Citation
Conservation Biology of Mitoura gryneus sweadneri (Lepidopertera: Lycaenidae)

Material Information

Title:
Conservation Biology of Mitoura gryneus sweadneri (Lepidopertera: Lycaenidae)
Creator:
PENCE, JAMES AKERS
Copyright Date:
2008

Subjects

Subjects / Keywords:
Butterflies ( jstor )
Cedars ( jstor )
Eggs ( jstor )
Female animals ( jstor )
Groves ( jstor )
Instars ( jstor )
Larvae ( jstor )
Pupae ( jstor )
Species ( jstor )
Trees ( jstor )
City of St. Augustine ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright James Akers Pence. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
12/31/2015
Resource Identifier:
658205211 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

CONSERVATION BIOLOGY OF Mitoura gryneus sweadneri (LEPIDOPTERA: LYCAENIDAE) By JAMES AKERS PENCE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORID A IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by James Akers Pence

PAGE 3

To Linda Pence: my best friend, my mirror , my wife, and my mo st dedicated field assistant

PAGE 4

iv ACKNOWLEDGMENTS First, I thank my wife Linda Pence for her support and encouragement in this work. I thank John V. Calhoun and Marc C. Minno for sharing specimens and personal observations about the butterfly. I thank Dr. Kurt Johnson for his work on the Mitoura that provided background and a studied perspective on the butterfly. I especially appreciate J. D. Tu rner of Huntsville, Alabama for his help in collecting Mitoura g. gryneus for my captive colonies and for cutting and sending fresh cedar to feed them all. I thank Dr. Jacqueline Y. Miller for confirming my microscopic discoveries . I thank Dr. Thomas C. Emmel for chairing my committee and Dr. Frank Sl ansky, Dr. James E. Maruniak, Dr. James E. Lloyd, and Dr. Jonathan Reiskind for evaluating this work. I thank Frank Cone, Steven P. Gilly, and Don West for their oral history of the Florida Division of Forestry. I thank Dr. T. J. Walker for the loan of temperature data loggers. I appreciate Jerry Wenzel and Nick Hostettler for help with field equipment. I owe a debt of gratitude to Deborah A. Hall, who helped me through the bureaucracy of graduate school; and to my friends Dennis and Rebekah Mahoney, who provided invaluable assist ance and computer expertise without which the project would not have become a reality.

PAGE 5

v TABLE OF CONTENTS Page ACKNOWLEDGME NTS.......................................................................................iv LIST OF T ABLES.................................................................................................ix LIST OF FI GURES...............................................................................................xi ABSTRACT.........................................................................................................xv CHAPTER 1 INTRODUCTION TO THE SP ECIES AND RELA TED TAXA.........................1 Higher Taxonomy of the Lyc aenidae ..............................................................5 Genus Mitoura ..............................................................................................10 Genus Callophrys .........................................................................................11 Original Description and Taxonomic History of Mitoura sweadneri ..............13 Mitoura sweadneri .................................................................................15 Mitoura gryneus sweadneri ....................................................................16 Mitoura grynea sweadneri .....................................................................16 Mitoura gryneus .....................................................................................16 Callophrys (Mitoura) gryneus smilacis ...................................................16 Callophrys (Mitoura) gryneus sweadneri ...............................................17 Callophrys gryneus sweadneri ...............................................................17 Other Possible Names...........................................................................17 Related Taxa and Thei r Host Pl ants............................................................19 2 LIFE HIST ORY.............................................................................................23 Methods and Materials of Life Histo ry Study: Immatu re Sta ges...................26 Eggs............................................................................................................. 31 Size........................................................................................................ 32 General Appear ance..............................................................................32 Ultrastruct ure.........................................................................................36 Period in Ovum......................................................................................38 First Instar....................................................................................................39 General Appear ance..............................................................................40 Head Capsul e........................................................................................43 Duration .................................................................................................44

PAGE 6

vi Second Inst ar...............................................................................................44 Second Head C apsule...........................................................................46 Duration .................................................................................................47 Molting No tes.........................................................................................48 Third Inst ar...................................................................................................50 Third Head C apsule ...............................................................................51 Duration .................................................................................................53 Fourth In star.................................................................................................53 Fourth Head C apsule.............................................................................56 Duration .................................................................................................57 Last Instar....................................................................................................57 The Feede r............................................................................................60 The Wander er........................................................................................62 Crawl off test...................................................................................63 Burrowing te st.................................................................................64 The Prepu pa..........................................................................................66 Final Head C apsule...............................................................................67 Duration .................................................................................................69 Pupa............................................................................................................. 70 Development and Duration ....................................................................72 Eclosion stra tegy.............................................................................73 Burrowing test re visited ...................................................................75 Audio Observ ations ...............................................................................76 Desidera ta.............................................................................................79 Adults In Vitro ...............................................................................................79 Sex Ratio...............................................................................................81 Size........................................................................................................ 81 Diet and Longev ity.................................................................................82 Fecundity ...............................................................................................86 Captive Copul ation................................................................................88 Larval Ho st...................................................................................................93 Specific Ta xonomy................................................................................94 Larval Host Spec ificity ...........................................................................98 Anthropocentric Ut ility of C edar........................................................... 100 Medicinal uses..............................................................................101 Other us es.....................................................................................101 Florida Cedar Today ............................................................................103 Life History Field Study: Adults in Situ .......................................................108 The Litera ture......................................................................................111 Methods and Materials of Life History Fiel d Study ...............................112 Data colle cted...............................................................................112 Effort..............................................................................................116 Results of Fi eld Work ...........................................................................120 Analysis of Fi eld Study ........................................................................126 Tree dimens ions............................................................................ 126 Tree sex ........................................................................................129

PAGE 7

vii Phenology .....................................................................................133 Adult behavi or...............................................................................138 Habitat Descr iption ..............................................................................146 Natural cedar groves .....................................................................147 Duff and the cedar mi croclimat e....................................................150 Importance of duff .........................................................................163 Detriments to habitat .....................................................................165 3 PAST AND PRESENT DISTRIBUT ION.....................................................170 Historic Ra nge............................................................................................170 Current Distri bution ....................................................................................175 Numbers ....................................................................................................178 Records from Other Researcher s........................................................178 My Field Re cords .................................................................................182 Comparison and Discussion ................................................................186 4 SPECIFIC PHENOTYPIC DISTINCTIVENESS, Mitoura gryneus sweadneri ...................................................................................................190 Historic Illust rations ....................................................................................192 Alar Characters: Mitoura gryneus ...............................................................194 Scudder, 1889a ...................................................................................196 Holland, 19 40......................................................................................199 Klots, 1951........................................................................................... 199 Harris, 1972......................................................................................... 199 Howe, 197 5.........................................................................................199 Johnson, 1980..................................................................................... 200 Opler and Kriz ek, 1984 ........................................................................200 Scott, 1986 ..........................................................................................200 Allen, 1997........................................................................................... 200 Alar Characters: Mitoura gryneus sweadneri .............................................204 Chermock, 1944.................................................................................. 206 Scudder, 1889 .....................................................................................206 Klots, 1951........................................................................................... 206 Howe, 197 5.........................................................................................207 Johnson, 1980..................................................................................... 207 Scott, 1986 ..........................................................................................207 Specificity of Alar Character States: M. g. sweadneri .................................219 Alar Characters: Poss ible Interg rades........................................................ 223 Genitalic Char acters...................................................................................227 Methods and Mate rials ........................................................................230 Results of Geni talic St udy.................................................................... 231 Significance of Genitalic Character States ...........................................238 Biological Diffe rences .................................................................................240 Larval Host s.........................................................................................241 Seasonality of Flights ...........................................................................242

PAGE 8

viii 5 HYBRIDIZATION EXPERIMENTS .............................................................244 Methods and Mate rials ...............................................................................246 Results of Hybridiz ing Experim ents............................................................248 Courtshi p.............................................................................................248 Flight Cage Experiments: M. g. gryneus with M. g. sweadneri ............249 Session observa tions.................................................................... 250 Hybrid pr ogeny.............................................................................. 252 Flight Cage Experimen ts: Hybrid s.......................................................255 Progeny............................................................................................... 258 Discussion and Su mmation ........................................................................262 6 ATTRIBUTES OF SUCCESSFUL COLONY H ABITATS...........................265 Dixie Co.: WP T 002.................................................................................... 268 Numbers ..............................................................................................273 Dispositio n...........................................................................................274 St. Augustine Lightho use: WPT 001.......................................................... 275 Numbers ..............................................................................................278 History and Dispos ition........................................................................279 Hollister, Putnam Co.: WPT 019................................................................281 Numbers ..............................................................................................283 Dispositio n...........................................................................................283 Specific Threats to Habitat and Conservation Concerns ............................287 Anthropocentric Threats ......................................................................288 Natural Ru in.........................................................................................290 Implications for Conserva tion of Biodive rsity.......................................291 7 DISCUSSION AND CONCLUSIO NS......................................................... 293 Taxonomic Distinc tiveness .........................................................................293 Life Histo ry...........................................................................................294 Phenotype Distinc tiveness ...................................................................295 Hybridizat ion........................................................................................296 Habitat Require ments................................................................................297 Conservation C oncerns ..............................................................................298 LIST OF REFE RENCES..................................................................................300 BIOGRAPHICAL SKETCH ...............................................................................313

PAGE 9

ix LIST OF TABLES Table page 1-1. List of North American Mitoura and their food plants...................................20 2-1. Dates and sites of gravid fema les taken for capt ive rearin g........................26 2-2. Period in ovum............................................................................................38 2-3. First instar duratio n......................................................................................44 2-4. Second instar duratio n................................................................................47 2-5. Third instar duration....................................................................................53 2-6. Fourth inst ar duratio n..................................................................................57 2-7. Observed depths of burrowing wanderers in vitro .......................................65 2-8. Head capsule (hc) si ze classes in mm........................................................69 2-9. Final instar duration.....................................................................................69 2-10. Total development time to p upation. .........................................................72 2-11. Times to eclosion. .....................................................................................73 2-12. Forewing lengths.......................................................................................82 2-13. Total Red Cedar seedlings sold..............................................................106 2-14. Cedar sites surveyed for Mitoura sweadneri ...........................................121 2-15. Frequency of observations delineated by tree sex ..................................132 2-16. Dates of butterflies o ccurrence multiple years .........................................135 2-17. First of seas on hairstreak s......................................................................139 3-1. Specimen l abel data I................................................................................180 3-2. Specimen l abel data II...............................................................................181

PAGE 10

x 3-3. My field study sight reco rds.......................................................................183 4-1. Tail lengths by subspecie s........................................................................222 4-2. Tail lengths of po ssible interg rades........................................................... 226 5-1. Flight cage experiments: M. g. gryneus with M. g. sweadneri ...................250 5-2. Hybrid progeny ..........................................................................................253 5-3. Flight cage experim ents: F1 Hy brids. ........................................................255 5-4. F2 Hybrids and backcross pr ogeny........................................................... 258 5-5. Flight cage experiments: F2 Hybrids and HYB/SW D backcros s...............260 6-1. Mitoura g. sweadneri observed abunda nce...............................................266

PAGE 11

xi LIST OF FIGURES Figure page 1-1. Distribution of ten Mitoura species and their larv al host pl ants....................21 2-1. Camera used to photograph immatu res......................................................31 2-2. Mitoura eggs from Flori da and Tenness ee..................................................33 2-3. Mitoura gryneus sweadneri egg..................................................................35 2-4. Annuli of Mitoura g. gryneus and M. g. sweadneri eggs..............................37 2-5. Feeding posture first instar Mitoura g. sweadneri ........................................40 2-6. First instar Mitoura g. sweadneri .................................................................41 2-7. First instar Mitoura g. gryneus .....................................................................42 2-8. First head c apsules. ....................................................................................43 2-9. Second instar M. g. sweadneri ....................................................................45 2-10. Second instar M. g. gryneus ......................................................................46 2-11. Second head capsules ..............................................................................47 2-12. Molting sequence ......................................................................................49 2-13. Third instar M. g. sweadneri ......................................................................51 2-14. Third instar M. g. gryneus ..........................................................................52 2-15. Third head c apsules. .................................................................................53 2-16. Fourth instar M. g. sweadneri ....................................................................54 2-17. Fourth instar M. g. gryneus ........................................................................55 2-18. Fourth head capsules ................................................................................56 2-19. Fifth instar M. g. sweadneri .......................................................................58

PAGE 12

xii 2-20. Fifth instar M. g. gryneus ...........................................................................59 2-21. Final M. g. sweadneri instar feeding posture.............................................61 2-22. The pink wanderer .....................................................................................62 2-23. Mitoura g. sweadneri prepupa ...................................................................66 2-24. Pupae of both butterflies............................................................................71 2-25. Eclosion strategy, Mitoura g. gryneus .......................................................75 2-26. Eclosion cups a rrayed on the hearth.........................................................78 2-27. Imagoes of bot h butterflie s........................................................................80 2-28. Ranges of the two larval hosts..................................................................96 2-29. Field dat a sheet .......................................................................................117 2-30. Fiel d handout ...........................................................................................119 2-31. Heights of marked tr ees.......................................................................... 127 2-32. Lowest widths of marked tr ees................................................................127 2-34. Juniperus v. silicicola flowers..................................................................130 2-35. Juniperus v . silicicola berries ...................................................................130 2-36. Marked trees delin eated by sex...............................................................131 2-37. My field records of M. g. sweadneri combined by date............................134 2-38. Days M. g. sweadneri were present from mu ltiple sour ces.....................136 2-39. Naturally occu ring edge hab itat...............................................................148 2-40. Young cedar s under oak .........................................................................149 2-41. Naturally occuri ng cedar gr ove................................................................149 2-42. Cedars measured for microclim ate..........................................................152 2-43. End of flight season mi croclimate temp eratures ......................................155 2-44. Holiday season microc limate temper atures .............................................156 2-45. Bellow freezi ng event II...........................................................................157

PAGE 13

xiii 2-46. Pre-eclosion temperatur es...................................................................... 158 2-47. Spring eclosion temperat ures..................................................................159 2-48. High summer in t he microclim ate............................................................160 3-1. Oldest range of the butterf ly...................................................................... 171 3-2. Ranges after Op ler and Kriz ek..................................................................172 3-3. Ranges afte r Scott .....................................................................................173 3-4. Refined historic ranges after Opler and Malikul .........................................174 3-5. Current known distribution of SweadnerÂ’s Ha irstreak ................................177 4-1. Papilio damon ............................................................................................192 4-2. Thecla smilacis ..........................................................................................193 4-3. Mitura damon ............................................................................................195 4-4. Mitoura g. gryneus .....................................................................................198 4-5. First flight Mitoura g. gryneus ....................................................................201 4-6. Summer brood Mitoura g. gryneus ............................................................203 4-7. Mitoura g. sweadneri First f light................................................................. 208 4-8. Hollister, Putnam Co. M. g. sweadneri ......................................................210 4-9. Mid-summer flight M. g. sweadneri ............................................................211 4-10. First flight males M. g. sweadneri ............................................................212 4-11. Second brood M. g. sweadneri ................................................................213 4-12. Mitoura g. sweadneri Gainesvi lle............................................................. 214 4-13. Late-summer va riations I.........................................................................215 4-14. Late-summer va riations II........................................................................216 4-15. Worn first flight M. g. sweadneri ..............................................................217 4-16. Mitoura g. sweadneri , Southern limit....................................................... 219 4-17. Intergrade, first fl ight, Liberty Co.............................................................224

PAGE 14

xiv 4-18. Intergrades, Mari anna, Jackson Co.........................................................225 4-19. Mitura , male abdominal appendages ......................................................228 4-20. Male Mitoura g. sweadneri ......................................................................228 4-21. Mitoura g. gryneus , male genita lia...........................................................232 4-22. Mitoura g. sweadneri , male genitalia, Putnam Co ...................................233 4-23. Mitoura g. sweadneri , male genitalia, Dixie Co ........................................234 4-24. Cornuti detail. ..........................................................................................234 4-25. Mitoura g. gryneus , female geni talia........................................................235 4-26. Mitoura g. sweadneri, female geni talia.................................................... 237 5-1. Hybr ids...................................................................................................... 254 5-2. F2 Hybrids and Backcross pr ogeny........................................................... 261 6-1. Original tree...............................................................................................269 6-2. Burned orig inal tr ee...................................................................................270 6-3. Grove after pine harve st............................................................................272 6-4. Cleared after t he pine harve st...................................................................274 6-5. St. Augustine Li ghthouse Pa rk..................................................................277 6-6. Light house.................................................................................................278 6-7. Possible habitat .........................................................................................280 6-8. Hollister, Pu tnam Co .................................................................................282 6-9. Destructi on WPT 019 ................................................................................285 6-10. New paved highway ................................................................................287 6-11. Palm Coast, Flagler Co ...........................................................................289 6-12. Other hai rstreaks .....................................................................................291

PAGE 15

xv Abstract of Dissertation Pr esented to the Graduate School of the University of Florida in Partial Fulf illment of the Requirements for t he Degree of Doctor of Philosophy CONSERVATION BIOLOGY OF Mitoura gryneus sweadneri (LEPIDOPTERA: LYCAENIDAE) By James Akers Pence December 2005 Chair: Thomas C. Emmel Major Department: Entomology and Nematology SweadnerÂ’s Hairstreak, a butterfly endemic to north central Florida, lives in Southern Red Cedar trees, Juniperus virginiana v. silicicola . The purpose of my study was to determine specific habitat requi rements essential for sustaining this species, its present conservation status, and its taxonomic relationship to the widespread nominate subspecies, Mitoura g. gryneus . The life histories of M. g. sweadneri from Florida and M. g. gryneus from Tennessee were studied in captive colonies. No physical differences were noted in any stadium of the immatu res. Differences in eclosi on strategies of the taxa may be adaptations to climatic differenc es across their distributional ranges. Mitoura g. sweadneri produces four broods per year; Mitoura g. gryneus is bivoltine, with extended partial eclosion of the second brood in staggered intervals of up to 120 days after pupation. Both butterflies overwinter as pupae in the duff at the base of their cedar host trees.

PAGE 16

xvi Minor differences in broadly variable alar characters of coloration and maculation were noted from broods of both taxa and from museum specimens. Though formerly considered distinctive, most characters showed too much variability in specimens from different locations and seasons to be reliable species-level indicators. Minor differenc es in the genitalic structures did not cause reproductive isolation in hybr idization experiments, which produced lab-viable progeny in F1 and backcross generations. A 5-year field study of active colo nies in seven Florida counties showed specific habitat requirements essential to survival of the species. Duff layer temperatures sheltered by conical c edar trees with branches intact to ground level, versus ambient temperature outsi de the trees, may signal eclosion times for overwintering pupae and maximize reproductive success. Damage by fire, flood, or degradation of cedar habitat resulting from natural succession of plant communities poses serious threats to the survival of established colonies. However, the most serious threat to t he survival of the Florida subspecies is from anthropocent ric manipulation of the landscape and wholesale destruction of cedar habitat. Ev en though the hairstreak only occurs in highly localized populations in low numbers, it is not likely to be considered for conservation listing because of its wide distribution in 27 Florida counties.

PAGE 17

1 CHAPTER 1 INTRODUCTION TO THE SPECIES AND RELATED TAXA SweadnerÂ’s Hairstreak, Mitoura gryneus sweadneri (Chermock), is a member of the largest and most diverse fa mily of butterflies, Lycaenidae, which comprises 30 to 40% of all named butte rfly species (New, 1993, Seufert and Fiedler, 1996). Robert Robbins (1982) esti mated the number of species of Lycaenidae worldwide at 6000-6900 (when Riodinidae are included) and this number may be conservative. The number of North American species is probably between 120 and 142 or so, depending on who is counting (Cushman and Murphy, 1993, Scott, 1986). A series of f our recent popular field guides to North American butterflies (Glassber g, 1993, 1999, 2001, Glassberg et al ., 2000) lists 129 lycaenid species and 25 subspecies. Several basic biological qualities of these butterf lies and their intricate adaptations to their environment make them important subjects in ecological studies, and consequently have attracted attention to this taxon since the importance of conserving biodiversity was first realized. The adults are a conspicuous, attractive f auna that has been well studied, at least in temperate zones. Forty years ago, some life hist ory information on 837 species had been collected, and at least one larval host plant for each of 98 species from the United States and Canada had been recor ded in the literature (Downey, 1962a). Lycaenids are generally rather small; t he worldÂ’s smallest butterfly may be a North American species. The Western Pygmy Blue, Brephidium exilis

PAGE 18

2 (Boisduval), has a forewing length of only 5-7 mm (Glassberg, 2001, Tilden and Smith, 1986). The next smallest is native to Afghanistan, Micropsyche ariana Mattoni, with a 7 mm wingspan (Mattoni , 1979). There are only a few big lycaenids; the largest is from eastern Asia, Liphyra brassolis Westwood, with an 8-9 cm wingspan (New, 1993). Lycaenids as a group are ra ther sedentary; however, Brephidium exilis is one of the few that is known to migrate (Scott, 1986). Some rarely stray from the immediate vicinity of their host plants. An interesting caveat is the case of lycaenids which have been dispersed on t he wind (Robbins and Small, 1981). Wind dispersal is a possible explanati on for the butterfly fauna occurring on remote islands, or colonizing devastated te rrestrial sites such as the Krakatoa island group of Indonesia since t hat volcano’s eruption in 1883 (New et al. , 1988). Despite their small size and general lack of vagility, the family is distributed throughout all major biomes fr om the tropics to t he temperate zones, and even the Arctic (Scott, 1986). They reac h their greatest diversity in the tropics, especially the rainforests of the Neotropics and southeastern Asia, with Africa supporting the third greatest divers ity. The Palearctic has about twice as many species as the Nearctic. Though t he butterflies of family Lycaenidae are well distributed, only a few species are widespread, and lycaenids show a high rate of endemism geographically and in local communities (New, 1993). Lycaenids use a wider range of larval hosts than species in any other butterfly family. Many have evolved beyond “an earlier pattern of feeding on dicotyledons” suggested by Ehrlich and Raven (1965, p. 598), and shifted to

PAGE 19

3 monocotyledons and in some cases gymnosperms such as conifers and cycads, rarely pteridophytes such as ferns or lycopsids (Singer et al ., 1971), or even orchids and mistletoe (Downey, 1962a). So me of the more unusual butterfly larval foods are lichens, green algae, and fungi (Cottrell, 1984). One North American species, Calycopis cecrops (Fabricius), the R ed-banded Hairstreak, is a detritivore and eats dead leaves and flower buds on the ground beneath its host plants (Scott, 1986). The most radica l lycaenid trophic adaptation is in the small portion of the family whos e larvae have abandoned plant foods for secretions, excretions, or regurgitations of other insects, and, in some cases, cannibalism, or preying on the larvae of ants, other Lepidopt era, or homopteran nymphs (Kitching, 1987). Lycaenidae is the only lepidopteran taxo n that exhibits regular or obligate aphytophagy. In a surv ey of 479 genera from eight lycaenid subfamilies, Cottrell (1984) found 22 genera from four subfamilies with species that were aphytophagous. This non-plant -feeding ranged from opportunistic or occasional cannibalism in some, through facultative aphytophagy in others, to two subfamilies (Liphyrinae and Miletinae) in which all genera with known host information are completely aphytophagous. Most lycaenid larvae are phytophagous an d specialize on specific plant parts, often reproductive structures: fl ower buds, flowers, and fruits (Downey, 1962a), or new leaf and stem growth. Several lycaenid taxa feed on nitrogen-rich plants and the females of some species may oviposit only on host plants grown in soil with a high nitrogen content (Pie rce, 1985), even though the same species of plants are widely distributed. In a laboratory experiment, female Jalmenus

PAGE 20

4 evagoras consistently chose their Acacia hosts grown in pots with nitrogenous fertilizer over the same plants that rece ived only water (Baylis and Pierce, 1991). This observation strongly suggests that so il index may be an impo rtant factor for determining which patches of host plant s support lycaenid larvae. Individual species are usually restricted to a narrow c hoice of food plants, often in the same genus, or a few species within the same plant family. A few of the known exceptions can be spectacularly polypha gous. For example, larvae of the Gray Hairstreak, Strymon melinus (Hübner), have been found f eeding on the flowers of 46 genera from 21 plant familie s, making it one of the most polyphagous of all butterflies (Robbins and Aiello, 1982, Sco tt, 1975). It is probably no coincidence that this butterfly is also one of the mo st widely distributed, ranging from British Columbia and Nova Scotia, across the entire U.S., through Mexico and Central America to Venezuela, South Americ a (Opler and Malikul, 1992, Tilden and Smith, 1986). A final special adaptation of the L ycaenidae is the amazing number of species found in association with ants. Downey (1962b) surveyed life history accounts of 833 lycaenid species and f ound that the larvae of 245 species exhibited myrmecophily. (Less than a dozen of these were native to the United States.) Seufert and Fiedler (1996) said more than half of all lycaenid species have some relationship with ants during thei r larval stages. Though most of these are facultative relationships, 1520% of these species are obligate myrmecophiles. In the most common fa cultative relationship, ants defend the lycaenid larva against parasitic Dipter a and/or Hymenoptera and receive a sweet

PAGE 21

5 reward exuded from special dorsal necta r glands of the larva. In the most extreme obligate relationships, the lyc aenid larvae are herded or carried into the ant’s nest and late instars eat ant brood. These ant-larv ae relationships are very old, evolutionarily speaking, and the l ycaenid larvae that participate may have evolved other special organs that emit chemical signals to ants (Fiedler et al ., 1996), or vibratory papillae that produce so unds that may call ants by substrate vibrations (DeVries and Poinar, 1997). This review of a few basic biological qual ities of lycaenids is a starting point toward understanding how many species of this taxon have very precise environmental requirements for survival. In some cases, these needs can only be met in a particular stage of plant succession. Butterflies whose abundance or very presence may depend on small incr emental changes in habitat are invaluable as environmental indicators. Mitoura gryneus sweadneri populations sometimes seem ephemeral. Kimball ( 1965, p. 48) said the subspecies “is evidently very local, and usually very rare .” Results of my field study show how closely M. g. sweadneri is adapted to its niche in nor th-central Florida and why it is an important indicator species. Higher Taxonomy of the Lycaenidae In the mid-eighteenth century, when t he Swedish botanist Carolus Linnaeus (Carl Von Linné) gave us the binomi nal system of nomenclature and the standard, minimal, descending hierarchy of class, or der, genus, and species in which to place each kind of plant and animal, he laid the foundation of modern-day taxonomy. By the time of the publication of the t enth edition of his Systemae Naturae in 1758 (since officially adopted as the beginning of zoological

PAGE 22

6 nomenclature), he had described near ly 2,900 insect species (Gordh and Headrick, 2001). Linnaeus placed all of the butterflies in the genus Papilio . This word is Classical Latin for butterfly or moth (Jaeger, 1955). Papilio is the principal genus of the family Papilionidae, the taxon that includes all swa llowtail butterflies, and also serves as the basis for the name of the superfamily Papilionoidea, the true butterflies. The Linnaean system began to be modified very soon after its inception. Two Austrian monks, Denis and Schiffe rmüller, published a list of butterflies found around Vienna in 1776. They recognized natural divisions of certain small butterflies as the “Blues”, whose dorsal wing surfaces were brilliant blue, the “Coppers” with predominately coppery-red dorsals, and the “Hairstreaks” that usually have one or more long, thin tails on their hind wings (Holland, 1940). But it was not until the Danish entomologist Johann Christian Fabricius, a student of Linnaeus from 1762 to1764, began to designat e new genera that these divisions were formalized (Gordh and Headrick, 2001). He described the genus Thecla in 1807, which in subsequent y ears became the catch-all ta xon for most hairstreaks that have tails (Howe, 1975). Thecla is New Latin for the Greek personal female name Thekla (Jaeger, 1955). Harry Zirlin, an environmental attorney, who writes about the etymology of taxonomic names, suggests (pers. com.) that Fabricius named the genus for Saint Thecla: B enedictine Abbess of Kitzingen and Ocshenfurt, illustrious for her sanctity and learning. Her date of birth is unknown, but she died ca. 790 A. D. However, the genus name may refer to a much older Saint Thecla of Iconium, t he first Christian martyr. S he personifies chastity. Her

PAGE 23

7 story and her travels with Saint Paul are recorded in a book of Apocrypha called "The Acts of Paul and Thecla.” The name survives today as subfamily Theclinae, and as the tribe Theclini (Eliot, 1973). Fabricius also described the genus Lycaena the same year. This name is from the Greek lykaon , meaning “a wolf-like animal” from lykos , “a wolf” (Jaeger, 1955, p. 146). Paul Opler speculated the genus was probably named for King Lycaon in Greek mythology who ruled t he Arcadian town, Lykosury (Opler and Krizek, 1984). My favorite interpretation is that Lycaena refers to Mt. Lycaeus, a mountain sacred to Zeus like Mt. Parnassu s, Mt. Helicon, and Mt . Olympus, all of which figure in butterfly names (Bird et al ., 1995). Today, Lycaena is the genus that includes many of the coppers and is the type-genus for subfamily Lycaeninae, and family Lycaenidae. Throughout the 19th century and the firs t part of the 20th, several more genera were added to Lycaenidae by various authors and several different schemes were tried for organizing them into subfamilies. A 1917 landmark paper by Barnes and McDunnough, revised by McDunnough (1938), established the order of most subfamilies re cognized today. The original divisions recognized by Denis and Schiffermüller were maintained, with many important revisions and additions, and nomenclatorial changes. Butterflies of the family Lycaenidae are commonly referred to as the gossamer wi ngs and include the Harvesters, Blues, Coppers, Hairstreaks, Elfins, and a few related genera. Shared adult characters include usually emarginated eyes (indent ed near the base of the antennae), and males with somewhat reduced forelegs, t hough female forelegs are more nearly

PAGE 24

8 normal size. The eggs are usually echino id, shaped like a sea urchin test or exoskeleton. The larvae are sluglike and somewhat dorsoventrally flattened, unlike most lepidopterous larvae which are more cylindrical. Lycaenid larvae also have head capsules that seem disproporti onably small, generally 50% or less of the width of the thorax. They also hav e a membranous cervix which allows the head to be extended inside plant tissue wh en early instars feed, or to be retracted into the first thor acic segment, which helps camouflage the later instars. The first instars usually have fairly l ong hair, but the succeeding instars are covered with a carpet of fine, short, dense setae. All lycaenid larvae have a peculiar fleshy lobe that interrupts the pa ttern of crochets on each proleg. They may have some, or all, of three types of myrmecophilous organs: pore cupola organs, tentacle organs, or a dorsal nectar organ (Downey, 1987, Opler and Krizek, 1984, Pierce et al ., 2002, Scott, 1986). Several authors have included the Meta lmarks, subfamily Riodiniae, among the Lycaenidae (Ehrlich and Ehrlich, 1961, New, 1993, Opler and Krizek, 1984, Scott, 1986, Robbins, 1982). However, R obbins (1988) pointed out significant differences in the male forelegs of Lycaenidae and Riodinidae, resolving them into separate families, and Harvey ( 1987) and Downey (1987) illustrated major differences in the larvae of these taxa so that today, as in the past, many authors recognize the Riodinidae as a distinc t, though closely related family (Bird et al ., 1995, Bridges, 1994, Calhoun, 1997, E liot, 1973, Glassberg, 1993, 1999, 2001, Glassberg et al ., 2000, Hodges et al ., 1983, Klots, 1951, McDunnough, 1938,

PAGE 25

9 Miller and Brown, 1981, Minno and E mmel, 1993, Opler and Malikul, 1992, Powell, 1975, Pyle, 1981, Tilden and Smith, 1986). Since Barnes and McDunnough, differ ent authors have arranged the North American Lycaenidae in slightly diffe rent subfamily, tribe, and sometimes subtribe, genus, and subgenus groupings (Ehrlich and Ehrlich, 1961, Howe, 1975, Miller and Brown, 1983, Opler and Krizek, 1984, Scott, 1986). Each of these arrangements depends upon the specif ic characters compared by the author(s) and applied to the North Amer ican fauna. Colonel John Eliot (1973) proposed a tentative arr angement of the Lycaenid ae of the world into subfamilies, tribes, and sections. He used a wide range of morphological characters, including genitalic features , secondary sexual characters such as scent brands, or androconial patches, and sca les, and erectile hair tufts, male fore tarsus features and the arrangement of spurs of the mid and hind tibiae, length and features of palpi, features of the proboscis, differences in the antennal club, as well as the number and arrangement of wing veins, and the presence or absence of tails and tornal lobes of the hind wings. He also considered characters of the larvae and pupae. EliotÂ’ s taxonomy has stood the test of time and is referenced by many researchers concerned with studies of the worldwide lycaenid fauna on such varied topics as taxonomic revision, phylogeny, and biogeography (Johnson, 1980), aphytophagy (C ottrell, 1984), ecology, evolution, and myrmecophily (Pierce et al ., 2002), and lycaenid co nservation (New, 1993). In fact, each of these cited authors begi n their papers with a phylogeny, or table of subfamilies and genera directly attribut ed to EliotÂ’s tentative arrangement of

PAGE 26

10 1973. Part of Eliot’s 134-page paper is a very carefully arranged set of keys, first to the subfamilies of Lycaenidae, and then to the tribes and sections of each of these taxa. Mitoura gryneus sweadneri keys out to subfamily Theclinae, tribe Eumaeini, Eumaeus section. Eliot lists 61 genera in this section including Callophrys , a near relative of Mitoura . The type-genus for Eumaeini is Eumaeus Hübner, 1819 and “is named after Eumaeus t he faithful swineherd of Odysseus” (Opler and Krizek, 1984, p. 87). It is interest ing to note that up to this point, all the higher taxonomic names for this butterfly, except Papilio , seem to have been chosen arbitrarily from classic mythology, or in the case of Thecla from uncanonical Christian legend, and provi de no intrinsic information about the organism. Genus Mitoura Genus Mitoura was designated by Samuel H. Scudder (1869). Its type species is Thecla smilacis Boisduval and LeConte 1833. The name was originally misprinted “ Mitouri ,” then corrected to Mitoura on an errata sheet inserted between pages 23 and 24. In the same paper, Scudder discusses genus Thecla and he carefully lists structural diff erences in the heads, antennae, antennal clubs, palpi, fore and middle tibiae, and wing venation between it and the new genus. The name Mitoura combines the Greek words mitos for thread, and oura for tail (Jaeger, 1955) to make “thread-tail” , the only descriptive word used in the taxonomy of Mitoura gryneus sweadneri that gives intrinsic information, for indeed all the species of Mitoura do have thread-like tails.

PAGE 27

11 Genus Callophrys The genus Callophrys was described by C. J. Bill berg in 1820 from the type Papilio rubi Linnaeus (Miller and Brown, 1981), a western European hairstreak species that is green and lacks tails. Bird et al . (1995) wrote that the name Callophrys is a compound of call from the Greek kalos for beautiful and ophrys for brow, meaning “beautiful brows”. Billberg may have been referring to white ventral hind wing markings that look lik e eyebrows, or maybe the striking white rings around the eyes common to many hairstreak species. The genus is Holarctic in occurrence; however, the genitalic characters of both males and females show marked differences between Palearctic and Nearctic species (Johnson, 1980). Johnson and Kruse (1997, p. 5) explained that the genitalia of both sexes of “European and North African true Callophrys . . . have complex and highly sculptured morphological elements while the American “ Callophrys ” have robust and simple configurations.” The genus is listed in McDunnough ( 1938) and includes four American species. Miller and Brown (1983) listed seve n species north of Mexico. Both of these taxonomic lists treat Callophrys, Mitoura, and Incisalia Scudder (the Elfins), as good genera. Ho lland (1940) treats Callophrys and Incisalia as subgenera of Thecla . Ziegler (1960) presented a revised sc heme of classification based upon male and female genitalic characters of hairstreaks considered permanent residents of North America north of Me xico, attempting to improve on Klots (1951) and McDunnough (1938). His paper is only five pages long and consists of two and a half pages of his new list and a page and a half of keys. He lumps

PAGE 28

12 Mitoura, Incisalia, and Callophrys into Callophrys based on two aspects of the aedoeagus (aedeagus) in one of two coupl ets for the male, and a single character of the cervix between the co rpus bursae and the ductus bursae in one of two couplets for the female. He also wrote that a definitive manuscript with complete details was being prepared, but it has never been published. The weakness of this arrangement is that Ziegler only dissected North American hairstreaks. Clench (1961) followed wit h a paper the next year, relegating Mitoura and Incisalia to subgeneric status and adding three more subgenera to Callophrys . The result of these two papers sparked debate among taxonomists that has lasted for 40 years. Howe (1975) follows the subgenera scheme and Scott (1986) dispenses with Incisalia and Mitoura completely, calling them all Callophrys . Johnson wrote his dissertation (1980) and several papers distinguishing the details of the species in the genera of the Callophryina of the world. He and Eric Quinter (Johnson and Quinter, 1982) point out the problems of nomenclature resulting from, among other things, the purely arbitrary decisions about lumping or splitting genera based on t heoretical preference in a comment on nomenclatural changes in the taxonomy for North American butterflies proposed by Ehrlich and Mur phy (1982). Kimball (1965), Klots (1951), Opler and Krizek (1984), and Opler and Malikul (1992) do not list any Callophrys in their indices. Pyle (1981) and Tilden and Smith (1986) use all three genera: Callophrys are tailless hairstreaks with green ventral wing surfaces, Incisalia is the genus of the Elfins, and Mitoura are tailed hairstreaks, most with green ventrals; all have androconial patches, and most species feed on junipers or

PAGE 29

13 cedars. It is interesting that Harry Clench (1975, p. 46), in his brilliant introduction to Howe’s landmark book, said: “There is no objective way to decide the merits of these opposing views, but time is probabl y on the side of the splitters.” Original Description and Taxonomic History of Mitoura sweadneri Mitoura sweadneri was described by noted lepi dopterist Frank H. Chermock 60 years ago. In the opening paragraph of his paper, he explains that he was rearranging his lycaenids according to “the McDunnough List” (Chermock, 1944, p. 213), which probably referred to t he McDunnough revision of 1938 cited above, when he realized that some of t he specimens were undetermined. He first describes four new races of Plebeius scudderi , two new races of Plebeius melissa, then a new Plebeius and a new Everes species, and two new races of Glaucopsyche lygdamus. In each of these descriptions, he explains how one race may be mistaken for another, or how his newly described butterfly fits with a particular plate in W. J. Holland’s Butterfly Book , or J. A. Comstock’s Butterflies of California . Then he explains which alar character(s) sets the new race, or species, apart from those previously named. For two of the Plebeius races and the new Plebeius species, he mentions differences in the genitalia but does not give specific characters. However, when he describes Mitoura sweadneri as a new species, though he mentions its al ar characters that are like other Mitoura , he does not specifically state which characte rs of this new butterfly distinguish it from other members of the genus, nor why he chose to call it a species instead of a race. These omissions have left ample room for assumptions and speculations by later writers and have contributed to conf usion about the status of this taxon.

PAGE 30

14 His description was based on 75 spec imens collected in June, 1940, and eight more taken in June, 1943 by his good friend, Dr. Walter R. Sweadner, Curator of Entomology of the Carnegi e Museum. All were captured at St. Augustine in St. Johns County, Florida, the type locality. This description is curiously the longest, most detailed one in his paper. Here is the complete text of that description; the spelling and grammar are ChermockÂ’s own. Expanse of both sexes, 25 mm. This spec ies is grayish brown, with a slight greenish iridescence above, with an extra-median white band showing through on the secondaries of some exam ples. The fringes of the primaries are somewhat lighter than the rest of the wings, and are almost white on the secondaries. The lower side of the prim aries are light brown, suffused with green. The extra-median interrupted wh ite line is complete and shaded inwardly with brown as in damon . The fringe is greyish brown. There is a narrow, sometimes broken band of whit e on the margin inside the fringe which on flown specimens may be mist aken for the fringe. The ground color of the secondaries is the same as t hat of the primar ies, but the green suffusion is heavier. The tails are black, tipped with white. The fringe is grey. The inner angle is composed of a large black patch above which is fair-sized white spot. A fine black marginal line occurs as in damon . Internally to this occurs a narrow but distinct, continuous white line from the outer angle to the inner angle. Inward from this there is a blue field overcast with white scales. This field is wi dest at the inner angle and tapers off toward the outer angle. On the inner edge of this field, between medial veins 1 and 2, occurs a large black patch as in siva . Another black patch occurs at the upper end of the blue fi eld. The zigzag extra-median white band is shaded like the same band in siva . It is shaded internally by a brown band. In the basal area there occurs two lar ge white spots. In rare cases these two spots fuse into o ne elongated spot. (Chermock, 1944, p. 216). He designated the holotype and allotype from the first series of 75 and all the rest as paratypes. There is some m ystery surrounding the location of the 83 specimens Chermock examined. Kurt Johns on found the holotype at the National Museum of Natural History (Johnson, 1980). Miller and Brown (1981) said the location of the type was unknown, but t hey speculated that it may be in the collection of the Carnegie Museum, Pi ttsburgh, Pa. Kimball (1965) gives a

PAGE 31

15 detailed list of material held at the Car negie reported to him by H. K. Clench and though he mentions microlepidoptera collected by Sweadner from southern Florida along with specimens from St. Johns and Flagler Counties from other collectors, there is no mention of an M. sweadneri type. The earliest reference found was from Rawson and Zieg ler (1950), who compared two M. sweadneri paratypes provided by W. P. Comstock of the American Museum of Natural History in New York with their newly described Mitoura hesseli, so apparently some of Chermock’s paratypes were hel d there. I was fortunate to find an M. sweadneri paratype at the Florida St ate Collection of Arthropods in Gainesville. It is a male collected at St. Augustine, 15 June 1921, and carries a yellow tag designating it as a paratype and signed by F. M. Chermock. It has the alar characters described for this butterfly. In the 60 years since its description, this butterfly has attracted little attention, so that the tota l literature in which it occurs could fit on less than ten pages with ample room for maps and phot ographs. A list of the significant references organized under the names applied by each author comprises the whole taxonomic history of Sweadner’s Hairstreak. Mitoura sweadneri Chermock (1944) is the original description. Rawson and Ziegler (1950) use Mitoura sweadneri on page 79 in their acknowledgments, but in the body of the paper, page 70, refer to it as “this species, or well-marked race of gryneus ” which provides the first printed suggestion this is not a species level taxon. Johnson (1980, p. 563) argues that sweadneri deserves species status. Bridges (1994, pp. II.75 and IX.79) lists sweadneri as a full species.

PAGE 32

16 Minno (1994, p. 623) in one line explains why the species is omitted from a list of endangered biota of Florida. Mitoura gryneus sweadneri Klots (1951, p. 141) refers to sweadneri as a subspecies of M. gryneus limited to Florida and disti nguished from the nominate species by the lack of orange-brown discal shading on the dor sal wings and with a larger black patch at the anal angle and the patch in cell Cu1 lacking orange on the ventral hind wing. dos Passos (1964, p. 54) says that sweadneri is a subspecies. Kimball (1965, p. 48), and Cushman and Murphy (1993, p. 39) were cited above. Miller and Brown (1981, p. 107) also use subspecies sweadneri . Opler and Krizek (1984, p. 97) refer to the Florida subspecies of Olive Hairstreak. Emmel (1993, p. 124) says a forthcomi ng paper will elevate the taxon to full species status again. Mitoura grynea sweadneri Miller and Brown (1983, p. 54) say it is a subspecies. Opler and Malikul (1992, p. 110) say M. g. sweadneri may be distinct enough to be considered a separate species. Calhoun (1996, p. 220, 1997, p. 45) re fers to the taxon as “generally recognized as a subspecies”, but ack nowledges that it may be a sibling species. Mitoura gryneus Pyle (1981, p. 439) does not grant t he butterfly subspecies status, but does acknowledge that it uses a different host, Juniperus silicicola , Southern Red Cedar, in Florida. Callophrys (Mitoura) gryneus smilacis Clench (1961, p. 207) first describ es characters of the subgenus Mitoura , including the rounded androconial scales, ventral wing colors, palpi proportions, and the unique spatulate co rnuti. Then he calls the Florida subspecies smilacis Boisduval and Le Conte 1833, formerly species sweadneri . Johnson (1980) explains that this smilacis = sweadneri

PAGE 33

17 synonymy is in error. The type locale of M. g. smilacis is “Georgia”, and the type locale M. g. gryneus is “Virginia” (Miller and Brown, 1981, p. 107). Howe (1975, p. 293) says “There is no orange-brown on the upperside of this subspecies in any brood.” He also notes the distribution as central and northern Florida. Callophrys (Mitoura) gryneus sweadneri dos Passos (1970, p. 32) says subspecies sweadneri as does Johnson (1976a, p. 7). Callophrys gryneus sweadneri Scott (1986, p. 374) says this Flor ida subspecies is similar to form smilacis “but lacks the red unh spot (a larger bl ack spot replaces it), the white unh postbasal bars are smaller, the whit e unh median line is straighter, and the ups lacks orange (but is sometimes yellow-orange in w Fla.)” He uses smilacis in reference to the summer form of C. gryneus across its range. Glassberg (1999, p. 82) and subsequently, Glassberg et al . (2000, p. 72) both use this subspecies name for the Florida butterfly, but do not recognize smilacis . Other Possible Names When Samuel Scudder (1869) described the new genus Mitoura , he introduced his work saying these are Amer ican butterflies, principally of New England. He designated Thecla smilacis the type species and gave its distribution as New England to Flor ida. Though Boisduval and Le Conte’s description of T. smilacis was based on a drawing by John Abbot (Miller and Brown, 1981) from Georgia, Scudder may have s een specimens from Florida, perhaps even from the same popul ation that Chermock called M. sweadneri . This suspicion is strengthened by his reference to Mitura damon (Scudder, 1889a) in which he says that a specimen “from the south” (vol. II, p. 863) has a uniform brown colored dorsal wing surface and long er tails than the northern butterflies.

