Systematics, conservation and morphology of the spider genus Tayshaneta (Araneae, Leptonetidae) in Central Texas Caves

Abstract The spider genus Tayshaneta is revised based on results from a three gene phylogenetic analysis (Ledford et al. 2011) and a comprehensive morphological survey using scanning electron (SEM) and compound light microscopy. The morphology and relationships within Tayshaneta are discussed and five species-groups are supported by phylogenetic analyses: the anopica group, the coeca group, the myopica group, the microps group and the sandersi group. Short branch lengths within Tayshaneta contrast sharply with the remaining North American genera and are viewed as evidence for a relatively recent radiation of species. Variation in troglomorphic morphology is discussed and compared to patterns found in other Texas cave invertebrates. Several species previously known as single cave endemics have wider ranges than expected, suggesting that some caves are not isolated habitats but instead form part of interconnected karst networks. Distribution maps are compared with karst faunal regions (KFR’s) in Central Texas and the implications for the conservation and recovery of Tayshaneta species are discussed. Ten new species are described: Tayshaneta archambaulti sp. n., Tayshaneta emeraldae sp. n., Tayshaneta fawcetti sp. n., Tayshaneta grubbsi sp. n., Tayshaneta madla sp. n., Tayshaneta oconnorae sp. n., Tayshaneta sandersi sp. n., Tayshaneta sprousei sp. n., Tayshaneta vidrio sp. n. and Tayshaneta whitei sp. n. The males for three species, Tayshaneta anopica (Gertsch, 1974), Tayshaneta devia (Gertsch, 1974) and Tayshaneta microps (Gertsch, 1974) are described for the first time. Tayshaneta furtiva (Gertsch, 1974) and Tayshaneta uvaldea (Gertsch, 1974) are declared nomina dubia as the female holotypes are not diagnosable and efforts to locate specimens at the type localities were unsuccessful. All Tayshaneta species are thoroughly illustrated, diagnosed and keyed. Distribution maps are also provided highlighting areas of taxonomic ambiguity in need of additional sampling.


Introduction
Tayshaneta are small spiders that belong to the family Leptonetidae, a group recognized for its association with caves and similar cryptic habitats (Ledford et al. 2004). Tayshaneta are widely distributed in caves of the Edward's Plateau (Fig. 3), an extensive limestone region in Central Texas that drains into the Edward's Aquifer and serves as the primary source of water for over 2 million people. The region is famous for its endemism and includes a high proportion of endangered and threatened species, many of which are subterranean specialists and known only from single springs or caves (Culver et al. 2003). Two Tayshaneta species are federally listed as endangered in Central Texas, T. microps (Gertsch, 1974) and T. myopica (Gertsch, 1974) and most others are of conservation concern (Bender et al., 2005;U. S. Fish andWildlife Service 1998, 2010). However, management and recovery efforts are limited by existing taxonomy which is poorly resolved and leaves the identity and distribution of Tayshaneta species ambiguous. Gertsch (1974) described the majority of the North American Leptonetidae and considered twelve species as part of a closely related Texas fauna. Although he originally described these species as congeneric with European Leptoneta, several publications (Brignoli 1977(Brignoli , 1979Platnick 1986) refuted this hypothesis and transferred the Texas fauna to the genus Neoleptoneta Brignoli, 1972. Two species were later added by Cokendolpher and Reddell (2001) and Cokendolpher (2004), who also provided details on their general biology. Recent phylogenetic work has shown that Neoleptoneta is paraphyletic and three additional genera, Chisosea, Ozarkia and Tayshaneta, were described . Tayshaneta presently includes eleven species restricted to Texas caves with close relatives in the Southeast, Southern Texas and Northern Mexico .
While Gertsch's (1974) study was the first to comprehensively treat the North American fauna, the taxonomic challenges of leptonetids frustrated him (D. Ubick, pers. comm.). Most species are represented by few specimens which in addition to being relatively small (1-2mm) are also delicate and easily damaged during exami-nation. Furthermore, the characters used to separate species are exceptionally fine and not often visible using conventional microscopy. European specialists, including Brignoli (1972Brignoli ( , 1974, Fage (1913) and Machado (1941Machado ( , 1945 relied heavily on compound light microscopy to produce detailed illustrations which Gertsch was reluctant to use. Consequently, most species remain poorly diagnosed and positive identification is only possible with topotypic material. Morphological homogeneity within female specimens is also problematic (Ledford 2004;Ledford and Griswold 2010) and although microscopy and preparation techniques have improved, leptonetid taxonomy remains dependent upon the details of male genitalia. Diagnostic features for Tayshaneta in particular are subtle and often require examination using scanning electron microscopy.
Recent studies on Cicurina spiders in Texas caves (Paquin and Hedin 2004;Paquin et al. 2008) have addressed similar problems by using molecular phylogenetic methods and fine scale geographic sampling to help resolve species limits. Although based on a single genetic locus, Paquin and Hedin (2004) clearly demonstrated that the integration of molecular data is a valuable aid to overcoming the difficulties of working with cave fauna, especially when specimens are rare or present diagnostic challenges. Studies of cave invertebrates are also underscored by conservation concerns, especially in Central Texas, where taxonomic identity can have profound socioeconomic impact. As emphasized by Paquin et al. (2008), the interaction between taxonomists, conservation biologists and development interests can be volatile and highlights the need for robust, integrative taxonomy based on multiple lines of evidence.
