2005. The Journal of Arachnology 33:101±109

DIVERSITY AMONG GROUND-DWELLING ASSEMBLAGES: HABITAT GENERALISTS AND SPECIALISTS

Rachael E. Mallis and Lawrence E. Hurd1: Department of Biology, Washington & Lee University, Lexington, Virginia 24450 USA. E-mail: [email protected]

ABSTRACT. We sampled assemblages of ground-dwelling with pitfall traps in six terrestrial habitats representing a successional gradient in southwestern Virginia, during the summer of 2002. Ap- proximately half of the 50 species trapped were habitat specialists with low abundance, found at only one of the sites, which is qualitatively consistent with the literature. Only four species, Schizocosa ocreata (Hentz 1844), Pirata insularis (Emerton 1885) Pirata aspirans (Chamberlain 1904) and mag- na (Keyserling 1887) were found at as many as four sites. A few species that were found in more than one study from disparate geographical communities, such as terricola (Thorell 1856) tended also to be relatively abundant habitat generalists. In general, the majority of spider species found in studies such as ours that examined multiple sites were habitat specialists and had low abundance. For our sample sites, there was no relationship between any measure of spider diversity (S, H', J') and successional age. Our results, and those of most other published studies, are consistent with the hypothesis that spider assemblages do not undergo succession and except for a few very common generalist species the com- position of these communities is unpredictable, and may depend more on stochastic colonization and speci®c resource requirements of specialists following immigration than on any predictable association with successional parameters. Keywords: Cursorial spiders, habitat specialization, spider diversity, succession

The importance of predators in the structure old ®eld, and forest litter communities (Hurd and function of natural ecosystems is becom- & Eisenberg 1990; Riechert & Bishop 1990; ing increasingly well documented (Terborgh Moran et al. 1996; Lawrence & Wise 2000). et al. 2001). Spiders are widespread and di- Given their demonstrated importance to the verse predators that are part of terrestrial ar- structure and function of many communities, thropod assemblages (Wise 1993) and arthro- it is important to gather information on the pods comprise more than half of known distribution and abundance of cursorial spider species (Wilson 1992). Cursorial spiders in species. Often it has been dif®cult to deter- particular are the dominant preda- mine what features of an environment deter- tors in many terrestrial communities, e.g., mine which, or how many, species of cursorial grasslands (Weeks & Holtzer 2000) and forest spiders will be present. For example, spider ¯oor litter (Uetz 1979). Their position in tro- diversity may not follow a trend toward in- phic structure of communities often is com- creasing diversity with increasing succession- plex: spiders in forest litter belong to both the al age (Hurd & Fagan 1992; Aitchison & decomposition and the grazing food webs be- Sutherland 2000; Buddle et al. 2000) that has cause they consume detritivores/fungivores been a traditional expectation for species of and herbivores (Uetz 1975; Wise et al. 1999). plants and during terrestrial succes- As larger species of wolf spiders mature, they sion (Odum 1969). prey more on herbivores that are part of the Spiders have legendary powers of dispersal grazing food web (Uetz 1975; McNabb et al. and often are among the ®rst colonizers of dis- 2001). Spiders have been experimentally dem- turbed sites (Hodkinson et al. 2001); the ®rst onstrated to exert important effects on the known colonist of Krakatoa was a spider populations of other in a variety of (Spiller et al. 1998). Many spiders have the experimental systems, including agricultural, ability to disperse by ``ballooning'' with silk at some point in their life cycles (Hodkinson 1 Corresponding author. et al. 2001). Lycosids and gnaphosids balloon

101 102 THE JOURNAL OF ARACHNOLOGY

