<%BANNER%>

Morphology, Molecules, and Delineation of the Gulf Coast Box Turtle

Permanent Link: http://ufdc.ufl.edu/UFE0023986/00001

Material Information

Title: Morphology, Molecules, and Delineation of the Gulf Coast Box Turtle
Physical Description: 1 online resource (80 p.)
Language: english
Creator: Butler, Jason
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: Wildlife Ecology and Conservation -- Dissertations, Academic -- UF
Genre: Wildlife Ecology and Conservation thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The eastern box turtle (Terrapene carolina) is a widespread and often abundant component of regional faunas. Despite the familiarity of T. carolina, little is known regarding the relationships among recognized subspecies. In the Florida panhandle, an area containing known genetic discontinuities in other turtles, four subspecies of T. carolina are distributed in close proximity. I implemented morphological and molecular analysis to understand the relationships among subspecific T. carolina taxa and explore how lineages are distributed across the Florida panhandle. I performed discriminate function analysis on 31 morphological characters from 723 individual T. carolina to develop canonical functions descriptive of T. carolina taxa. I validated these taxa as distinct evolutionary lineages by comparing mitochondrial DNA sequence fragments. I tested for intergradation between these taxa through analysis of a multilocus microsatellite dataset. I used this dataset to test for genetic structure among box turtles in the Florida panhandle. Morphological functions and mitochondrial DNA sequences may be used to discriminate between T. carolina subspecies. However, specimens from the Florida panhandle exhibit several distinct phenotypes and mitochondrial haplotypes. Despite this apparent geographic overlap of lineages, microsatellite analysis did not reveal genetic structure related to mitochondrial haplotypes, geographic distance or physiographic features. Spatial autocorrelation analysis suggests genetic structuring among box turtles may occur at relatively large geographic scales. The lack of fine-scale genetic structure may be attributed to the presence of transient males, juvenile dispersal or human relocation of T. carolina.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jason Butler.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Dodd, C. Kenneth.
Local: Co-adviser: Austin, James.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0023986:00001

Permanent Link: http://ufdc.ufl.edu/UFE0023986/00001

Material Information

Title: Morphology, Molecules, and Delineation of the Gulf Coast Box Turtle
Physical Description: 1 online resource (80 p.)
Language: english
Creator: Butler, Jason
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: Wildlife Ecology and Conservation -- Dissertations, Academic -- UF
Genre: Wildlife Ecology and Conservation thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The eastern box turtle (Terrapene carolina) is a widespread and often abundant component of regional faunas. Despite the familiarity of T. carolina, little is known regarding the relationships among recognized subspecies. In the Florida panhandle, an area containing known genetic discontinuities in other turtles, four subspecies of T. carolina are distributed in close proximity. I implemented morphological and molecular analysis to understand the relationships among subspecific T. carolina taxa and explore how lineages are distributed across the Florida panhandle. I performed discriminate function analysis on 31 morphological characters from 723 individual T. carolina to develop canonical functions descriptive of T. carolina taxa. I validated these taxa as distinct evolutionary lineages by comparing mitochondrial DNA sequence fragments. I tested for intergradation between these taxa through analysis of a multilocus microsatellite dataset. I used this dataset to test for genetic structure among box turtles in the Florida panhandle. Morphological functions and mitochondrial DNA sequences may be used to discriminate between T. carolina subspecies. However, specimens from the Florida panhandle exhibit several distinct phenotypes and mitochondrial haplotypes. Despite this apparent geographic overlap of lineages, microsatellite analysis did not reveal genetic structure related to mitochondrial haplotypes, geographic distance or physiographic features. Spatial autocorrelation analysis suggests genetic structuring among box turtles may occur at relatively large geographic scales. The lack of fine-scale genetic structure may be attributed to the presence of transient males, juvenile dispersal or human relocation of T. carolina.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jason Butler.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Dodd, C. Kenneth.
Local: Co-adviser: Austin, James.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0023986:00001


This item has the following downloads:


Full Text

PAGE 1

1 MORPHOLOGY, MOLECULES, AND DELINEATION OF THE GULF COAST BOX TURTLE By JASON M. BUTLER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Jason M. Butler

PAGE 3

3 To Sara

PAGE 4

4 ACKNOWLEDGMENTS I thank my supervisory committee (C. Kenneth Dodd, James Austin and James P. Ross) for their insight and support. I thank the members of the UF Conservation Genetics lab for inspiration and camaraderie. I th ank the staffs at the Florida Museum of Natural History, the National Museum of Natural History, the Sternbe rg Museum of Natural History, the University of Kansas Natural History Museum and the Florid a Integrated Science Center for their insight and assistance. I thank multiple individuals who provided guidance and support, whether they realize it or not, including Be n Atkinson, Matt Aresco, Frank Fontanella, Phil Spinks, Brad Shaffer, Joe Collins, Albert Meier, Larry Wilson and several employees of the Florida Fish and Wildlife Conservation Commission. I thank the Amer ican Museum of Natural History, James P. Ross and James Austin of the University of Fl orida and the Society for the Study of Amphibians and Reptiles for financial assist ance. This endeavor could not have been accomplished without the unconditional support of my wife, Sara Moore Butler.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4LIST OF TABLES ...........................................................................................................................7LIST OF FIGURES .........................................................................................................................8ABSTRACT ...................................................................................................................... ...............9 CHAP TER 1 INTRODUCTION .................................................................................................................. 11Regional Phylogeography .......................................................................................................11Phylogeography of the Box Turtle .........................................................................................12Relationships of the Box Turtle ..............................................................................................13Study Objectives .....................................................................................................................152 MATERIALS AND METHODS ...........................................................................................18Sampling Strategy ...................................................................................................................18Morphological Data Collection ..............................................................................................19Molecular Data Collection ......................................................................................................19Lineage Validation and Assignment ....................................................................................... 21Morphology .....................................................................................................................21MtDNA Sequences .......................................................................................................... 23Gene Flow ...............................................................................................................................233 RESULTS ....................................................................................................................... ........30Lineage Validation and Assignment ....................................................................................... 30Morphology .....................................................................................................................30MtDNA Sequences .......................................................................................................... 31Gene Flow ...............................................................................................................................314 DISCUSSION .................................................................................................................... .....43Lineage Validation and Assignment ....................................................................................... 43Intergradation and Gene Flow ................................................................................................46Genetic Structure ....................................................................................................................47Conclusion .................................................................................................................... ..........49APPENDIX A DISCRETE MORPHOLOGICAL CHARACTERS .............................................................. 52

PAGE 6

6 B MUSEUM AND TISSUE DATA OF TERRAPENE USED IN THIS STUDY .................... 59C MTDNA HAPOTYPES AND MICROSATELLITE ALLELES .......................................... 65LIST OF REFERENCES ...............................................................................................................74BIOGRAPHICAL SKETCH .........................................................................................................80

PAGE 7

7 LIST OF TABLES Table page 2-1 Primers and conditions used in this study .......................................................................... 253-1 Discriminate Function Analysis (DFA) models ................................................................ 333-2 Discriminate Function Analysis characters .......................................................................333-3 Genetic distances. ........................................................................................................ ......343-4 Allele summaries. ..............................................................................................................343-5 Autocorrelation statistics. ..................................................................................................34B-1 Museum specimens of Terrapene used in this study .........................................................59B-2 Localities of tissue specimens used in this study ...............................................................62C-1 Variable base pair loca lities in mtDNA haplotypes ...........................................................66

PAGE 8

8 LIST OF FIGURES Figure page 1-1 Hypothesized distributions of Terrapene taxa in southeastern North Am erica. ................ 172-1 Sampling localities for this study. ......................................................................................262-2 Sampling localities in the Florida Panhandle.. ................................................................... 272-3 Shell distances measured in this study ............................................................................... 282-4 Sampling hotspots ......................................................................................................... .....293-1 Canonical plots. .......................................................................................................... ........353-2 Morphological specimen assignments ............................................................................... 363-3 Florida panhandle morphological specimen assignments.. ................................................ 373-4 Mitochondrial haplotype network ......................................................................................383-5 Haplotype distributions ......................................................................................................393-6 Panhandle detail of haplotype distributions ....................................................................... 403-7 Autocorrelation plots.. ................................................................................................... ....413-8 Genetic surface. .......................................................................................................... ........424-1 Florida coastlines ........................................................................................................ .......51A-1 Marginal flare ............................................................................................................ .........56A-2 Posterior marginals ....................................................................................................... .....57A-3 Nucal shapes .............................................................................................................. ........58C-1 Microsatellite al lele frequencies ........................................................................................ 72

PAGE 9

9 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science MORPHOLOGY, MOLECULES, AND DELINEATION OF THE GULF COAST BOX TURTLE By Jason M. Butler December 2008 Chair: C. Kenneth Dodd Cochair: James Austin Major: Wildlife Ecology and Conservation The eastern box turtle ( Terrapene carolina ) is a widespread and often abundant component of regional faunas. Despite the familiarity of T. carolina, little is known regarding the relationships among recognized subspecies. In the Florida panhandle, an area containing known genetic discontinuities in other turtles, four subspecies of T. carolina are distributed in close proximity. I implemented morphological and mol ecular analysis to understand the relationships among subspecific T. carolina taxa and explore how lineages ar e distributed across the Florida panhandle. I performed discriminate function analysis on 31 morphological characters from 723 individual T. carolina to develop canonical f unctions descriptive of T. carolina taxa. I validated these taxa as distinct evolutionary li neages by comparing mitochondrial DNA sequence fragments. I tested for intergradation between these taxa through analysis of a multilocus microsatellite dataset. I used this dataset to test for genetic structure among box turtles in the Florida panhandle. Morphological functions and mitochondrial DNA sequences may be used to discriminate between T. carolina subspecies. However, specimens from the Florida panhandle exhibit several distinct phenotypes and mitoc hondrial haplotypes. Despite this apparent geographic overlap of lineages, micr osatellite analysis did not reve al genetic structure related to

PAGE 10

10 mitochondrial haplotypes, geographic distance or physiographic features. Spatial autocorrelation analysis suggests genetic structuring among box tu rtles may occur at rela tively large geographic scales. The lack of fine-scale genetic structure may be attributed to th e presence of transient males, juvenile dispersal or human relocation of T. carolina.

PAGE 11

11 CHAPTER 1 INTRODUCTION The eastern box turtle ( T errapene carolina ) is a charismatic and important component of eastern North America fauna. Despite the familiar ity of this turtle, rela tively little is known regarding the evolutionary history and relationships of th e many phenotypically variable Terrapene taxa. I investigated the relationships among Terrapene in the Florida panhandle, where the distributions of four subspecies are in proximity. Morphological and molecular techniques possess unique benefits and limitations for understanding in tra-specific relationships. I analyzed mitochondrial and nuclear DNA as well as morphology to understand the spatial genetic structure of Terrapene across the Florida panhandle. Regional Phylogeography Dozens of studies have exam ined concorda nce in distribution patterns, explored physiographic barriers to gene flow or attempted to elucidated mechanisms of speciation across a broad array of eastern North American reside nt taxa (reviewed by Soltis et al. 2006). Southeastern North America is a particularly interesting biogeographic area due to its physiographic complexity and habitat heterogene ity. The Coastal Plain has functioned as a refugium for taxa during Pleist ocene glaciations (Braun 1947, Chur ch et al. 2003, Austin et al. 2004) and xerothermic periods (Smith 1957). Th ese periodic rufugial events allowed for allopatric diversification and produced an ecologi cally important area. The Florida panhandle, in particular, contains many rare, pr otected and endemic organisms (Wol fe et al. 1988). Expansion following these refugial events f acilitated contact of related or ganisms the Coastal Plain is recognized as a major area of secondary contact for a broad array of taxa (Swenson and Howard 2005). The Apalachicola River, which bisects the Florida panhandle into eastern and western segments, is one of the most commonly rec ognized genetic discont inuities in various