PAGE 34

18 He also produced a range map for Mitura damon on plate 23 (Scudder, 1889b) that includes much of north-central Florida. In the description of a co llecting expedition to north Florida sponsored by the American Museum of Natural History in 1914, John Grossbeck details travel from Jacksonville to Pensacola from 25 September to 21 October, including stops at Devils Mill Hopper, East Gaine sville, Hogtown Creek, and Payne Prairie, or Alachua Lake (Grossbeck, 1917). He also was able to look over the collection at the University and was sent specimens from J. R. Watson, entomologist at the Agricultural Experiment Stat ion at Gainesville. On the list of lycaenids collected and recorded from the Univ ersity collection is Mitoura damon Cramer. The location is just given as Northern Florida, with a note that the larvae were found on Juniperus virginiana . He also remarks that Smilax rotundifolia , its recorded host, is “surely an error” (p. 24), and says the range of th e butterfly “extends to Massachusetts, Ontario, Dakota and Texas.” Thus M. damon could also refer to Sweadner’s Hairstreak before Chermock’s description, and Grossbeck’s report may be the first published record of the butterfly in Florida. Papilio damon Stoll in Cramer 1782 was the first name given to M. gryneus, but was preoccupied by a European lycaenid described by Denis and Schiffermüller in 1775 (Miller and Brown, 1981). That na me was replaced with Lycus gryneus by Hübner in 1819, probably based on a John Abbot discovery for which, as with Thecla smilacis, no type specimen exists (Miller and Brown, 1981). Paul Opler says Hübner named the species “after Grynea, a town in Mysia, northwest Asia Minor” (Opler and Krizek, 1984, p. 96).

PAGE 35

19 Final mention must be Holland ( 1940, p. 231 and pl. XXIX). Under the genus Thecla , he describes Thecla gryneus (Hübner), provides a drawing of its wing neuration, and calls it the type of Subgenus Mitoura . He gives the range of T. gryneus , the Olive Hair-streak, as “Ontario to Texas over the entire eastern half of the United States ” (pp. 231-232). This, too, includes Florida. Related Taxa and Their Host Plants There have been 13 Mitoura species named from No rth America north of Mexico, seven Mexican Mitoura and at least 15 subspecies and variations (Bridges, 1994). However, not all lepidopter ists agree on the status of all these individual taxa as evidenced by the diffe rences in lists from various authors. Bridges (1994) and Miller and Brown (1981) have very inclusive lists, but Brian Cassie et al . (1995) recognize only four species (as Callophrys ) , one of which has seven subspecies in all North Amer ica north of Mexico. Tilden and Smith (1986) list 11 Mitoura species from western North America and northern Mexico under the heading, Juniper Hairstreaks. Table 1-1 is a list of taxa related to Mitoura gryneus sweadneri and their known larval host plants. Of the 16 Mitoura spp. whose larval host plants are documented, 12 occur on only one shrub or tree in family Cupressaceae. The two butterflies whose larvae feed on mist letoe, family Loranthaceae, only choose dwarf mistletoes that are parasites of pine, fir, and juniper trees. Only two butterflies in the genus, M. gryneus and M. siva , have adapted to multiple hosts and all of those are junipers. These two al so have the widest distributions by far of all the North American Mitoura . This degree of host specialization can be a strong biological taxonomic character t hat would be missed by the lab-bound

PAGE 36

20 Table 1-1. List of North American Mitoura and their food plants* Taxon Author and year Type locale Larval host Family Cupressaceae barryi K. Johnson 1976 Union Co., Oregon Juniperus occidentalis Hook. byrnei K. Johnson 1976 Emida, Benewah Co., Idaho Thuja plicata Wats. cedrosensis Brown & Faulkner 1988 Isla de Cedros, Baja California Norte, Mexico J. californica Carr. dospassosi Clench 1981 Zimapan, Hildalgo, Mexicooccurs: oak-piñon-juniper scrub estela Clench 1981 El Salto, Durango, Mexico unknown gryneus (Hübner) 1819 not stated but Virginia in synonymy with damon J. virginiana L. J. pinchotti Sudw. J. ashi Buch. J. deppeana Stend. and some hybrids hesseli Rawson & Ziegler 1950 Lakehurst, Ocean Co., New Jersey Chamaecyparis thyoides (L.) B.S.P. loki (Skinner) 1907 Mt. Springs, San Diego Co., California J. californica Carr. macguireorum K. Johnson 1980 Saltillo, Coahuila, Mexico J. monosperma Torr.(Little) millerorum Clench 1981 Palos Colorados, Durango, Mexico unknown muiri (H. Edwards) 1881 Mendocino Co., California Cupressus sargentii Jeps. nelsoni (Boisduval) 1869 California Libocedrus decurrens Torr. plicataria K. Johnson 1976 Cameron Lake, British Columbia, Canada Thuja plicata Wats. rhodope (Godman & Salvin) 1887 northern Sonora, Mexico unknown siva W. Edwards 1874 Fort Windgate, New Mexico oligophagous on Juniperus spp including: J. scoplorum Sarg. J. virginiana L. J. horizontalis Moench. J. osteosperma (Torr.) Little J. monosperma (Engelm.)Sarg. J. deppeana Stend. J. californica Carr. J. occidentalis Hook. sweadneri Chermock 1944 St. Augustine, St. Johns Co., Florida J. silicicola (Small) Bailey thornei J. W. Brown 1983 Otah Mt., San Diego Co., California Cupressus forbesii Jeps. turkingtoni K. Johnson 1976 Namiquipa, Chihuahua, Mexico J. flaccida Schlecht. suspected Family Loranthaceae johnsoni (Skinner) 1904 Seattle, Washington and British Colombia, Canada Arceuthobium douglassi Engelm. and A. campylopodum Engelm. spinetorum (Hewitson) 1867 California Arceuthobium Beib. spp *This information is compiled from Bridges (1994), Brown (1983), Brown and Faulkner (1988) Clench (1981), Johns on (1976a, 1976b, 1978, 1980), and Miller and Brown (1981).

PAGE 37

21 systematist who only had access to pi nned specimens and the often-limited information that data tags provide. When Downey (1962a) cautioned against using host plant choice as taxonomic characters, it was because he considered the overall picture of what was known at the time about 837 lycaenid species, with one or more hosts known for only 56% of them. Certainly at that level, food plant records alone would be of little va lue in working out basic phylogenetic relationships of the group. However, Johnson (1976a, 1976b, 1980) showed that the very close association between these Mitoura taxa and their limited host choices has taxonomic value within this lycaenid group. Figure 1-1. Distribution of ten Mitoura species and their larval host plants This information is compiled from Bird et al . (1995), Brown (1983), Calhoun (1996), Glassberg (1999, 2001), Glassberg et al . (2000), Johnson (1976a, 1978, 1980), Opler and Malikul (1992), and Scott (1986).

PAGE 38

22 Distribution maps for ten North American Mitoura are combined in Figure 1-1. Of course, the distri bution of each butterfly mu st follow the range of its specific host though the plants may occupy wider ranges than the butterflies they host. At several points, three and occasional ly even four butterfly speciesÂ’ ranges overlap, and though they are sympatric, t hey do not share host plants. In some instances, sympatric species fly at the sa me time and may occasionally be seen at the same nectar so urces. For example, Mitoura gryneus and M. hesseli occur together in the Pine Barrens near Lakehur st, New Jersey (Rawson and Ziegler, 1950), on sand myrtle, Leiophyllum buxifolium (personal observation). The fact that these butterfly taxa maintain thei r host specialization contributes to the taxonomic value of host choice and may offe r clues to species radiation in this complex genus.

PAGE 39

23 CHAPTER 2 LIFE HISTORY Learning the life history of any butterfly is a fascinating study immersed in patient observation. If the species is of conservation concern, there is a certain urgency to grasp how the organism intera cts with its habitat and what qualities are essential to a habitable site. The mo re precisely the butterfly’s niche is understood, the greater the c hance of identifying, evaluat ing, and, if necessary restoring or creating habitat. I first encountered Sweadner’s Hairstreak in Conservation Biology of Lycaenidae (Butterflies) published by the International Union for Conservation of Nature and Natural Resources Species Survival Commission (New, 1993) where its status is listed as rare, threatened. The entry written by Emmel (1993, p. 124) described this butterfly as “considered very local and usually very rare.” I found it intriguing that “for un known reasons, the butterfly seems to be rather tightly adapted to coasta l climates and is rarely found very far inland” even though its host, Southern Red Cedar, grows all across north central Florida. Emmel notes the distribution as known in 1987, observed population size, basic habitat requirements, three annual broods, and gives a brief description of the mature larva. I carefu lly searched the literat ure published since the species was named by Chermock in 1944 and found that a complete life history of M. gryneus sweadneri had not been described. Chermock’s description of the imago was based on pinned specim ens. Though my primary interest has been working out habitat requirements in situ , in vitro study has allowed me to

PAGE 40

24 pose several questions and make plausible guesses about M. g. sweadneri Â’s adaptation to north central Florida, the most southeastern extension of the speciesÂ’ range. I had several objectives which I hoped to explore with a captive colony. What is the duration of development from egg to adult? How does this time compar e with the observed phenology? How many stadia do the larvae pass through to pupation? How does the sex ratio of a captive cohort compare with that observed in the field? Is the butterflyÂ’s host choi ce obligatory, and if so, is Juniperus silicicola the obligate host? Where does pupation take place and how might this affect survival during overwintering, temperature extremes of hot and cold, drought, or forest fire? If M. g. sweadneri were a candidate for captive propagation, could it be reared en masse ? Specifically: could the larv ae be raised in close quarters without succumbing to problems from moldy frass, or cannibalizing each other? Will captive-raised adults mate in captivity? Another major quest of my study was to test the status of M. g. sweadneri against that of M. g. gryneus within the biological species concept. Controlled mating experiments require virgin butterflies of both sexes and both presumed species in reproductive-ready state on the same day, in the same place. Raising M. g. gryneus for mating experim ents offered a chance to compare the life histories of these two butterflies, docum ent the differences/similarities of the immatures (eggs, larvae, and pupae), and to ponder the mystery of their eclosion strategies.

PAGE 41

25 Of fifteen Thecla species which Holland (194 0) listed in Subgenus Mitoura , he only knew of two with described life histories. One is a variety of Thecla siva Edwards, yclept Thecla juniperaria , for which the eggs, larvae, and chrysalis had been described by J. A. Comstock in his 1927 book, The Butterflies of California . The other was Thecla gryneus . In HollandÂ’s notes on the early stages of this latter species, he lists Juniperus virginiana as the host, claims two broods in the North and three in the South, and says t he life history has been described by several authors, though he does not give a spec ific reference. H. M. Tietz (1972), in his invaluable bibliographi c contribution to lepidopterology, listed 13 references containing life history information on Mitoura damon and eight of its synonyms or forms. Nine of these are descriptions of t he larva (presumably t he last instar) and chrysalis, two are brief descriptions of the larva only, and one is an almost complete life history by Scudder (1889a) . The Tietz index is an especially important reference because he began by rechecking the accuracy of all the references in a bibliography of life hi stories of Lepidoptera by Henry Edwards (1889), then added all the life history pub lications he could find from 1889 until about 1950 (Field and Clarke, 1972). The earliest reference to M. damon is from 1862, so Tietz effectively covers all the seminal life history publications on this butterfly. Johnson (1980) only found four references to the early stages of M. gryneus and these only covered the larva an d chrysalis. I chose Samuel Scudder (1889a), a renowned Harvard scientist, as my bas ic reference to the life history of Mitoura gryneus because of his careful attention to detail in the physical description, and generous illustrations (S cudder, 1889b) which include detailed

PAGE 42

26 drawings of diagnostic characters of immatures and adults as well as eighteen-stone color lithographs of the larv a and fifteen-stone lithog raphs of both male and female adults. He also prov ides some details of adult behavior, distribution, and known variations (inclu ding particulars of a southern specimen) which have been helpful in my study. Methods and Materials of Life Hi story Study: Immature Stages Gravid female Mitoura gryneus sweadneri were taken from three sites in Dixie Co. and one site in Putnam Co., Flor ida, with habitat that harbored reliable populations annually. These sites lie along a line that latitudi nally bisects the current known range of M. gryneus sweadneri . Gravid female Mitoura gryneus gryneus were taken from one site in Giles Co., Tennessee, which is near the geographic middle of the taxon’s range. Tabl e 2-1 is a complete list of capture Table 2-1. Dates and sites of gravid females taken for captive rearing Species(subspecies) Capture date Site Designation: Latitude/Longitude Site Location 5.Jul.1998 12.Jul.1998 002 N 29 ° 33.889’ W 083 ° 22.804’ FL: Dixie Co. nr. Jena E. side CR 361 7.5 mi S. of 358 18.Jun.1999 004 N 29 ° 33.261’ W 083 ° 22.877’ FL: Dixie Co. nr. Jena E. side CR 361 8.1 mi S. of 358 25.Sep.1999 002 3.Jul.2000 28.May.2001 020 N 29 ° 33.754’ W 083 ° 19.999’ FL: Dixie Co. nr. Jena Horse Shoe/Bowlegs/Mainline Rd. 3.27 mi E. of CR 361, S. of 358 14.Jun.2001 002 30.Jun.2001 020 21.May.2002 002 Mitoura g sweadneri 13.Apr.2003 31.May.2003 019 N 29 ° 37.565’ W 081 ° 47.668’ FL: Putnam Co., Hollister Etoniah Creek Office, Division of Forestry. SR 20 at Wipple Tree Rd. Mitoura g. gryneus 5.Apr.2000 7-8.Apr.2001 16.Apr.2003 022TF N 35 ° 05.316’ W 086 ° 51.909’ TN: Giles Co. nr. Elkton Turner Farms. Hwy 273 1.2 mi E. of I-65

PAGE 43

27 dates and locations for founders of captive colonies. The Tennessee site is 432 mi (695 km) northwest of the Dixie Co. sites and 479 mi (771 km) northwest of the Putnam Co. site as measured with GPS. Adult butterflies were held in 16 oz yellow plastic drink cups manufactured by Solo Cup Company of Urbana, Illinois. Each cup had two holes ~2.5 mm diameter drilled on opposit e sides between 2.3-3 cm down from the lip, and a single hole of 4.5-7 mm diameter in the center of its bottom. A Q-tip cotton swab with one end snipped off was inserted tight ly into one of the 2.5 mm holes and served as an artificial flower; the s haft offered a perch and the cotton-wrapped end as a nectary inside the cup. Adults were fed very dilute, freshly prepared honey/water, or commercially available fruit punch-flavored Gatorade applied to the swab with a hypodermic needle. Females were given a 15-20 cm growing tip of a cedar bough, usually from the tree they were captur ed on, or from the female Juniperus sp. native to their habitat. The side branches from the lower 3-4 cm of the stem were trimmed off and this cu t end inserted through the bottom hole from inside the cup so that the lower 2 cm pr otruded from the cup. A femaleÂ’s cup with cedar sprig was then placed inside another whole Solo cup which functioned as a reservoir for water so t hat the cedar stayed fresh for about two weeks. Each butterflyÂ’s cup was covered with a 23 cm square of nylon tulle (sheer net) held in place with a rubber band. Freshly cut cedar from Giles Co., T ennessee, was supplied by over-night express mail at two to three week inte rvals by J. D. Turner of Huntsville, Alabama, for the M. g. gryneus captive colony, and local cedar was cut as

PAGE 44

28 needed from a grove on SE 71st St., and from the Natural Area Teaching Lab at the University of Florida in Ga inesville to sustain the captive M. g. sweadneri . Both cedars kept well in water for up to tw o weeks or for much longer in air-tight plastic bags in the refrigerator. All rearing experiments were carried out at my home at 318 SE 71st St. in east Gainesville, principally in my st udy, a 3.2 m square bedroom, though as the colonies grew, adults were spread out into different rooms as needed to allow for separation of the sexes. The temperature was mainta ined at near constant 25.5 ° C by day and 25 ° C nights with sometimes mid-day peaks of 26.5 ° C in high summer, and occasionally cool nights as low as 20-19 ° C at the very beginning or near the end of the flight seas on of the Florida butterfly. Humidity was constant at 50% RH. Ambient temperature and humid ity were regulated by a Trane heat pump system. Temperature was monitor ed with four Acurite brand household thermometers from Chaney Instrument Company of Lake Geneva, Wisconsin. Females were induced to lay eggs either by placing their cups in a sunny window with partially opened blinds to prov ide bright dappled sunlight, or by positioning their cups 30-45 cm below 60 W incandescent bulbs which were controlled by a timer for measur ed intervals of light and dark. Soon after eggs hatched, the new larvae were transferred with a camel hair brush to 3-4 cm cedar sprigs placed inside air-tight 40 dram clear plastic vials, No.55-40 by Thornton Plastic Company of Salt Lake City, Utah. Fresh cut cedar stays moist and edible for a week to ten days inside these vials without any water and the growing larva can be observed without opening the co ntainer. As the

PAGE 45

29 larva proceeds through larger instars, cedar tips up to 12 cm can be tucked into the vials. Most final instars were allow ed to begin pupation in their plastic vials, when the pupae (or in some cases prepupae) were moved to Solo cups to complete development. These eclosion cups, covered with tulle but given no other accoutrements, were lined up in order of pupation date along a 3.66 m by 43 cm solid brick hearth against a brick wall with fireplace that received no direct sunlight. Thus development of the adult and any diapause proceeded at the same ambient temperature as the early stages except that, due to its brick construction, temperature on the hearth changed more slowly than that in the study. The only exception to this treatm ent was of small co horts of pupae that participated in two experiments to explor e/demonstrate the value of cedar duff described below. All plastic cups and vials were identified and dates of important milestones and page numbers noted on removable mult i-purpose labels manufactured by Avery Dennison. These labels are easily transferred to newly washed cups and vials so that important data on two to f our labels can accompany each individual from egg to adult. Detailed data and observations for each individual or container were recorded each time it was examined using loose-leaf pages added to threering binders for each species, for each season. The label data provided quick reference without having to flip through se veral notebook pages to find previous data. As each individual died, its labels were transferred to the final notebook page where its individual history was su mmed up; thus the labels serve as an index to the life of each butterfly.

PAGE 46

30 During rearing, the eggs, larvae, and pupae were studied with a 5-inch, optically ground 3-diopter magnifier lens lit by a 22 W circle daylight flourescent tube and, frequently, a 10 X, 15.8 mm triplet hand lens. In itial measurements of larger larvae and pupae were made with a Helios vernier caliper and later with a Digimatic electronic caliper model number 500-351 by Mitutoyo Corp. of Japan. Fine measurements of eggs and shed head capsules were made with a Wild model 308700 dissection microscope fi tted with a Bausch & Lomb ocular micrometer disc number S 58365 and calibra ted with a Bausch & Lomb stage micrometer number S 58436. The developmental stages of the butterf liesÂ’ life histories were documented in three different formats to provide illustrations for easy comparison. Color transparencies were made on Fujichrome Velvia, RVP 36, ISO 50 professional slide film exposed with a Nikon N8008 35 mm camera and a Micro-Nikkor 105 mm, f/2.8 lens, sometimes with various combinations of Nikon Teleconverter TC201 (2 X), Nikon extension tube PN-11 ( 52.5 mm), and extension ring PK-12 (14 mm). Lighting for macro photography was fr om either a Nikon SB-24 or SB-25 strobe light, or both in tandem. The came ra was mounted on a small tripod with the lens set at closest possible focus and a focusing rail improvised with large books with slick covers. Figure 2-1 illustrates this equipment set up to photograph hatchlings. Microscopic images of the immature stages were made with a Leica MZ 125 stereomicroscope with integrated video ca mera using Auto-Montage software by Syncroscopy of Synoptics, Ltd. Scanning electron micrographs of the

PAGE 47

31 ultrastructure of eggs and head capsul es were made on a Hitachi S-415A scanning electron microscope (SEM) using Polaroid 55 ISO 50 9 x 12 cm black and white instant sheet film which also produces a very high quality negative. The accelerating voltage for all SEM images was 15 kV. The egg specimens Figure 2-1. Camera used to photograph immatures The biology books function as an improvised focusing rail. were dried using hexamethyldisilazane in a treatment perfected by James L. Nation (1983) at the Depar tment of Entomology and Nematology, University of Florida. The SEM specimens were m ounted on stubs with double-sided tape and coated with gold-palladium 100 angstroms thick in a Denton Sputter Coater. Eggs Occasionally, a newly captured female M. g. sweadneri will lay a few eggs within a day or two after she is captured, but usually oviposit ion begins on the third day of captivity. The eggs are laid singly, often just back from the growing tip of a branchlet where the new growth emerges from the old. If the cup is

PAGE 48

32 stuffed with a robust branch tip, the fema le may begin to lay anywhere except the single palest green elongated primary tip. Wi th a more modest branch tip that does not fill the cup, so that the butterfly has more room to maneuver, she will begin oviposition on the lowest branchlet, nearest the bottom of the cup and work her way up from the more secondary br anchlets toward the apex. If not given a new branch over a few days, she will fill the cedar tip with eggs from bottom to top. Size Six fertilized eggs of M. g. sweadneri , and six of M. g. gryneus were measured and all found to be within 0.01 mm of a height of 0.37 mm, and width of 0.67 mm. These measurements are very close to the height, 0.32 mm, and width, 0.62 mm, reported in ScudderÂ’s ( 1889a) account. The 0.05 mm differences may be attributed to the methods of meas urement, or to how fresh the eggs are when they are measured. Freshly laid eggs are relatively plump. Unfertilized eggs shrink as they dehydrate, firs t losing height, then assuming a flattened doughnut shape as they become completely desiccated. General Appearance The freshly laid egg is green inside a white chorion, giving an overall appearance of pale green. As the embryo de velops, the green diminishes so that the egg looks almost completely white by the time it hatches. Figures 2-2 A and B are of a Mitoura gryneus sweadneri egg; 2-2 B was made the day after the egg was laid and 2-2 A was made two days later. Figure 2-2 C is of M. g. gryneus eggs. The one on the left is just about to hatch and the darkened area just to the left of the micropyle rosette is wher e the hatchling is chewing its way out.

PAGE 49

33 Figure 2-2. Mitoura eggs from Florida and Tennessee A and B) lateral aspect and top view of M. g. sweadneri eggs at 88 X normal size. C) is an M. g. gryneus egg about to hatch and a previo usly hatched egg at 60 X life size. All three ar e Auto-Montage images. The hatched egg on the right provides a view of the inside of the shell, showing the interior structure of the external prominences described by Scudder (1889a) C A B

PAGE 50

34 in his notes on Mitura damon and illustrated for him by J. H. Emerton (Scudder, 1889b, plate 65). Mitoura eggs are often described as echinoid-shaped. Both M. g. sweadneri and M. g. gryneus eggs are somewhat flattened on the top and much more flat on the bottom, with a s light depression in t he top, the annulus, with its inner ring or rosette surrounding t he micropyle. These features along with the prominences or tubercles contribut e to the overall sh ape resembling a sea urchin test. The chorion of insect eggs is produced by the follicle cells inside the ovary so that the surface of t he egg reflects the form of those cells (Chapman, 1969, Downey and Allyn, 1981). This acc ounts for the hexagonal shape of the prominences. The lacy appearance of the ridges and wrinkles in the chorion shown in the SEM micrograph of an M. g. sweadneri egg in Figure 2-3 is the result of unevenness of the laying down of the protein, chorionin, by these follicular cells (Chapman, 1969). Butterfly eggs that are laid on the surface of the host plant are glued in place by a secr etion of the accessory glands. This adhesive sometimes causes scales from th e female to be stuck to an egg as shown on the hatched egg in Figure 22 C, and also produces the “strings” associated with that same hatched egg and near the base of the egg in Figure 23. Scudder (1889a) recognize d the potential taxonomic value of the visual aspects of butterfly eggs and pioneered t heir use in his keys to genera. In his table of genera of Theclidi (1889a, p. 801), based on the egg, for instance, he separates Incisalia , the genus of the Elfins, and Mitura , from other genera that Ziegler (1960) later lumped into Callophrys . For the next 84 years, various

PAGE 51

35 Figure 2-3. Mitoura gryneus sweadneri egg, SEM micrograph Originally shot at 100 X magnification, reproduced here approximately 134 X life size. authors demonstrated the taxonomic val ue of egg characters. Coolidge (1924), and Comstock and Dammers (1932, 1938) not ed differences in size and shape of the prominences on the egg shells , and other characters in several descriptions of Mitoura eggs of species occurring in the western United States. Though many of these taxonomic decisi ons still stand today, Downey and Allyn (1981) point out the lim itations of drawings made wit h the light microscope due to the minute size of ridges on the chorion, complexity of sculpturing and how it is revealed by different light intensities or light from different directions, and the subjectivity of the artistÂ’s concept and rendering. They suggest we should be

PAGE 52

36 cautious in comparing the dr awings made by illustrators. Six different artists drew Scudder’s plates. Ultrastructure By the early 1970s, a few authors of life histories of Lycaenidae began using SEM images of eggs which showed deta ils of the chorion ultrastructure. In a study of the immature stages of Leptotes cassius theonus , Downey and Allyn (1979) used 15 SEM figures to provide astonishing detail and had already examined enough lycaenid eggs to recognize such generalities as “Most lycaenid species have three to five “petals” in the rosette, and intra-specific variability is not uncommon.” (1979, p. 1) . In a second paper (1980) they compared the eggs of 13 species of Riodinidae followed by a two part study (1981, 1984) of lycaenid eggs in which they treated 67 specie s from 23 lycaenid genera using 156 SEM micrographs. In seven of these micr ographs they demonstrated dramatic, species-significant differences in the ultrastructures of the eggs of Mitoura gryneus , M. nelsoni , M. siva , and M. hesseli . For the latter three they also list specific numbers of petals in their rosett es, four, four or five , and five unclosed, respectively (1984, p. 7). Scudder (1889a) said M. gryneus eggs had six of these petals surrounding the micropyle. I had hoped I could demonstrate a consistent difference in the number of petals between M. g. gryneus and M. g. sweadneri eggs, since their general appearance is so similar and the rosette of the M. g. sweadneri egg in Figure 2-2 B clearly has five petals. However, close examination of a sample of 15 M. g. gryneus rosettes revealed the number of petals can vary between three and six. Figure 2-4 is a collection of SEM micrographs of the annuli of the eggs of these two butte rflies. Figure 2-4 A shows

PAGE 53

37 Figure 2-4. Annuli of Mitoura g. gryneus and M. g. sweadneri eggs, SEM micrographs A) of a Tennessee female , B-E) from a Florida butterfly. All were shot at 500 X magnificati on, shown here 330 X natural size. the 6-petaled rosette of one M. g. gryneus egg, the form expected from ScudderÂ’s description. Figures 2-4 B through E are of M. g. sweadneri eggs which illustrate the full range of variabi lity observed in the rosettes of both

PAGE 54

38 butterfly subspecies. Even mo re importantly, the four M. g. sweadneri eggs were from a single female in a sample of 11 eggs that she laid on the same day, 12 June 2003. Even the number of closed cells in the annulus, including the joined cells of the micropyle rosette, varies wid ely from an observed high of 46 to a low of 19. With this much va riation in the configurati on of the annulus, clearly the number of petals surrounding t he micropyle is not a useful taxonomic character with which to separate these butterflies. Furthermore, superfici al differences in the sculptured surfaces of the tubercles in lycaenid eggs (which appear remarkably similar in the chorions of M. g. gryneus and sweadneri ) have little taxonomic significance (Downey and Allyn, 1984). Period in Ovum The time of development in the egg is usually five to six days for both Mitoura g. gryneus and M. g. sweadneri . Individual eggs were monitored at six to twelve hour intervals from time of ovi position until hatchlings eclosed. Data are from small cohorts from differ ent broods over three years of in vitro study. The samples include eggs from both wild-caught and captive females. Table 2-2 Table 2-2. Period in ovum compared for M. g. gryneus and M. g. sweadneri Mitoura g. sweadneri eggs Mitoura g. gryneus eggs sample size 156 137 mean 5.4 days 5.7 days standard deviation n-1 0.53 days 0.87 days shows the average period in ovum. The extremes for M. g. sweadneri eggs were only two that hatched in three days, none in four, and no eggs seven or more days old were observed to have hatched. The youngest M. g. gryneus eggs were

PAGE 55

39 four 4-day eggs, the oldest were thr ee 8-day eggs, and 24 7-day eggs, or about 17% of the sample, produced viable hatchlings. First Instar The new larva usually chews an opening in the top of the ovum, consuming the annulus and part of the upper edge of the chorion. About one in six takes part of the top of the egg and fini shes its exit with an opening in the side of the egg. The hatchling does not eat mo re than a third of its egg and does not attack other eggs nearby. For three M. g. sweadneri hatchings observed, it took 45 min from the time a hole first started to be vi sible to the naked eye until the hatchling ventured over the edge of its escape hole. No hatchling has been seen to leave a silk thread or trail on its initial exploration of the cedar sprig. Hatchlings usually crawl a short di stance from the egg and begin to feed, most within the first hour ex ovum . A few move about the cedar, making tentative feeding attempts, then moving on but settli ng into feeding within the first three hours. For hatchlings not in full feeding posture within six hours, the likelihood of survival declines. No first instar ob served wandering for eight hours or more without feeding was known to survive. Non-feeders may live 12 to 18 h before they expire. When a first in star feeds, it begins by chewing a hole in the cuticle and epidermis of the cedar leaf which is a little bit larger than its head. As it continues to feed upon the soft tissue insi de the leaf, it extends its whole head into the leaf on a long narrow ce rvix. Figure 2-5 is of an M. g. sweadneri first instar in full feeding posture on the growing tip of its host. At 32 h ex ovum this one has already made several of these bor es. When the bore is into the side of

PAGE 56

40 the leaf, it leaves a round hole whose edge stops the thoracic folds from following the head into the plant. Figure 2-5. Feeding posture first instar Mitoura g. sweadneri Auto-Montage image of a living larva was 25 X m agnification shown 55 X life size. General Appearance The freshly eclosed hatchling is almo st transparent pale yellow, with a flush of faint green visible inside the dorsu m from just behind the head through the thoracic segments and ext ending about half way down the abdominal segments. The head is very pale yellow-brown wit h darker brown in the area around the semi-circle of six stemmata. The setae of the first instar in Figure 2-5 appear to be of a nearly uniform length of 0.24 mm. However, examination at higher magnification, Figures 2-6 A and B, rev eals that these nearly equal-sized setae

PAGE 57

41 Figure 2-6. First instar Mitoura g. sweadneri : A) lateral aspect, B) dorsal aspect. A) Auto-Montage image at 40 X magnifi cation shown 85 X life size. B) same larva at 50 X magnification reproduced 82 X life size. are produced on a dorsal row of paired papi llae which also sport much shorter ones. There are also two lateral rows of much shorter spiculiferous setae above the spiracles. Below them is a lateral ro w of papillae, each with the longer setae.

PAGE 58

42 These observations are in agreement wit h ScudderÂ’s (1889a) characters for the tribe Theclidi and the genus Mitura . Figure 2-7 presents lateral and dorsal aspects of a slightly younger M. g. gryneus first instar which has the same setae pattern. Before the hatchlings are a day old, their heads have darkened to brown Figure 2-7. First instar Mitoura g. gryneus A) lateral aspect, B) dorsal aspect. Auto-Montage images are 52 X life size. and when the larvae have been feeding for about 30 h they begin to show a faint green wash of the dark green patte rn characteristic of these Mitoura larvae. By 58 h, the green has darkened to the colo r scheme of the mature larva. The M. g. gryneus hatchling, 03 1W1G1, of Figure 2-7 was eight hours ex ovum when frozen for Auto-Montage and measured 1.16 mm long by 0.37 mm at widest thoracic segment. Mitoura g. sweadneri first instar, 03 47SW5, of Figure 2-6 was 24 h ex ovum and measured 1.41 mm long, 0.5 mm wide. The measurements given for each larva are of that indivi dual only and are not intended as average sizes for that instar of ei ther butterfly. Because the size varies with the age of the caterpillar and whether or not it is ex tended when measured, the only reliable (and comparable) measur ement is of the scler otized head capsule.

PAGE 59

43 Figure 2-8. First head capsules, SEM micrographs A & C) M. g. sweadneri , B & D) M. g. gryneus . A & B shot at 300 X magn ification, shown 197 X and 203 X natural size. Lateral aspects, C & D, shot at 380 X magnification, shown ~250 X natural size. Head Capsule The first instar head capsules of both M.g. sweadneri and M. g. gryneus are of very uniform size, 0.24 mm measured ac ross the widest lateral extent of the genae. The size of this sclerotized structur e is limited by the internal dimensions of the egg. Figure 2-8 shows t he head capsules of both first instars at four slightly different angles that reveal the earliest views of the same setae visible on each succeeding head. Note the stemmata are in their most lateral position on the genae in this first head capsule.

PAGE 60

44 Duration Table 2-3 compares the duration of the first instars of both butterflies. The samples include larvae that produced both male and female butterflies fed on the host native to their population. Devel opment was monitored at approximately 6-hour intervals from hatching. All indivi duals in this sample began pupation from Table 2-3. First instar duration 1st instar M. g sweadneri M. g. gryneus sample size 15 15 mean 3.7days 3.7 days standard deviation n-1 0.53 days 0.46 days the fifth instar. The mean duration is appr oximately 3.7 days. The shortest time for both was 3 days; the longest was 4.5 days. Second Instar The early second stage larvae of both butterflies are immediately distinguishable from late first instars by the much greater num ber of setae. The paired dorsal papillae of the third thoracic segment, which supported three setae in the first instar, now produce six. Thes e papillae have lost a little of their sharp definition (Figures 2-9 and 2-10), as the larvae bec ome more robust and the coloration blurs the distinct shape. The setae, which seemed long and single in the first instar, are now only 0.15 mm long and appear to sprout in little bouquets. This is the only stage with this appearance. T he green is bright almost glowing in the newly molted larva, darkens some after a half dayÂ’s feeding, and becomes dull near the end of the stage. The new head is light, translucent yellow green with dark brown mouth parts and stemmatal area. The M. g. sweadneri second

PAGE 61

45 Figure 2-9. Second instar M. g. sweadneri A) lateral aspect, and B) dorsal aspect Auto-Montage images at 40 X magnification shown 45 X life size. instar, 03 47SW6, of Figure 2-9 was 36 h since first molt when frozen for Auto-Montage and measured 2.41 mm long by 1.06 mm wide after the image was made. Mitoura g. gryneus second instar, 03 1W2G18, was 50 h since molt when prepared for Auto-Montage, 2. 94 mm long, by 1.09 mm wide.

PAGE 62

46 Figure 2-10. Second instar M. g. gryneus A) is 45 X, and B) at 43 X life size. Some second instars still feed with the head inside the stem when the cedarÂ’s growth tip is large enough, though most eat with the head outside the leaf. The head is not visible in either case because it is covered as with a cowl by the first thoracic segment. Second Head Capsule Little change occurs in the second head capsule. The mandible sports a small knob distally. As the genae have bec ome larger, the apparent position of the stemmata is slightly more ventral t han in the first instar. In a small sample

PAGE 63

47 Figure 2-11. Second head capsules A) M . g. sweadneri , and B) M. g. gryneus SEM micrographs made at 200 X magnification, shown 134 X life size. of six of each, the second head capsule of M. g. sweadneri was 0.37 mm wide (one was 0.38 mm) and the second of M. g. gryneus was 0.39 mm (one was 0.40 mm) wide. These samples were from individuals that began pupation from the fifth instar. Duration Table 2-4 shows the average length of time that each larva spends as the second instar. The shortest time recorded for M. g. sweadneri was four days, Table 2-4. Second instar duration 2nd instar M. g. sweadneri M. g. gryneus sample size 15 15 mean 4.2 days 4.2 days standard deviation n-1 0.17 days 0.42 days the shortest for M. g. gryneus was 3.1 days, and the longest second instar recorded for both was 4.5 da ys. The times recorded for M. g. sweadneri larvae included a minimum of 16 h and a maximu m of 24 h as a sedentary pre-molt.

PAGE 64

48 Molting Notes When I first started to collect head capsules and exuviae to mark the progression of stadia, I puzzled over the tiny size of the pelt remaining on the cedar stem, or sometimes fallen to the bottom of the vial. Even after it was stretched out in a drop of water it nev er seemed big enough to have covered the previous instar. Just before molting begins, the head turns a uniform dark brown. In preparation for the molt, the la rva spins a delicate lattice of silk on a cedar stem, then crawls forward and gr asps it with the prolegs. The head is drawn into the cowl of the thoracic segments and the legs are withdrawn into the thorax so that the claws no longer grasp the cedar. As apolysis proceeds, the old cuticle becomes semi-transparent and the larva mo vement is limited to a very slight arching of the dorsum when t ouched. This premolt period usually lasts a little less than a day for transition to third instar, but may take 36 to 48 h in later instars. The actual moment of ecdysis happens very quickly. First, the head capsule pops off and the old skin splits from the front to the top of the thorax, then shrinks back toward the end of the abdomen as in Fi gure 2-12 A. The new instar crawls forward just far enough to turn around (2 -12 B), then enters the exuviae and begins eating it from the inside (2-12 C). In Figure 2-12 D, t he new third instar has finished all but about a quarter of t he exuviae, and two tracheae from the last pair of spiracles can be seen spiraling out from the remaining pelt. Usually more than half of the shed skin is eaten, and quite often the remainder is still held in place by the silk lattice. In this particu lar molt, the old head capsule stuck to the new setae while the larva turned around before it lodged on the exuviae.

PAGE 65

49 Figure 2-12. Molting sequence A) second instar head capsule pops off, B) new third instar turns around, and C) begins to eat the exuviae from within. D) Larva rests. Grid in A, B, and C is 1 mm.

PAGE 66

50 In some other molts I witnessed, the c apsule was propelled as much as 8 mm away as it popped off. None of the first four head capsules splits when shed, which may account for this springiness. The molting sequence documented in Fi gure 2-12 took 55 min, after which the larva rested for nearly an hour befor e beginning to feed on cedar. The only first instar molt I watched took 1 h 37 mi n as the new second instar took several rest breaks while it ate its fill of the exuviae. Third Instar The pattern of setae of the third inst ar becomes much more complex. The little mound on the dorsum of the third thoracic segment that clearly showed six setae in the second instar has become less distinct. The white dashes or crescents blur the margins so that the ear ly instar shows nine setae, but as the mound bulges and nearby pits deepen later in the stadium, 14 or even 17 setae could be included in the count. The set ae are a fairly uniform 0.14 mm long. The color of this instar remains the same as in the second stage. The two dorsal lines and lateral stripe may be whit e or yellowish white against the cedar green ground color. Third instar M. g. sweadneri , 03 47SW1, was 10.5 days old and had been feeding for 2.5 days since the se cond molt when it was frozen for Auto-Montage in Figure 2-13. It meas ured 5.67 mm long and 1.92 mm wide. The brown blemish on the right dorsal side in Fi gure 2-13 B is frostbite, an artifact of preparation for Auto-Montage. Mitoura g. gryneus larva, 03 1W3G1, of Figure 2-14 was about 38 h since the second molt, and measured 4.58 mm long by 1.67 mm wide when frozen. Both of these images show the be ginning of the

PAGE 67

51 Figure 2-13. Third instar M. g. sweadneri A) lateral, and B) dorsal images made at 20 X magnification and shown 26 X life size. sculptural change in conformation of this and later instars that is not readily apparent in the brand new thir d instar of Figure 2-12. Third Head Capsule The third head capsules of both butterflie s (Figure 2-15) are very close in size, though a sample of six of each taken from complete sets of head capsules of larvae that pupated from t he fifth instar show more individual variation than

PAGE 68

52 Figure 2-14. Third instar M. g. gryneus A) lateral and B) dorsal Auto-Montage images are 30 X life size. in previous instars. The mean width ac ross the widest point of the genae is 0.64 mm, and the largest 0.68 mm, with the smallest being 0.61 mm for both. Standard deviations are 0.032 mm for M. g. sweadneri , and 0.030 mm for M. g. gryneus . This head shows only minor setal changes from the second instar, for example, seta number 01 which arises from inside the ring of stemmata is slightly more prominent in Figure 2-15 than it is in Figure 2-11. The stemmata continue their migration down the genae and are mo re obviously ventral than in the previous head.

PAGE 69

53 Figure 2-15. Third head capsules A) M. g. sweadneri and B) M. g. gryneus SEM micrographs made at 100 X magnifi cation, shown 66 X life size. Duration The third instar lasts 3.5 days fo r both butterflies when sampled from Table 2-5. Third instar duration 3nd instar M. g. sweadneri M. g. gryneus sample size 15 15 mean 3.5 days 3.5 days standard deviation n-1 0.29 days 0.22 days individuals that began pupation from the fift h instar. The shortest duration from both samples was a 3-day larva and the longest was one 4-day larva. Fourth Instar The setae of the fourth instar are much more numerous and difficult to count because they are very close toget her and the sculpturing of the larvaÂ’s dorsum (Figure 2-16) makes approximating where one of the paired mounds on a thoracic segment stops and the next pl ane begins nearly impossible. About 24 setae sprout from the sa me spot counted in previ ous instars, and most are 0.17 mm long, with some up to 0.22 mm aro und the ventral lateral margins just above the legs and prolegs.

PAGE 70

54 Figure 2-16. Fourth instar M. g. sweadneri A) lateral and B) dorsal Auto-Montage images shot at 16 X magnification, shown 18 X life size. All Mitoura larvae have a diamond-shaped dorsal shield that Scudder (1889a) described as coriaceous, or leather y. Comstock and Dammers (1938) in their description of M. spinetorum refer to it as a cervical shield because of its location on the dorsum of t he first thoracic segment. The dorsum of the second segment has a deep groove, or sulcus. Th ese features are recognizable in the earlier stages, but especially visible in the later instars. The shield is easier to see in Figure 2-17 B, the dorsal view of M. g. gryneus , because it was frozen so

PAGE 71

55 Figure 2-17. Fourth instar M. g. gryneus A) lateral and B) dorsal Auto-Montage images at 17 X life size. quickly that it did not completely withdraw its head. The sculpturing of ridges and folds is more dramatic than in previ ous stages and each spiracle is surrounded by a “C”-shaped depression. Though all t he instars use silk and are capable of pulling themselves back to a branch if di slodged, the later stages seem to spin lifelines more often than do t he first through third instars. Fourth instar M. g. sweadneri , 03 47SW7, was about 60 h since third molt and measured 8.04 mm long by 2.92 mm wide when prepared for Auto-Montage. The blurred appearance of the first and second abdominal segments and faded white of the pattern are a result of bei ng frozen a little too long before the image

PAGE 72

56 was made. Mitoura g. gryneus fourth instar, 03 3W1G1, was 38 to 40 h old, 8.66 mm long and 2.66 mm wide in Figure 2-17. Fourth Head Capsule The fourth instar heads (Figure 2-18) are substantially the same as the previous head capsules except that t he setae along the clypeus and labrum are longer and the two semicircle s of stemmata are more ventral than in previous instars. These are especially noticeable in Figures 2-18 C and D, the left lateral views. In addition, except for a few s hort setae on the frons, virtually all the surfaces of the genae posterior and dorsal of the stemmata and the top of the Figure 2-18. Fourth head capsules A) and C) M. g. sweadneri , B) and D) M. g. gryneus are SEM micrographs made at 75 X magnification shown here at 66 X life size.

PAGE 73

57 head are completely smooth. This state is characteristic of lycaenid larvae that retract the head into the cowl of the first thoracic segmen t. The fourth instar head capsules of both butterflies are about 1.04 mm wide. In a sample of six of each subspeciesÂ’ head capsules, the smallest was 1.02 mm and the largest 1.06 or 1.07 mm wide when only individuals that pupated from the fifth instar were sampled. Duration The length of development for fourth instars is 4.5 days. Table 2-6 shows the means and standard deviati ons for both butterflies when the fifth instar was Table 2-6. Fourth instar duration 4th instar M. g. sweadneri M. g. gryneus sample size 15 15 mean 4.4 days 4.6 days standard deviation n-1 0.45 days 0.39 days the last stage before pupat ion. The shortest durat ion recorded was of one M. g. sweadneri that only took 3.6 days; the longest took 5 days. Last Instar The fifth instar is the last larval stage for about 80% of the individuals of these Mitoura butterflies. This is an exceptio n to Scott (1986) who said most lycaenids have only four larval stages. The folds and depressions of the body are at their deepest and are reminiscent of t he conformation of mammalian cerebral hemispheres (Figures 2-19 and 2-20). The setae range from 0.04 mm to 0.07 mm long, have increased in number about three fold, and cover the integument in a fine, velvety pile. The li ght dashes forming the dorsal and lateral lines are nearly white to yellowish whit e, and the juniper green ground color is the same for both. The head is green wit h brown around the mouth and stemmata.

PAGE 74

58 (In many of the Auto-Montage images the shades of the larval color pattern appear somewhat different in indivi dual images because I adjusted the brightness, contrast, and in some cases t he color balance of t he original files in order to emphasize the setae and/ or conformation of the larvae.) The M. g. sweadneri fifth instar, 03 45SW4, of Figure 2-19 froze so quickly that the prolegs are still par tially extended. Unfortunately, it also suffered frost bite to the lateral aspect of the anterio r of the first thorac ic segment (Figure 219 A). You can just see the posterior edge in an end-on view of the cervical shield in Figure 2-19 B. Figure 2-19. Fifth instar M. g. sweadneri A) lateral, B) dorsal Auto-Montage images made at 8 X magnification , shown about 9 X life size. The scale bar represents 5 mm.

