<%BANNER%>

Population Genetics of the Endangered Miami Blue Butterfly Cyclargus thomasi bethunebakeri

Permanent Link: http://ufdc.ufl.edu/UFE0024241/00001

Material Information

Title: Population Genetics of the Endangered Miami Blue Butterfly Cyclargus thomasi bethunebakeri Implications for Conservation
Physical Description: 1 online resource (173 p.)
Language: english
Creator: Saarinen, Emily
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: conservation, dna, ecology, lepidoptera, microsatellites
Entomology and Nematology -- Dissertations, Academic -- UF
Genre: Entomology and Nematology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: POPULATION GENETICS OF THE ENDANGERED MIAMI BLUE BUTTERFLY Cyclargus thomasi bethunebakeri: IMPLICATIONS FOR CONSERVATION The Miami blue butterfly (Cyclargus thomasi bethunebakeri) (Lepidoptera: Lycaenidae) is endemic to the state of Florida and, until the 1980s, was locally common throughout much of the southern mainland, western barrier islands and the Florida Keys. Over the last few decades, the butterfly s overall distribution and numerical abundance have been significantly reduced and a variety of factors have been identified as potential causes of overall decline. In 1999, the butterfly was rediscovered as part of a small breeding population in Bahia Honda State Park in the lower Florida Keys and in 2007 an additional population was found in Key West National Wildlife Refuge. A captive breeding colony from Bahia Honda State Park individuals was established in order to provide organisms for reintroduction and research. The current level of existing molecular diversity and gene flow of wild and captive populations are unknown, and I present data from twelve polymorphic microsatellite loci to assess the populations. Analysis of museum specimens shows that populations were historically widespread and genetically diverse. Some diversity has been lost in the current populations, however a greater genetic diversity (allelic diversity and heterozygosity) than originally predicted exists in both Bahia Honda State Park and Key West National Wildlife Refuge. Allelic frequencies also show short-range gene flow between habitat patches within Bahia Honda State Park. I show that the captive colony maintains representative genetic diversity from the wild colony, indicating that captive-bred individuals are genetically suitable for reintroduction to new sites. Genetic distances are increased between the wild colony and the captive colony when new wild stock is not integrated to the captive colony on a regular basis. Laboratory experiments with different numbers of founding individuals emphasize the importance of how stochastic demographic factors negatively impact small populations. The ultimate purpose of this project is to gain critical scientific information on this state-endangered butterfly in order to facilitate informed decision making regarding the short and long-term management of the taxon so as to secure a stable or increasing population at a level that does not require protection.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Emily Saarinen.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Daniels, Jaret.
Local: Co-adviser: Miller, Jacqueline Y.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024241:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024241/00001

Material Information

Title: Population Genetics of the Endangered Miami Blue Butterfly Cyclargus thomasi bethunebakeri Implications for Conservation
Physical Description: 1 online resource (173 p.)
Language: english
Creator: Saarinen, Emily
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: conservation, dna, ecology, lepidoptera, microsatellites
Entomology and Nematology -- Dissertations, Academic -- UF
Genre: Entomology and Nematology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: POPULATION GENETICS OF THE ENDANGERED MIAMI BLUE BUTTERFLY Cyclargus thomasi bethunebakeri: IMPLICATIONS FOR CONSERVATION The Miami blue butterfly (Cyclargus thomasi bethunebakeri) (Lepidoptera: Lycaenidae) is endemic to the state of Florida and, until the 1980s, was locally common throughout much of the southern mainland, western barrier islands and the Florida Keys. Over the last few decades, the butterfly s overall distribution and numerical abundance have been significantly reduced and a variety of factors have been identified as potential causes of overall decline. In 1999, the butterfly was rediscovered as part of a small breeding population in Bahia Honda State Park in the lower Florida Keys and in 2007 an additional population was found in Key West National Wildlife Refuge. A captive breeding colony from Bahia Honda State Park individuals was established in order to provide organisms for reintroduction and research. The current level of existing molecular diversity and gene flow of wild and captive populations are unknown, and I present data from twelve polymorphic microsatellite loci to assess the populations. Analysis of museum specimens shows that populations were historically widespread and genetically diverse. Some diversity has been lost in the current populations, however a greater genetic diversity (allelic diversity and heterozygosity) than originally predicted exists in both Bahia Honda State Park and Key West National Wildlife Refuge. Allelic frequencies also show short-range gene flow between habitat patches within Bahia Honda State Park. I show that the captive colony maintains representative genetic diversity from the wild colony, indicating that captive-bred individuals are genetically suitable for reintroduction to new sites. Genetic distances are increased between the wild colony and the captive colony when new wild stock is not integrated to the captive colony on a regular basis. Laboratory experiments with different numbers of founding individuals emphasize the importance of how stochastic demographic factors negatively impact small populations. The ultimate purpose of this project is to gain critical scientific information on this state-endangered butterfly in order to facilitate informed decision making regarding the short and long-term management of the taxon so as to secure a stable or increasing population at a level that does not require protection.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Emily Saarinen.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Daniels, Jaret.
Local: Co-adviser: Miller, Jacqueline Y.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024241:00001


This item has the following downloads:


Full Text

PAGE 1

1 POPULATION GENETICS OF THE ENDANGERED MIAMI BLUE BUTTERFLY Cyclargus thomasi bethunebakeri : IMPLICATIONS FOR CONSERVATION By EMILY VIRGINIA SAARINEN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FL ORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Emily Virginia Saarinen

PAGE 3

3 To my husband, Justin Arthur Saarinen and my daughter, Lila V irginia Saarinen for their steadfast love patience, and support

PAGE 4

4 ACKNOWLEDGMENTS I thank my committee chair Dr Jaret Daniels and co chair Dr. Jacqueline Miller for encouraging me to work with the Miami blue butterfly. Not only did they support my re search endeavors with their time and funding, but they led by excellent example how science should be conducted for the betterment of endangered taxa. I especially thank Drs. Lee and Jacqueline Miller for helping cultivate my interest in entomology at Ne w College of USF and for funding an early research trip to the Caribbean I thank my entire doctoral advising committee members, Dr. Pamela S. Soltis, Dr. Thomas C. Emmel, and Dr. Marc A. Branham, for continued support. My committee provided essential tr aining and support in the molecular laboratory critical conversations in evolutionary theory, and helpful editing of grants and manuscripts. I especially thank Dr. Daniels for his assistance in helping me to become a better writer and for guiding and enc ouraging me along the way. I am grateful for the environment provided to me in the laboratory of Drs. Jim and Al jandra Maruniak in the Department of Entomology and Nematology at the University of Florida Al taught me how to work successfully in the molecular laboratory and I will always be grateful for her help. The unwavering kindness and intellectual ability of both Drs. Maruniak helped me through the design and implementation of my molecular projects and I thank them both for listening to my pres entations, proofreading manuscripts, identifying the baculovirus in the Miami blue larvae and for helping me to build a bridge between ecology and genetics. I also thank Ginger Clark for her patience in answering my many microsatellite related questions and for running all of my samples so expertly through the UF ICBR laboratory. I also thank Dr. Jessica Light for training me in the Ancient DNA laboratory at the Florida Museum of Natural History

PAGE 5

5 I am thank ful for the support that I received from the D epartment of Entomology and Nematology at the University of Florida through both an Alumni Fellowship and the intellectual support from faculty and my fellow graduate students. I thank Debbie Hall profusely for her consummate care of all students and abil ity to keep my registration from falling through the cracks! I also thank my extended family of researchers at the McGuire Center for Lepidoptera and Biodiversity at the Florida Museum of Natural History. Christine Eliazar kept me sane and smiling and Ge orge Austin found valuable specimens in the collection and was always quick with a joke. I thank Valerie McManus for being a phenomenal laboratory and field assistant and Bret Boyd for taking care of the captive colony of butterflies. I thank my officema te Matt Trager for his insightful manuscript reviews and scholarly discussions of conservation biology. I thank John Calhoun for his helpful correspondence and relaying of important historical references. Conversations with John and other members of the valuable insight into the historical distribution and biology of the Miami blue butterfly. I also thank the staff and scientists at the many natural history museums that have loaned specimens or sent records for this project I thank Suzanne Rab Green (American Museum of Natural History ); George Austin Jackie Miller, Lee Miller, and Charlie Covell (McGuire Center for Lepidoptera and Biod iversity at the Florida Museum of Natural History) ; John R awlins and Vanessa Verdecia (Carnegie Museum of Natural History) ; Bob Robbins and Brian Harris (Smithsonian National Museum of Natural History ); and Jim Dunford for visiting the Field Museum. I esp ecially thank Mr. Ricardo Zambrano of the Florida Fish and Wildlife Conservation Commission ( FWC ) for his support of my research his ability to put science first, and for his

PAGE 6

6 actions as the consummate diplomat in complicated circumstances with this endang ered butterfly I also thank the members of the Miami Blue Working G roup for their passion and dedication to this butterfly. I thank Dr. Sue Perry and Elsa Alvear for lodging, transportation, and logistical support and scientific assistance while working in both Everglades and Biscayne National Parks. I profusely thank the Canon National Parks Science Scholars Program for funding my research and for believing in my abilities to succeed as a conservation scientist. Lastly, I thank my family for being supportive and silly and for helping me to remain well rounded throughout this project I am grateful to my parents and three brothers for their support and understanding during the last several years. The greatest thanks belong to my husband Justin, not only for his brilliant GIS help but for his unwavering ability to support our family and remain a solid partner and best friend

PAGE 7

7 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ......................... 10 LIST OF FIGURES ................................ ................................ ................................ ....................... 13 ABSTRACT ................................ ................................ ................................ ................................ ... 15 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 17 History of the Miami Blue Butterfly ................................ ................................ ...................... 17 History of the Genus Cyclargus Nabokov 1945 ................................ .............................. 18 Listing Status of Cyclargus thomasi bethunebakeri ................................ ........................ 19 Captive Colony ................................ ................................ ................................ ................ 19 Reintro duction Efforts ................................ ................................ ................................ ..... 20 Ecology of Cyclargus thomasi bethunebakeri ................................ ................................ ........ 20 Life History ................................ ................................ ................................ ..................... 20 Ant Associations ................................ ................................ ................................ .............. 23 Population Genetics ................................ ................................ ................................ ................ 23 F statistics ................................ ................................ ................................ ........................ 23 Founder Effects and Inbreeding ................................ ................................ ...................... 24 Microsatellites ................................ ................................ ................................ ................. 26 2 DEVELOPMENT AND CHARACTERIZATION OF POLYMORPHIC MICROSATELLIT E LOCI ................................ ................................ ................................ .... 29 Introduction ................................ ................................ ................................ ............................. 29 Methods ................................ ................................ ................................ ................................ .. 30 Results ................................ ................................ ................................ ................................ ..... 31 Discussion ................................ ................................ ................................ ............................... 32 3 UTILIZING MUSEUM SPECIMENS TO ASSESS HISTORICAL DISTRIBUTION AND GENETIC DIVERSITY ................................ ................................ ............................... 34 Introduction ................................ ................................ ................................ ............................. 34 Methods ................................ ................................ ................................ ................................ .. 37 Collection Database ................................ ................................ ................................ ......... 37 Genetic Sampling ................................ ................................ ................................ ............ 39 DNA Extractions and Microsatellite Analyses ................................ ................................ 39 Results ................................ ................................ ................................ ................................ ..... 42 Geographic Distribution of Museum Specimens ................................ ............................ 42

PAGE 8

8 Relationships Between Size, Sex, Year, and Collection Location of M useum S pecimens ................................ ................................ ................................ .................... 42 Comparison of Body Size in Historical and Current P opulations ................................ ... 43 Historical Genetic Diversity ................................ ................................ ............................ 43 Disc ussion ................................ ................................ ................................ ............................... 46 4 GENETICS OF WILD POPULATIONS ................................ ................................ ............... 63 Introduction ................................ ................................ ................................ ............................. 63 Meth ods ................................ ................................ ................................ ................................ .. 65 Habitat Maps ................................ ................................ ................................ ................... 65 Sampling ................................ ................................ ................................ .......................... 66 DNA Extractions and Microsate llite Amplification ................................ ........................ 67 Statistical A nalyses ................................ ................................ ................................ .......... 67 Estimation of Microsatellite Mutation Rate ................................ ................................ .... 69 Results ................................ ................................ ................................ ................................ ..... 70 Molecular Diversity ................................ ................................ ................................ ......... 70 Gene Flow in Bahia Honda State Park ................................ ................................ ............ 72 Mutation Rate of Microsatellites ................................ ................................ ..................... 74 Discussion ................................ ................................ ................................ ............................... 74 5 GENETICS OF THE CAPTIVE COLONY ................................ ................................ ........... 90 Introduction ................................ ................................ ................................ ............................. 90 Methods ................................ ................................ ................................ ................................ .. 91 Basis of Captive Colony ................................ ................................ ................................ .. 91 Incorporation of New Genetic Material ................................ ................................ .......... 92 Sampling of Captive Colony ................................ ................................ ........................... 93 Statist ical Analyses ................................ ................................ ................................ .......... 93 Results ................................ ................................ ................................ ................................ ..... 95 Genetic Changes in the Captive Colony o ver Time ................................ ........................ 95 Comparisons with Bahia Honda State Park ................................ ................................ ..... 96 Private Alleles ................................ ................................ ................................ ................. 98 Bottlenecks ................................ ................................ ................................ ...................... 99 Discussion ................................ ................................ ................................ ............................... 99 Private Alleles ................................ ................................ ................................ ............... 101 Bottlenecks ................................ ................................ ................................ .................... 102 Effective Population Size ................................ ................................ .............................. 102 Implications for Conservation ................................ ................................ ....................... 103 6 THE GENETIC AND FITNESS CONSEQUENCES OF FOUNDER EFFECTS ON LABORATORY POPULATIONS ................................ ................................ ....................... 112 Introduction ................................ ................................ ................................ ........................... 112 Methods ................................ ................................ ................................ ................................ 114 Experimental Treatment Lines ................................ ................................ ...................... 114 One pair founder lines ................................ ................................ ............................ 114

PAGE 9

9 Three pair founder lines ................................ ................................ ......................... 115 Control Groups ................................ ................................ ................................ .............. 115 Fitness Measurements and Statistical Analyses ................................ ............................ 115 DNA Extractions and PCR Amplifi cations ................................ ................................ ... 116 Genetic Statistical Analyses ................................ ................................ .......................... 117 Results ................................ ................................ ................................ ................................ ... 118 Conclus ions ................................ ................................ ................................ ........................... 120 Experimental Shortcomings and Fortuitous Discoveries ................................ .............. 120 Founder Effects on Heterozygosity, Genetic Diversity, and Evol utionary Potential .... 121 7 SYMBIOSIS BETWEEN MIAMI BLUE LARVAE AND ANTS ................................ ..... 129 Introduction ................................ ................................ ................................ ........................... 129 Methods ................................ ................................ ................................ ................................ 130 Results ................................ ................................ ................................ ................................ ... 131 Discussion ................................ ................................ ................................ ............................. 132 8 CONCLUSIONS ................................ ................................ ................................ .................. 139 Future Research ................................ ................................ ................................ .................... 143 Conservation Implications ................................ ................................ ................................ .... 144 APPENDIX A LIST OF ABBREVIATIONS ................................ ................................ ............................... 147 B DNA EXTRACTION METHODS ................................ ................................ ....................... 149 C ALLELE FREQUENCY TABLES ................................ ................................ ...................... 152 D PRIVATE ALLELE LIST ................................ ................................ ................................ .... 156 LIST OF REFERENCES ................................ ................................ ................................ ............. 158 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 173

PAGE 10

10 LIST OF TABLES Table page 1 1 Distribution and records of species of the genus Cyclargus ................................ ............. 28 1 2 Subs pecies of Cyclargus thomasi. ................................ ................................ ..................... 28 2 1 Characteristics of the twelve polymorphic microsatellite loci developed for C. t. bethunebakeri ................................ ................................ ................................ ................. 33 3 1 Human population estimates in south Florida counties. ................................ .................... 53 3 2 Museum specimens of C. t. bethunebakeri sampled for genetic diversity.. ...................... 54 3 3 Results of ANOVA analyses comparing historical (1927 1988 FLMNH) and contemporary (2002 2006 BHSP) C. t. bethunebakeri wing chord lengths (mm mean standard error) over several variables. ................................ ................................ ........... 59 3 4 Overview of molecular diversity measures for C. t. bethunebakeri populations at microsatellite locus CthB119. ................................ ................................ ............................ 59 3 5 Frequency of alleles at microsatellite locu s CthB119 in historical and current populations of C. t. bethunebakeri ................................ ................................ .................... 61 3 6 Pairwise population differentiation between each of the historical groups and contemporary populations of C. t. b ethunebakeri ................................ ............................. 62 4 1 Sampling regime for molecular studies of C. t. bethunebakeri on Bahia Honda State Park, Florida 2005 2006. ................................ ................................ ................................ ... 83 4 2 Genotyping error rates for each microsatellite locus of the C. t. bethunebakeri .............. 83 4 3 Summary statistics of allelic patterns for C. t. bethunebakeri in Bahia Honda State Park (BHSP), Florida.. ................................ ................................ ................................ ....... 84 4 4 Hardy Weinberg equilibrium (HWE) probabilities for C. t. bethunebakeri in BHSP a nd KWNWR populations by locus. ................................ ................................ ................. 84 4 5 Wilcoxon signed rank tests for excess (top) and deficiency (bottom) of heterozygosity for discrete breeding populations of wild and captive raised laboratory populations of C. t. bethunebakeri ................................ ................................ ... 85 4 6 Effective population size ( N E ) estimates of the BHSP population of C. t. bethunebakeri ................................ ................................ ................................ ................... 86 4 7 Results from Structure 2.0 for K values 1 5, data are from all BHSP subpopulatio ns ( N = 114) and are not pre assigne d to populations for analysis. ................................ ........ 88

PAGE 11

11 4 8 Results from Structure 2.0 for K values 1 5, data are from the BHSP 9 2005 ( N = 24) and KWNWR 2 2008 ( N = 27) populati ons. ................................ ................................ .... 89 5 1 Summary of C. t. bethunebakeri initial captive colony numbers from February 2003 to December 2005. ................................ ................................ ................................ ........... 106 5 2 Summary of second iteration captive colony numbers of C. t. bethunebakeri from April 2006 to December 2007.. ................................ ................................ ........................ 107 5 3 Overview of molecular diversity measures for Bahia Honda State Park (BHSP) and capt ive colony (CC) populations of C. t. bethunebakeri ( N ). ................................ .......... 107 5 4 Effective population size ( N E ) estimates of the 2006 captive colony of C. t. bethunebakeri ................................ ................................ ................................ .................. 108 5 5 D S ). ................................ ........... 109 5 6 D S ). ............. 109 5 7 Wilcoxon signed rank tests for excess (top) and deficiency (bottom) of heterozygosity for populations of captive raised laboratory populations of C. t. bethunebakeri ................................ ................................ ................................ .................. 111 6 1 Effects of founder lines on fitness measures (mean standard error). ............................ 124 6 2 Effects of founder lines on genetic variability measures. ................................ ................ 124 7 1 Ants of Everglades National Park (ENP) and Bahia Honda State Park (BHSP) in Florida collected in proximity to larvae of the C. t. bethunebakeri ................................ 138 C 1 Allele frequency for locus CthB11. ................................ ................................ ................. 152 C 2 Allele frequency for locus CthB101. ................................ ................................ ............... 152 C 3 Allele frequency fo r locus CthB103. ................................ ................................ ............... 152 C 4 Allele frequency for locus CthB106. ................................ ................................ ............... 153 C 5 Allele frequency for locus CthB115. ................................ ................................ ............... 153 C 6 Allele frequency for locus CthB117. ................................ ................................ ............... 153 C 7 Allele frequency for locus CthB119. ................................ ................................ ............... 154 C 8 Allele frequency for locus CthC12. ................................ ................................ ................. 154 C 9 Allele frequency for locus CthC116. ................................ ................................ ............... 154 C 10 Allele frequency for locus Cth C124. ................................ ................................ ............... 155

PAGE 12

12 C 11 Allele frequency for locus CthC127. ................................ ................................ ............... 155 C 12 Allele frequency for locus CthD7. ................................ ................................ ................... 155 D 1 Private allele list. ................................ ................................ ................................ .............. 156

PAGE 13

13 LIST OF FIGURES Figure page 3 1 Weather data averages for Key West, Florida ................................ ................................ 53 3 2 Museum specimens of C. t. bethunebakeri collected from 1900 1930, the frontier era as per Solecki (2001). ................................ ................................ ................................ ........ 55 3 3 Museum specime ns of C. t. bethunebakeri collected from 1931 1970, the development era as per Solecki (2001). ................................ ................................ ............. 56 3 4 Museum specimens of C. t. bethunebakeri collected from 1971 present, the globalization era as per Solecki (2001). ................................ ................................ ............. 57 3 5 Number of C. t. bethunebakeri specimens deposited in five major museum collections by Florida county from 1900 1990 ................................ ................................ .. 58 3 6 Number of C. t. bethunebakeri specimens deposited in five major museum collections for Dade and Monroe Counties, Florida ................................ .......................... 58 3 7 Ratio of H O /H E values for historical K ey Largo (yellow) and contemporary BHSP (green and red) and KWNWR (blue) populations of C. t. bethunebakeri ; circled points are from locus CthB119. ................................ ................................ ......................... 60 3 8 Frequency of alleles at microsatellit e locus CthB119 in historical C. t. bethunebakeri populations. ................................ ................................ ................................ ........................ 60 3 9 Allelic frequencies at microsatellite locus CthB119 of all populations of C. t. bethunebakeri ................................ ................................ ................................ .................... 61 4 1 Schematic diagram of Bahia Honda State Park, Florida C. t. bethunebakeri sampling locations ................................ ................................ ................................ ........................... 82 4 2 Removal of a wing fragment from an adult C. t. bethunebakeri butterfly for use in genetic analyses. ................................ ................................ ................................ ................ 83 4 3 Inbreeding coefficients (and standard error bars) for each metapopulation segment of C. t. bethunebakeri on Bahia Honda State Par k over time (2005 2006) ..................... 85 4 4 Pairwise F ST R ST and corresponding P values (after diagonal) for C. t. bethunebakeri on BHSP ................................ ................................ ................................ ........................... 87 4 5 Pairwise F ST R ST and corresponding P values (after diagonal) for C. t. bethunebakeri on BHSP in 2006.. ................................ ................................ ................................ ............. 88 4 6 Structure plot of all BHSP data ( N =114) when K = 2, data are from all B HSP subpopulations ( N = 114) and are not pre assigned to populations for analysis.. ............. 89

PAGE 14

14 4 7 Structure plot of KWNWR 2 2008 ( N = 27) and BHSP 9 2005 data ( N = 24) when K = 2 and prior populatio n assignment is incorporated into the model.. ............................... 89 5 1 Average allelic richness values of 12 microsatellite loci measured in discrete generations of the Bahia Honda State Park (BHSP) and capti ve colony (CC) of C. t. bethunebakeri ................................ ................................ ................................ ................. 108 5 2 ny (CC) butterflies over time. ................................ ................................ ............................. 110 5 3 captive colony (CC) butterflies over time. ................................ ................................ .... 110 5 4 Observed heterozygosi ty values averaged over 12 microsatellite loci for generations of C. t. bethunebakeri ; BHSP, Bahia Honda State Park population; CC, captive colony. ................................ ................................ ................................ .............................. 111 6 1 Distributions of alleles (bp) at l ocus CthB119 in an outbred population (top) and populations founded by one pair of adult C. t. bethunebakeri (five replicates). ............. 125 6 2 Distributions of alleles (bp) at locus CthC127 in an o utbred population (top) and populations founded by one pair of adult C. t. bethunebakeri (five replicates). ............. 126 6 3 Distributions of alleles (bp) at locus CthD7 in an outbred population (top ) and populations founded by one pair of adult C. t. bethunebakeri (five replicates). ............. 127 6 4 Cyclargus thomasi bethunebakeri larvae killed by baculovirus. ................................ ..... 128 7 1 Crematogaster ashmeadi ants tending a late instar Cyclargus thomasi bethunebakeri larva. ................................ ................................ ................................ .............................. 136 7 2 Details of Cyclargus thomasi bethunebakeri fourth i nstar caterpillar.. ........................... 137

PAGE 15

15 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy POPULATIO N GENETICS OF THE ENDANGERED MIAMI BLUE BUTTERFLY Cyclargus thomasi bethunebakeri : IMPLICATIONS FOR CONSERVATION By Emily Virginia Saarinen May 2009 Chair: Jaret C. Daniels Co chair: Jacqueline Y. Miller Major: Entomology and Nematology The Miami blue butterfly ( Cyclargus thomasi bethunebakeri ) (Lepidoptera: Lycaenidae) is endemic to the state of Florida and until the 1980s, was locally common throughout much of the southern mainland, western barrier islands and the Florida Keys Over the last few de cades, the and a variety of factors have been identified as potential causes of overall decline In 1999, the butterfly was rediscovered as part of a small breeding p opulation in Bahia Honda State Park in the lower Florida Keys and in 2007 an additional population was found in Key West National Wildlife Refuge A captive breeding colony from Bahia Honda State Park individuals was established in order to provide organi sms for reintroduction and research. The current level of existing molecular diversity and gene flow of wild and captive population s are unknown and I present data from twelve polymorphic microsatellite loci to assess the populations Analysis of museum specimens shows that populations were historically widespread and genetically diverse. Some diversity has been lost in the current populations, however a greater genetic diversity (allelic diversity and heterozygosity) than originally predicted exists in both Bahia Honda State Park and Key West National Wildlife Refuge. Allelic frequencies also show short range gene

PAGE 16

16 flow between habitat patches within Bahia Honda State Park. I show that the captive colony maintains representative genetic diversity from the wild colony, indicating that captive bred individuals are genetically suitable for reintroduction to new sites. Genetic distances are increased between the wild colony and the captive colony when new wild stock is not integrated to the captive colony on a regular basis. Laboratory experiments with different numbers of founding individuals emphasize the importance of how stochastic demographic factors negatively impact small population s The ultimate purpose of this project is to gain critical scienti fic information on this state endangered butterfly in order to facilitate informed decision making regarding the short and long term management of the taxon so as to secure a stable or increasing population at a level that does not require protection.

PAGE 17

17 C HAPTER 1 INTRODUCTION History of the Miami Blue Butterfly The Miami blue butterfly Cyclargus thomasi bethunebakeri (Comstock and Huntington) (Lepidoptera: Lycaenidae: Polyommatinae) is a small, brightl y colored butterfly endemic to s outh Florida, with addi tional species occurring throughout the Caribbean basin (Smith et al. range within Florida, but most authors agree that it once commonly occurred from the sou thern mainland through the Florida Keys and the Dry Tortugas (Klots 1964; Lenczewski 1980; Minno & Emmel 1993; Calhoun et al. 2002). Additional reports of C. t. bethunebakeri from the Bahamas, the Bi mini Islands, are unverifiable as there are no specimens located in museum holdings (L.D. Miller, personal communication ) These likely represent irregular, non breeding strays (Calhoun et al. 2002). Primarily a coastal species, C. t. bethunebakeri inhabits tropical coastal hammocks and beachside scrub. Over distribution and numerical abundance have been significantly reduced (Minno & Emmel 1994; Calhoun et al. 2002). Rapid urbanization and the associated loss of coastal habitat have eliminated C. t. bethunebake ri from the s outh Florida mainland. The butterfly has become considerably rarer and was thought to have been extirpated with no verified records from March 1992 to October 1999 (Calhoun et al. 2002). The taxon was rediscovered in November 1999 as part of a small population within the boundaries of Bahia Honda State Park (BHSHP) on Bahia Honda Key in the lower Florida Keys ( Ruffin & Glassberg 2000). Independent l epidopterists specializing in the taxonomy of this group have verified that the butterflies di scovered on BHSP are indeed the same subs pecies as historic specimens (L.D. and J.Y. Mill er, personal comm unication). I n December 2007, an additional population was discovered in Key West

PAGE 18

18 Nat ional Wildlife Refuge (KWNWR) 50 km west of Key West, Florida (C annon 2007). No taxonomic verification by genitalic dissection of these butterflies has occurred due to permit constraints of the KWNWR, but preliminary molecular evidence has confirmed that they are the same taxon (E.V. Saarinen, unpublished data). His tory of the Genus Cyclargus Nabokov 1945 The genus Cyclargus was split from the genus Hemiargus Hbner 1885 by Nabokov in 1945 (Nabokov 1945). Nabokov made this distinction based on structural differences in the male genitalia between the two groups and d esignated C. ammon Lucas 1857 from Cuba as the type speci es for the genus (Nabokov 1945). In his description of the new genus Cyclargus Nabokov designated the species C dominica Mschler 1886 from Jamaica, C. thomasi Clench 1941 from Cat Island (Bahamas ), and C. woodruffi Comstock & Huntington 1943 from Anegada (Virgin Islands) as members of the genus. Riley (1975) incorrectly re synonomized the genera Hemiargus and Cyclargus despite Nab Riley apparently s a nk Cyclargus back in Hemia rgus due to similar wing characteristics but failed to distinguish the genitalic distinctions recognized by Nabokov (Johnson & Blint 1995). In their revision of neotropical blue butterflies (Polyommatinae), Johnson and Matusik (1988) discovered this taxo nomic discrepancy and addressed the issue in later publications (Johnson & Blint 1995). However, several recent field guides ( Opler & Malikul 1992 ; Allen et al 2005 ; Minno et al. 2005). Blint and Johnson (1995) name eight species in Cyclargus ( ammon, erembis, thomasi, woodruffi dominica, kathleena, sorpresus and kathleena ). Treatment of the genus by Smith et al. (1994) includes a new species from Puerto Ri co, bringing the total to nine species (Table 1 1). The Miami blue butterfly Cyclargus thomasi bethunebakeri (Comstock and Huntington) is endemic to Florida with additional subspecies occurring in the Bahamas, Puerto Rico, and

PAGE 19

19 Hispaniola (Smith et al. 19 94) (Table 1 2). The type specimen of C. t. bethunebakeri is housed at the American Museum of Natural History (AMNH) and the type locale is Miami, Florida. This subspecies is often written as bethune bakeri but the original description does not hyphenat e the subspecies name (Comstock & Huntington 1943). Listing S tatus of Cyclargus thomasi bethunebakeri The decline of Cyclargus thomasi bethunebakeri may have begun gradually, but it has become dramatic within recent years. This butterfly is currently one of the most endangered taxa in the world due to its low population numbers and presence at a fraction of its original geographic range. Through the joint efforts of both state and federal agencies, it was given emergency status as a state endangered taxon in November 2002 by the Florida Fish and Wildlife Conservation Commission (FWC). An emergency management plan was later drafted and approved, and a captive propagation program was initiated (FWC 2003). As of 2008, C t bethunebakeri is still a candidat e to be listed as a federally endangered subspecies and is the only taxon in Florida that was listed due to public petition. Captive Colony The emergency management plan of 2002 (FWC 2003) called for the creation of a captive propagation program to be hous ed at the McGuire Center for Lepidoptera and Biodiversity of the Florida Museum of Natural History at the University of Florida in Gainesville, Florida (FLMNH). This colony was initiated utilizing eggs from the BHSP colony in December 2002. The purposes of this colony are to conserve the remaining genetic diversity of the taxon, safeguard the remaining population on BHSP from natural and human factors that may threaten its existence, and produce a significant number of individuals for reintroduction into the wild. As of 2008, 32 captive generations have been produced yielding in excess of 28,000 individuals ( J.C. Daniels unpublished data).

