THE SOUTHWESTERN NATURALIST 62(1): 1–13 MARCH 2017

GENETIC CHARACTERIZATION AND HABITAT USE OF EURYCEA

PTEROPHILA SALAMANDERS FROM JACOB’S WELL, HAYS COUNTY,
TEXAS

JEAN K. KREJCA,* DIANA J. MCHENRY, KRISTA M. MCDERMID, ZACHARY C. ADCOCK, AND MICHAEL R. J. FORSTNER

Zara Environmental LLC, 1707 West Farm to Market 1626, Manchaca, TX 78652 (JKK, KMM)
Department of Biology, Texas State University, 601 University Drive, San Marcos, TX 78666 (DJM, ZCA, MJRF)
*Correspondent: jean@zaraenvironmental.com

ABSTRACT—We sampled Eurycea pterophila, typically considered a spring-associated species, from unprece-
dented depths within Jacob’s Well, a natural cave in the Trinity Aquifer. Nearly all of the captures followed
high spring-flow events (25–70 cubic feet per second [cfs]), with almost no individuals seen or trapped during
low flow (<10 cfs). We analyzed the cytochrome b mitochondrial gene DNA sequences and identified four
haplotypes present at Jacob’s Well, two of which were unique to the site. However, this variation fit within the
range of genetic variation for E. pterophila, indicating there is only one species present. Comparisons with other
central Texas species indicate this population is part of the previously reported polytomy referred to as the
Eurycea neotenes complex. Because our study had a large sample size from a single locality, we were able to
detect discord between our phylogenetic tree, which showed uniformity among individuals from Jacob’s Well
and nearby populations, and our FST analyses, which showed no migrants among these populations. These
observations highlight the need for larger sample sizes to illuminate relationships among central Texas Eurycea
populations.

RESUMEN—Muestreamos a Eurycea pterophila, tı́picamente considerada una especie asociada con los
manantiales, de profundidades sin precedentes dentro de Jacob’s Well, una cueva natural en el acuı́fero
Trinity. Casi todas las capturas de la salamandra fueron después de eventos de alto flujo del manantial (25-70
pies cúbicos por segundo [cfs]), con casi ningún individuo visto durante flujo más bajo (<10 cfs). Analizamos
las secuencias del gene mitocondrial de ADN de citocromo b e identificamos cuatro haplotipos presentes en
Jacob’s Well, con dos únicos a esta sitio. Sin embargo, esta variación cuadró dentro de la gama de variación
genética para E. pterophila, indicando que hay sólo una especie presente. Las comparaciones con otras especies
del centro de Texas indican que esta población es parte de la politomı́a previamente conocida como complejo
de Eurycea neotenes. Porque nuestro estudio tuvo un gran tamaño de muestra de una sola localidad, fue posible
detectar discordia entre nuestro árbol filogenético, que mostró uniformidad entre individuos del Jacob’s Well
y de poblaciones cercanas, y nuestro análisis FST, que no mostró migrantes entre estas poblaciones. Estas
observaciones subrayan la necesidad de tomar muestras más grandes para iluminar las relaciones entre las
poblaciones de Eurycea del centro de Texas.

Jacob’s Well is a perennial spring that forms the collected from within the cave in 2007, exhibits a robust
headwaters of the lower third of Cypress Creek near cranial morphotype and unique haplotype (Glenn Long-
Wimberley, Texas. The spring opening can be entered ley, pers. comm.) in comparison to Eurycea collected from
using scuba gear, providing access to the second-longest the surface at the site. This historic specimen prompted
underwater cave in Texas (1,587 m), and is renowned for research into the salamander population at Jacob’s Well
consistent outflow from the Trinity Aquifer. Increased in order to determine whether the site hosts a single,
water use in the drainage basin threatens continued recognized species of Eurycea or whether there is a spring-
outflow at Jacob’s Well. The cave is inhabited by both associated morph as well as a cave-adapted morph
aquifer- and spring-adapted fauna, including a popula- present. Taxonomists report such cases of sympatry in
tion of Fern Bank salamanders, Eurycea pterophila, which other species of central Texas Eurycea, including E.
have been collected both from the bed of Cypress Creek sosorum and E. waterlooensis from Barton Springs in Travis
in the vicinity of the spring (Lucas et al., 2009) and from County (Hillis et al., 2001) and E. nana and E. rathbuni
within the cave itself (current study). from San Marcos Springs in Hays County (Bishop, 1941;
A single historical specimen of Eurycea, reportedly Chippindale, 2005; Chippindale and Fries, 2005).
2 The Southwestern Naturalist vol. 62, no. 1