PAGE 75

59 Each of these Auto-Montage images is made up of 16 to 19 separate digital files precisely aligned and superimposed to produce the 3-dim ensional view. To ensure that no movement occurs during t he process, each larva must be frozen. This caterpillar was 20 days old, four days since fourth molt, when prepared for Auto-Montage and measured 14.71 mm long by 4.70 mm wide. Mitoura g. gryneus fifth instar 03 8W1G1 (Figure 220) was 17 days old, 2.5 days since fourth molt, 13.20 mm long and 4.44 mm wide. This one froze more slowly than Figure 2-20. Fifth instar M. g. gryneus A) lateral, B) dorsal Auto-Montage images shown at 12 X life size.

PAGE 76

60 the M. g. sweadneri fifth instar, with the head, l egs, and prolegs fully withdrawn, and dorsum arched. Careful inspection of preserved final instars of both butterflies reveals possible pore cupola organs that Pierce et al . (2002) found on every lycaenid species they examined, with the possible exception of Liphyra brassolis . However, these Mitoura species lack tentacle or gans of the eighth abdominal segment and the dorsal nectary organ, or NewcomerÂ’s organ, of the seventh abdominal segment, either one or both of wh ich are found on lycaenid larvae with known ant associations (Downey, 1987, Fiedler, 1991, Pierce et al ., 2002). Furthermore, Callophrys ( M .) gryneus is listed as having no known ant association by Fiedler (1991). The progress of the final instar can be divided into three phases: feeding, wandering, and the prepupa. Behaviors and so me physiological differences in this stadium distinctly set the final instar apart from t he previous stadia. Observation of these differences led to clues of specific habitat requirements necessary for this animalÂ’s survival. The Feeder The feeding phase of the typical final instar lasts five to six days. The feederÂ’s head is rarely visible because it is completely covered by the first thoracic segment which also obscures the end of the cedar stem (Figure 2-21). In this posture, there is very little obv ious movement, just a gentle, nearly imperceptible undulation of t he blunt curve of the top of the mesothorax as the feeder chews. This instar is less parti cular about the cedar it eats than in previous instars. It eats further down t he growing stem, sometimes to its union

PAGE 77

61 with the woody stem, and will al so eat the palest, largest primary growth tip of the branch that ovipositing females avoid. At least two final instars ate the tops of juniper berries though this was not comm only observed in this stadium and never seen in any other instar. If pried from the stem with a paint brush and transferred to new host, this larva will often crawl qu ickly over the new cedar, leaving a trail of silk before settling to feed once more . This behavior was not observed in any other stadium. Figure 2-21. Final M. g. sweadneri instar feeding posture Single exposure AutoMontage at 8 X magnification, shown 6.3 X life size. Sometime during the fourth day, the wh ite or yellow markings of about half the final instars start to turn pink. By the fifth day, the color is quite striking as in the M. g. sweadneri of Figure 2-22. This color form is not related to sex, as larvae that have turned pink may become male or female butterflies. A half day or so later, the rate of feeding slows, signaling the end of this phase and the start of the next.

PAGE 78

62 The Wanderer The wandering phase may last from as few as 8 h to as long as 27 h in the lab. The wanderer will not feed, instead fo cusing its energy on getting off the host and finding a spot for pupation. This phase d oes not leave a silk trail. No life history account that I have s een stated specifically where Mitoura g. gryneus pupates. Howe (1975) said lycaenid pupae are usually in the debris at the base of the host plant, and Scott (1986) in hi s general discussion of characters of subfamily Lycaeninae said pupae may be found in leaf litter or in crannies such as loose bark, usually attached to the substrate with a silk girdle. Comstock and Dammers (1938) said captive Mitoura spinetorum pupated in a mass of the host Figure 2-22. The pink wanderer Fujichrome transparency, M. g. sweadneri final instar 4.5 X life size. plant without a silken attachment. Ho wever, in their life history of Mitoura loki they said “pupation takes place on the food plant, suspended from a silk button, and supported by a delicate silk girdle .” (Comstock and Dammers, 1932 p. 41).

PAGE 79

63 Finally, John Brown (1982) in his descrip tion of the new California species, Mitoura thornei , noted that pupation occurs in t he duff at the base of the host tree, Tecate Cypress, Cupressus forbesii Jepson. Left on their own, captive wanderers of both M. g. gryneus and M. g. sweadneri crawl off the cedar and after circling the bottom of the plasti c vials many times, hunker down on the plastic substrate to become prepupae. If touched or moved with a paintbrush, they will begin wandering again if they hav e not been settled too long. In an effort to learn where M. g. sweadneri pupates I conducted two in vitro experiments with individual larvae. Crawl off test In 17 individual trials, wanderers who had crawled off their last host branch were given new cedar branches, which t hey ignored in circling their vials or yellow Solo cups. When each wanderer was placed on the cedar, it immediately crawled off, no matter how many times the larva was replaced. In some of the trials, the branches were as long as 45 cm and held at different angles from vertical to horizontal. In every case, the wanderer crawled down, even if the tilt of the branch was changed as the larva mov ed. Figure 2-22 shows a test subject whose branch orientation has just been changed. In the typical response, this one came to the end of its branchlet, hel d on with the third pair of true legs and the prolegs, and reached out with the first two pairs of thoracic legs before turning around and proceeding down the st em. Wanderers who had come to rest on the bottoms of their cups, and been a llowed to settle long enough that they could no longer crawl when replaced on an hor izontal branchlet, could still arch

PAGE 80

64 their dorsa enough to fall off the branch. It is apparent that leaving the host is part of the strategy of preparing for pupation. Burrowing test In preparation for this test, a dry q uart of loose cedar duff was collected from beneath a large female cedar in t he U. F. Natural Area Teaching Lab and sterilized on a cookie sheet in a 350°F ov en for 30 min. Clean 40 dram plastic vials were marked to a depth of 3.5 cm from the bottom in 0.5 cm increments and filled to the top mark with sterilized duff. Nineteen wanderers were transferred to cedar sprigs in these experimental vials for observation. The typical subjec t crawled across the duff layer, then around the edge following the wall of the vial before pushing its way into the duff. With the head extended from under the thor acic folds, it burrowed into the duff almost vertically in a sort of headstand. One of the wanderers followed the clear cylindrical wall of the vial down a nd then up and down again for an hour or so before turning toward the center of the vial and out of view in the cedar duff. This burrowing effort was not excavation in that it did not involve digging, did not produce a tumulus of removed material, and did not leave a tunnel or even an obvious entry point on top of the duff. The wanderers were given four da ys to be sure they had begun pupation. Then, using forceps, the duff was carefully removed layer by layer from the top until the dorsum of the pupa was just vi sible. I could then determine how deep each wanderer had burrowed, using the gauge on the side of each vial. Due to the irregularity of the c edar pieces comprising the duff and the difficulty of removing them without disturbing the pupa, results reported in Table 2-7 are

PAGE 81

65 Table 2-7. Observed depths of burrowing wanderers in vitro Depth below surface in cm Number of wanderers 0.5 1 1.0 4 1.5 5 2.0 5 2.5 4 probably accurate to within 2 mm. None of these subjects settled on the bottom of the vial. At that le vel, the dorsum of the pupa would have been about 3.1 cm below the surface of the duff. The wanderer that settled at 0.5 cm down was little more then a single pupa height of 4 mm below the surface. Some limitations of this test to consi der include: the insi de diameter of the vial is 48 mm, which could limit how loose /tight the irregular duff is packed. This could vary from actual duff compactness in situ because the duff under a living tree has some larger sticks and other detritu s in particles that would not fit into the vials. Also, the duff overburden under a cedar is built up over several years so that the upper layers are looser and drier than the moist, more composted duff at the soil interface. These factors coul d affect how much effort the wanderer would have to expend to insinuate itself bet ween the cedar leaves and twigs. The relative ages of the wanderers when t hey began the experiment could also make a difference in how much time or energy each wanderer had to burrow. The fact that all 19 of the test-s ubject wanderes invested energy in burrowing seemed significant. The duff laye r under a cedar could provide refuge from predators, shade from the sun, or other s helter advantage to the pupae perhaps, in diapause for overwintering. I wondered how being buried in duff would affect eclosion of t he imago. After the measuremen ts were taken, the duff

PAGE 82

66 was returned to cover each pupa and the vials were covered with tulle squares to await eclosion. The results of this exper iment are continued in the discussion of eclosion. The Prepupa When the wanderer comes to rest, it does not spin a silk girdle as many other lycaenid larvae do, nor does it spin the silk lattice used by each of the previous instars of M. g. gryneus and M. g. sweadneri to secure attachment to Figure 2-23. Mitoura g. sweadneri prepupa A) lateral, B) dorsal aspect AutoMontage made at 10 X magnification shown 9.4 X life size.

PAGE 83

67 the host during premolt. It appears comple tely quiescent, but its whole shape starts to change slowly. It loses a few millim eters in length, gains a few in width, and the widest point shifts fr om the last thoracic segm ent to about the fourth abdominal segment as the dorsum arches . The head and legs are completely withdrawn; the prolegs are reduced to little rounded lumps, and as apolysis proceeds, the cuticle begins to clear so that the green color seems to come from within as in Figure 2-23. Whether this takes place on the bottom of an empty plastic vial, or in a duff-filled one, the prepupa is always resting venter down, and horizontally. This premolt may last 12 to 28 h before it defecates a final frass pellet and expells a pool of exuvial fluid. Th e final molt usually follows in 30 to 48 h. Mitoura g. sweadneri , 03 47SW2, became the prepupa of Figure 2-23 7.5 days after the fourth molt, 22 da ys after hatching. She was 14.36 mm long and 6.8 mm wide at the fourth abdominal segment. Because this phase is quiescent, she was not frozen for Auto-Montage. Final Head Capsule Unlike previous molts which bega n with the head capsule popping off in front of the larva, the final molt st arts when the head capsule splits along cleavage lines on either side of the frons continuing to the top of the head, or vertex, so that the genae are splayed apar t and the frons collapses caudad. As the final exuviae shrinks back over the dorsum of the new pupa, the caudal margins of the genae stay firmly attached to it and are first spread wide apart, then bunched back together as the shed ski n compacts like the pleated bellows

PAGE 84

68 of a concertina. The last head capsule is thus left twisted and misshapen so that measurement across the widest lateral extent of the genae is rendered moot. Finding what I took to be large, whole final head capsules led to the discovery that both M. g. gryneus and M. g. sweadneri had varying numbers of stadia. None of the partial life historie s I had consulted specified a number of instars for M. gryneus ; even ScudderÂ’s detailed descr iption (1889a) went from first instar to mature caterpillar. The mo st reliable authority I had during the first season I raised these butterflies was Scott (1986), who said that each species had a characteristic number and that mo st Lycaenidae had only four larval stadia. Initially, larvae were raised comm unally in their oviposition cups and the head capsules which I recovered from the frass usually fell into three broad size classes, with the occasional odd-sized big one. Since I did not always find the final exuviae with the pupae, I thought some final instars managed to shed the last head without splitting it. Then, duri ng the second captive brood, I started finding the tiny first instar head capsules which I had overlooked before and realized that every pupa which I had recorded from the fourth instar had actually been from the fifth. I was collecting about nine size classes of heads. When I segregated larvae into individual 40 dram vials, there was a much smaller volume of frass to search and I began to collect full sets from some individuals. From individual records and sorting piles of head capsules in generic trays for each subspecies, I have found that both butterflies commonly pass through five stadia with 17 to 23% of each brood pupating from the sixth instar. The much rarer life history is seven in stars, which I estimate to comprise a little less than

PAGE 85

69 2% of all development hist ories. Table 2-8 lists t he size classes for each alternative life history. Since these measurements define nine or ten categories into which head capsules of different stadia fit, they are only valuable for comparison when complete sets are collect ed. The smaller size of the first head Table 2-8. Head capsule (hc) size classes in mm Pupa from: 1st hc 2nd hc 3rd hc 4th hc 5th hc 6th hc 5th instar 0.24 0.37-0.40 0.61-0.68 1.02-1.07 6th instar 0.24 0.36-0.39 0. 56-.059 0.85-0.89 1.12-1.20 7th instar 0.21-0.22 0. 30-0.32 0.43-0.45 0.69-0. 71 0.92-0.95 1.15-1.19 capsule of 7-instar life histories shows t hat this is determined while the hatchling is still developing rather than in response to conditions ex ovum . Interestingly the 6-instar life history does not show the difference until the second stadium. The geometric growth of head capsules for both five and 6-instar life histories slightly exceeds the predicted factor of 1.4 of DyarÂ’s Rule (Dyar, 1890) for each molt. Of the 7-instar alternative, two molts fit, one exceeds, and the last two increase less than DyarÂ’s predictions. Duration This final instar feeds longer than the duration of any previous stage. The time wandering and the long intermolt befor e pupation bring the total duration of the final instar to nearly twic e that of the fourth instar . Times given in Table 2-9 Table 2-9. Final instar duration Time s are from fourth molt to pupa. 5th instars only M. g. sweadneri M. g. gryneus sample size 15 15 mean 8.3 days 8.8 days standard deviation n-1 0.79 days 0.86 days are only for final instars that molted to pupae from the fifth stage. The shortest time recorded for both was seven days, the longest was ten. The few complete

PAGE 86

70 records of butterflies that pupated from t he sixth instar have a mean duration of 8.4 days which is within the mean times of Table 2-9 but have a larger standard deviation, (1.60 days), due to one M. g. gryneus that pupated 12 days after the fifth molt. I have no precise record of the timing of the last stage for any of the pupae that were produced fr om the seventh instar. Pupa Immediately after the final molt, the new pupa is strikingly colored. The head and dorsum of the thorax is dark green, the wing covers are a much lighter, yellower green, and the abdomen is light yellow tan. The cuticle is noticeably tanning after three hours and usually comp letely sclerotized in six. The color becomes a dark nut brown with a halo of li ghter setae as in Figure 2-24, but may be dull black in some individuals. The ex uviae usually separates from the pupa with the spent head pointing away from the caudal abdomen. When found still touching on the bottom of a plastic vial, the attachment is very causal so that the hardened chrysalis rolls away when the vi al is handled. There was no obvious attachment to the dried cedar leaves when the wanderer was given duff in the burrowing test. The pupae of these two bu tterflies are identical. Scott (1986) wrote that lycaenid pupae frequently lack a cremaster and Edna Mosher (1916) said that no lycaenids have them alt hough the ventral surf ace of the abdomen often has groups of small hooked setae. The cremaster is an important feature for most Nymphalidae because it is t he hook that allows the pupa adheraena to hang head-down from a silk pad in the manner typical of that family.

PAGE 87

71 Figure 2-24. Pupae of both butte rflies A, B, and C) are ventral, lateral, and dorsal views of M. g. sweadneri . D, E, and F) are of M. g. gryneus . Auto-Montage images made at 10 X magnification shown 8 X life size. The bar in A) represents 10 mm. Once sclerotized, the pupal cuticle is durable: a carelessly transferred pupa can roll down the side of a Solo cup and bounce on the bottom with no ill effect. Pupae are resistant to entry by di sease organisms and somewhat water resistant. Individuals retrieved from moldy frass, setae covered in fine black mold

PAGE 88

72 spores are viable. Pupae do not do well in damp frass in airtight communal cups. If water condenses on the walls of the container, the pupae must be removed promptly to dry conditions or they will decay . This is easily solved in mass rearing if wanderers are allowed to leave their vials to pupate. Most pupae are 7.8-9.2 mm long and 45.2 mm wide although some as short as 7 mm have produced adults. Mitoura g. sweadneri , 03 47SW2, of Figure 2-24 began pupation nine days after fourth molt, 24 days since hatching, and 27 days since oviposition. She was 9.01 mm long and 5.03 mm wide. Mitoura g. gryneus , 03 6W3G6, of the same figure wa s nine days after fourth molt, 27 days ex ovum , and 33 days since oviposition. She measured 8.93 mm by 4.97 mm. Development and Duration The total time for development of thes e two butterflies is identical. Table 2-10 shows that average time from t he day of hatching until pupation began is within 4.8 h, and the interval from ovi position to pupa is within about 7 h for samples of 16 individuals monitored at 6-hour intervals that pupated from the Table 2-10. Total development time to pupation Both samples are from fifth instar life histories. M. g. sweadneri M. g. gryneus sample size 16 16 Since hatch: mean 25.8 days 25.6 days standard deviation n-1 2.10 days 0.81 days Since egg: mean 31.4 days 31.7 days standard deviation n-1 2.36 days 1.14 days fifth instar. The shortest time fr om hatch to pupa was from a single M. g. sweadneri that took 22 days; the longest from oviposition was from the same

PAGE 89

73 cohort and lasted 36 days. Available data from 6-instar individuals usually falls within the M. g. sweadneri range and may add only 6 -7.5 days to the averages in Table 2-10. Though these times are for in vitro culture with near-constant temperature of 25-26°C, they are useful for comparison of the duration of the biological process and show negligible differences. Eclosion strategy A more interesting comparison is the time in chrysalis. Table 2-11 compares records from closely monitor ed individuals of three broods of each butterfly with no regard to the number of instars, though about 20% of each was from 6-instar life historie s and none were from 7-instar individuals. If the first butterflies eclose as soon as they ar e mature, then the minimum time for metamorphosis is 11 to 12 days. The s hortest records may be an approximation of the development period needed to pr oduce the imago when only the first Table 2-11. Times to eclosion, M. sweadneri and M. gryneus Data for nondiapause adults of each cohort are given in days. M. g. sweadneri M. g. gryneus sample size 93 butterflies 120 butterflies Since pupa: range 12-16 11-16 mean 13.7 13.9 standard deviation n-1 1.03 1.44 Since hatch: range 36-53 37-50 mean 41.2 42.3 standard deviation n-1 3.13 3.25 Since egg: range 40-58 42-56 mean 46.6 48.4 standard deviation n-1 3.19 3.48 eclosers of broods are considered. Furt hermore, if these ti mes held true for the butterflies in situ , then we could expect to see new flights of both M. g. sweadneri

PAGE 90

74 and M. g. gryneus every 50 days or so throughout the flight season. Both butterflies overwinter as pupae, so we would expect the next springÂ’s first flight to be fully developed and to enter diapause to await favorable environmental conditions. Of all the M. g. sweadneri I raised with good records, only two eclosed after more than 16 days in the chrysalis. Imago, 01 19SW22, eclosed 28 days after final molt from fifth instar, 55 days ex ovum , and 61 days since oviposition, and 01 19SW4 took 27, 65, and 71 days respectively. Both of these individuals eclosed in late August and I think they might hav e waited until the next year if they were not subject to the constant conditions of my home. Captive-raised Mitoura g. gryneus from Tennessee stock demonstrate a different eclosion strategy in the lab. In a brood of 126 butterflies, 93 or 74% delayed eclosion, and in another smalle r brood, 33% eclosed long after the development period. A small F2 brood of 26 had 5 with delayed eclosion. Figure 2-25 illustrates the trend for 253 eclosions . The first 121 adults emerged within the 11-16 day range, leaving 52% in diapa use. These delayed eclosers dribbled out as singles, and as sma ll groups of two to sev en per day. The longest delayed eclosion of M. g. gryneus was 120 days. This eclosion strategy is the first significant difference observ ed in the life histories of Mitoura g. sweadneri and M. g. gryneus immatures. If environmental c onditions such as temperature, humidity, and ambient light on the brick hear th in the lab are constant for the pupae of both taxa and all the M. g. sweadneri of every brood eclose en masse within 12-16 days of pupation while in each brood of M. g. gryneus a third to a half eclose initially while t he remainder dribble out over the summer flight season,

PAGE 91

75 0 5 10 15 20 25 30 1115192327313539434751555963677175798387919599103107111115119Days since pupation beganNumber eclosed Figure 2-25. Eclosion strategy, Mitoura g. gryneus Time from pupation until eclosion in days compiled from data for 253 individuals from three broods. Tic marks of X-axis separate each of 110 days. then it seems this eclosion strategy in cludes some program for diapause that M. g. sweadneri lacks. The Florida butterflies eclose in an all-or-nothing shotgun strategy, while the Tennessee taxon ta kes a more conservative, spread-therisk/success-over-time, approach. This may be an adaptation to the climate, insolation, vagaries of weather, or other environmental conditions inherent in their habitat that dictate flight season or a consequence of genet ic from a history of varying selection. Burrowing test revisited All 19 of the pupae reburied at thei r original burrowing depth in the burrowing test above eclosed. Thirteen eclosed within the 12-16 day immediate ecloser range. Four of the original M. g. gryneus wanderers delayed eclosion. Butterfly, 03 1W2G13, emerged from 2.0 cm of duff after 30 days, 03

PAGE 92

76 1W2G16 from 2.5 cm after 39 days, 03 1W2G7 from 1.5 cm after 98 days, and 03 1W2G17 emerged from 2.5 cm after 99 days. In a follow-up trial, eight M. g. sweadneri pupae that began pupation on the bottom of empty vials were covered in 3.5 cm of sterile duff. The dorsum of each was at least 3.0 cm below the surface. Each of these eclosed in 13-15 days and, like the adults in the burrowing tes t, each escaped the pupa case cleanly and emerged above the duff with perfect wings. Though I did not witness any of them digging out, two were discovered on top of the duff, still pumping up wings with the antennae held very close together and par allel. Interestingly, adults that pupated in empty Solo cups sometimes had wi ng tips caught in the chrysalis or failed to eclose completely. Even with thes e small numbers of replication, these two experiments demonstrate that both w anderers are capable of burrowing into cedar duff, and that adults can succe ssfully dig out upon emergence. Audio Observations The pupae of both M. g. sweadneri and M. g. gryneus are capable of producing sounds soon after the final molt, even before the pupal cuticle is completely sclerotized to dark brown. These sounds may be produced with pegs and ribs between abdominal segments rubbing together when the segments move (Scott, 1986). I first noticed that pupae of both butterflies made chirping sounds when they were removed to eclosion cups from their larval vials. Specifically, these sounds occured when the pupa was touched with a pai ntbrush, but also sometimes when the vial was lifted and often when the vial was tapped. This correlation suggests

PAGE 93

77 that the pupal sounds were an alarm or deterrent to predat ors that might be startled by a chirping pupa. Sound production in lycaenid pupae has been studied since at least the 1960s (Downe y, 1966, and Downey and Allyn, 1978) and is reported in certain other families. But no explanation satisfactorily covers chirping in other situations observed frequently in in vitro culture of both of these Mitoura butterflies as described below. In a 17.5 by 26.5 cm cardboard tray holdi ng 15 plastic vials with late instars, a single new pupa sounds off with no appar ent provocation. When the pupa is taken out of the vial and placed on the lab bench, the sound becomes almost inaudible. The 40 dram vial amplifies the sound. In a 50 by 65 cm corrugated cardboard bo x holding 30 16 oz Solo cups with one M. g. sweadneri pupa in each, when one cup was removed for labeling and replaced, its resident began to chirp. Within moments, all the pupae in the box were chirping as if in response to the initial noisemaker. The cups in this box were close enough that most touched at least two neighbors. In a 40 L Rubbermaid laundry basket with 15 eclosion cups, a single pupa began chirping slowly and was joined by several others in a sound reminiscent of the chorusing of Tettigoniidae. The pupae chirping did not seem to be as well organized as the song of katydids. This was at 23:30. After the chorus wound down and fe ll silent, a single pupa began chirping and continued, alone for several minutes. When the pupae from the second example above were moved to the hearth, a few began to chirp and were joined by some other pupae of both M. g. sweadneri and M. g. gryneus that had already been on the hearth for days. Occasionally, when 70 to 126 eclosion c ups were lined up on the hearth as in Figure 2-26, an individual would chirp, sometimes joined by three to five others; then the next day, a small group of new adults eclosed. In these episodes I was not able to record exactly which pupae sounded off to compare with which new butterflie s emerged so I have not been able to document that relationship. When one pink wanderer was inadvert ently placed in a Solo cup that already held a pupa two different sounds were heard. First, as the wanderer crawled, it made a sort of grinding sound, like that made when a child puts playing cards in bicycle spokes and turns the wheel. The sound rises and falls rhythmically. When the wander er encountered the pupa and began to

PAGE 94

78 climb over it, the pupa started chirpi ng. As the wanderer crawled on, the chirping stopped until the caterpillar encountered the pupa again on the next round of the bo ttom of the cup. In some of the anecdotes above, the pupae were stimulated to sound off by touch. In some, the stimulus seems to be vibration traveling through the plastic cups; in others, the pupae initiate sounds on their own; and in others, pupae and/or larvae seem to stimulate s ound production in pupae. Larval sound production and its role in myrmecophilous ri odinid and lycaenid life histories has been carefully studied by Phil DeVrie s (1990, 1991a, 1991b, and DeVries, et al ., 1993), but DeVries does not mention the sounds made by pupae in the same context. As noted earlier, ther e are no known records of ant interactions with any of the Mitoura and the late-stage larvae la ck two out of three known Figure 2-26. Eclosion cups arrayed on the hearth Cups are close together but not touching.

PAGE 95

79 ant-associated organs. To date I have found no literature references that address pupal or larval sound production in these situations. However, conspecific acoustic communication has been studied in the Arched Hooktip Moth, Drepana arcuata Walker, where it is used specifically in resolving territorial disputes.The authors of this study speculated that larval signaling may be widespread in Lepidoptera (Yack et al. , 2001). Desiderata So, are the immature Mitoura communicating? Do the vibrations that are amplified as sounds in a plastic vial or cup travel through the duff beneath a cedar tree? Does a pupa chirp in response to a wanderer crawling over it, and if so, what does it mean to the wanderer? Do es a newly minted pupa call out to ask if it is alone? Do pupae respond to the chirping calls of each other? Do pupae perhaps communicate in order to coordinat e eclosion times as a way to improve chances of finding mates? These questions need investigation. The answers could reveal a degree of sophistic ated socialization unimagined in the Lepidoptera and prove basic to understandi ng the conservation biology of Mitoura g. sweadneri . The equipment specified by DeVries (1991b) and Yack et al . (2001) might be ideal for the investigat ion if they could be adapted for use in the field. Adults In Vitro Adults of both butterflies are quite similar in appearance. Figure 2-27 shows a living adult of each sex of M. g. sweadneri and M. g. gryneus . While specific alar characters are discussed in Chapter 4, it is important to note here that the

PAGE 96

80 green ground colors of the ventral wing surfaces are the same for both butterflies. Slight differences in t hese images are due to lighting when the Fujichrome originals were exposed and how these slides reproduced when digitalized. In freshly emer ged adults, the wings are a lighter, almost iridescent green which darkens slightly within six to eight hours. All the characters present in the Florida population are also expressed in individu als from Tennessee; the Figure 2-27. Imagoes of both butterflies A) female M. g. sweadneri , B) the male. C) and D) female and male M. g. gryneus . A, B, and C) shown 3 X, and D) 3.27 X life size. differences in subspecies are, for the most part, a matter of degree. These characters vary within broods in specif ic separate colonies. Males and females are usually indistinguishable in vent ral aspect, though the dorsals are easily separated by the prominent androconial patch in males which is sometimes

PAGE 97

81 visible as a faint textural outline near t he ventral forewing costal margin. Males usually hold the abdomen above the substrate, partially tucked between the hind wings with the venter parallel to their i nner margins, as in Figure 2-27 B. The abdomen of a gravid female hangs lower an d may be more obvious, as in Figure 2-27 A. Mitoura g. gryneus 01 G6E9 (Figure 2-27 C) was photographed during oviposition. Sex Ratio The sex ratio in captive-raised adults varies somewhat from brood to brood, with one sex comprising 42-58% of new adults. But overall, M. g. sweadneri are 49% male to 51% female, and M. g. gryneus 51% male to 49% female for all adults eclosed in captivity. This is remarkably close to the 50/50 benchmark which we would expect in a natural populat ion. Neither sex dominates the first emerging adults of either butterfly, so, in vitro , they do not exhibit protandry. And the delayed eclosion observed in M. g. gryneus is not favored by either sex. Neither sex shows a distinct difference in pupation time, and a 6-instar life history may produce either sex. Size In general, SweadnerÂ’s Hairstreak loo ks like a larger butterfly to human observers. However, when actual meas urements are compared in Table 2-12, the difference in mean forewing lengths ar e never as much as two millimeters. The mean size of M. g. sweadneri is larger than the maximum recorded size of M. g. gryneus in each of the categories compar ed. Notice that although the mean female length exceeds the mean male forewing in both butterflies by about

PAGE 98

82 Table 2-12. Forewing lengths, M. g. sweadneri and M. g. gryneus compared* FW mm FW mm M. g. sweadneri Sample: 65 68 range 13.3-15.7 12.7-16.2 mean 14.63 15.07 standard deviation n-1 0.53 0.76 M. g. gryneus Sample: 75 72 range 11.9-14.5 11.1-14.8 mean 12.93 13.30 standard deviation n-1 0.63 0.75 *Wing measurements were made with a Mitu toyo Digimatic electronic caliper. 0.4 mm, the female ranges have both the largest and smallest individuals. The StudentÂ’s T-test was used to compare the means of male wing lengths ( t = 13.786, P < 0.0001) and female wing lengths ( t = -11.242, P < 0.0001) which showed that the the longer forewing lengths of M. g. sweadneri are statistically significant differences between Tennessee and Florida butterflies for both male and female butterflies. Analysis was with SAS software. Measurements for these samples were made from living individuals and pinned specimens. The lengths given are fr om the base of the subcosta vein at the point where it emerges from the thorax to the apex of the forewing. These measurements should not be compared with those of authors who give total wing span because of the differences in how the wings of indivi dual specimens are arranged when spread and how the spec imen was handled when collected. None of the individuals in this study was pinched. Diet and Longevity None of the field guides or authoritativ e Lepidoptera references cited in the reference section gave detailed directions for feeding adults in captivity. Scott

PAGE 99

83 (1986) devotes over three pages (one is four plates of infrared photographs of flowers) to adult food choices and nutrition in the field, but only Holland (1940) in his description of the work of W. H. Edwards mentions a mix of honey and water to feed adults. When I embarked on exploring the life history of Mitoura g. sweadneri , T. C. Emmel and Jaret Daniels sugges ted different mixtures of honey and water, based on their successful reari ng projects with other butterflies, and Andrei Sourakov described holding a bu tterfly with forceps, and uncoiling the proboscis with a pin into a spoon of sugar mixed with water. I later found this method discussed and illustrated (New, 1997). The first season, I tried feeding adults dilute honey water using AndreiÂ’s method. This required catching and holding the struggling butterfly with forc eps at every feeding and then rinsing the butterflyÂ’s feet. Within a few days, t he wings were a tattered mess and after a week or so, the adult could no longer fly. My first innovation was to use a Q-tip swab as a nectary so that I could put t he butterfly on the stick instead of holding it. Soon, I punched holes in the walls of t he plastic cups to poke the cut ends of swabs through so that the butterflies need no t be removed from cups for feeding. The butterflies only sipped briefly of wa ter and honey at dilutions of 4:1 and 6:1 and did not live more than seven to nine days on such a sweet concentration. When the proportions were increased to 10 or 12:1, the butterf lies fed more and lived more than twice as long. Mitoura g. sweadneri stayed active for 15-17 days and some lived to the ripe old age of 23 days on the weaker dilution. Most adults died with swollen abdomens, some becom ing distorted and yellowish. When a bloated carcass was squeezed, the yellow li quid that leaked emitted the odor of

PAGE 100

84 mead. The honey had fermented. I wondered if the fermentation killed the butterflies, or if they had actually outlived the expected seven to nine days reported by Scott (1986). I wa s disturbed to discover that females often died with many eggs in their abdomens. A serendipitous event led to great ly improved nutrition and longevity in future Mitoura rearing. At the end of March 2 000, J. D. Turner called from Alabama to say M. g. gryneus were beginning to fly and I had an opportunity to collect stock to start a c aptive colony for mating ex periments. I planned to leave as soon as possible. He called again the next day saying he had already caught a few, and asked how to keep them until I arrived. I quickly described the double cups set-up with the Q-tip nectary and emphasized the importance of a weak honey-water dilution to feed the butterflies. When I arrived in Huntsville, I was shocked to see he had start ed them on fruit-punch-flavor ed Gatorade. J. D. used a hypodermic needle to moisten the Q-tip which greatly increased the ease of feeding a captive colony. The next four seasons, captive col onists were fed fruit-punch Gatorade and survival times jumped dramatically. Several M. g. sweadneri lived longer than 30 days, with maximums of 40, 46, and 47-day survivors. The M. g. gryneus colony sample (which was five ti mes the size of the total M. g. sweadneri colony) produced several 40+ day individuals with single records of 47, 54, and 60 days. Though there were some old males in both colonies, the record survivors were all female.

PAGE 101

85 By the end of the F2 generati on of 2000, I had abandoned the use of forceps for holding live adults except when measuring or marking them, or sometimes capturing escapees in the lab. In an effort to avoid wing damage and continual “capture stress”, I handled the forceps in a new way. I would replenish the Q-tip, touch the spade tip to the Gato rade, and then carefully offer it to the butterfly. The hairstreak would touch t he forceps with a fore tarsus, then step onto the tip and be slowly lifted up to the Q-tip, where it would usually step onto the stick for feeding. Female M. g. gryneus were the least wary and easiest to “train”, followed by gryneus males, then M. g. sweadneri females. The sweadneri males were the wariest and some took th ree days and several tries each day to train. At the beginni ng of the 2001 flight season, I st arted using an old sable hair watercolor paintbrush with a small di ameter, 23.5 cm wooden handle and a thin, tapered, split-away section at the end. This natural wood surface was more readily accepted as an elevator than steel spade-tip forceps. Most colonists of both hairstreaks were more docile in ev ery generation thereafter, which made mating experiments easier. That same summer, many captive butterflies showed signs of operant conditioning and began to respond to the removal of the tulle cover, or to the moment the Q-tip was refreshed, by uncoiling the proboscis wherever they were in the cup, and several became frequent self-feeders. When I could quickly squirt fresh Gatorade on each Q-tip for at least one feeding a day, it left more time for rearing larvae and al lowed me to maintain larger colonies. The remarkable ability of these butterflies to learn and remember this modified

PAGE 102

86 feeding behavior offers an opportunity to explore learning behavior in butterflies in future investigations. Fecundity Wild-caught females will begin to lay eggs by the third day in captivity. Newly eclosed virgin females will usually begin laying infertile eggs by the fifth or sixth day. Five different treatment s were tried to induce captive M. g. gryneus females to lay, or to increase the num ber of eggs laid. Qualities of light, photoperiod, and temperature were stimuli variables in each treatment. Oviposition cups were placed in a sunny west-facing window with partially closed blinds, providing sun and shade on each cup. They received at least six hours of continuous sunlight and temperatures up to 28.5°C. Cups placed in a north-facing window all day received bright light, but no direct sun and temperat ures at 25°-26.5°C. Cups placed 13 cm (5 in) below a pair of 20 W florescent shop lights on a time regimen of one hour on, 30 min o ff for 12 h a day. Temperature varied with normal indoor range of 25°-26°C. Cups placed 45 cm (18 in) below a 60 W incandescent bulb on a timer set for one hour on, 30 min off. Temperat ure varied from 25°-28°C during each light hour. Cups placed 122 cm (48 in) below a 250 W infrared bulb on a timer set for one hour on, one hour off for a total of six hours of infrared per day. Temperature climbed from 25° to 34°C during each light cycle. Results varied little in the first four treatments. Females started laying 5-6 eggs at first, then increased to a maximu m rate of 18-25 eggs per day before tapering off. The north window females nev er laid more than 18 eggs in a day. Females that had already began oviposition slowed to fewer than five per day under infrared light and many of the eggs fell from the cedar to the bottom of the cup in this treatment. These butterflies were agitated, wore away their wings on

PAGE 103

87 the cedar, and died with many eggs in their abdomens. M. g. sweadneri females were rotated into the safer treatments and achieved the same rates but smaller life-time egg totals than M. g. gryneus . In a sample of 22 M. g. sweadneri females, ten produced 70 or more eggs, five of those laid 92 or more, and the t op two producers laid 169 and 189 eggs. The hatch rate varied from about 89% to one female who laid 96 eggs in five days with a 99% hatching success. Out of a sample of 35 M. g. gryneus females, 25 produced 110 or more eggs, 17 of those laid 163 or more, nine laid 190+ and the top totals were 345, 375, 398, and 408 eggs. Interesti ngly, the top producer, 01 WG13, only lived 32 days and laid her whole clutch in 28 da ys, including three the day before she died. The female with the next highest fecundity, 01 WG13 , lived 30 days and laid 398 eggs over 25 days. T he longest lived butterfly, 01 WG7 , only laid 167 eggs over 50 of her record 60 days. A simple evaluation suggests that M. g. gryneus are more fecund than M. g. sweadneri , and while this may be so, there are extenuating circumstances. Two M. g. sweadneri , 02 SW30, and 02 SW31 laid five and three eggs, respectively, in 40 min starting at 16:46 in an outdoor, walk-in flight cage at the U. F. Endangered Species Lab during a mating experiment. The temperature ranged from 32.5°C in the shade to 34. 5°C in the sun inside the screened enclosure. If 02 SW30 had only maintained her ra te for six hours she would have laid 20% more than the best per diem observed in vitro . In addition to the

PAGE 104

88 lower temperature and lack of full insolati on, other indoor limit ing factors might be the relatively dry 50% RH and the effect of restricted movement in a 16 oz plastic cup. The Tennessee butterflies may have more easily adapted to these particular indoor-dictated conditions. Captive Copulation There is no mention in the available literature of breedi ng hairstreaks in captivity. Though there are species acc ounts in New (1993) that mention mass rearing in conservation efforts, and handpairing of Papilionidae is explained by New (1997) and others, no one specifically describes lycaenid copulation in captivity. Captive rearing usually begins wi th taking gravid females from the wild. This could be devastating to small populat ions of rare butterflies where one female’s brood could mean su ccess for the wild colony. Initially, I subjected unbred pairs of M. g. gryneus to the same treatments used to induce oviposition. The infrared treatment was modified at the suggestion of T. C. Emmel to increase the temperatur e to 38°C by lowering the lamp to 91.5 cm (36 in) above the breeding cups and cont inuing the same light regimen. None of the pairs mated in any treatment and the females began to lay infertile eggs. In a discussion with Andrei Sourakov, I learned that some of the butterflies he had raised only mated when freshly eclosed females were introduced to males, possibly due to pheromones the new fe males emitted. He also said pairs should be introduced to each other suddenly. So, as the next adults emerged, I segregated the sexes into different r ooms. My next experiment involved introducing new pairs into 1 x 1 x 2 ft cages supplied with cedar branches and placed in partial shade outside. The butte rflies were monitored at least hourly,

PAGE 105

89 fed three times a day and brought inside at dusk for several days. Again, the females only laid infertile eggs. I asked Robert K. Robbins of the Nati onal Museum of Natural History for a reference on captive breeding of lycaenids and he said that he did not know of anyone who had experimented with mating them. Furthermore, he said the problem was that males set up mating territo ries, a situation that would be very difficult to duplicate in the lab (pers. com.). The first successful captive breedi ng occurred when Jaret Daniels and I tried captive-raised M. g. sweadneri in a 6 x 6 x 8 ft outdoor walk-in flight cage at the Endangered Species Lab at U. F. t hat had been fitted with a 7.5 ft potted cedar tree. We released the butterflies at 09:15 and stayed with them until noon, when the midday heat and academic responsib ilities drove us from the cage. The hairstreaks seemed to ignore each other all morning, but the next day the single surviving female began to lay fertile eggs . In subsequent morning experiments, I never witnessed courting behavior or copul ation. The butterflies did not visit potted Spanish Needles, Bidens alba , provided for nectar, and they were captured, returned to separat e cups, and brought inside to avoid the midday heat which frequently proved lethal to the lab-raised butterflies. In searching field notes, I found that I had only seen M. g. sweadneri in copulo at 15:38, 16:10, and 17:27 in Florida. In Tennessee, J. D. Turner and I had observed M. g. gryneus in copulo at 14:30, 15:30, and 15:40, with a possible remembered fourth sighting at 13:30. Thes e times seem to corroborate ScottÂ’s

PAGE 106

90 observation that “Mating occurs only in the afternoon or early evening in many (Theclini) species” (Scott, 1986, p. 357). Protocols for a successful mating experiment began the afternoon before a mating day, when the ceiling, walls, and fl oor of the cage were thoroughly swept and the cedar tree rinsed with a hard stream of water to dislodge spiders and remove their webs. Just before the ex periment began, the cage and tree were once more policed for spiders, tree fr ogs, and anoles. Four feeding stations made from one-quarter lengthwise sections of Solo plastic cups were hung on the screen walls in shade and sun and eac h was fitted with a Q-tip nectary soaked with Gatorade. The specific conditions that wo rked for mating with both of these Mitoura evolved through multiple trials and errors. First, pupae were kept in individual eclosion cups. Each new adult was sexed on emergence, and segregated with others of its own sex to a specific room . Each subject was given a number with an Extra Fine Point Sharpie marker: male s on the left hindwing, females on the right. This allowed instant recognition of sexes and identification of individuals during flight cage sessions. Documentary photographs of in copulo pairs shot from one side showed which individuals were coupled, and shots from the other side showed the pair as they might appear in their habitat. The morning before an experiment, each butterfly was fed, then offered Gatorade again just before the trip to the flight cage. Males were released first and given ten minutes or so to sort themselves out and become acclimated to their flight cage surroundings. A minimum of three

PAGE 107

91 males worked, but five or six worked even be tter. Their first reac tion was to fly to the top of the cedar. As each new one was introduced the others flew out to whirl briefly in the air before settling again near the top of the tree. One to three males perched on limbs near the t op and extra males lit on the screen walls or top of the cage, usually on the northeast secti on of the cage, rarely on lower limbs. Females were then placed one at a time on the sticks of the fe eding stations to sip again if they desired and to bask befor e flying to the tree. As a female approached the cedar, one or more males fl ew out to meet her before settling again. Because of their rapid flight and near perfect camouflage on the cedar, flights of more than ten individuals were too difficult to monitor, so most experiments were limited to one or two fewer females than males flying at the same time. Courtship was a very rapid affair, us ually lasting less than ten seconds. After a female landed on an upper limb, s he would be joined by a male who lit nearby and usually jumped around br iefly before approaching her head-on actively fluttering his wings in a few shor t bursts of 1.5 s separated by intervals of one second or so. An unreceptive female might fly or simply step away. There was no observable response from a receptive female. If she did not leave, the male turned parallel to her, facing the same direction, and reached out with his abdomen waving like an elephantÂ’s trunk to clasp the tip of her abdomen. This was always the only observed contact. Soon after secure copula was established, the male carefully changed pos ition until he was in line behind her facing in the opposite direction. This pos ture was maintained for a few moments,

PAGE 108

92 after which the pair might shi ft position slightly but not usually as much as 90°. If left alone, they would usually stay in the same place. When disturbed, the female usually made a short flight to another branch, dragging her non-flying consort along. In a few instances, it was difficult to discern which individual of a disturbed pair actually flew. Copulation lasted from 41-47 min in most cases. The longest record was 54 min by a pair of M. g. gryneus . Once a couple had been united for a while, they could be moved into t he labeled cup of the female and another female released into the cage. No first-day males were ever flown in the cage. Males from a few days to over three weeks old mated successfully . Freshness was not a prerequisite for females. Males were equally attracted to new females who had never laid an egg or to females several days old who had already laid 50 or more virgin eggs. Though the male was the observably active suitor, females always selected their mates. Some males never courted and some females never mated. No previously mated female was ever courted. Optimal time to begin a mating cage session was between 15:30 and 16:30 when the temperature in direct sun in side the cage was between 28° and 37°C. Most captive copulations started between 16:00 and 17:20. The earliest was on a cloudy day about 14:27 and the latest starte d at 18:10. They might have mated even later, but once the sun slipped below t he tall pines in the U. F. Natural Area Teaching Lab, shade fell on the flight cage and the green butterflies became quiescent and difficult to find in the cedar tree. Sessions begun earlier in the afternoon resulted in higher mortality among the participants. Survival was

PAGE 109

93 greatly increased when the hairstrea ks fed during the hot afternoon. Occasionally, a butterfly would find a feeding station on its own, but usually I tried to feed each participant once every 40 mi n or so. This was easily accomplished with hairstreaks accustomed to the wooden paintbrush handle; they would step onto it to be transferred to a freshly r eplenished Q-tip. On a few hot afternoons a butterfly would alight on my face to si p perspiration. This presented a difficult challenge to remain focused on the action at hand, take notes, and resist the temptation to wipe the tickler away. Each butterfly was fed when returned to its cup and again when they were unpacked at home. Larval Host The recognized host for Mitoura g. sweadneri is Southern Red Cedar Juniperus silicicola (Small) L. H. Bailey and M. g. gryneus uses Eastern Red Cedar, Juniperus virginiana Linnaeus (Glassberg, 1999, Glassberg et al ., 2000, Johnson, 1978, 1980, Scott, 1986, and Scudder, 1889a). These very closely related trees are gymnosperms in the family Cupressaceae Bartlett along with Redwood, Sequoia sempervirens (D. Don) Endlicher, and Baldcypress, Taxodium distichum (Linnaeus) Richard. The family comprises 25-30 genera with 110-130 species worldwide of which nine genera and 30 species are represented in North America. Most Cupressac eae are monoecious and the majority of genera are monotypic. Juniperus species are usually dioecious, wind pollinated, and all produce berrylike seed cones whose s eeds are dispersed by birds. This is the most species-rich genus with a wide, almost continuous distribution across the Northern Hemisphere (Watson and Eckenwalder, 1993). (As gymnosperms the trees themselves (spor ophytes) are not sexual but produce either male or

