i ! &.J- The Journal of ARACHNOLOGY OFFICIAL ORGAN OF THE AMERICAN ARACHNOLOGICAL SOCIETY VOLUME 28 2000 NUMBER 1 THE JOURNAL OF ARACHNOLOGY EDITOR-IN-CHIEF: James W. Berry, Butler University MANAGING EDITOR: Petra Sierwald, Field Museum ASSOCIATE EDITORS: Matthew Greenstone, USDA; Robert Suter, Vassar College 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. Argentine de Ciencias Naturales; 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.; N. I. Platnick, American Mus. Natural Hist.; S. E. Riechert, Univ. Tennessee; A. L. Rypstra, Miami Univ, Ohio; M. H. Robinson, US. 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: Frederick A. Coyle (1999-2001), Department of Biology, Western Carolina University, Cullowhee, North Carolina 28723 USA. PRESIDENT-ELECT: Brent D. Opell ( 1 999-200 1 ), Department of Biology, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061 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, Department of Biology, University of Missis- sippi, University, Mississippi 38677 USA. BUSINESS MANAGER: Robert Suter, Dept, of Biology, Vassar College, Pough- keepsie, New York 12601 USA. SECRETARY: Alan Cady, Dept, of Zoology, Miami Univ, Middletown, Ohio 45042 USA. ARCHIVIST: Lenny Vincent, Fullerton College, Fullerton, California 92634. DIRECTORS: David Wise (1998-2000), Paula Cushing (1999-2001), Ann Rypstra (1999-2001). HONORARY MEMBERS: C. D. Dondale, H. W. Levi, A. R Millidge, W. Whit- comb. Cover photo: An ant mimic (Family Clubionidae: Myrmecium sp.) from Trinidad. {Photo by Joe Warfel of Arlington, Massachusetts) Publication date: 20 June 2000 @ This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 2000. The Journal of Arachnology 28:1-6 THE FAMILY GALLIENIELLIDAE (ARANEAE, GNAPHOSOIDEA) IN THE AMERICAS Pablo A. Goloboff: Consejo Nacional de Investigaciones Cientificas y Tecnicas, Instituto Miguel Lillo, Miguel Lillo 205, 4000 S. M. de Tucuman, Argentina ABSTRACT. Azilia leucostigma Mello-Leitao 1941 considered by Mello-Leitao as a metine (Tetrag- nathidae), is transferred to the gnaphosoid family Gallieniellidae, as the type species of the new genus Galianoella. The obliquely depressed endites, the flattened irregular posterior median eyes, and the conical anterior lateral spinnerets retaining a sclerotized distal ring, among other characters, clearly place the new genus in the family Gallieniellidae. Galianoella leucostigma is the only gallieniellid so far recorded from the Americas. This species has a specialized ant-preying behavior. Ant-preying may prove to be charac- teristic for all the family, as it was suspected in the with the modified chelicerae typical of the family. Keywords I Spiders, arachnids, Galianoella The family Gallieniellidae is a small family of dionychan spiders recorded so far only for Madagascar, the Comoro Islands, southern Af- rica, and Australia. Only a handful of papers have dealt with the group. Platnick (1984) summarized its taxonomic history and revised the family, describing several new species and the genus Legendrena\ Platnick (1990b) added some species in Gallieniella Millot 1947 and Legendrena\ Platnick (1990a) transferred Drassodella Hewitt 1916 from the Gnaphos- idae and mentioned the existence of undescri- bed genera of Australian gallieniellids. The gallieniellids have the obliquely de- pressed endites and flattened irregular poste- rior median eyes typical of the Gnaphosoidea, but they are probably the sister group of most other gnaphosoids because they have the an- terior lateral spinnerets conical and more closely set than in most other gnaphosoid fam- ilies and retain an apical segment (Platnick 1984, 1990a). These spiders are quite uncom- mon in collections, and very little is known of their biology. During recent years, several specimens of Gallieniellidae have been col- lected in dry and semi-arid habitats of north- western Argentina. These were first thought to represent an undescribed genus and species. The species, however, had been described (as a metine tetragnathid!) by Mello-Leitao (1941), with the name Azilia leucostigma, which is here designed as the type species of the new genus Galianoella. Although Mello- Madagascan Gallieniella', and it may be associated Leitao gave measurements for both the male and the female of A. leucostigma, the female was not illustrated, and the only specimen now available seems to be the male holotype. The abbreviations used in this study are standard for the Araneae. The notation for leg spines is as in Goloboff (1995). The speci- mens examined are deposited in the Instituto and Fundacidn Miguel Lillo, Tucuman, (FIML); Museo Argentine de Ciencias Natur- ales, Buenos Aires, (MACN); Museo de Cien- cias Naturales, La Plata (MLP); and American Museum of Natural History, New York (AMNH). Galianoella new genus Type species.-— leucostigma Mello- Leitao 1941. Etymology.— It is a pleasure to name the new genus after Marfa E. Galiano, who intro- duced me to arachnology, with gratitude for her continued help and advice, and in recog- nition of her contributions to arachnology. Diagnosis.- — Galianoella can be distin- guished easily from other gallieniellid genera by the membranose area on the cheliceral margins (Fig. 3), and the eyes occupying a wider portion of the head (Figs. 1, 2). Females can be distinguished also by the laterally de- pressed palpal tibia with strong prolateral spines (Fig. 4) and the wrinkled membranous posterior extensions of the epigynum (Fig. 6). Males can be recognized by the complex pal- 1 2 THE JOURNAL OF ARACHNOLOGY Figures 1-1 1. — Galianoella leucostigma. 1, Dorsal view, male; 2, Dorsal view, female; 3, Ventral view, female; 4, Right female palp, prolateral; 5, Spermathecae; 6, Epigynum; 7-9, Right male palp, bulb in resting position; 10, 11, Right male palp, bulb expanded. GOLOBOFF---AMERICAN GALLIENIELLIDS 3 Figures 12-14. — Galianoella leucostigma. 12, Female with egg-sac; 13, Cell; 14, Detail of cell entrance. pal femoral apophyses and keel (Figs. 7-9; absent in other genera of the family) and by the spiralled tegulum coiling aroung an exten- sive distal haematodocha (Figs. 8, 11). Description.— See species description. Relationships.— Their reduced number of cylindrical gland spigots suggests that Legen- drena and Gallieniella are sister groups; in different gnaphosoid families, as well as in Galianoella and Drassodella, those spigots are usually fairly numerous (and often ar- ranged in longitudinal rows; see Platnick 1990a). The eyes are set on a low tubercle in Gallieniella, Legendrena, and Drassodella, but are completely sessile and occupy most of the cephalic width in Galianoella and non- gallieniellid gnaphosoids, suggesting that Gal- ianoella is the sister group of the other three genera. The hypothesis of relationsips just pro- posed, however, is far from well supported, as it is contradicted by the elongated and inclined chelicerae of Gallieniella and Galianoella. Drassodella and Legendrena have shorter, more vertical chelicerae, which suggests that Galianoella and Gallieniella are sister genera. Besides the wider eye group, there are sev- eral other characters in which Galianoella dif- fers from all other gallieniellids, but they are most parsimoniously interpreted as its auta- pomorphies. These autapomorphies include the membranose cheliceral areas, the interior cheliceral faces with short spiniform setae, the undifferentiated cheliceral margins, the ven- tral spines on patellae, the wrinkled posterior epigynal extensions with a cavity, the ring- like tegulum with an extensive haematodocha, the modified female palpal tibia bearing strong prolateral setae, and the modified male palpal femur. Although its absence in the other gallieniellid genera cannot be ascertained from published illustrations of male palps, the subtegular projection is absent in Gallieniella (the only genus for which I have examined males). Thus, this projection may be another autapomorphy of Galianoella. Galianoella leucostigma (Mello-Leitao 1941) new combination Figs. 1-14 Azilia leucostigma Mello-Leitao, 1941: 155, fig, 50, pi. 7, fig. 33 (male holotype, from Argentina, Province of Salta, Salta, M. Biraben col., no date, MLP 14808, examined). Diagnosis.— See diagnosis for genus. Description.— Fewfl/e.- (From Chuscha). Total length 5.91 (6.78 with chelicerae). Cephalothorax (Fig. 2) 2.34 long, 1.83 wide. Cephalic region almost flat, 1.62 long, 1.38 wide. Eyes occupying 0.78 of head width, in two recurved rows; AME rounded, convex, on low common elevation; all other eyes com- pletely sessile; PME irregular, flattened, with diffuse limits. Chelicerae elongated (Fig. 3), with long fang, with about 60 small spiniform setae on interior face, larger and stronger near anterior face. Chelicerae distally narrow, with pro- and retromargins not well differentiated; with membranose area apically wider (visible only from below); single, large (retromargin- al?) tooth and 2 small denticles set on com- mon sclerotized elongate patch in the middle of this membranose area, 2 or 3 small (pro- marginal?) denticles on apex. Endites long, medially constricted; labium 0.56 long, 0.42 wide, flat; sternum 2.60 long, 1.20 wide. In- tercoxal bridges not visible. Palpi (Fig. 4) short; tibia short, laterally compressed, with strong prolateral erect setae. Legs long, slen- 4 THE JOURNAL OF ARACHNOLOGY Table 1. — Leg measurements (in mm) of Galianoella leucostigma. Female Femur Patella + Tibia Metatarsus Tarsus Total I 2.16 2.76 2.07 1.20 8.19 II 1.98 2.46 1.98 1.11 7.53 III 1.83 2.28 1.80 0.84 6.75 IV 2.28 2.91 2.43 1.08 8.70 Palp 0.72 0.72 — 0.66 2.10 Male Femur Patella + Tibia Metatarsus Tarsus Total I 2.12 2.76 2.12 1.24 8.24 II 2.00 2.28 2.00 1.08 7.36 III 1.88 2.24 1.80 0.88 6.80 IV 2.40 2.92 2.44 1.04 8.80 Palp 0.76 0.72 — 0.76 2.24 der; tarsi ascopulate, without claw-tufts; with few, weak spines. Measurements: See Table 1. Chaetotaxy: Femora: I, 1 D (1:3 AP), 1 P SUP (1:4 AP); II-IV, 1-1-0 D; palp, 1 p (1:4 ap). Patellae: I, II, 1 d ap (thin, erect, on the condyle joining patella and tibia), 1 v post b; III, IV, 1 d ap (as in I, II), 0 v; palp, 1 d ap (weaker than in I, II). Tibiae: I, II, 0; III, 1 d (1:4 ap), 1 V ANT (1:5 B), 2 V AP; IV, 1 d (1:4 ap), 1-1-1 V ANT, 1 V POST AP; palp, 3 P SUP A (erect, curved), 3 P (shorter). Metatarsi: I, II, 0; III, IV, 1 V ANT (1:3 B). Tarsi: I-IV, 0; palp, numerous erect thin setae on V and P, 1-1 D ANT B, 1-1 P (1:3 B). Palpal tibia laterally compressed (i.e., wider dorso-ventrally). Superior tarsal claws with single row of 3 teeth (basal one of III and IV bifid); third claw absent from all tarsi; palpal claw with 3 teeth, increasing in size from bas- al to distal. ALS conical, with distal sclero- tized ring, with 2 spigots (with short, rather wide shaft distinctly separated from base; pos- sibly corresponding to piriform glands, not SEM examined); PMS with larger spigot on apex (with base larger than shaft), plus 7 slightly smaller ones (possibly cylindrical, as they are absent in the male) on the dorsal face, in two alternate rows; PLS with 2 closely-set spigots on anterior edge, plus one medial spig- ot. Coloration: Cephalothorax and legs red- dish-brown; abdomen dorsally black, with two dorsal and single posterior yellow-cream are- as; venter pale, darker laterally and around spinnerets. Epigynum: (Fig. 6). Large sclero- tized plate, with posterior membranous wrin- kles, with posterior opening connecting to in- ternal cavity (apparently glandular). Spermathecae (Fig. 5) reniform; portion of copulatory ducts distal to spermathecae strongly sclerotized, spiralled; proximal por- tion twisted around most distal one, with me- dial glandular area (Fig. 5, arrow near strongly sclerotized edges of distal portion). Male: (Chuscha). As in female, except as noted: Total length 5.31 (5.91 with chelicer- ae). Cephalothorax (Fig. 1) 2.40 long, 1.83 wide. Cephalic region 1.56 long, 1.32 wide. Eyes occupying 0.77 of head width. Labium 0.56 long, 0.42 wide; sternum 2.43 long, 1.20 wide. Leg measurements: see Table 1. Chae- totaxy: As in female, except: Tibiae III, IV, 1- 1-2/1-0-1 V ANT, 1 V POST AP, 1 d (1:4 ap); palpal tibia with 5 setae on apex; cymbium with some thickened spiniform dorsal setae. Abdomen (Fig. 1) with dorsal anterior scutum. Spinnerets as in female; ALS spigots as in fe- male; PMS and PLS with single large apical spigot. Palp (Figs. 7-11): Three dorsal apical apophyses on femur: anterior one pointed and strongly sclerotized, posterior one blunt, mid- dle one rounded; retroventral apical keel on femur; retrolateral longitudinal keel on patel- la; tibia with dorsal long projection bearing strong setae. Bulb: Subtegulum very large, visible in prolateral view (Fig. 9), with small projection (Fig. 10) anchoring in the anterior basal rim of the cymbium; tegulum spiraled, almost continuous with strip-like spiralled em- bolus, coiling around large distal haematodo- cha occupying central position (Figs. 8, 11). Natural history,— Galianoella lives under stones or logs, in arid or semi-arid habitats. The specimens were found in small silk cells GOLOBOFF— AMERICAN GALLIENIELLIDS up to 2 cm long, 1 cm wide, sometimes cov- ered with debris and prey remains, with two entrances. The entrances had a peculiar struc- ture, with two small parallel flaps, each about 4 mm wide and with 15 small finger-like bars (formed by either a single thick thread, or sev- eral thinner threads compacted or cemented by some substance). These bars are about 1.5 mm long (half of which is imbedded in the silk mat), roughly parallel, and give the cell entrance the appearance of a double comb. Perhaps these peculiar combs help prevent ants from entering the refuge. The egg-sac is flattened, lenticular, 8-9 mm in diameter, with an internal white papery layer (as in many other gnaphosoids) and an outer layer of loose much thicker threads. Several cells contained more than a single egg- sac. The living specimens, although not defi- nitely myrmecomorph, have a strong ant-like appearance. They do not use their first legs as antennae (as many ant-mimicks do), but walk in a somewhat ant-like way; the resemblance to an ant is strengthened by the paler dots on the abdomen, which give the impression of a constriction. Especially because of the way they move, the first impression on seeing specimens in the field is of cell-living casta- neirine corinnids with unusually long chelic- erae and unusually bright PME (castaneirines do not live in cells, and have short vertical chelicerae and normal PME). In captivity, the specimens were observed to feed only on ants. A few other items of prey were offered, but ignored. The cells of Gali- anoella often contained Camponotus remains (Formicidae, Formicinae), and it is likely that adults prey upon them; in captivity, the spi- ders were fed Acromyrmex (a leaf-cutting ant, Myrmicinae) and soldiers of Pheidole (the Eu- ropean fire ant, also a myrmicine). The Mad- agascan Galieniella have been collected to- gether with ants, and it was suspected that they could prey upon them, but no actual ob- servations exist. The capture sequence in Galianoella is very stereotyped. The spider always attacks the ant from behind, placing its fangs on the sides of the ant’s thorax. When the ant was not facing away from the spider (the most common situation), the spider positioned her- self, moving around the ant sideways (i.e., al- ways facing the ant), in an arc about 1 cm in diameter; when the ant continued crawling 5 (which happened most of the times), the spi- der followed the ant in the same manner for a few centimeters until she could position her- self in a proper position to attack. While pur- suing the ant, the spider walked on six legs, with her chelicerae wide open, the palpi raised and retracted, and the extended anterior legs raised at an angle of about 45°. It is possible that the large tooth set on the sclerotized plate on the membranous chelic- eral patch and/or the spine-like setae on the inner chelicerae (together with the palpi, see below) play a special role in holding the ant after the attack. It is also possible that the spi- der fangs were not actually piercing the ant’s exoskeleton at this time (a couple catches were observed under the microscope, under low magnification); rather, the fangs seemed to embrace the ant’s thorax and coxae, holding the thorax pressed against the basal article of the chelicerae. Sometimes the spider held the ant in this way for only an instant, quickly releasing and following it until it died (always within a few seconds). Often, the spider did not release the ant at all; while holding the ant, the spider palpi were put downwards (with apical articles directed posteriorly), such that the tibial spines when pressed against the ant’s abdomen, and the ant’s abdomen could then only curve downwards. The spider palpi were then not visible from above. Careful ex- amination of the ant remains under light mag- nification (about 50X) revealed no holes, sug- gesting that the ant may be immobilized by a substance other than cheliceral venom. Jocque & Dippenaar-Schoeman (1992) have reported zodariid spiders subduing termites without biting them. Additional research is needed to determine whether that is the case in Gali- anoella, but it is not entirely unlikely that the unsclerotized cheliceral patches contain spe- cial glands that help in prey capture. The laroniine Eilica Keyserling 1891, an- other ant-catching gnaphosoid spider sympat- ric with Galianoella (collected at El Hongo and Chuscha), also has peculiar cheliceral modifications. The specimens of Eilica that were actually observed catching ants were mixed with other specimens of Eilica from the same localities, but later study revealed that two species {E. trilineata Mello-Leitao 1941 and E. modesta (Keyserling 1891)) coexist there; and one of the spiders was found eating a worker of Acromyrmex striatulus. The ant- 6 THE JOURNAL OF ARACHNOLOGY catching behavior of Eilica (so far unknown), however, is quite different from the cautious behavior of Galianoella: the spider quickly ran onto the ant’s head, bit the base of an an- tenna, quickly released the ant, and waited by the side for the ant to die (or at least, to be- come motionless; this occurred within a few seconds). The Laroniinae are characterized by a laminar (almost membranous) keel in the cheliceral margin. Callilepis Westring 1874, the other genus in the subfamily, has been re- ported to capture ants in a similar way (Heller 1976); and it is most likely that the cheliceral keel, synapomorphic for the subfamily, plays a special role here. Distribution.—Southem Salta and north- western Tucuman, in northwestern Argentina. The six known localities are all in two valleys which form part of a larger system of valleys, rather isolated from lower, more forested hab- itats. Relatively careful collecting in other parts of Salta and Tucuman has yielded no specimens of Galianoella, which may be re- stricted to these valleys. Other specimens examined. — ARGENTINA: Salta: Chuscha, 6 km NW Cafayate, 10 January 1995 (R Goloboff, C. Szumik), 27 (FIML); 18 April 1995 (R Goloboff, C. Szumik), 2 7 (FIML); 20 November 1995 (R Goloboff), 2S (FIML), Id (MACN), 1 7 (AMNH). El Kongo, 2 km S Ale- mania, July 1995 (M. Ramirez, R. Goloboff) 17 (MACN). La Salamanca, 3 km S Alemania, 19 Feb- ruary 1996 (R Goloboff, C. Szumik), 2d (FIML). Ruta Nacional 40, km. 1026, 6 km S Tolombon, 18 April 1995 (R Goloboff, C. Szumik), 17 3 juvs. (FIML). Tucuman: Amaicha del Valle, 10 January 1995 (R Goloboff, C. Szumik), 1 juv. (FIML). ACKNOWLEDGMENTS I am grateful to Norman Platnick for the loan of comparative specimens of Gallieniel- la, Legendrena, and Drassodella; to C. Sut- ton, A. Brescovit, and A. Bonaldo for making the type specimen of Azilia leucostigma avail- able; to Norman Platnick and Martin Ramirez for their critical comments on the manuscript; to the Consejo Nacional de Investigaciones Cientificas y Tecnicas for supporting my re- search; and to F. Cuezzo for identifying ant remains. LITERATURE CITED Goloboff, RA. 1995. A revision of the South American spiders of the family Nemesiidae (Ar- aneae, Mygalomorphae). Rart I: Species from Reru, Chile, Argentina, and Uruguay. Bull. American Mus. Nat. Hist. 224:12-189. Heller, G, 1976. Zum Beutefangverhalten der ame- isenfressenden Spinne Callilepis nocturna (Arachnida, Araneae, Drassodidae). Entomol. Germanica, 3:100-103. Jocque, R., and A.S. Dippenaar-Schoeman. 1992. Two new termite-eating Diores species (Araneae, Zodariidae) and some observations on unique prey immobilization. J. Nat. Hist., 26:1405™ 1412. Mello-Leitao, C. 1941. Las aranas de Cordoba, La Rioja, Catamarca, Tucuman, Salta y Jujuy. Rev. Mus. La Rlata (nueva serie, Zook), 2:99-198. Platnick, N. 1984. Studies on Malagasy spiders. 1. The family Gallieniellidae (Araneae, Gnaphoso- idea). American Mus. Nov., 2801:1-17. Platnick, N. 1990a. Spinneret morphology and the phylogeny of ground spiders (Araneae, Gnapho- soidea). American Mus. Nov., 2978:1-42. Platnick, N. 1990b. A new species of Legendrena (Araneae: Gallieniellidae) from Madagascar. J. New York. Entomol. Soc., 98:499-501. Manuscript received 1 September 1997, revised 14 January 1999. 2000. The Journal of Arachnology 28:7-15 DESCRIPTIONS AND NOTES ON THE GENUS PARADOSSENUS IN THE NEOTROPICAL REGION (ARANEAE, TRECHALEIDAE) Antonio D. Brescovit^; Josue Raizer^ and Maria Eugenia C. AmaraF: ^Lab. Artropodes Pegonhentos, Instituto Butantan, Av. Vital Brasil, 1500, CEP 05503-900, Sao Paulo, SP, Brazil; and ^Depto de Biologia, Centro de Ciencias Biologicas e de Saude, Universidade Federal de Mato Grosso do Sul, Campo Grande, MS, Brazil, C. Postal 549, CEP 79070-900, ABSTRACT. Three Brazilian species of the genus Paradossenus EO, Pickard-Cambridge 1903 are included in this paper: Paradossenus minimus (Mello-Leitao 1940), whose holotype was located and is here redescribed; Paradossenus corumba new species is described from Mato Grosso do Sul, Brazil and preliminary data on its biology are presented. Morphological data and new records of P. longipes (Tac- zanowski 1874) are included. Keywords: Paradossenus, Trachaleidae, Araneae, Neotropical region The genus Paradossenus F.O. Pickard- Cambridge 1903 was revised by Sierwald (1993) and includes three neotropical species: P. longipes (Taczanowski), P. pulcher Sier- wald 1993 and P. caricoi Sierwald 1993. In the same paper, she synonymized the mono- typic genus Xingusiella (type species X. min- ima), described by Mello-Leitao (1940) based on the illustration of the epigynum and char- acters presented in Mello-Leitao ’s description. The author also suggested that “P. minimus might be a fourth valid species in the genus Paradossenus” (Sierwald 1993). Recently the holotype of Xingusiella mini- ma was found in the MNRJ collection, mixed with other material of the family Pisauridae. The examination of this specimen confirms Sierwald ’s supposition that the specimen be- longs to this genus, and the species is here redescribed. While examining other Brazilian collections more P. longipes material was found, and its geographical distribution is here extended to include southern Brazil and north- ern Argentina specimens. A new species, P. corumba, much smaller than P. minimus, was collected by the second author (JR) in the pro- ject “Biodiversidade da Fauna Associada a Macrofitas Aquaticas”, which was being de- veloped in southern Pantanal floodplain, Co- rumba, Mato Grosso do Sul, Brazil and was organized by the third author (MEA). This new species is common in the study area, en- abling preliminary observations on its biolo- gy- METHODS The material examined belongs to the fol- lowing collections: IBSP, Instituto Butantan, Sao Paulo (A.D. Brescovit); MCN, Museu de Ciencias Naturais, Fundagao Zoobotanica do Rio Grande do Sul, Porto Alegre (E.H. Buck- up); MCTP, Museu de Ciencia e Tecnologia, Pontificia Universidade Catdlica do Rio Grande do Sul, Porto Alegre (A.A. Lise); MNRJ, Museu Nacional do Rio de Janeiro, Rio de Janeiro (A. Kury); ZUFMS, Colegao Zooldgica de Referencia da UFMS, Univer- sidade Federal de Mato Grosso do Sul, Campo Grande (L.O.I. Souza). The description format follows Brescovit & Hofer (1994) and the terminology used for the internal structures of the genitalia follows Sierwald (1993, 1996). All measurements are in millimeters. The epigyna were cleared in clove oil to study internal structures. Data on aquatic plant association, web type and prey capture strategies of Paradossenus corumba were obtained through field observations of 61 individuals, where 35 were studied by an animal focal method (sensu Lehner 1979; total of 175 minutes of observations divided in 35 sessions of five minutes each). These data were collected in temporary ponds in the 7 8 THE JOURNAL OF ARACHNOLOGY southern Pantanal floodplain (between 19°22'-19°33'S and 57°02'-57°03'W) from July 1994~April 1997. Paradossenus F.O. Pickard-Cambridge Paradossenus F.O. Pickard-Cambridge 1903: 155, (type species by original designation, Parados- senus nigricans F.O. Pickard-Cambridge [= Do- lomedes longipes Taczanowski 1874]). Sierwald 1993: 55. Xingusiella Mello-Leitao 1940: 23, (type species by original designation, X. minima Mello-Leitao. First synonymized by Sierwald 1993: 55, Carico 1993: 231. Paradossenus can be distin- guished from other trechaleids by at least four characters, three of which are presumably syn- apomorphies: male chelicerae with distinct elongated groove leading to the base of fang on the anterior surface of paturon (Figs. 6, 19; Sierwald 1993, fig, 11), leg I extremely long, and presence of a distal tegular projection, not pierced by the duct, in the male palp (Figs. 1, 17, 19; Sierwald 1990, fig. 35; 1993, fig. 12). An additional character would be the presence of slightly to moderately-recurved posterior eye row (Sierwald 1990, figs. 29, 30). The presence of four cheliceral teeth on the retro- margin, a character used as diagnostic by Sier- wald (1990), was inconsistent. The species in- cluded in this work had a retromargin with three teeth (Fig. 12). Paradossenus corumba Brescovit & Raizer new species Figs. 1-6; 11-17; 23 Types. — Male holotype from Corumba, Mato Grosso do Sul, Brazil, 1994, J. Raizer col., deposited in IBSP 6901; 1(3 & 19 par- atypes with same data of holotype, deposited in IBSP 6902 and 6903; 1 9 & 4 immatures from Passo do Lontra, Abobral Pantanal sub- region, Corumba, Mato Grosso do Sul, Brazil, 27 November 1994, J. Raizer col., deposited in IBSP 6904. Etymology. — The specific name is a noun in apposition taken from the type locality. Diagnosis. — The male of Paradossenus corumba is distinguished from P. longipes (Fig. 13; Sierwald 1990, figs. 34-36) by pres- ence of a retrolateral projection on the base of the cymbium (Figs. 2, 16) and a median apophysis with a bifid distal branch (Figs. 1, 17); the female of P. corumba differs from P. minimus by the sclerotized internal border of lateral lobes and middle field with a median depression (Figs. 3, 14). Description.— -Ma/e.* (holotype). Colora- tion: carapace orange with grayish border and with brown median dorsal band. Chelicerae yellow. Endites, labium and sternum yellow to white. Legs orange with brown longitudinal bands in all articles. Abdomen orange-brown, dorsally with grayish transversal bands and three pairs of longitudinal white spots. Ven- trally shiny white. Total length 2.65, Carapace 1.30 long, 1.20 wide. Clypeus 0.10, Eye di- ameters and interdistances: AME 0.10, ALE 0.08, PME 0.11, PLE 0,12; AME-AME 0.05, AME-ALE contiguous, PME-PME 0.08, PME-PLE 0.13, ALE-PLE 0.18. MOQ length 0.22, anterior width 0.21, posterior width 0.31. Chelicerae with elongated groove, deep, next to base of fang on the anterior surface (Figs. 6, 1 1) and 3 promarginal teeth being the median largest and 3 retromarginal denticles (Fig. 12). Labium 0.20 long, 0.17 wide. Ster- num 0.75 long, 0.67 wide. Abdomen 1.40 long. Leg measurements: I -femur 2.40; pa- tella 0.70; tibia 2.40; metatarsus 2.40; tarsus 1.00; total 8.90. II -2.00; 0.60; 1.90; 1.90; 0.80; 7.30. Ill -1.30; 0.40; 0.90; 1.00; 0,35; 3.95. IV -2.10; 0.50; 1.65; 2.10; 0.70; 7.05. Leg spination: tibia I-II v2-2-0; III-IV v2-2-2. Legs with plumose setae (Fig. 15). Bothrium of trichobothria with semicircular rim pre- senting longitudinal and slender striations (Fig. 13). Palp: retrolateral tibial apophysis subtriangular, very slender at tip; retrolateral ventral projection accentuated and globose (Figs. 1, 16); cymbium with retrolateral basal projection (Figs. 2, 16); tegulum with sperm ducts forming two loops; conductor incon- spicuous; median apophysis with two branch- es, one median rounded and the other distal bifid (Figs. 1, 17). Female: (IBSP 6904). Coloration as in male except legs with more accentuated bands on the articles and dorsum of abdomen darker. Total length 2.30. Carapace 1.20 long, 1.10 wide. Clypeus 0.07 high. Eye diameters and interdistances: AME 0.10, ALE 0.05, PME 0.12, PLE 0.11; AME-AME 0.03, AME-ALE 0.02, PME-PME 0.06, PME-PLE 0.12, ALE- PLE 0.21. MOQ length 0.23, front width 0.18, back width 0.26. Chelicerae not modified, with 3 promarginal teeth, the second basal be- ing larger than others and 3 large retromar- ginal teeth. Labium 0.15 long, 0.20 wide. BRESCOVIT ET AL.— NEOTROPICAL PARADOSSENUS 9 Figures l~6.—Paradossenus corumba new species, male. 1, Left palp, ventral view; 2, Left palp, retrolateral view; 3, Epigynum in ventral view; 4, Epigynum in dorsal view; 5, Epigynum in dorsal view (variation from Porto Cercado, Mato Grosso do Sul); 6, Male chelicera, anterior surface. Abbreviations: cp, basal projection of cymbium; dtp, distal tegular projection; e, embolus; II, lateral lobes; ma, median apophysis; mf, middle field; st, subtegulum; t, tegulum. Scale bars — 0.25 mm. Sternum 0.65 long, 0.60 wide. Abdomen 1.30 long. Leg measurements: I -femur 1.50; pa- tella 0.50; tibia 1,40; metatarsus 1.40; tarsus 0.55; total 5.35. 11 -1.40; 0.50; 1.25; 1.20; 0.50; 4.85. Ill -1.05; 0.30; 0.70; 0.80; 0.30; 3.15. IV -1.50; 0.45; 1.05; 1.55; 0.50; 5.15. Leg spination as in male. Epigynum: epigynal folds very narrow; middle field short, not cov- ering the epigastric furrow, with an anterior median depression; lateral lobes with narrow border sclerotized and rounded posteriorly (Figs. 3; 14). Vulva: wing of copulatory duct elongated, enlarged distally; true spermathe- cae slender, curved medially and with rounded head; elongated secondary spermathecae, transversally disposed (Fig. 4). Variation: Two males: total length 2.65- 2.70; carapace 1.20-1.30; femur I 2.00-2.40. Six females: total length 2.30-3.50; carapace 1.20-1.50; femur I 1.50-2.10. The females from Porto Cercado are darker, and the head of true spermathecae can be very slender (Fig. 5). Natural history,— Par ados senus corumba was observed associated with nine aquatic plants: Eichhornia azurea (Sw.) Kunth and E. crassipes (Mart.) Solms-Laub. (Pontederi- aceae), Echinodorus paniculatus Mich. (Alis- mataceae), Nymphaea amazonum Mart. & Zucc. (Nymphaeaceae), Salvinia auriculata Aublet (Salviniaceae), Phyllantus fluitans Miill. (Euphorbiaceae), Panicum mertensii Roth (Poaceae), Ludwigia inclinata (L.f.) Ra- ven (Onagraceae), and Pistia stratiotes L. (Ar- H— — - — -™-H Figures 7--10. — ^Species of Paradossenus, females. 7, % .—Paradossenus minimus. 7, Epigynum in ven- tral view; 8, Epigynum in dorsal view. 9, 10. P. longipes, variation of epigynum in dorsal view. 9. Reserva Florestal Adolfo Ducke, Manaus, Amazonas; 10, Sao Leopoldo, Rio Grande do Sul. Abbreviations: fd, fertilization duct; hs, head of true spermathecae; sec, secondary spermathecae; w, wing of copulatory duct. Scale bars = 0.25 mm. aceae). Spiders were most common in E. azur- ea (45.9% of 61 individuals). On the remaining plants occurrence was lower: 9.84% on E. crassipes, 6.56% on E. panicu- latus, 1.64% on N. amazonum, 6.56% on S. auriculata, 9.84% on P. fluitans, 11.48% on P. meriensii, 6.56% on L. inclinata, and 1.64% on P. stratiotes. Immatures and adults were found on aerial vegetative parts of the plants, but only im- matures were seen in retreats made on dam- aged or coiled plant leaves. Paradossenus cor- umba adults build irregular, horizontal webs (Fig. 23) which can be simple (observed only on Echinodorus paniculatus) or double (on Eichhornia azurea. Fig. 23, and E. crassipes), in this case without threads connecting the two parts. Spaces were observed between the threads and the plant petiole apex area (see arrows in the Fig. 23). In addition, the web has sticky silk threads. When the web is dou- ble, the spider walks under it, surrounding the plant petiole, and passing under each of the web parts through their spaces. In doing so, the spider is able to inspect the two parts of the web, sequentially. Some spiders were found walking on a plant or among plants. When walking on the plant, it patrols all its aerial parts. To move from one plant to another, spiders can walk on the water surface or attach silk threads be- tween plant leaves (in tall plants only, Ei- chhornia azurea, E. crassipes, Echinodorus paniculatus and Panicum mertensii). Paradossenus corumba can capture its prey in two ways. In the first way, a prey (an ara- neid) was captured actively while P. corumba walked on a plant. In this case, the hunting strategy is “search” (sensu Alcock 1979). In the second way, when a grasshopper nymph (probably Cornops sp., Acrididae) and a Dip- tera were captured, the spider stayed immobile on the plant leaf, near the water surface, keep- ing its cephalothorax oriented toward the wa- ter, and captured the preys that dropped in front of it. This behavior is characteristic of a BRESCOVIT ET AL.— NEOTROPICAL PARADOSSENUS 11 Figures 11-14.— Parados senus corumba new species, male. 11, Chelicerae, anterior surface; 12, Che- liceral teeth; 13, Tarsal tricobothrium, dorsal view; 14, Epigynum of female in ventral view (Scale bars for Figs. 11, 12, 14 = 100|xm; Fig. 13 - l|xm). “sit-and-wait” predator (Wise 1993). In both cases, the spider fed on the prey after immo- bilizing it with a single bite. The prey capture strategies observed for P. corumba indicate versatility in types of prey that are utilized. This versatility is poorly re- ported for spider species, with the exception of the salticid Portia fimbriata (Doleschall 1859) (see Jackson 1982; Jackson & Blest 1982) and the araneid Parawixia bistriata (Rengger 1836) (see Sandoval 1994). Distribution.— Mato Grosso do Sul, Brazil. Material examined.— BRAZIL. Mato Grosso do Sul: Porto Cercado, 4?2imm, August 1992 (A.A. Lise & A. Braul col.) (MCTP 2496; IBSP 6900); Corumbd Abobral sub-region, Passo da Lontra, 2dl$3imm, 1996, J. Raizer col. (IBSP 13757- 13759; ZUFMS). 12 THE JOURNAL OF ARACHNOLOGY Figures 15-18.- — Species of Paradossenus. 15-17. Paradossenus corumba, male. 15, Tibia I, lateral view, plumose setae; 16, Palpal tibia, retrolateral view; 17, Palpal bulb, ventral view. 18, P. longipes, male, palpal bulb, ventral view. (Scale bars for Fig. 15 = lOjxm; Figs. 16-18 = lOOjxm). BRESCOVIT ET AL.— NEOTROPICAL PARADOSSENUS 13 Figures l9~22.~Paradossenus longipes, male. 19, Chelicerae, anterior surface; 20, Tarsal organ; 21, Leg I, tarsal claws; 22, Epigynum of female from Mato Grosso, in ventral view. (Scale bars for Figs. 19; 21-22 - 100p.m; Fig. 20 - l^jim). Paradossenus minimus (Mello-Leitao) Figs. 7, 8 Xingusiella minima Mello-Leitao 1940: 23, fig. 1 (female holotype with egg sac, from Rio Xingu, Pard, Brazil, H. Leonardos col., MNRJ 585, ex- amined); Roewer 1954: 144. Paradossenus minimus: Sierwald 1993: 57. Diagnosis.— minimus is closest to P. corumba due to the rounded bor- der of lateral lobes, but may be distinguished by the epigynum with a short and narrow me- dian elevation on the middle field (Fig. 7) and the globose secondary spermathecae (Fig. 8). Description.— Fema/c.- (holotype). Colora- tion: carapace orange to gray (very discol- ored). Chelicerae red-brown. Endites and la- bium gray and white at tip. Sternum, legs and pedipalps yellowish. Abdomen dorsally gray- green, with an anterior dorsal grayish strip and a black band surrounding the spinnerets. Ven- trally white. Total length 3.50. Carapace 1.60 long, 1.20 wide. Clypeus 0.12 high. Eye di- ameters and interdistances: AME 0.08, ALE 0.07, PME 0.12, PLE 0.13; AME-AME 0.05, AME- ALE contiguous, PME-PME 0.12, PME-PLE 0.21, ALE-PLE 0.27. MOQ length 0.27, front width 0.11, back width 0.37. Che- licerae with 3 promarginal teeth and 3 retro- marginal denticles. Sternum 0.85 long, 0.55 wide. Abdomen 1.70 long. Leg measure- ments: I and II absent. Ill -femur 1.05; patella 0.35; tibia 0.80; metatarsus 1.00; tarsus 0.40; total 3.60. IV -1.90; 0.50; 1.40; 1.80; 0.65; 6.25. Spination: legs Ill-IV-tibia v2-2-2. Epi- gynum: epigynal folds broad, with an anterior widening, rounded; middle field posteriorly 14 THE JOURNAL OF ARACHNOLOGY Figure 23.— -Typical double web of Paradossen- us corumba adults, on leaf of Eichhornia azurea (Pontederiaceae). Observe the spaces between threads of each web parts and the plant leaf (ar- rows) used by the spider during the web inspection. Scale bar = 1 cm. short, with short and narrow median elevation; lateral lobes posteriorly rounded (Fig. 7). Vul- va: wing of copulatory duct slightly sclero- tized, short and subquadrangular; true spermathecae slender without distinct head; globose secondary spermathecae, without dis- tinct head stalk division (Fig. 8). Natural history.— The egg sac, reported by Mello-Leitao (1940) as globose, is similar to those found in the Lycosidae, attached to spin- nerets. Also present, as in P. longipes, are two discs, the upper larger than the lower disc, vault shaped, with the central scar where it was attached to spinnerets, the lower disk smaller and flat, with 20-25 shiny round eggs. Distribution.- — -North of Mato Grosso, Bra- zil. Material examined.— Only the type. Paradossenus longipes (Taczanowski) Figs. 9, 10; 18-22 Dolomedes longipes Taczanowski 1874: 88 ($ lec~ totype and S paralectotype, “Polska Academy of Sciences”, designated by Sierwald (1993), of Cayena, 04®o55'N, 52®18'W, Depto Cayena, Gui- ana Francesa, K. Jelski coL, not examined). Paradossenus nigricans Pickard-Cambridge 1903: 155 (male holotype and female paratype, “The Natural History, British Museum” BMNH~ 1898.5.5.101-2, from Buyassu, Parana e Breves, Maranhao, Brazil, not examined); Roewer 1954: 139; Bonnet 1958: 3325; Caporiacco 1948: 630; Sierwald 1990: 35; 1993: 59 (syn.). Paradossenus longipes: Caporiacco 1948: 630. Paradossenus taczanowskii Caporiacco 1948: 631 pis syntypes, “Muzeo di Zoologia di Specola, Firenze”, from Two Mouths, Essequibo, Guiana and Tibicuri-Cuyaha, Demerara, Guiana, not ex- amined); Sierwald 1990: 35. Morphoiogical notes.— Chelicerae in Fig. 19 showing the distinct elongated groove; tar- sal claws long, bearing 11-12 teeth, inferior tarsal claw on short tarsal onychium with slen- der ridges and presenting an elongated tooth (Fig. 21); tarsal organ oval with small and cir- cular opening (Fig. 20). Copulatory organs: no variation was found in the male palp col- lected in the northern region of South America (see Sierwald 1990, fig. 34) and those col- lected from the south of Brazil (Fig. 18). Among the females, no variation was found in the external plate of the epigynum (Fig. 22), but examining the internal structures, sig- nificant variation was detected in the form of the wings of copulatory ducts, which are very enlarged in the females from Manaus, Ama- zonas (Fig. 9) and narrowed in the females from Rio Grande do Sul (Fig. 10). Despite these variations we consider all specimens as P. longipes. Distribution.— Previously known from Venezuela, Guiana, Colombia, Bolivia, Peru and north of Brazil (Sierwald 1993: 62, 63). The new records extend the range of this spe- cies to south of Brazil and north of Argentina. New records.— BRAZIL. Acre: Serra do Divi- sor National Park (Camp), 1$, 14 November 1996 (R.S. Vieira coL) (IBSP 9305); Amazonas: Manaus, Reserva Florestal Adolfo Ducke, 1^, 8 August 1992 (S. Darwich coL) (MCTP 2846); 19,8 April 1992 (S. Darwich coL) (MCTP 2718); 12,8 April 1992 (U, Barbosa coL) (MCTP 2719); Mato Gros^ so: Coefluency Rivers Koluene and Xingu, (J.C. Carvalho col) (MNRJ 13446; IBSP 13756); Bahia: Iraquara, Pratinha (23®1FS, 48®12'W), 2 5 May 1998 (L.S. Rocha col.) (IBSP 20781); Sao Paulo: Mogi das Cruzes, Rio Tiete, 1?, July-Au- gust 1997 (R. Martins coL) (IBSP 11970); Parana: Candoi/Mangueirinha, Reservatorio do Rio Jordao, BRESCOVrr et al.=-neotropical paradossenus 15 Usina Hidreletrica de Segredo, 1$, 29 April 1996 (A.E Moraes & M.L. Javorowski col.) (IBSP 7142); Dois Vizinhos/Cruzeiro do Igua9u, Foz do Chopin, 1$, 8-15 November 1998 (Eq. IBSP col.) (IBSP 21247); Rio Grande do Sul: Rio Uruguai (Rodovia BR 153), Id, February 1989 (Eq. PUC col.) (MCTP 1296); Sao Leopoldo, 1$, 25 March 1983 (C.J. Becker col.) (MCN 11518); Triunfo, 19 with egg sac, 12 January 1989 (H.A. Gastal col.) (MCN 18086); ARGENTINA. Entre Misiones e Corri- entes: 19, 03-12 January 1989 (Eq. Garabi col.) (MCTP 1289). ACKNOWLEDGMENTS We would like to thank Prof. Pedro Kyo- hara and Miss Simone Perche de Toledo (USP) for the scanning electron micrographs, Cristina A. Rheims for the English language revision and the curators for loaning material for this study. The illustration of Paradossen- us corumba web drawing was provided by Vander M. Jesus. Thanks also to P. Sierwald and J, Berry for editorial review. This work was supported by CNPq grants (#530476/ 93.2; 522616/95.0 and 300169/96-5). LITERATURE CITED Alcock, J. 1979. Animal Behavior. An Evolution- ary Approach. Sinauer. Sunderland, Massachu- setts. Brescovit, A.D. & H. Hofer. 1994. Heidrunea, a new genus of the spider subfamily Rhoicininae (Araneae, Trechaleidae) from central Amazonia, Brazil. Andrias, 13:71-80. Caporiacco, L. 1948. Arachnida of British Guiana collected in 1931 and 1936 by Professors Beccari and Romiti. Proc. Zool. Soc. London, 118(3): 607-747. Carico, J.E. 1993. Revision of the genus Trechalea Thorell (Araneae, Trechaleidae) with a review of the taxonomy of the Trechaleidae and Pisauridae of the western hemisphere. J. ArachnoL, 21:226- 257. Jackson, R.R, 1982. The biology of Portia fim- briata, a web-building jumping spider (Araneae, Salticidae) from Queensland; intraspecific inter- actions. J. Zool. London, 196:295-305. Jackson, R.R, & A.D. Blest. 1982. The biology of Portia fimbriata, a web-building jumping spider (Araneae, Salticidae) from Queensland: utiliza- tion of webs and predatory versatility. J. Zool., London, 196:255-293. Lehner, P.N. 1979. Handbook of Ethological Meth- ods. Garland, New York. Mello-Leitao, C.F de. 1940. Aranhas do Xingu colhidas pelo dr. Henry Leonardos. Ann. Acad. Brasileira Sc„ 12(l):21-32. Pickard-Cambridge, F.O. 1903. On some new spe- cies of spiders belonging to the families Pisaur- idae and Senoculidae; with characters of a new genus. Proc. Zool. Soc. London, 1903 (1):151- 168. Sandoval, C.P. 1994. Plasticity in web design in the spider Parawixia bistriata: a response to var- iable prey type. Funct. EcoL, 8:701-707. Sierwald, P. 1990. Morphology and homologous features in the male palpal organ in Pisauridae and other spider families, with notes on the tax- onomy of Pisauridae (Arachnida: Araneae), Nemouria, 35:1-59. Sierwald, P. 1993. Revision of the spider genus Paradossenus, with notes on the family Trechal- eidae and the subfamily Rhoicininae (Araneae, Lycosoidea). Rev. ArachnoL, 10(3):53-74. Wise, D.H. 1993. Spiders In Ecological Webs. Cambridge Univ. Press, Cambridge. Manuscript received 6 October 1998, revised 1 July 1999. 2000. The Journal of Arachnology 28:16-22 OPTICAL STRUCTURE OF THE CRAB SPIDER MISUMENOPS FALLENS (ARANEAE, THOMISIDAE) Jose Antonio Corronca: CRJLAR-CONICET-UNLaR. Mendoza esq. Entre Rios. (5301) Anillaco, La Rioja, Argentina Hector R. Teranj Institute de Morfologfa Animal, Fundacion Miguel Lillo. Miguel Lillo 251. (4000) S.M. de Tucuman, Argentina ABSTRACT. We describe the histological structure of the eyes of Misumenops pallens (Araneae, Thom- isidae). We have carried out frontal, sagittal and transverse histological sections of the eyes. All the eyes have cuticular and laminar corneas and lenses. The anterior median eyes have two cellular types in the rhabdom; the remaining eyes have three cellular types. The anterior median eyes have a dark pigmented U-shaped mark in the middle of the retina. The indirect eyes have a dark pigmented band divided by a grate tapetum. The pathway of the optic nerves is also described. Our results suggest that Thomisidae may be a close relative of the superfamily Lycosoidea. RESUMEN. Se describe la estructura histologica de los ojos de Misumenops pallens (Araneae, Thom- isidae). Se realizaron cortes de los ojos en seccion frontal, sagittal y transversal. Todos los ojos tienen corneas y lentes cuticulares y laminares. Los ojos medios anteriores tienen dos tipos celulares en el rabdoma mientras que los restantes ojos tienen tres tipos celulares. Los ojos medios anteriores poseen, en el centro de la retina, una mancha de pigmento oscuro en forma de U, Los ojos de vision indirecta tienen una banda oscura de pigmento dividida por un tapete de tipo “grate.” Se estudia tambien el recorrido de los nervios opticos. Nuestros resultados sugieren que Thomisidae puede estar relacionado con la super- familia Lycosoidea. Keywords: Eyes, optic nerves, phylogenetic relationship Misumenops pallens (Keyserling 1880) (Thomisidae) are spiders that normally inhabit flowers and capture their prey by ambush. Their eyes are arranged in two recurved rows; in the anterior row the anterior median eyes (AME) are next to the bigger anterior lateral eyes (ALE) (Fig. 1). The posterior row eyes are equidistant, the posterior median eyes (PME) being smaller than the posterior lateral eyes (PLE). Lateral eyes are located on prom- inent tubercles. The dioptical apparatus of all the eyes of Misumenops pallens is formed by a cuticular cornea, a laminar lens and the “vit- reous body,” constituted by cone cells ar- ranged in a unique stratum that rests against a basal membrane. The eyes of Misumenops sp. have a dark pigmented ring called the “pu- pil,” a character shared with Lycosidae (Hom- ann 1971). The tapetum of the secondary eyes of Thomisidae is difficult to observe (see Levi 1982). In this study, the optic structure of Mis- umenops pallens is described in order to pro- vide new morphological characters that can be used in phylogenetic studies. METHODS Six adult females of Misumenops pallens, collected in March 1995 on soybean flowers in Burruyacu department (Tucuman, Argenti- na), were studied. Voucher specimens and his- tological slides are deposited in the arachnid collection of Fundacion Miguel Lillo, Tucu- man, Argentina (lot FML N° 2203). The spi- ders were anesthetized with chloroform. The cephalic regions of these spiders were dis- sected and were fixed in Bouin. The material was kept in n-Butyl alcohol during the time required to soften the cuticle, prior to embed- ding in Paraplast. Serial sections of 6 fxm thickness were cut, following the frontal, transverse and sagittal planes. Preparations were stained with Mal- lory-(Azan) Heidenhain and Haematoxylin- Eosin. Diagrams of optic nerves were prepared to 16 CORRONCA & TERAN— OPTICAL STRUCTURE OF MISUMENOPS 17 Figures 1, 2. — Misumenops pallens. 1. Ocular disposition; 2. Diagram showing the union of the ocular nerves in the cerebral ganglion -a. General view of the cerebral ganglion showing optic center, b. Trans- verse section showing the distribution of the optic nerves before their union with the cerebral ganglion, c. Transverse section showing the four optic centers, d. Transverse section of the optic center formed by the fusion of the four centers. Abbreviations: ale ™ anterior lateral eyes; ame = anterior median eyes; cad ” anterior right optic center; cai == anterior left optic center; co = optic center; coc — optic center of cerebral ganglion; cpd posterior right optic center; cpi = posterior left optic center; nj = optic nerve of anterior median eye; n2 = optic nerve of anterior lateral eye; nj = optic nerve of posterior median eye; n4 = optic nerve of posterior lateral eye. Scale bars: Fig. 1 - 0.42 mm; Fig. 2 = 66 p,m. trace their course as they leave each eye and enter the optic center of the cerebral ganglion; the nerves of each eye are designated as fol- lows: ni (AME), n2 (ALE), (PME) and n4 (PLE). RESULTS Anterior median eyes (AME).— (Figs. 3- 5). These eyes are pyriform, with their vertex towards the inner part. They have a cuticular and laminar cornea, with the outer surface formed by overlapping plates, separated by complete transverse grooves. The lens, located beneath the cornea, is laminar, ogival, with the greater convexity towards the inner part of the eye (Figs. 3, 5). It has few transverse grooves. Cone cells (Eakin & Brandenburger 1971) lie below the lens and contain a few irregular basal nuclei with homogeneous granular chro- matin (Figs. 4, 5). Cone cells are arranged in only one stratum and send out projections to- wards the lens. Cone cells rest against a thin basal membrane that separates them from the retina (Fig. 4). There is a wide dark pigmented ring (the “pupil”) in the anterior portion of the vitreous body (Figs. 4, 5). The retina is sub-conical and is formed by two cellular types, pigmented supporting cells and sensitive cells. Pigmented supporting cells are distributed in the central region of the rhabdom forming a U-shaped spot of dark pig- ment. There is a calyx-shaped layer of brown pigment between this spot and the pigmented ring of the dioptical apparatus (Figs. 4, 5). The pigmented layer is constituted by granules of brown and black pigment. Brown pigment dis- position is similar to the location of the dark pigment in the secondary eyes, while black pigment is located only in the central pig- mented zone of the retina. The function of each type of pigment has yet to be established. Each sensitive cell consists of a distal por- tion below the basal membrane, that forms a thin rhabdom located only in the central por- 18 THE JOURNAL OF ARACHNOLOGY Figures 3-7. — Misumenops pallens. 3-5. Anterior median eyes (AME); 3. Frontal section showing arrangement (250X); 4. Frontal section showing structural elements (400X); 5. Frontal section showing details of rhabdomeres and nuclei of sensitive cells (lOOOX). 6-7. Posterior median eyes; 6. Frontal section showning tapetum (250X); 7. Frontal sections showning structural elements (400X). Abbreviations: A = pigmented ring; C = cornea; CP = pigmented supporting cell; L = lens; MP = pigmented spot; NC = nucleus of cone cell; NS = nucleus of sensitive cell; nj = optic nerve of lateral anterior eyes; NP = nucleus of pigmented supporting cell; PP = brown pigment; RH — rhabdomeres; T = tapetum. Scale bars: Fig. 3 = 60 (xm, Fig. 4 = 4 pm. Fig, 5 = 1.6 pm. Fig. 6 = 66 pm and Fig. 7 = 42.6 pm. tion in front of the central pigmented spot (Fig. 5). The intermediate segment of the sen- sitive cells crosses the pigmented layer and ends in the nuclear portion, where the cell in- creases its volume (Fig. 5). The nuclei of sen- sitive cells are irregular, with homogeneous granular chromatin. They are located in a pe- ripheral basal stratum (Fig. 4). Anterior lateral eyes (ALE).- — (Fig. 9). These eyes are conical and situated in a an- terio-lateral position. Cornea, lens and cone cells are similar to those of the AME. The retina is formed by three cellular types, sen- sitive cells, pigmented supporting cells, and non-pigmented supporting cells (Fig. 9). Sen- sitive cells are arranged in at least two or three strata. They contain rounded nuclei, with granular chromatin homogeneously distribut- CORRONCA & TERAN—OPTICAL STRUCTURE OF MISUMENOPS 19 Figures 8-11. — Misumenops pallens. 8. Arrangement of posterior eyes and lateral anterior eyes, frontal section, showing disposition of optic nerves (250X); 9. Lateral anterior eyes showing principal structures (400X); 10-11. Lateral posterior eye; 10. Posterior lateral eye showing non-pigmented supporting cells (lOOOX); 11. Posterior lateral eyes showing optic nerve and structural elements (250X). Abbreviations: ale = anterior lateral eye; pme = posterior median eye; pie = posterior lateral eye; A - ring (“pupil”); CN = non-pigmented supporting cell; NC — nucleus of cone cell; NP — nucleus of pigmented supporting cell; NS = nucleus of sensitive cell; nj = optic nerve of anterior median eye; nj = optic nerve of anterior lateral eye; nj = optic nerve of posterior median eye; — optic nerve of posterior lateral eye; RT = rhabdomeres; T == tapetum. Scale bars: Figs. 8 and 11 = 66 |xm; Fig. 9 = 42.6 fxm; Fig. 10 = 15 |xm. ed, and clear cytoplasm around the nucleus. The intermediate segment extends from the soma of sensitive cells and continues in the parallel rhabdomeres (Fig. 9), whose projec- tions cross the “RT” type tapetum (according to Homann 1971) (Fig. 9). The few non-pig- mented supporting cells are large and have ovoid nuclei with homogeneous granular chromatin. They have abundant clear cyto- plasm with projections that can cross the rhab- domere layer (Fig. 9). Pigmented supporting cells contain cytoplasm with a great number of granules of concentrated pigment arranged in a dark layer. They are located between the rhabdomeres and the tapetum, and extend for- ward enclosing the vitreous body up to the lens base. There is a less pigmented wide lay- er below the tapetum. It is difficult to observe the nuclei of these cells due to the great amount of pigment, except in the peripheral lateral zone, where groups of nuclei of these cells can be observed (Fig. 9). Posterior median eyes (PME). — (Figs. 6, 7). These rounded eyes are located in a dorso- lateral position. Cornea, lens and cone cells are similar to those of the AME. Retina cells are similar to those of the ALE, except that the nuclei of sensitive ceLs are arranged in 20 THE JOURNAL OF ARACHNOLOGY two strata. Non-pigmented supporting cells are rare; they possess pyriform nuclei with ho- mogeneous granular chromatin and small cy- toplasm projections between the rhabdomeres. There are two layers of pigmented supporting cells, as in the ALE. These cell are separated by a well-developed “RT” type tapetum (Fig. 6). In a transverse section, rounded nuclei are visible, with homogeneous granular chromatin and some peripheral clear cytoplasm (Fig. 7). Posterior lateral eyes (PLE). — (Figs. 10, 11). These eyes are conical eyes and have cor- nea, lens and cone cells that are similar to those of the AME. The retina is similar to the retina of the ALE, except the nuclei of the sensitive cells are arranged in at least three strata. Pigmented cells, non-pigmented cells and shape of the tapetum are similar to those of the ALE (Figs. 10, 11). Trajectory of optic nerves.“”-Optic nerves from the AME run independently and parallel as they leave each eye, following the proso- mal median line (Fig. 12). The rest of the nerves emerge from the corresponding eyes, curve and run paired along the body median line (Fig. 13), between the poison glands. At the median region of the prosoma optic nerves remain paired. Their arrangement from the ventral to the dorsal part of the body is: nj, n2, n4 and n3 (Figs. 2b, 8). Posteriorly, optic nerves fuse to form four optic centers (Figs. 2c, 14). The two ventral optic centers corre- spond to the fusion of nj and n2 of their re- spective sides, while the dorsal optic centers, right and left, are formed by the fusion of the corresponding n4 and n^. The four optic cen- ters fuse in the posterior region of the proso- ma in a dorsal optic center located in the ce- rebral ganglion (Figs. 2d, 15). DISCUSSION Different cellular types were observed in the ocular structure of Misumenops pallens. This agrees with Eakin & Brandenburger’s (1971) description for Salticidae, Melamed & Trujillo Cenoz’s (1966) for Lycosidae and Corronca & Teran’s (1997) for Selenops La- treille (Selenopidae). However, the similarities found in the cellular types of the eyes of these families of spiders do not imply that there are no differences in their general structure. The tapetum and the anatomical structure of eyes are characters that can be used to recon- struct the phylogenetic relationships of spi- ders. Homann (1971) mentioned the presence of the “pupil” and the tapetum type as among the characters shared by both the Thomisidae and Lycosidae. Homann (1975) considered Thomisidae as sister group of the monophy- letic group Lycosoidea (Lycosidae, Senoculi- dae and Oxyopidae). Levi (1982) placed Thomisidae in their own superfamily together with Aphantochilidae, while Coddington & Levi (1991) considered Thomisidae in the Dionycha, even when its placement is not very clearly established. Philodromidae, Het- eropodidae and Selenopidae could be consid- ered as groups related with Thomisidae, be- cause of the presence of laterigrad legs and the locomotion type (Coddington & Levi 1991). Griswold (1993) studied the phyloge- netic relationships of Lycosoidea and consid- ered the copious and diverse anatomical and morphological characters, and established that the presence of an RT type of tapetum in at least one of the eyes is one of the two syna- pomorphies that supports the monophyly of this superfamily. The same author considered that the possible homology of the tapetum shape is the only evidence to include this fam- ily within Lycosoidea. Corronca & Teran (1997) suggested the probable relationship of Selenopidae with Lycosoidea. Results ob- tained from the study of the ocular structure of Misumenops pallens, and extrapolated to the rest of Thomisidae, show the existence of certain anatomical characters (presence of a well developed RT type tapetum in all sec- ondary eyes, except in PME where is reduced, and the “pupil”) shared with Lycosidae and of others (RT tapetum and sensitive cells in the secondary eyes arranged in al least two strata) with Selenopidae. These affinities sug- gest the probable relationship of both families (Thomisidae and Selenopidae) with Lycoso- idea. The U“Shaped spot of dark pigment, located in the central portion of AME retina in Thom- isidae, could be homologous with the V- shaped pigmented spot, typical of AME of Salticidae. Both structures present the same topology, but the four layers of receptive seg- ments that have been described by Eakin & Brandenburger (1971) for Salticidae are not present in Thomisidae. Recent observations by De la Serna & Spi- nelli (1995) for Latrodectus species (Theridi- idae) show that the four optic centers fuse to- CORRONCA & TERAN— OPTICAL STRUCTURE OF MISUMENOPS 21 Figures 12-15.-— Misumenops pallens. 12. Connection between anterior median eyes into the cerebral ganglion (250X); 13. Pathway of optic nerves of the anterior lateral eyes until fusion with the optic center, frontal section (160X); 14. Four optic centers, transverse section in the middle portion of the prosoma (250X); 15. Optic center formed by the fusion of the four centers, transverse section to the posterior portion of the prosoma (250X), Abbreviations: ame = anterior median eyes; ale = anterior lateral eye; CAD = anterior right optic center; CAI = anterior left optic center; CO = optic center; COC = optic center of cerebral ganglion; CPD “ posterior right optic center; CPI = posterior left optic center; GV = venom gland; NN = neuronal nucleus; nj, optic nerve of anterior median eye; n2 = optic nerve of anterior lateral eye. Scale bars: Figs. 12, 14 and 15 = 66 p.m. Fig. 13 = 94.2 |xm. 22 THE JOURNAL OF ARACHNOLOGY gether in an unique optic center in the cerebral ganglion. Our study of the trajectory of the optic nerves in Misumenops pallens agrees with these authors; however, the nerves do not fuse in their trajectory until they form the first optic centers. The presence of this character in Lycosidae and Salticidae should be studied. ACKNOWLEDGMENTS To Fundacion Miguel Lillo, INSUE and CRILAR-CONICET-UNLaR, for their sup- port and Mana Eugenia Morales for her help with the English version, LITERATURE CITED Coddington, J.A. & H.W. Levi. 1991. Systematics and evolution of spiders (Araneae). Annu. Rev. Ecol. Syst., 22:565-592. Corronca, J.A. & H.R. Teran. 1997. Estructura oc- ular de Selenops cache leti Simon (Araneae, Se- lenopidae). J. ArachnoL, 25:42-48. De la Serna de Esteban, C. & C.M. Spinelli. 1995. Los nervios opticos en cuatro especies de Lat- rodectus (Araneae, Theridiidae). J. ArachnoL, 23:31-36. Eakin, R.M. & J.L. Brandenburger. 1971. Fine structure of the eyes of jumping spiders. J. Ul- trastr. Res., 37:618-663. Griswold, C.E. 1993. Investigations into the phy- togeny of the lycosid spiders and their kin (Arachnida, Araneae, Lycosoidea). Smithsonian Contrib. ZooL, 539:1-39. Homann, H. 1971. Die Augen der Araneae. Ana- tomic, Ontogenie und Bedeutung fur die Syste- matik (Chelicerata, Arachnida). Z. Morph. Tiere, 69:201-272. Homann, H. 1975. Die Stellung der Thomisidae und der Philodromidae im System der Araneae. Z. Morphol. Tiere, 80:181-202. Levi, H.W. 1982. Araneae. Pp. 77-95, In Synopsis and Classification of Living Organisms, 2. (S.P. Parker, ed.). Melamed, J. & O. Trujillo-Cen6z. 1966. The fine structure of the visual system of Lycosa (Ara- neae, Lycosidae). Z. fiir Zellfor., 74:12-31. Manuscript received 30 November 1996, revised 4 June 1999. 2000. The Journal of Arachnology 28:23-28 MALE DIMORPHISM IN OEDOTHORAX GIBBOSUS (ARANEAE, LINYPHIIDAE): A MORPHOMETRIC ANALYSIS Stefan Heinemann and Gabriele Uhl; University of Bonn, Institute of Zoology, Department of Ethology, Kirschallee 1, D-53115 Bonn, Germany ABSTRACT. The linyphiid spiders Oedothorax gibbosus (Blackwall 1841) and Oedothorax tuberosus (Blackwall 1841) were formerly described as separate species due to marked differences in prosomal structures of the males. During the last decade it was demonstrated that they are two forms of a single species. However, it remained to be shown whether the former species represent two distinct morphs or extremes of a continuum of variation. A morphometric examination of 246 alcohol-preserved specimens revealed that individual spiders can clearly be assigned to one of two forms. No intermediates were found, demonstrating that there are two distinct morphs. Keywords; Species status, polymorphism, morphometry, sexual selection, gustatorial courtship Why individuals of some species occur in distinct varieties has been of considerable in- terest to evolutionary biologist (e.g., Clarke 1962). Dimorphism represents the simplest case of polymorphism, with two varieties maintained within the population. The most common case of dimorphism is the sexual di- morphism with males and females showing dimorphism in size (Anderson 1994). Behav- ioral or morphological dimorphism in one sex, usually occurring in the male sex, is known for a relatively large number of insects (Ham- ilton 1979; Thornhill & Alcock 1983; Dan- forth 1991; Alcock 1996; Eberhard & Gutier- rez 1991). To our knowledge, the only spider species investigated to date is the jumping spi- der, Maevia inclemens (Walckenaer 1837), in which the morphs show striking differences in body color and courtship behavior (Clark & Uetz 1992, 1993). Species with dimorphic males provide a unique opportunity to address questions about the importance of female choice (Gadgil 1972; Clark & Uetz 1992), male-male competition (Danforth 1991; Eber- hard & Gutierrez 1991), sensory exploitation (Clark & Uetz 1993), and alternative mating tactics with equal or unequal fitness (Austad 1984; Dominey 1984). However, it has yet to be shown that the varieties under consider- ation result from the expression of different developmental programs with a bimodal dis- tribution, excluding the differences that are simply extremes of a continuum of variation. The linyphiid spiders Oedothorax gibbosus (Blackwall 1841) and Oedothorax tuberosus (Blackwall 1841) were described as separate species due to differences in prosomal struc- tures of the males. In O. gibbosus, the male prosoma is raised to form a marked protuber- ance in front of which lays a deep notch sur- rounded by long black hairs. Protuberances, notches, grooves and poreplates frequently found in male linyphiid spiders were shown to function as gustatorial courtship devices in several species (Lopez & Emerit 1981; Schai- ble et al. 1986; Schaible & Gack 1987). Males of O. tuberosus on the other hand, lack the marked protuberance, notch and hair. How- ever, the division was doubted by several au- thors (Simon 1926; Locket & Millidge 1953; Wiehle 1960; Bosmans 1985; Roberts 1987) as neither the male pedipalps can be distin- guished nor are there differences in female so- matic and genitalic characteristics. Moreover, the two species almost always occur syntopi- cally (Wiehle 1960; Roberts 1987; Maelfait et al. 1990). Roberts (1987) strengthened this view by stating that: “occasional specimens seem to represent an almost intermediate state” and by including a drawing of a tub- erosus male with a slight notch. Not until a rearing study was undertaken by De Keer & Maelfait (1988) in which both male forms were reared from one egg-sac was it shown that O. gibbosus and O. tuberosus are two forms of one species. A more detailed rearing study supported this finding, demonstrating that the two forms are very ukely determined 23 24 THE JOURNAL OF ARACHNOLOGY by one major gene with a dominant and a re- cessive allele where the tuberosus phenotype is expressed in individuals carrying homozy- gotic recessive alleles (Maelfait et al. 1990), From this genetic system follows that the two forms must be discrete morphs which is in- compatible with the supposed intermediate forms. In a morphometric analysis, we ex- amine whether the gibbosus and tuberosus forms can be clearly distinguished on mor- phological grounds. This study provides the basis for the following investigations on fe- male mate choice. METHODS Oedothorax gibbosus occurs in North-, West- and Central Europe (Wiehle 1960). It is restricted to low productive, wet grassland and marshes that are frequently flooded during winter and requires high water quality, result- ing in a rather patchy distribution (De Keer & Maelfait 1989). We examined 246 alcohol-preserved spec- imens from the Institut Royal des Science Na- turelles de Belgique, Bruxelles, captured in pitfall traps from 1977 to 1991 at different locations in Belgium. We chose this collection for two reasons, 1) the most detailed study on the species was conducted in Belgium by Maelfait et al. (1990), and 2) this collection proved to be the largest one available, a pre- requisite for a solid morphometric investiga- tion. An example of each male form is illustrated in Figs. 1-4. For the morphometric analysis we took the following measures (in |xm)(Figs, 5-8): length of patella plus tibia of the first leg (a), height of the prosoma (b), width of the prosoma (c), length of the prosoma (d), dorsal line along the prosoma, when viewed from the side (e) and depth of the notch (f). To measure the height a perpendicular line was drawn from the highest point of the pro- soma. The dorsal line is a measure that in- cludes size and dimension of the notch and the hump. The height of the prosoma and the depth of the notch were measured additionally to examine both structures separately. The depth of the notch was measured by drawing a straight line over the notch from which a perpendicular was drawn to the deepest point of the notch. The width of the prosoma was measured at the widest part of the prosoma. To measure prosoma length, the length of a straight line from the front to the back of the prosoma, parallel to the sternum, was taken. The measure of patella plus tibia is frequently used as a measure of leg size and served as a measure independent of prosomal size. The measures were taken with a macro- scope (WILD M420) fitted with a CCD-cam- era (Pieper FK 5062), connected to a com- puter provided with the program NIH~Image (Version 1.60b7). SEM investigations were performed with a Hitachi S2460N using un- sputtered alcohol material under low vacuum mode. All statistical analyses were performed us- ing SPSS for Windows95, Version 8.0.1. The level of significance was set at 0.05. RESULTS The data were tested for normal distribu- tion: prosomal length, prosomal width, pro- somal height and length of the first leg showed a normal distribution (Kolmogorov- Smimov-one-sample-test: n = 246, (leg 1: n = 243), in all cases P > 0.05). The dorsal line of the prosoma {n = 246) and the depth of the notch {n = 219) were not normally dis- tributed (K-S test, both P < 0.01). In Principal Component Analysis using a correlation matrix and varimax rotation, two principal components with eigenvalues greater than 1 were extracted (Table 1). A clear sep- aration of the two morphs was possible along PCI which explains 48% of the variance. Characters highly correlated with this com- ponent are the dorsal line, the height of the prosoma and the depth of the notch, all char- acters whose presence is attributed to the gib- bosus form (Figs. 1, 2). The scatterplot of PC- scores shows two distinct distributions (Fig. 9), the left cloud representing the tuberosus form and the right one the gibbosus form. No intermediate forms were found. The character length of the dorsal line along the prosoma incorporates several prosomal measures. In order to exclude size effects, we used an index of the dorsal line relative to size as measured by prosoma length. The resulting histogram (Fig. 10) shows a clear bimodal dis- tribution and confirms that there are no inter- mediate forms. Thus we can safely assume the existence of two distinct morphs in O, gib- bosus. The mean of the dorsal line along the pro- soma of the gibbosus morph (I = 1948 |xm, HEINEMANN & UHL— MALE DIMORPHISM IN OEDOTHORAX GIBBOSUS 25 Table 1. — Rotated component matrix resulting from Principal Component Analysis using eigen- values greater than 1. PCI PC2 Dorsal line along the prosoma 0.974 0.158 Prosomal notch 0.973 -0.025 Prosomal height 0.944 0.107 Prosomal width 0.317 0.751 Leg 1 0.001 0.802 Prosomal length “0.341 0.863 Eigenvalues 2.89 1.99 Variance explained % 48.2 33.1 SD = 109) differs significantly from the one of the tuberosus morph (I = 1443 fxm, SD = 71) (Mann- Whitney- fZ-Test, ~ 141, n2 = 105, Z = 13.41, P < 0.001). The two morphs are significantly different in prosomal height, the prosoma of the gibbosus morph (x = 551 p.m, SD = 41) being higher than the one of the tuberosus morph (x = 425 p,m, SD = 43) (LTest, t = “23.35, #= 244, P < 0.001). Some males of the tuberosus morph showed a slight depression lacking hair on their prosoma. Comparison of depressions in the tuberosus morph (x = 14 pim, SD =10) with the notches of the gibbosus morph (x = 199 fxm, SD = 24) showed a significant dif- ference (C/-Test: = 114, ^2 = ^ = “12.781, P < 0.001). Furthermore, the pro- soma of the gibbosus morph is significantly broader (x = 810 p.m, SD = 34) as that of 500 pm Figures 1-4. — 1. Lateral view of prosoma of the gibbosus form of Oedothorax gibbosus', 2. Dorsal view of the gibbosus form; 3. Lateral view of prosoma of the tuberosus form; 4. Dorsal view of the tuberosus form. 26 THE JOURNAL OF ARACHNOLOGY Figures 5-8.— Schematic representation of the characters measured. 5. Prosoma of gibbosus form, lateral view; 6. Lateral view of prosoma of the tuberosus form; 7. Dorsal view of prosoma of the tuberosus form; 8. First leg: a = length of patella plus tibia of the first leg, b = height of the prosoma, c — width of the prosoma, d = length of the prosoma, e = dorsal line along the prosoma, when viewed from the side, and f = depth of the notch. the tuberosus morph (x = 793 p-m, SD = 34) (r-Test: t = -3.68, J/= 244, P < 0.001). The width of the prosoma significantly correlates with its height (Spearman rank correlation, r^ = 0.246, n = 246, P < 0.001). Although the gibbosus morph has a broader and higher prosoma than the tuberosus morph ^ • 3- tuberosus morph gibbosus morph C\J 2- 1 ° o. c (D c O ° Q3° o 1 - ° B □□ E o o ” Q. O c 0- B °§ p qj m °%iB ° o °° ° ® o° n ° ° •c Q. -1- □ p p p ^ BP qD □ ° B S ° °a ° -2- ° □ ° o □□ o -3 -2.0 -1.5 -1.0 -.5 0.0 .5 1.0 1.5 2.0 Principal component 1 Figure 9. — Scatter plot of scores resulting from Principal Component Analysis. the difference in the length of the prosoma is only marginally significant {gibbosus morph: X = 1007 fim, SD = 31, tuberosus morph: x = 1015 p-m, SD “ 38; f-Test: t = 1.85, df ~ 244, P = 0.066). Interestingly, the two morphs do not differ in overall body size as measured by the leg character {gibbosus morph: x = 861 p.m, SD = 32, tuberosus morph: x = 865 pfriVAKJ , MMUn€K10"Pieswpeima I -WMU^O^K^^Homoeomma jFt-B-MTT4~TrTrT-vji?A0wopg/ma ? ? 11111 Citharacanthus ^HBTTTTETO-Q- Cyclostemum Cyrtopholis W^^W[W\T‘\^'Vr\-‘Acanthoscurna Phormictopus Eupalaestrus Lasiodora Nhandu VUalius SA -J I S I 1 0-flJ 1 ?????? 7 aerobothria * ■Pamphobeteus Xenesthis Bmchypelma Sericopelma Megaphobema Theraphosa Pseudotheraphosa 1111111 Schizopelma* B^M-DQCHO“ Metriopeima * Hapalopus* 1111111 Hapalotremus* Figure 45,^ — Probable evolution of male palpal bulb keels in Theraphosinae. Bulb characters were mapped on the cladogram of Perez-Miles et al. (1996). Taxa which show incongruence between the character evolution proposed here and in the cladogram of Perez-Miles et al. (1996) are indicated by an asterisk. See text for further discussions. Abbreviations: 1 = information absent, refers to genera which did not have specimens examined in this work. A = apical keel; EC = embolus prolateral face extremely concave; PI = prolateral inferior keel; PS == prolateral superior keel; R = retrolateral keel; SA = subapical keel; SC = embolus prolateral face slightly concave above and under the prolateral keel. SC and EC were treated as additive multistate characters. Black rectangle = synapomorphy; white rectangle = reversal. Black square = present; white square = absent; upper black triangle in a square = character exhibiting interspecific variation. ventral subapical region (SGA) that seems to be a synapomorphy of these species and is called Acanthoscurria ferina and A. insubtilis subapical granular area. The opposite occurs in some species of the genera Cyriocosmus and Euathlus, where the four basic groups of keels are absent. The pal- pal bulb of Cyriocosmus (Figs. 17, 18) is highly modified and the species possess from short to large paraembolic apophysis (PA) (see Schiapelli & Gerschman de Pikelin 1973). In some species studied I have found some vestiges of the superior prolateral keel and I believe that, with an increased knowl- edge of the genus, there is the possibility that some basal species with a less modified em- bolus can be found, which still retain these keels plesiomorphically. In Euathlus truculentus, however, the only keel present is one ventral medial crest (VC, Figs. 7, 8) not found in other theraphosine species. There is no vestige of the prolateral keels that, as shown above, are the most ple- siomorphic ones in Theraphosinae. Because of the basal position in a politomy occupied by Euathlus in the cladistic analysis of Thera- phosinae by Perez-Miles et al. (1996) (Fig. 45), there are two possibilities: (1), The pro- BERTANI— THERAPHOSINAE MALE PALPAL BULBS 41 lateral keels are a synapomorphy of all ther- aphosine except Euathlus; in this case Euath- lus is the most basal taxon of all Theraphosinae, (2). The prolateral keels were lost in Euathlus truculentus. The two possi- bilities seem equally parsimonious. CONCLUSIONS As shown above, the theraphosine palpal bulbs present some basic groups of keels which are widespread among almost all ther- aphosine species. These five keels were ho- mologized through the classical criteria of ho- mology, i.e., they presented the same relative position in the bulbs; they presented morpho- logical similarity, considering that no extreme and improbable changes were seen; and they were in accordance with the other characters, in this case the other keels. The last one is the most powerful test of homology and the most important to systematics (Patterson 1982). Also, the co-occurence of the five proposed keels in some species is in accordance with the conjunction test of Patterson (1982), i.e., they constitute five homologous keels. Keels other than these basic ones were found and some morphological modifications were seen. I consider this only one more argument to jus- tify that these structures, overlooked for so long, could be valuable for taxonomy and sys- tematic work due to their great morphological interspecific variability. Of course, this work must be seen as an initial approach and surely many alterations will take place when more information on theraphosine morphology and hypothesis of relationship become available, a reason why no phylogenetic analysis was car- ried out here. However, when considering the cladogram of theraphosine genera proposed by Perez-Miles et al. 1996, I found concor- dance with the keels evolution proposed here, with two exceptions (Fig. 45). The first one is the genus Sphaerobothria which, as discussed earlier, has a very similar bulb when com- pared with some Aphonopelma species. The second is the branch including Schizopelma, Metriopelma, Hapalopus, and Hapalotremus. The position of these branches in this clado- gram is due in part to a distinct interpetration of male palpal bulb characters carried out in Perez-Miles et al. (1996) cladistic analysis. For example, in this paper the character “bul- bal keels smooth or absent” is considered primitive, while “bulbal keels serrated” is considered derived for Eupalaestrus, Nhandu, Vitalius, Lasiodora and Sphaerobothria. However, in Sphaerobothria the serrated (den- ticulated) keel is the PI (Fig. 24), while in the other four genera, the serrated keel is the SA (Fig. 32); thus these characters are non-ho- mologous and should be recoded. The reinter- pretation of these characters surely will cause some changes to this cladogram topology. ACKNOWLEDGMENTS I thank the following persons for the loan of specimens: Adriano B. Kury (MNRJ, Rio de Janeiro), Amo A. Lise (MCP, Porto Ale- gre), Fernando P6rez-Miles (MHNM, Monte- video), and Rick C. West (Victoria). Samuel Marshall and Dietmar Pinz kindly helped with some preserved specimens. Antonio D. Bres- covit, Hilton Japyassu, Pedro Gnaspini, Pedro 1. da Silva Jr, Ricardo Pinto-da-Rocha, Sergio A. Vanin, and Sylvia Lucas made important suggestions when the study was conducted. Antonio D. Brescovit, Fernando Perez-Miles, Norman 1. Platnick, Pablo Goloboff, and Rick C. West provided valuable comments on a previous draft of the manuscript. James Berry, Robert Raven, Petra Sierwald and an anony- mous reviewer also improved the manuscript considerably. I am also extremely grateful to Katia M. Faria who kindly made the illustra- tions. LITERATURE CITED Biicherl, W. 1957. Sobre a importancia dos bulbos copuladores e das apofises tibiais dos machos na sistematica das aranhas caranguejeiras (Orthog- natha). Anais Acad. Brasileira Cienc., 29:377- 416. Coddington, J.A. 1990. Ontogeny and homology in the male palpus of orb-weaving spiders and their relatives, with comments on phylogeny (Ar- aneoclada: Araneoidea, Deinopoidea). Smith. Contrib. Zook, 496:1-52. Coddington, J.A. & H.W. Levi. 1991. Systematics and evolution of spiders (Araneae). Annu. Rev. Ecol. Syst., 22:565-592. Comstock, J.H. 1910. The palp of male spiders. Ann. Entomol. Soc, America, 3:161-185. Cooke, J.A.L., V.D. Roth & EH. Miller. 1972. The urticating hairs of theraphosid spiders. American Mus. Nov., 2498:1-43. Gerschman de Pikelin, B.S. & R.D. Schiapelli. 1972. El genero Homoeomma Ausserer, 1871. Physis, 31:237-258. Gerschman de Pikelin, B.S. & R.D. Schiapelli. 1973. La subfamilia Ischnocolinae (Araneae, 42 THE JOURNAL OF ARACHNOLOGY Theraphosidae). Revta. Mus. Argentino Cienc. Nat. Bernardino Rivadavia (EntomoL), 4:43-77. Goloboff, RA. 1995. A revision of the South American spiders of the family Nemesiidae (Ar- aneae, Mygalomorphae). Part I: Species from Peru, Chile, Argentina, and Uruguay. Bull. American Mus. Nat. Hist., 224:1-189. Mello-Leitao, C.F. 1923. Theraphosoideas do Bras- il. Rev. Mus. Paulista, 13:1-438. Patterson, C. 1982. Morphological characters and homology. Pp. 21-74, In Problems of Phyloge- netic Reconstruction. (K.A. Joysey & A.E. Fri- day, eds.). Systematics Association Special Vol- ume. London, Academic Press. Perez-Miles, F, S.M. Lucas, PI. da Silva, Jr. & R. Bertani. 1996. Systematic revision and cladistic analysis of Theraphosinae (Araneae: Theraphos- idae). Mygalomorph, 1:33-68. Pickard-Cambridge, F.O. 1897. Arachnida-Aranei- da. Vol. II. Pp. 1-42, In Biologia Central!- Amer- icana (Godman, F.D. & O. Salvin, eds.). London. Pocock, R.I. 1903. On some genera and species of South-American Aviculariidae. Ann. Mag. Nat. Hist., ser. 7, 11:81-115. Prentice, T.R. 1997. Theraphosidae of the Mojave Desert west and north of the Colorado River (Ar- aneae, Mygalomorphae, Theraphosidae). J. Ar- achnoL, 25:137-176. Raven, R.J. 1985. The spider infraorder Mygalo- morphae (Araneae): Cladistics and systematics. Bull. American Mus. Nat. Hist., 182:1-180. Raven, R.J. 1990. Comments on the proposed pre- cedence of Aphonopelma Pocock 1901 (Arach- nida, Araneae) over Rhechostica Simon 1892. Bull. Zool. Norn., 47(2): 126. Remane, A. 1956. Die Grundlagen des naturlichen Systems, der vergleichenden Anatomie und Phy- logenetik. Leipzig: Geest und Portig. Schiapelli, R.D. & B.S. Gerschman de Pikelin. 1962. Importancia de las espermatecas en la sis- tematica de las arahas del suborden Mygalomor- phae (Araneae). Physis, 23:69-75. Schiapelli, R.D. & B.S. Gerschman de Pikelin. 1964. El genero Acanthoscurria Ausserer 1871 en la Argentina. Physis, 24:391-417. Schiapelli, R.D. & B.S. Gerschman de Pikelin. 1973. Genero Cyriocosmus Simon, 1903 (Ara- neae, Theraphosidae). Physis, 32:61-70. Schiapelli, R.D. & B.S. Gerschman de Pikelin. 1979. Las aranas de la subfamilia Theraphosinae (Araneae, Theraphosidae). Revta. Mus. Argenti- no Cienc. Nat. Bernardino Rivadavia, 5:287- 330. Sierwald, P. 1990. Morphology and homologous features in the male palpal organ in Pisauridae and other spider families, with notes on the tax- onomy of Pisauridae (Arachnida, Araneae). Nemouria, 35:1-59. Simon, E. 1892. Histoire Naturelle des Araignees, Paris: Roret, 1, 256 pp. Smith, A.M. 1995. Tarantula Spiders: Tarantulas of the U.S.A. and Mexico. Fitzgerald PubL, Lon- don. 196 pp. Valerio, C.E. 1980a. Aranas terafosidas de Costa Rica (Araneae, Theraphosidae). I. Sericopelma y Brachypelma. Brenesia, 18:259-288. Valerio, C.E. 1980b. Aranas terafosidas de Costa Rica (Araneae, Theraphosidae). III. Sphaero- bothria, Aphonopelma, Pterinopelma, Cithara- canthus, Crypsidromus y Stichoplastus. Rev, Biol. Trop., 28:271-296. Manuscript received 23 October 1998, revised 27 April 1999. 2000. The Journal of Arachnology 28:43-48 EXPLORING FUNCTIONAL ASSOCIATIONS BETWEEN SPIDER CRIBELLA AND CALAMISTRA Brent D. OpeO, Jamel S. Sandidge* and Jason E. Bond^: Department of Biology, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061 USA ABSTRACT. A spider’s calamistrum draws silk fibrils from its cribellum and helps combine them with supporting strands to form a cribellar prey capture thread. Despite the close functional association of these two features, this study shows that there is a great deal of variability in the ratio of cribellum width to calamistrum length. When the independent contrast method was used to examine these two features in 1 1 species representing seven families, no relationship was found. Likewise, no relationship was found among nine species representing seven genera of the family Uloboridae. Only among the 14 species of Mallos (Dictynidae) was calamistrum length directly related to cribellum width. This suggests that, above the genus level, differences in spinning behavior and morphological features such as leg length and abdomen size and shape influence the relationship of these two features. Keywords; cribellar thread, cribellate spiders, independent contrast method, functional linkage The outer surfaces of a cribellar capture thread are formed of thousands of fine, looped fibrils that are produced by spinning spigots on the cribellum (Eberhard & Pereira 1993; Opel! 1994a, 1995, 1996; Peters 1983, 1984, 1986, 1992). These fibrils are drawn from the cribellum and manipulated by a setal comb on the fourth walking leg, termed the calamis- trum, as they are combined with axial and, in some cases, paracribellar fibers to form a completed capture thread. This close function- al linkage between the calamistrum and the cribellum suggests that their features should also be closely related. The most obvious fea- tures to exhibit this relationship should be cal- amistrum length and cribellum width. We pre- dict that the calamistrum must be long enough to fully span the cribellum as it sweeps over it in a combing motion. However, cribellum width may not be the only factor that influ- ences calamistrum length. The effective length of a calamistrum is probably deter- mined by such factors as the angle at which the calamistrum passes over the cribellum and the lateral movement of the calamistrum dur- ing a combing stroke. Although these features ’ Current address: Dept, of Entomology, Univer- sity of Kansas, Lawrence, Kansas 66045 USA ^ Current address: Dept, of Zoology, Insect Di- vision, Field Museum of Natural History, Roose- velt Road at Lake Shore Drive, Chicago, Illinois 60605 USA and their relationships are poorly studied, they are likely to be affected by the length and width of a spider’s abdomen, the length of a spider’s fourth legs, by the manner in which the combing leg is supported (Eberhard 1988), and probably by other details of the combing behavior such as the length of each combing stroke. The diversity in cribellar thread-combing behavior documented by Eberhard (1988) sug- gests that the ratio of calamistrum length to cribellum width may differ considerably among cribellate taxa. The null hypothesis of this study is that this ratio is uniform for all cribellate taxa. Using the comparative method of phylogenetic systematics (Harvey & Pagel 1991), we test this hypothesis at three hier- archical levels: the interfamilial level, the in- trafamilial level, and the intrageneric level. The degree to which differences in behavior and other aspects of anatomy influence the ra- tio of calamistrum length to cribellum width will affect the level at which the null hypoth- esis will be rejected. As behavioral and mor- phological features should be most similar within members of the same genus, it should be more difficult to reject the null hypothesis at this level than at more inclusive levels, METHODS Measurements.— “The fourth legs and cri- bella of spiders were removed and mounted in water-soluble medium on microscope 43 44 THE JOURNAL OF ARACHNOLOGY Table 1. — Means and standard deviations of the ratio calamistrum length to cribellum width of rep- resentative species. Family Species n X SD Uloboridae Miagrammopes animotus 31 1.35 0.10 Uloborus glomosus 21 1.10 0.13 Octonoba sinensis 24 1.27 0.11 Dictynidae Mexitilia trivittata 6 1.67 0.24 Mallos bryantii 5 1.55 0.25 Mallo niveus 9 1.58 0.27 Mallos mians 8 1.49 0.19 slides, Calamistrum length and cribellum width were measured to at least the nearest 20fxm under a compound or dissecting micro- scope. Two indices can be used for calamis- trum length: the distance separating the tips of the proximal and distal-most setae of the cal- amistrum and the distance separating the proximal and distal-most setal bases. We chose the second index for two reasons. First, it can be more consistently measured and is not affected by missing setae. Second, it does not make any assumptions about the deflec- tion of calamistrum setae during cribellar fi- bril combing. In the case of those species with divided cribella, cribellum width included the central region that separated the two halves of the cribellum. We measured a single mature female per species. We reasoned that, as the cribellum and calamistrum must be function- ally linked throughout an individual’s devel- opment, these measurements would provide a more rigorous test of the hypothesis than would the use of mean values derived from several individuals of a species. Table 1 gives the variance of the ratio of calamistrum length to cribellum width for seven species included in this study. Phylogenetic analysis.— This study in- cludes representatives of the infraorder Ara- neomorphae, the family Uloboridae, and the dictynid genus Mallos O. Pickard-Cambridge 1902 (Figs. 1-3) and uses the phytogenies of Griswold et al. (in press), Coddington (1990), and Bond & Opell (1997), respectively. To an- alyze the relationships of calamistrum length and cribellum width in a phylogenetic context we used the independent contrasts method of Felsenstein (1985), as implemented by the Comparative Analysis of Independent Con- trasts program of Purvis & Rambaut (1995). All branch lengths were treated as equal. This method minimizes the influence of non-inde- pendence of the data due to phylogenetic re- lationship by analyzing directional changes in continuous characters. It does so by comput- ing differences between the features of sister taxa (both extant taxa and their inferred an- cestors). These differences are then normal- ized and relationships among the resulting in- dependent contrast values are examined using regression statistics (see Harvey & Pagel 1991 for a review of this approach). All known species of the genus Mallos were included in the analysis of the relation- ship between the calamistrum length and cri- bellum width. In contrast, analyses of the oth- er two clades included only some of the known members. We examined the conse- quences of partial sampling by analyzing the relationship between calamistrum length and cribellum width within subsets of the genus Mallos. We used a random number generator to select seven of the 14 species of Mallos. After constructing a pruned phytogeny that in- cluded these seven species and Mexitilia tri- vittata (Banks 1901) as an outgroup, we ran an independent contrast analysis for calamis- trum length and cribellum width. This proce- dure was repeated until a total of ten analyses had been run. We then repeated the entire pro- cedure a second time with nine species of Mallos being selected each time. RESULTS Values for calamistrum length and cribel- lum width are given in Figs. 1-3. Within the Araneomorphae, the ratio of calamistrum length to cribellum width ranged from 0.99- 2.57; and an independent contrast analysis showed that there was no relationship between the dimensions of these two features (F = 0.09, = 0.01, P = 0.77). Within the Ulo- boridae the ratio of calamistrum length to cri- bellum width ranged from 1.07-2.06 and an independent contrast analysis showed that there was no relationship between the dimen- sions of these two features {F = 0.63, = 0.10, F = 0.46). When this analysis is restrict- ed to orb-weaving uloborids of the genera Waitkera Opell 1979, Siratoba Opell 1979, Uloborus Latreille 1806, Octonoba Opell OPELL ET AL.— ^SPIDER CRIBELLA AND CALAMISTRA 45 Filistatldae — — - Kukulcmia Mbemmlis 479, 472 (HonU 1842) (0.99) Oscobfidae — — PlatoBcoblus florfdanus 218,392 (Banks 1896) (1.30) Uloboridaa Waitkera waitakeransSs 381,780 (Chamberiain 1946) (2.04) — . Uloborus glomosus 574, 672 (Waickenaar 1 837) (1.17) Dlctynidae — Mexitilia trivittata 448,816 (Banks 1901) (1.82) — Mallos bryanti Gertsch 1946 (l-S®) Amaurobildae — Callobius bennetti 686,1102 (Biackwall 1846) (1.61) Neolanidae — Neolana pallida 444, 1142 Foreter & Wilton 1 973 (2.57) Desfdae — — Matachla livor 840, 959 (Urquhart 1892) (1.14) f— Badumna insignia 697, 1499 (L. Koch 1872) (2.15) — ' Badumna hnginqua 664,1204 (L Koch 1867) (1.81) Figure 1. — Phylogeny of species representing seven families (from Griswold et al. 1999). Follow- ing each species is the width of its cribellum and the length of its calamistrum, both in p,m. Ratios of calamistrum length to cribellum width are in paren- theses. 1979, and Philoponella Mello-Leitao 1917 an independent contrast analysis still fails to show a relationship between calamistrum length and cribellum width (F = 0.11, 7?^ = 0.05, P = 0.77). Within the genus Mallos, the calamistrum length to cribellum width ratio ranged only from 1.26“L82 and an independent contrast analysis showed that there was a relationship (F = 8.40, F2 = 0.41, P = 0.013) between the dimensions of these two features (Fig. 4). However, in only three of the ten subsets that included seven Mallos species plus Mexitilia trivittata was there a significant relationship between calamistrum length and cribellum width (F - 8.91-22.75, - 0.64-0.82, P - 0.005-0.031). When the sample size was in- creased to include nine Mallos species, seven of the ten samples showed a relationship be- tween these features (F = 5.61-19.95, R} = 0.45-0.74, P = 0.050-0.003). DISCUSSION The size of a spider’s cribellum and the number of spigots that it bears are the main Wmitkora waitakennsis 381,780 (Chamberlain 1946) (2.04) SIratoba mf arena 223, 460 (Muma a Gertsch 1964) (2.06) Hyptiotes cavatus 390, 740 (Hentz 1847) (1.90) Hyptiotes gertschi 465, 820 Chamb.&lvie 1935 (1.76) — • Mlagrammopes animotus 540, 700 Chlckering 1968 (1.30) - — Miagrammopes species 426, 720 (1.69) Uloborus glomosus 574, 672 (Walekenaer 1837) (1.17) Octonoba sinensis §60, 600 (Simon 1880) (1.07) Philoponella aiizonica 409, 800 (Gertsch 1936) (1.96) Figure 2.— Phylogeny of species belonging to the family Uloboridae (from Coddington 1990). Fol- lowing each species is the width of its cribellum and the length of its calamistrum, both in jxm. Ra- tios of calamistrum length to cribellum width are in parentheses. factors that correlate with the stickiness of the cribellar thread that it produces (Opell 1994a, 1995, in press). However, differences in the way cribellar fibrils are combined with sup- porting fibers can alter thread stickiness (Opell 1994b), as can the deposition of linear cribellar threads in a looped manner when they are placed in the web (Opell, unpub. data). Although cribellum shape differs among taxa, spigot number is generally relat- ed to cribellum width. This evolutionary plas- ticity in cribellum width is reflected by dif- ferences in calamistrum length. The ratio of calamistrum length to cribel- lum width differs among taxa; but, with one exception, it always exceeds one. In Kulul- cania hibernalis (Hentz 1842) calamistrum length and cribellum width are essentially the same. This suggests that the production of a cribellar thread requires the calamistrum to span the complete width of the cribellum dur- ing a combing stroke. It is possible that a cal- amistrum could comb fibrils from only part of the cribellum spigots, but this seems unlikely for two reasons. First, as the spigots of the cribellum are probably not regionally con- trolled, non-calami strum setae on other parts 46 THE JOURNAL OF ARACHNOLOGY Mexitilla trivittata 448, 8 1 6 (Banks 1901) (1. 82) NIallos hmsperius 380, 600 (Chamberlin 1316) (1.58) Mallos margaretaa 380, 580 Gartsch 1946 (1.53) Mallos gregalls 380, 480 (Simon 1909) (1.26) Mallos bryanti 480, 760 Gertsch 1946 (1.58) Mallos kraussi 560, 780 Gertsch 1948 (1.39) Mallos dugesi 540, 980 (Becker 1886) (1.81) Mallos blandus 520, 743 Cham. & Gert. 1958 (1.42) Mallos macroilms 600, 900 Bond & Opel! 1997 (1.50) Mallos niveus 340, 540 (O. P.- Camb. 1902) (1.59) Mallos pallldus 400, 6S0 (Banks 1904) (1.70) Mallos pearcal 380, 600 Cham. & Gert 1958 (1.58) Mallos mlans 380, 620 (Chamberlin 1919) (1.63) Mallos gartsehl 640, 900 Bond & Opel! 1997 (1 41) Mallos chambarlini 480. 740 Bond & Opel! 1997 (1.54) Figure 3. — Phylogeny of the 14 known species of Mallos and a representative of its sister group Mexitilla (from Bond & Opell 1997). Following each species is the width of its cribellum and the length of its calamistrum, both in p,m. Ratios of calamistrum length to cribellum width are in paren- theses. Numbers near vertical lines denote the sister groups whose independent contrasts are given in Figure 4. of the combing leg that contacted cribellum spigots would tend to draw fibrils from them and these would become stuck to the leg or catch on the forming cribellar thread, thereby interfering with cribellar thread production. Second, cribellar thread is materially costly to produce (Opell 1997, 1998) and it seems un- likely that a cribellum with an unused lateral region would be retained. The apparent ease with which the cribellum itself is lost is doc- umented by a number of families, genera, and even species pairs (putative sister species) that have both cribellate and ecribellate members (Forster 1970; Forster & Wilton 1973). This study shows that at higher taxonomic levels, there is no uniform relationship be- tween cribellum width and calamistrum length. This suggests that the angle at which a calamisturm passes over a cribellum or the amount of lateral movement of the calamis- trum during a combing stroke differs greatly Figure 4.-— Regression of independent contrast values for cribellum width and calamistrum length for 14 Mallos species and Mexitilla trivittata. Num- bers identify the sister groups in Figure 3 from which these values were computed. among spiders. As noted in the introduction, a variety of morphological and behavioral fac- tors may influence the position and path of the calamistrum. Even among orb-weaving species of the family Uloboridae that support the combing leg in the same manner (Eberhard 1988; Opell unpub. obs. for Waitkera waitakerensis (Chamberlain 1946), Siratoba referena (Muma & Gertsch 1946), Uloborus glomosus (Walckenaer 1837), Octonoba sinensis (Simon 1880)) and share more similar body plans (ab- domen dimensions, leg lengths, and ratios of leg articles; Opell 1979), the ratio of calam- istrum length to cribellum width differs con- siderably. It is only within the genus Mallos that a clade-specific correlation between cal- amistrum length and cribellum width can be demonstrated. Even here this relationship is not exceedingly strong, as it begins to decay when sample size decreases. As comparisons of calamistrum length and cribellum width within the family Uloboridae and among families are based on small sam- ples, it is possible that an increased sample size would establish a significant relationship between these features. However, in compar- isons of other spider features similar phylo- genetic representation has been sufficient to demonstrate significant relationships (Opell 1994a, 1996, 1997, 1998, 1999, in press). Therefore, if there is a general relationship be- OPELL ET AL.— SPIDER CRIBELLA AND CALAMISTRA 47 tween calamistmm and cribellum features, it is weaker than those of other aspects of the phenotype. ACKNOWLEDGMENTS Material was collected during field studies conducted at the Archbold Biological Station, the Center for Energy and Environment Sci- ence’s El Verde field station in Puerto Rico, the Organization for Tropical Studies’ La Sel- va field station in Costa Rica, and the Amer- ican Museum of Natural History’s Southwest- ern Research Station in Arizona, U.S.A. Collecting permits for New Zealand species were granted by the Northland Conservancy Office of New Zealand’s Department of Con- servation and the Works and Services De- partment of the Whangarei District Council. This material is based upon work supported by the National Science Foundation under grant IBN-9417803. LITERATURE CITED Banks, N. 1896. New North American spiders and mites. Trans. American Entomol. Soc., 23:57-77. Banks, N. 1901. Some Arachnida from New Mex- ico. Proc. Acad. Nat. Sci., Philadelphia, 53:581- 590. Banks, N. 1904. Some Arachnida from California. Proc. California Acad. Sci., 3:331-376. Becker, L. 1886. Diagnoses de quelques Arachni- des nouveaux. Ann. Soc. Entomol. Belgique, 30: XXIII-XXVIL Blackwall, J. 1846. Notice of spiders captured by Professor Potter in Canada, with descriptions of such spiders as appear to be new to science. Ann. Mag. Nat. Hist,, 17:30-44, 76-82. Bond, J.E. & B.D. Opell. 1997. Systematics of the spider genera Mallos and Mexitilia (Araneae, Dictynidae). Zool. J. Linn. Soc., 119:389-445. Chamberlain, G. 1946. Revision of the Araneae of New Zealand (Part 2). Rec. Auckland Inst. Mus., 3:85-97. Chamberlin, R.V. 1916. Results of the Yale Peru- vian Expedition of 1911: The Arachnida. Bull. Mus. Comp. Zool., 60:177-299. Chamberlin, R.V. 1919. New Californian spiders. J. Entomol. Zool., Claremont, 12:1-17. Chamberlin, R.V. & W.J. Gertsch. 1958. The spider family Dictynidae in America north of Mexico. Bull. American Mus. Nat. Hist., 116:1-152. Chamberlin, R.V. & W. Ivie. 1935. Miscellaneous new American spiders. Bull. Univ. Utah, 26:1- 79. Chickering, A.M. 1968. The genus Miagrammopes (Araneae, Uloboridae) in Panama and the West Indies, Brev. Mus. Comp. Zool., 289:1-28. Coddington, J.A. 1990. Ontogeny and homology in the male palpus of orb-weaving spiders and their relatives, with comments on phylogeny (Ar- aneoclada: Araneoidea, Deinopoidea). Smith. Contrib. Zool., 496:1-52. Eberhard, W.G. 1988. Combing and sticky silk at- tachment behaviour by cribellate spiders and its taxonomic implications. Bull. British Arachnol. Soc., 7:247-251. Eberhard, W.G. & E Pereira. 1993. Ultrastructure of cribellate silk of nine species in eight families and possible taxonomic implications. (Araneae: Amaurobiidae, Deinopidae, Desidae, Dictynidae, Filistatidae, Hypochilidae, Stiphidiidae, Tengel- lidae). J. Arachnol., 21:161-174. Felsenstein, J. 1985. Phylogenies and the compar- ative method. American Nat., 125:1-15. Forster, R.R. 1970. The spiders of New Zealand, part III: Desidae, Dictynidae, Hahniidae, Amau- robioididae, Nicodamidae. Otago Mus. Bull., 3: 1-184. Forster, R.R. & C.L. Wilton. 1973. The spiders of New Zealand, part IV: Agelenidae, Stiphiidae, Amphinectidae, Amaurobiidae, Neolanidae, Ctenidae, Psechridae. Otago Mus. Bull., 4:1- 309. Gertsch, W.J. 1936. Further diagnosis of new American spiders. American Mus. Nov., 852:1- 27. Gertsch, W.J. 1946. Notes on American spiders of the family Dictynidae. Amererican Mus. Nov., 1319:1-21. Griswold, C.E., J.A. Coddington, N.I. Platnick, & R.R. Forster. 1999. Towards a phylogeny of en- telegyne spiders (Araneae, Entelegynae). J. Ar- achnol.,27(l):53-63. Harvey, PH. & M.D. Pagel. 1991. The Compara- tive Method in Evolutionary Biology. Oxford Univ. Press, New York. Hentz, N.M, 1842. Descriptions and figures of Ar- aneides of the United States. Boston J. Nat. Hist., 4:54-57. Hentz, N.M. 1847. Descriptions and figures of Ar- aneides of the United States. Boston J. Nat. Hist., 5:443-478. Koch, L. 1867. Beschreibungen neuer Arachniden und Myriapoden. Verh. Zool. Bot. Ges. Wien, 17:173-250. Koch, L. 1872. Die Arachniden Australiens, nach der Natur beschrieben und abgebildet. Part 2, Pp. 105-368. Niimberg. Latreille, P.A. 1806. Genera Crustaceorum et In- sectorum, Araneides, Part 1: Araneides, pp. 82- 127. Paris. Mello-Leitao, C.F. 1917. Generos e species novas de Araneidos. Arch. Esc. Sup. Agr, Med. Veter., 1:3-19. Muma, M.M. & W.J. Gertsch. 1964. The spider 48 THE JOURNAL OF ARACHNOLOGY family Uloboridae in North America north of Mexico. American Mus, Nov., 2196:1-43. Opell, B.D. 1979. Revision of the genera and trop- ical American species of the spider family Ulo- boridae. Bull. Mus. Comp. ZooL, 148:433-549. Opell, B.D. 1994a. Factors governing the sticki- ness of cribellar prey capture threads in the spi- der family Uloboridae. J. Morph., 221:111-119. Opell, B.D. 1994b. Increased stickiness of prey capture threads accompanying web reduction in the spider family Uloboridae. Funct. EcoL, 8:85- 90. Opell, B.D. 1995. Ontogenetic changes in cribei- lum spigot number and cribellar prey capture thread stickiness in the spider family Uloboridae. J. Morph., 224:47-56. Opell, B.D. 1996. Functional similarities of spider webs with diverse architectures. American Nat, 148:630-648. Opell, B.D, 1997. The material cost and stickiness of capture threads and the evolution of orb-weav- ing spiders. Biol. J. Linnean Soc., 62:443-458. Opell, B.D. 1998. Economics of spider orb-webs: the benefits of producing adhesive capture thread and of recycling silk. Funct EcoL, 12:613-624. Opell, B.D. 1999. Redesigning spider webs: stick- iness, capture area, and the evolution of modern orb-webs. Evol. EcoL Res. In press. Opell, B.D. In press. Changes in spinning anatomy and thread stickiness associated with the origin of orb-weaving spiders. Biol. J. Linnean Soc. Peters, H.M. 1983. Struktur und Herstellung der FangfMee cribellater Spinnen (Arachnida: Ara- neae). Verhand. Naturwissen. Ver., Hamburg, 26: 241-253. Peters, H.M. 1984. The spinning apparatus of Ulo- boridae in relation to the stracture and construc- tion of capture threads (Arachnida, Araneida), Zoomorph., 104:96-104. Peters, H.M. 1986. Fine structure and function of capture threads. Pp. 187-202, In Ecophysiology of Spiders (W. Nentwig, ed.). Springer Verlag, New York. Peters, H.M. 1992. On the spinning apparatus and structure of the capture threads of Deinopis sub- rufus (Araneae, Deinopidae). Zoomorph., 112: 27-37. Pickard-Cambridge, O. 1902. Arachnida. Aranei- da. Pp. 305-316, In Biologia Centrali- America- na. Purvis, A. & A. Rambaut. 1995. Comparative anal- ysis by independent contrasts (CAIC): an Apple Macintosh application for analysing comparative data. Comp. Appl. Biosci., 11:247-251. Simon, E. 1880. Etudes arachnologiques. lUMe- moire. XVII. Arachnides recueillis aux environs de Pekin par M.V, Collin de Plancy. Ann. Soc. Entomol. France, 10:97-128, Simon, E. 1909, Sur LAraignee Mosquero. Paris Academy of Sciences, 148:736-737. Urquhart, A.T 1892. Descriptions of new species of Araneae, Trans. New Zealand Inst, 25:165- 190, Walckenaer, C.A. 1837. Histoire naturelle des In- sectes. Apteres. Tome I, Pp. 1-682. Paris. Manuscript received 28 September 1998, revised 15 April 1999. 2000. The Journal of Arachnology 28:49-55 CHARACTERIZATION OF LIPOPROTEINS ISOLATED FROM THE HEMOLYMPH OF THE SPIDER LATRODECTUS MIRABILIS (ARANEAE, THERIDIIDAE) Monica Cunningham^ Alda Gonzalez^ and Ricardo Pollero^: ^Instituto de Investigaciones Bioquimicas de La Plata (INIBIOLP), Facultad de Ciencias Medicas, 60 y 120 (1900) La Plata, Argentina; ^CEPAVE. Facultad de Ciencias Naturales y Museo, 1900 La Plata, Argentina ABSTRACT. Two high density lipoprotein fractions (HDLj and HDL2) were isolated from the hemo- lymphatic plasma of the spider Latrodectus mirabilis (Holmberg 1876). For each, the hydrated density, the electrophoresis mobility of the apoproteins, and the lipid classes composition were determined. The HDL, fraction carried 80% of the total plasma lipids, which were predominantly composed of phospho- lipids, free fatty acids, and triacylglycerols. The apoprotein composition of this fraction showed two main bands, of 90 and 103 kDa. The HDL2 fraction was composed primarily of phospholipids, free fatty acids and cholesterol. This fraction contained hemocyanin as the principal apoprotein. When the HDLj fraction was separated into three subfractions, all of them contained hemocyanin, with the main subfraction con- taining the hexameric form of the respiratory pigment. With regard to triacylglycerol transport, lipid and apoprotein compositions and hemocyanin role in the lipid transport, these lipoproteins (HDLj, HDL2) show similarities and differences when compared to the two spider species already studied. Keywords; Lipoproteins, Latrodectus, hemolymph Lipids can not circulate freely in an aque- ous medium due to their hydrophobicity. Not- withstanding, they are transported by hemo- lymph from the sites of uptake or synthesis to the sites of storage and usage via water-solu- ble lipoproteins. Lipid circulation systems in invertebrates have been studied only in the phyla Arthropoda and Mollusca. Although the mechanisms of lipid circulation are well- known in arthropods such as insects and crus- taceans, in arachnids there is little available information on plasma lipoproteins. Lipopro- teins of high density have been detected in the hemolymph of spiders, scorpions, solpugids, and mites. According to their apoprotein com- ponents, lipoproteins in spiders, scorpions, and solpugids showed similar characteristics to those of insect lipophorins (Haunerland & Bowers 1989). In spiders, the lipoprotein lipid composition was extensively studied by Hau- nerland & Bowers (1987) in Eurypelma cali~ fornicum (Ausserer 1871) (Theraphosidae) and Polybetes pythagoricus (Holmberg 1874) (Heteropodidae) by Cunningham et al. (1994). In E. californicum, the high content of dia- cylglycerols and phospholipids also resembles the composition of insect lipophorins. In P. pythagoricus, three plasma lipoproteins were detected and characterized. One of them was of high density and evidenced similar apopro- teins to the ones called lipophorins. In con- trast, its lipid composition was rather differ- ent, containing a large amount of phospholipids and triacylglycerols. The other two lipoprotein fractions of P. pythagoricus, of high and very high density, also contained phospholipids and triacylglycerols as major lipids; but hemocyanin was the predominant apolipoprotein (Cunningham & Pollero 1996). The differences found in the lipid and apo- protein compositions of the two previously studied spiders led us to investigate the lipo- proteins in a third species, Latrodectus mira- bilis (Theridiidae), a widely distributed spe- cies. The literature reports studies on the biology and ecology (Gonzalez 1981; Estevez et al. 1984) and venom components (Flo et al. 1991). There is no available study on the bio- chemical and physiological aspects of the lip- id circulation. This study describes the com- position of the lipid and protein moities of two plasma lipoproteins isolated from the L. mir- abilis hemolymph. The role of triacylglycerols as circulating energetic lipids, of hemocyanin 49 50 THE JOURNAL OF ARACHNOLOGY as apolipoprotein, as well as composition sim- ilarities and differences between these lipo- proteins and those of other spider species, are discussed. METHODS Hemolymph collection and lipoprotein separation.— We collected adult females of Latrodectus mirabilis (deposited in the Mu- seum of Natural Sciences, La Plata) in sum- mer from the hills of Sierra de la Ventana, province of Buenos Aires, Argentina. After the legs were severed from the body, the spi- ders were placed in tubes and centrifuged at low speed in order to obtain hemolymph. Plasma was centrifuged in a gradient den- sity on 3 ml NaBr ""1.21 g/ml, with Trasylol as protease inhibitor, at 178,000 G for 22 hours in a Beckman L8 TOM centrifuge with a SW60 Ti rotor. As the density of the spider plasma was 1.006 g/ml, a saline solution of the same density was ran simultaneously as blank. The total volume of the tubes was frac- tionated from top to bottom into 0.3 ml frac- tions. The density and total proteins in each fraction were monitored by refractometry and light absorption at 280 nm, respectively. Lipid extraction and analysis,— Total lip- ids from the lipoprotein fractions were ex- tracted with chloroform/methanol (Bligh & Dyer 1959). Total lipids were analyzed on Merck high performance thin-layer chroma- tography (HP-TLC) plates. Hydrocarbons were separated from other neutral lipids by development in hexane-benzene (70:30 v/v). Polar lipids were resolved by developing the plates in chloroform/methanol/acetic acid/wa- ter (65:25:4:4 v/v) and hexane/diethyl ether/ acetic acid (80:20:1.5 v/v) for neutral lipids. Appropriate standards were used. Spots were visualized with 12 vapors and identified by comparison with known standards. The quantitative determination of the lipid classes was performed using a thin-layer chro- matograph coupled to a flame ionization de- tector (TLC-FID) system. A full description of this technique was given by Ackman (1990). FID scans were performed on an la- troscan TH-10 analyzer (latron Laboratories, Japan). The development solvent systems used were: hexane/benzene (70:30 v/v), ben- zene/chloroform/formic acid (70:25:2 v/v) and chloroform/methanol/water (70:25:3 v/v). Lipid classes were quantified by comparison with known amounts of standards ran under the same conditions and using monoacylgly- cerol as internal standard. Total lipids were calculated by the summation of individual lip- id weights. Gel permeation chromatograpfiy.—A very high density lipoprotein fraction separat- ed from the gradient was analyzed under na- tive conditions by preparative high-pressure liquid chromatography on a Superdex 200 HR 10/30 column (Pharmacia, Uppsala, Sweden) using 0.1 M Tris-HCI (pH 8.0), containing 10 mM CaCl2 and MgCl2, at the flow rate of 0.4 ml/min. Proteins were detected at 280 nm. Li- poprotein subfractions were eluted. The col- umn was calibrated for molecular weight us- ing thyroglobulie, ferritin, catalase, bovine serum albumin (BSA) and ribonuclease A (Pharmacia, Sweden) as protein markers. Characterization of apoproteins.— Total protein concentration in each fraction isolated from the density gradient was measured col- orimetrically (Lowry et al. 1951). These frac- tions and subfractions isolated by HPLC, were extensively dialyzed against 10 mM Tris-HCI (pH 6.8) and analyzed by electrophoresis un- der dissociating and native conditions. Sodi- um dodecyl sulfate-polyacrylamide gel elec- trophoresis (SDS-PAGE) was carried out in both, 8% continuous slab gels overlayered with 4% stacking gels and in gradients of 4- 20% gels (Laemmli 1970) in a mini-slab elec- trophoresis unit (8 X 10 cm). The resolving gel buffer was 0.375 M Tris-HCI and the stacking gel buffer was 0.125 M Tris-HCI. The electrode buffer contained Tris-glycine 0.025 M Tris, 0.192 M glycine (pH 8.3). Pro- teins were visualized by staining with Coom- assie Brilliant Blue. Molecular weight stan- dards (HMW, Pharmacia, Uppsala, Sweden and Markerkit, Sigma Chemical Co., St. Lou- is, Missouri) were ran in parallel lines. The presence of hemocyanin in the frac- tions was monitored by spectrophotometric scans from 200-700 nm, before and after sample treatment with 0.2 M KCN solution (Nickerson & Van Holde 1971). A DW-2000 UV-Vis spectrophotometer SLM Aminco was used. RESULTS Isolation of plasma lipoprotein frac- tions.-—After plasma centrifugation in density gradients, two colored bands appeared: a CUNNINGHAM ET M^.—LATRODECTUS HEMOLYMPH LIPOPROTEINS 51 Figure L— Total protein (absorbance at 280 nm) and density distribution in plasma fractions of Lat- rodectus mirabilis. Plasma was centrifuged in a NaBr gradient and fractionated. brownish one (HDLi) and a grey one (HDL2), whose densities were 1.13 g/ml and 1.19-1.20 g/ml, respectively. Measurements of absor- bance at 280 nm performed in each fraction from gradients showed a protein profile with two maxima, one of them (the smallest) cor- responded to HDLj and the major one to HDL2 (Fig. 1). Plasma fractions out of colored bands showed relatively low protein and no lipid concentrations. Both colored fractions were isolated and characterized separately. Hemocyanin was present in the HDL2 frac- Figure 2. — ^Hemocyanin characterization. Spec- trophotometric scans of hemocyanin from HDL2, before and after sample treatment with KCN. Bro- ken line = without KCN; solid line — with KCN. tion. The respiratory pigment was identified by modification of its characteristic absorption spectrum; the absorption band of 340 nm dis- appeared when samples were treated with KCN solution (Fig. 2). Lipid and protein characterization of HDL,.— The HDLj carries 80.4% of the total plasma lipids. Lipids in this fraction were an- alyzed in their component classes. Phospha- tidylcholine, phosphatidylethanolamine, free fatty acids, triacylglycerols, cholesterol and hydrocarbons were identified qualitatively us- ing HP-TLC. The quantitative lipid composition, deter- mined by TLC-FID, is shown in Table 1. The predominant lipids were phospholipids (35%) and free fatty acids (33%). Triacylglycerols Table 1 . — Composition of HDLj and HDL2 isolated from plasma of Latrodectus mirabilis. The lipoproteins were isolated by ultracentrifugation in density gradient. Lipids were identified after separation by HP-TLC and quantified by TLC-FID. Proteins were measured by colorimetry. Results are the average of three deter- minations (100 animals) ± SD. Data are expressed as weight percent of lipids as determined by TLC-FID. Component HDL, HDLj Lipid classes (percent weight/weight) Hydrocarbons 4.0 ± 0.3 14.1 ± 2.3 Triacylglycerols 24.1 ± 0.8 8.3 ± 2.1 Free fatty acids 33.0 ± 1.3 28.4 ± 4.2 Cholesterol 4.2 ± 0.3 20.1 ± 2.7 Diacylglycerols Traces Traces Phosphatidyl ethanolamine 3.6 ± 0.2 4.0 ± 0.5 Phosphatidyl choline 31.1 ± 0.3 25.1 ± 2.8 Total lipids (mg/ml hemolymph) 1.23 (20.3%) 0.3 (1.0%) Total proteins (mg/ml hemolymph) 4.83 (79.7%) 31 6 (99.0%) 52 Kd. 1 2 Figure 3.= — SDS-PAGE analysis (4”23% acryl- amide) of HDL, apoproteins from Latrodectus mir- abilis hemolymph, Kd: Molecular weights of stan- dard proteins expressed in kilodaltons. Lane 1: HDL, from L. mirabilis. Lane 2: Molecular weight standards (Kd). were quite abundant in this fraction whereas hydrocarbons and cholesterol were found in a low proportion. Traces of diacylglycerols were also detected. Other aliquots of HDL| were used to ana- lyze the constituent apoproteins by electro- phoresis. Figure 3 shows those results ob- tained from the electrophoretic analysis performed under dissociating conditions (SDS-PAGE). Among other proteins, two sharp bands of 90 and 103 kDa, respectively, were observed as the major HDLi apopro- teins. Lipid and protein characterization of HDL2.-— The HDL2 lipids were analyzed quantitatively and qualitatively. This lipopro- tein fraction carries 19.6% of total plasma lip- ids. The same lipid classes as those belonging to HDLj fraction were identified by HP-TLC and some differences were found in their rel- THE JOURNAL OF ARACHNOLOGY o’ Figure 4,— Elution profile from HPLC of HDL2 isolated from L. mirabilis plasma. Subfractions I, II and III were collected and analyzed separately. ative percentages. Phospholipids (29%) and free fatty acids (28%) were the predominant lipids, followed by cholesterol, hydrocarbons and minor quantities of triacylglycerols (Table 1). Aliquots of HDL2 were analyzed by HPLC under native conditions using columns of mo- lecular exclusion (Fig. 4). Three subfractions of Mr 440 kDa, 121 kDa and about 70 kDa, respectively, were found. They were eluted from the column, collected and analyzed by electrophoresis under denaturing conditions. Figure 5 (SDS-PAGE) shows two proteins of 76 and 67 kDa, respectively, in the three sub- fractions isolated from HDL2. DISCUSSION Centrifugation in a density gradient was ef- fective in separating two well-defined bands from Latrodectus mirabilis plasma which cor- responded to the high density lipoproteins HDLi and HDL2. HDLi has a density similar to that of lipophorins isolated from plasma of Eurypelma californicum (Haunerland & Bow- CUNNINGHAM ET AU—LATRODECTUS HEMOLYMPH LIPOPROTEINS 53 Figure 5. — SDS-PAGE analysis (4-23% acryl- amide) of HPLC-fractionated HDL2 from Latrodec- tus mirabilis hemolymph. Kd: Molecular weights of standard proteins expressed in kilodaltons. Lane 1: Molecular weight standards (Kd). Lane 2; HDLj subfraction 1. Lane 3: HDL2 subfraction 11. Lane 4: HDL2 subfraction III. ers 1987) and of Polybetes pythagoricus (Cunningham et al. 1994), which are the only arachnids which have been studied in detail. Its density is also similar to that of lipophorins found in insects (Chino et al. 1981). The HDL2, though its density is greater than the HDLj fraction, is also a high density lipopro- tein, and so it can be compared to the HDL previously isolated from plasma of P. pytha- goricus (Cunningham & Pollero 1996). HDLj is the main lipid carrier fraction in L. mirabilis hemolymph since more than three- fourths of the total plasma lipids are associ- ated to it. This quantitative importance in lipid transport locates it at the same level as that of E. californicum and above P. pythagoricus li- pophorin which only carries about 30% of cir- culating lipids. In contrast, when lipid classes found in this lipoprotein fraction are com- pared, similarities to P. pythagoricus and dif- ferences to E. californicum are evident. Phos- pholipids and fatty acids are the predominant lipids; and, as in P. pythagoricus lipophorin, triacylglycerols are the most abundant neutral lipids. This fact indicates that triacylglycerols together with free fatty acids are the main cir- culating energetic lipids in this species, in contrast to E. californicum and insects where the presence of large amounts of diacylgly- cerols characterizes the lipophorins. The protein moiety of HDLi is composed of two principal polypeptides with a molecu- lar weight of 90 and 103 kDa, respectively. This also differs when compared to E. cali- fornicum and P. pythagoricus apolipophorins, and to those found in other arachnids whose protein moieties have been studied (Hauner- land & Bowers 1989). In all these cases as well as in insects, lipophorin particles contain apoproteins of 80 and 250 kDa and a total weight of about 500 kDa. In short, due to its composition, the HDLi of L. mirabilis is sig- nificantly different when compared to HDLs of the same density in other invertebrates that are taxonomically close to it. For this reason, we think it shouldn’t be named lipophorin. In L. mirabilis, the HDL2 could play a sec- ondary role in the hemolymph transport of lip- ids due to the fact that the lipids associated to this lipoprotein are lesser that those ones bound to the HDLi. Nevertheless, its lipid composition, with relatively high amounts of hydrocarbons and cholesterol, suggests that HDL2 could be specialized in the transport of these lipid classes. Although this lipoprotein particle differs from the HDL of P. pythago- ricus not only in the lipid/protein ratio but also in the lipid classes it transports, both of them carry triacylglycerols but no diacylgly- cerols as the main neutral acylglycerides. The electrophoretic analysis of HDL2 under dissociating conditions, shows protein bands with molecular weights similar to the hemo- cyanin monomers found in other spiders (Schneider et al. 1977; Lamy et al. 1979; Markl 1986). The removal of copper by KCN treatment confirms this identification. Al- though we tried to stabilize the hemocyanin, it is very likely that, when handling the sam- ples, there would have been some dissociation of HDL2 native particles; and, consequently, subfractions of different size would have ap- peared after gel permeation chromatography. The loss of native conformation of hemocya- nin could be the result of changes in the pH, the divalent cation concentration during the processing samples, or due to the use of NaBr in the centrifugation procedure (van Holde & Miller 1986; Hepskovits & Villanueva 1986; Herskovits et al. 1991). Undoubtedly hemocyanin plays an apoli- poprotein role since it is part of this lipopro- tein particle as a principal protein. This func- tion of hemocyanin in spiders regarding the lipid transport, in addition to its classical role as respiratory pigment, has been recently re- 54 THE JOURNAL OF ARACHNOLOGY ported for P. pythagoricus plasma (Cuening- ham & Pollero 1996) where, however, other polypeptides associated with hemocyanin were also found. In this study, the three sub- fractions isolated from L. mirabiiis HDL2 un- der dissociating conditions only yielded he- mocyanin monomers. This corroborates the apoprotein function of hemocyanin in this li- poprotein. This apolipoprotein role of hemo- cyanin is not a constant in spiders, since no associated lipids could be detected in tarantula hemocyanin. A similar finding has been re- ported for molluscs where the hemocyanin of the cephalopod Octopus tehuelchus transports lipids (Heras & Pollero 1992), while that of the gasteropod Ampullaria canaiiculata does not (Garin & Pollero 1995). In P. pythagoricus plasma, we have char- acterized a third lipoprotein of very high den- sity which is the main carrier of circulating lipids, and which contains hemocyanin as the principal apoprotein (Cunningham & Pollero 1996). The existence of a VHDL has also been reported for Eurypelma californicum (Haunerland & Bowers 1989). In this case, however, it was a lipoprotein without hemo- cyanin which played a secondary role in lipid transport. In L. mirabiiis no particle with VHDL characteristics has been detected. In brief, L. mirabiiis contains two plasma lipoproteins; but their lipid and protein com- positions share only a few features with the hemolymph lipoproteins already described other spider species. Such differences in num- ber and composition of plasma lipoproteins in taxonomically close organisms make gener- alization difficult. ACKNOWLEDGMENTS This research was supported by grants from CONICET and CIC. BA, Argentina and Efa- mol Research Institute, Canada. M.C, is Fel- low of the CIC. BA; A.G. and R.P. are mem- bers of Carrera del Investigador Cientifico of the CONICET and CIC. BA, respectively. LITERATURE CITED Ackmao, R.G., C.A. McLeod & A.K. Banerjee. 1990. An Overview of analyses by Chromarod- latroscan TLC~FID. J. Planar Chromatogr., 3: 450=^489. Bligh, E.G, & WJ. Dyer. 1959. A rapid method of total lipid extraction and purification. Canadian J. Biochem. Physiol,, 37:911-917. Chino, H., R.G.H. Downer, G.R. Wyatt & L.L Gil- bert. 1981. Lipophorins, a major class of lipo- protein of insect hemolymph. Insect Biochem., 11:491. Cunningham, M., RJ. Pollero & A. Gonzalez. 1994, Lipid circulation in spiders. Transport of phospholipids, free acids and triacylglycerols as the major lipis classes by a high density lipopro- tein fraction isolated from plasma of Polybetes pythagoricus. Comp. Biochem. Physiol., 109B: 333-338. Cunningham, M. & R.J. Pollero. 1996. Character- ization of lipoprotein fractions with high content of hemocyanin in the hemolymphatic plasma of Polybetes pythagoricus. J. Exp. Zool., 274:275- 280. Estevez, A.L., A. GonzMez & J.A. Schnack. 1984. Estadisticas vitales en especies argentinas del ge- nero Latrodectus Walckenaer (Araneae, Theridi- idae). ll~Latrodectus antheratus (Badcock), Lat- rodectus coralUnus Abalos y Latrodectus diaguita Carcavailo. Physis, Sec. C, 42, 102: 2937. Flo, J., A. GonzMez & E. Massouh. 1991. Estudios electroforeticos e inmunologicos de especies del genero latrodectus Walckenaer. I-L, mirabiiis (Holmberg), L. antheratus (Badcock) y L. cor- allinus Abalos (Araneae, Theridiidae). Rev. Brasileira Entomol., 35(4):67 1-678. Garin, C.F. & RJ. Pollero. 1995. Isolation and characterization of a low-density lipoprotein fraction from plasma of the aquatic snail Ampul- laria canaiiculata. Comp. Biochem. Physiol., 111B:147-150. Gonzalez, A. 1981. Desarrollo postembrionario de Latrodectus mirabiiis, Latrodectus antheratus y Latrodectus coralUnus (Araneae, Theridiidae), Physis, 39(97):83-9L Haunerland, N.H. & W.S. Bowers. 1987. Lipopro- teins in the hemolymph of the tarantula, Eury- pelma californicum. Comp. Biochem. Physiol., 86B:57 1-574. Haunerland, N.H. & W.S. Bowers. 1989. Compar- ative studies on arthropod lipoproteins. Comp. Biochem. Physiol., 92B:137~14L Heras, H. & RJ, Pollero. 1992. Hemocyanin as an apolipoprotein in the hemolymph of the cepha- lopod Octopus tehuelchus. Biochim. Biophys. Acta., 1125:245-250. Herskovits, T.T. & G.B. Villanueva. 1986. Light- scattering investigation of the subunit structure and dissociation of Octopoda hemocyanins. Bio- chemistry, 25:931-939. Herskovits, T.T., C.J, Cousins & M.G. Hamilton. 1991. The stabilizing influence of divalent ions and Na+ on the di-decameric structure of Yoldia Umatula hemocyanin. Biochim. Biophys. Acta,, 1076:71-78. Laemmli, V.K. 1970. Cleavage of structural pro- CUNNINGHAM ET AI..—LATRODECTUS HEMOLYMPH LIPOPROTEINS 55 teins during the assembly of the head of bacte- riophage T4. Nature, 227:680-685. Lamy, J., J. Lamy, J. Weill, J. Markl, HJ. Schneider & B. Linzen. 1979. Hemocyanins in spiders, VIL Immunological comparison of the subunits of Eurypelma californicum. Hoppe-Seyler’s Z. Physiol. Chem., 360:889-895. Lowry, O.M., A.L. Rosebrough, A.L, Farr & R. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem., 193:265- 275. Markl, J. 1986. Evolution and function of struc- turally diverse subunits in the respiratory protein hemocyanin from arthropods. Biol. Bull., 171: 90-115. Nickerson, K.W. & K.E. Van Holde. 1971. A com- parison of molluscan and arthropod hemocyanin. I. Circular dichroism and absorption spectra. Comp. Biochem. Physiol., 39B:855-872. Schneider, H.J., J. Markl, W. Schartau & B. Linzen. 1977. Hemocyanins in spiders. IV. Subunits het- erogeneity of Eurypelma (Dugesiella) hemocya- nin and separation of polypeptide chains. Hoppe- Seyler’s Z. Physiol. Chem., 358:1133-1141. Van Holde, K.E. & K.I. Miller. 1986. P. 417, In Invertebrate Oxygen Carriers (Int. Conf.). (B. Lindzen, ed.). Springer, Berlin. Manuscript received 1 December 1997, revised 20 October 1998. 2000. The Journal of Arachnology 28:56-60 DIET-INDUCED AND MORPHOLOGICAL COLOR CHANGES IN JUVENILE CRAB SPIDERS (ARANEAE, THOMISIDAE) Victoria R. Schmalhofer^: Graduate Program in Ecology and Evolution, Department of Ecology, Evolution, and Natural Resources, Rutgers University, Cook College, 14 College Farm Road, New Brunswick, New Jersey 08901-8551 USA ABSTRACT. The effect of dietary pigments on abdominal color of juvenile spiders was examined in the laboratory using the flower-dwelling crab spiders Misumenops asperatus (Hentz 1847), Misumenoides formosipes (Walckenaer 1837), and Misumena vatia (Clerck 1757) (Thomisidae). Because these species lack hypodermal chromes, ingested prey pigments may show through the epidermis and affect opistho- somal coloration. Diet-induced color changes were restricted to the opisthosoma, and all three spider species responded similarly to dietary pigments. Opisthosomas of instars 2-4 fed red-eyed fruit flies turned pink, and the pink color faded back to the normal white over a period of 4-6 days. Opisthosomas of instars 5-7 fed red-eyed fruit flies remained white, as did opisthosomas of all instars fed white-eyed fruit flies (controls). In a field population of M. asperatus, 82% of spiders in July (instar 2), 93% of spiders in August (instars 3-4), and 8% of spiders in September (instar 5) had pink, orange, or brown opistho- somas. Yellow juveniles were also seen: 5% and 57% of M. asperatus observed in August and September, respectively, were yellow. Yellow juvenile M. formosipes were observed in the field as well. The yellow color did not result from dietary pigments, but was, rather, a morphological color change and included the prosoma and limbs, as well as the opisthosoma. Keywords: Flower spiders, opisthosoma, prey pigments, size-dependent effect The ability of certain species within the family Thomisidae (crab spiders) to undergo a reversible color change depending on their environmental substrate, a process referred to as a morphological color change (Holl 1987), has provoked interest among naturalists since the late nineteenth century (Angus 1882; Packard 1905; Gadeau de Kerville 1907; Ga- britschevsky 1927; Gertsch 1939; Weigel 1941). Most investigations of morphological color changes among thomisids have focused on the goldenrod spider, Misumena vatia (Clerck 1757) (e.g., Packard 1905; Gabrit- schevsky 1927; Millot 1926; Weigel 1941), and the ability to change color has been attri- buted only to adult females (Gabritschevsky 1927). Misumena vatia is typically white, but turns yellow when placed on a yellow substrate. Be- cause this species lacks hypodermal chromes and has a translucent cuticle, reflection of white light from guanine crystals in the intes- tinal diverticula causes M. vatia to appear * Current address: Arachnological Research and Consulting, 1131 South Ninth Street, South Plain- field, New Jersey 07080 USA white (Millot 1926; Weigel 1941). Under the stimulus of reflected yellow light (Gabrit- schevsky 1927; Weigel 1941), a yellow pig- ment is released into the hypoderm (Weigel 1941), and the yellow color becomes more in- tense the longer a spider remains on a yellow substrate (Packard 1905; pers. obs.). Morpho- logical color changes involve a spider’s entire body: prosoma, opisthosoma, and limbs take on a yellow hue. Similar morphological color changes have been reported in other thomisid species, including Misumenoides formosipes (Walckenaer 1837) and Misumenops aspera- tus (Hentz 1847) (Gertsch 1939; Schmalhofer 1996). Having a colorless, translucent integument has an interesting side-effect on juvenile flow- er-dwelling crab spiders: ingested material that is strongly pigmented may show through the epidermis, changing the color of a spider- ling’s opisthosoma. In the field, juvenile crab spiders having pink, orange, brown, green, yellow, or white opisthosomas have been ob- served (Schmalhofer 1996). Peck & Whit- comb (1968) observed similar diet-induced color changes in the clubionid Cheiracan- thium inclusum (Hentz 1847), a pale yellow 56 SCHMALHOFER— CRAB SPIDER COLOR CHANGES 57 spider with a transparent integument. They noted that spiders turned green and pink when fed pyralid larvae and red«eyed Drosophila, respectively; and they obtained a variety of opisthosomal colors by feeding spiders an ar- tificial diet containing dye (Peck & Whitcomb 1968). No information was provided, howev- er, concerning how long color changes lasted, their frequency of occurrence under natural conditions, or the instars affected. Using flow- er-dwelling crab spiders, I performed a labo- ratory experiment to determine the duration of diet-induced color changes and the instars af- fected. Field observations were also conduct- ed to determine the frequency of occurrence of diet-induced color changes in a natural crab spider population. METHODS Effects of Drosophila eye pigments on op- isthosomal color of juvenile crab spiders.— The effect of Drosophila eye pigments on op- isthosomal color of juvenile crab spiders was examined using M. asperatus, M. formosipes, and M. vatia. Adults of the three species are seasonally separated. Misumenops asperatus matures and females lay a single egg sac in spring, and spiderlings emerge in early sum- mer and overwinter as late instar juveniles (pers. obs.). Misumenoides formosipes ma- tures in midsummer, females produce a single egg sac in late summer or early autumn, and spiderlings generally overwinter in the egg sac (pers. obs.). Misumena vatia matures and fe- males produce a single egg sac in early-to- midsummer, and spiderlings emerge in late summer and overwinter as middle-instar ju- veniles (Fritz & Morse 1985). These species have seven juvenile instars, the first of which is spent in the egg sac (Gertsch 1939). Misumenops asperatus and M. formosipes were reared from egg sacs produced by fe- males collected in Middlesex and Somerset Counties, New Jersey. Adult specimens of M. asperatus and M. formosipes have been de- posited at the American Museum of Natural History. I have never found M. vatia in New Jersey, although the species is recorded as oc- curring in the state (Gertsch 1939). Misumena vatia used in the laboratory experiment orig- inated in Lincoln County, Maine, and egg sacs were provided by D. Morse. Egg sacs were maintained at room ambient temperature (25- 30 °C); and, after emergence, second instar spiderlings were placed in separate 4 dram shell vials with cotton plugs. Spiders were starved for seven days prior to a feeding trial to ensure that their guts were empty (Ander- son 1970; Nakamura 1987). During a feeding trial, a spider was supplied with either 10 red- eyed fruit flies (experimental group) or 10 white-eyed fruit flies (control group). Any flies not consumed after five hours were re- moved. After feeding, a spider’s opisthosomal color was subjectively categorized as bright pink, moderately pink, pale pink, or white. The number of days required for a spider’s opisthosoma to return to the normal white col- or was also noted. This protocol was repeated during each juvenile instar (instars 2-1) for each species. The number of spiders used in a feeding trial varied with species and instar (n == 6-20). Natural occurrence of juvenile Misumen- ops asperatus having colored opisthoso- mas.— Field observations focused on M. as- peratus and took place in a 2.8 ha^ field adjacent to the Busch campus of Rutgers Uni- versity in Middlesex County, New Jersey. I tagged 250 inflorescences each of Achillea millefolium (yarrow), Daucus carota (Queen Anne’s lace), and Solidago spp. (goldenrod), which were plant species commonly used by M. asperatus (Schmalhofer 1996). Solidago was the dominant plant species in the field, occupying approximately 76% of the area and having a density of 55.4 stems per m^. Achil- lea patches were interspersed among the Sol- idago and covered approximately 6% of the area. Achillea had a density of 3 1 .7 stems per m^. Daucus, with a density of 16.6 stems per m^, occurred at the field perimeter and cov- ered approximately 14% of the area. Bloom- ing in Achillea, Daucus, and Solidago oc- curred sequentially over the course of the summer, and flowering phenology in the three species showed little overlap (pers. obs.). Flowering in Achillea occurred from early June to mid-July, flowering in Daucus oc- curred from mid-July to mid-August, and flowering in Solidago occurred from mid- Au- gust through September. Over seven consec- utive days at the beginning of July, August, and September, I made daily surveys of tagged Achillea, Daucus, and Solidago, re- spectively. Observations occurred between 0900-1200 h, and I recorded the number of spiders per inflorescence and spider color. 58 THE JOURNAL OF ARACHNOLOGY RESULTS Effects of Drosophila eye pigments on op- isthosomal color of ju¥eeile crab spiders,-™ Misumenops asperatus, M. formosipes, and M. vatia responded similarly to Drosophila eye pigments. When fed red-eyed fruit flies, only a spider’s opisthosoma changed color: proso- ma and legs were unaffected by dietary pig- ments. I found that the opisthosomas of instars 2-4 fed red-eyed Drosophila turned pink, and the pink color slowly faded to the normal white over a period of 4-6 days. Intensity and duration of the color change varied with age of the spider. Second instar spiders turned bright pink, while older spiders took on a pale-to-moderate shade of pink. Intensity of opisthosomal color in instars 3~4 also seemed to vary with the number of Drosophila con- sumed: spiders capturing a single fly turning pale pink, while those capturing multiple prey (2-3 flies) took on a darker hue. Few spiders captured more than one fly. Opisthosomal col- or of older instars returned to normal more quickly (4 days) than did that of younger spi- ders (6 days). Opisthosomal color of instars 5-7 was not affected by Drosophila eye-pig- ments, regardless of the number of flies con- sumed. Opisthosomas of all spiders in the control groups fed white-eyed Drosophila re- mained white. Anecdotal observations indicated that the intensity and duration of dietary color changes and instars affected were also influenced by the causative pigment. For instance, the op- isthosoma of an instar 5 M asperatus that consumed a blood-fed mosquito turned dark brown, and the color faded over a six-day pe- riod. Opisthosomas of instars 4-5 of M. as- peratus found in the field feeding on uniden- tified green hemipterans turned brilliant green, but returned to normal after only two days. In all three species, spider size changed by more than an order of magnitude during the juvenile period. Mass of instar 2 spiders was less than 1 mg, while average mass of instar 7 (penultimate) female spiders was much greater: 48 mg (M. vatia, calculated from Fritz & Morse 1985; Morse 1988;' Morse & Ste- phens 1996), 42 mg (M. formosipes), and 24 mg (M. asperatus). Natural occurrence of juvenile Misumen- ops asperatus having colored opisthoso- nias,~-“The proportion of spiders showing di- etary color changes was very high in July and August (Table 1), and pink or orange opistho- somas were the most commonly seen varia- tions. The yellow color observed in juvenile M. asperatus in August and September was not dietarily induced, but was, rather, a mor- phological color change like that described for adult spiders (see introduction). The effects of dietarily acquired pigments were restricted to the opisthosoma, but yellow juveniles were fully colored; prosoma and limbs, as well as the opisthosoma, were yellow. Both male and female spiderlings were observed to turn yel- low. Juvenile M. formosipes also proved ca- pable of undergoing a morphological color change; 8% of juveniles seen in July (instars 5-6) and 50% of juvenile females seen in Au- gust (instar 7) were yellow. I have also ob- served yellow M. formosipes in sweepnet samples collected earlier in the season (May and June). The seasonal increase in the pro- portions of yellow juveniles in both M. as- peratus and M. formosipes populations reflect- ed an increase in the availability of plant species with yellow flowers (predominantly SoUdago) during the course of the summer. Spider position on inflorescences varied with plant species. On Achillea, most M. as- peratus were found on the underside of inflo- rescences; on Daucus, spiders occurred with similar frequencies on the upper surface and the underside of inflorescences; and on Soii- dago, most spiders wedged themselves be- tween the individual flowers comprising an in- florescence, Having a colored opisthosoma did not appear to influence spider position on inflorescences of any of the plant species ex- amined. This observation, however, was not quantified, DISCUSSION Although opisthosomal color provides some clues as to a juvenile crab spider’s recent feeding history, opisthosomal color should not be used as a means of categorizing juveniles in field populations as hungry or satiated. Too many variables affect opisthosomal color (e.g., number of prey ingested, spider age, time since ingestion, causative pigment, etc.) to make opisthosomal color a reliable indica- tor of hunger status. Also, many prey types captured by flower-dwelling crab spiders lack strong pigments and, thus, would not affect spider color. SCHMALHOFER-~-CRAB SPIDER COLOR CHANGES 59 Table 1.- — Proportions of Misumenops asperatus of various colors seen during the summer months in central New Jersey. Values presented are averaged over the seven days of observations each month. Spider densities are presented as mean number of spiders per inflorescence (±1 SD): 250 inflorescences of each plant species were surveyed. Flower color is indicated below each plant species. Spider colors marked with an * are diet-induced. Spider Instar density Spider color Month Plant species White Yellow Pink* Orange* Brown* July Achillea millefolium (white) 2 0.09 (0.06) 0.18 0.00 0.82 0.00 0.00 August Daucus carota (white) 3-4 0.83 (0.11) 0.01 0.05 0.50 0.38 0.05 September Solidago spp. (yellow) 5 0.07 (0.03) 0.35 0.57 0.01 0.07 0.00 The effect of dietary pigments on flower- dwelling crab spiders appears to be limited by spider size: smaller (younger) juveniles show the effects of prey pigments, while larger (old- er) juveniles are generally unaffected. Like older juveniles, adult females are unaffected by prey pigments (pers. obs.): mature female spiders fed red-eyed fruit flies ad libitum in the laboratory never displayed opisthosomal color changes, nor have I ever observed adult females in the field to be affected by pigments of ingested prey. In contrast to female crab spiders, adult males are small, typically 5 mg or less (Morse & Stephens 1996; pers. obs.). Because adult M. vatia and M. asperatus males largely lack opisthosomal chromes, ab- dominal color in males of these species has the potential to be affected by diet. However, this seems to be a rare occurrence. During eight years of field research, I have observed only a single adult male M. asperatus showing dietary pigments. The opisthosomal hypoderm of adult male M. formas ipes contains yellow chromes, which would obscure any ingested pigments. Thus, mature male M. formosipes are not subject to diet-induced color changes. Tlie differential effect of prey pigments on younger vj. older spiders, as seen in the lab- oratory experiment, can be explained by gut volume and feeding habits. Crab spiders begin feeding from their prey’s head (Pollard 1989, 1993; pers. obs.); therefore, eye pigments are ingested early during feeding. Compared to older (larger) spiders, younger (smaller) indi- viduals have correspondingly small gut capac- ities and their smaller stomach muscles prob- ably exert less force during feeding (this was not tested). Consequently, younger spiders ex- tract less material from a given prey item than do older spiders, and eye pigments compose a larger fraction of the ingested food. The ten- dency of crab spiders to discard one prey item before all the available material has been ex- tracted and to begin feeding on a new prey item when prey is abundant (Pollard 1989) may have enhanced the effect of Drosophila eye pigments on opisthosomal color of youn- ger juveniles. When offered an abundance of prey, Pollard (1989) found that crab spiders discarded the original prey item after a period of time corresponding to the length of time spent feeding from the head when only one prey item was provided. Morphological color changes in crab spi- ders are erroneously described as being re- stricted to adult female spiders (Gabritschev- sky 1927; Gertsch 1939; Hinton 1976; Holl 1987). This assertion is based on Gabritschev- sky’s (1927) experiments with laboratory- reared M. vatia, and does not hold true for M. asperatus, M. formosipes, or natural popula- tions of M. vatia. In the field, juvenile M. va- tia have been observed to undergo morpho- logical color changes (D. Morse pers. commun.). It is possible that Gabritschevsky’s results were due to the restricted diet {Dro- sophila) given to the spiderlings or to some other difference between the laboratory and field environment (e.g., light intensity, sub- strate character). Light quality or intensity in particular may be important in effecting mor- phological color changes: compared to color changes occurring under natural conditions, color changes occurring under artificial light- ing take longer to complete and a paler yellow hue results (pers. obs.). 60 THE JOURNAL OF ARACHNOLOGY The ability to turn yellow has obvious ben- efits for juvenile (and adult) crab spiders. By enhancing crypsis on yellow flowers, yellow spiders are less likely to be detected by pred- ators or prey. This capability would be partic- ularly useful for juvenile spiders if a large portion of the juvenile period coincided with a seasonal increase in the availability of yel- low-flowered plant species, as occurs in M. asperatus. Conversely, the impact of dietary color changes on crab spider fitness parame- ters, such as prey capture success and suscep- tibility to predators, is unknown, but presum- ably would be negative. At the study site, most plant species available in early-to-mid- summer had white flowers. Therefore ingested prey pigments could cause a spiderling to con- trast strikingly with its floral substrate. How- ever, since pink and orange were the predom- inant opisthosomal colors, apparency to insects may not have been strongly affected. Most insects are considered to be red-blind (Borror et al. 1989; Barth 1991), but this in- terpretation of insect visual systems has re- cently been challenged (Chittka & Waser 1997). Susceptibility to visual predators with good color perception/discrimination, such as birds, could be enhanced by spider ingestion of prey pigments. Further studies are needed to determine what, if any, impact diet-induced color changes have on crab spider fitness pa- rameters. ACKNOWLEDGMENTS D. Morse kindly provided the M. vatia used in the laboratory experiments. W. Schmalho- fer and an anonymous reviewer provided helpful commentary. LITERATURE CITED Anderson, J.E 1970. Metabolic rates of spiders. Comp. Biochem. Physiol., 33:51-72. Angus, J. 1882. Protective change of color in a spider. American Nat., 16:1010. Barth, EG. 1991. Insects and Flowers: The Biology of a Partnership. Princeton Univ. Press, Prince- ton, New Jersey. 408 pp. Borror, D.J., C.A. Triplehorn & N.F. Johnson. 1989. An Introduction to the Study of Insects. Saunders College Publ., Fort Worth, Texas. 875 pp. Chittka, L. & N.M. Waser. 1997. Why red flowers are not invisible to bees. Israel J. Plant Sci., 45: 169-183. Fritz, R.S. & D.H. Morse. 1985. Reproductive suc- cess and foraging of the crab spider Misumena vatia. Oecologia, 65:194-200. Gabritschevsky, E. 1927. Experiments on color changes and regeneration in the crab- spider, Mis- umena vatia. J. Exp. ZooL, 47:251-267. Gadeau de Kerville, H. 1907. Sur I’homochromie protectrice des femelles du Misumena vatia Clerck (Arachn.). Bull. Soc. Entomol. France, 1907:145-146. Gertsch, W.J. 1939. A revision of the typical crab- spiders (Misumeninae) of America north of Mex- ico. Bull. American Mus. Nat Hist, 76:277-442. Hinton, H.E. 1976. Possible significance of the red patches of the female crab-spider, Misumena va- tia. J. Zool. London, 180:35-39. Holl, A. 1987. Coloration and chromes. Pp. 16-25, In Ecophysiology of Spiders. (W. Nentwig, ed.). Springer- Verlag, Berlin. Millot, J. 1926. Contributions a I’histophysiologie des Araneides. Bull. Biol. France Belgique, 8:1- 283. Morse, D.H. 1988. Cues associated with patch- choice decisions by foraging crab spiders Misu- mena vatia. Behav., 107:297-313. Morse, D.H. 1993. Choosing hunting sites with lit- tle information: Patch-choice responses of crab spiders to distant cues. Behav. EcoL, 4:61-65. Morse, D.H. & E.G. Stephens. 