University of Texas at Tyler Scholar Works at UT Tyler

Biology Theses Biology

Spring 4-30-2012 Molecular Phylogenetics and Phylogeography of the American Box (Terrapene SPP.) Bradley T. Martin

Follow this and additional works at: https://scholarworks.uttyler.edu/biology_grad Part of the Biology Commons

Recommended Citation Martin, Bradley T., "Molecular Phylogenetics and Phylogeography of the American Box Turtles (Terrapene SPP.)" (2012). Biology Theses. Paper 15. http://hdl.handle.net/10950/74

This Thesis is brought to you for free and open access by the Biology at Scholar Works at UT Tyler. It has been accepted for inclusion in Biology Theses by an authorized administrator of Scholar Works at UT Tyler. For more information, please contact [email protected].

MOLECULAR PHYLOGENETICS AND PHYLOGEOGRAPHY OF THE

AMERICAN BOX TURTLES (TERRAPENE SPP.)

by

BRADLEY T. MARTIN

A thesis/dissertation submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Biology

John S. Placyk, Jr., Ph.D., Committee Chair

College of Arts and Sciences

The University of Texas at Tyler May 2012

The University of Texas at Tyler Tyler, TX This is to certifu that the Master's Thesis/Doctoral Dissertation of

BRADLEY T. MARTIN

has been approved for the thesisidissertation requirement on April4,2012 for the Master of Science degree

Approvals:

Thesis/Dissertation C

'ember: James

Chair, Department tlbhrS'o.- D.un r Acknowledgements:

I wish to give special thanks to my for supporting and encouraging me in my academic and life pursuits. It is without a doubt that I would not have made it this far without their support. My brother, Charley Martin, has always been and always will be my best friend. He has kept me sane throughout my academic career by being an amazing person to talk to and an exceptional brother and friend. My parents, Ben and

Karla Martin, are the kindest and most understanding, supportive, encouraging, and loving parents imaginable. I also wish to give special thanks to my advisor, John S.

Placyk, Jr., for taking me on as his graduate student and always being there when I needed him. Dr. Placyk is an exceptional advisor and person, and he has always been someone I could talk to, whether it be academic or personal matters. He has put forth more than his fair share of effort to support me in my academic endeavors, and I will always consider him a close friend. I wish to give special thanks to the undergraduate and graduate students with whom I have become close friends with during my tenure at the UT-Tyler. Their friendship and support means the world to me. I would like to give special thanks to my committee members for their advice, support, and time spent working with me on my M.S. thesis. Finally, this project would not have been possible without the countless volunteers, institutions, universities, wildlife organizations, and museums throughout much of the United States who kindly provided tissue samples for me so that the scientific community can better understand the genetics of the American box turtles and protect them from further population declines. Table of Contents

List of Tables ...... iii

List of Figures ...... v

Abstract ...... vi

Chapter One ...... 1 Introduction ...... 1 Current classification...... 2 Problems with the current Terrapene classification ...... 3 What is a ? ...... 4 Conservation implications from resolving Terrapene classification ...... 5 Importance of mitochondrial and nuclear DNA in phylogenetic analyses . 7 Phylogeography ...... 8 Summary ...... 9 Objectives ...... 10 Literature Cited ...... 10

Chapter Two ...... 16 Introduction ...... 16 Materials and Methods ...... 19 DNA extractions, PCR, and sequencing ...... 20 Sequence analysis and phylogenetic inference ...... 22 Combined data ...... 24 Haplotype networks ...... 24 AMOVA and SAMOVA analyses ...... 25 A priori AMOVAs ...... 26 A posteriori AMOVAs ...... 27 Cytb molecular clock analysis ...... 28 Results ...... 29 Cytb phylogenetic analysis ...... 29 T. c. carolina - T. c. triunguis and their associated taxa ...... 30 T. c. major - T. coahuila ...... 31 Cytb western clade ...... 31 GAPD phylogenetic analysis ...... 32 T. c. carolina - T. c. triunguis ...... 32 i Taxa associated with T. c. carolina ...... 33 GAPD western clade ...... 33 Combined mtDNA and nucDNA phylogeny ...... 33 Haplotype networks ...... 34 Population structure ...... 35 Molecular clock analysis ...... 35 Discussion ...... 36 Phylogenetic analyses ...... 36 Eastern clade ...... 37 Western clade ...... 37 Combined mtDNA and nucDNA data ...... 38 Polyphyly...... 39 Haplotype networks ...... 41 Population structure ...... 42 T. carolina population structure ...... 42 SAMOVA analyses ...... 43 Molecular clock ...... 44 T. carolina ...... 45 T. c. bauri and T. c. major ...... 45 T. c. triunguis, T. c. mexicana, and T. c. yucatana ...... 46 Comparisons with published divergence estimates ...... 47 Percent divergences ...... 47 Recommendations for classification revisions ...... 48 Conservation implications ...... 50 Literature Cited ...... 51

Appendices...... 91 Appendix A: A supplementary table listing sampling information included in the analyses for this study...... 91

Appendix B: Haplotype designations for the GAPD haplotype network, as determined using the Phase algorithm in DnaSP v 5.10.01...... 99

Appendix C: Cytochrome b DNA sequence alignment matrix as aligned in Clustal X...... 101

Appendix D: Glyceraldehyde-3-phosphate dehydrogenase DNA sequence alignment matrix as aligned in Clustal X...... 132

ii List of Tables

Table 1: The four currently recognized Terrapene and their associated subspecies, based on Minx (1996)...... 62

Table 2: GenBank Accession numbers for both the cytochrome b (Cytb) and glyceraldehyde-3-phosphate dehydrogenase (GAPD) genes...... 63

Table 3: Mean divergence time estimates, standard deviation (SD), and upper and lower 95% confidence intervals (CI) for the mtDNA cytochrome b gene...... 64

Table 4: Cytochrome b AMOVA with all 10 species/subspecies apportioned separately………………………...... 65

Table 5: Cytochrome b AMOVA based on the morphological data of Minx (1996)...... 66

Table 6: Cytochrome b AMOVA with the groups apportioned into eastern and western clades...... 67

Table 7: Cytochrome b AMOVA with T. o. ornata and T. o. luteola clumped together...... 68

Table 8: Cytochrome b AMOVA based the mtDNA and nucDNA phylogenies (Figures 4-5)...... 69

Table 9: Cytochrome b AMOVA apportioned based on the mtDNA and nucDNA phylogenies (Figures 4-5), with the exception of T. coahuila being grouped with T. carolina...... 70

Table 10: Cytochrome b AMOVA apportioned based on the mtDNA and nucDNA phylogenies (Figures 4-5), with the exception of T. c. bauri being placed into a unique group...... 71

Table 11: Glyceraldehyde-3-phosphate dehydrogenase AMOVA with all 10 species/subspecies separated into distinct groups...... 72

Table 12: Glyceraldehyde-3-phosphate dehydrogenase AMOVA based on morphological data of Minx (1996)...... 73

iii Table 13: Glyceraldehyde-3-phosphate dehydrogenase AMOVA with the groups apportioned into eastern and western clades...... 74

Table 14: Glyceraldehyde-3-phosphate dehydrogenase AMOVA with T. o. ornata and T. o. luteola combined into one group...... 75

Table 15: Glyceraldehyde-3-phosphate dehydrogenase AMOVA based on the mtDNA and nucDNA phylogenies (Figures ...... 76

Table 16: Glyceraldehyde-3-phosphate dehydrogenase AMOVA apportioned based on the mtDNA and nucDNA phylogenies (Figures 4-5), with the exception of T. coahuila being grouped with T. carolina...... 77

Table 17: Glyceraldehyde-3-phosphate dehydrogenase AMOVA apportioned based on the mtDNA and nucDNA phylogenies (Figures 4-5), with the exception of T. c. bauri being placed into a unique group...... 78

Table 18: Cytochrome b SAMOVA with the populations maximally differentiated into 19 groups...... 79

Table 19: Glyceraldehyde-3-phosphate dehydrogenase SAMOVA with the populations maximally differentiated into seven groups...... 80

Table 20: Percent divergences calculated using Jukes Cantor nucleotide divergences corrected for comparing populations...... 81

iv List of Figures

Figure 1: The Terrapene phylogeny of Minx (1996) based on 32 morphological characters...... 82

Figure 2: Phylogenies from previous phylogenetic studies assessing the classification of the subfamily , including the four currently recognized species of Terrapene (Gaffney and Meylan, 1988; Bickham et al., 1996; Burke et al., 1996; Feldman and Parham, 2002; Stephens and Wiens, 2003)...... 83

Figure 3: Sampling localties for the mitochondrial DNA (mtDNA) cytochrome b gene and the nuclear DNA (nucDNA) glyceraldehyde-3-phosphate dehydrogenase (GAPD) gene...... 84

Figure 4: The cytochrome b phylogram, generated using PhyML 3.0, consising of 253 sequences distributed into 104 haplotypes...... 85

Figure 5: The glyceraldehyde-3-phosphate dehydrogenase (GAPD) phylogram, generated using PhyML 3.0, consisting of 202 sequences distributed into 59 haplotypes...... 86

Figure 6: A combined mtDNA and nucDNA phylogram...... 87

Figure 7: A mtDNA cytochrome b (Cytb) chronogram in millions of years, generated in BEAST v1.6.2...... 88

Figure 8: Cytochrome b (Cytb) haplotype network with nine subgroups calculated using the 95% statistical parsimony method...... 89

Figure 9: Glyceraldehyde-3-phosphate dehydrogenase (GAPD) haplotype network with two subgroups calculated using the 95% statistical parsimony method...... 90

v Abstract

MOLECULAR PHYLOGENETICS AND PHYLOGEOGRAPHY OF THE AMERICAN BOX TURTLES (TERRAPENE SPP.)

Bradley T. Martin

Thesis/Disssertation Chair: John S. Placyk, Jr., Ph.D.

The University of Texas at Tyler May 2012

The classification of the American box turtles (Terrapene spp.) has remained enigmatic to systematists. Previous comprehensive phylogenetic studies have focused primarily on morphology. The goal of this study was to re-assess the classification of

Terrapene spp. by obtaining DNA sequence data from a broad geographic range and from all four species and 11 subspecies within the .

Tissue samples were obtained for all taxa except for T. nelsoni klauberi. DNA was extracted, and the mitochondrial DNA (mtDNA) cytochrome b (Cytb) and nuclear

DNA (nucDNA) glyceraldehyde-3-phosphate-dehydrogenase (GAPD) genes were amplified via polymerase chain reaction and sequenced. The sequence data were analyzed using maximum likelihood and Bayesian phylogenetic inference, a molecular clock, AMOVAs, SAMOVAs, haplotype networks, and pairwise percent divergence comparisons.

vi T. c. triunguis was paraphyletic to T. carolina and T. ornata ornata and T. o. luteola lacked distinction phylogenetically. T. nelsoni was confirmed to be the sister taxon of T. ornata, and T. c. major, T. c. bauri, and T. coahuila were not well resolved but were closely associated with T. c. carolina. T. c. mexicana and T. c. yucatana were closely associated with T. c. triunguis.

The results suggest that T. c. triunguis should be elevated to species status (T. mexicana), and mexicana and yucatana should be included in this group as subspecies.

In addition, T. o. ornata and T. o. luteola should not be considered separate subspecies.

Because conservation efforts are typically species-based, these results will be important for facilitating successful conservation management strategies.

vii

Chapter One

Literature Review

Introduction

The American box turtles (Terrapene spp.) are members of the family of Testudines, which is divided into two subfamilies: 1) Emydinae (New World Pond turtles; Terrapene, , , and Clemmys) and 2) (Chrysemys,

Deirochelys, , Malaclemys, Psuedemys, and ). Emydid turtles are characterized by having large plastrons, carapaces and plastrons that are usually contiguous, small skulls, plastral hinges in some genera (e.g., Terrapene), and various other morphological characters (Milstead and Tinkle, 1967; Ernst and Barbour, 1989;

Minx, 1992; Minx, 1996). In addition, many genera are aquatic or semi-aquatic (except for most of Terrapene), and as such have streamlined shells, membranous webs between their toes at the base (or at least remnants of toe webbing), and a low shell arch, again, with the exception of Terrapene.

Regarding Terrapene specifically, the common name box originates from their ability to close their plastron via moveable hinges (plastral shell kinesis). Their carapaces are characterized by a high-domed morphology thought to 1) aid in maintaining an upright position (in case the turtle is turned over on its back) and 2) act as an anti-predator defense by making it more difficult for predators to wrap their jaw 1 around the turtle’s shell (Domokos and Varkonyi, 2008). They are primarily terrestrial, with the exception of the semi-aquatic Coahuilan (Terrapene coahuila). As seen in other terrestrial members of Testudines (e.g., many ), they are also a very long-lived group (as are many turtles), often reaching an age of 50 years in the wild and are considered K-selected species due to their slow maturity (Dodd, 2001).

Current classification

Currently, four species and 11 subspecies of Terrapene are recognized (T. carolina, T. ornata, T. nelsoni, and T. coahuila; Milstead and Tinkle, 1967; Milstead,

1969; Minx, 1992; Minx, 1996). The subspecies associated with each species, along with their countries of inhabitance, are listed in Table 1.

In the last major phylogenetic assessment of the group, Minx (1996) evaluated 32 morphological characters and generated several hypotheses regarding evolutionary relationships within the genus (Figure 1). First, he suggested that there are distinct eastern (Terrapene carolina) and western () clades. Second, within the eastern clade, he found three close associations: 1) Terrapene c. bauri - T. c. carolina, 2)

T. c. triunguis - T. c. mexicana - T. c. yucatana, and 3) T. c. major - T. coahuila. In addition, Terrapene coahuila forms the sister species to T. carolina (Auffenberg and

Milstead, 1965; Milstead, 1969; Minx, 1992; Minx, 1996; Feldman and Parham, 2002;

Stephens and Wiens, 2003; Wiens et al., 2010). Third, Terrapene o. ornata and T. o. luteola are closely associated, and form the sister species to Terrapene nelsoni (Milstead and Tinkle, 1967; Minx, 1996; Feldman and Parham, 2002; Stephens and Wiens, 2003;

2 Wiens et al., 2010). However, Minx (1996) indicated that the eastern clade “forms an unresolved trichotomy,” and he called the relationships within that particular clade into question. This lack of clarity within the eastern clade has resulted in several relationships remaining unclear, and Minx (1996) suggested that molecular phylogenetic data may help to resolve these ambiguities.

Problems with the current Terrapene classification

That the current classification of the Terrapene group is not well-resolved may be due, in part, to intergradation between sympatric subspecies, a high level of inter- and intraspecific morphological variation compared to the amount of phenotypic variation seen in many other taxa, and paraphyly within T. carolina (Carr, 1940; Carr, 1952;

Milstead, 1969; Conant and Collins, 1991; Stephens and Wiens, 2003). Intergradation makes phylogenies based on morphology more difficult to resolve because intermediate phenotypic traits are common in hybridizing subspecies. Since previous phylogenetic data for Terrapene are mostly based on morphology (Milstead and Tinkle, 1967;

Milstead, 1969; Gaffney and Meylan, 1988; Minx, 1992; Minx, 1996), molecular data may shed light on which taxa and geographic regions that intergradation is occurring.

While Terrapene have been included in several molecular phylogenetic studies (Figure 2; e.g., Bickham et al., 1996; Burke et al., 1996; Feldman and Parham, 2002; Stephens and

Wiens, 2003; Herrmann and Rosen, 2009; Spinks and Shaffer, 2009; Wiens et al., 2010;

Butler et al., 2011), none have provided insight into the evolutionary history of the group due, in part, to limited sample sizes, primarily focusing on intergeneric level classification rather than interspecific level classification, focusing on species-level

3 classification that ignore subspecies, geographically limited sampling, and missing taxa

(i.e., none include every currently recognized species and subspecies). Since the current classification scheme is primarily based on morphology, a re-assessment using molecular phylogenetic data is needed in to either support the current classification of the group or to propose a new one.

What is a subspecies?

Because Terrapene currently contains multiple subspecies, a description of a subspecies and the roles that subspecies play in and conservation must be discussed. Mayr (1963) described a subspecies as an “aggregate of local populations of a species inhabiting a geographic subdivision of the range of the species and differing taxonomically from other populations of the species.” Frankham et al. (2002) stated that subspecies are on an evolutionary trajectory towards speciation. In other words, subspecies often represent diverging evolutionary lineages, frequently due to allopatry

(albeit not exclusively). The concept of a subspecies is important because it can increase our understanding of the evolutionary relationships at the intraspecific level, and it can indicate the evolutionary potential of a group of organisms (i.e., unique populations that may be diverging). There is, however, much disagreement among biologists as to what criteria should be used to classify a taxon as a species or subspecies, which can affect the conservation listing of the respective taxon (for a review, see Haig et al., 2006). Since conservation efforts are typically species-based, whether a taxon is listed as a species or a subspecies can affect the conservation management strategies and how much funding can be procured to implement these strategies.

4

Conservation implications from resolving Terrapene classification

American box turtle populations are declining throughout their range, in part, due to habitat loss resulting from increasing urbanization, collection from the wild for pet trade, and changes in predator pressures (Dodd, 2001). The 2011 International Union for

Conservation of Nature (IUCN) Red List classifies T. carolina as Vulnerable, T. ornata as Near Threatened, T. coahuila as Endangered with a Very High Risk of Extinction, and

T. nelsoni as Data Deficient (although it was listed as Threatened on the 2006 Red List).

In the United States, the various subspecies of T. carolina and T. ornata are listed as

Species of Special Concern in NH, CT, MI, TX, and MA, Protected in IN, and State

Endangered in ME, WI, IL. In order to prevent further declines in box turtle populations, it is essential that conservation efforts be employed; however, to develop and implement successful conservation strategies, additional knowledge is needed on the evolutionary history and underlying genetics of this taxon. Specifically, state and federal agencies will need to have knowledge of the level of classification (e.g., species or subspecies) for each group in order to make informed decisions about their . Since conservation efforts are typically species-based, it is important to have an understanding of their specific and subspecific relationships because the conservation strategies may differ depending on their classification. For example, the conservation management strategies may differ if certain groups, such as T. c. carolina and T. c. triunguis, were to be classified as separate species or as subspecies. Finally, it is important to have information regarding their underlying genetics in order to prevent the mixing of divergent genetic lineages into a population if translocation was being considered. The

5 mixing of divergent genetic lineages in a population can result in further population declines and even extinction, as was the case with the dusky seaside sparrow

(Ammodromus maritimus nigriscens).

