Phylogenetic Relationships and Systematic Revision of Central Texas Hemidactyliine Plethodontid Salamanders

HerpetologicalMonographs, 14, 2000, 1-80 ? 2000 by The Herpetologists'League, Inc.

PHYLOGENETIC RELATIONSHIPS AND SYSTEMATIC REVISION OF CENTRAL TEXAS HEMIDACTYLIINE PLETHODONTID SALAMANDERS

PAUL T. CHIPPINDALE,15 ANDREW H. PRICE,2 JOHN J. WIENS,3 AND DAVID M. HILLIS4

'Dept. of Biology, The University of Texas at Arlington, Arlington, TX 76019, U.S.A. 2Texas Parks and Wildlife Dept., 4200 Smith School Road, Austin, TX 78744, U.S.A. 3Carnegie Museum of Natural History, 4400 Forbes Ave., Pittsburgh, PA 15213-4080, U.S.A. 4Section of Integrative Biology and Institute for Cellular and Molecular Biology, The University of Texas at Austin, Austin, TX 78712, U.S.A.

ABSTRACT: Genetic variation and phylogenetic relationships of central Texas nontransforming spring and cave salamanders, genera Eurycea and Typhlomolge (Plethodontidae: Plethodontinae: Hemidactyliini), were examined using 25 allozyme loci and DNA sequence data for a maximum of 356 bp of the mitochondrial cytochrome b gene. Monophyly of the central Texas hemidactyliines is well supported. High levels of divergence occur among many populations and groups of populations, and there clearly are many more species in the group than previously recognized. Many have extremely restricted distributions in isolated islands of aquatic habitat. Several major monophyletic groups were identified that correspond to geographically circumscribed areas of the Edwards Plateau region. The deepest phylogenetic split in the group occurs between populations northeast versus southwest of the Colorado River. Species that have been assigned to the genus Typhlomolge are phylogenetically nested within the central Texas Eurycea; therefore, the genus Typhlomolge is placed in the synonymy of Eurycea. Continued recognition of the species E. latitans, E. nana, E. neotenes, E. pterophila, E. sosorum, E. tridentifera, and E. troglodytes is recommended, but E. neotenes appears to be restricted in range to a small geographic area, and is not widespread in the region as previously thought. The E. latitans and E. troglodytes species complexes are recognized; each consists of spring and cave populations that include those at the type localities of the latter two species, plus other populations to which they appear most closely related. Three new species from northeast of the Colorado River are described.

Key words: Caudata; Plethodontidae; Eurycea; Typhlomolge; Eurycea chisholmensis new species; Eurycea naufragia new species; Eurycea tonkawae new species; Phylogeny; Speciation; Central Texas; Allozymes; Cytochrome b

THE EDWARDS PLATEAU region of cen- other larval morphological features tral Texas is characterized by Cretaceous throughout their lives). Members of this limestones uplifted since at least mid-Ter- group exhibit a wide range of morpholo- tiary times, dissected and eroded to form gies that vary primarily according to numerous springs and caves (Sweet, whether they occupy surface or subterra- 1978a; Potter and Sweet, 1981; Abbot and nean habitats. Typhlomolge rathbuni, the Woodruff, 1986; Woodruff and Abbott, first member of the group to be described, 1986; and Veni, 1994 review the geologic was discovered after the drilling of an ar- history of the area). These habitat islands tesian well at San Marcos, Hays Co. in are inhabited by a variety of endemic 1895 (Stejneger, 1896). This large cave sal- aquatic organisms, many with extremely amander immediately captured the atten- restricted distributions. Predominant tion of the scientific community due to its among the aquatic vertebrate fauna of the bizarre morphology, including tiny non- region are plethodontid salamanders of the functional vestiges of eyes, loss of pigmen- genera Eurycea and Typhlomolge (tribe tation, long slender legs, and broad flat- almost all of which Hemidactyliini), are tened head; this species and its presumed perennibranchiate (i.e., retain gills and sister species T robusta exhibit some- of the most extreme cave-associated mor- 5 Author to whom correspondence should be ad- phologies known among vertebrates (Pot- dressed (email: [email protected]). ter and Sweet, 1981; Sweet, 1986). Rec- 2 HERPETOLOGICALMONOGRAPHS [Ala[No. 14

ognized as a plethodontid by Emerson ciated morphology (Mitchell and Reddell, (1905) and later placed in the tribe Hem- 1965; Mitchell and Smith, 1972; Sweet, idactyliini by Wake (1966), the relation- 1978a, 1984; Potter and Sweet, 1981). ships of Typhlomolge to other hemidacty- Based on biogeographic and geologic con- liines have remained controversial siderations, multiple invasions of subter- (Mitchell and Reddell, 1965; Mitchell and ranean habitat are likely to have occurred Smith, 1972; Potter and Sweet, 1981; (Mitchell and Smith, 1972; Sweet, 1978a, Lombard and Wake, 1986). For the re- 1982, 1984; Potter and Sweet, 1981), but mainder of this paper, we will use the ge- in at least some cases (especially in the ar- nus name Eurycea for these two species, eas inhabited by E. tridentifera, E. latitans, a taxonomic change proposed by Mitchell and perhaps E. troglodytes) there is the and Reddell (1965) and Mitchell and potential for subterranean gene flow Smith (1972). Molecular phylogenetic ev- among cave populations (Sweet, 1978a, idence presented here and by Chippindale 1984). (1995) supports this taxonomic shift. A key issue is whether the central Texas Additional central Texas hemidactyliines hemidactyliines are monophyletic. Wake were not recognized until 1937, when (1966) and Potter and Sweet (1981) sug- Bishop and Wright described Eurycea gested that members of what they consid- neotenes from a spring near Helotes in ered the genus Typhlomolge may be de- Bexar Co. In subsequent decades, several rived from an early Tertiary invasion of the more spring and cave species were de- area whereas Eurycea arrived later (per- scribed: E. nana Bishop 1941 from San haps in the late Miocene or Plio-Pleisto- Marcos Springs in Hays Co., E. latitans cene). According to this scenario, the ex- Smith and Potter 1946 from Cascade Cav- treme cave-associated morphologies of E. erns in Kendall Co., E. pterophila Burger, rathbuni and E. robusta (the two species Smith and Potter 1950 from Fern Bank formerly placed in Typhlomolge) may re- Springs in Hays Co., E. troglodytes Baker flect the long period of time spent under- 1957 from Valdina Farms Sinkhole in Me- ground. The range of (less extreme) cave- dina Co., E. tridentifera Mitchell and Red- associated morphologies seen in other dell 1965 from Honey Creek Cave in Co- central Texas subterranean Eurycea could mal Co., and E. sosorum Chippindale, then represent convergences with these Price and Hillis 1993 from Barton Springs species, varying in degree according to the in Travis Co. The status and relationships length of time spent underground and of these taxa have been problematic; most constrained by the putative common an- recently Sweet (1978a,b, 1984) questioned cestry of members of the Edwards Plateau the status of E. pterophila (which he syn- Eurycea group exclusive of E. rathbuni onymized under E. neotenes) as well as E. and E. robusta. Potter and Sweet (1981) latitans and E. troglodytes, which he con- found that the taxa they considered Ty- sidered hybrids between other species. phlomolge and some central Texas cave With the exception of Bogart's (1967) Eurycea exhibit the same general evolu- chromosomal studies, all inferences of re- tionary trends in head form (broadening lationships among the central Texas hem- and flattening of the skull). However, Pot- idactyliines have been based on morphol- ter and Sweet (1981) demonstrated that ogy, and no truly phylogenetic analysis of the osteological basis of these changes dif- the group has been attempted. Morpho- fers in the two groups. Given this result, logical variation in the group has proven Potter and Sweet (1981) recommended confusing; many surface dwellers from that the genus Typhlomolge continue to be throughout the region appear similar to recognized, a reasonable move given the one another based on external morphology information then available. (e.g., Mitchell and Smith, 1972; Hamilton, Extreme cave-associated morphologies 1973; Sweet, 1978a, 1982; Chippindale et are seen in some other non-Texan hemi- al., 1993), while subterranean dwellers dis- dactyliine plethodontids, including Typhlo- play a spectrum of degrees of cave-asso- triton spelaeus from the Ozark region and 2000] HERPETOLOGICAL MONOGRAPHS 3

Haideotriton wallacei from northern Flor- tyliine genera that are suspected to be ida and Georgia. We included these taxa closely related to or nested within Eury- in the present study; both are thought to cea. We based the choice of outgroup have close affinities to Eurycea and/or members on published morphological what has been considered Typhlomolge work (Wake, 1966; Sweet, 1977b; Lom- (Wake, 1966; Lombard and Wake, 1986). bard and Wake, 1986) and molecular data We also included representatives of all from a study in progress of higher-level species groups currently recognized within hemidactyliine relationships (Chippindale, Eurycea. Use of multiple outgroups is par- unpublished). The outgroup consisted of ticularly important because hemidactyliine E. bislineata (Renfrew Co. Ont.), E. wild- relationships are poorly understood, and erae (Watauga Co. NC), E. quadridigitata the closest relatives of the central Texas (yellow-bellied form, Tyler Co. TX), E. group are uncertain (but see Sweet, 1977b quadridigitata (silver-bellied form, for additional discussion). Charleston Co. SC), E. 1. longicauda (Bal- Given the nature of morphological var- timore Co. MD), E. m. multiplicata (Polk iation in the group and the potential for Co. AR), Haideotriton wallacei (Jackson parallel or convergent morphological evo- Co. FL), and Typhlotriton spelaeus (Stone lution, we used mainly molecular markers Co. MO). to investigate the evolutionary history and relationships of the central Texas hemidac- Allozyme Electrophoresis tyliines. Here we present the results of al- We examined a total of 357 individuals lozyme and mitochondrial DNA (mtDNA) of central Texas Eurycea for allozyme var- sequence studies that we have used to iation, representing 64 populations or taxa. characterize genetic variation and diversi- Early in the study, we homogenized many ty, identify species boundaries, and recon- salamanders whole in a solution of 0.001 struct the phylogenetic history of the M EDTA and 0.010 M Tris, pH 7.5, using group. We also offer a taxonomic revision approximately 1:1 w/v proportions of tissue of the group and descriptions of three new to grinding solution. We subsequently species. found that destruction of entire specimens was unnecessary, and for later allozyme MATERIALS AND METHODS work we used a combination of skeletal Sampling muscle, heart, liver, and gut homogenized Salamanders were collected from approximately 1:1 w/v in the above solu- springs and caves throughout the Edwards tion using an electric tissue grinder. Ho- Plateau region and returned to the labo- mogenates were spun for 3-5 min at about ratory alive, where they were dissected for 12,000 rpm in a microcentrifuge, and 8- appropriate tissues (see below) after an- 10 pl of the resulting supernatant was used aesthesia with MS-222 (Sigma). Specimens to soak filter paper wicks for electropho- were deposited into the Texas Natural His- resis. Electrophoretic methods and stain- tory Collection (Texas Memorial Museum, ing procedures generally followed Murphy University of Texas at Austin). Sampling et al. (1996); electrophoretic conditions localities are listed in Appendix I and used for resolution of different enzyme- mapped in Fig. 1A. The generalized range encoding loci are listed in Table 1. En- of members of the group with reference zyme system names and Enzyme Com- to counties and major cities is shown in mission numbers follow the Fig. IB. recommendations of the Expert Protein Analysis System, Swiss Institute of Bioin- Outgroup Taxa formatics (website: www.expasy.ch/en- For all parsimony analyses (see below), zyme/). We screened 25 loci for which eight outgroup taxa were used to root the banding patterns were readily interpret- trees. We chose taxa that span the range able and activity was strong, and rejected of morphological and genetic divergence many others for which activity levels were in the genus Eurycea, plus other hemidac- highly variable among individuals, resolu- 4 HERPETOLOGICAL MONOGRAPHS [No. 14

100? 990 98? 970 (a)

Texas 310

W_,

30??3

1010 1000 990 980 970

(b)

a _ ~ Gillespie ' |

Val Verde Edwards Kenda< ustin Real Bandera

Me xico Kinney Uvalde Medina 100km

l ______----- ,------_-_-_--_-9-San Antoniol

FIG. l.-A. The Edwards Plateau region of central Texas, with collection localities for salamanders sampled in this study. Filled circles represent surface spring localities and stippled squares represent cave localities. The hatched line represents the southem and eastem margin of the Edwards Plateau (the Balcones Escarp- ment). Allozyme data were obtained for all localities shown, and cytochrome b sequence data were obtained for a subset of these localities described in the text. B. Counties in which cave and spring hemidactyliine plethodontid salamanders are known or suspected to occur; collecting localities are shown in Fig. lA. Major cities are shown for reference and their outlines are greatly simplified. 2000] HERPETOLOGICAL MONOGRAPHS 5

TABLE 1.-Electrophoretic conditions used to resolve products of 25 enzyme-encoding loci in central Texas hemidactyliine plethodontid salamanders, plus outgroup taxa. Refer to Murphy et al. (1996) for composition of buffers 1, 3, 4, and 5, and Chippindale (1989) for composition of buffer 2. 1 = Tris-citrate II; 2 = "Poulik" pH 9.5; 3 = Tris-borate-EDTA pH 8.6; 4 = Tris-citrate-EDTA pH 7.0; and 5 = Tris-borate.

Electrophoretic Enzyme system Locus E.C. number conditions Aconitate hydratase Acon-1 4.2.1.3 1 Adenylate kinase Ak 2.7.4.3 1, 4 Aspartate aminotransferase (cytosolic) sAat 2.6.1.1 5 Creatine kinase Ck-1 2.7.3.2 1 Creatine kinase Ck-2 2.7.3.2 1 Leucyl (cytosol) aminopeptidase Cap 3.4.11.1 1 Glucose-6-phosphate isomerase Gpi 5.3.1.9 2 Glutathione reductase Gr 1.6.4.2 1 Glyceraldehyde-3-phosphate dehydrogenase Gapdh 1.2.1.12 3 Glycerol-3-phosphate dehydrogenase G3pdh 1.1.1.8 3 Isocitrate dehydrogenase (NADP+) Idh-1 1.1.1.42 1 Isocitrate dehydrogenase (NADP+) Idh-2 1.1.1.42 1 L-Lactate dehydrogenase Ldh-A 1.1.1.27 5 L-Lactate dehydrogenase Ldh-B 1.1.1.27 5 Malate dehydrogenase (NAD+) Mdh-1 1.1.1.37 1 Malate dehydrogenase (NAD+) Mdh-2 1.1.1.37 1 Malate dehydrogenase (oxaloacetate decarbox- ylating) (NADP+) Mdhp-1 1.1.1.40 1 Mannose-6-phosphate isomerase Mpi 5.3.1.8 2 Peptidase (cytosol nonspecific) (glycyl-L-leu- cine) Pep-A 3.4.13.8 3 Tripeptide aminopeptidase (L-leucylglycylgly- cine) Pep-B 3.4.11.4 1 XAA-Pro dipeptidase (L-phenylalanyl-L-pro- line) Pep-D 3.4.13.9 1 Phosphoglucomutase Pgm 5.4.2.2 3 6-Phosphogluconate dehydrogenase Pgdh 1.1.1.44 4 Pyruvate kinase Pk 2.7.1.40 4 Superoxide dismutase (cytosolic) sSod 1.15.1.1 2

tion was poor, or mobility patterns were stock (to which we added 12.5 ml Tris/HCl not consistently reproducible. Changes to pH 8.0 and 0.05 g Fast Blue BB): 500 ml the stain recipes described by Murphy et water, 0.37 g alpha-ketoglutaric acid, 1.33 al. (1996) are as follows: (1) Half volumes g L-aspartic acid, 2.50 g polyvinyl pyrroli- were used for all liquid stains except Gr, done, 0.5 g Na2EDTA, and 14.20 g G3pdh, Idh-1, Idh-2, and Pgdh, and third Na2HPO4. or quarter volumes were used for agar overlays; (2) For IDH stains, we substitut- DNA Amplification and Sequencing ed 0.08 g dry isocitric acid for the liquid Sequence data were gathered for single form, and used pH 7.0 Tris/HCl buffer, individuals from 34 populations of central because Idh-1 often showed little activity Texas Eurycea, including representatives at pH 8.0; (3) In stains for Ak, Ck, and Pk, from throughout the geographic range 200 U of hexokinase were used (rather sampled for allozyme variation, popula- than the recommended 20 U); (4) Glu- tions that proved substantially divergent cose-l-phosphate (Sigma G-1259) was based on the allozyme data, and all de- used as the Pgm substrate; (5) To stain for scribed species (whether currently recog- Gr, we used 13.0 ml of Tris/HCl pH 8.0, nized or not) except T. robusta, for which 0.5 ml 0.5 mg/ml DCIP, 0.002 g FAD, 0.01 the subterranean habitat is now inaccessi- g NADH, 0.02 g oxidized glutathione, and ble. We amplified a fragment of roughly 0.5 ml 5 mg/ml MTT; and (6) To stain for 400 bp from the 5' end of the mitochon- sAat, we used 12.5 ml of the following drial cytochrome b (cyt b) gene using the 6 HERPETOLOGICAL MONOGRAPHS [No. 14 polymerase chain reaction (PCR), with the 30 s, 50 C 30 or 45 s, 72 C 1 min (X 34)/ following primers (slightly modified from Step 3: 72 C 5 min (X 1). Moritz et al., 1992). Amplified DNA was purified using the method of Zhen and Swank (1993): 20-25 light strand: primer MVZ 15: GAACTA tul of the PCR sample was electrophoresed ATGGCCCACAC(AT)(AT)TACG(ACGT) a 1.5% AA through nonsubmerged agarose gel and the band of interest was removed strand: CB2H: CCCCTC heavy primer from a well cut in the 15% AGAATGATATTTGTCCTCA gel containing PEG 800 and 2X TAE. Cycle sequencing A map of the cyt b gene and the locations (e.g., Hillis et al., 1996) was usually used, of these primers is provided by Moritz et with 3' 32p end-labelled primers. Reactions al. (1992); CB2H is a truncated version of were electrophoresed through standard their cyt-b2 primer. 6% acrylamide DNA sequencing gels. Gels For most specimens, DNA was extract- were dried into Whatman 3M filter paper ed from tail or liver tissue using the STE without fixation and exposed to Kodak X- method described by Hillis et al. (1996); Omat AR or Biomax MR film for 1-4 days 1-2 1I of the resulting solution was then at room temperature without intensifying diluted in 50 1 of IX TE for PCR. For screens. Details of sequencing procedures Eurycea troglodytes (for which little allo- are given by Chippindale (1995). zyme data could be obtained, and which We sequenced each sample using both may now be extinct) we used supernatant MVZ 15 and CB2H as sequencing prim- from an allozyme sample prepared in the ers, with substantial overlap (typically 100- mid-1970's by S. Sweet and provided by 200 bp) in the middle region of the frag- D. Wake, and applied a modification of the ment for most samples. Some samples Chelex extraction method (Walsh et al., proved difficult to sequence using MVZ 1991). This method also was used for spec- 15, so we also designed an internal se- imens from Bat Well, Greenwood Valley quencing primer: 5'TC(ACT)TTTATTGA Ranch Springs, and Cloud Hollow Spring, (CT)CTCCCAGC 3'. Nucleotide sequenc- and E. tridentifera (Ebert Cave) and E. la- es were unambiguously alignable by eye, titans. The method is as follows: Fifty lI and no insertions or deletions were ob- of allozyme supernatant (E. troglodytes) or served. a tiny fragment of liver or muscle (other To confirm that we were working with specimens) was added to 500 ,ul of an au- the mitochondrial cyt b gene, we purified toclaved 5% solution of Chelex-100 (Bior- mtDNA from a specimen from the Suth- ad) in distilled water. Samples were placed erland Hollow population and compared in a 55 C water bath for about 3 h, shaking its cyt b sequence to that of another indi- occasionally. The samples were then vor- vidual from the same population for which texed briefly, heated to 95 C for 15 min, total cellular DNA was used. The sequenc- vortexed again, and centrifuged briefly to es were identical for all readable base precipitate the Chelex. pairs. PCR was performed using an MJ Re- Cyt b sequences were deposited in search PTC-100 or Ericomp thermal cy- GenBank (accession numbers AF252340- cler. PCR conditions that yielded the most 252380). consistent amplifications were as follows. Assessment Levels Variation and Reactions consisted of 3-6 [l dilute DNA of of (for Chelex extractions, 2 pl of DNA so- Cluster Analyses lution plus 2 ul of a 1 in 50 dilution of An IBM PC version of Swofford and Se- Chelex solution), 0.1 pxMeach primer, 40 lander's (1981) Biosys-1 program was used p,M dNTPs, standard Taq polymerase to calculate measures of allozyme variation buffer (1.5 mM MgCl2), and 1-2 U Taq and genetic distances, and to assess overall polymerase in a total volume of 50 ,tl. divergence based on allozymes. To assess Temperature cycling usually used was: deviations from Hardy-Weinberg equilib- Step 1: 94 C 1.5 min (X 1)/ Step 2: 94 C rium, we used chi-square tests and applied 20001200 HEPEOLGIAHERPETOLOGICAL MONOGRAPHSMOORAH 7

Levene's (1949) correction for small sam- TABLE 2.-Occurrence of alternative nucleotide ple size. We treated almost all localities as combinations in partial mitochondrial cytochrome b sequence, summarized by codon position. Values in separate OTUs for similarity clustering. cells represent total number of observations of com- However, we combined several localities binations of alternative nucleotides (or invariant nu- that were geographically proximal and cleotides) summed across the indicated positions. identical or near-identical in allelic com- Codon and as units to Nucleotide position position frequency single combination 1 2 3 yield a total of 59 "populations". These were: Barrow Hollow + Stillhouse Hollow A 24 21 15 C 16 25 4 Springs; Knight + Cedar Breaks Springs; G 25 18 0 Pedernales Spring 1 + Spring 2; Murphy's T 31 41 4 Spring + Sabinal Canyon Spring; Green- AC 3 0 0 wood Valley Ranch Springs 1 + 2 + 3; AG 4 2 11 AT 1 I 1 Cherry Creek Spring + Cloud Hollow CG 0 0 0 Spring; and E. rathbuni from Ezell's Cave CT 8 6 48 + Rattlesnake Cave + Diversion Spring. GT 0 0 0 No activity was observed for the Gr lo- ACG 1 1 4 cus in of the five E. rathbuni screened ACT 1 0 15 any AGT 1 0 7 for allozyme variation, nor could we detect ACGT 0 0 7 activity in any of the six Greenwood Valley Ranch Springs individuals for Mdh-2. For all analyses, we treated these individuals as homozygous for unique "null" alleles at in a step matrix, implemented using these loci. In doing so we assumed that PAUP* (version 4.0.d055 and 4.0b3; Swof- individuals in these populations possess a ford, 1998, 1999). Berlocher and Swofford unique form of each enzyme, and that the (1997) and Wiens (1995, 1999) provided a differences have a genetic basis. detailed description of this method. Overall divergence based on allozyme For parsimony analysis of sequence data was assessed using UPGMA (e.g., data, we partitioned characters into first, Sneath and Sokal, 1973) with Manhattan second, and third codon positions, fol- (Prevosti) distances (e.g., Wright, 1978); lowed by combinatorial weighting of we also used Nei's (1978) unbiased dis- changes among bases using Wheeler's tance and Rogers' (1972) distance for com- (1990) method with Rodrigo's (1992) cor- parison. To depict relative levels of se- rection for invariant positions. Values in quence divergence among central Texas the resulting three transformation matrices Eurycea, we calculated uncorrected se- were used as the numbers of steps among quence divergences among populations alternative bases in three PAUP* step ma- and taxa and performed neighbor-joining trices (corresponding to first, second, and cluster analysis using PAUP* v. 4.0b2 third codon positions). Frequencies of al- (Swofford, 1999). ternative nucleotide combinations are shown in Table 2 and transformation ma- Parsimony Analyses trices in Table 3. This approach allowed For parsimony analyses of allozyme incorporation of information on the fre- data, we employed a method of frequency- quency and direction of different kinds of based coding that involved treatment of changes, based on observed patterns in the each different observed array of allele fre- data set. Each step matrix was rendered quencies for a given locus (character) as a symmetric by averaging across the diago- unique state. Manhattan distances (D) nal to minimize effects of sampling error among states were calculated using Biosys- with respect to estimates of change fre- 1 (Swofford and Selander, 1981). These quency (see Swofford et al., 1996). For distances were then converted to whole parsimony analyses using combined allo- numbers (we rounded to two digits) and zyme and sequence data, we used the al- used as the numbers of steps among states lozyme characters, coded as described 8 HERPETOLOGICAL MONOGRAPHS [No. 14

TABLE 3.-Transformation matrices calculated from the observed nucleotide changes, scaled to a maximum of 100 steps for the rarest changes (see text for details).

Codon position 1 Codon position 2 Codon position 3 A c C G T A C G T A C G T A - 50 46 70 A - 100 58 100 A - 64 29 65 C 78 - 99 39 C 100 - 100 54 C 100 - 100 17 G 70 100 - 100 G 62 100 - i G 80 100 - 100 T 100 32 100 - T 89 44 i - T 89 13 84 above, plus the sequence data partitioned frequency. Membership of these groups is by codon position with the application of shown in Figs. 2 and 3, and localities are combinatorial weights. For these analyses, listed in Appendix I. Although most of we scaled the DNA step matrix values to these groups likely represent real evolu- 100 for equivalence with changes at each tionary units (species or monophyletic allozyme locus (because use of the Man- groups of species), several are somewhat hattan distance approach for allozymes al- arbitrary. Especially problematic assem- lowed up to 100 "steps", or distance units, blages of populations are the Carson Cave between states, i.e., alternative allele fre- group and the E. latitans complex. These quency arrays). issues, and the potential impact of the We conducted heuristic parsimony grouping approach on the analyses, are ad- searches using three treatments of the dressed below (see Discussion). data: 1) allozyme data only; 2) sequence data only; and 3) allozyme plus sequence Osteological Examination, Measurements, data. Each involved 50 random-taxon-ad- and Morphometric Analyses dition replicates to reduce the chances of Osteological variation among the three recovering only trees within a particular new species of Eurycea that we describe "tree island" (Maddison, 1991). We as- here was examined following clearing-and- sessed phylogenetic confidence in partic- staining of specimens using the method of ular nodes through use of nonparametric Dingerkus and Uhler (1977). Specimens bootstrapping (Felsenstein, 1985), imple- examined are listed below. For morpho- mented using PAUP*. One hundred heu- metric analyses, we used the following ristic bootstrap pseudoreplicates were con- measurements of external morphology: ducted for each data treatment; AG (axilla-groin length), ALL (anterior bootstrapping for data treatment 3 in- limb length, from anterior insertion to tip volved 10 random-taxon-addition se- of third toe), ED (eye diameter, measured quence replicates per bootstrap pseudo- as anterior-posterior diameter of the exter- replicate, but this was not feasible for data nally visible dark disc of the eye), HLA treatments 1 and 2 so only a single taxon- (head length A, distance from tip of snout addition-sequence was used within each to center of gular fold), HLB (head length pseudoreplicate for these analyses. B, distance from posterior margin of eye to posteriormost gill insertion), HLC Population Groups (head length C, distance from tip of snout The large number of populations in- to posteriormost gill insertion), HLL (hind cluded in the allozyme portion of this limb length, from groin to tip of third toe), study made it impractical to include each HW (head width immediately posterior to as a separate unit in parsimony analyses, jaw articulation), IOD (interocular dis- and sequencing of representatives from all tance), SL (standard length, distance from populations was impractical. To reduce the tip of snout to posterior margin of vent), number of working units, we constructed and TL (tail length, from posterior margin 24 population groups based on consider- of vent to tip). ED was measured at X64 ation of geographic location and proximity magnification with backlighting using an plus similarity in allelic composition and ocular micrometer, AG, SL, and TL were 2000] HERPETOLOGICAL MONOGRAPHS 9

Manhattandistance measured with digital calipers, and the re- 0.50 0.40 0.30 0.20 0.10 0.00 I . I . I I maining measures were taken at X7.5 a with an ocular micrometer. Cb LakeGeorgetown magnification Bat Well SaladoSprings Specimens measured were the following rb (all are from the Texas Natural History d - Collection, University of Texas at Austin). Jollyville Plateau North Eurycea tonkawae, Jollyville Plateau/

~bButtercup Creek caves Brushy Creek drainage: Balcones Park _ -- RoundRock Spring Kretschmarr Cave Spring (TNHC 55132-134); Barrow Hol- d ButtercupCr. caves Testudo Tube low Spring (TNHC 50933, 50938); Brushy Creek a Spring (TNHC 50988, 54225-226, Creek 55141- r d E. latitans Canyon Spring (TNHC eo complex 144); Krienke Spring (TNHC 53466, PedernalesSprings 53469, 53472, 53475-476); Schlumberger n ~.J^o Southeast _ d E pterophila Spring (TNHC 50985-986); Stillhouse ComalSprings Hollow (TNHC 50947, 50950, E. sosorom Spring _-_--.__?-b ' E neotenes 50952, 50956-957); Wheless Tract Spring b E. tridentifera 55155-157. c (TNHC 55150, 55152, Eury-

CampMystic cea naufragia, Georgetown area: Avant's Spring (TNHC 51027-029); Buford Hol- e CarsonCave group low Spring (TNHC 51008, 51013-014, rC L- 9 Southwest _ t GreenwoodValley Ranch Spr. 58860-861; Cedar Breaks Hiking Trail 176Spring SabinalSprings _L..X---- ' TuckerHollow Cave Spring (TNHC 51000-002). Eurycea E. nana San Marcos chisholmensis, Salado Springs: TNHC E. rathbuni 51139-142, 51143 (TL not measured), FIG. 2.-Similarity (UPGMA) clustering of central 51144 (HLA not measured), 51145-146, Texas hemidactyliine plethodontid salamanders based 58859. Summaries of measurements are Manhattan distances. on allozyme Major geographic in Table 4. regions are indicated. Within groups, populations are given as follows. Lake Georgetown group: a = Avant's We assessed the contributions of differ- Spring, b = Knight and Cedar Breaks Springs, c = ent variables to total morphometric vari- = Buford Hollow Spring; Jollyville Plateau group: a ance among individuals using principal b = Barrow Hollow and Still- Balcones Park Spring, (PCA) house Hollow Springs, c = Horsethief Hollow Spring, components analysis implemented d = New Bull Creek Spring, e = Schlumberger via STATISTICA v. 4.5 (1993, Statsoft, Spring, f = Wheless Spring, g = Hanks' Spring, h = Tulsa, OK); each variable was subjected to = Canyon Creek Spring, i Canyon Vista Spring; But- a log10 transformation prior to analysis. We Creek Caves a = Ilex b = But- tercup group: Cave, then used a one-way analysis of variance tercup Creek Cave, c = T.W.A.S.A. Cave, d Tree- house Cave; E. latitans complex: a = Bear Creek of individual scores on factor 2 to assess Spring, b = Cibolo Creek Spring, c = Honey Creek differences among the three new species Cave Spring, d = Kneedeep Cave Spring, e = Cherry we describe here, and performed a post Creek and Cloud Hollow Springs, f = Less Ranch hoc Scheffe test (Sokal and Rohlf, 1981) = Rebecca Creek h = Pfeiffer's Wa- Spring, g Spring, for of differences between ter Cave; E. pterophila: a = Boardhouse Springs, b significance = Peavey's Springs, c = Zercher Spring, d = Fern species. Bank Spring, e = T Cave, f = Grapevine Cave; E. neotenes: a = Helotes Creek Spring, b = Leon RESULTS Springs, c = Mueller's Spring; E. tridentifera: a = Honey Creek Cave, b = Ebert Cave, c = Badweather Intrapopulation Allozyme Variation Cave a = Carson b = West Pit; Carson group: Cave, ex- Nueces Spring, c = Sutherland Hollow Spring, d = Twenty-two of the 25 allozyme loci Smith's Spring, e = Robinson Creek Spring, f = WB amined displayed polymorphism among Spring, g = Trough Spring, h = Fessenden Springs. and/or within populations of central Texas hemidactyliines (Appendix II). However, levels of intrapopulation genetic variation generally were low. Direct-count heterozygosity (H) ranged from 0% in several 10 HERPETOLOGICAL MONOGRAPHS [No. 14