Figure 1.ÐMap of the Science Park on the campus of Washington & Lee University. The habitats used for sampling sites are described in the text: CG ϭ cedar grove; OF ϭ old ®eld; DR ϭ disturbance recovery; LH ϭ lowland hardwoods; UH ϭ upland hardwoods; PW ϭ Pine woods. as juveniles, while many linyphiids retain the der guilds is far from being answered, and will capability throughout adulthood (Mrzljak & require the accumulation of much more data Wiegleb 2000). However adept they are at ini- (Uetz et al. 1999). tially getting to new sites, cursorial spiders We report here on a ®eld study in which six should have speci®c habitat preferences that ®eld sites, representing different successional dictate which species will become established seres, are compared with respect to ground- and how abundant they will be. Some species dwelling spider species captured by pitfall are more particular than others: along a suc- trap sampling. We compared diversity among cessional gradient in Delaware, the lycosid sites, and the extent of habitat specialization Pirata insularis (Emerton 1885) was found (relative number of habitats in which species abundantly in all four communities examined, were found in this, and in previous studies) in whereas the gnaphosid Zelotes hentzi (Bar- these spiders. rows 1945) was rare and con®ned to the youn- gest successional site (Hurd & Fagan 1992). METHODS The structure of vegetation and some physico- Study Sites.ÐThe study sites we used for chemical habitat parameters may determine a this study are located in the 35ha Science Park spider's habitat choice (Mrzljak & Wiegleb of Washington & Lee University in Rock- 2000). Along forest litter gradients and in bridge Co, Virginia, USA (Fig. 1). The sites agroecosystems, lycosids manifest microhab- were chosen to represent different habitat itat preferences possibly based upon leaf litter, types and stages of secondary terrestrial suc- herbaceous vegetation and available moisture cession typical of this region, as described be- (Weeks & Holtzer 2000). The question of low in order of successional age. what determines the structure of cursorial spi- 1. Disturbance recovery (DR): This site is MALLIS & HURDÐCURSORIAL SPIDER DIVERSITY 103 a 0.5ha patch of level ground, surrounded by placed by hardwoods. The ground is mostly woods, that is being used in a long-term study in shade, with small patches of sunlight most of community succession. The ground had of the day. The soil is sandy and well-drained. been stripped of all native top soil prior to The litter layer consists of a Յ 1.0cm layer of 2001, leaving a bare, hard clay surface. Dur- dead cedar needles. We considered this site to ing fall and winter 2001 ground leaf mulch be the earliest forest community, between OF was applied to the bare clay and tilled into the and PW in age. upper 2±3cm, after which seeds of 19 native 4. Pine woods (PW): This site is in deep forb species and ®ve native grasses were shade, provided by an overstory of tall planted. The plant assemblage that sprouted (Ն10m) white pines (Pinus strobus) and during spring and summer 2002 constituted a mixed species of smaller hardwoods, espe- dense mixture of these intentionally planted cially red maple (Acer rubrum) and white ash species and incidental species that were pres- (Fraxinus americana), with an understory of ent as seed in the mulch (an especially con- saplings and shrubs. This species mix repre- spicuous member of the second category dur- sents an alternative intermediate stage of suc- ing late summer was ragweed, Ambrosia cession to the assemblage in CG; the relative artemisiifolia). The vegetation, which reached height of the hardwoods in this stand indicated a height of approximately 0.5m, was dense that PW may be the older of the two, at least enough to provide shade at the soil surface. 