Several geological areas are recognized on the Edward's Plateau, however most of the subterranean diversity is known from caves along the heavily faulted Balcones Escarpment (Fig. 3). The faulting serves to isolate regions of limestone and is likely correlated with the diversification patterns of cave invertebrates (White et al. 2009). Conservation biologists have used this fragmented geology to develop a conservation strategy based on "karst faunal regions" (KFR's), hypothesized as biologically discrete areas of cave habitat that are used to manage species recovery (U. S. Fish and Wildlife Service 1994; Veni 1992Veni , 1994. Three  are currently recognized in Bexar, Travis and Williamson Counties each of which includes large numbers of caves and encompasses the distributions of multiple endangered invertebrates. However, KFR boundaries are limited by existing taxonomy which in most cases does not accurately reflect species distributions (White and Carothers 2001).
This study revises the taxonomy of Tayshaneta based on the phylogenetic results of Ledford et al. (2011) and data collected from a morphological survey using scanning electron and compound light microscopy. Ten new species are described, along with three previously unknown sexes and all remaining species are imaged, diagnosed and keyed. The morphology and relationships within Tayshaneta are discussed and five species-groups are identified. Distribution maps are provided along with an evaluation of KFR's based on revised species distributions. The primary objective of this study is to produce a functional taxonomy for Tayshaneta that will facilitate conservation and management efforts and contribute to an understanding of the Texas cave fauna.

Taxon sampling
A resurgence of interest in Texas cave biology, driven largely by conservation efforts, has produced a wealth of new Tayshaneta specimens more than doubling records since Gertsch (1974). In order to prioritize collection sites, a database combining records for described species and all recent collections was developed. Collection sites were then selected to maximize sampling throughout known ranges with priority given to type localities. Outgroup selection was based on the most recent phylogenies of haplogyne spiders (Platnick et al. 1991;Ramirez 2000) and specimen availability. Between 1-10 individuals were collected from each site, placed directly into 95% ethanol and then transferred to storage at -20°C. Each specimen was assigned a unique voucher number and is accessioned in a database maintained at the California Academy of Sciences (CASC).
Voucher specimens for the study are deposited at the California Academy of Sciences (CASC), the Texas Memorial Museum (TMM), the Museum of Texas Tech University (TTU) and the Essig Museum, University of California, Berkeley (UCB).
Due to the sensitive nature of cave locations and in the spirit of respecting the rights of property owners and encouraging future research, precise locality information is not provided. Unless otherwise noted, all cave locations are limited to within 2 kilometers. Specimens used in this study along with their voucher codes are listed in Ledford et al. (2011) and a map highlighting the study area is provided in Fig 3. Distribution maps were produced using Arc GIS 10.0 (Environmental Systems Research Institute, CA). Karst faunal region boundaries were derived from shape files provided by Zara Environmental (K. O'Connor) through the U.S. Fish and Wildlife Service.

Morphology
Prior to examination with a Leo 1450VP Scanning Electron Microscope, all structures were cleaned with a fine brush or ultrasonicator and critical point dried. Best results were obtained by gradually dehydrating the specimen in increasing concentrations of ethanol for 24-48 hours prior to critical point drying. Dried specimens were then mounted on pin mount SEM stubs (Ted Pella Inc., Redding, USA) on copper-backed tape. Specimens were sputter coated for 120 seconds using a Denton Vacuum Sputter Coater. Large structures were photographed using a Nikon DMX1200 camera attached to a Leica MZ 16 stereomicroscope. Images were then montaged using Helicon Focus v. 4.2.1 (http://www/heliconsoft.com). For male specimens, the right palp was scanned and the left was maintained with the specimen post examination using compound light microscopy.
Vulvae were carefully excised and placed in a pancreatin solution for 24-48 hours to digest extraneous tissue (Alvarez-Padilla and Hormiga 2008) then placed in water and manually cleaned. Best results were obtained by removing the cuticle from the dorsal surface of the abdomen and digesting the entire structure. If the vulva remained unclear, it was stained for one minute with Chlorazol Black and reexamined. Images of each species were prepared using a Nikon DMX1200 camera attached to a Leica DM 4000 compound microscope. Genitalia were placed in Hoyer's solution and examined in well slides or temporary mounts following the procedure described by Coddington (1983).
Descriptions follow the format of Ledford and Griswold (2010) and Ledford (2004). Descriptions of previously unknown sexes were based upon individuals collected at the type locality. All measurements are in millimeters and quantify the structure at its widest or longest point. A summary of anatomical abbreviations used in the descriptions and keys is provided in Table 1. Individual images of all structures will be made available at the time of publication in Morphbank (www.morphbank.net) and species pages will be available in the Encyclopedia of Life (http://www.eol.org).