During the 1980s, in an attempt to conserve the genetic lineage of the dusky seaside sparrow, some of the few remaining representatives of this group were hybridized with Scott’s seaside sparrow (A. m. peninsulae), which was thought to be the most closely related subspecies to the dusky seaside sparrow based on morphological data.

However, Avise and Nelson (1989) later discovered via mitochondrial DNA (mtDNA) sequence analyses that Scott’s seaside sparrow was not the most closely related subspecies and that another more geographically distant, yet more genetically similar subspecies would have been a better choice for the breeding program. This failure coupled with various other factors resulted in conservation efforts that were unable to prevent the dusky seaside sparrow from going extinct. This example provides evidence that conservation efforts need to be based on sufficient knowledge of the species in question before management strategies are applied, and molecular data is essential to such strategies. In accordance with the need for sufficient data in order to make informed conservation management decisions, it is important to include both mtDNA and nuclear

DNA (nucDNA) sequence data to properly re-assess the classification of the group in question, as phylogenies based only one type of DNA can lead to inaccurate conclusions

(Wiens et al., 2010).

6 Importance of mitochondrial and nuclear DNA in phylogenetic analyses

MtDNA has been a useful marker for analyzing the phylogenetic relationships of many organisms for several reasons. First, divergence of mtDNA is commonly seen in instances of geographic separation (Avise et al., 1984a; Avise et al., 1986; Bermingham et al., 1986; Saunders et al., 1986; Hillis et al., 1996; Curole and Kocher, 1999).

Divergence patterns seen between groups for mtDNA can be highly variable over relatively short time periods when compared with nuclear DNA (nucDNA) depending on the gene sequenced, which makes mtDNA a useful indicator of population divergence and speciation events (Brown et al., 1979; Avise et al., 1987). Second, the maternal inheritance of mtDNA results in a lack of recombination and a haploid representation of the genetic information (Avise, 1994; Moore, 1995; Sunnucks, 2000; Avise, 2004), which makes analyzing mtDNA sequence data simpler than nucDNA. Although immensely useful in systematics and studies of phylogeography, mtDNA does have some limitations.

For example, introgression can give a misrepresentation of an individual or group (for a review, see Maddison, 1997; Parham et al., 2001; Stuart and Parham, 2007; Spinks and

Shaffer, 2009), and mtDNA can be subject to rapid rates of lineage extinction (Avise et al., 1984b). Differences in behavior between males and females can also potentially affect the evolution of mtDNA, as the two sexes can have different energy requirements, causing mtDNA to evolve at unequal rates based on these varying energy requirements

(FitzSimmons et al., 1997). Thus, additional data are necessary to supplement the phylogenies based on mtDNA. Wiens et al. (2010) found discordance between mtDNA and nucDNA phylogenies for Emydid turtles (including Terrapene) in which mtDNA without supplementation from other mtDNA genes and nucDNA genes may be

7 misleading; they also found discordance between different nucDNA phylogenies. Thus, it is important that both mtDNA and nucDNA data be utilized for a more complete picture.

Using both nucDNA and mtDNA for the phylogenetic analysis of a group can give a more complete picture of its evolutionary relationships and underlying genetics than using either type of data independently. The addition of a nucDNA phylogeny can also aid in overcoming some of the limitations seen in mtDNA. For example, nucDNA and mtDNA phylogenies that disagree can give an indication of hybridization for the mtDNA, or that different selection pressures are affecting diverging populations (Spinks and Shaffer, 2009). Alternatively, nucDNA and mtDNA phylogenies that agree can indicate a high level of support for the relationships that the phylogenies resolve. Like mtDNA phylogenies, nucDNA phylogenies also have limitations. For instance, it can take longer for variation to occur with nucDNA than with mtDNA (Birks and Edwards,

2002). NucDNA is also limited by recombination because alleles are inherited from both parents. However, when nuclear data are combined with mitochondrial data, they are more beneficial than being used independently (Rubinoff and Holland, 2005). Due to a generally slower rate of evolution for nucDNA, finer-scale phylogeographic approaches can more appropriately assess a nucDNA phylogeny that shows low variation.

Phylogeography

Because nucDNA in general has a slower rate of evolution than that of mtDNA in a wide variety of taxa (Prychitko and Moore, 1997; Groth and Barrowclough, 1999;

8 Prychitko and Moore, 2000; Birks and Edwards, 2002; Caccone et al., 2004; Engstrom et al., 2004; Fujita et al., 2004), nuclear phylogenies based on shorter time spans can be difficult to resolve. Accordingly, a phylogeographic approach can help elucidate intraspecific evolutionary relationships using sequences with low variability by assessing the population structure. In other words, a phylogeographic approach can indicate populations or regions with restricted gene flow and have acquired unique haplotypes

(Avise, 1992; 1994; for a review, see Avise, 2009). Spatial analyses of molecular variance and haplotype networks are useful in these situations, as they are able to segregate populations into geographic regions and illustrate which populations have unique haplotypes (Templeton, 1998; 2004). Having an understanding of the population structure of a group is important in order for conservation management strategies to be most effective, as these approaches can indicate the regions that are most phylogenetically similar. Introducing individuals to a population with the most similar genetic lineages can prevent detrimental genetic effects such as outbreeding depression

(for a review, see Edmands, 2007).

Summary

The genus Terrapene needs a comprehensive study to re-assess the current classification. Such a study will have conservation implications because, in order to most effectively develop and implement conservation management strategies, the evolutionary history and population genetics of the group must be investigated. Furthermore, a finer- scale phylogeographic study evaluating multiple populations and all of the species and subspecies will indicate the current status of their genetic diversity. A phylogeographic

9 approach can indicate isolated populations that may be in need of conservation attention.

Finally, both mtDNA and nucDNA genes need to be sequenced to appropriately assess their evolutionary history because only using one type of marker can lead to inaccurate conclusions. To date, a study that has comprehensively examined Terrapene classification with molecular phylogenetics to the extent that is needed in order to provide sufficient data to potentially facilitate successful conservation strategies has not been published.

Objectives

Given the need for additional molecular data in order to better understand the evolutionary history and underlying genetics of Terrapene, the goals of this study are to

1) re-assess the classification of the entire genus by including mitochondrial and nuclear

DNA sequence data from every representative taxon, and 2) include a wide geographic sampling in order to have a better understanding of the population structure and phylogeography of the Terrapene.

Literature Cited

Auffenberg, W. W. and W. W. Milstead. 1965. in the Quarternary of North America. In The Quarternary in the United States, (eds. H. E. Wright, Jr. and D. G. Frey). Princeton Univ. Press, Princeton, New Jersey.

Avise, J. C. 1992. Molecular population structure and the biogeographic history of a regional fauna: A case history with lessons for conservation biology. Oikos 63(1):62-76.

Avise, J. C. 1994. Molecular Markers, Natural History And Evolution. Chapman & Hall, New York. 10 Avise, J. C. 2009. Phylogeography: retrospect and prospect. Journal of Biogeography 36:3-15.

Avise, J. C. 2004. Molecular Markers, Natural History and Evolution (2nd ed.). Sinauer, Sunderland, MA.

Avise, J. C., E. Bermingham, L. G. Kessler and N. C. Saunders. 1984a. Characterization of mitochondrial-DNA variability in a hybrid swarm between subspecies of bluegill sunfish (Lepomis macrochirus). Evolution 38(5):931-941.

Avise, J. C., G. S. Helfman, N. C. Saunders and L. S. Hales. 1986. Mitochondrial-DNA differentiation in north-atlantic eels - Population genetic consequences of an unusual life-history pattern. Proceedings of the National Academy of Sciences of the United States of America 83(12):4350-4354.

Avise, J. C., J. Arnold, R.M. Ball, E. Bermingham, T. Lamb, J.E. Neigel, C. A. Reeb and N. C. Saunders. 1987. Intraspecific phylogeography: The mitochondrial DNA bridge between population genetics and systematics. Annual Review of Ecology and Sytematics 18:489-522.

Avise, J. C., J. E. Neigel and J. Arnold. 1984b. Demographic influences on mitochondrial DNA lineage survivorship in populations. Journal of Molecular Evolution 20(2):99-105.

Avise, J. C. and W. S. Nelson. 1989. Molecular genetic relationships of the extinct dusky seaside sparrow. Science 243(4891):646-648.

Bermingham, E., T. Lamb and J. C. Avise. 1986. Size polymorphism and heteroplasmy in the mitochondrial-DNA of lower-vertebrates. Journal of Heredity 77(4):249- 252.

Bickham, J. W., T. Lamb, P. Minx and J. C. Patton. 1996. Molecular systematics of the genus Clemmys and the intergeneric relationships of emydid turtles. Herpetologica 52(1):89-97.

Birks, S. M. and S. V. Edwards. 2002. A phylogeny of the megapodes (Aves : Megapodiidae) based on nuclear and mitochondrial DNA sequences. Molecular Phylogenetics and Evolution 23(3):408-421.

11 Brown, W. M., M. George and A. C. Wilson. 1979. Rapid evolution of animal mitochondrial-DNA. Proceedings of the National Academy of Sciences of the United States of America 76(4):1967-1971.

Burke, R. L., T. E. Leuteritz and A. J. Wolf. 1996. Phylogenetic relationships of emydine turtles. Herpetologica 52(4):572-584.

Butler, J. M., C. K. Dodd, Jr., M. Aresco and J. D. Austin. 2011. Morphological and molecular evidence indicates that the (Terrapene carolina major) is not a distinct evolutionary lineage in the Florida Panhandle. Biological Journal of the Linnaean Society 102:889-901.

Caccone, A., G. Gentile, C. E. Burns, E. Sezzi, W. Bergman, M. Ruelle, K. Saltonstall and J. R. Powell. 2004. Extreme difference in rate of mitochondrial and nuclear DNA evolution in a large ectotherm, Galapagos tortoises. Molecular Phylogenetics and Evolution 31(2):794-798.

Carr, A. E. 1940. A contribution to the herpetology of Florida. University of Florida Publication, Biological Science Series 3:100-101.

Carr, A. E. 1952. Handbook of Turtles: The Turtles of the United States and Canada and Baja California. Cornell University Press, Ithaca, New York.

Conant, R. and J. T. Collins. 1991. A Field Guide to the Reptiles and Amphibians of Eastern and Central North America (3rd ed.). Houghton Mifflin, Boston, Massachusetts.

Curole, J. P. and T. D. Kocher. 1999. Mitogenomics: digging deeper with complete mitochondrial genomes. Trends in Ecology & Evolution 14(10):394-398.

Dodd, K. C. 2001. North American Box Turtles, A Natural History. University of Oklahoma Press, Norman.

Domokos, G. and P. L. Varkonyi. 2008. Geometry and self-righting of turtles. Proceedings of the Royal Society B-Biological Sciences 275(1630):11-17.

12 Edmands, S. 2007. Between a rock and a hard place: evaluating the relative risks of inbreeding and outbreeding for conservation and management. Molecular Ecology 16(3):463-475.

Engstrom, T. N., H. B. Shaffer and W. P. McCord. 2004. Multiple data sets, high homoplasy, and the phylogeny of softshell turtles (Testudines: ). . Systematic Biology 53:693-710.

Ernst, C. H. and R. W. Barbour. 1989. Turtles of the World. Smithsonian Institution Press, Washington, D.C., and London.

Feldman, C. R. and J. F. Parham. 2002. Molecular phylogenetics of emydine turtles: Taxonomic revision and the evolution of shell kinesis. Molecular Phylogenetics and Evolution 22(3):388-98.

FitzSimmons, N. N., C. Moritz, C. J. Limpus, L. Pope and R. Prince. 1997. Geographic structure of mitochondrial and nuclear gene polymorphisms in Australian green turtle populations and male-biased gene flow. Genetics 147(4):1843-1854.

Frankham, R., J. D. Ballou and D. A. Briscoe. 2002. Resolving taxonomic uncertainties and defining management units. In Introduction to Conservation Genetics, (eds. R. Frankham, J. D. Ballou and D. A. Briscoe). Cambridge University Press, Cambridge, United Kingdom.

Fujita, M. F., T. N. Engstrom, D. E. Starkey and S. H. B. 2004. Turtle phylogeny: insights from a novel nuclear intron. Molecular Phylogenetics and Evolution 31:1031-1040.

Gaffney, E. S. and P. A. Meylan. 1988. A phylogeny of turtles. The phylogeny and classification of the tetrapods 1:157-219.

Groth, J. G. and G. F. Barrowclough. 1999. Basal divergences in birds and the phylogenetic utility of the nuclear RAG-1 gene. Molecular Phylogenetics and Evolution 12:115-123.

Haig, S. M., E. A. Beever, S. M. Chambers, H. M. Draheim, B. D. Dugger, S. Dunham, E. Elliott-Smith, J. B. Fontaine, D. C. Kesler and B. J. Knaus. 2006. Taxonomic considerations in listing subspecies under the US Act. Conservation Biology 20(6):1584-1594. 13 Herrmann, H. and P. C. Rosen. 2009. Conservation of aridlands turtles III: preliminary genetic studies of the and yaqui slider. Sonoran Herpetologist 22(4):38-43.

Hillis, D. M., C. Moritz and B. K. Mable. 1996. Molecular Systematics. Sinauer Associates, Sunderland, MA.

Maddison, W. P. 1997. Gene trees in species trees. Systematic Biology 46:523-536.

Mayr, E. 1963. Animal Species and Evolution. Belknap Press at Harvard University Press, Cambridge, MA.

Milstead, W. W. 1969. Studies on the evolution of the box turtles (genus Terrapene). Bulletin of Florida State Museum of Biological Sciences 14:1-113.

Milstead, W. W. and D. W. Tinkle. 1967. Terrapene of Western Mexico, with comments on species groups in the genus. Copeia 1967(1):180-187.

Minx, P. 1992. Variation in phalangeal formulas in the turtle genus Terrapene. Journal of Herpetology 26(2):234-238.

Minx, P. 1996. Phylogenetic relationships among the box turtles, Genus Terrapene. Herpetologica 52(4):584-597.

Moore, W. S. 1995. Inferring phylogenies from mtDNA variation - mitochondrial-gene trees versus nuclear-gene trees. Evolution 49(4):718-726.

Parham, J. F., W. B. Simison, K. H. Kozak, C. R. Feldman and H. T. Shi. 2001. New Chinese turtles: endangered or invalid? A reassessment of two species using mitochondrial DNA, allozyme electrophoresis and known-locality specimens. Animal Conservation 4:357-367.

Prychitko, T. M. and W. S. Moore. 1997. The utility of DNA sequences of an intron from the [beta]-fibrinogen gene in phylogenetic analysis of woodpeckers (Aves: Picidae). Molecular Phylogenetics and Evolution 8(2):193-204.

14 Prychitko, T. M. and W. S. Moore. 2000. Comparative evolution of the mitochondrial cytochrome b gene and nuclear ϲ-fibrinogen intron 7 in woodpeckers. Molecular Biology and Evolution 17(7):1101-1111.

Rubinoff, D. and B. S. Holland. 2005. Between two extremes: Mitochondrial DNA is neither the panacea nor the nemesis of phylogenetic and taxonomic inference. Systematic Biology 54(6):952-961.

Saunders, N. C., L. G. Kessler and J. C. Avise. 1986. Genetic variation and geographic differentiation in mitochondrial-DNA of the horseshoe-crab, Limulus polyphemus. Genetics 112(3):613-627.

Spinks, P. Q. and H. B. Shaffer. 2009. Conflicting mitochondrial and nuclear phylogenies for the widely disjunct Emys (Testudines: Emydidae) species complex, and what they tell us about biogeography and hybridization. Systematic Biology 58(1):1-20.

Stephens, P. R. and J. J. Wiens. 2003. Ecological diversification and phylogeny of emydid turtles. Biological Journal of the Linnean Society 79(4):577-610.

Stuart, B. L. and J. F. Parham. 2007. Recent hybrid origin of three rare Chinese turtles. Conservation Genetics 8(1):169-175.

Sunnucks, P. 2000. Efficient genetic markers for population biology. Trends in Ecology and Evolution 15(5):199-203.

Templeton, A. R. 1998. Nested clade analyses of phylogeographic data: testing hypotheses about gene flow and population history. Molecular Ecology 7(4):381- 397.

Templeton, A. R. 2004. Statistical phylogeography: methods of evaluating and minimizing inference errors. Molecular Ecology 13(4):789-809.

Wiens, J. J., C. A. Kuczynski and P. R. Stephens. 2010. Discordant mitochondrial and nuclear gene phylogenies in emydid turtles: implications for speciation and conservation. Biological Journal of the Linnaean Society 99:445-461.

15

Chapter Two

Re-assessing the Classification of the American Box Turtles (Terrapene spp.) Using

Mitochondrial and Nuclear DNA Markers

Introduction

The American box turtles, Terrapene, currently consist of four species and ten subspecies, but their classification is largely unresolved due to three main factors. First, paraphly within T. carolina has made phylogenetic inference of this species particularly problematic with Stephens and Wiens (2003) stating that T. carolina “might also consist of multiple species.” Second, a large amount of inter- and intraspecific morphological variation, and even high variation between individuals within the same populations, has made phylogenetic inference based on morphology, which is what the current classification is based on (e.g., Minx, 1996), less than useful. Third, intergradation between sympatric taxa imposes significant difficulties in resolving specific and subspecific relationships (Carr, 1940; Carr, 1952; Milstead, 1969; Conant and Collins,

1991; Butler et al., 2011). Therefore, assessing the evolutionary relationships within

Terrapene using molecular data is essential. Due to the wide distribution of the group and the high level of intraspecific variation seen within species, and even within populations, it is necessary to sample a wide range of individuals with sufficient sampling from each area to most appropriately assess their classification.

16

Currently, there are four recognized species within Terrapene: T. ornata, T. carolina, T. coahuila, and T. nelsoni (Milstead and Tinkle, 1967; Milstead, 1969; Minx,

1992; Minx, 1996). The subspecies currently associated with the aforementioned species of Terrapene, as well as their countries of inhabitance, are listed in Table 1. However, the phylogenetic relationships among some of these taxa are convoluted.

Terrapene currently consists of an eastern group (Terrapene carolina and

Terrapene coahuila) and a western group (Terrapene ornata and Terrapene nelsoni;

Minx, 1996). It is hypothesized that T. coahuila is basal to the eastern clade (Auffenberg,

1958; Legler, 1960; Minx, 1996), and the extinct Terrapene carolina putnami (giant box turtle; Hay, 1906) was ancestral to much of T. carolina, either by allopatric speciation or by hybridization events (Auffenberg, 1958; Auffenberg, 1959; Milstead, 1967; Milstead,

1969; Gillette, 1974). Terrapene ornata and T. nelsoni are thought to have diverged as a result of allopatric speciation (Milstead and Tinkle, 1967).