PAGE 110

94 female reproductive structures (gametophy tes). They do not make “flowers” in the strict sense of the word. Since the j unipers in my study are dioecious, I call them “male” or “female” trees throughout this paper and use the term “flower” for their reproductive structures for ease of reference.) Specific Taxonomy Linnaeus described Juniperus virginiana in 1753. Though its name literally means ”juniper of Virginia,” it referred to the upright, dominant red cedar of what is now the eastern United States mainla nd and set it apart from the European junipers. Sargent (1902) recognized J. virginiana as the eastern red cedar but identified the cedar found on the Atlantic coast of Georgia and Florida as J. barbadensis Linnaeus from Barbados, which also occurred in the Bahamas, Dominican Republic, montane Jamaica, and Antigua. A 175 years after Linnaeus, following a 1920 explor atory expedition to Florida, J. K. Small (1923) described Sabina silicicola as a distinct cedar species based on leaves that were thicker than those of S. virginiana and with a more blunt apex. He also distinguished this southern tree from the more northern one by the larger staminate ament, or male flower (akin to catkin ), 4-5 mm long and the smaller ovoid or ellipsoid-ovoid cone 3-4 mm l ong compared to the male flower of S. virginiana at 3-4 mm and its sub-globose cone at 5-6 mm long. The habitat and range of the new species was given as sand and clay soil in the coastal plain from Texas to Florida to South Carolina. The genus Sabina Haller was used at the time to distinguish c edar trees from junipers. Sabina is Latin for a kind of juniper also called savins and the s pecific name is from the Latin silic for flint, or siliceus meaning “of flint or limestone” with Latin colo meaning “to inhabit”

PAGE 111

95 (Jaeger, 1955). So, Small named this tree “cedar that lives on flint or limestone.” The Latin name is thus descriptive of its typical habit on calcareous substrate including (but not limited to) hammo cks, coastal beaches, and shell mounds (Wunderlin and Hausen, 2003). Correll and Johnston (1970) added to an evolving taxonomic confusion when they separated the species based on the branches being mostly pendulous in J. silicicola and horizontal in J. virginiana , then compared the female cones at 3-5 mm and 5-8 mm long, respectively. They dropped Small’s difference in fruit shape, kept his leaf differences, and said the ultimate twigs were usually less than 1 mm thick versus more than 1 mm thick. Little (1971) drew range maps, combined in Figure 2-28, for the United St ates Department of Agriculture which showed a break in the distribut ion of these species. Murr ay (1983) was the first to list J. silicicola as a subspecies, and Silba (1984) was the first to call J. silicicola a variety of J. virginiana , but neither of them publ ished data supporting their taxonomic decisions. In the popular Peterson field guide series, J. virginiana is said to have berries about 1/4 in in diam eter and live in old fields and dry soils compared to J. silicicola with fruits 1/8-3/16 in in diameter, more slender twigs, and inhabiting moist sites in the south (Petrides, 1988). In the most current, thorough revision of Juniperus in North America (Adams, 1993), J. silicicola is treated as a variety of J. virginiana . The taxonomic decisions are based on slight differences in the sizes of the pollen cones 4-5 mm and 3-4 mm long, seed cones 3-4 mm and 4-6(-7) mm long, seeds 1.5-3 mm and 2-4 mm long of J. v. var. silicicola and J. v. var. virginiana , respectively, plus the

PAGE 112

96 Figure 2-28. Ranges of the two larval hosts Drawn after maps 29-E and 31-E from Little, (1971). leaf apex shape, bluntly obtuse to acute for the former and acute in the latter.To the afore-mentioned characters, Adam s added a tree height to 10 m for J. v. var. silicicola , and to 30 m for J. v . var. virginiana . He notes that eastern red cedar hybridizes with J. horizontalis Moench and J. scopulorum Sargent but not with J. ashei J. Buchholz as some authors have suggested. He said the “southern variety of Juniperus virginiana appears to be restricted to coastal foredunes (and coastal river sandbanks) but differs little in morphology or leaf terpenoids from

PAGE 113

97 the upland J. virginiana and appears to intergrade with that variety in Georgia” (Adams, 1993, p. 418). His taxonomic decisions are based on multivariate analyses of morphology and terpenoids in the two trees and his determination that the characters attributed to J. virginiana var. silicicola are circumscribed within the range of va riation found within J. virginiana (Adams, 1986). In that same paper, the disjunct popul ations of southern Loui siana and Texas indicated in Little’s maps, Figure 2-28, were found to be J. virginiana . Furthermore, he acknowledged that the differences in the shapes of the crowns of trees growing near the coast may be the result of wind and salt spray damage rather than genetic-based characters. He also thought the patterns of variation could not definitively demonstrate gene flow and were suggestive of introgression. (Adams also pointed out that Little’s maps are of regions where the tr ee’s characters are distinctly expressed and do not indicate a Juniperus -gap between the ranges of these forms.) The most recent publications found no discrete disconti nuity between the two taxa (Wunderlin and Hausen, 2000) and tr eated them as conspecific, listing J. silicicola as a synonym of J. virginiana and referring to the tree as Red Cedar and dropping the “eastern” and “southern” qua lifiers in previous common names (Wunderlin and Hausen, 2003). Since trends in systematics and taxonomy come and go over time, I think that in order to preserve valuable historic data, the silicicola specific should be maintained as Juniperus virginiana var. silicicola , as Robert Adams (1986) suggested. Throughout this paper I use J. silicicola to refer

PAGE 114

98 to the cedars I found in M. g. sweadneri habitat whether in naturally occurring groves or planted shelterbelts throughout north-central Florida. Larval Host Specificity Host plant specificity is seen in most Mitoura species from the United States and Mexico where host plant data ar e known (Johnson, 1978, 1980, Brown, 1982, and Nice and Shapiro, 2001). It was suggested as a possible taxonomic character (Downey, 1962) and was used extensively and effectively as such along with morphological characters in Johnson’s revision of the Callophryina (Johnson, 1976a, 1978, 1980). In addition, Ni ce and Shapiro (2001) suggest that divergence among species M. nelsoni (Boisduval), M. muiri (Hy. Edwards), and M. siva (W. H. Edwards) is being driven by evolution of their use of different Cupressaceae host plants. They conclue ded that restricted gene flow between M. nelsoni and M. muiri may result partially from t he preference of the adults for mating on different host trees. This specificity does not, however , mean host exclusivity in all Mitoura species. Mitoura siva , and M. gryneus are two species that feed on different cupressaceous hosts across their ranges (Johnson, 1978). Review Chapter one, Table 1, four Juniperus species and some hybrid hosts are listed for M. gryneus and nine Juniperus hosts are known for M. siva . Furthermore, these two hairstreaks have wider ranges than all t he others mapped in Figure 1. Johnson stated that regional host specificity is maintained in these hairstreaks by the female’s oviposition preference for the hos t that she was raised on, according to the “Hopkins’ Host Principle” (Johnson, 1978). In perhaps t he earliest published experiment in host exclusivity in Mitoura , Remington and Pease (1955) raised

PAGE 115

99 small numbers of M. gryneus larvae (from eggs of a single female captured on Juniperus virginiana ) on J. virginiana and on Chamaecyparis thyoides (Linnaeus), Atlantic White-Cedar, the natural host of Mitoura hesseli . Not only did these larvae grow to maturity, the C. thyoides -fed adults mated and produced viable offspring, demonstrating that the white cedar host did not result in nutritional sterility of the adult butterflies. Sixteen years later, Eric Quinter at the American Museum of Natural History in New York successfully raised M. hesseli on J. virginiana but did not evaluate the nutritiona l sterility of the adults (Johnson, 1983). When I first handled cut J . virginiana and J. silicicola , I detected a slight difference in the odor of the two cedars. J. virginiana smelled a little stronger and more “spicy” than J. silicicola , which smelled more like a cedar chest or pencils remembered from childhood. Though this was not mentioned in Adams’ (1986) analysis of the specific terpenoids, I w ondered if the butterflie s could distinguish the difference. I had hoped to dem onstrate host specificity of J. v. silicicola for M. g. sweadneri , or lack of suitability of J. silicicola as host for M. g. gryneus , with my captive colonies, thinking this would strengthen the case for Florida endemism of M. g. sweadneri . During the first year that I raised the Tennessee Mitoura , I tried a single individual on J. silicicola and the first larva I observed took seven instars to reach pupation. Exci ted by this extra instar in the life history, I next tried moving eight M. g. gryneus hatchlings from J. virginiana onto J. silicicola , which yielded mixed results when fi ve of them would not feed but three survived to fly. The next season I found that females of either species

PAGE 116

100 would lay eggs on whichever cedar was placed in their cups. By alternating cedar sprigs each day in oviposition cups, I was able to get the hatchlings of each hairstreak to feed to maturity on whic hever cedar its egg was laid on. The resulting offspring proved fertile when mated within their own colonies. This demonstrated the lack of nutriti onal sterility in resultant o ffspring. One test I did not try was to give females of both butterflies a choice of cedar which might have revealed an oviposition preference in eit her species. No further experiment was done to determine food plant choices between the two cedars of already feeding larvae. However, when single larvae were presented with both male and female sprigs of their cedar host, they chose fema le sprigs over male sprigs more than 70% of the time, even if the female hos t had been cut longer and was drier than the male cedar. Though none of the cit ed cedar taxonomy papers mentioned sexual differences other than male flower s or female seed cones in either cedar, the female sprigs usually seemed to have longer and plumper, light green, growing tips than the male trees. This m eant that a feeding larva could eat more of the new softer growth on a female spri g before it encountered the transition to the dark green older growth on a stem and abandoned it for the new tip of the next available stem. Anthropocentric Utility of Cedar Junipers have volatile oils whic h contain terpenoids in high enough concentrations that all parts of the pl ants may be considered toxic (Foster and Duke, 1990). These aromatic compounds and other physical properties of the wood have made cedars an important staple of our ow n life history and ethnobotany for thousands of years. Cedar, cedar tree s, or cedar wood are

PAGE 117

101 mentioned 15 times in the King James translation of the Old Testament and cedar as incense has been a part of spirit ual practice and purification ritual in cultures from Asia to the Native Am ericans (Foster and Duke, 1990, Moerman, 1998). Medicinal uses Juniper berries have long been used as spice, to flavor tea and sloe gin, and in herbal apothecary, but the mo st diverse medicinal use of J. virginiana and J. silicicola has been documented from Native American peoples. Teas, infusions, and external applications of leaves or fruit have been used as an analgesic for stiff neck, backache, and heada che, and to treat coughs, colds, and to reduce fever. Young twigs have been used as a diuretic and as internal and external treatment of rheumatism. The berries have been used to expel internal parasites from the gut, to induce persp iration, and to induce or increase menstrual flow. The oil is an abortifacient and is known to have also caused vomiting, convulsions and even coma and death (Foster and Duke, 1990, Moerman, 1998, Petrides, 1988, and Weiner, 1991). Juniperus virginiana is said to have podophyllotoxin, an antitumor compound (Foster and Duke, 1990). Cedar oil, sometimes called cedar camphor, has found commercial application in soap, deodorant, perfume, insecticide, and moth repellant (Weimer, 1991) and more recently in aromatherapy. Other uses Horticulturists have traditionally valued Red Cedar trees planted as shelterbelt for wind protection, to reduc e erosion, and provide privacy screen in

PAGE 118

102 landscaping. Alone or in pairs, they ma ke striking ornamental specimen trees and are also grown for potted or cu t Christmas trees in Florida. The volatile oils make cedar an ex cellent though quick-burning fuel. As a Cub Scout, my flint-and-steel kit had a bi g wad of finely shredded cedar bark as tender. In earlier times the bark was used to soften cradles and in weaving floor mats (Moerman, 1998). The tree’s te rpenoids give the heartwood insect resistance, so it has long been used for f encing. The wood is straight-grained, durable, fragrant, and easily worked, so it is highly sought -after for cabinet making, particularly for cedar chests used to store winter clot hes for protection from moths (Petrides, 1988), but it has also been used for closet linings, flooring, lining saunas, and manufacturing scientif ic instruments. The red heartwood contrasts sharply with the white outer layers and ma kes attractive paneling. The greatest single commercial use of red cedar has been for making the casings of lead pencils. Exploitation for the pencil industry vi rtually eliminated Southern Red Cedar as an overstory s pecies in Florida by the late 19th century and it has never recovered (Ewel, 1990). Small (1933, p. 11) said of Southern Red Cedar “The natural supply has been exhausted through the manufacture of pencil wood.” Almost all of the forest in Florida is se condary or later growth. By 1850 most of the pre-European old growth Longleaf Pine, Pinus palustris , and other commercially valuable trees had been cut over. Though present-day cedar groves can be quite impressive, it takes some imagination to grasp what an oldgrowth cedar break must have been. Fort unately, a single specimen, the largest living Southern Red Cedar, grows near Archer in Alachua County, Florida. This

PAGE 119

103 record tree measures 158 cm (~5.2 ft) in diameter, 23 m (75.7 ft) tall, and has a crown spread of 16 m (52.5 ft) (Ward, 1997). It is esti mated to be 300-500 years old. One of the large branches is hollo w at the base and has a resident raccoon (personal observation). Florida Cedar Today The original description of Sabina silicicola (Small. 1923) said it was found near the edges of marshes and swamps in Florida. Juniperus silicicola is salt-tolerant and often found in coastal hydric hammocks near the boundary of brackish and freshwater wetlands and in hammocks on the banks of the St Johns River (Ewel, 1990). Zanoni (1978) said it ranged through the river swamps of the southeastern U. S. coastal plain, and a ccording to Adams (1993), it seemed to be restricted to coastal foredunes. Actually, the tree is found in a wide range of forest cover types (Eyre, 1980) and soils from acid to alkaline across Florida. It is adapted to varying degrees of moisture, though it must be well drained and can be very drought-tolerant when establis hed (Elias, 1980). It is easily damaged or killed by fire and modern-day fire suppression has helped J. silicicola reclaim interior sites in its natural historic range (Van Haverbeke and Read, 1976). Fire intolerance may be why it seems to have a wet reputation from earlier eras. Its current range is a little further south than illustrated in Little’s map, Figure 2-28, and a little north of the sout hern reach of U. S. Departm ent of Agriculture Plant Hardiness Zone 9 where average coldest te mperatures range from -7° to -1°C (20° to 30°F). Though not expressly st ated in the literat ure, these cold temperatures may define the tree’s s outhern boundary limit because the seed’s

PAGE 120

104 dormant embryos may wait for three y ears until cold stratification triggers germination (United States , Forest Service, 1974). The widespread occurrence of J. silicicola today may be attributed to two principal agents. Firs t, the fruits are known to be eaten by more than 50 bird species, opossum and, to a lesser ex tent, other wildlife (Petrides, 1988). The flowers of late winter and spring produce ripe fruit in the fall in time for migrating birds headed south. Seed stratification by passage through the gut of an animal may improve germination, just as soaking the seeds in citric acid before cold stratification increases the germination ra te beyond that of cold stratification alone (United States, Forest Service, 1974) . Bird dispersal could easily account for cedars that occur in old fields, along fence rows, and in edge habitat along margins of pine plantations , and around natural gaps in forest cover created by blow down. If southward-migrating bird s have been distributing cedar seeds from points north along their flyway for millennia, this could also explain the close similarity between J. virginiana and its southern variation, silicicola . Though man has long planted cedar on farms and near homes for uses detailed earlier, the se cond major agent spreading J. silicicola in the last 50 years has been the State of Flori da Department of Agriculture and Consumer Services and its Division of Forestry. In the y ears after the Second World War, the foresters of several southern states start ed growing pine trees for distribution to their constituents in an effort to stop erosion and stabilize soils. The Florida Division of Forestry supplied pines to i ndividuals and sold seedlings to timber interests to replant harvested stands. In 1955 with the est ablishment of the

PAGE 121

105 Division of ForestryÂ’s Andrews Nursery at Chiefland in Levy Co., they began to grow several other native trees incl uding Dogwood, Redbud, Baldcypress, Red Cedar and several hardwood species. Thes e seedlings were provided to 4H, FFA, garden clubs, Boy Scouts, and other organizations for Arbor Day planting programs and sold in lots of a thousand to landowners and commercial nurseries throughout Florida for the ornamental tr ade. In an effort to promote native ornamentals in landscaping, foresters packaged five seedlings of assorted species and sold them for pennies from the backs of state pickup trucks in grocery store parking lots (Frank Cone, pers. com.). Soon the Andrews Nursery was joined by two other Division of Fore stry pine nurseries, Munson Nursery at Milton, in Santa Rosa Co., near Pensacol a, and Herren Nursery near Lake Placid in Highlands Co. in growing non-pine species. Red Cedars were part of the first non-pine planting in 1955 and this first crop was distributed in 1956-1957. From 1955 to 1960, an employee of Andrews Nursery, Dale Donmore, drove to Se lma, Alabama every year to purchase Southern Red Cedar seeds (Steven P. Gilly, pers. com.) because Florida foresters did not have their own establis hed local cedar seed source until 1961 (Frank Cone, pers. com.). This influx of cedar from well north of the line on LittleÂ’s map (Figure 2-28), delin eating the southern boundary of J. virginiana explains why my own attempts to identify J. silicicola in Florida, based on characters described above ended in conf usion. Indeed, meas urements of cedar berries and observations of leaf shapes from cedars in several north central

PAGE 122

106 Florida locations seemed to directly cont radict the expected characters from the literature. One might wonder what e ffect nursery-grown cedars could have on the statewide distribution of this tree un til the sheer magnitude of the native ornamental effort is put into perspecti ve. From the first year, Red Cedar was always part of the annual crop and each year several hundred thousand cedar seedlings were sold. By 1959, Division of Forestry nurseries had produced their first billion trees, including all pine and non-pine species (Don We st, pers. com.). Table 2-13. Total Red Cedar seedlings sold in the last five years of production* Year sown Year sold Total seedlings 1985 1986-87 834,425 1986 1987-88 859,050 1987 1988-89 640,000 1988 1989-90 244,000 1989 1990-91 452,000 1989 1991-92 63,000 *Records courtesy of S. P. Gilly, Division of Forestry , Manager Andrews Nursery. Total numbers of Red Cedar seedlings di stributed during the last years of production are given in Table 2-13. Four times a year, each nursery delivered seedlings to all parts of Florida and R ed Cedar went to ev ery county from the Panhandle to south Florida as far sout h as Fort Lauderdale (Steven P. Gilly, pers. com.), well below the natural range of J. silicicola . The seedlings were very affordable at $20 per 1,000 in 1972 (Fr ank Cone, pers. com.), gradually increasing to $60 per 1,000 in 1987 and still only $125 per 1,000 in the last year of production (Steven P. Gilly, pers. com.). In the late 1980s the Division of Fore stry began to get complaints from commercial nurseries that the state agency was harming growers in the

PAGE 123

107 ornamental market by sell ing seedlings at non-compet itive prices. Ironically, many of these growers were cust omers who had purchased the stateÂ’s ridiculously low priced cedars in thous and lots. In 1990 Steve Gilly at Andrews Nursery got a phone call from Division of Forestry headquarters saying not to sow any cedars in October. The low numbers for the last seasons in Table 2-13 reflect selling off the last st ock. That last crop distri buted in the winter of 1991-92 marked the end of the production of native ornamental trees by the Division of Forestry. Since then, they have produc ed pines, Baldcypress, and Pondcypress exclusively. As late as 2004, Division of Forestry nurseries were still getting calls from customers seeking Red Cedar seed lings. Since private nurseries do not divulge their sales records, there is no concrete number available for Red Cedars sold since the state of Florida st opped producing them. Based on customer requests, the current estimate is that commercial nurseries are producing less than half the number of cedar seedlings grown by the Divis ion of Forestry in the late 1980s (Steven P. Gilly, pers. co m.). Between 1956 and 1992, an estimated minimum 22 million Red Cedars were sold by Florida Division of Forestry. These trees can still be seen today as wi ndbreaks along roads and highways, separating agricultural fields and pasture s, and marking property lines across Florida. They are also a familiar sight at the entrances to Division of Forestry offices and fire towers in every dist rict. Furthermore, their seeds have been disbursed by generations of birds and, together with the human planted trees, provide habitat for Mitoura g. sweadneri across its current range.

PAGE 124

108 The source of information about the fore stry divisionÂ’s 37-year history of growing native ornamentals was personal in terviews with foresters who lived it. Mr. Frank Cone joined the Florida Divis ion of Forestry in 1961 and was the Forest Area Supervisor at the Etoniah Creek district office at Hollister in Putnam Co. for many years before he retired. His home was on land just east of and adjacent to the Etoniah Creek office and the cedars in his yard and on the district office property next door comprised the best M. g. sweadneri colony in the state from 9 March 2000 until 15 October 2003. Steven P. Gilly has been the manager of the Andrews Nursery since 1987, and Don West is the district manager of Florida Division of ForestryÂ’s Waccasassa Forestry Center in Gainesville. All three men said there were no official Divis ion of Forestry publications or written directives concerning growing Red Cedar seedlings and, in fact, there was no official program to promote the cultur e of native ornamental s by the state of Florida. It was simply part of their job. Don West suggested that the only records of numbers grown during the 37-year history might be re corded in the minutes of the Forestry Council meetings. Life History Field Study: Adults in Situ In my initial search for a suitable subj ect for dissertation research in 1998, I made several field trips to a cedar habita t site in rural Dixie Co. near Jena, recommended independently, by Hugo Kons, Jr . and Jaret Daniels. I also visited a popular, well publicized cedar grove behind the Lions Club on 59th St. in Yankee Town, Levy Co. (Calhoun, 1996, and Glassberg et al ., 2000), and cedars in St. Augustine Lighthouse Park, St. Johns Co., the assumed type locality of Mitoura sweadneri (Chermock, 1944). As I drove through rural areas along the

PAGE 125

109 Gulf coast of Florida and tr aversed the state to the At lantic coast, I stopped at natural cedar groves, planted cedar rows , and individual trees to look for SweadnerÂ’s Hairstreak. I kept casual reco rds of the routes traveled and stops made. I had no preconceived ideas concerni ng habitat qualities or preferences of the butterfly. When practical, I hunted t he hairstreak using the method I learned in years of Xerces Society/North Am erican Butterfly Association annual 4th of July butterfly counts searching for Mitoura gryneus in New York: I thumped the trunks of cedar trees (Glass berg, 1993, Glassberg et al ., 2000, Opler and Krizek, 1984, Opler and Malikul, 1992, Scudder, 1889a, and Weed, 1926) with a heavy industrial mop handle fitted with the r ubber foot from a discarded crutch. Some of my initial observations were that there were many cedar sites with no M. g. sweadneri present. The few places where t he butterflies were found only harbored small numbers, even though there were often many cedar trees in close proximity. Cedar groves were oft en near-monocultures of trees of the same approximate size, with rarely any other tree species present except for an occasional pine usually towering above, and virtually no young cedar seedlings. Many cedars in groves and in plant ed shelterbelts, and individually grown specimens, shared a pyramidal symmetry modified somewhat by the nearness of other trees but recognizable as the classi c conical tree shape. However, trees in some Atlantic coast counties were some times gnarled and twisted into tortured shapes reminiscent of Japanese bonsai, and trees along driveways or in some planted rows were missing the lowest limbs often higher than six feet from the ground but SweadnerÂ’s Hairstreak might, or might not, be perched in any shaped

PAGE 126

110 cedar. Finally, cedars grew in several dry lo cations and a variety of wet situations along freshwater rivers, creeks and r oadside ditches, on the banks of ponds and flooded borrow pits and coquina pits, near brackish water, and on hammocks in high salt marsh. I wondered how proximit y to water affected butterfly habitat choice. I also wondered, if I took on this butterfly as a serious research subject, how I would ever be able to thump some of these cedars which frequently grew higher than the 3.7-6. 1 m (12-20 ft) trees IÂ’d most of ten searched in old fields in Westchester Co., New York and on Long Island. From t hese problems and biological puzzles, I began to form queries ab out the butterflyÂ’s life in its habitat that might be answered by d iligent observation in the fi eld which could illuminate essential elements necessary for its conservation. Here are some of the questions I sought to answer by direct observation of the butterfly in situ . What specific qualities or conditions constitute habitat for Mitoura g. sweadneri ? How do the butterflies use the habitat? When is the flight season? How many broods are produced each year? Are there regular peaks of abundance? Are there observable limiting factors like weather, availability of nectar sources, or predators t hat affect the population? Are there specific human activities t hat affect numbers of individuals in a colony? Are numbers of colonies or colony viability affected by specific human activities?

PAGE 127

111 Are publicly held lands providing mo re protected or better managed habitat for this butterfly than lands in private ownership? What is the current range of Mitoura g. sweadneri ? Should this butterfly be considered for conservation concern? Are there other species that coul d benefit from conservation measures, such as protecting habitat, for M. g. sweadneri ? The Literature Several authors give basic habitat requirements for these Mitoura (Clench, 1961, Daniels, 2003, Emmel, 1993, Glassberg et al ., 2000, Opler and Krizek, 1984, Opler and Malikul, 1992, Pyle, 1981, Scudder, 1889a, and Scott, 1986) as Red Cedar, or Southern Red Cedar trees , some mention nectar sources, but there are several different speculations about the number of broods and flight times. Almost all say that males perch on the host tree to await females. Clench (1961) suggested that M. gryneus prefers smaller, or stunted trees, and flies around the tops perching on branches, wh ile Johnson and Borgo (1976) in their study of perching behavior of M. siva and M. gryneus determined that the butterflies preferred taller trees for perch ing. The only recent, well researched, authoritative description of the range of the butterfly in Florida is by Calhoun (1996) and only two authors suggest M. g. sweadneri as a species of conservation concern (Daniels, 2003, and Emmel, 1993). I hoped to sort through this information and determine first-hand specific habitat requirements and specific behaviors of M. g. sweadneri that could lead to understanding the conservation biology of this butterfly. I have only seen M. g. gryneus flying wild five times; three were on trips to Tenne ssee to collect the nominate subspecies for captive colonies and two other time s were trips to the Pine Barrens near

PAGE 128

112 Lakehurst, New Jersey in the early 1990s to see M. gryneus and M. hesseli . I have not attempted to compar e specific behaviors of M. g. gryneus with those of M. g. sweadneri since the scope of this study is the conservation biology of the Florida butterfly Methods and Materials of Life History Field Study At least two of the sources ci ted above had not been published in 1999 when I began my field study in earnest. I had never met John Calhoun and was not aware of his interest in Sweadner ’s Hairstreak nor had I seen the Johnson and Borgo study of Mitoura perching behavior. I had no particular field study model to emulate or to suggest what data mi ght be useful in evaluating habitat or colony success. Data collected Before my first field trip, 6 February 1999, I decided to record the location of each cedar site I explored using a Ga rmin GPS 12 handheld global positioning satellite receiver from Garmin Internat ional, Olathe, Kansas and to mark each locale with M. g. sweadneri present as an electronically recorded three digit waypoint (WPT) in the receiver. Lat itude and longitude notation chosen was degrees, minutes, and decimal minutes to three places. I also recorded traditional location information, including county, nearest town, and approximate road/highway distances. The Florida Atlas & Gazetteer by DeLorme of Yarmouth, Maine was helpful in determining nearest towns and in navigating back roads in rural Florida. Additional abiotic data collected incl uded the observational time in and out at each locale, noted as 24-hour time. Temp erature in °F and moisture content of

PAGE 129

113 the air as relative humidity (%RH) were measured using a sling psychrometer model 12-7011 by Bacharach Instrument Company, Pittsburgh, Pennsylvania. Wind velocity was measured as miles per hour with a Dwyer wind meter by Dwyer Instruments, Inc. of Michigan City, Indiana. This instrument was held with the bottom at approximately 1.8 m (5 ft 11 in) from the ground. Wind direction and all other directional data were det ermined with a Silva backpacker compass type 27 by Silva of Sollent una, Sweden. Sun direction was noted relative to the observed zenith. A general site descripti on was included on the first page of each new entry, and general weather notes we re recorded for every site visit. Recording biotic information about t he butterflies and their host trees began with finding the butterflies. Thumping cedar trees using a heavy stick, battering ram fashion, to deliver a blow to the trunk works well for putting hairstreaks to flight. Apparently, these arboreal butte rflies have adapted to hang on to the trees in all sorts of wind and rain conditions , including storm winds common to our region. However, nothing in their long ev olution with the host has prepared them for holding on in the event of a sharp jolt. A solid thump to the trunk four to five feet up from the base of a cedar up to about 17.78 cm (7 in) in diameter sends a shudder out to the end of every branch; t he butterflies fall off their perches and must fly to regain them. Several cedars surveyed in north central Florida were larger, some as big as 49.3-55 cm (19.421.7 in) in diameter and little affected by thumping. To effectively jar these bigger tr ees, or any smaller cedar in a grove, I chose a telescoping fiberglass and alum inum painterÂ’s extension pole by Mr. Longarm International, Noordwijk, Netherland s. With this pole fully extended to

PAGE 130

114 5.3 m (17.75 ft), I could whack a branch 7. 24 m (23.75 ft) high st rongly enough to dislodge butterflies at twic e that height. My technique for shaking the tallest perches was to bang into a fork at the end of a larger limb so that the force was delivered directly toward the trunk. At 1. 36 kg (3 lb), the pole was light enough to handle effectively in the sometimes close quarters of a cedar grove. The end of the pole designed to receive a paint roller handle was modified to accept the threaded female end of a BioQuip Tropics Net and up to two handle extensions. Fully deployed, this gave me a somew hat unwieldy reach of 9.6 m (31.5 ft), though without the extensions, I could stil l reach butterflies at a respectable 8.3 m (27.25 ft) with much greater accura cy. The butt end of the pole was fitted with a large rubber foot from a child-siz e tempera easel to be used for thumping smaller trees. Fortunately, I rarely needed to capture the hairstreaks except to note wear of individuals, positively i dentify other species, or to provide progenitors for my captive colony. Mitoura g. sweadneri were easily identified using 10 X 42 mm binoculars by CabelaÂ’s of Sidney, Nebraska. This instrument was chosen for its bright field of vi ew, good for picking out a green butterfly against a green cedar, deep rubber eye cups that fold down so they can rest against my glasses, and especially for it s ability to focus down to 150 cm (59 in) from the subject which allows for intima te observation of butterfly behavior at close range. Though the binoculars weigh only 0.82 kg (29 oz ), the neck strap was replaced with a wide neoprene one by Op/Tech of Belgrade, Montana to make them comfortable in the fi eld for several hours at a time.

PAGE 131

115 Every tree found with a butterfly pres ent was marked with a streamer of lime green flagging tape and given a unique num ber, including the three digit site waypoint, date of original marking, and tr ee number of that day/site. This number was also written on the str eamer with a Sharpie perman ent marker for quick tree recognition on subsequent visits. The ID st reamers were originally attached to a branch end at shoulder level until late in that first season when the green of the tape proved fugitive in the intense Flor ida sun. The Sharpie notation was still legible so each streamer was moved in to the shade of the tree, usually tied around the trunk and each tape was replaced each season. Hiding the streamer within the tree also attracted less hum an attention and reduced vandalism. The circumference at breast height, 137 cm (~54 in), was measured with a vinyl coated dressmakerÂ’s tape for each butterf ly tree and its height was estimated by comparing it with the painterÂ’s pol e which had been marked in one-meter increments. The sex of eac h tree was noted when discerni ble and its total width, as spread of lowest limbs, was m easured with a 30.4 m (100 ft) Ultra-glass Premium double graduated fiberglass t ape on a reel, model OTR18M100, by Keson Industries, Inc. of Naperville, Illinois. The di stance and direction to the nearest tree were recorded and the gener al shape of each occupied tree was drawn. Nearest possible nectar sources were noted when present. When a tree is tapped, SweadnerÂ’s Hair streak almost always flies out and back in a tight U-shaped path. The original position of the butterfly was recorded as estimated height from the ground, or distance down from the top, and the side of the tree by compass direction. In some instances, perches of up to four

PAGE 132

116 butterflies were recorded for a single tree. In itially, all butterflies were recorded as “sighted” and sex was only recorded when the individual was netted. After a season of field trips and many hours of observation in the field, I learned to determine the sex of an individual by its posture and behavior. Behaviors such as whirling, copulation, or oviposition were noted and sometimes led to long periods of observation of an individual or small group of Sweadner’s Hairstreaks. Other butterfly species seen in and around c edar sites and nectar sources were recorded, with particular attention to butte rflies that actually landed on, perched in, or were flushed from the cedars. Pr esence or absence of some other insect orders were noted, specifically Odonata, Diptera, and Hymenoptera when they seemed remarkable. Sightings of other ani mals such as nesting birds, snakes, gopher tortoises and mammals were also not ed. All of these observations were recorded on a field data sheet, Figure 229, designed with ample space for butterfly sightings in three cedars, notes on behavior, and dra wings of the trees at initial discovery. At the top of the sheet is a line for the names of field assistants. Effort My original field research plan was to locate one active Mitoura g. sweadneri colony on the Gulf and one on the At lantic coast of Florida, then search for additional sites with the butte rfly present north and south along each coast in an effort to determine their pres ent-day range. By returning to a known site on each succeeding field trip and c ontinuing the search from there, I reasoned I would discover the flight ti mes of each brood and not mistake cedars explored between broods for uninhabited sites. Between 6 February 1999 and

PAGE 133

117 Figure 2-29. Field data sheet at 83% size

PAGE 134

118 8 May 2004, I made 120 field trips (only seven were in 2004), of which 51 were with an assistant, 69 were alone, and logg ed 286 visits to 90 recorded cedar sites. Over 2,780 individual trees were in spected in parts of 18 Florida counties comprising four Florida wate r management districts. The search included land in one national forest, nine state parks, two ci ty or county parks, five state wildlife management areas, three nati onal wildlife refuges, ar ound several boat ramps, and the grounds of two state museums. I routinely trespassed on timber company holdings and private hunt cl ubs, and whenever possible, sought permission from farmers, ranchers, and pr ivate landowners to inspect cedars along their borders and driveways. All natural cedar groves and planted shelterbelts were treated as potential Swea dnerÂ’s Hairstreak sites. In addition to my planned systematic search I solicited known colony sites from Dr. EmmelÂ’s research assistants, Peter Eliaza r and Stephanie Sanchez, and from Marc Minno, John Calhoun, Jeff Slotten, Hugo Ko ns, Jr., E. J. Ge rberg, and members of the Southern LepidopteristsÂ’ Societ y. I also read the tags on all the Mitoura g. sweadneri specimens in the Florida State Co llection of Arthropods (FSCA), but found this information too vague to find spec ific sites. The total recorded travel for these 120 trips was 16,688 miles and ac tual cumulative time in on-site observation was 646.15 h. Access to most cedar sites required parking along public thoroughfares, from four-lane state highw ays to tiny county roads. At each wildlife refuge or management area, I checked in with the local office for access information, allowable parking, and regulations c oncerning insect nets and flagging tape.

PAGE 135

119 Along the way, I met State Parks personne l, state foresters, wildlife officers, county sheriffs, and corrections officers from Dixie Co. hunting escapees as well as hunters, fishermen, boaters, landow ners, good Samaritans offering roadside assistance or wondering if I were simply out of my mind, and the curious public. When a site visit required leaving the car unattended for long periods, I left a printed dashboard sign that read “UF ENTOMOLOGY SURVEY” and my name. On more than one occasion I emerged from the bush carrying my 17 ft pole to find an official in a white state vehi cle parked near my car engaged in earnest radio communication. As a general icebr eaker for all these situations, I Figure 2-30. Field handout at 80% actual size Its name is Sweadner’s Hairstreak. It lives in Southern Red Cedar trees. I am researching its range, habitat, and population size. If you see it please note: Where? As specific as possible... What tree or flower? Date Time of day Have you seen this butterfly? This image is about life size Please contact: Akers Pence call (352) 376-8327 and leave your name, site information, and an evening phone number or, e-mail: pence@ifas.ufl.edu I will get in touch as soon as I can. I am a graduate student in the Entomology Dept. at UF Thank you!

PAGE 136

120 developed a sort of wanted poster handout printed here as Figure 2-30. Law enforcement, wildlife authorit ies and state employees accepted my story and a handout, and after we exchanged names and obs ervations about field conditions, we waved to each other on successive encounters. From local residents, I got advice on better places to fish, warnings about rattlesnakes and sinkholes, and, occasionally, directions to other c edar breaks. Though I must have distributed four dozen handouts over five years in t he field and several more at Southern LepidopteristsÂ’ Society and Sierra Club meetings, I never received a phone call or an e-mail about the butterfly. Results of Field Work Of the 90 cedar sites visited bet ween February 1999 and May 2004, 19 were found to have M. g. sweadneri present. At each of these marked sites, transects were established to facilitate or derly surveys of as many trees as could be reached. The complete list of sites is presented as Table 2-14. Waypoints 001-023 are the ones with the butterfly present, and all except WPT 021 supported colonies for at least part of the studyÂ’s duration. Site 021 was a planted row of nine trees plus a much older tree in a front yard at the end of my street, visited two to five times each week to cut food for colonies until eight trees of the row were cut 1 March 2004. Two individual M. g. sweadneri were seen here, one in May 1998 and another on 20 Ma y 2002. The butterfly was not seen at WPT 901-972. An additional five sites 973-977 were inspected before February 1999 and some several times sinc e, but were not included in the regular monitoring effort. Cedar si te 973 was a well known SweadnerÂ’s Hairstreak colony near the Royal Park St adium 16 Theatre in Gainesville through

PAGE 137

121 Table 2-14. Cedar sites surveyed for Mitoura sweadneri WPT Latitude Longitude Site Location Trees 001 N 29° 53.024' W 081° 17.173' St. Johns Co. St. Augustine Lighthouse on Anastasia Island 152 002 N 29° 33.889' W 083° 22.804' Dixie Co. near Jena, E side CR 361 7.5 mi S of 358 opposite Steinhatchee Falls Wildlife Management Area (WMA) 53 003 N 29° 33.594' W 083° 22.829' Dixie Co. near Jena, E side CR 361 7.8 mi S of 358 opposite Steinhatchee Falls WMA 46 004 N 29° 33.261' W 083° 22.877' Dixie Co. near Jena, E side CR 361 8.1 mi S of 358 opposite Steinhatchee Falls WMA 20 005 N 29° 33.164' W 083° 22.929' Dixie Co. near Jena, E side CR 361 8.3 mi S of 358 18 006 N 29° 42.015' W 083° 21.136' Taylor Co. Coolhey Island near Steinhatchee, end of Palm St. & River Ave. 37 007 N 29° 39.445' W 083° 21.334' Dixie Co. near Jena, on E side CR 361 0.2 mi S of 358, "Primitive Camping" 7 008 N 29° 29.954' W 081° 09.453' Flagler Co. near Flagler Beach 1.7 mi N of CR 100 on Colbert Ln. border of Misener Marine Construction Co & VacTec Center property opposite part of "Graham Swamp" St. Johns Water Management District land 16 012 N 29° 50.305' W 082° 37.568' Alachua Co. NW city limit of High Springs SW side Hwy 27 from Columbia Co line at NW 203 Place to city limits sign 160 013 N 29° 51.886' W 081° 16.494' St. Johns Co. Anastasia State Park near St. Augustine 88 014 N 29° 30.956' W 083° 15.953' Dixie Co. near Horseshoe Beach, 13.2 mi SW Cross City, W off CR 351 on Butler Rd. at Bud Gainy Rd. 8 015 N 29° 27.167' W 083° 12.088' Dixie Co. near Horseshoe Beach, Willie Locker Mainline Rd. Between CR 351 and 357 in Lower Suwannee National Wildlife Refuge (NWR) 10 016 N 29° 18.859' W 083° 01.983' Levy Co. near Cedar Key, S side CR 347 13.4 mi S of 336 in Lower Suwannee NWR 42 017 N 29° 15.139' W 082° 43.407' Levy Co., Gulf Hammock, Post Office at Hwy 19/98 and CR 326 32 018 N 29° 02.033' W 082° 42.085' Levy Co., Yankeetown, CR 40 2.5 mi W of Hwy 19/98 behind Lions Club and Coastguard 128 019 N 29° 37.565' W 081° 47.668' Putnam Co., Hollister, Division of Forestry, Etoniah Creek Office and Frank Cone's yard on SR 20 and Wipple Tree Rd. 92 020 N 29° 33.754' W 083° 19.999' Dixie Co. near Jena, 3.27 mi E. of CR 361 on Horse Shoe/Bowlegs/Mainline Rd. Palm Ridge Hunt Club became Cyprus Ridge Hunt Club in 2002 24 021 N 29° 38.899' W 082° 14.625' Alachua Co. near Gainesville at Newnans Lake, 329 SE 71st St. 10 023 N 29° 31.678 W 82° 11.863 Alachua Co. near Cross Creek Zane Hogan's Ranch CR 346 6.5 mi E of 441 22 901 N 29° 34.268 W 83° 23.224 Dixie Co. near Jena 0.7 mi W of 361 on Sink Creek Rd in Steinhatchee Falls WMA 27

PAGE 138

122 Table 2-14. Continued WPT Latitude Longitude Site Location Trees 902 N 29° 55.258 W 81° 19.424 St. Johns Co. St. Augustine, behind motel on San Marco Ave 32 903 N 29° 28.908 W 81° 21.026 Flagler Co. near Bunnell N side of Hwy 100 near CR 205 to Espanola 12 904 N 29° 27.737 W 81° 15.171 Flagler Co. near Bunnell US 1 at Carter St. 6 905 N 29° 24.318 W 81° 11.764 Flagler Co. near Korona Old Dixie Hwy & US1 & CR 325 E side of US 1 6 906 N 29° 26.650 W 81° 07.789 Flagler Co. Hwy 201 just north of VolusiaFlagler Co. lines South side of road 3 907 N 29° 33.600 W 83° 22.855 Dixie Co. near Jena W side CR 361 S of 358 23 908 N 29° 31.000 W 83° 22.256 Dixie Co. 361 W of Jena after pavement end past the boat ramp, hammock in salt marsh 96 909 N 29° 31.031 W 83° 22.192 Dixie Co. entrance to turnoff for hammock (WPT 908) 3.06 mi S of WPT 003 12 910 N 29° 31.852 W 83° 22.666 Dixie Co. near Jena both sides of 361 (E&W) 2.07 mi S of WPT 003 15 911 N 29° 41.877 W 83° 21.383 Taylor Co. near Steinhatchee SR 51 & River Ave 2 912 N 29° 43.891 W 83° 21.574 Taylor Co. near Steinhatchee SR 51 (Sandhill Hunt Club) 2.3 mi N of River Ave 1 913 N 29° 36.603 W 83° 05.064 Dixie Co. in Eugene CR 55A & Hwy 27 57 914 N 29° 46.412 W 82° 25.986 Alachua Co. Alachua both sides Hwy 441 at front entrance to Moltec 106 915 N 29° 53.344 W 82° 36.509 Columbia Co. near High Springs E side Hwy 441 1.8 mi WSW of O'Leno State. Park. 16 916 N 29° 56.854 W 82° 49.798 Suwannee Co. near Hildreth 276 Terr & 43 Rd30 917 N 29° 57.781 W 83° 46.513 Taylor Co. SW of Perry CR-356 meets Gulf boat ramp Aucilla Wildlife Mgmt. Area 15 918 N 29° 48.660 W 83° 34.782 Taylor Co. E side of CR 361 between Keaton Beech & Steinhatchee 7 919 N 29° 53.027 W 81° 19.433 St. Johns Co. St. Augustine US 1 N. English Landing N parking lot 33 920 N 30° 05.144 W 81° 27.434 St. Johns Co. near Durbin, Dexter Farm 10840 Old Dixie Rd 65 921 N 30° 07.774 W 81° 22.752 St. Johns Co. CR 210 Roscoe extension entrance of Guana River Wildlife Mgmt. Area 23 922 N 29° 32.346 W 83° 22.384 Dixie Co. near Jena 1.5 mi N of end CR 361 E side at Palm Coast Hunt Club 82 923 N 29° 39.151 W 83° 21.331 Dixie Co. near Jena CR 361 6.25 mi N of WPT 002 3 924 N 29° 36.698 W 82° 54.415 Gilchrist Co. Hwy 26 0.8 mi E of Wilcox at BP 45 925 N 29° 51.995 W 81° 18.161 St. Johns Co. near St. Augustine, Anastasia Island, E end of CR 312 bridge on St. Johns River (S under bridge) 28 926 N 29° 51.977 W 81° 16.569 St. Johns Co. Anastasia State Park nature trail 48 927 N 30° 07.086 W 81° 23.247 St. Johns Co. Palm Valley Rd at CR 210 42 928 N 30° 22.504 W 81° 24.304 Duval Co. Jacksonville, Kathryn Abbey Hanna Park 4 929 N 29° 26.278 W 83° 17.502 Dixie Co. town of Horseshoe Beach 29 931 N 29° 30.750 W 83° 15.942 Dixie Co. near Horseshoe Beech, Butler Rd NW of CR 351 (near WPT 014) 9

