HERPETOLOGICA VOL. 67 DECEMBER 2011 NO. 4 MORPHOLOGICAL DIFFERENTIATION IN OUACHITA MOUNTAIN ENDEMIC SALAMANDERS

Transcription

1 HERPETOLOGICA VOL. 67 DECEMBER 2011 NO. 4 Herpetologica, 67(4), 2011, E 2011 by The Herpetologists League, Inc. MORPHOLOGICAL DIFFERENTIATION IN OUACHITA MOUNTAIN ENDEMIC SALAMANDERS DONALD B. SHEPARD 1,5,6,KELLY J. IRWIN 2, AND FRANK T. BURBRINK 3,4 1 Sam Noble Oklahoma Museum of Natural History and Department of Zoology, 2401 Chautauqua Avenue, University of Oklahoma, Norman, OK 73072, USA 2 Arkansas Game & Fish Commission, 915 E. Sevier Street, Benton, AR 72015, USA 3 Department of Biology, 2800 Victory Boulevard, College of Staten Island City University of New York, Staten Island, NY 10314, USA 4 City University of New York, Biology Program, Graduate Center, 365 Fifth Avenue, New York, NY 10016, USA ABSTRACT: Morphology of some groups of organisms has been highly conserved over evolutionary time, resulting in genetically divergent taxa with relatively little morphological variation. Genetic studies on salamanders of the genus Plethodon have revealed a high level of morphologically cryptic diversity, and recent ecological and evolutionary studies have suggested that morphology may be informative for taxonomic purposes when combined with genetic data. We examined morphological variation in the Ouachita Mountain endemic salamanders P. and P. ouachitae to test whether these sister species, and recently identified phylogeographic lineages within species as defined by genetic data, can be discriminated using morphology. We found that P. and P. ouachitae differed significantly in morphology and could be classified correctly with greater than 90% accuracy. Individuals from a previously identified hybrid zone were morphologically intermediate to the two parental species. Phylogeographic lineages within each species (n 5 4 for P., n 5 7 for P. ouachitae) also exhibited significant morphological differences and were classified correctly in 57 72% of cases. Our results support genetic evidence that showed significant divergence between P. and P. ouachitae and among the different lineages within each species. The high levels of diversity within P. and P. ouachitae have important implications for conservation because these endemic species have small ranges and lineages are usually restricted to single mountains. Key words: Cryptic species; Hybridization; Morphology; Niche conservatism; Plethodon; Plethodontidae AN INDIVIDUAL9S morphology is the product of evolutionary history and environmental pressures (biotic and abiotic) acting during ontogeny (Foote, 1997; Wainwright and Reilly, 1994). Many morphological traits vary to some degree; however, the gain or loss of a particular trait in an ancestral species may severely constrain the range of potential morphological variation in its evolutionary descendents (Gould and Lewontin, 1979). Because of such 5 PRESENT ADDRESS: Bell Museum of Natural History and Department of Fisheries, Wildlife, and Conservation Biology, 1987 Upper Buford Circle, 100 Ecology Building, University of Minnesota, St. Paul, MN 55108, USA 6 CORRESPONDENCE: , dshepard@umn.edu phylogenetic constraints, morphology has been highly conserved in some groups of organisms over evolutionary time (e.g., turtles; Rieppel and Reisz, 1999), resulting in the uncoupling of morphological diversification and speciation (Adams et al., 2009; Kozak et al., 2005, 2006). Salamanders in the genus Plethodon are lungless, direct-developing ectotherms that require mesic environments for survival and reproduction; therefore, their distributions are strongly influenced by moisture and temperature (Grover, 2000; Jaeger, 1971; Petranka, 1998; Spotila, 1972). This narrow range of suitable environmental conditions is thought to constrain morphology while also promoting the isolation of populations, resulting in high 355

2 356 HERPETOLOGICA [Vol. 67, No. 4 FIG. 1. Map of the United States showing the location of the Ouachita Mountains (A), major mountains within the range of Plethodon and P. ouachitae showing elevations.500 m (B), and a digital elevation map showing the localities of specimens of P. (N) and P. ouachitae (#) used in Shepard and Burbrink (2008, 2009) and this study (C). Localities within the hybrid zone of Duncan and Highton (1979) are shown as 3. Elevation within this region ranges from a low of 135 m (black) to a high of 818 m (white). species diversity with little morphological variation among species (Adams et al., 2009; Kozak and Wiens, 2006; Kozak et al., 2005, 2006; Larson, 1989; Wake et al., 1983). The glutinosus group is the most speciose clade of Plethodon and many taxa are considered to be undifferentiated morphologically, having been described solely on the basis of genetic data (Highton, 1995; Highton et al., 1989; Highton and Peabody, 2000; Kozak et al., 2006; Wiens et al., 2006). Carr (1996), however, found significant morphological differences among many species in this group and reported a positive correlation between morphological divergence and genetic distance. In addition, several recent studies on Plethodon have found subtle differences in morphology associated with climatic variation, interactions with cooccurring closely related species (e.g., character displacement), and genetic drift following significant time in isolation (Adams, 2004, 2010; Adams and Rohlf, 2000; Adams et al., 2007; Arif et al., 2007; Myers and Adams, 2008; Wilson and Larsen, 1999). These findings suggest that despite a relatively high level of conservatism in Plethodon, morphology can still be useful in ecological and evolutionary studies, and has the potential to be informative for taxonomic purposes when combined with genetic data. Species in the Plethodon ouachitae complex are members of the glutinosus group and are endemic to high-elevation, mesic forest in the Ouachita Mountains of southeastern Oklahoma and west-central Arkansas, USA (Fig. 1; Duncan and Highton, 1979; Shepard and Burbrink, 2008, 2009, 2011; Trauth and Wilhide, 1999). The Ouachita Mountains, which comprise the southern portion of the Interior Highlands, are an ancient mountain system that shares a close geologic and biogeographic relationship with the Appalachian Mountains to the east, where Plethodon species diversity is highest (Hatcher et al., 1989; Mayden, 1985, 1988; Petranka, 1998; Thomas, 1985). Dunn and Heinze (1933) first described P. ouachitae from Rich Mountain in Polk County, Arkansas, followed by