In the last major classification assessment for Terrapene, Minx (1996) evaluated

32 morphological characters and found several close associations (Figure 1). First, he hypothesized the existence of an eastern group (T. carolina and T. coahuila) and a western group (T. ornata and T. nelsoni). Second, within the eastern clade, he found three close associations: 1) T. c. bauri - T. c. carolina, 2) T. c. triunguis - T. c. mexicana -

T. c. yucatana, and 3) T. c. major - T. coahuila, with T. coahuila as the sister species to T. carolina. Third, T. o. ornata and T. o. luteola form a close association, and represent the

17 sister species to T. nelsoni. However, he also indicated that several of these associations remain unresolved (e.g., the relationship between the T. c. bauri - T. c. carolina and T. c triunguis - T. c. mexicana - T. c. yucatana clades), and that molecular data were warranted in order to help resolve these ambiguities.

Whereas several phylogenetic studies have assessed Terrapene using molecular data to a limited extent (Figure 2; Bickham et al., 1996; Feldman and Parham, 2002;

Stephens and Wiens, 2003; Herrmann and Rosen, 2009; Spinks and Shaffer, 2009; Wiens et al., 2010; Butler et al., 2011), they have focused either on intergeneric classification, they do not include all taxa within Terrapene (e.g., all species and subspecies), or they are limited in geographic sampling. Thus, a re-assessment of the entire genus is needed to resolve the evolutionary history of the group. Because Terrapene are currently of conservation concern, this resolution will also be useful in the formulation of potential conservation management strategies.

American box turtle populations are declining throughout their range, in part, due to habitat loss resulting from increasing urbanization, collection from the wild for the pet trade, and changes in predator pressures (Dodd, 2001). The 2011 International Union for

Conservation of Nature (IUCN) Red List classifies T. carolina as Vulnerable, T. ornata as Near Threatened, T. coahuila as Endangered with a Very High Risk of Extinction, and

T. nelsoni as Data Deficient (although it was listed as Threatened on the 2006 Red List).

In the United States the various subspecies of the endemic T. carolina and T. ornata are state listed as Species of Special Concern in NH, CT, MI, TX, and MA, Protected in IN,

18 and Endangered in ME, WI, IL. Thus understanding their evolutionary history, as well as their ecology (e.g., range dynamics and habitat preferences) are particularly urgent. In order to prevent further declines in box turtle populations, it is essential that conservation efforts be employed. However, to develop conservation strategies, additional knowledge is needed on the evolutionary history and underlying genetics of this taxon. Specifically, state and federal agencies need to have knowledge of the level of classification (e.g., species or subspecies) for each group in order to make informed decisions about their conservation status. Because conservation efforts are typically species-based, it is important to have an understanding of their specific and subspecific status and relationships because the conservation strategies may differ depending on their classification. For example, the conservation management strategies may differ if certain groups, such as T. c. carolina and T. c. triunguis, were to be classified as separate species or lumped together as subspecies. It is also important to have information regarding their underlying genetics to prevent the mixing of divergent genetic lineages in a population.

The goals of this research are to 1) resolve the evolutionary history of the

Terrapene genus by assessing their classification using molecular phylogenetics and 2) assess the population structure of their entire range to evaluate their phylogeography.

Materials and Methods

Volunteers, museums, universities, private parties, and various other organizations supplied tissue samples in the form of tail tips, toenails, shell shavings, shell fragments, bone fragments, scutes, scales, muscle tissue, feces, and blood. The

19 tissue was collected from both living and dead individuals, with the less invasive collection methods (such as toenail clippings and shell shavings) being used with live specimens. Upon receipt, samples were stored in a -20º C freezer in either di-methyl sulfoxide (DMSO) buffer super-saturated with salt, or 95% ethanol. A total of 253 and

202 box turtle tissue samples were used for the mitochondrial DNA (mtDNA) and nuclear DNA (nucDNA) data analyses, respectively (Appendix A). At least three individuals for every species and subspecies and at least three individuals from every U.S. state within their range were chosen in order to include a wide geographic sampling

(Figure 3). In addition, at least three tissue samples were used in the data analyses for the

Mexican species and subspecies (except for T. c. yucatana and T. nelsoni due to the very limited amount of tissue samples available for these taxa).

DNA extractions, PCR, and sequencing

Genomic DNA was extracted from tissue samples with the illustra™ tissue & cells genomicPrep Mini Spin Kit (GE Healthcare). One mtDNA gene (i.e., Cytb) and one nucDNA gene (i.e., GAPD) were then amplified and sequenced. For Cytb (cytochrome b), the entire 1,097 base pair (bp) gene along with part of the adjacent tRNA-threonine

(tRNA-thr) gene (Saiki et al., 1988; Shaffer et al., 1997; Zamudio and Greene, 1997;

Lenk et al., 1999; Feldman, 2000; Rodriguez-Robles et al., 2001; Feldman and Parham,

2002; Stephens and Wiens, 2003; Spinks and Shaffer, 2009; Wiens et al., 2010) were amplified and sequenced using the forward primer CytbG and the reverse primer THR-8

(Spinks et al., 2004; Engstrom et al., 2007). For GAPD (glyceraldehyde-3-phosphate dehydrogenase), a 430 bp region of the gene including intron 11 and partial coding

20 region was amplified and sequenced using the forward primer GAPDL890 and the reverse primer GAPDH950 (Friesen et al., 1997; Dolman and Phillips, 2004). For both genes 20 µL PCR reactions were utilized and consisted of 7.1 µL H2O, 2.0 µL TopTaq

PCR buffer (Qiagen), 0.4 µL dNTPs, 2.0 µL Coral Load (Qiagen), 4.0 µL Q, 1.0 µL each primer, and 2.4 µL DNA. A negative control was included with each PCR to confirm that no contamination had occurred. The following parameters were used for the Cytb

DNA amplification: 35 cycles of 1 min denaturing at 94ºC, 1 min annealing at 51ºC, and

2 min DNA elongation at 72ºC. GAPD PCR parameters were as follows: initial denaturation for 5 min at 94ºC followed by 35 cycles of 30 sec denaturing at 94ºC, 1 min annealing at 63ºC, and 1 min 30 seconds extension at 72ºC. For the GAPD PCR, a final extension was performed for 5 min at 72ºC.

5 µL of each PCR product was then subjected to gel electrophoresis with each 1% agarose gel being prepared with tris-acetate-EDTA (TAE) buffer; the DNA was visualized in the gel with ethidium bromide to confirm successful PCR amplification. A molecular weight marker (1kb DNA ladder) was included with each gel to confirm that the amplified DNA was the appropriate size. Amplified DNA was purified with the

E.Z.N.A. Cycle Pure Kit (OMEGA bio-tek). Purified DNA was concentrated to the level recommended by Eurofins MWG Operon (20-40 ng/µL) and shipped to Eurofins MWG

Operon for sequencing reactions using BigDye® Terminator v 3.1 Cycle Sequencing kits

(Applied Biosystems).

21 Sequence analysis and phylogenetic inference

DNA was sequenced on an ABI 3730xl DNA sequencer at Eurofins MWG

Operon and manually proofread and edited using Sequencher 4.9 (Gene Codes

Corporation). Sequence alignments were conducted in Clustal X 2.0.11 (Thompson et al.

1997). Final editing was done using MacClade 4.08 (Maddison and Maddison, 1989).

When available, GenBank sequences from the literature were included in each analysis

(listed in Table 2). A T. o. luteola sequence for the GAPD gene was not available on

GenBank and sequences for T. c. bauri, T. c. major, T. c. mexicana, T. n. klauberi, and T. c. yucatana were not available for either gene.

Tajima’s D and Fu and Li’s D* and F* tests for neutrality were conducted for each gene using DnaSP v 5.10.01 (Librado and Rozas, 2009) to confirm that natural selection did not significantly influence the phylogenetic data and that the inferred phylogeny largely reflects the background rate of mutation (Tajima, 1989; Fu and Li,

1993). Phylogenies were inferred via maximum likelihood (ML; Felsenstein, 1981), and

Bayesian inference (BI; Smouse and Li, 1989; Rannala and Yang, 1996; Yang and

Rannala, 1997; Larget and Simon, 1999) methods. PhyML 3.0 was used to generate ML trees (Guindon et al., 2010), and BEAST v1.6.2 was used to infer BI trees (Drummond and Rambaut, 2007). Non-parametric bootstrap re-sampling (Felsenstein, 1985) was employed to quantify the statistical support for ML phylogenies and the Markov chain

Monte Carlo (MCMC) method was used to infer confidence values for BI (Mau, 1996;

Rannala and Yang, 1996; Mau and Newton, 1997; Yang and Rannala, 1997; Mau et al.,

1999). One thousand non-parametric bootstrap replications were used for ML trees

22 (Pattengale et al., 2009), and 50% majority-rule consensus trees were generated.

Bootstrap support values above 70% were considered well-supported (Hillis and Bull,

1993). BI analyses were run for 3.0 x 106 MCMC generations using default temperatures and with sampling trees occurring every 100 generations. The aforementioned 3.0 x 106

MCMC generations were chosen to make effective sample sizes (ESS) > 200 for the individual parameters in the analysis, as determined by Tracer v1.5 from the BEAST v1.6.2 software package (Drummond and Rambaut, 2007). The likelihood scores were monitored during each analysis until stabilization, and the samples obtained prior to stabilization were discarded as burn-in (Parham et al., 2006). Fifty percent majority-rule consensus trees were generated, and nodes having a Bayesian posterior probability (BPP)

≥ 95% were considered well-supported, as a 95% BPP corresponds to a 95% confidence interval (Huelsenbeck and Ronquist, 2001). For all analyses, jModelTest 0.1 was used to determine substitution model parameters using the Akaike Information Criterion corrected for small sample size (AICc; Posada, 2008). For Cytb, ML and BI analyses were conducted using the TPM2uf + I + G substitution model, with I = 0.4450 and G =

0.6160 and the sample size = 1,097. The rate matrix parameters were as follows: AC =

3.2798, AG = 21.7809, AT = 3.2798, CG = 1.0000, CT = 21.7809, and GT = 1.0000.

The base frequencies were set to 0.3059 (A), 0.3131 (C), 0.1213 (G), and 0.2598 (T).

GAPD ML and BI analyses were conducted using the K80 + I substitution model, with I

= 0.7080, the sample size = 430, and Kappa = 3.8280 (Ti/Tv = 1.9140). Each phylogeny was rooted with a published GenBank sequence from Clemmys, which is considered the sister genus to Terrapene (Bramble, 1974; Bickham et al., 1996; Feldman and Parham,

23 2002; Stephens and Wiens, 2003). Zero-length branches were collapsed into unique haplotypes using Collapse 1.2 to reduce clutter and computation time.

Combined data

The Cytb and GAPD DNA sequence data were concatenated into a single dataset in order to infer a combined mtDNA and nucDNA phylogram. The combined dataset contained 172 sequences condensed into 121 haplotypes. Prior to phylogenetic analysis, an Incongruence Length Difference (ILD; Mickevich and Farris, 1981; Farris et al., 1994) test with 100 replicates was performed using PAUP* v 4.0b10 (Swofford, 2003) to assess whether the topologies of the two trees were congruent. Each gene was partitioned separately for the combined analysis, and the model parameters for each partition were kept the same as for the individual analyses. Bayesian inference was conducted using

MrBayes v 3.1.2 (Ronquist and Huelsenbeck, 2003). The analysis was run for 3.0 X 106

MCMC generations with one cold chain and three heated chains using the default temperature. Log-likelihood values prior to stabilization were discarded as burn-in. BPP values ≥ 95% were considered well-supported.

Haplotype networks

Haplotype networks are useful for sequence variation within species or among closely related species because for recently diverged taxa there could potentially be a high number of mutational variants and because of the possibility of reversion to ancestral haplotypes for recently diverged subspecies (Crandall, 1994; Posada and

Crandall, 2001). Given that there are multiple subspecies in this dataset, haplotype

24 networks may further support the results of the ML phylogenies. Using the 95% statistical parsimony procedure (Templeton, 1998), TCS 1.13 (Templeton et al., 1992) was used to estimate a gene genealogy. Reticulation loops were removed a posteriori based on coalescent theory, which suggests that lower frequency haplotypes will be located externally on the haplotype network and higher frequency haplotypes will be located internally (Crandall, 1994). Haplotype bubbles were sized relative to the number of sequences within each haplotype. The haplotype bubbles connected by branches differed by one mutational step, and smaller bubbles were placed on the branches to represent missing intermediate steps. Because TCS 1.13 does not read ambiguity codes for heterozygous characters, the ambiguous characters were subjected to Phase analysis in DnaSP v 5.10.01 (Librado and Rozas, 2009). The Phase analysis determines the gametic phase of the DNA sequences and reconstructs the haplotypes without the ambiguous characters (Stephens et al., 2001; Stephens and Donnelly, 2003). Using Phase to reconstruct ambiguous haplotypes has been shown to be at least as accurate as the more time and cost intensive cloning method (Harrigan et al., 2008). The sequences, sans ambiguity codes, were then condensed into 41 unique haplotypes using Collapse 1.2 (see

Appendix B for haplotype information).

AMOVA and SAMOVA analyses

Spatial Analysis of Molecular Variance (SAMOVA) was performed using

SAMOVA 1.0 (Dupanloup et al., 2002) and Analysis of Molecular Variance (AMOVA;

Excoffier et al., 1992) was conducted using ARLEQUIN v. 3.11 (Excoffier et al., 2005) to examine population structures. Φ statistics, which are analogous to F-statistics

25 (Wright, 1951), were calculated from these analyses to assess how much variation is explained by groupings of populations or classification-based groupings. ΦCT values indicate the percent of variation explained among groups, ΦSC values indicate the percent of variation explained among populations within groups, and ΦST values indicate the percent of variation explained within populations. AMOVA was used to assess the population structure of molecular variation, and SAMOVA was used to assess whether geographically sympatric groups are maximally genetically isolated. In other words,

AMOVA and SAMOVA use the amount of variance explained among groups to assess whether there is a well-defined population structure. SAMOVA does not make assumptions about whether the populations are at Hardy-Weinberg equilibrium and a priori groups are not assigned to the populations, whereas AMOVA assigns populations into a priori groups. SAMOVAs can be less biased than AMOVAs because SAMOVAs assign groups based on geographic data with a user-defined number of groups, whereas the user must assign a priori AMOVA groups. Each SAMOVA and AMOVA was conducted with 1,000 simulated annealing permutations. For both Cytb and GAPD, a priori AMOVA groups were assigned several different ways in order to compare phylogenetic hypotheses using population structuring. Furthermore, different AMOVA tests were run both a priori and a posteriori to seeing the phylogenetic trees.

A priori AMOVAs

First, each subspecies (or species for monotypic groups) was apportioned separately into ten unique groups as follows: T. coahuila, T. n. nelsoni, T. c. carolina, T. c. triunguis, T. c. major, T. c. bauri, T. c. yucatana, T. c. mexicana, T. o. ornata, and T. o.

26 luteola. This ten-group analysis was conducted in order to assess whether there was greater or lower population structure when T. o. luteola and T. o. ornata were grouped together (see below under A posteriori AMOVAs). Second, groups were assigned based on the morphological data of Minx (1996), which included four groups based on species- level classification (T. carolina, T. ornata, T. coahuila, and T. nelsoni). This second analysis was conducted in order to compare population structuring between Minx’s

(1996) phylogenetic hypothesis and the relationships inferred from the mtDNA and nucDNA data (Figures 4-5). Lastly, groups were apportioned into one eastern (T. carolina and T. coahuila) and one western (T. ornata and T. nelsoni) group in order to assess whether the population structure would be better explained by eastern vs. western populations.

A posteriori AMOVAs

Several AMOVAs were performed based on the results or the phylogenetic analyses. First, each subspecies was assigned its own unique group except for T. o. ornata and T. o. luteola (totaling nine groups). T. o. ornata and T. o. luteola were combined in order to assess whether population structuring was greater or lower after they were clumped together. Second, the re-assessment provided by the mtDNA and nucDNA phylogenies (Figures 4-5) was compared with the hypothesis of Minx (1996) in order to evaluate which hypothesis indicated a higher population structure. The groups representing the presented mtDNA and nucDNA findings (five total groups) were as follows: T. c. carolina - T. c. major - T. c. bauri; T. coahuila; T. c. triunguis - T. c. mexicana - T. c. yucatana; T. o. ornata - T. o. luteola; and T. nelsoni. Third, T. coahuila

27 was apportioned into a group along with T. c. carolina and T. c. major in order to evaluate whether a greater population structure would be observed by grouping T. coahuila with the T. c. carolina group. Fourth, T. c. bauri was combined with T. c. carolina - T. c. major in order to evaluate whether a greater amount of population structure would be observed by grouping T. c. bauri with the T. c. carolina group.

Cytb molecular clock analysis

A relaxed, uncorrelated lognormal molecular clock was placed on the Cytb sequence data using the BEAST v1.6.2 software package in order to estimate divergence times (Drummond and Rambaut, 2007). A molecular clock was not inferred for GAPD due to a generally lower resolution of the ML and Bayesian phylograms. Clemmys guttata, Glyptemys muhlenbergii, Glyptemys insculpta, Emys orbicularis, and Emys marmorata were used to root the tree, as previous data indicated that these genera are most closely related to Terrapene within Emydinae (Bramble, 1974; Bickham et al.,

1996; Feldman and Parham, 2002; Stephens and Wiens, 2003). The molecular clock analysis was conducted using the Yule Process speciation tree prior and a GTR + I + G substitution model with the parameters and tree priors equal to those for the previously mentioned Cytb ML and BI phylogenetic analyses, as determined by jModelTest 0.1

(Posada, 2008). Fossil data and previously published divergence estimates were used to calibrate the molecular clock. The root was calibrated to 29.4 million years ago (mya), with the standard deviation (SD) = 2.01, as the most recent common ancestor (MRCA) to

Emydinae, and the MRCA to Glyptemys was calibrated to 17.0 mya (Spinks and Shaffer,

2009), with the SD = 2.90. Based on fossil data, the MRCA to T. ornata was calibrated

28 at 12.5 mya with the SD = 1.00, and the MRCA to T. carolina was calibrated to 10.0 mya with the SD = 1.00 (Holman and Fritz, 2005; Spinks and Shaffer, 2009). The molecular clock analysis was conducted using a Markov chain of 75.0 X 106 generations, with sampling occurring every 1,000 generations. As with the previously mentioned BI trees,

75.0 X 106 generations were chosen in order to bring the ESS > 200 for all parameters and tree priors in the analysis when analyzed using Tracer v1.5. The number of generations excluded from the analysis as burn-in was chosen based on visual inspection for stabilization of the log likelihood values.