40 years old. The site is located on a ¯ood There was no organic litter layer on the soil, plain, with relatively most soil. The litter layer owing to the absence of plants on the site dur- is mainly dead needles, but somewhat thicker ing the previous year. This site is surrounded (Ն1cm). by a wire fence to keep out deer, and is sur- 5. Upland hardwoods (UH): This site is rounded by woods. within the woodlot boundary of the DR site. 2. Old-®eld (OF): This site is a third year The overstory is of mixed hardwoods 10±15m successional sere, which was previously sub- high, especially white ash and tulip poplar jected to mowing once or twice annually. It is (Liriodendron tulipifera), with several species fully open to the sun, and consists of grasses of oaks (Quercus spp.), and red maple in the and forbs growing up to approximately 0.5m understory. This species mix is typical of the high typical of early successional old-®elds in community that replaces coniferous species, this region, including patches of emergent (ca i.e., perhaps 10±15 years older than site PW 1.5m high) late-season goldenrod (Solidago (i.e., 50±55 years old). The soil here is well- spp.), teasel (Dipsacus sylvestris), and rag- drained and rich in organic matter, with a rel- weed (Ambrosia artemisiifolia). Plant height atively deep (3±4cm) leaf litter layer. and soil shading at this site were similar to 6. Lowland hardwoods (LH): This is a ma- those at site DR (see above), but here there ture hardwood forest that has been subjected was a sparse and shallow (Յ0.5cm depth) lit- to little disturbance for at least the past 70 ter layer consisting of dead plant stalks and years, with canopy trees (tulip, maple, white leaves from the previous year's growth. An ash, and mixed oak) Ն 30m. This site is deep- abundant and diverse arthropod assemblage ly shaded and is on a ¯oodplain at the bottom inhabits this site, the most conspicuous of of a slope, downward from site UH. The leaf which are several species of grasshoppers litter is more compacted than and not as thick (Acrididae). The cedar grove site (CG, see be- as at site UH, probably because of the in- low) forms the western border of the old-®eld creased soil moisture and humidity. site. Sampling.ÐWe sampled ground-dwelling 3. Cedar grove (CG): This site consists of spiders at each site with six pitfall traps set a near monoculture of 6±8m high eastern red out in a 2 ϫ 3 array, such that no trap was cedar (Juniperus communis) from 30±40 closer than approximately 1.5m from its near- years old, with a sparse understory of hard- est neighbor. Trap arrays were at least 10m wood saplings and very little herbaceous from the edge of the habitats they sampled. ground cover. In our geographical area, cedars Each trap consisted of a 10cm diameter, 11cm grow in old-®elds until they shade out her- deep polypropylene cup ®tted into a perma- baceous vegetation, and then may dominate nent sleeve that was sunk into the ground the community until they senesce and are re- ¯ush with the soil surface. A cover for each 104 THE JOURNAL OF ARACHNOLOGY

Table 1.ÐTotal numbers of cursorial spiders by habitat type (sampling site): DR ϭ disturbance recovery; OF ϭ old ®eld; CG ϭ cedar grove; PW ϭ pine woods; UH ϭ upland hardwoods; LH ϭ lowland hard- woods. Superscripts denote other studies in which species were found: 1 ϭ Hurd & Fagan (1992); 2 ϭ Buddle et al. (2000); 3 ϭ Aitcheson & Sutherland (2000); 4 ϭ Bonte et al. (2002); 5 ϭ Gajdos & Toft (2000); 6 ϭ Uetz (1975); 7 ϭ Uetz (1976); 8 ϭ Uetz (1977); 9 ϭ Uetz (1979); 10 ϭ Buddle and Rypstra (2003); and 11 ϭ Draney (1997). Nomenclature follows Platnick (2003).