__ _ __.__ __ FIG. 3.-Population groups to which central Texas Eurycea were assigned for DNA sampling and parsimony analysis; these assignments were based largely on initial allozyme studies, the results of which are summarized in Fig. 2. The region outlined corresponds to the counties in which these salamanders are known to occur, detailed in Fig. 1. Filled circles represent surface spring localities and stippled squares represent cave localities. populations to 12.0% in the T.W.A.S.A. 2.92% (S.E. = 0.042). The percentage of Cave population (represented by a single polymorphic loci (P) ranged from 0% in specimen); the next highest observed H several populations to 32% in E. sosorum was 10.8% in Eurycea tridentifera from with a mean of 9.69% (SE = 0.128), and Badweather Pit (five specimens exam- the average number of alleles per locus (A) ined). Mean H across all populations was ranged from 1.0 to 1.2 with a mean of 1.09 (SE = 0.001). Sixteen significant deviations from Hardy-Weinberg equilibrium TABLE 4.-Summary statistics for measurements of were detected (chi-square test, P < 0.05), north central Texas species of Eurycea used in mor- of a total of 1600 X loci ex- phometric analyses. Sample sizes are in parentheses populations following species names; values given are mean (in amined. There were 145 cases in which a mm) ? 1 standard deviation. locus exhibited intrapopulation variation. Using this value to conduct sequential Species Bonferroni tests (Rice, 1989), at most four E. tonkawae E. naufraeia E. chisholmensis Variable (N = 29) (N = 13 (N = 7) deviations were significant. SL 30.49 (3.29) 28.97 (4.49) 32.93 (2.53) Variation AG 16.44 (2.09) 15.77 (2.87) 18.17 (1.84) Interpopulation Allozyme TL 23.67 (3.94) 21.04 (4.84) 24.58 (3.54) The UPGMA tree constructed from HLA 6.26 (0.76) 5.90 (0.71) 6.48 (0.67) Manhattan distances (Fig. 2) reveals a high HLB 5.13 (0.52) 5.17 (0.54) 5.79 (0.68) of differentiation HLC 8.08 (0.78) 8.16 (0.84) 8.90 (0.99) degree genetic among HW 4.67 (0.49) 4.50 (0.52) 5.01 (0.76) some populations and groups of popula- IOD 1.47 (0.26) 1.47 (0.19) 1.73 (0.27) tions. In particular, members of the HLL 6.30 (0.76) 6.39 (0.59) 7.20 (0.67) "northern" group (populations from north- ALL 5.60 (0.71) 5.33 (0.47) 6.25 (0.55) east of the Colorado River in Travis, Wil- ED 1.31 (0.10) 1.23 (0.16) 0.98 (0.14) liamson, and Bell counties) are extremely 2000]200IEPTLGCA HERPETOLOGICALMONOGRAPHS OORPS1 11 divergent from all other central Texas Eu- The neighbor-joining tree based on uncor- rycea. Average Manhattan D between the rected percent sequence divergence is northern and other populations exceed shown in Fig. 4. In most respects, the ma- 0.45, which correspond here to average jor patterns of divergence seen in the al- Nei's (1978) D over 0.65 and Rogers's lozyme-based cluster analysis also occur in (1972) D over 0.45. These distances reflect the DNA-based cluster analysis. The numerous differences in allelic composi- northern populations are strongly differ- tion from all other central Texas Eurycea entiated from all others, exhibiting over examined; most of these differences are 14% uncorrected sequence divergence on fully or nearly fixed, or mutually exclusive. average (Fig. 4). As in the allozyme-based South of the Colorado River, E. rathbuni cluster analysis, Eurycea rathbuni appears from the San Marcos region appears as the as next most distinct, with an average se- next most divergent member of the group quence divergence of approximately 9% with a Manhattan D exceeding 0.40 from from other non-northern populations. The all other taxa examined, and E. nana (also same division between southeastern and from the San Marcos area) is next most southwestern populations (exclusive of E. divergent (Manhattan D over 0.30). Of the rathbuni) occurs as in the allozyme-based remaining populations, there is a division cluster analysis, with average sequence di- between a "southeastern group" (all pop- vergences over 7%. A key difference be- ulations east of extreme eastern Kerr Co., tween the allozyme tree and that based on corresponding primarily to the southeast- sequence data is that in the DNA-based ern drainages of the Edwards Plateau) and tree, E. nana appears with the southeast- a "southwestern group" (corresponding ern group, whereas based on allozymes it primarily to southwestern drainages of the is strongly differentiated and appears out- plateau). While E. nana and E. rathbuni side all other non-northern populations ex- each possess unique alleles at several loci cept E. rathbuni. Most allozyme differenc- (Appendix II), differentiation among mem- es represent autapomorphies for E. nana bers of the southeastern and southwestern and thus are not informative in a parsi- groups primarily is based on allele frequen- mony context (see below). The other ma- cy variation. UPGMA trees constructed us- jor difference between the allozyme clus- ing Nei's (1978) unbiased distance and ter analysis tree and that based on DNA is Rogers's (1972) distance had topologies that the DNA tree reflects the extremely nearly identical to that of the Manhattan D low levels of sequence variation among tree except at the very lowest clustering lev- most taxa in the southeastern region. In els, and are not shown here. Major groups contrast, substantial allozyme variation is based on similarity analyses of allozyme present in the southeastern group (Fig. 2 data are shown in Fig. 3. and Appendix II). Levels and Patterns of Sequence Geographic Patterns of Genetic Variation Differentiation Across the maximum 356 bp of cyt b Cluster analyses of both the allozyme sequenced, 133 sites (37.5%) were variable and sequence data identify several major including outgroup taxa; within the in- groups of populations, most of which cor- group 101 sites (28.5%) varied (aligned se- respond to geographically circumscribed quences are shown in Appendix III). Of regions of the Edwards Plateau area: a the 118 codons examined, 15 (12.7%) ex- northern group, a southeastern group, a hibited amino acid variation considering southwestern group, Eurycea rathbuni all taxa; excluding the outgroup, 13 from San Marcos, and E. nana from San (11.0%) were variable. Base composition Marcos (E. nana is strongly differentiated of the light strand (G = 0.16, A = 0.26, T based on allozymes only). The latter four, = 0.36, C = 0.23) was very similar to that all from southwest of the Colorado River, reported for plethodontid salamanders of will be referred to collectively here as the the genus Ensatina by Moritz et al. (1992). "southern group". The named taxa E. neo- 12 HERPETOLOGICAL MONOGRAPHS LNo.[No. 14

E. rathbuni San Marcos E. nana E. sosorum Comal Springs E. pterophila _E. neotenes Southeast E. latitanscomplex E. tridentifera Pedernales Springs

E. troglodytes(topotypical) '-Sabinal group rI-|-- TuckerHollow Cave GreenwoodValley Ranch Spr. Southwest [-\1 176 Spring Camp MysticSpring Carson Cave group

rTestudo Tube ButtercupCreek caves Bat Well r-Jollyville Plateau group _- I* Round Rock spring North Salado Springs KretschmarrCave Lake Georgetowngroup - 2 substitutions FIG. 4.-Neighbor-joining tree for central Texas hemidactyliine plethodontid salamanders based on uncorrected cytochrome b sequence divergence. Rooting is based on positions of outgroup taxa (included in analysis but not shown). Species and informal groups correspond to those listed in text and shown in Figs. 2 and 3. Major geographic regions are indicated. tenes, E. latitans, E. pterophila, E. soso- latitans examined for allozyme variation rum, and E. tridentifera all cluster within displayed a unique Mdhp-1 allele that ap- the southeastern region based on allo- pears fixed or at very high frequency in the zymes and DNA. The only remaining three populations of E. tridentifera that we named taxon (exclusive of E. rathbuni and examined (Appendix II). As described E. nana) is E. troglodytes, which clusters above, analyses of sequence data place to- with members of the southwestern group potypical E. troglodytes with populations based on sequence data (allozyme data in the southwestern region and do not sup- were not available for this population). port a close relationship between this pop- Cave populations from the northern, ulation and the southeastern E. tridenti- southeastern, and southwestern regions all fera (but see Discussion for caveats cluster with spring populations in the same regarding use of mitochondrial sequence regions based on both allozymes and data). DNA, regardless of their degree of morphological divergence. The only exception Parsimony Analyses is E. rathbuni, which is strongly differen- Results of parsimony analyses are shown tiated both morphologically and based on in Figs. 5-7. Analysis of allozyme data allozymes and DNA from all other mem- alone yielded three equally parsimonious bers of the southern assemblage. trees of length 138.52 (we have divided We found no evidence that the cave- the number of steps in each tree by 100, dwelling taxa E. latitans and E. troglodytes so that the tree lengths reflect the number are of hybrid origin. None of the five E. of character changes, i.e., a fixed differ- 2000] HERPETOLOGICAL MONOGRAPHS 13

E. nana I San Marrsn%.f% E. rathbuni Testudo Tube 71 50 -Jollyville Plateau group Round Rock Spring E. tonkawae fi- ButtercupCreek caves KretschmarrCave North 91 p_ 4-- Lake Georgetowngroup E. naufragia Bat Well Salado Springs E. chisholmensis 52 GreenwoodValley Ranch Springs 176 Spring Southwest TuckerHollow Cave Sabinal Springs (E. troglodytes Carson Cave group complex) Camp MysticSpring E. sosorum L E. latitanscomplex -{ -E. pterophila E. neotenes Southeast Pedernales Springs Comal Springs E. tridentifera Haideotritonwallacei 71 - E. I. longicauda E. bislineata E. quadridigitataSC 56 - E. m. multiplicata Outgroup Typhlotritonspelaeus E. quadridigitataTX E. wilderae

FIG. 5.-Strict consensus of the three equally parsimonious trees for central Texas hemidactyliine plethodontid salamanders that resulted from a heuristic search (with 50 random taxon addition sequence replicates), using allozyme data alone with Manhattan distance step matrices. Bootstrap support (100 pseudoreplicates) is shown for nodes with 50% or greater support based on a separate nonparametric bootstrap analysis.

ence at one locus would result in one step All parsimony analyses support mono- between two taxa). The strict consensus of phyly of the northern group of popula- these trees is shown in Fig. 5. Analysis of tions. The weakest support occurs in the cyt b sequence data alone (with combina- allozymes-only trees, with a bootstrap val- torial weighting and partitioningby codon ue of 71% (Fig. 5). However, support for position) yielded 90 equally parsimonious all other nodes except ingroup monophyly trees of length 137.38; the strict consensus is weak using the allozyme data alone. is shown in Fig. 6. The single most-parsi- Members of the northern group not only monious tree (length 285.24) that resulted are characterized by diagnostic alleles at from combined analysis of the allozyme several allozyme loci and unique sequence and sequence data is shown in Fig. 7. substitutions, but also by unique amino Results of all three parsimony analyses acid substitutions at four codons. Unam- shown here are consistent with monophyly biguous and potential molecular synapo- of the central Texas hemidactyliines. The morphies for this and the other groups lowest bootstrapvalue for the node uniting recognized here are listed in detail by the Texas hemidactyliines is 59% using se- Chippindale (1995) and summarized in quence data alone; using allozymes alone the species accounts given in the Discus- the bootstrap value for this node is 91%, sion. and the combined analysis yields a value The southern group appears as mono- of 98%. phyletic in both analyses that include the 14 HERPETOLOGICAL MONOGRAPHS No.[No. 14

L L a k Lake Georgetowngroup E. naufragia 100 Bat Well E. naufragia Salado Springs E. chisholmensis 99 E-Jollyville Plateau group RoundRock Springs North 58-e Testudo Tube E. tonkawae 5X7J '- ButtercupCreek caves KretschmarrCave E. nana San Marcos E. latitanscomplex 99 74 E. pterophila 55 E. neotenes 72 - E. tridentifera 7255 Comal Springs Southeast E. sosorum Pedernales g,64 Springs GreenwoodValley Ranch Springs Carson Cave group 8 60 C E. troglodytes(topotypical) Southwest ~81 l/-Sabinal_ group ?Tucker Hollow (E. troglodytes -679 Cave Camp MysticSpring complex) 176 Spring E. rathbuniSan Marcos

. Typhlotritonspelaeus Haideotritonwallacei E. quadridigitataTX - 86, E m. multiplicata 58 Outgroup r ) -E. quadridigitataSC E /. longicauda I91 E E. bislineata E. wilderae

FIG. 6.-Strict consensus of the 90 equally parsimonious trees for central Texas hemidactyliine plethodontid salamanders that resulted from a heuristic search (with 50 random taxon addition sequence replicates), using cytochrome b data alone with partitioning by codon position and combinatorial weighting. Bootstrap support (100 pseudoreplicates) is shown for nodes with 50% or greater support based on a separate nonparametric bootstrap analysis. sequence data, but not in the tree based buni (bootstrap value 81%), and combi- on allozyme data alone (Fig. 5). This anal- nation of the allozyme and sequence data ysis places Eurycea rathbuni as the sister provides even stronger support (90%). Eu- lineage to the northern group, with mem- rycea rathbuni appears as the sister line- bers of the southwestern group appearing age to all other southern taxa in both anal- successively sister to this grouping (ren- yses that include the sequence data. dering the southwestern group nonmono- The southeastern populations plus Eu- phyletic). A monophyletic southeastern rycea nana appear as a well-supported group is sister to this grouping, and E. monophyletic group in the analyses that nana appears as sister to all other central include the sequence data (bootstrap val- Texas hemidactyliines. Support for these ues for this node are 99% based on se- relationships is weak (bootstrap values be- quence alone and 97% based on the com- low 50% for almost all relevant nodes), in- bined data). Using allozymes alone, the dicating little allozyme character support southeastern clade is present but the po- for any relationships except monophyly of sition of E. nana is problematic, as ex- all central Texas hemidactyliines and plained above. Similarly, relationships of monophyly of the northern group. Use of the southwestern populations are poorly the sequence data alone (Fig. 6) provides supported (and the group is not monophy- relatively strong support for monophyly of letic) based on allozymes alone, but they the southern group inclusive of E. rath- appear as a well-supported monophyletic 2000] HERPETOLOGICAL MONOGRAPHS 15

Testui do Tube 5 5 ( ButtercupCreek caves Kretschmlarr Cave E. tonkawae 'Jollyville PIateau group North RoundRoc ck Spring E. chisholmenssis (SaladoSprings) - E. naufragia(p)rovisional, Bat Well) *E.naufragia (I Lake Georgetown Springs)

E.rrathbuni San Marcos E. sosorum ComalSprings PedernalesSprings E. pterophila Southeast E. neotenes E tridentifera L E latitanscomplex E. nana San Marcos

a E. troglodytes(topotypical) 1SabinalSprings -TuckerHollow Cave Southwest - GreenwoodV. RanchSpr. - (E. troglodytes Highway176 Spring complex) >nCave group p MysticSpring

Outgroup

5 changes

FIG. 7.-The single most parsimonious tree that resulted from a heuristic search (with 50 random taxon addition sequence replicates) of combined allozyme and cytochrome b sequence data for central Texas hemidactyliine plethodontid salamanders (allozymes were coded using Manhattan distance step matrices and cytochrome b sequence data with partitioning by codon position and combinatorial weighting). Relative branch lengths are shown. Bootstrap support (100 pseudoreplicates) is shown above nodes with 50% or greater support based on a separate nonparametric bootstrap analysis.

group in the DNA-only (bootstrap 79%) b sequence data; Fig. 7) divides them into and combined (85%) analyses. Both anal- a monophyletic "northern" clade from yses that include the sequence data sup- northeast of the Colorado River, and a port a sister-group relationship between "southern" clade from southwest of the the southeastern and southwestern groups Colorado River. The southern clade is (bootstrap 64% for DNA only and 62% for composed of Eurycea rathbuni from San combined data). Marcos, a "southeastern" subgroup corresponding primarily to the eastern drain- Biogeographic and Taxonomic Patterns ages on the southern portion of the Ed- Based on Parsimony Analyses wards Plateau exclusive of the San Marcos Our preferred phylogenetic hypothesis region (E. nana from San Marcos appears for the central Texas hemidactyliines (that to be the sister lineage to this group), and based on the combined allozyme and cyt a "southwestern" subgroup generally cor- 16 HERPETOLOGICAL MONOGRAPHS [No. 14

TABLE 5.-Factor loadings and coefficients resulting from principal components analysis of external mea- surement variables for northern central Texas Eurycea. Two factors were extracted; the first explains 65.3% of the total variance, and the second explains 12.0%. Values shown are for unrotated axes, and have been rounded to three decimal places.

Factor 1 Factor 2 Variable Loading Coefficient Loading Coefficient SL -0.920 -0.128 -0.209 -0.158 AG -0.862 -0.120 -0.298 -0.226 TL -0.741 -0.103 -0.371 -0.280 HLA -0.885 -0.123 -0.026 -0.020 HLB -0.855 -0.119 0.275 0.208 HLC -0.887 -0.124 0.184 0.139 HW -0.857 -0.119 0.076 0.057 IOD -0.774 -0.108 0.351 0.265 HLL -0.808 -0.113 0.107 0.081 ALL -0.857 -0.119 0.037 0.028 ED -0.118 -0.016 -0.895 -0.677 responding to the western drainages in the based primarily on body size. The second southern portion of the Edwards Plateau factor explains another 12.0% of the total region. This pattern of phylogenetic rela- variance, and appears to be based primar- tionships is consistent in most ways with ily on eye diameter. A bivariate plot of in- the pattern of similarities revealed by the dividual factor scores (not shown) dem- cluster analyses of both allozyme and se- onstrated wide overlap between two of the quence data (except for the placement of three new species that we describe here E. nana based on the allozyme data). (those from the Jollyville Plateau and Georgetown areas), whereas most individ- Relationships of Outgroup Taxa uals from the Salado population occupied Among the outgroup taxa used, there a distinct region of factor space. For sim- was relatively little consistency in inferred plicity, we show a bivariate plot of ED vs. relationships among analyses. With the ex- AG in the species description for Eurycea ception of the allozymes-only analyses, chisholmensis (below). This includes two there generally was strong support for a additional animals (TNHC 51143 and that were not included in the factor sister-group relationship between Eurycea 51144) bislineata and E. wilderae, consistent with analysis because some measurements could not be taken. previous inferences that these taxa are part of a distinct within ANOVA revealed significant differences group Eurycea (e.g., the three new that we de- 1987). among species Jacobs, Eurycea quadridigitata scribe based on individual scores on from Texas and E. from here, quadridigitata factor 2 df = P < Pair- South Carolina never clustered (F-test, 46, 0.0001). together, wise between a that this nominal taxon be comparisons species, using suggesting may post-hoc Scheffe test, revealed that Eury- composed of at least two species. Depend- cea chisholmensis from Salado dif- on the the cave- Springs ing analysis, monotypic fered from E. tonkawae Jol- Haideotriton and significantly dwelling genera Typhlo- lyville Plateau/Brushy Creek drainage (P < triton occur in various in the tree, places 0.0001) and E. naufragia Georgetown (P but were nearly always imbedded within < 0.05), whereas the latter two species the genus Eurycea. were not significantly differentiated from each other based on Morphometric Analyses morphometric proportions alone. PCA extracted two factors. Loadings and coefficients for each variable on each DISCUSSION factor are listed in Table 5. The first ac- Intrapopulation Allozyme Variation counts for 65.3% of the total variance Shaffer and Breden (1989) found that among individuals, and appears to be nontransforming salamanders usually ex- 2000] HERPETOLOGICAL MONOGRAPHS 17

hibit lower levels of genetic variation (as Interpopulation Sequence Differentiation measured by allozyme markers) than do and Variability of Allozymes Versus transforming species. Our results for most Mitochondrial DNA populations of central Texas hemidactyli- The levels of nucleotide variation seen ines are consistent with this observation, for cyt b here are very similar to those re- given the mean heterozygosity (H) across ported by Moritz et al. (1992) for pletho- all populations of 2.9% (however, there are dontine plethodontids of the genus Ensa- several notable exceptions, particularly tina plus the outgroups Aneides and among cave-dwellers; see Appendix II and Plethodon. Including outgroups, we found below). Shaffer and Breden (1989) sug- 37.5% of the 356 positions surveyed vari- gested that this pattern may be due in part able (28.5% for ingroup members only), to the ephemeral nature of the habitats of while Moritz et al. (1992) reported 37.0% many nontransforming species: perenni- of positions variable for the 681 bp they branchiate salamanders often inhabit bod- sequenced across all Ensatina, Plethodon, ies of water in areas that are relatively hot and Aneides they examined. Moritz et al. and arid, and periodic drying of these (1992) found almost twice the amino acid aquatic habitats may result in genetic bot- sequence variation in their study as did we tlenecks. Our field observations (see Chip- (22.0% of amino acids variable including pindale, 1995) and those of Hamilton ingroup and outgroups, compared to (1973) and Sweet (1977b) support this hy- 12.7% here with outgroups and 11.0% pothesis for at least some populations of without). The reason for the apparently spring-dwelling central Texas hemidactyli- greater level of conservation in hemidac- ines. tyliine cyt b amino acid sequences is un- Sweet (1978a) suggested that many known. populations of cave hemidactyliines in cen- On a broad scale, the patterns of geo- tral Texas may have originated from graphic variation exhibited by both allo- spring-dwellers that followed the water zymes and cyt b sequences are similar. much more variation column underground when it dropped be- However, was revealed than b cause of drought, or when erosion led to by allozymes cyt sequences in the southeastern where most stream capture. Some of the highest levels region, and taxa exhibited identical or of heterozygosity observed were among populations near-identical This observation cave dwellers from subterranean aquatic sequences. demonstrates that even a mitochondrial systems that likely are relatively extensive gene (cyt b) generally regarded as fairly (especially Eurycea tridentifera from caves fast-evolving (e.g., Graybeal, 1993; Meyer, of the Cibolo Sinkhole Plain in Comal, 1994; Johns and Avise, 1998) may fail at Kendall, and Bexar counties); perhaps some level to exhibit useful variation, these numbers of in- systems support large while nuclear allozyme markers continue dividuals and/or are buffered the against to be informative. Although some studies drastic in water to changes availability of genetic variation in salamanders have which surface populations are subject. found the opposite (i.e., substantial mito- However, the relationship between hetero- chondrial variation but low levels of allo- zygosity and habitat type is far from clear- zyme variation; see Routman, 1993a re- cut: a few spring populations exhibit rela- garding Cryptobranchus and Routman, tively high levels of heterozygosity, and 1993b regarding Ambystoma), McKnight even within a given cave system, estimates et al. (1991) described a situation similar of heterozyosity can vary widely from site to that in the Texas Eurycea for the pleth- to site (e.g., for E. tridentifera from Honey odontid Phaeognathus hubrichti. Based on Creek Cave vs. Badweather Pit, and for a finding of substantial allozyme variation the various populations of E. tonkawae and low mitochondrial restriction site var- from the Buttercup Creek Cave system in iation, McKnight et al. (1991) suggested Williamson Co.). that a bottleneck may have been extreme 18 HERPETOLOGICALMONOGRAPHS [No. 14 enough to reduce effective mitochondrial thought by many to be one of the oldest population sizes to the point where diver- features of the Edwards Plateau (Abbott, sity was drastically diminished and partic- 1975; Sweet, 1978a; Veni and Associates, ular haplotypes became fixed, whereas ef- 1992), and probably has cut down through fective population sizes for nuclear genes the elevated limestones of the plateau remained large enough to maintain allo- throughout its existence, dividing the Ed- zyme diversity (see also Birky et al., 1983, wards (Balcones Fault Zone) Aquifer into 1989; Wade et al., 1994). Perhaps the an- two major sections with little or no hydro- cestor of the southeastern Edwards Pla- logic connection (Slade et al., 1986). Thus teau Eurycea experienced a similar bottle- it is not surprising that salamanders from neck. either side of the river are strongly differentiated; this pattern of vicariant isolation Geographic Patterns of Genetic by waterway also has been observed in Differentiation other groups of salamanders (e.g., Good Cluster analyses of both the allozyme and Wake, 1992). and sequence data (Figs. 2, 4) reflect the The similarity-based trees also illustrate high degree of genetic subdivision present the relatively high levels of genetic diver- in the group, particularly with respect to gence of the taxa from the San Marcos the groups of populations and taxa from Pool of the Edwards Aquifer in Hays Co. north versus south of the Colorado River. (southern region). Eurycea rathbuni and The large number of sequence substitu- E. nana clearly are distinct species long tions between the northern and southern isolated from gene flow with the other groups, coupled with cyt b amino acid sub- populations examined. This result is con- stitutions and numerous fixed or near-fixed sistent with the occurrence of numerous allozyme differences, indicates that these other endemic species of aquatic verte- groups are completely isolated from each brates, invertebrates, and plants at San other and probably have been for millions Marcos (e.g., Holsinger and Longley, of years. This conclusion is reinforced by 1980), and likely is related to the high de- the results of flow cytometric studies of gree of isolation of the San Marcos Pool nuclear genome size: C-values (nuclear of the Edwards Aquifer from the remain- DNA mass) are 12-13% higher on average ing southern portions of the aquifer (e.g., for members of the northern group than Potter and Sweet, 1981). the southern group, and there is no over- The subdivision of members of the lap in C-value distributions for the two southern group exclusive of the San Mar- groups [Chippindale and Lowcock, un- cos taxa (Eurycea rathbuni, plus E. nana published; Licht and Lowcock, 1991 (note for allozymes) into southeastern and that Licht and Lowcock erroneously listed southwestern components corresponds the mean C-value for members of the roughly to the eastern versus western northern group as 25.8; the correct value drainages of the southern plateau region. is 28.5)]. These drainages are of more recent origin Although Sweet (1978a) recognized the and less deeply incised than the Colorado potential of the Colorado River as a barrier River (e.g., see Sweet, 1978a, 1982; Veni, to gene flow, he (1978a, 1982) identified 1994). The southeastern and southwestern the few northern populations known then groups are not as strongly differentiated as Eurycea neotenes due to the high de- from one another as are the northern and gree of morphological similarity between southern groups, or the San Marcos taxa most surface-dwelling populations from compared to all others: most of the allo- the northern and southern areas. The type zyme-based differentiation constitutes al- locality for E. neotenes is in Bexar Co., in lele frequency variation rather than fixed the southern region. Based on the molec- differences. However, lack of current gene ular data, however, the northern group flow between the regions is especially ap- consists of strongly differentiated, long- parent based on the mitochondrial se- isolated species. The Colorado River is quence differences that separate all south- 2000] HERPETOLOGICALMONOGRAPHS 19 eastern populations from all southwestern formation on a locus by locus (i.e., individ- ones. ual character) basis while avoiding the Of the non-San Marcos southeastern peculiar sampling properties of some other populations, only peripheral ones from the commonly used genetic distances, such as northeastern edge of the area display sub- Nei's (1972) D (Hillis, 1984; Frost and Hil- stantial sequence differentiation, especially lis, 1990). Wiens (1995) and Berlocher and the recently-discovered populations from Swofford (1997) pointed out that this springs along the Pedernales River in Trav- method is equivalent to use of Swofford is Co. These populations are located in an and Berlocher's (1987) MANOB criterion, isolated outcrop of Cow Creek limestone previously deemed computationally im- and there is little potential for direct con- practical. Our choice of differential nection of this aquatic system with other weighting schemes for changes among se- drainages known to be inhabited by Eu- quence-data character-states [Rodrigo's rycea. (1992) correction of Wheeler's (1990) combinatorial weighting method] allows Phylogenetic Relationships and Species incorporation of more detailed informa- Boundaries tion on patterns of nucleotide change than We believe that the best estimates of would a simple transition versus transver- phylogeny are derived from treatments of sion treatment, and Hillis et al. (1994) the data that incorporate as much infor- demonstrated in simulation studies that mation as possible about evolutionary pro- symmetrically weighted parsimony per- cesses. Simulation and congruence studies forms substantially better in recovering the suggest that, in general, incorporation of true phylogeny than does uniform weight- realistic evolutionary parameters into char- ing of change probabilities. acter coding and weighting increases the The issue of whether or not to combine likelihood that the correct tree will be re- data sets in phylogenetic analysis has been covered, for both DNA sequence (e.g., the subject of considerable controversy Hillis et al., 1994; Huelsenbeck, 1995) and (reviewed by Chippindale and Wiens, allele frequency data (e.g., Wiens, 1998, 1994; Miyamoto and Fitch, 1995; Hillis, 2000; Wiens and Servedio, 1997, 1998). 1995; de Queiroz et al., 1995; Huelsen- Although application of frequency infor- beck et al., 1996). Most workers agree that mation in parsimony analysis has been it is desirable to examine trees reconstruct- controversial (e.g., see Buth, 1984; Croth- ed based on both separate and combined er, 1990; Murphy, 1993; Murphy and analyses, and we consider the results of Doyle, 1998), compelling arguments for both kinds of analyses here. Our preferred use of frequency-based methods have hypothesis of phylogeny is that derived been made (Swofford and Berlocher, from combination of the allozyme and 1987; Wiens, 1995, 2000). Simulation and DNA data, the former treated with fre- congruence studies indicate that methods quency-based coding and the latter with that incorporate frequency data tend to be combinatorial weighting and codon posi- more accurate than those that do not use tion partitioning. Despite the apparently this information (Cunningham, 1997; poor performance of the allozyme data Wiens, 1998, 2000; Wiens and Servedio, alone in resolving most relationships, their 1997, 1998). Given that results of simula- use in combination with the sequence data tion studies (Wiens and Servedio, 1997, substantially increased support for several 1998) were upheld by congruence analyses relationships indicated by the sequence of both morphological and allozyme data data alone (and in most cases in which (Wiens, 1998, 2000), claims that the sim- node support decreased as a result of data ulations are questionable because of un- combination, the changes were relatively realistic model assumptions (Kornet and minor). Turner, 1999) appear to be unwarranted. The major regional similarity groupings The Manhattan distance/step matrix ap- identified by cluster analyses of the allo- proach allows use of allele frequency in- zyme and DNA data appear in the phylo- 20 HERPETOLOGICALMONOGRAPHS [No. 14 genetic trees as monophyletic assemblages istence of monophyletic southeastern and (with the exception of the southwestern southwestern groups. Beyond this, there is group using allozymes alone), and all anal- little support based on the phylogenetic yses support monophyly of the central Tex- analyses for any particular pattern of re- as hemidactyliines. There is support for lationships within the major groups that monophyly of the northern group in all we have identified. For all treatments of analyses, and it is clear (as discussed the data, within-group bootstrap values are above) that these populations do not rep- nearly all below 70%, the level found in resent the species Eurycea neotenes. Con- simulation studies by Hillis and Bull sideration of all data supports recognition (1993) to correspond to a 95% probability of at least three separate species in the that a given clade is real. northern group; formal descriptions are With respect to taxonomy, none of our presented below. The basal split between analyses support the previous view (e.g., populations of Eurycea from north versus Baker, 1961; Brown, 1950, 1967c; Sweet, south of the Colorado is consistent with 1978a, 1982, 1984) of Eurycea neotenes as the view of the Colorado as an ancient, widespread throughout the region. We re- strong barrier to gene flow that has divided strict E. neotenes to springs in the area of the group into two major clades. Eurycea the type locality at Helotes Creek, Bexar rathbuni is phylogenetically embedded Co., resurrect the name E. pterophila Bur- within the Texas Eurycea in all analyses; ger, Smith and Potter 1950 for Blanco Riv- thus we concur with Mitchell and Reddell er drainage populations, continue to rec- (1965) and Mitchell and Smith (1972) that ognize the species E. tridentifera, E. the genus Typhlomolge should be sub- sosorum, and E. nana, and suggest that sumed within the genus Eurycea. Analyses many more, as yet unnamed, species exist. based on either DNA alone or combined All named species in the group occur in allozyme and sequence data indicate a sis- the southeastern region except E. troglo- ter-taxon relationship between E. rathbuni dytes from Valdina Farms Sinkhole in Me- (plus, presumably, E. robusta, which was dina Co. This taxon [which may now be unobtainable) and the remaining members extinct due to habitat modification (Veni of the southern group. This result suggests and Associates, 1987; G. Veni, personal an early divergence for the ancestor of E. communication to PTC)] was synonymized rathbuni and E. robusta. The other San by Sweet (1978a, 1984) under both E. neo- Marcos taxon, E. nana, is more problem- tenes and E. tridentifera because he con- atic, and its placement varies depending sidered this population a hybrid swarm. on which subset of the data is used. The Although allozyme data are not available, sequence data place it with the southeast- cyt b sequence for a specimen from the ern group, not surprising given its minimal type locality places this population within sequence divergence from them, and the the southwestern group, where it occurs combined data treatment provides very geographically. Given the maternal inher- strong support for this species as the sister itance of the mitochondrial genome, there taxon to the southeastern group (San Mar- conceivably could be male-based flow of cos is located in the southeastern Edwards genes (in the form of E. tridentifera from Plateau region but in many respects exhib- the southeast, one of the putative parent its a unique aquatic biota with many en- species), and thus the population could demic taxa; e.g., see Holsinger and Long- still consist of hybrids. However, we doubt ley, 1980). However, the high level of this, because there is no other evidence of divergence of E. nana based on allozymes widespread gene flow in this group. The leads to a weakly supported result in which other putative hybrid taxon whose name it is placed outside all other central Texas was synonymized by Sweet, E. latitans, hemidactyliines when the allozyme data does not appear to be a hybrid between E. alone are considered (Fig. 5). tridentifera and surface Eurycea based on Phylogenetic analyses based on the se- the allozyme data. The geographic location quence and combined data support the ex- of populations of E. latitans is much closer 2000] HERPETOLOGICALMONOGRAPHS 21 to the known range of E. tridentifera than cept sensu Wiley (1978) and Frost and is that of E. troglodytes, and gene flow in Hillis (1990). According to this view of the hydrologic system of this area seems species, one seeks evidence that a popu- much more plausible than between this relation or interbreeding group of popula- gion and the portion of the southwestern tions constitutes a distinct evolutionary lin- region in which E. troglodytes occurs, 75 eage that will continue to maintain a km distant from the known range of E. tri- separate identity from other such lineages dentifera. For these reasons, we recognize until it becomes extinct or undergoes ad- E. troglodytes and E. latitans. At a mini- ditional speciation. This evidence can be mum the name E. troglodytes should ap- based on many kinds of characters (e.g., ply to the Valdina Farms locality; we rec- morphological, molecular, behavioral, etc.) ommend that it be extended to include all and geographic considerations (i.e., the members of the southwestern group in the potential for isolation from gene flow) also "E. troglodytes complex" pending formal are relevant. By accepting the Evolution- description of other species in the region. ary Species Concept, we emphasize what Eurycea latitans applies at least to popu- species are (namely, distinct, independent- lations of the Cascade Caverns system of ly evolving historical lineages) rather than Kendall Co., and here we include several emphasizing any particular attributes of other populations in the area as members historical lineages (cf. Mayr, 1982; Pater- of a species complex containing this taxon. son, 1985). We also accept all evidence for the independence of historical lineages, Species Boundaries in the Central Texas rather than restricting ourselves to a lim- Eurycea ited or arbitrary set of criteria (cf. Cracraft, Prior to the molecular. work detailed 1989; Highton et al., 1989; Highton, 1990, here, nearly all assessments of species di- 1995). versity and boundaries in the central Texas Below, we present our interpretation of hemidactyliines were based on morpholog- species boundaries in the central Texas ical criteria. This approach resulted in the hemidactyliines and provide a brief ac- recognition of a relatively small number of count of each of the groups that we used taxa in the group, at most six species or for phylogenetic analyses. Species to which subspecies of Eurycea (e.g., see Baker, we assigned each of the populations sam- 1961; Brown, 1967a-c) plus two species of pled, and their geographic distributions, Typhlomolge. Several authors suggested, are shown in Fig. 8. Considerable ambi- however, that additional species remained guity remains with respect to fine-scale re- to be discovered in the group (e.g., Baker, lationships in the Texas Eurycea, and this 1961; Bogart, 1967; Brown, 1950, 1967c; is reflected by our tentative recognition of Mitchell and Smith, 1972). Sweet possibly heterogeneous groups such as the (1978a,b, 1982, 1984) assigned most "Carson Cave group" in the southwestern spring and cave populations from through- Edwards Plateau region. We identify two out the Edwards Plateau region to the spe- groups of populations as species complex- cies E. neotenes. This left three recognized es: the E. latitans complex in the south- species of Eurycea in the group (E. neo- eastern plateau region and the E. troglo- tenes, E. tridentifera, and E. nana), plus a dytes complex in the southwestern plateau fourth taxon from Barton Springs in Travis region. We provide synonymies for each of Co. that Sweet (1978a, 1984) considered a the species that we recognize, detail the distinct species based on morphology, later evidence that led us to separate them, rec- described by Chippindale et al. (1993) as ommend a formal name for members of E. sosorum. Potter and Sweet (1981) also the group if one currently is in use or is recognized two species in the genus Ty- available for resurrection, indicate what phlomolge, T. rathbuni and T. robusta, taxonomic problems remain, and suggest which we consider here to represent mem- ways to solve these problems. We also de- bers of the genus Eurycea. scribe three new species from the north- We follow the evolutionary species con- ern region. Our primary aims are to doc- m