PAGE 139

123 Table 2-14. Continued WPT Latitude Longitude Site Location Trees 932 N 29° 31.714 W 83° 14.144 Dixie Co. near Horseshoe Beech, CR 351 between 357 & Bowlegs Rd 11 933 N 29° 28.163 W 83° 16.469 Dixie Co. just E of 351 on Willie Lockler Mainline Rd 1 934 N 29° 27.992 W 83° 15.510 Dixie Co. Jena unit Big Bend Wildlife Mgmt. Area sign intersection of Rd # 18 & Rd #20 3 935 N 29° 23.652 W 83° 12.123 Dixie Co. Shired Island campground CR 357 Lower Suwannee National Wildlife Refuge 16 936 N 29° 23.582 W 83° 11.968 Dixie Co. Shired Island, W side of 357 at Gulf Lower Suwannee National Wildlife Refuge 18 937 N 29° 24.952 W 83° 10.741 Dixie Co. near Suwannee, Mainline Road between 357 & 349 Lower Suwannee NWR 10 938 N 29° 31.828 W 82° 30.373 Alachua Co. Archer, SW 154 St. 1.3 mi S of Hwy 24 9 939 N 29° 31.560 W 82° 49.332 Levy Co. S of Trenton, Hwy 49/129, 0.5 mi N of CR 349 191 940 N 29° 09.094 W 83° 02.941 Levy Co. Cedar Key, near Cedar Key State Museum 30 941 N 29° 25.000 W 82° 52.310 Levy Co. CR 347 & 341 S of Chiefland, W of Usher at 6050 NW 60 St 15 942 N 29° 56.976 W 81° 34.821 St. Johns Co. near Palmo, 8301 Colee Cove Rd (off SR 13) 12 943 N 29° 55.934 W 81° 33.870 St. Johns Co. near Palmo, Colee Cove Rd & SR 13(derelict nursery) 30 944 N 29° 00.206 W 82° 45.517 Levy Co., Yankeetown at County Park & boat ramp, end of C-40 29 945 N 29° 00.380 W 82° 45.268 Levy Co., Yankeetown, turn out on C-40, 0.5 mi N. of Levy County park 21 946 N 29° 13.213 W 82° 45.751 Levy Co. near Gulf Hammock C-326, 3.5mi W of 19/98, W end of last bridge before boat ramp 5 947 N 29° 14.686 W 82° 43.825 Levy Co. near Gulf Hammock, E side C-326 0.5 mi SW of 19/98 13 948 N 29° 24.542 W 81° 07.145 Volusia Co. Mound Grove, 201 & Highbridge Rd to boat ramp 4 949 N 29° 21.171 W 81° 06.114 Volusia Co. near Tomoka State Park entrance, SR 5A 32 950 N 30° 33.909 W 81° 49.667 Nassau Co. Callahan, SR A1A E of Hwy 1 near Stratton Rd. 7 951 N 30° 36.099 W 81° 45.476 Nassau Co. Nassau Wildlife Management Area SR A1A 5.5 mi E of Hwy 1 12 952 N 30° 43.327 W 81° 37.448 Nassau Co. Crandall, boat ramp at end of Crandall Rd. St. Marys River & GA border 56 953 N 30° 39.355 W 81° 26.432 Nassau Co. Fernandina, Amelia Island, Jasmine St. between SR A1A & Citrona 33 954 N 29° 38.869 W 83° 21.140 Dixie Co. near Jena, CR 361 0.91 mi S of 358 2 955 N 29° 31.428 W 83° 22.421 Dixie Co. near Jena, CR 361 10.5 mi S of 358 11 956 N 29° 39.276 W 83° 21.354 Dixie Co. near Jena, First Baptist Church of Jena, CR 361, 0.7 mi S of 358 8 957 N 29° 14.447 W 82° 43.974 Levy Co., Gulf Hammock 1.25 mi SW of 19/98 28 958 N 30° 36.293 W 85° 56.501 Walton Co. near Redbay, SR 81 1.5mi N 2

PAGE 140

124 Table 2-14. Continued WPT Latitude Longitude Site Location Trees 959 N 30° 48.379 W 85° 14.975 Jackson Co. near Marianna, Penn Ave turns into Bump Nose Rd. 1.6 mi N of Marianna City limit 22 960 N 30° 49.000 W 85° 15.573 Jackson Co. near Marianna, BumpNose Rd. 2.1 mi N Marianna 17 961 N 30° 37.586 W 81° 29.370 Nassau Co., O'Neil, on SR A1A under the bridge to Amelia Island 29 962 N 29° 38.171 W 81° 38.805 Putnam Co., Palatka, Ravine State Park & Gardens 33 963 N 28° 48.873 W 80° 51.969 Volusia Co. near Oak H ill, NW corner of US1 & Stacy Grove Rd. Turnbull Hammock St. Johns National Wildlife Refuge 22 964 N 28° 49.635 W 80° 51.118 Brevard Co. off US 1, public access rd., under RR bridge, near main entrance of Merritt Island National Wildlife Refuge 24 965 N 29° 29.254 W 82° 58.611 Levy Co. Manatee Springs State Park (cedars outside entrance to park, none inside) 17 966 N 28° 32.893 W 82 37.568 Hernando Co. near Bayport, N & S sides CR 550 & SR 50 49 967 N 28° 33.230 W 82° 38.277 Hernando Co. near Bayport CR 550 just E of CR 495 / 595 4 968 N 29° 40.148 W 81° 15.562 border St. Johns Co.& Flagler Co. at FaverDykes State Park, E of US 1, N of CR 204 4 969 Liberty Co. Apalachicola National Forest no cedars seen along SR 65 or CR 12 none 970 N 28° 24.944 W 82° 38.957 Hernando Co. & Pasco Co. line, near Aripeka, S end of CR 595 & US 19 found several cedar stands along 595 and Hernando Beach Rd. lots 971 N 28° 57.986 W 81° 20.459 Volusia Co. near Orange City Blue Springs State Park 23 972 N 29° 51.263 W 81° 16.751 St. Johns Co., St. Augustine Beach City Hall and Police Station, on Anastasia Island 24 973 N 29° 39.265 W 82° 22.777 Alachua Co., Gainesville, Royal Park Theatre parking lot 7833 Newberry Rd. 27 974 N 29° 38.054 W 82° 22.042 Alachua Co., Gainesville, UF Natural Area Teaching Lab, Natural Area Dr. 2 975 N 29° 38.546 W 82° 22.322 Alachua Co., Gainesville, UF campus W of Physical Plant Division parking lot, SW 34 St. & Radio Rd. 15 976 N 29° 37.640 W 82° 15.491 Alachua Co. near Gainesville at Newnans Lake, CR 329B & SR 20 9 977 N 29° 39.012 W 82° 14.558 Alachua Co. near Gainesville at Newnans Lake, 200 SE 71st Terr. 8 95 TOTAL sites TOTAL trees surveyed 2,780 the 1980s and early 1990s (Jeff Slotten and Marc Minno label data and personal communication), but none were found flying there in the 1998 flight season and in 1999 all but two of the original trees were destroyed when the theater was

PAGE 141

125 remodeled and its parking lot expanded. This site was visited several times since in search of the butterfly on the remain ing trees. Cedar site 974 consists of two female trees on the edge of the U. F. Na tural Area Teaching Lab across from the U. F. Entomology and Nematology Dep t., visited once or twice a week to cut cedar to feed captive colonies for three years during flight s eason. Waypoint 975 is today a fragment of decidu ous forest with a stand of 15 cedars intermixed at its edge where butterflies were released by Dr. EmmelÂ’s field assistants in an experiment to establish an on-campus colony in the late 1980s. Sites 976 and 977 are near my home and have been visited frequently since 1998 as part of my exercise regimen. Two additional WPT in Table 2-14 need clarification. Entry 969 has neither latitude nor longitude listed because it r epresents a large part of Apalachicola National Forest in Liberty County where I could find no c edar groves. It was listed because cedars were conspicuously abs ent from the edge habitat surveyed. Waypoint 970 represents my attempt to find the southernmost inhabited cedar site south of a colony recommended by Cla y Black (pers. com.) of the Southern LepidopteristsÂ’ Society which may represent the southern limit of the butterflyÂ’s range. Of 2,780 trees surveyed, 965 of them were in the 19 marked M. g. sweadneri sites. Of these 965 cedars, 130 had t he butterfly present on one or more of 207 visits to inhabited cedar sites and were measured, drawn, and marked with flagging tape.

PAGE 142

126 Over the course of my study , 472 individual sightings of M. g. sweadneri were recorded (470 in trees ). Of these, 380 were male , 48 were female, and the remaining 44 were of undetermined sex. Of the total sighted, 72 were captured, 50 males and 22 females, and except for individuals retained to establish captive colonies or document specific phenotypes , most were released into the same cedar site where they were ta ken within minutes of capture. Analysis of Field Study The life history of SweadnerÂ’s Hairst reak is inextricably bound to its Southern Red Cedar host. The specific c haracteristics of the individual trees used by the butterfly define critical qualities essential to the habitat of M. g. sweadneri . Three dimensions, height, width, and trunk diameter, were compared to get a general profile of what the butte rflyÂ’s habitat looks like. The sex of the trees provides another observable quality for evaluation. But the ecology (how the butterfly uses the cedar) of the bu tterfly and cedar provides the essential perspective from which the habi tat description is drawn. Tree dimensions Measurements of the heights of 126 of the 130 trees in which M. g. sweadneri were observed are compared in a histogram in Figure 2-31. Though the average height of trees us ed by the butterfly is 8.28 m, this statistic conveys little information without the distribution shown by the bars in the figure. The marked trees range from 3.3515.85 m tall but the vast ma jority, 86.5%, of these trees are in the height classes greater than 5.6 m and less than 11.6 m (18-38 ft). The four missing measurements were due to errors in data collection early in the survey and the destruction of the trees before the errors could be corrected.

PAGE 143

127 Figure 2-32 is a histogram comparing the widths, usually at the lowest branches, measured for 103 of the 130 ma rked trees. The missing width data for 27 trees are either from cedars standing cl ose together in a tightly packed grove, 1 4 6 16 32 16 19 16 10 3 2 1 0 0 5 10 15 20 25 30 353 . 3 5 4.6 5.5 6 . 5 7.62 8 . 5 3 9.5 10.67 11 . 5 12.5 14 . 5 15.85 Mo retree height classes in meterstree totals mean 8.28 m STDV n-1 2.12 m range 3.35-15.85 m sample size 126 Figure 2-31. Heights of marked trees histogram 11 6 24 17 14 12 7 10 6 3 2 0 0 5 10 15 20 25 300.61 0 . 9 1 1.52 2 . 6 3 . 6 6 4.7 5 . 4 9 6. 7 1 7 . 6 2 8 . 53 9 . 4 1 0 .8 Moretree width classes in meterstree totals mean 4.45 m STDV n-1 2.40 m range 0.61-10.8 m sample size 103 Figure 2-32. Lowest widths of marked trees histogram

PAGE 144

128 or from larger trees in a shelterbelt or forested edge wh ich have lost most of their lower limbs through human or natural pruning. In the interior of a close grove, the cedars grow into each other and the lo west limbs tend to be shaded out and spindly when under a closed cedar canopy. A gain, the average wi dth of 4.45 m is not as informative as the distribution bar s in Figure 2-32 which show that 90/103 or about 87% of the marked cedars are wider than 1.83 m but not greater than 8.53 m (6-28 ft) wide. 2 21 15 21 19 10 14 4 1 4 2 0 0 5 10 15 20 254 . 7 7 9.55 1 2 . 7 3 17.83 2 2 . 6 27.33 3 2 . 7 9 38.2 4 3 . 29 47. 7 5 57 . 93 Moretree diameter classes in centimeterstree totals mean 19.61 cm STDV n-1 11.09 cm range 4.46-15.93 cm sample size 113Figure 2-33. Diameters of marked trees histogram The last tree dimension assessed was diameter at chest height. This was measured, when possible, by wraping a dressmakerÂ’s tape around the tree and then converting the circumference to diam eter. Traditional direct measurement of diameter using a caliper was impractical because the cedar trunks, particularly of trees 20 years or older, are not symmetric ally round; rather they are shaped by bases of limbs bulged out in support of t heir weight as the tree grows interspaced

PAGE 145

129 with furrows and ridges, making accurate placement of the caliper points impossible. Secondly, the size of an instrument large enough for big trees is unwieldy amid the dense limbs of a health y cedar specimen. Data collection was limited by the density of lower limbs of some trees that di d not allow a close approach, or by the presence of Poison I vy, Toxicodendron radicans (L.) Kuntze, climbing the trunks of otherwise easily approachable butterfly cedars. For these reasons the sample of trees displayed in Figure 2-33 is limited to 113 of 130 marked cedars. About 88.5% of these diameters fall between 5.4 and 32.8 cm (2.13-13 in), which is a wide range of size . But the telling statis tic of Figure 2-33 is the unbiased standard deviation of 11.09 cm, suggesting the butterflies use trees of more varied di ameters than of any other measured dimension. Tree sex The sex of most mature J. v. silicicola is discernable during much of the year by their flowers and fruit. The tiny fl owers of late winter (Figure 2-34) could be overlooked at first glance, but the male trees are instantly recognizable by the cloud of golden brown pollen that the cones release on the wind, or when the tree is struck. For weeks a fter flowering is finished but before the fruit becomes obvious, the male trees rain spent c ones when tapped. From early summer until the end of the year, most sexually matu re female trees bear juniper berries (Figure 2-35), which can be quite dramatic in mast years. Usually, smaller trees in the lower ranges of height (< 6.7 m), and diameter (< 12. 73 cm) are immature, showing neither flowers nor fr uit. While sexual maturity is a function of the treeÂ’s age, height and particularly diameter are not necessarily so determined.

PAGE 146

130 Figure 2-34. Juniperus v. silicicola flowers A) Female, B) Male at 9 X life size Figure 2-35. Juniperus v. silicicola berries Trees of the same age growing in a grov e or planted shelterbelt may vary in diameter due to differences in growing cond itions such as slight differences in available nutrients or soil index. Figure 2-36 shows t he butterflyÂ’s tree choices broken down by sex, with 63% female over about 15% male. If half the

PAGE 147

131 82 19 29 020406080100 female male unknown tree totals sample size 130 Figure 2-36. Marked trees delineated by sex undetermined trees were male, that still represents a wide margin, 26% versus 74%, of nearly three times as many female as male cedars used by the hairstreak. The number of trees in a natur al grove or shelterbelt with M. g. sweadneri present varied from a minimum of se ven up to 160. Sheer numbers of trees however, do not signal highest habitat qua lity. The waypoints with the most consistent hairstreak colonies present fo r every brood during the survey included 7, 24, 53, and 92-tree transects, of which onl y 3, 7, 23, and 44, were ever seen to harbor the butterfly. Table 2-15 divides t he tree sex data of Figure 2-36 into the number of times the hairstreaks were s een in each tree sex. For instance, the eighth row shows that three of the 82 marked female trees held M. g. sweadneri on each of eight site visits, but only one male cedar was used that frequently. The added perspective of frequency of obser vations of butterflies in specific

PAGE 148

132 Table 2-15. Frequency of observations delineated by tree sex Observations* per tree Female trees Male trees Unknown trees Subtotals by frequency 1 33 6 26 65 2 18 6 2 52 3 13 2 1 48 4 6 1 28 5 3 2 25 6 3 18 7 1 7 8 3 1 32 9 1 9 11 1 11 18 1 18 tree totals 82 19 29 Obs, total = 313 *each observation may include one to five butterflies. cedars is evidence that the butterflies not only prefer female tr ees but were also choosing specific individuals of the 130 marked ones out of the 965 available trees in all 19 SweadnerÂ’s Hairst reak sites on multiple occasions. An explanatory note should be made on the large numbers of single and double frequency trees and some missing tree dimension data. From the beginning of the study, there were natural successional and especially anthropogenic changes to several study sites, ranging from pollarding of individual trees along drivew ays or power lines to wholesale clearing of cedar groves for pine plantations or other develop ment. The loss of many of the original 19 study sites forced more concentrated at tention and more frequ ent visits to the remaining thriving colonies and may have re sulted in more meaningful data from the smaller sample. The documented loss of habitat is an important part of the conservation biology of SweadnerÂ’s Hairst reak and is discussed at length in a later chapter.

PAGE 149

133 Phenology Of the references cited in the lit erature section above, the closest description of the phenology I observed is from Emmel (1993) who said M. g. sweadneri appeared in three seasonal broods: spring, summer, and fall. The most specific indications for finding the butterfly in situ were given in the Butterflies Through Binoculars field guide (Glassberg et al ., 2000) which recognized only two broods but divides the flight times and expected periods of abundance by region. In the Florida Panhand le from Jefferson County west, they said the season is from March to late September, but gave no period of abundance. In northern Florida from Levy, Marion, Putnam, and Flagler counties to the Georgia line and the central counties as far s outh as Hernando, northern Polk, and southern Volusia counties (judging from their range map), they said the season runs from early March to mid Oc tober and that the butterfy is most common from March to May, and in September-October. My field observations of 472 sightings , principally along north latitude 29° 02” to 53”, are diagramed in Figure 3-37. Abundance for each day ranges from one to a maximum of 44 individuals seen du ring the 5-year survey. Since a field trip might include surveys of one to someti mes as many as five marked sites, the numbers represent daily totals limited to Gulf , central, or Atlantic coast counties without regard to specific locations. Though I made several trips each January before the first flight began and follow-up tr ips after the flight season in late October and November to confirm defin ite beginning and ending dates, I culled these dates from the records for this c hart to save space. I also excluded dates the butterflies were absent from ephem eral colonies and did not repeat

PAGE 150

134 0 5 10 15 20 25 30 35 40 451-Feb 15-Feb 25-Feb 2-Mar 9-Mar 19-Mar 25-Mar 3-Apr 13-Apr 30-Apr 7-May 10-May 21-May 31-May 10-Jun 18-Jun 29-Jun 3-Jul 9-Jul 2-Aug 28-Aug 25-Sep 15-OctSite visit datesindividuals sighted Figure 2-37. My field records of M.g. sweadneri combined by date, 1999-2004 seasons compiled. the absent dates if the butterflies were pr esent in consecutive years on the same date. “Absent” dates, limited to no more t han three consecutive days in the chart, were included to illustrate possible gaps in the flight season. Coincidently, ten dates occur in more than one year (T able 2-16), and point to a couple of interesting observations. First, site s in the same county may have similar numbers flying in different y ears. Second, sites as far apart as Yankeetown, or Jena on the Gulf coast (WPT 018, and 002) and Hollister in Putnam Co. (WPT 019), may have close to the same totals, or very different numbers around the same date in the flight season. This is ev idence that the fli ght season across the whole studied range is not in perfect syn chrony, but rather varies by region somewhat as Glassberg et al . (2000) suggested.

PAGE 151

135 Table 2-16. Dates of butterflies occu rrence multiple years in Figure 2-37 Date Year # sighted Region County Site/WPT 19. Mar 2003 9 west Dixie 002, 007 20. Mar 1999 17 west Dixie 006, 007 2004 2 east Alachua 023 24. Mar 2003 15 west Dixie 002, 007, 020 25. Mar 2000 7 east Putnam 019 2002 18 east Putnam 019 10. Apr 1999 18 west Dixie, Taylor 003, 006 2000 6 west Dixie 002, 003, 007 13. Apr 2003 2 east Putnam 019 14. Apr 2001 1 west Dixie 020 2003 1 west Dixie 002 30. Apr 2000 1 west Dixie 007 2003 0 west Dixie 002 7. May 2003 2 east Putnam 019 8. May 2000 0 west Levy 017, 018 9. May 2001 0 west Dixie, Levy 002, 007, 020, 018 9. May 2004 0 east Alachua 023 29. May 2003 8 west Dixie, Levy 002, 007, 018 31. May 2003 5 east Putnam 019 22. Jun 1999 2 east St. Johns 001, 013 2003 2 east Putnam 019 29. Jun 2002 0 east Putnam 019 2003 4 west Dixie 002 30. Jun 2001 14 west Dixie 002, 020 2002 14 west Dixie 002, 007, 020 28. Aug 1999 9 west Levy 018 2003 18 east Putnam 019 Since my own field observations were not made on exactly the same days each year and I traveled to different sites more often in some months than in others, my own data are not perfectly even for July acro ss the state and are thus limited relative to my method and by the loss of several study sites. I compared my data with those of several others, including 41 published records collected from The LepidopteristsÂ’ Societ y Season Summary, and the Southern

PAGE 152

136 Lepidopterists’ News from 1974 to 1988 by John V. Calhoun plus 20 of his own sightings, and 105 records from T. C. Em mel and his field assistants P. Eliazar, J. L. Nation, Jr., D. Smith, M. Minno, and K. Weisner. I also collected the dates of 28 labeled specimens from B. Hollister, J. Slotten, and H. Kons, Jr. to add to 41 label dates from the FSCA, in cluding several from H. C. King, and C. F. Zeiger for a total of 235 records. These label data, in some cases from long-dead collectors, do not give total numbers seen, and 93 of the records only indicate that the butterfly was flying on that day, so I treated them a ll as days that the species was present and combined them with my own 472 sightings for Figure 2-38. This graph demonstrates that at least one observer and as many as five have seen the butterfly on 122 days of a possible 227-day flight season, 1 February to 15 October. Forty-three of the total “present” days are exclusively from my field work; six of the dates I did not find the but terfly at my study sites were “present” records from other lepidopterists, but 14 of my not-present dates stand. The earliest historic record in th is collection is of a worn male taken 0 1 2 3 4 51-Feb 19-Feb 26-Feb 2-Mar 7-Mar 10-Mar 17-Mar 20-Mar 26-Mar 31-Mar 3-Apr 6-Apr 10-Apr 14-Apr 19-Apr 22-Apr 29-Apr 4-May 8-May 16-May 21-May 27-May 30-May 6-Jun 13-Jun 17-Jun 20-Jun 25-Jun 30-Jun 9-Jul 19-Jul 28-Jul 12-Aug 23-Aug 28-Aug 3-Sep 12-Sep 17-Sep 23-Sep 28-Sep 7-Oct122 daysnumber of records Figure 2-38. Days M. g. sweadneri were present from multiple sources

PAGE 153

137 by F. M. Chermock at St. Augustine on 15 June 1921. The earliest occurance is mine on 1 February 2002 from Dixie Co. on t he Gulf coast. In fact, all but one of the February date records are mine. All of the late season Gulf coast records are from the last week in September except a single record from Levy Co. on 7 October 1979 by C. E. Stevens. The earlie st Atlantic coast date is 20 February 1982, Duval Co., C. F. Zei ger, repeated again in 1994 from Volusia Co. by J. V. Calhoun. All the rest of t he early east coast dates are from the first week in March. The late dates fr om that coast are all S eptember, with 12 September 1975, New Smyrna Beach, Volusia Co. by G. W. Rawson being the latest record. The paucity of records from interior Flor ida makes my early date of 26 February 2003, and my late one of 15 October 2003, both from Putnam Co., the brackets of the interior Florida season. I summarize the phenology of M. g. sweadneri as follows. The flight season usually begins about 15 February in the Gu lf coast counties, followed two to three weeks later by emergence in the in terior counties, Alachua and Putnam. The longest lag observed was 25 da ys. The season winds down in late September on the Gulf, but persists into early October in the interior. On the Atlantic coast, first emergence is usually in the first week of March and this season ends by mid September. The ear liest and latest dates reported above were the extremes. At least three flights, or broods occu r across the range every year, but four flights were common during my study. The ov iposition to eclosion times recorded in captive raising experim ents from Table 2-11 are almo st perfectly observed as

PAGE 154

138 45 to 50 day intervals in Figure 2-37. For example, eggs laid on 16 February produce the beginning of a new flight on 3 April. Thei r progeny emerge about 21 May, and so on throughout the season. The few individuals marking the end of a brood were usually very worn and faded males. Larval development and pupation may be a little faster in the high summer brood. When the last flight of the season (eclosing in late Augus t/September) reproduces, their progeny develops in time to overwinter as pupae to emerge the following year. The hairstreak was usually seen in low numbers, with 37 days having five or fewer adults in even the best habi tats and only 14 days with records of 10 or more. The two days when more than 20 were observed, 2 and 22 September 2002, were both at the Put nam Co. site and were my maximum records of 31 and 44, respectively. In general, my fiel d observations were disappointingly below Emmel’s (1993) report of small colo nies with 12 to 50 individuals normally present. Adult behavior Mitoura g. sweadneri was not seen basking any later than 09:14 recorded on a cool morning, 16 May 1999, when the temperature was 17.8°C (64°F). Basking posture is with the wings closed and tilted, broadside to the sun. For four to five weeks before the first flight in Dixie Co., the daytime high temperatures were 17.5°-24.5°C (63.5°-76°F) on preflight field trips in each year of the study, with late night temperatures dropping into the 3.3°-10.6°C (38°-51°F) range and occasional dips to near freezing. For eac h year except 2000, I witnessed the first flight of the season within one to seven day s after adults emerged at one or more marked sites. Opening day highs have been between 24.5°C and 28°C (76° to

PAGE 155

139 82°F) with the high 28°C, interestingly, recorded on the earliest beginning date for a season, 1 February 2002, in Dixie Co. The butterf lies probably bask at first light on cool days and thus were rarely seen in basking posture because the ambient temperature rises so quickly after sunrise. On six of the nine opening days recorded, only males were observed but both sexes were flying on the other thr ee (Table 2-17). Even though these are Table 2-17. First of season hairstreaks seen by sex Date Females Males Site/WPT 27. Feb. 1999 1 4 003 19. Feb. 2000 0 6 017 16. Feb. 2001 2 5 002, 007, 020 1. Feb. 2002 0 4 002, 007, 020 15. Feb. 2003 0 1 002 26. Feb. 2003 1 5 019 3. Mar. 2004 0 1 023 7. Mar. 2004 0 4 019 9. Mar. 2000 0 4 019 low numbers, the in situ observations reinforce the ev idence from captive rearing that M. g. sweadneri harbingers are not strictly prot androus, contrary to the rule for most butterflies (Ehrlich and Ehrlich, 1978, Scott, 1974). Males are almost always seen perching high in specific cedars of a grove or planted row. These trees are used brood after brood during a season and in many cases are the preferred trees year after year. Perching trees range from 4.57-15.85 m tall and average 8.4 m from all studied sites, standard deviation 1.55 m. The single tallest tree was 4. 6 m higher than the next tallest and the butterflies were rarely seen in trees 11 m or more tall. In most every site, these perching trees fit in near the top edge of the average height of the grove, but were not usually the tallest trees and, in fact, they lack a particular observable

PAGE 156

140 quality that sets them apart from the rest. Mature fema le cedars are chosen by a wide margin over male trees, as indicat ed previously in Figure 2-36. At only one continuously inhabited site, WPT 007 in Dixi e Co., was there only a single lekking tree that was the tallest. This female tree was over 2.5 m taller than its nearest challenger and over 3 m taller than any of the other trees of the grove. At the height of a brood with the most adults present at WPT 007, this tree held as many as four males awaiting female c onsorts. In very few instances, and then only briefly, were males perched at the apex of a cedar but usually within 1-1.5 m down from the top. The m ean distance from the top for 298 perching males was 1.55 m, standard deviation 0.98 m. In t he morning hours they were found on any side of the top with about equal frequency unt il 10:30-11:00, when they move to the side away from the sun. As the sun moves west from noon, they were most frequently seen perching on the northeast side of the top in more broken sunlight than that afforded by perches with wes t, or southwest exposure. This behavior held true for full sun and partly cloudy days, but was not as clearly defined on heavily overcast days. A second caveat is directly related to wind speed and direction. On afternoons when the wind was constant at any speed above 6-10 mi/h, or with sustained gusts over 11 mi/h measured at eye level, the butterflies sheltered on the leeward side of the tree regardless of the direction or intensity of the sun. Perching position is probably important in thermoregulation for these butterflies. The temperature di fferential between the direct sun and the shady side of a cedar easily varies from 2-6 Celsius degrees and may be as wide as 12 C° in cool weather.

PAGE 157

141 On early morning surveys, males were seldom seen interacting; however, at sites with several adults flying, males begin whirling in pairs and sometimes in groups of three to five by 10:40. Of 56 whirling incidents recorded, the frequency is about equal for every hour between 11:00 and 18:00 except for the period 14:00-14:40, when the total whirling incident s were twice that of any other hour. The latest single whirling pair was at 18: 02. Whirling behavior is apparently not defense of territory because one male never chased another from a tree. I agree with Scott (1974) that this is investigatory behavior. The males fly out to investigate flying objects of appropriate si ze in an effort to encounter mates. Though some authors describe fights between conspecific males of families with larger individuals like Papilion idae (Ledarhouse, 1982, Rutowski et al ., 1989, Pinheiro, 1990, and Wickman and Wiklund, 1983), at one time whirling behavior of genus Mitoura was considered play by some observers (Scudder, 1889a). Though not a territorial dispute, whirlin g may function as a test of stamina amoung males and give advantage to him who still has energy to greet the most females. I have timed several whirling se ssions and found they often lasted 11 to 15 min with the longest lasting over 21 mi n when the males who were involved when I arrived finally separated. As a brood ages at one particular site, the incidence of whirling and length of time decreases as the males become more worn. I wonder if another func tion of whirling is as a vi sual display, or perhaps pheromonal display that helps attract newly eclosed females to the tree. This perching and whirling behavior of M. g. sweadneri qualifies as lekking under E. O.WilsonÂ’s (1975) definition of a lek as a communal display or display area

PAGE 158

142 where males congregate to attract and court females and females come specifically to mate. Lekking is an effici ent system of mate location in a species of such low observed population density as with this hairstreak. Observed in situ copulation times were listed in the description of captive copulation above. Both the Tennessee and Fl orida butterflies mated in the mid to late afternoon in the field, and into the evening hours in captive mating experiments. Female M. g. sweadneri are much less conspicuous in the habitat than males. Heights were recorded for 45 fe males in 40 cedars ranging from 3.35 m (1.22 m shorter than the tree of any perch ing male) to one that was 11 m tall. All but seven were less than 9.5 m, and 19 were less than 8 m cedars. Average height for these 40 trees was 7.38 m (c ompare with 8.4 m for males) and the standard deviation was 1.84 m. The aver age distance down from the top was 3.09 m, or about twice that found for males. The gr eatest down-from-the-top measurement of a femaleÂ’s position was 8.69 m, also double that of any perching male. Oviposition behavior was recorded several times in every hour from 11:20 to 16:35. Captive females were someti mes seen going through the movement of tucking the tip of the abdomen below a cedar sprig but not depositing an egg so in situ oviposition was only recorded 15 time s and the eggs were found in 14 of those incidents. Observed oviposition height shows an even more dramatic division of habitat use by the sexes than perching. First, the cedars chosen for oviposition ranged from 3.359.14 m tall, averaged 6.38 m, with a relatively broad

PAGE 159

143 standard deviation of 2.11 m. Second, eggs were found at an average height from the ground of 2.39 m, standard deviation 1.28 m. Only one of the 15 was higher than 3.66 m and 11 were less than 2. 5 m. The lowest observed oviposition posture after which an egg was found wa s at 45 cm from the ground. Two females were seen to lay two eggs on the same tree. The eggs were usually found 1-5 cm back from the growing tip in situ , and were not hidden or necessarily placed under the cedar lea f. Nine oviposition cedars were female, four were not sexually mature, and two we re male. Ten of them were conical or shaped like rounded cones with the lowest limbs spread within 38 cm (15 in) above the ground. One tree had lowest branc hes, 1 m, one, 1.22 m, and one that received two eggs had all of its lim bs lower than 1.83 m removed. The non-conical trees either had double trunks, or had had their tops broken off in the past, giving them a rounded or lobed appearance. One of these was 3.35 m tall and shaped like a salad bowl 7.62 m wide. After laying an egg, most females flew to another cedar, but on two occasions they stepped inside the tree and disappeared. Females were perched in the same trees with males 15 times, perched in trees without males 15 times, ovipositing alone 11 times, and ovipositing in trees with males present four times. Several of the trees chosen for oviposition were never seen to have males present and most of these were smaller trees, but a few eggs were laid in lekking trees well below perching males. These observations suggest that both sexes use m any of the same trees and that the

PAGE 160

144 division of habitat is based more on the heigh t of trees, or, mo re importantly, the height in trees. Females may be less conspicuous if, lik e most butterflies, they only mate once (Alcock, 1983, 1987, Alcock and O Â’Neill, 1986, Ehrlich and Ehrlich, 1978, and Wickman and Wiklund, 1983) and are, theref ore, less frequently seen than the more obvious perching males. They may avoid oviposition near the tops of lekking trees because females at this stratum are frequently approached by males seeking mates (Scott, 1974).Though several authors suggest males live longer than females, in captive colonies the female M. g. sweadneri frequently survive males of the same cohort by severa l days. I think they initially lay eggs throughout their natal cedar grove and then disperse for the next nearest cedars. On only a few occasions have I found old females at study sites near the end of a brood. SweadnerÂ’s Hairstreak was almost a lways found on its host when adults were flying. Most flights we re very brief and took the fo rm of very short loops of 1.22-1.83 m (4-6 ft) from a branch tip perch and back in response to a sharp rap of the tree. Even when several individu als were flying, flights between trees rarely exceeded 7 m, and a few times, individuals crossed county road 361 in Dixie Co. for median records of 18 m ( 59 ft) The two longest distances were one of 35 m from tree to tree and another of 54 m ( 177 ft) from flowers to the nearest cedar. Flights of this distance and longer are probably commonplace because some study sites with stable colonies lack obvious nearby nectar resources, especially during the first flight.

PAGE 161

145 The butterfly was rarely seen at flow ers. At the beginning of the flight season, they were on Hog Plum, Prunus umbellata , and possibly, Chickasaw Plum, P. augustifolia , when little else was in bloom. I have seen one individual on Purple Deadnettle, Lamium purpureum , in March, and in late May, a single male visited Purple Coneflower, Echinacea purpurea , and Indian Blanket, Gaillardia pulchella , in my wildflower garden in Gaines ville. They were most frequently encountered on Spanish Needles, Bidens alba , the most commonly mentioned nectar source in field guides. Ger berg and Arnett (1989) reported them on Bidens pilosa , and Prunus americana . Emmel (1997) recommends planting composites, Asteraceae, to attract them to butterfly gardens. The most interesting records are from Clay Black (pers. com.) who found t hem nectaring on Coastal Plains Willow, Salix caroliniana , in February, on Saw Palmetto, Serenoa repens , and Sparkleberry, Vaccinium arboreum , April/May, and on milkworts, Polygala spp., and cultivated Cat Whiskers, Orthosiphon stamenis , September/October in Hernando Co. near the southern extent of the hairstreakÂ’s range. This spectrum of different flower shap es and families suggests M. g. sweadneri has eclectic taste in nectar. Though the butterfly was never seen on flowers in several study sites, the predominant blossoms Frog Fruit, Phyla nodiflora , ElephantÂ’s Foot, Elephantopus elatus , Blazing Star, Liatris spp., and Small Hop Clover, Trifolium dubium found nearby may have been used. Because of the long flight season encompassing the life cycles of several flow ering plants, this butterfly is of necessity a generalist nectarer.

PAGE 162

146 During long afternoon sessions in the mating cage, both M. g. gryneus and M. g. sweadneri have lit on my face and neck to sip perspiration. Also during captive breeding experiments, Jaret Daniel s and I have seen the butterfly briefly explore a cedar leaf with its proboscis, but I have not seen liquid on the cedar leaves. This behavior was not recognized in situ , though it could have easily been overlooked with even the best binoc ulars. Finally, on one occasion an individual was seen on mud. This was la te afternoon, at 16:35, on 25 February when shadows were long at WPT 003 in Dixie Co. Habitat Description As all of the afore-cited field guides suggest, Mitoura g. gryneus and M. g. sweadneri may be found on cedar trees. A popular guide for Florida even defines habitat as “Almost any areas with red c edar and nectar sources for both spring and fall broods” (Glassberg et al ., 2000, p. 71). Males ar e probably the most often seen because of their obvious perching and whirling behavior around specific lekking trees. However, in my study, single perching butterflies were sometimes found only twice in four years (WPT 021), or butterflies of both sexes seen in only two broods during one season of the four year survey (WPT 023). At first, I considered any site “habitat” if I observed the butterfly there. After evaluating notes from the fi rst season, I redefined the qualification as butterflies present in two consecutive broods and, late r on in the study, to butterflies present in two or more years. Singletons may be hopeful dispersing colonizers. An ephemeral colony may be established in a cedar site that harbors the butterfly for part of a season but lacks some quality needed to sustain a year-round population. Likewise, a specific cedar gr ove may improve to the quality that

PAGE 163

147 sustains a stable M. g. sweadneri population for years if colonized, or it may deteriorate through natural succession, natural disaster, or anthropogenic causation/manipulation so t hat the butterflyÂ’s specific micro niche requirements are lost. To determine the essential elements for good SweadnerÂ’s Hairstreak habitat, I compared field data, habitat notes, and tree sketches, gathered from naturally occuring cedar sites with ephem eral sightings and from groves with stable/successful colonies. The single obvious factor in all continuously inhabited cedar breaks is the presence of some conical cedars with their lowest boughs spreading just above and parallel with/to the ground. These trees can be along the edge of a cedar grove, on the margin of a pine plantation, along a water course, at the extreme edge of deciduous woods with no overarching leafy overstory, or in a per fectly straight, planted row of a shelterbelt. Obviously, if this were the only limiting factor, we should see Mitoura g. sweadneri along a number of county roads in 40 Florida counties and indeed, many of the 90 sites I surveyed had some conical cedars. This visual clue led to more detailed investigation. Natural cedar groves As noted previously, the seeds for natur al cedar breaks are distributed by birds so the seedlings may sprout under the limbs of any tree in which birds perch. While light shade is beneficial for the first growing season in silviculture (United States Forest Service, 1974), Juniperus silicicola is classed as intolerant to very shade-intolerant (Sargent, 1922) , which would explain why the trees occur frequently as edge habitat along margin s of taller, older pine trees. Under

PAGE 164

148 heavy overstory, like oaks, cedars that survive grow spindly, but along the margins of pine forest with good sun expo sure, they can form long lines and assume the conical form typical of the species, as in Figure 2-39 shot along CR 361 in Dixie Co. Notice that the dark green cedar in the right of the photograph is the farthest from the line of pines and it has the widest basal spread. Figure 2-39. Naturally occuring edge habitat A dramatic example of t he genesis of a non-linear grove is in an old pasture near Cross Creek in Alachua Co. By 1850, when most of the commercially valuable Southern Red Cedar had been eliminated as an overstory species in north central Florida and the pine forest had been cut over, much of what was left standing were large Live Oaks and associated trees. The heavy shade of these trees presumably prevented healthy cedar gr oves from reclaiming territory across their natural range (Wilhite, 1990). On Zane HoganÂ’s ranch near Cross Creek, there are three old oaks with developing th ickets of young cedars beneath their

PAGE 165

149 canopies. Figure 2-40 shows about th ree dozen of these young cedars in patches of sunlight beneat h one of them. The other two oaks each shade several dozen cedars, some as tall as 1.5 m. Mo st are stunted and scrawny with very few or no lower limbs. However, 60 m away is the huge skeleton of an oak with a Figure 2-40. Young cedars under oak Figure 2-41. Naturally occuring cedar grove

PAGE 166

150 grove of 22 thriving cedars below (Figur e 2-41). When this oak died in 1990-91 (Zane Hogan, pers. com.), the cedar s were still young enough to spread their lower limbs, some within 17 cm of the so il. As these trees grew, unchecked by the oakÂ’s shadow they assumed the coni cal conformation. These trees were about 15 years old when this picture was tak en. The tallest of them is 6.5 m and already 3.87 m at its widest point. The dense shade beneath such a cedar is even better at retarding the growth of co mpetitive plants than that of an oak. As these cedars continue to flourish, their lowest limbs will begin to interlace with those of their nearest neighbor s, as with the three cedars in the right foreground, and their combined shade will virtually guarantee a cedar-domin ated overstory. In fact, this shade is so deep that new cedar recruits will not add to the size of the grove unless they are planted by birds per ching in nearby non-cedar trees at the perimeter. The shade benefits t he cedarÂ’s shallow roots by protecting them from the drying effects of sun and wind, and a th ick layer of duff consisting of fallen cedar leaves further enhances moisture retention. Duff and the cedar microclimate In each generation the butterfly pupates in, and the progeny of the final flight of the season overwinters in, the duff layer beneath the larval host tree. In in vitro experiments, the wandering final in stars burrowed down through 2.5 cm of duff and butterflies emerged from pupae buried under 3-3.5 cm. I think the secret for a stable colony with the butte rfly present year-round lies in the quality of the duff layer and the temperature-miti gating effect of the cedar duff in the microclimate maintained under cone-shaped cedars.

PAGE 167

151 I first measured this effect on the last day the butterflies were seen flying, 6 October, of the 2002 flight season. The ambient temperature was 33.9°C (93°F) at 12:57, and was still ov er 33°C (91.4°F) at 15:52. Household thermometers were placed on top of the duff at the bas e of a cedar tree with no lower limbs, attached outside of the same tree in the shade near the end of a branch, and in the deep shade at the base of a cedar wit h all its ground level limbs intact. Temperatures were recorded at thr ee, 30-min intervals. These initial measurements showed a difference of as much as 2 C° between the ambient temperature outside the tree and the cool er shade at the bas e of the limbless tree, and a differential of 2. 5°-3.5 C° between ambient temperature and that in the deep shade of the cedar with limbs intact. More sophisticated gear was required to measure and record temperature parameters of the c edar microclimate over time. I used four Hobo Temp data loggers, model H01-001-01, by Onset Computer Co rporation of Bourne, Massachusetts. These lithium battery-power ed devices are 6 x 4.8 x 2 cm, record 1,800 measurements per r un with +/0.4°C accura cy at 25°C, and have a resolution of 0.1°C (0.18°F). They are launched and the data retrieved with Onset’s BoxCar software. When set fo r an 8-day run, they record the temperature at 225 intervals of 6 min 24 s every 24 h, and at 150, 9 min 36 s intervals per 24 h for a 12-day run. Th is offers detailed tracking of the temperature changes at different places in the microclimate of the cedar. The Division of Forestry site, WPT 019, at Hollister, Putman Co. was chosen for this study because its short di stance from my lab allowed for quick

PAGE 168

152 placement of the Hobos after launching at my computer. It had SweadnerÂ’s Hairstreak present for every brood sinc e its discovery, and its location on the grounds of a state office and employee resi dences offered some protection from vandalism or loss of the valuable field equipment, at least from human beings. Much to my chagrin, the presence of the fo resters, their children, and their dogs were completely ineffective at dissuadi ng nocturnal marauding opossums and raccoons in this rural environment. Two cedars of near the same age and size were needed for the temperature data I compared. The Du ff tree (Figure 2-42 A), had perching hairstreaks on eight site visits and was referenced as the 16th tree in the Division Figure 2-42. Cedars measured for microc limate Three Hobos were placed on the Duff tree, A), one at t he base of B) Limbless.

PAGE 169

153 of Forestry lane trans ect, ID# 2F-19-25-Mar-2000. It was a seedling planted as part of a shelterbelt between State Hig hway 20 and foresterÂ’s housing in 1967 (Frank Cone, pers. com.). It stood 9 m tall; it was 5.7 m wide at widest branch spread with a trunk diameter of 28.97 cm at chest height, and had basal branches within 15 cm of the ground. It was 36 years old when photographed.The trunk of a pi ne standing 7.5 m north of it is visible through the cone of cedar. The Limbless tree (Figure 2-42 B), is the 27th tree of ConeÂ’s East Side transect. It was 10 m tall, 7.2 m wide at its lowest limb, 257 cm above the ground, and 31.83 cm in diameter. It was planted in the early 1960s and its lower limbs were cut in the late 1980s. It was chos en as representative of cedars without lower limbs typical near houses and along driveways in urban settings, and occuring with deciduous species at natural sites. The duff layer within 80 cm of the base of this tree was 4-4.5 cm deep wit h a loose topping of cedar mixed with deciduous leaves. This forms a slight mound in the photograph with thin grass and shade pioneers. Eight meters behind this tree was a north-south row of cedars planted as seedlings about 2. 5 m apart in 1972 by Frank Cone. Though SweadnerÂ’s Hairstreak was not observed pe rching in this tree, it stood 8 m east of marked lekking cedars in the row and the Red-banded Hairstreak, Calycopis cecrops (Fabricius) was seen perching in, or flying to it on three occasions. Temperature Differential For each temperature sample, the Hobos were programmed and launched at home. In order to protect the electr onics from moisture, each was placed in a

PAGE 170

154 thin, zipper-sealed sandwich bag with most of the air withdrawn and a paper label. This bag was then wrapped around the data logger and placed inside a one quart Ziploc freezer bag with air removed. This outer bag was folded loosely around the instrument and held in place with tape. At th e site, one instrument labeled “loose duff” was placed under 2-3 cm of the l oose pinestraw-cedar topping snugly against the settled duff layer 61-89 cm east of the base of the Duff tree in shade below a basal limb 40 cm from the duff. This package was just barely visible from a meter away. To depl oy the “2 cm Duff” Hobo, the loose topping debris was removed, 2 cm of the duff layer was carefully scooped up, and the package was nestled into the settl ed duff 2-2.5 cm above the substrate. The top 2 cm duff was replaced and then co vered with the set aside topping. This Hobo was always placed 56-70 cm from the trunk under the same limb with “loose straw”. The “branch tip” Hobo wa s taped below an east-northeast oriented branch end 216-225 cm fr om the trunk and 208 cm above the ground, and, hopefully, out of direct sunli ght for the duration of the sample. This logger was to record approximate ambient temperature. The “limbless” Hobo was deployed 3667 cm east of the Limbless tree under 2 cm of settled cedar duff with mixed deciduous/cedar topping in a treatm ent similar to “2 cm Duff”. The first temperature sample was an eight-day run deployed 20-26 October 2002, 14 days after the last surviving adult s of the season were flying. By this time it was assumed that most of the fi nal flight’s progeny were finished feeding and beginning pupation. The temperature data was downloaded from the Hobos as text files and graphed with Microsoft Ex cel software to produce Figure 2-43.

