The spider species of the Great Lakes States.

ABSTRACT. Critical analysis of existing spider species lists for Wisconsin, Michigan, Ohio, Indiana and Illinois reveals 900 species recorded from the five-state region (284 genera, 40 families). All non-native, Palearctic, or otherwise questionable species records were scrutinized, and their status is discussed. The most speciose families in the region are the Linyphiidae (almost 24% of species), Salticidae (10.3%), Theridiidae (8.9%), Lycosidae (8.8%), and Araneidae (7.7%). All sources used for spider species names and species records are unambiguously quoted. Spider species records are presented in tables allowing comparison of family composition among the states, and prediction of number of heretofore unrecorded species. Richness among states is analyzed and found to be dependent on varying degrees of sampling effort. As a new tool, a Spider Species Name Concordance Table allows tracking previously published spider species names to the currently valid name of every species record. The study demonstrates the need for crucial pieces of scientific infrastructure, such as complete species catalogs, and the great utility of faunistic and taxonomic data to meet today's biodiversity challenges.

Keywords: Midwest spider fauna, checklist, faunistics, Araneae, gap analysis, Illinois, Indiana, Ohio, Wisconsin, Michigan

**********

In the past, faunal studies and alpha-taxonomic work played a significant role in biological research. Their importance and influence in biological science diminished during the past 6 decades (Wheeler 2004), especially in the developed (and supposedly well-studied) regions in the world, such as Europe and North America. The recent focus on biodiversity decline and conservation efforts demonstrates significant gaps in our faunistic knowledge. While species discovery for vertebrates is well advanced, the discovery of the majority of non-vertebrate species on the planet (Agenda 2000) lags far behind (Knapp et al. 2005). A 139% increase of recorded bivalve species from the Florida Keys between 1995 and 2004 illustrates the point (Bieler & Mikkelsen 2004).

Conservation efforts focus mainly on the estimated species-richness of habitats and the occurrence of endangered and threatened species (Mace 2004). Charismatic vertebrate species and their protection are most often invoked in conservation actions, whereas the majority of animal biodiversity, the invertebrates, remains on the sideline. We often know so little about species' ranges and abundance that species-richness estimates for any particular habitat or region are vague and uncertain.

Spiders, a mega-diverse group with 38,000 described species (Platnick 2005) are a case in point. As insect predators, they play a pivotal role in the regulation of insect populations in all terrestrial habitats. Species lists offer solid baseline data for large-scale biogeographic analyses, survey and monitoring efforts, and tracking of environmental changes. These lists form the foundation of species-richness estimates. Yet, reliable, up-to-date spider species lists for the 50 United States are not available, because faunistic research declined (and lost funding) before the job was done (Crawford 1988). Consequently, the current assemblage of spider species lists suffers from several problems, impeding biogeographic research as well as hampering their utility for ecological research and conservation efforts.

Spider species lists for various states and habitats were completed over a long period of time. The species records for the five states covered here are gleaned from the literature dating back 50 years for Wisconsin (Levi & Field 1954), while Indiana and Illinois were covered recently (Beatty 2002). Species names changed during this period. While such name changes are often somewhat naively decried as hampering the accessibility of the records (Golding & Timberlake 2003), it is imperative to assert that such name changes are the result of significant and badly needed progress in the systematics of the animals involved (see for example Froese & Pauly 2005). Since invertebrates in general, and spiders in particular are species-rich and require highly trained taxon-experts for identification, more name changes can be expected in the decades to come. Authors of previous species lists used various cross-reference methods to make lists compatible by citing older names in their lists, but these in turn became quickly out-dated. Therefore, employing such lists for various research and conservation tasks requires significant taxon-expertise and general users, especially within the ecological research and conservation communities, still face obstacles interpreting and using these records (Gotelli 2004).

Traditional forms of printed publications of individual species lists hamper regional and inter-state comparisons of species records. In connection with the significant time-gap between individual faunal lists and the inevitable nomenclatorial changes, such species lists do not support biogeographic research (Soberon & Peterson 2004), nor do they offer easy access to species-richness estimates. There are currently numerous activities with respect to GIS software development (e.g., Lifemapper, http://Lifemapper.org) to generate distribution maps. The goal is a '... predictive, electronic atlas of Earth's biological diversity,' the data for which must be retrieved from '... records of millions of plants and animals in the world's natural history museums.' Certainly, locality data taken directly from actual collection specimens would provide the best possible foundation of species-richness estimates and biogeographic research (Graham et al. 2004). Furthermore, vouchered museum specimens can be re-examined at any time for additional data, such as sex, size, abundance, and intraspecific variability; misidentifications will be rooted out over time. What stands between us and such a rich data source is the simple fact that the data of most invertebrate, especially arthropod collections, have not been electronically captured, and, at the current meager state of collection support, we cannot expect access to these data any time soon.

What measures, then, can be taken now to accelerate the rate of biodiversity discovery, either through sampling in nature or mining of museum collections? The existing faunistic and taxonomic literature harbors a wealth of biogeographic data; making these legacy data universally available, i.e., electronically and nomenclatorially updated, will serve a wide range of users, from researchers to land managers to collection managers. Various taxonomic authority files can be generated from such cross-referenced legacy data, e.g., generating species lists for states, regions, or habitats, aiding in museum collection management, and guiding future survey and monitoring efforts. The present study clearly demonstrates the great utility and predictive power of old-fashioned faunistics required for today's biodiversity research.

Our approach.--Two of the prevailing problems inherent to traditional faunal lists are the nomenclatorial changes due to time gaps and the isolation of lists from one another, hampering comparison and predictions. Using modern database technologies and electronic dissemination over the Web (Scoble 2004), these problems can be easily overcome today. However, even with these technologies, keeping checklists and nomenclature up-to-date requires expertise by sufficiently supported taxon-specialists. Spider systematics benefited from a relatively large number of active spider systematists, when compared to other rather neglected groups, such as many other terrestrial arthropods (e.g., the Myriapoda, see Milli-PEET, Field Museum Website). Furthermore, spider systematics is boosted by a rare piece of scientific infrastructure, an online world-wide taxonomic catalog (Platnick 2005). Such a catalog, generated and maintained by a taxon expert, forms the standardized base for all spider species names used here. The need for such standardization is readily acknowledged by the conservation and ecology user community (e.g., NatureServe's Central Databases, '... a "standard" name is selected and maintained ...' for every '... taxon tracked in its database').

To track name changes we employ the Spider Species Name Concordance Table (Appendix II). The table contains all species names and genus-species combinations present in the referenced species lists (see Literature Cited) with their currently valid name according to Platnick (2005). Appendix II is available online at http://www.uwgb.udu/ biodiversity/glspiders.

