Southeastern Naturalist 213 T.J. Hibbitts, S.A. Wahlberg, and G. Voelker 22001155 SOUTHEASTERN NATURALIST 1V4o(2l.) :1241,3 N–2o2. 02 Resolving the Identity of Texas Desmognathus Toby J. Hibbitts1,*, Scott A. Wahlberg2, and Gary Voelker1 Abstract - Dusky salamanders in Texas have long been identified as Desmognathus auriculatus (Southern Dusky Salamander). However, recent taxonomic work has challenged this notion, restricting D. auriculatus to areas east of the Apalachicola River in Florida and Georgia, while all Texas populations are now considered D. conanti. Regrettably, Texas specimens were not included in the studies upon which these proposed taxonomic changes were based. Dusky salamanders have become increasingly difficult to find in Texas and have only been observed in the state 3 times since 1980. We surveyed 18 sites of which 6 were historical localities. We found 4 extant populations of dusky salamanders at new locations in Texas, one of which was a county record. We were unable to find dusky salamanders at some historically robust populations. We also used mtDNA sequence data to verify that Texas dusky salamanders are indeed D. conanti. Introduction The taxonomic status of dusky salamanders (Desmognathus sp.) in East Texas has recently come into question (Beamer and Lamb 2008). Most authors have long considered Texas dusky salamanders to be Desmognathus auriculatus (Holbrook) (Southern Dusky Salamander) based on morphology and habitat (Cook and Brown 1974, Dixon 2013, Petranka 1998), although some early authors considered some Texas individuals to be D. brimleyorum (Stejneger) (Ouachita Dusky Salamander; Livezey 1950, Sanders and Smith 1949). Doubt about the validity of East Texas D. auriculatus arose when it was discovered that D. auriculatus from Mississippi were nested within D. conanti Rossman (Spotted Dusky Salamander) based on mtDNA sequence data (Kozak et al. 2005). A subsequent study also using mtDNA to better understand the relationships of dusky salamanders from the Southeastern Coastal Plain restricted the distribution of the Southern Dusky Salamander to Florida and Georgia east of the Apalachicola River (Beamer and Lamb 2008). Beamer and Lamb (2008) were unable to obtain samples from Texas, but samples from similar habitats to those found in East Texas in central and eastern Louisiana all were genetically identified as D. conanti. Therefore, they assigned all Texas specimens as D. conanti. Dusky salamanders are historically known from 18 counties in eastern Texas (Dixon 2013). However, recent collection records are sparse despite considerable sampling effort at historical localities (T.J. Hibbits, pers observ.). The apparent decline in Desmognathus populations in East Texas is alarming in light of similar 1Biodiversity Research and Teaching Collection, Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843-2258. 2USDA Forest Service, Southern Research Station, 506 Hayter Street, Nacogdoches, TX 75965. *Corresponding author - thibbitts@tamu.edu. Manuscript Editor: John Placyk Southeastern Naturalist T.J. Hibbitts, S.A. Wahlberg, and G. Voelker 2015 Vol. 14, No. 2 214 declines documented in Florida (Means and Travis 2007) and the decline of amphibians globally (Houlahan et al. 2000). The Texas habitats where dusky salamanders have been most recently found were small spring-fed creeks with sandy substrates as well as sloughs and cypress/gum swamps (T.J. Hibbits, pers observ.). Our study had 2 main objectives. First, we wanted to locate populations of Desmognathus in Texas. We chose to search in counties that historically had the largest number of specimens and collection sites to increase our chances of finding extant populations. Second, we wanted to determine the relationships of Texas dusky salamanders to other dusky salamanders of the Coastal Plains of the southeastern United States. We chose to use mtDNA to investigate these relationships Methods We searched for dusky salamanders at historic localities and other habitats that were suitable based on habitat descriptions for D. conanti and D. auriculatus (Petranka 1998). Historic localities were located by querying the VerpNet database (http://www.vertnet.org) and its 192 museum collections as well as the Biodiversity Research and Teaching Collection (TCWC), the Texas Natural History Collection (TNHC), and the Amphibian and Reptile Diversity Research Center (UTAA). Surveys for salamanders consisted of walking along spring runs, streams, and sloughs overturning logs and other debris. We also dip-netted the leaf litter and vegetation in these aquatic habitats and sifted through surface leaf litter, especially around springs and seeps. Each survey