PAGE 20

20 Reintroduction Efforts With a stable and productive captive colony, efforts began in 2004 to release individuals in to suitable habitats in Biscayne National Park (BNP) and Everglades National Park (ENP) as well as Dagney Johnson Key Largo Hammocks Botanical State Park (DJSP) Since May 2004, over 7,500 adults and larvae have been released into these protected parks, a reas which are not threatened by urbanization or development. In fact, because of widespread mosquito spraying and habitat destruction outside of the parks, these regions provide a refuge of the only remaining unspoiled beach scrub a nd hardwood hammock ha bitat in s outh Florida. Unfortunately, the extensive damage caused by the 2005 hurricanes has precluded any future releases in ENP. Efforts are currently focused on BNP and DJ SP and the extensive, endangered hardwood hammock present on the outlying islan ds of BNP. Ecology of Cyclargus thomasi bethunebakeri Life H istory Cyclargus thomasi bethunebakeri inhabits tropical hardwood hammocks and their associated margins, beachside scrub and tropical pine rocklands (Minno & Emmel 1993). Of particular interest in this taxon are changes in historic al and current use of larval host plan ts. Historic al records list the f abaceous host plants Caesalpinia bonduc (L.) Roxb. C pulcherrima L. Pithecellobium keyense Britton & Rose P. unguis cati L. and plants of the family Sapindaceae Cardiospermum halicacabum L. and C corindum L. (Matteson 1930; Klots 19 64 ; Howe 1975; Lenczewski 1980; Brewer 1982; Carroll & Loye 1987; Minno & Emmel 1993; Smith et al.1994; Calhoun et al. 2002; Caroll & Loye 2006). Earlier records (Klotz 1951 ; Howe 1975) denote Caesalpinia bonduc as the larval host plant of this taxon, with references of the last thirty years reporting Cardiospermum halicacabum (Lenczewski 1980 ; Leston et al. 1982 ; Minno & Emmel 1993 ; Smith et al. 1994 ; Calhoun et a l. 2002). There are numerous records

PAGE 21

21 from museum specimen s and private collector s of larvae in Florida feeding on Cardiospermum halicacabum ( Minno & Emmel 1993; Smith et al.1994; Calhoun et al. 2002), and possibly Cardiospermum corindum but there is some confusion between the plant taxa and larval feeding (Carroll & Loye 2006). Carroll & Loye (1987; 2006) and Carroll et al. (2003) have studied Cardiospermum spp. distribution in Florida and have documented feeding on this genus by other insects. They rep ort that although other authors report C. t. bethunebakeri larvae feeding on Cardiospermum halicacabum that it is more likely that larvae were feeding upon Cardiospermum corindum a native woody perennial. The species C. halicacabum has been considered an agricultural pest, and active eradication plans in Florida have removed many of the plants form the natural landscape. Cardiospermum corindum fruits year round and are predominant in the Upper Keys of Florida, making it the likely host plant of C. t. beth unebakeri larvae rather than C. halicacabum Carroll & Loye (2006) also report adult characters of the host plant key out to C. corindum (Wunderlin 1998). It is likely that authors writing about C. t. bethunebakeri larvae have cited past references and p erpetuated the taxonomic error in food plant choice in Cardiospermum and efforts to locate larvae were focused on this plant and not the previously noted hosts of Caesalpinia spp or Pithecellobium spp In order to feed on Cardiospermum spp ., C. t. bethun ebakeri larvae chew a hole into the inflated fruit (hence the common name balloon vine and heart seed vine) and eat the developing seed inside the balloon. C yclargus thomasi bethunebakeri larvae leave their entry hole into the fruit unplugged, allowing fo r ants to enter the pod to gather sugary secretions from the larvae (Minno & Emmel 1993). The mode of feeding is entirely d ifferent for larvae eating the f abaceous hosts Caesalpinia bonduc and Pithecollobium keyense as larvae feed on developing shoots, yo ung leaves, and flower buds (FWC 2003). Despite extensive surveys, no C. thomasi bethunebakeri larvae have

PAGE 22

22 been found on balloonvine since 1992 ( J.C. Daniels, personal communication). The current population of C. t. bethunebakeri on BHSP only feeds upon Caesalpinia bonduc as larvae, even though P. keyense is present (FWC 2003). The recently discovered population of butterflies in KWNWR is reported to only feed on P. keyense (Canon 2007), and no other host plants are present on these islands. Because of the large taxonomic division between Fabaceae and Sapindaceae and the underlying chemical differences, there has been great interest in the ability of larvae to exploit both host plants. Cyclargus thomasi bethunebakeri are multiply brooded on BHSP and individuals can be found throughout the year, although the life cycle does slow in winter months and a slight diapause is suspected ( E.V. Saarinen and J.C. Daniels, unpublished data ). The generation time is approximately one month. Adult butterflies mate and females lay eggs on developing shoots, foliage, and flower buds of Caesalpinia bonduc Eggs are laid singly, but may be clustered on developing leaves, shoot tips, and flower buds. Larvae chew out of eggs after several days of development. Larvae develop through four instar stages, the total larval development time lasting 3 4 weeks, with faster development times occurring in warmer and more humid months. Fourth instar larvae pupate in sheltered or inconspicuous areas on C. bonduc often underneat h leaf whorls or bracts. Adult butterflies eclose after 5 8 days, again depending on temperature and humidity. Adult butterflies live for 3 7 days in the field, and sex ratios are equal (J.C. Daniels, personal communication). Additional details of life history can be found in F WC (2003). The C. t bethunebakeri populations on both BHSP and KWNWR are metapopulations, although the two metapopulations do not interact due to over 70 km distance separating the two ( E.V. Saarinen and J.C. Daniels, unpublishe d data ) Immature stages are confined to discrete

PAGE 23

23 habitat patch areas of suitable host plant and development is completed in the natal patch. When adult butterflies eclose, they may migrate to nearby, suitable host plant patches in search of mates, food resources, and suitable habitat for egg laying. This dispersal is usually short distance as these butterflies are poor fliers (FWC 2003) but adult dispersal does facilitate gene flow between habitat patches. Habitat patches are continuously colonized by gravid females when they are of appropriate quality, and other patches may go extinct when environmental conditions are poor. Thus, there is a dynamic relationship between habitat patches that is driven by environmental factors and the biology of this but terfly. The term metapopulation depicts the dynamic nature of both organism movement through the landscape and the changeable landscape itself. Ant A ssociation s Historical, anecdotal records of C t bethunebakeri have mentioned larval associations with the Florida carpenter ant, Camponotus s p p Recent studies confirm that larvae associate with Camponotus floridanus (Buckley) as well as another member of the genus Camponotus planatus (Roger) Additionally, larvae have been observed tended by Crematog aster ashmeadi Emery Forelius pruinosus (Roger) and Tapinoma melanocephalum (Fabricius) Field surveys of remaining butterfly habitat and recent reintroduction sites reveal other potential ant associates ( Paratrechina longicornis (Latreille) and Paratre china bourbonica (Forel) ) and a host of potential ant predators (Saarinen & Daniels 2006). Detailed information is also presented about larval ant associated organs and their mediation of this facultative symbiosis. Population Genetics F statistics Wrig ht (1951) developed F statistics as a way to describe population genetic structure. These statistics are often used to infer gene flow and evaluate whether a population is in Hardy

PAGE 24

24 Weinberg equilibrium (HWE). Populations in HWE meet the following assumpt ions: infinitely large population size, random mating, no mutation, no selection, and no migration. Deviations from HWE may mean that one or more of these conditions are not met and use of F statistics may help determine which variations are due to inbree ding. The F ST statistic has a theoretical minimum of 0.0 and a theoretical maximum of 1.0. Low values indicate little genetic divergence while higher values indicate the fixation of alternative alleles (Hartl & Clark 1997). In 1978, Wright suggested gui delines for interpreting F ST values. F ST values vary considerably among utilized for relative comparison of values. He suggested the range 0 0.05 indicates litt le genetic differentiation, 0.05 0.15 indicates moderate genetic differentiation, 0.15 0.25 indicates great genetic differentiation, and F ST values above 0.25 indicate very great genetic differentiation (ital. in Hartl & Clark 1997). Values can also be ev aluated whether they are significantly greater than 0. Population differentiation may be underestimated by low F ST values, however, and considerable caution should be used when interpreting results. When population studies utilize microsatellite markers, the statistic R ST is often used in place of or addition to F ST The former measure specifically considers the stepwise mutation model (SMM) of microsatellite markers and the high rate of mutation in these markers (Slatkin 1995). Additional F statistics include F IS (the correlation between alleles within individuals and the subpopulation) and F IT (correlation between alleles within individuals and the total population). Founder Effects and Inbreeding ogists have been concerned with the effects of inbreeding on organism fitness (Darwin 1876). Early experiments showed the negative impacts of inbreeding due to increased homozygosity on the alleles that affect fitness (Charlesworth & Charlesworth 1999) An alternative hypothesis was also given that inbreeding

PAGE 25

25 depression is due to the expression of deleterious alleles that are unmasked as homozygotes. Wright (1977) showed that heterozygotes have an advantage at loci affecting fitness, and that in randomly mating populations these alleles will be retained as polymorphisms. An excess of homozygotes may occur in non randomly mating populations, such as those that are inbred. Sibling mating concentrates similar alleles in the same individual, and when these are deleterious recessives a less fit phenotype is realized and inbreeding depression is observed (Charlesworth & Charlesworth 1987). Additionally, deleterious, recessive alleles may arise in inbred populations via mutation, and these may also negatively impact fitness (Wright 1977). The statistic f relates the cumulative amount of inbreeding in a single population. F statistics and coefficient of inbreeding, f offer standard measures for comparing genetic data for inbred populations. Inbree ding depression is a major concern in endangered species management (Templeton 2002). As endangered species often exist in small, fragmented populations, instances of inbreeding are often quite high. Gene flow may be severely reduced, and the genetic var iation of metapopulations may be exceedingly low. Endangered species may suffer the negative effects of genetic drift, and all members of a population may exhibit decreased fitness measures as a result (Hedrick & Kalinowski 2000). Inbreeding need not alw ays be the culprit of reduced fitness, however. Deleterious alleles may be purged in small, inbred populations, resulting in very well adapted individuals (Byers & Waller 1999). Inbred individuals may be extremely homozygous and less capable of adapting to environmental change, but fitness in the current environment is high. Different species react differently to levels of inbreeding; mammals may exhibit inbreeding depression whereas invertebrates are capable of thriving at similar inbreeding coefficient s (Hedrick & Kalinowski 2000; Haikola et al. 2001). It is difficult to make

PAGE 26

26 generalizations about the role of inbreeding in endangered species management plans due to the complex nature and variety of animal and plant genomes. Individual studies are cruc ial to our understanding of reduced population mating systems, and the specific factors that affect traits related to fitness need to be examined in individual endangered species (Hedrick & Kalinowski 2000). Current statistical analyses and advances in mo lecular markers (e.g. microsatellites) have enabled us to detect previously hidden levels of inbreeding depression. Furthermore, inbreeding experiments allow for the direct detection of fitness component reduction in laboratory lines. There is no assump tion that inbred lines in the laboratory will behave the same as inbred wild individuals, but a correlation can be inferred (Charlesworth & Charlesworth 1999). Microsatellites Microsatellites are highly repetitive, tandem repeats of DNA. They are typical ly 1 6 base pairs (bp) in length and are found in tandem repeats in (typically) non coding regions of the genome. Because they are non coding regions of DNA, microsatellites can retain mutations and exhibit high mutation rates. Differences in the number of microsatellite repeats are useful in determining population structure, heterozygosity, geographic variation, and paternity. They have become the marker of choice in many butterfly population genetic studies owing to the high rate of mutation (Palo et a l. 1995; Saccheri et al. 1999; Anthony et al. 2001; Keyghobadi et al. 2002; Harper et al. 2003). Heterozygote deficiencies, geographic variation, and population structure were successfully analyzed with microsatellites in the endangered Karner blue butter fly, Lycaeides melissa samuelis (Nabokov) and such studies have set a precedent that genetic studies can be done on endangered species with minimal sampling effect and a great benefit to management plans (Nice et al. 2000; Anthony et al. 2001). Despite i ts usefulness in population level studies, the isolation of Lepidopteran microsatellite DNA has been difficult (Zhang 2004)

PAGE 27

27 but other techniques have shown successful amplification of loci using an enriched genomic library (Harper et al. 2000).

PAGE 28

28 Table 1 1 Distribution and records of species of the genus Cyclargus Species Species distribution Reference ammon Cuba, Grand Bahamas, Great Abaco, North Andros Lucas 1857 dominica Jamaica Mschler 1886 erembis Cayman Islands Nabokov 1948 kathleena Dominican R epublic Johnson & Matusik 1992 oualiri Puerto Rico Brevignon 2002 shuturn Jamaica Johnson & Blint 1995 sorpresus Dominican Republic Johnson & Matusik 1992 thomasi Florida, Bahamas, Turks and Caicos, Hispaniola Clench 1941 woodruffi Lesser Antillean i slands Comstock & Huntington 1943 Table 1 2 Subspecies of Cyclargus thomasi Subspecies Subspecies distribution Reference thomasi (nominate ssp.) Great Bahama Bank islands Clench 1941 bethunebakeri Florida* Comstock & Huntington 1943 bahamensis C rooked, Acklin, and Ragged Islands, and Plana Cays, Bahamas Clench 1942 noeli Hispaniola (Haiti and Dominican Republic) Comstock & Huntington 1943 clenchi Mayaguana, Turks and Caicos, Great and Little Inagua Miller, Simon, & Harvey 1992 *Smith et al. (1 994) note recordings of C. t. bethunebakeri from Bimini Islands, Bahamas, but no specimens in current museum holdings have been located to verify this.

PAGE 29

29 CHAPTER 2 DEVELOPMENT AND CHARACTERIZATION OF POLYMORPHIC MICROSATELLITE LOCI Introduction The Miami blue butterfly ( Cyclargus thomasi bethunebakeri ) (Lepidoptera: Lycaenidae) is endemic to the state of Florida and until the 1980s, was relatively widespread and locally common throughout much of the southern portion of the state. Over the last few decade s, the variety of factors have been identified as potential causes of decline, including habitat loss, fragmentation and degradation (Calhoun et al. 2002). Report s of verified records rapidly dwindled, and by the early 1990s C. t. bethunebakeri was presumed to be extirpated. In 1999, the butterfly was rediscovered as part of a small breeding population in Bahia Honda State Park (BHSP) in the lower Florida Keys. A rapid biological assessment yielded monthly abundance estimates ranging between 35.5 and 141.6 individuals during 2002 and 2003 ( J.C. Daniels, unpublished data .). Cyclargus thomasi bethunebakeri is currently a state endangered taxon in Florida and a cand idate for federal listing. As part of a larger conservation effort, a captive breeding colony was established at the at the University of Florida in Gainesville, Florida This captive colony was founded with individuals from the Bahia Honda State Park population, and is infused with new genetic material from the same source on a regular basis. Organisms from the captive breeding colony provide individuals for reintroducti on into suitable habitats in s outh Florida (FWC 2003). The microsatellite markers developed here will be used to analyze both wild and captive breeding populations as well as newly established colonies resulting from organism reintroductions.

PAGE 30

30 Methods A mi crosatellite library was created using the pooled DNA extracts from six adult butterfly thoraxes. These butterflies originated from the captive breeding colony at the University of Florida, and DNA was extracted with a DNeasy Qiagen Kit shortly after the butterflies expired naturally. DNA was sent to Genetic Identification Services, Inc. (GIS), Chatsworth, CA ( http://www.genetic id services.com/ ) for the creation of the microsatellite library. Enrichme nt was performed for four types of repeats, and GIS used a magnetic bead capture technology to capture GA, AAT, ATG, and TAGA repeats. A total of 62 different microsatellite containing clones were sequenced by GIS using universal M 13 primers (forward: 5' AGGAAA CAGCTATGACCA TG 3' and reverse: 5' ACGACGTTGTAAAACGAC GG 3' ). DNA Sequencing Kit (Amersham Biosciences P/N US81050), followed by electrophoresis on an Applied BioSystems Mo del 377 DNA Sequencer. Primers were designed for 33 microsatellite containing clones using DesignerPCR v. 1.03 (Research Genetics, Inc.). I tested primer pairs against pooled DNA used for microsatellite library development. I then tested twenty seven ca ndidate loci against seven butterfly individuals, and twelve loci were determined to be polymorphic and suitable for use in genotyping. DNA was extracted from 2 mm 2 wing fragments for genotyping of the 114 Bahia Honda State Park butterflies and stored in 95 100% ethanol at 80 C. Data are presented for all butterflies, representing several breeding populations sampled from September 2005 October 2006. Data is also shown for the September 2005 population; these butterflies were sampled over a five day per iod and represent a single breeding population. PCR conditions for these 12 loci were optimized on a gradient thermocycler (MJ Research PTC 200 Peltier Thermal Cycler). Amplification reactions were performed in a final volume of

PAGE 31

31 10.0 l containing approxi mately 2 ng of DNA 2.5x Eppendor f 5 PRIME Master Mix (1.25 U Taq DNA polymerase, 50 mM KCl, 30 mM Tris HCl, 1.5 mM Mg, 0.1% Igepal CA360, and 200 M of each dNTP), 0.1 mg/ml BSA and 0.6 M of each primer. The PCR program consisted of 94 C for 3 min; 34 cycles of 94 C for 40 s, 57 C for 40 s, 72 C for 30 s, and extension at 72 C for 4 min. Primers CthB117 and CthB119 had annealing temperatures of 54 C instead of 57 HEX, or TET (Sigma Genosys). PCR products were run with GeneS can LIZ labeled internal size standard (A pplied Biosystems, ABI) and analyzed using an ABI 3730 Automated DNA Sequencer and GeneS can 3.1.1 (ABI). Peaks were scored with GeneM apper version 4.0 ( ABI) and manually confirmed. Data were organized in Microsoft Excel and formatted using C onvert (Glaubitz 2004) to be used in other software packages. Arlequin version 3.01 (Excoffier et al 2005) was used to estimate the observed and expected heterozyg osities (H O and H E ) and to test for Hardy Weinberg equilibrium. FSTAT version 2.9.3.2 (Goudet 2001) was used to test for linkage disequilibrium. Micro Checker version 2.2.3 (van Oosterhout et al 2004) was used to test for the presence of null alleles. Sequential Bonferroni corrections for multiple comparisons were performed as per Rice (1988). Results Allele numbers varied from four to 14 and ranges of observed and expected heterozygosities were 0.02679 0.79630 and 0.06154 0.69565 respectively (Table 2 1). O nly o ne locus, CthC116 was determined to be in Hardy Weinberg equilibrium after sequential Bonferroni correction. I performed 66 pairwise comparisons of all loci for linkage disequilibrium using FSTAT, and after a sequential Bonferroni correctio n, no loci showed evidence of linkage disequilibrium Null alleles were detected at loci CthB101, CthB106,

PAGE 32

32 CthB117, CthB119, CthC12, CthC116, and CthC124 due to the presence of homozygote excess. T here was no evid ence of large allele drop out Analys is of a single breeding population of butterflies ( N = 24) sampled in September 2005 shows seven loci to be in Hardy Weinberg equilibrium after Bonferroni correction (Table 2 1). One locus (CthC124) was monomorphic in this population. No loci were in lin kage disequilibrium Null alleles were detected at loci CthB117, CthB119, and CthC12 due to the presence of homozygote excess. Discussion The butterflies genotyped for this study were from a very small population, which likely accounts for a significant deviation from Hardy Weinberg equilibrium at several loci. Homozygote excess was likely the result of generations of inbreeding. Population demographics are currently under investigation and further studies will employ the use of the microsatellite mark ers described here.

PAGE 33

33 Table 2 1 Characteristics of the twelve polymorphic microsatellite loci developed for C. t. bethunebakeri Primer sequence, F =F AM H =H EX T =TET ; T a optimized annealing temperature; Size range, observed size range of alleles in bp ; N, number of C. t. bethunebakeri successfully genotyped (out of 114); N A total number of alleles; H O and H E observed and expected heterozygosities, respectively; P HWE (total/ BHSP 9 2005), P value for Hardy Weinberg equilibrium across the whole sampl e set of 114 individuals (above diagonal) and for the single breeding population ( n =24) sampled in September 2005 from Bahia Honda State Park (below diagonal), those in italics are not in equilibrium. Locus Name GenBank Accession no. Repeat motif Primer s T a Size range (bp) N N A H O H E P HWE (total / BHSP 9 2005) CthB11 EU429568 (ATT) 12 F: F GTGACGAAAAGATAGCAACAG R:TTCCATCAGGTGACTCCTC 57 150 216 108 7 0.79630 0.69565 0.00040 / 0.03049 CthB101 EU429569 (AAT) 10 F: H TCACTGATGCTCGTCTCTCAC R:CCGA ACCGTTGGTAGGTC 57 138 201 107 6 0.51402 0.66237 0.00000 / 0.28371 CthB103 EU429570 (AAT) 10 F: T CTCCGAGATCACTCGAACT R:CGGAAATGAACCATGTTG 57 150 233 113 6 0.49558 0.57278 0.00000 / 0.16213 CthB106 EU429571 (ATT) 8 F: H CTGAGCCTAGTTTAGACTCCAG R:GGGATTTGAGAATG GTTTAGA 57 163 208 108 7 0.31481 0.41796 0.00000 / 0.00000 CthB115 EU429572 (ATT) 11 F: F TGGTGTGGTGTAAAGTATGTCA R:TCATGGTTCCTATCATCAGC 57 292 441 111 7 0.45946 0.59989 0.00000 / 0.09258 CthB117 EU429573 (ATT) 21 F: H CAGGGTAGGTAAATGCGTGTC R:GCTTTCCATCAGGTGA CTCC 54 152 189 114 9 0.20175 0.34222 0.00000 / 0.00000 CthB119 EU429574 (AAT) 11 F: F AGGGTAGGTAAATGCGTGTC R:CTTTCCATCAGGTGACTCCT 54 135 186 110 10 0.18182 0.31802 0.00000 / 0.00034 CthC12 EU429575 (GAT) 6 F: T GATCCGATTGCCTCTATCG R:TTACGCCTCATTCTTAACATCC 57 171 228 113 14 0.31858 0.65971 0.00000 / 0.00076 CthC116 EU429576 (CAT) 17 F: T TGCGAAACTACAGTCACCTC R:AAGACCATCCAAGAAACCAC 57 242 251 112 4 0.25000 0.31126 0.10107 / 0.57871 CthC124 EU429577 (CAT) 16 F: H GAACCACTGACCTCGTAACATG R:TGCAGGAAGGATAAACAAACAC 5 7 230 248 112 5 0.02679 0.06154 0.00000 / Monomorphic CthC127 EU429578 (GAT) 11 F: F AAATCGTGATTAGCTGTCGA R:TAAGGAACTCGGCATCATG 57 191 206 108 5 0.49074 0.60168 0.00261 / 0.02057 CthD7 EU429579 (TATC) 24 F: H TGGTAGAGCCGATTTAGGTG R:CAACCAGCTCAACTAACCAAC 57 221 313 110 13 0.59091 0.69124 0.00000 / 0.17686

PAGE 34

34 CHAPTER 3 UTILIZING MUSEUM SPECIMENS TO ASSESS HISTORIC AL DISTRIBUTION AND GENETIC DIVERSITY Introduction The results of conservation programs for imperiled species are substantially improved when approp riate ecological and genetic information are incorporated into research and management plans (IUCN 2008). However, even for historically widespread and abundant species, sufficient data may not be available prior to the decrease in the geographic range or population size of the focal taxa. Such population declines are frequently concomitant with losses of genetic variation that often occur first as responses to these declines and then as negative feedback that result in further decreases in population siz e (Frankham 2003). Consequently, conservation efforts may be severely hindered by both insufficient information on the current status of imperiled species and imprecise or entirely absent data regarding ecological and genetic aspects of extinct population s. A historical approach to reconstructing the ecology of imperiled taxa can offer invaluable knowledge regarding the geographic and genetic composition of these lost populations (Leonard 2008). Museum collections can be an important resource for such a pproaches (Harper et al. 2006; Winston 2007). Insect collections are generally kept at low relative humidities and constant temperatures, conditions which improve DNA preservation (Mandrioli et al. 2006; Leonard 2008). Some researchers have reported suc cess in genotyping museum specimens of insects (Goldstein & D eSalle 2003; Harper et al. 2003, 2006) whereas others have questioned the utility of such methods (Watts et al. 2007). However, recent advances in molecular genetic methods greatly improve the s uccess of such endeavors (e.g. Harper et al. 2006). Additionally, the s tudy of museum specimens allows researchers to plot historical occurrences of specimens, often with sufficient coverage to infer geographic range, and to identify potential changes in

PAGE 35

35 populations over time. distribution from specimen label data with genetic analysis of the corresponding specimen is a powerful new approach for understanding the spatio genetic structure of h istorical populations and may therefore have substantial contributions for directing future research and conservation efforts. The Miami blue butterfly ( Cyclargus thomasi bethunebakeri ) was historically common in the Florida Keys and southern Florida and a necdotal records show that its range extended northward along both coasts to Hillsborough and Volusia Counties ( Forbes 1941 ; Kimball 1965; Calhoun et al. 2002; Carroll & Loye 2006). Amateur and professional naturalists regularly encountered this butterfly in localized patches in south Florida, even referring to it as a nuisance when looking for rarer butterfly taxa (C. Covell, I. Finklestein, and R. Gilmore, personal communications). However, C. t. bethunebakeri populations began to disappear in the 1970s and by the 1980s it became increasingly difficult to find the butterfly even in apparently high quality habitat in southern Florida. By 1990 the butterfly had become considerably rarer and, with no verified records from March 1992 to October 1999, was th ought to be extinct (Calhoun et al. 2002). A small colony of butterflies was rediscovered in November 1999 within the boundaries of Bahia Honda State Park (BHSP) on Bahia Honda Key in the lower Florida Keys ( Ruffin & Glassberg 2000). In December 2007, an addit ional population w as discovered on several uninhabited islands within the Key West National Wildlife Refuge (KWNWR) (Cannon 2007). Despite formerly high population abundance and extensive geographic range throughout coastal southern Florida, wild po pulations of C. t. bethunebakeri are currently limited to a few small insular areas in the extreme southern portion of the state.

PAGE 36

36 A variety of anthropogenic and natural factors likely led to the decline of C. t. bethunebakeri with habitat loss, degradat ion and fragmentation driven by development and human population growth undoubtedly playing key roles. The human population of the state has increased from fewer than a million people in 1920 to over 18 million in 2007 (U.S. Bureau of Census, www.census.gov/population/cencounts/fl190090.txt ) Dade and Monroe Counties, two areas in which the butterfly was historically abundant, have exemplified this rapid growth (Table 3 1) Land use and land cover change have been correspondingly dramatic in this region, with agricultural and urban land currently predominating (Solecki 2001). Solecki (2001) divided the changes in s outh Florida land use into three eras: the frontier era (1870 1930), development era ( 1931 1970), and globalization era ( 1971 present) based on characteristics of political and economic changes occurring in the region The greatest land cover changes in Florida took place in the development era from 1953 1973 with the conv ersion of 5,800 km 2 of natural areas to agricultural and urban use areas (289.97 km 2 /year) (Solecki & Walker 2001). H abitat loss and fragmentation cannot be singled out as the sole causes of decline of C. t. bethunebakeri as there are a multitude of unkno wn factors in operat ion Additionally, natural forces have impact ed remaining areas of habitat suitable for C. t. bethunebakeri as evidenced by several major hurricanes in the 1920s and 1990s and extreme drought in the 1960s (Solecki 2001). The synergist ic effects between habitat loss and the unquantifiable losses from predation, parasitism, competition from invasive species, and other unidentified factors should not be discounted or overlooked (Calhoun et al. 2002). The goal of this study was to examin e h ow changes in the population size and geographic distribution of C. t. bethunebakeri have affected the genetic diversity of this historically widespread taxon that has experienced dramatic decreases in range and population size

PAGE 37

37 Specifically, I used an cient DNA techniques to assess changes in genetic diversity in a single population over time and compared historical genetic diversity with that found in the two remaining populations. Similar approaches have proven useful for other conservation projects (Ross et al. 2006; Leonard 2008). My study was motivated by the need to determine historical levels of genetic diversity over time in order to better understand and preserve the current genetic diversity present in extant populations. This study has furt her addressed whether the size of the butterflies have changed over time, season, and location, factors which may speak to the evolutionary potential and plasticity of the taxon. Methods Collection Database I directly transcribed or requested label data fr om C t bethunebakeri specimens held by five museums with preeminent Lepidoptera collections: the McGuire Center for Lepidoptera and Biodiversity at the Florida Museum of Natural History, Gainesville, Florida (FLMNH) ( N = 361); the American Museum of Natu ral History, New York City, New York (AMNH) ( N = 215); the Carnegie Museum of Natural History, Pittsburgh, Pennsylvania (CMNH) ( N = 179); the Smithsonian National Museum of Natural History, Washington, D.C. (SNMNH) ( N = 108); and the Field Museum, Chicago, Illinois (CNHM) ( N = 14). All specimens were verified in consultation with taxonomic experts as being C. t. bethunebakeri Taken together, label data were used to assess and develop the distributional range of the butterfly over time. I used label inf ormation from the collection database to create a GIS database for C. t. bethunebakeri Label data were georeferenced using Google Earth 4.3 ( http://earth.google.com/ ) and then converted into a geodatabase using the convert kml to shp tool from ArcGIS 9.2 (Environmental Syst ems Research Institute ESRI, Redlands, California). Distribution maps were created in ArcGIS 9.2. Not all label data from museum collections were usable for

PAGE 38

38 distribution level analyses. Som e labels were unreadable and others did not give a year or specific location record. Museum staff limitations prevented other data from being recorded beyond the county level. The total number of usable C. t. bethunebakeri collection data records was N = 689. Collection recor ds were divided into three eras: the frontier era (1870 1930), development era (1930 1970), and globalization era (1970 present) as per Solecki (2001) and plotted on base maps to assess differences in specimen records over time Da ta from FLMNH specimens were also used to examine relationships between body size (wing chord length), sex, and time and location of collection. Temporal data from collection records were analyzed by decade and by s outh Florida season (wet vs. dry). Wet s easons were delineated as months having more than 3.2 inches of rainfall and 76 90F average temperatures using data from National Oceanic and Atmospheric Administration (NOAA) data for Key West, Florida ( http://www.srh.noaa.gov/key/HTML/climate/eywclimate.html ) (Figure 3 1). The months May October were scored as wet months and dry months were January March and November December. Location data were organized both on a county and a regional level. At the regional level, locations were divided into seven groups: the lower Keys, the upper Keys, Everglades National Park, the Miami area, the Ft. Lauderdale area, the Naples area, and Sanibel Island. Data were also analyzed by grouping s populations, the former group encompassing samples from Everglades National Park, the Miami area, and the Ft. Lauderdale area and the latter group being comprised of samples from all of the Florida Keys. Wing cho rd measurements (forewing length) from the historical FLMNH samples were also compared with all current BHSP samples (2002 2003 and 2005 2006) to assess changes in

PAGE 39

39 wing length by sex, season, and year. ANOVA tests were run to assess relationships between wing chord length, sex, season, year collected, and location of collection in Microsoft Excel. Genetic Sampling I selected m useum samples (FLMNH only) for genetic analysis based on two criteria. First, wide geographic and time range s were required to max imize detection of allelic diversity over space and time. A broad geographic coverage is represented by samples FLMNH_004 FLMNH_ 008 (Table 3 2 ). These five samples formed the and were from multiple locations collected from 1966 1988. Secon d, numerous specimens taken over a series of days in the same location were necessary to identify individual population signatures For the se signatures I selected three groups of samples taken from Key Largo Florida, which has I found no historical specimens recorded from Bahia Honda State Park or Key West National Wildlife Refuge, the two sites with extant C. t. bethunebakeri populations, therefore, the extensive co llecting on Key Largo provided the best possible sampling to allow me to examine the change in a single population over time. The 1940s location represent ed eight butterflies collected in Key Largo in May in the 1940s. Label data for this group was indic collector, so the exact year is unknown. The 1963 population wa s represented by 14 organisms collected over two days in April 1963 on Key Largo. The last ten samples were collected from a population of C. t. bethunebakeri on North K ey Largo on a single day in March 1979. The se three sample groups presumably represent ed genetic diversity of a temporally continuous Key Largo population and allowed me to examine temporal changes in genetic structure at this site. DNA Extractions and Mi crosatellite Analyse s DNA was extracted from museum specimens in a dedicated ancient DNA laboratory at the University of Florida

PAGE 40

40 protocols to prevent contamination with contemporary DNA Neg ative controls were run with every extraction to ensure that current material did not contaminate museum samples A maximum of fifteen extractions were carried out at a time to minimize contamination between samples All samples were well labeled to keep data matched with museum samples. I carefully removed 1 2 adult butterfly legs and labial palps from FLMNH museum specimens Extraction methods followed the Qiagen DNeasy Blood and Tissue Kit protocol with modifications for historical material from Iudi ca et al. (2001). Butterfly legs and palps were washed three times with a 250l 1X PBS for 10 minutes at 55 C for each wash After washing, I add ed 180 l Buffer ATL and ground up the sample using a sterile, disposable pestle. I then a dd ed 20l Proteinas e K mix ed by vortexing, and incubate d at 55 C in a rotating rack for 24 48 hours. The remainder of the protocol followed the Qiagen Kit, with the addition of carrier nucleotides (yeast tRNA) to force precipitation as outlined in Kishore et al. (2006). D NA was eluted one time with 30 l warm Buffer AE I conducted PCR amplification s on ancient DNA templates in another dedicated laboratory using UV sterilized equipment and laboratory bench top to prevent contamination from current materials. PCR condition s for twelve polymorphic microsatellite loci followed those of Saarinen et al. (2009), with the modification that 5L of template DNA were used per 25 L PCR reaction and there was no multiplexing or multi pooling of samples. Data for current population genetic diversity were gathered from two different generations from the BHSP population (September 2005 and June 2006) as well as from the Marquesas Islands in KWNWR (February 2008). DNA extraction and PCR amplifications on contemporary samples followed S aarinen et al. (2009). O nly one locus (Cth B119 ) was successfully amplified from historical DNA in the majority of samples and was therefore the only locus used for analysis Microsatellite data from locus

PAGE 41

41 CthB119 from contemporary populations (BHSP and KW NWR) were checked for null alleles using Micro Checker version 2.2.3 (van Oosterhout et al. 2004) and significance was assessed after a Bonferroni correction (Rice 1988). All populations were tested for Hardy Weinberg equilibrium (HWE) in Arlequin version 3.01 (Excoffier et al. 2005) and a sequential Bonferroni correction was performed across multiple comparisons As data from a single locus are presented, no linkage disequilibrium tests were performed. Data from a previous study of C. t. bethunebakeri o n BHSP showed that no linkage exists between any of the twelve microsatellite markers described for this taxon (Saarinen et al. 2009). Analyses of observed ( H O ) and expected ( H E ) heterozygosity were calculated using Arlequin 3.01 (Excoffier et al. 2005) an d an analysis of the H O /H E ratio was graphed to assess changes over time. The number of alleles ( N A ) and number of effective alleles ( n e ) were calculated using GenAlEx 6.1 (Peakall & Smouse 2006). The calculation of effective population size ( N E ) require s either knowledge of consecutive generations of allelic data or knowledge of sex ratios, family size variance, and population size fluctuations (Frankham et al. 2003; Wang & Whitlock 2003) or data from multiple loci for linkage disequilibrium estimates (E ngland et al. 2006; Tallmon et al. 2008). Given these limitations, I have instead opted to compare the effective number of alleles ( n e ) of different populations over time, calculated as p i 2 ) (3 1) where p i is the frequency of the i th allele for the population (Peakall & Smouse 200 6 ). Lastly, I used GenAlEx 6.1 to estimate genetic divergence between populations using F ST (Weir & Cockerham 1984) and R ST (Slatkin 1995) and sign ificance levels were calculated using 999 permutations.