Comparisons between regions, habitats and states demonstrate differences in recorded species, allow predictions, and guide future species discovery. Explanations of recorded species differences can be sought; species-richness estimates can be deduced. Electronic manipulation of such species data allows sorting the data in various ways to answer questions about distribution and species-richness, and facilitates the discovery of faunal shifts. Recorded occurrences of species allow for prediction of future discovery of species in nearby regions (gap analysis). The species lists as presented in Table 8 and Appendices I and II are available online at http://www. uwgb.udu/biodiversity/glspiders. The data are also presented in database format at the same site. Maintaining such lists online will allow periodic updates. We hope that others will join the endeavor and add spider species lists for more U.S. states.

METHODS

The project brought together species records from a variety of sources (see Literature Cited) for five states surrounding the Great Lakes: A northern tier of two states, Wisconsin and Michigan, and a southern tier of three states, Illinois, Indiana, and Ohio (Fig. 1). Essentially the whole area of these states is part of the Great Lakes bioregion of eastern North America, which also includes portions of Minnesota, Pennsylvania, New York, and the Canadian province of Ontario. For four of the studied states, species lists exist in some form. Additional records were gleaned from regional and habitat surveys, from the alpha-taxonomic literature, and in some cases directly from identified specimens in collections.

[FIGURE 1 OMITTED]

Included species records.--Inclusion of a species record followed the criterion: Is the species likely to be "established" i.e., have reproducing populations in the area? Thus, exotic species are listed, as are species that live only in buildings, as long as it is reasonable to assume that they actually reproduce in the Great Lakes region. Casual importations, such as exotic tarantula species in the houses of pet owners are excluded from the list. Certain exotic species, such as the huntsman spider Heteropoda venatoria, are occasionally imported with lumber and other merchandise and are encountered by the public as repeated inquiries to the authors indicate. Such species records are discussed in the text, but are not included in the spider species list of the Great Lakes States (Table 8 and Appendix I). Species records of a non-established species that appear in the published literature are listed in the Spider Species Name Concordance Table (Appendix II). If intercepted guest spiders are known only from informal records and public inquiries, such species are discussed in the text. Uncertainties remain, such as the brown recluse spider (Loxosceles reclusa), which is native to southern Illinois, but reproducing populations have occasionally been encountered in some buildings in the region north of the species' natural range. Such species are included in the Great Lakes States spider species list. We further scrutinized the published spider species lists for questionable species records, mainly Palearctic species recorded for the Midwestern U.S. states. In each case a decision was based on the best of our knowledge whether the species was likely to have established populations in the Great Lakes area. Published records of clearly Palearctic species or species suspected to be misidentifications are not included in Table 8 and Appendix I, but can be found in the Spider Species Name Concordance Table (Appendix II) and are discussed in the text. It should be noted here that current geographic ranges of species, as they are noted in the Worldwide Spider Catalog (Platnick 2005), do not necessarily reflect the biogeographic origin of a species. Thus, a species listed as Holarctic may very well be introduced to a part of the Holarctic range. Several such cases are discussed below, and their biogeographic labels in Table 8 reflect our evaluation of their biogeographic status.

Compilation of species lists.--Spider species records are arranged alphabetically by family, genus, and species in Table 8 and Appendix I. Each species is found under its current genus-species combination (as listed in Platnick's on-line catalog 2005), which is not necessarily the combination used in the sources cited in Appendix I. Genera are listed in the family they are currently assigned to and disregard past family placements. Major family placement changes occurred with genera now assigned to the Agelenidae, Amaurobiidae, Liocranidae, Corinnidae, and Dictynidae. These, as well as a listing of old family names, are captured in the auxiliary Tables 6 and 7. Although we recognize that different opinions exist about genus placement or validity of species and synonyms, use of Platnick's nomenclature allows a stable foundation for all subsequent work and analyses conducted with the data presented. Table 8 lists presence-absence within states, and Appendix I lists sources and vouchers. For the literature source of the species records, a code is given, which can be found in the reference section. Name changes are dealt with in the Spider Species Name Concordance Table (Appendix II).

Illinois: The Illinois species records are taken from the faunistic and biogeographic literature (Kaston 1955; Beatty & Nelson 1976; Moulder 1966, and Beatty 2002) and a field guide (Moulder 1992). These are augmented by several sources. Spider records from the on-line database of Illinois species of the Illinois Natural History Survey (edited by S. Hill) were incorporated [URL: remains unavailable]. Also, spider material from pitfall traps and Berlese extractions in Cook and Lake Counties yielded additional records, primarily of ground spider species; these voucher specimens (Appendix I) are deposited at the Field Museum. The Cook County specimen collections were done biweekly from May to October in 1996 through 1999. Spider specimens were sorted and identified by co-author E. Lehman with additional determinations by P. Sierwald and M. Draney; the Cook county records have been databased (Field Museum of Natural History). The Lake County survey was conducted from 20 June-30 August, and 20, 24, 29 September 1999. Spider specimens were sorted and identified by co-author T. Prentice. The specimens are deposited at the Field Museum. We integrated the data of all vouchers from the five states housed in the collections of the Illinois Natural History Survey (University of Illinois, Champaign-Urbana) into our database (courtesy of C. Farvet); the majority of the INHS collection's spider specimens are from Illinois. The Illinois State Museum in Springfield (Illinois) maintains an on-line spider species database, which provides data on vouchered county records. Currently (2005) the Illinois State Museum database contains 224 spider species for Illinois. All vouchered species for which county records are cited are included in the Illinois species list.

Indiana and Ohio: The species records for Indiana are based on the most recent checklist by Beatty (2002, which incorporates the list by Parker 1969). The Ohio species records are adopted from an on-line state species list generated by Dr. Richard Bradley (Ohio State University, permission for inclusion granted) based on published records and voucher specimens of spider species from Ohio. No additional sources were drawn on for these two states.

Michigan: The Michigan data, initially compiled by Joan Jass (Curator of Invertebrates, Milwaukee Public Museum), are based on faunal studies of particular areas (Drew 1967; Chickering 1932, 1933, 1934), faunal studies restricted to particular habitats (Brady et al. 1991), and taxon-specific faunal lists by Chickering (1939, 1940, 1944, 1959) and Wolff (1984). Again, the alpha-taxonomic literature provided additional species records (Platnick & Dondale 1992; Proszynski 1968). Various statewide spider species lists were produced by Chickering & Bacorn (1933), Chickering (1935) and most recently by Snider (1991).

