kahola, 17 June 1965, R. F. Clarke (CAS). Chautauqua County, 4.8 km W Peru, 3 April 1933, C. E. Burt (NMNH). Cowley County, Winfield vie., 1933, C. E. Burt, B. Anderson (NMNH). Douglas County, Lawrence, B. C. Marshall (ANSP), 10 October 1947 (CAS), 28 April 1948 (INHS), and 22 September 1962 (SEM); Lawrence, no date, T. Yamashita (TY); prairie outside Lawrence, 3 May 1964, S. Roth (CAS); 4.8 km W Lawrence, Clinton Lake Rec. Area, 9 May 1993, B. Cutler (UMN); and Rock Cr., 3 May 1899 (NMNH). Lyon County, Emporia, 1 4 May and 2 1 Sep- tember 1966 (CAS). Osage County, 17 April 1966, R. F. Clarke (CAS). Reno County, Hutchinson, 27 July 1951 (CAS). Riley County, Manhattan, 10-19 Septem- ber 1 904 (MCZ) and 1 927, C. E. Burt (NMNH). Wilson County, Altoona, Neodesha Region, 11 August 1977, Gordon (FMNH). Woodson County, Toronto (CAS). Wyandotte County, Kansas City (NMNH). Sight re- cords (B. Cutler, in litt. to first author): Butler, Cher- okee, Clark, Elk, Trego, and Wabaunsee counties. LOUISIANA: Allen Parish, 16 February 1963, D. A. Rossman (LSU). Beauregard Parish, DeRidder, 1943, E. L. Bell (AMNH); and 41.6 km N Lake Charles, 10- 17 August 1941, E. L. Bell (AMNH). Caddo Parish, Ida, 11 June 1972, F. H. Eubanks (CAS); Blanchard, 14 February 1967, K. Howard (CAS); and Shreveport, 106 THE JOURNAL OF ARACHNOLOGY 24 March 1962 (MCZ). Caldwell Parish , 10.1 km E Columbia Heights, along LA hwy. 4, 3 April 1966, R. E. Tandy (FSCA). Claiborne Parish , Homer, 24 Jan- uary 1953, N. B. Causey (MCZ). Evangeline Parish, Chicot St. Pk., 25 November 1944, W. G. Moore (NMNH). Grant Parish, Kisatchie Nat. For., 1 April 1962, G. Pontiff (LSU); Kisatchie Nat. For., Longleaf Vista, no date, T. Yamashita (TY); 5.6 km N Williana, 10 May 1954, H. S. Dybas (FMNH); nr. Williana, 19 October 1953, Dybas (FMNH); and Dry Prong, 9 May 1954, Dybas (FMNH). Jackson Parish , Wyatt, 1 Sep- tember 1941 (AMNH). Lincoln Parish, Ruston, 10 July 1950, M. Cazier (AMNH) and 1 1 April 1960, D Cope- land (CAS); and 8 km E Ruston, 4 January 1 96 1 (LSU). Rapides Parish , Alexandria, H. W. Tobias (NMNH); Kisatchie Nat. For., Johnson Tract, 30 December 1958, D. J. Pirone (AMNH); Forest Hill, 1 1 November 1945, R. L. Wenzel (FMNH); Melder, Fall 1941, E. L. Bell (AMNH); and Camp Claiborne, 28 March 1969, F. C. Rabalais (LSU). Sabine Parish , 12.8 km NNW Many, 10 September aa 3 October 1942, E, C, Williams (FMNH). Vernon Parish , 6.4 km NE Leesville, 19 April 1 962, K. Arnold (LSU); and 4.3 km NE Caney, 8 April 1967, L. D. Wilson (FSCA). MISSOURI: Barry Coun- ty, Roaring River St. Pk., 9 May 1936, E. G. Fisher (ANSP); and Washburn, 14 May 1930, H. H. Shamel (NMNH). Benton Co., nr. Warsaw, 0.8 km N jet. MO hwys. 7 & UU, E. G. Riley (UMO); and 4.8 km NW Warsaw, 10 April 1968, J. R. Heitzman (FSCA). Carter County, Fremont, 6 January 1942 (CAS). Christian County, 2 August 1976, T. J. Riley (UMO). Cole Coun- ty, 9.6 km SE Russellville, 23 March 1958, H. D. Raithoe (CAS). Hickory County, 4.8 km E Wheatland, 20 April 1976, R. Sajdak, J. Buday (MPM). Holt Coun- ty, 8 km S Forest City, 12 May 1992, A. P. Bufalino (UMO). Howell County, Willow Springs, before De- cember 1959, N. Banks (MCZ). Iron County, Granite- ville, 20 March 1955, R. E. Crabill (NMNH); 16 km W Ironton, 25 August 1965, R. L. Mondale (CAS); and Ironton, 5 August 1965-31 July 1966, R. L. Mondale (CAS). Jackson County, Kansas City, 3 July 1955 (CAS). Jasper County , Joplin, Fall 1984 (WDS). Jefferson County, Pevely, 23 April 1956, J. M. Kingsolver (INHS). Madison County , Fredericktown, 29 July 1967-19 Au- gust 1968, R. L. Mondale (CAS). Miller County, 3.2 km SE Bromley, 1 6 June 1 967, M. A. Nickerson (CAS). Moniteau County, 5 September 1961, E. McDaniel (CAS). Oregon County , Alton, May 1956, W. F. Rush- ton (MCZ). Phelps County, Rolla, April 1964, J. R. Waring (AMNH). St. Clair County, 1 6 km NE Osceola, 11 July 1955, P. Anderson (AMNH); and Collins, 17 July 1956 (CAS). St. Genevieve County , 6 June 1937 (FMNH); Misplay Glade nr. Co. Rd. DD, no date, T. Yamashita (TY); and St. Genevieve, 31 May 1956, H. A. Lowenstam (INHS). St. Louis County, St. Louis, September 1 929, P. Paul (NMNH) and 29 March 1 963, K. Rhoads (CAS); Glencoe, 25 April 1962, K. Rhoads (CAS); Ranken, 21 July 1929-14 May 1933 and 29 July 1945, E. P. Meiner (NMNH, UMO); Kirkwood (NMNH); and Eureka, 6-7 March 1963, K. Rhoads (CAS). Stone County, 6 October 1973, C. R. Mappes (UMO); N Kimberling City, Table Rock L,, Joe Bald area, 25 May 1974, S. E. Thewke (UMO); Notch, 27 August-4 September 1924, A. B. Wolcott (FMNH); and Cape Fair, 14 April 1954, Bagby (MCZ). Taney County, Hollister, 19 August 1962, J. C. Johnson (CAS). Washington County , Washington St. Pk., 16 May 1943, C. J. Goodnight (AMNH). County unknown, Osage Bluff, 1940 (NMNH). NEBRASKA: Thayer County, Williams, 2 May 1948, Jones & Loomis (FSCA); 4 km N Gilead, 12 July 1982, W. F. Rapp (AMNH, WFR); Gilead, 22 September 1982 and 25 May 1983, Rapp (WFR); and Alexandria, 24 June and 22 September 1982, Rapp (AMNH, WFR). NEW MEXICO: Chaves County, 11.2 km E Roswell, along Pecos R , 29 July 1956, V. Roth, W. J. Gertsch (AMNH). DeBaca Coun- ty, 28.8 km S Taiban, 10 September 1958, D. Hall (CAS). Dona Ana County, University Park, 8 July 1967, B. A. Smith (CAS); and Las Cruces, 22 October 1975, T. Schowalter (UGA). Eddy County, Carlsbad, 1 9 June and 21 July 1964, H. T. Hoskins, R. W. Reeves (CAS); Whites City, 24 September 1950 (AMNH) and 5 Oc- tober 1961, W. J. Gertsch, W. Ivie (AMNH); Carlsbad Caverns Nat. Pk., 1 September 1947 (CAS); and 43.2 km SW Carlsbad, 18-25 July 1964, P. G. Sanchez, R. W. Reeves, P. F. Van Cleave (CAS). Guadalupe Coun- ty, 4.8 km N Vaughn, 18 August 1958, T. Marquez (CAS). Lincoln County, 9.6 km W Carrizozo, Malpais Lava Flow, 27 June 1947 (CAS). Otero County , La Luz, 10 August 1959, R. A. Miller (CAS). Quay Coun- ty, Ft. Bascoms, now in Tucumcari (NMNH). San Mi- guel County, Las Vegas (INHS), 3 September 1963, L. Nichols, and 13-15 1968, M. Gratten, H. L. Stahnke (CAS); 1.6 km S Las Vegas, 28 September 1965, H. Trujillo (CAS); nr Las Vegas (Hot Spgs.), August 1901, Schwarz (NMNH); and Montezuma, 27 August 1958, W. L. Smith (CAS). Socorro County, Mockingbird Gap, S of Oscura Mts., 5 August 1 967, C. H. Lowe (AMNH). OKLAHOMA: Alfalfa County , Cherokee, 6 Novem- ber 1961, B. Young (CAS), and 8 and 12.8 km S Cher- okee, 2 and 18 October 1961, J. Herrington (CAS). Beckham County, Sayre (CAS). Blaine County, Brown Nose St. Pk., R. L. Landie (OKSU). Carter County , Ardmore, October 1954 (CAS). Cherokee County, Hul- bert, 10 November 1954, R. N. Van Noy (OMNH). Cimarron County, 24 km N, 1 1.2 km W Boise City, 22 June 1966, D. C. Arnold (OKSU). Cleveland Coun- ty, Norman, 7 February 1932, H. Fisher (OMNH), 15 August 1959, J. Ward (MCZ), and 13 October 1975, C. Treaftig (OMNH); and L. Thunderbird, 16 October 1968, R. M. Waering (FSCA). Comanche County , Wichita Mtns. Nat. Wildlife Ref., 20 July 1932, H. Fisher (OMNH), 6 June 1939, E. Hixon (CAS), 3 Sep- tember 1 949, C. J. Goodnight (AMNH), 9 March 1963, G. L. Rotramel (INHS), and 30 March 1981, J. M. Carpenter (MCZ). Craig County, 15 April 1960, Os- ume (OKSU). Dewey County, 1.6 km N Taloga, along US hwy. 183 at S. Canadian R. 18 April 1980, S. K. SHELLEY & SISSOM- CENTR UROIDES VITTATUS & C. HENTZI 107 Wu, P. B. LaRochelle (UCO). Ellis County, L. Lloyd Vincent, 14 October 1967, D. C. Arnold (OKSU). Gar- field County, 9.6 km E, 3.2 km S Bison, 20 August 1989, R. L. Landie (OKSU). Greer County, Quartz Mtn. St. Pk., 20 September 1952 (NMNH) and 4 June 1954, P. W. Smith (INHS). Harper County, 4.8 km N, 3.2 km W Ft. Supply, 22 September 1957, Harper (OKSU). Haskell County, Kinto, 4 November 1988, L. Felchick (OKSU). Hughes County, 10 December 1933, J. R. Carpenter (OMNH). Kay County, Ponca City, 15 October 1975 (OKSU). Latimer County, 13 April and 1 1 June 1931, R. D. Bird (OMNH). LeFlore County, nr. Poteau, 16 May 1961 (OKSU). McClain County, Johnson’s Pasture, 16 February 1935 (OMNH). Marshall County, 3.2 km W Willis, along Cowan Cr., 27 June 1958, B. A. Branson (INHS); and L. Texoma, 16 August 1965, B. Rotramel (INHS). Murray County, Sulphur, 15-28 June and 20 August 1956 (CAS). Mus- kogee County, 3 August 1978, B. G. Hill (OKSU). Osage County, Hulah, 7 May 1985 (MCZ); and Osage Hills St. Pk., 12 October 1985, Blackwood (OKSU). Pawnee County, Pawnee, 33 September 1963, M. E. Sisk (OKSU). Payne County, Stillwater, 20 November 1990 and 24 November 1963, D. C. Arnold (OKSU) and spring 1970, L. T. Chapin (LSU); and nr. L. Black- well, 22 March 1990, M. Lee (OKSU). Pottawatomie County, Pearson, 26 February 1974, D. C. Arnold (OKSU); and Shawnee, 18 May 1952, J. M. W. (OMNH). Roger Mills County, Cheyenne, 5 August 1953, M. C. Sooter (FSCA). Seminole County, May 1930 (OMNH). Sequoyah County, nr. Sallisaw, 30 June 1961 (OKSU). Stephens County, 9.6-1 1.2 km E Dun- can, KX Ranch, 1981, R. E. Knight (MCZ); and Co- manche, 15 March 1972, D. C. Arnold (OKSU). Till- man County, 14.4 km S Davidson, along Red R., 15 May 1960, M. B. Lamb (OKSU). Tulsa County, Tulsa, 1 October 1951 (CAS), and 6 September 1966, R. M. Waering (FSCA). Woodward County, vie. Alabaster Caverns St. Pk., 5-11 October 1952 (AMNH). TEX- AS: Samples were examined which produced the fol- lowing 59 new county records: Anderson, Angelina, Atascosa, Bandera, Bastrop, Baylor, Brazoria, Brazos, Briscoe, Brown, Burnet, Calhoun, Camp, Cherokee, Comal, Coryell, Ector, El Paso, Fisher, Gray, Harris, Harrison, Hays, Houston, Hudspeth, Hutchinson, Jones, Kendall, Lamar, Lampasas, Leon, Liberty, Lla- no, Live Oak, Madison, Matagorda, McLennon, Mil- am, Mitchell, Montague, Moore, Newton, Nolan, Nu- eces, Palo Pinto, Polk, Potter, Randall, Refugio, Rob- ertson, Runnels, Rusk, San Augustine, Smith, Sterling, Tarrant, Upton, Walker, and Wilbarger (Depositories: AMNH, ANSP, INHS, FMNH, OMNH, SEM, SFA- SU, TAMU, UGA, UTEP, WDS, WTAMU). Sight records (K. J. McWest, in litt. to second author, 1994): Bell, Callahan, Collin, Eastland, Jasper, Montgomery, Nacogdoches, Rockwall, San Jacinto, and Shelby. These new records and observations bring to 135 the total number of counties in Texas from which C. vittatus has been recorded. MEXICO: CHIHUAHUA: Below Sierra Ponce, S of Santa Elena (across border from Castolon, Tx.), March 1991, P. Klawinski, P. Monk, R. Truss (SFAU). COAHUILA: Saltillo, 23 May 1952, Cazier, Gertsch, Schramme (NMNH); 20 km N Sal- tillo, 6 Jan 1977, Cokendolpher and Dalquest (MWSU); 40 km N Saltillo, 6 Aug 1972, J. Kaspar (MWSU); Los Pinas, 17.3 km S, 0.2 km E Arteaga, 14 July 1977, Liner, Chaney (FSCA); 24 November 1977, Liner, Bartlett (FSCA); and 1 7 July 1975, Liner (FSCA). Valle de Guerra, 8.6 km W Bunuelos, 15 July 1977, Liner, Chaney (FSCA). Campo Central, 48 km SE Boquillas (FMNH). 33.6 km NW Ciudad Melchor Muzquiz, 21 July 1972, Liner, Johnson, Chaney (FSCA). Sierra de Penetente, Saltillo to Diamante, 2,340 m, 13 July 1934 (ANSP). Tinajas de Chaves, 32 km S Boquillas, 8 April 1945, K. P. Schmidt (FMNH). NUEVO LEON: Km 888 Hualahuises, 10 January 1948 (AMNH). 9.1 km SSW Cerralvo, 14 July 1975, E. A. Liner et al. (FSCA). Montemorelos (FMNH). 32 km N Montemorelos, 16 June 1941, H. S. Dybas (FMNH). Monterrey, 14 June 1941, Dybas (FMNH). 1 .0 km S Portrero, Arroyo Mes- quital, 16 July 1974, Liner, R. M. Johnson, A. H. Chaney (FSCA). 3.2 km W, 2.2 km S San Antonio de las Alazanas, Cienega del Toro Rd., 24 November 1977, Liner, P. Bartlett (FSCA). 2.2 km SW San Isidro, 22 July 1976, Liner et al. (FSCA). 23 km E San Jose de Iturbide, 21 July 1976, Liner et al. (FSCA). 4.8 km S Galeana, 22 May 1973, D. A. Rossman (FSCA). Pi- cacho Mts., 9.8 km SW, 1 1.8 km NW Cerralvo, Ran- cho El Milagro, 1 1 July 1977, Liner, Chaney (FSCA). 2.9 km E San Juan Batista, 1 2 July 1977, Liner, Chaney (FSCA). San Juan Batista to La Cienega, Canon San Juan Batista, 1 5 July 1 974, Liner et al. (FSCA). Cienega de Flores, 14 June 1941, Dybas (FMNH). TAMAU- LIPAS: Reynosa, C. C. Hoffmann (AMNH). W of Ma- tamoros, 13 October 1985, H. R. Hermann (UGA). Ciudad Victoria, 17 May 1952 (NMNH) and 17 May 1952, W. J. Gertsch, M. Cazier, R. Schramme (AMNH). Abasolo, 17 May 1952, Cazier, Gertsch, Schramme (NMNH). El Tinieblo, 12 March 1972, B. D. Camp- bell, R. W. Mitchell (AMNH) and 6 March 1977, R. W. Mitchell (AMNH). Padilla, 17 May 1952, Cazier, Gertsch, Schramme (AMNH). Jimenez, 15 May 1952, Cazier, Gertsch, Schramme (NMNH). La Reforma, 1 5 October 1984, P. Sprouse (TMM). Sistemica Purifi- cacion, 26 November 1979, P. Sprouse (TMM). Human importation to others areas.— Specimens that are believed to represent accidental human importa- tions were examined from the following localities. The possibility also exists that some of the specimens below bear erroneous locality data. The specimen from Chi- cago was encountered in a street, and those from Dare and Wake counties, North Carolina, and Rankin Coun- ty, Mississippi, were taken in buildings. Juveniles have been collected in Rutherford County, Tennessee, in- dicating the probable existence of an established, re- producing population (Denise Due, Vanderbilt Univ., pers. comm, to second author, 1981). 108 THE JOURNAL OF ARACHNOLOGY CALIFORNIA: Contra Costa County, Richmond, Au- gust 1952 (CAS). ARIZONA: Maricopa County , Phoe- nix, 22 March 1952, S. Smith (CAS). County unknown, Kamah, P. A. Vestal (MCZ). COLORADO: Boulder County, Boulder, 8 June 1954 (CAS). IOWA: County unknown, Missouri Valley, 4 August 1941 (CAS). MISSOURI: Clark County, Fairmont (CAS). ILLI- NOIS: Cook County, Chicago, 3 June 1922, M. Jensen (FMNH). McHenry County, Woodstock, October 1951 (CAS). KENTUCKY: Marshall/Calloway Counties, Kentucky L., 15 August 1975 (INHS). TENNESSEE: Rutherford County, Tiger Hill nr. Murfreesboro, No- vember 1981 (WDS). Shelby County, Memphis, 13 August 1955 (CAS), and January 1963, O. E. Smith (CAS). LOUISIANA: East Baton Rouge Parish, Baton Rouge, November 1962, T. B. Murrell (LSU) and 12 October 1983, M. Villars (LSU). Orleans Parish, New Orleans, 31 July 1962, P. Esteve (CAS). County un- known, Mississippi River (NMNH). MISSISSIPPI: Lamar County, nr. Sumrall, summer 1989 (MMNS). Pike County, Summit, September 1966, J. D. Smith (CAS). Rankin County, Brandon, 7 August 1990, B. Tanner (MEM). NORTH CAROLINA: Dare County , Nags Head, 10 May 1986, L. Griffin (NCSU). Nash County, Rocky Mount, 26 July 1991 (NCSM). Wake County, Research Triangle Park, August 1991 (NCSM); and Raleigh, Wakefield St., 24 October 1986, M. A. Brittain (NCSU) and Bland Rd., 13 May 1991, J. Wig- more (NCSM). It is also noteworthy that C. vittatus may have been introduced abroad as well. At least, there are some specimens in museum collections bearing labels from locations in South America. Sissom & Louren^o (1987) discovered that the species C. dasypus Mello-Leitao described from Andahuaylas, Peru was in fact C. vit- tatus. These specimens were probably mislabeled, as the locality is deep in the mountainous interior of the country. There are also several specimens of C. vittatus from Caracas, Venezuela in the Field Museum of Nat- ural History, Chicago. Centruroides hentzi (Banks) Figs. 3, 5, 6, 7 Habitat.— According to Muma (1967), C. hentzi usually occurs under litter, logs, and stones in Florida; it can also be found under bark of dead trees and often enters houses. North of Flor- ida, specimens were encountered under pine and oak bark in Camden and Charlton counties, Georgia, respectively, and inside houses, con- dominiums, or dormitories in Charlton County, Georgia; Charleston County, South Carolina; and Durham, Carteret, and Brunswick counties, North Carolina. Of the six specimens seen at the South Carolina site, two have been preserved, one of which was in a sleeve of a robe and stung the collector when she tried to put on the robe (Langley 1994). Distribution.— In the United States, C. hentzi was previously known only from Florida, where it occurs statewide (Muma 1967); it can now be reported from adjacent parts of Alabama and Georgia, where it would logically be anticipated (Figs. 6, 7). In Georgia, the scorpion appears to be common in the southern tier of counties ad- jacent to Florida; it occurs offshore on Cumber- land Island, and these specimens constitute top- otypes of Buthus vittatus Say. The westernmost locality, Mobile, Alabama, establishes C. hentzi west of the Alabama River and suggests eventual discovery in southeastern Mississippi. Speci- mens from outside of Florida that are believed to represent natural occurrences were examined from the following localities: ALABAMA: Baldwin County, Bon Secour Nat. Wild- life Ref., along AL hwy. 1 80 ca. 19.2 km W Gulf Shores and 41.6 km W Florida State/Escambia County line, 17 April 1993, R. L. Brown (MEM); and Josephine, 29 December 1993 (NCSM). Mobile County, Mobile, H. P. Loding (AMNH). GEORGIA: Camden County, private land on Cumberland Island, 25 December 1993, C. Ruckdeschel, C. R. Shoop (CIM, NCSM, NMNH, WDS). Charlton County, 6.4 km W Folkston, 19 March 1936, F. Harper (NMNH); and Okefenokee Natl. Wild- life Ref., 13 June 1981, C. L. Smith, S. N. Brown (UGA) and Billy’s Island, 29 January 1978, D. H. Ha- beck (FSCA). Clinch County, 22.4 km N Fargo, 25 December 1 949, Smith & Smith (INHS). Cook County, Adel, Fall 1937, J. T. Dampier (NMNH). Lowndes County, 4.8 km NW Valdosta, March 1 976, D. Daleske (FMNH). Mitchell County, 6.4 km N Sale City, 27 November 1949, J. W. Crenshaw (FSCA). Thomas County, 20 April 1973 and 6 January 1976, W. T. Sedgwick (MCZ); Thomasville, 9 December 1903, M. Hebard (ANSP), March 1939, J. White (FMNH) and May 1 942, E. Ireland (FMNH); 1 6 km S Thomasville (Birdsong Plantation), 15 April 1945, D. C. Lowrie (FMNH); and Millpond Plantation, 3 March 1973 (MCZ). Ware County , Waycross, 8 May 1937, T. H. Hubbell (CAS); 16 km SE Waycross, 16 March 1963, H. W. & L. R. Levi (MCZ); and Laura Walker St. Pk., 19 February 1988, W. E. Steiner, J. B. Stribling (NMNH). County unknown, Clermont, 21 June 1955, A. W. Vasquez (NMNH). Human importation to other areas.— Specimens that are believed to represent accidental human introduc- tions were examined from the following localities. The samples from Georgia could conceivably represent nat- ural occurrence because these two counties, in the Fall Zone and outer periphery of the Piedmont Plateau, ca. 1 60 km from the most proximate locality, are not so remote as to be implausible indigenous records, par- SHELLEY & SISSOM - CENTR U ROIDES VITTATUS & C HENTZI 109 ticularly if C. hentzi occurs northward in the Gulf Coastal Plain.. However, they are detached from the clustered and unquestionably native records in south- ern Georgia and are therefore treated as introductions. The specimens from South Carolina and Brunswick County, North Carolina, were possibly imported with Florida palm trees that have been planted along the coast of the Carolinas (J. Morse, pers. comm). Repro- ducing populations have not been verified at any of the following sites. GEORGIA: Harris County, 10 May 1970 (CC). Mus- cogee County, Columbus, 17 May 1959 (CC). NORTH CAROLINA: Brunswick County, Bald Head L, Feb- ruary 1 993 (NCSM) and 1 .6 km E of Marina, July 1 992 (RNH). Carteret County, Bogue Banks, Emerald Isle, September 1993, D. McLuskey (NCSM). Durham County, Duke Univ., 8 September 1987, C. Brock (NCSU). SOUTH CAROLINA: Charleston County, Isle of Palms, 20 February 1994, S. Mims (NCSM). COMPARISON OF SCORPION DISTRIBUTIONS IN THE EASTERN UNITED STATES The known indigenous distributions of the scorpions in the United States east of the Central Plains are compared in Fig. 7; the ranges of C. gracilis and C. guanensis (= C. keysi), and V. carolinianus are adapted from maps published by Muma (1967) and Shelley (1994a), respec- tively. There is no known overlap between V. carolinianus and any of the buthids, although its range is only about 1 1 2 km north of that of C. hentzi in southern Alabama. According to Muma (1967), C guanensis is restricted to Collier, Dade, and Monroe counties, Florida. Centruroides gracilis is indigenous to the peninsula from Ala- chua County southward. To our knowledge, these distributions are still current, but we did discover the following two samples of C. gracilis, repre- senting accidental human importations, from well outside this area. MISSISSIPPI: Rankin County, in concrete debris at truck stop on US hwy. 49 just S 1-20, nr. Jackson, 27 October 1983, E. S. Olson (MMNS). TEXAS: Galves- ton County, Galveston, 1935 (NMNH). Additionally, there is an invidividual of C. gracilis from Dallas, Texas (NMNH), that was taken in 1956 “in produce from Central America,” and another col- lected in 1930 on a ship berthed at New Orleans (NMNH). ACKNOWLEDGMENTS We thank the following curators and collection managers for loaning specimens from the indi- cated institutional collections: AMNH, N. I. Platnick; ANSP, D. Azuma; CAS, C. E. Gris- wold; CC, G. E. Stanton; CIM, C. Ruckdeschel; FMNH, D. Summers and T. G. Anton; FSCA, G. B. Edwards; INKS, K. R. Methven; LSU, V. L. Moseley and D. A. Rossman; MCZ, H. W. Levi; MEM, T. L. Schiefer; MMNS, R. L. Jones; MPM, J. P. Jass; MWSU, N. V. Homer; NCSU, R. L. Blinn; NMNH, J. A. Coddington; OKSU, D. C. Arnold; OMNH, J. A. Droke; PMNH, R. J. Pupedis; SEM, R. W. Brooks; SFAU, K. J. McWest; TAMU, Ed Riley; TMM, J. R. Reddell; UCO, S. K. Wu; UGA, C. L. Smith; UMN, P. J. Clausen; UMO, R. W. Sites; and UTEP, W. MacKay. R. N. Henson, W. F. Rapp, and T. Yamashita kindly provided records and speci- mens from their private collections; D. C. Arnold assisted with Oklahoma literature records. Spe- cial thanks are extended to Jody Young, Foley, Alabama, for the Josephine, Alabama, specimen of C. hentzi\ to Carol Ruckdeschel (CIM), for collecting C. hentzi (= topotypes of B. vittatus Say) on Cumberland Island, Georgia; to B. Merle Shepard, Clemson University Coastal Research Center, Charleston, South Carolina, for the spec- imens of C. hentzi from Isle of Palms, South Carolina; and to K. J. McWest (SFASU) and T. Yamashita (Northwest Louisiana University) for providing personal observations and distribu- tional information on C. vittatus. John Morse of Clemson University provided information on the suspected mode of importation of C. hentzi along the coasts of North and South Carolina, and Denise Due of Nashville, Tennessee provided information regarding the population of C. vit- tatus in Murfreesboro. Doug Rossman of LSU assisted with finding localities in Mexico for the distributional maps. We are particularly grateful to Harley P. Brown for personally retrieving the samples at the OMNH for subsequent shipment to us; without his kind assistance this collection would have been unavailable. Page charges were paid jointly by the North Carolina State Museum of Natural Sciences and the Dept of Biology and Geosciences at West Texas A & M University. Figures 1-3 were prepared by R. G. Kuhler, NCSM scientific illustrator; Cathy Wood per- formed word processing chores. LITERATURE CITED Banks, N. 1901. Some Arachnida from New Mexico. Proc. Acad. Nat. Sci., Philadelphia, 53:574-595. Banks, N., M. Newport & R. D. Bird. 1932. Okla- homa spiders. Publ. Univ. Oklahoma Biol. Surv., 4:3-49. 110 THE JOURNAL OF ARACHNOLOGY Cokendolpher, J. C. & F. D. Bryce. 1980. Arachnids (excluding Acarina and Pseudoscorpionida) of the Wichita Mountains Wildlife Refuge, Oklahoma. Occas. Paps. Mus. Texas Tech Univ. No. 67:1-25. de Armas, L. F. 1981. El genero Centruroides Marx, 1889 (Scorpions: Buthidae), en Bahamas y Repub- lica Dominicana. Poeyana, 223:1-21. Diaz Najera, A. 1975. Listas y datos de distribucion geografica de los alacranes de Mexico (Scorpionida). Rev. Inst. Salud. Publica, Mexico, 35:1-36. Formanowicz, D. R. & L. R. Shaffer. 1993. Repro- ductive investment in the scorpion Centruroides vit- tatus. Oecologia 94: 368-372. Gentry, A., V. D. Roth & W. D. Sisson. 1991. Com- ments on the proposed conservation of the scorpion names Buthus vittatus Say, 1821, Centrums hentzi Banks, 1 904 and Buthus vittatus Guerin Meneville, [1838] (Arachnida, Scorpionida). Bull. Zool. No- men., 48:55-57. Hoffmann, C. C. 1932. Los Scorpiones de Mexico. Segunde Parte. Buthidae. An. Inst. Biol. Univ. Nac. Aut. Mexico, 3:243-361. Langley, L. 1991. Low country scorpions lurk in leaves and beds. The News and Courier/The Evening Post, Charleston, SC, 2 June:2E-3E. Langley, L. 1994. Scorpions here? Ask Mrs. Mims. The News and Courier/The Evening Post, Charles- ton, SC, 23 February: 1A-2 A. Lourengo, W. R. & O. F. Francke. 1984. The iden- tities of Tityus jloridanus and Tityus tenuimanus (Scorpiones: Buthidae). Florida Entomol., 67:424- 429. Muma, M. H. 1967. Scorpions, whip scorpions and wind scorpions of Florida. Arthropods of Florida and Neighboring Land Areas, 4:1-28. Opinion 1680. 1992. Buthus vittatus Say, 1821 (cur- rently Centmroides vittatus), Centrums hentzi Banks, 1 904 (currently Centruroides hentzi ) and Buthus vit- tatus Guerin Meneville, [1838] (currently Bothri- urus vittatus ) (Arachnida, Scorpionida): specific names conserved. Bull. Zool. Nomen., 49:163-164. Rapp, W. F. 1 987. Centruroides vittatus in south cen- tral Nebraska. American Arachnol. No. 36:9. Reddell, J. R. 1965. A checklist of the cave fauna of Texas. I. The Invertebrata (exclusive of Insecta). Texas J. Sci., 17:143-187. Reddell, J. R. 1970. A checklist of the cave fauna of Texas. IV. Additional records of Invertebrata (ex- clusive of Insecta). Texas J. Sci., 21:389-415. Rowland, J. M. & J. R. Reddell. 1976. Annotated checklist of the arachnid fauna of Texas (excluding Acarida and Araneida). Occas. Paps. Mus. Texas Tech Univ., No. 38:1-25. Say, T. 1821. An account of the Arachnides of the United States. J. Philadelphia Acad. Sci., 2:65-68. Shelley, R. M. 1994a. Distribution of the scorpion, Vaejovis carolinianus (Beauvois), a reevaluation. Brimleyana, 21:57-68. Shelley, R. M. 1 994b. Introductions of the scorpions, Centruroides vittatus (Say) and C. hentzi (Banks), into North Carolina, with records of the indigenous scorpion, Vaejovis carolinianus (Beauvois). Brim- leyana, 21:45-55. Sissom, W. D. & W. R. Louren^o. 1987. The genus Centruroides in South America (Scorpiones, Buth- idae). J. Arachnol., 15:11-28. Stahnke, H. L.& M. Calos. 1977. A key to the species of the genus Centruroides Marx (Scorpionida: Buth- idae). Entomol. News, 88:1 1 1-120. Stockwell, S. A. 1986. The scorpions of Texas (Arach- nida: Scorpiones). Master’s Thesis, Texas Tech Univ., Lubbock. Stockwell, S. A. & H. W. Levi. 1989. Case 2637. Buthus vittatus (currently Centruroides vittatus ; Arachnida, Scorpionida): proposed recognition of Wood (1863) as author of the specific name and designation of a neotype, and Centrums hentzi (cur- rently Centruroides hentzi) Banks, 1904: proposed conservation of the specific name. Bull. Zool. No- men., 46:233-235. Manuscript received 1 5 February 1995, revised 27 March 1995. 1995. The Journal of Arachnology 23:1 1 1-117 NATURAL HISTORY OF THE SPIDER GENUS LUTICA (ARANEAE, ZODARIIDAE) Martin G. Ramirez: Department of Biology, Bucknell University, Lewisburg, Pennsylvania 17837 USA ABSTRACT. Spiders of the genus Lutica are fossorial inhabitants of coastal dunes of southern California, Baja California and the California Channel Islands. They live in silk-lined burrows concentrated beneath dune vegetation. Lutica are sit-and-wait predators that subdue insects that walk near or over burrows. They are sedentary and do not engage in aerial dispersal via ballooning. Adult males abandon their burrows during the late summer and early fall to wander in search of females. Females produce eggsacs and guard them till they die; spiderlings emerge in the spring. Dune trapdoor spiders ( Aptostichus simus) prey on Lutica, while the larvae of a therevid fly are external parasites. Spiders of the genus Lutica are fossorial in- habitants of coastal dunes of southern California, Baja California and the California Channel Is- lands (Gertsch 1961, 1979; Ramirez 1988). Al- though described over 100 years ago (Marx 1891), little is known of their natural history. Gertsch (1979) stated that they are nocturnal and come to the surface at night to hunt various beetles and other insects that drop on the sand, and that they spin a loose tubular retreat deep in the cool, moist sand. Gertsch (1961) believed that they probably live for two to three years, with males maturing in the summer or fall, but admitted that little was known about the details of the . . lives and habits of these large, whitish spi- ders.” George Marx first described the genus Lutica from Klamath Lake, Oregon (Marx 1891). Gertsch (1961) corrected the type locality of Lu- tica maculata to Santa Rosa Island, California, and also described three new species: nicolasia (San Nicolas Island), clementea (San Clemente Island) and abalonea (Oxnard, Ventura County). Additional species have been described from In- dia (Tikader 1981), but these taxa are clearly misplaced (Jocque 1991). Gertsch (pers. comm.) has prepared a revision of Lutica based on mor- phological features, while Ramirez & Beckwitt (in press) have re-defined valid species and de- termined their phylogenetic relationships based largely on molecular characters. Since these works propose very different species designations than Gertsch (1961), species names in Lutica are un- certain at this time. My study of Lutica elaborates on the natural history of this obscure genus. METHODS From 1982 to 1987, 1 collected over 3000 Lu- tica from 20 different dune systems in southern California and Baja California (Fig. 1), including sites on all the Channel Islands except Anacapa (where they are not known to exist), as part of a study of the population genetics and biochemical systematics of this genus (Ramirez 1990). Spi- ders were collected by sifting dune sand beneath beach vegetation using geologic sieves with a minimum mesh size of 1.0 mm. All specimens were brought back to the laboratory alive, where they were either used for observations or pro- cessed for starch gel electrophoresis. Living spi- ders were maintained in small upright glass or plastic containers or in horizontal glass tubes, partly filled with beach sand. Water was added periodically with either an eye dropper or at- omizer. I fed them small arthropods, mainly fruit flies, house flies and beetle larvae (wireworms). For a mark-recapture experiment with Lutica in the field, I marked spiders on the dorsal surface of their abdomens with quick drying scale model paint (Testers Flat White), after first cooling the spiders in a refrigerator for 30 min to make them sluggish and easier to mark. After the spiders were warmed to ambient temperature, there was no visible difference in their behavior. RESULTS Burrow construction . — Individual Lutica readily constructed burrows in the laboratory af- ter being placed in sand-filled containers. Bur- rows consisted of silk-lined tunnels in the sand. 111 112 THE JOURNAL OF ARACHNOLOGY usually just below the surface and sometimes partly against the side of the glass container. This facilitated the observation (under subdued light) of activities within. Burrows had either open en- trances or no entrances. On two occasions, I observed burrow con- struction. In one case, the horizontal glass tube occupied by the spider was packed with moist sand in the sealed end. The spider moved about in a space between the sand and the lower side of the glass tube (Fig. 2). It moved its spinnerets from side to side and up and down, cementing fragments of sand together with silk, and slowly moved in a circle as it did so. It sometimes stopped this activity and moved over to the in- terior of the burrow wall where it pushed forward with its forelegs, pushing back the wall and ex- panding the burrow. It then resumed its circular spinning activity. I observed the spider until it suddenly halted its activity and did not resume work on its burrow. In the second case, also with a spider in a horizontal glass tube, the spider half- carried, half-pushed a pile of sand toward the entrance of its burrow. Before it reached the en- trance, it halted its activities and did not contin- ue. In the field, burrows were concentrated in and about stands of native dune vegetation, partic- ularly Abronia maritima and Franseria chant- issonis , and extended into the dune amidst litter and the root systems of the plants. On Santa Barbara Island, typical coastal dunes do not exist and these spiders live in the sandy soil and debris below vegetation growing on a sea cliff. While burrow entrances were normally not visible, one could often see small dimples on the open sur- faces of vegetated dunes after strong winds. These usually proved to be the entrances of Lutica bur- rows, composed of a delicate sand-covered, flap- like lid; this is consistent with Thompson’s (1 973) description of burrows on Santa Cruz and San Miguel Islands. Most burrows descended Into the sand at about a 45° angle, although some had portions of their length laying horizontally, just below the sand surface. On the other hand, at La Jolla Beach (Ventura County), I found four bur- rows that descended vertically into the sand. Lu- tica burrows were usually very fragile and quickly fell apart if the sand around them was removed. Individual burrows were usually from 2.5-1 5 cm in length, though W. Icenogie and I found a bur- row that was 25-30 cm long (occupied by a ma- ture female) at Little Harbor, Santa Catalina Is- land. Figure 1 . — Map of southern California and Baja Cal- ifornia, including the Channel Islands, showing Lutica sample sites. Population abbreviations are as follows: Channel Islands - Cuyler Harbor, San Miguel Island (SMI); Southeast Anchorage, Santa Rosa Island (SRI); Johnstons Lee, Santa Cruz Island (SCI); cliffs south of Signal Peak, Santa Barbara Island (SBI); Army Camp Beach (SNA), Dutch Harbor (SND), Red Eye Beach (SNE), San Nicolas Island; Little Harbor, Santa Cat- alina Island (CAT); Flasher Road Dunes, San Clemente Island (SCL); Mainland - Coal Oil Point Reserve (COP), Santa Barbara Co., California; McGrath State Beach (MG), Ventura Co., California; Oxnard Beach (OX), Ventura Co., California; La Jolla Beach (LJB), Ventura Co., California; Ball on a Wetlands (BA), Los Angeles Co., California; El Segundo Dunes, LAX (ESG), Los Angeles Co., California; Balboa Beach (NR), Orange Co., California; Ponto State Beach (PON), San Diego Co., California; Silverstrand State Beach (SVS), San Diego Co., California; Punta Estero (PE), Baja Cali- fornia Norte, Mexico; Guerrero Negro (GN), Baja Cal- ifornia Sur, Mexico. Prey capture.— Once they had constructed burrows in the laboratory, the spiders readily accepted small insects as food. An insect crawling about on the surface of the sand elicited an im- mediate response. The spider (hanging upside down) would rash about on the “ceiling” of its burrow, possibly trying to locate the exact po- sition of the insect by the vibrations caused by its activities. If the insect suddenly ceased its RAMIREZ— NATURAL HISTORY OF LUTICA 13 mesh cap burrow wall Figure 2.— Profile of horizontal Lutica rearing container, showing the orientation of a spider and its burrow. Containers were 2.5 cm in diameter and 9.5 cm long. movements, the spider would likewise stop its movements and would remain motionless until the insect started moving once again. Once the spider had positioned itself below the insect, it would lunge up and through the wall of the bur- row, grab the insect and pull it inside the burrow. Spiders sometimes left their burrow completely to pursue prey that initially escaped; they brought the prey back to the burrow either through the open entrance or through the hole created in leaving the burrow. Once insects ceased to strug- gle inside the burrow, the spider would leave the insect to patch up the hole in the burrow wall, and would then return to feed on the now dead prey. While the capture of prey through the burrow wall was typical, spiders sometimes emerged from their burrows at the first sign of prey vibrations and subdued their prey directly, before taking them inside the burrow. Most spiders deposited the prey remains outside the burrow following feeding. Field-collected burrows were always uniformly clean and a prey item was found inside a burrow on only one occasion. Timing of reproduction.— Lutica males molt to maturity and abandon their burrows to wan- der about in search of females. Based on a master list of collecting records of all Lutica specimens (available on request), the earliest record of an adult male is May 11(13, Oxnard Beach, Ventura County, 1968, M. Thompson) and the latest re- cords are November 3 (13, Oxnard Beach, Ven- tura County, 1982, M. Ramirez) and November 4(13, Silverstrand State Beach, San Diego Coun- ty, 1982, M. Ramirez). The largest number of records and actual numbers of males collected are for September and October. For example, pitfall traps set up in dunes at Pt. Mugu Naval Air Station, Ventura County collected 169 adult male Lutica between 3 1 August- 1 8 October 1981 (C. Nagano & J. Donohue pers. comm.). Thus, the peak of the breeding season appears to be late summer and early fall. Following mating, females produce eggsacs, though how soon is not known; there then fol- lows the period of spiderling growth and devel- opment. My earliest record of brood spiderlings was 15 April (El Segundo Dunes, Los Angeles International Airport, Los Angeles County, 1985), and I collected broods as late as August from both island and mainland populations. Females presumably guard eggsacs and young till they die; of 22 eggsacs or broods collected in 1985 and 1987, 12 were found in burrows along with the shrunken remains of adult females. Since bur- rows are destroyed during collection, scattering their contents, it is probable that the remains of adult females also may have been present with the other 10 eggsacs/broods. Dispersal.— Mark-recapture data suggest that non-reproductive dispersal is limited. At Coal Oil Point Reserve (Santa Barbara County), 170 spiders from a single dune were captured on 28 February 1984, marked on the dorsal surfaces of their abdomens and released into the dune from which they were taken. Seventy-seven (45.3%) of 1 70 spiders collected at that same dune a month later were marked. Assuming the loss of marks by individual spiders due to molts in the inter- vening period, actual site fidelity was probably greater. However, since I did not have an op- portunity to collect in dunes adjacent to the one in which the marked spiders were released, it is not known how many of the marked spiders I failed to recover may have moved to different dunes in the intervening month. Nonetheless, since their burrows were destroyed when the spi- ders were first collected, it is remarkable that 114 THE JOURNAL OF ARACHNOLOGY such a large percentage of them stayed in the same dune following release, Gertsch (1961) stated that Lutica do not bal- loon, as is common among many spiders (Decae 1987), and Lutica of all sizes instantly buried themselves in the sand if removed from their burrows. However, while sifting for Lutica dur- ing Santa Ana wind conditions, 1 often saw small specimens cling tenaciously to the mesh of the sieve; if they lost their grip, the smallest spiders would sometimes be blown up and out of the sieve. This is a highly unnatural situation, since the spiders do not normally move about at the surface during the day and would certainly not find themselves a foot or more above the sand surface. Dune vegetation is prostrate and I have never seen them climb about in plants. On the other hand, twice during Santa Ana winds, I saw a few immatures and adult females moving about on the surface of the dunes. Since I never ob- served Lutica moving about on the surface on any other occasions, it is possible that the wind had shifted the sand in the area where these spi- ders had made their burrows, eventually dis- lodging them. Thus, while they do not engage in ballooning behavior, it may be possible for the smallest instars of Lutica to be carried away in high winds. Predators. —Trapdoor spiders of the genus Ap- tostichus (Cyrtaucheniidae) are the only organ- isms known to prey on Lutica. One member of this genus, A. simus , is restricted to coastal dunes in southern California (Chamberlin 1917), in- cluding the California Channel Islands, and it lives in silk-lined burrows. In September 1979, W. Icenogle (pers. comm.) found the remains of an adult male Lutica (as well as an adult male Aptostichus ) in the burrow of an adult female Aptostichus in a coastal dune near Encinitas, San Diego County. Since only adult male Lutica would normally be expected to wander about on the surface of the sand, it is not likely that Aptosti- chus prey on non-male Lutica. Parasites.— Of the thousands of Lutica col- lected over six years, only a single spider was parasitized. In September 1983, 1 collected three Lutica from La Jolla Beach (Ventura County) which were paralyzed. Attached to the abdomen of one of the spiders was a small white larva. The larva eventually consumed the spider from the outside in, but unfortunately died without pupating. E. Schlinger (pers. comm.) identified the larva as that of a therevid fly (Diptera). No Therevidae have been reported previously as spi- der parasites (Eason et al. 1 967; E. Schlinger pers. comm.). Prey.— In the field, I recovered many Lutica with beetle larvae (wireworms) in their chelicerae and found one burrow which contained the dry remains of a wireworm. In the laboratory, Lutica readily attacked any small insects or spiders and never rejected any arthropod they were capable of subduing. If many prey items were supplied at once, most Lutica attacked and subdued all the arthropods in rapid succession before they began to feed on any of them. DISCUSSION Burrow construction.— The fossoria! lifestyle of Lutica is typical of the Zodariidae, most of which are ground or forest floor dwellers which often construct silken retreats, either burrows or silk-lined bags (Jocque 1991, 1993). Aside from Lutica, the construction of burrows with trap- doors has been reported among the Zodariidae in Antillorena (Gertsch 1961; Jocque 1991), Capheris (Hewitt 1914; Jocque 1991), Neosto- rena (Jocque 1991) and Psammorygma (Jocque 1991, 1993). Observations of burrow construc- tion have not been reported previously for a zo- dariid, although Harkness (1977) detailed the construction of a bag-type shelter by Zodarion frenatum. Prey capture. —The prey capture behavior in Lutica described herein is the first description of the sub-surface attack sequence of a burrow- dwelling zodariid. Since the orientation of bur- rows in the field ranged from nearly horizontal to vertical, it is probable that the sub-surface attack sequence described for Lutica only applies to those burrows which have at least some por- tion lying near the surface in a horizontal posi- tion, where arthropods can walk across them. With burrows situated at steeper angles, Lutica probably come out to attack passing insects, as did some laboratory spiders and as does Antil- lorena (Gertsch 1979). The sub-surface prey location and attack be- havior of Lutica strongly parallels that reported for the “purse web” spiders, Atypus , Calommata and Sphodros (Atypidae) (Bristowe 1958; Coyle 1986). These three spiders are all burrow dwell- ers which construct a tube-like, silken extension of the burrow (the “purse web”) that extends along the ground or vertically against a tree or other support. Prey are located when they walk or land on the purse web: the spider locates the position of the prey by its vibrations and once RAMIREZ — NATURAL HISTORY OF LUTICA 15 positioned below the insect, it then strikes through the silken tube, slits open the purse web and pulls the prey inside, much as Lutica does in its own burrows. However, while only the fangs of purse web spiders are extended though the tube wall to capture prey, Lutica may force much or all of its body through the burrow wall to do so. This similarity in attack sequence may be an example of convergence in behavior involving spiders in two very different families, due to the functional similarities of a purse web and a shallowly buried silk-lined burrow. Timing of reproduction.— The presence of males in the field largely in the summer and fall, coupled with the appearance of spiderlings by the spring, indicates that the production of egg- sacs and development of young takes place some- time between fall and spring. Bonnet (1935) not- ed that many spiders which mature and mate toward the end of the summer produce over- wintering eggsacs in the fall, with spiderlings emerging in the spring. Since brood spiderlings were collected in the field as early as April, it would appear that production of eggsacs and sub- sequent development of spiderlings is consistent with that of other spiders which mature in the late summer. When Lutica do reproduce, it is probable that the females guard the eggsacs and developing spiderlings in their burrows till they die, as evidenced by the regular collection of eggsacs or broods along with the remains of adult females. Dispersal. —Given the isolation of coastal dune systems along the southern California and Baja California coasts (Fig. 1) (Cooper 1967; Powell 1981), knowledge of the extent and timing of inter- and intra-dune dispersal by Lutica would be of great value in understanding the structure of their populations and patterns of genetic vari- ation within and among them (Ramirez 1990). On a local scale, a low dispersal rate among dif- ferent parts of a dune system [typical size 2-10 km2 (Powell 1981)] might lead to genetic sub- division, and possibly the evolution of micro- geographic races (Doyen & Slobodchikoff 1984). The results of the mark-recapture study sug- gest that non-reproductive terrestrial dispersal is low. Terrestrial dispersal is probably limited to wandering males and those spiders dislodged from their burrows by the shifting of dune sand. Nonetheless, dispersal on a local scale is appar- ently effective enough to maintain genetic ho- mogeneity among spiders in dunes on the same beach (Ramirez 1990). Ballooning is rare in fossorial spiders (Decae 1987) and has never been reported in the family Zodariidae (Jocque 1993). However, Robinson (1982) has suggested that spider aerial dispersal may sometimes be accidental More specifically, if a spider is small and light, it is possible that if it loses its hold of the substrate while exposed to wind of sufficient strength, it might become air- borne solely due to its favorable aerodynamic characteristics (Click 1939). This apparently happened with small Lutica in my sieves during Santa Ana winds. However, since ballooning spi- ders depend on wind borne silk threads for lift (Coyle 1983), it is unlikely that Lutica travel far even if they do become airborne, since they were never seen to pay out threads of silk into the wind or drop from elevated positions on drag- lines exposed to the wind, the two means spiders use to accomplish ballooning (Coyle 1983; Decae 1987). If aerial transport is a regular means of Lutica dispersal, one would expect that there would be few dune systems that they would not be capable of invading; yet, they are absent from most of the coastal region between Ventura County and Los Angeles (their absence from the well devel- oped dune system at Pt. Dume is particularly puzzling) and from Anacapa Island, the closest of the Channel Islands to the mainland (Fig. 1). Although Anacapa has no dune system, Lutica live on much more isolated Santa Barbara Island in a non-dune habitat. Thus, while it may be physically possible for Lutica to become air- borne, it is not likely that such a means of dis- persal has played a large part in creating present distributions. Predators.— Among the small but distinct ar- thropod fauna of California coastal dunes (Na- gano 1981; Powell 1981), Lutica and Aptostichus simus are the only predators to occupy silk-lined burrows. The record of a male Lutica from an Aptostichus burrow is not unexpected, since their burrows are often found side by side in the dunes. While no other case of predation on Lutica was observed, there are a few invertebrate and ver- tebrate insectivores that occupy California coast- al dunes and may potentially feed on Lutica , specifically windscorpions (Solpugida), side- blotched lizards ( Uta stansburiana ) and Califor- nia legless lizards (. Anniella pulchrd) (Hayes & Guyer 1981; Nagano et al 1981). However, such potential predators were only rarely encountered while collecting Lutica , Parasites. —The record of a there vid fly larva 116 THE JOURNAL OF ARACHNOLOGY consuming a paralyzed Lutica is highly unusual. Therevid larvae are predators of sand dune in- habiting insects and some may specialize on te- nebrionid larvae (Doyen 1976, 1984). Their in- teractions with spiders have not been reported previously. Spider wasps (Pompilidae), which are abundant in southern California, are spider spe- cialists and typically paralyze their prey (Waus- bauer & Kimsey 1985), so the three paralyzed spiders found at La Jolla Beach (Ventura County) were presumably the result of pompilid activity. If these spiders were indeed attacked by pompilid wasps, the absence of wasp eggs or larvae at- tached to the paralyzed bodies is puzzling; per- haps they were knocked off during sifting. The presence of a therevid larva attached to one of the three paralyzed spiders was probably the re- sult of a chance encounter with the immobile spider during the larva’s movements through the sand. Prey.