PAGE 42

42 Results Geographic Distribution of Museum Specimens I examined 877 C. t. bethunebakeri specimens from museum collections, of which 689 had sufficient label data for inclusion in my analyses. A large primarily coastal, geographic distribution of this butterfly is evident (Figure 3 2, 3 3, 3 4) During the early part of the century, most specimens were collected from the Miami area, with a few inland and northern county records, and no records from t he Florida Keys. Sp ecimens from mainland Monroe County as well as the Florida Keys began to appear in the 1930s, but most records during the development era were still from the Miami area and Dade Co unty During the globalization era (1970 present) the F lorida Keys became the major collection location, with very few records from Dade Co unty Specimen counts for eleven counties over time (1900 1990) showed that most specimens were collected in Dade and Monroe Co unties (Figure 3 5 ). Records from Dade Cou nty ( N = 212) (which included the Miami area) were most numerous in the 1930s and 1940s, and Monroe Co unty ( N = 387) records (including all of the Florida Keys) were most numerous in the 1970s (Figure 3 6). Relationships Between Size, Sex, Year, and C olle ction Location of Museum S pecimens There was a significant difference in wing chord length between male ( N = 206) and female ( N = 132) C. t. bethunebakeri (F 1, 336 = 6.344, P = 0.012) from the FLMNH collection. Males were smaller than females, with respec tive wing chord lengths of 11.221mm 0.075 (mean SE) and 11.515mm 0.088 (mean SE). There was no significant difference in wing chord length between wet ( N = 146) and dry ( N = 192) seasons (F 1, 336 = 0.932, P = 0.335), nor was there a significant ef fect of the decade of collection on wing chord length (1920s N = 15, 1930s N =13, 1940s N = 33, 1950s N = 8, 1960s N = 58, 1970s N = 141, 1980s N = 55; F6, 316= 0.913, P = 0.486). There were no significant differences in wing chord length between the seve n

PAGE 43

43 different location regions (Lower Keys N = 58, Upper Keys N = 196, Everglades National Park N = 15, Miami area N = 56, Ft. Lauderdale area N = 7, Naples area N = 4, Sanibel Island N = 2) (F 6, 331 = 1.27, P = 0.270). There were also no significant differ ences between eastern mainland Florida populations (Everglades National Park, Miami, Ft. Lauderdale N = 78) and specimens collected in the Florida Keys ( N = 241) (F 1, 318 = 0.022, P = 0.882). Lastly, there were no seasonal size differences between the mai nland Florida populations and those in the Keys either during the dry season (F 1, 176 = 0.659, P = 0.418) or between groups in the wet season (F 1, 141 = 0.405, P = 0.526). Comparison of Body Size in H istoric al and Current P opulations The historical C. t. b ethunebakeri specimens from FLMNH were compared with butterflies from BHSP (2002 2006) to determine whether there has been a change in wing chord length between historical populations and the extant BHSP population. Initial assessment of the BHSP populati on itself showed that females were significantly larger than males (for individuals measured in the wild, not reared in the laboratory) (F 1, 124 = 21.124, P < 0.001) (Table 3 3). Individuals sampled from BHSP in the wet season were also significantly larg er than dry season BHSP adults (F 1, 235 = 19.923, P < 0.001). When the BHSP data were compared to FLMNH data, there were several significant differences in wing chord lengths. BHSP individuals (11.585 0.054) were significantly larger than historical sa mples (11.297 0.058) (F 1, 599 = 12.900, P < 0.001). Females from BHSP were also significantly larger than historical samples of females (F 1, 233 = 12.756, P < 0.001) and BHSP adults measured in wet seasons were larger than those sampled in wet seasons i n museum collections (F 1, 347 = 17. 681, P < 0.001) (Table 3 3). Historic al Genetic D iversity I attempted PCR based amplifications on DNA extracted from museum samples at eight of the 12 microsatellite loci used to genotype modern specimens. However, o nly one locus

PAGE 44

44 (Cth B119 ) successfully amplified in the majority of historical DNA samples and was therefore the only locus used for analysis Loci CthB11 and CthB117 amplified for a few samples, whereas Cth B101, CthB103, CthB106, Cth C124 and CthD7 failed to amplify for any samples I lacked sufficient quantities of historical DNA templates to test the remaining four microsatellite loci Null alleles were detected in both contemporary populations (BHSP and KWNWR) at locus CthB119 Both contemporary populat ions showed homozygote excess but no signs of scoring error or large allele drop o ut. A nalyses of allelic data showed that Hardy Weinberg equilibrium (HWE) values at locus CthB119 were different between historical and contemporary populations. A ll histor ical populations were in HWE but no contemporary populations were in equilibrium (Table 3 4). Data from the twelve microsatellite loci developed for this taxon (Saarinen et al. 2009) were considered for the contemporary populations, and these results show that seven of 12 loci were in HWE in the September BHSP 2005 population, six of 12 in HWE in the June BHSP 2006 population, and ten of 12 in HWE in the February 2008 KWNWR population after Bonferroni correction. I also plotted the ratio of H O /H E for popu lations from each time period (Figure 3 7). This ratio in locus CthB119 is consistent in the historical populations, but is lower in 2005. The 2006 and 2008 populations show ratios similar to historical levels. The additional 11 microsatellite loci anal yzed in the contemporary populations show a range of H O /H E ratios between loci and years. The average ratios (across 12 loci) in each of the contemporary populations are similar to the ratios generated from the single locus CthB119 in the historical popul ations (Figure 3 7). Even with very small sample sizes, the historical populations had generally higher effective number of alleles ( n e ) and observed levels of heterozygosity ( H O ) than the extant BHSP

PAGE 45

45 population (Table 3 4). Additionally, the effective nu mber of alleles and observed heterozygosity values remained relatively stable or even increased across the three sampling periods in the historical Key Largo population (1940s, 1963, and 1979). The BHSP population showed reductions in both the effective n umber of alleles and observed levels of heterozygosity when compared to historical levels. The KWNWR population values were between historical levels and BHSP values (Table 3 4). Four newly detected alleles (135, 165, 171, and 180) were found in very l ow frequencies in the BHSP populations. Historic al populations showed a more diverse distribution of allelic frequency than the current BHSP population (Figure 3 8). The BHSP populations had one main allele (162) and low frequencies of other alleles. Hi storic al populations had a greater distribution of allelic frequencies across the alleles sampled thus far at locus CthB119 ( Figure 3 9 ). Allele 132 was only sampled from the KWNWR 2008 population (Table 3 5). Alleles 135, 153, 165, and 171 were only fo und in one of the BHSP populations, and all were at very low frequency. Only one allele (147) found in historical samples was not found in either the BHSP or KWNWR populations. This allele was present in three of the four historical populations with rela tively high frequency (0.082 0.167) but was absent from both current populations (Table 3 5). Allele 177 was present at a higher frequency in two historical samples (0.125 0.250) as well as in the KWNWR population (0.229), and was not detected in either B HSP sample. Allele 156 was also present in the Key Largo 1963 population (0.063) and KWNWR population (0.042), but was again rare in BHSP (0.028) and only found in the 2006 samples Allele 180, however, was found in all contemporary, but no historical pop ulation samples and allele 186 was found in all historical samples and KWNWR, but neither BHSP sample Overall the BHSP and KWNWR populations shared four of 11 alleles with frequencies

PAGE 46

46 0.020 and above (156, 159, 162, and 180). The KWNWR and historical po pulations shared five of eleven alleles when measured under the same criterion (156, 159, 162, 177, and 186). Genetic divergence between historical and contemporary populations is evident from both R ST and F ST measures (Table 3 6 ). The BHSP samples are significantly different from all historical populations as well as the extant KWNWR population ( P < 0.001). The KWNWR does not show significant differences from any of the historical populations. Discussion This study showed that C. t. bethunebakeri sp ecimens were common in museum collections from the early 1900s to the 1980s. The distribu tion of these butterflies over s outh Florida over time was further verified through examination of museum records and this wide distribution showed that butterflies w ere able to exist in a variety of locations in southern Florida and that they were found in high numbers in local patches. Collection records for Dade Co. mas sive land use change and urbanization (Solecki 2001). Monroe Co. records increased in later decades, peaking in the 1970s. Decreases in natural land areas could not be correlated with changes in the number of butterflies collected (or butterfly populatio n abundance) as there were several confounding factors involved. Growing human populations not only lead to a decrease in natural land areas but an increase in pesticide use for mosquito control and agriculture. Land use changes may also indirectly impac t native fauna by increasing the ranges of invasive plant and animal species (Chapin et al. 2000; Carroll & Loye 2006), and increased competition may further depress butterfly population numbers. Land management practices including fire suppression and ex otic species removal further affect remaining habitat and none of these factors should be underestimated nor pinpointed as the sole cause of decline of native taxa.

PAGE 47

47 It was not possible to quantify issues of collector bias when assessing the relative abun dance of historical specimens. Because of the wild nature of the Florida Keys, collecting restrictions, and the inaccessibility of certain islands, it was understandable why certain regions were visited more heavily than others by collectors. Additionall y, certain areas may have been targeted for collecting in order to observe other localized or rare species. Key Largo was one historical C. t. bethunebakeri a spot simply heavily visited by lepidopterists, or a combination of the two. The island of Key Largo may have represented one of several population centers and one that was easily accessible to collectors due to the road network. Other regions, especially those within protected lands like Everglades and Biscayne National Parks, may have held additional populations of butterflies that were not as heavily sampled by collectors. Analyses of wing chord data confirmed that females were larger than males in both historical and contemporary populations. Additional differences between historical and recent specimens showed that females from BHSP were larger than historical female specimens and that individuals from BHSP sampled in the wet season were larger th an historical specimens sampled during wet seasons. Overall analyses proved that BHSP individuals were larger than historical butterflies. These results may reflect that larger adults were positively selected for over time. Larger adult butterflies may be more capable of dispersing and finding food and mates. Larger adult size has also been correlated with increased fecundity and fitness, although this is only one measurement of potential fitness (Calvo & Molina 2005). Smaller adult size in butterflies is usually associated with limited food resources during larval development or abrupt termination of availability of food in the last larval instar, both conditions resulting in early pupation and a smaller adult size (T.C. Emmel, personal communication).

PAGE 48

48 PCR reactions failed to amplify 7 of 8 tested microsatellite loci in all individuals and additional loci could not be tested due to a paucity of historical DNA templates. There were several possible reasons for the failure of PCR amplification on histori cal templates. Historical DNA template may have been too degraded or the r egion to be amplified may have been too long to be amplified in sheared DNA. These issue s may be addressed by re designing primers to amplify short er regions of interest, as has be en done for larger bar coding endeavors (Hajibabaei et al. 2006). The length of microsatellite repeats may prevent this option at some microsatellite loci, but it may be possible at others. The utilization of single nucleotide polymorphisms ( SNPs ) to ana lyze historical specimens may represent a better approach for the re creation of historical population signatures (Goldstein & DeSalle 2003 ; Morin & McCarthy 2007; Svensson et al. 2007 ). A genomic approach to conservation genetics is also becoming increas ingly possible with the advent of affordable large scale sequencing technology (Leonard 2008). The genetic differences between the contemporary and historical populations were significant, but these should be interpreted carefully as they were from a sin gle microsatellite locus. Although other studies reconstructing historical populations have utilized three or four microsatellite markers (Hutchinson et al. 2003 ; Harper et al. 2006) the single locus used in this study did contain eleven alleles including several rare alleles. These alleles are useful in detecting genetic drift and as such, provide useful data (Frankel & Soul 1981). Measurements of HWE showed that historical populations remained in HWE over time and that the contemporary populations wer e in HWE at some loci. The temporal sampling employed by this study supports the conclusions that historical populations contained more genetic diversity than the extant BHSP population. This conclusion is supported by several different genetic measurem ents ( N A n e, H O, H E ) all of which suggest that

PAGE 49

49 historical populations were more diverse than either contemporary population. The historical populations showed higher observed heterozygosity than either of the contemporary (BHSP and KWNWR) populations. The number of effective alleles of historical populations was also higher, even with the limited sampling. Current overall allelic frequencies were reduced, with the BHSP population showing higher frequencies of fewer alleles. Because of the intensive sa mpling of the BHSP population, there was a greater presence of low frequency alleles. There were 12 total alleles at CthB119, and three of these were only found in the BHSP population with 0.014 frequencies. These may represent novel mutations as they we re not found in the historical populations or in the KWNWR population, although it is equally possible that they were simply not sampled in those groups. Only one allele was found in historical populations but was absent from both BHSP and KWNWR populatio ns. This allele (147) was historically present at a relatively high frequency (0.083 0.167). These data show that although allelic frequencies have changed over time, only one allele has not been found and is presumed to be lost from contemporary populat ions Other studies comparing historically common and currently endangered populations have shown a greater loss of alleles over time (Harper et al. 2006). The KWNWR population, found on the Marquesas Islands, had higher genetic diversity measures than t he BHSP population, and although little is known of this new population, it does lend encouragement for future conservation efforts. Both historical and contemporary populations show evidence of a metapopulation structure with interacting sub colonies (E. V. Saarinen and J.C. Daniels unpublished data). The metapopulation structure on KWNWR is more extensive than the metapopul ation structure present on BHSP, potentially providing better methods for retaining genetic diversity. Further review of population structure and the

PAGE 50

50 importance of metapopulation dynamics in this taxon are forthcoming and will be incorporated into management strategies. Overall patterns of genetic diversity in the BHSP population (mean overall observed heterozygosity of 39.5%) are sim ilar or slightly lower when compared with other non migratory considered endangered show a range of 40.97 48.37% mean overall observed heterozygosity (Keyghobadi e t al. 2002; Fauvelot et al. 2006 ; Sarhan 2006). This study is not the first to show high levels of genetic variation in an endangered taxon The Mauritius kestrel (Groombridge et al. 2000) and the greater one horned rhinoceros (Dinerstein & McCracken 199 0) have both recovered from extreme bottlenecks and show high hetero zygosity without the addition of new genetic material. Studies of an isolated population of the brown bear Ursus arctos also demonstrated high heterozygosity despite a severe, well docume nted bottleneck (Hartl & Hell 1994). All of these studies highlight the fact that bottleneck events resulting in low population numbers and the reduction of genetic diversity do not necessarily doom endangered taxa to extinction and that these taxa should not be written off as lost causes or denied funding on such grounds. This research speaks to the importance of studying historical populations in making assumptions of current populations (Leonard 2008; Shepherd & Lambert 2008), but grand assumptions c annot and should not be made from the frequency of alleles at a single locus. Furthermore, data characterizing historical populations should be calculated from numerous museum specimens to better estimate levels of historical diversity. This study showe d that museum collections represent a wealth of information that may prove invaluable for understanding the distribution of rare taxa and may be useful for conservation planning. Butterflies are particularly well represented in museum collections and

PAGE 51

51 sign ificant taxonomic and life history information is available for Lepidoptera making it ideal for studies such as this one. Similar historical approaches may be possible for a wide range of taxa that have small extant populations complemented by large muse um holdings from extinct populations. The extant C. t. bethunebakeri populations are geographically restricted to a few small islands in the Florida Keys and do not include the full genetic diversity formerly present when populations were more widespread and abundant. I have demonstrated here that the historical distribution and genetic variation of extinct populations can be reconstructed by careful and thorough examination of museum specimens. The habitat changes that have occurred in s outh Florida are neither unique nor reversible, but that does not mean that species conservation efforts cannot be successful in this region. Between the two contemporary populations, C. t. bethunebakeri has retained a significant amount of genetic diversity from its hist orical values. Through active and timely conservation efforts, this taxon holds promise to be restored to its once common status. This highlights the need for conservation of the current island populations as they represent the last members of the taxon and the remainder of historical diversity. Conservation efforts for this taxon should not be abandoned on the grounds of reduced molecular diversity or adaptive ability, as the genetic variation shown here is evidence of evolutionary potential. Other tax a with reduced population sizes should be genetically assessed and compared to historical populations through the analysis of museum specimens whenever possible With proper sampling and methodologies, such projects can assess historical levels of gene fl ow and estimate appropriate levels for contemporary populations. The utilization of museum label data and historical genetic analyses can also be incorporated into conservation studies in recommending new areas for reintroduction (Ross et al. 2006) Thes e efforts may help restore endangered species to common status by

PAGE 52

52 giving historical perspective, offering guidelines for reintroduction efforts and gene tic management, and providing a context for interpret ing present day genetic diversity.

PAGE 53

53 Table 3 1. H uman population estimates in south Florida counties. Source: U.S. Census Bureau 2008. Year Monroe Dade Florida total 1900 18,006 4,955 528,542 1910 21,563 11,933 752,619 1920 19,550 42,753 968,470 1930 13,624 142,955 1,468,211 1940 14,078 267,739 1,897,414 1950 29,957 495,084 2,771,305 1960 47,921 935,047 4,951,560 1970 52,586 1,267,792 6,789,443 1980 63,188 1,625,781 9,746,324 1990 78,024 1,937,094 12,937,926 2000 79,589 2,253,362 15,982,3 78 2007 73,223 2,387,170 18,251,243 Figure 3 1. Weather data averages for Key West, Florida

PAGE 54

54 Table 3 2. Museum specimens of C. t. bethunebakeri sampled for genetic diversity. Samples FLMNH_009 FLMNH_016 had data missing from the mus eum label. The data listed on the museum specimen labels are given in the table. Population group relates to the assignment of specimens for statistical analyses. Sample ID Location Date Population Group FLMNH_004 Middle Cape Sable, Everglades Nat'l. Pa rk April 7, 1966 Mixed FLMNH_005 Biscayne Nat'l. Monument May 10, 1973 Mixed FLMNH_006 Plantation Key March 1, 1981 Mixed FLMNH_007 Little Pine Key April 26, 1988 Mixed FLMNH_008 Big Pine Key May 29, 1979 Mixed FLMNH_009 Key Largo May 17 194_ 1940s F LMNH_010 Key Largo May 17 194_ 1940s FLMNH_011 Key Largo May 17 194_ 1940s FLMNH_012 Key Largo May 18 194_ 1940s FLMNH_013 Key Largo May 18 194_ 1940s FLMNH_014 Key Largo May 18 194_ 1940s FLMNH_015 Key Largo May 18 194_ 1940s FLMNH_016 Key Largo May 31 194_ 1940s FLMNH_017 Key Largo Key April 2, 1963 1963 FLMNH_018 Key Largo Key April 2, 1963 1963 FLMNH_019 Key Largo Key April 1, 1963 1963 FLMNH_020 Key Largo Key April 1, 1963 1963 FLMNH_021 Key Largo Key April 1, 1963 1963 FLMNH_022 Key Largo Key April 1, 1963 1963 FLMNH_023 Key Largo Key April 2, 1963 1963 FLMNH_024 Key Largo Key April 2, 1963 1963 FLMNH_025 Key Largo Key April 2, 1963 1963 FLMNH_026 Key Largo Key April 2, 1963 1963 FLMNH_027 Key Largo Key April 2, 1963 1963 FLMNH_028 Ke y Largo Key April 2, 1963 1963 FLMNH_029 Key Largo Key April 1, 1963 1963 FLMNH_030 Key Largo Key April 1, 1963 1963 FLMNH_031 North Key Largo March 15, 1979 1979 FLMNH_032 North Key Largo March 15, 1979 1979 FLMNH_033 North Key Largo March 15, 1979 1 979 FLMNH_034 North Key Largo March 15, 1979 1979 FLMNH_035 North Key Largo March 15, 1979 1979 FLMNH_036 North Key Largo March 15, 1979 1979 FLMNH_037 North Key Largo March 15, 1979 1979 FLMNH_038 North Key Largo March 15, 1979 1979 FLMNH_039 North Key Largo March 15, 1979 1979 FLMNH_040 North Key Largo March 15, 1979 1979

PAGE 55

55 Figure 3 2. Museum specimens of C. t. bethunebakeri collected from 1900 1930, the frontier era as per Solecki (2001).

PAGE 56

56 Figure 3 3. Museum specimens of C. t. bethunebakeri collected from 1931 1970, the d evelopment era as per Solecki (2001).

PAGE 57

57 Figure 3 4. Museum specimens of C. t. bethunebakeri collected from 1971 present, the g lobalization era as per Solecki (2001).

PAGE 58

58 Figure 3 5. Number of C. t. bethunebakeri specimens deposited in five major museum collections by Florida C ounty from 1900 1990 Figure 3 6. Number of C. t. bethunebakeri specimens deposited in five major museum collections for Dade and Monroe Counties, Florida

PAGE 59

59 Table 3 3. Results of ANOVA analyses co mparing historical (1927 1988 FLMNH) and contemporary (2002 2006 BHSP) C. t. bethunebakeri wing chord lengths (mm mean standard error ) over several variables. P values are italicized when significant at P < 0.05. Variable BHSP BHSP (mm) FLMNH FLMNH ( mm) P value Female 114 11.868 0.077 121 11.440 0.091 0.0004 0 Male 162 11.386 0.069 204 11.212 0.075 0.098 00 Wet season 203 11.687 0.059 146 11.273 0.083 0.00003 Dry season 34 10.971 0.177 179 11.317 0.082 0.089 00 Location (BHSP vs. h istorical Lower Keys) 276 11.585 0.054 58 11.462 0.141 0.357 00 Total 276 11.585 0.054 325 11.297 0.058 0.0004 0 Table 3 4. Overview of molecular diversity measures for C. t. bethunebakeri populations at microsatellite locus CthB119 N n umbe r of samples with positive PCR results; N A number of observed alleles; n e e ffective number of alleles; H O observed heterozygosity; H E expected heterozygosity, and P value of Hardy Weinberg equilibrium test significance. Tests are assessed for signific ance after a sequential Bonferroni correction; values not in equilibrium are in italics Population N N A n e H O H E P value Mixed 4 5.000 4.000 0.500 0.857 0.075 Key Largo 1940s 6 4.000 2.769 0.500 0.697 0.416 Key Largo 1963 8 5.000 3.459 0.500 0.75 9 0.126 North Key Largo 1979 6 4.000 3.789 0.667 0.803 0.043 BHSP Sept. 2005 24 4.000 1.417 0.083 0.301 0.000 BHSP June 2006 36 7.000 1.427 0.162 0.275 0.002 KWNWR Feb. 2008 16 4.000 2.798 0.250 0.663 0.000

PAGE 60

60 Figure 3 7. Ratio of H O /H E valu es for historical Key Largo (yellow) and contemporary BHSP (green and red) and KWNWR (blue) populations of C. t. bethunebakeri C ircled points are from locus CthB119. In contemporary populations, squares represent the remaining 11 individual microsatellite loci and the average ratio per year (across all 12 loci) are shown as a square with an asterisk Figure 3 8. Frequency of alleles at microsatellite locus CthB119 in all C. t. bethunebakeri populations. Data are from museum samples from Key Largo, Flo rida and current samples from BHSP, Bahia Honda State Park, Florida

PAGE 61

61 Figure 3 9. Allelic frequencies at microsatellite locus CthB119 of historical populations of C. t. bethunebakeri Table 3 5. Frequency of alleles at microsatellite locus CthB119 in historical and current populations of C. t. bethunebakeri Allele, size in base pairs; Populations ( N ) of Mixed Florida origin; KL, Key Largo, Florida; BHSP, Bahia Honda State Park, Florida, and KWNWR, Key West National Wildlife Refuge, Florida. A single historical allele is in italics Allele Mixed (4) KL 1940s (6) KL 1963 (8) North KL 1979 (6) BHSP 2005 (24) BHSP 2006 (36) KWNWR 2008 (16) 132 0.000 0.000 0.000 0.000 0.000 0.000 0.125 135 0.000 0.000 0.000 0.000 0.000 0.014 0.000 147 0.125 0.083 0.000 0.167 0.000 0.000 0.000 153 0.000 0.000 0.000 0.000 0.083 0.000 0.000 156 0.000 0.000 0.063 0.000 0.000 0.028 0.042 159 0.125 0.083 0.250 0.250 0.063 0.056 0.063 162 0.125 0.417 0.438 0.333 0.833 0.833 0.583 165 0.000 0.000 0.000 0.000 0.000 0.014 0. 000 171 0.000 0.000 0.000 0.000 0.000 0.014 0.000 177 0.250 0.000 0.125 0.000 0.000 0.000 0.229 180 0.000 0.000 0.000 0.000 0.021 0.042 0.042 186 0.375 0.417 0.125 0.250 0.000 0.000 0.042

PAGE 62

62 Table 3 6. Pairwise population differentiation betwee n each of the historical groups and contemporary populations of C. t. bethunebakeri F ST values (above the diagonal) and R ST values (below the diagonal) with respective levels of significance; P < 0.05, ** P < 0.001 Mixed 1940 1963 1979 BHSP KWNWR Mixed 0.000 0.000 0.000 0.420** 0.039 1940 0.000 0.000 0.000 0.287** 0.033 1963 0.000 0.000 0.000 0.350** 0.000 1979 0.000 0.000 0.000 0.378** 0.004 BHSP 0.059 0.216 0.542 ** 0.535 ** 0.305** KWNWR 0.006 0.000 0.000 0.000 0.462 **

PAGE 63

63 CHAPTER 4 GENETICS OF WILD POPULATIONS Introduction The Miami b lue butterfly ( Cyclargus thomasi bethunebakeri ) is a small, brightly colored butterfly endemic to southern Florida with two su bspecies occurring in the Bahamas and several species found throughout the Caribbean (Smith et al. 1994). This taxon has received considerable attention in the mainstream media due to its charismatic name and history. Once widespread and locally abundant in the south Florida mainland, C. t. bethunebakeri has been eliminated from much of its former range due to ever expanding urbanization and the associated loss of coastal habitat (Calhoun et al. 2002) V erified observations rapidly declined through the 20 th century and by the early 1990s C. t. bethunebakeri was presumed to be extirpate d Fortunately, the taxon was rediscovered on 29 November 1999 in a small population of less than 100 individuals within the boundaries of Bahia Honda State Park (BHSP) in the Lower Florida Keys ( Ruffin & Glassberg 2000 ). In December 2007, a second colon y of C. t. bethunebakeri was discovered on the Marquesas Islands in Key West National Wildlife Refuge (KWNWR), 50 km west of Key West (Cannon 2007). The two remaining populations are subject to an increased threat of extirpation from a variety of environme ntal and human factors including hurricanes, drought, predation illegal collecting, loss of genetic diversity, habitat alteration, and mosquito control practices. The BHSP population and the taxon as a whole are therefore in an extremely tenuous situation There exist no known additional neighboring populations capable of re colonizing habitat areas should they be impacted The BHSP and KWNWR populations are over 70 km apart, much of which is over open water, and it is unlikely that such a small butterfly could travel this distance in its short adult life span of less than a week (FWC 2003) The resulting low census numbers and

PAGE 64

64 population vulnerability make C. t. bethunebakeri one of the most critically imperiled insects in the United States Reduced gen etic diversity and reduction of habitat area are two potentially influential factors leading to species extinction (Avise 1996). It has been proposed that demographic factors may affect endangered species before the genetics and deleterious effects of inb reeding even have a chance to take effect ( Lande 1995), although other studies have shown that genetic factors impact the rate of extinction of endangered populations (Newman & Pilson 1997; Saccheri et al. 1998; Spielman et al. 2004). In this study, I det ermine and compare the levels of genetic diversity in both the BHSP and KWNWR populations and look for evidence of a genetic bottleneck. I also assess how diversity changes over time within the BHSP population. Lastly, I examine gene flow between the hab itat patches (subpopulations) of the BHSP metapopulation. These efforts will be useful in conservation efforts focused on this taxon as well as other endangered butterfly taxa of the region. I have selected several time scales over which to assess geneti c diversity in the BHSP population. I have done so to 1) increase statistical power by increasing sample sizes when samples can be pooled according to biologically sound reasoning ( as recommended by Frankham et al. 2003) and to 2) highlight any fine scale changes in molecular diversity that may occur throughout seasonal changes and with the flux of population numbers. The BHSP population exists a s a metapopulation, with butterflies and larvae occupying some but not all suitable habitat patches on the isla nd. The level of gene flow and connectivity between discrete habitat patches is currently unknown. I therefore conducted analyses of BHSP at several spatial scales to uncover the different levels of gene flow. I analyzed BHSP as a single population (wi th no subdivision), as individual colonies occupying discrete habitat patches (as several

PAGE 65

65 groups acting in a metapopulation structure), and as a division of clumped colonies versus other, more spatially distant colonies. In this manner, I hope to analyze genetic measures at all biologically relevant time and spatial scales. I used traditional genetic diversity measures and population level statistics to characterize the genetic diversity of the taxon. Loss of heterozygosity over time may indicate that a p opulation is undergoing inbreeding or otherwise losing evolutionary potential (Frankham et al. 2003). I also completed testing of Hardy Weinberg equilibrium (HWE), linkage disequilibrium, gene flow ( F ST and R ST ), and effective population size ( N e ). Wrig ht (1978) suggested guidelines for interpreting F ST values, but F ST values vary considerably among taxonomic groups and studies (review in Frankham et al. 2003). Inbreeding coefficients are also calculated according to methods of Weir & Cockerham (1984) t o compare relative levels of inbreeding between patches and populations. Methods Habitat Maps I took Global Positioning System (GPS) coordinates (WGS 1984 projection) and notes of habitat condition, quality, and extent of the locations of C. t. bethunebak eri colonies in Bahia Honda State Park (BHSP), Florida in September 2005 and April October 2006. I transformed these data into a geodatabase using ArcGIS 9. 2, Environmental Systems Research Institute Redlands ESRI, Inc., Redlands, California ). The poin t locations were then overlaid on a satellite photo of BHSP taken in 2005 obtained through the Florida Geographic Data Library ( http://www.fgdl.org/ ). I then digitally drew habitat extent onto fine scale aerial maps b y referencing my field notes (Figure 4 1).

