Hyde MO D 2021.Pdf

Development of Molecular and Morphological Resources for Identification and

Monitoring of Freshwater Mussel Species in the Genera Fusconaia and Pleurobema in

the Green River, Kentucky

Miluska Olivera-Hyde

Dissertation submitted to the faculty of the Virginia Polytechnic Institute and State University

in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

Fish and Wildlife Conservation

Eric M. Hallerman, Chair

Jess W. Jones, Co-Chair

Emmanuel Frimpong

Pawel Michalak

October 1st, 2020

Blacksburg, VA

Keywords: Freshwater Mussels, Fusconaia, Pleurobema, Morphometric and Phylogenetic

analyses

Development of Molecular and Morphological Resources for Identification and

Monitoring of Freshwater Mussel Species in the Genera Fusconaia and Pleurobema in

the Green River, Kentucky

Miluska Olivera-Hyde

TECHNICAL ABSTRACT

Freshwater mussel species in the genera Fusconaia and Pleurobema are particularly challenging to identify in the field. In this study, mussels from these genera were collected from the Green River, Kentucky for genetic and morphological analyses. I used molecular markers to detect any cryptic species within these genera and to test for genetic differentiation between two closely related nominal taxa P. rubrum and P. sintoxia using both mitochondrial

(ND1, COI, 16S rRNA) and nuclear (ITS1) DNA sequences. After species identification, I used microsatellite DNA markers to estimate genetic diversity and effective population sizes (Ne) of species of Pleurobema. I used microsatellite primers that were developed for P. clava and P. pyriforme in previous studies, as well as microsatellites that I developed for P. plenum. Finally,

I assessed morphological variation in my study species and developed dichotomous keys for the identification of both live mussels and shells. My results suggest that P. rubrum and P. sintoxia are the same species based on the mitochondrial DNA analyses, as there were few genetic differences between them. My results showed phylogenetically distinct lineages for F. flava, F. subrotunda, P. cordatum and P. plenum but no cryptic species were detected in the

Green River. Current and contemporary Ne showed that these species have large population sizes that should allow for avoiding inbreeding and maintaining their evolutionary potential.

Large genetic diversity as well as long-term effective population size could be the result of these species historically occurring as much larger assemblages that extended into the Ohio

River and its numerous tributaries. The last objective was to assess morphometrical differences

among these species. Using Canonical Variate Analysis, I found discernable morphological differences between the investigated species of Fusconaia and Pleurobema. The two

Fusconaia species were morphologically different from the Pleurobema species. However, the

Canonical Variate Analysis did not show differences among the Pleurobema species. I used decision tree analysis to develop a dichotomous tree, and random forest analysis was used to aid in the development of a dichotomous key by finding the most important diagnostic characters to distinguish these mussels. I then used the less subjective and easier to identify characters for the development of my dichotomous keys for live mussels and shells. However, both keys need to be tested in the field to determine their effectiveness. I could not separate P. rubrum and P. sintoxia mussels for morphometric analysis due to the lack of genetic differentiation and the inconsistent identification by the experts. However, I did describe a few individuals that look like P. rubrum and P. sintoxia to the eye of the experts. The description of these individuals matched previous descriptions of these mussels. Future studies need to assess taxonomic relationships among these species using genomics approaches, which might result in better node resolution. High genetic diversity and large effective population numbers for Pleurobema species suggest that these species’ populations are genetically healthy.

However, these results need to be interpreted carefully, and I therefore recommend additional studies to assess life history, habitat, host-fish availability, and current reproduction of these mussels in the Green River.

Development of Molecular and Morphological Resources for Identification and

Monitoring of Freshwater Mussel Species in the Genera Fusconaia and Pleurobema in

the Green River, Kentucky

Miluska Olivera-Hyde

POPULAR ABSTRACT

Freshwater mussels offer important ecosystem services for humans, to include water purification, nutrient storage and recycling, and mussels are part of the aquatic food web. In addition, freshwater mussels are indicators of ecosystem health. Because they rely on fish hosts to complete their complex life cycle, conservation of freshwater mussel species is particularly challenging. In this study, I focused my attention on a group of freshwater mussel species commonly known as “pigtoes”. These species are difficult to distinguish morphologically even by experts. Hence, first I used molecular genetic markers to rigorously identify these species.

Correct identification is needed to implement the most appropriate conservation plans for each species. Then I assessed the population viability by examining genetic diversity and estimating the size of the breeding population. . Finally, I developed two series of questions (called dichotomous keys) to help biologists identify either live mussel specimens or mussel shells.

My genetic results showed that there are five species of pigtoes in the Green River, KY. My results suggest that these mussel populations are large enough to survive and to adapt over time. Some of these species remain difficult to identify even by mussel experts who are unfamiliar with them. The dichotomous keys will support more accurate identification of these freshwater mussel species in the Green River, Kentucky, but need to be field-tested by mussel biologists.

ACKNOWLEDGEMENTS

I would like to dedicate my dissertation to my beloved husband Murray and my kids

Sebastian and Nathan. Both of my kids were born during my Ph.D. journey. Sebastian was with me during the writing of my dissertation proposal and Nathan was with me during my preliminary exams. Finally, I would like to dedicate my thesis to my family in Peru, to my parents Luis and Soledad, to my siblings David and Diana, and to my grandmother Luisa, because my professional journey has taken me so far away from them.

I would like to thank my committee members, Drs. Emmanuel Frimpong and Pawel

Michalak, for their input on my dissertation. I would also thank my advisors, Drs. Eric

Hallerman (Committee Chair) and Jess Jones (Committee Co-Chair), who gave me the opportunity to apply my previous experience on molecular biology to population genetics. I could not have asked for more insightful and understanding advisors. I would also like to thank the Department of Fish and Wildlife Conservation at Virginia Tech for helping to support my project.

The U.S. Fish and Wildlife Service, Frankfort, Kentucky Field Office sponsored this study. Funding was provided by Kentucky Aquatic Resources Fund (KARF), which is administered by the Kentucky Waterway Alliance (KWA). Field identification of mussels was performed by Leroy Koch (Senior Biologist, Kentucky Ecological Services Field Office – U.S.

Fish and Wildlife Service), Dr. Wendell Haag (Fisheries Research Biologist, U.S. Forest

Service stationed at the Center for Mollusk Conservation, Kentucky Department of Wildlife

Resources), Chad Lewis (Malacologist, Lewis Environmental Consulting, LCC), Dr. Monte

McGregor (State Malacologist, Center for Mollusk Conservation, Kentucky Department of

Wildlife Resources), and Adam Shephard (Icthyologist, Center for Mollusk Conservation,

Kentucky Department of Wildlife Resources). Field sampling, DNA collection, lab work, tagging, mussel collection, and photographs were performed with the help of the students and

v technicians at the Virginia Tech Freshwater Mollusk Conservation Center to include Aaron

Adkins, Anna Dellapenta, Alissa Ganser, Murray Hyde, Tim Lane, Katie Ortiz and Lee

Stephens and Caitlin Carey at the Conservation Management Institute, Virginia Tech.

TABLE OF CONTENTS

TECHNICAL ABSTRACT ...... ii

POPULAR ABSTRACT ...... iv

ACKNOWLEDGEMENTS ...... v

TABLE OF CONTENTS ...... vii

LIST OF TABLES ...... ix

LIST OF FIGURES ...... xi

LIST OF SUPPLEMENTAL TABLES ...... xiii

LIST OF SUPPLEMENTAL FIGURES ...... xiv

DICHOTOMOUS AND PHOTOGRAPHIC KEYS ...... xv

INTRODUCTION ...... 1 LITERATURE CITED ...... 3

Chapter 1 - Phylogenetic Assessment of Mussels (Bivalvia: Unionidae) in the Genera Fusconaia and Pleurobema in the Green River, Kentucky ...... 4 ABSTRACT ...... 4 INTRODUCTION ...... 6 METHODS ...... 9 Sample Collection ...... 9 Polymerase Chain Reaction ...... 10 Data Analysis ...... 12 RESULTS ...... 15 Amplification of Molecular Markers ...... 15 DNA Sequence Analysis ...... 16 Genetic Diversity ...... 18 Phylogenetic Analysis ...... 20 Haplotype Networks ...... 22 Split Networks...... 23 Pairwise Differentiation ...... 23 DISCUSSION ...... 25 Molecular markers ...... 26 Phylogenetic assessment ...... 28 Pyramid and Round pigtoes ...... 29 Genetic diversity ...... 31 Management Implications ...... 32 COLLABORATORS’ CONTRIBUTIONS...... 35 LITERATURE CITED ...... 36

Chapter 2 - Assessment of genetic diversity and effective population sizes (Ne) of Fusconaia and Pleurobema species in the Green River, Kentucky ...... 115 ABSTRACT ...... 115 INTRODUCTION ...... 116 METHODS ...... 118 Sample Collection ...... 118

vii

Molecular genetic markers ...... 119 Mitochondrial DNA Data Analysis ...... 121 Microsatellite DNA Data Analysis ...... 122 Long-Term and Contemporary Effective Population Size ...... 123 RESULTS ...... 124 Mitochondrial DNA Genetic Diversity ...... 124 Population Differentiation using mtDNA ...... 124 Long-term Female Effective Population Size Estimated using mtDNA ...... 125 Microsatellite Genetic Diversity ...... 125 Population Structure and Differentiation using Microsatellites ...... 127 Long-term Effective Population Sizes Using Microsatellites ...... 127 Contemporary Effective Population Number (Ne) ...... 128 DISCUSSION ...... 128 Genetic diversity ...... 128 Inbreeding Coefficient and Population Bottleneck...... 129 Effective Population Size and Viable Population Number ...... 130 Management Implications ...... 133 COLLABORATOR CONTRIBUTIONS ...... 134 LITERATURE CITED ...... 135

Chapter 3 - Quantification of morphological variation and development of morphology-based keys to identify species of Fusconaia and Pleurobema in the Green River, Kentucky ...... 176 ABSTRACT ...... 176 INTRODUCTION ...... 178 METHODS ...... 181 Sample Collection ...... 181 Identification Validation ...... 182 Geometric Morphometrics ...... 182 Decision Trees and Random Forest ...... 183 Dichotomous and Photographic Keys ...... 185 RESULTS ...... 185 Identification of Shells by Experts ...... 185 Geometric Morphometrics ...... 187 Decision Trees...... 188 Random-Forest Analyses ...... 189 Dichotomous Key ...... 190 Shell Photographic Key...... 191 Species Description ...... 192 DISCUSSION ...... 193 Expert identification ...... 193 Geometric Morphometrics ...... 194 Decision-Tree and Random-Forest Analyses ...... 195 Principal Issues in Mussel Identification ...... 196 Pyramid and Round Pigtoes ...... 196 Management Implications ...... 197 COLLABORATORS CONTRIBUTIONS ...... 199 LITERATURE CITED ...... 200

Chapter 4 Synthesis and Recommendations ...... 237 Molecular identification ...... 237 Estimation of genetic diversity and effective population size (Ne) ...... 238

viii

LIST OF TABLES Chapter 1. Phylogenetic Assessment of Mussels (Bivalvia: Unionidae) in the Genera Fusconaia and Pleurobema in the Green River, Kentucky

Table 1.1 Mitochondrial DNA (mtDNA) primers for NADH dehydrogenase 1 (ND1), cytochrome oxidase subunit I (COI), and 16S rRNA genes and nuclear DNA primers for the ribosomal internal transcribed spacer region subunit 1 (ITS1) used for amplification of DNA sequences and genetic analysis of freshwater mussel species belonging to the genera Fusconaia and Pleurobema...... 40 Table 1.2. Estimates of evolutionary divergence based on analysis of mitochondrial DNA COI sequence pairs between and within species of Fusconaia and Pleurobema ...... 43 Table 1.3. Estimates of evolutionary divergence based on analysis of mitochondrial DNA ND1 sequence pairs between and within species of Fusconaia and Pleurobema using p-distances (lower diagonal) and pairwise difference between species using the Tamura-Nei with gamma distribution (upper diagonal). ... 44 Table 1.4. Intraspecific variation of the mitochondrial DNA COI gene for species in the genera Fusconaia and Pleurobema...... 41 Table 1.5. Intraspecific variation of the mitochondrial DNA ND1 for species in the genera Fusconaia and Pleurobema...... 42

Chapter 2. Assessment of genetic diversity and effective population sizes (Ne) of Fusconaia and Pleurobema species in the Green River, Kentucky

Table 2.1. Distribution, host fish and conservation status of study species under the U.S. Endangered Species Act (ESA), the International Union for Conservation of Nature (IUNC). American Fisheries Society (AFS), and the Kentucky State Nature Preserves Commission (KSNP) for species belonging to the genus Fusconaia and Pleurobema...... 138 Table 2.2. Samples sizes for mtDNA (COI and ND1) and nuclear DNA microsatellites loci for species of Fusconaia and Pleurobema in the Green River, Kentucky...... 140 Table 2.3. The DNA sequences of primers, base-pair range of individual alleles, and repeat motif of the microsatellite PCR primers used in this study...... 141 Table 2.4. Intraspecific variation of the mitochondrial DNA cytochrome c oxidase subunit 1 (COI) and NADH dehydrogenase subunit 1 (ND1) about 1215 bp, for species of freshwater mussels in the genera Fusconaia and Pleurobema...... 143 Table 2.5. FST values estimated using mitochondrial DNA (mtDNA) and nuclear DNA (nDNA) data. Mussel specimens belonging to the genus Fusconaia and Pleurobema ...... 144 Table 2.6. Long-term effective population sizes estimated from combined mtDNA sequences of COI (471 bp) and ND1 (744 bp) and microsatellite data using approximate Bayesian computational approach implemented in DIYABC 2.1.0 ...... 145 Table 2.7. Genetic variation among microsatellites that were amplified for Pleurobema cordatum, Pleurobema plenum, and Pleurobema sintoxia/rubrum ...... 146 Table 2.8. Estimates of contemporary effective population sizes estimated from DNA microsatellite data that are bias-corrected using the linkage disequilibrium method...... 147

Chapter 3. Quantification of morphological variation and development of a morphology-based key to identify species of Fusconaia and Pleurobema in the Green River, Kentucky

Table 3.1. Sample sizes of mussel specimens used for conducting geometric morphometric analysis, decision tree and random forest analyses for the identification of species belonging to the genera Fusconaia and Pleurobema in the Green River, Kentucky...... 202 Table 3.2. Probability of five experts correctly identifying mussel species in the genera Fusconaia and Pleurobema ...... 203 Table 3.3. Confusion matrix showing the classification error for the five experts for mussel species in the genera Fusconaia and Pleurobema ...... 204 Table 3.4. Confusion matrix for decision trees for identification of live mussel specimens of Fusconaia and Pleurobema ...... 205 Table 3.5. Confusion matrix for decision trees for the identification of shells of species belonging to the genera Fusconaia and Pleurobema ...... 206 Table 3.6. Confusion matrix for random forest analysis of live mussel specimens and shells of species belonging to the genera Fusconaia and Pleurobema ...... 207

Table 3.7. Summary of the categorical and quantitative morphological variables used to describe and identify species belonging to the genera Fusconaia and Pleurobema ...... 208

LIST OF FIGURES Chapter 1. Phylogenetic Assessment of Mussels (Bivalvia: Unionidae) in the Genera Fusconaia and Pleurobema in the Green River, Kentucky

Figure 1.1. Sampling locations for freshwater mussel species in the genera Fusconaia and Pleurobema...... 45 Figure 1.2. Phylogenetic tree constructed using mitochondrial DNA COI sequences and Bayesian consensus trees...... 46 Figure 1.3. Phylogenetic tree constructed using mitochondrial DNA ND1 sequences and Bayesian consensus trees ...... 47 Figure 1.4. Phylogenetic tree constructed using mitochondrial COI + ND1 sequences and Bayesian consensus trees...... 48 Figure 1.5. Phylogenetic tree constructed using mitochondrial 16S rRNA sequences and Bayesian consensus trees...... 49 Figure 1.6. Phylogenetic trees constructed using ITS1 sequences with Clustal (A) and webPRANK (B) aligments...... 50 Figure 1.7. Haplotype network constructed using mitochondrial DNA COI sequences ...... 51 Figure 1.8. Haplotype network constructed using mitochondrial DNA ND1 sequences ...... 52 Figure 1.9. Haplotype network constructed using mitochondrial 16S rRNA sequences ...... 53 Figure 1.10. Haplotype network constructed using nuclear ITS1 sequences with Clustal alignment...... 54 Figure 1.11. Haplotype network constructed using nuclear ITS1 sequences with webPRANK alignment...... 55 Figure 1.12. Split phylogenetic network constructed using mitochondrial COI sequences...... 56 Figure 1.13. Split phylogenetic network constructed using mitochondrial ND1 sequences...... 57 Figure 1.14. Split phylogenetic network constructed using 16S rRNA sequences...... 58 Figure 1.15. Split phylogenetic network constructed using ITS sequences clustal alignment...... 59 Figure 1.16. Split phylogenetic network constructed using nuclear ITS sequences webPRANK alignment...... 60

Chapter 2. Assessment of genetic diversity and effective population sizes (Ne) of Fusconaia and Pleurobema species in the Green River, Kentucky

Figure 2.1. Sampling locations for freshwater mussel species in the genera Fusconaia and Pleurobema...... 148 Figure 2.2. Summary of intraspecific mitochondrial DNA COI+ND1 (1251 bp) variation for individuals of Fusconaia flava, Fusconaia subrotunda, Pleurobema cordatum, Pleurobema plenum, and Pleurobema sintoxia/rubrum...... 149 Figure 2.3. Effective female sizes estimated using mtDNA (COI + ND1) for Fusconaia flava, F. subrotunda, Pleurobema cordatum, P. plenum, and P. sintoxia/rubrum...... 150 Figure 2.4. Expected (He) and observed (Ho) heterozygosities for Pleurobema cordatum, Pleurobema plenum and Pleurobema sintoxia/rubrum...... 151 Figure 2.5. Number of alleles (A) at each locus for Pleurobema cordatum, Pleurobema plenum, and Pleurobema sintoxia/rubrum...... 152 Figure 2.6. M-ratios for freshwater mussels belonging to Pleurobema cordatum, P. plenum, and P. sintoxia/rubrum...... 153 Figure 2.7. Inbreeding coefficient (FIS) for freshwater mussels belonging to Pleurobema cordatum, P. plenum, and P sintoxia/rubrum...... 154 Figure 2.8. Long-term effective population sizes estimated using nuclear DNA microsatellites for Pleurobema cordatum, P. plenum, and P. sintoxia/rubrum...... 155

Chapter 3. Quantification of morphological variation and development of a morphology-based key to identify species of Fusconaia and Pleurobema in the Green River, Kentucky

Figure 3.1. Sampling locations for freshwater mussel species in the genera Fusconaia and Pleurobema ...... 210 Figure 3.2. Landmarks used for the geometric morphometric analysis...... 211 Figure 3.3. Morphology of a freshwater mussel. Length (mm) measurements of the shell that were used for decision-tree and random-forest analyses are illustrated...... 212 Figure 3.4. Categorical variables that were used to conduct the Decision Tree and Random Forest analyses for species in the genera Fusconaia and Pleurobema ...... 213 Figure 3.5. Bar graph showing the percentage of correctly identified mussel specimens and misidentifications for each species of Fusconaia and Pleurobema by each expert...... 214

Figure 3.6. Pie-chart showing the experts’ field identification of mussel specimens that were identified in the field to the Pleurobema sintoxia/rubrum clade ...... 215 Figure 3.7. Two representative shells for (A) Pleurobema rubrum and (B) Pleurobema sintoxia shell forms. . 216 Figure 3.8. Canonical Variate Analysis used to assess differences among species at the three size classes ...... 217 Figure 3.9. Canonical Variate Analysis used to assess differences among species at three size classes ...... 218 Figure 3.10. Shell landmark variation for species in the genera Fusconaia and Pleurobema at different size classes ...... 219 Figure 3.11. Normal distribution curve for ratios of quantitative morphological variables recorded for mussel specimens in the genera Fusconaia and Pleurobema...... 220 Figure 3.12. Decision tree showing external shell variables used to identify live mussel specimens of species in the genera Fusconaia and Pleurobema ...... 221 Figure 3.13. Decision tree showing external and internal shell variables used to identify shells only of Fusconaia and Pleurobema ...... 222 Figure 3.14. Most important morphological variables used for identifying mussels in the genera Fusconaia and Pleurobema ...... 223 Figure 3.15. Categorical variables and their proportion in each species belonging to the genera Fusconaia and Pleurobema...... 224 Figure 3.16. Ratios of quantitative morphological variables recorded for mussel specimens in the genera Fusconaia and Pleurobema ...... 225

xii

LIST OF SUPPLEMENTAL TABLES Chapter 1. Phylogenetic Assessment of Mussels (Bivalvia: Unionidae) in the Genera Fusconaia and Pleurobema in the Green River, Kentucky

APPENDIX Table 1.1. GenBank accession numbers for the mitochondrial DNA COI haplotypes for species in the genus Fusconaia and Pleurobema ...... 61 APPENDIX Table 1.2. GenBank accession numbers for the mitochondrial DNA ND1 haplotypes for species belonging to the genus Fusconaia and Pleurobema...... 64 APPENDIX Table 1.3. GenBank accession numbers for mitochondrial DNA 16sRNA haplotypes for Pleurobema sintoxia/rubrum and outgroups...... 68 APPENDIX Table 1.4. GenBank accession numbers for ITS haplotypes for Pleurobema sintoxia/rubrum and outgroups...... 69 APPENDIX Table 1.5. Haplotype combinations for mitochondrial DNA COI+ND1 sequences...... 70 APPENDIX Table 1.6. Variable nucleotide sites observed for Fusconaia flava in a 471 base-pairs (bp) section of the mitochondrial DNA gene COI ...... 75 APPENDIX Table 1.7. Variable nucleotide sites observed for Fusconaia subrotunda in a 471 base-pairs (bp) section of the mitochondrial DNA gene COI ...... 76 APPENDIX Table 1.8. Variable nucleotide sites observed for Pleurobema cordatum in a 471 base-pairs (bp) section of the mitochondrial DNA gene COI ...... 77 APPENDIX Table 1.9. Variable nucleotide sites observed for Pleurobema plenum in a 471 base-pairs (bp) section of the mitochondrial DNA gene COI ...... 81 APPENDIX Table 1.10. Variable nucleotide sites observed for Pleurobema sintoxia/rubrum in a 471 base-pairs (bp) section of the mitochondrial DNA gene COI ...... 82 APPENDIX Table 1.11. Variable nucleotide sites observed for Fusconaia flava in a 744 base-pairs (bp) section of the mitochondrial DNA gene ND1 ...... 83 APPENDIX Table 1.12. Variable nucleotide sites observed for Fusconaia subrotunda in a 744 base-pairs (bp) section of the mitochondrial DNA gene ND1 ...... 84 APPENDIX Table 1.13. Variable nucleotide sites observed for Pleurobema cordatum in a 744 base-pairs (bp) section of the mitochondrial DNA gene ND1 ...... 86 APPENDIX Table 1.14. Variable nucleotide sites observed for Pleurobema plenum in a 744 base-pairs (bp) section of the mitochondrial DNA gene ND1 ...... 92 APPENDIX Table 1.15. Variable nucleotide sites observed for Pleurobema sintoxia/rubrum in a 744 base-pairs (bp) section of the mitochondrial DNA gene ND1 ...... 93 APPENDIX Table 1.16. Variable nucleotide sites observed for Pleurobema sintoxia/rubrum in a 477 base-pairs (bp) section of mitochondrial DNA gene 16S rRNA ...... 95 APPENDIX Table 1.17. Variable nucleotide sites observed for Pleurobema sintoxia/rubrum in a 448 base-pairs (bp) section of nuclear DNA sequence ITS1...... 96 APPENDIX Table 1.18. Variable nucleotide sites observed for Pleurobema sintoxia/rubrum in a 448 base-pair (bp) section of nuclear DNA sequence ITS1...... 97 APPENDIX Table 1.19. The mitochondrial DNA gene COI sequences from Inoue et al. 2018 that were added to the phylogenetic tree and split network...... 98 APPENDIX Table 1.20. Sequences from other studies utilizing the mitochondrail DNA gene ND1 that were added to the phylogenetic tree and split network...... 103 APPENDIX Table 1.21. Evanno method using Delta K and Mean LnP(K) to support the number ofcluster in P. sintoxia/rubrum...... 105

Chapter 2. Assessment of genetic diversity and effective population sizes (Ne) of Fusconaia and Pleurobema species in the Green River, Kentucky

APPENDIX Table 2.1. Results of tests for the segregation of null alleles for species of Pleurobema cordatum, Pleurobema plenum, and P. sintoxia/rubrum...... 156 APPENDIX Table 2.2. Linkage disequilibrium tests for loci that did not show null alleles...... 159 APPENDIX Table 2.3. Microsatellite allele frequencies for Pleurobema cordatum, P. plenum, and P. sintoxia/rubrum...... 162

xiii

LIST OF SUPPLEMENTAL FIGURES Chapter 1. Phylogenetic Assessment of Mussels (Bivalvia: Unionidae) in the Genera Fusconaia and Pleurobema in the Green River, Kentucky

APPENDIX Figure 1.1. Phylogenetic tree constructed using mitochondrial DNA COI sequences and Bayesian consensus trees in MrBayes...... 108 APPENDIX Figure 1.2. Phylogenetic tree constructed using mitochondrial DNA ND1 sequences and Bayesian consensus trees in MrBayes...... 111 APPENDIX Figure 1.3. Phylogenetic tree constructed using mitochondrial DNA COI + ND1 sequences and Bayesian consensus trees in MrBayes...... 114

xiv

DICHOTOMOUS AND PHOTOGRAPHIC KEYS KEY A. Dichotomous key to identify live mussel specimens of the Fusconaia and Pleurobema species occurring in the Green River, Kentucky...... 226 KEY B. Dichotomous key to identify dead mussel specimens (shell without soft parts) of the Fusconaia and Pleurobema species occurring in the Green River, Kentucky ...... 229 KEY C. Photographic key of each investigated species occurring in the Green River, Kentucky ...... 231

INTRODUCTION

The Green River of central Kentucky (KY) is a tributary of the Ohio River and flows through the Interior Low Plateau physiographic province (Benke and Cushing 2011). It is one of most biodiverse tributaries of the Ohio River system (Master et al. 1998). The upper section of the Green River basin is used for agriculture and the river has several impoundments, while the middle section flows without impoundments for 100 miles through Mammoth Cave

National Park and eight counties (Benke and Cushing 2011; Master et al. 1999). The lower section of the Green river has several dams and locks for coal and bulk materials transport. The principal sources of pollution affecting water quality and freshwater mussel habitat involve agriculture, which occurs throughout the watershed. Biodiversity and water quality in the

Mammoth Cave underground section of the Green River are affected by siltation (due to deforestation), sedimentation and eutrophication from agriculture, streamflow regulation, and treated sewage discharge (Master et al. 1998). Preventing the spread of introduced species, such the zebra mussel, Dreissena polymorpha, is of high concern in the Green River (Master et al. 1998; Grabarkiewicz and Davis 2008).

The Green River, KY is characterized by high mussel biodiversity, with approximately

71 mussel species known (Grabarkiewicz and Davis 2008; Master et al. 1998; Benke and

Cushing 2011). Among them are two species in the genus Fusconaia and five in the genus

Pleurobema. Species of conservation concern are Rough Pigtoe (Pleurobema plenum, federally endangered), Clubshell (P. clava, federally endangered), Ohio Pigtoe (P. cordatum, near threatened), Pink Pigtoe (P. rubrum, near threatened), Round Pigtoe (P. sintoxia, least concern), Long-Solid (Fusconaia subrotunda, vulnerable), and Wabash Pigtoe (F. flava, least concern) (Master et al. 1998; Turgeon et al. 1998; Grabarkiewicz and Davis 2008; Haag and

Cicerello 2016; IUCN 2017). The external shell morphology of these species is quite similar and difficult to distinguish even by experts (Harper et al. 2000; Campbell et al. 2005; Campbell

1 and Lydeard, 2012; Jones et al. 2015). Past studies have described these species’ morphological characters, including reproductive traits, shell morphology, and soft anatomy (Ortmann 1921;

Heard and Guckert 1970). More recent studies have used genetic markers to clarify species boundaries within this group, as well as to help define management units (Harper et al. 2000;

Giribet and Wheeler 2002; Campbell et al. 2005; Campbell and Lydeard, 2012; Jones et al.

2015).

In these look-alike mussels, taxonomic characteristics useful for species identification overlap morphologically because of environmental factors, mussel age, and size. Hence, there is a need for development of a probabilistic, dichotomous key for identification of these mussel species, as well as an assessment of both current and long-term effective population sizes. In order to meet these needs, I needed to first correctly identify these mussels and define biologically based conservation units with the help of freshwater mussel experts and results of genetic analysis. After verification of mussel specimens genetically, different morphological characters then can be used for development of dichotomous and photographic keys. Finally, estimation of effective population sizes using mitochondrial and microsatellite DNA markers will help to assess the current conservation status of these species and to identify management units and develop management plans for the conservation of these look-alike species.

The phylogenetic analysis, estimation of genetic diversity, estimated population size, and morphological analysis of species belonging to the genera Fusconaia and Pleurobema in the Green River, KY is a first step leading to the identification of management units and estimation of minimum viable population (MVP) sizes. In addition, this study leads to a better morphological description of sympatric species occurring in this river system, and to the design and implementation of defensible management plans appropriate for critically endangered species, especially for P. plenum and P. clava.

LITERATURE CITED

Campbell, D. C., and C. Lydeard. 2012. Molecular systematics of Fusconaia (Bivalvia: Unionidae: Ambleminae). American Malacological Bulletin 30: 1-17.

Campbell, D. C., J. M. Serb, J. E. Buhay, K. J. Roe, R. L. Minton, and C. Lydeard. 2005. Phylogeny of North American amblemines (Bivalvia, Unionoida): prodigious polyphyly proves pervasive across genera. Invertebrate Biology 124:131-164.

Benke, A. C., and C. E. Cushing. 2011. Rivers of North America. Academic Press, New York.

Giribet, G., and W. Wheeler. 2002. On bivalve phylogeny: a high-level analysis of the Bivalvia (Mollusca) based on combined morphology and DNA sequence data. Invertebrate Biology, 121: 271-324.

Grabarkiewicz, J. D., and W. S. Davis. 2008. An introduction to freshwater fishes as biological indicators. U.S. Environmental Protection Agency, Office of Environmental Information. EPA-260-R-08-015, Washington, DC.

Haag, W. R., and R. R. Cicerello. 2016. A distributional atlas of the freshwater mussels of Kentucky. Scientific and Technical Series 8. Kentucky State Nature Preserves Commission, Frankfort, KY.

Harper, E. M., J. D. Taylor, and J. A. Crame. 2000. Unravelling the evolutionary biology of the Bivalvia: a multidisciplinary approach. Geological Society, London, Special Publications 177: 1-9.

Jones, J. W., N, Johnson, P. Grobler, D. Schilling, R. J. Neves, and E. M. Hallerman. 2015. Endangered rough pigtoe pearlymussel: assessment of phylogenetic status and genetic differentiation of two disjunct populations. Journal of Fish and Wildlife Management 6: 338- 349.

Master, L. L., S. R. Flack, and B. A. Stein. 1998. Rivers of life: critical watersheds for protecting freshwater biodiversity. Nature Conservancy, Arlington, Virginia: 71 pp. Ortmann, A. E. 1921. South American naiades: a contribution to the knowledge of the freshwater mussels of South America. Memoranda of the Carnegie Museum 8:451–668

Turgeon, D. D., J. F. Quinn, A. E. Bogan, E. V. Coan, F. G. Hochberg, W. G. Lyons, P. M. Mikkelsen, R. J. Neves, C. F. Roper, G. Rosenberg, and B. Roth. 1998. Common and Scientific Names of Aquatic Invertebrates from the United States and Canada: Mollusks. American Fisheries Society, Bethesda, MD.

Chapter 1 - Phylogenetic Assessment of Mussels (Bivalvia: Unionidae) in the Genera

Fusconaia and Pleurobema in the Green River, Kentucky

ABSTRACT

The purpose of this study was to phylogenetically characterize species in the genera

Fusconaia and Pleurobema of the Green River, KY using variation at the mitochondrial DNA

NADH dehydrogenase 1 (ND1), cytochrome oxidase subunit I (COI), and 16S rRNA genes and the nuclear DNA ribosomal internal transcribed spacer region subunit 1 (ITS1). Analysis of variation at these genetic markers was used to reconstruct phylogenetic history and to detect any cryptic biodiversity among these Green River mussels. Results from the mitochondrial

COI, ND1, COI+ND1 analyses showed five well-diverged groups that included F. flava, F. subrotunda, P. cordatum, and P. plenum as distinct clades, and with P. sintoxia and P. rubrum grouping together into a single clade. Thus, while these analyses did not distinguish P. sintoxia and P. rubrum as phylogenetically distinct species, the ypical shell forms of these two nominal taxa are very distinct. Further phylogenetic analysis using mitochondrial 16S rRNA and nuclear

ITS1 sequences also showed that P. sintoxia and P. rubrum were not distinct lineages. The highest haplotype diversity (h), average number of nucleotide differences (k), and nucleotide diversity (π) were reported for F. subrotunda at both the COI (h = 0.896, k= 3.805, π = 0.00808) and ND1 (h = 0.984, k= 6.595, π = 0.00886) markers. Molecular marker analyses did not result in discovery of cryptic species that are morphologically similar to the species of Fusconaia and

Pleurobema in the Green River, Kentucky. Future taxonomic study should include next- generation sequencing to increase the number of markers available. In addition, variation in shell morphology of all investigated species should be evaluated in different sections of the

Green River.

KEYWORDS: Fusconaia, Pleurobema, cytochrome oxidase subunit I (COI), NADH dehydrogenase 1 (ND1), mitochondrial 16S rRNA, ribosomal internal transcribed spacer region subunit 1 (ITS1)

INTRODUCTION

Freshwater mussel species in the genera Fusconaia and Pleurobema belong to the Tribe

Pleurobemini within the Family Unionidae (Subfamily Ambleminae) and are broadly distributed in river systems in central and eastern North America (Heard and Guckert 1970;

Campbell and Lydeard 2012b). Many species in these two genera look very similar in their shell morphology. The IUCN Red List status for the species of interest in this study range from critically endangered for the rough pigtoe (P. plenum) and the Clubshell (P. clava), vulnerable for long-solid (F. subrotunda), near-threatened for Ohio pigtoe (P. cordatum) and pink pigtoe

(P. rubrum), and least-concern for Wabash pigtoe (F. flava) and round pigote (P. sintoxia). The

Green River, Kentucky (KY) has one of the most diverse assemblages of mussels including species of Fusconaia and Pleurobema, In the Green River, all seven species are of management concern (Master et al. 1998). Complicating their management, the shell phenotypes of these species are particularly difficult to distinguish morphologically even by experts. Previous studies describing these species focused on a suite of shell and soft-body morphological characters. For example, some of the principal differences among species of Fusconaia and

Pleurobema are the number of gills charged when gravid (four for Fusconaia and two for

Pleurobema), conglutinate morphology (leaflike for Pleurobema and subcylindrical for

Fusconaia), and foot color (white for Pleurobema and generally orange for Fusconaia)

(Barnhart et al. 2008; Schilling 2015). However, shell and soft-body characters used for inter- and intra-specific morphological identification can overlap, depending on environmental and genetic factors, mussel age and size, and phenotypic plasticity (Inoue et al. 2014). In addition, similar phenotypes could be the result of all these Pleurobemini species being closely related, which likely has resulted in these species sharing traits.

Recent advances in development of molecular markers and the widening application of phylogenetic analyses have led to more reliable identification of mussels and to development

6 of more scientifically defensible management plans. Phylogenetic analysis of freshwater mussels in North America has focused increasingly on application of DNA sequence variation, including variation at the mitochondrial DNA (mtDNA) 16S rRNA, cytochrome oxidase subunit I (COI), cytochrome b (Cyt-b), and NADH dehydrogenase 1 (ND1) genes, as well as at nuclear genes, including large ribosomal subunit 28S rDNA and the ribosomal internal transcribed spacer region subunit 1 (ITS1) (Campbell et al. 2008; Campbell and Lydeard 2012a;

Graf and Cummings 2007; Jones et al. 2015). Several recent studies have characterized phylogenetic relationships among species in the genera Fusconaia and Pleurobema using these markers. For example, F. flava, F. cerina, and F. askewi were shown to be the same species based on COI and ND1 haplotypes (Campbell and Lydeard 2012a). Further, species in the genus Fusconaia tend to show low intrapopulation (F. subrotunda) and low interpopulation (F. flava/cerina) variation, i.e., mtDNA divergence is low both within and among species in this genus (Campbell and Lydeard 2012a). Relevant to my study of mussels in the Green River, phylogenetic relationships among F. flava, F. subrotunda, P. cordatum, P. clava, P. plenum,

P. sintoxia and P. rubrum were recently assessed using CO1, ND1, and ITS1 markers

(Campbell et al. 2008; Campbell and Lydeard 2012b; Jones et al. 2015; Inoue 2018;). Early studies suggested the existence of a P. cordatum group which included P. cordatum, P. plenum,

P. rubrum and P. sintoxia (Campbell et al. 2008; Campbell and Lydeard 2012b), and more extensive sampling by Jones et al. (2015) and Inoue et al. (2018) validated that P. cordatum and P. plenum are indeed different species, and utilizing ND1 and COI markers they showed that only a few nucleotide differences exist between individuals of P. sintoxia and P. rubrum, and that further phylogenetic assessment was needed to delineate these two nominal taxa.

While the shells of P. sintoxia generally are morphologically distinctive, they can occasionally be mistaken for P. rubrum and vice-versa. However, the umbos of P. rubrum are pointed and pronounced, resembling a pyramid, hence the common name “Pyramid pigtoe”, and this

7 species typically has a well-defined sulcus traversing the middle of each valve, especially in larger and older mussel specimens (Miller et al. 2008). Another relevant example of closely related species is P. clava and P. oviforme, a species endemic to the Tennessee and Cumberland

River watersheds, which showed that few molecular differences at mtDNA genes exist between these two taxa, but when assessed at nuclear ITS1, differences were observed (Campbell et al.

2008). Therefore, utilization of nuclear as well as mitochondrial DNA sequences is critical for assessing phylogenetic differentiation among closely related species.

Phylogenetic relationships among species in the genera Fusconaia and Pleurobema have been assessed across various geographic regions in North America using a suite of molecular markers. While these comparisons have been made with a large number of mussel specimens from various species belonging to the Tribe Pleurobemini (Graf and Cummings

2007; Campbell et al. 2008; Campbell and Lydeard 2012; Jones et al. 2015), molecular data are sparse for these species in the Green River, Kentucky. Rigorous phylogenetic assessment of morphologically similar species in these two genera was vital for development of a probabilistic dichotomous key for the Green River and the regional Ohio River mussel faunas.

The phylogenetic identification of mussel specimens from the Green River aimed to support the identification of cryptic species, management units and design and implementation of appropriate management plans for critically endangered species, especially for P. plenum.

In this study, DNA sequences of three mtDNA genes of importance, 16S rRNA, COI and ND1, were screened. The COI gene has been widely used in phylogenetic studies and DNA barcoding and evolves slowly relative to other mitochondrial protein-encoding regions

(Patwardhan et al. 2014). Similarly, the ND1 gene evolves slowly compared to nuclear DNA microsatellites and has been used to differentiate among different taxa that once were recognized only by morphological data (Abernethy et al. 2013). A combination of COI (~471 base-pair) and ND1 (~744 base-pair) genes (Graf and Cumming 2007) were used for assessing

8 the phylogeny of species of concern in this study. In addition, differences among closely related species, such as P. rubrum and P. sintoxia, were further evaluated using the mitochondrial 16S rRNA and nuclear ITS1 (Campbell et al. 2008). The execution of phylogenetic analysis, as well as the screening for any cryptic species belonging to the genera Fusconaia and Pleurobema in the Green River, KY supported the identification of biologically based management units, and assisted with morphological identification of these sympatric species occurring in this river system.

METHODS

Sample Collection

A total of 258 mussel specimens belonging to species in the genera Fusconaia and

Pleurobema were collected from two sites in the Green River, KY, Pool 4 during September

2015 (GPS coordinates = 37.18286, -86.6296; river mile = 149). A second sampling effort was conducted to increase the number of mussel specimens in the respective size-classes especially of smaller mussel specimens. This second sampling occurred in the Western Kentucky

University, Bio Preserve just upstream of Mammoth Cave National Park (GPS coordinates

37.17819, -86.1154; river mile = 197) during November 2017 (Figure 1.1). Additional 17 individuals were collected from the Clinch River. The locations in the Clinch River were Honey

Hole (GPS coordinates 36.523311, -83.204240), Frost Ford (GPS coordinates 36.534881, -

83.179205), and Kyle’s Ford (GPS coordinates 36.565230, -83.054863). In addition, five individuals were collected at other locations in the Tennessee River downstream of Pickwick

Dam, Hardin County, TN. These individuals were collected for use as outgroups to test for any cryptic species among collections from distinct watersheds. All mussel specimens collected from the Green River were physically tagged and kept at the Minor E. Clark Fish Hatchery, near Morehead, Kentucky, until data for the shell morphological analysis were collected. For

9 each species, tissue for DNA isolation was collected non-lethally by swabbing the mussel foot with a DDK-50 swab (Isohelix, Harriettsham, UK).

Polymerase Chain Reaction

I extracted DNA using an Isohelix DNA Isolation Kit. Concentration and purity of the double-stranded DNA were measured using a µLite PC spectrophotometer (Biodrop,

Cambridge, UK), and DNA was diluted to 10-30 ng/µl. All PCRs were performed in either a

T100TM or MyCycler TM thermocycler (both from Bio-Rad, Hercules, CA). PCR products were sent to the Fralin Life Sciences Institute (Blacksburg, VA) for Sanger sequencing. For ND1, I used two different pairs of primers to obtain amplified sequences for all species (Table 1.1).

Amplification products were obtained for most mussel specimens of F. flava, F. subrotunda,

P. cordatum, and P. plenum using primers LeuuurF and LoGlyR (Serb et al. 2003). For some mussel specimens (later identified as P. sintoxia and P. rubrum) it was necessary to use primers nadh1-F and nadh1-R (Buhay et al. 2002; Serb and Lydeard 2003). For both pairs of primers,

PCR reactions were conducted in a volume of 22 µl which contained 0.45X GoTaq Flexi

Buffer, 2.04 mM of MgCl2, 0.05 mM of each dNTP, 0.02 mg/ml of bovine serum albumin

(BSA), 0.09 µM of each primer, 0.02 units/µl of GoTaq® DNA polymerase (Promega,

Madison, WI), and 10-30 ng/µl of dsDNA template. The PCR protocol included an initial denaturation of 95oC for 5 min; followed by 30 cycles of 96oC for 20 sec, 53oC for 20 sec, and

72oC for 45 sec; a final extension at 72oC for 5 min; and a hold at 4oC.

Two slightly different forward primers were used to amplify COI sequences for all species (Table 1.1). Sequences for COI were obtained for mussel specimens of F. flava, F. subrotunda, P. cordatum, and P. rubrum and P. sintoxia using the primers LCO1490 (Folmer et al. 1994) and HCO700dy2 (Walker et al. 2006). The PCR amplification was conducted in a volume of 22 µl which contained 0.45X GoTaq Flexi Buffer, 2.05 mM of MgCl2, 0.045 mM

10 of each dNTPs, 0.02mg/ml of BSA, 0.09µM of each primer, 0.02 units/µl of GoTaq® DNA polymerase (Promega), and 10-30 ng/µl of dsDNA template. The PCR protocol included an initial denaturation of 94oC for 3 min; followed by 35 cycles of 94oC for 1 min, 57oC for 45 sec, and 72oC for 1 min; a final extension at 72oC for 5 min; and a hold at 4oC. In the case of

P. plenum, most sequences were obtained by using primers COIF (Campbell et al. 2005) and

HCO700dy2. The PCR reaction was conducted in a volume of 22 µl which contained 0.45X

GoTaq Flexi Buffer, 2.27 mM of MgCl2, 0.072 mM of dNTP mix, 0.02 mg/ml of BSA, 0.018

µM of each primer, 0.02 units/µl of GoTaq® DNA polymerase (Promega), and 10-30 ng/µl of dsDNA template. The PCR protocol included an initial denaturation of 94oC for 3 min; followed by 35 cycles of 94oC for 1 min, 55oC for 45 sec, and 72oC for 1 min; a final extension at 72oC for 5 min; and a hold at 4oC.

For 42 mussel specimens identified as P. sintoxia or P. rubrum with either COI and/or

ND1, sequences from the 16S rRNA region were amplified using primers reported in Table 1.1.

We also sequenced 16S rRNA for one mussel specimen of each of the other species (F. flava,

F. subrotunda, P. cordatum, and P. plenum). Finally, the 16S rRNA sequences of four P. sintoxia or P. rubrum mussel specimens from the Clinch River and three mussel specimens from the Tennessee River were added to provide outgroups for the phylogenetic analysis. PCR reactions were conducted in a final volume of 22 µl which contained 0.45X GoTaq Flexi

Buffer, 2.27 mM of MgCl2, 0.05 mM of each dNTP, 0.02 mg/ml of BSA, 0.18 µM of each primer, 0.02 units/µl of GoTaq® DNA polymerase (Promega), and 10-30 ng/µl of dsDNA template. The PCR protocol included an initial denaturation of 95oC for 5min; followed by 30 cycles of 96oC for 20 sec, 53oC for 20 sec, and 72oC for 45 sec; a final extension at 72oC for 5 min; and a hold at 4oC.

I amplified the nuclear ribosomal internal transcribed spacer region subunit 1 (ITS1) sequence for mussel specimens molecularly identified as P. rubrum and P. sintoxia (with

11 mitochondrial COI and/or ND1 markers) from the Green River (40 mussel specimens).

Sequences used as outgroups included P. sintoxia or P. rubrum from the Clinch River (four sequences) and Tennessee River (three sequences). Finally, two mussel specimens of F. flava, three F. subrotunda, two P. cordatum, and three P. plenum were sequenced and used as outgroups. PCR reactions were conducted in a final volume of 22 µl which contained 0.45X

GoTaq Flexi Buffer, 2.05 mM of MgCl2, 0.05 mM of each dNTPs, 0.02 mg/ml of BSA, 0.09

µM of each primer, 0.02 units/µl of GoTaq® DNA polymerase (Promega), and 10-30 ng/µl of dsDNA. The PCR protocol included an initial denaturation of 94oC for 5 min; followed by 35 cycles of 94oC for 40 sec, 64oC for 40 sec, and 72oC for 1 min; a final extension at 72oC for 5 min; and a hold at 4oC.