— Tenebrionids (Coleoptera) and their larvae (wireworms) are among the most abun- dant insects in California coastal dunes (Doyen 1976, 1984) and their numbers far exceeded the numbers of other insects recovered during sift- ing. Both Gertsch (1961) and Thompson (1973) suspected that Lutica preyed on tenebrionids and my capture of many of them with wireworms in their chelicerae has proven them correct. How- ever, save for the chance collection of Lutica with prey items, it will be difficult to determine wheth- er Lutica prey on adult beetles (or any other or- ganisms), given their rapid disposal of prey re- mains upon completion of feeding. The appli- cation of electrophoretic (Murray & Solomon 1978; Fitzgerald et al 1986) and serological (Greenstone 1977; Southwood 1978) analyses might distinguish, from a range of possible prey items, what Lutica are actually eating. ACKNOWLEDGMENTS I am particularly indebted to W. Gertsch for sparking my initial interest in Lutica and to M. Thompson for sharing with me his field note- books and maps based on his own Lutica col- lecting efforts in the 60’s and 70’s. C. Cutler, C. Drost, W. Icenogle, M. Wilson and members of the Arachnologists of the Southwest provided assistance in the field. F. Coyle, J. Cronin, J. Estes, L. Fox, C. Griswold, S. Marshall and D. Potts offered many constructive comments on various drafts of this manuscript. For providing collecting permits and access to various main- land localities, I thank: S. Clarke, Marine Science Institute, University of California, Santa Bar- bara; P. Principe, Department of Airports, City of Los Angeles; and the California Department of Parks and Recreation. For providing collect- ing permits and access to the California Channel Islands, and for providing lodging and transpor- tation while on the islands, I am grateful to: W. Ehom and F. Ugolini, Channel Islands National Park; S. Clarke, Marine Science Institute, Uni- versity of California, Santa Barbara; L. Laughrin, Santa Cruz Island Reserve; A. Propst and T. Martin, Santa Catalina Island Conservancy; S. Bennett and R. Turner, Catalina Island Marine Institute; J. Estes, U.S. Fish and Wildlife Service; R. Dow, J. Larson and L. Salata, U.S. Navy. Financial support was provided by grants from a variety of sources at the University of Califor- nia, Santa Cruz (Division of Natural Sciences, Graduate Studies and Research, Department of Biology, Patent Fund, Minority Biomedical Re- search Support Program) and by a grant from the Exline-Frizzell Fund for Arachnological Re- search (Grant No. 10). LITERATURE CITED Bonnet, P. 1 935. La longevite chez les Araignees. Bull Soc. Entomol. France, 40:272-277. Bristowe, W. S. 1958. The World of Spiders. Collins, London. Chamberlin, R. V. 1917. New spiders of the family Aviculariidae. Bull. Mus. Comp. Zool., 61:25-75. Cooper, W. S. 1967. Coastal dunes of California. Geol. Soc. America Mem., 104:1-131. Coyle, F. A. 1983. Aerial dispersal by mygalomorph spiderlings (Araneae: Mygalomorphae). J. Arach- nol, 11:283-286. Coyle, F. A. 1986. The role of silk in prey capture by nonaraneomorph spiders. Pp. 269-305, In Spi- ders: Webs, Behavior and Evolution. (W. A. Shear, ed.). Stanford Univ. Press, Stanford, California. Decae, A. E. 1987. Dispersal: Ballooning and other mechanisms. Pp. 348-356, In Ecophysiology of Spi- ders. (W. Nentwig, ed.). Springer- Verlag, New York. Doyen, J. T. 1976. Biology and systematics of the genus Coelus (Coleoptera: Tentyriidae). J. Kansas Entomol Soc., 49:595-624. Doyen, J. T. 1 984. Systematics of Eusattus and Con - isattus (Coleoptera: Tenebrionidae; Coniontini; Eu- satti). Occas. Pap. California Acad. ScL, 141:1-104. Doyen, J. T. & C. N. Slobodchikoff. 1984. Evolution of microgeographic races without isolation in a coastal dune beetle. J. Biogeogr., 1 1:13-25. Eason, R. R., W. B. Peck & W. H. Whitcomb. 1967. Notes on spider parasites, including a reference list. J. Kansas Entomol Soc., 40:422-434. Fitzgerald, J. D., M. G. Solomon & R. A. Murray. 1986. The quantitative assessment of arthropod RAMIREZ-NATURAL HISTORY OF LUTICA 17 predation rates by electrophoresis. Ann. Appl. Biol., 109:491-498. Gertsch, W. J. 1961. The spider genus Lutica. Senck- enbergia Biol., 42:365-374. Gertsch, W. J. 1979. American Spiders. 2nd ed. Van Nostrand Reinhold, New York. Glick, P. A. 1939. The distribution of insects, spiders and mites in the air. U. S. Dept. Agric. Tech. Bull., 673:1-150. Greenstone, M. H. 1977. A passive haemagglutina- tion inhibition assay for the identification of stom- ach contents of invertebrate predators. J. Appl. Ecol., 14:457-464. Harkness, R. D. 1 977. The building and use of “shel- ters” by a hunting spider ( Zodarium frenatum Simon) in Greece. Zool. Anz., 199:161-163. Hayes, M. P. & C. Guyer. 1981. The herptofauna of Ballona. Pp. H1-H80, In Biota of the Ballona Re- gion, Los Angeles County, California. (R. W. Schrei- ber, ed.). Report to the Los Angeles Co. Dept, of Regional Planning. Los Angeles Co. Mus. Nat. Hist., Los Angeles. Hewitt, J. 1914. Descriptions of new Arachnida from South Africa. Rec. Albany Mus., 3:1-37. Jocque, R. 1991. A generic revision of the spider family Zodariidae (Araneae). Bull. American Mus. Nat. Hist., 201:1-160. Jocque, R. 1993. “We’ll meet again”, an expression remarkably applicable to the historical biogeogra- phy of Australian Zodariidae (Araneae). Mem. Queensland Mus., 33:561-564. Marx, G. 1891. A contribution to the knowledge of North American spiders. Proc. Entomol. Soc. Washington, 2:28-37. Murray, R. A. & M. G. Solomon. 1978. A rapid technique for analyzing diets of invertebrate pred- ators by electrophoresis. Ann. Appl. Biol., 90:7-10. Nagano, C. D. 1981. California coastal insects: An- other vanishing community. Terra, 19:27-30. Nagano, C. D., C. L. Hogue, R. R. Sneiling & J. L. Donahue. 1981. The insects and related terrestrial arthropods of Ballona. Pp. E1-E89, In Biota of the Ballona Region, Los Angeles County, California. (R. W. Schreiber, ed.). Report to the Los Angeles Co. Dept. Regional Planning. Los Angeles County Mus. Nat. Hist., Los Angeles. Powell, J. A. 1981. Endangered habitats for insects: California coastal sand dunes. Atala, 6:41-55. Ramirez, M. G. 1988. Evolution and historical bio- geography of the spider genus Lutica (Araneae: Zo- dariidae) (abstract). American Zool., 28:10A. Ramirez, M. G. 1990. Natural history, population genetics, systematics and biogeography of the spider genus Lutica (Araneae: Zodariidae). Ph. D. disser- tation, Univ. California, Santa Cruz. Ramirez, M. G. & R. D. Beckwitt. In press. Phytog- eny and historical biogeography of the spider genus Lutica (Araneae, Zodariidae). J. Arachnol., 00:000- 000. Robinson, M. H. 1982. The ecology and biogeogra- phy of spiders in Papau New Guinea. Pp. 557-581, In The Biogeography and Ecology of New Guinea. (J. L. Gressit, ed.). Dr. W. Junk Publishers, The Hague. South wood, T. R. E. 1978. Ecological Methods. 2nd ed. Methuen & Co., London. Thompson, M. 1973. The spider genus Lutica. Whit- tier Narrows Nature Center Quart., 1:1-5. Wausbauer, M. S. & L. S. Kimsey. 1985. California spider wasps of the subfamily Pompilinae (Hyme- noptera: Pompilidae). Bull. California Insect Surv., 26:1-130. Manuscript received 5 January 1995, revised 6 June 1995. 1995. The Journal of Arachnology 23:1 18-124 REDESCRIPTION OF THE PENNSYLVANIAN TRIGONOTARBID ARACHNID LISSOMARTUS PETRUNKEVITCH 1949 FROM MAZON CREEK, ILLINOIS Jason A. Dunlop: Department of Earth Sciences, University of Manchester, Manchester M 1 3 9PL, UK ABSTRACT. The holotypes of the trigonotarbids Lissomartus carbonarius (Petrunkevitch 1913) and Lisso- martus schucherti (Petrunkevitch 1913) (Arachnida, Trigonotarbida) from the Pennsylvanian (Westphalian D) of Mazon Creek are redescribed. These forms may be synonymous, representing male/female or juvenile/adult dimorphs, but the two species are retained at present. A new reconstruction of Lissomartus schucherti is presented. A new family, Lissomartidae, is proposed for these species based on a combination of their lack of opisthosomal tuberculation and their opisthosomal segmentation pattern of tergites 2 + 3 fused and tergite 9 divided into median and lateral plates. Lissomartidae new family may be intermediate between Trigonotarbidae and Eophryn- idae + Aphantomartidae. Trigonotarbid arachnids ranged from the Up- per Silurian (PfidoH) (Jeram et al. 1990) to the Lower Permian (Asselian?) (Scharf 1924). Su- perficially spider-like animals, they lack silk-pro- ducing spinnerets, and are characterized by an opisthosoma with tergites divided into median and lateral plates. Trigonotarbids have been placed in the arachnid taxon Tetrapulmonata Shultz 1990, as the plesiomorphic sister group of the orders Araneae, Amblypygi, Uropygi and Schizomida (Shear et al. 1987). Trigonotarbids are most numerous in the coal deposits of North America and Europe and two specimens from Mazon Creek are redescribed here and inter- preted as cursorial predators on other arthro- pods. These specimens represent two species in a single genus, which is placed in a new family. PREVIOUS WORK One of the most productive areas for trigon- otarbid fossils is the Pennsylvanian (Westphal- ian D) locality of Mazon Creek, Illinois. Petrun- kevitch (1913) described two new arachnids from Mazon Creek (in what was then the order An- thracomarti): Trigonotarbus schucherti and Tri- go not arbus carbonarius. The genus Trigonotar- bus Pocock 1911 was rediagnosed by Petrunke- vitch ( 1 9 1 3) as trigonotarbids having a triangular carapace, lacking ornamentation, with a raised median region. Petrunkevitch (1913) differenti- ated T. schucherti from T. carbonarius and Po- cock’s type species, T. johnsoni from the West- phalian B of the British Middle Coal Measures, on account of the coxae touching along the mid- line in T. shucherti and coxae separated by a sternum in the other two species. He differenti- ated T. carbonarius from T. johnsoni by the shape of the stemite surrounding the anal operculum (a structure now interpreted as a pygidium, see below). In 1949 Petrunkevitch created a new genus, Lissomartus, for T. schucherti and T. carbon- arius. He created a new family, Trigonotarbidae, for T. johnsoni, but placed his new genus Lis- somartus in the family Trigonomartidae, a sub- stitute name for the family Aphantomartidae, proposed earlier by Petrunkevitch (1945). Pe- trunkevitch (1949) also created the order Tri- gonotarbi for some of the anthracomartid ma- terial, including Lissomartus (see Shear et al. (1987) for a discussion). Petrunkevitch (1949) diagnosed the family Trigonotarbidae as having an eight-segmented opisthosoma with the terminal tergite not divid- ed into median and lateral plates, while the Tri- gonomartidae was diagnosed as having an eight- segmented opisthosoma with a terminal tergite which was divided into median and lateral plates. It was on these grounds that Lissomartus, with a divided terminal tergite, was placed in the Tri- gonomartidae. The genus Lissomartus was de- fined by Petrunkevitch (1949) as trigonomartids with a smooth carapace and opisthosoma, the carapace being subtriangular, longer than wide and concave on each side anteriorly. This inter- pretation and systematic placement was retained by Petrunkevitch (1953, 1955) in his two further major reviews of the Trigonotarbida. The family 118 DUNLOP — THE TRIGONOTARBID LISSOMARTUS 19 name Trigonomartidae was rejected in favor of the original name, Aphantomartidae, by Selden & Romano (1983). A reappraisal of L. carbon- arius and L. schucherti was deemed necessary in the light of misinterpretations in Petrunkevitch’s morphological and taxonomic work (e. g., Selden & Romano 1983; Shear et al. 1987). METHODS The holotypes of Lissomartus schucherti (Pea- body Museum, Yale University (YPM), speci- men no. 169), and L. carbonarius (United States National Museum, Washington DC (USNM), specimen no. 37978) were whitened with am- monium chloride and studied under a binocular microscope. Drawings were prepared with the aid of a camera lucida. Both specimens are from Mazon Creek, Illinois, USA, which is dated at Pennsylvanian (Westphalian D) in age (see Ni- tecki 1979 for a geological interpretation of this locality). The holotype (British Museum, Natural His- tory (BMNH) In 31239), of Trigonotarbus John - soni and other specimens of this species were studied as the type and only species of the family Trigonotarbidae. The holotypes of Aphantomar- tus areolatus (British Geological Survey (GSM) 250 1 6-7) and Trigonomartus pustulatus (USNM 37984), were studied as representatives of the family Aphantomartidae. The holotype of Eo- phrynus prestvicii from the (Lapworth Museum, Birmingham University, UK (BU) 699) was studied as a representative of the Eophrynidae. MORPHOLOGICAL INTERPRETATION Both specimens of Lissomartus are preserved as external molds in clay-ironstone nodules. YPM 169 (Figs. 1-4) consists of part and counterpart showing the dorsal and ventral surfaces of the animal respectively, while USNM 37978 (Figs. 5, 6) consists of one half of a nodule only, the counterpart being unknown from the time of the original description, and shows the ventral sur- face. The carapace of YPM 169 (Figs. 1, 3, 7) shows the approximately triangular shape characteristic of many trigonotarbids, with a raised median region bearing a pair of eyes on a single tubercle. Additionally this median region also bears a pair of oval tubercles on this raised median region either side of the eye tubercle, and two less well defined tubercle pairs posterior to the eye tuber- cle, comprising a round and an elongate tubercle pair respectively. Some Devonian trigonotarbids show multi- faceted lateral eye tubercles in addition to the median eye tubercle (e. g., Shear et al. 1 987), and it is conceivable that the oval tubercles either side of the eye tubercle in YPM 169 are lateral eye tubercles, too. However, since lateral eye tu- bercles are not present in any of the other taxa interpreted as closely related to Lissomartus and would represent an uncharacteristically plesiom- orphic character in an otherwise rather derived trigonotarbid, I prefer to interpret these, with reservations, as simple tubercles (Fig. 7), as are observed in greater density on the carapaces of the eophrynids and aphantomartids. The opisthosomal morphology of the Penn- sylvanian trigonotarbids is interpreted in com- parison with the superbly preserved Devonian Rhynie chert material (Dunlop 1994). Interpre- tation of the Rhynie chert material indicates that trigonotarbids have an opisthosoma of 12 seg- ments with 9 dorsal tergites, the first of which is modified into a locking ridge which tucks under the carapace and is often very small (Dunlop 1994). Tergites 2 and 3 are fused into a single macrotergite in most trigonotarbids (Selden & Romano 1983; Shear et al. 1987). The last two segments (11, 12) are ring-like and form a py- gidium, with segment 10 forming a plate, not divided into tergites and stemites, surrounding this pygidium (Figs. 4, 7). Ventrally, stemite 1 is interpreted as being ab- sent in trigonotarbids (Dunlop unpubl. data). In comparison with Recent tetrapulmonate arach- nids (Shultz 1993) ’stemites’ 2 and 3 (the two anteriormost ventral sclerites in trigonotarbids) probably represent highly derived sutured-on lung-bearing appendages and are termed the an- terior and posterior operculae respectively (Shultz 1993). Stemite 4 is therefore the first visible tme stemite in trigonotarbids. Applying this inter- pretation to Lissomartus , its dorsal opisthosomal segmentation (Figs. 1, 3, 7) shows a first tergite without lateral plates, interpreted as the locking ridge which would have tucked under the cara- pace in life, and then subsequent divided tergites indicating a fused macrotergite 2 + 3. Tergite 9 is divided, but the division is not as strong as on the preceding tergites. Ventrally, in YPM 169 (Figs. 2, 4), there is a raised, bilobed structure apparently on the an- terior operculum. This is unusual among trigon- otarbids, which normally bear a similar raised structure on the posterior operculum. The bi- lobed structure is interpreted as being homolo- 1 20 THE JOURNAL OF ARACHNOLOGY Figures 1, 2.— The holotype of Lissomartus schucherti (Petrunkevitch 1913) (YPM 169). From the Pennsyl- vanian (Westphalian D) of Mazon Creek, Illinois, USA. 1, Part showing dorsal surface; 2, Counterpart showing ventral surface. Scale: 5 mm. Figures 3, 4.— Interpretative drawing of the specimen shown in Figures 1 and 2. 3, Dorsal surface; 4, Ventral surface. Cp = carapace, Et = eye tubercle, T = tergite with number, Lr = locking ridge, S = stemite with number, A. op = anterior operculum, P. op = posterior operculum, Vs? = ventral sacs?, Py= pygidium, Ch = chelicerae, L = walking leg with number, Pl= pedipalp, Cx = coxae, Tr = trochanter, Fe = femur, Pa = patella, Ti = tibia, Mt = metatarsus, Ts = tarsus, St = sternum. Scale: 5 mm. DUNLOP — THE TRIGONOTARBID LISSOMARTUS 121 Figure 5. —The holotype of Lissomartus carhonarius (Petrunkevitch 1913) (USNM 37978). From the Penn- sylvanian of Mazon Creek, Illinois, USA. Ventral sur- face only. Scale: 5 mm. gous with structures seen in some Recent arach- nids called ventral sacs whose function is obscure (Dunlop 1994), rather than a genital organ as Petrunkevitch (1949) suggested. However, it is worth noting that male amblypygids have a pair of gonopodi in this position associated with the genitalia (W. Shear, pers. comm.). The presence of a structure on the anterior operculum raises some doubts about the interpretation of the seg- mentation in this animal, but there is no visible segment in front of the anterior operculum and the overall segmentation pattern favors inter- preting these structures as belonging to the an- terior operculum. Whether they are ventral sacs or genitalia is impossible to determine, but since the genitalia of many Recent tetrapulmonates are concealed beneath the anterior operculum I fa- vor their interpretation as ventral sacs. Both specimens show a distinct deepening of the posterior opisthosoma posteriorly from the middle of stemite 5 (Figs. 2, 4-6). This could give the animal a relatively flat, narrow anterior opisthosoma with a deeper, bowl-like posterior opisthosoma (Fig. 7) in lateral view. The division between the ninth stemite and the tenth segment (not divided into a tergite and stemite) is present but poorly defined. Segment 10 surrounds a two- segmented pygidum. This stmcture is therefore not an anal operculum as interpreted by Petrunk- evitch (1949). 6 Figure 6.— Interpretative drawing of the specimen shown in Figure 5. Abbreviations as in Figures 3 and 4. Scale: 5 mm. The reconstruction of Lissomartus schucherti (Fig. 7) is based on YPM 1 69, with USNM 37978 (L. carhonarius) being used primarily for the coxo- stemal region. The claws and distribution of se- tae are hypothetical and based on the well-pre- served Devonian trigonotarbids (e. g., Shear et al. 1987) and comparisons with Recent arach- nids. The Lissomartus species are relatively large trigonotarbids and can be visualized as either ambushing or running down small arthropods on the floor of the coal forests. SYSTEMATIC PALEONTOLOGY Order Trigonotarbida Petrunkevitch 1949 Family Lissomartidae new family Type and only known genus. —Lissomartus Pe- trunkevitch, 1949. Diagnosis.— Trigonotarbids with a medially raised carapace bearing a pair of eyes on a me- dian tubercle. Carapace relatively smooth, but with slight lateral lobation and medial tubercu- lation. Opisthosoma smooth with tergite 1 pres- ent as a locking ridge, tergites 2 + 3 fused and tergite 9 divided into median and lateral tergites. 122 THE JOURNAL OF ARACHNOLOGY Stemite 5 large, with the opisthosoma deepening posteriorly. Discussion.— Lissomartus does not show the deep carapace lobation and heavily tuberculated dorsal surface which characterizes trigonotarbids such as Aphantomartus (e. g., Pocock 1911; Pe- tmnkevitch 1953; Selden & Romano 1983). On these grounds I reject Petrankeviteh’ s (1949) placement of Lissomartus in the family Aphan- tomartidae (his Trigonomartidae). Lissomartus is clearly related to Trigonotarbus (e. g., Pocock 1911; Petrankeviteh 1949) on ac- count of its overall carapace shape and lack of strong tuberculation. However, Lissomartus can be differentiated from Trigonotarbus by its car- apace ornamentation and opisthosomal segmen- tation. Specifically, Lissomartus shows fused ter- gites 2 + 3, a divided tergite 9 and unfused (bare- ly) stemite 9 and segment 10 whereas Trigono- tarbus has an unfused 2 + 3, an undivided tergite 9 and stemite 9 fused to segment 10 (unpubl. obs.). On these grounds I also reject Petrunk- evitch’s (1913) placement of Lissomartus in the Trigonotarbidae. Since Lissomartus cannot be placed in any existing family I am creating a new, monotypic family, Lissomartidae, to accom- modate the genus. This family is known only from the Westphalian D of Mazon Creek. Opisthosomal segmentation and ornamenta- tion patterns appear to be useful characters, vis- ible in most specimens, on which to base higher taxa in trigonotarbids. The Lissomartidae are clearly related to T. johnsoni in terms of their carapace shape and opisthosomal smoothness. Eophrynidae and Aphantomartidae are prob- ably sister groups, sharing a deeply lobed cara- pace and a heavily tuberculated dorsal surface. Lissomartidae may represent the plesiomorphic sister group of Eophrynidae + Aphantomartidae (with Trigonotarbidae perhaps the sister group to all three) since they do not have the, presum- ably derived, heavy tuberculation, but share with Eophrynidae + Aphantomartidae a division of tergite 9 (perhaps not fully complete in Lisso- martus). There is also the slight lobation of the carapace, reminiscent of that in aphantomartids and eophrynids, and the drawing out of the an- terior carapace of Lissomartus , similar to the pointed anterior spine of eophrynids. Genus Lissomartus Petrankeviteh 1 949 Type species. —Lissomartus schucherti (Pe- trunkevitch 1913). Included species.— L. schucherti, L. carbon- arius. Diagnosis.— As for the family. Lissomartus schucherti (Petrankeviteh 1913) Figs. 1-4, 7 Trigonotarbus schucherti Petrankeviteh 1913: 106, 107, figs. 63, 64, PL 10, figs. 53, 54. Lissomartus schucherti (Petrankeviteh). Petrankeviteh 1949: 257. Lissomartus schucherti (Petrankeviteh). Petrankeviteh 1953: 94. Lissomartus schucherti (Petrankeviteh). Petrankeviteh 1955: 113, fig. 80 (2a, b). Type.— Holotype and only known specimen YPM (169), part and counterpart. From the Pennsylvanian (Westphalian D) of Mazon Greek, Illinois. Diagnosis.— Lissomartids with a raised, bi- lobed structure of the anterior operculum. Ven- trally, anterior sclerites not pointed on the mid- line. Description.— Holotype 19.0 mm long; cara- pace 7.9 mm long, basal width 6.5 mm. Opis- thosoma 11.1 mm long with maximum width 9.0 mm. Carapace relatively flat, subtriangular, drawn anteriorly into a long, blunt point. Cara- pace with medial raised area bearing a pair of eyes on a tubercle, 3.0 mm from the front of the carapace. Slight raised nodes either side of, and posterior to, the eye tubercle, otherwise carapace smooth, but slightly lobed either side of the raised median region. Sternum present, but slightly displaced and not distinct in the fossil Coxae subtriangular, be- coming progressively larger posteriorly. Tro- chanters approximately as long as wide. Chelic- erae present, but indistinct. Other appendages relatively long and slender with a slight granular texture to the cuticle. Pedipalp shows an oblique articulation to the trochanteraafemur joint. Fo- domere lengths (in mm): Palp: Fe 2.9, Pa 2.7, Ti 2.0, Ts 2.9. Leg 1: Ti? 3.7. Leg 2: Fe 4.0, Pa 2.8, Ti 3.7. Leg 3: Fe 3.8, Pa 2.9, Ti 3.7. Leg 4: Fe 5.7, Pa 3.4, Ti 4.1, Mt 1.9, Ts 2.1 mm (abbre- viations as in Figs. 3, 4). Prosoma and opisthosoma slightly disarticu- lated in this fossil. Opisthosoma rounded, left hand margin being absent and the right hand tergites being obscured along their lateral mar- gins by poorly defined, superimposed stemites. With the exception of tergite 1 , tergites divided into median and lateral plates, median plates be- coming narrower posteriorly. Division of tergite DUNLOP T HE TRIGONOTARBID LISSOMARTUS 123 Figure 7.— Reconstruction of Lissomartus schucherti in dorsal, ventral and lateral view. Scale: 5 mm. 9 into median and lateral plates weaker than in the preceding tergites. Tergite lengths (in mm): 1: 0.7, 2 +3: 1.2, 4: 1.5, 5: 1.2, 6: 1.5, 7: 1.4, 8: 1.4, 9: 2.2. Ventrally, anterior sclerites are abbreviated, but are followed by large stemite 5. Anterior operculum bears a raised, bilobed structure on posterior margin. Stemite 5 bears a transverse division (not a segmental division) demarcating a deepening of the opisthosoma posterior to the division. Faint longitudinal folds on the ventral opisthosoma. Ventral sclerites lengths (in mm): anterior operculum: 0.7, posterior operculum: 1.2, stemite 4: 0.4, 5: 2.4, 6: 1.5, 7: 1.3, 8: 1.0, 9: 0.9. Pygidium diameter 0.9 mm. Lissomartus carbonarius (Petmnkevitch 1913) Figs. 5, 6 Trigonotarbus carbonarius Petmnkevitch 1913: 107, 8, fig. 65, PI. 10, fig. 55. Lissomartus carbonarius (Petmnkevitch). Petmnke- vitch 1949: 257. Lissomartus carbonarius (Petmnkevitch). Petmnke- vitch 1953: 94. Type.— Holotype and only known specimen, USNM 37978, one piece. From the Pennsylva- nian (Westphalian D) of Mazon Creek, Illinois. Diagnosis. — Lissomartids with no raised, bi- lobed structure on the anterior operculum. Ven- trally, anterior sclerites pointed anteriorly on the midline. Description.— Holotype 16.3 mm long; ven- tral opisthosoma 9.7 mm long maximum width 7.2 mm. Coxo-stemal region well preserved and shows a sternum, bluntly pointed at either end. Leg 4 coxae attach posterior to sternum, leg cox- ae 2 and 3 slot into recesses in sternum and leg coxae 1 attach anterior to sternum. Chelicerae present and wedge-shaped in ventral view and with the palpal coxae either side of them they define a small preoral cavity. Femur of leg 4 present and 5.2 mm long. Additional limbs ab- sent. The prosoma and opisthosoma are slightly dis- articulated in this fossil. Anterior segmentation of the opisthosoma clearly shows the abbreviated anterior sclerites pointed anteriorly on the mid- line and the large 5th stemite behind them. Lengths (in mm): anterior operculum: 0.4, pos- terior operculum: 0.7, stemite 4: 0.9, 5: 2.1, 6: 1.5, 7: 1.2. Bilobed structure, as in the anterior region of YPM 169, absent, but the deepening of the opisthosoma marked by a transverse di- vision of stemite 5 more pronounced than in 124 THE JOURNAL OF ARACHNOLOGY YPM 169. Stemites posterior to this become in- creasingly poorly defined. Lateral and posterior margins of opisthosoma presumed absent since the pygidium cannot be seen. Lateral margins of the opisthosoma show evidence of folding or wrinkling of the cuticle. Remarks.— Lissomartus schucherti and Lis - somartus carbonarius are very similar fossils and there is a strong possibility that they are syn- onymous. In this case L. carbonarius would be referred to L. schucherti (the first of the two spe- cies mentioned by Petrankevitch (1913)). The minor differences between these fossils could be the result of sexual dimorphism and/or ontoge- ny, as was suggested by Dunlop (1994) for the trigonotarbid Pleophrynus verrucosa. Differences in the anterior opisthosomal (genital) region are recorded within species of Amblypygi and Uro- pygi (W. Shear, pers. comm.) and there could be a ’straightening’ of the anterior sclerites between L. carbonarius and L. schucherti due to sexual maturation. However, since there are real morphological differences between the two monotypic species (the lack of a raised bilobed structure and the shape of the anterior ventral sclerites in the smaller L. carbonarius) I prefer to retain the spe- cies distinction with the reservations noted above; the dimorphic interpretation of Pleophrynus above was based on a wide range of specimens. Possibly, future finds of Lissomartus will give a clearer picture of intraspecific variation and clar- ify the position of these species. ACKNOWLEDGMENTS I thank I. Thompson (USMM) and R. D. White (YPM) for the loan of material in their care and R. Fortey and the staff of the (BMNH), S. Tun nicliffe (GSM) and P. Smith (BU) for their hos- pitality on my visits there and for the loan of material. I also thank P. Selden and W. Shear for criticisms of earlier versions of the manu- script and L. Anderson for useful discussions. This work was carried out under a UK Natural Environment Research Council studentship into early terrestrial ecosystems. LITERATURE CITED Dunlop, J. A. 1994. The palaeobiology of the Writh- lington trigonotarbid arachnid. Proc Geol Assoc., 105:287-296. Jeram, A. J., P. A. Selden & D. Edwards. 1990. Land animals in the Silurian: arachnids and myriapods from Shropshire, England. Science, 250: 658-661. Nitecki, M H. (ed.). 1979. Mazon Creek Fossils. Ac- ademic Press, New York. Petrankevitch, A. I. 1913. A monograph of the ter- restrial Palaeozoic Arachnida of North America. Trans. Connecticut Acad. Arts Sci, 18: 1-1 37. Petrankevitch, A. I. 1945. Palaeozic Arachnida of Illinois. An enquiry into their evolutionary trends. Illinois State Museum, Scientific Papers 3:1-72. Petrankevitch, A. I. 1949. A study of the structure, classification and relationships of the Palaeozoic Arachnida based on the collections of the British Museum. Trans. Connecticut Acad. Arts Sci., 37: 69-315. Petrankevitch, A. I. 1953. Paleozoic and Mesozoic Arachnida of Europe. Mem. Geol. Soc. America, 53:1-122. Petrankevitch, A. I. 1955. Arachnida. Pp. 44-175, In Treatise on Invertebrate Paleontology, Pt. P, Ar- thropoda 2. (R. C. Moore, ed.). Geol Soc. America and Univ. Kansas Press, Lawrence. Pocock, R. I. 1911. A monograph of the terrestrial Carboniferous Arachnida of Great Britain. Mono. Palaeontograph. Soc., London, 1-84. Scharf, W. 1924. Beitrag zur Geologic des Steink- holengebiets im Sudharz. Jb. halle. Verb. Erforsch. mitteldt. Bodenschatze, 4:404-437. Selden, P. A. & M. Romano. 1983. First Palaeozoic arachnid from Iberia: Aphantomartus areolatus Po- cock (basal Stephanian; prov. Leon, N. W. Spain), with remarks on aphantomartid taxonomy. Boln. Geol Min., 94:106-112. Shear, W. A., P. A. Selden, W. D. I. Rolfe, P. M. Bonamo & I. D. Grierson. 1987. New terrestrial arachnids from the Devonian of Gilboa, New York (Arachnida, Trigonotarbida). American Mus. Nov., 2901:1-74. Shultz, J. W. 1993. Muscular anatomy of the giant whipscorpion Mastigoproctus giganteus (Lucus) (Arachnida: Uropygi) and its evolutionary signifi- cance. Zool J. Linn. Soc., 108:335-365. Manuscript received 20 January 1995, revised 3 April 1995. 1995. The Journal of Arachnology 23:125-126 RESEARCH NOTES A NEW SYNONYM IN THE GENUS META (ARANEAE, TETRAGNATHIDAE) The orb-weaving spider Meta menardii (La- treille 1804) has been treated in the past as oc- cupying much of the holarctic region (Bonnet 1957; Roewer 1942; Levi 1980). Recently, how- ever, Marasik & Koponen (1992) concluded that three allopatric species are involved, namely, M. menardii in Europe, M. manchurica Marasik & Koponen 1992 in the Russian Far East, and M. americana Marasik & Koponen 1992 in eastern North America. A fourth related species, M. ja- ponica Tanikawa 1993, was subsequently de- scribed. My purpose here is to show that there is an available name older than americana for the eastern North American species. Gertsch (1933) described a species of cave spi- der from Indiana, proposing for it the new genus Auchicybaeus and placing it in the family Age- lenidae. Soon recognizing his error regarding the generic and familial placement, he synonymized the name Auchicybaeus under Meta. He also syn- onymized the specific name ovalis under men ardii (Gertsch & Ivie 1936). Bonnet (1957) re- corded the synonymy, but Roewer (1942) and Levi (1980) did not. I have examined Gertsch’s type specimen and have found it to match specimens of the eastern North American species formerly known as M. menardii and lately described as M. americana . The size, color pattern, cheliceral dentition, eye relations, leg trichobothria, and details of the epi- gynum and spermathecae of the type lead me to conclude that ovalis and americana represent a single species. The older name must be used, as formalized below. Meta ovalis (Gertsch) Meta menardii Emerton 1875:129 (part); Gertsch & Ivie 1936:20; Bonnet 1957:2787 (-/) (part); Levi 1980: 42, figs. 1 12-127, map 5; Platnick 1993:376 (-i) (part). Not menardii Latreille 1804. Auchicybaeus ovalis Gertsch 1 933: 1 1 , fig. 15. Holotype female from Marengo Spring Cave, Crawford Coun- ty, Indiana, 20 October 1911 (Arthur W. Henn), deposited in the American Museum of Natural His- tory, New York. Examined. Meta americana Marasik & Koponen 1992:138, figs. 1-4, 14. Holotype male from northeast of Jamison, Pennsylvania, June 1944 (W. Ivie), deposited in the American Museum of Natural History, New York, not examined. Male and female paratypes from Eganville, Ontario, 1 2 June 1 972 (S. Peck), deposited in the Canadian National Collection of Insects and Arachnids, Ottawa, Ontario, examined. NEW SYN- ONYM. Note: Platnick (1993:376) first used the combination Meta ovalis (Gertsch 1933). ACKNOWLEDGMENT I am grateful to Norman Platnick for lending me the type specimen of Auchicybaeus ovalis, LITERATURE CITED Bonnet, P. 1957. Bibliographia Araneoram. Vol 2, pt. 3, pp. 1927-3026. Douladoure, Toulouse. Emerton, J. H. 1875. Spiders common to New En- gland and Europe. Psyche, 1 : 1 29-13 1 . Gertsch, W. J. 1933. Diagnoses of new American spiders. American Mus. Nov it:., 637. 14 pp. Gertsch, W. J. & W. Ivie. 1936. Descriptions of new American spiders. American Mus. Novit, 858. 25 PP- Levi, H. W. 1 980. The orb- weaver genus Mecynogea, the subfamily Metinae and the genera Pachygnatha , Glenognatha and Azilia of the subfamily Tetrag- nathinae north of Mexico (Araneae: Araneidae). Bull. Mus. Comp. Zook, 149:1-75. Marasik, Y. M. & S. Koponen. 1992. A review of Meta (Araneae, Tetragnathidae), with description of two new species. J. Arachnol., 20:137-143. Platnick, N. L 1993. Advances in spider taxonomy 1988-1991, with synonymies and transfers 1940- 1980. New York Entomol. Soc. and American Mus. Nat. Hist., New York, New York. 846 pp. 125 126 THE JOURNAL OF ARACHNOLOGY Roewer, C. F. 1942. Katalog der Araneae von 1758 Agri-Food Canada, Ottawa, Ontario K1A0C6, bis 1940, bzw. 1954. Naturkunde und exakte Wis- Canada, senschaften Paul Budy, Bremen. Vol. 1, pp. 1-1040. Manuscript received 1 December 1994, revised 22 March 1995. Charles D. Dondale: Centre for Land and Bi- ological Resources Research, Agriculture and 1995. The Journal of Arachnology 23:127-129 THE NEW SPECIES PURUMITRA A USTRALIENSIS (ARANEAE, ULOBORIDAE) WITH NOTES ON ITS NATURAL HISTORY The genus Purumitra Lehtinen 1967 was pre- viously known only from specimens of Purum- itra grammicus (Simon 1893) collected on the Phillipine island of Luzon and the Caroline is- land of Ponape (Opell 1979). This paper de- scribes a second species of Purumitra that is found on continental islands of the Great Barrier Reef located off the east coast of Queensland, Austra- lia. For nomenclatural purposes, B. D. Opell is designated the author of this new species’ name. We thank the Queensland Department of En- vironment and Heritage for permission to work on the islands and transportation to some of them. T. W. Schoener thanks the University of Queens- land and Professor J. Kikkawa for arranging a visiting professorship, and the John Simon Gug- genheim Foundation for the fellowship support- ing this research. Robert Bennet made useful comments on the manuscript. Purumitra australiensi new species Figs. 1-5, Tables 1, 2 Types.— Female holotype and paratype from Pelican Island (nr. Brampton Island), 30 Septem- ber 1992 (T. Schoener, S. Keen); male paratype from Cow Island (near Whitsunday Island), 17 October 1992 (T. Schoener, S. Keen); in Queens- land Museum (see Fig. 1). The epithet of this species is an adjective derived from its known distribution. Diagnosis.— Purumitra australiensis is similar to P. grammicus in size, coloration, and general appearance (Figs. 2, 3; fig. 159 in Opell 1979). Female P. grammicus has an epigynum with a pair of lateral crypts and a median crypt that is subdivided by sclerotized ridges into a pair of anterior and a pair of posterior atria (fig. 1 60 in Opell 1979). In contrast, the epigynum of P. aus- traliensis has a pair of lateral crypts and a large, undivided median crypt (Fig. 5). The male pal- pus of P. grammicus has a median apophysis bulb (MAB) whose central depression is com- pletely divided by a narrow sclerotized ridge into a small region that is adjacent to the median apophysis spur (MAS) and a larger region near the MAB’s dorsal surface (fig. 1 57 in Opell 1 979). The width of this species’ concave median apophysis spur (MAS) is 0.7 x its length. In con- trast, the MAB of a palpus of P. australiensis (Fig. 4) has a central depression incompletely divided by a short sclerotized ridge into a large region ventral to the MAS (above the MAS in Fig. 4) and a small region near the base of the MAB. In P. australiensis the width of the concave MAS is only 0.4 x its length. Description.— Table 1 gives measurements of male and female specimens. As shown in Fig. 2, the carapace of a female is dark gray with median Figure 1.— The east coast of Queensland, Australia, showing the location of the islands on which specimens of Purumitra australiensis new species were collected. 127 128 THE JOURNAL OF ARACHNOLOGY Figures 2-5.— Purumitra australiensis new species. 2, dorsal view of female holotype; 3, dorsal view of male paratype; 4, retrolateral view of left palpal genital bulb of male paratype (MAB = median apophysis bulb, MAS = median apophysis spur); 5, ventral view of epigynum of female holotype. and lateral tan stripes. The carapace also has a narrow, dark gray border not easily seen in dorsal view. Sternum dark gray. Chelicerae with a pair of narrow, black, dorso-ventral stripes. First legs gray with dorsal white stripe extending full length of femur, proximal white ring on metatarsus, and faint tan ring at center of tibia. Abdomen in dor- sal view (Fig. 2) white with mottled gray sides; in lateral view mottled gray with a broad white stripe extending nearly its full length; in ventral view dark gray with a narrow, broken, median tan stripe and a pair of narrow, paraxial tan stripes. Male coloration (Fig. 3) similar to that of female except that the sternum has a light gray center and a dark margin and the first tibia is uniformly gray. Natural history.— All specimens were collect- ed from horizontal orb webs among understory/ edge vegetation (including ferns) within forest very near the shoreline. Webs were located near the ground; all had a stabilimentum (Table 2). Webs of juveniles were nearer the ground than those of mature females and were about as likely Table 1.— Measurements in mm of female holotype and male paratype of Purumitra australiensis new spe- cies. Female Male Total length 2.52 2.28 Carapace length 0.84 0.84 Maximum carapace width 0.72 0.72 Clypeus height 0.03 0.01 AME, ALE diameter 0.08, 0.05 0.08, 0.05 PME, PLE diameter 0.08, 0.06 0.06, 0.06 AME, PME separation 0.08, 0.14 0.08, 0.11 Sternum length 0.56 0.48 Maximum sternum width 0.44 0.36 Coxa-trochanter I 0.36, 0.36 0.32, 0.28 Femur I, IV 1.44, 0.96 1.32, 0.76 Patella I, IV 0.36, 0.28 0.32, 0.24 Tibia I, IV 1.24, 0.80 1.20, 0.64 Metatarsus I, IV 1.24, 0.70 1.12, 0.56 Tarsus I, IV 0.64, 0.56 0.56, 0.44 Calamistrum length 0.38 — Abdomen length 2.00 1.60 Maximum abdomen width 0.96 0.64 Cribellum width 0.20 - RESEARCH NOTES 129 Table 2.— Web placement and web features of Purumitra australiensis new species. Juveniles Adult females Height above ground in cm (mean ± 1 SD, n) Stabilimentum type: 20.3 ± 4.8, 5 27.3 ± 7.6, 4 No. with linear/ No. with circular 3/2 1 / 4 to have linear as circular stabilimenta. Webs con- structed by adult females usually had circular stabilimenta. Distribution.— Mature specimens were col- lected on the following continental islands from September 30-November 28 1992 (Fig. 1): Large Mausoleum (Newry Island Group, near Cape Hillsborough), Pelican Island (near Brampton Is- land), Cole and Cow Islands (both in Whitsunday Island area), Normanby Island (Frankland Group, south of Cairns), Garden Island (south of Family Islands). LITERATURE CITED Lehtinen, P. T. 1967. Classification of the cribellate spiders and some allied families, with notes on the evolution of suborder Araneomorpha. Ann. Zool. Fennici, 4:199-468. Opell, B. D. 1979. Revision of the genera and tropical American species of the spider family Uloboridae. Bull. Mus. Comp. Zool, 148:443-549. Simon, E. 1893. Arachnides. In Voyage de M. E. Simon aux iles Philippines (mars et avril 1 890). 6th Memoire. Ann. Soc. Ent. France, 62:65-80. Brent D. Opell: Department of Biology, Vir- ginia Polytechnic Institute and State Univer- sity, Blacksburg, Virginia 24061 USA. Thomas W. Schoener and Susan L. Keen: Sec- tion of Evolution and Ecology, Division of Biological Sciences, Storer Hall, University of California at Davis, Davis, California 95616 USA. Valerie T. Davies: Queensland Museum, P.O. Box 3300, South Brisbane, Queensland 4101, Australia. Manuscript received 22 March 1995. 1995. The Journal of Arachnology 23:130-133 A GYNANDROMORPHIC SCHIZOCOSA (ARANEAE, LYCOSIDAE) Individuals in which both sexes are discretely combined are termed gynandromorphs, whereas intersexuality is a condition in which portions of a body are intermediate between the sexes and are not clearly one sex or the other (White 1973; Roberts & Parker 1973). The earliest reference to a spider gynandromorph is given by Blackwall 1867 (cited in Bonnet 1945). Bonnet lists nu- merous other early citations of gynandromorphs. Roberts & Parker (1973) provide a classification of 1 4 types of gynandromorphs, which are com- binations of lateral and transverse gynandro- morphs, although they admit that several of these could never be externally recognized. Gynandromorphs probably occur in most taxa of animals including birds (Patten 1993), as well as in many insects and arachnids (Hannah-Alava 1960; Cokendolpher & Francke 1983; Brust 1966). They have been studied extensively in Drosophila (Wilkins 1993) and as early as the 1920’s, gynanders were used as a means of fate mapping cells. Several mutations in Drosophila are particularly prone to being gynandromorphic (White 1973). Only slightly more than 50 cases of gynandromorphy and intersexuality have been reported for spiders (Hull 1918; Bonnet 1934; Kaston 1961; Roberts & Parker 1973). It occurs but is perhaps equally rare in scorpions, solpug- ids and ticks, although reduced sexual dimor- phism may make detection more difficult (Cok- endolpher Sc Sissom 1988). Clarke & Rechav (1992) note that gynandromorphs are “wide- spread” in the Ixodidae, but they do not offer any estimates on the frequency of occurrence. Kaston (1961), in his summary of the spider gynandromorphs known to that point, suggested that gynandromorphy in spiders is “exceedingly rare.” Palmgren (1979) calculated a rough esti- mate of the frequency of gynandromorphism based on 69,970 adult spiders from his own col- lection and from the Zoological Museum of the Helsinki University Collection. He described four gynandromorphs and suggested that the phe- nomenon occurs about once per 17,000 normal spiders. One of the specimens described by Palmgren was from the genus Oedothorax. Holm (1941) noted that a disproportionate number of gynandromorphic specimens are in this genus. The causes of gynandromorphy have been in- vestigated for a variety of groups (but not spiders) and generally involve the nondisjunction in the X chromosomes early in development (White 1973). For example, in a species of tick studied by Homsher & Yunker (1981), the male tissue had the number of chromosomes typical for males (22 + X) while the chromosomes in the female tissue were consistent with that of a normal fe- male of that species (22 + XX). Presumably the mechanism of formation of gynandromorphs in spiders is the same as in these other groups; how- ever, there have been no published studies of the karyotype of gynandromorphic spiders (White 1973). Kaston (196 1) suggested that the phenom- enon of gynandromorphy would be less frequent in spiders than insects (particularly Drosophila) because of the chromosomal system found in spiders. In many spiders the males are “X1X20” or “X 1X2X30,” and females “XIX 1X2X2” or “X1X1X2X2X3X3” where females have two, three, or more chromosomes than males (White 1973; Hackman 1948; Wise 1983). Kaston sug- gested that the creation of a gynandromorph from a chromosomal female zygote would involve the loss of 2-3 chromosomes, rather than just one as in Drosophila and would thus be quite a bit less frequent in most spiders than in Drosophila. Although some behaviors in a few gynandro- morphs have been noted (e. g., Coelotes atropos produced an egg sac (Kaston 1961)), the most extensive behavioral description is that of the lycosid Alopecosa pulverulenta provided by Gack & von Helversen (1976). These authors de- scribed the individual as a “lateral gynandro- morph” in which the left side was male and the right side female, except for the right palp which was described as intersexual. The ventral opis- thosoma contained male sexual organs. When the gynandromorph was placed with a normal male spider of the same species, the male did not exhibit courtship behavior. This perhaps sug- gests that the gynander lacked pheromones that are often produced by female lycosids that elicit courtship in males. The gynander never mated with the male. However, the gynandromorph spider built an egg case that contained only a gelatinous fluid and was not carried on the spin- nerets. When placed with a normal female of the same species, the gynandromorph showed court- ship behavior typical of the males of its species, mounted and inserted the male palp in a manner 130 RESEARCH NOTES 131 Figures 1, 2. — Gynandromorphic Schizocosa ocreata from Hue, Hocking County, Ohio. 1, Dorsal view of gynandromorphic Schizocosa ocreata. Right side of individual shows male palp; left side shows female palp; 2, Ventral view showing differences in coloration of legs, sternum and venter. also typical of the species. It made no attempt to insert the other (intersexual) palp. Copulation was short and involved only one insertion. (The species typically shows several insertions). The gynandromorph mated with a second female. The authors do not report if either of the females laid an egg case following this mating. The present report provides a description of a gynandromorphic Schizocosa ocreata (Hentz) discovered in the summer of 1993 and a rough estimate of the frequency of gynandromorphism in this genus. This is the first report of a gynan- dromorph in the genus Schizocosa , although oc- currences have been reported in other lycosids (Exline 1938; Holm 1941; Kaston 1961; Mackie 1969; Gack & von Helversen 1976). The specimen of Schizocosa described here was from Ohio, Hocking County, Hue. It was col- lected in a house on 25 June 1993 by Lawrence M. King III (a former undergraduate student of Jerome Rovner) who suspected that it was a gy- nandromorph. The spider was given to J. Rov- ner, who noted that it walked in a manner typical of male S. ocreata by extending and tapping the front legs (pers. comm.). Description.— By using the classification sug- gested by Roberts & Parker (1973), I would sus- pect that this specimen is a regular Type 2 gy- nandromorph, although like many gynandro- morphs, it is not perfectly symmetrical (Figs. 1 , 2). The left side of the spider is female, its right side male. The total length is 7.8 mm and is within the range for both males and females of this species (Dondale & Redner 1978 report that the size range for males of S. ocreata is 5.65-8.3 mm; for females, 7.3-10.4 mm). The carapace length is 3.64 mm, and the carapace width 2.8 mm. The right (3) side of the carapace is 1 .3 mm from edge to the middle; the left (9) side is 1.5 mm, resulting in a slightly asymmetrical cara- pace. Likewise, the chelicera on the left side is slightly larger and the fang on this side also is longer. The spinnerets reflect the same pattern: on the left (9) side, the spinnerets are larger. The right side of the animal has a fully devel- oped and apparently normal 6 palp (Fig. 3). There is a stridulatory organ located on the tibio-tarsal joint of this palp. Leg I on the right side has a brush of bristles along the tibia, as is typical for males of this species (Fig. 1). The brush extends 132 THE JOURNAL OF ARACHNOLOGY Figures 3, 4. — Gynandromorphic Schizocosa ocreata. 3, Ventral aspect of spider’s right palp, scale bar = 0.5 mm; 4, Detail of spider’s epigynal area, scale bar = 0. 