PAGE 12

12 southeastern North American taxa (Soltis et al. 2006). The area adjacent to the river is a hotspot for diversity, endemism and phylogeographic br eaks (Stein et al. 2000, Swenson and Howard 2005). Genetic structuring, often associated with the Apalachicola River, is evident among many taxa including mammals (Ellsworth et al. 1994 ), fishes (Bermingham and Avise 1986) and insects (Maskas et al. 2000) with in the Florida panhandle. This region also contains a diverse herpetofaunal assemblage punctuated by a high de gree of endemism and intraspecific genetic diversity (Walker and Avise 1998). The distributions of many in tra-specific lineages reflect associations with current and hi storic physiographic features in the region. Numerous reptile taxa exhibit a genetic break across the Apalach icola River (Burbrink 2001, Means and Krysko 2001), including several turtles (Walker and Avise 1998). These examples illustrate the biogeographic and conservation importance of the southeastern coastal plain of North America and, more specifically, the Apalachicola region of the Florida panhandle. Phylogeography of the Box Turtle The eastern box turtle (T estudines: Emydidae: Terrapene carolina) is a functional component of the regional fauna. Aside from aiding in seed dispersal in ce rtain habitats (Liu et al. 2004), box turtles are known vectors for inve rtebrate parasites a nd prey items for many vertebrates (reviewed by Dodd 2001). Despite this ecological importance, little is known about the evolutionary histories, rela tionships or distributions of Terrapene lineages. Evolutionary studies often delineate unique lineages which may be important when considering management decisions. Several biological ch aracteristics advocate the use of Terrapene as an exemplar species for a phylogeographic study. Thes e attributes are outlined below. Terrapene is a moderately-sized organism with apparently low dispersal capability. Adult box turtles maintain relatively small home-ranges (Dodd 2001) and typically do not venture vast distances (Iglay et al. 2007, Schwartz 1974). Unlike birds or large mammals, box

PAGE 13

13 turtles do not autonomously disper se across landscapes. Additionall y, box turtles are too large to be dispersed by wind like many invertebrates and pl ants. The lack of potential for widespread dispersal suggests Terrapene should demonstrate genetic stru cturing at both landscape and range-wide scales. Unlike most members of the family Emydidae which are dependent on permanent water for survival, Terrapene carolina is terrestrially adapted and maintains an existence loosely tied to aquatic habitats (Carr 1952). B ox turtles can disperse independent of aquatic connections and their potential ranges are not greatly affected by stream capture or flooding (for biogeographic studies of aquatic taxa in s outheastern North America, see Bermingham et al. 1986, Gilbert 1987). Terrapene possesses a dense and unique shell that permits preservation and recognition in the fossil record. Pleistocene deposits in Florida commonly yield partial and complete Terrapene specimens (Hulbert 2001). Consideration and analysis of the fossil record contributes to our understanding of primary mechanisms of evolutionary processes (Gandolfo 2008). Appreciating when and where box turtles hi storically resided provides insight beyond biogeographic snapshots genera ted from modern range and habitat delineations. Box turtles exhibit high phenotypic variability. There are currently six recognized subspecies of eastern box turtle; four of these (T. c. carolina, T. c. bauri, T. c. major, T. c. triungus ) occur in the vicinity of the Florida Panha ndle and exhibit unique phenotypic affinities. The geographic variation in morphological char acters provides quantitative characters for developing phylogeographic hypothese s and comparing genotypic a nd phenotypic variation. Relationships of the Box Turtle The intergrade zone dynam ics and subspecies status of box turtles has been addressed and debated in multiple studies since the 19th century. Ditmars (1934) provided a comprehensive

PAGE 14

14 review of studies from 1820 to 1933 that described Terrapene taxa and addressed the relationships among them. Milstead (1969), Wa rd (1980) and Minx (1996) conducted intensive morphology-based investigations but reached no conclusive consensus of the inter-generic relationships of Terrapene In southeastern North America, intergrade zones among subspecies of T. carolina have been proposed exclusively on phenotype (Carr 1952; Conant and Collins 1991; Milstead 1969). Carr (1952) and Milstead (1969) described the co ntact zones between these taxa as geographically expansive areas cont aining phenotypic intermediates (Figure 1-1a). Ward (1980) disagreed and diagnosed the inte rmediates described by Milstead within the variation of specific taxa (Fi gure 1-1b). Recently published distribution maps do not reflect either of these descriptions (see Cona nt and Collins 1991, Minx 1996, Dodd 2001), but do exhibit the distributional proximity of four subs pecies to the Florida pa nhandle. The lack of concordance in results from phenotypic investigations of Terrapene relationships (Ward 1980, Milstead 1969, Minx 1996) s hows the need for a study us ing molecular techniques. Molecular methods provide powerful techniques to determine the nature and extent of contact zones (Harrison 1990), part icularly in taxa where phenotypi c characterizati ons are highly variable. Analysis of mitoc hondrial deoxyribonucleic acid (mt DNA) sequence data can reveal relationships between, and the distribution of, lineages of organisms. Southeastern North American reptiles exhibit divergent mtDNA sequ ences in lizards (Clark et al. 1999, Leache and Reeder 2002), snakes (Burbrink et al. 2001, Burbri nk et al. 2002) and turtles (Starkey et al. 2003, Walker and Avise 1998). A recent chelonian checklist includes four species of Terrapene ( T. carolina T. cohuila T. nelsoni T. ornata) with six subspecies ( T. carolina bauri T. c. carolina, T. c. major T. c. mexicana, T. c. triungus T. c. yucatana) within the T. carolina complex (Fritz and Havas 2007), each of which may possess diagnostic mtDNA haplotypes. Previous

PAGE 15

15 descriptions of T. carolina suggest intergrade zones occur ne ar the panhandle region of Florida (Carr 1952, Dodd 2001, Milstead 1969). Thes e observations imply mtDNA haplotypes representative of subspecific T.carolina lineages may occur in the Florida panhandle and clines, or gradation among characters (Huxl ey 1938), exist in the landscape between them (Figure 1-1a). The investigation by Ward (1980) concluded Terrapene lineages do not exhibit extensive intergradation and implies representative genotypic and phenotypic suites co-occur in narrow geographic ranges (Figure 1-1b). Both of these patterns may result from secondary contact of related lineages (Endler 1977). Th e conclusions of Milstead (1969) suggested gradual clines of characters in large geographic contact zones, whereas Wards (1980) proposal suggested steep clines in characters in narrow contact zones. Due to the recombinant nature of nu clear DNA and rapid mutational rates of microsatellites (Goldstein and Scholotterer 1999), these markers are ideal tools for investigating intra-specific relationships. Genetic variati on, flow and structure between closely related lineages of reptiles have been explored with frequency data of nuclear microsatellite alleles (Davis et al. 2002, Hoehn et al. 2007, MacAvoy et al. 2007). An sample of box turtles from the Florida panhandle should exhib it genetic structuring reflective of the intensity of intergradation. An expansive intergradation zone would exhibit limited genetic structure whereas structuring should be obviou s across a narrow geographic area of intergradation Study Objectives The prevalence of genetic discontinuities in a variety of organism s has encouraged a priori hypothesis testing of herpetofaunal studie s conducted in the Apalachicola region of Florida (Pauley et al. 2007). My study attempts to elucidat e the genetic structure of a widespread, putatively low-dispersi ng organism in a complex bioge ographic region. First, I test

PAGE 16

16 for the presence of multiple lineages associated with described T.carolina supspecies across the Florida panhandle. I examined mtDNA sequences and a suite of morphological characters to diagnose sub-specific T. carolina taxa. I then examine populati on structuring and gene flow between lineages in the Florida panhandle thr ough analysis of a multi-locus microsatellite dataset. I attempt to address the following questions: Are subspecies diagnosable using mtDNA haplotypes and statistical assignmen t of morphological characters? Is there evidence of a clinal pattern of intergradation among lin eages based on microsatellite lo ci and phenotypic variation? Finally, based on previous phylogeographic investigations of chelonians (Walker and Avise 1998), I ask whether the Apalachicola River has genetically structured Terrapene across the Florida panhandle?

PAGE 17

17 Figure 1-1. Hypothesize d distributions of Terrapene taxa in southeastern North America. A) adapted from Carr (1952) and B) compiled from the range descriptions of Ward (1980).

PAGE 18

18 CHAPTER 2 MATERIALS AND METHODS Sampling Strategy I exam ined 723 specimens of Terrapene carolina from four museum collections for a suite of discrete, meristic and color pattern characters (Appendix B). I selected specimens to maximize sampling coverage of southeastern No rth America (Figure 2-1) with emphasis on the Florida panhandle on either side of the Apalachic ola River (Figure 2-2). I focused on specimens preserved in fluid as opposed to skeletal speci mens which lose morphol ogical characters during maceration. When coordinates we re not available, I used DELORME TOPO 6.0 to georeference each specimen based on collection locality data Specimens whose locality could not be determined within 10 kilometers were omitted from analysis. Many museum specimens are geographically clustered around collection hotspots These hotspots typically contained several dozen specimens from an area less than 5,000 hectares. Specimens of Terrapene carolina in each hotspot were grouped together to permit in-depth multivariate analysis. I grouped between 20 and 30 individuals in each of th e eleven areas (Figure 2-4). Tissue samples were collected from museum and personal collections. I received the majority of Florida panhandle samples from M. Aresco, who began collecting Gulf Coast box turtle tissue nearly a decade ago. I also acquired tissue from specimens de ep within the ranges of each currently recognized subspecies (Fig 2-1). Most Terrapene tissue samples obtained from M. Aresco and other sources were collected fr om dead on the road (DOR) specimens. Tissue was collected from 125 Terrapene specimens (Appendix C). Fixation with formalin preserves museum specimens and denatures DNA. Most collectors have only recently be gun archiving tissue prior to pr eserving specimens. Therefore, tissue samples were not available from the ma jority of museum specimens I examined.

PAGE 19

19 Conversely, although an excellent source of DNA, DOR turtles are usually too damaged to permit accurate morphological examination. Due to the nature of sampling and limited tissue collections associated with fixed specimens, few morphological and tissue specimens originate from the same animal, although ther e is broad geographic overlap. Morphological Data Collection I exam ined several commonly used mo rphological characters and several novel characters. Twenty-three disc rete character observations, incl uding three sexually dichromatic characters, were made for each turtle. I also measured eight straight-line carapace and plastron lengths with vernier calipers (Figure 2-3). Each measurement was taken three times and the mean and variance of repeated measurements were calculated (Yezerinac et al. 1992). I scrutinized any average measurement with a va riation greater than 1.0 mm. Many of these highly variable means were associated with erroneous data entry (i.e. 102.5 mm versus 1002.5 mm) and corrected respectively. Specimens w ith means whose high variation could not be accounted for were omitted from analysis. For a full description of each character, including its potential for bias, see Appendix 1. Molecular Data Collection I extracted DNA fro m chelonian tissue using standard pheno l-chloroform techniques (Sambrook and Russel 2001) following a proteinase-K digestion. Although the time for completing an extraction may be longer, phenol-chloroform techniques cost less and provide higher DNA concentrations than kit extractions The concentration of template DNA was measured using a Nanodrop spectrophotomet er (Nanodrop ND-100, Wilmington, Deleware). However, DNA recovered from DOR samples can re sult in unreliable measures of template quality with spectrophotometry due to the high c oncentration of short, degraded DNA fragments.

PAGE 20

20 Rosenbaum et al. (2007) found varying phylogeographic information in three mtDNA genes applied to an Emydid turtle I examined the most variable of these genes, a fragment of the control region displacem ent loop (d-loop), in 117 Terrapene carolina specimens. King and Julian (2004) identified 22 polymorphic microsatellite loci in box turtles. I screened each of these loci for variability acro ss a subset of box turtle samples. I selected six loci for amplification across all Florida panhandle sample s based on ease of amplif ication and variability assayed from visualization on an Elchrom SEA-2000 (Cham, Switzerland) electrophoresis apparatus. I performed polymerase chain reaction (PCR) amplifications of double-stranded product on a Eppendorf thermocycler. PCR reactions for mtDNA and microsatellite amplification contained 5 l of 1 X PCR buffer (Promega, Madison, Wisc onsin), a final concentration of 2 mM MgCl2, 0.2 mM dNTPs, 0.2 units of Taq polymerase a nd varying concentrations of primer and template DNA in a total volume of 20 l. Cond itions for PCR differed between loci but all included an initial denaturation at 94C for 3 min, 35 amplificati on cycles of 45 s denaturation at 94C, 45 s annealing at varying te mperatures, 45 s extension at 72C and a final extension of 5 min at 72C. Primer sequences, primer and temp late concentrations and annealing conditions for PCR of individual genes are found in Table 2-1. Mitochondrial sequences were cleaned of unincorporated dNTPs through application of ExoSAP-IT (USB, Cleveland, Ohio) before followi ng standard sequencing protocol on an ABI 3130xl automated sequencer (Applied Biosystems, Foster City, California). I aligned d-loop sequences in the CLC DNA Workbench program (CLC bio, Katrinebjerg, Denmark) under default protocols.