22 HERPETOLOGICAL MONOGRAPHS [No. 14

YE. sosorum E. sp. Pedernales " Southeast ,5=^ LS E E. pterophila *^ * ~ E. robusta I * E. rathbuni San Marcos E. nana E. sp. Comal Springs

E. tridentifera Southeast E. neotenes E. latitanscomplex w m m

FIG. 8.-Geographic distribution of species of central Texas Eurycea recognized in this study. The region outlined corresponds to the counties in which these salamanders are known to occur, detailed in Fig. 1. Filled circles represent surface spring localities and stippled squares represent cave localities. ument the remarkable diversity in the cies formerly assigned to the genus Ty- group, identify populations or groups of phlomolge) from southwest of the populations that appear to represent dis- Colorado River, ranging from western tinct species, and provide a framework Travis County to eastern Val Verde Coun- within which future studies of this diver- ty. We have examined specimens from 12 sity can proceed. of the 14 counties in the southern region In the following accounts, we have or- in which salamanders are known or ganized the species and informal groups thought to occur. The exceptions are Kin- that we recognize by geographic region. ney Co., in which salamanders are likely infor- Appendix I provides precise locality to occur but no populations are known, mation for all examined. The populations and Val Verde Co. For the latter, repeated designation of a San Marcos region (in- visits to the two known salamander local- habited E. and by Eurycea nana, rathbuni, ities at Del Rio (Fourmile Cave and San E. is one of robusta) simply convenience, see failed because the taxa from this area are Felipe Springs; Sweet, 1978a) highly to We that these and the of E. nana is yield specimens. expect divergent placement salamanders would fall within the south- uncertain; no close relationship between western that we be- E. nana and the other two species at San group recognize (see based on their location. late in Marcos is implied. Museum acronyms follow) Very low Leviton et al. (1985). Distinguishing the study, Eurycea were discovered by molecular character states are listed; for Riley Nelson and co-workers north of Del sequence characters, these refer to the po- Rio in springs in the Devil's River drain- sition within cyt b relative to the first nu- age. Superficially, these appear similar to cleotide in the alignment of the sequence other surface populations from the south- we obtained (Appendix III). west, but we have not yet examined them for molecular markers. The southern SYSTEMATIC ACCOUNTS group is strongly supported as monophy- I. Southern Region letic in phylogenetic analyses of the mo- This assemblage includes all popula- lecular data that include mitochondrial tions of Eurycea (including the two spe- sequence. 2000] HERPETOLOGICALMONOGRAPHS 23

IA. San Marcos Region Comments.-Three species have at various been This area is located within Hays Co. in times regarded as members of the the southeastern region, but displays a genus Typhlomolge. Typhlomolge rathbuni unique aquatic fauna in many respects, Stejneger 1896 was the first to be de- and is and the salamanders found here (Eurycea scribed, known from several caves and rathbuni and E. robusta underground and wells that intersect the San Marcos Pool of E. nana on the surface) are highly diver- the Edwards Fault Zone Aquifer gent from others in the Edwards Plateau (Russell, 1976; Potter and Sweet, 1981; region based both on morphology and mo- Longley, 1986). In 1981, Potter and Sweet lecular markers. Synonymies and basic in- redescribed another species of Typhlom- formation are as follows: olge, T. robusta, based on a single specimen collected in 1951 from a hole drilled Eurycea nana in the bed of the Blanco River northeast of San Marcos. nana 1941:6. Longley (1978) uninten- Eurycea Bishop the Eurycea neotenes nana: Schmidt 1953:55. tionally provided original description of this taxon (e.g., see Potter and Sweet, Holotype.-UMMZ 89759. 1981) because in a government report he Type locality.-"Lake at the head of the used the name T robusta, originally ap- San Marcos River, Hays County, Texas". plied in a master's thesis by Potter (1963), Comments.-Chippindale et al. (1998) and offered an extremely brief description discussed the confusion in the literature of the type specimen while acknowledging regarding the status and distribution of Potter as the author of the name. Dixon this taxon and provided molecular and (1987) rejected Longley's description and morphological evidence that it is restricted cited several ways in which it deviated to the outflows of San Marcos Springs, from taxonomic practices specified by the Hays Co. All recent authors have recog- 1985 revised International Code of Zoo- nized this species, a position which we logical Nomenclature; authorship of the support. name remains controversial. Wake (1966) Distinguishing molecular character initially recognized a third species in the states are as follows. Allozymes: Unique a genus, T. tridentifera, originally described allele fixed at Ak; d allele fixed at Acon-1 by Mitchell and Reddell (1965) as Eurycea (an allele with this mobility is also seen in tridentifera. Sweet (1977a) listed addition- the outgroup taxa Haideotriton and Ty- al applications of this combination. How- phlotriton); unique e allele fixed at sAat; c ever, Wake's assessment was based on a allele fixed at Gr (also seen in the outgroup very limited number of specimens, and taxon Typhlotriton); d allele fixed at Pep- subsequent morphological work using ad- B (also present at low frequency in E. ditional material led Mitchell and Smith rathbuni and in the outgroup taxon E. (1972) to place members of the genus Ty- wilderae); unique h allele at Pep-D. Se- phlomolge in the synonymy of the genus quence: unique T at 24; unique A at 217. Eurycea. This conclusion is strongly supported by our molecular work. Eurycea rathbuni Since the discovery of Eurycea rathbu- Typhlomolge rathbuni Stejneger 1896:620. ni, relationships of the taxon have re- Eurycea rathbuni: Mitchell and Reddell mained controversial. Emerson (1905) first 1965:23. recognized that E. rathbuni was a plethodontid, and Wake (1966) assigned the ge- Holotype.-USNM 22686. nus to the tribe Hemidactyliini. Based on Type locality.-"subterranean waters pre-phylogenetic analyses, Wake (1966) near San Marcos, Texas" (the specimens regarded what was then considered Ty- examined by Stejneger were from the 188- phlomolge as part of a "pre-Eurycea" Mio- foot-deep Artesian Well at what is now cene radiation of hemidactyliine pletho- Southwest Texas State University in San dontids in southeastern and south-central Marcos). North America. Mitchell and Reddell 24 HERPETOLOGICALMONOGRAPHS [No. 14

(1965) and Mitchell and Smith (1972) Hays Co. (see Potter and Sweet, 1981 for viewed members of what had been consid- further discussion). ered Typhlomolge simply as extremes in a Distinguishing molecular character continuum of cave-associated morpholog- states are as follows. Allozymes: e allele ical changes in the Texas Eurycea. Potter fixed at Acon-1 (also seen in Eurycea sp. and Sweet (1981) concurred with Wake's from Comal Springs and the outgroup tax- (1966) interpretation of the time scale for on Typhlotriton); unique, presumed null invasion of the Edwards Plateau region by allele fixed at Gr; unique b allele fixed at ancestors of what they considered Ty- Mdh-2; unique g allele at Pep-D at 75% phlomolge and recommended (based on frequency. Sequence: T at 9 (shared with morphological analyses) that the genus Ty- Typhlotriton); unique A at 24; unique A at phlomolge should continue to be recog- 28; unique C at 36; T at 48 (shared with nized. Using parsimony methods, Lom- Typhlotriton); A at 69 (shared with Ty- bard and Wake (1986; and D.B. Wake, phlotriton); T at 78 (shared with some out- personal communication) investigated re- group members); C at 174 (shared with lationships among genera of plethodontids Greenwood Valley Ranch Springs, south- based primarily on characters of the feed- western group); T at 201 (shared with the ing apparatus (especially the tongue), but outgroup taxon E. bislineata). were unable to reliably reconstruct the po- Eurycea robusta sition of Typhlomolge except to verify that it belonged within the Hemidactyliini. Typhlomolge robusta Longley, in Potter Allozyme and mitochondrial sequence and Sweet 1981:70. data support the monophyly of the central Eurycea robusta: Petranka 1998:275. Texas inclusive of hemidactyliines Eurycea Holotype.-TNHC 20255. which is nested in the rathbuni, deeply Type locality.-"Beneath the Blanco group (E. robusta was unobtainable for River, 178 m elevation, 5 airline km NE of molecular but we assume based on studies, the Hays County Courthouse, San Marcos, that it is the sister to morphology species Hays County, Texas". E. rathbuni; see Potter and Sweet, 1981, Comments.-Above we recommend and Russell, 1976, for further details). Ty- synonymy of the genus Typhlomolge under cannot be retained as a phlomolge genus Eurycea, which renders this taxon Eurycea under current rules of nomenclature if it robusta. Petranka (1998) was the first to is a subset of Eurycea. In order to mini- formally apply the combination Eurycea mize the number of necessary taxonomic robusta, also used by Chippindale (1995) changes, we therefore synonymize the ge- in a Ph.D. dissertation. nus Typhlomolge under Eurycea. Such a move is consistent with the recommenda- IB. Southeastern Region tions of Mitchell and Reddell and (1965) This region encompasses western Travis, Mitchell and Smith and thus sim- (1972), Hays (except San Marcos, which we treat ply represents a reapplication of this com- separately; see above), Blanco, Comal, bination for E. rathbuni. Following the Kendall, Bexar, and extreme eastern Kerr recommendations of Chippindale (1995 counties. It contains the type localities for and personal communication), Petranka most of the central Texas Eurycea that (1998) also used the name Eurycea rath- have been recognized: Eurycea neotenes, buni, and synonymized T. robusta under E. latitans, E. pterophila, E. tridentifera, Eurycea. and E. sosorum. It is not clear whether the Frost (1985) gave the range of Eurycea problematic taxon E. nana (from San Mar- rathbuni as "Underground waters in Hays, cos Springs) should be considered part of Kendall, and Comal counties, central Tex- a monophyletic southeastern clade: this as, USA", but there is no reliable evidence species is highly divergent based on allo- that this species occurs anywhere but the zymes but falls within or sister to the San Marcos Pool of the Edwards Aquifer, southeastern populations based on cyto- 2000] HERPETOLOGICALMONOGRAPHS 25 chrome b sequence. Our preferred hy- E. neotenes have not generally been rec- pothesis of relationships (based on com- ognized since. Many authors (e.g., Baker, bined allozyme and sequence data) 1961; Behler and King, 1979; Brown, suggests that E. nana is the sister species 1950; Conant, 1958, 1975; Conant and of all southeastern taxa. Collins, 1991; Mitchell and Smith, 1972; We divide the southeastern clade into Schmidt, 1953; Sweet, 1977b, 1978a,b, seven taxa: Eurycea neotenes, E. pterophi- 1982, 1984) have regarded E. neotenes (or la, E. sosorum, E. tridentifera, Comal E. n. neotenes) as widespread in springs Springs, Pedernales Springs, and the E. la- and caves of the Edwards Plateau region. titans complex. It is likely that all except Based on our molecular evidence, we dis- the last represent single, distinct species. agree, and recommend restriction of the The evidence of species status is particu- name E. neotenes to spring populations lary strong for E. tridentifera, E. sosorum, from the vicinity of the type locality. The and the Pedernales populations, each of three populations that we examined which which exhibit diagnostic molecular mark- we place in E. neotenes are those from the ers and (for the former two) unique mor- type locality at Helotes Creek Spring, Bex- phological features (the Pedernales popu- ar Co., plus those at Leon Springs, Bexar lations have not yet been examined in Co., and Mueller's Spring, Kendall Co. detail morphologically). Eurycea neotenes, Distinguishing molecular character the E. latitans complex, E. pterophila, and states are as follows. Allozymes: These the Comal Springs population are distin- populations exhibit the b allele at high fre- guished primarily based on substantial al- quency at sAat, otherwise absent or at low lele frequency differences at allozyme loci. frequency in other southeastern populations and taxa (except E. tridentifera and Eurycea neotenes the Comal Springs and Pedernales popu- Eurycea neotenes Bishop and Wright lations); possess the b allele at Pgm at high 1937:142. frequency, otherwise seen only in the Co- Eurycea neotenes neotenes: Schmidt 1953: mal Springs and Camp Mystic (southwest- 55. ern group) populations and the outgroup member E. quadridigitata from Texas; and Holotype.-USNM 103161 (Cochran, exhibit the Mpi b allele at high frequency, 1961). otherwise seen only in the Pedernales pop- Type locality.-"Culebra Creek, 5 miles ulations, some members of the E. ptero- north of Helotes, Bexar County, Texas"; phila group, and the Fessenden and 176 corrected by Brown (1942) to the head- Springs populations from the southwest. spring of Helotes Creek in Bexar Co. Sequence: No diagnostic sequence substi- Comments.-Schmidt (1953) first sug- tutions were found. gested that Eurycea neotenes from the type locality, plus most other populations Eurycea pterophila of Eurycea in the Edwards Plateau region, Eurycea pterophila Burger, Smith, and be considered the subspecies E. n. neote- Potter 1950:51. nes, and he also recognized the subspecies Eurycea neotenes pterophila: Schmidt E. n. nana and E. n. pterophila (see above 1953:56. and below, respectively, for detailed ac- Eurycea neotenes: Sweet 1978b:106 (in counts of the latter two taxa). The subspe- part). cific designation was ignored by some authors (e.g., Baker, 1961). Brown (1967b,c) Holotype.-Floyd Potter Coll. No. A993 argued for recognition of E. nana as a full (private collection, now presumed lost). species but continued to recognize the Type locality.-"shallow stream flowing subspecies E. n. neotenes and E. n. pter- from Fern Bank Spring, 6.3 miles north- ophila. Sweet (1978b) formally synony- east of Wimberley on the Blanco River mized E. pterophila (or E. n. pterophila) Road, Hays County, Texas". under E. neotenes, and subspecies within Comments.-Sweet (1978b) demon- 26 HERPETOLOGICAL MONOGRAPHS [No. 14 strated that the morphological character Comments.-A detailed account of the states used by Burger et al. (1950) to dis- morphological and molecular features that tinguish Eurycea pterophila from other distinguish this species from other central Edwards Plateau Eurycea were either Texas Eurycea was provided in the original more widespread than previously thought description (Chippindale et al., 1993). or were erroneous (e.g., supposedly short Subsequent molecular work has reinforced digits were actually due to loss of tips the conclusion that this population repre- through bacterial infection). Sweet con- sents a distinct species isolated from gene cluded that there was no reason to rec- flow from all others. Our recognition of ognize the Fern Bank Spring population this taxon as a distinct species is consistent as a separate taxon, and relegated this tax- with Sweet's (1978a, 1984) conclusions on to synonymy with E. neotenes. Based based solely on morphology. on the information then available this was Distinguishing molecular character a logical assessment. Hamilton (1973) also states based on the data presented here was unable to distinguish Fern Bank are as follows. Allozymes: For Pep-A the Spring Eurycea from other populations in unique e allele is present at 83% frequen- the Blanco River drainage and elsewhere cy, and the other (d) allele present is oth- in the southeastern Edwards Plateau re- erwise seen only in members of the north- gion based on morphometric analyses. ern group, Eurycea nana, and the However, the allozyme evidence here Pedernales populations; and at Pep-D the shows that the Fern Bank Spring popula- a allele is fixed, otherwise seen only in geo- tion plus all others in the Blanco River graphically distant and otherwise very di- drainage share a high degree of similarity vergent members of the northern and in allele frequencies (although no alleles southwestern groups. Sequence: C at po- are diagnostic), suggesting recent or on- sition 195 is shared only with Pedernales, going gene flow. This, coupled with the re- Smith's Spring (southwest) and Carson striction of this group to a single drainage, Cave (southwest) populations and E. nana. leads us to recognize E. pterophila, especially because the allozyme and geographic Eurycea tridentifera evidence indicates that these populations Eurycea tridentifera Mitchell and Reddell almost certainly are isolated from gene 1965:14. flow with true (topotypical) E. neotenes. Typhlomolge tridentifera: Wake 1966:64. Further study will be necessary to clarify the status of this taxon. Populations that Holotype.-USNM 153780. we have examined and assign to this taxon Type locality.-"Honey Creek Cave, are: Fern Bank Spring, Zercher Spring, Comal Co., Texas". Boardhouse Springs, Peavey's Springs, Comments.-Mitchell and Reddell Grapevine Cave, and T Cave. (1965), Wake (1966), and Sweet (1977a, 1978a, 1984) regarded Eurycea tridenti- Eurycea sosorum fera as exhibiting a cave-associated mor- to of E. rathbuni Eurycea neotenes: Brown 1950:29 (in phology second only that E. in the part). (and robusta) extremity among Eurycea neotenes neotenes: Brown 1967c: Edwards Plateau hemidactyliines. The this led 36.1 (in part; mapped locality only). morphological features of taxon Eurycea sp.: Sweet 1984:429. Wake (1966) to transfer it to the genus Ty- work Mitchell Eurycea sosorum Chippindale, Price, and phlomolge, although later by Hillis 1993:249. and Smith (1972) suggested that the members of the genus Typhlomolge actually be- Holotype.-TNHC 51184. long within Eurycea, a move that we also Type locality.-"outflow of Parthenia support based on the molecular evidence. (Main) Springs in Barton Springs Pool, The molecular evidence also indicates that Zilker Park, Travis Co., Texas (30? 15' 49" E. tridentifera is not closely related to E. N, 97? 46' 14" W)." rathbuni. Sweet (1977a, 1978a, 1984) sug- 2000] HERPETOLOGICALMONOGRAPHS 27

gested that the caves of northern Bexar group, except in Comal Springs, Peder- Co. and the Cibolo Sinkhole Plain in nales, and Eurycea neotenes); at G3pdh which most known populations of this spe- the b allele appears fixed in the Honey cies occur are among the oldest in the pla- Creek Cave and Ebert Cave populations teau region, which would have allowed a (otherwise this allele is seen at low fre- long time for the evolution of cave-asso- quency in E. sosorum, E. rathbuni, and ciated features. Sweet (1978a, 1984) dem- members of the southwestern group; how- onstrated using morphometric analyses ever, it does not appear to be present in that populations from throughout the Badweather Pit E. tridentifera, suggesting known range of the species cluster togeth- that this population may in fact be isolat- er, and recommended recognition of this ed); and all three populations examined taxon as a single species with a relatively possess the unique Mdhp d allele at very wide subterranean range. The molecular high or 100% frequency. Sequence: No di- evidence supports this view for the popu- agnostic sequence substitutions were lations that we were able to examine; in found. addition, Bogart (1967) identified a chro- latitans mosomal nondisjunction that appears to Eurycea Complex be unique to the two populations of E. tri- The synonymy given below applies to dentifera that he examined (Badweather Eurycea latitans from the type locality Pit and Honey Creek Cave). Sweet only. We include numerous populations in (1977a) listed six known localities, and the E. latitans complex; all (except the pre- suggested that more likely would be found viously unknown Less Ranch Spring, in the Cibolo Sinkhole Plain and northern Cherry Creek Spring, and Cloud Hollow Bexar Co. As he predicted, salamanders Spring populations) were assigned by that appear to be this species have been Sweet (1978a, 1982) to E. neotenes. Some seen at Genesis Cave in northern Bexar of these may have been assigned to species Co. (D. Pearson, personal communication; earlier by Baker (1961) or Brown (1967c), see Veni, 1988 for details of the cave), and but for the populations in question these Chippindale and A.G. Grubbs have col- authors provided only maps, so the precise lected this species at Ebert Cave in Comal localities to which they referred are un- Co. [near Sweet's (1977a, 1978a, 1984) certain. Salamander Cave E. Kappelman locality]; Eurycea neotenes Wright and Wright was first collected there tridentifera by J. 1938:31 (in part) [assumed by Wright Reddell and M. This has Reyes. species and Wright to be E. neotenes based on also been found in caves of very recently a second-hand report; location given the Bullis base, Comal Co. Camp army (J. only as "a cave near Boerne", assumed Reddell and G. Veni, communi- personal by Smith and Potter (1946) to be Cas- cation to PTC), and several specimens cade Caverns; Bishop (1943) reported have been collected. Conant and Collins the presence of E. neotenes in Cascade restricted the distribution of this (1991) Caverns]. to the Creek species type locality, Honey Eurycea latitans Smith and Potter 1946: Cave, Comal Co., but this clearly was in 106. error. Eurycea neotenes latitans: Schmidt 1953; In addition to the distinctive morpho- 55. features exhibited this taxon logical by (see Eurycea tridentifera: Sweet 1984:438 (in Mitchell and Reddell, 1965; Wake, 1966; part) (Sweet regarded E. latitans as a ju- Mitchell and Smith, 1972; Sweet, 1977a, nior synonym of E. neotenes, but be- and the chromo- 1978a, 1984) potential lieved that the population contained in- somal described above, dis- autapomorphy trogressed genes from E. tridentifera). tinguishing molecular character states are as follows. Allozymes: At sAat, the b allele Holotype.-USNM 123594. is at medium to high frequency (this allele Type locality. -"the first large pool deep otherwise is rare in the southeastern within the recesses of Cascade Cavern, 4.6 28 HERPETOLOGICAL MONOGRAPHS [No. 14 miles by road (3 /2miles by airline) south- Comments.-The many characters that east of Boerne, Kendall County, Texas." distinguish this population from true Eu- Comments.-The status of this taxon is rycea nana (from San Marcos Springs, problematic. Sweet (1978a, 1984) dem- Hays Co.) are discussed by Chippindale et onstrated substantial morphological varia- al. (1998), and this population clearly does tion in this population and showed that not represent E. nana. It may represent a specimens from Cascade Caverns present- distinct species, although additional study ed a morphological spectrum from sur- of this population and its relationship to face-like morphologies to extreme troglob- others in the southeastern region is need- itic morphologies similar to those of ed. Further study of this putative taxon is Eurycea tridentifera. He hypothesized that critical, because its spring habitat is threat- this was the result of past introgression of ened by human demands on the waters of genes from the advanced troglobite E. tri- the southern Edwards Aquifer. dentifera into a cave population of E. neo- Distinguishing molecular character tenes. Our molecular data provide no evi- states are as follows. Allozymes: At Acon- dence of a hybrid origin for this 1, the e allele (otherwise seen only in Eu- population. Because the population that rycea rathbuni) is at medium frequency; lives in the underground system associated at sAat, the b allele (rare in the southeast with Cascade Caverns also does not appear except in E. tridentifera, E. neotenes, and to represent E. neotenes based on the mo- Pedernales) is at medium frequency; and lecular evidence, it seems reasonable to at Pgm, the b allele is at medium frequen- resurrect the name E. latitans for salaman- cy (otherwise seen only in E. neotenes, ders in this cave system. Here we assign Camp Mystic from the southwest, and the several other populations (Rebecca Creek outgroup member E. quadridigitata from Spring, Hays Co., Bear Creek Spring, Ci- Texas). Sequence: No diagnostic sequence bolo Creek Spring, Less Ranch Spring, characters were found. and Kneedeep Cave Spring, all in Kendall Co., Cherry Creek and Cloud Hollow Eurycea sp.-Pedernales populations Springs, Kerr Co., and Honey Creek Cave No previous taxonomic history. Spring, Comal Co.) to this species based and Hillis on similarity in allele frequencies and cy- Comments.-Chippindale found the first of two known tochrome b sequences. This group likely is populations a catch-all for members of the southeast- of this salamander in 1989, in two small on the northeast side of the Ped- ern group whose affinities are uncertain, springs ernales River in extreme western Travis and definitely needs further investigation. No molecular unite all Co. These springs are located in an isolat- synapomorphies ed band of Cow Creek limestone and are members of the group; the Rebecca Creek well separated geographically from all oth- population and Eurycea lati- topotypical er known populations of central Texas Eu- tans each exhibit potential sequence au- rycea. Although no detailed morphological tapomorphies. Specifically, Rebecca Creek studies have yet been conducted, these exhibits a T at position 99 otherwise seen salamanders appear to mature at a very only in Haideotriton. E. lati- Topotypical small size, and exhibit a relatively high fre- tans have a T at 82 (also seen in position quency of "gold" morphs (individuals in the southwestern Carson Cave population which the are and members E. bislineata and melanophores widely sepa- outgroup rated, yielding a light yellowish-gold col- E. wilderae). or). These salamanders possess unique combinations of allozyme and sequence Eurycea sp.-Comal Springs character states and almost certainly rep- Eurycea neotenes: Baker 1961:29 (in part). resent a distinct species; we expect to for- Eurycea neotenes neotenes: Brown 1967c: mally describe them as such pending com- 36.1 (in part; mapped locality only). pletion of additional molecular and Eurycea nana: Dixon 1987:60 (in part). morphological studies. 2000] HERPETOLOGICALMONOGRAPHS 29

Distinguishing molecular character cies. The southwestern group is strongly states are as follows. Allozymes: At Acon- supported as monophyletic in phylogenetic 1 the c allele (otherwise characteristic of analyses of the molecular data that include the northern group) occurs at low fre- mitochondrial sequence. quency; at sAat, the b allele is at medium frequency (otherwise rare in the southeast Eurycea troglodytes Complex except in Eurycea tridentifera, E. neote- The synonymy given below applies to nes, and Comal Springs); at Ldh-A the a Eurycea troglodytes from the type locality allele is at medium frequency (otherwise only. We include numerous populations in seen only in the T Cave population, a the E. troglodytes complex. All that we ex- member of the E. pterophila group); and amined (except the previously unknown at Mdhp the unique f allele occurs at low Greenwood Valley Ranch Springs and frequency. Sequence: T at position 75 West Nueces Spring localities) were as- (shared only with Smith's Spring in the signed by Sweet (1978a, 1982) to E. neo- southwestern group); C at 195 (shared tenes. Some of these may have been as- with E. sosorum, E. nana, and the south- signed to species earlier by Baker (1961) western Smith's Spring and Carson cave or Brown (1967c), but for the populations populations); unique G at 245; T at 246 in question these authors provided only (shared with some northern and south- maps, so the precise localities to which western group members as well as mem- they referred are uncertain. bers of the outgroup); C at 282 (shared E. Baker 1957:329. with E. rathbuni and members of the troglodytes: E. neotenes: Sweet 1984:438 (in part). northern and southwestern and groups E. tridentifera: Sweet 1984:438 (in part). outgroup); and C at 312 (shared with members of the southwestern and north- Holotype.-TNHC 21791. ern groups, E. rathbuni, and outgroup Type locality.-"a pool approximately members). 600 feet from the entrance of the Valdina Farms Sinkhole, Valdina Farms, Medina IC. Southwestern Region County, Texas." This region encompasses southern Gil- Comments.-Sweet (1978a, 1984) con- lespie Co., most of Kerr Co. (except the sidered this taxon a hybrid swarm derived easternmost extreme), Bandera, Real, and from Eurycea tridentifera and what he Edwards Counties, northern Medina, considered E. neotenes based on the mor- Uvalde, and probably Kinney Co., and phological variability that it displays. How- likely the populations in Val Verde Co. ever, this seems very unlikely based on Our molecular studies are only the begin- molecular evidence and geographic and ning for discovery of diversity in this clade, hydrologic considerations. Therefore, we and much more intensive sampling will be recommend continued recognition of this necessary to reliably identify species species, especially because it is the only boundaries in this group. Most known named member of the southwestern populations in the southwestern region group. Unfortunately, the population at have been considered Eurycea neotenes the type locality may have been destroyed (e.g., Baker, 1961; Brown, 1967c; Sweet, by human modification of its habitat (Veni 1978a, 1982), and only one other species and Associates, 1987; G. Veni, personal has been recognized in the southwestern communication to PTC). region (E. troglodytes Baker 1957; see be- Pending further investigation of rela- low). Both allozyme and sequence data tionships of the southwestern Edwards suggest that many members of this group Plateau Eurycea, we recommend that all are isolated from gene flow and probably members of the southwestern group col- represent distinct species. Our recognition lectively be referred to as the E. troglo- of a "Carson Cave group" is based on phe- dytes complex. This includes the following netic criteria and it may not represent a OTU's that we used in our phylogenetic monophyletic group of populations or spe- analyses: Tucker Hollow Cave, Greenwood 30 HERPETOLOGICAL MONOGRAPHS [No.[NJo 1414