Pairwise Jukes-Cantor DNA sequence divergences corrected for population comparisons (Jukes and Cantor, 1969) were calculated with DnaSP v 5.10.01 (Librado and Rozas, 2009) and used to calculate the percent sequence divergence between taxonomic groups. Interspecific mtDNA sequences for most Emydine turtles typically vary between 4-6%, with a mean of 5% for the mtDNA Cytb gene, and interspecific nuclear DNA sequences typically vary between ~0.2-4%, with a mean of ~1% in most freshwater turtles and tortoises, depending on the gene (Feldman and Parham, 2002;

FitzSimmons and Hart, 2007). Therefore these average values were used as references to compare the species and subspecies within Terrapene.

Results

Cytb phylogenetic analysis

The mtDNA Cytb phylogram contained 253 sequences distributed into 103 haplotypes (Figure 4; Appendix C). Out of the 1,097 characters, 254 were variable, and

29 191 were parsimoniously informative. Tajima’s D, Fu and Li’s D* and F* tests indicated that Cytb was not significantly being influenced by natural selection and that it largely reflects the background rate of mutation (Tajima’s D: -0.07205, P > 0.10; Fu and Li’s

D*: -1.14188, P > 0.10; Fu and Li’s F*: -0.80538, P > 0.10).

The phylogenetic analysis split Terrapene into eastern and western clades. The eastern clade contained T. carolina and T. coahuila, and the western clade contained T. ornata and T. nelsoni, which agrees with Minx (1996). However, several relationships within each clade differed from his phylogenetic hypothesis. Specifically, the relationships within the eastern “trichotomy” and between T. o. ornata and T. o. luteola were not apparent in the molecular phylogeny.

Within the eastern clade, the Cytb phylogram (Figure 4) supports a close association between 1) Terrapene c. triunguis, T. c mexicana, and T. c. yucatana, 2) T. c. carolina and T. c. bauri, and 3) T. c. major and T. coahuila. However, the data presented here indicate paraphyletic relationships within T. carolina that disagree with the morphological data from Minx (1996). The associations between specific Terrapene taxa for the Cytb phylogram are given below.

T. c. carolina - T. c. triunguis and their associated taxa

The Cytb phylogenetic data suggest that the T. c. triunguis - T. c. yucatana - T. mexicana clade is paraphyletic to T. carolina. Regarding the taxa associated with T. carolina, T. c. major; T. c. bauri; T. c. carolina; and T. coahuila form a monophyletic

30 clade. Finally, T. c. mexicana, T. c. yucatana, and T. c. triunguis form a monophyletic clade.

T. c. major - T. coahuila

Terrapene c. major is polyphyletic within T. carolina, as it is found in three different clades: 1) a clade consisting strictly of T. c. major, 2) a clade consisting of one T. coahuila haplotype that contained five sequences and several T. c. major haplotypes, and

3) within the T. c. carolina clade. In some cases the T. c. major found within the T. c. carolina clade are unique haplotypes, whereas others share haplotypes with T. c. carolina.

The clade consisting of T. c. major and T. coahuila was not basal to T. carolina, which is incongruent with the morphological data of Minx (1996).

Cytb western clade

Within the western clade, Terrapene o. ornata and T. o. luteola form a monophyletic clade. T. o. ornata and T. o. luteola lack the population structure that

Herrmann and Rosen (2009) suggested (they found unique groupings of T. o. ornata and

T. o. luteola). In the presented data, T. o. ornata and T. o. luteola do not share any haplotypes, but also do not show any apparent pattern of grouping, suggesting that they are lacking distinction genetically. T. nelsoni and T. ornata are sister taxa within the western clade.

31 GAPD phylogenetic analysis

The nucDNA GAPD phylogram contained 59 haplotypes and 202 sequences

(Figure 5; Appendix D). Thirty-seven of the 430 characters were variable, and 32 characters were parsimoniously informative. For GAPD, Tajima’s D, Fu and Li’s D* and

F* tests for neutrality indicated that natural selection is not significantly influencing the rate of mutation (Tajima’s D: -1.24918, P > 0.10; Fu and Li’s D*: 0.63148, P > 0.10; Fu and Li’s F*: -0.15751, P > 0.10).

The split of the eastern (T. carolina) and western (T. ornata) clades is well supported with bootstrap resampling but not BPP (Bayesian posterior probabilities). In discordance with the mtDNA data set, one individual that was morphologically identified as a T. c. carolina was found in the western clade and one individual that was morphologically identified as a T. o. ornata was found in the eastern clade.

T. c. carolina - T. c. triunguis

Within the eastern clade, the GAPD tree topology agrees with the mtDNA data in terms of the paraphyly between T. c. triunguis - T. c. mexicana and T. carolina, with the exception of two T. c. carolina individuals from SC being found within the T. c. triunguis clade (Figure 5). Terrapene c. mexicana is monophyletic with T. c. triunguis; however, T. c. yucatana is polyphyletic in the nucDNA tree, being separated into two haplotypes: 1) one within the T. c. triunguis clade and 2) the other within the T. carolina clade.

32 Taxa associated with T. c. carolina

The GAPD tree indicates that Terrapene c. bauri is paraphyletic to T. carolina, which is in discordance with the mtDNA tree. T. c. bauri was also more closely grouped with T. c. triunguis than with T. c. carolina. In concordance with the mtDNA tree, T. c. major is polyphyletic within T. carolina for the nucDNA GAPD tree, being distributed into four clades: 1) a clade consisting of solely T. c. major; 2) a clade consisting of one haplotype that shares T. c. major and T. coahuila; 3) a clade intermixed with T. c. carolina that shares a haplotype with T. c .carolina; and 4) one haplotype that is grouped with T. c. triunguis.

GAPD western clade

Within the western clade, Terrapene nelsoni is basal to T. ornata, which disagrees with the mtDNA data. In addition, as with the mtDNA tree, T. o. ornata and T. o. luteola appear to lack genetic distinction. They are not grouped in any apparent pattern for either gene, and for GAPD specifically T. o. luteola share some haplotypes with T. o. ornata.

Combined mtDNA and nucDNA phylogeny

Cunningham (1997) indicated that if P > 0.01 for the ILD test, the phylogenetic accuracy of the tree is improved, and if P < 0.001 the accuracy of the analysis is reduced.

Based on these values, the accuracy of the combined tree was not significantly improved but was also not significantly adversely affected (P = 0.01). The combined phylogeny contains 173 sequences distribute into 121 haplotypes (Figure 6; Appendix A)

33 and generally agrees with the Cytb topology. The BPP values of the combined phylogeny indicate greater support for the Terrapene coahuila - T. c. major clade but less support for the T. c. bauri - T. carolina and T. ornata - T. nelsoni clades.

Haplotype networks

The Cytb haplotype network consists of 103 haplotypes and is divided into nine subgroups that do not fall within 95% confidence intervals using the statistical parsimony procedure (Figure 8). Terrapene o. ornata and T. o. luteola are the only taxa that that fall within a 95% confidence interval and are located within the same subgroup. As is the case with the Cytb phylogram, T. c. major is polyphyletic, being distributed among three clades and sharing haplotypes with T. c. carolina in some cases. Terrapene coahuila is closely associated with T. c. major, and is separated from T. c. major by nine missing intermediate steps, suggesting the need for additional sampling for T. coahuila.

The GAPD haplotype network consists of 41 unique haplotypes after removing heterozygous characters and collapsing identical sequences (Figure 9). The network is split into two subgroups, with one containing T. nelsoni and the other containing the rest of Terrapene. Although several taxa within the GAPD network are difficult to interpret, five main clades are evident: T. c. carolina, T. c. triunguis, T. c. major - T. coahuila, T. ornata, and T. nelsoni. However it should be noted that some of these groups are polyphyletic for a few individuals. For example, individuals from T. c. carolina are present in haplotypes also containing T. c. triunguis, T. o. ornata, or T. o. luteola, and T. c. major is present in haplotypes containing T. c. triunguis or T. c. carolina. Some

34 haplotypes also consist of several taxa (e.g., haplotypes 2, 31, and 34; Figure 9; Appendix

B). Lastly, T. c. bauri is more closely associated with T. c. triunguis than with T. c. carolina, which differs from the Cytb data.

Population structure

A total of seven AMOVA hypotheses and one SAMOVA analysis were conducted for each respective mtDNA and nucDNA gene. For descriptions on each

AMOVA analysis and why they were conducted, see Materials and Methods. Each

AMOVA and SAMOVA analysis included three test statistics that provided the percentage of variation explained 1) among groups, 2) within groups but among populations, and 3) within populations. The three test statistics were as follows: 1) ΦCT

(among groups), 2) ΦSC, (within groups but among populations), and 3) ΦST (within populations). For clarity, only the ΦCT values and their associated percentages are reported. The AMOVA results are reported in Tables 4-17 and the SAMOVA results are reported in Tables 19 and 20.

Molecular clock analysis

The Cytb time-calibrated molecular clock analysis indicated an eastern/western divergence time estimate of ~15 mya (Table 3; Figure 7). The estimated divergence times for carolina/coahuila, carolina/bauri, and carolina/major splits were ~10.6 mya,

~10.3 mya, and ~7.2 mya, respectively. Lastly, the divergence estimates for triunguis/yucatana and triunguis/mexicana were ~9.2 and ~7.0 mya, respectively.

Divergence times for T. o. ornata and T. o. luteola could not be calculated because they

35 are so genetically similar and because the haplotypes containing these taxa are intermixed in no apparent pattern within the T. ornata clade.

Pairwise percent sequence divergences for the mtDNA Cytb gene ranged from

0.583% to 7.34%, while percent divergences for the nucDNA GAPD gene ranged from

0.233% to 3.09% (Table 20). For both Cytb and GAPD, T. c. carolina - T. c. triunguis, T. c. carolina -T. o. ornata, T. o. ornata - T. n. nelsoni, T. c. triunguis - T. c. yucatana, and T. o. ornata - T. c. triunguis, and T. c. carolina - T. c. bauri showed relatively high sequence divergences. For GAPD, the percent divergence between T. c. carolina and T. coahuila was relatively high; this comparison was low for Cytb. These high pairwise comparisons showed a greater percentage of sequence divergence than what Feldman and Parham

(2002) indicated as typically representing interspecific relationships within Emydine turtles for the mtDNA Cytb gene and what FitzSimmons and Hart (2007) indicated as interspecific for nucDNA in freshwater turtles and tortoises. The remaining groups showed relatively lower nucleotide divergences.

Discussion

Phylogenetic analyses

My data agree with Minx (1996) in several ways. First, the phylogenies support the monophyly of the eastern and western clades. Second, regarding the eastern clade, T. c. major is monophyletic with the T. carolina clade. Third, T. c. mexicana, T. c. yucatana, and T. c. triunguis form a monophyletic clade. Fourth, the sister relationship of T. ornata and T. nelsoni is supported. However, my data bring several currently

36 hypothesized relationships within Terrapene into question, while leaving some taxa unresolved (Figures 4-5).

Eastern clade

Several relationships proposed by Minx (1996) were not supported in the molecular phylogenies. First, the T. c. triunguis - T. c. yucatana - T. c. mexicana clade is paraphyletic to the T. c. carolina - T. c. major clade. Second, with respect to the sister relationship between T. carolina and T. coahuila as hypothesized by morphological and previous molecular data (Auffenberg and Milstead, 1965; Milstead, 1969; Minx, 1992;

Minx, 1996; Feldman and Parham, 2002; Wiens et al., 2010), the molecular phylogenetic analyses show that T. coahuila is associated with T. c. carolina, but the relationship between the two lacks resolution. T. coahuila and T. c. major also appear to be closely related, but T. coahuila is not the basal taxon to Terrapene as hypothesized by Minx

(1996) and T. c. major is polyphyletic, also being present within other T. carolina clades.

Third, the relationship of T. c. bauri to the rest of T. carolina is unclear, with the Cytb phylogeny supporting the monophyly of T. c. bauri with T. carolina but the GAPD phylogeny supporting the paraphyly of T. c. bauri and T. carolina.

Western clade

Terrapene o. ornata and T. o. luteola are monophyletic within T. ornata but are not grouped in any apparent pattern in this clade. Herrmann and Rosen (2009) found population structuring and a unique clade for T. o. luteola in their haplotype network and molecular phylogeny, respectively. However, for the phylogenies presented here these

37 two taxa are intermixed within the T. ornata clade and in some cases T. o. ornata and T. o. luteola share haplotypes, suggesting that they lack subspecific resolution.

Combined mtDNA and nucDNA data

The combined tree generally agrees with the topology of the Cytb tree, with the exceptions of the support values being greater for the T. coahuila - T. c. major clade and less for the T. c. bauri and T. ornata - T. nelsoni clades. The lower resolution of the T. c. bauri clade is logical because of the discordance of T. c. bauri between the Cytb and

GAPD phylogenies. The lower resolution between T. ornata and T. nelsoni is also logical becuase T. nelsoni forms the basal clade in the GAPD phylogeny, but it is not basal in the Cytb phylogeny. The Cytb gene contains far more variable and parsimoniously informative characters than the GAPD gene, and it is probable that the combined tree generally resembles the Cytb topology due to this character bias. It has been shown that multi-gene trees can provide greater phylogenetic accuracy than trees based on a single gene (Barrett et al., 1991; Olmstead and Sweere, 1994; Gadagkar et al.,

2005). However, while combining the Cytb and GAPD datasets did increase the sample size and improved the support of the T. coahuila - T. c. major clade, additional data from other nucDNA genes are needed to overcome the character bias imposed by Cytb.

These data suggest that classification revisions are needed for some groups, namely the relationships between the T. c. carolina and the T. c. triunguis - T. c. yucatana

- T. c. mexicana clades and between T. o. ornata and T. o. luteola. Other relationships, such as between T. coahuila and T. carolina, between T. c. bauri and T. carolina, and

38 between T. c. major, T. coahuila, and the rest of T. carolina are less clear. Before implementing classification revisions, however, it will be beneficial to obtain more DNA sequence data in the form of additional nucDNA genes, as the presented data are based on only two genes. Obtaining additional genes would allow a more thorough investigation of the molecular phylogenetic relationships of Terrapene.

Polyphyly

Polyphyly was present in the nucDNA GAPD gene that was not present in the mtDNA Cytb gene. For example, in the GAPD analysis one T. c. carolina from IL was found within T. ornata and two from SC in T. c. triunguis, one T. o. ornata from OK was found within the T. c. carolina clade, one T. c. yucatana was found within T. c. carolina, and one T. c. major from FL was found within T. c. triunguis. In addition, several T. c. major from various localities were found within T. c. carolina for both Cytb and GAPD.

For the IL T. c. carolina found within the T. ornata clade, it is possible that this individual is a hybrid because it was collected at a locality where the two subspecies were sympatric. Based on discordance between the nucDNA and mtDNA phylogenies, the T. c. major individuals found within T. c. carolina and T. c. triunguis and the T. c. bauri found within T. c. triunguis are possibly hybrids as their ranges are close together and often are sympatric. Since T. c. major is thought to have originated from a hybridization event between the extinct T. c. putnami (giant box turtle) and T. c. carolina based on similar morphological features (Auffenberg, 1958; Auffenberg, 1959; Dodd, 2001), it is possible that T. c. major represents to some extent the ancestral genetic lineage of the extinct T. c. putnami.

39

The presence of the T. o. ornata from OK, the two T. c. carolina from SC, and the

T. c. yucatana within the T. c. carolina clade are more difficult to explain. It was hypothesized by Milstead (1969) that T. c. yucatana originated via hybridization between the extinct T. c. putnami (giant box turtle), and it is possible that incomplete lineage sorting has occurred for T. c. yucatana as well as the other previously mentioned polyphyletic and discordant individuals (Avise et al., 1983; Neigel and Avise, 1986;

Maddison, 1997; Rosenberg, 2002). Incomplete lineage occurs in cases of short internodes coinciding with large effective population sizes. In other words, a speciation event is followed by a large demographic expansion, resulting in ancestral lineages being found within certain clades in a gene tree. Incomplete lineage sorting can occur even in populations that are allopatric (Avise et al., 1983; Neigel and Avise, 1986), and could explain the polyphyly found within GAPD. It is also possible that they are released pets, are translocated turtles, and/or are misidentified.

Many box turtles have been found far outside their normal range, and it is possible that some of the polyphly is due to misplaced turtles or hybrids being present in an intergradation zone (Dodd, 2001). Butler et al. (2011) similarly identified several individuals from GA that molecularly resembled T. c. carolina but morphologically resembled T. c. triunguis, giving an indication of hybridization. The sampling localities of their GA samples are also near a particularly messy intergradation zone that includes T. c. carolina, T. c. major, T. c. triunguis, and T. c. bauri. My SC samples are from the far southwest region of the state, and due to the close proximity of southwest SC and GA it is

40 possible that this GA hybrid T. c. triunguis - T. c. carolina population has spread to SC.

This intergradation zone may also at least in part be responsible for the T. c. major polyphyly seen for my dataset. Despite polyphyly within some clades and some discordance between the mtDNA and nucDNA phylogenies, both phylogenies are well resolved enough to interpret relationships between T. c. carolina - T. c. triunguis, T. ornata - T. nelsoni, and T. c. triunguis - T. c. mexicana - T. c. yucatana.

Haplotype networks

The Cytb haplotype network (Figure 7) was divided into nine subgroups that did not fall within 95% confidence intervals with the statistical parsimony method. The only two taxa that fell within a 95% confidence interval were T. o. ornata and T. o. luteola.

Accordingly, this mtDNA network indicates a high amount population structure at the subspecific level, excluding T. o. luteola. The GAPD haplotype network (Figure 9) is allocated into just two subgroups, T. nelsoni and all other taxa. Specifically, the GAPD network indicates high population structure for four main clades: T. c. carolina, T. c. triunguis, T. ornata, and T. nelsoni. T. coahuila is closely associated with T. c. major, but as with the Cytb and GAPD phylogenies T. c. major is polyphyletic and there are other T. c. major clades in addition to the one associated with T. coahuila, making the T. coahuila - T. c. major relationship convoluted. The presence of these main clades in the

GAPD network and the lack of a division between T. o. ornata and T. o. luteola within the T. ornata subgroup in the Cytb network suggests that 1) T. c. carolina and T. c. triunguis form distinct groups and as such their current classification status needs to be

41 amended and 2) T. o. ornata and T. o. luteola are very closely related and they are possibly not divergent enough to be considered separate subspecies.