Phylogeny
Detailed protocols for the extraction, amplification and sequencing of DNA are reported in Ledford et al. (2011). Three gene fragments were selected based on availability, prior use in systematics studies and amplification success. Mitochondrial cytochrome oxidase I (~800bp), nuclear histone 3 (~330bp) and 28s rDNA (~1000bp) were amplified following Ledford et al. (2011) and the primers and conditions used are reported in Table 3. Phylogenetic methods also follow Ledford et al. (2011) and both independent genes and concatenated data were analyzed under a variety of optimality criteria and conditions (Table 2). Sequence alignment was performed using CLUSTAL × v. 2.0 (Larkin et al. 2007) and additional 28s rDNA alignments were produced using Muscle v. 3.8 (Edgar 2004). Models of nucleotide evolution were selected using the Akaike Information Criterion (Akaike 1973) as implemented in MrModeltest v. 2.2 (Nylander 2004). Partitioning strategies for COI and histone 3 were evaluated using Bayes Factors (Brown and Lemmon 2007) for fully partitioned, partially partitioned and unpartitioned analyses. Bayesian analysis was performed using MrBayes v. 3.1.2 (Huelsenbeck and Ronquist 2001) using four independent runs until the standard deviation of split frequencies fell below 0.01. Stationarity was evaluated by examining the stability of posterior probabilities for nodes of each MCMC run using the Cumulative and Compare plots in Are We There Yet? (http://ceb.csit.fsu.edu/awty; Nylander et al. 2008) and the first 25% of trees were discarded from the posterior distributions of each analysis. Maximum likelihood analysis was performed using 1000 bootstrap replicates in RAxML v. 7.0.4 (Stamatakis 2006) and parsimony analyses were performed in PAUP* (Swofford 2003) using 1000 iterations of a heuristic search holding 100 trees for each iteration, with random taxon addition and tree bisection-reconnection (TBR) branch-swapping. Branches were collapsed using the default rule in PAUP* v.4.0 b10 (collapse if maximum length is zero). Nonparametric bootstrap support values were calculated using 1000 replicate searches with random taxon addition. Aligned data matrices and trees will be made available online in TreeBASE (http:// www.treebase.org/).

Morphology
Exemplars for each Tayshaneta species, including undescribed species discovered during the course of this study, were photographed using automontage, compound and scanning electron microscopy. Holotype specimens for each species were examined in order to confirm the identity of exemplars used in analyses. Images provided in this study are either taken directly from the holotype or from specimens collected at the type locality. Over 3,000 images were produced based on a set of standardized views and assembled into comparative plates. Careful attention was directed at diagnostic characters provided in Gertsch (1974) and to somatic features in order to assess variation in troglomorphic morphology.
Putative synapomorphies for Tayshaneta include a unique conformation of the female genitalia, with short spermathecal stalks bearing large heads (SH, and the recurved to straight retrolateral spine on the male palpal tibia (RTS, Figs 32A-F). Body color ranges from pale brown-yellow to depigmented with faint dark patterns surrounding the eyes and ocular area. The legs are covered in fine setae and bear few scattered spines. A ventroapical preening comb on metatarsus III was observed in each species examined (Fig. 12-13 in Ledford 2004). Patellar and tibial gland morphology was similar to that described by Platnick (1986) with triangular patellar plates bearing single small pores (Figs 30-31, 33, 38, 40, 46 in Platnick 1986). The abdomen lacks distinctive patterning, is sparsely setose and pale yellow to white in color. Spinning organs follow the descriptions of Leptoneta infuscata Simon, 1872 (Ledford and and Calileptoneta (Ledford 2004) with the exception of bearing fewer aciniform gland spigots (6-10) on the PMS and PLS .
In contrast to other leptonetine genera, the palpal morphology of Tayshaneta is relatively conserved and the bulb bears few spines, specialized setae, or accessory sclerites. The shape of the palpal tarsus is of two basic types; divided, as in T. fawcetti sp. n. (Fig. 31D) and tapering, as in T. coeca, . The depth of the division ranges from deeply divided as in T. fawcetti sp. n. and T. vidrio sp. n. (Figs 31D-E) to weakly divided or swollen as in T. madla sp. n. (Fig. 31F). An exposed tarsal organ is present dorsoapically and consists of a shallow circular base with a pair of round receptors (Figs 24G-H in Ledford et al. 2011). The embolus is weakly sclerotized, transparent and connected via a short tube to a large reservoir in the bulb (Figs 30A-D in Ledford et al. 2011). The sculpture along the margins of the embolus ranges from smooth as in T. coeca and T. myopica (Figs 36D,44D) to bearing tooth-like extensions and folds as in T. anopica (Fig. 33D). The embolus is typically curved or folded around the ventroapical portion of the bulb and bears a single, circular opening (Fig. 44F).