1996. The conse- quences of adult foraging success on the com- ponents of lifetime fitness in a semelparous, sit and wait predator. Evol. EcoL, 10:361-373. Nakamura, K. 1987. Hunger and starvation. Pp. 287-295, In Ecophysiology of Spiders.(W. Nen- twig, ed.). Springer- Verlag, Berlin. Packard, A.S. 1905. Change of color and protec- tive coloration in a flower-spider {Misumena va- tia Thorell). J. New York Entomol. Soc., 13:85- 96. Peck, W.B. & WH. Whitcomb. 1968. Feeding spi- ders an artificial diet. Entomol. News., 79:233- 236. Pollard, S.D. 1989. Constraints affecting partial prey consumption by a crab spider, Diaea sp. in- det. (Araneae: Thomisidae). Oecologia, 81:392- 396. Pollard, S.D. 1993. Little murderers. Nat. Hist., 102:58-65. Schmalhofer, V.R. 1996. The Effects of Biotic and Abiotic Factors on Predator-Prey Interactions in Old-Field Flower-Head Communities. Ph.D. Thesis, Rutgers Univ., New Brunswick. 178 pp. Weigel, G. 1941. Farbung und Farbwechsel der Krabbenspinne Misumena vatia (L.). Z. vergl. Physiol, 29:195-248. Manuscript received 22 August 1998, revised 1 July 1999. 2000. The Journal of Arachnology 28:61-69 COSTS AND BENEFITS OF FORAGING ASSOCIATED WITH DELAYED DISPERSAL IN THE SPIDER ANELOSIMUS STUDIOSUS (ARANEAE, THERIDIIDAE) Thomas C. Jones and Patricia G. Parker: Department of Evolution, Ecology, and Organismal Biology, The Ohio State University, 1735 Neil Ave., Columbus, Ohio 43210 USA ABSTRACT. In the theridiid spider, Anelosimus studiosus, most juveniles remain in their natal web, forming temporary colonies in which individuals cooperate in web maintenance and prey capture until they disperse at maturity. There is natural variation in age at dispersal, and subadult spiders removed from their natal webs build webs and continue to develop. To explore the costs and benefits of delayed dispersal, we compared the rate of prey capture and developmental rate for individuals in colonies and those isolated at the fourth instar. Rate of prey capture by colonies increased with colony size and age; this result was driven primarily by the enhanced capture of large prey by larger and older colonies. The presence of juveniles increased the overall productivity of webs, an effect which remained after the juveniles were removed from the web. Despite the overall increase in prey capture, per-individual prey capture decreased with colony size. The variance in prey capture success decreased significantly with colony size, but not with colony age. Spiders in colonies captured more prey per juvenile than singletons experimentally dispersed at the fourth instar; however, this did not result in increased development rate of colonial juveniles over isolated juveniles. These data suggest that juvenile A. studiosus benefit from delayed dis- persal by acquiring more resources and acquiring them more steadily. The productivity of webs of females whose juveniles were removed at the fourth instar remained higher than those of similarly aged females who never produced juveniles. This suggests that delayed dispersal of juveniles enhances the resources which the female could allocate to her next egg mass. Keywords: Parental investment, sub-sociality, risk-sensitivity, cooperative foraging Because spiders are generally limited by re- sources (Wise 1993), it is likely that any re- sources a mother spider provides to her ju- veniles would reduce her future egg production. Thus the behavior of maternal so- cial spiders would fit Trivers’ (1972) defini- tion of parental investment, in which a moth- er’s behavior enhances the survival of her current brood, at a cost to her production of future broods. However, if juveniles remain in their natal webs beyond an early altricial phase and become active in the web, their continued presence may enhance prey capture and/or defense. This in turn could enhance the mother’s production of future broods. In this way a mother may recoup her initial parental investment in terms of future reproductive success. The objective of this work is to de- scribe the relative costs and benefits of de- layed dispersal in Anelosimus studiosus (Hentz 1850), a spider in which the maternal- juvenile association is longer than in most ma- ternal social species. We used laboratory ex- periments to examine the effects of delayed dispersal on prey capture and development rate of late instar juveniles. We also examined the post-dispersal prey capture of webs in or- der to determine if delayed juvenile dispersal could enhance a mother’s future reproductive success. The effect of maternal care on the survival and growth of juveniles in maternal social spi- ders is well documented. Guarding of egg sacs is a relatively common form of maternal care in spiders, providing protection from preda- tion and parasitism (Foelix 1996). In colonies of the theridiid spider Theridion pictum (Wal- ckenaer 1802), unguarded egg sacs had dras- tically reduced hatching success, but juvenile size was not affected (Ruttan 1991). In about 20 described species, mothers actively provi- sion their offspring with paralyzed or regur- gitated prey (Foelix 1996). Mothers of the Eu- ropean agelenid spider Coelotes terrestris (Wider 1834) provision their offspring and protect them from predators and parasites un- 61 62 THE JOURNAL OF ARACHNOLOGY til the juveniles disperse after about one month (Horel & Gundermann 1992). Under laboratory conditions, the mother*s presence had a significant positive effect on juvenile survival. The mother’s parental investment, in terms of her ability to produce a second brood, was small relative to the enhanced survivor- ship of the current brood (Gundermann et al. 1997) . The 17 known species of non-territorial permanent- social spiders represent six fami- lies and are mostly found in the tropics (Avi- les 1997). Several studies have indicated that individual survivorship of colony members is greater than that of solitary individuals (Chris- tenson 1984; Riechert 1985; Aviles & Tufifio 1998) . Potential benefits of group living for spiders include reduced individual silk costs (Riechert et ah 1985; Tietjen 1986), capturing larger prey (Nentwig 1985; Rypstra 1990; Rypstra & Tirey 1990; Pasquet & Krafft 1992) and reduced predation (Henschel 1998), Fecundity in social spiders is lower than in solitary species (Riechert 1985; Vollrath 1986; Wickler & Siebt 1993). Female Anelo^ simus eximius in large colonies have lower fe- cundity than those in intermediate colonies (Keyserling 1884, Aviles & Thfino 1998). Po- tential costs of sociality for spiders include competition within the group (Rypstra 1993), increased incidence of parasitism (Aviles & Tufino 1998), and susceptibility to diseases (Henschel 1998). The social behavior of the theridiid spider, A. studiosus, is intermediate between the ma- ternal social and the non-territorial permanent- social spiders (Brach 1977), and the costs and benefits of delayed juvenile dispersal may go beyond simple parental investment. If web productivity is sufficiently enhanced by the presence of the late-instar, participating juve- niles, this enhancement could balance the costs of parental care to the mother, or even enhance her production of future broods. In this regard, A. studiosus may represent an evo- lutionary intermediate between maternal so- cial and non-territorial permanent-social spi- ders and, thus, could provide an important link in understanding the evolution of spider so- ciality. METHODS Study species.— Aneiosimus studiosus range from Argentina to New England and are typically found in open habitat, building webs at the tips of branches in low shrubs (Brach 1977). Adult females are fertilized before leaving the natal web or shortly after dispers- al. The mother produces and guards an egg case, feeds newly-emerged offspring through regurgitation, and provides second instar ju- veniles with paralyzed prey. As the juveniles develop beyond the second instai; they partic- ipate increasingly in prey capture and web maintenance (Brach 1977), Juveniles isolated at the fourth iestar or later can build their own webs, capture prey and continue to develop (Brach 1977; pers. obs.). Males are mature at the sixth post-emergent instar, and females at the seventh (pers. obs.). As the Juvenile fe- males mature, the mother becomes aggressive towards them, forcing them from the web (Brach 1977; but see Furey 1998). Adult males are always tolerated in the web by the mother; therefore, the maturing males appar- ently disperse of their own accord (Brach 1977). Female A. studiosus can produce up to three consecutive broods using the same web (pers. obs,). Rearing methods.— We collected 16 colo- nies from the Ocala National Forest in Florida in 1994 and 1995. We reared these colonies on live shrubbery within a 3.6 m X 2.4 m X 2.1 m enclosure in the Biological Sciences Greenhouses located at The Ohio State Uni- versity, maintained at temperatures between 23-32 °C, with a combination of natural light and supplemented light (on cloudy days) re- flecting the natural light cycle. Flying prey (Musca domestica. Drosophila melanogaster and D. hydei) were released into the enclosure three times a week, at which time the colonies were misted with distilled water. From the en- closure, we collected 72 adult females dwell- ing singly in newly-constructed webs in late March and early April 1997 and maintained them individually in 500 ml plastic containers. Each spider was provided a coiled twist-tie, which they used as a retreat. We fed them ad libitum, misted them three times a week, and exposed them to a male for 24 h within the week after they were collected. Voucher spec- imens are placed in The Museum of Biologi- cal Diversity at The Ohio State University. Experimental procedure,— Thirty-eight of the 72 isolated females produced egg cases. We placed these, with their egg sacs and re- treats, onto a small piece of artificial shrub- JONES & PARKER— FORAGING AND DISPERSAL IN SPIDERS 63 bery for 24 h while they constructed new webs. We then wired these new webs into the middle of larger arrangements of artificial shrubbery which were standardized by num- ber, size and positioning of the leaves. We housed the webs, individually, within cuboidal enclosures 46 cm on a side (these were screened on the four sides and solid on the top and bottom). Three times a week, we mist- ed the webs and released two M. domestica and ten D. melanogaster into the enclosure. We censused each web 48 h after prey release for the numbers and types of prey captured, as well as the numbers and age classes of ju- veniles present in the web. We removed the carcasses of captured prey from the webs and enclosures after each census. We assigned webs to two groups. In the treatment group we removed the juveniles from their natal web when the majority of them had reached the fourth instar, and indi- vidually placed three of the juveniles as sin- gletons into the experimental conditions de- scribed above. In the control group we removed the juveniles similarly, but immedi- ately replaced them and allowed them to de- velop and disperse naturally. We assigned webs to the two groups by first ranking them in order of number of juveniles in the web, then flipping a coin to decide the treatment of the first web, alternating the assignment of the remaining webs thereafter. We did this to en- sure a fair representation of the range in num- ber of juveniles in each treatment. There was no juvenile mortality or dispersal over the pe- riod for which the results are reported; thus, the number of juveniles remained constant within colonies. Seventeen females without juveiles were maintained under the experimental conditions for comparison with webs of similar age con- taining juveniles. Of these, ten did not pro- duce egg sacs, and seven produced egg sacs that did not hatch. If any of the adult females died during or within a week after the exper- imental period, we did not include data from their webs in the analyses. Twenty of the 38 egg sacs produced did not hatch, and six of the mothers died during the experiment. Data from seven control webs and five experimen- tally-dispersed webs were used. We estimated the amount of extractable re- sources for a given prey type as the average wet weight minus its average dry weight (13.1 mg for houseflies, 0.4 mg for Drosophila). Prey capture success was recorded as the number of each prey type times their extract- able weight. Due to asynchronous juvenile de- velopment, the age class of a web was de- scribed by the instar of the majority of the juveniles in it. Data analysis.-— In analyses exploring how colony size affects the amount of prey cap- tured, we calculated the mean per-trial prey capture over the period that juveniles were present. We estimated the per-juvenile prey capture by dividing the total mass of prey cap- tured in a trial by the number of juveniles in the colony. To analyze how colony size affects variation in prey capture, we used the coeffi- cient of variation (CV) among trials within colonies, in per-juvenile prey capture. We chose CV to standardize for the fact that we expect the variance to increase as the mean increases. We used regression analyses on the means and CVs of the colonies to test for ef- fects of colony size. In analyses of effects of colony age on foraging success we used data from the colonies multiple times (means and CVs at each instar within colonies), resulting in non-independence of the data. To account for this, we performed repeated measures analyses of covariance, with the instar of the majority of the juveniles as the covariate, and the individual colony as a random factor. RESULTS Effects of delayed dispersal on prey cap- ture.—-Across all webs, prey capture in- creased significantly with juvenile age (Fig. 1). In this plot, data from both the treatment and control colonies are factored into the means of the first three instars, because at that point both sets were intact and undisturbed. Only the control colonies are factored into the means of fourth through sixth instars. How- ever, we used only data from the control col- onies in the repeated measures ANCOVA. Mean per-trial prey capture also increased sig- nificantly with number of juveniles in the col- ony (Fig. 2). Despite the overall increased productivity of larger webs, there was less prey available to individual spiderlings as the number of juveniles increased (Fig. 3). The average coefficient of variation in per-juvenile prey capture showed no trend with respect to colony age (Fig. 4). There was, however, a significant decrease in the coefficient of vari- 64 THE JOURNAL OF ARACHNOLOGY Figure L— Average per-trial prey capture during the period juveniles were in the web vj stage of the colony. Plotted are the means for the colonies at a given instar with standard error bars (repeated mea- sures ANCOVA F = 4.07, P = 0.0035). Figure 3.— Average per-juvenile, per-trial, prey capture during the period juveniles were in the web vj number of juveniles in the web. Plotted are the means for each colony over all instars with standard error bars {m = 0.77, P - 0.01). ation in per-juvenile prey capture as colony size increased (Fig. 5). Much of the effects of colony size and age on foraging success were driven by the en- hanced ability of larger and more mature col- onies to capture the larger prey items. The av- erage number of houseflies captured per trial increased significantly with colony size {K^ — 0.79, P = 0.007; regression of the average number of houseflies captured per trial on the log of the number of juveniles in the colony). This increase was non-linear and asymptotic because the larger colonies depleted the avail- able flies. There was also a significant increase in the mean number of houseflies captured with colony age {F = 2.69, P = 0.04; repeated measures ANCOVA with juvenile instar as a cofactor). Effects of delayed dispersal on juvenile development,— The development rate of ju- veniles in colonies, as measured by the amount of time required to reach the fourth or sixth instars, was not related to prey capture per juvenile (Fig. 6). Similarly, when these de- velopment rates were compared to the coef- ficients of variation in per-juvenile prey cap- ture success, no trends were found (Fig. 7). Experimentally dispersed fifth instar single- tons captured fewer prey, on average, than the per-juvenile rate for a colony (Mann-Whitney U — 56.0, P < 0.01; Fig. 8A). The main cause of this difference was the fact that the single- tons captured only Drosophila while the col- onies were able to capture houseflies. The dif- ference in prey capture did not result in a difference in development rate, as measured by the duration of the fifth instar, between col- ony juveniles and singletons (Mann- Whitney U = 37.0, P = 0.92; Fig. 8C). Male singletons captured significantly less prey (Mann- Whit- ney U - 114.5, P = 0.002), and developed significantly more slowly in the fifth instar Figure 2.“-Average per trial prey capture during the period juveniles were in the web vj- number of juveniles in the colony. Plotted are the means for each colony over all instars with standard error bars (/?2 == 0.64, P = 0.003). Figure 4.— Coefficient of variation in per-juve- nile prey capture within instar, within colonies, vs stage of the colony. Plotted are the mean variances of the colonies at each instar with standard error bars (repeated measures ANCOVA F = 0.85, P - 0.81). JONES & PARKER— FORAGING AND DISPERSAL IN SPIDERS 65 Figure 5. —Coefficient of variation in per-juve- nile prey capture, within colonies, number of ju- veniles in the colony. Plotted are the variances for each colony pooled over all instars (R^ ~ 0.72, P - 0.017). (Mann-Whitney U = 49.0, P = 0.002), than female singletons (Figs. 8B, 8D), Effects of delayed dispersal on a moth- er’s future reproductive success.— To ex- amine potential foraging benefits to the moth- er associated with delayed dispersal of her offspring, we compared prey capture within and among the webs of females which did not produce egg cases (Group A, Table 1), webs in which females were guarding egg cases that did not hatch (Group B, Table 1), and the webs from which the juveniles had been ex- perimentally dispersed (Group C, Table 1). There were no differences in prey capture in the first week between any of the categories of webs, nor were the webs in which there were no juveniles more productive in the 5th week than they were at the first. Females who had had juveniles in their webs captured sig- nificantly more prey during the week after their offspring were dispersed (which on av- erage was around the fourth week after being placed on the plant) than did either of the two categories that had not had juveniles. Prey capture of females the week after their juve- niles were removed was not different than that of the week prior while the juveniles were still present. DISCUSSION The results presented here demonstrate that the presence of juveniles increased the overall productivity of webs, and that productivity in- creases with both the age (Fig. 1) and the number of juveniles in the web (Fig. 2). The majority of these effects were driven by the ability of larger and older colonies to capture Mean Per-Juvenite Prey Capture (g) Figure 6.— Colonial juvenile development vs mean per-juvenile prey capture. The points plotted are the times taken by colonies to reach the speci- fied instar (4th instar, = 0.012, P = 0.74.; 6th instar R^ - 0.075, P = 0.71). more houseflies, one of which has more ex- tractable resources than all ten of the Dro- sophila combined. These results are consistent with those found for several permanent- social spider species (Riechert et al. 1986; Tietjen 1986) including a congener of this species, A. eximius (Nentwig 1985; Rypstra 1990), as well as in colonial orb-weaving spiders (Uetz 1989). In these studies, social spiders captured larger prey and a wider range of prey sizes than solitary spiders of similar size. There was a significant decrease in the co- efficient of variation in per-juvenile prey cap- ture associated with the number of juveniles in the colony (Fig. 5). Reduced variance in foraging success has been identified as a po- tential benefit of spider coloniality in a dy- namic model (Caraco et al. 1995), and in co- lonial orb-weaving Metepeira spp. (Uetz 1988a, 1988b). These studies found that, un- der high prey densities, coloniality represents Coefficient of Variation in Per-Juvenile Prey Capture Success Figure 7. — Colonial juvenile development vs mean coefficient of variation in per-juvenile prey capture. The points plotted are the times taken by colonies to reach the specified instar (4th instar, R^ = 0.259, P = 0.30; 6th instar R^ = 0.005, P = 0.86). 66 THE JOURNAL OF ARACHNOLOGY fi- 7 S Ui S OJ ■M CL CO O >1 0? c/5 S' Q (U (0 k_ D Q CoSonia! Singleton Juveniles Juveniles n~ 9 n ^ 8 Female Male Singletons Singletons Figure 8. — Boxplots comparing prey capture and juvenile development between colonial and singleton juveniles. Plotted are the medians, inter-quartile ranges and standard ranges (see text for significance statistics). a ‘risk averse’ strategy in which the spiders trade a reduction in mean individual capture rate for a reduction in variance in capture rate. We found no relationships between mean or CV in per-juvenile prey capture and devel- opment rate (Figs, 6, 7), nor did the singleton juveniles develop more slowly than colonial individuals, despite the greatly-reduced prey capture in singletons (Fig. 8C). This suggests that, under these prey densities, the colonies were capturing considerably more prey than they could physiologically assimilate. Female singletons were more successful at capturing prey than male singletons (Fig. 9B). Though not measured directly, the female sin- gletons’ webs appeared larger and denser than those of the males. Among the non-territorial permanently-social spiders, males typically do not participate in web activities, and in such species the adult sex ratios are skewed to- wards females (Aviles 1997). These skewed sex ratios have apparently evolved through group selection, meeting the stringent condi- tions required to select for a trait which is ben- eficial to the colony but which, within the col- ony, reduces the fitness of individuals possessing it (Aviles 1986, 1993; Smith & Hagen 1996). The data presented here suggest that female A. studiosus may benefit by skew- ing their broods toward females. If web pro- ductivity increases with the proportion of fe- male juveniles, there may be an optimal brood sex ratio which balances the increased survi- vorship of female-biased broods, with Fisher’s (1958) selective pressure towards an equal in- vestment between male and female offspring. A female biased sex ratio was reported for this species in a Tennessee population (Furey 1998), but was not found among specimens from Ecuador (Aviles & Maddison 1991). The results presented here suggest that A. studiosus juveniles benefit from remaining in their natal web by obtaining more resources, and more consistent resources, than they JONES & PARKER— FORAGING AND DISPERSAL IN SPIDERS 67 Table 1.— Weekly web productivity averages, variances and specific comparisons (T statistics and P values) for three types of web. Group A females did not produce egg sacs, Group B females produced egg cases which did not hatch, and Group C females produced egg cases which hatched, and had their juveniles removed at the fourth instar. Group A No egg sac produced in = 8) Group B Eggs did not hatch {n = 6) Group C Juveniles removed at 4th instar in = 5) Week Wk 1 Wk 5 Wk 1 Wk 5 Wk 1 Wk 4 Wk 5 Mean (g) Variance 0.0017 3.6 E-6 0.0033 4.0 E-6 0.0014 4.5 E-6 0.0034 6.1 E-6 0.0032 1.2 E-5 0.0156 1.3 E-5 0.0150 5.4 &-5 A (wk 1) -1.63 P - 0.073 0.26 P - 0.40 -1.04 P = 0.16 A (wk 5) -0.021 P - 0.49 -8.0 P = 3 E-6 B (wk 1) _ -1.28 P = 0.13 -1.07 P - 0.156 B (wk 5) — -6.69 P = 5 E-5 C (wk 1) — -10.2 P = 0.0003 C (wk 4) — - -0.18 P = 0.43 C (wk 5) — would as singletons. However, because per- individual prey capture decreases with colony size (Fig. 3), for any given prey density there will be an upper limit to the number of ju- veniles a colony can support. Colony sizes in this experiment were lower than those report- ed for natural colonies (a mean of 36 juveniles at hatching; Brach (1977)). While the potential benefits of delayed dis- persal to the juveniles are relatively clear, there is indirect evidence that there are bene- fits to the mother as well. In this study, fe- males in webs that previously had juveniles captured more prey than those with webs of the same age that had not (Table 1), but webs that had had juveniles were no less productive during the week after the juveniles were re- moved than during the previous week with the juveniles present. This suggests that the ju- veniles’ main contribution to web productivity is in web construction rather than in subduing prey. While size of webs was not measured, webs with juveniles present became notice- ably larger than webs without. Because there is no observed aggression between a mother and her younger offspring, or among juveniles (Brach 1977), it is likely that captured prey is divided evenly (or at least randomly) among colony members. Ob- servations of interactions among colony mem- bers are limited for this species, and it is pos- sible for the mother or larger juveniles to dominate captured prey. Further work is need- ed to explore potential sibling rivalries and parent-offspring conflicts in this species. It should be kept in mind that, in this ex- periment, prey densities were artificial, stan- dardized, and depletable. Prey densities were chosen in an attempt to eliminate nutritionally related mortality, not to represent natural con- ditions. Therefore, the extent to which the pro- tocol reflects conditions associated with the evolutionary maintenance of A. studiosus be- havior is limited; however, the internal com- parisons of the experiment remain robust. The depletion of the prey in a given trial puts an upper limit on possible prey capture success (although in only two trials did a web capture all of the prey released). Prey density during a trial decreased as prey were captured, re- sulting in a decline in the probability of cap- turing more prey. Overall, prey depletion should have the effect of reducing the power of the experiment to detect factors that affect the mean capture rate of webs; prey depletion may also create a spurious reduction in vari- 68 THE JOURNAL OF ARACHNOLOGY ance measures as the more productive webs approach prey depletion. That neither colonial nor singleton juveniles appeared to be food-limited in this study is suggested by the stable growth rates of juve- niles regardless of group size or prey capture rate. These results would predict that under lower prey densities food limitation would af- fect the singleton juveniles more than colo- nials, except when the colony is so large that the per-juvenile prey capture is below that of singletons. As long as prey densities are high enough on average to support the colonies, the reduction in variance associated with cooper- ative foraging may allow the juveniles to as- similate the resources more efficiently. The data presented here suggest that de- layed dispersal of a brood could enhance the mother’s production of future broods by in- creasing the productivity of her web. The ex- perimental conditions were relatively mild, compared to natural conditions where webs are frequently damaged, particularly by rain- fall. Thus, cooperative web maintenance in this species may be even more important than this study would suggest. From these experimental data, it seems like- ly that cooperative foraging plays a significant role in the evolutionary maintenance of de- layed offspring dispersal in Anelosimus stu- diosus. While this work has identified several potential advantages of delayed dispersal, the specific nature of the costs and benefits would need to be tested under more natural condi- tions. This is also true for other factors which could influence the maintenance of delayed dispersal such as predation risk and parasit- ism. ACKNOWLEDGMENTS We thank G. Uetz, E. Marschall and T. Grubb and the members of the Parker lab for their assistance with the experimental design and manuscript preparation, and P. Doherty for statistical consultation. We also thank G. Keeney, A. Reynolds and the Jones family for help in collecting and maintaining specimens. Thanks to G. Miller, P. Sierwald, J. Berry and an anonymous reviewer for their helpful com- ments on this manuscript. Finally, special thanks to the late V. Roth for initial identifi- cation of specimens, and encouragement. LITERATURE CITED Aviles, L. 1986. Sex-ratio bias and possible group selection in the social spider Anelosimus eximius. American Nat, 128:1-12. Aviles, L, 1993. Interdemic selection and the sex ratio: a social spider perspective. American Nat., 142:320-345. Aviles, L. 1997. Causes and consequences of co- operation and permanent sociality in spiders. Pp. 476-498, In The Evolution of Social Behavior in Insects and Arachnids (J.C. Choe & B.J. Cres- pi, eds.). Cambridge Univ. Press, Cambridge, U.K. Aviles, L. & W. Maddison. 1991. When is the sex ratio biased in social spiders?: Chromosome studies of embryos and male meiosis in Anelo- simus species (Araneae, Theridiidae). J. Arach- nol., 19:126-135. Aviles, L. & P. Tufiho. 1998. Colony size and in- dividual fitness in the social spider Anelosimus eximius. American Nat, 152:403-418. Brach, V. 1977. Anelosimus studiosus (Araneae: Theridiidae) and the evolution of quasisociality in theridiid spiders. Evolution, 31:154-161. Caraco, T, G.W. Uetz, R.G. Gillespie & L. Giral- deau. 1995. Resource consumption variance within and among individuals: On coloniality in spiders. Ecology, 76:196-205. Christenson, TE. 1984. Behavior of colonial and solitary spiders of the theridiid species Anelosi- mus eximius. Anim. Behav., 32:725-734. Fisher, R.A. 1958. The Genetical Theory of Nat- ural Selection. 2nd ed. Pp. 158-160. Dover Pub- lications, New York. Foelix, R.F. 1996. Biology of Spiders. 2nd ed. Pp. 211-212. Oxford Univ. Press, New York. Furey, R.E. 1998. Two cooperatively social popu- lations of the theridiid spider Anelosimus studio- sus in a temperate region. Anim. Behav., 55:727- 735. Gundermann, J.L., A. Horel & C. Roland. 1997. Costs and benefits of maternal care in a subsocial spider, Coelotes terrestris. Ethology, 103:915- 925. Henschel, J.R. 1998. Predation on social and soli- tary individuals of the spider Stegodyphus dum- icola (Araneae, Eresidae). J. ArachnoL, 26:61- 69. Horel, A. & J.L. Gundermann. 1992. Egg sac guarding by the funnel-web spider Coelotes ter- restris: Function and development. Behav. Proc., 27:85-94. Nentwig, W. 1985. Social spiders capture larger prey: A study of Anelosimus eximius (Araneae: Theridiidae). Behav. Ecol & Sociobiol., 17:79- 85. Pasquet, A. & B. Krafft. 1992. Cooperation and prey capture efficiency in a social spider, Ane- losimus eximius. Ethology, 90:121-133. JONES & PARKER--™FORAGING AND DISPERSAL IN SPIDERS 69 Riechert, S.E. 1985. Why do some spiders coop- erate? Agelena consociata, a case study. Florida EntomoL, 68:105-116. Riechert, S.E., R.M. Roeloffs & A.C. Echternacht. 1986. The ecology of the cooperative spider Agelena consociata in equatorial Africa. J. Ar- achnoL, 14:175-191. Ruttan, L.M. 1991. Effects of maternal presence on the growth and survival of subsocial spider- lings (Araneae: Theridiidae). J. Insect Behav., 4: 251-256. Rypstra, A.L. 1990. Prey capture and feeding ef- ficiency of social and solitary spiders: a compar- ison. Acta Zool. Fennica, 190:339-343. Rypstra, A.L, 1993. Prey size, social competition and the development reproductive division of la- bor in social spider groups. American Nat., 142: 868-880. Rypstra, A.L. & R. Tirey. 1989. Observations on the social spider Anelosimus domingo (Araneae, Theridiidae), in southwestern Peru. J. Arachnol., 17:368. Smith, D.R. & R.H. Hagen. 1996. Population structure and interdemic selection in the coop- erative spider Anelosimus eximius. J. Evol. Biol., 9:589-608. Tietjen, W.J. 1986. Effects of colony size on web structure and behavior of the social spider Mallos gregalis (Araneae, Dictynidae). J. Arachnol., 14: 145-157. Trivers, R.L. 1972. Parental investment and sexual selection. Pp. 136-179, In Sexual Selection and the Descent of Man. (B. Campbell, ed.). Aldine- Atherton, Chicago. Uetz, G.W. 1988a. Risk-sensitivity and foraging in colonial spiders. Pp. 353-377, In Ecology of So- cial Behavior. (C.A. Slobodchikoff, ed.). Aca- demic Press, San Diego. Uetz, G.W. 1988b. Group foraging in colonial web-building spiders: Evidence for risk sensitiv- ity. Behav. Ecol. & SociobioL, 22:265-270. Uetz, G.W. 1989. The “ricochet effect” and prey capture in colonial spiders. Oecologia, 81:154- 159. Vollrath, E 1986. Eusociality and extrodinary sex ratios in the spider Anelosimus eximius (Araneae: Theridiidae). Behav. Ecol. & SociobioL, 18:283- 287. Wickler, W. & U. Seibt. 1993. Pedogenetic soci- ogenesis via the “sibling route” and some con- sequences for Stegodyphus spiders. Ethology, 95: 1-18. Wise, D.H. 1993. Spiders in Ecological Webs. Pp. 19-38. Cambridge Univ. Press, Cambridge. Manuscript received 10 February 1999, revised 10 July 1999. 2000. The Journal of Arachnology 28:70-78 RESOURCE PARTITIONING OF SPIDER HOSTS (ARACHNIDA, ARANEAE) BY TWO MANTISPID SPECIES (NEUROPTERA, MANTISPIDAE) IN AN ILLINOIS WOODLAND Kurt E. Redborg and Annemarie H, Redborg: Department of Biology, Coe College, Cedar Rapids, Iowa 52402 USA ABSTRACT. Two spider-boarding mantispids, Mantispa uhleri Banks 1943 and Climaciella brunnea (Say 1824), were found to be partitioning available spider egg resources in an Illinois woods based on vertical stratification. Mantispa uhleri was found to be phoretic on the philodromid Philodromus vulgaris (Hentz 1847), the salticid Metacyrba undata (De Geer 1867) and the anyphaenid Aysha gracilis (Hentz 1847) at levels of 75%, 26%, and 27% respectively. Ail of these spiders were collected from areas above the forest floor. In contrast, C brunnea was collected from 19% of leaf litter-inhabiting lycosids of the genus Schizocosa. There was no host range overlap within the woods, but in a grassy field without appreciable stratification of vegetation adjacent to the woods, both M. uhleri and C. brunnea were found aboard the lycosid Rabidosa punctulata (Walckenaer 1837) at levels of 2% and 7% respectively. A single larva of Mantispa pulchella (Banks 1912) associated with an anyphaenid from the woodland sample was also collected in this study. Mantispids are far more common than has been previously supposed and are likely an important factor in spider population dynamics and the evolution of spider behavior. Keywords: Mantispa uhleri, Climaciella brunnea, Mantispa pulchella The neuropteran family Mantispidae, sub- family Mantispinae, contains insects whose larvae are spider-egg predators (Redborg 1998). Larvae obtain eggs in one of two ways: (1) direct penetration of an egg sac by first instar larvae that search for egg sacs in the field, or (2) boarding adult female spiders and entering the spider’s egg sac while it is being constructed (Redborg & MacLeod 1985). In- side the egg sac, larvae pierce and drain the eggs with a sucking tube formed by modified mandibles and maxillae. After three larval in- stars, the developing mantispid spins a cocoon inside the egg sac using silk from the Malpi- ghian tubules. The species of spiders utilized by mantis- pids in North America are partially known for a number of species including Mantispa uhleri Banks 1943 (Redborg & MacLeod 1985), Mantispa fuscicornis Banks 1912 from Texas (Rice 1986; Rice & Peck 1991, cited as Man- tispa sayi Banks 1897) and Mantispa pul- chella (Banks 1912) (Hoffman & Brushwein 1989), Mantispa interrupta Say 1825 (Hoff- man & Brushwein 1990) and Mantispa viridis Walker 1853 (Brushwein et al. 1992) from South Carolina. Much of this work has not been quantitative. Moreover, little is known about how sympatric mantispids interact in their selection of hosts. Here we report on the resource use of two mantispid species in Illi- nois. The present study developed while we were collecting overwintering larvae of M. uhleri for a laboratory experiment from two of its host spiders, Philodromus vulgaris (Hentz 1 847) (Philodromidae) and Metacyrba undata (De Geer 1867) (Salticidae) in a small wood- land, and discovered unexpectedly high levels of infestation. In the spring we collected other types of spiders for comparison and discov- ered a second species of mantispid, Clima- ciella brunnea (Say 1824), aboard many of them. Because until that time no data had yet been reported documenting more than one mantispid species from the same study site, we continued our collections to see if there was any pattern to the kinds of spiders board- ed. It became apparent that these two mantis- pids were not boarding the same species of spiders. METHODS Mantispids are associated primarily with hunting spiders (Redborg & MacLeod 1985; Hoffman & Bmshwein 1989, 1990; Redborg 1998). We collected cursorial spiders from 70 REDBORG & REDBORG— SPIDER USE BY TWO MANTISPIDS 71 four microhabitats in a four-hectare, oak-hick- ory forest near Mahomet, Illinois known as Stidham Woods. Spiders were collected from (1) beneath tree bark during the winter, (2) on shrub-level foliage during the early spring, (3) from the woodland leaf litter during the late spring, summer, and early fall, and (4) from grassy fields bordering the woods during the late summer and early fall. Spiders were later anesthetized with CO2 and examined under a stereo microscope at 18X magnification for the presence of mantispid larvae. Mantispid boarding frequencies on spiders were ana- lyzed using Chi-square or the Fisher Exact Test. Bark-associated spiders.— Spiders were collected from beneath the bark of eight living shagbark hickories, Cary a ovata K. Koch, dis- tributed throughout the entire woods, between 18 December 1982 and 5 February 1983. Loose bark was removed from the trunk up to a height of 4 m. A white sheet was placed around the base of each tree to catch any spi- ders that fell from the bark. Between 12 June 1983 and 22 June 1983 additional shagbark hickories were examined for the presence of female spiders guarding egg sacs. Bark was pulled back and the un- dersurface of it examined. Egg sacs and as- sociated spiders were collected, brought into the laboratory, and the presence of mantispid cocoons and emergence of any adult mantis- pids were recorded. Low-level foliage spiders.— Spiders were located on the branches of small trees and shrubs with the aid of a headlamp and col- lected by hand on 16 May 1983, 19 May 1983 and 21 May 1983. Leaf litter lycosids.— Wolf spiders (Lycos- idae) were collected by hand on three nights in late spring between 21-28 May 1983 from the leaf litter within the woods at night aided by eye reflections from a headlamp. A second sample was taken in mid summer on 30 June. A third sample was taken in the fall on 19 & 26 September 1983. Egg sacs were obtained from, or collected with, some of these spiders. Spiders and egg sacs were maintained under ambient temper- ature and photoperiod on a screened porch. Field lycosids.— Wolf spiders were collect- ed from the grassy field to the north of the woods using a headlamp as described above on seven nights between 3 August-26 Sep- tember 1983. Adult Climaciella observations.— Since Climaciella adults are known to frequent flowers, the field area to the east and north of the woods was surveyed approximately once a week from 1 July 1983 to 4 September 1983 for the presence of Climaciella adults on flow- ers. This time frame was chosen based on our collection of Schizocosa adults with egg sacs in late June (see results). Voucher specimens of this study are depos- ited in the Field Museum of Natural History. RESULTS Bark-associated spiders.— Most of the spiders collected belonged to two species, Philodromus vulgaris and Metacyrba undata (Table 1). Philodromus sits loose beneath the bark and does not produce any type of silken retreat. Some of these spiders were concealed in cracks or crevices while others, aided by their flattened morphology, simply sat adhered to the under surface of the outer bark or the outer surface of the inner layers. Many spiders fell from the tree when the bark was removed and were recovered from the sheet. Specimens of M. undata were all contained within dense silken retreats. Cheiracanthium mildei (Clu- bionidae), Ariadna bicolor (Segestriidae) and Herpyllus ecclisiastica (Gnaphosidae) occu- pied silk retreats less dense than those of M. undata. Spiders carrying mantispid larvae were col- lected from all eight trees examined. A high frequency of 64 out of 85 P. vulgaris (29 sub- adult 9, 30 subadult S, 5 juveniles) had been boarded by at least one larva of M. uhleri. Eight of these spiders carried more than one larva (seven with two, one with three), most of which (68 out of 73) were tightly adhered to the dorsal, ventral, or lateral surface of the pedicel. Although larvae will enter the book lungs of sufficiently large species of spiders (Redborg & MacLeod 1985), none were found in this area on Philodromus. The remaining larvae were located at various positions around the leg bases or underneath the edge of the carapace. Mantispa uhleri larvae were also found aboard the salticid Metacyrba undata although its infestation frequency of 25 out of 85 (26%) was significantly lower (x^ = 43.19, P < 0.001) than the frequency of 64 out of 98 72 THE JOURNAL OF ARACHNOLOGY Table 1 .—Collections of cursorial spiders from four microhabitats in Stidham Woods, Illinois during 1982-83. (* = May collection; ** = June collection; *** = September collection.) No. of sub- No. of No. of No. of adult sub- adult No. of Micro- juven- fe- adult fe- adult habitat Species iles males males males males Total bark Philodromus vulgaris (Hentz 1847) 9 36 40 85 bark Metacyrba undata (De Geer 1867) 68 18 12 98 bark Cheiracanthium mildei C, L. Koch 1864 1 1 2 bark Ariadna bicolor (Hentz 1842) 10 10 bark Herpyllus ecclisiastica Hentz 1832 6 6 foliage Aysha gracilis (Hentz 1847) 1 6 4 11 foliage Anyphaena fraterna (Banks 1896) 1 19 18 38 litter* Schizocosa saltatrix (Hentz 1844) 1 1 litter* Schizocosa ocreata (Hentz 1844) 4 6 10 litter* Schizocosa rovneri Uetz and Dondale 1979 1 1 litter* Schizocosa sp. 55 53 108 litter** Schizocosa saltatrix 1 1 litter** Schizocosa ocreata 1 1 litter** Schizocosa ocreatal rovneri 12 12 litter** Schizocosa sp. 4 4 litter*** Schizocosa sp. 40 40 field Rabidosa punctulata (Hentz 1844) 38 19 23 59 46 185 field Rabidosa rabida (Walckenaer 1837) 5 1 6 field Hogna caroUnensis (Walckenaer 1837) 2 2 0 m t CO (/> 80 70 60 50 I 40 Q. CO 30 ^ 20 c 0 y 10 0 a. 0 □ M. uhleri C. brunnea Philodromus Metacyrba Anyphaenidae Rabidosa Schizocosa Figure L— Percentages of five groups of spiders boarded by first instar larvae of Mantispa uhleri and Climaciella brunnea (Neuroptera: Mantispidae) from an Illinois woodland. Collections of Aysha gracilis and Anyphaena fraterna were pooled and are designated Anyphaenidae. REDBORG & REDBORG— SPIDER USE BY TWO MANTISPIDS 73 (75%) on P. vulgaris. Twenty-five M. undata (5 adult 6 adult S, 14 juveniles) had been boarded by at least one larva of the mantispid and two of these spiders carried two larvae each. Most larvae were located on the dorsal, ventral, or lateral pedicel while no larvae oc- cupied the book lungs. One additional bark- associated spider was found boarded by a lar- va of M. uhleri, A subadult female Cheiracanthium mildei carried a larva in the right book lung. PMlodmmus egg sac collections.-— Five fe- male P. vulgaris guarding egg sacs were col- lected on 12 June 1983, 13 were collected on 16 June 1983, and eight were collected on 22 June 1983 for a total of 26 spiders. With the exception of one spider collected on 12 June, all spiders were guarding two egg sacs. In all cases, egg sacs were located on the inner sur- face of the removed piece of bark. The two sacs were close enough together so that the legs of the spider came in contact with both. The sequence of the two sacs was easy to de- termine by the differing developmental state of the eggs within and the second egg sac con- structed was always the smaller of the two. Eight of the 26 spiders were guarding sacs that contained a larva or cocoon of M, uhleri. The egg sac attacked was always the larger, first sac. Adult mantispids emerged from these sacs on 29 June, 30 June (4 mantispids), 2 July (2 mantispids), and 3 July 1983, Six of the spiders also had egg sacs containing wasp larvae, although which sac was involved var- ied. One or more wasps emerged from four of the older, first sacs. Wasps were in both of one spider’s egg sacs, and one spider that had a mantispid in its first sac had a wasp in the second. Although 75% of overwintering P. vulgaris had been boarded by M. uhleri lar- vae, only 31% of the spring egg sacs con- tained larvae. The frequencies associated with these two percentages are significantly differ- ent (x^ = 15.4, P < 0.001). Low-level foliage spiders.— All of the spi- ders collected from this microhabitat were an- yphaenids belonging to two species (Table 1). The overall infestation level of M. uhleri on Aiiyphaenids was 6% (3 out of 49), Three A. gracilis (1 2S) out of 11 carried a larva of M. uhleri with two of these larvae on the ped- icel and one in a book lung. No larvae were found aboard the 38 A. fraterna. The frequen- cies of M. uhleri on these two anyphaenid spe- cies were significantly different (Fisher Exact Test, P = 0.009). The infestation frequency of 3 out of 1 1 (27%) on A. gracilis, however, was no different (x^ ™ 0.06, P > 0.05) from that on M. undata but was significantly less (x^ = 8.50, P < 0.01) than that on P. vulgaris. A single larva of M. pulchella was also found associated with one of these anyphaen- ids. It unfortunately was dislodged from its host spider and found loose in the examination chamber with representatives of both spider species so that its exact host association could not be determined. Leaf litter lycosids.— A total of 120 spi- ders was collected in May with a second smaller sample of 18 spiders collected in late June (Table 1). Only adult spiders could be reliably identified to species. Schizocosa ocreata and S. rovneri are sibling species whose males can easily be distinguished by the tufts of black bristles on leg 1 of S. ocrea- ta but whose females are morphologically identical (Uetz & Dondale 1979). Some of these females were reliably identified by suc- cessfully mating them with a male of the ap- propriate species. If adult females could not be so mated, they are referred to as S. ocreatal rovneri as they could have been either species. Twenty-three of the 120 (19%) spiders (3? S. ocreata, 18 S. ocreata, 1 1 subadult 9 Schi- zocosa sp., ^ subadult 8 Schizocosa sp.) col- lected at the end of May 1983 had been board- ed by at least one larva of C brunnea. Two of these spiders (19 S. ocreata and 1 subadult 9 Schizocosa sp.) carried two larvae. Unlike larvae of M. uhleri, which are usually tightly adhered to the pedicel, these larvae were lo- cated along the edge of the carapace with their heads oriented toward the pedicel or toward the membranous area between the edge of the carapace and the coxae. Larvae of M. uhleri show no visible movement while attached to the pedicel, but larvae of Climacieiia could be seen to periodically move along the carapace and they could often be seen to seemingly push their mouthparts into the soft areas be- neath it. Small drops of what appeared to be spider hemolymph could sometimes be seen adjacent to larval mouthparts or in areas where larvae had recently been. Of the 18 spiders collected late June, four (29 Schizocosa sp., 2 subadult 9 Schizocosa sp.) carried a larva of C. brunnea. One of the adult females produced an egg sac on 3 July 74 THE JOURNAL OF ARACHNOLOGY Table 2. — ^Levels of Mantispa uhleri and Climaciella brunnea (Mantispidae) aboard the spider Rabidosa punctulata (Lycosidae) collected from grassy fields north of Stidham Woods, Illinois in 1983 (* one spider boarded by both mantispid species). Date No. of spiders No. of mantispids Juveniles Subadults Adults Total M. uhleri C. brunnea Aug. 3, 4 34 0 0 34 \* 1* Aug. 18 2 38 0 40 2 4 Aug. 31 1 4 35 40 1 3 Sept. 19, 21 1 0 32 33 0 2 Sept. 26 0 0 38 38 0 3 Total 38 42 105 185 4 13 1983. On 23 July 1983 an adult male C. brun- nea emerged from this sac. Three of the adult female spiders not carrying a mantispid larva had an egg sac when collected. Spiderlings emerged from these egg sacs on 7 July, 11 July, and 17 July 1983. Seven of the 40 juvenile Schizocosa col- lected in September (Table 1), presumably the offspring from July egg sacs, carried a single larva of C. brunnea. The three Schizocosa samples yielded no larvae of M. uhleri. The frequencies of Cli- maciella infestation were 19%, 22%, and 18%, respectively. They were not significantly different (x^ = 0.18, P > 0.05) from each oth- er. Field lycosids.— Collections were predom- inated by Rabidosa punctulata (Table 1). This species overwinters as an adult and produces egg sacs in the spring. Consistent with this scenario, juveniles were collected in early Au- gust, subadults in mid-August, and adults in late August and September. In contrast to the other spiders in this study, R. punctulata was boarded by larvae of both M. uhleri and C. brunnea (Table 2). In fact, one of these spiders carried a larva of both mantispid species. All of the other boarded spiders carried only a single larva. Of the 185 R. punctulata collected, four (2%) had been boarded by M. uhleri while 13 (7%) had been boarded by C. brunnea. Although the fre- quency of C. brunnea on R. punctulata was significantly greater than that of M. uhleri (x^ = 3.95, P < 0.05), there were significantly more C. brunnea on Schizocosa in May (x^ = 9.17, P < 0.005). No mantispids were found on either R. rabida or H. carolinensis. Adult Climaciella collections.— The fields to the north and west of the woods were sur- veyed for adult mantispids on flowering plants approximately weekly on 10 dates in July, Au- gust and September. In July, the most con- spicuous plants in bloom included red clover, Trifolium pratense L.; wild carrot, Daucus carota L.; ox-eye daisy. Chrysanthemum leu- canthemum L.; and common milkweed, Ascle- pias syriaca L. These were followed in Au- gust by thistle, Cirsium spp. and sunflower, Helianthus spp, and, in September, goldenrod, Solidago spp. The only mantispids found on these plants were five C brunnea. Three fe- males were observed on three different milk- weed plants on 17 July. A courting male was found associated with one of these females. The male’s behavior was similar to that de- scribed by Boyden (1983) for C. brunnea on milkweed in Minnesota. A sweet musk-like odor, presumably pheromone, from this male was quite apparent. A fourth female was ob- served, also on milkweed, a week later on 23 July. DISCUSSION It was actually more difficult to find a spec- imen of P. vulgaris from Stidham Woods not carrying a mantispid than it was to find one that did. While the exceedingly high levels of both M. uhleri and C. brunnea at this field site may seem excessive to some, we contend this is not an anomaly. An ongoing study of the distribution of mantispids in Iowa has so far involved the collection of over 5000 speci- mens of P. vulgaris and M. undata and, in areas in eastern Iowa where M. uhleri occurs, its levels on P. vulgaris range from 16-70% and on M. undata range from 8-33% (unpubl. data). Thus, while the levels of 75% and 26% REDBORG & REDBORG— SPIDER USE BY TWO MANTISPIDS 75 reported here are high, they are not incongru- ous and certainly comparable to levels found in Iowa. Scheffer (1992) recently reported as- sociations between Climaciella and Schizo- cosa from Cincinnati and northern Kentucky. Although she did not report frequencies, her report does not suggest that spiders bearing Climaciella larvae were difficult to find. Up to now, most mantispid studies have dealt with single species and have focused pri- marily on the documentation of spider hosts. Data are now needed involving sympatric spe- cies of mantispids collected from an area small enough to make some meaningful com- parisons regarding resource partitioning. A re- cent study reports two Japanese mantispids boarding two different groups of spiders in de- ciduous forests (Hirata et al. 1995). Larvae of Mantispa japonica were found on spiders col- lected on plants while Eumantispa harmandi were found aboard spiders associated with the forest floor. However, no levels of infestation were reported and no statistical comparisons made. Both M. uhleri and C. brunnea are spider boarders that overwinter on their respective host spiders and enter egg sacs when they are constructed the following year. Although M. uhleri will board a wide variety of hunting spiders, it is becoming increasingly apparent that P, vulgaris is its major host in much of the North American Midwest. Larvae of M. uhleri enter P. vulgaris egg sacs in May and June and emerge as adults in late June and early July. Newly-hatched M. uhleri larvae should begin appearing in mid-to-late July. While the spider M. undata is also an im- portant host for M. uhleri, its role pales in comparison to that of P. vulgaris. Hoffman & Brushwein (1989) hypothesized that M. pul- chella's greater association with anyphaenids, salticids, and clubionids as opposed to phil- odromids, oxyopids, and thomisids was due to the fact that the former spiders make silken retreats that perhaps enabled larvae to locate or board them more easily. There is no evi- dence for this in M. uhleri because P. vulgaris lacks retreats. Also, the flattened resting pos- ture of Philodromus would allow much leg and venter surface area to contact the sub- strate, thus facilitating larval contact. Both spider groups reported here (Schizo- cosa and Rabidosa) as hosts for C. brunnea are the same as those reported by Redborg & MacLeod (1983) in southern Illinois. Of the two host groups, the most important appears to be members of the genus Schizocosa. The infestation level of 19% on the nearly mature Schizocosa collected in May represents larvae that likely had overwintered on these spiders. Our collecting data suggest that egg sacs from Schizocosa probably are produced in late June and early July. The emergence of the C. brun- nea adult on 23 July from the egg sac of a Schizocosa collected 30 June corresponds with the appearance of adults on milkweed in the field. One might therefore expect newly- hatched Climaciella larvae to begin appearing in late July or early August. These would be the larvae that were then found on the juvenile Schizocosa in September. The almost identical level of 18% infestation on these spiders com- pared to those collected in May suggests that Climaciella population levels were fairly sta- ble. Within the wooded area, there was abso- lutely no overlap of host range between M. uhleri and C brunnea. The division of spider resources seems to be based on vertical strat- ification. All of the spiders associated with M. uhleri are foliage-inhabiting spiders while the main host for C. brunnea, Schizocosa, is usu- ally confined to the forest floor. One could ar- gue that this differential association is due to restricted host preferences on the part of the larvae, but we think it more likely due to dif- ferences in adult ovipositional or larval searching behavior. Mantispa uhleri will read- ily board lycosids under laboratory conditions and has been found at various times on vir- tually every group of hunting spiders includ- ing species of Schizocosa in southern Illinois (Redborg & MacLeod 1985), and Climaciella will board spiders other than lycosids in the laboratory (Redborg & MacLeod 1983). While it is true that all of the associations in this report, as well as all other published data, link Climaciella with lycosids, we think this can best be explained by behavioral factors which keep Climaciella larvae close to the ground. Although there is currently no direct evi- dence documenting ovipositional sites for M. uhleri in the field, its high levels on Philod- romus suggest that this mantispid may be lay- ing its eggs in the foliage or branches of the forest canopy. Our observations through the years suggest to us that P. vulgaris develops 76 THE JOURNAL OF ARACHNOLOGY in the tree canopy. For instance, each October, following the first frost, appreciable numbers of subadult Philodromus can be found col- lecting between the window frames and sills of the science building on the Coe College campus. These spiders are not evident during the summer on the walls of the building and there is no significant low-lying vegetation surrounding the building other than the fre- quently-mowed lawn. It seems reasonable that the spiders have been developing on the fo- liage or branches of the several oaks that line the grounds around the building. We regard the window sills and frames of the building as being the “urban” ecological equivalent of loose bark. Published data concerning the life history of this spider are crucially needed. In contrast, we suggest that C brunnea adults, although they may aggregate, mate and feed on nearby flowers, enter the woods and lay their eggs on or near the ground, Clima- della larvae do not actively search as do the larvae of M. uhieri but instead adopt a phor- etic posture in which they rear up on their tails and sway back and forth with legs out- stretched (Redborg & MacLeod 1983). It is thus not likely that larvae will travel a great distance from their egg clutch. Their strategy as obligate spider boarders is to wait for spi- ders to come to them. They would be most likely to come in contact with active spiders which certainly characterizes the Lycosidae. To produce infestation levels of 19% on Schi- zocosa, that waiting place, and by extension the site of adult ovlposition, is most probably on or near the ground. In light of these arguments, the occurrence of larvae of both species of mantispids on R, punctulata in the field adjacent to the woods is corroborative. The vegetation here has lim- ited vertical stratification of no more than a few feet. Mantispids of either species that at- tempted to oviposit in the field would wind up laying eggs in basically the same area—on the ground, on various grasses, or on the foliage of low-lying plants. The salt marsh in south- ern Mississippi studied by LaSalle (1986) would have been structurally somewhat rem- iniscent of the area studied here. He found Climaciella there laying eggs at the tips of leaf spikes of J uncus rushes. We can imagine similar ovipositional behavior here for both M. uhieri and C. brunnea, Rabidosa punctu- lata, along with its sibling species R, rabida. is usually found in grassy areas. Spiders were collected both on the ground and crawling along the foliage of grasses and other plants. Thus, whatever vertical stratification is present is probably completely traversed by tliis spi- der. And this, :ippro^>nately enough, is the one place where there is overlap of host range. We found larvae ol both species on this spider. In fact, one spider carried a larva of both M. uM~ eri and C. brunnea. This is, to our knowledge, the first documentation of two different spe- cies of mantispid aboard the same spider. It is important to note that the 2% infesta- tion level of M. uhieri and the 1% level of C brunnea on R. punctulata are both signifi- cantly lower than their respective levels on other spiders, suggesting that the field area is not the preferred ovipositional location for ei- ther species. Also strongly supported is the contention that C. brunnea is leaving its flow- er-inhabiting aggregation areas to lay its eggs within the woods. If adults were preferentially laying their eggs near their mating sites on milkweed, one would expect to find signifi- cantly more larvae on R. punctulata than Schi~ zocosa. Just the opposite is true. The signifi- cantly higher level of C brunnea on R. punctulata compared to that of M. uhieri is consistent with the sit-and-wait specialization of this mantispid that may favor the selection of lycosids as hosts. Although neither parasite favored the grassy field area for oviposition, C brunnea may wind up laying more eggs there because of the necessity to travel be- tween the two sites. While the infestation level of M uhieri on R. punctulata is slight, the infestation level of C. brunnea on this spider is more substantial. There is the potential for a complex, overlap- ping life cycle similar to those described for M. uhieri (Redborg & MacLeod 1985) and M pulchella (Hoffman & Brushwein 1989), Ra- hidosa punctulata females overwinter as adults and produce egg sacs early in the spring. Larvae that survived the winter on this spider would probably have emerged from egg sacs before we looked for adults in the summer. Climaciella offspring might appear early enough in the year to board Schizocosa spiders destined to spin egg sacs that same summer, or they might board immature R. punctulata that would not spin sacs until the following year. One or two generations per year are thus possible. Future study will be REDBORG & REDBORG=-=SPIDER USE BY TWO MANTISPIDS 77 necessary to assess the importance of this spi- der in the population dynamics of C. brunnea and vice-versa. The finding of a single larva of M. pul- chella on one of the anyphaenids in this study is intriguing. It is consistent with the findings of Hoffman & Bmshwein (1989) who found M. pulchella in South Carolina associated with small foliage-inhabiting wandering spi- ders. In fact, anyphaenids yielded the greatest number of M. pulchella larvae in their study with both A, fraterna and A. gracilis serving as hosts. More extensive collecting of small wandering spiders might have uncovered ad- ditional larvae of M. pulchella and the exis- tence of a third level of resource partitioning in Stidham Woods. It is also possible that M. pulchella is truly rare here, perhaps unable to compete successfully on its normal hosts due to the high level of competition from M. uhl- eri. If one focuses on the anyphaenids as a group, the infestation level of M. uhleri is only 6%, intermediate between the high levels on Philodromus/Metacyrba and the non-exis- tent level on Schizocosa. However, this may be misleading. All M. uhleri associated with this family were aboard A. gracilis and none aboard A. fraterna. It is possible that these two spiders, although collected from the same microhabitat early in the spring, may be oc- cupying different areas during the critical time when M. uhleri larvae are boarding them. Still, Redborg & MacLeod (1985) did find M. uhleri aboard A. fraterna in southern Illinois. More extensive sampling will be needed to answer this question. While 75% of overwintering P. vulgaris had been boarded by M. uhleri larvae, only 8 of 26 (31%) of egg-laying P. vulgaris were affected with a larva in their first sac. This significant difference shows that some larval mortality occurs between overwintering and egg sac production. There may be spider be- havioral mechanisms that reduce the number of larvae successfully entering egg sacs. Of particular interest here are the two egg sacs spun and guarded by this spider. Mantispid larvae were only found in the larger first egg sac. Eggs in the second egg sac escaped pre- dation, at least by mantispids. This is very high selective pressure which could have shaped the egg-laying strategy of P. vulgaris. Recent evidence (Vittitoe 1991) indicates that the second egg sac of P. vulgaris is an anti- mantispid mechanism evolved specifically to thwart M. uhleri predation. The small area of our study site may have affected the way in which these two mantis- pids interacted here. Perhaps in larger more extensive woods, C. brunnea is restricted to the interface between woods and field while M uhleri populations are more homogeneous- ly distributed in woods or even concentrated within the interior. Such spatial differences might be muddied in small woodlands. Thus resource partitioning between these two spe- cies may involve additional horizontal com- ponents. We acknowledge that our sample of hunting spiders from Stidham Woods deals with only a small number of species and is thus incom- plete, but the four different microhabitats they represent provide a good beginning for un- derstanding the differences between these two mantispids. Future work should focus on ad- ditional spider groups, particularly those of the surrounding fields, where greater overlap between M. uhleri and C. brunnea is predict- ed. In 1975 we attended the annual meeting of the American Arachnological Society and one of us (K.E. Redborg) presented some prelim- inary graduate student research on spider boarding behavior by larval mantispids. Find- ings suggested that these insects were much more common than had been previously sup- posed. Following the talk B.J. Kaston com- mented that, although he found the results in- teresting, his general impression through the years was that mantispids were “as scarce as hen’s teeth.” Later, H.W Levi informed us that he had commonly seen what appeared to be such larvae attached to the pedicel of spi- ders that he had collected in Wisconsin. A few weeks after the meeting J.E. Carrel wrote that, sparked by our discussions, he had examined a large container of preserved wolf spiders from his lab and discovered a “scum” of mantispid larvae floating on the top. Some textbooks still regard mantispids as being a novel but obscure group of insects, at least in temperate North America. The time has now arrived when these fascinating insects may no longer be regarded as rare but can more prop- erly be assessed as having an important im- pact on spider ecology and an important role in the evolution of spider behavior. 78 THE JOURNAL OF ARACHNOLOGY LITERATURE CITED Boyden, TC. 1983. Mimicry, predation and poten- tial pollination by the mantispid Climaciella brunnea van instabalis (Say) (Mantispidae: Neu- roptera). J. New York Entomol. Soc., 91:508= 511. Brushwein, J.R., K.M. Hoffman & J.D. Culin. 1992. Spider (Araneae) taxa associated with Mantispa viridis (Neuroptera: Mantispidae). J. ArachnoL, 20:153-156. Hirata S, M. Ishii & Y. Nishikawa. 1995. First in- star larvae of mantispids, Mantispa japonica MacLachlan and Eumantispa harmandi (Navas) (Neuroptera: Mantispidae), associating with spi- ders (Araneae). Japanese J. Entomol., 63:673- 680. Hoffman, K.M. & J.R. Brushwein. 1989. Species of spiders (Araneae) associated with the imma- ture stages of Mantispa pulchella Banks (Neu- roptera, Mantispidae). J. ArachnoL, 17:7-14. Hoffman, K.M. & J.R. Brushwein. 1990. Spider (Araneae) taxa associated with the immature stages of Mantispa interrupta (Neuroptera: Man- tispidae). Entomol. News, 101:23-28. LaSalle, M.W. 1986. Note on the mantispid Cli- maciella brunnea (Neuroptera: Mantispidae) in a coastal marsh habitat. Entomol. News, 97:7-10. Redborg, K.E. 1998. The biology of the Mantis- pidae. Ann. Rev. Entomol., 43:175-194. Redborg, K.E. & E.G. MacLeod. 1983. Climaciel- la brunnea (Neuroptera: Mantispidae): a mantis- pid that obligately boards spiders. J. Nat. Hist., 17:63-73. Redborg, K.E. & E.G. MacLeod. 1985. The de- velopmental ecology of Mantispa uhleri Banks (Neuroptera: Mantispidae). Illinois Biol. Mon- ogr., #53. 130 pp. Rice, M.E. 1986. Communal oviposition by Man- tispa fuscicornis (Say) (Neuroptera: Mantispi- dae) and subsequent larval parasitism on spiders (Arachnida: Araneida) in south Texas. J. Kansas Entomol. Soc., 59:121-126. Rice, M.E. & W.B. Peck. 1991. Mantispa sayi (Neuroptera: Mantispidae) parasitism on spiders (Araneae) in Texas, with observations on ovi- position and larval survivorship. Ann. Entomol. Soc. America, 84:52-57. Scheffer, S.J. 1992. Transfer of a larval mantispid during copulation of its spider host. J. Insect Be- hav., 5:797-800. Uetz, G.W. & C.D. Dondale. 1979. A new wolf spider in the genus Schizocosa (Araneae: Lycos- idae) from Illinois. J. ArachnoL, 7:86-87. Vittitoe, D.A. 1991. A possible anti-mantispid (Neuroptera: Mantispidae) behavioral mecha- nism in the spider Philodromus vulgaris (Hentz) (Araneae: Philodromidae). Unpubl. thesis. Coe College; Cedar Rapids, Iowa, USA. Manuscript received 19 August 1998, revised 1 Au- gust 1999. 2000. The Journal of Arachnology 28:79-89 EFFECTS OF FERTILIZER ADDITION AND DEBRIS REMOVAL ON LEAF-LITTER SPIDER COMMUNITIES AT TWO ELEVATIONS Angel J* Vargash Department of Biology, University of Puerto Rico, P.O, Box 23360, San Juan, Puerto Rico 00931-3360 USA ABSTRACT. This study investigates the indirect effects of primary productivity enhancement via fer- tilization, and the direct effects of environmental differences at two elevations, on the density and species richness of leaf-litter spiders. Litter was sampled in tabonuco forest (340-360 m elevation) and elfin forest (1051 m elevation) within the Luquillo Experimental Forest Long Term Ecological Research (LTER) site in Puerto Rico, Treatments consisted of three blocks with fertilization and control plots at both sites, and a one time removal of hurricane generated debris at tabonuco forest only. Treatments had no significant effect on spider density, species diversity, and species richness at either elevation. Elfin forest showed lower densities and lower species richness than tabonuco forest due to harsh environmental conditions. The thin litter layer and similar standing litter in the tabonuco forest suggest that spiders are limited by habitat, and also that they have successfully recolonized the debris cleared areas at this elevation. Harsh environmental conditions at elfin forest seem to be strong enough to counteract the effects of fertilizer addition on the measured variables. However, the high biomass of grasses in the fertilization plots at elfin forest could have caused an underestimation of spider densities. This study suggests that habitat availability is an important variable in bottom-up models for food web link control Keywords I Leaf-litter community, species diversity, primary productivity enhancement, tabonuco forest, Puerto Rico Most studies of indirect effects of primary productivity enhancement on spider densities, or studies on spider recolonization patterns, have focused on above-ground spiders be- cause they are easy to manipulate and count (Vince et aL 1981; Ehman & MacMahon 1996). Prey density may be affected by the bottom-up effects of nutrient addition in a food web (Power 1992). For example, the density of spiders of the Gulf of California is correlated negatively with island size (Polls & Hurd 1995). Higher marine productivity input to smallcj- islands, due to exposition of larger superficial area of small islands compared to larger ones, permits the support of higher ar- thropod prey densities and a higher density of web building spiders (Polis & Hurd 1995). In a salt marsh fertilization experiment, spiders showed a numerical response to an increased density of herbivores in the fertilization plots (Vince et aL 1981). An increase in prey triggers a density in- dependent aggregational and reproductive nu- * Current address: P.O. Box 132, San Antonio, Puerto Rico 00690 USA merical response in web-building spiders (Riechert & Lockley 1984; Wise 1993). These responses are said to be density independent because spiders have longer generation times and lower fecundity than most of their prey, and therefore can not track their prey popu- lations closely (Riechert & Lockley 1984). Spiders can quickly recolonize shrubs from which they are excluded by manipulation (Eh- man & MacMahon 1996). Differences in the recolonization pattern, with an initial colonist inhibiting the establishment of others (see Drake 1991; Law & Morton 1993), have been shown to be an important factor in community composition development (Ehmann & MacMahon 1996). As generalist predators, spiders constitute a very important group structuring leaf-litter communities (Clarke & Grant 1968; Moulder & Reichle 1972; Pfeiffer 1996). Leaf-litter ar- thropod communities can vary seasonally (Frith & Frith 1990), and along elevational gradients (Olson 1994). Variation in inverte- brate abundance can also be related to avail- ability of nutrients (Uetz 1976; Olson 1994) and fluctuations in environmental conditions 79 80 THE JOURNAL OF ARACHNOLOGY (Frith & Frith 1990). The poorer the available nutrients and/or the harsher the environment, the lower the abundance. In this study I focus on leaf-litter spiders to address the following questions: (1) How do harsh environmental conditions and lower pri- mary productivity at one of two elevations of a tropical rain forest adversely affect litter spi- ders density and richness of species? (2) How does enhanced productivity at the two eleva- tions, via fertilizer addition, favor higher den- sities of spiders?, and (3) How is species com- position affected by recolonization of spider depleted sites? I sampled litter spiders at two elevations in the Luquillo Experimental Forest (LEF). These two areas have subjected to fertilization treatments since 1989, after the strike of Hur- ricane Hugo. The low elevation site includes plots where all hurricane-generated debris was experimentally removed; and as a result, al- most all invertebrates were also removed. METHODS Study site. — This study was conducted in the tabonuco and elfin forests found in the Lu- quillo Experimental Forest Long-term Ecolog- ical Research (LTER) site in Puerto Rico. The tabonuco forest area is located near El Verde Field Station, at the eastern part of Puerto Rico (18°20'N, 65°49'W) and it is at an ele- vation between 340-360 m (Zimmerman et al. 1995). It is classified as a subtropical wet for- est (Ewel & Whitmore 1973). This area is dominated by Dacroydes excelsa Vahl, known as tabonuco, and Prestoea montana Nichols, known as the sierra palm (Walker et al. 1996). The tabonuco forest was heavily damaged by Hurricane Hugo in September 1989 (San- ford et al. 1991). The mass of fine litter (de- fined as all leaf, wood <1 cm in diameter, and miscellaneous plant material) that resulted from the hurricane was almost 400 times the daily average at El Verde and Bisley (Lodge et al. 1991). Input of nutrients via litterfall ap- pears to have altered nutrient cycles, increas- ing forest productivity and nutrient availabil- ity (Sanford et al. 1991). Canopy cover and height also decreased dramatically (Brokaw & Grear 1991). Invertebrate populations were greatly reduced (Alvarez & Willig 1993; Wil- lig & Camilo 1991). One of the elfin forest areas of the Luquillo Mountains is located at Pico del Este (18°16'N, 65°45'W), which is a summit area at 1051 m of elevation. Its vegetation is clas- sified as lower montane rain forest (Ewel & Whitmore 1973). The dominant species are Tabebuia rigida Urban, Ocotea spathulata Mez, and Calyptranthes krugii Kiaersk (Walk- er et al. 1996). This forest was heavily defo- liated by Hurricane Hugo (Brokaw & Grear 1991). Compared to pre-hurricane levels, mean annual litterfall was 1.9 times higher, and the annual fine litterfall input of N (1.5X), P (1.7X) and K (3.1 X) times higher (Lodge et al. 1991). Aside from these damages, there were no large structural changes at Pico del Este (Walker et al. 1996). Structural and dynamic features of the ta- bonuco forest are very different from the high altitude elfin forest. The number of trees per hectare, basal area, and soil organic matter are higher in the elfin forest; while specific leaf area, canopy height, tree diameter range, for- est volume and biomass, and species diversity are higher at the tabonuco forest (Weaver & Murphy 1990). Tree ingrowth and mortality, tree growth (includes biomass, volume, and diameter), litterfall, amount of loose litter, lit- ter turnover, herbivory, and net primary pro- ductivity are higher at the tabonuco forest (Weaver & Murphy 1990). Climatic condi- tions in the elfin forest at Pico del Este such as high humidity, soil saturation, relatively low temperatures, high winds, and soil leach- ing are thought to be influential to its struc- tural and dynamic features (Weaver et al. 1986, Weaver & Murphy 1990). Experimental design.-— The experimental blocks in the tabonuco forest were chosen at random. Each block was divided in three ex- perimental plots measuring 20 X 20 m each (Zimmerman et al. 1995). Plots were located on ridge tops to minimize water flow between plots (G.R. Camilo pers. comm.). The three treatments were: (1) one-time total debris re- moval, (2) fertilization, and (3) control. The one-time debris removal treatment occurred one month after the hurricane. Following the treatment, litter was allowed to accumulate naturally. Fertilizer treatment was first applied immediately after Hugo and then approxi- mately every three months. Fertilizer was add- ed at an annual rate of 300 kg/ha N, 100 kg/ ha P, 100 kg/ha K, 8 kg/ha B, 15.4 kg/ha Cu, 2.2 kg/ha Fe, 25 kg/ha Mn, 26 kg/ha Zn and 19 kg/ha Mg (Walker et al. 1996). These rates VARGAS—EFFECTS OF FERTILIZER AND DEBRIS REMOVAL 81 constitute N (3X), P (SOX) and K (2X) the mean annual inputs from fine litterfall (Lodge et al. 1991). The control plot was left intact, with no debris removed and no fertilizer ap- plied (Walker et al. 1996). Each block in elfin forest consisted of pairs of plots, located on ridge tops, randomly as- signed as control or fertilization (Walker et al. 1996). Debris removal treatment was not ap- plied due to the small amount of Hurricane generated debris at this forest (Zimmerman pers. comm.). Each plot measures 9 X 14 m. Fertilizer was first applied in April 1990 and then every 3 months to the present (Walker et al. 1996). Fertilizer constitute N (15X), P (166X), and K (SOX) the mean annual input from leaf litterfall (Lodge et al. 1991). For this study I used three adjacent blocks in elfin for- est to compare with three blocks in the tabon- uco forest. The litter spider community was sampled five times at each site between February 1996-January 1997. Each sample consisted of four random 0.25 m^ quadrants of leaf litter taken by hand from each experimental plot in each of three blocks at tabonuco forest and six randomly chosen plots at elfin forest. Leaf lit- ter was then taken to the laboratory and placed in Berlese funnels for 5 days, or until dry, to extract invertebrates. After invertebrates were removed the processed litter was returned to the site of collection. We placed four pitfall traps randomly in each plot in order to sample wandering noc- turnal spiders and other invertebrates that may not retreat into the leaf litter by daytime. Traps consisted of containers with openings of 10 cm in diameter and 18 cm deep. Each con- tainer was filled to less than half of its capac- ity with a 70% ethanol -5% ethylene glycol solution. The opening was covered with dis- posable dish to exclude rain water. The traps were left on the sites for two days. Spider samples obtained from Berlese and pitfall traps were preserved in 70% ethanol and were sorted by family and genus, and identified to species whenever possible using the appropriate literature (Petrunkevitch 1929, 1930a, 1930b; Bryant 1942; Chickering 1967, 1968, 1969, 1972a, 1972b). Juveniles were identified to family level only. Family and ge- neric names follow Platnick (1989). Collected specimens were deposited in the Biology Mu- seum of the University of Puerto Rico, Rfo Piedras Campus. ANOVAs for a two-factor split plot design laid off in localities (based on Ott 1993) were performed to determine differences in square root transformed density data (Zar 1984) and the number of species present between treat- ments and localities. Subplots within a locality were tested for treatment and time of sampling effects. In addition, two-factor repeated mea- sures ANOVA (Ott 1993) were performed on data from tabonuco forest to include data from the debris removal treatment, which is exclud- ed in the split plot ANOVA. The Morisita-Hom index was used to esti- mate the similarity (family level) among sites (Horn 1966; Wolda 1983; Russel-Smith & Stork 1995). A Multidimensional Scaling analysis was performed on the similarity ma- trix obtained from the index calculations to have a graphic representation of the dissimi- larities between plots. A Principal Compo- nents Analysis was performed on data from all sites to determine which families are more important to the dissimilarities between plots. RESULTS Density of spiders.— A split plot ANOVA performed on data from both sites showed that there is no effect of treatment on spider den- sity, but there is significant difference between localities (Table la). Densities per plot ranged from 0-15 ind./m^ at elfin forest and from 5™ 118 ind./m^ at tabonuco forest (Fig. 1). A re- peated measures ANOVA performed on data from tabonuco forest to account for debris re- moval treatment effects revealed no difference in density of spiders between treatments, but revealed effects of time (Table lb). Peak den- sities occurred between September and Octo- ber (Fig. 1). Species richness.— Based on adult individ- uals, there was a total of 31 species and 19 families identified from the two forest types (Table 2). A total of 27 species was found at tabonuco forest (Table 2). The dominant spe- cies in all treatments was Modisimus montan- us, followed by Theotima radiata, and Mas- teria petrunkevitchi. When juveniles and adults were taken together, the dominant fam- ily was Pholcidae (Table 2). Modisimus mon- tanus is the only adult species collected in the Pholcidae, therefore the juvenile individuals are probably of the same species. 82 THE JOURNAL OF ARACHNOLOGY 120 100 - 80 - TF-F TF-R TF-C EF-F EF-C Figure 1. — ^Density of spiders per treatment plot, elevation, and sampling date. TF = tabonuco, EF = elfin, F = fertilization, R = debris removal, C = control. Data from elfin forest yielded 16 species, four of these were found only at elfin forest; and all four of these were collected only once (Table 2). Pooled data for adult individuals from both treatments show that Mysmena car- ibbaea is the dominant species at this locality, followed closely by Oningis minutus and Cor- inna jayuyae (Table 2). Theotina radiata and M. petrunkevitchi were virtually absent at elfin forest. When juveniles are taken in consider- ation, along with adults, the dominant families are Pholcidae (presumably M. montanus) and Salticidae (most adults represented by O. min- utus) (Table 2). Of the seven species collected only in pitfall traps at tabonuco forest, Agriog- natha gloriae was a web builder not typical of the litter, but of the understory. At elfin forest there were two species captured only in pitfall traps (Table 2). Of the two species, a poorly-preserved male specimen of the genus Tetragnatha is typical of the understory, A split plot ANOVA showed that the num- ber of species differed between localities but not between treatments (Table 3 a). The num- ber of species was higher at tabonuco forest (see Table 2). A repeated measures ANOVA, performed on data from tabonuco forest to ac- count for debris removal effects, revealed no effect of this treatment on the number of spe- cies (Table 3b). Community similarity.— The Morisita- Table 1. — ANOVA analysis for the effects of treatment, time of sampling, and elevation on the density of leaf litter spiders. TF = tabonuco, EF = elfin. Source df Mean square F P A. Split plot ANOVA to compare treatment and time of sampling effects between TF and EF. Debris removal was excluded from this analysis. Between localities Time 4 5.402 1.72 >0.25 TF vs EF 1 228.32 72.64 <0.0025 Time X (TF vs EF) 4 3.143 Within localities Treatment 1 0.025 0.01 0.904 Treatment X Time 4 0.546 0.32 0.865 Error 45 1.723 B. Repeated measures ANOVA to compare effects of treatment (including debris removal) and time sampling (Time) at the plots of tabonuco forest. Between blocks Treatment 2 1.334 1.46 >0.25 Plots in treatment 6 0.910 Within blocks Time 4 11.66 3.01 0.035 Time X Treatment 8 0.935 0.24 0.979 Error 24 3.88 VARGAS— EFFECTS OF FERTILIZER AND DEBRIS REMOVAL 83 Horn community similarity index was calcu- lated using pooled data for family from all sampling times. Table 4 shows the similarity matrix obtained from this analysis. Multidi- mensional Scaling Analysis of the matrix shows good separation of sites based on lo- cality (Fig. 2). Plots from the same elevation and treatment tend to be most closely related, with the exception of the tabonuco forest, where removal plots show the greatest varia- tion in species composition (Fig. 2). Principal Component Analysis show that the first axis accounts for 92% of the variance between plots. The family with the highest absolute loading in this axis is Pholcidae (0.98), fol- lowed by Ochyroceratidae (0.15) and Heter- opodidae (—0.085). Axis 2 accounts for 4.2% of the variance between plots, with the highest absolute loading values for Ctenidae (—0.73), Heteropodidae ( — 0.66) and Pholcidae (-0.35). Axis 1 shows a clear separation of plots by elevation, except for two plots from tabonuco forest (control #2 and debris remov- al #3) that appear together with the plots from elfin forest (Fig. 3). These two plots from ta- bonuco forest have a lower density of Phol- cidae than the rest of the plots from this ele- vation. Axis 2 clearly separates one of the fertilized plots from elfin forest from all the other plots (Fig. 3). The prevalence of families with negative eigenvector values (namely Oonopidae, Ctenidae and Heteropodidae) is responsible for this separation. DISCUSSION Density of spiders, — The lack of a treat- ment response may be due to a lack of re- sponse from spider prey to treatments. Prelim- inary data for litter insects from the same experimental plots show no significant differ- ence between treatments at tabonuco forest or elfin forest (E. Nazario pers. comm). Standing litter was similar for all treatments at a given elevation, even though litter fall was higher in fertilization plots at each elevation site (Walk- er et al, 1996). This suggest that there is a higher density of decomposers in the fertiliza- tion plots. A factor opposing the bottom-up productiv- ity enhancement effects on spiders relates to features of the litter. The structure and depth of the litter have been shown to be very im- portant factors affecting the density and di- versity of litter arthropods (Uetz 1979; Bult- man & Uetz 1982, 1984). Spider density and diversity increase with higher litter depth and complexity (Uetz 1979; Bultman & Uetz 1982, 1984). Litter depth proved to be more important, in the short term, for spiders than nutrient content of the litter (Bultman & Uetz 1984). The fact that the litter layer is relatively thin at the tabonuco forest (Pfeiffer 1996) and elfin forest (pers. obs.), and that standing litter in our plots is similar in all treatments (Zim- merman et al. 1995; Walker et al. 1996), sup- ports the statement that spiders are habitat limited in our plots. The constant and rapid turnover of leaf litter (La Caro & Rud 1985) may limit habitat for litter spiders. Because leaves are constantly decomposing, M. mon- tanus will have to frequently switch to a new leaf. Consumption by vertebrate predators is not a very important factor opposing the bottom- up productivity enhancement effects on litter spiders in the tabonuco forest (Pfeiffer 1996). Diurnal predators concentrate foraging activ- ities to the arboreal layers (Reagan 1996); nocturnal predators forage in arboreal areas or near the ground (Stewart & Woolbright 1996). Eleutherodactylus portoricensis Schmidt is the only vertebrate that includes some litter arthropods in its diet (Stewart & Woolbright 1996). Non-anoline reptiles like the gecko Sphaerodactylus klauberi Grant (1 individual/ m^) may account for the majority of litter ar- thropod consumption (Pfeiffer 1996), which include, in order of quantity, Acari, Araneae, Collembola, Isopoda, and Coleoptera (Thom- as & Gaa Kessler 1996). However, unlike Eleutherodactylus frogs, we never collected S. klauberi in the litter. The one time debris removal from experi- mental plots in 1989 eliminated almost all lit- ter fauna and their respective habitats. The lack of differences in spider density, and the similarity of standing litter between treatments (Zimmerman et al. 1995), suggests that litter spiders were able to recolonize rapidly. Spi- ders near the debris removal plots had poten- tial free habitat to colonize from the moment when leaf fall began to cover the forest floor once again. The similarity of standing litter between treatments (Zimmerman et al. 1995) meant equal leaf-litter habitat availability in all plots. The relatively small size of our study plots (20 m^) permits rapid recruitment of col- 84 THE JOURNAL OF ARACHNOLOGY Table 2. — Pooled abundance for spider families and species found in all treatments at the sites of tabonuco (TF) and elfin (EF) forests. Data include total number adult and juvenile specimens collected in Berlese funnels for each family. Total number of individuals from a species is based on adult individuals only. Species found only in pitfall traps are indicated by an asterisk (F = fertilization, R = debris removal, C = control). Taxon TF-F TF-R TF-C EF-F EF-^C Pholcidae 351 426 358 36 4 Modisimus montanus Pet. 67 87 79 8 1 Ochyroceratidae 32 91 48 4 4 Ochyrocera sp. 0 0 1 1 2 Theotima radiata Simon 23 72 37 0 0 Dipluridae 44 25 47 0 0 Masteria petrunkevitchi (Chickering) 18 8 11 0 0 Corinnidae 46 46 45 11 7 Corinna jayuyae Pet. 6 8 9 3 2 Trachelas bicolor Keyserling* 1 0 0 0 0 Heteropodidae 44 13 35 6 4 Pseudosparianthis jayuyae Pet. 7 0 4 1 2 Salticidae 20 23 36 12 28 Corythalia gloriae Pet. 5 4 8 0 0 Emanthis portoricensis Pet. 0 0 0 0 1 Oningis minutus Pet. 0 1 3 5 7 Oonopidae 17 3 22 7 2 Close to Dysderina sp. 1 0 0 0 1 Oonops ebenicus Chickering 4 1 7 1 0 Oonops sp. 1 0 1 0 0 Close to Opopaea lutzi Pet. 3 0 3 4 0 Stenoonops sp.* 2 0 0 0 0 Ctenidae 11 6 4 8 7 Celaetycheus strennus Bryant 0 1 0 4 1 Oligoctenus ottleyi Pet. 1 1 0 0 0 Symphytognathidae 0 5 2 9 5 Mysmena caribbaea Gertsch 0 2 2 9 4 Barychelidae 16 7 5 1 0 Trichopelma corozali (Pet.) 8 7 5 1 0 Caponidae 0 0 3 4 1 Nops blanda (Bryant)* 0 0 1 2 0 Hahniidae 0 2 7 0 0 Neohahnia ernes ti (Simon) 0 2 0 0 0 Linyphiidae 0 7 0 0 0 Leptyphantes microserratus Pet. 0 6 0 0 0 Liocranidae 4 0 2 0 0 Phrurolithus insularus Pet. 3 0 2 0 0 Prodidomidae 1 0 1 0 0 Lygromma sp.* 1 0 0 0 0 Tetragnathidae 1 0 0 1 0 Agriognatha gloriae Pet.* 1 0 0 0 0 Tetragnatha sp.* 0 0 0 1 0 Theraphosidae 1 1 0 0 0 Ischnocolus culebrae Pet.* 0 1 0 0 0 VARGAS— EFFECTS OF FERTILIZER AND DEBRIS REMOVAL 85 Taxon Theridiosomatidae Baalzebub albinotatus (Pet.) Chthonas sp. Styposis luteus (Pet.) Thomisidae Epicaudus mutchleri Pet. Table 2. — Continued. TF-F 1 0 0 0 0 0 TF-R 5 0 2 0 1 1 TF-C 7 0 3 0 1 1 EF-F 0 0 0 0 EF-C 4 0 0 0 0 onizers from the surrounding habitat limited leaf litter. Difference in density of litter spiders be- tween tabonuco forest and elfin forest is con- sistent with a study that compared abundance and diversity of litter arthropods at different elevations in Panama (Olson 1994). In west- ern Panamanian forests, species diversity and number of individuals decline in the upward transition to cloud forests (Olson 1994). This decline is associated with harsher environ- mental conditions (Weaver et al. 1986; Olson 1994), lower productivity (Weaver & Murphy 1990), and low resource availability (Olson 1994) at high elevations. Some harsh climatic conditions at elfin forest include high humid- ity, moisture saturation, relatively low tem- peratures, high winds, and soil leaching (Weaver et al. 1986). Primary productivity (Weaver & Murphy 1990) and insect density (E. Nazario pers. comm.) are also lower at PE compared to EV. Leaf litterfall (Weaver & Murphy 1990) and standing litter (Walker et al. 1996) is also lower for PE. Thicker leaves at PE (Medina et al. 1981) should also be harder to curl than leaves at EV; and this could reduce the three-dimensional space of the litter, which is an important feature for spi- der habitat (Uetz 1979; Bultman & Uetz 1984). Species richness.^ — ^Another study done at tabonuco forest found a total of 22 spider spe- Table 3.— ANOVA analysis to compare the effects of treatment and time of sampling between tabonuco (TF) and elfin (EF) forest on the number of species. Source df Mean square F P A. Split plot ANOVA to compare treatment and time effects between elevations. Debris removal is ex- eluded from the analysis. Between elevations Time 4 3.52 2.32 >0.10 TF vs EF 1 104.017 68.58 <0.0025 Time X (TF vs EF) 4 1.52 Within elevations Treatment 1 3.75 0.17 0.68 Treatment X Time 4 3.08 1.94 0.12 Error 45 1.59 B. Repeated measures ANOVA to compare treatment (debris removal included) and time of sampling at the plots of tabonuco forest. Between blocks Treatment 2 1.09 0.23 >0.25 Block 6 2.48 Within blocks Time 4 2.86 1.47 0.24 Time X Treatment 8 2.01 1.03 0.43 Error 24 1.94 86 THE JOURNAL OF ARACHNOLOGY ^ e2 s .s o o ^ V, e G C bO B S p . ^ C £ 1-1 3 ON 00 Os 00 00 H-* O d d d t-H ON r- CM [^2 VO ON VO ON o^ SO VO 00 vn r- 00 as 00 o d d d d o vr> P ^ ON r-- ON ON P ON so r- OS vn 00 Os Os OS (M d d d d d cn m so CM vn P P Os ON 00 ON 00 On P so >o VO VO so O 00 o d d d d d d CM o ON VO cn P P Os ON On 00 On P PU 00 »o »o »o SO 00 o d d d d d d d hh SO CM VO CM VO >o 00 Os 00 00 oo 00 E p 00 m VO r- VO SO CO d d d d d d d d »-H CO pq U VO 00 5 VO VO VO ON 00 CM CO CM so VO CO r- >o VO 00 00 d d d d d d d d d p:, CNi m u CM 00 o r^ On r- r- r- VO VO VO 00 VO 00 VO »n VT) VO d d d d d d d d d d r— ( »— I CM VO Os r- r-- ON ON cn P n 00 00 VO VO VO Tt VO VO pq u 00 00 o d d d d d d d d d d d PP CO pq PP VO CM o o VO 00 OS CM CM o CM «n r- VO so VO m m CO vn VO vn \0 VO 00 d d d d d d d d d d d d pH CM pq PP o o Os r- CM CM 00 CM CM VO t-' 00 ■vt r- r- 00 00 00 00 00 00 VO d d d d d d d d d d d d d VO o VO CM Ov VO (O ON pp ^ r-- VO 00 00 VO VO r-> VO 00 pq PP o d d d d d d d d d d d d d d E 2E »-H CM fO ,-H U U U PP EE s 2 cn u CM U TF-C3 Ph PP PP PP pq pq W gq pq pq ii VARGAS— EFFECTS OF FERTILIZER AND DEBRIS REMOVAL 87 2 -| ■ TF-F A TF-R • TF-C □ EF-F _ O EF-C ^ A • • €N I I B Q 0 - A Dimension 1 Figure 2. — Multi-dimensional Scaling Analysis on the Morisita-Hom community similarity index (stress = 0.222). TF = tabonuco, EF ™ elfin, F = fertilization, R = debris removal, C = control. 20 10 H 0 -10 -20 -30 a B ■ • -40 - -50 -60 - □ ■ TF-F A TF-R • TF-C □ EF-F O EF-C -70 H 1 i — 1 — — — r 1 1 -100 -50 0 50 100 150 200 Axis 1 Figure 3.— The distribution of treatment plots in a two-dimensional morphospace, based on scores from the first two-family principal components axes. TF - tabonuco, EF = elfin, F = fertilization, R = debris removal, C ~ control. cies in the leaf litter (Pfeiffer 1996). The high- er number of species I found was the result of using pit-fall traps. The dominant species in tabonuco forest are consistent with the domi- nant species found by Pfeiffer (1996). The dif- ference in dominant species I found between elevations has been found in other studies (Ol- son 1994), and is related to the same condi- tions that limit density (see previous section). Community similarity .-“-Differences in community composition between elevations are related to differences in environmental conditions (Olson 1994). Dissimilarity be- tween debris removal plots is indicative of random differences in the recolonization pat- tern at each of the plots, with an initial colo- nist species inhibiting the establishment of others (see Drake 1991; Law & Morton 1993). Differences in the initial colonists arriving to a spider depleted plot have been shown, in the short term, to be important in the process of community development (Ehmann & Mac- Mahon 1996). Principal Components Analysis showed that differences between plots was mainly explained by differences in the abun- dance of Pholcidae. Lower Pholcidae abun- dance in a tabonuco forest removal plot and a control plot (Fig. 3) grouped them with plots from elfin forest. The difference of Pholcidae abundance in these two plots is likely to be due to the heterogeneity between plots. Higher numbers of individuals of some families at elfin forest compared to tabonuco forest (Table 2) are indicative of differences in community structure between both eleva- tions. Dissimilarity between elevations is con- sistent with the steady changes in community composition, or species turnover, found in an elevational gradient in Panamian forests (Ol- son 1994). Analysis of species composition at the intermediate forests of El Yunque (palm and Colorado forests) is necessary in order to determine if species turnover is constant. Morisita-Hom analysis was also applied to determine similarity between mean annual density (MAD) data from Pfeiffer (1996) and my pooled treatments mean density per sam- ple (PMD) data for tabonuco forest. The cal- culated index was 0.92. In order to determine how his data compare to mine in terms of den- sity of individuals from each family, I ran a simple regression analysis. This analysis showed a high correlation of his data with mine (r ” 0.90, P < 0.0005). However, Pfeif- fer (1996) found a higher density of spiders than I did (MAD = 5.94 + 1.90 * PMD). The higher density of spiders found by Pfeiffer (1996) can be attributed to selection of sample sites away from rock surfaces and his use of a vacuum aspirator to obtain his samples. Our treatment plots were located along ridge tops. 88 THE JOURNAL OF ARACHNOLOGY This feature can minimize the already thin lit- ter cover on the steeper areas, while concen- trating litter on the relatively flat areas, con- sequently minimizing litter obtained for analysis. ACKNOWLEDGMENTS I want to thank Catherine N. Duckett, T Mitchell Aide, Jess K. Zimmerman, Manuel J. Velez, and Gerardo R. Camilo for their ad- vice, encouragement, support and comments on early drafts of this paper. Identification of difficult species would have been impossible without the help of Robert L. Edwards, who also reviewed a later draft of this paper. Com- pletion of this project would have been less enjoyable without the help of Eduardo Naza- rio, and the rest of the “hojarasca"’ team: Ja- vier Blanco, Jose J. Reyes, Mitchell Chaar and Alejandro Molinelli. Janice Alers, Katherine Nieves, and Vicente Gomez were volunteer field assistants on various occasions. Thanks to the staff at El Verde Field station for their help and hospitality. The Department of Bi- ology and the Institute for Tropical Ecosystem Studies provided vehicles for transportation to Pico del Este. This paper is based on the Mas- ter’s thesis I presented to the Department of Biology, University of Puerto Rico, Rio Pie- dras. Financial support was provided by NSF grant #HRD-9353549 awarded to the CREST program. LITERATURE CITED Alvarez, J. & M.R. Willig. 1993. Effects of treefall gaps on the density of land snails in the Luquillo Experimental Forest of Puerto Rico. Biotropica, 25:100-110. Bryant, E.B. 1942. Additions to the spider fauna of Puerto Rico. J, Dept. Agric. Puerto Rico, 26: U19. Brokaw, N.V.L. & J.S. Grear. 1991, Forest struc- ture before and after Hurricane Hugo at three el- evations in the Luquillo Mountains, Puerto Rico. Biotropica, 23:386-392, Bultman, XL. & G,W. Uetz. 1982. Abundance and community structure of forest floor spiders fol- lowing litter manipulation, Oecologia, 55:34-41. Bultman, T.L. & G,W, Uetz. 1984. Effects of struc- ture and nutritional quality of litter on abundanc- es of litter-dwelling arthropods. American Midi. Nat,, 111:165-=172. Chickering, A.M. 1967. The genus Nops (Araneae, Caponiidae) in Panama and the West Indies, Bre- viora, 274:1-19. Chickering, A.M. 1968. The genus Dysderina (Ar- aneae, Oonopidae) in Central America and the West Indies. Breviora, 296:1-37. Chickering, A.M. 1969. The genus Stenoonops (Araneae, Oonopidae) in Panama and the West Indies. Breviora, 339:1-35. Chickering, A.M. 1972a. The genus Oonops (Ar- aneae, Oonopidae) in Panama and the West In- dies, part 3. Psyche, 78:104-115. Chickering, A.M. 1972b. The spider genus Track- elas (Araneae, Clubionidae) in the West Indies. Psyche, 79:215-230. Clarke, R.D. & P.R. Grant. 1968. An experimental study of the role of spiders as predators in a for- est litter community, part 1. Ecology, 49:1152- 1154. Drake, J.A. 1991. Community-assembly mecha- nisms and the structure of an experimental spe- cies ensemble. American Nat., 137:1-26. Ehmaen, W.J. & J.A. MacMahon. 1996. Initial tests for priority effects among spiders that co- occur on sagebrush shrubs. J. ArachnoL, 24:173- 185, Ewel, J.J. & Whitmore, J.L. 1973. The ecological life zones of Puerto Rico and the U.S. Virgin Islands. U.S. Forest Service Research Paper ITF~ 18, Inst, of Tropical Forestry, Rio Piedras, Puerto Rico. Frith, D. & C. Frith. 1990. Seasonality of litter invertebrate populations in an Australian upland tropical rainforest. Biotropica, 22:181-190. Horn, H.S. 1966, Measurement of “overlap” in comparative ecological studies. American Nat., 100:419-424. Law, R. & R.D. Morton. 1993. Alternative per- manent states of ecological communities. Ecol- ogy, 74:1347-1361. La Caro, F. & R.L. Rudd. 1985. Leaf litter disap- pearance rates in Puerto Rican montane forests. Biotropica, 17:269-276, Lodge, D.J., F.N. Scatena, C.E. Asbury, & M.J. Sanchez. 1991. Fine litterfall and related nutri- ent inputs resulting from Hurricane Hugo in sub- tropical wet and lower montane rain forests of Puerto Rico. Biotropica, 23:336-342. Medina, E., E. Cuevas & PL. Weaver. 1981. Com- posicion foliar de especies lefiosas de Pico del Este, Sierra de Luquillo, Puerto Rico. Acta Cient. Veeezolana, 32:159-165. Moulder, B.C, & D.E. Reichle. 1972. Significance of spider predation in the energy dynamics of forest-floor arthropod communities. Ecol. Mon- ogr., 42:473-498. Olson, D.M. 1994. The distribution of leaf litter invertebrates along a neotropical altitudinal gra- dient. J. Trop. EcoL, 10:129-150. Ott, R.L. 1993. An Introduction to Statistical Methods and Data Analysis, 4th ed. Duxbury Press, California. Petrunkevitch, A. 1929. The spiders of Porto Rico, VARGAS-=-EFFECTS OF FERTELIZER AND DEBRIS REMOVAL 89 part L Trans. Connecticut Acad. Arts Sci., 30:1- 158. Petrunkevitch, A. 1930a. The spiders of Porto Rico, part 2. Trans. Connecticut Acad. Arts Sci., 30:159-355. Petrunkevitch, A. 1930b. The spiders of Porto Rico, part 3. Trans. Connecticut Acad. Arts Sci., 31:1-191. Pfeiffer, W.J. 1996. Litter invertebrates. Pp. 138- 181,//! The Food Web of a Tropical Rain Forest. (D.P. Reagan & R.B. Waide, eds.). Univ. Chicago Press, Chicago. Platnick, N.L 1989. Advances in Spider Taxono- my: A Supplement to Brignoli’s “A Catalogue of the Araneae Described Between 1940 and 1981”. Manchester Univ. Press. New York. Polls, G.A. & S.D. Hurd. 1995. Extraordinarily high spider densities on islands: Flow of energy from the marine to terrestrial food webs and the absence of predation. Proc. Nat. Acad. Sci. USA, 92:4382-4386. Power, M.E. 1992. Top-down and bottom-up forc- es in food webs: Do plants have primacy? Ecol- ogy, 73:733-746. Reagan, D.P. 1996. Anoline lizards. Pp. 321-345, In The Food Web of a Tropical Rain Forest. (D.P. Reagan & R.B. Waide, eds.). Univ. Chicago Press, Chicago. Riechert, S.E. & T. Lockley. 1984. Spiders as bi- ological control agents. Ann. Rev. EntomoL, 29: 299-320. Russell-Smith, A. & N.E. Stork. 1995. Composi- tion of spider communities in the canopies of rainforest trees in Borneo. J. Trop. EcoL, 1 1:223- 235. Sanford, R.L., Jr., WJ. Parton, D.S. Ojima & DJ. Lodge. 1991. Hurricane effects on soil organic matter dynamics and forest production in the Lu- quillo Experimental Forest, Puerto Rico: Results of simulation modeling, Biotropica, 23:364-372. Steudler, P.A., J.M. Melillo, R.D. Bowden, M.S. Castro & A.E. Lugo. 1991. The effects of nat- ural and human disturbances on soil nitrogen dy- namics and trace gas fluxes in a Puerto Rican wet forest. Biotropica, 23:356-363. Stewart, M.M. & L.L. Woolbright, 1996. Amphib- ians. Pp. 273-320, In The Food Web of a Trop- ical Rain Forest. (D.P. Reagan & R.B. Waide, eds.). Univ, Chicago Press, Chicago. Thomas, R. & A. Gaa Kessler. 1996. Non-anolirie reptiles. Pp. 347-362, In The Food Web of a Tropical Rain Forest. (D.P. Reagan & R.B. Wai- de, eds.), Univ. Chicago Press, Chicago. Uetz, G.W. 1976. Gradient analysis of spider com- munities in a streamside forest. Oecologia, 22: 373-385. Uetz, G.W. 1979, The influence of variation in lit- ter habitats on spider communities. Oecologia, 40:29-42. Vince, S.W. , L Valiela & J.M. Teal. 1981. An ex- perimental study of the structure of herbivorous insect communities in a salt marsh. Ecology, 62: 1662-1678. Walker, L.R., J.K. Zimmerman, DJ. Lodge & S. Guzman-Grajales. 1996. An altitudinal compar- ison of growth and species composition in hur- ricane damaged forests in Puerto Rico. J. EcoL, 84:877-889. Weaver, P.L., E. Medina, D, Pool, K. Dugger, J. Gonzilez-Liboy, & E. Cuevas. 1986. Ecological observations in the dwarf cloud forest of the Lu- quillo mountains in Puerto Rico. Biotropica, 18: 79-85. Weaver, P.L. & P.G. Murphy. 1990. Forest structure and productivity in Puerto Rico’s Luquillo Mountains. Biotropica, 22:69-82. Willig, M.R. & G.R. Camilo. 1991. The effect of Hurricane Hugo on six invertebrate species in the Luquillo Experimental Forest of Puerto Rico. Biotropica, 23:455-461. Wise, D.H. 1993. Spiders In Ecological Webs. Cambridge Univ. Press, Cambridge. Wolda, H. 1983. Diversity, diversity indices and tropical cockroaches. Oecologia, 58:290-298. Woolbright, L.L. 1996. Disturbance influences long-term population patterns in the Puerto Ri- can frog, Eieutherodactylus coqui (Anura: Lep- todactylidae). Biotropica, 28:493-501. Zar, J.H. 1984. Biostatistical Analysis. 2nd ed. Prentice-Hall, Inc. New Jersey. Zimmerman, J.K„ W.M. Pulliam, D.J, Lodge, V. Quinones-Orfila, N. Fetcher, S. Guzman-Graja- les, J.A. Parrotta, C.E. Asbury, L.R. Walker & R.B. Waide. 1995. Nitrogen immobilization by decomposing woody debris and the recovery of tropical wet forest from hunicane damage. Oi- kos, 72:314-322. Manuscript received 28 September 1998, revised 1 July 1999. 2000. The Journal of Arachnology 28:90-96 A TWENTY- YEAR COMPARISON OF EPIGEIC SPIDER COMMUNITIES (ARANEAE) OF DANISH COASTAL HEATH HABITATS Peter Gajdos’’^ and S0ren Toft*; ^Department of Zoology, Institute of Biological Sciences, University of Aarhus, Universitetsparken, Building 135, DK-8000 Aarhus, Denmark; ^Institute of Landscape Ecology, Bratislava, Branch Nitra, Slovak Academy of Sciences, Akademicka 2, FOB ~23B, 949 01 Nitra, Slovak Republic ABSTRACT. The same epigeic spider communities of North-west Jutland coastal heath habitats (Den- mark, region Thy) initially surveyed by pitfall traps from 1977-79 were examined 20 years later (1997- 98). The heath plots were open sandy areas growing into Calluna heath, and the more stable Erica, Calluna/Empetrum, Molinia vegetation types. They have changed vegetatively only a little in those 20 years from natural succession. Though the spider communities of all areas showed only minor changes from the passage of time, these were larger than the differences attributable to the different habitat types despite large differences in soil humidity and vegetation structure. Keywords; Spiders, community composition changes, coastal heathland habitats, Denmark The impact of habitat management on spi- der communities is usually analyzed by com- paring variously treated or disturbed/undis- turbed sites at the same point in time. Other studies have followed changes in the spider fauna associated with specific disturbances and the subsequent successional recovery by collecting over a series of years (heathland fires: Merrett 1976; forest fires: Huhta 1971, Schaefer 1980). Studies of long-term changes in the spider fauna of a particular site due to human modification of the biotope, pollution, vegetational succession, climatic change, or other causes are rare. Schikora (1994) found relatively minor changes over 22 years in the spider species of a bog in spite of prominent vegetation changes due to drainage. However, the dominant spiders had changed from pho- tophilous to skotophilous species. Hanggi & Maurer (1982) compared the spider fauna of a Swiss raised bog after a 50-year interval, but the collecting methods were different. No known studies has analyzed the spider fauna of a specific locality by comparable methods over a long temporal scale. Our aim in the present study is to compare the epigeic spider communities of four adja- cent Danish coastal heathland habitats sam- pled by pitfall traps in the same locations with an interval of 20 years. We compare similar- ities in community composition both between habitats and between time periods. The veg- etation of the area was only slightly affected by human activity in the intervening time. One site, which had been disturbed at the time of the first investigation, had been changed by vegetational succession during the 20 years. METHODS Site descriptions.— The study was carried out at Tprvekjaer by Vester Vanned S0, Thy, Denmark (57°r30"N, 8°32'E). The area con- sisted of coastal heathland patches between an oligotrophic marsh to the south and sandy pastures and a coniferous plantation to the east and west. The plantation was established in the late 1950’s and has provided increasingly more protection from the wind as it matured. The North Sea coast is ca. 3.5 km to the west. Four points were sampled on a N-S transect perpendicular to low sandy ridges deposited by the prevailing westerly winds, which cre- ated alternating depressions and “hills” with varying vegetation. The differences in height between hills and depressions were never more than 0.5 m. However, the depressions could flood during winter, resulting in very divergent vegetational characteristics at the sites. Traps 1-2 were situated on a hilltop, 3 m from the edge of a spruce plantation. In 1977- 79 the traps were placed in a patch of bare 90 GAJDOS & TOFT^^HANGES IN HEATHLAND SPIDERS 91 sand created by human disturbance. In 1997 the vegetation had recovered completely and grown into a typical dry heathland patch, dominated by Calluna vulgaris (L.) (coverage 50%), Empetrum nigrum L. (30%), and moss as groundcover. Traps 3-4 were in a moist depression, 17 m from traps 1-2. The vegetation was domi- nated by Erica tetralix L, (90%) with moss covering the ground, and showed no recog- nizable changes between the two sampling pe- riods. Traps 5-6 were on a hilltop, 9 m from traps 3-4. The general character of the vegetation was dry dwarf-shrab heath. In 1977-79 it con- sisted of a mixture of Calluna and Empetrum; in 1997 Empetrum (70%) with moss ground- cover was clearly dominant. Traps 7“-8 were in the next depression, 17 m from traps 5-6. In both sampling periods the vegetation was a nearly~pure dense stand of the low grass Molinia caerulea (L.) (>90%), indicating a very moist soil. Whereas the first three sites were dwarf- shrub heaths, this site was better characterized as a meadow. Trapping.— Pitfall traps were used to mon- itor the active densities of ground-dwelling spiders at the selected sites. In 1911-19 glass jars (diameter 8 cm) were used, but in 1997 we used plastic beakers fitted into plastic flower pots (diameter 11 cm). A 3% formalin solution with ethylene glycol and detergent was used as a killing agent and preservative on both occasions. Trapping periods were 7 May 1977-23 March 1978, 13 May 1978-17 March 1979, and 11 May 1997-21 March 1998. The traps were emptied every 2-3 weeks during the warm seasons, and once a month or more infrequently during the winter periods. Pitfall traps were placed in pairs at each site ca. 1-2 rn apart. When trapping was repeated in 1997-98 the new traps were placed as close as possible to the same positions as was used earlier, all probably less than 2 m away. The spider material is deposited in the col- lection of Zoological Museum, Copenhagen. Weather,— We obtained weather informa- tion from the Danish Meteorological Institute. For 1977-79 data are from station Silstrup, for 1997-98 from Hprsted. Both are ca. 15 km from the study area. We used the monthly av- erages of temperature, sunny hours, and rain- fall (Fig. 1). Figure L— -Weather conditions during the three study periods, all plotted as monthly averages. A: Temperature (®C), B: Hours of sunshine, C: Rainfall (mm). Overall 1997-98 was somewhat sunnier and warmer and with less precipitation than 1977--79. The average differences between the three trapping periods are quite small, and the conditions of 1997-98 do not deviate more from either of the early periods than from each other (Fig. 1). Data analysis.— Comparison of the epigeic spider communities was made by Principal Component Analysis (PCA) using the CAN- OCO program (ter Braak 1987; cf. Jongmae et al, 1987). We compared the summed catch- es for each site and catching period and ap- plied the PCA to log-transfonned abundances of each species. All species were included in the analyses (and thus in determining the rel- ative distribution of the trap sites (Fig. 2)), but only the most dominant species (> 2% at one site and period) are presented in the species plot (Fig. 3), We also illustrate the species that disappeared or appeared between 1977-79 and 1997-98, as well as less dominant species (though > 0.4%) which showed substantial changes in relative abundance (Fig. 4). Ad- ditionally, we compared the dominance struc- ture and species composition of the habitats (Fig. 5), and analyzed the species changes be- tween the two periods. Two similarity indices for pairwise comparisons (Southwood 1966) were calculated: the S0rensen Quotient of 92 THE JOURNAL OF ARACHNOLOGY PC1 Figure 2. Principal Component Analysis (site plot) of summed catches from each of four trap sites for three trapping periods (1977-78, 1978-79 and 1997-98), Starting year and trap-numbers indicated next to each point. Traps 1-2: Bare sand (1977, 1978) or Calluna (1997); Traps 3-4: Erica; Traps 5-6: Empetrum/Cailuna; Traps 7-8: Molinia. similarity, QS = 2j/(a + b), where a and b are the number of species in the two samples, and j is the number of species common to both samples; and the Percentage of similarity, %S = Xi min (Pia, Pib), which sums the lowest vaL ues for the proportional abundances (p) of each species (i) in the two samples (a, b). RESULTS Faunistic characteristics*^— A total of 6368 specimens belonging to 113 species was collected, of which 23 species had a relative dominance of >2% in at least one site and year. The number of individuals and species at each site increased for 1997-98 compared to the earlier periods (Table 1). It can therefore Figure 3.— Principal Component Analysis (spe- cies plot, same analysis as Figure 2) of summed catches from each of four trap sites for three trap- ping periods (1977-78, 1978-79 and 1997-98), il- lustrating dominant species (>2%). Abbreviations of species names: see Appendix 1. be concluded that there has been no decline in the richness of the spider fauna over the 20 years. The number of species in common between the trapping periods was extremely stable (Ta- ble 2). It is remarkable that the number of spe- cies disappearing between 1977-78 and 1978- 79 was the same as between 1977-78 and 1997-98. This may indicate that “disappear- ance” does not necessarily mean extinction but rather reflects chance of capture. Given the low number of traps in each habitat, this effect is not surprising. More new species seem to have accumulated over the 20 year period than between 1977-78 and 1978-79, but this may also be due to the higher number of individ- uals caught in 1997-98. Table 1.— The number of individuals and species of spiders collected by two traps at each of four trapping stations during the three trapping periods. 1977-78 1978-79 1997-98 Ind. Species Ind. Species Ind. Species Traps 1-2 317 46 429 46 465 56 Traps 3-4 513 47 536 41 671 47 Traps 5-6 583 54 487 38 532 50 Traps 7-8 584 43 544 41 707 54 Total 1997 80 1996 71 2375 87 GAJDOS & TOFT— CHANGES IN HEATHLAND SPIDERS 93 Figure 4 —Principal Component Analysis (spe- cies plot, same analysis as Fig. 2) of summed catch- es from each of four trap sites for three trapping periods (1977-78, 1978-79 and 1997-98), illus- trating species appearing (App.) or disappearing (Dis.) between 1977-79 and 1997-98 (maximal dominance value (D) in year of presence indicated), and non-dominants showing substantial changes (increase/decrease) between these periods. Abbre- viations of species names: see Appendix 1. Principal component analysis. — Analysis of spider species abundances in each of the traps during every trapping period resulted in a plot of sites X years (Fig. 2) and correspond- ing plots of species distributions (Fig. 3, 4). The sites clustered in three groups (Fig. 2). One group consisted of traps 1-2 from 1977- 78 and 1978-79; a second group was formed by all other sites for the same two periods, and the third group included the four sites from 1997-98. Thus, except for traps 1-2 in the early periods, the habitats do not cluster together while the catching periods do. This means that the temporal changes in the fauna are more prominent than the differences be- tween the habitats. Axis 1 (PCI) mainly re- flects the changes resulting from vegetational succession at traps 1-2, while axis 2 (PC2) reflects the faunistic changes taking place at the remaining trapping sites over 20 years, which cannot be easily related to specific hab- itat changes. There seem to be no further re- lationships between the two PC-axes and characteristics of the habitats. The change of the spider community at traps 1-2 was ex- pected because this site was a disturbed patch of bare sand, which succession eventually turned into a plant community similar to that of site 5-6. We repeated this analysis for spring/summer (May-September) and autumn/ winter (October-March) catches separately. Both data sets gave the same pattern as for the full periods. The dominant species were concentrated in the central part of the PC-plot (Fig. 3), reflect- ing a high similarity in species composition between the habitats within a period. This plot and the following (Fig. 4) show the differenc- es in species composition responsible for the pattern in Fig. 2, and the axes should be in- terpreted similarly. Species in the upper left are those that increased in abundance after 20 years, while those to the right and in the lower part decreased. Species that either appeared or disappeared during the 20 years or had a rel- ative abundance of < 2% showed a clear sep- aration of increasing/appearing vs. decreasing/ disappearing (Fig. 4). Both types of similarity indices between years produced values between 70-80% (Ta- ble 2). The two early periods were not more similar than early versus late periods. Dominance structure.— The same few species were the dominants in all four habitats (Fig. 5), with Gnahosa leporina (20) and Cen- tromerita concinna (3) being at positions 1-3 Table 2.-— Comparison of spider population characteristics between trapping periods (catches from different habitats summed for each year). 77-78 vs. 78-79 77-78 vs. 97-98 78-79 vs. 97-98 Number species both periods 57 58 59 Number species disappearing 23 22 12 Number species appearing 14 27 28 Sprensen’s quotient of similarity 75.5 70.3 79.7 Percent similarity 69.0 79.3 69.5 94 THE JOURNAL OF ARACHNOLOGY Rank Figure 5.— Dominance curves for spider communities of four heathland trap-sites in three trapping periods (12 most abundant species only). Numbers indicate species identity (cf. Appendix 1). in all habitats and years (Fig. 5). Only in the Molinia habitat they were surpassed by Par- dos a pullata (14) and Agroeca proximo (16). Pardosa nigriceps (13) was codominant at the two “hilly” sites, while P. pullata had its highest dominance in the moist depressions. When comparing the four sites over the 20 years there are no indications of systematic changes in the dominance structure. Five spe- cies occurred with > 2% dominance in all years at all sites: C concinna, P. nigriceps, P. pullata, G. leporina, and Haplodrassus cu- preus. Species appearances and disappearanc- es.—At trap sites 1-2, several species prefer- ring bare sandy areas were caught during 1977-79 but disappeared later when the veg- etation closed. That was true for Arctosa per- ita, while Pardosa monticola, Aelurillus v-in- signitus and Zelotes electus decreased in abundance. Some vegetation-dwelling species, like Linyphia triangularis and Philodromus histrio, seemingly decreased. It is possible, however, that this is an artifact because they are found only on the soil surface if there is no vegetation. Macrargus rufus declined at some sites while M. carpenter! appeared in considerable numbers in 1997-98. We are unable to relate this shift to the small habitat or environmental changes between the sampling periods. In- creasing species (cf. Figs. 3, 4) include species that in Denmark are associated mainly with (dry) heathlands (P. ludicrum, Z. iatreiiiei, D. cupreus), while others are hygrophilous (M beata) or even moist heathland specialists {G. leporina). Thus, it is impossible to determine a direc- tion of change with respect to the ecological characteristics associated with species whose abundances changed. GAJDOS & TOFT—CHANGES IN HEATHLAND SPIDERS 95 DISCUSSION Structural characteristics of the vegetation are generally thought to be the most important factor for habitat selection of spiders and thus for determining the composition of the spider fauna (Duffey 1962, 1966, 1968; Curtis & Bignal 1980; Robinson 1981). We therefore expected spider communities in different hab- itats to show large differences relative to the temporal changes, especially at the two sites where no vegetational changes had occurred. We observed the opposite in spite of great dif- ferences in vegetational physiognomy be- tween some of the sites. The Molinia meadow and the Calluna/Empetrum heathland sites were very different both in vegetation struc- ture and soil moisture; the Erica and Molinia sites were similar in soil moisture but different in vegetation structure, and the Erica and the Calluna/Empetrum sites were somewhat sim- ilar in vegetational structure (all dwarf shrubs) but different in soil moisture. Yet, all were quite similar in their spider fauna. We found relatively large differences between bare and vegetated habitats, probably because bare sandy areas are without vegetational structure and also microclimatically extreme. Several xerophilic spider species are specialists of this habitat type. Temporal changes in the spider community composition were greater than differences be- tween habitats. This was not due to any dra- matic changes over the years in the compo- sition of the spider communities, however, because even these changes were quite small. This is not only evident from the high simi- larities, but also from a consideration of the specific changes. For example, the most abun- dant species that disappeared had a dominance score of only 2.2% at the site of highest abun- dance (Arctosa perita at traps 1-2). The ap- pearing species that became most abundant reached a dominance score of 2.4% {Macrar- gus carpenteri at traps 5-6). Among the dom- inants the greatest difference in dominance score between 1977-79 and 1997-98 (all sites combined) was <5%. Thus, viewed over the 20 years, the composition of the spider fauna has been very stable. On a still longer time scale these communities will certainly not be maintained; most likely the area will be in- vaded by shrubs (a process already started) and eventually trees, and thus the vegetation type will change completely, unless main- tained by management. This development is accelerated by the planting of the forest that surrounds the heathland area, creating a much milder microclimate than before and provid- ing invasive tree species. The reasons for the temporal changes should be considered. For trap-sites 1-2, veg- etational succession following a disurbance is the obvious cause. For the remaining sites the question is more difficult. We could see no pattern in the ecological preferences of spe- cies that decreased/disappeared or appeared/ increased. The weather in 1 997-97 was slight- ly warmer and dryer than before, but it is difficult to relate the specific faunistic changes to this fact, since the differences between the two early periods are as large as between early and late periods. LITERATURE CITED Curtis, DJ. & E.M. Bignal. 1980. Variations in peat bog spider communities related to environ- mental heterogeneity. Pp. 81-86, In Proc. 8th In- ternat. Congr. ArachnoL, Vienna. Duffey, E. 1966. Spider ecology and habitat struc- ture. Senckenbergiana Biol., 47:45-49. Duffey, E. 1968. An ecological analysis of the spi- der fauna of sand dunes. J. Anim. Eco!., 37:641- 674. Duffey, E. 1962. A population study of spiders in limestone grassland: the field-layer fauna. Oikos, 13:15-34. Huhta, V. 1971. Succession in the spider commu- nities of the forest floor after clear-cutting and prescribed burning. Ann. Zool. Fennica, 8:483- 542. Jongman, R.H.G., C.F.J. ter Braak & O.F.R. van Tongeren. 1987. Data Analysis in Community and Landscape Ecology. Pudoc, Wageningen. 299 pp. Merrett, P. 1976, Changes in the ground-living spi- der fauna after heathland fires in Dorset. Bull. British ArachnoL Soc., 3:214-221. Robinson, J.V. 1981. The effect of architectural variation in habitat on a spider community: an experimental field study. Ecology, 62:73-80. Schaefer, M. 1980. Sukzession von Arthropoden in verbrannten Kiefernforsten, II. Spinnen (Aranei- da) und Weberknechte (Opilionida). Forstwiss. CentralbL, 99:341-356. Schikora, H.-B, 1994. Changes in the terrestrial spider fauna (Arachnida: Araneae) of a North German raised bog disturbed by human influ- ence. 1964-1965 and 1986-1987: a comparison. Mem. Entomol. Soc, Canada, 169:61-71. 96 THE JOURNAL OF ARACHNOLOGY Southwood, TR.E. 1966. Ecological Methods. Me- thuen & Co. Ltd., London, ter Braak, CJ.F. 1987. CANOCO (version 2.1). Agricultural Mathematics Group, Wageningen. 95 pp. Manuscript received 18 May 1998, revised 12 Au- gust 1999. APPENDIX 1 Nomenclature, name abbreviations and number codes for species mentioned in text and figures. Ael.vin Arc. per Agr.pro Alo.pul Bat.gra Bol.lut Cen.con Che, err Che.vir Clu.tri Dra.cup Dra.pus Euo.fro Eva.mer Gna.lep Gon.rab Gon.viv Hap.sig Lep.eri Lep.men Lin.tri Mac. car Mac.raf Mei.bea Mei.rur Par.mon Par. nig Par. pul Pep.lud Phi. his Pho.gib Robiiv Saa.abn Sco.gra Tro.ter Typ.dig Wal.ant Wal.mon Wal.nud WaLuni Zel.ele ZeLlat Zor.spi Aelurillus v-insignitus (Clerck) Arctosa perita (Latr.) Agroeca proxima (O.P.-C.) (16) Alopecosa pulverulenta (Clerck) Bathyphantes gracilis (BL) Bolyphantes luteoius (BL) Centromerita concinna (Thor.) (3) Cheiracanthium erraticum (Walck.) Cheiracanthium virescens (Sund.) Clubiona triviaiis C.L.K. Drassodes cupreus (BL) (18) Drasyllus pusiUus (C.L.K.) (19) Euophrys frontalis (Walck.) Evamia merens O.P.-C. Gnaphosa leporina (L.K.) (20) Gonatium rubens (BL) (4) Gongylidiellum vivum (O.P.-C.) Hapiodrassus signifer (C.L.K.) (21) Lepthyphantes ericaeus (BL) (5) Lepthyphantes mengei Kulcz. (6) Linyphia triangularis (Clerck) Macrargus carpenteri (O.P.-C.) (7) Macrargus rufus (Wider) Meioneta beata (O.P.-C.) Meioneta rurestris (C.L.K.) Pardosa monticola (Clerck) Pardosa nigriceps (Thor.) (13) Pardosa pullata (Clerck) (14) Peponecranium iudicrum (O.P.-C.) (8) Philodromus histrio (Latr.) Pholcomma gibbum (Westr.) (1) Robertus lividus (BL) (2) Saaristoa abnormis (BL) Scotina graciiipes (BL) (17) Trochosa terricola Thor. (15) Typhocrestus digitatus (O.P.-C.) Walckenaeria antica (Wider) Walckenaeria monoceros (Wider) (10) Walckenaeria nudipalpis (Westr.) Walckenaeria unicornis O.P.-C. Zelotes electus (C.L.K.) Zelotes iatreiilei (Simon) (22) Zora spinimana (Sund.) (23) 2000. The Journal of Arachnology 28:97-106 HABITAT DISTRIBUTION, LIFE HISTORY AND BEHAVIOR OF TETRAGNATHA SPIDER SPECIES IN THE GREAT SMOKY MOUNTAINS NATIONAL PARK Marie Aiken and Frederick A. Coyle; Department of Biology, Western Carolina University, Cullowhee, North Carolina 28723 USA ABSTRACT. Habitat distribution patterns of five species of Tetragnatha Latreille 1804 were studied by analyzing 1163 one-hour samples collected at 17 focal sites representing 16 major biotic communities (habitats) in the Great Smoky Mountains National Park. Tetragnatha versicolor Walckenaer 1841 is a habitat generalist, being common over a wide range of elevations (520-1755 m) and in 10 of the 16 habitats, including seven forest habitats as well as wetland, high grass bald, and grassland habitats. Te- tragnatha laboriosa Hentz 1850 is virtually restricted to non-wetland grassy habitats, T elongata Wal- ckenaer 1805 to streams, T. viridis Walckenaer 1841 to hemlock trees, and T. straminea Emerton 1884 to non-forested wetlands (marshes). Microhabitat segregation exists in the high grass bald community between T. versicolor (prefers trees and shrubs) and T laboriosa (prefers herbs). Size frequency histograms of seasonal samples of T. straminea specimens indicate that this species has a one-year life cycle with six post-emergent instars, and that most individuals overwinter in the antepenultimate instar and mature and mate in May and June. Tetragnatha straminea is able to capture prey with or without using a web and adopts stick-like cryptic postures in three different contexts. Keywords; Tetragnatha, spider, habitat preference, life cycle, cryptic behavior Being highly diverse and abundant preda- tors, spiders are important regulators of ter- restrial arthropod populations (Riechert & Bishop 1990; Coddington & Levi 1991; Mor- an et al. 1996) and may prove to be useful indicators of the overall species richness and health of terrestrial communities (Noss 1990; Kremen et al. 1993; Colwell & Coddington 1994; Hanggi et al. 1995). But progress to- ward understanding the ecological roles of spiders is limited by a lack of knowledge of the habitat preferences and life histories of many species (Duffy 1978; Hanggi et al. 1995). Ecologists must know the autecology and life histories of important constituent spe- cies before they can gain key insights into food web dynamics and other aspects of a community’s dynamics (Olive 1980; Strong et al. 1984; Wilson 1992; Polls et al. 1996). Tetragnatha Latreille 1 804 may be the most widespread and abundant orb-weaving spider genus in the world (Levi 1981). Tetragnatha species live in tropical, temperate, and arctic climates and on all continents (except Antarc- tica) and many islands. On the Hawaiian Is- lands a major adaptive radiation of Tetrag- natha species has been discovered (Gillespie & Croom 1995). Fifteen Tetragnatha species are known from North America north of Mex- ico (Levi 1981), and some of these are nu- merically dominant spiders in particular hab- itats and over whole regions (Lowrie 1953; LeSar & Unzicker 1978). Despite the promi- nence of this genus, the life histories of only one North American species (7. laboriosa Hentz 1850) and a few species in other parts of the world have been rigorously analyzed and described (Juberthie 1954; Toft 1976; LeSar & Unzicker 1978), and knowledge of the habitat preferences of North American Te- tragnatha species consists of collecting re- cords and comments scattered widely in the literature. In this study, we describe the habitat distri- bution patterns of five Tetragnatha species found in the Great Smoky Mountains National Park Biosphere Reserve (GSMNP) by using large sets of spider samples collected from 16 major habitats with a standardized protocol used to inventory the spiders of the GSMNP. Located in the southern Appalachian Moun- tains, the GSMNP, due partly to its wide ele- vation range (275-2013 m), large size (207,000 ha), and low temperate latitude 97 98 THE JOURNAL OF ARACHNOLOGY (35°35'N), comprises a rich mosaic of biotic communities appropriate for investigating habitat preferences on a landscape scale. We also provide the first analysis and description of the life history, phenology, and behavior of Tetragnatha straminea Emerton 1884. Our main goal is to make this important assem- blage of spiders more accessible to ecologists. METHODS Habitat distribution.— -Teams of 3-5 (usu- ally 4) collectors used a modified Coddington sampling protocol (Coddington et al. 1996) to obtain the 1163 one-hour ground (408), aerial (310), beat (360), and sweep (85) samples of spiders used in this project. Ground sampling involved searching below knee level mostly on hands and knees, exploring leaf litter, logs, rocks, and plant surfaces. Aerial sampling in- volved searching foliage, branches, tree trunks, and spaces in between, from knee height up to maximum overhead arm’s reach. Beating consisted of striking vegetation with aim long stick and dislodging spiders onto a 0.5 m^ canvas sheet held horizontally below the vegetation. Hands and aspirators were used to collect the spiders into vials contain- ing 80% ethanol. One sample unit equaled one hour of uninterrupted effort using one of these three methods during which the collector at- tempted to collect every spider encountered. During each hour the team as a whole typi- cally used all three methods in the same area. In non-forest communities (grass bald, wet- land, and native grassland sites) one-hour sweep sampling was substituted for aerial and/ or beating methods; sturdy sweep nets with 38 cm diameter hoops were used, and the number of sweeps per hour (175-400, mean and SD — 268 ± 48) depended primarily on vegeta- tion structure and spider abundance. Two sets of samples (one in the spring and one in late summer) were collected in each of two years (1996 and 1997) from 15 sites and in 1995 from two other sites, the low grass bald and heath bald sites. These 17 focal sites were selected by GSMNP ecologists to rep- resent the 16 major habitat (community) types found in the GSMNP. Habitat type, locality data, collecting dates, and sampling effort for each focal site are given in the Appendix. Two montane wetland focal sites were chosen be- cause each one was too small to support the sampling effort judged necessary for this study. For a given focal site, the number of samples collected in the spring and summer were equal or very nearly so, as were the number of samples collected in 1996 and 1997. At each site (with the exception of the high grass bald and both montane wetland sites) nearly equal numbers of samples were collected with each of the methods employed. Descriptions of most of the sampled commu- nity types can be found in Whittaker (1956). Vegetation is being analyzed at each focal site by GSMNP botanists, and the results of these analyses will be posted in one or two years on the World Wide Web. Adult and juvenile Tetragnatha specimens were sorted from each sample and identified to species. By using eye arrangement, pigment pattern, and abdominal shape, we were able to identify all but about 1% of the juveniles. Tetragnatha versicolor Walckenaer 1841 and T. laboriosa juveniles cannot be separated by eye and body shape characters, but can be dis- tinguished by the following: in versicolor the black pigment area surrounding each lateral eye touches that of its neighboring lateral eye (clearly separate in laboriosa except in some of the youngest individuals), the abdominal venter is light (dark in laboriosa), and the sil- ver pigment dorsally on the abdomen of the smallest specimens is often interrupted by a median dorsal line of no pigment (not inter- rupted in laboriosa). All specimens will be deposited in the Smithsonian Institution. The relative abundance (mean number of individuals per one-hour sample) of each spe- cies in each year was computed for each of the 17 sites. It is important to note that this index of abundance does not reveal the often wide variation in number of individuals among one-hour samples at each site, varia- tion due largely to method bias to particular microhabitats, spatial environmental variation within each site, and seasonal changes in spi- der abundance correlated with species’ phe- nologies. An ANOVA (StatView 4.5 from Abacus Concepts) was used to examine the effect of year and method on spider abun- dance; P < 0.05 was our significance criteri- on. Life history.— We measured the length of the left tibia I (ITL) (along the dorsal surface) of all 220 T. straminea specimens collected at the two montane wetland sites, Meadow Branch marsh (15 May and 17 July 1996; 23 AIKEN & COTLE—TETRAGNATHA SPECIES IN THE GREAT SMOKY MOUNTAINS 99 May, 1 August, and 7 October 1997) and In- dian Creek marsh (27 May and 16 August 1996; 12 May and 29 July 1997). Toft (1976) demonstrated that ITL often distinguishes spi- der instars more clearly than does either the length or width of the carapace. Measure- ments were performed with a Wild M-5 ste- reomicroscope at 24 X and 12X magnification and are accurate to ± 0.077 mm. We used the StatView 4.5 computer program to generate ITL frequency distribution histograms. By ex- amining these histograms of seasonal subsets (spring, summer, and fall) of data pooled from both sites, it was possible to reveal phenology (seasonal timing of development) and gener- ation time (life cycle length). The histogram for all data pooled revealed the total number of instars. Behavior.-— We observed and photo- graphed live specimens in the field. Several T. straminea juveniles (antepenultimate instar) were placed in separate terraria and main- tained for several weeks on Drosophila flies while we observed prey capture and cryptic postures, sometimes using a hand-held mag- nifier. RESULTS Habitat distribution.-”Five species of Te- tragnatha were collected in the GSMNP: T. elongata Walckenaer 1805, T. laboriosa, T. straminea, T. versicolor, and T. viridis Wal- ckenaer 1841. Table 1 and Fig. 1 show the relative abundance of these species at each fo- cal site. Tetragnatha versicolor was found at 16 of the 17 focal sites and was common (rel- ative abundance = 0.5-2.0) or abundant (rel- ative abundance > 2.0) in 10 of the 16 habi- tats, including seven forest habitats as well as montane wetland, high grass bald, and native grassland habitats. It was especially abundant in mixed oak forest, Tetragnatha laboriosa was found at nine sites, but was rare at all but two of these sites, native grassland and high grass bald. Tetragnatha versicolor and T. la- boriosa were found over a wide elevational range (520-1830 m). Tetragnatha straminea was collected at only three sites, the montane wetland and native grassland sites, and was common or abundant at all three. Tetragnatha elongata was found only at the two sites through which streams flow. Tetragnatha vir- idis was found only at the two sites where hemlock trees are abundant. No Tetragnatha species were common at the spruce-fir, spruce, northern hardwood, low grass bald, or heath bald sites, and none were collected at the pine-oak (395 m) site (Table 1, Fig. 1). Sites with two or more common species of Tetragnatha were the high grass bald {versicolor and laboriosa), both wetlands {versicolor and straminea), and the native grassland {versicolor, laboriosa, and strami- nea) (Fig. 1). There were significant relative abundance differences between 1996 and 1997 for T. ver- sicolor at the mixed oak. Table Mountain pine, hemlock/hardwood cove, hardwood cove, and Meadow Branch wetland sites, and for T. laboriosa at the native grassland (Fig. 1). In each case, the relative abundance was higher in 1997. Microhabitat distribution.— At the high grass bald, T. laboriosa was more abundant in sweep samples (collected from herbaceous vegetation) than in beat samples (collected from shrubs and trees) {F = 5.64, df = I, P = 0.025), whereas T. versicolor was more abundant in beat than in sweep samples {F = 5.64, df^ 1, P = 0.025) (Fig. 2). At the In- dian Creek wetland site, T straminea was more abundant in sweep samples than in beat samples (F = 5.17, # = 1, P - 0.041), but r. versicolor was equally common in both sweep and beat samples (F = 0.24, df — P = 0.632) (Fig. 3). Although we were unable to make this kind of microhabitat comparison at the Meadow Branch wetland or native grassland sites (because the beat method was not used at these sites), we observed that T. straminea was more common in the low grassy vegetation of the wetter parts of these habitats than was T. versicolor. The few spec- imens of T. viridis that were found were col- lected only by beating the foliage of hemlock trees. Tetragnatha elongata was collected only over the small streams flowing through the hemlock and native grassland sites. Life history of T, straminea. — The size frequency histogram of all T. straminea indi- viduals collected at the wetland sites during both years indicates a total of six size/age classes and, therefore, six post-emergent in- stars (instars living outside the egg sac) (Fig. 4). As is typical for spiders (Toft 1976; Coyle 1985) the older the instar, the greater the var- iation in size. For two reasons, we suspect that the ITL frequency peak between 4.5 and 5.0 100 THE JOURNAL OF ARACHNOLOGY Table 1. — -Relative abundance of Tetragnatha species at 17 focal sites representing 16 biotic commu- nities in the Great Smoky Mountains National Park in both 1996 and 1997 (1996 and 1997 values are separated by a comma). Low grass and heath balds were sampled in 1995 only. Elevation (m) of each site is given in parentheses. Relative abundance value is underlined if at least one adult was collected. Relative abundance (mean number of individuals per sample) Habitat/focal site elongata laboriosa straminea versicolor viridis Spruce-fir (1830) 0, 0.04 0, 0.08 High grass bald (1755) 0.88, 1.67 1.46, 2.25 Spruce (1715) 0, 0.08 0,13,0.04 Beech gap (1645) 0.50, 0,21 Northern hardwood (1615) 0.16, 0.30 Red oak (1555) 0.40, 1.08 Low grass bald (1505) 0.17 0.40 Heath bald (1390) 0.10 Mixed oak (1115) 5.82, 18.0 Table Mtn. pine (1005) 0.02, 0 0.06, 0.58 Hemlock-hardwood cove (945) 1.15,2.75 0.04,0.06 Hemlock (885) 0.17, 0.19 1.73,3.36 0.02,0.03 Hardwood cove (740) 0, 0.02 0.43, 1.25 Wetland (Indian Cr.) (685) 0.06, 0 2.00,3.31 0.24, 1.38 Wetland (Meadow Br.) (535) 0, 0.19 2.00, 3.94 0.24,3.31 Native grassland (520) Pine-oak (395) 0,0.13 0.13,2.17 0.08, 0.80 0.04, 0.54 mm does not represent the modal value of one instar with a very broad size range, but is in- stead the result of size overlap between post- emergent instars IV and V: 1) The size range of adult females should be greater than that of any younger instar. 2) The ITL range of the penultimate male cohort (recognized by swol- len palpal tarsi) should approximate that of the penultimate females. Adult females were dis- tinguished by their protuberant genital area (and by fully developed spermathecae when- ever dissections were performed). Penultimate females (instar V) were distinguished on the basis of size and the absence of a protuberant genital area. Size frequency histograms of sea- sonal subsets of T. straminea specimens col- lected at both wetland sites show in late spring (12-27 May) adult and penultimate males, adult and penultimate females, and relatively large juveniles, most of which are presumably antepenultimate (Fig. 5). The summer (17 July-16 August) sample set contained a small- er number of adult females and younger ju- veniles (instars I-III) than were present in the spring. The fall (7 October) sample set (from Meadow Branch wetland) was composed only of a juvenile class (instars III-IV) with a mean ITL between that of the spring and summer samples. These seasonal patterns strongly sup- port a life history pattern of one generation per year with most individuals overwintering in the antepenultimate instar. Males and fe- males appear to mature and mate in May and June. Many adult females persist well into the summer months, but males are absent then, suggesting that they die soon after mating. Behavior of T s^aminea~ln the field, the orientation of T. straminea orbs varied from horizontal to diagonal. Some spiders were in the center of their web adopting a roughly stick-like posture (legs I and II extended for- ward fairly close to one another and legs III and IV extended backward near the sides of the abdomen). Others were stretched out on a twig or grass blade with legs I and II held together, the much shorter legs III surrounding and gripping the substrate, and legs IV ex- tended backward along the sides of the ab- domen. This second posture, in concert with the slender abdomen and pale yellow-brown color, made the spider exceedingly difficult for us to locate. Sometimes we could not find- the captive spiders that had adopted this very cryptic posture without jarring the dead grass stems in their containers. When disturbed in this way, the spider would sometimes drop AIKEN & COTLE^-TETRAGNATHA SPECIES IN THE GREAT SMOKY MOUNTAINS 101 spruce-fir (1830) high grass bald (1755) spruce (1715) beech gap (1645) northern hardwood (1615) red oak (1555) low gri^s bald (1505) heath bald (1390) mixed oak (1115) Table Mtn. pine (1005) hemlock-hardwood cove (945) hemlock (885) hardwood cove (740) wetland (Indian Cr.) (685) wetland (Meadow Br.) (535) native grassland (520) pine-oak (395) 0123401234 relative abundance (mean no, spiders per 1 hr sample) Figure L- — Relative abundance of the three most common Tetragnatha species in 1996 and 1997 at 17 focal sites representing 16 biotic communities in the Great Smoky Mountains National Park. Low grass and heath bald sites were sampled in 1995 only. Focal sites are listed in order from lowest to highest elevation (in meters within parentheses). An asterisk marks any bar representing a relative abundance value significantly higher than one for the same species and site in the other year (ANOVA, P < 0.05). 1996 22ZZZ □ ZZl m Z3 ■ laboriosa H straminea □ versicolor ’ZZZZ21 ZZl 5.8 1997 ZI IZl zzzzzzzzz:^ /77/A '/Z777 ZZZ3* 18.0 sss laboriosa versicoior BEAT SWEEP BEAT SWEEP (shrubs/trees) (herbs) (shrubs/trees) (herbs) Figure 2.~Microhabitat distribution of Tetrag- natha species at the high grass bald site, n ” 12 beat and 18 sweep samples. Standard error is shown on top of each bar. The P-value is generated by ANOVA; see text for test statistics. Straminea versicoior BEAT SWEEP BEAT SWEEP (shrubs/trees) (herbs) (shrubs/trees) (herbs) Figure 3.— Microhabitat distribution of Tetrag- natha species at the Indian Creek wetland, n = 8 beat and 7 sweep samples. Standard error is shown on top of each bar. The P-value is generated by ANOVA; see text for test statistics. THE JOURNAL OF ARACHNOLOGY 102 I TT TTl Tibia i iength (mm) - I Figure 4, — Size (ITL) frequency distribution histograms of all 220 Tetragnatha straminea individuals collected at the two montane wetland sites during 1996 and 1997. Females and individuals too young to be sexed are graphed separately from penultimate and adult males. Labeled horizontal bars indicate ITL ranges of putative and known (adults and penultimate males) posLemergent instars. Spring (12-27 May) wvr nenult. female.*! adult female*! Fall (7 October) in IV 4 5 6 7 Tibia I length (mm) 10 1 1 Figure 5. — Size (ITL) frequency distribution histograms of seasonal subsets of all 220 Tetragnatha |i straminea individuals collected at the two montane wetland sites during 1996 and 1997, Females and. 1 individuals too young to be sexed are graphed separately from penultimate and adult males. Labeled || horizontal bars indicate ITL ranges of putative and known (adults and penultimate males) postemergent ,|j instars. ^ AIKEN & COYl^E^TETRAGNATHA SPECIES IN THE GREAT SMOKY MOUNTAINS 103 and assume this cryptic stick-Mke posture while hanging suspended in mid-air from its dragline. Twice we were able to directly observe these spiders capture Drosophila flies without using a web. In the first observation, the fly was walking on the twig under which the spi- der was positioned cryptically. The fly ap- peared to hit one of the spider’s third legs (which were wrapped around the twig) and it was seized instantly. In the second observa- tion, the spider was ascending the side of its glass cage when a fly walked into it. The spi- der’s first two pairs of legs instantly surround- ed the fly for a brief moment until the spider could grasp it with its chelicerae. Silk was not used to immobilize either of these flies. Fol- lowing these and other less closely observed capture attempts not involving webs, the spi- der crushed and manipulated the prey with its chelicerae and pedipalps. We often observed individuals holding in their chelicerae 1-5 flies which had been captured without a web. Occasionally, such spiders with two or more flies in their mouthparts would capture addi- tional live flies (that we held in contact with the web) by grabbing them with the first legs and immediately wrapping them in silk with the hind legs. These immobilized flies were left attached to the web, and since we could not find them on the following day, we pre- sume they were eaten. DISCUSSION Habitat and microhabitat distribution,— Clearly, T, versicolor is a habitat generalist. Out finding that it is common or abundant over a wide elevation range in a wide variety of forest communities as well as wetland, grass bald, and grassland habitats, is consis- tent with collection records cited by Levi (1981), Although it appears to prefer woody vegetation and can thrive in dryer situations than many of its congeners, it can also be found on herbaceous vegetation in marshy ar- eas. Our observations, which are consistent with those of Comstock (1912), Lowrie (1953), Levi (1981), and Kaston (1981), show that T. laboriosa, like T, versicolor, often lives far from aquatic habitats, but, unlike T. ver- sicolor, rarely occurs in forests and is virtually restricted to non-wetland grassy habitats. In spite of this restriction, it thrives over a wide range of natural and agricultural communities and elevations (Levi 1981) and is the most abundant spider in New York alfalfa fields (Wheeler 1973) and central Illinois soybean fields (LeSar & Unzicker 1978). We suspect that the very few individuals of T, laboriosa collected at forest sites within the GSMNP were immigrants that had ballooned from non- forest habitats and would not have matured and reproduced where we found them; this view is supported by LeSar & Unzicker’s (1978) observations that early instars of T. la- boriosa are good ballooners and colonizers and by the fact that every forest-dwelling in- dividual we collected was an early instar ju- venile. Our data indicate that T, siraminea, T. vir- idis, and T. elongata are all habitat specialists. The restriction of T. straminea to non-forested wetlands in the GSMNP is consistent with collection records cited by Levi (1981), Levi’s (1981) observation that T. viridis is restricted to conifers matches our findings. We suspect that our data underestimates the abundance of T. viridis at the two sites where we found it because 1) it may frequent the large volume of hemlock canopy foliage above our sam- pling zone, 2) its green color and abandon- ment of web-building make it difficult to lo- cate visually, and 3) it may be especially difficult to dislodge (Levi 1981). Our obser- vation that r. elongata is strictly riparian and nearly always builds its webs over open water match those of Lowrie (1953), Levi (1981), Kaston (1981), and Gillespie (1987). Accord- ing to the distribution records in Levi (1981), there are only two other species of Tetrag- natha that we think might eventually be found in the GSMNP, T, guatemalensis O.P.