DR OF CG PW UH LH Species Family ;species Total Total Total Total Total Total Total Koch 1837 Agelenopsis pennsylvanica1 Koch 33 1843 Gertsch 1940 unicolor Hentz 11 2 1842 Corinnidae Karsch 1880 Castianeira cingulata1, 6, 8, 9, 11 11 Koch 1841 Castianeira longipalpus1, 6, 7, 8, 9, 11 134 Hentz 1847 Phrurotimpus alarius1, 6, 8, 9 Hentz 2 4 16 22 1847 Phrurotimpus borealis3, 6, 9 Emer- 325 ton 1911 Phrurotimpus minutus1, 6 Banks 112 1892 Phrurotimpus sp. Chamberlin & 112 Ivie 1935 Scotinella britcheri Petrunke- 11 vitch 1910 Scotinella formica1 Banks 1911 1 8 9 Scotinella sp. Banks 1911 1 1 Banks 1892 Cybaeus sp. Koch 1868 3 1 1 5 Dysderidae Koch 1837 Dysdera crocata4 Koch 1838 1 1 Gnaphosidae Pocock 1898 Drassyllus depressus Emerton 11 1890 Hahniidae Bertkau 1878 Hahnia cinerea3, 9 Emerton 1890 1 5 6 brunnea Emerton 1909 1 1 2 Hahnidae sp. 1 134 Neoantistea agilis3, 11 Keyserling 11 2 1887 Neoantistea magna Keyserling 1111 4 1887 Blackwall 1859 Drapetisca alteranda Chamber- 11 lin 1909 Subfamily: Linyphiinae Stemonyphantes lineatus Linnae- 11 us 1758 Tenuiphantes zebra Emerton 11 1882 Lycosidae Sundevall 1833 Allocosa funerea1, 11 Hentz 1844 2 1 3 Allopecosa aculeata Clerck 1757 1 1 Arctosa virgo Chamberlin 1925 1 1 Hogna helluo1, 6, 10 Walckenaer 11 2 1837 distincta1 Blackwall 21 3 1846 Pardosa milvina1, 10, 11 Hentz 11 11 1844 Pardosa saxatilis1, 6 Hentz 1844 16 16 Pardosa sp. 1 Koch 1847 1 1 Pardosa sp. 2 Koch 1847 7 7 Pirata aspirans1, 6 Chamberlin 12 1 2 6 1904 MALLIS & HURDÐCURSORIAL SPIDER DIVERSITY 105

Table 1.ÐContinued.

DR OF CG PW UH LH Species Family Genus;species Total Total Total Total Total Total Total Pirata insularis1, 3, 6 Emerton 25 37 1 141 204 1885 Rabidosa rabida1, 11 Walckanaer 55 1837 Schizocosa avida1 Walckanaer 34 7 1837 Lycosidae Sundevall 1833 Schizocosa bilineata1, 11 Emerton 22 1885 Schizocosa ocreata1, 6, 7, 8, 11 615 315 Hentz 1844 Trochosa terricola1, 2, 3, 4, 5 Tho- 21 3 rell 1856 Trochosa sp Kock 1847 1 1 Oonopidae Simon 1890 Orchestina saltitans Banks 1894 1 1 Oxyopidae Thorell 1870 Oxyopes salticus1, 11 Hentz 1845 2 2 Philodromidae Thorell 1870 Thanatus formicinus11 Clerck 11 1757 Pisauridae Simon 1890 tenebrosus Hentz 112 1844 Salticidae Blackwall 1841 Habronattus borealis Banks 11 1895 Neon nelii3 Peckham & Peckham 112 1888 Theridiidae Sundevall 1833 Euryopsis argentea Emerton 11 1882 Thomisidae Sundevall 1833 Ozyptila sp. Simon 1864 1 1 Xysticus ferox1, 7, 9, 11 Hentz 1847 1 1 2 Xysticus punctatus Keyserling 0 1880 Xysticus sp. Koch 1835 1 1 Habitat Total Abundance ϭ 50 35 41 47 31 178 382

trap was constructed using a Petri dish lid Linyphiidae), which were infrequently cap- with nails to elevate it 3cm above the lip of tured, and most of which were represented by the trap. These covers kept out rainwater and a single individual. At least one individual of falling debris. Sampling occurred at weekly each species collected (a male and female of intervals from early June to mid-August, and each, when available) was preserved in Kah- then once at the end of September 2002 (total le's ¯uid as part of a reference collection. ϭ 354 trap-days). Each time we sampled, we As with any ®eld study in a diverse species put approximately 2cm of 70% ethanol into assemblage, sampling ef®cacy is not likely to each trap during the afternoon (ca 1600h), and be equal among taxa. In the case of pitfall collected the samples 16±18h later. traps, for instance, the most active spiders All adult spiders collected from the traps (e.g., many lycosids) may have a tendency to were counted and identi®ed using taxonomic be disproportionately sampled relative to more keys (Kaston 1978, 1981; online taxonomic sedentary species (e.g., clubionids). There- updates http://kaston.transy.edu/spiderlist/ fore, species richness and relative abundance Kaston78.htm and http://kaston.transy.edu/ of captured spiders may not accurately re¯ect spiderlist/kast.htm; Roth 1993). Our nomen- the entire resident assemblage, but can be clature follows Platnick (2003). We did not used for comparisons among sites of those attempt to enumerate or identify to species taxa that are susceptible to pitfall trapping. spiders in the subfamily Erigoninae (family Data analysis.ÐWe compared sampling 106 THE JOURNAL OF ARACHNOLOGY

Table 2.ÐThe number of shared cursorial species between habitats, and species diversity (richness ϭ S; JЈϭevenness; Shannon's diversity ϭ HЈ) for each habitat based on pit trap samples. Sites arranged in order of increasing successional age from left to right and top to bottom. Site descriptions given in Methods.