Wisconsin: The Wisconsin data compiled initially by Joan Jass (Milwaukee Public Museum) are based on state faunal lists (Levi & Field 1954; Levi et al. 1958), regional surveys (Blaczyk et al. 1992), habitat surveys (LeSar & Unzicker 1978; Riechert & Reeder 1978) and the taxonomic literature of Dondale & Redner (1976, 1978, 1982, 1990) and Levi (1980). Species records and vouchers (Appendix I) were added from identified and computerized collections at the Milwaukee Public Museum.

Spider species name concordance table (Appendix II).--Numerous name changes have occurred in North American spiders during the almost 70 years since publication of the earliest of the literature sources drawn upon. In some cases, a single species may have been renamed several times as a result of synonymies and taxonomic revisions. Some species were originally misidentified until later taxonomic work clarified species identities. Some species were transferred to other genera; other names were synonymized. Such name changes are captured in Appendix II. The table lists genus and species names as they occurred in each of the literature sources listed (see Appendix I) for that species. Historical intermediary names not used in the faunal source lists are not included. Thus "Epeira cornuta Clerck 1757" and "Araneus cornutus (Clerck 1757)" would both indicate "transferred to Larinioides," the current araneid genus. Example, species transferred to a different genus: The Illinois Natural History Survey database (INHSD) listed the amaurobiid species: Coelotes juvenilis Keyserling. We added year of first description according to Platnick to ensure an unambiguous identification. The species is currently assigned to the genus Coras. Appendix I lists "Coras juvenilis (Keyserling 1881)" giving INHSD as a source, and indicates that it has also been recorded from Indiana, Michigan and Ohio, but not Wisconsin. Appendix II lists Coelotes juvenilis, the literature source (INHSD) and the current name Coras juvenilis in the last column.

Table 8 and Appendix I list the currently valid genus-species name combination, crediting the source, even if the source listed this species under what is now considered a junior synonym. Example, synonyms in faunal source lists: In 1933, Chickering and Bacorn added Anyphaena rubra to the known Michigan Anyphaenidae. Anyphaena rubra was determined to be a junior synonym of Hibana gracilis. Table 8 lists Hibana gracilis as a member of the Michigan fauna. Appendix I credits 'c&b' as a source for the Hibana gracilis (Hentz 1847) record in Michigan, although they originally listed the species as Anyphaena rubra. Appendix II notes this species synonymy. Minor typographical errors and discrepancies, such as the year of publication, and grammatically incorrect endings of species names (e.g., -us versus -a) are not included in Appendix II.

Statistical analysis.--Since this compilation represents "meta data" compiled from numerous sources and representing records originally amassed using various methodologies and for various purposes, no rigorous statistical analyses comparing richness or diversity among states were attempted. However, we did explore correlations of number of taxa that are now known from each state, with various basic properties of those states, including land area, population, and date that the state entered the union (as an index of how long the state has been intensively occupied by non-indigenous settlers). Properties of the states were obtained from the U.S. Census Bureau (2004). Relationships were explored using Pearson Correlation Coefficients (Systat 10, Systat, Inc., Evanston, IL). We also ran Chi-square tests to compare the number of species within each family with the regionwide average species-richness of each family, using Microsoft Excel. This was done to determine whether the composition of each state's species assemblage differed from the entire region's assemblage.

RESULTS

Species-richness analysis.--To date, 900 spider species have been recorded from the five-state region: 646 from Illinois, 383 from Indiana, 571 from Ohio, 479 from Wisconsin and 563 from Michigan. The 900 recorded spider species represent 284 genera and 40 families (Table 2). The most speciose family in the region is Linyphiidae, with 24% of species. This is over twice the species diversity of the next most diverse families, Salticidae (10.3%), Theridiidae (8.9%), Lycosidae (8.8%), and Araneidae (7.7%). Gnaphosidae and Thomisidae also make up 5% or more of the region's species; 10 other families make up 1% or more of the region's species, and the remaining 25 families each are represented by less than 1% of the total species (Table 2). The top 8 families contain over 75% of the region's species, and the top 13 families make up over 90%. The region's most speciose genera are listed in Table 1, and include Xysticus (25 species) followed by Clubiona, Philodromus, and Araneus, each represented by over 20 species. The ten most speciose genera mainly consist of globally speciose genera, including the type genera of four families (Clubiona, Philodromus, Araneus, and Theridion). It is not surprising that such large genera are represented by a number of species in our region. Most of these genera are represented in our region by only 3-10% of their total known species complement. One prominent exception is Ceraticelus; 47% of the world's species have been recorded from our region, making it a "hotspot" for the genus. Similar arguments could perhaps be made for Pirata and Drassyllus (our region includes 19% and 17% of the world's known species, respectively).

The number of species in each state varies from Indiana's 383 (42.5% of the region's total) to Illinois' 646 (71.7% of the region's total; Table 2). This shows that latitude and longitude do not predict the size of the current state lists, since the states with the highest and lowest numbers are both in the region's southern tier, and in fact are adjacent. Pearson correlation (Using Systat 10; Tables 3 and 4) shows that the number of species in each state is only weakly correlated with land area (0.609), but strongly correlated with the state's 2000 U.S. Census population (0.897). This supports our suggestion that the region's fauna is very incompletely known, and hence the size of the list is determined more by sampling effort (which has generally been higher in more populated states) than with biological or geographic differences between the states.

The number of families recorded from each state varied from 29 in Michigan and Wisconsin (72.5% of the region's total) to 39 in Illinois and Ohio (97.5% of the region's total). There is only a weak correlation between species and families recorded in each state (Pearson coefficient = 0.506), and a negative correlation between family richness and land area (-0.311). The correlation between state population size and family richness is weaker than with species (0.715). The strongest predictor of family richness in the region is clearly latitude; all three southern tier states (Illinois, Indiana, and Ohio) have recorded 4-10 more families than the two northern tier states (Michigan and Wisconsin). This suggests that several families reach their northernmost limits within our region, and that family level richness does indeed decrease with latitude across the region.

The proportional representation of species within families (i.e., species-rich versus species-poor families) within each state roughly reflects the region-wide pattern: A few diverse families contain most of the species. Linyphiidae and Salticidae are the most and second-most diverse families in all five states, and the region's top 6 most diverse families are the top six in each state, although the order varies somewhat (Table 2). However, the proportional representation of species within families is statistically different from the region average in some states. Chi-square tests compared the number of species within each family with the average species-richness of each family for the region by scaling the species-within-family richness to match the region-wide average richness. For these tests, families with less than 1% of the region's species (families ranked 16-40) were pooled into a single "rare families" category (yielding df = 15). The faunal structure of Ohio, Michigan, and Wisconsin did not statistically differ from this five-state average, but Illinois' fauna was different at P = 0.05 and Indiana's fauna was different at P = 0.001. These differences may be due to historical particulars rather than reflecting biological differences. For example, Illinois' list contains proportionally more corinnid species and fewer species of the "rare families" than the region's average, whereas Indiana's list contains far fewer linyphiid and gnaphosid species and more species of the "rare families" than the region's average (Table 2). We think it is easier to explain these differences as attributable to different collection methods or different research objectives of past studies than to put forward a reasonable biological explanation for these patterns.