PAGE 66

66 Sampling I developed a relatively non invasive method to assess molecular diversity in the Bahia Honda State Park (BHSP) and Key West National Wildlife Refuge (KWNWR) population s I gently capture d butterflies i n a net and remove d a 2 mm 2 wing fragment from the hind wing with small forceps ( Figure 4 2) Care was taken to clean forceps to prevent cross contamination between samples. The removal of such a small amount of wing material did not adversely affect but terfly behavior and butterflies were observed courting, mating, and feeding shortly after the removal of wing fragment. Wing tissue was preserved in 90 100% ethanol Sample vials were kept cool and in the dark until they were returned to the laboratory 8 0 C freezer for storage. Wing fragments for molecular analyses f rom the BHSP population of butterflies were sampled from September 2005 thru October 2006 ( N = 114) and represent multiple generations of this population ( Table 4 1). Individual months r epresent temporally discrete generations as these butterflies were sampled over a five day period in each month. The September 2005 samples were the only representation of BHSP in 2005. In 2006, data were pooled for samples ta ken in concurrent months (Ap ril and May, June and July, and September and October) to increase sample size and the power of statistical analyses. Only data from concurrent months were pooled as this represents a biologically probable breeding unit for this taxon that has a new gener ation approximately every 30 40 days ( FWC 2003 ). Annual data were pooled to show overall changes in diversity between years. All data from April 2006 October 2006 were pooled for 2006 results. The KWNWR population was sampled once in February 2008 and r epresents a discrete breeding population. Two samples were taken from Boca Grande Key and 27 samples were taken from Marquesas Key. Owing to the small sample size from Boca Grande Key, only samples from the Marquesas Keys were considered for genetic anal ysis of KWNWR.

PAGE 67

67 DNA Extractions and Microsatellite Amplification I extracted DNA from 2 mm 2 wing fragments using Qiagen DNeasy Blood and Tissue Kits modifications owing to the small size of the wing material I was extracting from ( Appendix A). I stored DNA extractions at 20 C until I proceeded with PCR amplifications of the twelve polymorphic microsatellite loci. All PCR reactions follow the protocols outlined in Saarinen et al. (2009). Be tween 5 10% of all BHSP samples were re genotyped at nine microsatellite loci to det ermine genotyping error rates for each locus All private alleles were individually assessed for validity and that they were not the result of genotyping or scoring error. Statistical A nalyses Microsatellite d ata were organized in Microsoft Excel and fo rmatted using Convert (Glaubitz 2004) to be used in other software packages. Micro Checker version 2.2.3 ( v an Oosterhout et al. 2004) was used to test for the presence of null alleles as problems of allelic drop o ut have been identified for PCR amplified products in Lepidopteran microsatellites ( Zhang 2004 ) as well as for studies utilizing noninvasive genetic samples of low quantity (Taberlet et al. 1999) I used Arlequin version 3.01 (Excoffier et al. 2005) to estimate the observed and expected heterozyg osities ( H O and H E ) and Hardy Weinberg equilibrium was tested for with an exact test using a Markov chain with 1,000,000 forecasted chain length and 100,000 dememorization (burn in) steps. Hardy Weinberg equilibrium significances were evaluated after a se quential Bonferroni correction (Rice 1988). I used FSTAT version 2.9.3.2 to test for linkage disequilibrium (Goudet 2001) and confirmed results with GDA 1.1 (Lewis & Zaykin 2001) I used GenAlEx 6.1 ( Peakall & Smouse 2006 ) to estimate the effective numbe r of alleles per locus and analyses of molecular variance across generational samples I used FSTAT to calculate allelic richness, a measure of the number of alleles independent of sample size

PAGE 68

68 (Goudet 2001) The independence from sample size allows for t he comparison of allelic richness values between generations or populations of varying size. F statistics, R ST values, and inbreeding coefficients were calculated in GenAlEx 6.1 (Peakall & Smouse 2006). Both F ST and R ST values were calculated for estimat es of gene flow as each measure is based on a different assumption of marker mutation rate. R ST has been recommended for microsatellite data as it follows a Stepwise Mutation Model (SMM) and may detect more population differentiation than F ST (Slatkin 199 5). F ST ( Wright 1965 ) uses allele identity rather than size corresponding to the Infinite Alleles Model (IAM) to measure population subdivision. Both F ST and R ST were calculated and tested for significance over 9,999 permutations. Sequential Bonferroni corrections were performed on all data with multiple comparisons to correct for Type 1 error (Rice 1988). The program Bottleneck (Cornuet & Luikart 1996) was used to look for genetic signatures typical of a population that has undergone a bottleneck. T he SMM, IAM, and TPM (Two Phase Model) of microsatellite evolution were calculated and 1000 iterations were run for each mod el For the latter model, I set the proportion of SMM at 70% and ran allowing 30% variance for TPM. I tested individual, discrete breeding populations (i.e. those samples collected within the same week ensuring no overlap of generations) that met the sample size criterion of 15 40 individuals (Cornuet & Luikart 1996). To ensure that the program worked with my data, I also tested a k nown population bottleneck using data from the captive breeding colony of C. t. bethunebakeri housed at the University of Florida. The October 2006 captive colony (CC 10 2006) had been allowed to breed without the introduction of new genetic material for three four generations, thus representing a bottlenecked population. I used the program NeEstimator ( Peel et al. 2004 ) to estimate the effective population size ( N E ) of the BHSP population. This program uses a moments based temporal model ( Waples 1989) t o estimate the effective population size

PAGE 69

69 over time. For these analyses, consecutive generations were required to estimate effective population sizes. Lastly, I used the program Structure 2.0 (Pritchard et al. 2000) to look for genetic structure across ge nerations and in the two C. t. bethunebakeri populations. This program assumes no specific mutation model and is appropriate for unlinked microsatellite data ( Pritchard et al. 2007 ). I examined all allelic data from BHSP ( N = 114) for population structur ing across generations. I then looked for population differentiation between BHSP and KWNWR by analyzing the September 2005 BHSP population and the February 2008 KWNWR population. The September 2005 BHSP sample was chosen for comparison with the KWNWR po pulation as both populations were similar in sample size. I used the population admixture model implemented in Structure with correlated allele frequencies among populations as recommended by Falush et al. (2003) for situations with subtle population stru cture. No prior population information was included in these analyses. I ran 5 replicates at each estimated population number ( K = 1 to K = 5) with a burn in of 50,000 iterations and collection of data over 500,000 iterations. Exploratory testing using the no admixture model showed similar results to the admixture model. The September 2005 BHSP and the February 2008 KWNWR population s were also run with these model parameters additionally allowing prior population information to assist with clustering. E stimation of Microsatellite Mutation Rate My goal in investigating the mutation rate of the microsatellite markers employed in this study was t o determin e whether a high mutation rate wa s responsible for observed genetic diversity. Microsatellite mutation rates have been measured across taxa with rates ranging from 10 6 in Drosophila mutation accumulation lines (Schug et al. 1998) to 10 3 in Gastropoda (Gow et al. 2005) and Insecta (Crozier et al. 1999). Extensive microsatellite surveys of the human genom e, utilizing over 5,000 microsatellites revealed a mutation rate of ~10 4 (Weissenbach et al.

PAGE 70

70 1992 ; Dib et al. 1996 ; Whittaker et al. 2003). A recent, extensive survey of the wheat genome using a gene genealogy approach showed a mean mutation rate of 1.68 x 10 3 across 21 microsatellite markers when experimental wheat populations were tracked over multiple generations (Raquin et al. 2008). My goal wa s to determine whether the microsatellite markers used in this study fe ll within this range of mutation rat es, or whether an elevated rate of microsatellite mutation exists and was responsible for an inflated value of molecular diversity. To determine the overall mutation rate of the 12 microsatellite markers used in this study, I genotyped two parents and 21 of their offspring. This family line was raised apart from other individuals so that there would be no mixing between family groups that could lead to erroneous results. By genotyping two parents and 21 offspring at 12 loci, I achieved 504 comparisons in which to look for mutations. As microsatellites are passed on by Mendelian inheritance, any change in allele size from parent to offspring can be considered a mutation. I manually confirmed all genotyping results within the program GeneMapper version 4. 0 (ABI) and re genotyped any results that yielded a mutation three times to confirm results were not a result of scoring error. Results Molecular Diversity Tests for linkage disequilibrium of the 12 microsatellite markers showed that loci were unlinked ( a fter a sequential Bonferroni c orrection) Therefore, the data represented here are from 12 independent loci in the C. t. bethunebakeri genome. Null alleles were inferred at seven loci (CthB101, CthB106, CthB115, CthB117, CthB119, CthC12, and CthC116) du e to the presence of homozygote excess. Micro Checker ( v an Oosterhout et al. 2004) showed no evidence of large allele drop out or scoring error due to stuttering I further verified genotyping results by re genotyping 5 10% of samples at nine loci (Table 4 2 ). Samples were accurately re

PAGE 71

71 genotyped at loci between 58.33 100.00%. In loci where successful re genotyping rates were less than 100%, samples were re genotyped a third time to verify results All statistical analyses were performed on this correc ted dataset. My first objective was to assess the molecular diversity within the Bahia Honda State Park (BHSP) population. The effective number of alleles ( n e ) in the BHSP remained relatively constant over time, at both a monthly (generational) and annu al scale (T able 4 3). Allelic richness was also stable over time in BHSP, with values ranging from 2.988 3.121 when averaged across the 12 microsatellite loci from September 2005 October 2006. Allelic richness in KWNWR was 3.790 in February 2008 highe r than BHSP values There were seven loci in Hardy Weinberg equilibrium (HWE) in each month sampled, but the individual loci in HWE differed between months (Table 4 4) Locus CthC124 was monomorphic in all BHSP generations except the June 2006 population (when it was in HWE ). The population on KWNWR is in HWE at 10 of 12 loci. Mean inbreeding coefficients were estimated across 12 microsatellite loci for each sampled group (subpopulation) over time ( Figure 4 3 ). These values are equivalent to the f of Weir & Cockerham (1984) as they were calculated on single populations from allelic frequency data. I calculated values in this manner to assess inbreeding within individual subpopulations. The most negative value ( 0.289, SE 0.076) was found in the Butto nwood 2006 group. The most positive values ( 0.285, SE 0.146 ) Bayside 2005, may denote inbreeding or undetected null alleles, both likely options given the nature of the subpopulation and the markers used to study them. Most subpopulation s analyzed (thre e of 11) showed a positive f value, denoting the possibility of inbreeding or null alleles, although individual subpopulation f values did not vary greatly over time. F IS values measure the heterozygosity reduction in an individual relative to its

PAGE 72

72 subpopu lation, and these values were consistent for the 2005 and 2006 BHSP populations; F IS =0.112 (SE 0.104) and F IS =0.120 (SE 0.070), respectively. Results from the program Bottleneck (Cornuet & Luikart 1996) do not show significant evidence of a genetic bottlen eck in any of the BHSP generations analyzed (Table 4 5 ). The Wilcoxon sign ed rank test s (Luikart & Cornuet 1997) for the September 2005, May 2006, and June 2006 samples failed to show excess heterozygosity under the three models (IAM, TPM, and SMM) Addi tionally, these three BHSP generations showed an L shaped allele frequency distributi on that is characteristic of a population in mutation drift equilibrium. Populations that have undergone a bottleneck show a mode shift in their allelic distribution. On ly the captive colony population (CC 10 2006) that h ad undergone a known bottleneck showed evidence of a mode shift. This population also showed heterozygosity excess under the IAM model. The KWNWR population is the only other population to show heterozy gote excess (also under the IAM, P = 0.003) but showed no evidence of a mode shift. Evidence of heterozygote deficiency was significant under the SMM in all BHSP populations. The CC 7 2006 population showed deficiency under the TPM and the SMM (Table 4 5 ). Effective population sizes of the Bahia Honda State Park population were evaluated using the program NeEstimator (Peel et al. 2004). Results were generated by comparing allelic frequencies in conse cutive generations in a moments based temporal model. These data are summarized in Table 4 6 and show an estimate of 26.8 (12.5 68.8 95% CI) for the effective population size between May 2006 to Octob er 2006 for the BHSP population. Estimates could not be calculated for the KWNWR population as I have measur ements from only one generation. Gene F low in Bahia Honda State Park Analyses of microsatellite frequencies were also used to assess gene flow between habitat patches ( Figures 4 4 and 4 5 ). No values were computed when only one sample was available in

PAGE 73

73 a subpopulation (no values are given for the Buttonwood population in 2005). When considering individual subpopulations on the South end of BHSP, only minor population subdivision is observed, with most pairwise values between neighboring segments at or ver y near 0.000. No F ST or R ST values are significant in 2005, but they are significant in 2006 (Figure 4 4). F ST and R ST values on the South end are higher in 2006 than in 2005, possibly due to the absence of individuals from the Bridge population. More p opulation subdivision is evident when considering the island of BHSP as a whole (Figure 4 5). The Buttonwood and Sandspur subpopulations are well linked by the Main Road population ( F ST and R ST effectively 0.000). The South end colonies show some divisio n from the other populations (Buttonwood, Main Road, and Sandspur). Again, many significance values are above P = 0.05 when comparing levels of gene flow. Interestingly, the smallest of these differences is between the South end colonies and Sandspur, wh ich is the farthest colony by geographic distance. These values still showed relatively little differentiation between subpopulations. Analyses of AMOVA showed high levels of variation among individuals in both 2005 and 2006. In 2005, 77% of molecular v ariance was among individuals, 23% within individuals, and 0% among populations. In 2006, 77% was among individuals, 20% within individuals, and 3% among populations when individual habitat patch populations were considered for each year. Analysis of al l BHSP allelic data using the program Structure (Pritchard et al. 2000) is summarized in Table 4 7. The highest Ln P ( X / K ) is for K = 2 (after 5 independent runs), suggesting that the BHSP data are derived from two source populations. Further analyses of the data show no structuring due to time of c ollection, sex, or location. The two populations show significant separation, but no immediate explanation for this division of BHSP data is evident (Figure 4 6).

PAGE 74

74 Estimations of genetic differentiation between BHSP and KWNWR (utilizing BHSP September 2005 and KWNWR February 2008 datasets, respectively) yield ed F ST = 0. 191 ( P = 0.0001 ) and R ST = 0. 275 ( P = 0.0001) after 9,999 permutations. Analysis from Structure 2.0 (Pritchard et al. 2000) showed that the esti mated log likelihood values increased with increasing K ( Table 4 8) This situation is described in Evanno et al. (2005) as a problem with the program Structure where likelihood values may increase with increasing K and continue to do so after the K has been reached. Additional analyses using the same model parameters but including prior population information about collection origin (BHSP or KWNWR) show the highest Ln P( X / K ) is for K = 2 ( Table 4 8) and show a clear distinction between the KWNWR and BHSP populations (Figure 4 7). Mutation Rate of Microsatellites All private alleles fit the expected mutation pattern and the 504 comparisons of microsatellites revealed one mutation at one locus. CthB103, a locus with a trinucleotide repeat, an of This mutation was therefore not generated in a step wise manner. It is unknowable whether the mutation occurred from the allele expanding from 153 to 162 bp or wheth er it shrank from 233 to 162 bp. This single mutation leads to a mutation rate of 1.98 x 10 3 This result is very comparable to the extensive survey of the wheat genome showing an average rate of 1.68 x 10 3 (Raquin et al. 2008). At the mutation rate o f the wheat genome, a mutation is expected every 595 comparisons on the average, and every 201 comparisons at the high end (4.97 x 10 3 ). Discussion Overall patterns of genetic diversity in the BHSP populations (mean overall observed heterozygosity of 39 .5%) are similar to or slightly lower than other non migratory butterfly

PAGE 75

75 endangered show a range of 40.97 48.37% mean overall observed heterozygosity (Keyghobadi et al. 2002; Fauvelot et al. 2006 ; Sarhan 2006). The federally endangered Karner blue butterfly ( Lycaeides m elissa samuelis ) exhibits 28.25% mean observed heterozygosity (Anthon y et al. 2001 ) but the endangered Polyommatus bellargus registers a mean obser ved heterozygosity value higher than non threatened species, 53.2% ( Harper et al. 2000). Perhaps concordantly, diversity in C. t. bethunebakeri was higher in the KWNWR population than the BHSP population (mean observed heterozygosity of 51% versus 39.5%). Allelic richness was also higher in KWNWR, and the KWNWR population may represent an important source of variation to be considered for future conservation efforts for this taxon. Heterozygosity may still be underestimated at some C. t. bethunebakeri mi crosatellite loci due to the presence of allelic dropout, even after my efforts to correct for this with re genotyping. This underestimation has multiple implications, such as the under representation of the number of loci in Hardy Weinberg equilibrium. Furthermore, as the program Bottleneck (Cornuet & Luikart 1996) uses the signal of heterozygote excess to interpret a population bottleneck, there may have been a population bottleneck that was undetected due to genotyping error. This shows the importance of re genotyping to ensure that results are correctly interpreted, as denoted by Pompanon et al. (2005). The validation of genotypic data is especially important in Lepidopte ra, as this group shows a high rate of PCR failure due to changes in flanking se quences (Zhang 2004). Because I re ran PCRs and re genotyped a percentage of samples that showed discord between genotyping scores, I am confident in the results I present. Lastly, the program Micro Checker detected null alleles at seven loci due to homo zygote excess. These may not be null alleles but rather representative of the reduced diversity of the population.

PAGE 76

76 Most researchers have assumed that the BHSP population has undergone a severe population bottleneck ( FWC 2003) but the data presented here show no signature of a bottleneck. It may be more likely that this C. t. bethunebakeri population has been subjected to a more historic bottleneck, one that is not detectable by my methods Only recent bottlenecks, within 4 N E generations, are able to b e dete cted by the program Bottleneck (Cornuet & Luikart 1996 ). If the effective population size is approximately 27 (as per NeEstimator results), then a bottleneck could only be detected if it occurred within 108 generations. As the C. t. bethunebakeri g eneration time is short (one generation per 30 40 days) it is likely that the population could have undergone a bottleneck in the 1980s and 1990s when populations were in severe decline (Calhoun et al. 2002), and then allelic levels returned to mutation dr ift equilibrium in the following months and years. The only wild population to show a sign of bottleneck is from KWNWR, which may be a more recently founded population. The new colonization of this area would certainly represent a bottleneck event, and c onfirms the ability of the program Bottleneck to detect such an event. The utility of this program is further confirmed as it has verified the bottleneck event that occurred in the captive colony of butterflies in October 2006. Wild butterfly stock had last been incorporated into the captive colony in June 2006, thus, four generations had elapsed with no detection of a bottleneck at this juncture highlights the import ance of infusing captive colonies with wild genetic stock on a regular basis. The detection of all bottlenecks in this study was only significant under the IAM, denoting that the TPM and SMM are poor at detecting past bottlenecks in these populations. Si milar results were found in the Adonis blue butterfly Polyommatus bellargus (Harper et al. 2003), highlighting the need to utilize all mutation models

PAGE 77

77 so as not to overlook significant results. Additionally, my analysis of microsatellite mutations in this taxon revealed a mutation event that did not occur in a stepwise manner. The SMM was most successful at detecting heterozygote deficiency, which is of great importance for such small, endangered populations, and which was detected in all BHSP populations analyzed. Even with the increased sampling from 2005 vs. 2006 ( N = 24 vs. N = 77) there was minimal increase in the effective number of alleles ( n e ) in the BHSP population. This confirms that my sampling practices for both years have effectively captur ed the diversity present in the population. These data show that the BHSP population has retained an adequate amount of genetic diversity to maintain the population despite perceived changes in overall population size. Census estimates (from daily transe ct counts) for 2005 are 5 133 individuals and 15 82 individuals for 2006 (J.C. Daniels, unpublished data). These data roughly agree with results found for effective population size based on genetic data which show the population supporting approximately 2 7 butterflies (13 70 individuals, 95% CI). Inbreeding coefficients ( f ) in individual habitat patches on BHSP and KWNWR vary greatly over space and time, from f = 0.289 0.285. The observation that inbreeding coefficients in discrete habitat patches (popul ation segments) change over time may speak more to habitat quality and the ability to disperse to better habitat. The inbreeding coefficient of the common but poorly dispersing lycaenid Arhopala epimuta is comparatively very high (denoting high levels of inbreeding), ranging from f = 0.407 0.503 (Fauvelot et al. 2006). Inbreeding coefficients in the non endangered, high dispersing butterfly Heliconius charitonia in the Miami area are all negative (Kronforst & Fleming 2001). The overall F IS values, howeve r, of BHSP C. t. bethunebakeri are consistent and relatively close to 0 and are similar to the swallowtail butterfly Papilio machaon ( F IS = 0.062) (Hoole et al. 1999). In both C. t. bethunebakeri and the

PAGE 78

78 swallowtail butterfly, high gene flow is evident, a nd the F IS value indicates neither an excess nor deficit of heterozygotes. The overall message appears to be that C. t. bethunebakeri is not particularly inbred in the wild and values are comparable to those of other butterflies with varying dispersal abi lities and population numbers. Care should still be taken in interpreting genetic results as several loci are still not in HWE on a regular basis and there appears to be a consistent heterozygote deficiency in the BHSP population. The population needs to be continuously monitored for drops in allelic diversity and changes in frequencies across loci. Such measurements offer a yardstick for overall population health. The l ow F ST and R ST values evaluated in this study demonstrate the high levels of gene f low between subpopulations (habitat patches) on BHSP. Such values are explained by either high levels of gene flow or a recent colonization or expansion event on BHSP. As this population was discovered in 1999, over 70 generations have been completed, al lowing sufficient ti me for allelic values to diverge after a founder event. This reinforces the hypothesis that low F ST and R ST values are due to high levels of gene flow between patches. In 2005, no significant differences were observed between habitat patches on th e South end of BHSP (Figure 4 4A ). Within the South end of BHSP, it appears that small habitat patches (like the Bridge patch) may be crucial in providing links between the other subpopulations in the area. When the Bridge patch was of poor quality and few butterflies resided there in 2006, estimates of population differentiation between the upper and lower walkway and Bayside areas increased and were significantly different (Figure 4 4 B ). The Main Road population similarly represents an im portant link between South end and Sandspur populations (Figure 4 5). The South end and Buttonwood populations, however, showed the highest F ST and R ST values (0.051 and 0.276, respectively), with significant values in the latter measure. These populatio ns are

PAGE 79

79 close in terms of geographic distance, but there may be habitat or demographic barriers to gene flow. The overall estimate of F ST in another endangered lycaenid butterfly, Polyommatus bellargus, is 0.127 in populations separated by much greater dis tance (Harper et al. 2003). Results from the Structure analysis showed two distinct populations on BHSP, but they were not structured by sex, time, or location. This suggests that there is an internal rescue occurring on BHSP and that gene flow between s ource patches may effectively rescue dwindling sub populations. Due to genetic mixing, exact patterns were not evident from the Structure results. Furthermore, d ata showing high gene flow and low levels of inbreeding in close habitat patches should be car efully considered when partaking in reintroduction exercises and conservation planning for C. t. bethunebakeri and taxa in similar metapopulation structures Additional analysis of the KWNWR population may uncover more metapopulation structuring in this r egion with future sampling and monitoring endeavors. The comparison of the BHSP and KWNWR population s ( September 2005 and February 2008 respectively ) shows F ST = 0. 191 ( P = 0.0 0 01 ) and R ST = 0. 275 ( P = 0.0001). This amount of population differentiation is perhaps lower than would be expected for populations that are over 70 km apart. This division is corroborated by results produced by Structure, showing the division between the BHSP and KWNWR populations. These values suggest the hypothesis that separ ation between the populations was recent. Populations between KWNWR and BHSP once existed, linkin g the two populations, and the recent absence of these populations has broken the gene flow. A nalysis of COI mitochondrial sequence data between individuals from BHSP and several islands of KWNWR show identical nucleotide sequences, lending further evidence that the se butterflies are the same taxon ( E. V Saarinen, unpublished data .). Divergence in microsatellite allele frequency has not yet extended to nucleo tide changes. Further confounding

PAGE 80

80 this story however, is the fact that larvae of KWNWR feed on host plant s of a different family from those of BHSP. Larvae on KWNWR feed on Pithecellobium spp. (Fabaceae) and BHSP larvae feed exclusively on Caesalpinia b onduc (Fabaceae), even though Pithecellobium spp. is found on BHSP. No C. bonduc has been found on the islands of KWNWR with C. t. bethunebakeri Historical data show that larvae were able to feed on both host plants as well as Cardiospermum spp. (Sapind aceae). An ancestral polymorphism explains how larvae of the same taxon can be feeding on very different plant families. Adaptation to specific hosts can lead to changes in allelic frequency ( Sunnucks et al. 1997 ) and may lead to speciation ( Templeton 19 81 ) and, t aken together, these results impress the need for further study. This study further clarifies the fact that the underlying genetic diversity of C. t. bethunebakeri is not due to an accelerated rate of mutation of microsatellite markers. The mutation rate for this taxon at these microsatellite loci is 1.98 x 10 3 which falls well within the range of published mutation rates across taxa and especially when compared to related invertebrate taxa; an ant at 1.8 x 10 3 (Crozier et al. 1999) and a freshwater snail at 1.1 x 10 3 (Gow et al. 2005). The extensive survey of wheat microsatellite marker mutations over 22 generations showed similar rates (0 4.97 x 10 3 per generation) and that the genetic diversity in the initial parental populations wa s conserved over time. Taken together, these results show that microsatellite mutations, i.e. the emergence of new diversity, balance the loss by selection and drift (Raquin et al. 2008). Loss of diversity by drift is of special concern in small populati ons, like those of C. t. bethunebakeri The BHSP population has persisted since at least 1999, despite subjugation to both major hurricane events and severe droughts between 2002 present. This taxon has essentially persisted in the face of natural disas ter yet has been unable to cope with the human induced disturbance

PAGE 81

81 and historical destruction of habitat that have affected its former habitat in s outh Florida. The data presented here speak to the fact that the habitat patches and metapopulation structur e allow for significant gene flow, maintenance of genetic diversity, and reduced levels of inbreeding; all of which are essential to the conservation of this taxon. This situation mirrors that of the Glanville fritillary Melitaea cinxia which went extinc t on the Finnish mainland in the 1970s and now occurs on neighboring islands in a fragmented, metapopulation structure. Extensive study of this system has shown how increased habitat fragmentation furthers colony extinction and how a balance between local extinctions and recolonizations are essential for taxon survival (Hanski et al. 1995; Hanski & Ovaskainen 2003). Additional studies of neotropical trees further highlight the importance of gene flow in maintaining genetic diversity and supporting populat ions over time (Lowe et al. 2005). Host plant quality may vary over time and space, but the presence of a variety of habitat patches ensures that some suitable host is always present for C. t. bethunebakeri butterflies and larvae. Fine scale habitat chan ges may influence the development of individuals, but the overall health of the taxon can be ensured by retaining a variety of such habitats.

PAGE 82

82 Figure 4 1. Schematic diagram of Bahia Honda State Park, Florida C. t. bethunebakeri sampling locations. A, overview; B, d etail of South end colonies B A

PAGE 83

83 Figure 4 2. Removal of a wing fragment from an adult C. t. bethunebakeri butterfly for use in genetic analyses. Table 4 1. Sampling regime for molecular studies of C. t. bethunebakeri on Bahia Honda State Park, Florida 2005 2006. Date N N ( pooled samples ) September 2005 24 24 April 2006 4 21 May 2006 17 June 2006 39 48 July 2006 9 September 2006 8 21 October 2006 13 Total 2005 24 Total 2006 90 Total 114 Table 4 2 Genotyping error rates for each microsatellite locus of the C. t. bethunebakeri N number of DNA samples with second round of PCR and genotyping; % matching, percentage of second genotyping results matching original score. All samples are from Bahia H onda State Park, Florida. Locus N C orrect allele matches I ncorrect allele matches % matching % failure due to allelic dropout CthB103 3 6 0 100.00 % NA CthB106 12 14 10 58.33 % 80 % CthB115 13 24 2 92.31 % 50 % CthB117 9 14 4 77.78 % 75 % CthB119 11 14 8 63.64 % 100 % CthC12 12 24 0 100.00 % NA CthC116 10 16 4 80.00 % 50 % CthC124 5 10 0 100.00 % NA CthD7 23 46 0 100.00 % NA Total 127 168 28 85.71 % 78.57%

PAGE 84

84 Table 4 3. Summary statistics of allelic patterns for C. t. bethunebakeri in Bahia Honda State Park (BHSP), Florida. Values are averaged across the 12 loci; N sample size; n e mean number of effective alleles; H O mean observed heterozygosity; H E mean expected heterozygosity. Data from Key West National Wildlife Refuge (KWNWR) are given last. Population ( N ) Total alleles Private alleles n e Mean H O Mean H E September 2005 (24) 52 5 2.064 0.3650.067 0.4580.057 April+May 2006 (21) 50 2 2.199 0.4210.075 0.4840.060 June 2006 (39) 64 10 2.134 0.4250.073 0.4630.060 July 2006 (9) 40 1 2.077 0.3500.068 0.4450.059 June+July 2006 (48) 67 13 2.125 0.4130.064 0.4630.060 September 2006 (8) 37 3 2.388 0.2680.084 0.4250.074 October 2006 (13) 42 7 2.272 0.4200.074 0.5130.054 Septe mber+October 2006 (21) 53 10 2.308 0.3620.063 0.5080.053 Total 2005 (24) 52 5 2.064 0.3650.067 0.4580.057 Total 2006 (90) 84 34 2.246 0.4030.060 0.4940.057 Total BHSP (114) 92 56 2.233 0.3950.060 0.4930.057 KWNWR (27) 58 22 3.010 0.5100.058 0. 6010.051 Table 4 4. Hardy Weinberg equilibrium (HWE) probabilities for C. t. bethunebakeri in BHSP and KWNWR populations by locus. N sample size in parentheses. Sequential Bonferroni corrections have been done for each population to correct Type 1 e rror. Locus CthC124 is monomorphic (M) in three populations. P values in italics are in HWE. All data are discrete populations and no data are pooled between months Locus BHSP September 2005 (24) BHSP May 2006 ( 21) BHSP June 2006 ( 39) BHSP October 2006 (13) KWNWR 2008 (27) Cth B11 0.029 0.201 0.028 0.516 0.549 Cth B101 0.280 0.005 0.000 0.071 0.174 Cth B103 0.147 0.058 0.403 0.600 0.010 Cth B106 0.002 0.004 0.003 0.036 0.080 Cth B115 0.045 0.186 0.001 0.000 0.000 Cth B117 0.000 0.043 0.126 0.001 0.010 C th B119 0.000 0.004 0.513 0.004 0.002 Cth C12 0.001 0.000 0.000 0.001 0.316 Cth C116 0.599 1.000 0.025 0.518 0.428 Cth C124 M M 0.027 M 0.136 Cth C127 0.021 0.420 0.561 0.393 0.347 Cth D7 0.184 0.077 0.000 0.365 0.897

PAGE 85

85 Figure 4 3. Inbreeding coeffici ents (and standard error bars) for each metapopulation segment of C. t. bethunebakeri on Bahia Honda State Park over time (2005 2006) Red, 2005; Blue, 2006; Green, 2008; UW, upper walkway; LW, lower walkway; Bayside, Florida Bay side of BHSP; Bridge, base of Old Flagler Railroad; Buttonwood Campground; Sandspur Campground; Main Road of BHSP. Data from the Key West National Wildlife Refuge (KWNWR) from the Marquesas Islands are also shown. Table 4 5. Wilcoxon signed rank tests for excess (top) and deficie ncy (bottom) of heterozygosity for discrete breeding populations of wild and captive raised laboratory populations of C. t. bethunebakeri BHSP, Bahia Honda State Park; CC, captive colony; KWNWR, Key West National Wildlife Refuge. Significant values ( P < 0.05) are in italics; IAM, infinite allele model; TPM, tow phase model; SMM, stepwise mutation model. Population Sample size IAM TPM SMM BHSP 9 2005 24 0.618/0.416 0.913/0.103 0.992/ 0.011 BHSP 5 2006 17 0.416/0.617 0.817/0.207 0.989/ 0.027 BHSP 6 2006 3 9 0.849/0.170 0.954/0.055 0.993/ 0.017 CC 7 2006 40 0.661/0.367 0.979/ 0.026 0.999/ 0.001 CC 10 2006 10 0.003 /0.998 0.103/0.913 0.289/0.740 CC 11 2006 12 0.150/0.875 0.327/0.715 0.545/0.500 KWNWR 2 2008 27 0.003 /0.997 0.088/0.924 0.689/0.339

PAGE 86

86 Table 4 6. Effective population size ( N E ) estimates of the BHSP population of C. t. bethunebakeri ; May 2006 October 2006. Results were generated by comparing allelic frequencies in consecutive generations in a moments based temporal model. Generation # generat ions N E 95% CI May June 2 26.6 10.0 534.6 June July 2 45.5 9.2 May July 3 25.1 8.2 September October 2 5.4 2.3 17.2 May October* 6 26.8 12.5 68.8 There is no data for August 2006

PAGE 87

87 Figure 4 4. Pairwise F ST R ST and corresponding P values (after diagonal) for C t. bethunebakeri on BHSP A, 2005; B, 2006 P value significance calculated from 9,999 permutations of 12 microsatellite loci. Low sample size prevented Bridge 2006 values from being calculated. A B

PAGE 88

88 Figure 4 5. Pairwise F ST R ST and corresponding P v alues (after diagonal) for C. t. bethunebakeri on BHSP in 2006. P value significance calculated from 9,999 permutations of 12 microsatellite loci. Table 4 7. Results from Structure 2.0 for K values 1 5, data are from all BHSP subpopulations ( N = 114) an d are not pre assigned to populations for analysis. The highest Ln P ( X / K ) in 5 runs is shown with mean inferred assignment to clusters. K Ln P ( X / K ) Cluster 1 Cluster 2 Cluster 3 Cluster 4 Cluster 5 1 2688.1 2 2570.4 0.462 0.538 3 2624.4 0. 298 0.336 0.366 4 2651.4 0.238 0.243 0.265 0.254 5 2780.8 0.197 0.188 0.210 0.206 0.199

PAGE 89

89 Figure 4 6. Structure plot of all BHSP data ( N =114) when K = 2, data are from all BHSP subpopulations ( N = 114) and are not pre assigned to populations for analysis. Each vertical line represents one individual butterfly; red and green represent two different source populations. Data are shown chronologically, with BHSP 9 2005 on the left and ending with BHSP 10 2006 on the right. Table 4 8. Results from Structure 2.0 for K values 1 5, data are from the BHSP 9 2005 ( N = 24) and KWNWR 2 2008 ( N = 27) populations. The highest Ln P ( X / K ) in 5 runs is shown with mean inferred assignment to clusters. Prior pop., samples divided into two source populations based on location; No prior pop., samples not pre assigned to a population for analysis. K Prior pop. Ln P ( X / K ) No prior pop. Ln P ( X / K ) Cluster 1 Cluster 2 Cluster 3 Cluster 4 Cluster 5 1 1327.0 1327.1 2 1187.7 1189.1 0.491 0.509 3 1243.6 1172 .4 0.385 0.280 0.335 4 1225.1 1155.4 0.263 0.235 0.240 0.262 5 1261.6 1135.6 0.210 0.200 0.191 0.226 0.173 Figure 4 7. Structure plot of KWNWR 2 2008 ( N = 27) and BHSP 9 2005 data ( N = 24) when K = 2 and prior population assignment is incorpor ated into the model. Each vertical line represents one individual butterfly; red and green represent two different source populations; 1, KWNWR; 2, BHSP.