Valley Ranch Springs, Sabinal group, Suth- Comments.-This is one of the few cave erland Hollow Spring, Trough Spring, and populations of Eurycea distant from the Carson Cave group (see Fig. 3 and Ap- Balcones Fault Zone, and almost certainly pendix I). This is a heterogeneous group is isolated from all or most other popula- that exhibits considerable molecular diver- tions. These salamanders, known only sity, and probably includes several distinct from two shallow pools in a hillside cave, species. However, the molecular evidence exhibit strong cave-associated morpholo- indicates that it probably is monophyletic, gies, including reduced eyes and pigmen- and it is clear that none of the members tation and broadened heads. Sweet of the group that we have examined are (1978a) provided a detailed morphological conspecific with any other central Texas description of the animals, and Sweet Eurycea. Below we offer a brief account (1984) included this population in a mor- of each of the OTUs from within this phometric analysis of central Texas cave group that we used for phylogenetic anal- Eurycea. J. R. Reddell (personal commu- yses. nication to PTC) considered this popula- Distinguishing molecular character tion a distinct species, although a formal states for this species complex are as fol- description was never published. Given lows. Allozymes: At sAat the b allele, ab- the morphological and molecular distinc- sent or at low frequency in many south- tiveness of this population and its apparent eastern populations, appears fixed in all isolation, this assessment may well be cor- southwestern populations; in all popula- rect. tions the Pep-B a allele appears fixed or at Distinguishing molecular character very high frequency (otherwise seen only states are as follows. Allozymes: The among the Texas Eurycea in topotypical E. unique c allele appears fixed at Idh-1; the latitans, E. tridentifera, and E. sosorum, a allele appears fixed at Gapdh (also seen and in the outgroup taxon E. quadridigi- at medium frequency in populations of the tata from South Carolina). Sequence Sabinal group, below). Sequence: C at po- (characters and states listed are those that sition 49 (shared with the southwestern distinguish members of the E. troglodytes Smith's Spring and Carson Cave popula- from members of the southeast- complex tions, southeastern populations, and some ern to which are group, they geographi- outgroup members); C at 186 (shared with A at 75 in all ex- cally adjacent): position E. rathbuni, Smith's Spring, Carson Cave, Smith's which has a T cept Spring, (A and some outgroup members); unique G shared with northern group and outgroup at 225. members); T at 126 (shared with members the northern E. and of group, rathbuni, Eurycea troglodytes Complex- some A at 198 outgroup members); Greenwood Valley Ranch Springs (shared with some members of the northern group and outgroup); A at 219 (shared No previous taxonomic history. with E. some members of the rathbuni, these sala- northern and members); Comments.-Superficially, group, outgroup manders similar to other south- T at 240 (shared with E. rathbuni, some appear members of the northern and out- western spring populations. Additional group, in this area of the is group members); A at 291 (shared with E. sampling range very desirable, as the status of these rathbuni, members of the northern group, populations is uncertain. Based on molecular di- and members of the outgroup), G at 321 could an undes- (shared with northern group); and A at vergence, they represent 361 (shared with E. nana, E. rathbuni, and cribed species. molecular character some outgroup members). Distinguishing states are as follows. Allozymes: Lack of Eurycea troglodytes Complex-Tucker detectable activity at Mdh-2 (treated as Hollow Cave null allele). Sequence: C at position 174 Eurycea neotenes: Sweet 1984:433 (in (shared with E. rathbuni); C at 323 part) (shared with Carson Cave). 2000] HERPETOLOGICALMONOGRAPHS 31

Eurycea troglodytes Complex-176 group, and some outgroup members); and Spring C at 204 (shared with Testudo Tube in the Eurycea neotenes: Sweet 1982:441 (in northern group and Trough Springs). part). Eurycea troglodytes Complex-Sabinal Comments.-Like many southwestern group this to be populations, population appears Eurycea neotenes: Sweet 1982:441 (in distinct based on molecular evidence, but part). sampling in the region is very limited and thus the status of this population is uncer- Comments.-These populations include tain. Distinguishing molecular character some of the few known naturally meta- states are as follows. Allozymes: Unique morphosing Eurycea in the Edwards Pla- Ldh-A e allele at high (88%) frequency; teau region (discussed by Bruce, 1976; fixed a allele at Pep-A (also seen in nearby Sweet, 1977b). They may represent a dis- Fessenden Springs). Sequence: Unique T tinct species. at position 171; unique T at 196; C at 303 Distinguishing molecular character (shared with northern populations, Eury- states are as follows. Allozymes: At Gapdh cea rathbuni, and some outgroup mem- the a allele (otherwise seen only in Tucker bers). Hollow Cave) is at medium frequency; at Mdhp the unique e allele is at high fre- Eurycea troglodytes Complex-Camp quency. Sequence: No diagnostic se- Mystic Springs quence substitutions were found. Eurycea neotenes: Sweet 1982:441 (in Eurycea troglodytes complex-Carson part). Cave group Comments. -This is another southwest- Eurycea neotenes: Sweet 1982:441 (in ern spring population that does not appear morphologically distinct, but which has part). many molecular character states that in- Comments.-Here we omit earlier au- dicate isolation from gene flow, at least thors' taxonomic assignments of popula- from the other populations examined. tions which may be closely related to Thus it may represent another distinct members of this group, because the species. Several sequence characters sug- group's composition is highly problematic. gest the possibility of close relationship Thus we cite only Sweet because he ac- with the Trough Spring population, here tually assigned most of the same popula- placed in the informal (and problematic) tions that we examined to Eurycea neote- Carson Cave group (below). Distinguish- nes. This assemblage (based on similarity ing molecular character states are as fol- in allozyme allele frequencies) definitely lows. Allozymes: Unique c allele at 90% needs further study. Even the relationships frequency at Mdh-1; fixed a allele at Pep- suggested by the sequence data are at D (shared with some northern popula- odds with this grouping in some respects tions, Eurycea sosorum, and nearby Fes- (e.g., for Trough Spring and Camp Mystic senden Springs); b allele fixed at Pk (an Spring), and there may be several species allele with the same mobility is seen in the involved. In future studies, the best strat- outgroup member Haideotriton). Se- egy probably will be to address relation- quence: C at position 83 (shared with ships within the southwestern group alone, some outgroup members); T at 138 treating as many populations as possible as (shared with members of the northern separate units of analysis. group and the outgroup member E. lon- The population for which this group is gicauda); T at 161 (shared with members named, Carson Cave, consists of morpho- of the northern group, some outgroup logically distinctive troglobites similar in members, and nearby Trough Springs); A some respects to those from Tucker Hol- at 203 (shared with Trough Springs, E. low Cave (see Sweet, 1978a, 1984 for rathbuni, Cedar Breaks in the northern more detailed morphological information). 32 HERPETOLOGICAL MONOGRAPHS [No. 14

It may represent a distinct species (as be- the dorsal surface of the tail in mature an- lieved by J. R. Reddell, personal commu- imals from the northern group is usually nication to PTC), although sequence data wider and more vivid than that seen in suggest a close relationship between this members of the southern group in which population and the nearby but morpholog- this feature is present. ically dissimilar Smith's Spring population. Identification of species boundaries in The populations included in this informal the northern group has proven difficult in group range widely in the southwestern some cases, especially with respect to the region and do not correspond to any single cave populations. We have treated most drainage. In this group we have included subterranean populations in the region as the populations from West Nueces Spring separate units for phylogenetic analysis be- and Smith's Spring (Edwards Co.), Carson cause of the uncertainty of their place- Cave and WB Spring (Uvalde Co.), Suth- ment; clearly additional sampling, and the erland Hollow Spring (Bandera Co.), Rob- use of new molecular markers, is desirable inson Creek Spring and Fessenden Spring to further elucidate relationships and spe- (Kerr Co.), and Trough Spring (Gillespie cies boundaries. Here we provide formal Co.). descriptions of three members of the northern group that clearly warrant status II. Northern Region as species distinct from one another, and Molecular evidence for monophyly of identify additional populations whose sta- the northern group is overwhelming. De- tus and affinities within the northern spite their high level of divergence, no group remain problematic. members of the group have been formally described as distinct species, mainly be- Eurycea tonkawae sp. nov. cause animals from the northern surface Fig. 9A populations known to previous workers ap- Eurycea neotenes: Baker 1961:30 (in part). peared very similar based on external mor- Eurycea neotenes neotenes: Brown, 1967c: phology to those from southern spring 36.1 (in part; mapped only). populations. Also, many of the northern locality populations were discovered during the Holotype.-TNHC 50952 (field number course of this study, and thus their diver- AHP 3240), an adult female collected by sity and the extent of their range previ- Andrew H. Price and Paul T. Chippindale ously was very poorly known. 12 September 1991 at the primary out- In addition to numerous allozyme and flows of Stillhouse Hollow Springs, Travis cytochrome b sequence characters, mem- Co. Texas, 30? 22' 28" N, 97? N45' 55" W. bers of the northern group are character- Paratypes.-TNHC 1802-816, 50933- ized (among the central Texas Eurycea) by 951, 50953-954, 50956-962, 50964-969, substantially larger genome sizes than all 50972-990, 50992-993, 51169-172, others [Chippindale and Lowcock, unpub- 53465-473, 53475-505, 53857, 54225-226, lished; Licht and Lowcock, 1991 (but see 54546-549, 55132-134, 55136, 55139-157, correction of Licht and Lowcock's data ta- 55387; MVZ 122695-703; UTA A-52989- ble above)], and all members examined so 52990. Cleared-and-stained: TNHC 50955, far exhibit a diagnostic ApaLl restriction 50963, 50991, 52759, 53474. site near the end of the 28S nuclear ri- Diagnosis.-The following diagnosis ap- bosomal DNA repeat unit (Chippindale, plies only to spring-dwelling populations unpublished). It is our subjective impres- of this species; see notes below under Ad- sion that the yellow stripe which occurs on ditional comments regarding cave-dwelling

FIG. 9.-Holotypes of northern central Texas species of Eurycea in life. A. Holotype of Eurycea tonkawae, adult female, TNHC 50952. B. Holotype of Eurycea naufragia, adult male, TNHC 58860. C. Holotype of Eurycea chisholmensis, adult male, TNHC 58859. 2000] HERPETOLOGICAL MONOGRAPHS 33

-t.W4

C 34 IIERPETOLOGICALHERPETOLOGICAL MONOGRAPHISMONOGRAPHS [No. 14 populations in the region that we provi- neck laterally just behind the angle of the sionally assign to this species. Based on ex- jaws, deep rich red (in life) and increasing ternal morphology, spring-dwelling speci- in size posteriorly. Gill insertions arranged mens of Eurycea tonkawae appear very in a stepped pattern with anterior pair far- similar to E. naufragia (described below). thest from and posterior pair closest to the However, the light areas surrounding each dorsal midline; the anterior pair project dorsolateral iridophore are square or ob- laterally from the neck and the posterior long in all but two of the specimens ex- pair arch over the neck. Dorsal ground amined in which this patterning was dis- color in life dark greenish-brown. Head cernible; two exhibited the rosette or and face anterior to the parietal region yel- starburst shaped light areas seen in all but low in life, the yellow extending backwards one of the 32 specimens of E. naufragia laterally on the cheeks to the gill inser- that were examined. In addition, the dark tions, leaving a dark diamond-shaped pa- lateral margins of the yellow-orange dorsal rietal mark connected posteriorly with the tail fin coloration generally are irregular dorsal ground color. A distinct dark canthal and chevron-shaped, whereas in E. nauf- line connects the anterior comer of each ragia they usually are regular and com- eye to each naris, and another diffuse dark plete. Eurycea tonkawae differs osteolog- line extends from the anterior point of the ically from the other two northern species parietal diamond forward along the mid- (E. naufragia and E. chisholmensis) in hav- line of the skull. A band of melanophores ing a relatively extensive separation of the is concentrated along the dorsal midline. frontals posteriorly, and a larger frontopa- Three distinct rows of iridophores are pre- rietal fontanelle. Eurycea tonkawae fur- sent, each iridophore centered in a square ther differs from E. naufragia in having or oblong-shaped area which was dull- the pars dorsalis of the premaxilla separat- white or cream-colored in life. The two ed (in contact or abutting medially in E. dorsolateral rows, the lateral-most just naufragia) and from E. chisholmensis in dorsal to the level of the limb insertions, having 16-17 presacral vertebrae (18 in E. extend posteriorly from the gills to the an- chisholmensis) and having the distal por- terior insertion of the dorsal tail fin. A tion of the ceratohyal mineralized (not third lateral row at the level of the limb mineralized in E. chisholmensis). insertions is confined to the trunk between Based on allozyme evidence, Eurycea the limbs. There are 16 costal grooves, tonkawae from springs of the Jollyville Pla- counting one each in the axilla and groin. teau and Brushy Creek is distinguished Tail relatively long and both dorsal and from other Eurycea that occur north of the ventral tail fins are well-developed. Dorsal Colorado River primarily by allele fre- surface of the tail and the dorsal tail fin quency differences; these include the from the level of the vent posteriorly were unique Mdhp a allele, present at varying bright yellow-orange in life (color 18 of frequencies (but not observed in all pop- Smithe, Naturalist's Color Guide, Ameri- ulations). Three nucleotide substitutions in can Museum of Natural History). Lateral cytochrome b (C at 56, G at 198, T at 281) margins of this area are a complex pattern were present that were not observed in of iridophores and melanophores forming any other populations of the northern dark chevrons, all connected as an irreg- clade of central Texas Eurycea. ular border. The ventral tail fin (bright yel- Description of holotype.-An adult fe- low in life) is narrow, and extends from the male (TNHC 50952) with large, well-de- vent posteriorly to the tip of the tail. The veloped eyes and a relatively broad head. ventral body surface is immaculate and Snout relatively blunt and rounded in dor- was translucent in life. sal aspect. Head widest where the upper Variation (in alcohol).-The dorsal col- and lower jaws meet; from there outline of or pattern of different specimens is vari- head continues straight to the level of the able, compounded by the fact that irido- eyes and then curves inward to the snout. phores and melanophores appear to be Three pairs of filamentous gills present on destroyed in some cases by initial formalin 2000] HERPETOLOGICAL MONOGRAPHS 35 fixation. Thus the holotype has faded to a other closely at their anterior ends but do very light color, and only the row of iri- not contact. The frontals are relatively nar- dophores (more accurately, their enclosing row, rounded posteriorly and overlapping light areas) along the lateral line is evident. the parietals and are pointed anteriorly The two dorsolateral rows of iridophores and overlapped by the frontal processes of may be complete and distinct, one or the the premaxillae. The partes dentales of the other may be incomplete or indistinct premaxillae are fused. The frontal pro- (e.g., TNHC 50972, 980-981), or both cesses (pars dorsalis) are well-separated may be indistinct (TNHC 55142, 55144) throughout their lengths. The premaxillae or absent (TNHC 50962). In some cases are dentate and bear 15 and 17 teeth in there is a distinct row of melanophores the two individuals from the type locality, concentrated along the lateral line with an and 11, 12, and 13 in the specimens from irregular array of iridophores on either Brushy Creek. The vomer and palatopter- side (e.g., TNHC 55139). In two cases ygoid are present, dentate, and unfused. (TNHC 54226, 55154) the dorsolateral ir- The parasphenoid lacks tooth-bearing idophores are scattered throughout the patches, and the coronoid is present and field and encompassed by starburst-shaped dentate. The elements of the hyobranchial light areas. In many cases (e.g., TNHC apparatus are primarily cartilaginous. All 53465-473, 53475-478; 55152) no dorsal specimens have the posterior end of the pattern remains, although iridophores may second basibranchial ossified and triradiate be scattered throughout the dorsal surface. (weakly ossified in TNHC 50991). Miner- Tail fins are not well-developed in two cas- alization is also present on the distal ends es (TNHC 54230, 55134). Costal grooves of ceratobranchials I-III. The distal end of are 14 (11 specimens), 15 (46 specimens), the ceratohyal is calcified only in the two or 16 (32 specimens). (Specimens exam- specimens from Stillhouse Hollow. There ined: 89; TNHC 6242; 50159; 50934-948, are 17 presacral vertebrae in four speci- 50950-951, 50953-954, 50956-959, mens (not counting the atlas), and 16 in 50961-962, 50964-965, 50967, 50972- one (TNHC 53474). The sacral diapoph- 976, 50978-981, 50987-988, 50990, yses are bicapitate in all individuals but 50992-993; 51172; 53465-473, 53475- one (TNHC 50963), in which they are 478; 54225-226, 54230-233; 55132-137, fused. The carpals and tarsals are cartilag- 55139-157). inous in all specimens. There are eight car- Osteology. -Five cleared-and-stained pals in three specimens (50963, 50991, specimens were examined; two from the 53474) and one specimen appears to have type locality (TNHC 50955, 52759), one seven carpals on one side through fusion from Krienke Spring (TNHC 50963), and of the prepollex and radiale. The cartilage two from Brushy Creek Spring (TNHC on the other side of this specimen and in 50991, 53474). Maxillae, septomaxillae, na- the other specimen from the type locality sals and prefrontals are absent. Orbito- are poorly-stained and could not be scored sphenoids are present. The frontoparietal for this feature. Four specimens were fontanelle is present and generally more scored for variation in tarsal morphology. extensive than in Eurycea chisholmensis or There are nine unfused elements present E. naufragia, and separation of the frontals in TNHC 50955 and on one side of two is greater in E. tonkawae than in the other other specimens (TNHC 50991 and two taxa. The medial border of the parie- 53474); on the other side of these two tals (forming the lateral margins of the specimens and in TNHC 50963 the num- frontoparietal fontanelle) is irregular in ber of elements is seemingly reduced to TNHC 50963, 50991, 53474, and one side eight by the fusion of tarsals four and five. of TNHC 52759 and is more straight- Phalangeal formulae are: 1-2-3-2 (hand) edged but scalloped at mid-length in and 1-2-3-3-2 (foot). 50955. The frontals are widely-separated Distribution.-Surface populations of and divergent posteriorly (but almost par- this species occur in springs of the Jolly- allel in TNHC 50991) and approach each ville Plateau region of Travis and William- 36 HERPETOLOGICAL MONOGRAPHS [No. 14 son Counties, Texas, and springs of nearby the easternmost known locality of Eurycea Brushy Creek, which drains eastward from tonkawae in Round Rock. Large quantities the Jollyville Plateau. The known range in- of foam of unknown chemical composition cludes the Brushy Creek, Bull Creek, Cy- have been observed issuing from Still- press Creek, Long Hollow Creek, and house Hollow Springs (the type locality), Walnut Creek drainages; the Spicewood and some individuals from this locality ex- Springs population provisionally assigned amined recently exhibited spinal deformi- to this species (see below) is located in the ties. City of Austin personnel have been Shoal Creek drainage. Populations that we monitoring populations of E. tonkawae assign to this species include those from since January, 1997, and have found that Balcones Community Park Spring, Barrow salamander density and abundance de- Hollow Springs, Bull Creek Spring, crease as water quality parameters associ- Brushy Creek Spring, Canyon Creek ated with urbanization increase (Beth Da- Spring, Canyon Vista Spring, Horsethief vis, personal communication). In general, Hollow Spring, Krienke Spring, Mc- the aquifers upon which this species de- Donald Well Spring (see comments be- pends are small and localized, and thus are low), New Bull Creek Spring, Schlumber- very susceptible to pollution, drying, or ger Springs, Stillhouse Hollow Spring, draining. Despite the location of some Spicewood Springs (see comments below), populations within City of Austin pre- and Wheless Spring. We also provisionally serves and the existence of what may rep- assign populations from Kretschmarr Sal- resent this species on the Travis County amander Cave, Travis Co., Testudo Tube, Audubon Sanctuary, the future of Eurycea Williamson Co., and caves of the Butter- tonkawae remains problematic. cup Creek system, Williamson Co. to this Additional comments.-We assign two species, but see comments below regard- spring populations, McDonald Well Spring ing the status of these populations. Precise and Spicewood Springs (both in Travis locations of these sites are given in Appen- Co.) to Eurycea tonkawae. Although nei- dix I. ther has yet been examined for molecular Etymology.-This species is named for markers, both occur within the known the Tonkawa tribe, who inhabited central range of E. tonkawae and appear morpho- Texas until the 1850's, when the tribe was logically similar to other specimens of this relocated to Indian Territory (now species. Specimens examined for the latter Oklahoma). The Tonkawa made camps two localities are MVZ 122705-711 around the springs that this salamander in- (McDonald Well Spring) and TNHC habits, and the stone tools and other arti- 54230-233 (Spicewood Springs). Four facts of the Tonkawa are still much in ev- specimens (TNHC 21640-643) collected idence around many of these springs. in 1956 by J. K. Baker are listed as from Tonkawa is derived from a Waco Indian "Travis: Dodd City: Jack Dies Ranch". Ac- word meaning "they all stay together," a cording to Mary Helen Bunton, who owns phrase which also describes the localiza- the property on which McDonald Well tion of Eurycea tonkawae around spring Spring is located (personal communication outflows. The name tonkawae is a noun in to AHP and PTC), the Jack Dies Ranch the genitive case (we treat Tonkawa as a was adjacent to this site (across FM 2769); latinized modern name). thus, salamanders from this locality prob- Conservation status.-Virtually the en- ably represent E. tonkawae. tire range of this species is encompassed We provisionally assign the population by a single USGS 7.5' Topographic Quad of Eurycea from Kretschmarr Salamander Map (Jollyville; ca. 143 km2). Successive Cave to E. tonkawae, based primarily on editions show this area of northwest Austin geographic distribution. This population undergoing tremendous development dur- occurs in a tiny stream cave on the Jolly- ing the last two decades, which continues ville Plateau in the vicinity of known apace to this day. For example, an office spring localities for E. tonkawae. Salaman- building was recently built directly above ders from this locality appear superficially 2000] HERPETOLOGICAL MONOGRAPHS 37 similar to animals from nearby spring lo- the Jollyville Plateau and Round Rock ar- calities. However, they are somewhat dis- eas. tinct based on allozymes and may be iso- We strongly suspect that salamanders of lated from gene flow with other the Buttercup Creek Caves system, and populations; further study is necessary. perhaps Testudo Tube, represent a distinct We provisionally assign Eurycea from evolutionary lineage, but additional study Testudo Tube Cave, and caves of the But- is needed. tercup Creek Cave system (Williamson Co.) to E. tonkawae, but note that the sys- Eurycea naufragia sp. nov. tematic status of these recently-discovered Fig. 9B cave populations is problematic. The few Eurycea neotenes: Sweet 1982:442 (in known from Testudo Tube specimens ap- part). pear similar to surface animals, and salamanders which probably represent E. ton- Holotype.-TNHC 58860 (field number kawae are known from springs on the PC 1998-10), an adult male collected by nearby Audubon Preserve (we were un- David M. Hillis and Laurie A. Dries, 14 able to collect at the Audubon locality al- August 1998 from the headsprings of Bu- though we have examined specimens in ford Hollow, a small tributary of the South the field). Testudo Tube may be separated San Gabriel River below Lake George- hydrologically from the nearby Buttercup town, 30? 39' 39" N, 97? 43' 36" W. Creek system (Russell, 1993) and thus sal- Paratypes.-TNHC 50999-51008, amanders from Testudo Tube may not be 51010-018, 51023-025, 51027-031, conspecific with those of the Buttercup 55386, 58861; MVZ 122775. Cleared-and- Creek caves. The recently discovered But- stained: TNHC 51026, 51009. tercup Creek Cave system, in the Cedar Diagnosis.-Based on external mor- Park area of Williamson Co., is a relatively phology, specimens of Eurycea naufragia extensive and probably hydrologically in- appear very similar to individuals from terconnected subterranean system (e.g., spring-dwelling populations of E. tonka- see Russell, 1993). However, the Eurycea wae. However, the light areas surrounding that inhabit these caves display consider- each dorsolateral iridophore are rosette or able variability in allozymes; this could starburst shaped in all but one of the spec- simply be an artifact of limited sampling. imens examined, unlike those of nearly all Most salamanders that have been observed specimens of E. tonkawae examined. In in the Buttercup Creek caves exhibit addition, the dark lateral margins of the strong cave-associated morphologies in- yellow dorsal tail fin coloration are regular cluding depigmentation, eye reduction, and complete, whereas in E. tonkawae and broadening and flattening of the head they usually are irregular. (personal observation and J. Reddell, per- Osteologically, Eurycea naufragia dif- sonal communication to PTC), but so few fers from Eurycea tonkawae and E. chish- specimens are available that generaliza- olmensis in having the pars dorsalis of the tions about the morphologies of these sal- premaxillae in contact or abutting medially amanders are difficult. One potential se- (although these processes are close in one quence synapomorphy unites the Testudo individual of E. chisholmensis). Eurycea Tube population with that from Ilex Cave, naufragia further differs from E. tonkawae the single member of the Buttercup Creek in having less extensive separation of the caves group for which sequence data are frontals posteriorly and a smaller fronto- available. Combined parsimony analysis of parietal fontanelle, and from E. chishol- all molecular data (Fig. 7) indicates a sister mensis in having 17 presacral vertebrae relationship between Testudo Tube and (18 in E. chisholmensis), the distal portion the Buttercup Creek Caves, with Kretsch- of the ceratohyal mineralized (not miner- marr salamander cave sister to this pair, alized in E. chisholmensis), distal tarsals 4 and this group in turn sister to spring- and 5 fused (unfused in E. chisholmensis), dwelling populations of E. tonkawae from and relatively irregular medial borders of 38 HERPETOLOGICAL MONOGRAPHS [No. 14 the parietal (relatively straight in E. chish- just behind the angle of the jaws, deep rich olmensis). red (in life), increasing in size posteriorly. Based on allozyme evidence, E. nauf- Gill insertions arranged in a stepped pat- ragia can be distinguished from other tern so that the anterior pair is farthest and northern central Texas Eurycea by posses- the posterior pair closest to the dorsal mid- sion of the unique Acon-I f allele at me- line; the anterior pair project laterally from dium allele frequency in all populations the neck and the posterior pair arch over examined except Avant's Spring and oc- the neck. Their stalks were gray (in life) currence of the unique Ck-1 c allele at me- with scattered melanophores. There are dium frequency in all populations exam- three more-or-less well-defined longitudi- ined. It is distinguished from E. tonkawae nal rows of iridophores on the body, one by high frequency of the Pep-D a allele along the lateral line and two beginning (also seen in E. chisholmensis), and from just dorsal to the anterior limb insertions E. tonkawae and E. chisholmensis by ap- and extending onto the tail. Each iridop- parent fixation of the G3pdh d allele (this hore sits at the center of a pale "rosette" allele also was seen in one individual from or starburst-shaped area, light cream- or the Buttercup Creek cave group that we flesh-colored in life, with the irregular provisionally assign to E. tonkawae). In E. margin of each defined by the unique in- naufragia from the Lake Georgetown area, teraction of its pigment and the encroach- we also found five nucleotide substitutions ing melanin. The row along the lateral line in cytochrome b that were not observed in may more appropriately be termed a field, any members of the northern clade of cen- as iridophores are dense and scattered tral Texas Eurycea, specifically T at 62, T throughout. Of the two remaining rows, at 88, G at 179, C at 213, and G at 300. the rosettes of the superior row are indis- Eurycea naufragia from springs of the tinct, especially anteriorly, whereas those Georgetown area is quite distinct from E. of the inferior row are more distinct, al- tonkawae and E. chisholmensis based on most rectangular, and almost form a con- cyt b sequence (more so than is either of tinuous longitudinal streak. Skin pigment these latter species from one another), de- ceases below the level of the lateral line, spite its geographic occurrence between and the ventral surface is immaculate and the ranges of these two species. (in life) translucent. There are 16 costal Description of holotype.-An adult male grooves, counting one each in the axilla with a gray dorsal ground color in life, ex- and groin. The dorsal tail fin, which begins cept for a light-brown streak extending at the level of the fourth caudal vertebra, posteriorly from a diamond-shaped mark is poorly developed. It is translucent along in the parietal area of the head along the the free margin and cream-colored to- midline to the base of the tail. There are wards its base, with melanophores scat- melanophores scattered throughout the tered throughout. In life, the base of the dorsal and lateral surfaces of the body, in- fin was extensively suffused with a rich cluding the limbs and digits, leading to the golden-yellow color which extended later- formation of a black, finely-reticulated, ally on to the tail itself, encompassing netlike pattern. There are a few scattered some of the iridophores there. Melano- melanophores on the palms of all four phores are densely concentrated along the feet, and along the margin of the lower lateral margins of this yellow field, forming jaw. Head broad but relatively short, wid- a distinct black border. The ventral surface est where the upper and lower jaws meet; of the tail is very finely mottled with me- snout slightly rounded and short. Eyes lanophores on either side of a weakly-de- large; iris gold in life. There are distinct veloped ventral tail fin, which begins at the melanophore concentrations forming a level of the third caudal vertebra and is black circle around each eye and a black pigmented similarly to the dorsal fin. canthal line from the anterior corner of Description of paratopotype in life.- the eye through each naris. Three pairs of An adult male (TNHC 58861) which dif- filamentous gills present on neck laterally fers from the holotype in the following 2000] HERPETOLOGICAL MONOGRAPHS 39 ways: an overall darker animal as the result frontal processes are in contact in TNHC of a greater concentration of melano- 51026 and are abutting in TNHC 51009. phores and a coarser dorsal pattern; dens- The premaxillae are dentate with 14 er melanophores on the palms of the feet (TNHC 51026) and 17 (TNHC 51009) but none on the lower jaw; iridophores teeth. The vomer and palatopterygoid are along the lateral line more distinct and in present, dentate, and unfused. The paras- a single row anteriorly; iridophores of dor- phenoid lacks tooth-bearing patches, and solateral rows distinct within their rosettes the coronoid is present and dentate. The only anteriorly; dorsal tail fin small but dis- elements of the hyobranchial apparatus tinct and beginning at level of sixth caudal are primarily cartilaginous. Both speci- vertebra; basal tail fin coloration dull or- mens have the posterior end of the second ange in life, not extending laterally on to basibranchial ossified and triradiate. Min- the tail, and lateral melanophore border eralization is also present on the distal less dense. ends of the ceratohyal and ceratobranchi- Variation (in alcohol).-The dorsal pat- als I-III, with mineralization at least three tern varies from an overall light appear- times more extensive in TNHC 51026. ance due to a paucity of melanophores There are 17 presacral vertebrae, not (e.g., TNHC 54556) to a very dense con- counting the atlas. The sacral ribs are bi- centration of melanophores with only scat- capitate. The carpals and tarsals are carti- tered iridophores present in no apparent laginous in all specimens. There are eight order (e.g., TNHC 51014, 54560). The carpals and eight tarsals; the number of dark eye ring remains in specimens of the tarsals is seemingly reduced to eight by the former type. One specimen (TNHC fusion of tarsals 4 and 5. Phalangeal for- 54558) has the margins of iridophore fields mulae are: 1-2-3-2 (hand) and 1-2-3-3- forming more regular ovoid or rectangular 2 (foot). outlines rather than a rosette or starburst Distribution.-This species is known shape. Costal grooves are 14 (5), 15 (16), only from springs and possibly one cave in or 16 (3). (Specimens examined: 24; Williamson County, Texas, associated with TNHC 50999-51005, 51007-008, 51010- drainages of the south, middle and north 017, 51023-025, 51027-028, 51030; forks of the San Gabriel river. We also pro- 55386). Of eight additional specimens visionally assign populations from the from Cowan Creek Spring (TNHC 54555- Cowan Creek drainage to this species. 562), tentatively assigned to this species, Cowan Creek drains into Berry Creek, three possess 15 costal grooves and five which drains into the San Gabriel River possess 16. below the city of Georgetown. The popu- Osteology. -Two cleared-and-stained lation from Bat Well, which we also pro- specimens were examined, TNHC 51009 visionally assign to this species, is located and 51026. Maxillae, septomaxillae, nasals in the Berry Creek drainage. and prefrontals are absent. Orbitosphe- Etymology.-The name of this species noids are present. The frontoparietal fon- is derived from the Latin naufragium, tanelle is present. The medial border of which means "remnants" or "remains." the parietals is irregular with short, finger- The name refers to the springs in which like projections extending into the fronto- this species lives, several of which have parietal fontanelle. The frontals diverge been destroyed or are threatened by hu- slightly posteriorly but are nearly parallel man development. for most of their lengths. They are in me- Conservation status.-Based on our ob- dial contact anteriorly in TNHC 51026 servations, the populations of this species and almost in contact in TNHC 51009. within the City of Georgetown proper The frontals are rounded posteriorly and probably are on the brink of extirpation. overlapping the parietals and are pointed The recently discovered populations on anteriorly and overlapped by the frontal Cowan Creek that we provisionally assign processes of the premaxillae. The partes to this species lie within Sun City George- dentales of the premaxillae are fused. The town, a new leisure and retirement com- 40 HERPETOLOGICAL MONOGRAPHS [No. 14 munity designed to accommodate 9,000 amanders from Salado Springs (although homes (about 18,000 people) at build-out. in other respects they appear to be speci- It is our understanding that management mens of E. naufragia); the others appear of Sun City Georgetown are aware of the in all respects (based on external mor- springs within their tract, and have no phology) to be specimens of E. naufragia. plans to physically impact them. The sev- Given this situation, and the problematic eral springs harboring Eurycea naufragia placement of the Bat Well population, fur- along the Middle Fork of the San Gabriel ther study of these and other populations River are in the immediate vicinity and of salamanders in the region between just downstream of a large quarrying op- Georgetown and Salado is needed to verify eration on leased land. The lessee, Capitol their taxonomic status. Aggregates, has consulted with one of us (AHP) after commissioning separate hy- Eurycea chisholmensis sp. nov. drogeological and salamander surveys of Fig. 9C the area, and it seems probable under cur- Eurycea neotenes: Sweet 1982:441 (in rent conditions that the quarrying activity part). will not jeopardize recharge and spring- flow. We believe there probably are undis- Holotype.-TNHC 58859 (field number covered localities for E. naufragia within PC 1998-9), an adult male collected by the San Gabriel River watershed west of Paul T. Chippindale on 1 August 1998, in Georgetown; therefore it is difficult to pro- approximately 10 cm of water at side ject the conservation status of this species. spring immediately adjacent to Main (= The realizable range of E. naufragia prob- Salado, Big Boiling, or Siren) Springs, Sal- ably is no greater than that of the Jollyville ado, Bell Co., Texas, 30? 56' 37" N, 97? 32' Plateau Salamander, however, and the de- 31" W., at approximately 1030 h. velopment pressure on the area that its Paratypes.-TNHC 51139-146. Cleared- range encompasses likely will soon catch and-stained: TNHC 52770-771. up with that around Austin. We assume Diagnosis. -Eurycea chisholmensis gen- that salamanders reported by Sweet erally exhibits reduced eyes in comparison (1978a, 1984) from springs in the city park to other spring-dwelling north central Tex- at Georgetown are conspecific with those as Eurycea (see Fig. 10). The dorsolateral that we examined from nearby springs in body fields lack well-defined iridophores the vicinity of Lake Georgetown. A single or melanophores, in contrast to E. tonka- individual that we found at the George- wae and E. naufragia. The dark eye ring town Park site in 1991 was a tiny juvenile seen in most individuals of E. naufragia that failed to yield successful results for and E. tonkawae is absent, and the upper key molecular markers. Here we provi- lip between the posterior border of the eye sionally assign salamanders from the Bat and the naris lacks dark pigment altogeth- Well and Cowan Creek Springs popula- er. Dorsal coloration typically is dark, with tions to E. naufragia, based largely on geo- a series of fine, lighter reticulations; in graphic proximity. The single individual contrast, other surface-dwelling north cen- available from Bat Well differs with re- tral Texas Eurycea usually exhibit varying spect to some molecular markers from densities of scattered melanophores on a other E. naufragia examined, although in lighter background. Dark lateral margins most respects it appears similar morpho- of yellow dorsal tail fin coloration are ab- logically. The Cowan Creek Spring popu- sent. lation was discovered late in the course of Morphometric differentiation between this study, and animals from this locality E. chisholmensis and the other new north- have not yet been examined for key mo- ern species with respect to the ratio of eye lecular markers. Eight specimens are avail- diameter to axilla-groin length is illustrated able (TNHC 54555-562). Two of these ex- in Fig. 10. hibit eye diameter: axilla-groin length Eurycea chisholmensis differs osteolog- ratios within the range of those seen in sal- ically from E. tonkawae and E. naufragia 2000] HERPETOLOGICAL MONOGRAPHS 41