Sites: DR OF CG PW UH LH DR 14 Ð ÐÐÐ Ð OF 6 18 ÐÐÐ Ð CG 0 3 12 ÐÐ Ð PW1249 ÐÐ UH116518 Ð LH2134913 HЈϭ 2.06 2.57 1.45 0.94 2.63 0.96 S ϭ 14 18 12 9 18 13 JЈϭ 0.78 0.89 0.58 0.43 0.91 0.37 sites with regard to diversity, measured as (1) 2). The pine stand (PW), representing the in- the number of species found, or species rich- termediate stage of succession, had the lowest ness, S, (2) Evenness of distribution of indi- spider diversity (Table 2). This site also had viduals among species, J', and (3) Shannon's the lowest apparent vegetational diversity diversity, H', which is a measure of the inter- among the six sites: there was almost no action between evenness and richness (Pielou ground cover vegetation, and the tree diversity 1969; Hill 1973). was limited to white pine and a few small de- RESULTS ciduous saplings. However, there were no oth- er apparent correlative trends between spider We collected a total of 50 species of diversity and site structure. The highest H' di- ground-dwelling spiders from our six sam- versity and J' evenness values we found were pling sites (Table 1). Twenty-six of these were in the nearly mature forest (UH), and the old habitat specialists, found at only one site; no ®eld (OF), our second to youngest site, yield- species was found at all six sites. The spiders ing virtually identical rank abundance patterns with the broadest distribution, found at four of (Fig. 2). Although species richness was the the six sites, were the lycosids Schizocosa same (18) for both of these sites, they only ocreata, Pirata insularis, Pirata arenicola and shared a single species, N. magna (Table 2). the hahniid Neoantistea magna. Both S. The climax forest (LH) had the lowest H' val- ocreata and P. insularis were found at all four ue even though species richness was about av- wooded sites and none was collected at either erage among the sites. This was because of of the open ®eld sites. However, P. arenicola the high dominance of a single species (Pirata and N. magna were found at combinations of wooded and open ®eld sites. insularis), which was re¯ected by the low val- Some of the spiders we found appear to ue of J' (Tables 1 & 2). have wide geographical distributions. Five Our most abundant species trapped was P. species in Table 1 were also reported in four insularis, accounting for more than half of all or more other studies from Denmark, Belgium spiders trapped (Table 1). The abundance of and Manitoba, as well as sites in the U. S. P. insularis in our traps was highest in mid- (Delaware, Ohio, Georgia and Virginia). June, decreasing to just two individuals caught However, there appears to be no reliable re- in August and September. From the beginning lationship between how broadly cursorial spi- of the sample season, the sex ratio of this spe- der species are distributed among geographic cies was highly male-biased. As the season sites, how many sites they occupy within a progressed it shifted to a female-biased ratio. study, or what kind of habitat (e.g., wooded July appeared to be the month of reproduc- or open) they prefer in those studies that sam- tion: females were caught with egg sacs on 2 pled more than one habitat type. July, and with juveniles riding on the dorsa of We found no relationship between succes- their abdomens on 10 and 19 July. sional age and any measure of diversity (Table We also were able to record some repro- MALLIS & HURDÐCURSORIAL SPIDER DIVERSITY 107 ductive data for S. ocreata and P. saxatilis. most abundant lycosids (Pardosa milvina and On 1 July we found a female S. ocreata with P. saxatilis) were con®ned to the most open a new (white) egg case. On 10 July a female site (DR). Buddle & Rypstra (2003) also not- with a gray (older) egg case was captured. The ed that Pardosa species achieve dense popu- case was dissected and almost fully developed lations in barren exposed habitats. Schizocosa eggs were found