The ventral sclerite (VS) is a single, spine-like projection that extends approximately half the length of the embolus. The position and length of the VS ranges from elongate and mesal as in T. fawcetti sp. n. (VS, Fig. 40E), to retroventral as in T. myopica (VS, Fig.  44B) and short as in T. sprousei sp. n. (VS, Fig. 48E) The VS is absent in several species, including T. coeca (Fig. 36E) and despite repeated efforts to determine whether this structure was related to expansion no VS was observed. The retrolateral sclerite (RS) is of two types, a shallow, pocket-like invagination as in T. fawcetti sp. n. (RS, Fig. 40E-F) or a distinctly separated, oval sclerite as in T. whitei sp. n. (RS,. The retrolateral tibial spine (RTS) is recurved to straight and ranges from short, occupying less than half the length of the palpal tarsus (RTS,D,36F), to elongate in which the spine extends greater than half the length of the palpal tarsus (RTS,Figs 31C,F,32C,F). The RTS is situated on a shallow to pronounced base and is moveable, possibly serving as a positioning structure during mating. A fine, comb-like sculpturing extends along the entire length of the RTS in most species, but may also be smooth near the base as in T. fawcetti sp. n. (Fig. 32D) and T. devia (Fig. 32B). Between three and four flattened setae are located near the base of the RTS (Figs 32A-F) along with several unmodified setae surrounding the base.
Examination of female genitalia using compound microscopy revealed relatively little variation among species and in most cases female specimens appear nearly identical in structural details . The preparation of female genitalia was problematic as the weakly sclerotized spermathecal stalks do not remain in a fixed position and slight differences in orientation can dramatically alter the structure's appearance. Even with careful preparation techniques the vulva is difficult to precisely position for comparison among individuals. The atrium is suboval to triangular and covered in fine pores. The spermathecal stalks are twisted and connect to the atrium basally via short sclerotized tubes. The spermathecal heads are swollen, circular (Figs 52A, C-F, 53A-B, D-F, 54A-C, E) to elongate (Figs 53C, 54D) and covered in fine pores.

Phylogeny
Results of phylogenetic analyses follow Ledford et al. (2011) and summary statistics for each analysis are presented in Table 2. Phylograms for concatenated analyses (Bayesian, maximum likelihood, parsimony) are presented in Figures 4-6 and independent gene trees are in Figures 7-9. Nodes with a posterior probability of 95% and greater are considered supported and all remaining nodes are collapsed. Nodes for maximum likelihood and parsimony analyses with bootstrap support values of 75% and greater are considered supported and all remaining nodes are collapsed.

Discussion
Among the most interesting results of the phylogenetic analyses is the contrast in branch lengths between Tayshaneta and the remaining North American genera (Figs 4-6). Although sampling and rate variation among genes (Figs 7-9) are known to affect branch lengths, the close relationships, morphological similarity and narrow geographic distributions of Tayshaneta suggest that it is a relatively recent radiation of species. Similar radiations are known for Cicurina spiders (Paquin and Dupérré 2009;Paquin and Hedin 2004) and Texella harvestmen Briggs 1992, 2004) both of which show similar biogeographic patterns and affinity for caves. Recent work has shown that the diversification patterns of Cicurina is correlated with the complex faulting in the region (White et al. 2009) and may serve as a general model to explain the diversity of the Texas cave fauna. On-going work has been directed at synthesizing the distributions for multiple cave invertebrates in order to develop a comprehensive understanding of the Texas fauna (Reddell et al. in prep.).
Although most Tayshaneta species have relatively conserved genitalic morphology, intraspecific variation in somatic features related to cave life (troglomorphism) is extreme and often includes a range of eye and pigment reduction. In T. myopica, for example, multiple morphotypes are often found within a narrow geographic distribution and range from darkly pigmented, large-eyed individuals (Figs 55E-F) to lightly pigmented, reduced-eyed forms ( Fig. 55A-C), to complete eye and pigment loss (Fig.  55D). While these differences likely indicate varying degrees of local adaptation to caves, the intergradient morphologies observed suggest that some species may have an adaptive cline from surface to cave-adapted morphotypes. Similar patterns of troglomorphic variation have been reported in Texella harvestmen that show multiple degrees of troglomorphic morphology between closely related species Briggs 1992, 2004). In T. reddelli and T. reyesi for example, species limits are often indistinct as specimens show a gradual reduction in eyes, pigment and tubercles on the carapace. One intriguing hypothesis is that populations are actively colonizing caves and becoming increasingly more troglomorphic, similar to the adaptive shift model proposed for Hawaiian isopods (Rivera et al. 2002).
Biogeographic relationships within Tayshaneta reflect the fragmented geology of region as distributions are allopatric and few cases of sympatry are known. However, distributions for most species remain poorly characterized and reflect incomplete sampling, especially of surface localities, which are rarely inventoried as part of cave surveys. Species distributions in Bexar and Travis Counties are particularly complex and several undetermined records (Fig. 61) likely represent range extensions or additional species, the identification of which will help resolve areas of taxonomic ambiguity. The most significant area of biogeographic ambiguity are caves and surface habitats in Comal and Hays Counties both of which remain poorly inventoried and are essential to resolving species limits, especially between T. coeca (Chamberlin & Ivie, 1942) and T. devia (Gertsch, 1974).