S 1.5 0 0 0 * 0 1.3 0.. S ? E e? o- _

E 1.1 .0 0 0. . .J m m 0 0.90 I

U 0.70 12 14 16 18 20 22

Axilla-groinlength (mm) FIG. 10.-The relationship between eye diameter and axilla-groin length in Eurycea tonkawae (solid circles), E. naufragia (open circles), and E. chisholmensis (grey squares). in having 18 presacral vertebra (as op- Eurycea examined, and it appears to lack posed to the usual 17 in E. tonkawae and the unique sequence substitutions that E. naufragia; one individual of E. tonka- characterize E. naufragia, spring popula- wae examined has 16), and by the absence tions of E. tonkawae, or cave populations of mineralization on the distal portion of provisionally assigned to E. tonkawae. the ceratohyal. Eurycea chisholmensis fur- Description of holotype. -An adult male ther differs from E. naufragia in having (TNHC 58859). Eyes are reduced and distal tarsals 4 and 5 unfused (fused in E. head is flat; in lateral profile eyes do not naufragia), relatively straight medial bor- project above the level of the head. Head ders of the parietal (irregular in E. nauf- is relatively long, and snout is blunt and ragia), and medial separation of the pars rounded in dorsal aspect. Head is broadest dorsalis of the premaxilla (in contact or where the two jaws meet, and narrows abutting in E. naufragia). Eurycea chish- very gradually to the level of the eyes and olmensis further differs from E. tonkawae then slightly more to the snout. There are in having less extensive separation of the three pairs of filamentous gills on the neck frontals posteriorly and a smaller fronto- laterally just behind the angle of the jaws; parietal fontanelle. these were reddish-brown (in life) and in- Based on allozyme evidence, Eurycea crease in size posteriorly; the most poste- chisholmensis can be distinguished from rior arch slightly over the neck. Gill inser- E. tonkawae and nearly all other northern tions are arranged in a stepped pattern so central Texas Eurycea provisionally as- that the anterior pair are farthest and the signed to this species by apparently fixed posterior pair closest to the dorsal midline. occurrence of the Pep-D a allele. This spe- Dorsal coloration nearly uniform grayish cies appears to lack any of the unique al- brown in life with a slight cinnamon tinge, leles found in E. naufragia, the nearest with occasional tiny, lighter flecks present taxon geographically. One nucleotide sub- at irregular intervals. Under magnification, stitution in cytochrome b (T at 252) also dorsal pigmentation appears as a series of distinguishes this taxon from other mem- fine reticulations. Laterally, the coloration bers of the northern clade of central Texas becomes a fine speckling on a cream (in 42 HERPETOLOGICAL MONOGRAPHS [No. 14 life) background, and the ventral surface is anterior pair project laterally from the largely translucent. Large, dark testes were neck but the proximal pair to the midline readily visible through the skin of the ven- do not arch over the neck. A dark parietal tral surface in life. Lateral iridophores are diamond-shaped mark is present but in- nearly indistinguishable and are uncount- distinct, as are the canthal lines connecting able; there is an ill-defined, faint series of the anterior corners of the eyes with the marks slightly paler than the surrounding nares. Dorsolateral fields of the trunk areas (this may correspond to iridophores) light-brown with faint light lines in life, present between the front and hind legs. and lack obvious iridophores or melano- Front and hind legs cream-colored in life, phores. The dorsal tail fin is well-devel- densely speckled with brown above; the oped posteriorly but small anteriorly, and undersides of the front legs are nearly free ends abruptly before it reaches the vent. of this speckling, and it is restricted pri- The ventral tail fin is weakly developed marily to the bases of the undersides of the and occupies only the posterior one-third hind legs. The top of the head has widely of the tail. Both tail fins translucent in life scattered flecks, cream-colored in life; the except for a scattering of melanophores. area immediately surrounding the gill in- Total circumference of the surface of the sertions was pink in life. The undersides anterior half of the tail yellow in life, but of the eyes are outlined in cream; this is this area lacks a defining border of any not visible from directly above because the kind. dorsal portions of the eyes are partly cov- Variation (in alcohol).-The degree of ered with skin. Iris of eye in life pale gold patterning of white lines within the brown with dark flecking. Very faint brown can- to gray dorsolateral fields is variable, from thal lines are visible but indistinct. The up- very few (TNHC 51141) to extremely per labial region lacks dark pigmentation dense (TNHC 51145). There may occa- from the posterior margin of the eye to the sionally be a few larger, irregularly-shaped, naris, while the lower labial region exhibits light areas superimposed on this basic pat- slight dark speckling at the posterior edge, tern (TNHC 51146). The dark canthal decreasing anteriorly. A prominent gold line, which is usually poorly defined, may stripe was present in life on the dorsal part be absent (TNHC 51143). The dorsal tail of the tail, narrow at the base and widen- fin may be essentially absent (TNHC ing toward the middle; by about 2/3 of the 51142). Costal grooves are 15 (5 speci- way along the tail, the gold color became mens) or 16 (2 specimens). (Specimens ex- increasingly fragmented, changing to amined: 7; TNHC 51139, 51141-146). speckling along the caudal fin on the pos- Osteology. -Two cleared-and-stained terior part of the tail. Sides of the tail were specimens were examined: TNHC 52770 also speckled with yellow in life. In life, and 52771. Maxillae, septomaxillae, nasals the underside of the tail exhibited a pale and prefrontals are absent. Orbitosphe- yellow line, extending from the posterior noids are present but are very small in opening of the vent along and onto the TNHC 52770. The frontoparietal fonta- ventral caudal fin. Both dorsal and ventral nelle is present. The medial border of the caudal fins are well developed on the pos- parietals is more-or-less straight and reg- terior portion of the tail. ular, but less so in TNHC 52771. The Description of paratopotype.-An adult, frontals are in contact medially near their sex uncertain (TNHC 51146) with a drab articulation with the premaxillae in TNHC ground color in life (color 27 of Smithe, 52771, and are slightly separated in TNHC Naturalist's Color Guide, American Mu- 52770. They are parallel for most of their seum of Natural History). There is a light lengths in TNHC 52770 and are parallel ring around each eye due to the absence for their anterior half in TNHC 52771. of melanophores in this area. Descriptions The frontals are rounded posteriorly and of head and gills are as for the holotype, overlapping the parietals and are pointed except that in this specimen the gills are and somewhat irregular anteriorly and reduced and were bright red in life; the overlapped by the frontal processes of the 20001200IEPTLGCA HERPETOLOGICAL MONOGRAPHSOORPS4 43 premaxillae. The partes dentales of the modified to some extent during the past premaxillae are fused. The frontal pro- 150 years, and the type locality is on the cesses are separate but in close proximity south bank of Salado Creek in a municipal in TNHC 52770. The premaxillae are den- park. Several groundwater contamination tate with 17 (TNHC 52771) and 18 incidents have occurred in the recent past (TNHC 52770) teeth. The vomer and pal- (Price et al., 1995) and the potential for atopterygoid are present, dentate, and un- more still exists. As we have suggested, fused. The parasphenoid lacks tooth-bear- however, additional localities may yet be ing patches, and the coronoid is present discovered, and we believe it is still pos- and dentate. The elements of the hyobran- sible to plan and implement conservation chial apparatus are primarily cartilaginous. measures for this species. Both specimens have the posterior end of Additional comments.-This is the the second basibranchial ossified and tri- northeasternmost known population of radiate. TNHC 52771 also has limited Eurycea in the Edwards Plateau region, mineralization near the posterior end of and salamanders from these springs are ceratobranchials I and III. There are 18 very elusive. Prior to this study, only a sin- presacral vertebrae, not counting the atlas. gle juvenile specimen was known (private The two heads and corresponding di- collection of B. C. Brown, Baylor Univer- apophyses of the presacral ribs are fused sity); we obtained most of the known spec- but still distinguishable in TNHC 52770 imens in 1989-1991 but (despite more whereas in TNHC 52771 two heads are than 20 additional visits to the type locality distinct on one side and completely fused between 1991 and 1998) found no others on the other. The carpals and tarsals are until 1998, when the holotype was collect- cartilaginous in all specimens, and there ed. are eight carpals and nine tarsals. Phalan- geal formulae are: 1-2-3-2 (hand) and 1- CONCLUSIONS 2-3-3-2 (foot). Many problems remain in the the taxo- Distribution.-This species is known nomic allocation of the central Texas Eu- only from Big Boiling (= Main, Salado, or rycea. This is not surprising, given the Siren) Springs and Robertson Springs at large number of populations, extreme geo- Salado, Bell Co., Texas. Recently, salaman- graphic fragmentation, and the complex ders that could represent this species have mixture of morphological conservatism been found in springs of nearby Butter- and apparent parallelism or convergence. milk Creek (G. Longley, personal com- Our findings are consistent with a general munication to PTC) but specimens are not pattern in plethodontid salamanders: in al- yet available for examination. most every wide-ranging "species" that has Etymology. -Jesse Chisholm laid out a been studied using molecular techniques, trail from the Canadian River in Indian numerous morphologically cryptic species Territory to Wichita, Kansas in 1865, and have been revealed (reviewed by Larson started driving cattle up the trail in 1866. and Chippindale, 1993). We see the work By 1867, the Chisholm Trail had been ex- described here as the basis for more de- tended south into Texas, and Salado was tailed studies of relationships and species an important stop on the trail because of boundaries within and among the taxa that its clean, clear springs, which are also the we have recognized. The identification of habitat of this species. The name chishol- the major monophyletic groups in this as- mensis is an adjective referring to the semblage will facilitate future systematic Chisholm Trail. and taxonomic work, as it will be possible Conservation status.-Determining the to focus on particular subsets of the Texas conservation status of this species is prob- Eurycea that represent historical (mono- lematical, given the difficulty with which phyletic) units. Previously, for example, a specimens have been acquired and lack of study of "E. neotenes" would necessarily knowledge of the extent of its range. Most have had to encompass the whole region of the spring outlets at Salado have been and what have proven to be numerous dis- 44 HERPETOLOGICAL MONOGRAPHS ENo.[No. 14 tinct species. Future work on the group LITERATURE CITED should include both morphological study and use of molecular ABBOTT,P. L. 1975. On the hydrology of the Edwards rapidly-evolving Limestone, south-central Texas. Journal of Hydrol- markers, and such work is in progress by ogy 24:251-269. the authors of this paper. We are optimis- ABBOTT,P. L., AND C. M. WOODRUFF,JR. (Eds). tic that further study will clarify many of 1986. The Balcones Escarpment: Geology, Hydrol- the systematic and taxonomic problems ogy, Ecology and Social Development in Central Texas. Geological Survey of America, San Diego, that remain in the Texas Eurycea, and California, U.S.A. hope that it will be possible to characterize BAKER,J. K. 1957. Eurycea troglodytes: a new blind and preserve the diversity in the group be- cave salamander from Texas. Texas Journal of Sci- fore much of it is lost due to human activ- ence 9:328-336. ities. . 1961. Distribution of and key to the neotenic Eurycea of Texas. Southwestern Naturalist 6:27-32. BEHLER, L., AND F. W. KING. 1979. The Audubon contributed to J. Acknowledgments.-Many people Field Guide to North American Reptiles the success of this Sam Sweet's detailed stud- Society project. and Alfred A. Inc., New York, ies of the central Texas and the information Amphibians. Knopf, Eurycea New York, U.S.A. that he to us formed much of the founda- provided BERLOCHER,S. H., AND D. L. SWOFFORD.1997. tion of this Grubbs' study. Andy caving expertise for trees under the fre- made it to visit subterranean habitats Searching phylogenetic possible many criterion: an us- that otherwise would have been inaccessible. quency parsimony approximation James 46: Reddell was a rich source of ing generalized parsimony. Systematic Biology knowledge regarding 211-215. central Texas salamanders. George Veni provided BIRKY,C. W, P. FUERST, AND T. MARUYAMA. many important insights into Edwards Plateau cave JR., 1989. Organelle gene diversity under migration, hydrogeology and arranged access to key caves. Bill mutation, and drift: equilibrium expectations, ap- Russell shared his knowledge of Texas caves and pro- proach to equilibrium, effects of heteroplasmic vided important specimens. David Cannatella, Mark cells, and to nuclear genes. Genetics Kirkpatrick, Tim Rowe, and Bull provided key comparison Jim 121:613-627. advice throughout this project. David Wake allowed BIRKY,C. W, T. MARUYAMA,AND P. FUERST. us to examine specimens from the herpetological col- JR., 1983. An to and lection of the Museum of Vertebrate Zoology at The approach population evolutionary for in mitochondria and chlo- University of California at Berkeley, and provided key genetic theory genes and some results. Genetics 103:513-527. tissue samples as well as insightful ideas. Others who roplasts, S. C. 1941. Notes on salamanders with de- contributed to the project by providing specimens, BISHOP, of several new forms. Occasional access to important localities, field companionship, or scriptions Papers valuable information include Bill Elliott, Lisa of the Museum of Zoology, University of Michigan 451:1-21. O'Donnell, Landon Lockett, Doyle Mosier, Robert 1943. Handbook of Salamanders: the Sala- Hansen, Glenn Longley, and Joel Hunter. We also . thank the numerous central Texas landowners who manders of the United States, of Canada, and of allowed us to search for salamanders on their prop- Lower California. Cornell University Press, Ithaca, erties; only once were we denied access to a sala- New York, U.S.A. mander locality. Wayne Van Devender, Paul Moler, BISHOP, S. C., AND M. R. WRIGHT.1937. A new ne- Robert Wilkinson, and Robert Miller provided spec- otenic salamander from Texas. Proceedings of the imens of many key outgroup taxa. Jewel Husti was an Biological Society of Washington 50:141-143. invaluable assistant in the laboratory. Marty Badgett BOGART, J. P. 1967. Life History and Chromosomes facilitated many aspects of the lab work. We thank of Some of the Neotenic Salamanders of the Ed- David Swofford for access to a test version of PAUP*. ward's Plateau. M.A. Thesis, University of Texas, Cynthia Chippindale assisted in many aspects of man- Austin, Texas, U.S.A. uscript preparation and was a great source of moral BREMER, K. 1988. The limits of amino acid sequence support. Brad Shaffer and Steve Tilley provided very data in angiosperm phylogenetic reconstruction. helpful reviews of an earlier version of this manu- Evolution 42:795-803. script. Parts of this work were funded by a Cave Re- BROWN, B. C. 1942. Notes on Eurycea neotenes. search Foundation Karst Fellowship to PTC, an NSF Copeia 1942:176. Dissertation Improvement Grant to PTC, and Sec- . 1950. An Annotated Check List of the Rep- tion 6 cooperative funding from the US Fish and tiles and Amphibians of Texas. Baylor University Wildlife Service and Texas Parks and Wildlife De- Press, Waco, Texas, U.S.A. partment. PTC also received funding for personal . 1967a. Eurycea latitans. Catalogue of Amer- support in the form of an NSERC postgraduate fel- ican Amphibians and Reptiles 34.1-34.2. lowship and several scholarships from the University . 1967b. Eurycea nana. Catalogue of Ameri- of Texas at Austin Department of Zoology and Grad- can Amphibians and Reptiles 35.1-35.2. uate School. Support to JJW during early phases of . 1967c. Eurycea neotenes. Catalogue of this work was provided by an NSF Graduate Fellow- American Amphibians and Reptiles 36.1-36.2. ship. BRUCE, R. C. 1976. Population structure, life history 2000] HERPETOLOGICAL MONOGRAPHS 45

and evolution of paedogenesis in the salamander clearing of alcian blue stained whole small verte- Eurycea neotenes. Copeia 1976:242-249. brates for demonstration of cartilage. Stain Tech- BURGER,W. L., H. M. SMITH,AND F. E. POTTER,JR. nology 52:229-232. 1950. Another neotenic Eurycea from the Edwards DIXON, J. R. 1987. Amphibians and Reptiles of Texas. Plateau. Proceedings of the Biological Society of Texas A&M University Press, College Station, Tex- Washington 63:51-58. as, U.S.A. BUTH, D. G. 1984. The application of electrophoretic DOWLING,T. E., C. MORITZ,AND J. D. PALMER. data in systematic studies. Annual Review of Ecol- 1990. Nucleic acids II: restriction site analysis. Pp. ogy and Systematics 15:501-522. 250-317. In D. M. Hillis and C. Moritz (Eds.), Mo- CHIPPINDALE,P. T. 1989. A high pH discontinuous lecular Systematics. Sinauer Associates, Sunder- buffer system for resolution of isozymes in starch land, Massachussetts, U.S.A. gel electrophoresis. Stain Technology 64:61-64. EMERSON, E. T., II. 1905. General anatomy of Ty- . 1995. Evolution, Phylogeny, Biogeography, phlomolge rathbuni. Proceedings of the Boston So- and Taxonomy of Central Texas Cave and Spring ciety of Natural History 32:42-76. Salamanders (Eurycea and Typhlomolge). Ph.D. FELSENSTEIN, J. 1985. Confidence limits on phylog- Dissertation, University of Texas, Austin, Texas, enies: an approach using the bootstrap. Evolution U.S.A. 39:783-791. CHIPPINDALE,P. T., A. H. PRICE,AND D. M. HILLIS. FROST, D. R. 1985. Amphibian Species of the World: 1993. A new species of perennibranchiate salaman- A Taxonomic and Geographical Reference. Allen der (Eurycea, Plethodontidae) from Austin, Texas. Press, Inc., and The Association of Systematics Col- Herpetologica 49:248-259. lections, Lawrence, Kansas, U.S.A. CHIPPINDALE,P. T., A. H. PRICE,AND D. M. HILLIS. FROST, D. R., AND D. M. HILLIS. 1990. Species in 1998. status of the San Marcos sala- Systematic concept and practice: herpetological applications. mander, Eurycea nana (Caudata: Plethodontidae). Herpetologica 46:87-104. Copeia 1998:1046-1049. GOOD, D. A., AND D. B. WAKE. 1992. Geographic CHIPPINDALE,P. T., AND J. J. WIENS. 1994. Weight- variation and speciation in the torrent salamanders ing, partitioning, and combining characters in phy- of the genus Rhyacotriton (Caudata: Rhyacotriton- logenetic analysis. Systematic Biology 43:278-287. idae). University of California Publications in Zo- 1961. of COCHRAN,D. M. Type Specimens Reptiles ology 126:xii + 91 p. and in the U.S. National Museum. Amphibians GRAYBEAL, A. 1993. The phylogenetic utility of cy- Bulletin of the U.S. National Museum 220:xv + tochrome b: lessons from bufonid frogs. Molecular 291 p. Phylogenetics and Evolution 2:256-269. CONANT,R. 1958. A Field Guide to and Am- Reptiles HAMILTON, A. L. 1973. Some Taxonomic Aspects of phibians of the United States and Canada East of Certain Paedogenetic Eurycea of the Blanco River the 100th Meridian. Houghton Mifflin Co., Boston, Drainage System in Hays and Blanco Counties, Massachusetts, U.S.A. Texas. M.S. Thesis, Southwest Texas State Univer- . 1975. A Field Guide to Reptiles and Am- sity, San Marcos, Texas, U.S.A. phibians of Eastern and Central North America. HIGHTON, R. 1990. Taxonomic treatment of geneti- Second Edition. Mifflin Co., Boston, Houghton differentiated 46: Massachussetts, U.S.A. cally populations. Herpetologica 114-121. CONANT,R., AND T. COLLINS.1991. A Field Guide J. . 1995. Speciation in eastern North American to Reptiles and Amphibians: Eastern and Central salamanders of the Plethodon. Annual Re- North America. Houghton Mifflin, Boston, Mas- genus view of and 26:579-600. sachussetts, U.S.A. Ecology Systematics HIGHTON,R., G. C. MAHA,AND L. R. MAXSON.1989. CRACRAFT, 1989. Speciation and its ontology: the J. Biochemical evolution in the salamanders of empirical consequences of alternative species con- slimy the Plethodon in the eastern cepts for understanding patterns and processes of glutinosus complex United States. of Illinois differentiation. Pp. 28-59. In D. Otte and J.A. En- University Biological 57:1-153. dler (Eds.), Speciation and Its Consequences. Sin- Monograph auer Associates, Sunderland, Massachussetts, HILLIS, D. M. 1984. Misuse and modification of Nei's U.S.A. genetic distance. Systematic Zoology 33:238-240. CROTHER,B. I. 1990. Is "some better than none" or . 1995. Approaches for assessing phylogenetic do allele frequencies contain phylogenetically use- accuracy. Systematic Biology 44:3-16. ful information? Cladistics 6:277-281. HILLIS, D. M., AND J. J. BULL. 1993. An empirical CUNNINGHAM,C. W. 1997. Is congruence between test of bootstrapping as a method for assessing con- data partitions a reliable predictor of phylogenetic fidence in phylogenetic analysis. Systematic Biology accuracy? Empirically testing an iterative proce- 42:182-192. dure for choosing among phylogenetic methods. HILLIS, D. M., J. P. HUELSENBECK, AND C. W. CUN- Systematic Biology 46:464-478. NINGHAM. 1994. Application and accuracy of mo- DE QUEIROZ,A., M. J. DONOGHUE,AND J. KIM. lecular phylogenies. Science 264:671-677. 1995. Separate versus combined analysis of phylo- HILLIS, D. M., B. K. MABLE,A. LARSON,S. K. DA- genetic evidence. Annual Review of Ecology and VIS, AND E. A. ZIMMER.1996. Nucleic acids IV: Systematics 26:657-681. sequencing and cloning. Pp. 321-381. In D. M. DINGERKUS,G., AND L. D. UHLER. 1977. Enzyme Hillis, C. Moritz, and B. K. Mable (Eds.), Molec- 46 HERPETOLOGICAL MONOGRAPHS [No. 14

ular Systematics, 2nd ed. Sinauer Associates, Sun- MCKNIGHT, M. L., C. K. DODD, JR., AND C. M. derland, Massachussetts, U.S.A. SPOLSKY.1991. Protein and mitochondrial DNA HOLSINGER, J. R., AND G. LONGLEY. 1980. The sub- variation in the salmander Phaeognathus hubrichti. terranean amphipod crustacean fauna of an arte- Herpetologica 47:440-447. sian well in Texas. Smithsonian Contributions in MEYER,A. 1994. Shortcomings of the cytochrome b Zoology 308:iii + 62 p. gene as a molecular marker. Trends in Ecology and HUELSENBECK, J. P. 1995. Performance of phyloge- Evolution 9:278-280. netic methods in simulation. Systematic Biology 44: MITCHELL,R. W., AND J. R. REDDELL. 1965. Eury- 17-48. cea tridentifera, a new species of troglobitic sala- HUELSENBECK, J. P., J. J. BULL, AND C. W. CUN- mander from Texas and a reclassification of Ty- NINGHAM. 1996. Combining data in phylogenetic phlomolge rathbuni. Texas Journal of Science 17: analysis. Trends in Ecology and Evolution 11:152- 12-27. 158. MITCHELL,R. W., AND R. E. SMITH.1972. Some as- JACOBS, J. F. 1987. A preliminary investigation of geo- pects of the osteology and evolution of the neotenic graphic genetic variation and systematics of the spring and cave salamanders (Eurycea, Plethodon- Two-lined Salamander, Eurycea bislineata (Green). tidae) of central Texas. Texas Journal of Science 23: Herpetologica 43:423-446. 343-362. JOHNS, G. C., AND J. C. AVISE. 1998. A comparative MIYAMOTO,M. M., AND W. M. FITCH. 1995. Testing summary of genetic distances in the vertebrates species phylogenies and phylogenetic methods with fom the mitochondrial cytochrome b gene. Molec- congruence. Systematic Biology 44:64-76. ular and Evolution 15:1481-1490. Biology MORITZ,C., C. J. SCHNEIDER,AND D. B. WAKE. KORNET,D. J., AND H. TURNER.1999. Coding poly- 1992. Evolutionary relationships within the Ensa- morphism for phylogeny reconstruction. Systematic tina eschscholtzii complex confirm the ring species Biology 48:365-379. interpretation. Systematic Biology 41:273-291. LARSON,A., AND P. T. CHIPPINDALE.1993. Molecular MURPHY,R. W. 1993. The phylogenetic analysis of to the of approaches evolutionary biology pletho- allozyme data: invalidity of coding alleles by pres- dontid salamanders. Herpetologica 49:204-215. ence/absence and recommended procedures. Bio- H. 1949. On a in LEVENE, matching problem arising chemical Systematics and Ecology 21:25-38. Annals of Mathematics and Statistics 20: genetics. MURPHY, R. W, AND K. D. DOYLE. 1998. Phylophe- 91-94. netics: frequencies and polymorphic characters in LEVITON,A. E., R. H. GIBBS, E. HEAL, AND C. JR., genealogical estimation. Systematic Biology 47: E. DAWSON. 1985. Standards in herpetology and 737-761. I. Standard codes for in- ichthyology: part symbolic MURPHY,R. W, W. SITES, D. G. BUTH, AND stitutional resource collections in and J. JR., herpetology C. H. HAUFLER.1996. Proteins: isozyme electro- ichthyology. Copeia 1985:802-832. phoresis. Pp. 51-120. In D. M. Hillis, C. Moritz, LICHT, L. E., AND L. A. LOWCOCK.1991. Genome and B. K. Mable (Eds.), Molecular Systematics, size and metabolic rate in salamanders. Compara- 2nd ed. Sinauer Associates, Sunderland, Massa- tive Biochemistry and Physiology 100B:83-92. chussetts, U.S.A. LOMBARD,R. E., AND D. B. WAKE. 1986. Tongue NEI, M. 1972. Genetic distance between populations. evolution in the lungless salamanders, family Pleth- American Naturalist 106:283-292. odontidae. IV. Phylogeny of plethodontid salaman- . 1978. Estimation of ders and the evolution of feeding dynamics. Sys- average heterozygosity and distance from a small number of in- tematic Zoology 35:532-551. genetic dividuals. Genetics 89:583-590. LONGLEY, G. 1978. Status of Typhlomolge (= Eury- cea) rathbuni, the Texas Blind Salamander. Endan- PATERSON,H. E. H. 1985. The recognition concept of 21-29. In E. S. Vrba gered Species Report (2):vi + 45 p. U.S. Fish and species. Pp. (Ed.), Species Wildlife Service Region 2, Albuquerque, New and Speciation. Transvaal Museum Monograph No. Mexico, U.S.A. 4, Pretoria, South Africa. . 1986. The biota of the Edwards Aquifer and PETRANKA,J. W. 1998. Salamanders of the United implications for paleozoogeography. Pp. 51-54. In States and Canada. Smithsonian Institution Press, and London. P. L. Abbott, and C. M. Woodruff, Jr. (Eds), The Washington Balcones Escarpment: Geology, Hydrology, Ecolo- POTTER,F. E., JR. 1963. Gross Morphological Varia- gy and Social Development in Central Texas. Geo- tion in the Genus Typhlomolge with a Description logical Survey of America, San Diego, California, of a New Species. M.A. Thesis, University of Texas, U.S.A. Austin, Texas, U.S.A. MADDISON,D. R. 1991. The discovery and impor- POTTER,F. E., JR., AND S. S. SWEET.1981. Generic tance of multiple islands of most-parsimonious boundaries in Texas cave salamanders, and a re- trees. Systematic Zoology 40:315-328. description of Typhlomolge robusta (Amphibia: MADDISON,W P., AND D. R. MADDISON.MacClade, Plethodontidae). Copeia 1981:64-75. version 3.05. Sinauer Associates, Sunderland, Mas- PRICE, A. H., P. T. CHIPPINDALE, AND D. M. HILLIS. sachussetts, U.S.A. 1995. A status report on the threats facing popu- MAYR,E. 1982. The Growth of Biological Thought. lations of perennibranchiate hemidactyliine pleth- Harvard University Press, Cambridge, Massachus- odontid salamanders of the genus Eurycea north of setts, U.S.A. the Colorado River in Texas. Final Section 6 report 2000] HERPETOLOGICAL MONOGRAPHS 47