Prior to this study, Eurycea collections at Jacob’s Well baited traps on each subsequent dive through May 2011 and did
used in genetic studies had occurred only at the mouth of not replace the trap during our last dive in June 2011.
the spring, and this fueled questions about the potential Between May 2009 and June 2011 we collected samples from
for a sympatric congener. Lucas et al. (2009) sampled at 46 Eurycea at Jacob’s Well and combined them with samples
Jacob’s Well and interpreted their population genetic obtained from other central Texas springs. The majority of the
salamanders obtained from Jacob’s Well were very small
results to indicate that central Texas spring-associated
juveniles that were recently dead or very nearly dead when
Eurycea do not show evidence of migration and that the retrieved from the driftnet. Samples were collected and stored
Jacob’s Well population is therefore isolated. More in 70% ethanol at -808C. We deposited tissues in the Michael R.
recently, Bendik et al. (2013) included two individuals J. Forstner frozen tissue catalog at Texas State University–San
from this site in a phylogenetic analysis of eastern Marcos. We sampled Eurycea under Department of the Interior,
Blepsimolge (a clade within central Texas Eurycea [Hillis U.S. Fish and Wildlife Service, Federal Fish and Wildlife Permit
et al., 2001]). Those authors found the Jacob’s Well Number TE676811-0 and Texas Parks and Wildlife Scientific
specimens to fall within a large Eurycea complex polytomy Permit Numbers SPR-0102-191, SPR-0290-022, and SPR-0390-
that includes salamanders from sites recognized to have 045, and under Institutional Animal Care and Use Committee
approvals 0715_0428_07, 04-3D2AAE71, 04-046 E25 EBSA, and
E. latitans, E. tridentifera, and E. pterophila as well as
1010_0501_09.
unrecognized diversity. This polytomy also includes the
Lab personnel used a DNeasyt Tissue Kit (Qiagen Inc.,
type localities for these three species. In an example with
Venlo, Netherlands) to isolate DNA from tissue (1–2 mm3) and
fish, researchers have detected genetic differentiation sequence a 1,026-base pair portion of the mitochondrial
even in cases of abundant migration (Tatarenkov et al., cytochrome b gene. We chose to use sequences from cyto-
2010). Hence, in addition to using genetic data to chrome b because the database for homologous sequences is
distinguish between possible cryptic species (‘‘surface- larger than that for ND4. Amplifications used primers MVZ15
adapted’’ and ‘‘cave-adapted’’ morphs), we sought to (Chippindale et al., 2000) and EURCB9 (Hillis et al., 2001) in
evaluate whether this location contains a unique lineage 50-ll reactions with 4 mM MgCl2, 0.1 mM dNTPs, 0.01 lM each
of evolutionarily distinct Eurycea. To answer that question primer, 2.5 units GoTaqt Flexi DNA polymerase (Promega
we reviewed the context with other populations, previous Corp., Fitchburg, Wisconsin), and a pH of 8.5 for amplification.
Polymerase chain reaction (PCR) parameters included an initial
studies, and additional data from specimens outside of
denaturing period of 958C for 5 min, then 35 cycles each
the study site. consisting of denaturing at 958C for 30 s, annealing at 558C for 1
min, and extension at 728C for 1 min, and a final extension
MATERIALS AND METHODS—We sampled individuals across period of 728C for 5 min. The PCR products were purified with
central Texas (Table 1), nonconsumptively where possible, and an AMPuret PCR Purification System (Agencourt Bioscience
used a combination of sequences generated by us and those Corporation, Beverly, Massachusetts) and then cycle-sequenced
available in GenBank for genetic analyses (see Supplemental with the above primers using a BigDyet Terminator v3.1 Cycle
Table available online at http://www.bioone.org/toc/swna/62/ Sequencing Kit (Applied Biosystems, Foster City, California).
1). We surveyed the surface habitat by examining the substrate Thermal cycling began with an initial denaturing period of 968C
in the bed of Cypress Creek from the spring opening to 15 m for 1 min, then 25 cycles each consisting of 968C for 1 min, 508C
downstream. We used scuba gear to survey the subsurface for 5 s, and 608C for 4 min. We cleaned the reactions of reagents
habitat (subsurface in this paper refers to all areas of the cave by ethanol precipitation and analyzed them on an ABI 3500xL
from the spring opening and below), including the series of Genetic Analyzer (Applied Biosystems). The resultant sequences
entrance shafts and areas beyond the constriction to about 42 m were edited in Sequenchere 4.5 (Gene Codes Corp., Ann Arbor,
depth and about 215 m penetration. We performed scuba Michigan).
surveys on 20 occasions from May 2009 to June 2011. Divers can To assess the phylogenetic relationship of Jacob’s Well
only enter the cave safely when the spring discharge is at or salamanders within central Texas Eurycea, we analyzed 153
below 20 cubic feet per second (cfs), and heavy rains between GenBank sequences in addition to 127 sequences generated by
January and April 2010 increased the flow of the spring to levels us (total = 280; Table 1). These 280 sequences included 69 from
that precluded scuba research during that time. We obtained Jacob’s Well: 45 subsurface samples from this study (we did not
discharge measurements from the United States Geological obtain a genetic sequence from one sample), 22 surface samples
Survey (USGS) gauging station on site (USGS 08170990) (Fig. (Lucas et al., 2009), and two samples from Jacob’s Well without
1). associated surface or subsurface designation (Bendik et al.,
A team of experienced, technically proficient divers collected 2013). We included all unique central Texas Eurycea cytochrome
Eurycea specimens using basters, dipnets, and centrifuge vials. In b sequences available on GenBank that were greater than 975
May 2009, divers deployed a driftnet at the first (vertical) base pairs and had locality data associated with the accessioned
constriction at about 25-m depth and 42-m penetration as well as sequence. In instances where the same individual salamander
eight salamander traps in the wide room beyond the constric- had multiple cytochrome b sequences available on GenBank
tion at about a 30-m depth and 70-m penetration. Divers baited from different studies, we chose the sequence with the least
inverted funnel bottle traps with Wellness Brand Fish and Sweet missing data. All sequences included in the analyses are labeled
Potato Formula Dog Food (Wellness Pet Food, Tewksbury, with their GenBank accession number or collector number and
Massachusetts), weighted with rocks from the site and affixed to collection locality (Figs. 2 and 3). We provide collection locality
2-lb weights. We retrieved and replaced traps with clean, freshly instead of taxon name because the species identity has not been
March 2017 Krejca et al.—Eurycea pterophila salamanders from Jacob’s Well, Hays County, Texas 3

TABLE 1—Taxon, sampling location, number of samples, and haplotypes for all cytochrome b sequences (GenBank and this study)
from Eurycea salamanders used in the phylogenetic and haplotype network analyses. Exact locality and haplotype data are only
provided for central Texas Eurycea, (—) indicates no data.

Taxona Localityb State: County n Haplotypesc

Eurycea species 3K Ranch Texas: Kendall 2 H39 (2)

Eurycea species Badweather Pit Texas: Comal 2 H20 (2)
Eurycea sosorum Barton Springs^ Texas: Travis 4 H42, H43, H44, H51
Eurycea species Bear Creek Spring Texas: Kendall 1 H20
Eurycea species Blowing Sink Cave Texas: Travis 2 H51 (2)
Eurycea species Boardhouse Springs Texas: Blanco 2 H08 (2)
Eurycea species Bucket of Toads Cave Texas: Bexar 4 H20 (2), H24 (2)
Eurycea species Bullis Bat Cave Texas: Comal 2 H14, H18
Eurycea species Bullis Spring 4C 018B Texas: Bexar 1 H26
Eurycea species Bullis Spring 8A Texas: Bexar 1 H23
Eurycea species Bullis Spring 9 083 Texas: Bexar 1 H40
Eurycea species Bullis Spring 9 29 Texas: Bexar 1 H40
Eurycea species Camp Bullis Cave #1 Texas: Comal 1 H20
Eurycea species Camp Bullis Cave #3 Texas: Comal 2 H14, H20
Eurycea latitans Cascade Caverns^ Texas: Kendall 3 H20 (3)
Eurycea species Cave Without a Name Texas: Kendall 2 H13 (2)
Eurycea species Cherry Creek Spring Texas: Kerr 1 H20
Eurycea species Cibolo Creek tributary Spring Texas: Kendall 3 H36, H37, H38
Eurycea species Cloud Hollow Spring Texas: Kerr 1 H20
Eurycea species CM Cave Texas: Comal 2 H03 (2)
Eurycea species Cold Spring Austin Texas: Travis 2 H50 (2)
Eurycea species Cold Spring Blanco Texas: Blanco 2 H08 (2)
Eurycea ‘Comal’ species Comal Springs+ Texas: Comal 2 H30, H35
Eurycea ‘Comal’ species Comal Springs+, Spring Run 1 Texas: Comal 3 H31, H32, H35
Eurycea ‘Comal’ species Comal Springs+, Spring Run 3 Texas: Comal 3 H31, H32, H35
Eurycea species Culebra Springs Texas: Bexar 1 H29
Eurycea species Devils Backbone Spring Texas: Comal 2 H20 (2)
Eurycea species Ebert Cave Texas: Comal 3 H11 (2), H20
Eurycea pterophila Fern Bank Springs^ Texas: Hays 14 H05, H06 (12), H15
Eurycea species Golden Fawn Water Cave Texas: Kendall 1 H19
Eurycea species Grapevine Cave Texas: Hays 1 H03
Eurycea ‘Pedernales’ species Hammett’s Crossing Spring+ Texas: Travis 2 H52 (2)
Eurycea species Hector Hole Texas: Bexar 1 H25
Eurycea neotenes Helotes Creek Spring^ Texas: Bexar 2 H21, H22
Eurycea species Hoffman Ranch Estavelle Texas: Kendall 1 H03
E. latitans Honey Creek Cave Spring Texas: Comal 1 H20
Eurycea tridentifera Honey Creek Cave^ Texas: Comal 5 H11 (2), H17, H20 (2)
E. latitans Honey Creek SNA Spring Texas: Comal 1 H20
Eurycea species Horsejump Spring Texas: Hays 2 H03
Eurycea species Hueco Springs Texas: Comal 7 H33, H34, H35 (5)
E. pterophila Jacob’s Well, surface Texas: Hays 22 H01 (2), H03 (20)
E. pterophila Jacob’s Well, 7 m Texas: Hays 10 H01 (4), H03 (6)
E. pterophila Jacob’s Well, 7 m–25m Texas: Hays 12 H01 (5), H03 (7)
E. pterophila Jacob’s Well, 25 m Texas: Hays 7 H03 (7)
E. pterophila Jacob’s Well, subsurfaced Texas: Hays 16 H01 (5), H02 (2), H03 (9)
E. pterophila Jacob’s Well, unknown location Texas: Hays 2 H03, H06
Eurycea species John Knox Ranch Spring Texas: Comal 1 H06
Eurycea species Knee Deep Cave Texas: Kendall 1 H12
Eurycea species Koch Spring Texas: Blanco 2 H08, H09
Eurycea species Leah’s Spring Texas: Bexar 1 H24
Eurycea species Leon Springs Texas: Bexar 1 H24
Eurycea species Less Ranch Spring Texas: Kendall 2 H12 (2)
Eurycea species Lewis Valley Spring Texas: Bexar 1 H23
Eurycea species Lost Dog Spring Texas: Bexar 1 H28
Eurycea species Magic Spring Texas: Comal 2 H03 (2)
4 The Southwestern Naturalist vol. 62, no. 1