A total of 237 of the 900 species (26.3%) are recorded from all five of the states. Similarly, 244 species (27.1%) are known from only one of the five states. Smaller numbers of species are known from 2 to 4 states: 171 species (19%) from two states, 125 species (13.8%) from three states, and 123 species (13.6%) from four states (Table 8). This pattern is not unexpected, since many species are widespread and common, and often appear in all states surveyed, and other rarely collected species are likely to occur only once until much sampling is done.

Also not surprisingly, states with more recorded species tend to have more regionally-unique species: The most species-rich state, Illinois, had 80 regionally unique species, followed by Michigan (66), Ohio (55), Wisconsin (28), and Indiana (15). Note this is the same order as species-richness, except that Michigan and Ohio switched order.

Gap analysis.--One of the important potential uses of sets of internally consistent species lists such as we present here is to predict the occurrence of species in areas from which they have not yet been recorded. Due to the region's history of repeated glaciation (which has resulted in a relatively newly colonized fauna in a topographically subtle region), spider species tend to have large ranges and few are narrowly endemic or endemic to small areas within this region. This empirical observation lends credence to our assumption that the geographic position of the states will tend to contain information about species distribution. For example, if a species occurs in states both east and west of a given state, it probably also occurs within that state. Similarly, if a species occurs in states with similar latitudes and other states with similar longitudes as the state in question, the species is also likely to occur in the state in question. Although it is certainly possible for a species to occur in any four of the states but not in a fifth, it is also true that such occurrences would usually necessitate very specific (and seemingly improbable) geographic distribution patterns. Thus, for many state occurrence combinations, we feel that a species is much more likely to exist within some unrecorded states than not to exist there. Using this sort of logic applied to our two-tiered, five state region (Fig. 1), we examined Table 8 and predicted occurrences of species within states from which they have not yet been recorded (Table 8); 'P' in a cell indicates a species is predicted to occur in a state, using the following criteria:

1) If a species has been recorded from 4 of the 5 states, we predict it also occurs in the fifth state. There were 123 such predicted occurrences.

2) Species found in 3 states should also occur in the others, except that (a) species found only in the southern tier states (Illinois, Indiana, and Ohio) are not predicted to occur in the northern tier states (Michigan and Wisconsin); (b) species found only in the western states (Illinois, Indiana, and Wisconsin) are not predicted to occur in the eastern states (Michigan and Ohio); and (c) species found only in the eastern states (Indiana, Michigan, and Ohio) are not predicted to occur in the western states (Illinois and Wisconsin). There were 182 such predicted occurrences.

3) We considered that species found in two states could predict occurrences in other states in certain circumstances (refer to Fig. 1). (a) Species from Illinois and Ohio should occur in Indiana. (b) Species from Illinois and Michigan should occur in Indiana. (c) Species from Indiana and Wisconsin should occur in Illinois. (d) Species from Ohio and Wisconsin should occur in the other three states.

Combinations we felt were not likely to be predictive of other state occurrences were: Illinois and Indiana; Illinois and Wisconsin; Indiana and Ohio; Indiana and Michigan; Ohio and Michigan; and Michigan and Wisconsin. There were 86 predicted occurrences using pairs of state occurrences.

Altogether, using 2642 recorded species occurrences, we predict 385 occurrences of species in states in which they are not yet recorded. If all these predictions are correct, this would increase the region's average state richness from 528 to 605 species, and raise individual state species totals by about 5% to about 53% (mean 17%, Table 5), even though the region's total species-richness would not increase with these additions. Note that the number of predicted occurrences in any one state depends both on having supposed gaps in species (inadequate sampling of a state) as well as the state's geographic position (peripheral areas benefit less than more central areas under our prediction criteria). Because of both of these factors, Indiana's fauna is predicted to increase the most using this gap analysis method; it is even predicted to have more species than Wisconsin, although this method does not predict all possible unrecorded species. Many areas (especially on the periphery of a region) may have unrecorded species that have not been recorded elsewhere in the region. Wisconsin, for example, may harbor unrecorded species that reach their southernmost limit within the state, and neither have been previously, nor in the future will be recorded in the southern tier of states. Note that we are not suggesting that predicted species do definitely occur in predicted states, only that it is more likely that they do occur there than that they do not. Until confirmed with specimens, our predictions should be considered to be hypotheses of species occurrence. Also note that failure to obtain a predicted occurrence does not mean that we predict a species will not occur in a state, but only that insufficient information exists in our matrix for us to venture a prediction.

Biogeographic analysis.--The known established spider fauna of the Great Lakes States is composed mainly of native Nearctic species (some of which may possibly actually be neotropical). In addition, 85 Holarctic species (9.4% of total, Table 8, denoted by 'H' in column D), common to both Old and New World temperate regions, count as native species. When referring to non-native taxa, we are generally discussing species that are not native to North America. New state or regional records, other than for most cosmopolitan and Palearctic taxa, are likely to be the result of new collecting efforts, rather than indicating range extensions for the respective species that are native to America and, occasionally, to southern Canada and northern Mexico.

However, in cases where specimens of native taxa are found great distances from their known ranges, accidental importation is often suspected. Whether or not a species has become established must be ascertained in order to define which species are actually representative of the spider fauna of a state or region. If a non-native species generates a reproducing population then it has become an integral part of the local community, regardless of the consequences of its presence. On the other hand, particular spiders are occasionally imported with shipments of agricultural and/or ornamental plants but are unable either to reproduce or maintain viable populations in a given region; these spiders should not be included on state or regional lists.

Four species, recorded only from Ohio, were undoubtedly imported with shipments to the state and are excluded from our five-state list (see Appendix II): Heteropoda venatoria Linnaeus, Cupiennius coccineus F.O.P.-Cambridge, Ctenus bilobatus F.O.P.-Cambridge, and Latrodectus hesperus Chamberlin & Ivie. Heteropoda venatoria and C. coccineus are tropical species, the former considered pantropical, the latter naturally occurring in Costa Rica and Panama. Both species are occasionally imported to the U.S. with produce or other floral shipments but have never become established within the country (although reports suggest that H. venatoria has become established in Florida (e.g., Gertsch 1949, Edwards & Marshall 2001). Ctenus bilobatus is known from the female type from Mexico (F.O.P.-Cambridge 1900). It is most likely tropical as most of the Ctenus species are. If indeed the species was correctly identified, specimens were probably imported with produce shipments from Mexico. It is highly unlikely that the species would be able to establish in the north-central part of the USA. Latrodectus hesperus (western black widow) occurs west of the approximate mid-line through Texas, Oklahoma, and Kansas north to the Canadian provinces. This spider is also occasionally shipped with produce and other materials but there is no evidence, to date, that the species has become established in Ohio.