Data Analysis

The consensus DNA sequences were obtained using Geneious® 7.0.6 (Biomatters, Inc.,

San Francisco, CA), and aligned using GeneStudio Version 2.2.0.0 (GeneStudio, Inc., Athens,

Georgia). DNA sequence variation metrics – such as polymorphic nucleotide sites, number of haplotypes, nucleotide diversity, and haplotype diversity – were calculated using DnaSP 5.10

(Rozas et al. 2009). Pairwise difference values within and between species were estimated using p-distances, and the most likely model of nucleotide substitution was identified using

MEGA6 (Tamura et al. 2013).

For construction of phylogenetic trees and networks, the most appropriate model of nucleotide substitution was selected using MrModeltest 2 (Nylander 2008), which works in an interface with PAUP 4.0 (Swofford 1998), by selecting the model with the lowest Akaike

Information Criterion (AIC). For COI, ND1 and COI + ND1 DNA sequences, the most appropriate model was the General Time Reversible (GTR+G+I) model with sites following a gamma distribution with a proportion of invariable sites. For 16S rRNA, the most appropriate

12 model was Hasegawa-Kishino-Yano (HKY+I) with invariable sites. Due to the presence of gaps caused by nucleotide insertions and deletions (indels) in the sequence alignments for the

ITS1 region, I performed DNA sequence alignment using both the ClustalW (Thompson et al.

2003) plug-in implemented in Geneious® 7.0.6 (Biomatters, Inc., San Francisco, CA) and also webPRANK (Löytynoja and Goldman 2010). After alignment, FastGap v1.2 (Borchsenius

2009) was used to encode observed indels. The most appropriate model was the symmetrical model (SYM+G) with gamma rates.

Phylogenetic trees were constructed using MrBayes 3.2.6 (Ronquist et al. 2012), and final trees were visualized using FigTree v1.4.2 (Rambaut 2014). To analyze the MCMC runs resulting from MrBayes, I used Tracer v 1.6.0 (Rambaut et al. 2009). In this software, the effective sample size (ESS) was >200 for all the trees. Phylogenetic trees were constructed for each mitochondrial marker, 16S rRNA, COI and ND1, using sequences from all the mussel specimens collected, and an additional tree was constructed using all markers combined. In addition, a phylogenetic tree for P. sintoxia and P. rubrum was constructed for the nuclear marker ITS1. These analyses incorporated four MCMC chains with trees sampled every 1000 generations ND1, and COI+ND1 and every 250 generations for COI. Finally, the 16S rRNA and ITS1 trees were sampled every 100 generations.

Species differentiation was assessed using the Automatic Barcode Gap Discovery

(ABGD) method (Puillandre et al. 2012) using ND1, COI, and COI + ND1 sequences.

Haplotype networks were constructed using the TCS (Clement et al. 2002) network as implemented in PopART 1.7 (Leigh and Bryant 2015) for COI, ND1, 16S rRNA, and ITS1 sequences (using both Clustal and WebPrank alignment). A phylogenetic network was constructed using SplitsTree4 (Huson and Bryant 2006) for COI, ND1, 16S rRNA, and ITS1 sequences. For COI and ND1 sequences, I included additional sequences from other species

(APPENDIX Tables 1.19 and 1.20). To construct the network, I used the most appropriate

13 evolutionary model to calculate the distances and the NeighborNet algorithm (Bryant and

Moulton 2002) for distance transformation. For COI, ND1, and combined COI+ND1 sequences, species delimitation was assessed by using ABGD. To assign mussel specimens to species, the Kimura (1980) two-parameter (K2P) distance model was used, where the minimum intraspecific genetic distance (Pmin) was set to 0.001 and the maximum intraspecific genetic distance (Pmax) was set to 0.1.

For construction of the COI tree and split network, I added DNA sequences reported by Inoue et al. (2018), The species added included Fusconaia askewi, F. cerina, F. chunii, F. flava, F. lanensis, F. masoni, F. subrotunda, Pleurobema cordatum, P. plenum, P. riddellii, P. rubrum, and P. sintoxia. Finally, a sequence of Pleuronaia dolabelloides was added as an outgroup for the phylogenetic tree (APPENDIX Table 1.19). These sequences were trimmed to match the length of the sequences for the phylogenetic analysis, resulting in many of them then being grouped into new haplotypes. Only one of these sequences for each new haplotype was added into the analysis used to reconstruct the phylogenetic tree. The added sequences as well as similar sequences, GenBank accession numbers and sampling locations for these added sequences are shown in APPENDIX Table 1.19. In addition, for construction of the ND1 tree and split network, I added sequences for Fusconaia askewi, F. lanensis, F. masoni, F. subrotunda, Pleurobema cordatum, P. plenum, P. rubrum, and P. sintoxia. In addition, I added a sequence of Pleurobema dolabelloides as an outgroup. These sequences, Genbank accession numbers, sampling locations and references are shown in APPENDIX Table 20.

I constructed a 16S rRNA tree to assess differences among haplotypes of P. rubrum and

P. sintoxia. The outgroups for this tree were collected from the Green River (Pool 4) and included F. flava (16S_Ffla), F. subrotunda (16S_Fsub), P. cordatum, (16S_Pcor) and P. plenum (16S_Pple). I constructed separate ITS trees using the Clustal and webPRANK sequence alignments. Outgroup taxa for the ITS1 analyses were collected in the Green River,

14 and included F. flava (ITS1_Ffla), F. subrotunda (ITS1_Fsub), P. cordatum, (ITS1_Pcor) and

P. plenum (ITS1_Pple).

Finally, mitochondrial DNA is not useful to assess species differentiation if there has been hybridization between these two putative species P. rubrum and P. sintoxia. Hence, in addition to ITS1 maker, I used nine DNA microsatellite markers in Chapter 2 to assess if the genetic differentiation and ancestry of these two species was indicative of species-level separation. I conducted this analysis using Structure 2.3.4 (Pritchard et al. 2000) and Structure

Harvester (Earle 2012), where I ran 1’000,000 MCMC generations with a burn in period of

100,000 generations and with 5 iterations for k=3. Fixation index (FST) using microsatellites was also assessed in Arlequin v3.5 (Excoffier and Lischer 2010) in Chapter 2.

RESULTS

Amplification of Molecular Markers

In the case of the COI primers, I used one reverse primer (HCO700dy2) and two forward primers. The first forward primer was LCO1490 (Folmer et al. 1994) and has been used in studies with different species of arthropods. Using this forward primer, I was able to amplify sequences for individuals belonging to F. flava, F. subrotunda, P. cordatum, and P. sintoxia/rubrum. In order to amplify sequences for individuals which were later identified as

P. plenum, I used a second forward primer COIF (Campbell et al. 2005). This primer has been used successfully for species in the tribe Pleurobemini (Burlakova et al. 2012, Campbell and

Serb 2012a, Campbell and Serb 2012b). For ND1, I used two pairs of primers to amplify a region of 744 bp from ND1. The first set included LeuuurF and LoGlyR (Serb et al. 2003) and was suitable to amplify sequences for individuals of F. flava, F. subrotunda, P. cordatum, and

P. plenum. These primers have been used successfully for a wide range of mussels and is cited in different studies (Schilling 2015, Serb et al. 2003, Smith et al. 2018). However, I needed

15 additional primers to amplify sequences of P. sintoxia and P. rubrum. These primers were nadh1-F and nadh1-F (Buhay et al. 2002, Serb and Lydeard 2003 (Table 1.1), which have been used in several other studies which included numerous species belonging to the tribe

Pleurobemini (Burlakova et al. 2012, Campbell and Serb 2012a, Campbell and Serb 2012b and

Campbell et al. 2008).

In the case of 16S rRNA, I did not have any issues amplifying the sequences using the primers 16SL1987 and 16Sbr-H (Table 1.1).These primers were useful to obtain sequences of

P. sintoxia/rubrum and for F. flava, F. subrotunda, P. cordatum, and P. plenum. Finally, ITS1 sequences were amplified using the primers 18S and 5.8S (King et al. 1999). These primers worked for most mussel specimens of P. sintoxia/rubrum and for additional outgroups sequences of F. flava, F. subrotunda, P. cordatum, and P. plenum. The principal problem when amplifying these sequences were the gaps among sequences which many times were due to an artifact in the sequencing purity and quality. To ensure the quality of the sequences, I re- sequenced the ones that presented extra bases and the ones that simply were not of high quality

(< 80% GC). Like Schilling (2015), I did not encounter length differences among sequences from the same mussel specimens. Hence, length differences were not quantified in my study or in Schilling (2015). Variation of ITS1 sequences in this study included only one ITS1 sequence for all individuals that were analyzed.

DNA Sequence Analysis

The haplotype names as well as the GenBank accession names, and the collection sites for all my sequences for COI, ND1, 16S rRNA, and ITS1 are listed in APPENDIX Table 1.1–

1.4. Combinations for COI and ND1 sequences are listed in APPENDIX Table 1.5.

The DNA sequences for COI, typically about 471 bp per mussel specimen, were obtained for all 258 mussels collected from the Green River (APPENDIX Table 1.1). These

16 sequences resulted in observation of 6 variable sites among haplotypes of F. flava (APPENDIX

Table 1.6), 14 for F. subrotunda (APPENDIX Table 1.7), 46 for P. cordatum (APPENDIX

Table 1.8), 11 for P. plenum (APPENDIX Table 1.9), and 21 for P. sintoxia and P. rubrum, collectively (APPENDIX Table 1.10). Molecular analysis using COI resulted in identification of 43 mussel specimens of F. flava representing 7 haplotypes, 22 mussel specimens of F. subrotunda representing 13 haplotypes, 117 mussel specimens of P. cordatum representing 43 haplotypes, 33 mussel specimens of P. plenum representing 8 haplotypes, and 43 mussel specimens P. sintoxia and P. rubrum representing 16 haplotypes. Mussel specimens collected from the Clinch and Tennessee rivers resulted in 9 additional haplotypes for P. plenum (COI-

Pple09 – Pple17) and 3 additional haplotypes for P. sintoxia and P. rubrum (COI-Psr17 –

Psr19).

Sequences for ND1 typically were 744 bp, exhibiting 14 variable sites among haplotypes of F. flava, 35 for F. subrotunda, 83 for P. cordatum, 12 for P. plenum, and 37 for

P. sintoxia and P. rubrum (APPENDIX Table 1.2, APPENDIX Table 1.11 to 1.15). Molecular analysis of ND1 sequences resulted in 42 mussel specimens of F. flava representing 13 haplotypes, 20 mussel specimens of F. subrotunda representing 17 haplotypes, 116 mussel specimens of P. cordatum representing 58 haplotypes, 32 mussel specimens of P. plenum representing 12 haplotypes, and 41 mussel specimens of P. sintoxia and P. rubrum collectively representing 17 haplotypes. Some haplotypes from mussel specimens of the respective species collected in the Green River were shared by mussel specimens collected from the Clinch and

Tennessee rivers. However, numerous additional haplotypes were observed; for F. subrotunda representing 1 additional haplotype, P. plenum 9 haplotypes and P. sintoxia and P. rubrum representing 4 haplotypes.

The different sample sizes of mussel specimens for ND1 and COI are due to not all the samples working for the two markers. Combining the DNA sequences of the COI and ND1

17 genes resulted in a 1,215-bp sequence, resulting in 15 haplotypic combinations for F. flava, 19 for F. subrotunda, 76 for P. cordatum, 15 for P. plenum, and 26 for P. sintoxia and P. rubrum, collectively. Mussel specimens collected from the Clinch and Tennessee rivers resulted in one additional haplotype for F. subrotunda, 13 for P. plenum, and 5 for P. sintoxia and P. rubrum.

For the 16S rRNA gene, a 477 bp DNA sequence was obtained for 42 mussel specimens of P. sintoxia and P. rubrum in the Green River, which showed 10 variable sites comprising nine haplotypes (16S_Psr01 – Psr09) (APPENDIX Table 1.16). In addition, one mussel specimen each of F. flava (16S_Ffla), F. subrotunda (16S_Fsub), P. cordatum (16S_Pcor), and P. plenum (16S_Pple) were sequenced from the Green River, and 7 mussel specimens of

P. sintoxia and P. rubrum collected from the Clinch and Tennessee rivers, which resulted in one additional haplotype (16S_Psr10).

For the nuclear ITS1 DNA sequences, both the Clustal and WebPrank alignments resulted in a 448-bp sequence containing 5 variable sites, which included 3 encoded gaps in both alignments (APPENDIX Table 1.17 – 1.18). However, the positions of the encoded gaps differed between the two alignments. Five haplotypes were observed for P. sintoxia and P. rubrum (ITS1_Psr03 to ITS1_Psr07) in the Green River samples, and two additional haplotypes were observed in samples from the Clinch and the Tennessee rivers (ITS1_Psr01 to

ITS1_Psr02). In addition, ten mussel specimens of other species from the Green River were used as outgroups in the phylogenetic analysis, including 2 mussel specimens of F. flava

(ITS1_Ffla01 to ITS1_Ffla02), 3 F. subrotunda (ITS1_Fsub01 to ITS1_Fsub03), 2 P. cordatum

(ITS1_Pcor01 to ITS1_Pcor02), and 3 P. plenum (ITS1_Pple01 to ITS1_Pple03).

Genetic Diversity

For F. flava, F. subrotunda, and P. cordatum, observed haplotype and nucleotide diversities were higher for ND1 DNA sequences than for COI DNA sequences (Table 1.2 and

1.4). For P. plenum and P. sintoxia/rubrum, haplotype diversity was higher for ND1 sequences, while nucleotide diversity was higher for COI sequences.

For COI sequences, the species with the highest haplotype diversity was F. subrotunda

(0.896), followed by P. sintoxia/rubrum (0.850), P. cordatum (0.767), P. plenum (0.701), and

F. flava (0.339) (Table 1.2). Nucleotide diversity was highest in F. subrotunda (0.00808), followed by P. plenum (0.00564), P. cordatum (0.00423), P. sintoxia/rubrum (0.00382), and

F. flava (0.00078). For ND1 sequences, the species with the highest haplotype diversity was F. subrotunda (0.984), followed by P. cordatum (0.930), P. sintoxia/rubrum (0.890), F. flava

(0.875), and P. plenum (0.768). The nucleotide diversity was highest in F. subrotunda

(0.00886), followed by P. cordatum (0.00501), P. sintoxia/rubrum (0.00370), P. plenum

(0.00368), and F. flava (0.00278)

Among the investigated species in this study, P. cordatum exhibited the highest haplotype diversities at COI (0.77) and at ND1 (0.93) (Table 1.4). These values for ND1 were slightly lower than those reported by Jones et al. (2015) in the Green River, KY (0.97) and

Tennessee River, TN (1.0). However, their sample size, 18 mussel specimens for both rivers, was considerably smaller than my sample size. In my study, nucleotide diversity for P. cordatum ranged from 0.004 at COI and 0.005 at ND1, which were slightly higher than values reported by Jones et al. (2015), between 0.003 in the Green River, KY and 0.00361 in the

Tennessee River, TN. For P. plenum, my nucleotide diversity ranged between 0.00368 and

0.00564 for ND1 and COI, respectively. This was comparable to the values reported by Jones et al. (2015), in which ND1 nucleotide diversity was between 0.003 and 0.005 for the Green

River, KY and the Clinch River, TN respectively.

Phylogenetic Analysis

Phylogenetic trees of DNA sequence haplotypes were constructed to visualize relationships among the respective lineages. The topology of the COI tree, showing species identifications and clades (Figure 1.2), was consistent with that of Inoue et al. (2018). In addition, haplotypes for each species collected from the Green River consistently grouped into the same species clades as haplotypes from mussel specimens collected in the Clinch and

Tennessee rivers, including one mussel specimen of F. subrotunda (collected from the Clinch

River), 14 P. plenum (12 from the Clinch River and 2 from the Tennessee River), and 7 P. sintoxia and P. rubrum (4 from the Clinch River and 3 from the Tennessee River). The additional sequences of Fusconaia flava in my study grouped together with sequences of F. flava and F. cerina from Inoue et al. 2008 (APPENDIX Table 1.19). My sequences grouped together with species of Fusconaia askewi, F. cerina, F. chunni, F. lanensis and F. masoni.

However, a lower prior maximal distance in ABGD (P = 1.00 e-3) resulted in the F. flava haplotypes grouping with the haplotypes of Inoue’s for F. flava and with F. cerina. The Green

River F. flava haplotypes were different from some F. flava from Arkansas. In addition, my mussel specimens of P. sintoxia and P. rubrum grouped together in the same clade with those of the same species reported by Inoue et al. (2008). This is particularly interesting as the authors added sequences from several locations where P. rubrum and P. sintoxia occur. The P. sintoxia/rubrum clade was paraphyletic with Pleurobema riddelli. The latter lineage seems to be closely related to P. sintoxia/rubrum. Phylogenetic trees showing names for all ND1 haplotypes combinations is available in APPENDIX Figure 1.1.

However, in the ND1 tree, clades were grouped together with sequences obtained from

Bertram et al. (2015), Burlakova et al. (2012), Jones et al. (2015), Marshall et al. (2018), and

Schilling (2015) (APPENDIX Table 1.20). Well-defined clades were observed for F. subrotunda, P. cordatum, P. plenum, and P. sintoxia/rubrum. The latter showed mussel

20 specimens from this and other studies grouping together, suggesting that P. sintoxia and P. rubrum are conspecific (Figure 1.3). In the case of F. flava, an initial partition in ABGD resulted in F. flava haplotypes grouping together with F. askewi, F. lanensis and F. masoni. A recursive partition separated the F. flava sequences from F. askewi, F. lanensis, and F. masoni.

Like results observed in the COI tree, F. askewi and F. lanensis grouped together. Phylogenetic trees showing names of the ND1 haplotypes combinations is available in APPENDIX Figure

1.2.

The phylogenetic tree constructed from combined COI and ND1 sequences resulted in five well-defined clades that included F. flava, F. subrotunda, P. cordatum, P. plenum, and P. sintoxia/rubrum (Figure 1.4). As the haplotype network for both COI and ND1 showed a low number of mutational steps, there is little evidence to support these clades as different species.

Phylogenetic trees showing names for all COI and ND1 haplotypes combinations is in

APPENDIX Figure 1.3.

The 16S rRNA tree did not show species-level differentiation between P. sintoxia and

P. rubrum (Figure 1.5). For ITS1 sequences, phylogenetic trees constructed with either the

Clustal or WebPRANK sequence alignments grouped haplotypes from P. cordatum and P. plenum together (Figures 1.6). There was clear separation, however, between haplotypes of F. flava and F. subrotunda.

If hybridization and backcrossing has occurred between P. sintoxia and P. rubrum, then mtDNA is not an appropriate marker to delineate these two species The fixation index (FST =

0.0028, P-value < 0.001) as reported in Chapter 2 was estimated from DNA microsatellites and showed that population differentiation between the sites located in Pool 4 and MCNP was low.

The results suggest that there were significant but minor differences between these two sampling locations. The Structure Harvester results suggested that the most appropriated number of clusters was K=2 according to the Delta-K method, however the mean LnP(K)

21 supported K=1 which matches the fixation index results (APPENDIX Table 1.21). Hence, the existence of only one cluster (K=1) was better supported. The nuclear DNA microsatellite results also suggest that these two taxa only represent one species in the Green River, Kentucky.

Haplotype Networks

For all four DNA sequence markers (COI, ND1, 16s rRNA, and ITS1) used in this study, the haplotype networks showed five distinct clades corresponding to the respective study species in the Green River, including F. flava, F. subrotunda, P. cordatum, P. plenum, and P. sintoxia and rubrum together (Figures 1.7 – 1.11, respectively). Mussel specimens within the

P. sintoxia/rubrum clade were not separated by a high number of mutational steps. The number of mutational steps between haplotypes ranged from 1-8 for COI, ND1, 16S rRNA, and ITS1.

This suggests that these mussel specimens may belong to a single species.

The COI haplotype network showed that a low number of nucleotide differences (15 nucleotides) separated F. flava and F. subrotunda (Figure 1.7). The number of nucleotide differences also was low between F. flava and F. subrotunda in the ND1 haplotype network

(Figure 1.8), with 32 nucleotide differences between these two species. In the case of F. subrotunda, the next most closely related species (after F. flava) was P. cordatum (14 nucleotide differences with COI and 27 with ND1). Results from 16S rRNA haplotype network matched results for COI and ND1 markers (Figure 1.9). Both ITS1 alignments showed only a few mutation steps (Figure 1.10-1.11).

In the ND1 haplotype network, we can observe that the haplotype that showed the highest number of mutational steps (8 mutational steps) from the nearest haplotype was

ND1_PSR09. This haplotype belongs to an individual collected from Pool 4 which was identified as a different species (F. subrotunda, P. plenum, and P. sintoxia) by each of the 3 experts that identify it in the field. In the 16S rRNA haplotype network, we observed that the

22 haplotype showing the largest number of mutational steps (3 mutational steps) was 16S_PSR07 which was amplified for only one individual (WG576) collected from Pool 4. This individual was identified 75% of the times as P. sintoxia by the experts. However, these number of mutational steps is too low in comparison to the number of mutational steps separation other species using the same markers (Figures 1.7 – 1.11).

Split Networks

Topologies of the split networks resulting from analysis of COI, ND1 and COI+ND1 sequences (Figures 1.12 through 1.16, respectively) were consistent and showed five distinct clades, F. flava, F. subrotunda, P. cordatum, P. plenum, and P. sintoxia/rubrum. In the COI split networks (Figure 1.12), I added sequences of P. sintoxia and P. rubrum from Inoue et al.

(2018), which clustered within my P. sintoxia/rubrum clade. The same results were observed when ND1 and 16S rRNA sequences of P. sintoxia and P. rubrum from Jones et al. (2015) were included (Figure 1.13 and 1.14). Split networks constructed from ITS1 sequences included both

P. sintoxia and P. rubrum, as well as outgroups from other species, such as F. flava, F. subrotunda, P. cordatum, and P. plenum. For both ITS1 alignments, one of the haplotypes of

P. sintoxia/rubrum (ITS1_Psr05) seemed to be particularly distinct from the other haplotypes in this clade. However, mussel specimens from the same outgroup species fell into different clades, making inference of species-level differentiation using ITS1 sequences unreliable

(Figure 1.15 and 1.16).

Pairwise Differentiation

In the case of the mitochondrial DNA COI gene, pairwise differentiation values using the Tamura-Nei with an invariable sites nucleotide mutation model for the COI sequences ranged between 0.043 between F. subrotunda and F. flava and 0.091 between P. plenum and

P. sintoxia/rubrum, whereas pairwise differentiation calculated using a p-distance model among species ranged from 0.041 between F. flava and F. subrotunda to 0.083 between P. sintoxia/rubrum and P. plenum. The p-distance pairwise differentiation within species was highest for P. plenum at 0.006, and lowest for F. flava at 0.001 (Table 1.4). In the case of NDI, pairwise differentiation values using the Tamura-Nei with invariable sites model ranged from

0.050 between F. subrotunda and F. flava and 0.108 between F. subrotunda and P. sintoxia/rubrum, whereas pairwise differentiation calculated using p-distances ranged from

0.046 between F. flava and P. cordatum to 0.098 between P. cordatum and P. sintoxia/rubrum.

Pairwise differentiation within species was highest for F. subrotunda at 0.009, while the lowest value was estimated for F. flava at 0.003. The pairwise p-distance results showed that mean intraspecific differentiation among individuals of F. subrotunda was 0.8% for COI and 0.9% for ND1, which were the highest intraspecific distances observed for the species studied in the

Green River. My results are similar to those reported by Schilling (2015), who reported an ND1 intraspecific difference of 1% among individuals of F. subrotunda in the upper Tennessee

River basin. Higher intraspecific differences values for F. subrotunda were reported by

Burlakova et al. (2012), in which COI pairwise differences ranged from 1.23 to 1.24%, while

ND1 pairwise differences ranged between 1.11 and 1.30%. Interspecific COI pairwise differences showed that the species most closely related to F. subrotunda was F. flava (4.1%), followed by P. cordatum (4.2%). These results were comparable to the ND1 pairwise differences, 4.7% among F. subrotunda and both F. flava and P. cordatum. In addition, pairwise differences paralleled results from the phylogenetic trees, split networks, and haplotype networks constructed using COI, ND1, and COI + ND1 sequences. Species differentiation tested with ABGD resulted in a well-differentiated F. subrotunda clade.

Intraspecific pairwise p-distances for mussel specimens of P. cordatum ranged between

0.4% for COI and 0.5% for ND1. My results were similar to those reported by Jones et al.

(2015), in which ND1 intraspecific distances for P. cordatum ranged between 0.4% and 0.7%.

Results for interspecific pairwise distances showed that the highest differences were between

P. cordatum and P. sintoxia/rubrum, with 6.9% for COI and 9.8% for ND1. The second-most differentiated species from P. cordatum was P. plenum, with 6.2% p-distance for COI and 7.8% for ND1.

Intraspecific variation among mussel specimens of P. plenum ranged between 0.4% and

0.6% for COI and ND1 sequences, respectively. My results were similar to those reported by

Jones et al. (2015), who reported ND1 intraspecific pairwise differences between 0.6% (Green

River, Kentucky) and 0.8% (Tennessee River, Tennessee) for P. plenum. In this study, intraspecific pairwise differences were highest between P. plenum and P. sintoxia/rubrum, which were between 8.3% and 8.8% for COI and ND1 sequences, respectively. Finally, intraspecific pairwise differences for P. sintoxia/rubrum were 0.4% for both COI and ND1 sequences, which were comparable to those of Jones et al. (2015), who reported ND1 intraspecific pairwise differences of 0.1% for P. sintoxia and 0.8% for P. rubrum.

DISCUSSION

Most mussel species show substantial differences in their shell morphology, life history, genetic diversity among one another and are easy to identify. However, species belonging to the genera Fusconaia and Pleurobema are particularly difficult to identify even by trained eyes. My phylogenetic assessment of Fusconaia and Pleurobema species in the

Green River, KY resulted in the identification of five well differentiated clades (F. flava, F. subrotunda, P. cordatum, P. plenum, P. sintoxia/rubrum). Study results did not uncover cryptic species and did not differentiate individuals of P. rubrum and P. sintoxia as separate species.

Finally, high observed haplotype diversity values suggest that these investigated populations in the Green River are genetically healthy.

Molecular markers

The idea behind using sequences to “barcode” species relies on intraspecific variation being clearly lower than interspecific variation for mitochondrial markers (COI, ND1, 16s rRNA) as well as nuclear marker (ITS1). The most used mitochondrial marker for DNA barcoding in eukaryotes is the COI gene (Bleidorn, 2017). The principal reasons for the use of mitochondrial markers for barcoding is that practically there are universal primers available, larger numbers of mtDNA copies per cell relative to nuclear DNA, and high interspecific variation that gives rise to so-called barcoding “gaps”. However, the use of only mitochondrial markers in phylogenetic studies has been criticized due to their solely maternal inheritance, inconsistent mutation rate, introgression, low effective population size, and heteroplasmy or pseudogenization (Bleidorn, 2017). In this study, for COI and ND1 primers, some mussel specimens of P. plenum and P. sintoxia/rubrum may have amplified or not amplified with one primer pair combination due to DNA sequence variation at the primer binding site.

The principal limitation for the 16S rRNA and ITS1 markers was that they did not have enough fixed nucleotide mutational steps to separate some of the study species. Species delimitation using AGBD was not possible for the 16S rRNA or ITS1 sequences because of their low nucleotide variability. Even though there were not as many variable sites separating species with 16S rRNA, species identification matched the identification obtained with the COI and ND1 genes. The limitation arises more with the use of the ITS1 sequences. The ITS1 marker over-splitted the sequences into groups, randomly separating species which were well defined with the mitochondrial markers, over splitting occurred for sequences of F. flava and P. cordatum. The sequences of P. sintoxia/rubrum also were over-splitted. The ITS1 sequence for the only individual in the P. sintoxia/rubrum clade (define by mitochondrial markers) that was consistently identified as P. rubrum by the 5 experts was ITS1_PSR4. Over splitting could be due to higher intraspecific variation relative to the interspecific variation. The sequencing of

26 this marker is challenging because more than one copy of this gene exists for each individual, as two copies are inherited from the parents. Other challenges for the ITS1 marker are the presence of gaps in the aligned sequences. The number of nucleotide differences was not high enough for successful use of species delimitation using ABGD. The two haplotype sequences for F. flava (ITS1_Ffla01 and ITS1_Ffla02) and F. subrotunda (ITS1_Fsub01 and

ITS1_Fsub02) came from two different sampling locations (Pool4 and MCNP) and this could explain the oversplitting within these two species. However, the data for these two species is limited to conclude that ITS1 sequences are appropriate for population level studies. The results from Elderkin, 2009 results suggested the contrary. In his study, the author amplified ITS1 sequences from Cumberlandia monodonta individuals. His ITS1 sequences showed that there is a higher genetic diversity within individuals due to heterozygous individuals compared to genetic variation among individuals (Elderkin 2009). In addition, separation of haplotypes of

F. flava (FST = -0.019, P = 0.776) and F. subrotunda (FST = 0.077, P = 0.077) per population is not supported by genetic differentiation analysis using COI + ND1 which suggested that individuals from Pool 4 and MCNP belong to the same population. There are four individuals in the P. sintoxia/rubrum clade that separated in two additional clades when COI and ND1 were combined together and species delimitation was assessed (Figure 1.4). Two of these individuals shared the ITS1_PSR03 (tag # WE779 from Pool 4, and tag # BLU009 from the

Clinch River) sequence. The other two individuals, one from Pool 4 (tag # WG591) and another from the Clinch River (tag#RubClinch) have ITS1_ PSR04 and ITS1_PSR01, respectively. The individuals WG591 was the only individual identified as P. rubrum by the 5 experts. My results for the ITS1 marker matched the results of Schilling (2015), who observed that some estimates for interspecific variation were lower than intraspecific variation estimates.

Phylogenetic assessment

My phylogenetic assessment resulted in the identification of five phylogenetic clades that corresponded to the respective investigated species, to include: F. flava, F. subrotunda, P. cordatum, P. plenum and a clade that included all specimens of P. sintoxia and P. rubrum for which not enough nucleotide differences were detected to consider them different species.

However, the main P. sintoxia and P. rubrum clade did contain sub-clade structuring that was comprised of four individuals. Two of these 4 individuals were collected from the Green River and one of them was consistently identified as P. rubrum (WG591) by the 5 experts while the other individual (WE779) was identified as P. sintoxia by 3 out of 4 experts. The assessment for cryptic species among watersheds did not suggest the discovery of any new species. That is, none of the DNA sequences from the Green River separated into their own clades or grouped with any of the DNA sequences that I added into the analysis from outside the drainage.

Delineation of species in the genera Fusconaia and Pleurobema is supported by a suite of morphological and life-history traits. In the Green River, one of the principal morphological difference which generally is reliable is foot color. Mussels in the genus Pleurobema generally have a white foot color, whereas mussels in the genus Fusconaia typically have an orange colored foot. In the case of life-history traits, two important characters for identification of

Fusconaia and Pleurobema species are conglutinate shape and the number of gills used to brood eggs and larvae. In the case of Fusconaia species, conglutinates are slender and subcylindrical and all four gills are used to brood eggs and larvae, whereas for Pleurobema conglutinates are leaf-like and can have different layers and only the outer pair of gills are used to brood eggs and larvae (Haag and Warren 2003; Barnhart et al. 2008). The two Fusconaia species in the Green River were characterized by distinct shell shapes, with F. flava being trapezoidal and F. subrotunda being more rounded and elongate.

Pyramid and Round pigtoes

In this study, I collected a relatively large number of individuals identified morphologically as P. sintoxia and P. rubrum from the Green River, KY to enhance the probability of delineating these two nominal taxa and for detecting any cryptic species which could have small populations and prove similar in appearance to these two species. By using a large sample size, the intraspecific and interspecific nucleotide differences are enhanced and better characterized, thus allowing for easier identification of species-level differences among taxa. Further, I added DNA sequences of P. sintoxia and P. rubrum from other studies, to include COI (Inoue et al. 2018) and ND1 (Jones et al. 2015), and all these sequences grouped together phylogenetically. Thus, even when utilizing all these DNA sequences I did not observe clear differentiation between nominal P. rubrum and P. sintoxia. When forcing the species delimitation for a higher species separation, I observed the formation of not two, but rather three clades when screening COI and ND1 sequences. One of the individuals was haplotype

COI_PSR14_and_ND1_PSR14, which was consistently identified as P. rubrum by the experts

(tag#WG591), the other individual (tag#WE779) was identified as P. sintoxia by 3 out of 4 experts. The ITS1 sequences for these individuals were ITS1_PSR3 and ITS1_PSR4, respectively. However, this phylogenetic separation was not corroborated by results of the other phylogenetic analysis methods applied, such as construction of haplotype networks and split networks. Morphologically, individuals of P. sintoxia and P. rubrum can look very distinct from each other, especially for the typical morphological shapes of the two shell forms. For P. rubrum, the shell shape is typically a scalene triangle with beaks facing forward and with a very marked sulcus that traverses the shell form near the umbo to the ventral margin, whereas for P. sintoxia, individuals are much more rounded in shape with beaks facing each other, and without a well-defined sulcus. Interestingly, separation of these two putative species seems to be supported by morphological differences in their glochidia (Culp et al. 2009). Thus, it is

29 important to investigate quantitatively if glochidial differences exist between the P. rubrum and P. sintoxia shell forms. Additional lines of evidence to test the species boundaries between these two taxa should include phylogenomic studies, which could test for potential phylogenetic differentiation at the genome level. In comparison to traditional phylogenetics which is conducted by sanger sequencing, phylogenomics uses much larger amounts of DNA sequence data, which increases the number of markers dramatically and reduces the sampling error. Hence, the use of phylogenomics presumably would result in better taxonomic resolution. Currently, the RAD sequencing approach is being used to assess phylogenomic differentiation between P. sintoxia and P. rubrum (N. Johnson, U.S. Geological Survey, personal communication). Another line of evidence to be tested in the future might include how the shape and size of the shell of P. sintoxia and P. rubrum changes morphologically due to changes in stream size from small to large river environments. Known as Ortmann’s (1920) law of stream position, such morphological changes have been described for F. flava for example, where the shell inflation index (width/length x100) increases from small streams to large rivers (Haag 2012).

Hybridization and backcrossing between individuals of P. rubrum and P. sintoxia may have resulted in introgression of the mtDNA genome between these species, making them indistinguishable at this marker. However, the results from the Structure analysis tested in this chapter and the FST values obtained in Chapter 2 suggested that there was only one multilocus genotypic cluster. Assuming that these two taxa represent distinct species in the Green River, hybridization processes may have resulted in genetic introgression. The interpretation of this data should be constrained to the individuals collected from the Green River only, and such hybridization and introgression processes should be tested for in the different drainages where these two putative species occur together. Other examples in the genus Pleurobema, exist of species that are phylogenetically indistinguishable at mtDNA but are morphologically distinct,

30 such as P. clava and P. oviforme, for example (C. Morrison, U.S. Geological Survey, personal communication).

Genetic diversity

The results from my phylogenetic analysis allowed me to estimate metrics of genetic diversity. In general, the investigated species’ ND1 and COI haplotype diversities in this study were high. One reason for this high genetic diversity could be that these species historically occurred in much larger interconnected populations that were linked to the main stem of the

Ohio River and its tributaries. The high genetic diversity currently observed among these species in the Green River could be the result of this demographic signal still being maintained in these populations. The effect of impoundments on these mussels is described by Haag and

Cicerello (2016), and may be one of the principal reasons for decline of species such as P. clava and many other pigtoe species throughout the Ohio River system, as many of these species are intolerant of the altered flow conditions caused by impoundments and dams.

However, other species such as P. cordatum seem more tolerant of impoundments and may adapt to these altered riverine systems. In contrast, P. plenum seems minimally tolerant of impoundments, and thus much of its historically suitable habitats throughout the Ohio River system have been altered or destroyed by dams, and likewise both P. rubrum and P. sintoxia are at best marginally tolerant to impounded stream conditions. Future studies should assess the effect of impoundment on shell shape variation among individuals of P. sintoxia/rubrum.

Individuals of F. flava seem to be able to adapt to a variety of habitats, and its population are generally stable throughout the species range, whereas F. subrotunda prefers unimpounded large stream environments and is declining throughout its range. Assessment of recruitment is important for all of these species, as unsuitable water quality and altered hydrology can decrease or even halt reproduction, and the high haplotype diversity that I observed could be a

31 measure of old non-recruiting and demographically imperiled populations. Recruitment failure in these populations would be catastrophic, as the extirpation of these populations from the

Green River would have serious consequences for the long-term conservation of these species.

Fortunately, there is evidence of recruitment for all five investigated species in the Green River, which is one of the best refuge strongholds for these and many other species in the Ohio River system. Periodic monitoring to assess the abundance, recruitment and genetic diversity of the

Green River mussel fauna will be critical for managing the population health of these species.

Management Implications

The IUCN Red List status for F. flava is “least concern” and surprisingly, the species showed the lowest haplotype and nucleotide diversity of my study species. The effective population size (Ne) has not been estimated for this species mainly due to a lack of PCR primers for DNA microsatellites specifically designed for this species or even for a closely related

Fusconaia species. The Fusconaia subrotunda clade was well supported phylogenetically and was the clade with the highest nucleotide and haplotype diversities in the study. Principal concerns regarding management of the species include continued demographic declines as large to medium sized free-flowing riverine habitats are lost (Haag and Cicerello 2016). Future efforts are needed to develop nuclear DNA genetic markers to estimate Ne as well as genetic diversity of this species in the Green River.

Pleurobema cordatum numbers also have declined range-wide likely due to the reduction of large river habitats. However, P. cordatum seems to be more tolerant to impoundments than F. subrotunda and other Pleurobema species (Haag and Cicerello 2016).

This species showed higher nucleotide diversity (π) and smaller haplotype diversities (h) than other P. cordatum populations reported for the Green and the Tennessee Rivers by Jones et al.

(2015).

The federally protected Pleurobema plenum has been listed as endangered since 1976 and its recovery plan approved in 1984. Pleurobema plenum is not very tolerant of impoundments and has been extirpated from most of its historical range. Because P. plenum is sensitive to habitat modification, its critical habitat (medium to large-sized rivers) must be protected. However, high haplotype diversity suggests that the P. plenum population is healthy and reasonably abundant in the Green River, KY. Recruitment and abundance of this species needs to be regularly monitored to ensure these values are not indicative of an aging, potentially non-recruiting population.

While both P. sintoxia and P. rubrum appear to belong to only one phylogenetic clade based primarily on mtDNA, these two nominal taxa may need to be treated as a single species in future management plans once additional morphological and nuclear DNA marker-based studies have been completed. Similar to most of the species collected from the Green River, mussel specimens belonging to the P. sintoxia/rubrum clade seem to be marginally tolerant to even intolerant of impounded riverine conditions (Haag and Cicerello 2016). However, there are ongoing studies assessing morphological differentiation between these two putative species by Dr. Monte McGregor at the Center for Mollusk Conservation of the Kentucky Wildlife

Resources Agency, who is currently assessing glochidial morphological differences. In addition, Dr. Nathan Johnson at the U.S. Geological Survey, Wetland and Aquatic Research

Center, Gainesville, Florida is currently using RAD-seq to delimit these two species.

Finally, in contrast to Schilling (2015) who performed a similar molecular marker- and morphology-based study of mussels in the Tennessee River basin and found evidence of cryptic species in the genus Pleurobema and Pleuronia, I did not find cryptic species in the Green

River. An additional species that has been reported for the Green River, KY, the endangered clubshell pearly mussel (P. clava) (Haag and Cicerello 2016), still occurs in the river upstream of the sampling sites, but was not found during the field collections in Pool 4 and MCNP. This

33 species was reported from the Green River (Kentucky) in Hart and Taylor counties by Watters

(1994). However, he did not find live mussels of this species, but rather only fresh-dead shells.

Thus, future studies are needed to monitor the recruitment, abundance and genetic diversity of this species in the river to determine its overall population health.

COLLABORATORS’ CONTRIBUTIONS

Sampling in the Green River, KY was conducted in collaboration with Chad Lewis and his crew at Lewis Environmental Consulting, LLC and Dr. Monte McGregor, Kentucky

Department of Wildlife Resources, Frankfort, KY. Species identifications were performed by

Leroy Koch, Dr. Wendell Haag, Chad Lewis, Dr. Monte McGregor, and Adam Shephard. DNA collection and tagging were done with help from Aaron Adkins, Anna Dellapenta, Jess Jones,

Tim Lane, and Lee Stephens. Murray Hyde helped with DNA collection, mussel tagging, and lab work. Insightful input for the methods and discussion was provided by committee members

Drs. Eric Hallerman, Jess Jones, Emmanuel Frimpong and Pawel Michalak. I aided in the DNA collection, mussel tagging, lab work, data analysis and prepared the dissertation manuscript with assistance from Eric Hallerman and Jess Jones.

LITERATURE CITED

Abernethy, E., E. McCombs, L. Siefferman, and M. Gangloff. 2013. Effect of small dams on freshwater mussel population genetics in two southeastern USA streams. Walkerana - The Journal of the Freshwater Mollusk Conservation Society 16:21-28.

Barnhart, M. C., W. R. Haag, and W. N. Roston. 2008. Adaptations to host infection and larval parasitism in Unionoida. Journal of the North American Benthological Society 27:370- 394.

Bertram, E. P. 2015. Confirmation of Potential Cyprinid Hosts for a State Threatened Freshwater Mussel of East Texas.

Bleidorn, C. 2017. Phylogenomics. Cham: Springer International Publishing, 35-36.

Borchsenius, F. 2009. FastGap 1.2. Department of Biological Sciences, University of Aarhus, Aarhus, Denmark. https://www.aubot.dk/FastGap_home.htm.

Bryant, D., and V. Moulton. 2002. NeighborNet: An agglomerative method for the construction of planar phylogenetic networks. International Workshop on Algorithms in Bioinformatics 375-391.

Buhay, J. E., J. M. Serb, C. R. Dean, Q. Parham, and C. Lydeard. 2002. Conservation genetics of two endangered unionid bivalve species, Epioblasma florentina walkeri and E. capsaeformis (Unionidae: Lampsilini). Journal of Molluscan Studies 68:385–391.

Burlakova, L. E., D. Campbell, A. Y. Karatayev, and D. Barclay. 2012. Distribution, genetic analysis and conservation priorities for rare Texas freshwater molluscs in the genera Fusconaia and Pleurobema (Bivalvia: Unionidae). Aquatic Biosystems 8:12.

Campbell, D. C., and C. Lydeard. 2012a. Molecular systematics of Fusconaia (Bivalvia: Unionidae: Ambleminae). American Malacological Bulletin 30:1-17.

Campbell, D. C., and C. Lydeard. 2012b. The genera of Pleurobemini (Bivalvia: Unionidae: Ambleminae). American Malacological Bulletin 20:19-38.

Campbell, D. C., P. D. Johnson, J. D. Williams, A. K. Rindsberg, J. M. Serb, K. K. Small, and C. Lydeard. 2008. Identification of ‘extinct’ freshwater mussel species using DNA barcoding. Molecular Ecology Resources 8:711-724.

Clement, M., Q. Snell, P. Walker, D. Posada, and K, Crandall. 2002. TCS: Estimating gene genealogies. Parallel and Distributed Processing Symposium, International Proceedings 2: 184.

Culp, J. J., Shepard, A. C., and McGregor, M. A. 2009. Fish hosts and conglutinates of the pyramid pigtoe (Pleurobema rubrum). Southeastern Naturalist, 8, 19-22.

Earl, D. A. 2012. Structure harvester: a website and program for visualizing structure output and implementing the Evanno method. Conservation Genetics Resources 4:359-361. Elderkin, C. L. 2009. Intragenomic variation in the rDNA internal transcribed spacer (ITS1) in the freshwater mussel Cumberlandia monodonta (Say, 1828). Journal of Molluscan Studies 75, 419-421.

Excoffier, L., and H. E. Lischer. 2010. Arlequin suite ver. 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology Resources 10: 564-567.

Folmer, O., M. Black, W. Hoeh, R. Lutz, and R. Vrijenhoek. 1994. DNA primers for amplification of mitochondrial cytochrome oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology 3:294-299.

Graf, D. L., and K. S. Cummings. 2007. Review of the systematics and global diversity of freshwater mussel species (Bivalvia: Unionoida). Journal of Molluscan Studies 73:291-314.

Haag, W. R. 2012. North American Freshwater Mussels: Natural History, Ecology, and Conservation. Cambridge University Press.

Haag, W. R., and M. L. Warren. 2003. Host fishes and infection strategies of freshwater mussels in large Mobile Basin streams, USA. Journal of the North American Benthological Society 22: 78–91.

Haag, W. R., and R.R. Cicerello. 2016. A Distributional Atlas of the Freshwater Mussels of Kentucky. Kentucky State Nature Preserves Commission, Frankfort, KY, pages 391- 395.

Heard, W. H., and R. H. Guckert. 1970. A re-evaluation of the recent Unionacea (Pelecypoda) of North America. Malacologia 10:333-355.

Huson, D. H. and D. Bryant. 2006. Application of phylogenetic networks in evolutionary studies. Molecular Biology and Evolution 23:254-267.

Inoue, K., A. L. McQueen, J. L. Harris, and D. J. Berg. 2014. Molecular phylogenetics and morphological variation reveal recent speciation in freshwater mussels of genera Acidens and Arkansia (Bivalvia: Unionidae). Biological Journal of the Linnaean Society 112: 535-545.

Inoue, K., D. M. Hayes, J. L. Harris, N. A. Johnson, C. L. Morrison, M. S. Eackles, T. L. King, J. W. Jones, E. M. Hallerman, A. D. Christian, and C. R. Randklev. 2018. The Pleurobemini (Bivalvia: Unionida) revisited: molecular species delineation using a mitochondrial DNA gene reveals multiple conspecifics and undescribed species. Invertebrate Systematics 32:689-702.

Kimura, M. 1980. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. Journal of Molecular Evolution 16: 111-120.

King T. L., M. S. Eackles, B. Gjetvaj, and W. R. Hoeh. 1999. Intraspecific phylogeography of Lasmigona subviridis (Bivalvia: Unionidae): Conservation implications of range discontinuity. Molecular Ecology 8: S65-S78.

Leigh, J. W., and D. Bryant. 2015. POPART: full-feature software for haplotype network construction. Methods in Ecology and Evolution 6: 1110–1116.

Löytynoja, A., N. Goldman. 2010. webPRANK: a phylogeny-aware multiple sequence aligner with interactive alignment browser. BMC Bioinformatics 11: 579.