inside, with fangs evident. ocreata was found in all four wooded sites. On 19 July a female was caught with juveniles This species is commonly found in leaf litter on her back. On 20 June we found a female of deciduous forests in eastern North America P. saxatilis, with an egg case. On two other (Wagner & Wise 1996). Uetz (1977) noted occasions (2 and 17 July) we found females that S. ocreata occurs in simple litter where with egg cases. On 26 July we found one with the leaves are compressed and the ground is spiderlings on her dorsum, and brought it back relatively moist. to the lab for observation. On 29 July spider- The diversity of sampled spiders in our lings were seen leaving the mother's dorsum study did not follow a successional gradient, and by 30 July they had all dispersed. a ®nding of other studies as well: a forest suc- cessional gradient in Delaware (Hurd & Fagan DISCUSSION 1992), and forests in Manitoba (Aitchison & As with previously reported studies, we Sutherland 2000; Buddle et al. 2000). Part of found that most ground-dwelling spider spe- the dif®culty may lie in the relative scarcity cies were habitat specialists, found at one or of studies that examine a wide range of suc- two sites, and very few were generalists. Be- cessional seres at a given locale. In any event, cause rare species may be present in such low attempting to ®nd predictable environmental numbers that they may be missed by sam- correlates to spider diversity have proved frus- pling, we cannot conclude that a species that trating for many researchers. In their 20 year did not show up in our samples was complete- study of coastal dunes Gajdos & Toft (2000) ly absent from a given site, but we can at least found that temporal changes in community score a species as present if we captured it in composition were greater than differences oc- a sample. This is a problem common among curring between habitats. It was impossible studies that report the presence of rare species, for them to determine what ways ecological many of which are represented by a single characteristics changed for those spider spe- trapped individual (e.g., 31 of 105 species re- cies in which abundance changed over time. ported by Buddle et al. 2000). Differences in the physical structure of leaf Both Aitchison & Sutherland (2000) and litter and its complexity can in¯uence species Hurd & Fagan (1992) found only three species composition, spider abundance and diversity that occupied four or more sites; we found generally increasing with increased litter only four species in that category. However, depth in some studies (Uetz 1975, 1977, 1979; very few of these are the same species. Six of Buddle & Rypstra 2003). Uetz (1975) found the species we found were also reported from that weather patterns, which could be tied to these two studies in Manitoba, one of which prey productivity for spiders, did not correlate (Trochosa terricola) was found in both Man- signi®cantly to any diversity measure. Instead, itoba studies and in our present study (Table he found that richness and evenness were re- 1) and has been reported to occur from as far lated to litter depth, and moderately well re- away as Finland (Aitchison & Sutherland lated to successional age and plant cover. 2000) and Belgium (Bonte et al. 2002). Not However, in our study we found as many spi- surprisingly, there were more spider species der species in the two open habitats with al- (20) in common between our present study most no litter (DR and OF) as we did in the and the geographically closer Delaware sites two hardwood forest habitats (UH and LH). of Hurd & Fagan (1992). The most abundant Mrzljak & Wiegleb (2000) presented evidence species in both the present Virginia study and that species richness and abundance are lim- the Delaware study was the lycosid, Pirata in- ited by vegetative strati®cation and height, sularis. e.g., tall grass