3 December 2011] HERPETOLOGICA 357 FIG. 2. Photographs of the different variants of Plethodon ouachitae, P., and their hybrids. See the online version of this article to view a color copy of this figure. Pope and Pope (1951), who described a second species, P. caddoensis, within this species complex from the Caddo Mountains in (Fig. 1). In a more extensive survey of the Ouachita Mountains, Blair and Lindsay (1965) showed that the range of P. ouachitae also included the higher elevations of Black Fork, Winding Stair, Buffalo,

4 358 HERPETOLOGICA [Vol. 67, No. 4 Kiamichi, Round, and Fourche mountains, and noted marked differences in dorsal color pattern among mountains, which led them to name four variants (Fig. 2): the Rich Mountain variant (Rich and Black Fork mountains), the Winding Stair variant (Winding Stair and Buffalo mountains), the Kiamichi Mountain variant (Kiamichi and Round mountains), and the Buck Knob variant (Fourche Mountain). Using allozymes, Duncan and Highton (1979) found that genetic differences between the Buck Knob variant and other populations of P. ouachitae were significant enough to warrant species recognition and described it as P.. They also found large genetic differences among the Kiamichi, Winding Stair, Round, Buffalo, Rich, and Black Fork Mountain populations of P. ouachitae, but did not recognize these as distinct species. Furthermore, Duncan and Highton (1979) found a narrow zone (approximately 1.8 km wide) on the western end of Fourche Mountain where P. and P. ouachitae hybridize. Within Plethodon, hybridization in contact zones is fairly common between closely related species (Chatfield et al., 2010; Weisrock et al., 2005; Weisrock and Larson, 2006; Wiens et al., 2006), leading some authorities who strictly follow the biological species concept to disregard many species, including P. (e.g., Petranka, 1998). Recent molecular phylogenetic studies, however, have supported the recognition of all three currently recognized species in the P. ouachitae complex (Kozak et al., 2006, 2009; Shepard and Burbrink, 2008, 2009, 2011; Wiens et al., 2006). Additionally, Shepard and Burbrink (2008, 2009) found that P. is composed of four geographically distinct lineages and that P. ouachitae is composed of seven lineages distributed across six major mountains (same six mountains sampled by Duncan and Highton). Because the lineages corresponded to specific mountains that were separated from other mountains by warmer, more xeric valleys that appeared to act as barriers to gene flow, Shepard and Burbrink (2008, 2009) suggested that each lineage might represent a distinct species. In light of past taxonomic uncertainty concerning P. and new phylogeographic data for both P. and P. ouachitae, this study aims to determine whether genetic differentiation among species and lineages is accompanied by morphological differences. First, we determined whether P. and P. ouachitae differ morphologically. Second, we assessed whether individuals from the hybrid zone, as defined by Duncan and Highton (1979), are morphologically intermediate to the two parental species. And third, we tested if lineages within P. and P. ouachitae, as identified by Shepard and Burbrink (2008, 2009), can be discriminated using morphology. Correspondence between genetic and morphological data would support recognizing species or lineages as distinct taxa and would provide insight into patterns of morphological diversification during the evolutionary history of this clade. MATERIALS AND METHODS Data Collection We examined the specimens of P. and P. ouachitae analyzed in the phylogeographic studies of Shepard and Burbrink (2008, 2009). Using these specimens permitted us to identify the evolutionary lineage of each individual, which included the following designations: P. ouachitae Kiamichi W, Kiamichi E, Round Mountain, Rich Mountain, Black Fork, Winding Stair, or Buffalo Mountain; P. W. Fourche, Blue Mountain, Buck Knob, or Little Brushy (Shepard and Burbrink 2008, 2009). Additionally, we used nine individuals of P. and two individuals of P. ouachitae that were not included in Shepard and Burbrink (2008, 2009), but were subsequently sequenced, analyzed, and assigned to one of their described lineages (Appendix I). We collected, preserved, and examined all specimens from 2004 to We preserved the specimens in 10% formalin and then transferred them to 70% ethanol for permanent storage in the Sam Noble Oklahoma Museum of Natural History at the University of Oklahoma, Norman, OK. We defined hybrids as those individuals collected from the zone delineated by Duncan and Highton (1979). Although we do not know with certainty that all of these individuals were hybrids, and hybrid individuals likely varied in the genetic contributions of each parental species, classifying