Non-native species: Non-native species consist of 17 cosmopolitan species (1.8% of total), seven Palearctic species (0.77% of total) inhabiting the Great Lakes region (Table 8, column D, indicated by 'C' and 'PA' respectively) and one species possibly introduced from the Pacific region and Australia (P/Au). Tegenaria domestica (Clerck) is an introduced cosmopolitan species in the family Agelenidae with a widespread distribution in both the Old and New Worlds. On the North American continent, it occurs at least as far north as Ellesmere Island in northern Canada (Roth 1968). The species is well established in all five of the Great Lakes states. Collection records indicate that the species is generally found in human habitations.

Three Palearctic (non-native) araneids, Aculepeira carbonaria, Araneus triguttatus, and Zygiella montana were recorded from Wisconsin; Z. montana is also reported from Michigan. The former two species are known from high elevations in the European mountains. Chamberlin & Ivie (1942) removed Aculepeira verae from the synonymy of A. carbonaria and both Levi (1951) and Levi & Field (1954) suggested that the specimen(s) collected from Wisconsin may have been A. verae (= A. packardi). Levi (1973) stated that there was no evidence that A. triguttatus occurs in the USA because the illustrated specimens of Epeira mayo (= A. triguttatus) were undoubtedly mislabeled. Since the original publication by Levi & Field (1954) linking these species to the Wisconsin fauna, no additional specimens have apparently been collected within the state and no voucher specimens have been located; we have removed the two species from our Great Lakes list (Appendix II). Zygiella montana is known from the European mountains at elevations above 1000 m. Most occurrences of the species in the USA are from mountainous areas or from northern latitudes. Voucher specimens of Zygiella montana are found at Michigan State University Entomological collection and in the Milwaukee Public Museum in Wisconsin. The species is also reported from Maine, North Carolina, and the Adirondacks and White Mountains. There seems to be little doubt that Z. montana is established in both Michigan and Wisconsin (Table 8, Appendix I). Two additional Palearctic araneid species, Araniella cucurbitina Clerck and Araneus angulatus Clerck, are reported only from Michigan; there are no known voucher specimens of either species. Snider (1991) notes that the former species is probably A. displicata (Hentz); and Levi (1971), in his revision of the diadematus Group, stated that no specimens of angulatus were in collections coming from North America and that the literature records of angulatus referred

to large specimens of various other native species (Appendix II). In light of the above, we have considered these records doubtful and have excluded the species from our list of the Great Lakes spider fauna. Larinioides sclopetarius (Clerck) is listed as Holarctic in Platnick (2005), but Levi (1974) judges it to be introduced into North America by its anthropochorous habitats here. It should be noted that since Levi's (1974) publication, the species seems to have become much more abundant in the Great Lakes region (Table 8 and M. Draney, unpubl, observ.). It is widely distributed in Eurasia (Levi 1974).

The Banded Argiope, Argiope trifasciata, is considered a cosmopolitan species although it is found in many non-disturbed areas throughout the USA. According to Levi (1968), the distribution of the species is not known but is nearly worldwide exclusive of regions occupied by the Eurasian species, A. bruennichi. Populations of A. trifasciata thrive in the five-state Great Lakes region. Gea heptagon may be an introduced species, possibly from the Pacific region or Australia (Levi, 2004).

The cosmopolitan wood-louse (Order Isopoda) specialist, Dysdera crocata C. L. Koch is widely distributed within the U.S. The species thrives in the more mesic habitats where there is an ample supply of isopods (most of which are also exotic according to Jass & Klausmeier 2000). To date, it is the only Dysdera species known from the U.S. Well-established populations occur in southern Wisconsin, Illinois, Indiana, and Ohio, but the species has apparently not yet been collected in Michigan. Both the Field Museum in Illinois and the Milwaukee Public Museum in Wisconsin house voucher specimens.

Three non-native gnaphosid species, Trachyzelotes lyonneti (Audouin), Urozelotes rusticus (L. Koch), and Zelotes subterraneus C. L. Koch have been reported from one or more of the five Great Lakes states. In their revision of Trachyzelotes and Urozelotes, Platnick & Murphy (1984) examined only one female specimen of T. lyonneti from the five-state region; this specimen was taken from under a rock in a backyard in Alton, Madison County, Illinois. The species is not a synanthropic/cosmopolitan species and appears to be more tolerant of drier conditions than U. rusticus. Until additional specimens are discovered, the species' establishment in Illinois is doubtful; we have removed the species on our updated species list (Table 8, Appendix I, see Appendix II). Urozelotes rusticus is considered a cosmopolitan species and reported from Wisconsin, Illinois, and Ohio. Platnick and Murphy (1984) examined material from both Wisconsin and Illinois but since that time, the species appears to have become established in Ohio as well. The Palearctic species Zelotes subterraneus is reported from all five states. However, Platnick & Shadab (1983) removed Z fratris Chamberlin from the synonymy of Z. subterraneus and stated that the species was often misidentified as Z. subterraneus because of their similarity. The native species is very widely distributed in the USA except in the southeast and south-central parts of the country; it is known from the northern parts of Illinois, Indiana, and Ohio and is widely distributed in Wisconsin and Michigan. It is very unlikely that Z. subterraneus occurs in North America; and that, instead, its multiple listings were the result of misidentifications. It is not included on our updated list.

Eight Palearctic or European linyphiid species were recorded from the five-state region. Five of these species were reported only from Michigan: Agyneta cauta (O.P.-Cambridge), Bathyphantes nigrinus (Westring), Centromerus serratus (O.P.-Cambridge), Pityohyphantes phrygianus (C.L. Koch), and Walckenaria acuminata Blackwall. Voucher specimens of four of these species are in the Michigan State University Entomology Collection: B. nigrinus, C. serratus, P. phrygianus, W. acuminata. Since there are no voucher specimens of A. cauta, we regard the record as a misidentification. Records and voucher specimen(s) of C. serratus and W. acuminata are from only one locality each and voucher specimen(s) of B. nigrinus from only one locality; the Walckenaeria acuminata specimen was actually from Europe and inadvertently listed by Snider (1991) (Snider pers. comm.). If specimens attributed to B. nigrinus and C. serratus were even identified correctly, it is highly unlikely that they have become established; no other specimens have surfaced since Snider's publication. Consequently, we removed these latter three species and A. cauta from the Great Lakes species list. Snider (1991) lists several localities for Pityohyphantes phrygianus but does not list P. costatus (Hentz), which has been consistently misidentified as P. phrygianus. We treat the records of P. phrygianus as records of P. costatus in Table 8 and Appendix I.