PAGE 171

155 16.0 20.0 24.0 28.0 32.018:04 2:04 10:04 18:04 2:04 10:04 18:04 2:04 10:04 18:04 2:04 10:04 18:04 2:04 10:04 18:04 2:04 10:04TimeTemperature in degrees Celsius 2 cm Duff loose duff branch tip limbless B A Figure 2-43. End of flight season microclimate tem peratures 20-26 October 2002, A) coldest, B) warmest records. Point A was the coolest recorded at 06:07, 21 October w hen the branch tip temperature was 16.7°C and the 2 cm Duff, loose duff, and limbless measurements hovered around 19.4°C recor ded by the 2 cm Duff logger. The differential between the ambi ent temperature and that of the other treatments was 2.7C°. The warmest event was the nex t afternoon high of 32.3°C recorded at the branch tip at 13:28 and shows a distinct differential of 7.5C° with the 24.8°C record from 2 cm Duff Hobo. At eac h of the following peaks, the limbess recorder’s blue line follows the branc h tip changes while the green and black lines of the duff Hobos demonstrate the insula ting effect of this part of the cedar microclimate. The third deployment was of 12 days over the Christmas and New Years holidays (Figure 2-44). As the ambient te mperature dropped to its lowest, -1.0°C

PAGE 172

156 from 06:43-07:40 the morning of 29 December, point A, the 2 cm Duff went down to 7.4°C from 06:43-08:48, a di fferential of 6.4 C°. The mitigating effect of 2 cm of duff was 4.1° warmer than the 3.3°C of the loose duff, but only 2.1° warmer than the 5.3°C record of the limbless poin t. Furthermore, both under-duff treatments change more slowly than the other two. The same effect was evident at point B, 14:04-14:14, when the branch tip peaked at 24. 8°C before starting to cool rapidly -2.0 2.0 6.0 10.0 14.0 18.0 22.0 26.018:04 6:04 18:04 6:04 18:04 6:04 18:04 6:04 18:04 6:04 18:04 6:04 18:04 6:04 18:04 6:04 18:04 6:04 18:04TimeTemperature in degrees Celsius 2 cm Duff loose duff branch tip limbless A B Figure 2-44. Holiday season microclimat e temperatures, 24 December 2002 to 3 January 2003, A) coldest, B) warmest records while the black line of the 2 cm Duff H obo held a steady plateau at 17.1°C from 14:04-19:02. While the maximum differ ence was 7.7C°, holding the steady temperature for five hours while loos e duff and limbless (green and blue lines, respectively) climb up and down, demonstrat es significant mitigation in this settled duff layer.

PAGE 173

157 The following week had more cold weather and the period 10-20 January 2003 had three near-freezing and two below-fr eezing mornings (Figure 2-45). On the near freezing mornings of the 2nd, 5th, 6th, and 7th days and the sub-zero mornings of the 9th and 10th days, the temperature di fferential between the 2 cm Duff and branch tip records was at least 6.8C° and as great as 9.8C° for the coldest morning, point B, and still 6.8C ° on the warnest afternoon, point A. -6.0 1.0 8.0 15.0 22.018:01 6:01 18:01 6:01 18:01 6:01 18:01 6:01 18:01 6:01 18:01 6:01 18:01 6:01 18:01 6:01 18:01 6:01 18:01Time Temperature in degrees Celsius 2 cm Duff loose duff branch tip limbless B A Figure 2-45. Bellow freezing event II , 10-20 January 2003, A) warmest, B) coldest records The last temperature sample before eclosion began recorded the coldest morning, point A, and the warmest after noon, point B, of overwintering (Figure 2-46). At 07:29 on 24 January the branch ti p record dropped to -7.3°C while the 2 cm Duff position was 10.1° warmer at 2.8°C. The warmest pre-eclosion afternoon recorded was 30 January between 15:00-15:58 when ambient temperature was 24.0°C and 2 cm Duff was 7.7° cooler at 16.3°C. From the

PAGE 174

158 beginning of this run to the end, the 2 cm Duff recorder shows a gradual warming trend moderated by the duff layer. No butterflies of any species were seen flying on ValentineÂ’s Day, 14 February 2003, at WPT 019 in Hollister, nor at WPT 001 in St. Augustine. However, on 15 February a single harbinger of the first flight was seen at WPT 002 in Dixie Co. When I arrived in Holl ister to deploy the data loggers for -8.0 0.0 8.0 16.0 24.018:02 6:02 18:02 6:02 18:02 6:02 18:02 6:02 18:02 6:02 18:02 6:02 18:02 6:02 18:02 6:02 18:02 6:02 18:02Time Temperature in degrees Celsius 2 cm Duff loose duff branch tip limbless A B Figure 2-46. Pre-eclosion te mperatures of the cedar microclimate, 23 January to 2 February 2003, A) coldest, B) warmest records the period 28 February to 7 March, six Sw eadnerÂ’s Hairstreaks were already out. Two of the males were whirlin g over a lekking tree in the D of F lane transect and one unmated female, 03 SW32, was captured 1400 m away in ConeÂ’s East Side transect in tree #2-19-25-Mar-2002 just 8 m due west of the limbless tree. Temperatures recorded for this period (Figure 2-47) are marked by less gradual

PAGE 175

159 change from the loose duff and 2 cm Duff deployments between broad plateaus evident in the duff layer. On the coldes t night, point A, the nadir was 10.9°C from 21:28-21:57, but the black lin e of the 2 cm Duff regist ered 13.3°C for over 6 h. The highest ambient temperat ure, point B, lasted more or less from 12:02-15:04, but the 2 cm Duff layer did not reac h its height of 21.7°C until 13:18 and 10.0 15.0 20.0 25.0 30.018:06 6:06 18:06 6:06 18:06 6:06 18:06 6:06 18:06 6:06 18:06 6:06 18:06 6:06 18:06 6:06 18:06 6:06 18:06Time Temperature in degrees Celsius 2 cm Duff loose duff branch tip limbless A B Figure 2-47. Spring eclosion temperatures in the microclimate, 26 February to 7 March 2003, A) coldest, B) warmest records maintained this plateau until 18:16, three ho urs after the precipitous drop outside the microclimate of the cedar cone, once again demonstrating the mitigating influence of the cedar duff. Temperature samples for the rest of the spring and most of the summer were plagued by animals digging up and chewing the Hobo data loggers deployed in duff treatments. In addition, the changing ar c of the sun found the

PAGE 176

160 branch tip recorder in some of the tr ials distorting the ambient temperature values. Both of these disruptions produced artificial highs that invalidated data. One final chart of data from three of four Hobos represents the combined heat-mitigating effects of the cedar’s c one shape and duff layer. Figure 2-48 plots temperature ranges of up to 13.5 C° (24.3 F°) in an 8-h period (points A and B), 21.0 25.0 29.0 33.021:07 9:07 21:07 9:07 21:07 9:07 21:07 9:07 21:07 9:07 21:07 9:07 21:07 9:07 21:07 9:07 21:07 9:07 21:07Time Temperature in degrees Celsius 2 cm Duff loose duff branch tip limbless B A Figure 2-48. High summer in the microclim ate, 21-31 July 2003, A) coolest, B) warmest records during high summer. The low ambient tem perature was 20.9°C at 05:26, point A, when the 2 cm Duff line held a short plat eau of 23.6°C. The ne xt afternoon, the branch tip indicated 34.4°C from 13:55-15: 02, but the 2 cm Duff record did not reach its peak of 27.1°C until 15:59, point B, for differentials of 2.7° and 7.3 C°, respectively. The green line of the loose duff Hobo follows the black 2 cm Duff plot closely through the lows, but separat es from it at t he heat extremes. The blue plot of the limbless logger exceeds highs of the black plot by several

PAGE 177

161 degrees in the first two days of the samp le, indicating a trend that may have been even more dramatic on 28 July. Unfortunat ely, around 02:18, 24 July, a tetrapod uncovered the limbless Hobo and left it in full sun for the remainder of this sample. The narrow width of the jaws indi cated by the teeth marks on the Ziploc bag suggests an opossum rather than a ra ccoon spoiled this record and the errant blue line was eliminated fr om the rest of the chart. Duff Explored The duff layer under tree #2F-19-25-Mar-2000 was explored during the winter in 18, 25 x 25 cm quadrats by my wife and me in hopes of finding overwintering hairstreak pupae. The effort took part of three field days and amounted to 21 man-hours. In each quadrat the loose straw on top of the duff layer was measured in place and inspected by removing it to a large shallow cardboard box. The depth of settled duff was measured down to sand and searched by removing layers of ~1.5 cm a little at a time to the box where it was sifted through twenty finger s. After each quadrat was thoroughly examined, the material was replaced in reverse order as close to the original position as possible. The settled duff out near the dr ip line of a branch was 2.6 cm deep to the sand below, covered by 2 cm of loose debris consisting of fallen cedar leaves, twigs, and some pine straw from a nearby tree. By the third quadrat, 75 cm in from the drip line, the settl ed duff had increased to 3.5 cm deep beneath the 2 cm loose straw topping. At 60 cm out from the trunk, the settled duff was 4.5-5 cm deep with a 3 cm loose straw t opping, and within 36 cm of the trunk it increased to over 5 cm deep with a 4-5 cm loose overburden. In quadrats right at

PAGE 178

162 the base of the tree, the settled duff is a nearly uniform 4 cm deep with up to 10 cm of loose debris above. The underly ing sand and the organic detritus/debris layer mounded up near the base of the trunk as under the Limbless tree. The loose straw and the top 1.5-2.5 cm of duff was generally dry, becoming more moist to nearly damp at the compos ted-duff/sand interface. When it rained, the water drained away through the less degraded upper duff and was more easily absorbed through the moist interfac e to the sand and cedar roots below than it would have been on bar e, dry sand. Rainy days were unsatisfactory for this work because the damp cedar debr is adhered to fingers and would not bounce and separate in the sifting box. In the 5th quadrat, 125 cm from the drip line and about the same distance from the trunk, we found a geometrid pupa 1.5 cm down into ~3.5 cm of settled duff and another at about the same depth clos er to the tree in quadrat 6. In the next three quadrats in toward the base of the tree, we found several fragments of insect puparia and pieces of beetle and roach bodies. In the 8th, 9th, and subsequent quadrats within 75 cm of the trunk, we reco vered six smooth, nearly featurelees, capsular puparia opened at one end, 8 mm long, determined as dipteran, and one as yet undetermined ope ned pupa. All of these were in 2-2.5 cm of dry, settled duff in a layer t hat varied 4-5.5 cm thick above the sand. At the sand-compost interface were se veral small empty snail shells and an occasional isopod. A small, r ed, torpid Southern Toad, Bufo terrestris , was uncovered 2 cm into the duff in quadrat 13. A month later, one of the geometrids eclosed inside the vial and was determined to be Patalene olyzonaris puber

PAGE 179

163 (Grote & Robinson, 1867), the Juniper Geometer. This is the most common lepidopteran encountered in this and severa l other cedar sites. I have found its twig-mimic larvae in local cedar cut for colony fodder and in cedar sent from Tennessee. Importance of duff The physical attributes of the duff laye r, its insulating quality that slows change and blunts the extremes of temperature, and its ability to both drain and retain moisture suggest qualities purposefully sought by the animals found there, particularly in the stratum 1.5-2.5 cm below the loose overburden. If length of photoperiod and/or change in temperat ure induce diapause, and a second change signals the production of horm ones that break diapause in the overwintering Mitoura g. sweadneri pupa, as in most other insects (Chapman, 1969), I think the plateaus on the edges of a specific limited temperature range as in Figure 2-47 may be part of the signa l for eclosion of the first flight. The photoperiods are much the same in a given month at this latitude year after year, but the ambient temperature, and, to a moderated extent, t he changes in the cedar’s microclimate, vary with weather conditions, as does the earliest eclosion date each spring. This arrangement would help assure favorable temperatures for the eclosing butterfly, a greater chanc e of finding a mate cued by the same abiotic signal, and a more certain supply of early blooming nectar sources. If the adults of subsequent broods eclose in situ soon after pupation is complete, as in my lab at near constant 25°-26.5°C, t hen the temperature range of high summer, between 23°-27°C (black line, Figure 2-47), ma y not be as critical a signal since

PAGE 180

164 that range, begun well before the July sample was collected, continues weeks after the last flight of the season (Figure 2-43). The limits of this single experiment are obvious. Only two trees were sampled, at only one site, through only one annual cycle. However, I think that with enough data loggers, a wide enough variety of cedars with basal limbs intact in stable colony sites, and a few more seasons to gather data, the minimum parameters of tree width and duff depth capable of ma intaining the essential overwintering microclimate for M. g. sweadneri could be determined and modeled for evaluating potential habitat for t he butterfly. I also believe that with enough eyes and fingers, the pupa in the “haystack” could be found and we could learn just how deep the wanderer burrows into the duff. These elementary observations sugges t a hypothetical explanation for ephemeral colonies that last through part of a flight season, but are absent the following spring. The thin duff layer under cedar #1-23-14-Mar-2004, on the far left in Figure 2-41, could be cover enough to produce a second flight adult, but not deep enough to harbor pupae through t he temperature extremes of succeeding broods. As it grows and expands basally along with its cohorts in the grove, it could attain essential microclim ate parameters and, if recolonized, this grove could sustain Sweadner’s Hairstreak for generations. This natural, old field grove also has the advantage of the proxim ity of the pine immediately behind and to the southeast of #1-23-14-Mar-2004 in the photograph. Its crown is already several meters taller than the cedars and as it sheds needles year after year, the

PAGE 181

165 pine straw contributes substantially to the fluffiness of the duff and the openness of the loose topping under every downwind cedar. The closeness of the cedars in a gr ove or planted row may enhance the tent effect covering a larger duff area than that of a singl e tree. It is not essential that all the cedars be coni cal to the ground. The cedars in the row behind the Limbless tree, in Figure 2-42 B, had a ll had their lower limbs cut 1.65-2.23 m above the ground, yet four out of the first ten in the ro w were marked perching or lekking trees. The definitive description of habitat for SweadnerÂ’s Hairstreak is that it must have some, preferably female, cedars with enough conical volume and deep enough duff to affect th e essential microclimate. Detriments to habitat Anything that degrades the duff under an otherwise apparent host tree degrades the habitat. Anything that opens t he base dimension, or interrupts too much of the surface of an individual cedarÂ’s cone can sabot age its microclimate. Though absence of flying adults is not alwa ys a reliable indication that a habitat is flawed, oft-visited sites with conical tr ees that never held the butterfly can show obvious faults, when evaluated with t he duff/microclimate evidence. These characteristics may prohibit a hairstreak col ony from being established, or be part of a process that degrades good habitat. An extreme example was WPT 908, a pine and cedar hammock west of CR-361 an d just a few miles from five once-sterling colonies along that road in Dixie Co. This coastal hammock surrounded by Black Needlerush, Juncus roemerianus , in high salt marsh just above mean high water line had a core of pines on the highest ground in the center encircled by 68 cedars, with an addi tional 28 standing just inside the outer

PAGE 182

166 circle. Needlerush protruded from beneat h, and insinuated between the cedars. Surrounding bare substrate with too little purchase for this plant harbored small active crab borrows. Southern Red Cedar is fa irly salt tolerant (Barrick, 1979, and Sargent, 1922) and often grows along margins of tidal marshes. However, the frequently inundated zone at the margin of this hammock did not allow the accumulated build up of a dry duff laye r, though it does collect flotsam during high water events. Likewise, cedars gro wing at waterÂ’s edge along the Tolomato River, part of the Intracoastal Wate rway near St. Augustine, had Smooth Cordgrass, Spartina alterniflora , growing below them. The presence of this most seaward, nonwoody salt marsh plant disqualifies the habitat for SweadnerÂ’s Hairstreak. A more subtle disqualifier of otherwis e beautiful brood trees along creeks, ditches, and the banks of some standing water was observed even at some very productive study sites. The cedars neares t the water course sometimes have perching males, but if too much of t he expanse of their lower boughs hang over the water, the duff would not make a laye r on that side of the tree, adjacent duff may get soggy too frequently for the pupae, and the presence of standing or moving water under the cone of the cedar c ould wreck its insulating value. A few meters away on higher ground, the duff makes a thick layer under the spreading cedars that sustain the colony. Another disqualifier is the too-prox imate presence of deciduous trees. Too rich a mix of deciduous leaves accumulating in the cedar duff changes its moisture-handling characteristics and the broad planes of the fa llen leaves might

PAGE 183

167 hinder a burrowing wanderer or an eclosing adult. This condition can occur over time as natural succession where decidu ous trees overgrow the more slowly growing cedars. Certain vines are another effect of natural succession detrimental to M. g. sweadneri habitat. Muscadine Grape, Vitis rotundifolia , and Greenbriar, Smilax spp., are listed as associated species in natural forest cover type Southern Redcedar, Society of American Forest ers Type 73 (Eyre, 1980, and Wilhite, 1990). Marked cedars used by the butterfly in some study sites had Greenbriar entwined about their trunks and some had t he ubiquitous Poison Ivy. A perching hairstreak might choose a cedar with Virginia Creeper, Parthenocissus quinquefolia , on its trunk, but not after the vine grows out through the cedarÂ’s cone. The most offensive vine to the butterfly is Muscadine. A lekking tree used for years during the study at WPT 017, near the Gulf Hammock Post Office, was abandoned when Muscadine succeeded its cr own. This vine was probably a significant habitat-degrading fa ctor at two other study sites, including the type locale, WPT 001 at St. Augustine. Though most of my sites were near hi ghways, or county roads, too close proximity to heavily traveled roads ma y be detrimental to habitat due to air disturbance or internal combustion by-pr oducts of passing traffic. Several cedar sites along four-lane divided highways did not harbor the butterfly during the survey even though the tree line along thes e roads is set back far from the pavement. The busiest roadway with a pr ime colony nearby was SR-20 which had two lanes running past WPT 019 in Putn am Co. Lekking trees in one of the

PAGE 184

168 most used transects at this site were in a row planted parallel to and 26.8-27.8 m from this highway. County r oad 361 in Dixie Co. is a very rural, low traffic paved road that ends in high salt marsh al ong the Gulf. Though lots of cedar grows along the southern end of t he road, good, frequently oc cupied habitat was at least 13 m from pavement edge, on ground just higher than the surrounding Rocky Creek Swamp, and no closer than 2 mi to the southern end of the road. Cedars on the east side of small unpaved, or limerock roads in Dixie and Levy Counties are sometimes coated in fine white dust during the dry season. The butterfly was never seen perching in dusty trees. A final anthropogenic detriment to potent ial SweadnerÂ’s Hairstreak habitat frequently encountered in suburban settings and on some park lands has been suggested above, but needs to be stated succinctly. As landscape esthetics change over time, and cedars become as stately and imposing as cedar tree #2F-19-25-Mar-2000 (Figure 2-42 A), whole rows often have all lower limbs cut high enough from the ground to enable unnatural succession of turfgrasses and the mowers to maintain them. This opens up the landscape at human height for lawns, picnic tables, and parking lots and banishes the butterfly. All but eight of the 130 trees used in active M. g. sweadneri colonies were less than 12.5 m tall and only a couple of specimens 11 m or taller had basal limbs intact at ground level. Though this is a small sample, it may indicate an upper height limit for trees currently used by M. g. sweadneri . We lack records of the butterfly in virgin cedar stands of the 19th century along Apalachee Bay that were more than 30 m tall (Eyre, 1980). T he tallest naturally occurring grove I

PAGE 185

169 found, at WPT 971, Blue Springs State Park, Volusia Co., wa s at the foot of a hill and surrounded by a mix of pine and deciduous forest. The core of this grove comprised 23 major cedars. All were over 13 m tall, 14 of them were at least 17 m and some exceeded 21.5 m. None had lower limbs below 7.3 m from the ground. They formed a nearly solid canopy and in the shade below grew wide cabbage palms and a few pines near the marg in. The understory was difficult to walk through and the ground was soggy with leaves, pine straw, and fallen cedar but no standing water. It may be that cedars naturally prune lower limbs when they are heavily shaded, just as pines in a pulpwood plantation do. This suggests that there is an upper age limit, size limi t, or a point of succession, beyond which the current productive Swead nerÂ’s Hairstreak habitats will no longer sustain the microclimate essential for the butterfly. The specificity of habitat requirements for SweadnerÂ’s Hairstreak may, in the future, be more narrowly defined, but these primary revelations of my field study already offer a partial explanation fo r the butterflyÂ’s listi ng as rare (Emmel, 1993), very rare (Kimball, 1965), lo cal and uncommon (Daniels, 2003, and Glassberg et al ., 2000). For SweadnerÂ’s Hairstreak to continue to inhabit its current range, it must have new cedar groves growing to replace productive habitat degraded through natural succession, natural disaster, and anthropogenic disturbance. Direct habitat modification and loss to rapid development, particularly along FloridaÂ’s coastline, exacerbates chronic loss to natural succession.

PAGE 186

170 CHAPTER 3 PAST AND PRESENT DISTRIBUTION The historic range and current distribution of Mitoura gryneus sweadneri are inextricably linked to taxonomic decisions made by each author publishing an account of the taxon. Even though the specific name, Mitoura sweadneri , was not designated for the Florida butterfly until F. H. Chermo ck named Dr. Walter SweadnerÂ’s specimens (Chermock, 1944) , the butterfly has been known and written about for well over 100 years. With each taxonomic change, almost all of the historic authors have given verbal de scriptions of the butterflyÂ’s range, but only a few have drawn maps. A review of the most widely circulated maps published since the genus was described 13 5 years ago is valuable taxonomic history and reflects the accumulated kno wledge of the range of the butterfly to date. Historic Range When Scudder (1869) described the new genus, Mitoura , based on the type species Thecla smilacis , he gave its range as New England to Florida. Twenty years later, he elaborated on th is American genus, calling it Mitura, represented by the single species, Olive Hair Streak, Mitura damon , in the eastern United States with its northern lim it in New England (Scudder, 1889a). His was the first map of the taxon and it is es pecially significant to this current study because he had seen a single specimen from the south (no details) with uniformly brown dorsal wing surfaces, lacking the reddish tint of his familiar New England butterfly

PAGE 187

171 and with a “longer tail of the hind wings 5.5 m long, or fully twice the length of the average northern specimens” (Scudder, 1889a p. 863). It is also significant that he identifies the butterfly’s host with t he current taxon of the host of the Figure 3-1. Oldest range of the butterfly After Scudder (1889b) Florida butterfly, Juniperus virginiana and that his map is a close approximation of the combined maps from Little (1971), illustrated previously as Figure 2-28. The noticeable difference is that Scudder’s butterf ly map lacks the presumed virginiana silicicola gap. When Chermock (1944) described Mitoura sweadneri , he only specified the type locale as St. Augustine, Florida. Klots (1951) called the butterfly a subspecies of M. gryneus , and limited its range to Florida, but neither he nor Mitura damon

PAGE 188

172 Clench (1961), who rejected the nomen, sweadneri , for Callophrys gryneus smilacis and gave central Florida as its range, drew a map. Howe (1975) followed ClenchÂ’s taxonomic lead and de scribed its range as central and northern Florida. Opler (Opler and Kriz ek, 1984) referred to the subspecies, M. g. sweadneri , as occurring in Florida, but drew a confusing range map (Figure 3-2). First, Opler left a big gap in distribution between the inte rior U. S. range of Figure 3-2. Ranges after Opler and Krizek (1984) M. gryneus and the more southerly coastal range of M. g. sweadneri ; second, he left a wide gap separating Gulf coast and At lantic coast ranges within Florida, and he extended the range up the Atlantic coast to South Carolina. This extended Atlantic range appears to cont radict his stat ed range limit of M. g. Mitoura gryneus M. g. sweadneri

PAGE 189

173 sweadneri . Interestingly, he said Juniperus virginiana is the only natural host, although his book was published nine years after LittleÂ’s government maps. Kurt Johnson, who championed the taxonomic signi ficance of the butterfliesÂ’ use of particular host trees, drew range m aps (Johnson, 1976, 1978, and 1980) for C. ( M ) gryneus and C. g. sweadneri that are much closer to the one Scott (1986) drew for these two subspecies (Figure 3-3) , except that Johnson did not consider the butterflies of the Atlant ic coast of Georgia to be sweadneri , citing Lucien Harris, Jr.Â’s (1972) authoritative obser vation that the dark summer form of C. g. gryneus found in Georgia was synonymous with C. g. gryneus f. smilacis . Figure 3-3. Ranges after Scott (1986) ScottÂ’s written description of the range of C. g. sweadneri lists only Florida as the distribution of the subspecies and ye t, he too drew an extended range to the C.g. gryneus C. g. sweadneri

PAGE 190

174 middle of coastal South Carolina. He acknowledged J. silicicola as the host for SweadnerÂ’s Hairstreak. The last historical map is Paul Op lerÂ’s refined version (Opler and Malikul, 1992) in Figure 3-4. By the time Ople r drew this map he had been compiling Figure 3-4. Refined historic ranges after Opler and Malikul (1992) records of the Florida butterfly by county for several years. Most of these, at least through 1983, were gathered directly from data originating in pinned specimens (Harry Pavulaan, pers. com.). This time, he changed the form of gryneus to grynea , but kept his former designation of t he Florida subspecies. He also edged the light gray range of the nominate spec ies farther inland from the Gulf of Mexico and the coast of South Caroli na and modified the northern reach of M. Mitoura grynea M .g. sweadneri

PAGE 191

175 grynea considerably. More importantly, he sa id that SweadnerÂ’s Hairstreak is found in peninsular Florida and redrew its range. First, he filled in the gap between FloridaÂ’s coasts suggesting he had f ound records from interior counties. Second, he drew the dark gray farther south of GeorgiaÂ’s border and extended it along the Big Bend region of the Gulf coas t. Finally, he shortened the butterflyÂ’s range along the Atlantic coas t of Florida but extended the dark gray region more northerly into S. Carolina as well as broadening its swath along the Atlantic coast. About 1988, he passed his distribution files to Harry Pavulaan for inclusion in the data base of the United St ates Geological Survey (USGS). Current Distribution At the beginning of my field research, I had planned to verify the current range of SweadnerÂ’s Hairstr eak on the ground, thinking that establishing a northern and southern limit on each Florida coast would de fine its distribution. An unforeseen problem was that lots of healthy cedar trees grew along the coasts but were not inhabited. Sometimes beautiful cedars grew within sight of occupied habitat, but as I now realize, they lack ed essential elements for the butterfly. Hunting down specific locations re commended by experienced collectors and researchers from the Souther n LepidopteristsÂ’ Society was frequently frustrating. Some cedar groves that were producti ve a few seasons before had morphed into strip malls, lost limbs to lawns, or simp ly been overgrown. On the other hand, my diligent search pattern turned up colonies in counties not mentioned by any of these collectors. John Calhoun (1996) published a Flor ida map with 19 counties where M. g. sweadneri had been collected, and one county where a colony of either M. g.

PAGE 192

176 grynea , or a possible intergrade, had been taken. Calhoun generously shared with me 33 records he had comp iled from reports in the literature or from other collectors. Only four of these records we re from the 1960s. Twelve were from the 1970s, and 17 were from 1981-89. He also provided 20 of his own personal records from the 1990s plus six of his re cords from the late 1990s from a site near Marianna in Jackson Co. which harbors a nonsweadneri phenotype. He told me of a report from Liberty Co. that may have been of an intergrade, but cautioned that he had not seen the specim en. For the past several years, Calhoun has compiled and coordinated recent Florida records for Pavulaan to include in the USGS data base. The USGS map (http://www.npwrc.usgs.gov/resource/distr /lepid/bflyusa/fl/294.htm) currently administered by Pavulaan includes five counties missing from CalhounÂ’s map and leaves out records from Marc and Mari a Minno from Camp Blanding in Clay Co. Several of the USGS county re cords absent from CalhounÂ’s map are pre-1988 and attributed to Paul Opler (P avulaan, pers. com.). Two remaining county records are unpublished data from Dave Baggett, and one, Washington Co., is from a single specimen collected by Ray Stanford in 2000 (Calhoun, pers. com.). To all these verified records I added the label data from specimens in the Florida State Collection of Arthropods and from records of other researchers cited in Chapter 2, plus my own sightings to make the Florida map of Figure 3-5. Dixie Co. was conspicuously missing from CalhounÂ’s map and listed as an unconfirmed or dubious record by USGS. Ei ght of my study sites were in Dixie Co. Though I searched for SweadnerÂ’s Hairstreak in 18 counties and only

PAGE 193

177 Figure 3-5. Current known distri bution of SweadnerÂ’s Hairstreak

PAGE 194

178 found it in seven of them, I consider Fi gure 3-5 to be an accurate chart of the current known range of the butterfly. I hav e also seen cedars that should provide good habitat in Gilchrist Co. which abuts five counties with good, stable colonies and in September 2005 M. g. sweadneri was photographed in O’Leno State Park adding Columbia Co. to the list of 27 Flor ida counties with records of Sweadner’s Hairstreak. Numbers The information traditionally inclu ded on the label of a pinned specimen gives date and location about the individual . Without additional field notes these data are of minimal value in guessing how many butterflies of a particular species a collector saw on that date, in that lo cality. There are no indications about the “the ones that got away”, specimens tak en in too poor condition to mount, or ones damaged in spreading that may have been discarded. Likewise, individual entries in data bases like USGS, or sight records in season summaries published by organizations of butterfly enthusiasts are, at best, valuable as records of the species’ presence. With Sweadner’s Hair streak, the problem of assessing such data is compounded by the fact that the butterfly perches in cedar trees very often beyond the reach of even the tallest collector without t he proper equipment and is rarely seen at flowers. Unfortunately, very few of the sp ecimen labels from researchers and butterfly collectors compile d for my study indicated whether the butterfly was collected on a host cedar, at a nectar source, or a puddle. Records from Other Researchers The first 33 entries in Table 3-1 are from specimen label data compiled from the collection of the FSCA. The last 12 ent ries are from private collections in

PAGE 195

179 Gainesville. Of the 45 entries in this table, 28 are of single specimens taken on 28 days, and only three entries (in bold, Ta ble 3-1) represent more than three specimens collected on a single day. If, for the sake of comparison, I assume that all these collectors are equally skilled netters and preparators, then the numbers of specimens from the same date could, to some extent reflect the abundance of the hairstreak on that day. This is not really a farfetched assumption because even the most skilled bu tterfly catcher would be hard put to collect every individual from among the c edars in a closely spaced grove. I think the entries in boldface represent unusually abundant days for Sweadner’s Hairstreak. This hypothesis is reinfo rced because two of the big numbers are from 14 entries representing the work of C. F. Zeiger and the re st of his entries are all of three or less captures. The big day, 26 June 1983, is also from a collector, L. C. D., who only c aptured small number s on other days. The data in Table 3-2 is from specim ens in the collectio n of the McGuire Center for Lepidoptera and Biod iversity at Florida Museum of Natural History in Gainesville. The first 12 entries are from label data of specimens collected by T. C. Emmel and his field assistants P. J. Elia zar, K. Weisner, J. L. Nation, Jr., D. Smith, and M. Minno over several y ears. These skilled netsmen were not recreational collectors but ra ther took specimens for Dr. Emmel’s long-term study of variations in the chromosomes of butte rflies from around the world. Ten of the 12 entries are from locations close to St. Augustine; “-LH” abbreviates Lighthouse for the historic St. Augustine light which is today a museum and popular tourist destination. Anastasia Is land is the geographical location of the

PAGE 196

180 Table 3-1. Specimen label data I Date Year locale County # sex Collector 15-Jun 1921 St.Augustine St.Johns 1 m F.M.Chermock 12-Mar 1979 St.Augustine St.Johns 2 m R.W.Bosco 21-Jun 1948 Daytona Volusia 1 f H.L.King 20-Jun 1948 Harbor Oak Volusia 1 f H.L.King 12-Sep 1975 New Smyrna Bch Volusia 1 m G.W.Rawson 10-Apr 1964 Ft. George Duval 1 m C.F.Zeiger 28-Jul 1963 Ft. George Duval 3 2m, 1f C.F.Zeiger 1-Sep 1962 Ft. George Duval 1 f C.F.Zeiger 20-Feb 1982 Ft. George Duval 2 1m, 1f C.F.Zeiger 17-Mar 1979 Ft. George Duval 1 f C.F.Zeiger 7-Mar 1964 Ft. George Duval 4 3m, 1f C.F.Zeiger 25-Mar 1978 Ft. George Duval 2 m C.F.Zeiger 6-Apr 1963 Ft. George Duval 1 f C.F.Zeiger 29-May 1964 Ft. George Duval 3 m C.F.Zeiger 28-Jul 1963 Ft. George Duval 10 5m, 5f C.F.Zeiger 1-Sep 1962 Ft. George Duval 2 m, f C.F.Zeiger 7-Sep 1981 Ft. George Duval 3 f C.F.Zeiger 17-Jul 1979 Ocala NF Marion 1 f C.F.Zeiger 8-Mar 1981 Ft. George Duval 3 2m, 1f C.M.Stevens 31-Mar 1963 Ozello Citrus 3 2m, 1f H.L.King 16-Apr 1967 Ozello Citrus 3 2m, 1f H.L.King 4-Apr 1964 Ozello Citrus 1 m H.L.King 31-Mar 1963 Ozello Citrus 1 m H.L.King 10-Apr 1967 Ozello Citrus 1 m H.L.King 8-May 1975 Istachatta Hernando 1 f H.L.King 2-May 1975 Yankeetown Levy 1 f L.C.D. 26-Jun 1983 Yankeetown Levy 6 2m, 4f L.C.D. 16-Jul 1983 Yankeetown Levy 1 f L.C.D. 12-Aug 1984 Yankeetown Levy 3 2m, 1f L.C.D. 16-Jul 1983 Crystal River Citrus 1 m L.C.D. 12-Aug 1984 Crystal River Citrus 1 f L.C.D. 13-Sep 1980 Yankeetown Levy 3 1m, 2f C.F.Zeiger 7-Oct 1979 Yankeetown Levy 1 f C.M.Stevens private collections 3-Sep 1993 Alachua 1 m J. Slotten 26-Mar 1978 Yankeetown Levy 2 m J. Slotten 25-Jun 1978 Yankeetown Levy 1 m J. Slotten 13-Jun 1980 Yankeetown Levy 1 m J. Slotten 8-Mar 1980 Yankeetown Levy 1 m J. Slotten 26-Mar 1978 Yankeetown Levy 1 m J. Slotten 25-Sep 1977 Hernando 1 f J. Slotten 20-Sep 1997 CR-361 nr Jena Dixie 1 m Hugo Kons, Jr. 25-Sep 1999 CR-361 nr Jena Dixie 1 f Hugo Kons, Jr. 26-Mar 2000 CR-361 nr Jena Dixie 1 f Hugo Kons, Jr. 24-Sep 2000 CR-361 nr Jena Dixie 1 f Hugo Kons, Jr. 28-Aug 1999 Gulf Hammock Levy 1 f Hugo Kons, Jr.

PAGE 197

181 Table 3-2. Specimen label data II Date Year locale County # sex Collector 17-Mar 1972 Anastasia Is. St.Johns 13 9m, 4f T.C.Emmel et al . 12-Apr 1982 St.Augustine LH St.Johns 2 m T.C.Emmel et al . 12-Apr 1983 St.Augustine St.Johns 6 2m, 4f T.C.Emmel et al . 12-Apr 1983 St.Augustine St.Johns 6 4m, 2f T.C.Emmel et al . 1-Apr 1984 Matanzas River Br St.Johns 3 m T.C.Emmel et al . 1-Apr 1984 St.Augustine LH St.Johns 8 4m, 4f T.C.Emmel et al . 1-Apr 1984 St.Augustine PD St.Johns 3 m T.C.Emmel et al . 3-Mar 1985 St.Augustine LH St.Johns 5 m T.C.Emmel et al . 9-Apr 1987 UF campus Alachua 3 m T.C.Emmel et al . 2-Apr 1988 Yankeetown Levy 5 m T.C.Emmel et al . 3-Apr 1988 St.Augustine LH St.Johns 1 m T.C.Emmel et al . 3-Apr 1988 Anastasia Is. St.Johns 1 m T.C.Emmel et al . other specimens 7-Mar 1976 New Smyrna Bch Volusia 2 m 20-Sep 1976 New Smyrna Bch Volusia 2 f 13-Apr 1980 Homosassa Citrus 8 6m, 2f B.Hollister 28-Mar 1986 Liberty 1 m M.Minno 23-Aug 1991 Silver Glen Sprgs Marion 1 m M.Minno 25-May 1991 UF campus Alachua 1 m M.Minno lighthouse shared with Anastasia State Park, and the Matanzas River Bridge connects Anastasia Island with mainland Fl orida just south of the city. The notation, “-PD”, is for the grounds of St. Augustine Beach city police department located 3 mi south of the St. Augustine Lighthouse. The last six entries in Table 3-2 are from the same colle ction but are not from the chromosome study. The dates in boldface type may represent days of unusual abundance of Sweadner’s Hairstreak. As stated in Chapter 2, Emmel (1993) found that isolated hairstreak coloni es were represented in small numbers of 12-50 adults so, 17 March 1972, 1 April 1984, and 13 April 1980 were probably dates with numbers near the higher end of Emmel’s observation. All of these records are of pinned specimens and so, at best, probably represent a subset of the butterflies flying on those field days.

PAGE 198

182 My Field Records Records of the current study fr om 6 February 1999 to 8 May 2004 are presented in Table 3-3. These data, arr anged in chronological order, leave out the many field trips when the butterflies were not seen. Each waypoint (WPT) entry represents at least one full survey of the transect established within the first few visits after an inhabited site was di scovered. Only a couple of the surveys were interrupted by rain and then only w hen minimal numbers were flying. When several consecutive surveys of a recorded site revealed M. g. sweadneri was no longer present, its status was changed to ephemeral. These ephemeral sites were visited less frequently so more fiel d time could be devoted to recording behavior in stable colonies. Sites ob literated in human e ndeavors were dropped from the list when no cedars were left standing. However, marked sites damaged by fire, or degraded through other natural causes were resurveyed periodically to determine the likelihood of their recovery. It is important to restate that all t he records of Table 3-3 are of total butterflies seen. These include the 70 or so individuals captured, most of which were released. In only ten of the 109 site vi sits (boldface in Table 3-3) were 11 or more adults sighted. Only 1-2 individual s were observed on 53 site visits, and on 44, or 40% of all visits, 3-9 were count ed. It is interesting to note different numbers of individuals at different site s on the same day. For instance, on 24 March 2003 three sites in Dixie Co. were surveyed; one had one and one had three adults flying but WPT 002 had a big day record of 11. While the sheer number of trees at a site can directly a ffect numbers of adults using the grove

PAGE 199

183 Table 3-3. My field study sight records* Date Year WPT County -Sight -Sight SIGHTED 27-Feb 1999 003 Dixie 1 1 2 27-Feb 004 3 3 20-Mar 006 Taylor 14 14 20-Mar 007 Dixie 2 1 3 3-Apr 001 St.Johns 5 5 3-Apr 008 Flagler 1 1 10-Apr 003 Dixie 8 3 11 10-Apr 006 Taylor 7 7 14-May 012 Alachua 3 3 16-May 012 1 1 6-Jun 001 St.Johns 1 1 13-Jun 001 1 1 18-Jun 004 Dixie 1 1 2 18-Jun 007 1 1 2 18-Jun 002 1 1 22-Jun 013 St.Johns 1 1 22-Jun 001 1 1 26-Jun 014 Dixie 1 1 5-Jul 014 3 3 9-Jul 015 Dixie 3 1 4 16-Jul 016 Levy 2 2 28-Aug 017 1 1 28-Aug 018 8 8 4-Sep 008 Flagler 1 1 25-Sep 007 Dixie 1 1 25-Sep 002 1 1 25-Sep 003 1 2 25-Sep 004 2 2 19-Feb 2000 017 Levy 6 6 25-Feb 003 Dixie 1 1 25-Feb 007 2 2 9-Mar 019 Putnam 5 5 25-Mar 019 5 7 2-Apr 002 Dixie 3 3 2-Apr 007 1 1 3 10-Apr 007 1 1 2 10-Apr 002 1 1 10-Apr 003 1 1 3 30-Apr 007 1 1 7-Jun 003 1 1 7-Jun 007 1 1 17-Jun 007 1 1 17-Jun 003 2 2 3-Jul 003 1 1 3-Jul 020 2 3 5 * Year and County columns om it consecutive repeats.

PAGE 200

184 Table 3-3. Continued* Date Year WPT County -Sight -Sight SIGHTED 16-Feb 2001 007 Dixie 1 1 16-Feb 020 1 3 16-Feb 002 1 1 3 28-Feb 007 3 3 28-Feb 002 4 4 28-Feb 003 6 7 28-Feb 004 4 5 2-Mar 002 7 7 14-Apr 020 1 1 28-May 002 6 7 28-May 020 3 3 14-Jun 020 4 4 14-Jun 002 1 1 30-Jun 002 4 9 30-Jun 020 2 2 5 1-Feb 2002 007 1 1 1-Feb 002 2 2 1-Feb 020 1 1 25-Mar 019 Putnam 17 1 18 20-May 021 Alachua 1 1 21-May 002 Dixie 4 2 6 21-May 007 1 2-Jun 020 2 2 2-Jun 002 1 1 30-Jun 020 5 5 30-Jun 002 3 2 8 30-Jun 007 1 1 24-Aug 002 1 2 24-Aug 007 2 1 3 2-Sep 019 Putnam 20 1 31 22-Sep 019 33 7 44 29-Sep 019 9 3 15 6-Oct 2002 019 3 3 15-Feb 2003 002 Dixie 1 1 26-Feb 019 Putnam 5 1 6 10-Mar 019 4 4 19-Mar 002 Dixie 6 6 19-Mar 007 3 3 24-Mar 007 3 3 24-Mar 020 1 1 24-Mar 002 8 11 4-Apr 002 4 4 13-Apr 019 Putnam 1 1 2 14-Apr 002 Dixie 1 1 7-May 019 Putnam 2 2 10-May 019 2 2

PAGE 201

185 Table 3-3 Continued* Date Year WPT County -Sight -Sight SIGHTED 24-May 019 Putnam 11 12 29-May 2003 018 Levy 1 1 29-May 002 Dixie 5 5 29-May 007 2 2 31-May 019 Putnam 4 1 5 10-Jun 019 7 7 22-Jun 019 2 2 29-Jun 002 Dixie 4 4 2-Aug 019 Putnam 1 1 16-Aug 019 13 1 15 28-Aug 019 15 3 18 28-Sep 019 2 2 15-Oct 2003 019 3 3 3-Mar 2004 023 Alachua 1 1 7-Mar 019 Putnam 4 4 14-Mar 023 Alachua 1 1 2 17-Mar 023 2 2 20-Mar 023 2 2 71 Days Totals 380 48 472 * Year and County columns om it consecutive repeats. this is not always the case. Seven site visits on 16 and 28 February 2001 (top of second page of Table 3-3) show WPT 007 which only had a single lekking tree, compared quite favorably with WPT 002 that had several. Two more aspects of Table 3-3 reveal subtle but significant details of the butterflyÂ’s distribution. First, days of high abundance may fall at the beginning or near the end of a flight season but 30 June of 2001 and 2002, and 24 May 2003 also show relatively large numbers and these are mid-season dates. This suggests that total numbers at specific sites do not rise and fall in an orderly pattern each year but fluctuate from fli ght to flight. Second, the numbers in the sighted column are frequently greater t han the totals of males and females observed which may in part reflect the diffi culty of judging t he sex of butterflies in situ . However, even if all the undetermined individuals were actually female,

PAGE 202

186 observed numbers of females are fa r less than the 50-50 sex ratio of Mitoura eclosed in captivity (Chapter 2). This re inforces my observation that females are less visible in a cedar grove and may ac tually be the primary dispersers. The capture records from Hugo Kons (Table 3-1) add another clue. On many days in the field with Hugo, at every site visited he always concentrated on nectar sources where the greatest diversity of Lepidoptera would be taken. Four out of five of his captures of M. g. sweadneri at flowers were females. It would be interesting to stake out a lush stand of Bidens near an active sweadneri colony from dawn to dusk for several days during the height of a flight and capture and mark every individual that came to nectar. This could reveal, as one would suspect, that both sexes visit flowers wit h the same frequency, but it might show that the sexes divide this resource tempor ally just as they use the host trees at different heights. Comparison and Discussion EmmelÂ’s field teams had a specific objective in seeking M. g. sweadneri for the chromosome study and so may have c oncentrated on bringing in numbers of this taxon. It would be helpful to know the dates they returned empty-handed. We do not know the objectives of the other collectors lis ted in Table 3-1 who may have been sampling the general diversity of Lepidoptera. In any case, their success depended on opportunity to capture the butterfly. Based only on the numbers of specimens in Tables 3-1 and 3-2, I consider 10% of the 61 entries to represent big days for local populations of the hairstreak. Of my own 109 site visits with carefully counted sight re cords, I consider 9% to have been big numbers. Big numbers observed during boom flights are the ex ception rather

PAGE 203

187 than the rule in my field study, and in t he literature. Scudder in his original life history of Mitura damon said it occurred “seldom in any great abundance” (Scudder, 1889a, p. 865). Klots acknowl edged the fluctuating numbers within M. gryneus colonies when he said this butte rfly is “sometimes not uncommon” (Klots, 1951 p. 141). Calhoun observ ed the same phenomenon when he described M. g. sweadneri populations as extremely localized but remarked that it could be plentiful at some loca tions (Calhoun, 1996), and Emmel (1993) had witnessed isolated colonies that someti mes numbered 50 individuals. I believe that boom flights occur when conditions ar e right in a specific location and that these flights produce dispersers that ma y consequently establish new colonies when they encounter good habitat. Low numbers alone may bear little rela tion to the overall health of a population. I think this species has adapt ed to survive in low numbers, due to specific, narrowly defined habitat requirement s. For example, lekking as the mate location adaptation could be predicted fo r this butterfly because it has a long breeding season, the operational sex rati o is heavily skewed, and the essential reproductive resources are limited (Eml en and Oring, 1977, and Bradbury, 1981, 1985). Sweadner’s Hairstreak biology is a model of these crit eria: males are seen perching for weeks at a time on spec ific trees, females are much less conspicuous in the habitat, and only certai n cedars satisfy the requirements for brood trees. Having a hardwired mate loca tion system is essential for such a small butterfly living in such big cedar trees in low population densities.