Master, L. L., S. R. Flack, and B. A. Stein. 1998. Rivers of life: critical watersheds for protecting freshwater biodiversity. Nature Conservancy, Arlington, Virginia: 71 pp.

Marshall, N. T., J. A. Banta, L. R. Williams, M. G. Williams, and J. S. Placyk. 2018. DNA barcoding permits identification of potential fish hosts of unionid freshwater mussels. American Malacological Bulletin 36 :42-56.

Miller, E. J., K. J. Couch, and J. Mason. 2008. A Pocket Guide to Kansas Freshwater Mussels. Friends of the Great Plains Nature Center, Wichita, Kansas 72 pp.

Nylander, J. A. A., 2008. MrModeltest 2.3. http://www.abc.se/~nylander/mrmodeltest2/mrmodeltest2.html.

Palumbi, S., A. Martin, S. Romano, W. O. McMillian, L. Stice, G. Grabowski, 1991. The Simple Fool’s Guide to PCR. Privately published document compiled by S. Palumbi, Dept. Zoology, University of Hawaii, Honolulu, HI, 96822.

Patwardhan, A., S. Ray, and A. Roy. 2014. Molecular markers in phylogenetic studies - a review. Journal of Phylogenetics and Evolutionary Biology 2: 1-9.

Pritchard, J.K., M. Stephens, and P. Donnelly. 2000. Inference of population structure using multilocus genotype data. Genetics 155:945-959.

Puillandre, N., A. Lambert, S. Brouillet, and G. Achaz. 2012. ABGD, Automatic barcode gap discovery for primary species delimitation. Molecular ecology 21: 1864-1877.

Rambaut, A., M. A. Suchard, D. Xie, and A. J. Drummond. 2009. MCMC trace analysis tool, version v1. 6.0. Institute of Evolutionary Biology, University of Edinburgh. http://tree.bio.ed.ac.uk/software/tracer/

Rambaut, A. 2014. FigTree v1. 4.2: Tree figure drawing tool. University of Edinburgh, Edinburgh. http://tree.bio.ed.ac.uk/software/figtree/

Ronquist, F., M. Teslenko, P. Van Der Mark, D. L. Ayres, A. Darling, S. Höhna, B. Larget, L. Liu, M. A. Suchard, and J. P. Huelsenbeck. 2012. MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Systematic Biology 61: 539-542.

Rozas, J. P. Librado, J. C. Sanchez-Delbarrio, X. Messeguer, and R, Rozas. 2009. DnaSP, version 5.10.00. Universita de Barcelona, Spain. http://www.ub.edu/dnasp/

Schilling, D. E. 2015. Assessment of morphological and molecular genetic variation of freshwater mussel species belonging to the genera Fusconaia, Pleurobema, and Pleuronaia in the upper Tennessee River basin. Master’s Thesis, Virginia Tech, Blacksburg, VA. 1-66

Serb, J. M. and C. Lydeard. 2003. Complete mtDNA sequence of the North American freshwater mussel, Lampsilis ornata (Unionidae): An examination of the evolution and phylogenetic utility of mitochondrial genome organization in Bivalvia (Mollusca). Molecular Biology and Evolution 20: 1854-1866.

Serb, J. M., J. E. Buhay, and C. Lydeard. 2003. Molecular systematics of the North American freshwater bivalve genus Quadrula (Unionidae: ambleminae) based on mitochondrial ND1 sequences. Molecular Phylogenetics and Evolution 28: 1-11.

Smith, C. H., N. A. Johnson, J. M. Pfeiffer, and M. M. Gangloff. 2018. Molecular and morphological data to facilitate future research on freshwater mussels (Bivalvia: Unionidae: Anodontinae). Data in Brief 17: 95-104.

Stansbery, D. H. 1967. A provisional classification of the Pleurobema cordatum complex in the Mississippi drainage basin of North America. Ohio State Museum 2.

Swofford, D. L. 1998. PAUP 4.0 beta version for Windows: phylogenetic analysis using parsimony. Sinauer Associates, Sunderland, Massachusetts.

Tamura K., G. Stecher, D. Peterson, A, Filipski, and S. Kumar. 2013. MEGA6: Molecular evolutionary genetics analysis version 6.0. Molecular Biology and Evolution 30: 2725-2729.

Thompson, J. D., T. J. Gibson, and D. G. Higgins. 2003. Multiple sequence alignment using ClustalW and ClustalX. Current protocols in bioinformatics 2-3.

Varela, E. S., C. R. Beasley, H. Schneider, I. Sampaio, N. D. S. Marques-Silva, and C. H. Tagliaro, 2007. Molecular phylogeny of mangrove oysters (Crassostrea) from Brazil. Journal of Molluscan Studies 73: 229-234.

Walker, J. M., J. P. Curole, D. E. Wade, E. G. Chapman, A. E. Bogan, G. T. Watters, and W. R. Hoeh. 2006. Taxonomic distribution and phylogenetic utility of gender-associated mitochondrial genomes in the Unionoida (Bivalvia). Malacologia 48: 265-282.

Watters, G. T. 1994. Clubshell (Pleurobema clava) and northern riffleshell (Epioblasma torulosa rangiana) recovery plan. Region Five, U.S. Fish and Wildlife Service Region 5 Office, West Hadley, MA. https://www.fws.gov/northeast/pafo/pdf/riffleshell_recovery_plan.pdf. 67 pages.

Table 1.1 Mitochondrial DNA primers for NADH dehydrogenase 1 (ND1), cytochrome oxidase subunit I (COI), and 16S rRNA genes and nuclear DNA primers for the ribosomal internal transcribed spacer region subunit 1 (ITS1) used for amplification of DNA sequences and genetic analysis of freshwater mussel species belonging to the genera Fusconaia and Pleurobema.

Gene Primer name Sequence Reference ND1 LeuuurF F: 5’- TGG CAG AAA AGT GCA TCA GAT TAA AGC - 3’ Serb et al. 2003 LoGlyR R: 5’- CCT GCT TGG AAG GCA AGT GTA CT - 3’ Serb et al. 2003 nadh1-F F: 5’ - TGG CAG AAA AGT GCA TCA GAT TTA AGC - 3’ Buhay et al. 2002, Serb and Lydeard 2003 nadh1-R R: 5’ - GCT ATT AGT AGG TCG TAT CG - 3’ Buhay et al. 2002, Serb and Lydeard 2003 COI LCO1490 F: 5’- GGT CAA CAA ATC ATA AAG ATA TTG G -3’ Folmer et al. 1994 CO1F F: 5’- GTT CCA CAA ATC ATA AGG ATA TTG G -3’ Campbell et al. 2005 HCO700dy2 R: 5’- TCA GGG TGA CCA AAA AAY CA -3’ Walker et al. 2006 16SL1987 F: 5’- GCC TCG CCT GTT TAC CAA AAA C - 3' Varela et al. 2007 16S rRNA 16Sbr-H R: 5’ - CCG GTC TGA ACT CAG ATC ACG -3’ Palumbi et al. 1991 18S F: 5’- AAA AAG CTT CCG TAG GTG AAC CTG CG- 3’ King et al. 1999 ITS1 5.8S R: 5’ - AGC TTG CTG CGT TCT TCA TCG -3’ King et al. 1999

Table 1.2 Intraspecific variation of the mitochondrial DNA COI gene for species in the genera Fusconaia and Pleurobema. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Average number of Number of Number of nucleotide Haplotype Species N Nucleotide diversity (π) variable sites haplotypes differences k diversity (h) (range)

Fusconaia flava 43 6 7 0.368 0.339 0.00078 Fusconaia subrotunda 22 14 13 3.805 0.896 0.00808 Pleurobema cordatum 117 45 43 1.991 0.767 0.00423 Pleurobema plenum 33 11 8 2.655 0.701 0.00564 Pleurobema sintoxia/rubrum 43 21 16 1.797 0.850 0.00382

Table 1.3. Intraspecific variation of the mitochondrial DNA ND1 gene for species in the genera Fusconaia and Pleurobema. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Average number of Number of Number of Haplotype Species N nucleotide Nucleotide diversity (π) variable sites haplotypes diversity (h) differences k (range)

Fusconaia flava 42 14 13 2.072 0.875 0.00278 Fusconaia subrotunda 20 35 17 6.595 0.984 0.00886 Pleurobema cordatum 116 54 58 3.724 0.930 0.00501 Pleurobema plenum 32 17 12 2.738 0.768 0.00368 Pleurobema sintoxia/rubrum 41 37 18 2.756 0.890 0.00370

Table 1.4. Estimates of evolutionary divergence based on analysis of mitochondrial DNA COI sequence pairs between and within species of Fusconaia and Pleurobema using p-distances (lower diagonal) and the Tamura-Nei with invariable sites model of nucleotide substitution (upper diagonal). Bold numbers represent estimates within species. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Fusconaia flava Fusconaia subrotunda Pleurobema cordatum Pleurobema plenum Pleurobema sintoxia/rubrum Fusconaia flava 0.001 0.043 0.053 0.074 0.077 Fusconaia subrotunda 0.041 0.008 0.044 0.069 0.070 Pleurobema cordatum 0.050 0.042 0.004 0.067 0.075 Pleurobema plenum 0.068 0.064 0.062 0.006 0.091 Pleurobema sintoxia/rubrum 0.071 0.064 0.069 0.083 0.004

Table 1.5. Estimates of evolutionary divergence based on analysis of mitochondrial DNA ND1 sequence pairs between and within species of Fusconaia and Pleurobema. Genetic divergence was computed using p-distances (lower diagonal) and the Tamura-Nei nucleotide mutation model with a gamma distribution (upper diagonal). Bold numbers represent estimates within species. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Fusconaia flava Fusconaia subrotunda Pleurobema cordatum Pleurobema plenum Pleurobema sintoxia/rubrum Fusconaia flava 0.003 0.050 0.050 0.089 0.098 Fusconaia subrotunda 0.047 0.009 0.051 0.089 0.108 Pleurobema cordatum 0.046 0.047 0.005 0.087 0.011 Pleurobema plenum 0.080 0.079 0.078 0.004 0.010 Pleurobema sintoxia/rubrum 0.086 0.094 0.098 0.088 0.004

Figure 1.1. Sampling locations for freshwater mussel species in the genera Fusconaia and Pleurobema. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

Figure 1.2. Phylogenetic tree constructed using mitochondrial DNA COI sequences and Bayesian consensus trees in MrBayes. The most appropriate model of nucleotide substitution using the Akaike Information Criterion (AIC) was the General Time Reversible (GTR+G+I) model that followed a gamma distribution with a proportion of invariable sites. The analysis was run with 11 million generations and trees were sampled every 250 generations, which generated a total of 66,002 trees. The final standard deviation of split frequencies was 0.009889 with a –ln likelihood of -3386.51. Posterior probabilities are indicated to the left of the respective nodes. The outgroup was Pleuronaia dolabelloides (MF962140). Species differentiation was assessed using the Automatic Barcode Gap Discovery (ABGD). To assign mussel specimens into the different hypothetical species, the Kimura (1980) two- parameter (K2P) distance model was used, where the minimum intraspecific genetic distance (Pmin) was set at 0.001 and the maximum intraspecific genetic distance (Pmax) was set at 0.1. Two partitions for species delimitation, partition (A), has a prior maximal distance of P = 4.64 e-3 while partition B with a maximal distance P = 1.00 e- 3.Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. Additional sequences for Fusconaia askewi, F. chunii, F. cerina, F. cerina, F. flava, F. lanensis, F. masoni, F. subrotunda, Pleurobema cordatum, P. plenum, P. riddellii, P. rubrum, and P. sintoxia were obtained from Inoue et al. (2018), With their respective GenBank accession numbers available in APPENDIX Table 1.19.

Figure 1.3. Phylogenetic tree constructed using mitochondrial DNA ND1 sequences and Bayesian consensus trees in MrBayes. The most appropriate model of nucleotide substitution using the Akaike Information Criterion (AIC) was the General Time Reversible (GTR+G+I) model that followed a gamma distribution with a proportion of invariable sites. The analysis was run with 11 million generations and trees were sampled every 1000 generations, which generated a total of 16502 trees. The final standard deviation of split frequencies was 0.007619 with a –ln likelihood of -4394.69. Posterior probabilities are indicated next to the respective nodes. The outgroup was Pleuronaia dolabelloides (KT188034). Species differentiation was assessed using the Automatic Barcode Gap Discovery (ABGD). Recurse partition was labeled as ABGD*. In this software, to assign mussel specimens into the different hypothetical species, the Kimura (1980) two-parameter (K2P) distance model was used, where the minimum intraspecific genetic distance (Pmin) was set at 0.001 and the maximum intraspecific genetic distance (Pmax) was set at 0.1. Two partitions for species delimitation, partition (A), has a prior maximal distance of P = 4.64 e-3 and partition B is its recursive partition with the same prior maximal distance. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens were collected from the Clinch River, Hancock County, TN, and Tennessee River downstream of Pickwick dam, Hardin County, TN. Additional reference sequences for Fusconaia askewi, F. lanensis, F. masoni, F. subrotunda, Pleurobema cordatum, P. plenum, and P. sintoxia were obtained from Bertram et al. 2015, Burlakova et al. 2012, Jones et al. 2015, Marshall et al. 2018, and Schilling 2015 with their respective accession numbers available in APPENDIX Table 1.20.

* **

Figure 1.4. Phylogenetic tree constructed using mitochondrial DNA COI + ND1 sequences and Bayesian consensus trees in MrBayes. The most appropriate model of nucleotide substitution using the Akaike Information Criterion (AIC) was the General Time Reversible (GTR+G+I) model that followed a gamma distribution with a proportion of invariable sites. The analysis was run with 6 million generations and trees were sampled every 1000 generations, which generated a total of 9002 trees. The final standard deviation of split frequencies was 0.009964 with a –ln likelihood of -5851.39. Posterior probabilities are indicated to the left of the respective nodes. The outgroup was Pleuronaia dolabelloides (MF962140 + KT188034). Species differentiation was assessed using the Automatic Barcode Gap Discovery (ABGD). To assign mussel specimens into the different hypothetical species, the Kimura (1980) two-parameter (K2P) distance model was used, where the minimum intraspecific genetic distance (Pmin) was set at 0.001 and the maximum intraspecific genetic distance (Pmax) was set at 0.1. Two partitions for species delimitation, partition (A), has a prior maximal distance of P = 1.67 e-3 and partition B is its recursive partition with the same prior maximal distance. From the 4 individuals in the two extra branches 2 were collected from Pool 4 and (**) one of them was consistently identified as P. rubrum by the 5 experts (WG591). (*) The other individual was mostly identified as P. sintoxia (WE779 by 3 out of 4 experts). (*, **) An additional 2 individuals were collected in the Clinch River, TN and were not identified by the experts. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky, and additional mussel specimens were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN.

Species Fusconaia flava Fusconaia subrotunda Pleurobema cordatum 16S Psr07 Pool4 Pleurobema plenum COI Pcor43 and ND1 Pcor12 Pool4 Pleurobema sintoxia/rubrum COI Pcor36 and ND1 Pcor12 Pool4 16SCOI Pcor05 Psr09 and ND1 MCNP Pcor51 Pool4 COI Pcor30 and ND1 Pcor12 Pool4 COI Pcor29 and ND1 Pcor12 Pool4 Posterior probabilities 16SCOI Pcor05 Psr08 and ND1 MCNP Pcor34 Pool4 COI Pcor05 and ND1 Pcor26 Pool4 0.500.50 -–0.900.90 COI Pcor11 and ND1 Pcor12 Pool4 COI Pcor05 and ND1 Pcor18 Pool4 0.900.90 -–0.950.95 16SCOI Pcor05 Psr06 and ND1 Pool4 Pcor15 CL Pool4 * 0.95>1.00 – 1.00 COI Pcor05 and ND1 Pcor12 Pool4 COI Pcor27 and ND1 Pcor41 Pool4 COI16S Pcor17 Psr04 and ND1 Pool4 Pcor25 Pool4 COI Pcor08 and ND1 Pcor11 Pool4 COI Pcor05 and ND1 Pcor08 MCNP COI Pcor05 and ND1 Pcor38 Pool4 0.001 COI16S Pcor41 Psr02 and ND1 Pool4 Pcor13 Pool4 MCNP COI Pcor05 and ND1 Pcor29 Pool4 COI Pcor19 and ND1 Pcor13 Pool4 COI Pcor16 and ND1 Pcor24 Pool4 16S Psr01COI Pcor05 Pool4 and ND1 MCNP Pcor20 Pool4 CL TN COI Pcor12 and ND1 Pcor13 Pool4 COI Pcor05 and ND1 Pcor13 Pool4 16S Psr05 Pool4COI CL Pcor31 and ND1 Pcor49 Pool4 COI Pcor05 **and ND1 Pcor35 Pool4 COI Pcor10 and ND1 Pcor16 Pool4 COI Pcor23 and ND1 Pcor32 Pool4 16S Psr03COI Pcor05 Pool4 and ND1 Pcor10 Pool4 COI Pcor07 and ND1 Pcor10 MCNP COI Pcor33 and ND1 Pcor53 Pool4 16S Psr10 CL COI Pcor05 and ND1 Pcor53 Pool4 COI Pcor05 and ND1 Pcor37 Pool4 COI Pcor20 and ND1 Pcor28 Pool4 COI Pcor03 and ND1 Pcor01 Pool4 COI Pcor03 and16S ND1 Pcor07 Pple MCNP Pool4 COI Pcor03 and ND1 Pcor03 MCNP COI Pcor01 and ND1 Pcor01 MCNP COI Pcor34 and ND1 Pcor54 Pool4 16S Pcor Pool4 COI Pcor06 and ND1 Pcor02 Pool4 COI Pcor06 and ND1 Pcor09 MCNP COI Pcor11 and ND1 Pcor22 Pool4 16S FsubCOI Pool4Pcor14 and ND1 Pcor21 Pool4 COI Pcor05 and ND1 Pcor06 MCNP COI Pcor04 and ND1 Pcor23 Pool4 COI Pcor04 and ND1 Pcor04 MCNP COI Pcor35 and ND116S Pcor02 Ffla Pool4 Pool4 COI Pcor02 and ND1 Pcor02 Pool4 MCNP COI Pcor42 and ND1 Pcor58 Pool4 COI Pcor40 and ND1 Pcor02 Pool4 COI Pcor39 and ND1 Pcor57 Pool4 Figure 1.5. Phylogenetic tree constructed using mitochondrial 16S COIrRNA Pcor37 sequences and ND1 Pcor04 and Bayesian Pool4 consensus trees COI Pcor05 and ND1 Pcor55 Pool4 that were constructed using MrBayes. The most appropriate modelCOI of Pcor05 nucleotide and ND1 Pcor52substitution Pool4 using the Akaike COI Pcor05 and ND1 Pcor04 Pool4 Information Criterion (AIC) was the Hasegawa Kishino-Yano (HKY+I)COI Pcor15 model and ND1with Pcor50 invariable Pool4 sites distribution. The analysis was run with 200,000 generations and trees were sampledCOI Pcor05 every and 100 ND1 generations, Pcor48 Pool4 which generated a COI Pcor05 and ND1 Pcor47 Pool4 total of 2993 trees. The final standard deviation of split frequencies wasCOI Pcor150.007909 and ND1 with Pcor46 a –ln Pool4 likelihood of -889.89. COI Pcor05 and ND1 Pcor45 Pool4 Posterior probabilities are indicated to the left of the respective COI nodes. Pcor05 The and ND1 outgroups Pcor43 Pool4 that were used were COI Pcor20 and ND1 Pcor42 Pool4 Fusconaia flava (16S_Ffla), F. subrotunda (16S_Fsub), PleurobemaCOI Pcor05 cordatum and ND1 Pcor39, (16S_Pco Pool4 r) and P. plenum COI Pcor25 and ND1 Pcor36 Pool4 (16S_Pple). In the tree, we identified the 4 individuals that were definedCOI Pcor24 in the and additionalND1 Pcor33 Pool4 2 clades defined with COI and ND1 markers. (*) One individual was collected from theCOI Pcor18 Clinch and River, ND1 Pcor27 TN Pool4 and had haplotyped COI Pcor15 and ND1 Pcor04 Pool4 16S_PSR06. (**) Two individuals were collected from Pool 4 (WE779COI Pcor13 and andWG591) ND1 Pcor19 as wellPool4 as one individual COI Pcor05 and ND1 Pcor17 Pool4 from the Clinch River shared the 16S_PSR05 haplotype0.005. Mussel specimenCOI Pcor05s and and ND1outgroups Pcor05 MCNP were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river COImile Pcor32 = 149) and ND1 and Pcor14 Mammoth Pool4 Cave National COI Pcor09 and ND1 Pcor14 Pool4 Park (MCNP) (GPS coordinates 37.17819, -86.1154; river mile =COI 197) Pcor22 in and the ND1 Green Pcor31 Pool4 River, Kentucky, and COI Pcor38 and ND1 Pcor56 Pool4 additional outgroups were collected from the Clinch River (CL), HancockCOI Pcor28 County, and ND1 Tennessee Pcor44 Pool4 and the Tennessee COI Pcor26 and ND1 Pcor40 Pool4 River downstream of Pickwick dam, Hardin County, Tennessee (TN)COI Pcor21. and ND1 Pcor30 Pool4 COI Fsub10 and ND1 Fsub02 Pool4 COI Fsub07 and ND1 Fsub10 Pool4 COI Fsub09 and ND1 Fsub08 Pool4 COI Fsub01 and ND1 Fsub08 Pool4 COI Fsub03 and ND1 Fsub04 Pool4 COI Fsub02 and ND1 Fsub03 MCNP 0.005 COI Fsub01 and ND1 Fsub16 Pool4 COI Fsub01 and ND1 Fsub01 MCNP COI Fsub01 and ND1 Fsub18 CL COI Fsub01 and ND1 Fsub12 Pool4 COI Fsub01 and ND1 Fsub07 Pool4 COI Fsub01 and ND1 Fsub02 MCNP COI Fsub04 and ND1 Fsub09 Pool4 COI Fsub05 and ND1 Fsub06 Pool4 COI Fsub04 and ND1 Fsub05 Pool4 COI Fsub11 and ND1 Fsub13 Pool4 COI Fsub08 and ND1 Fsub11 Pool4 COI Fsub13 and ND1 Fsub17 Pool4 COI Fsub11 and ND1 Fsub14 Pool4 COI Fsub12 and ND1 Fsub15 Pool4 COI Ffla01 and ND1 Ffla13 Pool4 COI Ffla01 and ND1 Ffla05 Pool4 MCNP COI Ffla01 and ND1 Ffla11 Pool4 49 COI Ffla06 and ND1 Ffla10 Pool4 COI Ffla07 and ND1 Ffla12 Pool4 COI Ffla04 and ND1 Ffla07 MCNP COI Ffla01 and ND1 Ffla04 Pool4 MCNP COI Ffla03 and ND1 Ffla03 Pool4 MCNP COI Ffla01 and ND1 Ffla03 Pool4 MCNP COI Ffla02 and ND1 Ffla02 Pool4 MCNP COI Ffla05 and ND1 Ffla06 Pool4 COI Ffla01 and ND1 Ffla06 Pool4 MCNP COI Ffla01 and ND1 Ffla09 Pool4 COI Ffla01 and ND1 Ffla08 Pool4 COI Ffla01 and ND1 Ffla01 Pool4 MCNP COI Pple10 and ND1 Pple07 CL COI Pple02 and ND1 Pple07 Pool4 COI Pple02 and ND1 Pple11 Pool4 COI Pple16 and ND1 Pple02 TN COI Pple16 and ND1 Pple20 CL COI Pple05 and ND1 Pple09 Pool4 COI Pple05 and ND1 Pple08 Pool4 COI Pple08 and ND1 Pple12 Pool4 COI Pple04 and ND1 Pple02 Pool4 COI Pple09 and ND1 Pple02 CL COI Pple03 and ND1 Pple05 Pool4 COI Pple17 and ND1 Pple21 TN COI Pple14 and ND1 Pple02 CL COI Pple02 and ND1 Pple10 Pool4 COI Pple06 and ND1 Pple02 Pool4 COI Pple03 and ND1 Pple06 Pool4 COI PPle02 and ND1 Pple04 Pool4 CL COI Pple02 and ND1 Pple02 Pool4 COI PPle01 and ND1 Pple19 CL COI Pple13 and ND1 Pple16 CL COI PPle01 and ND1 Pple15 CL COI PPle01 and ND1 Pple14 CL COI Pple07 and ND1 Pple01 Pool4 COI Pple01 and ND1 Pple03 Pool4 COI Pple01 and ND1 Pple01 Pool4 COI Pple12 and ND1 Pple17 CL COI Pple11 and ND1 Pple18 CL COI Pple15 and ND1 Pple13 CL COI PSR16 and ND1 PSR02 Pool4 COI PSR03 and ND1 PSR16 Pool4 COI PSR03 and ND1 PSR15 Pool4 COI PSR13 and ND1 PSR02 Pool4 COI PSR09 and ND1 PSR02 MCNP COI PSR05 and ND1 PSR02 MCNP COI PSR03 and ND1 PSR03 MCNP COI PSR04 and ND1 PSR02 MCNP COI PSR03 and ND1 PSR02 Pool4 MCNP COI PSR10 and ND1 PSR17 Pool4 COI PSR10 and ND1 PSR11 Pool4 COI PSR10 and ND1 PSR06 Pool4 MCNP COI PSR01 and ND1 PSR20 TN COI PSR01 and ND1 PSR06 TN COI PSR07 and ND1 PSR06 CL COI PSR01 and ND1 PSR07 MCNP COI PSR01 and ND1 PSR05 MCNP COI PSR01 and ND1 PSR18 MCNP COI PSR15 and ND1 PSR01 Pool4 COI PSR01 and ND1 PSR13 Pool4 COI PSR01 and ND1 PSR12 Pool4 COI PSR01 and ND1 PSR09 Pool4 COI PSR11 and ND1 PSR08 Pool4 COI PSR08 and ND1 PSR01 MCNP COI PSR07 and ND1 PSR01 MCNP COI PSR06 and ND1 PSR04 MCNP COI PSR01 and ND1 PSR01 Pool4 MCNP COI PSR19 and ND1 PSR19 CL COI PSR12 and ND1 PSR10 Pool4 COI PSR17 and ND1 PSR21 CL COI PSR14 and ND1 PSR14 Pool4 Pleuronaia dolabelloides MF962140 KT188034

ITS1_PSR07_Pool_4 ITS1_PSR07_Pool_4 ITS1_PSR04_MCNP_Poo4_TN * ITS1_PSR04_MCNP_Poo4_TN * ITS1_PSR02_CL 0.50 - 0.90 ITS1_PSR02_CL 0.90 - 0.95 ITS1_PSR03_MCNP_Pool 4_CL_TN 0.95 – 1.00 ** ITS1_PSR06_Pool 4 ITS1_PSR06_Pool 4 ITS1_PSR01_CL*** ITS1_PSR01_CL *** ITS1_PSR03_MCNP_Pool 4_CL_TN ** ITS1_Pple03_Pool 4 ITS1_Pple02_Pool 4 ITS1_Pple02_Pool 4 ITS1_Pple03_Pool 4 ITS1_Pple01_Pool 4 ITS1_Pple01_Pool 4 ITS1_PSR05_MCNP_Pool 4 ITS1_PSR05_MCNP_Pool 4 ITS1_Pcor02_Pool 4 ITS1_Pcor02_Pool 4 ITS1_Pcor01_Pool 4 ITS1_Pcor01_Pool 4 ITS1_Ffla01_MCNP ITS1_Ffla01_MCNP ITS1_Fsub01_MCNP ITS1_Fsub01_MCNP ITS1_Ffla02_Pool 4 ITS1_Ffla02_Pool 4 ITS1_Fsub02_Pool4 ITS1_Fsub02_Pool4

Figure 1.6. Phylogenetic trees constructed using ITS1 sequences with Clustal (A) and webPRANK (B) alignments and Bayesian consensus trees that were constructed using MrBayes. For both trees, the most appropriate model of nucleotide substitution using the Akaike Information Criterion (AIC) was the Symmetrical model (SYM+G) with gamma rates. The analysis was run with 400,000 generations and trees were sampled every 100 generations, which generated a total of 1785 trees. The final standard deviation of split frequencies was lower than 0.01 for both trees with a -ln likelihood of -982.06 and -1002.20 for the Clustal and webprank alignments, respectively. The outgroups that were collected from the Green River and included F. flava (ITS1_Ffla), F. subrotunda (ITS1_Fsub), P. cordatum, (ITS1_Pcor) and P. plenum (ITS1_Pple). Individuals grouping in the two additional clades in the COI+ND1 tree were identified in the tree: (*) Individual collected from the Clinch River, (**) Individuals collected from the Clinch River and Green River individual consistently identified as P. rubrum by the 5 experts (WG591), (***) Individual collected from the Green River which was identified as P. sintoxia by 3 out of 4 experts (WE779). Mussel specimens and outgroups were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional outgroups were collected from the Clinch River (CL), and Tennessee River (TN).

Figure 1.7. Haplotype network was constructed using mitochondrial DNA COI sequences using TCS (Clement et al. 2002) network as implemented in PopART 1.7 (Leigh and Bryant 2015). The size of each circle corresponds to the number of mussel specimens that share the same haplotype. The mutational steps are shown as hatch marks and the black circles represent inferred haplotypes. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens were collected from the Clinch River, Hancock County, TN, and Tennessee River downstream of Pickwick dam, Hardin County, TN. The outgroup was Pleuronaia dolabelloides (Genbank accession number = MF962140).

Figure 1.8. Haplotype network was constructed using mitochondrial DNA ND1 sequences using TCS (Clement et al. 2002) network as implemented in PopART 1.7 (Leigh and Bryant 2015). The size of each circle corresponds to the number of mussel specimens that share the same haplotype. The mutational steps are shown as hatch marks and the black circles represent inferred haplotypes. (*) Represents ND1_PSR09 haplotype that was amplified for only one individual (WE774) which was not consistently identified by three experts. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National (MCNP) Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens were collected from the Clinch River, Hancock County, TN, and Tennessee River downstream of Pickwick dam, Hardin County, TN. The outgroup was Pleuronaia dolabelloides (Genbank accession number = MF962140).

Figure 1.9. Haplotype network was constructed using mitochondrial DNA 16S rRNA sequences using TCS (Clement et al. 2002) network as implemented in PopART 1.7 (Leigh and Bryant 2015). The size of each circle corresponds to the number of mussel specimens that share the same haplotype. The mutational steps are shown as hatch marks and the black circles represent inferred haplotypes. Mussel specimens of Pleurobema sintoxia/rubrum were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (MCNP) (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. The outgroups were collected from the Green River and included Fusconaia flava (16S_Ffla), F. subrotunda (16S_Fsub), Pleurobema cordatum (16S_Pcor), and P. plenum (16S_Pple).

Figure 1.10. Haplotype network using the nuclear DNA ITS1 sequences constructed from the ClustalX alignment. Haplotype network was constructed using TCS (Clement et al. 2002) network as implemented in PopART 1.7 to show the number of mutational steps primarily among the Pleurobema sintoxia/rubrum sequences (Leigh and Bryant 2015). The size of each circle corresponds to the number of mussel specimens that share the same haplotype. The mutational steps are shown as hatch marks and the black circles represent inferred haplotypes. Mussel specimens of Pleurobema sintoxia/rubrum were collected 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. The outgroups were collected from the Green River and included Fusconaia flava (ITS1_Ffla01, ITS1_Ffla02), F. subrotunda (ITS1_Fsub01, ITS1_Fsub02), Pleurobema cordatum (ITS1_Pcor01, ITS1_Pcor02), and P. plenum (ITS1_Pple01 to ITS1_Pple03).

Figure 1.11. Haplotype network using nuclear DNA ITS1 sequences constructed from the webPRANK alignment. Haplotype network was constructed using TCS (Clement et al. 2002) network as implemented in PopART 1.7 to show the number of mutational steps primarily among the Pleurobema sintoxia/rubrum sequences (Leigh and Bryant 2015). The size of each circle corresponds to the number of mussel specimens that share the same haplotype. The mutational steps are shown as hatch marks and the black circles represent inferred haplotypes. Mussel specimens of Pleurobema sintoxia/rubrum were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. The outgroups were collected from the Green River and included Fusconaia flava (ITS1_Ffla01, ITS1_Ffla02), F. subrotunda (ITS1_Fsub01, ITS1_Fsub02), Pleurobema cordatum (ITS1_Pcor01, ITS1_Pcor02), and P. plenum (ITS1_Pple01 to ITS1_Pple03).

Figure 1.12. Split phylogenetic network using mitochondrial DNA COI gene sequences. Distances were calculated using GTR with a Gamma rates model and a proportion of invariable sites estimated with Splits Tree4 (Huson and Bryant 2006). The algorithm used for distances transformation was NeighborNet. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. The reference sequences were obtained from Inoue et al. 2018 with the associated accession numbers available in APPENDIX Table 1.19. The outgroup is Pleuronaia dolabelloides (GenBank accession number: MF962140)

Figure 1.13. Split phylogenetic network using mitochondrial DNA ND1 gene sequences. Distances were calculated using the GTR with Gamma rates model and a proportion of invariable sites estimated with Splits Tree4 (Huson and Bryant 2006). The algorithm used for distances transformation was NeighborNet. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. Additional reference sequences for Fusconaia askewi, F. lanensis, F. masoni, F. subrotunda, Pleurobema cordatum, P. plenum, and P. sintoxia were obtained from Bertram et al. 2015, Burlakova et al. 2012, Jones et al. 2015, Marshall et al. 2018, and Schilling 2015 with associated accession numbers available in APPENDIX Table 1.20.

Figure 1.14. Split phylogenetic network using mitochondrial DNA 16S rRNA gene sequences using the HKY model with invariable sites estimated with Splits Tree4 (Huson and Bryant 2006). The algorithm used for distance transformation was NeighborNet. Mussel specimens of Pleurobema sintoxia/rubrum were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. The outgroups were Fusconaia flava (16S_Ffla), F. subrotunda (16S_Fsub), Pleurobema cordatum (16S_Pcor), and P. plenum (16S_Pple) and also were collected from the Green River.

Figure 1.15. Split phylogenetic network using nuclear DNA ITS1 sequences constructed from the ClustalX alignment. Distances were calculated using the GTR with equal rates model estimated with Splits Tree4 (Huson and Bryant 2006). The algorithm used for distance transformation was NeighborNet. Mussel specimens of Pleurobema sintoxia/rubrum were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, - 86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. The outgroups were Fusconaia flava (ITS1_Ffla01 and ITS1_Ffla02), F. subrotunda (ITS_Fsub01 and ITS_Fsub02), Pleurobema cordatum (ITS1_Pcor01, ITS1_Pcor02), and P. plenum (ITS1_Pple01, ITS1_Pple02, and ITS1_Pple03) and also were collected from the Green River.

Figure 1.16. Split phylogenetic network using nuclear DNA ITS sequences constructed from the webPRANK alignment. Distances were calculated using the GTR with equal rates model estimated with Splits Tree4 (Huson and Bryant 2006). The algorithm used for distance transformation was NeighborNet. Mussel specimens of Pleurobema sintoxia/rubrum were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, - 86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. The outgroups were Fusconaia flava (ITS1_Ffla01 and ITS1_Ffla02), F. subrotunda (ITS_Fsub01 and ITS_Fsub02), Pleurobema cordatum (ITS1_Pcor01, ITS1_Pcor02), and P. plenum (ITS1_Pple01, ITS1_Pple02, and ITS1_Pple03) and also were collected from the Green River.

APPENDIX – SUPPLEMENTAL TABLES AND FIGURES APPENDIX Table 1.1. GenBank accession numbers for the mitochondrial DNA COI haplotypes for species in the genus Fusconaia and Pleurobema were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (MCNP) (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. Collection site Species Haplotypes GenBank Total Pool 4 MCNP Clinch Tennessee Fusconaia flava COI_Ffla01 MK255210 35 25 10 Fusconaia flava COI_Ffla02 MK255211 2 1 1 Fusconaia flava COI_Ffla03 MK255212 2 1 1 Fusconaia flava COI_Ffla04 MK255213 1 1 Fusconaia flava COI_Ffla05 MK255214 1 1 Fusconaia flava COI_Ffla06 MK255215 1 1 Fusconaia flava COI_Ffla07 MK255216 1 1 Fusconaia subrotunda COI_Fsub01 MK255217 7 5 2 Fusconaia subrotunda COI_Fsub02 MK255218 1 1 Fusconaia subrotunda COI_Fsub03 MK255219 1 1 Fusconaia subrotunda COI_Fsub04 MK255220 2 2 Fusconaia subrotunda COI_Fsub05 MK255221 1 1 Fusconaia subrotunda COI_Fsub06 MK255222 1 1 Fusconaia subrotunda COI_Fsub07 MK255223 1 1 Fusconaia subrotunda COI_Fsub08 MK255224 1 1 Fusconaia subrotunda COI_Fsub09 MK255225 1 1 Fusconaia subrotunda COI_Fsub10 MK255226 1 1 Fusconaia subrotunda COI_Fsub11 MK255227 2 2 Fusconaia subrotunda COI_Fsub12 MK255228 2 2 Fusconaia subrotunda COI_Fsub13 MK255229 1 1 Pleurobema cordatum COI_Pcor01 MK255230 1 1 Pleurobema cordatum COI_Pcor02 MK255231 6 3 3 Pleurobema cordatum COI_Pcor03 MK255232 4 2 2 Pleurobema cordatum COI_Pcor04 MK255233 2 1 1 Pleurobema cordatum COI_Pcor05 MK255234 56 53 3 Pleurobema cordatum COI_Pcor06 MK255235 5 4 1 Pleurobema cordatum COI_Pcor07 MK255236 1 1 Pleurobema cordatum COI_Pcor08 MK255237 1 1 Pleurobema cordatum COI_Pcor09 MK255238 3 3 Pleurobema cordatum COI_Pcor10 MK255239 1 1 Pleurobema cordatum COI_Pcor11 MK255240 2 2 Pleurobema cordatum COI_Pcor12 MK255241 1 1 Pleurobema cordatum COI_Pcor13 MK255242 1 1 Pleurobema cordatum COI_Pcor14 MK255243 1 1 Pleurobema cordatum COI_Pcor15 MK255244 3 3 Pleurobema cordatum COI_Pcor16 MK255245 1 1 Pleurobema cordatum COI_Pcor17 MK255246 1 1 Pleurobema cordatum COI_Pcor18 MK255247 1 1 Pleurobema cordatum COI_Pcor19 MK255248 1 1 61

APPENDIX Table 1.1. Continued. Collection site Species Haplotypes GenBank Total Pool 4 MCNP Clinch Tennessee Pleurobema cordatum COI_Pcor20 MK255249 2 2 Pleurobema cordatum COI_Pcor21 MK255250 1 1 Pleurobema cordatum COI_Pcor22 MK255251 1 1 Pleurobema cordatum COI_Pcor23 MK255252 1 1 Pleurobema cordatum COI_Pcor24 MK255253 1 1 Pleurobema cordatum COI_Pcor25 MK255254 1 1 Pleurobema cordatum COI_Pcor26 MK255255 1 1 Pleurobema cordatum COI_Pcor27 MK255256 1 1 Pleurobema cordatum COI_Pcor28 MK255257 1 1 Pleurobema cordatum COI_Pcor29 MK255258 1 1 Pleurobema cordatum COI_Pcor30 MK255259 1 1 Pleurobema cordatum COI_Pcor31 MK255260 1 1 Pleurobema cordatum COI_Pcor32 MK255261 1 1 Pleurobema cordatum COI_Pcor33 MK255262 1 1 Pleurobema cordatum COI_Pcor34 MK255263 1 1 Pleurobema cordatum COI_Pcor35 MK255264 1 1 Pleurobema cordatum COI_Pcor36 MK255265 1 1 Pleurobema cordatum COI_Pcor37 MK255266 1 1 Pleurobema cordatum COI_Pcor38 MK255267 1 1 Pleurobema cordatum COI_Pcor39 MK255268 1 1 Pleurobema cordatum COI_Pcor40 MK255269 1 1 Pleurobema cordatum COI_Pcor41 MK255270 1 1 Pleurobema cordatum COI_Pcor42 MK255271 1 1 Pleurobema cordatum COI_Pcor43 MK255272 1 1 Pleurobema plenum COI_Pple01 MK255273 19 16 3 Pleurobema plenum COI_Pple02 MK255274 10 9 1 Pleurobema plenum COI_Pple03 MK255275 2 2 Pleurobema plenum COI_Pple04 MK255276 1 1 Pleurobema plenum COI_Pple05 MK255277 2 2 Pleurobema plenum COI_Pple06 MK255278 1 1 Pleurobema plenum COI_Pple07 MK255279 1 1 Pleurobema plenum COI_Pple08 MK255280 1 1 Pleurobema plenum COI_Pple09 MK255281 1 1 Pleurobema plenum COI_Pple10 MK255282 1 1 Pleurobema plenum COI_Pple11 MK255283 1 1 Pleurobema plenum COI_Pple12 MK255284 1 1 Pleurobema plenum COI_Pple13 MK255285 1 1 Pleurobema plenum COI_Pple14 MK255307 1 1 Pleurobema plenum COI_Pple15 MK255308 1 1

APPENDIX Table 1.1. Continued. Collection site Species Haplotypes GenBank Total Pool 4 MCNP Clinch Tennessee Pleurobema plenum COI_Pple16 MK255288 2 1 1 Pleurobema plenum COI_Pple17 MK255289 1 1 Pleurobema sintoxia/rubrum COI_Psr01 MK255290 17 5 9 3 Pleurobema sintoxia/rubrum COI_Psr02 MK255291 1 1 Pleurobema sintoxia/rubrum COI_Psr03 MK255292 7 4 3 Pleurobema sintoxia/rubrum COI_Psr04 MK255293 1 1 Pleurobema sintoxia/rubrum COI_Psr05 MK255294 1 1 Pleurobema sintoxia/rubrum COI_Psr06 MK255295 1 1 Pleurobema sintoxia/rubrum COI_Psr07 MK255296 2 2 1 Pleurobema sintoxia/rubrum COI_Psr08 MK255297 1 1 Pleurobema sintoxia/rubrum COI_Psr09 MK255298 1 1 Pleurobema sintoxia/rubrum COI_Psr10 MK255299 7 6 1 Pleurobema sintoxia/rubrum COI_Psr11 MK255300 2 2 Pleurobema sintoxia/rubrum COI_Psr12 MK255301 1 1 Pleurobema sintoxia/rubrum COI_Psr13 MK255302 1 1 Pleurobema sintoxia/rubrum COI_Psr14 MK255303 1 1 Pleurobema sintoxia/rubrum COI_Psr15 MK255304 1 1 Pleurobema sintoxia/rubrum COI_Psr16 MK255305 1 1 Pleurobema sintoxia/rubrum COI_Psr17 MK255306 1 1 Pleurobema sintoxia/rubrum COI_Psr18 MK255286 1 1 Pleurobema sintoxia/rubrum COI_Psr19 MK255287 1 1

APPENDIX Table 1.2. GenBank accession numbers for the mitochondrial DNA ND1 haplotypes for species belonging to the genus Fusconaia and Pleurobema collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, - 86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. Collection site Species Haplotypes GenBank Total Pool 4 MCNP Clinch Tennessee Fusconaia flava ND1_Ffla01 MK255079 7 5 2 Fusconaia flava ND1_Ffla02 MK255080 2 1 1 Fusconaia flava ND1_Ffla03 MK255081 11 7 4 Fusconaia flava ND1_Ffla04 MK255082 3 1 2 Fusconaia flava ND1_Ffla05 MK255083 7 5 2 Fusconaia flava ND1_Ffla06 MK255084 4 3 1 Fusconaia flava ND1_Ffla07 MK255085 1 1 Fusconaia flava ND1_Ffla08 MK255086 1 1 Fusconaia flava ND1_Ffla09 MK255087 1 1 Fusconaia flava ND1_Ffla10 MK255088 1 1 Fusconaia flava ND1_Ffla11 MK255089 2 2 Fusconaia flava ND1_Ffla12 MK255090 1 1 Fusconaia flava ND1_Ffla13 MK255091 1 1 Fusconaia subrotunda ND1_Fsub01 MK255092 1 1 Fusconaia subrotunda ND1_Fsub02 MK255093 2 1 1 Fusconaia subrotunda ND1_Fsub03 MK255094 1 1 Fusconaia subrotunda ND1_Fsub04 MK255095 1 1 Fusconaia subrotunda ND1_Fsub05 MK255096 1 1 Fusconaia subrotunda ND1_Fsub06 MK255097 1 1 Fusconaia subrotunda ND1_Fsub07 MK255098 1 1 Fusconaia subrotunda ND1_Fsub08 MK255099 2 2 Fusconaia subrotunda ND1_Fsub09 MK255100 1 1 Fusconaia subrotunda ND1_Fsub10 MK255101 1 1 Fusconaia subrotunda ND1_Fsub11 MK255102 1 1 Fusconaia subrotunda ND1_Fsub12 MK255103 1 1 Fusconaia subrotunda ND1_Fsub13 MK255104 1 1 Fusconaia subrotunda ND1_Fsub14 MK255105 1 1 Fusconaia subrotunda ND1_Fsub15 MK255106 2 2 Fusconaia subrotunda ND1_Fsub16 MK255107 1 1 Fusconaia subrotunda ND1_Fsub17 MK255108 1 1 Fusconaia subrotunda ND1_Fsub18 MK255109 1 1 Pleurobema cordatum ND1_Pcor01 MK255110 3 2 1 Pleurobema cordatum ND1_Pcor02 MK255111 12 9 3 Pleurobema cordatum ND1_Pcor03 MK255112 1 1 Pleurobema cordatum ND1_Pcor04 MK255113 4 3 1 Pleurobema cordatum ND1_Pcor05 MK255114 1 1 Pleurobema cordatum ND1_Pcor06 MK255115 1 1 Pleurobema cordatum ND1_Pcor07 MK255116 1 1 Pleurobema cordatum ND1_Pcor08 MK255117 1 1