The majority of species described by Gertsch (1974) were known from single localities and was used as the primary justification for the endangered status of T. microps (Gertsch, 1974) and T. myopica (Gertsch, 1974) (U. S. Fish and Wildlife Service 1994, 2000. Recent sampling efforts, combined with the molecular and morphological data presented in this study, have shown that most species are more broadly distributed than expected but still of limited distribution. Furthermore, molecular data suggest that most troglobitic species are actively using subterranean microfissures and voids as corridors for dispersal between caves. The most striking examples are for the species T. anopica (Gertsch, 1974), T. myopica (Gertsch, 1974) and T. sandersi sp. n. each of which have populations in different caves that share identical haplotypes for the loci surveyed in this study. While these connections are not surprising given the geology of the area, they set a precedent for interpreting the distribution of other Tayshaneta species and are likely to effect conservation and management decisions.

Karst faunal regions
Karst faunal regions (KFR's) were originally developed as tools to aid the recovery of endangered karst invertebrates by identifying geologically independent regions that had a relatively high proportion of endemic species (Veni, 1992(Veni, , 1994. Although an evolutionary model was not explicitly proposed, the inherent reasoning is that the present distribution of the karst invertebrate fauna can be explained by the fragmented geology of the region (White and Carothers 2001). Although recent work has shown that phylogenetic divergence within Cicurina spiders is likely correlated with faulting, the distributions of most invertebrate groups are poorly understood which precludes a synthesis of biogeographic patterns in the area. Furthermore, the endemicity index used to help define KFR's is necessarily constrained by existing taxonomy, most of which is inadequately resolved or erroneous (Paquin and Dupérré 2009).
While the distribution of Tayshaneta is broader than anticipated, it is nevertheless highly restricted, especially when compared to other endangered invertebrate groups. Cicurina and Texella, for example, have highly active hunting lifestyles and are known to occur in far more caves. In contrast, Tayshaneta are more sedentary, spending most of their lives in webs with the exception of males that may leave the web upon maturity. Not surprisingly, the distribution of most Tayshaneta species closely corresponds to established KFR's. In Bexar County, T. microps is restricted to the Government Canyon KFR (Fig. 63) and despite extensive sampling no additional populations have been discovered. T. madla sp. n. and T. whitei sp. n., however, occur in multiple KFR's and although are not currently listed as endangered show that KFR's are not biologically exclusive as presently defined. T. myopica (Fig. 62) shows a similar pattern in Travis County, where most populations are known from the Jollyville KFR as well as in the McNeil/ Round Rock KFR.
Following the arguments of White and Carothers (2001), the presentation of these data is not designed to be a critique of the KFR strategy but rather highlights that the geological complexity and phylogenetic histories of invertebrates in the region make the delineation of boundaries a daunting task. From a conservation perspective, the use of KFR's have been successful at acquiring new cave habitat and establishing karst preserves, both of which are essential to the long-term protection of the karst invertebrate fauna. As recovery plans, local initiatives and monitoring continue to develop in the region taxonomic studies that integrate all available data will be essential to the successful implementation of the KFR conservation strategy.

Key to species of Tayshaneta
The key presented here relies heavily on fine details of the male and female genitalia, some features of which are not visible using conventional light microscopy or without special preparation techniques. Scanning electron and compound light microscopy is essential for positive identification and the females of most species are not diagnosable in the absence of associated males.  (Gertsch, 1974) showing karstic terrain B Entrance to Government Canyon Bat Cave, Bexar County, Texas, type locality for T. microps (Gertsch, 1974) (Gertsch, 1974), female, Geode Cave, Travis County, Texas B T. fawcetti sp. n., male and female in web, Fawcett's Cave, Val Verde County, Texas C T. myopica (Gertsch, 1974) (Gertsch, 1974) male, Government Canyon Bat Cave, carapace dorsal view B T. microps (Gertsch, 1974) male, Government Canyon Bat Cave, ocular area C T. coeca (Chamberlin and Ivie, 1942) male, New Braunfels, carapace dorsal view D T. myopica (Gertsch, 1974) male, Pedernales River, sternum E T. myopica (Gertsch, 1974) (Gertsch, 1974) male, Pedernales River, arrow to colulus B T. devia (Gertsch, 1974), MacDonald Cave, spinning field C T. devia (Gertsch, 1974) Ledford et al. 2011: 334-388 Type species. Leptoneta coeca Chamberlin & Ivie, 1942. Nomen dubium. Leptoneta furtiva (Gertsch, 1974) is described on on the basis of a single female specimen from Blackwell, Nolan County, Texas. The holotype is in poor condition, missing most of its appendages and genitalia. Efforts to recollect the species at the type locality have proven unsuccessful and the lack of diagnostic features prevents its diagnosis from any other Tayshaneta species. Leptoneta uvaldea (Gertsch, 1974) was described from Story Cave, Uvalde County, Texas, based on a single female specimen. While the holotype is in good condition, the genitalia are damaged and it cannot be separated from any other Tayshaneta species. Furthermore, the type locality, Story Cave, is widely recognized as a lost cave somewhere on the Marneldo Ranch (A. Gluesenkamp, pers. comm.). Given their lack of diagnostic features, both species are declared nomena dubia until additional specimens near the type localities can be obtained.
Diagnosis. Tayshaneta is separated from all other leptonetids by having males with a recurved to straight retrolateral spine on the palpal tibia (Figs 32A-F) and females with short spermathecal stalks bearing large circular to oval heads (Figs 52-54).