PAGE 21

21 Standard genotyping protocols were followed for fragment analysis of fluorescently labeled microsatellite product on an ABI 3130xl sequencer. I used GENEMARKER (Softgenetics, State College, Pennsylvania) to score microsatellite alleles based on ROX GS500 size standard (Applied Biosystems, Foster City, California). Seve ral loci from the microsatellite dataset were amplified and genotyped twice to address possible allelic dropout and assure consistent allelic scoring. Since degraded DNA wa s included in this study I tested for null alleles using the software MICRO-CHECKER (Oosterhout et al. 2004). Lineage Validation and Assignment Morphology I assessed the diagnostic util ity of m orphological characters in order to validate and detect differences between box turtle lineages. Diagnostic morphologi cal characters were developed through analysis of carapace and plas tron measurements corrected for body size. Milstead (1969) and Ward (1980) used measurement ratios to distinguish Terrapene taxa; however, the statistical behavior of ratios may invalidate analyt ical results (Atchley 1976). I corrected for size by regressing all characters against a size standard and reserving the residuals free from allometry (Reist 1986). I used shel l volume calculated from the product of depth, width and length measurements as a size standard. Step-wise discriminate function analyses (D FAs, JMP 7.0) were performed on combined and gender specific datasets to account for sexual dimorphism. I chose analysis models by excluding all characters which did not meet a di scriminating significance probability (p) greater than 0.05. Several assumptions und erlie the application of DFA, although large sample datasets may be robust to some violations (McGarigal et al. 2000). I inve stigated the viol ation of two of the assumptions that have the potential to invali date results. The canoni cal plots of groups that do not possess equal variance-covariance matrices may be distorted (Williams 1983). I used a

PAGE 22

22 logarithm transformation to standardize variab les so variances were equal between groups. Initial analysis of transformed and raw datasets did not produce differing results, so I used raw datasets for all analysis. Outliers may also influe nce results from DFAs (McGarigal et al. 2000). When I observed outliers in canonical plots, I removed them and reanalyzed the data. I evaluated sexually dimorphic characters to fi lter datasets by gender. I produced a male dataset by removing all specimens that did not ha ve an enlarged tail. The female dataset consisted of individuals without a concave posterior plastron lobe, enlarged tail or enlarged rear claws. Many specimens could not be sexed base d on these characters. I classified the sex of these individuals a posteriori by treating them as unknowns in an DFA of the specimens of known sex (Iverson and McCord 1991). I om itted unknown specimens that could not be assigned to a gender with at le ast 95% assignment probability. I used DFA to explore among versus within group morphological variation of box turtles from collection hotspots. I expected groups within the same geographic distribution of a subspecies to overlap in canoni cal distributions. I formed hi gher-level groups based on the overlap in canonical distributi ons and developed group-specific formulas from their canonical functions through DFA. The suitability of highe r and lower level group models was evaluated through comparisons of the amount of explained variation from the first two canonical functions and the percent of misidentified specimens. I chose the model with the most explanatory power to diagnose intraspecific box turtle lineages. Using this model, I assigned all morphological specimens to their most appropriate lineage. Specimens with a classification probability lower than 95% were omitted from further analysis. I visualized the distribution of morpho-lineages by mapping combined and gender-specific results. Finally, I examined the assignments of specimens from the Florida panhandle to determin e which lineages were present in the region.

PAGE 23

23 MtDNA Sequences The d-loop sequence data was also used to exam ine the divergence and monophyly of putative subspecies and to compare distributi ons of mtDNA with morphological lineages assigned to box turtle specimens from the Flor ida panhandle. Using sequences from each subspecies range and a subset of those from th e panhandle, I developed a haplotype network in TCS 1.13 under default settings. Exploring intraspecific relations hips with a network is often more appropriate than a phylogene tic tree due to the bifurcating assumptions of tree building and the limited resolution of intra-specific da ta (Posada and Crandall 2001). I recognized lineages based on the number of mutational steps between haplotypes and haplotype clusters. I compared corrected and uncorrected pairwise ge netic distances within and among these lineages using MEGA 4.0. Standard error of these dist ance were calculated through a 1000-replicate bootstrap. I assigned all specimens to a lineage and mapped these lineages over the specimen collection locality. A map of th e panhandle specimens was assessed to determine which lineages contribute to the region s genetic structure. Gene Flow I looked for overlap in micros atellite allele frequencies be tween box turtle lineages in the Florida panhandle to confirm or refute intergradation. I implemented a Fishers exact test with GENEPOP 4.0 to test the null hypothesis that all alle les in all lineages are from the same genetic population. Populations composed of hybrids may exhibit excess heteroz ygosity. I tested for excess heterozygosity and deviation from HardyWienberg equilibrium (HWE) in box turtle lineages alone and pooled with GENEPOP 4.0 (Raymond and Rousset 1995). I explored the suitability of the microsatellite dataset to detect natural clusters of related box turtles through a Bayesian approach implemented in STRUCTURE 2.1 (Pritchard et al. 2000).

PAGE 24

24 STRUCTURE attempts to detect natura l population clusters by simultaneously estimating cluster allele frequencies and assigning in dividuals to a cluster or a co mbination of clusters. These natural clusters may represent genetic patches or groups of individuals with more potential gene flow between each other than with individuals from other groups. The user inputs a genotypic dataset and a prediction for the number of clusters (k). STRUCTURE models the lik elihood of the observed dataset being produced by the predicted number of clusters through an admixture model using a set number of generations. To en sure equilibrium of th e proposed populations, I chose a model that incorporated 500,000 genera tions after a discarded burn-in period of 30,000 generations to calculate the lik elihood of K = 1 to 10 populations. I calculated the average likelihood for each potential number of populations from five iterations of each model. To explore the potential relationship between gene flow and geography, I used microsatellite genotypes to test for isolation by distance (IBD) across the Florida panhandle. I examined the regression of pairwise genetic differences against log transformed geographic distances (Rousett 2000) with a Ma ntel test in GenePop 4.0. I also looked for isolation by distance at a larger geographic sc ale through analysis of the discre te morphological characters. I implemented a spatial autocorrelation analysis using Alleles in Space (AIS, Miller 2005) to explore phenotypic similarity between indivi duals at various distance classes across the landscape. I plotted the slope of the autocorrelation statistic ( Ay) across 10, 20 and 30 distance classes and tested for its si gnificance against a 10 00 replicate randomization. Using AIS, I interpolated a genetic surface between morphol ogical box turtle samples across eastern North America to visualize potential areas of genetic discontinuity. I used a 100 X 100 grid and a distance weighting value of a = 0.5 in AIS to develop the surface.

PAGE 25

25 Table 2-1. Primers and conditi ons used in this study Locus Primer Sequences Flourescence/ Repeat Motif Primer Concentration Template (ng) Anealing Temperature ( C) Size Range (bp) mtDNA Starkey et al. 2003 DES1 (5'-3') GCA TTC ATC TAT TTT CCG TTA GCA 0.5 20 52 725 DES2 (3'-5') GGA TTT AGG GGT TTG ACG AGA AT 0.5 Microsatellites King and Julian 2004 GmuB08 (5'-3') CTC TGA GAC CCT TAT TCA CGT C HEX (green) / 0.15 20 60 189 208 R: (3'-5') AGC CTT TGT CTG TAA GCT GTT C TAC 0.15 GmuB12 (5'-3') TCA ATC TTC CAG CCT AAC TGT G FAM (blue) 0.3 20 60 177-190 R: (3'-5') AGG GAT GTG TTT TGC AAC TGG TAC 0.3 GmuD21 (5'-3') GCA GTT AGG CAT TAC TCA ACA TC TAMN (yellow) 0.75 40 55 149 190 R: (3'-5') AGG GTA TGA ATA CAG GGG TGT C ATCT 0.75 GmuD55 (5'-3') GTG ATA CTC TGC AAC CCA TCC FAM (blue) 0.45 20 55 154 212 R: (3'-5') TTG CAT TCA GAA TAT CCA TCAG ATCT 0.45 GmuD90 (5'-3') ATA GCA GGA CAA TTA CCA CCA G TAMN (yellow) 0.75 20 60 130 157 R: (3'-5') CCT AGT TGC TGC TGA CTC CAC ATCT 0.75 GmuD121 (5'-3') GGC AAA TAT CCA ATA GAA ATC C HEX (green) 0.9 20 55 123 163 R: (3'-5') CAA CTT CCT CGT GGG TTC AG ATCT 0.9

PAGE 26

26 Figure 2-1. Sampling localities fo r this study. Hollow circles ( ) represent museum samples used in morphological an alysis; soli d circles ( ) represent tissue sample localities. One circle may represent more than one specimen.

PAGE 27

27 Figure 2-2. Sampling localities in the Florida Panhandle. Hollow circles ( ) represent museum samples used in morphological analysis, solid circles ( ) represent tissue sample localities. One circle may represent more than one individual. The course of the Apalachicola river through the Florida Panhandle is outlined in dark grey.

PAGE 28

28 Figure 2-3. Shell distances measured in this study. A) Straight-line carapace. B) Shell height. C) Posterior plastral lobe. D) Interpectoral seam. E) Interhumeral seam. F) Interregular seam. G) Hinge. H) Shell width.

PAGE 29

29 Figure 2-4. Sampling hotspots. Circles represent hotspots of Terrapene carolina specimen localities from several museum collections. The locality centers of these hotspots are as follow: 1 Cherokee Co., Kansas, 2 Scott Co., Arkansas, 3 Forest Co., Mississippi, 4 Jackson Co., Mississippi, Calhoun Co., Flor ida, Liberty Co., Florida, 7 Alachua Co., Florida, 8 Dade Co., Flor ida, 9 Bibb Co., Georgia, 10 McMinn Co., Tennessee and 11 Prince Georges Co., Maryland. These specimens were considered groups in multivariate analyses.

PAGE 30

30 CHAPTER 3 RESULTS Lineage Validation and Assignment Morphology I rem oved 65 specimens whose sex could not be determined with at least 95% accuracy through DFA. The resulting dataset contained 340 males and 318 females. DFAs did not exhibit obvious differences between sexes so result s are presented for all specimens. The first lineage-based DFA identified a suite of 17 characters with a significant discriminating probability (p > 0.05). The mean confidence limit ellipse s (MCLE), or the area with a 95% probability of containing a groups average specimen, of many collection hotspots overlap considerably in canonica l space (Figure 3-1A). Many of the groups exhibiting overlap were located within the distribution of a co mmon subspecies. Specimens from hotspots one through three fall within the distribution of T. c. triungus and have overlapping MCLEs. Similarly, the two groups within the range of T. c. bauri overlap in canonical space. All DFA defined groups within the ranges of T. c. major and T. c. carolina overlap. The first two canonical functions of this model account for 71% of the discriminating power of the characters (Table 3-1). Under this model, 23.6% of the sp ecimens are misclassified (i.e. a specimen from group 1 should belong in group 2). Based on the observed overlap in the first lineage DFA, I developed two higher level models (Figure 3-1B and C) and compared th eir discriminating power. Both models were composed of clusters of groups which overl apped in MCLEs. I omitted groups within the Florida panhandle from one of the models (Fig ure 3-1C) to account fo r potential function bias introduced by intergradating specimens (phenotypic intermediates). The model with all eleven groups clustered into three higher levels did not exhibit overlapping MCLEs between groups

PAGE 31

31 (Figure 3-1A). Under this model, 1.7% of specimens are misclassif ied and the first two canonical functions account for 99% of the discri minating power (Table 3-1). The model that did not include panhandle specimens also did no t exhibit overlap between group MCLEs (Figure 3-1C). This model did not reveal any misclass ification of specimens and accounted for 100% of the discriminating power in the first two canonica l functions. A suite of 12 characters with significant discriminating power (Table 32) was used to develop this model. The final DFA assigned all but 84 specimens to one of the three clusters with 95% probability. The maps of the localities of each specimen assignment across the sampling range (Figure 3-2) roughly reflect recognized subspecific box turtle dist ributions (Figure 1-2). The map of assigned specimens in the Florida panhan dle (Figure 3-3) exhibits members from each cluster as well as many specimens whic h could not be reliably assigned. MtDNA Sequences The d-loop haplotype network reve aled four distinct clusters separated by at least seven mutational steps (Figure 3-4). The geographic localities of these cl usters roughly reflect recognized subspecific distributions with the ex ception of the largest, m ore geographically extensive cluster (Figure 3-5). Sp ecimens from two of the four cl usters were collected from the Florida panhandle (Figure 3-6). One of these cl usters is not represente d by specimens outside the panhandle. The uncorrected genetic divergen ce within these clusters ranges between 0.6 and 1.0% while differentiation between lineages is between 2.1% and 6% (Table 3-3). Gene Flow The software MICROCHECKER recognized the potential for null alleles in one microsatellite locus (GmuD121). This locus was not included in analysis. The two distinct dloop lineages (Figure 3-4 B, C) in the Florida pa nhandle share microsatellite alleles. The lineage restricted to the panhandle (lin eage C) does not possess any unique microsatellite alleles (Table

PAGE 32

32 3-4). Fishers exact tests of individual allele frequencies did not reject the null hypothesis that the specimens in the Florida panhandle are from the same population (p > 0.94). HWE tests in GENEPOP did not support heterozygotic deficit or excess across pooled sample. Global tests did not rejected HWE across pooled samples and loci. Structure analysis strong ly supported one cluster ( k = 1) in the Florida panhandle. Posterior probabilities of the aver age likelihoods from each simulated k were higher than 99% for k = 1 and less than 1% for all other values. I did not find evidence for IBD in the panhandle; the regression slope of genetic distance agains t geographic distance did not significantly differ from zero. Morphological analysis in AI S revealed significant spatia l autocorrelation (p > 0.95) in each set of distances (Table 3-5, Figure 3-7). The last distance class to exhibit significant autocorrelation was between 581 and 664 kilometers in the 30-class analys is. The interpolation of the genetic surface from AIS depicts one region of extremely high variability (Figure 3-8). This region corresponds with the Florida panhandle.