stands had more species than The majority of spiders we encountered be- short grass stands. Hurd & Fagan's (1992) longed to the family Lycosidae. While P. in- study of spider assemblages in Delaware sularis preferred wooded sites, the next two found the main difference to be between 108 THE JOURNAL OF ARACHNOLOGY woodland and open habitats and not age of ating emigration of two species (Ar- the habitats: diversity of cursorial spiders gen- anae:Lycosidae) in an agroecosystem. Environ- erally was greater in open ®eld habitats than mental Entomology 32:88±95. in woodlands. However, in our present study Buddle, C.M., J.R. Spence & D.W. Langor. 2000. we found no clear difference among sites Succession of boreal forest spider assemblages following wild®re and harvesting. Ecography 23: based on presence, absence, or extent of tree 424±436. cover. Diamond, J.M. 1975. Assembly of species com- It is apparently not dif®cult to predict the munities. Pp. 342±444. In Ecology and Evolu- presence in spider assemblages of some very tion of Communities (M.L. Cody and J.M. Dia- broad generalists such as Trochosa terricola, mond, eds.). Belknap Press, Massachusetts. but for most habitat specialists, such predic- Draney, M.L. 1997. Ground-layer species (Aranae) tion is problematic. Given the data so far, it is of a Georgia piedmont ¯ood plain agroecosys- hard to refute the null hypothesis that spider tem: species list, phenology and habitat selection. diversity within a site may be more a function Journal of Arachnology 25:333±351. of stochastic colonization opportunities of dif- Gajdos, P. & S. Toft. 2000. A twenty-year compar- ison of epigeic spider communities (Aranae) of ferent species rather than a set of intra-com- Danish coastal heath habitats. Journal of Arach- munity assembly rules (sensu Diamond 1975). nology 28:90±96. Other factors that can in¯uence species mem- Hanski, I.A. 1999. Metapopulation Ecology. Oxford bership in arthropod assemblages, including University Press, Oxford, U.K. spiders, are habitat features such as area, de- Hill, M.O. 1973. Diversity and evenness: a unifying gree of isolation, and movement patterns of notion and its consequences. Ecology 54:427± animals relative to their resource requirements 432. (Matter 1996, 2000; Hanski 1999; Marshall et Hodkinson, I.D., S.J. Coulson, J. Harrison & N R. al. 2000; Samu et al. 2003). Thus, changing Webb. 2001. What a wonderful web they weave: spider community composition over time is spiders, nutrient capture and early ecosystem de- velopment in the high ArcticÐsome counter-in- not really true succession at all, but rather re- tuitive ideas on community assembly. Oikos 95: peated colonization by opportunistic species. 349±352. The success of such colonists, once they in- Hurd, L.E. & R.M. Eisenberg. 1990. Arthropod vade a habitat, may well depend on competi- community responses to manipulation of a bi- tive abilities (Marshall et al. 2000) and the trophic predator guild. Ecology 71:2107±2114. changing environmental conditions that ac- Hurd, L.E. & W.F. Fagan. 1992. Cursorial spiders company plant succession (Mrzljak & Wie- and succession: age or habitat structure? Oecol- gleb 2000; Hodkinson et al. 2001), but as yet ogia 92:215±221. we are far from being able to predict cursorial Kaston, B.J. 1978. How to Know the Spiders. Wm. spider composition among seres with any de- C. Brown Company Publishers, Dubuque, Iowa. gree of precision. Kaston, B.J. 1981. Spiders of Connecticut. State Geological and Natural History Survey of Con- ACKNOWLEDGMENTS necticut, Hartford, Connecticut. Lawrence, K.L. & D.H. Wise. 2000. Spider preda- This study is part of ongoing long-term re- tion on forest-¯oor Collembola and evidence for search on succession and disturbance recovery indirect effects on decomposition. Pedobiologia in the Science Park of Washington & Lee Uni- 44:33±39. versity. Our research was supported by grants Marshall, S.D., S.E. Walker & A.L. Rypstra. 