TABLE 1—Continued.

Taxona Localityb State: County n Haplotypesc

Eurycea species Morales Spring Texas: Bexar 2 H24 (2)

Eurycea species Mullers Spring Texas: Kendall 1 H28
Eurycea species near Fern Bank Springs Texas: Hays 3 H06 (3)
Eurycea species Ott’s Spring Texas: Comal 6 H20 (6)
Eurycea species Peaveys Spring Texas: Kendall 1 H08
Eurycea species Pecan Springs Texas: Medina 1 H27
E. latitans Pfeiffer’s Water Cave Texas: Kendall 2 H12, H16
Eurycea species Preserve Cave Texas: Comal 3 H03 (3)
Eurycea species Rebecca Creek Texas: Comal 1 H04
Eurycea nana San Marcos Springs^ Texas: Hays 3 H45, H47 (2)
Eurycea species Sattler’s Deep Pit Texas: Kendall 1 H03
Eurycea species Sharon Spring Texas: Bexar 1 H24
E. nana Spring Lake^: below dam Texas: Hays 2 H45, H46
E. nana Spring Lake^: Diversion Springs Texas: Hays 2 H46, H49
E. nana Spring Lake^: Hotel site Texas: Hays 2 H45, H48
Eurycea species Stealth Cave Texas: Bexar 6 H20 (3), H24 (3)
Eurycea species Sycamore Spring Texas: Hays 2 H06 (2)
Eurycea species T Cave Texas: Blanco 3 H07, H08, H10
Eurycea species Taylor Springs Texas: Hays 1 H51
Eurycea species Upper Taylor Spring Texas: Hays 2 H51 (2)
Eurycea species White Springs Texas: Blanco 1 H41
Eurycea species Zercher Spring Texas: Blanco 1 H08
Eurycea species Carson Cave Texas: Uvalde 1 H66
Eurycea species Fessenden Spring Texas: Kerr 1 H64
Eurycea species Finnegan Spring Texas: Val Verde 1 H65
Eurycea species Less Ranch, Frederick Creek Texas: Kendall 1 H54
Eurycea species Lost Maples Texas: Bandera 6 H53 (5), H57
Eurycea species Love Creek Texas: Bandera 1 H56
Eurycea species RR335e Texas: Edwards 3 H58 (2), H59
Eurycea species Stockman Spring Texas: Kerr 5 H63 (5)
Eurycea troglodytes Sutherland Hollow Spring Texas: Bandera 1 H55
E. troglodytes Trough Spring Texas: Gillespie 1 H62
Eurycea species Zenner-Ahrens Roadf Texas: Gillespie 2 H60, H61
Eurycea waterlooensis Barton Springs^ Texas: Travis 2 H68 (2)
Eurycea rathbuni Rattlesnake Cave Texas: Hays 3 H67 (3)
Eurycea species Fessender Spring Texas: Kerr 1 H69
Eurycea tonkawae Balcones Canyonlands, above tributary 7 Texas: Travis 1 H73
E. tonkawae Balcones Canyonlands, tributary 5 Texas: Travis 2 H71, H73
Eurycea naufragia Buford Hollow Spring Texas: Williamson 1 H77
E. naufragia Cedar Breaks Hiking Trail Spring^ Texas: Williamson 1 H76
Eurycea chisholmensis Salado Springs^ Texas: Bell 1 H78
E. tonkawae SAS canyon Texas: Travis 1 H70
E. tonkawae Stillhouse Hollow Spring^ Texas: Travis 3 H72, H74, H75
Eurycea quadridigitata — Texas: Smith 1 —
E. quadridigitata — Texas: Nacogdoches 1 —
Eurycea chamberlaini — North Carolina: Duplin 1 —
E. chamberlaini — South Carolina: Colleton 2 —
E. quadridigitata — North Carolina: Robeson 1 —
E. quadridigitata — Georgia: Chatham 1 —
E. quadridigitata — Alabama: Macon 1 —
E. quadridigitata — Georgia: Marion 1 —
E. quadridigitata — Mississippi: Stone 1 —
E. quadridigitata — Alabama: Baldwin 1 —
Eurycea longicauda — Pennsylvania: Washington 1 —
E. longicauda — 1 —
Eurycea bislineata — New York: Chenango 1 —
E. bislineata — 1 —
March 2017 Krejca et al.—Eurycea pterophila salamanders from Jacob’s Well, Hays County, Texas 5

TABLE 1—Continued.

Taxona Localityb State: County n Haplotypesc

Eurycea multiplicata — Arkansas: Polk 2 —

E. multiplicata — Arkansas: Logan 1 —
Urspelerpes brucei — Georgia: Stephens 1 —
a
Species names as designated in previous studies.
b
Locations marked with ‘‘^’’ are type localities for described species (San Marcos Springs and Spring Lake are the same site). Locations marked
with ‘‘+’’ have Eurycea populations suggested to be separate species but which have not been formally described.
c
Number of sequences per haplotype are provided in parentheses, if greater than one.
d
Capture location is below surface but exact depth is unknown.
e
Spring location is 250 m west of RR335, approximately 8.5 road km south of the RR335 and Texas Highway 41 intersection.
f
Spring location is near Zenner-Ahrens Road, approximately 2.2 road km south of the White Oak Road and Zenner-Ahrens Road intersection.