1 mm. to the basitarsus. The left side has a palp resem- bling that of normal females, and on the opis- thosomal venter there is an epigynum that has a single large excavation (Fig. 4). This half of the epigynum looks normal for this species except that the median septum has an irregular border. The ventral aspect of the animal exhibits two different colorations (Fig. 2). On the spider’s right (a) side, the sternum is darker, although not in a straight line down the sternum. There is a dark band on the right side of the venter of the ab- domen, and the pattern of pigmentation on each side of the venter of the abdomen differs. The right side appears mottled, while the left side has distinct dark spots of pigment. The ventral surface of all four coxae and fem- ora on the right side is black (Fig. 2). On the left (9) side, the coxae and femora have patches of dark pigmentation but are overall much lighter. In a manner that is atypical for males of S. ocrea- ta,, the legs on the male side are uniformly dark. Curiously, the tibia of leg II has a rudimentary brush. Legs II-IV on the spider’s right side are more similar to a typical male leg I than they are to typical legs II-IV. From a dorsal view, the legs on the left side are annulated, while the legs on the right are mostly black with some lighter streaks. In most S. ocreata , legs II-IV of the males have annulations but are not uniformly dark. The venter of the abdomen has numerous spots of sclerotization that are more evident on the male side. The dorsum of the abdomen has a heart-mark and the pigmentation on the abdo- men is slightly asymmetrical. Estimation of frequency of gynandromorphs in Schizocosa . — My work for the past three years, including much done in collaboration with Gary L. Miller and Patricia R. Miller, provides a rough estimate as to the frequency of this phenomenon in the genus Schizocosa. For each of the summers of 1993 and 1994 we have maintained nearly 1000 specimens of Schizocosa and other lycosids in the laboratory for behavioral studies. We have also completed a year-long pitfall study, focusing primarily on the lycosids. Thus, we have iden- tified and/or observed behavior in close to 3000 individuals of Lycosidae (mostly Schizocosa ), and have never encountered a gynandromorph. In earlier studies done at Ohio, I raised or collected nearly an additional 2000 spiders. Thus, an es- timate of the frequency of gynandromorphs in this genus is that one may occur not more fre- RESEARCH NOTES 133 quently than once every several thousand spi- ders. The specimen is currently housed in the teach- ing collection of Jerome Rovner at Ohio Uni- versity. ACKNOWLEDGMENTS I am grateful to Lawrence M. King III for col- lecting the spider, to Jerome S. Rovner for bring- ing it to my attention, and to James C. Coken- dolpher for preparing the figures and for alerting me to some of the literature on gynandromorphs. Patricia Miller, Catherine Lamb and James Cok- endolpher all read earlier drafts of the manu- script. Financial support was from National Geographic Grant 5312-94 to G. Stratton and G. Miller. Deanna Tingley provided translations. Literature Cited Blackwall, J. 1867. Description of several species of East Indian spiders apparently new or little known to arachnologists. Annal. Mag. Nat. Hist., 19:387- 394. Bonnet, P. 1934. Le Gynandromorphisme chez les Araignees. Bull. Biol., 68:167-187. Bonnet, P. 1945. Bibliographia Araneorum. Tou- louse. Brest, R. H. 1966. Gynandromorphs and intersexes in mosquitoes (Diptera: Culicidae). Canadian J. ZooL, 44:91 1-921. Clarke, F. D. & Y. Rechav. 1992. A case of gynan- dromorphism in Amblyomma hebraeum (Acari: Ix- odidae). J. Med. Entomol., 29:1 13-1 14. Cokendolpher, J. C. & O. F. Franke. 1983. Gynan- dromophic desert fire ant, Solenopsis aurea Wheeler (Hymenoptera: Formicidae). J. New York Ent. Soc., 91:242-245. Cokendolpher, J. C. & W. D. Sissom. 1988. New gynandromorphic Opiliones and Scorpiones. Bull. British Arachnol. Soc., 7:278-280. Dondale, C. D. & J. H. Redner. 1978. Revision of the nearctic wolf spider genus Schizocosa (Araneida; Lycosidae). Canadian Entomol., 110:143-181. Exline, H. 1938. Gynandromorph spiders. J. Mor- phol., 63:441-475. Gack, C. & O. von Helversen. 1976. Zum Verhalten einer gynandromorphen Wolfspinne (Arachnida: Araneae: Lycosidae). Ent. Germanica, 3:109-118. Hackman, W. 1948. Chromosomenstudien an Ara- neen, mit besonderer Berecksichtigung der Ge- schlechts-chromosomen. Acta ZooL Fennica 54:1- 101. Hannah-Alava, A. 1960. Genetic mosaics. Sci. American, 202:118-130. Holm, A. 1941. Uber Gynandromorphismus und In- tersexualitat bei Spinnen. Zool. Bijdr., 20:397-413. Homsher, P. J. & C. E. Yunker. 1981. Bilateral gy- nandromorphism in Dermacentor andersoni (Acari: Ixodidae). J. Med. Entomol., 18:89-91. Hull, J. E. 1918. Gynandry in Arachnida. J. Genetics, 7:171-181. Kaston, B. J. 1961. Spider gynandromorphs and in- tersexes. J. New York Ent. Soc., 69:177-190. Mackie, D. W. 1969. A gynandromorph lycosid spi- der. Bull. British Arachnol. Soc., 1:40-41. Palmgren, P. 1979. On the frequency of gynandro- morphic spiders. Ann. Zool. Fennici, 16:183-185. Patten, M. A. 1 993. A probable bilateral gynandrom- ophic Black-Throated Blue Warbler. Wilson Bull., 105:695-698. Roberts, M. J. & J. R. Parker. 1973. Gynandry and intersexuality in spiders. Bull. British Arachnol. Soc., 2:177-183. White, M. J. D. 1973. Animal cytology and evolu- tion. Cambridge Univ. Press. Cambridge. 96 1 pp. Wilkins, A. D. 1993. Genetic analysis of animal de- velopment. Wiley-Liss, New York. 546 pp. Wise, D. 1983. An electron microscope study of the karyotypes of two wolf spiders. Canadian J. Genet. Cytol, 25:161-168. Gail E. Stratton1 : Dept, of Biology, Albion Col- lege, Albion, Michigan 49224 USA ‘Current address: Rhodes College, 2000 N. Parkway, Memphis, Tennessee 38112-1690. Manuscript received 4 January 1995, revised 20 March 1995. 1995. The Journal of Arachnology 23:134 Arachnological Research Fund The American Arachnological Society Fund for Arachnological Research (AAS Fund) is funded and administered by the American Arachnol- ogical Society. The purpose of the fund is to pro- vide research support for work relating to any aspect of the behavior, ecology, physiology, evo- lution, and systematics of any of the arachnid groups. Awards may be used for field work, mu- seum research (including travel), expendable supplies, identification of specimens, and/or for preparation of figures and drawings for publi- cation. Monies from the fund are not designed to augment or replace salary. Individual awards will not normally exceed $500.00, and preference will be given to students over part-time or ten- ured faculty. Up to five research awards will be made during each Winter-Spring or Summer-Fall granting period. Applications for support should be received by the chair of the review committee no later than May 30 or November 30, for fund- ing by June 30 and December 30, respectively. To be considered for an award from the AAS Fund, please submit three copies of a proposal of no more than five pages (including references) detailing your research project. Proposals should have three main parts: 1 ) an Introduction where background information is presented relative to the proposed work. The introduction should in- clude a section which places the proposed work in context with currently known relevant infor- mation, a section which provides justification for the proposed work, and a clear statement of the hypothesis(ses) to be tested or, in the case of systematic revisions, the type of synthesis that will be achieved and its significance; 2) a Meth- ods section where the methods, materials, ex- perimental design, and statistical or taxonomic analysis(ses) to be used are clearly and concisely presented, and 3) a Budget showing (in detail) how monies awarded will be spent in the pro- posed research. Proposals should be submitted to Dr. Craig S. Hieber, AAS Fund Chair, Dept, of Biology #1 742, St. Anselm College, Manchester, New Hamp- shire 03 1 02- 1310 USA. Proposals should be sub- mitted in English. Proposals may be FAXed (603- 641-7116), or sent electronically (chie- ber@hawk.anselm.edu) if it is appropriate or cost is prohibitive (out of country). 134 INSTRUCTIONS TO AUTHORS (revised July 1995) Manuscripts are preferred in English but may be ac- cepted in Spanish, French or Portuguese subject to availability of appropriate reviewers. Authors whose pri- mary language is not English may consult the Associate Editor for assistance in obtaining help with English manuscript preparation. All manuscripts should be pre- pared in general accordance with the current edition of the Council of Biological Editors Style Manual unless instructed otherwise below. Authors are advised to con- sult a recent issue of the Journal of Arachnology for additional points of style. Manuscripts longer than 1500 words should be prepared as Feature Articles, shorter papers as Research Notes. Send four identical copies of the typed material together with copies of illustrations to the Associate Editor of the Journal of Arachnology: Petra Sierwald, Associate Editor Division of Insects Dept, of Zoology Field Museum Roosevelt Road at Lakeshore Drive Chicago, IL 60605 USA [Telephone: (312) 922-9410, ext. 841; FAX: (312) 663-5397; Electronic mail: SIERWALD@FMNH.ORG]. Correspondence relating to manuscripts should also be directed to the Associate Editor. After the manu- script has been accepted, the author will be asked to submit the manuscript on a computer disc, preferably in MS DOS WordPerfect. FEATURE ARTICLES Title page. — The title page will include the com- plete name, address, and telephone number of the au- thor with whom proofs and correspondence should be exchanged, a FAX number 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 capital letters should be placed at the beginning of the first paragraph set off by a period. In articles written in English, a second abstract in an acceptable language may be included pertinent to the nationality of the author(s) or geo- graphic region(s) emphasized in the article. A second abstract in English must be included in articles not written in the latter language. 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 head 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 underlined and sep- arated 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 period regardless of language of the text. Citation of references in the text: Cite only papers already published or in press. Include within parenthe- ses the surname of the author followed by the date of publication. A comma separates multiple citations by the same author(s) and a semicolon separates citations by different authors, e.g., (Smith 1970), (Jones 1988; Smith 1993), (Smith 1986, 1987; Smith & Jones 1989; Jones et al. 1990). Literature cited section. — Use the following style: Lombardi, S. J. & D. L. Kaplan. 1990. The amino acid composition of major ampullate gland silk (dragline) of Nephila clavipes (Araneae, Tetragnath- idae). J. Arachnol., 18:297-306. Krafft, B. 1982. The significance and complexity of communication in spiders. Pp. 15-66, In Spider Communications: Mechanisms and Ecological Sig- nificance. (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. Running head. — The author surname(s) and an ab- breviated 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 con- tact the Associate Editor. Tables. — Each table, with the legend above, should be placed on a separate manuscript page. Only hori- zontal lines (usually three) should be included. Use no footnotes; instead, include all information in the leg- end. Make notations in the text margins to indicate the preferred location of tables in the printed text. Illustrations. — Address all questions concerning il- lustrations to the Editor of the Journal of Arachnology: James W. Berry, Editor Dept, of Biological Sciences Butler University Indianapolis, Indiana 46208 USA [Telephone (317) 940-9344; FAX (317)-940-9519; Electronic mail: BERRY@BUTLER.EDU] All art work must be camera-ready for reproduction. In line drawings, pay particular attention to width of lines and size of lettering when reductions are to be made by the printer. Multiple photos assembled on a single plate should be mounted with only a minimum of space separating them. In the case of multiple illus- trations 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. The author should indicate whether the illustrations should be one column or two columns in width. The overall dimen- sions should be no more than 11 inches (28 cm) X 14 inches (36 cm). Larger drawings present greater diffi- culty in shipping and greater risks of damage for which the JOA assumes no responsibility. Make notations in the text margins to indicate the preferred position of illustrations in the printed text. 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-us, 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, holotype male; 33, 34, A-us y-us, male. Scale = 1.0 mm. Assemble manuscript for mailing. — Assemble the separate sections or pages in the following sequence: title page, abstract, text, figure legends, footnotes, ta- bles with legends, figures. Page charges and reprints. — The current charge per journal page will be assessed as follows: Manu- scripts in the appropriate word processing program $ 10/page; others $35/page. Reprints are available only from the printer and should be ordered when the author receives the galley proof pages. Waiver of page charg- es: In exceptional cases, the Editor will consider re- quests for a waiver of page charges for manuscripts which have been accepted for publication. Requests should be made in writing to the Editor. The author must certify that neither grant nor institutional funds are available to cover all of the publication costs. The Society will normally waive no more than 80% of the page charges. RESEARCH NOTES Instructions above pertaining to feature articles ap- ply also to research notes, except that abstracts and most headings are not used and the author’s name and address follow the Literature Cited section. CONTENTS THE JOURNAL OF ARACHNOLOGY VOLUME 23 Feature Articles NUMBER 2 Description of the Spider Masoncus pogonophilus (Araneae, Linyphiidae), a Harvester Ant Myrmecophile Paula E. Cushing 55 A Test of the Central-Marginal Model Using Sand Scorpion Populations ( Paruroctonus mesaensis, Vaejovidae) Tsunemi Yamashita and Gary A. Polis 60 Natural History, Activity Patterns, and Relocation Rates of a Burrowing Wolf Spider: Geolycosa xera archboldi (Araneae, Lycosidae) Samuel D. Marshall 65 Observations of Habitat Use by Sarinda hentzi (Araneae, Salticidae) in Northeastern Kansas Stephen R. Johnson 71 Laboratory Studies of the Factors Stimulating Ballooning Behavior by Lin- yphiid Spiders (Araneae, Linyphiidae) Gabriel S. Weyman 75 Changes in Biomass of Penultimate-Instar Crab Spiders Misumena vatia (Araneae, Thomisidae) Hunting on Flowers Late in the Summer Doug- lass H. Morse 85 Redescription of the Scorpion Centruroides thorelli Kraepelin (Buthidae) and Description of Two New Species W. David Sissom 91 Distributions of the Scorpions Centruroides vittatus (Say) and Centruroides hentzi (Banks) in the United States and Mexico (Scorpiones, Buthidae) Rowland M. Shelley and W. David Sissom 100 Natural History of the Spider Genus Lutica (Araneae, Zodariidae) Martin G. Ramirez Ill Redescription of the Pennsylvanian Trigonotarbid Arachnid Lissomartus Pe- trunkevitch 1949 from Mazon Creek, Illinois Jason A. Dunlop 118 Research Notes A New Synonym in the Genus Meta (Araneae, Tetragnathidae) Charles D. Dondale 125 The New Species Purumitra australiensis (Araneae, Uloboridae) with Notes on its Natural History Brent D. Opell , Thomas W. Schoener, Susan L. Keen and Valerie T. Davies 127 A Gynandromorphic Schizocosa (Araneae, Lycosidae) Gale E. Stratton . . 130 Announcement Arachnological Research Fund 134 A The Journal of ARACHNOLOGY OFFICIAL ORGAN OF THE AMERICAN ARACHNOLOGICAL SOCIETY VOLUME 23 1995 NUMBER 3 THE JOURNAL OF ARACHNOLOGY EDITOR: James W. Berry, Butler University ASSOCIATE EDITOR: Petra Sierwald, Field Museum EDITORIAL BOARD: A. Cady, Miami (Ohio) Univ. at Middletown; J. E. Carrel, Univ. Missouri; J. A. Coddington, National Mus. Natural Hist.; J. C. Cokendolpher, Lubbock, Texas; F. A. Coyle, Western Carolina Univ.; C. D. Dondale, Agriculture Canada; W. G. Eberhard, Univ. Costa Rica; M. E. Galia- no, Mus. Argentino de Ciencias Naturales; M. H. Greenstone, BCIRL, Columbia, Missouri; C. Griswold, Calif. Acad. Sci.; N. V. Horner, Midwestern State Univ.; D. T. Jennings, Garland, Maine; V. F. Lee, California Acad. Sci.; H. W. Levi, Harvard Univ.; E. A. Maury, Mus. Argentino de Ciencias Naturales; N. I. Plat- nick, American Mus. Natural Hist.; G. A. Polis, Vanderbilt Univ.; S. E. Riechert, Univ. Tennessee; A. L. Rypstra, Miami Univ., Ohio; M. H. Robinson, U.S. National Zool. Park; W. A. Shear, Hampden-Sydney Coll.; G. W. Uetz, Univ. Cincinnati; C. E. Valerio, Univ. Costa Rica. The Journal of Arachnology (ISSN 0160-8202), a publication devoted to the study of Arachnida, is published three times each year by The American Arach- nological Society. Memberships (yearly): Membership is open to all those in- terested in Arachnida. Subscriptions to The Journal of Arachnology and American Arachnology (the newsletter), and annual meeting notices, are included with mem- bership in the Society. Regular, $30; Students, $20; Institutional, $80 (USA) or $90 (all other countries). Inquiries should be directed to the Membership Secretary (see below). Back Issues: Patricia Miller, PO. Box 5354, Northwest Mississippi Community College, Senatobia, Mississippi 38668 USA. Telephone: (601) 562- 3382. Undelivered Issues: Allen Press, Inc., 1041 New Hampshire Street, PO. Box 368, Lawrence, Kansas 66044 USA. THE AMERICAN ARACHNOLOGICAL SOCIETY PRESIDENT: Matthew H. Greenstone (1995-1997), Plant Science & Water Conservation Laboratory, USDA; Stillwater, Oklahoma 74075 USA. PRESIDENT-ELECT: Ann L. Rypstra (1995-1997), Dept, of Zoology, Miami University, Hamilton, Ohio 45011 USA. MEMBERSHIP SECRETARY: Norman I. Platnick (appointed), American Museum of Natural History, Central Park West at 79th St., New York, New York 10024 USA. TREASURER: Gail E. Stratton (1993-1995), Department of Biology, Rhoades College, Memphis, Tennessee 38112-1690 USA. BUSINESS MANAGER: Robert Suter, Dept, of Biology, Vassar College, Pough- keepsie, New York 12601 USA. SECRETARY: Alan Cady (1993-1995), Dept, of Zoology, Miami Univ., Mid- dleton, Ohio 45042 USA. ARCHIVIST: Vincent D. Roth, Box 136, Portal, Arizona 85632 USA. DIRECTORS: James Carico (1995-1997), Pat Miller (1993-1996), Robert Su- ter (1995-1997). HONORARY MEMBERS: C. D. Dondale, W. J. Gertsch, H. W. Levi, A. F. Millidge, W. Whitcomb. Cover illustration: Scanning electron microscope photograph of the abdomen of a mature female Philoponella vicina (Uloboridae). Photograph by Flory Pereira and William G. Eberhard. Publication date: 19 December 1995 © This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 1995. The Journal of Arachnology 23:155-144 '^vYTHSONjq^- JAN 1 9 1996 \ THE WEB-SPIDER COMMUMTCj^gj^BEAN AGROECOSYSTEMS IN SOUTRwESTEKN OHIO Ann L. Rypstra and Paul E. Carter: Department of Zoology, Miami University; Hamilton, Ohio 4501 1 and Oxford, Ohio 45056 USA ABSTRACT. We documented the web-spider community in a soybean agroecosystem over the entire growing season in 1990 and 1991 and over the period of peak spider abundance in August of 1993. Simultaneously a number of vegetational parameters were quantified in order to determine the extent to which the spider abundance was correlated with characteristics of the plant community. Web-spider abundance was higher in 1991 than in 1990 or 1993 and lower in 1993 than the other two years. The composition of the community in terms of web- types also differed among years with sheet webs (Linyphiidae, Agelenidae) being much more abundant in 1991 and orb webs (Araneidae) more abundant in 1990. In 1991, spider abundance was correlated with specific vegetation characteristics which suggests that the availability of habitat was important to spider colonization and establishment in that year. However, in 1 990 spider abundance was not correlated with any of the vegetation characteristics we measured. The late season spider density was positively correlated with weed biomass and the damage inflicted on the soybean leaves by herbivores was negatively correlated with the number of web- spiders across the three years. These data suggest that the web-spider community responds to some aspects of the plant community and that they have the potential to impact plant production by reducing the action of herbivores. Spiders are common generalist predators on arthropods in many agricultural systems. In spite of this fact, little data exist on their activities in these systems (Riechert & Lockley 1 984; NyfFeler & Benz 1987; Young & Edwards 1990; Wise 1993). Because spiders are generalist predators and most efforts to implement biological control have focused on predator or parasitoid interac- tions to reduce the impact of specific pests, spi- ders have not been considered seriously (Riech- ert & Lockley 1984; Nyffeler & Benz 1987). In addition, spiders have relatively long generation times and agricultural habitats are frequently dis- turbed by activities such as plowing and planting, which means that a large proportion of the spider community has to be re-established each season (Riechert & Lockley 1984; Young & Edwards 1990). It is not yet clear how predictable the spider community is from year to year and what factors might influence it most directly. Plants are important modifiers of the micro- climate for arthropods, specifically by moder- ating temperature and humidity extremes and by providing a more complex three-dimensional habitat (Cloudsley-Thompson 1962). The changes plants cause are dramatic in strongly sea- sonally systems like agricultural fields where ar- eas are specifically managed to proceed from no vegetation to a continuous cover of vegetation in a few months. As crop plants develop, the fields gradually become more hospitable to col- onization by spiders and other arthropods be- cause the plants provide structure, shade and help maintain moisture. Web-building spiders are particularly dependent on vegetation to provide suitable web-attachment sites (Greenstone 1 984; Rypstra 1986; Uetz 1991). Data suggest that spi- der colonization via aerial dispersal peaks early in the growing season (Bishop 1990; Bishop & Riechert 1990) so developmental rate and spac- ing pattern of crop plants should influence the establishment of the spider community in agro- ecosystems (Ferguson et al. 1984; Stinner & House 1990). In the United States, 262 species of spiders have been found in soybean ( Glycine max (L.) Merrill) fields (Young & Edwards 1990). In a given area, the number of spider species is likely to be less than that number, but soybeans typi- cally contain one of the most diverse commu- nities seen in any agricultural crop (LeSar & Un- zicker 1978; Young & Edwards 1990). Culm & Rust (1980) reported that the foliage community of spiders in soybeans responded across the sea- son to habitat space, a measure of plant devel- opment, whereas the ground dwelling spider community changed less over the season. Fur- ther evidence of the response of spiders to the soybean vegetation is apparent in comparisons among different cropping systems. Typically more 135 136 THE JOURNAL OF ARACHNOLOGY foliage-dwelling spiders have been found in fields where the plants were closer together and the cover more continuous than in other fields (Sprenkel et al. 1979; McPherson et al. 1982; Ferguson et al. 1984). In this study we focused on specific aspects of how the plant abundance and structure might be related to the web-spider community that de- veloped in soybeans. We attempted to determine how closely the spiders tracked vegetation de- velopment in this system by monitoring both spiders and plants weekly for two full seasons. In a third season, we collected data only at the time of peak web-spider and vegetational abun- dance in order to assess how predictable the web- spider guild composition was from year to year and to determine which vegetational character- istics might be most closely tied to spider abun- dance across years. METHODS Study area.— The study was conducted in soy- bean monoculture plots located at the Miami University Ecology Research Center, three miles north of Oxford, Butler County, Ohio, USA (Kemp & Barrett 1989). Each plot measured 60 x 70 m and contained 82 rows of soybeans plant- ed in an east-west direction. Experimental plots were bordered on all sides by a 15 m mowed grass strip. Four plots were randomly selected from an array of 12 in 1990 and three plots were selected from that array in 1991 and 1993. Herbicides were applied to all plots in all years. The pre-emergence herbicides, Lorox Plus® (lin- uron plus chlorimuron; 0.5 1 kg active ingredient/ hectare) and Dual 8E® (metolachlor; 1 .4 kg ac- tive ingredients/hectare), were applied to control for broadleaf and grassy weeds, respectively, in all three years. In 1991 and 1993, the post-emer- gence herbicide, Poast Plus® (sethoxyoim plus dash; 0.28 kg active ingredients/hectare) was ap- plied three weeks after planting to control grassy weeds. All herbicide treatments were applied be- fore we began sampling the plots. No insecticides were applied to the plots in any year. Plant characteristics.— Vegetation develop- ment was monitored weekly in each plot. We measured the height and width of five plants se- lected by generating random coordinates which determined meter points in a plot each week from three weeks after plant emergence until harvest. In order to obtain a measure of vertical leaf dis- tribution and therefore the availability of pos- sible web attachment points on the soybean plants, we placed a meter stick through the center of each of the five plants and recorded the height of each leaf touching the stick. Foliage height diversity (H') was then calculated from the leaf height measurements where H' = -Spjnp, (pt = the proportion of the total number of leaves with- in a 10 cm interval of soybean plant height) (Shannon & Weaver 1949). Total above ground production was measured just before the soybeans began to senescence (stage R6 as designated by Fehr & Caviness 1977) in all three years. On 7-10 September 1990, 16-20 August 1991 and 1993 four locations were se- lected randomly within each plot. One row meter of the soybean plants was clipped at the soil sur- face and collected. In addition the weeds in the area extending to the center point between rows on either side were clipped and collected. Plants were placed in a drying oven at 80 °C for at least seven days and then weighed. We assessed the cumulative damage inflicted by leaf chewing insects to soybean plants over the season by measuring leaf damage at the end of the season. Two sites in each plot were selected by generating random number coordinates. In 1 990 on 10 September, we traced 10 leaflets onto index cards at each site. On 24 August in 1991 and 1993, we traced 12 leaflets at each site. Leaf- lets areas were then determined to the nearest 0. 1 cm2 using a calibrated grid. Spiders.— We monitored the web-spider com- munity in the soybean plots by sampling four 1 m row lengths in each plot each week from the third week after plant emergence (early June) un- til harvest (late September) in 1990 and 1991 and from mid-season (mid July) to harvest in 1993. Specific row sections to be sampled were determined using different randomly generated coordinates each week. At each site, the plants and soil were visually searched for web-spiders between 0750-1050 h when dew made the webs most visible. Most web-spiders present in the fields were juveniles, which were difficult to iden- tify, especially without collection. Therefore, web- spiders were classified by web type (sheet, orb, or tangle). Sheet- web weavers (Linyphiidae, Age- lenidae) build webs that are characterized by a dense horizontal plane of silk frequently with a barrier web consisting of a tangle of silk sur- rounding the sheet to some degree. Orb- web weavers (Araneidae) consist of a circular plane of silk spirals with supporting spoke strands ra- diating from the hub. Tangle-web weavers (Ther- idiidae) build a three-dimensional and somewhat RYPSTRA & CARTER — SPIDERS IN SOYBEAN AGROECOSYSTEMS 137 Table 1.— The soybean growing season was separated into three time periods based on the developmental stage of the plants. Year Early Middle Late 1990 10 July-23 August 24 August-20 September 21 September- 13 October 1991 5 July-5 August 6 August-2 September 3 September-28 September 1993 28 June-24 July 25 July-28 August 29 August-25 September irregular mesh of strands connecting the vege- tation. Data analysis. —The soybean season in Ohio typically encompasses a four-month period from planting to harvest. The period of study shifted each year due to weather differences and the tim- ing of planting. For ease of comparison relative to the maturity of the plants, we divided each year into early, middle and late month-long time periods (Table 1). For years in which we had data for the entire season, a mean for each plot was generated for each of the three time periods. We then tested for differences between years and sea- son nested in year using a repeated measures analysis of variance. Comparisons of just the late season information across the three years were made using a one-way analysis of variance. The Tukey-Kramer Test was used to make pairwise comparisons among years. Mean spider abun- dance in the late season was regressed on those vegetation parameters that differed across all three years in an attempt to explain yearly variation in spider abundance. The 1990 and 1991 sea- sonal data were analyzed separately to dissect Table 2.— Summary of vegetation data from the soybean fields (Mean ± SD). ANOVA statistics for plant height, width, and H' are results of repeated measures test for differences among years and season nested in year. Statistics for biomass and leaf damage are from one-way ANOVA. 1990 1991 1993 Plant height (cm) Early 47.1 ± 2.9 41.2 ± 4.6 51.2 ± 3.7 Middle 74.9 ± 2.2 89.7 ± 2.4 92.6 ± 3.2 Late 72.5 ± 1.4 94.0 ± 2.0 95.2 ± 2.1 year, season (year): F = 168, 49.8; df = 2,6; P < 0.05 Plant width (cm) Early 47.5 ± 3.2 44.7 ± 3.3 42.3 ± 3.1 Middle 69.1 ± 1.7 65.8 ± 2.2 68.0 ± 2.3 Late 41.8 ± 3.1 61.3 ± 4.7 64.7 ± 5.7 year, season (year): F = 33.7, 7.9; df = 2,6; P < 0.05 Foliage height diversity (H') Early 1.44 ± 0.05 1.48 ± 0.03 1.54 ± 0.06 Middle 1.87 ± 0.03 1.88 ± 0.04 1.92 ± 0.05 Late 1.27 ± 0.14 1.67 ± 0.10 1.74 ± 0.12 year, season (year): F = 16.06, 10.2; df = 2,6; P < 0.05 Soybean biomass (g) Late 174 ± 32 356 ± 89 423 ± 97 year: F = 32.45; df — 2; P < 0.05 Weed biomass (g) Late 503 ± 153 987 ± 102 68 ± 49 year: F = 46.13; df = 2; P < 0.05 Leaf damage (%) Late 5.9 ± 0.6 2.2 ± 0.4 26.7 ± 4.5 year F = 56.56; df = 1 l,P< 0.05 138 THE JOURNAL OF ARACHNOLOGY (/) IT UJ G CL (/) CC Hi m 0 2 4 0 8 1 0 1 2 1 4 WEEK 1990 1991 1993 Figure 1 . — Spider abundance (per four one-meter row sections) in soybean plots over the entire growing season in 1990 and 1991 and the latter part of the season in 1993 (Mean ± SD). how closely the spiders tracked the changes in vegetation. Weekly values of the number of spi- ders were regressed on weekly values for vege- tation height, width, and foliage height diversity. RESULTS Vegetation.— Soybean plant size differed sig- nificantly among years but those differences were only really apparent in the middle and late por- tions of the season so there was a significant sea- sonal effect as well (Table 2). Plants were larger and more complex (as measured by foliage height diversity) in 1991 and 1993 than they were in 1990 (Table 2). Likewise, the above ground soy- bean biomass of one row meter at the peak of lushness (R6) was significantly greater in 1991 and 1993 than it was in 1990 (Table 2). The above ground biomass of weeds surrounding one meter of soybean plants weeds also varied among years (Table 2). The Tukey-Kramer Pairwise Comparisons Test revealed that there was greater weed biomass in 1991 than in 1990 or 1993 and weed biomass was higher in 1990 than it was in 1993 (Table 2). The proportion of each leaflet damaged was also different from year to year (Table 2). Chewing insects damaged the soybean plants much more dramatically in 1993 than in 1 990, and plants in both 1 990 and 1 993 received more damage than in 1991 (Table 2) (Tukey- Kramer Test, P < 0.05). Spiders.— There were significantly more web spiders found in 1991 than in 1990 across the whole season (Fig. 1) (repeated measures of year, season (year): F = 9.81, 6.15, df = 1, 5, P < <0.05). There was considerable overlap in the early and mid-season numbers but the abun- dances clearly separated by year in the late season (Fig. 1). We did not collect early or mid-season data for 1993, but there were significantly fewer spiders in the late season of that year than in the late season of either of the other two years (Fig. 1) (F= 60.91,#= 2, P< 0.05). Sheet webs were the most abundant web type in the soybean fields and, at the time of peak spider abundance in the late season, they com- prised over 40% of the spiders we observed in all three years (Fig. 2). Sheet webs were signifi- cantly more abundant in the 1991 season than they were in 1990 (Table 3). There were no sea- sonal differences in sheet web abundance in 1 990 or 1991 (Table 3). If we compare the late season data from all three years, there were significantly more sheet webs in the fields in 1991 than in the other two years (Table 3) (Tukey-Kramer Test, P < 0.05). More than 75% of the sheet-web builders in the plots in all three years belonged to five species (Table 4). Orb webs were second in abundance to sheet webs in 1990 and 1993 when they comprised more than 25% of the late season community, but they were very uncommon in 1991 (Fig. 2). Unlike sheet-web spinners, the orb-web weavers were significantly more abundant across the sea- son in 1990 than they were in 1991 (Table 3). RYPSTRA & CARTER — SPIDERS IN SOYBEAN AGROECOSYSTEMS 139 Table 3.— Abundances of the three common web types in soybean fields (Mean ± SD). ANOYA statistics are results of repeated measures test for differences between 1990 and 1991, and for season nested in year. Late season data indicated with were significantly different from other years by Tukey-Kramer Pairwise Com- parison Test (P < 0.05). 1990 1991 1993 Number of sheet webs Early 10.9 ± 6.0 23.0 ± 4.5 — Middle 11.6 ± 3.1 18.0 ± 3.1 _ Late 11.5 ± 2.7 26.6 ± 3.8* 6.3 ± 4.0 year, season (year): F = = 38.27, 1.7; df= 1,5; P < 0.05, P > 0.1 Number of orb webs Early 5.6 ± 1.3 3.7 ± 1.4 __ Middle 5.7 ± 1.4 2.0 ± 0.3 — Late 11.1 ± 2.2* 3.9 ± 0.5 3.3 ± 1.2 year, season (year): F = 50.13, 10.39; df = 1,5; P < 0.05 Number of tangle webs Early 0.8 ± 1.0 0.7 ± 0.6 — Middle 1.0 ± 0.5 0.3 ± 0.3 — Late 2.9 ± 0.7 4.6 ± 0.8 0.9 ± 1.0* year, season (year): F = 0.69, 22.21; df= 1,5; P > 0.1, P < 0.05 Additionally there was a significant seasonal in- crease in the number of orb webs that we were able to find in 1990 (Table 3). In a comparison of late season data of all three years, orb web abundance was significantly higher in 1990 than it was in either of the other two years (Table 3) (Tukey-Kramer Test, P < 0.05). Five species comprised better than 75% of the orb-web weav- ers that we observed in this habitat (Table 4). Tangle-web weavers were least abundant of the web-spinners, comprising less than 15% of the web-spider community in the soybean fields in the late season of all three years (Fig. 2). There was not a significant difference between the abun- dance of tangle web weavers in 1990 and 1991 (Table 3). However, the abundance of spiders building tangle webs was significantly greater lat- er in the season than it was in the early or middle portions in those years (Table 3). In a compar- ison of the late season data on tangle weaver abundance, there were significantly fewer in 1 993 than in the other two years of this study (Table 3). Four species of tangle-weavers were collected in all three years (Table 4). Spiders in relation to vegetation.— Total web- spider abundance across the season was corre- lated with many of the vegetation parameters in 1991, but was not correlated with any of these parameters in 1990 (Table 5). In 1991, the strongest correlation was between soybean plant width and web-spider abundance. However, fo- liage height diversity, and plant height were also significantly correlated with spider abundance in that year (Table 5). Table 4. —List of most common spider species found in the soybean fields categorized by web type. Sheet- web weavers Agelenidae Agelenopsis pennsylvanica (C. L. Koch) Linyphiidae Frontinella pyramiteia (Walck.) Meioneta micaria (Emerton) Tennesseellum formicum (Emerton) Microlinyphia pusilla (Sundevall) Orb-web weavers Araneidae Argiope aurantia Lucas A. trifasciata (Forskal) Cyclosa conica (Pallas) Neoscona arabesca (Walck.) Tetragnatha laboriosa Hentz Tangle- web weavers Theridiidae Achaearanea tepidariorum (C. L. Koch) Theridion frondeum Hentz T. neshamini Levi Theridula opulenta (Walck.) 140 THE JOURNAL OF ARACHNOLOGY LU CL >■ h- co LU LL o h- Z LU O cr LU Q. ■ SHEET E23 ORB □ TANGLE 1990 1991 YEAR 1993 Figure 2. —The relative abundance (percent of all web-spiders found) of sheet, orb, and tangle webs that comprised the web-spider community in soybeans during the late season of three years. The relative amount of weedy vegetation, leaf damage (which is a measure of the activity of herbivores) and web spiders in the late season across the three years of this study appear to be related. Weed abundance was a vegetation pa- rameter that was different in all three years, and it had a strong positive correlation with spider abundance across years (Fig. 3) (R2 = 95.5, P < 0.05). Leaf damage also differed among all years and it was negatively correlated with spider abundance (Fig. 4)(R2 = 79.5,P<0.05). Because of the strong correlation between weeds and spi- ders and between spiders and leaf damage, leaf damage was also negatively correlated with weed abundance in this data set ( R 2 = 74.0, P < 0.05). DISCUSSION The phenology of web-spiders in these Ohio soybean fields was similar in many respects to that observed in other north temperate studies (LeSar & Unzicker 1978; Culin & Rust 1980; Culin & Yeargan 1983; Ferguson et al. 1984). The variation among years is interesting in that the overall abundance was different in each of the three years and that difference is not reflected 1990 1991 1993 Figure 3.— The mean number of web spiders (per four one-meter row sections) found in the soybean fields late in the season as a function of the biomass of weeds surrounding one row meter. RYPSTRA & CARTER — SPIDERS IN SOYBEAN AGROECOSYSTEMS 141 NUMBER OF SPIDERS Figure 4. —Leaf damage (% removed by pest insects) experienced by soybean leaves as a function of the number of web-spiders (per four one-meter row sections) found in each field over three years. by parallel differences in the web types we ob- served. For example, the highest overall spider abundance was in 1991 when sheet webs dom- inated the community but orb webs were much more abundant in 1990 (Fig. 2). LeSar & Un- zicker (1978) observed that both Tetragnatha la- boriosa, an orb weaver, and Microlinyphia pus- ilia, a sheet-web weaver, were more abundant in a dry year than they were in a second wetter year and they attributed this differences to a negative effect of rainfall on web spinners. The sheet-web weavers at our site appeared to follow that pat- tern in that they were much more abundant in 1991, the driest of the three years studied (J. Klink pers. comm.). However, orb weavers, in- cluding T. laboriosa, were most abundant in 1990 (Fig. 2) which was the wettest of the three years under study (J. Klink pers. comm.). In 1991, sheet-web weavers were very abundant early in the season in relation to any other web type in any year (Table 3). It must be that, for some reason, they were able to disperse in precisely when microhabitat conditions were suitable and establish themselves in the fields early in that year. Since their webs are frequently three-di- mensional and require multiple attachment sites, high sheet web densities could have inhibited the colonization of orb-weavers. The lower numbers of spiders and differences in spider types present in the fields in 1990 than compared to 1991 may, in part, be due to the time of planting. The soybean season was about two weeks later in 1 990 that in the other years (Table 1). Since colonization occurs largely by ballooning and the greatest peak of ballooning is observed early in the summer (Bishop & Riech- Table 5.— Correlations between plant characteristics and web-spider abundance in soybean agroecosystems in 1990 and 1991. Variable Sign R df P 1990 Height + 0.234 43 >0.1 Width + 0.077 43 >0.1 Foliage height diversity + 0.234 43 >0.1 1991 Height + 0.491 47 <0.01 Width + 0.650 47 <0.01 Foliage height diversity + 0.565 47 <0.01 142 THE JOURNAL OF ARACHNOLOGY ert 1990), more potential colonizers would have found suitable habitat in 1991 than in 1990. Likewise, the timing of planting may have co- incided with the ballooning of sheet-web weavers to lead to a greater establishment of those spiders in that year. The season in 1993 was even earlier (Table 1) and the plants developed normally (Ta- ble 2), so this explanation for annual differences does not explain the overall low spider abun- dance observed in that year. The specific development of the web-spider community was more closely aligned with var- ious vegetational measurements in 1991 than in 1990 (Table 5). In many cases, the complexity of the habitat has been related to spider abun- dance (Greenstone 1984; Rypstra 1983, 1986; Dobeletal. 1990; Gunnarsson 1990; Uetz 1991). Yet in 1990 when overall spider abundance was lower, there were no significant correlations be- tween spider abundance and plant structure (Ta- ble 5). The lack of any such correlations, might suggest that the habitat was not saturated, i. e., that there were suitable unused web sites. Alter- natively, since sheet webs comprised such a large proportion of the web-spider community in 1 99 1 , it may be that our measures of vegetational com- plexity were better measures of habitat suitability for sheet-web weavers than for all web spiders (Fig. 2). One complication is that the overall spi- der community appeared to respond quite strongly across years to weed abundance which was also highest in 1991 (Fig. 3), yet our mea- sures of vegetational heterogeneity focused spe- cifically on the soybean plants and did not reflect changes in the developing weed community. The tighter correlations we observed in 1 99 1 in com- parison to 1 990, may mean that sheet-web weav- ers were more dependent on the soybean plants themselves for web sites than the other spiders we observed. Previous studies have demonstrated that spi- der numbers can be manipulated by altering the habitat structure available to them (Robinson 1981; Rypstra 1983; Carter & Rypstra 1995). Likewise, in no-till soybean systems, which tend to be more weedy than conventionally tilled fields, the greater diversity and abundance of beneficial arthropods, including spiders, have been attrib- uted to the greater structural diversity of the plant community (House & Stinner 1983). Ferguson et al. (1984) found greater spider numbers and diversity in soybean fields that were planted clos- er together and disturbed less. Gur data support those studies and suggest that high weed abun- dance is the basis for a more dense community of web spiders (Fig. 3). It has been demonstrated that spiders can re- duce the herbivory experienced by plants (Riech- ert & Bishop 1990; Carter & Rypstra 1995). The strong negative correlation between spider abun- dance and leaf damage we observed suggests that the spiders were having an impact on the action of herbivores in these fields as well (Fig. 4). We believe that weed abundance allowed a more dense community of spiders (Fig. 3) and that the reduction in leaf damage is due to the direct and indirect effects of the spiders on the herbivores. In experiments with introduced web-spiders in these same soybean fields, we observed a nega- tive correlation between the biomass of prey killed by spiders and leaf damage experienced by plants in localized areas (Carter & Rypstra 1995). Therefore we think that it is likely that the dif- ferences in damage we observed across years in this study are due to differences in the spider community. However, one cannot ignore the correlation between weed biomass and leaf dam- age. Weeds may offer polyphagous pests an al- ternative food source and, in that way, reduce their dependence on the crop plants. The results of studies on the interaction of weeds, foliage pest species, and crop plants are mixed, with some pest species inflicting more damage to the crop plants when weeds are abun- dant and some pest species inflicting less (Ham- mond et al. 1987; Stinner & House 1990). How- ever, more frequently it has been suggested that the reduction in pest damage in no-till, and there- fore weedy, agroecosystems is due to increased predation on or parasitism of herbivorous insects (Speight & Lawton 1976; House & Stinner 1983; Pavuk & Stinner 1992). Clearly more work is critically needed to uncouple these effects in or- der to understand the relationship between spi- ders, their prey and the plants in agroecosystems. In summary these data suggest that agricul- tural systems contain highly variable dynamic web spider communities in which the compo- sition parameters are related to vegetation de- velopment in some years. More information on the specifics of colonization and establishment of different spiders and their preferred plant as- sociations are critical to understanding this sys- tem. These data underscore the importance of understanding such interactions since it is be- coming increasingly clear that spiders are im- portant predators that can influence the action of pest insects in agroecosystems. RYPSTRA & CARTER — SPIDERS IN SOYBEAN AGROECOSYSTEMS 143 ACKNOWLEDGMENTS We thank J. R. Dobyns, D. L. Gorchov, M. H. Greenstone, S. D. Marshall, B. J. McNett, R. E. Lee, D. G. Pennock, M. J. Vanni and D. H. Wise for comments and suggestions on earlier drafts of this manuscript. D. M. Pavuk, M. Hooke and M. Meaner provided invaluable field and laboratory assistance. We also thank D. M. Pa- vuk and G. W. Barrett for the use of their fields supported by their USDA Grant 9 1 -37302-620 1 . John Klink generously provided the weather data. We are indebted to K. R. Cangialosi for numer- ous valuable discussions. This work was sup- ported by a Grant-in- Aid of Research from Sig- ma Xi, The Scientific Research Society and a grant from the Faculty Research Committee, Mi- ami University. LITERATURE CITED Bishop, L. 1990. Meteorological aspects of spider ballooning. Environ. Entomol., 19:1381=1387. Bishop. L. & S. E. Riechert. 1 990. Spider colonization of agroecosystems: mode and source. Environ. En- tomol., 19:1738-1745. Carter, P. E. & A. L. Rypstra. 1 995. Top-down effects in soybean agroecosystems: spider density affects herbivore damage. Oikos, 72:433-439. Cloudsley-Thompson, J. L. 1962. Microclimates and the distribution of terrestrial arthropods. Ann. Rev. Entomol, 7:199=222. Culin, J. D. & R. W. Rust. 1980. Comparison of the ground surface and foliage dwelling spider com- munities in a soybean habitat. Environ. Entomol., 9:577-582. Culin, J. D. & K. V. Yeargan. 1983. Spider fauna of alfalfa and soybean in central Kentucky. Trans. Kentucky. Acad. Sci., 44:40-45. Dobel, H. G., R. F. Denno & J. A. Coddington. 1 990. Spider (Araneae) community structure in an inter- tidal salt marsh: effects of vegetation structure and tidal flooding. Environ. Entomol., 19:1356-1370. Fehr, W. R. & C. E. Caviness. 1 977. Stages of soybean development. Coop. Ext. Serv., Agricul. & Home Econ. Exp. Stat., Iowa State Univ. Special Rep., 80: 1-12. Ferguson, H. J., R. M. McPherson & W. A. Allen. 1984. Ground- and foliage dwelling spiders in four soybean cropping systems. Environ. Entomol, 13: 975-980. Greenstone, M. H. 1984. Determinants of web spider species diversity: vegetation structural diversity vs. prey availability. Oecologia, 62:299-304. Gunnarsson, B. 1990. Vegetation structure and the abundance and size distribution of spruce-living spi- ders. J. Anim. Ecol, 59:743-752. Hammond, R. B. 1 987. Pest management in reduced tillage soybean cropping systems. Pp. 23-45 In Ar- thropods in Conservation Tillage Systems. (G. J. House & B. R. Stinner, eds.). ESA Misc. Publ, Col- lege Park, Maryland. House, G. J. & B. R. Stinner. 1983. Arthropods in no-tillage soybean agroecosystems: community composition and ecosystem interactions. Environ. Manag., 7:23-28. Kemp, J. C. & G. W. Barrett. 1989. Spatial pattern- ing: impact of uncultivated corridors on arthropod populations within soybean agroecosystems. Ecol- ogy, 70:1 14-128. LeSar, C. D. & J. D. Unzicker. 1978. Soybean spiders: species composition, population densities, and ver- tical distribution. Illinois Nat. Hist. Surv. Biol. Notes, 107:1-14. McPherson, R. M., J. C. Smith & W. A. Allen. 1982. Incidence of arthropod predators in different soy- bean cropping systems. Environ. Entomol, 1 1:685— 689. Nyffeler, M. & G. Benz. 1987. Spiders in natural pest control: a review. J. Appl. Entomol, 103:321-339. Pavuk, D. M., & B. R. Stinner. 