2000. from the R. E. Lee (REM) and John M. Glenn A test for a differential colonization and com- (LEH) research programs of the university. petitive ability in two generalist predators. Ecol- ogy 81:3341±3349. LITERATURE CITED Matter, S.F. 1996. Interpatch movement of the red Aitchison, C.W. & G.D. Sutherland. 2000. Diversity milkweed beetle, Tetraopes tetraophthalmus: in- of forest upland communities in Mani- dividual responses to patch size and isolation. toba taiga (Aranae, Opiliones). Canadian Field Oecologia 105:447±453. Naturalist 114:636±651. Matter, S.F. 2000. The importance of the relation- Bonte, D., L. Baert & J.P. Maelfait. 2002. Spider ship between population density and habitat area. assemblage structure and stability in a hetero- Oikos 89:613±619. geneous coastal dune system (Belgium). Journal McNabb, D.M., J. Halaj & D.H. Wise. 2001. Infer- of Arachnology 30:331±343. ring trophic positions of generalist predators and Buddle, C.M. & A.L. Rypstra. 2003. Factors initi- their linkage to the detrital food web in agroeco- MALLIS & HURDÐCURSORIAL SPIDER DIVERSITY 109

systems: a stable isotope analysis. Pedobiologia species diversity of wandering spiders (Aranae) 45:289±297. in deciduous forest litter. Environmental Ento- Moran, M.D., T.P. Rooney & L.E. Hurd. 1996. Top- mology 4:719±724. down cascade from a bitrophic predator in an Uetz, G.W. 1976. Pitfall trapping in ecological stud- old-®eld community. Ecology 77:2219±2227. ies of wandering spiders. Journal of Arachnology Mrzljak, J. & G. Wiegleb. 2000. Spider coloniza- 3:101±111. tion of former brown coal mining areasÐtime or Uetz, G.W. 1977. Coexistence in a guild of wan- structure dependent? Landscape & Urban Plan- dering spiders. Journal of Ecology 46: ning 1:131±146. 531±541. Odum, E.P. 1969. The strategy of ecosystem devel- Uetz, G.W. 1979. The in¯uence of variation in litter opment. Science 164:262±270. habitats on spider communities. Oecologia 40: Pielou, E.C. 1969. An Introduction to Mathematical 29±42. Ecology. Wiley, New York. Uetz, G.W., J. Halaj & A.B. Cady. 1999. Guild Platnick, N.I. 2003. The world spider catalog, ver- structure of spiders in major crops. Journal of sion 4.0. American Museum of Natural History, Arachnology 27:270±280. online at http://research.amnh.org/entomology/ Wagner, J.D. & D.H. Wise. 1996. Cannibalism reg- spiders/catalog81-87/index.html. ulates densities of young wolf spiders: evidence Riechert, S.E. & L. Bishop. 1990. Prey control by from ®eld and laboratory experiments. Ecology an assemblage of generalist predators: spiders in 77:639±652. garden test systems. Ecology 71:1441±1450. Weeks, Jr., R.D. & T.O. Holtzer. 2000. Habitat and Roth, V.D. 1993. Spider Genera of North America. season in structuring ground-dwelling spider American Arachnological Society. (Aranae) communities in a shortgrass steppe eco- Samu, F., A. Sziranyi & B. Kiss. 2003. Foraging in system. Environmental Entomology 29:1164± agricultural ®elds: local `sit-and-move' strategy 1172. scales up to risk-averse habitat use in a wolf spi- Wilson, E.O. 1992. The Diversity of Life. Harvard der. Animal Behaviour 66:939±947. University Press, Cambridge. Spiller, D.A., J.B. Cosos & T.W. Schoener. 1998. Wise, D.H. 1993. Spiders in Ecological Webs. Cam- Impact of a catastrophic hurricane on island pop- bridge University Press, New York. ulations. Science 281:695±697. Wise, D.H., W.E. Snyder & P. Tuntilbunpakul. Terborgh, J., L. Lopez, P. NunÄez, M. Rao, G. Sha- 1999. Spiders in decomposition food webs of habuddin, G. Orihuela, M. Riveros, R. Ascanio, agroecosystems: theory and evidence. Journal of G. H. Adler, T. D. Lambert & L. Balbas. 2001. Arachnology 27:363±370. Ecological meltdown in predator-free forest frag- ments. Science 294:1923±1926. Manuscript received 19 May 2003, revised 9 Feb- Uetz, G.W. 1975. Temporal and spatial variation in ruary 2004.