site was searched for at least 2 person hours with at least 2 observers. Dusky salamanders in Texas have been found year round based on museum records; therefore we did not focus surveys on any specific time of year, although most surveys were done in the spring months. When encountered, at least 2 individual Dusky Salamanders were collected per locality. We used the E.Z.N.A. tissue DNA kit (Omega) to extract whole genomic DNA from liver samples and the polymerase chain reaction (PCR) to amplify the mitochondrial cytochrome oxidase subunit c (cox1; 550 bp segment); primers and reaction conditions followed those in Beamer and Lamb (2008). We selected this gene because it allowed us to enhance our overall dataset by including extensive Desmognathus sequences available from GenBank (i.e., from Beamer and Lamb 2008). We first did an analysis with the complete dataset from Beamer and Lamb (2008), which included D. auriculatus, D. conanti, and our new material. We then pruned the clade using a subset of Desmognathus species that were most closely related to D. conanti based on the results of our initial analysis of the complete dataset, including: D. apalachicolae Means and Karlin (Apalachicola Dusky Salamander), D. brimleyorum, D. carolinensis Dunn (Carolina Mountain Dusky Salamander), D. fuscus, D. monticola Dunn (Seal Salamander), D. ocoee Nicholls (Ocoee Salamander), and D. santeetlah Tilley (Santeetlah Dusky Salamander) (Table 1). Automated sequencing was performed using BigDye chemistry (Applied Biosystems) and both forward and reverse sequencing products were run on an ABI 377 or 3730 sequencer at the University of Florida sequencing facility (Interdisciplinary Southeastern Naturalist 215 T.J. Hibbitts, S.A. Wahlberg, and G. Voelker 2015 Vol. 14, No. 2 Table 1. Dusky salamander (Desmognathus ssp.) samples and their collection localities. These specimens were used in our phylogenetic analysis to determine the taxonomic placement of the Texas dusky salamander samples. Specimen number Species State County EU311666 D. apalachicolae Florida Liberty EU311708 D. apalachicolae Florida Leon BLowe57 D. apalachicolae Florida Liberty FC11578 D. brimleyorum Oklahoma LeFlore RMB2327 D. brimleyorum Arkansas Nevada RMB2201 D. brimleyorum Arkansas Polk EU311713 D. carolinensis Tennessee Yancey EU311651 D. conanti Georgia Effingham EU311668 D. conanti South Carolina Barnwell EU311709 D. conanti Georgia Wayne EU311710 D. conanti Georgia Wayne TJR2470 D. conanti Georgia Richmond EU311698 D. conanti North Carolina Henderson EU311667 D. conanti Kentucky Livingston EU311712 D. conanti Alabama Lawrence EU311672 D. conanti Mississippi Jasper EU311685 D. conanti Mississippi Jasper ASU23806 D. conanti Mississippi Jasper EU311673 D. conanti Louisiana Washington EU311677 D. conanti Alabama Butler EU311679 D. conanti Florida Santa Rosa EU311678 D. conanti Alabama Baldwin EU311684 D. conanti Florida Washington EU311671 D. conanti Louisiana West Feliciana EU311674 D. conanti Mississippe Amite TJH2696 D. conanti Texas Tyler 94726 D. conanti Texas Tyler TJH2756 D. conanti Texas Newton TJH3263 D. conanti Texas Newton TJH3264 D. conanti Texas Newton TJH3265 D. conanti Texas Newton TJH3266 D. conanti Texas Sabine TJH3270 D. conanti Texas Sabine TJH3262 D. conanti Texas Newton TJH2757 D. conanti Texas Newton TJH2758 D. conanti Texas Newton TJH3269 D. conanti Texas Sabine CSA330 D. conanti Texas Newton EU311699 D. conanti Louisiana Grant EU311702 D. fuscus North Carolina Caldwell EU311662 D. fuscus North Carolina Wilkes EU311688 D. monticola North Carolina Watauga EU311690 D. monticola North Carolina Watauga EU311647 D. ocoee Georgia Lumpkin EU311652 D. ocoee Georgia Douglas EU311715 D. ocoee Georgia Union EU311676 D. santeetlah North Carolina Graham Southeastern Naturalist T.J. Hibbitts, S.A. Wahlberg, and G. Voelker 2015 Vol. 14, No. 2 216 Center for Biotechnology Research, Gainesville, FL). Sequences were aligned with Sequencher® version 5.0.1 sequence analysis software (Gene Codes Corp., Ann Arbor, MI), and we used MrBayes 3.1.2 (Huelsenbeck and Ronquist 2001) to generate a Bayesian tree. In MrBayes, we incorporated codon-position–specific models, as determined in MrModeltest 2.3 (Nylander 2004), and initiated 2 runs of 4 Markovchain Monte Carlo (MCMC) chains for 5,000,000 generations, each starting from a random tree and sampling every 100 generations. Default temperature-profile settings were used. Each run resulted in 50,000 trees and converged on the same topology. The first 5000 trees from each analysis were removed as “burn-in”; visualization of –ln scores versus generations confirmed that chain stationarity had been reached within 5000 trees. We