PAGE 90

90 CHAPTER 5 GENETICS OF THE CAPTIVE COLONY Introduction Numerous animal and plant species are endange red in the wild and now rely upon captive breeding for survival or as insurance against extinction (Soul et al. 1986; Frankham 2008). The maintenance of genetic diversity in a captive colony should be a priority when individuals produced therein will be released to found wild populations A captive colony must therefore actively capture and reflect the genetic diversity present in remaining wild populations. Genetic monitoring of captive colonies can be used to detect changes in genetic diversity over g enerations in captivity and to measure potential effects of inbreeding or drift (Lacy 1993). It is additionally important to genetically monitor individuals for signs of genetic adaptation to captivity as this is expected to have negative effects on reint roduction success (Frankham 2008). The Miami blue butterfly, Cyclargus thomasi bethunebakeri was once a common taxon throughout southern Florida but is currently only found on two island regions (Bahia Honda State Park and Key West National Wildlife Ref uge) in the extreme southern p ortion of the state (Klots 1964; Lenczewski 1980; Minno & Emmel 1993; Ruffin & Glassberg 2000; Calhoun et al. 2002; Cannon 2007). The documented decline may have begun gradually, but it has become dramatic within recent years Cyclargus thomasi bethunebakeri is currently one of the most endangered taxa in the United States due to its low population numbers presence at a fraction of its original geographic range and susceptibility to impact from disturbances such as hurrican es and tropical storms Through public petition, this taxon was given emergency status as a state endangered species in Florida in November 2002 by the Florida Fish and Wildlife Conservation Commission (FWC) and an emergency management plan was later draf ted. Cyclargus thomasi bethunebakeri is currently a candidate to be listed as a federally endangered species.

PAGE 91

91 The Florida Fish and Wildlife Conservation Commission ( FWC ) Department of Environmental Protection ( DEP ) U.S. Fish and Wildlife Service (US FWS ) and the University of Florida (UF) jointly agreed to establish a captive propagation program due to the extremely low estimated population size of less than 50 adult individuals remaining in Bahia Honda State Park (BHSP) The FWC authorized the collecti on and transport of 100 C. t. bethunebakeri eggs to the Unive rsity of Florida in Gainesville for the purpose of establishing a captive breeding population. The purposes of this colony are to conserve the remaining genetic diversity of the taxon, safeguard the existence of this taxon from natural and manmade factors that may threaten its existence, and to produce a significant number of individuals for reintroduction into the wild. Given the negative genetic effects that occur in captive breeding (inbreed ing depression, adaptation t o captivity, loss of diversity), it is important to assess the genetic constitution of a captive colony (Frankham 2008). It is not only important that the captive colony reflect the genetic diversity within the wild colony on B ahia Honda State Park, but that it continue s to mirror this wild population over time and that adaptation to captivity is minimized. I n this study I assess how genetic diversity is lost in captivity through comparison of captive generations over time as w ell as through direct comparison with the founding population on Bahia Honda I also evaluate the captive colony for novel mutations as well as changes in effective population size Methods Basis of Captive Colony A total of 100 eggs were collected from ten randomly captured female C t. bethunebakeri originating from several sites within BHSP over a four month period from February 2003 to May 2003. The resulting 100 eggs were used as the foundation for building the captive colony at the University of F lorida. In order t o avoid marked inbreeding depression, captive breeding was based on a founder number of at least ten unrelated pairs (Saccher i et al. 1996). To maintain

PAGE 92

92 genetic diversity over time and to compensate for lower founder numbers, captive po pulations were allowed to increase rapidly after establishment and future population bottlenecks were avoided as possible (Saccher i et al. 1996; v an Oosterhout et al. 2000; Saccheri et al. 2001). A significant decline in captive colony numbers was experi enced between November 2005 and February 2006 (Table 5 1). Subsequent analysis by University of Florida virologists Dr. James Maruniak and Dr. Al jandra Maruniak confirmed the presence of a baculovirus in the captive colony. The presence of baculovirus in fections can be common in lepidopteran populations, and can be particularly devastating and difficult to control in captive or laboratory situations (Crone et al. 2007). Due to the resulting significant decline in larval numbers, the FWC and University of Florida jointly agreed to terminate the initial captive colony after some 30 generations and start over with new founder stock from BHSP. C ollection of new genetic material from BHSP was initiated in early April 2006 and continued through June 2006. A tot al of 80 eggs were collected as initial founder stock from multiple locations within BHSP (Table 5 2) Incorporation of New Genetic Material T he captive colony received the addition of new genetic material when the removal of material from BHSP was possibl e without further depressing the population A total of 100 eggs (25 per quarter) were collected and transported to the University of Florida each year (FWC permit amendment no. WX02525a) when they were available from BHSP Due to large captive colony nu mbers typical of many insect populations, traditional studbook methods widely used for other breeding programs ( primarily mammalian) were unrealistic in this situation due to funding, time and manpower restrictio ns. Alternative methods, similar to those i n fish captive breeding programs, were therefore used to manage and conserve available genetic diversity in the butterfly captive colony (Fiumera et al. 1999) These methods include d the maintenance of

PAGE 93

93 detailed records on the origin and mating of founder lines. Males were allowed to mate only a single time and s ibling mating was restricted. Various life history and behavioral data, particularly as they relate d to overall organism fitness, were recorded when appropriate All changes or deviations from th e norm were noted and used as a yardstick to better evaluate overall colony health. Similarly, efforts were made to maintain captive colony numbers between 150 400 organisms and to reduce rapid fluctuatio ns in overall colony numbers. Sampling of Captive C olony A sample of butterflies that died of natural causes in the captive colony were saved and kept in a 20 C freezer beginning in 2004. These butterflies were used to assess the genetic diversity of the captive colony over time. Butterflies from several generations were stored in groups in 2004 and thus knowledge of discrete generations during these years wa s not available. Data from the 2004 captive colony therefore represents 1 12 generations (Table 5 1). Butterflies in subsequent years were kept separate to allow for analyses of discrete generations. DNA was extracted from the thora c es of butterflies an d twelve polymorphic microsatellite loci were PCR amplified according to methods in Saarinen et al. (2009). Data from captive colony butterflies were compared with BHSP data from 2005 2006. Statistical Analyses Microsatellite d ata were organized in Micros oft Excel and formatted using Convert (Glaubitz 2004) to be used in other software packages. GenAlEx 6.1 ( Peakall & Smouse 2006 ) was used to calculate the observed number of alleles, effective number of alleles per locus, lists of private alleles per locu s, and the observed and expected heterozygosities. Statistical difference between means of genetic diversity values are assessed by paired t tests as recommended by Nei (1987). I used FSTAT version 2.9.3.2 to test for allelic richness ( A R ) a measure of the number of alleles independent of sample size (Goudet 2001) The independence

PAGE 94

94 from sample size allows for the comparison between generations or populations of varying sizes. e measurements as per Nei (1972; 1978). Nei (1972) details D S a traditional genetic distance measure with relatively low variance that measures the accumulated allele differences per locus. Nei (1978) describes the unbiased estimator of genetic distance, and this statistic has b een recommended to correct for differences in low sample size (Paetkau et al. 1997) although Nei himself (1978) cautions that when genetic distances are minimal, more individuals need to be sampled. Nei (1978) further notes that the unbiased genetic dista nce estimator works well with most sample sizes as long as more than 30 loci are used. Hedrick (2005) further cautions against use of the unbiased estimator of genetic distance as it may give poor results when homozygosity and sample size are small. As t here are 12 polymorphic loci used in the study of C t. bethunebakeri I present both the traditional and The program Bottleneck (Cornuet & Luikart 1996) was used to look for genetic signatures typical of a po pulation that has undergone a bottleneck. The Stepwise Mutation Model (SMM), Infinite Alleles Model (IAM), and Two Phase Model (TPM) of microsatellite evolution were calculated, allowing 30% variance for TPM with a 70% proportion of the SMM and 1000 itera tions of all models. I tested a random sample of adults from different generations of the captive colony that approximately met the sample size criterion of 15 40 individuals (Cornuet & Luikart 1996). To ensure that the program worked with my data, I als o tested a known population bottleneck using data from the October 2006 captive colony. This iteration of the captive colony had been allowed to breed without the introduction of new genetic material for four generations, thus representing a known populat ion bottleneck (Table 5 2). I used the program NeEstimator ( Peel et al. 2004 ) to estimate the effective population size ( N E ) of the

PAGE 95

95 captive colony of butterflies. This program uses a moments based temporal model ( Waples 1989) to estimate the effective po pulation size over time. For these analyses, consecutive generations were required to estimate effective population sizes. As I did not have data from the captive colony for August or September 2006, these are not factored in the analyses. Estimates of the effective size of the captive colony were compared to estimates of the BHSP colony. Results The captive propagation program has been extremely productive since its inception in February 2003 yielding almost 25,000 viable C t. bethunebakeri adults (T able 5 1 and 5 2 ). It has served as the basis for the successful production of organisms for reintroduction as well as ecological, behavioral, and non target pesticide research. Genetic Changes in the Captive Colony over Time Observed heterozygosity remain ed relatively constant across different generations of the captive colony (Table 5 3). However, when looking at successive generations of the 2006 captive colony, observed heterozygosity did fall significantly from June ( H O = 0.410) to November ( H O = 0.29 7) ( P = 0.042, df = 11) One of the last generations of the initial captive colony (November 2005) had an o bserved heterozygosity of 0.373 and one year later in November 2006 observed heterozygosity had decreased to 0.297 although not significantly ( P = 0.105, df = 11). When genetic data were pooled and examined across years rather than generations, the observed heterozygosity remained relatively constant and no statistical significances were found between years ( H O = 0.332 in 2004, H O = 0.373 in 2005, a nd H O = 0.377 in 2006 ; P = 0.861, F 2 ,33 = 0.150 ). A general trend of loss of allelic richness was observed over time (Figure 5 1). Allelic richness in the November 2005 captive colony was 3.673 and it was 2.360 one year later in 2006 ( P = 0.0004, df = 1 1) Other measures, such as the number of

PAGE 96

96 effective alleles ( n e ) was relatively constant over both generations and years (Table 5 3), however low n e values were again seen in the November 2006 captive colony. The effective population size ( N E ) of the first iteration of the captive colony could not be calculated because materials were not saved from multiple discrete generations. In the 2006 captive colony, there was a low effective population size of each generation of the captive colony of less than 15 individuals (1.7 47.5 indivi duals, 95% CI). The effective population size decreased from the summer to the fall of 2006 as there was no new genetic material incorporated into the colony during this time (Table 5 4). Changes in the captive colony were a lso measured using the accumulated allele differences per locus to yield the genetic distance between generations. Measures of genetic distance are (traditional, Table 5 5 and unbiased, Table 5 6 ). These figures can be divided into three data quadrants. The lower right quadrant shows the genetic distances between discrete generations of the captive colony. Each column of data in a quadrant shows that genetic distance increases with tim e in both the traditional and unbiased measures. All unbiased measures of distance are lower than traditional values, although they show the same trend of increasing with comparisons of more temporally separate ge nerations The upper left quadrant shows changes in genetic distance in generations of butterflies from BHSP and is discussed in the following section. Genetic distance measurements from population matrices are plotted in Figures 5 2 and 5 3. Co mparisons with Bahia Honda State Park The BHSP po pulation and captive colony have held a similar number of alleles ( N A ) and simila r number of effective alleles ( n e ) over time Both BHSP and the captive colony held ~2 effective alleles from 2004 2006 (Table 5 3) Measures of observed heterozygosity were also

PAGE 97

97 similar, but no real temporal trends were evident (Figure 5 4). Allelic richness was lost over time in the captive colony but remained stable in BHSP (Figure 5 1). Changes in genetic distance between generations of the captive colony and BHS P population were evident ( Tables 5 5 and 5 6 lower left quadrants) Again, trends were similar in both tables although unbiased estimates of genetic distance were lower than traditional estimates. One generation was required for genetic material collec ted from BHSP to become incorporated and detectable in the captive colony. For this reason, a series of pairwise comparisons between months of generations is shown for BHSP and the captive colony. Generational data was also pooled for comparison of dista nce between the wild and the captive colony over years. Comparison of discrete generations showed an increased genetic distance between BHSP and captive colonies over time (Table 5 5 and 5 6 ). It is important to remember that a new captive colony was initiated in 2006. Comparison between the first captive colony and the BHSP showed traditional D S = 0.138 and unbiased D S = 0.113 (Tables 5 5 and 5 6 ). When generations (months) were compared in the second iteration of the captive colony (e.g. BHSP June 2006 vs. CC June 2006, BHSP July 2006 vs. CC July 2006, and BHSP October 2006 vs. CC October 2006), g enetic distance between the BHSP and captive colonies increased over time from June to September 2006. This increase in genetic distance over time between wild and captive populations was also evident when a lag time (allowing for incorporation of new material) was factored into the a nalysis (Table 5 5 and 5 6 ). A decrease in genetic distance between the October 2006 BHSP population and the November 2006 c aptive colony was also observed. Genetic distance values between generations were high (mean traditional D S = 0.114, mean unbiased D S = 0.083) when considered relative to the genetic distance calculated between years.

PAGE 98

98 When all BHSP 2006 data were pooled and compared against all CC 2006 data, distance values were 0.036 and 0.030, for traditional and unbiased estimations of genetic distance, respectively. When data were pooled over the two years 2005 2006, values decreased to 0.027 ( D S ) and 0.022 (unbiase d D S ). Genetic distance values over time within the BHSP population did not change, but the captive colony showed significant changes ( Figures 5 2 and 5 3 ). Not only were changes in genetic distance higher in the captive colony, but they also increase d over time. Again, changes in observed heterozygosity over time in wild and captive colony samples show ed no real trends (Figure 5 4 ). Private A lleles All private alleles conformed to base pair size expectations, but there were unique alleles only found in the BHSP population as well as private alleles only found in the captive colony (Appendix D) When considering the entire pool of alleles sampled from BHSP in 2005 2006 ( N = 114) vs. the alleles sampled from the captive colony during these same years ( N = 101), there were 29 alleles found only in BHSP and 17 alleles found only in the captive colony Of these 2 9 alleles, 25 were found at a low frequency of 0.004 0.009 in BHSP The remaining four alleles were found at the intermediate frequencies of 0. 013 0.022. In the captive colony, there were 17 alleles not found in BHSP. Of these alleles, only 6 were found at a low frequency (<0.009). Nine alleles were found at an intermediate to high frequency of 0.0 20 0.067. In summary, there were 29 alleles n ot represented in the captive colony from BHSP and there were 17 novel mutations in the captive colony not sampled or not present in BHSP. Of the private alleles in BHSP, 86% were found at a low frequency, versus 35 % of private alleles in the captive colo ny found at low frequency.

PAGE 99

99 Bottlenecks Different mutation models had an effect on the detection of heterozygote excess and deficiency (Table 5 7). The IAM was the only model able to detect a population bottleneck (denoted by heterozygote excess) in the O ctober 2006 captive colony. The TPM and SMM failed to detect any bottlenecks. The program Bottleneck also detected a number of heterozygote deficiencies (Table 5 7). The TPM detected a heterozygote deficiency in the July 2006 captive colony, the SMM det ected heterozygote deficiencies in the June and July 2006 captive colonies, and the IAM detected no heterozygote deficiencies. Discussion The captive colony has been successful at generating individuals for reintroduction to the wild and in safeguarding the remaining wild populations. The captive colony has had the additional benefit of elucidating life history characteristics for immature and adult stages including the identification of a baculovirus that was capable of decimating the captive colony an d could potentially destroy wild colonies as well. Information gathered from captive colony management is not always directly applicable to natural populations, but it may highlight various stochastic factors that are of importance in the wild as well su ch as the devastating effects of a pathogen The genetic measures used in this study ( N A n e H O A R D S and N E ) were useful in evaluating the ne gative consequences of captive breeding Together these methods are useful in elucidating the similar ities and differences in the captive colony and BHSP over time as well as genetic diversity lost in the captive colony. Individually, these measures can highlight novel mutations that have arisen in captivity and provide other estimates of genetic makeup such as effective population size. A concern in captive colony management is whether genetic diversity is lost over time (as a result of inbreeding or genetic drift) (Briscoe et al. 1992; Woodworth et al.

PAGE 100

100 2002; Frankham 2008). The number of alleles ( N A ) is biased by sample size, but when considering successive generations of similar sample size, a declining trend in the number of alleles in the captive colony is evident over time, and the number of effective alleles ( n e ) shows a similar trend. Values of observed heterozygosity fell significantly from November 2005 to November 2006 as did measures of allelic richness. Taken together, these data confirm that genetic diversity has been lost over time (between generations) of the captive colony. However, wh en years are compared instead of generations, all diversity measures are more stable. This is due to the influx of new genetic material during some generations and a complete lack of immigration during others. More effort was made in incorporating geneti c material from BHSP into the captive colony on a regular basis, but drought conditions in the summer of 2006 often prevented this. New genetic material was not added to the captive colony from July 2006 May 2007 as BHSP suffered from drought conditions a nd low population estimates (J.C. Daniels, unpublished data). The captive colony was not infused with new genetic material for approximately one year and evidence of genetic divergence from BHSP frequencies sampled was observable within three months. Decr eases in genetic diversity measures were seen during these times and fewer individuals were produced in the captive colony Allelic richness also decreased during this time. This denotes the importance of bringing new material into a captive colony on a generational or bi generational basis both to maintain genetic diversity and to prevent adaptation to captivity (Frankham & Loebel 1992). Another concern in captive colony management is how different ( in terms of genetic distance, allelic frequencies, and novel alleles ) the captive population becomes when compared to a wild population. Captive colony census numbers decreased in winter months and increased as conditions improved, much as they did in BHSP (J.C. Daniels, unpublished data ). There were

PAGE 101

101 high le vels of gene flow and, coupled with the effects of natural selection the BHSP population The captive colony is not subjected to natural selection allowing genetic drift to ha ve a greater impact on allelic frequencies and as such, result in greater genetic distances both between successive generations and when compared with the BHSP population. However, as the markers used in this study are neutral and presumably not linked to fitness measures, they would not be under selection in either the wild or the captive population and the differences in frequency are du e to random chance coupled with effective population size differences in each population. Private Alleles Differences b etween the BHSP population and the captive colony were due to differences in allelic frequencies as well as the presence of private alleles in both groups. These alleles wer e found only in single populations and usually occurred at low frequency. Private alleles found only in BHSP represent genetic diversity that has not been sampled and incorporated into the captive colony. Most private alleles in BHSP were present at a low frequency, indicating that they may be new mutations. These potentially novel m utations will eithe r be lost (by drift ) or retained in the population. It will be important to monitor these alleles over time and to incorporate them as necessary into the captive colony. Care should be taken to see where individuals with these pr ivate alleles are found on BHSP so that individuals from these habitat patches may be incorporated into the captive colony It is difficult to assess whether the novel alleles sampled from the captive colony represent real adaptation to captivity or simply neut ral variation as they are from neutral microsatellite markers These alleles underscore the importance of drift in captivity as they are the products of new mutation events that have drifted to a higher frequency in the absence of selection. In BHSP, all eles are lost during the natural selection process. In the captive colonies, non lethal molecular diversity is preserved. Private

PAGE 102

102 alleles in the captive colonies may not be selectively ad vantageous in a natural setting. In fact, these alleles may be neu tral or deleterious (or linked to a deleterious trait) underscoring the importance of keeping captive colonies genetically similar to wild populations in order to produce the best fit individuals for reintroduction M ore sampling has occurred in the capti ve colonies than from BHSP. A greater sampling coverage will return more alleles if all of the diversity was not captured by smaller sampling endeavors. The private alleles from the captive colonies may be present in BHSP at low levels but may not have b een sampled. Bottlenecks New genetic material was last incorporated into the captive colony in June 2006, so it is understandable that a bottleneck signature would appear four months later. It is unclear why there was no bottleneck detected in the Novembe r 2006 captive colony. The IAM was the best model for detecting a population bottleneck utilizing microsatellite data, while the SMM performed the best for detecting heterozygote deficiencies. Heterozygote deficiencies were detected the generation after new material was incorporated into the captive colony from the wild, but disappeared in subsequent generations. This confirms the conclusion made by Harper et al. (2003) that both the TPM and SMM perform poorly at detecting past bottlenecks and I emphasiz e the importance of utilizing several mutational models. Effective P opulation S ize The effective population size of the captive colony is extremely low and represents only a fraction of the census size of the captive colony. The effective population si ze of the captive colony is also lower than the BHSP effective population size a situation similar to other captive colonies (Bailey et al. 2007) Such figures are important as they show the importance of genetic diversity, not just maintaining a large c aptive population (Briscoe et al. 1992).

PAGE 103

103 Implications for Conservation The markers used in this study were presumably neutral and it is therefore not possible to differentiate whether allelic changes are due to drift or adaptation to captivity in the BHSP and captive colony populations. The fact that there are 17 alleles only found in the captive colony and that these alleles are found at moderate fr e quencies, speaks to their presence as an adaptation to captivity. A mutation in the wild has the chance of drifting to a higher frequency if there are no selection pressures against it, and this may be the case in captivity. These novel mutations may be able to drift to a higher frequency in captivity but may be deleterious in the wild. The accumulation of n ovel mutations such as these may eventual and fitness measures should be monitored closely in the captive colony (Woodworth et al. 2002). This study shows how, at neutral markers, a captive colony will diverge from a wild po pulation over time. This has many conservation implications if individuals produced in the captive colony are to be used for reintroduction purposes. Multiple procedures should be followed in order to reduce genetic adaptation to captivity. The first me thod is to introduce immigrants from the wild into the captive colony on a regular basis, once per generation if possible (Woodworth et al. 2002 ; Frankham et al. 2003 ). This is not possible for many endangered species programs as there are no remaining wi ld populations or to remove an individual would be unwise. The C. t. bethunebakeri program maximizes on the ability to incorporate wild stock into the captive program on a regular basis, and this study highlights the negative repercussions of not doing so Next, individuals should be raised in environments as similar to wild conditions as possible. Again, this is not possible due to funding and facility restrictions for most taxa Individuals from t he captive colony of C. t. bethunebakeri are allowed to mate in outdoor flight cages and are subjected to environmental changes as adults but other than that, conditions are kept very benign and do not afford any type of selection. Two other

PAGE 104

104 procedures that have been recommended by multiple authors (Margan et al. 1998; Fiumera et al. 1999; Woodworth et al. 2002; Frankham et al. 2008; Fraser 2008) have not yet been instituted in this captive management program. These recommendations are to 1) equalize family size and 2) fragment captive populations. Frankham et al. (2000) show that the equalization of family sizes effectively slows genetic adaptation to captivity in Drosophila and is a technique that could be easily done with butterflies. This procedure could be carried out in conjunction with the next object ive of fragmenting the population in multiple colonies. Family sizes could be equalized in different populations, thus removing the need to cull larvae from the program. Fragmenting the existing colony would help to safeguard it from future baculovirus o r other pathogen outbreaks as well as to slow genetic deterioration (Margan et al. 1998). Genetic adaptation to captivity is more significant in large populations than small populations (Woodworth et al. 2002) but small populations show greater inbreedin g depression. This balance may be achieved by subdividing a captive colony into several small medium sized captive populations. Smaller captive colonies are also easier for technicians to manage and smaller populations show less response to selective pre ssures (Woodworth et al. 2002). Lastly, because both wild and captive populations have private alleles, it is theoretically possible that they could enhance each others fitness through the introduction of novel genetic diversity (Bailey et al. 2007 but see Ford 2002 ). However local adaptations, both in captivity and in the wild, likely outweigh any benefit from the introduction of novel alleles as well as the potential breakdown by outbreeding depression. Theodorou and Couvet (2004) have shown that ca ptive bred individuals may have a positive effect on dwindling wild populations if a strict model is followed, including few captive breeders released per generation and adherence to captive breeding protocols. As such, captive bred individuals should onl y be used to found new

PAGE 105

105 colonies and should spend as few generations in captivity as possible. Captive breeding program s require extensive amounts of time and energy (Snyder et al. 1996) and if their ultimate goals are to produce reproductively fit offspri ng genetic variability must be maximized and genetic adaptation to captivity minimized.

PAGE 106

106 Table 5 1. Summary of C. t. bethunebakeri initial captive colony numbers from February 2003 to December 2005 N number of Bahia Honda State Park (BHSP) collected females refers from which eggs originated and were incorporated to the captive colony; N eggs collected from BHSP were directly collected from BHSP and added to the captive colony. Captive generation Month/Year N BHSP collected females N eggs collected fro m BHSP N viable adults produced 1 Feb/ 03 2 8 2 2 Mar/Apr 03 4 30 65 3 May 03 2 31 192 4 Jun 03 2 31 356 5 Jul/Aug 03 0 0 623 6 Aug/Sept 03 0 0 1 411 7 Sept/Oct 03 0 0 1 274 8 O ct/Nov 03 1 14 1 192 9 Nov/Dec 03 0 0 516 10 De c 03 /Jan 04 0 0 297 11 Jan/Feb 04 0 0 253 12 Feb /Mar 04 0 0 589 13 Mar/Apr 04 0 0 1 488 14 Apr /May 04 0 0 1 378 15 Jun 04 0 28 923 16 Jun/Jul 04 0 0 1 234 17 Aug 04 0 0 1 078 18 Aug/Sept 04 0 0 1 345 19 Sept 04 0 0 1 234 20 Oct 04 0 0 1 647 21 Nov 04 0 0 671 22 Dec 04/ Jan 05 0 0 113 23 Feb/Mar 05 0 0 24 24 Apr /May 05 0 23 32 25 Jun 05 0 0 109 26 Jul 05 0 56 787 27 Aug 05 0 0 80 28 Sept 05 0 0 87 29 Oct/Nov 05 0 0 54 30 Dec 05 0 0 63 Total 11 221 19 119

PAGE 107

107 Table 5 2. Summary of second iteration captive colony numbers of C. t. bethunebakeri from April 2006 to December 2007. N number of (Bahia Honda State Park) BHSP collected females refers to those from which eggs originated and were incorporated to the ca ptive colony; number of eggs collected from BHSP were directly collected from BHSP and added to the captive colony. Captive generation Month/Year N BHSP collected females N eggs collected from BHSP N viable adults produced 1 Apr/ May 06 0 46 28 2 J un 06 0 34 163 3 Jul 06 0 0 602 4 Aug/Sept 06 0 0 1 426 5 Sept/Oct 06 0 0 1 347 6 Nov/Dec 06 0 0 NA 6 Jan/Feb 07 0 0 NA 7 Mar/Apr 07 0 0 19 8 May/Jun 07 6 45 81 9 Jul 07 0 0 251 10 Aug 07 0 0 346 11 Sept 07 0 25 591 12 Oct 07 0 0 154 13 Nov/ Dec 07 0 0 NA Total 6 150 5 008 Table 5 3. Overview of molecular diversity measures for Bahia Honda State Park (BHSP) and captive colony (CC) populations of C. t. beth unebakeri ( N ) n mean amplifiable samples across all microsatellite loci; N A observed number of alleles ; n e effective number of alleles; H O observed heterozygosity; H E expected heterozygosity; *, data pooled and analyzed as one population. Population n N A n e H O H E BHSP 2005 (24) 23.750 4.417 2.068 0.365 0.457 BHSP 2006 (90) 86.833 7.000 2.246 0.403 0.494 BHSP Total (114*) 110.583 7.667 2.233 0.395 0.493 CC November 2005 (22) 19.500 4.750 2.776 0.373 0.573 CC June 2006 (15) 14.667 4.167 2. 341 0.410 0.513 CC July 2006 (40) 38.667 4.667 2.045 0.392 0.460 CC October 2006 (10) 9.667 3.000 2.143 0.361 0.448 CC November 2006 (12) 11.750 2.583 1.705 0.297 0.343 CC 2004 (25) 24.083 4.167 2.062 0.332 0.427 CC 2005 (24) 21.500 4.750 2.751 0.373 0.569 CC 2006 (77) 74.750 6.083 2.128 0.377 0.475 CC Total (126*) 120.333 7.167 2.440 0.368 0.533

PAGE 108

108 Figure 5 1. Average allelic richness values of 12 microsatellite loci measured in discrete generations of the Bahia Honda State Park (BHS P) and captive colony (CC) of C. t. bethunebakeri Equations for the linear function trend lines are y = 6E 05x + 5.2469 and R 2 = 0.0119 (BHSP) and y = 0.0031x + 124.71 and R 2 = 0.7891 (CC). Table 5 4. Effective population size ( N E ) estimates of the 2006 captive colony of C. t. bethunebakeri Results were generated by comparing allelic frequencies in consecutive generations in a moments based temporal model. Generation N generations N E 95% CI June July 2 14.8 6.7 47.5 October November 2 3.7 1.7 9.3 June November 6* 6.9 4.7 11.3 T he total number of elapsed generations June November is six real data exists only for 4 generations as there are no data for August or September 2006.