determined for several of the sampling locations. We also sequence per haplotype into GenBank for sequences generated
included outgroups from each major Eurycea complex (i.e., E. for this study: HQ713576–HQ713601.
bislineata, E. longicauda, E. multiplicata, E. quadridigitata; Harlan We aligned the 280 sequences using MUSCLE (Edgar, 2004)
and Zigler, 2009; Lamb and Beamer, 2012) and each major clade implemented in Geneious 6.0.6 (Biomatters Ltd., Auckland,
within Eurycea quadridigitata (i.e., western, central, E. chamberlai- New Zealand). We evaluated nucleotide substitution models
ni, E. quadridigitata, and Florida panhandle; Lamb and Beamer, with Akaike Information Criterion (AIC) using Modeltest
2012). We used Urspelerpes brucei as a non-Eurycea outgroup (Posada and Crandall, 1998) in PAUP* (Swofford, 2002)
(Lamb and Beamer, 2012; Steffen et al., 2014; Table 1). Our implemented in Geneious 6.0.6. We conducted maximum
analyses include sequences from Hillis et al. (2001): AY014841– likelihood (RAxML; Stamatakis, 2006) and Bayesian (MrBayes;
AY014857 (n = 16); Weins et al. (2003): AY260759 (n = 1); Ronquist and Huelsenbeck, 2003) phylogenetic analyses. We
Bonett and Chippindale (2004): AY528335–528403 (n = 6); implemented both analyses through Geneious, 9.1.4 and 6.0.6,
Lamb and Beamer (2012): JQ920618–JQ920719 (n = 20); respectively. The Modeltest results determined input parame-
Bendik et al. (2013): KC355860–KC355971 (n = 110); and this ters, and we used the general time reversible (GTR) substitution
study: MF10104–MF29815 (n = 127). We accessioned one model with gamma rate variation (Tavaré, 1986) for both

FIG. 1—The number of Eurycea collected from Jacob’s Well during fauna surveys compared to well discharge in cubic feet per
second (cfs). Two salamanders were collected in 2009 (over the course of eight surveys) and 44 were collected in 2010 (over the
course of nine surveys). No sampling took place between heavy rains in January 2010 and May 2010.
6 The Southwestern Naturalist vol. 62, no. 1

FIG. 2—Bayesian consensus phylogram of 279 Eurycea mitochondrial cytochrome b sequences rooted with Ursperlerpes brucei. The
Eurycea neotenes complex is collapsed and provided in Figure 3. Black vertical bars indicate clades of interest. Stars indicate posterior
probabilities ‡95. Bayesian posterior probabilities are provided above branches and maximum likelihood (RAxML) values are
provided below branches. Samples marked with ‘‘^’’ are type localities for described species (San Marcos Springs and Spring Lake are
the same site). Samples marked with ‘‘+’’ are suggested to be separate species but have not been formally described.

analyses. We estimated bootstrap values (Felsenstein, 1985) from (Excoffier et al., 2005) with 10,000 permutations and a
a heuristic search with random stepwise sequence addition for significance value of 0.05.
100 maximum likelihood iterations. The Bayesian analysis
parameters included four heated chains, five million Markov RESULTS—We collected all of our specimens between
chain Monte Carlo (MCMC) generations, saving every hun- the driftnet (25 m deep, 42 m penetration) and the
dredth tree, and with a burn-in of 500,000 generations.
bottom of the entrance shaft (7 m deep, 7 m penetra-
We further evaluated the population relationships among
tion). Divers, including one of the authors (J. K. Krejca),
central Texas Eurycea included in the phylogenetic analyses. We
trimmed all sequences to an equal size, resulting in a data set of saw an individual at 43 m depth and about 85 m
261 sequences of 911 base pairs with no missing data or gaps. We penetration, and other divers have recorded sightings
constructed a statistical parsimony network (Templeton et al., hundreds of meters into the cave. We captured most
1992) of mtDNA haplotypes for all central Texas Eurycea using specimens (including very small juveniles) in the driftnet
the TCS program (Clement et al., 2000) implemented in following high flow events, suggesting that reproduction
PopART (http://popart.otago.ac.nz). We further analyzed occurs in the cave upstream (i.e., deeper and farther
samples that resolved within the eastern Blepsimolge clade from penetration) from the driftnet. Divers captured other
the phylogenetic analyses for differences in allele frequencies.
individuals by hand, but we did not capture any
We included samples from eastern Blepsimolge type localities and
samples from localities with ‡3 sequences in a pairwise FST individuals in the bottle traps. We encountered salaman-
analysis (the proportion of total genetic variance contained in a ders on a variety of substrates including silt floors (deeper
subpopulation relative to the total genetic variance; Nei and Li, and farther in the cave) and underneath rocks and
1979). We conducted the pairwise FST analysis in Arlequin boulders (in the area of the entrance shaft). Over the
March 2017 Krejca et al.—Eurycea pterophila salamanders from Jacob’s Well, Hays County, Texas 7

FIG. 3—Bayesian consensus phylogram of 279 Eurycea mitochondrial cytochrome b sequences rooted with Ursperlerpes brucei. The
Eurycea neotenes complex and eastern Blepsimolge clade 1 are shown and all other nodes are removed (see Fig. 2). Stars indicate
posterior probabilities ‡95. Bayesian posterior probabilities are provided above branches and maximum likelihood (RAxML) values
are provided below branches. Samples marked with ‘‘^’’ are type localities for described species (San Marcos Springs and Spring Lake
are the same site). Samples marked with ‘‘+’’ are suggested to be separate species but have not been formally described.
8 The Southwestern Naturalist vol. 62, no. 1