1992. Influence of weed communities in com planting on parasitism of Ostrina nubilalis (Lepidoptera: Pyralidae) by Er- iborus terebrans (Hymenoptera: Icheumonidae). Biol. Control, 2:312-316. Riechert, S. E. & T. Lockley. 1984. Spiders as bio- logical control agents. Ann. Rev. Entomol, 29:299- 320. Riechert, S. E. & L. Bishop. 1990. Prey control by an assemblage of generalist predators: spiders in gar- den test systems. Ecology, 71: 1441=1450. Robinson, J. V. 1981. The effect of architectural vari- ation in habitat on a spider community: an exper- imental field study. Ecology, 62:73-80. Rypstra, A. L. 1983. The importance of food and space in limiting web-spider densities; a test using field enclosures. Oecologia, 59:312-316. Rypstra, A. L. 1986. Web spiders in temperate and tropical forests: relative abundance and environ- mental correlates. American Midi. Nat., 1 1 5:42-5 1 . Shannon, C. E. & W. Weaver. 1949. The Mathe- matical Theory of Communication. Univ. Illinois Press, Urbana, Illinois. Speight, M. R. & J. H. Lawton. 1976. The influence of weed-cover on the mortality imposed on artificial prey by predatory ground beetles in cereal fields. Oecologia, 23:211-223. Sprenkel, R. K, W. M. Brooks, J. W. Van Duyn & L. L. Deitz. 1979. The effects of three cultural vari- ables on the incidence of Nomurea rileyi, phytoph- agous Lepidoptera, and their predators on soybeans. Environ. Entomol, 8:334-339. Stinner, B. R. & G. J. House. 1990. Arthropods and other invertebrates in conservation-tillage agricul- ture. Ann. Rev. Entomol, 35:299-318. Uetz, G. W. 1991. Habitat structure and spider for- aging. Pp. 325=348, In Habitat structure, the phys- 144 THE JOURNAL OF ARACHNOLOGY ical arrangement of objects in space. (S. S. Bell, E. D. McCoy & H. R. Mushinsky, eds.). Chapman and Hall, New York. Wise, D. H. 1993. Spiders in ecological webs. Cam- bridge Univ. Press, Cambridge, England. Young, O. P. & G. B. Edwards. 1990. Spiders in United States field crops and their potential effects on crop pests. J. Arachnol., 18: 1-27. Manuscript received 8 July 1993, revised 24 July 1995. 1995. The Journal of Arachnology 23:145-150 MECHANISMS OF THE FORMATION OF TERRITORIAL AGGREGATIONS OF THE BURROWING WOLF SPIDER GEOLYCOSA XERA ARCHBOLDI MCCRONE (ARANEAE, LYCOSIDAE) Samuel D. Marshall1: Department of Zoology & Graduate Program in Ethology, University of Tennessee, Knoxville, Tennessee 37996 USA ABSTRACT. It has long been proposed that aggregations of Geolycosa wolf spiders form by limited dispersal from the maternal burrow. In this study I test for conspecific attraction and limited dispersal to account for the formation and maintenance of aggregations of Geolycosa xera archboldi McCrone, endemic to the scrub habitats of Highlands County in central Florida, USA. I found no evidence for conspecific attraction in either field tests or observations of natural relocation. I did confirm that hatchlings disperse a short distance from the maternal burrow. The distance hatchlings disperse is influenced by territorial interactions with siblings. Older spiders which were experimentally released also exhibited limited dispersal. Aggregation by territorial species presents a theoretical problem in that territoriality is by definition partitioning of space away from con- specifics. The hypothesized functions of terri- torial aggregation fall into two general categories: 1) ‘dear enemies’, where neighbors are beneficial and; 2) ‘conspecifics-as-cues’, where dispersing individuals use established conspecifics as cues to habitat quality. Getty’s (1981) competitive collusion model proposed that neighbors might ultimately serve to maximize territory size. This ‘dear enemies’ model suggests that individuals which settle next to neighbors will be less likely to lose territorial space to later immigrants to the patch. These later immigrants might insert their territories into the unclaimed space between randomly spaced territories and usurp peripheral space from these surrounding territories. Conspecific attraction occurs when dispersing or relocating individuals use the presence of es- tablished conspecifics as site-selection criteria when they choose a site to establish a territory. The use of conspecifics as evidence of habitat quality was proposed as early as 1961 by Orians to explain the phenomenon of territorial aggre- gation by nesting red-winged blackbirds ( Age - laius phoeniceus). Stamps (1987, 1988, 1991) has proposed this as the function of conspecific at- 1 Present address: Department of Zoology, Miami Uni- versity, Oxford, Ohio 45056. traction by anole lizards ( Anolis aeneus) on the Caribbean island of Grenada. I have documented both territoriality and ag- gregation in Geolycosa xera archboldi McCrone, a burrowing wolf spider (Marshall 1994). Clumped dispersion of Geolycosa burrows has been noted previously (G. rafaelana, Conley 1984; G. turricola, Miller 1989; G. missouriensis, Richardson 1990). It has been proposed that ag- gregations of Geolycosa wolf spiders form by the settlement of juveniles near the maternal burrow (Miller 1989). However, this has never been quantified nor have alternative hypotheses been tested. In this study I examine the mechanisms of aggregation formation in Geolycosa xera. Geo- lycosa xera is restricted to the scrub habitats of central Florida where it builds burrows in areas of exposed sand (Marshall 1994). These spiders are entirely dependant on their burrows for pro- tection from predators and climatic extremes. All foraging activity is centered on the burrow mouth and G. xera does not leave the immediate vicin- ity of its burrow unless dispersing. METHODS The present study was conducted at Archbold Biological Station in Highlands County, Florida. Archbold is a private research facility approxi- mately 10 km S of Lake Placid. Experimental tests for conspecific attrac- tion.—I tested for conspecific attraction in the summers of 1990 and 1991. I have previously 145 146 THE JOURNAL OF ARACHNOLOGY observed and quantified dispersal by these spi- ders in summer (Marshall in press). I used two approaches in my test for conspecific attraction: enclosure tests and field trials. In both these tests I examined the influence of established territory holders on burrow sites chosen by experimen- tally released individuals I will call ‘settlers’. For the enclosure tests I built five 1.0 x 2.0 m alu- minum flashing enclosures in an area of suitable microhabitat. The founding population was in- duced to dig burrows into the end of each enclo- sure I selected by covering the sand in the other end with leaf litter. I then removed the leaf litter and released one marked settler into each enclo- sure each night for a week at randomly selected points (using an X, Y coordinate system and a random number table). The morning after re- lease I noted where the settlers had dug their burrows. Settlers which dug burrows in the half of the enclosure with the founders were scored as having exhibited conspecific attraction, and those settlers which had dug burrows in the other end as not exhibiting conspecific attraction. For the field trials I created four rectangular, open sand patches 1.5 x 3.0 m in oak scrub by raking leaf litter and cutting vegetation to the ground surface, exposing the sand. Studies I have conducted on the ecology of G. xera indicate that barren sand is the sole requirement for burrow placement. Two of these were to be test patches with founders, and two were to be control patch- es without any founding population. Ten foun- ders were established in one end of the founder patches in a two by five array. This founding group was established by setting out 5 cm di- ameter approximately 20 cm tall clear acetate cylinders spaced 30 cm apart and placing a spider into each cylinder at dusk. It was hoped at this time that they would dig a burrow. Individuals which did not dig a burrow by morning were removed and replaced with another spider the following evening. These cylinders were re- moved after the spiders built burrows. I used older spiders (burrow diameter >5.0 mm) as founders, and younger spiders (burrow diameter <3.0 mm) as settlers (this followed the finding that younger individuals relocate more often; Marshall in press). The distinct size difference between founders and immigrants made it un- necessary to mark the spiders. Over six succes- sive evenings I introduced two settlers 1.0 m apart in the center of each sand patch (for a total of 12 per patch). The next morning I recorded the burrow sites (i. e., which end) they chose and removed them. To test for non-random dispersal I scored site choice by which half of the patch the settlers dug burrows in. If the settlers were exhibiting conspecific attraction they should preferentially choose the ends of the patches with founders. In the patches with no founding pop- ulation settlement should be random. To test for non-random settlement I used a Fisher’s exact test on frequency of burrow site choice scored by which end of the patch immigrants selected for each pair of patches for each treatment. In the founder patches this was either the half with founders or the half without founders. In the control patches this was either the eastern end or the western end. Census studies of immigration and recruit- ment.—In order to test for the influence of local population density on rates of immigration (spi- ders moving into the area) and recruitment (spi- ders hatching within the area), I collected census data on naturally-occurring local populations of G. xera. I set up 10 pairs of 1.0 m2 quadrats in 1 0 independent patches of sand in the scrub. A patch was considered independent if it had a well-defined leaf litter edge. I selected each cen- sus quadrat pair to represent the highest and the lowest spider densities found within each patch. 1 censused these patches on a weekly basis from 2 April- 10 July 1993. At each census new bur- rows were flagged and measured. Based on ob- servations of natal dispersal, I knew that all spi- ders with a burrow diameter of 2.0 mm or less were spiderlings and assigned them to the ’re- cruit’ category (in the sense of recruitment into the population by birth). All new spiders with larger burrows were considered to be settlers. Also, previous research (Marshall in press) has shown that these spiders will periodically close their burrows and that 90% of these closure pe- riods will last 14 days or less. For this reason I will consider only the census data from 1 6 April- 10 July in my analysis in order to minimize the counting of residents as settlers when they reopen their burrows during the early part of the census period. At each census any previously flagged burrows were checked, and it was noted whether they were open or closed. These data allow for an estimation of rates of immigration versus re- cruitment. Dispersal strategies.— Aggregations of Geoly - cosa wolf spiders have been proposed to form by the settlement of hatchlings in the vicinity of the maternal burrow (Conley 1984; Miller 1989). I tested this by recording the dispersal distances MARSHALL— GEOLYCOSA AGGREGATION FORMATION 147 of hatchlings from five separate sibling groups, I began by observing the burrows of females I knew to be incubating eggs in late June 1991. When I first observed spiderlings in the maternal burrow, I began to check the vicinity of the maternal burrow daily for the appearance of spiderling burrows. I measured the distances of each spi- derling burrow from the maternal burrow, its nearest-neighbor distance and its nearest-neigh- bor’s burrow mouth diameter (burrow mouth diameter is closely correlated with body size in Geolycosa ; McQueen 1983; Miller & Miller 1984; Marshall in press). As the maternal burrows were spaced widely apart, I feel confidant that all ob- served spiderlings were attributed to the appro- priate maternal burrow. In order to look for pat- tern in the data, I used regression analysis of the distance dispersed as a function of days since the onset of dispersal I predicted that a significant, positive relationship between day and distance dispersed would be evidence of territorial aggre- gation by the spiderlings. Individuals which dis- persed first would establish territories close to the maternal burrow. Siblings that dispersed later would be forced to walk further before digging a burrow by competition for space with previously settled siblings. Only the data from the first 10 days were used in order to limit the recounting of relocating spiderlings. Counting spiderlings twice would violate the assumption of indepen- dence of the regression model The high rate of relocation found in the 1991 summer field season (up to 3.2% of the popu- lation relocates per day; Marshall in press) cou- pled with the persistence of aggregations raised the question of how individual dispersal strate- gies might influence patterns of dispersion. In June of 1994 I marked and released 80 individ- uals in order to quantify individual dispersal dis- tance. The test subjects were juveniles arbitrarily collected from outside the study population. I marked the spiders with a fluorescent powder (Radiant Color, Magmder Color Co., Elizabeth, New Jersey) of a type which has been widely used for both invertebrates and vertebrates (Lemen & Freeman 1985; Fellers & Drost 1989; Morse 1993). I marked the spiders by placing them in a vial containing a small amount of the powder and gently shaking them so as to completely coat the spider. Spiders were held until release in a clean vial One advantage the use of this powder has over paint marking is that the spiders in- corporate the powder coating their bodies into their new burrows. These colored burrow mouths are very conspicuous in the white sand of the scrub. The spiders were released at sites in suit- able habitat at least 30 cm from larger coi] spe- cifics. I did this to reduce the chance of canni- balism due to my choice of release site. I released the spiders in early evening (approximately 1 800 h) which is the time of day I had most commonly observed relocating individuals. Spiders were re- leased by placing the vial containing the spider open on its side in the sand and then leaving the area. I found burrow sites the next morning by searching the entire sand patch. I also marked, released, and watched 14 ad- ditional individuals until they dug burrows. These individuals were marked with enamel paint on the carapace and released as detailed above. In- stead of leaving the area I stepped back to ob- serve from at least 4 m away. RESULTS Experimental tests for conspecific attrac- tion. —In the enclosure test there was no evidence for conspecific attraction. The mean proportion of settlers in the five enclosures choosing the end of the enclosures with founders was close to one- half (0.44). There was no evidence for conspecific attrac- tion in the field trials either. In both the patches with founders and without, settlers settled ran- domly (Fisher’s exact test: patches without foun- ders, P = 0.54; patches with founders, P = 0. 19). Census studies of immigration and recruit- ment.—-The mean densities for the weekly cen- suses of the paired plots were significantly dif- ferent for high versus low density local popula- tions (paired t = 5.M, df = 9, P < 0.001). How- ever, there was no significant difference for immigration rate (paired t = 1.08, df = 9, P > Table l —Summary of regression analyses of natal dispersal of 5 groups of Geolycosa xera archboldi at Archbold Biological Station. For the analyses the dis- tance from the maternal burrow that spiderlings built burrows was regressed on the number of days since the initiation of dispersal by the brood that the spiderling burrow appeared. N Sig. r2 First group 24 P= 0.01 0.26 Second group 19 P = 0.001 0.63 Third group 51 P = 0.006 0.14 Fourth group 20 P = 0.0043 0.37 Fifth group 22 P = 0.0006 0.45 1 48 THE JOURNAL OF ARACHNOLOGY DAYS Figure 1.— Cumulative mean dispersal distances (± SE) for five cohorts of hatchling Geolycosa xera archboldi at Archbold Biological Station, June- July 1991. The two lines represent the mean of values recorded for each spiderling burrow, the standard errors are for an n = 5 (for the five cohorts). 0.2) or recruitment rate (paired t = 0.04, df = 9, P > 0.5). Dispersal strategies. —In all five groups of hatchlings, there was a significant positive cor- relation between the days since the initiation of dispersal and the distance dispersed (Table 1). The low R2 values may be attributed to the spread of distances dispersed on the later dates. While the nearest-neighbor distance remained relative- ly constant, the distance from the maternal bur- DiSTANCE CATEGORY (CM) Figure 2.— Frequency distribution of dispersal distances for 68 marked juvenile Geolycosa xera archboldi at Archbold Biological Station, June 1994. Distances are from release site to new burrow site. MARSHALL— GEOLYCOSA AGGREGATION FORMATION 149 row increased sharply between days 3 and 6 (Fig. 1). Almost all of these nearest-neighbors were siblings. I found 68 of the 80 spiders I released. The marked spiders dispersed an average of 43.9 cm before building a burrow (Fig. 2, 43.9 ± 38.4 cm, n = 68, range 7-240 cm). My failure to find 12 of the released animals will bias my results to the shorter distance categories as these indi- viduals are likely to have dispersed further than most (an artifact of my owe searching behavior). Out of the 14 experimentally released spiders, 10 settled while under observation. The time to initiate burrow construction was 48:32 ± 15:23 min: sec (mean ± 1 SD) The remaining four took longer than 90 min, and I found them in new burrows the next morning. DISCUSSION Territorial aggregations of G. xera initially form and are maintained by limited dispersal Spi- derlings leaving the maternal burrow apparently disperse only as far as they have to avoid their territorial siblings. I found no evidence for con- specific attraction in this spider. However, ag- gregations persist in spite of relocation rates as high as 3.2% a day (Marshall in press). The rea- son for this seems to be the limited dispersal that these spiders exhibit even when relocating. Geo- lycosa xera is highly mobile on the sand (being an ambush predator), yet over half these spiders settled within 30 cm of the release site. I believe that this limited dispersal is an evolved strategy rather than a maladaptive lack of vagility. Dis- persal is assumed to be riskier than non-dispersal (South wood 1962; Gaines & McClenaghan 1980; Johnson & Gaines 1990). In the case of G. xera , an important potential cost of dispersal is the risk of mortality due to cannibalism. Geolycosa xera periodically close their burrows (e. g., when molting or after catching a large prey item). These periods of burrow closure last up to two weeks or longer (Marshall in press). Apparently dis- persing individuals are unable to assess the lo- cation of the closed burrow of a larger conspe- cific, and I have seen smaller spiders settling within the territory of larger individuals with closed burrows. The correlation of the abandon- ment of the burrow of the luckless settler with the re-opening of the larger resident's burrow is suggestive. This uncertainty associated with site choice underlies the risk of relocation, making it analogous to dispersing in a mine field. I found no evidence for any ecological predictors of bur- row site location within the microhabitat (Mar- shall 1 994). Thus, given open sand, I hypothesize that burrow sites are chosen only to avoid active larger conspecifics. There is no advantage to long- distance dispersal, given the uncertainty of the location of closed burrows and the risk implicit in crossing space defended by potentially can- nibalistic territory holders. While inbreeding has been hypothesized as a cost of limited dispersal (Johnson & Gaines 1 990), it is not likely an issue for G. xera . When male G. xera mature, they abandon their burrow and wander in search of mating opportunities. I have observed wandering adult male G. xera to move between patches, which I have not observed na- tal dispersers to do. Presumably, the greater dis- tances travelled by the males in search of matings reduce the probability of inbreeding within these patches of microhabitat. ACKNOWLEDGMENTS This work was vastly improved by the guid- ance of my doctoral committee: S. Riechert, C. Boake, G. Burghardt, A. Echtemacht, and D. Etnier. The manuscript benefitted from the crit- ical reviews of A. Rypstra and the other members of the Miami University Spider Group. The re- search was funded by NICHD Training Grant (T32-HD-07303), and by grants from the Theo- dore Roosevelt Memorial Fund, Sigma Xi, and Archbold Expeditions. LITERATURE CITED Conley, M. R. 1984. Population regulation and prey community impact of Geolycosa rafaelana (Cham- berlin) (Araneae: Lycosidae). Ph. D. dissertation, New Mexico State Univ., Las Cruces, New Mexico. Fellers, G. & C Drost. 1989. Fluorescent powder - a method for tracking reptiles. Herp. Rev., 20:91— 92. Gaines, M. & L. McClenaghan, Jr. 1980. Dispersal in small mammals. Ann. Rev. Ecol. Syst, 11:1 63— 196. Getty, T. 1981. Competitive collusion: the preemp- tion of competition during the sequential establish- ment of territories. American Nat., 1 18:426-431. Johnson, M. & M. Gaines. 1990. Evolution of dis- persal: theoretical models and empirical tests using birds and mammals. Ann. Rev. Ecol. Syst, 21:449- 480. Lemen, C. & P. Freeman. 1985. Tracking mammals with fluorescent pigments: a new technique. J. MammoL, 66:134-136. Marshall, S. D. 1994. Territorial aggregation in the burrowing wolf spider Geolycosa xera archholdi McCrone: formation, maintenance, and conse- 150 THE JOURNAL OF ARACHNOLOGY quences. Ph. D. dissertation, Univ. Tennessee, Knoxville. Marshall, S. D. 1995. Natural history, activity pat- terns, and relocation rates of a burrowing wolf spi- der: Geolycosa xera archboldi (Araneae, Lycosidae). J. Arachnol., 23:65-70. McQueen, D. J. 1983. Mortality patterns for a pop- ulation of burrowing wolf spiders, Geolycosa dom- ifex (Hancock). Canadian J. Zool., 61:1263-1271. Miller, G. L. 1989. Subsocial organization and be- havior in broods of the obligate burrowing wolf spi- der Geolycosa turricola (Treat). Canadian J. Zool., 67:819-824. Miller, G. L. & P. R. Miller. 1984. Correlations of burrow characteristics and body size in burrowing wolf spiders (Araneae: Lycosidae). Florida Ento- mol., 67:314-317. Morse, D. 1993. Some determinants of dispersal by crab spiderlings. Ecology, 74:427-432. Orians, G. H. 1961. Social stimulation within black- bird colonies. Condor, 63:330-337. Richardson, R. K. 1990. Life history, soil associa- tions, and contests for burrows in a burrowing wolf spider, Geolycosa missouriensis Banks. Disserta- tion, Univ. Oklahoma, Norman, Oklahoma, USA. Southwood, T. 1962. Migration in terrestrial arthro- pods in relation to habitat. Biol. Rev., 37:171-214. Stamps, J. A. 1987. Conspecifics as cues to habitat quality: A preference of juvenile lizards ( Anolis ae- neus ) for previously used territories. American Nat., 129:629-642. Stamps, J. A. 1988. Conspecific attraction and ag- gregation in territorial species. American Nat., 131: 329-347. Stamps, J. A. 1991. The effect of conspecifics on hab- itat selection in territorial species. Behav. Ecol. So- ciobiol., 28:29-36. Manuscript received 14 February 1995, revised 9 May 1995. 1995. The Journal of Arachnology 23:151-156 A COMPARISON OF POPULATIONS OF WOLF SPIDERS (ARANEAE, LYCOSIDAE) ON TWO DIFFERENT SUBSTRATES IN SOUTHERN FLORIDA David B. Richman: Dept, of Entomology, Plant Pathology and Weed Science, New Mexico State University, Las Cruces, New Mexico 88003 USA; Jan S. Meister1: Dept, of Entomology and Nematology, University of Florida, Gainesville, FL 326112; Willard H. Whitcomb2: Dept, of Entomology and Nematology, University of Florida, Gainesville, Florida 32611, and Leigh Murray: Dept, of Experimental Statistics, New Mexico State University, Las Cruces, New Mexico 88003 USA ABSTRACT. Wolf spiders were sampled from sandy and grassy substrates every month for one year at Archbold Biological Station, Lake Placid, Florida, from December 1981 through November 1982. It was found that the faunas were different in species composition, even though they were within a few meters of one another. Light-colored species, such as Lycosa ceratiola and Lycosa osceola, were more abundant on or restricted to the sandy surface, while darker colored species, such as Lycosa miami, Lycosa annexa, Lycosa abdita and Schizocosa crassipes were more abundant on or restricted to, the grassy substrate. A total of twelve species of lycosids were collected. Archbold Biological Station currently contains 2023.5 hectares of Lake Wales Ridge scrub, dominated by slash pine, Pinus elliotti Engleman, sand pine, Pinus clausa (Engleman) Sargent, sev- eral species of scrub oaks, Quercus spp., saw pal- metto, Serenoa repens (Bartr.) Small and Sabal etonia Swingler ex. Nash, among others (Vander Kloet 1979). The scrub region of south Florida is unique and the largest part of the remaining scrub on the Lake Wales Ridge occurs on the station. The wolf spider fauna is highly varied and contains some unique or nearly unique el- ements, such as Sosippus placidus Brady and Geolycosa xera McCrone (Franz 1982). Richman (1984) listed 15 known species for Highlands County. Most of the species involved are little known, other than their original descriptions. Two of the species which were recorded in the list are very abundant at the station. These are Lycosa ceratiola Gertsch & Wallace and Lycosa miami Wallace. Wallace (1942) included L. miami , but not L. ceratiola, in the lent a group; but both appear to be related, based on their morphology. An observation was made in 1981 by Richman and Whitcomb that there seemed to be a differ- 1 Current address: 11004 SW 67th St., Gainesville, Florida 32608. 2 Current address: 4013 NW 39th Way, Gainesville, Florida 32606. ence in the lycosid fauna on sand compared with the fauna on grass. To document this difference we decided to sample wolf spider populations on both substrates monthly for a year and then com- pare the results to test the hypothesis that there was a distinct difference between the faunas and that it persisted throughout the annual cycle. METHODS An area immediately adjacent to and south of the main building at Archbold Biological Station (27°20'N, 81°20'W) was selected for accessibility and for the presence of both grassy areas and sandy areas. The areas were approximately 10 x 100 m (1000 sq. m) and were perpendicular to each other. The grassy area was just south of the main building and was oriented along a north- south line, whereas the sandy area edge was just south of the south edge of the grassy area and was oriented along a east- west line (Fig. 1). Both substrates were sampled monthly at night from December 1981 through November 1982 (Table 1), using headlighting techniques (Wallace 1937). A minimum of 50 spiders was collected on each substrate during each sampling period (total minimum of 100 spiders/month), except for the first sample (December 1981) when a minimum of 24 was collected on each substrate and for grass in September 1982 when 29 were collected. Depending on the number of people searching 151 152 THE JOURNAL OF ARACHNOLOGY and the time of year, the samples were collected in anywhere from 1 0 minutes to two hours, but usually took 20-30 minutes. We used a relatively standard number of spiders because it was felt that this would give us a relative proportion of adults and immatures. At no time did we attempt to obtain absolute densities. Because the eye shine was used to collect all samples in both cases, rather than just searching for spiders, we feel that the proportions reasonably reflected the popu- lations present on the substrates. While the sand had little vegetation on it, the grass was short on the grassy area and presented no problems in seeing the eye shine. The spiders were preserved in alcohol and later identified and each carapace width measured (including immature speci- mens). Identifications were made by Allen Bra- dy, Hope College, Holland, Michigan; G. B. Ed- wards, Florida State Collection of Arthropods, Gainesville, Florida; and the senior author. Once adults were identified, it became possible to at least tentatively identify the majority of the im- mature specimens. Most of the immature spec- imens were identified by the senior author. Weather data were provided by Archbold Bio- logical Station. Samples of specimens collected are deposited at the Archbold Biological Station (complete synoptic collection), the Arthropod Museum of New Mexico State University (par- tial synoptic collection), and the Florida State Collection of Arthropods (partial synoptic col- lection and most extra and immature speci- mens). Statistical analyses included preliminary fre- quencies and time plots for counts of each species N S Sandy Area — 3* (Fire Lane) Archbold Main Bldg. 100 m r \ L Grassy Area i i i g » i rTHF^Ti |i 4 Railroad Tracks Figure 1 .—Map of Archbold Biological Station main building area, showing the two study sites. Scale is approximate. Map modified from Abrahmson et al. 1984. and substrate combination. Also chi-square tests of homogeneity were conducted to compare oc- currence of different species. These chi-square tests included: substrate by month comparisons for each species; substrate species comparisons for each month; and month by sex comparisons for each substrate and species combinations. Problems were experienced with sparseness in chi-square tables due to low counts for all but two species and all sexes. Finally, an analysis of variance was performed for each substrate and species combination to compare average size over months. All analyses were performed using SAS© (SAS Institute 1989, 1990). RESULTS A total of five species of Lycosa , three species of Rabidosa, one species of Gladicosa, one spe- cies of Pardosa, one species of Sosippus, and one species of Schizocosa was collected at the study site over the year (Table 2). Of these, eight species Table L— Collecting dates for Lycosa spp. at Archbold Biological Station, Highlands County, Florida (198 1— 1982). * Two teams used. Date Time of collection Grass substrate Sand substrate 17 December 1981 21 January 1982 22 February 1982 22 March 1982 20 April 1982 19 May 1982 21 June 1982 22 July 1982 18 August 1982 20 September 1982 26 October 1982 22 November 1982 Approximately 7:00-7:20 P.M. EST* 7:30-8:00 P.M. EST 7:45-8:05 P.M. EST 7:50-8:20 P.M. EST 9:05-9:15 P.M. EDT 9:00-9:30 P.M. EDT 9:00 PM EDT* 9:20-1:15 P.M. EDT 8:30-9:30 P.M. EDT* Approximately Approximately hour after sundown 7:20-7:36 P.M. EST 8:00-8:30 P.M. EST 8:05-8:30 P.M. EST 8:25-8:55 P.M. EST 9:15-9:30 P.M. EDT 9:30-10:00 P.M. EDT 9:00 P.M. EDT* 8:35-9:20 P.M. EDT 8:30-9:30 P.M. EDT* hour after sundown hour after sundown RICHMAN ET AL.-WOLF SPIDERS IN SOUTHERN FLORIDA 153 Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov MONTHS 1981-82 Figure 2.— Numbers of Lycosa ceratiola, Lycosa miami and other lycosids on sand at Archbold Biological Station, Highlands County, Florida, December 1981 to November 1982. were not listed by Richman (1984). Only two of the species collected in this study, L. ceratiola and L. osceola, were in the Archbold Biological Station Collection at the time of the study. In addition to the wolf spiders, at least one im- mature specimen of the giant crab spider, Het- eropoda venatoria (Linnaeus), and one of an un- identified gnaphosid were collected on the grassy surface over the year. The chi-square test of homogeneity of occur- rence of species on the two substrates indicated that there was indeed a significant difference (P < 0.005 for all months) between the frequency of Lycosa ceratiola and Lycosa miami on the two substrates. Only the samples from August were questionable because 25% of the cells had ex- pected counts of less than five. This sparseness was a common problem when more than these MONTHS 1981-82 Figure 3. —Numbers of Lycosa ceratiola, Lycosa miami and other lycosids on grass at Archbold Biological Station, Highlands County, Florida, December 1981 to November 1982. 154 THE JOURNAL OF ARACHNOLOGY Dec Feb Apr Jun Aug Oct Jan Mar May Jul Sep Nov MONTHS 1981-1982 High Temp Low Temp Rainfall Figure 4.— Weather data for Archbold Biological Station, Highlands County, Florida, December 1981 to November 1982. Temperatures are averages for each month in °c. Rainfall is total for month in cm. two species were included in the analysis, even though it is evident that there is a real difference between the species recovered on sand and those recovered on grass. Thus, no statistically valid conclusions can be made about the other species. Monthly counts (Figs. 2, 3) for the two species demonstrate this distinct difference very well. It is quite obvious that the sandy colored L. cer- atiola preferred sandy substrates, while the darker L. miami preferred the grassy substrate. Size analysis was only meaningful for L. cer- atiola on sand and L. miami on grass. These were present throughout the 12 months of the sam- pling period. The average carapace width for these two species each month is shown in Table 3. It is curious that while immatures were present at all times of the year, adult males of L. miami were found in only four scattered months (De- cember, March and July-August). Adult females of L. miami were collected in December and March-August, whereas males of L. ceratiola were collected from March-July and females in Jan- uary and March-No vember (Table 3). Some prey records were obtained, especially Table 2. —Species of Lycosa and other Lycosidae collected at Archbold Biological Station, Highlands County, Florida, 198 1-1982. Specimens of Lycosa ceratiola and Lycosa miami include those collected on both sand and grass and so do not match totals in Table 3. Identifications were by Allen Brady, G. B. Edwards and D. B. Richman. Species with “*” were not listed by Richman (1984). Species Substrate n and sexes Gladicosa pulchra (Keyserling)* Grass IS Lycosa abdita Gertsch* Grass, sand 26, 102, 3 imm. Lycosa annexa Chamberlin & Ivie* Grass 23S, 132, 41 imm. Lycosa ceratiola Gertsch & Wallace Sand, grass 23a, 552, 561 imm. Lycosa miami Wallace Grass, sand 12a, 222, 432 imm. Lycosa osceola Gertsch & Wallace Sand (1 on grass?) 32, 2 imm.? Pardosa littoralis Banks* Grass la, 12, 5 imm. Rabidosa hentzi (Banks)* Grass (2 on sand) la, 7 imm. Rabidosa punctulata (Hentz)* Grass 1 2, 3 imm. Rabidosa rabida (Walckenaer)* Grass 2 imm. Schizocosa crassipes (Walckenaer)* Grass (1 on sand) 3a, 102, 12 imm. Sosippus floridanus Simon Grass 1 imm. RICHMAN ET AL.^WOLF SPIDERS IN SOUTHERN FLORIDA 155 Table 3.— Mean carapace width (mm) for Lycosa ceratiola on sand and Lycosa miami on grass at Archbold Biological Station, December 198 1 to November 1982. Adult males and females present in the sample are noted by the symbols 6 or 2. Date Lycosa ceratiola (n) (Range) (SD) Lycosa miami (n) (Range) (SD) December 3.10 (29) (1.17-5.67) (1.36) 3.80 (21) (1.67-8.17) (1.91) 16, 12 January 2.62 (54) (1.17-5.83) (1.21) 28 2.66 (53) (1.33-5.17) (0.83) February 2.50 (46) (1.00-5.50) (1.05) 2.38 (34) (1.50-5.00) (0.77) March 4.01 (50) (1.16-7.50) (1.62) 8<5, 4$ 3.59 (52) (1.83-7.83) (1.39) U, 22 April 3.66 (59) (1.50-6.50) (1.47) 26, 42 4.06 (50) (1.83-8.83) (1.39) 42 May 3.68 (54) (1.67-6.67) (1.33) la, 32 4.23 (46) (2.50-8.67) (1.30) 19 June 4.43 (79) (1.00-7.50) (1.82) 66, 162 4.04 (35) (1.33-7.33) (1.74) 22 July 3.36 (52) (1.33-7.00) (1.67) 46, 8$ 5.25 (37) (1.67-8.50) (1.91) 66, 72 August 2.74 (54) (1.67-7.33) (1.14) 42 4.48 (23) (1.50-7.83) (2.04) 46, 42 September 2.56 (59) (0.83-5.33) (1.33) 42 5.37 (18) (1.67-7.00) (1.12) October 3.10 (42) (1.17-5.33) (1.15) 12 3.28 (29) (0.83-6.67) (1.44) November 3.39 (45) (1.33-6.00) (1.35) 5$ 2.90 (19) (1.17-5.33) (1.17) for Lycosa ceratiola. On 20 April 1982 at least 10 individuals of this species were collected with alates of the fire ant Solenopsis sp. (Formicidae). The same collection produced a click beetle (Ela- teridae), a scarab beetle Ataenius platensis (Blan- chard) (Scarabaeidae) and a leafhopper, Drae- culacephala inscripta Van Duzee. On 22 July 1 982 individuals of L. ceratiola had caught a muscid fly, two scarabs and a mirid bug in the genus Lygus. On the 1 8 August 1982 a L. ceratiola was taken with a beetle fragment. On the same date a specimen of L miami had captured a rove beetle (Staphylinidae). On 22 October 1982, a record for made of L miami with a male gryllid cricket, and on 22 November, with one psychid larva and a myrmycine ant. Weather data are summarized in Fig. 4. The lowest temperature was - 1 1 °C in January 1982 and the highest was 35.6 °c in August. The high- est rainfall was in June (27.85 cm) with the sec- ond highest in September (26.63 cm). December 1981 had the lowest rainfall (0.55 cm). With the exception of the cold periods in January and De- cember (to -=6.7 °c), the weather was very mild, even at night, and spiders were always found. DISCUSSION With these results we can make the following statements: 1), Lycosa ceratiola is most often found on the sandy substrate where it is most likely to be cryptic; 2), Conversely, Lycosa mia- mi is most often found on the grassy substrate where it is most likely to be cryptic; 3), The grassy substrate has a higher species diversity (2:1) than the sandy substrate; 4), There were also more specimens of species that were found on both substrates (other than L. ceratiola ) on grass than on sand; 5), The size analysis indicated that the various sizes were spread through the year, but that the smallest spiderlings were found in Au- gust for Lycosa ceratiola and in September for Lycosa miami. However, average carapace widths were the smallest for both species in February (Table 3). The low carapace widths in winter reflects the lower number of adults and the high numbers of (probably) third instar spiderlings. Thus most egg sacs were probably produced from the late summer and fall, although this may have been somewhat more spread out than for more northern species; finally, 6), prey, especially of L. ceratiola , appears to be quite varied and in- cludes ants, beetles, crickets, true bugs, leafhop- pers and caterpillars. Alate ants are readily at- tacked during mating flights. The apparent crypsis of the two major species of Lycosa may be somewhat puzzling, as these are obviously nocturnal spiders. However, it might be noted that on moonlight nights the sandy color of L. ceratiola made it nearly invisible, whereas specimens of L. miami were visible as almost shadow-like bodies against the white sand. The reverse was true on the grassy areas. Thus the crypsis may serve to conceal the spiders from attack by vertebrate predators, especially night birds, such as screech owls, which are known predators of large lycosid spiders (Ross 1969). The lack of female spiders carrying spiderlings or eggs is possibly a result of the larger, com- moner ( lenta group) species staying in their bur- rows during the egg laying and early spiderling 156 THE JOURNAL OF ARACHNOLOGY stages and, in the case of the smaller, rarer, spe- cies, may be a result of the low numbers taken. It is well established that members of the lenta species group are borrowers (Wallace 1942; Gertsch 1979). The less-collected species, some of which normally carry their eggs and first instar young with them, rarely came out on the grassy or sandy areas, but primarily stay on surfaces under the scrub canopy. The one specimen of Sosippus collected was obviously a stray, as these spiders build funnel webs in litter and shrubs (Brady 1962). Also, since only a few of the bur- rowing Lycosa osceola were collected, even on sandy surfaces, it is thought that they may have preferred sandy areas between plants in the near- by scrub. The scrub habitat was not sampled during the study, partly because of the difficulty of headlighting in that habitat. However, L. os- ceola was seen in the spaces between scrub plants on several occasions. The fact that some large species of Lycosa dif- fer in their substrate preference has been dem- onstrated in the past for two members of the lenta group, L. lenta Hentz and L. ammophila Wallace (Harper 1 97 1). In these species L. lenta preferred leaf-litter and L. ammophila preferred sand. As noted earlier, Wallace (1942) placed L. miami, but not L. ceratiola, in the lenta group. It seems likely, from our observations and from the ap- parently similar morphology and appearance (es- pecially the structure of the male palpi, female epigyna and general color pattern- L. ceratiola looks much like a very pale L. miami), that these two species, like L. lenta and L. ammophila, be- long in the same species group. Where these “Ly- cosa ” species will eventually be placed is unclear at the present time. However, it seems to be evident from both Harper (1971) and the current study that the various species of lycosids do par- tition their habitats in peninsular Florida. ACKNOWLEDGMENTS We would like to acknowledge the help of Anne Trambarulo, then a graduate student, Dept, of Entomology and Nematology, and Jonathan Re- iskind, Dept, of Zoology, University of Florida, Gainesville. Both caught spiders for us during at least one of our sampling periods. We especially want to thank Allen Brady, Hope College, Hol- land, Michigan, and G. B. Edwards, Paul Skelley and Frank Mead, all of the Florida State Collec- tion of Arthropods, Gainesville, Florida, for the identifications of the adult spiders and prey re- cords. We also thank the then director, James N. Layne, and the staff of Archbold Biological Sta- tion for their cooperation and support during this study. The two reviewers of this paper for the Journal of Arachnology, Gail Stratton and Jan Weaver, made a number of valuable suggestions which improved the presentation of data and the discussion and we are grateful for their efforts. LITERATURE CITED Abrahmson, W. G., A. F. Johnson & J. N. Layne. 1 984. Archbold Biological Station vegetation map. Archbold Biol. Stn. Brady, A. R. 1962. The spider genus Sosippus in North America, Mexico, and Central America (Ara- neae, Lycosidae). Psyche, 69:129-164. Franz, R. (ed.). 1982. Invertebrates. Rare and en- dangered biota of Florida. 6. Univ. of Florida Press. 131 P- Gertsch, W. J. 1979. American spiders. Van Nos- trand Reinhold. New York. 274 p. Harper, C. A. 1971. Comparative ecology of two sib- ling species of wolf spiders (Araneae, Lycosidae). Ph. D. Dissertation, Univ. of Florida, 99 p. Richman, D. B. 1984. A revised list of the spiders of Highlands County, Florida. Report to Archbold Biol. Stn. 3 p. Ross, A. 1969. Ecological aspects of the food habits of insectivorous screech-owls. Proc. Western Foun- dation Vert. Zool., 1:302-344. SAS Institute, Inc. 1989. S AS/ST AT User’s Guide, Version 6, 4th ed., Vol. 2, Cary, North Carolina. 846 p. SAS Institute, Inc. 1990. SAS Procedures Guide, Ver- sion 6, 3rd ed., Cary, North Carolina. 705 p. Vander Kloet, S. P. 1979. Florula Archboldiensis: Being an annotated list of the vascular plants of the Archbold Biological Station. Report for Archbold Biol. Stn., 65 p. Wallace, H. K. 1937. The use of the headlight in collecting nocturnal spiders. Entomol. News, 48: 107- 111. Wallace, H. K. 1 942. A study of the lenta group of the genus Lycosa, with descriptions of new species (Araneae, Lycosidae). American Mus. Nov., 1185, 21 p. Manuscript received 26 October 1994, revised 7 July 1995. 1995. The Journal of Arachnology 23:157-164 OBSERVATIONS ON THE NATURAL HISTORY OF AN UMMIDIA TRAPDOOR SPIDER FROM COSTA RICA (ARANEAE, OEM ZI DAE) Jason E. Bond1 and Frederick A. Coyle: Department of Biology, Western Carolina University, Cullowhee, North Carolina 28723 USA ABSTRACT. An Ummidia trapdoor spider species near San Vito, Costa Rica, prefers steep slopes and open, sparsely wooded, early successional stage habitats. This habitat preference and the paucity of small juvenile burrows near adult burrows are consistent with spiderling dispersal by ballooning, known to occur in other Ummidia species. The entrance and burrow architecture and prey capture and defensive behavior of this species are similar to those of the few other observed Ummidia species. Ummidia' s door-holding defensive behavior is described in detail for the first time. Two enigmatic phenomena were observed: door hinges were often tilted well away from the horizontal plane, and one spider was found on two successive afternoons with the anterior half of its body fully exposed as it held onto the inner surface of its fully open trapdoor. The trapdoor spider genus Ummidia, which is distributed widely across the southern United States and south through Mexico, the Caribbean, and Central America (Raven 1 985), may contain as many as 100 species (N. Platnick pers. comm.). Scattered accounts of burrow structure and con- struction behavior (Moggridge 1873; Atkinson 1886a, b, c; Coyle 1981), ballooning (Baerg 1928; Coyle 1985; Coyle et al. 1985), prey capture (Coyle 1981), and other facets of Ummidia nat- ural history (Gertsch 1979; Coyle 1981) have been published. We report here the first obser- vations of the biology of a Central American species, probably Ummidia rugosa (Karsch 1 880). METHODS Spiders were studied on the grounds of the Las Cruces Field Station of the Organization for Tropical Studies near San Vito, Puntarenas Province, in southwestern Costa Rica near the Panama border. The station is located at an el- evation of 1095 m, receives an average of 3600 mm of rain per year, and includes a remnant (about 100 ha) of old-growth premontane rain forest, the once widespread climax ecosystem of the region. Our observations were made on 2-3 March 1992, 3-6 March 1993, and 5-7 March 1995 during the dry season, which lasts from January through March. Four burrow entrances ‘Current address: Department of Biology, Virginia Polytechnic Institute and State University, Blacks- burg, Virginia 24061-0406 USA were observed by Coyle in 1992, 16 by Bond in 1993, and 31 by Coyle in 1995. Most entrances studied in 1992 and 1993 were observed and measured again in the subsequent studies. One adult female burrow was excavated, measured, and photographed in each year. Drawings of the body and spermathecae of the adult female col- lected in 1993 (Figs. 1, 2) and the following de- scription of this and a second (gravid) female collected in 1995 will help identify the species. Both specimens are deposited in the American Museum of Natural History. All measurements are given in millimeters. Values for the gravid female are in parentheses. Body length (not including chelicerae) 20.7 (21.4). Carapace 10.02 (9.21) long, 8.52 (7.89) wide. Abdomen 11.52 (12.69) long, 8.02 (8.71) wide. Deep procurved thoracic groove, 1.84(1 .72) wide and 6.72 (5.98) from anterior margin of carapace. Carapace uniform chestnut brown, pars cephalica with median longitudinal row of strong setae flanked by two longitudinal clusters of se- tae. Eye diameters: AME 0.28 (0.20), ALE 0.62 (0.57), PME 0.26 (0.22), PLE 0.28 (0.22). Eye interdistances: ALE-ALE 0.90 (0.86), AME-AME 0.24 (0.17), AME-ALE 0.44 (0.31), AME-PME 0.20 (0. 1 8), ALE-PLE 0. 1 8 (0.20), PLE-PLE 1.34 (1.14), PME-PME 0.68 (0.60), PME-PLE 0.10 (0.08). Endites, labium, and sternum light brown. Sternum 5.56 (4.90) long, 5.23 (4.73) wide. With 5-8 slit sensilla scattered on each side of sternum. Legs dark brown, no distinct markings. Pro- and retromargins of chelicerae each with 8-10 teeth. 157 158 Leg formula IV-I-II-III, with leg II only slightly longer than leg III. Leg I article lengths: coxa 3.24 (2.68), trochanter 1.41 (1.04), femur 5.98 (5.40), patella 3.74 (3.20), tibia 3.98 (3.56), metatarsus 2.49 (2.02), tarsus 1.58 (1.30). Many trichoboth- ria and 2-8 club-shaped bothria dorsally on tar- sus of each leg. Three to five club-shaped bothria on dorsal surface of palpal tarsi. MICROHABITAT, HABITAT, AND DEMOGRAPHY Burrows were found only on steeply sloping (60-90°) earthen trail banks, not on gently slop- ing or level ground. No burrows were found dur- ing a careful 30 min search on hands and knees for entrances in the level lawn immediately above the bank where most burrows occurred. The ap- parent preference of these spiders for steep banks, a preference not exhibited by western North Car- olina Ummidia (which are typically found on level and gently sloping ground), may be the re- sult of the heavy rains experienced by the Costa Rica population selecting against any proclivity to construct burrows in flood-prone ground. All but five of the burrows were situated in stable soil partly covered with moss and sheltered from rain and runoff by overhanging roots or other materials. The entrances of burrows found in less sheltered microhabitats showed evidence of ero- sion damage; two of these projected nearly 20 mm above the surrounding soil. Most burrows were located on the non-forest- ed grounds of the Station along the end of the trail leading to the “Mirador.” The rest were located in very young second growth forest with- in 20 m (“Jungle Trail”) and 200 m (trail to Rio Jaba) of the Station grounds. Searches along sta- ble trail and stream banks in older second growth and primary forest failed to locate any burrows. The apparent preference of these spiders for dis- turbed, sparsely wooded habitats is shared by Ummidia populations in western North Carolina (F. Coyle pers. obs.) and Arizona and Mexico (Gertsch 1979) and may be linked to a reliance on aerial dispersal by ballooning, a trait observed in populations in North Carolina (Coyle 1985; pers. obs.) and Arkansas (Baerg 1928). Such a relationship would fit Greenstone’s hypothesis that less predictable habitats select for higher rates of ballooning (Greenstone 1982; Coyle et al. 1985). The type of ballooning characteristic of Ummidia and other mygalomorphs, ballooning which apparently requires air currents stronger than gentle updrafts (Coyle 1983, 1985), might THE JOURNAL OF ARACHNOLOGY Figures 1, 2.