PAGE 204

188 Several authors have referred to t he butterfly as frequently found (though not necessarily/exclusively) in coastal habita t in either verbal descriptions of its distribution, or range maps (Calhoun, 1996, Daniels, 2003, Emmel, 1993, and Opler, 1984). Of the 61 entries in Tabl es 3-1 and 3-2, only three counties, Alachua, Marion, and Liberty, lack coas tline and the Alachua Co. records from the U. F. campus are dubious because they were more likely descendants of M. g. sweadneri released there to found an experimental colony. I think it is unlikely that the butterfly was historically re stricted to coastal habitat even though the earliest record from interior Florida I hav e seen is from Zeiger ’s 1979 capture in Marion Co. (Table 3-1). Gro ssbeck (1917) mentions a Mitoura damon specimen from north-central Florida which may have been from an interior location but does not give details. It is more likely t hat localized colonies existed before the mid-1800s and that the butterfly lost habitat in interior Florida due to widespread logging. Colonies in the middle of north and central parts of the state may have gone unnoticed until the 1970s simply becau se no one looked for them there. The biggest boost for the butte rfly in interior counties could be from the millions of Red Cedar seedlings produced and di stributed by the Florida Division of Forestry that were planted along roads and the boundaries of fields in rural areas. As these trees grew to a si ze and shape capable of maintaining the microclimate required by the pupa, they were colonized by dispersers from boom flights. These more visible planted rows eventually attracted the attention of lepidopterists who “discovered” butterf ly colonies in landlocked counties.

PAGE 205

189 Overall, the Florida butterflyÂ’s distri bution and redistribution across its range may function as a metapopulation (Gilpin and Hanski, 1991, 1997, and Hanski and Thomas, 1994). Local populations are colo nized by dispersers from booming flights in established colonies as habita t becomes available. As older habitat is destroyed or degraded through natural su ccession or other natural causes, once-stable populations wink out and their satellite colonies mature to become stable bases that produce new colonists in optimum years. This is a slow, long-term process linked to the pace of growing Southern Red Cedar trees.

PAGE 206

190 CHAPTER 4 SPECIFIC PHENOTYPIC DISTINCTIVENESS, Mitoura gryneus sweadneri The taxonomy of butterfly species is most frequently defined by morphological differences, or discontinui ties in their most obvious physical attribute: their wings. Alar characters, from Latin ala , a wing, may include the shape of the wing, and the a rrangement of its neuration, or venation, patterns of spots and stripes, colors and shading, specific features such as patches of scent scales called androconia, and even the shapes of individual scales. When alar characters are very similar, as is t he case in some species of Hesperiidae, taxonomists have resorted to close scrutin y of the sclerotized structures of the genitalia of these obscure/occult biological entities for specific characters with discontinuities consistent enough to def ine species. Fortunately, general wing shape and neuration are characte rs which are usually consistent at the level of genus or higher for most butterflies. Unfortun ately, these similarities place heavily weighted significance on the remaining alar characters for distinct separation of species. Since at least LinnaeusÂ’ time, the tax onomic description of a new butterfly has depended on the existence of a type s pecimen. This carefully selected individual was the one which the orig inal author had before him which exemplified the characters of the species. Since the mid-19th century, most descriptions have also included a type locale from which the specimen, or in the best cases, a series of specimens was taken. This type specimen, or holotype,

PAGE 207

191 and hopefully several paratypes, then rema in in some safe repository as a standard reference with which to compare newly discovered butterflies for species distinction. The necessity of a physical type specimen on a pin in a museum introduces a pl ethora of problems. The age and condition of the specimen( s) when captured. Flown butterflies lose scales. Even careful preparation of specimens can cause loss of scales and damage from pins inserted through the wings during the mounting process. Fugitive colors of living butterf lies fade in sunlight, and even pinned specimens in artificial light fade over time. Museum specimens are fragile and subj ect to humidity and a variety of museum pests. A single specimen or series may have been collected from only a single site in the butterflyÂ’s range and may thus repr esent a specific spatial form of the beast. Numerous butterflies, including some hairstreaks, are sexually dimorphic. These alar differences are frequently on the dorsal forewing surface, which is not apparent until the butterfly is spread. Some butterflies that are bivolti ne, or multivoltine, produce morphs exhibiting seasonal variation, introduc ing a temporal dimension to the phenotype. For these reasons and more, the best species descriptions are often based on captive-raised specimens which may have included representatives from several broods and potentially add life history di stinctions associated with specific alar characters. The researcher must depend on evaluation of information from pinned specimens from outlying regions if the butterfly occupies a broad range. Finally, there is the relative thoroughness of the written description from each author of a new species, which should be helpful for later generations. Arcane terms that lepidopterists use to describe distinct shades of colors have evolved

PAGE 208

192 and changed over time and may not have been used consistently over centuries or even decades. Even the light source used by a 19th century author to study his find could have a very different quality w hen reflected by the scales of butterfly wings, compared with the incandescent or florescent lamps of the mid-20th century, or with the techni cally advanced lighting available in modern research labs like the McGuire Center for Lepid optera and Biodiversity at the Florida Museum of Natural History. With t hese caveats in mind, I approached the phenotypic study of Mitoura g. sweadneri . Historic Illustrations The earliest extant depictions of what was to become the nominate species, Mitoura gryneus , were drawn from specimens in t he collection of Ca spar Stoll by Gerrit Wartenaar Lambertz, and published as hand-colored engravings by the amateur Dutch naturalist, Pieter Cramer, in what has become a famous historic Figure 4-1. Papilio damon (Stoll) in Cramer, 1782. D) ventral, C) dorsal. document in the history of entomology, De Uitlandsche Kapellen (Cramer, 1782). CramerÂ’s book, published in four volumes, contained descriptions of over 1,650 exotic butterfly species from several continents and was the first to use the

PAGE 209

193 then-new Linnaean system of taxonomic no menclature. Figure 4-1 illustrates the new butterfly Stoll named Papilio damon and recorded “Virginia” as its type locale (Stoll, 1782). Since the original type s pecimen is probably lost (Miller and Brown, 1981), as were many of these early spec imens of Cramer and Stoll, Lambertz’s drawings have special scientific status as ‘iconotypes’ (British Natural History Museum: (http://www.nhm.ac.uk/natur e-online/online-ex/art-t hemes/drawingconclusions/m ore/butterflies_more_info.htm)). The second-oldest published image of the hairstreak is from Boisduval and LeConte (pl. 33, 1833) which preceded Scud der’s original description of the genus Mitoura by 36 years (Scudder, 1869). Boisduval and LeConte named this Figure 4-2. Thecla smilacis (Boisduval and LeConte) 18 33 6) ventral, 5) dorsal butterfly Thecla smilacis , based on a drawing by John Abbot for which a type specimen probably never existed (Miller and Br own, 1981). The type locale given

PAGE 210

194 is simply ‘Georgia’. Though both of these illustrations are somewhat stylized, for example, Cramer’s colorist gave Papilio damon red eyes which are not characteristic of this insect (Figure 4-1) , taken together they illustrate two of the caveats mentioned above. First, the depiction of the alar c haracters in Figure 4-2, show solid brown dorsals with a distinctly different colored androconial patch on the forewings found only on the males of the genus, a sexual dimorphism not noted by the authors. Second ly, the solid brown dorsal surfaces are in sharp contrast to the orange discal areas surr ounded in black wing margins of Stoll’s P. damon in Figure 4-1. This is a seasonal characteristic of the second brood of M. gryneus . Harris referred to the Boisduval and LeConte illustration specifically as the dark summer form, C. g. gryneus f. smilacis , in his Butterflies of Georgia (Harris, 1972). When Scudder specified c haracters for the genus Mitoura , based on the type Thecla smilacis , he described comparative widths of the head, characters of the legs, antennae with 28 joints, and minor differences in neuration of the forewings of males and females, but gave no other alar details in this unillustrated work (Scudder, 1869). Tw enty years would pass before he described the alar characters of the Mitoura Alar Characters: Mitoura gryneus When Scudder described Mitura damon as the only species representing the genus in the eastern United States, he listed Papilio damon Cramer, 1782, Lycus gryneus Hübner, 1819, Thecla smilacis Boisduval and LeConte, 1833, and Mitoura smilacis Scudder, 1872 among the ten previous names this new

PAGE 211

195 designation was to replace (Scudder, 1889a). He wrote a meticulous description of the dorsal and ventral alar characte rs for this new taxon which covered Figure 4-3. Mitura damon Scudder, 1889b 17) female ventral & dorsal, 18) male v. & d. over two pages of single-spaced, 6-poi nt type and in his book of maps and illustrations to accompany the text (Scudder, 1889b), he published a 15-stone lithograph of drawings by J. Henry Blak e as plate 6 which included the two illustrations reproduc ed in Figure 4-3. In the 115 years since ScudderÂ’s detaile d description, several authors have abbreviated specific alar characters for the butterfly referred to as Mitoura gryneus in this paper, and a small subset of these authors has given brief distinct, discontinuities in thes e alar characters to separate M. g. sweadneri from the nominate subspecies. The terms used by each author to refer to a specific character have changed over time and even today there are no commonly accepted standard terms for some alar characters. Since ChermockÂ’s description of Mitoura sweadneri , a few characters, especially of the ventral hindwing, have taken on special emphasis. Interestingly, these characters are barely suggested

PAGE 212

196 in Figures 4-1, 4-2, and poor ly portrayed in Figure 4-3. In order to facilitate discussion and comparison of the work of several different authors, I have adopted the following terms and abbreviations for alar characters from several sources. VFW = ventral forewing(s). The postmedi an line obvious in the three figures as white, lined basad with a darker br own or red color is the PML. DFW = dorsal forewing(s). The center of the DFW is the discal area corresponding roughly with the orange pat ch surrounded by black in Figure 4-1 C, and with the same patch in Fi gure 4-3, 17, and 18 shown partially surrounded with dark brown, will be refe rred to as the disc. The lighter colored androconial patches below the co stal margin of the males (Figures 4-2, 5, and 4-3, 18) will be androconia. VHW = ventral hindwing(s). This wi ng also has a PML lined on the mesial side in red as in Figure 4-2, 6, plus tw o white postbasal lines, spots, or bars lined on their distal edges in red in the same figure. These basal bars, BB, are also indicated in the other fi gures with darker linings. The HW each have two tails. The longer one is at the end of the second Cubitus vein, Cu2, the shorter tail is produced at the end of Cu1, most clearly illustrated by Lambertz’s drawings (Figure 4-1). In the cell between these two veins (Cu1-Cu2), partway basad of the HW distal margin is a variable patch called the Thecla spot, TS, which is not well illus trated in any of the figures but indicated as two black spots in Fi gure 4-2, 6 where th e short tail at Cu1 is drawn too low. This terminology for designating veins follows “Comstock-Needham” (K lots, pp. 50-53, 1951). DHW = dorsal hindwing(s), sometime s has a faint white PML showing through. The disc on this surface includes the area immediately dorsal to the TS and corresponds roughly with the orange patch on the DHW in Figure 4-1 C. Abbreviated descriptions of alar characters for Mitoura gryneus from several authors follow which demonstrate the wide range of states of these characters. Scudder, 1889a Scudder said the DFW are a dull blackish brown with discs brassy ochraceous, , or tawny, , crossed by blackish brown veins. For DHW he

PAGE 213

197 specified discs as obscure brassy ochraceous, , or tawny, recognizing sexual dimorphism in these characters. The androconia are dust gray. VFW is green flecked with ferruginous scales, with the lower part covered by the HW pale slate brown. PML is very nearly strai ght, snowy white, bordered mesially with ferruginous, and broken and separated at its lowest segment which is moved basad. He referred to the PML of the VHW as me sial, and said it is “a very tortuous snow white stripe, bordered on the inner side with dark cinnamon” (Scudder, p. 862, 1889a). He referred to the BB as co lored like the extra-mesial band but with the colors reversed. Just distad of the mo st distal reach of the PML is a row of four or five blackish dots subparallel to the HW outer border. Just distad to one of these is another blackish spot with t he space between them filled with obscure orange. This second black spot together with its mesial orange lunule comprise the Thecla spot, though Scudder did not use this terminology. Scudder measured the longer tails of 20 , and 19 and said they ranged from 1.5-3.25 mm and averaged 2.25 mm. Finally, he observed a few alar particulars for a single southern specim en that will be repeated among the alar characters for M. g. sweadneri , and he mentioned a single female from Long Island with entirely blackish brown DFW and only a few tawny scales scattered over the DHW. This astute observation s uggests either that the solid black female form was not seen by him around Bo ston and Cambridge in which case it would be the first recognition of regional vari ation of this character, or, as is less likely, that he had not seen specimens of the second brood expressing this

PAGE 214

198 character state which would suggest t hat there were not prominent seasonal variations in dorsal wing coloration in his area. Figure 4-4. Mitoura g. gryneus , from the Pine Barrens of N. J. First flight, seasonal morph Specimens in Figure 4-4 were collected by W. B. Wright in t he vicinity of the type locale of Mitoura hesseli (Rawson, and Ziegler, 1950) and show typical alar characters of the first brood of th e nominate subspecies I have seen near Lakehurst, Ocean Co., N. J. The TS in t hese digital images are much clearer than in the illustrations of Figure 4-3, and the orange shadings of the discs of both DFW and DHW of the female are a c onsiderable variation of those in

PAGE 215

199 ScudderÂ’s plate (Figure 43, 17). These images of pinned specimens and those of the remaining alar figures were shot in sunlight with a Kodak EasyShare model Z760, 6.1 megapixel digital camera and reproduced at approxim ately 2.35 X life size. Holland, 1940 The discs of DFW and DHW are bright fulvous with the margins and veins of both wings blackish and darkest at the apex of the wing in Thecla gryneus . The white PML of the ventra ls of both wings are margi ned internally with brown. The BB are only described as white. Klots, 1951 The DFW and DHW of females have more orange color than those of the males and some females are nearly comple tely sooty brown. KlotÂ’s plate 16 shows a little orange in the TS. These di sc descriptions contradict Figure 4-4. Harris, 1972 As mention above, Harris said C. g. gryneus is double brooded in Georgia and that the (dorsals of) summer form adults are darker than those of the first spring flight. Howe, 1975 Dorsals have considerable orangebrown discs in both sexes. His illustrations of C. ( M. ) gryneus , plate 49, were from specimens collected 24 April 1954 at Bloomington, Indiana, and show dis cs as broad as those of the male in Figure 4-4 for both sexes. They also have orange about the TS. Judging by the date of capture, these were probably first brood adults.

PAGE 216

200 Johnson, 1980 The DFW and DHW varies distinctly in some populations across the range of Mitoura gryneus in the central and northern U. S. from solid dark Prouts brown to varying amounts of ochraceous tawney to tawney discs with margins of both wings always deep raw umber to fuscous. The discs of specimens from Texas and Arizona are broad and vary from c hamois to honey yellow. The VFW and VHW ground colors also vary regionally fr om bright deep green to yellow green, or lime green. The TS is orange bas ad and suffused blue distad. The VHW features a broken and irr egular PML and the BB vary from diminished to absent in the southwest U. S. to two or three in other parts of the butterflyÂ’s broad range. Opler and Krizek, 1984 The dorsals of M. gryneus have much orange (discs) in the spring flight and the summer form is black-brown. Their photograph (figure 82, plate 14) shows an individual from Fairfax Co., Virginia with the orange and black TS. Scott, 1986 ScottÂ’s alar characters for Callophrys gryneus include ventrals varying from brown to green with VHW PML as white zigzags and white BB. The dorsals have orange discs in the first brood and the summer broods (form smilacis ) are less orange. Allen, 1997 Allen clearly described two seas onal color forms for each sex of Mitoura grynea grynea in West Virginia. The first f light of spring has tawny orange, , to bronze, , discs followed by the summer forms: more tawny orange than bronze,

PAGE 217

201 , and dark brownish black , discs. All have the fa miliar red and white VHW PML. My own images are of captive-raised M. g. gryneus from stock collected from the approximate c enter of its range in Giles Co., Tenn. Figure 4-5 shows a Figure 4-5. First flight Mitoura g. gryneus , Giles Co., Tenn. male and female that were progeny of the second brood (captive F1), which should have overwintered but did not experie nce the chill of fall in their native Tennessee and consequently emerged in September 2003 in my warm living room. The DFW and HW of the male have t he discal characters typical of wild

PAGE 218

202 caught males from the first flight of Apr il in each year that I collected colony stock. The female dorsals show a dusti ng of orange in the DFW disc with a more pronounced orange DHW disc patch directly ov er the VHW TS also typical of the first brood. The black dots of the VHW Thecla spots happen to be somewhat larger than those of the Lakehurst, N. J. s pecimens (Figure 4-4), but this is also a variable character as evidenced by the even larger black spots in the VHW TS of F1 individuals in Figure 4-6. The male s of the second brood have much paler orange discs and several individuals had even less pronounced orange. The dorsals of the female of Figure 4-6 appear to be completely dark brown but actually have a smattering of orange sca les on the DHW. This character varies from solid dark brown, almost black, to various degrees of pale orange/dark yellow shading and some individuals only ha ve orange scales along certain veins of the DHW. This observation is in agreement with JohnsonÂ’s (1980) observations of the wide variability of t he disc character in the dorsals of the nominate species. These two examples (F igures 4-5, and 46) also demonstrate the Tennessee version of sexual dimor phism and seasonal variation of the butterfly in West Virginia (Allen, 1997). The colors of androconia are generally a shade of gray or brown but can vary widely from pale tan to nearly charcoal black or a very light gray in males produced from the eggs of a single female. T hese variations in androconia colors are evident in males from both broods t hough the lighter shades were more often noticed in second brood males in my captive colony of Tennessee stock.

PAGE 219

203 Figure 4-6. Summer brood Mitoura g. gryneus , Giles Co., Tenn. A closer look at the VHW PML characters of the M. g. gryneus figures revels a small sample of the variability of the widths of the white lines and their red mesial edges. These widths not only vary from across the range but can appear quite different from two individual s taken at the same location on the same day. Though the PML is consistently quite crooked in this limited selection of pictures, a small perc entage (perhaps 1:25-30) of ca ptive-raised individuals had nearly straight lines above HW vein M3. The width and prominence of the BB of Tennessee butterflies also vary fr om the images from Lakehurst, N. J.

PAGE 220

204 (Figure 4-4) and within their own cohort, as does the amount of reddish distal outline of these bars. The BB of the male in Figure 4-5, for example, are much less clearly defined than those of the fe male that eclosed on the same day and have very little red mesial lining compar ed to hers. By comparison, the summer brood female of Figure 4-6 has much reduc ed BB next to her counterpart of the same brood. A final comparison of the white-tipped, long Cu2 tails of the New Jersey hairstreaks with those of the Tennessee fi gures shows the tails of individuals from the middle of the speciesÂ’ range ar e longer. I was not able to quantify the differences in tail length because of t he limited number of carefully prepared M. g. gryneus specimens available to examine in the Florida State Collection of Arthropods (FSCA). (See Tennessee/ Flori da tails compared below.) Very often these long tails were carelessly bent, or perhaps, purposefully folded when a specimen was mounted in order to conser ve space. Crooked tails are nearly impossible to measure accurately because adding the lengths of short segments compounds even tiny errors made with the ca lipers to give artificially inflated results. Alar Characters: Mitoura gryneus sweadneri Though fewer authors have written specific descriptions of Mitoura gryneus sweadneri than have written about M. gryneus , each one, with the exception of Chermock, has limited comments to specific alar characters, or character states, viewed as distinct discontinuities betw een the two hairstreaks weather the author considered them species or subspecies. In ChermockÂ’s original description of Mitoura sweadneri , reproduced verbatim in Chapter 1, he listed nine characters,

PAGE 221

205 or character states, of the “primaries” or forewings; eight of which apply directly to the array of variable al ar character states of M. g. gryneus just described in the preceding section. One of these, “the extra-median in terrupted white line” (PML of the VFW), he even said is “s haded inwardly with brown as in damon ” (Chermock, p. 216, 1944), which was t he authoritative name of the nominate species (Scudder, 1889). As Chermock described the “secondaries” or hindwings, he used 11 characters, only two of which appear to apply solely to the VHW of M. g. sweadneri , and one of them varies considerably across the range of the Florida butterfly. In conclusion of the preamble to discussion of the alar particulars of M. g. sweadneri , I add two more terms and their abbreviations to the list of alar characters began above, neither of which have been used by any cited author in relation to this butterfly. TL = tornal lobe(s). At the anal an gle, or tornus of the hindwing of Mitoura , a small lobe of the wing bends out away from the plane of the wing at a 90° angle, lateral to the wing and the horiz ontal plane of the body. Viewed from above the perching hairstreak, or from below, the TL look like exaggerated mimics of the butterfly’s eyes and in combination with the long tails at Cu2 above, it may represent an enticing false head which could mislead the initial strike of a predator, giving the butterfly a second chance. This illusion is enhanced by the characteristic inte rmittent “grinding” of the hindwings when these hairstreaks perch which caus e the tails, or false antennae, to wave about. (This feature is not always apparent in museum specimens because preparators sometimes flatten t he TL to conform to the plane of the HW when a specimen is spread. It is discernable in the M. gryneus specimens in Figure 4-5, and by its s eeming absence in the dorsal view of the female in Figure 4-6 where the l obe is bent away from the camera, ventrad to the butterfly in its natural orientation.) TBL = thin brown line(s). This th in brown line lies half-way between the most caudad BB and red mesial lini ng of the PML of VHW of M. g. sweadneri specimens taken along north latitude 29° in my study sites. It is sometimes faintly represented when onl y a single scale wide but usually spans the full length of the most caudad BB.

PAGE 222

206 The following list of abbreviated descripti ons of character states begins with F. M. Chermock and includes those of all the previously cited authors writing about SweadnerÂ’s Hairstreak. The accom panying figures are my own digital images chosen to confirm, and to show variations of, these characters. Chermock, 1944 Chermock said the dorsals of M. sweadneri were grayish brown with a slight greenish iridescence, and that the VFW was light brown suffused with green. His two specific VHW characters were first, a large black patch on the inner angle, the TL, and a second large bl ack patch, the TS. He did not mention any orange associated with this second character. He described the VHW PML as a white zigzag with mesial shading brown, and the BB as two large white spots. He referred to the broad field mesi al of the VHW margin as blue with an overcast of white scales. Scudder, 1889 ScudderÂ’s single southern specimen ( no specific locale), described 55 years before Chermock, was diffe rent from all the northern Mitura damon specimens by dorsals of uniform brown with no reddish tint and tails 5.5 mm long, or twice as long as those of his familiar Olive Hair Streak. Klots, 1951 The Florida subspecies, sweadneri , lacks the orange-brown discs of M. g. gryneus , and the VHW has a larger black patch on the TL with the TS lacking orange.

PAGE 223

207 Howe, 1975 C. ( M. ) g. smilacis has no orange-brown ventral discs in any brood (my emphasis) and ranges across central and nort hern Florida. His Figure 2 of plate 51 shows no orange about the TS. Johnson, 1980 Johnson separated M. sweadneri from the nominate species by completely dark dull raw umber dorsals. His VHW PML character makes an important distinction of being “rather straight ac ross wing and jagging in a near “W” shape anad” lined in “broad red-brown basad” (J ohnson, p. 564, 1980). He also said the tails are longer and that sweadneri hardly ever has orange about TS (p. 566). Scott, 1986 Scott’s alar characters for C. g. sweadneri are that VHW lacks the red TS, the TS is black instead, BB are smaller, and the PML is strai ghter than that of C. g. gryneus . He specified the dorsals lacking orange discs but sometimes showing yellow-orange in we stern Florida. In my initial field season, Sweadner’s Hairstreak always appeared in very low numbers and I considered them much more valuable as live, in situ study subjects than as vouchers on pins. Consequ ently, I did not collect first flight adults in the first few years. The few fema les I took for captiv e propagation were always from later season’s broods when they were slightly more plentiful, and their progeny usually showed the solid dark brown to black ventrals cited in the literature. However, a Chermock paratype in the FS CA collected 15 June 1921 at St. Augustine, in worn condition, showed faded brown dorsals but with ten dull golden-orange scales just under the right androconia and seven more under the

PAGE 224

208 left one. Two males captured 10 April, and 7 March 1964 by C. F. Zeiger at Ft. George, Duval Co. had almost full orangegold DFW discs, and several females which he collected between 1963 and 1982 showed varying degrees of orange or gold shading of the DFW or DHW discs though many of his 33 specimens in FSCA are solid shades of brown. All of these specimens were taken just 40 mi north of the St. Augustine type locale . When I first saw T. C. EmmelÂ’s chromosome specimen, US-4024 , taken 17 March 1972 at St. Augustine, Figure 4-7. Mitoura g. sweadneri First flight, St. Augus tine, St. Johns Co. v) ventral, and d) dorsal as pects of each specimen.

PAGE 225

209 and a female from the same locale tak en 1 April 1984 (Figure 4-7), I realized the significance of the dates of Dr. Sweadner Â’s 83 specimens collected in June of 1940 and 1943. Chermock had only studied i ndividuals from one location and from one summer brood, so he missed the variation on the dorsals of EmmelÂ’s specimens in Figure 4-7. The dates of these two suggest t hey are late first flight individuals and though both initially appear ed to have solid dark brown dorsals, each one has a smattering of orange scales in the DFW discs which is a little more obvious in the female. The TL has a large black patch on each. The black TS lacks any orange scaling just as descr ibed by Klots, Howe, Johnson, and Scott above. Both specimens also have the TBL not mentioned in the literature which is a little more pronounced in the left VHW of the female. A first flight male and a late summer fe male collected 35.3 mi west of St. Augustine at my study site in Holliste r, Putman Co. show typical characters iterated by the authors cit ed above. Specifically, the male in Figure 4-8 has no orange on the dorsals, a nearly straight VHW PML, almost solid black TS, the big black spot of the TL and the long ta ils. The androconia are almost the same shade as the dorsal ground color, but show in the photograph because the androconial scales are a differ ent texture. The female in the lower half of the figure has no orange on the dorsals and the ventrals are faded to tan because she flew in the sunlight for weeks before capture. Her BB are much smaller than the maleÂ’s, but this is a variable character in sweadneri just as in the nominate subspecies. The VHW PML of both of t hese specimens have broad red mesial linings, the character state specified by Johnson. The TBL which I noted is visible

PAGE 226

210 in all of the SweadnerÂ’s Hairstreaks pictured. Two more specimens of the mid-summer flight from the same popul ation demonstrate fu rther degrees of Figure 4-8. Hollister, Putnam Co. M. g. sweadneri Male is first flight, female from late summer brood. variable character states that increas e in frequency and intensity in farther reaches of the Florida butterflyÂ’s range. The dorsals of both male and female in Figure 4-9 have a few scattered dull orange scales, expressed more on DFW of the , but more on DHW in the . The TS of the VHW has a slight lunule of dull orange brown on both wings, but the only expresses the character to a

PAGE 227

211 some what less obvious degree in the left TS. The BB vary in these two specimens as does the thickness of the TBL. Figure 4-9. Mid-summer flight M. g. sweadneri , Hollister, Putnam Co. Only 20 mi west of the Putman Co. si te, and 60 mi from St. Augustine, two males of the first flight from Cross Cr eek, Alachua Co. (Figure 4-10) have smaller black spots in the TS and neither has any orange associated with this character. But the 17 March has a thin row of brown scales between the TS eye and the black spot just basad of it. In the 14 March , this is reduced to just a few brown scales. The DHW of the 17 March has only a light gold dusting, while the DFW

PAGE 228

212 Figure 4-10. First flight males M. g. sweadneri , Cross Creek, Alachua Co. of the 14 March has this light gold dust only around the right androconia. The shade of the androconia is obviously di fferent in both specimens. The most remarkable change is in the width of the r ed mesial lining of the HW PML which is straighter in the 17 March . The TBL is much reduced in both and barely discernable in the 17 March (Figure 4-10). Figure 4-11 is of specimens captive-ra ised by M. C. Minno from stock taken at the now defunct Royal Park Theatre site in Gainesville, Alachua Co. This pair represents second brood indivi duals and they show remarkable differences in the

PAGE 229

213 Figure 4-11. Second brood M. g. sweadneri , Gainesville, Alachua Co. TS from the Cross Creek specimens from only 14 mi south. Though the black eyes of TS are large, their mesial bor ders are bright orange lunulae as in the M. g. gryneus specimens of Figures 4-4, 4-5, and 4-6. The red mesial linings of the VHW PML are thinner than those of any M. g. sweadneri previously figured. The row of black dots just basad of the TS in these individuals are expressed as black lines, as are those of the female M. g. gryneus of Figure 4-6. The TBL are very

PAGE 230

214 thin but present in both, and both have light brown dorsals without any orange scales. Figure 4-12. Mitoura g. sweadneri Gainesville, Alachua Co. Last brood male, early summer brood female. Two more of MinnoÂ’s specimens show furt her variations of character states in SweadnerÂ’s Hairstreaks from Gainesville. Minno raised the male in Figure 4-12 from Royal Park Theatre stock and it show s variable characters of the last fall flight in Alachua Co. The DFW discs ar e dusted in golden orange scales and the bright orange androconia are the only exam ple of this trait I have seen. The

PAGE 231

215 female in this figure was captured in the woods behind the Florida Division of Plant Industry building, now part of the U. F. Natural Area Teaching Lab tract. Her dorsals are solid brown. The white of the BB in both is indistinct and while both have the large black spot in the VHW TS, only the female shows as much bright orange as the previous examples from Gainesville (Figure 4-11). These two specimens are from different broods two years but less than two miles apart and both show VHW PML thicker than Mi nnoÂ’s specimens from 1987 (Figure 4-11). Figure 4-13. Late-summ er variations I, M. g. sweadneri , near Jena, Dixie Co.

PAGE 232

216 The next two figures are from my own captive colonies from Dixie Co. stock captured on the Gulf coast and por tray alar characters from the last flight of the 7-month Gulf coast season. The male in Figure 4-13 has light orange-gold dusting on both otherwise dark brown dorsa ls. The femaleÂ’s dorsals are the same solid dark brown without any orange-gol d scales. Both have the large black spot but with reduced orange of the TS. While the BB of both are reduced, the female has one BB without any white. S pecimens in Figure 4-14 are from the previous year. The male has more generously gold-dusted dorsals which are very faint in the female. Both have the large black spot of the TL, common across Figure 4-14. Late-summer variations II, M. g. sweadneri , near Jena, Dixie Co.

PAGE 233

217 the range, and the large black eye of the TS with only a thin slice of orange. The variable character of the HW PML is strai ghtest in the female in Figure 4-13, but also present in the male of Figure 414. The male of Figure 4-13 is the only example I have of M. g. sweadneri with such a poorly defined TBL of VHW. Figure 4-15. Worn first flight M. g. sweadneri , Yankeetown, Levy Co. The worn specimens in Figure 4-15 hav e lost too many scales and are too sun-bleached to show clear TBL characte rs though the partial remains of one are just visible on the right VHW. They do show the VHW PML of both, and the

PAGE 234

218 femaleÂ’s PML is bent distally in a fashion unlike the other M. g. sweadneri shown. The only examples of this distal curve I have are in the figures of M. g. gryneus from Tennessee and New Jersey. She is al so missing the BB that on all the other specimens lies nearest the costal margin of the VHW. These character states may be anomalies peculiar to this individual alone, but the distal curve of the PML may show an affinity with the same character of the nominate subspecies. These two specimens were part of a se ries of 18 captured at the Yankeetown site, waypoint (WPT) 018, in Levy Co., or raised from sto ck captured between 26 March 1978 and 28 June 2002 by Jeff Slotten. One of these, a male from 26 March 1978 had full yellow-gold ventral discs on both wings and two others had 2/3-3/4 yellow-gold discs of the DFW. The only cited au thor to recognize this particular character state was Scott (1986). The final two images of M. g. sweadneri are of specimens collected by John Calhoun in Polk Co. which repr esent characters from the south-central limit of the hairstreakÂ’s known range. The male (Figure 4-16) is probably from the first flight, if the timing in Polk Co. coincides with the first brood of Putnam Co., which is also in interior Florida. The dorsals of this fairly fresh specimen are heavily dusted in yellow-orange. The female is from a later brood and has solid brown ventrals without any orange or gold scales. Both have the large black spot of the TL and both have the VHW Thecla spot comprising the large black dot with mesial orange edge, accentuated in the fema le. The TBL is easily discernable in both specimens and the BB and VHW PML char acters are well within the range

PAGE 235

219 of variability demonstrated on other Sweadn erÂ’s Hairstreak specimens from across the state. Figure 4-16. Mitoura g. sweadneri , Southern limit, interior, Polk Co. Specificity of Alar Character States: M. g. sweadneri While there is obviously considerable variation in alar characters for M. g. gryneus found in the plates and description s from the several guides cited previously, particularly in dorsal shadi ng, this could be expected because the butterfly occupies such a broad range across the United States. The spread is so great from north to south that the first flight in Tennessee, which does not

PAGE 236

220 emerge until the last week of March and is not abundant until the first few days of April, is at least a full month ahead of conspecifics from the New York-New Jersey populations, which are rarely seen before the last week in April and are not recorded as common until May (Glass berg, 1993). The whole known range of M. g. sweadneri comprises just 27 counties of north and central Florida and the distances between my field sites are tiny , compared to the range of the nominate subspecies. The most western Dixie Co. site is 128 mi distant from the type locale at St. Augustine on the Atlantic coast, 60.5 mi west of the Gainesville Royal Park Theatre site, and only 95.5 mi from WPT 019 in Putnam County. (These are direct distances measured fr om recorded GPS waypoints during field research.) If indeed, the northern reach of Nassau Co . and the southern border of Polk Co. harbored SweadnerÂ’s Hairstr eak colonies, their total north-south range would only be a few minutes broader than three degrees of latitude, or roughly 221 mi wide. The fact that M. g. sweadneri wings are not uniform across the butterflyÂ’s range suggests restricted gene flow between isolated colonies due to the sessile nature of this butterfly reputed never to be found far from its Southern Red Cedar host trees. Though the dorsal alar characters of SweadnerÂ’s Hairstreak vary between the sexes and, to some extent, from spring through fall broods, the amount of orange, gold-orange, or yellow-orange scaling in the dorsal discs tends to be greatest and occur most frequently in s pecimens from the Gulf coast. Of 158 specimens from several sources compared for this character, 75 had some degree of this colored dusting; the rest were a nearly monotone shade of brown,

PAGE 237

221 or dark brown to nearly black. The whit e of the VHW PML shows faintly on the DHW in about half the specimens, and t hese are usually the lightest brown shade. Some of the ventral characters appear to be more prominent at specific locations. For instance, the widest red s hading mesial of the VHW PML is from interior Florida, Putnam Co., Polk Co ., and the southern end of Columbia Co. Though Johnson specified this character state for sweadneri , he only examined specimens which were from the Atlantic and Gulf coasts (Johnson, 1978, 1980) and so had missed the most pronounced stat e of this character. A final VFW character, the androconia color, is extr emely variable, as evidenced in just the few example figures, from dark ch arcoal to orange (Figure 4-12). From my own observations of th e ventral alar characters of M. g. sweadneri , including museum specimens, severa l from private collections, and my own captive colonies, the range of vari ability of character states is actually broader than those of all the previous authors combined. The ventral ground color of newly eclosed individuals darkens within a few hours to the medium green common to t he fresh captive specimens in the figures and only fades to yellowish tan with several days of exposure sunlight and loss of scales. The red details are a little less fugitive fading to brown, and the white characters last longest, often still discernable after no trace of green or red remains. The width of the white band and the red mesial lining of the VHW PML can be more than twice as wide in the br oadest examples than in the narrowest. This accentuated red band was not seen in any specimens nor illustrations of the nominate subspecies. This PML is straighter above the th ird branch of the Media HW vein, M3, than in all but less than 1% of M. g. gryneus specimens examined. The large black eye of the Thecla spot (TS) is larger in the Florida butterfly, but seems to gain the mesial orange l unule away from the type locale.

PAGE 238

222 Basal bars vary in size, shape, degree of white, and of red distad shading in no discernable pattern in different broods and within broods at the same location. The thin brown line (TBL) is almost exclusive to the Florida subspecies, though a shortened, faint suggestion of it is just visible in less than 2% of the Tennessee butterflies somew hat caudad of its position in sweadneri . The large black spot on the tornal lobe (TL) is bigger in all Florida specimens examined and the tornal lobe itself appears larger than that of any of my captive-raised Tennessee hairstreaks. The length of tails at Cu2 of the Fl orida subspecies is consistently longer and they are frequently wider than any of the tails of the Giles Co., Tennessee captive colony. Though I could not measure the tails of enough M. g. gryneus from across its entire range to pass judgment on the excl usivity of this character state to subspecies sweadneri , based on ScudderÂ’s observation of double length tails of his sole southern specimen (Scudder, 1889) and the reference to longer sweadneri tails from Johnson (1980), I think the tails of M. g. gryneus may get longer in populations occurring south of t he northern limit of its range. Table 4-1 compares tail lengths of Giles Co., T ennessee Olive Hairstr eaks with those of SweadnerÂ’s Hairstreak. Most of the M. g. gryneus that I raised in captivity were Table 4-1. Tail lengths by subspecies and sex in mm* subspecies: GRY GRY sample size 22 23 average 3.11 2.87 STDV n-1 0.36 0.44 range 2.33 3.69 2.19 3.90 subspecies SWD SWD sample size 59 78 average 5.32 4.92 STDV n-1 0.54 0.62 range 4.18 7.04 3.44 6.66 *Tail measurements were made with the same Mitutoyo Digimatic electronic caliper used to measure forewing lengths (Table 2-12).

PAGE 239

223 kept alive to determine life history spec ifics, or flown in captive breeding and hybridizing experiments, and consequently few were spread in good enough condition, with tails intact, to be measured for this character. However, the tail lengths of the sample of Tennessee butte rflies were longer minimum, maximum, and average lengths than Sc udderÂ’s reported lengths ranging 1.5-3.25 mm and averaging 2.25 mm from samples of 20 , and 19 (Scudder, 1889). They were also at least 2 mm shorter in every ca tegory, except possibly minimum male tail length, than those of the Florida sample. These tail length data were analyzed with SAS software using the StudentÂ’s T-test for the males of both subspecies ( t = 14.669, P < 0.0001) and for the females ( t = 16.391, P < 0.0001) which showed that these observed differences between the Tennessee population and M. g. sweadneri are statistically significant for both sexes. The discal coloration of first flight Tennessee M. g. gryneus may be a little more orange than first flight M. g. sweadneri , but the dorsal coloration of the summer, when not solid brown or black, is much closer to first flight sweadneri . The most exclusive alar character states of M. g. sweadneri are the larger tornal lobe, its big black spot, and tail length, fo llowed closely by the thin brown line of the Florida subspecies VHW and the usually straighter VHW PML. The width of the HW PML does not work as well bec ause of the frequent occurrence of its thinnest state in western parts of the Florida range. Alar Characters: Possible Intergrades When Calhoun wrote about the range of the St. Augustine Hairstreak, Mitoura ( g. ) sweadneri , he said it occurred in 20 Fl orida counties and that rare

PAGE 240

224 specimens with characters of the nomi nate subspecies had been found in Liberty Co. and in Jackson Co. of the Panhand le (Calhoun, 1996). In a recent communiqué he referred to the phenotype of these and possibly that of one specimen collected near Chipley in Washington Co. by Ray Stanford, 2 August 2000, as “intermediate” between M. grynea and M. g. sweadneri (Calhoun, pers. com.). In his dissertation, Johnson ( 1980) said two specimens collected in southern Georgia from Jekyll Island, Gl ynn Co., by Frank D. Fee, deposited in the National Museum of Natural History, expressed the phenotype of M. sweadneri . In a telephone interview in 2002, he said these two specimens might represent intergrades between the species (Johnson, pers. com.). On my limited surveys of the Jackson Co. site described by Calhoun, I was not able to find the hairstreak but a single specimen collected by Marc Minno in Liberty Co. (Figure 4-17) definitely shows affiniti es to characters of both M. g. gryneus and M. g. Figure 4-17. Intergrade, fi rst flight, Liberty Co. sweadneri . This butterfly has the large black eye of M. g. sweadneri ’s TS but with the wide orange lunule common to gryneus across its range. It also has the orange discs of first flight i ndividuals of the nominate subspecies, but the larger

PAGE 241

225 Figure 4-18. Intergrades, Marianna, Jackson Co.

PAGE 242

226 TL and its black spot diagnostic of the Florida subspecies. The VHW PML curves basad below M2, then distad towards M3 as in M. g. gryneus . Three specimens collected over two years by Calhoun (Fi gure 4-18) show this same mix of phenotypes. The VHW of the male in t he top row has wide red lining along the PML seen especially in interior Florida and the 18 June female in the bottom row has the thin VHW PML characteristic of M. g. gryneus. Two more specimens collected at this same site near Marianna in Jackson Co. by Jeff Slotten have the same intermediate phenotype with dark brow n dorsals that could be characters of either butterfly. All of these P anhandle specimens have the more tortured zigzag of the PML instead of the straight line of M. g. sweadneri . Though I did not photograph SlottenÂ’s specimens I was able to measure the tails of all of these butterflies. The results are in Table 42 below. Though this is a very small sample, it is interesting that the aver age tail length is about 1 mm longer than Table 4-2. Tail lengths of possible intergrades in mm INT lengths INT lengths combined stats 4.40 3.73 3.92 average 4.63 3.99 0.50 STDV n-1 4.02 3.22 3.22 4.63 range 3.42 that of the Tennessee hairstreaks, but 1 mm shorter than the average Florida specimen. The maximum tail length is 0.73 mm longer than the longest Tennessee specimen, but nearly 2.5 mm s horter than the longest Florida tail of Table 4-1. Finally, the minimum tail le ngth still exceeds the Giles Co., Tenn. minimum by a millimeter, but fits neatly into the range of the Florida character state.