course of this study (including both high flow and low individuals recovered in the eastern Blepsimolge clade
flow events), scuba divers trained in biological sampling were significant (Table 2), despite low sample size for
captured 0.38 salamanders per person-hour of searching, many sites. We detected significant genetic differentiation
and the driftnet captured 0.05 salamanders per million from the pairwise permutation test for 95 combinations of
gallons of water. sites (Table 2). Jacob’s Well, San Marcos Springs, Hueco
While we performed no formal morphometric analyses Springs, and Comal Springs were significantly different
on the samples, the large number of juveniles and a from all other sites included in the analysis.
handful of adult specimens seen did not exhibit a
troglomorphic appearance. Head shape and relative size, DISCUSSION—At 43 m below the water surface, this site
limb length and diameter, and coloration were typical for has the deepest recorded Eurycea observation in the state,
E. pterophila as assessed by casual examination of the and may be the deepest record for the genus thus far
specimens. reported. Eurycea sosorum is known from 38 m below the
We collected two salamanders during eight sampling water surface in a well (McDermid et al., 2015), and other
events prior to February 2010. Heavy rains during the localities for this genus in central Texas are from springs
spring of 2010 increased the discharge of the well to levels or cave streams that may be many meters below the
that were unsafe for scuba surveys (>20 cfs) until May ground surface but are from shallower water depths.
2010. We performed nine sampling events during the 4- Because very small juveniles were found in the driftnet
mo period following that high flow event, during which only after high flow events, it appears that reproduction is
time we captured 44 juvenile salamanders in the driftnet occurring upstream of the driftnet, meaning deeper than
(Fig. 1). 25 m and farther than 42 m into the cave. Given that this
The samples from Jacob’s Well were combined with spring-associated species uses the aquifer, it is likely that
Eurycea collected from seven counties, and from GenBank other spring-associated species not typically thought of as
accessions, for a total of 261 central Texas Eurycea aquifer-dwelling (e.g., E. nana, E. neotenes) also use this
individuals and 280 total salamanders analyzed for this subsurface habitat. A management consideration, given
study (Table 1). The model of evolution that best fit the these new data, is that salamander localities that are
data was GTR+I+G (chosen by AIC) as determined by springs would benefit from the inclusion of buffers for
Modeltest. Maximum likelihood and Bayesian phyloge- aquifer habitat that may serve as breeding grounds or vital
netic analyses resulted in similar topologies, and we refuge during drought. Currently, critical habitat is
present maximum likelihood results on the Bayesian mapped for spring-associated E. nana and includes only
phylogram (Figs. 2 and 3). We display the single Spring Lake and associated outflow and a buffer of 50 m
phylogenetic topology in two figures due to its large size downstream from Spring Lake Dam (U.S. Fish and
(Figs. 2 and 3). Figure 2 shows the relationships among Wildlife Service [USFWS], 1980), with no designation
the outgroups and major central Texas Eurycea clades given to potential subsurface habitat. More recently,
(i.e., Septentriomolge, Typhlomolge, western Blepsimolge, and USFWS (2013) designated not only surface critical habitat
eastern Blepsimolge; Hillis et al., 2001). One sample, at 33 known cave and surface E. waterlooensis and Eurycea
Fessender Spring (JQ920629), failed to resolve within a tonkawae locations but also designated subsurface critical
central Texas clade (Fig. 2). All Jacob’s Well samples, habitat that includes a 300-m circle around each
surface and subsurface, formed a polytomy within the documented cave and surface location.
Eurycea neotenes complex (Fig. 3). This study began during a severe drought, and through
The trimmed 911-base pair cytochrome b alignment of the end of 2009 we only saw (and collected) two
the 261 central Texas individuals from approximately 14 individuals. Several heavy rain events occurred during
species (12 named and 2 suggested) resulted in 78 unique the first several months of 2010, and we found the
haplotypes (Table 1). We present the statistical parsimony majority of individuals (96%) after those rain events. Our
network of these 78 unique mtDNA haplotypes in Figure results do not let us differentiate between an increase in
4, including the 52 unique haplotypes from 222 individ- detectability and an increase in actual numbers; however,
uals that resolved as the eastern Blepsimolge clade in the it is our speculation that reproduction, and therefore
phylogenetic analysis. Four haplotypes (i.e., H01, H02, salamander observations, increases soon after high
H03, and H06) were detected at Jacob’s Well and two aquifer flows, similar to patterns described for other
(i.e., H01 and H02) were unique to that site (Table 1). We Eurycea species studied by Gillespie (2011) and Dries et al.
observed haplotypes H01 and H03 across the surface and (2013). Rains likely push nutrients through the otherwise
subsurface depth profile (Table 1). We observed haplo- low-energy aquifer, and these sporadically available
types H03 and H06 at seven additional and four nutrients could feed salamanders directly or indirectly
additional sites, respectively (Table 1). We found sixty- via crustacean populations that responded to the nutri-
five haplotypes at only a single site and 13 haplotypes at ents. This scenario of nutrient flow is consistent with work
multiple sites (Table 1). by Simon and Benfield (2002), who demonstrated that
Most pairwise FST values calculated for groups of cave streams are carbon limited, and when carbon is
March 2017 Krejca et al.—Eurycea pterophila salamanders from Jacob’s Well, Hays County, Texas 9

FIG. 4—Statistical parismony network of 78 unique mitochondrial haplotypes in 261 sampled central Texas Eurycea. Circle size is
proportional to the number of individuals. Haplotype number is shown. Each hatch mark represents a single mutation and small,
unlabeled circles represent inferred median haplotypes. Inset shows a cladogram congruent with the central Texas Eurycea clades
recovered in the phylogenetic analysis (Fig. 2) and their corresponding haplotype numbers. Haplotypes 1, 2, 3, and 6 include
specimens from Jacob’s Well.

present it is quickly utilized. It also fits the model of (e.g., species-level) differences among specimens from
organic carbon distribution in caves presented by Simon Jacob’s Well, regardless of depth. We found two haplo-
et al. (2007). Researchers have documented other types (i.e., H01 and H03) in surface and subsurface
vertebrate cave species that cue reproduction from samples, and all four haplotypes at the site are similar.
flooding (e.g., Amblyopsis spelsea [Poulson, 1963]). The Both H01 and H02 are one mutation from H03, and H06
area on the surface around the spring opening increases is two mutations from H03 (Fig. 4). Jacob’s Well
from higher flows and is inundated or provides flowing individuals share haplotypes with Eurycea at 11 other
water, thus allowing individuals to feed in the higher- springs in the eastern Blepsimolge clade. However, Jacob’s
energy spring system. Additionally, the salamanders that
Well was significantly different from all other sites
divers captured were mostly juveniles, suggesting either
included in the pairwise analysis. Considering that these
that the rains might have triggered a breeding event or
populations were not distinct in the phylogenetic tree,
that the higher discharge rates pushed juveniles from the
and that three of the included sites shared a haplotype
deeper parts of the cave disproportionately relative to
adults. In terms of management considerations, this with Jacob’s Well, it was surprising that the FST results
dataset shows a positive correlation between salamander (Table 2) indicated they were also seemingly not
detections and spring flow, lending support for minimum exchanging successful migrants (FST values between
spring flow protections, as regulators are considering for 0.741–0.984). This incongruence is likely due to low
other Eurycea sites in central Texas (RECON Environ- sample size, and if more animals are analyzed from
mental Inc. et al., 2012, http://eahcp.org/documents/ surrounding springs, we may find evidence of migration.
Final%20HCP%20November%202012.pdf). We are cautious in concluding isolation within this
Our genetic analyses did not reveal any substantial cluster, given shared haplotypes across discrete sites and
 TABLE 2—Pairwise FST values (below diagonal) and P values (above diagonal) for all central Texas Eurycea type localities and all other springs with three or more samples.
Significant FST values are denoted with an asterisk. In several cases FST values are high (no gene flow) while P values show no population differentiation, and this discrepancy 10
can be explained by small sample sizes. (—) indicates not applicable because the column and row are the same site.

Cibolo near
San Creek Helotes Bucket Fern Fern Honey
Hammett’s Marcos Barton tributary Hueco Creek Comal of Toads Stealth Bank Bank Jacob’s Ott’s Preserve Ebert Creek Cascade
Crossing+ Springs^ Springs^ Springs Springs Spring^ Springs+ Cave Cave T Cave Springs^ Springs Well Spring Cave Cave Cave^ Caverns^
(n = 2) (n = 29) (n = 24) (n n = 23) (n = 27) (n=2) (n = 8) (n = 4) (n = 6) (n = 3) (n = 14) (n = 3) (n = 69) (n = 6) (n = 3) (n = 3) (n = 5) (n = 3)