— Adult female Ummidia collected in 1993 from Las Cruces, Costa Rica. 1, Dorsal view without pedipalps and legs; 2, Spermathecae. BOND & COYLE-NATURAL HISTORY OF AN UMMIDIA SPECIES 159 be especially ineffective and risky for a fossorial spider living in an old growth forest, where the necessary breezes are probably rare to nonexis- tent except at the top of the canopy. Despite much effort in 1993 and 1995 to find smaller burrow entrances, the great majority (94% in 1993 and 65% in 1995 ) were large (with door widths of 1 5-29 mm) and belonged to near adult or adult spiders, judging from the door widths (20-23 mm) of the three excavated burrows, all of which contained adult females. About half of these burrows were loosely clustered in groups of 2-4 burrows per m2; others were more isolat- ed. The only smaller entrances found were one with a 10 mm wide door (1993), two with 13 mm wide doors (1995), and nine much smaller entrances of very young individuals ( 1 995). These latter entrances were 1.5-3. 5 mm wide, and all but two of these burrows were unoccupied, as evidenced by severely damaged or missing doors. Two of these very small burrows were each about 50 mm from one adult burrow, another two were 10 mm and 15 mm from another adult burrow, and the other five were between 35 mm and 1 80 mm from a third adult burrow. This paucity of young burrows in locations where adults are common provides additional support for the hy- pothesis that spiderling ballooning is the primary dispersal mode of this species. Adult female bur- rows of non-ballooning fossorial spiders like the antrodiaetids are often surrounded by numerous burrows of early instars (Coyle 1971; Coyle & Icenogle 1994), whereas juveniles of a North Carolina Ummidia population known to balloon are almost never found near an adult burrow. The higher ratio of unoccupied/occupied very small burrows (0.78) than unoccupied/occupied large burrows (0.13) found in 1995 is consistent with the expectation that young juveniles expe- rience particularly high mortality rates because their high surface to volume ratio and shallow burrows make them especially vulnerable to en- vironmental crises like drought and erosion. ENTRANCE AND BURROW STRUCTURE The trapdoor is relatively thick and rigid with beveled edges (Figs. 3, 4, 10, 11; Table 1). When the door is closed, these edges fit snugly into the tough entrance rim which flares outward to form a complementary bevel. The entrance rim is usu- ally nearly flush with the surrounding soil but may extend as much as 5-10 mm above it. The door is composed of soil and silk. Its inner sur- face is covered with a thick, tough white layer of silk, and its outer surface, which is soil with bits of dead plant material and sometimes moss, re- sembles the surrounding ground surface (Figs. 3- 5, 10-12). The door is connected to the entrance rim by a broad hinge, the bulk of which is thick tough silk continuous with the entrance rim and burrow lining. On the outer surface of some doors are roughly concentric semi-circular ridges or flaps (Figs. 3, 4) which are probably old, smaller doors to which more soil and silk were added as the spider grew. Pieces of plant material and irreg- ular tabs of silk plus soil up to 10 mm long often extend from the entrance rim lips and door edges (Figs. 10, 11) and, like the linear litter and tabs attached to the entrances of other trapdoor spi- ders (Coyle 1986; Coyle et al. 1992; Coyle & Icenogle 1994), may serve to extend the spider’s prey-sensing radius. Data on burrow entrance structure is sum- marized in Table 1. The shape of the door is similar to that of many trapdoor taxa (Fig. 9). However, a most peculiar feature of these en- trances is their orientation. Contrary to the con- sistently near-horizontal hinges of most other trapdoor species living on steep slopes (Coyle 1986; Coyle & Icenogle 1 994), most of these Um- midia doors were oriented with the hinge tilted well away from the horizontal, and some were vertical or nearly so (Figs. 5, 11). In 1995, left tilting hinges (8) were less common than those tilting to the right (12). Figures 4, 5, and 6-8 illustrate the three ex- cavated burrows. All three extended roughly straight back (100-160 mm) into the trail bank, approximately perpendicular to the surface. As has been observed for other Ummidia species (Gertsch 1979), burrow diameter was fairly con- stant throughout each burrow’s length. The full length of each burrow was lined with silk, but the silk was much thicker near the entrance than elsewhere (Figs. 4, 5). The upper third of the longest burrow was lined with an especially thick leathery lining composed of several layers of silk and soil that were probably applied in response to the very loose soil in that spot (Fig. 5). PREY CAPTURE BEHAVIOR The foraging posture and prey capture behav- ior of these spiders are similar in form to those of the North American Ummidia studied by Coyle (1981). Approximately 30 min after sunset the Costa Rican spiders assumed the foraging pos- 160 THE JOURNAL OF ARACHNOLOGY Figures 3-5.— Photos of burrows of adult females of Ummidia at Las Cruces, Costa Rica. 3, Entrance with trapdoor propped open to show small broken silk seal and small old door attached to upper surface of functional trapdoor; 4, Side view of burrow and trapdoor excavated (same as in Fig. 3) in 1992, with female (upper wall of upper end of burrow has collapsed so that door is shifted from its normal position); 5, Side view of burrow excavated in 1993, with female, showing nearly vertical hinge and upper surface of partly open trapdoor. ture; the trapdoor is opened slightly (1-3 mm) and the spider is positioned just below the door with the tarsi of its pedipalps and first and second legs resting on the lip of the entrance rim (Fig. 10). Several arthropods were placed near en- trances to elicit prey capture responses. Only three capture attempts were observed; an opilionid was attacked when it touched the trapdoor with one BOND & COYLE — NATURAL HISTORY OF AN UMMIDIA SPECIES 161 Table 1.— Orientation, dimensions, and shape indices of Ummidia trapdoors from Las Cruces Field Station, Costa Rica. Range, mean, and standard deviation given. Only larger entrances (doors over 14 mm wide) included. 1995 sample (n = 20) was measured by Coyle and included many of the entrances in the sample (n = 15) measured by Bond in 1993. Door hinge index = door width/hinge width, and door shape index = door width/ door height; see Fig. 9. 1993 sample 1995 sample Door thickness (mm) 1. 5-3.0, 2.46 ± 0.57 Hinge angle (degrees) 0-90, 51.0 ± 33.4 0-85, 40.1 ± 24.0 Hinge width (mm) 13-21, 16.9 ± 2.0 14-25, 18.7 ± 3.0 Door width (mm) 15-27, 19.3 ± 3.0 16-29, 21.6 ± 3.0 Door height (mm) 11-22, 16.9 ± 2.6 Door hinge index 1.00-1.29, 1.14 ± 0.10 1.00-1.29, 1.16 ± 0.07 Door shape index 1.13-1.47, 1.29 ± 0.09 of its tarsi, a broad caterpillar covered with white scalelike hairs was attacked when it crawled within 1-2 mm of the trapdoor, and a small he- mipteran was captured when it walked within 3- 5 mm of the trapdoor. In all cases the door popped open and the spider lunged from the entrance, exposing all but the posterior half of its abdomen while holding onto the burrow with its fourth and possibly third legs. Prey were grabbed with the pedipalps and first legs. The opilionid and caterpillar were immediately released, but the pedipalps and legs flexed to pull the hemipteran close to the chelicerae for the strike, after which the spider immediately retreated into the burrow with prey. DEFENSIVE BEHAVIOR At night spiders were quick to pull their doors shut and hold them tightly closed when the en- trances were directly illuminated by a flashlight. During daylight hours, attempts to open (with a knife blade) or measure closed doors often caused the spiders to quickly pull the doors tightly closed; clearly many spiders monitor their entrances rather closely, even in the daytime when doors are closed and substrate vibrations may be the Figures 6-9.— Drawings of burrows of Ummidia females at Las Cruces, Costa Rica. 6-8, Side view drawings of the three adult female burrows that were excavated; 6, 1995 burrow; 7, 1992 burrow; 8, 1993 burrow; 9, Outline of trapdoor based on mean dimensions of 1 995 sample and showing the three door shape measurements. 162 THE JOURNAL OF ARACHNOLOGY Figures 10- 12. —Photos of burrow entrance behavior of adult females of Ummidia at Las Cruces, Costa Rica. 10, Spider in foraging posture at night; arrows point to claws, from left to right, of right leg II, right leg I, right pedipalp, left pedipalp, and left leg I touching entrance rim; 1 1 , Spider attempting to close door with fangs and claws of pedipalps, legs I, and legs II; 12, Spider attempting to close fully opened door with claws and fangs while anchoring itself in burrow with legs III. BOND & COYLE — NATURAL HISTORY OF AN UMMIDIA SPECIES 163 only sign of danger. As Gertsch (1979) observed in other Ummidia species, a great deal of force is required to pry open these secured trapdoors with a knife blade or pair of forceps. Even when the door is forced open the spider continues to maintain its grip, revealing its method (Figs. 1 1 , 1 2). With its venter facing the hinge, the spider holds the inner surface of the door with its fangs and the claws of its pedipalps and first and second legs, much as described by Gertsch (1979) and Coyle (198 1) for North American Ummidia spe- cies. Claw and fang marks are found on the un- dersurface of all doors (Fig. 12). The spider an- chors itself to the burrow by pressing the enlarged distal end of the saddle-shaped tibia and adjacent dorsal surface of the metatarsus of each third leg against the burrow wall (Fig. 12), as described for North American Ummidia populations by Coyle (1981), and by presumably holding onto the wall deeper in the burrow with the tarsal claws of its fourth legs. By opening doors in the daytime when spiders were not in contact with their doors, we could see that a spider usually closes its door with its tarsal claws before in- serting its fangs. While this mode of door-holding requires that a spider in foraging position first rotate 180° around its longitudinal axis, an extra maneuver which taxa that hold doors shut only with tarsal claws need not perform (Coyle et al. 1992), the Ummidia mode probably produces much more pulling power than if fangs are not used. Two spiders, one in 1992 and one in 1995, had each lightly sealed its door shut with a thin small patch of silk (Fig. 3) but moved up and closed the door after it was forced open. These seals appeared too weak to prevent predators from forcing the doors open, and instead may serve to hold the door more tightly against the entrance lip during the daytime to reduce the evaporative loss of burrow moisture or ensure crypsis. Another spider exhibited a curious, seemingly vulnerable, posture on two successive afternoons in 1993; it was found motionless and reaching out of the burrow entrance holding onto the undersurface of a fully opened door with its fangs and pedipalp and leg claws so that its pro- soma and anterior abdomen were fully exposed. Upon sensing the observer, it retreated into the burrow and pulled the door shut. ACKNOWLEDGMENTS This research was conducted during Western Carolina University (WCU) Tropical Biodiver- sity class field trips funded in part by WCU. Julie Waterman, Michele Lindsey, and Dan Pittillo kindly provided assistance in the field. We thank the Organization for Tropical Studies and Luis Diego Gomez for their hospitality at the Las Cru- ces Field Station. We also thank Robb Bennett and Bill Eberhard for their careful review of this manuscript. LITERATURE CITED Atkinson, G. F. 188 6a. A new trapdoor spider. Amer- ican Nat., 20:583-593. Atkinson, G. F. 1886b. A family of young trapdoor spiders ( Pachylomerus carolinensis Hentz). Ento- mol. America, 2:87-92. Atkinson, G. F. 1886c. Descriptions of some new trapdoor spiders: their nests and food habits. En- tomol. America, 2:109-117, 128-137. Baerg, W. J. 1928. Some studies of a trapdoor spider (Araneae: Avicularidae). Entomol. News, 39:1-4. Coyle, F. A. 1971. Systematics and natural history of the mygalomorph spider genus Antrodiaetus and related genera (Araneae: Antrodiaetidae). Bull. Mus. Comp. Zool., 141:29-402. Coyle, F. A. 1981. Notes on the behaviour of Um- midia trapdoor spiders (Araneae, Ctenizidae): bur- row construction, prey capture, and the functional morphology of the peculiar third tibia. Bull. British Arachnol. Soc., 5:159-165. Coyle, F. A. 1983. Aerial dispersal by mygalomorph spiderlings (Araneae, Mygalomorphae). J. Arach- nol., 1 1:283-286. Coyle, F. A. 1985. Ballooning behavior of Ummidia spiderlings (Araneae, Ctenizidae). J. Arachnol., 13: 137-138. Coyle, F. A. 1986. The role of silk in prey capture by nonaraneomorph spiders. Pp. 268-305, In Spi- ders: Webs, behavior, and evolution. (W. A. Shear, ed.). Stanford Univ. Press, Stanford, California. 492 pp. Coyle, F. A., R. E. Dellinger & R. G. Bennett. 1992. Retreat architecture and construction behaviour of an East African idiopine spider (Araneae, Idiopi- dae). Bull. British Arachnol. Soc., 9:99-104. Coyle, F. A., M. H. Greenstone, A. Hultsch & C. E. Morgan. 1985. Ballooning mygalomorphs: Esti- mates of the masses of Sphrodos and Ummidia bal- looners (Araneae: Atypidae, Ctenizidae). J. Arach- nol., 13:291-296. Coyle, F. A. & W. R. Icenogle. 1994. Natural history of the Californian trapdoor spider genus Aliatypus (Araneae, Antrodiaetidae). J. Arachnol., 22:225-255. Gertsch, W. J. 1979. American spiders. Van Nos- trand Reinhold, New York, 274 pp. Greenstone, M. H. 1982. Ballooning frequency and 164 THE JOURNAL OF ARACHNOLOGY habitat predictability in two wolf spider species (Ly- cosidae: Pardosa). Florida Entomol., 65:83-89. Karsch, F. 1880. Arachnologische Blatter (Decas I). Zeits. Gesam. Maturw., 53:373-409. Moggridge, I. T. 1873. Harvesting ants and trapdoor spiders. L. Reeve & Co., London. Raven, R. J. 1985. The spider infraorder mygalo- morphae (Araneae): Cladistics and systematics. Bull. American Mus. Nat. Hist., 182:1-180. Manuscript received 19 May 1995, revised 6 September 1995 . 1995. The Journal of Arachnology 23:165-170 CHIVALRY IN PHOLCID SPIDERS REVISITED Julie A. Blanchong, Michael S. Summerfield, Mary Ann Popson, and Elizabeth M. Jakob1: Department of Biology, Bowling Green State University, Bowling Green, Ohio 43403 USA ABSTRACT. Cohabiting pairs of adult spiders are likely to interact over prey, and the outcome of these interactions is likely to affect the reproductive success of both individuals. In two species of pholcid spiders, previous workers reported the occurrence of “chivalrous” behavior, in which males cede prey to females. We looked for the occurrence of chivalrous behavior in another pholcid spider, Holocnemus pluchei. Pairs of spiders were placed on a web and left overnight without prey. A housefly was then introduced onto the web equidistant from the spiders, and subsequent interactions were noted on audiotape. We found no evidence of chivalry in pairs of unknown mating status or in pairs that had recently mated: males and females were equally likely to win the prey, and intensity of interactions over prey was not influenced by the gender of the winner. The differences in our results compared to previously published work may be attributable to the fact that Holocnemus lives in unusually dense populations in nature. This, in combination with a pattern of last-male sperm priority, means that females may be difficult for males to monopolize, and a male will not substantially increase his own reproductive success by ceding prey to a female with which he has mated if others are also likely to mate with her. Web-building spider species vary tremen- dously in the duration of male and female co- habitation. Spiders may interact only during courtship and copulation (Robinson 1982), live permanently in the same colony (see reviews by Buskirk 1982; D’ Andrea 1987), or exhibit be- havior between these extremes (Suter & Wal- berer 1989). While sharing a web, males and fe- males can interact over incoming prey. These interactions also vary greatly, ranging from co- operative prey capture in some social species (Buskirk 1982), to forceful battles over prey (Su- ter 1985). Interactions over prey by cohabiting pairs are potentially of evolutionary importance because an increase in a female’s prey intake is likely to increase her fecundity (e. g., Turnbull 1962) and cohabitation can be relatively pro- longed. The extent to which a cohabiting male benefits from consuming prey himself or allow- ing a female to consume prey may be influenced by whether he has mated with the female, the number of other mates she has had, the pattern of sperm priority, the female’s defendability, and the number of other mating opportunities in the population. A particularly striking example of interactions over prey is the “chivalrous” behavior reported by Eberhard & Briceno (1983), in which male 1 To whom correspondence should be addressed. pholcids ceded prey to females. Blechroscelis sp. males, after attacking prey, would sometimes step aside and allow a female to take it. Occasionally, a female would vibrate her abdomen (a display also seen in male courtship, and interpreted by Eberhard & Briceno as “begging”); this behavior was often followed by the male ceding the prey to her. In two Modisimus species, males initiated most attacks on prey and then usually ceded the prey to females. In one species, the female would approach the male as he wrapped the prey, then he would step aside and allow the female to take the prey. In the other species, males usually com- pletely wrapped the prey, then plucked the web, and left the prey in the web for the female or carried it towards her. Eberhard & Briceno (1983) termed these behaviors “chivalrous” because males sometimes endured partial starvation while allowing females to feed, and we follow their terminology here. We looked for the occurrence of chivalrous behavior in another pholcid spider, Holocnemus pluchei. Several factors make Holocnemus a good species in which to study this behavior. First, it serves as a comparison to the pholcids studied by Eberhard & Briceno (1983). Holocnemus of- ten live in dense populations with interconnect- ing webs and a single sheet of silk may be shared by many spiders of all ages, in contrast to Ble- chroscelis and Modisimus , in which only adult pairs cohabit (Eberhard & Briceno 1985). Sec- 165 166 THE JOURNAL OF ARACHNOLOGY ond, interactions between Holocnemus spiders over prey are relatively well known (Jakob 1991, 1994), allowing us to interpret clearly their be- haviors. Finally, we have some information about the sperm priority pattern in Holocnemus . In Holocnemus , the second of a pair of males to mate with a female fertilizes 65-82% (95% con- fidence interval) of her eggs (Raster 1995). With this information, the implications of chivalrous behavior for a male’s reproductive success are more easily interpreted. In our laboratory study, we introduced prey onto webs shared by male and female pairs of adult spiders and noted subsequent interactions, watching especially for behavior patterns de- scribed by Eberhard & Briceno (1983). We define chivalrous behavior by male spiders to include the following: (1) upon the approach of a female, the male leaves the prey and stands aside as she takes possession, (2) the male wraps and then carries the prey to or in the direction of the fe- male and cedes it to her, or (3) both spiders si- multaneously wrap the prey, and the male sub- sequently moves away, with little or no aggres- sion by the female. If males are chivalrous, they might allow females to attack prey first, and fe- males may be more likely to win interactions over prey. Chivalrous males might cede prey without escalating interactions, so we also ex- amined the relationship between interaction in- tensity and the gender of the winner. We also analyzed relative frequency of other aggressive behaviors for males and females to look for any gender-specific differences. We studied two groups of spiders: pairs that may or may not have mated (with other spiders or with each other) prior to the test (mating status unknown, or MSU), and previously virgin pairs that were observed to mate with one another prior to the test. METHODS Adult and juvenile Holocnemus pluchei were collected in Davis, California in the summer of 1994 and shipped to our laboratory in Ohio. Rearing procedures follow Jakob & Dingle (1990) with the following exceptions. Juveniles were reared to maturity with three feedings per week of fruit flies (. Drosophila sp.) and flour beetle lar- vae ( Tribolium confusum). Adults were fed houseflies (Musca domesticus) and Tribolium larvae twice weekly. Spiders were maintained in a room with 16:8 I D cycle at a temperature of approximately 27 SC. The experimental arenas were four 52 x 37 x 22 cm plastic cages. Because Holocnemus are slow to build large webs in the laboratory, we introduced spiders onto webs built by eonspe- cifics. In the field, Holocnemus routinely use webs that other individuals have built (Jakob 1991), so this is a realistic approximation of field con- ditions. We performed two sets of trials. In the first set of 3 1 trials (mating status unknown, or MSU), field-caught spiders were randomly paired as adults, introduced into a web and left overnight. Because we did not monitor the spiders after pairing and prior to the test, we did not know whether these pairs mated in the laboratory. In the second set of 1 4 trials (mated pairs), mature virgin spiders were randomly paired and intro- duced into a test web, where we observed cop- ulation. Testing occurred approximately 24 hours after copulation. Within one-half hour before the start of all trials, we briefly removed spiders from the test arena without damaging the web, and weighed them on a Mettler balance. After their return to the web, spiders were given a few min- utes to acclimate before testing began. At the start of each trial, a housefly was anes- thetized with C02 and placed in the web with soft forceps approximately equidistant from each member of the pair, which were typically within 25 cm of each other. We made continuous voice recordings of observations with a microcassette recorder, beginning when the fly first moved. We noted whether both spiders oriented to the prey (turned to face the prey), which spider was first to attack the prey, if one spider relinquished the prey without fighting, and which spider ulti- mately won the prey. We also noted if the prey changed possession during the course of the in- teraction: that is, if it was first held by one spider in its chelicerae and later by the other. Trials ended when one spider was feeding on the prey and both spiders had been quiescent for at least 10 min. Previous observations suggest that spi- ders rarely steal prey after feeding begins (Jakob 1991). We classified interactions over prey into three levels after Jakob (1994). Interactions at lower levels are assumed to have lower risk of injury than interactions at higher levels. Level I: Ori- entation to con specific (spider turns its body to face conspecific), pushups (slow leg flexion) and abdomen twitching (fast dorsal/ventral twitching of the abdomen). Level II: Bouncing (sharply contracting its legs so body moves toward the BLANCHONG ET AL.- CHIVALRY IS DEAD 167 Table 1.— Number of trials in which males and females won the prey item, x2 goodness-of-fit tests are against expected values of 50:50. Calculations of the power of the test follow Cohen (1977). Male won Female won x2 P Power Mating status unknown 10 16 1.397 0.391 0.507 Mated pairs 9 5 1.143 0.428 0.513 All trials 19 21 0.100 0.752 0.764 web), approach conspecific, and web plucking (spider spreads its anterior pair of legs, pulls sharply downward on the web, and releases it so the silk snaps back). Level III: Chasing a con- specific, probing and contacting the conspecific with extended front legs, and grappling (locking chelicerae, intertwining legs and appearing to roll about the underside of the web sheet). From tape transcriptions, we calculated interaction time, excluding pauses between activities, for behav- iors of all levels combined and for behaviors of Levels II and III only. For results reported here, we omitted trials in which one of the spiders failed to respond to the prey at any time because we did not know if spiders had detected the vibrations of the prey. In MSU pairs, three males failed to respond to the prey at any time throughout the trial and two females failed to respond. In trials with mated pairs, the male did not respond in one trial and the female did not respond in another. Inclusion of these trials did not change the outcome of the analyses. We compared the frequencies that males and females attacked and captured prey against ex- pected frequencies of 50% with x2 goodness-of- fit tests. We used contingency tests to examine differences in the levels of escalation of inter- actions when males and females won. Contin- gency tables were analyzed with G-tests when cell sample sizes permitted and with x2 contingency analyses for other cases (Sokal & Rohlf 1981). We used nonparametric tests to examine whether Table 2.— Number of males and females that won the prey item for three relative weight classes. Larger Male within 10% of Smaller males female males Male 5 4 10 Female 2 1 18 interactions won by males were of the same du- ration as those won by females. RESULTS General descriptions of interactions.— Males and females were equally likely to attack the prey: in all trials combined, 20 females were first to attack the prey and 20 males were first to attack. MSU and mated trials did not differ significantly: in MSU pairs, 1 1 males and 1 5 females were first to attack, and in mated pairs, 9 males and 5 females were first (contingency table, G2 = 1.777, P > 0. 1 5). There were competitive interactions over prey in every trial, and every trial ended with one spider feeding on the prey. Winner of the interaction.— If males were chivalrous, females would be expected to win interactions over prey more often than males. However, males and females were equally likely to win the prey (Table 1). Mated and MSU pairs did not differ significantly (contingency table, G2 = 2.456, df= l,P> 0.12). We were interested in the combined effects of relative mass and prey on the outcome of inter- actions: chivalry might be occurring if males lost prey to females of smaller mass. This effect might be hidden in the data because, as a group, Hol- ocnemus males were significantly lighter than fe- males with which they were paired (Wilcoxon signed-rank test, Z = — 3.737, P = 0.0002). We categorized males as being within 10% of the mass of their partner (hereafter classified as same size), less than 10% of the female’s mass (small- er males), or greater than 10% of the female’s mass (larger males) (Table 2). Smaller males lost more interactions than expected, but this differ- ence was not significant (contingency table, G2 = 5.474, df= 2, 0.06 < P < 0.07). When we pooled same size and larger males, we found that they were significantly more likely to win fights over prey than were smaller (contingency table, G2 = 5.357, df = 1, P < 0.03). Thus, contrary to pre- dictions from chivalry, males tended to lose in- 168 THE JOURNAL OF ARACHNOLOGY Table 3.— Level of interactions reached by pairs of spiders. Higher interaction levels are considered to be of higher energetic cost and higher risk. Level I Level II Level III Mating status unknown Male wins 1 4 5 Female wins 0 5 11 Mated pairs Male wins 2 1 6 Female wins 0 0 5 teractions only when they were smaller than their partners. Interaction intensity.— Chivalry might occur in a more subtle way: perhaps males gave up prey without escalating interactions to their high- est level. For all trials combined, three interac- tions did not pass beyond Level I, 10 did not pass beyond Level II, and 27 reached Level III, the highest intensity level. There was no rela- tionship between the level of intensity that in- teractions reached and the gender of the winner for MSU pairs (x2 = 2.088, df= 2, P > 0.35), mated pairs (x2 = 2.121, df- 2, P > 0.3), or all trials combined (x2 = 3.836, df= 2, P > 0.14) (Table 3). MSU and mated pairs did not differ significantly in the level of interaction that was reached. No relationship was found between rel- ative masses of a pair (male within 10% of its partner’s mass, less than 10%, or greater than 10%) and interaction intensity (x2 = 1.744, df= 2 ,P> 0.78). Change of possession of the prey.— If chivalry occurs in this species, we would predict that prey would more often change from the possession of the male to the possession of the female. In five MSU pairs, the prey changed possession during the course of the interaction. In three trials, the prey was taken away from the male by the fe- male, and in two trials, the prey was taken away from the female by the male. In mated pairs, the prey changed possession from the male to the female once, and on one occasion, the prey changed from the female, to the male, then back to the female, who consumed it. Behavior.— No spider ever stood aside as an- other took the prey, wrapped the prey and then ceded it without aggression, or simultaneously wrapped the prey with another spider and then moved away. We also looked for more subtle evidence of chivalry by investigating whether females and males differed in their performance of specific agonistic behaviors. We counted the number of trials in which each behavior was performed at least once by the male or female. We found no significant differences in pushups, bouncing, ap- proach, chasing, web plucking or probing. How- ever, the numbers of trials in which the male abdomen twitched was significantly higher than the number of trials in which the female abdo- men twitched (contingency table analysis, male: 32 of 40 (60%); female: 14 of 40 (35%); G2 = 17.269, df= 1 ,P< 0.0001). Interaction duration.— If chivalrous males give up prey to females, interactions that are won by females may be shorter in duration. However, we found no significant relationship between gender of the winner and duration of all inter- actions (excluding pauses) or for duration of all behaviors of level II or III (Mann- Whitney U-tests) (Table 4). When MSU and mated pairs were analyzed separately, no significant differ- ence in any measure of duration was found. DISCUSSION Holocnemus pluchei were not chivalrous. Males and females were equally likely to attack and win prey. However, the power of our test comparing the frequency of winning for each gender (Table 1) indicates that we have a 24% probability of a Type II error, or accepting the null hypothesis when it is false; thus, this result alone does not Table 4.— Mean (±SE) interaction durations (s) in trials that females won compared to trials that males won. P values are derived from Mann-Whitney 17-tests. Duration of all interaction levels Duration of II & III Female won Male won P Female won Male won P Mating status unknown 543 ± 134 492 ± 126 NS 193 ± 76 110 ± 36 NS Mated 932 ± 312 257 ± 76 0.10 389 ± 186 79 ± 27 0.10 All trials 636 ± 128 381 ± 78 NS 171 ± 41 239 ± 73 NS BLANCHONG ET AL.- CHIVALRY IS DEAD 169 firmly establish that these spiders are not chiv- alrous. However, we saw none of the chivalrous behaviors described by Eberhard & Briceno (1983). Fights over prey often escalated. Prey changed hands from male to female approxi- mately as often as it changed from female to male. Males that were within 10% of the body mass of females or larger than females were likely to win prey, suggesting that when males lose prey, it is not because of chivalry but because of a lack of competitive ability. Few differences were found in male and female behavior patterns. Males were more likely to abdomen twitch; the meaning of this behavior is unclear, but we interpret it as a low-risk, low energy behavior. In sum, we found no evidence that males were allowing females an advantage in prey capture. Eberhard & Briceno (1983) suggest that, for Blechroscelis and Modisimus, it is to a male’s advantage to be chivalrous if it results in an in- crease in the number of eggs laid by the female that are sired by the male. Although mating was not directly observed in Eberhard & Briceno’s (1983) study, it is likely that their spiders had mated; only adult pairs cohabit, and paired males will fight with males that are introduced onto the web (Eberhard & Briceno 1985), which is con- sistent with the idea that females are a valuable resource worthy of defense (e. g., Parker 1984). The last pholcid male to mate with a female may father many of her eggs: Austad (1984) predicts that, based on the cul-de-sac shape of the sper- matheca, haplogyne spiders such as pholcids should show last male sperm priority or sperm mixing. Eberhard et al. (1993) found that in an- other pholcid, Physocyclus globosus, sperm pri- ority pattern for twelve females that were each mated with two males did not differ from that expected for random sperm mixing. If Austad’s prediction proves true for Blechroscelis and Modisimus , it should benefit a male to cede prey to a female with which he has just mated. Why, then, are Holocnemus males not chiv- alrous? Kaster (1995) found, using the technique of sterilization by irradiation, that the sperm pri- ority pattern in Holocnemus is highly variable: the second male of a pair of males fertilized be- tween 2.6 and 100% of a female’s eggs. High variability in sperm precedence is common in insects and is as yet unexplained (Lewis & Austad 1990). However, in most of Raster’s pairs of males, the second male fathered most of the eggs (x = 73.7%, 95% confidence interval 65.8-8 1 .6%). It seems clear that the first male to mate with a virgin female is not guaranteed to fertilize the bulk of her eggs if the female has subsequent mates. A male that cedes prey to a female may not gain much benefit in reproductive success if another male mates after he does. Holocnemus differs from the species that Eber- hard & Briceno studied in that Holocnemus fe- males may not be a defendable resource. Hol- ocnemus populations are extremely dense: for example, there may be over 600 spiders on a 3 m x 15 m juniper bush (Jakob unpubl. data). Both males and females move frequently from one web to another (Jakob 1991), so the intrusion rate of potential competitors for a female’s at- tentions and the rate of female encounters with new males are both likely to be high. In the lab- oratory, females readily remate: Kaster (1995) removed males when they finished copulating with a virgin female and immediately introduced a new male, and found that copulation began again in an average of 437 sec ( SE = 101.91, n = 20). The populations of Blechroscelis and Mod- isimus are less dense than those of Holocnemus , and webs of individuals or pairs are discrete (W. G. Eberhard pers. comm.). Male Blechroscelis and Modisimus are more likely to be able to successfully defend females from competitors. An additional effect of the high population density of Holocnemus is that males are likely to have other mating opportunities, which would deval- ue any one mating and make it profitable for a male to increase its energetic intake to allow fur- ther searching for mates. Other variables that would affect a male’s probability of finding more than one mate include predation risk while searching and the male’s expected lifespan. Nei- ther of these, to our knowledge, has been mea- sured for any pholcid in the field. However, Hol- ocnemus males can live for over a year in the laboratory (unpubl. data), which suggests they may indeed have ample opportunity to remate in the field. ACKNOWLEDGMENTS This work was supported by BGSU Under- graduate Research Scholarships to JAB and MSS, and a Faculty Research Challenge award and NSF grant IBN 94-07357 to EMJ. We thank A. Porter for statistical advice, S. Vessey, A. Porter, A. Rypstra, and W. G. Eberhard for comments on the manuscript, and B. Randall for her tolerance of spiders in the Animal Facility. 170 THE JOURNAL OF ARACHNOLOGY LITERATURE CITED Austad, S. N. 1984. Evolution of sperm priority pat- terns in spiders. Pp. 223-249, In Sperm Competi- tion and the Evolution of Animal Mating Systems. (R. L. Smith, ed.). Academic Press, London. Buskirk, R. E. 1982. Sociality in the Arachnida. Pp. 281-367, In Social Insects, vol. II, (H. R. Hermann, ed). Academic Press, London. Cohen, J. 1977. Statistical Power Analyses for the Behavioral Sciences. Academic Press, New York. D’ Andrea, M. 1987. Social behaviour in spiders (Arachnida, Araneae), Italian J. ZooL, Monograph 3. Eberhard, W. G. & R. D. Briceno. 1983. Chivalry in pholcid spiders. Behav. Ecol. Sociobiol, 13:1 89— 195. Eberhard, W. G. & R. D. Briceno. 1985. Behavior and ecology of four species of Modisimus and Ble- chroscelis (Araneae, Pholcidae). Rev. Arachnol., 6:29-36. Eberhard, W. G., S. Guzman-Gomez & K. M. Catley. 1993. Correlation between spermathecal morphol- ogy and mating systems in spiders. Biol. J. Linn. Soc., 50:197-209. Jakob, E. M. 1991. Costs and benefits of group living for pholcid spiders: losing food, saving silk. Anim. Behav., 41:711-722. Jakob, E. M. 1994. Contests over prey by group- living pholcids. J. Arachnol., 22:39-45. Jakob, E. M. & H. Dingle. 1990. Food level and life history characteristics in a pholcid spider. Psyche, 97:95-110. Raster, J. 1995. Sperm priority pattern in the spider Holocnemus pluchei (Pholcidae). Master’s thesis, Bowling Green State Univ. Lewis, S. M. & S. N. Austad. 1990. Sources of intra- specific variation in sperm precedence in red flour beetles. American Nat., 135:351-359. Parker, G. A. 1984. Evolutionary stable strategies. Pp. 30-6 1 , In Behavioural Ecology, An Evolution- ary Approach. 2nd edition. (J. R. Krebs & N. B. Davies, eds.). Sinauer Associates, Sunderland, Mas- sachusetts. Robinson, M. H. 1982. Courtship and mating be- havior in spiders. Ann. Rev. Entomol., 27:1-20. Sokal, R. R. & F. J. Rohlf. 1981. Biometry, 2nd ed. W. H. Freeman and Co., New York. Suter, R. 1985. Intersexual competition for food in the bowl and doily spider, Frontinella pyramitela (Linyphiidae). J. Arachnol., 13:61-70. Suter, R. B. & L. Walberer. 1989. Enigmatic cohab- itation in bowl and doily spiders, Frontinella pyr- amitela (Araneae, Linyphiidae). Turnbull, A. L. 1 962. Quantitative studies of the food of Linyphia triangularis Clerck (Araneae: Linyphi- idae). Canadian Entomol., 94:1233-1249. Manuscript received 16 March 1995, revised 6 June 1995. 1995. The Journal of Arachnology 23:171-176 GENERIC PLACEMENT OF THE EMPIRE CAVE PSEUDOSCORPION, MICROCREAGRIS IMPERIALIS (NEOBISIIDAE), A POTENTIALLY ENDANGERED ARACHNID William B. Much more: Department of Biology, University of Rochester, Rochester, New York 14627 USA James C. Cokendolpher: Adjunct Professor, Department of Biology, Midwestern State University, Wichita Falls, Texas 76308 USA ABSTRACT. Types, topotypes, and some other material of the pseudoscorpion Microcreagris imperialis Much- more have been studied, and the species is transferred to the genus Fissilicreagris Curcic. Supplemental description and illustrations are presented, including the first information about females. This species is known only from three caves in Cave Gulch, Santa Cruz County, California. The cave habitat for F. imperialis is threatened by vandalism, development, and closure; and the U. S. Fish and Wildlife Service has proposed this pseudoscorpion as a candidate for listing as an endangered or threatened species. More than 1500 species and subspecies of an- imals in the United States are proposed by the U. S. Fish and Wildlife Service for listing as en- dangered or threatened (Drewry 1994). In most instances, these taxa have been submitted for consideration in the absence of any validation of taxonomic status. One of the candidate species of pseudoscorpions being reviewed for possible addition to the List of Endangered and Threat- ened Wildlife under the Endangered Species Act of 1973, as emended, is Microcreagris imperialis Muchmore, from Empire Cave, Santa Cruz County, California. As Mahnert (1979) and Curcic (1983) have demonstrated, the genus Microcreagris Balzan is restricted to two species in China and Afghani- stan, and the numerous American species which had been placed in that genus are improperly assigned. Curcic (1 978-1 989) created quite a few new genera and placed many of the American species in them; he was, however, unable to make definite generic assignments for some species, in- cluding M. imperialis (1984: 165). The purpose of this paper is to clarify the tax- onomic position of M. imperialis so that this potentially endangered species can be identified properly in the literature. Furthermore, we wish to provide sufficient identification characteristics so that field biologists might more easily recog- nize this species in its native habitat. METHODS Specimens have been borrowed from the American Museum of Natural History, New York, New York (AMNH); California Academy of Sciences, San Francisco, California (CAS); D. Ubick personal collection, San Francisco, Cali- fornia (CDU); and the Florida State Collection of Arthropods, Gainesville, Florida (FSCA). Unless otherwise stated, the specimens have been dissected, cleared, and mounted in Canada balsam on microscope slides. Fissilicreagris imperialis (Muchmore), new combination Empire Cave Pseudoscorpion Figs. 1-9 Microcreagris imperialis Muchmore 1969:13-15, 21, fig. 10; Arnett 1984:21666; Briggs & Ubick 1988:44; Drewry 1989:566; Coddington, Larcher & Coken- dolpher 1 990: 1 1 ; Drewry 1991:58833; Harvey 1991: 342; Drewry 1994;59025. ‘Microcreagris’ imperialis: Curcic 1984:164, 165, figs. 20, 41. Pseudoscorpion: Briggs 1990:180. Type locality.— Empire Cave, in Cave Gulch, one mile NW of Santa Cruz, Santa Cruz County, California. Material examined. — Holotype male and one paratype male from Empire Cave, 26 August 171 172 THE JOURNAL OF ARACHNOLOGY Figures 1-4.— Fissilicreagris imperiaiis (Muchmore), male holotype. 1, Tip of movable finger of chelicera, with galea; 2, Cheliceral flagellum; 3, Central part of stemite 3; 4, Stemites 5-9, showing chaetotaxy (setae omitted). 1 963, R. E. Graham, mounted on slides (AMNH); one paratype male, same data (FSCA); one top- otype female from Empire Cave, September 1972, R. Lem, mounted on slide (CAS); one topotype male from Empire Cave, 8 July 1989, D. Ubick, et al., in alcohol (CDU); two topotype males from Empire Cave, 8 September 1991, D. Ubick and S. Fend, mounted on slides (CAS); one female from Dolloff Cave, across Cave Gulch from Em- pire Cave, 22 April 1979, D. C. Rudolph et ah, mounted on slide (CAS); three females from IXL Cave, in Cave Gulch one-half mile S of Dolloff Cave, 21 April 1979, D. C. Rudolph et al., two mounted, one in alcohol (CAS). Supplementary description.— The topotypes and specimens from Dolloff* and IXL caves are generally similar to the holotype and paratypes. Dimensions, proportions, and chaetotaxies of body and appendages vary slightly from the val- ues given in the original description, but all ap- pear conspecific with the types. A few features of the additional specimens are, however, worth mentioning. Apex of palpal coxa (manducatory process) bears three setae in all specimens, as in types. Cheliceral galea short and twice bifid (Fig. 1); galeae of types also like this, not just “with four or five terminal spinules” as characterized in original description (Muchmore 1969:15). Chel- iceral flagellum composed of 7-8 serrate setae (Fig. 2). Sternites 6, 7, and 8 with two setae on face near middle (discal setae) as in types (one topotype with two setae on face of stemite 5); stemites 9 and 1 0 with two corresponding setae slightly anterior to the marginal row (Fig. 4). Tri- chobothria on palpal chela of holotype as shown in Fig. 5 (and Curcic 1984: fig. 41); there is a little variation in position of trichobothria on fixed fingers of other specimens (Fig. 6). Genital opercula (stemites 2 and 3) of male about as illustrated for holotype (Curcic 1984: fig. 20); stemite 3 of holotype with 22 setae scattered broadly (paratypes and topotypes with 1 8-22 se- tae); sternite 3 of holotype with five small setae near middle and 1 1 larger setae along posterior margin (paratypes and topotypes with 5-9 small setae near middle and 12-13 along margin); an- terior margin of stemite 3 slightly concave at middle (Fig. 3), but not as distinctly indented as in Fissilicreagris chamherlini (Beier) (see Curcic 1 984: fig. 3). Genital opercula of female as shown in Fig. 8, similar to those of F. chamherlini (see Curcic 1984: fig. 4); stemite 2 with 8-11 small setae in two groups, on either side of midline; sternite 3 with 11-14 setae along posterior mar- gin. In both sexes, on stemites 3 and 4, there are 4-7 small setae on each spiracular plate. Internal genitalia of holotype male shown in Fig. 7; gen- erally similar to those of Saetigerocreagris phyl- lisae (Chamberlin) (see Chamberlin 1962: fig. 12), Tartarocreagris texana (Muchmore) (see Much- more 1992: fig. 4), and Fissilicreagris macilenta (Simon) (see Muchmore 1994: fig. 4); the dorsal sacs are thin-walled and not as clearly separate MUCHMORE & COKENDOLPHER— EMPIRE CAVE PSEUDOSCORPION 173 Figures 5-8 .—Fissilicreagris imperialis (Muchmore). 5, Left palpal chela of holotype (male), lateral view, showing positions of trichobothria (darkened areoles are underneath); 6, Right palpal chela of paratype (male), lateral view; 7, Internal genitalia of holotype (male), ventral view (dorsal genital sacs stippled); 8, Central parts of stemites 2 and 3 of female (IXL Cave). as in the three species mentioned; lateral sacs long and narrow. Internal genitalia of female not distinguished. Measurements (mm).— Male: Figures given first for holotype, followed in parentheses by ranges for the two paratypes and two topotypes. Body L 3.50 (2.50=3.45). Carapace L 0.935 (0.89= 0.96). Chelicera L 0.52 (0.495=0.56). Palp: tro- chanter 0.58 (0. 56-0. 63)/0. 215 (0.205=0.23); fe- mur 1.12 (1.08-1. 19)/0.23 (0.215=0.24); patella 1.07 (1.00=1.1 1)/0.29 (0.28=0.305); chela (with- out pedicel) 1.73 (1.69-1.86)/0.435 (0.43-0.47); hand (without pedicel) 0.73 (0.70-0.8 15)/0. 385 (0.38=0.43); pedicel L 0. 1 6 (0. 1 6-0. 1 8); movable finger L 1.05 (1.03-1.15). Leg IV: femur + pa- tella 0.89 (0.81~0.92)/0.215 (0.205-0.23); tibia 0.835 (0.79-Q.89)/Q. 1 1 5 (0. 1 1-0. 125); basitarsus 0.29 (0.295-0.3 1)/0.095 (0.08-0.095); telotarsus 0.43 (0.40-0.445)/0.07 (0.07-0.075). Female: Figures given first for topotype, fol- lowed in parentheses by ranges for three speci- Figure 9. — Fissilicreagris imperialis (Muchmore), fe- male (IXL Cave). Dorsal view (setae are transparent and can only been seen with a microscope, not visible in field examinations). 174 THE JOURNAL OF ARACHNOLOGY mens from Dolloff and IXL caves. Body L 3.33 (3.52-3.62). Carapace L 0.89 (1.00-1.05). Che- licera L 0.525 (0.53-0.585). Palp: trochanter 0.55 (0.59-0.63)/0.215 (0.22-0.26); femur 1.03 (1.09- 1.21)/0.215 (0.24-0.26); patella 0.95 (1.03-1.14)/ 0.265 (0.29-0.325); chela (without pedicel) 1.67 ( 1.76—1 .9 1)/0.43 (0.48-0.55); hand (without pedicel) 0.70 (0.76-0. 87)/0.385 (0.43-0.50); pedicel L 0. 1 5 (0. 1 7-0.20); movable finger L 0.96 (1.04-1.16). Leg IV: femur + patella 0.79 (0.87- 0.96)/0. 18 (0.21-0.235); tibia 0.74 (0.85-0.925)/ 0.11 (0.12-0.125); basitarsus 0.27 (0.30-0.32)/ 0.08 (0.09-0. 105); telotarsus 0. 39 (0.4 1 5-0.445)/ 0.075 (0.08). Remarks.— Attempting to place M. imperialis in a genus, Curcic (1984:165) stated “In most of its diagnostic characters (shape of flagellum, presence of anterior discal setae on abdominal stemites, chaetotaxy of manducatory process, and trichobothriotaxy), it is closest to the genus Aus- tralinocreagris Curcic 1984].” This is generally correct, but it is also true that in these same characters M. imperialis is very similar to rep- resentatives of the genus Fissilicreagris, which are found in the same general area of California (see Curcic 1984:154-156; Muchmore 1994:63- 64). In addition, the internal genitalia of male M. imperialis are more like those of F. macilenta (see Muchmore 1994) and F. chamberlini than those of Australinocreagris grahami (Muchmore) (unpubl. obs.); in particular, the dorsal genital sac of the latter species is entire and round in outline, while that of the first two species and M. imperialis is bilobed or divided into two separate round sacs. Microcreagris imperialis is also sim- ilar to the two species of the genus Saetigero- creagris Curcic in respect to the male genitalia, but it differs from them in proportions of body and appendages and in the chaetotaxies of var- ious parts. Its relation to the genus Tartarocrea- gris, which also has similar male genitalia, but is presently known only from Texas, is uncertain. It is concluded that Microcreagris imperialis Muchmore is most similar to Fissilicreagris ma- cilenta and F. chamberlini and should be con- sidered congeneric with them. Its major differ- ence from them is its lack of eyes, a condition which is presumably an adaptation to life in caves. Field recognition.