-Cam- bridge 1889 and T. pallescens F.P, Cambridge 1903. If these two are living in the GSMNP, they are not common. The finding that T. versicolor is distributed among more habitats in the GSMNP and else- where than are T. straminea, T. viridis, and T. elongata, and the observation that this species has a higher (67®N) and larger (54') latitudinal and geographic (ca. 20.7 billion km^) range than the other three species (46-57°N; 16-34'; 1.6-7. 8 billion krn^) (Levi 1981), appear to fit a taxonomically widespread biodiversity pat- tern where habitat generalists in many taxa tend to occupy broader latitudinal and geo- graphical ranges than do habitat specialists (Stevens 1989; Wilson 1992). However, T. la- 104 THE JOURNAL OF ARACHNOLOGY boriosa, which appears from our data to be less of a habitat generalist than T. versicolor, has much the same geographic range as ver- sicolor. Apparently, the ability of T. laboriosa to colonize and reproduce in open habitats suits it well to utilizing a wide array of edaph- ic and early successional non-forest habitats which have proliferated because of increased human impact on landscapes and which are simply not well represented in the GSMNP. In other words, its status as a habitat generalist cannot be fully expressed in the GSMNP land- scape. Our results indicate that the coexistence of r. versicolor and T. laboriosa at the high grass bald site involves microhabitat segre- gation in a patchy community; versicolor lives primarily in the shrubs and small trees that are scattered within and surround the open areas of grass and other herbs where laboriosa lives. It is puzzling why no adults of T. ver- sicolor were collected here despite the abun- dance of juveniles (Table 1). Perhaps this pop- ulation is largely or wholly maintained by aerial immigration from .high density forest- dwelling populations at lower elevations; this hypothesis remains to be tested. The beat vs, sweep data from the Indian Creek wetland site suggest that the T. versicolor population there is not as distinctly segregated from the stra- minea population. However, observations dur- ing an autumn sampling effort in the Meadow Branch wetland, as well as T. versicolors ability to prosper away from aquatic habitats, suggest to us that an appropriate sampling de- sign would reveal that the straminea popula- tion is concentrated in grasses and other herbs in the wetter part of these wetlands while the versicolor population is chiefly found on taller and more sturdy vegetation in the dryer areas. The significantly higher relative abundance values in 1997 as compared to 1996 for T. versicolor at several sites and for T. laboriosa at one site may be the result of population increases. However, we suspect that the 1997 sampling team devoted more effort to collect- ing small juveniles (particularly from beating sheets and sweep nets) than did the 1996 team, thus creating a bias which might have caused these relative abundance differences. Life history.— Ours is the first life history analysis of T. straminea. This and other life history analyses of north temperate Tetrag- natha species show that one-year life cycles may be the rale in this genus; Finnish popu- lations of T. extensa (Linnaeus 1758), T. ob- tusa C.L. Koch 1837, and T, montana Simon 1874, and Illinois populations of T. laboriosa all have annual cycles (Toft 1976; LeSar & • Unzicker 1978), Much like T, straminea, these ' species overwinter in mid-to-late juvenile in- j; stars and mature and mate in late spring or ; early summer. However, Juberthie (1954) | showed that in southern France Tetragnatha j. species may have two generations per year. ; LeSar & Unzicker (1978) found that lab- i reared T. laboriosa has eight posternergent in- I stars, rather than the six our field data indicate 1 for T. straminea, but the natural phenologies ; of these two species are very similar. Behavior.— The cryptic, stretched~out I stick-like postures of T. straminea (on its web, I' on vegetation, or hanging in mid-air), like : similar postures adopted by other species of Tetragnatha and unrelated spiders like Dei- nopis MacLeay 1839 (Comstock 1912; Bris- to we 1958; McKeown 1963; Forster & Forster 1973; Levi 1981; Kaston 1981; Gillespie & t Groom 1995; Getty & Coyle 1996), surely I must serve to reduce an individual’s chances ‘ of being detected or recognized as prey by f visual predators. The remarkably flexible prey capture behavior we have observed in T. stra- |; minea— -the ability to catch prey both with and without the use of a web— has also been ob- served by Luezak & Dabrowska-Prot (1966) in a Eurasian species, T. montana. This ver- satile capture program, which may be more [I widespread in the genus than is currently ap- ' predated, may help explain the origin of non- ■ web-building cursorial spiny-legged lineages represented by T. viridis (Levi 1981) and sev- eral Hawaiian species (Gillespie & Croom 1995). i ACKNOWLEDGMENTS ; Robert Edwards, Jeff Stiles, Ricky Wright, : Doug Toti, Jeremy Miller, Melinda Davis, and i; Ian Stocks all helped sample and process ; specimens, Richard Brace, Jonathan Codding- ton, Michael Lewder, Denise McNabb, Trevor ; Bundle, and an anonymous reviewer provided helpful comments on drafts of this paper. This [ research was supported by a Western Carolina University Undergraduate Research Grant to : MA and National Science Foundation (DEB- 9626734) and National Park Service Chal- lenge Cost Share grants to FAC. ■■ AIKEN & COYLE— -TETRAGNATHA SPECIES IN THE GREAT SMOKY MOUNTAINS 105 LITERATURE CITED Bristowe, W.S, 1958. The World of Spiders. Col- lins, London. Coddington, J.A. & H.W. Levi. 1991. Systematics and evolution of spiders (Araneae). Annu. Rev. Ecol. Syst, 22:565-^592. Coddington, J.A., L.H. Young, & F.A. Coyle. 1996. Estimating spider species richness in southern Appalachian cove hardwood forest. J. ArachnoL, 24:ll--28. Colwell, R.K. & J.A. Coddirigton. 1994. Estimat- ing terrestrial biodiversity through extrapolation. Phil. Trans. Roy. Soc. London, B, 345:101-118. Comstock, J.H. 1912. The Spider Book, Cornell Univ. Press, Ithaca, New York. Coyle, EA. 1985. Two-year life cycle and low pal- pal character variance in a Great Smoky Moun- tain population of the lamp-shade spider (Ara- neae, Hypochilidae, Hypochilus). J. ArachnoL, 13:211-218. Duffy, E. 1978. Ecological strategies in spiders in- cluding some characteristics of species in pioneer and mature habitats. Symp. ZooL Soc. London, (1978) No. 42, 109-123. Forster, R.R. & L.M. Forster. 1973. New Zealand Spiders: An Introduction. Collins, Auckland, New Zealand. Getty, R.M. & F.A. Coyle. 1996. Observations on prey capture and anti-predator behaviors of ogre- faced spiders (Deinopis) in southern Costa Rica (Araneae, Deinopidae). J. ArachnoL, 24:93-100. Gillespie, R.G. 1987. The mechanism of habitat selection in the long jammed orb-weaving spider Tetragnatha elongata (Araneae, Araneidae). J. ArachnoL, 15:8^90. Gillespie, R.G, & H.B, Croom. 1995. Comparison of speciation mechanisms in web~bui!ding and non-web-biiilding groups within a lineage of spi- ders. Pp, 121-146. In Hawaiian Biogeography: Evolution on a Hot Spot Archipelago. (W.L. Wagner & V.A. Funk, eds.). Smithsonian lestit. Press, Washington, D.C. Hanggi, A., E. Stockli & W. Nentwig. 1995. Hab- itats of central European spiders. Misc. Faun. Helvetiae, No. 4. Juberthie, C. 1954, Sur les cycles biologiqiies des araignees. Bull. Soc. Hist, Nat. Toulouse, 89: 299-318. Kaston, B.J. 1981. Spiders of Connecticut, (rev. ed,), State Geol. and Nat. Hist. Surv. Connecti- cut, Bulletin No. 70. Kremen, C., R.K. Colwell, TL. Erwin, D.D. Mur- phy, R.F. Noss & M.A. Sanjayan. 1993, Terres- trial arthropod assemblages: their use in conser- vation planning. Conserv. Biol., 7(4):796-808. LeSar, C.D. & J.D. Unzicker, 1978. Life history, habits, and prey preferences of Tetragnatha ia- boriosa (Araneae: Tetragnathidae). Environ. En- tomoL, 7:879-884. Levi, H.W. 1981. The American orb-weaver genera Dolichognatha and Tetragnatha north of Mexico (Araneae: Araneidae, Tetragnathidae). Bull. Mus. Comp. ZooL, 149(5):271-318. Lowrie, D.C. 1953. The orb-weaving spiders of the Chicago area. American Midi. Nat., 49(3):846- 861. Luczak, J. & E. Dabrowska-Prot. 1966. Experi- mental studies on the reduction of the abundance of mosquitoes by spiders. Bull. Acad. Polish Sci, CI.2, 14:315-^320. McKeowe, K.C. 1963. Australian Spiders. Angus and Robertson, Sydney. Moran, M.D., TP. Rooney & L.E. Hurd. 1996. Top-down cascade from a bitrophic predator in an old-field community. Ecology, 77(7):2219™ 2227. Noss, R.F. 1990. Indicators for monitoring biodi- versity: A hierarchical approach. Conserv, BioL, 4(4):355-364. Olive, C.W. 1980. Foraging specializations in orb- weaving spiders. Ecology, 61(5):1 133™! 144. Polis, G.A., R.D. Flolt, B.A. Menge & K.O. Wine- miller. 1996. Time, space, and life history: In- fluences on food webs. Pp. 435™461. In Food Webs: Integration of Patterns and Dynamics. (G.A. Polis & K.O. Winemiller, eds.). Chapman & Hail. Reichert, S.E. & L. Bishop. 1990. Prey control by an assemblage of generalist predators: Spiders in garden test systems. Ecology, 71:1441-1450. Stevens, G.C. 1989. The latitudinal gradient in geographical range: Flow so many species co- exist in the tropics. American Natur., 133(2): 240-256. Strong, D.R., D. Simberloff, L.G. Abele & A.B. Thistle (eds.). 1984, Ecological Communities: Conceptual Issues and the Evidence. Princeton Univ. Press., Princeton, New Jersey. Toft, S. 1976. Life histories of spiders in a Danish beech wood. Nat. Jutlandica, 19:5-40. Wheeler, A.G, 1973, Studies on the arthropod fau- na of alfalfa, V. Spiders (Araneida). Canadian EntomoL, 105:425™432. Wilson, E.O. 1992. The Diversity of Life. Harvard Univ, Press, Cambridge, Massachusetts. Manuscript received 2 October 1998, revised 27 April 1999. APPENDIX Habitat type, locality data, collecting dates, and sampling effort for each of the 17 focal sites (listed in order from highest to lowest elevation). Number of ground, aerial, beat, and sweep samples given in parentheses after total number of one-hour samples. 106 THE JOURNAL OF ARACHNOLOGY Spruce-fir forest: NORTH CAROLINA: Swain County, 0.5 km SW Mt. Collins, N & 5 sides of Appalachian Trail, E2755, N39403, 1815-1845 m elev., 26 June 1996, 14 Septem- ber 1996, 11 June 1997, 23 August 1997. 48 samples (16-16-16-0). High grass bald: NORTH CAROLINA: Swain County, Andrews Bald, E2738, N39354, 1755 m elev., 27 June 1996, 22 Sep- tember 1996, 12 June 1997, 6 September 1997. 48 samples (18-0-12-18). Spruce forest: NORTH CAROLINA: Swain County, just SW junction of Noland Divide Trail and road to pumping station, E2755, N39382, 1715 m elev., 20 June 1996, 7 Sep- tember 1996, 10 June 1997, 23 August 1997. 48 samples (16-16-16-0). Beech gap forest: NORTH CAROLINA: Swain County, in hog exclosure below Ap- palachian Trail at 350 m E Road Prong Trail- head, E2786, N39433, 1645 m elev., 14 June 1996, 15 August 1996, 10 June 1997, 13 Au- gust 1997. 48 samples (16-16-16-0). Northern hardwood forest: NORTH CAR- OLINA: Haywood County, Cataloochee Di- vide just above Hemphill Bald Trail at 200 m E Garrett’s Gap, E3055, N39359, 1615 m elev., 12 and 15 June 1996, 14 August 1996, 6 June 1997, 12 August 1997. 84 samples (29-27-28-0). Red oak forest: NORTH CAROLINA: Swain County, Roundtop Knob, E of Noland Divide Trail about 2 mi SE Clingman’s Dome Road, E2770, N39364, 1555 m elev., 24 June 1996, 31 August 1996, 4 June 1997, 11 Au- gust 1997. 88 samples (30-28-30-0). Low grass bald: NORTH CAROLINA: Swain County, Gregory Bald, E2401, N39343, 1505 m elev., 3-5 June 1995, 29-30 Septem- ber 1995. 72 samples (24-0-24-24). Heath bald: TENNESSEE: Sevier County, Inspiration Point on Alum Cave Trail, E2789, N39461, 1390 m elev., 25-25 May 1995, 23- 24 September 1995. 72 samples (24-24-24- 0). Mixed oak forest: TENNESSEE: Sevier County, E, S, & W slopes of Chinquapin Knob, E2639, N39512, 1083-1144 m elev., 13 June 1996, 13 August 1996, 2 June 1997, 7 August 1997. 85 samples (29-26-30-0). Table Mountain pine forest: TENNESSEE: Sevier County, about 200 m N of route 441 loop NW of Chimneys picnic area, E2738, N39471, 976-1037 m elev., 6 June 1996, 6 August 1996, 27 May 1997, 6 August 1997. 64 samples (23-18-23-0). Hemlock-hardwood cove forest: TENNES- SEE: Sevier County, N & E Grotto Falls Trail- head at Roaring Fork Motor Trail, P. White I veg. plot, E2772, N39512, 945 m elev., 22 May 1996, 30 July and 1 August 1996, 19 May 1997, 4 August 1997. 96 samples (32- 32-32-0). Hemlock forest: NORTH CAROLINA: Haywood County, Cataloochee, 150 m S ' mouth of Palmer Branch at Caldwell Fork, E3107, N39436, 854-915 m elev,, 4 June 1996, 5 August 1996, 18 May 1997, 1 June 1997, 10 and 24 August 1997. 84 samples ' (29-26-29-0). Hardwood cove forest: TENNESSEE: Se- vier County, along Porter’s Creek Trail at 200 paces above bridge over Porter’s Creek, ! E2830, N39508, 740 m elev., 18-19 June ' 1996, 24-25 August 1996, 21-22 May 1997, 31 July 1997. 116 samples (39-37-40-0). Wetland (Indian Creek): NORTH CARO- LINA: Swain County, marsh between Indian Creek Trail and Indian Creek at 2 mi, NE of junction with Deep Creek Trail, E2817, N39296, 685 m elev., 27 May 1996, 16 Au- gust 1996, 12 May 1997, 29 July 1997. 33 samples (14-4-8-7). Wetland (Meadow Branch): TENNESSEE: ' Blount County, marsh along Meadow Branch at 0.5 km ENE of Dosey Gap, E2527, N39470, 535 m elev., 23 May 1996, 1 August 1996, 15 May 1997, 17 July 1997. 33 samples (13-8-0-12). Native grassland: TENNESSEE: Blount County, Cades Cove, S side Abrams Creek about 0.3 mi. upstream from Cades Cove Loop Road bridge, E2426, N39423, 520 m elev., 5 June 1996, 8 August 1996, 15 May , 1997, 17 July 1997. 48 samples (24-0-0-24). ' Pine-oak forest: TENNESSEE: Blount County, 300 m N of junction of Tabcat Creek and Maynard Creek, E2301, N39347, 395 m elev., 28-29 May 1996, 2 August 1996, 14 May 1997, 15 July 1997. 96 samples (32-32- 32-0). 2000. The Journal of Arachnoiogy 28:107“! 14 SPIDER BIODIVERSITY IN CONNECTION WITH THE VEGETATION STRUCTURE AND THE FOLIAGE ORIENTATION OF HEDGES Frederic Ysnel and Alain Canard; Laboratoire de Zoologie et d’Ecophysiologie, UMR CNRS 6553, Universite de Rennes I, Campus de Beaulieu, 35042 Rennes Cedex, France ABSTRACT. The relationship between the structure of spider communities and an index of hedge ecological quality (based on an analysis of vegetation architecture using vegetation diversity and foliage cover) was investigated. The comparison deals with six hedges each of low, medium and high ecological value. The species richness and species composition of dominant spiders was the same for hedges of different quality. Thus it is concluded that these two simple parameters cannot reflect the diversity of the hedge foliage. Indicating species of the differences between ecological quality of two hedges could be required among the groups of species absent from one type of hedge. However, the foliage orientation of the hedges may induce substitution of spider species; thus special attention must be paid to the foliage orientation when comparing the spider communities inhabiting the hedges. Keywords; Foliage cover, foliage orientation, species richness Shrubby and raised hedges constitute one of the major elements supporting faunistic di- versity within rural landscapes. In western France in particular, studies undertaken on woody areas have shown the close associa- tions between the vegetation architecture of the hedges and the diversity or density of the hedge-inhabiting fauna, especially birds, small mammals, reptiles and insects (Saint Girons 1994; Constant & Eybert 1995; Burel 1996). On a regional scale, classifications, based on vegetation, have been established to define the suitability of hedges for certain fauna, such as game birds (Brown & Aubineau 1989). To achieve the goals of hedgerow management (maintenance of biodiversity, wood produc- tion, amelioration of climatic effects and im- provement of water quality), a special index of classification for the ecological value of the hedges was developed by Roze (1995). This index was based on an analysis of the major architectural characteristics of hedges (vege- tation diversity and percentage of foliage cov- er). Numerous workers have detailed the strong relationship between vegetation struc- ture and the composition of spider communi- ties; and it is often argued that this is the most important parameter involved in web site se- lection (Wise 1993). Consequently, it is ex- pected that the diversity of spiders and the species composition of the dominant spiders in hedge foliage can reflect the hedge ecolog- ical value when that value is defined by an index of quality integrating the vegetation ar- chitecture, The aim of this work was to in- vestigate the relationship between the varia- tion in the ecological index proposed by Roze (1995) and the variation in the associated spi- der communities. A comparison of the spider communities inhabiting hedges of different ecological values is presented. METHODS Study area and index of hedge quality The area investigated was situated in an ag- ricultural landscape of Brittany (western France) consisting of fallow-fields (24.5 ha) surrounded by raised hedges for which density reached 1700m/10 ha. The plot was in the dis- trict of Cande-La Brocherie at 1°2'W, 47°34' N. The evaluation method used to assess hedge quality took into account the floristic composition and structure of the hedges (Roze 1995; Table 1). A high biological value was allotted to the hedges when they were estab- lished on a complex of ditches or slopes, when the foliage cover of the shrubby and arbores- cent layers was high, and when brambles and nettles were wanting. Additional points were allotted when species, which were not very 107 108 THE JOURNAL OF ARACHNOLOGY 1 Table L — Card-data for the evaluation of the bi- ological quality of one hedge. The number of points is indicated in parentheses. 1) Slope/Ditch complex ditch (1) slope (1) double hedge (1) ditch elevation (>1 m) (2) 2) Trees percentage of re-covering <20% (0) 20 < % <50 (1) >50% (2) spontaneous species (oak. . .) (1) not frequent species (alder, hornbeam. . .) (1) seedlings (1) 3) Shrubs percentage of re-covering <20% (0) 20 < % < 50 (1) >50% (2) specific diversity 2-3 sp. (1) >4 sp. (2) original vegetation (spindle tree. . .) (1) 4) Edge vegetation Endymion non scripus & Anemone nemorsa (3) Umbilicus rupestris & Polypodium vulgare (2) Ruscus aculeatus & Rubis perenigra (2) Teucrium scorodonia & Stellaria holos- tea (1) Juncus effusus + hydrophilous vegetation (1) Rubus fruticosus & Dactylis glomerata (0) Pteridium aquilinum (0) Rubus fruticosus & Dactylis glomerata (0) Urtica dioica (“”1) frequently distributed at a regional scale, were present. The range of hedge ecological quality values varied from 1™20 which provided a comparative index for the biological quality of each hedge. Collection of spiders and data analysis.”— Our previous investigations into the spider communities inhabiting shrub lay- ers in western France have demonstrated that there was no considerable variation between the species composition of the “spring com- munity” and the “annual community” for successive years (Canard 1979; Canard 1984; Ysnel et. al. 1996). These results concerning the temporal stability of the spider commu- 1.44 1.28 1.12 0.96 0.80 Figure L— Dendogram of hedge similarity (UGPMA clustering method) concerning foliage spider communities. nities justified our comparison here of differ- ent hedge communities during spring. Six hedges (A-F) were selected according to their index of biological quality to provide two hedges in each of three categories: hedge A (index of 20) and B (19) to “high”; hedge C (11) and D (9.5) to “medium”; and hedge E and F (6.5) to “low.” All the selected hedg- es were situated in a complex of three contig- uous fallow-fields. The foliage spiders were collected by six series of branch-beating dur- ing spring 1997 (March, April, May) using a beating tray of 0.7 and a walking stick. Two people collected the spiders, one person was beating while another one was collecting the spiders from the tray with the aid of a pooler. Spiders were sampled over a total of 75 m for each hedge, which was comprised of 15 samples, with each sample of 5 linear me- ters at each of three heights: low (ground lev- el), medium (at 1 m) and high (at 2 m). For each 5 meter sample at a given height, five tablecloths were placed on the ground and three whacks per tray were given to help achieve an “equal beating effort” across all samples. Since the beating method collects spiders only during their diurnal activity pe- riod (McCaffrey et. al. 1984), the timing of beating was randomly distributed across all i; hedges and heights sampled. This method ob- ' viously undersampled nocturnal spiders. To estimate the potential influence of foliage ori- entation, sampling was carried out along 20 m YSNEL & CANARD^SPIDER BIODIVERSITY AND HEDGEROWS 109 Table 2. — Foliage spider communities and ecological value of hedges (ind/m: mean number of individ- uals per linear meter; sp/m: mean number of species per linear meter; J': Shannon evenness index). Hedge quality Specific diversity sp/m ±SD ind/m ±SD J' Foliage orien- tation High hedge A 47 2 ± 0.56 14.3 ± 2.7 0.68 NE hedge B 35 1.5 ± 0.35 6.2 ± 2.1 0.87 W Total specific diversity: 53 Medium hedge C 34 2.1 ± 0.34 11.3 ± 3.4 0.84 E hedge D 41 2.1 ± 0.30 9.7 ± 1.9 0.81 SE Total specific diversity: 46 Low hedge E 44 1.8 ± 0.73 6.6 ± 2.7 0.81 N hedge F 30 1.3 ± 0.22 5.2 ± 2.4 0.79 W Total specific diversity: 51 of each side of hedge D during September 1997. The nomenclature of spiders used follows Platnick (1997). The juveniles of the follow- ing genera were considered to belong to only one species: Agroeca sp, Episinus sp., Zora sp., Pirata sp., Evarcha sp., Micaria sp., Cheiracanthium sp., Zelotes sp., Zygiella sp., Tibellus sp., Xysticus sp. Clubiona sp , Heiiophanus sp, Leptyphan- tes sp. Ozyptila sp., Robertus sp. and Tetrag- natha sp. were counted as species where there were only juveniles in what was collected. In order to simplify the comparison of the spider communities, the “dominant species” refers to species represented by at least 4.5% of the total individuals collected in one hedge. A cluster analysis was performed by NTSYS-PC program with the use of the UGPMA method. The similarity matrices for community analyses were derived using the chi-squared distance by means of the follow- ing formula: dij = J'Z (Xki/Xi - Xkj)Vxk An ANOVA (multiway factor analysis) was also performed by STATGRAPHIC-PC pro- gram to test the differences between the hedg- es. This analysis considered variation from two factors: hedge quality and number of in- dividual per species. The variables examined were first transformed in percentages for the anova analysis and the data from couple of hedges of low, medium or high value were pooled. RESULTS Species richness and index of density.™ A total of 72 genera and species was identified from the foliage of the six hedges studied (Appendix 1). The average species richness of the three categories of hedges remained vir- tually identical (Table 2), and a hedge of low ecological value could harbor a species rich- ness greater than that in a hedge of high value. Table 3.— -Percentage of shared species between hedges (T = total number of species; s.s. = shared species). A-B A-C A-D A-E A-F B-C B-D B-E B-F T 53 51 59 58 51 41 52 52 30 s.s. 24 27 29 25 21 22 26 22 20 % 45.2 53 49.1 43.1 41.2 5J.7 50 42.3 66.7 C-D C-E C-F D-E D-F E-F T 46 43 40 58 50 48 s.s. 25 27 23 28 24 23 % 54.3 62.7 57.5 48.3 48 47.9 no THE JOURNAL OF ARACHNOLOGY Table 4. — Comparison between the spider communities inhabiting the two faces of the same hedge (relative abundance of species is given in parentheses; * = 1 individual). Foliage orientation W-NW E-SE Number of species 38 32 Total individuals 257 210 Dominant species Mangora acalypha (13.2%) Zilla diodia (13.2%) Anyphaena accentuata (6.5%) Not common Clubiona brevipes (2.7%) Theridion mystaceum (1.5%) Theridion pallens (1.5%) Atea triguttata (1.5%) Hyptiotes paradoxus (1.2%) Meta segmentata (0.7%) Leptyphantes tenuis (0.7%) Araneus umbraticus (*) Microlinyphia pusilla (*) Synaema globosum (*) Philodromus cespitum (24.8%) Nigma puella (6.6%) Heliophanus sp. (8%) Theridion tinctum (1.4%) Anelosimus sp. (1.9%) Bathyphantes gracilis (*) There is no significant differences between the average number of species collected by linear meter in hedges B, C, D and E. The average density of individuals collected fell consider- ably for the two hedges of low value and for one of the hedge of high ecological value (B). The difference in the mean number of spiders collected between the two hedges A and B (high value) was strongly related to the pres- ence of numerous immatures of four species or genera {Zygiella sp., Nigma puella (Simon 1870), Araneus diadematus Clerck 1758, Dic- tyna uncinata Thorell 1856) in hedge A. This was confirmed by the low value of the Shan- non evenness index for that spider assem- blage. Influence of foliage orientation.^ — The ori- entation of the foliage may influence the struc- ture of the spider communities since the per- centage of shared species is higher between two hedges of the same foliage orientation (hedges B and F, west orientation) than be- Table 5.— ANOVA analysis of three community categories (low, medium, and high). Source df MS F-ratio P-value Main effects A; species 72 53.2 12.95 0.000 B: hedge type 2 <0.01 0.00 1.000 Interactions AB 146 2.44 0.6 0.1 Residual 222 4.11 tween two of the same ecological value (Table 3). This hypothesis is supported by the com- parison of the spider communities sampled on the two faces of the same hedge (Table 4). We observed a substitution among the three dom- inant species and 12 of the species collected on this hedge were not common to both faces of the hedge investigated. Moreover, the UGPMA analysis separated the six spider communities into three clusters which were not congruent with the respective ecological value of the hedges (Fig. 1). This can also be related to the foliage orientation since the cluster analysis separated group of hedges (A,D, or C) sampled on their eastern face.Thus, variation in the relative abundance of individuals observed among the six com- munities could not be correlated with the eco- logical value of the hedges. Specific composition of spider communi- ties.—-The ANOVA shows that there is a sig- nicant difference between the relative abun- dance of each species in the three types of hedges (source A: P-value < 0.05), but the relative abundance of a same species collected in the three type of hedges is not significantly different (source B; P-value > 0.05). Further- more, there is no significant interaction amongst the two factors which strongly sug- gests the lack of relationship between hedge type and the structure of the spider community associated (Table 5). This has to be connected with the fact that 90% of the individuals col- YSNEL & CANARD-SPIDER BIODIVERSITY AND HEDGEROWS 1 Table 6. — ^Dominant species in each hedge with relative abundance (in percentage). A B C D E F Zygiella sp. 23 4.5 18 17.5 9.7 13 Nigma puella 16.6 9.3 14 15.6 22.5 Philodromus sp. 5.7 5.7 9 9.9 5.9 Zilla diodia 5 12 6.3 9.6 7.5 5 Dictyna uncinata 6 6.5 6 8.5 Araniella opisthographa 9.8 11 9 Anyphaena accentuata 4.5 Araneus diadematus 9.4 Heliophanus sp. 15 Paidiscura pallens 11 lected belonged to shared species (Appendix 1). Among the 10 dominant species collected in each hedge (Table 6), 5 are the dominant species in all hedges. The dominance of A. diadematus and of Heliophanus sp. has to be related to the numerous immatures collected in only one of the hedges of high ecological value. The same remark can be made con- cerning the dominance of A. accentuata (hedge C) and P. p aliens (hedge E). There- fore, if we consider the representation of adult spiders, there were no dominant species which were characteristic of hedges of low, medium, or high ecological value. In addition, the anal- ysis of species distribution according to func- tional groups did not reveal a significant dif- ference in the representativeness of the various groups (Table 7). Very few species (Table 8) were collected on only one of the six hedges, and each was represented by only 1 or 2 individuals. Some species were absent from hedges of high value (e.g., Lathy s hu- mills Blackwall 1855, Araneus triguttatus (Fa- bricius 1775) or, on the contrary, species were always absent from hedges of low value (e.g., Table 1. — ^Number of species according to hunt- ing habits for the different group of hedges. High value Medi- an value Low value Total Orb-web spiders 13 14 10 14 Frame-web spiders 11 11 13 16 Sheet-web spiders 9 6 9 16 Ambush hunters 11 8 9 12 Diurnal wanderers 12 10 10 16 Nocturnal wanderers 6 5 7 8 Gibbaranea gibbosa (Walckenaer 1802), Sai- tis barbipes Simon 1868). DISCUSSION Very few comparative studies have been made on the spider communities of the hedge- row networks, and they are mainly based on the analysis of ground living spiders (Petto 1990; Bergthaler 1996). This first approach to investigating foliage spider communities shows that there were no direct relationships between spider biodiversity and an index that described hedge habitat quality based on the analysis of the vegetation architecture. There- fore, concerning the spiders inhabiting the fo- liage, easy field indicator parameters of hedge quality, as for instance spider species com- position or relative abundance of species, are not useful. By artificially modifying the density of the foliage of a big sage {Artemisia tridentata), Hatley & MacMahon (1980) demonstrated that spider species diversity and the number of guilds were positively correlated with in- dicators of shrub volume and foliage diversity. These variations were observed on spider communities which were colonizing a shrubby layer composed by only one vegetal species. W& also found that the hedge type may influ- ence the composition of spider assemblage in the foliage. But, in the present case, because the architecture of the foliage is too diverse, whatever the ecological value of the hedge is, the spider specific richness remains almost the same for hedges of high or low ecological val- ue. Moreover, it can be argued that foliage ori- entation, which was not incorporated into the index of vegetation quality, induced substitu- tion of spider species, further limiting again 112 THE JOURNAL OF ARACHNOLOGY Table 8. — Single species in three categories of hedges (* Genus present in the two other types of hedges). Hedge quality High Medium Low Diurnal Salticus scenicus Pi rata sp. Bianor aurocinctus Pardosa hortensis Alopecosa accentuata Nocturnal Micaria sp. Agroeca sp. Clubiona terrestris* Frame web Robertus arundineti* Robertus lividus* Philodromus dispar* Theridion impressum* Theridion tinctum* Episinus sp. Ambush-hunters Ozyptila praticola* Tibellus sp. Sheet-weavers Agyneta affinis Agyneta subtilis Pelecopsis parallela Walckenaeria acuminata Agyneta rurestris Microlinyphia pusilla Ceratinella brevipes Collinsia submissa Leptyphantes ericaeus Oedothorax fuscus Panamonops sulcifrons the ability of the index to reflect changes in spider diversity. This study also demonstrates that one hedge has to be carefully sampled on its two faces in order to identify the whole spider species inhabiting the foliage. As density and specific diversity of spiders do not correspond to the general vegetal qual- ity of hedges, are there any indicating species which show the habitat quality? The dominant species did not vary among hedges of differ- ent quality, which supports our former obser- vations on the relatively stable composition of dominant species colonizing the shrubby lay- ers within the same macroclimatic sector (Ysnel et al. 1996). However, the indicator species for the ecological quality of the hedg- es could be identified, not among the domi- nant species, but on the contrary, by consid- ering the single species collected in one hedge. However, these species were poorly represented in the samplings and their absence from hedges of other quality could be sam- pling artifact or could be related to the foliage orientation of the hedge investigated. Some species are missing from the category of hedges with a high or low ecological value. These species, then, are likely to be more in- dependent of the orientation of the hedges and their absence could be connected to the struc- ture of the vegetation. Further investigations in other hedges of different ecological value are required to clarify these indicators. Con- cerning the maintenance of spider biodiversi- ty, we must notice that the presence of hedge , groups of different index on one area will lead to bigger specific diversity than the presence of only one edge group of high index. S ACKNOWLEDGMENTS We are grateful to M.C. Eybert and T. Ges- lin for providing the biological value of the I hedges investigated. This work was supported by the Conseil Cynegetique Regional des Pays de Loire. i j LITERATURE CITED ; Bergthaller, G.J. 1995. Preliminary results on the colonization of a newly planted hedgerows by j epigeic spiders (Araneae) under the influence of adajacent cereal fields. Proceedings of the Xlllth Intern. Congr. Arachnology (Geneva). Rev. Suis- se Zook, Vol. h.s. II:61™70. Brun, J.C. & J. Aubineau, 1989. La classification cynegetique des haies: une methode adaptee aux operations d’amenagement rural. Notes tech- niques. Bulletin mensuel ONC.N^ISS, fiche N°54. Burel, F. 1996. Hedgerows and their role in agri- cultural landscapes. Crit. Rev. in Plant Sci., 15(2):169-190. Canard, A. 1979. Donnees ecologiques sur qu- elques araneides d’une lande seche armoricaine. Rev. ArachnoL, 2(6):303-312. Canard, A. 1984. Contribution h la connaissance du developpement, de I’dcologie et de Fecophysiologie des araneides de landes armor- icaines. These de Doctorat es-Sciences, Univer- site de Rennes L YSNEL & CANARD---SPIDER BIODIVERSITY AND HEDGEROWS 13 Constant, P. & M.C, Eybert. 1995. Cavifaune et la haie. Penn ar bed, 153/154:85--93. Hatley, C.L. & J.A. MacMahon. 1980. Spider com- munity organization: seasonal variation and the role of vegetation architecture. Environ. Ento~ mol., 9:632-639. Petto, R, 1990. Abundance and prey of Coelotes terrestris (Wider) (Araneae, Agelenidae) in hedges. Bull. British Arachnol. Soc., 8(6): 185” 193. Piateick, N.L 1997, Advances in Spider Taxonomy (1992-1995) with Redescriptions 1940-1980, New York EetomoL Soc., 976 pp. McCaffrey, J.P., M.P. Parrella & R.L, Horsburgh. 1984. Evaluation of the limb-beating method for estimating spider (Araneae) populations on apple trees. J, ArachnoL, 11:363-368. Roze, E 1995. Methode d’ evaluation de ITnteret biologique et 6cologique des haies et talus en Bretagne. Botanica Rhedonica, Nouvelle sdrie, 3: 46”54. Saint-Girons, H. 1994. Ecologie et repartition des reptiles: role des haies et talus plantes. Penn ar bed, 153/154:78”84. Wise, D.H. 1993. Spiders in Ecological Webs, Cambridge Univ. Press, 328 pp. Ysnel, E, A. Canard & G. Tiberghien. 1995. The shrub layer spider communities: variation of composition and structure of the gorse clump communities in western France. Proc. Xlllth In- tern. Congr. Arachnol. (Geneva). Rev. Suisse ZooL, VoL h.s. II, 691-700. Manuscript received 3 October 1998, revised 5 July 1999. Appendix 1. —Total list of species with number of individuals collected in all the hedges. High Medium Low Biological value of hedges Diurnal wanderers Alopecosa accentuata (Latreille 1817) Anyphaena accentuata (Walckenaer 1802) Ballus biimpressus (Doleschall 1852) Bianor aurocinctus (Ohlert 1865) Ero aphana (Walckenaer 1802) Evarcha sp, Heliophanus cupreus (Walckenaer 1802) Heliophanus sp. Macaroeris nidicolem (Walckenaer 1802) Pardosa hortensis (Thorell 1872) Pardos a sp. Pirata sp, Pisaura mirabilis (Clerck 1758) Saitis barbipes Simon 1868 Saiticus scenicus (Clerck 1758) Zora sp. Nocturnal wanderers Agroeca sp, Cheiracanthium sp. Clubiona brevipes (Biackwall 1841) Clubiona compta Koch C.L. 1839 Clubiona terrestris Westring 1851 Clubiona sp. Micaria sp. Zelotes sp. Frame~web spiders Anelosimus vittatus (Koch C.L. 1836) Dictyna uncinata Thorell 1856 Episinus sp. Lathys humilis Biackwall 1855 Nigma puelia (Simon 1870) Paidiscura pallem (Biackwall 1834) 1 33 11 35 20 5 5 14 13 4 7 7 1 2 6 8 5 1 1 1 2 13 1 19 61 2 15 4 13 3 1 13 8 3 1 1 1 1 2 1 1 1 2 1 3 1 1 1 1 3 1 3 1 6 5 5 1 13 1 28 19 39 21 26 7 1 1 5 1 2 14 3 17 7 20 4 63 12 51 38 6 30 1 25 1 6 172 39 108 98 5 80 9 5 12 21 52 12 114 THE JOURNAL OF ARACHNOLOGY Appendix 1. — Continued. High Medium Low Biological value of hedges A B C D E F Robertas arundineti (Cambridge O.R 1871) 2 Robertas lividas (Blackwall 1836) 1 Robertas sp. 3 1 Theridion impressam Koch C.L. 1881 1 Theridion mystaceam Koch L. 1870 25 1 11 6 9 14 Theridion tinctam (Walckenaer 1802) 2 Theridion varians Hahn 1831 7 2 21 1 7 3 Theridion sp. 50 20 102 16 51 26 Orb-weavers Araneas diadematas Clerck 1758 98 18 2 12 2 1 Araneas sturmi (Hahn 1831) 1 2 1 1 5 Araneas triguttatas (Fabricius 1775) 1 4 3 Araniella opisthographa (Kulckzynski 1905) 28 41 30 22 56 32 Argiope bruennichi (Scopoli 1772) 3 2 1 Cyclosa conica (Pallas 1772) 1 1 Gibbaranea bitabercalata (Walckenaer 1802) 3 7 24 5 1 Gibbaranea gibbosa (Walckenaer 1802) 1 1 2 4 Larinioides cornutus (Clerck 1758) 2 5 1 M angora acalypha (Walckenaer 1802) 18 12 17 20 11 4 Tetragnatha montana Simon 1874 2 1 Tetragnatha sp. 17 11 16 11 21 7 Zilla diodia (Walckenaer 1802) 52 50 49 60 37 17 Zygiella sp. 240 19 140 110 48 46 Sheet-weavers Agyneta affinis (Kulckzynski 1898) 1 Agyneta rarestris (Koch C.L. 1836) 2 Agyneta subtilis (Cambridge O.P. 1863) 3 Bathyphantes gracilis (Blackwall 1841) 1 1 Ceratinella brevipes (Westring 1851) 2 Collinsia submissa (Koch L. 1879) 1 Hypomma cornutum (Blackwall 1833) 2 5 Lepthyphantes ericaeas (Blackwall 1853) 1 Lepthyphantes tenais (Blackwall 1852) 8 1 2 1 Lepthyphantes sp. 14 1 8 4 3 2 Microlyniphia pasilla (Sundevall 1830) 1 Oedothorax fascas (Blackwall 1834) 1 1 Panamonops sulcifrons (Wider 1834) 1 Pelecopsis parallela (Wider 1834) 1 1 Porrhomma oblitam (Cambridge O.P. 1870) 1 1 1 1 Walckenaeria acaminata (Blackwall 1833) 1 Ambush-hunters Diaea dorsata (Fabricius 1777) 1 1 Misumenops tricaspidatus (Fabricius 1775) 8 1 5 2 2 4 Ozyptila praticola (Koch C.L. 1837) 3 Ozyptila sp. 1 1 12 4 1 Philodromus cespitam (Walckenaer 1802) 3 5 2 2 5 Philodromas dispar (Walckenaer 1802) 1 Philodromus rufus (Walckenaer 1802) 1 1 1 2 6 Philodromus sp. 60 24 29 57 49 21 Synaema globosum (Fabricius 1775) 1 4 1 Tibellus sp. 1 Tmarus stellio Simon 1875 2 1 Xysticus sp. 6 3 1 2 I 2000. The Journal of Arachnology 28:115-122 EFFECT OF RIVER FLOW MANIPULATION ON WOLF SPIDER ASSEMBLAGES AT THREE DESERT RIPARIAN SITES Erik J. Wenninger*: Department of Biology, University of Toledo, Toledo, Ohio 43606 USA William F. Fagan: Department of Biology, Arizona State University, Tempe, Arizona 85287^1501 USA ABSTRACT. The distribution, abundance, and diversity of wolf spider (Lycosidae) assemblages were investigated via pitfall trapping at three sites near Granite Reef Dam outside Phoenix, Arizona. These three sites featured different moisture and temperature regimes due to the dam, which diverts the Salt River into an urban canal system. Site 1 was a natural riparian area above the dam along the Salt River, Site 2 was adjacent to a man-made diversion canal, and Site 3 was adjacent to the dry riverbed below the dam. Four lycosid species were found at Site 1, with Pardosa vadosa Barnes 1959 dominating. Two species each, though very few total individuals, were found at Sites 2 and 3. Simpson’s index of diversity (of lycosids and of all other terrestrial arthropods) was higher for Site 1 than for Sites 2-3. Prey availability was comparable among sites, but Site 1 had significantly higher relative soil moisture levels and less extreme substrate and air temperature conditions than did Sites 2 and 3. Spider abundance at each site was independent of prey availability, but instead depended chiefly upon moisture and temperature regimes among sites. The results suggest that wolf spiders experienced a significant effect from disturbance of their habitat by the dam, and that abiotic habitat attributes such as moisture and temperature may be more important for wolf spider abundance than prey availability alone in desert riparian systems. Keywords: Pardosa, Salt River, Arizona In comparison to habitats featuring less hu- man impact, urbanization can have significant effects on the environmental conditions, pop- ulations, and community structures of ecolog- ical systems (McDonnell et al. 1997). While vertebrate populations often may decline due to the anthropogenic pressures and habitat loss associated with urbanization (for example: Hoi Leitner 1989; Gill & Williams 1996), many invertebrate species exhibit an ability to establish alternative ecological relationships allowing them to persist or even flourish in urban environments (Frankie & Ehler 1978; Dreistadt et al. 1990). As a result, arthropod populations and assemblages may be similar among natural and disturbed sites (Frankie & Ehler 1978). As one might expect, however, urbanization has also been shown to have ad- verse effects on some invertebrate populations (Nowakowski 1986; Sawoniewicz 1986; Ruszczyk & Mellender 1992). Frankie & Eh- ’ Current address: Dept, of Biological Sciences, Idaho State University, Pocatello, Idaho 83209- 8007 USA ler (1978) point out that perhaps one of the few generalizations which can be made about terrestrial invertebrate populations in urban environments is that the distribution and di- versity of such species often reflect different moisture regimes. As part of the newly-funded Urban Long Term Ecological Research site in central Ari- zona, we set out to compare the distribution and diversity of assemblages of wolf spiders (Lycosidae) in three Sonoran Desert riparian areas featuring different environmental re- gimes as a function of river flow manipula- tion. We sought to investigate the relation- ships between wolf spider distribution and abundance patterns to prey availability, tem- perature regimes (air temperature, substrate temperature, and variation between the two), and relative soil moisture. Most wolf spiders do not build webs, but rather are vagrant hunters, and spend most of their time near the ground surface. They may wander or remain stationary while hunting un- til a prey item is detected by visual or vibra- tory cues, at which point they attack (Kaston 115 T 116 1978; Kronk & Riechart 1979; Cady 1984; Persons & Uetz 1996). Different species of Pardosa, the dominant genus found in this study, have been variously described as either sit-and-wait or cursorial hunters (Morse 1997). A large body of research has demon- strated that wolf spiders exhibit habitat selec- tion and distribution and abundance patterns based on a variety of factors, including: prey availability, capture efficiency, mating proba- bility (in males), herbaceous vegetation cover, temperature, humidity, and soil moisture con- tent (e.g., Cherrett 1964; Hallander 1967, 1970; Lowrie 1973; Kronk & Riechert 1979; Bultman 1992; Cady 1984; Moring & Stewart 1994). Microenvironmental factors such as vegetation cover, temperature, humidity, and prey availability can be directly related to sub- strate moisture levels. Based on these studies, we expected that mid-summer censuses (when the abiotic con- ditions of the desert were at their most ex- treme) would result in wolf spider assemblag- es that varied as a function of habitat. In particular, we expected wolf spider abundance and species diversity to depend sensitively on soil moisture as in Kronk & Reichert's (1979) study of Rabidosa santrita (Chamberlin & Ivie 1935) and as in Agnew & Smith’s (1989) study of spiders in irrigated and drought- stressed peanut fields. Experiments demon- strating the inability of Pirata piraticus (Clerck 1757) to tolerate desiccation (Cherrett 1964), as well as the association of many western Pardosa species with moist habitats (Lowrie 1973) further supported our expecta- tions. STUDY SITES AND METHODS The study sites were two desert riparian ar- eas adjacent to the Salt River and one area along a canal, running through Tonto National Forest near Granite Reef Dam, 22 km east of downtown Phoenix, Arizona. Completed in 1908, Granite Reef Dam is the point where the Salt River is diverted into man-made ca- nals for eventual human use (Higgs 1995). The presence of this water resource is one of the key factors that has facilitated explosive growth of the Phoenix metropolitan area in the last several decades. Because of the river di- version, the riverbed below the dam is nearly completely dry for much of the year. Prior to the completion of Granite Reef Dam, down- stream reaches were well-watered and fea- THE JOURNAL OF ARACHNOLOGY tured desert riparian vegetation typical of up- stream areas today (see below). However, Granite Reef Dam is only the most recent modification to the river and surrounding ri- parian corridor: this portion of the river was also the site of large-scale water diversions into irrigation canals by the Hohokam culture (AD 700-1450) (Gregory 1991). Site 1 was a strip of riparian area approxi- mately 7 km upstream of the dam in a semi- natural area designated for recreational use. Immature willow {Salix gooddingii and Saiix exigua), cottonwood {Populus fremontii), and tamarisk (Tamarix spp.; invasive exotics) trees as well as understory riparian vegetation grew along the river bank; the substrate was primarily rock cobble and sand. Site 2 was about 2 km downstream of the dam adjacent to one of the diversion canals. Because the canals were constructed of concrete, which al- . lows for little lateral movement of water out- ward from the sides of the canal, plant cover at this site, even that immediately adjacent to the canal, was typical upper Sonoran Desert vegetation featuring saguaro cactus (Carnegia gigantea) and palo verde (Cercidium micro- phyllum). The substrate was primarily densely packed sand. Site 3 was an area about 0.5 km downstream of the dam running along the dry riverbed where the river formerly flowed. It featured a mixture of upper Sonoran vegeta- tion and riparian species able to persist on the water and disturbance regime provided by low-volume, irregular releases of water from the dam; the substrate was primarily a mixture of sand and cobbles. At two locations in each of the three sites, we placed a set of ten pitfall traps (spaced 2 m apart in two rows of five) with the first row located about 2 m away from the adjacent wa- ter source (or edge of dry riverbed) and run- ning parallel to it. Traps (plastic drinking cups [“Dixie®”] 9 cm in diameter) were buried in the ground with the rim set flush with the sur- face. A second cup, with the top 3 cm cut off, was placed in each buried cup for periodic removal of specimens. Forest service regula- tions, concerns over public access (especially pets), and the intense heat and evaporative po- tential of the Sonoran Desert region during the summer, mandated that we use “dry” pitfall n traps (e.g., Hurd & Fagan 1992) rather than traps containing chemical preservatives. To provide a vertical dimension to the trap (and WENNINGER & FAGAN— -RIVER FLOW MANIPULATION AND WOLF SPIDERS 117 thus refugia for captured animals), we placed a loosely crumpled piece of toweling paper in the bottom of each trap. Trapped spiders and other arthropods were collected every 3-6 days. Spiders were sorted to species (using Kaston 1978 and Roth 1993), and broken down into age (sub-adult, adult) and sex cat- egories. Species identifications based on rep- resentative specimens were established by Dr. David Richman (New Mexico State Univer- sity). Voucher specimens have been deposited in the Central Arizona Phoenix LTER’s ar- thropod collection, which is associated with other natural history collections at Arizona State University (ASU). Other arthropods were sorted to family or order, as possible. We also used dry cup pitfall trapping to provide an estimate of available prey, which included counting all soft-bodied arthropods that did not exceed the average length of the largest wolf spider species found (as in Moring & Stewart 1994). This means that we counted only small, immature, and soft-bodied indi- viduals of Formicidae, Dermaptera, and Co- leoptera. Our estimate of available prey thus may be an underestimate for large bodied wolf spiders that have sometimes been observed feeding on hard-bodied insects (e.g., Coleop- tera, Orthoptera [Nyffeler & Benz 1988]). Al- though the inability of lycosids to climb up the smooth surfaces of pitfall traps does not preclude the use of dry pitfall trap data for wolf spiders, many potential insect prey may walk or fly out of such traps or be preyed upon by lycosids while in the traps, which means that our arthropod data are likely un- derestimates. Traps were in place from 11 June 1998-13 July 1998, although a rising river level behind the dam (due to early arrival of the monsoon season in the Sonoran Desert) washed out all traps at Site 1, forcing the early termination of arthropod collection on 30 June 1998 at that site. After finding (1) no discemable dif- ferences between trapped arthropods at Sites 2 and 3 between the periods 1 1 June-30 June 1998 and 30 June-13 July 1998 and (2) no temporal trends in abundance at Site 1, we corrected for the different numbers of trap days by multiplying all counts at Site 1 by 32/ 17. Our results are comparable if we restrict our analyses to data taken from all three sites between 11 June-30 June 1998. Temperature readings were taken at selected locations near each set of traps at each site over four non- consecutive, sunny days, with four readings being taken at each plot every hour between the hours of 0700-1100 h. Both ground tem- peratures and air temperatures (with the ther- mometer held 2 cm above the ground) were taken. Soil moisture readings were taken with a soil moisture probe (measuring relative per- cent soil moisture) on one day with five mea- surements being taken at Sites 2 and 3. Six readings were taken at Site 1 (three at each sub-site) as more variable soil moisture levels were found. RESULTS Pardosa vadosa Bames 1959 was by far the most common lycosid in the vicinity of Gran- ite Reef Dam, comprising well over 90% of the individual lycosids captured (Table 1). Pardosa vadosa (5-6 mm as adults) was also the only lycosid found at all three sites. For this species, 54% of the mature, identifiable individuals were female, indicating a relative- ly balanced sex ratio during the sampling pe- riod. In addition, P. vadosa was the only spe- cies for which a large number of sub-adults was collected. This is potentially important because it could indicate that other lycosids may reproduce at different times of the year than P. vadosa, which could lead to markedly different abundance patterns through time. Arctosa littoralis (Hentz 1844) (adult size 12- 15 mm), which was found only at Site 1, was the next most common lycosid as determined by pitfall trap collections. Sosippus californi- cus Simon 1898 (adult size 12-16 mm) was also found only at Site 1, but in low numbers. Allocosa subparva Dondale & Redner 1983 (adult size 4-5 mm) was found at both Sites 1 and 3, but in low numbers at the latter site, while Pardosa sp. #2 was found only at Site 2, again in low numbers. After lycosids, the Gnaphosidae was the next most common fam- ily of spiders caught in the pitfall traps. Pitfall trapping indicated wolf spiders were more abundant at Site 1 than at Sites 2-3 (Ta- ble 1). This pattern held for male, female, sub- adult, and unidentifiable individuals (sex un- identifiable due to severe desiccation and/or cannibalism in traps). Roughly 16% of col- lected wolf spiders appeared to have been at- tacked by other spiders while inside the dry pitfall traps. Other arthropods commonly represented at 118 THE JOURNAL OF ARACHNOLOGY Table 1.^ — ^Total counts of each arthropod group at each site, with lycosids separated into species. All Site 1 traps were destroyed on day 18. Site 1 specimen counts are corrected for differential trap-days by multiplying by 32/17. Site 1-A* Site LB* Site 1 (pooled*) Site 2-A Site 2-B Site 2 (pooled) Site 3-A Site 3-B Site 3 (pooled) Lycosidae Pardosa vadosa (total) 602 652 1254 3 0 3 0 2 2 Female 171 168 339 2 0 2 0 1 1 Male 139 149 288 0 0 0 0 0 0 Sub-adult 136 288 424 1 0 1 0 1 1 Sex unidentifiable 156 47 203 0 0 0 0 0 0 Pardosa sp. 2 (total) 0 0 0 3 0 3 0 0 0 Female 0 0 0 0 0 0 0 0 0 Male 0 0 0 0 0 0 0 0 0 Sub-adult 0 0 0 3 0 3 0 0 0 Arctosa Uttoralis (total) 32 10 42 0 0 0 0 0 0 Female 11 6 17 0 0 0 0 0 0 Male 19 4 23 0 0 0 0 0 0 Sub-adult 0 0 0 0 0 0 0 0 0 Sex unidentifiable 2 0 2 0 0 0 0 0 0 Allocosa subparva (total) 14 0 14 0 0 0 0 1 1 Female 6 0 6 0 0 0 0 1 1 Male 6 0 6 0 0 0 0 0 0 Sub-adult 0 0 0 0 0 0 0 0 0 Sex unidentifiable 2 0 2 0 0 0 0 0 0 Sosippus californicus (total) 6 2 8 0 0 0 0 0 0 Female 2 2 4 0 0 0 0 0 0 Male 4 0 4 0 0 0 0 0 0 Sub-adult 0 0 0 0 0 0 0 0 0 Gnaphosidae 9 21 30 16 2 18 0 2 2 Salticidae 0 0 0 1 0 1 1 4 5 Clubionidae 0 2 2 1 1 2 0 0 0 Oxyopidae 0 0 0 1 0 1 0 0 0 Theridiidae 0 0 0 0 0 0 1 3 4 Unknown spiders 2 0 2 11 10 21 4 8 12 Formicidae 1020 446 1466 521 782 1303 1298 1193 2491 Coleoptera 200 597 797 126 163 289 107 171 278 Isopoda 1316 85 1401 39 7 46 30 35 65 Acarina 2 184 186 74 55 129 132 60 192 Collembola 0 0 0 78 100 178 11 25 36 Dermaptera 171 32 203 0 0 0 0 0 0 Scorpiones 0 21 21 5 25 30 9 7 16 Miscellaneous available prey 2 17 19 43 26 69 9 31 40 Total available prey 1067 632 1699 625 708 1333 1258 1121 2379 these sites included members of the taxa: For- micidae, Isopoda, Coleoptera, Acarina, Col- lembola, Dermaptera, and Scorpiones. For- micids comprised the dominant group at all sites. Kendall’s rank correlation analyses of the relative abundance of the top ten arthropod groups found at each subsite indicated greater intrasite variability at Site 1 than at Sites 2-3 (Table 2). In addition rank correlation analy- ses indicated substantial differences in relative abundance of different arthropod groups be- tween Site 1 and Sites 2-3. However, Sites 2- 3 harbored strikingly similar arthropod assem- blages overall (Table 2). WENNINGER & FAGAN=-™RIVER FLOW MANIPULATION AND WOLF SPIDERS 119 Table 2.— Rank correlation coefficients for ar- thropod assemblages within and among pitfall sam- pling sites. Analyses involve the 10 most common arthropod groups except for analyses involving Site 3 in which only 9 groups were sufficiently common for analysis, * == significant at P — 0.05, ** — significant at F ~ 0.01. Sites compared Rank correlation coefficient Sub-sites at Site 1: 0.547* Sub-sites at Site 2: 0.786** Sub-sites at Site 3: 0.983** Site 1 vs. Site 2: 0.442 Site 1 vs. Site 3: 0.569 Site 2 vs. Site 3: 0.940** Dominance-diversity curves (Fig. 1) also reveal striking differences among sites. Ar- thropod collections at Site 1 are dominated by four groups of arthropods (Formicidae, Iso- poda, Pardosa vadosa, and Coleoptera), whereas Formicidae are clearly dominant at Sites 2 and 3. Calculating Simpson’s index of diversity also indicates higher terrestrial ar- thropod diversity at Site 1 (0,781) compared with Sites 2 and 3 (0.553 and 0349, respec- tively). Wolf spider abundance at each site showed no correlation with available prey (Kendall’s rank correlation; Fig. 2). Total wolf spiders collected at Site 1 far exceeded those collected at Sites 2 and 3, but available prey varied only slightly between sites. Average morning air and substrate temper- atures at Site 1 were lower than comparable averages from Sites 2 and 3 (MANOVA, Wilks’ Lambda = 0.572, P < 0.001). In ad- dition, substrate temperatures at Sites 2 and 3 were on average 2.4 °C and 1.5 °C higher, re- spectively, than corresponding air tempera- tures, while at Site 1 (the natural river site) average air and substrate temperatures were virtually identical. At Site 1, substrate tem- peratures on the cobblestones were generally warmer than the air and the soil was generally cooler. The relative abundance of lycosids de- creased as air temperature, substrate temper- ature, and the temperature difference between air and substrate increased (Fig. 3). Relative abundance of wolf spiders also increased with increasing relative soil moisture among sites. At Site 1 , where relative soil moisture ranged from 50-70%, wolf spiders represented be- tween 20-35% of the pittrap-collected fauna. Rank Figure L— Dominance diversity curves of the 10 most abundant groups of arthropods at each site. Counts from Sitel are sums of actual and projected counts. In contrast, at Sites 2 and 3, where relative soil moisture ranged from 0-10%, wolf spi- ders represented less than 2% of the pittrapped specimens. DLSCUSSION Overall, abiotic conditions and the diversity of available prey appear to influence wolf spi- der diversity and abundance in riparian and pseudo-riparian areas near Granite Reef Dam in central Arizona. In particular, the less ex- treme moisture and temperature regimes of the riparian habitat at Site 1 likely facilitated the greater abundance of wolf spiders there. Although substrate temperature was consis- tently higher than air temperature at Sites 2 and 3, substrate temperature differed little from air temperature at Site 1 , where high soil moisture levels likely contributed to a cooling effect. Experimental studies of microhabitat Figure 2.— Total lycosids trapped at each site compared to total available prey. Available prey in- cluded all soft-bodied arthropods that did not ex- ceed the average length of the largest wolf spider species found. Note logarithmic y-axis. 120 THE JOURNAL OF ARACHNOLOGY 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 ♦ Site ! A Site 2 O Site 3 0 12 3 4 Teniperature Difference (°C) Figure 3=— The proportion of wolf spiders (all species) at each sub-site as a function of tempera- ture regimes. Temperature readings were taken in early to mid-moming, before the most thermally- oppressive part of the day. Air temperature was tak- en approximately 2 cm above the ground surface. Error bars give standard errors of temperature read- ings. selection at Site 1 could support or refute the idea that spiders are actually selecting the riv- erside for (at least in part) its less extreme temperatures. Wolf spider abundance and diversity were also positively related to relative soil moisture among sites. These results agree strongly with our prediction that we would find the greatest abundance and diversity of wolf spiders by the natural river upstream of the dam, due to the importance of proximity to water and moist soil for iycosid distributions (Cherrett 1964; Kronk & Riechert 1979; Cady 1984; Agnew & Smith 1989; Bultman 1992). Indeed, Low- rie (1973) has shown that moisture is a key factor in the fairly specific habitat preferences of many species of Pardosa, the dominant ge- nus found in this study. In general, prey availability has been shown to be an important aspect of spider habitat as- sociation (Kronk & Riechert 1979; Moring & Stewart 1994; Henshel & Lubin 1997). But, because wolf spiders are known to require a variety of food items in order to reach matu- rity (Uetz et al. 1992), higher species diversity at Site 1 as opposed to the overall abundance of prey species (Fig, 2) may contribute to in- creased wolf spider abundance there. Similar plant species composition at Sites 2 and 3 likely contributes to the strong rank correla- tion among groups of arthropods between those sites, especially with respect to herbiv- orous insects (Table 2). All analyses of prey availability, however, must be viewed in the light that pitfall trap- ping is a sampling method with differential capture success among species. For instance, pitfall traps sample not the true density, bet rather the “active density” of wandering ar- thropods in an area over a given time (Uetz & Unzicker 1976; Uetz 1977). The dry pitfall traps likely under- sampled potential insect prey, as mentioned above, especially Diptera, which may comprise a significant component of Pardosa diet (Hallander 1970; Morse 1997; Nyffeler & Benz 1988; Nyffeler & Breene 1990). Predation by the numerous spiders in the dry traps at Site 1 may also have reduced prey availability at that site, possibly account- ing for the similarity in prey abundance col- lected at each site. Pitfall traps are still useful, however, in estimating the number of species of wandering spiders present over a wide range of habitats (Uetz & Unzicker 1976). Although we lack data on spider distribu- tions prior to dam construction, the results of this study suggest that wolf spider assemblag- es may have been substantially affected by dam construction, water diversion, and sub- sequent changes of the riparian vegetation in the vicinity of Granite Reef Dam. Our results support the hypotheses that desert riparian wolf spider-habitat associations are strongly influenced by soil moisture and substrate-air temperature regimes and that abundance of available prey alone may not be a good pre- dictor of wolf spider distributions. The impacts of urbanization on spider as- WENNINGER & FAGAN—RIVER FLOW MANIPULATION AND WOLF SPIDERS 121 semblages are worth investigating because spiders are not only an important food source for birds, lizards, wasps, and other species; but, when viewed as an assemblage of gen- eralist predators, they may also play an im- portant role in the regulation of insect popu- lations (Riechert & Lockley 1984; Settle et al. 1996; Morse 1997; Skerl 1997). Overall, the study of the ecological consequences of ur- banization for particular groups of plants and animals is important because it can indicate the degree of disturbance of their environ- ments and may be useful in developing strat- egies for conservation (Ruszczyk & Mellen- der 1992). Although this research was specifically designed as a summer study, when the desert environment was at its most ex- treme, it would be interesting to investigate if the striking patterns observed here persist within and among years, when the desert ri- parian sites experience a greater range of en- vironmental conditions. ACKNOWLEDGMENTS We are especially grateful to Dr. David Richman (New Mexico State University) for his species determination of the wolf spiders we studied. We thank Dr. Diane Hope and Rick Prigge for field site location assistance; Maggie Tseng for arthropod identification as- sistance; Jessamy Rango for literature refer- ences; and, for field work: Andy Chan, Tarek Eldin, Aaron McDade, and Lewis Rosenberg. E.W. received support from an NSF REU sup- plement to the Central Arizona -Phoenix Long-Term Ecological Research project at Ar- izona State University, funded by Grant #DEB-9714833. LITERATURE CITED Agnew, C.W. & J.W. Smith, Jr. 1989. Ecology of spiders (Araneae) in a peanut agroecosystem. En- viron. EntomoL, 18:30-42. Bultman, TL. 1992. Abundance and association of cursorial spiders from calcareous fens in south- ern Missouri. J. ArachnoL, 20:165-172. Cady, A.B. 1984. Microhabitat selection and lo- comotor activity of Schizocosa ocreata (Wal- ckenaer) (Araneae, Lycosidae). J. ArachnoL, 11: 297-307. Cherrett, J.M. 1964. The distribution of spiders on the Moor House National Reserve, Westmorland. J. Anim. EcoL, 33:27-47. Dreistadt, S.H., D.L. Dahlsten & G.W. Frankie. 1990. Urban forests and insect ecology. Bio- Science, 40:192-198. Frankie, G.W. & L.E. Ehler. 1978. Ecology of in- sects in urban environments. Ann. Rev. Ento- moL, 23:367-387. Gill, A.M. & J.E, Williams, 1996. Fire regimes and biodiversity: The effects of fragmentation of southeastern Australian eucalypt forests by ur- banisation, agriculture and pine plantations. For. EcoL & Manage., 85:261-278. Gregory, D.A. 1991. Form and variation in Ho- hokam settlement patterns. Pp. 159-193, In Cha- co and Hohokam: Prehistoric Regional Systems in the American Southwest. (PL. Crown & WJ. Judge, eds.). School of American Research Press, Santa Fe, New Mexico. Hallander, H. 1967. Range and movements of the wolf spiders Pardosa chelata (O.F. Muller) and Pardosa pullata (Clerck). Oikos, 18:360-364. Hallander, H. 1970. Prey, cannibalism, and micro- habitat selection in the wolf spiders Pardosa che- lata (O.F. Muller) and Pardosa pullata (Clerck). Oikos, 21:337-340. Henschel, J.R. & Y.D. Lubin. 1997. A test of hab- itat selection at two spatial scales in a sit-and- wait predator: A web spider in the Namib Desert dunes. J. Anim. EcoL, 66:401-413. Higgs, J.A. 1995. Granite Reef Diversion Dam Hy- draulic Model Study. Memorandum Report for The Salt River Project, Phoenix, Arizona, http:// ogee.do.usbr.gov/jhiggs/granite/granite.html Hoi Leitner, M.K. 1989. Changes In The Mam- malian Fauna Of The Neusiedler Lake region Austria Over The Last Three Decades. Bonner Zoologische Monographien 29. Hurd, L.E. & WE Fagan. 1992. Cursorial spiders and succession: age or habitat structure? Oecol- ogia, 92:215-221. Kaston, B.J. 1978. How To Know The Spiders. 3rd ed. Wm. C. Brown Co., Publishers, Dubuque, Iowa. Kronk, A.E. & S.E. Riechert. 1979. Parameters af- fecting the habitat choice of a desert wolf spider, Lycosa santrita Chamberlin and Ivie. J. Arach- noL, 7:155-166. Lowrie, D. 1973. The microhabitats of western wolf spiders of the genus Pardosa. EntomoL News., 84:103-116. McDonnell, M.J., S.TA. Pickett, P. Groffman, P. Bohlen, R.V. Pouyat, W.C. Zipperer, R.W. Par- melee, M.M. Carreiro, K. Medley. 1997. Eco- system processes along an urban-to-rural gradi- ent. Urban Ecosystems, 1:21-36. Moring, J.B. & K.W. Stewart. 1994. Habitat par- titioning by the wolf spider (Araneae, Lycosidae) guild in streamside and riparian vegetation zones of the Conejos River, Colorado. J. ArachnoL, 22: 205-217. Morse, D. 1997. Distribution, m.ovement, and ac- tivity patterns of an intertidal wolf spider Par- 122 THE JOURNAL OF ARACHNOLOGY dosa lapidicina population (Araneae, Lycosidae). J. ArachnoL, 25:1-10. Nowakowski, E. 1986. Structure of soil click bee- tle (Coleoptera, Elateridae) communities in urban green areas of Warsaw. Mem. ZooL, 41:81-102. Nyffeler, M. & G. Benz. 1988. Feeding ecology and predatory importance of wolf spiders (Par- dosa spp.) (Aranae, Lycosidae) in winter wheat fields. J. AppL Entomol., 106:123-134. Nyffeler, M. & R, Breene. 1990. Evidence of low daily food consumption by wolf spiders in mead- owland and comparison with other cursorial hunters. J. Appl. EntomoL, 110:73-81. Persons, M. & G. Uetz. 1996. The influence of sensory information on patch residence time in wolf spiders (Araneae: Lycosidae). Anim. Behav. 51:1285-1293. Riechert, S.E. & T. Lockley. 1984. Spiders as bi- ological control agents. Ann. Rev. EntomoL, 29: 299-320. Roth, V.D. 1993. Spider Genera of North America. 3rd ed. American ArachnoL Society, Gainesville, Florida. Ruszczyk, A. & A. Mellender. 1992. Gradients in butterfly species diversity in an urban area in Brazil. J. Lepid. Soc., 46:255-264. Sawoniewicz, J. 1986. Structure of Ichneumonidae (Hymenoptera) communities in urban green ar- eas of Warsaw. Mem. ZooL, 41:103-124. Settle, W.H., H. Ariawan, E.T. Astuti, W. Cahyana, A.L. Hakim, D.H. Hindayana, A.S. Lestari, Pa- jarningsih and Sartanto. 1996. Managing tropi- cal rice pests through conservation of generalist natural enemies and alternative prey. Ecology, 77:1975-1988. Skerl, K.L. 1997. Spider conservation in the Unit- ed States. Endangered Species Update. 14:9-14. http://www.umich.edu/'~esupdate/library/97.03- 04/skerLhtml Uetz, G.W. 1977. Coexistence in a guild of wan- dering spiders. J. Anim. EcoL, 46:531-542. Uetz, G.W. & J.D. Unzicker. 1976. Pitfall trapping in ecological studies of wandering spiders. J. Ar- achnoL, 3:101-111. Uetz, G.W, J. Bischoff & J. Raver. 1992. Survi- vorship of wolf spiders (Lycosidae) reared on different diets. J. ArachnoL, 20:207-211. Manuscript received 20 November 1998, revised 5 February 1999. 2000. The Journal of Arachnology 28:123-127 RESEARCH NOTE EXTENDED NEST RESIDENCE AND CANNIBALISM IN A JUMPING SPIDER (ARANEAE, SALTICIDAE) Keywords: Sociality, parental care, matriphagy, sub-social Menemerus bracteatus (L, Koch 1879) is a large unldentate Australian salticid that nests under the bark of eucalypt trees (Davies & Zabka 1989). During an earlier work (Rienks 1992) several nests were found on the scribbly gum, Eucalyptus racemosa Cavanilles. While studying the microhabitats of a wide range of salticid species, it was noted that sometimes a single nest of this species was occupied by numbers of large juveniles and the dead and shrunken body of a conspecific adult female, possibly the mother. It is common for juvenile salticids to remain with the mother for the first instar after emerging from the postembryo stage (Richman & Jackson 1992), but nest sharing by larger juveniles and an adult fe- male is unusual. My observations suggested the juveniles of M. bracteatus may feed on their mother, a behavior known as matriphagy. Matriphagy, although known in a variety of spider families, has not been documented for salticids. In the present paper, I provide data on nest structure, nest residence and predators other than conspecifics. Also, I examine the hypothesis that juveniles of M. bracteatus cannibalize their siblings and practice matri- phagy. Voucher specimens have been lodged with the Queensland Museum (QM S. 47 193). Four study sites in forests in which the scribbly gum was common were selected in the Brisbane metropolitan area. Three sites were in Toohey Forest, Griffith University Campus; and one was in woodland adjacent to Tingalpa Reserve, In each site, I sampled the occupants of as many nests as possible. No nests were found in the first search which was made when the scribbly gums had just begun shedding bark (late October to early November). A search between late December and early January revealed 35 nests in various stages of construction, 24 of which contained at least one clutch. A further seven nests found were from the previous season (evi- dence: constructed on older bark attached to the shedding bark, and contained shed exo- skeletons only). Description of nests.— -Nests were con- structed in the curve of the shedding bark and had tough roof and outer wails which strongly adhered to both the shedding bark and the tree trank. Each nest had two reinforced entrances with projecting lip-like flaps above and below the entrance slit which may hinder access by predators and parasites. Nests were very strongly constructed and could be removed in- tact by carefully pulling the loose bark piece away from the tree trank. Most nests appeared to be in an early stage of constraction. Of the 11 nests which con- tained no clutches, seven appeared to have been just begun and consisted of the outer walls only (three of these contained adult fe- males), while the other four contained what appeared to be preyed-upon clutches (stained mass in which some individual chorions could be distinguished) and may have been aban- doned. Counts were made of the number of clutches, and the number of eggs per clutch in each nest for 23 nests. A total of 54 clutch- es (median of two per nest) were found: 26% of the 23 nests contained one clutch, 30% con- tained two, 30% contained three and 13% contained four or more. Nests containing four or more clutches were completely filled with clutches and densely packed with loose sheets of very sticky silk laid down between succes- sive clutches. In contrast, nests with fewer clutches were only partially filled with a con- spicuous gap between the nest contents and the nest roof. The number of eggs per clutch was between 9-45 (mean 23.8, SD 6.9, n = 32) Eggs were 123 124 THE JOURNAL OF ARACHNOLOGY Table 1. — Stages found in previous season’s nests of Menemerus bracteatus. Numbers marked with an asterisk are nests in which adult remains were found. Nest 1, which may have been from the 1995-6 season, had been subjected to substantial insect attack; and its contents were almost entirely gone, leaving the outer nest wall only. Remains of individuals that had apparently been preyed upon are indicated by the number found, followed by “p”. Instar determination is based on size of the carapace of the shed exuviae. Nest Larva First instar Second in star Third in star Fourth instar Fifth instar Sixth instar 1 0 0 0 0 0 1 0 2* 19 25 32 + Ip 11 + Ip 0 0 0 3 27 26 35 22 0 0 1 4 27 25 + 2p 18 7 0 0 0 5* 35 19 24 18 4 1 0 6 36 36 26 29 2 0 0 7 26 + Ip 27 + Ip 26 + Ip 13 10 0 2 pale orange in color, did not adhere to each other, and were enclosed in a loose bag of non-sticky silk. Regardless of whether they were developing or apparently preyed-upon, clutches were included in calculations if the original numbers of eggs could be accurately determined. Numbers of eggs per clutch (clutch size) did not vary significantly with apparent order of laying (outermost clutch taken to be the most recent). Of the 23 nests in which clutches were examined in detail, the oldest stage of development included the egg (57% of nests), embryo (9%), prelarva (4%), larva (13%), first instar (13%) and second in- star (4%) (terminology after Foelix (1996) which follows that of Vachon (1957)). In total, 43% of the nests had one or more clutches showing signs of development. Three nests contained a clutch consisting of eggs, developing eggs, and prelarvae and/or larvae, suggesting that development of eggs within a clutch tends not to be synchronous. In two nests containing four or more clutches, clutch development showed a cohort effect with sep- aration of cohorts by up to one instar (i.e., modal numbers at every second stage). One of these nests contained 2-3 clutches of eggs (total of 71 eggs), 31 larvae and 34 first in- stars. The second nest contained one clutch of eggs (25 eggs), 19 prelarvae, 7 larvae, 31 first instars and 5 second instars. Nest residence by juveniles. — The seven nests collected that were from the previous (1996-7) season contained shed carapaces and exoskeletons of, in total, seven distinct stages, including the larval stage and six instars (Ta- ble 1). The largest carapace was considerably smaller than adult-sized. The numbers of car- apaces at each stage was more or less constant from the larval stage through to the third and sometimes the fourth instar (as shown by sec- ond and third instar carapaces). Numbers then declined rapidly, suggesting that dispersal had occurred in the third and fourth instars. If nests of this species usually contain about four clutches then it appears that the number of juveniles that survived to disperse as fourth (or occasionally third or fifth) instars, was equivalent to 1-1.5 full clutches. The presence of carapaces of large juveniles (fifth instar and older) suggested that juveniles may use the natal nest as a retreat for five or more instars. Cannibalism in the nest— Sometimes en- tire clutches, still enclosed in the silk bag, contained empty chorions, and had apparently been eaten. Such clutches were found in the nests from both seasons. Also, I found several apparently preyed-upon individuals (larvae and later stages) in three nests from the pre- vious season (Table 1). Two of the current sea- son’s nests that contained four clutches and were more developed than the other nests were examined in more detail for signs of can- nibalism. All first and later instars in both of these nests had grossly enlarged abdomens, consistent with having recently fed. All larvae had small abdomens which were similar in size to those of the prelarvae, suggesting that they had not fed. It appears that more clutches are laid than survive to disperse (see above). Since the number of larval carapaces in the previous season’s nests never exceeded 36 (the equiv- alent of just over one clutch), it is likely that RIENKS~-EXTENDED NEST RESIDENCE IN A SALTICID 125 the older instars preyed upon prelarvae and larvae in addition to eggs. Two nests from the previous season con- tained what were apparently adult remains, in one case the dorsal part of the carapace of an adult-sized individual, and in the other case, trachea attached to fragments of abdominal cuticle. It was not possible to determine whether these remains were those of adult fe- males. Predation.— Of the 23 nests collected in the 1997-8 season that were examined in de- tail, 26% contained one or two larvae of Aus~ tromantispa imbecilla (Gerstaecker) (Neurop- tera: Mantispidae). The mantispid larvae from each of these nests, all of which initially con- tained either two or three clutches, were reared until pupation. In all cases, only a few eggs and larvae survived, the rest apparently being consumed by the mantispid. Four other nests contained clutches that had apparently been preyed upon by other predators. In total, 43% of 23 nests contained preyed-upon clutches with some nests having one (22% of nests), two (17%) or three (4%) clutches af- fected. The young of M bracteatus postpone dis- persal from the natal nest until between the third and fifth instars, far later than is ob- served for most salticids. Another example of extended nest residence may also occur in Hy- paeus cucullatus Simon 1900, because fe- males and groups of juveniles of various sizes have been observed to share nests in this Cen- tral American salticid (Jackson 1989), but ex- amples of juveniles cohabiting beyond the first and second instar are better known in ma- ternal-social web-building spiders (Tretzel 1961, in Shear 1970; Kullmann 1972), and in spiders with more extended sociality (Jacson & Joseph 1973; Seibt & Wickler 1987; Evans et al. 1995). In M. bracteatus, the extended nest resi- dence, and the consequent large size attained by juveniles before dispersing, may be facili- tated by the laying of multiple clutches in the same nest. This provides opportunity for ju- veniles to feed upon sibling eggs and probably also larvae and prelarvae. The finding of adult remains in old nests suggests that, as in many maternal-social (Bristowe 1958; Tretzel 1961, in Shear 1970; Kullmann 1972) and perma- nent social species from families other than the Salticidae (Jacson & Joseph 1973; Seibt & Wickler 1987; Evans et al. 1995), matri- phagy may occur in this salticid species. Per- haps the adaptive significance of the long du- ration of nest residence by juveniles may be primarily facilitation of matriphagy. The laying of multiple clutches in the same nest probably does, however, have drawbacks. For other salticids comparable to M bractea- tus in size (females ranged from 9.5- 11.5mm), the interval between oviposition of successive clutches tends to be 20-30 days. Assuming that the inter-clutch interval is com- parable for M. bracteatus it is probable that a maternal female would need to make inter- mittent feeding forays away from the nest dur- ing the time span required for multiple ovi- position. While at the nest, the female may be able to guard her eggs against the attacks of predators and parasites, but leaving the nest to feed would be likely to expose her broods to higher risks of attack by other spiders, ants, beetles and acrocerid flies and so forth (Austin 1985). The tough nest construction and com- plex, dense sticky silk packing of M. bractea- tus nests may provide an exceptionally diffi- cult barrier for enemies to penetrate when the maternal females is away (see Austin 1985), but the protection provided appears to be lim- ited. The finding of preyed-\;tpo:n clutches in many nests, including some in nests that ap- peared to have been abandoned early during construction, suggests that predation while the female is away may be significant. Nests of M. bracteatus were also vulnerable to attack by A. imbecilla, a mantispid and a specialist predator of spider eggs. Mantispid larvae con- sumed virtually all the eggs in the nests ex- amined (see also Downes 1985) suggesting that maternal M. bracteatus, by laying all their clutches in the one nest, potentially place at risk their entire season’s, and perhaps life- time’s, reproductive effort. Females of M. bracteatus may lay all their clutches in one nest because overlying bark on scribbly gum tranks is both sparse and ephemeral, and so nest sites are in short sup- ply. Theory suggests that egg cannibalism and delayed juvenile dispersal may arise because the oviposition sites of females are widely separated from the juvenile habitat (which may be the case in M bracteatus) and that an unknown fitness advantage accrues to females by producing fewer, larger young (Crespi 1992). Alternatively, this delay, and the con- 126 THE JOURNAL OF ARACHNOLOGY sequent larger size of juveniles at dispersal, may be a fortuitous outcome of the opportu- nity to cannibalize siblings (and possibly, the mother) afforded by constraints on nesting sites. Cannibalism of siblings in the nest is com- mon in many solitary spider species (Krafft 1982) and it also occurs in some species with extended sociality (Evans et al 1995, but see Brach 1975). It has been argued that sociality in spiders evolved in some via an extension of an initial tolerant phase in the egg sac (Avi- les 1997). The occurrence of sibling canni- balism in M. bracteatus is therefore interest- ing because it exists alongside a tolerance amongst larger juveniles. Studies have shown that colonial-living web-building spiders capture more prey than solitary individuals, but that they are also sub- ject to “costs” unique to this way of life, i.e., an increased vulnerability to predators and parasites as the size of the colony increases (see Uetz & Hieber 1 997 and references there- in). Most social spider species occur in the tropics (Aviles 1997) where numbers of spe- cialist predators and parasitoids are very high (Begon et al. 1996). The occurrence in the sol- itary maternal social sub-tropical M. bractea- tus of high rates of nest predation by a spe- cialist mantispid egg predator raises the possibility that high rates of predation or par- asitism could be a cause rather than simply a consequence of group-living in spiders. ACKNOWLEDGMENTS This study was undertaken during study leave from the University of the South Pacific, whose support I gratefully acknowledge. I am most grateful to Robert Jackson for extensive comments on the manuscript, and to Petra Sierwald, Robert B. Suter and several anon- ymous reviewers for their helpful comments. Dr. Carla Catterall and the Australian School of Environmental Studies at Griffith Univer- sity in Brisbane very kindly provided research facilities for this study. I also wish to thank Drs. Valerie Todd Davies and Chris Burwell for identifying the salticid and the mantispid, respectively, and for providing copies of rel- evant articles. LITERATURE CITED Austin, A.D, 1985. The function of spider egg sacs in relation to parasitoids and predators, with spe- cial reference to the Australian fauna. J. Nat. Hist., 19:359-376. Aviles, L. 1997. Causes and consequences of co- operation and permanent sociality in spiders. Pp. 476-498, In The Evolution of Social Behaviour in Insects and Arachnids. (J.C. Choe & B.J. Crespi, eds.). Cambridge Univ. Press, Cam- bridge, Begon, M., J.L. Harper, & C.R. Townsend. 1996, Ecology, Individuals, Populations and Commu- nities. 3rd ed. Blackwell Scient. Publ., Boston, Oxford. Bristowe, W,S. 1958. The World of Spiders. Col- lins, London. Brach, V. 1975. The biology of the social spider Anelosimus eximius (Araneae: Theridiidae). Bull. Southern California Acad. Sci., 74:37-41. Crespi, B.J. 1992. Cannibalism and trophic eggs in subsocial and eusocial insects. Pp. 176-213, In Cannibalism. Ecology and Evolution among Di- verse Taxa. (M.A. Elgar & B.J. Crespi, eds). Ox- ford Univ. Press, Oxford. Davies, V.T. & M. Zabka. 1989. Illustrated keys to the general of jumping spiders (Araneae: Salti- cidae) in Australia. Mem. Queensland Mus., 27(2): 189-266. Downes, M.F. 1985. Emergence of Austromantispa imbecilla (Gerstaecker) (Neuroptera: Mantispi- dae) from the retreat web of Mopsus pencillatus (Araneae: Salticidae). Australian Entomol. Mag. 12(3):54. Evans, T.A., E.J. Wallis & M.A. Elgar. 1995. Mak- ing a meal of mother. Nature, 376:299. Foelix, R.F. 1996. The Biology of Spiders. 2nd ed. Oxford Univ. Press, New York, Oxford. Jackson, R.R. 1989. An unusual nest built by Hy- paeus cucullatus, a jumping spider (Araneae, Salticidae) from Costa Rica. Bull. British Arach- nol. Soc., 8(l):30-32. Jacson, C.C. & K.J. Joseph. 1973. Life history, bi- onomics and behaviour of the social spider Ste- godyphus sarasinorum Karsch. Insectes Sociaux, 20(2): 189-204. Krafft, B. 1982. The significance and complexity of communication in spiders. Pp. 15-66, In Spi- der Communiation: Mechanisms and Ecological Significance. (P.N. Witt & J.S. Rovner, eds.). Princeton Univ, Press, Princeton, New Jersey. Kullmann, E.J. 1972. Evolution of social behavior in spiders (Araneae; Eresidae and Theridiidae). American ZooL, 12:419-426. Richman, D. & R.R. Jackson. 1992, A review of the ethology of jumping spiders (Araneae, Sal- ticidae). Bull. British. Arachnol. Soc., 9:33-37. Rienks, J.H. 1992, Influences of microhabitat structure on the colour patterns of jumping spi- ders (Araneae: Salticidae). Ph.D. Thesis, Griffith University, Brisbane, Australia. Seibt, U. & W, Wickler. 1987. Gerontophagy ver- RIENKS—EXTENDED NEST RESIDENCE IN A SALTICID 127 sus cannibalism in the social spiders Stegodyphus mimosarum Pavesi and Stegodyphus dumicola Pocock. Anim, Behav., 35(6): 1903-1 904. Shear, W.A. 1970. The evolution of social phenom- ena in spiders. Bull. British Arachnol. Soc,, 1(5): 65-76. Uetz, G.W. & C.S. Hieber. 1997. Colonial web- building spiders: Balancing the costs and benefits of group-living. Pp. 458-474, In The Evolution of Social Behaviour in Insects and Arachnids. (J.C. Choe & B.J. Crespi, eds.). Cambridge Univ. Press, Cambridge. Jane H. Rienks: Biology Department, The University of the South Pacific, Suva, Fiji Manuscript received 20 April 1998, revised 27 Jan- uary 1999. 2000. The Journal of Arachnology 28:128-130 RESEARCH NOTE EGG SACS OF PITYOHYPHANTES PHRYGIANUS ARE NOT AFFECTED BY ACID RAIN Keywords: Air pollution, reproduction, spider embryos Acidic precipitation is one of the most im- portant air pollution problems today, causing ecological as well as physiological effects on terrestrial and aquatic animals (Newman et al. 1992). However, the effects on terrestrial ar- thropods are poorly known. In their review of pollution and insects, Heliovaara & Vaisanen (1993) found only three studies focussed on the direct effects of acid rain on terrestrial ar- thropods. In one of the studies, the growth rate of juvenile spiders exposed to simulated acid rain was examined (Gunnarsson & Johnsson 1989). However, earlier stages during devel- opment may be exposed to acid rain as well. Spiders deposit their eggs within an egg sac. The sac is a shelter for the eggs and made of spider silk, consisting of protein with alanine and serine as two major components (Foelix 1996). Each developing spider egg is protect- ed by the chorion layer (Foelix 1996). This means that in order to damage the embryonic spiders, the acidic water must penetrate not only the silken egg case, but also the chorionic egg shell. Here I examine the effects of simulated acid rain on egg sacs of the spruce-living (Picea abies (L.)) sheetweb spider Pityohyphantes phrygianus (C.L. Koch 1836), In south Swe- den, it has a biennial life-cycle. The males ma- ture before the females in late spring (Gun- narsson & Johnsson 1990), and mating takes place in May. In late June, the females start reproducing, placing their egg sacs directly on spruce branches. This means that the egg sacs in most areas of south Sweden are exposed to ambient concentrations of air pollutants, in- cluding acid rain, for about three weeks until hatching starts. Adult females were collected from spruce branches at different sites in coniferous forests 20-40 km east of Goteborg in SW Sweden. The females were collected at the end of June, when egg production starts. In the laboratory, the females were placed in 0.5 liter plastic vi- als with spruce twigs. They were fed with ves- tigial wing fruit flies {Drosophila melanogas- ter) ad libitum. The vials were sprayed with tapwater at regular intervals to maintain the humidity. All experiments were performed at room temperature (21-25 °C) and under nat- ural photoperiod. The females produced a first egg sac, which was attached to a twig or to the inside wall of the vial. The egg sacs were carefully removed and placed individually in 10 ml plastic vials, which were closed with a cotton ball. Ap- proximately 70% of the females produced a second egg sac. All further treatments of the egg sacs were randomized. Each egg sac, once a day, was gently sprayed with water of a particular acidity, which formed a cover of small drops on the egg sac and the insides of the vials. The spray- ing was done in a standardized fashion that was similar in all treatment groups. The con- trol group was sprayed with tapwater of pH 7 and the experimental groups with water of pH 4.0 (simulated acid rain; mean of bulk depo- sition in south Sweden is pH 4.3, see Balsberg P^hlsson & Bergkvist 1995), and pH 2.2. The solutions were obtained by using a stock so- lution of tap water for all treatments. Parts of this stock solution were mixed with diluted sulphuric acid. The pH of the solutions was checked at regular intervals and found to re- main constant. However, new solutions were prepared once during the experiment. This ex- periment is referred to as the “main experi- ment.” The egg sacs were checked in two ways. First, spiderlings that had emerged from the egg sac were recorded. If no spiderlings were 128 GUNNARSSON— -EGG SACS AND ACID RAIN observed, the egg sac was opened 25 days af- ter its deposition. Second, the hatching suc- cess was established by counting the numbers of hatched spiderlings and dead/undeveloped eggs. In egg sacs where spiderlings emerged spontaneously, the spraying of water was ceased on the day of emergence and all spi- derlings and any remaining eggs were checked after another two days. In the presentation of data, means are given together with their standard deviations. Non- parametric statistical methods were used since non-normality was observed in hatching data and transformation did not change this. All tests were two-tailed. To provide a comparsion with experimental results, egg sac production in a natural pop- ulation 40 km east of Goteborg was recorded in July. The number of eggs was counted and used for comparison with the experimental sit- uation. Egg sacs from the wild were not used in any experiment. A field-collected egg sac contained, on average, 43.2 ± 15.9 eggs {n = 17). However, there was a negative correlation between the collecting date and the number of eggs in the natural population (Spearman rank correlation test; g = —0.589, P = 0.019, n = 17). This suggests that females in the wild produced smaller clutches later in the season, possibly because there are fewer eggs in a sec- ond egg sac. It is known from other species that females produce fewer eggs in successive egg sacs (Foelix 1996). The mean number of eggs in an egg sac in i the main experiment was 36.4 ± 12.8 {n = 88). The egg numbers in the first and second egg sac were similar (Wilcoxon matched-pairs i signed-ranks test; z — —0.74, P = 0.46, n ~ j 31), and not correlated (Spearman, = 0.162, I P = 0.36, n = 31). Egg production in the lab- i oratory was similar to the natural population I (Mann-Whitney (7-test; z = —1.48, P = 0.14, ! rii = 88, «2 = 17). Spiderlings emerged spontaneously from egg sacs sprayed with water of different acid- ity except for those treated with water of pH 2.2. A comparison of egg sacs with sponta- neously emerging spiderlings in the main ex- ' periment showed a highly significant differ- I ence between the treatments (x^ = 26.20, df I = 2, F = 0.001): spiderlings emerged in I 72.7% {n = 33) of the egg sacs in the control I (pH « 7), 65.5% (n - 29) in pH 4.0, and 0% 1 {n = 17) in pH 2.2. 129 In the main experiment, the hatching suc- cess of the spiderlings in the first and second egg sac was similar within each treatment (Mann-Whitney I/-tests; 0.51 < P < 0.75). Consequently, first and second egg sacs were pooled in the analyses. Comparisons between the treatments (pHs « 7 (control), 4.0 (simu- lated acid rain), 2.2) showed that the hatching success differed significantly (Kruskal- Wallis one-way ANOVA; H = 13.43, df ^ 2, P = 0.0012). Multiple comparisons at the 5% level (Siegel & Castellan 1988), showed that the mean hatching rate in pH 2.2 (13.7% ± 17.0%, n = 11) differed from control (51.8% ± 37.2%, n = 33) and from simulated acid rain (43.1% ± 38.3%, n = 29), but there was no difference between the two latter treat- ments. Pooling the treatments of pH » 7 and 4.0 revealed a negative correlation between the number of eggs in each egg sac and the hatching success (Spearman, g = —0.504, P = 0.0001, n = 62). This was, however, not the case in pH 2.2 (g = 0.091, P = 0.11, n - 17). In an additional, small scale experiment one year after the main experiment, treatments with pH « 7 (control), pH 4.0, pH 3.5 and pH 3.0 solutions were performed as in the main experiment. The reason for doing this additional experiment was to examine the ef- fects of another two acidic solutions (pH 3.5 and 3.0), and test for a possible threshold be- low pH 4.0. This experiment was analyzed separately since it was performed at room temperature, i.e., there were slightly different conditions between years. In the additional experiment, approximately similar percentages (67-78%) of egg sacs with emerging spiderlings were observed among the groups (pHs » 7, 4.0, 3.5, and 3.0). The hatching success of spiderlings in the treatments pH — 7 (mean 80.9% ± 36.6%, n = 5), pH 4.0 (75.5% ± 35.2%, n == 8), pH 3.5 (83.3% ± 31.9%, n = 9), and pH 3.0 (68.8% ± 54.0%, « = 3) was similar (Krus- kah Wallis one-way ANOVA; H = 0.25, df = 3, P = 0.97). There was no correlation be- tween the number of eggs in each egg sac and the hatching success (Spearman, r, = 0.161, P = 0.43, n - 25). Obviously developing embryos are rather well protected against acid rain since only egg sacs treated with water of pH 2.2 showed a statistically significant deviation from the con- 130 THE JOURNAL OF ARACHNOLOGY trol. Examination of the egg sacs suggested that the outside structure of the sacs was af- fected at this low pH. The silk formed a dense mass of threads, which were glued together but with minute openings in between, in con- trast to the loose structure of threads in the unaffected egg sacs. This had two conse- quences: (1) the hatched spiderlings could not emerge from the egg sac, possibly because they could not find their way out of walls con- sisting of threads glued together; (2) the hatching of spiderlings was affected negative- ly, suggesting that acidic water entered the damaged egg sac and reached the developing embryos. The correlation between egg numbers on hatching success of spiderlings may be an ar- tifact due to disturbance. Removal of egg sacs from the deposition points may have caused unfavorable position changes of eggs within clutches. It is also possible that the water spraying was insufficient to support all eggs in large clutches with enough moisture. In nat- ural populations, the mean hatching success of spiderlings seems to be >90% (pers. obs.). The experimental hatching success was low even in the control, suggesting that the labo- ratory conditions affected the results, at least in the main experiment. However, the egg numbers per egg sac in the natural population and in the experiment were similar. The pH of throughfall water in spruce was slightly higher than bulk deposition in south Sweden, averaging 4.3--4.6 (Balsberg P&hls- son & Bergkvist 1995). Thus, there is no ev- idence suggesting that acid rain affects the de- velopment of embryos within spider egg sacs, unless under extreme conditions. Similar re- sults were obtained for growing juveniles of P. phrygianus (Gunnarsson & Johnsson 1989). In the present system, indirect effects of acid rain are more important. For instance, accelerated needle-loss is causing changes in predator-prey interactions, involving spiders and their predators (Gunnarsson 1995, 1996; Sundberg & Gunnarsson 1994). I thank K. Hellstrom, J. Johnsson and K. Madsen for laboratory assistance. This study was supported by the National Swedish En- vironment Protection Board. LITERATURE CITED Balsberg PShlsson, A-M. & B. Bergkvist. 1995. Acid deposition and soil acidification at a south- west facing edge of Norway spruce and Euro- pean beech in south Sweden. In Effects of acid deposition and tropospheric ozone on forest eco- systems in Sweden. (H. Staaf & G. Tyler, eds.). Ecol. Bull., 44:43-53. Foelix, R. 1996. Biology of Spiders. 2nd ed. Ox- ford Univ. Press, New York. 330 pp. Gunnarsson, B. 1995. Arthropods and passerine birds in coniferous forest: The impact of acidi- fication and needle-loss. In Effects of acid de- position and tropospheric ozone on forest eco- systems in Sweden. (H. Staaf & G. Tyler, eds.). Ecol. Bull., 44:248-258. Gunnarsson, B. 1996. Bird predation and vegeta- tion structure affecting spruce-living arthropods in a temperate forest. J. Anim. Ecol., 65:389- 397. Gunnarsson, B. & J. Johnsson. 1989. Effects of simulated acid rain on growth rate in a spruce- living spider. Environ. Pollut., 56:311-317. Gunnarsson, B. & J. Johnsson. 1990. Protandry and moulting to maturity in the spider Pityohy- phantes phrygianus. Oikos, 59:205-212. Heliovaara, K. & R. Vaisanen. 1993. Insects and Pollution. CRC Press, Boca Raton. 393 pp. Newman, J.R., R.K. Schreiber & E. Novakova. 1992. Air pollution effects on terrestrial and aquatic animals. Pp. 177-233, In Air Pollution Effects on Biodiversity, (J.R. Barker & D.T. Tin- gey, eds.). Van Nostrand Reinhold, New York. Siegel, S. & N.J. Castellan. 1988. Nonparametric Statistics. 2nd ed. McGraw-Hill, New York. 399 PP- Sundberg, I. & B. Gunnarsson. 1994. Spider abun- dance in relation to needle density in spruce. J. Arachnol., 22:190-194. Bengt Gunnarsson: Department of Zoolo- gy, Box 463, and Department of Applied Environmental Science, Box 464, Goteborg University, SE 405 30 Goteborg, Sweden Manuscript received 25 February 1999, revised 10 June 1999. 2000. The Journal of Arachnology 28:131--132 RESEARCH NOTE WHICH SPERMATHECA IS INSEMINATED BY EACH PALP IN THERAPHOSIDAE SPIDERS?: A STUDY OF OLIGOXYSTRE ARGENTINENSIS (ISCHNOCOLINAE) Keywords; Theraphosid copulation, insertion side, mono-palpectomized males Which female receptacle is reached by a particular (right or left) palpal organ and how deep the embolus is inserted are unresolved problems in mygalomorph spiders. Despite evidence for the sperm storage function of spermathecae in some haplogyne spiders (in- cluding Mygalomorphae) (Coyle et al. 1983), literature dealing with these questions is scarce. The complementarity between male and female genital structures has been, until now, the only useful evidence regarding the lateral correspondence and the depth of inser- tion in Mygalomorphae (Coyle et al. 1983; Costa & Perez-Miles 1998) studied both is- sues using copulations by mono-palpectomi- zed males and consequent histological iden- tification and location of sperm masses in the two spermathecal receptacles. Oligoxystre argentinensis (Mello-Leitao 1941) is a medium-sized theraphosid from temperate South America. Male palpal organs have a very long embolus (Fig. 1). The adult females have, attached to the bursa copulatrix, two separated spermathecal receptacles, each one consisting of a long stalk with a spherical fundus (Fig. 2). Three males and six females of this species were collected in Sierra de las Animas (34°45'S, 55°2rW), Maldonado, Uruguay. They were reared in petri dishes containing moist cotton, and fed mainly with larvae of Tenebrio sp. (Coleoptera, Tenebrion- idae). Females molted in the laboratory, thus they had empty spermathecae. The right palpal organ of each male was covered with a drop of paraffin to prevent its use, but the covered palps were autotomized one or two days after manipulation. After a week each male was placed together with a “virgin” female for the first series of copu- lations. At least a week after these encounters, each male was placed with another “virgin” female for a second series of copulations. As soon as copulations were finished females were mechanically sacrificed by a cephalotho- rax puncture, and their spermathecae were im- mediately removed by dissection. Spermathe- cae were fixed in paraformaldehyde, impregnated with osmium tetroxide and em- bedded in araldite. Longitudinal sections were stained with toluidine blue and examined with an optical microscope. Voucher specimens were deposited in the arachnid collection of the Facultad de Ciencias, Montevideo. In the first series of copulations each male inserted his only palp (the left) 2, 3 and 4 times, respectively. In the second series each of the three males performed two insertions. In the first series two females each had both spermathecal receptacles completely filled with sperm (Fig. 3) (the third specimen, cor- responding to the two-insertion copulation, was damaged). In the second series, one fe- male had her left spermathecal receptacle filled with sperm and the right one empty; while in the other two females, both sperma- thecal receptacles were empty. The only male which had inseminated a female in the second series had performed only two insertions in the first series. The availability of both filled and empty spermathecal receptacles made it possible for us to study and compare them. We observed in the spermathecal wall the presence of ori- fices and features that resemble pores and glands as described by De Carlo (1973) in species of Grammostola and Acanthoscurria. Sections of four sperm-filled receptacles and sections of five empty receptacles were mea- sured (in mm), with an accuracy of 0.01 mm. Mean total width (measured in the middle of 131 132 THE JOURNAL OF ARACHNOLOGY Figures 1, 2.~OUgoxystre argentinensis. 1, Left male palpal organ (ventral view); 2, Female sper- mathecae (ventral view). Scale = 1 mm. Figure 3.— Longitudinal section of a spermathe- cal receptacle of Oligoxystre argentinensis filled with sperm mass. their length) of filled receptacles was 030 (± 0.00 SD), while empty receptacles measured 0.274 (± 0.046 SD). The Student’s utest showed significant differences between them (? = 6.14, P < 0.001). Mean wall width (in- cluding inner cuticle and epithelial layer, fol- lowing De Carlo 1973) showed no significant differences between filled and empty recep- tacles (0.125 ± 0.17 SD and 0.126 ± 0.17, respectively). The mean width of the lumen was 0.063 (± 0.010 SD) in sperm-filled re- ceptacles, and 0.023 (± 0.022 SD) in empty ones. The Student’s utest showed significant differences between them {t = 3.63, P < 0.01). Results indicate that a given palp is able to inseminate either or both spermathecal recep- tacles in O. argentinensis. Unexpectedly, there is no evidence of morphological or ethological constraints which prevent a palp from deliv- ering sperm to either receptacle. Our findings also suggest that sperm are directly deposited by the embolus deep into the spermathecal re- ceptacles, since females were sacrificed soon after mating and no immediate sperm transfer mechanisms along the spermathecae are known. The increased receptacle lumen width in filled spermathecae resulted from the stretching (expansion) of the spermathecal wall rather than the reduction of wall thick- ness. Finally, the low insemination level ob- served in the second series lead us to suspect that these males had difficulties recharging the palpal organs after their first copulations. One possible explanation could be that the absence of one palp disturbed the sperm induction be- havior of these males. ACKNOWLEDGMENTS We are grateful to F. Coyle, W.E. Eberhard and B.A. Huber for their critical reading of the manuscript, and to P. Sierwald and J. Ber- ry for the editorial corrections. We thank O. Trajillo-Cenoz, A. Fernandez and G. Casa- nova (IIBCE) for their technical support. LITERATURE CITED Coyle, F.A., F.W. Harrison, W.C. MacGimsey & J.M. Palmer. 1983. Observations of the structure and function of spermathecae in haplogyne spi- ders. Trans. American Microsc, Soc., 102(3): 272--280. Costa, EG. & F. Perez-Miles. 1998. Behavior, life cycle and webs of Mecicobothrium thorelU (Ar- aneae, Mygalomorphae, Mecicobothriidae). J. ArachnoL, 26:317-329. De Carlo, J.M. 1973. Anatomia microscopica de las espermatecas de los generos Grammostola y Acanthoscurria (Araneae, Theraphosidae). Phy» sis, Secc. C. B. Aires, 32(85):343-350. Fernando G. Costab Fernando Perez- Miles^* 2 and Sylvia Corte^: ^Etologfa, Zoologia Experimental, IIBCE, Av. Italia 3318, Montevideo, Uruguay; ^Seccion En- tomologia and ^Secciqon Etologia, Facultad de Ciencias, Igua 4225, 1 1400 Montevideo, Uruguay Manuscript received 22 August 1998, revised 1 July 1999. INSTRUCTIONS TO AUTHORS (revised October 1999) 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 photocopies of illustra- tions to the Managing Editor. Do not send original handmade illustrations until the manuscript has been ac- cepted. Mail to: Petra Sierwald, Managing Editor; Division of In- sects, Dept, of Zoology, The Field Museum of Natural History, 1400 South Lakeshore Drive, Chicago, IL 60605 USA [Telephone: (312) 665-7744; FAX: (312) 665-7754; E-mail: psierwald@lmnh.org] The Managing Editor will either forward your man- uscript to one of the associate (subject) editors for the review process or initiate the review process directly. You will receive correspondence aclmowledging the receipt of your manuscript by the responsible associate or managing editor with the manuscript number of your manuscript. Please use this number in all corre- spondence regarding your manuscript. Correspondence regarding the review process should be directed to the respective associate or managing editor. Correspond- ence regarding all other matters should be directed to the Editor. After the manuscript has been accepted, the author will be asked to submit the manuscript on a computer disc. It must be in a widely-used program. Indicate clearly on the computer disc the word pro- cessing program used and the type of computer (Mac or PC). 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. 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 italicized and sepa- rated 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 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: James W. Berry, Editor-In-Chief; Dept, of Biolog- ical Sciences, Butler University, Indianapolis, Indi- ana 46208 USA [Telephone (317) 940-9344; FAX (317) 940-9519; E-mail: jwberry@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 of original artwork should be no more than 1 1 inches (28 cm) ;ts 14 inches (36 cm). Photocopies in review manuscripts should be reduced to the exact measurements that the author wants to appear in the final publication. Larger drawings present greater dif- ficulty in shipping and greater risks of damage for which the JOA assumes no responsibility. Make no- tations 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^. — A-us x-us, male from Timbuktu. 1, Left leg; 2, Right chelicera; 3, Dorsal aspect of geni- talia; 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, proofs and reprints. — There are no page charges, but authors will be charged for changes made in the proof pages. Authors will receive a reprint order form along with their page proofs. Reprints will be billed by the Allen Press. 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 28 Feature Articles Number 1 The Family Gallieniellidae (Araneae, Gnaphosoidea) in the Amerieas by Pablo A. Goloboff 1 Descriptions and Notes on the Genus Paradossenus in the Neotropical Region (Araneae, Trechaleidae) by Antonio D. Brescovit, Josue Raizer & Maria Engenia C. Amaral 7 Optical Structure of the Crab Spider Misumenops pallens (Araneae, Thomisidae) by Jose Antonio Corronca & Hector R. Teran 16 Male Dimorphism in Oedothorax gibbosus (Araneae, Linyphiidae): A Morphometric Analysis by Stefan Heinemann & Gabriele Uhl 23 Male Palpal Bulbs and Homologous Features in Theraphosinae (Araneae, Theraphosidae) by Rogerio Bertani 29 Exploring Functional Associations Between Spider Cribella and Calamistra by Brent D. Opell, Jamel S. Sandidge & Jason E. Bond 43 Characterization of Lipoproteins Isolated from the Hemolymph of the Spider Latrodectus mirabilis (Araneae, Theridiidae) by Monica Cunningham, Alda Gonzalez & Ricardo Pollero 49 Diet-Induced and Morphological Color Changes in Juvenile Crab Spiders (Araneae, Thomisidae) by Victoria R. Schmalhofer 56 Costs and Benefits of Foraging Associated with Delayed Dispersal in the Spider Anelosimus studiosus (Araneae, Theridiidae) by Thomas C. Jones & Patricia G. Parker 61 Resource Partitioning of Spider Hosts (Arachnida, Araneae) by Two Mantispid Species (Neuroptera, Mantispidae) in an Illinois Woodland by Kurt E. Redborg & Annemarie H. Redborg 70 Effects of Fertilizer Addition and Debris Removal on Leaf-Litter Spider Communities at Two Elevations by Angel J. Vargas 79 A Twenty- Year Comparison of Epigeic Spider Communities (Araneae) of Danish Coastal Health Habitats by Peter Gajdos & Soren Toft 90 Habitat Distribution, Life History and Behavior of Tetragnatha Spider Species in the Great Smoky Mountains National Park by Marie Aiken & Frederick A. Coyle 97 Spider Biodiversity in Connection with the Vegetation Structure and the Foliage Orientation of Hedges by FrMerickYsnel & Alain Canard 107 Effect of River Flow Manipulation on Wolf Spider Assemblages at Three Desert Riparian Sites by Erik J. Wenninger & William F. Fagan 115 Research Notes Extended Nest Residence and Cannibalism in a Jumping Spider (Araneae, Salticidae) by Jane H. Rienks 123 Egg Sacs of Pityphantes phrygianus Are Not Affected by Acid Rain by Bengt Gunnarsson 128 Which Spermatheca Is Inseminated by Each Palp in Theraphosidae Spiders?: A Study of Oligoxystre argentinensis (Ischnocolinae) by Fernando G. Costa, Fernando Perez-Miles & Sylvia Corte 131 pWT The Jou ARACHNOLOGY OFFICIAL ORGAN OF THE AMERICAN ARACHNOLOGICAL SOCIETY VOLUME 28 2000 NUMBER 2 THE JOURNAL OF ARACHNOLOGY EDITOR-IN-CHIEF: James W. Berry, Butler University MANAGING EDITOR: Petra Sierwald, Field Museum SUBJECT EDITORS: Ecology — ^Matthew Greenstone, USDA; Systematics — Mark Harvey, Western Australian Museum; Behavior and Physiology — Robert Suter, Vassar College 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; 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; N. I. Platnick, American Mus. Natural Hist.; S. E. Riechert, Univ. Tennessee; A. L. Rypstra, Miami Univ, Ohio; M. H. Robinson, US. 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 dind 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: Frederick A. Coyle (1999—2001), Department of Biology, Western Carolina University, Cullowhee, North Carolina 28723 USA. PRESIDENT-ELECT: Brent D. Opell ( 1 999—200 1 ), Department of Biology, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061 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, Department of Biology, University of Missis- sippi, University, Mississippi 38677 USA. SECRETARY: Alan Cady, Dept, of Zoology, Miami Univ, Middletown, Ohio 45042 USA. ARCHIVIST: Lenny Vincent, Fullerton College, Fullerton, California 92634. DIRECTORS: Bruce Cutler (2000-2002), Paula Cushing (1999-2001), Ann Rypstra (1999-2001). HONORARY MEMBERS: C. D. Dondale, H. W. Levi, A. F. Millidge, W. Whit- comb. Cover photo: SEM of cuspule of inner proximal maxmilla surface from exuvium of sub-adult Brachypelma boehmi (Araneae, Theraphosidae). Greatest width of cuspule=45 pm. {Photo by Bruce Cutler of the University of Kansas) Publication date: 29 September 2000 0 This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 2000. The Journal of Arachnology 28:133-140 DESCRIPTION OF THE MALE OF SOSIPPUS PLACIDUS, WITH NOTES ON THE SUBFAMILY SOSIPPINAE (ARANEAE, LYCOSIDAE) Petra Sierwald: Department of Zoology, Insects, Field Museum of Natural History, 1400 South Lake Shore Drive, Chicago, Illinois 60605 USA ABSTRACT. The male of the Florida funnel-web building wolf spider species Sosippus placidus Brady 1972 is described and figured for the first time. Analysis of the male palp’s morphological structure reveals that Sosippus possesses a median apophysis like other members of the Araneoclada, but which of the three additional tegular apophyses is the conductor cannot be determined at present. The study demonstrates that the palea, the putative key apomorphy of the clade Venoniinae-Allocosinae-Pardosinae-Lycosinae requires further morphological analysis. The genus Porrimosa is a close relative of the genus Sosippus based on shared characters in the male palp. The ontogeny of the female copulatory organs of Sosippus agrees with that of other members in the RTA clade. Hippasella nitida Mello-Leitao 1944, placed by Capocasale (1990) in the genus Sosippus, is not recognized as a congener. Keywords: Porrimosa, lycosid subfamilies, male palp structure Members of the North American wolf spi- der genus Sosippus Simon 1888 build rather large funnel-shaped capture webs in shrubs and herbaceous vegetation. The webs are strikingly similar to the “typical” agelenid webs. The web-building habit of Sosippus and a few other lycosid genera is frequently cited as “unusual” for the hunting spiders, as Ly- cosidae are often called (e.g., Gertsch 1949: 194). Simon (1898: 322) placed Sosippus and some of the other lycosid genera with long posterior spinnerets in the Hippasae (later sub- family Hippasinae), whereas he assigned Au- lonia C.L. Koch 1848, which also has long posterior spinnerets, to the Lycosae (later sub- family Lycosinae). Roewer (1959: 7) suggest- ed transferring all lycosid genera with long posterior spinnerets to the Hippasinae. Lehti- nen & Hippa (1979: 3) argued that the funnel- web is “simply a plesiomorphic character for a wide group of families within the Amauro- biomorpha,” rendering the Hippasinae as de- fined by Roewer polyphyletic. Citing genitalic characters, they placed Sosippus in the Lycos- inae. Dondale (1986: 329) introduced the new ly- cosid subfamily Sosippinae, containing Sosip- pus, “. . . Porrimosa and its relatives,” but did not provide a listing of all genera to be in- cluded. He also proposed a new subfamily system for the Lycosidae, placing the Sosip- pinae as sister taxon to the four other subfam- ilies, the Venoniinae, Allocosinae, Pardosinae and Lycosinae. Groups and nodes of Donda- le’s subfamily system are supported exclu- sively by morphological characters of the male palp. Since the Lycosidae is a species- rich family with considerable morphological diversity of the copulatory organs and the sis- tergroup to the Lycosidae is not yet known, the characters cited by Dondale require further analysis with regard to polarization (e.g., “loss” of terminal apophysis) and homology status. Zyuzin (1985, 1993) proposed a some- what different subfamily system for the Ly- cosidae, stressing the importance of characters derived from the copulatory organs as well. He did not discuss Dondale’s proposal of the new subfamily Sosippinae. In the present study, the male of S. placidus is described for the first time, and the struc- tural relationships of the sclerites in the gen- ital bulb are analyzed. Comparison with scler- ite and apophyses structure in palps of other lycosid groups will establish testable homol- ogy hypotheses required for further phyloge- netic analyses of the Lycosidae. The ontogeny of the female organs is also illustrated. Taxonomic history of the genus. — In the first revision of the genus, Brady (1962: 131) placed the then-known North and Central American Sosippus species in two groups: one 133 134 THE JOURNAL OF ARACHNOLOGY group with an eastern North American distri- bution, including S. floridanus Simon 1898, S. mimus Chamberlin 1924, and S. texanus Bra- dy 1962; and the other group with a western and Central American distribution, including S. californicus Simon 1898, S. mexicanus Si- mon 1888, S. michoacanus Brady 1962, S. plutonus Brady 1962, and S. agalenoides Banks 1909. In a subsequent study of the east- ern North American species floridanus species group), Brady (1972) described the additional species, S. janus and S. placidus. The latter species was based on female spec- imens alone; and its current known distribu- tion is restricted to Highlands County in cen- tral Florida, near Lake Placid. Capocasale (1990) transferred the South American Hip- pasella nitida Mello Leitao 1944 to the genus Sosippus. METHODS During studies in Florida, the author ob- tained nine juvenile S. placidus specimens from Dr. M. Deyrup, who collected them at Archbold Biological Station near Lake Placid at the original type locality. The specimens were reared in the lab; they built capture webs and took prey readily. Three males matured in April 1987. The molted exoskeletons of all specimens were collected. Exoskeleton sec- tions from between the book lungs were re- moved from the juvenile and subadult fe- males’ molts and mounted ventral side up on SEM stubs. The samples were air-dried and sputter-coated. SEM photographs were taken with several different scanning electron mi- croscopes at the Field Museum and at the Na- tional Museum of Natural History (Washing- ton, DC). Sosippus specimens of other species were borrowed from institutions listed in the acknowledgments. All measurements are in mm. In 1994, it was suggested that S. placidus be placed on the list of endangered and threat- ened species (U.S. Dept, of the Interior. Fed- eral Register 59(219): 58982), but this pro- posal was not adopted (U.S. Dept, of the Interior. Federal Register 61(40)). Sosippus placidus Brady 1972 Figs. 1-9 Sosippus mimus [in part], -Brady 1962: 156, figs. 34, 35. Sosippus placidus Brady 1972: 46, figs. 25-21 , 39, map 1. 9 holotype: USA, Florida, Highlands County, 6 miles W of Lake Placid, Brady & To- othaker coll.; MCZC. Brignoli 1983: 458. Diagnosis,— Sosippus placidus can be dis- tinguished from all other members of the ge- nus by the bright orange ventral coloration of sternum, legs and abdomen; large adult size (? 19-32 mm, S' 16-24 mm); and the three retromarginal cheliceral teeth, of which the in- nermost is twice as large as the other two. Sosippus texanus Brady 1962 is very similar to S. placidus in body size, number and size ratio of cheliceral teeth and morphology of the copulatory organs, but it lacks the bright or- ange ventral coloration. Almost the entire middle field of the epigynum (also called sep- tum by other authors) is very broad in S. tex- anus (see Brady 1962, figs. 21, 22), whereas it widens only posteriorly in S, placidus (Fig. 1). In the male palp, apophysis a is tapered with a round tip in S. texanus (see Brady 1962, figs. 37-39), whereas it carries a dis- tinctly swollen tip in S. placidus (Fig. 7). Relationships. — The eastern North Ameri- can species S. mimus, S. janus, and S. flori- danus possess four retromarginal cheliceral teeth (with some individual variation). The re- maining species in the genus have three che- liceral teeth; in S. texanus and S. californicus these have the same size ratio as in S. placidus (unknown for the remaining species). Detailed studies into such morphological characters may provide support for the delineation of species groups. Description. — Male: Measurements (3(3): body 16.0-24 long, carapace 6.2™7.5 long, 5. 3- 7.0 wide; sternum 3. 0-4.5 long, 2. 1-3.7 wide; labium 0.6-0.9 long, 0.2-0.5 wide. Right leg IV, femur 7.5-10.5 long, patella-tib- ia 10.0-12.0, metatarsus 10.0-12.5, and tarsus 3. 3- 6. 5; total leg length: 34.0-41.5. Males slightly smaller than females with longer legs than females (see below). Leg formula IV, I- II, III; length: leg IV 34-41.5; legs I and II 30-32; Leg III, 21-25. Spination of legs (see Table 1): spination of femur and patella iden- tical in all species of the genus (with some individual variation); spination of tibia and metatarsus with intraspecific variation espe- cially regarding the dorsal tibial spines on legs III and IV. Color pattern: Carapace orange brown (rust), eye region dark with eyes circled in black; a black thin stripe lining the periph- ery of the carapace. Chelicerae brownish- SIERWALD— MALE OF SOSIPPUS PLACIDUS 135 Figures 1-6. — Sosippus placidus, female copulatory organs; SEM. 1. External features; co = copulatory opening, ll = lateral lobe, mf = middle field; 2. Internal organs; bs = base of the spermatheca, fd = fertilization duct, hs = head of spermatheca, ss = stalk of spermatheca; vc = vulval chamber; 3. Head of spermatheca enlarged, showing pores; 4-6. Anlagen of the female copulatory organs, dorsal view, molts. 4. Penultimate instar; 5. Penultimate instar, head of spermatheca enlarged showing pores; 6. Antepenul- timate instar. Scale bars: Figs. 1, 2 = 0.2 mm; Figs. 3, 6 = 0.05 mm; Fig. 4 = 0.1 mm; Fig. 5 = 0.001 mm. 136 THE JOURNAL OF ARACHNOLOGY Figure 7. — Sosippus placidus, left male palp, ventral view; SEM. Abbreviations: a, b and c = tegular apophyses; dtp = distal tegular projection, ma = median apophysis, pr = palea region, st = lunar plate of subtegulum, t = tegulum. Scale bar: 0.5 mm. black; sternum pale orange to yellow (in aL cohol); labium and endites darker orange to reddish“black. Ventral surface of abdomen tan to orange, laterally darker orange to red with black hairs; dorsum light brown to tan. Legs brown with alternating light and dark bands; coxae and trochanters bright orange to yellow, covered with white hairs. Palp (Figs. 7, 9): apophysis a (labeled conductor by Brady (1962), see discussion below) distinctly swol- len and tip sclerotized (Fig. 7). Female: Measurements (8?): body 19.0- 32.0 long; carapace 6. 8-8.4 long, 6. 0-7.0 wide; sternum 3. 5-4. 7 long, 2.4-3. 7 wide; la- bium 0.6-1. 1 long, 0.3-0.6 wide. Right leg IV: femur 5.0-8.4 long, patella-tibia 8.4-10.0, metatarsus 4.2-8, 7, tarsus 3. 3-5.0. Total leg length: 24.0-30.3. Leg formula, spination and color pattern as in male (see also Brady 1972: 47), except for two dorsal rows of five white spots along the axis of the abdomen (see Bra- dy 1972, fig. 39). Epigynum (Fig. 1) with a wide posterior section of the middle field (sep- tum), and large copulatory openings in ante- rior region of the epigynal folds; internal or- gans (Figs. 2, 8, and see below) consist of true spermatheca with base, stalk and head, and a kidney-shaped sclerotized chamber (labeled VC vulval chamber). Specimens examined: Sosippus californicus: f MEXICO; Sonora, San Pedro Bay, 19, 17 July |i 1921, colL J.C. Chamberlin (paratype of S. prag- | maticus Chamberlin 1924); CASC. Sosippus flori- | danus: UNITED STATES: Florida, Alachua ^ County, Gainesville, 29 1 (5^ (immature), 16 Novem- i ber 1935, colL W.A. Murrill; MCZC. Alachua ; County, 16, 8 May 1934, colL A.F. Carr, det. Bra- I dy; AMNH. Highland County, Highland Hammock State Park, Sebring, 19, 24 March 1938, coll. Gertsch, det. Brady; AMNH. Sosippus janus: Flor- . ida, Alachua County, NW shore of Lake LocMoosa, ' 3 9 , 10 June 1968, coll. A.R. Brady and J. Toothak- er; MCZC. Alachua County, 1916, 18 April 1935, |' coll. A.F. Carr; AMNH. Sosippus mimus: Florida, t Liberty County, Blountstown, 66, 17 April 1938, l| coll. Gertsch; AMNH. Texas, Hidalgo County, Ed- inburg, 19, September-December 1933, coll. Mu- I laik; MCZC. Florida, Highland County, Lake Plac- J id, 1 9 ,1 6(immatere), 1943, coll. M. Cazier, det. Brady; AMNH. Liberty County, Blountstown, 16, 4 9 (immature), 17 April 1938, coll. Gertsch, det. i' Brady; AMNH. Sosippus placidus: Florida, High- !' land County, Lake Placid, Archbold Biological Sta- [ tion, 6936, September 1986, coll. M. Deymp; j; USNM, CASC, FMNH. Highland County, Lake Amiz, 29, 25 August 1975, coll. Brach; USNM, | MCZC. Sosippus texanus: Texas, Aransas County, i Goose Island State Park, 2916, 15 June 1961, coll. | A.R. Brady; MCZC. Hidalgo County, Edinburg, |' 19, September-December 1933, coll. Gertsch, det, ? Brady; AMNH. Porrimosa harknessi (Chamberlin [ 1916): PERU, Huadquina 5000 ft, 16 holotype, July 1911, Yale Peruvian Expedition; MCZC. | Structure and ontogeny of the Sosippus | vulva. — ^The female copulatory organs devel- op via the formation of paired longitudinal in- ‘ vaginations, termed epigynal folds (epf), . above the epigastric furrow (Figs, 4, 6). Such > folds have been observed in several families j in the RTA-clade sensu Coddington & Levi > (1991) (see Sadana 1972; Lachmuth et [ aL1985; Sierwald 1989). The internal female | organs (Fig. 2) consist of the true, tri-partite 5 spermatheca with base (bs), stalk (ss) and | head (hs) as identified for many Lycosoidea (Sierwald 1989; Griswold 1993). The head of the true spermatheca is clearly recognizable by its pores on the top (Fig. 3). Attached to the base of the spermatheca is a kidney- shaped, sclerotized chamber (vc. Fig. 2; la- beled B by Brady 1962, fig. 20), whose an- terior tip lies ventrally of the stalk of the spermatheca (Fig. 8). The copulatory opening is formed by the elongated epigynal folds; the internal sections of the folds are membranous SIERWALD— MALE OF SOSIPPUS PLACIDUS 137 Figures 8, 9, — Schematic drawings of copulatory organs. Stippling indicates sclerotized areas, lines indicate membranous sections; 8. Trajectory of ducts in vulva, schematic; abbreviations as in Fig. 2, cd = copulatory duct; 9. Sclerites of left Sosippus genital bulb in ventral view, schematic; a, b, c = tegular apophyses; p = protuberance of apophysis a; dtp = distal tegular projection, e — embolus, ma = median apophysis, pr = palea region, t = tegulum. Embolic division tilted out of original position, arrow indicates direction of tilt. and rather wide, thus resulting in a wide cop- ulatory opening (Fig. 1). The sclerotized base of the spermatheca is attached to the posterior end of the epigynal folds, enclosing the cop- ulatory duct (Figs. 2, 8). The copulatory duct branches into the duct of the spermathecal stalk and the duct leading into the kidney- shaped vulval chamber. The fertilization duct branches off from the vulval chamber duct. Antepenultimate and penultimate molts (Figs. 4-6) possess anlagen of the female or- gans. The ontogeny of these organs follows the same pattern observed in various Pisauri- dae (Sierwald 1989) and corresponds closely Table 1. — ^Leg spination in Sosippus placidus. Abbreviations: 1 = spine normal length, i = short spine, [ ] = common variation, [variations] = different variations common in this location. Femur Patella Tibia Metatarsus Leg I dorsal 11 i 0 0 0 Prolateral 11 1 1 1 1 i [variations] Retrolateral ill 1 1 1[0] 1 i [variations] Ventral 0 0 [1]11 11 ii 11 11 i Leg II dorsal 11 i 0 0 [1 1] 0 Prolateral 11 1 1 1 1 i [variations] Retrolateral ill 1 1 1 [0] 1 i [variations] Ventral 0 0 11 11 ii 11 11 i Leg III dorsal 11 i 0 0 [1 1] 0 ii Prolateral 11 1 1 1 1 1 i Retrolateral ill 1 1 1 1 1 i Ventral 0 0 11 11 ii 11 11 i Leg IV dorsal 11 i 0 0 [1 1] 0 ii Prolateral 11 1 1 1 1 1 i Retrolateral i 1 1 1 1 1 i Ventral 0 0 11 11 ii 11 11 i 138 THE JOURNAL OF ARACHNOLOGY to the one observed in Lycosa chaperi Simon 1885 by Sadana (1972). Anlagen are formed by paired longitudinal invaginations, with the future head of the spermatheca recognizable in early instars by its pores (penultimate an- lage, Figs. 4, 5; antepenultimate anlage. Fig. 6.). In the penultimate anlage (Fig. 4) the kid- ney-shaped vulval chamber is already recog- nizable. Structural analysis of the Sosippus palp. — The tegulum of the Sosippus genital bulb is ring-shaped (see Figs. 7, 9) as in many other lycosoids and agrees in its basic struc- ture with the pisaurid palp (see Sierwald 1990; fig. 2). The sperm duct enters the te- gulum dorsally, runs retrolaterally and turns into the ventral section of the tegular ring. Brady’s (1962: fig. 36) figure of a partially inflated S. californicus palp labels a median apophysis, conductor, basal haematodocha, lateral apophysis of conductor, mesal apoph- ysis of tegulum, the tegulum, and the embo- lus. The tegulum appears to carry four con- spicuous apophyses labeled here a, b, c and the median apophysis ma (Figs. 7, 9). Apoph- yses a and b originate from the dorsal section of the tegular ring. Apophysis a (labeled con- ductor by Brady 1962) is long, slender and finger-shaped. At its base it carries a small bi- lobed fleshy protuberance p, which is not vis- ible in the unexpanded bulb. Apophysis b (la- beled lateral apophysis of conductor by Brady) originates also in the dorsal section of the tegulum next to apophysis a. Apophysis b is long and sickle-shaped and lies transversely on the ventral surface in the unexpanded bulb. Apophysis c (not labeled but figured by Brady 1962 in fig. 36) originates from the ventral section of the bulb as an outgrowth of the teg- ular wall and is broad and flat. This apophysis is a thin, very translucent, triangular- shaped lamella, which may be difficult to discern un- der light microscopy. It is unclear at this point if any of these apophyses is a homologue to the pisaurid conductor. The fourth apophysis arises from the mem- branous center of the tegular ring and is most likely the homologue of the median apophysis ma (labeled mesal apophysis of tegulum by Brady 1962 in fig. 36; the identity of the part he labeled median apophysis is unclear). It is strongly sclerotized; and its dorsal rim is at- tached to a fringed lamella (Fig. 9), which lies in the notch below protuberance p in the non- inflated bulb (see Dondale 1986, fig. 2). The elongated tips of the sickle-shaped apophysis b, the median apophysis and apophysis c point in the same direction in the non-inflated bulb, with the tip of the embolus sandwiched be- tween the median apophysis and apophysis b. The section of the tegulum preceding the embolic division becomes very broad and strongly sclerotized (located prolaterally in the non-inflated bulb directly above the lunar plate of the subtegulum, Fig. 7) and corre- sponds to the distal tegular projection (dtp) in the pisaurid palp. The embolic division is con- nected to the tegulum by a rather narrow, mostly membranous stalk. In the unexpanded palp, this stalk is bent dorsally and retrolater- ally, bringing the embolic division over the tegulum, with the tip of the embolus pointing prolaterally. The base of the embolic division is a wide sac, its walls consisting of partially sclerotized and partially membranous sec- tions. The location of this large sac in the em- bolic division indicates that it is most likely homologous to the basal membranous tube and the distal sclerotized tube of the pisaurid bulb (see Sierwald 1990: fig. 3) and to the palea of other lycosids (labeled palea region (pr) in Figs. 7, 9). The embolus is spine-like, thin, curved and rather short, describing an incomplete loop. In Sosippus the embolic di- vision carries no apophyses as they occur in other lycosids. This study confirms the presence of the me- dian apophysis in the Sosippus palp as it has been proposed for the Araneoclada (see Cod- dington 1990: 10; Sierwald 1990: 44). How- ever, the status of the “conductor” in the So- sippus palp is unclear at this point. The conductor, as an outgrowth of the tegular wall, can be found in various families of the Ara- neoclada (e.g., Anyphaenidae, Pisauridae, Amaurobiidae, Psechridae, Araneidae and others). In the Sosippus palp, there are three tegular outgrowths (apophyses a, b, and c), each of which may represent the homologue of the Araneoclada conductor. The other two then represent evolutionary novelties. The long finger-shaped apophysis a is shared by all members of the genus and rep- resents a synapomorphy for Sosippus (see Brady 1962, figs. 34-47). Figures of the male palp of Hippasella nitida Mello-Leitao 1944 (Capocasale 1990: figs. 12, 13, Mello-Leitao 1944, fig. 32) indicate that this species does SIERWALD— MALE OF SOSIPPUS PLACIDUS not possess the finger-shaped apophysis, and as far as the figures can be interpreted, its palps have no close similarity with the Sosip- pus palp in general. In addition, Mello-Lei- tao’s description (1944: 343) of the size ratio of the eyes (anterior eyes larger than posterior eyes in H. nitida) exclude this species from the genus Sosippus (posterior median eyes distinctly larger than anterior eyes in Sosip- pus). Sosippus shares characters with Porrimosa Roewer 1960 (Brady 1962, fig. 33; Capoca- sale 1982, figs. 6~10). The embolic division is similar, consisting of a sac tilted dorsally and retrolaterally and a short, spine-like em- bolus, whose tip is sandwiched between the strongly sclerotized apophysis b and the tip of the median apophysis in the unexpanded palp. The median apophysis is smaller and less strongly sclerotized in Porrimosa than in So- sippus. Apophysis a is present, but it is short, broad and forms a hump (not long and finger- like, labeled conductor in Capocasale 1982, figs. 6“10). Apophysis c is represented by a low ridge arising from the ventral section of the tegulum. The shared characters in the palps of both genera support the close rela- tionship of both genera as mentioned by Don- dale. DISCUSSION Dondale’s subfamily proposal forms a valu- able starting point for the analysis of the ly- cosid interrelationship. The characters Don- dale employed for his analysis of lycosid subfamilies will require further analyses of the respective palpal structures and additional, in- dependent character systems should be includ- ed. According to his proposal the characters “terminal apophysis lost, tegular groove func- tioning as a conductor,” and “embolus laying in a cluster of tegular apophyses” are apo- morphies for the Sosippinae. Since the sister- group of the Lycosidae is not known yet, it is unclear at this point whether the absence of the terminal apophysis in the Sosippus palp represents a synapomorphy or is simply the plesiomorphic condition. The present study demonstrates that the “cluster of tegular apophyses” requires further detailed study in other lycosid groups to develop homology hy- potheses, especially to clarify the presence or absence of the Araneoclada conductor. The character “tegular groove functioning as a 139 conductor” cannot be evaluated at this point since the actual function of various parts of the palp is unclear (see Zyuzin 1985, 1993 for a detailed discussion). The character “palea developed,” the putative key apomorphy for the taxon Venoniinae-Allocosinae-Pardosinae- Lycosinae, equally requires further refine- ment, since it was demonstrated here that the large membranous sac at the base of the em- bolus in Sosippus consists of scleritized and membranous parts with similarity to the de- veloped palea in other lycosids. A detailed study of the palea morphology will provide further insight into this putative key apomor- phy for other lycosid groups. ACKNOWLEDGMENTS I wish to thank Dr. M. Deyrup (Archbold Biological Station, Florida) for the Sosippus specimens. Preserved material for this study was kindly loaned by Dr. J. Coddington and Scott Larcher (National Museum of Natural History, Washington, DC; USNM), Dr. H.W Levi (Museum of Comparative Zoology, Cambridge; MCZC), Dr. Norman I. Platnick (American Museum of Natural History, New York; AMNH), and Dr. C.E. Griswold (Cali- fornia Academy of Sciences; CASC). The SEM laboratories of the National Museum of Natural History (Washington, DC) and The Field Museum provided the use of their facil- ities. I am grateful to Drs. Bennett, Dondale, Stratton, Coddington and an anonymous re- viewer for their candid comments on earlier drafts of this manuscript. This study was fund- ed in part by a German Science Foundation grant to the author. Mr. Tariq Farooqui, an un- dergraduate student from North Park College, Illinois, collected the descriptive data on the males of Sosippus placidus. His participation was made possible through an NSF-Intemship grant to the Field Museum (DEB93- 17449). LITEATURE CITED Brady, A.R. 1962. The spider genus Sosippus in North America, Mexico, and Central America (Araneae, Lycosidae). Psyche, 69(3): 129-164. Brady, A.R. 1972. Geographic variation and spe- ciation in the Sosippus floridanus species group (Araneae: Lycosidae). Psyche, 79(l/2):27-47. Brignoli, P.M. 1983. A Catalogue of the Araneae, Described Between 1940 and 1981. Manchester Univ. Press. 755 pp. Capocasale, R.M. 1982. Las especies del genero 140 THE JOURNAL OF ARACHNOLOGY Porrimosa Roewer, 1959 (Araneae, Hippasinae). J. ArachnoL, 10:145-156. Capocasale, R.M. 1990. Las especies de la subfam- ilia Hippasinae de America del Sur (Araneae, Ly= cosidae). J. ArachnoL, 18:131-141. Coddington, J.A. 1990. Ontogeny and homology in the male palpus of orb=weaving spiders and their relatives, with comments on phylogeny (Ar- aneoclada: Araneoidea, Deinopoidea). Smithson- ian Contr. ZooL, 496:1-52. Coddington, J.A. & H.W. Levi. 1991. Systematics and the evolution of spiders (Araneae). Ann. Rev. Ecol. Syst., 22:565-592. Dept, of the Interior, US Fish and Wildlife Service. 1994. Endangered and threatened wildlife and plants; animal candidate review. 50 CFR Part 17. Federal Register, 59(219):58982-59028. Dept, of the Interior, US Fish and Wildlife Service. 1995. Endangered and threatened wildlife and plants. 50 CFR Part 17. Federal Register, 61(28): Candidate categories review -candidate taxa re- classification 7457-7463; plant and animal taxa: 7596-7613. Dondale, C.D. 1986. The subfamilies of wolf spi- ders (Araneae: Lycosidae): Pp. 327-332, In Ac- tas X Congreso Intemacional de Aracnologia (J.A. Barrientos, ed.). Jaca, Spain. Gertsch, W.J. 1949. American Spiders. D. van Nostrand Co., Toronot, New York, London. 285 pp. Griswold, C.E. 1993. Investigations into the phy- logeny of the lycosoid spiders and their kin (Arachnida, Araneae, Lycosoidea). Smithsonian Contr. ZooL, 539:1-39. Lachmuth, U., M. Grasshoff & EG. Barth. 1985. Taxonomische Revision der Gattung Cupiennius Simon 1891 (Arachnida: Araneae: Ctenidae). Senck. Biol., 65(3/6):329-372. Lehtinen, P. & H. Hippa. 1979. Spiders of the Ori- ental-Australian Region. I. Lycosidae: Venoni- ; inae and Zoicinae, Ann. ZooL Fennici, 16:1-22. i Mello-Leitao, C. 1944. Aranas de la Provincia de ' Buenos Aires. Rev. Mus. La Plata (N.S.) (ZooL), [ 3(24):3 11-393. 1; Roewer, C.-E 1959. Araneae Lycosaeformia Ila. i Lycosidae: 1-518. Exploration du Parc National de rupemba, Mission G.F. de Witte. Bruxelles. [ Sadana, G.L. 1972. Studies on the postembryonic j development of the epigynum of Lycosa chaperi Simon (Lycosidae: Araneida). Research Bulletin, | Panjab University, 23 (3/4); 243-247. j Sierwald, R 1989. Morphology and ontogeny of j female copulatory organs in American Pisauri- j' dae, with special reference to homologous fea- [ tures (Arachnida: Araneae). Smithsonian Contr. i ZooL, 484:1-24. Sierwald, P. 1990. Morphology and homological {, features in the male palpal organ in Pisauridae :: and other spider families, with notes on the tax- ' onomy of Pisauridae (Arachnida: Araneae). ;■ Nemouria, 35:1-59. j' Simon, E. 1888. Descriptions d’especes et de genres nouveaux de FAmerique centrale et des Antilles. Ann. Soc. Entomol. France, 8(6):203- ; 216. I; Simon, E. 1898. Histoire Naturelle des Araignees. | Roret, Paris 2(2): 193-380. j; Zyuzin, A. A. 1985. [Generic and subfamilial cit- ; eria in the systematics of the spider family Ly- !. cosidae (Aranei), with the description of a new genus and two new subfamilies]. Proc. ZooL j; Inst., Leningrad, 139:40-51 [in Russian]. Zyuzin, A. A. 1993. Studies on the wolf spiders ^ (Araneae: Lycosidae). 1. A new genus and spe- cies from Kazakhstan, with comments on the Ly- ; cosinae. Mem. Queensland Mus., 33(2):693-700. j tl Manuscript received 6 July 1999, revised 18 Jan- : uary 2000. ; 2000. The Journal of Arachnology 28:141-148 IRACEMA CABOCLA NEW GENUS AND SPECIES OF A THERAPHOSID SPIDER FROM AMAZONIC BRAZIL (ARANEAE, THERAPHOSINAE) Fernando Perez-Miles: Seccion Entomologia, Facultad de Ciencias, Igua 4225, 11400 Montevideo, Uruguay ABSTRACT. The new genus Iracema (Araneae, Theraphosidae, Theraphosinae) comprising the only species Iracema cabocla, from the Amazonic state of Roraima, Brazil, is described. The cladistic rela- tionships of this genus within the Theraphosinae are analyzed. Keywords: Theraphosid phylogeny, Amazonic spider, systematic s Theraphosidae is the most diverse family of the Mygalomorphae, comprising around 80 genera and 800 known species (Coddington & Levi 1991). The subfamily Theraphosinae only occurs in the New World, mainly in the Neotropics, and has been revised recently by Perez-Miles (1992, 1998) and Perez-Miles et al. (1996). Following the reasoning of Cod- dington & Levi (1991) that one third of all spider genera occur in the Neotropics and only 20% of world fauna is described, a large num- ber of Theraphosinae taxa are expected to be discovered, especially considering the poor knowledge of the group. Examining the spider collection of the INPA (National Institute for Amazonic Research, Manaus, Brazil) four specimens of Theraphosidae from Maraca, Roraima, Brazil, were found. These spiders did not fit with any known theraphosid genus, suggesting that they represent a new genus. The study of these spiders showed that they share the main synapomorphies of the Ther- aphosinae: extended subtegulum, keel on pal- pal bulbs, theraphosine types of urticating hairs and unilobular spermathecae, which en- couraged me to place this new genus within this subfamily. The addition of Iracema to a previous cladistic analysis showed that it would be the sister group of Cyriocosmus Si- mon 1903. METHODS Abbreviations: AME = anterior median eyes, ALE = anterior lateral eyes, PME = posterior median eyes, PLE = posterior lateral eyes, OQ = ocular quadrangle (including lat- eral eyes), d = dorsal; p = prolateral, r = retrolateral, v = ventral; INPA (Institute Na- cional de Pesquisas Amazonicas). All mea- surements are in mm and were taken using an ocular micrometer. Drawings were made with a camera lucida. Cladistic analysis was based on the previous matrix of Theraphosinae gen- era (Perez-Miles 1998, Perez-Miles et al. 1996) using the Pee-Wee (version 2.5.1) pro- gram, developed by Goloboff (1993); multi- state characters are considered as additive be- cause they are part of logically ordered transformation series or morphoclines. Other cladistic techniques follows Perez-Miles et al. (1996). Iracema new genus Type species. Iracema cabocla new spe- cies. Etymology. — Iracema (feminine) is an an- agram of America and the title of the most famous novel of the Brazilian writer Jose de Alencar, which describes the destruction and oppression of native Amazonic people through contact with civilization. Diagnosis. — Iracema differs from most theraphosid genera in the presence of a pro- cess in the retrolateral face of male palpal tib- ia. Additionally differs from several genera of Theraphosinae in the presence of Type IV ur- ticating hairs and in the very reduced number of labial cuspules. Females lack Type III ur- ticating hairs which are present in males. Ir- acema differs from Cyriocosmus in the palpal organ by lacking a paraembolic apophysis and in the spermathecae by the lack of a spiral neck and a caliciform fundus. Iracema differs 141 142 THE JOURNAL OF ARACHNOLOGY Figures 1-7. — Iracema cabocla new species. 1-5, Holotype male from Brazil, Roraima, Maraca. 1, Dorsal view (scale = 10 mm); 2, Right palpal tibia showing the retrolateral process (scale = 5 mm); 3, Right tibia I, distal portion showing the prolateral tibial apophysis (scale 5 mm); 4, Left palpal organ, prolateral view (scale = 1 mm); 5, Left palpal organ, retrolateral view (scale = 1 mm). 6, 7, Paratype female from Brazil, Roraima, Maraca. 6, Dorsal view (scale = 10 mm); 7, Spermathecae, ventral view (scale = 1 mm). PEREZ-MILES— NEW GENUS AND SPECIES OF THERAPHOSIDAE 143 Table 1. — Iracema cabocla new species. Male holotype and (male paratype), length of leg and palpal segments. I II III IV Palp Femur 10.8 (10.3) 9.5 (8.8) 8.5 (8.0) 11.2(10.5) 5.7 (5.7) Patella 5.0 (5.0) 4.7 (4.7) 4.0 (4.0) 4.7 (4.5) 3.3 (3.6) Tibia 9.2 (9.3) 7.2 (7.0) 6.5 (6.2) 9.3 (8.8) 5.3 (4.8) Metatarsus 8.2 (7.8) 7.8 (7.6) 8.6 (8.8) 12.5 (12.0) — Tarsus 4.8 (5.2) 5.0 (4.5) 5.0 (4.6) 5.4 (5.5) 1.9 (2.0) from Grammostola Simon 1892 by the ab- sence of stridulatory hairs and from Plesi- opelma Pocock 1901 by the absence of a nod- ule on the male metatarsus L The palpal organ of Iracema differs from that of Homoeomma Ausserer 1871 by the absence of a digitiform apophysis and from that of Paraphysa Simon 1892 by the presence of a process on the re- trolateral face of palpal tibia. It also differs from Paraphysa by the very reduced number of labial cuspules (3, being more than 10 in Paraphysa) and by the divided tarsal scopu- lae. This character was seriously questioned by Perez-Miles (1994) because it could main- ly reflect differences in size. All generic char- acters are coded in Table 3. Iracema cabocla new species Figs. 1-7; Tables 1-2 Types.— Holotype male, from Maraca, Ro- raima Brazil, 18 July 1987 (Steve Bowles in pit-fall trap). Paratypes: 24 July 1987 (Steve Bowles in pit-fall traps), 1 <32 $ from the same locality of the holotype. All specimens are de- posited in the collection of the INPA, Manaus, Brazil. Etymology, — The specific epithet is a noun in aposition from the Portuguese feminine word “cabocla” which refers to the people (women) from the Amazonic forests. Tradi- tionally it refers to the de-tribalized Indians and diverse racial mixture with Indian blood. Diagnosis. — The diagnostic generic char- acters of this monotypic genus can also be used to recognize the species Iracema cabo- cla. Description.— (holotype). Total length, not including chelicerae nor spinnerets 25.6; carapace length 11.2, width 10.33. An- terior eye row slightly procurved, posterior slightly recurved. Eyes sizes and interdist- ances: AME 0.38, ALE 0.43, PME 0.25, PLE 0.30, AME-AME 0.25, AME-ALE 0.20, PME-PME 0.82, PME-PLE 0.05, ALE-PLE 0.20, OQ length 0.9, width 1.7, clypeus 0.25. Fovea transverse, straight, width 1.7. Labium length 1 .4, width 1 .9 with 3 cuspules, maxillae with 66 cuspules. Sternum length 4.9, post- stemal sigilla oval, submarginal. Chelicerae with 9 teeth on the promargin (5 proximal of them smaller). Tarsi LIV densely scopulated, scopulae divided by a stripe of longer, thicker setae; this stripe is narrow in forelegs to wide in hindlegs. Metatarsi I and II scopulate on distal half. III apically scopulate, IV ascopu- late. Palpal tibia with a process on the retro- lateral face in distal portion (Fig. 2); ventrally two fields of spiniform hairs present (prola- teral and retrolateral). Tibia I with prolater- oventral, distal double apophysis (Fig. 3). Flexion of metatarsus I between tibial apoph- ysis. Palpal organ piriform, as in Figs. 4-5. Length of leg and palpal segments given in Table 2. — Iracema cabocla new species. Female paratypes (described), length of leg and palpal seg- ments. I II III IV Palp Femur 6.5 (6.7) 5.7 (5.7) 5.1 (5.3) 7.2 (7.1) 4.9 (4.9) Patella 4.0 (4.0) 3.4 (3.5) 3.1 (3.3) 3.5 (3.8) 2.9 (2.7) Tibia 5.3 (5.2) 4.2 (4.2) 3.7 (4.0) 5.7 (5.8) 3.3 (3.5) Metatarsus 4.2 (4.1) 4.2 (4.0) 4.6 (4.7) 7.3 (7.6) — Tarsus 2.8 (2.9) 2.7 (2.9) 3.1 (3.0) 3.4 (3.5) 3.5 (3.5) 144 THE JOURNAL OF ARACHNOLOGY , 0 OUTGROUP L67^14 HEMIRRHUGUS 1—66 9 EUATHLUS k-45-^19 MELLOLEITAOINA 1—32 tmesiphant:es (-56-,— 37-P-7 CYRIOCOSMUS L.16 IRACQ4A k-44=p-15 HOMOEOMMA 1—25 PLESIOPELMA 11 GRAMMOSTOLA 1—23 PARAPHYSA 65y-36=^5 CLAVOPELMA I L35-P-2 APHONOPELMA 1—4 CITHARACANTHUS 64-P-6 CYCLOSTERNUM L63t— 22 PAMPHOBETEUS L -40-r-31 THRIXOPELMA L.39 39-f— 1 ACANTHOSCURRIA 1— 38-j— 8 CYRTOPHOLIS L24 PHORMICTOPUS 10 EUPALAESTRUS L-60-P-17 LASIODORA L2I NHANDU 1—53.