Population structure

The AMOVAs used to assess the separation or clumping of T. o. ornata and T. o. luteola (Table 4; 7; 11; 14) indicated a low amount of population structure for T. o. ornata and T. o. luteola. This suggests that T. o. luteola and T. o. ornata be clumped into one group as T. ornata. For both Cytb and GAPD, the lowest amount of between-group population structure was found when assigning just two a priori groups consisting of the eastern and western clades (Table 6; 13), supporting that the east/west classification is not adequate for describing the population structure within Terrapene.

T. carolina population structure

The a priori AMOVA based on the presented mtDNA and nucDNA phylogenies

(Table 8; 15) indicates a higher level of population structure than that of the morphological data of Minx (1996; Table 3; 10). Specifically, a higher amount of population structure is seen when assigning the T. c. triunguis - T. c. mexicana - T. c. yucatana clade into a unique group. This suggests that the current classification needs to be revised by elevating the triunguis - mexicana - yucatana clade to a separate species from T. carolina.

The apportionment of T. coahuila as a unique group was supported due to a lower amount of population structure seen when clumping T. coahuila into T. carolina (Table

42 9; 16), which supports the sustaining of T. coahuila as its own species. The population structure for T. c. bauri was higher when represented as a unique group (Table 10; 17), suggesting that this group may be more divergent from T. c. carolina than Minx (1996) hypothesized. These results suggest that T. coahuila should remain a monotypic species and that T. c. bauri may need to be elevated to species status.

SAMOVA analyses

The Cytb SAMOVA showed maximal genetic apportionment when split into 19 groups (Table 18). However, the GAPD SAMOVA apportioned Terrapene into seven groups (Table 19). For GAPD, several taxa shared haplotypes in some cases, which reduced the number of groups. For example, a large haplotype consisting mostly of T. c. triunguis also contained a small number of sequences from T. c. yucatana, T. c. mexicana,

T. c. major, and one T. c. carolina. This was the only haplotype in which T. c. mexicana was found, and was one of only two haplotypes in which T. c. yucatana was found. The

Cytb SAMOVA grouped T. coahuila into its own unique group, but the GAPD

SAMOVA combined T. coahuila with T. c. major from FL. The GAPD SAMOVA placed T. c. major, T. c. bauri, T. c. triunguis - T. c. mexicana - T. c. yucatana, T. c. carolina, and T. nelsoni, into unique groups, while T. o. luteola and T. o. ornata were allocated into the same group for both Cytb and GAPD.

The Cytb gene showed much more variation than GAPD. The increased number of unique groups seen for Cytb suggests that there is a barrier to gene flow within taxa.

43 While this barrier may not have been present long enough to cause deep divergences for intraspecific relationships with nucDNA, it has affected faster evolving genes.

These findings further support 1) separating T. c. triunguis -T. c. mexicana - T. c. yucatana from T. carolina, 2) that T. c. bauri may constitute its own species, 3) that T. coahuila was closely associated with some T. c. major haplotypes, and 4) that T. o. ornata and T. o. luteola should be clumped together as T. ornata (without subspecific designations).

Molecular clock

Several historical geological and climatic events may explain the chronogram and estimated divergence times (Table 3; Figure 7). First, the divergence time for the split between the eastern and western groups occurred ~15 mya. At this time in the middle - late Miocene (Barstovian Age), the climate in central North America (e.g., Kansas and

Nebraska) was becoming warmer, and mesic areas were becoming interspersed with grasslands (Berry, 1918; Chaney and Elias, 1936; Hesse, 1936; ; for a review, see

Axelrod, 1985; Wolfe, 1985). In addition, the earliest known fossil box turtles were found in Barstovian deposits from ~15 mya in Nebraska (Holman and Corner, 1985;

Holman, 1987). Since T. o. ornata are typically a more grassland-oriented species and T. carolina typically inhabit mesic woodlands (Dodd, 2001), it makes sense that the divergence between the eastern and western clades would have occurred ~15 mya in the

Barstovian Age where savannah-like grasslands were becoming more abundant. The earliest fossils resembling T. ornata were also dated to ~14.5 mya and were found in

44 Barstovian deposits (Holman and Fritz, 2005), which further supports the estimated divergence dates given in Table 3 and Figure 7.

T. carolina

In comparison, T. carolina tend to inhabit mesic woodlands (with the exception of

T. coahuila; Dodd, 2001). Progressing towards the late Miocene (~10 mya), much of eastern North America consisted of deciduous forests that was gradually being separated by emerging grassland in the southeast (Graham, 1965; Webb, 1983; Woodburne, 2004), and the northeast was predominantly deciduous forest (Wolfe, 1975; Tiffney, 1985a;

1985b; Graham, 1993; Mai, 1995; Janis et al., 1998; for a review, see Manchester, 1999).

This coincides with the divergence of the T. carolina group because separation from the western group as a result of diverging habitat requirements may have resulted in speciation once box turtles began to migrate eastward. The divergence of the ancestral T. c. carolina lineage ~7 mya in very late Miocene or early Pliocene climatic conditions makes sense geologically due to the woodland habitat seen in the in the northeastern part of the United States and the temperate climate at the time (Woodburne, 2004). Lastly, during the very late Miocene or early Pliocene, climate change caused an increase in provincialism in the North American biomes (Webb, 1977), which could have contributed to the speciation of Terrapene taxa.

T. c. bauri and T. c. major

While T. c. bauri tend to inhabit mesic woodlands, they also are often found in savannah-like biomes (Dodd, 2001). Because Florida developed more mesic habitats

45 over time, it makes sense that T. c. bauri, which originated in savannah and salt marsh biomes that were present ~10 mya in northern Florida and the Gulf Coast (Webb, 1977;

Woodburne, 2004), would be more adaptive with their habitat preferences. In the middle to late Miocene (~12.5 mya - ~5mya), sea levels were generally receding and sediment from the Appalachian Mountains filled the channel separating central and lower Florida from the Gulf Coast (Randazzo and Jones, 1997). This land connection proabably allowed the ancestor to T. c. bauri to migrate southward into useable habitats that were present in southern Florida during the mid-late Miocene and early Pliocene (Wolfe, 1985).

It is also possible that after box turtle migrations to peninsular Florida, vicariance events due to the rising and receding of sea levels resulted in the separation and subsequent speciation of an isolated ancestral population of T. c. bauri. Lastly, the MRCA to T. c. major and T. c. carolina diverged ~7 mya, and this relatively recent divergence makes sense due to some non-monophyly being found between the two taxa.

T. c. triunguis, T. c. mexicana, and T. c. yucatana

The T. c. yucatana lineage was estimated to have diverged from the T. c. triunguis lineage earlier than T. c. mexicana, which is supported by Milstead’s (1969) explanation of the origin of T. c. mexicana. He postulated that T. c. mexicana originated through an intergrade between T. c. yucatana and T. c. triunguis. It is possible that the MRCA of T. c. mexicana and the T. c. triunguis lineage was isolated from the populations in what is now the U.S.A. due to the development of rainforests between the ancestral populations.

(Stehli and Webb, 1985). It is also possible that the MRCA to T. c. yucatana was genetically isolated after migrating to the Yucatan Peninsula via the Isthmus of

46 Tehuantepec land bridge in the Pliocene and Pleistocene from ~8 mya until ~2.5 mya, when fluctuations in sea levels may have caused their isolation (Beard et al., 1982;

Bryant et al., 1991; Mulcahy and Mendelson, 2000; Mulcahy et al., 2006).

Comparisons with published divergence estimates

Divergence estimates for Terrapene have been reported in the literature (Near et al., 2005; Spinks and Shaffer, 2009), but these analyses are mostly focused on intergeneric classification and no studies have been published that comprehensively analyzed the majority of the taxa within the genus. Thus, this chronogram provides the first divergence time estimates for most taxa within Terrpaene and can help in our understanding of the climatic and geographic processes by which these groups diverged.

The pairwise percent divergences also shed light on how divergent the taxa within

Terrapene are.

Percent divergences

Percent divergences were relatively high between T. c. carolina - T. c. triunguis, T. c. carolina - T. c. bauri, T. c. carolina - T. o. ornata, T. o. ornata - T. c. triunguis, and T. o. ornata - T. nelsoni for both mtDNA and nucDNA (Table 20). For only Cytb, percent divergences were relatively high for T. c. triunguis - T. c. yucatana, and for only GAPD T. c. carolina - T. coahuila were relatively divergent. Each of these pairwise comparisons was equivalent to or greater than what Feldman and Parham (2002) and FitzSimmons and

Hart (2007) considered as representing separate species for the mtDNA Cytb gene in

Emydid turtles and nucDNA in freshwater turtles, respectively. In fact, many of these

47 comparisons are in the range for what is typically considered inter-family divergence levels. It is evident that for both the mtDNA and nucDNA these taxa are more divergent than previously believed, and revisions within Terrapene are necessary.

Recommendations for classification revisions

The phylogenetic analyses, the AMOVAs, the haplotype networks, and the percent divergences all suggest that classification revisions are in order for the Terrapene.

With respect to the T. carolina group, all analyses support the splitting of T. c. carolina and T. c. triunguis. I recommend that Terrapene c. triunguis be elevated to full species status as Terrapene mexicana triunguis. Terrapene c. mexicana and T. c. yucatana should be placed within T. mexicana as T. m. mexicana and T. m. yucatana. The species should be named T. mexicana because mexicana was the earliest to be described (Gray,

1849). Terrapene c. bauri is relatively highly divergent from T. c. carolina according to the AMOVA and SAMOVA analyses and percent divergences, but in the phylogenies the clade is not well-supported with bootstrap resampling or BPP. Thus, the data were inconclusive for T. c. bauri. Data were inconclusive for T. c. major as well because of polyphyly. Due to the historical affinity of T. c. major and T. c. bauri with T. c. carolina, these taxa should remain in T. carolina pending further analyses.

It is certainly possible that the polyphyly for T. c. major and incongruences between the Cytb and GAPD phylogenies are due to introgression and hybridization.

Terrapene coahuila were closely associated with T. c. major but were not well-resolved in the phylogenies. The retention of T. coahuila as the sister species to T. carolina was

48 supported by the Cytb haplotype network, the GAPD percent divergences, the Cytb

SAMOVA, and both the Cytb and GAPD AMOVAs. However, due to the lack of resolution for T. coahuila in the phylogenies and the convoluted association with T. c. major, I recommend that T. coahuila maintain its current specific status and remain the sister clade to T. carolina until additional data are available.

Terrapene o. ornata and T. o. luteola are very closely related, and may not be divergent enough to be considered separate subspecies. I recommend that these taxa should be clumped together as T. ornata (sans subspecific designations), which disagrees with Herrmann and Rosen (2009), who found population structuring between T. o. ornata and T. o. luteola. Finally, T. nelsoni should remain the sister species to T. ornata.

It is important to note that these data are based on just one mtDNA and one nucDNA gene, and data for additional genes would be beneficial. Furthermore, sequencing additional genes could be helpful with resolving some of the more ambiguous clades, such as T. c. bauri, T. c. major, and T. coahuila. Although in some instances the analyses may not be well-resolved enough to make sensible classification interpretations

(e.g., T. c. major, T. coahuila, and T. c. bauri), others were clearer. For example, the splitting of T. carolina - T. triunguis was supported by both the Cytb and GAPD phylogenies, percent divergences, AMOVAs, SAMOVAs, and haplotype networks. In addition, the clumping of T. o. ornata and T. o. luteola and the sister relationships of T. ornata and T. nelsoni were supported by all of the aforementioned analyses. The placement of the triunguis subspecies into T. mexicana was supported by all analyses;

49 yucatana, however, was polyphyletic for GAPD. Because 1) the data are limited for yucatana, 2) the Cytb data support placing yucatana into T. mexicana, and 3) previous phylogenetic analyses support the affinities within the mexicana - yucatana - triunguis clade (Milstead, 1969; Minx, 1996), the taxa within this clade should maintain their affinities with one another.

Conservation implications

It is essential to have an understanding of the evolutionary history of a group in need of conservation management, and these data have shed light on some of the evolutionary relationships within Terrapene. Because conservation efforts are typically species-based and tend to ignore subspecies, the splitting of T. carolina (with 6 subspecies) into T. carolina and T. mexicana (with three subspecies each) will be important for future conservation management strategies. In addition, the Cytb

SAMOVA analysis indicated restricted gene flow for some intraspecific populations.

Having an understanding of the underlying genetics of intraspecific populations is very important for successfully facilitating conservation management strategies. As such, further population-level analyses are warranted in order to assess the genetic “health” of individual populations. Parameters including effective population sizes, inbreeding coefficients, and levels of heterozygosity can shed light on the conservation status of

Terrapene, and while some work in this regard has already been done (Kuo and Janzen,

2004; Howeth et al., 2008; Buchman et al., 2009; Cureton et al., 2009), a wider range of populations and geographic localities need to be sampled. Most previous work has focused on limited geographic ranges, and it will be useful to have an understanding of

50 the conservation status of each taxon throughout their range. Thus, while these data provide support for revising the Terrapene classification scheme, there is still much work to be done in terms of finer-scale population genetics.

Future research for Terrapene should focus on two areas. First, additional mtDNA and nucDNA genes need to be sequenced in order to improve sample sizes. This may help to resolve some of the poorly supported clades, and to provide further support for those that are supported. Second, finer-scale population genetic analyses should be performed in order to better assess the conservation status of Terrapene and to provide insight into their underlying genetics so that successful conservation management strategies can be employed. It is my hope that the data within as well as additional data from future work will aid not only in our general understanding of the evolutionary history of Terrapene, but also with their conservation.

Literature Cited

Auffenberg, W. 1958. Fossil turtles of the genus Terrapene in Florida. Bulletin of Florida State Museum of Biological Sciences 3:53-92.

Auffenberg, W. 1959. A Pleistocene Terrapene hibernaculum, with remarks on a second complete box turtle skull from Florida. Quarterly Journal of the Florida Academy of Science 22:49-53.

Auffenberg, W. W. and W. W. Milstead. 1965. Reptiles in the Quarternary of North America. In The Quarternary in the United States, (eds. H. E. Wright, Jr. and D. G. Frey). Princeton Univ. Press, Princeton, New Jersey.

51 Avise, J. C., J. F. Shapira, S. W. Daniel, C. F. Aquadro and R. A. Lansman. 1983. Mitochondrial DNA differentiation during the speciation process in Peromyscus. Molecular Biology and Evolution 1(1):38-56.

Axelrod, D. I. 1985. Rise of the grassland biome, central North America. The Botanical Review 51(2):163-201.

Barrett, M., M. J. Donoghue and E. Sober. 1991. Against consensus. Systematic Zoology 40:486-493.

Beard, J. H., J. B. Sangree and L. A. Smith. 1982. Quaternary chronology, paleoclimate, depositional sequences, and eustatic cycles. American Association of Petroleum Geologists Bulletin 66:158-169.

Berry, E. W. 1918. Fossil from the late Tertiary of Oklahoma. Proceedings of the U.S. National Museum 54:627-636.

Bickham, J. W., T. Lamb, P. Minx and J. C. Patton. 1996. Molecular systematics of the genus Clemmys and the intergeneric relationships of emydid turtles. Herpetologica 52(1):89-97.

Bramble, D. M. 1974. Emydid shell kinesis: Biomechanics and evolution. Copeia 1974:707-727.

Bryant, W. R., J. Lugo, C. Cordova and A. Salvador. 1991. Physiography and Bathymetry. In The Gulf of Mexico Basin: the Geology of North America, (eds. A. Salvador). Geological Society of America, Boulder, CO.

Buchman, A. B., R. Deaton, C. P. Randle, T. Brummel, E. D. Wilson and W. I. Lutterschmidt. 2009. Isolation and characterization of polymorphic microsatellite loci for the three-toed box turtle (Terrapene carolina triunguis) and cross- amplification in other Terrapene species. Molecular ecology resources 9(4):1169- 1171.

Butler, J. M., C. K. Dodd, Jr., M. Aresco and J. D. Austin. 2011. Morphological and molecular evidence indicates that the Gulf Coast box turtle (Terrapene carolina major) is not a distinct evolutionary lineage in the Florida Panhandle. Biological Journal of the Linnaean Society 102:889-901.

52 Carr, A. E. 1940. A contribution to the herpetology of Florida. University of Florida Publication, Biological Science Series 3:100-101.

Carr, A. E. 1952. Handbook of Turtles: The Turtles of the United States and Canada and Baja California. Cornell University Press, Ithaca, New York.

Chaney, R. W. and M. K. Elias. 1936. Late Tertiary floras from the High Plains. Publication of the Carnegie Institue of Washington 476:1-72.

Conant, R. and J. T. Collins. 1991. A Field Guide to the Reptiles and Amphibians of Eastern and Central North America (3rd ed.). Houghton Mifflin, Boston, Massachusetts.

Crandall, K. A. 1994. Intraspecific cladogram estimation - accuracy at higher levels of divergence. Systematic Biology 43(2):222-235.

Cunningham, C. W. 1997. Can three incongruence tests predict when data should be combined? Molecular Biology and Evolution 14(7):733-740.

Cureton, J. C., II, A. B. Buchman, C. P. Randle, W. I. Lutterschmidt and R. Deaton. 2009. Characterization of ten novel microsatellite loci for the threatened , Terrapene ornata. Conservation Genetics Resources 1(1):141-143.

Dodd, K. C. 2001. North American Box Turtles, A Natural History. University of Oklahoma Press, Norman.

Dolman, G. and B. Phillips. 2004. Single copy nuclear DNA markers characterized for comparative phylogeography in Australian wet tropics rainforest skinks. Molecular Ecology Notes 4(2):185-187.

Drummond, A. J. and A. Rambaut. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC evolutionary biology 7:214.

Dupanloup, I., S. Schneider and L. Excoffier. 2002. A simulated annealing approach to define the genetic structure of populations. Molecular Ecology 11(12):2571-2581.

53 Engstrom, T. N., T. Edwards, T. F. Osentoski and E. M. Myers. 2007. A Compendium of PCR Primers for mtDNA, Microsatellite, and Other Nuclear Loci for Freshwater Turtles and Tortoises. Chelonian Research Monographs 4:124-141.

Excoffier, L., G. Laval and S. Schneider. 2005. Arlequin (version 3.0): An integrated software package for population genetics data analysis. Evolutionary Bioinformatics 1:47-50.

Excoffier, L., P. E. Smouse and J. M. Quattro. 1992. Analysis of molecular variance inferred from metric distances among dna haplotypes - application to human mitochondrial-dna restriction data. Genetics 131(2):479-491.

Farris, J. S., M. Kallersjo, A. G. Kluge and C. Bult. 1994. Testing significance of incongruence. Cladistics 10(3):315-319.

Feldman, C. R. and J. F. Parham. 2002. Molecular phylogenetics of emydine turtles: Taxonomic revision and the evolution of shell kinesis. Molecular Phylogenetics and Evolution 22(3):388-98.