— Although preserved ma- terial is required for positive identification, per- sons conducting a census of cave faunas or other work associated with protection of this species can be fairly certain that they have Fissilicreagris imperialis if: (1) it is in a cave in Cave Gulch, and (2) it appears like Fig. 9, eyeless, and about 3.0-3. 5 mm in length. The only other pseudo- scorpion known from Cave Gulch is an unde- scribed, blind species of Neochthonius Cham- berlin (Chthoniidae) in Empire Cave (unpubl. obs.); this is easily distinguished from F. imper- ialis by its much smaller size (only one-third as long as the latter). There are other species un- doubtedly present at surface locations, but these should have eyes. DISCUSSION Fissilicreagris imperialis is known only from Empire, Dolloff, and IXL caves in Cave Gulch, Santa Cruz County, California. It may occur in one or more of the other caves in Cave Gulch, but it is certainly restricted to this small, isolated karst area. The three caves are all within one- half mile of each other. In addition to the new records listed under specimens examined, it is important to note that D. Ubick (pers. comm. 1995) observed but did not collect more than six specimens in Empire Cave on 3 July 1993. Microcreagris imperialis was first listed by the U.S. Fish and Wildlife Service as a candidate for review as an endangered or threatened species over a decade ago (Arnett 1984). It is still only a candidate species and therefore it receives no substantive or procedural protection under the Endangered Species Act. Drewry (1989, 1991, 1994) continued to list this species, without a change in its status of review. The history of Empire Cave is a tragic one. It has been known and vandalized for over 120 years (Halliday 1962). According to Graham (1967), during August 1962 the entrance to the cave was capped by a cement barrier through which a small portal (about one meter square) allowed access to the cave. This change in the entrance greatly decreased the available light in the entrance and presumably restricted air flow and increased humidity. By August 1963, a dra- matic shift was noted in the distribution of cave arthropods and gastropods (Graham 1967, 1968a). Presumably the restricted entrance also altered the energy input into the cave. Graham (1968a) also noted that in 1966 the cave was blackened and filled with a strong odor, possibly gasoline. He further stated that this is the most heavily vandalized cave in the state. Despite re- peated attempts to seal the cave, it has been dug open in each case (Graham 1968a). Adamson (1982) referred to the caves in Cave Gulch as small, badly vandalized, and often lit- MUCHMORE & COKENDOLPHER - EMPIRE CAVE PSEUDOSCORPION 175 tered and trashed caves. She went on to report on a clean-up trip to the caves (including Empire and DollofF caves) during March of 1 982. At that time Empire Cave was “really filled with trash including papers, wood, cigarette butts, orange peels, etc. and most dangerous, the shards of many a beer bottle. This poor cave is reputedly used for parties etc. by UCSC students.” As in so many efforts to clean caves, apparently no attention was given to the fauna and how this “trash” may be affecting them. While some of the collections of this pseudoscorpion have been taken from the undersides of rocks (D. Ubick pers. comm. 1995), most specimens thus far reported from Empire Cave have been taken on wood in the cave. How many pseudoscorpions or their prey items were accidentally removed with the wood during the 1 982 clean-up? It is hoped that future efforts will be better guided. Briggs & Ubick (1988) stated that Cave Gulch on the Gray Whale Ranch and nearby Empire Cave were in danger. At that time there were plans to log the area. Those authors felt that this activity could collapse caves and disturb root systems on which primary consumers feed, alter drainage, and block entrances. If the areas were logged it could also lead to further development. Briggs (1990) continued to state that the cave habitat for this species was threatened by devel- opment and closure. A Timber Harvest Plan to log the Gray Whale Ranch and an adjacent area on Empire Grade was approved in late 1994 by the California Department of Forestry (Anony- mous 1995). According to the plan, harvesting can start at anytime. The owner of the ranch has stated that the entire ranch will be logged within four years (Anonymous 1995). Little data on the biology of pseudoscorpions in Cave Gulch caves are available. The topotype female reported herein from Empire Cave was taken under wood at 20 °C on 25 September 1972. The type series was collected “about mid- dle of twilight zone. Each captured on the floor either on side or bottom of wood, or in one case traveling over dripstone on floor. Temp. 53.5 °F, air saturated, floor very damp.” (Muchmore 1969). There are a few other observations avail- able on the habitat, but nothing specific to the pseudoscorpion. These caves are in limestone (Graham 1966). The Cave Gulch is subject to intermittent flooding. During heavy flooding, al- most all of the 75 m of Empire Cave fills with water (Halliday 1962). Dolloff Cave’s entrance is almost at the level of an intermittent side stream and must be re-excavated after every major flood (Halliday 1962). The lower portion of Empire Cave has clay floors (Graham 1967). Graham (1967, 1968a, 1968b) provided maps to Empire Cave and recorded the conditions: 29 January 1960 (6. 9-7. 3 °C, 94-99% R. H.), 7 August 1962 (9.0-9. 3 and 12°C, 100%R.H.), 26 August 1963 (9.3-10.9 and 12.8 °C, 96-100% R. H.). In the same publications, he also recorded the Dolloff Cave conditions as: 28 August 1963 (8.6-10.2 and 12.6 °C, 90% R. H.), 16 October 1966 (16 °C, 85% R. H.). ACKNOWLEDGMENTS We are indebted to Darrell Ubick, California Academy of Sciences, San Francisco, California; Norman I. Platnick, American Museum of Nat- ural History, New York, New York; and G. B. Edwards, Florida State Collection of Arthro- pods, Gainesville, Florida for the loan of mate- rial. We sincerely thank Norman V. Horner (Midwestern State University, Wichita Falls, Texas) and Darrell Ubick (California Academy of Sciences, San Francisco) for reviewing the manuscript. LITERATURE CITED Adamson, M. 1982. Santa Cruz clean-up. Devil’s Advocate, 15:34-35. Anonymous. 1995. Gray Whale Ranch in holding pattern; the bad news. Updates from Save the Gray Whale Parklands, Sierra Club, 7:1. Arnett, G. R. 1 984. Endangered and threatened wild- life and plants; review of invertebrate wildlife for listing as endangered or threatened species. Dept. Interior Fish & Wildlife Serv. 50 CFR, part 1 7. Fed. Reg., 49:21664-21675. Briggs, T. S. 1990. Biology of northern California. Protecting California cave biology. P. 180 In NSS [Nat. Speleo. Soc.l 1990 Convention Guidebook (V. Johnson, ed.), Yreka, California. Briggs, T. S. & D. Ubick. 1988. Cavemicoles from Cave Gulch, Santa Cruz County. California Caver, 38:43-44. Chamberlin, J. C. 1962. New and little-known false scorpions, principally from caves, belonging to the families Chthoniidae and Neobisiidae (Arachnida, Chelonethida). Bull. American Mus. Nat. Hist., 123: 299-352. Coddington, J. A., S. F. Larcher, & J. C. Cokendolpher. 1990. The systematic status of Arachnida, exclu- sive of Acari, in North America north of Mexico. Pp. 5-20 In Systematics of the North American insects and arachnids: Status and needs. (M. Kosz- tarab & C. W. Schaefer, eds.), Virginia Agricul. Exp. Stat. Info. Ser. 90-1, Virginia Polytech. Inst. & State Univ., Blacksburg, Virginia. 176 THE JOURNAL OF ARACHNOLOGY Curcic, B. P. M. 1978. Tuberocreagris, a new genus of pseudoscorpions from the United States (Arach- nida, Pseudoscorpiones, Neobisiidae). Fragm. Bal- canica, 10:1 1 1-121. Curcic, B. P. M. 1983. A revision of some Asian species of Microcreagris Balzan, 1892 (Neobisiidae, Pseudoscorpiones). Bull. British Arachnol. Soc., 6:23-36. Curcic, B. P. M. 1984. A revision of some North American species of Microcreagris Balzan, 1892 (Arachnida: Pseudoscorpiones: Neobisiidae). Bull. British Arachnol. Soc., 6:149-166. Curcic, B. P. M. 1 989. Further revision of some North American false scorpions originally assigned to Mi- crocreagris Balzan (Pseudoscorpiones, Neobisi- idae). J. Arachnol., 17:351-362. Dre wry , G. 1989. Endangered and threatened wildlife and plants; animal notice of review. Dept. Interior Fish & Wildlife Serv. 50 CFR, part 17. Fed. Reg., 54:554-579. Dre wry , G. 1991. Endangered and threatened wildlife and plants; animal candidate review for listing as endangered or threatened species. Dept. Interior Fish & Wildlife Serv. 50 CFR, part 17. Fed. Reg., 56: 58804-58836. Dre wry. G. 1994. Endangered and threatened wildlife and plants; animal candidate review for listing as endangered or threatened species. Dept. Interior Fish & Wildlife Serv. 50 CFR, part 17. Fed. Reg., 59: 58982-59028. Graham, R. E. 1966. Observations on the roosting habits of the big-eared bat, Plecotus townsendii, in California limestone caves. Cave Notes, 8:17-22. Graham, R. E. 1 967. The subterranean niche of Pseu- dometa biologica (Arachnida; Araneidae) in the Santa Cruz caves, California, with comments on ecological equivalence in the cave environment. Caves Karst, 9:17-22. Graham, R. E. 1968a. Spatial biometrics of subter- ranean demes of Triphosa haesitata (Lepidoptera: Geometridae). Caves Karst, 10:21-28. Graham, R. E. 1968b. The twilight moth, Triphosa haesitata, (Lepidoptera: Geometridae) from Cali- fornia and Nevada caves. Caves Karst, 10:41-48. Halliday, W. R. 1962. Caves of California. Privately printed, Seattle, Washington. 1 94 pp. Harvey, M. S. 1991. Catalogue of the Pseudoscor- pionida. Manchester Univ. Press, Manchester, En- gland, 726 pp. Mahnert, V. 1979. The identity of Microcreagris gi- gas Balzan (Pseudoscorpiones, Neobisiidae). Bull. British Arachnol. Soc., 4:339-341. Muchmore, W. B. 1969. New species and records of cavemicolous pseudoscorpions of the genus Micro- creagris (Arachnida, Chelonethida, Neobisiidae, Ideobisiinae). American Mus. Novit., 2392:1-21. Muchmore, W. B. 1992. Cavemicolous pseudoscor- pions from Texas and New Mexico (Arachnida: Pseudoscorpionida). Texas Mem. Mus., Speleol. Monogr., 3:127-153. Muchmore, W. B. 1994. On four species of pseu- doscorpions from California described by E. Simon in 1878 (Pseudoscorpionida: Neobisiidae, Cheme- tidae, Cheliferidae). J. Arachnol., 22:60-69. Manuscript received 15 July 1995, revised 12 August 1995. 1995. The Journal of Arachnology 23:177-193 PHYLOGENY AND HISTORICAL BIOGEOGRAPHY OF THE SPIDER GENUS LUTICA (ARANEAE, ZODARIIDAE) Martin G. Ramirez: Department of Biology, Bucko ell University, Lewisburg, Pennsylvania 17837 USA Richard D. Beckwltt: Department of Biology, Framingham State College, Framingham, Massachusetts 01701 USA ABSTRACT. Spiders of the genus Lutica from 19 populations in southern California and Baja California, including all the California Channel Islands except Anacapa, were compared electrophoretically on the basis of variability at 1 5 gene loci. Fixed allelic differences clearly define two species: new species A [Santa Barbara and Ventura Counties, northern Channel Islands (San Miguel, Santa Rosa, Santa Cruz), southern Channel Islands (San Nicolas, Santa Barbara, Santa Catalina)] and new' species C [Guerrero Negro, central Baja California], while morphological features define two others: new species B [Los Angeles, Orange and San Diego Counties, northern Baja California] and clementea [San Clemente Island]. Phylogenetic analysis of the electrophoretic data using a variety of methods revealed that evolutionary rates among the populations sampled have been very unequal. The phylogenetic relationships among populations consistently supported by the electrophoretic cladograms generally correspond with the geological history of the Channel Islands and adjacent mainland and suggest certain likely colonization events involving some of the islands. Genetic similarities and differences among populations can be used to assess specific hy- potheses about biogeography and evolution. Populations on islands are especially useful be- cause they frequently are discrete entities with little gene flow among islands; for some islands the geologic history also is known. This type of analysis is especially powerful for sedentary spe- cies in which chances for dispersal among pop- ulations are minimal (Carlquist 1981). The California Channel Islands are an excel- lent system for addressing questions of evolu- tionary and biogeographic history. These eight islands vary in size, topography and physical iso- lation (Fig. 1). While the geologic history of the islands and their surroundings is complex, it has clearly involved the northward transport of these island landmasses on crustal blocks (terranes) caught in the tectonics of the Pacific/North American plate margin (Homafius et al. 1986). It has also involved extensive changes in sea level which have repeatedly submerged some islands while possibly leaving the highest areas of others continuously above water since the Oligocene (Vedder & Howell 1980; Haq et al 1987). Bio- geographic studies of biologically old taxa on the Channel Islands need to consider both vicariance and dispersal as factors in producing contem- porary distributional patterns. The spider family Zodariidae and many of its genera have existed since at least the Oligocene (Petrunkevitch 1942, 1952) and they have been exposed to the geological and climatic changes of the last 30 million years. Worldwide, 47 gen- era have been described, mainly from the trop- ical and temperate regions of the Old World (Joe- que 1991). Spiders of the genus Lutica are the only native representatives of the Zodariidae in the continental United States and they live in restricted insular and coastal dune communities in southern California and Baja California, in- cluding all the California Channel Islands except Anacapa (Ramirez 1995). Their preferred habi- tat is a sand dune covered by native beach veg- etation located well behind the high tide line and the influence of sea water (Gertsch 1961; Ra- mirez 1995). On Santa Barbara Island, typical coastal dunes do not exist and these spiders live in the sandy soil and debris below vegetation growing on a sea cliff (Ramirez 1995). They live in silk-lined burrows they construct in the sand, emerging only at night to feed and, during Au- gust-October, to mate (Gertsch 1961; Ramirez 1 995). When dislodged from their burrows, these 177 178 THE JOURNAL OF ARACHNOLOGY Figure 1 . —Map of southern California and Baja Cal- ifornia, including Channel Islands, showing Lutica sample sites. Population abbreviations follow Table 1. spiders actively burrow into the loose sand and are quickly lost from sight. Lutica does not use ballooning (aerial transport on wind blown silk threads) as a means of dispersal at any point in its life cycle; ballooning is rare in other fossorial spiders (Decae 1987) and has never been re- corded in the family Zodariidae (Jocque 1993). N on - reproducti ve terrestrial dispersal may be minimal (Ramirez 1995). Males wandering in search of females can be found in great numbers in September; how far they actually range is not known. Lutica is thought to have a lifespan of two to three years (Gertsch 1961; W. Icenogle pcis. comm.). In this study, we present the results of a survey of allozyme variation among Lutica populations from the Channel Islands and mainland of south- ern California and Baja California. The primary objective was to use allozyme variation to iden- tify valid species of Lutica and to determine their phylogenetic relationships. A second objective was to use the phylogenetic relationships indi- cated by the electrophoretic data to discuss the historical biogeography of this genus, with par- ticular attention to estimating probable patterns of colonization among island and mainland pop- ulations. METHODS Systematic background.— George Marx first described the genus Lutica from Klamath Lake, Oregon (Marx 1891). Gertsch (1961) corrected the type locality of Lutica maculata to Santa Rosa Island, California, and also described three new species: nicolasia (San Nicolas Island), clementea (San Clemente Island) and abalonea (Oxnard, Ventura County). Additional species have been described from India (Tikader 1981), but these taxa are dearly misplaced (Jocque 1991). Collections.— During 1985 and 1987, we col- lected Lutica from 1 9 sites in southern California and Baja California, covering a range of 865 km (Fig. 1). Sample sizes ranged from 20-50 spiders per population for a total of 812 spiders (Table 1). In the laboratory, they were starved for at least a week and then frozen at - 70 °C until they were prepared for electrophoresis. Electrophoresis. —A survey of 60 enzymes on 2-7 buffer systems revealed consistently scorable activity for 1 5 loci on three buffer systems; elec- trophoretic techniques and staining protocols are described in Ramirez (1990). No significant dif- ferences in the banding patterns of spiders of different ages or sex were ever detected, making it possible to examine spiders of all instars. All genotypes were inferred from the appearance of the staining patterns and the known subunit structure of the enzymes (Harris & Hopkinson 1976; Richardson et al. 1986). Species identification.— In this study, the de- tection of fixed allelic differences was the crite- rion for species identification, in accord with the biological species concept [Le., a fixed difference reflects the separate gene pools of two non in- terbreeding taxa (Mayr 1970)] and following the recommendations of Farris (1981) and Richard- son et al. (1986). Since it has been shown that a sample of three individuals each from two dif- ferent populations is sufficient to reveal a fixed allelic difference between the populations (Rich- ardson et al 1986), the mean sample sizes per locus in this study, which ranged from 33-46 for 1 6 populations and 1 9-22 for three populations, were certainly adequate for the detection of fixed differences and the identification of species. In cases where diagnostic loci could not be found for a group of populations, reference was made to the morphological taxonomy of Gertsch (1961, pers. comm.) for evidence that might suggest val- RAMIREZ & BECKWITT — PH YLOGENY AND BIOGEOGRAPHY OF LUTICA 179 Table 1.— Summary of collections of Lutica. Samples include spiders of all instars. Locality (abbreviation) Sample size Dates of sampling Coal Oil Point Reserve (COP) (Santa Barbara County) 48 May 12, 1985 McGrath State Beach (MG) (Ventura County) 48 June 1 1 & August 15, 1985 Oxnard Beach (OX) (Ventura County) 36 June 11, 1985 La Jolla Beach (LJB) (Ventura County) 48 May 27, 1985 San Miguel Island (SMI) Cuyler Harbor 48 August 13, 1985 Santa Rosa Island (SRI) Southeast Anchorage 48 July 1, 1987 Santa Cruz Island (SCI) Johnstons Lee 48 August 17, 1985 Santa Barbara Island (SBI) Cliffs south of Signal Peak San Nicolas Island 48 July 9-10, 1987 Army Camp Beach (SNA) 24 July 31, 1985 Dutch Harbor (SND) 22 July 30, 1985 Red Eye Beach (SNE) 20 July 31, 1985 Santa Catalina Island (CAT) Little Harbor 48 August 23, 1985 San Clemente Island (SCL) Flasher Road Dunes 48 August 21, 1985 Ballona Wetlands (BA) (Los Angeles County) 48 June 9, 1985 El Segundo Dunes, LAX (ESG) (Los Angeles County) 36 April 15, 1985 Balboa Beach (NB) (Orange County) 48 April 14, 1985 Silverstrand State Beach (SVS) (San Diego County) 48 July 13, 1985 Punta Estero (PE) (Baja California Norte, Mexico) 48 October 15, 1985 Guerrero Negro (GN) (Baja California Sur, Mexico) 50 October 18, 1985 id groupings. Gertsch has recently reviewed mor- phological variation in this genus and all refer- ences to morphological differences are based on personal communication with him. Phylogenetic analysis.— The problem of esti- mating phylogenetic trees from electrophoretic data has generated a wealth of divergent opinion, some of it couched in very strong language (re- viewed by Felsenstein 1982; Buth 1984). While many methods for phylogenetic tree construction from electrophoretic data have been proposed (Felsenstein 1982; Swofford & Olsen 1990), none has been universally accepted (Quicke 1993; Av- iso 1994). Because of this lack of agreement, we used a variety of methods to analyze the elec- trophoretic data set for Lutica , using allele fre- quencies, alleles as discrete characters and ge- netic distances. These methods are based on the two main approaches that do not assume a con- stant rate of molecular evolution across all taxa being compared, maximum parsimony (Edwards & Cavalli-Sforza 1963) and maximum likeli- hood (Edwards & Cavalli-Sforza 1964; Felsen- stein 1981). Comparison of the trees generated by the various methods indicates those portions of the phylogeny that are unaffected by the dif- ferent assumptions of each method (i.e., different methods may yield similar branching patterns for some or all taxa) and which therefore may be assumed to represent more accurately actual evolutionary relationships (Lanyon 1985; Avise 1 994). Computer programs to carry out each of these methods are readily available. The partic- ular programs/packages and specific computa- tional procedures which were used are as follows: Maximum parsimony: FREQPARS (version 1.0) (Swofford 1988) was used to conduct fre- quency parsimony analysis (Swofford & Ber- locher 1987) of alleles at all loci, except those which were monomorphic across all populations (n) or n - 1 populations and therefore were phy- logenetically uninformative. FREQPARS 1 .0 has a very limited ability to search for the most par- simonious tree(s), so 1 9 runs of the data set were performed, with each OTU (operational taxo- nomic unit) in turn being placed first in the input file (following Rohlf & Wooten 1 988). The short- est (i.e., most parsimonious) tree generated was retained. HENNIG86 (version 1.5) (Farris 1988, 1989) was used for Wagner parsimony (Kluge & Farris 1969; Farris 1970) analysis of alleles as discrete characters, using presence/absence (1/0) coding for both complete (all alleles at frequencies > 0) and reduced (all alleles at frequencies > 0.05) data sets (following Mickevich & Mitter 1981; C. Griswold pers. comm.), for all informative loci. The implicit enumeration (IE*) option of HENNIG86 was used to generate all most par- simonious trees for the complete and reduced 180 THE JOURNAL OF ARACHNOLOGY data sets, and then a strict consensus tree (Nelson 1979; Rohlf 1982) was computed for each set of most parsimonious trees. BIOSYS-1 (version 1.7) (Swofford & Selander 1981, 1989) was used to perform distance Wag- ner analysis (Farris 1972; Swofford 1981) on a matrix of Rogers (1972) genetic distances for the Lutica populations; since Nei (1972, 1978) dis- tances are non-metrical, which can result in neg- ative branch lengths (Farris 1 972; Nei 1 987), they are not appropriate for use with distance Wagner analysis (Swofford 1981). Specifically, the DIS- WAG step call was invoked, with the multiple addition criterion (Swofford 1981) (maxtree = 30), Prager & Wilson’s (1976) F goodness of fit criterion and outgroup rooting (Farris 1 972) [with Guerrero Negro (GN) as outgroup] options. The shortest (i.e., most parsimonious) tree generated was retained. Maximum likelihood: PHYLIP (versions 3.1 and 3.2) (Felsenstein 1988, 1989a, b) was used to generate maximum likelihood trees from the allele frequency data for the 19 Lutica popula- tions using the CONTML program, with the G (global branch swapping), J (jumble addition, i.e., each OTU is added to the developing tree in random order) and O [outgroup rooting, with Guerrero Negro (GN) as outgroup] options in- voked. Since CONTML does not perform ex- haustive enumeration and evaluation of all pos- sible tree topologies, the data set was run 19 times, with different random number seeds for thej option, following Felsenstein’s ( 1981, 1989b) recommendation. Of the trees generated, the tree with the greatest likelihood was kept and the others were discarded. Congruence. —Congruence among the phytog- enies generated by the different methods was de- termined by the construction of consensus trees (reviewed by Rohlf 1982; Mickevich & Platnick 1989). In the present study, consensus trees were constructed for multiple trees generated by a sin- gle method (i.e., Wagner parsimony analysis of alleles as discrete characters with HENNIG86), as well as among the cladograms representing the best or consensus tree for each method. Because methods which generate multiple, equally likely trees might lead to misleading results with Ad- ams (1972) consensus, strict consensus (Nelson 1979; Rohlf 1982) was used in such cases. On the other hand, in an effort to maximize taxo- nomic information, Adams consensus was used to determine congruence among the final (best or strict) trees produced by each method. To quantitatively assess congruence among the phy- logenies generated by the different methods, two consensus measures were calculated (following Rohlf 1982 and Rohlf et al. 1983): the normal- ized consensus fork (CF) index of Colless (1980) and the index of Rohlf (1982). Both congru- ence measures can range from 0.0 (totally dis- similar topologies) to 1 .0 (identical topologies). The CONTREE program included with the PAUP (version 2.4.1) (Swofford 1985) computer package was used to generate consensus trees and calculate the consensus indices. RESULTS There were 43 alleles identified at the 15 ge- netic loci; two loci (APK-2 and G-3-PDH) were monomorphic across all populations and one lo- cus (TPI-2) was autapomorphic (variable in only a single population) (Table 2). Tables of inter- population genetic distances, mean genetic dis- tances within and between Lutica species, and the data matrix of alleles coded as character states, which were used as the basis for some of the analyses reported herein, are available on request from the senior author. Species identification. —The most striking fea- ture of the allelic data (Table 2) is the genetic distinctness of the Guerrero Negro (GN) popu- lation: there are fixed differences at the FUM, FIK and LDH loci; nearly fixed differences at the GPI (frequency of GPI-A - 0.990), IDH (fre- quency of IDH-C = 0.990) and PGM (frequency of PGM-A = 0.940) loci; and unique alleles at the AAT (AAT-A) and IDH (IDH-D) loci. In addition, the IDH-B allele, appearing at a fre- quency of 0.021 in only one other population [San Cruz Island (SCI)], is found at a frequency of 0.796 at Guerrero Negro. In light of the three fixed and three nearly fixed differences, the pop- ulation of Guerrero Negro clearly represents a distinct species, our new species C. Spiders from this region are also a distinct group morpholog- ically. Due to its genetic distinctness and the lack of certainty about which zodariid taxon would serve as an suitable outgroup for Lutica , Guer- rero Negro was used as the outgroup in the phy- logenetic analyses reported here. In analyzing the data for the remaining pop- ulations for valid phylogenetic groups, the fixed difference at the NP locus is clearly indicative of common ancestry (and is not contradicted by data at other loci): populations 1-12 are fixed for NP-A and comprise our new species A. These are the mainland populations of Santa Barbara RAMIREZ & BECKWITT-PHYLOGENY AND BIOGEOGRAPHY OF LUTICA 181 and Ventura Counties, as well as the populations of the northern Channel Islands (San Miguel, Santa Rosa, Santa Cruz) and three of the south- ern Channel Islands (San Nicolas, Santa Barbara, Santa Catalina). With the exception of Santa Bar- bara and Santa Catalina Islands, this is also a valid group morphologically. As for the remaining populations (13-18), the electrophoretic data provide little basis for spe- cies decisions. The population of San Clemente Island (SCL), representing clementea, was not characterized by any fixed differences (Table 2) but did possess two unique alleles, though one was very rare: APK- 1 -A was found at a frequency of 0.132 and TPI-l-A was found at a frequency of 0.03 1 . Among the mainland populations of southern California and northern Baja Califor- nia, there were likewise no fixed differences that would conclusively indicate species status for any of these populations, though one locus indicated a close relationship between the populations of the Ballona Wetlands (BA) and El Segundo Dunes (ESG): at the PGM locus, the PGM-B allele was found at frequencies of 0.696 and 0.750 respec- tively and is found in only two other populations at frequencies of less than 0.05. Since the elec- trophoretic data are neutral with regard to the status of clementea, it will be accepted as a valid species. Likewise, while there are no allelic dif- ferences that would unite the mainland popu- lations of southern California and northern Baja California as a group, morphological features de- fine these populations as a distinct group (see also Thompson 1973). As such, they will be ac- cepted as a valid species, our new species B. Phylogenetic analysis.— We produced five es- timates of the phylogeny of Lutica using methods which make no assumptions about evolutionary rates among taxa (i.e., frequency parsimony, dis- tance Wagner, Wagner parsimony analysis of al- leles as discrete characters and maximum like- lihood). Trees generated by these methods have branch lengths which are proportional to the amount of evolutionary change which has oc- curred along each branch (Nei 1987; Swofford & Berlocher 1987). The trees generated by these methods for Lutica had branch lengths which were very unequal among the populations being compared. The distance Wagner tree (Fig. 2) is typical of the branch length variability which was present in all the trees; some populations (i.e., COP, MG, OX, LJB) have undergone consid- erable differentiation, while others (i.e., PE, NB) have changed much less. The unevenness of the branch lengths indicate that allelic evolution in Lutica has certainly not been clocklike. In order to simplify comparisons among the phytogenies produced by the four rate indepen- dent methods, they are presented as cladograms in which only the branching patterns are shown (following Richardson et al. 1986) (Figures 3-7). These cladograms are consistent in the definition of two monophyletic groups: A) the large group A (= new species A) appears in all the cladograms [in that based on Wagner parsimony analysis of alleles as discrete characters using the complete data set (alleles > 0.0) (Fig. 5), the population of San Clemente Island (SCL) is also included in this group]; B) the Los Angeles County popula- tions of the Ballona Wetlands (BA) and the El Segundo Dunes (ESG) form a clade that appears in all the cladograms. Within group A, two clades are found in all the cladograms: one consisting of the population of Coal Oil Point Reserve (COP), Santa Barbara County, and the Ventura County populations of McGrath State Beach (MG), Oxnard Beach (OX) and La Jolla Beach (LJB) [reflecting their common possession of the PEP-C allele at frequencies ranging to fixation (Table 2)]; and the other comprised of the pop- ulations of San Nicolas Island [Red Eye Beach (SNE), Dutch Harbor (SND), Army Camp Beach (SNA)] and at least one of the northern Channel Islands, usually Santa Rosa (SRI) [reflecting their common possession, except for Santa Cruz Is- land (SCI), of the LDH-C allele in frequencies ranging to fixation (Table 2)]. These relationships are accurately represented in the Adams consen- sus tree for these cladograms (Fig. 8). The populations of new species B (including the BA - ESG clade) and clementea are placed in various positions among the five cladograms, with no consistent pattern of relationship, not surprising given the inconclusiveness of the elec- trophoretic data for these populations. San Cle- mente Island (< clementea ) is placed as the sister group to new species A in two cladograms (Figs. 3, 7); as sister group to new species A along with the populations of Balboa Beach (NB) and Punta Estero (PE) in one (Fig. 4); and as sister group to new species A along the with populations of the Ballona Wetlands (BA) - El Segundo Dunes (ESG) clade, Balboa Beach (NB) and Punta Es- tero (PE) in another (Fig. 6). As mentioned ear- lier, the remaining cladogram (Fig. 5) places San Clemente as part of new species A. The Adams consensus tree (Fig. 8) reconciles these differ- ences by placing SCL, NB and PE as sister group 182 THE JOURNAL OF ARACHNOLOGY U > s d d « g < | I S 04 © M B ■S3 H a PM « to 2 ‘o a m £ £ X VO o ^ pq io sa '- f 5 * O m E3 m •< V, pg Tf §2 m 2 5 m ^ « ^r .4 xr X rn o ^ o «n s * o 3 d < O VO t — < Ov QV o r-. — < 0© 04 *?r in os o o o *-i QV 04 o- o ov o\ ^ t- r-4 ov o OV -h vo m Oh O d d — < Os r- 04 ov o d d 04 oo m vo *=< o© d d i-i Os O' 04 Os O d d o o o o o o o Ov —1 o o q os o q 4-4 d d o o — i OV o o o o 04 O' o o o q O Ov q q — 1 d d d d o o ro O' o w o o O© i—i o VO q q O Ov q 00 — 1 d d d o o o o o o o o o o q q q q q ,-4 — < MO o o o o — < o o o o o\ o q q q q d d d d d d o o o o o o o o — < Qv o o q q O ov q q 1 1 d d o o o o o o o o Os — < o o q q Ov O q q — 1 — 1 d d o o Os O o »n o o 04 VO — i o O' q q O Os O q oo d d odd d d o o o o o o o o o o q q q q q — • — * _< — • o o o o o o o o o o q q q q q — < o o 04 O© o Ov o o ■sf in o O' q q O ov q q a d d d d d VO o o o o o OV o — < Ov o o vo q O ov q q d ’-4 d d d o o o o o o o o o o q q q q q — < »-* i-1 ^ m o OV o o 04 VO o O' 04 o o Os O q OV O q q d d d d 1 o o Q© 04 o o o o «n -vt o o q q Ov O q q d d d d d d o o ov ^ OV o d d < eq U O £ < eq O 0.05), due to the ex- clusion of practically all alleles under such a re- striction. The cladograms produced by each method are shown in Fig. 9, A-D and the Adams consensus tree is shown in Fig. 9E. For three OTUs, there are three possible relationships [A(BQ, C(AB), B(AQ] and all three are seen for the ingroup taxa among the cladograms in Fig. 9, A-D, although those produced by frequency parsimony (A) and maximum likelihood (D) are identical and place new species B and dementea as sister groups, as might be expected given their minimal genetic distance [Nei (1978) unbiased genetic distance: 0.017 (Ramirez 1990)]. Since the topologies of these cladograms covered all the possibilities for a three taxon statement, the Adams consensus tree presents the relationships among the three ingroup species as a unresolved trichotomy, although the consensus indices for this tree were fairly good [Colless’ (1980) CF = 0.500 and Rohlfs (1982) Cl, = 0.667], reflecting the perfect agreement between two cladograms. Thus, while two of the cladograms agreed in the placement of new species B and dementea as sister groups, these results were contradicted by the topologies of the other two cladograms, so RAMIREZ & BECKWITT — PHYLOGENY AND BIOGEOGRAPHY OF LUTICA 185 Figure 3.— Cladogram for Lutica based on the short- est tree generated using the frequency parsimony meth- od (Swafford & Berlocher 1987) and outgroup rooting (Farris 1972), with Guerrero Negro (GN) as outgroup. The frequency parsimony method minimizes tree length in the Manhattan metric (Sneath & Sokal 1973); total length of shortest tree = 29.509. Alphabetic designa- tion “A” denotes new species A. the electrophoretic data were not able to conclu- sively determine phylogenetic relationships among the ingroup species. DISCUSSION Electrophoresis and morphology. —The genus Lutica occupies a long geographic range (ap- proximately 1857 km) yet is relatively invariant morphologically. Gertsch (pers. comm.) uses fea- tures of the male palpi to discriminate species and only in clementea and the populations of central and southern Baja California are the dif- ferences in these structures clearly distinct. An analysis of variation among Lutica specimens involving 23-29 morphological characters (Thompson 1973) did not find statistically sig- nificant differences (M. Thompson pers. comm.). The genus Lutica is also relatively invariant genetically: Ramirez (1990) found low levels of genetic variability among Lutica populations, as well as a general trend toward within population homozygosity. As a presumably old genus (Ra- mirez 1990), the existence of low genetic vari- ability was unexpected and an analysis of the genetic structure of each species suggests that in- breeding, a spatial Wahlund effect due to local Figure 4. —Cladogram for Lutica based on the short- est tree generated using the distance Wagner method (Farris 1972; Swofford 1981), which is shown in Fig. 2. Alphabetic designation “A” denotes new species A. probabilities of random mating and environ- mental homogeneity associated with a subter- ranean existence in coastal dune ecosystems may be the most likely causes of low variability in Lutica (Ramirez 1990). The electrophoretic data define an outgroup [Guerrero Negro (GN), new species C] and two nested ingroups: first, populations 1™18, and nested within that, populations 1-12 (new spe- cies A). Morphological data indicate the specific distinctness of the population of San Clemente Island (clementea) (Gertsch 1961, pers. comm.), whereas the electrophoretic data were neutral. The electrophoretic data were likewise inconclu- sive with regard to the status of the mainland populations of southern California and northern Baja California but since they are morphologi- cally a valid group, they are assigned to new spe- cies B. Future electrophoretic studies involving more loci (only 1 2 of the 1 5 loci were phyloge- netically informative) may eventually result in the discovery of diagnostic loci for these main- land populations (new species B), as well as for clementea . The morphological systematics of the genus Lutica has been in a state of flux for many years (M. Thompson pers. comm.; W. Gertsch pers. comm.) and electrophoretic variation (particu- larly fixed allelic differences) is probably a more 186 THE JOURNAL OF ARACHNOLOGY COP MG OX UB CAT SNE SND SNA SRI SMI SCI SBI SCL BA ESG NB PE SVS GN Figure 5. — Cladogram for Lutica based on strict con- sensus tree (Nelson 1979; Rohlf 1982) of 10 trees of 43 steps each with consistency indices of 0.442 gen- erated using Wagner parsimony (Kluge & Farris 1969; Farris 1970), with alleles treated as characters with frequency greater than 0 = present. Consistency index = 0.410. Figure 7.— Cladogram for Lutica based on the tree of highest likelihood generated by the restricted max- imum likelihood method (Felsenstein 1981) and out- group rooting (Farris 1972), with Guerrero Negro (GN) as outgroup. Ln Likelihood = 879.074. Alphabetic des- ignation “A” denotes new species A. COP MG OX LJB SRI SNE SNA SND SMI SCI SBI CAT BA ESG SCL NB PE SVS GN Figure 6. —Cladogram for Lutica based on strict con- sensus tree (Nelson 1979; Rohlf 1982) of eight trees of 21 steps each with consistency indices of 0.610 gen- erated using Wagner parsimony (Kluge & Farris 1969; Farris 1970), with alleles treated as characters with frequency greater than or equal to 0.05 = present. Con- sistency index = 0.590. Alphabetic designation “A” denotes new species A. reliable indicator of taxonomic relationships than morphological features for this genus. An obvi- ous disagreement between our species assign- ments and those of Gertsch (1961) concerns the status of the populations of San Nicolas and Santa Rosa Islands and Oxnard, Ventura County: each is considered a distinct species ( nicolasia , ma- culata and abalonea, respectively), while we place them all in new species A. Due to the fixed dif- ference at the NP locus, the assignment of these populations to new species A is unambiguous on genetic grounds. Gertsch (pers. comm.) began a revision of Lutica prior to his deteriorating health and so a detailed comparison of the population groupings indicated by morphological and elec- trophoretic characters will have to await its com- pletion and publication. Phylogeny and speciation in Lutica . — The fact that the genetic distance between new species B and clementea (0.017) is several orders of mag- nitude less than the other inter-specific estimates [0.138-0.796, all Nei (1978) unbiased distances (Ramirez 1990)] would suggest that these taxa were originally a single species that only recently diverged. If this is the case, one would predict that these species should be placed as sister groups in any phylogeny, as a clade that is the sister group to new species A. However, while this was RAMIREZ & BECKWITT-PHYLOGENY AND BIOGEOGRAPHY OF LUTICA 187 "A" COP MG OX LJB SMI SRI SNE SNA SND SCI SBI CAT SCL NB PE BA ESG SVS GN Figure 8. —Adams (1972) consensus tree based on cladograms of Figs. 3-7. Consensus indices for this tree are: Colless’ (1980) CF = 0.471 and Rohlfs (1982) Cl, = 0.400. Alphabetic designation “A” denotes new spe- cies A. the case in two of the species cladograms (Fig. 9A, D), these relationships were contradicted by the topologies of the other two cladograms (Fig. 9B, C). It should be noted that the two cladograms which depict new species B and clementea as sister groups (Fig. 9A, D) are the products of phylogenetic methods (frequency parsimony and maximum likelihood) which use allele frequency data directly. Methods which use allele frequen- cies may be superior because they avoid the loss of phylogenetic information and the procedural/ theoretical complexities associated with the re- duction of such data to distances or characters (Berlocher 1984; Swofford & Berlocher 1987). On the other hand, allele frequencies are subject to the effects of random drift and/or selection and can vary over time, and so may not provide reliable information for analysis (Crother 1990). Given the continuing controversy about allele frequencies and the potential superiority of phy- logenetic methods which make direct use of them (e.g., Shaffer et al. 1991; Jones et al. 1 993), a firm conclusion concerning the relationship of new species B and clementea within Lutica will have to await a future phylogenetic analysis. Biogeography of Lutica in southern California and Baja California.— The phylogenetic rela- tionships among the populations consistently A) FREQUENCY PARSIMONY I New Species B — ' clementea New Species A New Species C B) DISTANCE WAGNER New Species A — clementea New Species B New Species C C) WAGNER PARSIMONY OF ALLELES >0.0 I New Species A ' New Species B clementea New Species C D) MAXIMUM LIKELIHOOD I New Species B I clementea New Species A New Species C E) ADAMS CONSENSUS New Species A New Species B clementea New Species C Figure 9. —Cladograms for Lutica species. (A) Clado- gram based on the shortest tree generated using the frequency parsimony method (Swofford & Berlocher 1987) and outgroup rooting (Farris 1972), with new species C (Guerrero Negro) as outgroup. The frequency parsimony method minimizes tree length in the Man- hattan metric (Sneath & Sokal 1973); total length of shortest tree = 18.160. (B) Cladogram based on the shortest tree generated using the distance Wagner method (Farris 1972; Swofford 1981), with multiple addition criterion (Swofford 1981) and outgroup root- ing (Farris 1 972), with new species C (Guerrero Negro) as outgroup. The distance measure used was Rogers (1972). Total length of shortest tree = 0.616 and Prager & Wilson’s (1976) F= 1.857. (C) Cladogram based on the shortest tree generated using Wagner parsimony (Kluge & Farris 1969; Farris 1970), with alleles treated as characters with frequency greater than 0 = present. Total length of shortest tree = 1 6 steps and consistency index = 0.680. (D) Cladogram based on the tree of highest likelihood generated by the restricted maxi- mum likelihood method (Felsenstein 1981) and out- group rooting (Farris 1972), with new species C (Guer- rero Negro) as outgroup. Ln Likelihood = 53.788. (E) Adams (1 972) consensus tree based on cladograms pre- sented in A-D. Consensus indices for this cladogram are: Colless’ (1980) CF = 0.500 and Rohlfs (1982) Cl, B 0.667. 