5 December 2011] HERPETOLOGICA 359 individuals from this zone as hybrids is the most conservative approach. We measured only individuals.43 mm snout vent length (SVL), which gave us sample sizes of 111 for P., 221 for P. ouachitae, and 25 for their hybrids. Size at sexual maturity for P. ouachitae was reported as mm SVL for males and mm SVL for females (Highton, 1962; Pope and Pope, 1951). No specific information on size at sexual maturity for P. was available, but we assumed it to be similar to P. ouachitae because they are sister taxa and adult individuals are similar in size (Duncan and Highton, 1979; Kozak et al., 2006, 2009; Wiens et al., 2006). We determined the sex of individuals via dissection and examination of gonads. It was not necessary to dissect individuals that had an obvious mental gland because only adult males in species of Plethodon have a well-developed mental gland during the breeding season. For each individual, we used digital calipers to measure the following variables to the nearest 0.01 mm: SVL distance from the tip of the snout to the posterior margin of the cloaca, head width (HW) width of the head at its widest point, head length (HL) distance from the midline of the gular fold to the tip of the snout, head height (HH) height of the head at its tallest point, canthus rostralis length (CR) distance from the anterior edge of the eye to the posterior edge of the nostril, interorbital distance (IOD) shortest distance between the eyes taken dorsally, body width (BW) width of the trunk at its widest point, body height (BH) height of the trunk at its tallest point, axilla groin length (AGL) distance from the anterior edge of the hind limb insertion to the posterior edge of the front limb insertion, humerus length (HUM) distance from the insertion of the forelimb to the tip of the elbow, and femur length (FEM) distance from the insertion of the hindlimb to the tip of the knee. These measurements are commonly used in studies of morphological variation in plethodontid salamanders (Adams et al., 2009; Carr, 1996; Wilson and Larsen, 1999). Characters of the feet and digits have also been used (e.g., Carr, 1996; Wilson and Larsen, 1999); however, species in the P. ouachitae complex have mite infestations, unlike their syntopic congeners, which often deform their feet and digits and make them unreliable for use as characters in morphological studies (Duncan and Highton, 1979; Pope and Pope, 1951; Winter et al., 1986). All measurements were made by the same person (DBS) to avoid between-researcher variation, and bilateral variables (CR, AGL, HUM, and FEM) were taken on the specimen s right side, except on the rare occasion when damage or deformity necessitated measuring the character on the specimen s left side. Statistical Analyses We tested for differences in SVL between P. and P. ouachitae and between males and females using a two-way ANOVA with species and sex as classes and SVL as the dependent variable. Because all morphological measurements are related to overall body size and covary with one another, we created sizeadjusted, independent morphological variables to determine whether species differed in other morphological attributes (e.g., shape). To do this, we first pooled all data, regressed each variable against SVL, and calculated the standardized residuals to remove the effect of body size. We then used principal components analysis (PCA) on the correlation matrix of these new size-adjusted variables to convert them to a set of uncorrelated (orthogonal) variables. We retained all components and used the scores for each as dependent variables in a multivariate ANOVA (MANOVA) to test for morphological differences between species and sexes. Significant multivariate effects were followed by univariate tests for each component. Hybrid individuals were included in the regressions used to size-standardize variables and in the PCA in order to calculate PC scores for them, but they were excluded from the ANOVA and MANOVA. Next, we performed a discriminant-function analysis (DFA) with SVL and the sizeadjusted morphological variables to calculate the probability of correctly classifying the two species using morphology. Prior probabilities of group membership were based on group sample sizes and separate analyses were conducted for males and females because the previous analysis indicated significant sexual dimorphism in both species. We excluded hybrids from the discriminant analyses, but

6 360 HERPETOLOGICA [Vol. 67, No. 4 TABLE 1. Means (6 SD) of morphological variables (mm) measured for Plethodon, P. ouachitae, and their hybrids. F: female, M: male, n: sample size, SVL: snout vent length, HW: head width, HL: head length, HH: head height, CR: canthus rostralis length, IOD: interorbital distance, BW: body width, BH: body height, AGL: axilla groin length, HUM: humerus length, FEM: femur length. P. P. ouachitae Hybrids F M Total F M Total F M Total Sex n SVL HW HL HH CR IOD BW BH AGL HUM FEM used the resulting functions to classify them to one of the two parental species. We tested for differences in SVL among lineages within P. (n 5 4) and P. ouachitae (n 5 7) using separate ANOVAs for each species, with lineage as the class and SVL as the dependent variable. A significant overall ANOVA was followed by post hoc Tukey honestly significant difference (HSD) tests to determine which lineages were significantly different from each other. To test whether lineages within species differed in other aspects of morphology, we first recalculated size-adjusted variables as above for each species separately. For each species, we then used these variables in a PCA and used the scores for each component as dependent variables in a MANOVA with lineage and sex as classes. Significant multivariate effects were followed by univariate tests and post hoc comparisons among lineages were made using Tukey HSD tests. Next, we performed a DFA for each species using SVL and the sizeadjusted morphological variables to calculate the probability of correctly classifying the different lineages using morphology. Prior probabilities of group membership were based on group sample sizes. We conducted separate DFAs for males and females of each species and excluded hybrids from these analyses. We conducted all statistical analyses using SPSS v.17.0 (SPSS Inc., 2009) and log 10 - transformed data. RESULTS Interspecies Differentiation Snout vent length differed significantly between P. and P. ouachitae (F 1, , P, 0.001), with P. being larger on average (Table 1). Males and females did not differ significantly in SVL (F 1, , P ) in either species (i.e., nonsignificant interaction: F 1, , P ; Table 1). Principal components analysis on the size-adjusted morphological variables produced 10 orthogonal components, of which the first seven explained more than 90% of the total variation (Table 2). A MANOVA using the scores of the 10 PCs as dependent variables showed significant multivariate differences between P. and P. ouachitae