Felsenstein, J. 1981. Evolutionary trees from DNA-sequences - a maximum-likelihood approach. Journal of Molecular Evolution 17(6):368-376.

Felsenstein, J. 1985. Confidence-limits on phylogenies - an approach using the bootstrap. Evolution 39(4):783-791.

FitzSimmons, N. N. and K. M. Hart. 2007. Genetic studies of freshwater turtles and tortoises: a review of the past 70 years. Chelonian Research Monographs 4:15-46.

Friesen, V. L., B. C. Congdon, H. E. Walsh and T. P. Birt. 1997. Intron variation in marbled murrelets detected using analyses of single-stranded conformational polymorphisms. Molecular Ecology 6:1047–1058.

Fu, Y. X. and W. H. Li. 1993. Statistical tests of neutrality of mutations. Genetics 133(3):693-709.

Gadagkar, S. R., M. S. Rosenberg and S. Kumar. 2005. Inferring species phylogenies from multiple genes: concatenated sequence tree versus consensus gene tree.

54 Journal of Experimental Zoology (Molecular and Developmental Evolution) 304B:64-74.

Gillette, D. D. 1974. A proposed revision of the evolutionary history of Terrapene carolina triunguis. Copeia 1974(2):537-539.

Graham, A. 1965. Origin and evolution of the biota of southeastern North America: evidence from the fossil record. Evolution 18:571-585.

Graham, A. 1993. History of the Vegetation Cretaceous (Maastrichtian) -Tertiary. In Flora of North America North of Mexico, (eds. Oxford University Press, New York.

Gray, J. E. 1849. Description of a new species of box from Mexico. Proceedings of the Zoological Society of London 17:16-17.

Guindon, S., J.-F. Dufayard, V. Lefort, M. Anisimova, W. Hordijk and O. Gascuel. 2010. New Algorithms and Methods to Estimate Maximum-Likelihood Phylogenies: Assessing the Performance of PhyML 3.0. Systematic Biology 59(3):307-321.

Harrigan, R. J., M. E. Mazza and M. D. Sorenson. 2008. Computation vs. cloning: evaluation of two methods for haplotype determination. Molecular ecology resources 8(6):1239-1248.

Hay, O. P. 1906. Descriptions of two new genera (Echmatemys and Xenochelys) and two new species (Xenochelys formosa and Terrapene putnami) of fossil turtles. Bulletin of the American Museum of Natural History 22:27.

Herrmann, H. and P. C. Rosen. 2009. Conservation of aridlands turtles III: preliminary genetic studies of the desert box turtle and yaqui slider. Sonoran Herpetologist 22(4):38-43.

Hesse, C. J. 1936. Lower Pliocene vertebrate fossils from the Ogallala Formation (Laverne Zone) of Beaver County, Oklahoma. Publication of the Carnegie Institue of Washington 467:47-72.

Hillis, D. M. and J. J. Bull. 1993. An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analyses. Systematic Biology 42:182-192. 55 Holman, J. A. 1987. Herpetofauna of the Egelhoff site (Miocene: Barstovian) of North- Central Nebraska. Journal of Vertebrate Paleontology 1987:109-120.

Holman, J. A. and R. G. Corner. 1985. A Miocene Terrapene (Testudines, Emydidae) and other barstovian turtles from south-central Nebraska. Herpetologica 41(1):88- 93.

Holman, J. A. and U. Fritz. 2005. The box turtle genus Terrapene (Testudines : Emydidae) in the Miocene of the USA. Herpetological Journal 15(2):81-90.

Howeth, J. G., S. E. McGaugh and D. A. Hendrickson. 2008. Contrasting demographic and genetic estimates of dispersal in the endangered : a contemporary approach to conservation. Molecular Ecology 17(19):4209-4221.

Huelsenbeck, J. P. and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17(8):754-755.

Janis, C. M., K. M. Scott and L. L. Jacobs. 1998. The Evolution of Tertiary Mammals in North America. Cambridge University Press, New York.

Jukes, T. H. and C. R. Cantor. 1969. Evolution of protein molecules. In Mammalian Metabolism III, (eds. N. H. Muntro). Academic, New York.

Kuo, C. H. and F. J. Janzen. 2004. Genetic effects of a persistent bottleneck on a natural population of ornate box turtles (Terrapene ornata). Conservation Genetics 5(4):425-437.

Larget, B. and D. L. Simon. 1999. Markov chain Monte Carlo algorithms for the Bayesian analysis of phylogenetic trees. Molecular Biology and Evolution 16(6):750-759.

Legler, J. M. 1960. Natural history of the ornate box turtle, Terrapene ornata ornata Agassiz. University of Kansas Public Museum of Natural History 11:527-669.

Lenk, P., U. Fritz, U. Joger and M. Wink. 1999. Mitochondrial phylogeography of the , Emys orbicularis (Linnaeus 1758). Molecular Ecology 8(11):1911-1922.

56 Librado, P. and J. Rozas. 2009. DnaSP v5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25:1451-1452.

Maddison, W. P. 1997. Gene trees in species trees. Systematic Biology 46(3):523-536.

Maddison, W. P. and D. R. Maddison. 1989. Interactive analysis of phylogeny and character evolution using the computer program MacClade. International Journal Of Primatology 53:190-202.

Mai, D. H. 1995. Tertiare Vegetationsgeschichte Eueropas. Gustav Fischer, Jena, Germany.

Manchester, S. R. 1999. Biogeographical relationships of North American tertiary floras. Annals of the Missouri Botanical Garden 86:472-522.

Mau, B. and M. Newton. 1997. Phylogenetic inference for binary data on dendrograms using Markov chain Monte Carlo. Journal of Computational and Graphical Statistics 6:122-131.

Mau, B., M. A. Newton and B. Larget. 1999. Bayesian phylogenetic inference via Markov chain Monte Carlo methods. Biometrics 55(1):1-12.

Mickevich, M. and J. S. Farris. 1981. The implications of congruence in Menidia. Systematic Zoology 30(3):351-370.

Milstead, W. W. 1967. Fossil box turtles (Terrapene) from central North America, and box turtles of eastern Mexico. Copeia 1967:168-179.

Milstead, W. W. 1969. Studies on the evolution of the box turtles (genus Terrapene). Bulletin of Florida State Museum of Biological Sciences 14:1-113.

Milstead, W. W. and D. W. Tinkle. 1967. Terrapene of Western Mexico, with comments on species groups in the genus. Copeia 1967(1):180-187.

Minx, P. 1992. Variation in phalangeal formulas in the turtle genus Terrapene. Journal of Herpetology 26(2):234-238.

57 Minx, P. 1996. Phylogenetic relationships among the box turtles, Genus Terrapene. Herpetologica 52(4):584-597.

Mulcahy, D. G. and J. R. Mendelson, III. 2000. Phylogeography and speciation of the morphologically variable, widespread species Bufo valliceps, based on molecular evidence from mtDNA. Molecular Phylogenetics and Evolution 17:173-189.

Mulcahy, D. G., B. H. Morrill and J. R. Mendelson, III. 2006. Historical biogeography of lowland species of toads (Bufo) across the Trans- Mexican Neovolcanic Belt and the Isthmus of Tehuantepec. Journal of Biogeography 33:1889-1904.

Near, T. J., P. A. Meylan and H. B. Shaffer. 2005. Assessing concordance of fossil calibration points in molecular clock studies: an example using turtles. The American Naturalist 165(2):137-146.

Neigel, J. E. and J. C. Avise. 1986. Phylogenetic relationships of mitochondrial DNA under various demographic models of speciation. . In Evolutionary Processes and Theory, (eds. E. Nevo and S. Karlin). Academic Press, New York.

Olmstead, R. G. and J. A. Sweere. 1994. Combining data in phylgenetic systematics - an empirical approach using 3 molecular data sets in the Solanaceae. Systematic Biology 43:467-481.

Parham, J. F., C. R. Feldman and J. L. Boore. 2006. The complete mitochondrial genome of the enigmatic bigheaded turtle (Platysternon): description of unusual genomic features and the reconciliation of phylogenetic hypotheses based on mitochondrial and nuclear DNA. BMC evolutionary biology 6:11.

Pattengale, N. D., M. Alipour, O. R. P. Bininda-Emonds, B. M. E. Moret and A. Stamatakis. 2009. How Many Bootstrap Replicates Are Necessary? In Research in Computational Molecular Biology, Proceedings, (eds. S. Batzoglou).

Posada, D. 2008. jModelTest: Phylogenetic model averaging. Molecular Biology and Evolution 25:1253-1256.

Posada, D. and K. A. Crandall. 2001. Intraspecific gene genealogies: trees grafting into networks. Trends in Ecology and Evolution 16(1):37-45.

58 Randazzo, A. F. and D. S. Jones. 1997. The Geology of Florida. University Press of Florida, Gainesville.

Rannala, B. and Z. H. Yang. 1996. Probability distribution of molecular evolutionary trees: A new method of phylogenetic inference. Journal of Molecular Evolution 43(3):304-311.

Rodriguez-Robles, J. A., G. R. Stewart and T. J. Papenfuss. 2001. Mitochondrial DNA- based phylogeography of North American rubber boas, Charina bottae (Serpentes: Boidae). Molecular Phylogenetics and Evolution 18:227-237.

Ronquist, F. and J. P. Huelsenbeck. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19(12):1572-1574.

Rosenberg, N. A. 2002. The probability of topological concordance of gene trees and species trees. Theoretical Population Biology 61(2):225-247.

Saiki, R. K., D. H. Gelfand, S. Stoffel, S. J. Scharf, R. Higuchi, G. T. Horn, K. B. Mullis and H. A. Erlich. 1988. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487-491.

Shaffer, H. B., P. Meylan and M. L. McKnight. 1997. Tests of turtle phylogeny: Molecular, morphological, and paleontological approaches. Systematic Biology 46(2):235-268.

Smouse, P. E. and W. H. Li. 1989. Likelihood analysis of mitochondrial restriction- cleavage patterns for the human chimpanzee-gorilla trichotomy. Evolution 43:1162-1176.

Spinks, P. Q. and H. B. Shaffer. 2009. Conflicting mitochondrial and nuclear phylogenies for the widely disjunct Emys (Testudines: Emydidae) species complex, and what they tell us about biogeography and hybridization. Systematic Biology 58(1):1-20.

Spinks, P. Q., H. B. Shaffer, J. B. Iverson and W. P. McCord. 2004. Phylogenetic hypotheses for the turtle family . Molecular Phylogenetics and Evolution 32(1):164-182.

59 Stehli, F. G. and S. D. Webb. 1985. The Great American Biotic Interchange. Plenum Press, New York.

Stephens, M. and P. Donnelly. 2003. A comparison of bayesian methods for haplotype reconstruction from population genotype data. The American Journal of Human Genetics 73(5):1162-1169.

Stephens, M., N. J. Smith and P. Donnelly. 2001. A new statistical method for haplotype reconstruction from population data. The American Journal of Human Genetics 68(4):978-989.

Stephens, P. R. and J. J. Wiens. 2003. Ecological diversification and phylogeny of emydid turtles. Biological Journal of the Linnean Society 79(4):577-610.

Swofford, D. L. 2003. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4. Sinauer Associates, Sunderland, Massachusetts.

Tajima, F. 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123(3):585-595.

Templeton, A. R. 1998. Nested clade analyses of phylogeographic data: testing hypotheses about gene flow and population history. Molecular Ecology 7(4):381- 397.

Templeton, A. R., K. A. Crandall and C. F. Sing. 1992. A cladistic-analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and dna-sequence data .3. cladogram estimation. Genetics 132(2):619-633.

Tiffney, B. H. 1985a. The Eocene North Atlantic land bridge: its importance in Tertiary and modern phytogeography of the Northern Hemisphere. Journal of the Arnold Arboretum 66(2):243-273.

Tiffney, B. H. 1985b. Perspectives on the origin of the floristic similarity between eastern Asia and eastern North America. Journal of the Arnold Arboretum 66

Webb, S. D. 1977. A history of the savanna vertebrates in the New World. Part I: North America. Annual Review of Ecology and Systematics 8:355-380.

60 Webb, S. D. 1983. The rise and fall of the late Miocene ungulate fauna in North America. In Coevolution, (eds. M. H. Nitecki). University of Chicago Press, Chicago.

Wiens, J. J., C. A. Kuczynski and P. R. Stephens. 2010. Discordant mitochondrial and nuclear gene phylogenies in emydid turtles: implications for speciation and conservation. Biological Journal of the Linnaean Society 99:445-461.

Wolfe, J. A. 1975. Some aspects of plant geography of the Northern Hemisphere during the Late Cretaceous and Tertiary. Annals of the Missouri Botanical Garden 62:264-279.

Wolfe, J. A. 1985. Distribution of major vegetation types during the Tertiary. Geophysical Monograph 32:357–375.

Woodburne, M. O. 2004. Late Cretaceous and Cenozoic Mammals of North America: Biostratigraphy and Geochronology. Columbia University Press, New York.

Wright, S. 1951. The genetical structure of populations. Annals of Eugenetics 15:323-354.

Yang, Z. H. and B. Rannala. 1997. Bayesian phylogenetic inference using DNA sequences: A Markov Chain Monte Carlo method. Molecular Biology and Evolution 14(7):717-724.

Zamudio, K. R. and H. W. Greene. 1997. Phylogeography of the bushmaster (Lachesis muta : Viperidae): implications for neotropical biogeography, systematics, and conservation. Biological Journal of the Linnean Society 62(3):421-442.

61 Tables and Figures

Table 1: The four currently recognized Terrapene species and their associated subspecies, based on Minx (1996). The country of inhabitance for each respective taxon is listed in parentheses beside the taxonomic name (USA = United States of America and Mex = Mexico).

T. carolina T. coahuila T. ornata T. nelsoni carolina (USA) N/A (Mex) ornata (USA) nelsoni (Mex) triunguis (USA) luteola (USA, Mex) klauberi (Mex)

major (USA)

bauri (USA)

mexicana (Mex)

yucatana (Mex)

62 Table 2: GenBank Accession numbers used in this study for both the cytochrome b (Cytb) and glyceraldehyde-3-phosphate dehydrogenase (GAPD) genes.

Taxon Accession # Source(s) Cytb GAPD

Terrapene carolina carolina AF258871 GQ896138 Feldman and Parham, 2002; Wiens et al., 2010

Terrapene carolina triunguis FJ770616 GQ896139 Spinks et al., 2009; Wiens et al., 2010

Terrapene coahuila - AF258872 GQ896140 Feldman and Parham, 2002; Wiens et al., 2010

Terrapene ornata ornata GQ896203 GQ896142 Wiens et al., 2010

Terrapene ornata luteola AF258874 N/A Feldman and Parham, 2002

Terrapene nelsoni nelsoni AF258873 GQ896141 Feldman and Parham, 2002; Wiens et al., 2010

Clemmys guttata FJ770591 GQ896113 Spinks et al., 2009; Wiens et al., 2010

63 Table 3: Mean divergence time estimates, standard deviation (SD), and upper and lower 95% confidence intervals (CI) for the mtDNA cytochrome b gene. Letters refer to specific nodes on Figure 7. Mean SD Upper 95% CI Lower 95% CI A 15.3 0.054 19.5 11.8

B 10.6 0.022 12.8 8.3

C 10.3 0.025 12.7 7.7

D 7.2 0.038 9.6 5.0

E 9.2 0.031 11.4 6.9

F 7.0 0.035 9.2 4.8

G 23.0 0.26 30.6 15.5

64 Table 4: Cytochrome b AMOVA with all 10 species/subspecies apportioned separately. Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI TolAZ Tnels Tcoah TcmAL TccDE TcbFL TctKS Tcy TcMX TooIL TolNM TcmFL TccMA TctMO TooCO TcmLA TccMD TctTX TooNM TcmMS TccPA TctLA TooSD TccME TctAR TooNE TccNY TctMS TooKS TccCT TctOK TooIA TccWV TooTX TccGA TooMO TccOH TooOK TccAL TccKY TccNC TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL ______ΦCT = 0.8594 % of variation explained among groups = 85.94 P < 0.0001

65 Table 5: Cytochrome b AMOVA based on the morphological data of Minx (1996). Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI Tnels Tcoah TccDE TooIL TccMA TooCO TccMD TooNM TccPA TooSD TccME TooNE TccNY TooKS TccCT TooIA TccWV TooTX TccGA TooMO TccOH TooOK TccAL TolAZ TccKY TolNM TccNC TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL TcbFL TcmFL TcmAL TcmLA TcmMS TctKS TctMO TctTX TctLA TctAR TctMS TctOK Tcy TcMX ______ΦCT = 0.6387 % of variation explained among groups = 63.87 P < 0.0001

66 Table 6: Cytochrome b AMOVA with the groups apportioned into eastern and western clades. Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI TccDE TooIL TccMA TooCO TccMD TooNM TccPA TooSD TccME TooNE TccNY TooKS TccCT TooIA TccWV TooTX TccGA TooMO TccOH TooOK TccAL TolAZ TccKY TolNM TccNC Tnels TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL TcbFL TcmFL TcmAL TcmLA TcmMS Tcoah TctKS TctMO TctTX TctLA TctAR TctMS TctOK Tcy TcMX ______ΦCT = 0.6208 % of variation explained among groups = 62.08 P < 0.0001

67 Table 7: Cytochrome b AMOVA with T. o. ornata and T. o. luteola clumped together. Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI Tnels Tcoah TcmAL TccDE TcbFL TctKS Tcy TcMX TooIL TcmFL TccMA TctMO TooCO TcmLA TccMD TctTX TooNM TcmMS TccPA TctLA TooSD TccME TctAR TooNE TccNY TctMS TooKS TccCT TctOK TooIA TccWV TooTX TccGA TooMO TccOH TooOK TccAL TolAZ TccKY TolNM TccNC TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL ______ΦCT = 0.8611 % of variation explained among groups = 86.11 P < 0.0001

68 Table 8: Cytochrome b AMOVA based the mtDNA and nucDNA phylogenies (Figures 4- 5). Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI Tnels Tcoah TccDE TctKS TooIL TccMA TctMO TooCO TccMD TctTX TooNM TccPA TctLA TooSD TccME TctAR TooNE TccNY TctMS TooKS TccCT TctOK TooIA TccWV Tcy TooTX TccGA TcMX TooMO TccOH TooOK TccAL TolAZ TccKY TolNM TccNC TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL TcbFL TcmFL TcmAL TcmLA TcmMS ______ΦCT = 0.7979 % of variation explained among groups = 79.79 P < 0.0001