PAGE 109

109 Table 5 5. Pairwise population matrix of Ne D S ). Bahia Honda State Park (BHSP) and captive colony (CC) populations of C. t. bethunebakeri BHSP Sept 2005 BHSP May 2006 BHSP June 2006 BHSP July 2006 BHSP Sept 2006 BHSP Oct 2006 CC Nov 2005 CC June 2006 CC July 2006 CC Oct 2006 0.033 BHSP May 2006 0.050 0.032 BHSP June 2006 0.074 0.061 0.044 BHSP July 2006 0.077 0.099 0.075 0.068 BHSP Sept 2006 0.097 0.084 0.088 0.088 0.112 BHSP Oct 2006 0.138 0.117 0.158 0.131 0.192 0.113 CC Nov 200 5 0.078 0.050 0.051 0.054 0.139 0.125 0.116 CC June 2006 0.087 0.054 0.042 0.066 0.130 0.112 0.119 0.038 CC July 2006 0.142 0.098 0.096 0.102 0.199 0.165 0.182 0.075 0.058 CC Oct 2006 0.156 0.118 0.112 0.106 0.196 0.102 0.170 0.116 0.080 0.098 CC Nov 2006 Table 5 (unbiased D S ). Bahia Honda State Park (BHSP) and captive colony (CC) populations of C. t. bethunebakeri BHSP Sept 2005 BHSP May 2006 BHSP June 2006 BHSP July 2006 BHSP Sept 2006 BHSP Oct 2006 CC Nov 2005 CC June 2006 CC July 2006 CC Oct 2006 0.012 BHSP May 2006 0.035 0.014 BHSP June 2006 0.039 0.023 0.013 BHSP July 2006 0.041 0.061 0.043 0.016 BHSP Sept 2006 0.065 0.049 0.059 0.039 0. 063 BHSP Oct 2006 0.113 0.089 0.136 0.089 0.150 0.074 CC Nov 2005 0.050 0.019 0.026 0.009 0.094 0.083 0.080 CC June 2006 0.072 0.036 0.031 0.034 0.098 0.083 0.097 0.014 CC July 2006 0.110 0.063 0.068 0.053 0.150 0.119 0.143 0.033 0.030 C C Oct 2006 0.135 0.094 0.094 0.068 0.158 0.067 0.142 0.086 0.062 0.063 CC Nov 2006

PAGE 110

110 Figure 5 (CC) butterflies over time. Equations for the power function trend lines a re y = 0.077x 0.0971 and R 2 = 0.0328 (BHSP) and y = 0.0532x 0. 4248 and R 2 = 0.6304 (CC). Figure 5 captive colony (CC) butterflies over time. Equations for the power function t rend lines are y = 0.023x 0.2524 and R 2 = 0.1184 (BHSP) and y = 0.024x 0.6509 and R 2 = 0.6035 (CC).

PAGE 111

111 Figure 5 4. Observed heterozygosity values averaged over 12 microsatellite loci for generations of C. t. bethunebakeri ; BHSP, Bahia Honda State Park popula tion; CC, captive colony. Table 5 7. Wilcoxon signed rank tests for excess (top) and deficiency (bottom) of heterozygosity for populations of captive raised laboratory populations of C. t. bethunebakeri ; CC, captive colony. Significant values ( P < 0.0 5) are in italics; IAM, infinite allele model; TPM, tow phase model; SMM, stepwise mutation model. Population Sample size IAM TPM SMM CC June 2006 15 0.455/0.575 0.830/0.190 0.993/ 0.017 CC July 2006 40 0.661/0.367 0.979/ 0.026 0.999/ 0.001 CC October 200 6 10 0.003 /0.998 0.103/0.913 0.289/0.740 CC November 2006 12 0.150/0.875 0.327/0.715 0.545/0.500

PAGE 112

112 CHAPTER 6 THE GENETIC AND FITNESS CONSEQUENCES OF FOUNDER EFFECTS ON LABORATORY POPULATIONS Introduction Small population size, bottleneck events, and inbreeding go hand in hand yet are often treated separately in population level studies. Bottlenecks result in reduced population size and can be caused by founder events, disease outbreaks, or habitat loss. Single generation bottlenecks, known as founde r events, lead to small populations with reduced genetic diversity and adaptive potential due to alleles lost as a result of the founder event (Frankham et al. 2003). Studies with model organisms have shown that not only are rare alleles lost in founder e vents but that overall allelic frequencies are distorted in the resulting populations (England et al. 2003). These populations will be of reduced size and subjected to elevated levels of genetic drift and inbreeding. Inbreeding depression often results i n small populations and is realized as a loss of reproductive fitness, further contributing to population decline and extinction (Newman & Pilson 1997; Saccheri et al. 1998). This cascade of events is all initiated by founding populations with a few indiv iduals, and is of great concern in both the management of endangered species populations and the creation of captive colonies. Founder events are a severe form of bottleneck event (England et al. 2003) but are nonetheless a natural part of metapopulation dynamics (Hanski 1999; Haikola et al. 2004). Organisms that migrate to unoccupied habitat patches may found new colonies, often from single or very few individuals. In addition to environmental variables, the continued occupancy of such patches depends on the genetic variability of founders as well as that from new migrants. In many endangered taxa, however, there are few to no neighboring populations from which to receive new migrants. Newly founded populations are therefore dependent upon the genetic

PAGE 113

113 variability inherent in its founding individuals and are subjected to high levels of inbreeding due to the limited mating options. The Miami blue butterfly, Cyclargus thomasi bethunebakeri is a Florida listed endangered taxon that persists in a metapopu lation structure at two locations in the lower Florida Keys; Bahia Honda State Park (BHSP) and Key West Nat ional Wildlife Refuge (FWC 2003; E.V. Saarinen, unpublished data ). Previous research with this taxon has shown that gene flow is high between neighb oring patches, but that butterflies do not readily colonize new patches further than 12m ( J.C. Daniels unpublished data ). There is also no gene flow between the BHSP and KWNW R populations as they are over 70 km apart, separated by water barriers, and hav e no known intermediary, stepping stone populations (FWC 2003; E.V. Saarinen, unpublished data.). Suitable habitat patches within the metapopulation frameworks are regularly colonized, go extinct, and are subsequently re colonized by successive generation s of butterflies. Habitat patches may be colonized by single gravid females or by multiple immigrants. The goals of this study are to assess the effect of founder events on various levels of genetic diversity with the ultimate goal of evaluating evoluti onary potential of a population after such an event. In order to investigate the fitness and genetic consequences of new colony founding, I established artificial laboratory founder lines of butterflies so as to mimic the beginning of new populations by a limited number of adults. I then measured fitness in the resulting generation of butterflies and time to extinction of each of the populations. I used heterozygosity and other molecular measurements to assess genetic variability in newly founded laborat ory populations.

PAGE 114

114 Methods Experimental Treatment Lines One pair founder lines I created experimental lines of butterflies originating from wild caught BHSP females from 2005 2007. Females were only removed from BHSP when field numbers were high and in c ompliance with FWC permits. When BHSP population estimates were low, experimental lines were founded from individuals from an already established University of Florida captive colony. In these instances, only organisms that had been in captivity for one generation were used to found experimental lines. All female butterflies were kept alive in individual oviposition cups and supplied with ample artificial nectar (Gatorade on a cotton swab) and fresh terminal growth of host plant Caesalpinia bonduc (L.) Roxb. (Fabaceae) As all wild caught and captive colony females were assumed to be gravid, I stimulated oviposition with an alternating light regime in the laboratory. Eggs from these females were kept separately and the resulting larvae were allowed to complete development. As adult eclosions occurred, full sibling individuals were moved to individual flight cages outside of the McGuire Center for Lepidoptera and Biodiversity, Gainesville, Florida (FLMNH) and allowed to mate. Only full siblings were pr esent in each flight cage. All flight cages were monitored throughout the day at regular intervals of 15 minutes between 9am and 6pm and recorded mating pairs of butterflies were removed. All butterflies used in this study mated for at least 45 minutes a nd, after natural separation, females were placed in individual oviposition cups with ample artificial nectar (Gatorade on a cotton swab) and fresh terminal growth of C bonduc Males were not returned to the flight cage. The eggs that resulted from eac h full sibling mated female were removed and maintained in individually labeled containers. In this manner, the developing larvae, pupae, and adults that resulted from each full sibling mating were maintained as discrete groups. Each

PAGE 115

115 group was supplied w ith ample host plant and conditions were identical to those for raising the captive colony. Twelve replicates of single founder lines were completed in 2005 2006. Three pair founder l ines Three pairs of unrelated females and males (all from the Univer sity of Florida captive colony) were introduced to individual flight cages for the establishment of three pair experimental founder lines. Cages were regularly monitored and matings by the three females in each replicate were verified. The population of each line was again given ample food at each life stage. Lines founded by three pairs of adults were replicated three times in 2007. Control Groups Butterflies from the University of Florida captive colony were used as the control group for comparison wi th the different artificial founder line treatments as it was not possible to track individuals in the BHSP wild population. During the time of the experiments, the captive colony was regularly infused with new genetic material from the wild BHSP populati on, ensuring its status as an outbred population. Fitness Measurements and Statistical Analyses Measurements of fitness were recorded from individuals in all experimental treatments as well as from a subset of individuals from the University of Florida captive colony (control group). Fitness measurements and genetic samples were taken from a random sampling of individuals in the captive colony. All measurements were taken during the same time period that the treatment experiments were conducted, ensuri ng that larvae and adults were raised under identical environmental conditions and given host material of similar quality and from the same source. Life history traits typically show greater inbreeding depression than morphometric traits (Coltman & Slate 2003), and the former were measured in each generation. I counted the

PAGE 116

116 number of eggs that each female butterfly laid for the duration of her lifetime in each population of treatment and control butterflies. I tested the number of days over which females laid their eggs to account for any differences caused by long lived females in control and treatment groups. I also recorded the number of adults that successfully eclosed in each population. To account for differences in the starting number of eggs fro m each female, this statistic was converted to a ratio of the number of adults that successfully eclosed from the starting number of eggs. Egg laying and viability is a well documented fitness measurement for Lepidoptera (Saccheri et al. 1996) and the rat io of successfully eclosed adults generated from initial eggs encompasses the entire life cycle. The variance of each population was tested with an F test and sample t tests assuming unequal variances when appropriate) compared the means of the one pair founder line populations and concordantly run controls, the three pair founder lines and their respective controls, and the pooled treatments vs. the pooled cont rols. All tests were performed using Minitab statistical software, version 13.32 ( http://www.minitab.com ) DNA Extractions and PCR Amplifications I developed a relatively non invasive method to assess molecular d iversity in C. t. bethunebakeri in all experimentally founded lines as well as from individuals from the University of Florida captive colony. I gently held and removed a 2 mm 2 fragment of each The removed po rtion of the wing was placed directly in a vial with 90 100% ethanol and I carefully clean ed the forceps to prevent cross contamination between samples. Sample vials were kept in a cool, dark place until they could be placed in a 8 0 C freezer for stora ge. I extracted DNA from butterfly wing fragments using Qiagen DNeasy Blood and Tissue Kits (Qiagen, USA). I protocols for kit extractions with minor modifications owing to the small size of the wing

PAGE 117

117 material I used for extra ction I macerated wing material with sterile, disposable micro pestles and performed a final one time elution of each DNA template in 50L of warm 10mM TRIS. I stored DNA extractions at 20 C until I proceeded with PCR amplifications. All PCR reactions follow the protocols o utlined in Saarinen et al. (2009 ) for amplification of 12 polymorphic microsatellite loci All private alleles were individually assessed for validity and that they were not the result of genotyping or scoring error. Genetic Stati stical Analyses Analyses of observed and expected heterozygosity and effective number of alleles were assessed using GenAlEx 6.1 (Peakall & Smouse 2006). I used FSTAT version 2.9.3.2 (Goudet 2001) to assess gene diversity and allelic richness in control and treatment groups of butterflies. Allelic richness is one measurement of diversity unbiased by sample size; this measurement compares the number of alleles per population and allows for sample sizes to be different. Gene diversities were estimated at and were then averaged across the twelve loci for each population as recommended by England et al. (2003). Differences in gene diversity and allelic richness between treatment and control groups were assessed by ANOVA tests. The loss of heterozygosity after a founder effect was calculated with the equation H = H 1 H 0 (6 1) where H 1 is the heterozygosity after the founder event, H 0 H is the diff erence in the two generations. The empirical loss of heterozygosity was compared with the expectation that 25% of heterozygosity will be loss after a founder event of two individuals (England et al. 2003; Frankham et al. 2003).

PAGE 118

118 Results The results from these studies show that full siblings of C. t. bethunebakeri will mate with each other and produce viable offspring. Additionally, observations of the mating cages did not indicate any full sibling mating avoidance behaviors. However, because of other d emographic factors (presence of a baculovirus that killed many larvae, failure for males and females to emerge at overlapping times, etc.) all lines of one pair and three pair founded populations went extinct after one round of full sibling mating. There was no difference in the number of days over which any females (of any treatment or control groups) laid eggs ( F 1,32 = 1.448, P = 0.238). Populations that were founded by one pair of butterflies did not show a reduction in the number of eggs laid per fema le when compared to the outbred captive colony ( T = 1.08, df = 20, P = 0.294) ( Table 6 1). Populations that were founded by three pairs of butterflies did show a reduction in the number of eggs laid per female ( T = 2.71, df = 17, P = 0.015). When the o ne and three pair founder lines were pooled and compared against the control groups, there was a reduction in egg laying in the treatment group ( T = 2.70, df = 37, P = 0.010). The lifetime fitness measure of the number of adults surviving from eggs was c ompared and all treatment groups were significantly lower than the respective control groups. One pair founded populations showed reduced adult/egg ratios ( T = 2.57, df = 8, P = 0.033). Three pair founded populations were also reduced when compared to t heir controls ( T = 4.70, df = 9, P = 0.001) as were the pooled founder number treatments vs. pooled controls ( T = 4.86, df = 20, P = 0.000). After one generation of full sibling inbreeding (founding by a single pair of adults) there was no significan t difference in gene diversity ( P = 0.413) or allelic richness ( P = 0.080) with an outbred population (Table 6 2). Heterozygosity measures were consistent, showing no significant differences in observed heterozygosity in samples from populations founded b y one

PAGE 119

1 19 pair of adults vs. samples from an outbred population ( P = 0.084). Although these results did not show significant differences between one pair founded populations and the outbred individuals, allelic frequencies still changed after one round of inb reeding ( Figure s 6 1, 6 2, 6 3). These three microsatellite loci (CthB119, CthC127, and CthD7) are indicative of other microsatellite loci and showed how low frequency alleles are lost after a founder event and how high frequency alleles may become fixed. No statistics were calculated for the three pair treatments as only one adult butterfly was produced from the three replicates. These results show a decrease in heterozygosity of 15.8% after a new population was founded by one pair of butterflies versus an outbred population These results are not as severe as the expected value derived from the hypothesis that a 25% of loss of heterozygosity per generation will result from a population founded by only two reproducing individuals (Frankham et al. 2003). An unexpected result of this experiment was the discovery of a baculovirus affecting the larvae of C. t. bethunebakeri Larvae from both experimental treatments (one pair and three pair founded populations) were observed to become slow moving, paler in color, and the appearance of a yellow saddle was evident in the middle of the larvae ( Figure 6 4). All larvae with these symptoms died within 24 hours, and larvae raised in the same cups as infected larvae also became afflicted. Identification of the bac ulovirus was performed by Drs. James and Al jandra Maruniak, virologists at the University of Florida. Several larvae in the captive colony (2007) also became symptomatic of the virus, and measures were quickly taken to quarantine these individuals and to bleach the laboratory to prevent further spread of the virus. As both treatment and control individuals were affected, infection with this virus is not assumed to be due to reduced viral resistance in inbred individuals.

PAGE 120

120 Conclusions Experimental Shortcom ings and Fortuitous Discoveries This study affirmatively showed a decrease in fitness measures between populations founded by reduced numbers of individuals vs. the outbred group, but was unable to distinguish any difference between one pair or three pair founded populations because of poor sample size in the latter group. It is unclear why only one adult successfully eclosed from the three replicates of three pair founded populations. Sixteen butterflies eclosed from the 12 replicates of one pair founded populations. This study also revealed larval death from baculovirus infection. This infection occurred in ex perimental lines founded by one pair of butterflies in 2005 and 2006 as well as in all of the three pair founded populations in 2007. Individua ls from the 2006 and 2007 captive colony also had individuals afflicted with the baculovirus and as such, a direct correlation between inbreeding and lack of resistance to this virus cannot be inferred. Other insect conservation captive breeding programs have had problems with pathogens (Cunningham & Frank 1993; Pearce Kelly et al. 1998; Mattoni et al. 2003) but this is the first time a baculovirus has affected an endangered Lepidoptera captive breeding colony. Knowledge of the virulence of this pathogen has led to greater monitoring and prevention regimes in the management practices of the captive colony. An additional piece of ecological knowledge was gained insofar as I observed no avoidance of full sibling mating. These results are concordant with t hose found in the Glanville fritillary butterfly ( Melitaea cinxia ) (Haikola et al. 2004). In a small population, not mating with close relatives likely means not mating at all and therefore not reproducing. Thus, it is not surprising that both M. cinxia and C. t. bethunebakeri two taxa that exhibit metapopulation structuring with small populations colonized by low numbers of founders, have not developed

PAGE 121

121 strategies to avoid mating with siblings. Further knowledge of the ecology of C. t. bethunebakeri sho ws that adults are poor dispersers (J.C. Daniels, unpublished data) and are as such unable to migrate to distant habitat patches to mate with non relatives. Studies of model organisms have shown that m ating with relatives has likely led to the purging of deleterious alleles in homozygotes (Charlesworth & Charlesworth 1987 ; Crnokrak & Barrett 2002) Although still preliminary, this study suggests that the inability to detect or failure to refuse close siblings as mates may actually help save endangered tax a such as C. t. bethunebakeri and may have led to the purging of deleterious alleles. Founder Effects on Heterozygosity, Genetic Diversity, and Evolutionary Potential Work by Frankham et al. (1999) and England et al. (2003) have shown how population bottle necks affect the evolutionary potential and short term adaptability of populations of Drosophila melanogaster Furthermore, these studies showed how, in the short term, allelic diversity is lost rather than heterozygosity In the current study, significa nt amounts of heterozygosity and allelic diversity were actually retained in bottlenecked populations, but the sample sizes of these studies were small and grand inferences can not be drawn. Additionally, this study was successful only over one generation, whereas other studies using a model organism included data from multiple generations. This study is, however, important in that it showed how an endangered taxon reacts to experimental founding in the laboratory; data that are directly applicable to rein troductions in the field as well as to the beginnings of new populations for captive breeding. Severe bottlenecks result in an immediate loss of rare alleles, causing a reduction in allelic diversity faster than loss of heterozygosity (Nei et al. 1975; All endorf 1986). Results from this study support the previous statement, showing that rare alleles were immediately lost and that heterozygosity decreased by 15.8% after a new population was founded by one pair of butterflies

PAGE 122

122 versus the outbred control The se results were not as severe as the expected value derived from the hypothesis that a 25% of loss of heterozygosity per generation will result from a population founded by only two reproducing individuals (Frankham et al. 2003). However, this butterfly e xhibits a generation time of approximately one month (FWC 2003) and within several months a large amount of genetic variability and evolutionary potential could easily be lost. Results from this study support the idea that a bottleneck (single generation f ounder event) does not cause an immediately significant reduction in heterozygosity. Prolonged reduction in population size will have this effect (Frankham et al. 2003). Results from the current study do show how rare alleles are lost after founder event s, resulting in lower allelic richness, but not a significant reduction after one generation. The distribution of microsatellite alleles changes after a founder event and low frequency alleles are lost. Individual alleles are the source of evolutionary p otential and have been shown to be important in insecticide and disease resistance ( McKenzie & Batterham 1994; England et al. 2003). However, in mutation selection balance, there is an equilibrium between losing rare alleles that are fuel for evolutionary adaptability and losing alleles that are neutral or slightly deleterious (Kimura 1983). In small populations, such deleterious alleles will be lost and result in a population with lower genetic diversity but higher fitness than a larger, more outbred pop ulation. In a 2003 meta analysis, Reed and Frankham show that population genetic diversity is highly correlated with population fitness, reinforcing the idea that the loss of heterozygosity has a detrimental effect on a population. Because of the link b etween heterozygosity and population size (Frankham 1996; Frankham et al. 2001) heterozygosity can provide insight into a However this study has shown not on ly how founder events may produce populations of

PAGE 123

123 significantly reduced fitness but that fitness declines may be observed before decreases of neutral genetic diversity can be detected. It is therefore important to monitor both direct measurements of fitnes s as well as neutral genetic diversity as the latter will not as accurately predict pending population crashes. Considering the forces (genetic drift and inbreeding) that impact small populations is important in endangered species management as these fact ors lead to both heterozygosity and fitness declines (Frankham et al. 1999). Populations founded by low numbers of individuals become extinct at higher levels, thus denoting the importance of avoiding population bottlenecks in species of conservation conc ern.

PAGE 124

124 Table 6 1. Effects of founder lines on fitness measures (mean standard error). There were no significant differences between the one pair and three pairs treatments for either variable and there were no significant differences between controls Treatment Replicates Eggs laid/female Adults/eggs 1 pair 12 77.58321.859 0.0350.024 1 pair control 8 114.00026.107 0.1990.059 3 pairs 3 61.23020.515 0.0010.001 3 pairs control 10 154.50027.665 0.1630.035 Table 6 2. Effects of founde r lines on genetic variability measures. N sample size of butterflies genotyped; N A total number of alleles; N E effective number of alleles; H O observed heterozygosity; H E expected heterozygosity. The three pair treatment resulted in only one female butterfly. Treatment N N A N E H O H E Genetic diversity Allelic richness 1 pair 14.917 2.917 1.783 0.234 0.391 0 .411 2.797 1 pair control 40.667 4.750 2.094 0.392 0.470 0.477 3.619 3 pairs 0.917 1.083 1.083 0.167 0.083 NA NA 3 pairs control 8. 750 3.500 2.439 0.415 0.529 0.572 3.223

PAGE 125

125 Figure 6 1. Distributions of alleles (bp) at locus CthB119 in an outbred population (top) and populations founded by one pair of adult C. t. bethunebakeri (five replicates).

PAGE 126

126 Figure 6 2. Distributions of alleles (bp) at locus CthC127 in an outbred population (top) and populations founded by one pair of adult C. t. bethunebakeri (five replicates).

PAGE 127

127 Figure 6 3. Distr ibutions of alleles (bp) at locus CthD7 in an outbred population (top) and populations founded by one pair of adult C. t. bethunebakeri (five replicates).

PAGE 128

128 Figure 6 4. C yclargus thomasi bethunebakeri larvae killed by baculovirus.

PAGE 129

129 CHAPTER 7 SYMBIOS IS BETWEEN MIAMI BLUE LARVAE AND ANTS Introduction Over 75 percent of lycaenid larvae with known life histories associate with ants (Pierce et al. 2002). Such myrmecophilous relationships may be mutualistic to varying degrees or even parasitic whereby lar vae are predatory in ant nests (Pierce & Mead 1981; Fiedler & Maschwitz 1988; New 1993). The resulting communication between larvae and ants is mediated by a complex array of tactile, chemical, and often audible signals (DeVries 1990). Specifically, larv ae possess highly specialized organs that can extrude alarm, reward or appeasement chemicals. In response, tending ants often protect the surrounding larvae from a variety of natural predators and parasitoids, and thus can potentially provide a benefit fo r survival (Thomas 1980; Webster & Nielson 1984; Pierce & Easteal 1986; Savignano 1994). The presence of an ant guard has also led to larger adults in the related butterfly Hemiargus isola (Reakirt) When larvae are reared with ants they may expend less energy than when they are untended and therefore develop into larger potentially more fecund adults (Wagner & Martinez del Rio 1997). Cushman & Murphy (1993) suggest that ant associations may also play an important role in the persistence of lycaenid popu lations. They additionally propose that species with a dependence on ants, whether facultative or obligatory, display an increased sensitivity to environmental change, and thus are more susceptible to endangerment than species that lack ant associations. The Miami blue, Cyclargus thomasi bethunebakeri (Comstock & Huntington) (Lycaenidae), state endangered. Once commonly found in tropical coastal hammocks and their associated margins, beachside scrub and tropical pine rocklands fro m the southern Florida mainland south through the Florida Keys to Key West and the Dry Tortugas

PAGE 130

130 numerical abun dance has now been reduced to two remaining metapop ulation s within the boundaries of Bahia Honda State Park in the Lower Keys and the islands of Key West National Wildlife Refuge (Klots 1964; Kimball 1965; Lenczewski 1980; Minno & Emmel 1993; Ruffin & Glassberg 2000 ; Calhoun et al. 2002 ; FWC 2003 ). Develo ping larvae of C. thomasi bethunebakeri have be en shown to be tended by ants in the genus Camponotus but the extent of the relationship has remained poorly understood (Minno & Emmel 1993). Recent population studies of the butterfly at Bahia Honda State Pa rk and additional reintroduction sites within Everglades National Park confirm a continued association. Here, I provide a more detailed account of this relationship, identify additional ant associates and potential predatory ant species, and discuss the c orresponding implications for the conservation and recovery of this critically imperiled butterfly. Methods the Flamingo Campground, Rowdy Bend Trail, and Bear Lake Road sites in southern portions of Everglades National Park during daylight hours on 24 27 May 2004 and 31 July 2 August 2004. These areas contain small numbers of C. thomasi bethunebakeri either as part of the remaining natural metapopulation or as relocat ed individuals. Ants were surveyed on and around Caesalpinia bonduc (L.) Roxb. (Fabaceae) the current primary larval host plant of the Miami blue, by extensive hand collecting and baiting. Sugar baits consisting of a series of at least ten index cards e ach with approximately 10g of crushed pecan cookie, were placed along a transect passing along the base of C. bonduc plants. Baits were left out for one hour, at which time the resulting cards with ants were collected in Ziploc style plastic bags. Additi onally, when larvae of C. thomasi bethunebakeri were found in association with ants, 1 2 ant specimens were

PAGE 131

131 collected from the tended larvae. This research was conducted under permits EVER 2004 SCI 0038 and 5 04 58 for Everglades National Park and Bahia H onda State Park. Finally, to provide additional detail on the structure of the larval ant organs, three C. thomasi bethunebaker i larvae from the captive colony maintained at the University of Florida were preserved and used for SEM and Auto Montage photo graphic analysis. Larvae were placed in near boiling water for 60 seconds, transferred to 25% ethanol for two hours, 50% ethanol for another two hours, and stored in 75% ethanol before they were photographed. Results A total of 18 ant species were collect ed in Everglades National Park and Bahia Honda State Park and results are summarized in Table 7 1. Of the 18 collected species, five were confirmed tending larvae of Cyclargus thomasi bethunebakeri : Camponotus floridanus, Camponotus planatus, Crematogaster ashmeadi, Forelius pruinosus, and Tapinoma melanocephalum. C. floridanus were found tending larvae every time larvae were encountered in the field when other ants were not present. All other ants were encountered 1 2 times in the field, and no two speci es of ants tended larvae at the same time. There were typically two ants tending a larva at a time, with the exception being Crematogaster ashmeadi which tended larvae in higher numbers (Figure 7 1) An additional two species, Paratrechina longicornis and Paratrechina bourbonica have been named as potential ant associates. The former species has been found in proximity to C. thomasi bethunebakeri larvae and appears to tend them, although encounters are brief. The latter species was observed tending l arvae of another lycaenid, Strymon martialis ( Herrich Schffer ) on Caesalpinia bonduc at Bahia Honda State Park. No predation by these ants was observed in the field.

PAGE 132

132 Details of the ant organs of Cyclargus thomasi bethunebakeri are shown in Figure 2. Se cond through fourth instars possess a dorsal nectary organ (=honey gland) with associated pore cupolas (=lenticles) on A7 and a pair of eversible tentacle organs on A8. Abdominal segments A7 and A8 are fused dorsally. Tentacle organs were observed to eve rt independently in the field when stimulated by attendant ants and liquid droplets from the honey gland were actively imbibed by all species of ants. Camponotus floridanus became excited and agitated, evidenced by increased body and antennal movements, w hen the tentacle organs were everted. D iscussion This study definitively finds Camponotus floridanus as the species in primary attendance of the endangered Cyclargus thomasi bethunebakeri larvae. C. floridanus is a native species occurring throughout F lorida and is nocturnal; they are commonly found foraging on nickerbean and tending C. thomasi bethunebakeri larvae in both the Everglades and Bahia Honda locations. Camponotus planatus is a diurnal species, but is not commonly found tending larvae (it ha s never been found tending larvae in the Everglades and only once in Bahia Honda). This diurnal species may tend larvae when they are more abundant, and C. planatus may be important in protecting the species during the day. Buckley & Gullan (1991) have s hown that more aggressive ants provide better protection for soft scales and mealybugs, and as a corollary the large and potentially aggressive Camponotus species may prove effective in deterring predators and parasitoids from harming larvae (Axn 2000). Crematogaster ashmeadi have been found in great numbers tending i ndividual larvae in Bahia Honda but have not yet been observed in the Everglades. Again, these ants are not commonly found tending larvae, but when they are, there are at least five ants fo und, demonstrating the quality of their trailing and recruitment signals. Other Crematogaster species have been found worldwide to tend lycaenid larvae and this genus seems predisposed to lycaenid

PAGE 133

133 symbioses (Atsatt 1981; Fiedler 1991; Pierce et al. 2002; Saarinen 2006). These ants are equipped with a flexible abdomen and attached sting; despite the small size they are capable of defending larvae from other ants or harmful invertebrates. Both Forelius pruinosus and Tapinoma melanocephalum may be opportuni stic ants imbibing food rewards from C. thomasi bethunebakeri larvae. Their behavior suggests no elevated protection for the larvae they tend; however, their presence may potentially deter predators. Both Paratrechina longicornis and bourbonica, along wi th Tapinoma melanocephalum may not provide demonstrative protection of larvae, they may also defend the larvae as any other food source. At the very least, they tolerate these larvae and coincidentally tend instars that feed on nickerbean flowers and buds where the ants are also gathering nectar. In an assessment of potential ant partners, it was shown that Tapinoma sessile (Say) widely distribute d North American lycaenid Glaucopsyche lygdamus (Doubleday) providing no significant cost or benefit (Fraser et al. 2001). No ant association should be overlooked or underestimated, however. Researchers of the critically imperiled European lycaenid butt erfly Maculinea rebeli (Hirschke) host ant species prior to further management conservation strategies in order to avoid failure of management programs or even damage to populations on the edg 2003). The facultative nature of the ant associations are clear, encompassing four genera (five including Paratrechina ) and three subfamilies; Camponotus and Paratrechina are Formicinae, Crematogaster is Myrmicinae, and T apinoma and Forelius are Dolichoderinae. Larvae may

PAGE 134

134 by Henning (1983). These chemicals, primarily from the tentacle organs and potentially from the pore cupo las, may serve to either alarm, excite, or appease the ants. Further study into the chemical secretions of all organs, including the dorsal nectary organ but specifically of the ons. Further comparisons of each ant species own alarm and attractant pheromones with those isolated from larval volatiles may further elucidate ant larval relationships, including whether ants are specifically, chemically targeted and whether others are simply opportunistic tenders. No interactions between the other identified ants and larvae were observed. Several of these ants, however, may be predating on the larvae at other times. Species of special concern include Solenopsis invicta Buren and Was mannia auropunctata (Roger) in Bahia Honda State Park. All three Pseudomyrmex species may also potentially be predators, possibly excepting P. simplex (F. Smith) due to its small size. It is fortunate that in these locations, there are also abundant colo nies of Camponotus floridanus. Further field observations, especially at night when C. floridanus are most active, need to be carried out to assess interactions of the ant mosaic when C. thomasi bethunebakeri larvae are present. This study shows the per sistence of Wasmannia auropunctata on Bahia Honda State Park (first recognized there by Deyrup et al. 1988). This invasive tramp ant is native to the New World tropics so its presence in the Florida Keys is not a surprise, but it may be a cause for concer n. W. auropunctata also known as the little fire ant is an opportunistic feeder that forages day and night and that bears a painful sting. This ant has displaced several endemic invertebrates in the Galapagos Islands, and it has the potential to do so el sewhere (Meier 1994). It, along with the red imported fire ant Solenopsis invicta, is associated with a decline in endemic biodiversity (Meier, 1994; Wojcik 1994). Neither ant has been found near relocated colonies of C. t.