APPENDIX Table 2.2. Continued. Collection site Species Haplotypes GenBank Total Pool 4 MCNP Clinch Tennessee Pleurobema cordatum ND1_Pcor09 MK255118 1 1 Pleurobema cordatum ND1_Pcor10 MK255119 3 2 1 Pleurobema cordatum ND1_Pcor11 MK255120 1 1 Pleurobema cordatum ND1_Pcor12 MK255121 26 26 Pleurobema cordatum ND1_Pcor13 MK255122 11 11 Pleurobema cordatum ND1_Pcor14 MK255123 4 4 Pleurobema cordatum ND1_Pcor15 MK255124 1 1 Pleurobema cordatum ND1_Pcor16 MK255125 1 1 Pleurobema cordatum ND1_Pcor17 MK255126 1 1 Pleurobema cordatum ND1_Pcor18 MK255127 2 2 Pleurobema cordatum ND1_Pcor19 MK255128 1 1 Pleurobema cordatum ND1_Pcor20 MK255129 1 1 Pleurobema cordatum ND1_Pcor21 MK255130 1 1 Pleurobema cordatum ND1_Pcor22 MK255131 1 1 Pleurobema cordatum ND1_Pcor23 MK255132 1 1 Pleurobema cordatum ND1_Pcor24 MK255133 1 1 Pleurobema cordatum ND1_Pcor25 MK255134 1 1 Pleurobema cordatum ND1_Pcor26 MK255135 1 1 Pleurobema cordatum ND1_Pcor27 MK255136 1 1 Pleurobema cordatum ND1_Pcor28 MK255137 1 1 Pleurobema cordatum ND1_Pcor29 MK255138 1 1 Pleurobema cordatum ND1_Pcor30 MK255139 1 1 Pleurobema cordatum ND1_Pcor31 MK255140 1 1 Pleurobema cordatum ND1_Pcor32 MK255141 1 1 Pleurobema cordatum ND1_Pcor33 MK255142 1 1 Pleurobema cordatum ND1_Pcor34 MK255143 1 1 Pleurobema cordatum ND1_Pcor35 MK255144 1 1 Pleurobema cordatum ND1_Pcor36 MK255145 1 1 Pleurobema cordatum ND1_Pcor37 MK255146 1 1 Pleurobema cordatum ND1_Pcor38 MK255147 1 1 Pleurobema cordatum ND1_Pcor39 MK255148 1 1 Pleurobema cordatum ND1_Pcor40 MK255149 1 1 Pleurobema cordatum ND1_Pcor41 MK255150 1 1 Pleurobema cordatum ND1_Pcor42 MK255151 1 1 Pleurobema cordatum ND1_Pcor43 MK255152 1 1 Pleurobema cordatum ND1_Pcor44 MK255153 1 1 Pleurobema cordatum ND1_Pcor45 MK255154 1 1 Pleurobema cordatum ND1_Pcor46 MK255155 1 1 Pleurobema cordatum ND1_Pcor47 MK255156 1 1

APPENDIX Table 2.2. Continued. Collection site Species Haplotypes GenBank Total Pool 4 MCNP Clinch Tennessee Pleurobema cordatum ND1_Pcor48 MK255157 1 1 Pleurobema cordatum ND1_Pcor49 MK255158 1 1 Pleurobema cordatum ND1_Pcor50 MK255159 1 1 Pleurobema cordatum ND1_Pcor51 MK255160 1 1 Pleurobema cordatum ND1_Pcor52 MK255161 1 1 Pleurobema cordatum ND1_Pcor53 MK255162 2 2 Pleurobema cordatum ND1_Pcor54 MK255163 1 1 Pleurobema cordatum ND1_Pcor55 MK255164 1 1 Pleurobema cordatum ND1_Pcor56 MK255165 1 1 Pleurobema cordatum ND1_Pcor57 MK255166 1 1 Pleurobema cordatum ND1_Pcor58 MK255167 1 1 Pleurobema plenum ND1_Pple01 MK255168 15 15 Pleurobema plenum ND1_Pple02 MK255169 7 4 3 Pleurobema plenum ND1_Pple03 MK255170 1 1 Pleurobema plenum ND1_Pple04 MK255171 2 1 1 Pleurobema plenum ND1_Pple05 MK255172 2 2 Pleurobema plenum ND1_Pple06 MK255173 1 1 Pleurobema plenum ND1_Pple07 MK255174 4 3 1 Pleurobema plenum ND1_Pple08 MK255175 1 1 Pleurobema plenum ND1_Pple09 MK255176 1 1 Pleurobema plenum ND1_Pple10 MK255177 1 1 Pleurobema plenum ND1_Pple11 MK255178 1 1 Pleurobema plenum ND1_Pple12 MK255179 1 1 Pleurobema plenum ND1_Pple13 MK255180 1 1 Pleurobema plenum ND1_Pple14 MK255181 1 1 Pleurobema plenum ND1_Pple15 MK255182 2 2 Pleurobema plenum ND1_Pple16 MK255183 1 1 Pleurobema plenum ND1_Pple17 MK255184 1 1 Pleurobema plenum ND1_Pple18 MK255185 1 1 Pleurobema plenum ND1_Pple19 MK255186 1 1 Pleurobema plenum ND1_Pple20 MK255187 1 1 Pleurobema plenum ND1_Pple21 MK255188 1 1 Pleurobema sintoxia/rubrum ND1_Psr01 MK255189 10 2 8 Pleurobema sintoxia/rubrum ND1_Psr02 MK255190 9 4 5 Pleurobema sintoxia/rubrum ND1_Psr03 MK255191 1 1 Pleurobema sintoxia/rubrum ND1_Psr04 MK255192 1 1 Pleurobema sintoxia/rubrum ND1_Psr05 MK255193 2 2 Pleurobema sintoxia/rubrum ND1_Psr06 MK255194 7 3 1 1 2 Pleurobema sintoxia/rubrum ND1_Psr07 MK255195 1 1

APPENDIX Table 2.2. Continued. Collection site Species Haplotypes GenBank Total Pool 4 MCNP Clinch Tennessee Pleurobema sintoxia/rubrum ND1_Psr08 MK255196 2 2 Pleurobema sintoxia/rubrum ND1_Psr09 MK255197 1 1 Pleurobema sintoxia/rubrum ND1_Psr10 MK255198 1 1 Pleurobema sintoxia/rubrum ND1_Psr11 MK255199 1 1 Pleurobema sintoxia/rubrum ND1_Psr12 MK255200 1 1 Pleurobema sintoxia/rubrum ND1_Psr13 MK255201 2 2 Pleurobema sintoxia/rubrum ND1_Psr14 MK255202 1 1 Pleurobema sintoxia/rubrum ND1_Psr15 MK255203 1 1 Pleurobema sintoxia/rubrum ND1_Psr16 MK255204 1 1 Pleurobema sintoxia/rubrum ND1_Psr17 MK255205 1 1 Pleurobema sintoxia/rubrum ND1_Psr18 MK255206 1 1 Pleurobema sintoxia/rubrum ND1_Psr19 MK255207 1 1 Pleurobema sintoxia/rubrum ND1_Psr20 MK255208 1 1 Pleurobema sintoxia/rubrum ND1_Psr21 MK255209 1 1

APPENDIX Table 1.3. GenBank accession numbers for mitochondrial DNA 16sRNA haplotypes for Pleurobema sintoxia/rubrum and outgroups. Mussel specimens for each species were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (MCNP) (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. Collection site Species Haplotypes GenBank Total Pool 4 MCNP Clinch Tennessee Fusconaia flava 16S_Ffla MK394978 1 1 Fusconaia subrotunda 16S_Fsub MK394976 1 1 Pleurobema cordatum 16S_Pcor MK394979 1 1 Pleurobema plenum 16S_Pple MK394977 1 1 Pleurobema sintoxia/rubrum 16S_Psr01 MK394981 30 10 16 1 3 Pleurobema sintoxia/rubrum 16S_Psr02 MK394982 7 6 1 Pleurobema sintoxia/rubrum 16S_Psr03 MK394983 1 1 Pleurobema sintoxia/rubrum 16S_Psr04 MK394984 1 1 Pleurobema sintoxia/rubrum 16S_Psr05 MK394985 3 2 1 Pleurobema sintoxia/rubrum 16S_Psr06 MK394986 2 1 1 Pleurobema sintoxia/rubrum 16S_Psr07 MK394987 1 1 Pleurobema sintoxia/rubrum 16S_Psr08 MK394988 2 2 Pleurobema sintoxia/rubrum 16S_Psr09 MK394989 1 1 Pleurobema sintoxia/rubrum 16S_Psr10 MK394980 1 1

APPENDIX Table 1.4. GenBank accession numbers for nuclear DNA ITS1 haplotypes for Pleurobema sintoxia/rubrum and outgroups. Mussel specimens for each species were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (MCNP) (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. Collection site Species Haplotype GenBank Total Pool 4 MCNP Clinch Tennessee Pleurobema sintoxia/rubrum ITS_Psr01 MK268755 1 1 Pleurobema sintoxia/rubrum ITS_Psr02 MK268756 1 1 1 Pleurobema sintoxia/rubrum ITS_Psr03 MK268757 22 11 7 2 2 Pleurobema sintoxia/rubrum ITS_Psr04 MK268758 19 7 11 1 Pleurobema sintoxia/rubrum ITS_Psr05 MK268759 2 1 1 Pleurobema sintoxia/rubrum ITS_Psr06 MK268760 1 1 Pleurobema sintoxia/rubrum ITS_Psr07 MK268761 1 1 Fusconaia flava ITS1_Ffla01 MK268762 1 1 Fusconaia flava ITS1_Ffla02 MK268763 1 1 Fusconaia subrotunda ITS1_Fsub01 MK268764 2 2 Fusconaia subrotunda ITS1_Fsub02 MK268765 1 1 Pleurobema cordatum ITS1_Pcor01 MK268766 1 1 Pleurobema cordatum ITS1_Pcor02 MK268767 1 1 Pleurobema plenum ITS1_Pple01 MK268768 1 1 Pleurobema plenum ITS1_Pple02 MK268769 1 1 Pleurobema plenum ITS1_Pple03 MK268770 1 1

APPENDIX Table 1.5. Haplotypes for mitochondrial DNA COI+ND1 concatenated sequences. Mussel specimens for each species were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens of P. rubrum were collected from the Clinch River, Hancock County, TN, and the Tennessee River downstream of Pickwick dam, Hardin County, TN. Collection site Species Haplotype combinations Total Pool 4 MCNP Clinch Tennessee Fusconaia flava COI_Ffla01 and ND1_Ffla01 7 7 2 Fusconaia flava COI_Ffla01 and ND1_Ffla03 9 6 3 Fusconaia flava COI_Ffla01 and ND1_Ffla04 3 1 2 Fusconaia flava COI_Ffla01 and ND1_Ffla05 7 5 2 Fusconaia flava COI_Ffla01 and ND1_Ffla06 3 2 1 Fusconaia flava COI_Ffla01 and ND1_Ffla08 1 1 Fusconaia flava COI_Ffla01 and ND1_Ffla09 1 1 Fusconaia flava COI_Ffla01 and ND1_Ffla11 2 2 Fusconaia flava COI_Ffla01 and ND1_Ffla13 1 1 Fusconaia flava COI_Ffla02 and ND1_Ffla02 2 1 1 Fusconaia flava COI_Ffla03 and ND1_Ffla03 2 1 1 Fusconaia flava COI_Ffla04 and ND1_Ffla07 1 1 Fusconaia flava COI_Ffla05 and ND1_Ffla06 1 1 Fusconaia flava COI_Ffla06 and ND1_Ffla10 1 1 Fusconaia flava COI_Ffla07 and ND1_Ffla12 1 1 Fusconaia subrotunda COI_Fsub01 and ND1_Fsub01 1 1 Fusconaia subrotunda COI_Fsub01 and ND1_Fsub02 1 1 Fusconaia subrotunda COI_Fsub01 and ND1_Fsub07 1 1 Fusconaia subrotunda COI_Fsub01 and ND1_Fsub08 1 1 Fusconaia subrotunda COI_Fsub01 and ND1_Fsub12 1 1 Fusconaia subrotunda COI_Fsub01 and ND1_Fsub16 1 1 Fusconaia subrotunda COI_Fsub01 and ND1_Fsub18 1 1 Fusconaia subrotunda COI_Fsub02 and ND1_Fsub03 1 1 Fusconaia subrotunda COI_Fsub03 and ND1_Fsub04 1 1 Fusconaia subrotunda COI_Fsub04 and ND1_Fsub05 1 1 Fusconaia subrotunda COI_Fsub04 and ND1_Fsub09 1 1 Fusconaia subrotunda COI_Fsub05 and ND1_Fsub06 1 1 Fusconaia subrotunda COI_Fsub07 and ND1_Fsub10 1 1 Fusconaia subrotunda COI_Fsub08 and ND1_Fsub11 1 1 Fusconaia subrotunda COI_Fsub09 and ND1_Fsub08 1 1 Fusconaia subrotunda COI_Fsub10 and ND1_Fsub02 1 1 Fusconaia subrotunda COI_Fsub11 and ND1_Fsub13 1 1 Fusconaia subrotunda COI_Fsub11 and ND1_Fsub14 1 1 Fusconaia subrotunda COI_Fsub12 and ND1_Fsub15 2 2 Fusconaia subrotunda COI_Fsub13 and ND1_Fsub17 1 1 Pleurobema cordatum COI_Pcor01 and ND1_Pcor01 1 1 Pleurobema cordatum COI_Pcor02 and ND1_Pcor02 6 3 3 Pleurobema cordatum COI_Pcor03 and ND1_Pcor01 2 2 Pleurobema cordatum COI_Pcor03 and ND1_Pcor03 1 1

APPENDIX Table 1.5. Continued. Collection site Species Haplotype combinations Total Pool 4 MCNP Clinch Tennessee Pleurobema cordatum COI_Pcor03 and ND1_Pcor07 1 1 Pleurobema cordatum COI_Pcor04 and ND1_Pcor04 1 1 Pleurobema cordatum COI_Pcor04 and ND1_Pcor23 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor04 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor05 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor06 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor08 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor10 2 2 Pleurobema cordatum COI_Pcor05 and ND1_Pcor12 21 21 Pleurobema cordatum COI_Pcor05 and ND1_Pcor13 8 8 Pleurobema cordatum COI_Pcor05 and ND1_Pcor15 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor17 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor18 2 2 Pleurobema cordatum COI_Pcor05 and ND1_Pcor20 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor26 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor29 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor34 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor35 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor37 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor38 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor39 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor43 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor45 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor47 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor48 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor51 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor52 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor53 1 1 Pleurobema cordatum COI_Pcor05 and ND1_Pcor55 1 1 Pleurobema cordatum COI_Pcor06 and ND1_Pcor02 4 4 Pleurobema cordatum COI_Pcor06 and ND1_Pcor09 1 1 Pleurobema cordatum COI_Pcor07 and ND1_Pcor10 1 1 Pleurobema cordatum COI_Pcor08 and ND1_Pcor11 1 1 Pleurobema cordatum COI_Pcor09 and ND1_Pcor14 3 3 Pleurobema cordatum COI_Pcor10 and ND1_Pcor16 1 1 Pleurobema cordatum COI_Pcor11 and ND1_Pcor12 1 1 Pleurobema cordatum COI_Pcor11 and ND1_Pcor22 1 1 Pleurobema cordatum COI_Pcor12 and ND1_Pcor13 1 1 Pleurobema cordatum COI_Pcor13 and ND1_Pcor19 1 1

APPENDIX Table 1.5. Continued. Collection site Species Haplotype combinations Total Pool 4 MCNP Clinch Tennessee Pleurobema cordatum COI_Pcor14 and ND1_Pcor21 1 1 Pleurobema cordatum COI_Pcor15 and ND1_Pcor04 1 1 Pleurobema cordatum COI_Pcor15 and ND1_Pcor46 1 1 Pleurobema cordatum COI_Pcor15 and ND1_Pcor50 1 1 Pleurobema cordatum COI_Pcor16 and ND1_Pcor24 1 1 Pleurobema cordatum COI_Pcor17 and ND1_Pcor25 1 1 Pleurobema cordatum COI_Pcor18 and ND1_Pcor27 1 1 Pleurobema cordatum COI_Pcor19 and ND1_Pcor13 1 1 Pleurobema cordatum COI_Pcor20 and ND1_Pcor28 1 1 Pleurobema cordatum COI_Pcor20 and ND1_Pcor42 1 1 Pleurobema cordatum COI_Pcor21 and ND1_Pcor30 1 1 Pleurobema cordatum COI_Pcor22 and ND1_Pcor31 1 1 Pleurobema cordatum COI_Pcor23 and ND1_Pcor32 1 1 Pleurobema cordatum COI_Pcor24 and ND1_Pcor33 1 1 Pleurobema cordatum COI_Pcor25 and ND1_Pcor36 1 1 Pleurobema cordatum COI_Pcor26 and ND1_Pcor40 1 1 Pleurobema cordatum COI_Pcor27 and ND1_Pcor41 1 1 Pleurobema cordatum COI_Pcor28 and ND1_Pcor44 1 1 Pleurobema cordatum COI_Pcor29 and ND1_Pcor12 1 1 Pleurobema cordatum COI_Pcor30 and ND1_Pcor12 1 1 Pleurobema cordatum COI_Pcor31 and ND1_Pcor49 1 1 Pleurobema cordatum COI_Pcor32 and ND1_Pcor14 1 1 Pleurobema cordatum COI_Pcor33 and ND1_Pcor53 1 1 Pleurobema cordatum COI_Pcor34 and ND1_Pcor54 1 1 Pleurobema cordatum COI_Pcor35 and ND1_Pcor02 1 1 Pleurobema cordatum COI_Pcor36 and ND1_Pcor12 1 1 Pleurobema cordatum COI_Pcor37 and ND1_Pcor04 1 1 Pleurobema cordatum COI_Pcor38 and ND1_Pcor56 1 1 Pleurobema cordatum COI_Pcor39 and ND1_Pcor57 1 1 Pleurobema cordatum COI_Pcor40 and ND1_Pcor02 1 1 Pleurobema cordatum COI_Pcor41 and ND1_Pcor13 1 1 Pleurobema cordatum COI_Pcor42 and ND1_Pcor58 1 1 Pleurobema cordatum COI_Pcor43 and ND1_Pcor12 1 1 Pleurobema plenum COI_Pple01 and ND1_Pple01 14 14 Pleurobema plenum COI_Pple01 and ND1_Pple03 1 1 Pleurobema plenum COI_Pple01 and ND1_Pple14 1 1 Pleurobema plenum COI_Pple01 and ND1_Pple15 1 1 Pleurobema plenum COI_Pple01 and ND1_Pple19 1 1 Pleurobema plenum COI_Pple02 and ND1_Pple02 2 2

APPENDIX Table 1.5. Continued. Collection site Species Haplotype combinations Total Pool 4 MCNP Clinch Tennessee Pleurobema plenum COI_Pple02 and ND1_Pple04 2 1 1 Pleurobema plenum COI_Pple02 and ND1_Pple07 3 3 Pleurobema plenum COI_Pple02 and ND1_Pple10 1 1 Pleurobema plenum COI_Pple02 and ND1_Pple11 1 1 Pleurobema plenum COI_Pple03 and ND1_Pple05 2 2 Pleurobema plenum COI_Pple03 and ND1_Pple06 1 1 Pleurobema plenum COI_Pple04 and ND1_Pple02 1 1 Pleurobema plenum COI_Pple05 and ND1_Pple08 1 1 Pleurobema plenum COI_Pple05 and ND1_Pple09 1 1 Pleurobema plenum COI_Pple06 and ND1_Pple02 1 1 Pleurobema plenum COI_Pple07 and ND1_Pple01 1 1 Pleurobema plenum COI_Pple08 and ND1_Pple12 1 1 Pleurobema plenum COI_Pple09 and ND1_Pple02 1 1 Pleurobema plenum COI_Pple10 and ND1_Pple07 1 1 Pleurobema plenum COI_Pple11 and ND1_Pple18 1 1 Pleurobema plenum COI_Pple12 and ND1_Pple17 1 1 Pleurobema plenum COI_Pple13 and ND1_Pple16 1 1 Pleurobema plenum COI_Pple14 and ND1_Pple02 1 1 Pleurobema plenum COI_Pple15 and ND1_Pple13 1 1 Pleurobema plenum COI_Pple16 and ND1_Pple02 1 1 Pleurobema plenum COI_Pple16 and ND1_Pple20 1 1 Pleurobema plenum COI_Pple17 and ND1_Pple21 1 1 Pleurobema sintoxia/rubrum COI_Psr01 and ND1_Psr01 6 1 5 Pleurobema sintoxia/rubrum COI_Psr01 and ND1_Psr05 2 2 Pleurobema sintoxia/rubrum COI_Psr01 and ND1_Psr06 2 2 Pleurobema sintoxia/rubrum COI_Psr01 and ND1_Psr07 1 1 Pleurobema sintoxia/rubrum COI_Psr01 and ND1_Psr09 1 1 Pleurobema sintoxia/rubrum COI_Psr01 and ND1_Psr12 1 1 Pleurobema sintoxia/rubrum COI_Psr01 and ND1_Psr13 2 2 Pleurobema sintoxia/rubrum COI_Psr01 and ND1_Psr18 1 1 Pleurobema sintoxia/rubrum COI_Psr01 and ND1_Psr20 1 1 Pleurobema sintoxia/rubrum COI_Psr03 and ND1_Psr02 4 2 2 Pleurobema sintoxia/rubrum COI_Psr03 and ND1_Psr03 1 1 Pleurobema sintoxia/rubrum COI_Psr03 and ND1_Psr15 1 1 Pleurobema sintoxia/rubrum COI_Psr03 and ND1_Psr16 1 1

APPENDIX Table 1.5. Continued. Collection site Species Haplotype combinations Total Pool 4 MCNP Clinch Tennessee Pleurobema sintoxia/rubrum COI_Psr04 and ND1_Psr02 1 1 Pleurobema sintoxia/rubrum COI_Psr05 and ND1_Psr02 1 1 Pleurobema sintoxia/rubrum COI_Psr06 and ND1_Psr04 1 1 Pleurobema sintoxia/rubrum COI_Psr07 and ND1_Psr01 2 2 Pleurobema sintoxia/rubrum COI_Psr07 and ND1_Psr06 1 1 Pleurobema sintoxia/rubrum COI_Psr08 and ND1_Psr01 1 1 Pleurobema sintoxia/rubrum COI_Psr09 and ND1_Psr02 1 1 Pleurobema sintoxia/rubrum COI_Psr10 and ND1_Psr06 4 3 1 Pleurobema sintoxia/rubrum COI_Psr10 and ND1_Psr11 1 1 Pleurobema sintoxia/rubrum COI_Psr10 and ND1_Psr17 1 1 Pleurobema sintoxia/rubrum COI_Psr11 and ND1_Psr08 2 2 Pleurobema sintoxia/rubrum COI_Psr12 and ND1_Psr10 1 1 Pleurobema sintoxia/rubrum COI_Psr13 and ND1_Psr02 1 1 Pleurobema sintoxia/rubrum COI_Psr14 and ND1_Psr14 1 1 Pleurobema sintoxia/rubrum COI_Psr15 and ND1_Psr01 1 1 Pleurobema sintoxia/rubrum COI_Psr16 and ND1_Psr02 1 1 Pleurobema sintoxia/rubrum COI_Psr17 and ND1_Prub01 1 1 Pleurobema sintoxia/rubrum COI_Psr19 and ND1_Psr19 1 1

APPENDIX Table 1.6. Variable nucleotide sites observed for Fusconaia flava in a 471 base-pairs (bp) section of the mitochondrial DNA COI gene, where a total of 6 variables sites and 7 haplotypes were observed in 43 mussel specimens collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Variable sites (bp)

124 169 212 253 406

Haplotypes N COI Ffla01 35 A A T T A G COI Ffla02 2 . . . C . . COI Ffla03 2 . . . . G . COI Ffla04 1 . G . . . . COI Ffla05 1 G . . . . . COI Ffla06 1 . . C . . . COI Ffla07 1 . . . . . A

APPENDIX Table 1.7. Variable nucleotide sites observed for Fusconaia subrotunda in a 471 base-pairs (bp) section of the mitochondrial DNA COI gene, where a total of 14 variables sites and 13 haplotypes were observed in 22 mussel specimens collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, - 86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Variable sites (bp)

130 139 214 238 325 469

25 34 40 46 58 94

4 8

Haplotypes COI Fsub01 7 T T A C G A A G T A A G T A COI Fsub02 1 . . G ...... COI Fsub03 1 . C G ...... G . . . . COI Fsub04 2 C . . . . . G A . G . A . G COI Fsub05 1 C . . . . . G A . G . . . G COI Fsub06 1 . . . T . G . . A . . A C G COI Fsub07 1 C . . . . . G A . G . . . . COI Fsub08 1 ...... A . G COI Fsub09 1 ...... G ...... COI Fsub10 1 C . . . . . G A . G . A . . COI Fsub11 2 . . . . T . . A . . . A . G COI Fsub12 2 . . . . . G . A . . . A . G COI Fsub13 1 ...... A . . G A . G

APPENDIX Table 1.8. Variable nucleotide sites observed for Pleurobema cordatum in a 471 base-pairs (bp) section of the mitochondrial DNA COI gene, where a total of 46 variables sites and 43 haplotypes were observed in 117 mussel specimens collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, - 86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

115 Variable149 sites157 (bp)158 171 190 191 193 236 238 239 244 247 251 262 271

15 22 23 25 28 31 34 58 61 64 93 97

Haplotypes N COI Pcor01 1 G G T A G G A A G C T G A A T A C A T A T A T A C T G G COI Pcor02 6 ...... T ...... C . . COI Pcor03 4 ...... T ...... COI Pcor04 2 . . . G . . . . . T ...... COI Pcor05 56 ...... T ...... COI Pcor06 5 ...... T ...... C . . . T . . . COI Pcor07 1 ...... T . . . . . G ...... COI Pcor08 1 ...... T ...... COI Pcor09 3 ...... T ...... T . . . COI Pcor10 1 A A . G A . . . . T ...... COI Pcor11 2 ...... T ...... C ...... COI Pcor12 1 ...... T ...... COI Pcor13 1 ...... G . . T ...... G ...... COI Pcor14 1 . . . . . A . . . T ...... COI Pcor15 3 ...... G . T ...... COI Pcor16 1 . . . G . . . . . T ...... C G . . . . COI Pcor17 1 ...... T ...... COI Pcor18 1 ...... T ...... COI Pcor19 1 ...... T . . . G ...... COI Pcor20 2 ...... T ...... COI Pcor21 1 ...... T ...... G . . T . . . COI Pcor22 1 ...... T ...... T . . . COI Pcor23 1 ...... A T ...... COI Pcor24 1 ...... T ...... A . COI Pcor25 1 ...... T ...... COI Pcor26 1 ...... T . A . . . . T ...... T . . . COI Pcor27 1 ...... T ...... COI Pcor28 1 ...... T ...... T . . . COI Pcor29 1 ...... T ...... COI Pcor30 1 ...... T ...... C

APPENDIX Table 1.8. Extended.

Variable sites (bp)

292 322 328 352 366 367 373 391 393 397 406 409 418 421 427 428 430 469

Haplotypes N COI Pcor01 1 A G A G C A C C T T G G G T A T G G COI Pcor02 6 . A . A ...... COI Pcor03 4 ...... COI Pcor04 2 . A . A ...... COI Pcor05 56 . A . A ...... COI Pcor06 5 . . . A . G ...... COI Pcor07 1 . A . A ...... COI Pcor08 1 G A G A ...... COI Pcor09 3 G A . A ...... T . A . . . . . COI Pcor10 1 . A . A . . . . . C ...... COI Pcor11 2 . A . A ...... COI Pcor12 1 . A . A . . . . C ...... COI Pcor13 1 . A . A ...... COI Pcor14 1 . A G A ...... COI Pcor15 3 . A . A ...... COI Pcor16 1 . A . A ...... COI Pcor17 1 . A . A . . . . . C ...... COI Pcor18 1 . A ...... COI Pcor19 1 . A . A ...... COI Pcor20 2 . A . A ...... A . . . . . COI Pcor21 1 . A . A . . T T ...... COI Pcor22 1 G A . A ...... A . . . . . COI Pcor23 1 . A . A ...... COI Pcor24 1 . A . A ...... COI Pcor25 1 . A . A . . . T . . . . . C . . . . COI Pcor26 1 . A . A ...... COI Pcor27 1 . A . A ...... A COI Pcor28 1 . A . A ...... T . . . . . COI Pcor29 1 . A . A T ...... COI Pcor30 1 . A . A ......

APPENDIX Table 1.8. Continued.

Variable sites (bp)

115 149 157 158 171 190 191 193 236 238 239 244 247 251 262 271

15 22 23 25 28 31 34 58 61 64 93 97

Haplotypes N COI Pcor31 1 ...... T ...... COI Pcor32 1 ...... T ...... G . . T . . . COI Pcor33 1 ...... G . T . . G ...... COI Pcor34 1 ...... T ...... G . . C . . . T . . . COI Pcor35 1 ...... T ...... C . . COI Pcor36 1 ...... T ...... COI Pcor37 1 ...... T ...... T . . . COI Pcor38 1 ...... T . . . . C . . . . G . . . . T . . . COI Pcor39 1 ...... T ...... COI Pcor40 1 . . C ...... T ...... COI Pcor41 1 ...... T ...... COI Pcor42 1 ...... T ...... COI Pcor43 1 ...... T C ......

APPENDIX Table 1.8. Continued and extended.

Variable sites (bp)

292 322 328 352 366 367 373 391 393 397 406 409 418 421 427 428 430 469

Haplotypes N COI Pcor31 1 . A . A ...... T . COI Pcor32 1 G A . A ...... T . A . . . . . COI Pcor33 1 . A . A ...... COI Pcor34 1 . . . A . G ...... COI Pcor35 1 . A G A ...... COI Pcor36 1 . A . A ...... G . . . COI Pcor37 1 . A . A ...... COI Pcor38 1 . A . A ...... T ...... COI Pcor39 1 . A . A ...... C . . COI Pcor40 1 . A . A ...... COI Pcor41 1 . A . A . G ...... COI Pcor42 1 G A . A ...... COI Pcor43 1 . A . A ......

APPENDIX Table 1.9. Variable nucleotide sites observed for Pleurobema plenum in a 471 base-pairs (bp) section of the mitochondrial DNA COI gene, where a total of 11 variables sites and 8 haplotypes were observed in 33 mussel specimens collected in 2015 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) in the Green River, Kentucky. Variable sites (bp)

115 145 190 217 241 295 328 424 433 447

Haplotypes N COI Pple01 16 G G G G C G G A A C C COI Pple02 9 A A . A . . . . . T . COI Pple03 2 A A . A A . . . . T . COI Pple04 1 A A . A . . . . G T . COI Pple05 2 A A . A . . A . . T . COI Pple06 1 A A A A . . . . . T . COI Pple07 1 ...... A COI Pple08 1 A A . A . A . G G T .

APPENDIX Table 1.10. Variable nucleotide sites observed for Pleurobema sintoxia/rubrum in a 471 base-pairs (bp) section of the mitochondrial DNA COI gene, where a total of 21 variables sites and 16 haplotypes were observed in 43 mussel specimens collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Variable sites (bp)

130 157 166 190 193 199 214 236 244 251 322 382 399 448 454 457

10 28 52 58 70

Haplotypes N COI Psr01 14 A G T A G T G A A A T A C G T G A C G C A COI Psr02 1 . . . . . G ...... A . . . COI Psr03 7 ...... G ...... COI Psr04 1 ...... C G ...... COI Psr05 1 . . . . . A . . . . . G ...... COI Psr06 1 . . . G ...... COI Psr07 2 ...... G ...... COI Psr08 1 ...... T . COI Psr09 1 G ...... G ...... COI Psr10 7 ...... G . . . . COI Psr11 2 ...... G ...... COI Psr12 1 . . C ...... T . C A . . A . . COI Psr13 1 ...... G . A ...... COI Psr14 1 . A . . . . A . G . . . T ...... G COI Psr15 1 . . . . C ...... COI Psr16 1 ...... G . . G ......

APPENDIX Table 1.11. Variable nucleotide sites observed for Fusconaia flava in a 744 base-pairs (bp) section of the mitochondrial DNA ND1 gene, where a total of 14 variables sites and 13 haplotypes were observed in 42 mussel specimens collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Variable sites (bp)

129 160 217 222 226 231 239 259 265 293 394 455 721 742

Haplotypes N ND1 Ffla01 7 T T C T C T G C T T T T G T ND1 Ffla02 2 . . . C T . . . . . C . . . ND1 Ffla03 11 ...... C . . . ND1 Ffla04 3 ...... C . C . . . ND1 Ffla05 7 . . A ...... C . A . ND1 Ffla06 4 ...... C ND1 Ffla07 1 ...... C C . . ND1 Ffla08 1 ...... T ...... ND1 Ffla09 1 ...... A ...... ND1 Ffla10 1 ...... C . A . ND1 Ffla11 2 . . . . . C . . . C C . A . ND1 Ffla12 1 C ...... C . . . ND1 Ffla13 1 . C A ...... C . A .

APPENDIX Table 1.12. Variable nucleotide sites observed for Fusconaia subrotunda in a 744 base-pairs (bp) section of the mitochondrial DNA ND1 gene, where a total of 35 variables sites and 17 haplotypes were observed in 20 mussel specimens collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, - 86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Variable sites (bp)

101 121 130 193 206 211 226 244 265 293 367 376 394 397 398 410 433 470 475 485 496 535 544 550 556

73 88 89

Haplotypes N ND1_Fsub01 1 A G T T A G T C T C A T G A A T T G C T G C T A A T A A ND1_Fsub02 2 ...... C . . . C T T ...... ND1_Fsub03 1 ...... C . . . C T T . A ...... ND1_Fsub04 1 . . . C ...... C . . . C T T ...... C . . ND1_Fsub05 1 . A C ...... C . . . C T T C . . . . G . G . ND1_Fsub06 1 . A C C ...... C C G . . C T T . . . . . G . G . ND1_Fsub07 1 ...... C . . . G . C . . . C T T ...... ND1_Fsub08 2 . . . . . A ...... C . . . C T T ...... ND1_Fsub09 1 . A C ...... C . . C C T T . . . . . G . G . ND1_Fsub10 1 ...... C . . . . . C . . . C T T ...... ND1_Fsub11 1 C A C ...... C . . . C T T . . . . . G . G . ND1_Fsub12 1 ...... C . . . C T T ...... G ND1_Fsub13 1 . A C . . A . T . . . . C . G . C T T . . . . . G . G . ND1_Fsub14 1 . . C . G . . . . T . . C . . . C T T . . T . G . . . . ND1_Fsub15 2 G . C . G ...... C . . . C T T . . . C . . . . . ND1_Fsub16 1 ...... C . . . . T ...... ND1_Fsub17 1 . . C . . . . . C T . . C . G . C T T ......

APPENDIX Table 1.12. Extended.

Variable sites (bp)

568 577 598 604 613 615 625 625

Haplotypes ND1_Fsub01 C T C T C T A A ND1_Fsub02 ...... ND1_Fsub03 ...... ND1_Fsub04 . . . . . C . . ND1_Fsub05 . . T . . . G G ND1_Fsub06 ...... G G ND1_Fsub07 . C . C . . . . ND1_Fsub08 T ...... ND1_Fsub09 ...... G G ND1_Fsub10 . C . C . . . . ND1_Fsub11 ...... G G ND1_Fsub12 ...... ND1_Fsub13 ...... G G ND1_Fsub14 . . . . A . . . ND1_Fsub15 ...... G G ND1_Fsub16 ...... ND1_Fsub17 ......

APPENDIX Table 1.13. Variable nucleotide sites observed for Pleurobema cordatum in a 744 base-pairs (bp) section of the mitochondrial DNA ND1 gene, where a total of 83 variables sites and 58 haplotypes were observed in 116 mussel specimens collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, - 86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

101 121 130 137 139 Variable147 sites154 (bp)159 161 164 193 205 211 222 224 226 247 262 265 268

31 42 61 70 71 89

7 9

Haplotypes N ND1_Pcor01 3 C T A T A T A C T A G T T A A T C A T A T T C A A A T T ND1_Pcor02 12 ...... ND1_Pcor03 1 ...... ND1_Pcor04 4 . . G ...... ND1_Pcor05 1 . . G ...... G A ...... ND1_Pcor06 1 . . G ...... T . . . . ND1_Pcor07 1 ...... ND1_Pcor08 1 ...... ND1_Pcor09 1 ...... C . ND1_Pcor10 3 ...... ND1_Pcor11 1 . C ...... ND1_Pcor12 26 ...... ND1_Pcor13 11 ...... ND1_Pcor14 4 ...... ND1_Pcor15 1 ...... ND1_Pcor16 1 ...... A ...... ND1_Pcor17 1 ...... T . . . . . ND1_Pcor18 2 ...... ND1_Pcor19 1 ...... ND1_Pcor20 1 . . G ...... G ...... ND1_Pcor21 1 . . G ...... T . . . . ND1_Pcor22 1 ...... T . . . . ND1_Pcor23 1 . . G ...... ND1_Pcor24 1 . . . . . C ...... C . C C ...... ND1_Pcor25 1 ...... ND1_Pcor26 1 ...... ND1_Pcor27 1 . . G ...... G ...... ND1_Pcor28 1 ...... ND1_Pcor29 1 ...... ND1_Pcor30 1 ......

APPENDIX Table 1.13. Extended.

Variable sites (bp)

275 280 293 295 297 298 308 319 331 337 355 370 373 376 386 388 400 414 415 427 430 431 433 452 470 479 484 490

Haplotypes ND1_Pcor01 T C T A T A G G C G C T G A C A C T T C A G A G G G A T ND1_Pcor02 ...... ND1_Pcor03 ...... A ...... ND1_Pcor04 ...... ND1_Pcor05 ...... ND1_Pcor06 ...... A . . . T ...... ND1_Pcor07 ...... ND1_Pcor08 ...... A T . . . A ...... ND1_Pcor09 ...... ND1_Pcor10 ...... ND1_Pcor11 ...... T . . . A ...... G ...... ND1_Pcor12 ...... A ...... ND1_Pcor13 ...... C ...... ND1_Pcor14 . . . . . G ...... ND1_Pcor15 ...... A ...... ND1_Pcor16 ...... T ...... ND1_Pcor17 . . . . C ...... A A . . ND1_Pcor18 ...... A ...... G . ND1_Pcor19 C ...... A . . . ND1_Pcor20 ...... C ...... ND1_Pcor21 ...... T ...... ND1_Pcor22 ...... T ...... ND1_Pcor23 ...... G ...... ND1_Pcor24 . . . G . . . T ...... ND1_Pcor25 ...... T . . . A ...... ND1_Pcor26 ...... A . . . . C ...... ND1_Pcor27 ...... T ...... ND1_Pcor28 ...... A ...... ND1_Pcor29 ...... C ...... G . . . . . ND1_Pcor30 ...... T ......

APPENDIX Table 1.13. Extended.

Variable sites (bp)

493 496 515 526 541 547 553 574 583 586 597 613 619 625 631 634 638 652 658 667 685 694 698 706 719 721 733

Haplotypes ND1_Pcor01 T G G C C T G A C G C C A A C T T C C A G C G C G G C ND1_Pcor02 ...... T ...... ND1_Pcor03 ...... ND1_Pcor04 ...... T ...... ND1_Pcor05 ...... T ...... ND1_Pcor06 ...... G . . . C . T G ...... ND1_Pcor07 . . . T ...... A ...... ND1_Pcor08 ...... T ...... ND1_Pcor09 ...... T ...... ND1_Pcor10 . A ...... T ...... ND1_Pcor11 ...... T ...... ND1_Pcor12 ...... T ...... T ...... ND1_Pcor13 ...... T ...... ND1_Pcor14 ...... T ...... T ...... A ND1_Pcor15 ...... T ...... T . . . . T . . . ND1_Pcor16 . A ...... C . . T . . . A . . . . ND1_Pcor17 ...... T T ...... ND1_Pcor18 ...... T ...... T ...... ND1_Pcor19 ...... T ...... ND1_Pcor20 ...... T ...... ND1_Pcor21 ...... C . T G ...... ND1_Pcor22 ...... C . T G ...... ND1_Pcor23 ...... T ...... ND1_Pcor24 ...... T ...... ND1_Pcor25 ...... T ...... ND1_Pcor26 ...... T ...... T ...... ND1_Pcor27 ...... T . A ...... ND1_Pcor28 ...... G ...... T ...... ND1_Pcor29 ...... T ...... A . ND1_Pcor30 . . A . . C . . T ...... C . . T ......

APPENDIX Table 1.13 Continued.

Variable sites (bp)

101 121 130 137 139 147 154 159 161 164 193 205 211 222 224 226 247 262 265 268

31 42 61 70 71 89

7 9

Haplotypes N ND1_Pcor31 1 . . G ...... G . . ND1_Pcor32 1 ...... ND1_Pcor33 1 ...... G C ...... ND1_Pcor34 1 ...... C ...... ND1_Pcor35 1 ...... ND1_Pcor36 1 . . G . . . . T ...... ND1_Pcor37 1 ...... ND1_Pcor38 1 ...... C ...... ND1_Pcor39 1 T ...... ND1_Pcor40 1 . . G ...... G ...... T . . . . . ND1_Pcor41 1 . . G ...... G ...... ND1_Pcor42 1 ...... C ND1_Pcor43 1 . . G ...... C ...... ND1_Pcor44 1 ...... ND1_Pcor45 1 ...... ND1_Pcor46 1 ...... ND1_Pcor47 1 ...... G . . . ND1_Pcor48 1 ...... ND1_Pcor49 1 ...... ND1_Pcor50 1 . . . C G ...... ND1_Pcor51 1 ...... ND1_Pcor52 1 . . G ...... ND1_Pcor53 2 ...... ND1_Pcor54 1 ...... C ...... ND1_Pcor55 1 . . G . . . G ...... C ND1_Pcor56 1 ...... C ...... ND1_Pcor57 1 . . G ...... ND1_Pcor58 1 . . G . . . . . C ......

APPENDIX Table 1.13. Continued and extended.

Variable sites (bp)

275 280 293 295 297 298 308 319 331 337 355 370 373 376 386 388 400 414 415 427 430 431 433 452 470 479 484 490

Haplotypes ND1_Pcor31 . . . . . G ...... ND1_Pcor32 ...... C ...... ND1_Pcor33 ...... ND1_Pcor34 ...... A ...... ND1_Pcor35 ...... T ...... ND1_Pcor36 ...... ND1_Pcor37 ...... G ...... ND1_Pcor38 ...... A ...... ND1_Pcor39 ...... ND1_Pcor40 ...... ND1_Pcor41 ...... T . . . A ...... ND1_Pcor42 ...... ND1_Pcor43 . T ...... ND1_Pcor44 C ...... A . . . . ND1_Pcor45 ...... C ND1_Pcor46 ...... A ...... ND1_Pcor47 ...... ND1_Pcor48 ...... C ...... ND1_Pcor49 ...... T ...... A . . . ND1_Pcor50 ...... ND1_Pcor51 ...... A ...... ND1_Pcor52 ...... ND1_Pcor53 ...... ND1_Pcor54 ...... ND1_Pcor55 ...... ND1_Pcor56 . . C ...... C ...... ND1_Pcor57 ...... A ...... ND1_Pcor58 ......

APPENDIX Table 1.13. Continued and extended.

Variable sites (bp)

493 496 515 526 541 547 553 574 583 586 597 613 619 625 631 634 638 652 658 667 685 694 698 706 719 721 733

Haplotypes ND1_Pcor31 ...... T ...... T ...... ND1_Pcor32 . A ...... T ...... ND1_Pcor33 ...... T . . . T ...... ND1_Pcor34 ...... T ...... T ...... ND1_Pcor35 . A ...... C . . T . . . A . . . . ND1_Pcor36 ...... T ...... ND1_Pcor37 ...... G ...... T ...... ND1_Pcor38 ...... T ...... ND1_Pcor39 ...... T ...... ND1_Pcor40 ...... T ...... T . . T . . . . . ND1_Pcor41 ...... T ...... ND1_Pcor42 ...... T ...... ND1_Pcor43 ...... T ...... ND1_Pcor44 . . . . T . . . T . . A ...... T . . . . . A . . ND1_Pcor45 ...... G . . . . T ...... ND1_Pcor46 ...... A ...... T ...... ND1_Pcor47 ...... T ...... ND1_Pcor48 ...... A ...... T ...... ND1_Pcor49 . A ...... C . . T . . . A . . . . ND1_Pcor50 ...... T ...... ND1_Pcor51 . A ...... T ...... T ...... ND1_Pcor52 . A ...... T ...... ND1_Pcor53 ...... G ...... T ...... ND1_Pcor54 ...... T ...... ND1_Pcor55 ...... T ...... ND1_Pcor56 ...... T ...... T ...... ND1_Pcor57 C ...... T ...... ND1_Pcor58 ...... T ......

APPENDIX Table 1.14. Variable nucleotide sites observed for Pleurobema plenum in a 744 base-pairs (bp) section of the mitochondrial DNA ND1 gene, where a total of 17 variables sites and 12 haplotypes were observed in 32 mussel specimens collected in 2015 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) in the Green River, Kentucky. Variable sites (bp)

121 133 193 268 337 370 439 512 520 580 637 664 709

65 67 97

Haplotypes N ND1_Pple01 15 T G A T G C T C A T C C A T C T C ND1_Pple02 4 ...... C . . . . T . T ND1_Pple03 1 ...... C ...... ND1_Pple04 1 . A . . . . . T . C . . . . T . T ND1_Pple05 2 . . G C . . . . . C . . . . T . T ND1_Pple06 1 C ...... C . . . . T . T ND1_Pple07 3 ...... T C T ND1_Pple08 1 . . . . A A . . . C T . G . T . T ND1_Pple09 1 . . . . A . . . . C . . G . T . T ND1_Pple10 1 ...... G C . . . . T . T ND1_Pple11 1 ...... T . T ND1_Pple12 1 ...... C . T . C T . T

APPENDIX Table 1.15. Variable nucleotide sites observed for Pleurobema sintoxia/rubrum in a 744 base-pairs (bp) section of the mitochondrial DNA ND1 gene, where a total of 37 variables sites and 18 haplotypes were observed in 41 mussel specimens collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Variable sites (bp)

127 150 151 160 187 241 242 262 275 337 358 433 439 447 475 505

23 24 27 33 42 49 51 52 55 73 82 89

Haplotypes N ND1_Psr01 10 G C T A T G G C A T G C A T A C A C C G T A G A C T C C ND1_Psr02 9 ...... A ND1_Psr03 1 ...... A ND1_Psr04 1 ...... G ...... ND1_Psr05 2 ...... A . . . . . ND1_Psr06 4 ...... A ...... ND1_Psr07 1 ...... A . . . . . ND1_Psr08 2 . . . . C ...... ND1_Psr09 1 C T G . . . . G . . . A T . T ...... ND1_Psr10 1 ...... C A . . C . T ...... G . . A . ND1_Psr11 1 ...... A ...... ND1_Psr12 1 . . . G . . C ...... ND1_Psr13 2 ...... ND1_Psr14 1 ...... A . . . . T G . . . C . . . T . . . ND1_Psr15 1 ...... A ND1_Psr16 1 . . . . . A . G C ...... G T ...... C . A ND1_Psr17 1 ...... A ...... ND1_Psr18 1 ......