^33 VITALIUS L.! 52-P-29 SPHAEROBOTHRIA 1— 51-P-34 XENESTHIS L_50t— 3 BRACHYPELMA L_49.r-"18 RffiGAPHOBEMA 8 18-P-28 SERICOPELMA L4' L4 .^46^26 PSEUDOTHERAPHOSA I L.30 THERAPHOSA 3-p-27 SCHIZOPELMA -42 -r— 20 METRIOPELMA 1— 41-p-12 HAPALOPUS L.13 HAPALOT REMUS Figure 8. — Tree of genera of Theraphosinae, including Iracema new genus (fit 128.0, 80 steps). Table 1 . Femur III swollen. Spination: Femora MV and palp I 2P; II 2P; III 3R; IV IR; Palp IP. Patellae MV and palp 0. Tibiae I 2P, 2V; II 2P, 6V, IR; III 2P, 4V, 2R; IV 4P, 4V, 2R, Palp 2P Metatarsi I IV; II IP, 5V, IR; III 5P, 3-4 V, 2R, 2D; IV 7P, 4V, 6R, 0-lD. Tarsi I- IV without spines. Color: Cephalothorax red- dish-brown; legs and abdomen dark brown. Types III and IV urticating hairs present. Female: (paratype). Total length, not in- cluding chelicerae nor spinnerets 23.5. Ceph- alothorax length 9.5, width 8.5. Anterior eye row straight to slightly procurved, posterior row slightly recurved. Eye sizes and inter- distances: AME 0.40, ALE 0.45, PME 0.23, PLE 0.25, AME- AME 0.20, AME- ALE 0.10, PME-PME 0.78, PME-PLE 0.05, ALE-PLE 0.15, OQ length 0.7 width 1.3, clypeus 0.13. Fovea procurved width 2.0. Labium length 1.50, width 1.95 with 3 cuspules, maxillae with 90 cuspules. Sternum length 4.3, post- sternal sigilla oval, narrow, submarginal. Che- licerae with 9 teeth on the promargin (4 of them proximal, smaller). Tarsi densely scop- ulate, scopulae divided by a stripe of longer, thicker setae; this stripe is narrow in forelegs to wide in hindlegs. Metatarsi I and II scop- ulate on distal half. III apically scopulate, IV ascopulate. Length of leg and palpal segments in Table 2. Spination: Femora I-IV and palp, I IP; II IP; III IP, ID, IV ID; palp IP. Patellae I-IV and palp 0. Tibiae I-IV and palp, I 3V; II IP, 3V; III 2P, 3V, 2R; IV 2V, IR; palp IP, 3V Metatarsi MV, I 4V, II 2P, 4V, ID; III 2P, 4V, 2R; IV 2P, 8V, 5R, 2D. Tarsi MV and palp 0. Cephalothorax and legs light reddish- brown, abdomen grey-brown. Only Type IV urticating hairs present. Spermathecae with two receptacles only partly fused (Fig. 7). Distribution. — Iracema cabocla is only known from the type locality, Maraca, Rorai- ma, Brazil, with no further information avail- able. Cladistic relationships. — Including Irace- ma in the matrix of Perez-Miles et al. (1996: 37, table 2) modified in the analysis of Perez- Miles 1998a (Table 3); the most parsimonious tree of total fit 128.0 (43% ) and 83 steps was found (Fig. 8 ), in which Iracema was re- solved as the sister group to Cyriocosmus. Both share the retrolateral process on the pal- pal tibia (Perez-Miles 1998b). List of synapomorphies. — Acanthoscur- ria: char 4: 1 2, char 6: 0 1; Brachy- PEREZ-MILES— NEW GENUS AND SPECIES OF THERAPHOSIDAE 145 Table 3. — Character matrix of genera of Theraphosinae. Characters and (states) as follow: 0 Apical region of palpal bulb: subcylindrical (0), subconical (1), concave-convex (2). 1 Relative width of sclerites lU-III of bulb: narrow (less than 10% of length) (0), wide (1). 2 Paraembolic apophysis: absent (0), present (1). 3 Bulbal keels: smooth (0), serrated (1). 4 Bulbal keels: two subequal (0), two inequal (1), one peripheric (2), one peripheric plus supernumeraries (3). 5 Subtegulum: not extended (0), large extended (1). 6 Male tibial apophysis (leg I); double (0), one (1), absent (2). 7 Digitiform apophysis of bulb: absent (0), present (1). 8 Metatarsus I of male: without basal process (0), with basal process (1). 9 Male palpal tibia: without retrolateral process (0), with retrolateral process (1). 10 Male palpal tibia: without retrolateral cluster of spines (0), with retrolateral cluster of spines (1). 11 Male palpal tibia without prolateral process (0) , with prolateral process (1). 12 Flexion of metatarsus I (males): on outer side of tibial spurs (0), between tibial spurs (1). 13 Spermathecae: with two receptacles separated or only partly fused (0), widely fused (1), single semicircular receptacle (2), single oval receptacle (3). 14 Spermathecae multilobular in each side (0), unilobular at least in each side (1). 15 Femur III: not incrassate (0), incrassate (1). 16 Tibia IV: not incrassate (0), incrassate (1). 17 Femur IV: without retrolateral scopula (0), with retrolateral scopula (1) . 18 Urticating hairs type I: absent (0), present (1). 19 Urticating hairs type III absent (0), present (1). 20 Urticating hairs type IV: absent (0), present (1). 21 Trochanteral “stridulatory” hairs: absent (0), present (1). 22 Coxal “stridulatory” hairs: absent (0), present (1). 23 Coxal spinules: absent (0), present (1). 24 Labial cuspules: numerous (more than 15) (0), few or none (0). 25 Fovea: normal (0), with spheroid process (1). 26 Metatarsus I of males: normal (0), strongly curved (1). 27 Urticating hairs on prolateral palpal femur: absent (0), present (1). 28. Urticating hairs type VI: absent (0), present (1). 29 Coxae: normal (0), retrolaterally extended (1). 0 1 2 3 4 5 6 7 8 9 1 0 1 2 3 4 OUTGROUP 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Acanthoscurria 1 1 0 0 2 1 1 0 0 1 0 0 ? 0 1 Aphonopelma 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 Brachypelma 2 1 0 0 4 1 0 0 0 0 0 0 0 1 1 Citharacanthus 0 0 0 0 1 1 0 0 0 0 0 0 1 7 ? Clavopelma 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 Cyclosternum 1 1 0 ? 1 1 0 0 0 0 0 0 0 0 1 Cyriocosmus 0 0 1 0 1 1 0 0 0 1 1 0 0 0 1 Cyrtopholis 1 1 0 0 1 1 0 0 0 1 0 1 0 0 1 Euathlus 0 0 0 0 1 1 0 0 0 0 0 0 0 0 7 Eupalaestrus 1 1 0 1 1 1 0 0 0 0 0 0 0 0 1 Grammostola 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 Hapalopus 1 1 0 0 2 1 0 0 0 0 1 0 1 3 1 Hapalotremus 0 0 0 0 1 1 0 0 0 0 0 0 1 3 1 Hemirrhagus 7 7 7 ? ? ? ? 7 ? ? ? ? ? 0 1 Homoeomma 0 7 0 1 1 1 0 1 1 0 0 ? 1 0 1 Iracema 0 0 0 0 0 1 0 0 0 1 0 0 1 0 1 Lasiodora 1 1 0 1 2 1 0 0 0 0 0 0 0 1 1 Megaphobema 2 1 0 0 4 1 0 0 0 0 0 0 0 2 1 Melloleitaoina 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 Metriopelma 1 1 0 0 2 1 2 0 0 0 0 0 7 3 1 Nhandu 1 1 0 1 2 1 2 0 0 0 0 0 0 1 1 Pamphobeteus 2 1 0 0 1 1 0 0 0 0 0 0 1 0 1 Paraphysa 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 Phormictopus 2 1 0 0 ? 1 0 0 0 7 0 0 0 0 1 Plesioplema 0 0 0 0 1 1 0 0 1 0 ? 0 1 0 1 Pseudotheraphosa 2 1 0 0 3 1 0 0 0 0 0 0 0 2 1 Schizopeima 1 1 7 0 2 1 1 0 0 0 ? 0 ? ? ? Sericopelma 2 1 0 0 4 1 2 0 0 0 0 0 7 2 1 Sphaerobothria 2 1 0 1 3 1 0 0 0 0 0 0 0 1 1 Theraphosa 2 1 0 0 3 1 2 0 0 0 0 0 7 2 1 Thrixopelma 1 1 0 0 1 1 0 0 0 1 0 0 0 0 1 Tmesiphantes 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 Vitalius 1 1 0 1 2 1 0 0 0 0 0 0 0 1 1 Xenesthis 2 1 0 0 3 1 0 0 0 0 0 0 1 1 1 146 THE JOURNAL OF ARACHNOLOGY Table 3. — Extended. 5 6 7 8 9 2 0 1 2 3 4 5 6 7 8 9 OUTGROUP 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Acanthoscurria 0 0 1 1 1 0 1 0 0 0 0 0 0 0 0 Aphonopelma 7 0 0 1 0 0 0 0 1 0 0 0 0 0 0 Brachypelma 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 Citharacanthus 0 0 0 1 0 0 1 0 1 0 0 0 0 0 0 Clavopelma 0 0 0 1 1 0 0 0 1 0 0 0 0 0 0 Cyclosternum 0 0 0 ? 1 0 0 0 0 0 0 0 0 0 0 Cyriocosmus 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 Cyrtopholis 0 0 1 1 0 0 1 0 ? 0 0 0 0 0 0 Euathlus 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 Eupalaestrus 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 Grammostola 0 0 0 0 1 1 0 1 0 0 0 0 0 0 0 Hapalopus 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 Hapalotremus 0 0 0 0 1 0 0 0 0 1 0 1 0 0 0 Hemirrhagus 0 0 0 0 0 0 0 ? 0 0 0 ? 0 1 1 Homoeomma 0 0 0 0 1 1 0 0 0 0 0 7 0 0 0 Iracema 1 0 0 0 1 1 0 0 0 1 0 0 0 0 0 Lasiodora 0 0 1 1 1 0 0 1 0 0 0 0 0 0 0 Megaphobema 1 0 1 1 1 0 0 0 0 0 0 0 0 0 0 Melloleitaoina 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0 Metriopelma ? 0 0 ? ? 7 0 0 0 0 0 0 0 0 0 Nhandu 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 Pamphobeteus 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 Paraphysa 0 0 0 0 1 1 0 0 0 7 0 0 0 0 0 Phormictopus 0 0 ? 1 0 0 1 1 0 0 0 0 0 0 0 Plesiopelma 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 Pseudotheraphosa 0 0 1 0 1 0 0 1 0 0 0 0 0 0 0 Schizopelma 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 Sericopelma 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 Sphaerobothria ? 0 0 1 0 0 1 0 0 0 1 0 0 0 0 Theraphosa 0 0 1 0 1 0 0 1 0 0 0 0 0 0 0 Thrixopelma 0 0 1 7 7 7 0 1 1 0 0 0 7 0 0 Tmesiphantes 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 Vitalius 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 Xenesthis 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 pelma: char 17: 1 ->0; Cithar acanthus: char 12: 0 -A 1; char 21: 0 1; Cyriocosmus: char 2: 0 — > 1, char 10: 0 -a 1, char 19: 1 -A 0; Cyrtopholis: char 11: 0 — > 1; Eupalaestrus: char 15: 0 —> 1, char 16: 0 1; Grammostola: char 22: 0-^1; Hapalopus: char 10: 0 -A 1; Hapalotremus: char 0: 1 — > 0, char 1: 1 “A 0, char 4: 2 -A 1, char 24: 0 -a 1; Hemirrhagus: char 28: 0 -a 1, char 29: 0-^1; Homoeomma: char 3: 0 -A 1, char 7: 0 -A 1; Iracema: char 4: 1 -A 0, char 12: 0 -a 1, char 15: 0 -a 1, char 24: 0 -A 1; Lasiodora: char 22:0 -a 1; Megaphobema: char 15: 0 -A 1; Mellolei- taoina: char 24: 0^1; Metriopelma: char 6: 1 -A 2; Nhandu: char 6: 0 A 2; Pamphobe- teus: char 0: 1 A 2, char 12: 0 A 1; Phor- mictopus: char 0: 1 A 2, char 22: 0 a 1; Pseudotheraphosa: char 6: 1 A 0; SericopeP ma: char 6: 1 A 2; Sphaerobothria: char 17: 1 A 0, char 21: 0 a 1, char 25: 0 a 1; Ther~ aphosa: char 6: 1 a 2; Thrixopelma: char 22: 0 A 1, char 23: 0 A 1; Xenesthis: char 12: 0 A 1 . Node 35: char 19: 1 A 0; Node 36: char 23: 0 A 1; Node 37: char 9: 0 A 1; Node 38: char 19: 1 A 0; Node 39: char 21: 0 A 1; Node 40: char 9: 0 A 1; Node 41: char 6: 1 A 0, char 26: 0 a 1; Node 42: char 17: 1 A 0; Node 43: char 0: 2 a 1, char 4: 3 a 2; Node 44: char 8: 0 a 1, char 12: 0 A 1; Node 45: char 15: 0 A 1; Node 46: char 22: 0 A 1; Node 47: char 4: 4 A 3, char 18: 1 A 0; PEREZ-MILES— NEW GENUS AND SPECIES OF THERAPHOSIDAE 147 — 0 OUTGROUP L65^14 HEMIRRHAGUS L-64=^9 EUATHLUS .40-j=^18 MELLOLEITAOINA L3I TMESIPHANTES E_54^7 CYRIOCOSMUS —11 GRAMMOSTOLA —22 PARAPHYSA -39-^15 HOMOEOMMA L24 PLESIOPELMA j— 0 OUTGROUP L65-p»14 HEMIRRHAGUS ^54^7 CYRIOCOSMUS Lsa^ll GRAMMOSTOLA L22 PARAPHYSA 1-39-^15 HOMOEOMMA L24 PLESIOPELMA 64-r— 9 EUATHLUS 40=^18 MELLOLEITAOINA L3I TMESIPHANTES 10 _ 0 OUTGROUP L65t— 14 HEMIRRHAGUS 54-^7 CYRIOCOSMUS 11 GRAMMOSTOLA 22 PARAPHYSA 39^15 HOMOEOMMA L24 PLESIOPELMA U-40-P-18 MELLOLEITAOINA L_31 TMESIPHANTES U9 EUATHLUS 11 Figures 9-11. — Alternative resolutions of node 65 of three trees of maximum fit, from the matrix of Perez-Miles 1998 (unpublished). Node 54 in- cludes genera which share the apomorphic presence of Type IV urticating hairs. Node 48: char 6: 0 1; Node 49: char 13: 1 2; Node 50: char 4: 3 —» 4, char 19: 0 1; Node 51: char 3: 1 -a 0; Node 52: char 0: 1 -A 2, char 4: 2 3; Node 53: char 19: 1 0; Node 56: char 20: 0 1; Node 60: char 4: 1 2, char 13: 0 1; Node 61: char 3: 0^1; Node 63: char 17: 0 -A 1; Node 64: char 0: 0 “> 1, char 1: 0 — > 1; Node 65: char 18: 0 -> 1; Node 66: char 19: 0 1; Node 67: char 14: 0 -4 1. DISCUSSION these nodes relate Cyriocosmus, Grammosto= la, Paraphysa, Homoeomma and Plesiopelma based on the synapomorphic presence of Type IV urticating hairs, also present in Iracema. The sexual dimorphism of Iracema with re- spect to the presence of urticating hairs is re- markable; the male has types III and IV while female has only type IV. Recently Bertani (1997) found that in several theraphosid spe- cies the males have types I and III urticating hairs while females have only type I. The study of juveniles and the ontogeny of urti- cating hair types in these species seems to be crucial in determining if Types I and III are lost in females or gained by males during de- velopment. This fact has ecological implica- tions related to adult life strategies, mainly re- lating to theraphosid males being more errant than females. Also, this kind of study could enlighten the phylogenetic aspects of these important characters as well as the possible different functions of different hair types. The presence of a process in the retrolateral face of palpal tibiae in Iracema, somewhat re- sembles the process of Cyriocosmus indicated by Perez-Miles (1998b), but in the latter genus the process includes a field of spines absent in Iracema. This process is also present in oth- er genera group which lack Type IV hairs: Acanthoscurria Ausserer 1871, CyrtophoUs Simon 1892 and Phormictopus Pocock 1901. Another striking character of Iracema is the extreme reduction of labial cusps found only in other two theraphosid genera: Hapalotre- mus Simon 1903 and Hapalopus Ausserer 1875 (the last without labial cusps); these gen- era are far from Iracema considering other characters (Table 3). They also lack Type IV urticating hairs. In a previous analysis of the Theraphosinae, Perez-Miles (1998) obtained three trees of maximum fit (131.5 and 79 steps). These trees show differences in the internal relationships of the node involving genera with Type IV urticating hairs (Figs. 9-11). The present cla- distic analysis resulted in only one tree which is better resolved with the inclusion of Ira- Iracema is included in Theraphosinae by sharing the synapomorphies of the subfamily. In the cladogram (Fig. 8) Iracema was related with the node 56 which includes the group of genera of node 43 in the cladogram of Perez- Miles et al. (1996: 43, fig. 2). In both trees cema. ACKNOWLEDGMENTS I am indebted to Catarina da Silva Motta (INPA) for the loan of the specimens here de- scribed; to EG. Costa, M. Simo, J. Berry, P. 148 THE JOURNAL OF ARACHNOLOGY Sierwald and two anonymous reviewers for the critical reading of the manuscript. LITERATURE CITED Bertani, R. 1997. Estudo comparative dos pelos urticantes em Theraphosinae (Araneae, Thera- phosidae). Act. Primer Encuentro Aracnologos Cono Sur, Montevideo, p. 29. Coddington, J.A. & H.W. Levi. 1991. Systematics and evolution of spiders (Araneae). Annu. Rev. Ecol. Syst., 22:565-592. Goloboff, P.A. 1993. Pee-Wee, version 2.00. Pro- gram and documentation, distributed by the au- thor, San Miguel de Tucuman. Perez-Miles, F. 1992. Analisis cladistico preliminar de la subfamilia Theraphosinae (Araneae; Ther- aphosidae). Bol. Soc. Zool. Uruguay (2a. epoca), 7:11-12. Perez-Miles, F. 1994. Tarsal scopula condition in Theraphosinae (Araneae, Theraphosidae): Its systematics significance. J. ArachnoL, 22:46-53. Perez-Miles, F. 1998a. A phylogenetic analysis of Theraphosinae (Araneae, Theraphosidae). XIV Internal. Congr. ArachnoL, Chicago, p. 28. Perez-Miles, F. 1998b. Revision and phylogenetic analysis of the Neotropical genus Cyriocosmus Simon, 1903 (Araneae, Theraphosidae). Bull. British ArachnoL Soc., 1 1(3):95-103. Perez-Miles, F, S.M. Lucas, PI. da Silva, Jr. & R. Bertani. 1996. Systematic revision and cladistic analysis of Theraphosinae (Araneae: Theraphos- idae). Mygalomorph, 1:33-68. Manuscript received 10 April 1999, revised 1 Oc- tober 1999. 2000. The Journal of Arachnology 28:149-157 RESPIRATORY SYSTEM MORPHOLOGY AND THE PHYLOGENY OF HAPLOGYNE SPIDERS (ARANEAE, ARANEOMORPHAE) Martin J. Ramirez: Laboratorio de Atropodos, FCEyN, Universidad de Buenos Aires, Pabellon II Ciudad Universitaria (1428), Buenos Aires, Argentina, and Museo Argentino de Ciencias Naturales, Av. Angel Gallardo 470 (1405), Buenos Aires, Argentina ABSTRACT. The morphology of the respiratory system of basal araneomorph spiders, the Haplogynae and of Entelegynae with female haplogyne genitalia, is reviewed. The homology of cuticular respiratory structures is discussed in light of evidence from abdominal muscles and ontogeny. Ten morphological characters (13 transformations) were coded, mainly from the posterior pulmonary (or tracheal) segment, and other 7 non-respiratory characters here added. The new data were combined with those of a previously published analysis, resulting in a data matrix of 82 characters scored for 44 terminals. The evolution of the tracheal system is traced through the phylogeny of basal spiders and the Haplogynae, and new syna- pomorphies are provided. Elongate 3rd abdominal entapophyses are a synapomorphy of Araneomorphae. True median tracheae are a synapomorphy of Entelegynae (convergently with Austrochilinae), as is the extreme posterior displacement and narrowing of the tracheal spiracle. Tetrablemmidae, Pholcidae, Di- guetidae and Plectreuridae are united by the absence of tracheae; and these taxa are united with Scytodidae, Sicariidae and Drymusidae by the fusion of 3rd entapophyses. Keywords: Tracheae, cladistics, abdominal muscles Since the seminal and detailed works of Bertkau (1872, 1878), much attention has been devoted to the respiratory system of spi- ders. Although the morphology and diversity of respiratory structures was repeatedly used in classifications (e.g., Bertkau 1878; Petrunk- evitch 1933; Forster 1970), most attempts to depict the evolution of the respiratory organs in spiders were discouraging because of in- congruity with other character systems, which led some authors even to negate the value of the respiratory organs to define higher groups (Lamy 1902; Levi 1967). The efforts were un- able to overcome the obstacle of evaluating all character systems simultaneously. Fortu- nately, cladistic theory has provided the tools to manage all data globally; and the difficult task was recently achieved for basal araneo- morphs and haplogyne spiders (Platnick et al. 1991). The aim of this contribution is to in- vestigate once again the evolutionary trans- formations of the respiratory system through spider phylogeny, testing previous hypotheses of relationships in the light of new data. Homology and ontogeny of respiratory structures. — Purcell (1909) convincingly demonstrated that lateral tracheae of araneo- morph spiders originate as modifications of the posterior book lungs, and median tracheae as modifications of the entapophyses of the same segment. Median tracheae are distin- guished from hollowed entapophyses (also called apodemal lobes) by their much more elongate shape, and by their thin cuticle; in some cases they still retain their connection with abdominal muscles (Lamy 1902). There has been some confusion in the literature about the “transverse duct” or “interpulmon- ary” or “inter- tracheal canal of communica- tion.” In many spiders, the minute projections lining respiratory cuticles (called “spicules”) also extend to cover the innermost part of the interpulmonary or inter-tracheal furrow. For the tracheal segment, Purcell (1909: 65) called this “intertracheal canal of communication,” defined as “a canal connecting the median trunks with one another and with the lateral trunks at their base,” and identified the struc- ture as serially homologous with the interpul- monary canal of communication. Other au- thors (e.g., Forster & Platnick 1984) called the same structure “transverse duct.” If not to- 149 150 THE JOURNAL OF ARACHNOLOGY pologically definite as a “duct” (as discussed by Hormiga 1994), this canal becomes a func- tional duct because the spicules prevent the smooth anterior and posterior walls of the fur- row or tracheal vestibule from collapsing to- gether (Purcell 1909: fig. 26). I will follow here the original and accurate wording of Pur- cell. METHODS Tracheae and other cuticular structures were observed after digestion of tissues with a 10~ 20% KOH solution at approximately 100 °C in a double boiler or hot plate. Dissections for muscle observations were made on regular al- cohol-fixed specimens. Small structures were mounted in lactic acid or clove oil, and ob- served with a compound microscope. This analysis complements Platnick et al. (1991), and so numbers for characters follow that pa- per. RESPIRATORY SYSTEMS OF THE REPRESENTATIVE TAXA Most data on tracheae, entapophyses and muscle attachments were extracted from the general works by Lamy (1902), Purcell (1909, 1910), Kastner (1929), and references therein. Data on particular groups were found in For- ster et al. (1987: Austrochiloidea, Hypochilo- idea), Ramirez & Grismado (1997: Filistati- dae), Forster (1995: Scytodidae, Drymusidae, Sicariidae and Periegopidae), Platnick (1989: Diguetidae), Forster & Platnick (1985: Dys- deroidea), Forster & Platnick (1984: Palpi- manoidea), Platnick et al. (1999: Palpimani- dae), and Forster (1970: Entelegynae). The new data are discussed below. Austrochilinae: There is a wide furrow linking three paired structures (Fig. 4), de- scribed by Forster et al. (1987): “the inner pair are in fact apodemes [...]. The middle pair of tubes (those immediately lateral to the apodemal lobes) could be homologous with one of the book lung lamellae, but the outer pair are more likely to represent the marginal extensions of the original atrial pouch, which in most spiders [. . .] tend to be arcuate.” Their interpretation agrees with my observa- tions. The inner pair connects with the median longitudinal muscles. In the early instars of Thaida peculiaris Karsch 1880 the interme- diate pair arises during ontogeny as a flat out- growth of the more lateral pair (Fig. 3). All these structures are lined with spicules, in- cluding an inter-tracheal canal. In subsequent stages the modified entapophyses are indistin- guishable from the true median tracheae found in Entelegynae. Sicariidae: In Loxosceles laeta (Nicolet 1849) and Sicarius Walckenaer 1847 spp. (from Argentina), there is a median structure homologous with the two fused entapophyses, similar to that found in Drymusa Simon 1891 and Scytodes Latreille 1804, but more elon- gate and thick. I found in both sicariids the expected attachment of the median longitudi- nal muscles that converge on the fused enta- pophyses (Fig. 10). Tetrablemmidae: Platnick et al. (1991) cod- ed the respiratory characters of Caraimatta Lehtinen 1981 according to the description of Brignoliella Shear 1978 given by Forster & Platnick (1985). It seems that they confused the ducts of the female genitalia, or the paired pits of the preanal plate, with tracheae or spi- racles. In Brignoliella cf. carmen Lehtinen 1981 (from New Caledonia), and in Carai- matta cf. cambridgei (Bryant 1940), the only remnant of tracheal system is a median apo- deme (Fig. 6), in agreement with Shear (1978). I also found a similar apodeme in an unidentified Pacullinae from Borneo. Diguetidae and Plectreuridae: A transverse external mark indicates the place where lon- gitudinal muscles attach, on a wide line of the abdominal cuticle (Fig. 7). The entapophyses appear to have lost, in some degree, their function of main site of muscle attachment. In Kibramoa Chamberlin 1924 (Fig. 8) and Plec- treurys Simon 1893 (Fig. 9) the entapophyses are still recognizable as a short median lobe. In Diguetia catamarquensis (Mello-Leitao 1941) and Segestrioides tofo Platnick 1998 the marks on the cuticle are similar to those of Fig. 9, but the median lobe is almost unrec- ognizable. Telemidae: My dissections of Usofila sp. (from California) showed a tracheal pattern like that of Telema Simon 1882, as described by Fage (1913). Ochyroceratidae: Ochyrocera Simon 1891 sp. has two groups of 4-5 tubes each arising from each anterior comer of a characteristic trapezoidal vestibule (Fage 1912: fig. 73), one of them posteriorly directed. In the space be- tween these groups, I found a pair of short RAMIREZ— RESPIRATORY SYSTEM OF HAPLOGYNE SPIDERS 151 Figures 1-10. — Posterior respiratory system and abdominal structures. 1. Liphistius sumatranus Thorell 1890, exuvia of female, detail of 3rd abdominal entapophyses on posterior interpulmonary furrow; 2. Hy- pochilus cf. gertschi Hoffman 1963, female from Virginia, Giles County, posterior respiratory system and spinneret’s bases; 3. Thaida pecuUaris, first free instar, posterior respiratory system and spinneret’s bases; 4. Thaida pecuUaris, subadult male, detail of posterior respiratory system; 5. Ochyrocera sp., female from Minas Gerais, dissected and cleared abdomen, showing median longitudinal muscles and tracheal system; 6. Caraimatta cf. cambridgei, female from Costa Rica, digested abdomen, dorsal view; 7. Diguetia catamar- quensis, female, dissected abdomen, anterior-lateral view, showing insertion of median longitudinal muscles; 8. Kibramoa sp., female from California, 3rd entapophysis and muscle insertion area; 9. Plectreurys sp,, female from Costa Rica, muscle insertion area; 10. Loxosceles laeta, posterior respiratory system showing muscle insertions (lateral tracheae broken). Abbreviations: 3rdEnc = entochondrite at hind end of third median longitudinal muscle; 3rdEnt = third entapophysis; 3rdML = third median longitudinal abdominal muscle; 4thML = fourth median longitudinal abdominal muscle; BL = book lung; ITC == inter-tracheal canal; LT = lateral tracheae. 152 THE JOURNAL OF ARACHNOLOGY entapophyses, where the longitudinal muscles connect (Fig. 5). Archaeidae: The reduced tracheal system of Archaea workmani (O. P. -Cambridge 1881) consists of two separate spiracles each leading to a slender median tracheae, without a trans- verse furrow (Forster & Platnick 1984). I found the apex of these structures widened and fibrose, typical of muscle insertions. CLADISTIC ANALYSIS The present data matrix includes the 43 ter- minals from Platnick et al. (1991), plus Pi- kelinia roigi Ramirez & Grismado 1997 (Fil- istatidae, Prithinae) and a root vector, all scored for 80 characters. The first 67 charac- ters are those used in that paper; only modi- fications and additional characters are listed below. The root vector specifies the states ple- siomorphic for Mygalomorphae and Liphis- tiomorphae. Polymorphisms were used to ex- press variability in the taxa represented by the selected exemplars, and internal steps were added to account for the homoplasy while computing weights. If a representative species does not has a condition known to occur in the family it represents, I followed a strategy similar to that of Platnick et al. (1991), but coding polymorphic entries. Polymorphisms were assigned according to notes in Platnick et al. to characters 23 (in Oecobius Lucas 1846), 36 (in Dysdera Latreille 1804 and Otiothops Macleay 1839), and 65 (in Pholcus Walckenaer 1805), and checked to ensure none required illogical optimizations. Except as noted, all characters were treated as unor- dered. Character 1: Cribellum: present (0); absent (1). Gradungula Forster 1955 and Pianoa For- ster 1987 are coded as 1, although the primi- tive state for the gradungulids should be 0. This coding does not produce an illogical op- timization, as the lost cribellum appears as synapomorphy of both genera. Character 16: Posterior book lungs or modifications: pair of normal book lungs (0); pair of book lungs re- duced to two lamellae (1); pair of lateral tra- cheae (2); absent (3). Filistatines are coded [012] because the homology of their short, flattened lateral structures are unclear (Purcell 1910: 558; Forster et al. 1987: 93). Character 18: Opening(s) of posterior respiratory sys- tem, or position of 3rd abdominal entapo- physes: about midway between anterior book lungs and spinnerets (0); just behind openings of anterior respiratory system (1); just anterior to spinnerets (2). The root is coded as [02] because the openings of posterior book lungs are just anterior to the spinnerets in Liphis- tiomorphae, but separated from them in My- galomorphae. Character 20: Cheliceral gland mound: absent (0); present (1). The putative parallelism in Crassanapis Platnick & Forster 1989 was coded as 1 (Platnick & Forster 1989: fig. 11). Character 32: Posterior spira- cles or origin of 3rd abdominal entapophyses: separate (0); contiguous (1); fused (2). This character expresses the degree of fusion of the formerly bilateral posterior respiratory organs, and is, accordingly, coded as ordered. The po- sition of apodemes serves to discriminate be- tween states in those cases where there is a median transverse furrow, but two interpreta- tions (a wide median spiracle, or two spiracles linked by a furrow) are possible. Diguetia Si- mon 1895 and Segestrioides Keyserling 1883 are coded as uncertain because they lack def- inite cuticular apodemes, and the longitudinal muscles insert on a wide line. Appaleptoneta Platnick 1986 is also coded as uncertain be- cause its respiratory system is unknown, and Leptoneta Simon 1872 has no evidence of apodemes (Lamy 1902: fig. 16). Otiothops is coded [12] because of the variability found in Otiothopinae (Platnick et al. 1999). Character 45: Cribellum: entire (0); divided (1). Gra- dungula and Pianoa Forster 1987 are coded as inapplicable, with the same provisions as in character 1. Gray (1995) noted the curious optimization of the entire cribellum as primi- tive, given that it is homologous with paired anterior median spinnerets. Interestingly, first free instars of Thaida peculiaris show a bi- lobate cribellum, with only one spigot on each side (Fig. 3). Character 67: 3rd abdominal en- tapophyses: short, flat or absent (0); elongate (Fig. 2) (1). I added one internal step to the character because other pholcids lack the en- tapophyses (Lamy 1902). Character 68: Shape of fused 3rd abdominal entapophyses: short, slender (0) (Lamy 1902: fig. 14); elon- gate, broad (1) (Fig. 10). Character 69: Me- dian tracheae: absent (0); present (1). Char- acter 70: Transverse furrow between posterior spiracles: present (0); absent (1). The furrow is present in arachnid outgroups and Liphis- tiomorphae (Fig. 1), but absent in all Myga- lomorphae (e.g., Purcell 1910: 525; Forster et RAMIREZ— RESPIRATORY SYSTEM OF HAPLOGYNE SPIDERS 153 al. 1987: 93). It is coded as present in those groups with a single median spiracle when- ever it is still possible to discern a furrow not lined with spicules. Some authors that over- looked that furrow interpreted the structures as two separate spiracles (e.g., Millidge 1986; revised by Hormiga 1994). Character 71: In- ter-tracheal canal: absent (0); present (1). Scored as uncertain in those terminals without spicules through the tracheal system. Malle- colobus Forster & Platnick 1985 is coded [01], as the canal is present in Orsolobus Si- mon 1893 and Falklandia Forster & Platnick 1985, but absent in Mallecolobus and other orsolobids (Forster & Platnick 1985: 225). The same is true for Segestria Latreille 1804, as the canal is present in Ariadna Audouin 1826 (op. cit.). Character 72: Dysderoid lat- eral tracheae: absent (0); present (1). Each tra- cheal spiracle connects with a broad trunk an- teriorly directed. At its base arises a small trunk that provides tracheoles to the posterior part of the abdomen. Also present in caponiids (Purcell 1910). Character 73: Bunch of pro- somal tracheoles on lateral tracheae: absent (0) ; present (1). Typical of dysderoids and Ca- poniidae. Character 74: Anterior book lungs: present (0); transformed into tracheae (1). Ochyrocera is coded as [01], as Theotima sp. (from Argentina) have tracheae (pers. obs.), but at least some Ochyrocera have lung leaves still recognizable. Character 75: 3rd dorso- ventral abdominal muscles: present (0); absent (1) . Although present in Liphistiomorphae and related arachnids, it is coded as [01] for the root, because some Mygalomorphae (at least) seem to lack these muscles (Acanthogonatus centralis Goloboff 1995, and unidentified Theraphosidae, pers. obs.). Abdonfinal mus- culature was studied in only a few taxa. The muscles were not found in normal dissections of Gradungula sorenseni Forster 1955, Scy- todes sp. (from Buenos Aires), Diguetia ca- tamarquensis, Mecysmauchenius segmentatus Simon 1884 and Otiothops birabeni Mello- Leitao 1945, but these observations must be considered preliminary until more refined techniques are employed. Filistatids were cod- ed according to Ramirez «fe Grismado (1997). All other codings are from Millot (1936). Character 76: Leg autospasy: between coxa and trochanter (0); between patella and tibia (1). Hypochilus Mark 1888 is coded as un- certain, because it lacks definite regions for leg autospasy (Petrunkevitch 1933: 347). Character 77: Excavation between male pal- pal femur and trochanter, into which the em- bolus fits (Ramirez & Grismado 1997): absent (0) ; present (1). Character 78: Three syna- pomorphies for Filistatidae (Gray 1995; Ra- mirez & Grismado 1997): absent (0); present (1) . Character 79: Supra- anal organ: absent (0); present (1). A synapomorphy of Digueti- dae (Lopez 1983; Platnick 1989). Character 80: Bipectinate claws: absent (0); present (1). Coded as [01] in Dysdera because the single row of teeth in dysderids seems to retain trac- es of two rows (Forster & Platnick 1985: 218). Character 81: Proprioceptor bristles on tarsi: absent (0); present (1). A synapomorphy of orsolobids plus at least some oonopids (For- ster & Platnick 1985: 219, 227; Platnick et al. 1991: 67). The data matrix of Table 1 was analyzed under parsimony using implied weights (Go- loboff 1993, 1995), using Pee- Wee version 3.0 (Goloboff 1999). This program assigns lower weight to characters with more homoplasy. In- ternal steps of characters were assigned as im- plied by polymorphic terminals with com- mand ccode=. The same tree of Fig. 11 is found for any value of the constant of con- cavity K {I < K < 6). Under K — 3, 80% of the independent replications of Wagner trees followed by TBR branch swapping (command mult^N;) produces the same optimal tree, thus it is likely an exact solution. The tree is 243 steps long, which is two steps longer than the 20 trees obtained under equal weights with Nona (Goloboff 1999). In these trees, steps are saved in some homoplasious characters (like the anterior median eye loss, and the in- ter-tracheal canal) at expenses of less homo- plasious ones (independent acquisition of re- trolateral tibial apophysis, and reversion to a primitive tapetum). DISCUSSION Forster (1995) discussed the phylogeny of haplogyne spiders proposed by Platnick et al. (1991) in the light of additional characters from the tracheal system. He proposed the group Sicarioidea coincident with Simon’s (1893) Sicariidae, composed by Sicariidae, Scytodidae, Periegopidae, Drymusidae, Plec- treuridae, and Diguetidae, all united by the fu- sion of the third entapophyses. The present analysis that takes into account all characters 154 THE JOURNAL OF ARACHNOLOGY Table 1. — Modifications and additional characters for the data matrix of Platnick et al. (1991). Pikelinia scores as Kukulcania for all characters not shown here, v = [01], w = [012], x = [12], y = [02], ? = unknown, - = inapplicable. Prior weight applied as: character 27 (weight 10), 28(14), 28(2), 51(5), 76(3). Internal steps implied by polymorphisms as: character 23, 32, 33, 39, 65, 67, 74, 80 (1 step); 36, (2 steps); 70 (4 steps); 71 (3 steps). Character 16 18 32 67 70 75 80 root 0 y 0 000 00000 vOOOO 00- Hypochilus 0 0 0 1-0 00000 0-000 000 Ectatostisca 0 0 0 1-0 0?000 10000 007 Gradungula 0 0 0 1-0 00000 10000 00- Pianoa 0 0 0 1-0 0?000 70000 00- Hickmania 0 0 0 1-0 0?000 70000 007 Austrochilus 1 0 0 1-1 01000 71000 001 Thaida 1 0 0 1-1 01000 71000 001 Pikelinia 3 0 0 1-0 1-000 01100 007 Filistata w 0 0 1-0 01000 OHIO 000 Kukulcania w 0 0 1-0 01000 OHIO 000 Scytodes 2 0 2 100 01000 10000 10- Sicarius 3 0 2 110 — 000 10000 00- Drymusa 2 0 2 100 01000 70000 00- Loxosceles 2 0 2 110 01000 10000 00- Diguetia 3 0 p 000 — 000 10001 00- Segestrioides 3 0 7 000 — 000 70001 00- Plectreurys 3 0 2 000 — 000 70000 00- Kibramoa 3 0 2 000 — 000 70000 00- Pholcus 3 0 1 000 — 000 10100 00- Caraimatta 3 0 2 100 — 000 70000 00- Nops 2 1 0 000 01111 70000 00- Ochyrocera 2 0 1 000 O-OOv 70000 00- Segestria 2 1 0 000 vvlio 10000 00- Dysdera 2 1 0 000 10110 10000 vO- Mallecolobus 2 1 0 000 vvlio 70000 11- Dysderina 2 1 0 000 OHIO 70000 11- Appaleptoneta 2 0 7 7 - 7 0?000 71000 00- Usofila 2 0 0 7-7 10000 70000 00- Archaea 3 0 0 1-1 10000 70000 00- Mecysmauchenius 2 0 1 1-1 01000 10000 00- Trice Hina 2 2 1 1-1 01001 70000 00- Huttonia 2 0 2 111 01000 70000 00- Othiotops 2 0 X 000 01000 10000 00- Waitkera 2 0 1 1-1 01000 00000 001 Tetragnatha 2 2 1 1-1 01000 10000 00- Crassanapis 2 2 1 1-1 01000 70000 00- Oecobius 2 2 1 1-1 01000 00000 001 Stegodyphus 2 2 1 1-1 01000 00000 001 Deinopis 2 2 1 1-1 01000 70000 001 Dictyna 2 0 1 1-1 01000 00000 001 Callobius 2 2 1 1-1 01000 70000 001 Araneus 2 2 1 1-1 01000 00000 00- Mimetus 2 2 1 1-1 01000 70000 00- Pararchaea 2 2 1 1-1 01000 70000 00- from both sources (but revises some obser- vations), yields intermediate results. In agree- ment with Forster’s hypothesis, my analysis retrieves a monophyletic group with fused en- tapophyses, but including Tetrablemmidae, af- ter the re-examination of their tracheal system. However, the placement of Pholcidae coin- cides with that of Platnick et al. 1991. It must RAMIREZ— RESPIRATORY SYSTEM OF HAPLOGYNE SPIDERS 155 (7.5) FUistata Kukulcania Pikelinia Nops Dysdera Segestria Mallecolobus Dysderina Usofiia Appaleptoneta Ochyrocera Sicarius Loxosceles Scytodes Drymusa Caraimatta Pholcus Diguetia Segestrioides Plectreurys Kibramoa Figures 1 1 . — Optimal cladogram for the representative taxa. Bremer support in terms of Fit are given on each node. be noticed that the differences between my re- sults and those of Platnick et al. involve groups with relatively low Bremer support (Bremer 1994; values on Fig. 11), which might be the most prone to change should new characters (e.g., from female genitalia) or rep- resentatives (e.g., from Pacullinae and Theo- timinae) be added. The elongate entapophyses (char. 67) are a synapomorphy of Araneomorphae, with a sub- sequent reversion in the Haplogynae other than filistatids (node 6), and regain in Scyto- didae, Sicariidae and Tetrablemminae as a central, fused element (see below). Confirm- ing the hypothesis of Purcell (1909), the short apodemes of Segestriidae (and their relatives) are reduced entapophyses rather than reduced median tracheae. As supposed by the same au- thor, the loss of the transverse furrow (char. 70) is a synapomorphy of the suborder My- galomorphae, with parallelisms in some iso- lated araneomorph groups. Although homo- plasy seems to be rampant in this character, no parallel gains of a transverse furrow have been mapped. The inter-tracheal canal appears in Araneoclada or Neocribellatae (ambiguous optimization), and is independently lost in several araneocladan clades. Lateral tracheae (char. 16 -state 2) are a synapomorphy of Ar- aneoclada (node 2), whereas the reduction of posterior book lungs to two pulmonary leaves (char. 16-1) is a synapomorphy of Austrochil- inae. Within the Haplogynae, filistatines (node 5) were repeatedly described as having some rel- ict of book lungs instead of lateral tracheae. Because the optimization of the character gives state 2 at the base of Filistatinae, the congruence criterion suggests that these struc- tures are homologous with lateral tracheae. The 3rd dorsoventral abdominal muscles (char. 75) have been lost several times in this tree, but were never found in haplogynes other than Filistatidae. The loss of lateral tracheae (char. 16-3) is a synapomorphy of node 17, with parallelism at least in Prithinae {Pikelinia Mello-Leotao), Sicarius, and dictynids. The advanced spiracles (char. 18-1) are a synapo- morphy of caponiids {Nops MacLeay 1838) and Dysderoidea (node 8), but the placement of Tetrablemmidae {Caraimatta) is different from that of Platnick et al. because of the re- examination of the tracheal system of tetra- blemmids. The fused entapophyses (char. 32- 2) are a synapomorphy of node 13 plus Periegopidae; this last group was not included here but seems to be the undisputed sister group of Scytodidae (Forster 1995). For this data matrix there is a reversion to state 1 in Pholcus, but conditions in other pholcids range from a pair of contiguous entapophyses linked by a furrow, to the smooth concave cu- 156 THE JOURNAL OF ARACHNOLOGY tide serving directly as the site for muscle attachment. Further elongation of the fused entapophyses is a synapomorphy of Sicari- idae. All book lung reductions (char. 74) have independent origin for this data set. Three characters of the respiratory system are synapomorphies of Entelegynae: The first is the extreme posterior displacement of the spiracle (char. 18-2), with homoplasy in sev- eral palpimanoids, dictynids, Uloboridae, and many derivative groups not included in the analysis. The second is the contiguous median tracheae (homologous with 3rd entapophyses, char. 32-1), although the same state appears to arise convergently (but without true median tracheae) in Ochyrocera and Pholcus. The third is true median tracheae (char. 69), with a notable convergence in Austrochilinae. A scenario of the morphological transfor- mations leading to the median tracheae can be traced by optimizing characters on the phy- logeny. Basal spiders (and closer outgroups) have hollowed thick entapophyses, arising from an interpulmonary furrow. The entapo- physes elongated in Araneomorphae. The spicules typical of respiratory cuticles extend- ed from posterior book lungs (in an ancestor of the Neocribellatae) or from lateral tracheae (in an ancestor of Araneoclada) to line the fur- row, forming an inter- tracheal canal. At the same time, or later in some ancestor of the Entelegynae, the spicules lined also the inte- rior surface of entapophyses, that became elongated and slender, with thin cuticle, form- ing the median tracheae. This transformations series was hypothesized by Purcell as early as 1909. ACKNOWLEDGMENTS Helpful comments on versions of the man- uscript were provided by Jonathan Codding- ton, Pablo Goloboff, Marfa Elena Galiano, Brent Opell, Norman Platnick, Jim Berry, and two anonymous reviewers. Charles Griswold, N. Platnick and J. Coddington provided spec- imens for this study. Support for this project was provided by a graduate fellowship and EXO085 fund from the Universidad de Buen- os Aires, a Short-Term Visitor Award from the Smithsonian Institution, and Collection Study Grants from the American Museum of Natural History and the California Academy of Sci- ences. LITERATURE CITED Bertkau, P. 1872. Ueber die Respirationsorgane der Araneen. Arch. Naturg., 38(l):208-233. Bertkau, P. 1878. Versuch einer natiirlichen Anord- nung der Spinnen, nebst Bemerkungen zu ein- zelnen Gattungen. Arch. Naturg., 44:351-410. Bremer, K. 1994. Branch support and tree stability. Cladistics, 10:295-304. Page, L. 1912. Etudes sur les Araignees cavemi- coles. 1. Revision des Ochyroceratidae (n. fam.). Biospelogica, XXV, Arch. Zool. Exper., 10(5): 97-162. Page, L. 1913. Etudes sur les Araignees cavemi- coles. II. Revision des Leptonetidae. Biospelo- gica, XXIX, Arch. Zool. Exper., 10(5):479-576. Forster, R.R. 1970. The spiders of New Zealand, Part III: Desidae, Dictynidae, Haniidae, Amau- robioididae, Nicodamidae. Otago Mus. Zool. Bull., 3:1-184. Forster, R.R. 1995. The Australasian spider family Periegopidae Simon, 1893 (Araneae: Sicarioi- dea). Rec. Western Australian Mus., 52:91-105. Forster, R.R. & N.I. Platnick. 1984. A review of the archaeid spiders and their relatives, with notes on the limits of the superfamily Palpima- noidea (Arachnida, Araneae). Bull. American Mus. Nat. Hist., 178(1): 1-106. Forster, R.R. & N.I. Platnick. 1985. A review of the austral spider family Orsolobidae (Arachnida, Araneae), with notes on the superfamily Dysde- roidea. Bull. American Mus. Nat. Hist., 181:1- 230. Forster, R.R., N.I. Platnick & M.R. Gray. 1987. A review of the spider superfamilies Hypochiloidea and Austrochiloidea (Araneae, Araneomorphae). Bull. American Mus. Nat. Hist., 185(1): 1-1 16. Goloboff, P.A. 1993. Estimating character weights during tree search. Cladistics, 9:83-91. Goloboff, P.A. 1995. Parsimony and weighting: a reply to Turner and Zandee. Cladistics, 11:91- 104. Goloboff, P.A. 1999. Pee-Wee and Nona, versions 3.0, programs and documentation. . Gray, M.R. 1995. Morphology and relationships within the spider family Filistatidae (Araneae: Araneomorphae). Rec. Western Australian Mus. SuppL, 52:79-89. Hormiga, G. 1994. Cladistics and the comparative morphology of linyphiid spiders and their rela- tives (Araneae, Araneoidea, Linyphiidae). Zool. J. Linn. Soc., 111:1-71. Kastner, A. 1929. Bau und Funktion der Facher- tracheen einiger Spinnen. Zeits. Morph. Okol. Tiere, 13(3-4):463-557. Lamy, E. 1902. Recherches anatomiques sur les trachees des Araignees. Ann. Sci. Nat. Zool., (8)15:149-280. RAMIREZ— RESPIRATORY SYSTEM OF HAPLOGYNE SPIDERS 157 Levi, H.W. 1967. Adaptation of respiratory system of spiders. Evolution, 21:571-583. Lopez, A, 1983. Some observations on the internal anatomy of Diguetia canities (McCook, 1890) (Araneae, Diguetidae). J. ArachnoL, 11(3):377- 384. Millidge, A.F. 1986. A revision of the tracheal structures of the Linyphiidae (Araneae). Bull. British ArachnoL Soc., 7(2):57-6L Millot, J. 1936. Metamerisation et musculature ab- dominale chez les Araneomorphes. Bull. Soc. Zool. France, 61:181-204. Petrunkevitch, A. 1933. An inquiry into the natural classsification of spiders, based on a study of their internal anatomy. Trans. Connecticut Acad. Arts. Sci., 31:303-389. Platnick, N.I. 1989. A revision of the spider genus Segestrioides (Araneae, Diguetidae). American Mus. Nov., 2940:1-9. Platnick, N.I., J.A. Coddington, R.R. Forster & C.E. Griswold. 1991. Spinneret morphology and the phytogeny of haplogyne spiders (Araneae, Ara- neomorphae). American Mus. Nov., 3016:1-73. Platnick, N.I. & R.R. Forster. 1989. A revision of the South American and Australasian spiders of the family Anapidae (Araneae, Araneoidea). Bull. American Mus. Nat. Hist., 190:1-139. Platnick, N.I., C.J. Grismado & M.J. Ramirez. 1999. On the genera of the spider subfamily Otiothopinae. American Mus. Nov., 3257:1-25. Purcell, W.E 1909. Development and origin of re- spiratory organs in Araneae. Quart. J. Microsc, Sc. (N.S.), 54:1-110. Purcell, W.E 1910. The phylogeny of trachea in Araneae. Quart. J. Microsc. Sci., 54:519-563. Ramirez, M.J. & C.J. Grismado. 1997. A review of the spider family Filistatidae in Argentina (Arachnida, Araneae), with a cladistic reanalysis of filisitatid genera. Entomol. Scandinavica, 28: 319-349. Shear, W.A. 1978. Taxonomic notes on the ar- mored spiders of the families Tetrablemmidae and Pacullidae. American Mus. Nov., 2650: 1-46. Simon, E. 1893. Histoire Naturelle des Araignees. Roret, Paris, l(2):257-488. Manuscript received 11 May 1999, revised 4 Oc- tober 1999. 2000. The Journal of Arachnology 28:158-168 EFFECTS OF CLIMATE AND PREY AVAILABILITY ON FORAGING IN A SOCIAL SPIDER, STEGODYPHUS MIMOSARUM (ARANEAE, ERESIDAE) T.E. Crouch: Department of Zoology and Entomology, University of Natal, Pietermaritzburg, South Africa and Durban Natural Science Museum, RO. Box 4085, Durban 4000, South Africa. Y. Lubin: Mitrani Department of Desert Ecology, Blaustein Institute for Desert Research, Ben Gurion University of the Negev, Sede Boqer Campus, 84990 Israel ABSTRACT. Tropical areas with favorable climatic conditions, high prey availability and large prey size are assumed to favor sociality in spiders. Notwithstanding, the three social species of Stegodyphus (Eresidae) inhabit arid and semi-arid habitats with marked daily and seasonal variation in climate. The nests of the social spider Stegodyphus mimosarum Pavesi commonly occur in dry Acacia savanna in southern Africa. We investigated the abiotic conditions to which the nests of S. mimosarum are exposed and the changes in availability of potential insect prey at different times of year and over the daily cycle. We used these data to determine the extent to which prey availability and climatic conditions explain seasonal and daily variation in the activity of the spiders. Data were collected during four sampling periods a year over two years from nests of S. mimosarum located on the Mkomazi River Bridge (KwaZulu-Natal, South Africa). We measured ambient and nest temperatures and in a sample of nests, spider growth rate, prey availability, foraging activity and activity on the web at night. Spiders had two periods of increased growth rate occurring in early and late summer, at times of year when ambient temperature rarely falls below 20 °C. Temperatures inside the nest were generally higher than ambient throughout the day and night. Foraging response, measured as the numbers of individuals responding to the vibrations of a tuning fork, was significantly higher by night than by day. In summer, foraging response decreased with increasing temperature during the day, whereas in winter, there was a positive correlation between foraging response and temperature at night. Potential prey, measured as mean numbers of insects trapped in a sample of webs, were more abundant during the day than at night, despite the fact that the spiders were most active on the web at night. Nocturnal insects, however, were larger than diurnal ones and spiders handled sig- nificantly more large prey both during the day and at night. Correlation and partial correlation analyses indicate that ambient temperature and windspeed play a direct role in influencing foraging and other activity on the web. Nonetheless, the predominance of nocturnal activity in both summer and winter could not be explained by climatic conditions and prey availability alone. Some other factor (e.g., predation or parasitism) may be involved. Keywords: Climate, prey availability, foraging, social spider Most of the 18 or so known species of so- cial spiders (also referred to as cooperatively group-living or permanently social) are trop- ical, and most are found in the wet tropics (D’ Andrea 1987; Aviles 1997). Sociality may occur with greater frequency in the tropics be- cause the benign climate allows activity to be maintained year-round, and thus a colony can be maintained continuously over several gen- erations, or because potential insect prey are available year-round, also allowing continu- ous activity (Riechert 1995). Additionally, large insects, which can be captured more ef- ficiently by a group of spiders than by solitary spiders of a similar size (Nentwig 1985), are more abundant in the tropics (Rypstra 1990). Notwithstanding, the three social species of Stegodyphus (Eresidae) (and indeed, most of the remaining 17 solitary species of the genus) are sub-tropical and live in arid, semi-arid and seasonally wet savannas of Africa and the In- dian subcontinent (Kraus & Kraus 1988). The two African species, S. dumicola and S. mi- mosarum, occur largely in dry thombush (Acacia) savanna, where sununer tempera- tures are high and winter is generally cold, 158 CROUCH & LUBIN— FORAGING IN STEGODYPHUS MIMOSARUM 159 with little plant growth or insect activity. Ste- godyphus colonies have a strongly seasonal developmental cycle, which is linked to the local seasonal regime (Seibt & Wickler 1988). Consequently, we expect to find a strong cor- relation between variation in the physical and biotic environment and both daily and season- al activity of spiders in Stegodyphus colonies. We investigated the abiotic conditions to which the nests of S. mimosarum are exposed at different times of year and the changes in availability of potential insect prey. We used these data to examine the hypothesis that prey availability and abiotic conditions explain sea- sonal and daily variation in the activity of these spiders. METHODS Natural history and study area.™ Nests of S. mimosarum often occur near water, in the canopy of thorny acacia and other trees, as well as on man-made structures such as utility poles, road signs, fences and bridges (Kraus & Kraus 1988; Seibt & Wickler 1988; pers. obs.). Our study population consisted of nests that occupied the railing of the Mkomazi River bridge, 23 km west of Richmond in KwaZulu-Natal (29°54'3r'S, 30°05'35''E). The bridge spans 66 m and is 8 m wide. The 354 vertical aluminum struts on each side of the bridge support a horizontal handrail at a height 1 m above the ground. The nests oc- cupy the underside of the railing between the vertical struts along both sides of the bridge. At the start of the study in January 1995, there were 615 nests on both sides of the bridge combined. At this time these nests were at most nine years old, as the 1987 floods de- stroyed the railings together with any nests. Nests occurred also in the canopies of trees downstream from this site and on trees grow- ing on pylons below the bridge. These latter nests were difficult to access. The annual rainfall for the area for 1994 was 630 mm, 1041 mm for 1995 and 1112 mm for 1996. Most rain fell in summer (Oc- tober-February). Summer temperatures regu- larly exceeded 35 °C and during winter dropped below 0 °C. Data collection.^ — The study was conduct- ed from January 1995 to November 1996. We measured body size (length from the tip of the prosoma to the tip of the abdomen) of indi- viduals from 20-30 randomly selected nests at different times of year. Measurements of abiotic and biotic factors were conducted over a 3 -day period, once every 4 months (Febru- ary, May, August, November). Forty to sixty nests were randomly selected for each obser- vation period (20-30 nests from each side of the bridge). We measured nest, web and am- bient temperatures, windspeed, spider activity and prey availability. Diurnal data were col- lected over three days during each month sam- pled in 1995. Nocturnal data were collected during two nights each in May and August 1996, and from a single night each during No- vember and February (1996). Time of day is local time (GMT +2 hours). Nest conditions: Temperatures were mea- sured from a single nest on the south side of the bridge. Measurements were taken inside the nest, about 2 cm below the surface and 3 cm below the nest in the capture web on the north and south sides. These measurements were taken using copper-constantan thermo- couples. A temperature probe and anemome- ter were placed at nest height to measure am- bient temperature and windspeed respectively. Temperatures and windspeeds were recorded at three five-minute intervals every hour by an MCS 120-02 datalogger (M C Systems, Steenberg, Cape Town, South Africa) and hourly means were then calculated and sum- marized separately for day and night periods. Prey availability: Nests were surveyed at two-hour intervals for new prey items that were either trapped on the web or were being handled by the spiders. The numbers of spi- ders handling the prey, prey size (mm) and identity to order level were noted. Foraging response: Foraging response was assessed as the number of spiders responding to the vibrations emitted from a musical tun- ing fork (440 Hz) which are similar to vibra- tions produced by buzzing insects trapped in the web (Henschel et al. 1992). The stimulus was applied to the capture web 4 cm below the nest. The number of spiders emerging from the nest or approaching the vibrating tuning fork within 5 seconds was counted at two-hour intervals throughout the observation period. This behavior provided a relative mea- sure of the readiness of spiders to attack prey caught in the web and allowed us to compare the spider’s response to a standardized stim- ulus under different ambient conditions. Activity on the web at night: Spiders that 160 THE JOURNAL OF ARACHNOLOGY emerged from the nest at night were observed under red light. Activities on the nest surface and on the capture web included construction (spinning), maintenance (the removal of old prey and silk) and prey capture. In addition, some individuals were stationary on the nest surface or on the web. The number of spiders on the web and nest surface was recorded at two-hour intervals prior to measuring foraging effort with the tuning fork. Statistical analysis. — Temperatures mea- sured at different locations (inside the nest, on the web, ambient) were compared using paired r-tests (two-tailed) on the mean hourly temperatures for the three days or two nights of each sample period. Data for each month were tested separately both here and in all oth- er comparisons. As the same null hypothesis was being tested on each of the three days sampled (e.g., nest temperature did not differ from ambient), a Bonferroni correction was used and the acceptable level of significance (P = 0.05) was divided by k, the number of non-independent tests (k = 3 and 2 for diurnal and nocturnal data, respectively) (Haccou & Meelis 1994). Chi-squared tests for indepen- dence were conducted for prey data where the variables included in the analysis were the type of prey (order), prey size classes and the number of spiders handling prey (Zar 1984). Relationships between the variables (ambi- ent temperature, windspeed, prey, foraging re- sponse and spider activity) were tested with Pearson’s product-moment correlation and Spearman’s rank correlation. Partial correla- tion analysis (Zar 1984) was used to deter- mine the correlation between any two vari- ables while maintaining all others constant. Data on prey availability and foraging ac- tivity required logarithmic transformation pri- or to analysis and values were replaced by log (x + 1) (Elliot 1983). RESULTS Growth rate and seasonal development of spiders. — Stegodyphus mimosarum ap- peared as juveniles in February, were sub- adult from October to December, and reached maturity in summer from December to Feb- ruary when mating and egg laying took place (Fig. 1). Little growth occurred during the winter months (May to August). In 1996, in- dividual growth during the winter months (May- July) was less than 5% per month (av- eraging 1. 7-4.3%, in body length), whereas in I summer spiders grew 13.3-16.6% per month in body length. Seasonal changes in temperature and windspeed. — Diurnal conditions: During |! February, which is late summer and the hot- test month sampled, the temperature inside the nest was on average 2.5 °C higher than am- bient (t = 4.64, P < 0.001) (Fig. 2). During February a maximum of 41.7 °C was recorded within the nest and 36.6 °C for ambient tem- perature. Temperatures below the nest, on the jj south and north sides of the railing, were not jj significantly different from ambient. However, they were on average 3.0 °C (r = 5.65, P < I! 0.001) and 2.8 °C (t = 5.17, P < 0.0001) re- spectively, lower than temperatures within the nest. At high ambient temperatures (± 30-35 ; °C), spiders were observed sitting below the nest in a layer of loose silk on the nest surface, : as well as just inside the nest entrances. j| Throughout the remainder of the year (i.e., j May, August and November), temperatures I measured inside the nest did not differ signif- ji icantly from ambient, nor from those on the ; north and south side of the nest. However, j, temperatures within the nest were generally } above those on the web. Maxima recorded for !, ambient and nest temperature were 33.2 °C jj and 37.7 °C, respectively for May, and 32.2 || °C and 32.8 °C for November. Windspeed was || highest during the summer months of Novem- |1 ber and February (Fig. 2). Nocturnal conditions: During February and August nighttime temperatures within the nest were only slightly, but significantly higher than ambient (Fig. 2). Temperatures inside the nest were on average 1.0 °C higher than am- bient {t = 2.62, P = 0.01) in February and 0.9 °C in August {t = 2.50, P = 0.013). In winter (May) and early summer (November) nest and ambient temperatures showed similar ! patterns, but they were not significantly dif- ferent. Nest temperatures were at a minimum of 7.1 °C in August, when minimum ambient was 6.1 °C, and at a maximum of 27.7 °C in ' February, when maximum ambient reached 27 °C. Nighttime windspeed was highest in the : summer months (November-February, Fig. 2). ‘I Daily changes in temperature and wind- I speed. — Mean ambient and nest temperatures ii reach a maximum between 1200-1400 h and were at a minimum before sunrise (Fig. 3). ‘ CROUCH & LUBIN— FORAGING IN STEGODYPHUS MIMOSARUM 161 JUVENILE SUBADULT ADULT Figure 1. — Mean and 95% confidence intervals (dots above bars) for body length of Stegodyphus mimosarum and the period of occurrence of different life stages on the bridge. During winter (May-August) the temperature differences between day and night were ex- treme, resulting in rapid loss of heat from the nest from midday to sunset, and rapid heat gain from sunrise to midday (Fig. 3). Nest and ambient temperatures in winter (August) re- mained well below 25 °C throughout the day. During the hottest month (February), temper- atures inside the nest were above 30 °C from 1000-1400 h. Daytime windspeeds peaked at 1600 h and were both greater and more variable than those measured at night. Wind and tempera- ture were not significantly correlated, apart from nighttime records in February and May (Spearman rank correlation, R = 0.465, P = 0.004, n = 36 andR = 0.374, P = 0.009, n = 75, respectively). Prey Availability.- — Prey numbers: Throughout the year the greatest numbers of prey were found on the web between 0800- 1000 h; the greatest numbers of prey per web were in the summer months (May-No vember; Fig. 4). With the exception of November, very few prey items were observed in the webs during the night (1900-0500 h), despite the fact that the spiders were most active at this time (see below). In all months, more insects were trapped in S. mimosarum webs (= avail- able prey) during the day than at night. A sin- gle insect was trapped at night in a survey of approximately 50 webs in each of the sam- pling months of February and August (late summer and late winter, respectively). In May (early winter) and November (early summer) respectively, 4% and 12% of the insects trapped were nocturnal. While these figures represent insects that landed on the web and were available to spiders, not all of these in- sects were actually captured by the spiders (Table 1). There was no significant difference between the numbers of insects actually han- dled by day and night. However, proportion- ally more nocturnal insects were handled (83% and 78% of trapped insects in May and November, respectively), whereas only 5-25% 162 THE JOURNAL OF ARACHNOLOGY Figure 2. — Mean windspeed (km/h), ambient temperature (°C) and temperatures inside and below the nest on the north and south side. of diurnal insects were handled by the spiders (Table 1). We compared the sizes of insects trapped in the webs with those of prey actually handled by the spiders, and similarly, the types of in- sects trapped and handled. Because of small sample sizes in some months, we pooled the data from all sampling dates for the statistical analyses. Prey size: Most insects trapped in the cap- ture webs were < 3 mm in body length (Table 1). Despite their availability, insects of this size class were rarely handled by the spiders. Medium sized prey (3-6 mm) and larger in- sects (> 6 mm) were handled significantly more often than expected from their abun- dance in the webs by day (all seasons com- bined, x" = 164.8, df 3, P = 0.0001). A large proportion of the prey available at night was greater than 3 mm in length and there was no significant difference between the size of prey available and those handled at night (all seasons combined, ~ 1.32, df = 2, P > 0.05). Prey type: The prey taxa available changed throughout the year. Diptera were common in all samples, Ephemeroptera were most com- mon in February; Hemiptera and Coleoptera in May and November, and Hymenoptera in November (Table 2). There were significant differences in the distribution of major taxa available in the web and those handled by the spiders during the day (all seasons combined, X^ = 32.98, df = 4, P = 0.001). By day, more Coleoptera and Diptera, and fewer Hemiptera, were handled than expected. Noctumally, prey taxa available in the web and those handled by spiders did not differ statistically. Three- Table 1 . — The distribution of size classes of prey available to the spiders (insects trapped in webs) and the corresponding percentage handled by them. February May August November Size classes Day Night Day Night Day Night Day Night Prey available in webs (%) <1 to 3 mm 58.1 0 64 33.3 37.2 100 80.9 7.3 3.1 to 6 mm 19.3 100 12.8 50.1 37.6 0 16 21.8 6.1 to 15 mm 19.3 0 20.5 8.3 25.5 0 2.7 56.4 >15 mm 3.2 0 2.7 8.3 0 0 0.4 14.5 Total prey available 31 1 326 12 43 1 406 55 Prey handled (%) <1 to 3 mm 0 0 0.6 25.1 0 100 0.7 0 3.1 to 6 mm 3.2 100 2.5 41.6 2.3 0 0 14.5 6.1 to 15 mm 16.1 0 8.8 8.3 2.3 0 0.7 49.18 >15 mm 3.2 0 2.5 8.3 0 0 0 14.5 Total prey available 7 1 47 10 2 1 6 43 Prey handled as % of total prey available 22.5 100 14.4 83.3 4.6 100 1.4 78.2 CROUCH & LUBIN— FORAGING IN STEGODYPHUS MIMOSARUM 163 MAY NOVEMBER Time Figure 4. — Mean diurnal and nocturnal numbers of insects trapped in webs (available prey) ± SD. Figure 3. — Mean diurnal and nocturnal ambient temperature, nest temperature (°C) and windspeed (km/h). way contingency tables (partial independence) tested for interactions between prey size and type for all seasons. We found for both day and night there was a lack of independence between prey size and type in influencing whether the prey was handled (day: ~ 197.6, df=H,P = 0.0001; night: x" = 15.58, df=6, P = 0.016). Foraging response.— The response of spi- ders to a prey stimulus (tuning fork) was greater at night than during the day: February, t = —17.3, P = 0.001; August, t = —4.2, P = 0.012 and November, /=— 10.3,P = 0.001 (Fig. 5). For all of these P < 0.013, the Bon- ferroni- adjusted level of alpha. Diurnal for- 164 THE JOURNAL OF ARACHNOLOGY Table 2. — The distribution of taxa of prey available to the spiders (insects trapped in webs) and the corresponding percentage handled by them. February May August November Prey type Day Night Day Night Day Night Day Night Prey available in webs (%) Coleoptera 3.2 100 16.8 16.6 0 0 3.2 47.7 Diptera 29 0 49.6 50 76.7 100 13.3 10.9 Hemiptera 22.5 0 26.9 0 4.6 0 38.4 3.6 Hymenoptera 9.6 0 3.3 25 0 0 22.1 34.5 Other 35.7 0 4.3 8.4 18.7 0 23 3.3 Total prey available 31 1 326 12 43 1 406 55 Prey handled (%) Coleoptera 3.2 100 2.7 8.3 0 0 0.2 38.1 Diptera 9.6 0 9.8 41.6 4.6 100 0.2 3.6 Hemiptera 0 0 1.2 0 0 0 0.4 1.8 Hymenoptera 9.6 0 0.3 25 0 0 0 32.7 Other 0 0 0.3 8.3 0 0 0 1.8 Total prey handled 7 1 47 10 2 1 6 43 aging response was higher in May than other months and spider activity was not signifi- cantly different by night and day {t = —3.62, P > 0.013). With the exception of February, foraging response decreased in the second half of the night, from about 0100 h. A similar pattern was observed when we used the pro- portion of nests in which spiders responded rather than mean number of spiders respond- ing. The diurnal response levels varied con- siderably, with peaks occurring at different times of the day throughout the year (Fig. 5). February, May and August had higher re- sponse levels than November. This is at least in part attributable to the presence of young spiders in the nests during this period, where- as in November most spiders were subadult or adult; and the colonies contained fewer indi- viduals owing to mortality during the growth phase. Nocturnal activity on the web. — Shortly after sunset, spiders emerged from the nest and dispersed over the nest surface and cap- ture web where they engaged in web cleaning, construction, or were motionless on the nest or web. The numbers emerging from the nests were highest between 0300-0400 h in Feb- ruary and November (summer) and between 1900-2100 h during the winter months (May and August) (Fig. 6). All spiders returned to the nest shortly before sunrise. Relationships between activity, prey availability and abiotic factors. — Foraging response was correlated with climatic vari- ables and prey availability in some instances and not in others (Table 3). By day foraging was negatively related to windspeed in all four sampling periods, with the probability of 0.5^ = 0.063 of a negative relationship occurring by chance alone in all four samples. There was no significant correlation between forag- ing response and windspeed at night. Daytime foraging response was negatively correlated with ambient temperature in November (sum- mer), while at night foraging response showed a strong positive correlation with ambient temperature in August, which was the coldest month. Prey availability and foraging re- sponse were significantly positively correlated only in February (daytime sample), however all 6 correlation coefficients were positive, with a probability of this occurring by chance alone of 0.5^ == 0.016. Partial correlation analysis allowed for the comparison of two variables whilst holding constant the influence of other variables on the two in question. These results show a similar pattern to that obtained for the simple corre- lation (Table 3). Spider activity on the web at night was positively correlated with ambient temperature in August, as was the foraging re- sponse at night, and foraging response and ac- tivity were strongly positively correlated. Dur- ing August the partial correlation coefficients for both foraging response and spider activity with ambient temperature were positive and CROUCH & LUBIN— FORAGING IN STEGODYPHUS MIMOSARUM 165 Figure 5. — Mean diurnal and nocturnal spider foraging activity (number of spiders approaching vibrating tuning fork in 5 seconds) ± SD. significant, suggesting that nighttime activity is strongly dependent on ambient temperature during the cool season. DISCUSSION Seasonality of spider growth. — The sea- sonality of growth and the range of spider siz- es observed here was similar to those of col- onies observed in other parts of KwaZulu-Natal (unpubl. data) and by Seibt & Wickler (1988). There was little spider growth in winter (May, August), when very small young were present in the nest and after the females had died. Growth to maturation and egg-laying occurred in the summer months. FEBRUARY p 1-4 I 1.2 © r 1 I I g 0.6 ;| 0.4 ! 0.2 (9 I 0 MAY 23 i AUGUST Figure 6. — Mean number of spiders on the web at night ± SD. The difference in the growth rate of spiders during winter and summer months corre- sponds to the initial slow increase and the ex- ponential phase, respectively, of a typical sig- moid growth curve. The food requirements of a colony are expected to be greatest during the period of exponential growth of juveniles, i.e., during early summer. Consequently, condi- tions should be more favorable for growth in the summer months. This was largely the case for the abiotic conditions as well as the avail- ability of prey. Abiotic conditions: In May and August, mean nest temperatures were below 25 °C during the day and less than 15 °C at night. February was the hottest month and in both 166 THE JOURNAL OF ARACHNOLOGY Table 3. — Correlations (Pearson's) between mean number of spiders responding to prey stimulus (for- aging), insects trapped on 40-60 webs (prey), nocturnal activity on the web (On web), wind and ambient temperature (Tan.^)- the following symbols have been used; only one prey item was recorded in night samples and therefore omitted from the analysis (*), appropriate for nocturnal data only (**), activity on the web was not included in the analysis of foraging response (***)^ significant partial correlations co- efficient (P < 0.05)(§). Correlation coefficient (P-value) Wind X ^ amb Prey On Web** February Day Foraging -0.054 (NS) -0.387 (NS) 0.652§ — {n = 15) Prey -0.040 (NS) -0.506§ (P < 0.01) _ Night* Foraging -0.824§ (P < 0.05) -0.479 (NS) 0.526 (NS) (« = 6) On Web {P < 0.05) -0.503 (NS) -0.371 (NS) — — May Day Foraging -0.467§ 0.161 (NS) 0.046 (NS) — {n = 14) Prey (P = 0.09) -0.253 (NS) -0.472 (NS) _ _ Night Foraging 0.356 (NS) 0.497 (NS) 0.493 (NS) 0.493 (NS) {n = 12) Prey 0.215 (NS) 0.429 (NS) — — On Web 0.438 (NS) 0.267 (NS) 0.348 (NS) — August Day Foraging -0.036 (NS) 0.050 (NS) 0.077 (NS) — {n = 15) Prey -0.182 (NS) -0.188 (NS) — — Night* Foraging 0.421 (NS) 0.955§ — 0.898 {n = 12) On Web 0.476 (NS) (P < 0.001) 0.961§ (P < 0.001) (P < 0.001) November Day Foraging -0.529§ -0.452§ 0.326 (NS) — {n = 18) Prey (P < 0.05) -0.349 (NS) (P < 0.05) 0.074 (NS) _ _ Night Foraging*** 0.098 (NS) 0.516 (NS) 0.472 (NS) — {n = 6) Prey 0.294 (NS) 0.829§ — — On Web 0.095 (NS) (P < 0.05) 0.431 (NS) 0.399 (NS) — February and November nighttime tempera- tures rarely fell below 20 °C. Although the nests were positioned on an exposed bridge, strong winds were not recorded during our ob- servation periods. The windspeed was gener- ally lower in the winter months of May and August and higher during November, both during the day and night. Nest and ambient temperatures peaked between 1200-1400 h and were lowest just before dawn. While the difference between day and nighttime nest temperatures was often > 15 °C during the winter months, in November there was Mttle difference between the maximum and mini- mum nest temperatures recorded (±5 °C). Temperatures inside the nest were nearly al- ways higher than ambient, as found also by Seibt & Wickler (1990). Thus, in the summer months, spiders inside the nest might suffer excessive heat loads during mid-day, but they can cool convectively by moving out of the nest. Convective cooling may be enhanced by the prevalence of stronger afternoon winds during the summer months. Seibt & Wickler (1990) showed that S. mimosarum actively avoided temperatures above 41 °C. During CROUCH & LUBIN— FORAGING IN STEGODYPHUS MIMOSARUM 167 one hot day in December 1997, when ambient temperature exceeded 42 °C at 0900 h, we ob- served spiders moving onto the web into the shadow cast by the nest, and some females moved their egg sacs onto the web as well. This behavior was observed frequently in the social S. dumicola (Seibt & Wickler 1990; pers. obs.). A similar response to high mid- day nest temperatures occurs in the solitary Stegodyphus lineatus (Henschel et al. 1992) and in a widow spider Latrodectus revivienis (Lubin et al. 1993), both web-building species of desert habitats. In all of these cases, the silken structure of the nest does not protect the spiders from high daytime temperatures (see also Seibt & Wickler 1990), rather the spiders must use behavioral methods of ther- moregulation. Prey availability: Although insect abun- dance was highest during the day, the re- sponse of spiders to web vibrations (simulated prey) was greater at night. Furthermore, spi- ders handled a greater proportion of insects trapped at night than during the day. Noctur- nal insects constituted only 8% of the total number of insects available on the web, but 47% of the prey actually handled by the spi- ders. The distribution of insect sizes suggests an explanation for this anomaly: more than half of the diurnal insects trapped were very small (< 3 mm body length), whereas more than half of the nocturnal insects were > 6 mm. Using an approximate conversion for in- sect body length to biomass (mass “ 0.0305 X length^ ^^; Rogers et al. 1976), we estimated that nocturnal insects constituted 28% of the biomass of available prey and 46% of the bio- mass of insects handled by the spiders. Thus, in terms of energy intake, nocturnal insects were more profitable than diurnal prey. The prey taxa available changed throughout the year; and there were significant differences in the distribution of major taxa available in the web and those handled by the spiders, sug- gesting that the spiders fed selectively. Owing to the lack of independence between prey type and prey size in our data, we cannot determine whether selection was for particular types or size classes of prey, or both factors combined. Ward (1986) analyzed prey remains from nests of S. mimosarum, finding similar sea- sonal differences in composition as well as a predominance of large prey items (beetles and orthopteroid insects). Prey exoskeletons may bias the results toward the larger insects, which are less likely to become fragmented. Thus, our observations of prey handled by the spiders confirm Ward’s conclusion, that S. mb mosarum preferentially takes large prey, even when most insects available are small. Foraging activity as a function of climat- ic conditions and prey availability. — Both web maintenance and prey capture occurred mainly at night. This strong diel pattern of activity could not be explained by climatic conditions and prey availability alone. Anoth- er important factor might be the risk of pre- dation or parasitism. From September to Feb- ruary substantial mortality occurred in colonies, largely from parasitism by Pseudo- pompilus funereus (Hymenoptera, Pompili- dae) (pers. obs.). Predation by a Red-billed Woodhoopoe (Phoeniculus purpureus) was observed on a colony of S. mimosarum at a different site. Both of these predators are di- urnal. Ambient temperature played a direct role in foraging and web-maintenance activities, while wind appeared to have less of an influ- ence. Humidity inside and outside the nest was not measured, but may influence activity as well. Typically, low humidity and high am- bient temperatures would coincide during midday (see Seibt & Wickler 1990). High am- bient temperatures during the day reduced for- aging response, as did low nighttime temper- atures. In August, the coldest month, there was little nocturnal activity. Similarly, in S. lineatus, both the speed and frequency of re- sponses to a prey stimulus (tuning fork) was lower at low ambient temperatures (Henschel et al. 1992). Another solitary eresid, Seothyra henscheli, from the Namib desert, showed very limited foraging response at temperatures below 20 °C. In general, spiders adapted to hot climates, may be constrained more by low ambient temperatures than high ambient tem- peratures, especially if foraging activity is largely nocturnal, as is the case in S. mimo- sarum. However, above-ambient temperatures inside the nest at night may act to buffer the low ambient temperature and thereby increase the time available to the spiders for foraging activity in cold winter months. Furthermore, large nests with greater thermal mass are bet- ter buffered against low temperature effects (Weldon 1997). Small colonies and newly es- tablished nests, however, may be sensitive to 168 THE JOURNAL OF ARACHNOLOGY immediate climatic conditions, as well as to the indirect effects of climate on prey avail- ability. One of the consequences of low tempera- tures at the start of the post-winter growth phase is its potential to delay maturation. In the solitary S. lineatus, delayed maturation has a strong negative effect on fitness, as the occurrence of wasp parasitism increases with time in the season and juvenile survival de- creases if emergence is delayed (Henschel et al. 1992; Schneider & Lubin 1997; Ward & Lubin 1993). In the latter species, the width of the window of time for development is de- termined by climatic factors. Long-term mon- itoring of changes in numbers and sizes of colonies of S. mimosarum will provide infor- mation on the extent to which growth and sur- vival vary with changing biotic and abiotic conditions. ACKNOWLEDGMENTS We thank Barry and Lyn Porter for provid- ing us with accommodation on their farm whilst conducting fieldwork. Tessa Hedge and Neil Crouch kindly helped with the fieldwork at night. Funding for 1995 was provided through the Foundation for Research and De- velopment and an Israel/South Africa bilateral exchange to Y.D. Lubin and M. Lawes. We are grateful to M. Lawes and M. Perrin. This is Publication No. 295 of the Mitrani Depart- ment of Desert Ecology. LITERATURE CITED Aviles, L. 1997. Causes and consequences of co- operation and permanent-sociality in spiders. Pp. 476-498, In Social Behavior in Insects and Arachnids (J.C. Choe & B.J. Crespi, eds.). Cam- bridge Univ. Press, Cambridge. D’ Andrea, M. 1987. Social behaviour in spiders (Arachnida, Araneae). Ital. J. Zool. New Series Monogr., 3:1. Elliot, J.M. 1983. Some methods for the statistical analysis of samples of benthic invertebrates. Freshwater Biol. Asso. Sci. Publ. No. 25, Titus Wilson & Son Ltd., Kendal. Haccou, P. & E. Meelis. 1994. Statistical Analysis of Behavioural Data. Oxford Univ. Press, Ox- ford, New York, Tokyo. Henschel, J.R., D. Ward & Y. Lubin. 1992. The importance of thermal factors for nest-site selec- tion, web construction and behaviour of Stego- dyphus lineatus (Araneae: Eresidae) in the Negev Desert. J. Therm. Biol., 17:97-106. Kraus, O. & M. Kraus. 1988. The genus Stegody- phus (Arachnida, Araneae). Sibling species, spe- cies groups, and parallel origin of social living. Ver. naturwiss. Ver. Hamburg, 30:151-254. Lubin, Y.D., S. Eller & M. Kotzman. 1993. Web relocation and habitat selection in a desert widow spider. J. Arachnology, 19:40-48. Nentwig, W 1985. Social spiders catch larger prey: A study of Anelosimus eximius (Araneae: Ther- idiidae). Behav. Ecol. Sociobiol., 17:79-85. Riechert, S.E. 1995. Why do some spiders coop- erate? Agelena consociata, a case study. Florida Entomol., 68:105-116. Rogers, L.E., WT. Hinds & R.L. Buschbom. 1976. A general weight vs. length relationship for in- sects. Ann. Entomol. Soc. America, 69:387-389. Rypstra, A.L. 1990. Prey capture and feeding ef- ficiency of social and solitary spiders: a compar- ison. Acta Zool. Fennica, 190:339-343. Schneider, J. & Y. Lubin. 1997. Does high adult mortality explain semelparity in the spider Ste- godyphus lineatus (Eresidae). Oikos, 79:353- 100. Seibt, U. & W. Wickler. 1988. Bionomics and so- cial structure of “Family Spiders” in the genus Stegodyphus, with special reference to the Afri- can species S. dumicola and S. mimosarum (Ar- aneida, Eresidae). Verb, naturwiss. Ver. Ham- burg, 30:255-303. Seibt, U. & W. Wickler. 1990. The protective func- tion of the compact silk nest of social Stegody- phus spiders (Araneae, Eresidae). Oecologia, 82: 317-321. Ward, D. & Y. Lubin. 1993. Habitat selection and the life-history of a desert spider, Stegodyphus lineatus (Eresidae). J. Anim. EcoL, 62:353-363. Ward, PI, 1986. Prey availability increases less quickly than nest size in the social spider Ste- godyphus mimosarum. Behaviour, 97:213-225. Weldon, D. 1997. Thermoregulatory function of nest size of the social spider Stegodyphus mi- mosarum. Unpubl. Honours thesis. Dept, of Bi- ology, Univ. of Natal, South Africa. Zar, J.H. 1984. Biostatistical Analysis. 2nd Ed. Prentice-Hall, Inc. Englewood, New Jersey. Manuscript received 18 November 1998, revised 10 January 2000. 2000. The Journal of Arachnology 28:169-179 THE IMPACT OF HABITAT FEATURES ON WEB FEATURES AND PREY CAPTURE OF ARGIOPE AURANTIA (ARANEAE, ARANEIDAE) C. Neal McReynolds^; Natural Science Division, Blue Mountain College, Blue Mountain, Mississippi 38610 USA ABSTRACT. Prey capture by the orb=web spider, Argiope aurantia Lucas 1833, depends on the type of the web=site selected. I analyzed A. aurantia web sites in open field and adjacent forest edge habitats to identify habitat features associated with web characteristics and prey capture. In the open field, the use of herbs or grass for web attachment was associated with smaller web diameters, and lower attachment heights and web heights. In both forest edge and open field, the distance to the nearest flower was less when web attachments were on composites. In the open field, webs attached to grass captured more orthopteran prey, and webs attached to herbs and composites captured more hymenopteran prey. The mean number of prey captured and the proportion of hymenopteran prey increased with higher web attachments in the open field habitat. Close proximity of webs to goldenrod in bloom in the open field habitat increased the mean number of prey captured and the proportion of hymenopteran prey. In the forest edge habitat, the presence of goldenrod was associated with more hymenopteran and orthopteran prey and with a higher mean prey number captured. Generally, webs in the open field habitat had more hymenopteran and or- thopteran prey and higher mean prey number captured than the forest edge habitat. The web-site providing the greatest probability for encountering and capturing prey is predicted to be one with a tall composite plant for web attachment near goldenrod in bloom. Keywords: Habitat selection, old-field habitat, predation, web-site Web building spiders can increase prey cap- ture by selecting sites providing high prey availability (Turnbull 1973; Riechert 1976; Riechert & Luczak 1982; Bradley 1993). Many factors determine web- site quality in- cluding thermal stress on the spider (Riechert & Tracy 1975; Tolbert 1979), web structure (Coleboum 1974; Greenstone 1984), and prey availability (Olive 1980; Howell & Ellender 1984). Web-site quality could be determined by habitat features of the web-site that influ- ence prey encounter and capture. Therefore, a spider may select a high quality web -site by choosing habitat features associated with high prey capture rate. Differences in habitat use can change a spi- der’s diet (Brown 1981; Horton & Wise 1983) by changing prey availability (Olive 1980, 1981a, 1981b, 1982) and/or web characteris- tics (Greenstone 1984). If flowers in bloom attract insect pollinators to a habitat, then flowers close to a web-site can increase pol- ' Current address: Dept, of Natural Sciences, Tex- as A&M International University, 5201 University Boulevard, Laredo, Texas 78041 USA. linators (e.g., Hymenoptera) encountering the web (see Howell & Ellender 1984; McRey- nolds & Polls 1987). Two habitat features that influence impor- tant web characteristics are the type and height of plant used for web attachment (En- ders 1973, 1975; Pasquet 1984). If sturdy plants such as trees and shrubs support larger stronger webs, then larger, more powerful prey items (e.g., Orthoptera) can be captured compared to webs on slighter plants such as grasses (Uetz et al. 1978; McReynolds & Po- lls 1987). If the flying insects (e.g., Hyme- noptera) are at greater heights in vegetation where there is more open space for flight, then increasing the height of plant used for web attachment (thus increasing web height) can increase encounters with the web by flying in- sects (McReynolds & Polls 1987). For habitat selection to be effective, differ- ent habitats or microhabitats must differ in ef- fect on individual fitness, and the individual must be able to select the higher quality hab- itat based on some environmental cue or cues (Orians & Wittenberger 1991). However, tem- poral and spatial variations in habitat quality 169 170 THE JOURNAL OF ARACHNOLOGY make it difficult to find and choose a high quality site (Orians & Wittenberger 1991), and the risk of movement from a web site increas- es the expediency of remaining in a lower quality site (Vollrath 1985). Argiope aurantia must select a web-site ensuring a high prey encounter rate in a heterogeneous old-field habitat with spatial variation in vegetation and flowers in bloom and temporal variation of flowers blooming and prey availability. This paper describes associations between habitat features and estimates of prey capture for the orb- web spider, A. aurantia. The four habitat features considered were: plant type for web attachment, web attachment height on plant, nearest flower in bloom and nearest flower distance to web. In a heterogeneous en- vironment of an old-field, these habitat fea- tures are possible cues for the spider to select a web-site with a high probability of prey cap- ture. The plants chosen for web attachment could be the most influential habitat feature for the spider building a web. Therefore, a comparison among the various plant types chosen by spiders for the highest web attach- ment with other habitat features and web char- acteristics of A. aurantia could help establish associations. The main questions I address are: How do vegetative habitat features influ- ence A. aurantia's web characteristics and the number and type of prey captured? What hab- itat features are potential cues that could be used by the spider during web-site selection to choose a web-site with high probability of future prey capture? METHODS Study animal. — Argiope aurantia builds a large vertical orb- web on vegetation in old- field habitats. The diurnal spider then sits at the web hub to wait for prey snared in the web (Reed et al. 1969). Spiders capture large prey encountering the web by wrapping the prey in silk before delivering a bite (Robinson 1969; Robinson et al. 1969; Hardwood 1974). Wrapped prey remain on the web until carried to the hub for feeding. The female spiders reach maturity and produce eggs in September and October (Olive 1980; Horton & Wise 1983). The spiderlings survive the winter in the egg sac and emerge in April and May (Tolbert 1977). Habitat. — Habitat utilization by adult fe- male A. aurantia was investigated from 4 Sep- tember-1 October 1989, 22 September-25 October 1990, 14 September-13 October 1991, and 13 September- 17 October 1992 in early successional old-field habitats located on the property of Blue Mountain College, Blue Mountain, Mississippi (1 km N of Blue Mountain on Tippah County Road 805). I di- vided the old-field into two habitats, open field and forest edge. Open field habitat was old pastures, and forest edge habitat was the margin between woods and mowed lawns for a softball field and golf course. Both habitats had a mixed grass-herbaceous vegetation of an early successional stage. The herbaceous vegetation included many species that bloom in the late summer and early autumn, such as goldenrod (Solidago spp.), boneset {Eupato- rium perfoliatum), ironweed (Vernonia sp.), fleabane {Erigeron spp.), sunflower {Helian- thus spp.), other composites (Asteraceae), honeysuckle {Lonicera japonica) and par- tridge pea (Cassia sp.). Shrubs (e.g., black- berry, Rubus sp. and pasture rose, Rosa sp.) and some early successional trees (e.g., sweet- gum, Liquidambar styraciflua; Sassafras al- bidum; and sumacs, Rhus spp.) were also common in both habitats. Willows (Salix sp.) occurred in a boggy area of the open field. The two habitats mainly differed in the pres- ence of canopy trees. The open field had sap- lings of early successional trees but very few large trees to shade the other vegetation, while the forest edge had canopy trees shading the grass-herbaceous vegetation daily. Data collection.^ — Habitat and web char- acteristics of adult female A. aurantia spiders were gathered by walking through the open field or along the forest edge and finding a spider at the web hub. This search was not considered to be a census of the spider pop- ulation in either habitat. The animals were collected in batches, uniquely marked on the dorsal abdomen with a permanent marker, and released within 24 hours on vegetation of the open field or forest edge habitat. The marked spiders were found on a web after release and observed as long as they remained at the web- site. Foraging data were collected by observ- ing captured and wrapped prey in the web. If the web- site was abandoned, attempts were made to find again the marked spider and con- tinue to record data at the new web-site. For- aging data were collected at several web-sites at one time for a total of 88 web-sites in the McREYNOLDS— HABITAT UTILIZATION BY ARGIOPE AURANTIA 111 forest edge and 57 web-sites in the open field. Additional data were collected on web and habitat characteristics from spiders that es- caped collection for marking, were not found again after mark and release, or were found later near marked spiders. Habitat and web parameters measured were: (1) the plant used for the highest web attachment point (grass, composite, herb, shrub, or tree), (2) web attachment height on that plant, (3) taxon of nearest flower in bloom to the web hub (goldenrod, boneset, ironweed, fleabane, sunflower, other composites, honey- suckle, or partridge pea) or, if no flower was within four meters, then recorded as “no flow- er,” (4) distance from nearest flower blossom to the web hub, (5) web height at the orb hub, and (6) vertical web diameter. Ironweed, flea- bane, sunflower, boneset, and other compos- ites were pooled into “composite flower” class in the forest edge habitat, and honey- suckle and partridge pea were pooled into “other flower” class. In the open field habitat, all flowers except goldenrod and boneset were pooled into the “other flower” class. Foraging data were collected by observing webs of marked individuals between 1600- 1900 h to record any prey wrapped (i.e., cap- tured) by the spider during the day. Foraging variables for each marked individual included: (1) number of wrapped prey present in the web, (2) prey taxon, and (3) prey size (length of body and width of abdomen). The mean number of days of collecting foraging data of marked individuals at a particular web-site was 2.6 days in the forest edge and 5.6 days in the open field. The mean prey number cap- tured per day for each marked individual could then be calculated. I identified to order each prey item while on the web and mea- sured the length when the condition of the re- mains allowed. Orthoptera and Hymenoptera had the highest proportions, with other insect orders (Coleoptera, Diptera, Lepidoptera, He- miptera, Homoptera, Odonata, and Mecop- tera) and arachnid orders (Araneae and Opi- liones) pooled into “other prey” class because of low numbers expected in contingency ta- bles. To reduce disturbance to the spider, prey items were not removed from the web. Un- identifiable prey were pooled with “other prey” class. Data analyses. — Comparisons between certain habitat features and other habitat fea- tures, web characteristics, and spider prey cap- ture were performed. Comparisons of relative proportion in a contingency table of a habitat feature and prey taxa captured used the ad- justed G-test for independence. The data from individual spiders were pooled in habitat clas- ses of the contigency table. Habitat classes or prey taxa classes were pooled when the as- sumption of expected values greater than five was violated. Comparisons of means were performed using analysis of variance (ANO- VA) after using the Barlett’s test (corrected) for homogeneity of variance test. If the class variances were heterogeneous, the Kruskal- Wallis test (corrected for ties) compared three or more classes and the Mann- Whitney U-test (corrected for ties) tested differences between two classes. Unplanned comparisons of a sig- nificant ANOVA were performed using the Student-Newman-Keuls Multiple Compari- sons test (Sokal & Rohlf 1981). Associations between two variables were determined using a parametric test (product-moment correla- tion) if the assumption of linearity was not violated. RESULTS Plant types used for web attachment. — In the forest edge, mean web attachment height, web height, and web diameter were not significantly different among plants used for web attachment (Table 1 A). Nearest flower distance to the web was significantly different among those plants used for attachment in for- est edge habitats and that distance was shorter with the web attached to a composite instead of grass, herbs, shrubs, or trees (Table lA). The mean web attachment heights of webs on grass and herbs were significantly lower than with shrubs, trees, and composites in the open field (Table IB). Web heights were sig- nificantly different among the types of plants used for web attachment in the open field, with webs using shrubs higher than those us- ing grass or herbs (Table IB). Web diameters were also different among the types of plants used for web attachment in the open fields, with webs attached to composites larger than webs attached to herbs (Table IB). In addi- tion, the variances of nearest flower distance among web attachment plants in the open field were significantly heterogeneous; and the mean distance to the nearest flower was great- 172 THE JOURNAL OF ARACHNOLOGY Table 1. — Parameters associated with plant types for attachment of Argiope aurantia webs in forest edge and open field habitats. All mean values ± standard error (SE). Means that are followed by the same letter are not significantly different (unplanned comparisons, P < 0.05). Mean attachment height (cm) n Mean hub height (cm) n Mean web diameter (cm) n Mean nearest flower distance (cm) n A. Forest edge Grass-Herbs 108.3 ± 5.8 30 63.6 ± 4.1 25 44.6 ± 2.1 25 186.9 ± 27.3a 16 Composites 118.1 ± 10.6 16 62.1 ± 5.0 12 38.8 ± 2.9 12 38.8 ± 19.5b 16 Shrubs 110.2 ± 6.5 23 63.8 ± 4.8 20 45.8 ± 2.1 20 120.0 ± 25.0a 20 Trees 124.6 ± 5.3 54 69.0 ± 3.9 49 48.1 ± 2.1 48 133.2 ± 14.2a 45 ANOVA F3. 119 — 1.61 F3,io2 = 0.51 F3,9, = 2.00 F3 93 = 6.28 ns ns ns P < 0.001 B. Open field Grass 97.6 ± 3.9ac 17 66.5 ± 3.8a 17 40.6 ± 2.8ab 17 109.4 ± 23.6 18 Herbs 90.6 ± 7.8a 9 57.5 ± 4.6a 8 34.4 ± 4.8a 8 55.7 ± 22.7 7 Composites 121.8 ± 7.3b 14 76.8 ± 6.6ab 14 50.4 ± 2.6b 14 25.4 ± 16.2 14 Shrubs 127.5 ± 6.2b 26 88.3 ± 5.0b 26 42.9 ± 2.4ab 26 103.5 ± 26.5 24 Trees 130.0 ± 9.7bc 6 75.0 ± 8.3ab 6 50.0 ± 7.0ab 6 75.0 ± 18.9 6 ANOVA F4.67 = 5.97 ^4,66 = 4.34 F4.66 = 2.89 P < 0.001 P < 0.01 P < 0.05 Bartlett statistic 14.03, P < 0.01 Kruskal- Wallis 12.98, P < 0.05 er when the attachment plants were grasses or shrubs (Table IB). The diet in the web of marked A. aurantia was compared among the plants used for web attachment as the mean number of prey cap- tured per day, the taxa of prey captured, and prey size of taxa. The mean prey numbers among classes of plants for web attachment were not significantly different at the forest edge (Table 2A) or in the open field (Table 2B). However, variances in mean prey number among classes of plants for web attachment were significantly heterogeneous for both hab- itats (Table 2). The proportions of prey taxa captured among the various attachment plants were significantly different for the open field but not for the forest edge (Fig. 1). In the open field, webs attached to herbs captured a higher proportion of hymenopteran prey and when attached to grass a higher proportion of or- thopteran prey (Fig. IB). The size of orthop- teran prey among classes of plants for web attachment was not significantly different in either habitat (Fig. 2). Hymenopteran prey siz- es and other prey sizes were significantly dif- ferent among classes in the forest edge but not in the open field (Fig. 2). For both hymenop- teran prey and other prey, mean prey size was greater in the herb-grass than the tree-shrub class in the forest edge habitat though not as predicted. In a comparison among prey taxa, mean prey size of orthopteran prey was sig- nificantly larger than hymenopteran or other prey in the forest edge (^2^51 = 29.0, P < 0.001) and open field (F2J67 — 109.6, P < 0.001) (see Fig. 2). Web attachment height. — Web character- istics and diet were compared to web attach- ment height in both habitats. Web height (in cm) was positively correlated with web at- tachment height (in cm) in forest edge (y = 0.494X + 9.4, r2 - 0.575, n = 102, F = 139.58, P < 0.001) and open field (y = 0.614X + 5.47, F - 0.596, n = 11, F = 101.82, P < 0.001). Web diameter (in cm) was positively correlated with web attachment height (in cm) in both habitats, but the rela- tionship is not as strong for web diameter as web height in forest edge (y ~ 0.1 7 lx + 25.9, F - 0.256, n = 104, F - 35.1, P < 0.001) and open field (y = 0.152x -I- 25.9, F = 0.124, n = ll,F = 9.77, P < 0.01). A more McREYNOLDS— HABITAT UTILIZATION BY ARGIOPE AURANTIA 173 o a o k. a Trees Shrubs Herbs-Grass Trees-Shrubs Herbs Grass Plant for Web Attachment Figure L — The proportion (%) of prey taxa captured in the webs of Argiope aurantia among plant types for web attachment in two habitats. (A) In the forest edge, the frequency of prey taxa was not significantly different among classes (Ga^j = 6.53, ns, = 4, « = 73). (B) In the open field, the frequency of prey taxa was significantly different among classes (Ga^j = 26.13, P < 0.001, df = 4, n ^ 177). Table 2. — Mean prey number captured per day per individual Argiope aurantia for plant types for attachment of webs in forest edge and open field habitats. Mean SE n A. Forest edge Herbs-Grass 0.32 0.08 23 Composites 0.61 0.15 14 Shrubs 0.64 0.25 14 Trees 0.28 0.08 36 Bartlett statistic 17.13, P < 0.001 Kruskal-Wallis statistic 5.08, ns B. Open field Herbs 0.64 0.16 7 Grass 0.30 0.10 15 Composites 1.01 0.30 12 Shrubs 0.45 0.09 17 Tree 1.38 0.79 6 Bartlett statistic 39.74, P < 0.001 Kruskal-Wallis statistic 8.16, ns 174 THE JOURNAL OF ARACHNOLOGY 25 0. 10 (A) T 109 i I □ 11 i t(B) 28 • 24 ( 4 • Orthoptera ♦ Hymenoptera □ Other Prey Herbs Tree-Shrub Herbs Grass Tree-Shrub Plant for Web Attachment Figure 2. — The mean size (±SE, n) of orthop- teran prey, hymenopteran prey, and other prey cap- tured in the webs of Argiope aurantia among plant types for web attachment in two habitats. (A) In the forest edge, prey size was significantly different among classes for hymenopteran prey (F, ,7 = 7.48, P < 0.05) and other prey (F, 21 = 6.16, P < 0.05) but not orthopteran prey (F, 20 2.20, ns). (B) In the open field, prey size was not significantly dif- ferent among classes for orthopteran prey (F2 70 = 0.19, ns), hymenopteran prey (F2 ^4 = 103, ns), and other prey (F2 21 = 0.44, ns). direct relationship between web attachment height and web height can exist because web attachment height determined the maximum web height, but web heights below maximum did occur. Web attachment height had a sig- nificant effect on these web characteristics but was not the only factor. The observed diet of A. aurantia was com- pared among web attachment height classes as the number of prey captured per day per in- dividual, and the taxa of prey caught. The mean prey number was not significantly dif- ferent among web attachment height classes in the forest edge (Table 3A) but was in the open field, with the most prey captured in higher web attachments (Table 3B). The var- iances in prey number for web attachment heights were significantly heterogeneous for both habitats (Table 3). The proportions of prey taxa captured among the web attachment height classes were not significantly different in the forest edge but were in the open field (Fig. 3) where higher webs captured a high proportion of hymenopteran prey and a low proportion of orthopteran prey (Fig. 3B). Nearest flower. — The nearest flower in bloom was compared to spider diet in both habitats. The mean number of prey captured was significantly different among the four near- est flower classes in the forest edge, with lower number of prey captured per day with no flow- er near the web than with goldenrod nearby (Table 4A). There was no difference in mean number of prey captured among the three flow- er classes in the open field (Table 4B). The variances in prey number among nearest flower classes were significantly heterogeneous for both habitats with a high variance in prey num- ber for the goldenrod class (Table 4). The pro- portions of prey taxa were significantly differ- ent among nearest flower classes in the forest edge and open field (Fig. 4). In the forest edge, the proportions of orthopteran and hymenop- teran prey were higher with goldenrod nearby; but in the open field, the proportion of hyme- nopteran prey was higher with goldenrod, and orthopteran prey proportion was highest with the other flower class. Nearest flower distance. — The nearest flower distance was compared to spider diet in both habitats. In the forest edge, mean num- ber of prey captured was not significantly dif- ferent between nearest flower distance classes of 0-50 cm and > 50 cm for any of the near- est flower taxa: goldenrod, composites, or oth- er flowers (Table 5A). In the open field, only goldenrod had a significant difference in mean number of prey captured between nearest flower distance classes, with more prey caught by spiders near goldenrod than spiders > 50 cm from goldenrod (Table 5B). The propor- tions of prey taxa captured were not signifi- cantly different among the nearest flower dis- tance classes for the forest edge but were significantly different for the open field (Fig. 5), with the proportion of hymenopteran prey increasing when the web was closer to a flower. Habitat comparisons. — A measure of hab- itat quality was estimated by comparing spider diets between the two habitats. In the forest edge, the mean number of prey captured per day was significantly less (mean ±SE = 0.40 McREYNOLDS— HABITAT UTILIZATION BY ARGIOPE AURANTIA 175 30 20- 10 (A) H Orthoptera S Hymenoptera M Other Prey (B) “T 25-75 75-100 100-125 125-200 25-75 75-100 Web Attachment Height (cm) 100-125 125-200 Figure 3. — The proportion (%) of prey taxa captured in the webs of Argiope aurantia among web attachment height classes in two habitats. (A) In the forest edge, the frequency of prey taxa was not significantly different among classes (G^dj = 10.25, m, df — 6, n — 73). (B) In the open field, the frequency of prey taxa was significantly different among classes (G^dj = 37.86, P < 0.001, df 6, n = 173). Table 3. — Mean prey number captured per day per individual Argiope aurantia for web attachment heights in forest edge and open field habitats. Mean SE n A. Forest edge 50-100 cm 0.42 0.08 31 100-125 cm 0.42 0.16 23 125-150 cm 0.41 0.15 19 150-200 cm 0.31 0.11 14 Bartlett statistic Kniskal-Wallis statistic 1.50, ns 7.60, P < 0.05 B. Open field 50-100 cm 0.44 0.08 16 100-125 cm 0.49 0.17 24 125-200 cm 1.11 0.29 16 Bartlett statistic Kruskal-Wallis statistic 9.66, P < 0.01 19.62, P < 0.001 176 THE JOURNAL OF ARACHNOLOGY H Orthoptera S Hymenoptera @ Other Prey 1 r other FL GO Nearest Flower Other FL Figure 4. — The proportion (%) of prey taxa captured in the webs of Argiope aurantia among nearest flower classes in two habitats. (A) In the forest edge, the frequency of prey taxa was significantly different among classes (Ga^j = 6.37, P < 0.05, df = 2, n = 73). (B) In the open field, the frequency of prey taxa was significantly different among classes (Ga^j = 16.4, P < 0.01, df — A, n = ill). Abbreviations: BO = boneset, GO = goldenrod. Other FI = other flowers. Table 4. — Mean prey number captured per day per individual Argiope aurantia for nearest flower in bloom in forest edge and open field habitats. Mean SE n A. Forest edge Composite Flowers 0.37 0.11 26 Goldenrod 0.53 0.12 32 Other Rowers 0.51 0.16 14 No Flower 0.08 0.06 16 Bartlett statistic Kruskal- Wallis statistic 5.08, P < 0.01 13.71, P < 0.01 B, Open field Boneset 0.54 0.14 16 Goldenrod 0.80 0.24 20 Other Flowers 0.59 0.19 21 Bartlett statistic Kruskal- Wallis statistic 1.05, ns 6.96, P < 0.05 McREYNOLDS— HABITAT UTILIZATION BY ARGIOPE AURANTIA 111 © g Pm Oithoptera IS Hymenoptera @ Other Prey I r 25-50 50-200 Nearest Flower Distance (cm) Figure 5. — The proportion (%) of prey taxa captured in the webs of Argiope aurantia among nearest flower distance classes in two habitats. (A) In the forest edge, the frequency of prey taxa was not signif- icantly different among classes = 4.34, ns, df = 4, n == 73). (B) In the open field, the frequency of prey taxa was significantly different among classes (Ga^j = 20.26, P < 0.01, df = 6, n = 177). ±0.06, n = 88), than in the open field (0.65 ±0.12, n = 57) {U = 1924.0, P < 0.05). The variances in prey number between habitats were significantly heterogeneous (F = 2.33, df = 56, 87, P < 0.001). The proportions of prey taxa captured were significantly different between the two habitats (Fig. 1, = 18.51, df = 2, P < 0.001), with the proportions of both hymenopteran and orthopteran prey higher in the open field. Therefore, the forest edge habitat had lower quality prey capture sites for A. aurantia than the open field. DISCUSSION A possible explanation for the difference in prey capture between the two habitats is the differences in the relative density of grass and herbaceous vegetation affecting prey avail- ability (Olive 1980, 1981a) and/or the pres- ence of flowers in bloom that attract A. au- rantia prey. The two habitats also differ in abiotic environmental factors (e.g., the pres- ence of shade) that could influence the spider directly or through prey availability (Riechert & Tracy 1975). Enders (1973) observed that A. aurantia shifts from closed sites with pe- rennials (description similar to the forest edge habitat) to open sites (i.e., open field) as they enter adulthood. However, more adult A. au- rantia in this study remained in the forest edge habitat because mowed lawns could act as a barrier to their movement. Habitat utilization can determine the struc- ture and size of the web (Coleboum 1974; Pasquet 1984; Lubin et al. 1993). Web height at the hub influenced the diet of A. aurantia through an increase in the proportion of Hy- menoptera and a decrease in Orthoptera cap- tured as the prey capturing surface was posi- tioned higher (McReynolds & Polis 1987). The present results are consistent: a similar association was found between web attach- ment height and proportions of prey taxa, and a positive correlation existed between web at- tachment height and web height. This increase in web attachment height was also associated with increased prey capture. Maybe spiders select web-sites providing high web attach- ments to increase web height. This, in turn, increases the frequency of encounter and cap- ture of higher flying Hymenoptera and in- creases the total number of prey captured f 178 THE JOURNAL OF ARACHNOLOGY Table 5. — Mean prey number captured per day per individual Argiope aurantia for distance to nearest flower in bloom of different flower types in forest edge and open field habitats. Mean SE n Mann- Whitney U' A. Forest edge Composite Flowers 0-50 cm 0.42 0.20 11 U' = 72.0 >50 cm 0.24 0.11 12 P = 0.73, ns Goldenrod 0-50 cm 0.6 0.19 11 U' = 131.5 >50 cm 0.37 0.09 20 P = 0.38, ns Other Flowers 0-50 cm 0.64 0.17 6 t/' = 28.5 >50 cm 0.46 0.29 7 P = 0.29, ns B. Open field Boneset 0-50 cm 0.6 0.12 10 U' = 44.5 >50 cm 0.43 0.32 6 P = 0.12, ns Other Flowers 0-50 cm 0.69 0.48 8 U’ = 42.5 >50 cm 0.67 0.13 7 P = 0.09, ns Goldenrod 0-50 cm 1.32 0.43 10 U' = 89.5 >50 cm 0.28 0.08 10 P < 0.01 (McReynolds & Polis 1987). However, these results do not support the prediction that stur- dier plants used for web attachment support stronger, larger webs and therefore capture larger and stronger prey such as orthopterans. The presence of flowers near the web site may directly affect prey capture of A. aurantia by attracting insect pollinators, herbivorous insects, and their arthropod predators. Results suggest that proximity to goldenrod increases prey capture probability more than any other flower. In both habitats, Hymenoptera were captured near goldenrod, maybe because this plant attracts more insect pollinators than oth- er flowers in old-field habitats during late summer and autumn. In the forest edge, the capture of Orthoptera also increased near goldenrod, maybe because goldenrod with as- sociated grass or herbaceous vegetation also attracts more herbivorous insects than the trees and shrubs that are common at the forest edge. Nearest flower and nearest flower dis- tance appear to be good indicators of prey capture and may be predictors of prey avail- ability and web-site quality, although nearest flower and nearest flower distance do not in- dicate the presence and density of other flow- ers in bloom near the web site. Further re- search is required to test the above predictions on the effect of goldenrod on prey availability and web- site quality. Prey capture at a web-site can fluctuate (Ja- netos 1982; Bradley 1993; VoUrath 1985), and the risk to a spider in selecting a web-site can increase with temporal and/or spatial variation in prey availabihty (Caraco & Gillespie 1986; GiUespie & Caraco 1987; Smallwood 1993). The data on within web-site variance needed to evaluate the decisions made by individual spi- ders on their tenure at web-sites (see Caraco et al. 1995) are not available in this paper. How- ever, when based on the between web-site var- iance, web-site quality is highly variable within habitat classes (e.g., high mean and variance of prey number in the goldenrod class of the open field habitat). One explanation for spatial and temporal variabihty among web-sites is that at- tractiveness of the flowers to insect poUinators around the web-site changes over time, chang- ing prey availabihty at various web-sites. These hypotheses need further testing. The predicted high quality web-site for A. aurantia (i.e., one that shows a high mean prey number) is a combination of habitat fea- tures including a tall (> 125 cm) plant for web attachment near (< 50 cm) goldenrod in bloom. However, with the high variance, there is a risk that an individual will not capture the minimum energy requirements. Caraco et al. (1995) predict that solitary spiders such as A. aurantia should be more risk-prone by select- ing highly variable foraging sites because these places would occasionally yield suffi- cient energy for survival and reproduction while less variable (with the average below the minimum) sites rarely or never yield suf- ficient energy. Therefore, a spider should se- lect a web-site with certain habitat features — not to ensure constant prey availability — but to increase the probability of occasional high , prey capture. In addition, selection of a web- j site by A. aurantia with the above habitat fea- tures should increase the probability of suffi- | cient prey capture for survival and reproduction. The major emphasis of further research is to establish whether A. aurantia does select or prefer web-sites with these pre- dicted habitat features. L McREYNOLDS— HABITAT UTILIZATION BY ARGIOPE AURANTIA 179 ACKNOWLEDGMENTS I dedicate this paper to the memory of Gary A. Polis. He improved this paper by reviewing an early draft, and he encouraged me to con- tinue research at Blue Mountain College. He will be missed. In addition, Alan Cady and anonymous reviewers improved the paper im- measurably. LITERATURE CITED Bradley, R.A. 1993. The influence of prey avail- ability and habitat on activity patterns and abun- dance of Argiope keyserlingi (Araneae: Aranei- dae). J. ArachnoL, 21:91-106. Brown, K.M. 1981. Foraging ecology and niche partitioning in orb- weaving spiders. Oecologia, 50:380-385. Caraco, T & R.G. Gillespie. 1986. Risk- sensitivi- ty: foraging mode in an ambush predator. Ecol- ogy, 67:1180-1185. Caraco, T, G.W. Uetz, R.G. Gillespie & L. Giral- deau. 1995. Resource consumption variance within and among individuals: On coloniality in spiders. Ecology, 76:196-205. Coleboum, P.H. 1974. The influence of habitat structure on the distribution of Araneus diade- matus Clerck. J. Anim. EcoL, 43:401-409. Enders, F. 1973. Selection of habitat by the spider Argiope aurantia Lucas (Araneidae). American Midi. Natur., 90:47-55. Enders, F. 1975. Effects of prey capture, web de- struction and habitat physiognomy on web-site tenacity of Argiope spiders (Araneidae). J. Ar- achnoL, 3:75-82. Gillespie, R.G. & T, Caraco. 1987. Risk-sensitive foraging strategies of two spider populations. Ecology, 68:887-899. Greenstone, M.H. 1984. Determinants of web spi- der species diversity: Vegetation structural diver- sity v.y. prey availability. Oecologia, 62:299-304. Harwood, R.H. 1974. Predatory behavior of Argiope aurantia. American Midi. Natur., 91:130-139. Horton, C.C. & D.H. Wise. 1983. The experimen- tal analysis of competition between two syntopic species of orb- web spiders (Araneae: Araneidae). Ecology, 64:929-944. Howell, EG. & R.D. Ellender. 1984. Observation on growth and diet of Argiope aurantia Lucas (Araneidae) in a successional habitat. J. Arach- nol., 12:29-36. Janetos, A.C. 1982. Foraging tactics of two guilds of web-spinning spiders. Behav. Ecol. SociobioL, 10:19-27. 1 Lubin, Y, S. Ellner & M. Katzman. 1993. Web j relocation and habitat selection in a desert widow j spider. Ecology, 74:1915-1928. j McReynolds, C.N. & G.A. Polis. 1987. Ecomor- j phological factors influencing prey use by two sympatric species of orb-web spiders, Argiope aurantia and Argiope trifasciata (Araneidae). J. ArachnoL, 15:371-383. Olive, C.W 1980. Foraging specialization in orb- weaving spiders. Ecology, 61:1133-1144. Olive, C.W 1981a. Optimal phrenology and body size of orb- weaving spiders: foraging constraints. Oecologia, 49:83-87. Olive, C.W 1981b. Co-adapted foraging traits in a guild of orb- weaving spiders. Oecologia, 49:88-91. Olive, C.W 1982. Behavioral response of a sit- and-wait predator to spatial variation in foraging gain. Ecology, 63:912-920. Orians, G.H. & J.F Wittenberger. 1991. Spatial and temporal scales in habitat selection: A landscape perspective on population dynamics. American Natur., 137:S29-S49. Pasquet, A. 1984. Predatory- site selection and ad- aptation of the trap in four species of orb-weav- ing spiders. Biol. Behav. 9:3-19. Reed, C.F., P.N. Witt & M.B. Scarboro. 1969. The orb-web during the life of Argiope aurantia (Lu- cas). Devel. PsychobioL, 2:120-129. Riechert, S.E. 1976. Web-site selection in the desert Agelenopsis aperta. Oikos, 27:311-315. Riechert, S.E. & J. Luczak. 1982. Spider foraging: Behavioral responses to prey. Pp. 353-389, In Spider Communication: Mechanisms and Ecolog- ical Significance. (P.N. Witt & J.S. Rovner, eds.). Princeton Univ. Press, Princeton, New Jersey. Riechert, S.E. & C.R. Tracy. 1975. Thermal bal- ance and prey availability: Bases for a model re- lating web-site characteristics to spider reproduc- tive success. Ecology, 56:265-284. Robinson, M.H. 1969. Predatory behavior of Ar- giope argentata (Fabricius). American ZooL, 9: 161-173. Robinson, M.H., H. Mirick & O. Turner. 1969. Predatory behavior of some araneid spiders and the origin of immobilization wrapping. Psyche, 76:487-501. Smallwood, P.D. 1993. Web-site tenure in the long- jawed spider: Is it risk-sensitive foraging or con- specific interactions? Ecology, 74:1826-1835. Tolbert, WW 1977. Aerial dispersal behavior of two orb-weaving spiders. Psyche, 84:13-27. Tolbert, WW 1979. Thermal stress of the orb- weaving spider Argiope trifasciata (Araneae). Oikos, 32:386-392. Turnbull, A.L. 1973. Ecology of the true spiders (Ar- aneomorphae). Ann. Rev. Entomol., 18:305-348. Uetz, G.W., A.D. Johnson & D.W Schemske. 1978. Web placement, web structure, and prey capture in orb-weaving spiders. Bull. British Ar- achnol. Soc., 4:141-148. Vollrath, F. 1985. Web spider’s dilemma: A risky move or site dependent growth. Oecologia, 68: 69-72. Manuscript received 3 August 1998, revised 20 De- cember 1999. 2000. The Journal of Arachnology 28:180-184 EVALUATION OF FORMULAE TO ESTIMATE THE CAPTURE AREA AND MESH HEIGHT OF ORB WEBS (ARANEOIDEA, ARANEAE) Marie Elisabeth Herberstein: Department of Zoology, University of Melbourne, Parkville 3010, Australia I-Min Tso: Department of Biology, Tunghai University, Taichung 407, Taiwan ABSTRACT. We evaluated several formulae to estimate the capture area (the area of the web covered by capture spirals) and the mesh height (the distance between capture spirals) of orb webs constructed by Argiope keyserlingi Karsch. The accuracy of the various formulae was estimated through regression anal- yses. Accordingly, we propose two new formulae specifically suited for asymmetric orb webs, which provide accurate estimates of capture area and mesh height. Keywords: Web architecture, web design, Araneidae, Argiope keyserlingi The fundamental unit of behavior in orb- web spiders is the construction and design of the web. Web size and design can vary due to prey size (Sandoval 1994), food availability (Herberstein et al. 2000; Sherman 1994; Tso 1999), developmental stage (Higgins 1995; Heiling et al. 1998; Heiling & Herberstein 1998), physiological status (Eberhard 1988), web site (Eberhard 1989) and various abiotic factors (Vollrath et al. 1997). These web var- iations can directly influence the number and types of prey entangled. For example, a larger web will increase the rate of prey interception (Chacon & Eberhard 1980; Higgins & Bus- kirk 1992; Herberstein & Elgar 1994). Simi- larly, the distance between the capture spirals (mesh height) may affect the visibility of the web (Rypstra 1982; Craig 1986) and the size of prey entangled (Uetz et al. 1978; Murakami 1983; Miyashita & Shinkai 1995; Herberstein & Heiling 1998). While the geometric nature of orb webs aids the measurement and consequent com- parison of web elements such as web size and mesh height, these are sometimes difficult to obtain, particularly in the field. Therefore, some studies have used the length of the web radius (Higgins & Buskirk 1992) or web di- ameter (McReynolds & Polis 1987) as a very rough approximation of web size. Several re- cent studies have estimated web area with the help of formulae that require only a few mea- surements of the web (e.g., Nentwig 1985; Walker 1992; Sherman 1994). Regrettably, those studies do not provide a detailed de- scription of the formulae used, nor do they estimate the accuracy of the generated values. Recently Tso (1996) investigated the orb webs of Argiope trifasciata ForskM 1775 and estimated the capture area of the web {= the area covered by sticky spirals) and the mesh height using two formulae. Despite the de- tailed description of these formulae, Tso (1996) did not provide an account of how ac- curate the estimates were. Here we test the accuracy of several formulae to estimate cap- ture area and mesh height by comparing the values derived from the formulae with exact values. Those tests will help validate surrogate variables and provide ecologists and etholo- gists with appropriate tools for estimating orb web parameters in the field. METHODS We used the webs of 1 1 adult female Agrio- pe keyserlingi Karsch 1878 (built in 40 cm X 50 cm X 8.5 cm frames in the laboratory). The spiders were collected from suburban gar- dens in Brisbane, Australia and transferred to the laboratory in Melbourne, Australia. Each spider constructed one web, which was used for analysis {n = 11). Exact mesh height was obtained by measuring each distance between the spirals in the vertical upper and lower sec- tor (Fig. 1). The values for both the upper and lower web halves were averaged for the mesh 180 HERBERSTEIN AND TSO— WEB FORMULAE EVALUATION 181 upper vertical sector (a) {dJlYix The 'Vertical Radii - Hub’ formula (b) es- timates the ‘true’ capture area by subtracting the hub area, which is calculated using the vertical hub diameter (H). This diameter ex- tends vertically from the innermost spiral in the upper web half to the innermost spiral in the lower web half (Fig. 1). (b) (d,/2)% - (H/2)2it The ‘Ellipse’ formula (c) assumes an ellip- tical approximation of the web and estimates both radii from the vertical and the horizontal diameter (d^), respectively, but includes the hub area in its estimation. The ‘Ellipse — Hub’ formula (d) subtracts the hub area using the vertical hub diameter. Figure 1 . — A schematic representation (modified from Heiling & Herberstein 1998) of an asymmetric orb-web, defining the parameters used in the equa- tions given by Tso (1996) and in this study. See text for symbols used. height of the whole web. The exact capture area was obtained by summing the area cov- ered by spirals in each web sector. The indi- vidual sector areas were calculated by treating each sector as a trapezoid, where the inner- and outermost spirals were assumed parallel. Although the inner and outer spirals may not always be perfectly parallel, we expect the consequent biases to be minimal. The exact capture area excluded the area of the hub, which is not covered by sticky spirals and therefore does not function in capturing prey. To estimate the capture area of the webs, we considered several scenarios, which dif- fered in the number of measurements taken from the webs. For example, a researcher may only know a single web diameter or may know all four web radii and the hub radii. We then developed formulae that are based on the available information and tested their predic- tive powers. The ‘Vertical Radii’ formula (a) assumes a circular approximation of the web and esti- mates the radius from the vertical web diam- eter (dy), which extends from the outermost spiral in the upper web half vertically through the hub to the outermost spiral in the lower web half (Fig. 1). The hub area is included in this formula. (c) (d,/2)(dh/2)iT (d) (d,/2)(dh/2)iT - {WiyiT The capture area formula (e; ‘Tso — Hub’) used by Tso (1996) calculates the web area of the upper and lower web halves separately us- ing semi-circle approximations. It requires the upper (r^) and lower (q) vertical radii, which extend from the hub to the outermost spiral in the upper and lower web half, respectively (Fig. 1). The area of the hub is calculated us- ing the vertical hub diameter and subtracted to estimate the capture area. The ‘Adjusted Radii — Hub’ formula (f) is a modification of the ‘Tso — Hub’ formula. It also assumes a circular approximation treating each web half as semi-circles, but it adjusts the vertical radii by taking the horizontal di- ameter into consideration. Additionally, the hub area is calculated using the upper (Hr^) and lower (Hq) hub radii separately. For this formula we required the upper and lower ver- tical radii, the horizontal diameter, the upper vertical hub radius and the lower vertical hub radius. (f) + ■irn ■Tr(Hq)^ 182 THE JOURNAL OF ARACHNOLOGY Table 1. — The mean ± SE of the actual and the estimated capture area using various formulae which either include ( + ) or exclude (-) the area of the hub. The functional relationships between the actual and the estimated values are indicated using linear regression models with the SE of the regression slope given in parentheses. The F value indicates the significance of the regression model (Wilkinson 1992). Mean ± SE (cm^) Estimate w = 11 Functional relationship Significance Actual capture area 555.8 ± 40.8 Vertical Radii + Hub 628.2 ± 47.3 y = 207.9 + 0.6 (0.22) x; F = 6.3; P = 0.03 = 0.347 Ellipse -1- Hub 572.8 ± 33.6 y = -103.9 ± 1.2 (0.13) x; F = 82.2; P = 0.0001 = 0.890 Vertical Radii — Hub 547.2 ± 41.7 y = 206.2 + 0.6 (0.03) x; F = 6.6; P = 0.03 = 0.360 Tso — Hub 637.5 ± 48.5 y = 160.4 + 0.6 (0.19) x; F = 10.7; P = 0.01 /?2 = 0.493 Ellipse - Hub 491.9 ± 29.7 y = -96.8 + 1.3 (0.12) x; F = 124.5; P = 0.0001 = 0.925 Adjusted Radii — Hub 513.6 ± 30.7 y = -116.1 + 1.3 (0.08) x; F = 273.3; P = 0.0001 R~ = 0.965 The adjusted upper (r^u) and lower (r^j) vertical web radii are: performed using SYSTAT 5.2 for the Macin- tosh (Wilkinson 1992). r au We tested two different formulae to esti- mate the average mesh height in orb-webs. The first (g) was previously published by Tso (1996) and it requires the upper and lower web radii, the hub diameter and the number of sticky spirals in the upper (S^) and lower (S,) web halves counted in the vertical sector directly above and below the hub (Fig. 1). We modified this formula (h), using the upper and lower vertical hub radii rather than the hub diameter. (h) - Hr, r, - HrA 2\(S, - 1) (S, - l)j The formulae for capture area and mesh height were evaluated using regression anal- yses between exact values and their equivalent estimates generated by the formulae. Accord- ingly, an accurate estimate generates a high correlation coefficient {R^). All analyses were RESULTS AND DISCUSSION Generating the capture area from the ver- tical diameter alone does not yield accurate estimates (Table 1). In contrast, estimates cal- culated by the ‘Ellipse’ formula are greatly improved. This is most likely to be due to the asymmetric nature of A. keyserlingi webs and indeed many other orb webs (Vollrath & Mor- en 1985; Vollrath 1987; Foelix 1992; Herber- stein & Heiling 1999). Generally, orb webs are vertically elongated, particularly in the lower web half and the horizontal radii are shorter. Thus considering the horizontal di- ameters will improve estimates for asymmet- ric webs. Subtracting the hub area from the ‘Vertical Radii’ and ‘Ellipse’ formulae further improved these estimates (Table 1). Thus ex- cluding the area of the hub from a capture area estimate is warranted for A. keyserlingi and species with similar webs. In those species, however, where the hub only takes up a small- er proportion of the web, it may be of minor importance. Despite incorporating more web parameters than the ‘Ellipse — Hub’ formula, the ‘Tso — Hub’ formula did not yield as accurate esti- mates (Table 1). This is primarily due to web asymmetry, which also affects the hub region. Consequently, the capture area is generally HERBERSTEIN AND TSO— WEB FORMULAE EVALUATION 183 Table 2, — The mean ± SE of the actual and the estimated mesh height using formulae given in Tso (1996) and this study. The functional relationships between the actual and the estimated mesh height are indicated using linear regression models with the SE of the regression slope given in parentheses. The F value indicates the significance of the regression model (Wilkinson 1992). Mean ± SE (cm) n = ll Functional relationship Significance Actual Tso (1996) This study 0.45 ± 0.02 0.39 ± 0.02 0.45 ± 0.02 y = 0.13 + 0.83 (0.17) x; = y = 0.02 + 0.95 (0.07) x; = 0.66 0.95 F = 19.96; P = 0.002 F = 199.13; P = 0.0001 overestimated, particularly in the lower web half. The most accurate estimates are gener- ated by the ‘Adjusted Radii - Hub’ formula, because vertical asymmetry is being consid- ered by incorporating the horizontal radii as well as calculating the upper and lower hub region separately (Table 1). Additionally, this formula generates separate values for the up- per and lower web regions, which can be used for further analyses. The mesh height formula used by Tso (1996) was not as accurate as our modified formula (Table 2) for two main reasons. First, Tso’s (1996) formula uses the vertical hub di- ameter rather than the upper and the lower vertical hub radii separately, which introduces a bias in asymmetric webs. Second, the sector length covered by the sticky spirals is divided by the number of spirals, a common mistake (e.g., Sandoval 1994). Instead, this length should be divided by the number of spacings between the spirals, which equals the number of spirals minus one. This is particularly im- portant for webs with few spiral spacings. Ob- viously, the accuracy of a mesh height for- mula could be further improved by sampling and incorporating additional web sectors. The appropriateness of any web formula largely depends on the geometric nature of the web. Circular approximations such as the ‘Vertical Radii — Hub’ or the ‘Tso — Hub’ formulae, may accurately estimate capture area in symmetric and circular webs. Asym- metric webs with large hub areas however re- quire more complex approximations, such as the proposed ‘Adjusted Radii — Hub’ for- mula. ACKNOWLEDGMENTS We are very grateful for the helpful com- ments provided by Norbert Milasowszky, Mark Elgar and the reviewers and editors of the Journal of Arachnology. Robert Raven provided helpful information about the loca- tion of the spiders. Doug and Sue Thiele gave permission to collect the spiders from their gardens. Diana Fisher and Simon Blomberg provided logistic support. Astrid Heiling and Volker Framenau helped with the formulae and the web graphic. John Mackenzie and Ja- net Yen provided flies for the spiders. MEH is supported by the Austrian Science Foun- dation through the postdoctoral grant J13 18- BIO. LITERATURE CITED Chacon, R & W.G. Eberhard. 1980. Factors af- fecting numbers and kinds of prey caught in ar- tificial spider webs, with considerations of how orb webs trap prey. Bull. British Arachnol. Soc., 5:29-38. Craig, C.L. 1986. Orb-web visibility: The influ- ence of insect flight behaviour and visual phys- iology on the evolution of web design within the Araneoidea. Anim. Behav., 34:54-68. Eberhard, W.G. 1988. Behavioral flexibility in orb web construction: Effects of supplies in different glands and spider size and weight. J. Arachnol., 16:295-302. Eberhard, W.G. 1989. Effects of orb web orienta- tion and spider size on prey retention. Bull. Brit- ish Arachnol. Soc., 8:45-48. Foelix, R.R 1992. Biologic der Spinnen. Georg Thieme Verlag, Stuttgart. Heiling, M.A. & M.E. Herberstein. 1998. The web of Nuctenea sclopetaria (Araneae, Araneidae): Relationship between body size and web design. J. Arachnol., 26(l):91-96. Heiling, A.M., M.E. Herberstein & G. Spitzer. 1998. Calculation of the capture thread length in orb webs (Araneae: Araneidae): Evaluation of new formulae. Ann. Entomol. Soc. America, 91(1):135-138. Herberstein, M.E., C.L. Craig & M.A. Elgar, 2000. Foraging strategies and feeding regimes: Web and decoration investment in Argiope keyserlingi 184 THE JOURNAL OF ARACHNOLOGY Karsch (Araneae: Araneidae). Evol. Ecol. Res., 2:69-80. Herberstein, M.E. & M.A. Elgar. 1994. Foraging strategies of Eriophora transmarina and Nephila plumipes (Araneae: Araneoidea): Nocturnal and diurnal orb-weaving spiders. Australian J. Ecol., 19:451-457. Herberstein, M.E. & A.M. Heiling. 1998. Does mesh height influence prey length in orb-web spiders? European J. Entomol., 95:367-371. Herberstein, M.E. & A.M. Heiling. 1999. A sym- metry in spider orb- webs: A result of physical constraint? Anim. Behav., 58:1241-1246. Higgins, L.E. 1995. Direct evidence for trade-offs between foraging and growth in a juvenile spider. J. Arachnol., 23:37-43. Higgins, L.E. & R.E. Buskirk. 1992. A trap-build- ing predator exhibits different tactics for differ- ent aspects of foraging behaviour. Anim. Behav., 44:485-499. McReynolds, C.N. & G.A. Polis. 1987. Ecomor- phological factors influencing prey use by two sympatric species of orb-web spiders, Argiope aurantia and Argiope trifasciata (Araneidae). J. Arachnol., 15:371-383. Miyashita, T. & A. Shinkai. 1995. Design and prey capture ability of webs of the spiders Nephila clavata and Argiope bruennichii. Acta Arachnol., 44:3-10. Murakami, Y. 1983. Factors determining the prey size of the orb-web spider Argiope aurantia (Koch) (Argiopidae). Oecologia, 57:72-77. Nentwig, W. 1985. Prey analysis of four species of tropical orb weaving spiders (Araneae: Aranei- dae) and a comparison with araneids of the tem- perate zone. Oecologia, 66:580-594. Rypstra, A.L. 1982. Building a better trap; An ex- perimental investigation of prey capture in a va- riety of spider webs. Oecologia, 52:31-36. Sandoval, C.P. 1994. Plasticity in web design in the spider Parawixia bistriata: A response to variable prey type. Funct. Ecol., 8:701-707. Sherman, P 1994. The orb- web: An energetic and behavioural estimator of a spider’s dynamic for- aging and reproductive strategies. Anim. Behav., 48:19-34. Tso, I.M. 1996. Stabilimentum of the garden spider Argiope trifasciata: A possible prey attractant. Anim. Behav., 52:183-191. Tso, I.M. 1999. Behavioral response of Argiope trifasciata to recent foraging gain: A manipula- tive study. American Midi. Nat., 141:238-246. Uetz, G.W., A.D. Johnson & D.W. Schemske. 1978. Web placement, web structure, and prey capture in orb- weaving spiders. Bull. British Ar- achnol. Soc., 4:141-148. Vollrath, F. 1987. Altered geometry of webs in spi- ders with regenerated legs. Nature, 328:247-248. Vollrath, F. & W. Mohren. 1985. Spiral geometry in the garden spider’s orb web. Naturwissen- schaften, 72:666-667. Vollrath, E, M. Downes & S. Krackow. 1997. De- sign variability in web geometry of an orb- weav- ing spider. Physiol. Behav., 62:735-743. Walker, J.R. 1992. What do orb webs catch? Bull. British Arachnol. Soc., 9:95-98. Wilkinson, L. 1992. SYSTAT: Statistics. Version 5.2. Evanston. Manuscript received 10 May 1999, revised 20 Sep- tember 1999. 2000. The Journal of Arachnology 28:185-194 POPULATION STRUCTURE, SEASONALITY, AND HABITAT USE BY THE GREEN LYNX SPIDER PEUCETIA VIRIDANS (OXYOPIDAE) INHABITING CNIDOSCOLUS ACONITIFOLIUS (EUPHORBIACEAE) Angelica M. Arango and Victor Rico-Gray: Departamento de Ecologia Vegetal, Institute de Ecologia, A.C.; Apdo. 63; Xalapa, VER 91000 Mexico Victor Parra-Tabla: Departamento de Ecologia, EM.V.Z. Universidad Autonoma de Yucatan; Apdo. 4-1 16; Merida (Itzimna) YUC 97000 Mexico ABSTRACT. For one year we studied the habitat use of Peucetia viridans living on Cnidoscolus acon- itifolius, in a pasture land in Merida, Yucatan, Mexico. Highest spider density was recorded in August (total 118, adults 77), and lowest in May (total 7, adults 2). Spider density was significantly higher in isolated plants and lower in plants in a patch. Sex ratio (9:3) varied from 1:1.5 in April, to 1:1 in May, and to 1:0.1 in September. The dominant instar (both sexes) changed during the study. Throughout the study more spiders were recorded in ’repose’ than performing any other activity. Foraging and feeding were more intense between July and September, when their prey, flower visitors, were more abundant. The number of spiders on plants varied spatially and temporally due to the combined effects of distance of the individual plant to the nearby forest, monthly precipitation, plant height, and number of panicles in anthesis. Forty-eight percent of the spiders were found living on plants with 20-30 panicles in anthesis (2% of the plant population). Most of the spiders (except for adult females) were found either below or above leaves. There were no significant differences in the distribution of most stadia respective to plant height. Positive significant correlations were found between the number of spiders and the abundance of floral visitors when the data were compared shifted-back one month, and between the number of spiders and the number of panicles in bloom when the data were compared shifted-back two months. When the abundance of spiders, floral visitors and number of panicles in bloom were correlated to monthly precip- itation, we found a positive significant correlation for spider abundance when the data were compared shifted-back three months, a significant negative correlation for floral visitors when the data were compared shifted-back two months, and a nonsignificant correlation for the number of panicles in bloom, although both (panicles and floral visitors) peaked in May. Keywords: Peucetia viridans, Cnidoscolus aconitifolius , population structure, seasonality Animals which are mobile during a period of their life actively select a site for nesting, feeding and reproduction. The latter may be influenced by parental habitat occupation, high density of competitors, or habitat avail- ability at certain times of the year. The new site has to provide enough food, adequate nesting conditions, and protection against en- emies and adverse weather conditions. Food resources are patchily distributed for most an- imals, regulating their feeding behavior, pop- ulation dynamics, fitness and ultimately their evolution (Bronstein 1995). Thus, the fitness of an animal should be directly influenced by its ability to find a suitable habitat, which is based on an innate preference for certain high- quality environmental characteristics (e.g., ab- sence of enemies and availability of food and shelter). Object organization in space is used to locate such habitats (McCoy & Bell 1991). Environmental characteristics exert a strong influence on habitat selection in spiders (Uetz 1991). For example, spiders depend on the structure of the environment because: (1) they need attachment sites for their webs, and (2) their sensory organs are based on the recog- nition of tactile vibrations of the substrate (Rovner & Barth 1981; Uetz & Stratton 1982). Spider populations show certain asso- ciations between their structure and the het- erogeneity and/or structural complexity of the plant community (Chew 1961; Riechert & 185 186 THE JOURNAL OF ARACHNOLOGY Reeder 1970). Certain spiders have highly specific associations with plants. Thus their abundance and richness depend directly on the availability of specific plant species. The as- sociation between spiders and plant commu- nity structure suggests stratification of species or habitat partitioning, which should decrease interspecific competition. Spiders select sites based on the level of protection against ex- treme temperatures and the destruction of webs and nests, maximizing foraging time on the web. Similarly, they use environmental el- ements as indicators of prey availability, e.g., plant flowering (Morse 1984; Pollard et al. 1995). Our field observations in the Yucatan Pen- insula, Mexico, have shown a close associa- tion between the green lynx spider (Peucetia viridans Hentz, Oxyopidae) and Cnidoscolus aconitifolius (Mill.) I.M. Johnstone (Euphor- biaceae). However, the characteristics that de- termine this habitat selection are largely un- known. The purpose of this research was to describe quantitatively this spider-plant inter- action, and to explain the physical and spatial characteristics of the habitat used by the spi- der. In particular, we addressed the following questions: (1) Which plant parts does the spi- der use more frequently? (2) Which plant characteristics determine the presence of the spider? (3) Are all the stages in the life cycle of the spider accomplished on C. aconitifol- ius? (4) How does the population of the spider vary through time? and (5) Is there synchrony between the flowering time of the plant and the life cycle of the spider? METHODS Study site and organisms. — Field work was conducted in a 13,000 m^ grassland owned by Universidad Autonoma de Yucatan, located 15.5 km south of Merida, Yucatan, Mexico (20°58'N, 89°37'W, elevation 9 m), which is surrounded on three sides by grass- land, and on one by tropical lowland dry for- est (canopy height is ca. 15 m). The genus Cnidoscolus is characterized by the presence of urticant compounds which contribute to plant defense against herbivores (Harbome & Turner 1984). Cnidoscolus acon- itifolius has extrafloral nectaries which are visited by ants, flies, bees and wasps. The in- florescence is a panicle with feminine and her- maphroditic flowers (both flowers may be pre- sent in one panicle) and has no specific pollinators (Carbajal-Rodriguez 1998). In the study site, C. aconitifolius is distributed in clumps of up to 12 individuals, but solitary individuals are common. The green lynx spider {Peucetia viridans) is a cursorial hunting spider, foraging by day and night on a wide variety of prey, common- ly living on wild flowers, grasses, low shrubs or weeds (Whitcomb & Eason 1967; Nyffeler et al. 1987a, b, 1992; Weems & Whitcomb 1977; Simon 1980; Van Niekerk & Dippe- naar-Schoeman 1994; Whitcomb et al. 1966). It is the dominant polyphagous predatory ar- thropod in certain systems. Its diet includes several insect orders, spiders (including its own species), and at times it preys on indi- viduals up to 2.5 times larger than itself (Nyf- feler et al. 1988a, 1992). In Texas and Florida, P. viridans is frequently associated with Cro- ton capitatus (Euphorbiaceae), and with relat- ed genera like Gossypium (Malvaceae) and Helianthus (Asteraceae), where it plays an im- portant role as predator of noxious fauna (Randall 1982; Simpson 1995). Peucetia vir- idans is considered an annual univoltine spe- cies, with a reproductive season during the summer. Oviposition (25-600 eggs) is during the autumn, hatching and dispersal of juve- niles by ballooning takes place during the winter; and growth of juveniles takes place in spring (Exline & Whitcomb 1965; Whitcomb & Eason 1965). Sampling design and statistics. — Field ob- servations were made between April and Sep- tember of 1997 during the last 10 days of each month; a typical day started at 0800 h and finished at 1300 h. In the first visit we marked all Cnidoscolus aconitifolius individuals {n = 183) in the sampling site. For each plant we recorded height, cover (see below), number of panicles in anthesis, distance to the forested area, and their aggregation pattern (i.e., whether isolated or in a patch, see below). To estimate plant cover, we used the formula for an ellipse (C 0.25ttZ)iZ)2, where and D2 are two perpendicular diameters crossing the center of the plant) rather than a circle, be- cause C. aconitifolius shrubs are quite irreg- ular and fit better an oval shape. A plant was considered in a patch when its leaves over- lapped with another individual and/or the dis- tance between the base of their stems was no more than 40 cm; if these parameters were not ARANGO ET Ah.—PEUCETIA VIRIDANS: PHENOLOGY AND HABITAT USE 187 met the plant was considered as solitary. The distance to the forested area was considered important because (1) it is probably the source of young spiders colonizing C. aconitifolius individuals in the grasssland, and (2) because environmental conditions are different closer to the forest (e.g., more shade and humidity, and less insolation). On each visit we counted all Peucetia vir- idans individuals present per plant, and for each spider recorded: sex, activity (repose, foraging, feeding, care of offspring or egg sacs, courtship), location on the plant (on the stem, above or below a leaf, among new leaves, among the inflorescence, on a panicle), height above the ground, and size. To estimate size we used the width of the cephalothorax, which is relatively flxed per developing in- stars (nine instars for females, and eight for males) (Brady 1964; Killebrew & Ford 1985; Louda 1982; Randall 1978; Van Niekerk & Dippenaar-Schoeman 1994; Whitcomb et al. 1966). To accomplish the above, spiders were not removed from the plants. Instars were es- timated using previously collected and mea- sured individuals, which were organized by size in a cotton- stuffed vial and preserved in 70% alcohol. The vial was placed near a spi- der and size was established by comparison. We estimated the abundance of floral visi- tors per month using five inflorescences on each of 10 isolated and 10 grouped individuals randomly selected. All visitors were counted when they made physical contact with the flowers at the time of peak activity (1200- 1230 h). A three-way analysis of variance (SigmaStat 1995) was used to determine dif- ferences on the abundance of floral visitors among months and between isolated and grouped plants; the data was transformed by obtaining the square root of the value plus one (Zar 1996). We used a two-way analysis of variance (SigmaStat 1995) to determine if P. viridans exhibits (1) vertical stratification on the plant, (2) location preferences among in- stars and over time, (3) changes on activity intensity over time, and (4) comparison of spi- der abundance per plant grouping over time. A log-linear model was fitted with the GLIM- 4 statistical system package (Francis et al. 1993) to test the hypothesis that spider pres- ence is correlated to plant characteristics (number of panicles), and that synchrony ex- ists between plant phenology, the life cycle of the spider (abundance and instar- structure per month), and the precipitation pattern of the study site. Because we used “count data,” the goodness-of-fit was evaluated with a lest using the G statistic and a Poisson error dis- tribution. With Poisson errors, the change in variance can be compared directly with ta- bles to assess its significance (Crawley 1993). In order to estimate the synchrony between plant phenology and the life cycle of the spi- der, we compared the number of blooming panicles of Cnidoscolus and the number of floral visitors to the abundance of Peucetia per month. As organisms usually need time to re- spond to changes in their environment (e.g., Ogata et al. 1996), these correlations (Pear- son) were computed following a time lag scheme, which consisted in taking the resul- tant spider abundances for a specific month and correlating them with the blooming pan- icles and/or floral visitors abundances of the preceding months. Correlations were comput- ed at one, two and three months time lag. RESULTS Population parameters. — Highest spider density was recorded in August (total 118, adults 77), and lowest in May (total 7, adults 2) (Table 1). Spider density was significantly higher in isolated plants and lower in plants in a patch {F = 9.849; P = 0.026). Sex ratio (9:3^) varied from 1:1.5 in April, to 1:1 in May, and to 1:0.1 in September (Table 1). The dominant instar (both sexes) changed during the study. For example, instar IV in April, in- star V in May, instars VI and IX in June, in- stars VII and IX in July, instars VIII (mature males) and IX (mature females) in August, with the onset of the reproductive season and the appearance of instar I; while in September the number of mature males (instar VIII) de- creased and instars I, II, and III increased (Fig. 1).^ Activity.