69 Table 9: Cytochrome b AMOVA apportioned based on the mtDNA and nucDNA phylogenies (Figures 4-5), with the exception of T. coahuila being grouped with T. carolina. Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI Tnels TccDE TctKS TooIL TccMA TctMO TooCO TccMD TctTX TooNM TccPA TctLA TooSD TccME TctAR TooNE TccNY TctMS TooKS TccCT TctOK TooIA TccWV Tcy TooTX TccGA TcMX TooMO TccOH TooOK TccAL TolAZ TccKY TolNM TccNC TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL TcbFL TcmFL TcmAL TcmLA TcmMS Tcoah ______ΦCT = 0.7781 % of variation explained among groups = 77.81 P < 0.0001

70 Table 10: Cytochrome b AMOVA apportioned based on the mtDNA and nucDNA phylogenies (Figures 4-5), with the exception of T. c. bauri being placed into a unique group. Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI Tnels Tcoah TcbFL TccDE TctKS TooIL TccMA TctMO TooCO TccMD TctTX TooNM TccPA TctLA TooSD TccME TctAR TooNE TccNY TctMS TooKS TccCT TctOK TooIA TccWV Tcy TooTX TccGA TcMX TooMO TccOH TooOK TccAL TolAZ TccKY TolNM TccNC TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL TcmFL TcmAL TcmLA TcmMS ______ΦCT = 0.8220 % of variation explained among groups = 82.20 P < 0.0001

71 Table 11: Glyceraldehyde-3-phosphate dehydrogenase AMOVA with all 10 species/subspecies separated into distinct groups. Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI TolAZ Tnels Tcoah TcmAL TccDE TcbFL TctKS Tcy TcMX TooCO TolNM TcmFL TccMA TctMO TooNM TcmLA TccMD TctTX TooSD TcmMS TccPA TctLA TooNE TccME TctAR TooKS TccNY TctMS TooIA TccCT TctOK TooTX TccWV TooOK TccGA TccOH TccAL TccKY TccNC TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL ______ΦCT = 0.7860 % of variation explained among groups = 78.60 P < 0.0001

72 Table 12: Glyceraldehyde-3-phosphate dehydrogenase AMOVA based on morphological data of Minx (1996). Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI Tnels Tcoah TccDE TooCO TccMA TooNM TccMD TooSD TccPA TooNE TccME TooKS TccNY TooIA TccCT TooTX TccWV TooOK TccGA TolAZ TccOH TolNM TccAL TccKY TccNC TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL TcbFL TcmFL TcmAL TcmLA TcmMS TctKS TctMO TctTX TctLA TctAR TctMS TctOK Tcy TcMX ______ΦCT = 0.7019 % of variation explained among groups = 70.19 P < 0.0001

73 Table 13: Glyceraldehyde-3-phosphate dehydrogenase AMOVA with the groups apportioned into eastern and western clades. Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI TccDE TooIL TccMA TooCO TccMD TooNM TccPA TooSD TccME TooNE TccNY TooKS TccCT TooIA TccWV TooTX TccGA TooMO TccOH TooOK TccAL TolAZ TccKY TolNM TccNC Tnels TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL TcbFL TcmFL TcmAL TcmLA TcmMS Tcoah TctKS TctMO TctTX TctLA TctAR TctMS TctOK Tcy TcMX ______ΦCT = 0.6732 % of variation explained among groups = 67.32 P < 0.0001

74 Table 14: Glyceraldehyde-3-phosphate dehydrogenase AMOVA with T. o. ornata and T. o. luteola combined into one group. Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI Tnels Tcoah TcmAL TccDE TcbFL TctKS Tcy TcMX TooCO TcmFL TccMA TctMO TooNM TcmLA TccMD TctTX TooSD TcmMS TccPA TctLA TooNE TccME TctAR TooKS TccNY TctMS TooIA TccCT TctOK TooTX TccWV TooOK TccGA TolAZ TccOH TolNM TccAL TccKY TccNC TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL ______ΦCT = 0.7905 % of variation explained among groups = 79.05 P < 0.0001

75 Table 15: Glyceraldehyde-3-phosphate dehydrogenase AMOVA based on the mtDNA and nucDNA phylogenies (Figures 4-5). Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI Tnels Tcoah TccDE TctKS TooCO TccMA TctMO TooNM TccMD TctTX TooSD TccPA TctLA TooNE TccME TctAR TooKS TccNY TctMS TooIA TccCT TctOK TooTX TccWV Tcy TooOK TccGA TcMX TolAZ TccOH TolNM TccAL TccKY TccNC TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL TcbFL TcmFL TcmAL TcmLA TcmMS ______ΦCT = 0.7444 % of variation explained among groups = 74.44 P < 0.0001

76 Table 16: Glyceraldehyde-3-phosphate dehydrogenase AMOVA apportioned based on the mtDNA and nucDNA phylogenies (Figures 4-5), with the exception of T. coahuila being grouped with T. carolina. Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI Tnels TccDE TctKS TooIL TccMA TctMO TooCO TccMD TctTX TooNM TccPA TctLA TooSD TccME TctAR TooNE TccNY TctMS TooKS TccCT TctOK TooIA TccWV Tcy TooTX TccGA TcMX TooMO TccOH TooOK TccAL TolAZ TccKY TolNM TccNC TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL TcbFL TcmFL TcmAL TcmLA TcmMS Tcoah ______ΦCT = 0.7169 % of variation explained among groups = 71.69 P < 0.0001

77 Table 17: Glyceraldehyde-3-phosphate dehydrogenase AMOVA apportioned based on the mtDNA and nucDNA phylogenies (Figures 4-5), with the exception of T. c. bauri being placed into a unique group. Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI Tnels Tcoah TcbFL TccDE TctKS TooCO TccMA TctMO TooNM TccMD TctTX TooSD TccPA TctLA TooNE TccME TctAR TooKS TccNY TctMS TooIA TccCT TctOK TooTX TccWV Tcy TooOK TccGA TcMX TolAZ TccOH TolNM TccAL TccKY TccNC TccIN TccTN TccVA TccSC TccMI TccRI TccNJ TccMS TccIL TcmFL TcmAL TcmLA TcmMS ______ΦCT = 0.7644 % of variation explained among groups = 76.44 P < 0.0001

78 Table 18: Cytochrome b SAMOVA with the populations maximally differentiated into 19 groups. Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TccKY TccIL TccGA TccDE TccVA TccWV TcbFL TctLA TccNC TccTN TccAL TccMA TctKS TccSC TccMD TctMO TccMS TccPA TctAR TccIN TccME TctMS TccOH TccNY TctOK TccMI TccCT TccRI TccNJ ______TctTX TcMX Tcy Tcoah TcmLA TcmAL TcmFL Tnels TolAZ TcmMS TooNM

TolNM TooTX ______TooCO TooWI TooSD TooKS TooNE TooOK TooIA TooMO ______

ΦCT = 0.9101 % of variation explained among groups = 91.01 P < 0.0001

79 Table 19: Glyceraldehyde-3-phosphate dehydrogenase SAMOVA with the populations maximally differentiated into seven groups. Samples were labeled by scientific name followed by the state from which they were collected. Too = T. o. ornata, Tol = T. o. luteola, Tnels = T. nelsoni, Tcoah = T. coahuila, Tcm = T. c. major, Tcc = T. c. carolina, Tcb = T. c. bauri, Tct = T. c. triunguis, Tcy = T. c. yucatana, and TcMX = T. c. mexicana. ______TooWI Tnels Tcoah TcmAL TccDE TcbFL TctKS TooCO TcmFL TcmMS TccMA TctMO TooNM TccMD TctTX TooSD TccPA TctLA TooNE TccME TctAR TooKS TccNY TctMS TooIA TccCT TctOK TooTX TccWV TcMX TolAZ TccGA Tcy TolNM TccOH TccSC TccAL TooOK TccKY TcmLA TccNC TccIN TccTN TccVA TccMI TccRI TccNJ TccMS TccIL ______ΦCT = 0.8278 % of variation explained among groups = 82.78 P < 0.0001

80 Table 20: Percent divergences calculated using Jukes Cantor nucleotide divergences corrected for comparing populations. Cytochrome b GAPD

carolina - bauri 6.10 0.965

carolina - coahuila 0.583 1.02

carolina - triunguis 6.71 0.927

triunguis - mexicana 2.56 0.233

triunguis - yucatana 6.59 0.567

ornata - carolina 6.83 1.34

ornata - nelsoni 6.08 3.09

ornata - luteola 0.583 0.404

ornata - triunguis 7.34 1.12

81

Figure 1: The Terrapene phylogeny of Minx (1996) based on 32 morphological characters. The phylogeny includes all four currently recognized species and all 11 currently recognized subspecies of Terrapene.

82

Figure 2: Phylogenies from previous phylogenetic studies assessing the classification of the subfamily Emydinae, including the four currently recognized species of Terrapene (Gaffney and Meylan, 1988; Bickham et al., 1996; Burke et al., 1996; Feldman and Parham, 2002; Stephens and Wiens, 2003).

83

Figure 3: Sampling localties for the mitochondrial DNA (mtDNA) cytochrome b gene and the nuclear DNA (nucDNA) glyceraldehyde-3-phosphate dehydrogenase (GAPD) gene.

84

Figure 4: The cytochrome b phylogram, generated using PhyML 3.0, consising of 253 sequences distributed into 103 haplotypes. As determined by the Akaike Information Criterion corrected for small sample size (AICc), TPM2uf + I + G was used as the nucleotide substitution model, with I = 0.4450 and G = 0.6160. The rate class and base frequency parameters were as follows: AC = 3.2798, AG = 21.7809, AT = 3.2798, CG = 1.0000, CT = 21.7809, GT = 1.0000; freqA = 0.3059, freqC = 0.3131, freqG = 0.1213, freqT = 0.2598, and sample size = 1,097. Bayesian posterior probabilities (BPP; above branches) and non-parametric bootstrap resampling values (below branches) were considered supported at ≥ 95% and ≥ 70%, respectively.

85

Figure 5: The glyceraldehyde-3-phosphate dehydrogenase (GAPD) phylogram, generated using PhyML 3.0, consisting of 202 sequences distributed into 59 haplotypes. As determined by the Akaike Information Criterion corrected for small sample size (AICc), K80 + I was used as the substitution model. The following parameters were used for phylogenetic inference: I = 0.7080, kappa = 3.8280 (Ti/Tv = 1.9140), and sample size = 430. Bayesian posterior probabilities (BPP; above branches) and non-parametric bootstrap resampling values (below branches) were considered supported at ≥ 95% and ≥ 70%, respectively.

86

Figure 6: A combined mtDNA and nucDNA phyogram. The Cytb and GAPD gene regions were partitioned separately, and the substitution model parameters for each partition were kept the same as they were for the corresponding individual analyses (Figure 4-5). Bayesian posterior probabilities (BPP) were inferred as support values, with BPP ≥ 0.95 considered well-supported.

87

Figure 7: A mtDNA cytochrome b (Cytb) chronogram in millions of years, generated in BEAST v1.6.2. The nucleotide substitution model was GTR + I + G, with I = 0.4450 and G = 0.6160. The rate class and base frequency parameters were as follows: AC = 3.2798, AG = 21.7809, AT = 3.2798, CG = 1.0000, CT = 21.7809, GT = 1.0000; freqA = 0.3059, freqC = 0.3131, freqG = 0.1213, freqT = 0.2598, and sample size = 1,097. Bayesian posterior probabilities (BPP; above branches) were considered supported at ≥ 95%. Fossil data and published divergence time data were used for node calibration (black squares). The root (MRCA to Emydinae) was calibrated to 29.4 mya with a standard deviation (SD) of 2.01, the MRCA to Glyptemys was calibrated to 17.0 mya with a SD of 2.9, the MRCA to T. carolina was set to 10.0 mya with a SD of 1.0, and the MRCA to T. ornata was set to 12.5 mya with a SD of 1.0. The grey bars represent the upper and lower 95% confidence intervals. The letters refer to estimated divergence times in Table 3. 88

Figure 8: Cytochrome b (Cytb) haplotype network with nine subgroups calculated using the 95% statistical parsimony method. Haplotypes are smaller when containing fewer sequences and larger when containing more sequences. Each branch represents a single mutational step, and the small black dots represent missing intermediate steps.

89

90

Figure 9: Glyceraldehyde-3-phosphate dehydrogenase (GAPD) haplotype network with two subgroups calculated using the 95% statistical parsimony method. Haplotypes are smaller when containing fewer sequences and larger when containing more sequences. Each branch represents a single mutational step, and the small black dots represent missing intermediate steps.

Appendices

Appendix A: A supplementary table listing sampling information included in the analyses for this study. The data are given for the respective subspecies (as identified by the collectors); locality data (state, county, and geographic coordinates); sample size (n) for the mitochondrial (mt) nuclear (nuc; Figure 4-5), and combined mtDNA and nucDNA phylogeny (Figure 6); the haplotype numbers (Hap#) associated with each phylogeny (Figure 4-6); and the collectors who provided the tissue samples. Abbreviations for each collector are as follows, with their organizational affiliations (if provided) listed in parentheses: Lori Erb (LE), Lori Johnson, Liz Willey, and Catalina-Lopez Ospina (CLO) with the MA Division of Fisheries and Wildlife (MDFW), Danielle O’Delle (DO) with the Nantucket Conservation Foundation (NCF), Scott Smyers (SS) with the Oxbow Associates (OA), James Lee with the Camp Shelby Nature Conservancy (CSNC), Nathan Nazdrowicz with the Ashland Nature Center (ANC), Roger D. Birkhead (RDB), W. Birkhead (WB), Zach Felix (ZF), Susan Matthews (SM) with the Jug Bay Wetlands Sanctuary (JBWS), the Wildlife Center of VA (WCVA), the Illinois Natural History Survey (INHS), John Palis (JP), Ellen Emmerich (EE), Erin Smithies-Baker (ESB), Warren Duzak (WD), G.J. Watkins (GJW), A.A. Leenders (AAL), B.T. Roach (BTR), L Colwell (LC), Mike Martin (MM), Chris Tabaka with the Binder Park Zoo (BPZ), Aaron Gooley (AG), Jane Wyche (JW) with the Merchant’s Millpond NC State Park (MMSP), Ann Summers (AS) with the University of North Carolina at Greensboro (UNCG), Jerry Reynolds (JR) with the NC State Museum of Natural History (NCMNS), Pete Senchyshak (PS), Garry Brian (GB), Andrew Townsend (AT), Harriet Forrester (HF), Dennis Quinn (DQ), the RI Department of Environmental Management, Division of Wildlife and Fisheries (RIDWF), Joe McGavin (JG), Sally Ray (SR), Jonathan Mays (JM), Jim Koukl (JK), Angela Roe Frost (ARF), Alan Byboth (AB), Suellyn Martin (SMa), Michael Brodt (MB), Joshua Jagels (JJ) with the KS Department of Wildlife and Parks (KDWP), Amity Bass (A_Bass), Carl Franklin (CF), Steve Shively (SSm), Beth Millig (BM), Becky Rosamond (BR) with the U.S. and Wildlife Service (USFW), Moses Michelsohn (MM), the Museum of Vertebrate Zoology (MVZ) with the University of California at Berkely (UCB), Kenneth Krysko (KK), Matt Greene (MG), David Steen (DS), Matt Aresco (MA), V. Jo (VJ), Sean Graham (SG), Daniel Parker (DP), Anthony Flanagan (AF), Paul E. Moler (PEM), Deanne Molar (DM), Bob Hay (BH), Ann-Elizabeth Nash (AEN), Alessandra Higa (AH), John Iverson (JI), Greg Pauly (GP), Ian Murray (IM), Lisa Schmidt (LS), Neil Bernstein (NB), Fred Janzen (FJ), Renn Tumlison (RT), Tom deMaar (TdM) with the Gladys Porter Zoo (GPS), the Arizona- Sonora Desert Museum (ASDM), Zoo Atlanta (ZA), and Joe Mendelson (JMen). The state designations for the Mexican taxa are as follows: Cuatro Cienegas (CC), Sonora (SA), Tamaulipas (TL), and the Yucatan Peninsula (YU).