APPENDIX Table 1.15. Extended.

Variable sites (bp)

583 589 624 626 638 664 682 719 743

Haplotypes N ND1_Psr01 10 C C A T T T G G C ND1_Psr02 9 ...... ND1_Psr03 1 ...... A T ND1_Psr04 1 ...... ND1_Psr05 2 ...... ND1_Psr06 4 ...... ND1_Psr07 1 . T ...... ND1_Psr08 2 ...... ND1_Psr09 1 A . . A . . . . . ND1_Psr10 1 ...... ND1_Psr11 1 . . . A . . . . . ND1_Psr12 1 ...... ND1_Psr13 2 . . . . . C . . . ND1_Psr14 1 ...... ND1_Psr15 1 . . G ...... ND1_Psr16 1 ...... ND1_Psr17 1 ...... A . . ND1_Psr18 1 . . . . C . . . .

APPENDIX Table 1.16. Variable nucleotide sites observed for Pleurobema sintoxia/rubrum in a 477 base-pairs (bp) section of mitochondrial DNA 16S rRNA gene, where a total of 10 variables sites and 9 haplotypes were observed in 42 mussel specimens collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, - 86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

Variable sites (bp)

5 39

Haplotypes N 24

118 295 304 311 320 324 346

16S Prs01 26 A C T T C T C G A G 16S Prs02 7 . . . . A . . . . . 16S Prs03 1 ...... T A . . 16S Prs04 1 ...... A 16S Prs05 2 ...... T . . . 16S Prs06 1 . . . C ...... 16S Prs07 1 G A G ...... 16S Prs08 2 . . . . . C . . . . 16S Prs09 1 ...... G .

APPENDIX Table 1.17. Variable nucleotide sites observed for Pleurobema sintoxia/rubrum in a 448 base-pairs (bp) section of nuclear DNA ITS1 sequence. Sequences were aligned using ClustalX. Variable sites that were not encoded by FastGap were in the positions 55, 292, 296, 323, and 339. A total of 5 haplotypes were observed in 45 mussel specimens collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

Base pairs (bp) Coded gaps

54 54

- -

50 51 52 53 54 55

292 296 301 323 339 301

50 52 N ITS_Psr03 22 G A G G C A C A C C ITS_Psr04 19 . . . . . C . . . . ITS_Psr05 2 . . G A A A . . A . T . A A ITS_Psr06 1 A A . . . C - . ITS_Psr07 1 . . . . G C . . . .

APPENDIX Table 1.18. Variable nucleotide sites observed for Pleurobema sintoxia/rubrum in a 448 base-pair (bp) section of nuclear DNA ITS1 sequence. Sequences were aligned using webPRANK. Coded gaps are shown as: 0 = gap absent, 1 = gap present, and “-” = gap unknown. Variable sites that were not encoded by FastGap were in the positions 52, 292, 296, 323, and 339. A total of 5 haplotypes were observed in 45 mussel specimens collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

Base pair (bp) Coded gaps

51 56

- -

50 51 52 54 55 56

292 296 298 323 339 298

50 54

ITS_Psr03 22 G A G G C A C A C C ITS_Psr04 19 . . . . . C . . . . ITS_Psr05 2 . . . A A A . . A . T . A A ITS_Psr06 1 A A . . . C . . ITS_Psr07 1 . . . . G C . . . .

APPENDIX Table 1.19. The mitochondrial DNA COI gene sequences from Inoue et al. (2018) that were added to the phylogenetic tree and split network analyses. The sequences include the species (Fask = Fusconaia askewi, Flan = Fusconaia lanensis, Fcer = Fusconaia cerina, Fmas = Fusconaia masoni, Fsub = Fusconaia subrotunda, Fchu = Fusconaia chunii, Pcor = Pleurobema cordatum, Pple = Pleurobema plenum, Prid = Pleurobema riddellii, Prub = Pleurobema rubrum, and Psin = Pleurobema sintoxia) followed their respective GenBank accession number and their state location. After reference sequences were trimmed to match the length of my Green River sequences, they grouped together. These are reported as similar sequences.

Species Sequences added in the tree Similar sequences

Fusconaia askewi Fask_MF961812_TX Fusconaia askewi Fask_MF961814_TX Flan_MF961935_TX Fusconaia askewi Fask_MF961815_TX Flan_MF961932_TX

Fusconaia askewi Fask_MF961817_TX Fusconaia askewi Fask_MF961818_TX Fask_MF961826_TX, Flan_MF961931_TX Fusconaia askewi Fask_MF961819_TX Fask_MF961821_TX, Fask_MF961831_TX, Fask_MF961838_TX Fusconaia askewi Fask_MF961825_TX Fusconaia askewi Fask_MF961828_TX Fask_MF961830_TX

Fusconaia askewi Fask_MF961832_TX

Fusconaia askewi Fask_MF961836_TX Fusconaia askewi Fask_MF961839_TX Fask_MF961840_TX Fusconaia askewi Fask_MF961813_TX Fask_MF961816_TX, Fask_MF961820_TX, Fusconaia lanensis Fask_MF961823_TX, Fask_MF961824_TX, Fask_MF961827_TX, Fask_MF961829_TX, Fask_MF961834_TX, Fask_MF961835_TX, Fask_MF961837_TX, Flan_MF961933_TX, Flan_MF961934_TX, Flan_MF961936_TX, Flan_MF961937_TX, Flan_MF961938_TX, Fusconaia cerina Fcer_MF961846_LA

Fusconaia cerina Fcer_MF961848_LA

Fusconaia cerina Fcer_MF961849_MS Fusconaia cerina Fcer_MF961850_MS Ffla_MF961896_LA, Ffla_MF961900_AL Fusconaia cerina Fcer_MF961851_AL Fcer_MF961852_AL Fusconaia cerina Fcer_MF961843_LA Ffla_MF961879_AR, Ffla_MF961883_AR, Fusconaia flava Ffla_MF961885_AR, Ffla_MF961886_AR, Ffla_MF961887_AR, Ffla_MF961888_AR, Ffla_MF961889_AR, Ffla_MF961890_AR, Ffla_MF961895_OK, Ffla_MF961907_OH, Ffla_MF961908_OH, Ffla_MF961918_AR, Ffla_MF961919_AR, Ffla_MF961923_KY_IN, Ffla_MF961925_AR, Ffla_MF961926_AR Fusconaia cerina Fcer_MF961844_LA Fcer_MF961845_LA, Fcer_MF961847_LA, Fusconaia flava Ffla_MF961913_AR Fusconaia flava Ffla_MF961867_AR

Fusconaia flava Ffla_MF961869_AR

Fusconaia flava Ffla_MF961870_AR

APPENDIX Table 1.19. Continued.

Species Sequences added in the tree Similar sequences Fusconaia flava Ffla_MF961871_AR Ffla_MF961872_AR, Ffla_MF961874_AR

Fusconaia flava Ffla_MF961873_AR

Fusconaia flava Ffla_MF961875_AR Fusconaia flava Ffla_MF961876_KY Ffla_MF961877_KY, Ffla_MF961878_KY, Ffla_MF961917_AR Fusconaia flava Ffla_MF961880_MN Fusconaia flava Ffla_MF961881_AR Ffla_MF961882_AR, Ffla_MF961884_AR, Ffla_MF961927_AR Fusconaia flava Ffla_MF961891_TX Ffla_MF961893_TX, Ffla_MF961904_AR

Fusconaia flava Ffla_MF961892_TX

Fusconaia flava Ffla_MF961894_AR

Fusconaia flava Ffla_MF961897_LA

Fusconaia flava Ffla_MF961898_AR Fusconaia flava Ffla_MF961899_LA Ffla_MF961910_AR Fusconaia flava Ffla_MF961901_LA Ffla_MF961902_LA Fusconaia flava Ffla_MF961903_AR Ffla_MF961920_AR

Fusconaia flava Ffla_MF961905_AR

Fusconaia flava Ffla_MF961906_AR

Fusconaia flava Ffla_MF961911_AR

Fusconaia flava Ffla_MF961912_AR

Fusconaia flava Ffla_MF961914_AR

Fusconaia flava Ffla_MF961915_AR

Fusconaia flava Ffla_MF961916_AR

Fusconaia flava Ffla_MF961922_AR

Fusconaia flava Ffla_MF961924_LA

Fusconaia flava Ffla_MF961928_AR

Fusconaia flava Ffla_MF961929_AR

Fusconaia flava Ffla_MF961930_AR

Fusconaia masoni Fmas_MF961939_VA

Fusconaia masoni Fmas_MF961940_VA Fusconaia masoni Fmas_MF961941_NC Fmas_MF961942_NC

Fusconaia subrotunda Fsub_MF961948_KY Fusconaia subrotunda Fsub_MF961949_KY Fsub_MF961950_KY

Fusconaia subrotunda Fsub_MF961951_NC

Fusconaia subrotunda Fsub_MF961952_NC

Fusconaia subrotunda Fsub_MF961953_NC

Fusconaia subrotunda Fsub_MF961954_NC

APPENDIX Table 1.19. Continued.

Species Sequences added in the tree Similar sequences Fusconia chunii Fchu_MF961853_TX Fchu_MF961854_TX, Fchu_MF961855_TX, Fchu_MF961856_TX Pleurobema cordatum Pcor_MF961959_KY Pleurobema cordatum Pcor_MF961960_KY Pcor_MF961963_KY Pleurobema cordatum Pcor_MF961961_KY Pcor_MF961965_OH, Pcor_MF961968_OH

Pleurobema cordatum Pcor_MF961962_KY

Pleurobema cordatum Pcor_MF961964_OH

Pleurobema cordatum Pcor_MF961966_AR

Pleurobema cordatum Pcor_MF961967_OH

Pleurobema cordatum Pcor_MF961969_AL

Pleurobema dolabelloides Pdol_MF962140 Pleurobema plenum Pple_MF961970_TN Pple_MF961972_TN

Pleurobema plenum Pple_MF961971_TN

Pleurobema plenum Pple_MF961973_TN Pleurobema riddellii Prid_MF961974_AR Prid_MF961976_AR, Prid_MF961978_AR, Prid_MF961979_AR, Prid_MF961981_AR, Prid_MF961982_AR, Prid_MF961983_AR, Prid_MF961984_AR, Prid_MF961986_AR, Prid_MF961988_AR Pleurobema riddellii Prid_MF961975_AR Prid_MF961977_AR

Pleurobema riddellii Prid_MF961980_AR

Pleurobema riddellii Prid_MF961985_AR

Pleurobema riddellii Prid_MF961987_AR

Pleurobema riddellii Prid_MF961989_AR Pleurobema riddellii Prid_MF961990_AR Prid_MF961991_AR

Pleurobema riddellii Prid_MF961993_AR Pleurobema riddellii Prid_MF961994_TX Prid_MF961995_TX, Prid_MF961996_TX, Prid_MF961997_TX, Prid_MF962002_TX, Prid_MF962003_TX, Prid_MF962004_TX Pleurobema riddellii Prid_MF961998_TX Pleurobema riddellii Prid_MF961999_TX Prid_MF962000_TX, Prid_MF962001_TX

Pleurobema rubrum Prub_MF962005_TN

Pleurobema rubrum Prub_MF962006_TN

Pleurobema rubrum Prub_MF962007_KY

Pleurobema rubrum Prub_MF962010_AR

Pleurobema rubrum Prub_MF962012_AR

Pleurobema rubrum Prub_MF962013_AR

Pleurobema rubrum Prub_MF962014_AR

Pleurobema rubrum Prub_MF962015_AR

Pleurobema rubrum Prub_MF962018_AR

100

APPENDIX Table 1.19. Continued.

Species Sequences added in the tree Similar sequences

Pleurobema rubrum Prub_MF962021_AR

Pleurobema rubrum Prub_MF962024_AR

Pleurobema rubrum Prub_MF962026_AR

Pleurobema rubrum Prub_MF962029_AR

Pleurobema rubrum Prub_MF962030_AR

Pleurobema rubrum Prub_MF962033_TN Pleurobema rubrum Prub_MF962008_AR Prub_MF962016_AR, Prub_MF962020_AR, Pleurobema sintoxia Prub_MF962025_AR, Prub_MF962027_AR, Psin_MF962035_AR, Psin_MF962064_AR, Psin_MF962065_AR, Psin_MF962080_AR, Psin_MF962081_AR, Psin_MF962087_AR Pleurobema rubrum Prub_MF962009_AR Prub_MF962011_AR, Prub_MF962017_AR, Pleurobema sintoxia Prub_MF962022_AR, Prub_MF962028_AR, Psin_MF962038_TN, Psin_MF962039_TN, Psin_MF962043_MN, Psin_MF962050_KY, Psin_MF962053_KY, Psin_MF962054_PA, Psin_MF962055_PA, Psin_MF962069_AR, Psin_MF962076_KY, Psin_MF962079_AR, Psin_MF962084_AR, Psin_MF962085_AR, Psin_MF962092_AR, Psin_MF962097_AR Pleurobema rubrum Prub_MF962019_AR Psin_MF962090_AR, Psin_MF962091_AR Pleurobema sintoxia Pleurobema rubrum Prub_MF962031_TN Psin_MF962045_KY Pleurobema sintoxia Pleurobema rubrum Prub_MF962032_TN Psin_MF962056_PA, Psin_MF962061_PA, Pleurobema sintoxia Psin_MF962073_TN, Psin_MF962074_TN, Psin_MF962075_TN Pleurobema rubrum Prub_MF962034_AR Psin_MF962094_AR, Psin_MF962095_AR, Pleurobema sintoxia Psin_MF962096_AR, Psin_MF962098_AR, Psin_MF962099_AR, Psin_MF962101_AR, Psin_MF962102_AR, Psin_MF962103_AR Pleurobema sintoxia Psin_MF962036_AR Pleurobema sintoxia Psin_MF962037_TN Psin_MF962072_TN

Pleurobema sintoxia Psin_MF962040_PA

Pleurobema sintoxia Psin_MF962041_WI

Pleurobema sintoxia Psin_MF962042_WI Pleurobema sintoxia Psin_MF962046_KY Psin_MF962060_PA Pleurobema sintoxia Psin_MF962047_KY Psin_MF962049_KY

Pleurobema sintoxia Psin_MF962048_KY Pleurobema sintoxia Psin_MF962051_KY Psin_MF962052_KY Pleurobema sintoxia Psin_MF962057_PA Psin_MF962058_PA, Psin_MF962059_PA Pleurobema sintoxia Psin_MF962062_AR Psin_MF962104_AR

Pleurobema sintoxia Psin_MF962063_AR

Pleurobema sintoxia Psin_MF962066_AR

101

APPENDIX Table 1.19. Continued.

Species Sequences added in the tree Similar sequences

Pleurobema sintoxia Psin_MF962067_AR

Pleurobema sintoxia Psin_MF962068_AR

Pleurobema sintoxia Psin_MF962070_AR

Pleurobema sintoxia Psin_MF962071_AR Pleurobema sintoxia Psin_MF962077_KY Psin_MF962078_KY

Pleurobema sintoxia Psin_MF962082_AR

Pleurobema sintoxia Psin_MF962083_AR

Pleurobema sintoxia Psin_MF962086_AR

Pleurobema sintoxia Psin_MF962088_AR

Pleurobema sintoxia Psin_MF962100_AR

102

APPENDIX Table 1.20. Sequences from other studies utilizing the mitochondrial DNA ND1 gene that were added to the phylogenetic tree and split network analyses. The sequences include the species (Fask = Fusconaia askewi, Flan = Fusconaia lanensis, Fmas = Fusconaia masoni, Fsub = Fusconaia subrotunda, Pcor = Pleurobema cordatum, Pple = Pleurobema plenum, Prub = Pleurobema rubrum, Psin = Pleurobema sintoxia, Pdol = Pleurobema dolabelloides) followed their respective GenBank accession number or haplotype name for Jones et al. 2015 sequences, and their sampling location.

Species Name in the tree Waterbody Reference Fusconaia askewi Fask_JN180976_TX Sabine drainage, Texas Burlakova et al. 2012 Fusconaia askewi Fask_JN180977_TX Sabine drainage, Texas Burlakova et al. 2012 Fusconaia askewi Fask_KY442832_TX Texas Bertram et al. 2015 Fusconaia askewi Fask_KY442833_TX Texas Bertram et al. 2015 Fusconaia askewi Fask_MG020448_TX Texas Marshall el al. 2018 Fusconaia lanensis Flan_JN180980_TX Texas Burlakova et al. 2012 Fusconaia lanensis Flan_JN180982_TX Texas Burlakova et al. 2012 Fusconaia masoni Fmas_KT187973_VA Craig Creek, Virginia Schilling 2015 Fusconaia masoni Fmas_KT187974_VA Craig Creek, Virginia Schilling 2015 Pleurobema rubrum Prub_KT188097_DR Duck River, Tenessee Schilling 2015 Pleurobema rubrum Prub_KT188096_DR Duck River, Tenessee Schilling 2015 Pleurobema rubrum Prub_KT188095_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187998_ND Nolichucky Drainage, Tenessee Schilling 2015 Fusconaia subrotunda Fsub_KT187997_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187996_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187995_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187994_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187993_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187992_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187991_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187990_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187989_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187988_PO Powell Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187987_CL Powell Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187986_PO Powell Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187985_PO Powell Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187984_PO Powell Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187983_PO Powell Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187982_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187981_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187980_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187979_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187978_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187977_CL Clinch Drainage, Tennessee Schilling 2015 Fusconaia subrotunda Fsub_KT187976_CL Clinch Drainage, Tennessee Schilling 2015

103

APPENDIX Table 1.20. Continued. Species Name in the tree Waterbody Reference Fusconaia subrotunda Fsub_KT187975_CL Clinch Drainage Schilling 2015 Pleurobema cordatum Pcor_PcGreen01_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcGreen02_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcGreen03_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcGreen04_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcGreen05_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcGreen06_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcGreen07_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcGreen08_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcGreen09_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcGreen10_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcGreen11_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcGreen12_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcGreen13_GR Green River, Kentucky Jones et al. 2015 Pleurobema cordatum Pcor_PcTenn14_TN Tennessee River Basin Jones et al. 2015 Pleurobema cordatum Pcor_PcTenn15_TN Tennessee River Basin Jones et al. 2015 Pleurobema plenum Pple_PpClinch01_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema plenum Pple_PpClinch02_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema plenum Pple_PpClinch03_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema plenum Pple_PpClinch04_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema plenum Pple_PpClinch05_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema plenum Pple_PpClinch06_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema plenum Pple_PpClinch07_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema plenum Pple_PpClinch08_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema plenum Pple_PpClinch09_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema plenum Pple_PpClinch10_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema plenum Pple_PpClinch11_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema plenum Pple_PpGreen12_GR Green River, Kentucky Jones et al. 2015 Pleurobema plenum Pple_PpGreen13_GR Green River, Kentucky Jones et al. 2015 Pleurobema plenum Pple_PpGreen14_GR Green River, Kentucky Jones et al. 2015 Pleurobema plenum Pple_PpGreen15_GR Green River, Kentucky Jones et al. 2015 Pleurobema plenum Pple_PpGreen16_GR Green River, Kentucky Jones et al. 2015 Pleurobema rubrum Prub_PrClinch01_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema rubrum Prub_PrClinch02_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema rubrum Prub_PrClinch03_CL Clinch River, Tennessee Jones et al. 2015 Pleurobema sintoxia Psin_PsGreen01_GR Green River, Kentucky Jones et al. 2015 Pleurobema sintoxia Psin_PsGreen02_GR Green River, Kentucky Jones et al. 2015 Pleurobema dolabelloides Pdol_KT188034 Schilling 2015

104

APPENDIX Table 1.21. Evanno method using Delta K and Mean LnP(K) to support the number of clusters in P. sintoxia/rubrum The clusters were inferred using microsatellites in the software STRUCTURE. Individuals were collected in in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

K Reps Mean LnP(K) Stdev LnP(K) Ln'(K) Delta K

1 10 -1823.12 0.66 - -

2 10 -1823.58 0.81 -0.46 2.62

3 10 -1826.17 3.37 -2.59

105

APPENDIX Figure 1.1. Figure continues in the following page.

106

APPENDIX Figure 1.1. Continued

107

APPENDIX Figure 1.1. Phylogenetic tree was constructed using mitochondrial DNA COI gene sequences and Bayesian consensus trees in MrBayes. The most appropriate model of nucleotide substitution selected using the Akaike Information Criterion (AIC) was the General Time Reversible (GTR+G+I) model that followed a gamma distribution with a proportion of invariable sites. The analysis was run with 11 million generations and trees were sampled every 250 generations, which generated a total of 66,002 trees. The final standard deviation of split frequencies was 0.009889 with a –ln likelihood of -3386.51. Posterior probabilities are indicated to the left of the respective nodes. The outgroup taxa was Pleuronaia dolabelloides (MF962140). Species differentiation was assessed using the Automatic Barcode Gap Discovery (ABGD). To assign mussel specimens into the different hypothetical species, the Kimura (1980) two-parameter (K2P) distance model was used, where the minimum intraspecific genetic distance (Pmin) was set at 0.001 and the maximum intraspecific genetic distance (Pmax) was set at 0.1. Two partitions for species delimitation, partition (A), has a prior maximal distance of P = 4.64 e-3 while partition B with a maximal distance P = 1.00 e-3. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (MCNP) (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens were collected from the Clinch River (CL) in Hancock County, TN, and from the Tennessee River (TN) downstream of Pickwick dam, Hardin County. Additional sequences for Fusconaia askewi (Fask), Fusconaia chunii (Fchu), Fusconaia cerina (Fcer), Fusconaia flava (Ffla), Fusconaia lanensis (Flan), Fusconaia masoni (Fmas), Fusconaia subrotunda (Fsub), Pleurobema cordatum (Pcor), Pleurobema plenum (Pple), Pleurobema riddellii (Prid), Pleurobema rubrum (Prub), and Pleurobema sintoxia (Psin) were obtained from Inoue et al. (2018) with the accession numbers available in APPENDIX Table 1.19.

108

APPENDIX Figure 1.2. Figure continues in the following page.

109

APPENDIX Figure 1.2. Continued.

110

APPENDIX Figure 1.2. Phylogenetic tree was constructed using mitochondrial DNA ND1 gene sequences and Bayesian consensus trees in MrBayes. The most appropriate model of nucleotide substitution selected using the Akaike Information Criterion (AIC) was the General Time Reversible (GTR+G+I) model that followed a gamma distribution with a proportion of invariable sites. The analysis was run with 11 million generations and trees were sampled every 1000 generations, which generated a total of 16502 trees. The final standard deviation of split frequencies was 0.007619 with a –ln likelihood of -4394.69. Posterior probabilities are indicated next to the respective nodes. The outgroup was Pleuronaia dolabelloides (KT188034). Species differentiation was assessed using the Automatic Barcode Gap Discovery (ABGD). Recurse partition was labeled as ABGD*. In this software, to assign mussel specimens into the different hypothetical species, the Kimura (1980) two-parameter (K2P) distance model was used, where the minimum intraspecific genetic distance (Pmin) was set at 0.001 and the maximum intraspecific genetic distance (Pmax) was set at 0.1. Two partitions for species delimitation, partition (A), has a prior maximal distance of P = 4.64 e-3 and partition B is its recursive partition with the same prior maximal distance. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional mussel specimens were collected from the Clinch River (CL) in Hancock County, TN, and from the Tennessee River (TN) downstream of Pickwick dam, Hardin County. Additional reference sequences for Fusconaia askewi (Fask), Fusconaia lanensis (Flan), Fusconaia masoni (Fmas), Fusconaia subrotunda (Fsub), Pleurobema cordatum (Pcor), Pleurobema plenum (Pple), Pleurobema rubrum (Prub), and Pleurobema sintoxia (Psin) were obtained from Bertram et al. 2015, Burlakova et al. 2012, Jones et al. 2015, Marshall et al. 2018, and Schilling 2015 with the accession numbers available in APPENDIX Table 1.20.

111

APPENDIX Figure 1.3. Figure continues in the following page

112

APPENDIX Figure 1.3. Continued.

113

APPENDIX Figure 1.3. Phylogenetic tree constructed using mitochondrial COI + ND1 genes sequences and Bayesian consensus trees in MrBayes. The most appropriate model of nucleotide substitution selected using the Akaike Information Criterion (AIC) was the General Time Reversible (GTR+G+I) model that followed a gamma distribution with a proportion of invariable sites. The analysis was run with 6 million generations and trees were sampled every 1000 generations, which generated a total of 9002 trees. The final standard deviation of split frequencies was 0.009964 with a –ln likelihood of -5851.39. Posterior probabilities are indicated to the left of the respective nodes. The outgroup was Pleuronaia dolabelloides (MF962140 + KT188034). Species differentiation was assessed using the Automatic Barcode Gap Discovery (ABGD). To assign mussel specimens into the different hypothetical species, we used the Kimura (1980) two-parameter (K2P) distance model, where the minimum intraspecific genetic distance (Pmin) was set as 0.001 and the maximum intraspecific genetic distance (Pmax) was set as 0.1. Two partitions for species delimitation, partition (A), has a prior maximal distance of P = 1.67 e-3 and partition B is its recursive partition with the same prior maximal distance. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, - 86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky, and additional mussel specimens were collected from the Clinch River (CL), and Tennessee River (TN).

114

Chapter 2 - Assessment of genetic diversity and effective population sizes (Ne) of Fusconaia and Pleurobema species in the Green River, Kentucky

ABSTRACT

Reduction in genetic diversity is often correlated with a decline in effective population size

(Ne), a process that frequently occurs in populations of imperiled species. Using mitochondrial

DNA (mtDNA) sequences and nuclear DNA (microsatellites) markers, I estimated genetic diversity, current Ne using a linkage disequilibrium method, and long-term Ne using an approximate Bayesian computation method for Pleurobema cordatum, P. plenum, and P. sintoxia/rubrum populations in the Green River, Kentucky. In addition, I estimated the long-term female effective population size Nef using mtDNA sequences for Fusconaia flava, F. subrotunda,

P. cordatum, P. plenum, and P. sintoxia/rubrum. Because this was the first study characterizing variation at microsatellite loci in P. plenum, ten polymorphic loci were developed specifically for this species. In addition, microsatellite markers developed for P. clava and P. pyriforme were used to assess genetic variation and estimate Ne for my study species. For the three species of

Pleurobema, all loci were polymorphic, and the number of alleles ranged between 7 and 42 alleles per locus. Observed heterozygosities ranged between 0.465 and 0.969 for the respective

Pleurobema species. The largest contemporary Ne was estimated for P. plenum at 303 individuals

(95% CI =68.5 – infinite) while the lowest was for P. sintoxia/rubrum at 132 individuals (95% CI

=39.5 – infinite). The largest long-term Ne was estimated for P. cordatum at 19,000 individuals

(95% CI = 8,570 – 34,300), while the lowest was for P. plenum at 5,850 individuals (95% CI =

3,100 – 8,950). Finally, the largest long-term Nef was also for P. cordatum at 420,000 individuals,

(95% CI 325,000 – 517,000), while the lowest was for P. plenum at 9,370 individuals (95% CI

8,290 – 9,910). My results suggest that study populations in the Green River, KY are

115 demographically healthy, as they show high genetic diversity and large effective population numbers. Future studies should take into consideration life-history traits, habitat availability, and recruitment of these populations to further assess the conservation status needs of these species in the Green River.

KEYWORDS: Fusconaia, Pleurobema, freshwater mussels, cytochrome oxidase subunit I (COI),

NADH dehydrogenase 1 (ND1), effective population size, microsatellites.

INTRODUCTION

The effective population size (Ne) is the size of an “ideal” population in which the change in heterozygosity or allele frequencies over time is the same as those in the observed population

(Luikart et al. 2010). The magnitude of Ne affects population viability, extinction risks, and the framing of conservation and management decisions (Luikart et al. 2010). In endangered species with small population sizes, reduction of genetic diversity is directly related to a reduction in Ne.

The magnitude of Ne is a function of various factors, such as census population size (Ne differs from census population size as a real population departs from an ideal Wright-Fisher population) and demographic parameters (i.e., sex ratio, reproductive success, mode of inheritance, and mating system). Incorrect assumptions regarding these demographic parameters can result in imprecise estimation of Ne (Caballero 1994; Wang and Caballero 1999; Wang 2005). In recent years, however, there have been advances in molecular marker-based methods for estimation of Ne, due to both higher availability of molecular markers and development of better estimation algorithms

(Luikart et al. 2010). Some of these algorithms are based on using observed linkage disequilibrium or excess heterozygosity to estimate the size of the population that would have given rise to those

116 departures from Hardy-Weinberg estimated genotype frequencies (Waples and Do 2010). Use of coalescent methods also may prove useful because the approach reduces the effect of missing genetic diversity (Charlesworth et al. 1995; Kuhner 2006).

Assessment of Ne is important for predicting how maintenance of genetic variation affects fitness and likelihood of survival of imperiled species. Among the freshwater mussel species in this study, Pleurobema plenum is of particular interest as it has been classified as Endangered under the U.S. Endangered Species Act (ESA). Conservation status and distribution for each study species under the U.S. ESA and the International Union for Conservation of Nature (IUNC) are listed in Table 2.1. The conservation status for P. sintoxia and P. rubrum has been assessed separately for these two mussels, as they are considered different species by the U.S. Fish and

Wildlife Service and the International Union for the Conservation of Nature (IUCN). However, recent studies using mitochondrial NADH dehydrogenase 1 (ND1) (Jones et al. 2015) and cytochrome oxidase subunit I (COI) (Inoue et al. 2018) DNA sequence markers reported that P. sintoxia and P. rubrum are genetically very similar (<1% divergence), suggesting that they are the same species and hence their conservation status needs to be reassessed. Finally, because the estimation of Ne is useful for projecting the likely survival and viability of species with small populations (Wang 2005) and helps to support management planning, I estimated both current and long-term Ne for the Green River populations of these species using both DNA microsatellite and mtDNA data.

The purpose of this study was to estimate the genetic diversity and Ne of genetically identified mussel specimens of F. flava, F. subrotunda, P. cordatum, P. plenum, and P. rubrum/sintoxia in the Green River, Kentucky. Estimation of genetic diversity and Ne is important to support assessment of conservation status of focal populations for species of interest. Moreover,

117 the estimation of long-term Ne is important to support reconstruction of the demographic history of these species. Analyses of Ne will allow for design of management actions to ensure preservation of genetic diversity within populations of the species of interest. In addition, the development of microsatellite DNA markers for P. plenum will be useful for genetic monitoring for this species and others in the genus Pleurobema.

METHODS

Sample Collection

Mussel specimens of F. flava, F. subrotunda, P. cordatum, P. plenum, and P. sintoxia/rubrum were collected from two sites in the Green River, KY, Pool 4 at river mile (RM)

149 (37.18286, -86.6296) and Mammoth Cave National Park (MCNP) at RM 197 (37.17819, -

86.1154) during 2015 and 2017 (Figure 2.1). Mussels were tagged and kept at the Minor E. Clark

Fish Hatchery near Morehead, Kentucky until collection of morphological data was completed.

Samples of tissue for DNA extraction were collected non-lethally using SK1 buccal swabs

(Isohelix, Harrietsham, UK), and DNA was isolated using the Isohelix DDK-50 kit. DNA concentration and purity were quantified using a µLite PC spectrophotometer (Biodrop,

Cambridge, UK). DNA concentration was diluted to 10-300 ng/µl for all PCRs. Additional individuals of Fusconaia subrotunda (n= 1), Pleurobema plenum (n = 12) and Pleurobema sintoxia/rubrum (n = 4) were collected from the Clinch River. The locations in the Clinch River were Honey Hole (GPS coordinates 36.523311, -83.204240), Frost Ford (GPS coordinates

36.534881, -83.179205), and Kyle’s Ford (GPS coordinates 36.565230, -83.054863). In addition,

Pleurobema plenum (n = 2) and Pleurobema sintoxia/rubrum (n = 3) were collected at other locations in the Tennessee River downstream of Pickwick Dam, Hardin County, TN.

118

Molecular genetic markers

The number of mussel specimens collected, the number of DNA sequences observed at the mitochondrial COI+ND1 genes, and numbers of nuclear microsatellite DNA genotypes for each mussel species collected from the Green River are listed in Table 2.2. All PCR amplifications were performed using either a T100TM or MyCycler TM thermocycler (both from Bio-Rad, Hercules,

CA).

In the case of COI, I used two forward and one reverse primer sequences. The forward primer sequences were LCO1490 (5’- GGT CAA CAA ATC ATA AAG ATA TTG G -3’) and

CO1F (5’- GTT CCA CAA ATC ATA AGG ATA TTG G -3’), while the reverse sequence was

HCO700dy2 (5’- TCA GGG TGA CCA AAA AAY CA -3’) (Campbell et al. 2005; Folmer et al.

1994; Walker et al. 2006). The PCR protocol using the primers LCO1490 (Folmer et al. 1994) and

HCO700dy2 (Walker et al. 2006) was used for amplification of COI of F. flava, F. subrotunda, P. cordatum, and P. sintoxia/rubrum. The PCR amplification was conducted in a volume of 22 µl which contained 0.45X GoTaq Flexi Buffer, 2.05 mM of MgCl2, 0.045 mM of each dNTPs, 0.02 mg/ml of BSA, 0.09 µM of each primer, 0.02 units/µl of GoTaq® DNA polymerase (Promega), and 10-50 ng/µl of dsDNA template. The PCR protocol included an initial denaturation of 94oC for 3 min; followed by 35 cycles of 94oC for 1 min, 57oC for 45 sec, and 72oC for 1 min; a final extension at 72oC for 5 min; and a hold at 4oC. In the case of P. plenum, most sequences were obtained by using primers COIF (Campbell et al. 2005) and HCO700dy2. The PCR reaction was conducted in a volume of 22 µl which contained 0.45X GoTaq Flexi Buffer, 2.27 mM of MgCl2,

0.072 mM of dNTP mix, 0.02 mg/ml of BSA, 0.018 µM of each primer, 0.02 units/µl of GoTaq®

DNA polymerase (Promega), and 10-30 ng/µl of dsDNA template. The PCR protocol included an

119 initial denaturation of 94oC for 3 min; followed by 35 cycles of 94oC for 1 min, 55oC for 45 sec, and 72oC for 1 min; a final extension at 72oC for 5 min; and a hold at 4oC. PCR products were sent to the Fralin Life Sciences Institute (Blacksburg, VA) for Sanger sequencing.

In the case of ND1, I used two pairs of primers for the amplification of ND1, LeuuurF (5’-

TGG CAG AAA AGT GCA TCA GAT TAA AGC - 3’) and LoGlyR (R: 5’- CCT GCT TGG

AAG GCA AGT GTA CT - 3’) and nadh1-F (5’ - TGG AG AAA AGT GCA TCA GAT TTA

AGC - 3’) and nadh1-R (5’ - GCT ATT AGT AGG TCG TAT CG - 3’) (Buhay et al. 2002; Serb et al. 2003). Amplification products were obtained for most mussel specimens of F. flava, F. subrotunda, P. cordatum, and P. plenum using primers LeuuurF and LoGlyR (Serb et al. 2003).

For some P. sintoxia/rubrum mussel specimens, however, it was necessary to use primers nadh1-

F and nadh1-R (Buhay et al. 2002; Serb and Lydeard 2003). For both pairs of primers, PCR reactions were conducted in a volume of 22 µl which contained 0.45X GoTaq Flexi Buffer, 2.04 mM of MgCl2, 0.05 mM of each dNTP, 0.02 mg/ml of BSA, 0.09 µM of each primer, 0.02 units/µl of GoTaq® DNA polymerase (Promega, Madison, WI), and 10-50ng/µl of dsDNA template. The

PCR protocol included an initial denaturation of 95oC for 5 min; followed by 30 cycles of 96oC for 20 sec, 53oC for 20 sec, and 72oC for 45 sec; a final extension at 72oC for 5 min; and a hold at

4oC.

Microsatellite library development for P. plenum was completed at the Savannah River

Ecology Laboratory at the University of Georgia. A genomic library was prepared with inserts size-selected to range from 300–600 bp. Paired-end reads were sequenced on an Illumina HiSeq sequencer. The DNA sequences obtained were analyzed using PAL-Finder_V0.02.03 (Castoe et al. 2012) to locate microsatellite repeat-bearing loci. Candidate PCR primer pairs annealing to sequences just outside the repeat-bearing loci were designed and synthesized (Integrated DNA

120

Technologies, Coralville, Iowa). The criteria used to select these primer pairs were polymorphism of the loci amplified, repeat motif, and primer annealing temperature for use in subsequent multiplexing. Optimization of PCR amplification was completed for 10 loci by testing mussel specimens each of P. plenum, P. cordatum, and P. sintoxia/rubrum. In addition, microsatellite variation was screened using primer pairs developed for P. clava and P. pyriforme by Moyer and

Williams (2011) and Jones et al. (2015) (Table 2.3). For all microsatellite loci, PCR reactions were prepared in a total of 22 µl. The reagents and their concentrations in each reaction were: 0.5x

GoTaq Flexi buffer, 2.5 mM MgCl2, 0.05 mM dNTP mix, 0.02 mg/ml BSA, 0.2 µM of each primer, 0.025 units/µl of GoTaq® DNA polymerase (Promega, Madison, WI), and 10-30 ng/µl of

DNA sample. The PCR protocol included an initial denaturation of 94oC for 3 minutes; followed by 35 cycles of 94oC for 40 seconds, 58oC for 40 seconds, and 72oC for 1 minute. There was a final extension at 72oC for 5 minutes and a hold at 4oC. PCR products were verified by visualization of fluorescence bands in an ethidium bromide-stained agarose gel under UV light.

PCR products were sent to the Institute of Biotechnology at Cornell University, Ithaca, New York for DNA fragment-size analysis.

Mitochondrial DNA Data Analysis

For the mtDNA sequences, consensus sequences were obtained using Geneious® 7.0.6

(Biomatters, Inc., San Francisco, CA), and alignments were constructed using GeneStudio Version

2.2.0.0 (GeneStudio, Inc., Athens, Georgia). To test for population differentiation between various combinations of populations, I estimated the fixation index (FST) in Arlequin v3.5 (Excoffier and

Lischer 2010). Once the respective populations were defined, estimates of genetic diversity (휃휋) for mtDNA were calculated using DnaSP 5.10 (Rozas et al. 2009).

121

Microsatellite DNA Data Analysis

Raw microsatellite data were scored for amplicon size using GeneMarker (SoftGenetics,

State College, PA), and MICRO-CHECKER (Van Oosterhout et al. 2004) was used to estimate the likelihood of any genotyping errors and to infer presence of null alleles. Arlequin v3.5

(Excoffier and Lischer 2010) was used to test for locus-by-locus departures of genotype frequencies from those expected under Hardy-Weinberg equilibrium. To assess the significance of any departures, I used 1 million Markov chain steps and 100,000 dememorization steps. Linkage disequilibrium was assessed in Arlequin with 10,000 permutations after excluding related individuals (to full siblings) to avoid any influence of family structure among individuals which could result in false positive linkage disequilibrium test results between loci. Family relationships among individuals for each species were assessed using ML-relate (Kalinowski et al. 2006).

Population differentiation was estimated using FST values. These were calculated in

Arlequin v3.5 (Excoffier and Lischer 2010). For P. sintoxia/rubrum, I estimated FST values to compare population differentiation between MCNP and Pool 4, and between the Green River with

Tennessee River. After populations were defined, Arlequin v3.5 (Excoffier and Lischer 2010) was used to quantify nuclear DNA diversity metrics (expected and observed heterozygosities, numbers of alleles observed per locus). Estimates of genetic diversity (휃퐻) for microsatellites was calculated using the average heterozygosity in each population as implemented in Arlequin v3.5 (Excoffier and Lischer 2010).

To test for the effects of random genetic drift, Arlequin v3.5 (Excoffier and Lischer 2010) was used to calculate the M-ratio (Garza and Williamson 2001); M-ratios lower than 0.7 are characteristic of populations that have recently suffered a bottleneck. Finally, the inbreeding coefficient FIS and its significance were calculated using Arlequin v3.5 (Excoffier and Lischer

122

2010). Negative FIS values indicate a higher frequency of heterozygotes, and positive values a higher frequency of homozygotes than expected under Hardy-Weinberg expectation, suggesting inbreeding. Inbreeding coefficients close to 0 indicate panmixia within a population.

Long-Term and Contemporary Effective Population Size

The long-term Ne was estimated using current genetic variation from both mitochondrial

DNA (COI and ND1) and nuclear DNA microsatellites. In the case of mtDNA, genetic variation was estimated from the number of haplotypes, mean number of nucleotide differences between sequences, and number of segregating nucleotide sites (Wang 2005; Chapter 1). Genetic variation at microsatellite loci also was quantified in terms of heterozygosity (Wang 2005). The long-term effective population size (Ne) was estimated using DIYABC 2.1.0 (Cornuet et al. 2014), which applies a coalescent approach to microsatellite DNA data. The analysis was conducted using 1 million simulated data sets and assumed an equal female/male ratio.

To estimate the long-term effective population size for females, I used combined mitochondrial COI+ND1 sequences. The best-supported mutation models were assessed using

MEGA (Tamura et al. 2007) for each defined population using AICc values (Corrected Akaike

Information Criterion). The best supported model was used to estimate the long-term effective population number of females Nef using DIYABC 2.1.0 (Cornuet et al. 2014). The analysis was conducted using 1 million simulated data sets and assumed an equal female/male ratio. The

Hasegawa-Kishino-Yano model of mutation was the best-supported model for F. flava (MCNP and Pool 4), and P. plenum (Pool 4). The Tamura-Nei model was the best-supported model for F. subrotunda (MCNP and Pool 4), P. cordatum (MCNP and Pool 4), and P. sintoxia/rubrum (MCNP and Pool 4).

123

Contemporary Ne was estimated using NeEstimator V2.1 (Do et al. 2014) for microsatellite

DNA data using both the linkage disequilibrium (assuming random mating) and heterozygote excess methods. Confidence intervals (95%) were estimated using the jackknife method. The

NeEstimator software utilizes a bias correction for small sample sizes.

RESULTS

Mitochondrial DNA Genetic Diversity

The final sequence length for COI gene was about 471 base-pairs (bp), while the final sequence length for the ND1 gene was about 744 bp. The sample sizes per species per gene are shown in Table 2.4, where P. cordatum had the highest number of mitochondrial COI+ND1 DNA haplotypes with 76, followed by P. sintoxia/rubrum with 26. Fusconaia subrotunda had the highest average number of nucleotide differences among sequences with k = 10.21, while the lowest was observed in F. flava with k = 2.45. Further, F. subrotunda showed the highest haplotype diversity with h = 0.995, while P. plenum had the lowest with h = 0.806, and the nucleotide diversity was highest for F. subrotunda with π = 0.0084, while the lowest was in F. flava with π =

0.00202 (Table 2.4, Figure 2.2).

Population Differentiation using mtDNA

Using mitochondrial DNA sequence data, I estimated population differentiation between the two sampling locations in the Green River (MCNP and Pool 4) for F. flava, F. subrotunda, P. cordatum, and P. sintoxia/rubrum. Populations of F. subrotunda showed the highest differentiation between the two sampling locations (MCNP and Pool 4), with FST = 0.077 (p

<0.001), suggesting that these two demes are only one population. In the case of P. cordatum, the

124

FST = 0.039 (p < 0.001) does not suggest strong differentiation between Pool 4 and MCNP (Table

2.5). Samples collected from the MCNP and Pool 4 for F. flava and P. sintoxia/rubrum exhibited only minor non-significant differentiation and thus were considered as one population (Table 2.5).

In the case of P. plenum, I only had samples collected from Pool 4 in the Green River and 14 additional sequences from the Tennessee River watershed, that were collected primarily from the

Clinch River, TN, with a differentiation of FST = 0.037 (p <0.001) between the two populations.

Regarding the mtDNA results, I caution that only a handful of individuals were collected from the

Tennessee River drainage, and that these collections included specimens from various locations.

Because sampling from the Tennessee River was sparse, population structure between the two sites in this drainage was not tested.

Long-term Female Effective Population Size Estimated using mtDNA

The highest long-term Nef was for P. cordatum (MCNP + Pool 4), with 419,000 individuals

(95% CI = 325,000 – 517,000), followed by F. subrotunda (MCNP + Pool 4) with 160,000

(112,000 – 196,000), F. flava (MCNP + Pool 4) with 69,700 (39,500 – 109,000), P. sintoxia/rubrum (MCNP + Pool 4) with 19,900 (19,600 – 20,000), and P. plenum (Pool 4) with

9.530 (8,290 – 9,910) (Table 2.6, Figure 2.3); prior and posterior probabilities are reported in Table

2. 6.

Microsatellite Genetic Diversity

After removing data for microsatellite loci at which null alleles were segregating

(Appendix Table 2.1), I retained 11 loci for P. cordatum and 9 loci for both P. plenum and P. sintoxia/rubrum (Table 2.7). After removing the individual data for P. cordatum for relationships

125 at the full-sibling and parent/offspring levels, a positive linkage disequilibrium test resulted in removal of all data for locus PpyD10 from subsequent analysis, which showed linkage to loci

Ppl08, Ppl09, PclD104, and PclD106 (Appendix Table 2.2). However, all other linkage disequilibrium tests did not result in the removal of data for any of the remaining loci for P. plenum

(Appendix Table 2.2).

Assessing the presence of null alleles and linkage disequilibrium was more complex for populations of P. sintoxia/rubrum. This was because the data set included individuals collected from two locations in the Green River (Pool 4 and MCNP), as well as seven specimens from various locations in the Tennessee River system. Hence, I first tested for the presence of any potential null alleles segregating within the two putative populations (Appendix Table 2.1). For linkage disequilibrium testing, I used all loci that did not show presence of null alleles in at least one of the two putative populations (Appendix Tables 2.2); linkage disequilibrium did not indicate the need to remove any additional loci. Thus, the loci used for assessment of population structure included Ppl01, Ppl03, Ppl05, Ppl07, Ppl08, PclB11, PclD104, PclD106, PclD9, and PpyD9.

After the subsequent population differentiation analysis resulted in no differentiation between the

Green River and Tennessee River populations, I again tested for segregation of null alleles and for linkage disequilibrium among all P. sintoxia/rubrum collected from the Green River, KY. The removal of loci showing null alleles resulted in retention of nine loci (Table 7). Further, testing for linkage disequilibrium did not result in removal of any additional loci from the dataset (Appendix

Table 2). Loci that were used for subsequent analyses (not including population structure) were

Ppl01, Ppl03, Ppl07, Ppl08, PclB11, PclD104, PclD106, PclD9, and PpyD9.