— Throughout the study more spi- ders were in ’repose’ than in any other activ- ity. Foraging and feeding were more intense between July and September when their prey, flower visitors, were more abundant. The care of egg sacs and offspring also follows a sim- ilar pattern (Fig. 2). Habitat selection. — The number of spiders on plants of C. aconitifolius varied spatially and temporally due to the combined effects of distance of the individual plant to the nearby 188 THE JOURNAL OF ARACHNOLOGY Table 1 . — Abundance and sex ratio per month of Peucetia viridans living on Cnidoscolus aconitifolius. Month Number of spiders Spider Number of olants Spider density per plant Number of spiders per plant aggrega- tion Imma- ture Adult Total sex ratio 9 sampled Immature Adult Total Isolated Grouped April 13 5 18 1:1.5 95 0.14 0.05 0.19 16 2 May 5 2 7 1:1 181 0.03 0.01 0.04 5 2 June 20 13 33 1:0.6 178 0.11 0.07 0.19 19 14 July 34 39 73 1:0.4 183 0.19 0.21 0.40 48 25 August 41 77 118 1:0.6 183 0.22 0.42 0.64 61 56 September 31 34 65 1:0.1 183 0.17 0.19 0.36 36 26 forest, monthly precipitation, plant height, and number of panicles in anthesis. The general- ized linear model fitted explained 9.59% of the variation (Table 2). Spider abundance was significantly and positively associated with plant height (x^ = 23.07, df = 1; P < 0.01; 1.98% of total variance). On the other hand, spider abundance was significantly and nega- tively correlated to distance to the nearby for- est (x^ = -43.53, df = 1; P < 0.01; 3.73% of the total variance), monthly precipitation (X" = -23.35, df = 1; P < 0.01; 2.17% of total variance), and the number of panicles in bloom (x^ = -9.72, df = 1; P < 0.01; 0.83% of total variance). The interaction between distance to nearby forest and precipitation was also positively correlated with spider abun- dance (x^ = 6.28, df = 1; P < 0.01; 0.54% of total variance); at the onset of the rainy season spiders were found near the forest, and as precipitation increased, the distance to the forest at which spiders were found also in- creased. We also found a positive significant correlation between the interaction of precip- itation X number of panicles in bloom, and spider abundance (x^ = 3.98, df = 1; P < 0.01; 0.34% of total variance). Forty-eight percent of the spiders were found living on plants with a range of 20-30 panicles in anthesis; which only represents 2% of the C. aconitifolius population. Most of the spiders were found either below or above leaves. We did not find significant differences among spider location sites on the plant, ex- cept for below and above leaves compared with those less used sites (i.e., fruits and pan- icles, F — 4.613; P < 0.01). Likewise, most developmental stadia did not show structure preferences. Instar IX (adult females) differed significantly from the other instars (F = 2.166; P = 0.044) because quite frequently they were found living below the leaves. There were no significant differences in the distribution of most instars respective to plant height because most spiders were found be- tween 60-80 cm. Again, only the location of instar IX was statistically different from the rest {F = 7.519; P < 0.001), usually nesting at heights between 1-2 m. Synchrony between phenologies. — The number of flower visitors, the number of pan- icles in anthesis, the number of spiders, and the precipitation data per month are presented in Fig. 3. Grouped plants had significantly more floral visitors than isolated plants, and peak visitation was in July; we did not find differences in number of visitors among plant individuals either isolated or in groups (Table 3). There is a clear displacement in time among the peaks of blooming panicles (May), floral visitors (July), spiders (August), and precipitation (May and September). Positive significant correlations were found between the number of spiders and the abundance of floral visitors when the data was compared with one month time lag (r = 0.891, Pearson (g)o.o5(2),6 = 0.755, P = 0.042), and between the number of spiders and the number of pan- icles in bloom when the data was compared with two months time lag (r == 0.93, Pearson (g)o.o5(2),6 ^ 0.811, P < 0.05). When the abun- dance of spiders, floral visitors and number of panicles in bloom were correlated to monthly precipitation, we found a positive significant correlation for spider abundance when the data was compared with three months time lag (r = 0.949, Pearson {rX.05a\6 = 0 '77, P < 0.05), a negative significant correlation for ARANGO ET AL.—PEUCETIA VIRIDANS: PHENOLOGY AND HABITAT USE 189 c 50 0) E 40 0) Q. 3Q I I „ a I i Jit. I i a -13 i=a EUL 1 IV V VI VII VIII IX Instars CO Figure 1. — Instar distribution of the population of Peucetia viridans through the year in Merida, Yu- catan, Mexico. Instars I to VIII correspond to males and females, instar IX only adult females (see text, sampling design). floral visitors when the data was compared with two months time lag (r = —0.775, Pear- son (rJo.o5(2),6 = 0.77, P < 0.05), and a non- significant correlation for the number of pan- icles in bloom (r = 0.042, Pearson (g)oo5(2),6 = 0.77, P > 0,05), although both peaked in May. DISCUSSION The life cycle of Peucetia viridans has been reported (Florida, Texas and Baja California) to start with the mating season in July, eggs are laid in September, hatching and dispersal between November and early January, and growth from January to June, when males and 190 THE JOURNAL OF ARACHNOLOGY 100% - 90% - 80% - 70% - . . ' 60% - . 50% - ... 40% - Ill: 30% - 20% - . < 10% - „ ^ 0% - April May Rivi June August September a Mating ■ Feeding □ Guarding H Resting B Foraging Figure 2. — Frequency of the different activities recorded for Peucetia viridans inhabiting Cnidoscolus aconitifolius in Merida, Yucatan, Mexico. females reach their mature state (Brady 1964; Whitcomb & Eason 1966; Louda 1982; Van Niekerk & Dippenaar-Schoeman 1994). In Yucatan the cycle is similar, although dis- placed two months due to differences in the climatic patterns between our study site and the areas where the latter studies were accom- plished. In Yucatan, the abundance of P. vir- idans increases when precipitation increases and temperature decreases, courtship and mat- ing start in May, and mating peaks between June and August. One female was found guarding an egg sac in April, none were re- corded in May, while this activity increased through August (26 females guarding egg sacs and progeny); finally, hatching and dispersal occurred between August and September. Feeding behavior increased in May which co- incides with the pre-mating season, pre-adult maturation and growth of juveniles. Foraging behavior was well represented throughout, ex- cept for April and June. In summary, despite localities and changes in weather patterns, it seems that the phenology of the spider closely follows the changes in the physical environ- ment of each site. Louda (1982) found that P. viridans was associated with the larger individuals of Hap- lopappus venetus (Asteraceae) rather than on younger plants or on those with taller inflo- rescences. Our population of C. aconitifolius differed in the number of panicles in bloom Table 2. — Summary of results from the generalized linear models fitted to the data on plant physical characteristics, distribution pattern, and number of spiders present. Source of variation df % of variation P Distance to forest (A) -43.53 1 3.73 <0.01 Precipitation per month (B) -25.35 1 2.17 <0.01 Plant height 23.07 1 1.98 <0.01 Panicles in anthesis (C) -9.72 1 0.83 <0.01 A * B 6.28 1 0.54 <0.01 B * C 3.98 1 0.34 <0.01 Error 1051.87 1 90.38 Total 1163.8 100 ARANGO ET Al^.—PEUCETIA VIRIDANS: PHENOLOGY AND HABITAT USE 191 Figure 3. — Abundance of Peucetia viridans, number of panicles in bloom of Cnidoscolus aconitifolius, floral visitors and precipitation per month at the study site. per plant, and more spiders were found on taller plants, with more panicles in bloom, more cover, and closer to the nearby forest patch. Sixty-five percent of the plants had be- tween 0“5 panicles in bloom, but only 2% of the spiders were found on these plants, where- as 48% of the spiders inhabited plants with 20-30 panicles in bloom. We suggest that this pattern could be the result of the dispersal be- havior (ballooning) of the spiderlings of P. viridans, who may take refuge in the nearby forest patch during the first months of their development, and then move back to C. acon- itifolius individuals. Morse (1993) has proved that spiderlings may balloon more than once, increasing their probability of placement on a satisfactory hunting site. Peucetia viridans could be selecting larger plants (i.e., with more panicles in bloom) which will attract more floral visitors and where the spider will gain protection from the extended plant cover; while they may select isolated plants in order to avoid or decrease competition for space. Crab spiders rely heavily on cues from the environment, such as the quantity of nectar in a flower, or the number of flowers present (Morse & Fritz 1982). We suggest that P. vir- idans, guided by color recognition and/or the amount of floral nectar, chooses plants based on the number of panicles in bloom, which increases the number of visitors, and thus spi- der survivorship. Morse (1991) demonstrated that Misumena vatia actively chooses its ter- ritory, moving from poor to high quality in- florescences. Crab spiders do not seem to re- spond to unopened flowers or to the number of nectar-secreting flowers, instead, they direct their response to the number of insects attract- ed to plants (Morse 1988). Our results suggest that P. viridans is selecting plants based main- ly on the number of panicles in bloom and plant height. Table 3. — Summary of the results of the three-way ANOVA comparing the abundance of floral visitors per month, visiting isolated and grouped plants. Source of variation df MS F P Date (month) 5 16.43 11.60 < 0.001 Plants (grouped/isolated) 1 99.81 70.51 < 0.001 Differences among plants 4 0.940 0.664 = 0.624 192 THE JOURNAL OF ARACHNOLOGY We expected to find an association between spider age and their vertical distribution on the plant, since adult females were found living and foraging on taller branches and inflores^ cences, and juveniles (3rd and 4th instars) on lower locations. However, with the exception of adult females who sought high places to lay their egg sacs, the rest of the developmental stages did not show a preferred nesting or for- aging height. We did not find either any pref- erence for location on the plant among instars. Most of the spiders were found above (34%) or below (37%) leaves, and only 15% of the spiders were found on inflorescences; but there was no preference by instar or time of year. Morse (1993) found that crab spiders choose specific leaf areas to build their nests, particularly close to favorable spiderling hunt- ing sites. Gravid females of P. viridans choose the more shaded plants nearest to the forest patch, decreasing the probability of spiderling desiccation and increasing the potential food resources (more panicles in bloom). Crab spi- ders are very efficient in choosing the umbel with the largest number of white flowers, and the fact that their presence resembled the fre- quency with which insects visited umbels, rather than the number of flowers visited on these umbels, suggests that the mere appear- ance of an insect, however fleeting, provides the spider with the single largest amount of information required to make a choice (Morse & Fritz 1982). Thus the frequency of spider attack on their prey should provide us with useful information on site quality (Morse & Fritz 1982). Louda (1982) found in Baja California that flowering was correlated with the relative abundance of spiders and floral visitors. In Yucatan the flowering peak occurred in May, the peak of flower visitors in June, and the peak of spider abundance in August. However, the resulting significant correlations (either with one, two or three months of time lag) among the above variables, suggest that or- ganisms need time to respond to changes in their environment. The latter could also be an escape strategy of the plant, since blooming panicles are available to pollinators when the abundance of P. viridans is low. Not all individuals of C. aconitifolius were inhabited by P. viridans. Interestingly, these were visited by geometer caterpillars which heavily damage leaves (Parra-Tabla & Car- bajal-Rodriguez, unpubl. data). Freitas & Oliveira (1996) have demonstrated that but- terflies visually recognize potential egg pred- ators, such as ants, and actively choose sites that are better for egg-laying, thereby reducing the risk of death of their offspring. It is quite possible that the geometer moths ovipositing on C. aconitifolius recognize P. viridans and thus only oviposit on those plants without spi- ders. Even though spiders prey heavily on the plant’s pollinators, they may also impede ovi- position by the moth on the plant and thereby reduce potential leaf damage, benefiting the plant (see also Louda 1982). Finally, our results suggest that Peucetia viridans uses high-quality portions of its hab- itat, choosing those plants offering better sources of food, shelter, and favorable envi- ronmental conditions. The study of tritrophic- level interactions (e.g., plant-herbivores and/ or pollinators-predators, such as spiders) should be pursued because they may yield more information on how different clustering of relationships between species affect the ecology and evolution of interactions (Price et al. 1980; Thompson 1994). ACKNOWLEDGMENTS We appreciate the help during field work of Miguel Carbajal-Rodriguez and Ascencidn Capistran. We thank the authorities of Facul- tad de Medicina Veterinaria y Zootecnia at Universidad Autonoma de Yucatan for the fa- cilities to use their pasture land to accomplish this study. This research was supported by CONACYT No. 90679 to AMA and No. 95- 0137 to VRG, and Instituto de Ecologia, A.C. No. 902-16. LITERATURE CITED Brady, A.R. 1964. The lynx spiders of North America, north of Mexico (Araneae, Oxyopidae). Bull. Mus. Comp. Zool., 131:506-518. Bronstein, J.L 1995. The plant-pollinator land- scape, Pp. 256-288, In Mosaic Landscapes and Ecological Processes. (L. Hansson, L. Fahrig & G. Merriam, eds.). Chapman and Hall, London. Carbajal-Rodriguez, M. 1998. Biologia reproduc- tiva y efectos de la hebivoria foliar en la expre- sion sexual y el exito reproductivo en la planta ginomonoica Cnidoscolus aconitifolius (Mill.) LM. Johnstone (Euphorbiaceae). B.Sc. thesis. Universidad Autonoma de Yucatan, Mexico. Chew, R.M. 1961. Ecology of the spiders of a de- sert community. J. New York Entomol. Soc., 9: 5-41. ARANGO ET Al.,—PEUCETIA VIRIDANS: PHENOLOGY AND HABITAT USE 193 Crawley, M.J. 1993. GLIM for Ecologists. Black- well Scientific Publ., Oxford. Exline, H. & W.H. Whitcomb. 1965. Clarification of the mating procedure of Peucetia viridans (Araneidae: Oxyopidae) by a microscopic ex- amination of the epigynal plug. Florida Ento- moL, 48:169-171. Francis, B., M. Green, & C. Payne (eds.). 1993. The GLIM system, release 4 manual. Clarendon Press, Oxford. Freitas, A.V.L. & PS. Oliveira. 1996. Ants as se- lective agents on herbivore biology: Effects on the behaviour of a non-myrmecophilous butter- fly. J. Anim. EcoL, 65:205-210. Harbome, J.B. & B.L. Turner. 1984. Plant Che- mosystematics. Academic Press, New York. Killebrew, D.W. & N.B. Ford. 1985. Reproductive tactics and female body size in the green lynx spider, Peucetia viridans (Araneae, Oxyopidae). J. ArachnoL, 13:375-382. Louda, S.M. 1982. Inflorescence spiders: a cost/ benefit analysis for the host plant, Haplopappus venetus Blake (Asteraceae). Oecologia, 55:185- 191. McCoy, E.D. & S.S. Bell. 1991. Habitat structure: The evolution and diversification of a complex topic, Pp. 3-27, In Habitat Structure. (S.S. Bell, D.E. McCoy & R.H. Mushinsky, eds.). Chapman and Hail, New York. Morse, D.H. 1984. How crab spiders (Araneae, Thomisidae) hunt at flowers. J. ArachnoL, 12: 307-316. Morse, D.H. 1988. Relationship between crab spi- der Misumena vatia nesting success and earlier patch-choice decisions. Ecology, 69:1970-1973. Morse, D.H, 1991. Choosing hunting sites with lit- tle information: Patch-choice responses of crab spiders to distant cues. Behav. EcoL, 4:61-65. Morse, D.H. 1993. Some determinants of dispersal by crab spiderlings. Ecology, 74:427-432. Morse, D.H. & R.S. Fritz. 1982. Experimental and observational studies of patch-choice at different scales by the crab spider Misumena vatia. Ecol- ogy, 63:172-181. Nyffeler, M., D.A. Dean & W.L. Sterling. 1987a. Predation by green lynx spider, Peucetia viridans (Araneae: Oxyopidae), inhabiting cotton and woolly Croton plants in east Texas. Environ. En- tomoL, 16:355-359. Nyffeler, M., D.A. Dean, & W.L. Sterling. 1987b. Evaluation of the importance of the striped lynx spider, Oxyopes salticus (Araneae: Oxyopidae), as a predator in Texas cotton. Environ. Entomol. 16:1114-1123. Nyffeler, M., D.A. Dean & W.L. Sterling. 1992. Diets, feeding specialization, and predatory role of two lynx spiders, Oxyopes salticus and Peu- cetia viridans (Araneae, Oxyopidae), in a Texas cotton agroecosystem. Environ. Entomol., 21: 1457-1465. Ogata, N., V. Rico-Gray & D. Nestel. 1996, Abun- dance, richness, and diversity of Myxomycetes in a neotropical forest ravine in Mexico. Biotro- pica, 28:627-635. Pollard, S.D., M.W Beck & G.N. Dodson. 1995. Why do male crab spiders drink nectar? Anim. Behav., 49:1443-1448. Price, P.W, C.E. Bouton, P. Gross, B.A. McPheron, J.N. Thompson & A.E. Weis. 1980. Interaction among three trophic levels: Influence of plants on interactions between insect herbivores and natural enemies. Annu. Rev. EcoL Syst., 11: 41-65. Randall, J.B. 1978. New observation of maternal care exhibited by the green lynx spider, Peucetia viridans (Araneida: Oxyopidae) Psyche, 84:286- 291. Randall, J.B. 1982. Prey records of the green lynx spider, Peucetia viridans (Hentz) (Araneae, Ox- yopidae). J, ArachnoL, 10:19-22. Riechert, S.E. & WG. Reeder. 1970. Effects of fire on spider distribution on southwestern Wisconsin prairies, Pp. 73-90, In Proceedings of the 2nd Midwest Prairie Conference. (J.H. Zimmerman, ed.). Rovner, J.S. & EG. Barth. 1981. Vibratory com- munication through living plants by a tropical wandering spider. Science, 214:464-466. Simon, E. 1980. Etudes arachnologiques 22e Me- moire. 34 etude sur les Arachnides deT Yemen. Ann. Soc. Entomol. France (Ser. 6), 10:77-124. Simpson, M.R. 1995. Covariation of spider egg and clutch size: The influence of foraging and parental care. Ecology, 76:795-800. SigmaStat. 1995. SigmaStat/w 2.0, statistical soft- ware. Jandel Scientific Software, San Rafael, California. Thompson, J.N. 1994. The Coevolutionary Pro- cess. Univ. Chicago Press, Chicago. Uetz, G.W 1991. Habitat structure and spider for- aging, Pp. 325-348, In Habitat Structure. (S.S. Bell, D.E. McCoy & R.H. Mushinsky, eds). Chapman and Hall, New York. Uetz, G.W & G.E. Stratton. 1982. Acoustic com- munication and reproductive isolation in spiders, Pp. 123-159, In Spider Communication. (P.N. Witt & J.S. Rovner, eds.). Princeton Univ. Press, Princeton. Van Niekerk, P. & A.S. Dippenaar-Schoeman. 1994. A revision of the Afrotropical species of Peucetia (Araneae: Oxyopidae). Entomol. Mem. Dept. Agric. Repub. South Africa, 89:1-50. Weems, H.V. & WH. Whitcomb. 1977. The green lynx spider, Peucetia viridans (Hentz) (Araneae: Oxyopidae). Entomol. Circular No. 181. Florida: Dept, Agriculture and Consumer Services. Whitcomb, W.H. & R. Eason. 1965. The mating 194 THE JOURNAL OF ARACHNOLOGY behavior of Peucetia viridans (Araneidae: Oxy- opidae). Florida Entomol., 48:163-167. Whitcomb, W.H., M. Hite & R. Eason. 1966. Life history of the green lynx spider Peucetia viri- dans (Araneae: Oxyopidae). J. Kansas Entomol. Soc., 39:259-267. Whitcomb, W.H. & R. Eason. 1967. Life history and predatory importance of the striped lynx spi- der (Araneidae; Oxyopidae). Proc. Arkansas Acad. Sci., 21:54-59. Zar, J.H. 1996. Biostatistical Analysis. Prentice- Hall, New Jersey. Manuscript received 22 November 1998, revised 16 December 1999. 2000. The Journal of Arachnology 28:195-200 FOOD CONSUMPTION RATES AND COMPETITION IN A COMMUNALLY FEEDING SOCIAL SPIDER, STEGODYPHUS DUMICOLA (ERESIDAE) Nava Amir, Mary E. A. Whitehouse^ and Yael Lubin: Mitrani Department of Desert Ecology, Blaustein Institute for Desert Research, Ben Gurion University of the Negev, 84990 Sede Boker, Israel ABSTRACT, A major factor which affects an animal’s consumption rate is competition for food items. Competition usually results in a drop in consumption rate; however, this may be counteracted if the animals can exploit the foraging efforts of others, as could occur in social spiders when feeding on the same prey item. Spiders digest prey extra-orally and might utilize the enzymes or digesta produced by other indi- viduals feeding from the same prey item. We investigated prey consumption in the social spider Stego- dyphus dumicola to determine if the rate of consumption of individual spiders changed in the presence of competitors. We found that when one spider fed on small prey, food consumption rate decreased with feeding duration. When the prey was larger in relation to the spider there was an initial delay in con- sumption. There was no apparent advantage for a second spider to feed on a prey item already being consumed: the second spider fed for less time and gained less mass. These results indicate that social spiders compete during the process of food ingestion and the presence of another spider reduces the value of the prey item to a subsequent forager. Keywords: Competition, sociality, foraging Foraging theory indicates that the rate at which food is consumed at a patch strongly influences the residual value of that patch to an animal (Krebs et al. 1974; Chamov 1976; Iwasa et al. 1981). Competition among con- specifics can reduce consumption rate by re- ducing the residual amount of food in a patch available to the forager, or by reducing the amount of time available for foraging owing to time lost in direct physical confrontation (Sasvari 1992). In group-feeding social species, competi- tion during foraging and feeding is expected to be less extreme than in solitary species. So- cial spiders are those that live in communal webs in which there are no individually de- fended territories (D’ Andrea 1987; Aviles 1997; Whitehouse & Jackson 1998). Social spiders cooperate in capturing prey which is then consumed by a group of individuals. By cooperating, they can handle larger prey than most similar-size solitary species (Nentwig 1985; Rypstra & Tirey 1991; Rypstra 1993; Pasquet & Krafft 1992). ' Current address: Department of Zoology and Entomology, The University of Queensland, Bris- bane, Queensland 4072, Australia Spiders feed using extra-oral digestion in which they pump enzymes into the body of the prey and then ingest the emulsified con- tents (Collatz 1987; Cohen 1995). Extra-oral digestion affects the rate at which food can be consumed by such a predator during a feeding bout. As enzymes need time to digest prey, the predator may not ingest much food in the initial stages of feeding, but it can consume food at a fast rate later on, once the prey is digested. In social spiders many individuals can feed on the same prey, which may mean they have access to each other’s enzymes. This could result in spiders exploiting en- zymes and digesta of other individuals (Ward & Enders 1985; Whitehouse & Lubin 1999). In this situation, the presence of conspecifics feeding concurrently on a food item may ac- tually increase the value of the food item and increase the rate of consumption for the “ex- ploiting” spider. The timing of feeding by an individual within a group foraging event could influence its rate of prey consumption. If the prey is initially digested slowly and then later digest- ed quickly, it may be advantageous to feed from the prey later in the foraging event, after 195 196 THE JOURNAL OF ARACHNOLOGY other spiders have injected enzymes and di- gestion has begun. Alternatively, if the prey is digested quickly and consumption is initially fast but quickly drops off, then the first to feed will gain the most and it may be advantageous to lead the attack on the prey in order to se- cure the best feeding position. Attacking first, though, is potentially hazardous. The attacker must subdue the prey, possibly depleting its poison reserves and putting itself at risk. Thus, the rate of prey consumption can influ- ence both the attack and the feeding strategies of group-feeding spiders. Factors which have been shown to affect consumption rate in ex- tra-orally digesting predators include the size of the prey relative to the predator (Cohen & Tang 1997; Erickson & Morse 1997), the type of prey involved (Leborgne et al. 1991), and the size of the feeding group (Ward & Enders 1985). We studied the feeding behavior of the so- cial spider Stegodyphus dumicola (Eresidae) to determine the influence of the presence of a conspecific on the trajectory of prey con- sumption. Stegodyphus dumicola occurs in southern Africa in colonies of up to several hundred individuals. The spiders cooperate in nest construction, care of young and prey cap- ture, and readily feed together in large groups (Seibt & Wickler 1988; Wickler & Seibt 1993). We examined the consumption rate of “groups” consisting of only two animals feeding on small grasshoppers of half to two- thirds their body size and compared consump- tion rates of members of a pair and of solitary individuals. While a group size of two indi- viduals is unusual, such groups do occur in nature (Henschel 1991/1992); and even in larger nests, small prey items are often at- tacked by only a few individuals (Lubin pers. obs.). METHODS Colonies of S. dumicola containing juve- niles were collected in Namibia in January 1996 and housed in Sede Boker, Israel, in a climate-controlled room at 27 °C. and a pho- toperiod similar to outside conditions. Exper- iments were conducted from July 1996 to March 1997, and all spiders used in the ex- periments were derived from the same colony. The spiders were all juvenile females weigh- ing about 40 mg, or about two-thirds adult size. Voucher specimens are deposited at the Mitrani Department for Desert Ecology. Food consumption pattern of single spi- ders.— Consumption rates were determined for spiders feeding alone in two tests. Because the tests were separated by a few weeks, spi- ders in the second test were larger than those in the first. In the first test, 51 individuals of similar body size were drawn from the colony and put in individual plastic containers (a cyl- inder 30 cm long, diameter 12 cm) with sup- ports for web building, where they were given seven days to acclimate. After a week, each spider was weighed on an analytical balance to the nearest 0. 1 mg, and then fed one grass- hopper nymph. We recorded the time until the spider attacked the prey, and the length of time the spider fed (excluding pauses in feed- ing). Different individuals were allowed to feed for predetermined durations (15, 30, 60, 90, 120, 180, and 240 min) after which feed- ing was stopped and each spider was re- weighed. In the second test, conducted concurrently with the test of pairs of spiders (see below), 23 individuals were allowed to feed for dif- ferent durations, as in the first test. There was a small, but significant difference in body siz- es of spiders between the two tests {t = —4.6, df= 72, P < 0.001; average body mass in the first test: 42.15 ±5.7 mg, second test: 48.6 ± 5.3 mg). The prey mass was increased in the second test (t = —\\.6,df= 72, P < 0.001; average prey mass in the first test: 19.3 ±3.6 mg, second test: 29.5 ± 3.2 mg). The ratio of prey mass to spider mass was higher in the second test (0.61 ± 0.05) than in the first (0.465 ± 0.1; arcsin transformed ratios, t = -6.85, df = 72, P < 0.001). Food consumption of pairs. — Twenty-one pairs of spiders were matched for size (body mass: 46.7 ± 6.3 mg; average mass difference between pairs = 3.5 mg, range 0-14.6 mg). To distinguish between pair members, bee numbers (numbers designed for use in api- aries) were glued to the abdomen with trans- parent nail polish. The pairs were placed in plastic containers and left for seven days to acclimate. Before the experiment each spider was weighed, and each pair was given one grasshopper nymph. We recorded the time un- til the first spider attacked the prey, and the duration of feeding (excluding pauses in feed- ing). Once the first spider had fed for a pre- AMIR ET AL.— FOOD CONSUMPTION IN STEGODYPHUS 197 Figure 1. — Relative consumption rates (% in- crease in body mass) of spiders feeding alone: a comparison of two ratios of prey/spider mass. Filled triangles (A), heavy line: low ratio = 0.465 ±0.1; open squares (□), thin line: high ratio = 0.62 ± 0.07. The polynomial regression of the low-ratio curve is: y = -x^ -h 0.19x - 2.03, = 0.74; the regression of the high-ratio curve is: y = x^ - 12x + 0.09, - 0.78 (percentages were arcsine trans- formed for the regressions). determined length of time (either 15, 30, 60, 90, 180, or 240 min) the experiment was stopped. If the second individual had begun to feed, we recorded the time it began to feed and the duration of feeding. RESULTS Food consumption pattern of spiders feeding alone.—The rate of food consump- tion by spiders feeding alone was examined in the two different tests. When spiders fed in the absence of conspecifics (first test), their body mass increased with the time spent feed- ing (Fi 23 = 39.8; P < 0.001). However, there was a significant difference in mass gain be- tween the first and second tests (ANCOVA: final body mass with initial mass as covariate, F 1 71 = 7.94, P = 0.006; Fig. 1) which was caused by differences in the trajectories of mass increase experienced by the two groups. In the first test, the relative change in body mass was initially linear and began to asymp- tote after two hours. In the second test, the coefficient changed sign, and the spiders be- gan to show an increase in body mass only after an hour of feeding. The difference be- tween the two tests is explained in part by the different prey mass/spider mass ratios. In a general linear model, both prey mass and feeding time were significant (P < 0.001, n = 74, with initial spider mass as covariate), to- gether explaining 87.4% of the variance in fi- nal spider body mass for both tests. Food consumption of spiders in pairs. — When all 21 pairs of spiders were considered, the trajectories of prey consumption of first and second spiders did not differ (ANCOVA, P > 0.1; combined regression, y = 0.002x + 0.055, Fig. 2). However, in 12 instances (57.1%), only a single spider of the pair fed. To establish whether one spider in a pair fed alone significantly more often than both spi- ders together, we needed to take into account the fact that we stopped spiders at different times after they started to feed. In the above 21 pairs, the maximum time taken for the first spider to attack the prey was 170 minutes. If we assume that the second spider responded to the prey in the same manner as the first spider, it should also have a maximum delay of 170 minutes before beginning to feed. Con- sequently, we removed the eight tests in which the experiment was stopped before it had run for 170 minutes. Of the remaining 13 tests, although the first spider fed in all of them, the second spider fed in only six instances (com- parison of first and second spiders, Fisher’s exact test, P = 0.005). There was a short but variable delay be- tween the attack of the first and second spider (median = 16 min, range = 8-164 min, n = 9). The first spider always fed longer than the second spider (Wilcoxon signed ranks test, z = —2.67, P = 0.008, n = 9), and there was a trend for the first spider to gain more mass than the second (Wilcoxon signed ranks test, z = — 1.7, F = 0.086, n = 9). We tested for differences in the consumption rates of the first and second spiders that fed together by comparing the regressions of final body mass on net feeding time, with initial body mass as CO variate. The consumption rate of the first spider was greater than that of the second spi- der to feed (ANCOVA, 15 = 4.244, P = 0.057). Mass loss.—Some spiders lost mass during the feeding trials. The mass lost by the spiders was always larger than the measurement error due to weighing inaccuracy, which was cal- culated at 0.08 mg. In the first test with single spiders, four spiders (7.8%, n = 51) lost mass (median = —0.6 mg, range = —0.1 to -1.3 mg); all had fed for 15-30 min. Twelve spi- 198 THE JOURNAL OF ARACHNOLOGY Feeding duration, min Figure 2. — Relative consumption rates of spiders feeding in pairs: percent increase in body mass against net feeding time of each spider. First spider to begin feeding: open diamonds (o), solid line; second spider: closed squares (■), dashed line. The regression equations are: first spider, y = 0.092x — 3.98, = 0.58; second spider, y = + 0.06x — 1.13, /?- = 0.655 (arcsin transformed percentages). ders lost mass in the second test (median = — 1.25 mg, range == —0.2 to —3.0 mg), with feeding durations ranging from 15-118 min (median = 33 min). The feeding duration of spiders that lost mass was significantly shorter than of those that gained mass (median =120 min, range 52-180 min; Mann-Whitney U = 6.5, P < 0.001, n = 23). In spiders that fed in pairs, a decrease in body mass occurred in three first and nine sec- ond spiders. As the sample sizes were small, we used bootstrapping (Simon 1995) to deter- mine the probability that the observed differ- ence between the median mass loss in the two groups occurred by chance alone. There was no difference in mass loss between second spiders that fed (n = 3) and second spiders that did not feed (n = 6, P = 0.21). However, first spiders that fed and lost mass (n = 4) tended to lose more than second spiders that fed and lost mass (n = 3, P = 0.087) and more than the single spiders of the concurrent second test (n = 12, P = 0.078). DISCUSSION When solitary spiders fed on small prey items, their body mass increased with feeding time. In the first test with single spiders, using relatively small prey (prey/spider mass = 0.465), the gain followed a typical curve of diminishing returns, similar to that shown by the spider Zygiella x-notata (Araneidae) feed- ing on cricket nymphs (prey/spider mass = 0. 1-0.3; Leborgne et al. 1991). Thus, small prey items, less than half the mass of the spi- der, are rapidly depleted. Another spider at- tempting to feed on the same prey item would gain no advantage by waiting, and would ob- tain more food by joining early in the feeding bout. With larger prey (prey/spider mass = 0.6), there was a delay in the spider’s consumption, resulting in a feeding trajectory with the op- posite sign to that above (Fig. 1). The lag be- fore the initial increase in food intake might be due to the time necessary for enzymes to take effect in digesting the larger meal. The delay was more pronounced when spiders fed alone than when they fed in pairs. This sug- gests that the presence of conspecifics caused spiders to increase their consumption rate. During the initial period on the prey, when venom and enzymes are presumably being pumped into the prey, spiders may even lose mass. Although sample sizes were small, mass AMIR ET AL.— FOOD CONSUMPTION IN STEGODYPHUS 199 loss was greater in first spiders than in second spiders or spiders feeding alone (comparing only those individuals that lost mass). Thus, with large prey it may be advantageous for a second spider to join later and capitalize on enzymes injected by the first spider (Ward & Enders 1985). In tests with pairs of spiders, however, we found that the second spider tended to join early in the feeding bout. In spite of possible advantages of such “enzyme piracy” (Whitehouse & Lubin 1999), second spiders fed for less time than the first spiders and had lower consumption rates. The advantage shown for the first spider to feed agrees with other studies of group feed- ing in social spiders. Willey & Jackson (1993) found that in Stegodyphus sarasinorum, when tested in groups of 10 individuals, spiders that attacked first fed for longer duration than those that arrived later. In Stegodyphus mi- mosarum (Ward & Enders 1985), the first spi- der of a pair to attack did not feed longer than its partner, but fed more frequently from the thorax and head of the prey, body parts which yield the highest reward (Robinson 1969), while its partner showed no feeding site pref- erence. Likewise, in a group of five individ- uals of S. dumicula matched for size, the first spider that attacked the prey tended to obtain more food, but did not feed for longer (White- house & Lubin 1999). In the latter study the individuals that gained the most mass were those that fed longest during the middle part of a foraging bout, although they also tended to initiate the attack (Whitehouse & Lubin 1999). Competition over prey occurs in coopera- tive group-living spiders (Ward & Enders 1985, Whitehouse & Lubin 1999), but it is apparent mainly in differences in rates of food consumption. In the social Stegodyphus, there is little evidence of active competition in the form of aggressive interactions over prey. In this study, we found that when the prey item is smaller than the spider, often only a single spider will attack and feed, and when two in- dividuals do feed together, the second obtains less food from the prey. The results of this study suggest that “piracy” of enzymes or di- gesta may occur, and that spiders may adjust the timing of feeding and their consumption rate to compensate for losses due to other in- dividuals. These considerations as well as dif- ferences in possible trajectories of food con- sumption, e.g., in relation to the relative size of prey and spider, may influence the deci- sions to join an individual feeding on a prey item. Further studies of the dynamics of group feeding and the physiology of food ingestion are needed to understand the costs and bene- fits of group feeding in social spiders. ACKNOWLEDGMENTS We thank Joh Henschel for providing us with the colonies and Alain Pasquet and Ray- mond Leborgne for commenting on the man- uscript. This is contribution #289 from the Mitrani Department for Desert Ecology. LITERATURE CITED Aviles, L. 1997. Causes and consequences of co- operation and permanent-sociality in spiders. Pp. 476-499, In The Evolution of Social Behaviour in Insects and Arachnids. (J.C. Choe & B.J. Crespi, eds.). Cambridge Univ. Press, Cam- bridge. Cohen, A.C. 1995. Extra-oral digestion in preda- ceous terrestrial arthropoda. Annu. Rev. Ento- moL, 40:85-103. Chamov, E.L. 1976. Optimal foraging, the margin- al value theorem. Theor. Popul. Biol., 9:129- 136. Cohen, A.C. & R. Tang. 1997. Relative prey weight influences handling time and biomass ex- traction in Sinea confusa and Zelus renardii (Heteroptera: Reduviidae). Environ. EntomoL, 26:560-565. Collatz, K.G. 1987. Structure and function of the digestive tract. Pp. 229-238, In Ecophysiology of spiders (W. Nentwig, ed.). Springer- Verlag, Berlin. D’Andrea, M. 1987. Social behaviour in spiders (Arachnida, Araneae). Monit. Zool. Italiano Monogr., 3:1-156. Henschel, J.R. 1991/1992. Is solitary life an alter- native for the social spider Stegodyphus dumi- colal Namibia Scient. Soc., Windhoek, Namibia, 43:71-79. Iwasa, Y., M. Higashi & N. Namaura. 1981. Prey distribution as a factor determining the choice of optimal foraging strategy. American Nat., 117: 710-723. Krebs J.R., A. Kacelnick & P. Taylor. 1974. Hunt- ing by expectation or optimal foraging? A study of patch use by chickadees. Anim. Behav., 22: 953-964. Leborgne, R., A. Pasquet & M.A. Sebrier. 1991. Modalities of feeding behaviour in an orb-weav- ing spider Zygiella x-notata (Clerck) (Aranei- dae). Behaviour, 117:3-4. Nentwig, W. 1985. Social spiders catch larger prey: 200 THE JOURNAL OF ARACHNOLOGY A study of Anelosimus eximius (Araneae: Ther- idiidae). Behav. EcoL SociobioL, 17:79-85. Pasquet, A. & B. Krafft. 1992. Cooperation and prey capture efficiency in a social spider, Ane- losimus eximius (Araneae, Theridiidae). Etholo- gy, 90:121-133. Robinson, M.H. 1969. Predatory behaviour of Ar- giope argentata (Fabricius). American Zool., 9: 161-173. Rypstra, A.L. 1993, Prey size, social competition, and the development of reproductive division of labor in social spider groups. American Nat., 142:868-880. Rypstra, A.L. & R.S. Tirey. 1991. Prey size, prey perishability and group foraging in a social spi- der. Oecologia, 86:25-30. Sasvari, L. 1992. Great tits benefit from feeding in mixed-species flocks: A field experiment. Anim. Behav., 43:289-296. Seibt, U. & W. Wickler. 1988. Bionomics and so- cial structure of ‘family spiders’ of the genus Ste- godyphus, with special reference to the African species Stegodyphus dumicola and Stegodyphus mimosarum (Araneida, Eresidae). Verb, naturw. Ver. Hamburg, 30:255-304. Simon, J.L. 1995. Resampling Stats (software). Resampling Stats, Inc. Arlington, Virginia. Ward, PI. & M.M. Enders. 1985, Conflict and co- operation in the group feeding of the social spi- der Stegodyphus mimosarum. Behaviour, 94: 167-182. Whitehouse, M.E.A. & R.R. Jackson. 1998. Pred- atory behaviour and parental care in Argyrodes flavipes, a social spider from Queensland. J. Zool., London, 244:95-105. Whitehouse, M.E.A. & Y.D. Lubin. 1999. Strategic competition during foraging in the social spider Stegodyphus dumicola. Anim. Behav., 58: 677- 688. Wickler, W & U. Seibt. 1993. Pedogenetic soci- ogenesis via the “sibling-route” and some con- sequences for Stegodyphus spiders. Ethology, 95: 1-18. Willey, M.B. & R.R. Jackson. 1993. Predatory be- havior of a social spider, Stegodyphus sarasino- rum (Araneae: Eresidae): Why attack first? Ca- nadian J. Zool., 71:2220-2223. Manuscript received 10 May 1999, revised 10 Au- gust 1999. 2000. The Journal of Arachnology 28:201-210 PREDATOR AVOIDANCE ON THE WATER SURFACE? KINEMATICS AND EFFICACY OF VERTICAL JUMPING BY DOLOMEDES (ARANEAE, PISAURIDAE) Robert B. Suter and Jessica Gruenwald: Department of Biology, Vassar College, Poughkeepsie, New York 12604 USA ABSTRACT. Vertical jumps of fishing spiders {Dolomedes sp.) from the water surface have been pre- sumed to be evasive behaviors directed against predatory fish. We used high-speed videography to analyze the jumps of fishing spiders and then constructed a numerical model to assess the effectiveness of these jumps in evading predatory strikes by trout. Jump height (mean = 3.7 cm) and duration (mean = 0.17 sec) were similar across spider masses (0.05-0.66 g) but latency to jump increased significantly with mass. To accomplish jumps of similar height, more massive spiders had to generate more force during the propulsive phase of the jump than did smaller spiders; and the contribution of fluid drag to the total force used in jumping was substantially greater for large spiders than for smaller ones. Our model juxtaposing the jumps of spiders and the attacks of trout revealed that jump heights and durations were inadequate: only the most lethargic strikes by trout could be successfully evaded by jumping vertically from the water surface. Keywords: Hydrodynamics, aquatic locomotion, predation, spider, Dolomedes Fishing spiders {Dolomedes sp.; Araneae, Pisauridae), noted for their locomotion on the water surface (e.g., Barnes & Barth 1991; Shultz 1987), are adept at predation both on land and on the surface of ponds and slowly flowing streams (Gorb & Barth 1994). While on the water surface, they can also become prey, captured not only by animals that detect them from above (e.g., frogs, birds) but also by submerged predators (fish). One of the best studied of the fishing spiders, D. triton (Wal- ckenaer 1837), has two well-known responses to danger from above: it either disappears un- der the water surface by climbing downward on submerged vegetation (McAlister 1959; pers. obs.) or it rapidly gallops away across the water surface (Suter & Wildman 1999; Gorb & Barth 1994). We and others (G. Mill- er, pers. comm.; Suter 1999) have observed that, when startled by sudden water-borne vi- brations while at rest on the water surface, these spiders jump vertically and then either gallop away or return to rest. Our working assumption is that the jump functions to de- crease the probability of capture by fish. Is this a reasonable assumption? At first glance, a vertical jump from the wa- ter surface would seem to be ineffective as evasive behavior because the spider would land exactly where it started, presumably ex- actly where the attacking fish had aimed (Fig. 1). But fish (e.g., trout, Oncorhynchus spp.) accelerate rapidly when lunging at prey (Do- menici & Blake 1997), and rarely do so from directly below their intended victims. Thus an attacking fish usually has a non-vertical tra- jectory, and a vertical jump by a spider would be effectively evasive if it began in time, were high enough, and were of long enough dura- tion. The height and duration of a jump, closely linked to each other by the physics of gravi- tation, depend upon the acceleration the spider can impart to itself by pushing down against the substrate (water, in this case). In earlier studies (Suter et al. 1997; Suter 1999; Suter & Wildman 1999) our laboratory established that the locomotion of fishing spiders on the water surface is based on fluid drag: during horizontal rowing, for example, the dimples in the water surface (caused by the downward push of the spider’s hydrophobic legs) move backward as the spider strokes, encounter re- sistance due to drag, and thereby impart a for- ward acceleration to the spider. In the current study, we looked closely at the forces in- volved in jumping because, as with rowing, the forces generated by the interactions of spi- 201 202 THE JOURNAL OF ARACHNOLOGY Figure 1. — Diagram of the evasion model. See text for explanation. der motions and water ultimately govern the qualities of a jump. METHODS Spiders. — Dolomedes triton are found throughout temperate North America where they inhabit the edges of ponds and slowly flowing streams (Gertsch 1979). The subjects for this study were collected from small ponds in Mississippi and were held in our laboratory under conditions described elsewhere (Suter et al. 1997). Voucher specimens are deposited with the Mississippi Entomological Museum at Mississippi State University. All observa- tions and experiments were conducted at lab- oratory temperatures between 20-23 °C dur- ing local daylight hours. Kinematics of jumping. — The arena for our studies was a 38 liter aquarium filled to a depth of approximately 4 cm with distilled water. In a trial, we placed the test spider in the arena and used a slender glass rod or gen- tle puffs of air to maneuver the animal into the center of the arena. At that position its body was at the point of sharpest focus of ei- ther a high-speed video camera (Kodak model EktaPro EM- 1000) or a 35 mm single lens reflex still camera (Nikon N70). We elicited a jump by sharply hitting the table that sup- ported the aquarium at a distance approxi- mately 2 cm from one comer of the aquarium (approximately 0.3 m from the spider). The images from the high-speed video were collected at 1000 images per second and stored in S-VHS format (Sony model 9500 MDR). We analyzed the spider’s motion in the vertical plane by displaying each digitally- paused video frame on top of a computer-gen- erated x-y cursor grid (NIH Image software, version 1.55 f) by means of a video scan con- verter (Digital Vision, Inc., model TelevEyes/ Pro, connected to an Apple Corporation com- puter, Power Macintosh 7100/80AV). We then manually digitized the coordinates of the body’s approximate center of mass (midway between the top and the bottom of the poste- rior margin of the cephalothorax) every 5 ms for the duration of a jump, and used the co- ordinates to calculate the displacement of the spider through time; velocity was calculated for each 5 ms interval. We selected several trials on which we used the same techniques to digitize the locations of the tips of the legs nearest to the camera. To measure latency, the time between de- livery of the stimulus and the first jumping movements, we placed a 2 X 3 cm mirror in the frame of view of the camera and angled it so that the high-speed video would show both the production of the stimulus and the motion of the spider. The vibrations caused by the stimulus could have reached the spider via transmission through about 0.28 m of water (assuming it propagated from the comer of the aquarium) or through only 4 cm of water (as- suming it propagated first along the glass base of the aquarium and then vertically through the water directly below the spider. In either case, the vibratory stimulus would have reached the spider in < 0.2 ms (given a trans- mission velocity of 1497 m/s in distilled wa- ter; Weast 1985). To make still photographs (35mm) of spi- ders during jumps, we used a percussion sen- sor and an electronic short-interval timer (LPA Design, LPA Time Machine) to trigger an electronic flash (Vivitar, 285HV) after a known post-stimulus delay. Calculation of vertical forces. — We used two methods to calculate the vertical forces produced by spiders during jumping. First (the “acceleration method”), we calculated aver- age force production by applying Newton’s ' second law (F "= ma) to our data on spider mass and acceleration (above). Second (the “leg-motion method”), analyses of the kine- matics of jumping gave us information about the dynamics of the sub-surface portion of the SUTER & GRUENWALD— VERTICAL JUMPING IN DOLOMEDES 203 legs. We assumed that, although the sub-sur- face portion of a leg was not entirely sur- rounded by water, its motion through the wa- ter created a drag force identical to that created by a fully submerged leg segment of the same length and moving at the same av- erage velocity. This is a plausible assumption for two reasons. First, the drag on a sub- merged cylinder is proportional to the frontal surface area (Denny 1993). And second, our earlier work (Suter & Wildman 1999) showed that Denny’s equation for drag on a sub- merged cylinder, which incorporates both drag coefficients and Reynolds numbers (equation 4.29 in Denny 1993), fit the force data for the legs of spiders galloping across the water. We used Denny’s equation to calculate the total thrust force exerted by that leg segment in a direction perpendicular to the leg’s long axis, and used trigonometry to resolve that vector into its horizontal and vertical components. For this study, the horizontal component was ignored because the horizontal forces gener- ated by opposing legs (e.g., left I vs. right IV) are approximately equal in magnitude and op- posite in direction (hence the verticality of the jump). Evasion model.- — The premise underlying our evasion model was that an attack by a fish could be evaded by a fishing spider if the spi- der’s jump occurred at the correct time relative to the attack and if the jump were high enough. In the geometric model (Fig. 1): (a) a fish attacked in a straight line at an angle (a) to the water surface and at a constant ve- locity (V„); (b) throughout the attack, the tra- jectory of the fish was “aimed” at the location of the spider at rest on the water surface; that is, the center of the fish’s open mouth fol- lowed a line that would have, had the spider remained stationary, intersected the center of mass of the spider when the spider was at rest; (c) the spider detected the approach of the fish at a distance 0-2 cm), using sensors on the part of the spider (body or appendages) nearest to the fish, and began its vertical jump with a latency dictated by data collected in this study; (d) the spider jumped to a height (and with a duration) dictated by data collect- ed in this study; (e) the attacking fish, a trout with attack velocities comparable to published fast start velocities of trout (Oncorhynchus mykiss: Domenici & Blake 1997) and with at- tack angles varying between 20-80°, attacked with its mouth open and circular (radius: r, 1- 2 cm); and (f) a successful evasion was de- fined as one in which the spider’s center of gravity was outside of the fish’s mouth > 0) at the moment when the center of the mouth crossed the line representing the ver- tical trajectory of the spider. The attack angle (e, above) was constrained at the lower end by the fact that the spider would be invisible to the fish at angles less than the critical angle of the air- water interface, 48° (Denny 1993); we chose 20° both because we didn’t want to underestimate a fish’s ability to detect surface distortions even when it could not see through the surface, and because the fish’s angle rel- ative to the spider increases as the fish comes close to the spider. At the upper end, the at- tack angle was constrained by the recognition that, as the angle approaches 90°, a spider jumping vertically could not escape even if its jumps were 3X the highest jumps actually measured. The model addressed two questions: for what angles of attack (a) and attack velocities (V^) is spider jumping effective, and how do these parameters compare with actual veloci- ties of attack by fish in the range of angles tested? RESULTS Data from high speed videography. — Videography at 1000 images/sec revealed that a jumping Dolomedes uses all eight legs, ac- celerated simultaneously downward, to propel itself into the air above the water surface (Fig. 2). During the propulsive part of the jump, each leg moves so rapidly (angular velocity = 3.36 ± 1.02 degrees/ms; mean ± 1 S.D.) that an air-filled cavity persists behind it through- out the interaction of leg and water (Fig. 3). The peak height that a spider’s body reaches during a jump is determined primarily by its velocity at the end of the propulsive part of the jump. That velocity, in turn, is a conse- quence of the acceleration produced when a force exerted downward by the spider (= up- ward by the water) moves the mass of the spi- der. Thus, while the spider is in the air, its center of gravity should follow a parabolic path, decelerated by gravity as the spider rises and accelerated by gravity as the spider falls toward the water. In our study, the spider’s paths were nearly perfectly parabolic (Fig. 4), with a characteristic small depression of the 204 THE JOURNAL OF ARACHNOLOGY Figure 3. — Details of the sub-surface shapes of cavities formed during the propulsive phase of jumping by a smaller (0.32 g) female are revealed i in an image captured on 35 mm film with electronic flash illumination (top). The tips of some tarsi pro- trude very slightly into the surrounding water. The ' location of two legs within air-filled cavities can be 'j seen most clearly in the enhanced (bottom) image i- derived from the enlargement (center). j; II I I I I lusc aiiu icii ICIULIVC lu liic uouy uuiiiig a ^ jump (Figs. 2, 4). | The digitization of body height as a func- | tion of time (Fig. 5, upper graphs) made it ;; possible to calculate velocity (Aheight/Atime) and plot velocity as a function of time (Fig. ; 5, lower graphs). During a jump, we used the il downward motion of a leg tip (upper graph, dashed lines) to define the time during which j Figure 2. — High-speed lateral views of Dolo- medes jumping vertically from the water surface. In an analysis of videographic images of a large (0.67 g) female, captured at 1000 frames/sec, the propul- sive phase of the jump was completed within the j first 90 ms, peak elevation was reached at about 134 !! ms, and the spider was out of contact with the water j for about 141 ms. l! SUTER & GRUENWALD— VERTICAL JUMPING IN DOLOMEDES 205 Figure 4. — Digitized tracks of a spider’s approx- imate center of mass (solid line, “ — ”), the tarsus of a leg I (filled circles, and the tarsus of a leg IV (open circles, “O”), during a typical vertical jump from the water surface. The trajectory of the spider’s center of mass follows nearly perfectly the parabola (dashed line, “ — ”) expected from gravi- tational mechanics. The sub-surface locations of the tarsi, during the initial 0.06 sec, indicate the pro- pulsive phase of the jump. propulsive acceleration occurred. For every two adjacent points in these graphs, we cal- culated the change in height as a function of elapsed time (vertical velocity). Plots of ve- locity versus time (lower graphs) showed roughly linear accelerations (slopes) for the propulsive and free-fall phases of the jumps: during propulsion, accelerations were rapid, approximately four times the acceleration of gravity; during freefall, calculated accelera- tions were within 5% of what was expected (9.8 m/s^) for objects under the influence of gravity alone. In the jump of the larger spider, the steep negative acceleration that occurred between 0.015 and 0.045 sec is the result of the spider’s legs rising from below to above its body (see Figs. 2 & 4), causing a rise in the spider’s center of mass without a corre- sponding rise in the position of the spider’s body. Having measured multiple jumps of spiders of five different sizes, we were able to assess performance (i.e., jump height or time in free- fall) as a function of mass. A regression of time in freefall on mass (Fig. 6) revealed no Dolomedes triton 0.233g Dolomedes triton 0.475g Figure 5. — Calculation of the accelerations due to the propulsive actions of the legs and due to gravity during free-fall. Changes in the height of the spider’s center of gravity (upper graphs, solid lines) over time are caused initially by the downward push of the legs (upper graphs, dashed lines) and subsequently by the pull of gravity. 206 THE JOURNAL OF ARACHNOLOGY 0.30-1 ^ 0.25- 0.20- 0.05; duration = 0.167 ± 0.046 sec, mean ± 1 S.D.) and could not be measured in our data from the single male (open circles, “O”). significant relationship. Because time aloft and jump height are physically linked (h = v^t — gfl2, where h is height, g is the acceleration of gravity, and t is time in the air), it follows that jump height is also relatively constant across sizes. This result is consistent with aL lometric measurements of jumping height in terrestrial mammals (Hill 1950; Pennycuick 1992). Latency to jump (the time between the de- livery of the stimulus and the first detectable downward movements of the spider’s legs) did vary significantly with spider mass (Fig. 7): the largest spiders we tested were about 33% slower to respond than the smallest. Because jump height and time in the air were approximately uniform across spider siz- es (Fig. 6) and because larger spiders have more mass to accelerate, we assumed that the forces exerted by spiders during jumping would rise linearly with mass. This assump- tion was confirmed by our measurement of the force/leg used by spiders jumping vertically (Fig. 8, upper graph). The force used to ac- celerate a spider upward (’acceleration meth- od”) rose significantly with mass (upper graph, solid line: for the pooled sexes, force = 4.52 mass + 0.039, r^ = 0.970, n = 5, P < 0.01). To investigate the contribution that surface tension may make to the water’s resistance to the motion of the legs (and hence the spider’s ability to push off from the water surface), we Figure 7. — Latency, the time between the stim- ulus and the first propulsive motions of the legs, rose significantly with mass (pooled sexes; latency = 0.006 mass -h 0.008, P = 0.967, n = 5, P < 0.01). made calculations based upon the following premises: (a) the eight legs contribute equally to the support and vertical propulsion of the spider; (b) about half of each leg is in contact with the water during the propulsive phase of jumping (Fig. 2); (c) leg length is predictably related to spider mass (Suter & Wildman 1999); (d) maximum dimple depth is 3.8 mm (Suter & Wildman 1999). Our calculations, using vertical forces de- rived from the “acceleration method,” re- vealed (Fig. 8, Table 1) that spiders of mass < 0.3 g could become airborne by simply pushing against the resistance caused by the dimples’ combination of surface tension and buoyancy (curved, dashed line in Fig. 8, upper graph). Larger spiders, however, had to rely on drag resistance to generate the force nec- essary to propel them vertically. This differ- ence in the importance of surface tension was also apparent in our force calculations, using vertical forces derived from the “leg-motion method,” concerning the jumps of a 0.05 g spider and a 0.75 g spider (Fig. 8, lower graphs). The vertical component of the force vector produced by the propulsive parts of the legs varied strongly with the angle of each leg relative to horizontal: as a leg approached 90°, the proportion of the force it could generate in a vertical plane approached zero. Not sur- prisingly, therefore, most of the useful force generation during a jump occurred when the legs were moving fast enough (e.g., not at the very beginning of a downward stroke) and were not at too steep an angle. For the larger of these spiders, the “submerged” portion of SUTER & GRUENWALD— VERTICAL JUMPING IN DOLOMEDES 207 Figure 8. — Upper graph: The force used to accelerate a spider upward (“acceleration method”) rose significantly with mass (solid line, r^ = 0.970, n = 5, P < 0.01), but only for spiders < 0.3 g was that force available from pushing against the resistance of the dimple (curved, dashed line). Lower graphs: The vertical component of the force vector produced by the propulsive parts of the legs (“leg motion method”) varies strongly with the angle of each leg relative to horizontal. Horizontal dashed lines represent the average force (“acceleration method”) required for a spider of the given mass to perform a jump of average height and duration. Vertical dashed lines in the upper graph mark the masses of the two spiders depicted in the lower graphs. each leg (the part visible below the water sur- face in Fig. 3) made a major contribution to vertical force needed for a jump, whereas for the small spider, the submerged portion of the leg made a very small contribution. Spider jumps in the context of fish strikes. — -Our geometrical model (Fig. 1), de- signed to assess the efficacy of the vertical jump as a fish evasion behavior, combined our data on jump kinematics and latency with published data on trout fast start velocities (Domenici & Blake 1997). In the model, a successful evasion was one in which the spi- der’s center of mass was outside of the fish’s mouth as the strike trajectory of the fish crossed the jump trajectory of the spider. When we plotted evasion distance (d^^, cm) as a function of the angle of attack (a, degrees) and attack velocity (V^, m/s), taking all ^ev, > 0 as successful evasions, we found that even large variations in detection distance (0-=2 cm) and spider size (0.06-1.0 g) did not render the spider safe at steep angles of attack or at strike velocities > 1 m/s (Fig. 9). The maximum fast start velocities of trout (Domenici & Blake 1997), averaging 1.66 ± 0.48 (S.D.) m/s, are higher than the strike ve- locities at which spiders jumping vertically are safe (Fig. 10). Assuming that strikes have ap- proximately the same peak velocities as fast starts, we conclude that only the most lethargic strikes by trout could be evaded by spiders. 208 THE JOURNAL OF ARACHNOLOGY (/) O) E "w ra w? o c Increasing detection distance Figure 9. — Success of attack evasion. Evasion was deemed successful if the center of mass of the spider was outside of the trajectory of the trout’s mouth as the trout passed through the vertical trajectory of the spider. Thus, all positive values of evasion distance (vertical axis) constitute successful evasions. The evasion distance cm) was plotted as zero for negative values of (spider captured) to emphasize the difference between successful evasion and failure. In all situations, the velocity of the attacking fish and its angle of attack strongly influenced the efficacy of evasive jumping. Increases in the distance at which fish attacks could be detected, d^^, substantially increased the evasion success footprint, and increases in spider size had a similar effect. Table 1. — Jumping from the water surface requires an upward force sufficient both to resist the down- ward pull of gravity and to accelerate the spider upward. Only for small spiders is the resistance offered by a dimple (surface tension plus buoyancy) sufficient for both (compare last two columns). Values in the last four columns are for a single leg and assume that all eight legs participate in vertical propulsion, that about half of each leg provides thrust, and that maximum dimple depth is 3.8 mm. Column 2, from equation 2, Suter & Wildman 1999; column 4, from regression in Fig. 8; column 6, from Suter & Wildman 1999. Spider mass g Estimated leg length mm Force required for static support mN Force required for jumping mN Total resistive force required mN Force available from dimple mN 0.050 11.5 0.061 0.265 0.326 0.875 0.150 17.8 0.184 0.717 0.900 1.271 0.250 20.8 0.306 1.169 1.475 1.437 0.350 22.7 0.429 1.620 2.049 1.555 0.450 24.1 0.551 2.072 2.623 1.662 0.550 25.3 0.674 2.524 3.198 1.733 0.650 26.3 0.796 2.976 3.772 1.792 0.750 27.1 0.919 3.428 4.346 1.816 SUTER & GRUENWALD— VERTICAL JUMPING IN DOLOMEDES 209 Fish Attacks Combination Figure 10. — Evasion success (top) and the actual attack dynamics of trout (middle), when combined, reveal a very small area of overlap on the velocity vs. angle-of-attack plane (bottom). This example depicts results for a 0.25 g spider with a 2 cm detection distance. DISCUSSION We began this study under the assumption that vertical jumps from the water surface by D. triton function to decrease the probability of capture by fish attacking from below. Im- plicit in our assumption was the presumed role that natural selection had played in shaping both jump latency and jump height (and du- ration), with the result that vertical jumping from the water surface, as currently practiced by fishing spiders, would be an effective eva- sive behavior. We have demonstrated, on the contrary, that jumping could save spiders in only a very small fraction of attacks by fish (Fig. 10). At the root of this ineffectual ca- pacity are the size-independent maximum height of jumps (about 3.67 cm) and their cor- respondingly brief duration (about 0.17 sec), and at the root of the limited jump height is the quality of the interaction between the spi- ders' legs and the water. Fluid drag ultimately provides the resis- tance against which the spider pushes (Fig. 8; Suter & Wildman 1999). It follows that ana- tomical modifications to a spider’s legs such as lateral expansions (via hairs or cuticular shape changes), which would expand the area of the surface perpendicular to the direction of the legs’ motion during jumping, would in- crease drag and allow a more rapid upward acceleration of the spider. The more rapid ac- celeration would cause both jump height and jump duration to rise and would render the spider less vulnerable to predation by fish. The absence of such expansions suggests (a) that predation by fish constitutes a relatively mild selective force on these fishing spiders, (b) that contrary selective pressures (e.g., those fostering efficient rowing or prey cap- ture) prevail, or (c) that lateral expansion is phylogenetically constrained. We have no data that allow us to discriminate among these three possibilities and note that any or all of them could operate simultaneously. ACKNOWLEDGMENTS We thank Patricia Miller, Gail Stratton and Edgar Leighton for providing us with the spi- 210 THE JOURNAL OF ARACHNOLOGY ders used in this study, Erin Murphy for some of the data collection and analysis, and John Long for the use of the high-speed videogra- phy equipment (purchased by JL under grant #N000 14-97- 1-0292 from the Office of Naval Research). The study was supported in part by funds provided by Vassar College through the Undergraduate Research Summer Institute and the Class of ’42 Faculty Research Fund, LITERATURE CITED Bames, W.J.R & EG. Barth. 1991. Sensory control of locomotor mode in semi-aquatic spiders. Pp. 105-116, In Locomotor Neural Mechanisms in Arthropods and Vertebrates (D.M. Armstrong & B.M.H. Bush, eds.) Manchester Press, Manches- ter. Denny, M.W. 1993. Air and Water: The Biology and Physics of Life’s Media. Princeton Univ. Press, Princeton. Domenici, P. & R.W. Blake. 1997. The kinematics and performance of fish fast-start swimming. J. Exp. Biol., 200:1165-1178. Gertsch, W.J. 1979. American Spiders (2nd Ed.). Van Nostrand Reinhold Company, New York. Gorb, S.N. & EG. Barth. 1994. Locomotor behav- ior during prey-capture of a fishing spider, Do- lomedes plantarius (Araneae: Araneidae): Gal- loping and stopping. J. Arachnol., 22:89-93. Hill, A.V 1950. The dimensions of animals and their muscular dynamics. Sci. Prog., 38:209- 230. McAlister, W.H. 1959. The diving and surface- walking behaviour of Dolomedes triton sexpunc- tatus (Araneida: Pisauridae). Anim. Behav., 8: 109-111. Pennycuick, C.J. 1992. Newton Rules Biology: A Physical Approach to Biological Problems. Ox- ford Univ. Press, Oxford. Shultz, J.W. 1987. Walking and surface film loco- motion in terrestrial and semi-aquatic spiders. J. Exp. Biol., 128:427-444. Suter, R.B. 1999. Walking on water. American Sci., 87:154-159. Suter, R.B., O. Rosenberg, S. Loeb, H. Wildman & J.H. Long, Jr. 1997. Locomotion on the water surface: Propulsive mechanisms of the fisher spi- der Dolomedes triton. J, Exp. Biol., 200:2523- 2538. Suter, R.B. & H. Wildman. 1999, Locomotion on the water surface: Hydrodynamic constraints on rowing velocity require a gait change. J. Exp. Biol., 202: 2771-2785. Weast, C. 1985. CRC Handbook of Chemistry and Physics. CRC Press, Boca Raton. Manuscript received 10 January 2000, revised 5 May 2000. 2000. The Journal of Arachnology 28:211-216 PREDATORY INTERACTIONS BETWEEN MUD-DAUBER WASPS (HYMENOPTERA, SPHECIDAE) AND ARGIOPE (ARANEAE, ARANEIDAE) IN CAPTIVITY Todd A. Blackledge^ and Kurt M. Pickett: Department of Entomology, The Ohio State University, 1735 Neil Avenue, Columbus, Ohio 43210 USA ABSTRACT. We report on efforts to maintain two common sphecid wasps, Chalybion caeruleum (Saus- sure 1867) and Sceliphron caementarium (Drary 1773), in field and laboratory enclosures in order to observe their predatory interactions with the orb- weaving spiders Argiope aurantia Lucas 1833 and A. trifasciata (ForskM 1775). Both species of wasps seemed to locate webs primarily by chance while flying along the tops of the vegetation but differed greatly in their hunting tactics once webs were located. Sceliphron caementarium was most successful at capturing spiders that had dropped out of webs in re- sponse to the wasp’s hitting the web. But, C. caeruleum often employed a type of aggressive mimicry: it landed in the web or used its middle legs to pluck the web, luring the spider to the wasp. Argiope did not differ in their defensive response to C. caeruleum and S. caementarium. Most Argiope dropped out of webs in response to attacks rather than using other defensive behaviors such as shuttling between sides of webs or vibrating webs. Keywords: Sympatry, competition, niche partitioning Sphecid wasps are common predators of orb-weaving spiders. Because individual wasps capture several spiders to provision each cell in a nest and build multiple cells over their lives (Coville 1987), mud-dauber wasps can act as a particularly potent selective force on the evolution of spider defensive be- haviors. Many studies have examined the numbers and species of spiders provisioned in wasp nests, providing insight into which spi- ders may be most vulnerable to wasps (e.g., Muma & Jeffers 1945 and references in Krombein et al. 1979). These studies indicate that different species of wasps that hunt in the same habitat, such as Chalybion caeruleum and Sceliphron caementarium, often catch dif- ferent prey. This suggests that sympatric spe- cies of sphecids may employ different preda- tory tactics, perhaps due to niche partitioning. There are few, mostly anecdotal, observations on the hunting tactics of sphecids (Peckham & Peckham 1905; Rau 1928, 1935; Eberhard 1970; Endo 1976; Coville 1987; Rayor 1997). But, there has been no comparative study of the hunting behaviors of sympatric C. caeru- leum and S. caementarium. I ' Current address: Insect Biology-ESPM, 201 I Wellman Hall, University of California, Berkeley. , Berkeley, California 94720-3 1 12 USA. Little is known about the primary and sec- ondary defensive behaviors orb-web spiders use against sphecids. Yet, it is the interaction of spider defensive behaviors and the preda- tory tactics of wasps that determine if indi- vidual spiders survive predation attempts (Cloudsley-Thompson 1995; Edmunds & Ed- munds 1986; Tolbert 1975). There are two de- tailed studies of wasp-spider interactions, but these focus on wasps hunting nocturnal or co- lonial orb- weaving spiders (Eberhard 1970; Rayor 1997). What is missing, therefore, are studies of the interactions of wasps with sol- itary, diurnal spiders, such as Argiope. Argiope is among the most intensively stud- ied genera of spiders and is likely to be par- ticularly vulnerable to visually-hunting pred- ators because it rests at the center of its web during daylight. Argiope is also an important model for testing hypotheses concerning pos- sible defensive functions of structures such as barrier webs (Higgins 1992) or stabilimenta (Blackledge & Wenzel 1999). Here we report on our efforts to maintain two species of sphe- cid wasps (C. caeruleum and S. caementar- ium) in field and laboratory enclosures and our observations of their predatory interac- tions with the orb- weaving spiders Argiope aurantia and A. trifasciata. 211 212 THE JOURNAL OF ARACHNOLOGY METHODS We observed the hunting behaviors of C. caeruleum and S. caementarium in one indoor enclosure (1998 and 1999) and three outdoor enclosures (1999). All wasps were collected as adults in the field (Dublin, Ohio), except for a single C. caeruleum that emerged from a previously collected nest during the 1998 study. The collection site consisted of old bams surrounded by old fields. The primary prey caught by wasps at this site were im- mature A. trifasciata (pers. obs.). Individual wasps were distinguished by paint on the tho- rax or abdomen. The 3.4 X 2.7 X 2.2 m screened indoor enclosure was located in Ohio State Univer- sity’s Insectary, Columbus, Ohio, in a green- house room with light and temperature main- tained near outdoor levels. Assorted plants, including flowering Echinacea (Asteraceae) and Lantana (Verbenaceae), were scattered throughout the enclosure to provide resting places for wasps. The plants also simulated the natural background of foliage in which wasps hunt spiders, a potentially important feature of the study because background may influence the conspicuousness of spider silks to insects (Blackledge 1998a; Blackledge & Wenzel 2000). A 20 X 30 cm plastic pan was placed in one comer of the enclosure and con- tained a layer of earth from the same pond at the field site where wild S. caementarium col- lected mud for their nests. The pan was par- tially filled with water and then tilted to create a moisture gradient from completely saturated to nearly dry, simulating the bank of the pond. Mud nests of S. caementarium, collected at the field site, were glued to wooden boards in the upper comers of the enclosure to encour- age building of new nest cells by S. caemen- tarium. These nests also provided vacant cells for C. caeruleum, which nests only in aban- doned S. caementarium cells (Rau 1928). In 1998, petri dishes containing a sucrose and honey mixture were placed on the floor of the cage to provide wasps with a nectar source. In 1999, a plastic hummingbird feeder filled with a 1:1 honey:water solution was used in- stead. The honey water was changed every two days to prevent fermentation. The three outdoor enclosures consisted of nylon screening over wood frames (3.8 X 2.3 X 2.0 m) and were located in a field at Ohio State University’s Rothenbuhler Honeybee Laboratory, Columbus, Ohio. We found it necessary to cover the bottom edge of the screening with thick layers of bark mulch and stone to prevent wasps from crawling under the edges of the enclosures. The natural ground cover consisted of various grasses (Po- aceae) and thistle (Asteraceae), with a thick layer of thatch. There were some naturally oc- curring A. trifasciata in the surrounding field. Again, each enclosure had a 20 X 30 cm plas- tic pan containing mud and water, wooden boards with mud S. caementarium nests glued to them, and a hummingbird feeder as a nectar source. Immature A. aurantia and A. trifasciata were collected from roadside ditches in and around Columbus. Most of the spiders were uniquely marked and weighed immediately af- ter collection. Spiders were allowed to build their webs in 35 X 35 X 10 cm wooden frames as described in Blackledge (1998b) but modified with both plastic sides being remov- able. We placed individual frames containing spiders within the enclosures to observe wasp- spider interactions. We recorded our obser- vations on audio tape and also video-taped a few of the encounters. We also include some observations on A. trifasciata, in webs on nat- ural plant supports, which we placed in the same outdoor enclosures and one of us (TAB) used for a second study examining the role of stabilimenta as wasp defenses. We released a variety of araneid, linyphiid and tetragnathid spiders into the indoor enclosure to provide alternative prey, while the outdoor enclosures naturally contained a variety of agelenids, sal- ticids and thomisids as well as Cyclosa conica (Pallas 1772) and Uloborus glomosus (Wal- ckenaer 1841), Because we later found few individuals of these species in wasp nests (10 of 142 excavated spiders) and we never di- rectly observed a predation event involving these species, we exclude them from further discussion. RESULTS In the indoor enclosure, we observed 24 at- | tempted predation events during 20 days of > observation (between 4-28 August 1998 and | between 28 July- 17 August 1999). In the out- door enclosures, we observed 50 predation at- tempts during observations every day between 21 August and 11 September 1999. Chalybion BLACKLEDGE & PICKETT— WASP VERSUS SPIDER 213 Table 1. — Predatory tactics of two species of sphecid wasp, C caeruleum and S. caementarium, and the common defensive responses by immature A. aurantia and A. trifasciata. Observations were made on 3 individuals of C caeruleum and 5 individuals of S. caementarium. The heading “Spider approached wasp” includes approaches by spiders to either wasps landing in webs or plucking webs. Defensive responses of spiders were not mutually exclusive. Asterisks denote significant differences, using binomial probability, between species of wasps in frequency of behaviors (*P < 0.05, **P < 0.01, ***P < 0.005). C. caeruleum S. caementarium Observed attacks 48 26 Location of capture: web center 6 3 capture zone or frame threads 14 6* ground below web 4 Total 24 20 Wasp landed in web 22 Wasp plucked web 11 0** Spider approached wasp 15 Response of spider: drop from web 21 15 abandon web 7 7 move to web periphery 15 6 Mass of spiders captured: mean ± standard deviation 0.04±0.01 mg 0.04±0.02 mg range 0.02-0.07 mg 0.02-0.08 mg caeruleum opened their nests and began hunt^ ing between 1000-1200 h and resealed their nests between 1400-1700 h or, if no spiders were captured, after only 30 min. Sceliphron caementarium typically opened nests for the entire day (1000-1700 h). Like other sphe- cids, both C. caeruleum and S. caementarium often did not hunt on overcast, rainy days and became active much later than normal on cooler days (see also Freeman & Johnston 1978; Powell 1967). Encounters were some- times brief— lasting only a few seconds if spi- ders were caught at the centers of webs, and sometimes much longer, lasting 2-3 min if spiders attempted to escape by dropping and then moving rapidly through the grass. We combined all of the data for each species of wasp (Table 1) and, within each species, we had approximately the same number of obser- vations for each individual wasp. We only in- cluded observations on predation attempts on spiders that were within the size range cap- tured by wasps during the experiment (Table 1). Both wasp species seemed to locate webs by chance while flying along the top of the vegetation in a seemingly haphazard flight path. However, S. caementarium and C. ca- eruleum differed greatly in their hunting tac- tics once webs were located (Table 1). Sceli- phron caementarium bumped into webs while flying, but then flew off without seeming to react to webs as anything other than physical barriers. But, these wasps vigorously pursued spiders that dropped from webs, spending as much as 2-3 min crawling around the thatch and grass stems under webs in gradually en- larging circular patterns until either spiders were located or wasps began flying again. In contrast C. caeruleum often landed in webs or on the substrate supporting webs and then used their middle legs to pluck the silk. When in a web, C. caeruleum sometimes con- tracted its entire body every few seconds for up to two minutes. In 68% of these instances, spiders ran to wasps after wasps had landed in or plucked at webs. Many of these spiders (70%) were caught as they approached wasps or as wasps chased them back to the centers of webs, but others immediately dropped out of webs upon contacting wasps. Captured spiders were stung between the carapace and sternum in the posterior of the cephalothorax. Paralysis appeared to be in- 214 THE JOURNAL OF ARACHNOLOGY stantaneous, but spiders were occasionally stung multiple times, stings lasting up to a few seconds. Wasps carried spiders by holding the pedipalps in their mandibles, with the venters of spiders facing toward the venters of wasps. Wasps commonly pressed their mandibles against the chelicerae of spiders for a few sec- onds after capture, perhaps drinking hemo- lymph. After about 25% of captures, both spe- cies of wasp drank hemolymph from the chelicerae or coxae of spiders for periods of up to 1 min. Four of those spiders were sub- sequently discarded instead of being used to provision a nest. We observed 9 instances (not in Table 1) where a wasp attacked a spider, grasped the spider with its legs, wrapped its abdomen around the spider as though stinging it, but then released the spider and flew away. In each instance the spider was still alive and ran away when touched by one of us. All but two of those spiders weighed within the mean ±2 standard deviations of Argiope captured dur- ing the study. DISCUSSION Eberhard (1970) concluded that contrast be- tween a spider and the background upon which it rested was one of the most important cues used by S. caementarium to locate Lar- inioides (Araneus) cornutus (Clerk 1757), which were hiding in retreats near webs. In our study, both C. caeruleum and S. caemen- tarium often alighted upon dark spots of de- bris or the shadows of insects or spiders on the opposite side of the screen tent, which supports Eberhard’s hypothesis that wasps re- spond to contrast. However, S. caementarium attacked very few spiders at the centers of webs, instead seeming to stumble into and out of webs without regard for the possible pres- ence of spiders. Ccalybion caeruleum and S. caementarium often flew within 2 cm of spi- ders on webs or grass, without reacting to the spiders, but quickly chased spiders once spi- ders dropped from or moved within webs. Both of these observations suggest that con- trast was not actually used to locate Argiope in our study. There are at least two potential explanations for this difference with Eber- hard’s findings. The light-colored bodies of ju- venile Argiope may reflect significant UV light (Craig & Ebert 1994), and this may pro- vide a poor contrast against natural back- grounds to insects, much as stabilimentum silk can (Blackledge 1998a; Blackledge & Wenzel 2000). Another possible explanation is that motion may be an important cue in eliciting attacks by S. caementarium. This second ex- planation seems particularly likely because S. caementarium pounced on small moving in- sects or even falling debris, particularly when wasps were searching for spiders flushed from webs. Sceliphron caementarium aggressively pur- sued spiders that dropped from webs, catching most prey by chasing spiders on the ground, while C. caeruleum used aggressive mimicry to catch spiders that were still in webs (Table 1). Chalybion caeruleum landed in webs and then plucked at the silk in webs, luring spiders to themselves. In almost 70% of encounters where C caeruleum landed in or plucked webs, spiders approached wasps; and most of those spiders were captured with little chase. We even observed one instance where a spi- der, which had dropped out of its web into the grass, proceeded to crawl back up its dragline to the web center and then to a C. caeruleum as the wasp plucked the web. This plucking behavior is similar to that described for Chal- ybion spp. (Schwarz, in Howard 1901; Coville 1976) and Trypoxylon sp. (Rau 1926; pers. obs.) and may be a particularly effective method to hunt retreat dwelling spiders (Co- ville 1976). One vespid is also thought to use vibrations caused by tapping with its antennae to lure spiders to the hubs of webs (MacNulty 1961). Sceliphron caementarium nests contain a wider range of spider prey than the nests of C. caeruleum. Sceliphron caementarium pro- visions nests with both web-building and cur- sorial spiders, while the nest contents of C. caeruleum are largely restricted to orb and tangle web-building spiders (Krombein et al. 1979; Muma & Jeffers 1945). These differ- ences in nest provisioning likely reflect the different hunting tactics used by these two species of wasps. The use of old Sceliphron nests by Chalybion (Rau 1928) restricts Chal- ybion to hunting in habitats occupied by Sce- liphron. Thus, competition has likely been an important selective factor in the evolution of Chalybion and Sceliphron hunting behaviors. Therefore, the specialization on web-building spiders by Chalybion could be due to niche partitioning. BLACKLEDGE & PICKETT— WASP VERSUS SPIDER 215 Argiope used similar defensive behaviors against both species of wasps (Table 1). The most common response to attacks was for spi- ders to drop from webs (50% of encounters) and then either freeze or run to nearby cover. Spiders often maintained contact with their webs via draglines and returned 2-10 min lat- er. But, spiders sometimes abandoned webs completely, moving up to 1 m away, in deep grass. Argiope trifasciata on natural webs built in the grassy outdoor enclosures also sometimes abandoned webs when attacked. They would then build webs in new locations the next day, without having consumed the abandoned web. These observations suggest that field researchers should use caution when assuming that abandoned webs always indi- cate predation, because abandoning webs is itself a defensive strategy. Occasionally a spider ran to the top or side of its web (30% of encounters), remaining motionless for up to several minutes before returning to the web center. Spiders that re- mained at web hubs often stilted, holding their bodies far out from webs and angling their abdomens away from the plane of webs. We suggest that these defensive behaviors might be relatively specialized responses to wasp predators (see also Cushing & Opell 1990), because spiders did not engage in other com- mon defensive behaviors such as web flexing or shuttling (Cloudsley-Thompson 1995; Ed- munds & Edmunds 1986; Tolbert 1975). Web flexing is often initiated when humans ap- proach webs (pers. obs.) and may function against salticid predators (Tolbert 1975) but was never used against wasps. While our ob- servations supplement descriptive works on the behavioral interactions of wasps and spi- ders, we hope that the use of enclosures will also facilitate a more experimentally-based approach to the study of wasp- spider interac- tions. ACKNOWLEDGMENTS J.W. Wenzel provided critical support and advice during the study. We thank everyone at the OSU Insectary, Rothenbuhler Honeybee Laboratory, and T.C. Jones for use of their en- closures and other supplies. Dr. D. Bunner kindly allowed us to spend many afternoons collecting wasps on his farm. We also appre- ciate many helpful comments on this manu- script from an anonymous reviewer, P. Sier- wald, and R. Suter. This study was supported by funding (to TAB) from the American Ar- achnological Research Fund, an Animal Be- havior Society Research Grant, a Graduate Student Alumni Research Award and a Pres- idential Fellowship from Ohio State Univer- sity, a Grant-in- Aid of Research from the Na- tional Academy of Sciences, through Sigma Xi, and a National Science Foundation Grad- uate Research Fellowship. LITERATURE CITED Blackledge, T.A. 1998a. Signal conflict in spider webs driven by predators and prey. Proc. Roy. Soc. London, B., 256:1991-1996. Blackledge, T.A. 1998b. Stabilimentum variation and foraging success in Argiope aurantia and Ar- giope trifasciata (Araneae: Araneidae). J. Zool., 246:21-27. Blackledge, T.A. & J.W. Wenzel. 1999. Do stabi- limenta in orb webs attract prey or defend spi- ders? Behav. EcoL, 10:372-376. Blackledge, T.A. & J.W. Wenzel. 2000. The evo- lution of cryptic spider silk: a behavioral test. Behav. Ecol., 11:142-145. Cloudsley-Thompson, J.L. 1995. A review of the anti-predator devices of spiders. Bull. British Ar- achnol. Soc., 10:81-96. Coville, R.E, 1976. Predatory behavior of the spi- der wasp, Chalybion calif ornicum. Pan-Pacific EntomoL, 52:229-233. Coville, R.E. 1987. Spider-hunting sphecid wasps. Pp. 309-318, In Ecophysiology of Spiders. (W. Nentwig, ed.). Springer, New York. Craig, C.L. & K. Ebert. 1994. Colour and pattern in predator-prey interactions: The bright body colours and patterns of a tropical orb-spinning spider attract flower-seeking prey. Func. Ecol., 8: 616-620. Cushing, RE. & B.D. Opell. 1990. Disturbance be- haviors in the spider Uloborus glomosus (Ara- neae, Uloboridae): Possible predator avoidance strategies. Canadian J. Zool., 68:1090-1097. Eberhard, W, 1970. The predatory behavior of two wasps, Agenoideus humilis (Pompilidae) and Sceliphron caementarium (Sphecidae), on the orb weaving spider Araneus comutus (Aranei- dae). Psyche, 77:243-251. Edmunds, M. & J. Edmunds. 1986. The defensive mechanisms of orb weavers (Araneae: Aranei- dae) in Ghana, West Africa. Pp. 73-89, In Proc. Ninth Intern. Congr. Arachnol., Panama 1983. (W.G. Eberhard, Y.D. Lubin & B.C. Robinson, eds.). Smithsonian Inst. Press, Washington, D.C. Endo, A. 1976. Factors influencing the prey selec- tion of a spider wasp, Episyron arrogans (Smith) (Hymenoptera: Pompilidae). Physiol. Ecol., 17: 335-350. 216 THE JOURNAL OF ARACHNOLOGY Freeman, B.E. & B. Johnston. 1978. The biology in Jamaica of the adults of the sphecid wasp Sce- liphron assimile Dahlbom. Ecol. Entomol., 3:39- 52. Higgins, L.. 1992. Developmental changes in bar- rier web structure under different levels of pre- dation risk in Nephila clavipes (Araneae: Tetrag- nathidae). J. Insect Behav. 5:635-655. Howard, L.O. 1901. The Insect Book. Doubleday, Page & Co., New York. Krombein, K.V., P.D. Hurd, D.R. Smith & B.D. Burk. 1979. Catalog of Hymenoptera in Amer- ica North of Mexico. Smithsonian Inst. Press, Washington, D.C. MacNulty, B.J. 1961. Field observations on some West African Hymenoptera. Proc. S. London En- tomol. Nat. Hist. Soc., 1961:118-133. Muma, M.H. & W.F Jeffers. 1945. Studies of the spider prey of several mud-dauber wasps. Ann. Entomol. Soc. America, 38:245-255. Peckham, G.W & E.G. Peckham. 1905. Wasps So- cial and Solitary. Houghton, Mifflin and Co., Boston. Powell, J.A. 1967. Behavior of ground nesting wasps of the genus Nitelopterus, particularly N. califomicus (Hymenoptera: Sphecidae). J. Kan- sas Entomol. Soc., 40:331-346. Rau, P. 1926. The ecology of a sheltered clay bank; a study in insect sociology. Trans. Acad. Sci. St. Louis, 25:121-276. Rau, P. 1928. The nesting habits of the wasp, Chal- ybion caeruleum. Ann. Entomol. Soc. America, 21:25-35. Ray or, L.S. 1997. Attack strategies of predatory wasps (Hymenoptera: Pompilidae; Sphecidae) on colonial orb web-building spiders (Araneidae: Metepeira incrassata). J. Kansas Entomol. Soc., 69(4) suppl., 67-75. Shafer, G.D. 1949. The Ways of the Mud-dauber. Stanford Univ. Press, Stanford. Tolbert, WW. 1975. Predator avoidance behaviors and web defensive structures in the orb weavers Argiope aurantia and Argiope trifasciata (Ara- neae, Araneidae). Psyche, 29-52. Manuscript received 2 July 1999, revised 20 March 2000. 2000. The Journal of Arachnology 28:217-222 SPIDERS IN ROCKY HABITATS IN CENTRAL BOHEMIA Vlastimil Ruzickai Institute of Entomology, Czech Academy of Sciences, Branisovska 31, 370 05 Ceske Budejovice, Czech Republic ABSTRACT. Spiders of andezite and limestone rocks in Central Bohemia were studied. The material was collected using hanging desk traps. Rocky habitats are inhabited by a well-established spider assem- blage. A lower slope angle, and consequently more diverse terrain, probably support a higher species diversity. Some species inhabit exclusively rocky habitats. Segestria bavarica and Theridion betteni occur in the Czech Republic exclusively on rocky habitats. Erigonoplus jarmilae, Zelotes puritanus, and Altella biuncata appear to occur primarily on rocky habitats. Anyphaena furva appears to live on trunks of trees growing on sun-exposed rocks. Some thermophilous species narrow their ecological niche exclusively to southern exposed rocky habitats with a warm microclimate towards the north. Keywords: Spiders, rocks, vegetation-free habitats. The lowlands of Central Europe were orig- inally covered mostly with closed forests, al- though isolated islet-like natural non-forest habitats occasionally occurred. Today the sur- face of some of these non-forested areas is composed of bare bedrock or products of its erosional breakdown (without a soil layer) and are typified by gravel and sand banks, sand dunes, scree slopes and rock outcrops. Ac- cording to their specific substratum and mi- croclimate, these habitats can harbor special- ized spider species. There has been no comprehensive study of the spiders of gravel banks in the Czech Re- public. However, useful data were obtained during the grid mapping of lycosid distribu- tion by Buchar (1995); and gravel and sand banks are known to harbor several specific in- habitants. Of these, Oedothorax agrestis (Blackwall 1853) is the most common, Par- dosa morosa (L. Koch 1870) occurs sporadi- cally, while Arctosa cinerea (Fabricius 1777) and Arctosa maculata (Hahn 1822) are very rare. There is also a need for a detailed study of the spiders of sand dunes in the Czech Re- public. However, the specific arachnofauna of this habitat is known. Arctosa perita (Latreille 1799), Steatoda albomaculata (De Geer 1778), and Attulus saltator (Simon 1868) rep- resent specialized inhabitants of sand dunes (Miller 1971; Buchar 1995). New investiga- tions of sand dunes in southern Moravia have resulted in several new records for the Czech thermophilous species Republic (Ruzicka 1998): Uloborus walcken- aerius Latreille 1806, Mecynargus foveatus (Dahl 1912) and Titanoeca psammophila Wunderlich 1993. Scree slopes have been intensively studied over the past years, and microclimate condi- tions and spider assemblages have been de- scribed (Ruzicka & Zacharda 1994; Ruzicka et al. 1995). Acantholycosa norvegica sudeti- ca (L. Koch 1875), Bathyphantes simillimus buchari Ruzicka 1988, Lepthyphantes impro- bulus Simon 1929 and Wubanoides uralensis (Pakhorukov 1981) are the most specific in- habitants of boulder accumulations in the Czech Republic (Ruzicka 1996; Ruzicka & Hajer 1996; Ruzicka & Zacharda 1994). Rock faces, rock walls, solitary rock out- crops and rocky slopes in deeply-cut river val- leys remain some of the most inaccessible and unknown habitats. Due to the difficulty of ex- ploiting them economically, these habitats have remained unchanged over the entire Ho- locene. The plant and animal communities in- habiting them represent edaphic climaxes. The aim of this study was to describe and evaluate the species composition of a spider assem- blage in two different rocky habitats in the warmest territory in Central Bohemia. METHODS Study sites.— Both localities studied lie in Central Bohemia, on the border of Thermo- phyticum and Mesophyticum, and both are at similar elevations with similar exposures. Nezabudicke Skaly (rocks) Nature Reserve is 217 218 THE JOURNAL OF ARACHNOLOGY Figure 1. — Traps, as shown above, hanging against the creviced and rough rock surfaces, were used to capture the spiders. They contained a mix- ture of formaldehyde and glycerol. situated in the Kfivoklatsko Biosphere Re- serve, near Nezabudice village, about 50 km west of Prague, elevation 290 m. The rocks are composed of andezite and form a south- easterly exposed amphitheater. The western part of this amphitheater is formed by bare rocks, small scree fields and narrow scree ledges with plant tussocks, shrubs and solitary trees. The rocky slope is about 60 m high with a slope angle of about 45°. Kotyz National Nature Monument is situ- ated in the Bohemian Karst Protected Land- scape Area, near Koneprusy village, about 30 km southeast of Prague, elevation 380 m. The vertical southern exposed rock wall is made of limestone. It is about 40 m high and is part- ly overgrown by plant tussocks. We automated the collecting of spiders on rocks using hanging desk traps (Ruzicka & Antus 1997) (Fig. 1). The traps, made of rigid plastic, consist of a desk (25 X 20 cm, which formed an artificial horizontal surface) and a can (13 cm high and 10.5 cm in diameter) inserted in the center of the desk. The traps contained a mixture of 7% formaldehyde and 10% glycerol with a few drops of a surfactant. Each trap was hung from a hooked nail. A band of emery tape was stuck on the back edge of the trap and shaped to form a con- nection, or transit, between the desk and the rock surface. We hung the traps in marginal. creviced, and rough parts of the rocks in the mosaic of bare rock surface and vegetation tussocks. Six traps were placed in Nezabudi- cke Skaly from May 1996 to April 1997, and in Kotyz from May to October 1996. Five ad- ditional desk traps with a half-circle back mar- gin were hung on old oaks growing at Neza- budicke Skaly from April to June 1997. The material was evaluated with respect to the occurrence of the species in phytogeo- graphical regions and in habitats of various degree of originality (Buchar 1993) (see leg- end of Table 1). The nomenclature follows the check list of spiders of the Czech Republic (Buchar et al. 1995). RESULTS AND DISCUSSION Species diversity. — A total of 218 deter- minable spider individuals belonging to 48 species was collected on rocks at Nezabudicke Skaly. A total of 102 determinable spider in- dividuals belonging to 27 species was col- lected at Kotyz. A total of 30 spider individ- uals belonging to 10 species was collected on tree trunks at Nezabudicke Skaly (Table 1). Fourteen common species and a common dominant species, Drassodes lapidosus, re- flect the similarity of both sites. The localities studied differ in the type of rock and in the slope angle. The higher number of species and individuals at Nezabudicke Skaly is probably caused by a greater diversity of terrain con- ditions resulting in more niche diversity. Low- er slope angle allows the occurrence of small scree fields, small ledges, and consequently rich plant tussocks, solitary shrubs and trees. The frequency of specimens of species oc- curring primarily in natural habitats amounts to 0.41 at Nezabudicke Skaly, and 0.45 at Ko- tyz, which is considerably higher than the minimal value of 0.20 that is characteristic for protected regions in the Czech Republic (Ruz- icka 1987). High frequency of these species indicates original habitats. Rock as a habitat. — Hanggi et al. (1995) distinguish 85 habitat types in their classifi- cation of Central European habitats. They in- clude only the habitat “Alpine rocks” in the category of “Alpine habitats.” The rocks of lower elevations are omitted. However, there are spider species that live occasionally, pri- marily, or exclusively on rocks. Heimer & Nentwig (1991) and also Miller (1971) de- scribed rocks as living habitat for about 20 RUZICKA— SPIDERS ON ROCKS 219 Table L — Survey of material collected (d/9/j) by hanging desk traps on rocks at Nezabudicke Skaly (A), on tree tranks at Nezabudicke SkMy (B), and on rocks at Kotyz (C). T = occurring primarily in Thermophyticum, M = in Mesophyticum, O = in Oreophyticum, N = non-specific; 1 = occurring in natural habitats corresponding to climatic or edaphic climax, 2 = capable of occupying some shadow and wet secondary, semi-natural habitats (cultural forests, shrubs, cultivated wetlands), 3 = capable of forming viable populations in artificially deforested, man-made habitats (fields, meadows, urban habitats). A B c Segestriidae T 1 Segestria bavarica C. L, Koch 1843 3/- - -/i/i Dysderidae N 2 Harpactea hombergi (Scopoli 1763) 1/- 1/9 1/- Eresidae T 1 Eresus cinnaberinus (Olivier 1789) - - 2/- Theridiidae T 1 Dipoena melanogaster (C. L. Koch 1837) - - 1/1 ? 1 Dipoena nigroreticulata (Simon 1879) -/I - - N 2 Episinus truncatus Latreille 1809 - - -n T 1 Theridion betteni Wiehle 1960 -/I - - N 2 Theridion tinctum (Walckenaer 1802) - 1/- - Linyphiidae N 3 Araeoncus humilis (Blackwall 1841) -/I - - N 2 Bathyphantes nigrinus (Westring 1851) -/I - - O 2 Centromerus sellarius (Simon 1884) 1/1 - - N 3 Centromerus sylvaticus (Blackwall 1841) 1/- - - N 3 Dicymbium nigrum (Blackwall 1834) 2l\ - - T 1 Erigonoplus Jarmilae (Miller 1943) 15/5 - - N 2 Lepthyphantes flavipes (Blackwall 1854) 21- -n - T 1 Lepthyphantes keyserlingi (Ausserer 1867) 21- - 2/- N 3 Lepthyphantes mengei Kulczynski 1887 1/- - - N 2 Lepthyphantes pallidus (O. P.-Cambridge 1871) -/I - - N 3 Linyphia triangularis (Clerck 1757) -/I - - N 3 Micrargus subaequalis (Westring 1851) - - 1/- N 2 Microneta viaria (Blackwall 1841) 2/- - - N 1 Panamomops affinis Miller & Kratochvil 1939 l/_ - - Tetragnathidae N 3 Pachygnatha degeeri Sundevall 1830 - - 1/- Araneidae T 1 Gibbaranea bituberculata (Walckenaer 1802) - 1/- - N 3 Mangora acalypha (Walckenaer 1802) -1-12 - -/I Lycosidae T 2 Alopecosa accentuata (Latreille 1817) 5/1 - - T 1 Arctosa figurata (Simon 1876) “ - 1/- ? 1 Pardosa alacris (C. L. Koch 1833) 2/2 - 5/3 T 1 Pardosa bifasciata (C. L. Koch 1834) - - 3/3 T 1 Trochosa robusta (Simon 1876) 1/1 - - M 3 Trochosa ruricola (De Geer 1778) 6/1 - - N 2 Xerolycosa nemoralis (Westring, 1861) 6/- - 4/- Agelenidae O 2 Histopona torpida (C. L. Koch 1834) 2/- - - N 2 Textrix denticulata (Olivier 1789) 6/-/1 — — 220 THE JOURNAL OF ARACHNOLOGY Table 1. — Continued. A B C Dictynidae T 1 Altella biuncata (Miller 1949) 51- - - Amaurobiidae O 2 Callobius claustrarius (Hahn 1833) 11- - - O 2 Coelotes inermis (L. Koch 1855) 41- - - N 2 Coelotes terrestris (Wider, 1834) 1/- - - Titanoecidae T 1 Titanoeca quadriguttata (Hahn 1833) 1/- - 7/2 Anyphaenidae M 1 Anyphaena furva Miller 1967 - 3/- - Liocranidae N 2 Apostenus fuscus Westring 1851 1/- - - N 3 Liocranum rupicola (Walckenaer 1830) - 1/- -1-12 N 2 Phrurolithus festivus (C. L. Koch 1835) - - ll- Clubionidae N 1 Clubiona comta C. L. Koch 1839 - 1/2 - Zodariidae T 1 Zodarion gennanicum (C. L. Koch 1837) 1/1 - - Gnaphosidae T 1 Callilepis schuszteri (Herman 1879) 2/3 - - N 2 Drassodes lapidosus (Walckenaer 1802) 34/4 6/2 37/4 T 1 Drassyllus villicus (Thorell 1875) 1/1 - - N 1 Echemus angustifrons (Westring 1862) 1/- - - T 1 Gnaphosa opaca Herman 1879 5/1 - - T 1 Zelotes erebeus (Thorell 1870) 2/1 1/- - ? 1 Zelotes exiguus (Muller & Schenkel 1895) 5/3 - - T 1 Zelotes puritanus Chamberlin, 1922 7/4 - - Thomisidae M 1 Ozyptila blackwalli Simon, 1875 - - 1/- T 1 Ozyptila nigrita (Thorell 1875) 21- - - M 3 Xysticus kochi Thorell 1872 6/3 - 1/- T 1 Xysticus ninnii Thorell 1872 - - 31- Salticidae T 2 Aelurillus v~insignitus (Clerck 1757) 14/3/2 1/- - ? 1 Heliophanus aeneus (Hahn 1831) 6/- - - T 2 Heliophanus cupreus (Walckenaer 1802) 1/- - 1/- T 1 Pellenes tripunctatus (Walckenaer 1802) - - 1/- T 1 Philaeus chrysops (Poda 1761) 31- - 61- T 1 Phlegra festiva (C. L. Koch 1834) - - 11- N 3 Salticus scenicus (Clerck 1757) 2/2/1 - ll- T 1 Sitticus penicillatus (Simon 1875) - - -11 M 3 Sitticus pubescens (Fabricius 1775) 2/2 — 1/- spider species. Ruzicka (1992) described the spider assemblage inhabiting sandstone rocks in northern and northeastern Bohemia. Bath- yphantes simillimus inhabits, exclusively, these sandstone rocks in Central Europe, Lep- thyphantes pulcher inhabits not only sand- stone, but also granite and limestone rocks. Drassodes lapidosus was the dominant spe- cies in both localities studied; this species is generally considered to live under stones. The RUZICKA— SPIDERS ON ROCKS 221 occurrence of Segestria bavarica, Theridion betteni, Textrix denticulata, and Salticus scen- icus on rocks is mentioned by Miller (1971). The first two species occur exclusively on rocks in the Czech Republic. Abundant oc- currence of Titanoeca quadri guttata, Callile- pis schuszteri, Gnaphosa opaca, Zelotes ex- iguus, and Aelurilus v~insignitus at localities studied indicates that these species are well able to colonize rocky habitats. The abun- dance of Erigonoplus jarmilae, Zelotes puri- tanus, and Altella biuncata at Nezabudicke Skaly indicates that they occur primarily on rocky habitats. During intensive research of xerotherm localities in the Czech Republic, six specimens of Zelotes puritanus were col- lected at forest steppes (Miller & Buchar 1977; Sinkova 1973). Smaha (1983) collected 12 specimens at steppe slopes with isolated rocks in Kfivoklatsko Biosphere Reserve, while we obtained 11 specimens; 14 speci- mens of Erigonoplus jarmilae were collected at rock steppes (Miller 1947; Valesova 1962), we obtained 20 specimens; 5 specimens of Al- tella biuncata were collected on rock steppes and rocky slopes (Miller 1949; Buchar 1989; Dolansky 1997), we obtained 5 specimens. The occurrence of Zelotes puritanus (= Ze- lotes kodaensis Miller & Buchar 1977) in Eu- rope is known in the Czech Republic, Poland (Star^ga 1972) and Austria (Thaler 1981). This species inhabits exclusively original hab- itats, rocks and rock steppes here. This species cannot represent a recent introduction into Eu- rope (Platnick & Shadab 1983). In North America Zelotes puritanus inhabits a wider range of habitats. Specimens have been col- lected in pitfall traps in aspen, fir, scrub oak, lodgepole, ponderosa pine, black spruce for- ests, in beach litter, meadows, pastures, prai- ries, sagebrush, and under logs and rocks. Anyphaena furva was described by Miller (1967) from one male collected on a rock wall in the Zadielska Dolina valley, Slovakia. Sma- ha (1985) collected one male in a scree field under Tyfovska Skala rock in Kfivoklatsko Reserve. Our finding of three males on tree trunks at Nezabudicke Skaly coincides to the biology of the closely related Anyphaena ac- centuata (Walckenaer 1 802), and suggests that Anyphaena furva inhabits tree trunks on sun- exposed rocks and can occasionally move onto such rocks. Rock as a habitat of thermophilous spe- cies.— The frequency of specimens of ther- mophilous species amounts 43% at Nezabu- dicke SkMy, and 38% at Kotyz. Together, 26 thermophilous species belonging to 11 fami- lies were recorded. The surface of sand dunes and scree slopes can be heated to high temperatures. This effect is caused by isolating air interlayers. The specificity of arachnofauna of sand dunes is well known. In contrast, we found no specific inhabitants of upper overheated margins of boulder accumulations. This is probably the result of the very low humidity of these sites (Ruzicka et al. 1995). Rocks are the third bare, natural habitat, which can also overheat. Potential surface temperature and heat ac- cumulation capacity are considered as the most important characteristics of thermal be- havior of rocks. Different values of physical constants of particular rocks actually suggest that the rocky substratum may play a decisive role in the thermal balance of habitats domi- nated by larger exposed rock. Both andezite and limestone are considered to be “warm, calorific” rock with a predisposition to har- boring isolated populations of thermophilous plant and animal species (Rejmanek 1971). Thermophilous Segestria bavarica inhabits rocky habitats (in Switzerland and Austria), and also forests, where it was collected in pit- fall traps and by hand-picking under bark (No- flatscher 1991; Maurer & Hanggi 1990). In the Czech Republic, it occurs exclusively on rocks. This case supports a hypothesis that, in some northern locations, thermophilous spe- cies narrow their ecological niche exclusively to south-exposed rocky habitats, and they can reach the northernmost range of their distri- bution in these habitats. For example, Jonsson (1995) recorded the most northern occurrence of several thermophilous species in Sweden on rocky habitats. ACKNOWLEDGMENTS I thank Petr Antus for his help with col- lecting spiders on rocks, and Prof. Jan Buchar for his constructive criticism of the manu- script. This research was supported by the Grant Agency of the Czech Republic (Project No. 206/96/0326 and 206/99/0673). LITERATURE CITED Buchar, J. 1989. The knowledge of the present Bo- hemian arachnofauna and its improvement to 222 THE JOURNAL OF ARACHNOLOGY evaluation of natural conditions. Thesis, Charles University, Praha. 206 pp. (in Czech). Buchar, J. 1993. Kommentierte Artenliste der Spinnen Bohmens (Araneida). Acta Univ. Caro- linae Biol., 36 [1992]:383-428. Buchar, J. 1995. Bohemian wolf spiders (Araneida: Lycosidae). Acta Univ. Carolinae Biol., 39:3-28. Buchar, J., V. Ruzicka & A. Kurka. 1995. Check list of spiders of the Czech Republic. Pp. 35-53, In Proc. 15th European Colloquium of Arach- nology. (V. Ruzicka, ed.). Institute of Entomol- ogy, Ceske Budejovice. Dolansky, J. 1997. Thermophilous community of spiders from east- west part of Zelezne Hory Mts. Vc. sb. pnr. Prace a studie, 5:121-126 (in Czech, English sununary). Hanggi, A., E. Stockli & W. Nentwig. 1995. Le- bensraume mitteleuropaischer Spinnen. Centre suisse de cartographic de la faune, Neuchatel. 460 pp. Heimer, S. & W. Nentwig. 1991. Spinnen Mitte- leuropas. Paul Parey. Berlin and Hamburg. 543 pp. Jonsson, L.J, 1995. Cheiracanthium elegans, a new spider to Northern Europe, with a brief summary of the genus in Sweden. Entomol. Tidskr., 116: 55-58. Maurer, R. & A. Hanggi. 1990. Katalog der schweizerischen Spinnen. Sweizerischer Bund fiir Naturschutz. 411 pp. Miller, E 1947. Spiders of the serpentine rocky steppes near Mohelno. Acta Soc. pro cognitione et conservatione naturae in Moraviae Silesiaque, 7:1-128 (in Czech, French summary). Miller, F. 1949. The new spiders from the serpen- tine rocky heath near Mohelno (Moravia occ.). Entomol. Listy, 12:88-98. Miller, E 1967. Studien iiber die Kopulationsor- gane der Spinnengattung Zelotes, Micaria, Rob- ertus und Dipoena nebst Beschreibung einiger neuer oder unvollkommen bekannter Spinnenar- ten. Acta Sc. Nat. Brno, 1(7):25 1-298. Miller, F. 1971. Order Spiders Araneida. Pp. 51- 306, In Key to the Fauna of the Czechoslovakia IV. (M. Daniel & V. Cemy, eds.). CSAV, Praha (in Czech). Miller, F. & J. Buchar. 1977. Neue Spinnenarten aus der Gattung Zelotes Gistel und Haplodrassus Chamberlin (Araneae, Gnaphosidae). Acta Univ. Carolinae Biol., 1974:157-171. Noflatscher, M.-T. 1991. Beitrage zur Spinnenfau- na Siidtirols. Ill: Epigaische Spinnen an Xero- therm-Standorten am Mitterberg, bei Neustift und Sterzing (Arachnida: Aranei). Ber. nat.-med. Verein Innsbruck, 78:79-92. Platnick, N.I. & M.U. Shadab. 1983. A revision of the American spiders of the genus Zelotes (Ar- aneae, Gnaphosidae). Bull. American Mus. Nat. Hist., 174:97-192. Rejmanek, M. 1971. Ecological meaning of the thermal behaviour of rocks. Flora, 160:527-561. Ruzicka, V. 1987. Biodiagnostic evaluation of epi- geic spider conununities. Ecology (CSSR), 6: 345-357. Ruzicka, V. 1992. Current results of an arachno- logical survey of some sandstone rock sites in Bohemia (so-called “rock cities”). Arachnol. Mitt, 3:1-13. Ruzicka, V. 1996. Species composition and site distribution of spiders (Araneae) in a gneiss mas- sif in the Dyje river valley. Rev. Suisse de Zook, vol. hors serie, pp. 561-569. Ruzicka, V. (ed.). 1998. Spiders of southeastern Moravia. Sbomik Pnrodovedneho klubu v Uh. Hradisti 3:23-35 (in Czech, English summary). Ruzicka, V. & P. Antus. 1997. Collecting spiders from rocky habitats. Newsl. British Arachnol. Soc., 80:4-5. Ruzicka, V. & J. Hajer. 1996. Spiders (Araneae) of stony debris in North Bohemia. Arachnol. Mitt., 12:46-56. Ruzicka, V., J. Hajer & M. Zacharda. 1995. Arach- nid population patterns in underground cavities of a stony debris field (Araneae, Opiliones, Pseu- doscorpionidea, Acari: Prostigmata, Rhagidi- idae). Pedobiologia, 39:42-51. Ruzicka, V. & M. Zacharda. 1994. Arthropods of stony debris in the Krkonose Mountains, Czech Republic. Arctic and Alpine Res., 26:332-338. Sinkova, H. 1973. Spiders of the hills Sance and Hradiste at Zavist near Zbraslav. BSc thesis, Charles University, Praha. 67 pp. (in Czech). Smaha, J. 1983. Beitrag zur Erkenntnis der Arach- nofauna einiger Biozonosen des K0ivoklat-Ge- bietes (Mittelbbhmen). Vfst. es. Spolec. zooL, 47:126-136. Smaha, J. 1985. Einige Ergebnisse der Arachno- faunaforschung im Staatlichen Schutzgebiet Ty- fov. Bohem. cent., 14:189-224 (in Czech, Ger- man summary). Strar^ga, W. 1972. Fiir die Fauna Polens neue und seltenere Spinnenarten (Aranei), nebst Beschrei- bung von Lepthyphantes milleri sp. n. Fragmenta Faun., 18:55-98 (in Polish, German summary). Thaler, K. 1981. Bemerkenswerte Spinnenfunde in Nordtirol (Osterreich) (Arachnida: Aranei). Ver- offentl. Mus. Ferdinandeum, 61:105-150. Valesova, E. 1962. Spiders of the steppe locality Lochkov-Radotm. BSc thesis, Charles Univerity, Praha, 128 pp. (in Czech). Manuscript received 6 May 1998, revised 24 Sep- tember 1999. 2000. The Journal of Arachnology 28:223-226 RESEARCH NOTE ANYPHAENIDAE IN MIOCENE DOMINICAN REPUBLIC AMBER (ARACHNIDA, ARANEAE) Keywords: Anyphaenidae, Miocene amber, Dominican Republic Anyphaenids have a worldwide distribution but are particularly common in the neotropics. They are medium to large, long-legged spiders with claw tufts formed from several rows of lamelliform setae, and the tracheal spiracle sit- uated considerably more anteriorly than in other spiders; however, the latter character varies between genera. The family contains fast, active hunters, usually found on vegeta- tion, particularly tree foliage. In a revision of North American anyphaen- id genera Platnick (1974) stated that the tax- onomy of the approximately 375 Neotropical species was unclear. In a phylogenetic study, Ramirez (1995) established three anyphaenid subfamilies (Malenellinae Ramirez 1995, An- yphaeninae Bertkau 1878, Amaurobioidinae Hickman 1949), but considered the interfa- milial relationships unclear. The subfamilies were delimited by the position of the tracheal spiracle, the structure of the tegulum and me- dian haematodocha, and the structure of the female palpal tarsus. Recent papers (Brescovit 1996 and referenc- es therein) have delimited many of the Recent Neotropical anyphaenid genera. Brescovit (1996) revised the Neotropical Anyphaeninae at the generic level, creating 14 new genera (new total 32), 12 new synonymies, and 70 new combinations. This paper newly combines the amber species Anyphaeniodes bulla (Wun- derhch 1988) (= Aysha bulla) and Lupettiana ligula (Wunderhch 1988) (= Teudis ligula) in the light of Brescovit’s (1996) revision (which omitted fossil taxa). The Miocene Dominican Republic amber specimens studied, which are the only known representatives of the species concerned, were obtained from the Senckenberg Museum, Frankfurt (SMF, courtesy of Dr. M. GraBhoff). This amber is considered to be approximately 15-20 million years old (Iturralde-Vinent & MacPhee 1996). Anyphaenoides bulla (Wunderlich 1988) new combination Fig. 1 Aysha bulla Wunderlich 1988: 220, figs. 599-602, 764, holotype and only known specimen: male, SMF 38160, in Miocene Dominican Republic amber, examined. Emended diagnosis.— Males of A. bulla can be recognized by the following combi- nation of characters: embolus long, not form- ing a broad subcircular loop in the distal half of the cymbium, lacking a median constriction and a basal embolic process; large hook- shaped median apophysis with a broad base; tibia long with a simple retrolateral tibial apophysis. Female unknown. Remarks. — This species can be excluded ftom Aysha Keyserling 1891 by having a sim- ple palpal tibia (Wunderlich 1988: fig. 602) lacking complicated apophyses (e.g., Brescov- it 1996: fig. 259). Lupettiana ligula (Wunderlich 1988) new combination Fig. 2 Teudis ligula Wunderlich 1988: 221, figs. 603-605, 765, holotype and only known specimen: male, SMF 38152, in Miocene Dominican Republic amber, examined. Emended diagnosis.— Males of L. ligula can be recognized by the following combi- nation of characters: embolus (or conductor — see remarks) long, projecting ventrally; ven- tral tegular projection with pointed tip; retro- 223 224 THE JOURNAL OF ARACHNOLOGY Figure 1 . — Right pedipalp of Anyphaenoides bul- la new combination, male holotype, SMF 38160. Scale = 0.2 mm. Abbreviations: csp = cymbial spine, e = embolus, ma == median apophysis, t = tegulum, vtp = ventral tegular projection. lateral tibial apophysis long, with rounded apex; palpal tibia without dorsal cusps. Fe- male unknown. Remarks. — In this specimen not all the palpal sclerites are visible because of the po- sition in which the spider is preserved. The only view of the sclerites possible is that shown in Fig. 2, and it is not clear whether the anterior projection is the embolus or the conductor. However, the structure of the retro- lateral tibial apophysis (Fig. 2; Wunderlich 1988: figs. 604-605) is a synapomorphy of the genus and is sufficient evidence for the proposed new combination. Eskov (1990) has commented that the amber spider fauna is tax- onomically subequal to Recent faunas, and the certainty with which pattern-based species can be recognized in the fossil record is less than that for extant organisms (Smith 1994). This species can be excluded from Teudis O.R- Cambridge 1896 by having a palpal tibia lack- ing short conical projections (e.g., Brescovit 1996: fig. 68). Figure 2. — Right pedipalp of Lupettiana ligula new combination, male holotype, SMF 38152. Scale = 0.2 mm. Abbreviations: cs = cymbial se- tae, e/c? = embolus or conductor (see remarks un- der L. ligula), fe I = femur I, rta = retrolateral tibial apophysis, s = subtegulum, t = tegulum, ti = palpal tibia, vtp = ventral tegular projection. Wulfila spinipes Wunderlich 1988 Wulfila spinipes Wunderlich 1988: 218, figs 589- 598, 762-763, holotype male SMF 38136, and one male paratype, SMF 38144, both in Miocene Dominican Republic amber, examined. Remarks. — Wulfila as currently delimited contains approximately 40 species with Ne- arctic and Neotropical distributions (Brescovit 1996). The interspecific relationships are un- clear and the genus is in need of revision. The specimens described by Wunderlich are re- tained in Wulfila due to the structure of the complicated retrolateral tibial apophysis, the long ventral tegular projection and the conical ventral coxal projections. Unfortunately, legs I are missing in both holotype and paratype, so it is impossible to determine their lengths relative to legs II; however, the specimens do not possess the large and distinct ventral che- PENNEY— MIOCENE DOMINICAN REPUBLIC AMBER ANYPHAENIDAE 225 liceral tooth present in Wulfilopsis (e.g., Bres- covit 1996: fig= 34). Wunderlich’s diagnosis serves only to separate this species from the other described Dominican Republic amber spiders, and is not sufficient to separate it from all the extant Wulfila species. An emend- ed diagnosis will have to wait, pending revi- sion of the extant species or, preferably, the amber specimen would be included in such a revision. DISCUSSION These are the first fossil records of the gen- era Anyphaenoides and Lupettiana, taking them back 15-20 million years. As a result of the new combinations, Aysha and Teudis are not known in the fossil record. Lupettiana is represented on Hispaniola by two, and Wulfila by three extant species, whereas Anyphaenoides is not recorded from the Recent Hispaniolan fauna (Penney 1999a). Brescovit’s (1992) revision of the genus ex- tended the known geographical range of An- yphaenoides from Peru, Equador and the Ga- lapagos Archipelago, to include Panama, Venezuela, Surinam, Brazil and northern Ar- gentina. Baert (1995) added Cocos Island in the Pacific. Hispaniola is unique in terms of its known spider fauna in that more families are recorded from fossil species in amber than are known from extant species (Wunderlich 1988; Penney 1999b). There have been 291 Recent species in 155 genera and 40 families recorded from Hispaniola (e.g.. Banks 1903; Bryant 1943, 1945, 1948; Penney 1999a), but this fauna has not been intensively investigat- ed using a variety of collecting techniques. Evidence from sedimentary and geomor- phic data, alluvial terraces and albedo reflec- tivity indices suggest that the Dominican Re- public was not drastically affected by the Pleistocene glaciations (Schubert 1988), and the Tertiary Hispaniolan spider lineages have probably suffered no major habitat disruption that would cause their extinction. This is sup- ported by the high degree of similarity be- tween the species composition of the known Tertiary fauna and the Recent fauna (Penney 1999b). Anyphaenoides is recorded from the amber and is a component of the Recent Neo- tropical fauna; it can be predicted that this ge- nus has at least one undiscovered Recent spe- cies present on Hispaniola. ACKNOWLEDGMENTS Thanks to Manfred GraBhoff and Uli Schreiber (SMF) for providing amber spiders for study, to Paul Selden (University of Man- chester), and reviewers for their comments on the manuscript, and to Antonio D. Brescovit (Instituto Butantan, Brazil) for providing re- prints of his publications. LITERATURE CITED Baert, L. 1995. The Anyphaenidae of the Galapa- gos Archipelago and Cocos Island, with a rede- scription of Anyphaenoides pluridentata Borland 1913. Bull. British Arachnol. Soc., 10:10-14. Banks, N. 1903. A list of Arachnida from Hayti, with descriptions of new species. Proc. Acad. Nat. Sci. Philadelphia, 55:340-345. Brescovit, A.D. 1992. Revisiao das aranhas neo- tropicais do genero Anyphaenoides Borland (Ar- aneae, Anyphaenidae). Rev. Brasileira. Entomol., 36:741-757. Brescovit, A.D. 1996. Revisao de Anyphaeninae Bertkau a nivel de generos na regiao Neotropical (Araneae, Anyphaenidae). Rev. Brasileira Zool., 13:1-187. Bryant, E.B. 1943. The salticid spiders of Hispan- iola. Bull. Mus. Comp. Zool., 92:445-521. Bryant, E.B. 1945. The Argiopidae of Hispaniola. Bull. Mus. Comp. Zool., 95:357-442. Bryant, E.B. 1948. The spiders of Hispaniola. Bull. Mus. Comp. Zool., 100:331-459. Eskov, K.Y. 1990. Spider palaeontology: Present trends and future expectations. Acta Zool. Fen- nica, 190:123-127. Iturralde-Vinent, M.A. & R.D.E. Macphee. 1996. Age and palaeogeographical origin of Dominican amber. Science, 273:1850-1852. Penney, D. 1999a. Dominican Republic amber spi- ders and their contribution to fossil and Recent ecology. Unpubl. Ph.D. thesis, Univ. of Man- chester. 377 pp. Penney, D. 1999b. Hypotheses for the Recent His- paniolan spider fauna based on the Dominican Republic amber spider fauna. J. Arachnol., 27: 64-70. Platnick, N.I. 1974. The spider family Anyphaen- idae in America north of Mexico. Bull. Mus. Comp. Zool., 146:205-266. Ramairez, M.J. 1995. A phylogenetic analysis of the subfamilies of Anyphaenidae (Arachnida, Ar- aneae), Entomol. Scandinavica, 26:361-384. Schubert, C. 1988. Climatic changes during the last glacial maximum in northern South America and the Caribbean: A review. Interciencia, 13: 128-137. Smith, A.B. 1994. Systematics and the fossil re- cord, documenting evolutionary patterns. Black- well Sci. PubL, London. 178 pp. 226 THE JOURNAL OF ARACHNOLOGY Wunderlich,;. 1988. Die fossilen Spinnen im dom- Manchester, Oxford Road, Manchester, inikanischen Bernstein. Beitr. Araneol., 2:1-378. M13 9PL, United Kingdom David Penney: Invertebrate Zoology, The Manuscript received 10 February 1999, revised 1 Manchester Museum, The University of October 1999. 2000. The Journal of Axachnology 28:227-230 RESEARCH NOTE EREMOPUS ACUITLAPANENSIS, A NEW SPECIES (SOLIFUGAE, EREMOBATIDAE, EREMOBATINAE) FROM GUERRERO, MEXICO Keywords: Solifugae, Guerrero, Mexico A medium-to-deep fondal notch is usually present below the fixed finger on the chelic- erae of male eremobatine solifugids. While studying a series of these arachnids from Guerrero, Mexico, collected by members of the Instituto de Biologia of Universidad Na- cional Autonoma de Mexico, we discovered a new eremobatine species whose uniquely modified fixed cheliceral finger bears a me- soventral flange that covers the fondal notch. Vazquez (1986) illustrated this species, but he did not offer a formal description. We now present a description. Nomenclature and measurements (mm) were made as described in Muma (1951). The ratios CL/CW, PW/PL and A/CP, as defined by Brookhart & Muma (1981) and ECCS se- tae are also provided. Colors are from speci- mens preserved in alcohol. Type depositories are abbreviated as follows: IB UN AM = La- boratorio de Acarologia, Instituto de Biologia, Universidad Nacional Autonoma de Mexico, Coyoacan 04510, D.E, Mexico; AMNH = American Museum of Natural History, New York. Eremopus acuitlapanensis new species Figs. 1-9, Tables 1, 2 Types.- — Holotype, male: Mexico: Guerre- ro: Acuitlapan, 25 km East of Taxco (17°35'N, 99n5'W), 13 October 1976, Mel- gar, Novelo, Chavarria, collectors. Deposited in IBUNAM. Paratypes: Mexico: Guerrero: Acuitlapan, 13 October 1976, Melgar, Novelo, Chavarria, IS (IBUNAM); Las Granadas, 28 October 1980, [collector unknown], Id (AMNH); Acuitlapan, 25 km east of Taxco (17°35'N, 99n5'W), 13 October 1976, E. Cer- vantes, 1 $ (IBUNAM); Acuitlapan, 23 Janu- ary 1978, 1. Espejel, 1 9 (AMNH). Etymology.— The species name acuitlapa- nensis refers to Acuitlapan, Guerrero state, Mexico, the place where the specimens were collected. Diagnosis.— The presence on males of a wide, shallow, almost indistinct mesal crease on the fixed cheliceral finger, an anterior tooth on the movable cheliceral finger, and four long, stout needle-like ctenidia on the first post-spiracular abdominal stemite suggest a close relationship of this species to Eremopus montezuma Roewer 1934 and Eremopus fus- cus (Muma 1987). Males of Eremopus acu- itlapanensis new species are easily distin- guished from other members of the genus by the presence of a mesoventral flange on the basal one-third of the fixed cheliceral finger. The flange almost covers the entire fondal notch. The genital opercula of females of Er- emopus acuitlapanensis new species are sim- ilar to those of E. fuscus females, but the lat- eral concavities of E. acuitlapanensis opercula are a little closer together and the posterior margins are well-defined. Description.— Ma/c.' Propeltidium yellow, wider than long, with a narrow, dusky band on anterior margin (measurements in Table 1). Eye tubercle dark; eyes separated by almost two diameters. Dorsal opisthosoma dark brown-to-black with pleural membranes grey- to-brown. First post-spiracular abdominal ster- nite provided with four long, stout needle-like ctenidia (Fig. 6), these extending beyond the middle of the succeeding stemite. Chelicera (Figs. 1-5) robust, yellow; dentition reddish- brown. Fixed finger straight, slightly down- 227 228 THE JOURNAL OF ARACHNOLOGY Table 1 . — Measurements (mm) of male holotype and male paratypes of Eremopus acuitlapanensis. The data are taken from the holotype and two paratype specimens. Structure Length Width Ratios Holotype Paratype Holotype Paratype Holotype Paratype Chelicerae 7.5 7.5, 7.7 3.7 3.7, 3.7 CL/CW = 2.03 2.03, 2.08 Propeltidium 4.0 4.0, 4.0 5.5 5.5, 5.7 PW/PL = 1.37 1.37, 1.42 Palpi 22.2 21.5, 22.2 A/CP = 6.09 5.83, 5.79 Legs I 18.4 18.0, 16.5 Legs IV 29.5 27.5, 29.0 turned distally, with part of the basal one-third forming a mesoventrally directed flange that covers almost the entire fondal notch (see ar- row in Fig. 3). Mesal surface of fixed finger with a wide, shallow, crease extending from the tip of the finger a point above the upper- most tooth (I) of the ectal fondal tooth row (Fig. 3). Fondal teeth arranged in two rows of four teeth each, both graded I, III, II, IV in size. Movable finger with low, flattened an- terior tooth, large principal tooth, and two small intermediate teeth contiguous with the principal tooth; small, distinct mesal tooth present. Flagellum complex with dorsal series of simple tubular setae and ventral series of slightly striate to plumose setae. Apical bris- tles of flagellum complex not conspicuously enlarged or flattened. Ectal cheliceral cluster setae as in Fig. 5. Palpi yellow, each with many cylindrical spine-like setae and long whip-like setae; scopula absent. Legs yellow, without dark markings. Malleoli white. Female: Similar in form and coloration to male, but differing as follows. Opercula of genital segment (Fig. 9) with lobes extended laterally, slightly separated anteriorly and pos- teriorly, with a well-defined posterior border, and with a concavity on the ectal side of each lobe. First post-spiracular sternite without cte- nidia. Chelicerae (Fig. 7) proportionally larger than those of male. Fixed finger with a large principal tooth, a slightly smaller medial tooth, an anterior tooth which is half the size of the principal tooth, a well-developed inter- mediate tooth between the anterior and medial teeth, and three small intermediate teeth con- tiguous with the anterior margin of the prin- cipal tooth. Movable finger with a large prin- cipal tooth, a well-developed anterior tooth, and three small intermediate teeth contiguous with the anterior margin of the principal tooth; small, distinct mesal tooth present. Ectal che- liceral cluster setae as in Fig. 8. This contribution is dedicated to the mem- ory of Dr. Leonila Vazquez Garcia, Curator of Arachnology, Instituto de Biologia (IBUN- AM), for providing us with facilities in which to study the museum’s specimens. Thanks also to Dr. Anita Hoffmann, Laboratorio de Acar- ologia “Anita Hoffmann,” Facultad de Cien- cias, UNAM, for the opportunity to work with her arachnological collection deposited at IBUNAM. We are also grateful to Dr. Warren E, Savary for his critical review of the man- uscript. Table 2. — Measurements (mm) of female paratypes of Eremopus acuitlapanensis; the museum acronyms are given. Length Width Ratios Structure IBUNAM AMNH IBUNAM AMNH IBUNAM AMNH Chelicerae 9.5 10.0 4.0 4.0 CL/CW = 2.40 2.50 Propeltidium 4.0 4.5 7.0 7.0 PW/PL = 1.75 1.55 Palpi 20.0 20.0 A/CP - 4.44 4.55 Legs I 16.0 17.0 Legs IV 24.0 29.0 VAZQUEZ & GAVINO-ROJAS— NEW SOLIFUGID FROM MEXICO 229 2.5 mm Figures 1-9. — Eremopus acuitlapanensis new species. 1-6, Male holotype. 1. Ectal view of right chelicera; 2. Mesal view of right chelicera; 3. Mesal view of right chelicera (detail); 4. Frontal view of right chelicera; 5. Holotype ECCS setae of right chelicera; 6. Ventral view of abdominal ctenidia. 7-9. Female paratype. 7. Mesal view of left chelicera; 8. ECCS setae of left chelicera; 9. Genital opercula. 230 THE JOURNAL OF ARACHNOLOGY LITERATURE CITED Brookhart, J.O. & M.H. Muma. 1981. The pallipes species-group of Eremobates Banks (Solpugida: Arachnida) in the United States. Florida Ento- mol., 64(2):283-308. Muma, M.H. 1951. The arachnid order Solpugida in the United States. Bull. American Mus. Nat. Hist., 97:31-141. Muma, M.H. 1985. A new possibly diagnostic character for Solpugida (Arachnida). Nov. Ar- thropodae, 2(2): 1-5. Muma, M.H. 1987. New species and records of Solpugida (Arachnida) from Mexico, Central America and the West Indies. Printed by South- west Offset, Silver City, New Mexico. 31 pp. Roewer, C.F.R. 1934. Solifugae, Palpigradi. In Dr. H.G. Bronn’s Klassen und Ordnungens des Tier- reichs 5,4(4): 1-723. Vazquez, I.M. 1986. Nuevos Eremobatidos Mexi- canos (Arachnida: Solpugida). Proc. 9th Intemat. Congr. ArachnoL, Panama, 1983:281-284. Ignacio M. Vazquez and Rafael Gavino-Ro- jas: Laboratorio de Acarologia “Anita Hoffmann,” Facultad de Ciencias, Univer- sidad Nacional Autonoma de Mexico, Coy- oacan 04510, D.E, Mexico. Manuscript received 2 December 1998, revised 1 October 1999. 2000. The Journal of Arachnology 28:231-236 RESEARCH NOTE EXTERNAL MORPHOLOGY AND ULTRASTRUCTURE OF THE PREHENSILE REGION OF THE LEGS OF LEIOBUNUM NIGRIPES (ARACHNIDA, OPILIONES) Keywords! Chemoreception, harvestmen, locomotion, setae Species of harvestmen (Arachnida, Opili- ones, Palpatores) in the family Sclerosoma- tidae frequently employ prehensile flexion of the telotarsus during locomotion. Kaestner (1968) described the ability of these arach- nids to anchor themselves to objects such as blades of grass by wrapping their legs around these objects. We have observed both Leiob- unum nigripes (Weed 1892) and L. vittatum (Say 1821) moving across surfaces by form- ing coils at the end of their legs, especially the second pair (Figs. 1-4). While moving across a smooth substrate, these harvestmen cast the coiled regions of their legs about un- til they catch on a structure. Similar strate- gies are also employed by harvestmen during climbing, with the exception being that once a purchase is obtained with a coil, the free legs often wrap around and climb up the an- chored leg. In addition, we have also ob- served harvestmen in aggregations wrapping their legs around the legs of adjacent indi- viduals (Fig. 2). Movement of the legs in harvestmen has been hypothesized to occur through a combi- nation of muscle action and a hydraulic pump mechanism (Shultz 1989; Foelix 1996). Ac- cording to this hypothesis, hemolymph is pumped into the legs by contraction of either the musculi laterales or the endostemal mus- cles (the primitive condition: Shultz 1991) of the prosoma (Parry 1960), resulting in leg ex- tension. For harvestmen, Shultz (1989) re- ported that the basitarsus and telotarsus of the ^Current address: Department of Biology, 411 SW 24th Street, Our Lady of the Lake University, San Antonio, Texas 78207-4689, USA. leg are traversed by two tendons arising from muscles that are used to move the tarsal claw. The telotarsus is subdivided by numerous adesmatic joints (>50: Kaestner 1968) that impart a prehensile character to the tarsus when flexed (Figs. 5-7). Flexion at the ades- matic joints can occur only ventrally in L. ro- tundum (Latreille 1795) because the ventral joint membranes are shorter than the dorsal joint membranes (Kaestner 1968). In this pa- per we describe the external morphology and ultrastructure of the prehensile region of the legs of juveniles of Leiobunum nigripes (Scle- rosomatidae). We collected juvenile Leiobunum nigripes from Chicot State Park, Evangeline Parish, Louisiana on 8 March 1997 and housed them in screened aquaria for approximately one week prior to preservation. Within 48 h after molting, specimens were fixed in cold (4 °C) Trump’s fixative (a mixture of sodium caco- dylate buffer, formalin, and glutaraldehyde) overnight, rinsed in 0.2 M sodium cacodylate buffer (pH 7.4) and postfixed in 2% OSO4 for 90 min at room temperature. Specimens were then dehydrated in a graded ethanol se- ries and chemically dried with hexamethyldi- silazane (Nation 1983), mounted on aluminum stubs, and sputter-coated for 2 min with —20 nm of gold. We examined and photographed these specimens with a JEOL 6300-F field emission scanning electron microscope at ac- celerating voltages of 15-20 kV. Specimens examined with transmission electron microscopy (TEM) were fixed and dehydrated using the same protocol described above for scanning electron microscopy 231 232 THE JOURNAL OF ARACHNOLOGY iungB!>S »8S^ Figures 1-4. — Adults of the harvestman Leiobunum nigripes on hardware cloth (mesh size 6 mm X 6 mm) showing the prehensile ability of the tarsi. 1. An individual anchored to the substrate; 2. A small aggregation of harvestmen in which one individual has wrapped one of its leg around the leg of another; 3, 4. Dorsal views of the prehensile region of the tarsi, showing the wrapping of the legs around individual metal wires. Arrows in each figure indicate regions of flexion in the distal tips of the telotarsus. (SEM). Following dehydration, specimens were slowly infiltrated in Spurr’s low viscosity standard resin (Spurr 1969) over four days and sectioned with a diamond knife. Thin sections were collected on carbon- stabilized 200 iJim thin bar grids, stained sequentially with meth- anolic uranyl acetate and aqueous lead citrate, and observed with a Hitachi H-7000 trans- mission electron microscope at 75 kV. On each leg, L. nigripes has a single, smooth tarsal claw that is not toothed (Fig. 5). Smaller setae, or sensilla trichodea (Schneider 1964; Spicer 1987), and larger primary spines, or sensilla chaetica (Schneider 1964; Spicer 1987), are denser on the ventral surface of the telotarsus than on the dorsal surface (Fig. 6). The sensilla trichodea lie nearly parallel with the surface of the leg and have no specialized basal articulating membrane (Figs. 6, 7). The sensilla chaetica are nearly perpendicular to the leg surface and have a specialized basal articulating membrane (Figs. 11, 12), with blunt tips and whorled striae (Fig. 14), unlike those of sensilla trichodea (Figs. 6, 7). There is no evidence of trichobothria, mechanore- ceptors that are common to most arachnids (Reissland & Corner 1986; Foelix 1996). The adesmatic joints are easily distinguished from true joints (Fig. 8) on the basis of their small size. Cross sections examined with TEM con- firmed the earlier anatomical observations of Kaestner (1968); i.e., no muscles were found between the segments of the telotarsus (Fig. 9). We observed only a single tendon (Fig. 9) connecting the tarsal claw to the claw-flexing GUFFEY ET AL.— PREHENSILE REGION OF HARVESTMEN LEGS 233 Figures 5-8. — External morphology of tarsus IV of Leiobunum nigripes. 5. Lateral view of the telotarsus and tarsal claw. Scale bar = 127 mm; 6. The adesmatic joints on the ventral surface of the telotarsus. Scale bar = 64 pm; 7. The sensilla trichodea (st) and sensilla chaetica (sc) on the ventral surface of the telotarsus near an adesmatic membrane (am) of an adesmatic joint. Scale bar = 23 pm; 8. A lateral view of a true joint between the two most distal segments of a leg, the basitarsus and the telotarsus. Scale bar = 73 pm. musculature located in the tibia. In the telo- tarsus, we also observed epidermal cells lining the innermost portions of the cuticle (hypo- dermis) and occurring in clusters within the leg hemocoel (Fig. 10). During the course of our TEM studies of the internal features of the leg, several sec- tions provided information concerning the innervation of the sensilla chaetica (Figs. 6, 7, 11). Apical pores, a common feature of sensilla chaetica among arthropod chemo- receptors (reviewed in Zacharuk 1980), were not observed in our specimens. This sensillum is innervated by many presumably chemoreceptive dendrites (Figs. 12, 13). The dendrites originate from enveloping cells within the hypodermis (Fig. 13; inset) which do not attach to the cuticular wall of the bas- al articulating membrane. Instead, the sheath containing the dendrites passes directly through the center of the setal shaft (Fig. 12, 13), a common feature of arthropod che- moreceptors (Altner & Prillinger 1980; Za- charuk 1980). The external morphology of the prehensile region of the legs of Leiobunum nigripes is similar to that reported by Kaestner (1968) for L. rotundum and by Holmberg & Cokendol- pher (1997) for Togwoteeus biceps (Thorell 1877). Our observations of the sensilla on the tarsi of L. nigripes are also similar to those reported by Spicer (1987) for the palps of L. townsendi. The most numerous sensory or- gans on the legs of L. nigripes appear to be 234 THE JOURNAL OF ARACHNOLOGY Figures 9-14. — Ultrastmcture of the telotarsus of leg IV of Leiobunum nigripes. 9. TEM micrograph of a cross section of the telotarsus revealing a single tendon (t) within a hemocoelic space (hs) and showing no muscle or tendon attachments with the inner surface of the cuticle (c). Scale bar = 6 pm; 10. TEM micrograph of the epidermal cells lining the innermost portion of the cuticle. Scale bar = 3 pm; 11. SEM micrograph of the specialized basal articulating membrane (bm) of a sensilla chaetica (s) from the ventral surface of the telotarsus. Scale bar = 3 pm; 12, TEM micrograph of a basal articulating membrane and shaft of a sensilla chaetica revealing the dendrites (d) and dendritic sheath (ds) within the shaft of the GUFFEY ET AL.— PREHENSILE REGION OF HARVESTMEN LEGS 235 sensilla chaetica (primary spines) and sensilla trichodea (setae). Unlike the palps of L. town- sendi, however, these sensilla appear to be more numerous on the ventral surface of the telotarsus than the dorsal surface. Spicer (1987) also reported two types of sensilla chaetica (types I and II) based on the differing lengths of the sensilla. We observed only one type of sensilla chaetica in L. nigripes. We also did not observe any pores that are char- acteristic of chemoreceptors on the sensilla chaetica (Slifer 1970), but the structure of the dendrites innervating them (e.g., many den- drites and lack of attachment to the basal ar- ticulating membrane) indicates that they may function in chemoreception. Spicer (1987) in- ferred that the row of spines found on the ven- tral surface of the palps of L. townsendi were chemoreceptors and such receptors have been reported for other species of harvestmen (e.g., Foelix 1985). ACKNOWLEDGMENTS Funding for this study was provided by a grant to C. Guffey from the American Arach- nological Society Fund for Graduate Student Research, Louisiana Board of Regents Doc- toral Fellowship grant LEQSF[ 1994-99] -GF- 29 to C. Guffey through R.G. Jaeger, United States Department of Agriculture grant USDA-CREEF 9501834 to B.E. Felgenhauer, and research grants from The University of Southwestern Louisiana Graduate Student Or- ganization to C. Guffey and V.R. Townsend. Assistance with the identification of species was provided by J. Cokendolpher. We thank J. Marshall and two anonymous reviewers for critically reviewing an earlier draft of this manuscript and T. Pesacreta for assistance with the transmission electron microscope and the scanning electron microscope at the Elec- tron Microscopy Center at The University of Southwestern Louisiana. LITERATURE CITED Altner, H. & L. Prillinger. 1980. Ultrastmcture of invertebrate chemo-, thermo-, and hygrorecep- tors and its functional significance. Int. Rev. CytoL, 67:69-139. Foelix, R.F 1985. Mechano- and chemoreceptive sensilla. Pp. 118-137, In Neurobiology of Arachnids. (EG. Barth, ed,). Springer- Verlag, Berlin. Foelix, R.F. 1996. Biology of Spiders, 2nd ed. Ox- ford Univ. Press, New York. Holmberg, R.G. & J.C. Cokendolpher. 1997. Re- description of Togwoteeus biceps (Arachnida, Opiliones, Sclerosomatidae) with notes on its morphology, karyology and phenology. J. Arach- nol., 25:229-244. Kaestner, A. 1968. Invertebrate Zoology, Vol. II: Arthropod Relatives, Chelicerata, Myriapoda. (Translated by H.W. Levi & L.R. Levi.) Intersci- ence Publishers, New York. Nation, J.L. 1983. A new method for using hexa- methyldisilazane for preparation of soft insect tissue for scanning electron microscopy. Stain TechnoL, 58:347-351. Parry, D.A. 1960. Spider hydraulics. Endeavor, 19: 156-162. Reissland, A. & P. Gomer. 1986. Trichobothria. Pp. 138-160, In Ecophysiology of Spiders. (W. Nen- twig, ed.). Springer- Verlag, New York. Schneider, D. 1964. Insect antennae. Ann. Rev. En- tomoL, 9:103-122. Shultz, J.W. 1989. Morphology of locomotor ap- pendages in Arachnida: Evolutionary trends and phylogenetic implications. Zool. J. Linn. Soc., 97:1-56. Shultz, J.W. 1991. Evolution of locomotion in Arachnida: The hydraulic pressure pump of the giant whipscorpion, Mastigoproctus giganteus (Uropygi). J. Morph., 210:13-31. Slifer, E.H. 1970. The structure of arthropod chemoreceptors. Ann. Rev. Entomol., 15:121- 142. Spicer, G.S. 1987. Scanning electron microscopy of the palp sense organs of the harvestman Leiobunum townsendi (Arachnida: Opiliones). Trans. American Microsc. Soc., 106:232-239. Spurr, A.R. 1969. A low- viscosity epoxy resin em- <- seta. Scale bar = 2 [xm; 13. TEM micrograph of a sensilla chaetica, the cuticle and the underlying structure within the telotarsus. The inset is of a nerve from a chemoreceptive setae, revealing multiple dendrites within a single dendritic sheath. Scale bar = 5 ixm; 14. SEM micrograph of the distal tip of a sensilla chaetica revealing the whorled striae on the external surface and the lack of a discemable pore at the tip. Scale bar = 3 fxm. 236 THE JOURNAL OF ARACHNOLOGY bedding medium for electron microscopy. J. Ul- trastruc. Res., 26:31-43. Zacharuk, R.Y. 1980. Ultrastructure and function of insect chemosensilla. Ann. Rev. EntomoL, 25: 27-47. Cary Guffey ^ Victor R. Townsend, Jr., and Bruce E. Felgenhauer: Department of Biology, P.O. Box 42451, University of Southwestern Louisiana, Lafayette, Louisi- ana, 70504^2451 USA Manuscript received 18 December 1998, revised 5 January 2000. 2000. The Journal of Arachnology 28:237-240 RESEARCH NOTE A COMPARATIVE STUDY OF SEXUAL BEHAVIOR IN TWO SYNMORPHIC SPECIES OF THE GENUS LYCOSA (ARANEAE, LYCOSIDAE) AND THEIR HYBRID PROGENY Keywords! Spider hybrid courtship, hybrids, species-specific recognition Lycosa thorelli Keyserling 1877 3.nd Lycosa carbonelli Costa & Capocasale 1984 are syn- morphic, synchronic and sympatric species (Costa & Capocasale 1984) that, however, can be collected in different microhabitats (un- publ. data). Males are slightly different in their female- searching behavior, but differ clearly in their courtship in front of females, thus avoiding interbreeding (Costa & Capo- casale 1984). Costa & Francescoli (1991), us- ing anaesthetized females, obtained an excep- tional hybrid brood (from L. thorelli male and L. carbonelli female) and two conspecific con- trol broods. These three broods were raised to adulthood (Francescoli & Costa 1992) and then used in the present study. Costa et al. (1997), analyzing the behavior of parental (L. thorelli and L. carbonelli) and hybrid males elicited by female hybrid pher- omone, found that hybrid males showed an intermediate behavioral pattern between both parental species and a low activity level. Pa- rental males showed an intermediate activity level when compared to the activity elicited by conspecific and heterospecific stimuli. In this paper, we analyzed (1) the behavior of the above mentioned males elicited by the parental species’ sexual pheromone, and (2) the direct interactions among the males and the females of the three groups. This study allows a thorough comparison of the sexual behavior of the three groups released by dif- ferent sexual stimuli. Also, these data facili- tate both a deeper analysis of the function of reproductive isolation mechanisms and the formulation of hypotheses about signal effec- tiveness, mechanisms of heritability, and the possible evolutionary paths taken by the courtship behavior of these species. We used 5 S and 2 ? L. thorelli, 8 S and 3 9 L. carbonelli, and 9S and 8 9 hybrids. Also, one wild-caught female L. thorelli and one wild-caught L. carbonelli were used as sex pheromone donors. Spiders were housed in the same conditions as in Costa et al. (1997). Voucher specimens were deposited in the en- tomological collection of the Facultad de Ciencias, Montevideo. During the experiments the females were kept in glass containers (15 cm diameter X 5 cm high) with sand as a substrate. Two types of experiments were done. Males were ob- served in: (1) the presence of one parental sex pheromone, and (2) in the presence of both the female spider and the corresponding sex pheromone. In the first type, the females re- mained in the containers at least 24 h and were taken out immediately before the intro- duction of the male. The male was gently in- troduced, and his behavior was observed for 5 minutes. In the second type, we introduced a female at least 24 h before the experiment and the behavior of the male was then record- ed after visual or tactile contact with the fe- male. To reduce the probability of attacks, the male was introduced behind an opaque barrier. Experiments were ended when: (1) the male performed 60 minutes of sexual activity with- out copulation; (2) no sexual behavior was ob- served for 20 minutes; (3) copulation was completed; (4) the female attacked the male. Room temperature during the experiments was 23.4 ± 1.5 °C. Forty-three trials were done, 15 in the con- text of female sex pheromone only and 28 in the context of the female and sex pheromone. The trials with pheromone only were (H = hybrids, C — L. carbonelli, T = L. thorelli', the first letter corresponds to the male and the second to the female): HC (4 trials), CC (2), TC (2), CT (2), TT (2), HT (3); data from HH, CH and TH were taken from Costa et al. (1997). The trials with females were: HH (5 trials), HC (3), HT (3), CH (3), CC (3), CT 237 238 THE JOURNAL OF ARACHNOLOGY Table 1. — Single records of rhythms and angles for leg movements. Data taken only from videos that allowed clear, on-screen measurements. Rhythms measured in movements/second; angles covered by leg movements measured in degrees. Groups were identified with two letters, the first corresponding to the male species and the second corresponding to the female (or pheromone donor): H = hybrids, C = L. carbonelli, T = L. thorelli. “ — ” denotes no data. Experimental group Behavior HC CC CT TT Leg waving Rhythm 8.2 3.6 3.5 — Angle 13.8 22.9 21.3 — Rubbing Rhythm — — — 11.8 Drumming Rhythm 7.0 4.0 5.0 4.6 Leg tapping Rhythm — 2.7 3.3 — Angle — 28.0 25.1 — (2), TH (3), TC (3) and TT (3). Individuals were used randomly, avoiding consecutive tri- als for the same individual. The low number of observations was because of the extremely limited number of available individuals (Fran- cescoli & Costa 1992) and the risk involved in direct sexual encounters. The trials were video-taped and analyzed using 19 behaviors. Some behaviors are composed by more than one act that occur simultaneously. The behav- iors observed in this study were: Abdominal vibrations (AV), Attack (At), Copulation (Co), Drumming (Dr), Explosive locomotion (EL), Immobility (Im), Leg tapping (LT), Leg wav- ing (LW), Locomotion (Lo), Locomotion- with-Drumming (Lo/Dr), Locomotion-with- Leg tapping (Lo/LT), Locomotion-with-Leg waving (Lo/LW), Locomotion-with-Leg wav- ing-with-Drumming (Lo/LW/Dr), Locomo- tion-with-Palpation (Lo/Pa), Locomotion- with-Palpation-with-Leg tapping (Lo/Pa/LT), Palpation (Pa), Positioning (Po), Rest posture (RP) and Rubbing of legs (Ru). Comparisons using data obtained here and data from Costa et al. (1997) were made. The mean repertoire size comparisons used both sexual behaviors and all behaviors. Mean rep- ertoire size was the average number of differ- Table 2. — Repertoire size in experimental groups responding to parental pheromone. “ — ” denotes no data. HH, CH and TH data were taken from Costa et al. (1997). Group Reper- toire size Num- ber of obser- Repertoire size values as X (SD) vations All units Sexual units HH 21 46 5.80 (3.59) 3.78 (3.19) CH 19 16 7.90 (3.20) 5.69 (3.28) TH 22 15 8.13 (5.59) 6.07 (5.26) HC 6 4 2.75 (2.22) 1.00 (2.00) CC 6 2 3.00(1.41) 2.00 (0.00) TC 1 2 1.00 (0.00) — HT 5 3 2.33 (1.53) 1.33 (1.53) CT 8 2 5.00(1.41) 2.00 (2.83) TT 9 2 5.00 (4.24) 3.00 (4.24) ent behaviors presented in any experiment for each group. Single measurements of rhythms and angles for leg movements in some behav- iors were obtained (Table 1). The repertoire sizes of males stimulated by parental sex pheromone were smaller in relation to those elicited by hybrid pheromone (Table 2). Comparisons of mean repertoire size for the same kind of male and for the same kind of pheromone were made. In the first comparisons, HH was significantly different than HT {t = 3.37, 0.01 > P> 0.001) and than HC {t = 2.48, 0.02 > P > 0.01), using all behaviors. Using sexual behaviors, HH was different from HT (t = 2.45, 0.02 > P > 0.01) and from HC (r = 2.52, 0.02 > P > 0.01). TH showed the biggest mean repertoire size and TC showed the small- est one (all behaviors; t = 4.94, P < 0.001). CH was significantly different than CC (all behav- iors: t = 3.0, 0.01 > P > 0.001; sexual behav- iors: t = 4.5, P < 0.001). In the second type of comparisons, HH was significantly different than CH (all behaviors: t = 2.19, 0.05 > P > 0.02; sexual behaviors: t = 2.02, 0.05 > P > 0.02). In the intraspe- cific trials, sexual behaviors such as Leg wav- ing, Drumming and Rubbing (and Explosive locomotion in L. thorelli male) were usually performed. In the interspecific trials only L. carbonelli males performed some sexual be- haviors. Hybrid males performed Leg waving. Drumming, Palpation and Leg tapping as sex- ual behaviors in the presence of parental sex- ual pheromones. COSTA ET AL.