7 December 2011] HERPETOLOGICA 361 TABLE 2. Factor loadings for morphological variables for each principal component (PC) and the percentage of the total variation that each PC explains (%Var. exp.) for the analysis comparing Plethodon and P. ouachitae. Values at the bottom of each PC column are the P-values from an ANOVA testing for differences between species and sexes for that PC. Significant results are shown in bold. PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8 PC9 PC10 HW HL HH CR IOD BW BH AGL HUM FEM %Var. exp Species, , Sex,0.001,0.001, Species 3 sex (Wilk s l , F 10, , P, 0.001). Males and females also differed significantly (Wilk s l , F 10, , P, 0.001) in both species (i.e., nonsignificant interaction: Wilk s l , F 10, , P ). Follow-up univariate tests showed that species differed significantly in PC1, PC3, and PC6, whereas sexes differed in PC1, PC2, PC3, PC5, and PC6 (Table 2). Plots of individual scores for PC1 and PC3 showed that species separated primarily along the PC3 axis in both sexes, but still overlapped a moderate amount (Fig. 3). Hybrid individuals appeared intermediate between the parental species in both females and males; however, overlap with P. was greater than overlap with P. ouachitae (Fig. 3). On the basis of factor loadings for PC1 and PC3, males had greater HW and HH than females in both species, and P. ouachitae had greater IOD and CR but smaller HW and HH than P. (Fig. 3). Discriminant-function analysis using SVL and the size-adjusted morphological variables correctly classified 93.2% of females and 93.5% of males to species. Of the 12 female hybrids, 8 (66.7%) were classified as P. and 4 (33.3%) were classified as P. ouachitae. Of the 13 male hybrids, 8 (61.5%) were classified as P. and 5 (38.5%) were classified as P. ouachitae. Intraspecies Differentiation Lineages within P. did not differ significantly in SVL (F 3, , P ). Size-adjusted morphological variables subjected to PCA produced 10 orthogonal components, of which the first seven explained more than 90% of the total variation (Table 3). A MANOVA using the scores of the 10 PCs as dependent variables showed significant multivariate differences between lineages of P. (Wilk s l , F 30, , P, 0.001). Males and females also differed significantly (Wilk s l , F 10, , P, 0.001) and this was consistent across lineages (i.e., nonsignificant interaction: Wilk s l , F 30, , P ). Follow-up univariate tests showed that lineages differed significantly in PC1 and PC7, whereas sexes differed in PC1, PC2, and PC4 (Table 3). All lineages were significantly different (all P # 0.01) from one another in PC1 except for the Blue Mountain and Little Brushy lineages (P ) and the Blue Mountain and Buck Knob lineages (P ). For PC7, only the Blue Mountain and Little Brushy lineages were significantly different (P ). Plots of individual scores for PC1 and PC2 showed that lineages within P. separated primarily along the PC1 axis, with one lineage also diverging along the PC2 axis in males; however, lineages still overlapped considerably in morphospace (Fig. 4). On the basis of factor loadings for PC1 and PC2, males typically had greater IOD and HW than females and lineages differed in BW, BH, HH, and IOD (Fig. 4). Discriminant-function analysis using SVL and the size-adjusted

8 362 HERPETOLOGICA [Vol. 67, No. 4 FIG. 3. Scatter plot of principal component (PC) scores for PC1 and PC3 for female (A) and male (B) Plethodon, P. ouachitae, and their hybrids. In (C), mean scores for the two species and their hybrids are plotted by sex along with the variables that had the highest loadings on PC1 and PC3. The angle and length of arrows indicate the direction and magnitude of the correlation of the variable with each axis and show how species and sexes differ with respect to the variables. morphological variables correctly classified 72.2% of females and 61.4% of males to one of the four lineages. Lineages within P. ouachitae significantly differed in SVL (F 6, , P, 0.001). Post hoc Tukey HSD tests showed that individuals from the Kiamichi E, Kiamichi W, and Round Mountain lineages were significantly smaller on average (P, 0.05) than individuals from the Black Fork and Buffalo Mountain lineages (Fig. 5). A PCA on the size-adjusted morphological variables produced 10 orthogonal components, of which the first eight explained.90% of the total variation (Table 4). A MANOVA using the scores of the 10 PCs as dependent variables showed significant multivariate differences between lineages of P. ouachitae (Wilk s l , F 60, , P, 0.001). Males and females also differed significantly (Wilk s l , F 10, , P, 0.001) and this was consistent across lineages (i.e., nonsignificant interaction: Wilk s l , F 60, , P ). Follow-up univariate tests showed that lineages differed significantly in PC1, PC2, PC7, PC9, and PC10, whereas sexes differed in PC1 and PC2 (Table 4). Although ANOVA revealed a significant overall effect of lineage on PC1, no lineages were significantly different from each other in post hoc tests (all P. 0.05). For PC2, the Buffalo Mountain lineage was significantly different from the Black Fork, Kiamichi E, and Rich Mountain lineages (all P, 0.002) and the Kiamichi E and Round Mountain lineages were also different (P ). The Buffalo Mountain and Rich Mountain lineages were significantly different in PC7 (P ), the Kiamichi W lineage was significantly different from the Round Mountain, Black Fork, Winding Stair, and Rich Mountain lineages in PC9 (all P, 0.01), and the Winding Stair and Rich Mountain lineages differed significantly in PC10 (P ). Plots of individual scores for PC1 and PC2 showed that lineages within P. ouachitae separated primarily along the PC2 axis; however, lineages overlapped greatly in morphospace (Fig. 4). On the basis of factor loadings for PC1 and PC2, males typically had greater IOD, HL, and CR than females, and lineages differed in BW, BH, HW, and HH (Fig. 4). Discriminant-function