part III, Texas Parks and Wildlife Department, the Texas Eurycea (Amphibia: Plethodontidae). Austin, Texas, U.S.A. Ph.D. Dissertation, University of California, Berke- RICE, W. R. 1989. Analyzing tables of statistical tests. ley, California, U.S.A. Evolution 43:223-225. . 1978b. On the status of Eurycea pterophila RODRIGO,A. G. 1992. A modification to Wheeler's (Amphibia: Plethodontidae). Herpetologica 34: combinatorial weights calculations. Cladistics 8: 101-108. 165-170. . 1982. A distributional analysis of epigean ROGERS,J. S. 1972. Measures of genetic similarity populations of Eurycea neotenes in central Texas, and genetic distance. Studies in Genetics VII. Uni- with comments on the origin of troglobitic popu- versity of Texas Publication 7213:145-153. lations. Herpetologica 38:430-444. ROUTMAN,E. 1993a. Mitochondrial DNA variation . 1984. Secondary contact and hybridization in in Cryptobranchus alleganiensis, a salamander with the Texas cave salamanders Eurycea neotenes and extremely low allozyme diversity. Copeia 1993:407- E. tridentifera. Copeia 1984:428-441. 416. . 1986. Caudata. Pp. 734-740. In L. Botosa- . 1993b. Population structure and genetic di- neanu (Ed.), Stygofauna Mundi. Leiden, E.J. Brill/ versity of metamorphic and paedomorphic popu- Dr. W. Backhuys. lations of the tiger salamander, Ambystoma tigrin- SWOFFORD,D. L. 1999. PAUP*: Phylogenetic Anal- um. Journal of Evolutionary Biology 6:329-357. ysis Using Parsimony and Other Methods, version RUSSELL,W. H. 1976. Distribution of troglobitic sal- 4.0d55. Smithsonian Institution, Washington, D.C., amanders in the San Marcos area, Hays Co., Texas. U.S.A Report 7601, Texas Association for Biological In- SWOFFORD,D. L. 1999. PAUP*: Phylogenetic Anal- vestigation of Troglobitic Eurycea, Austin, Texas, ysis Using Parsimony and Other Methods, version U.S.A. 4.0b2. Sinauer Associates, Sunderland, Massachus- RUSSELL,W. H. 1993. The Buttercup Creek Karst, setts, U.S.A. Travis and Williamson Counties, Texas: Geology, SWOFFORD,D. L., AND S. H. BERLOCHER.1987. In- Biology, and Land Development. Report to the ferring evolutionary trees from gene frequency data University Speleological Society, University of Tex- under the principle of maximum parsimony. Sys- as, Austin, Texas, U.S.A. tematic Zoology 36:293-325. SCHMIDT,K. P. 1953. A Check List of North Amer- SWOFFORD,D. L., G. J. OLSEN, P. J. WADDELL,AND ican Amphibians and Reptiles, 6th ed. American D. M. HILLIS.1996. Phylogeny inference. Pp. 407- Society of Ichthyologists and Herpetologists, Uni- 514. In D. M. Hillis, C. Moritz, and B. K. Mable versity of Chicago Press, Chicago, Illinois, U.S.A. (Eds.), Molecular Systematics, 2nd ed. Sinauer As- SHAFFER,H. B., AND F. BREDEN. 1989. The rela- sociates, Sunderland, Massachussetts, U.S.A. tionship between allozyme variation and life his- SWOFFORD,D. L., AND R. B. SELANDER. 1981. tory: non-transforming salamanders are less vari- BIOSYS-1: a Fortran program for the comprehen- able. Copeia 1989:1016-1023. sive analysis of electrophoretic data in population SLADE,R. S., JR., M. E. DORSEY,AND S. L. STEWART. genetics and systematics. Journal of Heredity 72: 1986. Hydrology and water quality of the Edwards 281-283. Aquifer associated with Barton Springs in the Aus- VENI, G. 1988. The Caves of Bexar County, 2nd ed. tin area, Texas. Water Resource Investigation Re- Speleological Monograph, Texas Memorial Muse- port 86-4036, U. S. Geological Survey, Austin, Tex- um, University of Texas, Austin, Texas, U.S.A. as, U.S.A. . 1994. Geomorphology, Hydrology, Geo- SMITH,H. M., AND F. E. POTTER,JR. 1946. A third chemistry, and Evolution of the Karstic Lower neotenic salamander of the genus Eurycea from Glen Rose Aquifer, South-Central Texas. Ph.D. Texas. Herpetologica 3:105-109. Dissertation, Pennsylvania State University, Uni- SNEATH,P. H. A., AND R. R. SOKAL.1973. Numerical versity Park, Pennsylvania, U.S.A. Taxonomy: the Principles and Practice of Numeri- VENI, G., AND ASSOCIATES.1987. Valdina Farms cal Classification. W. H. Freeman and Co., San Sinkhole: Hydrogeologic & biologic evaluation. Francisco, California, U.S.A. Unpublished Report, Edwards Underground Wa- SOKAL,R. R., AND F. J. ROHLF. 1981. Biometry: the ter District, San Antonio, Texas, U.S.A. Principles and Practice of Statistics in Biological . 1992. Geologic controls on cave develop- Research, 2nd ed. W. H. Freeman and Co., San ment and the distribution of cave fauna in the Aus- Francisco, California, U.S.A. tin, Texas region. Unpublished Report, U.S. Fish STEJNEGER,L. 1896. Description of a new genus and and Wildlife Service, Austin, Texas, U.S.A. species of blind tailed batrachians from the subter- WADE, M. J., M. L. MCKNIGHT,AND H. BRADLEY ranean waters of Texas. Proceedings of the U.S. SHAFFER.1994. The effects of kin-structured col- National Museum 18: 619-621. onization and stepping-stone range expansion on SWEET,S. S. 1977a. Eurycea tridentifera. Catalogue nuclear and cytoplasmic gene diversity. Evolution of American Amphibians and Reptiles 199.1-199.2. 48:1114-1120. . 1977b. Natural metamorphosis in Eurycea WAKE, D. B. 1966. Comparative osteology and evo- neotenes, and the generic allocation of the Texas lution of the lungless salamanders, family Pletho- Eurycea (Amphibia: Plethodontidae). Herpetolo- dontidae. Memoirs of the Southern California gica 33:364-375. Academy of Sciences 4:1-111. . 1978a. The Evolutionary Development of WALSH, P. S., D. A. METZGER, AND R. HIGUCHI. 48 HERPETOLOGICAL MONOGRAPHS [No. 14

1991. Chelex 100 as a medium for simple extrac- binal River watershed, 29? 49' 26" N, 99? 34' 01" W. tion of DNA for PCR-based typing from forensic (Southwest, E. troglodytes complex, Sabinal group); material. BioTechniques 10:506-513. Sutherland Hollow Spring, west prong Medina River, WHEELER, W. C. 1990. Combinatorial weights in 29? 44' 58" N, 99? 25' 36" W. (Southwest, E. troglo- phylogenetic analysis: a statistical parsimony pro- dytes complex, Carson Cave group). cedure. Cladistics 6:269-275. Bell County.-Salado (Big Boiling, Main, or Siren) WIENS, J. J. 1995. Polymorphic characters in phylo- Springs, Salado Creek, 30? 56' 37" N, 97? 32' 31" W. genetic systematics. Systematic Biology 44:482- (North, E. chisholmensis-type locality); Salado (Rob- 500. ertson) Springs, Salado Creek, 30? 56' 37" N, 97? 32' . 1998. Testing phylogenetic methods with 39" W. (North, E. chisholmensis). tree-congruence: phylogenetic analysis of polymor- Bexar County.-Helotes Creek Spring, Medina phic morphological characters in phrynosomatid River watershed, 29? 38' 15" N, 98? 41' 40" W. lizards. Systematic Biology 47:427-444. (Southeast, E. neotenes-type locality); Leon Springs, . 1999. Polymorphism in systematics and com- Leon Creek, Medina River watershed, 29? 39' 46" N, parative biology. Annual Review of Ecology and 98? 38' 14" W. (Southeast, E. neotenes). Systematics 30:327-362. Blanco County.-Boardhouse Springs, Blanco Riv- . 2000. Reconstructing phylogenies from al- er watershed, 30? 06' 40" N, 98? 18' 07" W. (South- data: method lozyme comparing performance with east, E. pterophila); T-Cave, Blanco River watershed, congruence. Biological Journal of the Linnaean So- 30? 04' 36" N, 98? 19' 46" W. (Southeast, E. ptero- ciety: In press. phila); Zercher Spring, Blanco River watershed, 30? WIENS, J. J., AND M. R. SERVEDIO. 1997. Accuracy 06' 10" N, 98? 27' 25" W. (Southeast, E. pterophila). of phylogenetic analysis including and excluding Comal County.-Badweather Pit, Cibolo Creek polymorphic characters. Systematic Biology 46: watershed, 29? 45' 21" N, 98? 37' 13" W (Southeast, 332-345. E. tridentifera); Comal Springs, headwaters of the WIENS, J. J., AND M. R. SERVEDIO. 1998. Phyloge- Comal River, 29? 42' 49" N, 98? 08' 13" W. (South- netic analysis and intraspecific variation: perfor- east, E. sp.); Ebert Cave, Cibolo Creek watershed, mance of parsimony, likelihood, and distance meth- 29? 45' 06" N, 98? 23' 28" W. (Southeast, E. triden- ods. 47:228-253. Systematic Biology tifera); Honey Creek Cave, Guadalupe River water- E. 0. 1978. The WILEY, evolutionary species concept shed, 29? 50' 50" N, 98? 29' 30" W. (Southeast, E. reconsidered. 27:17-26. Systematic Zoology tridentifera-type locality); Rebecca Creek Spring, C. AND P. L. ABBOTT. 1986. WOODRUFF, M., JR., Guadalupe River watershed, 29? 55' 28" N, 98? 22' Stream and evolution of the Edwards piracy Aqui- 22" W. (Southeast, E. latitans complex). fer along the Balcones Escarpment, central Texas. Edwards County.-Smith's ( = Dutch Creek) 77-89. In P. L. Abbott, and C. M. Woodruff, Pp. Spring, Nueces River watershed, 29? 39' 09" N, 100? The Balcones Jr. (Eds.), Escarpment: Geology, Hy- 06' 12" W. (Southwest, E. complex, and Social in Cen- troglodytes drology, Ecology Development Carson Cave group); West Nueces River Spring, 29? tral Texas. of America, San Di- Geological Survey 43' 20" N, 100? 24' 51" W. (Southwest, E. California, U.S.A. troglodytes ego, complex, Carson Cave group). WRIGHT,A. H., AND A. A. WRIGHT.1938. Amphib- Gillespie Pedernales Riv- ians of Texas. Transactions of the Texas County.-Trough Spring, Academy er watershed, 30? 08' 36" N, 99? 04' 40" W (South- of Science 21:1-35, 2 unnum., 3 pl. west, E. troglodytes complex, Carson Cave group). WRIGHT, S. 1978. Evolution and the Genetics of Pop- Hays County.-Ezell's Cave, San Marcos River wa- ulations, Vol. 4. Variability Within and Nat- Among tershed, 29? 52' 27" N, 97? 57' 34" W (San Marcos, ural Populations. of Chicago Press, Chi- University E. rathbuni); Fern Bank (Little Arkansas) Springs, cago, Illinois, U.S.A. Blanco River watershed, 29? 59' 00" N, 98? 00' 49" ZHEN, L., AND R. T. SWANK. 1993. A simple and high W. (Southeast, E. locality); Grapevine yield method for recovering DNA from agarose pterophila-type Cave, Blanco River watershed, approximately 30? 02' gels. BioTechniques 14:894-898. 30" N, 98? 12' 45" W. (Southeast, E. pterophila); Rat- tlesnake Cave, San Marcos River watershed, 29? 54' APPENDIX I 07" N, 97? 55' 17" W. (San Marcos, E. rathbuni); San Localities for populations of central Texas Marcos (Aquarena) Springs, pipe outflow at subma- hemidactyliine salamanders examined in this study rine theater, headwaters of the San Marcos River, 29? 53' 35" N, 97? 55' 50" W. (San Marcos, E. rathbuni); Watersheds in which each is located are locality San Marcos Springs, headwaters of the where not obvious based on the name. (Aquarena) given locality San Marcos River, 29? 53' 35" N, 97? 55' 50" W. (San In cases in which formal taxon names have been as- Marcos, E. to from those nana-type locality). signed populations particular localities, Kendall Creek Medina Riv- names are used. Informal to which County.-Bear Spring, groups popula- er watershed, 29? 48' 15" N, 98? 52' 10" W (South- tions have been assigned based on this study are list- east, E. latitans complex); Cibolo Creek Tributary ed where appropriate. Spring, Cibolo Creek watershed, 29? 49' 03" N, 98? Bandera County.-Murphy's Spring (Wedgeworth 51' 43" W. (Southeast, E. latitans complex); Knee- Creek South Spring), Sabinal River watershed, 29? deep Cave Spring, Guadalupe River State Park, 29? 48' 00" N, 98? 33' 31" W. (Southwest, E. troglodytes 52' 31" N, 98? 29' 05" W. (Southeast, E. latitans com- complex, Sabinal group); Sabinal Canyon Spring, Sa- plex); Less Ranch Spring, Guadalupe River water- 2000] HERPETOLOGICAL MONOGRAPHS 49 shed,29? 46' 40" N, 98? 50' 52" W. (Southeast, E. la- 30? 20' 28" N, 98? 08' 14" W (Southeast, E. sp.); titans complex); Mueller's Spring, Medina River Hammett's Crossing Spring #2 (Pedernales Spring watershed, approximately 29? 44' N, 98? 47' 30" W. #2), Pedernales River, 30? 20' 23" N, 98? 08' 15" W. (Southeast, E. neotenes); Peavey's Springs, headwa- (Southeast, E. sp); Horsethief Hollow Spring, Bull ters of the Blanco River, approximately 30? 05' 30" Creek watershed, 30? 24' 31" N, 97? 49' 00" W. N, 98? 39' 30" W. (Southeast, E. pterophila); Pfeiffer's (North, E. tonkawae); Kretschmarr Cave, Colorado Water Cave, Guadalupe River watershed, 29? 45' 44" River watershed, 30? 24' 47" N, 97? 51' 10" W. N, 98? 39' 59" W. (Southeast, E. latitans-subterra- (North, provisionally E. tonkawae); Schlumberger nean extension of type locality). Spring, headwaters west fork Bull Creek, 30? 25' 15" Kerr County.-176 Spring, Guadalupe River wa- N, 97? 50' 24" W. (North, E. tonkawae); Stillhouse tershed, 30? 05' 18" N, 99? 19' 14" W. (Southwest, E. Hollow Springs, Bull Creek watershed, 30? 22' 28" N, troglodytes complex); Cherry Creek Spring, Guada- 97? 45' 55" W. (North, E. tonkawae-type locality); lupe River watershed, 29? 50' 41" N, 98? 56' 52" W. Wheless Springs, Long Hollow Creek, Colorado Riv- (Southeast, E. latitans complex); Cloud Hollow er watershed, 30? 27' 42" N, 97? 52' 28" W. (North, Spring, Medina River watershed, 29? 50' 36" N, 98? E. tonkawae). 57' 14" W (Southeast, E. latitans complex); Edmun- Uvalde County.-Carson Cave, West Nueces River son Creek (Camp Mystic) Springs, Guadalupe River watershed, 29? 28' 50" N, 100? 04' 44" W. (Southwest, = watershed, 30? 00' 21-3" N, 99? 21' 43-54" W. E. troglodytes complex, Carson Cave group); WB ( Wetback) Sabinal River watershed, 29? 35' (Southwest, E. troglodytes complex); Fessenden Spring, Springs, Guadalupe River watershed, 30? 10' 00" N, 12" N, 99? 36' 14" W (Southwest, E. troglodytes com- Carson Cave 99? 20' 32" W. (Southwest, E. troglodytes complex, plex, group). Williamson Carson Cave group); Robinson Creek (Highway 16 County.-Avant's (Capitol Aggregates) middle fork San Gabriel River, 30? 38' 44" South) Spring, north prong Medina River watershed, Spring, N, 97? 44' 11" W. E. Bat Well, Cow- 29? 54' 55" N, 99? 15' 08" W. (Southwest, E. troglo- (North, naufragia); an Creek watershed, San Gabriel River 30? dytes complex, Carson Cave group). drainage, 42' 10" N, 97? 42' 59" W. (North, provisionally E. Real County.-Greenwood Valley Ranch Spring Creek (Round Rock) 30? #1, east Nueces River, 29? 57' 20" N, 99? 58' naufragia); Brushy Spring, prong 31' 00" 97? 39' 38" W. E. Bu- 17" W. E. Green- N, (North, tonkawae); (Southwest, troglodytes complex); ford Hollow below Lake wood Ranch #2, east Nueces Riv- Springs, just Georgetown Valley Spring prong Dam, north fork San Gabriel River, 30? 39' 39" N, er, 29? 59' 11" N, 99? 57' 51" W. (Southwest, E. trog- 97? 43' 36" W. (North, E. naufragia-type locality); lodytes complex); Greenwood Valley Ranch Spring Buttercup Creek Cave (Buttercup River Cave), But- #3, east prong Nueces River, 29? 59' 22" N, 99? 57' tercup Creek Karst, Brushy Creek watershed, ap- 13" W. (Southwest, E. Tucker troglodytes complex); proximately 30? 29' 33" N, 97? 50' 44" W. (North, Hollow Cave, Frio River watershed, 29? 44' 33" N, provisionally E. tonkawae); Cedar Breaks Hiking Trail 99? 46' 42" W. E. (Southwest, troglodytes complex). Spring, south shore of Lake Georgetown, north fork Travis County.-Balcones Community Park San Gabriel River, 30? 39' 36" N, 97? 45' 02" W. Walnut Creek 30? 24' 45" 97? Spring, watershed, N, (North, E. naufragia); Ilex Cave, Buttercup Creek 43' 02" W. (North, E. tonkawae); Barrow Hollow Karst, Brushy Creek watershed, approximately 30? Spring, Bull Creek watershed, 30? 22' 33" N, 97? 46' 29' 28" N, 97? 50' 50" N. (North, provisionally E. 02" W (North, E. tonkawae); Barton Springs, Barton tonkawae); Knight (Crockett Garden) Spring, south Creek, 30? 15' 49" N, 97? 46' 14" W. (Southeast, E. shore of Lake Georgetown, north fork San Gabriel sosorum-type locality); Bull Creek (Hanks Tract) River, 30? 39' 50" N, 97? 45' 04" W (North, E. nauf- Spring, north fork Bull Creek, 30? 25' 38" N, 97? 49' ragia); T.WA.S.A. Cave, Buttercup Creek Karst, 08" W. (North, E. tonkawae); Bull Creek Spring Pool Brushy Creek watershed, approximately 30? 29' 49" (New Bull Creek Spring), west fork Bull Creek, 30? N, 97? 50' 48" W. (North, provisionally E. tonkawae); 24' 59" N, 97? 49' 00" W. (North, E. tonkawae); Can- Testudo Tube, Buttercup Creek Karst, Brushy Creek yon Creek Spring, north fork Bull Creek, 30? 25' 33" watershed, approximately 30? 29' 35" N, 97? 51' 23" N, 97? 48' 51" W. (North, E. tonkawae); Canyon Vista W. (North, provisionally E. tonkawae); Treehouse Spring, Bull Creek watershed, 30? 25' 51" N, 97? 46' Cave, Buttercup Creek Karst, Brushy Creek water- 55" W. (North, E. tonkawae); Hammett's Crossing shed, approximately 30? 29' 55" N, 97? 50' 07" W Spring #1 (Pedernales Spring #1), Pedernales River, (North, provisionally E. tonkawae). A-xPPENDIX II Part A Allele Frequencies in Populations of Central Texas Eurycea and Outgroup Members N is the number of individuals from a given population in which a particular locus was resolved. Loci are numbered from least to most anodally migrating, and alleles are designated by letters that in most cases increase with increasing anodal mobility. H =direct-count heterozygosity per locus per individual, P = percentage polymorphic loci, and A =number of alleles/locus. Standard errors are in brackets. NC = not calculated. Locus Acon-1 Ak sAat Ck- 1 Ck-2 Cap Gpi Gr Gapdh G3pdh NORTH Jollyville Plateau E. tonkawae Balcones Park Spring N= 5 5 5 5 5 5 4 4 4 5 C: 1. 000 b: 1.000 b: 1. 000 a: 1. 000 a: 1.000 a: 1.000 a: 1. 000 a: 0. 625 b: 1.000 a: 1.000 b: 0.375 H Barrow Hollow Spring N= 5 5 5 4 4 4 5 5 4 5 C: 1. 000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 b: 1. 000 a: 1.000 Bull Creek Spring N= 5 5 5 4 5 5 5 5 5 5 C: 1. 000 b: 1.000 b: 1. 000 a: 1. 000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 b: 1.000 a: 1. 000 Canyon Creek Spring N= 3 3 3 3 3 3 3 3 3 3 c: 1. 000 b: 1.000 b: 1. 000 a: 1. 000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 b: 1.000 a: 1. 000 0 N= 4 4 4 3 3 3 4 4 2 4 z- Canyon Vista Spring 0 C: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 b: 1. 000 a: 1.000 0 Horsethief Hollow Spring N= 1 1 1 1 1 1 1 1 1 1 C: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 b: 1.000 a: 1.000 1I New Bull Creek Spring N= 2 2 2 2 2 2 2 2 1 1 C: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1. 000 a: 1. 000 b: 1. 000 b: 1. 000 a: 1. 000 Schlusmberger Spring N= 6 6 6 6 6 6 6 6 6 6 C: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1. 000 a: 1.000 a: 1. 000 b: 1.000 b: 1. 000 a: 1. 000 StiullhouseHollow Springs N= 14 14 14 13 14 14 14 14 14 14 C: 1.000 b: 1. 000 b: 1. 000 a: 1. 000 a: 1.000 a: 1. 000 a: 1.000 b: 1.000 b: 1. 000 a: 1. 000 Wheless Springs N= 5 5 5 5 5 5 5 5 5 5 C: 1. 000 b: 1.000 b: 0. 900 a: 1. 000 a: 1. 000 a: 1.000 a: 1. 000 b: 1. 000 b: 1. 000 a: 1. 000 C: 0. 100 Round Rock Spring N 9 9 9 9 9 9 9 9 9 9 E. tonkawae C: 1.000 b: 1. 000 b: 1. 000 a: 1. 000 a: 1. 000 a: 1. 000 a: 1. 000 b: 1. 000 b: 1. 000 a: 1.000 Lake Georgetown Area N= 4 4 5 5 4 5 3 4 3 4 E. naufragia Avant's Springs C: 1. 000 b: 1. 000 b: 1. 000 a: 0. 700 a: 1. 000 a: 1.000 a: 0.667 a: 0.375 b: 1.000 d: 1.000 c: 0.300 b: 0.333 b: 0.625 Buford Hollow Springs N 8 8 9 8 9 9 9 7 7 9 c: 0.688 b: 1. 000 b: 1.000 a: 0.688 a: 1.000 a: 1.000 a: 0. 944 a: 0.929 b: 1.000 d: 1.000 f: 0.313 c: 0.313 b: 0.056 b: 0.071 0l APPENDIX II 0z Part A-Continued. Locus Acon-1 Ak sAat Ck-1 Ck-2 Cap Gpi Gr Gapdh G3pdh Cedar Breaks Spring N= 5 5 5 4 5 5 5 5 4 5 C: 0.900 b: 1.000 b: 1.000 a: 0.625 a: 1.000 a: 1. 000 a: 1.000 a: 0. 100 b: 1.000 d: 1.000 f: 0.100 c: 0.375 b: 0.900 Knight Spring N= 5 5 5 5 5 5 5 4 4 5 c: 0.200 b: 1.000 b: 1.000 a: 0.500 a: 1.000 a: 1.000 a: 1.000 a: 0.375 b: 1.000 d: 1.000 f: 0.800o c: 0.500 b: 0. 625 Salado Springs N-= 8 8 S S 8 8 5 5 S 8 E. chisholmensis C: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 b: 0.875 a: 1.000 c: 0. 125 Bat Well N= 1 1 1 1 1 1 1 1 1 1 E. naufragia C: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 a: 1.000 Testudo Tube N= 2 2 2 2 2 2 2 2 2 2 E. tonkawae b: 1.000 b: 1.000 b: 0.250 a: 1.000 a: 1. 000 a: 1. 000 a: 1.000 b: 1. 000 b: 1.000 a: 1.000 c: 0. 750 z Kretsclhmuarr Cave N= 4 4 4 4 4 4 3 4 4 4 0 E. tonkawaee C: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 b: 1.000 b: 1.000 a: 1. 000 0 Buttercup Creek Caves E. tonkawae Buttercup Creek Cave N= 3 3 3 3 3 3 3 3 2 3 h: 0. 167 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 b: 1.000 a: 1.000 c: 0.833 Ilex Cave N= 2 2 2 2 2 2 2 2 2 2 C: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 b: 1.000 a: 1.000 Treehouse Cave N= 1 1 1 1 1 1 1 1 I 1 b: 1.000 b: 1.000 b: 0.500 a: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 b: 1.000 a: 1.000 c: 0.500 T.WA.S.A. Cave N= 1 1 1 1 1 1 1 1 1 1 C: 1.000 b: 1.000 b: 0.500 a: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 b: 1.000 a: 0.500 c: 0.500 d: 0.500 SAN MARCOS E. nana N= 12 13 13 13 13 13 13 13 11 13 d: 1.000 a: 1.000 e: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1.000 c: 1.000 E. mathbuni N= 5 5 5 4 5 5 5 0 5 5 e: 1.000 b: 1.000 b: 0.600 b: 1.000 b: 1.000 a: 1.000 a: 1.000 c: 1.000 b: 0.700 c: 0.400 c: 0.300 CA' APPENDIX II bo: Part A-Continued. Locus

Acon-I Ak sAat Ck-1 Ck-2 Cap Gpi Cr Gapdh C3pdh SOUTHEAST E. latitans complex E. latitans (type locality) N= 3 4 5 4 4 3 5 5 2 2 b: 1.000 b: 1.000 a: 0.900 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1. 000 b: 0. 100 Bear Creek Spring N= 6 6 6 6 6 6 6 6 6 6 b: 1.000 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1.000 Cherry Creek Spring N= 2 2 2 2 2 1 2 2 2 2 Z b: 1.000 b: 1.000 a: 1. 000 b: 1.000 a: 1'.000 a: 1.000 a: 1. 000 a: 1.000 c: 1.000 c: 1.000 Cloud Holow Spring N= 5 5 5 5 5 5 4 5 4 5 b: 1.000 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1. 000 Cibolo Creek Spring N= 8 8 8 8 8 8 7 8 8 8 b: 1.000 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1. 000 a: 1.000 c: 1.000 c: 1.000 0 Kneedeep Cave Spiring N= 6 5 6 6 6 5 5 6 6 6 a: 0.083 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 0.900 a: 1.000 c: 1.000 c: 1.000 b: 0.917 b: 0. 100 Honey Creek Cave Spring N= 1 1 1 1 1 1 1 1 1 1 z b: 1.000 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1. 000 a: 1.000 c: 1.000 c: 1.000 0 Less Ranch Spring N- 2 2 2 2 2 2 2 2 2 2 b: 1.000 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1. 000 c: 1.000 Rebecca Creek Spring N= 6 6 3 5 5 6 5 6 5 6 b: 1.000 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 b: 1.000 a: 1.000 c: 1.000 c: 1.000 E. neotenes Helotes Spiring (type locality) N= 11 9 11 11 10 9 10 11 6 9 b: 1.000 b: 1.000 a: 0. 136 b: 1.000 a: 1.000 a: 1.000 a: 0. 700 b: 1.000 c: 1.000 c: 1.000 b: 0. 864 c: 0.100 d: 0.200 Leon Springs N= 2 2 2 2 2 2 2 2 2 2 b: 1.000 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 0.500 b: 1.000 c: 1.000 c: 1.000 b: 0.500 Mueller's Spring N= 4 4 4 4 4 4 4 4 4 3 b: 1.000 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 0.875 b: 1.000 c: -1.000 c: 1. 000 d: 0. 125 APxxPENDIX II Part A-Continued. Locus