— SEXUAL BEHAVIOR IN HYBRID LYCOSIDS 239 Table 3. — Some behaviors observed in direct male-female encounters. Only presence of each unit in the experiences are listed. No sex = absence of sexual behavior. Experimental groups identified as in Table 1. Behavior abbreviations are LW = leg waving, Ru = rubbing of legs, Dr = drumming, AV = abdominal vibrations, EL = explosive locomotion, RP = rest posture. At = attack. One experiment of the CT group was deleted due to the absence of sexual behavior during the 20 minute period. Group (n) No sex Male Female Copulation LW Ru Dr AV EL LW Dr RP At HH (5) 0 4 2 4 1 0 0 0 0 4 0 HC (3) 2 1 1 1 1 0 1 1 1 2 0 HT (3) 1 2 1 1 0 0 1 0 1 2 0 CH (3) 2 0 0 1 0 0 0 0 0 2 0 CC (3) 0 3 1 3 1 0 1 1 1 2 1 CT (2) 2 0 0 0 0 0 0 0 0 1 0 TH (3) 1 1 0 1 0 0 0 0 0 2 0 TC (3) 1 0 0 2 0 0 0 0 0 2 0 TT (3) 0 3 0 1 1 3 0 0 0 1 1 Data for male-female encounters are showed in Table 3. In one HC experiment, the male per- formed four unsuccessful mount atttempts. Our results suggest that Leg waving, Drum- ming and Rubbing may be essential visual and acoustic signals for species recognition. These behaviors had constant species-typical char- acteristics (rhythms and angles) even when exposed to different pheromones. Parental fe- males would discriminate slight differences in movement frequencies and angles from the signalling tnales. In Lycosa malitiosa Tullgren 1905, for example, the males were not rec- ognized by conspecific females when the sex- ual signalling frequencies were experimental- ly changed (Costa & Sotelo 1983). Taking into account the complete precopulatory isolation between L. thorelU and L. carbonelU (Costa & Capocasale 1984), the absence of recogni- tion of hybrid males by parental females was expected. However, the intermediate charac- teristic of the hybrid male signals elicited less intense rejections by parental females than the heterospecific males. In the present study males showed a nar- rower repertoire than when exposed to hybrid pheromone (Costa et al. 1997) in both sexual and all behaviors. The absence of palpation in all its forms in males exposed to parental pheromones is remarkable, because of its oc- currence in the presence of hybrid pheromone (Costa et al. 1997). This fact could be ex- plained assuming that the hybrid pheromone is composed of species-specific tactochemical elements from both parental species, then in- creasing the male repertoires. In agreement with Costa & Capocasale (1984), L. carbo- nelli and L. thorelli males showed a poor rep- ertoire when tested with the heterospecific pheromone. The absence of reaction in males in the two TC cases also supports this view. In direct male-female encounters, male L. carbonelU were best at discriminating, be- cause they displayed low sexual activity in re- sponse to L. thorelli and hybrid females (Table 3). This is in agreement with the results re- ported by Costa & Francescoli (1991) using anesthetized females. Hybrid males were the least discriminating, but they did not succeed in obtaining copulation. The low attack level in female L. thorelli could be considered as indicative of sexual re- ceptivity. The low level of sexual displaying in female L. thorelli does not indicate non-recep- tivity because these females are usuaUy passive (Costa & Capocasale 1984). Our results show the absence of receptivity in hybrid females test- ed with the three types of males, and in parental females tested with heterospecific males. Strat- ton & Uetz (1986) reported similar responses in hybrid females of Schizocosa ocreata and S. rovneri, and rejection of hybrid males by paren- tal females. The moderate tolerance of parental females to hybrid males would be based on the presence of some elements firom both parental courtship behaviors. The occurrence of copulations in conspe- cific experimental groups indicates that the laboratory conditions did not affect sexual communication. Thus, the absence of copu- 240 THE JOURNAL OF ARACHNOLOGY lations in the other groups suggests that nat- ural hybrids — if they occur — will not repro- duce. However, a L. carbonelli female received an intense courtship and repeated mounting attempts from one hybrid male. This female was receptive probably due to the rec- ognition of some species-specific signals; but, at the mounting attempt, she could have de- tected chemotactile information from the males’ integument, allowing rejection. The characteristics of both parental species’ courtship displays agree with the hypothesis from Bristowe & Locket (1926) on the origin of those displays by rituafization of searching movements. Furthermore, both species show similar behaviors when exposed to sex phero- mone, but in presence of conspecific females, L. thorelli males change their behavior while L. carbonelli males maintain the searching pattern (Costa & Capocasale 1984). The common an- cestor would have had a similar pattern to that of L. carbonelli because the pattern is performed in the searching phase by both species. Although sympatric, L. thorelli are captured mainly in sunny short-grass areas, whereas L. carbonelli are captured in tail-grass areas, in- cluding dark and humid places. The Explosive locomotion performed by a L. thorelli male would only be seen in open areas. L. carbo- nelli shows a pattern fitted to dark and closed areas with multiple obstacles, consisting of “cautious” locomotion, and a high occurrence of Leg waving using their long legs. These two different habitats may have determined the distinctive characteristics of the observed courtship patterns. The precopulatory isolation between L. tho- relli and L. carbonelli could have evolved by a process of alteration in the communication codes, from a mutation or recombination of the genes responsible of the signalling frequency. Indeed, movement frequencies during some displays were greater in L. thorelli than L. car- bonelli (Costa et al. 1997). Also, Explosive lo- comotion could have originated by a Lo/LW frequency increase alternating with prolonged immobility periods. In this process the well- known high selectivity level of the female should play the main role (Suwa 1985). Strat- ton & Uetz (1986) suggested that Schizocosa ocreata and S. rovneri speciated by alterations in the courtship pattern of their ancestor. In those species these authors postulated a model involving a mutation in “single autosomal loci.” Results fromL. thorelli andL. carbonelli suggest that more complex genetic determina- tion mechanisms are involved. LITERATURE CITED Bristowe, W.S. & G.H. Locket. 1926. The court- ship of British lycosid spiders, and its probable sigrificance. Proc. Zool. Soc., 22:317-347. Costa, EG. & R.M. Capocasale. 1984. Lycosa car- bonelli sp. nov.: una etoespecie simpatrida, si- bilina de Lycosa thorelli (Keyserling) (Araneae, Lycosidae). J. ArachnoL, 11:423-431. Costa, EG. & G. Francescoli. 1991. Analyse ex- perimentale de I’isolement reproductif entre deux especes jumelles et sympatriques d’araignees: le Lycosa thorelli (Keyserling) et le Lycosa car- bonelli Costa et Capocasale. Canadian J. Zool., 69:1768-1776. Costa, EG. & J.R Sotelo. 1983. Estudio preliminar sobre la interaccion macho-hembra de Lycosa malitiosa (Araneae, Lycosidae) a distintas tem- peraturas experimentales. Resumenes y Comun- icaciones de las III Jomadas de Ciencias Natur- ales, Montevideo (Uruguay). Pp. 42-44. Costa, EG., G. Francescoli & C. Viera. 1992. Estudio preliminar del comportamiento de machos de Ly- cosa thorelli, Lycosa carbonelli y sus hibridos en presencia de feromona sexual hibrida (Araneae, Ly- cosidae). Bol. Soc. Zool. Uruguay, 7:3-4. Costa, EG., C. Viera & G. Francescoli. 1997. Male sexual behaviour elicited by a hybrid pheromone: A comparative study on Lycosa thorelli, L. car- bonelli, and their hybrid progeny (Araneae, Ly- cosidae). Canadian J. Zool., 75:1845-1856. Francescoli, G & EG. Costa. 1992. Postemergence development in Lycosa carbonelli Costa and Ca- pocasale, L. thorelli (Keyserling), and their hybrid progeny (Araneae, Lycosidae): A comparative laboratory study. Canadian J. Zool, 70:380-384. Stratton, G.E. & G.W. Uetz. 1986. The inheritance of courtship behavior and its role as a reproduc- tive isolating mechanism in two species of Schi- zocosa wolf spiders (Araneae: Lycosidae). Evo- lution, 40:129-141. Suwa, M. 1985. Why does the ability to distinguish the mating partner of the same species in the wolf spider differ between the male and the fe- male? J. Ethol. (Kyoto), 3:79-82. Fernando G* Costa: Etologia, Division Zoologia Experimental, IIBCE. Av. Italia 3318, Montevideo, Uruguay Carmen Viera and Gabriel Francescoli: Seccion Entomologia and Etologia, Facul- tad de Ciencias. Igua 4225, Montevideo 11400, Uruguay Manuscript received 20 August 1998, revised 13 July 1999. 2000. The Journal of Arachnology 28:241-242 RESEARCH NOTE ESTIMATING FORAGING INTAKE: A COMMENT ON TSO AND SEVERINGHAUS (1998) Keywords: Energy intake, allometry, digestible biomass Tso & Severinghaus (1998) have recently drawn attention to the problems of using Schooner’s (1980) length- weight equations for insects as a means of estimating foraging in- take by spiders. They pointed out that a bio- mass estimate calculated from Schooner’s equations includes all tissues irrespective of whether they are digestible or not. As preda- tors do not utilize largely indigestible material (e.g., the exoskeleton — -but note that at least some spider species can produce chitinase (Collatz 1987)) this will inevitably lead to in- accuracies in determining digestible biomass acquisition by a foraging spider. Tso & Sev- eringhaus (1998) therefore measured the ac- tual biomass removed from a variety of prey items of different sizes and taxonomic groups by caged female Argiope trifasciata (Forskal 1775). They did this by simply subtracting the weight of the discarded exoskeleton from the initial wet weight of the prey. Dry weights of the prey items used were estimated by insert- ing prey lengths into Schoener’s (1980) equa- tions for each of the various taxonomic groups. Plots of ingested biomass and of dry weights of prey against prey length yielded curves that increasingly diverged towards higher prey lengths. Tso & Severinghaus as- sume dry weight is composed of digestible biomass -f exoskeleton; ingested biomass is composed of digestible biomass + water; the relative proportions of these three components are constant and the absolute contribution of each is a function of size. This being so, be- cause water comprises a large proportion of the wet weight (and therefore, ingested bio- mass) the absolute difference between ingest- ed biomass and dry weight will be a positive function of size — the plotted curves will di- verge with increasing body length. The con- clusion is that because many large spiders take a great range of prey sizes “. . . the relative energy content of large prey would be greatly underestimated if determined by dry weight alone.” They recommend that “future studies should consider using ingestible biomass of prey in estimating the foraging intake of spi- ders.” Tso & Severinghaus (1998) argue that using dry weights will tend to underestimate the in- gested biomass, and disproportionally so with increasing prey sizes. However, the ingested biomass they suggest measuring is still not a good estimate of the energy derived from the prey because a large proportion of this bio- mass will be the water responsible for the di- vergence of the plotted curves; water is not a source of metabolic energy. In very dry hab- itats, the water content of the biomass ingest- ed from a prey item may indeed be of great importance, and the total volume of liquified food ingested will certainly be a factor in de- termining satiation level in situations where prey is not limiting. Wet biomass ingested will only be proportional to energy intake if the separate components of water and digestible biomass are in constant proportions (as as- sumed by Tso & Severinghaus) in different sized prey. This will only be the case if water content and digestible biomass both scale with size in exactly the same way. One might ex- pect both to be approximately proportional to volume (i.e., oc length^) but the exact expo- nents would have to be determined empirical- ly (see Schoener 1980), and their coefficients (0.7 for water and 0.1 for digestible macro- molecules in the equation of Tso & Severin- ghaus) checked for constancy across prey size range. 241 242 THE JOURNAL OF ARACHNOLOGY An appropriate measure that is likely to be a direct function of energy intake from a prey item is total dry weight (= digestible dry weight + exoskeleton) less the weight of the dry exoskeleton rejected after feeding. The absolute intake of digestible dry weight must, of necessity, always be less than the total dry weight of the prey and will therefore fall be- low the lower curves in Tso & Severinghaus’s fig. 1. Within this constraint, the shape of the digestible dry weight curve will depend on its allometric relationship with absolute size. AL though digestible dry weight probably scales approximately oc length^ (but, again, see Schoener 1980) exoskeleton probably reflects more closely surface area (i.e., oc length^). As total dry weight increases with size, one would therefore expect a greater proportion to be represented by digestible material in larger prey items. Some evidence for this is provided by Rees (1986) who investigated the relation- ship between the fraction of total (wet) mass attributable to dry skeletal mass and total wet mass across taxa within six beetle families. The slopes of all six plots were negative (two- tailed sign test, P = 0.03), although only one was individually significant. Total mass was measured as wet rather than dry weight, but if the degree of tissue hydration is constant or, if variable, not a function of beetle size, these data suggest that skeletal mass decreases and, as a consequence, the remainder (digestible mass) increases with total beetle mass (size). This is in direct contrast to the conclusions of Tso & Severinghaus quoted above — the use of total dry weight as a surrogate for energy availability will produce an underestimate that decreases with increasing prey size. If energy intake is the currency of interest when inves- tigating spider foraging, ingested dry weight is the appropriate, and direct, measure to use. I would like to thank Peter Mayhew and Chris Rees for discussion. LITERATURE CITED Collatz, K.-G. 1987. Structure and function of the digestive tract. Pp. 229-238, In Ecophysiology of Spiders (W. Nentwig, ed.). Springer- Verlag, Berlin. Rees, C.J.C. 1986. Skeletal economy in certain herbivorous beetles as an adaptation to a poor dietary supply of nitrogen. Ecol. Entomol., 11: 221-228. Schoener, TW. 1980. Length-weight regression in tropical and temperate forest-understory insects. Ann. Entomol. Soc. America, 73:106-109. Tso, I-M. & L.L. Severinghaus. 1998. Ingested biomass of prey as a more accurate estimator of foraging intake by spider predators. J. ArachnoL, 26:405-407. G.S. Oxford: Department of Biology, Uni- versity of York, P.O. Box 373, York YOlO 5YW, U.K. Manuscript received 6 March 1999, revised 3 No- vember 1999. 2000. The Journal of Arachnology 28:243-244 RESEARCH NOTE A TEST OF POLLEN FEEDING BY A LINYPHIID SPIDER Keywords: Araneae, nutrition, Pinus, cage For several weeks starting in January or February each year, fallout of pine pollen forms a yellow coating on the upper surface of almost everything present in the xeric, up- land habitats on the Lake Wales Ridge in south-central Florida. Juvenile orb-weavers, Araneus diadematus Clerck (Araneidae), and crab spiders, Thomisus onustus Walckenaer (Thomisidae), at high latitudes are known to greatly enhance their life expectancy in spring by feeding on pollen (Smith & Mommsen 1984; Vogelei & Greissl 1989). Hence, we reasoned that web spiders in Florida scrub might consume pine pollen adhering to their silk for added nutrition when insects typically are in short supply. We tested this idea using individually housed, female bowl and doily spiders, Frontinella pyramitela (Walckenaer) (Linyphiidae). We selected this spider because it is locally abundant on the Lake Wales Ridge in winter and its sheet webs become exten- sively coated with pine pollen. We collected adult and subadult F. pyram- itela (n = 36) at the Archbold Biological Sta- tion, Highlands County, Florida in February 1995 and transported them alive back to the laboratory in Missouri. We weighed the spi- ders to the nearest 0.01 mg and then placed them individually in cages made from recy- cled, 2-liter plastic carbonated beverage bot- tles. Each cage contained four vertical glass rods (20 cm X 4 mm o.d.) arranged in a square ~5 cm on a side to provide support for a spider’s web (Fig. 1). A cage was prepared by cutting off the bottom part of a transparent bottle, embedding the rods in a 2 cm thick layer of patching plaster poured into the bot- tom section, and taping the capped top section back on the bottom section after the plaster was dry. Once sealed in this manner, the in- expensive bottle cage proved to be mold-free and almost airtight. Two days after their introduction into the bottle cages at 22-26 °C under constant illu- mination, all spiders had spun typical sheet webs on the glass rods. On days 3, 8, and 13 we misted the contents of each cage with wa- ter sprayed through the bottle’s orifice. On day 4 we assigned equal numbers of spiders (n = 12) at random to one of three treatments: Un- fed, Pollen-Fed, and Fly-Fed. On days 4, 9, and 14 we uncapped the cage of each Pollen- Fed spider and manually stripped much pollen from two ripe strobili of the South Florida slash pine, Pinus elliottii Engelm. var. densa Little & Dorman, sufficient to coat the entire web. To retain nutrients, the strobili were kept frozen at —20 °C in plastic bags after collec- tion at the Archbold Biological Station, On the same three days we fed 5-7 adult Dro- sophila melanogaster Meigen to each spider in the Fly-Fed group. On day 19, we opened every cage and re- weighed the spiders. The initial masses of the spiders were high- ly uniform (Mean ± S.E.M. = 2.26 ± 0.14 mg; Coefficient of Variation = 0.0080). But at the end of the tests, spiders given a diet of fruit flies had gained an average of 2.29 ± 0.61 mg. In contrast, the Pollen-Fed spiders each had lost 0.29 ± 0.07 mg, an amount sta- tistically equivalent to the mass lost by an Un- fed spider (0.24 ± 0.0.14 mg). The Fly-Fed spiders were significantly heavier than spiders in either of the other two groups (ANOVA, F = 20.79, df ^ 2, P < 0.0001). Hence, we conclude that F. pyramitela did not consume pine pollen in amounts sufficient to maintain body mass. In addition, extensive observa- tions never revealed any behavior that might suggest this spider was actively consuming 243 244 THE JOURNAL OF ARACHNOLOGY Figure 1 . — Side view of the sheet web of F. pyr- amitela suspended from four glass rods inside a cage made from a transparent, 2-liter plastic, car- bonated beverage “soda” bottle. The top of the bot- tle is held in place by wide adhesive tape. Removal of the tape allows one to have free access to the spider on its web. pollen or removing pollen from its silk, al- though we often saw spiders attack fruit flies entrapped in their webs. Pine pollen is an abundant, although short- lived resource that is known to be the pre- ferred dietary component for the rare blister beetle, Lytta polita Say, which emerges in winter in south-central Florida (Carrel et al. 1990). This large insect consumes staminate Pinus cones, much as Americans eat com-on- the-cob. Furthermore, chemical analysis showed that pine pollen is nearly as nutritious as pollen collected by honeybees foraging at nearby flowers (J.E. Carrel & J. Bull unpubl. data). Hence, even though bowl-and-doily spi- ders did not gain weight when offered pine pollen, it is likely that other species, in partic- ular orb-weavers that ingest silk as they take down their adhesive spirals, consume pine pollen trapped in their webs. The soda bottle cages have proven to be very suitable for long term studies of linyphiid spiders. For example, we reared several gen- erations of F. pyramitela in these cages, al- lowing us to rapidly repeat work on web building, predation, and the pheromonal basis of courtship in this spider (Suter & Renkes 1982, 1984; Suter & Hirscheimer 1986). In addition, we now are testing the chemical ba- sis of prey discrimination by small araneids housed in the bottle cages. Because the cages are disposable, there is no possibility of carry- over of chemical residues from test to test. We thank the Archbold Biological Station for providing research facilities, Jessica Whit- ed for making the illustration, Jan Weaver for technical help, and the University of Missouri Research Board for partial funding of this pro- ject. LITERATURE CITED Carrel, J.E., J.M. Wood, Z. Yang, M.H. McCairel «fe E.E. Hindman. 1990. Diet, body water, and hemolymph content in the blister beetle Lytta polita (Coleoptera: Meloidae). Environ. Ento- mol., 19:1283-1288. Smith, R.B. & TR Mommsen. 1984. Pollen feed- ing in an orb-weaving spider. Science, 226: 1330-1332. Suter, R.B. & G. Renkes. 1982. Linyphiid spider courtship: releaser and attractant functions of a contact sex pheromone. Anim. Behav., 30:714- 718. Suter, R.B. & G. Renkes. 1984. The courtship of Frontinella pyramitela (Araneae, Linyphiidae): Patterns, vibrations and functions. J. ArachnoL, 12:37-54. Suter, R.B. & A.J. Hirscheimer. 1986. Multiple web-bome pheromones in a spider Frontinella pyramitela (Araneae: Linyphiidae). Anim. Be- hav., 34:748-753. Vogelei, A. & R. Greissl. 1989. Survival strategies of the crab spider Thomisus onustus Walckenaer 1806 (Chelicerata, Arachnida, Thomisidae). Oec- ologia, 80:513-515. James E. Carrel, Heather K. Burgess and Dennis M. Shoemaker: Division of Bi- ological Sciences, University of Missouri, Columbia, Missouri 65211-7400 USA Manuscript received 28 May 1999, revised 5 Oc- tober 1999. 2000. The Journal of Arachnology 28:245-247 RESEARCH NOTE COMPARISON OF THE FERTILITY BETWEEN LOXOSCELES INTERMEDIA AND LOXOSCELES LAETA SPIDERS (ARANEAE, SICARHDAE) Keywords: Fertility, sinanthropy, loxoscelism, egg-sac production Envenomation by brown spiders of the ge- nus Loxosceles Heinecken & Lowe 1832 of North America, the Middle East, South Africa and South America commonly results in a lo- cal necrotic skin lesion and sometimes causes systemic effects that can lead to the death of the patient (Denny et al. 1964; Efrati 1969; Newlands 1982; Gerstch 1967; Gerstch & En- nik 1983; Futrell 1992). Loxosceles spp. are the most poisonous spiders in Brazil and chil- dren who develop the more severe systemic effects after envenomation nearly always die. At least three different Loxosceles species of medical importance are known in Brazil: L. intermedia Mello-Leitao 1934, L. laeta (Nic- olet 1849) and L. gaucho Gertsch 1967. More than 1500 cases of envenomation by L. inter- media alone are reported each year. Because of a lack of understanding of the mechanism of action of the venom, an effective treatment is not available. Loxosceles are nocturnal and non-aggres- sive spiders. In the natural environment, they live under rocks, inside tree holes and other places that may serve as shelter. While some occupy hot and arid regions, others inhabit relatively damp areas. They also live in dark, dry places in houses, such as doorsteps, wall cracks, spaces behind pictures, furniture or even curtains, as well, in household rubbish and buildings (Gertsch 1967; Gertsch & Ennik 1983). Loxosceles intermedia prevails in the urban environment of the states of Parana and Santa Catarina (south region of Brazil) (Fischer 1994; Mattosinho et al. 1997). This species is restricted to the southern regions of South America including Brazilian Federal District (middle west region), the states of Rio de Ja- neiro and Sao Paulo (southeast region), Rio Grande do Sul (south region), and also in Ar- gentina. The distribution of L. laeta is much wider, and it can be found throughout South America including Peru, Chile, Ecuador, Bra- zil (from the state of Paraiba to the state of Rio Grande do Sul, from the northeast region to south region), Uruguay and Argentina. Ac- cording to Gerstch (1967), L. laeta has also spread to some parts of North America, being found in Massachusetts and other locations due to its sinanthropy (Levi & Spielman 1964). In Brazil, L. laeta is also found in the same States as L. intermedia. It prevails in the south of Santa Catarina State (south region) (Mattosinho et al. 1997) and, in Curitiba city (Parana State, south region) during June and July, although being less abundant than L. in- termedia. Although L. intermedia and L. laeta can be both found in the south region of Brazil, there has been a significant increase in the number of Loxosceles bites mainly associated with L. intermedia which seems to be positively cor- related with the expansion of this species’ range (Ribeiro et al. 1993). The present study was performed to compare the fertility of the two species reared in laboratory to better un- derstand expansion of the L. intermedia pop- ulation in the south region of Brazil. This study was conduced in “Biotmo de Criagao e Manutengao de Aranhas” of the Im- munochemistry Laboratory, Butantan Insti- tute, Sao Paulo, Brazil. The spiders used in this study were collected in the town of Cam- po Alegre (Santa Catarina State, south region, Brazil) from June to August. The sampled group of females, fertilized in the natural en- vironment, comprised 108 L. intermedia and 245 246 THE JOURNAL OF ARACHNOLOGY 47 L. laeta. They were transferred to plastic boxes (9.5 cm diameter X 5.5 cm high) and kept in the laboratory under normal environ- mental temperature and relative humidity (19.3 °C ± 2.8 and 81.3% RH ± 2.07). The spiders were fed with cockroach nymphs {Pycnoscellus surinamensis, Dictyopthera, Blaberidae) or with darkling beetle larvae (Te~ nebrio mollitor, Coleoptera, Tenebrionidae) twice a month. All specimens were observed weekly for eight months. During this period, the follow- ing variables were evaluated: number of egg sacs per spider, total number of eggs per egg sac, total number of spiderlings hatched per egg sac and time for spiderlings to hatch. The mean of the values was compared using a two-tailed t-test at a significance level of 0.05. The results show that the differences be- tween the mean number of egg sacs per spider of L. intermedia [1.79 ± 0.83] and L. laeta [1.67 ± 0.84] were not statistically significant (Fig. lA). However, the mean number of eggs per egg sac per spider and as well as the total number of eggs was significantly higher for L. laeta (Fig. IB). Mean times to hatching for L. laeta spiderlings were significantly greater for L. laeta than L. intermedia (Fig. 1C). The per- centage of hatched spiderlings was high but did not reveal statistically significant differ- ences between the two species. The mean number of the egg sacs produced per female was similar for both species, the maximum was five egg sacs for L. intermedia and four for L. laeta; the minimum was one egg sac for both species. These results differ from those of Galiano & Hall (1973) who de- scribed up to 15 egg sacs per female of L. laeta. However, those females were mated un- der laboratory conditions, which makes it pos- sible to record all the egg sacs produced per female. Nevertheless, it cannot be excluded that, because they were not feeding in the nat- ural environment, they may possibly have had enhanced fertility. Hite et al. (1966) described up to five egg sacs per female of L. reclusa, while Fischer (1996) observed up to three egg sacs for L. intermedia. As in our study, these authors observed adult females collected in their natural environment, and therefore the possibility that they had produced previous egg sacs could not be excluded. The Loxos- celes spiders can live from 3-7 years (Galiano & Hall 1973; Lowrie 1980, 1987). The age of Figure 1. — Comparison of the fertility of Lox~ osceles intermedia and Loxosceles laeta collected as fertilized adults. (A) Number of egg sacs per spider, (B) Number of eggs per egg sac, and (C) Time to hatching of spiderlings. The results ae ex- pressed as mean ± SD. the spiders collected can also affect the quan- tity of egg sacs produced per spider. The analysis showed that L. laeta exceeded L. intermedia in both the total number of eggs produced per female and per egg sac. These results may reflect differences in body weight between the two species. The females of L. laeta were larger and heavier [1.161 cm ± 0.52; 0.2115 g ± 0.026] than L. intermedia [1.096 cm ± 0.093; 0.1260 g ± 0.035] (Cris- tina de Oliveira et al, 1999; G. de Andrade GONSALVES DE ANDRADE ET AL.— FERTILITY IN LOXOSCELES 247 unpubl. data), and such differences might al- low the former species to have a greater ovi- position potential. It is well-known that fecun- dity tends to be correlated with body mass for female invertebrates, including spiders (Hig- gins 1992; Fischer 1996; Schneider 1996). Under the conditions of this study, the means of the total number of eggs produced per spider and per egg sac were greater for L. laeta which suggests that a greater fertility could be ascribed to L. laeta than to L. inter- media. If so, these considerations suggest that the significant expansion of L. intermedia in the south region of Brazil is not due to a great reproductive rate of that species. Studies on the ecological aspects of the sinanthropy of both species, as well as the possible environ- mental alterations in the south region of Bra- zil, may explain the predominance of the L. intermedia spiders. LITERATURE CITED Cristina de Oliveira, K., R.M. Gongalves de Andra- de, A.L. Giusti, W. Dias da Silva & D.V. Tam- bourgi. 1999. Sex-linked variation of Loxosceles intermedia spider venoms. Toxicon, 37:217-221. Denny, W.E, C.J. Dillaha & P.N. Morgan. 1964. Hemolytic effect of Loxosceles reclusa venom: in vivo and in vitro studies. J. Lab. Clin. Med., 64:291-298. Efrati, R 1969. Bites by Loxosceles spiders in Is- rael. Toxicon, 6:239-241. Fischer, M.L. 1994. Levantamento das especies de Loxosceles Heinecken & Lowe, 1832 no muni- cipio de Curitiba, Parana, Brasil. Estudos de Biologia, 3:63-88. Fischer, M.L. 1996. Biologia e Ecologia de Lox- osceles intermedia, Mello & Leitao, 1934 (Ara- nea: Sicariidae) no munfcipio de Curitiba, PR. (Dissertagao de mestrado. Ciencias Biologicas, Ufpr J3Y Futrell, J.M. 1992. Loxoscelism. J. Med. Sci., 304(4):261-267. Galiano, M.E. & M. Hall. 1973. Datos adicionales sobre el ciclo vital de Loxosceles laeta (Nicolet, 1849) (Araneae). Physis, 32(85):277-288. Gertsch, W.J. 1967. The spider genus Loxosceles in South America (Araneae: Scytodidae). Bull, Mus. Nat. Hist, 136:119-183. Gertsch, W.J. & E Ennik. 1983. The spider genus Loxosceles in North America, Central America and the West Indies (Araneae: Loxoscelidae). Bull. Mus. Nat. Hist., 175:263-360. Higgins, L.E. 1992. Developmental plasticity and fecundity in the orb-weaving spider Nephila cla- vipes. J. ArachnoL, 20:94-106. Hite, M.J., W.J. Gladney, J.L. Lancaster, Jr. & WH. Whitcomb. 1966. Biology of brown recluse spi- der. Arkansas Agric. Exp. Stat. Bull., 711:2-26. Levi, H.W. & A. Spielman. 1964. The biology and control of the South American brown spider Lox- osceles laeta (Nicolet), in a North American fo- cus. J. Trop. Med. Hyg., 13:132-136. Lowrie, D.C. 1980. Starvation longevity of Lox- osceles laeta (Nicolet) (Araneae). Entomol. News, 91(4): 130-132. Lowrie, D.C. 1987. Effects of diet on the devel- opment of Loxosceles laeta (Nicolet) (Araneae, Loxoscelidae). J. ArachnoL, 15:303-308. Mattosinho, S.G., U.M. Sezerino, M. Zannin, M, Grando, J.L. C. Cardoso, V.R.D. von Eickstedt & F.O.S. Fran§a. 1997. Geographic distribution of Loxoscelism in Santa Catarina (Brazil) and species of Loxosceles sp, involved and existent in the state. J. Ven. Anim. Tox., 3(1):99. Newlands, G., C. Isaacson & C. Martindale. 1982. Loxoscelism in the Transvaal, South Africa. Trans. Royal Soc. Trop. Med. Hyg., 76(5):610- 615. Ribeiro, L.A., V.R.D. von Eickstedt, G.B.G. Rubio, J.F. Konolsaisen, Z. Handar, M. Entres, V.A.F.P de Campos & M.T. Jorge. 1993. Epidemiologia do acidente por aranhas do genero Loxosceles Heinecken & Lowe no Estado do Parana. Mem. Inst. Butantan, 55:19-26. Schneider, J.M. 1996. Differential mortality and relative maternal investment in different life stages in Stegodyphus llineatus (Araneae, Eresi- dae). J. ArachnoL, 24:148-154. Rule Maria Gonsalves de Andrade: Lab- oratory of Immunochemistry, Butantan In- stitute, Av. Vital Brazil, 1500, 05503-900, Sao Paulo, SP, Brazil Wilson R. Lourengo: Laboratoire Zoologie (Arthropodes) Museum National d’Histoire Naturelle’61, rue de Buffon 75005, Paris, France Denise Vilarinho Tambourgi: Laboratory of Immunochemistry, Butantan Institute, Av. Vital Brazil, 1500, 05503-900, Sao Paulo, SP, Brazil Manuscript received 10 July 1998, revised 6 Oc- tober 1999. 2000. The Journal of Arachnology 28:248-250 RESEARCH NOTE GROUP DISPERSAL IN JUVENILE BRACHYPELMA VAGANS (ARANEAE, THERAPHOSIDAE) Keywords: Tarantula behavior, spiderling aggregration Cutler & Guarisco (1995) summarized the literature accounts of juvenile dispersal by mygalomorph spiders and noted that such ob- servations were rare. None of those reports involved the Theraphosidae because juvenile behaviors associated with dispersal had not been described for this family. This report documents three instances of aggregative ju- venile dispersal by theraphosid spiders ob- served in the Lamanai Archaeological Re- serve, Orange Walk District, Belize (17°45'08"N, 88°39'25"W). On 26 May 1998, and 8-9 June 1999, ju- venile Brachypelma vagans (Ausserer 1875) were observed in large aggregations and were apparently dispersing from their natal burrows in a manner not previously described for any spider. Only one other large theraphosid oc- curs in the area, Crassicrus lamanai Reichling & West 1996, and juveniles of this species are distinguished by their uniform light gray col- or, as opposed to juvenile B. vagans which are pale with a black spot on the abdomen. In addition, C. lamanai prefers cleared habitat and is rare in dense forest where these obser- vations were made. Each encounter took place at night between 2015-2115 h on a dirt road leading into old growth secondary forest. In each instance, groups of spiderlings (number- ing 72, 76, and 135 respectively) were walk- ing in single file, forming a line which slowly snaked its way along the road. From a dis- tance the processions resembled a column of ants, and the largest of these aggregations formed a line 1.09 m in length (Fig. 1). The spiders maintained close proximity to one an- other while walking, often lightly touching the abdomen of the individual ahead of them with their front legs. The spiderlings were observed for 8-12 m as they moved diagonally across the road and into the vegetation. A thorough daytime search of the surrounding area re- vealed that the nearest burrows occupied by adult B. vagans were —50 m from the site where the June observations were made. The spiders were disturbed by the direct beam of a flashlight; or if approached too closely, they stopped their progression and scattered slightly. Once the disturbance ceased they reassembled in single file and proceeded as before. On two occasions road dust was sprinkled across the spiderlings’ path in a small gap which had formed in the line. When they reached the road dust the spiderlings stopped and began milling about as the ones ahead of them continued on their way. After a minute, the spiderlings began moving in the same general direction as before and appeared to recapture the trail beyond the dust. The tarantula occupying the front of the line changed frequently. As the leading spider took a slight turn to the left or right, the spider behind it would move ahead and take over the lead while the previous leader would insert itself farther back in line. This replacement of the leading spider occurred every 7-10 cm. Terrestrial theraphosid spiders often occur in dense, local aggregations, with burrows abundant in some locations but absent in ad- jacent sites which represent similar habitat (Baerg 1958). These assemblages exhibit un- derdispersed distribution patterns, with nearest neighboring burrows in closer proximity than would be predicted by chance alone (Reich- ling 1999). Some fossorial lycosids also occur in clusters. Geolycosa xera archboldi Mc- Crone 1963, a sandhill endemic of central Florida, typically disperse less than one meter 248 REICHLING— GROUP DISPERSAL IN BRACHYPELMA 249 from their maternal burrow and settle within one hour, leading Marshall (1995) to propose that the burrow aggregations characteristic of this taxon are due to highly restricted juvenile dispersion distances. This mechanism does not appear satisfactory for explaining thera- phosid aggregations in light of the consider- able distance from potential maternal burrows that dispersing B. vagans were seen. The observations described here suggest a plausible explanation for the clustered spatial patterns of theraphosid spiders. With the ex- ception of mature males, terrestrial theraphos- ids are rarely observed far from their burrow, and it is likely that an individual’s burrow site remains close to the location where it was first established by the juvenile. If the mass move- ment of single clutches of B. vagans continues until the juveniles settle and establish resi- dence, it would result in the aggregations characteristic of tarantulas in Belize and else- where. The hypothesis that these clusters are composed primarily of siblings can be tested. I thank the Conservation Division of the Belize Forest Department for permission to conduct field research. This work was sup- ported by the Lamanai Field Research Center, Indian Church Village, Belize, and I am grate- ful to its owners, Mark and Monique Howells. I thank Ann Reichling for her encouragement and support. LITERATURE CITED Figure 1 . — Line of juvenile Brachypelma vagans (arrow) moving across a dirt road in the Lamanai Archaeological Reserve, Orange Walk District, Be- lize. Some spiderlings have scattered due to the beam of a flashlight. Ausserer, A. 1875. Zweiter Beitrag zur Kenntnis der Arachniden -Familie der Territelariae Thorell (Mygalidae autor). Verhandl. K.K. Zool.-Bot. Gesell. Wien, 25:125-206. Baerg, WJ. 1958. The Tarantula. Univ. of Kansas Press, Lawrence, Kansas. Cutler, B. & H. Guarisco. 1995. Dispersal aggre- gation of Sphodros fitchi (Araneae, Atypidae). J. Arachnol., 23:205-206. Marshall, S.D. 1995. Mechanisms of the forma- tion of territorial aggregations of the burrowing wolf spider Geolycosa xera archboldi McCrone (Araneae, Lycosidae). J. Arachnol., 23:145- 150. McCrone, J.D. 1963. Taxonomic status and evo- lutionary history of the Geolycosa pikei com- plex in the south-eastern United States (Ara- neae, Lycosidae). American Mid. Nat., 70:47- 73. Reichling, S.B. 1999. Nearest neighbor relation- ships among theraphosid spiders in Belize. Southwestern Nat., 44:518-521. Reichling, S.B. & R.C. West. 1996. A new genus 250 THE JOURNAL OF ARACHNOLOGY and species of theraphosid spider from Belize Galloway Avenue, Memphis, Tennessee (Araneae, Theraphosidae). J. ArachnoL, 24:254- 38112 USA 261. Manuscript received 27 July 1999, revised 13 Oc- tober 1999. Steven B. Reichling: Memphis Zoo, 2000 2000. The Journal of Arachnology 28:251-253 RESEARCH NOTE HUNTING AND FEEDING BEHAVIOR OF ONE HETEROPODA SPECIES IN LOWLAND RAINFOREST ON BORNEO (ARANAE, SPARASSIDAE) Keywords: Heteropoda, hunting behavior Spiders in the genus Heteropoda Latreille 1804 (Sparassidae) are distributed in tropical Asia and Australia, with the exception of the cosmopolitan Heteropoda venatoria (Linnae- us 1758) (see Roewer 1954; Brignoli 1983; Platnick 1989, 1993, 1998; R Jager per. com- mun.). With the exception of the Australian species (Davis 1994), the genus Heteropoda as well as the entire family are unrevised (Ja- ger 1998). To date, two nominal species of Heteropoda have been described from Bor- neo, H. hosei Pocock 1897 and H. obtusa Thorell 1890; a generic revision is necessary to distinguish between these species (Jager pers. commun.). Heteropoda sp. from lowland rainforest in Kinabalu Park (Sabah, Malaysia) were observed in the field and laboratory. Voucher specimens, including two females, one male, and two juveniles, are deposited at the Field Museum of Natural History. Between April and July 1998, one of us (SA) conducted nightly surveys along streams in lowland rainforest (600 m elevation) near Poring Hot Springs in Kinabalu Park (Sabah, Malaysia, 6°03'N, 116°42'E). Surveys were conducted along 100 m transects at three sites in Kinabalu Park and one site in an agricul- tural area. Visual surveys were conducted by walking slowly along each edge of the stream, searching for spiders in, under, and on all sub- strates between the stream banks. We recorded the size of each spider and its position on the transect. We recorded the substrate type se- lected by each spider, the horizontal distance from the edge of the stream, the vertical dis- tance from the ground, and the nearest aquatic microhabitat type. We encountered a mean of 6.5 ±3.1 spi- ders per night on 100 m stream transects in primary forest. Significantly fewer spiders (2.2 ± 1.6) occurred along streams in an ag- ricultural area nearby (independent samples T = 3.85, df - 13.5, P - 0.002). Spiders were unevenly distributed, with a mean distance be- tween spiders of 16.8 ± 18.1 m in primary forest and 19.1 ± 23.3 m in the agricultural area. Spiders exhibited stereotyped microhab- itat selection. Ninety-three percent of the spi- ders in primary forest {n = 154) and 75% of the spiders in the agricultural area (w = 22) perched, facing downward, on boulders and small rocks at the edge of streams. In both primary forests and the agricultural area, spi- ders perched less than a meter from the water and the ground. On average, spiders perched a distance of 0.93 ± 1.27 m from the water in primary forest, and 0.69 ± 0.80 m in the agricultural area. Spiders perched a mean dis- tance of 0.44 ± 0.57 m from the surface of the water or the ground in both habitats. Spi- ders remained motionless unless disturbed. At one instance we observed a spider jumping from its perch on a rock and diving into the water. There it remained submerged for a pe- riod of time long enough for us to lose sight of it. We did not observe prey capture in the field. Thirteen spiders were housed in the labo- ratory for up to four weeks, each in 15 gallon (57 liters), clear, plastic cages with screen lids. Every five days the spiders were fed three prey items in cafeteria trials. Prey items in- cluded three species of frog larvae, Lepto- brachium montanum (Megophryidae), Meris- togenys orphnocnemis (Ranidae), and Rana signata (Ranidae), one species of fish, Glan- 251 252 THE JOURNAL OF ARACHNOLOGY Table 1. — Prey selection by Heteropoda sp. in laboratory trials. Prey Species n offered n eaten Cockroaches unknown 20 15 Fish Glanyops hanitschii 20 10 Large tadpoles Leptobrachium montanum 30 3 Small tadpoles Rana signata 10 0 Meristogenys orphnocnemis 10 0 yops hanitschii, and an unidentified species of cockroach. During each trial all spiders were given the same three prey. Live prey were placed in shallow, plastic trays (10X15X1 cm) filled with water on the floor of each cage. Cockroaches were released on the floor of the cage. We observed behavior of the spiders for up to 1 hour following the initiation of a feed- ing trial. We recorded attack, capture and prey handling behaviors. Between trials, spiders consistently oriented themselves above shallow trays of water in their cages. Spiders rested vertically on cage walls, facing downward, with their pedipalps and two front appendages resting lightly in the water. During cafeteria trials, spiders remained motionless until movement of the prey elicited a predatory response. Spiders generally at- tacked prey in the water by quickly lurching forward and piercing the skin of the prey with their fangs. After a successful capture, spiders climbed the cage wall to begin prey manipu- lation. Spiders used the front appendages and chelicerae to “fold” the prey in half, using silk to reinforce the fold. Spiders then fas- tened the prey to the cage wall with silk, re- leased hold of the prey, and began a stereo- typed weaving procedure. Spiders straddled the prey and moved in a counter-clockwise di- rection, rotating the body 360° directly above the prey while encircling the prey with silk. The spiders continued weaving until prey items were wrapped in tight packages of silk. After prey capture and manipulation, spiders began feeding on the prey. Spiders generally completed feeding less than 24 hours after prey capture and discarded the shrunken body of the prey at the bottom of the cage. In cafeteria trials, spiders consumed 75% of the cockroaches, 50% of the fish, and 10% of the large tadpoles (L. montanum) offered. Spi- ders did not capture or consume small tad- poles (M. orphnocnemis and R. signata). In trials with small and large tadpoles, spiders captured and consumed only large tadpoles. In trials with fish and large tadpoles, both prey were taken. In trials with cockroaches and tad- poles, or cockroaches and fish, spiders cap- tured and consumed cockroaches. Capture of cockroaches was significantly different from capture of aquatic vertebrate prey. Spiders were alerted by movement of the prey, and they attacked the prey with a swift and precise jump. One spider attacked a cock- roach with a vertical jump of over 20 cm (pers. obs.). Spiders did not use silk to subdue invertebrate prey. No terrestrial attacks were observed in the field, although we observed one spider in the process of consuming a moth. Hunting on the water surface has so far been reported from three spider families in various parts of the world (Pisauridae, Tre- chaleidae, and Lycosidae). Hunting on the wa- ter surface and feeding on aquatic and non- aquatic prey is known from three pisaurid genera; the worldwide Dolomedes Latreille 1804 (Bleckmann & Rovner 1984; Williams 1979), the African- Asian Thalassius Simon 1885 (Abraham 1923; Sierwald 1988), and the South American Ancylometes Bertkau 1880 (Schiapelli & Gerschman 1970); among the Trechaleidae it is known for members of the South American genus Trechalea Thorell 1869 (Berkum 1982). Among the Lycosidae, members of the genus Pirata Sundevall 1832 live in marshes and move over the water sur- face (Bristowe 1923; Ehlers 1939). Diving be- havior is reported for Dolomedes species and T. spinosissimus (Sierwald 1988). This rep- resents the first report of members of the fam- ily Sparassidae hunting on the water surface. We are grateful to the Malaysian govern- ment for permission to work in Sabah. We thank the director of Sabah Parks, Datuk Lam- ri Ali and the deputy director, Francis Liew, for permission to work in Kinabalu Park and for temporary staff housing at the field site. AIRAME & SffiRWALD— HUNTING AND FEEDING IN HETEROPODA 253 We thank Matthew Chatfield, Jacqueline Schlosser, and Frederick Francis for their valuable assistance in the field and laboratory. We gratefully acknowledge Peter Jager con- firming the genus identification. Field work was supported by grants from the Environ- mental Protection Agency Graduate Fellow- ship to Satie Airame and the GANN training grant to the Department of Ecology and Evo- lution at the University of Chicago. LITERATURE CITED Abraham, N. 1923. Observations on fish and frog- eating spiders of Natal. Ann. Natal Mus., 5(1): 89-95. Berkum, EH, van. 1982. Natural history of a trop- ical, shrimp-eating spider (Pisauridae). J. Arach- nol., 10:117-121. Bleckmann, H. & J. Rovner. 1984. Sensory ecol- ogy of a semi-aquatic spider (Dolomedes triton). 1. Roles of vegetation and wind-generated waves in site selection. Behav. Ecol. Sociobiol., 14: 297-301. Brignoli, P.M. 1983. A catalog of the Araneae de- scribed between 1940 and 1981. Manchester Univ. Press. Manchester. Bristowe, W.S. 1923. Spiders found in the neigh- bourhood of Oxshott. Proc. South London En- tomol. Nat. Hist. Soc. 1922/1923:1-11. Davies, V.T. 1994. The huntsman spiders Hetero- poda Latreille and Yiinthi gen. nov. (Araneae, Heteropodidae) in Australia. Mem. Queensland Mus., 35:75-122. Dippenaar-Schoeman, A.S. & R. Jocque. 1997. Af- rican Spiders: An Identification Manual. Plant Protection Research Institute, Pretoria, South Af- rica. Ehlers, M. 1939. Untersuchungen fiber Formen ak- tiver Lokomotion bei Spinnen. Zool. Jb., Abt. Syst, 72:373-499. Jager, P. 1998. First results of a taxonomic revision of the SE Asian Sparassidae (Araneae). Pp. 53- 59. In Proc. 17th European Coll, Arachnol. (P. Selden, ed.) Edinburgh. 1997. Platnick, N.I. 1989, Advances in Spider Taxono- my: 1981-1987. Manchester Univ. Press. Man- chester, UK. Platnick, N.I, 1993. Advances in Spider Taxono- my: 1988-1991, with Synonymies and Transfers 1940-1980. New York Entomol. Soc. & Amer- ican Mus. Nat. Hist. New York. Platnick, N.I. 1998. Advances in Spider Taxono- my: 1992-1995 with Redescriptions from 1940- 1980. New York Entomol. Soc. & American Mus. Nat. Hist. New York. Roewer, C.F. 1954. Katalog der Araneae von 1758 bis 1940, bzw. 1954. 2. Band. Abt. a (Lycosae- formia, Dionycha [excl, Salticiformia]). Inst. Royal Sci. Nat. Belgique. Bruxelles. Schiapelli, R.D. & B.S. Gerschman, 1970. Coesi- deraciones sobre el genero Ancylometes Bertkau 1880 (Araneae: Pisauridae). Acta Zoo. Lilloana, 27:155-175. Sierwald, P. 1988. Notes on the behavior of Tha- lassius spinosissimus (Arachnida: Araneae: Pi- sauridae). Psyche, 95(3/4):243-252. Williams, D.S. 1979. The feeding behavior of New Zealand Dolomedes species (Araneae: Pisauri- dae). New Zealand J. Zool., 6:95-105. Satie Airame: Committee on Evolutionary Biology, University of Chicago, 1101 E. 57th Street, Chicago, Illinois 60637, and the Field Museum, Zoology, 1400 S. Lake Shore Drive, Chicago, Illinois 60605 USA Petra Sierwald: Division of Insects, De- partment of Zoology, Field Museum of Nat- ural History, 1400 South Lakeshore Drive, Chicago, Illinois 60605 USA Manuscript received 24 February 1999, revised 10 December 1999. 2000. The Journal of Arachnology 28:254-255 BOOK REVIEW Lycosids in China. C.M. Yin, X.J. Peng, L.R Xie, Y.H. Bao & J.F. Wang. 317 pp. Hunan Normal University Press, Changsha 1997. ISBN 7-81031- 599-4/Q.017. The arachnological research activities in China have increased considerably during the last decades, promoted by and associated with increasing awareness of the role of spiders in agricultural issues (Song 1996). The literature on Chinese spiders has grown substantially mainly due to the studies performed by vari- ous research groups in the country. One of them is active at the Hunan Normal University in Changsha under the leadership of Prof. Changmin Yin, and a number of comprehen- sive books on spiders have originated from this group {Spiders in China, One Hundred New and Newly Recorded Species of the Fam- ilies Araneidae and Agelenidae including pa- pers by C.M. Yin et al. and J.F. Wang et al. (1990), Salticids in China by X.J. Peng, L.P Xie, X.Q. Xiao & C.M. Yin (1993), Fauna Sinica: Arachnida Araneae: Araneidae by C.M. Yin, J.F Wang, M.S. Zhu, L.P. Xie, X.J. Peng & YH. Bao (1997)]. A recent contribution from this group is Ly- cosids in China. As stated in the foreword, the book does not encompass all wolf spider spe- cies currently known to occur in China but is based on the material available in the collec- tion of the Department of Biology at Hunan Normal University. Most material therefore comes from more southern parts of this vast country. Regrettably, for ’outsiders’, the book is addressed mainly to Chinese reading arach- nologists. No English summary is given, but there are bilingual (Chinese/English) figure legends. Descriptions of 135 species distributed among 13 ’traditional’ genera are given, in- cluding lists of synonyms, comments on affin- ities with other species, habitat (for some of the species) and distribution, particularly within China. It is to be noted that the genus Ocyale Audouin 1826 is now represented in China by the recently (1997) described O. qiongzhongensis Yin & Peng which is includ- ed in the book. Keys to subfamilies, genera and species are supplied. [The subfamily Hip- pasinae is maintained, encompassing a mix- ture of genera {Ocyale, Pirata Sundevall 1833, Venonia Thorell 1894, Hippasa Simon 1885) currently allocated to other subfamilies, in despite of Hippasinae presently being placed as a junior synonym of Lycosinae {cf Dondale 1986; Zyuzin 1993)]. Illustrations are provided for all species, at least of the habitus (not very informative) and the copulatory or- gans. For some species the copulatory organ of only one sex is shown despite both sexes are known (the other sex not present in the collection). For a number of species additional illustrations are given, showing, e.g., the in- flated bulbus. Details of the macerated female receptacular complex are given for several species; information which hardly has been given as a routine in comparable monographic treatments. For many species more details of the male palp are still wanting, i.e., the con- figuration of the terminal part of the bulbus and the detailed shape of the embolus. The number of species within each genus as treated in the book is: Evippa Simon 1882 (2), Xerolycosa Dahl 1908 (1), Hippasa (4), Ocyale (1), Pirata (12), Venonia (1), Alope- cosa Simon 1885 (13), Arctosa C.L. Koch 1847 (17), Hogna Simon 1885 (2), Lycosa La- treille 1804 (16), Trochosa C.L. Koch 1847 (8), Pardosa C.L. Koch 1847 (56), Wadicosa Zyuzin 1985 (2). Several species are described as “sp. nov.” though the names were already introduced in original descriptions (in Roman letters) by various author groups elsewhere (in issues of 254 BOOK REVIEW 255 either Acta Arachnologica Sinica or Korean Arachnology from 1997, antedating the pub- lication date, 1 December 1997, of the present book). Only four of the species treated seem to have been originally described in this book, VIZ. '"Alopecosa disca Tang et al., sp. nov.,” ''Alopecosa wenxianensis Tang et al.,” ‘Arc- tosa Uujiapingensis sp. nov.,” and '"Pardosa alboannulata sp. nov.” (names cited as they appear in the book). From the illustrations it is apparent that there are a number of misidentifications. The following serve as examples only and is not meant to be a complete listing (for which the reviewers have insufficient knowledge): The figures referring to certain species, e.g., Xe~ rolycosa nemoralis (Westring 1861), were ap- parently drawn from material belonging to other species. The epigynum in ventral view attributed to Pardosa schenkeli Lessert 1904 reminds one more of P. hanrasanensis Jo & Paik 1984 (from Korea); the latter on the other hand is listed as a synonym of what is stated to be Pardosa bifasciata (C.L. Koch 1834). Pardosa anchoroides Yu & Song 1988 was recently synonymized with P. adustella Roewer 1951 (by Logunov & Marusik 1995). The illustrations meant to show the Pardosa atrata (Thorell 1873) male were clearly drawn from another species, and the drawings attri- buted to the Pardosa lapponica (Thorell 1872) female were made from another, pos- sibly undescribed, species. The illustrations ascribed to Pardosa uncifera Schenkel 1963 do not match the type material examined by us. Without details of the terminal apophysis of the bulbus, it is impossible to judge wheth- er the authors really had Pardosa monticola (Clerck 1757) at hand. The illustrations of Pardosa multivaga Simon 1880 make us sus- pect that this species may not even belong in Pardosa. There are scattered misspellings, e.g., ''kro- tochvillV' instead of kratochvili throughout (in Alopecosa), '"dividV" instead of davidi (syno- nym of Alopecosa licenti), etc., author of Par- dosa astrigera is L. Koch, not his father C.L. Koch. Several of the references given in the foreword and the introdutory chapter do not appear in the list of literature cited at the end. Despite the linguistic problems and the lim- ited coverage of lycosid species from northern China — the title is accordingly ''Lycosids in China'" not “The Lycosids of China” — this book is a valuable tool for researchers outside China interested in taxonomic problems of East Asian wolf spiders. It will, above all, serve as a useful iconotheca and a source for taxonomic inspiration for those of us who do not master Chinese. Both reviewers are grateful to Prof. Yin for copies of the book and to Mrs. Fang Fang, ichthyologist at the Swedish Museum of Nat- ural History, for translation of certain passag- es in the book. LITERATURE CITED Dondale, C.D. 1986. The subfamilies of wolf spi- ders (Araneae: Lycosidae). Actas X Congr. Int. AracnoL, Jaca, Espana, 1:327-332. Logunov, D.V. & Marusik, YM. 1995. Spiders of the family Lycosidae (Aranei) from the Sokhon- do Reserve (Chita Area, East Siberia). Beitr. Ar- aneoL, 4:109-122. Song, D.X. 1996. Aspects of spider research in China. Revue Suisse Zool., vol. hors sme:611- 615. Zyuzin, A. A. 1993. Studies on the wolf spiders (Araneae: Lycosidae). I. A new genus and spe- cies from Kazakhstan, with comments on the Ly- cosinae. Mem. Queensland Mus., 33:693-700. Torbjdrn Kronestedt: Department of En- tomology, Swedish Museum of Natural History, Box 50007, SE-104 05 Stockholm, Sweden Yuri M. Marusik: Institute for Biological Problems of the North, Karl Marx prospect 24, Magadan 685010, Russia Manuscript received 5 March 1999, revised 2 No- vember 1999. 2000. The Journal of Arachnology 28:256 ARACHNOLOGICAL RESEARCH FUND The AAS Fund for Arachnological Re- search (AAS Fund) is funded and admin- istered by the American Arachnological So- ciety. The purpose of the fund is to provide research support for work relating to any aspect of the behavior, ecology, physiology, evolution, and systematics of any of the arachnid groups. Awards may be used for field work, museum research (including travel), expendable supplies, identification of specimens, and/or preparation of figures and drawings for publication. Monies from the fund are not designed to augment or re- place salary. Individual awards will not exceed $1000.00, and, although open to all stu- dents and faculty with less than $500.00 per year research budget, preference will be given to students. A total of $6000.00 is available for awarding during each funding year. Available monies could be expended for three large proposals, a greater number of partially funded proposals, and/or a num- ber of smaller, less expensive proposals. The final funding pattern is at the discretion of the review committee. Applications for support should be re- ceived by the chair of the review committee no later than January 15. To be considered for an award from the AAS Fund, please submit four copies of a proposal of no more than five pages (including references) de- tailing your research project. Proposals should have three main parts: 1) an INTRODUCTION where background information is presented relative to the pro- posed work. The introduction should in- clude a section which places the proposed work in context with currently known rel- evant information, 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 revi- sions, the type of synthesis that will be achieved and its significance; 2) a METH- ODS section where the methods, materials, experimental design, and statistical or tax- onomic analysis(ses) to be used are clearly and concisely presented, and 3) a BUDGET explaining (in detail) how monies awarded will be spent in the proposed research. Proposals should be submitted to: Dr. Deborah Smith, AAS Fund Chair Department of Entomology Haworth Hall University of Kansas Lawrence, KS 66045 USA Proposals must be submitted in English. The four copies of the proposal must be in the hands of the Eund chair by the appro- priate deadlines to be considered. Electron- ic submission and/or FAX submission of proposals with hard copies to follow is ac- ceptable only if the other copies arrive be- fore the stated deadlines. If these submis- sion rules are difficult or prohibitive be- cause of cost, erratic postal services, or re- mote location (remote field stations or sites), other methods of submission may be acceptable. For other submission possibili- ties, please contact the chair of the Fund at the above address, or electronically at dsmith@kuhub.cc.ukans.edu. Alternative submissions will be accepted only if the chair has been previously contacted, and all deadlines will still apply. 256 INSTRUCTIONS TO AUTHORS (revised August 2000) Manuscripts are accepted in English only. Authors whose primary language is not English may consult the editors for assistance in obtaining help with manuscript preparation. All manuscripts should be prepared in general accordance with the current edition of the Council of Biological Editors Style Manual unless instructed otherwise below. Authors are advised to con- sult a recent issue of the Journal of Arachnology for additional points of style. Manuscripts longer than three printed journal pages should be prepared as Feature Articles, shorter papers as Short Commun- ications. Send four identical copies of the typed mate- rial together with copies of illustrations to the Managing Editor of the Journal of Arachnology: Petra Sierwald, Managing Editor; Division of In- sects, Dept, of Zoology, The Field Museum of Na- tural History, 1400 South Lakeshore Drive, Chicago, IL 60605 USA [Telephone: (312)-665-7744; FAX: (312)-665-7754; E-mail psierwald@fmnh.org]. The Managing Editor will forward your manuscript to one of the Subject Editors for the review process. You will receive correspondence acknowledging the receipt of your manuscript from the responsible subject editor or the managing editor, with the manuscript number of your manuscript. Please use this number in all correspondence regarding your manuscript. Corres- pondence relating to manuscripts should be directed to the Managing Editor or the appropriate Subject Editor. After the manuscript has been accepted, the author will be asked to submit the manuscript on a computer disc in a widely-used word processing program. Indicate clearly on the computer disc the word processing pro- gram and the type of computer (Mac or PC). FEATURE ARTICLES Title page. — The title page will include the complete name, address, and telephone number of the author with whom proofs and correspondence should be exchanged, a FAX number and electronic mail address if available, the title in capital letters, 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. A second abstract, in a language pertinent to the nationality of the author(s) or geographic region(s) emphasized, may be included. Keywords. — Give 3-5 appropriate ke5rwords fol- lowing the abstract. Text. — Double-space text, tables, legends, etc. throughout. Three categories of headings are used. The first category (METHODS, RESULTS, etc.) is typed in capitals and on a separate line. The second category of heading, in bold type, begins a paragraph with an indent and is separated from the text by a period and a dash. The third category of heading may or may not begin a paragraph but is italicized and separated fi-om 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 (e.g.,-0.123). 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). Include a letter of permission from any person who is cited as providing impublished data in the form of a personal communication. Literature cited section. — Use the following style, and include the full unabbreviated journal title. Lombardi, S.J. & D.L. Kaplan. 1990. The amino acid composition of major ampullate gland silk (drag- line) of Nephila clavipes (Araneae, Tetragnathidae). Journal of Arachnology 18:297-306. Krafft, B. 1982. The significance and complexity of communication in spiders. Pp. 15-66, In Spider Communications: Mechanisms and Ecological Significance. (P.N. Witt & J.S. Rovner, eds.). Princeton University Press, Princeton, New Jersey. Footnotes. — Footnotes are permitted only on the first journal page to indicate current address or other information concerning the author. These are placed together on a separate manuscript page. Tables and fig- ures may not have footnotes. Running head. — The author’s sumame(s) and an abbreviated title should be typed all in capital letters and must not exceed 60 characters and spaces. The running head should be placed near the top of the title page. Taxonomic articles. — Consult a recent taxonomic article in the Journal of Arachnology for style, or contact the Subject Editor for Systematics. Papers containing the original taxonomic description of the focal arachnid taxon should be listed in the Literature Cited section. Tables. — Each table, with the legend above, should be placed on a separate manuscript page. Only horizon- tal lines (usually three) should be included. Tables may not have footnotes; instead, include all information in the legend. Make notations in the text margins to indi- cate the preferred location of tables in the printed text. Illustrations. — Address all questions concerning illustrations to the Editor of the Journal of Arach- nology: James W. Berry, Editor; Dept, of Biological Sciences; Butler University, Indianapolis, Indiana 46208 USA. [Telephone (317)-940-9344; FAX (317)- 940-9519; E-mail: jwberry@butler.edu]. All art work must be camera-ready (mounted and labeled) for reproduction. Figures should be arranged so that they fit (vertically and horizontally) the printed journal page, either one column or two columns, with a mini- mum of wasted space. When reductions are to be made by the printer, pay particular attention to width of lines and size of lettering in line drawings. Multiple photos assembled on a single plate should be mounted with only a minimum of space separating them. In the case of multiple illustrations mounted together, each illus- tration must be munbered sequentially rather than given an alphabetic sequence. Written on the back should be the name(s) of author(s) and an indication of top edge. Indicate whether the illustration should be one column or two columns in width. The overall dimensions should be no more than 1 1 inches (28 cm) X 14 inches (36 cm). Larger drawings present greater difficulty in shipping and greater risks of damage for which the JoA assumes no responsibility. In manu- scripts for review, photocopies are acceptable, and should be reduced to the exact measurements that the author wants to appear in the final publication. Make notations in the text margins to indicate the preferred position of illustrations in the printed text. Color plates can be printed, but the author must assume the full cost, currently about $600 per color plate. Legends for illustrations should be placed together on the same page(s) and separate from the illustrations. Each plate must have only one legend, as indicated below: Figures \-4. — A-us x-us, male from Timbuktu: 1. Left leg; 2. Right chelicera; 3. Dorsal aspect of geni- talia; 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 holot3q)e 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, footnotes, tables with legends, figure legends, figures. Page charges and reprints. — There are no page charges, but the author will be charged for changes made in the proof pages. Reprints are available only from the Allen Press and should be ordered when the author receives the proof pages. Allen Press will not accept reprint orders after the paper is published. SHORT COMMUNICATIONS The above instructions pertaining to Feature Articles apply also to Short Communications, which should be prepared in the same manner as regular Feature Articles. Short Commimications are usually limited in length to three journal pages, including tables and figures. They will be printed in a smaller (10 point) typeface. CONTENTS The Journal of Arachnology Volume 28 Feature Articles Number 2 Description of the male of Sosippus placidus, with notes on the subfamily Sosippinae (Araneae, Lycosidae) by Petra Sierwald 133 Iracema cabocla new genus and species of a theraphosid spider from Amazonic Brazil (Araneae, Therephosinae) by Fernando Perez-Miles 141 Respiratory system morphology and the phytogeny of haplogyne spiders (Araneae, Araneomorphae) by Martin J. Ramirez .... 149 Effects of climate and prey availability on foraging in a social spider, Stegodyphus mimosarum (Araneae, Eresidae) by T.E. Crouch & Y. Lubin 158 The impact of habitat features on web features and prey capture of Argiope aurantia (Araneae, Araneidae) by C. Neal McReynolds 169 Evaluation of formulae to estimate the capture area and mesh height of orb webs (Araneoidea, Araneae) by Marie Elisabeth Herberstein & I-Min Tso 180 Population structure, seasonality, and habitat use by the green lynx spider Peucetia viridans (Oxyopidae) inhabiting Cnidoscolus aconitifolius (Euphorbiaceae) by Angelica M. Arango, Victor Rico-Gray & Victor Parra-Tabla 185 Food consumption rates and competition in a communally feeding social spider, Stegodyphus dumicola (Eresidae) by Nava Amir, Mary E.A. Whitehouse & Yael Lubin 195 Predator avoidance on the water surface? Kinematics and efficacy of vertical jump- ing by Dolomedes (Araneae, Pisauridae) by Robert B. Suter & Jessica Gruenwald 201 Predatory interactions between mud-dauber wasps (Hymenoptera, Sphecidae) and Argiope (Araneae, Araneidae) in captivity by Todd A. Blackledge & Kurt M. Pickett 211 Spiders in rocky habitats in central Bohemia by Vlastimil Ruzicka 217 Research Notes Anyphaenidae in Miocene Dominican Republic amber (Arachnida, Araneae) by David Penney 223 Eremopus acuitlapanensis, a new species (Solifugae, Eremobatidae, Eremobatinae) from Guerrero, Mexico by Ignacio M. Vazquez & Rafael Gavino-Roj as 227 External morphology and ultrastructure of the prehensile region of the legs of Leiobunum nigripes (Arachnida, Opiliones) by Cary Guffey, Victor R. Townsend, Jr. «& Bruce E. Felgenhauer 231 A comparative study of sexual behavior in two synmorphic species of the genus Lycosa (Araneae, Lycosidae) and their hybrid progeny by Fernando G. Costa, Carmen Viera & Gabriel Francescoli 237 Estimating foraging intake: A comment on Tso and Severinghaus (1998) by G.S. Oxford 241 A test of pollen feeding by a linyphiid spider by James E. Carrel, Heather K. Burgess & Dennis M. Shoemaker 243 Comparison of the fertility between Loxosceles intermedia and Loxosceles laeta spi- ders (Araneae, Sicariidae) by Rute Maria Gon9alves de Andrade, Wilson R. Louren^o & Denise Vilarinho Tambourgi 245 Group dispersal in juvenile Brachypelma vagans (Araneae, Theraphosidae) by Steven B. Reichling 248 Hunting and feeding behavior of one Heteropoda species in lowland rainforest on Borneo (Araneae, Sparassidae) by Satie Airame & Petra Sierwald 251 Other Lycosids In China (by C.M. Yin, X.J. Peng, L.P. Xie, Y.H. Bao & J.F. Wang) reviewed by Torbjorn Kronestedt & Yuri M. Marusik 254 Arachnological Research Fund 256 VX V liSB /OT The Journal of ARACHNOLOGY OFFICIAL ORGAN OF THE AMERICAN ARACHNOLOGICAL SOCIETY THE JOURNAL OF ARACHNOLOGY EDITOR-IN-CHIEF: James W. Berry, Butler University MANAGING EDITOR: Petra Sierwald, Field Museum SUBJECT EDITORS: Ecology — Matthew Greenstone, USDA; Systematics — Mark Harvey, Western Australian Museum; Behavior and Physiology — Robert Suter, Vassar College 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. Argentine de Ciencias Naturales; 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.; N. I. Platnick, American Mus. Natural Hist.; S. E. Riechert, Univ. Tennessee; A. L. Rypstra, Miami Univ., Ohio; M. H. Robinson, US. 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: Frederick A. Coyle (1999-2001), Department of Biology, Western Carolina University, Cullowhee, North Carolina 28723 USA. PRESIDENT-ELECT: Brent D. Opell ( 1 999-200 1 ), Department of Biology, Virginia Pol34:echnic Institute and State University, Blacksburg, Virginia 24061 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, Department of Biology, University of Missis- sippi, University, Mississippi 38677 USA. SECRETARY: Alan Cady, Dept, of Zoology, Miami Univ., Middletown, Ohio 45042 USA. ARCHIVIST: Lenny Vincent, Fullerton College, Fullerton, California 92634. DIRECTORS: Bruce Cutler (2000-2002), Paula Cushing (1999-2001), Ann Rypstra (1999-2001). HONORARY MEMBERS: C. D. Dondale, H. W. Levi, A. F. Millidge, W. Whit- comb. Cover photo: Ballooning pisaurid (Araneae, Pisauridae) from Upper Souris National Wildlife Refiige, North Dakota. {Photo by Bryan Reynolds) Publication date: 4 December 2000 0 This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 2000. The Journal of Arachnology 28:257-260 A NEW SPECIES OF THE GENUS KIMULA (OPILIONES, MINUIDAE) FROM THE DOMINICAN REPUBLIC Abel Perez Gonzalez and Luis F. de Armas: Depaitamento de Ivertebrados, Institute de Ecologia y Sistematica, A.R 8029, C.P. 10800, Habana 8, Cuba ABSTRACT. Kimula cokendolpheri new species is described from the Central Range of the Dominican Republic, West Indies. It is the first non-fossil species of this genus recorded from Hispaniola, Keywords: Opilionids, Kimula, West Indies, Dominican Republic The Antillean genus Kimula Goodnight & Goodnight 1942 includes species that are found on the islands of Puerto Rico {K. elon- gata Goodnight & Goodnight 1942), Cuba (K. tuberculata Goodnight & Goodnight 1943, K. levii Sihavy 1969, K. banksi Sihavy 1969, K. goodnightorum Sihavy 1969, K. turquiensis Sihavy 1969; and K. botosaneanui Avram 1973) (Cokendopher & Camilo-Rivera 1989). On St. Johns, U.S. Virgin Islands, it is repre- sented by an undescribed species, and in the Dominican Republic a female of Kimula? was found in amber that has a confirmed date of 25-40 MYA (Cokendolpher & Poinar 1992). Kimula cokendolpheri new species is the first known living species of the genus Kimula from Hispanola. However, if speciation of this group in Hispanola is similar to that on Cuba (where there are other undescribed species, pers. obs.), it is likely that additional new spe- cies of Kimula will be found in Hispaniola. METHODS We studied material from the collections of the invertebrates of Hispanola that is depos- ited in the Instituto de Ecologia y Sistematica (lES) of the Ministerio de Ciencia, Tecnologia y Medio Ambiente, Havana, Cuba. The no- menclature of the dorsum follows the usage of Maury (1991). We denote the body divi- sions as: prosoma, mesotergum (areas I, II, III, and IV), lateral margin, and posterior margin (denoted as area V by other authors). Dorsal scute is the sum of thee mesotergum and its posterior margin. Measurements are given in mm and were made with a dissecting micro- scope equipped with an ocular micrometer. Kimula cokendolpheri new species (Figs. 1-9) Type specimens. — Male holotype and two male and four female paratypes collected in Casabito, Constanza, provincia La Vega, Do- minican Republic beneath stones on 27 Sep- tember 1987 by A. Abud and L.F. de Armas. Deposited in the lES. Distribution. — Known only from the type locality. Etymology. — The specific epithet is a pa- tronym in honor of James C. Cokendolpher, who has studied the opilionid fauna of the An- tilles. Diagnosis. — Total length 5.00 with area I lacking a median line and armed with a stout median spine similar to that of area II. Femurs of the pedipalps armed dorsally with a series of tubercles terminating in setae. Trochanter IV armed with a blunt ventroproximal tuber- cle. Tarsal formula: 4, 9-13, 5, 6. Distinctive male genitalia as shown in Figs. 7-9. Kimula cokendolpheri new species has two characters that are recorded for the first time in this ge- nus: the median spine of areas I and II and the presence of tubercles on the dorsal surface of the femoral palp. These characters permit clear separation of this species from others in this genus. Description. — Male: Body brown. The chelicerae, pedipalps, legs, and prosoma bear yellowish patches and in some aspects appear reticulated. Trochanters of legs 1, II, and III with dorsal yellowish markings. Ocular tuber- cle prominent, granular, armored with an erect, anteriorly inclined spine between the eyes (Fig. 1). Areas of the mesotergum dis- tinct, covered irregularly by granules and tu- 257 258 THE JOURNAL OF ARACHNOLOGY Figures 1-6. — External morphology of Kimula cokendolpheri new species. 1. Lateral view; 2. Dorsal view; 3. Lateral view of opisthosoma; 4. Retrolateral view of right pedipalp; 5. Prolateral view of right pedipalp; 6. Medial view of right chelicera. Scale bars == 1 mm. bercles, terminated in a small apical setae, in- ner areas with similar projections in the form of large spines: the first situated on the central portion of the posterior margin of area I and the second, smaller, occupying the same po- sition in area IT Area I without a transverse median line. Areas I and II separated by an incomplete furrow. Dorsal scute with sinuous lateral margins and a straight posterior mar- gin; these margins present a longitudinal row PEREZ & DE ARMAS— NEW SPECffiS OF KIMULA 259 Figures 7-9. — Penis of Kimula cokendolpheri new species. 7. Ventrolateral view (scale = 1 mm); 8. Lateral view of extreme distal region (scale = 0.1 mm); 9. Dorsal view of extreme distal region (scale = 0.1 mm). of tubercles that are more evident and dentic- ulate at the level of furrow II; this region achieves its maximum width at the level of furrow 11. Tergites free with a longitudinal row of tubercles terminating in setae. Free ter- gite III has a median spine (Fig. 2); anal oper- culum with numerous short tubercles (Fig. 3). Retrolateral surface of coxa III with a distal tubercle, coxa IV very well developed and strongly tuberculated on its prolateral surface. Ventrally, all coxae granulated. Stemites free with a longitudinal row of granules; free ster- nite IV with a median spiny apophysis; free stemite V with two rows of longitudinal tu- bercles one at the anterior margin and another on the posterior, separated by a furrow (Fig. 3). Pedipalp (Figs. 4, 5): the coxa has a ventral tubercle with apical setae, dorsally there are two proximal tubercles, one external and one internal; trochanter with five tubercles that possess apical setae, one dorsal and four ven- tral; femur with five or six small dorsal tu- bercles that bear fine apical setae, ectolaterally with a tuberculate proximal spine and with four or five tubercles and, on the anterior half of the internal surface, with tuberculate spines having large bases; patella unarmed; tibia ven- trally with four tuberculate spines on its ex- 260 THE JOURNAL OF ARACHNOLOGY ternal border and two tuberculate spines on its internal border; tarsus with four ventral tuber- culate spines on its external border and three ventral tuberculate spines on its internal bor- der. Chelicerae (Fig. 6): basal article with a strong distal elevation on whose outer poste- rior border there is a small tubercle; distal ar- ticles with small tubercles terminated in setae. Legs lack tarsal processes and scopulae; legs I and II with all their articles, except the tarsae (which are unarmed), covering by tubercules that reach their greatest development in the ventral region of the femur. Leg IV is the most well-developed and is strongly armored with tubercles and spines, except for the tarsus, the trochanter has a characteristic blunt ventral tu- bercle (Fig. 1) and the femur is notably en- larged with a ventral row of strong spines. The patella and the tibia are strongly tuberculate and at their apices have very enlarged and glo- bose ventral tubercles. Tarsal formula: 4(2), 9-13(2), 5(3), 6(3). Male genitalia as shown in Figs. 7-9. Measurements of the male ho- lotype: total length = 5.6; prosoma + scutum = 4.7; maximum width 3.4; leg \ — 1 1 (tro- chanter 0.5, femur 1.7, patella 0.8, tibia 1.3, metatarsus 2.0, tarsus 1.4); leg II = 11.3 (tro- chanter 0.7, femur 2.4, patella 1.1, tibia 1.8, metatarsus 2.5, tarsus 2.8); leg III = 8.3 (tro- chanter 0.6, femur 1.7, patella 0.8, tibia 1.4, metatarsus 2.3, tarsus 1.5); leg IV = 12.6 (tro- chanter I.O, femur 2.8, patella 1.7, tibia 2.5, metatarsus 2.9, tarsus 1.7). Female: Similar to the male in appearance, but smaller. The spines are reduced and femur IV differs markedly and is not as enlarged. Trochanter IV has a ventrodistal spine rather than the blunt tubercle characteristic of the male. The free stemite lacks the spiny median apophysis. Measurements of the one female paratype: total length = 4.9; prosoma + scu- tum = 3.9; maximum width 2.8; leg I = 7.7 (trochanter 0.6, femur 1.5, patella 0.8, tibia 1.2, metatarsus 1.9, tarsus 1.7); leg II = 10.7 (trochanter 0.8, femur 2.2, patella 1.1, tibia 1.6, metatarsus 2.3, tarsus 2.7); leg III = 7.8 (trochanter 0.7, femur 1.5, patella 0.8, tibia 1.3, metatarsus 2.0, tarsus 1.5); leg IV = 11.2 (trochanter 1.1, femur 2.4, patella 1.2, tibia 2.1, metatarsus 2.8, tarsus 1.6). Natural history. — The specimens studied were collected at an elevation of approxi- mately 1000 m above sea level, beneath stones and in forest litter in a very humid for- est at the margins of a stream that abounded in tree ferns. ACKNOWLEDGMENTS We are grateful to J.C. Cokendolpher for sending literature and, particularly, to the late Emilio E. Maury, Museo Argentino de Cien- cias Naturales “Bernardino Rivadavia,” for the training provided to the first author and for the valuable literature he provided. Nor- man Platnick kindly lent the holotype of Ki- mula elongata. I especially want to thank Brent D. Opell for the translation to English and for his help in the publication of the pre- sent article. LITERATURE CITED Avram, S. 1973. Recherches sur les opilionides de Cuba. IL Phalangodidae: Kimula (Metakimula) botosaneanue n. sg., n. sp. In Resultats de Ex- peditions Biospeologiques Cubano-roumaines a Cuba. Bucuresti 1:253-258. Cokendolpher, J.C. & G.R. Camilo-Rivera. 1989. Annotated bibliography to the harvestmen of the West Indies (Arachnida: Opiliones). Occasional Papers of the Florida State Collection of Arthro- pods 5:1-20. Cokendolpher, J.C. & G.O. Poinar, Jr. 1992. Ter- tiary harvestmen from Dominican Republic am- ber (Arachnida: Opiliones: Phalangodidae). Bul- letin of the British Arachnological Society 9(2): 53-56. Goodnight, C.J. & M.L. Goodnight. 1942. Phal- angids from Central America and the West In- dies. American Museum Novitates 1184:1-23. Goodnight, C.J. & M.L. Goodnight. 1943. Three new phalangids from tropical America. Ameri- can Museum Novitates 1228:1-4. Maury, E.A. 1991. Gonyleptidae (Opiliones) del bosque subanartico Chileno-argentino I. El ge- nero Acanthoprocta Loman, 1899. Boletin de la Sociedad de Biologia de Concepcion (Chile) 62: 107-117. Silhavy, V. 1969. The genus Kimula Goodnight and Goodnight from Cuba (Arachnoidea, Opi- lioniodea). Acta Entomologica Bohemoslovaca 66(6):399-409. Manuscript received 15 May 1998, revised 20 May 2000. 2000. The Journal of Arachnology 28:261-292 SYSTEMATICS OF THE GENUS DYSDERA (ARANEAE, DYSDERIDAE) IN THE EASTERN CANARY ISLANDS Miquel A. Arnedo^; Departament de Biologia Animal, Universitat de Barcelona. Av. Diagonal 645, E08028 Barcelona, Spain Pedro Oromi: Departamento de Biologia Animal, Universidad de La Laguna, Tenerife, Spain Carles Ribera: Departament de Biologia Animal, Universitat de Barcelona. Av. Diagonal 645, E08028 Barcelona, Spain ABSTRACT. The circum-Mediterranean spider genus Dysdera has undergone an outsanding species radiation in the volcanic archipelago of the Canary Islands. The present study deals with the endemic species that inhabit the older and ecologically distinct islands of Fuerteventura, Lanzarote and their nearby islets. A new species, Dysdera sanborondon, is described. The male of D. spinidorsum Wunderlich 1991, is described for the first time. Five species are redescribed: D. alegranzaensis Wunderlich 1991; D. Ian- cerotensis Simon 1907; D. longa Wunderlich 1991; D. nesiotes Simon 1907, and D. spinidorsum Wun- derlich 1991. The species D. liostethus Simon 1907 is proposed to be a senior synonym of D. clavisetae Wunderlich 1991 and its presence in the eastern islands is considered to be doubtful. A neotype is des- ignated for D. nesiotes. The distribution of D. alegranzaensis is extended to Lanzarote and the other northern islets. Dysdera nesiotes is reported for the first time in the eastern Canaries. Morphological affinities and distribution patterns are discussed. The remarkably lower number of endemic species har- bored by the eastern islands, when compared with other Canarian islands similar in size but younger in age, is proposed to be the result of a major extinction event in the eastern Canaries due to climatic change. Keywords: Spider taxonomy, oceanic islands, colonization, extinction Studies in oceanic archipelagos have be- come crucial in the rise and development of evolutionary thinking and the present Darwin- ian paradigm. To date, the role played by the different islands has been highly biased in fa- vor of the Pacific Archipelagos (the Hawaiian Islands and the Galapagos). Nevertheless, in the last few years a growing number of studies on the systematic s of such diverse groups as lizards (Thorpe et al. 1994, 1995; Gonzalez et al. 1996; Rando et al, 1997), beetles (Juan et al. 1995, 1996a, 1996b, 1998) or plants (Bohle et al. 1996; Francisco-Ortega et al. 1996; Kim et al. 1996; Mes & T’Hart 1996) have re- vealed an additional excellent model for the study of biodiversity in the Atlantic region: the Macaronesian archipelagos, and in partic- ular the Canary Islands. * Current Address: Division of Insect Biology, ESPM, University of Califomia-Berkeley, 201 Well- man Hall, Berkeley, California 94720-3112, USA The genus Dysdera Latreille 1804 compris- es more than 200 species of nocturnal wan- dering spiders spread over the circum-Medi- terranean region. About a quarter of these species have been described from the Maca- ronesian archipelagos (Fig. 1), representing the most species-rich spider genus reported in them. Nevertheless, the Macaronesian endem- ics are far from being equally distributed. The Canary Islands harbor 43 of these endemics, while five endemics have been documented from Madeira (Denis 1962; Wunderlich 1994). The Azores, Cabo Verde and Selvagens Islands each have a single species (Amedo un- publ. data; Berland 1936; Kulczynski 1899). The unusually large number of species in the Canaries suggests many evolutionary and eco- logical questions. A research program is cur- rently underway to resolve some of the prob- lems posed by the genus in the archipelago (Ribera & Amedo 1994; Amedo & Ribera 261 262 THE JOURNAL OF ARACHNOLOGY WESTERN CENTRAL LA PALMA ^ TENERIFE LACX)MERA ^ ^ o 0 LANZAROTE FUERTEVENTURA EASTERN GRAN CANARIA N A ALEGRANZA ^ JTANA CLARA ^ DEL ESTE LAGRACIOSA ^ FAMARA LANZAROTE FUERTEVENTURA CENTRAL EDIFICE jandIa Figures 1-3. — 1, Location of the Canaries and the remaining Macaronesian archipelagos; 2, The Canary Islands; 3, Fuerteventura, Lanzarote and the islets. 1996; Arnedo et al. 1996; Arnedo & Ribera 1997). Although the geological processes that cre- ated the Canary Islands are still a matter of debate (Anguita & Hernan 1975; Carracedo et al. 1998a 1998b), the first island most likely arose about 25 My ago. The seven main is- lands lie 100 km from the northwestern coast of Africa in a roughly straight line (Fig. 2). A geographical gradation in their geological age exists, the islands being older in the east and becoming younger to the west. The estimated geological age for each island is: Fuerteven- tura 20-22 My, Lanzarote 15-19 My, Gran Canaria 14-16 My, Tenerife 11.6-14 My, La Gomera 10-12 My, La Palma 1.6-2 My and El Hierro 0.8-1 My (Cantagrel et al. 1984; Mitchell-Thome 1985; Ancochea et al. 1990; Coello et al. 1992; Fuster et al. 1993; Anco- chea et al. 1994, 1996). Unlike some well- known oceanic archipelagos such as the Ha- waiian Islands, the growth of the islands extended over long periods of time (Coello at al. 1992), and volcanic activity is cyclic and is not restricted to the younger islands. These features together with the absence of a sub- duction region which would promote subsi- dence of the older islands, as is the case in several Pacific archipelagos (Paulay 1994), al- low the islands to reach later stages of eco- logical succession. The eastern Canaries are the emergent regions of a volcanic ridge, run- ning parallel to the African coast in a NNE- SSW direction (Coello et al 1992). It com- prises two main islands, Fuerteventura at the SSW and Lanzarote at the NNE end, and sev- eral islets: Lobos, between the two big islands, and La Graciosa, Roque del Este, Roque del Oeste, Montana Clara and Alegranza, to the north of Lanzarote (Fig. 3). The maximum ocean depth between these islands is barely 40 m and thus it is very likely that they were connected during glaciation periods. The is- lands are the result of five volcanic complexes that arose from the ocean in a temporal suc- cession: the peninsula of Jandia 20.7 Mya, the Central edifice 22.5 Mya, the northern edifice 17.0 Mya in Fuerteventura (Ancochea et al. 1996) and Aj aches 15.5 Mya and Famara 10.2 Mya in Lanzarote (Coello et al. 1992). The eastern Canaries have undergone several sub- aerial cycles of volcanic activity. A major gap in activity between the Miocene and the Pli- ocene periods brought about an extensive ero- sion of the edifices. Postmiocene activity was limited to central and northern Fuerteventura and Lanzarote (Coello et al. 1992). In these regions, recent volcanic activity, and historical eruptions in the case of Lanzarote, have been documented. Apart from the lack of recent volcanic activity, the peninsula of Jandia, in southern Fuerteventura, is characterized by its ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 263 “ecological” isolation. It is separated from the rest of the island by an isthmus which is ex- tensively covered with eolic sands. The geological structure of the sea floor be- tween the eastern Canaries and Africa is ob- scured by thick sediments. Surprisingly, sub- fossil ostrich eggs have been found in the islands. These data have driven some authors to claim a continental origin for the eastern Canaries with subsequent episodes of volcanic activity (Sauer & Rothe 1972). However, geo- logical data accumulated during the last few years strongly disagree with this view, point- ing to a strictly oceanic origin of the islands. Before the present study five Dysdera spe- cies were reported to be present in the eastern Canaries (Wunderlich 1991; Amedo et al. 1996): Dysdera longa Wunderlich 1991, and D. spinidorsum Wunderlich 1991 from Fuer- teventura; D. liostethus Simon 1907 from Lanzarote; D. alegranzaensis from the islet of Alegranza and D. lancerotensis Simon 1907, reported from the two major islands. Two of these species, D. liostethus and D. spinidor- sum, were known from single specimens: a male and a female respectively. After taking into account their age and size, the number of Dysdera endemic species har- bored by the eastern Canaries is remarkably low when compared with the remaining is- lands in the archipelago. The three endemic species from Fuerteventura represent less than half the number of endemic species known from the similarly sized but younger Gran Ca- naria, and much less than the 21 endemic spe- cies from the slightly larger but younger Ten- erife. Lanzarote has the same number of endemic species as La Palma, which is similar in size but ten-fold younger, while eight spe- cies are known from La Gomera, only half its size and age. It is possible that the depauper- ate species composition in the eastern islands is the result of undersampling or, more gen- erally, of the poor taxonomic knowledge of these islands. However, if these disparities are in fact real, the ecological and evolutionary processes that underlay them need to be elu- cidated. METHODS Material was made available from scientific institutions (as well as personal collections) and collection expeditions to the islands by the authors. The following colleagues and mu- seums kindly supplied material for the present study: Dr. E. Enghoff from the Zoologisk Mu- seum of Copenhagen (ZMK), O. Escola from the Museu de Zoologia de Barcelona (MZB), Dr. P.D. Hillyard from the Natural History Museum of London (BMNH), Dr. P. Oromi from the Universidad de La Laguna (UL), Dr. G. Ortega from the Museo de Ciencias Natur- ales de Santa Cruz de Tenerife (MCNT), Dr. C. Holland from the Museum National d’Histoire Naturelle de Paris (MNHN) and Miguel Villana (MNCN). Material from the authors’ expeditions is stored in the collection of Arachnids of the University of Barcelona, Spain (UB). Character definition and terminology. — Characters were examined under a Wild Heer- brugg (12~100X magnification) dissecting mi- croscope and measurements were taking using an ocular micrometer. Female vulvae were re- moved and muscle tissues were digested using a KOH (35%) solution before observation. Male bulbi and spinnerets were removed, cleaned by means of ultrasound and exanfined using a HITACHI S-2300 Scanning Electron Microscopy at 10-15 Kv. Drawings of dorsal carapace, ventral chelicera, male palp and fe- male endogyne were made with the aid of a drawing grid. Characters exanfined for taxonomic revi- sion and their diagnostic resolution have been discussed elsewhere (Amedo et al. 1996; Ar- nedo & Ribera 1997). Leg spination was re- corded using the codification method fully de- scribed in Amedo & Ribera (1997). Stmctures of the male bulbus and female endogyne were mostly named following Deeleman-Reinhold & Deeleman (1988). However, after exami- nation of a large number of continental rep- resentatives it was realized that some of the terms included very different and probably non-homologous characters. With the aim of clarifying character terminology a full de- scription and definition of characters are pro- vided for Dysdera male and female genitalia (see also Amedo & Ribera 1997). Male bulbus: (Figs. 4-10) The genus Dys- dera has one of the most complex bulbs in the whole family Dysderidae. Schult (1980, 1983) was the first to establish the homologies be- tween the Dysdera bulb and the spider ground plan as suggested by Kraus (1978). In Dys- dera, the basal and medial haematodochae as well as the sclerites I (= subtegulum) are very 264 THE JOURNAL OF ARACHNOLOGY Figures 4-10. — Diagrams showing the male cop- ulatory bulbus characters discussed in text. 4, Fron- tal view; 5, External view; 6, Posterior view. 7-10, Different types of DD; 7-9, Frontal view; 10, DD type 9, Distal tip internal view. reduced and hardly visible. On the other hand, the sclerite II or tegulum (T) is very well- developed, representing in most cases half of the bulbus, and holds a posterior apophysis (P). The T externally covers the spermophore (= reservoir or sperm duct) (SP). The distal division (DD) of the bulb includes the mem- branous distal haematodocha (DH), which in- cludes the seminiferous duct (SD) inside, and the sclerite III (= embolic division). The DH is usually truncated at its distal tip, where the seminiferous duct opening is found. Some- times, the internal distal tip of the DH projects as a finger-like structure. Sclerite III or em- bolic division, which is located on the anterior side of the bulb, is divided into two branches, the internal branch or internal sclerite (IS) and the external one (ES). Schult (1980, 1983) proposed that the posterior apophysis (P) and the external sclerite (ES) were homologues to the median apophysis and the conductor, re- spectively, of the Araneoclada. However, both the median apophysis and the conductor are developed from the claw fundamental dorsal lobe, which separates early in its ontogeny from the rest of the bulb. As a consequence, the two structures should be intimately related and more or less independent from the rest of the sclerites (Coddington 1990), Nevertheless, the posterior apophysis and external sclerite of Dysdera are not only clearly separated one from the other but they form part of the scler- ites, the tegulum and the embolic division, re- spectively, that are supposed to be indepen- dently derived during ontogeny. Therefore, the posterior apophysis and the external scler- ite are better considered as apomorphic fea- tures of Dysdera. The relative development and degree of fusion of the external and in- ternal sclerites is variable. The IS is usually more or less straight. A frontal apophysis (FA) is sometimes present in IS proximal region. In some species, an expansion of the distal in- ternal part of the DH has been observed. When this happens the IS usually covers the external and anterior sides of the expansion, thus assuming the appearence of a crest, here referred to as the “DD internal expansion.” However, this structure is different from some crest-like ridges that may be present on the anterior distal part of the IS. These ridges may be straight and parallel to the IS, which char- acterizes the Canarian Dysdera species, or arch-like and opened to the distal tip. Here- after, the former crest is referred to as C while the second one is simply called “arch-like ridge” (AR). The distal external margin of the IS may be already expanded. This expansion is sheet- like and laterally projected over the ES and is called the “lateral fold” (LF, not shown). The lateral fold has several levels of development. In some Canarian species, it is very reduced and only visible at the distal tip of the DD, being called the “additional crest” (AC). In other instances, the LF is strongly sclerotized and apophysis-like, and is referred to as the “medial apophysis” (MA). The ES is markedly bent in the middle, going from the anterior side to the posterior one. There- fore, the distal part of the DH is anteriorly covered by the IS and posteriorly covered by the ES. The ES is usually laterally expanded in a sheet-like structure called the “lateral sheet” (L). The external margin of this struc- ture may be sclerotized. The degree of devel- opment of the L is very variable. In some ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 265 Figures 11-16. — Diagrams showing the vulva characters discussed in text. 11, Sagital section of the female genital region; 12, Anterior diverticle, dorsal view; 13, Anterior diverticle (AVD absent), ventral view; 14, Anterior diverticle (AVD present), ventral view; 15, Transversal section of the anterior diverticle (AVD present), posterior view; 16, Pos- terior diverticle, dorsal view. Canarian species, a small apophysis, anteri- orly projected, has been recorded, and is named the “lateral sheet apophysis” (LA). Posteriorly, the ES border may be fused to the DH or may form a rim, which is called the “additional lateral sheet” (AL). The border of this rim is generally smooth, although some species have a toothed margin. Finally, in some species the distal tip of the AL is pro- jected in a flagellum (F). Vulva: (Figs. 11-16) The female genitalia are entirely internal. Mcheidze (1972) coined the term “endogynum” to refer to this struc- ture in contrast to the “epigynum” or external female genital structures of the entelegyne spiders, although the more general term vulva was preferred in this study. The genitalic fur- row (G), located in the anterior ventral region, gives rise to the internal bursa (B) which is divided into two diverticles, an anterior div- erticle (AD) and a posterior one (PD). These two pouches are also separated dorsally by the oviduct opening (O). The posterior diverticle is usually more developed than the anterior one and is mostly membranous with the single exception of the transversal bar (TB). This structure is located on the anterior dorsal mar- gin of the posterior diverticle. There is a semi- circular sheet-like expansion on its anterior border, the “bursal valve” (V), which fits with the anterior diverticle, closing the oviduct opening to the bursa. The anterior diverticle holds nearly all the female genitalic characters used in the taxonomy not only of the genus but of the entire family. The anterior diverticle is further divided into two pouches, a dorsal diverticle and a ventral one, by a middle in- vagination of its lateral walls. This fold is called the “major fold” (MF). The dorsal an- terior diverticle is commonly highly sclero- tized, and is referred to as the “dorsal arch” (DA). The dorsal side of the DA, called the “dorsal fold” (DF), is responsible for locking the V. Additional lateral folds may be found in the DA. The ventral diverticle is called the “ventral arch” (VA) in contrast to the DA. It roughly corresponds to the “ventral plate” de- fined by Deeleman-Reinhold & Deeleman (1988). The anterior part of the VA is bent upwards, limiting the most anterior margin of the DA. An additional lateral fold of the VA, resulting in an “additional ventral diverticle” (AVD), has been reported in some Canarian Dysdera. The level of sclerotization of the VA is very variable and is very useful in both tax- onomy and phylogeny. Unfortunately, draw- ings of the ventral vulva are very scarce in the taxonomic studies of the Dysderidae. Finally, a T-shaped, completely sclerotized spermathe- ca (S) is found in the anterior ventral region of the VA. Spinnerets and associated spigot glands were assigned after Platnick et al. (1991). All taxonomic characters were recorded in DEL- TA format (Dallwitz 1980, 1993). All mea- surements are in mm. Abbreviations used in text and figures are as follows. Eyes: AME: anterior medial eyes, PME: posterior medial eyes, PLE: posterior lateral eyes; cheliceral teeth: B: basal tooth, M: medial tooth, D: dis- tal tooth; male copulatory bulb: T: tegulum, SP: spermophore, DD: distal division, IS: in- ternal sclerite, FA: frontal apophysis, ES: ex- ternal sclerite, DH: distal haematodoca, SD: 266 THE JOURNAL OF ARACHNOLOGY seminiferous duct, C: crest, AR: arch-like ridge, MA: medial apophysis, AC: additional crest, LF: lateral fold over L, between internal and external sclerites, L: lateral sheet, LA: lat- eral sheet apophysis, AL: additional lateral sheet at back internal border, F: flagellum, P: posterior apophysis; female genitalia: G: gen- italic furrow, B: internal bursa, AD: anterior diverticle, PD: posterior diverticle, O: oviduct opening, DA: dorsal arch, DF: dorsal fold, MF: major fold, S: spermatheca, TB: trans- versal bar, V: bursal valve, VA: ventral arch, AVD: additional ventral diverticle; spinnerets: ALS: anterior lateral spinnerets, PMS: poste- rior medial spinnerets, PLS: posterior lateral spinnerets, ms: major ampulate gland spigot, ps: polar pyriform gland spigot. In order to test if the eastern islands were significantly poorer in number of endemic species that the remaining Canaries, the log- transformed number of Dysdera species in each island was plotted against the log-trans- formed island age. The regression coefficient and a 95% confidence interval were calculated for the whole set of islands and with the east- ern islands removed. RESULTS Family Dysderidae Genus Dysdera Latreille 1804 Dysdera alegranzaensis Wunderlich 1991 Figs. 17-22, 23-26, 27, 28 Dysdera alegranzaensis Wunderlich 1991: 287- 288, figs. 7-9 (Holotype male; from the ridge of the Caldera, Alegranza; June 1990; P. Oromi leg.; #02748, stored at UL; examined). Diagnosis. — This species closely resembles D. longa, D. nesiotes and D. spinisorsum in somatic morphology and genitalia. Males are distinguished from the former species by showing a remarkable reduction in size of the bulb crest (C) (Fig. 24) and lacking the fla- gellum (Fig. 23). In females, vulva DA is dis- tinctly shortened in length and back lateral margins are truncated (Fig. 20). Additionally, males can be distinguished from the sympatric D. nesiotes by having a distal division (DD) not bent in relation to the tegulum (T) (Fig. 19) and having the lateral sheet apophysis (LA) expanded over the lateral sheet (L) (Fig. 23). Description. — Male holotype: (Figs. 17— 19, 23-24). Carapace (Fig. 17) 4.48 long; maximum width 3.43; minimum width 2.31. Brownish-red, frontally darker, becoming lighter towards back; slightly foveate at bor- ders, slightly wrinkled with small black grains mainly at front. Frontal border roughly trian- gular, from 1/2-% carapace length; anterior lat- eral borders convergent (very slightly); round- ed at maximum dorsal width point, back lateral borders straight; back margin wide, straight. AME diameter 0.25; PLE 0.2; PME 0.16; AME on edge of frontal border, sepa- rated from one another by about % diameter, close to PLE; PME very close to each other, about Vs PME diameter from PLE. Labium trapezoid-shaped, base wider than distal part; longer than wide at base; semicircular groove at tip. Sternum orange, frontally darker, be- coming lighter towards back; very slightly wrinkled, mainly between legs and frontal border; uniformly covered in slender black hairs. Chelicerae (Fig. 18) 1.96 long, about Vs of carapace length in dorsal view; fang medium- sized, 1.4; basal segment dorsal, ventral side completely covered with piligerous granula- tions. Chelicera inner groove short, about Vs cheliceral length; armed with three teeth and lamina at base; B > D > M (similar in size); D round, located roughly at center of groove; B close to basal lamina; M at middle of B and D. Front legs dark orange, back legs yellow. Lengths of male described above: fel 3.73; pal 2.56; til 3.77; mel 3.45; tal 0.7; total 14.21; fe2 3.4; pa2 2.33; ti2 3.62; me2 3.54; ta2 0.79; total 13.68; fe3 2.61; pa3 1.44; ti3 1.72; me3 2.47; ta3 0.63; total 8.87; fe4 3.54; pa4 2; ti4 2.65; me4 3.4; ta4 0.79; total 12.38; relative length: 1 -2-4-3; fe palp 2.23; pa palp 1.12; ti palp 0.93; ta palp 0.88; total 5.16. Spi- nation: palp, legl, leg2 spineless. Fe3d spines in one row: 2-3; ti3d spines arranged in two bands: proximal 1.2.1; distal 1.0.1; ti3v spines arranged in two bands: proximal 1.0.1; distal 1.0.0; with two terminal spines. Fe4d spines in two rows: anterior 3; posterior 6; ti4d spines arranged in two bands: proximal 1.1.1; distal 1.0.1; ti4v spines arranged in two bands: proximal 1.0.1; distal 1.0.1; with two terminal spines. Dorsal, ventral side of pedipalp cov- ered with small piligerous grains (scarcely); very long hairs on back legs as well as on pedipalps. Claws with 8 teeth or less; hardly larger than claw width. Abdomen 10.7 long; ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 267 Figures 17-22. — Dysdera alegranzaensis. 17, Carapace, dorsal; 18, Left chelicera, ventral; 19, Right male bulbus, external. 20-22, Vulva anterior diverticle: 20, Dorsal; 21, Ventral (S separated); 22, Lateral. Scale bars in mm. cream-colored; cylindrical. Abdominal dorsal hairs 0.144 long; thick, roughly straight, com- pressed, lanceolate; uniformly, thickly distrib- uted. Male copulatory bulbus (Fig. 19) T as long as DD; external, internal distal border sloped backwards. DD not bent in lateral view; in- ternal distal border markedly expanded. ES wider, more sclerotized than IS; IS continuous to tip. DD tip (Figs. 23-25) straight in lateral view. C present, short; distal end on DD in- ternal tip; poorly developed; located close to DD distal tip; proximal border sharply de- creasing; distal border truncated, upper tip not projected, rounded, external side smooth. LF absent. L well-developed; external border sclerotized, laterally markedly folded, distally projected; distal border divergent, continuous. LA present, sheet-like; as long as L, distally not fused. F absent. AL present, well-devel- oped; proximal border in posterior view smooth, not fused with distal haematodoca. P (Fig. 26) fused to T; perpendicular to T in lat- eral view; lateral length from of T width; ridge present, perpendicular to T; distinctly expanded, right-angled; upper margin smooth; not distally projected; back margin not folded. Female: (Figs. 20-22, 27, 28). All charac- ters as in male except: Carapace 5.25 long; maximum width 4.02; minimum width 2.83. Deep red. Back lateral borders straight. AME diameter 0.25; PLE 0.21; PME 0.2. Chelicerae 2.33 long; fang 1.57. D = B > M (similar). Legs dark orange. Lengths of fe- male described above: fel 4.19; pal 2.89; til 4.47; mel 3.73; tal 0.74; total 16.02; fe2 3.63; pa2 2.61; ti2 3.45; me2 3.62; ta2 0.7; total 14.01; fe3 2.98; pa3 1.81; ti3 2.09; me3 2.98; ta3 0.74; total 10.6; fe4 3.96; pa4 2.28; ti4 2.89; me4 3.86; ta4 0.84; total 13.83; relative length 1 -2-4-3; fe palp 2.14; pa palp 1.21; ti palp 0.98; ta palp 1.16; total 5.49. Spination: 268 THE JOURNAL OF ARACHNOLOGY Figures 23-28. — Dysdera alegranzaensis, right male bulbus. 23, DD frontal; 24, DD external; 25, DD posterior; 26, P internal. 27-28, Dysdera alegranzaensis, spinnerets. 27, Right ALS; 28, Right PLS. palp, legl, leg2 spineless. Fe3d spines in one row: 1; ti3d spines arranged in two bands: proximal 1.2.1; distal 1.0.1.; ti3v spines ar- ranged in two bands: proximal 1.0.0; distal 1.0.0; with two terminal spines. Fe4d spines in two rows: anterior 1; posterior 5; ti4d spines arranged in two bands: proximal 1.1.1; distal 1.0.1; ti4v spines arranged in two bands: proximal 1.0.1; distal 0.0. 1 ; with two terminal spines. Abdomen 10.74 long. Abdominal dorsal hairs 0.18. Vulva (Fig. 20-22) DA not distin- guishable from VA; rectangular; DA twice as wide as long; DF wide in dorsal view. MF well-developed, completely sclerotized. VA frontal region completely sclerotized; poste- rior region sclerotized at most anterior area; tooth- shaped expansion from internal back border, not joined to lateral sclerotization, about half of DF lateral margins; AVD absent. S attachment projected under VA; arms as long as DA, straight; tips dorsally projected; neck as wide as arms. TB usual shape. ALS (Fig. 27) with PS; remaining piriform spigots ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 269 Table 1. — Intraspecific spination variability of Dysdera alegranzaensis. Proximal Medial-proximal Medial-distal Distal Tibia 3 dorsal 1.2-4. 1 0 0 l.O-l.l Tibia 4 dorsal O-l.O-l.l 0 0 O-LO-l.O-l Tibia 3 ventral LO.0-1 O.O.O-l 0 O-l.O.O-l Tibia 4 ventral 1.0.1 O.O.O-l. 0 O-l.O.l Number of rows Number of spines Femur 3 dorsal 0-1 0-1 Femur 4 dorsal 2 1-3/2-6 more external than MS, arranged in two rows; 8+1 piriform gland spigots; PMS, PLS (Fig. 28) with 10-15 aciniform gland spigots. Intraspecific variation.^ — Male cephalo- thorax ranges in length from 3.99-4.48, fe- male from 3.57-5.25. AME separation from PLE-PME from PME diameter to y2- Carapace ornamentation somewhat reduced, nearly smooth. Chelicera relative size up to % of the carapace length. Distal reduction of the chelicera granulations in some female speci- mens. Relative size of B and D variable, M always the smallest. Some female palps with ventral granulation. Spination variability in Table 1. Additional material examined. — Alegranza: El Faro, 6 April 1993, 13, (P. Oromi, 2530 UL). Inside the Caldera, June 1990, ljuv., (P, Oromi, 2735 UL). Unknown locality, 3rd week March 1995, 13, (P Oromi, 4106 UB); June 1990, 19, 3 juv., (P Oromi, 2733 UL). La Graciosa: Caldera de Pedro Barba, 30 March 1996, 13, (P. Oroiru, 3134 UB). Montana del Mojon, 30 March 1996, 1 9, (P. Oromi, 3137 UB). Lanzarote: Han'a: Mon- tanas de Famara, around Mirador del Rfo, Novem- ber 1988, 19, (A. Enghoff, 2670 ZMK); 22 Feb- ruary 1995, 33 (Arnedo, Ribera & Oromf, #2858-59, 4076 UB); 39, (Arnedo, Ribera & Oroim, #4080, 4104-5 UB). Yaiza: Montanas de Femes, Atalaya de Femes; 22 February 1995, 29, (Amedo, Ribera & Oroim, #4089-90 UB). Distribution.— Endemic species from Lan- zarote and Northern islets. Comments. — This species had only been reported from the rocky island of Alegranza before the present study. Dysdera lancerotensis Simon 1907 Figs. 29-34, 36, 38-40, 41, 42 Dysdera crocata lancerotensis Simon 1907: 258. (Types; 3339; unknown locality, Lanzarote; Ch. Alluaud leg.; #15586, stored at MHNP; exam- ined). Dysdera crocota lancerotensis: Denis 1941: 108. — Schmidt 1973: 360-361. Dysdera lancerotensis: Wunderlich 1991: 296-298, figs. 50-52 [3 9]. Diagnosis. — This species strongly differs from any other Canarian endemics. It closely resembles the cosmopolitan species D. cro- cota C.L. Koch 1839, from which both sexes can be distinguished by a spiny fel (although is not always so), males by the shape of the distal division (DD) tip in frontal view (Fig. 36-38) and the presence of two or three ridges on the posterior apophysis (P) upper margin (Fig. 39), and females by the dorsal shape of the dorsal arch (DA), the frontal projection of the ventral arch (VA) under the dorsal one (DA) and the presence of a tiny strip con- necting frontally the dorsal arch with the sper- matheca (S) attachment (Figs. 32, 35). Description. — Male: (Figs. 29-31, 36, 38- 39). Carapace (Fig. 29) 3.43 long; maximum width 2.87; minimum width 2.1. Uniformly dark red, slightly foveate at borders, slightly wrinkled with small black grains mainly at front. Frontal border roughly round, about f carapace length; anterior lateral borders con- vergent (slightly); rounded at maximum dorsal width point, back lateral borders rounded; back margin wide, bilobulated; slightly stepped in lateral view. AME diameter 0.2; PLE 0.18; PME 0.14; AME slightly back from frontal border, separated from one another by about % diameter, close to PLE; PME very close to each other, less than PME diameter from PLE. Labium trapezoid- shaped, base wider than distal part; as long as wide at base; semicircular groove at tip. Sternum uniformly orange; very slightly wrinkled, mainly be- 270 THE JOURNAL OF ARACHNOLOGY Figures 29-35. — Dysdera lancerotensis. 29, Carapace, dorsal; 30, Left chelicera, ventral; 31, Right male bulbus, external. 32-34, Vulva anterior diverticle: 32, Dorsal; 33, Ventral; 34, Lateral. 35, Dysdera crocota, vulva anterior diverticle, dorsal. Scale bars in nun. tween legs and frontal border; uniformly cov- ered in slender black hairs. Chelicerae (Fig. 30) 1.82 long, about Vi of carapace length in dorsal view; fang long, 1.54; basal segment dorsal side completely covered with piligerous granulations (sparse), ventral side smooth. Chelicera inner groove long, about Vi chelic- eral length; armed with three teeth and lamina at base; D=B>M; D trapezoid, located rough- ly at centre of groove; B close to basal lamina; M close to B. Legs orange. Lengths of male described above: fel 2.56; pal 1.58; til 2.24; mel 2.33; tal 0.65; total 9.36; fe2 2.28; pa2 1.4; ti2 1.96; me2 2.1; ta2 0.65; total 8.39; fe3 2; pa3 1.16; ti3 1.3; me3 1.77; ta3 0.56; total 6.79; fe4 2.47; pa4 1.3; ti4 1.91; me4 2.33; ta4 0.65; total 8.66; relative length: 1-4-2-3; fe palp 1.67; pa palp 0.93; ti palp 0.79; ta palp 0.93; total 4.32. Spination: palp, legl, leg2 spineless. Fe3d spineless; ti3d spines arranged in two bands: proximal 1.0.1; distal 1.0.1; ti3v spines arranged in one band: proximal 0.1.0; with two terminal spines. Fe4d spines in one row: 3; ti4d spines arranged in two bands: proximal 1.0.1; distal 1.0.1; ti4v spines ar- ranged in one band: proximal 0.0- 1.0; with two terminal spines. Dorsal side of frontal legs covered with small piligerous grains; ventral side covered with hairs, lacking grains. Claws with 8 teeth or less; hardly larger than claw width. Abdomen 4.48 long; whitish; cy- lindrical. Abdominal dorsal hairs 0.036 long; thin, roughly straight, not compressed, blunt, tip enlarged; uniformly, scantly distributed. Male copulatory bulbus (Fig. 31) T as long as DD; external distal border straight; internal projected at middle. DD bent about 45° in lat- eral view; internal distal border not expanded. ES wider, more sclerotized than IS; IS contin- uous to tip. DD tip (Figs. 36, 38-39) straight in lateral view; posterior (lower) sheet pro- jected under frontal (upper) one; posterior ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 271 Figures 36-42. — Dysdera lancerotensis, right male bulbus. 36, DD frontal; 37, DD frontal of Dysdera crocota; 38, DD, external; 39, DD posterior; 40, P external. 41-42, Dysdera lancerotensis, spinnerets. 41, Right ALS; 42, Right PLS. sheet distal internal margin sloped; arch-like ridge present. MA present; hook-like; single pointed projection at internal base. C absent. L absent or hardly visible. LA absent. F ab- sent. AL absent. P (Fig. 40) not fused to T; parallel to T on its proximal part, perpendic- ular on distal; lateral length from ^3-% Of T width; ridge present, parallel to T; not ex- panded; upper margin markedly toothed, on its distal part, very few teeth (1-3); not dis- tally projected; back margin not folded. Female: (Figs. 32-34, 41, 42). All charac- ters as in male except: carapace 3.85 long; maximum width 3.22; minimum width 2.38. AME diameter 0.21; PLE 0.18; PME 0.16. Chelicerae 2.03 long; fang long, 1.89. Lengths of female described above: fel 2.8; pal 1.72; til 2.33; mel 2.33; tal 0.6; total 9.78; fe2 2.56; pa2 1.49; ti2 2.1; me2 2.19; ta2 0.56; total 8.9; fe3 1.96; pa3 1.16; ti3 1.4; me3 1.91; ta3 0.56; total 7; fe4 2.61; pa4 1.4; ti4 1.86; me4 2.56; ta4 0.65; total 9.08; relative length 1 -4-2-3; fe palp 1.86; pa palp 0.83; ti palp 0.79; ta palp 1.26; total 4.74. Spination: palp 272 THE JOURNAL OF ARACHNOLOGY Table 2. — Intraspecific spination variability of Dysdera lancerotensis. Proximal Medial-proximal Medial-distal Distal Tibia 3 dorsal LOT 0 0 1.0.1 Tibia 4 dorsal O-l.OT O-l.O.O 0 O-l.OT Tibia 3 ventral O-LO-2.0 0 0 0 Tibia 4 ventral O-LO-2.0 0 0 0 Number of rows Number of spines Femur 1 frontal distal 2 0-2 Femur 2 frontal distal 1 0-1 Femur 3 dorsal 0 0 Femur 4 dorsal 2 0-1/0-3 spineless. Pel two terminal spines on anterior margin. Fe2 one terminal spine on the anterior margin. Fe3d spineless; ti3d spines arranged in two bands: proximal 1.0.1; distal 1.0.1; ti3v spines arranged in one band: proximal 0.1.0; with two terminal spines. Fe4d spines in one row: 3-2; ti4d spines arranged in two bands: proximal 0.0.1; distal 1.0. 1-0; ti4v spines ar- ranged in one band: proximal 0.1.0; with two terminal spines. Dorsal, ventral side of pedi- palp covered with hairs, lacking grains. Abdomen 5.95 long; whitish; cylindrical. Abdominal dorsal hairs 0.054 long; thin, roughly straight, not compressed, blunt, tip enlarged; uniformly, scantly distributed. Vul- va (Fig. 32”34) DA clearly distinguishable from VA; DA slightly wider than long; DF narrow in dorsal view. MF margins not fused, poorly developed, membranous. VA rectan- gular; projected under DA; frontal region with a narrow sclerotized band connecting S at- tachment to DA; posterior region not sclero- tized; AVD absent. Ventral narrow dark bands developed from S attachment. S attached to membranous VA; arms as long as DA, clearly curved; tips not projected; neck as wide as arms. TB usual shape. ALS (Fig. 41) with PS; remaining piriform spigots more external than MS, arranged in three rows; 12+1 piriform gland spigots; PMS, PLS (Fig. 42) with 10- 15 aciniform gland spigots. Intraspecific variation* — Male cephalo- thorax ranges in length from 2.81-4.06, fe- male from 2.94--4.69. AME separation rang- ing from % diameter to 1. PLE-PME ranging from Va PME diameter to f. Sternum moder- ately wrinkled. D from markedly larger than B to as large as B. One specimen from La Graciosa has D under groove middle point. P transversal ridges reduced to two. DA frontal border sometimes straight. S shape somewhat variable. Spination variability in Table 2. Additional material examined. — Alegranzai unknown locality, 3rd week March 1995, 23, (P. Oromi, #4115 UB, #2892 UL); unknown date, 19; (P. Oromi, #4173 UB). Fuerteventuraj La Oliva: E from Punta Ballena, N from Cotillo; 6 September 1990, 19, (H. Enghoff & M. Baez, #2631 ZMK). Cotillo-Los Lagos; 10 February 1997, 19, (P. Oromi, 3185 UL), Malpais de Bayuyo, 20 February 1995; 13, (Amedo, Ribera & Oromi, #2855 UB); 29, (Amedo, Ribera & Oromi, #2856, 4071 UB). Pdjara: Bco. del Ciervo, Morro de Cavedero N from Morro Jable, Jandia, 4 January 1990, 43, (H. Enghoff & M. Baez, #2633-35 ZMK); 19, (H. Enghoff & M. Baez, #2632 ZMK); ljuv., (H. Engh- off & M. Baez, #2633 ZMK); 17 Febmary 1995, 23, (Amedo, Ribera & Oromi, #2840, 4057 UB). La Graciosa: Caleta del Sebo; 31 March 1996, 1 9, (R Oromi, 3135 UB). Playa Lambra, 1 April 1996, ljuv., (R Oromi, 3136 UB). Lanzarote: Haria: Fa- mara, Mirador del Rio, 15 March 1995, 23, (un- known, #4103, 4179 UB). Yaiza: Montanas de Fe- mes, Atalaya de Femes, 22 Febraary 1995, 23, (Amedo, Ribera & Oromi, #2869, 4092 UB); 19, (Amedo, Ribera & Oromi, #2870 UB). Montana Clara: La Caldera, 23 Febmary 1995, 23, (Ame- do, Ribera & Oromi, #2871, 2872 UB); ljuv., (Ar- nedo, Ribera & Oromi, #4178 UB). Distribution. — ^Endemic species from the eastern Canaries. Dysdera Uostethus Simon 1907 Dysdera Uostethus Simon 1907: 261, fig. 4E [3]. (Type lost). D. clavisetae Wunderlich 1991: 291-292, figs. 24- 27 [3,9] (Holotype female; Mirador de Frontera, El Golfo, El Hierro, 8 July 1973, J. Wunderlich leg., not examined. Paratypes; 13, Mirador de Frontera, El Golfo, El Hierro, 8 July 1973, J. ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 273 Wunderlich leg., #03842, stored at UL, examined. Id, MSS Salvador-3, El Hierro, 19 August 1987, A.L. Medina leg., #H-C3-378, stored at UL, ex- amined). -Amedo et al. 1996: 247-251, figs. 6A- D, 7A-D and 8A-B [d,9]. New synonymy. Distribution. — Widely-spread species in the islands of La Gomera and El Hierro (Wun- derlich 1991; Amedo et al. 1996). Its presence in Lanzarote is considered to be doubtful. Comments. — The only known material as- signed to this species was a male used in the original description (Simon 1907). With the sole exception of D. lancerotensis, all male types of the seven Canarian species described by Simon, which were supposed to be stored at MHNP, seem to have been lost (Wunderlich 1987). These type material could not be lo- cated either in other museums where Simon’s type material from Iberian and north African species (MNCN and BMNH) or other Can- arian types (MCNT and UL) were kown to be stored. Finally, the late arachnologist Dr. P. Brignoli had been loaned a significant amount of type material from various European mu- seums. Because Dr. Brignoli had published a number of papers on the family Dysderidae, there was a chance that some of Simon’s ma- terial from the Canaries was in his possession. However, the current curators of his personal collection were unable to locate these speci- mens. Most of characters given in the original de- scription of D. liostethus are not species-di- agnostic for Canarian Dysdera. However, the spination pattern is, in this case, very infor- mative. This species is said to share a similar chaetotaxia with D. rugichelis Simon 1907. Femora with numerous spines arranged in two assymetric rows and a strongly spinate pos- terior tibiae characterize the latter species. This spination pattern is very particular and has only been observed in D. clavisetae Wun- derlich 1991, D. enghoffi Amedo, Oromf & Ribera 1996, D. hirguan Amedo, Oromf & Ribera 1996, from La Gomera, D. ratonensis Wunderlich 1991, from La Palma and D. ver- neaui Simon 1883, from Gran Canaria. Dys- dera verneaui could be removed from the list because it was described by the same author and a synonymy is very unlikely. Dysdera ra- tonensis and D. hirguan are very large species (more than 14 mm in total length), which does not fit with the total length reported for D. liostethus (8 mm). Finally, in D. enghojfi the dorsal side of the basal segments of the che- licerae is completely covered with granula- tions and its copulatory bulbus is character- ized by a T and a DD of equal size. In contrast, D. liostethus is supposed to have chelicerae in which the basal segment is scarcely covered with granulations, and in the drawing of the male palp, a markedly longer DD than T can be observed. The only re- maining species D. clavisetae fits these fea- tures perfectly. However, there are still two arguments against the synonymy. First, the P of the male bulbus in Simon’s drawing is very short while D. clavisetae has a long P. How- ever, P development has been shown to be polymorphic in other Canarian endemic Dys- dera, e.g., Dysdera macra Simon 1883 (Ar- nedo & Ribera 1999). The second problem has to do with the original type locality. How- ever, this argument is not against this synon- ymy in particular but to any presence of this kind of male genital pattern in the eastern Ca- naries. The drawing of the male palp of D. liostethus suggest a combination of characters that has only been observed in endemic spe- cies from the central and western islands. This genitalic pattern is characterized by a tegulum (T) slightly smaller than the distal division (DD), a short but well-developed crest (C), which is located at the DD distal tip, a well- developed lateral sheet (L) with a membra- nous external lateral border and without apophysis (LA), and a poorly developed AC. Moreover, additional cases of mistakenly assigned localities in the same article have been demonstrated (Arnedo et al. 1996). Therefore, the original type locality of D. lios- tethus is considered to be doubtful, at least. Finally, a synonymy of both species is con- sidered to be preferable to an unnecessary proliferation of names. Dysdera longa Wunderlich 1991 Figs. 43-49, 50-53, 54, 55 Dysdera longa Wunderlich 1991: 298, figs. 53-56 [(3,9]. (Holotype male; Morro de Cavedero N from Mono Jable, Pajara, Fuerteventura; 4 Jan- uary 1990; H. Enghoff & M. Baez leg.; #298, stored at ZMK; examined. Paratypes: 1(3, 19,2 juv.; Cumbres de Jandia, Pajara, Fuerteventura; 27 February 1990; P. Oromf leg.; #2710, stored at UL; examined). Diagnosis. — Very large Dysdera similar to remaining eastern species, apart from D. Ian- 274 THE JOURNAL OF ARACHNOLOGY 1 Figures 43-49. — Dysdera longa. 43, Carapace, dorsal; 44, Left chelicera, ventral; 45, Right male bulbus, external; 46, Male abdomen, lateral. 47-49, Vulva anterior diverticle: 47, Dorsal; 48, Ventral (S separated); 49, Lateral. Scale bars in mm. cerotensis, especially in genitalic pattern. Both sexes can be distinguished from the for- mer species by its larger size (carapace length > 6), the dorsal projection of the distal region of the abdomen (mainly in males) (Fig. 46), and the lanceolated hairs not being posteriorly curved. Males have bulb tegulum (T) mark- edly larger than the distal division (DD) (Fig. 45) and having a sheet-like crest (C) laterally expanded (Fig. 51), while in females the vulva dorsal (DA) and ventral (VA) archs lateral borders are separated (Fig. 49). Description.^ — Male holotype: (Figs. 43- 46, 50-53). Carapace (Fig. 43) 7.07 long; maximum width 5.53; minimum width 3.29. Reddish-orange, frontally darker, becoming lighter towards back; slightly foveate at bor- ders, slightly wrinkled with small black grains mainly at front. Frontal border roughly round, from !/2-% carapace length; anterior lateral borders convergent; pointed at maximum dor- sal width, back lateral borders straight; back margin wide, straight. AME diameter 0.36; PLE 0.31; PME 0.25; AME on edge of frontal border, separated from one another by about 36 diameter, close to PLE; PME very close to each other, less than Va PME diameter from PLE. Labium trapezoid-shaped, base wider than distal part; longer than wide at base; semicircular groove at tip. Sternum reddish- orange, frontally darker, becoming lighter to- wards back; very slightly wrinkled, mainly between legs and frontal border; uniformly covered in slender black hairs. Chelicerae (Fig. 44) 3.29 long, about % of carapace length in dorsal view; fang medium- sized, 2.5; basal segment dorsal, ventral side completely covered with piligerous granula- tions. Chelicera inner groove short, about 36 cheliceral length; armed with three teeth and lamina at base, additional ventral tooth on left chelicera; B > D = M (similar); D round. ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 275 located roughly at center of groove; B close to basal lamina; M at middle of B and D. Front legs dark orange, back legs yellow. Lengths of male described above: fel 5.81; pal 3.91; til 6.16; mel 5.81; tal 1.12; total 22.81; fe2 4.9; pa2 3.5; ti2 4.97; me2 4.55; ta2 1.02; total 18.94; fe3 3.64; pa3 2.33; ti3 3.64; me3 2.59; ta3 0.84; total 13.04; fe4 4.83; pa4 3.03; ti4 4.13; me4 4.69; ta4 1.07; total 17.75; relative length: 1-2-4-3; fe palp 3.49; pa palp 1.63; ti palp 1.77; ta palp 1.63; total 8.52. Spination: palp, legl, leg2 spineless. Fe3d spineless; ti3d spines arranged in two bands: proximal 1.0.0; distal 1.0.1; ti3v spines arranged in one band: proximal 0.0.1; with two terminal spines. Fe4d spines in one row: 13; ti4d spines arranged in two bands: proxi- mal 0.0. 1 ; distal 0.0. 1 ; ti4v spines arranged in one band: proximal 0.0.1; with two terminal spines. Dorsal, ventral side of pedipalp cov- ered with hairs, lacking grains; very long hairs on back legs as well as on pedipalps. Claws with 8 teeth or less; hardly larger than claw width. Abdomen 1 1 long; cream-colored; back end projected upwards in lateral view (Fig. 46). Abdominal dorsal hairs 0.108 long; thick, roughly straight, compressed, lanceo- late; uniformly, thickly distributed. Male copulatory bulbus (Fig. 45) T twice as long as DD; external, internal distal border sloped backwards. DD bent about 45° in lat- eral view; internal distal border not expanded. ES wider, more sclerotized than IS; IS contin- uous to tip (slim). DD tip (Figs. 50-52) straight in lateral view. C present, long; distal end beside DD internal tip; distal border trun- cated, toothed, markedly expanded, projected over DD external part. LF absent. L well- developed; external border sclerotized, lat- erally markedly folded, distally projected; distal border divergent, continuous. LA pre- sent, hook-like; shorter than L. F present, straight, proximally fused to DD. AL present, well-developed, joined to flagellum; proximal border in posterior view smooth, not fused with distal haematodoca. P (Fig. 53) fused to T; perpendicular to T in lateral view; lateral length from Vi-Vs of T width; ridge present, perpendicular to T; distinctly expanded, right- angled; upper margin smooth; not distally pro- jected; back margin not folded. Female paratype: (Figs. 47-49, 54, 55). All characters as in male except: Carapace 6.79 long; maximum width 5.25; minimum width 3.78. Back lateral borders straight. AME di- ameter 0.36; PLE 0.32; PME 0.27; AME on edge of frontal border, separated from one an- other by about % diameter, close to PLE; PME very close to each other, less than V4 PME di- ameter from PLE. Chelicerae 3.12 long; fang medium-sized, 2.9; B > D = M (similar). Legs dark orange-colored. Lengths of female described above: fel 8.26; pal 5.6; til 7.21; mel 7.21; tal 1.4; total 29.68; fe2 6.65; pa2 5.18; ti2 6.02; me2 6.02; ta2 1.47; total 25.34; fe3 5.25; pa3 3.15; ti3 3.85; me3 5.04; ta3 1.26; total 18.55; fe4 7; pa4 3.92; ti4 5.6; me4 6.58; ta4 1.75; total 24.85; relative length 1- 2-4-3; fe palp 4.9; pa palp 2.66; ti palp 2.1; ta palp 2.8; total 12.46. Spination: palp, legl, leg2 spineless. Fe3d spineless; ti3d spines ar- ranged in two bands: proximal 1.1.0; distal 1.0.1; ti3v spines arranged in one band: prox- imal 1.0.0; with two terminal spines. Fe4d spines in one row: 11-10; ti4d spines arranged in two bands: proximal 0.0.1; distal 0.0.1; ti4v spines arranged in one band: proximal 1.0.1; with two terminal spines. Dorsal side of fron- tal legs covered with small piligerous grains (sparse). Abdomen 1 1 long; cream-colored; back end projected upwards in lateral view (slightly). Abdominal dorsal hairs 0.56 long; thick, roughly straight, compressed, lanceolate; uni- formly, thickly distributed. Vulva (Figs. 47- 49) DA clearly distinguishable from VA; DA slightly wider than long; DF wide in dorsal view. MF margins not fused, well-developed, anterior region sclerotized. VA rectangular, pointed expansion at middle frontal part; pro- jected under DA; frontal region completely sclerotized; posterior region sclerotized at lat- eral margins; AVD absent. S attachment pro- jected under VA; arms as long as DA, straight; tips not projected; neck as wide as arms. TB usual shape. ALS (Fig. 54) with PS; remain- ing piriform spigots more external than MS, arranged in two rows; 13 + 1 piriform gland spigots; PMS, PLS (Fig. 55) with more than 20 aciniform gland spigots. Intraspecific variation. — Male cephalo- thorax ranges in length from 6.30-7.21, fe- male from 6.02-7.35. AME separation from Vs diameter to Vi. PLE-PME from Vs PME dia- mater to %. Sternum ornamentation sometimes reduced. Relative size of cheliceral teeth var- iable although no large differences in size. P 276 THE JOURNAL OF ARACHNOLOGY Figures 50-55, — Dysdera longa, right male bulbus. 50, DD frontal; 51, DD external; 52, DD posterior; 53, P internal. 54-55, Dysdera longa, spinnerets. 54, Right ALS; 55, Right PLS. Table 3. — Intraspecific spination variability of Dysdera longa. Proximal Mediahproximal Medial-distal Distal Tibia 3 dorsal 1.0-2.0-1 0 0 l.O.O-l Tibia 4 dorsal O-LO.0-1 0 O-LO.O O-LO.0-1 Tibia 3 ventral l.O-LO 0 0 O-LO.O-1 Tibia 4 ventral O-LO.O-1 0 0 0 Number of rows Number of spines Femur 3 dorsal 0 0 Femur 4 dorsal 1 8-13 ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 277 back margin slightly folded. Spination vari- ability in Table 3. Additional material examined. — Fuerteventu- ra: Pdjara: Bco. del Ciervo, Cumbres de Jandia, N slope, 17 February 1995, 2d, (Amedo, Ribera & Oromi, #2836 UB, 2838 UL); 79, (Amedo, Ribera & Oromi, #2837, 3183 UL, 4054-56, 4058, 4117 UB); lOjuv., (Amedo, Ribera & Oromi, #2831-35, 2839, 4050-53 UB); 27 Febmary 1990, 19, (P. Oromi, #2621 MCNT). Distribution.— Endemic species from the Jandia peninsula, at southern Fuerteventura. Dysdera nesiotes Simon 1907 Figs. 56-63, 64-67, 68, 69 Dysdera nesiotes Simon 1907: 260-261, fig. 4G [<3] (Type lost). — Reimoser 1919: 200. — Denis 1963: 37-38.— Schmidt 1973: 360-361.— Ram- bla 1978: 132-133.— Amedo et al. 1996. Dysdera wollastoni Blackwall 1864 nec. Kulczyn- ski 1899: 23-26. fig. 22-24 [d]. — Reimoser 1919: 200.— Berland & Denis 1946: 224. Wun- derlich 1991: 312. Fig. 129 [d] Dysdera wollastoni nesiotes Simon 1912: 59-60. — Denis 1941: 108. Types. — Neotypes, by present designation, ld,19, 3juv.; label states: ''Dysdera wolias- toni Blackwall, Ins. Salvages (Garreta leg.)”; #B 536 (jar number), stored at MHNR Diagnosis.— This species strongly resem- bles D. spinidorsum. Males can be distin- guished from the latter by the short lateral apophysis (LA) (Fig. 64), the moderately ex- panded crest (C) (Fig. 65), and the presence of a fold between the additional lateral sheet (AL) and the flagellum (F) (Fig. 66). Female vulva has the backwards projection of the me- dial fold (MF) not so well developed (Fig. 59) and displays posterior sclerotization of the ventral arch (VA) (Fig. 60). Description. — Male neotype: (Figs. 56-58, 64-67). Carapace (Fig. 56) 4.23 long; maxi- mum width 3.71; minimum width 2.2. Dark brownish-orange, frontally darker, becoming lighter towards back; smooth with some small black grains mainly at front. Frontal border roughly triangular, from s-f carapace length; anterior lateral borders convergent; rounded at maximum dorsal width point, back lateral bor- ders straight; back margin wide, straight. AME diameter 0.27; PLE 0.21; PME 0.18; AME on edge of frontal border, separated from one another by about % diameter, close to PLE; PME very close to each other, about Vi PME diameter from PLE. Labium trape- zoid-shaped, base wider than distal part; lon- ger than wide at base; semicircular groove at tip. Sternum orange, frontally darker, becom- ing lighter towards back; very slightly wrin- kled, mainly between legs and frontal border; uniformly covered in slender black hairs. Chelicerae (Fig. 57) 1.82 long, about Vs of carapace length in dorsal view; fang medium- sized, 1.05; basal segment dorsal, ventral side completely covered with piligerous granula- tions. Chelicera inner groove short, about Vs cheliceral length; armed with three teeth and lamina at base; D = B > M (similar); D round, located roughly at centre of groove; B close to basal lamina; M at middle of B and D. Front legs dark orange, back legs yellow. Lengths of male described above: fel 3.5; pal 2.45; til 3.5; mel 3.29; tal 0.63; total 13.37; fe2 3.08; pa2 2.1; ti2 2.8; me2 2.94; ta2 0.7; total 11.62; fe3 3.26; pa3 1.4; ti3 1.75; me3 2.17; ta3 0.7; total 9.28; fe4 3.29; pa4 1.68; ti4 2.7; me4 3.15; ta4 0.7; total 11.52; relative length: 1 -2-4-3; fe palp 2.1; pa palp 1.12; ti palp 1.13; ta palp 1.13; total 5.48. Spination: palp, legl, leg2 spineless. Fe3d spineless; ti3d spines arranged in two bands: proximal 1.0.1; distal 1.0.1; ti3v spines arranged in two bands: proximal 1.0.0; distal 1.0.0; with two terminal spines. Fe4d spines in two rows: anterior 4; posterior 6-7; ti4d spines arranged in two bands: proximal 0.0.1; distal 0.0.1; ti4v spines arranged in two bands: proximal 1.0.1; distal 0-1. 0.0-1; with two terminal spines. Dorsal side of frontal legs covered with small pili- gerous grains; ventral side covered with hairs, lacking grains; very long hairs on back legs as well as on pedipalps. Claws with 8 teeth or less; hardly larger than claw width. Abdomen 6.86 long; whitish; cylindrical. Abdominal dorsal hairs 0. 1 1 long; thick, roughly straight, compressed, lanceolate; uniformly, thickly distributed. Male copulatory bulbus (Fig. 58) T as long as DD; external, internal distal border sloped backwards. DD bent about 45° in lateral view; internal distal border markedly expanded. ES wider, more sclerotized than IS; IS continuous to tip (diffused). DD tip (Figs. 64-67) straight in lateral view; frontal (upper) sheet internal part markedly projected above posterior (low- er) sheet. C present, long; distal end beside DD internal tip; distal border rounded, smooth, markedly expanded, perpendicular to 278 THE JOURNAL OF ARACHNOLOGY Figures 56-63. — Dysdera nesiotes. 56, Carapace, dorsal; 57, Left chelicera, ventral; 58, Right male bulbus, external. 59-63, Vulva anterior diverticle: 59, Dorsal; 60, Ventral; 61, Lateral. 62, 63, Variability, ventral. Scale bars in mm. DD. LF absent. L well-developed; external border sclerotized, laterally markedly folded; distal border divergent, continuous. LA pre- sent, hook-like; shorter than L. F present, tip bent backwards, proximally fused to DD. AL present, well-developed, not joined to flagel- lum; proximal border in posterior view smooth, not fused with distal haematodoca. P (Fig. 67) fused to T; perpendicular to T in lat- eral view; lateral length from of T width; ridge present, perpendicular to T; distinctly expanded, rounded; upper margin slightly toothed, mainly on external side, along its ex- tent, few teeth (4-6); not distally projected; back margin not folded. Female: (Figs. 60, 61, 68, 69). All charac- ters as in male except: carapace 4.55 long; maximum width 3.71; minimum width 2.38. AME diameter 0.27; PLE 0.21; PME 0.18; AME separated from one another by about % diameter. Chelicerae 1.92 long; fang medi- um-sized, 1.19. B>D>M (similar). Front legs dark orange, back legs yellow. Lengths of female described above: fel 3.36; pal 2.38; til 2.94; mel 2.8; tal 0.63; total 12.11; fe2 3.86; pa2 2.1; ti2 2.66; me2 2.66; ta2 0.63; total 11.91; fe3 2.24; pa3 1.4; ti3 1.75; me3 2.31; ta3 0.63; total 8.33; fe4 3.5; pa4 1.68; ti4 2.66; me4 3.22; ta4 0.7; total 11.76; rela- tive length 1-2-4-3; fe palp 2.2; pa palp 0.98; ti palp 0.84; ta palp 1.19; total 5.21. Spination: palp, legl, leg2 spineless. Fe3d spineless; ti3d spines arranged in two bands: proximal 1.0.1; distal 1.0.0; ti3v spines arranged in two bands: proximal 1.0.0; distal 1 -0.0.0; with two ter- minal spines. Fe4d spines in two rows: ante- ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 279 Figures 64-69. — Dysdera nesiotes, right male bulbus. 64, DD frontal; 65, DD external; 66, DD pos- terior. 67, P external. 68-69, Dysdera nesiotes, spinnerets. 68, Right ALS; 69, Right PMS (lower) and PLS (upper). rior 1; posterior 6-5; ti4d spines arranged in two bands: proximal 0.0.1; distal 0,0.1; ti4v spines arranged in two bands: proximal 1.0.1; distal 1 -2.0.0- 1; with two terminal spines. Abdomen 6.86 long; whitish; cylindrical. Abdominal dorsal hairs 0.126 long; thick, roughly straight, compressed, lanceolate; uni- formly, thickly distributed. Vulva (Fig. 60, 61) DA not distinguishable from VA; rectangular; DA twice as wide as long; DF wide in dorsal view. MF well-developed, completely sclero- tized, projected backwards, shorter than DA lateral length. VA frontal region completely sclerotized; posterior region sclerotized in most anterior area; tooth-shaped expansion from internal back border; not joined to lateral sclerotization, about half of DF lateral mar- gins; AVD absent. S attachment not projected under VA; arms as long as DA, slightly curved; ends projected forwards; neck hardly visible. TB usual shape. ALS (Fig. 68) with PS; remaining piriform spigots more external 280 THE JOURNAL OF ARACHNOLOGY Table 4. — Intraspecific spination variability of Dysdera nesiotes. Proximal Medial-proximal Medial-distal Distal Tibia 3 dorsal 1.0-2.0-1 0 0 1. 0.0-1 Tibia 4 dorsal O-l.O.l 0 0 O-l.O.l Tibia 3 ventral 0-2.0.0-1 0 0 O-l.O.O Tibia 4 ventral 0 b b 0 0 O-l.O.O-l Number of row Number of spines Femur 3 dorsal 0-1 0-2 Femur 4 dorsal 2 1-6/4-7 than MS, arranged in two rows; 10+1 piri- form gland spigots; PMS, PLS (Fig. 69) with 10-15 aciniform gland spigots. Intraspecific variation. — Male cephalo- thorax ranges in length from 3.64-4.48, fe- male from 3.92-5.46. AME separation from Vs diameter to PLE-PME from Vs PME di- ameter to V2. In general, B largest, D clearly above groove middle point and M position variable. Some female specimens have ab- dominal hairs that are not clearly lanceolated. An unusual range of variability in DA shape can be observed. Two extreme types can be recognized although several intermediate forms have been recorded. The first of them (Fig, 62) is distinguished by a markedly wide DA in dorsal view, with rectangular anterior lateral borders, tooth-like ventral sclerotiza- tion which is restricted to the frontal region, and S as long as DA. The second one (Fig. 63) shows a moderately wide DA, with its an- terior frontal margins rounded, more devel- oped sclerotization of the frontal region with tooth-like projection hardly noticeable, and S markedly shorter than DA. Female specimens from the Selvagens Islands as well as a single specimen from northeastern Lanzarote fit the first type, while the second one is spread over the remaining localities. Spination variability in Table 4. Additional material examined. — Alegranza: unknown locality, 3rd week March 1995, Id, (P. Oromi, #2890 UB); 39, (P. Oromi, #2891, 4109, 4107 UB). Lanzarote: Haria: Malpais de la Co- rona, Charcos de marea, 25 February 1995, Id, (Amedo, Ribera & Oronif, 2887 UB). Montanas de Famara, around Mirador de Haifa; 22 February 1995, Id, (Amedo, Ribera & Oromi, 2866 UB); 1 $ , (Amedo, Ribera & Oromi, 4087 UB). Montan- as de Famara, around Mirador del Rfo, 22 Febmary 1995; 6d, (Amedo, Ribera & Oromi, #2861, 2863, 4072-3 UL, 4075, 4077 UB); 79, (Amedo, Ribera & Oromi, #2857, 2860, 2862 UL, 2936, 4082, 4084-5 UB). Yaiza\ Montanas de Femes, Atalaya de Femes, Id, 22 Febmary 1995, (Amedo, Ribera & Oromi, 2868 UB); 1 9 , (Amedo, Ribera & Oromi, 2867 UB). Montana Clara: La Caldera, 23 Febmary 1995, 4d, (Amedo, Ribera & Oromf, #2873, 2878, 2888-9 UB); 89, (Amedo, Ribera & Oromf, #2818, 2874, 2876, 2879, 2880, 4093-95 UB). Ilhas Selvagens: 3d, Id subad., 19, ljuv.; label states: ‘‘'Dysdera verneaui Simon, Grant coll.”; #BM1897.10.18.41-46 BMNH. Distribution. — This species is spread over Lanzarote, the northern islets and the Selva- gens Islands. Comments. — The male type material of this species, which is the only type known since the females were found to be a wrong identification (Amedo & Ribera 1999), has been lost. Comments for D. liostethus are equally appliable to this species. Due to the taxonomic confusion that has surrounded D. nesiotes, and according to article 75 of the ICZN (4'*^ edition), a neotype was designated. The neotype was selected from a series of specimens studied by Simon (a label stating this is included in the specimen’s vial), the original author of D. nesiotes. Simon identi- fied these specimens as D. wollastoni sensu Kulczynski 1899, which was, subsequently, considered a junior synonym of D. nesiotes (Denis 1963). The locality of the neotypes does not match the original type locality. Howevere, the last is considered to be doubt- ful (see discussion below). Before the present study, it was suggested that D. nesiotes was present in the Canarian islands of La Palma and Tenerife and in the Selvagens Islands, a group of three islets lo- cated between Madeira and the Canaries about 150 km north of Tenerife. Nevertheless, no specimens assigned to this species have ever ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 281 been reported from La Palma or Tenerife after the original description (Simon 1907). The supposed presence of this species in Tenerife could be explained by a misidentification. Si- mon transferred three females, originally as- signed to Dysdera insulana Simon 1883, to D. nesiotes. After examination of these fe- males by us, they turned out to belong to the species Dysdera propinqua Ribera, Ferrandez & Blasco 1986 (Amedo & Ribera 1999). The latter species is widely distributed in Tenerife. Probably, this locality was erroneously as- signed after misidentification of additional la- beled female material. The presence of D. ne- siotes in La Palma is even more difficult to explain. However, other cases of possible wrongly assigned localities have been pro- posed in other Canarian Dysdera described by Simon, e.g., the presence of D. insulana in La Palma and Lanzarote (Arnedo & Ribera 1997). Moreover, the geographical distribution of certain morphological characters (e.g., the presence of LA and F is restricted to endemic species from the eastern Canaries), give sup- port to the absence of D. nesiotes from the western and central Canaries. Dysdera wollastoni Blackwall 1864 was considered a junior synonym of D. crocota by Denis (1963), based on revision of the type material. Recently, Wunderlich (1991) has re- jected this synonymy based solely on the fact that D. crocota is so well known that it would be unlikely that a trained arachnologist would commit such nustake. However, Wunderlich himself has described a new Canarian ende- mism that was subsequently synonymized with D. crocota (Amedo & Ribera 1999). In any case, BlackwalTs actual description cor- responds to D. crocota. Kulczynski (1899) published a thorough and nicely illustrated re- description of what he wrongly identified as D. wollastoni, based on specimens also col- lected in the Selvagens. This redescription was similar to Simon’s original description of D. nesiotes (Simon 1883) to such an extent that Simon subsequently considered D. nesi- otes as a subspecies of D. wollastoni (Simon 1912). Lately, Denis (1963) claimed to have found no morphological evidence to justify a subspecies status for the Canarian specimens, and, because Kulczynski’s redescription was based on a wrong identification, D. nesiotes was the senior synonym. Recently, Wunder- lich (1991) has considered that the synonymy of D. wollastoni and D. nesiotes is also based on a misidentification. Nevertheless, we have been unable to find any diagnostic difference between the studied populations of D. nesiotes from the Selvagens Islands and those from the eastern Canaries, and thus we consider them as allopatric populations of the same species. Dysdera sanborondon new species Figs. 70-75, 76-79, 80, 81 Types.