188 THE JOURNAL OF ARACHNOLOGY supported by the electrophoretic cladograms and depicted in the Adams consensus tree (Fig. 8) reflect the evolutionary relationships of these fos- sorial spiders and suggest probable scenarios for the historical colonization of some of the Chan- nel Islands. In most instances, the electropho- retic data correspond well with the known geo- logical history of the islands and adjacent main- land. During the late Pleistocene, eustatically lowered sea levels united the four northern Chan- nel Islands (San Miguel, Santa Rosa, Santa Cruz, Anacapa) into a single land mass, Santarosae (Orr 1968). Santarosae began its final breakup only about 16,000 years ago (Vedder & Howell 1980; Johnson 1983). The former physical connection of San Miguel, Santa Rosa and Santa Cruz Is- lands appears to be reflected in the close genetic relationship of the Lutica populations of these islands: all three island populations are members of new species A and in four of the five clado- grams (Figs. 3-5, 7), at least two of these three islands are placed in the same clade within new species A. In contrast, the southern Channel Is- lands (San Nicolas, Santa Barbara, Santa Cata- lina, San Clemente) were never physically con- nected (Vedder & Howell 1980; Johnson 1983). Since two of these islands (San Nicolas and Santa Barbara) were submerged during the middle Pleistocene (Johnson 1983), they derived their biota from other sources since that time. The fact that the Lutica populations of the two south- ern islands which were not submerged, Santa Catalina and San Clemente, are considered to be very different on both electrophoretic and mor- phological grounds is indicative of the absence of significant gene flow that would have been provided by an inter-island connection and may reflect the fact that these islands were originally far apart, prior to San Clemente’s arrival at its present location due to terrane transport (Crouch 1979; Homafius et al. 1986). The mainland populations of new species A and B are only about 57 km apart at their south- ern and northern boundaries respectively [be- tween La Jolla Beach (LJB), Ventura County and the Ballona Wetlands (BA), Los Angeles County] yet spiders from these regions are members of different taxa. This disjunction may simply re- flect the fact that there are no relatively contin- uous dune systems in the intervening coastal area between Ventura County and Los Angeles (Coo- per 1967; Powell 1981) which might act as a corridor for gene flow between these species. On the other hand, this disjunction may be associ- ated with geologic changes that occurred in this region beginning in the Pliocene. During this time the Los Angeles basin was flooded (Murphy 1983a), which, coupled with the northward ex- tension of the Sea of Cortez, caused complete isolation or severe restriction of the movements of organisms to and from Baja California at its northern end (Durham & Allison 1960), a situ- ation which lasted till the Pleistocene (Murphy 1983a). This so called San Gorgonio Barrier has been implicated as a historic biogeographic ob- stacle for the movement of certain xeric-adapted reptiles (Murphy 1983b) and may be at least partly the cause of the disjunction between the main- land populations of Lutica from its northern (new species A) and central (new species B) mainland regions. The Vizcaino Peninsula has alternately been united with and separated from Baja California by sea level changes since the Eocene (Durham & Allison 1960; Murphy 1983a). Since an arid desert lies between this region and the northern portion of Baja (Crosswhite & Crosswhite 1982), it is probable that the divergence between the Lutica populations of the Vizcaino Peninsula and those of northern Baja California is an ancient one, as has been shown for the vegetation of these regions (Axelrod 1979, 1980). The considerable genetic and morphological differences between the population of Guerrero Negro (new species C) and populations to the north is consistent with the geologic history outlined above and indicates a long absence of gene flow between spiders of these two regions. Patterns of colonization. —Some of the genetic relationships are indicative of certain likely col- onization events involving the Channel Islands. These will be reviewed for each island or group of islands in the sections which follow. Northern Channel Islands, Santa Barbara and San Nicolas Islands: The populations of San Ni- colas Island were consistently most closely grouped with one of the northern Channel Is- lands (usually Santa Rosa), indicating probable colonization of this formerly submerged south- ern Channel Island from the islands 80 km to the north. Santa Barbara Island was also sub- merged in the Pleistocene and is also a member of new species A. Since there is no particular population(s) with which it is consistently grouped, all that can be deduced is that colonists of Santa Barbara were derived from one of the new species A populations, most of which are located to the north. In the case of both San RAMIREZ & BECKWITT — PHYLOGENY AND BIOGEOGRAPHY OF LUTICA 189 Nicolas and Santa Barbara, rafting colonists from the north would have been aided by south flow- ing ocean currents and prevailing northwest winds that have been implicated in the dispersal of oth- er organisms in this region (examples in Power 1980; Cowen 1985), as well as the south flowing longshore current (Ledig & Conkle 1983). How- ever, the ocean current patterns in this region are not invariant and the southward flowing Cali- fornia Current is known to reverse its direction during El Nino events (Cowen 1985), perhaps making it possible for propagules to drift north- ward from Santa Catalina Island to Santa Bar- bara Island. The fact that the Lutica populations of the northern Channel Islands and adjacent mainland of Santa Barbara and Ventura Counties are members of the same species is typical of the close relationships that have been reported for island and mainland populations of other organ- isms in this region (e. g., sand crickets, Rentz & Weissman 1973; Weissman & Rentz 1976; deer mice, Ashley & Wills 1987). Indeed, 89% of the orthopteran fauna of the northern Channel Is- lands also occurs in the Santa Monica Mountains (Rentz & Weissman 1981). While the general interpretation of such distributions and relation- ships has been that colonists from the adjacent mainland founded the island populations (Rentz & Weissman 1981; Ashley & Wills 1987), the geologic relationships among the islands and mainland were considerably different in the past, rendering considerations of dispersal among what may be recent subdivisions possibly suspect. For example, the northern Channel Islands were sit- uated as much as 8° to the south of their present locations during the middle Miocene, prior to northward transport on a terrane (Kamerling & Luyendyk 1985). The southern origin for these islands may mean that the actual colonists of these islands came from mainland populations of southern California or Baja California. On the other hand, the populations of these islands and the adjacent mainland were perhaps established at more or less the same time by colonists from San Clemente Island or mainland populations of new species B. While these and other coloniza- tion scenarios may be plausible, the electropho- retic data do not allow one to determine direc- tions of colonization between the island and mainland populations of new species A, nor do they establish the actual sister group of this spe- cies, rendering consideration of mainland-island colonization scenarios involving populations of this species unwarranted at this time. Santa Catalina Island: Since Santa Catalina Island (along with Santa Rosa and Santa Cruz Islands) may have been continuously above wa- ter since the Oligocene (Vedder & Howell 1980; Haq et al 1987), has remained relatively sta- tionary during this period (Luyendyk et al. 1 985) and is close to the southern California mainland, one would expect that its closest biotic relation- ship should be with populations from the adja- cent Los Angeles-San Diego coastal strip. How- ever, we have shown that the Lutica population of Santa Catalina is most closely related to new species A populations rather than populations on the southern California mainland. The most parsimonious explanation for such a relationship is that spiders from the adjacent mainland never colonized Santa Catalina and so new species A spiders were the first and only colonists. On the other hand, such a pattern of relationships may be due to the extinction of an original insular form derived from the mainland prior to colo- nization by new species A spiders or because new species A spiders proved to be superior in com- petition with the native insular form. Separate colonizations of Santa Catalina Island from the northern Channel Islands and southern Califor- nia mainland have been proposed for Channel Island deer mice, Peromyscus maniculatus , due to mitochondrial DNA restriction fragment polymorphisms found among Santa Catalina Is- land mice (Ashley & Wills 1987). Extinction has also been suggested to explain the distribution of the island night lizard, Klauberina , which is found on San Clemente, San Nicolas and Santa Barbara Islands but not on Santa Catalina Island (Crother et al. 1986; Bezy & Sites 1987). San Clemente Island: A biogeographic rela- tionship between San Clemente Island and Baja California has been proposed by Crother et al. (1986), based on a cladistic study of morphology and karyology within the lizard family Xantu- siidae and a vicariance model based on terrane movement linking San Clemente and central Baja California. Based on geophysical evidence (Crouch 1979; Homafius et al. 1986), it is clear that San Clemente Island was in close proximity to central Baja California up to about 18 million years ago, when the terrane on which it is situated started moving north along the San Clemente Island Fault, eventually reaching its present po- sition about 5-8 million years ago. Crother et al. (1986) hypothesize that relatively sedentary taxa 190 THE JOURNAL OF ARACHNOLOGY (like xantusiid lizards) which occupied San Cle- mente Island and the adjacent mainland of Baja California prior to the time of northward move- ment (and whose descendants continue to oc- cupy these areas today) should be closely related. Since Lutica is clearly sedentary and has dis- tinct species which occupy San Clemente Island {demented) and central Baja California (new spe- cies C), we hoped to be able to test Crother et al.’s (1986) vicariance hypothesis using the elec- trophoretic data reported herein. However, given the need to use Guerrero Negro (new species C) as an outgroup in our phylogenetic analyses, in the absence of an appropriate zodariid taxon, it was not possible to determine whether dementea is indeed most closely allied with new species C of central Baja California. As such, a final de- cision concerning Lutica' s involvement in the biogeographic hypothesis of Crother et. al. (1986) will have to await a future phylogenetic analysis using an actual outgroup taxon. SUMMARY The geological history of the California Chan- nel Islands and mainland of southern California and Baja California involves extensive sea level changes and the movement of terranes. These geomorphic changes may have influenced the evolution of taxa in this region, particularly if they are sedentary and biologically old. Analysis of the results of an electrophoretic survey of populations of the spider genus Lutica from much of its range revealed fixed allelic dif- ferences that clearly define two species: new spe- cies A [Santa Barbara and Ventura Counties, northern Channel Islands (San Miguel, Santa Rosa, Santa Cruz), southern Channel Islands (San Nicolas, Santa Barbara, Santa Catalina)] and new species C [Guerrero Negro, central Baja Califor- nia], While diagnostic loci were not found for the population of San Clemente Island {elemented) or the mainland populations of southern Cali- fornia and northern Baja California, they are morphologically recognizable units according to Gertsch, so clementea was accepted as valid, while the mainland populations were assigned to new species B. Phylogenetic analysis of the electrophoretic data using a variety of methods revealed that evolutionary rates among the 19 populations sampled have been very unequal. The phyloge- netic relationships among populations consis- tently supported by the electrophoretic clado- grams generally correspond with the geological history of the Channel Islands and adjacent mainland and suggest certain likely colonization events involving some of the islands. A future electrophoretic study of Lutica in- volving more loci and a zodariid taxon as an outgroup (chosen in light of Jocque 1991) is needed to A) genetically validate the species sta- tus of the mainland populations of southern Cal- ifornia and northern Baja California, as well as of clementea , and B) to resolve the phylogenetic relationships among the four species (new species A, B, C, clementea). Further systematic studies of other monophyletic taxa (particularly those which are biologically old and poor dispersers) occupying the California Channel Islands and mainland of southern California and Baja Cali- fornia are needed to better understand the geo- logic and biogeographic evolution of this region. Given the lack of even basic knowledge con- cerning many taxonomic groups in this area, par- ticularly among the invertebrates, this will be a fruitful area for the conduct of systematic and biogeographic studies. ACKNOWLEDGMENTS We are particularly indebted to W. Gertsch for sparking our initial interest in Lutica and to M. Thompson for sharing with us his field note- books and maps based on his own Lutica col- lecting efforts in the 60’s and 70’s. C. Cutler, C. Drost, W. Icenogle, M. Wilson and members of the Arachnologists of the Southwest provided assistance in the field. K. Akiyama, A. Jimenez and M. Sandoval, participants in the National Institutes of Health Minority High School Stu- dent Research Apprentice Program, were expert co-workers in the lab in the summer of 1987, while R. Garthwaite was a continuing source of technical advice concerning electrophoretic pro- cedures. T. Cranford and K. Huie provided gen- eral computer advice and D. Wake and M. Fre- low helped with analysis of our data using BIO- SYS-1. R. Bennett, J. Cronin, J. Estes, L. Fox, C. Griswold and D. Potts offered many construc- tive comments on various drafts of this manu- script. For providing collecting permits and access to various mainland localities, we thank: S. Clarke, Marine Science Institute, University of Califor- nia, Santa Barbara; P. Principe, Department of Airports, City of Los Angeles; and the California Department of Parks and Recreation. For pro- viding collecting permits and access to the Cal- ifornia Channel Islands, and for providing lodg- RAMIREZ & BECKWITT — PH YLOGENY AND BIOGEOGRAPHY OF LUTICA 191 ing and transportation while on the islands, we are grateful to: W. Ehom and F. Ugolini, Channel Islands National Park; S. Clarke, Marine Science Institute, University of California, Santa Bar- bara; L. Laughrin, Santa Cruz Island Reserve; A. Propst and T. Martin, Santa Catalina Island Conservancy; S. Bennett and R. Turner, Catalina Island Marine Institute; J. Estes, U.S. Fish and Wildlife Service; R. Dow, J. Larson and L. Sal- ata, U.S. Navy. Financial support was provided by grants from a variety of sources at the Uni- versity of California, Santa Cruz (Div. of Natural Sciences, Graduate Studies and Research, Dept, of Biology, Patent Fund, Minority Biomedical Research Support Program) and by a grant from the Exline-Frizzell Fund for Arachnological Re- search (Grant No. 10). LITERATURE CITED Adams, E. N. III. 1972. Consensus techniques and the comparison of taxonomic trees. Syst. Zool., 21: 390-397. Ashley, M. & C. Wills. 1987. Analysis of mitochon- drial DNA polymorphisms among Channel Island deer mice. Evolution, 41:854-863. Avise, J. C. 1994. Molecular Markers, Natural His- tory and Evolution. Chapman and Hall, New York. Axelrod, D. I. 1979. Age and origin of Sonoran Desert vegetation. Occas. Pap. California Acad. Sci., 132: 1-74. Axelrod, D. I. 1980. History of the maritime closed- cone pines, Alta and Baja California. Univ. Cali- fornia Publ. Geol. Sci., 120:1-143. Berlocher, S. H. 1984. Insect molecular systematics. Annu. Rev. Entomol., 29:403-433. Bezy, R. L. & J. W. Sites. 1987. A preliminary study of allozyme evolution in the lizard family Xantu- siidae. Herpetologica, 43:280-292. Buth, D. G. 1984. The application of electrophoretic data in systematic studies. Ann. Rev. Ecol. Syst., 15:501-522. Carlquist, S. 1981. Chance dispersal. American Sci., 69:509-516. Colless, D. H. 1980. Congruence between morpho- metric and allozyme data for Menidia species: A reappraisal. Syst. Zool., 29:288-299. Cooper, W. S. 1967. Coastal dunes of California. Geol. Soc. America Mem., 104:1-131. Cowen, R. K. 1985. Large scale pattern of recruit- ment by the labrid, Semicossyphus pulcher. Causes and implications. J. Mar. Res., 43:719-742. Crosswhite, F. S. & C. D. Crosswhite. 1982. The Sonoran Desert. Pp. 163-319, In Reference Hand- book of the Deserts of North America. (G. L. Bend- er, ed.). Greenwood Press, Westport, Connecticut. Crother, B. I. 1990. Is “some better than none” or do allele frequencies contain phylogenetically useful information? Cladistics, 6:277-281. Crother, B. I., M. M. Miyamoto & W. F. Presch. 1 986. Phylogeny and biogeography of the lizard family Xantusiidae. Syst. Zool., 35:37-45. Crouch, J. K. 1979. Neogene tectonic evolution of the California Continental Borderland and western Transverse Ranges. Geol. Soc. America Bull., 90: 338-345. Decae, A. E. 1987. Dispersal: Ballooning and other mechanisms. Pp. 348-356, In Ecophysiology of Spi- ders. (W. Nentwig, ed.). Springer- Verlag, New York. Durham, J. W. & E. C. Allison. 1960. The geologic history of Baja California and its marine fauna. Syst. Zool, 9:47-91. Edwards, A. W. F. & L. L. Cavalli-Sforza. 1963. The reconstruction of evolution. Heredity, 18:553. Edwards, A. W. F. & L. L. Cavalli-Sforza. 1964. Re- construction of evolutionary trees. Pp. 67-76, In Phenetic and Phylogenetic Classification. (V. V. Heywood & J. McNeill, eds.). Publ. No. 6, System- atics Association, London. Farris, J. S. 1970. Methods for computing Wagner trees. Syst. Zool., 19:83-92. Farris, J. S. 1 972. Estimating phylogenetic trees from distance matrices. American Nat., 106:645-668. Farris, J. S. 1981. Distance data in phylogenetic anal- ysis. Pp. 3-23, In Advances in Cladistics: Proc. First Meeting of the Willi Hennig Society. (V. A. Funk & D. R. Brooks, eds.). New York Botanical Garden, New York. Farris, J. S. 1988. HENNIG: HENNIG86 reference, version 1.5. Published by the author. Farris,!. S. 1989. HENNIG86: A PC-DOS program for phylogenetic analysis. Cladistics, 5:163. Felsenstein, J. 1981. Evolutionary trees from gene frequencies and quantitative characters: Finding maximum likelihood estimates. Evolution, 35: 1 229- 1242. Felsenstein, J. 1982. Numerical methods for inferring evolutionary trees. Q. Rev. Biol, 57:379-404. Felsenstein, J. 1988. PHYLIP - Phylogeny inference package (version 3.1). Univ. Washington, Seattle. Felsenstein, J. 1 989a. PHYLIP - Phylogeny inference package (version 3.2). Cladistics, 5:164-166. Felsenstein, J. 1989b. PHYLIP - Phylogeny inference package (version 3.2). Univ. Washington, Seattle. Gertsch, W. J. 1961. The spider genus Lutica. Senck- enb. Biol., 42:365-374. Haq, B. U., J. Hardenbol & P. R. Vail. 1987. Chro- nology of fluctuating sea levels since the Triassic. Science, 235:1156-1167. Harris, H. & D. A. Hopkinson. 1976. Handbook of Enzyme Electrophoresis in Human Genetics. North Holland Publishing, Amsterdam. Homafius, J. S., B. P. Luyendyk, R. R. Terres & M. J. Kamerling. 1986. Timing and extent of Neogene tectonic rotation in the western Transverse Ranges, California. Geol. Soc. America Bull., 97:1476-1487. 192 THE JOURNAL OF ARACHNOLOGY Jocque, R. 1991. A generic revision of the spider family Zodariidae (Araneae). Bull. American Mus. Nat. Hist, 201:1-160. Jocque, R. 1993. “We’ll meet again”, an expression remarkably applicable to the historical biogeogra- phy of Australian Zodariidae (Araneae). Mem. Queensland Mus., 33:561-564. Johnson, D. L. 1983. The California continental bor- derland: Landbridges, watergaps and biotic disper- sals. Pp. 482-527, In Quaternary Coastlines and Marine Archaeology: Towards the Prehistory of Landbridges and Continental Shelves. (P. M. Mas- ters & N. C. Fleming, eds.). Academic Press, Lon- don. Jones, T. R., A. G. Kluge & A. J. Wolf. 1993. When theories and methodologies clash: A phylogenetic reanalysis of the North American ambystomatid salamanders (Caudata: Ambystomatidae). Syst. Biol., 42:92-102. Kamerling, M. J. & B. P. Luyendyk. 1985. Paleo- magnetism and Neogene tectonics of the northern Channel Islands, California. J. Geophys. Res., 90: 12485-12502. Kluge, A. G. & J. S. Farris. 1969. Quantitative phy- letics and the evolution of anurans. Syst. Zook, 18: 1-32. Lanyon, S. M. 1985. Detecting internal inconsisten- cies in distance data. Syst. Zook, 34:397-403. Ledig, F. T. & M. T. Conkle. 1983. Genetic diversity and genetic structure in a narrow endemic, Torrey Pine ( Pinus torrey ana Parry Ex Carr). Evolution, 37: 79-85. Luyendyk, B. P., M. J. Kamerling, R. R. Terres & J. S. Homafius. 1985. Simple shear of southern Cal- ifornia during Neogene time suggested by paleo- magnetic declinations. J. Geophys. Res., 90:12454- 12466. Marx, G. 1891. A contribution to the knowledge of North American spiders. Proc. Entomol. Soc. Washington, 2:28-37. Mayr, E. 1970. Populations, Species and Evolution. Harvard Univ. Press, Cambridge, Massachusetts. Mickevich, M. F. & C. Mitter. 1981. Treating poly- morphic characters in systematics: A phylogenetic treatment of electrophoretic data. Pp. 45-60, In Ad- vances in Cladistics: Proc. First Meeting of the Willi Hennig Society. (V. A. Funk & D. R. Brooks, eds.). New York Botanical Garden, New York. Mickevich, M. F. & N. I. Platnick. 1989. On the information content of classifications. Cladistics, 5:33-47. Murphy, R. W. 1 983a. The reptiles: Origins and evo- lution. Pp. 130-158, In Island Biogeography in the Sea of Cortez. (T. J. Case & M. L. Cody, eds.). Univ. California Press, Berkeley. Murphy, R. W. 1983b. Paleobiogeography and ge- netic differentiation of the Baja California herpto- fauna. Occas. Pap. California Acad. Sci., 137:1-48. Nei, M. 1 972. Genetic distance between populations. American Nat., 106:283-292. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of indi- viduals. Genetics, 89:583-590. Nei, M. 1 987. Molecular Evolutionary Genetics. Co- lumbia Univ. Press, New York. Nelson, G. 1979. Cladistic analysis and synthesis: Principles and definitions, with a historical note on Adanson’s Families des Plantes (1763-1764). Syst. Zook, 28:1-21. Orr, P. C. 1968. Prehistory of Santa Rosa Island. Santa Barbara Mus. Nat. Hist., Santa Barbara, Cal- ifornia. Petrunkevitch, A. 1942. A study of amber spiders. Trans. Connecticut Acad. Arts Sci., 34:1 19-464. Petrunkevitch, A. 1952. Macroevolution and the fos- sil record of the Arachnida. American Sci., 40:99- 122. Powell, J. A. 1981. Endangered habitats for insects: California coastal sand dunes. Atala, 6:41-55. Power, D. M. (ed.) 1 980. The California Islands: Proc. Multidisciplinary Symposium. Santa Barbara Mus. Nat. Hist., Santa Barbara, California. Prager, E. M. & A. C. Wilson. 1976. Congruence of phytogenies derived from different proteins: A mo- lecular analysis of the phylogenetic position of cra- cid birds. J. Mol. Evok, 9:45-57. Quicke, D. L. J. 1993. Principles and Techniques of Contemporary Taxonomy. Chapman & Hall, New York. Ramirez, M. G. 1990. Natural history, population genetics, systematics and biogeography of the spider genus Lutica (Araneae: Zodariidae). Ph.D. disser- tation, Univ. California, Santa Cruz. Ramirez, M. G. 1995. Natural history of the spider genus Lutica (Araneae, Zodariidae). J. Arachnok, 23:111-117. Rentz, D. C. & D. B. Weissman. 1973. The origins and affinities of the Orthoptera of the Channel Is- lands and adjacent mainland California. Part I: The genus Cnemotettix. Proc. Acad. Nat. Sci. Philadel- phia, 125:89-120. Rentz, D. C. & D. B. Weissman. 1981. Faunal affin- ities, systematics and bionomics of the Orthoptera of the California Channel Islands. Univ. California Publ. Entomoh, 94:1-240. Richardson, B. J., P. R. Baverstock & M. Adams. 1986. Allozyme Electrophoresis: A Handbook for Animal Systematics and Population Studies. Academic Press, New York. Rogers, J. S. 1972. Measures of genetic similarity and genetic distance. Studies in Genetics VII, Univ. Tex- as Publ., 7213:145-153. Rohlf, F. J. 1982. Consensus indices for comparing classifications. Math. Biosci., 59:131-144. Rohlf, F. J. & M. C. Wooten. 1988. Evaluation of the restricted maximum-likelihood method for es- RAMIREZ & BECKWITT — PH YLOGENY AND BIOGEOGRAPHY OF LUTICA 193 timating phylogenetic trees using simulated allele- frequency data. Evolution, 42:581-595. Rohlf, F. J., D. H. Colless & G. Hart. 1983. Taxo- nomic congruence re-examined. Syst. Zool., 32: 1 44— 158. Shaffer, H. B., J. M. Clark & F. Kraus. 1991. When molecules and morphology clash: A phylogenetic analysis of the North American ambystomatid sal- amanders (Caudata: Ambystomatidae). Syst. Zool., 40:284-303. Sneath, P. H. A. & R. R. Sokal. 1973. Numerical Taxonomy. Freeman, San Francisco. Swofford, D. L. 1981. On the utility of the distance Wagner procedure. Pp. 25-43, In Advances in Cla- distics: Proc. First Meeting of the Willi Hennig So- ciety. (V. A. Funk & D. R. Brooks, eds.). New York Botanical Garden, New York. Swofford, D. L. 1985. PAUP: Phylogenetic analysis using parsimony, version 2.4. 1 . Illinois Natural His- tory Survey, Champaign. Swofford, D. L. 1988. FREQPARS, release 1.0. Il- linois Natural History Survey, Champaign. Swofford, D. L. & S. H. Berlocher. 1987. Inferring evolutionary trees from gene frequency data under the principle of maximum parsimony. Syst. Zool., 36:293-325. Swofford, D. L. & G. J. Olsen. 1990. Phylogeny re- construction. Pp. 411-501, In Molecular System- atics. (D. M. Hillis & C. Moritz, eds.). Sinauer, Sun- derland, Massachusetts. Swofford, D. L. & R. B. Selander. 1981. BIOSYS-1: A FORTRAN program for the comprehensive anal- ysis of electrophoretic data in population genetics and systematics. J. Hered., 72:281-283. Swofford, D. L. & R. B. Selander. 1989. BIOSYS-1: A computer program for the analysis of allelic vari- ation in population genetics and biochemical sys- tematics, release 1.7. Illinois Nat. Hist. Surv., Champaign. Thompson, M. 1 973. The spider genus Lutica. Whit- tier Narrows Nature Center Quart., 1:1-5. Tikader, B. K. 1981. Studies on spiders of the genus Lutica Marx (Family-Zodariidae) from India. J. Bombay Nat. Hist. Soc., 78:139-142. Vedder, J. G. & D. G. Howell. 1980. Topographic evolution of the Southern California Borderland during late Cenozoic time. Pp. 7-31, In The Cali- fornia Islands: Proc. Multidisciplinary Symposium. (D. M. Power, ed.). Santa Barbara Mus. Nat. Hist., Santa Barbara, California. Weissman, D. B. & D. C. Rentz. 1976. Zoogeography of the grasshoppers and their relatives (Orthoptera) on the California Channel Islands. J. Biogeogr., 3:105-114. Manuscript received 26 November 1994, revised 6 Sep- tember 1 995. 1995. The Journal of Arachnology 23:194-198 REDESCRIPTION OF STENOSTYGNUS PUSIO SIMON AND SYNONYMY OF CARIBBIANTINAE WITH STENOSTYGNINAE (OPILIONES: LANIATORES, BIANTIDAE) Ricardo Pinto-da-Rocha: Museu de Zoologia, Universidade de Sao Paulo, Caixa Postal 7172, Sao Paulo, SP, Brazil, 01064-970 ABSTRACT. Based on the lack of the tarsal process and genitalic features, Stenostygnus pusio Simon 1879 is transferred from the Stygnidae to the Biantidae. This is the first record of the family Biantidae from Brazil, Colombia and Ecuador. The subfamily Caribbiantinae is synonymized with Stenostygninae. Genera and species other than Stenostygnus pusio , formerly placed in the Stenostygninae are transferred to Heterostygninae (Styg- nidae). The convergences between Stygnidae and Biantidae are presented and discussed. Stenostygnus pusio Simon 1879 was described based on a single female from Tefe (Amazonas, Brazil) and included in the then created new sub- family of Gonyleptidae, Stygninae. Roewer (1913) established the subfamily Stenostygninae for Stenostygnus Simon and Stenostygnellus Roew- er. Sorensen (1932) erected the family Stygnidae for the Gonyleptidae subfamilies without a com- mon eye mound (Heterostygninae, Phareinae, Stenostygninae and Stygninae). Up to now, S. pusio was known only by the holotype. Capo- riacco (1951) described Stenostygnus magnus from Venezuela. The study of Stenostygnus pusio , especially of the male genitalia, showed me that this species belongs to the Biantidae. The external characters included it in the Caribbiantinae, which is a ju- nior synonym of Stenostygninae. The specimens studied are deposited in: Mu- seu de Zoologia da Universidade de Sao Paulo, Sao Paulo (MZSP); Museu de Ciencias Naturais, Porto Alegre (MCN); Museum of Comparative Zoology, Cambridge (MCZ); Museum National d’Histoire Naturelle, Paris (MNHN); and the National Museum of Natural History, Washing- ton (USNM). All measurements are in milli- meters. Biantidae Thorell Biantidae Thorell 1889: 670; Sorensen, 1932: 224; Martens, 1978: 358; Shear, 1982: 107. Diagnosis. —Cephalothorax almost as wide as abdomen. Eye mounds separate. Chelicera seg- ment I smooth, condyle-like. Tarsi III-IV with or without scopulae; without tarsal process. Pe- nis with titillators and conductors; stylus long and thin. Subfamilies included.— Biantinae, Dibuninae, Stenostygninae and Zairebiantinae. However, the opiliologists are not unanimous as regards com- position of the subfamilies of the Biantidae. Roewer (1923) considered all subfamilies as members of Phalangodidae, Sorensen (1932) considered Dibuninae as a family, but Martens (1978) and Shear (1982) did not consider Car- ibbiantinae (now Stenostygninae) as a Biantidae and finally, Martens (1978) considered Dibuni- nae as a Phalangodidae member. Stenostygninae Roewer Stenostygninae Roewer 1913: 163; 1923: 459; Mello- Leitao 1932: 418. Caribbiantinae Silhavy 1973: 133; Avram 1977: 123. SYN. N. Diagnosis.— Eye mounds near groove I. Four areas on dorsal scute, area I with or without me- dian groove. Male chelicera swollen or not swol- len. Pedipalps long and thin; dorsally unarmed; coxa conical; patella with or without mesoapical seta. Tarsi III-IV with scopulae of spatulate hairs; claws smooth and opposite; distitarsus I with three and II with four segments. Metatarsus III spindled. Penis without membranous region in the ventral plate. Stenostygnus Simon Stenostygnus Simon: 1879: 219, 224; Roewer 1913: 163; 1923: 460; Mello-Leitao 1932: 418 (type spe- cies: Stenostygnus pusio Simon 1 879, by monotypy). Diagnosis.— Differs from the other Stenostyg- ninae genera by the lack of spines or prominent 194 PINTO-DA-ROCHA — REDESCRIPTION OF STENOSTYGNUS 195 Figures 1-4.— Stenosty gnus pusio Simon, female holotype. 1, Habitus, dorsal view; 2, Lateral view; 3, Ventral view of left pedipalp, tibia and tarsus; 4, Dorsal view of distal part of penis. Scale bar: Figs. 1-3 = 1.0 mm; Fig. 4 = 0.5 mm. tubercles on dorsal scute; male chelicera not swollen; pedipalpal patella without apical setae; male tarsus I with 8 segments, instead of 5-7. Stenostygnus pusio Simon Figs. 1-7 Stenostygnus pusio Simon 1879: 224; Roewer 1913: 163, fig. 71 (redescription); 1923: 460, fig. 578 (re- description); Mello-Leitao: 1923: 133 (citation); 1932: 419 (redescription), (holotype female “Teffe, coll. Simon, Paris, n° 4007, type” MNHN, examined). Male description.— (USNM). Measurements: Dorsal scute 1.54 long, 1.22 wide; cephalothorax 0.66 long, 1.10 wide. Chelicera: segment II 0.66 long, III 0.22 long. Dorsal scute : (Figs. 1, 2). Cephalothorax and eyes mounds smooth. An- terior margin slightly raised laterally. Areas I-IV undivided and with minute tubercles. Lateral margin with one row of large tubercles from coxa III to posterior margin. Posterior margin with one row of large tubercles, median row of tu- bercles smaller. Free tergites I— III with one row of small tubercles. Anal opercle tuberculate. Ven- ter: Coxae I III with minute tubercles, lacking apically; coxa IV smooth. Stigmatic area smooth, long; stigmae concealed; free stemites with one transversal row of minute tubercles. Anal opercle tuberculate. Chelicera : Segment I smooth, II not swollen; chelicera fingers with teeth. Pedipalps: Coxa with two ventral tubercles; trochanter smooth; femur straight, without ventrobasal tu- bercle; patella swollen at apical third; tibia (Fig. 3) with one ventrobasal seta and three longer setae on each side. Tarsus (Fig. 3) with two long setae on each side. Legs: Long, straight and thin. Patela IV with three prolateral tubercles and one dorsoapical. Tarsal claws smooth, double and lying perpendicular to the axis of the leg. Scop- ulae dense on last segment, with spatulate hairs. Tarsal segmentation: 8(3), 12(4), ?, 6. The spec- imens examined lacked the third pair of legs. Penis : (Figs. 5-7). Ventral plate with an apical cleft “U” shaped; with five retrolateral pairs of setae and two subapical ventral pairs of bifid setae. With titillators; stylus and conductors con- cealed by the titillators. Coloration: Brownish, pedipalpus yellowish. Female supplemental description (holotype). — Measurements: Dorsal scute 1.40 long, 1.92 wide; cephalothorax 0.52 long, 1.16 wide. Chelicera: segment II 0.52 long, III 0.26 long. (Measure- ments of appendages in Table I.) Female similar to the male. Lateral margin of dorsal scute with a row of large tubercles from coxa II to posterior margin. Patella IV without tubercles. Tarsal segmentation: 6(3), 9(3), 6, 6. 196 THE JOURNAL OF ARACHNOLOGY Figures 5-1 .—Stenosty gnus pusio Simon. Scanning electron micrographs of distal part of penis (USNM). 5, Dorsal view; 6, Lateral view; 7, Ventral view. Scale bar = 0.2 mm. Type locality . — Tefe, Amazonas, Brazil (03°22'S - 64°42'W). Distribution.— Recorded from Amazon basin (Colombia, Ecuador and Brazil). All records were made from localities near the tributaries of Am- azon/Solimdes river. Roewer (1913, 1923) re- corded the “type” as being from Cayenne (French Guiana) and a second specimen, not belonging to the type series, from Tefe (Brazil). I examined the holotype and it was labeled “Teffe” as in Simon’s description. There are no other speci- mens of S. pusio in the MNHN, and I think that Roewer was mistaken. Material examined.— BRAZIL. Amazonas : Tefe, 9 holotype (MNHN); Solimoes River, 30 April 1966, 3 6 (USNM); Alto Solimoes, 20 De- cember 1979, A. Lise leg , 19 (MCN). COLOM- BIA. Amazonas: Amacayacu (National Parq 48 km NW from Leticia, 03°48'S, 70°16'W), 90-100 m, 3 October 1985, 19 (MCZ). ECUADOR. Napo: Pompeya (Napo river), May 1965, Pena leg, 19 (MCZ). DISCUSSION Simon (1879) described Stenostygnus pusio within the Gonyleptidae, subfamily Stygninae. Simon’s subfamily, characterized by the lack of a common eye mound, was divided by Roewer (1913) into four separate subfamilies: Heteros- tygninae, Phareinae, Stenostygninae and Styg- ninae. Stenostygninae was established for two monotypic genera, Stenostygnellus (with one species, S.flavolimbatus Roewer 1913) and Sten- ostygnus. Later, the following genera and species were described in the Stenostygninae: Bunistyg- nellus beebei Goodnight & Goodnight 1949; B. macrochelis Roewer 1916; B. ornatus Roewer 1943; Dichobunistygnus ephippiatus Roewer 1915; Hoplostygnus albicinctus Roewer 1915; Stenostygnellus praetiosus Caporiacco 1951; Xanthostygnus fractus Mello-Leitao 1949 and another species of Stenostygnus (S. magnus Ca- poriacco 1951). During the review of the family Stygnidae (Pinto-da-Rocha unpubl. data), I examined the holotype and some other specimens of S. pusio. I didn’t observe a tarsal process on legs III-IV and from the examination of the genitalic fea- tures, I concluded that there is no evidence to place Stenostygnus in the Stygnidae. Simon (1879) stated that the holotype had a well developed pseudonychium (= tarsal process) and pectinate claws, but I observed only a long hair and smooth claws. The presence of a tarsal process is syna- porn orphic for Cranaidae + Gonyleptidae + Cosmetidae + Stygnidae (Kury 1992). The tarsal process is historically an important feature to distinguish the families of Gonyleptoidea with- out the common eye mound: Podoctidae/Bian- tidae/Stygnommatidae (tarsal process absent) from Stygnidae (present). However, some species of Stygnidae have secondaryly lost the tarsal pro- cess (e. g., Auranus, Pickeliana, unpubl. data). The genitalic features (presence of titillators) confirm the position of S. pusio in Biantidae. The Biantidae are currently divided in four subfamilies: Biantinae, Stenostygninae (= Car- ibbiantinae), Dibuninae and Zairebiantinae. Biantinae have scopulae, four areas on dorsal scute and parallel claws; Stenostygninae have four PINTO-DA-ROCHA— REDESCRIPTION OF STENOSTYGNUS 197 Table 1.— Appendage measurements of the female holotype of Stenostygnus pusio Simon. Tr = trochanter, Fe = femur, Pt = patella, Ti = tibia, Mt = metatarsus, Ta = tarsus. Measurements are in mm. Tr Fe Pt Ti Mt Ta Total Leg I 0.16 1.40 0.32 1.00 1.48 0.92 5.28 II 0.26 3.36 0.44 2.96 4.60 1.76 13.38 III 0.20 2.28 0.36 1.20 2.50 0.78 7.32 IV 0.26 3.20 0.48 1.62 3.96 0.82 10.34 Pedipalp 0.22 1.10 0.64 0.56 - 0.40 2.92 areas, scopulae and opposite claws; Dibuninae have three areas, no scopulae and parallel claws. Zairebiantinae seem not to be a Biantidae be- cause the eyes are placed on two distinct mounds close to each other and located in the middle of the cephalothorax (Kauri 1985) instead of far apart and near the line I. Unfortunately, the gen- italia of Zairebiantes microphthalmus are poorly known and the drawing of Kauri (1985, fig. 249) is difficult to relate to any other family/subfamily of the Gonyleptoidea. There are no cladistic hy- potheses for Biantidae subfamilies; but I believe that opposite claws and spindled metatarsi III are synapomorphic for the Stenostygninae. The presence of a membranous region ventrally in the ventral plate (Schwellkorper of Martens 1978, 1986) and the small number of articles in tarsi I (usually three) are synapomorphic for the Bian- tinae. The genitalia of the Dibuninae species are unknown. Members of the Stenostygninae have been recorded from the Antilles (Silhavy 1973; Avram 1977), the Biantinae are recorded from the Oriental region (Roewer 1923; Martens 1978) and the Ethiopian region (Roewer 1923; Rambla 1982; Kauri 1985) and the Dibuninae from the Oriental region (Roewer 1923). Stenostygnus pu- sio is the first species of Biantidae recorded from Brazil, Colombia and Ecuador and is the second known representative of the family in South America. Gonzalez-Sponga (1992) recorded an unidentified species from Venezuela (specific lo- cality not mentioned). Stenostygnus pusio is closely related to the “Caribbiantinae” by presence of opposite claws and large number of articles on tarsi I; and it lacks the membranous region ventrally on ven- tral plate (plesiomorphic). Another synapomor- phy of the Stenostygninae, male metatarsi III spindled, couldn’t be observed in S. pusio be- cause all males studied were without third legs. The convergences between some Biantidae and some Stygnidae are remarkable. Those of note are the opposite claws (Stenostygninae, Heter- ostygninae), scopulae with spatulate hairs (Bian- tinae, Stenostygninae, Heterostygninae), last tar- sal article III-IV flattened (Stenostygninae, Het- erostygninae), eye mound divided and situated back in the cephalothorax, dorsal scute rectan- gular/slightly elliptical, and pedipalpal coxae/fe- mur/patella lengthened. Based on the convergences mentioned (that are synapomorphic for Heterostygninae) and by the presence of male genitalia with slender ventral plate and posterior claws pectinated, the other genera and species formerly placed in the Sten- ostygninae are transferred to the Heterostygni- nae. Stenostygnus magnus Caporiacco will be the type species of a new genus of Heterostygninae (Pinto-da-Rocha unpubl. data). ACKNOWLEDGMENTS I am grateful to Dr. Arturo Munoz-Cuevas (Museum National d’Histoire Naturelle) for sending me the holotype of S . pusio, to Dr. Ad- riano B. Kuri (Universidade do Rio de Janeiro), Dr. Sonia Casari (Museu de Zoologia da Univ- ersidade de Sao Paulo) and James C. Cokendol- pher for suggestions on the manuscript. I am also grateful to Dr. Alberto A. G. F. C. Ribeiro, head of the Laboratorio de Microscopia Eletronica (Institute de Biociencias da Universidade de Sao Paulo) and to Marcio V. Cruz for the help in the use of the electron microscope. This study was supported by a grant from Funda^ao de Amparo a Pesquisa do Estado de Sao Paulo (#91/4054- 7). LITERATURE CITED Avram, S., 1977. Recherches sur les opilionides de Cuba III. Genres et especes nouveaux de Carib- biantinae (Biantidae, Gonyleptomorphi). Resultats Exp. Biospeol. Cubano-Roumaines a Cuba, 2:123- 136. Caporiacco, L. di. 1951. Studi sugli Aracnidi del Ven- ezuela raccolti dalla Sezione di Biologia (Universida Centrale del Venezuela). I Parte: Scorpiones, Opi- 198 THE JOURNAL OF ARACHNOLOGY liones, Solifuga y Chemetes. Acta biol. Venezuelica, 1:1-46. Gonzalez-Sponga, M. A. 1992. Aracnidos de Vene- zuela. Opiliones Laniatores II. Familia Cosmetidae. Biblioteca de la Academia de Ciencias Fisicas, Ma- tematicas y Naturales, Caracas, vol. 26, 432 pp. Kauri, H. 1985. Opiliones from Central Africa. Ann. Mus. r. l’Afrique Centrale, 8° serie, 245:1-168. Kury, A. B. 1992. The genus Spinopilar Mello-Lei- tao, 1940, with notes on the status of the family Tricommatidae (Arachnida, Opiliones). Steenstru- pia, 18:93-99. Martens, J. 1978. Opiliones aus dem Nepal-Hima- laya. IV. Biantidae (Arachnida). Senckenbergiana Biol., 58:347-414. Martens, J. 1 986. Die Grossgliederung der Opiliones und die Evolution der Ordnung (Arachnida). Actas X Congr. Int. Aracnol. Jaca/Espana, 1:289-310. Mello-Leitao, C. F. 1923. Opiliones Laniatores do Brasil. Arq. Mus. Nac. Rio de Janeiro, 24:105-197. Mello-Leitao, C. F. 1932. Opiliones do Brasil. Revta Mus. Paulista, 17 (2° parte): 1-505, + 61 plates. Rambla, M. 1982. Contributions a l’etude de la faune terrestre des iles granitiques de l’archipel des Se- chelles. Opiliones (Arachnida). Ann. Mus. Royal L’Afrique Centrale, 8° serie, 242:86. Roewer, C. F. 1913. Die Familie der Gonyleptiden der Opiliones-Laniatores. Arch. Naturg., 79A: 1-256. Roewer, C. F. 1923. Die Weberknechte der Erde. Systematische Bearbeitung der bisher bekannten Opiliones. Jena, Gustav Fisher. 1, 1 16 p. Shear, W. 1982. Opiliones. Pp. 104-110, vol. 2 In Synopsis and classification of living organisms. (S. P. Parker, ed.). McGraw-Hill Book Co., New York. Silhavy , V. 1973. Two new systematic groups of gon- yleptomorphid phalangids from the Antillean-Ca- ribbean region, Agoristenidae fam. n. and Carib- biantinae subfam. n. (Arachn.: Opilionidea). Ves- tinik Ceskoslovenske Spol. Zool., 37:1 10-143. Simon, E. 1879. Essai d’une classification des Opi- liones Mecostethi. Ann. Soc. Entomol. Belgique, 22: 183-241. Sorensen, W. 1932. Descriptions Laniatorum (Ar- achnidorum opilionum Subordinis). Kongl. Danske Vidensk. Selsk. Skr. Naturvidensk. Math. Afd. (Ko- benhavn), (9) 3 (4): 199-422. Thorell, T. 1889. Aracnidi Artrogastri Birmani rac- colti da L. Fea nel 1885-1887. Ann. Mus. Civ. Stor. Nat. Genova, 27:521-729. Manuscript received 31 October 1994, revised 6 June 1995. 1995. The Journal of Arachnology 23:199-201 RESEARCH NOTES REPORT ON A RARE DEVELOPMENTAL ANOMALY IN THE SCORPION, CENTRUROIDES VITTATUS (BUTHIDAE) While perusing scorpion samples to delineate the distribution of Centruroides vittatus (Say) in the central United States, we discovered a spec- imen in the collection of the Emerson Entomo- logical Museum of Oklahoma State University with two fully formed metasomas and telsons (Figs. 1, 2). This anomaly, involving complete duplication of the metasoma and telson, has been reported very rarely in the past. Although known in ancient times (see Vachon 1953), it was first described in the modem literature by Pavesi (1881) in the chactid, Euscorpius germanus (Koch). Duplicate metasomas and telsons have since been reported in E. carpathicus (Linnaeus) and in several buthids, including Buthacus lep- tochelys (Hemprich & Ehrenberg), Androctonus crassicauda (Olivier), Hottentotta (= Buthotus) alticola (Pocock), Centruroides infamatus (Koch), C. gracilis (Latreille), C. margaritatus (Gervais) and C. exilicauda (Wood) (= C. sculpturatus Ew- ing) (Berland 1913; Brauer 1917; Campos 1918; Franganillo 1934; Millot & Vachon 1949; Va- chon 1953; Williams 1971; Armas 1977). In a few cases (i. e., in E. germanus, B. leptochelys , and C. gracilis), the duplication involves part of the mesosoma as well, with the bifurcation aris- ing at the level of mesosomal segment III or IV. Brauer (1917) demonstrated that the condition results from splitting of the posterior part of the embryonic germ band. In his study of 5000 em- bryos of E. carpathicus, duplication-type anom- alies appeared in 13 specimens; of these, only one (or 0.02%) involved potential duplication of the metasoma. 1 2 Figures 1, 2. —Anomalous specimen of Centruroides vittatus. 1, Dorsal view; 2, Ventral view. Scale line = 10 mm. 199 200 THE JOURNAL OF ARACHNOLOGY Figures 3, 4.— Mesosomal segment VII and proximal metasomal segments of anomalous Centruroides vittatus. 3, Dorsal view; 4, Ventral view. The present specimen of C. vittatus , an im- mature female, most likely in the fifth instar, was collected by L. Feldick on 4 November 1988 at Kinta, Haskell County, Oklahoma. It is very sim- ilar to the specimens of A. crassicauda, H. alti- cola, and C. gracilis illustrated in Millot & Va- chon (1949), Vachon (1953), and Armas (1977), respectively. It also generally matches the spec- imen of C. exilicauda described by Williams (1971). The two metasomas and telsons are fully formed and the carination of the metasomal seg- ments is normal. Each metasoma bears an anus ventrally at the end of the fifth segment. It is quite probable that each metasoma and telson was fully functional. The seventh mesosomal segment is quite ab- normal, as would be expected in order for it to accommodate two metasomas. The tergite (Fig. 3) terminates posteriorly in a distinct triangular projection between the origins of the two me- tasomas. The lateral keels of the tergite are rel- atively normal, but the median keel is posteriorly bifurcate. The stemite is even more aberrant (Fig. 4) . Its posteromedial margin bears a narrow, deep, concave indentation, and there is an irregularly- shaped boss and ridge along the posterior mid- line. This structure is flanked laterally by four pairs of keels (normal specimens have only four keels total - two submedians and two laterals - on the same stemite). Five pairs of setae are sym- metrically placed on the stemite as shown in Fig. 4. We thank D. C. Arnold, curator of the Em- erson Entomological Museum, Oklahoma State University, for loan of the specimen and our wives for assistance with translating the litera- ture. The photographs were taken by D. J. Lyons of the NCSM Exhibits Department; we are grate- ful for his efforts. LITERATURE CITED Berland, L. 1913. Note sur un Scorpion muni de deux queues. Bull. Soc. Entomol. France, 18:251-252. Brauer, A. 1917. Ueber Doppelbildungen des Skor- pions Euscorpius carpathicus L. Sitz. Ak. Wiss. Ber- lin, 1917:208-221. Campos, F. 1918. Algunos casos teratologicos ob- servados en los Artropodos. Ann. Ent. Soc. Amer- ica, 1 1:97-98. de Armas, L. F. 1977. Anomalias en algunos Buth- idae (Scorpionida) de Cuba y Brasil. Poeyana, 176: 1-6. Franganillo, P. 1937. Un monstruo aracnologico. Mem. Soc. Cubano Nat. Hist., 1 1:55. Millot, J. & M. Vachon. 1949. Ordre des Scorpions. Pp. 386-436 In Traite de Zoologie (P.-P. Grasse, ed.), Masson et Cie, Paris, 6:1-979. Pavesi, P. 1881. Toradelfia in uno Scorpione. Rend. Inst. Lombardo, vol. 14. Vachon, M. 1953. The biology of scorpions. En- deavour, 12:80-87. Williams, S. C. 1971. Developmental anomalies in RESEARCH NOTES 201 the scorpion Centruroides sculpturatus (Scorpioni- da: Buthidae). Pan Pacific Entomol., 47:76-77. W. David Sissom: Department of Biology & Geosciences, WTAMU Box 808, West Texas A & M University, Canyon, Texas 790 1 6-000 1 USA Rowland M. Shelley: North Carolina State Museum of Natural Science, P.O. Box 29555, Raleigh, North Carolina 27626-0555 USA Manuscript received 15 February 1995, revised 22 April 1995. 1995. The Journal of Arachnology 23:202-204 DISPERSAL MECHANISMS OF STEGODYPHUS (ERESIDAE): DO THEY BALLOON? Silk is used by spiders in dispersal in different ways. One method, bridging, is to cast a line into the breeze and, when it catches on a distant ob- ject, to climb out on the line to its end. Bridging may also be accomplished by dropping on a line and swinging on it to reach a new site (Barth et al. 1991). A second method is to balloon: when the extruded thread and the spider get enough lift from updrafts (usually thermals or vertical wind-velocity gradients), the spider will be lifted off the substrate and carried through the air (De- cae 1986; Eberhard 1987; Greenstone 1990; Su- ter 1991, 1992; Weyman 1993; Follner & Klar- enberg 1995). Stegodyphus species (Eresidae) have been re- ported to disperse both by bridging and balloon- ing. Details of ballooning in Stegodyphus were reported by Wickler & Seibt (1986). They ob- served a gravid female S. mimosarum, one of three social species of Stegodyphus , ballooning over a distance of 1 8 m, having started at a height of 2 m. The spider had a body length of 10 mm with an estimated mass of 85-150 mg and was hanging from 3-4 silk threads of 60-80 cm length, i.e., the estimated total silk length was 1.8-3. 2 m. We compare the dispersal methods of bridging and ballooning for Stegodyphus in terms of their aerodynamics, consequences and relative im- portance. With this analysis we evaluate previ- ous interpretations of dispersal of Stegodyphus in relation to current knowledge of their ecology and population biology. We use evidence from the literature and our own general observations made during several years of fieldwork with four species of Stegodyphus , namely, the social S. dumicola Pocock 1898 and S. mimosarum Pa- vesi 1883, and subsocial S. lineatus Latreille 1817 and S. bicolor (O. Pickard-Cambridge 1869). We estimated the vertical air velocity required for Wickler & Seibt’s (1986) S. mimosarum to remain airborne, using Suter’s (1991) equation 7: Vsb = W -r (11.5 x L x W0-094 + 1.94 x W0-366) where Vsb = vertical air velocity in m/sec acting on the spider’s silk and body, W = weight in (1 « 0. 