The Palearctic species Gonatium rubens Blackwall has been recorded from Wisconsin, Illinois, and Michigan. Millidge (1981), however, stated that all North American specimens labeled 'G. rubens' in the American Museum of Natural History were, without exception, specimens of G. crassipalpum Bryant. This endemic species is widely distributed on the North American continent except in the extreme southern parts. In our region of concern, the latter species has been recorded from only Illinois and Ohio. It seems a reasonable assumption that the G. rubens specimens from the three states were probably misidentified and that the species does not occur in North America. We removed G. rubens from Table 8 and Appendix I and attribute its records to G. crassipalpum. A similar case of misidentification applies to Great Lakes records of Pocadicnemis pumila (Blackwall). The species was formerly widely reported in eastern North America, but Millidge (1975) demonstrated that two closely related species, P. americana Millidge and P. occidentalis Millidge were confused with P. pumila. Millidge only found one bona fide P. pumila record from northeastern North America. Wisconsin and Michigan records of P. pumila before 1975 are suspect, and here we consider them to be records of P. americana. Unless new specimens are collected, we do not consider that it occurs in our region.

Eperigone fradeorum (Berland) is considered a cosmopolitan species; its origin is unknown. In his revision of Eperigone, Millidge (1987) suggested that the species "may" be endemic to the eastern seaboard (particularly Florida). He further stated that E. fradeorum has undergone widespread dispersal, as least partially through the agency of human travel. It appears that the species has become established in the southern portion of Illinois but has not been recorded yet in Indiana or Ohio. Maso gallicus Simon (Europe to Azerbaijan) and Stemonyphantes lineatus (Linnaeus) (Palearctic) are recorded from Wisconsin, the latter also from Michigan. There are no voucher specimens of the former species and no additional reports of the species since its listing by Levi et al. (1958). It was long believed that the American species, S. blauveltae Gertsch, was the same as the Palearctic species S. lineatus; and specimens of the former species have repeatedly been misidentified as S. lineatus. The Wisconsin and Michigan records of S. lineatus are regarded as misidentifications. Maso gallicus and S. lineatus are, consequently, removed from our updated species list (see Appendix II).

The Palearctic lycosid species, Arctosa cinerea (Fabricius), was recorded in 10 regions, all of which are in the state of Michigan. However, prior to 1976 (Roth & Brown 1976) when A. littoralis (Hentz) was removed from the synonymy of A. cinerea, specimens of the former species were generally misidentified as A. cinerea. On the Michigan list (Snider 1991) localities for both species are identical with the exception of one additional Chickering record (year not provided) of A. cinerea from Cheboygan County. Furthermore, voucher specimens are indicated for all A. littoralis localities but are lacking for the identical A. cinerea localities. In light of the above, we consider the records of the Palearctic species A. cinerea to be instead, duplicate records of A. littoralis.

Ero furcata (Villers), the Palearctic mimetid species, was recorded from all five Great Lakes states. However, Kaston (1977) realized that our species was different from the Palearctic species and removed E. leonina (Hentz) from the synonymy of E. furcata with which it had been confused for so long. The native E. leonina is only recorded in our region from Wisconsin and Michigan. Snider (1991) commented that the Michigan record of the Palearctic species was a misidentification and referred to Kaston (1981). Therefore, we consider the records of the Palearctic species E. furcata as misidentifications and note that the native species occurs in all five states.

The miturgid Cheiracanthium mildei is listed in Platnick (2005) as Holarctic, but it is a Mediterranean native. It has spread throughout most of the eastern United States since it was first found there in 1949 (Bryant 1951), and is recorded from Illinois, Ohio, and Wisconsin in our region.

The nesticid, Eidmannella pallida (Emerton) is a cosmopolitan species known from Wisconsin, Illinois, Indiana (Gertsch 1984), and Ohio. Reports from the four states indicate that the species is well established. The species is known also from other northern localities in New Jersey, Massachusetts, Oregon, and Ontario, Canada. Two cosmopolitan species of the family Oecobiidae, Oecobius cellariorum (Duges) and O. navus Blackwall are recorded from Ohio; the former species is also known from Illinois. Both species are believed to be established but appear to be restricted to indoor habitats. In the more southern regions, both species can be found both in and outside of buildings.

Records for the Palearctic philodromid species Philodromus aureolus (Clerck) include Illinois, Wisconsin, and Michigan. However, Dondale (1961) elevated P. cespiticolis Walckenaer from a subspecies of P. aureolus. In his publication, P. aureolus, referred to by Levi & Field (1954) and by Chickering (1940), was synonymized with P. cespiticolis, which was later placed in the synonymy of P. cespitum (Walckenaer) (Dondale & Redner 1976). LaSar & Unzicker (1978) listed the Palearctic species as also occurring in Illinois but there are no voucher specimens. Therefore, we regard the records of P. aureolus from the three states as misidentifications.

Two pholcid species, Pholcus phalangioides (Fuesslin) and Spermophora senoculata (Duges), are recorded from the Great Lakes region; both species have apparently been introduced into the Great Lakes states. Considered a cosmopolitan species, P. phalangioides occurs in all five states and, throughout its range in this region, has become well established, primarily in houses and other buildings. Populations also occur in several additional northern states. Spermophora senoculata (Duges) has been found in all of the Great Lakes states except Michigan. This species is listed by Platnick (2005) as Holarctic. However, it appears that within the Great Lakes region, the species has a strictly anthropochorous distribution, residing in or on houses and other buildings, and is assumed to be introduced (Huber 2000).