9 December 2011] HERPETOLOGICA 363 TABLE 3. Factor loadings for morphological variables for each principal component (PC) and the percentage of the total variation that each PC explains (%Var. exp.) for the analysis comparing the four lineages within Plethodon. Values at the bottom of each PC column are the P-values from an ANOVA testing for differences between lineages and sexes for that PC. Significant results are shown in bold. PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8 PC9 PC10 HW HL HH CR IOD BW BH AGL HUM FEM %Var. exp Lineages, Sex 0.039, Lineages 3 sex analysis using SVL and the size-adjusted morphological variables correctly classified 57.7% of females and 69.4% of males to one of the seven lineages. DISCUSSION Strong niche conservatism in Plethodon has facilitated the generation of high species diversity, but at the same time appears to have constrained morphological variation (Adams et al., 2009; Kozak et al., 2006, 2009; Kozak and Wiens, 2006; Larson, 1989; Wake et al., 1983). Consequently, distinguishing closely related species of Plethodon is difficult without genetic data (Highton, 1995; Highton et al., 1989; Highton and Peabody, 2000). The degree of morphological variation among species in the glutinosus group is positively related to genetic distance (Carr, 1996); therefore, morphological differences are predicted to be smallest for recently diverged taxa. Genetic data indicate that P. and P. ouachitae diverged approximately 2.2 million years ago (mya) and lineage diversification within each species occurred during the Middle Pleistocene (approximately mya; Shepard and Burbrink, 2008, 2009). Our results here showed that P. and P. ouachitae differ morphologically and can be discriminated with greater than 90% accuracy. Therefore, recognition of the two species as suggested by genetic studies is well supported by our morphological data (Duncan and Highton, 1979; Kozak et al., 2006, 2009; Shepard and Burbrink, 2008, 2009; Wiens et al., 2006). Individuals from the hybrid zone were morphologically intermediate between the two parental species, but the majority (64%) was more similar to P.. Shepard and Burbrink (2009) found that individuals from the westernmost locality in the hybrid zone had mitochondrial deoxyribonucleic acid (mtdna) of P. ouachitae, but individuals from other localities farther east in the hybrid zone had mtdna of P.. Furthermore, they noted that individuals in the eastern part of the hybrid zone had the typical dorsal pattern of P., whereas the dorsal pattern of individuals from the westernmost locality was highly variable and not typical of either parental species (Shepard and Burbrink, 2009; Fig. 2). Duncan and Highton (1979) also documented intermediate color patterns and a high level of individual variation in hybrids, and found that frequencies of P. ouachitae alleles sharply decreased as one moved eastward within the hybrid zone. Hybridization is common between closely related species of Plethodon that presumably diverged in allopatry and then come back into secondary contact (Chatfield et al., 2010; Weisrock et al., 2005; Weisrock and Larson, 2006; Wiens et al., 2006). Limited hybridization in contact zones is not enough to obscure the historic effects of prolonged isolation in Plethodon because gene flow between

10 364 HERPETOLOGICA [Vol. 67, No. 4 FIG. 4. Scatter plot of principal component (PC) scores for PC1 and PC2 for the four lineages within Plethodon (A: females, B: males) and the seven lineages within P. ouachitae (D: females, E: males). In (C) and (F), mean scores for each lineage are plotted by sex along with the variables that had the highest loadings on PC1 and PC2 for each species. The angle and length of arrows indicate the direction and magnitude of the correlation of the variable with each axis and show how lineages and sexes differ with respect to the variables. species typically does not penetrate much beyond these narrow contact zones (Chatfield et al., 2010; Duncan and Highton, 1979; Weisrock et al., 2005; Weisrock and Larson, 2006). Furthermore, hybridization in contact zones is likely a crucial step in the evolution of reproductive isolation through reinforcement, and therefore the occurrence of hybridization should not necessarily preclude recognition of taxa as distinct species (Coyne and Orr, 2004; Wake, 2006). Lineages within P. showed no difference in overall body size (SVL); however, lineages within P. ouachitae from the