Hammett’s — 0.017 0.065 0.099 0.026 0.342 0.021 0.063 0.034 0.096 0.009 0.098 0.001 0.036 0.098 0.104 0.046 0.102
Crossing+
San Marcos 0.953* — 0.001 0.004 <0.001 0.017 <0.001 0.002 <0.001 0.004 <0.001 0.004 <0.001 <0.001 0.005 0.006 <0.001 0.004
Springs^,a
Barton 0.719 0.714* — 0.056 0.003 0.068 0.003 0.026 0.006 0.059 <0.001 0.060 <0.001 0.005 0.060 0.059 0.008 0.057
Springs^
Cibolo Creek 0.971 0.915* 0.485 — 0.008 0.105 0.006 0.030 0.025 0.103 0.002 0.101 <0.001 0.012 0.100 0.101 0.017 0.107
tributary
Springs
Hueco 0.966* 0.925* 0.662* 0.802* — 0.027 0.016 0.006 0.003 0.008 <0.001 0.010 <0.001 0.001 0.010 0.010 0.001 0.009
Springs
Helotes 0.971 0.921* 0.446 0.820 0.626* — 0.024 0.131 0.137 0.108 0.009 0.096 <0.001 0.037 0.100 0.103 0.046 0.101
Creek
Spring^
Comal 0.964* 0.927* 0.689* 0.805* 0.303* 0.650* — 0.006 0.001 0.006 <0.001 0.004 <0.001 0.000 0.008 0.006 0.001 0.005
Springs+,b
Bucket of 0.934 0.896* 0.446* 0.549* 0.331* 0.385 0.415* — 1.000 0.089 0.005 0.090 <0.001 0.128 0.081 0.231 0.088 0.425
Toads Cave
Stealth Cave 0.934* 0.897* 0.521* 0.563* 0.340* 0.425 0.421* -0.263 — 0.023 <0.001 0.023 <0.001 0.187 0.022 0.057 0.079 0.467
T Cave 0.942 0.899* 0.419 0.600 0.727* 0.708 0.740* 0.385 0.424* — 0.002 0.104 <0.001 0.011 0.098 0.101 0.017 0.100
Fern Bank 0.966* 0.932* 0.718* 0.735* 0.780* 0.844* 0.790* 0.540* 0.533* 0.631* — 1.000 <0.001 <0.001 0.001 0.004 0.001 0.012
The Southwestern Naturalist

Springs^
near Fern 1.000 0.922* 0.465 0.818 0.810* 0.903 0.808* 0.489 0.478* 0.600 -0.192 — <0.001 0.014 0.099 0.105 0.016 0.101
Bank
Springs
Jacob’s Wellc 0.984* 0.968* 0.898* 0.863* 0.873* 0.922* 0.880* 0.751* 0.741* 0.815* 0.766* 0.805* — <0.001 0.631 <0.001 <0.001 <0.001
Ott’s Spring 1.000* 0.936* 0.590* 0.848* 0.820* 0.942* 0.817* 0.442 0.400 0.622* 0.528* 1.000* 0.690* — 0.013 0.086 0.062 1.000
Preserve Cave 1.000 0.922* 0.465 0.818 0.810* 0.903 0.808* 0.489 0.478* 0.600 0.665* 1.000 -0.058 1.000* — 0.099 0.015 0.105
Ebert Cave 0.968 0.901* 0.412 0.636 0.732* 0.777 0.742* 0.304 0.330 0.400 0.542* 0.667 0.749* 0.547 0.667 — 0.678 0.402
Honey Creek 0.971* 0.912* 0.520* 0.657* 0.736* 0.815* 0.747* 0.322 0.330 0.441* 0.489* 0.605* 0.699* 0.211 0.605* -0.071 — 0.462
Cave^,d
Cascade 1.000 0.918* 0.425 0.750 0.761* 0.887 0.763* 0.250 0.250 0.429 0.453* 1.000 0.671* 0.000 1.000 0.333 0.032 —
Caverns^
^ Type localities for central Texas Eurycea species.
+
Localities with Eurycea populations suggested to be separate species but that have not been formally described.
a
San Marcos Springs and Spring Lake are the same site. Samples include all localities sampled within Spring Lake (i.e., Hotel site, below dam, Diversion Springs).
b
Comal Springs samples include all sampled spring runs (i.e., 1 and 3).
c
Jacob’s Well samples include all sampled depths (i.e., surface, 7 m, 7–25 m, 25 m, subsurface, and unknown location).
d
Honey Creek Cave samples only include cave samples and exclude Honey Creek Cave Spring and Honey Creek SNA Spring.
vol. 62, no. 1
March 2017 Krejca et al.—Eurycea pterophila salamanders from Jacob’s Well, Hays County, Texas 11

the work of Tatarenkov et al. (2010), wherein differenti- Large, geographic-scale genetic analyses of central
ation can be found despite abundant migration. Texas Eurycea are now possible because a decade of
These results are consistent with other authors. Bendik sampling the geographic breadth of their distribution has
(2006) found evidence of introgression between geo- been achieved. Although the genus has 13 named species
graphically close populations and two Jacob’s Well occurring within an approximately 300-mi area in Texas,
specimens, and Bendik et al. (2013) analyzed both this study is among one of the first that incorporated
genetics and morphometrics from larger sample sizes of dense, single-locality sampling into a larger species
other cave and spring populations in this clade, and context. The lack of dense sampling at any given locality
found no evidence of sympatric cave and spring-adapted is largely due to the difficulty of habitat access (e.g., small
congeners. In the evaluation by Lucas et al. (2009), the springs on private land and underwater caves), fluctuat-
authors noted the presence of unique haplotypes for ing drought conditions which can dry up ephemeral
several sites. While we also found haplotypes that were springs that would otherwise be accessible, and the short
unique to Jacob’s Well and to Ott’s Spring, Lucas et al. duration that most studies are able to continue sampling
(2009) found shared haplotypes at these sites for another versus the duration necessary to collect a large number of
mtDNA gene, ND4. In addition, while both studies samples for analysis.
recovered unique haplotypes at Comal Springs and Jacob’s Well is one of a handful of moderately large to
Hueco Springs, we also found a haplotype shared large springs that still flow in central Texas (Brune, 1975).
between those two sites (Table 1). Overall, the results This site flowed continuously through the drought of
indicate that the population from Jacob’s Well has either record (1956), and was recorded to stop for the first time
current or historical contact with one of the parental in 2000 (D. Baker and N. Hauwert, pers. comm.). The
populations. reduction of spring flow is a state-wide phenomenon and
The compilation of all available sequence data reveals one that is unlikely to reverse trend. In particular, this site
that all Jacob’s Well samples and the sequences available is within Groundwater Management Area 9, a region set
for E. latitans, E. tridentifera, and E. pterophila fall within a with Desired Future Conditions to include 30-ft average
single large clade, a finding similar to results provided by declines in water levels through 2060 (State of Texas,
Bendik (2006). Bendik et al. (2013) demonstrated that 2010). If this condition is realized, Jacob’s Well would
those three species were not supported by the mitochon- stop flowing and the ‘spring’ habitat, the wetted area
drial dataset; however, they did not suggest a taxonomic exposed to sunlight, would no longer be wetted or
revision. It is likewise outside the scope of this manuscript accessible to salamanders. This could potentially impact
to suggest taxonomic revisions to central Texas Eurycea, the salamander abundance on the surface and possibly
but the following discrepancies are important in regard to change the patterns of nutrient flow through the cave.
the taxonomic status of Jacob’s Well Eurycea. The type While E. pterophila from Jacob’s Well may not represent a
localities for E. latitans (Cascade Caverns) and E. distinct evolutionary lineage from others in the area,
tridentifera (Honey Creek Cave) share a haplotype (H20) based on our analyses of mtDNA sequence data, they do
and are one mutation from H06, one of the haplotypes at provide genetic variation unique to the species and, as
Fern Bank Springs, the type locality for E. pterophila (Table such, the site should be preserved as a discrete
1 and Fig. 4). Two locations, Bucket of Toads Cave management unit as suggested in Lucas et al. (2009).
(‘‘Bucket o’ Toads’’ in Bendik et al., 2013) and Stealth This spring is unique in its large size as compared to
Cave, cause paraphyly among the two major branches of adjacent localities, and it is one of the only central Texas
the E. neotenes complex (Fig. 3), as they also did in the springs with an associated cave that is accessible enough
phylogenetic analysis in Bendik et al. (2013). These to study the subterranean aquatic fauna in situ at depths
populations highlight the importance of locating sites (both water depth and penetration into the aquifer)
with observable gene flow among populations and inaccessible at other spring sites. These concomitant
question the taxonomic division of E. latitans-tridentifera- benefits increase the overall value of this site as a stable
pterophila from E. neotenes. Thus, currently we assign locality of Eurycea, with all the attendant benefits to future
Jacob’s Well salamanders to part of a geographically research and conservation goals toward a better under-
broad group of E. neotenes populations, congruent with standing of this unique group of salamanders in central
Bendik et al. (2013). One resolution to the observed Texas.
paraphyly is that the E. neotenes complex may represent a
We thank A. Sansom (River Systems Institute), D. Baker
single species. This solution is congruent with Sweet’s
(Wimberley Valley Watershed Authority), A. Price, the Way
(1978) conclusion that E. pterophila is not a separate
Family Foundation, Coypu, J. Brown, G. Tatum, and other
species from E. neotenes. More work will be required volunteer divers, Zara Environmental LLC, and Texas State
before the taxonomy can accurately portray the underly- University. These entities provided the funding to complete this
ing evolutionary relationships among the salamanders project, unlimited access to the study area, support throughout
from these localities, and future work using many more all phases of the research, and the use of laboratory equipment
individuals per locality may help resolve the taxonomy. and time.
12 The Southwestern Naturalist vol. 62, no. 1

