2007 SOUTHEASTERN NATURALIST 6(4):571–576 Multivariate Analyses of Shell Morphology of Putative Hybrid Box Turtles William I. Lutterschmidt1,*, Sian A. Escobar1, and Everett D. Wilson1 Abstract - Terrapene carolina triunguis (Three-toed Box Turtle) and Terrapene ornata ornata (Ornate Box Turtle) are sympatric in areas of eastern Texas. Usually, these species are easily distinguished by shell shape and pigmentation. However, many turtles from areas of eastern Texas have characteristics of both Three-toed and Ornate Box Turtles, which we presume to be a result of hybridization. We collected, marked, and measured 103 box turtles from Walker County for morphometric analyses. Because of our limited number of Ornate Box Turtles sampled (n = 6), an additional 68 specimens from eastern Texas in the Texas Cooperative Wildlife Collection (TCWC) were examined, and the carapace and plastron measured to characterize species-specific shell morphology for Ornate Box Turtles. Multivariate analyses indicated that all morphometrics of carapace shape (length, width, curvature length, and curvature width) loaded heavily (r = 0.929, 0.897, 0.936, and 0.955, respectively) on the first principle component (Factor 1) and explained 86.41% of the variation in shell morphology and best distinguished observed differences between species and putative hybrids. Putative hybrids demonstrated a shell morphology similar to that of Three-toed Box Turtles and differed significantly from that of Ornate Box Turtles. Of the total 177 turtles examined, we considered 78 to be Three-toed Box Turtles, 74 to be Ornate Box Turtles, and 25 to be hybrids between the two parent species. Introduction Species that are closely related phylogenetically and that occur sympatrically have a potential to hybridize. Both interspecific (e.g., Capula 1993, 2002) and intergeneric (e.g., Barber et al. 2003, Harding and Davis 1999) hybridization have been documented in reptiles. More specifically, hybridization between Terrapene carolina triunguis (Agassiz) (Three-toed Box Turtle) and Terrapene ornata ornata (Agassiz) (Ornate Box Turtle) occurs in sympatric populations (see Ernst and McBreen 1991, Ernst et al. 1994). Shannon and Smith (1949) described an individual that displayed intermediate characteristics of a Three-toed Box Turtle and an Ornate Box Turtle, and Rodeck (1949) reported the mounting of a female Three-toed Box Turtle by a male Ornate Box Turtle; copulation was interrupted. Intermediate characteristics may include skin pigmentation and mottling that is typical of Ornate Box Turtles, but with a keel-shaped shell and the presence of three toes on the hind feet similar to that in Three-toed Box Turtles (Fig. 1). However, the character of three toes is not always present in Three-toed Box Turtles (Conant and Collins 1998), creating more difficulty in distinguishing species based on these morphologic characters. We collected Three-toed Box Turtles and morphologically intermediate individuals (Fig. 1) that appear to be putative hybrids of Three-toed 1Department of Biological Sciences, Sam Houston State University, Huntsville, TX 77341. *Corresponding author - lutterschmidt@shsu.edu. 572 Southeastern Naturalist Vol.6, No. 4 Box Turtles and Ornate Box Turtles. Natural occurrences of interspecific hybridization are most likely when species utilize similar resources. Both Three-toed Box Turtles and Ornate Box Turtles are reported (see Dixon 2000) from Walker County, TX, due to the county’s mosaic of grass lands, lowland hardwoods, and pine forests. Although the Three-toed Box Turtle prefers different habitats than does the Ornate Box Turtle (forests, meadows, second-growth scrubs, and riparian areas versus prairies, savannas, woodlands, and swamps, respectively; Dodd 2001), interspersion of these preferred habitats in Walker County eliminates ecological reproductive barriers and provide an opportunity for hybridization (Ward 1968). We investigated the shell morphology of these morphologically intermediate turtles suspected of hybridization. These putative hybrid turtles collected from Walker County, TX were used to address the question: Do putative hybrid box turtles demonstrate a shell morphology that is more similar to that of the Three-toed Box Turtle or that of the Ornate Box Turtle? Figure 1. Summary of morphological characters of T. c. triunguis (Three-toed Box Turtle), T. o. ornata (Ornate Box Turtle), and putative hybrids demonstrating an intermediate morphology. 