The Palearctic jumping spider Evarcha falcata (Clerck) was listed as occurring only in Michigan (Snider 1991). Because of the close similarity of E. hoyi and E. falcata and the fact that there are no voucher specimens for the Ohio record of E. falcata, we consider the listing of this Palearctic species as a misidentification. Hasarius adansoni (Audouin) is an introduced synanthropic species with known populations in several northern states including Wisconsin, Illinois, Indiana, New York, and Massachusetts (Cutler 1990). Myrmarachne formicaria (DeGeer) is a Palearctic species that was recorded in Ohio alone among the Great Lakes states. To our knowledge, this has been the only record of the species from the USA; and there are apparently no voucher specimens verifying the Ohio record. Therefore, we consider M. formicaria a doubtful record, but maintain this species for the time being as a member of the Great Lakes fauna. The Palearctic salticid species, Phlegra fasciata (Hahn) was recorded from Wisconsin and Michigan. However, Chickering's (1944) specimens of P. hentzi from Michigan were misidentified as P. fasciata, as were Levi & Field's (1954) specimens from Wisconsin (see Platnick 2005). The latter authors referred to P. fasciata as "leopard spider", Attus leopardus Hentz, a synonym of P. hentzi. Phlegra hentzi (Marx) was removed from the synonymy of P. fasciata by Logunov & Koponen (2002). We list Phlegra hentzi as a Midwest spider species and removed Phlegra fasciata from the Midwest spider list. Sitticus fasciger (Simon), possibly of Asian origin, is established in Quebec, Canada and in several states in the USA, including Wisconsin and Illinois. Platnick (2005) lists the species as occurring in Russia, China, Korea, Japan, and USA. The Zebra Jumping Spider, Salticus scenicus (Clerck) is listed as Holarctic in Platnick (2005), but is also apparently introduced to North America (Gertsch 1949). It is now recorded from all five states in our region.

The sicariid species, Loxosceles rufescens (DuFour) is considered a cosmopolitan species, and its occurrence in the USA is strictly confined to buildings (R. Vetter, pers. communication). Within the Great Lakes region, the species appears to be established in the Argus Building on the University of Michigan campus, in at least one building in the Cincinnati region of Ohio, and was collected in 2002 by the senior author in basements of buildings in downtown Chicago.

Meta menardi (Latreille), a Eurasian tetragnathid species, is reported from all Great Lakes states except Michigan. However, Marusik & Koponen (1992) recognized that within the Holarctic distribution of M. menardi there were three allopatric species involved, one of which occurred in North America, M. americana Marusik & Koponen, which Dondale (1995) synonymized with M. ovalis (Gertsch 1933). Consequently, the M. menardi records in Wisconsin, Illinois, Indiana, and Ohio were actually records of M. ovalis and are treated in our tables as such.

Seven introduced species in the family Theridiidae are well established in two or more of the five Great Lakes states. Three of the seven species are in the genus Steatoda: S. albomaculata (De Geer), S. grossa (C. L. Koch), and S. triangulosa (Walckenaer). Steatoda albomaculata and S. triangulosa occur in all five states, S. grossa is found only in Indiana and Ohio. Achaearanea tepidariorum (C. L. Koch) and Theridula opulenta (Walckenaer) are also prevalent in four of the five states (Levi & Field 1954; see discussion below under Theridula emertoni). Enoplognatha ovata (Clerck) is listed as Holarctic in Platnick (2005), but is apparently introduced to North America from Europe, as suggested by Levi (1957). Further evidence of recent introduction is that the distribution map for the species in Levi (1957) showed only four records for eastern North America, and none from our five-state region. Today, E. ovata is one of the most abundant species in the understory of deciduous woods throughout our region (M. Draney, pers. obs.). In light of this recent and dramatic expansion, its status as introduced is all but confirmed. Finally, Achaearanea tabulata Levi is also listed as Holarctic in Platnick (2005), but Dondale et al. (1994) consider it to be introduced. It is now recorded from Illinois as well as Ontario.

The 25 species that were removed from our five-state species list, either because the taxon had not become established (i.e., Heteropoda venatoria), or were misidentified (i.e., Philodromus aureolus), or constitute doubtful records (i.e., Centromerus serratus) are listed in our Spider Species Name Concordance Table (Appendix II).

Native species misidentifications: In light of the taxonomic advances within the past quarter century, misidentifications are often discovered when reviewing some of the older literature. Three particular species that were listed in the Great Lakes states are in question: Crustulina guttata, Euryopis californicus, and Uloborus glomosus. Crustulina altera was not described until 1942, and Chickering obviously recognized the difference between the species he called guttata in 1933 and the one he recorded as sticta in 1935. The abdominal colorations of altera and guttata are more alike than those of guttata and sticta (T.P. personal observation in California). By process of elimination we believe that the Chickering specimens identified as C. guttata were C. altera.

The species that Levi & Field referred to as Euryopis californica was, with little doubt, E. pepini. The Levi & Field publication was dated April 1954 and the Levi revision of Euryopis came out in June 1954. Euryopis pepini (Pepin County, Wisconsin) was described in the June issue and, of course, was not listed in Levi & Field 1954. Pepin was one of the same localities that Levi & Field stipulated for E. californica; also in E. pepini the conductor of the palp has an elbow as in californica, which was mentioned in the Levi & Field publication. Therefore we attribute the Wisconsin specimens of this species to E. pepini. Levi & Field (1954) listed Theridula sphaerula (now considered a synonym of the cosmopolitan Theridula opulenta, see above) as a member of the Wisconsin spider fauna, but identified these specimens as the native Theridula emertoni Levi 1954 in a later publication (Levi et al. 1958).

The salticid Habronattus agilis, reported by several sources for four of the five states (see Table 8 and Appendix I), represents a possible misidentification for H. cognatus (pers. communication B. Cutler). According to Griswold (1987), H. agilis is restricted to the eastern seaboard of the U.S. The records of Pelegrina montana reported from the Milwaukee Public Museum and by Chickering for Michigan may represent misidentifications of P. insignis (B. Cutler, pers. communication). The distribution for this species given by Maddison (1996) supports this assumption. Uloborus glomosus is the name that supplanted U. americanus, an unavailable name according to the rules of the ICZN; Muma & Gertsch (1964) were fairly clear on this issue in their publication.

DISCUSSION

The known spider fauna of the Great Lakes States includes 2.4% of the 38,834 known species worldwide, and 7.9% of the 3593 genera (Platnick 2005). This may seem like a tiny fraction of the world fauna unless the region's land area is taken into account: The Great Lakes states make up only about 0.5% of the world's ice-free land area. The region is actually quite diverse considering its latitude, topography and glaciation history.

Although we here report that only about 3% of the established species in the region are exotic species, we are certain this represents an underestimate of the total proportion of introduced species. Many taxa are insufficiently well known, in terms of their ecology, distribution, and systematics, to be able to evaluate their biogeographic origins with any certainty, and many simply have not been evaluated at all. Arachnological faunistic works have traditionally given scant attention to biogeographic origin, a situation which will hopefully change as arachnologists come to appreciate the critical importance of the native/non-native distinction in conservation contexts, as well as in understanding the ecology and evolution of these animals.