Acon-1 Ak sAat Ck-1 ck-2 Cap Gpi Cr Gapdh G3pdh

Fern Bank Spring (type loc.) N= 10 10 10 10 10 10 10 10 10 10 b: 1.000 b: 1.000 a: 0.950 b: 1.000 a: 1.000 a: 1. 000 a: 1.000 b: 1. 000 C: 1.000 C: 1.000 b: 0.050 Boardhiouse Springs N= 11 11 11 11 11 11I 11 11I 11 11 b: 1.000 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 0.955 b: 1.000 C: 1.000 C: 1.000 b: 0.045 Grapevine Cave N= 2 2 2 2 2 2 2 2 2 2 b: 1.000 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 C: 1.000 C: 1.000 N= Peavey's Springs 5 5 5 5 5 5 5 5 5 5 0.6 b: 1.000 b: 1.000 a: 1. 000 b: 1.000 a: 1.000 a: 1. 000 a: 1.000 h: 1. 000 0: 1.000 0: 1.000 T Cave N= 5 5 5 5 5 4 5 5 5 5 b: 1.000 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 C: 1.000 C: 1.000 E. sosorum N= 13 14 14 14 14 14 14 13 12 14 b: 0.923 b: 1.000 a: 0.964 b: 1.000 a: 1.000 a: 1.000 a. 0.464 a. 1.000 C: 1.000 b: 0. 107 c: 0.077 b: 0.036 b: 0.536 c: 0.893 0 E. tridentifera 0 Honey Creek Cave (type locality) N= 5 5 5 5 5 4 5 5 5 5 b: 1.000 b: 1.000 a: 0.300 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 C: 1.000 b: 1.000 Cr' b: 0. 700 Badweather Pit N= 5 5 5 5 5 5 5 5 5 5 b: 1.000 b: 1.000 a: 0.400 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 0.900 C: 1. 000 C: 1.000 b: 0.600 b: 0. 100 Ebert Cave N= 6 6 6 6 6 6 6 6 6 6 b: 1.000 b: 1. 000 a: 0.500 b: 1.000 a: 1.000 a: 1. 000 a: 0.583 a: 1.000 C: 1. 000 b: 1.000 b: 0.500 b: 0.417 Comal Springs N- 11 12 11 11 11 12 11 11 9 9 b: 0.545 b: 1.000 a: 0.591 b: 1.000 a: 1.000 a: 1.000 a: 0.545 a: 1.000 C: 1.000 C: 1.000 e: 0.455 b: 0.409 b: 0.455 Pedernales Spring 1 N= 7 7 6 7 7 6 5 7 7 7 b: 0.857 b: 1.000 a: 0.333 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 C: 1.000 C: 1.000 c: 0. 143 b: 0.667 Spring 2 7 7 7 7 7 7 7 7 7 7 b: 0.857 b: 1.000 a: 0.714 b: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1.000 c: 0. 143 b: 0.286 APDPENDIX 1I Part A-Continued. Locus Acon-I Ak sAat Ck-1 Ck-2 Cap Gpi Cr Gapdh G3pdh SOUTHWEST (E. troglodytes complex) Carson Cave Group Carson Cave N= 5 5 5 5 5 S S 5 4 5 b: 1.000 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 C: 1.000 b: 1.000 S S Fessenden Spring N= 8 8 8 8 8 8 8 8 a: 0.063 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 C: 1.000 b: 0.938 Robinson Creek Spring N= 4 4 4 4 4 4 4 4 4 4 b: 1.000 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 C: 1.000 C: 1.000 H 0 N- 14 14 14 14 14 14 14 14 14 14 Smith's Spring 0 b: 1.000 b: 1.000 b: 1. 000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 C: 1.000 b: 0.536 c: 0.464 Sutherland Hollow Spring N= 11 11 11I 11 11 11 11 11 11 11 b: 1.000 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1. 000 a: 0. 955 b: 1.000 C: 1.000 b: 0.455 b: 0.045 c: 0.545 0 Trough Spring N= 10 10 10 10 10 10 10 10 10 10 0 b: 1.000 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 C: 1.000 C: 1.000 1 1 1 1 1 West Nueces Spring 1= 1 I 1 1 b: 1.000 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 C: 1.000 b: 1.000 WB Spring N= 4 4 4 4 4 4 4 4 4 4 b: 1.000 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 0. 750 b: 1.000 C: 1.000 C: 1.000 b: 0.250 10 Camp Mystic Spring N= 10 10 10 10 10 10 10 10 10 b: 1.000 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 C: 1.000 C: 1.000 Greenwood Valley Ranch Springs N= 6 6 6 6 6 6 6 6 6 6 b: 1.000 b: 1.000 b: 1. 000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1. 000 C: 1.000 b: 1. 000 0 176 Spring N= 4 3 4 4 4 4 4 4 4 4 a: 0. 125 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 C: 1.000 b: 1.000 b: 0.875 Sabinal Group Murphy's Spring N- 4 4 4 4 4 2 4 4 4 4 b: 1.000 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 b: 1.000 b: 1.000 a: 0.500 b: 1.000 c: 0.500 Sabinal Canyon Spring N= 11 11 11 11 11 11 11 11 10 11 b: 1. 000 b: 1.000 b: 1. 000 b: 1.000 a: 1.000 a: 1.000 a: 0.091 b: 1. 000 a: 0.950 b: 0.955 b: 0.909 c: 0.050 c: 0. 045 APxUPENDIX I Part A-Continued. Locus

Acon-I Ak sAat Ck-1 Ck-2 Cap Gpi Cr Gapdh C3pdh Tucker Hollow Cave N= 6 5 6 6 6 6 6 6 6 6 b: 1.000 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1. 000 a: 0.667 b: 1. 000 a: 1.000 b: 1.000 b: 0.333 H OUTGROUP 0 Eurycea m. multiplicata N= 2, 2 2 2 2 2 2 2 2 2 0: 1.000 b: 1.000 1: 1.000 d: 1.000 c: 1.000 c: 1.000 b: 1.000 e: 1.000 x: 1.000 a: 1.000 2 E.l.longicauda N= 2 2 2 2 2 2 2 2 2 x: 0.500 b: 1.000 y: 1.000 d: 1.000 e: 1.000 a: 1.000 b: 1.000 a: 1.000 y: 1.000 a: 1.000 ~z y: 0.500 0 E. wilderae N= 3 3 0 0 2 3 3 3 3 3 e: 1.000 b: 1.000 f: 1.000 a: 1.000 b: 1.000 a: 1.000 y: 1.000 a: 1.000 0 E. bislinteata N= 1 1 1 0 1 1 1 1 1 1 i: 0.500 b: 1.000 y: 1.000 e: 1.000 a: 1.000 b: 1.000 a: 1.000 y: 1.000 a: 1.000 k: 0.500 2 E. quadridigitata TX N= 2 2 2 2 2 2 2 2 2 c: 1.UUU b: 1.uUu f. u.500 C: 1.000 a: 1.000 a: 1.000 b: 1.000 e: i.OuO y: 1.000 a: 1.000 g: 0.500 2 E. quadridigitata SC N= 2 2 2 2 2 2 2 2 2 k: 1. 000 b: 1. 000 b: 1.000 y: 1. 000 a: 1.000 a: 1. 000 b: 1.000 e: 1.000 y: 1. 000 b: 1.000 Haideot,riton wallacei N= 1 1 1 1 1 0 0 1 1 0 d: 1.000 b: 1.000 x: 1.000 d: 1. 000 c: 1.000 a: 1.000 j: 1. 000 2 Typhlotriton spelaeus N= 2 2 2 0 2 2 2 2 2 d: 0.500 b: 1.000 b: 1.000 c: 1.000 c: 1. 000 b: 1.000 b: 0. 750 d: 1.000 x: 1.000 e: 0.500 c: 0.250 APxxPENDIX II Part B Locus ldhI 1 ldh-2 Ldh-A Ldh-B Mdh-1 Mdh-2 Mdhp Mpi Pep-A Pep-B NORTH Jollyville Plateau E. tonkawae Balcones Park Spring N- 5 5 5 5 5 5 5 5 5 5 a: 1.000 b: 1. 000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 Barrow Hollow8Spring N= 5 5 5 5 4 4 5 4 5 5 a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 Bull Creek Spring N= 5 5 5 5 5 5 5 5 5 5 a: 1.000 b: 1.000 c: 1.000 a: 1. 000 a: 1. 000 a: 1.000 a: 0. 100 c: 1.000 d: 1. 000 b: 1.000 b: 0.900 Canyon Creek Spring N= 3 3 3 3 3 3 3 3 2 3 a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1. 000 a: 1.000 a: 0.333 c: 1.000 d: 1. 000 b: 1.000 ~ b: 0.6670 Canyon Vista Spring N= 4 4 4 4 4 4 4 4 4 4 0 a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 0.875 c: 1.000 d: 1.000 b: 1.000 b: 0. 125 Horsethief Hollow N= 1 1 1 1 1 1 1 1 1 1 a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 New Bull Creek Spring N= 2 2 2 2 2 2 2 2 2 2 0 a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 0 Scbilumberger Spring N= 6 6 6 6 6 6 6 6 6 6 0 a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 Stilihouse Hollow N= 14 14 14 14 14 14 14 14 14 14 a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1. 000 a: 1.000 a: 0.214 c: 1.000 d: 1.000 b: 1.000 i b: 0.786 Wheless Springs N= 5 5 5 5 4 4 5 5 2 5 a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 Round Rock Spring N= 9 9 9 9 9 9 9 9 9 9 E. tonkawae a: 1.000 h: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 a: 0.063 c: 1.000 d: 1.000 b: 1.000 b: 0.938 Lake Georgetowrn Area E. naufragia Avant's Spring N= 5 5 5 5 5 5 4 5 4 5 a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 Buford Hollow Springs N= 9 9 9 9 9 9 9 9 9 9 a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 0. 944 b: 1.000 C: 1.000 d: 1.000 h: 1.000 c: 0.056 JAPPENDIX II Part B-Continued. Locus

ldh-1 ldh-2 Ldh-A Ldh-B MdTh-1 Mdhi 2 Mdhp Mpi Pep-A Pep-B Cedar Breaks Spring N= 5 5 5 5 5 5 5 5 5 5 a: 1. 000 b: 1.000 c: 1.000 a: 1. 000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 Knight Spring N= 5 5 5 5 5 5 5 5 4 5 a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 C: 1.000 d: 1.000 b: 1.000 Salado Springs N= 8 8 8 8 8 8 8 8 8 8 E. chisholmensis a: 1. 000 b: 1.000 c: 1.000 a: 1. 000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 Bat Well N= 1 1 1 1 1 1 1 1 1 1 E. naufragia a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 Testudo Tube N- 2 2 2 2 2 2 2 2 2 2 E. tonkawae a: 1. 000 b: 1. 000 c: 1. 000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 b: 1.000 b: 1.000 Kretschmarr Cave N= 4 4 4 4 4 4 3 4 4 4 H E. tonkawae a: 1.000 b: 1.000 c: 1.000 a: 1. 000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 0 Buttercup Creek Caves 0 E. tonkawae Buttercup Creek Cave N- 1 3 3 3 3 3 3 3 3 3 1 a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 Ilex Cave N= 2 2 2 2 2 2 2 2 2 2 a: 1. 000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1.000 b: 1.000 0 Treehouse Cave N= 1 1 1 1 1 1 1 1 1 1 o a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 d: 1. 000 b: 1. 000 0 T.WA.S.A. Cave N= 1 1 1 1 1 1 1 1 1 1 ~ a: 1.000 b: 1.000 c: 1.000 a: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 b: 0.500 b: 1.000 d: 0.500 SAN MARCOS E.nana N= 13 13 13 13 13 13 13 12 12 13 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1. 000 a: 1.000 c: 1.000 c: 1.000 b: 0.375 d: 1.000 d: 0. 625 E. rathbuni N= 5 5 5 5 5 5 5 5 5 5 b: 1.000 b: 1.000 b: 1.000 a: 1.000 a: 1.000 b: 1.000 c: 1.000 c: 1.000 b: 1.000 b: 0.900 d: 0.100 SOUTHEAST E. latitans complex E. latitans (type locality) N= 3 5 5 5 5 5 5 4 2 3 b: 1.000 a: 1.000 b: 1.000 a: 1. 000 a: 1. 000 a: 1.000 C: 1. 000 c: 1.000 b: 1. 000 a: 1.000 Bear Creek Spring N- 5 6 6 6 6 6 6 6 6 6 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1.000 b: 1.000 b: 1.000 cr APPENDIX II Part B-Continued. Locus ldhi1 ldh-2 Ldh-A Ldh-B MdhlI Mdh-2 Mdhp Mpi Pep-A Pep-B Cherry Creek Spring N= 2 2 2 2 2 1 2 1 2 2 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 0: 1.000 b: 1.000 b: 1.000 CloudlHollow Spring N= 5 5 5 5 5 5 5 5 5 5 b: 1. 000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1.000 b: 1.000 b: 1.000 Cibolo Creek Spring N- 8 8 8 8 8 8 8 8 8 8 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 0: 1.000 c: 1.000 b: 1.000 b: 1.000 Kneedeep Cave Spring N= 6 6 6 6 6 6 6 6 6 6 b: 1.000 a: 1. 000 b: 1.000 a: 1.000 a: 0.667 a: 1.000 c: 1.000 c: 1.000 b: 1.000 b: 1.000 d: 0.333 Honey Creek Cave Spring N- 1 1 1 1 1 1 1 1 1 1 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 C: 1.000 b: 1.000 b: 1. 000 Less RanchiSpring N= 2 2 2 2 2 2 2 2 2 2 0 b: 1.000 a: 1. 000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1.000 b: 1.000 b: 1.000 Rebecca Creek Spring N= 6 6 6 6 6 6 6 6 6 6 0 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1.000 b: 1. 000 b: 1.000 E. neotenes Helotes Spring (type locality) N- 9 11 11 11 9 11 10 11 11 11 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 b: 1.000 b: 1.000 b: 1.000 0 Leon Springs N= 2 2 2 2 2 2 2 2 2 2 Z b: 1.000 a: 1.000 b: 1.000 a: 1. 000 a: 1. 000 a: 1. 000 c: 1. 000 b: 0.570 b: 1.000 b: 1.000 0: 0.250 Mueller's Spring N- 4 4 4 4 4 4 4 4 4 4 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1. 000 b: 1.000 b: 1.000 b: 1.000 E. pterophila Fern Bank Spring (type locality) N= 10 10 10 10 10 10 10 10 10 10 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 b: 0.600 b: 1.000 b: 1.000 c: 0.400 Boardhouse Springs N= 11 11 11 11 11 11 11 11 11 11 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1. 000 b: 1. 000 b: 1. 000 Grapevine Cave N= 2 2 2 2 2 2 2 2 2 2 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 b: 1.000 b: 1.000 b: 1.000 Peavey's Springs N= 5 5 5 5 5 5 5 5 5 5 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1.000 b: 1.000 b: 1. 000 T Cave N- 5 5 5 5 5 5 5 5 5 5 b: 1. 000 a: 1.000 a: 0.500 a: 1.000 a: 1.000 a: V.000 c: 1.000 c: 1.000 b: 1.000 b: 1.000 d: 0.500 APPENDIX II Part B-Continued. Locus

ldh-1 Jdh-2 Ldh-A Ldhs B Mdh-1 Mdh-2 Mdhp Mpi Pep-A Pep-B E. sosorum N= 12 14 14 14 14 14 14 14 12 14 b: 1.000 a: 1.000 b: 0.964 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1.000 d: 0.167 a: 0.071 d: 0.036 e: 0.833 b: 0.929 E. tridentifera Honey Creek Cave (type locality) 4 4 5 5 5 5 5 5 3 4 b: 1.000 a: 1.000 b: 1. 000 a: 1.000 a: 1.000 a: 1.000 d: 1. 000 C: 1.000 b: 1.000 a: 0.750 N= b: 0.250 Badweather Pit 5 5 5 5 5 5 5 5 5 5 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 0.~900 d: 1.000 C: 1.000 b: 1. 000 a: 0.400 N- C: 0.100 b: 0.600 H Ebert Cave 6 6 6 6 6 6 6 6 6 6 b: 1.000 a: 1.000 b: 0.500 a: 1.000 a: 1.000 a: 1.000 c: 0.083 C: 1.000 b: 1.000 a: 1.000 d: 0.917 0 Comal Springs N= 11 12 12 11 11 12 12 11 12 11 C'l) b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 C: 1.000 C: 1.000 b: 1.000 b: 1.000 0 Pedemales Spring 1 7 7 7 7 7 7 7 7 6 7 z b: 1.000 a: 1.000 a: 0.429 a: 1.000 a: 1.000 a: 1.000 c: 0.929 b: 0.500 b: 0.917 b: 1. 000 b: 0.571 f: 0.071 c: 0.500 d: 0.083 0 0 Spring 2 N= 7 7 7 7 7 7 7 7 7 6 b: 1.000 a: 1.000 a: 0.429 a: 1.000 a: 1.000 a: 1.000 c: 0.929 b: 0.429 b: 1.000 b: 1.000 b: 0.571 f1:0.071 c: 0.571 SOUTHWEST (E. troglodytes complex) N- Carson Cave Group Carson Cave 5 5 5 5 5 5 5 5 5 5 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 C: 1.000 C: 1.000 b: 0.900 a: 1.000 C: 0. 100 Fessenden Spring 8 8 8 8 8 8 7 8 7 8 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 C: 1.000 b: 0.063 a: 0.929 a: 1.000 c: 0.938 b: 0.071 Robinson Creek Spring 4 4 4 4 4 4 4 4 4 4 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 C: 1.000 C: 1.000 b: 1.000 a: 1.000 Smith's Spring 12 14 14 14 14 14 14 14 13 14 b: 1.000 a: 0.857 b: 1.000 a: 1.000 a: 0.393 a: 1.000 C: 1.000 C: 1.000 b: 0.654 a: 1.000 b: 0. 143 b: 0.607 c: 0.346 AP&7EPENDIXII Part B-Continued. Locus ldh-I Idh-2 Ldh-A Ldh-B MdhlI Mdh-2 Mdhp Mpi Pep-A Pep-B Sutherland Hollow Spring b: 1.000 a: 1.000 b: 0.650 a: 1.000 a: 1.000 a: 1.000 c: 0.909 c: 1.000 b: 1.000 a: 1.000 d: 0.350 e: 0.091 Trough Spring N= 10 10 10 10 10 10 10 10 9 9 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 d: 1.000 b: 1.000 a: 1.000 West Nueces Spring N= 1 1 1 1 1 I 1 1 I 1 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 1.000 c: 1.000 b: 1.000 a: 1.000 WB Spring N= 4 4 4 4 4 4 4 4 4 4 b: 1.000 a: 1. 000 b: 1.000 a: 1. 000 a: 1. 000 a: 1.000 C: 1.000 c: 1-.000 b: 1.000 a: 0.875 H b: 0. 125 0 Camp Mystic Spring 10 10 10 10 1 0 10 10 10 10 10 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 0. 100 a: 1.000 C: 1. 000 c: 1.000 b: 1.000 a: 1.000 N= C: 0.900 Greenwood Valley Ranch Springs 6 6 6 6 6 0 6 6 6 6 0 d: 1. 000 a: 1.000 b: 1.000 a: 1.000 c: 1.000 e: 1.000 C: 1.000 C: 1.000 a: 1. 000 zl- 0 176 Spring N= 4 4 4 3 4 4 4 4 3 4 0 b: 1.000 a: 1.000 h: 0. 125 a: 1.000 a: 1.000 a: 1.000 C: 1.000 b: 1.000 a: 1.000 a: 1. 000 e: 0.875 Sabinal Group Murphy's Spring N= 4 4 4 4 4 4 4 4 4 4 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1. 000 a: 1.000 c: 0. 125 C: 1.000 b: 1. 000 a: 1.000 e: 0.875 Sabinal Canyon Spring N= 11 11 11 11 11I 11 11 11 11I 11 b: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 c: 0.409 a: 0.045 b: 1.000 a: 1.000 e: 0.591 c: 0.955 Tucker Hollow Cave N= 5 6 6 6 6 6 6 6 6 4 c: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1.000 a: 1.000 C: 1.000 C: 1.000 C: 1.000 a: 1.000 OUTGROUP Eurycea m. multiplicata N= 2 2 2 2 2 2 2 2 0 2 b: 1. 000 e: 1.000 b: 1. 000 b: 1.000 a: 1.000 e: 1.000 X: 1.000 d: 0. 750 P: 1.000 i: 0.250 E. 1. longicauda N= 2 2 2 2 2 2 2 2 0 2 b: 1.000 e: 1.000 x: 1.000 a: 0.500 a: 1.000 g: 1.000 k: 0. 750 e: 1.000 1: 1.000 0 c: 0.500 m: 0.250 APPENDIX II Part B-Continued. Locus Idh-i ldh-2 Ldh-A Ldh-B MdhIi Mdh-2 Mdhp Mpi Pep-A Pep-B H E. wilderae N= 3 3 3 3 3 3 3 3 0 3 0 b: 1.000 e: 1.000 f: 1.000 a: 1.000 a: 1.000 g: 1.000 m: 0.670 c: 0.250 b: 0.250 n: 0.330 y: 0. 750 d: 0. 750 E. bislineata N= 1 1 1 0 1 1 1 1 0 1 b: 1.000 e: 1.000 f:- 1.000 a: 1.000 g: 1.000 g: 1.000 e: 1.000 b: 1.000 E. quad-ridigitata TX N= 2 2 2 2 2 2 2 2 0 2 0 e: 1.000 e: 1.000 f: 1.000 a: 1.000 a: 1.000 g: 1.000 g: 1.000 C: 0.500 j: 0.500 0 e: 0.500 1: 0.500 E. quadridigitata SC N= 2 2 2 2 2 2 2 2 0 2 b: 1.000 x: 1.000 f: 1.000 c: 1.000 x: 1.000 d: 1.000 g: 1.000 d: 1.000 a: 1.000 Haideotriton wallacei N= 1 1 1 0 1 1 1 1 0 1 x: 1.000 b: 1.000 f:. 1.000 a: 1.000 k: 1.000 m: 1.000 c: 1.000 i: 0.500 k: 0.500 Typhlot,riton spelaeus N= 2 2 2 2 2 2 2 2 0 2 b: 0. 750 e: 1.000 f: 1.000 x: 1.000 a: 1.000 i: 1.000 x: 0. 750 z: 1.000 1: 1.000 g: 0.250 g: 0.250 APPENDIX II Part C Locus

Pep-D Pgm Pgdh Pk sSod H P A NORTH Jollyville Plateau E. tontkawae Balcones Park Spring 5 5 5 5 5 e: 0.900 a: 1. 000 e: 1.000 a: 1.000 a: 1.000 0.018 8.0 1.l N= f: 0.100 [0.013] [0.1] Barrow Hollow Spring 4 4 5 4 5 e: 1.000 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.000 0.0 1.0 N= [0.000] [0.01 Bull Creek 5 5 5 5 5 H Spring 0 e: 0.900 a: 0.900 e: 1.000 a: 1.000 a: 1.000 0.024 12.0 1.1 f: 0.100 C: 0. 100 [0.013] [0.1] Canyon Creek'Spring 2 3 3 3 3 e: 1.000 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.027 4.0 1.0 [0.027] [0.01 0 4 4 4 4 4.0 1.0 z-1 Canyon Vista Spring N= 4 0 e: 1.000 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.010 [0.0] C) [0.01] Horsethief Hollow Spring N= 1 1 1 1 1 0.0 1.0 e: 1.000 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.000 [0.01 [0.000] New Bull Creek Spring N= 1 2 2 2 2 e: 1.000 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.000 0.0 1.0 [0.000] [0.0] Schlumberger Spring 6 6 6 6 6 e: 1.000 a: 1.000 e: 1.000 a: 1. 000 a: 1.000 0.000 0.0 1.0 [0.000] [0.0] Stillhouse Hollow Springs 14 14 14 14 14 e: 1.000 a: 1. 000 e: 1.000 a: 1.000 a: 1. 000 0.006 4.0 1.0 [0.006] [0.0] Wheless Springs 4 4 5 5 5 e: 0.875 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.018 8.0 1.1 f?:0. 125 [0.013] [0.1] Round Rock Spring 9 9 9 9 9 E. tonkawae c: 0. 944 a: 1.000 b: 0.056 a: 1. 000 a: 1.000 0.014 12.0 1.1 e: 0.056 e: 0.944 [0.008] [0.1] 0l APPENDIX II Part C-Continued. Locus Pk sSod H P A Pep-D Pgm Pgdh o Lake Georgetown Area N= 2 5 o E. naufragia a: 1.000 a: 1.000 0.038 12.0 1.1 H~ Avant's Springs a: 1.000 a: 1.000 e: 1.000 [0.031] [0.1] 9 9 24.0 1.2 Buford Hollow Springs N- 9 9 9 a: 0.944 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.049 [0.1] e: 0.056 [0.022] 5 5 Cedar Breaks Spring N= 4 5 5 a: 1.000 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.026 12.0 1.1 [0.014] [0.1] 5 5 5 Knight Spring N= 5 5 a: 1.000 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.034 12.0 1.1 [0.020] [0.1] z 8 Salado N= 8 8 8 7 Springs a: 1.000 0.010 4.0 1.0 E. chisholmensis a: 1.000 a: 1.000 e: 1.000 a: 1.000 [0.010] [0.0] 0 1 Bat Well N= 1 1 1 1 a: 1.000 a: 1.000 0.000 0.0 1.0 E. naufragia a: 1.000 a: 1.000 e: 1.000 [0.000] [0.0] 2 Testudo Tube N= 2 2 2 2 a: 1.000 0.020 4.0 1.0 E. tonkawae e: 1.000 a: 1.000 e: 1.000 a: 1.000 [0.020] [0.0] 3 Kretschmarr Cave N= 3 4 4 4 a: 1.000 0.000 0.0 1.0 E. tonkawae f: 1.000 a: 1.000 e: 1.000 a: 1.000 [0.000] [0.0] Buttercup Creek Caves E. tonkawae 3 3 3 Buttercup Creek Cave 3 3 e: 0.167 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.027 8.0 1.1 f: 0.833 [0.018] [0.1] 2 Ilex Cave 2 2 2 2 e: 0.250 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.020 4.0 1.0 f: 0.750 [0.020] [0.0] 1 Treehouse Cave N= 1 1 1 1 f: 1.000 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.040 4.0 1.0 [0.040] [0.0] APPENDIX 11 Part C-Continued. Locus

Pep-D Pgm Pgdh Pk sSod H P A T.WA.S.A. Cave N= 1I 1 1 I 1.000 a: 1.000 e: E: 1.000 a: 1.000 a: 1.000 0.120 12.0 1.1 [0.066] [0.1] SAN MARCOS E. nana N= 12 11 13 8 12 h: 1.000 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.017 4.0 1.0 [0.017] [0.0] E. rathbuni N= ~ 4 5 5 4 4 d: 0.200 a: 1. 000 e: 1.000 a: 1.000 a: 1.000 0.048 16.0 1.2 g: 0.800 [0.024] [0.1] H SOUTHEAST E. latitans complex E. latitans (type locality) N= ~ 4 5 5 5 4 c: 0.625 a: 1.000 b: 1.000 a: 1. 000 a: 1.000 0.018 8.0 1.1 d: 0.375 [0.013] [0.1] Bear Creek Spring N= 6 6 6 6 6 0 C: 1.000 a: 0.917 d: 1.000 a: 1.000 a: 1.000 0.007 4.0 1.0 c: 0.083 [0.007] [0.0] 0 Chersry Creek Spring N= ~ 2 2 2 2 1 C: 1.000 a: 1. 000 b: 1.000 a: 1.000 a: 1.000 0.000 0.0 1.0 [0.000] [0.0] Cloud Hollow Spring N= 4 5 5 4 5 c: 1.000 a: 1.000 b: 1.000 a: 1. 000 a: 1.000 0.000 0.0 1.0 [0.000] [0.0] Cibolo Creek Spring N 8 8 8 8 7 C: 1.000 a: 1.000 a: 0.063 a: 1.000 a: 1.000 0.015 4.0 1.2 b: 0.063 [0.015] [0.2] c: 0.063 d: 0.250 e: 0.563 Kneedeep Cave Spring N= 6 6 6 5 5 C: 1.000 a: 1.000 b: 0.250 a: 1.000 a: 1.000 0.048 16.0 1.2 e: 0. 750 [0.028] [0.1] Honey Creek Cave Spring N 1 111 1 C: 1. 000 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.000 0.0 1.0 01 [0.000] [0.0] I? APPENDIX II Part C-Continued. Locus Pep-D Pgm Pgdh Pk sSod H P A Less Ranch Spring N= 2 2 2 2 2 c: 0.750 a: 1.000 1. 000 b: a: 1.000 a: 1.000 0.020 4.0 1.0 tTl d: 0.250 [0.020] [0.0] Rebecca Creek Spring N= 5 6 6 6 6 C: 0.900 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.008 4.0 1.0 d: 0. 100 [0.008] [0.000] E. nieotenes Helotes Spring (type locality) N- 9 10 10 10 11 c: 0.667 a: 0.400 c: 0.650 a: 1.000 a: 1.000 0.081 20.0 1.2 d: 0.333 b: 0.600 e: 0.350 [0.036] [0.1] L1- Leon Springs 2 2 2 2 2 H C: 1.000 a: 0.500 C: 0. 750 a: 1.000 a: 1.000 0.040 16.0 1.2 0- N= 0 b: 0.500 e: 0.250 [0.028] [0.1] 07 Mueller's Spring 3 4 4 4 4 C: 1.000 a: 0. 125 b: 0.500 a: 1.000 a: 1. 000 0.020 12.0 1.1 0 N- b: 0.875 e: 0.500 [0.014] [0.1] z E. pterophila Fern Bank Spring (type locality) 10 10 10 10 7 C: 1.000 a: 1. 000 b: 1.000 a: 1.000 a: 1.000 N- 0.028 8.0 1.1 [0.024] [0.1] Boardhiouse Springs 11 11 11 11 11 C: 0.955 a: 1.000 b: 1.000 a: 1.000 a: 1.000 0.007 N= 8.0 1.1 d: 0.045 [0.005] [0.1] Grapevine Cave 2 2 2 2 2 c: 0.250 a: 1.000 b: 1.000 a: 1.000 a: 1.000 0.020 4.0 1.0 d: 0.750 [0.02] [0.0] Peavey's Springs N= 5 5 5 4 5 0I C: 1.000 a: 1.000 b: 1.000 a: 1.000 a: 1. 000 0.000 0.0 1.0 [0.000] [0.0] T Cave 5 5 5 5 5 C: 0.500 a: 1.000 b: 1.000 a: 1.000 a: 1.000 0.032 8.0 1.1 d: 0.500 [0.025] [0.1] E. sosorum 11 14 14 12 11 a: 1.000 a: 1.000 b: 0.214 a: 1. 000 a: 1.000 0.053 32.0 1.3 e: 0. 786 [0.0231 [0.1] APDPENDIX 1I Part C-Continued. Locus

Pep-D Pgsn Pgdh Pk sSod H P A

Honey Creek Cave (type locality) N= 5 5 5 5 4 a: 0.100 a: 1.000 b: 0.200 a: 1.000 a: 1.000 0.032 16.0 1.2 d: 0.900 e: 0.800 [0.019] [0.1] 0) 0 Badweather Pit 4 5 5 5 5 C: 0.250 a: 1.000 b: 0. 700 a: 1.000 a: 1.000 0.108 24.0 1.2 N= d: 0.750 e: 0.300 [0.040] [0.1] Ebert Cave 4 6 6 6 6 d: 0.625 a: 1.000 b: 0.833 a: 1.000 a: 1.000 0.097 20.0 1.2 e: 0.375 e:0. 167 [0.046] [0.1] Comal Springs 12 12 1 1 1 1 12 C: 1.000 a: 0.375 b: 1.000 a: 1.000 a: 1.000 0.057 16.0 1.2 b: 0.625 [0.030] [0.1] z Pedernales Spring 1 7 7 7 5 6 N= c: 0.857 a: 1.000 b: 1. 000 a: 1. 000 a: 1.000 0.090 28.0 1.3 0 d: 0. 143 [0.038] [0.1] Spring 2 7 7 7 6 7 C: 1.000 a: 1.000 b: 0.929 a: 1.000 a: 1.000 0.057 24.0 1.2 e: 0.071 [0.027] [0.1] SOUTHWEST (E. troglodytes complex) Carson Cave Group Carson Cave 5 5 5 5 5 C: 1.000 a: 1.000 e: 1.000 a: 1. 000 a: 1.000 0.008 4.0 1.0 [0.008] [0.0] Fessenden Spring 8 8 8 8 8 a: 0.063 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.026 16.0 1.2 d: 0.875 [0.0131 [ol]1 e: 0.063 Robinson Creek Spring 3 4 4 4 4 d: 0.333 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.027 4.0 1.0 N= e: 0.667 [0.027] [0.0] Smith's Spring 14 14 14 14 14 C: 0.929 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.044 20.0 1.2 d: 0.071 [0.018] [0.1] APPENDIX 1I Part C-Continued. Locus

Pep-D Pgm Pgdh Pk sSod H P A 11 Sutherland Hollow Spring N= 10 11 11 11~~~~lx C: 1.000 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.046 I 0- n I C) [0.029] [0.1] 10 Trough Spring 9 10 10 10 C: 0. 778 a: 1.000 d: 0.050 a: 1.000 a: 1.000 0.013 8.0 1.1 N= d: 0.222 e: 0.950 [0.010] [0.1] West Nueces 1 1 1 1 Spring 0.0 1.0 C: 1.000 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.000 N= [0.000] [0.0] 4 tz WB Spring 4 4 4 4 H c: 0. 625 a: 1.000 e: 1. 000 a: 1. 000 a: 1.000 0.060 12.0 1.1 d: 0.375 [0.036] [0.1] 10 Camp Mystic Spring 4 10 10 10 a: 1.000 a: 0.900 e: 1. 000 b: 1.000 a: 1.000 0.008 8.0 1.1 N- b: 0,100 [0.008] [0.1] 0 4 6 6 6 6 z- Greenwood Valley Ranch Springs 0 c: 0.250 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.020 4.0 1.0 d: 0. 750 [0.020] [0.0] 2 4 4 4 4 176 Spring N= 0 d: 0.750 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.040 12.0 1.1 e: 0.250 [0.024] [0.1] Sabinal Group N= 4 4 Murphy's Spring 4 4 4 c: 0.875 a: 1.000 e: 1. 000 a: 1.000 a: 1.000 0.040 12.0 1.1 d: 0. 125 [0.024] [0.1] N= 1 1 1 0 11 1 1 11 Sabinal Canyon Spring 1.2 C: 0.909 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.037 24.0 d: 0.091 [0.020] [0.1] Tucker Hollow Cave 6 4 6 6 6 c: 0.500 a: 1.000 e: 1.000 a: 1.000 a: 1.000 0.040 8.0 1.1 ~-q d: 0.500 [0.029] [0.1] OUTGROUP 2 Eurycea m. multiplicata 0 2 2 2 a: 0. 750 C: 1.000 x: 0. 75 f: 1. 000 N C NC NC x: 0.250 d: 0.25 2 E. 1. longicauda 0 2 2 2 a: 1.000 0: 1.000 d: 1.000 f: 1.000 NC NC NC JAP3PENDIX II Part C-Continued.