11 December 2011] HERPETOLOGICA 365 FIG. 5. Average snout vent length (SVL) for the seven lineages within Plethodon ouachitae. Error bars are 95% confidence intervals. Means with the same letter above (a,b) are not significantly different from each other (P. 0.05). Sample sizes are listed below. Kiamichi Mountains (Kiamichi W, Kiamichi E, Round Mountain lineages) were significantly smaller in SVL than lineages of P. ouachitae on other mountains. The smaller size of P. ouachitae in the Kiamichi Mountains (including Round Mountain) has been reported previously (Blair and Lindsay, 1965; Duncan and Highton, 1979). In the Kiamichi Mountains, P. ouachitae is syntopic with P. kiamichi, another member of the glutinosus group, and they occur in near equal numbers (Duncan and Highton, 1979; D. Shepard, personal observation). Other sympatric members of the glutinosus group (e.g., P. albagula) are largely absent from the higher elevations on Rich, Black Fork, Winding Stair, and Buffalo mountains (Duncan and Highton, 1979; Trauth and Wilhide, 1999; D. Shepard, personal observation), and behavioral experiments have shown that P. ouachitae is highly aggressive and able to exclude the larger P. albagula (Anthony et al., 1997). Because of intense interspecific competition, morphological character displacement is a common phenomenon in Plethodon where closely related species co-occur, and body size is a major axis along which plethodontids diversify (Adams, 2004, 2007, 2010; Adams and Rohlf, 2000; Adams et al., 2007; Kozak et al., 2005, 2009; Hairston, 1951). The smaller body size of P. ouachitae in the Kiamichi Mountains may have evolved to reduce competitive interactions with the larger, syntopic P. kiamichi. Because the Kiamichi Mountains were the most probable ancestral area for P. ouachitae and dispersal northward to the other mountains occurred more recently in its evolutionary history (Shepard and Burbrink, 2008), the evolution of larger body size in lineages on the other mountains may be the result of competitive release or divergent selection pressures associated with competitive interactions with P. albagula. Despite high overall similarity, significant differences in morphology independent of body size were found among lineages within each species. This subtle variation allowed lineages to be discriminated with a moderate level of accuracy (57 72%) given their recent TABLE 4. Factor loadings for morphological variables for each principal component (PC) and the percentage of the total variation that each PC explains (%Var. exp.) for the analysis comparing the seven lineages within Plethodon ouachitae. Values at the bottom of each PC column are the P values from an ANOVA testing for differences between lineages and sexes for that PC. Significant results are shown in bold. PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8 PC9 PC10 HW HL HH CR IOD BW BH AGL HUM FEM %Var. exp Lineages 0.021, Sex, Lineages 3 sex

12 366 HERPETOLOGICA [Vol. 67, No. 4 origins (Shepard and Burbrink, 2008, 2009). Whether morphological differences among lineages are the result of genetic drift or are due to divergent selective pressures associated with variation in environmental factors on different mountains is unknown and warrants further investigation. Lineages within P. ouachitae also vary in dorsal color pattern (Fig. 2); however, similar patterns are shared by some lineages (i.e., the so-called variants of P. ouachitae; Blair and Lindsay, 1965; Duncan and Highton, 1979; D. Shepard, personal observation). No geographic variation in dorsal pattern has been reported in P.,but the size and color of the dorsal spots can vary substantially among individuals (D. Shepard, personal observation). These observations provide little insight into the evolution of color pattern differences and highlight the need for additional work to determine whether it has any adaptive significance. Much of the species diversity in Plethodon is morphologically cryptic, which has important implications for conservation because many genetically distinct taxa have small ranges and may be restricted to single mountaintops (Highton, 1995; Highton et al., 1989; Shepard and Burbrink, 2008, 2009, 2011). Identification and recognition of independent evolutionary lineages is imperative to conserve biodiversity. The Appalachian Mountains have the highest diversity of Plethodon species, and most genetic work to identify morphologically cryptic species has been conducted there (Highton, 1995; Highton et al., 1989; Highton and Peabody, 2000). The Ouachita Mountains have a smaller number of species; however, comparatively little work has been conducted to examine patterns of genetic diversity within species until recently. The Ouachita Mountain endemic salamanders P. and P. ouachitae are genetically and morphologically differentiated, and each species is composed of multiple geographically distinct and genetically divergent lineages that exhibit a moderate level of morphological differentiation (Blair and Lindsay, 1965; Duncan and Highton, 1979; Shepard and Burbrink, 2008, 2009; this study). These results suggest that lineages within P. and P. ouachitae likely warrant recognition as species (Wiens and Penkrot, 2002). However, additional work using multiple independent nuclear loci and populationlevel sampling of all lineages is needed to conclusively support such a decision. This work is currently underway. Although the entirety of the range of P. and most of the range of P. ouachitae lie within the boundaries of the Ouachita National Forest, which affords some protection, forest management practices (e.g., timber harvesting, controlled burning, etc.) may need to be modified to ensure longterm conservation of all evolutionary lineages. Acknowledgments. For help with fieldwork, we thank R. Bartlett, T. Colston, G. Costa, S. Filipek, D. Filipek, A. Fink, T. Gendusa, T. Guiher, L. Irwin, R. Lewis, K. Roberts, S. Ruane, B. Timpe, and the Arkansas Herpetological Society. We also thank R. Bastarache, S. Cochran, B. Crump, J. Davis, and E. Sharp of the Ouachita National Forest and M. Howery of the Oklahoma Department of Wildlife Conservation for facilitating our project. We thank D. Adams, J. Caldwell, G. Costa, N. Franssen, I. Schlupp, and L. Vitt for constructive comments on an earlier version of this paper. Funding was primarily provided by a State Wildlife Grant through the Arkansas Game and Fish Commission. All collecting was done under permits to DBS or FTB from the State of Oklahoma, State of Arkansas, and US Forest Service. This work was completed in partial fulfillment of DBS s requirements for a Ph.D. at the University of Oklahoma, and was conducted under the approval of the Animal Care and Use Committees of the University of Oklahoma (#R06-007) and College of Staten Island (12-Y3-08). LITERATURE CITED ADAMS, D. C Character displacement via aggressive interference in Appalachian salamanders. Ecology 85: ADAMS, D. C Organization of Plethodon salamander communities: Guild-based community assembly. Ecology 88: ADAMS, D. C Parallel evolution of character displacement driven by competitive selection in terrestrial salamanders. BMC Evolutionary Biology 10:72. ADAMS, D. C., AND F. J. ROHLF Ecological character displacement in Plethodon: Biomechanical differences found from a geometric morphometric study. Proceedings of the National Academy of Sciences of the United States of America 97: ADAMS, D. C., M. E. WEST, AND M. L. COLLYER Location-specific sympatric morphological divergence as a possible response to species interactions in West Virginia Plethodon salamander communities. Journal of Animal Ecology 76: ADAMS, D. C., C. M. BERNS, K.H.KOZAK, AND J. J. WIENS Are rates of species diversification correlated with rates of morphological evolution? Proceedings of the Royal Society B-Biological Sciences 276: ANTHONY, C. D., J. A. WICKNICK, AND R. G. JAEGER Social interactions in two sympatric salamanders:

13 December 2011] HERPETOLOGICA 367 Effectiveness of a highly aggressive strategy. Behaviour 134: ARIF, S., D. C. ADAMS, AND J. A. WICKNICK Bioclimatic modelling, morphology, and behaviour reveal alternative mechanisms regulating the distributions of two parapatric salamander species. Evolutionary Ecology Research 9: BLAIR, A. P., AND H. L. LINDSAY, JR Color pattern variation and distribution for two large Plethodon salamanders endemic to the Ouachita Mountains of Oklahoma and Arkansas. Copeia 1965: CARR, D. E Morphological variation among species and populations of salamanders in the Plethodon glutinosus complex. Herpetologica 52: CHATFIELD, M. W. H., K. H. KOZAK, B. M. FITZPATRICK, AND P. K. TUCKER Patterns of differential introgression in a salamander hybrid zone: Inferences from genetic data and ecological niche modelling. Molecular Ecology 19: COYNE, J. A., AND H. A. ORR Speciation. Sinauer, Sunderland, Massachusetts, USA. DUNCAN, R., AND R. HIGHTON Genetic relationships of the eastern large Plethodon of the Ouachita Mountains. Copeia 1979: DUNN, E. R., AND A. A. HEINZE A new salamander from the Ouachita Mountains. Copeia 1933: FOOTE, M The evolution of morphological diversity. Annual Review of Ecology and Systematics 28: GOULD, S. J., AND R. C. LEWONTIN The spandrels of San Marco and the Panglossian paradigm: A critique of the adaptationist programme. Proceedings of the Royal Society of London B 205: GROVER, M. C Determinants of salamander distributions along moisture gradients. Copeia 2000: HAIRSTON, N. G Interspecies competition and its probable influence upon the vertical distribution of Appalachian salamanders of the genus Plethodon. Ecology 32: HATCHER, R. D., JR., W. A. THOMAS, AND G. W. VIELE The Appalachian Ouachita Orogen in the United States. The Geology of North America, Volume F-2. Geological Society of North America, Boulder, Colorado, USA. HIGHTON, R Revision of North American salamanders of the genus Plethodon. Bulletin of the Florida State Museum 6: HIGHTON, R Speciation in eastern North American salamanders of the genus Plethodon. Annual Review of Ecology and Systematics 26: HIGHTON, R., AND R. B. PEABODY Geographic protein variation and speciation in salamanders of the Plethodon jordani and Plethodon glutinosus complexes in the southern Appalachian Mountains with the description of four new species. Pp In R. C. Bruce, R. G. Jaeger, and L. D. Houck (Eds.), The Biology of Plethodontid Salamanders. Kluwer Academic/ Plenum, New York, New York, USA. HIGHTON, R., G. C. MAHA, AND L. R. MAXSON Biochemical evolution in the slimy salamanders of the Plethodon glutinosus complex in the eastern United States. Illinois Biological Monographs 57: JAEGER, R. G Moisture as a factor influencing the distribution of two species of terrestrial salamanders. Oecologia 6: KOZAK, K. H., AND J. J. WIENS Does niche conservatism promote speciation? A case study in North American salamanders. Evolution 60: KOZAK, K. H., A. LARSON, R. M. BONETT, AND L. J. HARMON Phylogenetic analysis of ecomorphological divergence, community structure, and diversification rates in dusky salamanders (Plethodontidae: Desmognathus). Evolution 59: KOZAK, K. H., D. W. WEISROCK, AND A. LARSON Rapid lineage accumulation in a nonadaptive radiation: Phylogenetic analysis of diversification rates in eastern North American woodland salamanders (Plethodontidae: Plethodon). Proceedings of the Royal Society B- Biological Sciences 273: KOZAK, K. H., R. W. MENDYK, AND J. J. WIENS Can parallel diversification occur in sympatry? Repeated patterns of body-size evolution in coexisting clades of North American salamanders. Evolution 63: LARSON, A The relationship between speciation and morphological evolution. Pp In D. Otte and J. Endler (Eds.), Speciation and its Consequences. Sinauer, Sunderland, Massachusetts, USA. MAYDEN, R. L Biogeography of Ouachita Highland fishes. Southwestern Naturalist 30: MAYDEN, R. L Vicariance biogeography, parsimony, and evolution in North American freshwater fishes. Systematic Zoology 37: MYERS, E. M., AND D. C. ADAMS Morphology is decoupled from interspecific competition in Plethodon salamanders in the Shenandoah Mountains, USA. Herpetologica 64: PETRANKA, J. W Salamanders of the United States and Canada. Smithsonian Institution Press, Washington, DC, USA. POPE, C. H., AND S. H. POPE A study of the salamander Plethodon ouachitae and the description of an allied form. Bulletin of the Chicago Academy of Sciences 9: RIEPPEL, O., AND R. R. REISZ The origin and early evolution of turtles. Annual Review of Ecology and Systematics 30:1 22. SHEPARD, D. B., AND F. T. BURBRINK Lineage diversification and historical demography of a sky island salamander, Plethodon ouachitae, from the Interior Highlands. Molecular Ecology 17: SHEPARD, D. B., AND F. T. BURBRINK Phylogeographic and demographic effects of Pleistocene climatic fluctuations in a montane salamander, Plethodon. Molecular Ecology 18: SHEPARD, D. B., AND F. T. BURBRINK Local-scale environmental variation generates highly divergent lineages associated with stream drainages in a terrestrial salamander, Plethodon caddoensis. Molecular Phylogenetics and Evolution 59: SPOTILA, J. R Temperature and water in the ecology of lungless salamanders. Ecological Monographs 42: SPSS, INC SPSS v SPSS Inc., Chicago, Illinois, USA. THOMAS, W. A The Appalachian-Ouachita connection: Paleozoic orogenic belt at the southern margin of North America. Annual Review of Earth and Planetary Sciences 13:

14 368 HERPETOLOGICA [Vol. 67, No. 4 TRAUTH, S. E., AND J. D. WILHIDE Status of three plethodontid salamanders (Genus Plethodon) from the Ouachita National Forest of southwestern Arkansas. Journal of the Arkansas Academy of Science 53: WAINWRIGHT, P. C., AND S. M. REILLY Ecological Morphology: Integrative Organismal Biology. University of Chicago Press, Chicago, Illinois, USA. WAKE, D. B Problems with species: Patterns and processes of species formation in salamanders. Annals of the Missouri Botanical Garden 93:8 23. WAKE, D. B., B. G. ROTH, AND M. H. WAKE On the problem of stasis in organismal evolution. Journal of Theoretical Biology 101: WEISROCK, D. W., AND A. LARSON Testing hypotheses of speciation in the Plethodon jordani species complex with allozymes and mitochondrial DNA sequences. Biological Journal of the Linnean Society 89: WEISROCK, D. W., K. H. KOZAK, AND A. LARSON Phylogeographic analysis of mitochondrial gene flow and introgression in the salamander, Plethodon shermani. Molecular Ecology 14: WIENS, J. J., AND T. A. PENKROT Delimiting species using DNA and morphological variation and discordant species limits in spiny lizards (Sceloporus). Systematic Biology 51: WIENS, J. J., T. N. ENGSTROM, AND P. T. CHIPPINDALE Rapid diversification, incomplete isolation, and the speciation clock in North American salamanders (Genus: Plethodon): Testing the hybrid swarm hypothesis of rapid radiation. Evolution 60: WILSON, A. G., JR., AND J. H. LARSEN, JR Morphometric analysis of salamanders of the Plethodon vandykei species group. American Midland Naturalist 141: WINTER, D. A., W. M. ZAWADA, AND A. A. JOHNSON Comparison of the symbiotic fauna of the family Plethodontidae in the Ouachita Mountains of western Arkansas. Proceedings of the Arkansas Academy of Science 40: Accepted: 20 August 2011.Associate Editor: Michael Freake APPENDIX I. Specimens used in this study in addition to those in Shepard and Burbrink (2008, 2009) with voucher numbers (DBS: Donald B. Shepard; KJI: Kelly J. Irwin) and locality data (FR 5 United States Forest Service Road). Individuals were sequenced for cytb or ND4 (or both) and analyzed to place them into their respective lineage (D. Shepard, personal observation). Elevation (Elev.) is in meters above sea level. Geographic coordinates are based on the NAD83 map datum. ID# Species Lineage Locality Latitude Longitude Elev. Sex DBS 1632 KJI 1120 KJI 1121 KJI 1123 KJI 1124 DBS 2045 DBS 2046 DBS 2047 DBS 2048 DBS 1262 DBS 1563 Plethodon Blue Mountain Plethodon ouachitae Plethodon ouachitae Blue Mountain, N slope along FR 54, Scott County, Arkansas Round Mountain Lynn Mountain, N slope below FR 6025 along tributary to Pigeon Creek, LeFlore County, Oklahoma Rich Mountain Spring Mountain, 3.8 km E of Hwy 1/Talimena Drive on Spring Mtn Rd/FR 6007, LeFlore County, Oklahoma Male Female Male Female Male Female Male Male Female Male Male