Putative synapomorphies. Species of Tayshaneta are united by the unique conformation of the female genitalia, with short spermathecal stalks bearing large heads (Figs 52-54) and the recurved to straight retrolateral spine on the male palpal tibia (Figs 32A-F).
Description. Complete description of female in Gertsch (1974: 172). Habitus of female in Figs 12D-F, genitalia as in Fig. 52A and images of egg-sac in Figs 2D, 52B.

Variation (n = 2).
Natural History. An egg-sac for this species was found with a female specimen from Corn Cobb's Cave (Figs 2D, 52B). The egg-sac was found hanging by a single thread covered with small pebbles and contained two eggs. Diagnosis. Tayshaneta archambaulti can be separated from all Tayshaneta species that lack a ventral sclerite, except T. coeca and T. devia, by the following combination of characters: embolus oval to quadrate, lacking sculpture along its margin (E, Fig.  34D); retrolateral tibial spine short, occupying less than 0.50× the length of the palpal tarsus (RTS, Fig. 34A). Separated from T. devia by having a retrolateral tibial spine with sculpture along its entire length and from T. coeca by having the embolus curved distally and extending beyond the apical portion of the bulb (E, Fig. 34E).
Distribution. This species is known only from Burnett Ranch Cave and Grapevine Cave in southwestern Hays County (Fig. 58).
Natural History. Individuals for this species were collected throughout Grapevine Cave, however, most specimens were encountered at the base of the cave's vertical entrance in the twilight area under stones. They were collected in fine sheet webs similar to other Tayshaneta species. Diagnosis. Tayshaneta bullis can be separated from all other Tayshaneta species that lack a ventral sclerite by having an elongate retrolateral tibial spine at least 0.5× the length of the palpal tarsus (RTS, Fig. 35A) and a distinctly quadrate shaped embolus (E, Fig. 35D).
Description. Complete description in Cokendolpher (2004: 65). Habitus of male and female in Figs 14A-F, scanning electron micrographs of male palp in Figs 35A-F and female genitalia in Fig. 52D.
Distribution. Known from two caves in Bexar County, Up the Creek Cave on Camp Bullis and Hills and Dale's Pit (Fig. 59). Natural History. Cokendolpher (2004) reported on the shape of the egg-sac for this species along with details on their general biology. The egg-sac was covered in small pebbles or detritus similar to that observed for T. anopica (Fig. 2D, 53B). Females were observed to retain sperm for several months and the egg-sacs contained few, relatively large eggs. Type data. Male holotype from Heidrich's Cave, New Braunfels, 20-June-1938, Comal County, Texas, 20-June-1938.70N, 98.10W, (AMNH, formerly in the University of Utah collection, examined).
Notes. Heidrich's Cave was the name used by Chamberlin and Ivie (1942) for Brehmmer Cave in the original description of the species (Reddell and Cokendolpher 2004). Gertsch (1974) considered specimens from Natural Bridge Caverns as conspecific with T. coeca, however, no illustrations or diagnostic details were provided. Female specimens from Natural Bridges Caverns show similar somatic morphology and genitalia, but cannot be confidently determined in the absence of associated males. While male specimens are reported in Gertsch (1974) they were not located in collections. Given its proximity to the type locality and morphological similarity the specimens are tentatively maintained as conspecific. In several cases, specimens of T. devia were difficult to separate from T. coeca except by the fine details of the retrolateral tibial spine and embolus. Given the geographic disjunction between populations in Comal and Williamson Counties, additional sampling is required in these area, especially on the surface, in order to refine species limits.
Other material examined. USA: Texas: Comal County: Brehmmer Cave (=Heidrich's Cave), 5mi. W. of New Braunfels, 19-March-1960, W. Gertsch, W. Ivie, Sch- Diagnosis. Tayshaneta coeca can be separated from other Tayshaneta species that lack a ventral sclerite, except T. archambaulti and T. devia, by having a short retrolateral tibial spine, occupying less than 0.5× the length of the palpal tarsus (RTS, Fig. 36F) and a rectangular embolus that lacks sculpture along its margin (E, Fig. 36D). Separated from T. devia by having a retrolateral tibial spine with sculpture along its entire length (RTS, Fig. 36C, F) and from T. archambaulti by the distinctive shape of the embolus (E, Fig. 36D).
Description. Complete description in Gertsch (1974: 170-171). Habitus of male and female in Figs 15A-F, scanning electron micrographs of male palp in Figs 36A-F and female genitalia in Fig. 52E.
Distribution. Caves and surface localities in Hays and Comal Counties (Fig. 58). Notes. Gertsch (1974) included a single female specimen from Stark's North Mine in Travis County as conspecific with T. concinna although it is unclear which characters he based this decision upon. Stark's North Mine is a unique feature in the Austin chalk formation and appears to be largely artificial, probably carved out by local residents. Recent inventories at the site have recovered additional Tayshaneta Figure 16. Tayshaneta concinna (Gertsch, 1974), Lost Gold Cave, Travis County, Texas (male holotype, AMNH), habitus. A T. concinna male, dorsal B T. concinna male, ventral C T. concinna male, lateral. specimens, including adult males, which share the genitalic morphology of T. concinna and are recovered as part of the concinna clade (Clade A, Fig. 4). Given the highly disturbed nature of the habitat, it is likely that T. concinna also occurs on the surface. Although adult males are not available from the populations in Seibert Sink (Travis County) or County Line Bat Cave (Williamson County), molecular analyses support them as close relatives of T. concinna and they are tentatively assigned to the species pending the discovery of males.