Locus. Pep-D Pgm Pgdh Pk sSod H P A H E. wilderae N= 0 3 3 3 3 a: 1.000 e: 0.500 d: 1.000 f: 1.000 NC NC NC x: 0.333 y: 0.167 E. bislineata N= 0 1 1 1 1 z'l0 a: 1. 000 b: 1.000 d: 1.000 f: 1. 000 N C N C N C 0 E. quadridigitata TX N 0 2 2 2 2 0 b: 1.000 e: 1. 000 e: 1.000 f: 1.000 NC NC NC E. quadridigitata SC N= 0 2 2 2 2 a: 1.000 b: 1.000 X: 1.000 f: 1. 000 NC N C N C Haideotriton wallacei N= 0 1 1 1 1 a: 1. 000 e: 1.000 b: 1.000 f: 1. 000 N C N C N C Typhiotriton spelaeus N= 0 2 2 2 2 a: 1. 000 d: 0.250 d: 1.000 f: 1. 000 N C N C N C e: 0.750 APPENDIX III members Partial cytochrome b sequence for central Texas Eurycea and outgroup Positions shown are 43 to 398 within the gene. Texas group members are listed by major geographic region of occurrence. Northern E. tonkawae (Testudo Tube) ?TTAACAACTCATTTATTGATCTCCCAGCCCCTTCTACCTTATCCTACT E. tonkawae (Ilex Cave) ???? ...... E. tonkawae (Kretschmarr Cave) ??????? ...... N...... * ...... E. tonkawae (Stillhouse Hollow Springs) ???????...... N.G E. tonkawae (Horsethief Hollow Spring) ??????? ...... NNG...... H E. tonkawae (Round Rock Spring) ????? ...... NN.G...... G ...... E. naufragia (Cedar Breaks Spring) ?????...... E. naufragia (Bat Well) ?????...... C...... NNNG...... E. chisholmensis (Salado Springs) ???? ...... N .G ...... 0 San Marcos 0- E. rathbuni ?????...T..C ...... C..A... A ...C..C.A...N..T..T. 0 E. nana ?????...... C ...... C..T...N .... G..A.A.C.N..T.... Southeastern E. sosorum (Barton Springs) ???...... C ...... C ...... G..A.AN . ..T...C E. tridentifera (Badweather Pit) ?????...... C ...... C.....N.. ...C.NNNNT... ..T...C E. tridentifera (Honey Creek Cave) ???????...... C...... C..A.NNN.. ..T...C . E. tridentifera (Ebert Cave) ?????? C...... C. . .C. .A.ANN.. ..T...C

. . . .A.ATN. . ..T...C * .C. E. pterophila (Boardhouse Spring) ?????...... C...... C. *N. E. latitans (Pfeiffer's Water Cave) ?????? ..... C ...... C. o..o.o. . .C..A.AT... ..T...C E. latitans complex (Cibolo Creek Spring) ??????? ...... C C. ...oooo...... G..A.NN... ..T...C C. ...C..A.AT... ..T...C E. latitans complex (Cloud Hollow Spring) ?????????..NN...... * .N.-.. E. latitans complex (Honey Creek Cave Spring) ????? ...... C. .o.o . .. . .C..A.AT... ..T...C E. latitans complex (Rebecca Creek Spring) ??????? ..C...... C. o.oooo, . .C.NANNN ... ..T...C E. neotenes (Helotes Creek Spring) ????????...C...... C. . G....ANNN... ..T...C E. sp. (Comal Springs) ?? ... ..C ...... G. A.ATN ....T... .C APPENDIX III -- Continued.

E. sp. (Pedernales Springs) ??????????.C ...... C NN.N ...G.NA.ANNNN..T.. .C Southwestern E. troglodytes (Valdina Farms Sinkhole) ????????..C ...... C ...... ANN .T .... E. troglodytes complex (176 Spring) ?? ????? ...... C...... G. .A.ANC ...... T E. troglodytes complex (Camp Mystic Spring) ?????...... C...... C...... G .A .A .C ....T.... E. troglodytes complex (Carson Cave) ????????...... C...... G. .A.ATC.N ...... C E. troglodytes complex (Greenwood Valley Ranch) ????????...... C...... G . A.ANC....T .... E. troglodytes complex (Sabinal Canyon Spring) ?????? ...... C...... A.A.C ....T .... E. troglodytes complex (Smith's Spring) ?????????..C ...... C. .... N.N. .NG.NA.NNN.N ...... C E. troglodytes complex (Sutherland Hollow Spr.) ???????.... C...... NA A.ANN ....T.... E. troglodytes complex (Trough Spring) ???????.... C...... G. .A.ANC ....T .... H E. troglodytes complex (Tucker Hollow Cave) ?????????..C. C. ....C .A..A. NNC ....T...C Outgroup E. bislineata .T...... T ...... c... .C .....A. ...A.C.N..T...C E. longicauda ?????T .....C. .C.....A. .A.GN...... C E. multiplicata ?????...... C...... C... .C .....A. .A.G.C.G.C....C 0 E. SC z'l quadridigitata ?????...... C...... c.... .C .....A. .A.G.C.G.C....C 0 E. quadridigitata TX ?????T .....T. .CA ....A. ...GTC.G.C..... E. wilderae ???????....T...... c ... .C.....A. . .NGNN.N. .T.... Haideotriton wallacei A.C..T .....T. .CA....A. ...GTC.G.C..... 0 Typhlotri ton spelaeus ??????..T.... .Co...... CA... .G. .A.G.C.G.CT..T.

Northern E. tonkawae (Testudo Tube) TATGAAATTTTGGCTCTCTCTTAGGAGTCTGCCTAATTTCACAAATTCT E. tonkawae (Ilex Cave) ...... o.. ...oo . . * * . ,. o o o .o . o. E. tonkawae (Kretschmarr Cave) .o...... o...o...... o...o......

E. tonkawae (Stillhouse Hollow Springs) .oo....C ...... ?...

E. tonkawae (Horsethief Hollow Spring) ...... C ...... o.o. .o. . .o.o.... . E. tonkawae (Round Rock Spring) ...... C...... E. naufragia (Cedar Breaks Spring) E. naufragia (Bat Well) .T...... E. chisholmensis (Salado Springs) San Marcos E. ra thbuni ...... C.....A .C A ...... C.T...... AT......

APPENDIX III Continued.

E. nana .. ...A...T ..... CA ...... A ...... T. Southeastern E. sosorum (Barton Springs) ..... A..T .....CA ...... A ...... T. E. tridentifera (Badweather Pit) ...... T .....CA . . A...... T. E. tridentifera (Honey Creek Cave) . ....A...T .....A ...... A ...... T. E. tridentifera (Ebert Cave) ...... C .....A .....T .....CA ...... A ...... T. E. pterophila (Boardhouse Spring) ...... TA .....CA . ..A ...... T. E. latitans (Pfeiffer's Water Cave) ...... C .....A .....T .....CA .....T .....A ...... T. E. latitans complex (Cibolo Creek Spring) ...... A...... T .....CA...... A ...... T. E. latitans complex (Cloud Hollow Spring) .....A...T.....ACA ...... A...... T. E. latitans complex (Honey Creek Cave Spring) ..... C ...... A...... A ...... T. H E. latitans complex (Rebecca Creek Spring) .....A...T .....A ...... A ...... T. 0 E. neotenes (Helotes Creek Spring) C...... T.....CA ...... A ...... T. E. sp. (Comal Springs) C...... A . T .....CA ...... A ...... T. E. sp. (Pedernales Springs) .N. .N. C.....A .....T .....TA ...... A...... T. ~z Southwestern 0 E. troglodytes (Valdina Farms Sinkhole) C.....A..C..T ...... A ...... T. 0 E. troglodytes complex (176 Spring) ...... C...... C...... A ...... T. E. troglodytes complex (Camp Mystic Spring) ...... A..C..T ...... CA ...... T. E. troglodytes complex (Carson Cave) C.A..C..T.AT...... A...... TC E. troglodytes complex (Greenwood Valley Ranch) .....A..C..T...... A ...... T. E. troglodytes complex (Sabinal Canyon Spring) C.....A..C..T...... A ...... T. E. troglodytes complex (Smith's Spring) ...... C .... A..C..T .... .T.NNNN ...... A ...... T. E. troglodytes complex (Sutherland Hollow Spr.) C.....A..C..T..A ...... T. E. troglodytes complex (Trough Spring) ...... C. .T ...... A...... T. E. troglodytes complex (Tucker Hollow Cave) ...... C...... C..T ...... A ...... T. Outgroup E. bislineata ...... CC.T ...... T .....A ...... CT. E. longicauda .C...... C ...... A.T...... CA..CT. E. multiplicata ...... A..CC.T ...... CA ... .G. .CT, E. quadridigitata SC ...... A..CC.T ...... CA ... .G .CT. E. quadridigitata TX ...... A. .CC.T ...... A.T ...... A ...... T. E. wilderae ...... C...... CA.T ...... A . ..T .....A ...N ...CT. Haideotri ton wallacei ...... C. .A.. CC.T...A.T ...... A. .. A ...... CT. Typhlotri ton spelaeus ...... C..C..A. .. C.A ...... A ...... A...... C.. APPENDIX III Continued. Northern E. tonkawae (Testudo Tube) AACAGGCCTATTTCTCGCAATACACTATACTGCAGATACTACTTCCGCA E. tonkawae (Ilex Cave) .C*.. C...... E. tonkawae (Kretschmarr Cave) ...... C..C..C ...... C E. tonkawae (Stillhouse Hollow Springs) .o ...... o.o.oooo...... o. E. tonkawae (Horsethief Hollow Spring) ...... E. tonkawae (Round Rock ***...... o oo*.o o..... Spring) ...... T...... E. naufragia (Cedar Breaks Spring) ...G...T...... T..* , ...... C ...... E. naufragia (Bat Well) ...G...T...... T...... T.. C...... C...... A...G. E. chisholmensis (Salado Springs) ...G...T. o...... C...... C ..C ..C ..A ... San Marcos H E. ra thbuni ...... T ...... T ...... C .C .C ..A ... E. nana .G.....C..C..C..A..... Southeastern ~z E. sosorum (Barton Springs) . . G...T ...... T...... T..C ...... C..C..C..A . .. E. tridentifera (Badweather Pit) . ..G. . .T ...... T...... T. .C...... C..A ... 0 0 E. tridentifera (Honey Creek Cave) ...G..T...... T...... T..C...... C..C..C..A ... E. tridentifera (Ebert Cave) ...G...T ...... T ...... T..C ...... C..C..C..A... E. pterophila (Boardhouse Spring) ... .T ...... T...... T..C...... C..C..C..A... E. latitans (Pfeiffer's Water Cave) ...G...T...... T...... T..C...... C..C..C..A... E. latitans complex (Cibolo Creek Spring) ... T...... T...... T...... C..C. .A... E. latitans complex (Cloud Hollow Spring) G ...T ...... T..C.C..C..C..A... E. latitans complex (Honey Creek Cave Spring) G...T ...... T...... T..C ...... C..C..C..A... E. latitans complex (Rebecca Creek Spring) ...G...T...... T...... T..C ...... C..C..C..A... E. neotenes (Helotes Creek Spring) ...T ...... T ...... T..C ...... C..C..C..A .. E. sp. (Comal Springs) ...G . T ...... T. .C ...... C C. .C. .A... E. sp. (Pedernales Springs) ...G...T ...... T...... T..C...... C..C..C..A... Southwestern E. troglodytes (Valdina Farms Sinkhole) ...G... T...... T...... T...... C..C..C..A... E. troglodytes complex (176 Spring) E. troglodytes complex (Camp Mystic Spring) G...T . T. T... C.... C..A... E. troglodytes complex (Carson Cave) ...G...T...... T...... T...... C..C..C..A... E. troglodytes complex (Greenwood Valley Ranch) APPENDIX III Bo Continued. 0z

E. troglodytes complex (Sabinal Canyon Spring) ...G...T ...... T...... T...... C..C..C..A... E. troglodytes complex (Smith's Spring) ...... T...... TT ...... C..C..C..A... E. troglodytes complex (Sutherland Hollow Spr.) . T.T..... T ...... C..C..C. .A... E. troglodytes complex (Trough Spring) ...... T. T ..T...... C..C..C..A... E. troglodytes complex (Tucker Hollow Cave) ...G..T ...... T . C..C..A... Outgroup E. bislineata ..... T...... G.....C.....T...... C. E. longicauda ...... AT...... T...... T ...... C .... . C..... C...... E. multiplicata ...... GT....C..T ...... C..C.....C..CGTC..A... E. quadridigitata SC ...... C T ...... C. ....C..CT.C..A.... E. quadridigitata TX ...... T...... T ...... C..C..C...... C. .C z E. wilderae ...... A ...C...... G..C...... C. .C ...... ~tT Haideotriton wallacei 0 T .....AT. T...... C..C.G..C..CT.C..... 0 Typhlotri ton spelaeus ...... C..T...... C..C.....C..CTC...... i-i 0 m

Northern 0Q E. tonkawae (Testudo Tube) TTTTCCTCCGTGGCCCACATTTGCCGTGATGTAAATTATGGCTGACTTG z0 E. tonkawae (Ilex Cave) ...... 0 E. tonkawae (Kretschmarr Cave) 0 ...... 0 E. o ...... tonkawae (Stillhouse Hollow Springs) ...... > E. tonkawae (Horsethief Hollow Spring) * ***o - o ...... **...... E. tonkawae (Round Rock Spring) ... . CA3 E. naufragia (Cedar Breaks Spring) ...... co E. naufragia (Bat Well) '"????"'?*?N'...... C ... ..T ..A ...... E. chisholmensis (Salado Springs) San Marcos E. rathbuni ?.C .....T . .A ...... C ...... C ...... E. nana Southeastern E. sosorum (Barton Springs) ..C .....T ..A E. tridentifera (Badweather Pit) .T..A ...... E. tridentifera (Honey Creek Cave) .C .....T..A ...... E. tridentifera --I (Ebert Cave) .C .....T..A ...... C.0 E. pterophila (Boardhouse Spring) T..A. APPENDIX III Continued. 1

E. latitans (Pfeiffer's Water Cave) ..C.. ...T..A...... ****.. .*......

E. latitans complex (Cibolo Creek Spring) * .C.. ...T..A. ..* .*......

E. latitans complex (Cloud Hollow Spring) ..C.. ...T..A...... * ...... *.*......

E. latitans complex (Honey Creek Cave Spring) * .C.. ..T. .A...... *...... *... * ...... E. latitans complex (Rebecca Creek Spring) . .C.. ...T..A...... * E. neotenes (Helotes Creek Spring) ..C.. ...T..A. * * ** * * * . .* *...... N. . * . .. E. sp. (Comal Springs) . .C.. ...T..A. * ...... *** ...... ?o.o...** .o.. E. sp. (Pedernales Springs) . .C.. . ..T..A. * * ...... **** ** . *. *. * * .....C . Southwestern E. troglodytes (Valdina Farms Sinkhole) . ....T ...... G...c...... E. troglodytes complex (176 Spring) C .....T..A ...... T.. . .T H E. troglodytes complex (Camp Mystic Spring) ...... T..AAT ...... 0 E. troglodytes complex (Carson Cave) .....T..A ...... C. E. troglodytes complex (Greenwood Valley Ranch) .C .... . T.....T ...... C...... E. (Sabinal troglodytes complex Canyon Spring) ..C.... . T.....T...... *.* ~z E. troglodytes complex (Smith's Spring) ..C.....T..A ...... N...... C, E. troglodytes complex (Sutherland Hollow Spr.) ..C...... T...... 0 ...... C...... 0 E. troglodytes complex (Trough Spring) ...... T..AAT......

E. troglodytes complex (Tucker Hollow Cave) .C .....T .....T ...... C..N ... .. Outgroup E. bislineata ..C.....T. .A. .T. .T...... A..C .....C. ....A .....G. E. longicauda ..C.....T..A.....T. .C .....A.. .C. .T ...... T .....A. E. multiplicata ..C..T. .C .....A..C...... A . E. quadridigitata SC .C.....T..A .....T. .C .....A. .C ...... A . E. quadridigitata TX ...... A..T...... C .....A ...... C..T...... E. wilderae . .C ...... T ...... A.....GGN...... T...... Haideotri ton wallacei ...... T. . T...... A ...... T .....A. Typhlotri ton spelaeus . . . . . T..A. .A. .A. .T. .C .....A...... C. .A ....G.

Northern E. tonkawae (Testudo Tube) TACGCAGCATTCACACTAATGGAGCATCACTATTCTTTATTTGTATGTA E. tonkawae (Ilex Cave) E. tonkawae (Kretschmarr Cave) .A...... APPENDIX III to Continued. C)

.....A.. E. tonkawae (Stillhouse Hollow Springs) .G...... G...... N..C ..C .....A .. E. tonkawae (Horsethief Hollow Spring) .G ....T ...... G ...... N ...... A .. E. tonkawae (Round Rock Spring) .G .....T...... A .. E. naufragia (Cedar Breaks Spring) ...... AT ...... C ...... A.. E. naufragia (Bat Well) ,. .T...... C ..... A .. H~ E. chisholmensis (Salado Springs) ..... T ...... San Marcos A.C ...... C.....T.. E. ra thbuni .G. .T.AT ...... G...... C .C. .T.. E. nana .G .....T...... CA.C .....CA.C Southeastern CA.C ...... C..C..T.. E. sosorum (Barton Springs) .G .....T ...... C..C .. . .. CA.C ...... C. .C. .T.. E. tridentifera (Badweather Pit) .G.....T ...... C .. . .. CA.C ...... C. .C. .T.. E. tridentifera (Honey Creek Cave) .G.....T ...... C ...... C. .C. .T.. E. tridentifera (Ebert Cave) .G .....T ...... C..C.....CA.C .. . ..CA.C ...... C..C..T.. E. pterophila (Boardhouse Spring) .G .....T ...... C..C ~z ...... C. .C. .T.. 0 E. latitans (Pfeiffer's Water Cave) .G .....T...... C..C .. . ..CA.C . C.CA.C..T.N .....C ... E. latitans complex (Cibolo Creek Spring) .G .....T...... C. 0 ...... C. .C. .T.. E. latitans complex (Cloud Hollow Spring) .G .....T ...... C..C.....CA.C CA.C ...... C. .C. .T.. E. latitans complex (Honey Creek Cave Spring) .G .....T ...... C.C ..... CA.C ...... C..C..T.. E. latitans complex (Rebecca Creek Spring) .G ...T ... C...... C CA.C ...... C..C..T.. E. neotenes (Helotes Creek Spring) .G .....T ...... C..C ...... G ...... C.. . ..CA.C ...... C. .C. .T.. E. sp. (Comal Springs) .G .....T ...... C.....CA.C ...... C..C..T.G E. sp. (Pedernales Springs) Southwestern C-c ...... CA .C ...... T .. E. troglodytes (Valdina Farms Sinkhole) ...... T...... CA.C ...... T.. E. troglodytes complex (176 Spring) ...... T...... A...... T. .CA.C ...... T.. E. troglodytes complex (Camp Mystic Spring) ...... C...... CACC...... C .....T.. E. troglodytes complex (Carson Cave) ...... T...... T. CA .C ...... T .. E. troglodytes complex (Greenwood Valley Ranch) ...... T...... CA .C ...... T .. E. troglodytes complex (Sabinal Canyon Spring) ...... G..CACC ...... C .....T.. E. troglodytes complex (Smith's Spring) ...... T...... T...... CA.C.CT ...... T.. E. troglodytes complex (Sutherland Hollow Spr.) ...... A...... CA.C.CT ...... T.. E. troglodytes complex (Trough Spring) ...... T...... T..CG.CT.. E. troglodytes complex (Tucker Hollow Cave) Outqroup APPENDIX III Continued.

E. bislineata .G..T.AT .....T ...... C..TA.T..T ...... T...... C. .T.. E. longicauda .G ...... T ...... T..TA ... CT. .N ...... T.. E. multiplicata ...... C .. C...... C T.. E. quadridigitata SC ...... C ...... C. . TA . T . T. . C . . . . . C . . T . . E. quadridigitata TX ..A T..C . E. wilderae ...... AT .....T.....C .....C..TA.T. .T ...... T.. Haideotri ton wallacei .G....AT ...... C..TA.T. .T..C .....C..T.. . C Typhlotri ton spelaeus ...... ATG ...... C.A.T ...... T..

H Northern E. tonkawae (Testudo Tube) TCTACACATTGGACGGGGCCTATATTATGGCTCTTACATATTTAAAGAA E. tonkawae (Ilex Cave) ...... o...... E. tonkawae (Kretschmarr Cave) . o ...... o...... o...... T 0 E. tonkawae (Stillhouse Hollow Springs) .A...... E. tonkawae (Horsethief Hollow Spring) .A ...... T ...... 0 E. tonkawae (Round Rock Spring) .A ...... o.. ... T ...... E. naufragia (Cedar Breaks Spring) 0 Well) C T...... E. naufragia (Bat .T...... * E. chisholmensis (Salado Springs) San Marcos E. rathbuni C..T...... A.. E. nana C..T..T ...... T ...... G.A...G. Southeastern .T. .G.A...G... E. sosorum (Barton Springs) C..T..T ...... T...... A. E. tridentifera (Badweather Pit) C..T..T...... T...... A. .T ..G. G .. E. tridentifera (Honey Creek Cave) C..T..T ...... T ...... A..T..G.A...G... E. tridentifera (Ebert Cave) C..T..T ...... T ...... A..T..G.A...G... E. pterophila (Boardhouse Spring) C..T..T ...... T ...... A..T..G.A...G... E. latitans (Pfeiffer's Water Cave) C..T..T ...... A ...... G ... E. latitans complex (Cibolo Creek Spring) C..T..T ...... T ...... A..T..G.A...G... E. latitans complex (Cloud Hollow Spring) C..T..T ...... T...... T ...... A..T..G.A...G...... A..T..G.A ... G... E. latitans complex (Honey Creek Cave Spring) C..T..T ...... T...... A ..T..G.A...G... E. latitans complex (Rebecca Creek Spring) C T T...... T...... A..T. .G.A ... G... APPENDIX III Continued.

E. neotenes (Helotes Creek Spring) C..T..T...... T...... A ..G . . .. G...G E. sp. (Comal Springs) C. .T..T...... T...... G.A...G... E. sp. (Pedernales Springs) . ..T..T ...... T...... A .....G.A ...G... Southwestern E. troglodytes (Valdina Farms Sinkhole) ...T..T .....G ...... -G-. - - ...... A. .T....A..... E. troglodytes complex (176 Spring) C..T..T ...... A...... A..... E. troglodytes complex (Camp Mystic Spring) C..T. .T .....G ...... A ...... A. .. . . E. troglodytes complex (Carson Cave) C..T..T ...... A. .T ....A..... E. troglodytes complex (Greenwood Valley Ranch) C. .T. .T .....G ...... A. .T....A..... E. troglodytes complex (Sabinal Canyon Spring) .. .T...... G...... A ...... A. .. . . E. (Smith's C. .T. .T ...... A. troglodytes complex Spring) ...... E. troglodytes complex (Sutherland Hollow Spr.) ...T..T .....G...... G...... A . .T....A..... E. troglodytes complex (Trough Spring) C. .T. .T .....G ...... A. E. troglodytes complex (Tucker Hollow Cave) C..T. .T .....G...... A...... Outgroup z E. bislineata .. .T...... C..A. .... T..G..T ...... 0- E. longicauda .. .T ...... A...... A .....G..C... 0 E. multiplicata oooT...... T...... A ..T....A .... E. quadridigitata SC . .T ...... A. ....N..A..T ...... (In E. quadridigitata TX ...G . .C .....A...... A..T.....N... E. wilderae ... C..T .....C. .A. .N. . .C. .... T..A ...... N... Haideotri ton wallacei .T..T.... .CG..A. ... .G...... T...... Typhlotri ton spelaeus ..T ...... A ...... A..T ....A...... G

Northern E. tonkawae (Testudo Tube) ACCTGAAACATCGGAGTTATTCTTTTGTTTTTAGTAATAGCGACAGCAT E. tonkawae (Ilex Cave) ..... G...... N. N ...... E. tonkawae (Kretschmarr Cave) ...... G ...... NN E. tonkawae (Stillhouse Hollow Springs) ..... G...... N ...... E. tonkawae (Horsethief Hollow Spring) ...... N...... E. tonkawae (Round Rock Spring) ...... E. naufragia (Cedar Breaks Spring) .....G ...... N ...N ... E. naufragia (Bat Well) ...... NNNN APPENDIX III Continued. oo

E. chisholmensis (Salado Springs) ...... N ... San Marcos E. ra thbuni ..T ...... T .....C .....A ..A ...... T .....A ...... E. nana ..T. T...... T...... A..A ...... A ...... Southeastern E. sosorum (Barton Springs) .....T..T.T ...... G.. A...... T .....A ...... E. tridentifera (Badweather Pit) .T... ..T..T ...... A...... T...N.A ..N..N E. tridentifera (Honey Creek Cave) .T.....T..T...... A...... T.....A...N.?? E. tridentifera (Ebert Cave) .T.....T..T ...... A...... T...N.A ...... C E. pterophila (Boardhouse Spring) .T.....T..T...... A...... T .....A...???? E. latitans (Pfeiffer's Water Cave) .T.....T..T ...... A...... T...N.A...N..? H E. latitans complex (Cibolo Creek Spring) .T .....T..T ...... A...... E. latitans complex (Cloud Hollow Spring) . ..T...... A...... T. . N.A ..... C E. latitans complex (Honey Creek Cave Spring) .T .....T..T ...... A...... T ..... A ...... E. latitans complex (Rebecca Creek Spring) ...... A...... T ..... A.....?? E. neotenes (Helotes Creek Spring) ...... T T. .T...... G A...... T .....A..NN.?? E. (Comal Springs) T .. .T. .T...... GA...... T..... A ...... 0 sp. T.. 0-1 E. sp. (Pedernales Springs) .T...... T..TT...... C..... G. A...... T..... A ...... Southwestern 00 E. troglodytes (Valdina Farms Sinkhole) .T.....T. .T . .. .C...T.A...... E. troglodytes complex (176 Spring) .T ...... T .... .C...T.A...... A .C .....A...... E. troglodytes complex (Camp Mystic Spring) .T.....T. .T . . . .C...T.A...... A..NN.N E. troglodytes complex (Carson Cave) .T .....T. .T .... .C...T.AN. .C...... T .....A ...... E. troglodytes complex (Greenwood Valley Ranch) .T .....T. .T .... .C...T.A.. .C...... C. ..N.A ...... E. troglodytes complex (Sabinal Canyon Spring) .T .....T. .T . . .. .C...T.A...... A ...... E. troglodytes complex (Smith's Spring) .T .....T. .T . .. .C...T.A...... NN ...... E. troglodytes complex (Sutherland Hollow Spr.) .T.....T. .T . . .. .C...T.A...... A ...... E. troglodytes complex (Trough Spring) .T .....T. .T .... .C...T.A...... N . . ? E. troglodytes complex (Tucker Hollow Cave) .T .....T. .T . ... .C...T.A...... A ...... Outgroup E. bislineata ...... T .. ..A . .... C.A ...... A ...... E. longi cauda .T ...... CN.AC.A...... G .....A. E. multiplicata .T .... .C..C..CC.A ...... C..G..T. E. quadridigitata SC . . . . *. T* . .T .....C..C... C.A...... A . E. quadridigitata TX . . . . . T. .T.... .C..C..NC.A...... A . APPENDIX III Continued.

E. wilderae ...... N .C .A ...... A ...... Haideotri ton wallacei ...... T..T .....C ... ..CC ...... A ....N.. Typhlotri ton spelaeus ...... A . ....CC ....C.....A ......

toI Northern HZ E. tonkawae (Testudo Tube) TTGTTGGGTA??? E. tonkawae (Ilex Cave) ...... ?????.. E. tonkawae (Kretschmarr Cave) ????????????? E. tonkawae (Stillhouse Hollow Springs) ...... TGT E. tonkawae (Horsethief Hollow Spring) ...... TGT E. tonkawae (Round Rock Spring) ...... ????? ? ? ? ? E. naufragia (Cedar Breaks Spring) .????????????7 7 7 7 7 7 E. naufragia (Bat Well) ?????????????...... E. chisholmensis (Salado Springs) ????????????? ~z San Marcos E. ra thbuni .C .....??????? 0 E. nana .N...???????? 0 Southeastern z) E. sosorum (Barton Springs) .... ????????? E. tridentifera (Badweather Pit) NC??????????? E. tridentifera (Honey Creek Cave) ?????????????

* * E. tridentifera (Ebert Cave) ?????????????* * * * 7* * . E. pterophila (Boardhouse Spring) ????????????? E. latitans (Pfeiffer's Water Cave) NC??????????? E. latitans complex (Cibolo Creek Spring) .C??????????? E. latitans complex (Cloud Hollow Spring) ????????????? E. latitans complex (Honey Creek Cave Spring) .C..C.??????? E. latitans complex (Rebecca Creek Spring) ?????????????I? ? 1 7 7 7? ? 7 7 7 7 7 E. neotenes (Helotes Creek Spring) NC. . ????????? E. sp. (Comal Springs) C ....??????? E. sp. (Pedernales Springs) .C..C..A..TG? Southwestern E. (Valdina Farms Sinkhole) -4 troglodytes cc E. troglodytes complex (176 Spring) [No. 80 80 HERPETOLOGICAL MONOGRAPHS [No. 14~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~14

APPENDIX III Continued.

E. troglodytes complex (Camp Mystic Spring) NC.????????? E. troglodytes complex (Carson Cave) .C..C..A?????C,C? A9'9 ...... E. troglodytes complex (Greenwood Valley Ranch) ????????????? E. troglodytes complex (Sabinal Canyon Spring) .C..C.??????? E. troglodytes complex (Smith's Spring) .C..C..??????...... E. troglodytes complex (Sutherland Hollow Spr.) .C??????????? E. troglodytes complex (Trough Spring) ????CC?????? E. troglodytes complex (Tucker Hollow Cave) ?????????????...... ?...... 7 ...... Outgroup E. bislineata ...... ??????? E. longicauda ????????????? E. multiplicata E. quadridigitata SC E. quadridigitata TX E. wilderae Haideotriton wallacei .C..A..... ??? Typhlotri ton spelaeus .C..A .....TG?