used the remaining 90,000 trees to generate a 50% majority-rule consensus tree. Results Our query of HerpNet and the major Texas herpetological collections revealed 68 localities where 361 specimen records of dusky salamanders have been found in eastern Texas since 1940. The most recent collection date was 1995, but only 3 specimens from 2 localities were found after 1975. We searched 18 localities in 9 counties of East Texas. Six of these localities historically supported populations of dusky salamanders (111 of the 361 specimen records for the state). We found 4 localities that contained populations of dusky salamanders (TCWC 94274, 94726–27, 95134–36, 97079–85, 98122; Fig. 1). All of these localities were in clear, springfed, first-order streams, and none of these localities were historical collection sites. Our initial Bayesian analysis of Desmognathus (not shown) included all of the samples used by Beamer and Lamb (2008) in addition to our new samples from Texas. This analysis resulted in a very similar phylogeny to that of Beamer and Lamb (2008), and showed that our new Texas specimens were sister to D. conanti, and not closely related to D. auriculatus. Therefore we pruned the dataset to include just those species that were most closely related to D. conanti, and this analysis produced a generally well-supported tree (Fig. 2). Most importantly, the Texas Desmognathus populations formed a monophyletic, albeit moderately supported (0.86 posterior probability) group. This clade falls inside the larger D. conanti clade. Within the D. conanti clade, 3 general clades can be identified (Fig. 2). First is a well-supported clade distributed along the Atlantic Coastal Plain of Georgia and the Carolinas. The second is another well-supported clade from Kentucky, Alabama, eastern Mississippi and the panhandle of Florida. The third clade included our samples from Texas and samples collected just east of the Mississippi River in Louisiana and Mississippi. However, if the moderate to weakly supported clades are collapsed, then the third clade is better recognized as a polytomy comprising 3 strongly supported clades: 1) eastern Louisiana and Mississippi, 2) Newton County, TX, and 3) Tyler County, TX. Geographically, these 3 areas are separated by potentially strong river barriers: the Mississippi River, the Sabine River, and the Neches River. Southeastern Naturalist 217 T.J. Hibbitts, S.A. Wahlberg, and G. Voelker 2015 Vol. 14, No. 2 Discussion It appears that populations of Desmognathus in Texas may have suffered from local extirpations. Recent searches for the species in the state previous to our efforts Figure 1. Survey sites for dusky salamanders (Desmognathus spp.) in East Texas. Open circles are sites where dusky salamanders were not observed and filled circles are sites where they were collected. Southeastern Naturalist T.J. Hibbitts, S.A. Wahlberg, and G. Voelker 2015 Vol. 14, No. 2 218 have been unsuccessful at a number of historical localities (Beamer and Lamb 2008). After the initial rediscovery in the Big Thicket National Preserve Canyonlands Unit (P. Crump, Houston Zoo, Houston, TX, pers comm.), we concentrated our efforts on first-order, spring-fed streams with sandy substrates, and all populations were found in these sorts of habitats. Undoubtedly more populations exist in similar conditions, but access on private property is not readily available and not all 18 counties in which they were historically found (or surrounding counties) were surveyed for this study. Thus, additional data is needed before we can say with absolute certainty that there has been a range reduction or population declines. We did, however, survey without any success the sites that had the most known specimen records. It is clear from our results that Texas populations of Desmognathus should be referred to as D. conanti, as proposed by Beamer and Lamb (2008). Also the habitat associations of the specimens that we found fit those of D. conanti in other areas. Figure 2. Bayesian phylogram of the mtDNA cox1 gene for Desmognathus in east Texas. Species included in this analysis were: Desmognathus conanti, D. apalachicolae, D. brimleyorum, D. carolinensis, D. fuscus, D. monticola, D. santeetlah, and D. ocoee. Asterisks at nodes denote Bayesian posterior probabilities ≥ 0.95, and numbers indicate posterior probabilities ≥ 50%. Numbers appended to non-Texas samples indicate sequences obtained from GenBank. Southeastern Naturalist 219 T.J. Hibbitts, S.A. Wahlberg, and G. Voelker 2015 Vol. 14, No. 2 Desmognathus have also been found in cypress sloughs and swamps in Texas, but we found no specimens in any of these habitats. It is possible that old