PAGE 33

33 Table 3-1. Discriminate Function Analysis (DFA) models Model Canonical Discriminating Power Misclassified Clusters Groups Function 1Function 2 Total 11 11 38% 33% 71% 23.6% 3 11 55% 44% 99% 1.7% 3 9 52% 48% 100% 0% These percentages represent model summaries from DFA of morphological specimens. The first model represents each of the sampling groups analyzed separately while the second and third models represent combined sampling groups that overlap in canonical space. Table 3-2. Discriminate Func tion Analysis characters Character Frequency Mean Residuals Group Cluster Head/Leg Pattern Notched Beak Dark Carapace Background No. of rear toes Vertebral Ridge Streaks Spots Blotches Posterior Lobe Interregular Seam Interhumeral Seam Depth 1,2,3 0.67 0.740.23 3.030.320.240.440.18-3.510.98 -1.72-0.87 7,8 0.84 0.291.00 3.430.840.940.680.013.11-1.80 2.694.49 9,10,11 0.64 0.280.98 3.970.490.210.510.88-1.26-0.05 -0.75-3.06 These discriminating characters from DFA were significant between clustered groups.

PAGE 34

34 Table 3-3. Genetic distances Lineage Within Pairwise A B C D E A 0.010 + 0.003 B 0.008 + 0.002 0.038 + 0.007 C 0.004 + 0.002 0.033 0.029 + 0.007 + 0.007 D 0.006 + 0.003 0.054 + 0.009 0.057 + 0.009 0.059 + 0.010 E 0.029 + 0.007 0.028 + 0.007 0.029 + 0.007 0.060 + 0.010 F 0.033 + 0.007 0.041 + 0.008 0.036 + 0.008 0.051 + 0.009 0.021 + 0.006 Within and between group uncorrected geneti c distances with bootstrap values from 1000 replicates of similar cluste rs of d-loop haplotypes Table 3-4. Allele summaries gmuB12 gmuB08 gmuD21 gmuD55 d-loop lineage n No. of Alleles Unique No. of Alleles Unique No. of Alleles Unique No. of Alleles Unique B 56 8 2 8 2 12 6 23 11 C 15 6 0 6 0 6 0 12 0 Microsatellite allele summaries for the m itochondrial lineages in the Florida panhandle Table 3-5. Autocorrelation statistics # of classes Distance per class (km) 1 2 3 4 5 6 7 5 500 0.39* 0.47 0.53 0.53 0.57 10 250 0.35* 0.42* 0.46 0.49 0.52 0.53 0.53 20 125 0.31* 0.39* 0.42* 0.42* 0.45* 0.48 0.47 30 83 0.29* 0.35* 0.41* 0.41* 0.43* 0.42* 0.44* These values are the spatial autocorrelation statis tics for four sets of dist ance classes. Asterisks indicate statistics significantly different from a random distribution.

PAGE 35

35 Figure 3-1. Canonical plots. These canonical plots are from three di scriminate function analysis models. A) the plot from the model of all groups separate, B) and C) plots from the models of clustered groups.

PAGE 36

36 Figure 3-2. Morphological specimen assignments. Specimens that could not be assigned through discriminate function analysis with 95% probability are represented by Xs.

PAGE 37

37 Figure 3-3. Florida panhandle mor phological specimen assignments. Specimens that could not be assigned with 95% probabi lity are represented by Xs.

PAGE 38

38 Figure 3-4. Mitochondrial haplot ype network. This haplotype network was derived from d-loop sequences across the sampling range. Each square represents a base pair mutational step. Hollow squares indicate a network c onnection with less than 95% probability. Lineages A, B, C and D represent wester n, eastern, panhandle and peninsular clades respectively. Lineages E and F represent one specimen each from Escambia County, Florida and Terrapene coahuila respectively.

PAGE 39

39 Figure 3-5. Haplotype dist ributions. Specimen localities are la beled based on clades from d-loop sequences. Lineage A is found prim arily within the distribution of T. c. triungus, Lineage B within the distribution of T. carolina carolina Lineage C within the Florida Panhandle and Lineag e D within the range of T. c. bauri Several lineage B specimens are found within the range of T. c. triungus and in the Florida panhandle.

PAGE 40

40 Figure 3-6. Panhandle detail of ha plotype distributions. Specimen localities are labeled based on clades from d-loop sequences. Lineages B and C co-occur in this area and do not appear to possess geographic affinities. The Apalachicola River as it flows through Florida is outlined in grey.

PAGE 41

41 Figure 3-7. Autocorrelation plots. These plots show the spatial autocorrelation of morphological characters across the sampling range. The x axis depicts pairwise geographic distances and the y axis depicts the autoco rrelation statistics for A) 5, B) 10, C) 20 and D) 30 distance classes. The dashed line is the average spatial autocorrelation statistic and its intercept with the obs erved autocorrelation line indicates the maximum distance of significan t spatial autocorrelation.

PAGE 42

42 Figure 3-8. Genetic surface. This interpolation shows the range-wide genetic surface from the phenotypic dataset. The bottom right corner and the upper left corner are equivalent to the specimen localities from the southeast and northwest respectively.

PAGE 43

43 CHAPTER 4 DISCUSSION Lineage Validation and Assignment In eastern N orth America, three of the four currently recognized subspecies of Terrapene carolina possess unique morphological and mitochondr ial characters. The DFAs revealed differences in the morphology of a western, an ea stern and a Florida lineage representative of the currently recognized T. c. triungus T. c. carolina and T. c. bauri respectively. These lineages also possess unique mitochondria l haplotypes (Figure 3-5, clad e A, clade B and clade D respectively). The distributions of specimens a ssigned to these lineages are nearly concordant with current subspecies distributional maps. I do not recognize the validity of the Gulf Coast box turtle as a distinct evolutionary lineage. I was unable to distinguish sp ecimens within the recognized range of T. c. major from the eastern lineage. Many of the phenotypic charact ers used to describe gulf coast box turtles are highly variable and also occur among eastern box turtles. Additionally, d-loop sequences from most specimens within the range of T. c. major are the same as those of the eastern lineage. However, another unique mtDNA ha plotype (lineage C, Figure 3-6) occurs in this area, and many male box turtles from the gulf coast regi on seem to possess unique characteristics including a white head (Milstead 1969), a deeply cusped beak a nd a thick zygomatic arch (Minx 1996). These morphological characters and th e lineage C mtDNA haplotype may have been inherited from the extinct giant box turtle T. c. putnami (Milstead 1969). Interestingly, specimens exhibiting classic T. c. carolina phenotypes possess lineage C d-loop haplotypes and specimens exhibiting T. c. major phenotypes possess eastern dloop haplotypes (M. Aresco, upublished data). These results corrobor ate Milsteads (1969) hypothesis that T. c. major was not a distinct lineage, but in stead a mixture of extant subspecies plus the extinct T. c. putnami.

PAGE 44

44 The distinct geographic dist ributions, character suites and divergent mitochondrial haplotypes suggest the western, eastern and Florida lineages of box turtles underwent allopatry and divergence. The divergence of clade C may al so be a product of historic vicariance due to physiographic features. Avise (1992) estimated mtDNA divergence rate s of turtles to be between 0.2-0.4% per million years. Under this estimate, the eastern, western and Florida clades diverged between 9.5 and 28.5 mya. The earliest known box turtle fossils allocated to T. carolina are from the late Miocene (~5 mya, Holman unpublished datasee Dodd 2001). Either much older T. carolina fossils remain undiscovered or the mtDNA dive rgence estimates of Avise (1992) are not applicable to the T. carolina mitochondrial control region Rates in mtDNA variation may differ severalf old among turtle taxa (Walker and Avise 1998), so I considered fossil evidence and historic geologic events to calibrate divergence rates in T. carolina. Eustatic estimates indicate sea levels of the coast of South Carolina were approximately 35 meters higher than current le vels during the middle Pliocene (3.5 2.5 mya, Dowsett and Cronin 1990). Sea levels of this he ight would have inundate d much of Florida, leaving an island disjunct from continental North America (Figure 4-1). Box turtle specimens are known from peninsular Florida prior to this event (Hulbert 2001). If box turtles persisted on the Pliocene island they would have diverged fr om those of the mainland, perhaps forming the Florida clade observed in this st udy. Using this event as a calib ration point, the divergence rate of the control region of T.carolina mtDNA is between 1.54 to 2.36% per million years. This estimate overlaps the conventional 2.0% divergen ce rate estimated for non-turtle vertebrates (Wilson et al. 1985).

PAGE 45

45 Under this estimate, eastern and western box turtles diverged between 1.61 2.47 mya while lineage C diverged from eastern box tu rtles between 1.23 and 1.88 mya.. These dates correspond with glaciations associ ated with the divergence of Chrysemys (Starkey et al. 2003). During interglacial periods, the cool waters and large floodplain of the Mississippi River may have prevented dispersal of rat sn akes (Burbrink et al. 2000). Similarly, divergence of eastern and western box turtle clades may have resulted fr om vicariance due to river valley inundations. The haplotype and phenotype distributions appear to be associated more with the Mobile basin than the Mississippi basin. Th e ancestral Mobile River basin drained a larger area than the present day basin, including the Appalachian drainages that curre ntly feed the Tennessee River (Mayden 1988). During glaciatio ns, the Mobile River basin may have produced vicariance in box turtle populations. Islands created during high sea levels in the early Pleistocene may have provided refugia for box turtles. A unique lineage of kingsnakes is thought to have dive rged while isolated on Pleistocene islands in the panhandle of Florida (Means and Krysko 2001). The unique mitochondrial haplotypes of clade C may represent a box turtle lineage which diverged from the eastern clade by persisting on islands in the Coastal Plain during times of high sea level. The divergence of clade B and C lineages c ould also be the resu lt of landscape-level habitat affinities. During high s ea levels the Coastal Plain may be reduced to a series of islands, however during glaciations and low sea levels the Coastal Plain would extend seaward beyond its current distribution. T. c. putnami may have flourished in these low, coastal areas. As sea levels advanced to modern levels, the habitat of T. c. putnami would have been reduced. The higher elevation areas of the Coasta l Plain would provide refugia for T. c. putnami but also promote mixing with lineages adapted to the Piedmont.