LITERATURE CITED central Texas paedomorphic salamanders. Herpetologica

57:266–280.
BENDIK, N. F. 2006. Population genetics, systematics, biogeogra- LAMB, T., AND D. A. BEAMER. 2012. Digits lost or gained? Evidence
phy, and evolution of the southeastern central Texas Eurycea for pedal evolution in the dwarf salamander complex
clade Blepsimolge (Plethodontidae). M.S. thesis, The Uni- (Eurycea, Plethodontidae). PLoS ONE 7:e37544. doi:10.
versity of Texas at Arlington, Arlington, Texas. 1371/journal.pone.0037544.
BENDIK, N. F., J. M. MEIK, A. G. GLUESENKAMP, C. E. ROELKE, AND P. LUCAS, L., Z. GOMPERT, J. OTT, AND C. NICE. 2009. Geographic and
T. CHIPPINDALE. 2013. Biogeography, phylogeny, and morpho- genetic isolation in spring-associated Eurycea salamanders
logical evolution of central Texas cave and spring salaman- endemic to the Edwards Plateau region of Texas. Conserva-
ders. BMC Evolutionary Biology 13:201. Available at: http:// tion Genetics 10:1309–1319.
www.biomedcentral.com/1471-2148/13/201. MCDERMID, K., P. SPROUSE, AND J. KREJCA. 2015. Geographic
BISHOP, S. C. 1941. Notes on salamanders, with descriptions of distribution. Eurycea sosorum (Barton Springs salamander).
several new forms. Occasional Papers of the Museum of Herpetological Review 46:556–557.
Zoology 451, University of Michigan, Ann Arbor, Michigan. NEI, M., AND W. H. LI. 1979. Mathematical model for studying
BONETT, R. M., AND P. T. C HIPPINDALE . 2004. Speciation, genetic variation in terms of restriction endonucleases.
phylogeography and evolution of life history and morphol- Proceedings of the National Academy of Sciences of the
ogy in plethodontid salamanders of the Eurycea multiplicata U.S.A.76:5269–5273.
complex. Molecular Ecology 13:1189–1203. POSADA, D., AND K. A. CRANDALL. 1998. MODELTEST: testing the
BRUNE, G. 1975. Major and historical springs of Texas. Texas model of DNA substitution. Bioinformatics 14:817–818.
Water Development Board, Technical Report #189, Austin, POULSON, T. L. 1963. Cave adaptation in amblyopsid fishes.
Texas. American Midland Naturalist 70:257–290.
CHIPPINDALE, P. T. 2005. Eurycea rathbuni (Stejneger, 1896) Texas
RONQUIST, F., AND J. P. HUELSENBECK. 2003. MrBayes 3: Bayesian
blind salamander. Pages 760–762 in Amphibian declines: the
phylogenetic inference under mixed models. Bioinformatics
conservation status of United States species (M. Lannoo,
19:1572–1574.
editor). University of California Press, Berkeley.
SIMON, K. S., AND E. F. BENFIELD. 2002. Ammonium retention and
CHIPPINDALE, P. T., AND J. N. FRIES. 2005. Eurycea nana (Bishop,
whole-stream metabolism in cave streams. Hydrobiologia
1941) San Marcos salamander. Pages 755–756 in Amphibian
482:31–39.
declines: the conservation status of United States species (M.
SIMON, K. S., T. PIPAN, AND D. C. CULVER. 2007. A conceptual
Lannoo, editor). University of California Press, Berkeley.
model of the flow and distribution of organic carbon in
CHIPPINDALE, P. T., A. H. PRICE, J. J. WIENS, AND D. M. HILLIS. 2000.
caves. Journal of Cave and Karst Studies 69:279–284.
Phylogenetic relationships and systematic revision of central
STAMATAKIS, A. 2006. RAxML-VI-HPC: maximum likelihood-based
Texas Hemidactyliine plethodontid salamanders. Herpeto-
phylogenetic analyses with thousands of taxa and mixed
logical Monographs 14:1–80.
models. Bioinformatics 22:2688–2690.
CLEMENT, M., D. POSADA, AND K. A. CRANDALL. 2000. TCS: a
computer program to estimate gene genealogies. Molecular STATE OF TEXAS. 2010. Groundwater management area 9
Ecology 9:1657–1659. resolution #072610-01. Designation of desired future condi-
DRIES, L. A., C. HERRINGTON, L. A. COLUCCI, N. F. BENDIK, D. A. tions for groundwater management area 9 aquifers (July 26,
CHAMBERLAIN, D. JOHNS, AND E. PEACOCK. 2013. Major amend- 2010). p. 23–25 of August 26, 2010 Letter to Kevin Ward,
ment and extension of the habitat conservation plan for the Executive Director, Texas Water Development Board from
Barton Springs salamander (Eurycea sosorum) and the Austin Ronald Fieseler, General Manager Blanco Pedernales GCD
blind salamander (Eurycea waterlooensis) to allow for the and Groundwater Management Area 9 Coordinator. Avail-
operation and maintenance of Barton Springs and adjacent able at: http://www.twdb.texas.gov/groundwater/docs/
springs. City of Austin Watershed Protection Department, DFC/GMA9_DFC_Adopted_2010-0726.pdf. Accessed 14 Sep-
Austin, Texas. tember 2016.
EDGAR, R. C. 2004. MUSCLE: multiple sequence alignment with STEFFEN, M. A., K. J. IRWIN, A. L. BLAIR, AND R. M. BONETT. 2014.
accuracy and high throughput. Nucleic Acids Research Larval masquerade: a new species of paedomorphic salaman-
32:1792–1797. der (Caudata: Plethodontidae: Eurycea) from the Ouachita
EXCOFFIER, L., G. LAVAL, AND S. SCHNEIDER. 2005. Arlequin ver. 3.0: Mountains of North America. Zootaxa 3786:423–442.
an integrated software package for population genetics data SWEET, S.S. 1978. On the status of Eurycea pterophila (Amphibia:
analysis. Evolutionary Bioinformatics Online 1:47–50. Plethodontidae). Herpetologica 34:101–108.
FELSENSTEIN, J. 1985. Confidence limits on phylogenies: an SWOFFORD, D. L. 2002. PAUP*: phylogenetic analysis using
approach using the bootstrap. Evolution 39:783–791. parsimony (*and other methods), version 4.0b10. Sinauer
G ILLESPIE , J. H. 2011. Ecology and conservation of the Associates, Sunderland, Massachusetts.
endangered Barton Springs salamander (Eurycea sosorum). TATARENKOV, A, C. I. HEALEY, AND J. C. AVISE. 2010. Microgeo-
Ph.D. dissertation, University of Texas at Austin. graphic population structure of green swordtail fish: genetic
HARLAN, A. J., AND K. S. ZIGLER. 2009. The phylogenetic position differentiation despite abundant migration. Molecular Ecol-
of Eurycea lucifuga, the cave salamander, and the evolution of ogy 19:257–268.
cave-adapted species within Eurycea. Journal of the Tennessee TAVARÉ, S. 1986. Some probabilistic and statistical problems in
Academy of Science 84:77–82. the analysis of DNA sequences. Lectures on mathematics in
HILLIS, D. M., D. A. CHAMBERLAIN, T. P. WILCOX, AND P. T. the life sciences. American Mathematical Society 17:57–86.
CHIPPINDALE. 2001. A new species of subterranean blind TEMPLETON, A. R., K. A. CRANDALL, AND C. F. SING. 1992. A cladistic
salamander from Austin, Texas, and a systematic revision of analysis of phenotypic associations with haplotypes inferred
March 2017 Krejca et al.—Eurycea pterophila salamanders from Jacob’s Well, Hays County, Texas 13