1 mg), L = silk length in m. To apply this equation, we assumed that the physical pa- rameters remain constant beyond the boundaries for which Suter’s equation was developed. By substituting the maximum silk length and min- imum S. mimosarum mass (above), a vertical velocity component of 9.2 m/s was necessary; minimum silk length and maximum spider mass requires 21.6 m/s. A wind of this strength would not have been described as a “barely perceptible breeze” (Wickler & Seibt 1986; p. 628), which may refer to a 0.1 -1.0 m/s wind. A horizontal wind of 3 m/s near the ground is the maximum at which thermals, suitable for ballooning, can be maintained (Greenstone 1 990; Weyman 1 993). For controlled ballooning, we assume a maxi- mum vertical wind component of 3 m/s, follow- ing Stull’s (1988) description of boundary layer meteorology. If Wickler & Seibt’s (1986) S. mi- mosarum experienced vertical winds of 0.1-3 m/s, it would have required 12-655 m of silk to become airborne. It is possible that the drag line was longer than reported because its distal end can be difficult to see (Eberhard 1 987). Furthermore, the above cal- culation does not take into account that spiders can change the drag on their bodies by several orders of magnitude when changing posture in relation to the direction of air flow (Suter 1992). Wickler & Seibt’s (1986) S. mimosarum was at least 3-6 times the maximum mass (25.5 mg) found in 2800 ballooning species investigated by Greenstone et al (1987). It clearly requires ex- treme conditions for this spider to experience lift. These conditions are well outside the boundaries used in aerodynamic models of spider ballooning (Humphrey 1987; Suter 1991, 1992). Ballooning is the domain of small spiders (Weyman 1993) and over 90% of them are < 1 mg in size (Green- stone et al. 1987). When members of large taxa, such as mygalomorphs, balloon, it is the small spiderlings (< 2.2 mg with an outlier of 5.8 mg) that do so (Coyle et al 1985). Social Stegodyphus are typically big at first dis- persal, as they usually disperse only when adult, or, more rarely, subadult. Dispersing female S. dumicola weigh 1 03-2 1 3 mg and males some 23- 48 mg (Henschel et al. 1995). By contrast, sub- 202 RESEARCH NOTES 203 social S. lineatus are 3-8 mg at first dispersal (Schneider 1992), within the size range of some heavy ballooning spiders. Nonetheless, S. linea- tus have not been recorded ballooning. On three occasions, one of the authors (JH) observed S. dumicola parachuting downwards when his presence disturbed spiders that were casting silken lines, evidently for bridging. The spiders landed several meters away on the ground. Size may explain why airborne Stegodyphus drop and land a short distance from the start. Perhaps other observations of ballooning by S. dumicola (in the laboratory, S. Kurpick pers. comm.) and by S. sarasinorum Karsch (Jambunathan 1905; Jacson & Joseph 1973) also occurred upon dis- turbance caused by the observers’ presence. If escape were not the case, the observation of sev- eral S. sarasinorum ballooning together on one gossamer (Jacson & Joseph 1973) would be puz- zling in view of the weight handicap this imposes on the conditions for remaining airborne. We have seen Stegodyphus casting bridging lines on numerous occasions over distances of several meters. The slightest vertical air move- ment is sufficient for silken threads to be airborne (Suter 1991). Thus, silken threads easily cross gaps, enabling even large spiders to move on them when the lines snag on objects. Bridging was used by over 88% of the dispersing S. dum- icola recorded by Henschel (pers. obs.). The adoption of the tiptoe posture and the location of nests on the windward side of trees were interpreted as indirect indications of aerial dispersal (Wickler & Seibt 1986; Seibt & Wickler 1988). However, we have observed Stegodyphus standing tiptoe when casting bridging lines and reaching windward destinations via these bridg- ing lines. This stance and location can therefore not serve as independent evidence of dispersal by ballooning. Both social and subsocial species of Stegody- phus appear to face high risks of predation during dispersal (Ward & Lubin 1993; Henschel pers. obs.). Ballooning must be regarded as a hazard- ous method of dispersal for such slow-moving spiders that are ungainly off the web. By bal- looning, they immediately forgo the possibility of backtracking to the safety of their colony of origin if the new site turns out to be unsuitable, especially if it is occupied by predators such as ants. For this reason, bridging is likely to be safer. Occasionally, social spiders are carried to new sites by agents beyond their control. Storms translocate spiders, large potential prey escape with spiders attached, mammals and birds pass through webs bearing spiders, and Gabar gos- hawks carry occupied spider nests onto their own nests (Seibt & Wickler 1988; Henschel et al. 1992a, b; Riechert & Roeloffs 1993). Social spi- ders may colonize new regions by such fortui- tous, hazardous translocations, but the signifi- cance at the population level is unknown. The extreme population subdivision indicated by protein allozyme electrophoresis (Smith & Engel 1994) constitutes additional evidence that Stegodyphus have poor powers of dispersal. Col- onies of S. sarasinorum in India had low fre- quencies of polymorphic loci with most of the variation occurring between different subpopu- lations. The known distances of dispersal of Ste- godyphus during their lifetime are short: 1-26 m for S. dumicola (Henschel pers. obs.); 1-83 m for S. lineatus (Ward & Lubin 1993). Most Ste- godyphus disperse no more than a few meters at a time (op. cit.; Schneider 1992). Furthermore, the patchy distribution patterns of webs suggests that most dispersal is over very short range (Wickler 1973; Seibt & Wickler 1988; Schneider 1992; Ward & Lubin 1993; Henschel pers. obs.). We conclude that there is insufficient evidence to support the impression expressed by authors citing the Wickler & Seibt (1986) report (e. g., Seibt & Wickler 1988; Riechert & Roeloff 1993; Smith & Engel 1994; Aviles 1995) that Stego- dyphus uses ballooning regularly and therefore is expected to have greater powers of dispersal than other social spiders. Current evidence indicates that Stegodyphus are conservative dispersers that normally employ bridging lines to move only short distances from their natal nest. Rarely, spi- ders may become airborne when they are dis- turbed and drop while casting bridging lines. They escape and land a short distance away, no further than when bridging. ACKNOWLEDGMENTS JH is a visiting scientist at the Blaustein In- ternational Center and a fellow of the Alexander- von-Humboldt Foundation, and JS is a VAT AT post-doctoral fellow at Ben Gurion University. Klaus Follner, Matthew Greenstone, Robert Su- ter and Mary Whitehouse commented on the manuscript. This is contribution No. 207 of the Mitrani Centre for Desert Ecology, Blaustein In- stitute for Desert Research, Sede Boqer, Israel. LITERATURE CITED Aviles, L. in press. Causes and consequences of co- operation and permanent-sociality in spiders: a re- 204 THE JOURNAL OF ARACHNOLOGY view of non-territorial permanent-social spiders. In Social competition and cooperation in insects and arachnids. Vol.2, Evolution of sociality. (B. Crespi & J. Choe, eds.). Princeton Univ. Press, Princeton. Barth, F. G., S. Komarek, J. A. C. Humphrey & B. Treidler. 1991. Drop and swing dispersal behavior of a tropical wandering spider: experiments and nu- merical model. J. Comp. Physiol A, 169:313-322. Coyle, F. A., M. H. Greenstone, A.-L. Hultsch & C. E. Morgan. 1985. Ballooning mygalomorphs: Es- timates of the masses of Sphodros and Unmidia bal- looners (Araneae: Atypidae, Ctenezidae). J. Arach- nol, 13:291-296. Decae, A. E. 1986. Dispersal: ballooning and other mechanisms. Pp. 348-356, In Ecophysiology of spi- ders. (W. Nentwig, ed.). Springer Verlag, Berlin. Eberhard, W. G. 1987. How spiders initiate airborne lines. J. Arachnol, 15:1-9. Follner, K. & A. J. Klarenberg. 1995. Aeronautic behaviour in the wasp-like spider, Argiope bruen- nichi (Scopoli) (Araneae, Argiopidae). Pp. 66-72, In Proc. 15th European Colloquium of Arachnol. (V. Ruzicka, ed.). Inst. Entomol., Ceske Budejovice. Greenstone, M. H. 1990. Meteorological determi- nants of spider ballooning: the roles of thermals vs. the vertical windspeed gradient in becoming air- borne. Oecologia, 84:164-168. Greenstone, M. H., C. E. Morgan & A.-L. Hultsch. 1987. Ballooning spiders in Missouri, USA, and New South Wales, Australia: family and mass dis- tributions. J. Arachnol, 15:163-170. Henschel, J. R. in press. Group-living reduces risk of predation in the spider Stegodyphus dumicola (Eresidae). Zool J. Linn. Soc. Henschel, J. R., Y. D. Lubin & J. Schneider. 1995. Sexual competition in an inbreeding social spider, Stegodyphus dumicola (Araneae: Eresidae). Insectes Soc., 42:419-426. Henschel, J. R., J. M. Mendelsohn & R. E. Simmons. 1992a. Is the association between the Gabar gos- hawk and social spiders Stegodyphus mutualism or theft? Gabar, 6:57-59. Henschel, J. R., R. E. Simmons & J. M. Mendelsohn. 1992b. Gabar goshawks and social spiders revis- ited: untangling the web. Gabar, 7:49-50. Humphrey, J. A. C. 1987. Fluid mechanic constraints on spider ballooning. Oecologia, 73:469-477. Jacson, C. C. & K. J. Joseph. 1973. Life-history, bi- onomics and behaviour of the social spider Stego- dyphus sarasinorum Karsch. Insect Soc., 189-204. Jambunathan, N. S. 1905. The habitats and life his- tory of a social spider {Stegodyphus sarasinorum Karsch). Smithson. Misc. Coll, Washington, 47, 2:265-372. Riechert, S. E. & R. M. Roeloffs. 1993. Evidence for and consequences of inbreeding in the cooperative spiders. Pp. 283-303, In Natural history of inbreed- ing and outbreeding. (N. Thornhill, ed.). Univ. Chi- cago Press, Chicago. Schneider, J. 1 992. Die Wurzeln des Soziallebens bei der subsozialen Spinne Stegodyphus lineatus (Er- esidae). Ph. D. dissertation, Ludwig-Maximillian- Universitat, Miinchen, Germany, 135 pp. Seibt, U. & W. Wickler. 1988. Bionomics and social structure of ’Family Spiders’ of the genus Stegody- phus, with special reference to the African species S. dumicola and S. mimosarum (Araneida, Eresi- dae). Verh. naturwiss. Ver. Hamburg, 30:255-303. Smith, D. R. & M. S. Engel. 1994. Population struc- ture in an Indian cooperative spider, Stegodyphus sarasinorum Karsch (Eresidae). J. Arachnol, 22:108- 113. Stull, R. B. 1988. An introduction to boundary layer meteorology. Kluwer Academic Publishers, Dor- drecht. Suter, R. B. 1991. Ballooning in spiders: results of wind tunnel experiments. Ethol, Ecol. & Evol, 3:13- 25. Suter, R. B. 1992. Ballooning: data from spiders in freefall indicate the importance of posture. J. Ar- achnol., 20:107-113. Ward, D. & Y. Lubin. 1993. Habitat selection and the life history of a desert spider, Stegodyphus li- neatus (Eresidae). J. Anim. Ecol, 62:353-363. Weyman, G. S. 1993. A review of the possible caus- ative factors and significance of ballooning in spi- ders. Ethol, Ecol & Evol., 5:279-291. Wickler, W. 1973. Uber Koloniegrundung und so- ziale Bindung von Stegodyphus mimosarum Pavesi und anderen sozialen Spinnen. Z. Tierpsychol, 32: 522-531. Wickler, W. & U. Seibt. 1986. Aerial dispersal by ballooning in adult Stegodyphus mimosarum. Na- turwiss., 73:628-629. Joh R. Henschel: Desert Ecological Research Unit of Namibia, P.O.Box 1592, Swakop- mund, Namibia Jutta Schneider: Max-Planck-Institut fur Ver- haltensphysiologie, Seewiesen, D-82319 Stamberg, Germany Yael D. Lubin: Mitrani Centre for Desert Ecol- ogy, Blaustein Institute for Desert Research, Ben Gurion University of the Negev, 84993 Sede Boqer Campus, Israel Manuscript received 3 April 1995, revised 21 July 1995. 1995. The Journal of Arachnology 23:205-206 DISPERSAL AGGREGATION OF SPHODROS FITCHI (ARANEAE, ATYPIDAE) Observations of dispersal aggregations of my- galomorph spiderlings have been rarely reported in the literature. Spiderlings of the European purseweb, Atypus qffinis Eichwald 1830 (Atypi- dae), were discovered climbing up vegetation on warm, spring days (Bristowe 1939). As several reached the top of a garden stake, the wind dis- lodged them and their draglines, which became attached to other objects. The young of another species of purseweb spider, Sphodros rufipes (La- treille 1829), performed similar preballooning behavior in the laboratory (Muma & Muma 1945). Coyle (1983) was fortunate in observing the dispersal of purseweb spiderlings in North Carolina. Sphodros fitchi Gertsch & Platnick 1980 is a purseweb spider found in the central plains states from Nebraska to Oklahoma and Arkansas (Gertsch & Platnick 1980). It appears to be the rarest of three species of Sphodros which occur in northeastern Kansas. Therefore, the natural history information of S. fitchi is limited to a few anecdotal observations (Fitch 1963; Guarisco 1988; Morrow 1985; Teeter 1984). This is the first report of a dispersal aggregation of spider- lings of this species. At noon on 2 April 1995, a dispersal aggre- gation of immature Sphodros fitchi was discov- ered by the first author in the highest branches of a 1 m tall eastern red cedar (Juniperus virgi- niana L.) on the west campus of the University of Kansas in Douglas County, Kansas. The small tree was located at the edge of a lawn and second growth woods composed predominantly of osage orange trees {Madura pomifera (Raf.) Schneid.) 3-5 m in height. A group of 1 4 immatures and two dense, silk mats, each 1 cm2 in area, were seen on the tips of two branches. During 30 min- utes of observation, the spiders slowly walked on the silk mats and the silk strands between the branches and the adjacent branch tips. Four ap- peared to let go, fell a few centimeters on drag- lines to lower branches, then slowly climbed back to the top. Ballooning behavior in Sphodros and other mygalomorphs consists of descending on a dragline until the force of the wind breaks it near its attachment point. Then the spider and dragline are carried by the breeze (Coyle 1983). Therefore, the four observations may represent unsuccessful ballooning attempts. The sky was partly cloudy, the temperature was about 23 °C, and there were westerly gusty winds. Later that day, we returned to the site, and the second author discovered a large, white silk tube attached to the base of the red cedar tree. No spiderlings were observed. The tube and the ce- dar trunk were each approximately 2.5 cm in diameter. The tube extended 1 7 cm up the side of the tree. The next day, the silk tube was ex- cavated and it contained an adult S. fitchi female. No egg-sac or egg remains were found inside the tube, which was 33.5 cm in total length. The length of the female, including chelicerae, was 2.7 cm. The lengths of five spiderlings from the dispersal aggregation, including chelicerae and spinnerets, ranged from 2.25-2.50 mm. The av- erage width of the prosoma at the anterior edge was 0.68 mm {n = 5, range 0.60-0.76). Two of the immatures were placed in vials containing moist soil and vertical twigs. Small silk tubes covered with soil particles were discovered the next day. One was located on the side of the vial and the second was located along a twig. The aerial portion of each tube was about 1 cm in length. The following day, the web along the twig was 2 cm in length. The dispersal aggregation described here re- sembles those described by other authors. Al- though no actual ballooning was witnessed, the behavior of four spiderlings was consistent with that described by Coyle (1983). The immatures were capable of independent living, based on the construction of their own webs in captivity. We would like to thank Dr. Hampton Shirer, Lawrence, Kansas for temperature information. LITERATURE CITED Bristowe, W. S. 1939. The Comity of Spiders. Vol. I. Ray Society, London. 228 pp. Coyle, F. A. 1983. Aerial dispersal by mygalomorph spiderlings (Araneae, Mygalomorphae). J. Arach- nol., 11:283-286. Fitch, H. S. 1 963. Spiders of the University of Kansas 205 206 THE JOURNAL OF ARACHNOLOGY Natural History Reservation and Rockefeller Ex- perimental Tract. Univ. Kansas Mus. Nat. Hist. Misc. Publ., 33:1-202. Gertsch, W. J. & N. I Platnick. 1980. A revision of the American spiders of the family Atypidae (Ara- neae, Mygalomorphae). American Mus. Nov., No. 2704. 39 pp. Guarisco, H. 1988. Predation of Achaearanea tepi- dariorum (Araneae, Theridiidae) upon Sphodros fit- chi (Araneae, Atypidae). J. ArachnoL, 16:390-391. Morrow, W. 1985. Two species of atypid spiders (Araneae, Atypidae) in eastern Kansas: male emer- gence times and notes on natural history. Master’s Thesis, Univ. Kansas. 47 pp. Muma, M. H. & K. E. Muma. 1945. Biological notes on Atypus bicolor Lucas (Arachnida). Entom. News, 56:122-126. Teeter, M. M. 1984. The role of slope orientation in nest-site selection by Sphodros spp. (Araneae, Atyp- idae): field and experimental observations. Master’s Thesis, Univ. Kansas. 40 pp. Bruce Cutler: Electron Microscopy Labora- tory, and Department of Entomology, Uni- versity of Kansas, Lawrence, Kansas 66045- 2106 USA. Hank Guarisco: Kansas Biological Survey, 204 1 Constant Avenue, Lawrence, Kansas 66047 USA. Manuscript received 21 June 1995, revised 13 August 1995. 1995. The Journal of Arachnology 23:207-208 PREDATION BY MISUMENOPS ASPERATUS (ARANEAE, THOMISIDAE) ON THE METALLIC PITCH NODULE MOTH, RETINIA METALLICA (LEPIDOPTERA, TORTRICIDAE) Forest entomologists have long suspected that spiders play important roles in the population dynamics of forest insects because of their pred- atory habits and abundance on trees (e. g., see Loughton et al 1963). Despite this potential im- portance, however, few observations of spiders actually feeding on tree pests have been reported. Some exceptions include spiders observed prey- ing on destructive bark beetles (Jennings & Pase 1975) and on forest-tree defoliators (Jennings & Houseweart 1989). Possible reasons for the scar- city of observed predatory bouts by spiders in forests and tree plantations include: (1) the di- minutive size of spiders compared to the tree; (2) the cryptic habits of some spiders, especially those that employ hunter-ambusher tactics; and (3) the low prey-capture success of some species (Jackson 1977). Hunting spiders are less apt to be observed with prey than web-spinners. Large orb weavers and other web-spinners that “store” prey in their webs offer an easier means of prey assessment. Nentwig (1987) noted that nonweb- building spiders handle only one prey at a time; consequently, their hunting success is relatively low, and ingestion time short. Hence, a low per- centage of hunting spiders are found with prey at any specific time in a population (Nentwig 1987). Here we describe predation by an ambushing crab spider on a destructive insect pest of pon- derosa pine, Pinus ponderosa (Laws.), in a shel- terbelt of Nebraska. This is the first recorded instance of spider predation on the metallic pitch nodule moth, Retinia metallica (Busck), in North America. Larvae of this moth bore into the new growth of pine stems, twigs, and branches (Fur- niss & Carolin 1977; Dix et al. 1986). During July, the larvae produce a nodule or lump of pitch and frass at the point of attack. Such feeding stunts tree growth and frequently kills the tips. Heavily infested trees have excessive branching. On 22 May 1987, the senior author observed a crab spider feeding on a small female moth near the apex of a ponderosa pine branch (1.2m high). The tree was approximately 5 m high and was growing in a multi-row farmstead shelterbelt (Hollst Farm) near Mead, Saunders County, Ne- braska (41°16'N, 96°28'W). The spider with cap- tured prey was collected, photographed (Fig. 1), and then preserved in 70% ethanol for later iden- tification. The crab spider, an adult female Misumenops asperatus (Hentz), was identified by the junior author. The specimen will be deposited in the arachnid collections of the U. S. National Mu- seum of Natural History, Washington, D. C. This species of crab spider hunts by stealth and ambush (Gertsch 1939). Branch apices are hunting sites where these crab spiders can wait for flying insects such as moths to land (pers. obs.). Because R. metallica moths frequent branch apices and similar microhabitats, they are sus- ceptible to predation by ambushing crab spiders like M. asperatus. However, the frequency and extent of predation by M. asperatus on R. me- tallica are unknown. This insect, and similar lep- idopterous species whose larvae live inside the twigs of trees, is most susceptible to predation by spiders during the moth-flight period. In the collection locale, the adult flight of R. metallica spans three weeks during May and early June (Dix unpubl. data). We suspect that hunting spiders are more suc- cessful at capturing small moths like R. metallica than are web-spinners. Moth scales provide a means of escape from spider webs (Eisner et al. 1964); however, such defenses are ineffectual against ambushing crab spiders. Juillet (1961) found that wandering spiders of the families Sal- ticidae and Thomisidae killed three times as many adults of the European pine shoot moth, Rhy- acionia buoliana (Schiff.), as did web spinners of the family Araneidae. Although M. asperatus is common on young ponderosa pines in Nebraska shelterbelts, the density of its populations on shelterbelt trees is unknown. It is found in both old field (e. g., Berry 1970) and arboreal habitats. In South Carolina, M. asperatus preyed on both larvae and adults of the Nantucket pine tip moth, Rhyacionia frus- trana (Comstock), another destructive insect pest of pine plantations (Eikenbary & Fox 1968). 207 208 THE JOURNAL OF ARACHNOLOGY Figure 1.— A female Misumenops asperatus feeding on the metallic pitch nodule moth, Retinia metallica, on a ponderosa pine tree in Nebraska. Ponderosa pine may gain some protection from the predatory habits of spiders like M. asperatus. Such mortality would be particularly important when spiders and other predators kill gravid moths of R. metallica. No doubt other spider species also capture and feed on R. metallica ; however, this potential source of moth mortality has not been fully investigated. Similar obser- vations of spider predation on other insect pests of ponderosa pine (e. g., scarab beetles (Jennings 1974), Southwestern pine tip moth (Jennings 1975; Lawson et al. 1983), and pine butterfly (Jennings & Toliver 1976)) support our conclu- sion. ACKNOWLEDGMENTS We thank Jon Keller and James Kalish for the photograph of Misumenops asperatus with Re- tinia metallica prey. We are grateful to Drs. Bruce Cutler, Richard R. Mason, and Gail E. Stratton for their constructive comments on an earlier draft of this research note. LITERATURE CITED Berry, J. W. 1970. Spiders of the North Carolina Piedmont old-field communities. J. Elisha Mitchell Sci. Soc., 86:97-105. Dix, M. E., J. E. Pasek, M. O. Harrell, & F. P. Bax- endale. 1986. Common insect pests of trees in the Great Plains. Nebraska Coop. Ext. Serv. EC 86= 1548. 44 pp. Eikenbary, R. D. & R. C. Fox. 1 968. Arthropod pred- ators of the Nantucket pine tip moth, Rhyacionia frustrana. Annl. Entomol. Soc. America, 61:1218- 1221. Eisner, T., R. Alsop & G. Ettershank. 1964. Adhe- siveness of spider silk. Science, 146:1058-1061. Fumiss, R. L. & V. M. Carolin. 1 977. Western Forest Insects. U. S. Dep. Agr., For. Serv., Misc. Publ. No. 1339. U. S. Govt. Printing Office, Washington, D. C. 654 pp. Gertsch, W. J. 1939. A revision of the typical crab- spiders (Misumeninae) of America north of Mexico. Bull. American Mus. Nat. Hist., 76:277-442. Jackson, R. R. 1 977. Prey of the jumping spider Phi- dippus johnsoni (Araneae: Salticidae). J. Arachnol., 5:145-149. Jennings, D. T. 1974. Crab spiders (Araneae: Thom- isidae) preying on scarab beetles (Coleoptera: Scar- abaeidae). Coleopterists Bull., 28:41-43. Jennings, D. T. 1975. Life history and habits of the southwestern pine tip moth, Rhyacionia neomexi- cana (Dyar) (Lepidoptera: Olethreutidae). Annl. En- tomol. Soc. America, 68:597-606. Jennings, D. T. & M. W. Houseweart. 1989. Sex- biased predation by web-spinning spiders (Araneae) on spruce budworm moths. J. Arachnol., 17:17 9— 194. Jennings, D. T. & H. A. Pase III. 1975. Spiders prey- ing on Ips bark beetles. Southwestern Nat., 20:225- 229. Jennings, D. T. & M. E. Toliver. 1976. Crab spider preys on Neophasia menapia (Pieridae). J. Lepi- dopterists’ Soc., 30:236-237. Juillet, J. A. 1961. Observations on arthropod pred- ators of the European pine shoot moth, Rhyacionia buoliana (Schiff.) (Lepidoptera: Olethreutidae), in Ontario. Canadian Entomol., 93:195-198. Lawson, H. R., L. A. Yost & D. T. Jennings. 1983. Southwestern pine tip moth: notes on larval descent behavior, predators, and associated shoot borer in northern Arizona. Southwestern Nat., 28:95-97. Loughton, B. G., C. Derry & A. S. West. 1963. Spi- ders and the spruce budworm. Pp. 249-268. In The dynamics of epidemic spruce budworm popula- tions. (R. F. Morris, ed.). Mem. Entomol. Soc. Can- ada, 31. 332 pp. Nentwig, W. 1 987. The prey of spiders. Pp. 249-263, In Ecophysiology of Spiders. (W. Nentwig, ed.). Springer- Verlag, Berlin. 448 pp. Mary Ellen Dix: Rocky Mountain Forest and Range Experriment Station, National Agro- forestry Center, East Campus, University of Nebraska, Lincoln, Nebraska 68583-0822 USA Daniel T. Jennings: Northeastern Forest Ex- periment Station, 180 Canfield Street, Mor- gantown, West Virginia 26505 USA Manuscript received 21 April 1995, revised 7 July 1995. 1995. The Journal of Arachnology 23:209-21 1 ON MALES OF CALIFORNIAN TALANITES (ARANEAE, GNAPHOSIDAE) Shortly after the revision of the North Amer- ican species of Talanites (Platnick & Shadab 1 976, as Rachodrassus ), we discovered two additional species in California. These were recently de- scribed by Platnick & Ovtsharenko (1991), as T. moodyae and ubicki, on the basis of females. Here we report the newly discovered males. One noteworthy observation on these speci- mens is the presence of a ventroapically pro- jecting process at the base of the embolus (Figs. 2, 4). This process is not recorded for the species described in the above papers, and we were un- able to observe it in specimens of T. echinus (Chamberlin) or exlineae (Platnick & Shadab). The Californian Talanites have already been shown to be somatically distinct in having un- usually small eyes (Platnick & Ovtsharenko 1991). Perhaps these character states are derived. However, given that the Californian fauna is rife with relictual arachnids, plesiomorphy cannot be completely ruled out. In fact, the two species have a typically relictual distribution pattern and, despite an intense collecting effort by the authors, are known from only a few localities. (Given the rarity of the species, we list below all of their known localities.) The Talanites species from the eastern United States and Mexico, in contrast, are each much more widely distributed, showing a combined range from San Luis Potosi and Tex- as to Florida, north to Arkansas and Virginia. Following the format in the above papers we describe here the males of the two species. The unique male specimens have been deposited at the California Academy of Sciences (CAS), other Figures 1-2. — Talanites moodyae Platnick & Ovtsharenko. 1 , Left male palp, ventral view. 2, Same, retrolateral view, (setae omitted) 209 210 THE JOURNAL OF ARACHNOLOGY Figures 3-4 .—Talanites ubicki Platnick & Ovtsharenko. 3, Left male palp, ventral view. 4, Same, retrolateral view, (setae omitted) specimens are in the collections of the American Museum of Natural History (AMNH), Califor- nia Department of Food and Agriculture (CDFA), and D. Ubick (CDU). We thank Charles E. Gris- wold of the California Academy of Sciences and Norman I. Platnick of the American Museum of Natural History for reading and commenting upon an earlier draft of this paper. Special thanks go to Joy Boutin and Bill Tyson for helping col- lect specimens of Talanites ; Joy Boutin also as- sisted in rearing the males of T. ubicki. We thank the California Department of Food and Agri- culture for helping defray the publication costs. All measurements are in mm. Talanites moodyae Platnick & Ovtsharenko Figs. 1,2 Talanites moodyae Platnick & Ovtsharenko 1991: 119, figs. 15, 16 (female holotype in AMNH, not exam- ined). Diagnosis. —Males are distinguished from all other North American Talanites by the com- bined presence of a ventroapically directed pro- cess on the embolar base, a small second point on the median apophysis, and reduced eyes. Description. —Male: Total length 7.1. Cara- pace length 3.4; width 2.7. Femur II length 2.6. Eye sizes and interdistances: AME 0.08, ALE 0.09, PME 0.05 x 0.07, PLE 0.06 x 0.08, AME- AME 0.07, AME-ALE 0.08, PME-PME 0.15, PME-PLE 0.15, ALE-PLE 0.08, MOQ length 0.22, front width 0.21, back width 0.26. Leg spi- nation: femora: I,IIp0- 1 -2, d 1 - 1 - 1 , r 1 - 1 - 1 ; IIIpO- 1-1, dl-1-1, rl-1-1; IVpl-1-1, dl-1-1, rl-1-1; tibiae: Ipl-1-1, v2-2-2, rl-1-2; IIpl-1-1, v2-2-2, r2-l-l; metatarsi: IIIp2-2-2, v2-2-2, rl-1-1. Pal- pal tibial apophysis flattened, directed retrola- terally; embolus distally pointed with ventroap- ically directed basal process; median apophysis with small, basal second point. Female: Described by Platnick & Ovtsharenko (1991). Material examined.— USA. California: Fres- no County: Granite Hill, 1.5 mi NE Navelencia, 1 February 1994 (W. H. Tyson, CDFA), 1 ju- venile. E slope Smith Mountain, 4 mi E Reedley, UBICK & MOODY— CALIFORNIAN TALANITES 211 grassland, under granite, 29 March 1991 (D. Ubick, CDU), 5 juveniles, 19 January 1994 (W, H. Tyson, CDFA), 2 females. Tulare County: S slope Smith Mountain, under serpentine, 1 9-2 1 January 1994 (W. H. Tyson, CDFA), 5 females, 19 juveniles. Bacon Hill, 5 February 1992 (M. J. Moody, CAS), 1 male. Twin Buttes, 2 Decem- ber 1993 (W. H. Tyson, CDFA), 1 juvenile. Ven- ice Hills, 12 March 1991, under rocks (W. H. Tyson, CDFA), 1 juvenile. Rocky Hill, E of Ex- eter on Hwy 130, grassland, under serpentine, 26 January 1991 (D. Ubick, CDU), 1 penulti- mate female. Notes.— Three penultimate males were placed in a 10 x 8.5 cm plastic jar containing soil to a depth of 3 cm. The soil was maintained moist and a piece of cardboard was placed to provide cover. The spiders were fed termites; two of the spiders died, the third matured by 3 June 1992. Talanites ubicki Platnick & Ovtsharenko Figs. 3, 4 Talanites ubicki Platnick & Ovtsharenko 1991: 120, figs. 3, 4 (female holotype in AMNH, not examined). Diagnosis.— Males are distinguished from all other North American Talanites by the com- bined presence of a ventroapically directed pro- cess on the embolar base, a large second point on the median apophysis, and reduced eyes. Description. —Male: Total length 4.6. Cara- pace length 2.1; width 1.7. Femur II length 1.6. Eye sizes and interdistances: AME 0.05, ALE 0.09, PME 0.05 x 0.07, PLE 0.05 x 0.08, AME- AME 0.05, AME-ALE 0.04, PME-PME 0.08, PME-PLE 0.08, ALE-PLE 0.04, MOQ length 0.17, front width 0.15, back width 0.21. Leg spi- nation: femora: Ip0-l-2, dl-1-1, r0-l-0; Ilp0-1- I, dl-1-1, r0-l-0; III,IVpl-l-l, dl-1-1, r0-l-l; tibiae: I, Up 1 - 1 - 1 , v2-2-2, r 1 11; metatarsi: IIIp 1 - 1 - 1 , v 2-2-2, r 1 - 1 - 1 . Palpal tibia! apophysis pointed, directed retrodorsally; embolus distally broad with ventroapically directed basal process; median apophysis with large second point. Female: Described by Platnick & Ovtsharenko (1991). Material examined.— USA. California: Marin County: Novato, San Marin Drive, 1 8 April 1 992 (D. Ubick & I. Boutin, CAS), 1 male. December to March 1982-1992 (D. Ubick, CDU), females and juveniles. Notes. —The male specimen was one of four penultimate males collected at the type locality, a serpentine grassland on the SW slope of Burdell Mm, just north of Simmons Lane, Novato. (The correct spelling of the type locality is Novato, not 44Novata” as given by Platnick & Ovtshar- enko 1 99 1 .) The Talanites were found under large serpentine floats at the edge of a seepage where the soil conditions were moist. In the lab, the spiders were placed in terraria on moist serpen- tine soil clumps. A variety of insect prey was introduced during the captivity. The spiders were reluctant feeders but were observed feeding on embiids, termites, and Drosophila . Of the four spiders, two died without molting, one died dur- ing molting in late May, and one molted suc- cessfully on 30 May 1992. LITERATURE CITED Platnick, N. I. & M. Shadab. 1976. A revision of the spider genera Rachodrassus, Sosticus, and Scopodes (Araneae, Gnaphosidae) in North America. Amer- ican Mus. Novit, 2594:1-33. Platnick, N. I. & V. I. Ovtsharenko. 1991. On Eur- asian and American Talanites (Araneae, Gnaphos- idae). J. ArachnoL, 19:1 15-121. Darrell Ubick: Associate, Department of En- tomology, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 USA. Marjorie J. Moody: Entomologist, California Department of Food and Agriculture, P. G. Box 3468, Visalia, California 93278 USA. Manuscript received 19 April 1995, revised 21 July 1995 . 1995. The Journal of Arachnology 23:212-214 CHEMICAL ATTRACTION OF MALE CRAB SPIDERS (ARANEAE, THOMISIDAE) AND KLEPTOPARASITIC FLIES (DIPTERA, MILICHIIDAE AND CHLOROPIDAE) After the first day of a study testing the at- traction of scavenging flies (Diptera, Milichiidae and Chloropidae) to defensive chemicals of true bugs (Heteroptera), it was apparent that males of one type of crab spider (Thomisidae) were also attracted to the chemical treatments. Therefore, the original goal of the study was abandoned in 1993 in favor of a full-time investigation of spi- der attraction. Traps were constructed of transparent cylin- drical containers (20.2 x 19.7 cm; Tri-State Molded Plastic, Dixson, Kentucky) by cutting two 9 -cm diameter holes in opposite sides, and covering each hole with an inwardly projecting screen funnel (Aldrich et al. 1984). On 8 June 1993, nine traps were hung from stakes in a 0.9 ha fallow field 1 0 m apart in contact with foliage (mixed grasses, goldenrod, and bush-clover), al- ternating the following three treatments: (E)- 2- octenal, (Ts)-2-decenal, and unbaited controls. Chemicals were dispensed from cotton swabs (Q- Tips®, Chesebrough Pond’s Co., Greenwich, Connecticut) dipped in neat standards (ca. 200 n\; Bedoukian Research, Inc., Danbury, Con- necticut). Traps were inspected every 1-2 days, and rebaited every 2-3 days. From 9-18 June, 200 males of Xysticus ferox (Hentz) (Thomisidae) were caught in traps baited with alkenals; none were captured in controls (Table 1, Field 1). Capture of X. ferox males was variable for both (^j-2-octenal-baited traps (to- tals of 28, 33, 59 males/trap) and (^-2-decenal- baited traps (7, 28, and 45 males/trap) such that attraction to -2-octenal was not significantly greater than to (E)- 2-decenal (P = 0.33). From 18-20 June 1993, captures of Xysticus males dropped to almost zero in Field 1 ; there- fore, on 21 June the traps were redeployed in another field (0.5 ha) that had been lightly sown with a mixed ground cover including red clover and vetch (Table 1, Field 2). Vegetation in Field 2 was sparse, so traps were placed directly on the ground. From 22 June-15 July, a total of 74 males of four Xysticus spp. was caught in traps baited with (E)- 2-octenal or (^-2-decenal: X. ferox, X. discursans Keyserling, X. triguttatus Keyserling, and X. auctificus Keyserling. No Xys- ticus females were caught, and one Xysticus in- dividual was caught in a control trap. Xysticus spp. from Field 2 were grouped because we could not reliably separate the species; however, 1 5 out of the 25 specimens determined were X. aucti- ficus. Again, the attraction to (E)- 2-octenal was greater, but not significantly so, than to (E)-2- decenal (Table 1, P = 0.35). In order to determine if other compounds were involved in spider attraction or if the known attractants acted synergistically, additional sets of traps were deployed in Field 2 baited in a similar manner with (Ty-2-octenoic acid, octan- oic acid, 1-octanol, f'£'y)-2-octenal/(^E)-2-decenal (1:1 blend), (E)~ 2-nonenal, (Ej-2-decenal ace- tate, and (E)- 2-hexenal butyrate (Aldrich Chem- ical Co., Milwaukee, Wisconsin; or Bedoukian Research, Inc.). Standards of octenal and decenal contained impurities of the corresponding acids (1.48% (£y-2-octenoic acid; 1.58% (2^-2-decen- oic acid; analyzed by standard methods, e. g., Aldrich et al. 1984), and the acids predominated in extracts of Q-Tips after 24 h field exposure in hot, sunny weather (88% and 65% (TT)-2-octenoic acid and (Ts)-2-decenoic acid, respectively). Nev- ertheless, there was no indication that (E)- 2-oc- tenoic acid or octanoic acid are attractive to Xys- ticus species. Octanol was a common minor im- purity (< 1%) in both octenal and decenal stan- dards, but it seemed inactive. Similarly, there was no indication of synergism between octenal and decenal, no evidence that (Ty-2-nonenal is attractive, and common esters of stink bugs, (E)- 2-decenyl acetate, and plant bugs (Miridae), (E)- 2-hexenyl butyrate, appeared inactive. In 1 994, traps were deployed earlier in Field 1 (14 May-30 June), and an additional set of traps was baited with another alkenal commonly produced by heteropterans, (Ts) -2-hexenal (Be- doukian Research, Inc.) (Aldrich 1988). Xysticus individuals were caught from the first day of the experiment in numbers greater than the previous year (Table 1 ). The results for Xysticus (identified to genus only) corroborated previous results for male-specific attraction to (E)- 2-octenal and (E)~ 2-decenal, but (TTj-2-hexenal was not attractive. In 1994 we also decided to collect the trapped 212 RESEARCH NOTES 213 Table 1. —Total numbers of Xysticus species males and females caught in traps in 1993 and 1994. Within a column, sums followed by the same letter are not significantly different at the 5% level (three traps/treatment; one-way ANOVA of rank transformed sums/trap/trapping period, fit separately for each field). Treatment 1993 1994 Field 1 Field 2 Field 1 Male Female Male Female Male Female (2s)-2-Hexenal _ __ — _ la 3a (£>2-Octenal 120a 0a 52a 0a 224b 0a (ii)-2-Decenai 80a 0a 22a 0a 115b la Control 0b 0a lb 0a 0a 0a milichiid and chloropid flies, several of which are kleptoparasitic on true bugs caught in spider webs (Eisner et al 1991; Landau & Gaylor 1987). Exclusively females of one milichiid, Milichiella arcuata (Loew), were significantly attracted to ®-2-hexenal-baited traps, but not to the other alkenal-baited or control traps (Table 2). Fifteen chloropid species were caught, totalling 2269 in- dividuals (predominantly females), but Ocella trigramma (Loew) accounted for over 95% of the total, with O. cinerea (Loew) and O parva (Ad- ams) comprising about 1%. Chloropids were at- tracted to all three alkenals compared to control traps; however, f7s)-2-octenal was most attrac- tive, followed by f£/'-2»decenai, and (E)- 2-hex- enal was the least attractive (Table 2). These data suggest that scavenging milichiid and chloropid flies use allomones from dying bugs in spider webs (and probably elsewhere), not just to find food, but also to discriminate between heterop- teran species (see also Sugawara & Muto 1974). The surprising discovery that male Xysticus species are attracted to ®-2-octenal and (E)~ 2- decenal is difficult to explain. Xysticus species chemically attracted to alkenals are brown, ground-dwelling spiders that probably seize most of their prey after laying-in-wait in the litter zone (Morse 1983). A variety of heteropterans are among the natural prey of litter-inhabiting crab spiders (Nyffeler & Breene 1990; Nentwig 1986; Araya & Haws 1 988); therefore, it is possible that Xysticus use alkenals as host-finding kairomones as do scavenging flies. This seems unlikely, though, because Heteroptera constitute only a small portion of the prey taken by ground-dwell- ing Xysticus species (Nyffeler & Breene 1990; Nentwig 1986), and esters of Heteroptera were unattractive. Most importantly, only four fe- males and no immatures were caught in chem- ically baited traps, compared to 615 adult Xys- ticus males. Behavioral studies have shown that both web- building and hunting spiders communicate with pheromones (Barth 1993; Pollard et al 1987; Rovner 1991; Suter & Hirscheimer 1986; Tietjen 1979). To date, there has been only one spider sex pheromone chemically identified: unmated females of the sheet- web weaving spider, Liny- phia triangularis (Clerck) (Linyphiidae), deposit the dimer of (R) • 3 - h ydroxybu ty ric acid on their webs which, after breaking down to the more volatile monomer, elicits the web reduction be- havior of males leading to copulation (Schulz & Toft 1993). Discovery of an acidic pheromone for a spider suggested that acidic impurities in alkenal standards might be responsible for at- traction of Xysticus males. Nonetheless, traps baited with high and low (10 id) doses of (E)~ 2- octenoic acid were unattractive in the field. In summary, the exact role of (Js)-2-decenal and (2s)-2-octenal in attraction of Xysticus males is not yet clear, but our results suggest that the alkenals, or impurities in the synthetic standards, are related or identical to sex pheromone com- ponents of these spiders. This is apparently the Table 2. —Milichiid and chloropid flies caught in traps in 1994. Within a column, sums followed by the same letter are not significantly different at the 5% level (three traps/treatment; one-way ANOVA of rank trans- formed sums/trap/trapping period). Treatment Milichiidae Chloropidae (L’)-2-Hexenal 136a 33a (Js)-2-OctenaI lb 2069b (L) -2 .Decenal 7b 167c Control 0b Od 214 THE JOURNAL OF ARACHNOLOGY first report of spiders being attracted into traps baited with synthetic chemicals. ACKNOWLEDGMENTS We thank Mr. Gustavo Hormiga of the Mary- land Center for Systematic Entomology (Uni- versity of Maryland and Smithsonian Institu- tion) for identifying the spiders and reviewing the manuscript, and Dr. C. W. Sabrosky (USDA- ARS, Systematic Entomology Laboratory) for determinations of Diptera. We are also grateful to Dr. Douglass H. Morse, Division of Biology and Medicine, Brown University, for a helpful discussion of thomisid biology and for reviewing an early draft of the manuscript. Mention of a company name does not imply endorsement by the U. S. Department of Agriculture. LITERATURE CITED Aldrich, J. R. 1988. Chemical ecology of the Het- eroptera. Ann. Rev. Entomol., 33:211-238. Aldrich, J. R., J. P. Kochansky & C. B. Abrams. 1984. Attractant for a beneficial insect and its parasitoids: Pheromone of the predatory spined soldier bug, Podisus maculiventris (Hemiptera: Pentatomidae). Environ. Entomol., 13:1031-1036. Araya, J. E. & B. A. Haws. 1988. Arthropod pre- dation of black grass bugs (Hemiptera: Miridae) in Utah ranges. J. Range Management, 41:100-103. Barth, F. G. 1993. Sensory guidance in spider pre- copulatory behaviour. Comp. Biochem. Physiol, 104A:717-733. Eisner, T., M. Eisner & M. Deyrup. 1991. Chemical attraction of kleptoparasitic flies to heteropteran in- sects caught by orb- weaving spiders. Proc. Natl. Acad. Sci., 88:8194-8197. Landau, G. D. & M. J. Gaylor. 1987. Observations on commensal Diptera (Milichiidae and Chloropi- dae) associated with spiders in Alabama. J. Arach- nol, 15:270-272. Morse, D. H. 1983. Foraging patterns and time bud- gets of the crab spiders Xysticus emertoni Keyserling and Misumena vatia (Clerck) (Araneae: Thomisi- dae) on flowers. J. Arachnol., 1 1:87-94. Nentwig, W. 1986. Non-webbuilding spiders: Prey specialists or generalists? Oecologia, 69:571-576. Nyffeler, M. & R. G. Breene. 1 990. Spiders associated with selected European hay meadows, and the ef- fects of habitat disturbance, with the predation ecol- ogy of the crab spiders, Xysticus spp. (Araneae, Thomisidae). J. AppL Entomol, 110:149-159. Pollard, S. D., A. M. MacNab & R. R. Jackson. 1987. Communication with chemicals: Pheromones and spiders, Pp. 133-141, In Ecophysiology of Spiders (W. Nentwig, ed.). Springer, New York. Rovner, J. S. 1991. Turning behaviour during pher- omone-stimulated courtship in wolf spiders. Anim. Behav., 42:1015-1016. Schulz, S. & S. Toft. 1993. Identification of a sex pheromone from a spider. Science, 260:1635-1637. Sugawara, R. & T. Muto. 1974. Attraction of several dipterous insects to aliphatic esters (Diptera: Mili- chiidae, Chloropidae and Ceratopogonidae). Appl. Entomol. Zool, 9:11-18. Suter, R. G. & A. J. Hirscheimer. 1986. Multiple web-borne pheromones in a spider Frontinella pyr- amitela (Araneae: Linyphiidae). Anim. Behav., 34: 748-753. Tietjen, W. J. 1979. Tests for olfactory communi- cation in four species of wolf spiders (Araneae, Ly- cosidae). J. Arachnol., 6:197-206. Jeffrey R. Aldrich: Insect Chemical Ecology Laboratory, USDA-ARS, Bldg 007, Agricul- tural Research Center-West, Beltsville, Mary- land 20705 USA Tev M. Barros: Department of Entomology, University of Maryland, College Park, Mary- land 20742 USA. Manuscript received 13 February 1995, revised 21 July 1995. INSTRUCTIONS TO AUTHORS (revised July 1995) Manuscripts are preferred in English but may be ac- cepted in Spanish, French or Portuguese subject to availability of appropriate reviewers. Authors whose pri- mary language is not English may consult the Associate Editor for assistance in obtaining help with English manuscript preparation. All manuscripts should be pre- pared in general accordance with the current edition of the Council of Biological Editors Style Manual unless instructed otherwise below. Authors are advised to con- sult a recent issue of the Journal of Arachnology for additional points of style. Manuscripts longer than 1 500 words should be prepared as Feature Articles, shorter papers as Research Notes. Send four identical copies of the typed material together with copies of illustrations to the Associate Editor of the Journal of Arachnology: Petra Sierwald, Associate Editor Division of Insects Dept, of Zoology Field Museum Roosevelt Road at Lakeshore Drive Chicago, IL 60605 USA [Telephone: (312) 922-9410, ext. 841; FAX: (312) 663-5397; Electronic mail: SIERWALD@FMNH.ORG]. Correspondence relating to manuscripts should also be directed to the Associate Editor. After the manu- script has been accepted, the author will be asked to submit the manuscript on a computer disc, preferably in MS DOS WordPerfect. FEATURE ARTICLES Title page. — The title page will include the com- plete name, address, and telephone number of the au- thor with whom proofs and correspondence should be exchanged, a FAX number 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 capital letters should be placed at the beginning of the first paragraph set off by a period. In articles written in English, a second abstract in an acceptable language may be included pertinent to the nationality of the author(s) or geo- graphic region(s) emphasized in the article. A second abstract in English must be included in articles not written in the latter language. 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 head 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 underlined and sep- arated 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 period regardless of language of the text. Citation of references in the text: Cite only papers already published or in press. Include within parenthe- ses the surname of the author followed by the date of publication. A comma separates multiple citations by the same author(s) and a semicolon separates citations by different authors, e.g., (Smith 1970), (Jones 1988; Smith 1993), (Smith 1986, 1987; Smith & Jones 1989; Jones et al. 1990). Literature cited section. — Use the following style: Lombardi, S. J. & D. L. Kaplan. 1990. The amino acid composition of major ampullate gland silk (dragline) of Nephila clavipes (Araneae, Tetragnath- idae). J. Arachnol., 18:297-306. Krafft, B. 1982. The significance and complexity of communication in spiders. Pp. 15-66, In Spider Communications: Mechanisms and Ecological Sig- nificance. (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. Running head. — The author surname(s) and an ab- breviated 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 con- tact the Associate Editor. Tables. — Each table, with the legend above, should be placed on a separate manuscript page. Only hori- zontal lines (usually three) should be included. Use no footnotes; instead, include all information in the leg- end. Make notations in the text margins to indicate the preferred location of tables in the printed text. Illustrations. — Address all questions concerning il- lustrations to the Editor of the Journal of Arachnology: James W. Berry, Editor Dept, of Biological Sciences Butler University Indianapolis, Indiana 46208 USA [Telephone (317) 940-9344; FAX (317)-940-9519; Electronic mail: BERRY@BUTLER.EDU] All art work must be camera-ready for reproduction. In line drawings, pay particular attention to width of lines and size of lettering when reductions are to be made by the printer. Multiple photos assembled on a single plate should be mounted with only a minimum of space separating them. In the case of multiple illus- trations 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. The author should indicate whether the illustrations should be one column or two columns in width. The overall dimen- sions should be no more than 11 inches (28 cm) X 14 inches (36 cm). Larger drawings present greater diffi- culty in shipping and greater risks of damage for which the JOA assumes no responsibility. Make notations in the text margins to indicate the preferred position of illustrations in the printed text. 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 -us, 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, holotype male; 33, 34, A-us y-us, male. Scale = 1.0 mm. Assemble manuscript for mailing. — Assemble the separate sections or pages in the following sequence: title page, abstract, text, figure legends, footnotes, ta- bles with legends, figures. Page charges and reprints. — The current charge per journal page will be assessed as follows: Manu- scripts in the appropriate word processing program $ 10/page; others $35/page. Reprints are available only from the printer and should be ordered when the author receives the galley proof pages. Waiver of page charg- es: In exceptional cases, the Editor will consider re- quests for a waiver of page charges for manuscripts which have been accepted for publication. Requests should be made in writing to the Editor. The author must certify that neither grant nor institutional funds are available to cover all of the publication costs. The Society will normally waive no more than 80% of the page charges. RESEARCH NOTES Instructions above pertaining to feature articles ap- ply also to research notes, except that abstracts and most headings are not used and the author’s name and address follow the Literature Cited section. CONTENTS The Journal of Arachnology Volume 23 Feature Articles Number 3 The Web-Spider Community of Soybean Agroecosystems in Southwestern Ohio by Ann L. Rypstra and Paul E. Carter 135 Mechanisms of the Formation of Territorial Aggregations of the Burrowing Wolf Spider Geolycosa xera archboldi McCrone (Araneae, Lycosidae) by Samuel D. Marshall 145 A Comparison of Populations of Wolf Spiders (Araneae, Lycosidae) on Two Different Substrates in Southern Florida by David B. Richman, Jan S. Meister, Willard H. Whitcomb and Leigh Murray 151 Observations on the Natural History of an Ummidia Trapdoor Spider from Costa Rica (Araneae, Ctenizidae) by Jason E. Bond and Frederick A. Coyle 157 Chivalry in Pholcid Spiders Revisited by Julie A. Blanchong, Michael S. Summerfield, Mary Ann Popson and Elizabeth M. Jakob 165 Generic Placement of the Empire Cave Pseudoscorpion, Microcreagris imperialis (Neobisiidae), a Potentially Endangered Arachnid by William B. Muchmore and James C. Cokendolpher 171 Phylogeny and Historical Biogeography of the Spider Genus Lutica (Araneae, Zodariidae) by Martin R. Ramirez and Richard D. Beckwitt 177 Redescription of Stenostygnus pusio Simon and Synonymy of Caribbiantinae with Stenostygninae (Opiliones: Laniatores, Biantidae) by Ricardo Pinto-da-Rocha 194 Research Notes Report on a Rare Developmental Anomaly in the Scorpion, Centruroides vittatus (Buthidae) by W. David Sissom and Rowland M. Shelley 199 Dispersal Mechanisms of Stegodyphus (Eresidae): Do They Balloon? by Joh R. Henschel, Jutta Schneider and Yael D. Lubin 202 Dispersal Aggregation of Sphodros fitchi (Araneae, Atypidae) by Bruce Cutler and Hank Guarisco 205 Predation by Misumenops asperatus (Araneae, Thomisidae) on the Metallic Pitch Nodule Moth, Retinia metallica (Lepidoptera, Tortricidae) by Mary Ellen Dix and Daniel T. Jennings 207 On Males of Californian Talanites (Araneae, Gnaphosidae) by Darrell Ubick and Majorie J. Moody 209 Chemical Attraction of Male Crab Spiders (Araneae, Thomisidae) and Kleptoparasitic Flies (Diptera, Milichiidae and Chloropidae) by Jeffrey R. Aldrich and Tev M. Barros 212