PAGE 46

46 A study incorporating carefully -structured sampling across th e Coastal Plain of North American could test these pot ential scenarios. An islandrefugia hypothesis would suggests several distinct but closely rela ted lineages exist across the Coastal Plain, with each of these distinct lineages representing a re fugial island availabl e during high sea-levels. Alternatively, a Coastal Plain habitat hypothesis would suggests one major lineage across the entire landscape. Intergradation and Gene Flow The m ost phenotypically diverse area across the sampling range, as depicted by the genetic surface interpolation from AIS, is the Florida panhandle (Figure 3-8). Phenotypic assignment through DFA showed that box turtles from the Florida panhandle are composed of western, eastern and Florida lineages (Figure 33), as well as many specimens which could not be assigned with confidence. However, of th ese three lineages, only eastern mtDNA haplotypes occur in the Florida panhandle samples. Based on the mtDNA evidence from the samples in this study, Terrapene in the Florida panhandle are apparently not composed of intergrades between western, eastern and Florida lineag es as proposed by Milstead (1969) However, this conclusion is not definitive since the evid ence is based solely on one mtDNA gene. This pattern could also occur through biological selec tion; the maternally inherited mtDNA from the eastern lineage may be more prevalent if female eastern Terrapene are more reproductively fit. Corroboration of this conclusion can be confirmed through c ongruence with an unlinked gene (i.e., nuclear intron). Although western and Florida lineages are not represented by panhandle specimens, lineage C haplotypes occur sympatrically with easte rn haplotypes in the Florida panhandle. Recolonization and secondary cont act of these lineages could produce the observed pattern of mtDNA haplotypes in the Florida panhandle. Analys is of the microsatellite dataset suggests these two lineages are not repr oductively isolated. Although lineages B and C may have

PAGE 47

47 diverged through allopatry, these lineages have since reticulated. The prevalence of eastern lineage specimens and the lack of unique micros atellite alleles in lineage C suggest complete swamping, at least in the Florida panhandle. One potential scenario involve s eustatic allopatry the ancestral stock to lineages B and C were se parated by high sea levels. The eastern lineage maintained a geographically expansive distribution while lineage C was confined to Coastal Plain islands. The lineages diverged; lineage B into T. c. carolina and lineage C into T. c. putnami As sea levels receded, box turtles coloni zed new, dry areas and the lineages made secondary contact. Another possible scenario i nvolves landscape-level allopatr y through habitat affinities and segregation; lineage B repr esenting a taxon inhabiting the Pied mont and lineage C inhabiting the Coastal Plain. Under this scenario, divergence would occu r during glaciations and low sea level and mixing would be promot ed as sea levels advanced. Regardless of the mechanism, it is apparent that the eastern li neage swamped the now extinct T. c. putnam i, at least in the Fl orida panhandle. Swamping may have occured by T. c. carolina outnumbering T. c. putnami but selection likely contribute d to the extinction of the giant box turtle. The larger T. c. putnami would have been selected against by predators (including man) and fire. Genetic Structure Tests for IB D and cluster assignment did not reveal structuring among box turtles within the Florida panhandle. These results, in addition to the distribution of d-loop lineages B and C, imply the Apalachicola River does not and may have never acted as a barrier to gene flow in box turtles. The lack of discon tinuity associated with the Ap alachicola River separates the phylogeographic pattern of box turtles from the pa ttern of many other members of the family Emydid, and aligns it closer to the mtDNA pa ttern exhibited by the gopher tortoise ( Gopherus

PAGE 48

48 polyphemus, Walker and Avise 1998). Interestingly, the several Emydid taxa that do exhibit genetic discontinuities associated with the Apalachicola River inhabit aquatic environments, while the gopher tortoise is fully terrestrial. The Apalachicola River genetic discontinuity does not seem to apply to terrestrial chelonians. Analysis of microsatellite loci did not reve al structuring within the Florida panhandle. Although the lack of apparent structure may be due to size homoplasy in informative microsatellite markers, limited structure could also result from high gene flow at the observed geographic scale. I examined sample-wide spat ial autocorrelation of phenotypic characters to estimate larger geographic scales of genetic structure. Significant spatial autocorrelation occurred between pairs of individuals at least 581 kilometers apart. Autocorrelation at this scale suggest the area of genetic patc hes may greatly exceed levels suggested by the sedentary nature of many box turtles. The home ranges of most box turtles rema in constant over several years (Nichols 1939). The box turtles that do not maintain a stable home range may be responsible for the lack of small-scale genetic structuring appa rent from this study. My results corroborate the suggestions of Kiester (1982) that proposed enha nced gene flow by transient box turtles would result in a lack of subdivided populations of box turtles except at large scales. Additionally, no studies have investigated the dispersal of hatchling and juvenile box turtles from their nest site. Young box turtles dispersing large distances before settling into home ranges may also produce a lack of small-scale genetic structure. Aside from natural movements, the lack of genetic structuring in box turtles may be due to human relocation. Box turtles are often remove d from the wild and kept as pets (Dodd 2001). These turtles may later escape or be release d. During their captiv ity, box turtles can be transported great distances. A three-toed box tu rtle collected in Guam serves as an extreme

PAGE 49

49 example of this process (McCoid 1992). Human re locations of box turtles have likely disrupted the genetic integrity of the Florida panhandle. Conclusion Distinct evo lutionary lineages of Terrapene can be diagnosed through statistical analysis of morphological suites and iden tification of mtDNA haplotypes. At least four divergent mtDNA lineages are present in southeastern No rth America, although morphological analysis recognized three distinct clades. Western, eastern and peninsular Florida lineages are represented by both unique mtD NA haplotypes and morphological suites. The fourth mtDNA haplotype occurs sympatrically with eastern haplotypes within th e described range of the Gulf coast box turtle, Terrapene carolina major Most specimens from this region are morphologically assignable to th e eastern box turtle, although several individuals cannot be confidently assigned. A microsatellite dataset was compiled from Terrapene specimens in the Florida panhandle. Analysis of this dataset did not show genetic structure associated with unique mtDNA haplotypes or the Apalachicola River. The lack of structur e may be a result of microsatellite size homoplasy, although the spatia l autocorrelation of morp hological traits at a large geographic scale suggests a biological explanation. Gene flow across large geographic distance may be facilitated by transient males or extensive dispersal of juvenile box turtles. Additionally, genetic structure has likely been disrupted by human relocations of Terrapene Future phylogeographic studies could elucidat e the contact zone dynamics between other Terrapene taxa. These studies should incorporat e mitochondrial and nuclear markers from samples across a large geographic scale. Additi onal studies are needed to better understand the dispersal capability of Terrapene. These studies should focus on hatchling and transient male

PAGE 50

50 turtles. The influence of human relocations on natural Terrapene populations should also be examined.

PAGE 51

51 Figure 4-1. Florida coastlines. Black areas indica te the sections of present day Florida which would remain above the sea if levels rose to those estimated during the Pliocene (+ 35m, Dowsett and Cronin, 1990)

PAGE 52

52 APPENDIX A DISCRETE MORPHOLOGICAL CHARACTERS The following descriptions detail each charac ter or m easurement a nd include the source and potential diagnostic problems when relativ e. Figures 2-3 and A-1 through A-3 depict character states and delineate landmar ks of osteological measurements. Head coloration: The coloration of fixed specimens fades with time and preservation. The presence of dark or light co loration is typically still obvious. I recorded whether the head of a specimen exhibited dark, light or both colorations. Head and leg pattern: The head of Florida box turtle s are usually patte rned with two light stripes on either side, where those of gulf coast box turtles may be uniformly black (Carr 1952). I recorded whether specimens heads and front limbs were patter ned or plain. This character may also fade with time in preserved specimens. Notched beak: Many box turtles exhibit a notch or cleft in the premaxilla. The notch in the beak of some specimens may be up to a centi meter deep while the beaks of others may be completely flat. Ditmars (1934) considered the presence of a notched max illary beak a variable character and Minx (199 6) found differences in the fre quency of notched beaks between subspecies of Terrapene carolina. In life, the beak of a turtle may wear with time, potentially obscuring the presence of this character. I ob served this character across all age groups, indicating its usefulness for most specimens. Flared anterior carapace margin: Carr (1952) reported flaring of the rear carapace marginals in T. c. carolina and T. c. major The rear marginals of box turtles are extremely variable and range from perpendicular to the plastron to upwardly curved like a gutter (see Figure 6E in Milstead 1969). I c onsidered the marginal of specimens to be flared if its outer periphery was parallel or upturned to the plastron (Figure A-1). Carapace background color: The background carapace color of Terrapene carolina is typically dark, although T.carolina triungus may possess a lighter horn-colored carapace background (Ditmars 1934). I recorded whether th e carapace background was dark or light. Rear toe count: The number of rear toes or claws varies between the subspecies of T. carolina (Minx 1996). I recorded whether specimens possessed three or four toes on each rear foot. Some specimens possessed three toes on one rear foot and four toes on the other. I recorded when this phenotype was observed, but unfortunately did not record which foot possessed the observed number of toes. Specimens missing rear feet were noted. Imbricate anterior carapace margin: While examining specimens, I noticed that some exhibit an imbricate or serrated anterior margin of the carapace (Figure A-2). I recorded the presence or absence of this character. Vertebral Ridge: Many specimens exhibit a prominent vertebral ridge, or middorsal keel a thin (< 1.0 cm), raised section of carapace bisecting the animal dorsally (Minx 1996). I

PAGE 53

53 recorded the presence of a ridge when apparent across at least one-half of the carapace. Shell wear may reduce the prominence of a verteb ral ridge, although I typically found ridges observable in very old specimens. Vertebral Stripe: Taylor (1895) originally described the Florida box turtle as possessing a yellow keel. I recorded the presence of a light vertebral strip when it was apparent over at least one half of the shell. Shell wear and scarri ng can obscure the presence of this character. Carapace pattern: Box turtles exhibit remarkable variation in carapace pattern. Differences in pattern between subspecies have been noted (Carr 1952). I recorded the presence of streaks, spots or blotches on the carapace of specimens. I defined streaks as numerous straight, light lines at least 3 cm in length; spots as numerous round, light areas of pigmentation less than 1 cm in diameter; and blotches as num erous irregular light markings greater than 1 square centimeter in area. Many specimens exhibited combinations of these patterns and were noted accordingly. This character could not be a ccurately determined for several badly scarred or poorly preserved specimens. Plastron pattern: Eastern and Florida box turtles may exhibit a patterned plastron while the plastrons of gulf coast and three-toed box turtles may be uniformly colored (Dodd 2001). I recorded whether the plastron of each specimen was uniform or patterned. With age, the plastron of a turtle may wear smooth and obscu re any obvious patterns. I addressed this potential bias by comparing the frequency of patterned plastrons between specimens with a low number of annuli and specimens w ith annuli completely worn over. Plastron coloration: The plastrons of gulf coast box turtles are dark brown or black while three toed box turtles usually possess a uniformly yellow plastron (Dodd 2001). Although plastron coloration may fade in preserved spec imens, I easily recorded whether specimens possessed a light or dark plastron. Urn-shaped nucal: Auffenberg (1967) described the pres ence of an urn-shaped anterior vertebral bone (or nucal) as a di stinguishing characteristic between T.carolina major and T.carolina subspecies. I recorded the presence of an urn-shaped nucal. A diagram of this character along with other commonly observed nucal shapes is pr ovided (Figure A-3). Sexually Dimorphic Characters Adult box turtles exhibit sexual dimorphism. The tail of male turtles is typically larger and longer than those of similarly sized females. Several studies report variation in sexually dimorphic characters between subspecies of box turt les. I recorded the presence of an enlarged tail along with several other sexually dimorphic characters described below. Concave plastron: The concave posterior plastron of male Terrapene carolina facilitates fitting of the shell of females during reproduction (Dodd 2001). Male three-toed box turtles do not usually have a plastron concavity (M ilstead 1969). I recorded the plastron state for each examined specimen.

PAGE 54

54 Depth of plastron concavity: The relative depth of the plastron concavity differs between subspecies of T. carolina with gulf coast box turtles concavities being the deepest (Dodd 2001). I considered the depth of a plaston to be deep if it was more than 20% of the length of the concave region. Shape of concave region: The plastron concavity of male box turtles is typically round, although gulf coast box turtles often possess an el ongate concavity (Milstead 1969). I recorded concavities as elongate if their lengt h was at least 110% of their width. Enlarged rear toes: The claws on the rear feet of male adult turtles of ten grow much larger than the rear claws of females of compar able size. Carr (1952) reported this character across many box turtle taxa. I record ed rear toes of a specimen as en larged if their size appeared at least twice that of toes on the front feet. I noted when I could not dete rmine the state of this character due to specimens missing their rear feet. Continuous Characters I measured straight-line distances three tim es per character for most specimens. Measurements were made to the nearest 1/100th of a millimeter with a 20 cm digital vernier caliper. Accurate measurements of some specimens were inhibited due to scarring of the animal or awkward position of the limbs. I recorded when these difficulties occurred so their potential for influencing analytical abnormalities could be examined. Posterior lobe of plastron: This distance is the length of the posterior half of the plastron, measured from the center of the hinge to the posterior end of the interanal seam. Interregular seam: I recorded this measure as the dist ance between the pair of the most anterior scutes on the plastron. Interhumeral seam: This measurement is the length of the seam between the pair of scutes in the center of the frontal lobe of the plastron. Interpectoral seam: The most posterior pair of scutes on the frontal lobe of the plastron connect to form the interpectoral seam. I measured the length of this seam. Hinge: I took this measurement acro ss the posterior po rtion of the frontal lobe along the hinge in the center of the plastron. As men tioned earlier, many museum specimens have a detached plastron. Taking this measurement along the posterior end of the frontal lobe allowed for consistent measurements across all specimens. Carapace width: I measured the width of the she ll in line with the plastron hinge. Shell depth: I measured the depth of the shell be tween the hinge and the top of the carapace directly above the hinge. The plastrons of some specimens were detached, likely for diet or reproduction studies. I recorded when plastrons were detached and fitted the plastron of these specimens snuggly against the carapace befo re taking depth measurements. The position

PAGE 55

55 of greatest shell depth differs between some box tu rtle subspecies (Carr 1952). Unfortunately, I did not record where the greatest sh ell depth occurred on each specimen. Carapace Length: I measured the straight-lined di stance from the notch between the pair of anterior marginals and the notch be tween the two most posterior marginals.