2007 W.I. Lutterschmidt, S.A. Escobar, and E.D. Wilson 573 Materials and Methods We collected, marked, measured, and released 103 box turtles (78 Threetoed Box Turtle; 25 putative hybrids) within Walker County, TX from April 2004 to June 2005 (sampling was under Texas Parks and Wildlife Scientific Collectors Permit #0499-028 issued to W.I. Lutterschmidt). Turtles were sexed, and body mass and the following shell morphology measurements recorded: carapace length (CL), carapace width (CW), carapace curvature Figure 2. Graphed results of Factors 1 and 2 of the PCA for shell morphology. The top two panels represent the analysis for all individuals, middle and bottom panels are for male and female analyses, respectively. (A) Red points represent T. c. triunguis (Threetoed Box Turtle), blue points represent T. o. ornate (Ornate Box Turtle), and open triangles represent putative hybrids. (B) The 95% confidence intervals for Factor 1 and 2 graphed on a smaller scale indicate that the shell morphology of putative hybrid box turtles is similar to that of theThree-toed Box Turtle, while both putative hybrids and Three-toed Box Turtles differ significantly from Ornate Box Turtles. 574 Southeastern Naturalist Vol.6, No. 4 length (CCL), carapace curvature width (CCW), and plastron length (PL). Blood from the femoral vein of each box turtle was drawn and immediately frozen and stored at -70 °C for future molecular analyses. Female turtles were x-rayed to determine reproductive condition. Because only six Ornate Box Turtles were collected, an additional 68 specimens from the Texas Cooperative Wildlife Collection (TCWC) at Texas A&M University were measured for shell morphology. Only adult specimens from eastern Texas were measured and used in the analysis to avoid potential influences of geographic variation in shell morphology. Statistical analyses of shell morphology were conducted using SPSS® 8.0 statistical software, and all analyses were considered to be significant at P < 0.05. Results Multivariate analyses were used to investigate possible morphometric differences in shell morphology among Three-toed Box Turtles, Ornate Box Turtles, and putative hybrids. Preliminary analyses indicated that body mass and plastron length did not distinguish differences between Three-toed Box Turtles, Ornate Box Turtles, and putative hybrids (P > 0.05) and were therefore removed from the final analysis. Secondly, due to preservation techniques, the museum specimens of Ornate Box Turtles cannot offer accurate measures of body mass, thereby making comparisons of body mass with living specimens invalid. Thus, morphometric data consisted of measures characterizing carapace shape (i.e., CL, CW, CCL, and CCW) for each turtle species and individuals identified as putative hybrids. The principle components analysis (PCA) indicated that there was a significant effect of species and sex on shell morphology (Table 1). Eigen values indicated that all morphometrics of carapace shape (CL, CW, CCL, and CCW) loaded heavily (r = 0.929, 0.897, 0.936, and 0.955, respectively) on the first principle component (Factor 1) and explained 86.41% of the variation in shell morphology among species groups (Table 2). Carapace width (CW) was the only morphometric that correlated slightly (r = 0.438) with the second principle component (Factor 2), which explained only 6.97% of Table 1. Reported results (SPSS® 8.0, 1997) of the analysis of variance within the principle components analysis (PCA) of). Factor 1 and Factor 2 explain 86.41% and 6.97% of the variation in shell morphology, respectively. Dependent Source variable df SS MS F P Species Factor 1 2 342.88 171.44 62.39 less than 0.001 Factor 2 2 28.84 14.42 22.51 less than 0.001 Sex Factor 1 1 39.40 39.40 14.34 less than 0.001 Factor 2 1 4.08 4.08 6.37 0.012 Species*sex Factor 1 2 10.64 5.32 1.94 0.147 Factor 2 2 6.74 3.37 5.26 0.006 Within (error) Factor 1 171 469.90 2.75 Factor 2 171 109.56 0.641 Corrected total Factor 1 176 841.79 Factor 2 176 145.64 2007 W.I. Lutterschmidt, S.A. Escobar, and E.D. Wilson 575 the variation in shell morphology (Fig. 2a). The first principle component (Factor 1) best distinguished species groups and shows that putative hybrids have a shell morphology most similar to that of the Three-toed Box Turtle. The 95% confidence intervals for Factor 1 and Factor 2 overlap for the Three-toed Box Turtle and putative hybrids; the Ornate Box Turtle does not overlap and differs significantly in shell morphology (Fig. 2b). Discussion Pigmentation and basic shell morphology are commonly used to distinguish Three-toed Box Turtles and Ornate Box Turtles. A solid yellow plastron and a narrow and keeled carapace with flaring on the posterior portion of the shell identify the Three-toed Box Turtle. The carapace of the Ornate Box Turtle is dorsally compressed, unkeeled, and strongly pigmented with five to nine light stripes on the second costal scute. The plastron is also heavily pigmented with dark mottling. Individuals with a morphology intermediate between two closely related species are assumed to occur from hybridization between those species (e.g., Azevedo et al. 2003, Kasahara et al. 1998, Sullivan 1986). Our morphometric analysis indicates that individuals with intermediate characteristics are more similar to the shell morphology of the Three-toed Box Turtle than that of the Ornate Box Turtle. Putative