— Holotype male from Montanas de Tegu, Betancuria, Fuerteventura; 18 Febmary 1995, (Amedo, Ribera & Oromi, #2850 UB). Paratype female from Cuchillos de Jacomar, between Valle de Jacomar and Valle de los Toneles, Tuineje, Fuerteventura; 19 Febmary 1995, (Amedo, Ribera & Oromi; #2852 UB). Etymology. — The name in apposition of this species refers to San Borondon, the fas- tasy island that the first Spanish settlers of the 15th and 16th centuries believed they saw from the Canaries on extremely clear days. Diagnosis.— Very small Dysdera (carapace length < 3). Even though this species shows a similar genitalic pattern to the remaining eastern species (with the exception of D. lan- cerotensis) both sexes can be easily distin- guished by its smaller size and lack of lan- ceolate abdominal hairs. Males have neither lateral sheet apophysis (LA) (Fig. 76) nor ad- ditional lateral sheet (AL) (Fig. 78), and in females, the posterior region of the ventral arch (VA) is markedly sclerotized (Fig. 74). Description. — Male holotype: (Figs. 70- 72, 76-79). Carapace (Fig. 70) 2.33 long; maximum width 1.72; minimum width 1.12. Uniformly dark brownish-orange, heavily wrinkled, foveate, covered with small black grains. Frontal border roughly triangular, from 5-f carapace length; anterior lateral borders convergent; rounded at maximum dorsal width point, back lateral borders straight; back margin narrow, straight. AME diameter 0.16; PLE 0.14; PME 0.11; AME on edge of frontal border, separated from one another by less than 14 diameter, close to PLE; PME very close to each other, less than Va PME diameter from PLE. Labium trapezoid- shaped, base wider than distal part; as long as wide at base; semicircular groove at tip. Sternum dark or- ange, uniformly distributed; wrinkled; uni- formly covered in slender black hairs. Chelicerae (Fig. 71) 1.09 long, about V3 of carapace length in dorsal view; fang medium- 282 THE JOURNAL OF ARACHNOLOGY Figures 70-75. — Dysdera sanborondon new species. 70, Carapace, dorsal; 71, Left chelicera, ventral; 72, Right male bulbus, external. 73-75, Vulva anterior diverticle: 73, Dorsal; 74, Ventral; 75, Lateral. Scale bars in mm. sized, 0.74; basal segment dorsal, ventral side completely covered with piligerous granula- tions. Chelicera inner groove medium- size, about f cheliceral length; armed with three teeth and lamina at base; D = B > M (simi- lar); D triangular, located roughly at center of groove; B close to basal lamina; M at middle of B and D. Legs orange. Lengths of male described above: fel 1.86; pal 1.16; til 1.54; mel 1.44; tal 0.42; total 6.42; fe2 1.54; pa2 1.02; ti2 1.35; me2 1.35; ta2 0.42; total 5.68; fe3 1.26; pa3 0.7; ti3 0.84; me3 1.12; ta3 0.32; total 4.24; fe4 1.77; pa4 0.88; ti4 1.4; me4 1.63; ta4 0.42; total 6.1; relative length: 1-4- 2-3; fe palp 0.93; pa palp 0.51; ti palp 0.46; ta palp 0.56; total 2.46. Spination: palp, legl, leg2 spineless. Fe3d spineless; ti3d spines ar- ranged in two bands: proximal 1.0.1; distal 1.0.0; ti3v spines arranged in one band: prox- imal 1.0.0; with two terminal spines. Fe4d spines in two rows: anterior 1; posterior 2; ti4d spines arranged in two bands: proximal 1.0.1; distal 0.0.1; ti4v spines arranged in one band: proximal 1.0.1; with two terminal spines. Dorsal side of frontal legs covered with small piligerous grains; ventral side cov- ered with hairs, lacking grains. Claws with 8 teeth or less; hardly larger than claw width. Abdomen 2.4 long; whitish. Abdominal dorsal hairs 0.027 long; medium thickness, roughly straight, not compressed, blunt, tip not en- larged; uniformly, thickly distributed. Male copulatory bulbus (Fig. 72) T slightly smaller than DD; external distal border straight; internal sloped backwards. DD bent about 45° in lateral view; internal distal border markedly expanded. ES wider, more sclero- tized than IS; IS continuous to tip (slim). DD tip (Fig. 76-78) straight in lateral view. C pre- sent, long; distal border rounded, smooth. ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 283 Figures 76-8 L — Dysdera sanborondon new species, right male bulbus. 76, DD external; 77, DD fron- tal; 78, DD posterior; 79, P internal. 80-81, Dysdera sanborondon new species, spinnerets. 80, Right ALS; 81, Right PMS (lower) and PLS (upper). slightly expanded, perpendicular to DD. LF absent. L well-developed; external border sclerotized, not folded, distally projected; dis- tal border divergent, continuous. LA absent. F present, distally curved to external side, not fused to DD. AL absent. P (Fig. 79) fused to T; markedly sloped on its proximal part, per- pendicular on distal; lateral length from of T width; ridge present, perpendicular to T; not expanded; upper margin markedly toothed, on its distal part, few teeth (4-6); not distally projected; back margin not folded. Female paratype: (Figs. 80, 81). All characters as in male except: Carapace 2.79 long; maximum width 2.05; minimum width 1.35. AME diameter 0.16; PLE 0.16; PME 0.12. Chelicerae 1.3 long; fang medium- sized, 0.93; basal segment dorsal, ventral side com- pletely covered with piligerous granulations (distally slightly reduced). B > D — M (sim- ilar). Legs yellow. Lengths of female de- scribed above: fel 2; pal 1.35; til 1.68; mel 1.68; tal 0.39; total 7.1; fe2 1.77; pa2 1.26; 284 THE JOURNAL OF ARACHNOLOGY ti2 1.63; me2 1.63; ta2 0.42; total 6.71; fe3 1.49; pa3 0.84; ti3 0.98; me3 1.35; ta3 0.42; total 5.08; fe4 2.1; pa4 1.02; ti4 1.49; me4 1.86; ta4 0.42; total 6.89; relative length 1-4“ 2-3; fe palp 1.12; pa palp 0.6; ti palp 0.51; ta palp 0.7; total 2.93. Spination: palp, legl, leg2 spineless. Fe3d spineless; ti3d spines arranged in two bands: proximal 1.0.1; distal 1.0.0; ti3v spines arranged in one band: proximal 1.0.1; with two terminal spines. Fe4d spines in one row: 2; ti4d spines arranged in two bands: proximal 0.0.1; distal 0.0.1; ti4v spines ar- ranged in two bands: proximal 1.0.1; medial- proximal 0.0.1; with two terminal spines. Ab- domen 6.8 long; whitish; cylindrical. Abdominal dorsal hairs 0.063 long; thin, curved, compressed, pointed; uniformly, thickly distributed. Vulva (Fig. 73-75) DA not distinguishable from VA; rectangular, pointed expansion at middle frontal part; DA slightly wider than long; DF wide in dorsal view. MF margins not fused, well-developed, completely sclerotized. VA frontal region completely sclerotized; pos- terior region sclerotized at lateral margins; AVD absent. S attachment projected under VA; arms as long as DA, straight; tips not projected; neck as wide as arms. TB usual shape. ALS (Fig. 80) with PS; remaining pir- iform spigots more external than MS, ar- ranged in one row; 4+1 piriform gland spig- ots; PMS, PLS (Fig. 81) with 5-10 aciniform gland spigots. Intraspecific variation. — Unknown. Distribution. — Endemic species from cen- tral Fuerteventura. Dysdera spinidorsum Wunderlich 1991 Figs. 82-87, 88-91, 92, 93 Dysdera spinidorsum Wunderlich 1991: 307-308, figs. 101-102 [9]. (Holotype female; NE road to Betancuria (550 m), Betancuria, Fuerteventura; 5 January 1990; H. Enghoff, M. Baez, 9, leg.; #307, stored at ZMK; examined.) Diagnosis. — Both sexes of this species can be distinguished from the sympatric D. san- borondon by its larger size (carapace length > 4) and lanceolate abdominal hairs. Males display both a well-developed lateral sheet apophysis (LA) (Fig. 88) and additional lateral sheet (AL) (Fig. 90). In females, the posterior region of the ventral arch (VA) is membra- nous (Fig. 86). Males differ from the morpho- logically closely related D. nesiotes by having the tegulum (T) longer than the distal division (DD) (Fig. 84), the lateral sheet apophysis (LA) being frontally projected (Fig. 88) and the crest (C) distinctly expanded (Fig. 89). In females, the medial fold (MF) is markedly projected backwards (Fig. 85) and the ventral arch (VA) shows a reduction of its ventral sclerotization (Fig. 86). Description. — Male: (Figs. 82-84, 88-91). Carapace (Fig. 82) 4.9 long; maximum width 3.64; minimum width 2.59. Reddish-orange, frontally darker, becoming lighter towards back; slightly foveate at borders, slightly wrinkled with small black grains mainly at front. Frontal border roughly triangular, from 5-f carapace length; anterior lateral borders convergent; pointed at maximum dorsal width, back lateral borders straight; back mar- gin wide, straight. AME diameter 0.23; PLE 0.22; PME 0.17; AME on edge of frontal bor- der, separated from one another by about 36 diameter, close to PLE; PME very close to each other, about Vs PME diameter from PLE. Labium trapezoid-shaped, base wider than distal part; longer than wide at base; semicir- cular groove at tip. Sternum orange-yellow, frontally darker, becoming lighter towards back; very slightly wrinkled, mainly between legs and frontal border; uniformly covered in slender black hairs. Chelicerae (Fig. 83) 2.1 long, about % of carapace length in dorsal view; fang medium- sized, 1.4; basal segment dorsal, ventral side completely covered with piligerous granula- tions. Chelicera inner groove medium-size, about % cheliceral length; armed with three teeth and lamina at base; B > D = M (simi- lar); D round, located roughly at center of groove; B close to basal lamina; M at middle of B and D. Legs yellow. Lengths of male described above: fel 3.82; pal 2.56; til 3.82; mel 3.49; tal 0.74; total 14.43; fe2 3.49; pa2 2.37; ti2 3.35; me2 3.21; ta2 0.84; total 13.26; fe3 2.7; pa3 1.58; ti3 1.86; me3 2.51; ta3 0.74; total 9.39; fe4 3.45; pa4 1.96; ti4 2.65; me4 3.4; ta4 0.79; total 12.25; relative length: 1-2- 4-3; fe palp 2.1; pa palp 1.16; ti palp 0.93; ta palp 1.12; total 5.31. Spination: palp, legl, leg2 spineless. Fe3d spineless; ti3d spines ar- ranged in two bands: proximal 1.2.1; distal 1.0.1; ti3v spines arranged in two bands: prox- imal 1.0. 1-0; distal 1 -0.0.0; with two terminal spines. Fe4d spines in two rows: anterior 1; posterior 5; ti4d spines arranged in two bands: ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 285 Figures 82-87. — Dysdera spinidorsum. 82, Carapace, dorsal; 83, Left chelicera, ventral; 84, Right male bulbus, external; 85-87, Vulva anterior diverticle: 85, Dorsal; 86, Ventral; 87, Lateral. Scale bars in mm. proximal 0.0.1; distal 0.0.1; ti4v spines ar- ranged in two bands: proximal 0~ 1.0.1; distal 0-1. 0.0; with two terminal spines. Dorsal side of frontal legs covered with small piligerous grains; ventral side covered with hairs, lacking grains; very long hairs on back legs as well as on pedipalps. Claws with 8 teeth or less; hardly larger than claw width. Abdomen 4.9 long; cream-colored; cylindrical. Abdominal dorsal hairs 0.2 long; thick, roughly straight, compressed, lanceolate; uniformly, thickly distributed. Male copulatory bulbus (Fig. 84) T slightly longer than DD; external, internal distal bor- der sloped backwards. DD bent about 45° in lateral view; internal distal border markedly expanded. ES wider, more sclerotized than IS (slightly); IS continuous to tip (diffuse). DD tip (Fig. 88-90) straight in lateral view; fron- tal (upper) sheet internal part markedly pro- jected above posterior (lower) sheet. C pre- sent, long; distal end beside DD internal tip; distal border rounded, smooth, markedly ex- panded, perpendicular to DD. LF absent. L well-developed; external border sclerotized, laterally markedly folded; distal border diver- gent, continuous. LA present, sheet-like; as long as L, completely fused. F present, tip di- vided, proximally fused to DD, AL present, well-developed, joined to flagellum; proximal border in posterior view smooth, not fused with distal haematodoca, P (Fig. 91) fused to T; perpendicular to T in lateral view; lateral length from of T width; ridge present, perpendicular to T; distinctly expanded, rounded; upper margin markedly toothed, along its extent, numerous teeth (more than 10); not distally projected; back margin not folded. Female holotype: (Figs. 85-87, 92, 93). All characters as in male except: Carapace 4.9 long; maximum width 3.85; minimum width 2.73. AME diameter 0.29; PLE 0.23; PME 0.2; PME less than V4 PME diameter from PLE. Chelicerae 2.38 long, about % of carapace length in dorsal view; fang medium-sized, 1.47. Lengths of female described above: fel 3.63; pal 2.56; til 3.17; mel 3.08; tal 0.7; total 13.14; fe2 3.4; pa2 2.42; ti2 3.12; me2 2.8; ta2 0.7; total 12.44; fe3 2.84; pa3 1.63; ti3 1.77; me3 2.66; ta3 0.74; total 9.64; fe4 3.77; pa4 2.1; ti4 2.8; me4 3.45; ta4 0.84; total 12.96; relative length 1 -4-2-3; fe palp 2.23; pa palp 1.16; ti palp 0.93; ta palp 1.3; total 5.62. Spination: palp, legl, leg2 spineless. Fe3d spineless; ti3d spines arranged in two bands: proximal 1.2.1; distal 1.0.1.; ti3v spines ar- 286 THE JOURNAL OF ARACHNOLOGY Figures 88-93. — Dysdera spinidorsum, right male bulbus. 88, DD frontal; 89, DD external; 90, DD posterior; 91, P internal. 92-93, Dysdera spinidorsum, spinnerets. 92, Right ALS; 93, Right PLS. ranged in two bands: proximal 1.0.0; distal 1.0.0; with two terminal spines. Fe4d spines in two rows: anterior 2; posterior 6; ti4d spines arranged in two bands: proximal 0.0.1; distal 0.0.1; ti4v spines arranged in two bands: proximal 1.0.1; distal 1.0.1; with two terminal spines. Abdomen 5.88 long; whitish; cylindri- cal. Abdominal dorsal hairs 0.37 long; thick, roughly straight, compressed, lanceolate; uni- formly, thickly distributed. Vulva (Figs. 85- 87) DA not distinguishable from VA; rectan- gular; DA twice as wide as long; DF wide in dorsal view. MF well-developed, completely sclerotized, projected backwards, longer than DA lateral length. VA frontal region com- pletely sclerotized; posterior region sclero- tized at most anterior area; AVD absent. S at- tached to membranous VA; arms as long as DA, clearly curved; ends projected forwards; neck hardly visible. TB usual shape. ALS (Fig. 92) with PS; remaining piriform spigots more external than MS, arranged in two rows; 11 -I- 1 piriform gland spigots; PMS, PLS (Fig. 93) with 10-15 aciniform gland spigots. Intraspecific variation. — Male cephalo- thorax ranges in length from 4.41-4.69, fe- ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 287 Table 5. — Intraspecific spination variability of Dysdera spinidorsum. Proximal Medial-proximal Medial-distal Distal Tibia 3 dorsal 1. 1-2.1 0 0 1.0.1 Tibia 4 dorsal O-l.O-l.l 0 0 0.0.1 Tibia 3 ventral LO.0-1 0 0 O-l.O.O Tibia 4 ventral l.O.O-l 0 0 O-l.O.O Number of rows Number of spines Femur 3 dorsal 0 0 Femur 4 dorsal 2 1-2/4-8 male from 4.55-5.67. PLE-PME from i PME diameter to Cheliceral granulations distally reduced in some females. Spination variability in Table 5. Additional material examined. — Fuerteventu- ra: Antigua: Montanas de Tegu, road Antigua-Be- tancuria; N slope, 18 February 1995, 26, (Amedo, Ribera & Oronif, #2841 UL, 2842 (description) UB); 69, (Amedo, Ribera & Oromi, #2843-45 UL, 2849, 4066 UB); 10 juv., (Amedo, Ribera & Oromi, #2846-48 UL, 4059-65 UB). Betancuria: Betancu- ria, around village, 18 Febmary 1995, 29, (Amedo, Ribera & Oromi, #2851, 4067 UB). La Oliva: N. of La Oliva (175 m), 6 January 1990, 1 juv., (H, Enghoff & M. Baez, 2668 ZMK). Puerto del Ro- sario: La Matilla, near village, 20 Febmary 1995, 29, (Amedo, Ribera & Oromi, #2854, 4069); 1 juv., (Amedo, Ribera & Oromi, #4070 UB). From Montana Muda to La Matilla, 6 January 1990, 1 juv., (H. Enghoff & M. Baez, 2664 ZMK). Tuineje: Cuchillos de Jacomar, between Jacomar and Tone- les Valley 19 Febmary 1995, 2 juv., (Amedo, Ri- bera & Oromi, #2852, 4068 UB). Distribution. — Endemic species from cen- tral and northern Fuerteventura. When the log-transformed number of Dys- Table 6. — Number of endemics and the estimated age of each Canarian island. The values in brackets in the island age column are the ones used in the regressions. Island Num. of endemics Island age (Mya) Fuerteventura (F) 4 20-22 (21) Lanzarote (L) 3 15-19 (17) Gran Canaria (C) 10 14-16 (15) Tenerife (T) 21 11.6-14 (13) La Gomera (G) 8 10-12 (11) La Palma (P) 3 1.6-2 (1.8) El Hierro (H) 2 0.8-1 (0.9) dera species in each Canarian island is plotted against the log-transformed island age, a clear linear relationship is observed (Fig. 94). How- ever, the statistical regression obtained was very poor (r^ = 0.317). When a 95% confi- dence interval was considered, two islands seemed to depart from the general trend: Ten- erife had more species than expected while Lanzarote was poorer in species than expect- ed. Removing both values did not result in a much better fit for the regression (r^ = 0.620). Nevertheless, when the two eastern islands were removed from the analysis (Fig. 95), the relationship was markedly improved (r^ = 0.866). DISCUSSION Morphological characters, and male geni- talia in particular, suggest that with the excep- tion of D. lancerotensis, the eastern endemic species are very closely related. Putative syn- apomorphies of this group include: the pres- ence in the male bulb of a well-developed crest (C) along the internal margin of the an- terior distal half region of the distal division, the presence of a lateral sheet apophysis and the presence of thick, lanceolate abdominal dorsal hairs. D. sanborondon only shares the first of these characters, which may indicate that it is basal. The presence of a flagellum in the male bulb, found in all but D. alegran- zaensis, can also be found in some continental taxa, and is therefore likely to be a plesiom- orphic state. Several characters support a sis- ter-species relationship between D. nesiotes and D. spinidorsum: the enlargement of the internal margin of the anterior distal region of the distal division of the male bulb, and the posterior projection of the major fold (ME) lateral margins of the vulva. Male bulb characters such as the presence 288 THE JOURNAL OF ARACHNOLOGY Log island age Figures 94, 95. — Regression plots of the log-transformed number of endemics against log-transformed island ages (upper and lower curves representing 95% confidence interval). 94, all the islands included in the regression; 95, Eastern islands excluded from the regression, (r^ = regression coefficient, m = slope, b = constant). Abbreviations as in Table 6. of the medial apophysis (MA), an arch-like ridge at the distal division anterior distal tip, and a free posterior apophysis, ally the eastern endemic D. lancerotensis to the “crocota” group of species proposed by Deeleman-Rein- hold & Deeleman (1988), which are distrib- uted through southwestern Europe and north- ern Africa. Therefore, D. lancerotensis could be the result of an independent colonization event of the eastern islands. A striking feature of Canarian Dysdera is that their distribution areas largely overlap (Arnedo et al. 1996; Amedo & Ribera 1997). This pattern is also present in the eastern is- lands. In Fuerteventura, D. spinidorsum is found in the two known localities of D. san- borondon, while in Lanzarote and the northern islets D. alegranzaensis and D. nesiotes co- occurred in most of the collection localities. Although clear size segregation exists be- tween D. sanborondon and D. spinidorsum, no morphological differentiation that would seem relevant to ecological differentiation is found between D. alegranzaensis and D ne- siotes. The species D. longa from Fuerteven- tura presents the only case of intra-island al- lopatric distribution, as it is restricted to the peninsula of Jandia. Several examples of al- lopatric distributions in arthropods and slugs between Jandia and the remaining regions of Fuerteventura have also been reported (Hut- terer 1989; Juan et al. 1998). Finally, D. lan- cerotensis is spread throughout the eastern is- lands and is sympatric with all other eastern endemics. While the sympatric distribution of D. lan- cerotensis may have secondarily resulted from an independent colonization, this does not seem to be the case for the remaining species. As described above, the eastern Canaries are part of a single volcanic ridge parallel to the African coast. The depths between the islands are small (less than 40 m between Fuerteven- tura, Lanzarote and the islets) and it is prob- able that all were connected several times dur- ing glaciation events. This would help to explain the presence of the same three Dys- dera species inhabiting both Lanzarote and the northern islets. However, this geological scenario raises some questions regarding spe- cies distributions on the two main islands. Only one of the six eastern endemic species is shared between Lanzarote and Fuerteven- tura. This pattern could suggest several rounds of species exchange, with some recent enough to preclude morphological differentiation D. lancerotensis provides an example. In con- trast, the allopatric distribution of D. nesiotes- D. spinidorsum sister species pair suggests a more ancient vicariance event. The presence of D. nesiotes in the Selva- gens Islands is also unusual. Other examples ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 289 of shared species between Lanzarote-northem islets and the Selvagens are also known: the spiders Oecobius lampeli (Araneae, Oecobi- idae) and Ozyptila atlantica (Araneae, Thom- isidae) have been reported from the Selvagens and the eastern Canaries (Wunderlich 1991); the beetle Macrocoma oromiana (Coleoptera, Chrysomelidae) can be found both in Selva- gens and Alegranza; the genus Ifnidius (Co- leoptera, Malachiidae) includes one species from Ifni, one species from Lanzarote-Ale- granza and one species from the Selvagens; and the Selvagens endemic Cardiophorus oromii (Coleoptera, Elateridae) has its closest relatives in the eastern Canaries. Although the origin of the Selvagens Islands traces back to the Oligocene, most of the present-day emer- gent lands is likely to be the result of quater- nary volcanic activity after a long period of immersion under the ocean (Bravo & Coello 1978). Therefore, D. nesiotes probably colo- nized the Selvagens Islands from Lanzarote- northem islets in relatively recent times. Surprisingly, none of the material studied here could be assigned to the cosmopolitan species D. crocota. This species is widely dis- tributed in places disturbed by human activity not only in the remaining Canaries but throughout the world. The same result was found by Wunderlich (1991), who also con- sidered the only known report of this species in the eastern islands (Schmidt 1975) to be doubtful. Competitive exclusion by the pres- ence of the very closely related D. lancero- tensis may explain the absence of D. crocota in these islands. The species diversity of an oceanic island is the product of colonization and local diver- sification (Paulay 1994). The relative contri- bution of each process to the actual species number is heavily influenced by parameters like the island area, the distance from biota sources and the geological age. Plots of num- ber of species, after a substantial improvement of taxonomic knowledge, against island age indicates that the eastern islands harbor a sig- nificant lower number of endenfic Dysdera species than the rest of the Canaries. A major ecological differentiation exists between the eastern Canaries and both the central and western ones. In the Canaries, a zone of tem- perature inversion is formed at an altitude of roughly 1000 meters. This is the result of the joint effect of the humid and cool trade winds of the NE and the dry trade winds from the NW. An almost permanent cloud belt is formed in this zone. This cloud belt is the main water supply of the islands. Due to their greater age, the eastern islands have been strongly eroded and their mountains rarely reach altitudes above 800 m. This fact pre- vents these islands from capturing the clouds and the humid trade winds. Moreover, a very dry and dusty wind blows from the nearby Sahara desert. This climatic regime brings about a lower diversity of habitats in the east- ern islands compared to the central and west- ern Canaries. The low number of endemics in the eastern islands could therefore be ex- plained by extinction mainly related to the major environmental change that took place on these islands. The distribution of the east- ern endemic specimens seems to support this hypothesis. Most of the specimens were col- lected from sites located on the northern slopes of massifs over 400 m high. These places represent the wettest parts of these is- lands. The single specimen (belonging to D. nesiotes) found in a dry habitat (the sand dunes of Malpais de Corona) was captured by night. Nevertheless, Dysdera lancerotensis constitutes an exception to the rule. This spe- cies is spread over most of the island habitats, from mountain summits to lava fields, includ- ing places disturbed by human activity. An ex- tremely high level of tolerance to a wide range of environmental conditions has already been reported for a closely related species, D. cro- cota (Cooke 1968). ACKNOWLEDGMENTS For loan of material, we would like to thank E. Enghoff (ZMK), O. Escola (MZB), P.D. Hillyard (BMNH), G. Ortega (MCNT), C. Holland (MNHN) and Miguel Villana (MNCN). Gonzalo Giribet, Salvi Carranza and Andy Bohonak provided valuable com- ments on the manuscript, Ariel Fluhr translat- ed original species descriptions from German and Nuria Agusti helped with the artwork. We are also grateful to people of the Serveis Cien- tifico Teenies of the Universitat de Barcelona for their help with the SEM work. The auton- omous government of the Canaries supplied technical assistance in the expedition to Mon- tana Clara. This research was supported by projects DGICYT PB93-0811 and 2192-PGC 94A and grants FI grant from the Generalitat 290 THE JOURNAL OF ARACHNOLOGY and a “Ajut per a la finalitzacio de la tesi doctoral” of the Universitat de Barcelona (to M.A.). LITERATURE CITED Ancochea, E., J.L, Brandle, C.R. Cubas, E Heman & MJ. Huertas. 1996. Volcanic complexes in the eastern ridge of the Canary Islands: the Mio- cene activity of the island of Fuerteventura. Jour- nal of Volcanology and Geothermal Research 70: 183-204. Ancochea, E., J.M. Fuster, E. Ibarrola, A. Cendrero, J. Coello, E Heman, J.M. Cantagrel & C. Ja- mond. 1990. Volcanic evolution of the island of Tenerife (Canary Islands) in the light of the new K-Ar data. Journal of Volcanology and Geother- mal Research 44:231-249. Ancochea, E., E Heman, A. Cendrero, J.M. Can- tagrel, J.M. Fuster, E. Ibarrola & J. Coello. 1994. Constmctive and destmctive episodes in the building of a young oceanic island. La Palma, Canary Islands, and the genesis of the Caldera de Taburiente. Journal of Volcanology and Geo- thermal Research 60:243-262. Anguita, F. & F. Heman. 1975. A propagating frac- ture model versus a hot spot origin for the Ca- nary Islands. Earth and Planetary Science Letters 27:11-19. Amedo, M.A., P. Oronu & C. Ribera. 1996. Ra- diation of the genus Dysdera (Araneae, Haplo- gynae, Dysderidae) in the Canary Islands: The western Islands. Zoologica Scripta 25:241-274. Amedo, M.A. & C. Ribera. 1996, Dysdera rato- nensis Wunderlich, 1991 (Arachnida, Araneae) a troglomorphic species from La Palma, Canary Is- lands: Description of the male and redescription of the female. Revue Arachnologique 11:109- 122. Amedo, M.A. & C. Ribera. 1997. Radiation of the genus Dysdera (Araneae, Haplogynae, Dysderi- dae) in the Canary Islands: The island of Gran Canaria. Zoologica Scripta 26(3):205-243. Amedo, M.A. «fe C. Ribera. 1999. Radiation of the genus Dysdera (Araneae, Dysderidae) in the Ca- nary Islands: The island of Tenerife. Journal of Arachnology 27:604-662. Borland, L. & J. Denis. 1946. Les araignees des iles de I’Atlantique. Memoires de la Societe de Biogeographie. Paris 8:219-237. Blackwall, L. 1864. Notice of spiders indigenous to the Salvages, received from the Barao do Cas- tello de Paiva. Annals and Magazine of Natural History 14:74-180. Bohle, U., H.H. Hilger & W.E Martin. 1996. Island colonization and evolution of the insular woody habit in Echium L. (Boraginaceae). Proceedings of the National Academy of Sciences, USA 93: 11470-11745. Bravo., T. & J. Coello. 1978. Aportacion a la geo- logia y petrologia de las islas Salvajes. Pp. 15- 35, In Contribucion al estudio de la historia nat- ural de las Islas Salvajes. Santa Cmz de Tenerife: Aula de Cultura de Tenerife. Cantagrel, J.M., A. Cendrero, J.M, Fuster, E. Ibar- rola & C. Jamond. 1984. K-Ar chronology of the vulcanic emptions in the Canarian archipel- ago: Island of La Gomera. Bulletin of Volcanol- ogy 47:597-609. Carracedo, J.C., S.J. Day, H. Guillou, E. Rodriguez, J.A. Canas & EJ. Perez. 1998a. Origen y evol- ucion del volcanismo de las Islas Canarias. In Ciencia y Cultura en Canarias. OACIMC, Santa Cmz de Tenerife. 165 pp. Carracedo, J.C., S. Day, H. Guillou, E.R. Badiola, J.A. Canas & F.J.P. Torrado. 1998b. Hotspot vol- canism close to a passive continental margin: The Canary Islands. Geological Magazine 135(5):591-604. Coddington, J.A. 1990. Ontogeny and homology in the male palpus of orb-weaving spiders and their relatives, with comments on phylogeny (Ar- aneoclada: Araneoidea, Deinopoidea). Smithson- ian Contributions to Zoology 496:1-50. Coello, J., J.M. Cantagrel, F. Heman, J.M. Fuster, E. Ibarrola, E. Ancochea, C. Casquet, C. Jamond, J.R. Diaz de Teran & A. Cendrero. 1992. Evo- lution of the eastern volcanic ridge of the Canary Islands based on new K-Ar data. Journal of Vol- canology and Geothermal Reseach 53:251-274, Cooke, J.A.L. 1968. Factors affecting the distri- bution of some spiders of the genus Dysdera (Ar- aneae, Dysderidae). Entomologist’s Monthly Magazine 103:221-225. Dallwitz, M.J. 1980. A general system for coding taxonomic descriptions. Taxon 29:41-46. Dallwitz, M.J., T.A. Paine & E.J. Zurcher, 1993. User’s guide to the DELTA system: A general sytem for processing taxonomic descriptions, v.4.01. CSIRO. Canberra. Deeleman-Reinhold, C. & Deeleman, P.R. 1988. Revision des Dysderinae. Tijdschrift Entomolo- gie 131:141-269. Denis, J. 1941. Les araignees des lies Canaries. Annales de la Societe Entomologique de France 110:105-130. Denis, J. 1962. Les araignees de I’Archipel de Madere. Publ^oes do Institute de Zoologica -Dr. Auguste Nobre 79:9-118. Denis, J. 1963. Spiders from the Madeira and Sel- vage Islands. Boletim do Museu Municipal do Funchal 17:29-48. Francisco Ortega, J,, R.K. Jansen & A. Santos Guerra. 1996. Chloroplast DNA evidence of colonization, adaptive radiation, and hybridiza- tion in the evolution of the Macaronesian flora. Proceedings of the National Academy of Scienc- es USA 93:4085-4090. Fuster, J.M., F. Heman, A. Cendrero, J. Coello, J.M. ARNEDO ET AL.— THE GENUS DYSDERA IN THE EASTERN CANARIES 291 Cantagrel, E. Ancochea & E. Ibarrola. 1993. Geocronologia de la Isla de El Hierro (Islas Can- arias). Boletin de la Real Sociedad Espanola de Historia Natural Seccion Geologica 88:85-97. Gonzalez, R, E Pinto, M. Nogales, J. Jimenez- Asensio, M. Hernandez & V.M. Cabrera. 1996. Phylogenetic relationships of the Canary Islands endemic lizards genus Gallotia (Sauria: Lacerti- dae). Molecular Phylogenetics and Evolution 6: 63-71. Hutterer, R. 1989. Recent and fossil slugs of the genus Parmacelia in the Canary Islands, with the description of three new species. Archiv fuer Molluskenkunde 120:73-93. Juan, C., P. Oromi & G.M. Hewitt. 1995. Mito- chondrial DNA phylogeny and sequential colo- nization of Canary Islands by darkling beetles of the genus Pimelia (Tenebrionidae). Proceedings of the Royal Society of London, B Biological Sciences 261:173-180. Juan, C., P Oromi & G.M. Hewitt. 1996a. Phylog- eny of the genus Hegeter (Tenebrionidae, Cole- optera) and its colonization of the Canary Islands deduced from cytochrome oxidase I mitochon- drial DNA sequences. Heredity 76:392-403. Juan, C., K.M. Ibrahim, P Oromi & G.M. Hewitt. 1996b. Mitochondrial DNA sequence variation and phylogeography of Pimelia darkling beetles on the Island of Tenerife (Canary Islands). He- redity 77:589-598. Juan, C., K.M. Ibrahim, P. Oromi & G.M. Hewitt. 1998. The phylogeography of the darkling beetle Hegeter politus in the eastern Canary Islands. Proceedings of the Royal Society of London B Biological Sciences 265:135-140. Kim, S.C., D.J. Crawford, J. Francisco Ortega & A. Santos Guerra. 1996. A common origin for woody Sonchus and five related genera in the Macaronesian islands: Molecular evidence for extensive radiation. Proceedings of the National Academy of Sciences USA 93:7743-7748. Kraus, O. 1978. Liphistius and the evolution of spi- der genitalia. Arachnology; Symposia of the Zoological Society of London 42:235-254. Kulczynski, W. 1899. Arachnoidea opera Rev. E. Schmitz collecta in insulis Maderianis et un in- sulis Selvages dictis. Bulletin International de I’Academie des Sciences de Cracovie. pp. 36-37. Latreille, P.A. 1804. Tableau methodique des insec- tes. In Dictionnaire (Noveau) d’histoire Naturel- le, Paris 24:129-200. Mcheidze, TS. 1972. Novije Vide paukov roda Harpactocrates (Dysderidae). Bulletin of the Academy of Sciences of Georgia (SSR) 68:741- 743. Mes, T.H.M. & H. T’Hart. 1996. The evolution of growth-forms in the Macaronesian genus Aeon- ium (Crassulaceae) inferred from chloroplast DNA RFLPs and morphology. Molecular Ecol- ogy 5:351-363. Mitchell-Thome, R.C. 1985. Radiometric studies in Macaronesia. Boletim do Museu Municipal do Funchal 37:52-85. Paulay, G. 1994. Biodiversity on oceanic islands: Its origin and extinction. American Zoologist 34: 134-144. Platnick, N.I., J.A. Coddington, R.R. Forster & C.E. Griswold. 1991. Spinneret morphology and the phylogeny of haplogynae spiders (Araneae, Ar- aneomorpha). American Museum Novitates 3016:1-73. Rambla, M. 1978. Aracnidos de las Islas Salvajes (Opiliones y Araneidos). Pp. 129-137, In His- toria Natural de las Islas Salvajes. Tenerife: Aula de Cultura de Tenerife. Rando, J.C., E. Hernandez, M. Lopez & A.M. Gon- zalez. 1997. Phylogenetic relationships of the Canary Islands endemic lizard genus Gallotia in- ferred from mitochondrial DNA sequences: In- corporation of a new subspecies. Molecular Phy- logenetics and Evolution 8:114-116. Reimoser, E. 1919. Katalog der echten Spinnen (Araneae) des Palaarktischen Gebietes. Abhan- dlungen der Zoologisch-botanischen Gesellschaft in Wien 10:1-180. Ribera, C. & M.A. Amedo. 1994. Description of Dysdera gollumi (Araneae, Haplogynae), a new troglobitic species from Tenerife, Canary Islands, with some comments on Canarian Dysdera. Me- moires de Biopeologie 21:115-119. Ribera, C., M.A. Ferrandez & A. Blasco. 1985. Ar- aneidos cavemicolas de Canarias II. Memoires de Biospeologie 12:51-66. Ribera, C. & M.A. Ferrandez. 1986. Tres nuevas especies de Disderidos (Arachnida: Araneae) cavemicolas de la Peninsula Iberica. Publica- ciones del Departamento de Zoologica (Barce- lona) 12:51-58. Sauer, E.G.F. & P. Rothe. 1972. Ratite eggshells from Lanzarote. Canary Islands. Science 176:43- 45. Schmidt, G. 1973. Zur Spinnenfauna von Gran Ca- naria. Zoologische Beitraege 19:347-392. Schmidt, G. 1975. Zur Spinnenfauna von Lanza- rote (Kanarischen Inseln). Zoologische Beitraege 21:239-245. Schult, J. 1980. Die Genitalstmkturen haplogyner Araneae unter phylogenetischem Aspekt (Arach- nida). D. Phil. Thesis, Hamburg. Schult, J. 1983. Taster haplogyner Spinnen unter phylogenetischem Aspekt (Arachnida: Araneae). Abhandlungen des Naturwissenschaftlichen Ver- eins in Hamburg 26:69-84. Simon, E. 1883. Etudes Arachnologiques XIV Me., materiaux pour servir a la faune arachnol- ogique des lies de P Ocean Atlantique. Annales 292 THE JOURNAL OF ARACHNOLOGY de la Societe Entomologique de France 6:294- 298. Simon, E. 1907. Etude sur les Araignees de la sous-section des Haplogynes. Annales de la So- ciete Entomologique de Belgique 51:246-264. Simon, E. 1912. Arachnides recueillis par M.L. Garreta a I’ile Grande-Salvage. Bulletin de la So- ciete Entomologique de France 2:59-61. Thorpe, R.S., D.R McGregor, A.M. Gumming & W.C. Jordan. 1994. DNA evolution and coloni- zation sequence of island lizards in relation to geological history: mtDNA RFLP, cytochrome B, cytochrome oxidase, 12S rRNA sequence, and nuclear RAPD analysis. Evolution 48:230-240. Thorpe, R.S., A. Malhotra, H. Black, J.C. Daltry & W. Wuester. 1995. Relating geographic pattern to phylogenetic process. Philosophical Transac- tions of the Royal Society of London B Biolog- ical Sciences 349:61-68. Wunderlich, J. 1987. Die Spinnen der Kanarischen Inseln und Madeiras. Taxonomy and Ecology 1: 1-435. Wunderlich, J. 1991. Die Spinnen-fauna der Mak- aronesischen Inseln. Beitrage zur Araneologie 1: 1-619. Wunderlich, J. 1994. Zu Okologie, Biogeographie, Evolution und Taxonomic einiger Spinnen der Makaronesischen Inseln. Beitrage zur Araneolo- gie 4:385-439. Manuscript received 25 October 1999, revised 1 June 2000. 2000. The Journal of Arachnology 28:293-299 NEW SPECIES AND RECORDS OF KLEPTOCHTHONIUS FROM INDIANA (PSEUDOSCORPIONIDA, CHTHONIIDAE) William B. Muchmore: Department of Biology, University of Rochester, Box 270211, Rochester, New York 14627=0211 USA ABSTRACT. New records and supplemental data are given for the troglobitic species Kleptochthonius packardi; and two new epigean or troglophilic species are described, K. griseomanus and K. lewisorum. Some comments are made on the status of the genus. Keywords: Pseudoscorpionida, Kleptochthonius, cavemicoles, Indiana In 1879, H. Hagen described Blothrus pack- ardi from Wyandotte Cave in Crawford Coun- ty, Indiana. This was the first cavemicolous pseudoscorpion known in North America. Re- study of the type collection (Muchmore 1963) revealed that the species belongs in the genus Kleptochthonius Chamberlin 1949. In 1994 I reported an isolated Kleptochthonius palp from Wilson’s Cave in Jefferson County, In- diana. No other material pertaining to the ge- nus has been reported from the state until re- cently. Over the past several years, intensive search by J.J. Lewis and his colleagues in caves of southern Indiana has turned up new material of K. packardi, together with several other species of pseudoscorpions. The present paper reexamines Kleptochthonius packardi and describes two new species in the genus. METHODS Most of the specimens studied here were dissected, cleared, and mounted on micro- scope slides for detailed examination. Speci- mens are deposited in the Florida State Col- lection of Arthropods, Gainesville, Florida (FSCA) and the Museum of Comparative Zo- ology, Harvard University, Cambridge, Mas- sachusetts (MCZ). Some abbreviations are used in the text: L = length; L/B = ratio, length/breadth; L/D = ratio, length/depth; T = tactile seta. SYSTEMATICS Genus Kleptochthonius Chamberlin Apochthonius (Heterochthonius) Chamberlin 1929: 153; Beier 1932:42. Heterochthonius Chamberlin: Hoff 1945:313; Hoff 1949:434. Kleptochthonius Chamberlin 1949:4; Hoff 1958:7; Malcolm & Chamberlin 1961:2-3; Muchmore 1965:1; Muchmore 1990:510; Harvey 1991:177; Muchmore 1994a: 13. Chamberlinochthonius Vachon 1952:105; Hoff 1958:7. Kleptochthonius (Chamberlinochthonius) Vachon: Malcolm & Chamberlin 1961:16; Muchmore 1965:1; Muchmore 1990:510; Harvey 1991:179. Kleptochthonius was described originally by J.C. Chamberlin (1929) as Heterochthon- ius, a subgenus of Apochthonius Chamberlin. The name Kleptochthonius was first applied by Chamberlin (1949), after he discovered that the name Heterochthonius had been used previously by Berlese (1910) for a genus of Acarina. At that time, only two species were known, K. crosbyi (Chamberlin 1929) and K. multispinosus (Hoff 1945), both epigean forms from North Carolina. In 1952, M. Va- chon erected a new, allied genus, Chamberli- nochthonius, the type species, C. henroti, be- ing a troglobitic form from a cave in West Virginia. Malcolm & Chamberlin (1961) de- scribed two new epigean species of Klepto- chthonius from Oregon and eight new trog- lomorphic species from caves in eastern states; the latter they placed in Chamberlino- chthonius, which they regarded as an “artifi- cial but convenient” subgenus of Klepto- chthonius. Subsequently, Muchmore (1963, 1965, 1976, 1994a, 1994b) and other authors (e.g., Harvey 1991) followed Malcolm & Chamberlin in assigning species to the sub- genera Kleptochthonius or Chamberlino- chthonius. However, subgeneric designations are not used in the present report, because they are currently undergoing reevaluation (see Discussion). 293 294 THE JOURNAL OF ARACHNOLOGY The genus Kleptochthonius was well de= fined by Malcolm & Chamberlin (1961). Sup- plementary discussions of specialized sensory setae on the palpal chela and of the dorsal process of the movable chelal finger were giv- en by Muchmore (1965, 1976, 1994a). Kleptochthonius packardi (Hagen) Figs. 1, 4 Blothrus packardi Hagen 1879:399. Chthonius packardi (Hagen): Hagen 1880:83-84; Hubbard 1880:37, 79, 84 (in part); Banks 1895: 13 (in part); Blatchley 1897:170; Coolidge 1908: 114 (in part); Vachon 1952:111 (in part). Chthonius packardii Hagen: Packard 1888:43-48, figs. 12a-g, PI. XI figs. 3, 3a-j (in part). Chtonius [sic] packardii Hagen: Blatchley 1897: 205 (in part). Cthonius [sic] packardi Hagen: Blatchley 1897:171. Chthonius(l) packardi (Hagen): Beier 1932:61 (in part); Roewer 1937:240 (in part); Hoff 1958:4 (in part). Genus? packardi Hagen: Hoff 1949:443 (in part). ''Chthonius'^ packardi (Hagen): Malcolm & Cham- berlin 1961:1. Kleptochthonius (Chamberlinochthonius) packardi (Hagen): Muchmore 1963:2-5, figs. 1-2; Much- more 1965:2, 7; Muchmore 1976:211; Harvey 1991:181 (in part); Muchmore 1994b:319-320. Not Chthonius packardii Hagan [sic]: Giovannoli 1933:604 (misidentification). Type material examined. — Lectotype male (No. 1), allotype female (No. 2), and 4 paralectotype males of Blothrus packardi Ha- gen [also labeled '"Chamberlinochthonius packardi (Hagen), det. W.B. Muchmore”], on slides, in MCZ. Type locality. — Wyandotte Cave, Craw- ford County, Indiana. [Note: Harvey (1991: 181) erroneously gives “Mammoth Cave, Kentucky, U.S.A.” as the type locality of K.(C.) packardi]. Diagnosis. — A large, eyeless species of Kleptochthonius with very slender appendag- es (length of palpal chela 1.3 mm or greater, chela usually 7.0 or more times as long as broad); all parts light brown or lighter; dorsal process on proximal end of movable chelal finger long, cylindrical; a short, stout sensory seta on medial side of fixed finger near base. Additional material examined. — INDIANA: Crawford County: Route 66 Cave, 6 km S of Har- dinsburg, “hand-collected from raccoon scat (with Collembola) on mudbank, end of main passage of cave about 60 meters from entrance, dry upper lev- el”, 26 October 1996 (J.J. Lewis, James Lewis, V. Lewis), 1 9 . Harrison County: Binkley Cave, about 1.5 km S of Corydon, in pitfall trap about 300 m into dark zone, 16 November 1997 (J.J. Lewis, T Sollman), 1 tritonymph; Coon Cave, 6.5 km SW of Corydon, “from rock on top of baited (cheese) pit- fall trap, dark zone”, 17 May 1997 (J.J. Lewis, S. Rafail), 1 S ; Maucks Cave, Harrison-Crawford State Forest, “from flowstone in lower level of cave”, 14 September 1996 (J.J. Lewis, James Lewis, V. Lew- is), 1 tritonymph; Twin Domes Cave, Twin Domes Nature Preserve, in pitfall traps, 31 May 1998 (J.J. Lewis), 2$. Orange County: Murray Spring Cave, Paoli Country Club, in pitfall trap, 30 April 1998 (J.J. Lewis, S. Rafail), 1$; Saltpeter Cave, 6 km NNE of Marengo, 2 March 1997 (J.J. Lewis, James Lewis, V. Lewis), 2 9 . (All on slides, in FSCA.) Supplemental data. — All parts of animals pale in color. Eyes absent in all. Chaetotaxy of carapace 4-4-4-2-4 =18, except in one fe- male where there are 3, rather than 4, setae on posterior margin. Tergal chaetotaxy somewhat variable, but usually much like the types, i.e., 2-3:2-3:2-3:2-4:4:4:5-6:6:-. Internal genitalia of male similar to those of K. crosbyi (see Malcolm & Chamberlin 1961: fig. 3A). Ap- pendages of adults very long and slender. Pal- pal femur 1. 55-1.65 X and chela 2.3-2.45 X as long as carapace; L/B of femur 6.85-7.15, pa- tella 2.35-2.55, chela 7. 1-7.9; L/D of hand 2.75-3.05; movable finger 1.55-1. 7 X as long as hand. Leg I: femur 2.25-2.5 X as long as patella. Leg IV: L/D of femur + patella 3.45- 3.85, tibia 5.25-5.7. The dorsal process on the proximal end of movable finger of chela is long, cylindrical (Fig. 1). There is a short sen- sory seta on the medial side of the fixed finger of the palpal chela, at or just distad of level of trichobothrium ist (Fig. 4). Tritonymph much like adult but smaller and with slightly less slender appendages; with only 7 trichobothria on hand and fixed chelal finger and 3 on movable finger. Short sensory seta on fixed chelal finger as in adult. Measurements (mm). — Adult: Figures given first for the single male, followed in pa- rentheses by ranges for 6 females. Body L 1.90 (1.81-2.63). Carapace L 0.605 (0.59- 0.695). Chelicera L 0.53 (0.51-0.57). Palp: trochanter 0.30 (0.28-0.32) / 0.155 (0.13- 0.17); femur 1.00 (0.97-1.11) / 0.14 (0.14- 0.155); patella 0.36 (0.355-0.39) / 0.15 (0.15- 0.16); chela 1.50 (1.42-1.60) / 0.19 (0.19- 0.22); hand 0.60 (0.55-0.63) / 0.195 (0.19- 0.22); movable finger L 0.94 (0.89-1.04). Leg MUCHMORE— CAVERNICOLOUS KLEPTOCHTHONIUS FROM INDIANA 295 Figures 1-6. — Species of Kleptochthonius. 1-3. Proximal end of movable finger of palpal chela, lateral view, showing dorsal process (dorsal at top; areole is that of trichobothrium b). 1. Kleptochthonius pack- ardi, lectotype male; 2. Kleptochthonius griseomanus new species, holotype female; 3. Kleptochthonius lewisorum new species, holotype female. 4. Kleptochthonius packardi, lectotype male: Left palpal chela, dorsal view, with enlargement of sensory seta on fixed finger (other setae omitted). 5, 6, Kleptochthonius griseomanus new species, holotype female. 5, Right palp, dorsal view, with enlargement of sensory seta on fixed finger; 6. Left chela, lateral view (base broken; setae omitted). I: femur 0.62 (0.585-0.665) / 0.095 (0.08- 0.11); patella 0.265 (0.235-0.295) / 0.09 (0.08-0.09). Leg IV: femur + patella 0.865 (0.86-0.955) / 0.245 (0.23-0.26); tibia 0.60 (0.555-0.63) / 0.105 (0.105-0.12). Tritonymph: Two specimens. Body L 1.45, 1.68. Carapace L 0.49, 0.495. Chelicera L 0.40, 0.40. Palp: femur 0.75, 0.76 / 0.125, 0.125; patella 0.28, 0.29 / 0.125, 0.13; chela 1.13, 1.14 / 0.17, 0.17; hand 0.46, 0.45 / 0.17, 0.17; movable finger L 0.69, 0.71. Leg IV: femur + patella 0.605, 0.64 / 0.16, 0.18; tibia 0.42, 0.45 / 0.09, 0.09. Remarks.— The newly collected specimens appear to be slightly larger than the types from Wyandotte Cave, as reported in 1963. However, these differences are probably due in part to changes in measuring techniques be- tween that time and the present. In any event, the species, as here recognized, is rather var- iable in size and chaetotaxy. Further collecting and study may reveal that more than one spe- cies is represented. In addition to the type locality, Wyandotte Cave, Kleptochthonius packardi has been found in several caves in neighboring Craw- 296 THE JOURNAL OF ARACHNOLOGY Figures 7, 8. — Kleptochthonius lewisorum new species, holotype female. 7. Left palp, dorsal view but chela twisted showing medial surface, with enlargement of sensory seta on fixed finger (darkened areoles are underneath, i.e., on lateral side of chela); 8. Right chela, lateral view (setae omitted). ford, Harrison, and Orange Counties, Indiana. Two females were collected within Twin Domes Cave, Harrison County, where K. gri- seomanus new species is also present at the entrance. A tritonymph, apparently belonging to K. packardi, was taken some 300 m into the eastern end of the Binkley Cave System, the largest known cave in Indiana, while the holotype of K. lewisorum new species (see be- low) was found in the Baelz Cave section at the western end of the system (about 5.5 km away, straight-line distance) (see Lewis & Sollman 1998). Kleptochthonius{l) sp. A single, detached, left palp of an adult pseudoscorpion was collected in Wilsons Cave, Jefferson County, Indiana; it has been tentatively identified as belonging to an un- known species of Kleptochthonius (Much- more 1994b). If it is indeed a Kleptochthonius, it appears most closely related to K. sheari Muchmore (1994a), with a relatively long. sensory seta at the base of the fixed chelal finger. From the attenuation of the palp, it ap- pears to be a troglomorphic species. No other representative of the genus has been found in this part of Indiana. Kleptochthonius griseomanus new species Figs. 2, 5, 6 Type material. — Holotype female (WM8208.01001) from Indian Cave, (a sand- stone cave in the Hemlock Cliffs area of Hoo- sier National Forest), about 6.5 km SSE of Taswell, Crawford County, Indiana, 5 July 1997 (J.J. Lewis, S. Rafail); allotype male (WM8240.02001) from leaf litter at base of entrance pit. Twin Domes Cave, Twin Domes Nature Preserve, Harrison County, Indiana, 31 May 1998 (J.J. Lewis, R. Bums, E. Bums, H. Huffman, E. Jacquart) (mounted on slides, in FSCA). Diagnosis. — A smaller, less slender spe- cies, with palpal chela 1.05 mm long, 4.7- 5.05 X as long as broad; 4 comeate eyes; MUCHMORE— CAVERNICOLOUS KLEPTOCHTHONIUS FROM INDIANA 297 mostly light brown, but hand of chela distinct- ly gray; dorsal process on proximal end of movable chelal hnger small, roughly bilobed; a short, stout sensory seta on medial side of hxed finger near base. Kleptochthonius gri- seomanus appears most closely related to K. inusitatus Muchmore (1994a) from eastern Ohio. The two are similar in size and propor- tions, but K. griseomanus differs in having a distinctly gray palpal chela, fewer setae on the terga, a smaller, less strongly bilobed process on the base of the movable finger of the chela, and the small sensory seta on the fixed finger closer to the level of trichobothrium ist. Description. — Representative of Klepto- chthonius as discussed above, and with the following particular features. Male and female much alike. Hand of palpal chela gray; chelal fingers and other palpal segments, carapace and chelicera tan; other parts lighter. Carapace about as long as broad; epistome barely per- ceptible; 4 comeate eyes; chaetotaxy 6-4-4-2- 4 = 20. Coxal area typical; each coxa I with 5 coxal spines. Tergal chaetotaxy of holotype 4:4:7:6:8:9:10:9:?:?;T2T:0, allotype similar. Sternal chaetotaxy of holotype (female) 8: (3)8(3):(3)8(3):12:14:14:13:13:1 1:0:2; ster- nites 2--5 of male 13:11-10 / (3)6(3):(3)9(3): 1 1 . Internal genitalia of male similar to those of K. crosbyi (see Malcolm & Chamberlin 1961: fig. 3A). Chelicera 0.8 as long as car- apace; hand with 7 setae; flagellum of about 7 setae; galea a very low elevation. Palp (Fig. 5) long and slender; femur 1.3-1. 35 X and chela 1.9-1.95X as long as carapace. L/B of trochanter 1.8-1.85, femur 5.5-5.85, patella 2.1-2.15, and chela 4.7-5.05; L/D of hand 1.95-2.05; movable finger 1.5X as long as hand. Trichobothria as shown in Fig. 6. A short, sensory seta is present distad of tricho- bothrium ist on medial side of fixed finger (Fig. 5). Dorsal process on base of movable finger small, roughly bilobed (Fig. 2). Fixed finger of holotype with 21 tall, spaced macro- denticles and 10 very small, rounded micro- denticles alternating distally; movable finger with 11 tall, spaced macrodenticles, 6 very small alternating microdenticles, and 10 low, rounded teeth proximally. Legs rather long and slender. Leg I with femur 2.1-2.2X as long as patella. Leg IV: L/D of femur + pa- tella 2.9-3. 0, tibia 4.9-5. 1. Measurements (mm). — Figures given first for holotype female, followed in parentheses by those for allotype male. Body L 2.11 (1.87). Carapace L 0.555 (0.54). Chelicera L 0.45 (0.42). Palp: trochanter 0.235 (0.23) / 0.13 (0.125); femur 0.725 (0.73) / 0.13 (0.125); patella 0.32 (0.295) / 0.15 (0.14); chela 1.04 (1.06) / 0.22 (0.21); hand 0.43 (0.435) / 0.22 (0.21); movable finger L 0.64 (0.66). Leg I: femur 0.39 (0.415) / 0.075 (0.075); patella 0.185 (0.185) / 0.075 (0.075). Leg IV: femur + patella 0.615 (0.66) / 0.20 (0.23); tibia 0.435 (0.46) / 0.09 (0.09); basi- tarsus 0.245 (0.235) / 0.075 (0.065); telotarsus 0.415 (0.445) / 0.05 (0.05). Etymology. — The species is named griseo- manus in reference to the distinctly gray hand of the palpal chela. Remarks.- — Two specimens of K packardi were collected within Twin Domes Cave, in the entrance pit of which the allotype of K. griseomanus was found (see above). The for- mer is certainly a troglobite, whereas the latter is at best a troglophile, or an epigean species only accidentally associated with the cave. Kleptochthonius lewisorum new species Figs. 3, 7, 8 Type material.— Holotype female (WM8207.01001) from the “underside of a stone lying in leaf litter with some raccoon droppings, in the company of some troglobitic Sinella alata Christiansen (Collembola), twi- light zone,” Baelz Cave, Binkley Cave Sys- tem, Harrison County, Indiana, 28 June 1997 (J.J. Lewis, FA. Pursell) (mounted on slide, in FSCA). Diagnosis.— A medium-sized species (pal- pal chela 1.15 mm long), with moderately slender palps (chela 4.6X as long as broad); 4 eyes, posterior pair smaller than anterior pair; all parts, including palps, light brown or lighter; process on proximal end of movable finger of palpal chela small, irregularly round- ed; a moderately long sensory seta on medial side of fixed finger near middle. Description of female. — (Male unknown). Representative of the genus Kleptochthonius as discussed above, and with the following particular features. Palps very light brown, carapace and chelicerae tan, other parts ligh- ter. Carapace with epistome very small; 4 cor- neate eyes, posterior pair smaller; chaetotaxy 9-4-4-2-4 = 23. Coxal area typical of the ge- nus; each coxa I with 3 coxal spines. Tergal chaetotaxy 4:4:4:6:7:9:9:9:9:7:T2T:0.; sternal 298 THE JOURNAL OF ARACHNOLOGY chaetotaxy 8:(4)6(4):— ?— T1T2T1T:2. Che= licera 0.75 as long as carapace; hand with 7 setae; flagellum of 7-8 setae; galea a low el- evation. Palp (Fig. 7) long, slender: femur 1.3X and chela 1.95X as long as carapace. L/ B of trochanter 1.9, femur 5.05, patella 1.75, and chela 4.6; L/D of hand 1.6; movable fin- ger 1.85X as long as hand. Trichobothria as shown in Fig. 8. A sensory seta of moderate length on fixed finger near middle of medial side (Fig. 7). Dorsal process on base of mov- able finger small, irregularly rounded (Fig. 3). Fixed finger with about 20 tall, spaced macro- denticles, decreasing in size to very small proximally, and 10 moderately large micro- denticles alternating distally (in two of the in- tervals between macrodenticles there are two microdenticles rather than one); movable fin- ger with 10 tall, spaced macrodenticles, 6 moderately large alternating microdenticles, and 10 low, rounded teeth proximally. Legs moderately slender: leg I with femur 2. 1 X as long as patella; leg IV with L/D of femur + patella 2.95, of tibia 4.5. Measurements (mm). — Body L 2.47. Car- apace L 0.54. Chelicera L 0.445. Palp: tro- chanter 0.265/0.14; femur 0.755/0.15; patella 0.295/0.17; chela 1.15/0.25; hand 0.42/0.265; movable finger L 0.78. Leg I: femur 0.39/ 0.08; patella 0.185/0.08. Leg IV: femur + pa- tella 0.615/0.21; tibia 0.43/0.095; basitarsus 0.235/0.08; telotarsus 0.43/0.045. Etymology. — The species is named for Jul- ian J. Lewis and his sons, James J. Lewis and Victor M. Lewis, who for the past several years have been leading the way in studies of the invertebrate faunas in Indiana caves. Remarks. — Because of the moderately long sensory seta near the middle of the fixed chelal finger, Kleptochthonius lewisorum ap- pears related to one or more, as yet unidenti- fied, species from the southeastern states. It differs from them in size, proportions, and body chaetotaxy. Baelz Cave, the type locality, consists of a short passage in the bluff of Indian Creek near the Seven Springs resurgence of Binkley Riv- er; it is a route for floodwater overflow out of the Binkley Cave System. This is about 6 km from the site of capture of a specimen of K. packardi in the eastern end of the system (see above). DISCUSSION Like Tyrannochthonius Chamberlin 1929 (see Muchmore & Chamberlin 1995; Much- more 1996) the genus Kleptochthonius con- tains species with quite varied morphologies. On the one hand, Kleptochthonius crosbyi (type species of the genus) is a small, four- eyed, epigean chthoniid, with moderately slender palps, while K. henroti (Vachon) (type species of Chamberlinochthonius Vachon) is a large, blind, troglobitic species with very at- tenuated palps. In their revision of Klepto- chthonius, Malcolm & Chamberlin (1961) recognized the close relationship of these var- ied species by “considering Chamberlino- chthonius Vachon as an artificial but conve- nient subgenus comprising essentially the cavemicolously modified forms of Klepto- chthonius"" (p. 3). Now, when there are some 10 epigean species and over 30 cavemicolous species known in the genus, it is perfectly clear that a generic or subgeneric distinction based on size, eyes, coloration, or slenderness of appendages is not warranted. However, it does appear possible that the nature and num- ber of special sensory setae on the fixed finger of the palpal chela (Muchmore 1976, 1994a), and perhaps some other characters (Much- more 1965), will provide evidence of separate evolutionary lines within the genus. Restudy of all species in the genus is in progress. ACKNOWLEDGMENTS I am greatly indebted to the following for the new collections of pseudoscorpions from Indiana caves on which this work is based: Julian J. Lewis and his associates, Elizabeth Bums, Ronnie Bums, Henry Huffman, Ellen Jacquart, James J. Lewis, Victor M. Lewis, E Allen Pursell, Salisa T. Rafail and Thomas R Sollman. J.J. Lewis made many useful sug- gestions for the improvement of the manu- script. The study was aided in part by a grant from the Indiana Department of Natural Re- sources, Division of Nature Preserves. LITERATURE CITED Banks, N. 1895. Notes on the Pseudoscorpionida. Journal of the New York Entomological Society 3:1-13. Beier, M. 1932, Pseudoscorpionidea I. Subord. Chthoniinea et Neobisiinea. Das Tierreich 57:1- 258. Blatchley, W.B. 1897. Indiana caves and their fau- MUCHMORE— CAVERNICOLOUS KLEPTOCHTHONIUS FROM INDIANA 299 na. Annual Report, Indiana Department of Ge- ology and Natural Resources 21:121-212. Chamberlin, J.C. 1929. A synoptic classification of the false scorpions or chela-spinners, with a re- port on a cosmopolitan collection of the same. Part 1. The Heterosphyronida (Chthoniidae) (Arachnida-Chelonethida). Annals and Magazine of Natural History (10)4:50-80. Chamberlin, J.C. 1949. New and little-known false scorpions from various parts of the world (Arachnida, Chelonethida), with notes on struc- tural abnormalities in two species, American Museum Novitates 1430:1-57. Chamberlin, J.C. & D.R. Malcolm. 1960. The oc- currence of false scorpions in caves with special reference to cavemicolous adaptation and to cave species in the North American fauna (Arachnida- Chelonethida). American Midland Naturalist 64: 105-115. Coolidge, K.R. 1908. A list of the North American Pseudoscorpionida. Psyche 15:108-114. Giovannoli, L. 1933. Invertebrate life of Mammoth and other neighboring caves. American Midland Naturalist 14:600-623. Hagen, H. 1879. Hoehlen-Chelifer in Nord- Amer- ica. Zoologischer Anzeiger 2:399-400. Hagen, H. 1880. (Untitled). Pp. 83-84, In Two days’ collecting in the Mammoth Cave, with contributions to a study of its fauna. (H.G. Hub- bard). American Entomologist. Vol. 3. Harvey, M.S. 1991. Catalogue of the Pseudoscor- pionida. Manchester Univ. Press, Manchester, England. Hoff, C.C. 1945. Pseudoscorpions from North Car- olina. Transactions of the American Microscop- ical Society 64:311-327. Hoff, C.C. 1949. The pseudoscorpions of Illinois. Bulletin of the Illinois Natural History Survey 24:407-498. Hoff, C.C. 1958. List of the pseudoscorpions of North America north of Mexico. American Mu- seum Novitates 1875:1-50. Hubbard, H.G. 1880. Two days’ collecting in the Mammoth Cave, with contributions to a study of its fauna. American Entomologist 3:34-40, 79- 84. Lewis, J.J. & TP. Sollman. 1998. Groundwater monitoring in significant aquatic caves that lie beneath impending residential developments in the Blue River basin of southern Indiana. Final Report to U.S. Fish & Wildlife Service, unpub- lished, 89 pp. Malcolm, D.R. & J.C. Chamberlin. 1961. The pseudoscorpion genus Kleptochthonius Cham- berlin (Chelonethida, Chthoniidae). American Museum Novitates 2063:1-35. Muchmore, W.B. 1963. Redescription of some cav- emicolous pseudoscorpions (Arachnida, Chelo- nethida) in the collection of the Museum of Comparative Zoology. Breviora 188:1-16. Muchmore, W.B. 1965. North American cave pseudoscorpions of the genus Kleptochthonius, subgenus Chamberlinochthonius (Chelonethida, Chthoniidae). American Museum Novitates 2234:1-27. Muchmore, WB. 1976. New cavemicolous species of Kleptochthonius, and recognition of a new species group within the genus (Pseudoscorpion- ida: Chthoniidae). Entomological News 87:211- 217. Muchmore, W.B. 1990. Pseudoscorpionida. Ch. 18, pp. 503-527, In Soil Biology Guide (D.L. Dindal, ed.). John Wiley & Sons, New York. Muchmore, W.B. 1994a. Three unusual new epi- gean species of Kleptochthonius (Pseudoscor- pionida: Chthoniidae). Jeffersoniana 6:1-13. Muchmore, W.B. 1994b. Some pseudoscorpions (Arachnida: Pseudoscorpionida) from caves in Ohio and Indiana, U.S. A. Transactions of the American Microscopical Society 113:316-324. Muchmore, W.B. 1996. The genus Tyrannochthon- ius in the eastern United States (Pseudoscorpion- ida: Chthoniidae). Part II. More recently discov- ered species. Insecta Mundi 10:153-168. Muchmore, W.B. & J.C. Chamberlin. 1995. The genus Tyrannochthonius in the eastern United States (Pseudoscorpionida: Chthoniidae). Part 1. The historical taxa. Insecta Mundi 9:249-257. Packard, A.S. 1888. The cave fauna of North America, with remarks on the anatomy of the brain and origin of the blind species. Memoirs of the National Academy of Science 4:1-156. Roewer, C.E 1937. Chelonethi oder Pseudoskor- pione. Pp. 161-320. In Klassen und Ordnungen des Tierreichs, 5, IV, 6 (2). (H.G. Bronns, ed.). Akademische Verlagsgesellschaft m.b.H., Leip- zig. Vachon, M. 1952. A propos d’un Pseudoscorpion cavemicole decouverte par M. le Dr H. Henrot, dans une grotte de la Virginie occidentale, en Amerique du Nord. Notes Biospeologiques 7: 105-112. Manuscript received 27 July 1999, revised 10 De- cember 1999. 2000. The Journal of Arachnology 28:300-308 SPIDER SIZE AND LOCOMOTION ON THE WATER SURFACE (ARANEAE, PISAURIDAE) Robert B. Suter and Jessica Gruenwald: Department of Biology, Vassar College, Poughkeepsie, New York 12604 USA ABSTRACT. Newly emerged fishing spiders, Dolomedes triton (Walckenaer 1837), can achieve rowing velocities as high as those of adults despite an approximately 600-fold difference in mass (1.7 mg v.?. LI g). In contrast, when velocity is measured in relative terms (body lengths/sec), small spiders move much more rapidly than adults, with Uei mass'® This surprising performance of very small spiders can be attributed both to their very high stride frequency (f^ « mass “0.43) and to the high angular velocity of their propulsive legs (w oc mass'® 3^). Calculations of leg tip velocities, based on measurements of both angular velocities and leg lengths, reveal that maximum leg tip velocities are achieved by spiders of about 33 mg, nineteen times more massive than the smallest spiders we tested. Some very small spiders perform con- spicuously and consistently less well than do others of the same size. A detailed dissection of the motion of these underachievers reveals that a disproportionate amount of their rowing effort goes into vertical as opposed to horizontal work: the ratio of vertical to horizontal work during rowing is 1.03 ± 0.89 : 1 in normal fishing spiders and 5.18 ± 1.73 : 1 in the underachievers. Keywords: Allometry, locomotion, size, spider, aquatic Locomotion across the surface of water is performed both by fishing spiders (Arachnida, Araneae, Pisauridae) (McAlister 1959; Shultz 1987) and by water striders (Insecta, Hemip- tera, Gerridae) (Anderson 1976). This form of locomotion involves not only support by the water’s surface tension but also drag resis- tance to leg motion as the animals push back- ward to propel themselves forward. Crucial for the rowing gait in this semi- aquatic loco- motion is the dimple formed in the water sur- face as the animal’s leg pushes first down and then backward, because it is the leg with its dimple that encounters resistance as it moves horizontally, and it is that resistance against which the animal pushes to achieve forward acceleration (Suter et al. 1997). When the leg pushes too deeply into the water or moves horizontally too rapidly, the dimple disinte- grates, leaving the leg alone, without the dim- ple, to provide forward thrust. The leg without the dimple encounters far less drag resistance because its effective frontal surface area is much smaller (Suter & Wildman 1999). This unorthodox form of locomotion, avail- able only to relatively small creatures, is per- formed by pisaurid spiders of all ages and consequently of all sizes. In one of the most common pisaurids in North America, Dolo- medes triton (Walckenaer 1837), adult females at about 1 g are approximately 600 times as massive as hatchlings, a difference that should, in theory, result in substantial differ- ences in the efficacy of rowing locomotion. Is there a relatively simple allometric relation- ship between size and rowing efficacy in pi- saurid spiders? The allometric properties of aerial, terres- trial, and submerged locomotion have been studied in considerable detail (for references, see Pennycuick 1992; Calder 1984; Peters 1983). The allometry of locomotion on the water surface, however, has received very lit- tle attention: both Vogel (1994) and Denny (1993) have speculated about the difficulties that may be faced by very small organisms (e.g., juvenile water striders) attempting to push against a very slippery water surface, and we have published some data on allome- tric relationships (Suter et al. 1997; Suter & Wildman 1999) in support of specific argu- ments about the biomechanics of rowing in fishing spiders. Heretofore, there has been no explicit investigation of the relative efficacy of rowing locomotion for very small vs. much larger arthropods. Thus, the primary reason for undertaking the empirical studies on which we report here is the need to fill that gap in our knowledge. The secondary reason for these studies is derived from the observation 300 SUTER & GRUENWALD— SIZE AND AQUATIC LOCOMOTION 301 (Suter unpubl. obs.) that very young fishing spiders in the genus Dolomedes hunt primarily in terrestrial and emergent vegetation whereas adolescents and adults hunt more frequently at the water surface. The current study, if it demonstrates substantial size- specific differ- ences in rowing efficiency, could help to ex- plain the motivation for changing foraging be- havior during maturation. METHODS Organisms. — The species of fishing spider used in this study, Dolomedes triton (Araneae, Pisauridae), can become quite large (adult fe- males: to 1.5 g, 2 cm body length, and 9 cm leg span), and normally inhabit marshes and the edges of ponds and streams throughout much of North America (Gertsch 1979). The larger subjects (> 0.1 g) for these experiments were collected from small ponds in Mississip- pi and held in our laboratory (maintenance and experimentation at 22“25 °C) in 3.8 liter plastic aquaria containing water (about 2 cm deep) and an inverted clay flower pot to pro- vide a solid substrate. We fed these spiders assorted insects and changed their water ap- proximately once a week. The smaller sub- jects (1.7“100 mg) were hatched in the labo- ratory from an egg case borne by an adult female captured (as above) in Mississippi. The locomotion of the smallest subjects (1.7 mg) was studied while they were still 2nd instar hatchlings and had not yet eaten. These and the other small spiders were first reared com- munally in a 19 liter aquarium containing 2 cm of water and several rocks to furnish solid substrate (although they seldom left the glass walls of the aquarium), and were provided with live fruit flies {Drosophila melanogaster and D. virilis) and each other ad lib. The largest of the adult fishing spiders (1.05 g) was 610X as large as the smallest of the hatchling spiders (1.72 mg), providing us with more than 2.5 orders of magnitude in mass variation against which to scale the several pa- rameters of surface locomotion. High-speed videography. — Because most of the motion involved in the spiders’ rowing movements across the surface of water occurs in the horizontal plane, we videotaped their locomotion from directly above. The arena used in videotaping the locomotion of all but the smallest spiders consisted of a white por- celain-surfaced tray, a smooth, circular plastic barrier to prevent the spider’s escape, and a layer of water at least twice as deep as the deepest dimple we had observed for a spider of the size of the test spider. We used the bot- tom section of a small petri dish (6.0 cm di- ameter) as the arena for videotaping the lo- comotion of the smallest spiders. The arenas were lit with an incandescent point source (subtending an angle of 0.28° when 60 cm from the spider), mounted at 45° above and to one side of the videotaped part of the arena. We adjusted the camera’s aperture to obtain sufficient depth of field to allow both the spi- der and its shadow (on the porcelain surface of the arena or on a white sheet of paper under the petri dish) to be in sharp focus. During a trial, we placed a test spider in the arena, recorded its movements at 1000 fps with a Kodak EktaPro EM- 1000 video record- er, and stored the images in S-VHS format. We analyzed the spider’s motion in the hori- zontal plane by using Image (NIH shareware) to digitize and record the coordinates of the anterior and posterior tips of the body and the angles of legs III and II (relative to the body’s long axis) either every 1 ms (for small spi- ders) or every 5 ms. We used the body coor- dinates to measure the spider’s length, to an- alyze the displacement of the spider through time in both absolute and relative terms, and to calculate the pitch (p, degrees) of the body [p = cos“' (apparent length/true length)] as it changed during the rowing stride cycles. We used changes in leg angles over time to esti- mate the angular velocity of the leg and the velocity of its tip during the power phase of rowing locomotion. To estimate the horizontal component of the force exerted by a spider during a rowing stroke, we graphed the spider’s velocity (m/s) as a function of time (s) and used the slope of the linear part of the line as our estimate of acceleration (Suter et al. 1997). We then applied Newton’s second law (F = ma) to cal- culate the average net force exerted by the spi- der during acceleration in the horizontal plane. To estimate the horizontal work done dur- ing a rowing stroke, we multiplied the hori- zontal component of force by the distance the spider traveled during the application of the force (w — mad). To estimate the vertical work done during a rowing stroke, we used the measurements of pitch to calculate the maximum change in the height of the spider’s 302 THE JOURNAL OF ARACHNOLOGY Mass (g) Figure 1. — The relationship between Dolomedes triton mass and velocity (mean ± SD). Especially for very small spiders, between-individual variance is high; standard deviations indicate substantial var- iability within individuals. center of mass (Fig. 9). That change in height required work against gravity (w = mgh). RESULTS Actual horizontal velocities achieved by spiders rowing across the surface of distilled water averaged 0.107 ± 0.038 m/sec (mean ± SD of the means of 14 individuals, 51 trials total) and showed no obvious systematic changes with spider mass (Fig. 1). Within-in- dividual variation was high for some spiders, likely a consequence of differences in effort as reflected in differences in the angular ve- locities of the propulsive legs (Suter et al. 1997; Suter & Wildman 1999). When the log- arithms of velocities were plotted against the logs of spider masses, however, it became clear that inter-individual differences were substantially greater for very small spiders than for larger ones (Fig. 2). Two individuals, both with average velocities < 0.05 m/sec, were outliers in a population of spiders whose average velocities otherwise remained be- tween 0.08 m/sec and 0.15 m/sec. Because the outliers had masses < 0.03 g, we included in our analysis a treatment of two groups that were defined by spider size (the five spiders with masses < 0.03 g, two of which were the outliers, and the nine larger spiders). Although absolute rowing velocities did not Figure 2. — The log-log relationship between spi- der mass and velocity. Spiders greater than 0.03 g (•) had mean rowing velocities > —1.2 log m/s whereas spiders with masses less than 0.03 g (o) had a bimodal distribution of mean rowing veloci- ties with three individuals attaining velocities com- parable to those of the larger spiders and two in- dividuals having mean velocities < — 1.4 log m/s. To elucidate the differences among the large and the two modes of small spiders, the rowing loco- motion of three spiders (indicated by “ + a, 1.7 mg; b, 2.0 mg; c, 308 mg) is characterized in detail in Figure 5 and is highlighted in Figures 3, 4, and 10). vary systematically with mass, relative veloc- ities, (body lengths/sec), decreased sub- stantially and significantly with increasing mass (Fig. 3). The relationship for only the larger spiders (> 0.03 g) was Vrei “ (r^ = 0.653, n = 9, P < 0.01) and that relation- ship changed only slightly when the smaller spiders were included in the analysis (Vrei ^ ^-0.31; ^2 = 0.785, n = 14, P < 0.01). This strong relationship may be a consequence, in part, of the fact that stride frequency, (strides/sec), also decreased significantly with increasing mass (Fig. 4; for larger spiders, = and for all spiders, In an effort to understand the causes of the quantitatively very different performances of individuals of the same small mass and the very similar performances of individuals of divergent masses, we analyzed in detail the locomotor behavior of three spiders (a-c in Figs. 1-3). In this trio, a, b, and c had masses SUTER & GRUENWALD— SIZE AND AQUATIC LOCOMOTION 303 Figure 3. — The log-log relationship between spi- der mass and relative velocity measured in body lengths traveled per second. A linear fit to all points indicates LogjoV = —0.305 Logiom {n = 13, r^ = 0.785, P < 0.01). When we removed spiders small- er than 0.03 g from the analysis, the slope of the linear fit changed only slightly, to —0.320 {n = 9, r2 = 0.653, P < 0.01). of 1.7, 2.0, and 308 mg respectively, but had mean absolute velocities of 0.038, 0.109, and 0.143 m/sec respectively. The spiders’ hori- zontal velocities varied over time, rising lin- early during the propulsive phase of each stride and decreasing during the recovery phase (Fig. 5, top row). The acceleration was, in each case, accompanied by a rapid rise in the area of the shadows cast by the dimples made by legs III and II (Fig. 5, middle row), indicating a corresponding rise in the depth of each dimple (shadow area, in mm^, is a nearly perfect linear correlate of dimple depth, in mm: for a 13.8 mm length of hydrophobic wire, for example, area = 34.6 depth - 6.2, ^ — 8, r^ = 1.00). For the two smallest spiders {a, b), the horizontal acceleration was also ac- companied by a distinct rise in body pitch (the angle between the animal’s body and the wa- ter surface), but such an association was much less evident for the largest spider (c) (Fig. 5, bottom row). Data from an earlier study indicated that leg length in these spiders is directly proportional to the log of mass (Fig. 6, modified from Suter & Wildman 1999, fig. 7 A) and that the log of the angular velocities of the propulsive legs is Figure 4. — Stride frequency during rowing is closely related to spider mass. A linear fit to log- transformed data yields the equation Logjo/s = -0.432 Logiom + 0.145 (n = 13, r^ = 0.877, P < 0.01). When we removed spiders smaller than 0.03 g from the analysis, the slope of the linear fit changed to —0.357 (n = 9, r^ = 0.676, P < 0.01). inversely proportional to the log of mass (Fig. 7, modified from Suter & Wildman 1999, fig. 7B). Further analysis of these data relating an- gular velocity to mass indicates that maximum (i.e., achievable) angular velocity is probably also a linear function of mass when both var- iables are log-transformed, but with a more steeply negative slope (Fig. 7). Because the propulsive force available during rowing is strongly dependent on the velocity of the leg tips, we used the equations relating mass to leg length and angular velocity (Figs. 6, 7) to elucidate the relationship between leg tip ve- locity and spider mass (Fig. 8). Achievable leg tip velocity rose steeply with mass until it reached its maximum at about 0.8 m/s in 33 mg spiders, and then fell with further increas- es in mass. In contrast, average leg tip veloc- ity also rose rapidly to its peak (about 0.3 m/ s) in 141 mg spiders but then declined only very slightly with further increases in mass. DISCUSSION General observations. — The principles governing the relationships between size and locomotion remain controversial despite sev- eral decades of careful measurements (re- viewed, for example, in Peters 1983; Garland 1983; Calder 1984; Pennycuick 1992). Much of the controversy revolves around the scaling 304 THE JOURNAL OF ARACHNOLOGY Figure 5. — Analysis of the rowing locomotion of the three spiders first identified in Figure 2. Velocities (top row of graphs; three-point running averages) vary periodically with time, indicating accelerations and decelerations that correspond to the propulsive and the glide phases of each rowing stroke; the horizontal acceleration during each stroke was calculated as the slope of the velocity versus time segment of the propulsive phase. The change in dimple shadow area (middle row of graphs) created by legs II (solid lines) and III (dashed lines) during the strokes graphed in the top row; ellipses indicate the relative sizes of maximum dimple shadows for the three different spiders. The change in body pitch (bottom row of graphs) during the same strokes; body pitch, referring to the angle formed between the body’s long axis and the horizontal, varied substantially more during rowing by spider a than it did during the rowing of the other two spiders. SUTER & GRUENWALD— SIZE AND AQUATIC LOCOMOTION 305 Figure 6. — Leg length in fishing spiders varies linearly with the log of mass. This relationship [L, = 13.22 LogioW + 28.73; n = 62, two legs each (• and o) for 31 spiders, r^ = 0.946, P < 0.001] is surprising because, for most other organisms in which it has been measured, it is the log-log rela- tionship that is linear. parameter that is to be held constant while other parameters are changed: different pre- dictions follow from models based on geo- metric similarity, for example, than from those based on similarity of musculoskeletal elastic- ity or stress in the propulsive limbs. Not sur- prisingly, the reviewed literature concerns ter- restrial locomotion, flying, and swimming, and ignores locomotion on the water surface. The present paper is the first to explore these relationships in an organism that regularly in- habits the water surface, although size and lo- comotion on the water surface by basilisk liz- ards has received some attention (Glasheen & McMahon 1996a, 1996b). A comparison of the results of this study with expectations from other (principally ver- tebrate) studies reveals that models based on geometric similarity (Hill 1950) work well for rowing locomotion as performed by D. triton (Table 1): absolute velocity is approximately scale-invariant (Fig. 2) while relative velocity varies with mass~®^* (Fig. 3) and stride fre- quency varies with mass ®"^^ (Fig. 4). The ab- solute vs. relative velocity findings are partic- ularly interesting for two reasons. First, the fact that very small spiders can achieve the same rowing velocities as adults despite a 610-fold difference in mass is remarkable. That achievement is made possible by both the rapid rise in stride frequency with decreas- Log mass (g) Figure 7. — Log-transformed angular velocities of the propulsive legs vary approximately linearly with the log-transformed masses of the spiders (LogioW = —0.328 Logio»i — 0.256, n = 50, r^ = 0.379, P < 0.01). In an effort to estimate the an- gular velocities of which spiders are capable, we fit a line to the maxima (circled data points; LogioW = — 0.623 LogioW — 0.266; n ~ 9, = 0.991, P < 0.01). Figure 8. — Leg tip velocity is maximum when spiders are very small. The relationship between leg tip velocity and mass was calculated from the al- lometric relationships between leg length and mass (Fig. 6) and between leg angular velocity and mass (Fig. 7). The solid line indicates the relationship based upon average angular velocities (lower curve fit. Fig. 7) and the broken line indicates the rela- tionship based upon maximum angular velocities (upper curve fit. Fig. 7). Peak leg tip velocities oc- curred at 0.141 g and 0.033 g for the solid and dashed lines, respectively. 306 THE JOURNAL OF ARACHNOLOGY Table 1. — Allometric relationships predicted in the literature and determined in this study. Proportionality Relationship Expected Source Observed Source absolute velocity vs. mass y oc mO V oc mO-24 Hill 1950 Heglund et al. 1974 y a Fig. 2 relative velocity vs'. mass V (X m-0'33 y a m-009 Hill 1950 Calder 1984 y (X m-o-31 Fig. 3 stride frequency vs. mass f, a m-0'33 /, oc Hill 1950 Alexander 1982 /, oc Fig. 4 ing size (Fig. 4) and the fact that leg tip ve- locity is at a maximum in very small spiders (Fig. 8). The achievement also bespeaks the efficacy of the drag-based propulsive system used by even the smallest of the spiders (Suter & Wildman 1999). And second, relative ve- locity may be a better predictor of success at evading predation than is absolute velocity (Van Damme & Van Dooren 1999), hypothet- ically making newly hatched spiderlings far more difficult to capture than their parents since the spiderlings’ relative velocities can be up to lOX as great (Fig. 3). Conspicuous differences among small spiders. — The generalizations discussed above ignore the conspicuously poor perfor- mances of the two very small spiders with av- erage rowing velocities < 0.05 m/sec (Fig. 2). To understand their relative deficit, we ana- lyzed one of them (Figs. 2-4) in detail, and compared the results to those of a similarly small but fast spiders {b) and to those of a much larger adult (c). In theory, the low ve- locity of a could be caused by any number of physiological or biomechanical deficits such as low power output by the muscles that move the legs, low stride frequency due to muscular or neural deficits, sub-optimal stroke direc- tion, and so forth. The difference between a and b in stride frequency (Figs. 4, 5), although in the right direction, is only about 12%, far too little to account for the more than three-fold differ- ence in absolute velocity (Fig. 5, top row: a, 0.046 m/s; b, 0.16 m/s). In contrast, there are substantial and revealing differences between Figure 9. — The change in pitch of the long axis of the spider’s body was measured both at the beginning of the rowing stroke (angle a) and at the peak of thrust production (angle (B). The change in pitch allowed us to estimate the change in height of a spider’s center of mass [h ^ I (sin|3 — sina) where I is the length of the body] and therefore to calculate the vertical work done during the rowing stroke. Because a spider can raise its body without changing the body’s pitch (Suter 1999), the use of pitch changes to estimate work done against gravity underestimates vertical work. SUTER & GRUENWALD— SIZE AND AQUATIC LOCOMOTION 307 the dimple areas recorded during the power strokes of the three spiders in this comparison (Fig. 5, middle row): the 308 mg spider (c) made dimples in the water surface that were approximately lOOX as large (in the horizon- tal plane) as those of the 2.0 mg spider (b), but this 2.0 mg spider’s dimples were about lOX as large as those of the 1.7 mg spider (a). Because a is the outlier in performance (Fig. 2), we took the relationship between spider mass and dimple area to be best indicated by the data from b and c. From that perspective, the dimple area of a is low by a factor of between 7 and 8 (expected maximum dimple area, 2.01 mm^; observed maximum dimple area, 0.26 mm^). Because the relationship between dimple area and dimple depth is linear (see Results), the observed deficit in dimple area for a cor- responds to a seven- to eight-fold deficit in dimple depth. That difference, in turn, is more than enough to account for the observed ve- locity difference between spiders a and be- cause the horizontal thrust force that can be generated by a spider on the water surface is strongly influenced by dimple depth (Suter et al. 1997; Suter & Wildman 1999). To spider a, therefore, the water surface would appear to be very slippery, offering much less resis- tance to the backward motion of the propul- sive legs than would be encountered if the dimples were deeper. The spider could com- pensate behaviorally for the deficit in dimple depth (and resistance) in either of two ways: it could increase the angular velocities of the propulsive legs or it could deflect some of its leg displacement downward in an attempt to increase dimple depth, either of which would have the effect of increasing drag (Suter & Wildman 1999). Our estimates of the ratios of vertical to horizontal work during rowing strokes (Fig. 10, derived from pitch measure- ments like those in Fig. 5, bottom row) indi- cate that spider a was doing about four times as much work to raise itself against gravity as it was to propel itself forward, confirming its attempts to increase dimple depth. Not sur- prisingly, the spider with the highest ratio of vertical to horizontal work, d (Fig. 10), was also very small and the slowest of all of the spiders tested (Fig. 2). The underlying cause of the difference in dimple sizes that appears to be at the root of the locomotor difficulties experienced by spi- Figure 10. — The ratio of vertical work (w = mgh, mJ) to horizontal work (w = mad, mJ) in relation to the log of spider mass. The high ratio measured for spider a (four times as much work going into vertical displacement as into horizontal locomotion during each rowing stroke) provides an explanation for the conspicuously slow horizontal locomotion of that spider (Fig. 2). That inefficiency is surpassed by only one spider {d), which is similarly conspic- uous in horizontal velocity (Fig. 2, lowest velocity). der a but not by spider b remains obscure. Because the two spiders differed little in mass (1.7 vs. 2.0 mg), the mass difference is un- likely to explain the differences in dimple siz- es. We suspect, instead, that variations either in the hydrophobicity of the spider cuticle or in the structures of hairs on the tarsi (J. Rov- ner and R Sierwald, pers. comm.) may ac- count for the unexplained differences. ACKNOWLEDGMENTS We thank Edgar Leighton, Patricia Miller and Gail Stratton for providing us with the subjects of this study, and we thank John Long both for sharing his considerable bio- mechanics expertise and for the use of the high-speed videography equipment (provided to JL by grant #N000 14-97- 1-0292 from the Office of Naval Research). Erin Murphy’s col- lection of the pilot data that led to this study is also appreciated. The study was supported in part by funds provided by Vassar College through the Undergraduate Research Summer Institute and the Class of ’42 Faculty Research Fund. 308 THE JOURNAL OF ARACHNOLOGY Appendix 1. — Symbols used in the text and figures. Symbol Meaning Units a acceleration m/s^ d distance m F force mN fs stride frequency strides/sec 8 acceleration due to gravity m/s^ h height m 1 length of body m m mass g P pitch of longitudinal axis degrees Ke. relative velocity body lengths/sec W work mJ LITERATURE CITED Alexander, R. McN. 1982. Size, shape, and struc- ture for running and flying. In A Companion to Animal Physiology (C.R. Taylor, K. Johansen & L. Bolis, eds.). Cambridge Univ. Press, New York. Anderson, N.M. 1976. A comparative study of lo- comotion on the water surface in semiaquatic bugs (Insects, Hemiptera, Gerromorpha). Viden- skabelige Meddelelser fra Dansk Naturhistorisk Forening 139:337-396. Calder, W.A. III. 1984. Size, Function and Life History. Harvard Univ. Press, Cambridge. 431 pp. Denny, M.W. 1993. Air and Water: The Biology and Physics of Life's Media. Princeton Univ, Press, Princeton. 341 pp. Garland, T. Jr, 1983. The relation between maximal running speed and body mass in terrestrial mam- mals. Aumal of Zoology 199:157-170. Gertsch, W.J. 1979. American Spiders (2nd ed.). Van Nostrand Reinhold Company, New York. 274 pp. Glasheen, J.W. & T.A. McMahon. 1996a. A hydro- dynamic model of locomotion in the basilisk liz- ard. Nature 380:340-342. Glasheen, J.W. & T.A. McMahon. 1996b. Size-de- pendence of water-running ability in basilisk liz- ards (Basiliscus basiliscus). Journal of Experi- mental Biology 199:2611-2618. Heglund, N.C., C.R. Taylor & T.A. McMahon. 1974, Scaling stride frequency and gait to animal size: Mice to horses. Science 186:1112-1113. Hill, A.V. 1950. The dimensions of animals and their muscular dynamics. Science Progress 38: 209-230. McAlister, W.H. 1959. The diving and surface- walking behaviour of Dolomedes triton sexpunc- tatus (Araneida: Pisauridae). Animal Behaviour 8:109-111. Pennycuick, C.J. 1992. Newton Rules Biology: A Physical Approach to Biological Problems. Ox- ford Univ. Press, Oxford. Ill pp. Peters, R.H. 1983. The Ecological Implications of Body Size. Cambridge Univ. Press, Cambridge. 329 pp. Shultz, J.W. 1987. Walking and surface film loco- motion in terrestrial and semi-aquatic spiders. Journal of Experimental Biology 128:427-444. Suter, R.B. 1999. Cheap transport for fishing spi- ders: The physics of sailing on the water surface. Journal of Arachnology 27:489-496. Suter, R.B., O. Rosenberg, S. Loeb, H. Wildman & J.H. Long, Jr. 1997. Locomotion on the water surface: Propulsive mechanisms of the fisher spi- der Dolomedes triton. Journal of Experimental Biology 200:2523-2538. Suter, R.B. & H. Wildman. 1999. Locomotion on the water surface: Hydrodynamic constraints on rowing velocity require a gait change. Journal of Experimental Biology 202:2771-2785. Van Damme, R. & T.J.M. Van Dooren. 1999. Ab- solute versus per unit body length speed of prey as an estimator of vulnerability to predation. An- imal Behaviour 57:347-352. Vogel, S. 1994. Life in Moving Fluids (2nd ed.). Princeton Univ. Press, Princeton. 467 pp. Manuscript received 23 August 1999, revised 4 January 2000. 2000. The Journal of Arachnology 28:309-318 CHEMICAL CUES FROM ANTS INFLUENCE PREDATORY BEHAVIOR IN HABROCESTUM PULEX, AN ANT-EATING JUMPING SPIDER (ARANEAE, SALTICIDAE) Robert J. Clark and Robert R. Jackson: Department of Zoology, University of Canterbury, Private Bag 4800, Christchurch, New Zealand Bruce Cutler: Electron Microscope Laboratory, Haworth Hall, University of Kansas, Lawrence, Kansas 66045-7534 USA ABSTRACT. The ability of Habrocestum pulex, a myrmecophagic jumping spider, to detect olfactory and contact chemical cues from ants was investigated experimentally. When given a choice between walking over clean soil or soil that had housed ants, H. pulex spent significantly more time on ant-treated soil. However, H. pulex did not appear to discriminate between clean blotting paper and blotting paper over which ants had walked. In tests using a Y-shaped olfactometer, when given a choice between an experimental arm containing air from a cage containing ants, or 6-methyl-5-hepten-2-one, and a control arm containing clean air, H. pulex moved into the experimental arm significantly more frequently than the control arm. When on soil that had previously housed ants, agitated walking, undirected leaping, posturing with body raised, and perching on top of corks were each significantly more prevalent than when H. pulex was on clean soil. Chemical cues left by ants on soil also affected H. pulex’s attention to visual cues from ants: when on treated soil, H. pulex initiated and completed stalking sequences more often, and after shorter latency, than when on control soil. Keywords: Prey detection, myrmecophagy, kairomone, Salticidae, Habrocestum pulex Unique, complex eyes and acute vision in jumping spiders (Salticidae) have led to the evolution of intricate, vision-guided courtship and predatory tactics (Crane 1949; Drees 1952; Land 1969a, 1969b; Forster 1982; Blest et al. 1990; Jackson & Pollard 1996, 1997). However, salticids are not restricted to reli- ance on optical cues, as tactile, auditory and substrate-vibration cues also influence salticid courtship, either concurrent with or as alter- natives to, visual communication (Richman & Jackson 1992; Jackson & Pollard 1997). Pher- omone-based intraspecific communication is also widespread in the Salticidae (Crane 1949; Jackson 1987; Pollard et al. 1987; Willey & Jackson 1993; Clark & Jackson 1994a, 1994b, 1995a, 1995b), but little is known about whether salticids are influenced by kairomo- nes (chemicals that provoke a response ben- eficial to the receiver but not the sender of the signal, where the sender and receiver belong to different species; Brown et al. 1971). Ants are one of the most abundant prey-size arthropods in the habitats of most spiders (Holldobler & Wilson 1990), but their defens- es (strong mandibles, formic acid and poison- injecting stings: Wray 1670; Edmunds 1974; Holldobler & Wilson 1990; Blum 1992) ap- pear to present spiders with formidable chal- lenges. Yet a minority of spiders has over- come the ant’s defenses, thereby gaining access to this exceptionally numerous prey (Mackay 1982; Oliviera & Sazima 1985; Nyf- feler et al. 1988; Elgar 1993; Cushing 1997). Within the Salticidae, 21 ant-eating (myr- mecophagic) salticids have been studied in de- tail: Aelurillus aeruginosus (Simon 1871), A. cognatus (O.R-Cambridge 1872), A. kochi Roewer 1951, six undescribed species of Chalcotropis Simon 1902, Chrysilla lauta Thorell 1887, Corythalia canosa (Walckenaer 1837), Habrocestum pulex (Hentz 1846), Siler semiglaucous Simon 1901, Siler sp. Simon 1889, three undescribed species of Natta Karsch 1879, two undescribed species of Xen- ocytaea Berry, Beatty, Prozynski 1998 (for- merly called ''Euophrys'') and Zenodorus or- biculatus (Keyserling 1881) (Edwards et al. 1974; Cutler 1980; Jackson & van Olphen 1991, 1992; Li et al. 1996, 1999; Jackson et al. 1998). Although these species feed on a wide variety of insects, they have all been 309 310 THE JOURNAL OF ARACHNOLOGY shown in standardized tests to prefer ants over other prey and to have ant-specific prey-cap- ture behavior (Li & Jackson 1996). Except for Corythalia canosa and Zenodorus orbicula- tus, each of these species has been shown to prefer ants as prey and to use ant-specific prey-capture behavior even when tested with motionless lures (dead insects mounted in life- like posture on corks), implying that optical cues pertaining to shape and form enable them to distinguish ants from other insects (Li & Jackson 1996; Li et al. 1996; Jackson et al. 1998). However, the ability to rely solely on vision for detecting ants does not preclude the possibility that chemical cues also influence the predatory behavior of myrmecophagic sal- ticids. In the present paper, we investigate how Habrocestum pulex, a previously studied myr- mecophagic salticid from North America, re- sponds to chemical cues from ants. Habroces- tum pulex lives in leaf litter, a microhabitat in which numerous visual obstructions might of- ten hinder early visual detection of prey. Abil- ity to detect chemical cues from ants might play an important role in preparing H. pulex to respond appropriately to its unusually dan- gerous prey. In earlier studies (Cutler 1980; Li et al. 1996), H. pulex was tested with prey in a sim- ple laboratory environment. In the present study, we first observe H. pulex's predatory behavior in an environment with leaf litter present, thereby simulating nature more close- ly than previously. We next consider three hy- potheses concerning how H. pulex might react to contact chemical cues when in an environ- ment recently occupied by ants. Habrocestum pulex might do any combination of the follow- ing: remain in the environment, adopt behav- ior and posture appropriate for capturing ants, or exhibit heightened attention to optical cues from ants. We consider the role of both olfac- tory and contact chemical cues from ants in moderating the prey-capture behavior of H. pulex. METHODS General. — Except for minor modifications, maintenance procedures, cage design and data analysis were as in earlier studies (Jackson & Hallas 1986). All experiments were carried out in New Zealand using laboratory cultures of H. pulex, originally collected in Kansas, USA. Each individual salticid was used in a maximum of two tests for any one experi- ment, and there was no evidence that the iden- tity of individual salticids influenced test out- come. Data from males and females, not being statistically different, were pooled. Body lengths of adults were 3-5 mm. Statistical methods were from Sokal & Rohlf (1995). In observations and experiments with live ants, we used Monomorium antarcticum Smith 1858, a myrmicine ant native to New Zealand (Ettershank 1966; Bolton 1987). The most common prey of H. pulex in nature ap- pear to be Lasius spp. Fabricius 1804 (For- micinae) (Cutler unpubl. data), which were not available in New Zealand. To test for re- sponses which might be specific to Lasius spp., we conducted olfactometer tests using commercially available 6-methyl-5-hepten-2- one (Sigma Chemical Co.), an alarm phero- mone of Lasius spp. and other ants (Duf field et al. 1977; Blum 1981; Turker 1997a, 1997b). Monomorium antarcticum and other myrmicine ants appear not to make this pher- omone (Holldobler & Wilson 1990). Predation on ants in a complex environ- ment.— The environment was a plastic box (length 170 mm, width 110 mm, depth 60 mm) filled to a depth of 15 mm with soil. Leaf litter was scattered about on top of the soil, covering about 30% of the box surface. Four small corks on which H. pulex could stand were spaced within the box, providing perches above the level of leaf litter. Observations were staged by putting H. pulex in this envi- ronment in the presence of 10-20 prey, where (depending on the test) prey were either ants or vestigial-winged fruit flies {Drosophila me- lanogaster Meigen 1804). The goal was to get qualitative information on how H. pulex cap- tured prey in approximately natural environ- ments. Choice tests using blotting paper.^ — We adopted, after minor modification, procedures devised earlier for testing the ability of salti- cids to discriminate between the draglines of different conspecific individuals (Clark & Jackson 1994a, 1995a, 1995b). In each test, H. pulex was offered a choice between treated (had been in contact with ants) and untreated (clean) blotting paper. Treated blotting paper was prepared by leaving four ants in a plastic petri dish (diameter 90 mm) for two hours, with one circular piece of blotting paper taped CLARK ET AL.— CHEMICAL CUES FROM ANTS 311 to the top and another to the bottom. During the twO“hour period, ants actively walked about in the petri dish, repeatedly moving over both pieces of blotting paper. Immediately afterward, each piece of blot- ting paper was cut in half and the test chamber was prepared. The test chamber was another petri dish (diameter 90 mm) with one half piece of treated blotting paper taped to the top of the dish and another half-piece of treated blotting paper taped to the bottom of the dish directly below the top piece. The other half of the test chamber had control blotting paper taped to the top and bottom. A 15 -mm trian- gle, cut out of the blotting paper and surround- ed by a horseshoe-shaped metal divider, served as a “neutral area” into which the test spider was introduced before testing. Having the metal divider in place meant that the sal- ticid could not, all at once, view the entire space within the petri dish (see Clark & Jack- son 1994a). A test was defined as having start- ed when the spider moved out of the neutral area and onto the blotting paper. This always happened within 1 min. The test ended 10 min later. For each test, a difference score was ob- tained (time spent on treated paper minus time spent on control paper). Maximum and mini- mum possible scores were +600 sec (spent entire time on ant-treated blotting paper) and — 600 sec (spent entire time on control blot- ting paper), respectively. Choice tests using soil. — Commercial pot- ting mix was placed in a square (160 mm X 160 mm, height 80 mm) plastic storage con- tainer filled to a depth of 20 mm and nficro- waved (900 W) for 10 min, then held in the container (kept closed) for a waiting period of 20-30 days. Treated soil was prepared by keeping about 100 ants in the closed container during the waiting period. Potential contami- nants from feeding material were avoided by not feeding the ants during this time. The ants survived the fasting period. Control soil was kept ant free. The test chamber was a plastic box (length 170 mm, width 110 mm, height 60 mm) filled to a depth of 15 mm with control soil. Two watch glasses (inner diameter 50 mm, inner height 7 mm; outer diameter 65 mm, outer height 15 mm) were placed 10 mm apart (measured from nearest edges) in the center of the box. The watch glasses were filled with soil, then embedded in the surrounding soil (soil level with rim of watch glass). To facil- itate seeing whether test spiders were in the watch glass, the rim of each glass was kept clear of soil. Treated soil was placed in the experimental watch glass (ants removed im- mediately beforehand) and control soil was placed in the control watch glass. Whether treated soil was on the left or right was de- cided at random for each test. To start a test, a spider was placed on the soil between the two watch glasses. For the next 60 min, we recorded how much time the test spider spent in each watch glass. Time spent outside the watch glasses was ignored. Effect of chemical cues in soil on behav- ior and posture. — Control and treated soils were prepared as in the experiment on choice of soil. Each test spider was tested on one day with treated soil and on the previous or next day (order decided at random) with control soil. During 15 -min tests, the test spider’s be- havior was recorded in detail, but we present data here only where there was statistical ev- idence of behavior being influenced by soil treatment. The test chamber was a cylindrical plastic dish (diameter 90 mm, height 40 mm) with soil covering the bottom to a depth of 10 mm. Four corks (diameter 9 mm at the narrow end) were embedded with the upper 5 mm of cork (narrow end) extending above the soil. Corks were evenly spaced in a square centered in the middle of the dish (center of each cork 20 mm from the center of the nearest neighboring cork). Evenly spread around the dish between the corks were four convex 10 X 10 mm piec- es of leaf litter (Oak, Quercus spp. Linnaeus 1753), each positioned so that the test spider could walk under it. Effect of chemical cues in soil on atten- tion to optical cues. — We investigated wheth- er H. pulex's attention to optical cues from ants is affected by the presence of chemical cues from ants. Preparation of soil and the test chamber was as described for the experiment on how chemical cues affect behavior and posture, except that no leaf litter was present and there was a glass vial (65 mm long, inner diameter 10 mm) containing two ants on the soil centered between the corks. Latencies to initiate and complete stalking sequences di- rected at the ants were recorded. Stalking was initiated when the test spider turned toward an ant and began to move steadily toward it, and 312 THE JOURNAL OF ARACHNOLOGY completed when the test spider touched the vial. Test spiders were allowed 15 min to be- gin stalking and subsequently allowed 15 min to complete the stalking sequence. Olfactometer tests. — A Y-shaped olfac- tometer (Fig. 1) with airflow adjusted to 1000 ml/min (Matheson FM-1000 flowmeter) was used to assess H. pulex's response to airborne odors from ants. At this airflow setting, there was no evidence that H. pulex's locomotion was impaired. Air flowed from a tap through two separate flowmeters into a stimulus cham- ber (which contained an odor source) and a control chamber (which was empty). During experimentation, whether the experimental chamber was on the left or right side of the olfactometer was decided at random. Air moved from the stimulus chamber to the stim- ulus arm and from the control chamber to the control arm. Collectively, the stimulus and control arms are referred to as the “choice arms.” Air flowed from each “choice arm” into a single test arm. At one end of the test arm, there was a holding chamber into which a spider was placed prior to testing. A metal barrier, positioned in a slit between the hold- ing chamber and the test arm, blocked the spi- der’s entry into the test arm. Thirty min before each test, an odor source (depending on the experiment, either four ants or 10 p.1 of 6- methyl-5-hepten-2-one) was placed in the ex- perimental chamber. This 30-min period al- lowed the air to circulate evenly and ensured that air pressure was comparable throughout the olfactometer. During testing, spiders tended to walk about actively in the olfactometer, sometimes enter- ing the experimental or control arm, or both, several times but staying only briefly. For each spider, we recorded both the first and fi- nal choice. The first arm the spider entered was its first choice regardless of how long it stayed. By definition, a spider made its final choice when it entered an arm and remained there for a minimum of 30 sec. A maximum of 60 min was allowed for the spider to make a final choice after leaving the holding cham- ber. Between tests, the olfactometer was dis- mantled and cleaned first with 80% ethanol and then with water. This was a precaution against the possibility that spiders might be affected by draglines or chemical traces from previously tested spiders. T O Figure 1. — Olfactometer. Arrows indicate direc- tion of airflow. SC = stimulus chamber (contains odor source); CC = control chamber (empty); H = holding chamber (location of test spider at start of test); TA = test arm; CA = control arm; SA = stimulus arm; MS = metal screen fitted in slit (blocks spider’s entry into test arm before test be- gins); T = tap from which air enters olfactometer; B = opaque barrier (prevents test spider from see- ing ants); RS = rubber stopper; O = air leaves ol- factometer; EB = edge of box enclosing olfactom- eter. Diagram not to scale. See text for details. RESULTS Predation on ants in a complex environ- ment.— Habrocestum pulex tended to leap on fruit flies from any orientation, but attacked ants by repeatedly approaching head on, mak- ing stabs with its fangs, then backing away (Fig. 2). Once the ant was more or less qui- escent, H. pulex approached slowly, grasped the ant and began feeding. During and im- mediately prior to attacking an ant, the spi- CLARK ET AL.— CHEMICAL CUES FROM ANTS 313 der’s palps were retracted to the sides of the chelicerae, but palps tended not to be retracted during attacks on flies. Locomotion, when it occurred during tests with flies, tended to be by slow, continuous stepping, and the normal posture was adopted with the body ca. 1 mm above the substrate and legs only moderately extended. With ants, prey-capture sequences were normally preced- ed by distinctive preliminary behavior which included agitated walking, undirected leaping and posturing with the body raised. These se- quences were often preceded by periods dur- ing which H. pulex simply watched (main- tained orientation towards) an ant. Agitated walking was a distinctive style of motion in which H. pulex repeatedly spurted forward for ca. 0.5 sec at 30-50 mm/sec, paused and then spurted forward again. Habrocestum pulex made undirected leaps by suddenly propelling itself more or less straight upward with no tar- get being evident. When in the body-elevated posture, H. pulex stood with its legs more ex- tended than normal, so that its body was 2-3 mm off the substrate. When predation was delayed or failed to occur in tests with flies, H. pulex spent much of the time sheltering under leaf litter, but H. pulex rarely sheltered under leaves in tests with ants. A common preliminary to predation on both ants and flies was for H. pulex to stand on corks and watch prey active on the soil below (Fig. 3). Attacks were often made by rushing down from a cork, after which H. pulex usually returned to the top of the same cork to feed. Choice tests using blotting paper. — Scores were spread more-or-less evenly over the range of possible values, providing no ev- idence that H. pulex discriminated between treated and control blotting paper (Fig. 5). Choice tests using soil. — Habrocestum pu- lex spent more time on treated, rather than control, soil (Fig. 6). In 20 tests, one spider spent more time on control soil, one spent equal time on treated and control soil, and the remaining 18 spent more time on treated soil (McNemar test comparing the number that spent more time on treated versus control soil; P < 0.001, n = 19). Effect of chemical cues in soil on behav- ior and posture. — Agitated walking, undi- rected leaping, the body-raised posture and perching on corks were more prevalent when H. pulex was in experimental chambers rather than control chambers (Table 1). Effect of chemical cues in soil on atten- tion to optical cues. — When on treated soil, H. pulex initiated and completed (Fig. 4) stalking sequences against ants more often than when on control soil (Table 1). The la- tency to initiate and to complete stalking was shorter on treated than control soil (Fig. 7). Olfactometer tests. — When tested with ants in the stimulus chamber, the first choice was the stimulus arm in 1 1 tests and the con- trol arm in four tests (binomial, NS). The final choice was the stimulus arm in 13 tests and control arm in two tests (binomial, NS). In all tests in which the stimulus arm contained 6- methyl-5-hepten-2-one, the first and final choices were identical: the stimulus arm in 10 tests and the control arm in one test (binomial, NS). There was no statistical evidence of a relationship between latency to choose and whether the choice was the control or the stimulus arm or, if it was the stimulus arm, whether the stimulus was pheromone or an ant (Mann- Whitney rank-sum tests, NS; Fig. 8). DISCUSSION Habrocestum pulex apparently detects and responds adaptively to chemical cues from ants. Our findings support the following hy- potheses: (1) /7. pulex chooses to remain on soil containing chemical cues from ants (choice of soil); (2) ant-derived chemical cues in soil stimulate H. pulex to adopt posture and behavior appropriate for capturing ants, even in the absence of optical cues from ants (effect of chemical cues on behavior and posture); (3) ant-derived chemical cues in soil heighten H. pulex’s attention to optical cues from ants (ef- fect of chemical cues in soil on attention to optical cues); and (4) H pulex is attracted by olfactory cues from ants (olfactometer tests). Failure to show a preference for treated over control blotting paper in a petri dish suggests that blotting-paper choice tests are excessively artificial. Rather than demonstrating responses to the particular ant species on which H. pulex preys most often in nature, our results suggest that H. pulex has evolved the ability to detect and respond adaptively to chemicals secreted by a broader range of ants. In all experiments, we used Monomorium antarcticum, a New Zea- land myrmicine ant which would not be en- 314 THE JOURNAL OF ARACHNOLOGY Figures 2-4. — 2. Habrocestum pulex (on right) slowly approaches ant {Monomorium antarticum) (on left). Ant now quiescent, having been repeatedly stabbed by H, pulex; 3. Habrocestum pulex on top of cork watching ant (not in photograph) moving about on soil; 4. Habrocestum pulex completes stalking sequence in tests of effect of chemical cues in soil on attention to optical cues (see text). Ant in glass vial (lower right), H. pulex (above, left) faces ant and touches glass. CLARK ET AL.— CHEMICAL CUES FROM ANTS 315 Difference score (sec) Figure 5. — Distribution of difference scores (time spent on treated blotting paper minus time spent on control blotting paper) from experiment on choice of blotting paper. See text (data more-or-less evenly spread). No statistical evidence of prefer- ence (Wilcoxon test for paired comparisons, NS). Difference score (sec) Figure 6. — Distribution of difference scores (time spent in experimental watch glass minus time spent in control watch glass) from experiment on choice of soil, showing preference for treated soil (Wilcoxon test for paired comparisons, P < 0.001). Note: There was only one negative score. countered by H. pulex in nature. Habrocestum pulex preys especially often in nature on Las- ius spp., which are formicines. In our exper- iments, H. pulex also was influenced by 6- methyl-5“hepten-2"One, a ketone characteristic of the mandibular gland secretions of many formicine ants and the anal gland secretions of dolichoderine ants (Duffield et al. 1977). In ants, use of chemically-similar pheromones by different species is common (Gabba & Pavan 1970). The ketone 6-methyl-5“hepten-2-one ap- pears to be a kairomone not only for H. pulex but also for Habronestes bradleyi Walckenaer, a myrmecophagic zodariid spider. When test- ed in a Y-shaped olfactometer, with a choice between chemical cues from disturbed doli- choderine ants (Iridomyrmex purpureus Smith 1858) and clean air, Habronestes bradleyi most often moved toward the cues from in- jured or disturbed ants (Allan et al. 1996). Gas chromatography revealed that 6-methyL5- hepten-2“One is released in high concentra- tions by injured or disturbed Iridomrymex purpureus. When retested in the Y-shaped ol- Table 1. — Results from experiments on effects of chemical cues in soil on Habrocestum pulex. A. Behavior and posture. B. Attention to optical cues. Each spider tested one day on treated soil (had been in contact with ants) and on alternate day on control soil (had not been in contact with ants). Compared to when on control soil, H. pulex on treated soil: A. performed more agitated walking, undirected leaping, holding body raised and perching on wall. B. More often initiated and completed stalking. See text for details. Data analysis: McNemar test for significance of changes (for these tests, only the first two columns of data are used). Experi- ment Response On treated soil only On control soil only On both types of soil On neither type of soil McNemar test A Agitated walking 8 1 9 2 P < 0.05 Undirected leaping 12 1 2 4 P < 0.01 Holding body raised 12 0 4 4 P < 0.01 Perching on cork 11 1 5 3 P < 0.01 B Initiate stalking 11 1 7 1 P < 0.01 Complete stalking 12 2 4 2 P < 0.01 316 THE JOURNAL OF ARACHNOLOGY Figure 7. — Latencies (median in sec) to initiate (I) and complete (C) stalking sequence (see text for definitions) in experiment testing for effect of chemical cues in soil on attention to optical cues. Latencies when on treated soil (been in contact with ants) shorter than latencies when on control (clean) soil (Wilcoxon tests for paired comparisons, P < 0.005 for both initiating and completing stalking). factometer, test spiders moved into olfactom- eter arms which contained 6-methyl-5-hepten- 2-one more often than into the clean arms (Allan et al. 1996), implying that this ketone is at least one of the chemicals used by Ha= bronestes bradleyi to locate /. purpureus. Detecting 6-methyl-5-hepten-2-one is un- likely to be how H. pulex detects Monomo- rium antarcticum. Whether M. antarcticum uses alarm pheromones is unknown. Other myrmicine ants are known to do so, but they use another closely related ketone, 4-methyl-3- heptanone (Gabba & Pavan 1970; Holldobler & Wilson 1990), instead of 6-methyl-5-hepten- 2-one. It may be that, for myrmecophagic spi- ders and for ants, sensory systems are not nar- rowly tuned to particular ketones, but instead respond to a range of structurally related chemicals (see Tiirker 1997a, b). Perhaps, H. pulex has evolved chemoreceptors sensitive to a series of structurally related chemicals, rath- er than those secreted by any particular set of ant species. Broad-sensitivity sensors would assist H. pulex in predatory sequences against a wide range of ant species, including even New Zealand ants it would never encounter in nature. Kairomone detection appears to function not only to bring H. pulex into proximity with its prey, but also to elicit changes in behavior, body posture and locomotion that prepare H. pulex for predation on ants before an ant is Latency (sec) Figure 8. — Latency for test spiders to enter the experimental arm in olfactometer tests. Choice was between experimental arm (contained either live ants (“ant”; n = 15) or 6-methyl-5-hepten-2-one (“pheromone”; n = 11) or control arm. Instances of choosing control arm are not shown. seen. In particular, cues from ants caused H. pulex to move to higher ground (i.e., perch on corks), where its ability to detect optical cues from ants might be enhanced; and H. pulex often launched attacks on ants from elevated positions. Habrocestum pulex illustrates that the evo- lution of complex eyes and exceptionally in- tricate vision-based predatory behavior in sal- ticids is not incompatible with the evolution of kairomone-detection abilities and intricate chemical-mediated predatory behavior in myrmecophagic salticids. In salticids, a vi- sion-based perceptual and behavior system ap- pears to have only minimal, if any, cost to proficiency at using a chemical-based percep- tual and behavior system (Jackson & Pollard 1996, 1997). In H. pulex, the ways in which chemoreception influences predatory behavior are as intricate as those known for any non- salticid spider. Independently of optical cues, H. pulex not only appears to use kairomones for locating and preparing to prey on ants. Kairomones also appear to influence attention to optical cues. When ant-derived cues were present, H. pulex located ants faster than when they were absent. This suggests that the chem- ical and vision-based perceptual systems of salticids may have reached a remarkable level of integration. CLARK ET AL.— CHEMICAL CUES FROM ANTS 317 ACKNOWLEDGMENTS We thank Simon Pollard, David Blest, Duane Harland and Philip Taylor for com- ments on the manuscript. This research was partly funded by a grant from the New Zea- land Marsden Fund (UOC512). Voucher spec- imens of the ants and spiders used are depos- ited at the Florida State Collection of Arthropods (RO. Box 147100, Gainesville, Florida, USA) and the Canterbury Museum (Christchurch, New Zealand). LITERATURE CITED Allan, R.A., M.A. Elgar & R.J. Capon. 1996. Ex- ploitation of an ant chemical alarm signal by the zodariid spider Habronestes bradleyi Walcken- aer. Proc. R. Soc. B., 263:69-73. Attygalle, A.B. & E.D. Morgan. 1985. Ant trail pheromones. Adv. Insect Physiol., 18:1-30. Blest, A.D., D.C. O’Carroll & M. Carter. 1990. Comparative ultrastructure of layer 1 receptor mosaics in principal eyes of jumping spiders: The evolution of regular arrays of light guides. Cell Tissue Res., 262:445-460. Blum, M.S. 1966. The source and specificity of trail pheromones in Termitopone, Monomorium and Huberia, and their relation to those of some other ants. Proc. R. Entomol. Soc. London, Ser. A (Gen. Entomol.), 41:155-160. Blum, M.S. 1981. Chemical Defenses of Arthro- pods. Academic Press, New York. 562 pp. Blum, M.S. 1992. Ant venoms: Chemical and pharmacological properties. J. Tox. Toxin Rev., 11:115-164. Blum, M.S. & G.N. Ross. 1965. Chemical releas- ers of social behavior. V, Source, specificity, and properties of the odour trail of Tetramorium gui- neense (E) (Formicidae: Myrmicinae). J. Insect Physiol., 11:857-868. Bolton, B. 1987. A review of the Solenopsis genus group and review of the afro-tropical Monomo- rium Mayr (Hymenoptera, Formicidae). Bull. British Mus. Nat. Hist. (Entomol.), 54:263-452. Brown, W.L., T. Eisner & R.H. Whittaker. 1971. Allomones and kairomones: Transpecific chemi- cal messengers. BioScience, 20:21-22. Clark, R.J. & R.R. Jackson. 1994a. Self recogni- tion in a jumping spider: Portia labiata females discriminate between their own draglines and those of conspecifics. Ethol. Ecol. EvoL, 6:371- 375. Clark, R.J. & R.R. Jackson. 1994b. Portia labiata, a cannibalistic jumping spider, discriminates be- tween own and foreign eggsacs. Int. J. Comp. Psychol., 7:38-43. Clark, R.J. & R.R. Jackson. 1995a. Dragline-me- diated sex recognition in two species of jumping spiders (Araneae, Salticidae) Portia labiata and Portia fimbriata. Ethol. Ecol. EvoL, 7:73-77. Clark, R.J. & R.R. Jackson. 1995b. Araneophagic jumping spiders discriminate between the drag- lines of familiar and unfamiliar conspecifics. Ethol. Ecol. EvoL, 7:185-190. Crane, J. 1949. Comparative biology of salticid spiders at Rancho Grande, Venezuela. Part IV. An analysis of display. Zoologica, 34:159-214. Cushing, RE. 1997. Myrmecomorphy and myr- mecophily in spiders: A review. Florida Ento- moL, 80:165-193. Cutler, B. 1980. Ant predation by Habrocestum pu- lex (Hentz) (Araneae: Salticidae). Zool. Anz., 204:97-101. Drees, O. 1952. Untersuchungen fiber die ange- borenen Verhaltensweisen bei Springspinnen (Salticidae). Z. TierpsychoL, 9:169-207. Duffield, R.M., J.M. Brand & M.S. Blum. 1977. 6- Methyl-5-Hepten-2-one in Formica species: Identification and function as an alarm phero- mone (Hymenoptera: Formicidae). Ann. Ento- mol. Soc. America., 70:309-310. Edmunds, M. 1974. Defence in Animals: A Survey of Antipredator Defences. Longmans, New York. 357 pp. Edwards, G.B., J.R Carroll & W.H. Whitcomb. 1974. Stoidis aurata (Araneae: Salticidae), a spi- der predator of ants. Florida Entomol., 57:337- 345. Elgar, M.A. 1993. Inter- specific association involv- ing spiders: Kleptoparasitism, mimicry and mu- tualism. Mem. Queensland Mus., 33:411-430. Ettershank, G. 1966. A generic revision of the world myrmicinae related to Solenopsis and Pheidologetron (Hymenoptera: Formicidae). Australian J, ZooL, 14:73-171. Forster, L.M. 1982. Vision and prey-catching strat- egies in jumping spiders. American Sci., 70:165- 175. Gabba, A. & M. Pavan. 1970. Researches on trail and alarm substances in ants. Pp. 161-203, In Advances in Chemoreception. Vol. 1. (J.W John- ston, D.G. Moulton & A.I. Turk, eds). Appleton- Century Crofts, New York. Hangartner, W 1967. Spezifitat und Inaktivierung des Spurpheromons von Lasius fuliginosus (Latr.) und Orientierung der Arbeiterinnen im Duftfeld. Z. Vergl. Physiol., 57:103-136. Holldobler, B. & E.O. Wilson. 1990. The Ants. Springer Verlag, Berlin. 732 pp. Jackson, R.R. 1987. Comparative study of releaser pheromones associated with the silk of jumping spiders (Araneae, Salticidae). New Zealand J. Zool., 14:1-10. Jackson, R.R. & S.E.A. Hallas. 1986. Comparative biology of Portia africana, P. albimana, P. fim- briata, P. labiata, and P. schultzi, araneophagic web-building jumping spiders (Araneae: Saltici- 318 THE JOURNAL OF ARACHNOLOGY dae): Utilisation of webs, predatory versatility, and intraspecific interactions. New Zealand J. Zool., 13:423-489. Jackson, R.R. & A. van Olphen. 1991. Prey-cap- ture techniques and prey preferences of Cory- thalia canosa and Pystira orbiculata, ant-eating jumping spiders (Araneae, Salticidae). J. ZooL, London, 223:577-591. Jackson, R.R. & A. van Olphen. 1992. Prey-cap- ture techniques and prey preferences of ChrysiU la, Natta and Siler, ant-eating jumping spiders (Araneae, Salticidae) from Kenya and Sri Lanka. J. Zool., London, 227:163-170. Jackson, R.R. & S.D. Pollard. 1996. Predatory be- havior of jumping spiders. Ann. Rev. EntomoL, 41:287-308. Jackson, R.R. & S.D. Pollard. 1997. Jumping spi- der mating strategies: Sex among cannibals in and out of webs. Pp. 340-350, In Mating Sys- tems in Insects and Arachnids. (J. Choe & B. Crespi, eds.). Cambridge Univ. Press, Cam- bridge. Jackson, R.R., D. Li, A.T. Barrion & G.B. Edwards. 1998. Prey-capture techniques and prey prefer- ences of nine species of ant-eating jumping spi- ders (Araneae: Salticidae) from the Philippines. New Zealand J. Zool., 25:249-272. Kaston, B.J. 1936. The senses involved in the courtship of some vagabond spiders. Entomol. Americana, 16:97-160. Land, M.E 1969a. Structure of the retinae of the principal eyes of jumping spiders (Salticidae: Dendryphantinae) in relation to visual optics. J. Exp. Biol., 51:443-470. Land, M.E 1969b. Movements of the retinae of jumping spiders (Salticidae: Dendryphantinae) in response to visual stimuli. J. Exp. Biol., 51:471- 493. Li, D. & R.R. Jackson. 1996. Prey-specific capture behaviour and prey preferences of myrmeco- phagic and araneophagic jumping spiders. Rev. Suisse Zool., hors serie, pp. 423-436. Li, D., R.R. Jackson & B. Cutler. 1996. Prey-cap- ture techniques and prey preferences of Habro- cestum pulex, an ant-eating jumping spider (Ar- aneae, Salticidae) from North America. J. Zool. London, 240:551-562. Li, D., R.R. Jackson & D.P. Harland. 1999. Prey- capture techniques and prey preferences of Ae- lurillus aeruginosas, A. cognatus and A. kochi. ant-eating jumping spiders (Araneae: Salticidae) from Israel. Israel J. ZooL, 45:341-359. Mackay, W.P. 1982. The effect of predation of western widow spiders (Araneae: Theridiidae) on harvester ants (Hymenoptera: Formicidae). Oec- ologia, 53:404-411. Millot, J. 1946. Sens chemiques et sens visuel chez les araignees. Annee BioL, 22:1-21. Nyffeler, M., D.A. Dean & W.L. Sterling. 1988. The southern black widow, Lactrodectus mac- tans (Araneae, Theridiidae) as a predator of the red imported fire ant, Solenopsis invicta (Hy- menoptera, Formicidae), in Texas cotton fields. J. Appl. Entomol., 106:52-57. Oliveira, PS. & I. Sazima. 1985. Ant-hunting be- haviour in spiders with emphasis on Strophius nigricans. Bull. British Arach. Soc., 6:309-312. 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- Verlag, Ber- lin. Richman, D.B. & R.R. Jackson. 1992. A review of the ethology of jumping spiders (Araneae, Sal- ticidae). Bull. British Arach. Soc., 9:33-37. Robinson, S.W., J.L. Moser, M.S. Blum & E. Amant. 1974. Laboratory investigations of the trail-following responses of four species of leaf- cutting ants with notes on the specificity of a trail pheromone of Atta texana (Buckley). Insectes Sociaux, 21:87-94. Sokal, R.R. & EJ, Rohlf. 1995. Biometry. Free- man, San Francisco. 887 pp. Tiirker, L. 1997a. A topological study of certain chemicals eliciting the alarm response in Irido- myrmex pruinosus. Turkish J. Biol., 21:149-152. Tiirker, L. 1997b. Topological effects on the activ- ity of compounds acting as alarm pheromones of Pogonomyrmex badius. Turkish J. Biol., 21:153- 157. Willey, M.B. & R.R. Jackson. 1993. Olfactory cues from conspecifics inhibit the web-invasion behavior of Portia, web-invading araneophagic spiders (Araneae: Salticidae). Canadian J, Zool., 71:1415-1420. Wray, J. 1670. Concerning some un-common ob- servations and experiments made with an acid juyce to be found in ants. Phil. Trans. R. Soc. London, 5:2063-2066. Manuscript received 12 May 1999, revised 15 De- cember 1999. 2000. The Journal of Arachnology 28:319-328 LIFE HISTORY OF PARDOSA MOESTA AND PARDOSA MACKENZIANA (ARANEAE, LYCOSIDAE) IN CENTRAL ALBERTA, CANADA Christopher M* Buddie: Department of Biological Sciences, University of Alberta, Edmonton, Alberta T6G 2E9 Canada ABSTRACT. The density, fecundity, and life-cycle of Pardosa moesta Banks 1892 and Pardosa mack- enziana (Keyseriing 1877) were studied in a deciduous forest in central Alberta, Canada. Density estimates were lower than reported for other Pardosa species; they ranged from 0.46 per m^ for male P. mackenziana to 2.99 per nP for immature P. mackenziana. Adult female densities were below 1 per m^ for both species. Clutch sizes were highly variable and averaged (± SE) 33.06 ± 1.29 for P. moesta and 48.37 ± 1.67 for P. mackenziana. Although clutch size was positively related to female size, little of the variation was adequately explained by female size alone. Several lines of evidence suggest that P. moesta and P. mack- enziana require two years to mature in central Alberta, with a peak reproductive period in May and June. Females carry egg sacs into the summer months and immature spiders overwinter following the first growing season when they are still less than 5 mg in weight. After a second summer of growth, sub- adults overwinter and maturation occurs early in the spring. Keywords: Lycosidae, density, fecundity, phenology, life-cycle Wolf spiders in the genus Pardosa C.L. Koch 1847 are among the most conspicuous and abundant of the ground-dwelling spiders. However, little is known about the life history of many northern species of this genus in North America, even though 46 species are found in Canada, at least eight of which are distributed widely across the country (Don- dale & Redner 1990). Two of these species, Pardosa moesta Banks 1892 and Pardosa mackenziana (Keyseriing 1877), have been noted as being among the most abundant wolf spiders collected in deciduous forests of north-central Alberta (Buddie et al. 2000). Significant progress has been made in un- derstanding the ecology and biology of many Pardosa species in Europe, Japan, and south- ern latitudes in North America (e.g., Hallander 1967; Vlijm & Kessler-Geschiere 1967; Mi- yashita 1968, 1969; Edgar 1971a, b, 1972; Dondale 1977; Greenstone 1980; Graze et al. 1989; Samu et al. 1998). It is conunonly thought that most spiders living in temperate zones have annual life-cycles (Gertsch 1979), and this is true for many Pardosa from vari- ous regions including Europe, southern Can- ada, and the United States (Vlijm & Kessler- Geschiere 1967; Schmoller 1970; Dondale 1977; Graze et al. 1989). However, several Pardosa species studied from high altitudes, northern latitudes, and under cooler conditions require more than one year to complete their development (Leech 1966; Schmoller 1970; Edgar 1971b). Gther characteristics such as natural densi- ties of Pardosa species and estimates of clutch size are known for many species in the United States and some regions in Canada (e.g., Ea- son 1969; Schmoller 1970; Dondale 1977; Lowrie & Dondale 1981), and various species from Europe (e.g., Edgar 1971b; Kessler 1971). For example, Dondale (1977) reported densities of P. saxatilis (Hentz 1844) between 0. 8-4.4 per m^ in southern Gntario; and as part of a detailed study of P. lugubris (Wal- ckenaer 1802) in Scotland, Edgar (1971b) re- ported densities of various life stages between 1.7-6. 2 per m^. There is also a variety of pub- lished records on the average clutch size for many Pardosa species, and these range from as low as 25.5 eggs/female for the small spe- cies P. saxatilis to a high of 82.0 eggs/female for the larger P. amentata (Clerck 1757) (Mar- shall & Gittleman 1994). During 1998 and 1999 I studied life history characteristics of P. moesta and P. macken- ziana. The objectives were to determine the natural densities of these species, establish 319 320 THE JOURNAL OF ARACHNOLOGY their clutch sizes and assess whether the num- her of offspring is determined by female size, and to ascertain the life cycles of P. moesta and P. mackenziana in deciduous forests of central Alberta, Canada. METHODS Study site and species descriptions. — This work was done at the George Lake Field Site located 75 km northwest of Edmonton, Al- berta (ca. 53°57'N, 114°06'W). There are ap- proximately 180 ha of continuous hardwood forest at the field site, which is surrounded by agricultural land to the south and west, a lake to the east, and more than 500 ha of contin- uous deciduous forest to the north. Dominant tree species include trembling aspen {Populus tremuloides Michx.), balsam poplar {Populus balsamifera L.), birches (Betula papyrifera Marsh, and B. neoalaskana (Sarg.)), and patches of white and black spruce {Picea glauca (Moench) Voss and P. mariana (Mill.) BSP). The study area for this research was a 2.2 ha area of upland aspen forest (Niemela et al. 1992). Pardosa moesta and P. mackenziana are among the most abundant wolf spiders found on the forest floor at George Lake; other Par- dosa species encountered less frequently in- clude P. xerampelina (Keyserling 1877), P. fuscula (Thorell 1875), P. distincta (Blackwall 1846) and P. ontariensis Gertsch 1933. Par- dosa moesta has general habitat affinities in- cluding meadows, hayfields, marshes, bogs, lawns, gravel pits, clear-cuts, rocky shores, and deciduous forests (Wolff 1981; Dondale & Redner 1990; Buddie et al. 2000). Lowrie (1973) has suggested that western populations of P. moesta occur more often in wet habitats from various elevations. Pardosa macken- ziana is usually associated with coniferous forests although known to inhabit salt marsh- es, bogs, beaches, and deciduous forests (Lowrie 1973; Dondale & Redner 1990; Bud- die et al. 2000). In a study of spider assemblages in north- central Alberta, Buddie et al. (2000) found that P. moesta and P. mackenziana co-occur in a variety of different age-classes of decid- uous forest stands. The proportions of the two species, however, differed depending on whether the forest stand had a closed canopy. In open stands, 67.3% of the total catch of the two species was P. moesta. In closed canopy stands the situation was reversed as 67.8% of the total catch of the two species was P. mack- enziana. Pardosa moesta and P. mackenziana are easily distinguished in the field based on their size and coloration. Pardosa moesta is the smaller of the two species, with an average length of 4.95 mm for males and 5.64 mm for females, whereas the average length for P. mackenziana is 5.91 mm for males and 6.85 mm for females (Dondale & Redner 1990). Adult and sub-adult male P. moesta have a dark, shiny carapace in contrast to the lighter brown carapace with its lighter median band on male and female P. mackenziana. Female P. moesta have a dark carapace with faint me- dian and submarginal bands. It is also possible to distinguish immature stages of the two spe- cies based on subtle difference in the colora- tion and patterns on the carapace; immature P. mackenziana have distinct white setae that outline a V-pattem on the median region of the carapace. The white setae on the carapace of immature P. moesta are arranged in a more scattered pattern on the carapace. Additional- ly, the carapace of inunature P. mackenziana is a deeper brown color than the carapace of immature P. moesta. Voucher specimens of both species are deposited in the Strickland Entomological Museum, University of Alber- ta, Edmonton, Alberta, Canada. Density estimates. — Densities of Pardosa were estimated by haphazardly placing on the forest floor an upright bucket (28 cm diameter, 23 cm height) with its bottom removed. After the bucket was firmly placed on the forest floor the enclosed leaf-litter was searched for wolf spiders (similar to quadrat sampling used by Edgar (1971a, b)). Individual Pardosa were identified, counted and brought to the laboratory. This procedure was repeated 241 separate times between 23 April-15 June 1999. Three different life-stages were classi- fied for both species: immatures, sub-adults, males and females. Results from the bucket estimates were extrapolated to the number of Pardosa of different life stages per m^ of for- est floor, separated into three sampling periods of approximately equal lengths (23 April- 10 May, 11-27 May and 28 May-15 June). Fecundity. — Female P. moesta and P. mackenziana carrying egg sacs, or those ap- pearing to be gravid (i.e., found with swollen abdomens), were collected on an opportunistic BUDDLE— PA/?Z)05A LIFE HISTORY 321 basis during the spring of 1998 and 1999 in order to assess fecundity and relationships be- tween female size and clutch size. Many Par- dosa species are known to produce more than one egg sac in a given season (Miyashita 1969; Edgar 1971b; Wolff 1981). However, all collections were made early in the season, en- suring that catches did not contain females with second egg sacs, which are known to contain fewer eggs (Miyashita 1969; Edgar 1971a). Data about size and fecundity were collected for a total of 66 P. moesta and 73 P. mackenziana. Live females were gently held between a piece of soft foam and a clear plastic petri dish and their carapace width (CW) was measured to the nearest 0.01 mm using an ocular micrometer. CW is easily measured and thought to be a good indicator of overall spider size, as has been shown for both web-building and hunting spiders (Hags- trum 1971; Spiller & Schoener 1990; Wise & Wagner 1992; Zinmiermann & Spence 1992). Spiders were held at 25 °C under long-day photoperiod (16 h light: 8 h dark) in clear film canisters with moistened plaster-of-Paris on the bottom to maintain humidity (similar to the procedure outlined by Wise & Wagner (1992)). Many females produced egg sacs in captivity, and for most female spiders, spider- lings were allowed to hatch to determine clutch size. Due to time constraints, however, some of the specimens were placed immedi- ately in 70% ethanol, and egg sacs were later dissected for measures of clutch size. Linear regression was used to assess the relationships between female size and number of offspring produced for each species. Life cycle.^ — Adult population dynamics: The activity of adult wolf spiders can be as- sessed by using a sampling technique such as pitfall trapping. Pitfall trap catches depend on spider activity so absolute density estimates, for example, are not possible with such data. However, pitfall trap data can be used to infer the peak reproductive period for spiders as during this time male and female spider activ- ity increases. In the present study, data gen- erated from live-trapping and mark recapture using pitfall traps are used only to assess the activity of adult Pardosa, the peak reproduc- tive period and the duration of female surviv- al. This work was completed from May to Au- gust 1998 using enclosures previously used for experiments with ground beetles (see Nie- mela et al. (1997)). Enclosures were located 50-60 m from the area where density esti- mates were obtained. Three sets of enclosures measuring 4 X 24 m in length (subdivided into six compartments per enclosure, each measuring 4 X 4 m) were made in 1989 by sinking % inch (ca. 2.0 cm) plywood 30 cm into the ground, leaving 40-45 cm above ground. All seams were sealed with caulking and a strip of aluminium flashing 10 cm wide was screwed or nailed to the top part of the walls. Experiments were designed based on the assumption that Pardosa species would be unable to move between compartments. How- ever, both P. moesta and P. mackenziana were observed climbing between compartments; nevertheless, it was still possible to monitor the population dynamics of adult wolf spiders within the enclosures and to assess the length of female survival. Eight pitfall traps without preservative were placed in each of the 18 compartments. Traps were 1 liter plastic containers sunk into the ground so that the trap lip was flush with the substrate. Funnels were placed in the traps to prevent spiders from escaping. Traps were opened and monitored three to four times per week from early May until mid-July and about once per week until the end of August. I re- corded the sex of captured P. moesta and P. mackenziana, and recorded whether females carried egg sacs. Sixty P. moesta and 48 P. mackenziana fe- males carrying egg sacs were marked and re- leased on 16 June into the aforementioned compartments. Spiders were marked with a small dot of enamel paint on the carapace. A small hole was drilled into a petri dish that was placed over the spider being gently held on a piece of foam; females were maneuvered on the foam pad so their carapace was directly below the hole and a toothpick dipped in paint was inserted through the hole to place paint on the carapace. Marked females were moni- tored along with other live trap catches in the compartments in order to estimate how long individual P. moesta and P. mackenziana fe- males survive in the field. Juvenile growth and development: Under- standing population dynamics of adult spiders is insufficient for an adequate understanding of life-cycles; additionally, it is essential to determine the growth of juvenile spiders through the course of the summer and to es- 322 THE JOURNAL OF ARACHNOLOGY Table 1 . — Density (number per m^) of immature (IM), sub-adult (S A), male, and female Pardosa moesta and P. mackenziana obtained from 241 samples (14.94 total sampling area) between 23 April- 15 June 1999. Sample period Area , (mri Pardosa moesta Pardosa mackenziana IM SA 6 $ IM SA 3 9 23 April- 10 May 4.4 2.73 0.68 0 0.23 2.04 1.13 0 0 11 May-27 May 6.2 1.29 111 0.48 0.81 2.42 0.48 0 0.65 28 May- 15 June 4.3 1.61 0 0.92 1.61 2.99 0 0.46 1.15 Average 1.87 0.81 0.47 0.88 2.48 0.54 0.15 0.60 tablish the overwintering stage. As part of an experiment investigating the competitive in- teractions between P. moesta and P. mack- enziana, a number of newly dispersed spider- lings were released into small arenas. The arenas were white buckets, with the bottoms removed, measuring 28 cm in diameter and 23 cm in height. The buckets were sunk 5-7 cm into the ground on 8 July 1998 and covered with fine mesh to prevent inunigration and enfigration. Newly dispersed spiderlings were obtained from female P. moesta and P. mack- enziana used for fecundity estimates. Spider- lings from more than 10 females of each spe- cies were bulk weighed in groups of ten. A total of 237 P. moesta and 234 P. macken- ziana was placed in 12 arenas between 13 July-21 July 1998. In September 1998, the leaf litter from within the arenas was sifted and searched for Pardosa specimens. These were weighed and then immediately returned to the arenas. As soon as the snow melted in the spring of 1999, the litter within the arenas was searched a final time and Pardosa were counted and weighed. Spring cohorts: To better understand what life-stages of P. moesta and P. mackenziana overwinter in central Alberta, the specimens retained from the density estimates were weighed to the nearest mg. Some additional specimens were collected on an opportunistic basis through until 30 June 1999 to increase the sample size for these estimates. It was as- sumed that if these species are annual, only one weight class of individuals would be pre- sent following the overwintering period. Spe- cies requiring two years to complete devel- opment should show two size classes of individuals at the time of spring emergence, and three size classes of individuals during the reproductive period (Dondale 1961; Edgar 1972). RESULTS Density estimates.— A total of 117 P. moesta and P. mackenziana was counted dur- ing the 241 density estimate samplings. Im- mature Pardosa represented the most fre- quently encountered spiders and had the highest density estimates during most sam- pling periods (Table 1). Densities of sub- adults were highest between 23 April-27 May and decreased in the final sampling period; adults increased in density in the last two sam- pling periods (Table 1). Males of both species were encountered infrequently during the sur- vey and thus their density estimates were low in comparison to other life stages (Table 1). Female densities averaged 0.88 per m^ for P. moesta and 0.60 per m^ for P. mackenziana. Fecundity. — Pardosa moesta was the smaller of the two species with a mean (± SE) CW of 2.07 ± 0.02 mm, and its average clutch size was 33.06 ± 1.29 eggs or spider- lings per egg sac. Pardosa mackenziana had an average CW of 2.73 ± 0.02 mm and a mean clutch size of 48.37 ± 1.67, Both spe- cies showed a significantly positive relation- ship between female size and clutch size using linear regression (Fig. lA, B). However, very little of the variation in clutch size was ex- plained by female size as indicated by the low values (especially for P, mackenziana (Fig. IB)). Life cycle.— Adult population dynamics: Live-trapping data show that male and female P. moesta were most active in mid-May and early June (Fig. 2A). Peak activity of P. mack- enziana males and females was slightly later; they were most frequently caught between late May and mid-June (Fig. 2B). Spider activity is known to vary with temperature (Dondale & Binns 1977). The high variation in live catches of adult Pardosa in May and June was BUDDLE— PAi?Z)05A LIFE HISTORY 323 A Pardosa moesta B Pardosa mackenziana Carapace width (mm) Figure 1 . — Linear regression of clutch size (num- ber of spiderlings or eggs per egg sac) against car- apace width (mm) for Pardosa moesta, n = 66 (A) and P. mackenziana, n = 73 (B). partially explained by variation in the mean daily temperatures during the spring (temper- atures were obtained from a weather station at Sion, Alberta, 14 km south-west of the George Lake Field Site); warm days often correspond- ed to peaks in adult Pardosa activity (Fig. 2). Females carrying egg sacs were caught from 3 June“^25 August for P. moesta and from 21 May-25 August for P. mackenziana (Fig. 2). Therefore, spiderlings could be active from late spring and into the autumn months for both species. The late season catches of females carrying egg sacs likely corresponded to the production of a second egg sac. Marked females were released on 16 June, and individuals of both species were re-cap- tured at various times throughout the summer (Fig. 2). Two marked P. moesta were re-cap- tured on 1 1 August, showing that females live at least 56 days in the field after being col- lected, marked and returned to the enclosures A Pardosa moesta Figure 2. — Number of Pardosa moesta (A), and Pardosa mackenziana (B) collected by live trapping in enclosures between 11 May-29 August 1998. Solid lines represent catches of males, dashed lines are females. Solid circles (•) are females carrying egg sacs, open circles (o) are re-captures for marked females (released 16 June, solid arrow). Open tri- angles (A) are mean daily temperatures (°C) for May and June. on 16 June. Female P. mackenziana were not found in the enclosures as long as P. moesta', the latest re-capture for P. mackenziana was 21 July, 35 days after release. Juvenile growth and development: Spider- lings released into arenas at the beginning of this experiment (13 July-21 July) had an av- erage weight of 0.45 ± 0.03 mg for P. moesta and 0.58 ± 0.04 mg for P. mackenziana. Weights in September were 1.30 ± 0.25 mg for P. moesta {n = 29), and 1.28 ± 0.12 mg for P. mackenziana (n = 34). Thus, Pardosa moesta spiderlings gained on average 2.8 times their weight, and P. mackenziana 2.2 times their weight between mid-July and Sep- tember 1998. Leaf-litter from the arenas was 324 THE JOURNAL OF ARACHNOLOGY 14 May-30 June 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 Weight class (mg) Figure 3. — Frequency of Pardosa moesta (open bars) and Pardosa mackenziana (solid bars) by weight class (mg) collected from 23 April-30 June 1999. Period of spring emergence given as 23 April- 13 May, period of reproduction given as 14 May-30 June. Solid arrow indicates average weight for adult Pardosa moesta, dashed arrow indicates average weight for adult Pardosa mackenziana. sifted and searched again on 23 April 1999; four P. moesta with an average weight of 1.25 ± 0.25 mg and five P. mackenziana with an average weight of 1.40 ± 0.25 mg overwin- tered in the arenas. Although the arenas only approximated natural conditions, some spiders survived the winter and did not gain weight between September 1998 and April 1999. Spring cohorts: Spider weights from indi- viduals retained from the density estimates in- dicate that two life stages, and a few larger individuals, were present immediately follow- ing winter (23 April-13 May) (Fig. 3). During the peak reproductive period (13 May-30 June), three life stages were present for both species (Fig. 3). The majority of specimens collected fell below the average weight of adult specimens (Fig. 3). Although adult P. moesta and P. mackenziana showed a peak in activity from mid-May until late June, which would correspond to the reproductive period Figure 4. — Generalized life cycle of Pardosa moesta and Pardosa mackenziana in deciduous for- ests of central Alberta, Canada showing immatures (IM), sub-adults (SA), males, females, reproductive period (R), and overwintering period (OW). (Fig. 2.), many smaller instars (i.e., < 5 mg in size) of both species were also present on the forest floor during this time (Fig. 3). Taken together, results from the population dynamics of adults, juvenile growth and de- velopment, and from the weight classes of spring cohorts suggest P. moesta and P. mack- enziana take two years to complete develop- ment in central Alberta. Both species have the same generalized life-cycle (Fig. 4); the only notable difference in life-cycles between the two species is that P. moesta has an earlier reproductive period than P. mackenziana (Fig. 2). Both species appear to have at least two overwintering periods: one as immatures and one as sub-adults (Fig. 4). DISCUSSION Density. — Densities of sub-adult and adult P. moesta and P. mackenziana were below 2.0 per m^ during all sampling periods and im- mature densities were all below 3.0 per m^; these estimates were lower than has been re- ported for other species of Pardosa. In Scot- land, for example, Edgar (1971b) reported im- mature P. lugubris densities as high as 6.2 per m^ for shaded areas in the spring. However, immature P. lugubris were also found to have low densities in clearings (Edgar 1971b); dif- ferent life-stages of Pardosa may utilize dif- ferent habitats and their densities would thus vary depending on habitat type. Immature P. BUDDLE— /"AjRDOM LIFE HISTORY 325 iugubris move from clearings to overwinter- ing areas in the autumn, and female P. iugub- ris carrying egg sacs may search for open ar- eas in which to sun their egg sacs and deposit their young (Edgar 1971a, b). Adult P. moesta are known to attain high populations in open, grassy regions (e.g., Dondale & Redner 1990; Buddie et al. 2000), which are common in the agricultural landscape within 100-“200 m of the George Lake study area. Although P. moesta can certainly maintain populations in a closed canopy deciduous forest, densities of this species may be higher in more open hab- itats. Similarly, P. mackenziana may have higher densities in coniferous forests where this species is reported to be most commonly collected (Dondale & Render 1990). Although densities of immature P. moesta and P. mackenziana remained between 1.29- 2.99 per m^ during all three sampling periods at George Lake, sub-adult and adult densities varied more dramatically by sampling period. Sub-adult densities of both species decreased as spring progressed as sub-adults molted to sexually mature adults during the peak repro- ductive period from mid-May to late June. Male densities were low for both species, which may reflect their higher mobility; males may have been better able to escape when the bucket was placed on the forest floor. Fecundity .