The microsatellite data showed high levels of expected heterozygosity for my Pleurobema study species (Table 2.7, Figure 2.4). The highest expected heterozygosity was for P. plenum (Pool

126

4) at 0.867, followed by P. cordatum (Pool 4) at 0.849, and P. sintoxia/rubrum (MCNP and Pool

4) at 0.817. However, my results were lower those of Jones et al. (2015), who reported expected heterozygosities between 0.91 (Clinch River) and 0.92 (Green River) for P. plenum. The expected heterozygosity for P. cordatum (Green River), 0.86, was similar to that reported by Jones et al.

(2015). The highest number of alleles was shown for PclD106, followed by PpyD9 (Figure 2.5).

Mean values of the M-ratio were above 0.7 (Table 2.7, Figure 2.6), suggesting that there has not been a recent population bottleneck in any of the populations, P. cordatum – Pool 4, P. plenum – Pool 4, and P. sintoxia/rubrum – MCNP and Pool 4. In addition, mean values of FIS were close to 0, meaning that genotype values approximated Hardy-Weinberg expectations with little to no inbreeding (Table 2.7, Figure 2.7).

Population Structure and Differentiation using Microsatellites

Significant but minor differentiation was observed between the two populations of P. sintoxia/rubrum in the Green River and between populations in the Green River and the Tennessee

River drainage (Table 2.5). For P. sintoxia/rubrum, the FST value between Pool 4 and MCNP was

0.00282 (p-value < 0.001), while the FST value between the Green River and the Tennessee River were 0.0461 (p-value < 0.001).

Long-term Effective Population Sizes Using Microsatellites

Long-term Ne values estimated using the Bayesian DIYABC algorithm showed the highest long-term Ne for P. cordatum, with 17,500 individuals (95% CI 8,570 – 34,300), followed by P. sintoxia/rubrum with 8,370 (5,300 – 11,100), and P. plenum with 5,740 (3,100 – 8,950) (Table

2.6, Figure 2.8), with prior and posterior probabilities available in Table 2.6.

127

Contemporary Effective Population Number (Ne)

Using the linkage disequilibrium method, the highest contemporary Ne was estimated for

P. plenum at 303.3 individuals (95% CI = 68.5-infinity), followed by P. cordatum with 231.4 individuals (41.6-infinity), and P. sintoxia/rubrum with 131.07 individuals (39.5-infinity) (Table

2.8). These numbers were calculated with a lowest allele frequency of 0.05; estimates at least doubled when using a lowest allele frequency of 0.02 (Table 2.8).

DISCUSSION

The principal objective of this chapter was to assess the genetic diversity of mussel species in the genera Fusconaia and Pleurobema in the Green River, KY. Processes affecting genetic diversity in these populations, such as inbreeding and the occurrence of population bottlenecks, also have been assessed to gain insights into genetic and demographic processes operating within and among populations. In addition, I estimated the effective population number (contemporary

Ne, long-term Ne, and long-term female effective population size Nef) for these species, estimates which are useful for the identification of vulnerable populations of conservation interest and the estimation of the viable population number. The importance of this study relies on the relatively large number of mussel specimens collected in comparison to previous studies. As these mussel species are difficult to identify in the field even by the experts, this was the first study assessing these genetic parameters in populations in the Green River.

Genetic diversity

The genetic diversity of my study species was relatively high in terms of numbers of alleles and expected heterozygosities for species of Pleurobema in the Green River. These high numbers

128 are supported by the high mtDNA haplotype diversity estimated for Pleurobema and Fusconaia species. The high genetic diversity of these species in the Green River could be result of these species being demographically connected historically to much larger populations occurring in the mainstem Ohio River and other tributary streams. Dam construction and impoundments likely have isolated and affected these populations in different ways depending on their tolerance to impoundments and the altered flow conditions that they cause. A reduction in the population size of P. clava in the Green River for example has been explained by its low tolerance of impoundments. For the species of interest in this study, their tolerances to such altered flow conditions range from low (e.g., P. clava) to marginally or even moderately tolerant (e.g., P. cordatum).

Inbreeding Coefficient and Population Bottleneck.

Inbreeding may impact the viability of a population by reducing the fitness of individuals

(Allee effect) and affecting their ability to adapt to environmental changes. Inbreeding does not, however, seem to be occurring in populations of P. cordatum (Pool 4), P. plenum (Pool 4) and P. sintoxia/rubrum (MCNP + Pool 4). These inferences match those of Jones et al. (2015) for P. cordatum and P. plenum collected in the Green River, Kentucky. High FIS values observed in these populations could be the result of facultative hermaphroditism in freshwater mussels (van der

Schalie 1970, Jones et al. 2015). Hermaphroditism was reported to occur occasionally for P. cordatum (van der Schalie 1970). However, I have not found studies that report hermaphroditism for P. plenum or P. sintoxia/rubrum, but my observations of M-ratios above the criterion value of

0.7 (Garza and Williamson 2001) suggest that none of these three populations have suffered a recent genetic bottleneck. These results did not match those of Jones et al. (2015) for P. cordatum

129 and P. plenum collected in the Green River, Kentucky, who reported lower M-ratio values for these two species.

Effective Population Size and Viable Population Number

The viable population number (VPN) or minimum viable population number (MVP) is commonly assessed by using the “50/500 rule”. The short-term Ne (50) relates to the population size needed to avoid inbreeding depression, while the long-term Ne (500) relates to the population size needed to maintain evolutionary potential. However, the applicability of this rule might not be supported for mussel species, as it was developed for mammals. A larger VPN may be needed for mussels due to these animals generally being broadcast spawners. This means that there is a need for a high density of potential spawners and partners for most eggs to be fertilized (Jones et al. 2020). This is known as the Allee effect and can decrease the Ne depending on both the location and number of individuals. Discrepancies between past and current Ne can therefore be the result of environmental conditions affecting population sizes in different ways over time. In addition, different tolerances to environmental conditions could explain a larger long-term Ne for some species in comparison to other species. Another important aspect to consider in freshwater mussel populations is the population viabilities of their host fish populations. These mussel species specialize on parasitizing a variety of host fishes (Table 2.1). The assessment of population viability of host fish species and their conservation is important for maintaining freshwater mussel population sizes (Haag 2012). While multiple fish species may support glochidial transformation to the juvenile stage for a specific mussel species, one host fish species usually works better than others. This is the case for P. plenum, which transforms at higher frequencies when utilizing

Striped Shiner compared to Blacknose dace or Spottail shiners (T. Lane, VDWR, unpublished

130 data). A more drastic reduction of some mussel populations than in others could be the result of declines in their preferred host fishes (Modesto et al. 2018). Thus, an accurate estimation of the

VPN also should consider habitat characteristics, life-history strategies, and population threats, among other characteristics. However, these traits can vary greatly among different species, and

VPN size has not been estimated for any mussel species or population, and hence the 50/500 criterion cannot be applied generally to all species of mussels equally (Haag 2012). Applying the

50/500 rule to my results would suggest that P. cordatum (Ne = 231.4, 95%, CI = 41.6-infinity),

P. plenum (Ne = 303.3, 95% CI = 68.5-infinity) and P. sintoxia/rubrum (Ne = 131.07, 95%, CI =

39.5-infinity) have a large enough contemporary Ne to avoid inbreeding depression resulting in adequate individual fitness and the capability of the mussels to survive variation in environmental conditions. Moreover, these estimates of contemporary Ne are large enough to maintain fitness and are supported by the near-zero estimated inbreeding coefficients. The estimates for contemporary and long-term Ne are larger than the ones obtained for other species using the same linkage disequilibrium and coalescent methods reported by Lane et al. (2019) for Venustaconcha trabalis and Venustaconcha troostenesis. In the case of long-term Ne, all my values were much >500 for

P. cordatum (Ne = 17,500; 95% CI = 8,570 – 34,300), P. plenum (Ne = 5,740; 95% CI = 3,100 –

8,950) and P. sintoxia/rubrum (Ne = 8,370; 95% CI = 5,300 – 11,100), suggesting that these populations have the capacity to maintain their long-term evolutionary potential. These large estimated Ne population sizes match the results for high genetic diversity and could be result of these species having once occurred in populations contiguous with much larger ones in the Ohio

River system. However, in order to test for variations in Ne, at a broader geographic scale, additional sampling should include populations of these species in the Green River, the mainstem

Ohio River, and other key tributary rivers, such as the nearby Licking River, KY. It is important

131 to notice that there are inconsistences between long-term Ne and contemporary Ne estimates. I observed that the species with the largest contemporary Ne was P. plenum but this species had the smallest long-term Ne . These inconsistences could be because Ne as well as VPN differ among mussel species because of their different life-history traits.

In the Green River, the principal issues affecting mussel declines are poorly understood, but the most important factors driving declines are thought to be dams and the effects of coal mining, gas and oil extraction, stream channelization, and invasive species (Haag and Cicerello

2016). While water quality is without doubt one of the most important factors affecting mussel densities, its effect has not been well studied (Haag and Cicerello 2016). Water pollution can reduce reproduction of adults, recruitment of juveniles, or both. Hence, an estimate of the VPN could be misleading after assessing Ne in an older generation of mussels in the absence of recruitment of young mussels to the population. Finally, larger estimated long-term Nef in comparison to long-term overall Ne could be the result of an artifact among the respective estimation methods. Mitochondrial DNA is less mutable than microsatellite DNA, which makes mitochondrial variation a more reliable record of what happened in the more distant past, i.e., thousands of years ago. In the software packages that I used to estimate both of these parameters, it may have taken longer for the coalescence to occur with the mtDNA than with the microsatellite data. This faster coalescence time can be the result of homoplasy in which microsatellites with the same fragment length actually may have had different mutational histories, leading to more rapid apparent coalescence than was actually the case, leading to lower Ne estimates than for mitochondrial DNA (Bos et al. 2008). My estimates for contemporary Ne were not large in comparison to the estimates of long-term Ne which may be due to contemporary Ne being estimated using the linkage disequilibrium method. Single-nucleotide polymorphisms (SNPs) seem to be a

132 more appropriate tool to use for estimation of Ne from linkage disequilibrium data because it provides a larger number of genetic markers than microsatellites (Barbato et al. 2015).

Management Implications

The results of this study suggest the need for reassessment of conservation status for F. flava, F. subrotunda, P. cordatum, P. plenum, and P. sintoxia/rubrum. Using the 50/500 rule of thumb, my results suggest that these species meet the classical criteria for VPN. This rule of thumb has been developed for mammals, however, which makes it questionable for application to this suite of mussel species. The investigated populations of Fusconaia and Pleurobema in the Green

River do not appear to be suffering from inbreeding depression, and long-term Ne estimates suggest that each respective population is large enough to maintain evolutionary potential for multiple generations into the future. However, there is a need for continued genetic and demographic monitoring, study of life-history traits such as spawning periods, identification of host fishes, and habitat availability in order to reach defensible inference that these populations are secure. Finally, there is a need to develop additional molecular genetic markers for species in the genus Fusconaia for the assessment of population differentiation, genetic diversity, recent population bottlenecks, among other useful measures for conservation assessment of species in this genus.

133

COLLABORATOR CONTRIBUTIONS

Mussel sampling in the Green River, KY was conducted in collaboration with Chad Lewis and his crew at Lewis Environmental Consulting, LLC. DNA collection and tagging were conducted with the help from Aaron Adkins, Anna Dellapenta, Jess Jones, Tim Lane, and Lee Stephens. Murray

Hyde helped with DNA collection, mussel tagging, and lab work. Insightful input for the methods and discussion was provided by Eric Hallerman, Jess Jones, Emmanuel Frimpong and Pawel

Michalak. I aided in the DNA collection, mussel tagging, lab work, data analysis and prepared the manuscript with assistance from Eric Hallerman and Jess Jones.

134

LITERATURE CITED

Barbato, M., P. Orozco-terWengel, M. Tapio, and M. W. Bruford, M. W. 2015. SNeP: a tool to estimate trends in recent effective population size trajectories using genome-wide SNP data. Frontiers in genetics 6, 109.

Bos, D. H., D. Gopurenko, R. N. Williams, and J. A. DeWoody. 2008. Inferring population history and demography using microsatellites, mitochondrial DNA, and major histocompatibility complex (MHC) genes. Evolution: International Journal of Organic Evolution 62, 1458-1468.

Caballero, A. 1994. Developments in the prediction of effective population size. Heredity 73: 657-679.

Castoe, T. A., A. W. Poole, A. J. de Koning, K. L. Jones, D. F. Tomback, S. J. Oyler-McCance, J. A. Fike, S.L. Lance, J.W. Streicher, and E.N. Smith. 2012. Rapid microsatellite identification from Illumina paired-end genomic sequencing in two birds and a snake. PLoS ONE 7:e30953.

Charlesworth, D., B. Charlesworth, and M. T. Morgan. 1995. The pattern of neutral molecular variation under the background selection model. Genetics 141: 1619-1632.

Cicerello, R. R., and Schuster, G. A. 2003. A guide to the freshwater mussels of Kentucky. Kentucky State Nature Preserves Commission.

Cornuet, J. M., J. Veyssier, P. Pudlo, A. Dehne-Garcia, M. Gautier, R. Leblois R., J. M. Marin, and A. Estoup. 2014. DIYABC v2.0: a software to make approximate Bayesian computation inferences about population history using single nucleotide polymorphism, DNA sequence and microsatellite data. Bioinformatics 30: 1187-1189.

Do, C., R. S. Waples, D. Peel, G. M. Macbeth, B. J. Tillett, and J. R. Ovenden. 2014. NeEstimator V2: re-implementation of software for the estimation of contemporary effective population size (Ne) from genetic data. Molecular Ecology Resources 14: 209-214.

Excoffier, L., and H. E. Lischer. 2010. Arlequin suite ver. 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology Resources 10: 564-567.

135

Freshwater Mussel Host Database. 2017. The freshwater mussel host database, Illinois Natural History Survey & Ohio State University Museum of Biological Diversity, 2017. http://wwx.inhs.illinois.edu/collections/mollusk/data/freshwater-mussel-host-database. (October 2020).

Garza, J. C., and E. G. Williamson. 2001. Detection of reduction in population size using data from microsatellite loci. Molecular Ecology 10: 305-318.

Haag, W. R. 2012. North American freshwater mussels: natural history, ecology, and conservation. Cambridge University Press 391-421.

Haag, W. R., and R.R. Cicerello. 2016. A Distributional Atlas of the Freshwater Mussels of Kentucky. Kentucky State Nature Preserves Commission, Frankfort, KY 136-201.

Hill, W. G. 1981. Estimation of effective population size from data on linkage disequilibrium. Genetical Research 38: 209-216.

Jones, J. W., W.F. Henley, A.J. Timpano, E. Frimpong, and E.M. Hallerman. 2020, Spawning and gravidity of the endangered Oyster Mussel (Epioblasma capsaeformis) in captivity for production of glochidia. Invertebrate Development and Reproduction. (In Press)

Kalinowski, S. T., A. P. Wagner, and M. L. Taper, M. L. 2006. ML‐Relate: a computer program for maximum likelihood estimation of relatedness and relationship. Molecular Ecology Notes 6: 576-579.

Kuhner, M. K. 2006. LAMARC 2.0: maximum likelihood and Bayesian estimation of population parameters. Bioinformatics 22: 768-770.

Lane, T. W., E. M. Hallerman, and J. W. Jones, J. W. 2019. Population genetic assessment of two critically endangered freshwater mussel species, Tennessee bean Venustaconcha trabalis and Cumberland bean Venustaconcha troostensis. Conservation Genetics, 20, 759-779.

Luikart, G., N. Ryman, D. A. Tallmon, M. K. Schwartz, and F. W. Allendorf. 2010. Estimation of census and effective population sizes: the increasing usefulness of DNA-based approaches. Conservation Genetics 11: 355-373.

Modesto, V., M. Ilarri, A. T. Souza, M. Lopes‐Lima, K. Douda, M. Clavero, and R. Sousa. 2018. Fish and mussels: Importance of fish for freshwater mussel conservation. Fish and Fisheries 19, 244-259.

136

Moyer, G. R., and A. S. Williams. 2011. Isolation and characterization of microsatellite loci for the oval pigtoe mussel, Pleurobema pyriforme (Lea 1857). Conservation Genetics Resources 3 :255-257.

Rozas, J. P. Librado, J. C. Sanchez-Delbarrio, X. Messeguer, and R, Rozas. 2009. DnaSP, version 5.10.00. Universita de Barcelona, Spain. http://www.ub.edu/dnasp/.

Tamura, K., J. Dudley, M. Nei, and S. Kumar. 2007. MEGA4: molecular evolutionary genetics analysis (MEGA) software, version 4.0. Molecular Biology and Evolution 24: 1596-1599.

Van der Schalie, H. 1970. Hermaphroditism among North American freshwater mussels. Malacologia 10: 93-112.

Van Oosterhout, C., W. F. Hutchinson, D. P. Wills, and P. Shipley. 2004. MICRO-CHECKER: software for identifying and correcting genotyping errors in microsatellite data. Molecular Ecology Notes 4: 535-538.

Wang, J. 2005. Estimation of effective population sizes from data on genetic markers. Philosophical Transactions of the Royal Society B: Biological Sciences 360: 1395–1409.

Wang, J., and A. Caballero. 1999. Developments in predicting the effective size of subdivided populations. Heredity 82: 212-226.

Waples, R.S. 2006. A bias correction for estimates of effective population size based on linkage disequilibrium at unlinked gene loci. Conservation Genetics 7: 167-184.

Waples, R. S., and C. H. I. Do. 2010. Linkage disequilibrium estimates of contemporary Ne using highly variable genetic markers: a largely untapped resource for applied conservation and evolution. Evolutionary Applications 3: 244-262.

137

Table 2.1. Distribution, host fish and conservation status of study species in the Green River, Kentucky under the U.S. Endangered Species Act (ESA), the International Union for Conservation of Nature (IUNC), American Fisheries Society (AFS), and the Kentucky State Nature Preserves Commission (KSNPC) for species belonging to the genus Fusconaia and Pleurobema.

Status Species General distribution Host fish ESA IUNC Others Black Crappie, Blackside darter, Bluegill, Mississippi River basin, Ohio River basin, Creek Chub, Goldfish, Johnny Darter, Green River, Great Lakes basin. Species Largemouth Bass, Logperch, Longear Fusconaia may have colonized Upper Hudson River - Least concern - Sunfish, Rainbow Darter, Pumpkinseed, flava and Lake Ontario basins. Southern Hudson Silver Redhorse, Silver Shiner, Spotfin Bay basin (Cicerello and Schuster 2003, Shiner, White Crappie (Freshwater Mussel Haag and Cicerello 2016). Host Database 2017). Endemic to the Ohio River Basin, including No known host species. AFS Special Green River and Tennessee River. Concern, Fusconaia Under Historically it has been reported for the Vulnerable KSNPC subrotunda Review Maumee River system of western Lake Erie Special (Cicerello and Schuster 2003, Haag and Concern. Cicerello 2016). Banded Darter, Blacknose Dace, Blackside darter, Bluegill, Bluntnose Minnow, Central Stoneroller, Common Shiner, Creek Chub, Occurs in the Green River and other Fathead Minnow, Golden Redhorse, AFS tributaries to the Ohio River and Great Goldfish, Greenside Darter, Largemouth Pleurobema Critically Endangered, Lakes region. Historically occurred in the Endangered Bass, Logperch, Longear Sunfish, Mountain clava endangered KSNPC Tennessee River and Cumberland River Redbelly Dace, Northern Hog Sucker, Endangered (Cicerello and Schuster 2003, Haag and Rainbow Darter, River Chub, Rock Bass, Cicerello 2016). Striped Shiner, Tippecanoe Darter, Variegate Darter (Freshwater Mussel Host Database 2017). Blacknose Dace, Bluegill, Bluntnose Endemic to and extant in the Ohio River Minnow, Brook Stickleback, Creek Chub, Pleurobema AFS Special Basin. Occurs in the Green River (Cicerello - Near-threatened Guppy, Rosefin Shiner, Scarlet Shiner, cordatum Concern and Schuster 2003, Haag and Cicerello Spotfin Shiner, White Sucker (Freshwater 2016). Mussel Host Database 2017). AFS Endemic to the Ohio River Basin. Extant in Striped Shiner, Blacknose and Spottail Pleurobema Critically Endangered, Tennessee River, Green River and Clinch Shiner (T. Lane, personal communication). Endangered plenum endangered KSNPC River (Cicerello and Schuster 2003, Haag Endangered and Cicerello 2016).

138

Table 2-1. Continued.

Status Species General distribution Host fish ESA IUNC Others AFS Mississippi River Basin, Ohio River Basin Scarlet Shiner, Silver Shiner, Spotfin Shiner, Pleurobema Under Threatened, broadly, including Tennessee River and Streamline chub (Freshwater Mussel Host Near-threatened rubrum Review KSNPC Green River (Cicerello and Schuster 2003, Database 2017). Endangered Haag and Cicerello 2016). Bigeye Shiner, Blacknose Dace, Blacktail Shiner, Bluegill, Bluenose Minnow, Mississippi River Basin, Ohio River Basin, Bluntnose Minnow, Brook Stickleback, Pleurobema Great Lakes basin, Green River (Cicerello Central Stoneroller, Common Shiner, Creek - Least concern - sintoxia and Schuster 2003, Haag and Cicerello Chub, Northern Redbelly Dace, Red Shiner, 2016). Southern Redbelly Dace, Spotfin Shiner, Whitetail Shiner (Freshwater Mussel Host Database 2017).

139

Table 2.2. Samples sizes for mtDNA for cytochrome c oxidase subunit 1 (COI) and NADH dehydrogenase subunit 1 (ND1) about 1215 bp and nuclear DNA microsatellites loci for investigated species of Fusconaia and Pleurobema in the Green River, Kentucky. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. mtDNA DNA microsatellite loci Collection sites N COI + ND1 Ppl01 Ppl03 Ppl07 Ppl08 Ppl09 PclB11 PclD104 PclD106 PclD9 PpyC106 PpyD9

Fusconaia flava Pool 4 and MCNP 43 42 ------Fusconaia subrotunda Pool 4 and MCNP 22 20 ------Pleurobema cordatum Pool 4 and MCNP 117 116 32 32 32 32 32 32 32 32 - 28 32 Pleurobema plenum Pool 4 33 32 33 33 33 32 32 33 - - 33 29 33 Pleurobema sintoxia/rubrum Pool 4 and MCNP 43 41 43 43 43 43 - 43 42 42 43 - 38

140

Table 2.3. The DNA primer sequences, base-pair range of individual alleles, and repeat motif of the microsatellite PCR primers used in this study. Ten PCR primers were developed for Pleurobema plenum. PCR primers developed for Pleurobema clava and Pleurobema pyriforme in previous studies also were used. Unavailable GenBank accession numbers were labeled as “NA”.

Primer Sequence Base-Pair range Repeat Motif Reference Genbank no. Ppl01 F: 5' - CTT TCC ACA TGC ACA CAA GG - 3' 141-186 AAT This study NA R: 5' - TAA GTG ACC TCT TGA CAC GC - 3' Ppl02 F: 5' - CAG CAA GAC CAA ATA GCA GC - 3' 126-222 AAT This study NA R: 5' - ACG ATG AAG TGT AGG CAG ATG - 3' Ppl03 F: 5' - CTC GTG CTA CTT GGT TGT CC - 3' 114-156 AAT This study NA R: 5' - TGT GGA CGT GTT CTG TCA TTA G - 3' Ppl04 F: 5' - CGG TGT ATT AAA CTG CAT CCA C - 3' 129-234 AAT This study NA R: 5' - TTC AAA CGC CAG ACT TTC AAG - 3' Ppl05 F: 5' - GTC TCT CAC TGT TTC CAC GTG - 3' 168-231 AAT This study NA R: 5' - CAG TGG CTT TAG AGA TGC AGC - 3' Ppl06 F: 5' - TTG GGT GCT ATA CAA AGT CCT C - 3' 156-240 AAT This study NA R: 5' - TGT CAA ACG AGT TCA AAT GTC C - 3' Ppl07 F: 5' - AGC AGG TAC CTA TGA GCT CAC - 3' 138-180 AAT This study NA R: 5' - TTG GCT AAT TAT GTC ACC CAG G - 3' Ppl08 F :5' - GTA AAG TCA AGG TGG TGT CCC - 3' 99-153 ATC This study NA R: 5' - CGC ATG ATA GGC TAC GCA TG - 3' Ppl09 F: 5' - TCC TGG ATG ATG AGT AGC ACA G - 3' 129-165 ATC This study NA R: 5' - CAA AGG GCA GCA GAG AAC AG - 3' Ppl10 F: 5' - ACT GTA TCG GTC ACA GAA TGT C -3' 159-210 AAG This study NA R: 5' - TGT TAT TCA AAT CCA CGC TTC C -3' PclB11 F: 5’ - GGC GTA GTT TGA ACC ATT C - 3’ 163-223 Dinucleotide Jones et al. 2015 NA R: 5’ - TTG AAA TCT GCC CCA TAA C - 3’ PclC121 F: 5’ - TCA AAA GAC CCT CTT AGC ATA G - 3’ 126-272 Dinucleotide Jones et al. 2015 NA R: 5’ - AGA TTG GGA GCC TAT CAC A - 3’ PclD104 F: 5’ - TAT CAA CCC CAGA CA TTA CCA G - 3’ 88-244 Tetranucleotide Jones et al. 2015 NA R: 5’ - GTT TCT GAT GAC AAA TCC CTT C - 3’

141

Table 2.3. Continued. Primer Sequence Base-Pair range Repeat Motif Reference Genbank No. PclD106 F: 5’ - TGC TGT AAT AAA CAT TTG TCA CCT AC - 3’ 148-346 Dinucleotide Jones et al. 2015 NA R: 5’ - CAA ACA TTG TGT GCA GTT AGG AC - 3’ PclD9 F: 5’ - AAT TCC TTA ACG TCA AGT TCC TC - 3’ 173-271 Dinucleotide Jones et al. 2015 NA R: 5’ - GCA ATA TAA GCA ACA CGA TAC G - 3’ PpyD10 F: 5’ - CCT TAT CTC ATC GCC ACT ATG - 3’ 250-342 TAGA Moyer & Williams 2011 HQ285942 R: 5’ - ATT GCC AGC ACC CAA ATA - 3’ PpyC106 F: 5’ - CAT GCC AAA TTC ATT CGT G - 3’ 202-322 CATC Moyer & Williams 2011 HQ285940 R: 5’ - TCT CGG ACT GTC TGT CTA TAC G - 3’ PpyD9 F: 5’ - GCT GGA TAA ATG AGC AAA TG - 3’ 148-288 TAGA Moyer & Williams 2011 HQ285938 R: 5’ - CAC CTA AGC ACA ATG GTA AAT C - 3’ PpyD101 F: 5’ - CCA GAT TCC CATC AGT TAC G - 3’ 138-318 TAGA Moyer & Williams 2011 HQ285939 R: 5’ - CTC CAC CC ACTT TCA CCT - 3’

142

Table 2.4. Intraspecific variation of the mitochondrial DNA cytochrome c oxidase subunit 1 (COI) and NADH dehydrogenase subunit 1 (ND1) genes of about 1215 bp in total length, for species of freshwater mussels in the genera Fusconaia and Pleurobema. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (MCNP) (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

Number of Number of Average number of Haplotype Nucleotide Species Site location N variable sites haplotypes nucleotide differences k diversity (h) diversity (π)

Fusconaia flava Pool 4 and MCNP 42 20 15 2.448 0.899 0.00202 Fusconaia subrotunda Pool 4 and MCNP 20 46 19 10.211 0.995 0.0084 Pleurobema cordatum Pool 4 and MCNP 116 129 76 5.732 0.96 0.00472 Pleurobema plenum Pool 4 32 28 15 5.415 0.806 0.00446 Pleurobema sintoxia/rubrum Pool 4 and MCNP 41 57 26 4.5 0.962 0.0037

143

Table 2.5. Genetic differentiation (FST) values estimated using mitochondrial DNA (mtDNA) and nuclear DNA (nDNA) microsatellite data. Mussel specimens belonging to the genus Fusconaia and Pleurobema were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (MCNP) (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. mtDNA nDNA

Species Sites FST P-value FST P-value Fusconaia flava MCNP / Pool 4 -0.019 0.776 - - Fusconaia subrotunda MCNP / Pool 4 0.077 0.077 - - Pleurobema cordatum MCNP / Pool 4 0.039 < 0.001 - - Pleurobema plenum Pool 4 / Tennessee River 0.037 0.023 - - Pleurobema sintoxia/rubrum MCNP / Pool 4 0.014 0.1 0.0028 < 0.001 Pleurobema sintoxia/rubrum Green River / Tennessee River 0.12 < 0.001 0.0462 < 0.001

144

Table 2.6. Long-term female effective population (Nef) sizes estimated from combined mtDNA sequences of cytochrome c oxidase subunit 1 (COI=471) and NADH dehydrogenase subunit 1 (ND1=744 bp) and long-term (Ne) estimated using nuclear DNA (nDNA) microsatellite data using an approximate Bayesian computational approach implemented in DIYABC 2.1.0 (Cornuet et al. 2014). Mussel specimens belonging to the genus Fusconaia and Pleurobema were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (MCNP) (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Prior distribution Posterior distribution 95% Confidence Species Sites Parameter N Min Max Distribution Median Intervals

Fusconaia flava MCNP / Pool 4 Nef 42 10 130,000 Uniform 69,000 39,500 – 109,000

Fusconaia subrotunda MCNP / Pool 4 Nef 20 10 208,000 Uniform 160,000 112,000 – 196,000

mtDNA Pleurobema cordatum MCNP / Pool 4 Nef 116 10 560,000 Uniform 419,000 325,000 – 517,000

Pleurobema plenum Pool 4 Nef 32 10 10,000 Uniform 9,530 8,290 – 9,910

Pleurobema sintoxia/rubrum MCNP / Pool 4 Nef 41 10 20,000 Uniform 19,900 19,600 – 20,000 Pleurobema cordatum Pool 4 Ne 32 10 45,000 Uniform 17,500 8,570 – 34,300 nDNA Pleurobema plenum Pool 4 Ne 33 10 14,650 Uniform 5,740 3,100 – 8,950 Pleurobema sintoxia/rubrum MCNP / Pool 4 Ne 43 10 12500 Uniform 8,370 5,300 – 11,100

145

Table 2.7. Genetic variation among DNA microsatellites that were amplified for Pleurobema cordatum, Pleurobema plenum, and Pleurobema sintoxia/rubrum, where: N = number of individuals, He = expected heterozygosity, Ho = observed heterozygosity, A = number of alleles, Range = base-pairs between the largest and the shortest alleles, M-ratio = the ratio of the number of alleles to range in allele size, FIS = inbreeding coefficient, HW = test of departure from Hardy-Weinberg equilibrium. Mussel specimens belonging to the genus Pleurobema were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (MCNP) (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

Species Locus N Ho He A Range (bp) M-ratio FIS HW Pleurobema Ppl01 32 0.688 0.763 9 27 0.900 0.101 0.163 cordatum Pool 4 Ppl03 32 0.938 0.876 12 42 0.800 -0.072 0.179 Ppl07 32 0.719 0.786 9 33 0.750 0.086 0.926 Ppl08 32 0.781 0.906 12 33 1.000 0.140 0.008 Ppl09 32 0.781 0.862 12 36 0.923 0.095 0.104 PclB11 32 0.469 0.580 10 30 0.625 0.194 0.012 PclD104 32 0.813 0.901 12 60 0.750 0.100 0.198 PclD106 32 0.969 0.946 28 124 0.444 -0.024 0.830 PpyC106 28 0.857 0.918 20 76 1.000 0.067 0.344 PpyD9 32 0.906 0.951 25 120 0.806 0.048 0.431 Mean 31.600 0.792 0.849 14.900 58.100 0.800 0.074 0.320 SD 1.300 0.146 0.113 6.900 36.900 0.172 0.076 0.323 Pleurobema Ppl01 33 0.758 0.822 11 30 1.000 0.079 0.163 plenum Ppl03 33 0.788 0.876 11 39 0.780 0.102 0.312 Pool 4 Ppl07 33 0.939 0.847 9 24 1.000 -0.111 0.687 Ppl08 32 0.906 0.893 12 51 0.667 -0.015 0.525 Ppl09 32 0.781 0.719 7 18 1.000 -0.088 0.465 PclB11 33 0.939 0.893 11 60 0.355 -0.053 0.968

PclD9 33 0.848 0.904 13 56 0.867 0.062 0.751 PpyC106 29 0.793 0.924 14 120 0.452 0.144 0.274 PpyD9 33 0.909 0.929 21 108 0.750 0.022 0.355 Mean 32.300 0.851 0.867 12.111 56.200 0.764 0.016 0.500 SD 1.300 0.073 0.066 3.919 35.800 0.238 0.088 0.260 Pleurobema Ppl01 43 0.744 0.818 11 36 0.846 0.091 0.266 sintoxia/rubrum Ppl03 43 0.907 0.913 14 42 0.933 0.006 0.421 MCNP and Pool 4 Ppl07 43 0.744 0.782 10 33 0.833 0.050 0.390 Ppl08 43 0.512 0.593 11 33 0.917 0.138 0.011 PclB11 43 0.721 0.779 10 28 0.667 0.076 0.182 PclD104 42 0.929 0.957 26 128 0.788 0.030 0.716 PclD106 42 0.905 0.980 42 150 0.553 0.078 0.025 PclD9 43 0.465 0.569 12 62 0.375 0.184 0.046 PpyD9 38 0.947 0.966 29 140 0.408 0.019 0.640 Mean 42.200 0.764 0.817 18.300 72.400 0.702 0.075 0.300 SD 1.600 0.179 0.155 11.400 51.400 0.213 0.058 0.262

146

Table 2.8. Estimates of contemporary effective population sizes (Ne) estimated from DNA microsatellite data that are bias-corrected using the linkage disequilibrium method (Hill 1981; Waples 2006; Waples and Do 2010). Mussel specimens belonging to the genus Fusconaia and Pleurobema were collected in Pool 4 and Mammoth Cave National Park (MCNP) in the Green River, Kentucky. Lowest allele frequency used 0.05 0.02

Pleurobema cordatum Estimated Ne 231.4 719.8 Pool 4 95% CIs for Ne (Jackknife) 41.6- Infinite 106.0 - Infinite

Pleurobema plenum Estimated Ne 303.3 2035.4 Pool 4 95% CIs for Ne (Jackknife) 68.5 – Infinite 156.8 - Infinite

Pleurobema sintoxia/rubrum Estimated Ne 131.7 261.3

MCNP and Pool 4 95% CIs for Ne (Jackknife) 39.5 – Infinite 82.2 – Infinite

147

Figure 2.1. Sampling locations for freshwater mussel species in the genera Fusconaia and Pleurobema. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

148

A. Number of variable sites B. Number of haplotypes C. Average number of nucleotide differences k

140 Number of haplotypes 12 80 120 10 70 100 60 8

80 50

k 6 60 40 30 4 40

Number of variable sites variable of Number 20

Number Number of haplotypes 2 20 10

0 0 0 Fusconaia flava Fusconaia Pleurobema Pleurobema Pleurobema Fusconaia flava Fusconaia Pleurobema Pleurobema Pleurobema Fusconaia flava Fusconaia Pleurobema Pleurobema Pleurobema subrotunda cordatum plenum sintoxia/rubrum subrotunda cordatum plenum sintoxia/rubrum subrotunda cordatum plenum sintoxia/rubrum

D. Haplotype diversity (h) E. Nucleotide diversity (π) 1.2 0.009

0.008 1 0.007

0.8 0.006

0.005

h 0.6 π 0.004

0.4 0.003

0.002 0.2 0.001

0 0 Fusconaia flava Fusconaia Pleurobema Pleurobema Pleurobema Fusconaia flava Fusconaia Pleurobema Pleurobema Pleurobema subrotunda cordatum plenum sintoxia/rubrum subrotunda cordatum plenum sintoxia/rubrum

Figure 2.2. Summary of intraspecific mitochondrial DNA of combined sequences (1251 bp) of cytochrome c oxidase subunit 1 (COI) and NADH dehydrogenase subunit 1 (ND1) variation for individuals of Fusconaia flava, Fusconaia subrotunda, Pleurobema cordatum, Pleurobema plenum, and Pleurobema sintoxia/rubrum. Estimated parameters are number of variable sites (A), number of haplotypes (B), average number of nucleotide differences k (C), haplotype diversity h (D), nucleotide diversity π (E). Individuals were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, - 86.1154; river mile = 197) in the Green River, Kentucky.

149

Figure 2.3. Effective female population sizes (Nef) estimated using mtDNA of combined sequences (1251 bp) of cytochrome c oxidase subunit 1 (COI) and NADH dehydrogenase subunit 1 (ND1) for Fusconaia flava, F. subrotunda, Pleurobema cordatum, P. plenum, and P. sintoxia/rubrum. Error bars represent 95% confidence intervals. This parameter was estimated in DIYABC.Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

150

Figure 2.4. Expected (He) and observed (Ho) heterozygosities for Pleurobema cordatum , Pleurobema plenum and Pleurobema sintoxia/rubrum estimated from DNA microsatellites. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

151

Figure 2.5. Number of alleles (A) at each locus from DNA microsatellites for Pleurobema cordatum, Pleurobema plenum, and Pleurobema sintoxia/rubrum. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

152

Figure 2.6. Mean M-ratios values for freshwater mussels belonging to Pleurobema cordatum, P. plenum, and P. sintoxia/rubrum. Error bars represent standard deviation of the mean. M-ratio values were estimated from DNA microsatellites, where values below 0.7 suggest the recent occurrence of a recent genetic bottleneck. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile =149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

153

0.200

0.150

) IS

0.100

0.050

0.000 Inbreeding coefficient (F Inbreeding -0.050

-0.100 Pleurobema Pleurobema Pleurobema cordatum plenum sintoxia/rubrum Species

Figure 2.7. Mean inbreeding coefficient (FIS) for freshwater mussels belonging to Pleurobema cordatum, P. plenum, and P sintoxia/rubrum estimated from DNA microsatellites. Error bars represent standard deviation of the mean. FIS values close to 0 indicate conformance of genotype ratios to Hardy-Weinberg expectation. Positive FIS suggest inbreeding and values that equal 1 suggest that there are no heterozygotes in the population. Finally, FIS values that equal to -1 suggest that there is an excess of heterozygous in the population. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

154

Figure 2.8. Long-term effective population sizes (Ne) estimated using nuclear DNA microsatellites for Pleurobema cordatum, P. plenum, and P. sintoxia/rubrum. Error bars represent 95% confidence interval. These parameters were estimated using DIYABC software. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

155

APPENDIX – SUPPLEMENTAL TABLES AND FIGURES

APPENDIX Table 2.1. Test results for the segregation of null alleles for species of Pleurobema cordatum, Pleurobema plenum, and P. sintoxia/rubrum. Null allele frequencies were estimated by using the method of Oosterhout at al. (2004). Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. . Null Estimated Species Locus Alleles N HW Frequency Present Pleurobema cordatum Ppl01 no 0.039 32 0.163 Pool 4 Ppl02 yes 0.431 28 0 Ppl03 no -0.048 32 0.179 Ppl06 yes 0.316 19 0 Ppl07 no 0.038 32 0.926 Ppl08 no 0.059 32 0.008 Ppl09 no 0.042 32 0.104 Ppl10 yes 0.411 31 0 PclB11 no 0.075 32 0.012 PclC121 yes 0.185 32 0 PclD104 no 0.039 32 0.198 PclD106 no -0.02 32 0.83 PclD9 yes 0.269 26 0 PpyD10 no -0.033 31 1 PpyC106 no 0.027 28 0.344 PpyD9 no 0.017 32 0.431 PpyD101 yes 0.103 32 0.005 Mean 0.115 30.2 0.247 SD 0.152 3.4 0.347 Pleurobema plenum Ppl01 no 0.031 33 0.163 Pool 4 Ppl02 yes 0.277 32 0 Ppl03 no 0.045 33 0.312 Ppl04 yes 0.119 33 0 Ppl05 yes 0.288 33 0 Ppl06 yes 0.129 28 0.001 Ppl07 no -0.065 33 0.687 Ppl08 no -0.021 32 0.525 Ppl09 no -0.063 32 0.465 Ppl10 yes 0.426 25 0 PclB11 no -0.034 33 0.968 PclC121 yes 0.197 33 0.001 PclD104 yes 0.14 33 0

156

APPENDIX Table 2.1. Continued.

Null Estimated Species Locus Alleles N HW Frequency Present PclD106 yes 0.102 33 0 PclD9 no 0.025 33 0.751 PpyD10 yes 0.313 29 0 PpyC106 no 0.064 29 0.274 PpyD9 no 0.004 33 0.355 PpyD101 yes 0.053 33 0.083 Mean 0.107 31.7 0.241 SD 0.138 2.3 0.306 Pleurobema Ppl01 yes 0.105 21 0.139 sintoxia/rubrum Ppl02 yes 0.429 18 0 MCNP Ppl03 no 0.109 21 0.4 Ppl04 yes 0.225 20 0 Ppl05 yes 0.146 21 0.049 Ppl07 no 0.04 21 0.071 Ppl08 no 0.092 21 0.137 Ppl10 yes 0.4 19 0 PclB11 no -0.013 21 0.576 PclC121 yes 0.245 21 0 PclD104 no 0.023 21 0.048 PclD106 no 0.052 21 0.017 PclD9 no 0.069 21 0.198 PpyD10 yes 0.163 15 0.008 PpyC106 yes 0.208 19 0 PpyD9 no -0.003 20 0.648 PpyD101 yes 0.305 17 0 Mean 0.153 19.9 0.135 SD 0.133 1.8 0.208 Pleurobema Ppl01 no -0.03 22 0.21 sintoxia/rubrum Ppl02 yes 0.338 9 0 Pool 4 Ppl03 no 0.031 22 0.901 Ppl04 yes 0.363 19 0 Ppl05 no 0 22 - Ppl07 no -0.015 22 0.84 Ppl08 no 0.05 22 0.066 Ppl10 yes 0.233 18 0 PclB11 no 0.05 22 0.145 PclC121 yes 0.322 22 0

157

APPENDIX Table 2.1. Continued.

Null Estimated Species Locus Alleles N HW Frequency Present PclD104 no -0.013 21 0.911 PclD106 no 0.004 21 0.323 PclD9 no 0.027 22 0.223 PpyD10 yes 0.271 18 0 PpyC106 yes 0.154 21 0 PpyD9 no -0.009 18 0.764 PpyD101 yes 0.315 0 Mean 0.123 20.063 0.274 SD 0.148 3.356 0.361 Pleurobema Ppl01 no 0.04 43 0.266 sintoxia/rubrum Ppl02 yes 0.404 27 0 MCNP and Pool 4 Ppl03 no -0.003 43 0.421 Ppl04 yes 0.297 39 0 Ppl05 yes 0.117 43 0.021 Ppl07 no 0.016 43 0.39 Ppl08 no 0.073 43 0.011 Ppl10 yes 0.326 37 0 PclB11 no 0.024 43 0.182 PclC121 yes 0.289 43 0 PclD104 no 0.009 42 0.716 PclD106 no 0.033 42 0.025 PclD9 no 0.083 43 0.046 PpyD10 yes 0.231 33 0 PpyC106 yes 0.19 40 0 PpyD9 no 0.003 38 0.64 PpyD101 yes 0.315 35 0 Mean 0.144 39.8 0.16 SD 0.139 4.6 0.24

158

APPENDIX Table 2.2. Linkage disequilibrium tests for loci that did not show presence of null alleles. The criterion significance level was set at 0.05. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. + = significant test and - = non-significant test.

A. Pleurobema cordatum – Pool 4

Ppl01 Ppl03 Ppl07 Ppl08 Ppl09 PCLB11 PclD104 PclD106 PpyD10 PpyC106 PpyD9 Ppl01 * Ppl03 - * Ppl07 - - * Ppl08 - - - * Ppl09 - - - + * PclB11 - - - - + * PclD104 ------* PclD106 ------* PpyD10 - - - + + - + + * PpyC106 - + - - - - - + - * PpyD9 ------+ *

B. Pleurobema plenum – Pool 4

Ppl01 Ppl03 Ppl07 Ppl08 Ppl09 PclB11 PclD9 PpyC106 PpyD9 Ppl01 * Ppl03 - * Ppl07 - - * Ppl08 + - + * Ppl09 + + - + * PclB11 - - - - - * PclD9 ------* PpyC106 ------* PpyD9 ------*

159

APPENDIX Table 2.2. Continued.

C. Pleurobema sintoxia/rubrum – MCNP

Ppl01 Ppl03 Ppl05 Ppl07 Ppl08 PclB11 PclD104 PclD106 PclD9 PpyD9 Ppl01 * Ppl03 + * Ppl05 - - * Ppl07 - - - * Ppl08 - - - - * PclB11 - - - - + * PclD104 - - - - + - * PclD106 ------* PclD9 - - - - + - + - * PpyD9 - - + - - - + + - *

D. Pleurobema sintoxia/rubrum – Pool 4

Ppl01 Ppl03 Ppl05 Ppl07 Ppl08 PclB11 PclD104 PclD106 PclD9 PpyD9 Ppl01 * Ppl03 - * Ppl05 - - * Ppl07 - - - * Ppl08 - - - - * PclB11 - - - - - * PclD104 - + - + - + * PclD106 - - - - - + - * PclD9 - - - - - + - - * PpyD9 ------*

160

APPENDIX Table 2.2. Continued.