PAGE 135

135 bethunebakeri in the Everglade s nor has either ant been directly observed killing larvae or adults in Bahia Honda. S. invicta mounds are not as large or as extensive on Bahia Honda as they are in extremely disturbed habitats, and perhaps mound density is not large enough to impact the C. t bethunebakeri metapopulations. Further field work will need to examine predation rates by ants, and the specific impact that these invasive ants may have on endemic butterfly species, especially species of special concern.

PAGE 136

136 Figure 7 1 Cremato gaster ashmeadi ants tending a late instar Cyclargus thomasi bethunebakeri larva. Several other C. ashmeadi ants were present but not visible in this photo. Reprinted with permission from Saarinen, EV and J.C. Daniels. 2006 Miami blue butterfly larvae (L epidoptera: Lycaenidae) and ants (Hymepnoptera: Formicidae): new information on the symbionts of an endangered taxon. Florida Entomologist 89 :69 74.

PAGE 137

137 Figure 7 2. Details of Cyclargus thomasi bethunebakeri fourth instar ca terpillar a, dorso posterior abdominal segments (A7 A8) showing ant organs; b, detail of cuticular setae; c, tentacle nectary organ inverted; d, tentacle nectary organ everted (cuticular setae omitted); e, dorsal nectary organ bordered by pore cupolas.

PAGE 138

138 Table 7 1 Ants of Everglades National Park (ENP) and Bahia Honda State Park (BHSP) in Florida collected in proximity to larvae of the C. t. bethunebakeri 1, ants collected by larvae in present study; 2 ants listed in Deyrup et al. (1988); 3 ants listed in Ferster & Prusak (1994); Ant Status; S, confirmed symbiont of MBB ; pS, potential symbiont; u unknown; P, potential predator. ENP BHSP Ant Status Subfamily Pseudomyrmicinae Pseudomyrmex elongatus 1 2 P Pseudomyrmex gracilis 1 P Pseudom yrmex simplex 1 P Subfamily Myrmicinae Crematogaster ashmeadi 1, 2 S Monomorium floricola 1 2 u Pheidole dentata 1 2 u Pheidole floridana 1 1 u Solenopsis invicta 1, 2 P Solenopsis geminata 1 1, 2 u Tetramorium simillimum 1 u Wasma nnia auropunctata 1 1, 2 P Subfamily Dolichoderinae Forelius pruinosus 1 1, 2 S Tapinoma melanocephalum 1 S Subfamily Formicinae Brachymyrmex obscurior 1 1, 2 u Camponotus floridanus 1, 3 1, 2 S Camponotus planatus 1 1 S Paratrechina bour bonica 1 pS Paratrechina longicornis 1 1, 2 P

PAGE 139

139 CHAPTER 8 CONCLUSIONS This study successfully investigated the population genetics and corresponding conservation implications for the endangered Miami blue butterfly Cyclargus thomasi bethunebakeri Suc h knowledge is crucial for developing effective conservation and management strategies for this taxon and can help guide the decision making process for other threatened invertebrate taxa. I began this project by creating a microsatellite library to deve lop markers appropriate for addressing population level questions Developing microsatellites for butterflies can be difficult because of the high occurrence of null alleles (Zhang 2004), but microsatellites are the most appropriate markers for discerning gene flow a nd molecular variation within population s Additionally, I utilized and tested a relatively noninvasive sampling technique that allowed me to sample living butterflies without negatively impacting them Coupled together, the twelve polymorphi c microsatellite markers and noninvasive sampling method allowed me to assess genetic diversity within populations of this imperiled taxon (Chapter 2). I employed the use of microsatellite markers to explore historical genetic diversity by genotyping museu m specimens of butterflies using single legs and labial palps (Chapter 3) Only one other study has successfully amplified microsatellite DNA from historical butterfly specimens but this study destructively sampled the heads (Harper et al. 200 6). Howeve r, the Harper et al. ( 2006 ) study was able to amplify three microsatellite loci while I was only able to amplify one My study of C t bethunebakeri from museum collections quantified the diversity present in a historical population present in Key Largo, Florida. According to the results from a single microsatellite locus, d iversity was consistently high within the historical Key Largo population My study then expanded beyond the genetic field to include an analysis of the

PAGE 140

140 geographic distribution of th is butterfly. Through examination of several museum collections and discussions with lepidopterists and collectors I discovered that C t bethunebakeri were locally common in some areas of southern Florida and the Florida Keys up until the 1980s. Durin g favorable years, small populations of butterflies were also established in more northern counties and the y even colonized some inland areas, although none of these remained for more than a few seasons. The C t bethunebakeri population centers appear t o be in south east Florida and the Florida Keys We can use the k nowledge of historical distribution and population locations to plan future reintroduction events with captive bred butterflies. We can also continue to search for additional wild population s of butterflies within areas that showed a high historical incidence of these butterflies. The historical distribution of butterflies throughout the Florida Keys further sugg ests a metapopulation structure. This involve d partially isolated populations of C t bethunebakeri that were loosely connected via gene flow and that underwent localized extinction and colonization events along the chain of suitable habitat in the Florida Keys. Knowledge of historic al levels of gene flow and population distributi on will be useful in planning future reintroduction measures While I could not assess historical gene flow between populations, I was able to confirm the presence and distribution of past populations by cataloguing museum specimens and through conversati ons with lepidopterists. The genetic diversity of the historical population in Key Largo was high, but diversity within this population was observed to decrease over the decades that I sampled (1940 1979) A c omparison of historical diversity from the Ke y Largo population with the extant population on Bahia Honda State Park (BHSP) revealed differences in allelic frequencies, but only minor differences in the overall number of alleles. T he historical population had a greater distribution of medium high fr equency alleles, while the BHSP

PAGE 141

141 population show ed dominance of a few alleles (Chapter 3). Again, this study was hindered by the presence of data from only one microsatellite locus, but this is an issue when dealing with museum materials and degraded DNA. My investigations into the genetic diversity of the BHSP and Key West National Wildlife Refuge ( KWNWR ) populations reveal ed adequate levels of diversity within the 12 polymorphic microsatellite marker loci used to assess the populations. There was no ev idence of a recent bottleneck within the BHSP population, but there may have been one in or prior to the 1980s that wa s undetectable with current methodologies. The KWNWR population did show signs of a bottleneck, which provides support for the hypothesis that it is a newly founded population. The BHSP population had higher levels of diversity than originally predicted from the low census estimates (Chapter 4) Gene flow between habitat patches on BHSP is crucial for the maintenance of this diversity and likely imperative for the long term survival of the population. Future reintroduction efforts need to be mindful of this and mimic the metapopulation structure present on BHSP. There is gene flow between distant habitat patches on BHSP and the patches a long the Main Road of BHSP appear to be especially important in maintaining this gene flow. journey between the far ends of BHSP. met Honda, one that is similar to the historical metapopulation structure and distribution of populations on a larger scale. The captive colony has been extremely successful at producing in dividuals for reintroduction. However, g enetic monitoring showed that the captive c olony significantly diverged from the BHSP population after only a few generations with no new infusion of material from the wild (Chapter 5). Measures of genetic potential, such as allelic richness,

PAGE 142

142 remained st able in the BHSP population but were observed to decrease over time in captivity. I have also recorded the appearance of novel alleles in the captive colony, which are of concern as they may represent variation that would be detrimental to individ uals rei ntroduced to the wild. Individuals introduced from the wild will help prevent inbreeding depression in the captive colony and likely boost the fitness of organisms bred for reintroduction efforts (Tallmon et al. 2004). Laboratory investigations into the effects of a small number of individuals founding new colonies did not go as expected, but did yield information which will be useful in other conservation e ndeavors. The discovery of a novel baculovirus highlights the vulnerability of captive bred popul ations and has led to modified laboratory techniques. The importance of demographic factors like eclosion time, availability of suitable mates, etc. highlighted the importance of stochastic factors in colony founding and population persistence. The genet ics of newly founded colonies were not able to be well studie d in these experiments due to small sample size but the demographic factors highlighted how populations may be vulnerable before or while molecular issues are involved (Chapter 6) The presence of symbiotic ant partners adds to the charismatic nature of the C t bethunebakeri story. My surveys showed that several ant genera and species were found tending C t bethunebakeri larvae and no predation by ants was directly observed. The presence o f invasive ant taxa should be monitored or addressed by future study as these ants, while they might not directly prey on C t bethunebakeri larvae, may be responsible for disrupting the mosaic ant community which may eventually negatively affect immature C t bethunebakeri (Chapter 7)

PAGE 143

143 Future Research The markers I developed in this study will be useful for future conservation efforts with captive and wild populations of this butterfly and may even be applied to other members of the genus Cyclargus If other members of the genus are imperiled in the Caribbean, these microsatellite markers may be useful in elucidating population structure and assessing genetic diversity. The microsatellite markers were useful in discerning the genetic diversity present in the newly found population i n KWNWR This discovery on the remote, uninhabited islands of the refuge again speaks and lends hope for the recovery of the taxon Despite extensive surveying efforts, I still believe it is likely that additional populations of butterflies exist in the Florida Keys, although these may be ephemeral and low in population number. The KWNWR population offers additional opportunities for study, not only o n the biology of this organism but also on greater theoretical issues in conservation biology. Studies into host plant use and preference can be continued with this taxon as the larvae of the KWNWR population feed on a plant different from that on BHSP, a n issue that may lead to ecological speciation or represent an ancestral polymorphism. Additionally, concerns of outbreeding depression are great within the conservation community, and experimental breeding between BHSP and KWNWR individuals may shed ligh t on this topic. Conservation projects are often stuck in a dilemma between inbreeding depression and outbreeding depression (Edmands 2006) and the nature of these two populations coupled with the ability to raise this taxon in captivity may clarify this issue. From a management perspective, the KWNWR population al so offers new genetic diversity, diversity that may prove essential in the long term solution for returning fit individuals to the wild.

PAGE 144

144 Recent theoretical models have been proposed to explain (Tallmon et al. 2006) and data from C t bethunebakeri may be useful in testing these models. uced variability. This may have been how the metapopulation structure of C t bethunebakeri worked historically, with immigrant butterflies supported by theoret ical and experimental work with other metapopulations (Thrall et al. 1998; Ingvarsson 2001) and provide an interesting avenue for future research with C t bethunebakeri Concerns of genetic adaptation to captivity may be addressed by dividing the curre nt captive colon y into multiple captive colony segments These colonies could be housed at additional institutions or maintained as discrete entities within the McGuire Center for Lepidoptera and Biodiversity. If a new virus or pathogen emerged and destr oyed one colony, the taxon would be safeguarded by the presence of additional colonies in other locations. The c ontinuous release of captive bred individuals into the wild should be maintained in order to impro ve the chances of successfully founding new w ild colonies. Additionally, newly founded populations should be monitored regularly and definitive data about the success of larval vs. adult releases gathered and evaluated A comparison of the successes of reintroductions fro m captive bred individuals vs. translocated individuals from BHSP would also aid in reintroduction efforts Given the expense of captive colonies and number of threatened taxa, such data could prove useful for other endangered invertebrate management projects. Conservation Implic ations The Xerces blue butterfly ( Glaucopsyche xerces ) was a lycaenid butterfly endemic to the coastal dunes of northern California. The development of the dune habitat undermined the remaining populations of these butterflies, and eventually led to its e xtinction in the 1940s. We

PAGE 145

145 ar e in a time of mass extinction, with species such as the Xerces blue butterfly lost every day. Insect extinctions comprise the neglected majority of extinctions in our ongoing biodiversity crisis (Dunn 2005) and m any insect and arthropod taxa are lost before they are even known to science. Recovery and conservation programs may be able to salvage some insect taxa, but the remaining diversity is lost forever. While recovery pro grams such as that developed fo r C t bethunebak eri are noteworthy it is more important that the factors leading to imperilment are identified and mitigated before species become endangered. Conservation endeavors should act more preemptively and aim to preserve species before they become threatened. This begin s with the protection of habitat and a the importance of knowledge of an organism s biology and ecology to it s long term persistence These butterflies exist in a metapopulation framework for continued survival in the dynamic environment of s outh Florida one that is dominated by major environmental changes precipitated by hurricanes and droughts. This butterfly has been able to exist in this system for a long time before people colonized the region, and its continued persistence in the region depends on a habitat network with colo nies linked by gene flow between patches. Basic biological field work has uncovered the importance of habitat connectivity a nd the significance has been supported by my molecular studies The captive colony has been a pr oductive endeavor in that we have learned useful biological information and have been successful at producing numerous organisms for reintroduction However, t he bureaucracy surrounding the reintroduction efforts and conflicting stakeholder interests may have hindered the initial success of the reintroduction goals The development of the Miami Blue Working Group has helped mitigate some of these conflicts

PAGE 146

146 This group can serve as a model for other programs that aim to minimize discord between stakeholders and result in successful collaborations and reintroduction efforts (Daniels 2009). A unified approach to conservation encompassing genetics, ecology, and b ehavior will offer a more comprehensive program with a higher chance of success, and the C t bethunebakeri recovery program has made great strides in encompassing all of these fields. As scientists we should further use the data we generate to make reco mmendations for pra ctical solutions in management, especially where genetic data are concerned. Guidelines for the use of genetic data in both the listing process and management plan ning need to be drafted (Haig et al. 2006; Fallon 2007; Waples et al. 200 7 ; Vernesi et al. 2008 ). Given the magnitude of the biodiversity crisis, there will only be more taxa listed in the future. New guidelines to assist in the effective use of genetic data will help standardize and streamline the listing process and aid in future decisions that will benefit the multitude of threatened organisms.

PAGE 147

147 APPENDIX A LIST OF ABBREVIATIONS ABI: Applied Biosystems, Inc. AMNH: American Museum of Natural History A R : Allelic richness BHSP: Bahia Honda State Park, Florida BNP: Biscayn e National Park bp: Basepairs BSA: Bovine Serum Albumin CC: Captive colony (of Cyclargus thomasi bethunebakeri ) CMNH: Carnegie Museum of Natural History CNHM: Chicago Natural History Museum (the Field Museum) DEP: Department of Environmental Protection DNA: Deoxyribonucleic acid D S : Genetic distance DJSP: Dagney Johnson Key Largo Hammocks Botanical State Park ENP: Everglades National Park ESRI: Environmental Systems Research Institute (Redlands, California) f: Inbreeding coefficient FAM: Carboxyfl uorescein (blue) FLMNH: Florida Museum of Natural History F IS : The proportion of inbreeding in a population due to inbreeding within the subpopulation. This is the correlation between homologous alleles in an individual relative to the subpopulation. F I T : The proportion of inbreeding in a population due to inbreeding within the subpopulation and differentiation among subpopulations. This is the correlation between homologous alleles in an individual relative to the total population. F ST : The most inclu sive measure of population substructure, it measures the proportion of inbreeding in the subpopulation relative to the total population (a level of genetic divergence among subpopulations). This value is used to estimate gene flow and population subdivisi on. FWC: Florida Fish and Wildlife Conservation Commission formerly abbreviated as FFWCC GIS: Geographic Information System GIS, Inc.: Genetic Identification Services, Inc. (Chatsworth, California) H O : Observed heterozygosity in a population based on t he Hardy Weinberg equilibrium principle. H E : Expected heterozygosity in a population based on the Hardy Weinberg equilibrium principle. HEX: Hexachlorofluorescein (green) HWE: Hardy Weinberg equilibr ium. The equilibrium state of genotype frequencies ach ieved in an infinitely large, idealized population that is not subjected to mutation, migration, or selection. H O and H E values are used to test whether allelic frequencies conform to HWE. IAM: Infinite alleles model ; used to explain microsatellite mutati on patterns. N : Number of samples N A : Number of observed alleles N E : Effective population size.

PAGE 148

148 n e : Number of effective alleles NPS: National Parks Service P HWE: Probability that a population is in Hardy Weinberg equilibrium. This value of often 2 test, but I use the exact test as recommended by Excoffier et al. (2005) 2 test to a table of arbitrary size and uses a modified Markov chain walk to produce a more computationally efficient result (Guo & Th ompson 1992). PCR: Polymerase chain reaction Rpm: Rotations per minute (of centrifuge). R ST : F ST analogue for microsatellite markers (developed by Slatkin 1995) SEM: Scanning electron microscope SMM: Stepwise mutation model used to explain microsatelli te mutation patterns. SMNH: Smithsonian Natural History Museum T a : Annealing temperature for PCR T m : Melting temperature TET: Tetrachlorofluorescein (yellow) TPM: Two Phase Model ; used to explain microsatellite mutation patterns. USFWS: United States Fish and Wildlife Service UV : Ultraviolet

PAGE 149

149 APPENDIX B DNA EXTRACTION METHODS Day 1 E xtraction from whole butterfly: 1) Measure the forewing from the wing tip to where the wing meets the body. Record this in the spreadsheet and the lab notebook. 2) Remove t he wings and abdomen with a razorblade. Store the wings and abdomen in a labeled tube. 3) Slice open the thorax with the razorblade and expose tissue by continuing to cut the thorax into pieces, taking care not to lose any of the thorax pieces. Only proces s one specimen at a time and make sure not to cross contaminate samples. 4) Put cut up thorax into a clean, labeled tube (use the 2mL tubes). Put the thorax into the tube by using the razorblade or forceps. Add 180 L Buffer ATL. Macerate the tissue with a sterile blue pestle. Grind up the tissue as much as possible in the tube and keep as much of the tissue as possible in the tube. 5) Add 20 L Proteinase K. Mix by vortexing for 30 seconds. If necessary, do a short centrifuge to make sure that all the ti ssue is in the liquid and is not stuck on the side of the tube or on the lid. Incubate at 55 C overnight. PROCEED TO DAY 2 E xtraction from butterfly wing fragment : 1) Very carefully, remove the alcohol from the tube containing the wing fragment. I recom mend centrifuging the sample for 2 minutes (at 13,200 rpm) to affix the wing fragment to the tube. Remove all alcohol by pipette. 2) Use the vacuum for 5 minutes to remove the rest of the alcohol. 3) Add 180 L Buffer ATL and add 2 copper BBs (only use auto claved BBs, never use small glass beads). 4) Vortex each sample for 30 seconds, ensuring that the BBs macerate the wing fragment. Tilt the tube around in the vortex to make sure that the BBs move all around the tube. 5) Very carefully, pour the BBs out of t he tube into a waste cup. This is best done by pouring the BBs into the lid of the 2 mL tube and then dumping them out of the lid. You may want to practice pouring BBs out of a tube with water in it (and not a valuable sample). If it is very difficult t o remove the BBs, switch to macerating wing tissue with the sterile blue pestles.

PAGE 150

150 6) Add 20 L Proteinase K. Centrifuge for a short spin (or for 1 minute if necessary) to make sure that all the fragment is in the liquid and not stuck on the sides or lid of the tube. Incubate at 55 C overnight. PROCEED TO DAY 2 Day 2 1) Remove the rack of tubes from the 55 C water bath. Turn the bath up to 70 C. 2) Add 20 L (of 20 mg/mL concentration) RNAse to each sample. Let it sit for 5 minutes at room temperature. Mi x by inverting tubes. Continue to Step 4 for wing fragment extractions. There is no need to centrifuge the tubes with wing fragments as there is not enough tissue to clog the spin columns. 3) Whole butterfly/thoracic extractions only : centrifuge the tube f or 5 min. at 13,200 rpm. Transfer the supernatant to a new tube. Do this with a P1000 for most of the liquid and then get the remaining liquid with a P200 (yellow tip). Try to get as much liquid as possible into the new tube while minimizing the amount of tissue. Discard the tube with the tissue in it. 4) Add 200 L Buffer AL to the sample and mix well by vortexing. Incubate at 70 C for 10 minutes. Flick or invert the tube a few times after removing it from the bath. Return the bath to 55 C. 5) Add 2 00 L 100% ethanol and vortex for 15 seconds. 6) the mixture from step 4 into the DNeasy mini spin column placed in a 2 mL collection tube. Centrifuge at 8,000 rpm for 1 minute. Discard flow through and collection tube. 7) Place the DNeasy mini spin column in a new 2mL collection tube (provided). Add 500 L Buffer AW1 and centrifuge for 1 minute at 8,000 rpm. Discard flow through and collection tube. 8) Place the DNeas y mini spin column in a new 2mL collection tube (provided). Add 500 L Buffer AW2 and centrifuge for 3 minutes at 13,200 rpm. Discard flow through and collection tube. 9) Carefully remove the column so that the column does not come into contact with the flow through. If this does happen, empty the collection tube and reuse it in another centrifuge step for 1 min. at 13,200 rpm (rotations per minute) 10) During the 3 minutes that the samples are centrifuging, put the Buffer EB into the warm water bath. 11) Plac e the DNeasy mini spin column in a clean, labeled 2mL microcentrifuge tube (not provided). For whole body/thorax extractions : p ipette 100 L of warm Buffer EB (=10mM TRIS) directly onto the DNeasy membrane. For wing fragment extractions : pipette 50 L of warm Buffer EB (=10mM TRIS) directly onto the DNeasy membrane.

PAGE 151

151 Incubate at room temperature for 3 5 minutes. Centrifuge for 1 minute at 8,000 rpm to elute. Throw away the spin column and save the labeled tube in the 20 C freezer.

PAGE 152

152 APPENDIX C ALLELE FREQUENCY TABLES Table C 1. Allele frequency for locus CthB11. Table C 2. Allele frequency for locus CthB101. Table C 3. Allele frequency for locus CthB103.

PAGE 153

153 T able C 4. Allele frequency for locus C thB106. Table C 5. Allele frequency for locus CthB115. Table C 6. Allele frequency for locus CthB117.

PAGE 154

154 Table C 7. Allele frequency for locus CthB119. Table C 8. Allele frequency for locus CthC12. Table C 9. Alle le frequency for locus CthC116.

PAGE 155

155 Table C 10. Allele frequency for locus CthC124. Table C 11. Allele frequency for locus CthC127. Table C 12. Allele frequency for locus CthD7.

PAGE 156

156 APPENDIX D PRIVATE ALLELE LIST Table D 1. Private allele list. Pop ulation Locus Allele (bp) Freq uency BHSP 2005 2006 B11 171 0.005 BHSP 2005 2006 B103 216 0.004 BHSP 2005 2006 B106 208 0.009 BHSP 2005 2006 B115 301 0.018 BHSP 2005 2006 B115 313 0.005 BHSP 2005 2006 B117 152 0.018 BHSP 2005 2006 B11 7 155 0.013 BHSP 2005 2006 B117 176 0.009 BHSP 2005 2006 B117 189 0.004 BHSP 2005 2006 B119 135 0.005 BHSP 2005 2006 B119 165 0.009 BHSP 2005 2006 B119 171 0.005 BHSP 2005 2006 B119 180 0.005 BHSP 2005 2006 C12 189 0.004 BHSP 2005 2006 C12 204 0.02 2 BHSP 2005 2006 C12 207 0.009 BHSP 2005 2006 C12 210 0.004 BHSP 2005 2006 C12 228 0.004 BHSP 2005 2006 C124 230 0.009 BHSP 2005 2006 C124 239 0.004 BHSP 2005 2006 C124 245 0.009 BHSP 2005 2006 C124 248 0.009 BHSP 2005 2006 D7 221 0.005 BHSP 2005 2006 D7 225 0.009 BHSP 2005 2006 D7 241 0.005 BHSP 2005 2006 D7 249 0.009 BHSP 2005 2006 D7 285 0.005 BHSP 2005 2006 D7 293 0.005 BHSP 2005 2006 D7 313 0.009 CC 2005 2006 B11 233 0.067 CC 2005 2006 B11 241 0.010 CC 2005 2006 B101 192 0.005 CC 2005 2006 B106 199 0.025 CC 2005 2006 B115 235 0.005 CC 2005 2006 B115 262 0.005 CC 2005 2006 B115 307 0.010 CC 2005 2006 B117 143 0.005 CC 2005 2006 B119 150 0.020 CC 2005 2006 B119 201 0.005 CC 2005 2006 C12 201 0.043 CC 2005 2006 C124 242 0.031 CC 2005 2006 C127 176 0.027

PAGE 157

157 Table D 1. Continued Population Locus Allele (bp) Frequency CC 2005 2006 C127 182 0.005 CC 2005 2006 C127 185 0.043 CC 2005 2006 C127 188 0.022 CC 2005 2006 C127 200 0.027

PAGE 158

158 LIST OF REFERENCES Allen, T.J., J.P. Brock, and J. Glassberg. 2005. Caterpillars in the field and garden. A field guide to the butterfly caterpillars of North America. Oxford University Press, Inc., New York, NY. Allendorf, F.W. 1986. Genetic drift and the loss of alleles versus heterozygosity. Zoo Biolog y 5 :181 190. Anthony, N., G. Gelembiuk, D. Raterman, C. Nice, and R. Ffrench Constant. 2001. Isolation and characterization of microsatellite markers form the endangered Karner blue butterfly Lycaeides melissa samuelis (Lepidoptera). Hereditas 134 :271 273. Atsatt, P.R. 1981. Lycaenid butterflies and ants: selection for enemy free space. The American Naturalist 118 :638 654. Avise, J. C. 1996. Introduction: the scope of population genetics. Pages 1 9 in J.C. Avise and J. L. Hamrick, editors. Conservation G enetics : case histories from nature. Chapman and Hall, New York, N ew Y ork Axn, A.H. 2000 Variation in behavior of lycaenid larvae when attended by different ant species. Evolutionary Ecology 14 :611 625. Bailey, N.W., C.M. Garcia, and M.G. Ritchie. 2007 Beyond the point of no return? A comparison of genetic diversity in captive and wild populations of two nearly extinct species of Goodeid fish reveals that one is inbred in the wild. Heredity 98 :360 367. Blint Z., and K. Johnson. 1995. Neotropical Poly ommatinae diversity and affinities. I. Relationships of the higher taxa (Lepidoptera: Lycaenidae). Acta Zoologica Academiae Scien tiarum Hungaricae 41 :211 235. Brevignon, C. 2002. Cyclargus Nabokov 1945 (Lepidoptera, Lycaenidae, Polyommatinae, Polyommatini). Lambillionea CII :97 102. Sanibel Island, Florida. Sanibel Captiva Conservation Foundation. Briscoe, D.A., J.M. Malpica, A. Robertson, G.J. Smith, R. Frankham, R.G. Banks, and J.S.F. Barker. 1992. Rapid loss of genetic variation in large captive populations of Drosophila flies: implications for the genetic management of captive populations. Conservation Biology 6 :416 425. Buckley, R. and P. Gullan. 1991. More aggressive ant species (Hymenoptera: Formicidae) provide better protection for soft scales and mealybugs (Homoptera: Coccidae, P seudococcidae). Biotropica 23 :282 286.

PAGE 159

159 Byers D.L., and D.M. Waller 1 999 Do plant populations purge their genetic load? Effects of population size and mating history on inbreeding depression. Annual Review of Ecological Systems 30 :479 513. Calhoun J., J. Slotten, and M. Salvato. 2002. The rise and fall of tropical blues in Florida: Cyclargus ammon and Cyclargus thomasi bethunebakeri Holartic Lepidoptera 7 :13 20. Calvo D., and J.M. Molina 2005 Fecundity body size relationship and other reproductive aspects of Streblote panda (Lepidoptera: Lasiocampidae). Annals of the Entomological Society of America 98 :191 196. Cannon, P. 2007. Rarest of the very rare: local photographer finds colonies of Miami blue butterflies. Keys Sunday, 7 January 2007:16 18. Carroll, S.P., and L.E. Loye. 1987. Specialization of Jadera species (Hem iptera: Rhopalidae) on seeds of the Sapindaceae, and coevolution of defense on attack. Annals of the Entomological Society of America 80 :373 378. Carroll, S.P., M. Marler, R. Winchell, and H. Dingle. 2003 Evolution of cryptic f light m orph and l ife h istor y d ifferences d uring h ost r ace r adiation in the s oapberry b ug, Jadera haematoloma Herrich Schaeffer (Hemiptera: Rhopalidae) Annals of the Entomological Society of America 96 :135 143. Carroll, S.P., and J. Loye. 2006. Invasion, colonization, and disturbanc e; historical ecology of the endangered Miami blue butterfly. Journal of Insect Conservation 10 :13 27. Chapin, S.F., E.S. Zavaleta, V.T. Eviner, R.L. Naylor, P.M. Vitousek, H.L. Reynolds, D.U. Hooper, S. Lavorel, O.E. Sala, S.E. Hobbie, M.C. Mack, and S. D iaz. 2000. Consequences of changing biodiversity. Nature 405 :234 242. Charlesworth D., and B. Charlesworth. 1987 Inbreeding depression and its evolutionary consequences. Annual Review of Ecological Systems 18 :237 268. Charlesworth B., and D. Charleswor th 1999 The genetic basis of inbreeding depression. Genetics Research 74 :329 340. Clench, H. K. 1941. A new race of Hemiargus for the Bahamas (Lepidoptera: Lycaenidae). Memorias de la Sociedad de Historia Natural XV (4):407 408. Clench, H. K. 1942. The Ly caenidae of the Bahama Islands (Lepidoptera, Rhopalocera). Psyche 9/12 :52 60. Coltman, D.W. and J. Slate. 2003. Microsat ellite measures of inbreeding: a meta analysis. Evolution 57 :971 983. Comstock, W. P. and E. I. Huntington. 1943. Lycaenidae of the Ant illes (Lepidoptera: Rhopalocera). Annals of the New York Academy of Sciences XLV (2): 49 130.

PAGE 160

160 Cornuet J.M. and G. Luikart. 1996. Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics 1 44 :2001 2014. Crnokrak, P., and S. Barrett. 2002. Purging the genetic load: A review of the experimental evidence. Evolution 56 :2347 2358. Crone, E.E., D. Pickering, and C.B. Shultz. 2007. Can captive rearing promote recovery of endangered butterflies? An assessment in the face of uncertainty. Biological Conservation 139 :103 112. Crozier R.H., B. Kaufmann, M.E. Carew, and Y.C. Crozier. 1999. Mutability of microsatellites developed for the ant Camponotus consobrinus Molecular Ecology 8 :271 276. Cunningham A.A., and J.M. Frank. 1993. Veterinary observations of an arthropod mycosis with comparison of the efficacy of topical povidone iodine, nystatin, and chlortrimazole treatments. Journal of Zoo and Wildlife Medicine 24 :508 514. Cushman, J.H., and D.D. Mur phy. 1993. Conservation of North American lycaenids an overview. Pages 37 44 in T.R. New, editor. Conservation biology of Lycaenidae (butterflies). IUCN, The World Conservation Union, Gland, Switzerland. Daniels, J.C. 2009. Cooperative conservation effo rts to help recover an endangered south Florida butterfly. Insect Conservation and Diversity 2:62 64. Darwin C.R. 1876 The effects of cross and self fertilization in the vegetable kingdom. John Murray, London, England. DeVries, P.J. 1990. Enhancement of symbioses between butterfly caterpillars and ants by vibrational communication. Science 248 :1104 1106. Deyrup, M., Carlin, N., Tager, J., and Umphrey, G. 1988. A review of the ants of the Florida Keys. Florida Entomologist 71 :163 176. Dib, C., S. Faure, C Fizames, D. Samson, N. Drouot, A. Vignal, P. Millasseau, S. Marc, J. Hazan, E. Seboun. 1996. A comprehensive genetic map of the human genome based on 5,264 microsatellites. Nature 380 :152 154. Dinerstein, E., and G.F. McCracken. 1990. Endangered greater one horned rhinoceros carry high levels of genetic variation. Conservation Biology 4 :417 422. Dunn, R.R. 2005. Modern insect extinctions, the neglected majority. Conservation Biology 19 :1030 1036. Edmands, S. 2006. Between a rock and a hard place: evaluat ing the relative risks or inbreeding and outbreeding for conservation and management. Molecular Ecology 16 :463 475.