PAGE 56

56 Figure A-1. Marginal flare. Medial view of the posterior carapace of two Terrapene A) flared rear marginal, B) rear marginal not flared.

PAGE 57

57 Figure A-2. Posterior marginals. A dorsal view of the posterior carapace of Terrapene shows the differences in posterior marginals A) exhibits smooth rear marginals, B) represents serrated rear marginals.

PAGE 58

58 Figure A-3. Nucal shapes. These nucal shapes are common in Terrapene carolina including the urn-shape described by Auffenberg (A, 1952).

PAGE 59

59 APPENDIX B MUSEUM AND TISSUE DATA OF TERRAPENE USED IN THIS STUDY Table B-1. Museum specimens of Terrapene used in this study. KU 2540 KU 46722 KU 46805 KU47350 KU51452 KU 217160 KU 3013 KU 46752 KU 46807 KU47352 KU51453 KU 218795 KU 3014 KU 46753 KU 46808 KU47353 KU51454 KU 289712 KU 3063 KU 46758 KU 46810 KU47354 KU51455 UF 19 KU 3068 KU 46759 KU 46811 KU47355 KU51456 UF 659 KU 3093 KU 46760 KU 46812 KU47357 KU51457 UF 1411 KU 3142 KU 46761 KU 46813 KU47358 KU51458 UF 2349 KU 3143 KU 46762 KU 46814 KU47360 KU51460 UF 2377 KU 3144 KU 46763 KU 46815 KU47361 KU51461 UF 2378 KU 3396 KU 46765 KU 46816 KU47362 KU61854 UF 2379 KU 3832 KU 46766 KU 46817 KU47363 KU70969 UF 2380 KU 15828 KU 46767 KU 46818 KU47364 KU75135 UF 3328 KU 15829 KU 46768 KU 46819 KU47365 KU75137 UF 4019 KU 15886 KU 46769 KU 46820 KU47366 KU88838 UF 4020 KU 15890 KU 46770 KU 46821 KU47368 KU88841 UF 4225 KU 17367 KU 46771 KU 46822 KU47369 KU91357 UF 4226 KU 18387 KU 46773 KU 46823 KU47370 KU143854 UF 4227 KU 19343 KU 46774 KU 46824 KU47371 KU143855 UF 4228 KU 19344 KU 46775 KU 46825 KU47372 KU144569 UF 4230 KU 19348 KU 46776 KU 46826 KU47373 KU151894 UF 4231 KU 19478 KU 46778 KU 46827 KU47374 KU151895 UF 4232 KU 19738 KU 46780 KU 46828 KU47478 KU153637 UF 4233 KU 19739 KU 46781 KU 46829 KU47479 KU177148 UF 4234 KU 19741 KU 46784 KU 46830 KU47482 KU177149 UF 4235 KU 20936 KU 46786 KU 46831 KU47483 KU177217 UF 4236 KU 20937 KU 46787 KU 46851 KU48240 KU192259 UF 4237 KU 22818 KU 46788 KU 46853 KU48241 KU192404 UF 4239 KU 23039 KU 46789 KU 46854 KU48242 KU197234 UF 4241 KU 23337 KU 46790 KU 46893 KU48243 KU203625 UF 4245 KU 23338 KU 46791 KU 46894 KU48244 KU207136 UF 4247 KU 23340 KU 46793 KU 46917 KU48245 KU208133 UF 4406 KU 23341 KU 46794 KU 46918 KU48246 KU214312 UF 4407 KU 23342 KU 46797 KU 46919 KU48247 KU214313 UF 4408 KU 23343 KU 46798 KU 47341 KU48249 KU214317 UF 4409 KU 23344 KU 46799 KU 47342 KU48250 KU214318 UF 4410 KU 23345 KU 46800 KU 47343 KU48258 KU214320 UF 4412 KU 23346 KU 46801 KU 47344 KU50505 KU214321 UF 4413 KU 23348 KU 46802 KU 47346 KU50507 KU214322 UF 4414 KU 23349 KU 46803 KU 47347 KU50752 KU214323 UF 4415 KU 41570 KU 46804 KU 47349 KU51433 KU214324 UF 4416

PAGE 60

60 Table B-1. Continued UF 4418 UF 7865 UF 10127 UF21400UF47872UF 48161 UF 4419 UF 7874 UF 10128 UF21476UF47874UF 48162 UF 4420 UF 7907 UF 10147 UF21477UF47879UF 50789 UF 4422 UF 8464 UF 10181 UF21478UF47883UF 50802 UF 4423 UF 8465 UF 11120 UF21479UF47887UF 50813 UF 4423 UF 8473 UF 11121 UF27612UF47888UF 53406 UF 4424 UF 8590 UF 12012 UF27613UF47890UF 54660 UF 4425 UF 8591 UF 12348 UF27614UF47896UF 54661 UF 4426 UF 8592 UF 12447 UF30177UF47898UF 54662 UF 4427 UF 8619 UF 14199 UF30178UF47900UF 63785 UF 4428 UF 8739 UF 14666 UF30179UF47901UF 63786 UF 4429 UF 8905 UF 14668 UF30180UF47902UF 63788 UF 4430 UF 8906 UF 14669 UF30181UF47903UF 63790 UF 4431 UF 8937 UF 14670 UF30182UF47904UF 63791 UF 4432 UF 9253 UF 14671 UF30183UF47905UF 63792 UF 4433 UF 9350 UF 14672 UF30185UF47906UF 63793 UF 4434 UF 9390 UF 16224 UF30186UF47907UF 63793 UF 4437 UF 9391 UF 19736 UF30187UF47908UF 63795 UF 4438 UF 9392 UF 19737 UF30188UF47910UF 65954 UF 4439 UF 9409 UF 20544 UF30189UF47960UF 66321 UF 4440 UF 9493 UF 21171 UF30256UF47961UF 66322 UF 4441 UF 9497 UF 21172 UF30257UF48121UF 66324 UF 4442 UF 9498 UF 21173 UF30259UF48122UF 66352 UF 4443 UF 9499 UF 21174 UF30275UF48123UF 66353 UF 4444 UF 9500 UF 21175 UF30276UF48124UF 66354 UF 4445 UF 9501 UF 21176 UF47779UF48125UF 66423 UF 4446 UF 9502 UF 21181 UF47780UF48127UF 66426 UF 4447 UF 9704 UF 21182 UF47781UF48128UF 66591 UF 4448 UF 9708 UF 21183 UF47782UF48131UF 66593 UF 4450 UF 9710 UF 21185 UF47784UF48133UF 67752 UF 4452 UF 9711 UF 21186 UF47785UF48134UF 67756 UF 4453 UF 9731 UF 21190 UF47786UF48135UF 73785 UF 6510 UF 9732 UF 21191 UF47787UF48138UF 73786 UF 6511 UF 9734 UF 21192 UF47788UF48140UF 73787 UF 6512 UF 9735 UF 21194 UF47789UF48145UF 74704 UF 6513 UF 9736 UF 21195 UF47790UF48146UF 87651 UF 6514 UF 9738 UF 21196 UF47791UF48147UF 89498 UF 6515 UF 9739 UF 21198 UF47792UF48150UF 91092 UF 6610 UF 9740 UF 21200 UF47793UF48151UF 123198 UF 6611 UF 9742 UF 21201 UF47794UF48153UF 137227 UF 7443 UF 9757 UF 21202 UF47795UF48154UF 141802 UF 7444 UF 9787 UF 21204 UF47796UF48155UF 141876 UF 7446 UF 9815 UF 21396 UF47797UF48156UF 141880

PAGE 61

61 Table B-1. Continued UF 149147 UF 10126-2 USNM45774 USNM288236 USNM 326347 UF 149148 UF 10126-3 USNM55589 USNM288237 USNM 326348 UF 149150 UF 9492-1 USNM55590 USNM288238 USNM 326349 UF 149152 UF 9492-2 USNM55648 USNM292080 USNM 326350 UF 149153 UF 9689-1 USNM60898 USNM292081 USNM 326351 UF 149154 UF 9756-1 USNM60899 USNM304355 USNM 326352 UF 149156 UF 9756-2 USNM60900 USNM323057 USNM 326353 UF 149157 UF 7535 USNM60901 USNM323058 USNM 326354 UF 149158 UF 7537 USNM64600 USNM323073 USNM 326355 UF 149159 UF 7804 USNM64989 USNM323074 USNM 326356 UF 149160 UF 9832 USNM69547 USNM326204 USNM 326357 UF 149161 UF 9872 USNM79370 USNM326285 USNM 326358 UF 149162 UF 9881 USNM81032 USNM326294 USNM 326359 UF 149163 UF 21397 USNM94375 USNM326295 USNM 326360 UF 149164 UF 21398 USNM95329 USNM326297 USNM 326361 UF 149165 UF 21399 USNM95331 USNM326298 USNM 326362 UF 149166 UF 47798 USNM95332 USNM326299 USNM 326363 UF 149167 UF 47799 USNM95333 USNM326301 USNM 326364 UF 149168 UF 47866 USNM95335 USNM326319 USNM 326365 UF 149169 UF 48157 USNM95337 USNM326320 USNM 326366 UF 149170 UF 48158 USNM95338 USNM326321 USNM 326367 UF 149171 UF 48160 USNM95346 USNM326322 USNM 326368 UF 149173 UF 141890 USNM95348 USNM326324 USNM 326369 UF 149174 UF 141912 USNM95349 USNM326325 USNM 326370 UF 149175 UF 149146 USNM95353 USNM326327 USNM 326371 UF 149176 USNM 53 USNM95357 USNM326328 USNM 326388 UF 149177 USNM 11613 USNM95358 USNM326330 USNM 327978 UF 149178 USNM 19481 USNM99841 USNM326331 USNM 328081 UF 149179 USNM 22340 USNM100359 USNM326332 USNM 336217 UF 149180 USNM 22502 USNM100519 USNM326333 USNM 497306 UF 149181 USNM 22681 USNM101060 USNM326334 USNM 83992 UF 149182 USNM 28436 USNM118169 USNM326335 USNM 84447 UF 149183 USNM 29211 USNM142098 USNM326336 USNM 84448 UF 149184 USNM 45302 USNM142099 USNM326337 USNM 84450 UF 149197 USNM 45303 USNM142100 USNM326339 USNM 84451 UF 151190 USNM 45308 USNM197487 USNM326340 USNM 84452 UF 151365 USNM 45317 USNM197490 USNM326341 USNM 84457 UF 151368 USNM 45322 USNM218797 USNM326342 USNM 84882 UF 151372 USNM 45338 USNM218798 USNM326343 USNM 86443 UF 151373 USNM 45342 USNM220293 USNM326344 USNM 91226 UF 151375 USNM 45343 USNM288178 USNM326345 USNM 92019 UF 10126-1 USNM 45772 USNM288235 USNM326346 USNM 94372 KU=University of Kansas Natural History Muse um, UF=Florida Museum of Natural History, USNM=Smithsonian Institution National Museum of Natural History.