-^Measures of both female spi- der size and fecundity varied considerably in P. moesta and P. mackenziana. Overall, how- ever, both species were substantially larger than the average for Canada. Pardosa moesta has previously been reported as having an av- erage CW (± 1 SD) of 1.91 ± 0.14 mm (w = 20) and the average CW for P. mackenziana has been reported as 2.55 ± 0.17 mm {n = 136) (Lowrie & Dondale 1981; Dondale & Redner 1990). The clutch size of 48.37 for P. mackenziana is close to the estimate of 50 re- ported by Lowrie & Dondale (1981) but was substantially lower than the estimate of 57.5 provided by Schmoller (1970) for alpine pop- ulations in Colorado. In general, the average female size of a spi- der species is positively correlated with av- erage clutch size (Marshall & Gittleman 1994). Using data from Schmoller (1970), Lowrie & Dondale (1981), Marshall & Gittle- man (1994), and unpublished data for P. xer- ampelina, I used linear regression to assess the strength of this relationship for Pardosa Figure 5. — Linear regression of clutch size against carapace width (mm) for 14 species of Par- dosa (A) using data from Schmoller (1970), Lowrie & Dondale (1981), Marshall & Gittleman (1994) and C. Buddie (unpubl. data). Shaded area repre- sents the 95% confidence limits for the regression line. Solid triangle (A) represents a published esti- mate for Pardosa mackenziana (from Lowrie & Dondale (1981)). Estimates for Pardosa moesta (o) and Pardosa mackenziana (•) from George Lake are averages with one standard deviation (solid hor- izontal and vertical lines) and range (dashed hori- zontal and vertical lines) for both carapace width and clutch size. species. Using data for 14 species of Pardosa, there is a positive relationship between spe- cies size and clutch size, and close to two- thirds of the variation in clutch size is ex- plained by species size (R^ = 0.62, Fig. 5). Clutch size for P. moesta at George Lake is close to what can be expected based on its size alone. However, the estimates for P. macken- ziana from George Lake fell farther below what was expected, and out of the 95% con- fidence limits for the regression line (Fig. 5). Thus, understanding variation in fecundity de- mands more than simply an understanding of size. Within a species, however, there are strong relationships between female size and fecun- dity for both web-building and hunting spiders (e.g., Wise 1979, 1993; Enders 1976; Beck & Conner 1992; Simpson 1995). Although pos- itive relationships characterized both P. moes- ta and P. mackenziana, female size is clearly not the only determinant of fecundity. Kessler (1971) showed that food shortages can affect the number of eggs in two species of Pardosa. Furthermore, in a study of food limitation on the reproductive output of the pisaurid Dolo- 326 THE JOURNAL OF ARACHNOLOGY medes triton Walckenaer 1837, Spence et al. (1996) also showed that food limitation may be important in determining clutch size, but that these effects may vary with female size. Clutch size is dependent on the individual condition of the female; and this will vary de= pending on various factors such as environ^ mental conditions, prey availability, and hab- itat type. Life cycle. — The population dynamics of adult P. moesta and P. mackenziana were in- ferred from live pitfall trapping, a technique that depends on the activity of individual spi- ders. Focussing on these data, it appears that P. moesta and P. mackenziana follow a pat- tern typical for life histories of Pardosa in temperate zones: sub-adults must overwinter since mating occurs early in the spring, males die shortly after the reproductive period, and females carry egg sacs into the summer months (Turnbull 1966; Edgar 1971a). There were only two notable differences in adult ac- tivity between the two species: female P. moesta may live longer than P. mackenziana^ and the reproductive period for P. moesta is slightly earlier than for P. mackenziana, a finding also noted by Wolff (1981). By itself, the phenological data for adult populations could be interpreted to mean that both species have annual life-cycles. Howev- er, data about juvenile growth and develop- ment and weight classes of spring cohorts es- tablish that more than one year is required for these species to mature. The numerous small (i.e., < 5 mg) individuals of P. moesta and P. mackenziana found in the early spring would require at least one more year to complete their development. Also, P. moesta and P. mackenziana spiderlings held in outdoor are- nas did not reach the sub-adult stage in their first growing season and would require an ad- ditional overwintering period to complete their development. During the period of spring emergence, weights of immature Pardosa specimens did not fall into a single weight class but were spread over several weight classes (Fig. 3). This may reflect the different cohorts pro- duced from early (i.e., mid-May until June) compared to mid-season (i.e., late June until July) egg sacs from the previous summer. Spi- derlings dispersing from mid-season egg sacs would not have the same potential for growth and development before the onset of cooler conditions compared to spiderlings dispersing from early season egg sacs. This suggests that overwintering for immature P. moesta and P. mackenziana may be facultative rather than obligatory; immature spiderlings may over- winter at different stages in their development. However, the reproductive period for both species is early in the spring, suggesting that the second overwintering stage primarily con- sists of sub-adults. To ensure synchrony of the mating period, spiderlings from mid-season egg sacs would have to gain proportionally more size during their second summer compared to those from early season egg sacs. Pardosa may accom- plish this by altering the number of instars to reach maturity, as instar number is flexible in many spider species (e.g., Miyashita 1968; Edgar 1972; Toft 1976; Zimmermann & Spence 1998). Edgar (1971a) also showed that although the second egg sacs of P. lugubris had fewer eggs, the eggs themselves were heavier, possibly in preparation for cooler winter conditions. A small number of female P. moesta and P. mackenziana carry egg sacs much later in the season than the majority of the populations (i.e., late August, Fig. 2). Since spiderlings emerging from these egg sacs would be sub- stantially smaller than those emerging earlier in the season, it is possible that spiderlings from late season egg sacs may slow down their development and stretch their life cycle over two additional growing seasons. By im- plementing a three year life cycle, synchrony of mating would be ensured. However, be- cause only a small proportion of female P. moesta and P. mackenziana carry egg sacs in late August, it is unlikely that many individ- uals in the central Alberta populations of these species would exhibit three year life-cycles. Most egg sacs are carried in early or mid-sea- son, suggesting the majority of individuals of P. moesta and P. mackenziana have biennial life cycles. A two year life cycle for P. moesta and P. mackenziana is similar to that found for P. lugubris in central Scotland (Edgar 1971b), and for several species living at high altitudes (Schmoller 1970). Further south it is probable that P. moesta and P. mackenziana have an- nual life cycles. Schmoller (1970), for exam- ple, suggested that in high altitude regions of Colorado, P. mackenziana exhibits annual life BUDDLE— PA/^DO^A LIFE HISTORY 327 cycles. Pardosa lugubris has an annual life- cycle on the European mainland (Vlijm et al. 1963), and a biennial life-cycle in central Scotland (Edgar 1971b). The difference in life-cycle is attributed to cooler conditions in Scotland. However, Edgar (1972) also showed that the life cycle of P. lugubris in the Neth- erlands may vary from annual to biennial de- pending on environmental conditions and the timing of spiderling dispersal. A mixed an- nual-biennial life cycle has also been sug- gested for P. tesquorum (Odenwall 1901) in central Saskatchewan (D.J. Buckle unpubl. data). Another variation in Pardosa life-cycles has been shown for P. agrestis (Westring 1861) in central Europe. Here, Samu et al. (1998) report a bimodal life-history pattern, with reproductive periods in May and August. Undoubtedly, Pardosa life-cycles are remark- ably flexible, and this may aid in explaining their dominance in many terrestrial ecosys- tems. ACKNOWLEDGMENTS Thanks to Alice Graham for her outstand- ing help with field and laboratory work, and J.R. Spence for inspiration and encourage- ment. Funding was provided by the University of Alberta (Province of Alberta Graduate Fel- lowship) and the Natural Science and Engi- neering Research Council of Canada (NSERC) in the form of a post-graduate scholarship to the author and an operating grant to J.R. Spence (Department of Biologi- cal Sciences, University of Alberta). Com- ments from D.J. Buckle, I.C. Robertson and J.R. Spence greatly improved earlier drafts of the manuscript. LITERATURE CITED Beck, M.E. & E.F. Conner. 1992. Factors affecting the reproductive success of the crab spider Mis- umenoides formosipes: The covariance between juvenile and adult traits. Oecologia 92:287-295. Buddie, C.M., J.R. Spence & D.W. Langor. 2000. Succession of boreal forest spider assemblages following wildfire and harvesting. Ecography 23: 424-436. Dondale, C.D. 1961. Life histories of some com- mon spiders from trees and shrubs in Nova Sco- tia. Canadian Journal of Zoology 39:777-787. Dondale, C.D. 1977. Life histories and distribution patterns of hunting spiders (Araneida) in an On- tario meadow. Journal of Arachnology 4:73-93. Dondale, C.D. & M.R. Binns. 1977. Effect of weather factors on spiders (Araneida) in an On- tario meadow. Canadian Journal of Zoology 55: 1336-1341. Dondale, C.D. & J.H. Redner. 1990. The Insects and Arachnids of Canada. Part 17. The Wolf Spi- ders, Nursery web Spiders, and Lynx Spiders of Canada and Alaska (Araneae; Lycosidae, Pisaur- idae, and Oxyopidae). Agriculture Canada Pub- lication 1856. Eason, R.R. 1969. Life history and behavior of Pardosa lapidicina Emerton (Araneae: Lycosi- dae). Journal of the Kansas Entomological So- ciety 42:339-360. Edgar, W.D. 1971a. Seasonal weight changes, age structure, natality and mortality in the wolf spi- der Pardosa lugubris Walck. in central Scotland. Oikos 22:84-92. Edgar, W.D. 1971b. The life-cycle, abundance and seasonal movement of the wolf spider, Lycosa (Pardosa) lugubris, in central Scotland. Journal of Animal Ecology 40:303-322. Edgar, W.D. 1972. The life cycle of the wolf spider Pardosa lugubris in Holland. Journal of Zoology 168:1-7. Enders, F. 1976. Clutch size related to hunting manner of spider species. Annals of the Ento- mological Society of America 69:991-998. Gertsch, WJ. 1979. American Spiders. 2nd ed. Van Nostrand Reinhold Company, New York, New York. Greenstone, M.H. 1980. Contiguous allotopy of Pardosa ramulosa and Pardosa tuoba (Araneae, Lycosidae) in the San Francisco Bay region, and its implications for patterns of resource partition- ing in the genus. American Midland Naturalist 104:305-311. Hagstrum, D.W. 1971. Carapace width as a tool for evaluating the rate of development of spiders in the laboratory and the field. Annals of the En- tomological Society of America 64:757-760. Hallander, H. 1967. Range and movements of the wolf spiders, Pardosa chelata (O.F. Muller) and P. pullata (Clerck). Oikos 18:360-364. Kessler, A. 1971. Relation between egg production and food consumption in species of the genus Pardosa (Lycosidae, Araneae) under experimen- tal conditions of food-abundance and food-short- age. Oecologia 8:93-109. Leech, R.E. 1966. The spiders (Araneida) of Hazen Camp 80°41'N, 71°8LW Quaestiones Entomo- logicae 2:153-212. Lowrie, D.C. 1973. The microhabitats of western wolf spiders of the genus Pardosa. Entomolog- ical News 84:103-116. Lowrie, D.C. & C.D. Dondale. 1981. A revision of the nigra group of the genus Pardosa in North America. Bulletin of the American Musuem of Natural History 170:125-139. Marshall, S.D. & J.L. Gittleman. 1994. Clutch size 328 THE JOURNAL OF ARACHNOLOGY in spiders: Is more better? Functional Ecology 8: 118-124. Miyashita, K. 1968. Growth and development of Lycosa t-insignita Boes. et Str. (Araneae: Lycos- idae) under different feeding conditions. Applied Entomology and Zoology 3:81-88. Miyashita, K. 1969. Seasonal changes of popula- tion density and some characteristics of overwin- tering nymph of Lycosa t-insignita Boes. et Str. (Araneae: Lycosidae). Applied Entomology and Zoology 4:1-8. Niemela, J., J.R. Spence & D.H. Spence. 1992. Habitat associations and seasonal activity of ground-beetles (Coleoptera, Carabidae) in central Alberta. Canadian Entomologist 124:521-540. Niemela, J., J.R, Spence & H. Carcamo. 1997. Es- tablishment and interactions of carabid popula- tions: An experiment with native and introduced species. Ecography 20:643-652. Graze, M.J., A. A. Grigarick & K.A. Smith. 1989. Population ecology of Pardosa ramulosa (Ara- neae, Lycosidae) in flooded rice fields of north- ern California. Journal of Arachnology 17:163- 170. Samu, E, J. Nemeth, E Toth, E. Szita, B. BCiss & C. Szinetar. 1998. Are two cohorts responsible for the bimodal life-history pattern in the wolf spider Pardosa agrestis in Hungary? Pp, 215- 221, In Proceedings of the 17th European Col- loquium of Arachnology. (P.A. Seldon, ed.). Ed- inburgh, 1997. Schmoller, R. 1970. Life histories of alpine tundra Arachnida in Colorado. American Midland Nat- uralist 83:119-133. Simpson, M.R. 1995. Covariation of spider egg and clutch size: The influence of foraging and parental care. Ecology 76:795-800. Spence, J.R., M. Zimmermann & J.R Wojcicki. 1996. Effects of food limitation and sexual can- nibalism on reproductive output of the nursery web spider Dolomedes triton (Araneae: Pisauri- dae). Oikos 75:373-382. Spiller, D. & T.W. Schoener. 1990. Lizards reduce food consumption by spiders: Mechanisms and consequences. Oecologia 83:150-161. Toft, S. 1976. Life histories of spiders in a Danish beech wood. Natura Jutlandica 19:5-40. Turnbull, A.L. 1966. A population of spiders and their potential prey in an overgrazed pasture in eastern Ontario. Canadian Journal of Zoology 44:557-583. Vlijm, L., A. Kessler & C.J.J. Richter. 1963. The life-history of Pardosa amentata (Cl.) (Araneae, Lycosidae). Entomologische Berichten (Amster- dam) 23:75-80. Vlijm, L. & A. Kessler-Geschiere. 1967. The phe- nology and habitat of Pardosa monticola, P. ni- griceps, and P. pullata (Araneae, Lycosidae). Journal of Animal Ecology 36:31-56. Wise, D.H. 1979. Effects of an experimental in- crease in prey abundance upon the reproductive rates of two orb-weaving spider species (Ara- neae: Araneidae). Oecologia 41:289-300. Wise, D.H. 1993. Spiders in Ecological Webs. Cambridge Univ. Press, Cambridge, England. Wise, D.H. & J.D. Wagner. 1992. Evidence of ex- ploitative competition among young stages of the wolf spider Schizocosa ocreata. Oecologia 91:7- 13. Wolff, R.J. 1981. Wolf spiders of the genus Par- dosa (Araneae: Lycosidae) in Michigan. Great Lakes Entomolologist 14:63-68. Zimmermann, M. & J.R. Spence. 1992. Adult pop- ulation dynamics and reproductive effort of the fishing spider Dolomedes triton (Araneae, Pi- sauridae) in central Alberta. Canadian Journal of Zoology 70:2224-2233. Zimmermann, M. & J.R. Spence. 1998. Phenology and life-cycle regulation of the fishing spider Do- lomedes triton Walckenaer (Araneae, Pisauridae) in central Alberta. Canadian Journal of Zoology 76:295-309. Manuscript received 10 January 2000, revised 20 May 2000, 2000. The Journal of Arachnology 28:329-345 A STRUCTURED INVENTORY OF APPALACHIAN GRASS BALD AND HEATH BALD SPIDER ASSEMBLAGES AND A TEST OF SPECIES RICHNESS ESTIMATOR PERFORMANCE Douglas S. Toti^ Frederick A. Coyle, and Jeremy A. Miller^; Department of Biology, Western Carolina University, Cullowhee, North Carolina 28723 USA ABSTRACT. The current rate of species attrition necessitates the development of quick and accurate sampling protocols and species richness estimators. Four time-based and one area-based methods were used to sample spiders of a grass bald and a heath bald in the Great Smoky Mountains National Park in late spring and early fall of 1995. Eighty-four samples were collected at each site; 1853 adults and 91 species were found in the grass bald, 573 adults and 60 species in the heath bald. The data were analyzed with 11 species richness estimators: Chao & Lee 1, Chao & Lee 2, ACE, ICE, bootstrap, Chao 1, Chao 2, first-order jackknife, second-order jackknife, Michaelis-Menten runs, and Michaelis-Menten means. All but the Chao & Lee estimators generated richness estimates that clustered within a reasonable range, 106- 160 species for the grass bald and 68-90 species for the heath bald. The failure of the observed species accumulation curve to level off for our data sets showed that more sampling would be needed to determine the number of species present as adults during the two sampling seasons. Although this prevented us from rigorously testing richness estimator performance, we found that the Michaelis-Menten means estimator performed better than the other estimators when judged by two indirect criteria of good estimator perfor- mance— the estimator curve should reach an asymptote with fewer samples than are required for the observed species accumulation curve to reach an asymptote, and the estimates should be close to reason- able visual extrapolations of the asymptote of the observed species accumulation curve. We postulate that the differences we found in species richness and taxon and guild composition between the spider assem- blages of these two bald communities are, at least in part, a consequence of striking differences in the physiognomy, richness, and taxonomic composition of the plant associations of the two communities. Keywords^ Spiders, species richness, richness estimators, Appalachian balds In order to know how and where to protect biodiversity, it is imperative that we learn more about the patterns of diversity of terres- trial arthropods, which may comprise 80% or more of the Earth’s species but have too often been neglected by resource managers and con- servation planners (Wilson 1988, 1992; Kre- men et al. 1993; Colwell & Coddington 1994; Longino 1994). Spiders, which include about 36,000 described species and are estimated to number 60,000-170,000 species (Coddington & Levi 1991; Platnick 1999), comprise a sig- nificant portion of this terrestrial arthropod di- versity. Spiders are abundant and ubiquitous, ‘ Current address: Ecology Group, Leidy Labo- ratories, University of Pennsylvania, Philadelphia, Pennsylvania 19104 USA. ^ Current address: Department of Biological Sci- ences, George Washington University, Washington, DC 20052; and Department of Entomology, National Museum of Natural History, NHB-105, Smithsonian Institution, Washington, DC 20560 USA. employ a remarkable diversity of predation strategies, occupy a wide array of spatial and temporal niches, are characterized by high within-habitat taxonomic diversity, exhibit taxon- and guild-specific responses to envi- ronmental change, and are relatively easy to sample and identify. They are important reg- ulators of insect populations (Riechert & Lockley 1984; Riechert & Bishop 1990; Wise 1993) and may prove to be useful indicators of the overall species richness and health of biotic communities (Kremen et al. 1993; Col- well & Coddington 1994; Norris 1999). Coddington et al. (1991) pioneered the de- velopment of a sampling protocol and esti- mation procedure for rapid assessment of spi- der diversity at tropical forest sites. This and similar protocols can be structured to provide replicated data sets that reflect the relative abundance of species in the sites and habitats studied and may therefore provide comparable views of species richness, taxonomic compo- 329 330 THE JOURNAL OF ARACHNOLOGY sition, and guild structure across diverse com= munities and regions (Coddington et al. 1996; Silva & Coddington 1996; Dobyns 1997). Colwell & Coddington (1994) reviewed and explored the performance of richness estima- tors and emphasized the need to test these with real data sets. Balds, natural tree-less communities located on well-drained high-elevation sites below the climatic tree-line, are among “the most dis- tinctive and celebrated” communities of the southern Appalachian Mountains (Whittaker 1956; Mark 1958). Despite considerable re- search, there is no clear understanding of the factors responsible for the origin and mainte- nance of these communities (Cain 1930; Whittaker 1956; Billings & Mark 1957; Mark 1958; Stratton & White 1982). Grass balds, because of their high plant species richness, asthetic appeal, and shrinking size (due to for- est encroachment), are currently the focus of special monitoring and management efforts by resource specialists in the Great Smoky Mountains National Park (GSMNP) (Keith Langdon pers. com.). Heath balds, which are dense thickets of evergreen ericaceous shrubs on highly acidic soil, support far fewer plant species and a more homogeneous architecture than grass balds, but attract considerable at- tention because of their colorful floral dis- plays. In the current study we employ a modified Coddington protocol and eleven richness es- timator algorithms and other analytical meth- ods to provide the first estimates of the species richness and structure of spider assemblages in a grass bald and a heath bald. Additionally, we use these data sets to evaluate the perfor- mance of the richness estimators. METHODS Study sites. — The two sites are 40 km apart in the GSMNP. Gregory Bald, the grass bald site, covers a very gently rounded peak (UTM grid coordinates: E2400, N39343) and ranges from 1490-1510 m elevation. This bald cov- ers about 3 hectares and contains 175 vascular plant species, more than any other bald in the GSMNP (Stratton & White 1982). It consists of large open grassy areas interrupted by patches of shrubs (up to 2 m tall) and, occa- sionally near its edge, small trees (up to 15 m tall). The dominant grasses are mountain oats (Danthonia spp.) and blue grasses (Poa spp.); the dominant shrubs are blueberries (Vaccin- ium spp.), hawthorns (Cretaegus spp.) and azaleas {Rhododendron spp.). The ground sur- face is covered by thin mats of dead grasses and sedges in grassy areas and a thin layer of leaf litter below the shrubs. The heath bald site (UTM grid coordinates: E2788, N39460) covers 0.5 hectares at 1380- 1410 m elevation on the southeast-facing slope both above and below a 50 m stretch of Alum Cave Trail immediately below Inspira- tion Point (which is on a ridge extending south from Peregrine Peak). Cain (1930) found that nearby heath balds supported only 12 plant species. The heath bald at this site is a homogeneous, dense, woody mass of inter- woven ericaceous shrubs about 3-4 m tall; rhododendron {Rhododendron catawbiense) and mountain laurel {Kalmia latifoUa) domi- nate; Vaccinium, Carolina rhododendron {R. minus), greenbriar {Smilax), and sand myrtle {Leiophyllum buxifolium) are also present. The ground, virtually devoid of herbaceous plants, is covered by a thick layer of leaf litter (interrupted in places by patches of short com- pact moss) over thick, moist, spongy humus. Data collection* — Our sampling procedure included five methods chosen to access all mi- crohabitats in these two communities: aerial hand collection, ground hand collection, beat- ing, Tullgren funnel litter extraction, and sweep-netting. The first four methods were used in the heath bald; sweep-netting was sub- stituted for aerial hand collection in the grass bald due to the predominance of low vegeta- tion in that community. The aerial and ground hand collection methods are synonymous with the “looking up” and “looking down” meth- ods, respectively, of Coddington et al. (1991). Aerial sampling involves searching leaves, branches, tree trunks, and spaces in between, from knee height up to maximum overhead arm’s reach. Ground collection involves searching on hands and knees, exploring the leaf litter, logs, rocks, and plants that are be- low knee level. Beating consists of striking vegetation at any level with aim long stick and catching the falling spiders on a tray held horizontally below the vegetation. Because the dense maze of shrub branches throughout the heath bald and in some parts of the grass bald made it difficult to maneuver the standard 0.5 m^ beating sheet, we used instead a small- er (0.24 m^), rigid, rectangular (57 X 42 cm) TOTI ET AL.— SPIDER ASSEMBLAGES & SPECIES RICHNESS ESTIMATORS 331 plastic tray with a 1.5 cm high rim. The open- ing of the heavy sweep net used for sweep- netting was 0.37 m in diameter; at 225-425 sweeps per hour (mean = 327.5) and a mean sweep length of 1.4 m, an average of 49 m^ of habitat volume was sampled per hour. For all of these methods, fingers, glass vials, and aspirators were used to collect spiders into 80% ethanol. Each litter sample consisted of 1 m^ of leaf litter and underlying loose humus that was placed in a plastic bag, transported to the lab, and processed in 50-60 cm diam- eter Tullgren funnels fitted with 6-8 mm mesh screens and 60 watt light bulbs for two-four days until the litter was dry. In grassy areas of the grass bald, where much of the litter was interlocked with grass and low herbs, a long knife was used to cut away and collect thin sections of sod. Except for the litter samples, time was used to partition sampling effort into replicate sam- ples; one sample unit equals one hour of un- interrupted time during which a collector at- tempts to collect every spider encountered that is not obviously a juvenile. During any one sampling hour each of the three collectors (the authors) used only one method, but the team as a whole employed all three time-based methods in the same portion of the site. Sam- pling effort was distributed so that, in each sampling season, no area within the site was sampled more than once with a given method and nearly all of the available habitat area was sampled. Because of the density and height of the heath vegetation, the area sampled per hour was much smaller in the heath bald than in the grass bald. It should be noted that since sweeping was substituted for aerial hand sam- pling on the grass bald and since it took more time and effort to maneuver in the heath (which biases time-based samples), consider- able caution must be exercised when making between-community comparisons of the abun- dance or relative abundance of taxa. In partic- ular, these differences may bias Kulczynski’s index of similarity (see below). Night sam- pling was tried (3 one-hour samples in the grass bald), but, since the rate of capture of adults was so low (2.3 per hour), sampling was limited to daylight hours. Sampling was conducted in the spring and fall of 1995: 25- 26 May and 23-24 September in the heath bald; 3-5 June and 29 September-1 October in the grass bald. Forty- two samples (36 one- hour samples distributed equally among the three time-based methods and six litter sam- ples) were collected at each site in each sea- son. Although many juveniles end up in each sample, only adults were counted, identified, and used in analyses because identifying ju- veniles to species is often impossible. The specimens, which are being curated temporar- ily at Western Carolina University, will even- tually be deposited in the National Museum of Natural History of the Smithsonian Insti- tution. Data analysis. — The computer program Estimates (Version 5.0.1) (Colwell 1997) was used to evaluate the performance of the fol- lowing 1 1 species richness estimators with our data sets: Chao & Lee 1, Chao & Lee 2, ACE, ICE, Chao 1, Chao 2, first-order jackknife, second-order jackknife, bootstrap, Michaelis- Menten runs, and Michaelis-Menten means. The two Michaelis-Menten estimators use the same equation, a two-parameter hyperbolic function first used to describe enzyme kinet- ics, to directly extrapolate the species accu- mulation curve, but they differ in computa- tional format (Colwell & Coddington 1994; Colwell 1997). The other nine estimators are non-parametric algorithms which estimate the number of species yet-to-be-collected based on a quantification of rarity. Chao & Lee 1, Chao & Lee 2, ACE, and ICE are coverage- based richness estimators based on the statis- tical concept of sample coverage. ACE (abun- dance-based coverage estimator) (Chao et al. 1993) and ICE (incidence-based coverage es- timator) (Lee & Chao 1994) are modified ver- sions of the two Chao & Lee (1992) estima- tors, which have been found to consistently overestimate richness, especially with small samples (Colwell & Coddington 1994; Col- well 1997). Chao & Lee 1, Chao & Lee 2, ACE, and Chao 1 are all abundance-based es- timators; that is, they use abundance to quan- tify rarity (for example, the number of single- tons and doubletons, which are the number of species represented by only one or two indi- viduals in the entire data set). ICE, Chao 2, both jackknife estimators, and the bootstrap estimator are incidence-based; they rely on in- cidence (presence/absence) data to quantify rarity (for example, the number of uniques and duplicates, which are the number of spe- cies found in only one or two samples in the entire data set). We used the Coleman curve. 332 THE JOURNAL OF ARACHNOLOGY which plots the expected richness for random subsamples of the entire data set, to determine whether the samples are uniform enough to justify use of the Michaelis-Menten estimators (Colwell & Coddington 1994; Colwell 1997). See Colwell & Coddington (1994), Colwell (1997), and Chazdon et al. (1998) for descrip- tions and discussions of these estimator al- gorithms and for a demonstration of how Estimates tracks changes in each richness es- timate as samples accumulate. From a species- by-sample abundance matrix, the program se- lects a sample, calculates the richness estimates based on that sample, selects a sec- ond sample, recomputes the estimates using the data from both samples, and so on until all samples are included. By randomizing sample order (we chose 100 randomizations) and computing the mean richness estimate for each sample accumulation level, the program removes the effect of sample order and gen- erates a smoother species accumulation curve, thereby permitting closer comparison of esti- mator performance. The fact that the Coleman estimator curve was nearly identical to the species accumulation curve in all of our data sets indicates that our samples are not espe- cially heterogeneous, and that randomization of sample accumulation order is therefore jus- tified (Colwell & Coddington 1994). Using the same randomization protocol, we also plotted the mean number of singletons, uniques, doubletons, and duplicates against sample number. Percentage complementarity (Colwell & Coddington 1994), Kulczynski’s index of sim- ilarity (also called the Bray-Curtis index) (Bray & Curtis 1957), and Sorensen’s index of similarity (Kent & Coker 1992) were used to compare the taxonomic composition of the two bald communities. Percentage comple- mentarity = 100(x/y), where x = number of unique species (collected in only one com- munity or the other), and y = total number of species collected in both communities (com- bined species richness). Kulczynski’s index of similarity (K) = 2w/(a + b), where a = num- ber of individuals collected in community A, b = number of individuals collected in com- munity B, and w = sum of the lesser abun- dances for those species present in both com- munities. Sorensen’s index of similarity (S) = 2e/(c + d), where c = number of species col- lected in community A, d = number of spe- cies collected in community B, and e = num- ber of species common to both communities. Percentage complementarity is a measure of difference. Kulczynski’s index is a measure of similarity and, because it uses abundance data, emphasizes the importance of common spe- cies. Sorensen’s index, also a measure of sim- ilarity, does not emphasize the importance of common species. RESULTS A total of 2426 adult spiders representing 22 fanrnlies, 89 genera, and 128 species was present in the 168 samples collected in this study (see Table 1 for breakdowns by com- munity and season). The number of adults col- lected and the observed richness were much higher in the grass bald than in the heath and were higher in the spring than in the fall in both communities (Table 2). Sampling inten- sity, the ratio of adults to species, was higher for the grass bald than for the heath bald (Ta- ble 2). The inventory completeness index (the percentage of species that is not singletons), another indication of how well a community has been sampled, was slightly lower for the grass bald than for the heath bald (Table 2). Species richness estimates. — For none of the six sample sets (the total sample for each community and the two seasonal subsets for each community) does the mean, randomized, observed species accumulation curve reach an asymptote (Figs. 1-6), although these curves for the three heath conununity data sets (Figs. 2, 5, 6) appear to more closely approach an asymptote than do those of the corresponding grass bald data sets (Figs. 1, 3, 4). The Mi- chaelis-Menten, ICE, and Chao 2 estimator curves approach an asymptote more rapidly as sample number increases than do the other es- timator curves (Figs. 1-6). In all six data sets, the Michaelis-Menten estimate appears to ap- proach an asymptote more closely than do the other estimates. The second-order jackknife estimates climb more steeply for every data set than do the first-order jackknife estimates. The shape of the bootstrap estimator curve de- parts relatively little from the observed spe- cies accumulation curve. As predicted by Col- well (1997), the Michaelis-Menten runs estimator generated especially high and erratic richness estimates early in the curve. Since this estimator leveled off to nearly the same values as the Michaelis-Menten means, only TOTI ET AL.— SPIDER ASSEMBLAGES & SPECIES RICHNESS ESTIMATORS 333 Table L — Species collected in bald communities; numbers of adults given for spring and fall sample sets (42 samples per set). Classification follows Platnick (1997), except that linyphiids are divided into subfamilies. Guild designations (based on our collecting data and literature): AW = aerial web-builder, AH = aerial hunter, GW = ground web-builder (web in, or attached to, ground litter), GH = ground hunter; AG and GA mean, respectively, primarily aerial or primarily ground. Erigonine linyphiids and leptonetids were assigned to web-building guilds even though for many of these species it is not known whether webs are used in prey capture. Singleton status designations (based on identified GSMNP collec- tions): C = common in GSMNP (in one or more other habitats), U = apparently uncommon in GSMNP. Taxon Grass bald Spring Fall Heath bald Spring Fall Guild Single- ton status Agelenidae Agelenopsis utahana (Chamb. & Ivie) 0 3 0 4 AW Amaurobiidae Callobius bennetti (Blackw.) 0 0 0 1 AW C Cybaeopsis armipotens (Bishop & 4 0 0 0 GW Crosby) Cybaeopsis pantoplus (Bishop & Cros- 0 0 1 14 GW by) Coras montanus (Emerton) 0 0 1 2 GW Wadotes calcaratus (Keys.) 0 3 0 0 GW Wadotes hybridus (Emerton) 0 1 0 0 GW C Wadotes tennesseensis Gertsch 0 0 3 6 GW Antrodiaetidae Antrodiaetus unicolor (Hentz) 0 8 12 9 GW Anyphaenidae Wulfila saltabunda (Hentz) 0 1 0 0 AH u Araneidae Araneus trifolium (Hentz) 0 12 0 0 AW Araniella displicata (Hentz) 2 0 0 0 AW Argiope aurantia Lucas 0 2 0 0 AW Argiope trifasciata (Forskal) 0 1 0 0 AW c Eustala cepina (Walck.) 0 0 1 0 AW u Gea heptagon (Hentz) 0 3 0 0 AW Hyposinga rubens (Hentz) 13 0 0 0 AW Neoscona arabesca (Walck.) 0 3 0 0 AW Neoscona hentzi (Keys.) 0 3 0 0 AW Clubionidae Clubiona canadensis Emerton 4 0 3 0 AH Clubiona kastoni Gertsch 1 0 0 0 AH c Clubiona rhododendri Barrows 66 34 2 3 AGH Clubiona spiralis Emerton 0 0 7 10 GAH Clubiona sp. A 1 0 0 0 AH u Cybaeidae Cybaeus patritus Bishop & Crosby 0 0 0 4 GW Dictynidae Cicurina breviaria Bishop & Crosby 1 2 1 2 GW Cicurina brevis Emerton 0 0 1 0 GW c Cicurina minima Chamb. & Ivie 0 0 6 10 GW Dictyna maxima (Banks) 0 0 1 0 AW u Lathys immaculata Chamb. & Ivie 1 0 1 12 GW c Gnaphosidae Zelotes hentzi Barrows 4 0 0 0 GH 334 THE JOURNAL OF ARACHNOLOGY Table 1. — Continued. Taxon Grass Spring bald Fall Heath bald Spring Fall Guild Single- ton status Hahniidae Calymmaria persica (Hentz) 0 0 2 6 GW Cryphoeca montana Emerton 0 0 6 1 GW Neoantistea agilis (Keys.) 1 7 0 0 GW Neoantistea magna (Keys.) 2 11 5 15 GW Leptonetidae Appaleptoneta coma (Barrows) 0 0 0 1 GW C Appaleptoneta silvicultrix (Crosby & Bishop) 0 0 0 2 GW Linyphiidae Erigoninae Blestia sarcocuon (Crosby & Bishop) 4 36 2 3 GW Ceraticelus alticeps Fox 564 434 3 6 AW Ceraticeius carinatus Emerton 12 77 6 14 GW Ceratinella brunea Emerton 2 3 0 0 GW Ceratinops carolinus Banks 1 0 1 0 GW C Ceratinopsidis formosa Banks 0 0 0 1 AW C ColUnsia oxypaederotipus (Crosby) 128 2 85 0 GW Eperigone trilobata (Emerton) 1 0 0 0 GW U Eridantes erigonoides Emerton 8 0 0 0 GW Erigone autumnalis Emerton 11 2 0 0 GW Erigone brevidentata Emerton 4 0 0 0 GW Floricomus praedesignatus Bishop & Crosby 10 1 11 11 GW Gonatium crassipalpum Bryant 0 3 0 0 AW Grammonota pictilis (O.P.-Cambr.) 0 0 2 0 AW Maso sundevallii (Westring) 6 0 0 0 GW Pelecopsis moesta (Banks) 12 12 0 0 GW Pocadicnemus americana Milledge 3 0 17 0 GAW Scylaceus palUdus (Emerton) 0 7 0 0 GW Waickenaeria digitata (Emerton) 4 0 0 0 GAW Walckenaeria directa (O.R-Cambr.) 1 1 3 5 GW Waickenaeria minuta (Emerton) 0 0 3 1 GW Walckenaeria pallida (Emerton) 0 0 1 0 GW C Waickenaeria spiralis (Emerton) 4 0 0 0 GW Erigoninae sp. A 0 0 0 1 GW u Erigoninae sp. B 0 2 0 0 AW Erigoninae sp. C 0 1 0 0 AW u Linyphiinae Bathyphantes bishopi Ivie 4 50 16 2 GW Bathyphantes pallidus (Banks) 0 1 0 0 GW u Centromerus denticulatus (Emerton) 0 0 0 1 GW C Florinda coccinea (Hentz) 39 7 0 0 AW Frontinella pyramitela (Walck.) 0 0 0 1 AW C Lepthyphantes zebra (Emerton) 17 0 56 29 GW Meioneta micaria (Emerton) 8 1 2 0 GAW Meioneta semipallida Chamb. & Ivie 0 5 0 0 GAW Neriene radiata (Walck.) 0 0 9 0 AW Neriene redacta Chamb. 0 5 0 0 GAW Neriene variabilis (Banks) 0 0 0 1 GAW c Tapinopa bilineata Banks 0 0 0 1 GW u Taranucnus ornithes (Barrows) 0 0 1 1 GW TOTI ET AL.— SPIDER ASSEMBLAGES & SPECIES RICHNESS ESTIMATORS 335 Table 1. — Continued. Taxon Grass bald Spring Fall Heath Spring bald Fall Guild Single- ton status Liocranidae Phrurotimpus borealis (Emerton) 3 0 18 0 GH Scotinella sp, A 0 0 2 0 GH Lycosidae AUocosa funerea (Hentz) 0 1 0 0 GH U Arctosa virgo (Chamb.) 3 0 0 0 GH Pardosa atlantica Emerton 1 0 0 0 GH u Pardosa milvina (Hentz) 1 0 0 0 GH c Pardosa saxatilis (Hentz) 1 0 0 0 GH c Pirata hiteorum Wallace & Exline 0 6 0 0 GH Pirata montanus Emerton 1 2 0 0 GH Schizocosa bilineata (Emerton) 1 0 0 0 GH u Varacosa avara (Keys.) 1 0 0 0 GH c Nesticidae Nesticus reclusus Gertsch 0 0 2 2 GW Oxyopidae Oxyopes salticus Hentz 5 0 0 0 AH Philodromidae Philodromus montanus Bryant 2 0 0 0 AH Salticidae Eris marginata (Walck.) 0 0 0 2 AH Evarcha faicata (Clerck) 5 3 0 0 AH Ghelna canadensis (Banks) 5 4 0 0 GH Habrocestum puiex (Hentz) 1 0 2 0 AH c Habronattus coecatus (Hentz) 1 0 0 0 AH c Hentzia mitrata (Hentz) 2 0 0 0 AH Maevia inclemens (Walck.) 12 0 0 0 AH Maevia sp. A 2 1 14 4 AH Neon nellii Peckham & Peckham 17 3 21 0 GH Pelegrina montana (Emerton) 1 0 0 0 GH c Pelegrina proterva (Walck.) 18 8 0 1 AGH c Pelegrina sp. A 1 0 0 0 AH u Phidippus clarus Keys. 0 3 0 0 AH Phidippus mystaceus (Hentz) 0 1 0 0 AH u Talavera minuta (Banks) 1 0 0 0 GH u Thiodina puerpera (Hentz) 1 0 0 0 AH c Zygoballus bettini Peckham 2 0 0 0 AH Zygoballus sexpunctatus (Hentz) 1 0 0 0 AH c Tetragnathidae Tetragnatha iaboriosa Hentz 5 0 0 0 AW Zygiella dispar (Kulczynski) 0 0 0 2 AW Theridiidae Dipoena nigra (Emerton) 1 0 0 0 AW c Pholcomma barnesi Levi 0 0 1 0 GW u Pholcomma hirsuita Emerton 0 0 1 1 GW Robertus frontatus (Banks) 1 0 0 0 GW c Theridion differens Emerton 0 0 2 0 AW Theridion frondeum Hentz 0 1 0 0 AW c Theridion intervallatum Emerton 0 1 0 0 AW u 336 THE JOURNAL OF ARACHNOLOGY Table 1. — Continued, Grass bald Heath bald Single- ton Taxon Spring Fall Spring Fall Guild status Theridion lyricum Walck. 0 0 7 0 AW Theridion neshamini Levi 1 0 0 0 AW U Theridion sexpunctatum Emerton 0 0 30 0 AW Theridula opulenta (Walck.) Theridiosomatidae 23 2 0 0 AW Theridiosoma gemmosum (L. Koch) 0 0 7 0 GW Thomisidae Misumena vatia (Clerck) 0 0 1 0 AH U Misumenoides formosipes (Walck.) 0 1 0 0 AH c Misumenops asperatus (Hentz) 3 0 0 0 AH Misumenops oblongus (Keys.) 5 0 0 0 AH Ozyptila distans Dondale & Redner 0 0 0 3 GH Xysticus triguttatus Keys. 1 0 0 0 GH u Total 1072 781 378 195 Table 2. — Richness estimates and other summary values for each bald community and for each seasonal sample set from each community. Each richness estimate represents the mean (and, for some estimators, the SD) for 100 randomizations of sample order. Sampling intensity is the ratio of individuals to species. Inventory completeness is the percentage of species that are not singletons. Adjusted estimate range is the range of all but the Chao & Lee richness estimate values divided by the observed number of species. Grass bald Heath bald All samples Spring samples Fall samples All samples Spring samples Fall samples Richness estimates Chao & Lee 1 948.1 558.3 262.7 106.5 90.4 59.5 Chao & Lee 2 9147.8 4307.1 1260.1 150.9 144.1 73.0 ACE 120.9 93.6 59.4 75.6 57.5 53.6 ICE 125.3 99.8 65.0 76.8 57.9 52.5 Chao 1 159.6 ± 36.7 110.6 ± 25.5 62.0 ± 9.9 72.8 ± 8.4 54.4 ± 6.8 51.1 ± 8.8 Chao 2 139.2 ± 22.7 101.6 ± 18.2 71.1 ± 14.8 82.6 ± 13.9 61.6 ± 11.0 53.1 ± 9.6 first-order jackknife 124.6 ± 6.0 93.3 ± 5.9 65.6 ± 4.2 78.8 ± 4.5 59.6 ± 3.7 53.6 ± 3.9 second-order jack- knife 146.2 109.8 76.2 89.6 67.4 60.5 bootstrap 105.8 77.9 56.0 68.3 51.7 45.7 MM runs 112.4 85.6 72.9 70.7 59.5 59.5 MM mean 110.8 83.1 69.4 72.3 58.9 55.9 Observed richness 91 66 48 60 45 39 No. of samples 84 42 42 84 42 42 No. of adults 1853 1072 781 573 378 195 No. of adults/sample 22.1 25.5 18.6 6.8 9.0 4.6 No. of singletons 31 25 14 16 13 13 No. of doubletons 7 7 7 10 9 7 No. of uniques 34 28 18 19 15 15 No. of duplicates 12 11 7 8 7 8 Sampling intensity 20.4 16.2 16.3 9.6 8.4 5.0 Inventory complete- ness 66 62 71 73 71 67 Adjusted estimate range 0.59 0.50 0.42 0.37 0.33 0.38 TOTI ET AL.— SPIDER ASSEMBLAGES & SPECIES RICHNESS ESTIMATORS 337 the curves of the latter estimator are presented here. Chao 1, Chao 2, and ICE estimator curves were also especially erratic, even at high sample numbers. Because the two Chao and Lee estimators gave unrealistically high estimates (Table 2), their curves are not pre- sented in Figs. 1-6. Plots of singletons and uniques rise quickly, level off, and do not de- cline. There are always more uniques than sin- gletons. Plots of doubletons and duplicates rise more slowly, level off, and, in some data sets (grass total and grass fall), begin to fall. The richness estimates generated by the 1 1 estimators varied widely (Table 2). The two Chao & Lee estimates were always distinc- tively high (especially for the grass bald). The bootstrap estimates were consistently the low- est of the remaining nine estimators, and the second-order jackknife and, occasionally, Chao 1 produced the highest estimates. The estimates of the other six estimators tended to cluster more tightly and varied in rank de- pending on community and season. The rang- es spanned by these six estimates (ACE, ICE, Chao 2, first-order jackknife, and Michaelis- Menten runs and means) are smaller for the heath bald data sets than for the corresponding grass bald sets. These six richness estimates (106-160 for the grass bald and 68-90 for the heath) and the observed richness (9 1 and 60) indicate that more spider species live in the grass bald community than in the heath bald community. This conclusion is also supported by the ob- servation that the heath bald data set produced a smaller adjusted estimate range (the ratio of the range of all but the two Chao & Lee es- timators divided by the observed richness; Ta- ble 2), which suggests that the heath bald in- ventory is more nearly complete than is the grass bald inventory. The conclusion is further reinforced by the observation that the ob- served species accumulation curves for the grass bald data sets appear to be further from reaching an asymptote than are those for the corresponding heath bald data sets (Figs. 1- 6). We found this same pattern when we plot- ted species accumulation curves using number of specimens for the independent variable in- stead of number of samples; at an x-axis value of 573 specimens (the total number found in the heath bald sample set), the grass bald curve is steeper than the heath bald curve. This suggests that the observed difference in species richness between the two spider as- sembalges is not a result of sampling bias due to reduced sampling maneuverability in the heath bald. Community structure. — Values of the complementarity and similarity indices (Table 3) show that these two communities differ greatly in spider species composition; only 23 species were common to both communities. Even if the effect of “rare” species is reduced by deleting, before computing these indices, all singleton species that were found in only one community, the index values still indicate a large (although reduced) difference in spe- cies composition. In addition, there is a con- siderable, although smaller, difference be- tween spring and fall samples within each community in the species present as adults (Table 3). Both assemblages exhibit the commonly encountered skewed frequency distribution of few common species and many rare ones (Williams 1964) (Table 1). In the grass bald only 7 of the 91 observed species each com- prise 2% or more of the adults collected. Of these 7 “dominant” species, one (Ceraticelus alticeps, an erigonine linyphiid) is superabun- dant, comprising 54% of all adults collected at the site. In the heath bald 19 of the 60 ob- served species each comprise 2% or more of the adults collected. The two most abundant of these “dominants”, ColUnsia oxypaedero- tipus (an erigonine linyphiid) and Lepthy- phantes zebra (a linyphiine linyphiid), make up 15% and 14% respectively of all adults collected. In both communities, linyphiids were far more common than any other family in terms of numbers of species and adults (Table 4). The next three most species-rich families in the grass bald (Salticidae, Lycosidae, and Ar- aneidae) were much less well represented in the heath bald; the absence of lycosids and the presence of only one araneid species in the heath samples are particularly noteworthy. Very small juveniles of Araneus orbweavers (probably A. nordmanni) were common in the heath; we saw only two or three large orb webs, but did not find their owners. Two fam- ilies were notably more species-rich in the heath samples than in the grass bald: Dictyn- idae and Leptonetidae. In the grass bald, the percentages of aerial (47) and ground-dwelling (53) species are Cumulative Number of Species Cumulative Number of Species 338 THE JOURNAL OF ARACHNOLOGY 170 160 150 140 130 120 no 100 90 80 70 60 50 40 30 20 10 0 100 90 80 I 70 60 50 40 30 20 10 0 TOTI ET AL.— SPIDER ASSEMBLAGES & SPECIES RICHNESS ESTIMATORS 339 Grass Bald (spring) ,chao i "^^jackknife 2 T 10 20 30 40 50 Cumulative Number of Samples 120- 100- 90- a m 0 70- 1 g 60- u 50- 3 40- 6 U 30. 20" 10 0 4- Grass Bald (fall) uniques singletons duplicates doubletons 120 110 100 90 80 70 60 50 40 30 20 10 0 0 10 20 30 40 50 Cumulative Number of Samples 10 20 30 40 50 Cumulative Number of Samples Cumulative Number of Samples Figures 3-6. — Plots comparing the performance of eight estimators of species richness with the ob- served species accumulation curve, using data from the four sets of spider samples. 3. Spring samples from the grass bald; 4. Fall samples from the grass bald; 5. Spring samples from the heath bald; 6. Fall samples from the heath bald. Scales, line symbols, variables, and computation protocols are the same as for Figure 1. Figures 1, 2. — Plots comparing the performance of eight estimators of species richness with the observed species accumulation curve, using data from all 84 samples (spring and fall) of spiders from the grass bald (Figure 1) and the heath bald (Figure 2). The species accumulation curve (S observed) plots the observed number of species as a function of the number of pooled samples. The eight curves above the species accumulation curve show the estimated species richness based on successively larger numbers of samples. The estimators used are ACE, ICE, Chao 1, Chao 2, first-order jackknife (jackknife 1), second- order jackknife (jackknife 2), bootstrap, and Michaelis-Menten means (MMMeans). All values were gen- erated by Estimates, version 5.0.1 (Colwell 1997). The four curves at the bottom of the graph plot mean numbers of singletons, doubletons, uniques, and duplicates as a function of cumulative number of samples. For all 13 curves, each point is the mean of 100 values based on 100 randomizations of sample accu- mulation order. 340 THE JOURNAL OF ARACHNOLOGY Table 3. — Values of complementarity and similarity indices for the two communities and for the spring and fall data sets of each community. See text for definitions of the indices. Index values in parentheses were generated after deleting all singletons found in only one community. Grass bald to heath bald Grass bald spring to fall Heath bald spring to fall % Complementarity 82 (73) 75 60 Kulczynski’s index of similarity 0.196 (0.229) 0.584 0.363 Sorensen’s index of similarity 0.305 (0.422) 0.397 0.571 about equal, and the ground web-building guild is the most species-rich (35% of all spe- cies present) (Table 5). In the heath bald, 67% of the species are ground-dwellers, with the great majority of these (58.3% of all species present) probably being ground web-builders. Pronounced seasonal changes occurred in the (adult) guild composition of the grass bald community but not in the heath community (Table 5). The number of aerial web-builder species present as adults increased and the number of aerial and ground hunter species present as adults decreased between spring and fall in the grass bald. Collecting methods: taxonomic charac- terization of yields. — The complementarity matrices reveal — for both communities — gen- erally large differences among the methods in the taxonomic composition of the spider sam- ples these methods yield (Table 6). The small- est differences are those between ground and litter samples, beating and sweeping samples, and beating and aerial samples. In terms of number of species per sample, it appears that aerial and beating methods are the least pro- ductive (Table 7). Each method yielded some unique species not collected by any other method; aerial hand collecting in the heath yielded the smallest number of unique species (Table 7). From 33-60% of these unique-to- method species were singletons. DISCUSSION Species richness estimates. — The best way to test the performance of species richness es- timators is to use data sets from a site where the actual species richness is known; the ger- minating seed bank data set used by Colwell & Coddington (1994) and Butler & Chazdon (1998) essentially meets this requirement. Un- fortunately, we cannot use this direct approach to evaluate estimator performance because none of our observed species accumulation curves reached an asymptote; evidently we have not collected all the species present as adults at either site during the seasons when we sampled. However, we can employ other less rigorous (indirect) ways to assess esti- mator usefulness — observing how rapidly es- timation curves approach an asymptote as sample number increases (Colwell & Cod- dington 1994; Coddington et al. 1996; Chaz- don et al. 1998), looking for a consensus among a majority of estimators (Coddington et al. 1996), and comparing the estimator curves to subjective visual extrapolations of the possible asymptotes of an observed spe- cies accumulation curve. A good estimator 1) should reach (or at least closely approach) a stable asymptote with fewer samples than are required for the observed species accumula- tion curve to reach an asymptote, 2) is un- likely to yield estimates that differ widely from those of all other estimators, and 3) should give estimates that are close to reason- able visual extrapolations of the asymptote of the observed species accumulation curve. The two Chao & Lee estimators generated unrealistically large estimates, especially so with the grass bald data sets. Colwell & Cod- dington (1994) observed the same tendency of these two estimators to overestimate species richness with a seed bank data set. The newer, modified coverage-based estimators, ACE and ICE, the latter of which performed especially well in a recent study by Chazdon et al. (1998), generate much more realistic richness estimates for our sample sets than do the Chao & Lee estimators. Although the rankings of richness values generated by all 1 1 estimators vary somewhat among our data sets and from study to study (Coddington et al 1996; Silva & Coddington 1996; Dobyns 1997; Chazdon et al. 1998), the relatively tight clustering of the ACE, ICE, Chao 2, first-order jackknife, TOTI ET AL.— SPIDER ASSEMBLAGES & SPECIES RICHNESS ESTIMATORS 341 Table 4. — Percent (of community total) of species and adults collected in each family. Number of species is in parentheses. Grass bald Heath bald % of species % of adults % of species % of adults Agelenidae 14 (1) 0.2 1.7 (1) 0.7 Amaurobiidae 3.3 (3) 0.4 6.7 (4) 4.7 Antrodiaetidae 1.1 (1) 0.4 1.7 (1) 3.7 Anyphaenidae 1.1 (1) 0.1 Araneidae 8.8 (8) 2.1 1.7 (1) 0.2 Clubionidae 4.4 (4) 5.7 5.0 (3) 4.4 Cybaeidae 1.7 (1) 0.7 Dictynidae 2.2 (2) 0.2 8.3 (5) 5.9 Gnaphosidae IT (1) 0.2 Hahnidae 2.2 (2) 1.1 5.0 (3) 6.1 Leptonetidae 3.3 (2) 0.5 Linyphiidae 30.8 (28) 80.6 36.7 (22) 51.7 Liocranidae 1.1 (1) 0.2 3.3 (2) 3.5 Lycosidae Nesticidae 9.9 (9) 1.0 1.7 (1) 0.7 Oxyopidae 1.1 (1) 0.3 Philodromidae 1.1 (1) 0.1 Salticidae 18.7 (17) 5.0 8.3 (5) 7.3 Tetragnathidae 1.1 (1) 0.3 1.7 (1) 1.2 Theridiidae 6.6 (6) 1.6 8.3 (5) 7.3 Theridiosomatidae 1.7 (1) 1.2 Thomisidae 4.4 (4) 0.5 3.3 (2) 0.7 and MichaeliS“Menten estimator values sug- gests that they are either estimating the same real value or are being biased in the same manner. When we apply the above-mentioned three criteria of a potentially good estimator to the performance of the estimators with all six of our data sets, the Michaelis-Menten es- timator appears to perform best. ICE and ACE also perform rather well but do not approach an asymptote as quickly as the Michaelis- Menten estimator. Chao 2 and the first-order jackknife show some promise, but the former is sometimes quite unstable and neither close- ly approach an asymptote. The poor perfor- mance of the bootstrap estimator on our data sets echos the findings of others (Colwell & Coddington 1994; Chazdon et al. 1998). Al- though they used tropical forest seed, seed- ling, and sapling data sets that differed greatly from ours, Colwell & Coddington (1994), Butler & Chazdon (1998), and Chazdon et al. (1998) did not come to conclusions that were radically different from ours about the perfor- mance of these richness estimators. However, they did give ICE (Chazdon et al. 1998) and Chao 2 (Colwell & Coddington 1994; Chaz- don et al. 1998) the highest overall ratings. The failure of the observed species accu- mulation curve and most of the estimator curves to reach an asymptote with our data Table 5. — Percentage of species in each guild for each community and for spring and fall samples from each community. Any species in two guilds (see Table 1) was assigned to its primary guild. Grass Heath Grass Heath Spring Fall Spring Fall Aerial web builders 22.0 21.7 12.1 31.3 17.8 15.4 Ground web builders 35.2 58.3 37.9 41.7 62.2 69.2 Aerial hunters 25.3 11.7 28.8 16.7 11.1 10.3 Ground hunters 17.6 8.3 21.2 10.4 8.9 5.1 342 THE JOURNAL OF ARACHNOLOGY Table 6. — Percent complementarity of the samples collected by different methods. See text for definition of percent complementarity. Number of sample units in parentheses. Grass bald Beating (24) Sweeping (24) Ground (24) Litter (12) Beating — 70 80 87 Sweeping — — 80 87 Ground — — — 67 Heath bald Beating (24) Aerial (24) Ground (24) Litter (12) Beating — 72 78 87 Aerial — — 84 90 Ground — — — 57 sets is directly related to the fact that the num- bers of singleton and unique species failed to decline as sample size increased (Figs. 1-6). Indeed, the relatively steep “final” slopes of the Chao 1, Chao 2, and first- and second- order jackknife curves for the grass bald data sets (Figs. 1, 3, 4) were caused by this failure of singletons and uniques to decrease with in- creased collecting effort while doubletons and, to a lesser degree, duplicates decreased. Since rare species (especially singletons and uniques) play such an important role in gen- erating most of these estimates, it may be in- structive to examine the ecological and taxo- nomic status of singletons in our data sets. Most of the singletons in each community (55% of bald and 63% of heath singletons) are common in other habitats in the GSMNP, and most of the rest appear to be common only in regions beyond the GSMNP boundary (Table 1). Consequently, many of these species may not be permanent (breeding year after year) members of these bald communities. However, it is also possible that some or many of these singletons may be temporal singletons, artifacts of temporally patchy sampling; we will not know without collecting early spring, summer, and late fall samples of adults from these balds. For those famihes represented by five or more species in either or both bald communities, the percentage of that family’s species that are sin- gletons ranges from 22-67%: Araneidae (22%), Linyphiidae (28%), Amaurobiidae (29%), Clubionidae (40%), Salticidae (50%), Thomisidae (50%), Theridiidae (55%), Dictyn- idae (60%), Lycosidae (67%). Why lycosid species should more often be rarer than araneid or linyphiid species in these habitats, especially in the meadow-like grass bald, is not obvious. Sampling bias is not a likely explanation, be- cause our ground and litter sampling methods collect large numbers of lycosid individuals and/or species in some other non-forest and forest communities in the GSMNP. Table 7. — Comparison of total number of species and number of unique species (here defined as species collected by only one method) sampled by each method in each community. Number of samples in parentheses after each method heading. The number of unique species which are singletons is given in parentheses after the number of unique species. Grass bald No. of species No. of species per sample No. of unique species Heath bald No. of species No. of species per sample No. of unique species Beating (24) 27 1.1 7 (4) Beating (24) 23 1.0 9 (4) Sweeping (24) 44 1.8 20 (11) Aerial (24) 14 0.6 4 (2) Ground (24) 45 1.9 15 (9) Ground (24) 38 1.6 18 (6) Litter (12) 32 2.7 11 (5) Litter (12) 27 2.3 7 (3) TOTI ET AL.— SPIDER ASSEMBLAGES & SPECIES RICHNESS ESTIMATORS 343 As in similar inventories of spider assem- blages (Coddington et aL 1996; Dobyns 1997), it is difficult to judge from these data sets and estimates the true species richness at either study site, primarily because we iden- tified and counted only adults and therefore do not know how many resident species pop- ulations consisted only of juveniles during the two brief sampling periods at each site. The diversity of phenologies in a typical spider community is so great (Toft 1976) that it may be difficult or impossible to estimate true rich- ness until sampling bouts for adults are dis- tributed more evenly throughout the annual cycle of seasons or juveniles are identified to species. This is demonstrated by the differ- ences in species richness (observed and esti- mated) and species composition between spring and fall samples at each of our bald sites. Accurate estimation of true richness from snapshot sampling may only be feasible where spider faunas are extremely well known, because it will require either accurate sampling and identification of all age classes or the use of estimator formulas derived in part from extensive knowledge of the life cy- cle patterns of relevant spider assemblages. It is possible that late spring and early fall are the two best times to inventory spiders in tem- perate communities, and that two such sam- ples will prove adequate for estimating spe- cies richness, but these possibilities need to be tested by year-round sampling. One goal of species richness inventories should be to help predict how many samples are required for a complete (observed species accumulation curve reaches asymptote) or ad- equate (accurate estimate of true richness) sur- vey. Indices like sampling intensity or inven- tory completeness may be useful. Our results indicate that one drawback of the sampling intensity index is that a superabundant species (like Ceraticelus alticeps in the grass bald) can inflate the index; even though our sam- pling intensity at the grass bald was 20.4, the species accumulation curve generated for that site was not as close to its asymptote as was the curve for the heath bald, which had a much lower sampling intensity (9.6). Exclud- ing species with abundances of more than 100 or 200 would be a way to avoid such inflated sampling intensity values. As noted by Cod- dington et al. (1996), the rough estimate by Coddington et al. (1991) that a sampling in- tensity of 10 should be adequate for an ac- curate survey is low, at least for some spider assemblages. The inventory completeness in- dices for our two fall data sets (71, 67) are comparable to that for the Ellicott Rock forest fall data set (71) (Coddington et al. 1996). While the latter data set consisted of three times as many samples as either of our fall data sets, it also contained over twice as many species. The similar slopes of the observed species accumulation curves for the forest and both fall bald data sets suggest that all three inventories may have reached roughly the same degree of completeness. It is clear that more sampling is needed at both bald sites to determine whether any of these estimators can provide meaningful esti- mates of the species richness of these spider assemblages. However, a few results suggest that our heath bald inventory is more nearly complete than the grass bald inventory, in spite of roughly equal sampling effort; the heath data set yields 1) observed and estimated richness curves that are more closely approaching an asymptote, 2) smaller gaps between observed and estimated richness curves, and 3) a smaller interval between the lowest and highest rich- ness estimates. This last result is expressed by the adjusted estimate range (Table 2), which may be a useful indicator of inventory com- pleteness. The intensity and seasonal frequency of sampling needed to generate samples of adult spiders that may yield useful estimates of species richness will only be deterimined by analysis of data from concerted year-round sampling effort at a particular site and will cer- tainly differ from region to region and habitat to habitat. Community structure. — The large differ- ences between these two bald spider assem- blages in both taxon and guild composition are not surprising considering the big differ- ences in community physiognomy and plant species composition. The grass bald contains large patches of low grass and herb-dominated meadow habitat not found in the heath bald; the impact of this difference in the plant com- ponent of the communities on the spider com- ponent is demonstrated by the observation that 19 (26%) of the 76 grass bald spider spe- cies found elsewhere in the GSMNP are found only in non-forested meadow habitats whereas in the heath bald this is true for only one (2%) of 54 species. The fact that many of these 344 THE JOURNAL OF ARACHNOLOGY meadow species are salticids, lycosids, and ar- aneids helps explain the much better represen- tation of these three families in the grass bald. We suggest that the very low richness and abundance of adult orb weavers in the heath bald, in spite of moderate numbers of very young juveniles, may be due to a paucity of flying insects, which we noticed while sam- pling in the heath. Perhaps juvenile orb weav- ers colonize the heath from adjacent habitats but find survival difficult. The striking domi- nance of ground-dwelling guilds, and espe- cially the ground web-builders, in the heath bald may be due in part to a litter and humus layer that is thick and well-shaded (and there- fore probably relatively stable microclimati- cally) and to the low diversity and abundance of herbivorous insects supported by the rela- tively unpalatable ericaceous foliage. Other studies demonstrate the positive correlation between litter depth, litter nficroclimate sta- bility, and ground spider species richness (Uetz 1979; Coyle 1981). Furthermore, while sampling spiders, we observed a high density of detritivore arthropods in this heath litter. Perhaps the diterpene antifeedents and insec- ticides that make ericaceous leaves unpalat- able to many herbivores (Rosenthal & Janzen 1979; Klocke et al. 1991; Harbome 1993) are leached out of the litter or reabsorbed by the plant before leaf abscission. For the reasons given earlier in the Results section, we feel confident that the differences between the two sample sets in observed (91 v.y. 60) and estimated (106-160 vs. 68-90) species richness mean that the grass bald spi- der assemblage is significantly richer than that of the heath bald. The much higher plant spe- cies richness and much more varied physiog- nomy (patches of meadow and shrubs, and scattered trees) of the grass bald, the domi- nance of relatively unpalatable foliage in the heath bald, and the greater diversity and abun- dance of herbivorous insects in the grass bald, are likely to be important (and interrelated) causes of this marked difference in spider spe- cies richness. The apparent temporal shift in taxononfic and guild structure from spring to fall within each of these two spider assemblages is, of course, an artifact of our ignorance of juvenile spiders. The distinct differences in the life cy- cles and adult phenologies within an assem- blage of spider species (Toft 1976) guarantees that adult-only samples taken in one season will be different taxonomically from those taken from the same site at another season. The increase in aerial web-builders and the equally marked decrease in hunting guilds be- tween spring and fall (adult) samples at the grass bald are consistent with the tendency of most north temperate araneids to be late sum- mer and fall breeders and most north temper- ate hunting guild taxa to breed in the spring and early summer (Toft 1976; Gertsch 1979). Collecting methods. — Longino & Colwell (1997) stressed the importance of using sam- pling methods that collect complementary sets of species. The large differences among our five collecting methods in the taxonomic com- position of the samples these methods yielded, as well as the fact that even the least produc- tive method (aerial hand sampling) collected four species not collected by any other method in the heath bald, justify their continued use in future sampling in these habitats. The very high productivity and distinctiveness of the sweep samples suggests that we were justified in substituting this method for aerial hand sampling in the grass bald. Such a substitution in the heath bald would not be appropriate; the physiognomy of the heath bald makes sweeping very difficult and is such that sweeping would probably sample the same taxa that beating does. ACKNOWLEDGMENTS Keith Langdon of the GSMNP provided lo- gistic support. Jonathan Coddington, Matt Greenstone, and two anonymous reviewers provided helpful comments on a draft of this paper. This research was supported by a Na- tional Park Service Challenge Cost-Share Grant and a National Science Foundation Grant (DEB-9626734) to FAC. LITERATURE CITED Billings, W.D. & A.E Mark. 1957. Factors in- volved in the persistence of montane treeless balds. Ecology, 38:140-142. Bray, J. & J. Curtis. 1957. An ordination of the upland forest communities of southern Wiscon- sin. Ecol. Monog., 27(4):325-349. TOTI ET AL.— SPIDER ASSEMBLAGES & SPECIES RICHNESS ESTIMATORS 345 Butler, BJ, & R.L. Chazdon, 1998. Species rich- ness, spatial variation, and abundance of the soil seed bank of a secondary tropical rain forest. Biotropica, 30(2):2 14-222. Cain, S.A. 1930. An ecological study of the heath balds of the Great Smoky Mountains. Butler Univ. Botanical Stud., 1:177-208. Chazdon, R.L., R.K. Colwell, J.S. Denslow & M.R. Guariguata. 1998. Statistical methods for esti- mating species richness of woody regeneration in primary and secondary tropical forests of NE Costa Rica. Pp, 285-309, In Forest Biodiversity Research, Monitoring and Modelling: Concep- tual Background and Old World Case Studies. (F. Dallmeier & J.A. Comiskey, eds.). Pantheon Press, Paris. Coddington, J.A., C. Griswold. D. Davila, E, Pen- aranda & S. Larcher. 1991. Designing and test- ing protocols to estimate biodiversity in tropical systems. Pp. 44-60, In The Unity of Evolution- ary Biology. Proc. IV Intern. Congr. Syst. EvoL Biol. (E. Dudley, ed.). Discorides Press, Port- land, Oregon. Coddington, J.A. & H.W. Levi. 1991. Systematics and evolution of spiders. Annu. Rev. EcoL Syst., 22:565-592. Coddington, J.A., L.H. Young & F.A. Coyle. 1996. Estimating spider species richness in a southern Appalachian cove hardwood forest. J. ArachnoL, 24:111-128. Colwell, R.K. 1997. Estimates: Statistical Esti- mation of Species Richness and Shared Species from Samples. Version 5. User’s Guide and ap- plication published at: http://viceroy.eeb. ucoen. edu/estimates. Colwell, R.K. & J.A. Coddington. 1994. Estimat- ing terrestrial biodiversity through extrapolation. Phil. Trans. Royal Soc. (Ser. B), 345:101-118. Coyle, EA. 1981. Effects of clearcutting on the spider community of a southern Appalachian for- est. J. ArachnoL, 9:285-298. Dobyns, J.R. 1997. Effects of sampling intensity on the collection of spider species and the esti- mation of species richness. Environ. EntomoL, 26(2): 150-162. Gertsch, W.J. 1979. American Spiders. Van Nos- trand Reinhold, New York. Harbome, J.B. 1993. Introduction to Ecological Chemistry, Academic Press, New York. Kent, M. & P. Coker. 1992. Vegetation Description and Analysis: A Practical Approach. CRC Press, Boca Raton, Florida. Klocke, J.A., Mei-Ying Hu, Shin-Foon Cuiu, & L Kubo. 1991. Grayanoid diterpene insect antifee- daiits and insecticides from Rhododendron molle. Phytochemistry, 30(6): 1797-1800. Kremen, C., R.K.Colwell, T.L. Erwin, D.D. Mur- phy, R.F. Noss & M.A. Sanjayan. 1993. Terres- trial arthropod assemblages: Their use in conser- vation planning. Conserv. Biol., 7(4):796-808. Longino, J.T 1994. How to measure arthropod di- versity in a tropical rainforest. Biol. Intemat., 28: 3-13. Longino, J.T. & R.K. Colwell. 1997. Biodiversity assessment using structured inventory: Capturing the ant fauna of a tropical rain forest. EcoL Appl- ic., 7(4): 1263-1277. Mark, A.F. 1958. The ecology of the southern Ap- palachian grass balds. EcoL Monogr., 28:293-336. Norris, K.C. 1999. Quantifying change through time in spider assemblages: Sampling methods, indices, and sources of error. J. Insect Conserv., 3:1-17. Platnick, N.I. 1997. Advances in Spider Taxonomy 1992-1995. New York EntomoL Soc., New York. Platnick, N.I. 1999. Dimensions of biodiversity: Targeting megadiverse groups. Pp. 33-52, In The Living Planet in Crisis: Biodiversity Science and Policy. (J. Cracraft & FT. Grifo, eds.). Columbia Univ. Press, New York. Riechert, S.E. & L. Bishop. 1990. Prey control by an assemblage of generalist predators: Spiders in garden test systems. Ecology, 71(4): 1441-1450. Riechert, S.E. & TC. Lockley. 1984. Spiders as biological control agents. Annu, Rev, EntomoL, 29:299-320. Rosenthal, G. & D.H. Janzen. 1979. Herbivores: Their Interaction with Secondary Plant Metabo- lites. Academic Press, New York. Silva, D. & J.A. Coddington. 1996. Spiders of Pak- itza (Madre de Dios, Peru): Species richness and notes on community structure. Pp. 253-311, In The Biodiversity of Southeastern Peru. (D.E. Wilson & A. Sandoval, eds.). Smithsonian Instit., Washington D.C. Stratton, D.A. & P.S. White. 1982. Grassy balds of Great Smoky Mountains National Park: Vascular plant floristics, rare plant distributions, and an assessment of the floristic data base. Res./Resour. Manag. Rep. SER-58, National Park Service, Nat. Sci. Res. Div,, Atlanta, Georgia. Toft, S. 1976. Life histories of spiders in a Danish beech wood. Nat. Jutlandica, 19:6-40. Uetz, G.W. 1979. The influence of variation in lit- ter habitats on spider communities. Oecologia, 40:29-42. Whittaker, R.H. 1956. Vegetation of the Great Smoky Mountains. EcoL Monogr., 26:1-80. Williams, C.B. 1964. Patterns in the balance of na- ture. Academic Press, New York. Wilson, E.O. 1988. The current state of biological diversity. Pp. 3-17, In Biodiversity. (E.O. Wil- son, ed.). Nat. Acad. Press, Washington, D.C. Wilson, E.O. 1992. The Diversity of Life, Harvard Univ. Press, Cambridge, Massachusetts. Wise, D.H. 1993. Spiders in Ecological Webs. Cambridge Univ. Press, Cambridge, U.K. Manuscript received 2 September 1999, revised 21 January 2000. 2000. The Journal of Arachnology 28:346-350 DOES THE PRESENCE OF POTENTIAL PREY AFFECT WEB DESIGN IN ARGIOPE KEYSERLINGI (ARANEAE, ARANEIDAE)? Marie E. Herberstein, Anne C. Gaskett, Deborah Glencross, Simon Hart, Sue Jaensch and Mark A. Elgar: Department of Zoology, University of Melbourne, 3010, Victoria, Australia ABSTRACT. Orb-web spiders may anticipate their future prey environment by detecting the presence of prey and adjusting their web building behavior accordingly. Here we investigate the effect of different prey sizes and density on the web size and mesh height of the orb webs constructed by Argiope keyserlingi. The experimental design allowed the transmission of prey vibrations but prevented any capture. We found that A. keyserlingi constructed webs more frequently in the presence of prey, but did not alter the web size or mesh height of their webs. Keywords: Orb web, mesh height, foraging, behavior Orb-web spiders (Araneae, Araneidae) em- ploy remarkable flexibility in their foraging behavior. For example, following periods of starvation, orb-web spiders increase the size of their webs and attack prey less selectively while sated spiders reduce web size and reject less profitable prey (e.g., Sherman 1994; Her- berstein et al. 1998; Herberstein et al. 2000). Web construction is energetically the most ex- pensive component of a spider’s foraging ef- fort (Opell 1998), and webs cannot be modi- fied following completion. Decisions made during web construction influence subsequent capture success until a new web is built. Thus, it may be advantageous to design a web in anticipation of the future prey environment, rather than simply relying on past events. Web-building spiders may make some pre- emptive foraging decisions in response to the density or size of potential prey. Sandoval (1994) concluded that the orb- web spider, Parawixia bistriata is able to exploit swarms of unusually large termite prey. Parawixia bistriata typically constructed small, finely meshed webs at night that trapped tiny dip- teran prey (Sandoval 1994). At the onset of the rainy season, the spiders dramatically changed their activity patterns and web de- sign. At this time, they built additional webs during the day with increased web area and mesh height (the average distance between capture spirals). Interestingly, the spiders seemed to anticipate the timing of the swarms: they changed their web design before the ter- mites emerged, potentially using rainfall and humidity as cues (Sandoval 1994). Experi- mental evidence also suggests that spiders vary mesh height due to the presence of dif- ferently sized prey (Schneider & Vollrath 1998). In a similar case, Zygiella x-notata (Pasquet et al. 1994) anticipated prey density before web construction. More abundant po- tential prey induced the construction of small- er webs earlier in the evening. Presumably, smaller webs were finished more quickly, al- lowing prey capture to commence earlier. Here, we examine the effect of the size and number of potential prey on the web building behavior of Argiope keyserlingi Karsch 1878. We predict that larger potential prey will in- duce increased mesh height, and that higher prey density will decrease web area. Experiments were conducted in March and April 1998 and January 1999, using adult Ar- giope keyserlingi collected in Sydney and Brisbane, Australia. In the laboratory, spiders were housed in upturned plastic cups (13.5 X 9X9 cm) where they were watered and fed blow flies (Lucilia cuprina, Diptera) ad libi- tum. The spiders were unable to construct a functional web in the upturned cups apart from a few supporting threads. Thus, prey capture did not involve a web. Instead, the spiders generally grasped the flies buzzing around in the cup. The spiders were starved for four days prior to experimentation. This ensured that the spi- ders’ energetic status was uniform. Addition- 346 HERBERSTEIN ET AL.— PREY AND WEB DESIGN 347 ally, by depriving spiders of web=building space we minimized the influence of previ- ously built webs on the foraging decisions made during experimentation (see Herberstein et al. 2000). The spiders were weighed and transferred to three-dimensional frames (40 X 50 X 8.5 cm) and allowed to construct com- plete webs in the presence of different sizes and densities of potential prey. Frames either contained 30 Drosophila (Diptera), one blow fly, or 30 blow flies. Prey were held in iden- tical plastic jars (diameter: 4.7 cm, height: 6.8 cm), covered by fine mesh. This setup allowed prey movement and the transmission of air- borne vibrations created by the buzzing of the flies, but prevented capture. We selected the two prey types because they differ in body length (blow flies: 7.8 ± 0.12 mm, n = 20; Drosophila: 2.5 ± 0.06 mm, n = 20). To control for differences in weight and therefore energy return, treatment one consisted of 30 Drosophila per jar. This approximated the weight of one blow fly per jar as used in treatment two (one blow fly: 0.022 ± 0.0006 g, n = 39; 30 Drosophila: 0.021 ± 0.0004 g, n ^ 21). The third treat- ment, 30 blow flies, allowed comparison of the webs built for different prey densities, and for different prey types. Only the first web spun by each individual was measured and used to evaluate the effects of the prey treat- ments. This minimized the influence of pre- vious foraging history on web design. We es- timated the web area and the mesh height using various formulae that only require a few measurements (Herberstein & Tso 2000). Statistical analyses were conducted using Systat 5.2 (Wilkinson 1992) and G^Power (Buchner et al. 1997). Data were log trans- formed if they were not normally distributed (Kolmogorov-Smimov). Web area, mesh and spider weight were compared using ANOVA with treatment and year as factors. All values are mean ± SE unless stated otherwise. Data from 49 spiders were included in the analyses. There was no significant difference in body weight between the spiders used in 1998 and 1999 (for 1998/1999: 30 blow flies 0.255 ± 0.028 g / 0.266 ± 0.023 g, one blow fly 0.269 ± 0.020 g / 0.293 ± 0.029 g, 30 Drosophila 0.253 ± 0.019 g / 0.219 ± 0.038 g; Fi 43 == 0.00001, P > 0.05). The weight of spiders allocated to the three treatments was similar (F2, 43 = 1.37, P > 0.05), and there 100 - 1 large prey 30 small prey 30 large prey Figure 1. — The mean (± SE) area of webs con- structed in the presence of one large prey, 30 small prey and 30 large prey in 1998 (•) and 1999 (o). was no interaction effect of year and treat- ment (F2,43 0.61, P > 0.05). Web area (Fig. 1) did differ between the two years (Fj 43 = 30.79, P < 0.01): in 1999 spiders constructed larger webs compared to the previous year. This is probably because spiders were main- tained in the laboratory for approximately two months before use in 1998, whereas the ex- periment was commenced within two weeks of collection in 1999. Varying the size and density of potential prey did not affect web size (F2 43 = 0.007, P > 0.05), nor was there an interaction effect between year and treat- ment (F2 43 = 0.79, P > 0.05). The size of the frame, and thus the available web building space may have limited the foraging decision of the spiders. However, the maximum web size observed (approx. 850 cm^) was less than half of that available (2000 cm^). Mesh height (Fig, 2) was similar in both years (Fj 43 ^ 1.40, P > 0.05) and was un- affected by prey treatment (F2 43 = 0.34, P > 0.05). Contrary to prediction, the presence of large prey did not result in larger mesh height compared to small prey. Power analysis re- vealed that our sample size was sufficient to detect a treatment effect (1 — (3 == 0.68). Again, the interaction between year and treat- ment was not significant (F2 43 == 1.63, P > 0.05). These results are contrary to both of our predictions, and the results of previous studies (Schneider & Vollrath 1998; Pasquet et al. 1994) that found a relationship between the size and density of prey and web design. However, these previous experiments released prey into the web-building frames with the spiders. In the laboratory, we frequently ob- 348 THE JOURNAL OF ARACHNOLOGY 0.8 -1 0.3 - 0.2 -\ 1 i 1 1 large prey 30 small prey 30 large prey Figure 2. — The mean (± SE) mesh height for webs constructed in the presence of one large prey, 30 small prey and 30 large prey in 1998 (•) and 1999 (o). serve orb-web spiders housed in both frames and cups grasping and consuming prey with- out webs. As the spiders in these previous ex- periments (e.g., Schneider & Vollrath 1998) had the opportunity to capture prey during web building, it is unclear whether their webs represented an anticipatory prey assessment, or past experience. In the present study, en- closing the prey in mesh-covered jars pre- vented such confounding effects. However, the absence of any significant difference in web design between the prey treatments sug- gests two explanations; either our experimen- tal design did not allow the spiders to detect the prey, or A. keyserlingi does not make pre- emptive adjustments to web mesh size and area to suit varying sizes and numbers of po- tential prey. To distinguish between these two explana- tions, we repeated the experiments in January and May 2000 using identical methods but in- cluding a control treatment (no flies), where we measured web area and mesh height in a sub- sample and the frequency of web con- struction in a larger sample of individuals. We predicted that, if these spiders can detect air- borne vibrations created by the enclosed flies, we should find differences between treatments that included no blow flies (empty container), one blow fly and 30 blow flies. Any difference in the web-building behavior between the no- fly treatment versus the fly treatments would indicate that the spiders were able to detect the presence or absence of prey in the con- tainers. We found no significant differences in mesh size (F2 34 == 1.28, P > 0.05) or web Table 1. — The mean (±SE) for the web area and mesh height of spiders constructing webs when there are no flies, one fly or 30 flies enclosed with the spider. Treat- ment Sam- ple size Web area (cm^) Mesh height (cm) No flies 11 1053.0 ± 112.6 0.517 ± 0.02 1 fly 12 1015.3 ± 108.4 0.503 ± 0.04 30 flies 14 1086.8 ± 100.3 0.517 ± 0.02 area (F2, 34 = 0.29, P > 0.05; Table 1) be- tween the treatments. However, fewer spiders constructed a web when no flies were present (16 out of 24 spiders). In contrast, almost all spiders (21 out of 22) presented with a jar of 30 blow flies, and 19 of 26 spiders presented with only one blow fly, built a web. We com- pared these frequencies using a contingency table, which revealed that the likelihood to build a web was significantly different be- tween the three treatments (x^ = 6.3, P = 0.044). These results indicate that our exper- imental design allowed the spiders to detect the presence of potential prey, and they ad- justed the frequency of web construction ac- cordingly (see also Pasquet et al. 1994), but not web size or design. It may be that spiders are unable to detect differences between the airborne vibrations created by different sizes and densities of prey. Alternatively, spiders may be able distinguish between prey densi- ties and sizes, but do not alter the web design in response. Behavioral tests, such as those presented here, cannot distinguish between these two alternatives. Adjusting web building frequency in re- sponse to the presence of prey may reflect risk sensitivity, where foragers react to variation in prey encounter rates by changing web sites or web size (e.g., Herberstein et al. 2000; Gilles- pie & Caraco 1987). Web building spiders in- vest a substantial amount of energy into silk production and web construction (e.g., Peakal & Witt 1976; Higgins & Buskirk 1992), and rely on prey coming into contact with the web. As such, prey encounter can be highly unpre- dictable and spiders may conserve energy by not building a web when there is little indi- cation of abundant prey. In contrast, when prey is in close proximity and in high density. HERBERSTEIN ET AL.— PREY AND WEB DESIGN 349 increased web building activity may allow these spiders to exploit abundant prey. Numerous field studies have also failed to find a consistent relationship between mesh height and prey size (McReynolds & Polis 1987; Herberstein & Elgar 1994; Herberstein & Heiling 1998). Simulations (Eberhard 1986) and laboratory manipulations (Nentwig 1983) further confirm that orb- webs do not function as “sieves.” Mesh height may fulfill alternative functions. A narrow mesh may fa- cilitate the retention of larger prey, as more threads are in contact with the item (Eberhard 1990) . However, more spiral turns also reflect more light thus increasing the visibility of the web to prey (Craig 1986; Craig & Freeman 1991) . Mesh height may therefore indicate a compromise between prey retention and web visibility. A larger capture area results in a higher prey interception rate (Chacon & Eber- hard 1980) and by increasing the distance be- tween sticky spirals spiders may enlarge over- all capture area without greater energy expenditure. Accordingly, food deprived spi- ders commonly increase web area to enhance prey encounter (Sherman 1994; Herberstein et al. 2000). Finally, it seems unlikely that spi- ders would tailor their webs for small and pos- sibly unprofitable prey. Spiders often ignore small prey entangled in the web (Uetz & Hart- sock 1986; Herberstein et al. 1998) which may subsequently escape. Logically, any web should target profitable prey items worthy of attack and more permanent retention through silk wrapping. Web design reflects several trade-offs be- tween the different functions of various web elements and is influenced by internal physi- ological states and previous experience. Inter- preting orb-webs as size filters is likely to oversimplify this complex foraging invest- ment. We thank Fleur de Crespigny and Sharada Ramamurthy for their technical help and sup- port; Simon Blomberg, Diana Fisher and Mat- thias Herberstein for logistic support; John Mackenzie and Janet Yen for providing the flies, the Austrian Science Foundation for fi- nancial support to MEH (J1318~BIO and J1500-BIO) and the Australian Research Council to MAE (ARC 19930103). LITERATURE CITED Buchner, A., F. Paul & E. Erdfelder. 1997. G«Power: a priori, post-hoc, aedcompromise power analyses for the Macintosh (Version 2.1.2; Computer program). University of Trier: Ger- many. Chacon, R & W.G Eberhard. 1980. Factors affect- ing numbers and kinds of prey caught in artificial spider webs with considerations of how orb webs trap prey. Bulletin of the British Arachnological Society 5:29-38. Craig, C.L. 1986. Orb- web visibility: The influ- ence of insect flight behaviour and visual phys- iology on the evolution of web designs within the Araneoidea. Animal Behaviour 34:54-68. Craig, C.L. & C.R. Freeman. 1991. Effects of predator visibility on prey encounter: A case study on aerial web weaving spiders. Behavioral Ecology & Sociobiology 29:249-254. Eberhard, W.G. 1986. Effects of orb- web geometry on prey interception and retention. Pp. 70-100, In Spiders — Webs, Behavior, and Evolution. (W.A. Shear, ed.). Stanford Univ. Press. Stanford, California. Eberh.ard, W.G. 1990. Function and phylogeny of spider webs. Annual Review of Ecological Sys- tems 21:341-372. Gillespie, R.G. & T. Caraco. 1987. Risk-sensitive foraging strategies of two spider populations. Ecology 68:887-899. Herberstein, M.E., K.E. Abemethy, K. Backhouse, H. Bradford, EE. de Crespigny, P.R. Luckock & M.A. Elgar. 1998. The effect of feeding history on prey capture behaviour in the orb-web spider Argiope keyserlingi (Araneae: Araneidae). Ethol- ogy 104:565-571. Herberstein, M.E,, C.L. Craig & M.A. Elgar. 2000. Foraging strategies and feeding regimes: Web and decoration investment in Argiope keyserlingi Karsch (Araneae: Araneidae). Evolutionary Ecology Research 2:69-80. Herberstein, M.E. & M.A. Elgar, 1994. Foraging strategies of Eriophora trammarina and Nephila plumipes (Araneae: Araneoidea): Nocturnal and diurnal orb-weaving spiders. Australian Journal of Ecology 19:451-457. Herberstein, M.E. & A.M. Heiling. 1998. Does mesh height influence prey length in orb-web spiders? European Journal of Entomology 95: 367-371. Herberstein, M.E. & I.M. Tso. 2000. Evaluation of formulae to estimate the capture area and mesh height of orb webs (Araneoidea: Araneae). Jour- nal of Arachnology 28(2): 180-1 84. Higgins, L.E. & R.E. Buskirk. 1992. A trap-build- ing predator exhibits different tactics for differ- ent aspects of foraging behaviour. Animal Be- haviour 44:485-499. McReynolds, C.N. & G.A. Polis. 1987. Ecomor- phological factors influencing prey use by two sympatric species of orb-web spiders, Argiope 350 THE JOURNAL OF ARACHNOLOGY aurantia and Argiope trifasciata (Araneidae)„ Journal of Arachnology 15:371-383. Nentwig, W. 1983. The non-filter function of orb webs in spiders. Oecologia 58:418-420. Opell, B.D. 1998. Economics of spider orb-webs: The benefits of producing adhesive capture thread and of recycling silk. Functional Ecology 12:613-624. Pasquet, A., A. Ridwan & R. Leborgne. 1994. Presence of potential prey affects web-building in an orb- weaving spider Zygiella x-notata. An- imal Behaviour 47:477-480. Peakal, D.B. & P.N. Witt. 1976. The energy budget of an orb web-building spider. Comparative Bio- chemistry & Physiology A 54:187-190. Sandoval, C.P. 1994. Plasticity in web design in the spider Parawixia bistriata: A response to variable prey type. Functional Ecology 8:701- 707. Schneider, J.M. & E Vollrath, 1998. The effect of prey type on the geometry of the capture web of Araneus diadematus. Naturwissenschaften 85: 391-394. Sherman, P. 1994. The orb web: An energetic and behavioural estimator of a spider’s dynamic for- aging and reproductive strategies. Animal Be- haviour 48:19-34. Uetz, G.W. & S.P. Hartsock. 1987. Prey selection in an orb-weaving spider: Micrathena gracilis (Araneae: Araneidae). Psyche 94:103-116. Wilkinson, L. 1992. 'Systat: Statistics.’ Version 5.2, Systat: Evanston, Manuscript received 20 August 1999, revised 30 June 2000. 2000. The Journal of Arachnology 28:351-352 SHORT COMMUNICATION ADANSONIA IS A BAOBAB TREE, NOT A THERIDIID SPIDER Ingi Agnarsson; Department of Biological Sciences, George Washington University, Washington, D.C. 20052 USA. ABSTRACT. The name Adansonia Saville-Kent was erroneously introduced into spider taxonomy by Bonnet in 1939 and still appears in the literature. Saville-Kent was referring to a tree, not describing a spider. Keywords: Theridiidae, Araneae, Adansonia, nomenclature, systematics Most biologists are familiar with the genus Ad- ansonia Linnaeus 1753, which contains the mag- nificent Baobab trees. The Baobabs are particularly prolific in Madagascar but also widely distributed in continental Africa, and one species is native to northwestern Australia. These trees are also popular in botanical gardens and parks in other parts of the world. Less well known is the mollusk Adansonia Pallary 1902. Pallary validly proposed this name as a subgenus of Donovania Bucquoy Dautzenberg & Dollfus, which in turn is now considered a junior synonym of the buccinid snail Chauvetia Monter- osato. Very few are aware of yet another use of the name Adansonia, but Adansonia Saville-Kent 1897 is currently listed as a generic name of the spider family Theridiidae (Platnick 1997). The latest use of the name is an error that can be traced back to a cataloging mistake by Bonnet (1939). This note is written to clarify the situation and prevent further inclusion of the name Adansonia Saville-Kent in spider taxonomy. The acclaimed author of the name is William Sa- ville-Kent (1845-1908), whose works include “The Great Barrier Reef of Australia” (Saville-Kent 1893) and “The Naturalist in Australia” (Saville- Kent 1897). In the latter he was discussing a ther- idiid spider: “A remaining spider form included in Chromo-Plate IX. invites brief notice. It is rep- resented by Figs. 12 to 15 [these show the de- tails of the egg cocoon and the general habitus of the spider]. This type is apparently referable to the genus Theridium, and is remarkable more especially with relation to its habits and the singular environments of its egg cocoon. It was observed by the writer in the neighbour- hood of Derby, at the head of King’s Sound, Western Australia, taking up its abode in the fissures of the gnarled trunks of the older Bao- bab or Bottle-trees, Adansonia rupestris. The spider, a small brown one, presents no special features of interest, and neither does the web, which is of the irregularly meshed order. Sus- pended in the snare, however, there is gener- ally present a little cupola-shaped mass, which, on near examination, is found to be composed superficially of the emptied skins and disjoint- ed limbs of a small species of black ant upon which this spider habitually feeds. The interior of this ant aggregation is hollow, and is found to contain in its upper confines the spherical silken egg cocoon of its fabricator, which it has most effectively and ingeniously concealed from view” (Saville-Kent 1897:261). It is clear from Saville-Kent’s text that he did not intend to describe a new species, and thus gives the spider no name, he is simply sharing some inter- esting observations with the reader. Bonnet (1939), however, mistakenly connected Saville-Kent’s de- scription of the spider to the Latin name of the Bao- bab and listed Adansonia Saville-Kent, as a new genus and Adansonia rupestris Saville-Kent as a new species (which he designated as the type spe- cies, by monotypy), in the family Theridiidae (Bon- net 1939:158)! Bonnet’s error does not appear in Levi & Levi’s (1962) exhaustive work on therediid genera, nor in the catalogs of Roewer (1942), Brig- noli (1983) or Platnick (1988), but Adansonia is listed as a theridiid genus in the two most recent spider taxonomy catalogs under the heading of “No 351 352 THE JOURNAL OF ARACHNOLOGY Entries” (Platnick 1993:180; 1997:248). Thus no one has used this name since Bonnet’s error. The argument might be made that Bonnet’s error can be considered as providing availability to Ad- ansonia rupestris Saville-Kent. To be available, ev- ery new name published after 1930 must be accom- panied by a bibliographic reference to a description that states in words, characters that are purported to differentiate the taxon (International Commission of Zoological Nomenclature, 1999: Art. 13.1.1 & 13.1.2). Thus Bonnet would make Adansonia ru- pestris available because he provides a reference to a description. However, Saville-Kent does not pro- vide a description of the spider, “The spider . . . presents no special features of interest and neither does the web . . .” (Saville-Kent 1897:261) but rather of the egg cocoon. The egg cocoon is the work of an animal and is clearly excluded from zoological nomenclature (International Commission of Zoological Nomenclature, Art. 13.6.2). There- fore, it is clear that by the publication of Adansonia in Bonnet’s (1939) catalog Bonnet was not making a new name available; he was merely cataloging what he thought was Saville-Kent’s new name {Ad- ansonia rupestris). But as there is no new Saville- Kent name. Bonnet does not accidentally validate a name. It is furthermore clear that Adansonia was in use for a mollusk genus (Pallary 1902) at the time of Bonnet’s publication and is therefore unavailable regardless of other things due to the principle of homonymy (International Commission of Zoologi- cal Nomenclature, 1999: Art. 52.1, 52.2 & 52.3). Adansonia Saville-Kent is thus a nomen nudum and Adansonia rupestris is a tree. After the removal of Adansonia there are cur- rently 73 valid genera in the family Theridiidae (Platnick 1997; Tanikawa 1998). I would like to thank Mark Harvey for making a copy of Saville-Kent’s paper available to me and with his help in locating Bonnet’s error. Chris Thompson provided valuable help on nomenclatural issues. Gustavo Hormiga, Mark Harvey, Chris Thompson, Jonathan A. Coddington, Petra Sier- wald, Matjaz Kuntner and Jeremy Miller provided comments on the manuscript. Support for this re- search was provided by a National Science Foun- dation grant to Gustavo Hormiga and Jonathan Coddington (DOEB 9712353) and the USIA Ful- bright Program. LITERATURE CITED Bonnet, P. 1939. Bibliographia Araneorum. Tome II. Faculty Des Sciences, Toulouse, France. 918 pp. Brignoli, PM. 1983. A Catalogue of the Araneae Described Between 1940 and 1981. Manchester Univ. Press. Manchester. 755 pp. International Commission of Zoological Nomencla- ture. 1999. International Code of Zoological No- menclature (4th ed )- International Trust for Zoo- logical Nomenclature, London. 306 pp. Levi, H.W. & L.R. Levi. 1962. The genera of the spider family Theridiidae. Bulletin of the Muse- um of Comparative Zoology 127(1): 1-71. Linnaeus, C. 1753. Species Plantarum. Stockholm. 1200 pp. Pallary, P. 1902. Liste des mollusques testaces de la Bale de Tanger. Journal de Conchy liologie 50(1):13. Platnick, N.I. 1988. Advances in Spider Taxonomy 1981-87. Manchester Univ. Press. New York. 673 pp. Platnick, N.I. 1993. Advances in Spider Taxonomy 1988-91. New York Entomological Society. New York. 846 pp. Platnick, N.I. 1997. Advances in Spider Taxonomy 1992-1995. New York Entomological Society. New York. 976 pp. Roewer, C. Fr. 1942. Katalog der Araneae von 1758 bis 1940, 1. Band. Kommissions-Verlag von Natura, Bremen. 1040 pp. Saville-Kent, W. 1893. The Great Barrier Reef of Australia; Its Products and Potentialities. W.H. Allen, London. 387 pp. Saville-Kent, W. 1897. The Naturalist in Australia. Chapman & Hall, Limited. London. 421 pp. Tanikawa, A. 1998. The new synonymy of the spi- der genus Argyrodes (Araneae: Theridiidae) and a description of a new species from Japan. Acta Arachnologica 47(l):21-26. Manuscript received 23 October 1999, revised 16 April 2000. 2000, The Journal of Arachnology 28:353-356 SHORT COMMUNICATION METOPIA SINENSIS (DIPTERA, SARCOPHAGIDAE), AN UNUSUAL PREDATOR OF LIPHISTIUS (ARANEAE, MESOTHELAE) IN NORTHERN THAILAND Peter J. Scliwendinger: Department of Arthropodology and Entomology I, Natural History Museum, P.O. Box 6434, CH-1211 Geneva, Switzerland Thomas Pape: Department of Entomology, Swedish Museum of Natural History, Box 50007, SE- 10405 Stockholm, Sweden Keywords: Liphistius, Metopia, predator, Thailand Knowledge on predators, parasites and par- asitoids of the Mesothelae is sparse and frag- mentary. Various authors reported on gamasid and erythraeid mites, mermithid nematodes and Rickettsia infesting either adult mesothe- lid spiders or their egg cases (Bristowe 1933; Yoshikura 1954; Platnick & Sedgwick 1984; Haupt 1979 & pers. comm.). Sedgwick & Platnick (1987) found a wasp pupa in an aban- doned burrow of L. endau Sedgwick & Plat- nick 1987 in Malaysia, and Bristowe (1976) reported the observation of an unidentified pompilid wasp larva attached to the abdomen of a paralysed L. desuitor Schiodte 1849 found inside its burrow on Penang Island, Ma- laysia. The latter author further mentioned empty pupal cases of a fly among the old frag- ments of L. bristowei Platnick & Sedgwick 1984 (misidentified as L. birmanicus Thorell 1897) in Thailand, probably on Doi Suthep (Mount Suthep), Chiang Mai, northern Thai- land (c/ Bristowe 1975). Recently one of us (PJS) found puparia and imagines of Milichia sp. (Diptera, Milichiidae, det. by J. Chainey, The Natural History Museum, London) among partly devoured eggs of L. yamasakii Ono 1988 in northern Thailand. Also puparia of another cyclorrhaphan fly were collected from several empty burrows of L. bristowei on Doi Suthep and nearby Doi Inthanon, but no imagines could be obtained for a proper identification (Schwendinger 1990). Later Schwendinger (1998) reported a second, more fortunate find. Imagines of Metopia sinensis Pape 1986 (Figs. l-=3; specimens deposited at the Swedish Museum of Natural History, Stockholm), a miltogrammine flesh fly, were raised from larvae infesting three or four car- casses of L. lahu Schwendinger 1998 inside the spider burrows at Doi Angkhang, about 150 km north of Doi Suthep. The fly species was at that time known only from a single specimen from southern China (Pape 1986a, 1996). The observation was cautiously inter- preted as a case of carrion feeding rather than predation or parasitism (Schwendinger 1998) because species of Metopia Meigen 1803 are known to be kleptoparasites in nests of soli- tary aculeate Hymeeoptera (Pape 1986b, 1987; Spofford et al. 1989). The fly larvae feed on prey stored for the host progeny. New observations reported and discussed in the fol- lowing, however, indicate that M. sinensis very likely is a primary predator. In December 1997 PJS again found fly pu- paria of M. sinensis in two empty Liphistius burrows (near the Thai-Myanmar border in Mae Hong Son Province, northern Thailand) in a colony of an undescribed species closely related to L. lahu. All puparia had already hatched; but one of the spiders collected, a lively and seemingly healthy female with a new egg case, carried six tiny fly larvae ven- trally between its leg coxae. The next day (af- ter transferring the spider to a laboratory in Chiang Mai) the spider was found motionless 353 354 THE JOURNAL OF ARACHNOLOGY Figures 1-5. — Metopia sinensis. 1. Male, habitus, dorsal view; 2. Male, frontal view of head; 3. Male, left lateral view of head. Note very large eyes, taking up almost entire side of head, and strongly-receding head profile with numerous facial bristles along the antenna. 4. Larvae of Metopia sinensis feeding on a dying Liphistius sp. female, second day after collecting; 5. Same, on third day after collecting. Scales: 1 = 2.0 mm; 2, 3 - 0.8 mm. and apparently dying, and by evening the lar- vae had moved onto the posterior part of its carapace (Fig. 4). On the fourth day the fully grown larvae (Fig. 5) abandoned the dead spi- der after having devoured most of the tissues inside the prosoma and anterior opisthosoma. On the fifth day all larvae pupariated, four di- rectly on the spider carcass, the remaining two in the container, a short distance away from the spider. Seventeen days later five imagines hatched, the sixth in the morning of the 18th day. The two male and four female flies were kept alive for two more days, during which they behaved quite atypically for sarcophagid flies: flying clumsily and unwillingly, most of the time hiding among the substrate. This may, however, be an aberrant behavior due to unnatural conditions in the laboratory. A search for M. sinensis at the same locality dur- ing two days in December 1998 was unsuc- cessful. All spiders examined were unaffected, no puparia were found and no adult flies were attracted to three live Liphistius females (de- posited in the Natural History Museum of Ge- neva) dug out of their burrows and placed as bait in uncovered containers. Species of Metopia are known to be klep- toparasites in nests of various aculeate wasps and bees of the families Pompilidae, Spheci- dae and Halictidae (very rarely also Vespi- dae), and the particular fly species seem to have a broad spectrum of hosts (Spofford et al. 1989). The flies have been classified as “hole-searchers” (e.g., Evans 1970; Spofford & Kurczewski 1990), which means that fe- males search for host nest entrances rather than trail the wasps themselves. The female flies may larviposit into the host burrow, ei- ther standing on the rim or flying low over the hole, and the larvae then wriggle down to the stored prey of the wasp. Or, the flies may enter the burrow to larviposit near or even onto the food source. The odor of the wasp presumably triggers gravid flies to larviposit after they SCHWENDMGER & PAPE—DIPTERAN PREDATOR ON LIPHISTIUS 355 have located the entrance of a host burrow (Endo 1980a,b). Being ‘‘hole-searchers,” spe- cies of Metopia pay only little attention to prey specimens that are dragged or otherwise transported by a potential host wasp. How- ever, individual wasps dragging prey close to the nest, and even more so those excavating burrows, may be attractive to female Metopia, The prey itself is not used as substrate for lar- viposition before it is deposited in the burrow, but female flies may occasionally larviposit directly onto the adult wasp. The latter appar- ently always turns out to be fatal for the fly larvae (Endo 1980a), The present observations are considered un- ambiguous evidence that the association be- tween M. sinensis and Liphistius spp. is not simple carrion-feeding. The possibility that the Metopia larvae were deposited on a spider left insufficiently paralyzed by a pompilid wasp is not considered very likely as the in- fested spider appeared in full vigor. Also, the repeated finds of flesh-fly puparia, here ten- tatively attributed to M, sinensis, in Liphistius burrows show that the association is persistent in time and not just a haphazard or freak lar- viposition. Note that the find of fly puparia by Bristowe (1976) may also refer to Metopia si- nensis. We have decided to classify M. sinen- sis as a predator rather than a parasite (or par- asitoid), following Price’s (1980) definition, which states that a parasite is primarily “an organism living in or on another living organ- ism.” As the larvae of M. sinensis apparently kill the spider and complete most of their lar- val life on the carcass, they behave more like predators, even if grossly outsized by their prey. The predator-prey association between Me- topia sinensis and Liphistius most likely de- veloped from kleptoparasitism. Sparse infor- mation from the literature and observations in Thailand by PJS indicate that pompilid wasps attack and paralyze Liphistius directly inside the spider burrows. In the case of M. sinensis, however, it appears that the wasps have com- pletely lost their role as food providers for the fly larvae. Predation appears to be rare or local and, at least in northern Thailand, confined to only a few Liphistius species. Other Liphistius spe- cies in the same area (i.e., L. yamasakii and L. lannaianus Schwendinger 1990) and spe- cies elsewhere in Thailand were repeatedly observed and collected in moderate numbers by PJS during more than seven years, yet none of them was ever seen affected by M. sinensis. In this context, it is interesting that the in- fested burrows of L. lahu at Doi Angkhang were only about 2 km away from a thriving population of L. lannaianus. On Doi Inthaeon, fly puparia (presumably of M. sinensis) were collected from scattered burrows of L. bris- towei at 1250 m, but not found in the dense colonies of L. yamasakii 350^530 m higher up. From the same mountain, at 1000 m, 11 flies were also raised from the carcass of a mygalomorph spider, Damarchus sp. (Neme- siidae), found inside its burrow. The flies were identified as Metopia sp. by Nigel Wyatt (The Natural History Museum, London) and pos- sibly also belong to M. sinensis (specimens unfortunately lost after identification). While local prey specificity cannot be ruled out, a broader prey spectrum seems very like- ly considering the known distributional range of M. sinensis, which is much larger than that of its known prey. Liphistius bristowei, L. lahu and the related undescribed species are at present known only from northern Thai- land; the latter two probably also occur across the border in Myanmar. We thank Joachim Haupt (Berlin) for pro- viding literature and for sharing his unpub- lished observations on pathogens and para- sites of mesothelid spiders with us. Nigel P. Wyatt and John Chainey (both London) pro- vided help with certain identifications, and Torbjom Kronestedt (Stockholm) and Nikolaj Scharff (Copenhagen) kindly commented on the manuscript. Mrs. Elisabeth Binkiewicz skillfully produced the illustrations of Metopia sinensis. LITERATURE CITED Bristowe, W.S. 1933. The liphistiid spiders. With an appendix on their internal anatomy by J. Mil- let. Proc. Zool. Soc. London, 1932:1015-1057. Bristowe, W.S. 1975. An interesting spider found in Thailand. Nat. Hist. Bull. Siam Soc., 26:166- 167. Bristowe, W.S. 1976. A contribution to the knowl- edge of liphistiid spiders. J. ZooL, London, 178: 1-6. Endo, A. 1980a. The behaviour of a miltogram- mine fly Metopia sauteri (Townsend) (Diptera, Sarcophagidae) cleptoparasitizieg on a spider wasp Episyron arrogans (Smith) (Hymenoptera, Pompilidae). Kontyu, 48(4):445-457. 356 THE JOURNAL OF ARACHNOLOGY Endo, A. 1980b. On the host-cleptoparasite rela- tionship between the spider wasp Episyron ar- rogans (Smith) (Hymenoptera, Pompilidae) and the miltogrammine fly Metopia sauteri (Town- send) (Diptera, Sarcophagidae). Japanese J. Ecol., 30(2): 117-132. (in Japanese with English summary) Evans, H.E. 1970. Ecological-behavioral studies of the wasps of Jackson Hole, Wyoming. Bull. Mus. Comp. ZooL, 140(7):45 1-511. Haupt, J. 1979. Lebensweise und Sexual verhalten der mesothelen Spinne Heptathela nishihirai n. sp. (Araneae, Liphistiidae). Zool. Anz., 202(5/6): 348-374. Pape, T. 1986a. A revision of Oriental and eastern Palaearctic species of Metopia Meigen (Diptera: Sarcophagidae). Stuttgarter Beitr. Naturk. A: Biol., 395:1-8. Pape, T. 1986b. Afrotropical species of Metopia (Insecta, Diptera, Sarcophagidae). Steenstrupia, 124(4):73-84. Pape, T. 1987. Revision of Neotropical Metopia Meigen (Diptera: Sarcophagidae). Syst. Ento- mol., 12:81-101. Pape, T. 1996. Catalogue of the Sarcophagidae of the world (Insecta: Diptera). Mem. Entomol., In- ternational, 8:1-558. Platnick, N.I. & W.C. Sedgwick. 1984. A revision of the spider genus Liphistius (Araneae, Meso- thelae). American Mus. Novit., 2781:1-31. Price, PW. 1980. The evolutionary biology of par- asites. Princeton Univ. Press, Princeton, New Jer- sey. xi + 237 pp. Schwendinger, PJ. 1990. On the spider genus Li- phistius (Araneae: Mesothelae) in Thailand and Burma. Zool. Scripta, 19(3):33 1-351. Schwendinger, P.J. 1998. Five new Liphistius spe- cies (Araneae, Mesothelae) from Thailand. Zool. Scripta, 27(1): 17-30. Sedgwick, W.C. & N.I. Platnick. 1987. A new spe- cies of Liphistius (Araneae, Mesothelae) from Jo- hore, Malaysia. Malayan Nature J., 41:361-363. Spofford, M.G. & EE. Kurczewski. 1990. Com- parative larvipositional behaviours and clepto- parasitic frequencies of Nearctic species of Mil- togrammini (Diptera: Sarcophagidae). J. Nat. Hist., 24:731-755. Spofford, M.G., EE. Kurczewski & W.L. Downes, Jr. 1989. Nearctic species of Miltogrammini (Diptera: Sarcophagidae) associated with species of Aculeata (Hymenoptera: Vespoidea, Pompi- loidea, Sphecoidea, Apoidea). J. Kansas Ento- mol. Soc., 62(2):254-267. Yoshikura, M. 1954. Embryological studies on the liphistiid spider, Heptathela kimurai. Kumamoto J. Sci. B, 1954:41-48. Manuscript received 1 July 1999, revised 17 Feb- ruary 2000. 2000. The Journal of Arachnology 28:357-360 SHORT COMMUNICATION THE USE OF FRUITS BY THE NEOTROPICAL HARVESTMAN NEOSADOCUS VARIABILIS (OPILIONES, LANIATORES, GONYLEPTIDAE) Glauco Machado: Museu de Historia Natural, Instituto de Biologia, Universidade Estadual de Campinas, CP 6109, 13083-970, Campinas, SP, Brazil Marco A. Pizo: Departamento de Botanica, Instituto de Biociencias, Universidade Estadual Paulista. CP 199. 13506-900. Rio Claro. SP. Brazil Keywords: Diet, fruits, harvestmen, Neosadocus Harvestmen are solitary, nocturnal foragers that have a variety of feeding habits, ranging from scavenging to predation (see review in Gnaspini 1996). Although harvestmen seem to be generalist omnivorous arthropods, ac- cepting both plant and animal matter, several species show a tendency to camivory (Bris- towe 1949; Capocasale & Bruno-Trezza 1964; Anuradha & Parthasarathy 1976; Gnaspini 1996; Machado et al. 2000). Reports of fru- givory in harvestmen are scarce and in gen- eral are restricted to captive animals (Capo- casale & Bruno-Trezza 1964; see also Gnaspini 1996). In this paper we provide the first detailed account of frugivory by a har- vestman species, and investigate if fruit size and chemical content of the fleshy portion can influence fruit use by the harvestmen. The study was conducted from October 1995 to February 1997 in the lowland forest of the Parque Estadual Intervales (24°14'S, 48°04'W), a 490 km^ reserve located in the Ribeira Valley, Sao Paulo state, southeast Bra- zil. The study site (Saibadela Research Sta- tion, elevation 70 m) receives about 4200 mm of rainfall a year, with no month receiving less than 100 mm. Rainfall, however, is less in- tense and less frequent between April and Au- gust, when the temperature may drop to nearly 10 °C (mean ± SD = 20.8 °C ± 2.5 for the study period). This period contrasts with the wetter period (September-March) when tem- peratures may reach 42 °C (25.7 °C ± 2.8). The vegetation is predominantly composed of old-growth forest (sensu Clark 1996) with an open understory and trees reaching up to 30 m. The fruits of the following trees were used to investigate frugivory in harvestmen: Virola oleifera (Myristicaceae), Eugenia stictosepala (Myrtaceae), Cabralea canjerana (Meli- aceae), Citharexylum myrianthum (Verbena- ceae), Alchornea glandulosa and Hyeronima alchorneoides (Euphorbiaceae), throughout this paper referred to only by their generic names. Besides their availability, these fruits were selected for study because (1) they fall within three discrete size classes commonly found in tropical forests (Corlett 1996; see Ta- ble 1); (2) all of them are covered by a thin skin which allows the exploitation by har- vestmen, and (3) they fit within two distinct extremes relative to the lipid content of their fleshy portions; the arils of Virola, Cabralea and Alchornea are lipid-rich, while the pulps of Eugenia, Citharexylum and Hyeronima are lipid-poor (Table 1). The fruits of Eugenia, Citharexylum and Hyeronima are drupes bear- ing one {Eugenia and Hyeronima) or two seeds {Citharexylum). The fruits of the re- maining species are capsules that open to ex- pose the 1-12 fruits, i.e., seeds coated by red {Virola and Alchornea) or orange {Cabralea) arils. These fruits are eaten by birds, monkeys and/or bats which frequently drop many fruits under the parent plants (Galetti 1996; Pizo 357 358 THE JOURNAL OF ARACHNOLOGY Table 1. — Fruit maturation period, morphology, size class (following Corlett 1996) and chemical com- position of the six fruits studied. Morphological values are means ± SD. At least 20 fruits of each species were weighed. L = lipids, P = protein, TC = total carbohydrate (i.e., soluble + structural carbohydrates). Lipids, proteins and ashes were analyzed according to the methods described in Bligh & Dyer (1959), AAC (1995, method # 46-13) and AOAC (1984, method # 22027), respectively. Total carbohydrates were obtained by difference. Fruit Maturation period Morphology Fresh weight Total weight (g) of pulp/aril (g) Size class Hyeronima Mar-Apr 0.05 ± 0.01 0.03 ± 0.01 small Alchornea Oct-Nov 0.09 ± 0.01 0.03 ± 0.01 small Citharexylum Feb-Mar 0.9 ± 0.2 0.7 ± 0.2 medium Cabralea Sep-Dec 0.9 ± 0.3 0.09 ± 0.02 medium Eugenia Apr-May 5.8 ± 1.2 2.1 ± 1.0 large Virola Jul-Oct 3.5 ± 1.2 1.1 ± 0.5 large 1997). The period of fruit maturation for the six plant species is presented in Table 1. Voucher specimens of the harvestman were deposited in the Museu de Zoologia da Univ- ersidade de Sao Paulo (MZUSP), and plants at the herbarium of the Universidade Estadual Paulista at Rio Claro (HBRC). We made diurnal and nocturnal censuses (35 days) of the harvestmen attending fresh fruits placed on the forest floor along a tran- sect established 1-2 m off one of the trails that crossed the study site. One hundred fruits of Virola and 50 fruits of the other five species were set along the transect 5 m apart. Each fruit was protected from vertebrate removal by wire cages (15 X 15 X 10 cm, 1.5 cm mesh) closed on the top and staked to the ground. Plastic wraps placed on the top of cages pro- tected fruits and harvestmen from being dis- turbed by light to moderate rains. No census was conducted under heavy rains for which the plastic shelters were useless. Fruits were set on the transect at 0800 h and checked at four-hour intervals throughout a 24 hour pe- riod. The daily light period at the study site span from 0600-1800 h, thus rendering two diurnal and two nocturnal censuses. Only one harvestman species, Neosadocus variabilis (Mello-Leitao 1935), was recorded on the fruits. Individuals of N. variabilis were observed exploiting fruits of Cabralea, Al- chornea, Eugenia, and Virola (Table 1). No harvestman was seen consuming fruits of Citharexylum and Hyeronima. Large fruits as a whole were more exploited than smaller ones (x^ = 10.94, df = 1, P < 0.001; medium and small fruits combined). Although there is a tendency for the harvestmen to exploit lipid- rich fruits as compared to lipid-poor ones (5.5% vs. 2.0%, respectively), the difference did not reach statistical significance (x^ = 2.73, df = 1, P = 0.09). Neosadocus variabilis seems to be a strictly nocturnal forager since it was recorded only during the night census- es, i.e., from 2000-0400 h. Individuals fed on the pulp or aril of the fruit on the spot, never displacing them. Although Walker (1928) stated that har- vestman diet can consist of fruit juices and other plant-derived matter, few studies have documented fruit use by species of the order. Edgar (1971) observed the palpatorid Leiob- unum vittatum (Say 1821) feeding on a ripe wild raspberry and, among the Laniatores, Acanthopachylus aculeatus (Kirby 1819) ac- cepts papaya in the laboratory (Capocasale & Bruno-Trezza 1964), while Neosadocus var- iabilis was seen eating fallen fruits in the field (Gnaspini 1996; this study). Despite the scar- city of records, the exploitation of fallen fruits by ground-dwelling harvestmen is possibly more common than previously thought, espe- cially for those species inhabiting tropical rainforest where a great amount of fleshy fruits is produced on a year-round basis (Jor- dano 1993). At our study site, for example, more than 500 kg/ha/year of fleshy fruits reach the forest floor (Pizo unpubl. data) al- most continuously through the year (Morellato et al. 1999). Results regarding the choice of fruits by N. variabilis based on their size and lipid content MACHADO & PIZO— USE OF FRUITS BY NEOSADOCUS VARIABILIS 359 Table 1. — ExteEded. Percent of , Chemical composition (percent of dry mass) % of harvestmen water L P TC visiting 85.6 7.9 6.3 — 0 43.3 68.4 7.6 21.7 2 81.4 6.3 6.8 82.7 0 47.7 70.8 10.3 16.5 2 77.9 5.2 8.5 85.5 6 62.7 61.8 4.6 32.1 9 must be interpreted cautiously since small fruits are rapidly removed by ants (Pizo & Oliveira 2000), thus becoming unavailable for harvestmen. In any case, large fruits may rep- resent more attractive food sources because they bear a great amount of fleshy material, either pulp or aril (Table 1). The use of lipid- rich fruits by harvestmen, on the other hand, deserves further investigation. In our study, N. variabilis exploited all the three lipid-rich fruits tested, and only the largest lipid-poor fruit. The fruits tested also differ in their car- bohydrate content. This is expected since lip- ids and carbohydrates are highly negatively correlated in our fruit sample (Spearman rank correlation: r^ = -=0.90, n = 3, P = 0.03), as usually occurs for fleshy fruits in general (Jor~ dano 1993; Pizo & Oliveira 2000). There is no a priori reason to suspect that harvestmen would avoid carbohydrate-rich fruits. The em- phasis on the role of the lipid content of the fruits in their use by harvestmen, on the con- trary, is justified because it has been shown that lipid-rich fruits serve as food for carniv- orous arthropods such as ponerine ants (Horv- itz & Beattie 1980; Pizo & Oliveira 1998, 2000), and also attract other non-fragivorous arthropods, e.g., cockroaches and grasshop- pers (Pizo unpubl. data). Carroll & Jaezen (1973) hypothesized that, from the ants’ view- point, the lipid-rich fruits may be chemically analogous to their insect prey, an idea sup- ported by the comparison made by Hughes et al. (1994) between the fatty acid composition of elaiosomes, lipid-rich food bodies of typi- cal myrmecochorous fruits, and insects. Given that elaiosomes and the arils of lipid-rich fruits are chemically and morphologically similar structures (Hughes et al. 1993), it is possible that harvestmen use these fruits more often than we have previously suspected. We are grateful to PS, Oliveira, A.V.L. Freitas and two anonymous reviewers for helpful comments on the manuscript; to the Fundayao Florestal do Estado de Sao Paulo for permitting our work at Parque Intervales. The study was supported by fellowships from FAPESP to M.A. Pizo, and CAPES to G. Ma- chado. LITERATURE CITED Anonymous. 1984. Official Methods of Analysis. Association of Official Analytical Chemists, Washington. Anonymous. 1995. American Association of Ce- real Chemists. St. Paul, Minnesota. Anuradha, K. & M.D. Partharasathy. 1976. Field studies on the ecology of Gagrellula saddlana Roewer (Palpatores, Opiliones, Arachnida) and its behaviour in the laboratory condition. Bull. Ethol. Soc. India, 1:68-71. Bligh, E.G. & W.J. Dyer. 1959. A rapid method of total lipid extraction and purification. Canadian J. Biochem. Phys., 37:911-917. Bristowe, W.S. 1949. The distribution of harvest- men (Phalangida) in Great Britain and Ireland, with notes on their names, enemies and food. J. Anim. Ecol., 18:100-114. Capocasale, R. & L.B. Brano-Trezza. 1964. Biol- ogia de Acanthopachylus aculeatus (Kirby, 1819), (Opiliones: Pachylinae). Rev. Soc. Uru- guay a EntomoL, 6:19-32. Carroll, C.R. & D.H. Janzen. 1973. Ecology of for- aging by ants. Ann. Rev. Ecol. Syst., 4:231-257. Clark, D.B. 1996. Abolishing virginity. J. Trop. EcoL, 12:735-739. Corlett, R.T. 1996. Characteristics of vertebrate- dispersed fruits in Hong Kong. J, Trop. Ecol., 12: 819-833. Edgar, A. L. 1971. Studies on the biology and ecol- ogy of Michigan Phalangida (Opiliones). Misc. Publ. Mus. Zool. Univ. Michigan, 144:1-64. Galetti, M. 1996. Fruits and fragivores in a Bra- zilian Atlantic forest. Ph.D. thesis, University of Cambridge, Cambridge, UK. 360 THE JOURNAL OF ARACHNOLOGY Gnaspini, R 1996. Population ecology of Gonio- soma spelaeum, a cavernicolous harvestman from southeastern Brazil (Arachnida: Opiliones: Gonyleptidae). J. Zool., 239:417-435. Horvitz, C.C. & A.J. Beattie. 1980. Ant dispersal of Calathea (Marantaceae) seeds by carnivorous ponerines (Formicidae) in a tropical rain forest. American J. Bot., 67:321-326. Hughes, L., M. Westoby & D. Johnson. 1993. Nu- trient costs of vertebrate- and ant-dispersed fruits. Funct. Ecol., 7:54-62. Hughes, L., M. Westoby & E. Jurado. 1994. Con- vergence of elaiosomes and insect prey: Evi- dence from ant foraging behaviour and fatty acid composition. Funct. EcoL, 8:358-365. Jordano, P. 1993. Fruits and frugivory. Pp. 105- 156, In Seeds: The ecology of regeneration in plant communities (M. Fenner, ed.). CAB Inter- national, Wallingford. Machado, G., R.L.G. Raimundo & PS. Oliveira. 2000. Daily activity schedule, gregariousness, and defensive behaviour in the Neotropical har- vestman Goniosoma longipes (Opiliones: Gony- leptidae). J. Nat. Hist., 34:587-596. Morellato, PC., D.C. Talora, A. Takahasi, C.C. Bencke, E.C. Romera & V.B. Zipparro. In press. Phenology of Atlantic rain forest trees: A com- parative study. Biotropica. Pizo, M.A. 1997. Seed dispersal and predation in two populations of Cabralea canjerana (Meli- aceae) in the Atlantic forest of southeastern Bra- zil. J. Trop. Ecol., 13:559-578. Pizo, M.A. & PS. Oliveira. 1998. Interaction be- tween ants and seeds of a nonmyrmecochorous neotropical tree, Cabralea canjerana (Meli- aceae), in the Atlantic forest of southeast Brazil. American J. Bot., 85:669-674. Pizo, M.A. & PS. Oliveira. 2000. The use of fruits and seeds by ants in the Atlantic forest of south- east Brazil. Biotropica. Walker, M.E. 1928. A revision of the order Phal- angida of Ohio. Bull. Ohio Biol. Surv., 19:149- 175. Manuscript received 10 June 1999, revised 10 Jan- uary 2000. 2000. The Journal of Arachnology 28:361-363 SHORT COMMUNICATION HOMALONYCHUS THEOLOGUS (ARANEAE, HOMALONYCHIDAE): DESCRIPTION OF EGGSACS AND A POSSIBLE DEFENSIVE POSTURE Richard S. Vetter: Department of Entomology, University of California, Riverside, California 92521 USA James C. Cokendolpher: 2007 29th Street, Lubbock, Texas 79411 USA ABSTRACT. Presented here is a description of the cryptic, sand-covered eggsacs of Homalonychus theologus. Additionally, when this species is gently harassed, it adopts a rigid, paired-leg position which may be a defensive posture functioning in immobility and possibly mimicking cactus spines. Keywords: Spider, eggsac, defensive behavior, immobility The spider family Homalonychidae is represent- ed by a single North American genus consisting of two species. Homalonychus spiders are found in the deserts of extreme southeastern California, the southern tip of Nevada, southwestern Arizona, northwestern Sonora, and Baja California. These spiders are not commonly encountered, and the sparse information that is known regarding its nat- ural history was presented by Roth (1984). We had the opportunity to examine a few individuals of H. theologus Chamberlin 1924 and present our obser- vations on two aspects. Eggsacs. — Roth (1984) mentioned only one eggsac for the genus that was collected in April but “nothing is recorded regarding either the placement of the egg sac or its description.” In the original description of H. positivus (= H. selenopoides Marx 1891), Chamberlin (1924) reported a collec- tion from Guaymas, Sonora, Mexico, in which three females were “taken under stones with egg sacks” on 12 April 1921. Because the specimens reported by Chamberlin were type specimens which Roth might have studied for his revision, it is possible that the eggsac mentioned by Roth was from the same collection. Apparently, Roth overlooked the information presented by Chamberlin. Two female H. theologus were captured (Cali- fornia: San Bernardino County, 9 #1, off Amboy Rd by Sheep Hole Pass, 640 m, 16 February 1997, M. Holman, O. Trout; 9 #2, 5 km S Amboy, 150 m, on salt flat under a board, 25 April 1998, R. Vetter) and maintained in 2.5 liter plastic containers with sand substrate. (9 #1 had beach sand and 9 #2 had sand from its natural habitat.) Crumpled pa- per toweling served as refugia, and females were maintained until death. Upon cleaning out their containers within two days of each female’s death, it was discovered that in the hidden recesses of the paper towels, each spider had produced two round- ed, sand-covered eggsacs (Fig. 1). The eggsacs were about 18 mm in diameter and smooth inside, being lined with silk (Fig. 2). Eggsacs from 9 #1 were transferred to a 4 liter plastic jar and maintained in the first author’s home at temperatures of 24-30 °C. Spiderlings were first noticed on the toweling 53 days later although the development time probably was longer because the date of oviposition was unknown. Seventeen spi- derlings were collected, and examination of the eggsacs revealed 22 shed skins in one eggsac and no evidence of shed skins, spiderlings nor infertile eggs in the other. Eggsacs of 9 #2 were removed from the female’s container and examined. One contained 17 shed skins and 8 desiccated, presum- ably infertile eggs; it is unknown where the spider- lings dispersed. The other eggsac contained 20 spi- derlings and one desiccated egg. Sixteen of the 20 spiderlings appeared to have died molting from the 1st to 2nd instar; of the remaining four, two were dead and two were moribund. No spiderlings were found outside of the eggsacs. Although the maintenance of these spiders was artificial, the eggsacs were similar in construction to one observed under a rock in Punta Diggs (17 km S San Felipe, Baja California Norte, Mexico, under rocks on sandy desert soil, S. Johnson, pers. 361 362 THE JOURNAL OF ARACHNOLOGY Figures 1, 2. — Eggsacs of Homalonychus theologus. 1. Sand-covered eggsac of Homalonychus theo- logus suspended from the underside of a paper towel; 2. Close-up of eggsac with view of inside. An egg is visible in the interior of the sac. (Photos by P. Kirk Visscher) comm.). In our artificial setting, H. theologus pro- duced 21, 22, and 25 eggs per eggsac. These fe- males were probably fed only about once per month and, therefore, the egg total here may be low due to the sparse food supply. Because we collected one specimen at night while it roamed around in sparse- ly vegetated desert yet saw none during the day while conducting experiments over extensive peri- ods of time in the same locale, we speculate that H. theologus spiders spend daytime in rodent bur- rows and under rocks where they easily could affix their eggsacs. The sand and silk covering would probably aid in humidity/temperature control as well as camouflaging the eggsac to avoid detection by potential predators or parasites. Although Hom- alonychus spiders are found partially buried in sand (Roth 1984; pers. obs.), they do not appear to con- struct burrows nor remain hidden in the sand during the daylight hours. Potential defensive posture. — When at rest, Homalonychus spiders position themselves with all legs spread out from one another (Fig. 3; also see Roth 1984: fig. 14). When disturbed, H. theologus shifts its legs to a rigid “paired-leg” formation (first two legs forward, hind two legs rearward) (Fig. 4). When a mature (12 mm body length) female (Cal- ifornia: Riverside County, Cactus City, 17 km W Chiriaco Summit off I- 10, 400 m, at night wander- ing, 23 March 1997, R. Vetter) was held by her legs with a pair of forceps she could be rotated in all positions without becoming limp or attempting to run. This behavior can also be elicited by touching the spider with a pencil or forceps, when at rest or while moving, day or night. However, a 3 mm Homalonychus juvenile from the same locale did not adopt this posture when chased for several min- utes in two separate trials. Therefore, propensity to display this behavior may be size dependent (i.e., the spider’s potential as a prey item). When a pen- ultimate H. theologus female (9 mm body length) was uncovered under a rubber tire, (California: San Bernardino County, 5 km S Amboy, 150 m, 26 April 1998, R. Vetter), she moved several cm from her initial spot, became immobile, adopted the “paired leg” stance (which was maintained while being maneuvered into a 40 dram vial), slid down the length of the vial and did not abandon this po- sition until she was slid back out of the vial into the collector’s hand. We did not have sufficient numbers of specimens to attempt additional tests and, therefore, we can only speculate on the mechanism of the behavior if, indeed, it is defensive in function. The primary defenses of H. theologus are nocturnal activity and crypsis including burial in sand, the cryptic aspect of which is enhanced by the spider’s dorsal hairs VETTER & COKENDOLPHER^-i/OMALOiVFOTra BEHAVIOR 363 Figures 3, 4. — -Postures of Homalonychus theologus. 3. Characteristic resting posture of Homalonychus with legs held fiat and spread equidistant from one another. This penultimate male molted and was not placed back on sand. Hence, it shows its natural coloration without sand trapped amongst hairs; 4, Posture of a female H. theologus with rigid body and paired legs, possibly a defense mechanism. (Photos by J. C. Cokendolpher) which trap small sand grains (except in mature males) (Roth 1984). Immobility is a common gen- eral defense among animals (Cott 1940). Cloudsley- Thompson (1995) mentions death-feigning (thana- tosis) in an exhaustive review of spider defensive behaviors; however, no behavior such as we have seen in H. theologus is mentioned. Additionally, thanatosis in spiders usually involves holding the legs tucked in close to the body. The rearrangement of the legs in. H. theologus is somewhat puzzling as two legs held together would seem to increase the spider's conspicuousness, and hence belie its cryp- tic nature. Possibly, this leg orientation provides a novel image to a predator accustomed to eating spi- ders. Predators are known to avoid novel stimuli (Cott 1940) although we doubt that predators will care whether its prey have “4” or 8 legs. We would like to offer one additional speculative hypothesis. Because H.theologus is both nocturnal and a desert dweller, possibly the immobility in concert with paired-leg posture mimics the appearance of de- tached spines of dead cactus which could be an ef- fective defense in the desert at night when visibility is poor. Voucher specimens are housed at the California Academy of Sciences. We thank M. Holman and O. Trout for providing us the first ovipositing female and S. Johnson for sharing information on his observations of an egg- sac in Mexico. LITERATURE CITED Chamberlin, R.V 1924. The spider fauna of the shores and islands of the Gulf of California. Proc. California Academy of Sciences, 4th Ser. 12:561-694. Cloudsley-Thompson, J.L. 1995. A review of anti- predator devices of spiders. Bulletin of the Brit- ish Arachnological Society 10:81-96. Cott, H.B. 1940. Adaptive Coloration in Animals. Methuen, London. Roth, V.D. 1984. The spider family Homalonychi- dae (Arachnida, Araneae). American Museum Novitates #2790, 11 pp. Manuscript received 1 July 1999, revised 16 April 2000. 2000. The Journal of Arachnology 28:364 ARACHNOLOGICAL RESEARCH FUND The AAS Fund for Arachnological Re- search (AAS Fund) is funded and admin- istered by the American Arachnological So- ciety. The purpose of the fund is to provide research support for work relating to any aspect of the behavior, ecology, physiology, evolution, and systematics of any of the arachnid groups. Awards may be used for field work, museum research (including travel), expendable supplies, identification of specimens, and/or preparation of figures and drawings for publication. Monies from the fund are not designed to augment or re- place salary. Individual awards will not exceed $1000.00, and, although open to all stu- dents and faculty with less than $500.00 per year research budget, preference will be given to students. A total of $6000.00 is available for awarding during each funding year. Available monies could be expended for three large proposals, a greater number of partially funded proposals, and/or a num- ber of smaller, less expensive proposals. The final funding pattern is at the discretion of the review committee. Applications for support should be re- ceived by the chair of the review committee no later than January 15. To be considered for an award from the AAS Fund, please submit four copies of a proposal of no more than five pages (including references) de- tailing your research project. Proposals should have three main parts: 1) an INTRODUCTION where background information is presented relative to the pro- posed work. The introduction should in- clude a section which places the proposed work in context with currently known rel- evant information, 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 revi- sions, the type of synthesis that will be achieved and its significance; 2) a METH- ODS section where the methods, materials, experimental design, and statistical or tax- onomic analysis(ses) to be used are clearly and concisely presented, and 3) a BUDGET explaining (in detail) how monies awarded will be spent in the proposed research. Proposals should be submitted to: Dr. Deborah Smith, AAS Eund Chair Department of Entomology Haworth Hall University of Kansas Lawrence, KS 66045 USA Proposals must be submitted in English. The four copies of the proposal must be in the hands of the Fund chair by the appro- priate deadlines to be considered. Electron- ic submission and/or EAX submission of proposals with hard copies to follow is ac- ceptable only if the other copies arrive be- fore the stated deadlines. If these submis- sion rules are difficult or prohibitive be- cause of cost, erratic postal services, or re- mote location (remote field stations or sites), other methods of submission may be acceptable. For other submission possibili- ties, please contact the chair of the Fund at the above address, or electronically at dsmith@kuhub.cc.ukans.edu. Alternative submissions will be accepted only if the chair has been previously contacted, and all deadlines will still apply. 364 INSTRUCTIONS TO AUTHORS (revised August 2000) Manuscripts are accepted in English only. Authors whose primary language is not English may consult the editors for assistance in obtaining help with manuscript preparation. All manuscripts should be prepared in general accordance with the current edition of the Council of Biological Editors Style Manual unless instructed otherwise below. Authors are advised to con- sult a recent issue of the Journal of Arachnology for additional points of style. Manuscripts longer than three printed journal pages should be prepared as Feature Articles, shorter papers as Short Commun- ications. Send four identical copies of the typed mate- rial together with copies of illustrations to the Managing Editor of the Journal of Arachnology: Petra Sierwald, Managing Editor; Division of In- sects, Dept, of Zoology, The Field Museum of Na- tural History, 1400 South Lakeshore Drive, Chicago, IL 60605 USA [Telephone: (312)-665-7744; FAX: (312)-665-7754; E-mail psierwald@fhinh.org]. The Managing Editor will forward your manuscript to one of the Subject Editors for the review process. You will receive correspondence acknowledging the receipt of your manuscript from the responsible subject editor or the managing editor, with the manuscript number of your manuscript. Please use this number in all correspondence regarding your manuscript. Corres- pondence relating to manuscripts should be directed to the Managing Editor or the appropriate Subject Editor. After the manuscript has been accepted, the author will be asked to submit the manuscript on a computer disc in a widely-used word processing program. Indicate clearly on the computer disc the word processing pro- gram and the type of computer (Mac or PC). FEATURE ARTICLES Title page. — The title page will include the complete name, address, and telephone number of the author with whom proofs and correspondence should be exchanged, a FAX number and electronic mail address if available, the title in capital letters, 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. A second abstract, in a language pertinent to the nationality of the author(s) or geographic region(s) emphasized, may be included. Keywords. — Give 3-5 appropriate keywords fol- lowing the abstract. Text. — Double-space text, tables, legends, etc. throughout. Three categories of headings are used. The first category (METHODS, RESULTS, etc.) is typed in capitals and on a separate line. The second category of heading, in bold tj^e, begins a paragraph with an indent and is separated from the text by a period and a dash. The third category of heading may or may not begin a paragraph but is italicized and separated from the text by a colon. Use only the metric system unless quoting text or referencing collection data. All decimal fractions are indicated by the period (e.g.,-0.123). 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). Include a letter of permission from any person who is cited as providing unpublished data in the form of a personal communication. Literature cited section. — ^Use the following style, and include the fiill unabbreviated journal title. Lombardi, S.J. & D.L. Kaplan. 1990. The amino acid composition of major ampullate gland silk (drag- line) of Nephila clavipes (Araneae, Tetragnathidae). Journal of Arachnology 18:297-306. Krafft, B. 1982. The si^ificance and complexity of communication in spiders. Pp. 15-66, In Spider Communications: Mechanisms and Ecological Significance. (P.N. Witt & J.S. Rovner, eds.). Princeton University Press, Princeton, New Jersey. Footnotes. — Footnotes are permitted only on the first journal page to indicate current address or other information concerning the author. These are placed together on a separate manuscript page. Tables and fig- ures may not have footnotes. Running head. — The author’s sumame(s) and an abbreviated title should be typed all in capital letters and must not exceed 60 characters and spaces. The running head should be placed near the top of the title page. Taxonomic articles. — Consult a recent taxonomic article in the Journal of Arachnology for style, or contact the Subject Editor for Systematics. Papers containing the original taxonomic description of the focal arachnid taxon should be listed in the Literature Cited section. Tables. — ^Each table, with the legend above, should be placed on a separate manuscript page. Only horizon- tal lines (usually three) should be included. Tables may not have footnotes; instead, include all information in the legend. Make notations in the text margins to indi- cate the preferred location of tables in the printed text. Illustrations. — Address all questions concerning illustrations to the Editor of the Journal of Arach- nology: James W. Berry, Editor; Dept, of Biological Sciences; Butler University, Indianapolis, Indiana 46208 USA. [Telephone (317)-940-9344; FAX (317)- 940-9519; E-mail: jwberry@butler.edu]. All art work must be camera-ready (mounted and labeled) for reproduction. Figures should be arranged so that they fit (vertically and horizontally) the printed journal page, either one column or two columns, with a mini- mum of wasted space. When reductions are to be made by the printer, pay particular attention to width of lines and size of lettering in line drawings. Multiple photos assembled on a single plate should be mounted with only a minimum of space separating them. In the case of multiple illustrations mounted together, each illus- tration must be numbered sequentially rather than given an alphabetic sequence. Written on the back should be the name(s) of author(s) and an indication of top edge. Indicate whether the illustration should be one column or two columns in width. The overall dimensions should be no more than 1 1 inches (28 cm) X 14 inches (36 cm). Larger drawings present greater difficulty in shipping and greater risks of damage for which the JoA assumes no responsibility. In manu- scripts for review, photocopies are acceptable, and should be reduced to the exact measurements that the author wants to appear in the final publication. Make notations in the text margins to indicate the preferred position of illustrations in the printed text. Color plates can be printed, but the author must assume the full cost, currently about $600 per color plate. Legends for illustrations should be placed together on the same page(s) and separate from the illustrations. Each plate must have only one legend, as indicated below: Figures 1-4. — A-us x-us, male from Timbuktu: 1. Left leg; 2. Right chelicera; 3. Dorsal aspect of geni- talia; 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, footnotes, tables with legends, figure legends, figures. Page charges and reprints. — There are no page charges, but the author will be charged for changes made in the proof pages. Reprints are available only from the Allen Press and should be ordered when the author receives the proof pages. Allen Press will not accept reprint orders after the paper is published. SHORT COMMUNICATIONS The above instructions pertaining to Feature Articles apply also to Short Communications, which should be prepared in the same manner as regular Feature Articles. Short Communications are usually limited in length to three journal pages, including tables and figures. They will be printed in a smaller (10 point) typeface. CONTENTS The Journal of Arachnology Volume 28 Feature Articles Number 3 A New Species of the Genus Kimula (Opiliones, Minuidae) from the Dominican Republic by Abel Perez Gonzalez & Luis F. de Armas .... 257 Systematics of the Genus Dysdera (Araneae, Dysderidae) in the Eastern Canary Islands by Miquel A. Arnedo, Pedro Oromi & Carles Ribera 261 New Species and Records of Kleptochthonius from Indiana (Pseudoscor- pionida, Chthoniidae) by William B. Muchmore 293 Spider Size and Locomotion on the Water Surface (Araneae, Pisauridae) by Robert B. Suter & Jessica Gruenwald 300 Chemical Cues from Ants Influence Predatory Behavior in Habrocestum pulex^ an Ant-eating Jumping Spider (Araneae, Salticidae) by Robert J. Clark, Robert R. Jackson & Bruce Cutler 309 Life History of Pardosa moesta and Pardosa mackenziana (Araneae, Lycosidae) in Central Alberta, Canada by Christopher M. Buddie . . 319 A Structured Inventory of Appalachian Grass Bald and Heath Bald Spider Assemblages and a Test of Species Richness Estimator Performance by Douglas S. Toti, Frederick A. Coyle & Jeremy A. Miller 329 Does the Presence of Potential Prey Affect Web Design in Argiope keyser- lingi (Araneae, Araneidae)? by Marie E. Herberstein, Anne C. Gaskett, Deborah Glencross, Simon Hart, Sue Jaensch & Mark A. Elgar 346 Short Communications Adansonia Is a Baobab Tree, Not a Theridiid Spider by Ingi Agnarsson . . 351 Metopia sinensis (Diptera, Sarcophagidae), an Unusual Predator of Liphistius (Araneae, Mesothelae) in Northern Thailand by Peter J. Schwendinger & Thomas Pape 353 The Use of Fruits by the Neotropical Harvestman Neosadocus variabilis (Opiliones, Laniatores, Gonyleptidae) by Glauco Machado & MarcoA. Pizo 357 Homalonychus theologus (Araneae, Homalonychiidae): Description of Eggsacs and a Possible Defensive Posture by Richard S. Vetter & James C. Cokendolpher 361 Announcement Arachnological Research Fund 362 i