from restriction endonuclease mapping and DNA sequence designation of critical habitat for the Austin blind and
data. III. Cladogram estimation. Genetics 132:619–633. Jollyville Plateau salamander. Federal Register 161:51328–
UNITED STATES FISH AND WILDLIFE SERVICE (USFWS). 1980. 51379.
Endangered and threatened wildlife and plants; listing of WEINS, J. J., P. T. CHIPPINDALE, AND D. M. HILLIS. 2003. When are
the San Marcos salamander as threatened, the San Marcos
phylogenetic analyses misled by convergence? A case study in
gambusia as endangered, and the listing of critical habitat for
Texas cave salamanders. Systematic Biology 52:501–514.
Texas wild rice, San Marcos salamander, San Marcos
gambusia and fountain darter. Federal Register 45:47355–
47364. Submitted 14 January 2016. Accepted 7 December 2016.
USFWS. 2013. Endangered and threatened wildlife and plants; Associate Editor was Neil B. Ford.
THE SOUTHWESTERN NATURALIST 62(1): 14–28 MARCH 2017

LATE PLEISTOCENE AND HOLOCENE

BISON OF THE COLORADO PLATEAU

JEFF M. MARTIN,* RACHEL A. MARTIN, AND JIM I. MEAD

Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843 (JMM)
Ecosystem Science and Management, Texas A&M University, College Station, TX 77843 (RAM)
The Mammoth Site, Hot Springs, SD 57747 (JIM)
Don Sundquist Center of Excellence in Paleontology, East Tennessee State University Gray Fossil Site and Museum,
Gray, TN 37615 (JMM, RAM, JIM)
*Correspondent: jeff.m.martin@exchange.tamu.edu

ABSTRACT—Fossils of Bison (bison) are scarce on the Colorado Plateau, especially in and around the Greater
Grand Canyon Region. Because of poor preservation and collection biases in the region, various resource
managers have erroneously designated bison a nonnative and human-(re)introduced species. This decision
directly impacts an extant herd of approximately 400 bison that periodically meander onto Grand Canyon
National Park lands from neighboring U.S. Forest Service and State of Arizona lands. We re-examined and
verified paleozoological museum specimens of this large mammal confirming a prehistoric and early historic
presence of Bison on the Colorado Plateau. Our findings indicate that Bison bison should be considered a
native species on the Colorado Plateau because they have a nearly continuous record of inhabitance in the
region. This record is supported by 74 Bison-bearing fossil, subfossil, and historical localities since the latest
Pleistocene.

RESUMEN—Los fósiles de Bison (bisontes) son escasos en la meseta del Colorado, especialmente dentro y
alrededor de la región del Gran Cañón. Debido a la pobre preservación de los fósiles y parcialidad en las
colecciones de la región, varios administradores de recursos naturales han denominado equivocadamente
bisontes como especies exóticas e introducidas por humanos. Esta decisión impacta directamente a la
población actual de cerca de 400 bisontes que a veces traspasan a los terrenos del Servicio Forestal de E.E.U.U.
y del Estado de Arizona, hasta los terrenos del Parque Nacional del Gran Cañón. Re-examinamos y verificamos
especı́menes paleozoológicos de museos de este gran mamı́fero, confirmando la presencia prehistórica e
histórica temprana de Bison en la meseta del Colorado. Nuestros resultados indican que Bison bison debe ser
considerado como especie nativa en la meseta del Colorado por su casi continuo registro de habitación dentro
de la región. Este record es respaldado por 74 fósiles y subfósiles relacionados con Bison, y por localidades
históricas desde el Pleistoceno tardı́o.

Perhaps no other extant animal in North America (Paleoindians; Koch and Barnosky, 2006). Second, Bison
possesses the cultural, spiritual, ecological, economic, survived the threat of Americans of European descent,
political, and natural history attributes that are emblem- who purposefully hunted the bison nearly to extinction,
atic of the North American bison (Bison bison; Artiodac- during the late 1800s Common Era (Hornaday, 1889).
tyla, Bovidae, Bovini; Isenberg, 1997; Sanderson et al., The most recent comprehensive analysis of Bison
2008). Certainly, no other animal boasts the story of distribution in continental North America illustrates an
surviving the brink of extinction twice. This story does not apparent geographic ‘gap’ of bison records for much of
stop at the North American plains subspecies (Bison bison the arid Southwest, especially on the Colorado Plateau
bison), but it also applies to the North American woods (CP; Mc Donald, 1981; Fig. 1). In the western United
bison subspecies (Bison bison athabascae) along with the States, there are well-documented occurrences of Qua-
European bison species (Bison bonasus; wisent). First, ternary-aged Bison in physiographic provinces neighbor-
Bison survived the megafaunal extinction at the end of the ing the CP, such as the Great Basin (Jennings, 1978; Scott
Pleistocene approximately 11,700 calendar years Before and Cox, 2008), southern and middle Rocky Mountains
Present (cal yr BP) and, in North America, the genus (McDonald, 1981), Wyoming Basin (McDonald, 1981),
outlived mammoths (Mammuthus), mastodons (Mammut), Arizona deserts (Agenbroad and Haynes, 1975; Wolff,
horses (Equus), ground sloths (Megalonyx et al.), and 2013), Bonneville Basin (Madsen et al., 2001), and Rio
other megafauna while coexisting with early Americans Grande Rift basin (Harris, 2014).
Reproduced with permission of copyright owner. Further reproduction
prohibited without permission.