PAGE 62

62 Table B-2. Localities of tissue sp ecimens used in this study id St County Lat Long tcar001 FL Liberty 30.3877 -84.6714 tcar002 FL Leon 30.4371 -84.5119 tcar003 FL Liberty 30.1702 -85.0660 tcar004 FL Liberty 30.3740 -84.8058 tcar005 FL Liberty 30.3740 -84.8058 tcar006 FL Liberty 30.1101 -85.0905 tcar007 FL Liberty 30.3492 -84.8324 tcar008 FL Leon 30.4308 -84.2580 tcar009 FL Leon 30.4445 -84.2929 tcar010 FL Leon 30.4631 -84.3253 tcar011 FL Liberty 30.3739 -84.8059 tcar012 FL Liberty 30.2020 -84.9104 tcar013 FL Liberty 30.3693 -83.8118 tcar014 FL Liberty 30.3872 -84.6621 tcar015 FL Leon 30.4584 -84.3235 tcar016 FL Leon 30.5275 -84.3572 tcar017 FL Leon 30.5275 -84.3572 tcar018 FL Leon 30.4853 -84.3098 tcar019 FL Leon 30.4324 -84.3605 tcar020 FL Liberty 30.2451 -84.5272 tcar021 FL Leon 30.4584 -84.3244 tcar022 FL Jefferson 30.4730 -84.0215 tcar023 FL Leon 30.5275 -84.3572 tcar024 FL Leon 30.4929 -84.3207 tcar025 FL Leon 30.5275 -84.3572 tcar026 FL Leon 30.4584 -84.3235 tcar027 FL Leon 30.4937 -84.3121 tcar028 FL Leon 30.4584 -84.3235 tcar029 FL Leon 30.4955 -84.3256 tcar030 FL Leon 30.5007 -84.3283 tcar031 FL Jefferson 30.4306 -84.0197 tcar032 FL Leon 30.4554 -84.3331 tcar033 FL Gadsden 30.6384 -84.3619 tcar034 FL Gadsden 30.6066 -84.3639 tcar035 GA Decatur 30.7725 -84.4833 tcar036 GA Clay 31.6151 -85.0479 tcar037 FL Leon 30.4959 -84.3257 tcar038 FL Liberty 30.0417 -85.0175 tcar039 FL Liberty 30.1609 -84.9589 tcar040 FL Leon 30.4793 -84.3633 tcar041 FL Franklin 29.6261 -84.9327 tcar042 FL Leon 30.4092 -84.3053 tcar043 FL Leon 30.4720 -84.4051 tcar044 FL Leon 30.4603 -84.3137

PAGE 63

63 Table B-2. Continued id St County Lat Long tcar045 FL Leon 30.5653 -84.3586 tcar046 FL Leon 30.4841 -84.3085 tcar047 FL Leon 30.5421 -84.2109 tcar048 FL Leon 30.5233 -84.3534 tcar049 FL Leon 30.4513 -84.3229 tcar050 FL Leon 30.4880 -84.3089 tcar051 FL Leon 30.4622 -84.3239 tcar052 FL Leon 30.4926 -84.3199 tcar053 FL Walton 30.4736 -85.9809 tcar054 FL Leon 30.5310 -84.3609 tcar055 FL Walton 30.4703 -85.9582 tcar056 FL Walton 30.4747 -85.9988 tcar057 FL Walton 30.6062 -85.9417 tcar059 FL Walton 30.4537 -86.0426 tcar060 FL Gadsden 30.5863 -84.7140 tcar061 FL calhoun 30.4354 -85.3718 tcar062 FL Walton 30.6601 -85.9366 tcar063 FL Walton 30.4770 -86.0553 tcar064 FL Walton 30.5555 -85.9599 tcar065 FL Walton 30.4755 -86.0131 tcar066 FL calhoun 30.4432 -85.0314 tcar067 FL Walton 30.4665 -85.9449 tcar100 FL Alachua 29.7060 -82.3990 tcar101 FL Alachua 29.7060 -82.3990 tcar102 FL "panhandle" tcar103 KY Breckenridge 37.6614 -86.5994 tcar104 KY Breckenridge 37.6614 -86.5994 tcar105 KY Breckenridge 37.6614 -86.5994 tcar106 FL St. Johns 30.0541 -81.5153 tcar107 FL St. Johns 30.0541 -81.5153 tcar108 FL Okaloosae 30.6246 -86.7667 tcar109 FL Okaloosae 30.5263 -86.4976 tcar110 FL Gulf 29.8524 -85.3357 tcar111 FL Holmes 30.9643 -85.9666 tcar112 FL Hernando 28.5930 -82.3708 tcar113 FL Escambia 30.5696 -87.3996 tcar114 FL Hendry 26.5963 -80.9826 tcar115 FL Gulf 29.8140 -85.2957 tcar116 FL Gulf 29.8140 -85.2957 tcar117 KY Todd 36.6467 -87.1726 tcar118 NC Chatham 35.5907 -78.9109 tcar120 KY Hart 37.2403 -86.0087 tcar121 KY Hart 37.2464 -85.9961 tcar122 KY Lyon 37.0806 -88.0737

PAGE 64

64 Table B-2. Continued id St County Lat Long tcar123 NC Avery 35.8589 -82.0385 tcar124 MS Scott 32.3637 -89.5646 tcar125 VA Shenandoah 38.9376 -78.4523 tcar126 VA Surry 37.0333 -76.9574 tcar127 NC Craven 34.9324 -77.0738 tcar128 AL Sumter 32.8160 -88.1660 tcar129 NJ Cumberland 39.2891 -74.7408 tcar130 VA Prince Edward 37.2131 -78.4438 tcar131 VA Prince Edward 37.1471 -78.3327 tcar132 IL Johnson 37.4056 -88.9516 tcar133 KS Linn 38.2626 -94.7029 tcar134 KS Linn 38.2626 -94.7029 tcar135 MO Bates 38.2560 -94.5330 tcar136 MO Boone 38.8763 -92.2544 tcar137 KS Cowley 37.1071 -96.5719 tcar138 KS Chatauqua 37.0018 -96.3030 tcar139 OK Cherokee 35.9785 -95.1369 tcar140 OK McCurtain 33.8449 -94.7916 tcar141 FL Franklin 29.9358 -84.3422 tcar142 KS Crawford 37.6463 -94.8250 tcar143 KS Crawford 37.6446 -94.8134 tcar144 KS Bourbon 38.0084 -94.7670 tcar145 FL Gulf 29.9016 -85.0887 tcar146 FL Gulf 29.8714 -85.2319 tcar147 FL Franklin 29.7840 -84.9106 tcar148 FL Gulf 29.8064 -85.2918 tcar149 FL Gulf 29.8912 -85.0720 tcar150 FL Calhoun 30.5825 -84.9913 tcar151 FL Calhoun 30.4982 -85.0420 tcar152 FL Calhoun 30.2848 -85.2303 tcar153 FL Walton 30.6969 -85.9935 tcar154 FL Alachua 29.6813 -82.3533 tcar155 VA Floyd 37.1046 -80.1298 tcar156 FL Liberty 30.0588 -85.0200 tcar157 FL Jefferson 30.1901 -84.0548

PAGE 65

65 APPENDIX C MTDNA HAPOTYPES AND MICROSATELLITE ALLELES

PAGE 66

66 Table C-1. Variable base pair localities in mtDNA haplotypes Base Pair Location HAP 3 6 22 46 67778084879296104 120 131 132 138 169 193 215 254 1 G C T T A A A C C A T G T G A A C A 2 C T C G T T A C T G T G 3 C T . G T T T A C T G T G 4 C T . G T T A C T G T G 5 T . . T C G G T 6 T . G T . G G T 7 T . G T C G G T 8 T . G T . G G T 9 T . G T . G T 10 T . G T C G G T 11 T . G T C G G T 12 T . G T C G T 13 T . G T . G G T 14 T . G T C G G T 15 T . G T . G G T 16 T . G T C G G T 17 T . G T . G G T 18 T . G T C G G T 19 T . G T C G G T 20 T . G T C G G T 21 T . G T C G G T 22 T . G T . G G T 23 T . G T C G G T 24 T C G G . G 25 T . G T T . G 26 T . G T T . G 27 T . G T T . G 28 T . G G T T . G

PAGE 67

67 Table C-1. Continued Base Pair Location HAP 3 6 22 4667778084879296104120 131132138169193215254 29 T . G T T . G 30 T . G T T . G 31 A . G T T . G T 32 A . G G T T T . G 33 A . G T T . G 34 A . G T T . G 35 A . G G T T T . G

PAGE 68

68 Table C-1. Continued Base Pair Location HAP 274 277 286331332333334336337345372396 397400412417419420421 1 A A C C T T A T A T A G C G G T T G A 2 T T C C . A C C A 3 T C . C A C C A 4 T T C C . A C C A 5 G . G C A T A A C C A G 6 . G C A T A A C C A G 7 G . G C A T A A C G 8 . G C A T A A C C A G 9 G . G C A T A A C C A G 10 G . G C A T A A C C A G 11 . G C A T A A C C A G 12 G . G C A T A A C C A G 13 . G C A T A A C C A G 14 . G C A T A A C C A G 15 . G C A T A A C C A G 16 . G C A T A A C C A G 17 G . G C A T A A C C A G 18 G . G C A T A A C C A G 19 . . C A T A A C C A G 20 G . C C A T A A C C A G 21 G . G C A T A A C C A G 22 . G C A T A A C C A G 23 . G C A T A A C C A G 24 . . C A T . A G 25 . C A T A C A G 26 . C A T A C A G 27 . C A T A C C A G 28 . C A T A A C A G 29 G . C A T A C A G

PAGE 69

69 Table C-1. Continued Base Pair Location HAP 274 277 286331332333334336337345372396 397400412417419420421 30 . C A T A C A G 31 . . C G A T C A 32 . A G C G A T . A 33 . . G C G A T A C A 34 . . G C G A T C A 35 . . G C G A T . A

PAGE 70

70 Table C-1. Continued Base Pair Location HAP 431 454 460 462463470472482491509510 511519520521522526527 1 G G T T T T G T C T G A A T G T T A 2 A A C T . A C 3 A A C T . A C 4 A A C T . A C 5 A C . C A G . 6 A C C C A . 7 A C A C A . 8 A C C . A . 9 A C C . A . 10 A C C C A . 11 A C C C A . 12 A C A C A C . 13 A . C C A . 14 A . C C A . 15 A . . C A . 16 A . . C A G C . 17 A C C C A . 18 A C C C A . 19 A C C C A . 20 A C . C A . 21 A C . C A . 22 A C C . A . 23 A C C C A . 24 . . A . 25 A C . C A C . 26 A C . C A C . 27 A C . C A C . 28 A C . C A C . 29 A C . C A C .

PAGE 71

71 Table C-1. Continued Base Pair Location HAP 431 454 460462463470472482491509510 511519520521522526527 30 A C . C A . 31 A . . G A 32 A . . G . G 33 A . . G A 34 A . A . G A 35 A . . G .

PAGE 72

72 A 0 2 4 6 8 10 12 14 16122123127128130131132135136139140141144146148149150152153154158162163 Allele sizeFrequenc y B 0 2 4 6 8 10 12 14 16170174178179181182183186187190191194195198200202203206207210211212214216221 Allele sizeFrequenc y C 0 5 10 15 20 25 30 35 40 45149150154158162166170182186190 Allele sizeFrequenc y Figure C-1. Microsatelli te allele frequencies. A) Gm uD121, B) GmuD55, C) GmuD21, D) GmuB08 and E) GmuB12.

PAGE 73

73 D 0 5 10 15 20 25 30 35189192193195198201202205208 Allele sizeFrequenc y E 0 10 20 30 40 50 60177179180183185186188191 Allele sizeFrequenc y Figure C-1. Continued

PAGE 74

74 LIST OF REFERENCES Am ato ML, Brooks RJ, Fu J (2008) A phylogeogra phic analysis of popul ations of the Wood Turtle (Glyptemys insculpta) throughout its range. Molecular Ecology 17, 570-581. Aresco MJ (2005) Mitigation measures to reduc e highway mortality of turtles and other herpetofauna at a north Florida lake. Journal of Wildlife Management 69, 549-560. Atchley WR, Gaskins CT, Anderson D (1976) Statis tical Properties of Ratios. I. Empirical Results. Systematic Zoology, 25, 137-148. Auffenberg W (1967) Further notes on fossil box turtles of Florida. Copeia 2, 319-325. Austin JD, Lougheed SC, Boag PT (2004) Discor dant temporal and ge ographic patterns in maternal lineages of eastern north American frogs, Rana catesbeiana (Ranidae) and Pseudacris crucifer (Hylidae). Molecular Phylogenetics and Evolution 7, 371-379. Beamer DA, Lamb T (2008) Dusky salamanders ( Desmognathus Plethodontidae) from the Coastal plains multiple independent lineages and their bearing on the molecular phylogeny of the genus. Molecular Phylogenetics and Evolution 47, 143-153. Bermingham E, Avise JC (1986) Molecular zoogeography of freshwater fishes in the southeastern United States. Genetics 113, 939-965. Braun EL (1947) Development of the deciduous forests of eastern North America. Ecological Monographs, 17, 211-219. Burbrink FT (2001) Systematics of the eastern rat snake complex ( Elaphe obsoleta ). Herpetological Monographs 15, 1-53. Burbrink FT (2002) Phylogeographic analysis of the corn snake (Elaphe guttata ) complex as inferred from maximum likeli hood and Bayesian analyses. Molecular Phylogenetics and Evolution 25 465-476. Burbrink FT, Fontanella F, Pyron RA, Guiher TJ, Jimenez C (2008) Phylogeography across a continent: the evolutionary and demographic hist ory of the North American Racer (Serpentes: Colubridae: Coluber constrictor ). Molecular Phylogeneti cs and Evolution 47, 274-288. Burbrink FT, Lawson R, Slowinski JB ( 2000) Mitochondrial DNA phylo geography of the polytypic north American rat snake ( Elaphe obsoleta ): a critique of the subspecies concept. Evolution 54, 2107. Caro TM, ODoherty G (1999) On the use of surrogate species in conservation biology. Conservation Biology 13 805-814 Carr AF (1952) Handbook of Turtles: the Turtles of the United States, Canada and Baja California Cornel University Pr ess, Ithaca, New York.