Notes. Shultz
Cave is commonly referred to as MacDonald Cave and is located approximately 2.5mi. NE of Volente in Travis County. Although the male for this species was not available to Gertsch (1974), recent inventories of caves in this area have produced the first male specimens and added several new records from nearby caves. Of special interest are records from leaf litter near the entrance of Tooth Cave (type locality for T. myopica), approximately 2 miles south of MacDonald Cave. Although Gertsch (1974: 171-172) originally described T. devia as a troglobite based on the type specimen's reduced eyes and pigment, the discovery of surface populations suggests that the species is a widespread troglophile although some populations may be locally adapted to caves. One record from Williamson County (Village Idiot Cave) is tentative as diagnostic structures on the male palp are partially obscured.
Diagnosis. Tayshaneta devia may be separated from other Tayshaneta species that lack a ventral sclerite, except T. archambaulti and T. coeca, by having a short retrolateral tibial spine, occupying less than 0.5× the length of the palpal tarsus (RTS, Fig. 38A) and an apically tapering subquadrate embolus that lacks sculpture along its margin (E, Fig. 38D). Separated from T. archambaulti and T. coeca by having a retrolateral tibial spine with a base that lacks distinctive sculpture (RTS, Fig. 31B) and by the unique shape of the embolus (E, Fig. 38D).
Description. Complete description of female in Gertsch (1974: 171-172). Habitus of male and female in Figs 17A-E and female genitalia in Fig. 53A.
Female (Fawcett's Cave). Body length 1.4, carapace 0.60 long, 0.50 wide, length 1.17× width. Pigmentation and setation same as for male, except ocular area with a faint dark pattern enclosing the AER (Figs 19D-F). Legs elongate and thin, femur I 1.6× carapace length, covered in fine setae and with few scattered spines. Atrium trapezoidal, length 0.73× width, spermathecae with short twisted stalks and elongate heads (Fig. 53C). Abdomen pale brown, without pattern, 0.80 long, 0.58 wide, covered in fine setae.
Distribution. Known only from Fawcett's Cave in the Devil's River State Natural Area, Val Verde County, Texas (Fig. 60).
Natural History. Individuals of T. fawcetti were photographed during a 2009 expedition to Fawcett's Cave (Fig. 2B) where they were observed to make fine sheet webs similar to other leptonetid spiders. Male and female pairs were often found in the same web and the egg-sacs were suspended near the web margins. Most specimens were found at the base of the cave's vertical entrance in twilight under loose rocks and breakdown material. Tayshaneta grubbsi sp. n. urn:lsid:zoobank.org:act:22A96E29-F1DE-4AEB-8B90-5DF97342F67A http://species-id.net/wiki/ Tayshaneta_grubbsi  Figs 20A-C, 32E, 41A-F, 60 Type data. Male holotype from Litterbarrel Cave, 5mi. southeast of Comstock, Val Verde County, Texas, 1-September-1974, S. Sweet, M. Reaka, 29.65N, 101.16W, (AMNH).
Etymology. This species is named in honor of Andy Grubbs, a remarkable collector of several new Tayshaneta species throughout Texas.
Note. The coloration of this specimen has likely been affected by its preservation conditions.
Distribution. Known only from Litterbarrel Cave, Val Verde County, Texas (Fig. 60). Etymology. This species name is taken in apposition to the type locality and honors the Madla family, owners of Madla's Cave and the surrounding property. Diagnosis. Tayshaneta madla may be separated from all Tayshaneta species, except T. bullis and T. microps, by having males with an elongate retrolateral tibial spine (Figs 31F, 32F), more than 0.5× length of the palpal tarsus and lacking a ventral sclerite (Figs 42B, E). Separated from T. bullis and T. microps by the unique shape of the embolus with an enlarged basal tooth (E, Fig. 42D, F).
Notes. Tayshaneta microps was listed under the Endangered Species Act in 2001 (U. S. Fish and Wildlife, 2010) due to pressure from urbanization in areas surrounding San Antonio, Texas. Two records are currently reported for the species, Government Canyon Bat Cave and Surprise Sink, both of which are in Northern Bexar County. The two specimens from Surprise Sink were examined in detail and while they share reduced eyes similar to T. microps, both specimens are immature cannot be confirmed as this species in the absence of associated males.
Diagnosis. Tayshaneta microps may be separated from all Tayshaneta species, except T. bullis and T. madla, by having males with an elongate retrolateral tibial spine (RTS, Figs 31F, 32F), more than 0.5× length of the palpal tarsus and lacking a ventral sclerite (Figs 42B, E). Separated from T. bullis and T. madla by the unique shape of the embolus (Fig. 43D).
Natural History. One adult male specimen was collected for DNA extraction and scanning electron microscopy in November 2009. Although only a single male was found, immature and female specimens were commonly observed in small sheet webs under breakdown material and at the base of walls on opposite sides of the cave entrance.