E. Pleurobema sintoxia/rubrum – MCNP and Pool 4

Ppl01 Ppl03 Ppl07 Ppl08 PclB11 PclD104 PclD106 PclD9 PpyD9 Ppl01 * Ppl03 - * Ppl07 - - * Ppl08 - - - * PclB11 - - - + * PclD104 - - - - + * PclD106 ------* PclD9 - - - - + - - * PpyD9 ------*

161

APPENDIX Table 2.3. Nuclear DNA microsatellite allele frequencies for Pleurobema cordatum, P. plenum, and P. sintoxia/rubrum. Mussel specimens were collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (MCNP) (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky. Additional individuals were collected in the Clinch River. The locations in the Clinch River were Honey Hole (GPS coordinates 36.523311, -83.204240), Frost Ford (GPS coordinates 36.534881, -83.179205), and Kyle’s Ford (GPS coordinates 36.565230, -83.054863). In addition, more individuals were collected at other locations in the Tennessee River downstream of Pickwick Dam, Hardin County, TN. Pleurobema Pleurobema Pleurobema Pleurobema Pleurobema Pleurobema sintoxia/ sintoxia/ sintoxia/ sintoxia/ Locus Allele cordatum plenum rubrum rubrum rubrum rubrum Pool 4 Pool 4 MCNP Pool 4 Clinch Tennessee Ppl01 141 - 0.03 - - - -

144 0.016 0.015 - - - - 147 - 0.03 - - - - 150 0.266 0.167 0.357 0.341 0.375 0.167 153 0.031 0.349 0.167 0.114 0.125 - 156 0.109 0.136 0.167 0.159 - - 159 0.391 0.091 0.048 0.091 0.125 - 162 0.047 0.061 0.024 0.068 - - 165 0.094 0.091 0.119 0.091 0.125 0.5 168 0.031 0.015 0.119 0.023 0.125 - 171 0.016 0.015 - - - - 174 - - - 0.023 0.125 0.167 177 - - - - - 0.167 180 - - - 0.023 - - 183 - - - 0.046 - - 186 - - - 0.023 - - Ppl02 126 - 0.031 - - - -

135 - - 0.056 - - - 147 0.411 - 0.389 0.222 - - 150 - - 0.056 - 1 - 159 0.036 0.109 0.194 0.111 - - 162 - 0.125 0.056 - - - 165 0.036 0.031 0.25 0.333 - - 168 0.357 0.078 - - - - 171 0.071 0.156 - - - - 174 - 0.125 - - - - 177 - 0.078 - - - - 180 - 0.219 - 0.111 - - 183 - - - 0.056 - - 186 - 0.031 - 0.167 - - 189 0.036 - - - - - 195 0.054 - - - - - 222 - 0.016 - - - - Ppl03 114 0.031 - 0.024 - - -

117 0.063 0.03 0.095 0.023 0.125 0.167

162

APPENDIX Table 2.3. Continued.

Pleurobema Pleurobema Pleurobema Pleurobema

Pleurobema sintoxia/ sintoxia/ sintoxia/ sintoxia/ Locus Allele Pleurobema plenum rubrum rubrum rubrum rubrum cordatum Pool 4 MCNP Pool 4 Clinch Tennessee Pool 4

120 0.109 - 0.071 0.091 0.25 0.167 123 0.047 0.061 0.191 0.046 0.125 0.333 126 0.25 0.046 0.191 0.114 - - 129 0.156 0.136 0.095 0.091 0.25 - 132 0.109 0.091 0.095 0.169 - 0.167 135 0.125 0.212 0.095 0.136 0.125 0.167 138 0.063 0.182 0.024 0.114 - - 141 - 0.152 0.048 0.091 - - 144 - 0.046 0.024 0.046 0.125 - 147 0.016 0.03 0.024 0.046 - - 153 0.016 - 0.024 0.023 - - 156 0.016 0.015 - 0.023 - - Ppl04 129 - - - 0.053 - -

132 - 0.03 0.025 - - - 135 - - 0.375 0.316 0.25 - 138 - 0.061 - - - - 144 - 0.136 - - - - 150 - 0.046 - - - - 153 - 0.152 0.15 0.105 0.5 - 156 - 0.152 0.1 0.026 - - 159 - 0.136 - - - - 162 - 0.091 0.075 - - 0.5 165 - 0.03 0.025 0.053 0.25 - 168 - 0.015 0.025 0.026 - - 171 - 0.121 0.025 - - - 174 - 0.015 0.025 0.026 - 0.25 177 - 0.015 0.05 0.026 - 0.25 183 - - - 0.053 - - 186 - - 0.05 - - - 189 - - 0.025 0.237 - - 192 - - 0.025 - - - 201 - - - 0.053 - - 234 - - 0.025 0.026 - - Ppl05 168 - 0.015 - - - -

171 - 0.03 - - - - 174 - 0.03 - - - -

163

APPENDIX Table 2.3. Continued.

Pleurobema Pleurobema Pleurobema Pleurobema

Pleurobema sintoxia/ sintoxia/ sintoxia/ sintoxia/ Locus Allele Pleurobema plenum rubrum rubrum rubrum rubrum cordatum Pool 4 MCNP Pool 4 Clinch Tennessee Pool 4

177 - - 0.024 - - - 180 - 0.046 - - - - 183 - 0.455 0.905 1 1 1 186 - 0.106 - - - - 189 - 0.076 - - - - 192 - 0.136 - - - - 195 - 0.106 - - - - 201 - - 0.048 - - - 231 - - 0.024 - - - Ppl06 156 0.395 - - - - -

159 0.053 - - - - - 162 0.053 - - - - - 183 0.132 0.161 - - - - 195 0.026 - - - - - 198 0.053 0.054 - - - - 201 0.158 0.071 - - - - 207 - 0.089 - - - - 210 - 0.196 - - - - 213 - 0.179 - - - - 216 - 0.107 - - - - 219 - 0.054 - - - - 222 - 0.036 - - - - 225 - 0.054 - - - - 231 0.053 - - - - - 240 0.079 - - - - - Ppl07 138 0.016 - - - - -

141 0.031 - - - - - 147 0.313 0.076 0.048 0.023 - - 150 0.063 0.076 0.143 0.136 0.125 - 153 0.281 0.258 0.119 0.046 0.5 0.833 156 0.203 0.212 0.405 0.409 - 0.167 159 0.047 0.167 0.048 0.023 0.125 - 162 - 0.03 0.024 0.046 - - 165 0.031 0.091 0.119 0.205 - - 168 - 0.076 0.095 0.023 - - 171 0.016 0.015 - 0.023 - -

164

APPENDIX Table 2.3. Continued.

Pleurobema Pleurobema Pleurobema Pleurobema

Pleurobema sintoxia/ sintoxia/ sintoxia/ sintoxia/ Locus Allele Pleurobema plenum rubrum rubrum rubrum rubrum cordatum Pool 4 MCNP Pool 4 Clinch Tennessee Pool 4

174 - - - - 0.125 - 177 - - - - 0.125 - 180 - - - 0.068 - - Ppl08 99 0.063 - - 0.091 - -

102 0.047 0.094 0.667 0.591 0.375 0.667 105 0.141 0.094 0.095 0.023 0.375 0.333 108 0.063 0.094 - 0.023 - - 111 0.188 0.031 - 0.023 - - 114 0.047 0.063 - - 0.125 - 117 0.125 0.016 0.048 - - - 120 0.109 0.234 0.024 0.068 - - 123 0.063 0.078 0.071 0.114 - - 126 0.078 0.141 0.048 0.046 - - 129 0.047 0.063 - 0.023 0.125 - 132 0.031 - 0.048 - - - 135 - 0.063 - - - - 153 - 0.031 - - - - Ppl09 129 0.016 - - - - -

132 0.031 0.031 - - - - 135 0.047 0.063 - - - - 138 0.109 0.484 - - - - 141 0.063 0.156 - - - - 144 0.141 0.109 - - - - 147 0.125 0.125 - - - - 150 0.297 0.031 - - - - 153 0.063 - - - - - 156 0.063 - - - - - 162 0.031 - - - - - 165 0.016 - - - - - Ppl10 159 0.016 - - - - -

162 - 0.06 - - - - 165 0.032 0.02 0.053 0.056 - - 168 - 0.08 0.053 - - - 171 0.065 - - 0.083 0.5 - 174 0.032 0.06 0.105 0.194 0.25 - 177 0.29 - 0.158 0.167 0.125 -

165

APPENDIX Table 2.3. Continued.

Pleurobema Pleurobema Pleurobema Pleurobema

Pleurobema sintoxia/ sintoxia/ sintoxia/ sintoxia/ Locus Allele Pleurobema plenum rubrum rubrum rubrum rubrum cordatum Pool 4 MCNP Pool 4 Clinch Tennessee Pool 4

180 0.129 0.08 0.079 0.222 0.125 0.333 183 0.081 0.12 0.132 0.056 - 0.333 186 0.129 0.1 0.132 0.056 - - 189 0.113 0.08 0.026 - - 0.167 192 0.081 0.08 0.026 0.111 - - 195 - 0.16 0.158 - - - 198 0.032 0.04 0.079 - - 0.167 201 - 0.04 - - - - 204 - - - 0.056 - - 207 - 0.04 - - - - 210 - 0.04 - - - - PclB11 163 - 0.03 - - - -

185 0.016 - - - - - 187 - - - - 0.125 - 189 0.047 - - 0.046 - - 191 0.078 - - 0.023 - - 193 0.641 - 0.071 0.091 - - 197 0.063 0.106 0.286 0.386 0.125 0.5 199 0.063 0.197 0.262 0.296 - - 201 - - 0.024 - - 0.167 203 0.016 0.121 0.191 0.136 0.25 0.333 205 0.031 0.106 0.048 - - - 209 0.016 0.091 0.095 0.023 - - 211 - 0.091 - - 0.125 - 215 0.031 0.152 - - 0.125 - 217 - 0.046 0.024 - - - 221 - 0.03 - - 0.125 - 223 - 0.03 - - 0.125 - PclC121 126 0.031 - - - - -

132 0.047 - - - - - 134 - - 0.024 - - - 136 0.016 - - - - - 138 0.016 - - - - - 140 0.016 - - - - - 142 0.031 - - - - - 146 0.031 - - - - -

166

APPENDIX Table 2.3. Continued.

Pleurobema Pleurobema Pleurobema Pleurobema

Pleurobema sintoxia/ sintoxia/ sintoxia/ sintoxia/ Locus Allele Pleurobema plenum rubrum rubrum rubrum rubrum cordatum Pool 4 MCNP Pool 4 Clinch Tennessee Pool 4

148 0.031 - - 0.023 - - 150 0.063 - - - - - 154 0.016 - - - - - 156 0.141 - - 0.068 - 0.333 158 0.031 - - - - - 160 0.031 - - 0.023 - - 162 0.078 - - - - - 164 0.031 0.015 0.048 - - - 166 0.047 - 0.024 0.046 0.125 - 168 0.063 - 0.071 0.227 - - 170 0.016 - 0.071 - - - 172 0.016 - 0.024 - - - 174 - - 0.048 0.023 - - 176 0.031 - 0.024 - - - 178 0.063 0.076 0.048 - - - 180 - - 0.071 0.023 0.25 - 182 0.016 - - - 0.125 - 184 0.016 - - - - - 186 0.016 0.015 0.071 0.023 - - 188 0.031 0.03 - 0.046 - - 190 - 0.015 0.024 0.046 - - 192 0.031 - - - 0.125 0.333 194 - 0.03 - - - - 196 - 0.03 - - - - 198 - 0.106 - 0.114 - 0.333 200 - 0.046 - - 0.125 - 202 - 0.046 0.071 0.046 - - 204 - 0.046 0.024 0.023 - - 206 - 0.106 - - - - 208 0.016 0.03 0.071 0.091 0.125 - 210 0.016 0.03 0.048 - - - 212 - 0.061 0.071 0.023 - - 214 - 0.061 - - - - 216 - 0.03 0.048 0.068 0.125 - 218 - 0.015 - - - - 220 - - - 0.091 - -

167

APPENDIX Table 2.3. Continued.

Pleurobema Pleurobema Pleurobema Pleurobema

Pleurobema sintoxia/ sintoxia/ sintoxia/ sintoxia/ Locus Allele Pleurobema plenum rubrum rubrum rubrum rubrum cordatum Pool 4 MCNP Pool 4 Clinch Tennessee Pool 4

222 0.016 0.046 - - - - 224 - 0.015 0.071 - - - 226 - 0.03 - - - - 230 - 0.015 - - - - 232 - 0.015 - - - - 238 - 0.015 - - - - 242 - 0.03 - - - - 248 - 0.015 - - - - 250 - 0.015 - - - - 262 - 0.015 - - - - 272 - - 0.048 - - - PclD104 88 0.063 - - - - -

92 0.063 0.015 - - - - 96 0.063 0.015 - - - - 100 0.094 0.046 0.024 - - - 104 0.109 0.015 - - - - 108 0.188 0.091 - - - - 112 0.156 0.03 0.024 - 0.375 - 116 0.109 0.091 0.071 - - - 120 0.078 0.091 0.071 0.119 0.125 - 124 0.047 0.046 0.071 0.048 - 0.167 128 - 0.03 0.048 0.024 - 0.167 132 - 0.106 0.071 0.048 0.125 0.333 136 - 0.106 0.048 0.024 - - 140 - 0.076 0.071 0.048 0.125 - 144 0.016 0.046 - - - - 148 0.016 0.061 - - - -

152 - 0.046 0.024 0.048 0.125 -

156 - 0.046 - - - -

160 - - 0.024 0.071 - -

164 - 0.015 0.024 - - -

168 - - 0.048 0.048 - -

172 - - 0.024 0.024 - -

176 - 0.03 - - - -

180 - - 0.024 0.024 - -

184 - - - 0.048 - -

168

APPENDIX Table 2.3. Continued.

Pleurobema Pleurobema Pleurobema Pleurobema Pleurobema Pleurobema sintoxia/ sintoxia/ sintoxia/ sintoxia/ Locus Allele cordatum plenum rubrum rubrum rubrum rubrum Pool 4 Pool 4 MCNP Pool 4 Clinch Tennessee

192 - - 0.024 0.024 - - 196 - - 0.024 0.024 - - 200 - - 0.048 0.048 - - 204 - - 0.024 0.048 - - 208 - - 0.071 0.143 - - 212 - - 0.048 0.095 - - 216 - - 0.024 - 0.125 - 220 - - - 0.048 - - 224 - - 0.048 - - - 228 - - 0.024 - - - 236 - - - - - 0.167 244 - - - - - 0.167 PclD106 148 - - 0.071 - - - 152 - - 0.024 0.048 - - 156 - - 0.024 0.024 - - 158 - - 0.048 - - - 160 - - - 0.024 - - 164 - - 0.048 0.024 - - 166 - 0.015 - - - 0.167 168 - - - 0.024 - - 170 0.016 0.03 - - 0.125 - 172 - - - 0.024 - 0.167 174 - 0.015 0.048 - - - 176 - - 0.071 0.048 - - 178 0.016 0.03 - - - - 180 - - - 0.048 0.125 - 182 - 0.03 0.024 0.024 - - 184 - 0.03 - 0.024 - 0.167 186 - 0.046 - - - - 188 - - 0.024 - - -

190 - 0.015 0.024 0.048 - - 192 - - - 0.024 - - 194 - 0.015 0.048 0.024 - - 196 - 0.03 - 0.024 - - 198 - 0.061 0.024 - - - 200 - - 0.024 0.024 0.125 -

169

APPENDIX Table 2.3. Continued.

202 - 0.046 0.024 - - - 204 - - 0.071 0.024 0.125 - 206 - 0.015 - 0.024 - - 208 0.016 - 0.071 0.071 0.125 0.333 210 - 0.046 0.048 - - - 212 - 0.03 - - - - 214 - - 0.024 0.024 - - 216 - 0.015 0.024 0.024 - - 218 - 0.03 - 0.095 - - 220 - - 0.024 0.024 - - 222 - 0.076 0.024 0.024 - - 224 0.016 0.015 - 0.071 - - 226 0.016 - - - 0.125 - 228 0.031 - - 0.024 0.125 - 230 0.047 0.106 - - - - 232 0.016 - - - - - 234 0.016 0.015 - - - - 236 - - 0.024 - - - 238 - 0.015 - - 0.125 - 240 0.047 - - - - - 242 0.172 0.03 - - - - 244 0.031 - - - - - 246 0.078 0.03 - 0.024 - - 250 0.016 0.046 0.024 0.024 - - 252 0.016 0.03 - - - - 254 0.016 - - - - - 256 - - - 0.024 - - 258 0.109 - - - - - 260 0.016 - - - - - 262 0.063 - - - - - 264 - - 0.024 0.024 - - 266 0.031 0.015 - - - - 268 - - - - - 0.167 270 0.016 - - - - - 272 - - 0.048 - - - 274 0.031 - - - - -

170

APPENDIX Table 2.3. Continued.

Pleurobema Pleurobema Pleurobema Pleurobema

Pleurobema sintoxia/ sintoxia/ sintoxia/ sintoxia/ Locus Allele Pleurobema plenum rubrum rubrum rubrum rubrum cordatum Pool 4 MCNP Pool 4 Clinch Tennessee Pool 4

276 - - 0.024 - - - 278 0.031 - - - - - 282 0.031 - - - - - 284 - - 0.024 - - - 286 0.047 - 0.024 0.024 - - 288 0.016 - - - - - 290 0.031 - - - - - 294 0.016 - - - - - 296 - 0.03 - - - - 298 - - - 0.024 - - 322 - 0.061 - - - - 326 - 0.015 - - - - 346 - 0.015 - - - - PclD9 173 - - 0.048 - - -

175 - - 0.071 - - - 177 0.019 - - - - 0.167 179 0.039 - 0.5 0.796 0.667 0.5 181 0.019 - - 0.046 - - 183 0.096 - 0.048 0.023 - 0.167 187 0.135 - 0.048 0.023 - 0.167 191 0.077 - - - - - 195 0.019 0.03 0.048 0.068 - - 199 - - 0.024 - 0.167 - 203 0.077 0.046 0.071 0.046 0.167 - 207 0.039 0.152 0.024 - - - 211 0.135 0.121 0.095 - - - 215 0.019 0.046 - - - - 219 0.096 0.182 - - - - 223 0.019 0.061 - - - - 227 - 0.121 - - - - 231 0.019 0.061 - - - - 235 - 0.061 0.024 - - - 239 0.019 0.091 - - - - 243 - 0.015 - - - - 251 0.039 0.015 - - - - 255 0.019 - - - - -

171

APPENDIX Table 2.3. Continued.

Pleurobema Pleurobema Pleurobema Pleurobema

Pleurobema sintoxia/ sintoxia/ sintoxia/ sintoxia/ Locus Allele Pleurobema plenum rubrum rubrum rubrum rubrum cordatum Pool 4 MCNP Pool 4 Clinch Tennessee Pool 4

259 0.058 - - - - - 263 0.019 - - - - - 271 0.039 - - - - - PpyD10 250 0.968 - - - - -

254 - 0.035 - - - - 258 - 0.035 0.033 - - - 266 0.016 - - 0.111 - - 270 - 0.017 0.167 - - - 274 - 0.121 0.167 0.056 0.25 - 278 - 0.017 0.1 0.083 - 1 282 0.016 0.069 0.1 0.111 - - 286 - 0.035 0.1 0.056 0.25 - 290 - 0.259 0.033 0.056 - - 294 - 0.086 0.033 0.25 0.25 - 298 - 0.069 0.033 0.083 - - 302 - 0.052 0.067 0.111 - - 306 - 0.086 - 0.028 0.25 - 310 - 0.086 - - - - 314 - 0.035 0.067 - - - 330 - - 0.033 - - - 334 - - 0.067 - - - 342 - - - 0.056 - - PpyC106 202 - - - 0.119 - -

206 - - - 0.048 - - 210 - - 0.026 0.024 - - 214 - - 0.053 0.024 0.375 0.167 218 0.018 - - 0.048 0.25 0.167 222 0.018 0.017 0.026 0.167 - 0.5 226 0.071 0.069 0.053 0.143 0.125 - 230 0.036 0.103 0.132 0.024 - - 234 0.018 0.086 0.053 0.024 0.125 - 238 0.089 0.121 0.053 0.048 - - 242 0.232 0.121 0.026 0.024 0.125 - 246 0.018 0.103 0.079 0.095 - - 250 0.036 0.121 - - - - 254 0.036 0.086 0.026 0.071 - -

172

APPENDIX Table 2.3. Continued.

258 0.125 0.035 0.184 0.048 - - 262 0.036 0.052 0.053 - - 0.167 266 0.036 - - 0.048 - - 270 0.018 - - - - - 274 0.036 - - - - - 278 0.036 - 0.105 0.024 - - 282 0.036 - - - - - 286 0.054 0.035 0.053 - - - 290 0.036 - 0.053 - - - 294 0.018 - - 0.024 - - 298 - - 0.026 - - - 318 - 0.035 - - - - 322 - 0.017 - - - - PpyD9 148 - - 0.025 - - -

156 0.016 - - - - - 160 0.016 - - - - - 164 0.078 - 0.025 - - - 168 0.094 0.03 - - - - 172 0.125 0.015 - - - - 176 0.016 - - - - - 180 0.063 0.061 - - - - 184 0.031 0.03 0.025 - - - 188 0.094 0.015 0.025 0.028 - - 192 0.078 0.046 0.025 0.167 0.125 - 196 0.031 0.076 0.075 0.056 - 0.167 200 - 0.03 0.025 0.028 0.125 - 204 0.047 0.076 0.05 - - - 208 0.016 0.015 - 0.028 - 0.167 212 - 0.197 0.025 - - - 216 0.016 0.076 - 0.028 0.125 - 220 0.016 0.046 0.1 0.028 0.375 - 224 - 0.106 0.075 0.028 - - 228 0.031 0.061 0.05 - - - 232 0.016 - 0.05 - - - 236 - 0.03 0.05 0.056 - 0.167 240 0.047 - 0.025 0.028 0.125 -

173

APPENDIX Table 2.3. Continued.

244 0.016 0.03 - 0.028 - - 248 0.016 - 0.025 0.028 - 0.167 252 - 0.015 0.025 0.111 - 0.333 256 - 0.015 0.075 0.028 - - 260 0.016 - 0.05 0.028 - - 264 0.031 - 0.05 0.056 0.125 - 268 0.047 - - 0.028 - - 272 0.031 0.015 - 0.083 - - 276 0.016 0.015 0.025 0.028 - - 280 - - 0.075 0.028 - - 284 - - - 0.028 - - 288 - - 0.025 0.056 - - PpyD101 138 0.016 - - - - -

150 - 0.046 0.088 - - - 154 0.031 - 0.059 - - - 158 0.047 - - - - - 162 0.094 0.046 0.029 0.056 - - 166 0.109 0.046 - 0.083 - - 170 0.094 - 0.029 - - - 174 0.109 - - - - - 178 0.172 - - 0.028 - - 182 0.063 0.015 - 0.056 0.25 - 186 0.063 - - 0.028 - - 190 0.031 0.061 0.088 - - - 194 0.047 - 0.029 0.056 - - 198 0.016 0.03 - - 0.125 - 202 0.031 0.061 0.088 0.083 - 0.333 206 0.016 0.015 - 0.028 0.125 - 210 - 0.061 - 0.083 - - 214 0.016 0.046 0.059 - - - 218 0.016 0.046 0.118 0.083 - - 222 - 0.015 0.059 0.028 - - 226 - 0.046 0.088 - - - 230 0.016 - 0.118 0.167 0.25 0.333 234 - 0.091 0.029 0.056 0.25 - 238 - 0.076 - 0.028 - -

174

APPENDIX Table 2.3. Continued.

242 0.016 0.061 - - - 0.333 246 - 0.015 - - - - 250 - 0.015 - - - - 254 - - 0.059 - - - 258 - 0.015 0.059 0.056 - - 262 - - - 0.028 - - 270 - 0.015 - - - - 274 - - - 0.056 - - 278 - 0.061 - - - - 282 - 0.015 - - - - 286 - 0.03 - - - - 290 - 0.03 - - - - 294 - 0.015 - - - - 306 - 0.015 - - - - 318 - 0.015 - - - -

175

Chapter 3 - Quantification of morphological variation and development of morphology- based keys to identify species of Fusconaia and Pleurobema in the Green River, Kentucky

ABSTRACT

This study quantified morphological variation among genetically identified mussel specimens of Fusconaia flava, F. subrotunda, Pleurobema cordatum, Pleurobema plenum and P. sintoxia/rubrum inhabiting the Green River, Kentucky. Because species in these genera are morphologically similar in shell phenotype and thus are difficult to identify, mussels of each species were first identified using molecular markers. Molecular identification then was compared with identifications made by experts to assess the reliability of phenotype-based identification of members of this group. On average, experts were able to correctly identify 70% of the mussel specimens collected from the Green River. In particular, I used field identification by the experts to try to separate the putative species P. rubrum and P. sintoxia, where identification by the experts was not consistent for either species. However, I was able to identify a few individuals that were consistently identified as either P. rubrum or P. sintoxia. Geometric morphometric analysis used landmarks on shells to analyze morphological differences among all investigated species, as well as among individuals within species in three size-class categories (small 20–60 mm; medium 60–

100 mm; and large > 100 mm mussels). Canonical Variate Analysis (CVA) was used to assess differences between mussel specimens in each size class within species. My results showed that the most difficult-to-identify species at each size-class category were mussel specimens belonging to Genus Pleurobema. The CVA showed that species in the Genus Fusconaia were well differentiated morphologically and hence much easier to identify. A multi-variable decision-tree analysis was conducted to determine the best suite of morphological variables for use to identify both live mussels and shells to the species level. The cross-validation error rates for these analyses

176 were 12.6% and 4.14% for live mussels and shells, respectively. The most important variables were presence of a sulcus and shell shape (e.g., trapezoidal, circular, oval, equilateral triangle, isosceles triangle), which were identified by random forest and decision-tree analyses.

Dichotomous keys for identifying both shells and living mussels were developed based on the morphological variables identified in the decision-tree and random-forest analyses. Key variables for identification of mussel species in the Green River included foot color, beak direction, and beak position to the anterior margin, as these variables are relatively easy to record and to identify mussels in the field. Ultimately though, identification of these species may still rely on molecular methods.

KEYWORDS: Freshwater mussels, Fusconaia, Pleurobema, Green River, Kentucky, shell morphology, decision trees, random forest, dichotomous keys

177

INTRODUCTION

Identification of species in the genera Fusconaia and Pleurobema typically is challenging for biologists because morphological characteristics of the shell and soft body parts can overlap and vary based on the age and size of the respective mussel specimen. Further, environmental factors such as stream size and productivity can influence size and shape of the shell (Ortmann

1920), making species identification even more difficult. In addition, morphological similarities among species in these genera could be the result of shared traits among closely related species.

Comparative morphological analysis is essential for understanding phylogenetic relationships among bivalves and for identifying characters useful for their identification. Shell morphology is important for describing mussel species and especially for identifying them in the field (Zieritz and Aldridge 2009). Shell morphology analyses include measurements such as hinge length, beak cavity depth, relative height, wet weight, and width. Morphological characters useful for identification of freshwater mussel species include shell shape and outline, shell color and ray pattern, and reproductive traits of the soft-body anatomy, such as number and color of gills used to brood larvae (glochidia).

Species can show considerable variation in their shell morphology, depending on the size and sex of the mussel specimen (sexual dimorphism) or environment (including clinal variation of the shell along the length of the stream) (Ortmann 1920; Inoue et al. 2013). Sexual dimorphism is absent in Fusconaia, while it is absent or weak in Pleurobema (Grabarkiewicz and Davis, 2008).

Analysis of the shell outline also is helpful for quantifying morphological variation among phylogenetic lineages, and has been conducted using different methods, including the Fourier method (Crampton and Maxwell 2000). Landmark positions can be determined using photographs of shells and a series of Fourier coefficients that are compared statistically like any other variable

178

(Zieritz and Aldridge 2009). In this study, I used Canonical Variate Analysis (CVA) to assess differences between landmarks in mussel specimens at different size classes and species. Principal component analysis (PCA) was used to quantify shell landmark variation in Fusconaia and

Pleurobema species of the Green River in Kentucky. A useful characteristic for the identification of the two genera is foot color, for example, which generally is orange for Fusconaia species and white for Pleurobema species. However, even this trait is known to be variable in some cases

(Schilling 2015).

For purposes of species identification in freshwater mussels, some of the most important reproductive traits are the shape of the conglutinate, and number of gills used as marsupia to brood eggs and glochidia (Haag and Warren 2003; Barnhart et al. 2008; Grabarkiewicz and Davis 2008).

Species in the genera Fusconaia and Pleurobema produce conglutinates containing eggs and glochidia, which can differ morphologically among species within these genera (Haag and Warren

2003; Barnhart et al. 2008; Grabarkiewicz and Davis 2008). These aggregations of eggs and glochidia are released into the water column, where they are eaten by host fish, infesting their gills in the process (Grabarkiewicz and Davis 2008). Conglutinates are slender and sub-cylindrical in the genera Fusconaia and broad and leaf-like with numerous egg layers in the genus Pleurobema.

In the case of F. flava, larval threads have been observed to help suspend conglutinates in the water column (Haag and Warren 2003; Barnhart et al. 2008). Another principal difference between these two genera is that species belonging to genus Fusconaia use all four gills as marsupia to brood eggs and glochidia, while those in Pleurobema only use the outer pair of gills. The number of gills that are used as marsupia seems to be homoplastic, with the ancestral condition being use of all four gills and the derived condition being the use of only the outer gills (Lydeard et al. 1996;

Lydeard et al. 2000).

179

Morphological characters that can be used in the development of a dichotomous key in these look-alike species can be obtained from the species descriptions (Cicerello and Schuster

2003; Pennsylvania Chapter American Fisheries Society 2018). In the case of P. plenum, the nacre is white, and the shell has a tall, triangular shape. Pleurobema cordatum is characterized by having a slightly twisted shell with white nacre, presence of a sulcus, and the beak is located at or behind the anterior margin of the shell, as in P. rubrum. However, P. rubrum can have either white or pink nacre. The principal difference between P. rubrum and P. cordatum is that P. rubrum often has the shape of a scalene triangle, while P. cordatum has the shape of an equilateral triangle. In individuals of P. sintoxia, there is a lack of sulcus, and they show a shallow or not-very-deep beak cavity, and nacre can be either white or pink. This species also is more rounded than F. flava, which has a more triangular shape, deeper beak cavity, prominent beaks that face each other, white nacre, and presence of a wide, shallow sulcus. Fusconaia subrotunda has a round-to-oval profile, especially when young and small, prominent anterior beaks, compressed and deep beak cavities, no sulcus, and the shell elongates with age. From these general descriptions, important characters of the shell (e.g., beak position, presence of sulcus, nacre and periostracum color) and soft-body anatomy (e.g., foot color and number of marsupial gills) can and have been used traditionally to identify mussel specimens in the field, and also can be used to construct a dichotomous key (Watter and Byrne, 2016). In addition, other characters such as foot color, hinge length, hinge width, umbo elevation, and periostracum color can be used to describe and identify species of interest (Schilling,

2015). In this study, I used Decision Tree analysis to build a tree with the most important variables for the identification of live mussel specimens and shells. Finally, I used Random Forest analysis to find the most important variables to use for identification of mussels to each respective species.

180

METHODS

Sample Collection

Individuals of F. flava, F. subrotunda, P. plenum, P. rubrum, P. sintoxia, and P. cordatum were collected from two main sites in the Green River, KY, including Pool 4 (GPS coordinates =

37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates

37.17819, -86.1154; river mile = 197) (Figure 3.1). The collected mussel specimens were tagged and kept alive at the Minor E. Clark Fish Hatchery (Morehead, KY, operated by the Kentucky

Department of Fish and Wildlife Resources) until morphological data could be collected to complete the morphological analyses. The numbers of individuals that were used for each analysis are shown in Table 3.1. For each species, mussel specimens from three size-classes were collected, including small (20–60 mm), medium (60–100 mm), and large (>100 mm) individuals. Mussel specimens used for analysis were genetically identified using mitochondrial DNA (mtDNA) sequences (see Chapter 1) before the characters were utilized for construction of dichotomous and photographic keys. Foot color was recorded as the principal soft-body character of the mussel. Soft parts were collected from selected individuals of F. flava, F. subrotunda, P. cordatum, P. sintoxia/rubrum and stored in 75% alcohol in 16-oz. containers. Because my mtDNA phylogenetic analysis conducted in Chapter 1 showed that P. sintoxia and P. rubrum were the same taxon, I used the field identification made by the experts to try to further separate these two shell forms.

Foot color for the endangered P. plenum was observed in the field or the hatchery without sacrificing the mussel. The soft part color, as well as shell characters were used to characterize each species morphologically in the decision-tree and random-forest analyses and to develop the identification keys. Individuals of F. flava, F. subrotunda, P. cordatum, and P. sintoxia/rubrum

181 were sacrificed in order to record additional categorical (nacre color) and quantitative (beak depth) variables.

Identification Validation

In October 2015, sampling in Pool 4 resulted in the collection of 209 mussel specimens, which were identified by five experts, Leroy Koch (Senior Biologist, Kentucky Ecological

Services Field Station – U.S. Fish and Wildlife Service), Dr. Wendell Haag (Fisheries Research

Biologist, U.S. Forest Service stationed at the Center for Mollusk Conservation, Kentucky

Department of Wildlife Resources), Chad Lewis (Malacologist, Lewis Environmental Consulting),

Dr. Monte McGregor (State Malacologist, Center for Mollusk Conservation, Kentucky

Department of Wildlife Resources), and Adam Shephard (Icthyologist, Center for Mollusk

Conservation, Kentucky Department of Wildlife Resources). The experts’ probability of correctly identifying a mussel as well as the identification error for each species from these two genera was calculated.

Geometric Morphometrics

Photographic images of the external shell were used for the geometric morphometric analysis, which was conducted by analyzing photographs of numerous specimens of F. flava, F. subrotunda, P. cordatum, P. plenum, and P. sintoxia/rubrum (Table 3.1). Mussel specimens were divided into three different size classes to assess possible differences among individuals of the same species, but from different size-classes.A digital semicircle divided by 15-degree increments was overlain onto each photograph. The anterior part of the umbo was aligned to landmark 1, while the end of the hinge ligament was aligned with landmark 3 (Figure 3.2). Landmarks were obtained

182 by using tpsDig2 v 2.25 (Rohlf 2001). For each species, the data were checked for outliers, and then a covariance matrix was used to perform a Principal Component Analysis (PCA). The PCA was used to assess variation in how the landmarks differed per size-class. Only principal component 1 was used to show geometric variation in each size-class. In addition, I performed a

Canonical Variate Analysis (CVA) to determine whether there were differences among sizes- classes within each species. This analysis also was performed to test whether there were differences among species size-classes. I used MorphoJ version 1.06 (Klingenberg 2011) to perform PCA and CVA for the respective landmark analyses.

Decision Trees and Random Forest

Decision trees analysis was performed using the party 1.3-3 (Hothorn et al. 2019) and rpart

4.1-15 (Therneau et al. 2015) packages implemented in R. I conducted a random forest analysis using randomForest (Liaw 2018) and caret (Kuhn 2012) packages implemented in R. Both analyses allowed me to assess the most appropriate morphological characters for the development of the dichotomous key. The dichotomous key was supplemented with original photographs.

A suite of shell characters used for the decision-tree and random-forest analyses included quantitative characters such as maximum length, hinge length, beak cavity depth, perpendicular height, width, and wet weight. These measurements were used to calculate the following morphometric ratios: (1) hinge length:maximum length, (2) beak cavity depth:maximum length,

(3) relative height (perpendicular height:maximum length), (4) shell obesity (width:maximum length), and (5) wet weight:maximum length. These characters were measured in millimeters or in grams using a caliper or a digital scale. These ratios were tested for normality by observing their distribution on a normal probability curve. Categorical characters – including beak direction, beak

183 position in relation to the anterior margin, shell shape, sulcus presence, nacre color, and foot color

– were used in the decision-tree analysis. For the analysis of live mussel specimens, I included all these variables but without nacre color and the ratio of beak depth:maximum length. For shell analysis, I did not use foot color and wet weight:maximum length ratio. The number of mussel specimens used for these analyses are reported in Table 3.1. Shell measurements and morphological characters that were recorded are shown in Figures 3.3 and 3.4. In order to record data for the beak position I aligned the hinge parallel to a horizontal line. The sulcus was considered broad if it represented at least 2/3 of the shell (Figure 3.4).

I constructed two decision trees, one for classification of live mussels and another for shells. The study species (F. flava, F. subrotunda, P. cordatum, P. plenum, and P. sintoxia/rubrum) was the response variable in the decision trees. For both the live mussel specimen and shell decision trees, the data were partitioned as 80% training data set (119 observations) and

20% validating data set (23 observations). In order to trim the tree, the confidence level was set to

90% because the numbers of mussel specimens per species and for the three different size-classes were not large. I set up the tree analysis so that a branch would split into two if the number of mussel specimens was at least 1, and K-fold cross-validation with K=10 was performed in order to validate the model.

I performed two random-forest tree analyses, one for live mussel specimens and another for shells. For random forest analysis, I used 70% (100 mussel specimens) of the data as a training dataset and 30% (42 mussel specimens) of the data as a validating dataset. The number of trees to be analyzed was determined after plotting the out-of-bag (OOB) error and observing how it declined and then stabilized. For live mussel specimens, I analyzed 900 trees while for shells I analyzed 200 trees.

184

Dichotomous and Photographic Keys

The dichotomous key was developed using the most important variables identified by decision-tree and random-forest analyses. In addition, I added characters that are easier to observe in the field and were less subjective.The mussels were photographed using a Pentax K-50 Digital

SRL camera inside a lightbox with a color wheel to quantify the true color of the shell and a caliper as a measurement gauge. Photographs of each species and individuals within species were used to document morphological variation and for use in the decision-tree and random-forest analyses, and to help to develop the dichotomous key. In addition, photographs were organized for mussel specimens with the small, medium, and large size-classes to compare morphological differences within and among species and to develop a photographic key.

RESULTS

Identification of Shells by Experts

Mussel specimens collected from Pool 4 and MCNP and held at the Minor Clark Fish

Hatchery were identified by five mussel experts to their respective nominal species. Afterwards, all mussel specimens were identified to the species level using the mitochondrial COI and ND1 genes (see Chapter 1), which revealed some misidentifications by the experts. On average, experts correctly identified a mussel 70% of the time (Figure 3.5 and Table 3.2). The identification errors are shown in table 3.3.

Of my samples, F. flava was correctly identified by the experts’ 95% of the time (Table

3.2; Figure 3.5). These results support the current understanding based on field survey data that only two species in the genus Fusconaia occur in the Green River. While species in this genus typically have an orange-colored foot, for the F. flava samples that I analyzed in the Green River,

185 the foot can be either orange or white. This observation helps explains why some F. flava mussel specimens were erroneously identified as Pleurobema species, which typically have a white foot color, for example P. plenum or P. sintoxia/rubrum.

My data also showed that F. subrotunda (Classification error = 22%) often was erroneously identified as either F. flava or P. sintoxia/rubrum. Schilling (2015) showed that F. subrotunda can have a white to orange foot color. However, his specimens were collected in rivers throughout the upper Tennessee River basin, primarily from the Clinch River in Tennessee. In my study, all F. subrotunda mussel specimens that I collected in the Green River in Kentucky had an orange foot.

However, the Pleurobema species were more difficult to identify by the experts than the two

Fusconaia species. The classification errors for the Pleurobema species were as follows: P. cordatum = 41%, for P. plenum = 35%, and P. sintoxia/rubrum = 20% (Table 3.3). These three species all had a white-colored foot, meaning that their identification mostly relied on morphological characters of the shell.

Because the molecular identification resulted in low to not enough genetic differentiation between P. rubrum and P. sintoxia (Chapter 1), I used the experts’ identification to try to further separate mussel specimens belonging to this group. The experts identified mussel specimens in this group as either F. subrotunda, P. cordatum, P. plenum, P. sintoxia, and P. rubrum (Figure

3.6). However, a total of 74% of the identifications made in this genetically identified group were identified as P. sintoxia. The individuals in this group were mostly misidentified as P. plenum

(13% of the identifications). Experts identified these mussel specimens as P. rubrum only 7% of the time and only one individual was consistently identified as P. rubrum by the five experts. This mussel specimen separated into an additional clade when its COI and ND1 haplotypes

(COI_PSR14_and_ND1_PSR14) were combined together. However, the DNA sequence

186 divergence was still very low and not enough to consider it a different species as stated in Chapter

1. The shell of this individual had a “shallow and broad” sulcus, the beaks faced forward and passed the anterior margin of the shell, the shape was triangular (isosceles), its foot was white and the nacre (which was not observed at the time of field identification) was pink. These characteristics are typical of P. rubrum (Figure 3.7A) as traditionally understood by the experts.

However, the high error rate of the experts and only one representative mussel specimen was not enough to create an additional group for the morphometric, decision trees, and random forest analyses. Hence, I grouped all of the genetically identified specimens into a single P. sintoxia/rubrum group. In addition, I have morphometric data for 6 individuals that were identified as P. sintoxia by the experts. These mussel specimens showed a rounded triangular shell (most of the time equilateral), absence of sulcus, the beaks faced each other and were not close to the anterior margin, the foot was white and the nacre (which was not observed at the time of field identification) was either white or pink (Figure 3.7B).

Geometric Morphometrics

The results of CVA showed differentiation among shell size-classes for F. flava, F. subrotunda, P. cordatum, and P. sintoxia/rubrum; however, many of the observations were outside of the 0.9 confidence interval, which suggests that differentiation was non-significant among the three sizes classes for these species. Morphological separation was more defined among size classes for F. subrotunda, and a similar but weaker level of differentiation occurred among the size classes of F. flava and P. sintoxia/rubrum. For P. cordatum, there was no well-defined differentiation among mussel specimens between the large and medium size classes (Figure 3.8).

The CVA of the different shell size-classes showed that F. flava and F. subrotunda were distinct

187 at all three size-classes. However, mussel specimens belonging to P. cordatum, P. plenum, and P. sintoxia/rubrum did not show strong differentiation among species for the respective size-classes

(Figure 3.9). Finally, I observed significant variation in each shell-landmark position from the mean for each species’ size classes. It is important to note that variation at each shell landmark can occur independently of variation at other landmarks (Figure 3.10). Deviation from average shell landmarks was later used to select mussel specimens that had landmark positions representative of the “average” landmark positions for the species as a whole to develop the photographic key.

Decision Trees

As the data distribution of the morphological variables had normal bell-shaped curves, I did not proceed to transform the data before running the decision trees (Figure 3.11). In the decision tree for live mussel specimens, the cross-validation error rate was 12.61%, and the misclassification error rate using the training data set was 10.92%, while the error rate with the validating data set was 8.70% (Table 3.4). In the decision tree for shell classification, the cross- validation error rate was 4.14%, and the misclassification error rate with the training data set was

4.20%, while the error rate with validating data set was 0% (Table 3.5). Both decision trees show the probability of correctly identifying a species by following the different branches. The decision tree for live mussel specimens showed that the most important variables for identifying the species were sulcus presence, beak direction, and shell shape (Figure 3.12). The decision tree for shells showed that the most important variables for identification of these species were sulcus presence, shell shape, nacre color, shell width and beak position (Figure 3.13). Misclassification rates for live mussel specimens and for shells in the three Pleurobema species were higher than between the two Fusconiaia species. Importantly, occasional misclassifications of live mussel specimens

188 or shells to the wrong species will occur even if the user of either key correctly distinguishes the mussels’ characters.

Random-Forest Analyses

For live mussel specimens, the K-fold validation with K=10 resulted in an error rate of

3.5%. The number of variables that were tested at each split was set at 2, as this was the optimal value with respect to the out-of-bag (OOB) error rate estimate. The OOB error rate estimate was

6%, meaning that the model has about 94% accuracy in identifying live mussel specimens to the correct species. Classification errors occurred mostly for P. cordatum (6.5%), P. plenum (13.3%) and P. sintoxia/rubrum (7.7%) (Table 3.6). This analysis also allowed me to identify which morphological variables were most important for making accurate species identifications. Species predictions made using the validating data set resulted in an accuracy of 1 (95% CI = 0.96 -1), meaning that all live mussel specimens were correctly classified. On the other hand, predictions using the testing data set had an accuracy of 0.90 (95% CI = 0.77 – 0.97). The mean decrease in accuracy and the mean decrease in the Gini graphs (Figure 3.14) showed that the most important morphological variables for identification of these species were sulcus presence followed by beak direction and shape.

For shells, the K-fold validation with K=10 resulted in an error rate of 2.8%. The number of variables that were tested at each split was set at 3, which was the optimal value with respect to the OOB error rate estimate. The OOB error rate estimate was 3%, meaning that the model had approximately a 97% accuracy for identification of shells to the correct species. However, classification errors occurred for P. plenum (6.7%) and P. sintoxia/rubrum (7.7%), with the model predicting that both of these species can be confused with P. cordatum (Table 3.6). This analysis

189 also allowed me to identify which morphological variables were most important for classification of shells to the correct species identification. Model predictions using the validating data set resulted in an accuracy of 1 (95% CI = 0.96 -1), meaning that all shells were classified to the correct species. On the other hand, predictions using the testing data set had an accuracy of 0.90

(95% CI = 0.77 – 0.97). The mean decrease in accuracy and the mean decrease in the Gini graphs

(Figure 3.14) showed that the most important variable for identification of shells to their respective species were sulcus presence followed by shape, nacre color, and beak direction.

Dichotomous Key

The dichotomous key was developed using the most important morphological variables identified in the decision-tree analysis. The variables identified by the decision-tree analysis were the same as the most important variables identified by the random-forest analysis. I provide a table summarizing the most important categorical and quantitative variables for identifying each species

(Table 3.7), graphs showing the proportion of expression of these characters in each species

(Figure 3.15), and box plots of the quantitative variables (Figure 3.16). Collectively, this information allowed me to develop two separate dichotomous keys for identification of live mussel specimens and shells. In both dichotomous keys, I utilized the variables that were easiest to identify and use in both field and hatchery settings. For the dichotomous key for the identification of live mussel specimens, the most important character was foot color (KEY A), a character that is easy to identify in the field. Generally, the foot color for mussel specimens of the genus Fusconaia is orange and for mussel specimens of the genus Pleurobema it is white. However, for the Green

River, I also recorded mussel specimens of F. flava with a white foot. A character that was not used in the decision tree was beak position to the anterior margin, but this character was useful to

190 separate mussel specimens of P. plenum and P. sintoxia/rubrum with “beak extending beyond the anterior margin” and “beak does not extend beyond to the anterior” margin, respectively (Figure

3.4). In the case of the dichotomous key for shell identification (KEY B), shape was the most important character to separate F. subrotunda (circular or oval) from the other species, which typically had a more triangular (equilateral and isosceles) or trapezoidal shape. Since it is easy to confuse trapezoidal shape with triangular shape, I used beak direction to separate F. flava and P. sintoxia/rubrum (beaks face each other) from P. cordatum, P. plenum and P. sintoxia/rubrum

(beaks face forward). As was shown by the geometric morphometric analysis, mussel specimens of the genus Pleurobema seem more difficult to distinguish from each other. I used sulcus presence and beak position relative to the anterior margin to separate these species. It is important to emphasize that these dichotomous keys are useful for the identification of mussel specimens in the

Green River, Kentucky and may not apply to other river systems. Finally, errors may be the result of subjectivity of some of the characters. The use of the dichotomous keys can be supplemented with use of the photographic keys (KEY C as well as Table 3.7 and Figures 3.15 and 3.16), which shows and summarizes visually the proportions of mussel specimens with certain characters in each species.