hybrids also retain the presence of three toes on the hind feet as in Three-toed Box Turtles, but demonstrate shell pigmentation patterns that closely resemble Ornate Box Turtles. Although, museum (TCWC) and county records (Dixon 2000) document the occurrence of Ornate Box Turtles within Walker County, our sampling provided no Ornate Box Turtles while providing 78 individuals of Three-toed Box Turtles. Second, based upon putative hybrid shell morphology being more similar to that of the Three-toed Box Turtle, we believe that these hybrids are most likely Three-toed Box Turtles that have obtained shell pigmentation from occasional hybridization events with Ornate Box Turtles (Ernst and McBreen 1991). Future work will analyze the collected blood samples with molecular techniques to investigate the degree of hybridization for individual box turtles with characteristics intermediate between Three-toed Box Turtles and Ornate Box Turtles. Although such techniques, including the use of allozymes and microsatellites, are helpful in evaluating hybridization, morphometric analyses are used often and are adequate to identify and investigate hybridization in turtles (e.g., Seidel and Palmer 1991, Seidel et al. 1999) and other amphibians and reptiles (e.g., Arevalo et al. 1993, Gergus et al. 1999). Table 2. Reported Eigen values and % variance for each of the four principle components of the PCA . Component Eigen values % of variance 1 (CL, CW, CCL, CCW) 3.456 86.409 2 (CW) 0.279 6.971 3 0.165 4.125 4 0.099 2.494 576 Southeastern Naturalist Vol.6, No. 4 Acknowledgments We thank the curators and staff of the Texas Cooperative Wildlife Collection at Texas A&M University as well as J. Kelly McCoy, Diane L. Neudorf, Monte L. Thies, and anonymous reviewers for helpful comments that have improved this manuscript. Literature Cited Arevalo, E., G. Casas, S.K. Davis, G. Lara, and J.W. Sites, Jr. 1993. Parapatric hybridization between chromosome races of the Sceloporus grammicus complex (Phrynosomatidae): Structure of the Ajusco transect. Copeia 1993:352–372. Azevedo, M.F.C., F. Foresti, P.R.R. Ramos, and J. Jim. 2003. Comparative cytogenetic studies of Bufo ictericus, B. paracnemis (Amphibia, Anura), and an intermediate form in sympatry. Genetics and Molecular Biology 26:289–294. Barber, R.C., C.T. Fontaine, J.P. Flanagan, and E.E. Louis, Jr. 2003. Natural hybridization between a Kemp’s Ridley (Lepidochelys kempii) and Loggerhead Sea Turtle (Caretta caretta) confirmed by molecular analysis. Chelonian Conservation and Biology 4:701–704. Capula, M. 1993. Natural hybridization in Podarcis sicula and P. wagleriana (Reptilia: Lacertidae). Biochemical Systematics and Ecology 21:373–380. Capula, M. 2002. Genetic evidence of natural hybridization between Podarcis sicula and Podarcis tiliguerta (Reptilia: Lacertidae). Amphibia-Reptilia 23:313–321. Conant, R., and J.T. Collins. 1998. Reptiles and Amphibians of Eastern/Central North America. Houghton Mifflin Co., New York, NY. 616 pp. Dixon, J.R. 2000. Amphibians and Reptiles of Texas. Texas A&M University Press, College Station, TX. 421 pp. Dodd, K. 2001. North American Box Turtles: a Natural History. University of Oklahoma Press, Norman, OK. 231 pp. Ernst, C.H., and J.F. McBreen. 1991. Terrapene carolina. Catalog of American Amphibians and Reptiles 512:1–13. Ernst, C.H., R.W. Barbour, and J.E. Lovich. 1994. Turtles of the United States and Canada. Smithsonian Institution Press, Washington, DC. 578 pp. Gergus, E.W.A., D.B. Malmos, and B.K. Sullivan. 1999. Natural hybridization among distantly related toads (Bufo alvarius, Bufo cognatus, Bufo woodhousii) in central Arizona. Copeia 1999:281–286. Harding, J.H., and S.K. Davis. 1999. Clemmmys insculpta (Wood Turtle) and Emydoidea blandingii (Blanding’s Turtle) hybridization. Herpetological Review 30:225–226. Kasahara, S., A.P.Z. Silva, and S.L. Gruber. 1998. Use of lymphocyte cultures for BrdU replication banding patterns in anuran species (Amphibia). Genetics and Molecular Biology 21:471–476. Rodeck, H.G. 1949. Notes on box turtles in Colorado. Copeia 1949:32–34. Seidel, M.E., and W.M. Palmer. 1991. Morphological variation in turtles of the genus Pseudemys (Testudines: Emydidae) from central Atlantic drainages. Brimleyana 17:105–135. Seidel, M.E., J.N. Stuart, and W.G. Degenhardt. 1999. Variation and species status of slider turtles (Emydidae: Trachemys) in the southwestern United States and adjacent Mexico. Herpetologica 55:470–487. Shannon, F.A., and H. Smith. 1949. Herpetological results of the University of Illinois field expedition, spring 1949. I. Introduction, Testudines, Serpentes. Transactions of the Kansas Academy of Sciences 52:494–509. Sullivan, B.K. 1986. Hybridization between the toads Bufo microscaphus and Bufo woodhousii in Arizona: Morphological variation. Journal of Herpetology 20:11–21. Ward, J.P. 1968. Presumed hybridization of two species of box turtle. Copeia 1968:874–875.