91

Appendix A: (Continued)

Subspecies State County Coordinates n (mt) n (nuc) n (comb.) mt Hap# nuc Hap# comb. Hap# Collector(s) carolina MA Hampshire 42.26 N, -72.52 W 1 1 1 28 12 33 LE (MDFW) carolina MA Hampshire 42.34 N, -72.66 W 4 4 1 28 1, 6, 8, 9 24, 27, 28, 38 LE, LJ (MDFW) carolina MA Barnstable 41.76 N, -70.49 W 2 1 1 28 1 24 LE (MDFW) 2, 3, 7, 8, 9, carolina MA Barnstable 41.94 N, -70.03 W 9 7 7 28 26, 27, 32, 35, 38, 43, 47 LE, CLO (MDFW) 11, 39 carolina MA Barnstable 41.83 N, -69.97 W 3 2 2 28 1 24 LE (MDFW) carolina MA Barnstable 41.99 N, -70.05 W 1 1 1 28 2 35 LE (MDFW) carolina MA Hampden 42.10 N, -71.32 W 2 2 2 28 9, 12 27, 33 LE (MDFW) carolina MA Hampden 42.14 N, -72.63 W 1 2 1 28 13, 16 46 LE, LW (MDFW)

92 carolina MA Bristol 41.77 N, -71.03 W 1 1 1 28 9 27 LE (MDFW)

carolina MA Plymouth 41.66 N, -70.82 W 3 2 2 28, 60 1 24, 25 LE, CLO (MDFW) carolina MA Plymouth 41.99 N, -70.74 W 3 3 3 28, 58, 59 2, 9, 39 27, 44, LE, CLO (MDFW) LE (MDFW), DO (NCF), carolina MA Nantucket 41.28 N, -70.08 W 2 2 2 28 9, 11 25, 36 SS (OA) carolina MA Suffolk 42.36 N, -71.06 W 0 1 0 - 39 - LE, LW (MDFW) carolina MS Toshimingo 34.74 N, -88.24 W 1 1 1 76 24 77 JL (CSNC) carolina DE New Castle 39.80 N, -75.66 W 4 4 4 28, 77 1, 40, 41, 43 41, 42, 24, 29 NN (ANC) carolina AL Tallapoosa 32.87 N, -85.81 W 2 2 2 25 8 39, 84 RDB carolina AL Tallapoosa 32.87 N, -85.82 W 1 0 0 35 - - RDB carolina AL Lowndes 32.21 N, -86.55 W 1 0 0 25 - - RDB carolina AL Russell 32.47 N, -85.20 W 1 1 1 25 12 34 RDB carolina AL Russell 32.26 N, -85.35 W 1 0 0 35 - - RDB carolina AL Russell 32.26 N, -85.42 W 1 1 1 103 12 86 RDB

Appendix A: (Continued)

Subspecies State County Coordinates n (mt) n (nuc) n (comb.) mt Hap# nuc Hap# comb. Hap# Collector(s) carolina AL Macon 32.51 N, -85.61 W 3 1 1 30, 35, 37 38 76 RDB carolina AL Macon 32.45 N, -85.81 W 1 0 0 39 - - RDB carolina AL Elmore 32.57 N, -86.03 W 1 0 0 30 - - RDB carolina AL Lee 32.54 N, -85.59 W 1 0 0 35 - - RDB carolina AL Bullock 32.08 N, -85.69 W 1 1 1 35 9 51 RDB carolina AL Bullock 32.20 N, -85.50 W 0 1 0 - 9 - RDB carolina AL Barbour 32.03 N, -85.09 W 1 1 1 35 50 53 RDB carolina AL Barbour 31.88 N, -85.46 W 0 1 0 - 9 - RDB carolina GA Calhoun 31.53 N, -84.60 W 0 1 0 - 51 - RDB carolina GA Harris 32.76 N, -84.88 W 2 2 2 30 41 48, 85 WB carolina GA Harris 32.63 N, -85.00 W 1 1 76 25 81 WB

93 carolina GA Cherokee 34.23 N, -84.49 W 1 0 0 70 - - ZF

carolina MD Anne Arudel 38.78 N, -76.70 W 4 6 2 26, 28 24, 34 24, 78 SM (JBWS) carolina VA Albermarle 38.03 N, -78.56 W 3 2 2 34, 36 18, 38 73, 74 WCVA carolina VA Augusta 38.20 N, -79.12 W 1 0 0 17 - - WCVA carolina VA Fluvanna 37.84 N, -78.28 W 1 1 1 33 1 72 WCVA carolina IL Will 41.44 N, -87.98 W 1 2 1 32 1, 42 67 INHS carolina IL Clinton 38.61 N, -89.42 W 1 2 1 15 38, 49 66 INHS carolina IL Williamson 37.73 N, -89.06 W 1 1 1 20 39 75 JP carolina NY Ulster 41.89 N, -74.26 W 1 0 0 28 - - EE carolina NY Orange 41.51 N, -74.05 W 1 1 1 28 40 41 ESB carolina NY Sullivan 41.72 N, -74.76 W 1 1 1 28 2 35 ESB carolina NY Westchester 41.15 N, -73.76 W 1 1 1 76 43 61 ESB carolina NY Westchester 41.29 N, -73.87 W 0 1 0 - 16 - ESB

Appendix A: (Continued)

Subspecies State County Coordinates n (mt) n (nuc) n (comb.) mt Hap# nuc Hap# comb. Hap# Collector(s) carolina TN Davidson 36.17 N, -86.78 W 2 2 18, 35 12, 13 52, 64 WD carolina KY Laurel 37.11 N, -84.12 W 1 1 1 39 2 71 GJW, AAL, BTR, LC carolina KY Leslie 37.09 N, -83.38 W 1 1 1 19 12 65 GJW, AAL, BTR carolina KY Carter 38.33 N, -83.05 W 1 1 1 87 12 56 GJW, AAL, BTR, LC carolina SC Beaufort 32.33 N, -80.70 W 4 4 3 85. 89 10, 28 78, 80 MM carolina SC Port Royal 32.40 N, -80.70 W 1 0 0 89 - - MM carolina SC Beaufort 32.35 N, -80.71 W 1 1 0 89 37 - MM carolina SC Jasper 32.45 N, -81.11 W 1 0 0 89 - - MM carolina MI Kalamazoo 42.24 N, -85.53 W 2 1 0 89 40 - CT (BPZ) carolina MI Calhoun 42.25 N, -85.00 W 3 2 2 89 38, 43 60, 63 CT (BPZ) carolina WV Lewis 38.91 N, -80.58 W 2 1 1 28, 39 1 24 AG

94 carolina WV Hampshire 39.32 N, -78.46 W 1 1 1 35 9 51 AG

carolina WV Cabell 38.58 N, -82.26 W 1 1 1 39 24 69 AG carolina WV Roane 38.54 N, -81.33 W 1 1 1 39 1 70 AG carolina NC Gatesville 36.44 N, -76.68 W 1 1 1 86 18 49 JW (MMSP) carolina NC Orange 36.04 N, -79.20 W 1 0 0 89 - - AS (UNCG) carolina NC Johnston 35.68 N, -78.47 W 1 2 1 90 12, 41 87 JR (NCMNS) carolina OH Ross 39.23 N, -82.98 W 1 0 0 39 - - AG carolina OH Lawrence 38.41 N, -82.48 W 1 1 1 39 1 70 AG carolina OH Meigs 39.04 N, -82.13 W 1 1 1 89 39 62 PS carolina OH Washington 39.42 N, -81.45 W 1 1 1 89 61 59 PS carolina IN Morgan 39.43 N, -86.42 W 2 2 2 39, 89 43, 60 57, 68 GB carolina IN Scott 38.74 N, -85.80 W 3 2 2 89, 106 13, 63 55, 58 AT carolina IN Jefferson 38.76 N, -85.42 W 2 2 2 89 61 59 AT

Appendix A: (Continued)

Subspecies State County Coordinates n (mt) n (nuc) n (comb.) mt Hap# nuc Hap# comb. Hap# Collector(s) carolina NJ - 40.00 N, -74.50 W 3 5 3 28, 76 3, 33, 38, 42 26, 79 HF carolina CT Fairfield 41.49 N, -73.42 W 6 4 4 28, 108 1, 29, 34, 43 24, 30, 31, 40 DQ carolina RI Kent 41.67 N, -71.64 W 1 1 1 28 1 24 RIDWF carolina RI Kent 41.63 N, -71.66 W 1 1 1 28 60 45 RIDWF carolina RI Washington 41.49 N, -71.67 W 2 2 2 28 38, 39 37, 43 RIDWF carolina PA Carbou 40.87 N, -75.74 W 2 2 2 28 1, 38 24, 37 JMG carolina PA York 40.07 N, -76.88 W 2 1 1 28, 89 62 54 SR carolina ME Knox 44.26 N, -69.38 W 1 1 1 28 38 37 JM triunguis TX Dallas 32.97 N, -96.72 W 4 2 2 1, 3, 4, 6 10, 27 115, 116 JK triunguis TX Smith 32.26 N, -95.20 W 3 0 0 5, 13, 14 - - JK, ARF triunguis TX Smith 32.30 N, -95.21 W 1 1 1 14 53 117 AB, JK

95 triunguis TX Smith 32.23 N, -95.17 W 0 1 0 - 57 - JK

triunguis TX Collin 33.18 N, -96.58 W 1 0 0 14 - - JK, SMa triunguis TX Red River 33.62 N, -95.05 W 1 0 0 78 - - ? triunguis MO Jefferson 38.22 N, -90.40 W 3 3 3 7, 109, 110 51, 54 107, 113, 118 MB triunguis MO St. Louis 38.56 N, -90.38 W 1 0 0 9 - - MB triunguis KS Crawford 37.58 N, -94.97 W 1 0 0 8 - - JJ (KDWP) triunguis KS Crawford 37.56 N, -94.92 W 2 1 1 11, 12 10 104 JJ (KDWP) triunguis KS Crawford 37.65 N, -94.87 W 0 1 0 - 51 - JJ (KDWP) triunguis KS Crawford 37.60 N, -95.00 W 0 1 0 - 51 - JJ (KDWP) triunguis LA Morehouse 32.71 N, -91.93 W 1 1 1 8 27 105 A_Bass triunguis LA Beauregard 30.65 N, -93.34 W 1 0 0 63 - - CF triunguis LA Vernon 30.99 N, -93.16 W 1 1 1 66 51 112 SSm triunguis LA Rapides 31.21 N, -92.59 W 1 1 1 68 51 114 SSm

Appendix A: (Continued)

Subspecies State County Coordinates n (mt) n (nuc) n (comb.) mt Hap# nuc Hap# comb. Hap# Collector(s) triunguis LA Rapides 31.20 N, -92.58 W 0 1 0 - 52 - SSm triunguis LA Grant 31.52 N, -92.53 W 0 1 0 - 10 - SSm triunguis AR Pulaski 34.83 N, -92.49 W 5 4 4 8, 10, 96, 97 10, 27 103, 109, 110 BM triunguis MS Perry 31.19 N, -89.20 W 1 1 1 95 18 106 JL (CSNC) triunguis MS Grenada 33.78 N, -90.04 W 0 1 0 - 51 - BR (USFW) triunguis OK - 35.22 N, -97.42 W 2 4 2 78, 80 10, 53 108 DL major FL Franklin 29.94 N, -85.01 W 1 1 1 16 22 89 Mmich major FL Franklin 29.97 N, -84.98 W 1 0 0 22 - - Mmich major FL Franklin 29.80 N, -84.83 W 1 1 1 30 22 82 Mmich major FL Wakulla 30.06 N, -84.57 W 1 1 1 24 22 50 MVZ (UCB) major FL Holmes 30.87 N, -85.81 W 1 0 0 22 - - KK

96 major FL Walton 30.44 N, -85.95 W 1 0 0 22 - - MG

major FL Walton 30.49 N, -85.94 W 1 0 0 71 - - MA major FL Okaloosa 30.67 N, -86.63 W 1 0 0 35 - - DS major FL Okaloosa 30.75 N, -86.56 W 2 0 0 76 - - DS major FL Gulf 29.90 N, -85.24 W 1 1 1 69 22 83 KK major FL Gulf 30.06 N, -85.19 W 1 0 0 69 - - WB major FL Gulf 29.82 N, -85.28 W 1 0 0 76 - - WB major FL Escambia 30.61 N, -87.33 W 1 0 0 72 - - KK major AL Houston 31.24 N, -85.12 W 1 1 1 22 1 88 RDB, VJ major AL Barbour 32.01 N, -85.40 W 1 0 0 22 - - SG major AL Baldwin 30.63 N, -87.91 W 2 2 2 72 55, 56 95, 96 RDB major AL Mobile 30.55 N, -88.12 W 1 1 1 102 32 94 RDB major LA Concordia 31.77 N, -91.44 W 1 1 1 73 55 90 A_Bass

Appendix A: (Continued)

Subspecies State County Coordinates n (mt) n (nuc) n (comb.) mt Hap# nuc Hap# comb. Hap# Collector(s) major MS Perry 31.19 N, -89.20 W 2 2 2 72, 92 10, 32 92, 97 JL (CSNC) hybrid MS Forrest 31.19 N, -89.20 W 1 1 1 93, 101 21. 66 93 JL (CSNC) bauri FL Alachua 29.64 N, -82.35 W 1 1 1 74 19 99 WB bauri FL Hernando 28.59 N, -82.37 W 1 1 1 74 47 101 KK, DP bauri FL Hendry 26.60 N, -80.98 W 1 1 1 75 19 100 KK, AF bauri FL Taylor 29.77 N, -83.57 W 1 1 1 74 10 102 KK, PEM, DM ornata WI Columbia 43.45 N, -89.35 W 2 1 1 41, 55 48 19 BH ornata WI Iowa 42.96 N, -90.13 W 4 1 1 48, 54, 55 48 18 BH ornata WI Dane 43.19 N, -89.77 W 1 0 0 49 - - BH ornata WI Sauk 43.18 N, -90.10 W 3 2 2 50, 54, 55 48 18, 19 BH ornata CO Weld 40.29 N, -104.48 W 8 4 4 40, 53, 99 4 5 AEN

97 ornata CO Jefferson 39.59 N, -105.25 W 2 1 1 51, 53 48 6 AEN

ornata CO El Paso 38.84 N, -104.52 W 1 0 0 52 - - AEN ornata SD Pennington 44.08 N, -103.23 W 5 3 3 23 48, 49 9, 10 AH ornata NE Garden 41.62 N, -102.34 W 3 3 3 23, 27 4 11, 13, 14 JI ornata NE Sheridan 42.50 N, -102.43 W 1 1 1 23 4 12 JI ornata TX Montague 33.67 N, -97.73 W 1 1 1 57 4 13 CF ornata TX Stonewall 33.18 N, -100.25 W 1 1 1 57 20 14 GP ornata TX Stonewall 33.27 N, -100.25 W 1 1 1 57 4 13 GP ornata TX Winkler 31.86 N, -103.05 W 0 1 0 - 4 - MVZ (UCB) ornata NM Chaves 33.36 N, -104.47 W 1 1 1 44 45 12 IM ornata NM San Miguel 35.47 N, -104.83 W 0 1 0 - 4 - IM ornata KS Douglas 38.76 N, -95.15 W 1 1 1 80 4 20 LS ornata KS Meade 37.29 N, -100.37 W 1 0 0 42 4 - JJ (KDWP)

Appendix A: (Continued)

Subspecies State County Coordinates n (mt) n (nuc) n (comb.) mt Hap# nuc Hap# comb. Hap# Collector(s) ornata KS Meade 37.17 N, -100.45 W 1 1 1 80 4 20 JJ (KDWP) ornata KS Meade 37.38 N, -100.17 W 1 1 1 80 4 20 JJ (KDWP) ornata KS Shawnee 38.91 N, -95.56 W 1 1 1 80 4 20 LS ornata IA Johnson 41.67 N, -91.59 W 2 3 2 81, 82 4, 48 3, 4 NB ornata IL Carroll 41.96 N, -90.10 W 4 0 0 54, 99, 100 - - FJ ornata MO Boone 38.94 N, -92.33 W 1 0 0 54 - - WB hybrid OK - 35.22 N, -97.42 W 1 1 1 80 64 21 DL hybrid AR Conway 35.26 N, -92.70 W 1 1 1 107 65 111 RT luteola NM Socorro 34.02 N, -106.93 W 5 4 4 43, 45 46, 47 4, 46, 59 11, 15, 16, 17 IM luteola AZ Cochise 31.63 N, -109.20 W 2 2 2 83, 84 4 1, 2 MVZ (UCB) luteola AZ Cochise 31.76 N, -109.26 W 1 1 1 84 4 1 MVZ (UCB)

98 luteola AZ Cochise 31.91 N, -109.15 W 0 1 0 - 4 - MVZ (UCB)

coahuila CC - 26.99 N, -102.07 W 4 4 3 94 22 98 TdM (GPS) nelsoni SA - 29.65 N, -110.87 W 1 1 1 105 67 23 ASDM mexicana TL - 26.40 N, -99.03 W 6 5 4 65, 98 10 119, 120 Jmen (ZA) yucatana YU - 18.85 N, -89.13 W 2 2 2 79 10, 17 121, 122 JB, PP (CRI)

Appendix B: Haplotype designations for the GAPD haplotype network, as determined using the Phase algorithm in DnaSP v 5.10.01. The left column (GAPD TCS Network Haplotype #) refers to the haplotype numbers in the GAPD haplotype network (Figure 9). The right column (GAPD Phylogeny Haplotype #) refers to the corresponding haplotype numbers from the GAPD phylogenetic analysis (Figure 5).

GAPD TCS Network Haplotype # GAPD Phylogeny Haplotype # h1 h59-1, h59-2 h2 h4-1, h4-2, h48-1, h49-1 h3 h5-2, h20-1, h20-2, h45-1, h46-1 h4 h5-1 h5 h46-2 h6 h48-2 h7 h45-2 h8 h49-2 h9 h3-1, h3-2, h6-2, h7-2, h24-2, h50-1 h10 h1-1, h1-2, h9-1, h13-1, h15-1, h16-1, h38-1, h39-1, h42-1 h11 h39-2, h41-2, h60-1 h12 h50-2 h13 h2-1, h2-2, h13-2, h60-2 h14 h11-1, h11-2, h15-2 h15 h12-1, h12-2, h43-1, h61-1, h62-2 h16 h34-1, h34-2, h61-2, h63-2 h17 h6-1, h9-2, h8-1, h29-1, h33-1, h40-1, h41-1, h62-1, h63-1 h18 h35-1 h19 h35-2 h20 h24-1, h28-1, h38-2, h40-2 h21 h28-2 h22 h33-2 h23 h43-2 h24 h29-2, h42-2 h25 h7-1, h8-2, h16-2 h26 h22-1, h22-2 h27 h32-1 h28 h32-2 h29 h17-1, h17-2 h30 h18-1, h18-2, h54-1, h64-1, h66-1 h31 h25-1, h55-1, h56-1 h32 h51-2, h52-1, h55-2, h66-2

99 Appendix B: (Continued)

GAPD TCS Network Haplotype # GAPD Phylogeny Haplotype # h33 h53-2, h56-2, h64-2 h34 h10-1, h10-2, h21-1, h25-2, h27-1, h51-1, h53-1, h54-2 h35 h19-1, h19-2, h47-1 h36 h21-2 h37 h27-2, h52-2, h65-2 h38 h65-1 h39 h57-1, h57-2, h58-1, h58-2 h40 h47-2 h41 h67-1, h67-2

100

Appendix C: Cytochrome b DNA sequence alignment matrix as aligned in Clustal X. Each row represents a haplotype (h) followed by an associated number. Sampling information corresponding to each haplotype is contained in Appendix A.

101

Appendix C: (Continued)

102

Appendix C: (Continued)

103

Appendix C: (Continued)

104

Appendix C: (Continued)

105

Appendix C: (Continued)

10

6

Appendix C: (Continued)

107

Appendix C: (Continued)

108

Appendix C: (Continued)

109

Appendix C: (Continued)

110

Appendix C: (Continued)

111

Appendix C: (Continued)

112

Appendix C: (Continued)

113

Appendix C: (Continued)

114

Appendix C: (Continued)

115

Appendix C: (Continued)

116

Appendix C: (Continued)

117

Appendix C: (Continued)

118

Appendix C: (Continued)

119

Appendix C: (Continued)

120

Appendix C: (Continued)

121

Appendix C: (Continued)

122

Appendix C: (Continued)

123

Appendix C: (Continued)

124

Appendix C: (Continued)

125

Appendix C: (Continued)

126

Appendix C: (Continued)

127

Appendix C: (Continued)

128

Appendix C: (Continued)

129

Appendix C: (Continued)

130

Appendix C: (Continued)

131

Appendix D: Glyceraldehyde-3-phosphate dehydrogenase DNA sequence alignment matrix as aligned in Clustal X. Each row represents a haplotype (h) followed by an associated number. Sampling information corresponding to each haplotype is contained in Appendix A.

132 Appendix D: (Continued)

133

Appendix D: (Continued)

134

Appendix D: (Continued)

135

Appendix D: (Continued)

136

Appendix D: (Continued)

137

Appendix D: (Continued)

138

Appendix D: (Continued)

139

Appendix D: (Continued)

140