PAGE 161

161 England, P.R., G.H.R. Osler, L.M. Woodworth, M.E. Montgomery, D.A. Briscoe, and R.A. Frankham. 2003. Effects of intense versus diffuse popu lation bottlenecks on microsatellite genetic diversity and evolutionary potential. Conservation Genetics 4 :595 604. Evanno, G., S. Regnault, and J. Goudet. 2005. Detecting the number of clusters of individuals using the software Structure: a simulation s tudy. Molecular Ecology 14 :2611 2620. Excoffier L., G. Laval, S. Schneider. 2005. A rlequin version 3.0: an integrated software package for population genetics data analysis. Evolutionary Bioinformatics Online 1 : 47 50. Fallon, S.M. 2007. Genetic data and t he listing of species under the U.S. Endangered Species Act. Conservation Biology 21:1186 1195. Falush, D., M. Stephens, and J.K. Pritchard. 2003. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequenci es. Genetics 164 :1567 1587. Fauvelot, C., D.F.R. Cleary, and S.B.J. Menken. 2006. Short term impact of 1997/1998 ENSO induced disturbance on abundance and genetic variation in a tropical butterfly. Journal of Heredity 97 :367 380. Ferster, B., and Z. Prusa k. 1994. A preliminary checklist of the ants (Hymenoptera: Formicidae) of Everglades National Park. Florida Entomologist 77 :508 512. Fiedler K. 1991. Systematic, evolutionary, and ecological implications of myrmecophily within the Lycaenidae (Insecta: Lepi doptera: Papilionoidea). Bonner zoologische Monographien, Nr. 31 Zoologisches Forschungsinstitut und Museum Alexander Koenig, Bonn, Germany. Fiedler, K., and U. Maschwitz. 1988. Functional analysis of the myrmecophilous relationships between ants (Hymeno ptera: Formicidae) and lycaenids (Lepidoptera: Lycaenidae). II. Lycaenid larvae as trophobiotic partners of ants a quantitative approach. Oecologia 75 :204 206. Fiumera, A.C., L. Wu, P.G. Parker, and P.A. Fuerst. 1999. Effective population size in the cap tive breeding program of the Lake Victoria cichlid Paralabidochromis chilotes Zoo Biology 18 :215 222. Florida Fish and Wildlife Conservation Commission (FWC) 2003 Management Plan: Miami Blue Cyclargus =( Hemiargus) thomasi bethunebakeri Tallahassee, Flo rida. Accessed 11 February 2009. Available from http://myfwc.com/imperiledspecies/plans/Miami%20blue%20management%20plan.pdf Forbes, W.T.M. 1941. The Lepidoptera of the Dry Tortugas. Psyche 48 :147 148. Ford, M.J. 2002. Selection in captivity during support ive breeding may reduce fitness in the wild. Conservation Biology 16 :815 825.

PAGE 162

162 Frankel, O.H., and M.E. Soul. 1981. Conservation and evolution. Cambridge University Press, Cambridge, United Kingdom. Frankham, R. 2008. Genetic adaptation to captivity in spec ies conservation programs. Molecular Ecology 17 :325 333. Frankham, R. 1996. Relationship of genetic variation to population size in wildlife. Conservation Biology 10 :1500 1508. Frankham, R., J. D. Ballou, and D. A. Briscoe. 2003. Introduction to conservati on g enetics Cambridge University Press, Cambridge, United Kingdom. Frankham, R. D.M. Gilligan, D. Morris, and D.A. Briscoe. 2001. Inbreeding and extinction: effects of purging. Conservation Genetics 2 :279 284. Frankham, R., H. Manning, S.H. Margan, and D .A. Briscoe. 2000. Does equalization of family sizes reduce genetic adaptation to captivity? Animal Conservation 3 :357 363. Frankham, R., K. Lees, M.E. Montgomery, P.R. England, E. Lowe, and D.A. Briscoe. 1999. Do population size bottlenecks reduce evoluti onary potential? Animal Conservation 2 :255 260. Frankham, R., and D.A. Loebel. 1992. Modeling problems in conservation genetics using captive Drosophila populations: rapid genetic adaptation to captivity. Zoo Biology 11 :333 342. Fraser, D. 2008. How well c an captive breeding programs conserve biodiversity? A review of the salmonids. Evolutionary Applications 1 :535 586. Fraser, A.M., A.H. Axn, and N.E. Pierce. 2001. Assessing the quality of different ant species as partners of a myrmecophilous butterfly. Oe cologia 129 :452 460. Glaubitz J. 2004. Convert : a user friendly program to reformat diploid genotypic data for commonly used population genetic software packages. Molecular Ecology Notes 4 : 309 310. Goldstein, P.Z., and R. DeSalle. 2003. Calibrating phylo genetic species formation in a threatened insect using DNA from historical specimens. Molecular Ecology 12 :1993 1998. Google. 2008. Google Earth Mapping Service, 4.0.2737. Mountain View, CA, USA. Accessed 29 June 2008. Available from http://earth.google.com/ Goudet, J. 2001 FSTAT a program to estimate and test gene diversities and fixation indices (version 2.9.3.2). Accessed 2 February 2008. Available from h ttp://www.unil.ch/izea/softwares/fstat Gow, J.L. L.R. Noble, D. Rollinson, and C.S. Jones. 2005. A high incidence of clustered microsatellite mutations revealed by parent offspring analysis in the African freshwater snail, Bulinus forskalii (Gatropoda, P ulmonata). Genetica 124 :77 83.

PAGE 163

163 Mauritius kestrel. Nature 403 :616. Guo, S.W., and E.A. Thompson. 1992. Performing the exact test for Hardy Weinberg proportion for mu ltiple alleles. Biometrics 48 :361 372. Haig, S.M., E.A. Beever, S.M. Chambers, H.M. Draheim, B.D. Dugger, S. Dunham, E. Elliott Smith, J.B. Fontaine, D.C. Kesler, B.J. Knaus, I.F. Lopes, P. Loschl, T.D. Mullins, and L.M. Sheffield. 2006. Taxonimic consider ation in listing subspecies under the U.S. Endangered Species Act. Conservation Biology 20:1584 1594. Haikola and I. Hanski. 2001 Inbreeding depression and the genetic maintenance of a genetic load in Melitaea cinxia populations. Conservation Genetics 2 :235 335. Haikola, S., M.C. Singer, and I. Pen. 2004 Has inbreeding depression led to avoidance of sib mating in the Glanville fritillary butterfly ( Melitaea cinxia )? Evo lutionary Ecology 18 :113 120. Hajibabaei, M., M.A. Smith, D.H. Janzen, J.J. Rodriguez, J.B. Whitfield, and P.D.N. Hebert. 2006. A minimalist barcode can identify a specimen whose DNA is degraded. Molecular Ecology Notes 6 :959 964. Hanski, I., T. Pakkala, M. Kuussaari, and G. Lei. 1995. Metapopulation persistence of an endangered butterfly in a fragmented landscape. Oikos 72 :21 28. Hanski, I. 1999. Metapopulation ecology. Oxford University Press, Oxford, United Kingdom. Hanski, I., and O. Ovaskainen. 2003. Metapopulaiton theory for fragmented landscapes. Theoretical Population Biology 64 :119 127. Harper, G. L., S. Piyapattanakorn, D. Goulson, and N. Maclean. 2000. Isolation of microsatellite markers from Adonis blue butterfly ( Lysandra bellargus ). Molecular Ecology 9 :1919 1920. Harper, G. L., N. Maclean, and D. Goulson. 2003. Microsatellite markers to assess the influence of population size, isolation and demographic change on the genetic structure of the UK butterfly Polyommatus bellargus Molecular Ecology 12 :3349 3357 Harper, G. L., N. Maclean, and D. Goulson. 2006 Analysis of museum specimens suggests extreme genetic drift in the adonis blue butterfly ( Polyommatus bellargus ). Biological Journal of the Linnean Society 88 :447 452. Hartl, D. L., and A. G. Clark. 1997. Principles of population g enetics. Sinauer Associates, Inc., Sunderland, M assachusetts.

PAGE 164

164 Hartl, G.B., and P. Hell. 1994. Maintenance of high levels of allelic variation in spite of a severe bottleneck in population size: the brown bear ( Ursus a rctos ) in the Western Carpathians. Biodiversity and Conservation 3 :546 554. Hedrick. P.W. 2005. Genetics of populations, 3 rd edition. Jones and Bartlett Publishers, Sudbury, M assachusetts. Hedrick, P.W. and S.T. Kalinowski. 2000. Inbreeding depressio n in conservation biology. Annual Review of Ecological Syst ems 31 :139 162. Henning, S.F. 1983 Chemical communication between lycaenid larvae (Lepidoptera: Lycaenidae) and ants (Hymenoptera: Formicidae). Journal of the Entomological Society of South Africa 46 : 341 366. Hoole, J.C., D.A. Joyce, and A.S. Pullin. 1999. Estimates of gene flow between populations of the swallowtail butterfly, Papilio machaon in Broadland, UK and implications for conservation. Biological Conservation 89 :293 299. Howe, W.H. 1975. The b utterflies of North America. Doubleday and Company, Garden City, New York. Hutchinson, W.F., C. van Oosterhout, S.I. Rogers, and G.R. Carvalho. 2003. Temporal analysis of archived samples indicates marked genetic changes in declining North Sea cod ( Gadus m orhua ). Proceedings of the R oyal Society of London B. 270 :2125 2132 Ingvarsson, P.K. 2001. Restoration of genetic variation lost the genetic rescue hypothesis. Trends in Ecology and Evolution 16 :62 63. Acce ssed: 5 January 2009. Available from http://cmsdata.iucn.org/downloads/state_of_the_world_s_species_factsheet_en.pdf Iudica, C.A., W.M. Whitten, and N.H. Williams. 2001. Small bones from dried mammal museum specimens as a reliable source of DNA. BioTechni ques 30 :732 736. Johnson, K., and D. Matusik. 1988. Five new species and one new subspecies of butterflies from the Sierra de Baoruco of Hispaniola. Annals of the Carnegie Museum 10 :221 254. Johnson, K. and D. Matusik. 1992. Genera and species of the neot like hairstreak butterflies (Lepidoptera, Lycaenidae, Theclinae). Taxonomic a dditions to recent studies of the neotropical butterflies. Reports of the Museum of Natural History, University of 23 :3 5. Johnson, K., a nd Z. Blint. 1995. Distinction of Pseudochrysops, Cyclargus, Echinargus, and Hemiargus in the neotropical Polyommatini (Lepidoptera, Lycaenidae). Reports of the 54 :1 14. Keyghobadi, N., J. Roland, and C. Strobeck. 2002. Isolation of novel microsatellite loci in the Rocky Mountain Apollo butterfly Parnassius smintheus Hereditas 136 :247 250.

PAGE 165

165 Kimball, C. P. 1965. The Lepidoptera of Florida : an annotated c hecklist. Arthropods of Florida an d n eighboring and a reas. Volume 1. Division of Plant Industry, Florida Department of Agriculture, Gainesville, Florida Kimura, M. 1983. The n eutral t heory of m olecular e volution. Cambridge University Press, Cambridge United Kingdom. Kishore, R., W.R. Har dy, V.J. Anderson, N.A. Sanchez, and M.R. Buoncristiani. 2006. Optimization of DNA extraction from low yield and degraded samples using the BioRobot EZ1 and BioRobot M48. Journal of Forensic Science 51 :1055 1061. Klots, A. B., 1964. A field guide to the b utter flies of North America, east of the Great Plains Houghton Mifflin, Boston Massachusetts Kronforst, M.R., and T.H. Fleming. 2001. Lack of genetic differentiation among widely spaced subpopulations of a butterfly with a home range behavior. Heredit y 86 :243 250. Lacy, R.C. 1993. Impacts of inbreeding in natural and captive populations of vertebrates: implications for conservation. Perspectives in Biology and Medicine 36 :480 496. Lande, R. 1995. Mutation and conservation. Conservation Biology 9 :782 79 1. Leonard, J.A. 2008. Ancient DNA applications for wildlife conservation. Molecular Ecology 17 :4186 4196. Lenczewski, B. 1980. Butterflies of Everglades National Park. Report T 588 National Park Service, South Florida Research Center, Homestead Florida. Lewis, P. O., and Zaykin, D. 2001. Genetic Data Analysis: Computer program for the analysis of allelic data. Version 1.0 (d16c). Accessed 2 February 2008. Available from h ttp://lewis.eeb.uconn.edu/lewishome/software.html Lowe, A.J., D. Boshier, M. Ward C.F.E. Bacles, and C. Navarro. 2005. Genetic resource impacts of habitat loss and degradation; reconciling empirical evidence and predicted theory for neotropical trees. Heredity 95 :255 273. Luikart G. and J. M. Cornuet. 1997. Empirical evaluation of a test for identifying recently bottlenecked populations from allele frequency data. Conservation Biology 12 :228 237. Mandrioli, M., F. Borsatti, and L. Mola. 2006. Factors affecting DNA preservation from museum collected lepidopteran specimens. Entomologi a Experimentalis et Applicata 120 :239 244. Margan, S.H., R.K. Nurthen, M.E. Montgomery, L.M. Woodworth, E.H. Lowe, D.A. Briscoe, and R. Frankham. 1998. Single large or several small? Population fragmentation in the captive management of endangered species. Zool Biology 17 :467 480.

PAGE 166

166 Matteson, J.H. 1930. Anaea portia the leaf wing and a list of Rhopalocera of Miami, Florida. Privately printed, 16pp. Mattoni, R., T. Longcore, Z. Krenova, and A. Lipman. 2003. Mass rearing of the endangered Palos Verdes blue b utterfly ( Glaucopsyche lygdamus palosverdesensis : Lycaenidae). Journal of Research on the Lepidoptera 37 : 55 67. McKe nzie, J.A., and P. Batterham. 1994. The genetic, molecular and phenotypic consequences of selection for insecticide resistance. Trends in E cology and Evolution 9 :166 169. Meier, R.E. 1994. Coexisting patterns and foraging behavior of introduced and native ants (Hymenoptera: Formicidae) in the Galapagos Islands (Ecuador). Pages 44 62 in D.F. Williams D.F., editor. Exotic ants: biology, impact, and control of introduced species. Westview Press, Boulder, Colorado. Miller, L. D., M. J. Simon, and D. J. Harvey 1992 The butterflies (Insecta: Lepidoptera) of Crooked, Acklins and Mayaguana Islands, Bahamas, with a discussion of the biogeographical a ffinities of the southwestern Bahamas and a description of a new subspecies by H.K. Clench. Annals of the Carnegie Museum 61 :1 31. Minno, M. C. and T. C. Emmel. 1993. Butterflies of the Florida Keys. Scientific Publishers, Gainesville, Florida Minno, M.C. J.F. Butler, and D.W. Hall. 2005 Florida butterfly caterpillars and their host plants. University Press of Florida, Gainesville, Florida. Morin, P.A., and M. McCarthy. 2007. Highly accurate SNP genotyping from historical and low quality samples. Molecul ar Ecology Notes 7 :937 946. Mschler H. B. 1886 Beitrge zur Schmetterlings Fauna von Jamaica. Abhandlungen von der Senckenbergischen naturforschenden Gesellschaft, Frankfurt 14 :25 88. Nabokov, V. 1945. Notes on neotropical Plebejinae (Lycaneidae: Lepid optera). Psyche 52 :1 61. Nabokov, V. 1948. A new species of Cyclargus Nabokov (Lycaenidae, Lepidoptera). Entomology 81 :273 280. Nei M. 1972 Genetic distance between populations. American Naturalist 106 :283 292. Nei M. 1978 Estimation of average heteroz ygosity and genetic distance from a small number of individuals. Genetics 89 :583 590. Nei, M., T. Maruyama, and R. Chakraborty. 1975. The bottleneck effect and genetic variability in populations. Evolution 29 :1 10. Nei, M. 1987. Molecular evolutionary gene tics, Columbia University Press, New York, New York.

PAGE 167

167 New, T.R. 1993. Conservation Biology of the Lycaenidae (Butterflies). IUCN Publ ished Report 8, Gland, Switzerland. Newman, D., and D. Pilson. 1997. Increased probability of extinction due to decreased genetic effective population size: experimental populations of Clarkia pulchella Evolution 51 :354 362. Nice, C.C., G. Gelembiuk, N. Anthony, and R. Ffrench Constant. 2000. Population genetics and phylogeography of the butterfly genus Lycaeides USFWS, G reen Bay Office, Wisconsin, i 43. Accessed 11 February 2009. Available from http://www.fws.gov/Midwest/endangered/insects/kbb/kbb final rp2.pdf NOAA. 2008. Florida Keys C limatology, National Weather Service Forecast Office, Key West, Florida. Accessed 5 December 2008. Available from http://www.srh.noaa.gov/key/HTML/climate/eywclimate.html Opler, P .A., and V. Malikul. 1992 A field guide to the eastern bu tterflies. Houghton and Mifflin Company, New York, New York. Paetkau, D., L.P. Waits, P.L. Clarkson, L. Craighead, and C. Strobeck. 1997. An empirical evaluation of genetic distance statistics using microsatellite data from bear (Ursidae) populations. Genetics 147 :1943 1957. Palo, J., S. L. Varvio, I. Hanski, and R. Vainola. 1995. Developing microsatellite markers for insect population structure: complex variation in a checkerspot butterfly. Heredit as 123 :295 300. Passera, L. 1994. Characteristics of tramp ant species. Pages 23 43 in D.F. Williams editor. Exotic ants: biology, impact, and control of introduced species. Westview Press, Boulder, Colorado. Peakall, R., and P. Smouse. 2006. Genalex 6 : g enetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes 6 : 288 295. Pearce Kelly, P., R. Jones, D. Clarke, C. Walker, P. Atkin, and A.A. Cunningham. 1998. The captive rearing of threatened Orthoptera: a comp arison of the conservation potential and London. Journal of Insect Conservation 2 :201 210. Peel, D., J.R. Ovenden, and S.L. Peel. 2004. NeEstimator: software for esti mating effective population size, Version 1.3 Queensland Government, Department of Primary Industries and Fisheries, Brisbane, Australia. Accessed 8 July 2008. Available from http://www.dpi.qld.gov .au/fishweb/11637.html Pierce, N.E. and S. Easteal. 1986. The selective advantage of attendant ants for the larvae of a lycaenid butterfly, Glaucopsyche lygdamus Journal of Animal Ecology 55 :451 462.

PAGE 168

168 Pierce, N.E. and P.S. Mead. 1981. Parasitoids as se lective agents in the symbiosis between lycaenid butterfly larvae and ants. Science 211 :1185 1187. Pierce, N.E., M.F. Braby, A. Heath, D.J. Lohman, J. Mathew, D.B. Rand, and M.A. Travassos. 2002. The ecology and evolution of ant association in the Lycaenid ae (Lepidoptera). Annu al Review of Entomology 47 :733 771. Pompanon, F., A. Bonin, E. Bellemain, and P. Taberlet. 2005. Genotyping errors: causes, consequences and solution. Nature Reviews Genetics 6 :847 859. Pritchard, J.K., M. Stephens, and P. Donnelly. 2 000. Inference of population structure using multilocus genotype data. Genetics 155 :945 949. Pritchard, J.K., X. Wen, and D. Falush. 2007. Documentation for structure software: version 2.2. Accessed 12 March 2008. Available from http://pritch.bsd.uchicago. edu/software. Raquin, A. L., F. Depaulis, A. Lambert, N. Galic, P. Brabant, and I. Goldringer. 2008. Experimental estimation of mutation rates in a wheat population with a gene genealogy approach. Genetics 179 :2195 2211. Reed, D.H., and R. Frankham. 2003. Correlation between fitness and genetic diversity. Conservation Biology 17 :230 237. Rice, W.R. 1988. Analyzing tables of st atistical tests. Evolution 43 :223 225 Riley, N.D. 1975. A field guide to the butterflies of the West Indies. Collins, London, Englan d. Ross, J.D., A.D. Arndt, R.F.C. Smith, J.A. Johnson, and J.L. Bouzat. 2006. Re examination of the historical range of the greater prairie chicken using provenance data and DNA analysis of museum collections. Conservation Genetics 7 :735 750. Ruffin J., a nd J. Glassberg 2000 Miami blues still fly. American Butterflies 8 :28 29. Saarinen, E.V., J.C. Daniels, and J.E. Maruniak. 2009. Development and characterization of polymorphic microsatellite loci in the endangered Miami blue butterfly ( Cyclargus thomasi bethunebakeri ). Molecular Ecology Resources 9 :242 244. Saarinen, E.V. 2006. Life history and myrmecophily of Neomyrina nivea periculosa 59 :112 115. Saarinen, E.V., and J. Daniels. 2006. Miam i blue butterfly larvae (Lepidoptera: Lycaenidae) and ants (Hymepnoptera: Formicidae): new information on the symbionts of an endangered taxon. Florida Entomologist 89 :69 74. Saccheri, I.J., P.M. Brakefield, and R.A. Nichols. 1996. Severe inbreeding depres sion and rapid fitness rebound in the butterfly Bicyclus anynana (Satyridae). Evolution 50 :2000 2013.

PAGE 169

169 Saccheri, I., M. Kuussaari, M. Kankare, P. Vikman, W. Fortelius, and I. Hanski. 1998. Inbreeding and extinction in a butterfly metapopulation. Nature 392 : 491 494. Saccheri, I. J., I. J. Wilson, R. A. Nichols, M. W. Bruford, and P. M. Brakefield. 1999. Inbreeding of bottlenecked butterfly populations: estimating using the likelihood of changes in marker allele frequencies. Genetics 151 :1053 1063. Sacch eri, I.J., R.A. Nichols, and P. M. Brakefield. 2001. Effects of bottlenecks on quantitative genetic variation in the butterfly Bicyclus anynana Genetics Research Cambridge 77 :167 181. Sagra, R. de la. 1857. Historia Fsica Poltica y Natural de la Isla de Cuba. Segunda Parte. Historia Natural. 7. Crustaceos, Aragnides Insectos, 202 280. Paris [dated 1856 but not published or completed until 1857]. Sarhan, A. 2006. Isolation and characterization of five microsatellite loci in the Glanville fritillary butterfly ( Melitaea cinxia ). Molecular Ecology Notes 6 :163 164. Savignano, D.A. 1994. Benefits to Karner Blue butterfly larvae from association with ants. Pages 37 46 in D.A. Andrew, R.J. Baker, and C.P. Lane, editors. Karner blue butterfly: a symbol of a vanishing landscape. Minnesota Agricultural Experiment Station, St. Paul, Minnesota. Schug, M.D., C.M. Hutter, M.A. Noor, and C.F. Aquadro. 1998. Mutation and evolution of microsatellites in Drosophila melanogaster Genetica 102 103 :359 367. Shepherd, L.D., and D.M Lambert. 2008. Ancient DNA and conservation: lessons learned from the endangered kiwi of New Zealand. Molecular Ecology 17 :2174 2184. Slatkin, M. 1995. A measure of population subdivision based on microsatellite allele frequencies. Genetics 139 :457 462. Smith, D. S., L. D. Miller, and J. Y. Miller. 1994. The b utt erflies of the West Indies and s outh Florida. Oxford University Press, Oxford, United Kingdom. Snyder, N.F.R., S.R. Derrickson, S.R. Beissinger, J.W. Wiley, T.B. Smith, W.D. Toone, and B. Miller. 1996. Limitations of captive breeding in endangered species recovery. Conservation Biology 10 :338 348. Solecki, W.D. 2001. South Florida The role of global to local linkages in land use/land cover change in South Florida Ecological Economics 37 :339 356. Solecki, W.D., and Walker, R.T. 2001. Creating the South Florida landscape. Pages 237 273 in Gordon Wolman, editor. Growing populations, changing landscapes, s tudies from India, China, and the United States Nat ional Academy Press, Washington, D.C. Soul, M.E., M. Gilpin, W. Conway, and T. Foose. 1986. The millennium ark: how long a voyage, how many staterooms, how many passengers? Zoo Biology 5 :101 113.

PAGE 170

170 Spielman, D., B. W. Brook, and R. Frankham. 2004. Most species are not driven to extinction before gene tic factors impact them. Proceedings of the National Academy of Sciences 101 :15261 15264. Steiner, F.M., M. Sielezniew, B.C. Schlick Steiner, H. Httinger, A. Stankiewicz, and A. Grnicki. 20 03. Host specificity revisted: n ew data on Myrmica host ants of the lycaenid butterfly Maculinea rebeli Journal of Insect Conservation 7 :1 6. Sunnucks, P. P.J. De Barro, G. Lushai, N. Maclean, and D. Hales. 1997. Genetic structure of an aphid studied using microsatellites: cyclic parthenogenesis, differentiated linea ges and host specialization. Molecular Ecology 6 :1059 1073. Svensson, E.M., C. Anderung, J. Baubliene, P. Persson, H. Malmstrom, C. Smith, M. Vretemark, L. Daugnora, and A. Gotherstrom. 2007. Tracing genetic change over time using nuclear SNPs in ancient a nd modern cattle. Animal Genetics 38 :378 383. Taberlet, P., L.P. Waits, and G. Luikart. 1999. Noninvasive sampling: look before you leap. Trends in Ecology and Evolution 14 :323 327. Tallmon, D.A., G. Luikart, and R.S. Waples. 2004. The alluring simplicity and complex reality of genetic rescue. Trends in Ecology and Evolution 19 :489 496. Tallmon, D.A., A. Koyuk, G. Luikart, and M.A. Beaumont. 2008. ONeSAMP: a program to estimate effective population size using approximate Bayesian computation. Molecular Ecol ogy Resources 8 : 299 301. Templeton, A.R. 1981. Mechanisms of speciation a population genetics approach. Annual Review of Ecological Systematics 12:23 48. importance of multilocus genetic diversity in conservation biology: response to Kalinowski et al. Conservation Biology 16 :1151 1155. Theodorou, K., and D. Couvet. 2004. Introduction of captive breeders to the wild: harmful or beneficial? Conservation Genetics 5 :1 12. T homas, J.A. 1980. Why did the large blue become extinct in Britain? Oryx 15 :243 247. Thrall, P.H., C.M. Richards, D.E. McCauley, mand J. Antonovics. 1998. Metapopulation collapse: the consequences of limited gene flow in spatially structured populations. P ages 83 104 in J. Bascompte and R.V. Sol, editors. Modeling spatiotemporal dynamics in ecology. Springer Verlag, New York, N ew Y ork v an Oosterhout, C., G. Smit, M.K. van Heuven, and P.M. Brakefield. 2000. Pedigree analysis on small laboratory population s of the butterfly Bicyclus anynana : t he effects of selection on inbreeding and fitness. Conservation Genetics 1 :321 328.

PAGE 171

171 v an Oosterhout C., W. F. Hutchinson, D. P. M. Wills, P. Shipley. 2004. Micro Checker: software for identifying and correcting genotypi ng errors in microsatellite data. Molecular Ecology Notes 4 : 535 538. Vernesi, C., M.W. Bruford, G. Bertorelle, E. Pecchioli, A. Rizzoli, and H.C. Hauffe. 2008. 804. Wagner, D. and C. Martinez del Rio. 1997. Experimental tests of the mechanism for ant enhanced growth in an ant tended lycaenid butterfly. Oecologia 112 :424 429. Wang, J., and M.C. Whitlock. 2003. Estimating effective population size and migration rates from genetic samples over space and time. Genetics 163 :429 446. Waples, R.S., P.B. Adams, J. Bohnsack, and B.L. Taylor. 2007. A biological framework for evaluating whether a species is threatened or endangered in a significant portion of its range. Conservation Biolog y 21:964 974. Waples, R.S. 1989. A generalized approach for estimating effective population size from temporal changes in allele frequency. Genetics 121 :379 391. Watts, P.C., D.J. Thompson, K.A. Allen, and S.J. Kemp. 2007. How useful is DNA extracted from the legs of archived insects for microsatellite based population genetic analyses? Journal of Insect Conservation 11 :195 198. Webster, R.P., and M.C. Nielson. 198 Satyrium edwardsii 38 :124 133. Weir, B.S., and C.C. Cockerham. 1984. Estimating F statistics for the analysis of population structure. Evolution 38 :13 58 1370. Weissenbach, J., G. Gyapay, C. Dib, A. Vignal, J. Morrissette, P. Millasseau, G. Vaysseix, and M. Lathrop. 1992. A second generation linkage map of the human genome. Nature 359 :794 801. Whittaker, J.C., R.M. Harbord, N. Boxall, I. Mackay, G. Daws on, and R.M. Sibly. 2003. Likelihood based estimation of microsatellite mutation rates. Genetics 164 :781 787. Winston, J.E. 2007. Archives of a small planet: the significance of museum collections and museum based research in invertebrate taxonomy. Zootaxa 1668 :47 54. Woodworth, L.M., M.E. Montgomery, D.A. Briscoe, and R. Frankham. 2002. Rapid genetic deterioration in captivity: causes and conservation implications. Conservation Genetics 3 :277 288. Wojcik, D.P. 1994. Impact of the red imported fire ant on n ative ant species in Florida. Pages 269 281 in D.F. Williams, editor. Exotic ants: biology, impact, and control of introduced species. Westview Press, Boulder, Colorado.

PAGE 172

172 Wright, S. 1951. The genetical structure of populations Annals of Eugenics 15 : 323 3 54 Wright, S. 1965. The interpretation of population structure by F statistics with special regard to systems of mating. Evolution 19 :395 420. Wright, S. 1977. Evolution and the genetics of populations. 3. Experimental results and evolutionary deductions University of Chicago Press, Chicago, Illinois. Wright, S. 1978. Evolution and the genetics of p opulations. 4. Variability w ithin and a mong n atural p opulations. University of Chicago Press, Chicago, Illinois. Wunderlin, R.P. 1998. Guide to the vascular p lants of Florida. University of Florida Press, Gainesville, Florida. Zhang, D. X. 2004. Lepidopteran microsatellite DNA: redundant but promising. Trends in Ecology and Evolution 19 :507 509.

PAGE 173

173 BIOGRAPHICAL SKETCH Emily Virginia Saarinen was born in upstate New York She completed school in the Fayetteville Manlius public school district and attended Syracuse University as a linguistics and 1997. As her interests in the biological sciences began to grow, University of South Florida. It was in Sarasota, Florida at New College that her interests in entomology and the natural sciences really blossomed. Under the expert tutelage of Drs. Lee and Jackie Miller and Sandra Gilchrist, Emily embarked on a study of the coevoution between a local moth species and its host plant. She graduated from New College in 1999 with degree in b iology and a desire to explore the natural world. Emily spent the next 12 months travelling around the world, visiting Denmark, Oman, Sri Lanka, Indonesia, Thailand, and Malaysia. She returned to the U.S. with a desire to pursue a degree in e ntomology after her experiences overseas. After working in industry for a year, she began her M.S. degree at the University of Florida in the Department of Entomology and Nematology. She was awarded a Fulbright Fellowship from the U.S. State Department and conducted field and laboratory research at the Forest Research Institut e of Malaysia in 2002 2003. Her research on the symbiotic relationship between highly specialized weaver ants and lycaenid butterflies in the rainforest of Malaysia solidified her desire to obtain a doctorate in entomology. She finished her M.S. degree i n 2004 and quickly began her Ph.D. research on the Miami blue butterfly, focusing on the interface between ecology and genetics. She received her Ph.D. from the University of Florida in the spring of 2009. Emily continues to explore the natural world and is joined by her husband and daughter are to continue teaching and conducting research around the world and to apply her knowledge to conservation projects.