specimens from cypress swamps represent D. auriculatus or another undescribed species, but we believe that is unlikely. Desmognathus auriculatus, sensu Beamer and Lamb (2008), is restricted to east of the Apalachicola River in Georgia and Florida. No populations sampled to date west of the Apalachicola River are genetically more similar to D. auriculatus than to D. conanti, even with fairly thorough sampling (Beamer and Lamb 2008, this study). As in several other amphibian species, it appears that major rivers are dispersal barriers (Vieites et al 2006, Zhao et al 2009). Within Texas, the salamanders to the east of the Neches River are all more closely related to one another than to those west of the Neches, although our specimens from central Newton County are closer in straight line distance to those salamanders from Tyler County than the ones from southern Sabine County. Specimens are known from west of the Trinity River, but we were unable to find an extant population. Our results would suggest that if these populations are indeed extirpated, then we have already lost genetically distinct population segments. Our results further highlight the importance of the population to the west of the Neches River and the need to locate additional populations in this area, if any in fact exist. Fortunately, this area is under the protection of the national park system. In conclusion, our data support the assertion made by Beamer and Lamb (2008) that the Texas dusky salamanders are D. conanti. We documented 4 extant populations in eastern Texas, and Texas populations can be genetically distinguished from other populations of D. conanti; however, at some Texas localities that seemed to contain robust populations in the past (judged by specimen collection numbers at localities), D. conanti are apparently now either extirpated or in very low numbers. Acknowledgments We would like to thank Don Shepard for tissues of several species of Desmognathus as well as discussions about the project. The following people helped with field surveys: Rich Adams, Jarred Melvin, Matthijs Hollanders, Matt Buckingham, Taylor Hall, and Bob Dink. The Texas A&M University Institutional Animal Care and Use Committee approved this research under AUP 2013-0081, and this project was also completed under the authority of a permit issued by the Texas Parks and Wildlife Department (SPR-0506-662). This is publication number 1489 of the Biodiversity Research and Teaching Collections at Texas A&M University. Literature Cited Beamer, D.A., and T. Lamb. 2008. Dusky salamanders (Desmognathus, Plethodontidae) from the coastal plain: Multiple independent lineages and their bearing on the molecular phylogeny of the genus. Molecular Phylogenetics and Evolution 47:143–153 Cook, M.L., and B.C. Brown. 1974. Variation in the genus Desmognathus (Amphibia: Plethodontidae) in the western limits of its range. Journal of Herpetology 8:93–105. Dixon, J.R. 2013. Amphibians and Reptiles of Texas. Texas A&M University Press, College Station, TX. 447 pp. Southeastern Naturalist T.J. Hibbitts, S.A. Wahlberg, and G. Voelker 2015 Vol. 14, No. 2 220 Houlahan, J.E., C.S. Findlay, B.R. Schmidt, A.H. Meyer, and S.L. Kuzmin. 2000. Quantitative evidence for global amphibian population declines. Nature 404:752–755. Huelsenbeck, J.P., and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754–755. Kozak, K.H., A. Larson, R.M. Bonett, and L.J. Harmon. 2005. Phylogenetic analysis of ecomorphological divergence, community structure, and diversification rates in dusky salamanders (Plethodontidae: Desmognathus). Evolution 59:2000–2016. Livezey, R.L. 1950. An intergrade population of the Texas subspecies of Desmognathus fuscus. American Midland Naturalist 3:600–604. Means, D.B., and J. Travis. 2007. Declines in ravine-inhabiting dusky salamanders of the southeastern US Coastal Plain. Southeastern Naturalist 6:83–96. Nylander, J.A.A. 2004. MrModeltest, Version 2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University, Uppsala, Sweden. Petranka, J.W. 1998. Salamanders of the United States and Canada. Smithsonian Institution Press, Washington, DC. 587 pp. Sanders, O., and H.M. Smith. 1949. Some noteworthy records of amphibians in Texas. Transactions of the Kansas Academy of Science 52:28–29. Vieites, D.R., Y. Chiari, M. Vences, F. Andreone, F. Rabemananjara, P. Bora, S. Nieto- Román, and A. Meyer. 2006. Mitochondrial evidence for distinct phylogeographic units in the endangered Malagasy Poison Frog, Mantella bernhardi. Molecular Ecology 15:1617–1625. Zhao, S., Q. Dai, and J. Fu. 2009. Do rivers function as genetic barriers for the Plateau Wood Frog at high elevations? Journal of Zoology 279:270–276.