Shell Photographic Key

Photographs in this study included one individual that show the different parts of a mussel shell, including the measurements that were taken, such as maximum length and perpendicular height (Figure 3.3). As mussel shape proved to be most important character for mussel identification, additional mussel photographs were arranged to show the respective shapes

(circular, equilateral, isosceles, oval, and trapezoidal) that were used to describe the mussels. In

191 addition, I took photographs showing sulcus characters (absent, broad and deep, broad and shallow, narrow and deep, and narrow and shallow). Moreover, photographs for all recorded characters also were included (Figure 3.4). These characters included nacre color (pink and white), foot color (orange and white), position of the beak regarding the anterior margin, and beak direction. Finally, the photographic key included mussel specimens of different sizes for each species that showed the least divergence from the mean shell landmarks (KEY C).

Species Description

The following species descriptions were developed using the most important variables for the description of each species as assessed in the decision-tree and the random-forest analyses, with categorical and quantitative variables of each species summarized in Table 3.7 and in Figures

3.12 and 3.13. Fusconaia flava has a trapezoidal shell shape with the beaks facing each other.

However, the beaks were eroded and destroyed in many of the mussel specimens and their direction was not always easy to determine. Beaks are not close to the anterior margin. Foot color generally is orange; however, sometimes it was white. Nacre color was white most of time but occasionally it is pink. The only other species showing pink nacre was P. sintoxia/rubrum. Finally, the sulcus was broad and deep (KEY C. 1). Fusconaia subrotunda has a circular shell shape that becomes oval as the mussel grows larger into an adult. The foot color is orange, the beaks face forward, and nacre is white, and the shell does not have a sulcus (KEY C. 2). When Pleurobema cordatum individuals are small they have an equilateral shape; however, the shell becomes isosceles triangular in shape as it grows larger into an adult. Foot and nacre are white, the beaks usually face forward, but sometimes they can face each other, and the sulcus is usually shallow and narrow but distinct (KEY C. 3). Pleurobema plenum generally has a shell that is equilateral

192 triangular in shape. However, the shell can occasionally be isosceles triangular in shape. The beaks face forward at each other, which can make P. plenum easy to confuse with P. cordatum. However, the beaks usually pass the anterior margin, which is not common in P. cordatum or P. sintoxia/rubrum. Foot and nacre color are white (KEY C. 4). Pleurobema sintoxia/rubrum: species can be equilateral or isosceles triangular for older mussels. Beak cavity is shallow in comparison to the other species. Foot color is white and nacre can be white or pink. Beaks are not close to the anterior margin. Beaks can either face forward or face each other (KEY C. 5). However, the description of this group is challenging as the molecular markers were not able to separate these two species (Chapter 1). Extreme forms of the shells were consistently identified by the experts as either P. rubrum or P. sintoxia. The only mussel specimen identified as P. rubrum by all the experts had a triangular shape (isosceles), a sulcus (shallow and broad), the beaks passed the anterior margin and faced forward, the foot color was white and the nacre was pink (Figure 3.7A). In addition, the six mussel specimens consistently identified as P. sintoxia showed a more equilateral shape with rounded edges, the beaks faced forward but they did not pass the anterior margin, the foot color was white and the nacre was either white or pink (Figure 3.7B).

DISCUSSION

Expert identification

Misidentification frequencies for the mussels considered in this study was between 5-41%.

The easiest mussels to identify were Fusconaia species because they commonly have an orange foot. The harder identification problem arises with species in the genus Pleurobema.

Morphometric analyses revealed that Pleurobema species are difficult to tell apart, which led to a high rate of classification error by the experts (20-41%). These misidentifications are due to

193 various factors but fundamentally stem from the close morphological similarity of the shells and foot color among the Pleurobema species. In addition, the similarity of these traits is likely due to each species’ close genetic relatedness and the phenotypic plasticity common among species in these two genera (Inoue et al. 2013).

Geometric Morphometrics

My study results showed that the mussel specimens’ shell landmarks changed with size, as they grew from small to large sized indivduals. The size and shape of mussels also can be affected by river nutrient richness. For example, mussels with smaller and thinner shells may come from rivers that have lower amounts of key shell-building nutrients such as calcium and bicarbonate, which can compromise growth (Haag and Rypel 2011). Other factors possibly affecting shell size and mass are river size (depth and width), sediment type, stream flow variation, wind and current exposure, and food availability (Haag 2012). The results showed that the most difficult mussels to identify were those belonging to the genus Pleurobema; similar shapes among these species could be the result of evolutionary convergence upon consensus shapes, or traits shared because of a common ancestor.

On the other hand, species from the genus Fusconaia (F. flava and F. subrotunda) had very different shell shapes and soft-body color from each other. Foot color is the most distinctive character consistently differentiating species of Pleurobema from species of Fusconaia. However, in some cases, F. flava foot color can be white. Differences among collected mussels of F. flava could be the result of Ortmann’s (1920) observation that stream position causes clinal variation of shell shape within a species, resulting in inflated shells in large streams and compressed shells in

194 small streams (Haag 2012). The principal limitation to test Ortmann’s law in the Green River system was my small sample sizes, and their limited geographic range within the system.

Decision-Tree and Random-Forest Analyses

These analyses allowed me to simultaneously use categorical and continuous variables to find differences among the study species. Characters that were used in the descriptions of these species and that were easy to identify in the field were recorded and used for analyses. The most important morphological traits influencing tree accuracy and the Gini values were mussel shape and sulcus presence. However, field assessment and quantifications can vary depending on the judgment of the person recording the data. Someone who is very experienced in mussel inditification may result in more accurate in the scoring of categorical variables and hence in correctly identifying a species. While white nacre color was most frequently recorded for the study specimens among each respective species, pink nacre was not infrequent and has been hypothesized to be the result of staining by chemical elements in the water (Rosenberg and

Henschen 1986). However, in Venustaconcha trabalis and V. troostensis, the nacre color has traditionally been an important character to tell apart these species, and recent hatchery-based evidence suggest a genetic basis for this color trait in these two species (Parmalee and Bogan 1998,

Lane et al. 2019). In addition, P. sintoxia has been described as having polymorphic nacre color, which can be either white, pink, or orange (Haag 2012). In this study, nacre color was found to be either pink or white for P. sintoxia/rubrum and F. flava, whereas individuals of F. subrotunda, P. cordatum, and P. plenum only had white nacre.

195

Principal Issues in Mussel Identification

In addition, umbos of many mussel specimens in the medium and large size-classes were at times eroded. Which if severe can affect shell shape designation as well as positions of the landmarks. A character that was not taken into consideration was sex, as sexual dimorphism has been described as absent in species belonging to the tribe Pleurobemini (Haag and Rypel 2011).

One character that did not exhibit much variation among the study species was color of the periostracum. However, smaller specimens sometimes had a lighter color than the larger mussels, but shell color ranged between light brown to almost black among these species. In addition, foot color is an easy character to identify and is useful for separating Fusconaia and Pleurobema species in the Green River, KY. My results indicated that mussels of F. subrotunda in the Green

River have an orange-colored foot, and so mussels of this species should not be erroneously identified as any of the Pleurobema species in the Green River. On the other hand, mussel specimens of F. flava had either a white or orange foot. Hence, white-footed individuals of F. flava in the Green River can be erroneously identified as a Pleurobema species.

Pyramid and Round Pigtoes

The morphological assessment of P. sintoxia and P. rubrum was particularly difficult as the molecular identification did not separate individuals into distinct groups or clades. However, extreme forms of these shells resulted in consistent identification of the P. rubrum and P. sintoxia shell forms by the experts. The recorded morphological characters matched the description from other studies for the Green River and other watersheds (Cicerello and Schuster 2003, Pennsylvania

Chapter American Fisheries Society 2018) and three of the most important characters to differentiate these two shell forms is the presence (P. rubrum) or absence (P. sintoxia) of a sulcus,

196 the isosceles triangle shape (P. rubrum) or equilateral triangle shape with rounded edges (P. sintoxia), and the beaks passing (P. rubrum) or not passing (P. sintoxia) the anterior margin.

However, the lack of genetic differentiation and the limited number of mussel specimens available in this study for the P. rubrum shell form was not enough data to make strong conclusions about these two shell forms or even to morphologically characterize the different size-classes of these two shell forms.

Management Implications

Species of freshwater mussels belonging to the genera Fusconaia and Pleurobema are particularly difficult to identify even for the experts. Decision trees and random forest analysis identify the most important variables to identify these mussels. However, field researchers must be familiar with the different shell shape categories in order to be able to correctly identify mussels of these species. A mussel identification workshop would help test the dichotomous keys before and after mussel biologist have been trained. It would be interesting to see if the expert’s identification scores would increase after the workshop. Thus, the proposed dichotomous keys still need to be tested by experts. However, the high morphological variability of the mussel shapes as well as the shared traits among these species will likely result in a high error rate even among experts. In many cases then of management importance, the correct identification of these species should be corroborated by use of molecular markers.

The use of phylogenomics to assess the species status of P. sintoxia and P. rubrum is recommended to inform future conservation plans. However, even if these two nominal taxa were shown to be separate species, the use of molecular markers would still be needed to identify them, as most individuals belonging to species of Pleurobema were very similar in their morphometric

197 analysis. In addition, the assessment of shell variation depending on their position in a stream is also important to assess whether shell shapes differ upstream versus downstream. The possibility of hybridization among these putative species should be tested using nuclear molecular marker to determine whether intermediate shell forms (which are not consistently identified as P. rubrum or

P. sintoxia by the experts) are the result of hybridization.

As these species may appear together in other watersheds, it is important to clarify that the dichotomous keys should only be used for mussels in the Green River, KY. Mussels in other watersheds even of the same species may show other characters and/or shapes that was not considered in this study. Hence, a dichotomous key should be developed for each watershed or state as needed.

198

COLLABORATORS CONTRIBUTIONS

Hyde helped with DNA collection, mussel tagging, and lab work. Field identification of the species were performed by Leroy Koch, Dr. Wendell Haag, Chad Lewis, Dr. Monte McGregor, and Adam

Shephard. Insightful input for the methods and discussion was provided by Eric Hallerman, Jess

Jones, Emmanuel Frimpong and Pawel Michalak. I aided in the DNA collection, mussel tagging, lab work, data analysis and prepared the manuscript with assistance from Eric Hallerman and Jess

Jones.

199

LITERATURE CITED

Pennsylvania Chapter American Fisheries Society. 2018. A Field Guide to Pennsylvania’s Freshwater Mussels. Available online: https://pa.fisheries.org/wp- content/uploads/2018/02/Mussel-ID-workshop-field-guide-2-9-18.pdf. (accessed on 23 October 2019).

Barnhart, M. C., W. R. Haag, and W. N. Roston. 2008. Adaptations to host infection and larval parasitism in Unionoida. Journal of the North American Benthological Society 27: 370-394.

Cicerello, R. R., and Schuster, G. A. 2003. A guide to the freshwater mussels of Kentucky. Kentucky State Nature Preserves Commission.

Crampton, J. S., and P. A. Maxwell. 2000. Size: all it’s shaped up to be? Evolution of shape through the lifespan of the Cenozoic bivalve Spissatella (Crassatellidae). Geological Society, London, Special Publications 177: 399-423.

Haag, W. R. 2012. North American Freshwater Mussels: Natural History, Ecology, and Conservation. Cambridge University Press.

Haag, W. R., and A. L. Rypel. 2011. Growth and longevity in freshwater mussels: evolutionary and conservation implications. Biological Reviews 86: 225-247.

Haag, W. R., and M. L. Warren. 2003. Host fishes and infection strategies of freshwater mussels in large Mobile Basin streams, USA. Journal of the North American Benthological Society 22: 78–91.

Hothorn, T., A. Zeileis, and K. Hornik. 2019. Package ‘party’. Available online: cran.r- project.org/web/packages/party/party.pdf (accessed on 23 October 2019).

Inoue, K., D. M. Hayes, J. L. Harris, and A. D. Christian. 2013. Phylogenetic and morphometric analyses reveal ecophenotypic plasticity in freshwater mussels Obovaria jacksoniana and Villosa arkansasensis (Bivalvia: Unionidae). Ecology and Evolution 3: 2670-2683.

Klingenberg, C. P. 2011. MorphoJ: an integrated software package for geometric morphometrics. Molecular Ecology Resources 11: 353-357.

Kuhn, M. 2012. The caret package. Available online: cran.r-project.org/package=caret. (accessed on 28 October 2019).

Lane, T. W., E. M. Hallerman, and J. W. Jones. 2019. Population genetic assessment of two critically endangered freshwater mussel species, Tennessee bean Venustaconcha trabalis and Cumberland bean Venustaconcha troostensis. Conservation Genetics 20: 759-779.

200

Liaw, A. 2018. Package ‘randomForest’. Available online: cran.r- project.org/web/packages/randomForest/randomForest.pdf (accessed on 28 October 2019).

Lydeard, C., Mulvey, M., and G. M. Davis. 1996. Molecular systematics and evolution of reproductive traits of North American freshwater unionacean mussels (Mollusca: Bivalvia) as inferred from 16S rRNA gene sequences. Philosophical Transactions of the Royal Society of London B: Biological Sciences 351: 1593-1603.

Lydeard, C., R. L. Minton, and J. D. Williams. 2000. Prodigious polyphyly in imperiled freshwater pearly-mussels (Bivalvia: Unionidae): a phylogenetic test of species and generic designations. Geological Society, London, Special Publications, 177: 145-158.

Ortmann, A. E. 1920. Correlation of shape and station in fresh-water mussels (Naiades). Proceedings of the American Philosophical Society 59: 269-312.

Parmalee, P.W., and Bogan, A.E. 1998. The freshwater mussels of Tennessee. University of Tennessee Press, Knoxville

Rohlf, F. J. 2001. TPSDig2: a program for landmark development and analysis. http://life. bio. sunysb. edu/morph/index. Html.

Rosenberg, G. D., and Henschen, M. T. 1986. Sediment particles as a cause of nacre staining in the freshwater mussel, Amblema plicata (Say) (Bivalvia: Unionidae). Hydrobiologia, 135, 167- 178.

Therneau, T., B. Atkinson, B. Ripley, and Ripley, M. B. 2015. Package ‘rpart’. Available online: cran.r-project.org/web/packages/rpart/rpart.pdf (accessed on 23 October 2019).

Watter, G. T. and C. Byrne. 2016. Freshwater mussel identification workshop. Museum of Biological Diversity. The Ohio State University, Columbus, Ohio.

Zieritz, A., and D. C. Aldridge. 2009. Identification of ecophenotypic trends within three European freshwater mussel species (Bivalvia: Unionoida) using traditional and modern morphometric techniques. Biological Journal of the Linnaean Society 98: 814-825

201

Decision trees and random forest analyses

Categorical variables Quantitative variables

Beak Geometric Shell position to Beak depth : Hinge length : Wet weight : morphometric size anterior Beak Foot Nacre Sulcus Maximum Maximum Relative Shell Maximum analysis classes margin direction color color Shape presence length length height obesity length small 12 8 8 8 8 8 8 8 8 8 8 1 medium 26 19 19 19 19 19 19 19 19 19 19 12 Fusconaia flava large 1 0 0 0 0 0 0 0 0 0 0 0 total 39 27 27 27 27 27 27 27 27 27 27 13 small 2 2 2 2 2 2 2 2 2 2 2 2 Fusconaia medium 7 7 7 7 7 7 7 7 7 7 7 0 subrotunda large 12 6 6 6 6 6 6 6 6 6 6 0 total 21 15 15 15 15 15 15 15 15 15 15 2 small 19 7 6 7 7 7 7 7 7 7 7 4 Pleurobema medium 76 29 28 29 29 29 29 29 29 29 29 7 cordatum large 18 6 6 6 6 6 6 6 6 6 6 0 total 113 42 40 42 42 42 42 42 42 42 42 11 small 2 2 0 2 0 2 2 0 0 2 2 0 Pleurobema medium 30 17 5 17 0 17 17 0 0 17 17 0 plenum total 32 19 5 19 0 19 19 0 0 19 19 0 small 22 22 22 22 22 22 22 22 22 22 22 13 Pleurobema medium 17 15 15 15 15 15 15 15 15 15 15 7 sintoxia/rubrum large 2 2 2 2 2 2 2 2 2 2 2 1 total 41 39 39 39 39 39 39 39 39 39 39 21

202

Table 3.2. Probability of five experts correctly identifying mussel species in the genera Fusconaia and Pleurobema collected in 2015 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) in the Green River, Kentucky. All mussels were held at the Minor E. Clark Fish Hatchery, Morehead, KY and were subsequently identified to species genetically using the mtDNA gene ND1.

Pleurobema sintoxia/ Fusconaia flava Fusconaia subrotunda Pleurobema cordatum Pleurobema plenum All species Pleurobema rubrum Expert 1 0.97 1 0.37 0.73 0.69 0.65 Expert 2 0.89 0.33 0.4 0.8 0.81 0.57 Expert 3 1 0.78 0.82 0.73 0.81 0.83 Expert 4 0.95 0.92 0.67 0.22 0.74 0.66 Expert 5 0.96 1 0.71 0.84 0.8 0.8 N 5 5 5 5 5 5 Average 0.95 0.81 0.59 0.67 0.77 0.7 SD 0.04 0.28 0.2 0.25 0.06 0.11 CI (95%) 0.04 0.25 0.17 0.22 0.05 0.1 CI (99%) 0.05 0.32 0.23 0.29 0.06 0.12

203

Table 3.3. Confusion matrix showing the classification error for the five experts for mussel species in the genera Fusconaia and Pleurobema collected in 2015 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) in the Green River, Kentucky. All mussels were held at the Minor E. Clark Fish Hatchery, Morehead, KY and were subsequently identified to species genetically using the mtDNA gene ND1. “-” indicates no value recorded for a given pairwise combination.

Experts Identification

Fusconaia Fusconaia Pleurobema Pleurobema Pleurobema Classification flava subrotunda cordatum plenum sintoxia/rubrum error

Fusconaia flava 0.95 0.01 - 0.02 0.02 5%

Fusconaia subrotunda 0.01 0.78 - - 0.20 22% Pleurobema cordatum 0.01 0.01 0.59 0.11 0.27 41%

Pleurobema plenum 0.02 - 0.04 0.65 0.29 35% Molecular

identification Pleurobema - 0.04 0.01 0.14 0.80 20% sintoxia/rubrum

204

Table 3.4. Confusion matrix obtained from the decision tree analysis for identification of live mussel specimens of Fusconaia and Pleurobema species in the Green River, Kentucky. Data validations included Hold-Out validation which splits the data into training (80%, 119 observations) and validation data (20%, 23 observations). The confusion matrix shows predicted and actual identifications made to show the misclassification for each species. Misclassification for training data was 10.92% and for validation data it was 8.70%. “-” indicates no value recorded for a given pairwise combination.

Actual Fusconaia Fusconaia Pleurobema Pleurobema Pleurobema

flava subrotunda cordatum plenum sintoxia/rubrum

Fusconaia flava 25 - - - - Fusconaia subrotunda - 12 - - - Pleurobema cordatum - - 32 - 3

Pleurobema plenum - - 2 7 - Trainingdata Pleurobema sintoxia/rubrum - - - 8 30

Fusconaia flava 2 - - - - Prediction Fusconaia subrotunda - 3 - - - Pleurobema cordatum - - 8 - - Pleurobema plenum - - - 2 -

Validation data Pleurobema sintoxia/rubrum - - - 2 6

205

Table 3.5. Confusion matrix obtained from the decision tree analysis for the identification of shells of species belonging to the genera Fusconaia and Pleurobema in the Green River, Kentucky. Data validations included Hold-Out validation which splits the data into training (80%, 119 observations) and validation data (20%, 23 observations). The confusion matrix shows predicted and actual identification to show the misclassifications for each species. Misclassification for the training data was 4.20% and for validation data it was 0%. “-” indicates no value recorded for a given pairwise combination.

Actual Fusconaia Fusconaia Pleurobema Pleurobema Pleurobema

flava subrotunda cordatum plenum sintoxia/rubrum

Fusconaia flava 25 - - - - Fusconaia subrotunda - 12 - - - Pleurobema cordatum - - 34 - 3

Pleurobema plenum - - - 14 1 Trainingdata Pleurobema sintoxia/rubrum - - - 1 29

Fusconaia flava 2 - - - - Prediction Fusconaia subrotunda - 3 - - - Pleurobema cordatum - - 8 - - Pleurobema plenum - - - 4 -

Validation data Pleurobema sintoxia/rubrum - - - - 6

206

Table 3.6. Confusion matrix for random forest analysis of live mussel specimens and shells of species belonging to the genera Fusconaia and Pleurobema collected from the Green River, Kentucky. The confusion matrix presents the predicted and actual identifications to show the misclassifications for each species. Classification error for each species shows the percentage of misclassified mussel specimens. “-” indicates no value recorded for a given pairwise combination.

Predicted Fusconaia Fusconaia Pleurobema Pleurobema Pleurobema Classification

flava subrotunda cordatum plenum sintoxia/rubrum error

Fusconaia flava 17 - - - - 0

Fusconaia subrotunda - 11 - - - 0 Pleurobema cordatum - - 29 2 - 6.50%

Actual Pleurobema plenum - - 1 13 1 13.30% Live mussels Pleurobema sintoxia/rubrum - - 2 - 24 7.70%

Fusconaia flava 17 - - - - 0

Fusconaia subrotunda - 11 - - - 0 Pleurobema cordatum - - 31 - - 0

Actual Pleurobema plenum - - 1 14 - 6.70% Shells Shells only Pleurobema sintoxia/rubrum - - 2 - 24 7.70%

207

Table 3.7. Summary of the categorical and quantitative morphological variables used to describe and identify species belonging to the genera Fusconaia and Pleurobema in the Green River, Kentucky. Morphological variables were used to conduct the Decision Tree and Random Forest analyses. Pleurobema Variable Fusconaia flava Fusconaia subrotunda Pleurobema cordatum Pleurobema plenum sintoxia/rubrum Beaks face each other or are Beaks face forward, sometimes Beaks face forward. Beak Beaks face forward or are Beaks face each other or well destroyed, sometimes beaks face each other or are Sometimes beaks face each direction destroyed face forward beaks face forward well destroyed other or are destroyed Beak Most of the time beaks pass Most of the time beaks are position Beaks are usually not close to Beaks are either not close anterior margin. Sometimes not close to the anterior with Beaks are not close to anterior the anterior margin, sometimes to anterior margin or at beaks are at the anterior margin. Sometimes beaks respect to margin beaks are at anterior margin or anterior margin margin or pass anterior pass anterior margin or are anterior pass anterior margin margin at anterior margin margin Foot Usually orange and sometimes Orange White White White color white Nacre White and sometimes pink White White White White or pink color Most small sized mussel specimens have equilateral Small sized mussel specimens Small sized mussel shape but sometimes they have equilateral shape. Medium Small mussel specimens specimens have equilateral can be isosceles. Most sized mussel specimens are are circular while medium shape. Most medium sized medium sized mussel Shape Trapezoidal most of the time equilateral and and large sized individuals mussel specimens also have specimens have isosceles sometimes isosceles. Most are oval equilateral shape but shape but sometimes can large sized mussel specimens sometimes can be isosceles be equilateral. Large sized are isosceles mussel specimens are isosceles Most of the time the sulcus is absent in mussel Most of the time broad and specimens from the Green Sulcus Narrow and deep or narrow and shallow or narrow and Broad and deep Absent River. However, a few presence shallow, nearly always present shallow. Sometimes sulcus is mussel specimens may absent or narrow and deep have a narrow and deep or narrow and shallow sulcus Beak depth: Range 0.04 – 0.11 Range 0.04 – 0.1 Range 0.03 – 0.11 Range 0.06 – 0.11 Range 0.02 - 0.07 maximu m length

208

Table 3.7. Continued.

Pleurobema Variable Fusconaia flava Fusconaia subrotunda Pleurobema cordatum Pleurobema plenum sintoxia/rubrum Hinge length: Range 0.26 – 0.38 Range 0.34 – 0.47 Range 0.29 – 0.48 Range 0.36 – 0.48 Range 0.32 – 0.53 maximum length Small mussel specimens’ Small mussel specimens’ Medium mussel specimens’ ratio range 0.86 – 0.99 Small mussel specimens’ ratio range 0.79 – 0.94 ratio range 0.66 – 0.85 Medium mussel specimens’ Medium mussel specimens’ ratio range 0.86 – 1.03 Relative height Medium mussel specimens’ Large mussel specimens’ ratio range 0.81 – 0.99 ratio range 0.81 – 1.12 Medium mussel specimens’ ratio range 0.82 – 0.92 ratio range 0.59 – 0.77 Large mussel specimens’ ratio range 0.70 – 0.87 ratio range 0.74 – 0.85 Shell obesity Range 0.50 – 0.76 Range 0.41 – 0.66 Range 0.49 – 0.79 Range 0.56 – 0.75 Range 0.48 – 0.86 Small mussel specimens’ Small mussel specimens’ Wet weight: Medium mussel specimens’ ratio range 0.47 – 0.29 ratio range 0.38 – 1.07

maximum length ratio range 0.84 – 2.86 Medium mussel specimens’ Medium mussel specimens’ ratio range 1.38 – 1.75 ratio range 0.95 – 1.92

209

Figure 3.1. Sampling locations for freshwater mussel species in the genera Fusconaia and Pleurobema collected in 2015 and 2017 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) and Mammoth Cave National Park (GPS coordinates 37.17819, -86.1154; river mile = 197) in the Green River, Kentucky.

210

Figure 3.2. Landmarks used for the geometric morphometric analysis. Two main landmarks were aligned to the anterior part of the umbo (Point 1) and to the end of the hinge ligament (Point 3). A total of 11 landmarks separated by 15 degrees were used for obtaining shell measurement data.

211

Figure 3.3. Morphology of a freshwater mussel. Length (mm) measurements of the shell that were used for decision-tree and random-forest analyses are illustrated. A. Exterior of left valve; B. Interior of left valve.

212

A. Beak direction B. Beak position with respect to anterior margin

C. Foot color D. Nacre color

E. Shell Shape

F. Sulcus

Figure 3.4. Categorical variables that were used to conduct the Decision Tree and Random Forest analyses for species in the genera Fusconaia and Pleurobema sampled from the Green River, Kentucky. Variables used include: (A) beak direction, (B) beak position to anterior margin, (C) foot color, (D) Nacre color, (E) Shell shape, and (F) Sulcus. These characters were recorded for mussel specimens of species belonging to the genera Fusconaia and Pleurobema collected from the Green River, Kentucky.

213

Figure 3.5. Bar graph showing the percentage of correctly identified mussel specimens and misidentifications for each species of Fusconaia and Pleurobema by each expert. Mussel specimens were collected in 2015 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) in the Green River, Kentucky.

214

Figure 3.6. Pie chart showing the experts’ field identification for mussel specimens that were morphologically identified to the clade Pleurobema sintoxia/rubrum. Mussel specimens were collected in 2015 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) in the Green River, Kentucky.

215

A. Pleurobema rubrum shell form B. Pleurobema sintoxia shell form

Figure 3.7. Two representative shells for (A) Pleurobema rubrum and (B) Pleurobema sintoxia shell forms. These mussel specimens were identified consistently as these two shell forms by all experts. Mussel specimens were collected in 2015 from Pool 4 (GPS coordinates = 37.18286, -86.6296; river mile = 149) in the Green River, Kentucky.

216

Figure 3.8. Canonical Variate Analysis used to assess differences among three size classes [small (20-60 mm), medium (60-100 mm), and large sized mussel specimens (>100 mm)] of each species. Ellipses represent confidence intervals of the means, probability = 0.9. This analysis was conducted to illustrate and test for differences among mussel specimens within species at the three different size classes. For the analysis shell landmarks were used (see Figure 3.2) to measure the shells of each mussel specimen from the four respective species including: (A) Fusconaia flava, (B) Fusconaia subrotunda, (C) Pleurobema cordatum, (D) Pleurobema sintoxia/rubrum in the Green River, Kentucky. There was not enough individuals of Pleurobema plenum to be included in this analysis.

217

Fusconaia flava Fusconaia subrotunda Pleurobema cordatum Pleurobema plenum Pleurobema sintoxia/rubrum

A. Small size B. Medium size

Canonical variate 2 variate Canonical Canonical variate 2 variate Canonical

Canonical variate 1 Canonical variate 1

C. Large size Canonical variate 2 variate Canonical

Canonical variate 1

Figure 3.9. Canonical Variate Analysis used to assess differences among species at three size classes (A) small (20-60 mm), (B) medium (60-100 mm), (C) large sized mussel specimens (>100 mm) of each respective species. Ellipses represent confidence intervals for the means, probability = 0.9. This analysis compares all species at different sizes to show whether certain species are more similar to another species at a certain size class or not. For this analysis I used shell landmarks for mussel specimens of each species in the genera Fusconaia and Pleurobema sampled from the Green River, Kentucky.

218

Fusconaia Fusconaia flava Fusconaia flava subrotunda Fusconaia subrotunda

Pleurobema cordatum Pleurobema cordatum Pleurobema cordatum Pleurobema plenum

Pleurobema sintoxia, Pleurobema sintoxia, Pleurobema rubrum Pleurobema rubrum

Figure 3.10. Shell landmark (see Figure 3.2) variation for species in the genera Fusconaia and Pleurobema at different size classes [small (20-60 mm), medium (60-100 mm), large mussel specimens (>100 mm)] collected from the Green River, Kentucky. For Fusconaia flava, graphs explain 36.30% of geometric variation for small size mussel specimens, and 39.44% for medium mussel specimens. For Fusconaia subrotunda, graphs explain 60.03% of geometric variation for medium size mussel specimens, and 56.16% for large mussel specimens. For Pleurobema cordatum, graphs explain 42.81% of geometric variation for small size mussel specimens, 55.10% for medium mussel specimens, and 56.16% for large size mussel specimens. For medium mussel specimens of Pleurobema plenum, graphs explain 61.76% of geometric variation. Finally, shells landmark variation for P. sintoxia/rubrum. Graphs explain 51.90% of geometric variation for small size mussel specimens, and 52.90% for medium mussel specimens.

219

A. B. C.

Frequency Frequency Frequency

Beak depth : Maximum length Hinge length : Maximum length Perpendicular height : Maximum length D. E.

Fusconaia flava Fusconaia subrotunda

Frequency Pleurobema cordatum Frequency Pleurobema plenum Pleurobema sintoxia/rubrum

Width : Maximum length Wet weight : Maximum length

Figure 3.11. Normal distribution curves for ratios of quantitative morphological variables recorded for mussel specimens in the genera Fusconaia and Pleurobema collected from the Green River, Kentucky to include: (A) beak depth:maximum length, (B) hinge length:maximum length, (C) relative height, perpendicular height:maximum length (D), obesity, width:maximum length and (E) Wet weight:maximum length.

220

Species Sulcus presence Fusconaia flava p<0.001 Fusconaia subrotunda Pleurobema cordatum Pleurobema plenum Broad and deep Absent, broad and shallow, narrow and deep, narrow and shallow Pleurobema sintoxia/rubrum

Sulcus presence p<0.001

Narrow and deep, Absent, broad and narrow and shallow shallow

Beak direction Shape p<0.001 p<0.001

Face each other, Equilateral, unknown Circular, oval face forward isosceles

Beak direction p<0.001 Unknown, face forward, destroyed Face each other

0.09 0.22 0.33

0.67 0.91 0.78 n=25 n=35 n=9 n=24 n=14 n=12 Probabilities: Probabilities: Probabilities: Probabilities: Probabilities: Probabilities: Fusconaia flava = 1 Pleurobema Pleurobema Pleurobema Pleurobema Fusconaia cordatum = 0.91 plenum = 0.78 sintoxia/rubrum = sintoxia/rubrum = subrotunda = 1 Pleurobema Pleurobema 0.67 1 sintoxia/rubrum = cordatum = 0.22 Pleurobema 0.09 plenum = 0.33 Figure 3.12. Decision tree analysis showing external shell variables used to identify live mussel specimens of species in the genera Fusconaia and Pleurobema in the Green River, Kentucky. The tree was constructed using 80% of the data as training data and 20% as validation data. The 10 K-fold cross validation was performed in order to validate the model. Number of mussel specimens analyzed is represented by “n”. 221

Figure 3.13. Decision tree analysis showing external and internal shell variables used to identify shells only of Fusconaia and Pleurobema in the Green River, Kentucky. Decision tree was constructed using 80% of the data as training data and 20% as validation data. The 10 K- fold cross validation was performed in order to validate the model. Number of mussel specimens analyzed represented by “n”.

222

A. B.

C. D.

Figure 3.14. Most important morphological variables used for identifying mussels in the genera Fusconaia and Pleurobema in the Green River, Kentucky. Variable importance was determined by a mean decrease in identification accuracy for (A) living mussel specimens and (B) shells. The mean decrease in Gini (node purity, which measure the probability of misclassification) for (C) living mussel specimens (D) shells found using random forest analysis. Shell obesity represents shell width (mm): maximum length, while relative height represents height: maximum length.

223

Figure 3.15. Categorical variables and their proportion in each species belonging to the genera Fusconaia and Pleurobema. The categorical variables that were recorded for each mussel specimen were (A) beak direction, (B) beak position to anterior margin, (C) foot color, (D) nacre color, (E) sulcus description, and (F) shell shape per size-class. Morphological characters were recorded for mussel specimens belonging to species in the genera Fusconaia and Pleurobema collected from the Green River, Kentucky.

224

A. B.

Beak depth : Maximum Beak depth : Maximum length Hinge Hinge length Maximum : length

Fusconaia Fusconaia Pleurobema Pleurobema Pleurobema Fusconaia Fusconaia Pleurobema Pleurobema Pleurobema flava subrotunda cordatum plenum sintoxia/rubrum flava subrotunda cordatum plenum sintoxia/rubrum

C. D. Wet weight Wet weight : Maximum length

Perpendicular Perpendicular height : Maximum length Fusconaia Fusconaia Pleurobema Pleurobema Pleurobema Fusconaia Fusconaia Pleurobema Pleurobema flava subrotunda cordatum plenum sintoxia/rubrum flava subrotunda cordatum sintoxia/rubrum

length length

Maximum

: : Width

Fusconaia Fusconaia Pleurobema Pleurobema Pleurobema sintoxia/rubrum flava subrotunda cordatum plenum Figure 3.16. Ratios of quantitative morphological variables recorded for mussel specimens in the genera Fusconaia and Pleurobema collected from the Green River, Kentucky to include: (A) beak depth:maximum length, (B) hinge length:maximum length, (C) perpendicular height:maximum length, (D), wet weight:maximum length, and (E) width:maximum length. 225

DICHOTOMOUS AND PHOTOGRAPHIC KEYS

KEY A. Dichotomous key to identify live mussel specimens of the Fusconaia and Pleurobema species occurring in the Green River, Kentucky. Disclaimer: This dichotomous key has been developed to be used only in Fusconaia and Pleurobema species collected from the Green River, Kentucky 1) A. Foot: orange (Figure 1a) …………………………………………………………2 B. Foot: white (Figure 1b) …………………………………...………………...…...3

1a 1b

Figure 1. Mussel foot color: white (1a) and orange (1b).

2) A. Shell shape: trapezoidal (Figure 2a) …………………………... Fusconaia flava B. Shell shape: circular (Figure 2b) or oval (Figure 2c) …….Fusconaia subrotunda

2a. 2b. 2c.

2d. 2e.

Figure 2. Mussel shape: (2a) trapezoidal, (2b) circular, (2c) oval, (2d) equilateral, and (2e) isosceles shape.

3) A. Beak direction: Facing each other (Fig. 3a)………………….…………………..4

226

B. Beak direction: Facing forward (Fig. 3b) ………………………………………..5

3a 3b

Figure 3. Beak direction: (3a) facing each other and (3b) facing forward.

4) A. Shell shape: trapezoidal (Fig. 2a)………………………………Fusconaia flava B. Shell shape: equilateral (Fig. 2d) or isosceles (Fig. 2e)...………………………6 5) A. Shell sulcus: absent (Fig. 4a) ……………………..Pleurobema sintoxia/rubrum B. Shell sulcus: Present and either broad and shallow (Fig. 4b) or narrow and deep (Fig. 4c) or narrow and shallow (Fig. 4d) …………….………………....……..7 4A 4B

4C 4D

Figure 4. Sulcus presence: (4A) Absent, (4B) broad and shallow, (4C) narrow and deep, and (4D) narrow and shallow.

6) A. Shell sulcus: absent (Fig. 4a) ………………………Pleurobema sintoxia/rubrum B. Shell sulcus: Narrow and shallow (Fig. 4d) ……………....Pleurobema cordatum

227

7) A. Position of beak with respect to anterior margin: does not extend beyond anterior margin (Fig. 5a) ……………………………………..……Pleurobema cordatum B. Position of beak with respect to anterior margin: extends beyond anterior margin (Fig. 5b)………………………………………………………Pleurobema plenum

5a 5b

Figure 5. Position of the beak with respect to the anterior margin: (5a) beak does not extend beyond the anterior margin and (5b) beak extends beyond anterior margin.

228

KEY B. Dichotomous key to identify dead mussel specimens (shell without soft parts) of the Fusconaia and Pleurobema species occurring in the Green River, Kentucky. Disclaimer: This dichotomous key has been developed to be used only in Fusconaia and Pleurobema species collected from the Green River, Kentucky

1) A. Shape: circular (Figure 1a) or oval (Figure 1b)………….. Fusconaia subrotunda B. Shape: equilateral (Figure 1c), isosceles (1d), or trapezoidal (Figure 1e) .……...2

1a 1b. 1c.

1d. 1e.

Figure 1. Shell shape: (1a) circular, (1b) oval, (1c) equilateral, (1d) isosceles, and (1e) trapezoidal shape.

2) A. Beak direction: Face each other (Fig. 2a) ………………….………………..…..3 B. Beak direction: Face forward (Fig. 2b) ………………………………………….4

2a 2b

Figure 2. Beak direction: (2a) facing each other and (2b) facing forward.

3) A. Shell shape: trapezoidal (Fig. 1a) ……………...………………Fusconaia flava B. Shell shape: equilateral (Fig. 1d) or isosceles (Fig. 1e) triangular.……..………6 4) A. Shell sulcus: absent (Fig. 3a) …………………….Pleurobema sintoxia/rubrum B. Shell sulcus: Broad and shallow (Fig. 3b) or narrow and deep (Fig. 3c) or narrow and shallow (Fig. 3d) ………………………………………………..…………..7

229

3a 3b

3c 3d

Figure 3. Sulcus presence: Absent (3a), broad and shallow (3b), narrow and deep (3c), and narrow and shallow (3d).

5) A. Shell sulcus: absent (Fig. 3a) …………………….Pleurobema sintoxia/rubrum B. Shell sulcus: Narrow and shallow (Fig. 3d) ……………..Pleurobema cordatum 6) A. Position of beak with respect to anterior margin: not close to anterior margin (Fig. 4a) ………………………………...... …………………Pleurobema cordatum B. Position of beak with respect to anterior margin: Pass anterior margin (Fig. 4b)……………………………………………………………Pleurobema plenum

4a 4b

Figure 4. Position of the beak with respect to the anterior margin: (4a) beak is not anterior the anterior margin and (4b) beak pass anterior margin.

230

KEY C. Photographic key

231

KEY C. 1. Small and medium sized mussel specimens of Fusconaia flava collected from the Green River, Kentucky. Mussel specimens of small and medium size classes have a trapezoidal shape, with a ventral margin that is slightly more enlarged in medium sized mussel specimens compared to small mussels. All individuals have a sulcus that was broad and could vary from shallow to deep. Typically, the nacre color was white. However, occasionally mussel specimens had pink colored nacre. Foot color for most individuals was orange but occasionally some specimens had a white colored foot.

232

KEY C. 2. Mussel specimens of Fusconaia subrotunda of small, medium, and large sizes sampled in the Green River, Kentucky: Small individuals have a circular shape. However, medium and large sized individuals typically show a more oval shape where the ventral margin is more elongated than in small size mussels. There is an absent of sulcus in all sizes. The nacre color is white and the foot color is orange.

233

KEY C. 3. Mussel specimens of Pleurobema cordatum of small, medium and large size individuals sampled from the Green River, Kentucky. Small size mussels typically show an equilateral shape, whereas medium size individuals can have either equilateral or isosceles shape. In addition, all large size individuals have an isosceles shape. The ventral margin usually gets larger (in comparison to the dorsal and anterior margins) as the mussel grows. The sulcus is either “narrow and shallow” or “narrow and deep”. The color of the nacre and the foot is white.

234

KEY C. 4. Mussel specimens of Pleurobema plenum of small and medium sized individuals in the Green River, Kentucky. Mussel specimens of P. plenum typically show and equilateral shape, but occasionally, they have an isosceles shape. The sulcus is not an important variable to describe this species, as species show absence of sulcus or it is “broad and shallow”, “narrow and shallow”, or “narrow and deep”. As this is an endangered species, the nacre was recorded only for 4 mussel specimens and it was white. Foot color is usually white.

235

KEY C. 5. Mussel specimens of Pleurobema sintoxia/rubrum of small, medium, and large sized individuals sampled from the Green River, Kentucky. Mussel specimens of all sizes were typically equilateral but a few individuals have an isosceles shape. Mussel specimens of medium and large sizes may become isosceles in shape as the ventral margin is enlarged in contrast with small mussel specimens. Most of the mussel specimens did not have a sulcus. However, some mussel specimens have a “narrow and shallow” or “narrow and deep sulcus”. The nacre color can be either white or pink and the foot color is always white.

236

Chapter 4 Synthesis and Recommendations

To inform management decision making, I aimed to achieve molecular identification, estimation of genetic diversity and effective population size, and morphometric analysis of species of Fusconaia and Pleurobema in the Green River, Kentucky (Pool 4 and MCNP). My results and management recommendations are synthetized in the following paragraphs.

Molecular identification

I used three mitochondrial markers (COI, ND1, 16s rRNA) and a nuclear marker (ITS1) to identify species of Fusconaia and Pleurobema in the Green River, Kentucky. All sampled mussels were sequenced for COI and ND1. Additional DNA sequences that have been reported for COI

(Inoue et al. 2018) and ND1 (Burlakova et al. 2012, Bertram et al. 2015, Jones et al. 2015, Schilling

2015, Marshall et al. 2018) were added to my analyses to construct phylogenetic trees and split networks to confirm species identification and to test for any cryptic species. Results from

Bayesian phylogenetic trees, split phylogenetic networks, haplotype networks, species delimitation, and pairwise distance performed with COI, ND1, COI + ND1 sequences suggested that five species were detected in the Green River, KY – including Rough Pigtoe (Pleurobema plenum, federally endangered), Ohio Pigtoe (P. cordatum, near threatened), Long-Solid

(Fusconaia subrotunda, vulnerable), and Wabash Pigtoe (F. flava, least concern) and a single clade that included mussel specimens of Pink Pigtoe (P. rubrum, near threatened) and Round Pigtoe (P. sintoxia, least concern). Additional 16s rRNA and ITS1 markers were used to test further whether

P. rubrum and P. sintoxia were two different species or conspecific. Study results from 16s rRNA showed no difference between these two nominal species, whereas the ITS1 phylogenetic analysis did not show enough genetic differentiation among the two taxa to clearly support designation as

237 separate species. My study did not result in the identification of cryptic species in the Green River of Kentucky.

Estimation of genetic diversity and effective population size (Ne)

A concatenated sequence for COI and ND1 was used to estimate genetic diversity as well as the long-term female effective population size (Nef) for Green River populations of F. flava, F. subrotunda, P. cordatum, P. plenum, and P. sintoxia/rubrum. I developed primers for microsatellites for P. plenum and I used additional microsatellites originally developed for P. clava

(Jones et al. 2015) and P. pyriforme (Moyer and Williams 2011) to estimate genetic diversity as well as contemporary and long-term effective population sizes (Ne). Following the “50/500 rule”, the estimations for contemporary Ne for P. cordatum (Ne = 231.4, 95%, CI = 41.6-infinity), P. plenum (Ne = 303.3, 95% CI = 68.5-infinity) and P. sintoxia/rubrum (Ne = 131.07, 95%, CI =

39.5-infinity); together with the estimates for long-term Ne for P. cordatum (Ne = 17,500; 95% CI

= 8,570 – 34,300), P. plenum (Ne = 5,740; 95% CI = 3,100 – 8,950) and P. sintoxia/rubrum (Ne =

8,370; 95% CI = 5,300 – 11,100) suggest that the populations in the Green River are large enough to avoid inbreeding depression and maintain their evolutionary potential. However, to estimate the minimum viable population size (MVP) of any species, it is necessary to understand its life history, habitat, and current reproduction. My results for Nef were F. flava (Nef = 69,000; 95% CI = 39,500

– 109,000), F. subrotunda (Nef = 160,000; 95% CI = 112,000 – 196,000), P. cordatum (Nef =

419,000; 95% CI = 325,000 – 517,000), P. plenum (Nef = 9,530; 95% CI = 8,290 – 9,910), and P. sintoxia/rubrum (Nef = 19,900; 95% CI = 19,600 – 20,000). Surprisingly, these estimates were higher than Ne estimations using biparental microsatellite DNA markers. Our contemporary and long-term effective population sizes for P. cordatum, P. plenum, and P. sintoxia/rubrum seemed

238 to meet the “50/500 rule”. However, the MVP size has not been determined for any mussel species.

Hence, there is a need to assess life-history, habitat, current reproductive processes (gravidity) before assuring that these species have healthy populations that will continue to adapt over time.

Many mussel species have disappeared as older individuals have died, as poor water quality may have stopped their reproduction

Morphometric analysis

I asked five experts to identify mussel specimens of Fusconaia and Pleurobema collected from Pool 4 in the Green River, Kentucky. On average, experts were able to correctly identify these mussels 70% of the time. The easiest to identify were mussels of F. flava and F. subrotunda, which on average were correctly identified of 95% and 81%, respectively. Species in the genus

Pleurobema were more challenging to identify by the experts (range of correct identification, 59%

- 77%). This study aimed to find morphological difference among these species as well as to create a dichotomous key to improve the correct identification of shells and live mussels. Mussel specimens for all of the study species were divided into three sizes-classes (small, 20–60 mm; medium, 60–100 mm; and large, >100 mm mussel specimens). I used Canonical Variate Analysis

(CVA) to test for differences between species at different size-classes, and then used CVA to test for differences in each species at difference size-classes. The results showed that Pleurobema species were more difficult to identify among themselves than from species in the genus

Fusconaia. Decision trees and random forest analyses resulted in the most-appropriate variables to use in a dichotomous key as well as construction of dichotomous trees for species identification.

Two dichotomous keys were created, one for the identification of shells and another for live mussels using the more easily identified characters from results of decision and random tree

239 analyses. However, these keys still need to be tested in the field. Principal challenges to the identification of these mussels seem to be higher when separating mussel specimens of

Pleurobema than for Fusconaia. In this study, I have described mussel specimens collected from the Green River, KY but in some instances, field identification may need to be supported by a molecular identification. Finally, the proposed dichotomous keys should only be used only for the

Green River watershed.

240