The known spider fauna varies among the Great Lakes states. These differences can be attributed to several environmental factors, such as climate, habitats and the varying extent of undisturbed habitats. Such factors affect the actual spider fauna present in any region at a particular point in time. Climate changes, and more recently, various and extensive habitat alterations through human activity cause faunal changes. Unfortunately, we can neither observe nor measure faunal changes, because our current lists of 'known' spider species records are not complete. Instead they reflect the varying degree of past and ongoing faunistic work in the Midwest states. Especially the low documented diversity of the Indiana spider fauna appears to be attributable to lack of collecting and research effort, as this fact has been documented in numerous faunal studies (Palmer et al. 2002). Our results above demonstrate clearly the impact of past research efforts by individual arachnologists such as Chickering for Michigan, Levi for Wisconsin and Beatty for Illinois. Ongoing faunistic studies are still being supported by a few individuals, such as Draney and Bradley for Wisconsin and Ohio.

However imperfect our current spider species lists may be, these lists are the starting point of biodiversity, biogeography, systematic and evolutionary research. In fact the generation of such species lists, taxonomically updated and available online, is demanded as an essential global resource (Knapp et al. 2005) by a variety of end-users, such as environmental agencies, governmental bodies and researchers far beyond the taxon-specialist (Steenkamp & Smith 2003). Perfecting and updating such legacy species lists is of central importance for their utility and several recent studies suggest a holistic approach, incorporating a variety of different data sources for the improvement of species lists. Current species lists and species gaps (such as we defined above) guide sampling efforts in under-sampled habitats or regions (Palmer et al. 2002). Disparate data sets from different types of surveys can be combined (Crosier & Stohlgren 2004). Bieler & Mikkelsen (2004, see above) demonstrated the value of critically analyzing grey literature, such as governmental technical reports and amateur lists, in combination with focused field work and mining museum collections.

For invertebrates and especially for arthropods, natural history collections represent a vast, yet largely untapped biodiversity information source, as it is now widely recognized e.g., by NSF initiatives such as LINNE (http://www.flmnh.ufl.edu/linne/news.htm). Whereas vertebrate collections, and to a certain degree mollusk collections are computerized, terrestrial arthropod collections lag far behind, mainly due to the sheer number of specimens in these collections. Data models and database development have come a long way (e.g., BIOTA, Ke-EMu [http://www.kesoftware.com/emu/], Specify and others). However, the main hurdle remains the enormous and as of yet unfunded task of data entry. In numerous taxon groups, we may not even know where the specimen collections are housed, e.g., see Sierwald & Reft 2004. The U.S. spider fauna is a case in point; specimens are housed in various U.S. collections, but tracing the U.S. spider fauna of a particular region is almost impossible, since none of the U.S. spider collections are computerized.

We attempted to locate spider collections containing a significant proportion of species from the Great Lakes Region. The following collections are likely to harbor at least some Midwest spider material: Field Museum of Natural History, Chicago (curator Petra Sierwald); Milwaukee Public Museum, Milwaukee (curator Joan Jass); Illinois Natural History Survey (collection manager C. Farvet); Illinois State Museum (curator E.D. Cashatt); Purdue University, Entomology Dept. (curator Arwin Provonsha); Earlham College (Leslie Bishop); Emporia State University, Emporia, KS (curator John Richard); Snow Entomological Museum, University of Kansas (collection manager Z.H. Falin); Ball State University (Gary Dodson); Entomology Museum of Michigan State University (about 5500 lots, adjunct professor Dr. R. J. Snider). Chickering's collection of Michigan spiders was deposited at the Museum of Comparative Zoology at Harvard. The University of Michigan Museum of Zoology does not contain a substantial spider collection. H.K. Wallace made a list of spiders from the George Reserve of Michigan University, but there are no voucher specimens from this study. The majority of the University of Michigan Museum's spider collection was transferred to the Florida State Collection of Arthropods, Gainesville (fide N. I. Platnick). Private collections: J. A. Beatty (Carbondale, Southern Illinois University); J. L. Kaspar (Oshkosh, Wisconsin); R.A. Bradley (The Ohio State University, Marion Campus); M.L. Draney (University of Wisconsin-Green Bay).

In the near future, we will continue to maintain the database and concordance table online and include documented changes submitted by users. Ultimately, this database will become more useful in more varied contexts by its planned expansion along several axes. First, species will continue to be added as we capture the remaining legacy data from published spider species lists, and as identification of newly collected and museum specimens proceeds. Clearly, we have recorded only a fraction (albeit probably a substantial fraction) of the species that occur in each of these states. Secondly, our existing database framework can be extended to adjacent states in our region, and eventually may come to encompass much or all of North America. Lastly, we plan to improve the spatial resolution of much of the data. Almost all of our data can be tracked to county level, and its input into the database can provide much finer spatial information about spider distribution across the region. The centroid location of each county can be used as a location index for many biogeographic purposes. Additionally, many of the more recent records have latitude and longitude data attached (or can be get-referenced post-hoc with a fair degree of accuracy). Fine-scale location data on at least a subset of the recorded records can be used to address landscape-scale questions about entire assemblages of organisms. Extensive regional-scale spatial data on the complete complement of a region's species has never existed for spiders. We want to make such data publicly accessible because the potential applications of such a dataset are limited only by the imaginations of researchers.

ACKNOWLEDGMENTS

Numerous individuals supported this effort. We are greatly indebted to Joan Jass (Milwaukee Public Museum), who provided us with unpublished spider voucher lists from the collections of that Museum. She also permitted us to incorporate the species lists based on literature data that she compiled for Wisconsin and Michigan. She contributed greatly to the section on native species of the five Great Lakes States. Colin Favret, collection manager at the Illinois Natural History Survey made a list of the entire Midwest spider collection at the Survey available to us. Dr. Richard Bradley kindly permitted us to use his spider species data from Ohio. Clearly, such efforts as the one presented above depend heavily on the work of others and the collaboration between taxonomists. Dr. R. Waltz (State Entomologist, Division of Entomology & Plant Pathology, Indianapolis), Dr. Gail Stratton (University of Mississippi, Mississippi), Dr. D. H. Cameron (University of Michigan, Ann Arbor), L. Leibensperger (Museum of Comparative Zoology, Harvard University) and Dr. James Berry assisted us in locating spider collections harboring Midwest spiders. The Cook County Oak Savannah survey 1996-1999 was organized through the Environmental and Conservation Program (ECP) at the Field Museum of Natural History, and funded through grants by the Illinois Department of Natural Resources (administered through the Chicago Wilderness Coalition), with collecting permits issued by the Illinois Nature Preserves Commission and the Cook County Forest Preserve District. Reviews by Drs. B. Cutler and C. D. Dondale greatly improved the manuscript.