Description. Complete description in Gertsch (1974: 169-170). Habitus of male and female in Figs 23A-F, scanning electron micrographs of male palp in Figs 44A-F and female genitalia in Fig. 53F.
Distribution. Known from caves in Travis and Williamson Counties, Texas (Fig. 57). Natural History. Individuals in Geode Cave and Tooth Cave were observed suspended beneath sheet webs at the bases of stable rocks and breakdown material (Figs 2A,2C). When disturbed, individuals would drop from their webs and fold their legs in a protective posture similar to that reported for Calileptoneta (Ledford, 2004).
Etymology. This species is named in honor of Kathleen O' Connor, fellow caver and biologist who helped collect many exciting Tayshaneta specimens.
Notes. A single adult male collected from Cathy's Cave, Hays County, Texas shares the genitalic morphology of T. oconnori but was damaged during examination and only the right palp remains. The specimen was highly troglobitic and is tentatively assigned to T. oconnori until additional specimens can be collected.
Diagnosis. Tayshaneta oconnori may be separated from all Tayshaneta species, except T. anopica and T. sandersi, by having the following combination of characters: pigmentation and eyes entirely absent (Figs 24A-C); legs extremely long and thin, femur I 1.8-1.9× carapace length; embolus with a distinctive apical bifurcation (E, Fig. 45D). Separated from T. anopica and T. sandersi by having the ventral sclerite straight and short, not extending past the base of the embolus (VS, Fig. 45E) and by the unique shape of the embolus (E, Fig. 45D).
Description. Complete description in Cokendolpher (2001: 46). Habitus of male and female in Figs 25A-F, scanning electron micrographs of male palp in Figs 46A-F and female genitalia in Fig. 54A.
Distribution. Caves of Fort Hood, Bell County, Texas and surface localities in Blanco, Burnett, Travis and Williamson Counties, Texas (Fig. 57). Etymology. This species is named in honor of Mark Sanders, fellow caver, biologist, and collector of several Tayshaneta species in Texas.
Notes. The only known adult male for T. sandersi is from Whirlpool Cave and is missing most of its appendages and the carapace. Individuals from District Park Cave, Slaughter Creek Cave and Whirlpool Cave are genetically identical suggesting that the species may occur more broadly in the Onion Creek watershed of Barton Springs.
Natural History. Three individuals were found deep in District Park Cave in fine sheet webs under loose rocks. The single male individual was found wandering among loose rocks in Whirlpool Cave.
Distribution. Known from three caves in Travis County, Texas (Fig. 56).    Diagnosis. Tayshaneta valverdae may be separated from all other Tayshaneta species, except T. emeraldae, T. fawcetti, T. grubbsi and T. vidrio by having the male palpal tarsus divided apically (TS, Fig. 31D) and by having a mesoapically positioned ventral sclerite on the palpal bulb (VS, Fig. 49E). Separated from T. emeraldae, T. fawcetti, T. grubbsi and T. vidrio by the unique shape of the embolus with a prominent basal tooth (Fig. 49D).
Description. Complete description in Gertsch (1974: 173). Habitus of male and female in Figs 28A-F, scanning electron micrographs of male palp in Figs 49A-F and female genitalia in Fig. 54C.
Distribution. Known from caves and surface localities in Bandera, Uvalde and Val Verde Counties, Texas (Fig. 60). Etymology. This species name is derived from the Spanish name for the Glass Mountains "Sierra del Vidrio" in West Texas. The name is to be treated as a noun in apposition.
Diagnosis. Tayshaneta vidrio may be separated from all other Tayshaneta species, except T. emeraldae, T. fawcetti, T. grubbsi and T. valverdae by having the male palpal tarsus divided apically (Fig. 31D) and by having a mesoapically positioned ventral sclerite on the palpal bulb (VS, Fig. 50E). Separated from T. emeraldae, T. fawcetti, T. grubbsi and T. valverdae by having an oval embolus that is smooth along its margins and a ventral sclerite with a distinct apical division (VS, Fig. 50D-F).
Female (400ft. Cave). Body length 1.49, carapace 0.63 long, 0.50 wide, length 1.25× width. Pigmentation, setation and eyes same as for male. Legs elongate and thin, femur I 1.57× carapace length, covered in fine setae and with few scattered spines. Atrium oval, length 1.5× width, spermathecae with twisted stalks and elongate heads (Fig. 54D) Etymology. This species is named in honor of Kemble White, fellow caver, geologist and collector of many Tayshneta species in Texas.
Other Diagnosis. Tayshaneta whitei may be separated from all Tayshaneta species, except T. bullis and T. microps, by having a combination of males with an elongate retrolateral tibial spine, more than 0.5× length of the palpal tarsus and lacking a ventral sclerite (Figs 51B, E). Separated from T. bullis and T. microps by the unique shape of the embolus (E, Fig. 51D) and the distinctive retrlolateral sclerite (RS, Figs 51D-E).
Natural History. Several individuals of T. whitei were collected under loose stones near the bases of walls in Lithic Ridge Cave, Bexar County Texas.