PAGE 75

75 Church SA, Kraus JM, Miche ll JC, Church DR, Taylor DR (2003) Evidence for multiple Pleistocene refugia in the postglacial e xpansion of the eastern tiger salamander, Ambystoma tigrinum tigrinum Evolution 57, 372-383. Clark AM, Bowen BW, Branch LC (1999) Effect s of natural habitat fragmentation on an endemic scrub lizard (Sceloperus woodi): an historical perspec tive based on a mitochondrial DNA gene genealogy. Molecular Ecology 8, 1093. Conant R, Collins JT (1991) A Field Guide to the Reptiles and Amphibians of Eastern and Central North America 3rd ed. Houghton Miffli n, Boston, Massachusetts. Crandall KA, Bininda-Emonds ORP, Mace GM, Wa yne RK (2000) Considering evolutionary process in conservation biology. Trends in Ecology and Evolution 15,290-295. Davis LM, Glenn TC, Stickland DC, Guillette LJ Elsey RM, Rhodes WE, Dessauer HC, Sawyer RH (2002) Microsatellite DNA analyses s upport and east-west phylogeographic split of American alligator populations. Journal of Experimental Zoology, 294, 352. Devitt TJ, LaDuc TJ, McGuire JA (2008) The Trimorphodon biscutatus (Squamata: Colubridae) species complex revisi ted: a multivariate st atistical analysis of geographic variation. Copeia 2, 370-387. Ditmars RL (1934) A Review of the box turtles. Zoologica (Proceedings of the New York Zoological Society) 17, 1-44. Dodd Jr CK (2001) North American Box Turtles. A Natural History. University of Oklahoma Press, Norman, Oklahoma. Dodd Jr CK, Franz R (1993) The need for status information on common herpetofaunal species. Herpetological Review 24, 47-50. Dowsett HJ, Cronin TM (1990) High eustatic sea level during th e middle Pliocene: evidence from the southeastern U.S. Atlantic coastal plain. Geology 18, 435-438. Ellsworth DL Honeycutt RL, Silvy NJ, Bick man JW, Klimstra WD (1994) Historical biogeography and contemporary patterns of mitochondrial DNA variatio n in white-tailed deer from the southeastern United States. Evolution, 48, 122-136. Elmer KR, Davila JA, Lougheed SC (2007) Cryptic diversity and deep divergence in an upper Amaxonian leaflitter frog, Eleutherodactylus ockendeni. BMC Evolutionary Biology 7, 247. Endler JA (1977) Geographic Variation, Speciation and Clines Princeton University Press, Princeton, New Jersey. Epperson BK (2003) Geographical Genetics Princeton University Press. Princeton, New Jersey.

PAGE 76

76 Fontanella FM, Feldman CR, Siddall ME Burbrink FT (2008) Phylogeography of Diadophis punctatus: Extensive lineage diversity and repeated pa tterns of historical demography in a transcontinental snake. Molecular Phylogenetics and Evolution 46, 1049-1070. Fritz U, Havas P (2007) Checklist of Chel onians of the World. Vertebrate Zoology, 57, 149-368. Gandolfo MA, Nixon KC, Crepet WL (2008) Selection of fossils for ca libration of molecular dating models. Annals of the Missouri Botanical Garden 95, 34-42. Gilbert CA (1987) Zoogeography of the freshwat er fish fauna of s outhern Georgia and peninsular Florida. Brimleyana, 13, 25-54. Goldstein DB, Schlotterer C (1999) Microsatelites: Evol ution and Application Oxford University Press, Oxford, UK. Harrison RG (1990) Hybrid Zones: windows on evolutionary processes. Oxford Surveys in Evolutionary Biology, 7, 69-128. Hoehn M, Sarre SD, Henle K (2007) The tales of two geckos: does dispersal prevent extinction in recently fragmented populations? Molecular Ecology 16, 3299. Hulbert Jr. RC (2001) The Fossil Vertebrates of Florida 1st edn. University Press of Florida, Gainesville, Florida. Huxley JS (1938) Clines: An Auxiliary Taxonomic Principle. Nature, 142 219-220. Iglay RB, Bowman JL, Nazdrowi cz NH (2007) Eastern box turtle ( Terrapene carolina carolina ) movements in a fragmented landscape. Journal of Herpetology 41, 102-106. Iverson JB, McCord WP (1994) Variation in east Asian turtles of the genus Mauremys (Bataguridae, Testudines). Journal of Herpetology 28, 178. Kiester AR, Schwartz CW, Scwart z ER (1982) Promotion of gene flow by transient individuals in an otherwise sedentary population of box turtles ( Terrapene carolina triunguis ). Evolution 36, 617-619. King PB, Beikman HM (1974) Geologic map of th e United States (exclusive of Alaska and Hawaii). Reston, Va., U.S. Geologica l Survey, 3 sheets, scale 1:2,500,000. King TL, Julian SE (2004) Conservation of mi crosatellite DNA flanking sequence across 13 Emydid genera assayed with novel bog turtle ( Glyptemys muhlenbergii ) loci. Conservation Genetics, 5, 719-725. Kuo CH, Ja nzen FJ (2004) Geneti c effects of a persistent bottl eneck on a natural population of ornate box turtle (Terrapene ornata ). Conservation Genetics, 5, 425-437.

PAGE 77

77 Lamb T, Lydeard C, Walker RB, Gibbons JW (1994) Molecular systematics of map turtles (Graptemys): a comparison of mitochondr ial restriction site versus sequence data. Systematic Biology, 43, 543-559. Leache AD, Reeder TW (2002) Molecular sy stematics of the eastern fence lizard ( Sceloperus undulatus): a comparison of parsimony, like lihood and Bayesian approaches. Systematic Biology, 51, 44-68. Liu H, Platt SG, Borg CK (2004) Seed dispersal by the Florida box turtle ( Terrapene carolina bauri ) in pine rockland forests of the Florida keys, United States. Oecologia 138, 539-546. MacAvoy ES, McGibbon LM, Sainsbury JP, Lawrence H, Wilson CA, Daugherty CH, Chambers GK (2007) Genetic variation in island populations of tuatara ( Sphenodon spp.) inferred from microsatellite markers. Conservation Genetics 8, 305-318. Maskas SD, Cruzan MB (2000) Patterns of intraspecfic dive rsification in the Piriqueta caroliniana complex in southeastern No rth America and the Bahamas. Evolution 54, 815-827. Mayden, RL (1988) Vicariance biogeography, pa rsimony and evolution in North American freshwater fishes. Systematic Zoology 37, 329-355. McCoid MJ (1992) Geographic distribution: Terrapene carolina triungus Herpetological Review 23, 25. McGarigal K, Cushman S, Stafford SG (2000) Multivariate statistics for Wildlife and Ecology Research. Springer-Veriag, New York, New York. Means DB, Krysko KL (2001) Biogeography an d pattern variation of kingsnakes, Lampropeltis getula, in the Apalachicola region of Florida. Contemporary Herpetology, 5, 1-33. Miller MP (2005) Alleles In Sp ace (AIS): Computer software for the joint analysis of interindividual spatial and genetic information. Journal of Heredity 96, 722-724. Milstead WW (1969) Studies on the Evolution of Box Turtles (Genus Terrapene ). Bulletin of the Florida State Museum, 14, 1-113. Minx P (1996) Phylogenetic Relationsh ips among the Box Turtles, Genus Terrapene. Herpetologica 52, 584-597. Nichols JT (1939) Range and hom in g of individual box turtles. Copeia 1939, 125-127. Oosterhout C, Hutchinson WF, Wills DPM, Shipley P (2004) Micro-checker: software for identifying and correcting genotyping e rrors in microsatellite datasets. Molecular Ecology Notes 4, 535.

PAGE 78

78 Paquette SR, Lapointe FJ (2007) The use of shell morpometrics for the management of the endangered Malagasy radiated tortoise ( Geochelone radiata ). Biological Conservation 134, 3139. Pauley GB, Piskurek O, Shaffe r HB. (2007) Phylogeographic conc ordance in the southeastern United States: the flatwoods salamander, Ambystoma cingulatum as a test case. Molecular Ecology, 16, 415-429. Posada D, Crandall KA (2001) Intraspecific gene genealogies: trees gr afting into networks. Trends in Ecology and Evolution 16, 37-45. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics, 155 945-959. Raymond, M and Rousett F (1995) An exact test for population differentiation. Evolution 49, 1280-1283. Rieseberg LH, Archer MA, Wayne RK (1999) Transgressive segr egation, adaptation and speciation. Heredity 83, 363-372. Reist JD (1986) An empirical evaluation of co efficients used in residual and allometric adjustment of size covariation. Canadian Journal of Zoology 64, 1363-1368. Rosenbaum PA, Robertson JM, Zamudio KR ( 2007) Unexpectedly low genetic divergences among populations of the th reatened bog turtle (Glyptemys muhlenbergii ). Conservation Genetics, 8, 331. Rousett F (2000) Genetic differe ntiation between individuals. Journal of Evolutionary Biology, 13, 58. Sambrook J, Russell DW (2001) Molecular cloning: a laboratory manual 3rd edn. Cold Springs Harbor Laboratory Press, New York, New York. Schwartz CW, Schwartz ER (1974) The threetoed box turtle in central Missouri: Its population, home range and movements. Missouri Department of Conservation Terrestrial Series 5, 1-28. Smith PW (1957) An analysis of post-Wisconsin biogeography of the prairie peninsula region based on distributional phenomena among terrestrial vertebrate populations. Ecology, 38, 205218. Soltis DE, Morris AB, McLachlan JS, Manos PS, Soltis PS (2006) Comparative phylogeography of unglaciated eastern North America. Molecular Ecology 16, 42614293.

PAGE 79

79 Starkey DE, Shaffer HB, Burke RL, Forstner MR J, Iverson JB, Janzen FJ, Rhodin AGJ, Ultsch GR (2003) Molecular systematics, phylogeography a nd the effects of Plei stocene glaciation in the painted turtle ( Chrysemys picta ) complex. Evolution, 57, 119-128. Stein BA, Kutner SL, Adams JS (2000) Precious heritage: the status of biodiversity in the United States Oxford University Press, Oxford, UK. Swenson NG, Howard DJ (2005) Clustering of contact zones, hybrid zones, and phylogeographic breaks in North America. American Naturalist, 166, 581. Taylor WE (1895) The box turtles of North America. Proceedings of the Un ited States National Museum 17, 573-588. Telles MPD, Deniz-Filho JAF, Bastos RP, Soares TN, Guimaraes LD, Lima LP (2007) Landscape genetics of Physalaemus cuvier i in Brazilian Cerrado: correspondence between population structure and patterns of human occupation and habitat loss. Biological Conservation 139, 37-46. Walker D, Avise JC (1998) Principles of phylogeography as illustrated by freshwater and terrestrial turtles in th e southeastern United States. Annual Review of Ecology and Systematics, 29, 23-58. Ward JP (1980) Comparative cranial morphologoy of the fr eshwater turtle subfamily Emydinae: An analysis of the feedin g mechanisms and systematics Ph.D. dissertation, North Carolina State University, Ralei gh, North Carolina. Wilson AC, Cann RL, Carr SM, George M, Gyllensten UB, Helm-Bychowski KM, Higuchi RG, Palumbi SR, Prager EM, Sage RD, Stoneking M (1985) Mitochondrial DNA and two perspectives on evolutionary genetics. Biological Journal of the Linnean Society, 26, 375400. Wolfe SH, Reidenauer JA, Means DB (1988) An ecological characterization of the Florida Panhandle. U.S. Fish and Wildlife Service Biolog ical Report, Washi ngton, District of Columbia. Yezerinac SM, Lougheed SC, Handford P (1992) M easurement error and morphometric studies: an assessment of statistical power and the effect of observer experience using avian skeletons. Systematic Biology 41, 471-482.

PAGE 80

80 BIOGRAPHICAL SKETCH Jason Michael Butler was born in 1981, in Harford Kentucky. The oldest of four children, he grew up mostly in Philpot, Kentucky. He earned his B.S. in biology at Western Kentucky University in 2005. He began working toward a M.S. in wildlife ecol ogy and conservation in 2006. After completing his masters program, Jason will return to Kentucky to pursue a career in conservation and ecological research. Jason has b een married to Sara N. Moore Butler for two years.