PAGE 243

227 This Marianna, Jackson Co. site, WPT 960, is 142 mi north of WPT 002 in Dixie Co. and 309 mi south of the Turner Farms site, WPT 022, in Giles Co., Tennessee where I collected M. g. gryneus founders for my captive colonies. I think these rare specimens represent an intergrade between the two subspecies. They may be part of a cline of alleles expr essed infrequently as far south as Levy Co. To really confirm this idea, I would need a series of specimens from different broods collected from sites in each count y of the Big Bend along FloridaÂ’s Gulf coast, as well as vouchers in good condition from all the counties in the center of Alabama from the Florida border to Tenness ee. That would be difficult to achieve even for a well funded team of highly mo tivated field researchers. First, the butterfly often occurs in low numbers in the best of habitats, and second, the path of the cline would run along the Interstate 65 corridor which includes the highly developed sprawling urban areas of thr ee major Alabama cities, Montgomery, Birmingham, and Decatur-Huntsville. Still, it is an attractive hypothesis made more alluring by the cedars occasiona lly seen along Interstate 65 north of Montgomery. Genitalic Characters Very little has been publish ed about the genitalia of Mitoura g. sweadneri and most of that was written by a si ngle author, Kurt Johnson. When Scudder elaborated on the genus Mitura , he made very little reference to genitalic characters, writing barely two lines about the genus, and little more about the male Mitura damon . He did, however, include a drawing by Edward Burgess as figure 28 on plate 34 of his self-publis hed master work (Scudder, 1889b). Figure 4-19 is a digital reproducti on of the original lithograph. The basic genitalic

PAGE 244

228 Figure 4-19. Mitura , male abdominal append ages After Scudder, 1889b Figure 4-20. Male Mitoura g. sweadneri , labeled after Klots, 1970. Auto Montage image, dissection by author. structures of the male ar e drawn in a somewhat sty lized portrayal, but are easily recognizable in the Auto Montage image of the whole sclerotized structure

PAGE 245

229 dissected from a male Mitoura g. sweadneri taken at site WPT 019 in Putnam Co. (Figure 4-20). Scudder wrote in an era before a standardi zed vocabulary for these characters was widely accept ed. Alexander Klots had been working on terminology for male genitalic features bef ore the publication of his popular field guide (Klots, 1951) in which he included a se parate diagram of t he structures of male Lycaenidae that separated them from most other butterfly families but said that much work was left to be done on t he genitalia of female butterflies. For ease of reference, I labeled Figure 4-20 with terms Klots had chosen a few years later which still are in wide use today (K lots, 1970), and I employ these along with his terms for female genitali a in discussion of these two Mitoura hairstreaks. Scudder’s diagnosis of Mitura said the (vinculum) is “cleft above in such a way that the notch terminates sharply”, the lateral arms are not as delicately pointed as those of Thecla , and that the (valvae) are “gibbous at base, tapering rapidly beyond to a finely drawn out point” (Scudder, p. 857, 1889a). These characters are visible in his illustrati on (Figure 4-19), and the base of the right valva and its lateral outline show through t he transparent right half of the brightly lit vinculum in the digital image (Figure 4-20). Of Mitura damon , he said the (tegument) is rounded and has a lobe (uncus , or labis (Klots, 1951)), and the (valvae) are thin with needl e-like points (Scudder, 1889a). In his revision of the Callophryina, J ohnson (1976, 1980) said that valvae of M. gryneus were slightly concave, and shouldered caudad, and that the aedeagus and cornuti were thin. The distinctive characters for male M. sweadneri were concave (my emphasis) valvae, extremely shouldered caudad, and the

PAGE 246

230 aedeagus and cornuti were long and thin. He specified female character states for M. gryneus as having no signa in the walls of the corpus bursae, and lamellae as long as broad without ridges or convol utions where they join. The different states for M. sweadneri included one “U” shaped signum, lamellae with ridges and convolutions, and broadly spatulate with outstanding ventral invaginations (Johnson, 1976, 1980). He had dissected 46 M. gryneus specimens, including both sexes, and 17 M. sweadneri including males and females. His map indicates that all the Florida butterflies were from Gulf and Atlantic coastal locations. For my study of M. g. sweadneri genitalic characters, I included specimens from interior Florida, as well as from coastal populations. My dissections of specimens from Putnam Co. repres ent material would not have seen. Methods and Materials I dissected six male M. g. gryneus from captive stock and five females. The Florida specimens included four Dixie Co. males, two from Putnam Co., and two from Ft. George, Duval Co. Nine female specimens included five from Dixie Co., and two each from Putnam and Duval counties. I followed a standard method for specimen preparation, macerating dried abdomens with KOH (Winter, 2000), and all dissections were done at the E ndangered Species Lab oratory at the University of Florida. Specimens were studied using a Wild dissection microscope and the images for the figures were made using Auto Montage equipment in the Department of Entomo logy and Nematology as described in Chapter two. The genitalic structures of the males were posed suspended in a small wells carved into a Plexiglas s lide filled with glyceri n. The sclerotized

PAGE 247

231 structures of the female genitalia are rather flat and were posed by immersing them in a droplet of glycerin placed on t op of the slide. Scale bars were included for most of the images. The Auto Mont age images were subjected to a minimum of manipulation of brightness and contrast to illuminate details. Results of Genitalic Study All of the following figures of genitalic characters are reproduced here at about 22 X life size, except for the details of cornuti and signa. Because each digital image is built up of from 8 scans, for , to 15 scans, for , they reveal a depth-of-field sometimes difficult to portr ay with camera lucida. The whole male structure, Figure 4-21 w, is not intended to be typical of the Giles Co., Tennessee population, just as the figures of Flori da males (Figures 4-20, 4-22, 4-23) do not portray specific character states from t hose populations. All of the figures, rather, show a variety of variations observed in samples of both taxa. The length of the aedeagus was measur ed in six Tennessee butterflies and varied between about 2.68–3 X the length of the base of the valva caudad to the saccus. This same proportion was noted for the range of eight M. g. sweadneri specimens. The Dixie Co. male (Figure 423 w) was one of the longest at 2.96 : 1, and the Putnam Co. male (Figure 4-22 w) one of the shortest with a ratio of 2.71 : 1, aedeagus to valva. The shape and orientation of the vinculum and its lengt h to the base of the tegument varies, as does the thickness of the uncus and the turn of its ventral tips. Falces of both species vary slightly in length, but are as long as the valvae tips which protrude caudad to variable l engths past the unci. When viewed from

PAGE 248

232 the ventral or dorsal aspect, the curve and orientation of the falces in these figures do not nearly show the range of states of this character. The uncus of each specimen bears t he same setae, but the length and thickness of the setae themselves are so variable that the Putnam Co. male Figure 4-21. Mitoura g. gryneus , male genitalia Giles Co. Tenn. w) whole structure with aedeagus, l) right latera l aspect, v) ventral, and d) dorsal aspects. (Figure 4-22) appears nearly bald compared to the Tennessee male and the one from Dixie Co. (Figures 4-21, 4-23). The length and curve of the saccus are observable in the lateral view of each s pecimen, and the bread th of the base of the saccus is visually different in each of the ventral and dorsal views. The cephalad taper of each saccus is also a quite variable character. JohnsonÂ’s

PAGE 249

233 character state of the more shouldered caudad shape of the valvae seems to hold true for the Florida specimens seen in lateral view with the aedeagus removed of Figures 4-22 l, and 2-23 l. The thickness of the aedeagus in each Figure 4-22. Mitoura g. sweadneri , male genitalia, Putnam Co. of the M. g. sweadneri specimens, including those from Duval Co. seemed as thick, and in some thicker than those of the sample of six dissected from Tennessee hairstreaks (Figure 2-21 w). Ho wever, the widths of the caudad ends of the aedeagi, just before they flare, ar e of even width in Figure 4-24 GRY, and SWD, at 52 X life size. This figure was incl uded specifically to show the spiny tips of the cornuti. Each butterfly has six of these spines on each of two tips. When

PAGE 250

234 the vesica is everted inside the bursa of the female during copulation, or as frequently occurs as a male is dying in captivity, these spines are quite widely separated and spread along the cephalad e dge of the semi-globular inflated vesica. I was not able to evert the vesica of a specimen for preservation, though I Figure 4-23. Mitoura g. sweadneri , male genitalia, Dixie Co. Figure 4-24. Cornuti det ail Caudad tips, aedeagi of both subspecies.

PAGE 251

235 tried several times using differ ent sizes of hypodermic syringes. A ventral view of each of two dissections of Tennessee females is presented in Figure 4-25, and three of the Florida dissections are pictured in Figure 4-26. The broad fan shape to the ri ght in each image has been referred to as the sterigma by some authors and as the genital plate by others. Because Figure 4-25. Mitoura g. gryneus , female genitalia, Giles Co. Tenn. Two individuals GRY 1, GRY 2, and signa of GRY 2. these terms sometimes have broader applic ation to include all the externally visible female features, I have used Kl otsÂ’ chosen term, lamella, for this sclerotized structure (Klo ts, 1970). The long tube cephalad of the lamella, reminiscent of a celery stalk, is t he ductus bursae which leads to the filmy remains of the sac-like corpus bursae at the left of each female figure.

PAGE 252

236 Lamellae of both Tennessee females show strong ridges and deep notches high on their dorsal lateral margins. The fan of these lamellae are some what irregular, and that of GRY 2 (Figure 4-25) has a slight invagination in the caudad center. The enlargement (Figure 4-25 signa) is 150 X life size of the two signa just visible on the walls of the corpus bursae of Figure 4-25 GRY 2. The one on the lower right of the lower panel is on the ventral wall and the more dimly represented signum in the extreme left of this panel is seen from inside the dorsal wall of the bursa looking through the nearly transparent membrane. All five of the Mitoura g. gryneus females had these characters in varying degrees of prominence. The dissections of three female sweadneri in Figure 4-26 show variation in the degree of the state of ri dges of the lamella. The lamella in Figure 4-26 SWD 1 has a stronger ridge on the right hand portion of the fan (lower edge in the figure) than the one on the left. Female SWD 2 (Figure 4-26) has two ridges on either side, and the specimen from Putnam Co. has almost no ridge at all. The general shape of each of these lamellae is different ; though I think the straight line on the left caudad edge of the plate in SWD 1 ma y have been broken in dissection since part of the curve is abruptly missing. All three of thes e dissections have notches in the margins of the lamellae. Those of SWD 1 are more subtly represented than those more dramatic notches of the M. g. gryneus specimens, or of those of SWD 2. The single specimen pictured from Putnam Co. only has a single notch on the right lateral edge and it is reinforc ed with a slight ridge. The signa from the bursa of the Putnam Co. specimen (Figur e 4-26 signa, Putnam Co.) are very

PAGE 253

237 Figure 4-26. Mitoura g. sweadneri, female genitalia Two from Dixie Co. SWD 1, SWD 2, one from Putnam Co., and signa below. faint brown sclerotizations magnified 110 X. The signa detailed (Figure 4-26 signa SWD 2) are visible in the lower left corner of the whole structure dissection

PAGE 254

238 (Figure 4-26 SWD 2). These are shown app roximately 340 X life size. The upper signum in the picture is inside the vent ral wall of the bursa and the lower one is inside the dorsal wall. These two were the most developed signa that I found and were shaped like two halves of a bird’s beak. All nine of the female Florida butterflies had two signa, some more defi ned than others. Ziegler, in his key to species groups of Callophrys specified “Corpus bursa with signum vestigial or absent” (Ziegler, p. 23, 1960) for the “ gryneus group”. Since the next step in the key refers to paired signa, I interpre t Ziegler to mean he had not seen paired signa in Mitoura before he redefined the North American Hairst reak genera. Significance of Genitalic Character States Though I measured aedeagus length in my dissections, after reading Robert Robbins’s analysis of the range of variability of this character state and the range of the degree of cu rvature of the aedeagi in M. spinetorum , and M. millerorum (Robbins, 1990), I decided that si ze was not a species-specific genitalic character in Mitoura . I represented the range of this character as a proportion of the length of the base of the valvae following Clench (1981) in a paper finished posthumously for him by Lee D. Miller and Jacqueline Y. Miller. Robbins also demonstrated t he wide variations in shape of the valvae and sacci (Robbins, 1990), which appear to have limit ed value in separating western Mitoura species. All of the male genitalic features I compared had at least the range of shapes and sizes illustrated in t he figures of the preceding section. When J. W. Brown looked for species distinctions in M. thornei , M. loki , M. nelsoni , and M. n. muiri , he concluded that there was too much variation in genitalic characters to provide reliable means for separating them and instead,

PAGE 255

239 chose colors of alar characters and ma culation, larval host specificity, and asynchronous flight periods as specie s-level characters (Brown, 1982). He acknowledged, though, that wit h large sample sizes specific shapes of valvae and of cornuti might yield quantitative differences. When I found the sclerotized spots on t he walls of the corpus bursae of both hairstreaks, I was hesitant to declare them signa because of the authoritative pronouncement s of Ziegler and of cited above. I asked Jacqueline Miller, then at the Allyn Mus eum in Sarasota, Florida, to examine samples of my dissected female specimens from two Fl orida counties and from Tennessee. Dr. Miller came to the Endangered Spec ies Lab on 26 September 2003 and after looking at the paired structures in several depression plates with the same microscope used in the dissections, confir med that they were indeed signa. This left only the character of the shapes of the valvae (Johnson, 1980) as possibly distinct traits of the genitalia of M. gryneus and M. sweadneri . Brown and Faulkner (1988) after noting the variabilit y of the shape of va lvae and lamellae in Mitoura cedrosensis from California, concluded that variations in these characters may be too broad to be of spec ies-level taxonomic value, although the shape of the saccus might be helpful in diagnosing some of the California Mitoura . It is obvious that some specific geni talic differences ma tter to individual taxonomists who find them useful di stinctions with particular groups of Mitoura . With quantitative analysis, some char acter states may be helpful in demonstrating how closely (or distantly) species are related. In my brief

PAGE 256

240 investigation of the genitalic characters of M. g. gryneus , and M. g. sweadneri , I found wide variations in even small sample s, which occurred in both butterflies. I think the real biological value of each trai t lies in whether it functions alone or in conjunction with others, as a mechanism of reproductive isolation of species. From the evidence suggesting alar characters of these two Mitoura may express a cline in northern Florida, and from re sults of hybridization experiments, reported in Chapter 5, I think the minor differences in the genitalia are not sufficient, alone, as species distinctions. Biological Differences My comparison of deta iled life histories of Mitoura g. gryneus and Mitoura g. sweadneri in Chapter 2 is arguably limited by my choice of breeding stock of the nominate subspecies form only one population in Giles Co., Tennessee. The fact that the summerÂ’s progeny would not ov erwinter in the narrow temperature range of my lab meant that I had to collect new colonists from the first flig ht adults for three years in Tennessee. The captive colo nists from Florida were collected from mostly late spring and summer broods over five years from four Florida study sites, and three of those were from Gulf coast populations in Dixie Co. (review Table 2-1). However, the advantages of using stock from only one site in Tennessee may be significant. First, Giles Co. is located in the approximate geographic middle of the range of M. g. gryneus so the alleles of that gene pool may be far less extreme phenotypic forms than those of, say, M. g. gryneus populations at the northeastern limit in New England, assuming that gene flow is limited by the sedentary natur e of these hairstreaks. Second, the Tennessee site is far enough north of the Florida sites, and the broods asynchronous enough, to

PAGE 257

241 insure that individuals from these populations ar e isolated temporally, and spatially. Still, details of the immature stages from egg characters, sizes of head capsules, duration of larval stadia, and coloration, to behavior of the wanderers are virtually identical. However, at least two other facets of the biology of these hairstreaks need further discussion and explanation. Larval Hosts The larval host specificity of Mitoura species is a fascinating aspect of hairstreak biology. The range maps of t he butterflies and the accompanying table in Chapter 1 (Figure 1, Table 1) are ev idence that many species either choose, or are restricted to, specific host specie s. In addition to the early work cited previously on raising certain Mitoura on Cupressaceae species not normally selected by the females, the l ong lists of host species used by M. siva and M. gryneus attest to the adaptabilit y of these two species over their wide ranges. Current work on the effects of host species specificity in several Mitoura , most notably by Matthew Forister (2004, 2005a, and 2005b) are opening the way to in-depth study of this biology. Host specificity, or use of Juniperus silicicola by Mitoura g. sweadneri , has been mentioned (or touted) by severa l authors (Emmel, 1993, Glassberg et al ., 2000, Johnson, 1976, 1978, 1980, Scott, 1986 and others) as a biological character of the Florida butterfly. Recent refinements in cedar tree taxonomy by Adams (1986, 1993) widely accepted by bot anists today show the differences once thought to separate J. silicicola from J. virginiana are within the range of variations expressed by the nominate c edar. The idea that se eds of the more northern form of the cedar are regularly distributed s outh by many species of

PAGE 258

242 migrating birds and the fact that millions of seedlings grow n from seed collected north of Little’s range of J. silicicola (Figure 2-28) by Florida foresters have been planted across the range of M. g. sweadneri attests to the adaptability of the Florida butterfly to this variety of cedar. Some of these trees still harbor Sweadner’s Hairstreak colonies today. My experiments raising both butterflies on the alternate variations of Red Cedar show ed that it did not interfere with mating or with reproductive success of the but terflies. This preponderance of data demonstrates that cedar taxonomy is more important to taxonomists than it is the butterfly and should not, in this case, be used as a species-specific character. Seasonality of Flights The emergence of the first seasonal flights of M. g. gryneus occur earlier in southern populations than in those from mo re northern climes and several of the afore-cited authors have reported at least on e additional brood of M. g. sweadneri in Florida. I found that four broods occur in the flight season lasting at least 227 days along north latitude 29°. The difference in eclosion strategies between the two butterflies seem s to be temperature-driven. M. g. gryneus uses an extended and delayed second brood approac h, while Sweadner’s Hairstreak takes advantage of the longer growing s eason and favorable temperatures to be multivoltine. My observati on that after the initial flight of F1, remaining M. g. gryneus eclosed in small numbers at near cons tant temperatures in the lab over the remainder of the summer, suggests that this strategy is to some extent genetically controlled and ma y be one of the few dependable non-alar characters separating the two subspecies. The o ccurrence of individuals expressing a possible intergrade phenotype in a few north Florida counties blurs the

PAGE 259

243 species-level distinction. T he distinct alar characters of SweadnerÂ’s Hairstreak centered on the St. Augustine type locale and more vividly expressed in Putnam Co. butterflies modulate somewhat ac ross its range with di stance from the Atlantic coast. This plasticity of phenot ype, taken together with the results of hybridization experiments, suggests Mitoura gryneus sweadneri works well as a subspecies.

PAGE 260

244 CHAPTER 5 HYBRIDIZATION EXPERIMENTS When Mitoura sweadneri was described as a new species (Chermock, 1944), taxonomic decisions were based on characters of type specimens, which in this case were collected from summer flight phenotypes taken from one location. Johnson’s revision of the classification of Mitoura was based on phylogenetic systematics and cladistic classification (Johnson, 1980). While cladistics uses terminology such as “autapomorphy” for a character that distinguishes one species from anot her, and “synapomorphy” for characters shared by related taxa, both of these te rms refer to characters (or character states), so ultimately these analyses are based on typology. Today, cladistics has not yet had a major effect on the systematics of butterflies and new Lepidoptera are still being described as they were 200 years ago (Shapiro, 2002). Even very sophisticated genetic analysis of some Mitoura species has offered little help in clarifying their ta xonomic relationships. Nice and Shapiro compared allozyme allele frequencies, and the mt-DNA cytochrome oxidase subunits from Mitoura nelsoni , M. muiri , and M. siva and found that Mitoura populations in California lacked any signifi cant differentiation (Nice and Shapiro, 2001). In Chapter 4, I demonstr ated the wide variation in alar characters of Sweadner’s Hairstreak across its range and th rough different periods of the flight season. The few characters that may be distinct to the Florida butterfly are

PAGE 261

245 themselves variable in nature and might not prove as durable if the phenotypes of the two butterflies do intergrade into each other at points where the range of the taxa meet. The alar distinctiveness c ould only be evaluated if larger samples from the theoretical border populations were examined. Unless longer tails and larger tornal lobes with bigger black s pots represent species-level distinctions, the typological method fails to distingu ish these two butterflies as species. My assessment of the genitalic characte rs of both sexes of both butterflies revealed only minor differences in the shape of the valvae, first described by Johnson (1980) as a possible distinction. In an array of genita lic structures, the other structures showed a wide range of vari ability even in my small sample. This minor difference in the valvae may or may not withstand closer scrutiny by future investigators. The ultimate relevance of the shapes of genitalic structures is whether or not they are different enough to present a prezygotic isolating mechanism which would prevent butterflies of the tw o phenotypes from interbreeding. The Biological Species concept (sometimes referred to as the Isolation Species concept) articulated by Ernst Mayr (1963) states that species are groups of potentially interbreeding natural populations that are reproductively isolated from other groups. The Tennessee M. g. gryneus population from which I collected gravid females for captive raisin g (to compare life histories with M. g. sweadneri ) is isolated from the Florida butterflies spatially by several hundred miles and temporally by asynchronous broods. The likelihood that indi viduals from these populations would ever meet and attempt to interbreed is remote at best. The

PAGE 262

246 purpose of these hybridizing experim ents was to determine whether these populations were so different as to be unable to produce offspring that could live in the sheltered conditions of my lab colonies. This in vitro “reproductive isolation” could possibly be caused by differences in pheromones, courtship behavior, differences in genitalic structures, gametic, or genetic incompatibility. I tested the reproductive isolation of these widely separated populations in several experiments with three basic obj ectives. First, I wished to determine whether individuals that had been raised on their native host plants would court in captivity. Second, would they copulat e? The third objective, assuming the first two gave positive results, was to det ermine whether these experimental copulations would produce hybrid offs pring, which could then in turn be evaluated for viability of the immature stages and reproductive viability of any adults produced. The first four factor s would represent possible prezygotic mechanisms, while the last one would invo lve postzygotic isolating mechanisms. I have found no published records of hybr idization experiments comparing these two Mitoura , nor have I found hybrid comparis ons of any other species in the genus. As far as I can tell, this is an original study. Methods and Materials I had already determined that captive-raised progeny of each butterfly species would mate in flight cages under certain conditions, and that captive colonies could be maintained. The me thods that worked for intraspecific reproduction were employed as described in Chapter 2. Spec ifically, virgin individuals that had been segregated sexually since eclosion were brought to the flight cage in the late afternoon and intr oduced for the first time. Males were

PAGE 263

247 released first and given time to sort themselves out before the females were introduced to the cage. Details of courtship behavior were recorded. Most of the captive-raised butterflies were accustomed to being handled in the lab with the broken paint brush handl e described in Chapter 2 and did not require handling with forceps except so metimes when being collected at the end of a flight cage session. Using this te chnique, I was able to move individuals around the flight cage relatively unobtru sively and make introductions of some individuals to others. I tried both in troducing females to males and males to females. All the eggs of females who mated we re counted; the number of eggs hatched, pupae produced, and adults eclosed were also carefully noted along with daily observations of the lives of immatures. Usually, each mated female was given sprigs of Florida cedar alter nating with Tennessee cedar every two to three days of oviposition. Most hatchli ngs were transferred to the same small sealed plastic vials used in raising spec ies and some of each cohort were given the other cedar to test host acceptance by the larvae. The greatest challenge in arranging in terspecific mating experiments was having reproductively ready adults of bot h butterflies on hand at the same time. The different timing of their summer broods limited the num bers available and therefore the number of ex periments. On one occasion, I was able to capture a few wild male M. g. sweadneri to supplement insufficient numbers of captive-raised males. For the most part, I tried to keep captive-raised adults of

PAGE 264

248 Florida butterflies on hand and wait for the sporadic eclosions of the incomplete summer brood of M. g. gryneus to produce sufficient adult subjects. Results of Hybridizing Experiments Experiments in mating Tennessee hairstr eaks with the Florida native taxon were conducted over two flight seasons and yielded mixed results, ranging from no courtship at all, to courtship in whic h suitors were reject ed, to copulation. Reproductive success from interspecific copulations also varied in progeny produced and in the number of furt her experiments conducted with these progeny. Results of experiments with M. g. gryneus and M. g. sweadneri are presented in a separate section from fli ght cage experiments conducted with their progeny. To facilitate clear communication, I have used abbreviations for the different taxa which take less space in t able cells than writing out their full names in italics. These names are repres ented by the following abbreviations. GRY = Mitoura g. gryneus from the captive colony of the Giles Co., Tennessee population. SWD = Mitoura g. sweadneri captive bred from Flor ida populations except in one instance in which wild Dixie Co. males were introduced in the flight cage. HYB = F1 hybrids, the first f ilial progeny of a GRY/SWD mating. H/H = F2 hybrids, the second generation of hybrid progeny resulting from interbreeding in HYB/HYB crosses. H/SWD = Progeny resulti ng from the backcross of HYB/SWD . Courtship Males are the obviously active partner in courtship. Courting behavior as described in Chapter 2 for intraspecific mating was the same observed in these interspecies experiments. Briefly, the male lands usually head to head with the

PAGE 265

249 female; he approaches, fluttering his fo rewings a few times, upon which the female either accepts or rejects the offe r. She may step away indifferently, or take flight. Prior to the wi ng fluttering signal, a male ma y take a few excited steps on the cedar branch, which may be an act of final orientation to the female. These steps are noteworthy because males ar e very seldom seen walking, even in the confines of a drink-cup cage. A femaleÂ’s flight to the tree may be the first stimulus because males almost always fly out to meet approaching butterflie s of either sex in natural habitat. However, in the confined space of the flight cage, the distance which newly released females fly to approach the cedar is restricted to only a few feet from the point of release, which may modify the strength of the stimulus. When first released into the flight cage, males rarely exhibited more than a brief version of the protracted whirling bouts obs erved of males on the lek in situ . I think this abbreviated whirling behavior is an artifact of working in a cage only 8 ft tall. When a female accepts her suitor, t here is no observable signal. She simply maintains her perch as the male positions himself alongside her to reach out with open valvae to secure a union. Flight Cage Experiments: M. g. gryneus with M. g. sweadneri M. g. gryneus and M. g. sweadneri were flown together specifically to test their courtship and mating possibilities in se ven flight cage experiments. I did not try interspecies couples in any of the unsuccessful exper imental treatments investigated for single species. The re sults from all seven sessions are listed chronologically in Table 5-1. The dates of each experiment reveal that all the test subjects were from summer broods of bot h species. The participantsÂ’ column

PAGE 266

250 lists the numbers, sex, and taxa of t he butterflies flown together. Times chosen for each flight cage corresponded with t he observed copulation times for both species in previous work. The times listed in the in copulo notes are records of when copulation started and minutes entri es record how long each copulation lasted. Table 5-1. Flight cage experiments: M. g. gryneus with M. g. sweadneri . Date Participants Time Copulation? in copulo 25 August 2000 4 GRY , 6 SWD 17:1518:15 (GRY)00 yc27B X (SWD)00 spv1Ca1 17:23, 47 min 5 June 2001 5 GRY , 4 GRY , 4 SWD 15:4017:07 NO 6 June 2001 9 GRY , 1 SWD 7 GRY , 4 SWD 15:4017:39 GRY X GRY 16:14, 52 min (GRY)01 G13A4 X (SWD)01 SW 17 (wild) 16:46, 55 min 20 June 2001 12 GRY , 8 GRY , 3 SWD 14:4417:30 GRY X GRY 16:01, 37 min 3 July 2001 5 GRY , 1 SWD 10 GRY , 3 SWD 14:5618:30 NO 7 August 2001 7SWD , 7 GRY 15:5117:30 NO 15 August 2001 2SWD , 3 GRY 10:0813:00 NO Session observations The first entry in Table 5-1 was the first interspecies experiment and it is remarkable for several reasons. All six males were released by 17:15 and the GRY females were released a minute apar t after males had selected perches at the top of the cedar and the screen enclosure. Tennessee female GRY 00 yc27B was released at 17:20 and courtship must have begun as soon as she flew to the tree. She was discovered in copulo three minutes later. In the GRY colony

PAGE 267

251 notes of the morning before the flight cage session, I had observed that she was already 12 days old and had laid 164 virg in eggs on Tennessee cedar, and I had decided she was really too old for the experiment but included her to have more activity in the experiment. Not only was she the only female that mated, none of the other butterflies even courted. T he session was ended as a thunderstorm began. The second experiment of 2001 was dr amatic because even though it only lasted 2 h, 10 were flying with 11 , offering 110 possible pairings. Four wild SWD had been captured a few days before to give the seven captive GRY competition. SWD were all flying by 15:40 and all the females were free by 15:54. The butterflies were perched and most ly still, basking, or seeking shade after the initial flurry of release. No courtship was noted. At 16:05 the GRY were released and the cage whirled for a few minutes, until the GRY couple mated at 16:14. All the butterflies took refuge in the shade of the cedar or north corners of the enclosure from the direct afternoon sun, recording 42°C in direct exposure and 38° (110°F) in the shade. At ~5 min intervals, 5-6 hairstreaks would jump up for a moment then settle ba ck into the shade. At 16:46, I had just tried for the third time to introduce GRY 01 G13A4 to a cedar-perched SWD on the tip of the paintbrush handle when wild 01 SW17 , after a very brief courtship, stepped onto the handled to jo in with her and begin copulation. I transferred them to her cup in the shade and they remained in copulo until 17:41, the longest record in this series. I induced a second SWD/GRY courtship after a

PAGE 268

252 few attempts at 16:57; she did not fl y and he did not reach for her, but approached her abdomen with his head. He s pent some minutes uncoiling and recoiling his proboscis before they flew apart. The session was called at 17:39, just in time to collect t he rest of the participants bef ore thunder at 18:00 signaled the approaching storm. The temperatur e dropped to 28.5°C before the rain began. Hybrid progeny The data for GRY 00 yc27B is listed in Table 5-2. Her hybrid egg total is in the upper middle of the long-laying GRY fema le range of oviposition, but when her previously laid 164 virgin eggs are added, her life-time egg production of 345 is among the top four contenders of the species (Chapt er 2). The eggs were laid over 28 days of her 35-day life, includ ing 19 oviposition days post copulation. She was fecund. Her eggs were laid in clutches of 2169 in five separate oviposition cups, three with Tennessee cedar, two with local J. v. silicicola . The hatch success was also good at 84%, but larva survival wa s not good. Many of the hatchings died before they were moved to individual pl astic vials and about half of those moved died before the molt to second instar. Seve ral did not progress to third instar. The larvae from eggs laid on the local cedar fared a little better and produced most of the pupae. I do not attribute t he early larval death rate directly to host intolerance by the hatchlings on J. virginiana because their dam was raised on it, imported fresh from Tennessee. This cedar was not as plump as the local cedar to begin with and it may have dried too quickly for the comple te nourishment of the hatchlings. I do not have enough data to det ermine if the death rate was due to

PAGE 269

253 hybrid inviability, particularly because so many survived embryogenesis to hatch as first instars, and some that did not liv e to pupate made it as far as third instar. It is more interesting that almost all the pupae were from the sixth instars, except for two fifths, and one, 00 hyb4F , that fed for seven instars and rested in diapause for 107 days to be the first captiv e-raised butterfly to overwinter in my lab. Usually only 17-23% of pupae are from the sixt h instar and the seventh instar pupa is rare, at less that 2%. Table 5-2. Hybrid progeny Couple Eggs laid Hatched Pupae Adults 00 yc27B X 00 spv1Ca1 181 152 28 23, 8 , 15 01 G13A4 X 01 SW 17 355 249 66 53, 27 , 26 The second GRY to mate with a Florida butterfly was also rather fecund. Hairstreak 01 G13A4 laid 355 fertilized eggs over 36 days post copulation for a life-time total of 379, including virgin eggs. She lived for 44 days, which is near the record for my captive colonists. The hatch success was 70% and half of the immatures that survived to pupate fed on Tennessee cedar. Of the 66 pupae, 81% eclosed adults. These butterflies were t he stock for most of the rest of the hybrid experiments. Figure 5-1 shows two examples of GR Y/SWD hybrid wing characters. The female has larger tornal lobes with big black spots associated with Florida butterflies and long tails. Her Thecla spot has a large bla ck center but with the orange lunule associated with Gulf coast M. g. sweadneri and Tennessee butterflies. The femaleÂ’s VHW PML has the more crooked look of M. g. gryneus ,

PAGE 270

254 while the maleÂ’s is slightly straighter . His TS is typical of the Tennessee population and his tails are also much shorter than hers. As with the alar character figures of Chapter 4, these are not necessarily typical of this cross. Indeed, specimens from these two broods s how at least as much variability as that of any species brood. Most of the hybrid progeny was flown in mating experiments and did not keep as well as t hese two. It is interesting that both exhibit the dorsals typical of the firs t flight Tennessee population. The male (Figure 5-1) is hybrid 00 hyb4F that ov erwintered in the lab. The female is marked for the flight cage; t he brown stain of her right VHW is from fluid lost after the specimen was pinned. Figure 5-1. Hybrids, GRY X SWD , v) ventral, d) dorsal aspects.

PAGE 271

255 Flight Cage Experiments: Hybrids While it was unlikely that representat ives of the two distant populations would ever have met or copulated, t heir reproductive success demonstrates physical and gametic compatib ility, and evidence that th ey are not reproductively isolated. Data presented in Table 5-3 show that their hybrid progeny were willing to mate in five out of seven trials. Two of the 45 possible mating pairs flown in the 2000 season did copulate, and all of the participants of that series were F1 hybrid siblings, progeny of GRY 00 yc27B . These experiments were less than ideal Table 5-3. Flight cage exper iments: F1 Hybrids, (GYR X SWD ). Date Participants Time Copulation? in copulo 25 October 2000 3 HYB , 3 HYB 16:1017:45 NO 27 October 2000 6 HYB , 5 HYB 15:4517:40 00 hyb1J X 00 hyb3D 00 hyb2D X 00 hyb2J 16:15, 43 min 16:48, 44 min 8 November 2000 2 HYB , 3 HYB 15:1517:00 NO 17 August 2001 5 HYB , 4 GRY , 2 SWD , 4 HYB , 1 GRY , 4 SWD , 17:1519:25 01 13A4hyB17 X 19 SW 11 01 13A4hybB20 X 01 13A4hybB18 17:44, 40 min 18:13, 57 min 24 August 2001 6 HYB , 7 HYB , 4 SWD 16:5819:05 01 13A4hybE16 X 01 13A4hybG20 17:24, 29 min 1 September 2001 6 HYB , 8 HYB 16:0819:32 01 13A4hybC4 X 01 13A4hybF6 01 13A4hybG13 X 01 13A4hybG4 01 13A4hybH1 X 01 13A4hybC3 17:00, 36 min 17:34, 41 min 17:35, 46 min 16 September 2001 2 GRY , 2 HYB 17:2019:10 01 G6F2 X 01 13A4hybE4 18:01, 52 min

PAGE 272

256 because they were conducted later in the season than either species is known to fly. The ambient temperatures for both the October and November trials were right around 27° C, only slightly warmer than the constant temperature of the lab. Much of the time in all three flight cage sessions was devoted to basking, or sheltering from the moder ate autumn breeze. Earlier sunsets also made for shorter sessions. For the last four flight cage ex periments of August and September 2001 (Table 5-3) at least one of the participant s in every mated pair was an offspring of GRY 01 G13A4 . These experiments were run in favorable temperatures, in generally sunny conditions and all but the last one (16 September 2001) had the advantage of abundant males. The most inte resting aspect of these sessions is the inclusion of species individuals am ong the hybrid participants allowing the possibility of backcross mating. The 17 August experiment had more HYB than either species, and the HYB were one shy of a majority of ma les. With 20 butterflies flying and 99 possible mating combinations, there was littl e time to try introductions. With so many participants, I had anticipated a liv ely session, but the butterflies were mostly passive after the sorting flight s following release, until about 18:00. The session began with 34.5°C in the shade and the temperature rose to 36°C ( 47°C in direct sun) inside the enclosure, so m any of the captive-raised butterflies took refuge in the shade, as wild SW D do on hot afternoons. The HYB /SWD backcross started at 17:44 after a very br ief courtship, and a few of the females laid virgin eggs halfway up the cedar. At 18:04 something changed and individual

PAGE 273

257 butterflies who made short flight s stirred most of the male s to fly up and resettle. This happened several times until 18:47. The HYB/HYB mating was unremarkable except that it lasted 57 min, setting a new record duration. Overall, very little courtship was observed in this crowded experiment. The 24 August and 1 September are inte resting because in the first, 16 hairstreaks were flown, including four SWD offering 66 possible pairings, yet only one mating occurred in over 2 h. By contrast, the 1 S eptember session had only 14 all HYB participants, offered 48 ma ting combinations, and three couples joined. All of the participants in both sessions were in good condition. A new behavior was observed on 24 August when one of the SWD became very excited, did the jerky walking, and brief flu tter of courtship inside the tulle cover of his cup while it was still on the fl oor of the flight cage. The HYB were released first at the beginning of the session, then the females. The other three SWD followed, but the last one had been forgotten. His solo courtship behavior was noticed right after the HYB/HYB copulation began and he was released immediately. The last flight cage session on Table 53 had the least participants of any session that resulted in copulation. It was conducted specifically to try what resulted in a successful GRY/HYB backcro ss. It is interesting that only two female M. g. gryneus were flown with only two male hybrids and a mating occurred. Results of my previous experiments suggested courtship and copulation were most likely to occur when three or more, or and access of males were flown.

PAGE 274

258 Progeny The results from these nine matings of Table 5-3 are listed in Table 5-4. Two of the eight mated HYB laid no eggs at all. This is unusual because most all captive-raised females begin laying vi rgin eggs by 5-8 days after eclosion. They both developed eggs and died with dramat ically swollen abdomens. Both of these females were of the 2001 HYB c ohort. One of them, 01 13A4hybE16, was part of the 24 August copula lasting 29 min, the shortest recorded. Two of the remaining HYB laid eggs that did not hatch and the rest of them had very low hatch rates. Interest ingly, 18 of 23 F2 hybrid pupae eclosed adults. So, if there was a lack of game tic compatibility among these hybrid siblings, it was not complete. Table 5-4. F2 Hybrids and backcross progeny Couple Eggs laid Hatched Pupae Adults 00 hyb1J X 00 hyb3D 61 0 0 0 00 hyb2D X 00 hyb2J 67 9 1 F2 HYB 1 01 13A4hyB17 X 19 SW 11 206 7 2 (HYB X SWD) 1 , 1 01 13A4hybB20 X 01 13A4hybB18 108 21 14 11, F2 HYB 5 , 6 01 13A4hybE16 X 01 13A4hybG20 0 01 13A4hybC4 X 01 13A4hybF6 102 0 01 13A4hybG13 X 01 13A4hybG4 0 01 13A4hybH1 X 01 13A4hybC3 111 21 8 6, F2 HYB 4 , 2 01 G6F2 X 01 13A4hybE4 156 131 48 33, (GRY X HYB) 17 , 15 , 1 ?

PAGE 275

259 The more telling result was from the HYB/SWD backcross. Female 01 13hybB17 laid 206 eggs but just over 3% hatched. The SWD may not have made good sperm, but considering the evidence from low success of the sib couples and the respectable results of the GRY/HYB backcro ss, I think the problem may have been more with the HYB or the eggs they produced. The successful hatching of 131/156 eggs laid by GRY 01 G6F2 indicates that her HYB partner produced viable gametes and he was one of the HYB sibs. Her total adult progeny might have been substantially better; at one time there were 105 live, feeding larvae. At least 50 of thes e died soon after a white mold developed on the cedar inside the plastic larval vials. The 18 F2 hybrids and the single pai r of HYB/SWD backcross butterflies eclosed in the fall of 2001 after the normal observed flight season of M. g. sweadneri and even later than that of the T ennessee population. Sixteen of the F2 HYB were flown in the flight cage in seven sessions, and the backcrosses were flown once alone. The temperatures were lower than desirable for the last six sessions, ranging between25-26°C. Cour tship was observed in four of these six trials, including between the H/SWD hybrid couple but none led to copulation. Temperature may have been the influenci ng factor in leading to this low performance. However, the courting ma les in these sessions were more persistent than in any previous observati ons. In the 26 October flight cage, with only one pair flying, there were three atte mpts each lasting 2-3 min. One of these was on a horizontal support of the screen enclosure wh ere he seemed to grab her with his forelegs while waving hi s abdomen wildly, valvae agape. The 10 and

PAGE 276

260 11 November sessions had more of this persistent courtship, including one attempt with a female resting on my left thumb, and another on my glasses frames and the bridge of my nose. For the first two sessions (19 and 21 Oc tober), the temperatures were 30° and 32°C –within, though on the lower end of, the range at which copulation occurred. Only the first session produced a mating, and the female, F2 HYB 01 h/h hybB20D7, died seven days later without laying any eggs. Table 5-5. Flight cage experimen ts: F2 Hybrids and HYB/SWD backcross Date Participants Time Copulation? in copulo 19 October 2001 3 H/H , 3 H/H 16:5318:54 01 h/h hybB20 D7 X 01 h/h hybB20 D8 18:14, 52 min 21 October 2001 2 H/H , 3 H/H 16:5718:21 NO courtship 26 October 2001 1 H/SWD , 1 H/SWD 16:2018:09 courtship only 4 November 2001 3 H/H , 2 H/H 16:0117:08 NO courtship 7 November 2001 3 H/H , 3 H/H 15:5017:08 courtship only 10 November 2001 5 H/H , 3 H/H 15:2516:38 courtship only 11 November 2001 4 H/H , 3 H/H 14:2516:55 courtship only 26 November 2001 3 H/H , 2 H/H 15:1516:28 NO courtship Figure 5-2 is a collection of ventral shot s of some of the variations in wing maculation of F2 hybrids and the backcro sses. The dorsal wings were mostly solid dark brown in the F2 hybrid s pecimens, though some had varying amounts of gold dusting. There were only two progeny of the HYB X SWD backcross and both had solid brown dorsals without any dusting. However, among the 33 offspring of the GRY X HYB backcross, every degree of gold, from full discs to light gold dusting, to solid dark brown dorsals, was evident.

PAGE 277

261 Figure 5-2. F2 Hybrids and Backcross progeny The ventral characters also varied widely, but no particular patterns were evident in these small samples. The rela tively straight VHW PML most frequently seen on the Florida butterflies is shown next to the more crooked VHW PML of M. g. gryneus manifested to some degree in ever y pair. The large tornal lobe with

PAGE 278

262 its larger black spot persisted in most specimens, but the longer tails accompanying this character in M. g. sweadneri were expressed more consistently in the F2 hybrids, and in the small HYB X SWD backcross sample. The Thecla spot shows a variety of presentat ions, with the black eye generally a bit larger than in the Tennessee hairstrea ks but differing emphasis of the orange lunule so that on this character alone, so me of these specimens would fit in with the Putnam Co., Florida colony and some might be seen in Giles Co., Tennessee. The BB varied from absent in the male of the HYB X SWD backcross to emphasized in the male GRY X HYB specimen. Finally, the thin brown line (TBL) which I tracked in the M. g. sweadneri specimens (Chapter 4) seemed to be missing or only vary slight in most of the hybrid progeny, but appeared with some emphasis in about half of the GRY X HYB backcross specimens. Discussion and Summation These flight cage experiments dem onstrate some degree of success in each of my stated objectives. Courtship responses were observed in every taxon combination flown except by a GRY to a SWD . This was the most difficult combination to test becaus e only four virgin SWD were available for three out of the 10 sessions with full species participants. Seven GRY flew with only one SWD in one session, and 10 GRY were flown with another single female (Table 5-1). One session of the primar ily F1 test series had a single GRY

PAGE 279

263 faced with two SWD in a mass of 20 participants. I think it is quite likely this combination would have mated if greater numbers of SWD had been flown. At least one of every comb ination that courted did proceed into copulation including the F2 hybrid tests (Table5-5). So , slight genitalic differences were not a physical barrier to reproduction. The fact that both of the interspecies copulations produced large numbers of fertile eggs suggests gametic compatibility, at least in this small te st. That so many of the hatchlings died without feeding, or failed to pass the sec ond molt, could signal low hybrid larval viability except that part of the lack of success may be attributed to cedar quality. This F1 hybrid survival rate would need to be tested in several repetitions before making conclusions; even full species larvae have poor survivability at low humidity, or with poor food quality. The F1 hybrid copulations (Table 5-4) show the greatest range of results in numbers of eggs laid, 0-111, with only two crosses producing near the middle of the range. While 100 eggs is not near the high range of either species, only 60 or so eggs were laid by some SWD in my captive colony, and a few females of both species never laid eggs. The fact that 51 of 449 F1 HYB femaleÂ’s eggs hatched (11.4%) does seem significant. It could be that the gametes of F1 HYB are not very compatible, or that since all of the partici pants were siblings, this could be a result of inbreeding depression. The alar character photographs (Figures 5-1, 5-2) point to a problem any experienced taxonomist woul d face when confronted with a Cornell drawer full of my specimens. Is there enough distinctivene ss in the two taxa to tell them apart

PAGE 280

264 on sight, or does the presence of po ssible intergrades and maybe hybrids increase the variability of alar character st ates too greatly to define the points of distinction? With the demonstrated lack of reproductive isolation between these two distant populations, I think all thes e character states are possible to encounter in populations distributed between Florida and Tennessee.

PAGE 281

265 CHAPTER 6 ATTRIBUTES OF SUCCESSFUL COLONY HABITATS In the report of my fiel d study, I presented a list of 90 cedar sites surveyed for SweadnerÂ’s Hairstreak (Chapter 2, Table 2-14). The list includes naturally occurring cedar groves, planted shelter belts, and some urban landscapes. A few of the rows of cedars had been planted in rural areas that had become suburban settlements before I began the study. An y one of the sites visited between February 1999 and May 2004 in which the butterfly was found was given a waypoint (WPT) number that began with 0, and each identified site was visited at least five more times during the survey. Some of the accessible 900-series WPT of Table 2-14 were visited numerous times in hopes of seeing sweadneri . Almost all of the early expl orations began with a visit to an inhabited site to verify that the hairstreak was flying, before c ontinuing my search for new locations. Before the end of the 1999 s eason, I realized that with study organisms usually occurring in low numbers, it was reasonable that not all presumed colonies would have observable individuals present every day that they we re found at another site. I surely missed colonies at some cedar stands because I did not return a third or fourth time. Kathy C. Malone, a butterfly enthusiast at the Florida Museum of Natural History, sent me a digital photograph of M. g. sweadneri taken at OÂ’Leno State Park (WPT 915) thr ee years after I had quit looking there. The advantage of starting at a known inhabit ed site was that t hat I could drive 6075 mi north or south, stopping briefly at previously hunted cedar stands, before I

PAGE 282

266 found a new cedar grove with butte rflies and gauge the distance between colonies by GPS reckoning. Surveys did not take long if no hairstreaks were flying. Thumping each cedar yields inst ant results; a perching SweadnerÂ’s Hairstreak flies out and back in a characte ristic sharp U-turn. The green ventrals and bright white PML are easily veri fied with good binoculars. So, an empty grove of 47 trees could be surveyed in 20 min and a planted row of 100 in less than 30 min. Each tree with the per ching butterfly was marked with green flagging tape and given a number for record keeping. Table 6-1. Mitoura g. sweadneri observed abundance WPT County Times present Max seen Total trees Marked trees 001 St. Johns 4 5 152 4 002 Dixie 22 11 53 26 003 Dixie 9 11 46 12 004 Dixie 4 5 20 2 005 Dixie 2 3 18 3 006 Taylor 2 8 37 8 007 Dixie 18 3 7 3 008 Flagler 2 2 16 2 012 Alachua 2 3 160 2 013 St. Johns 1 1 88 1 014 Dixie 2 3 8 2 015 Dixie 1 4 10 3 016 Levy 1 2 42 2 017 Levy 2 6 32 2 018 Levy 2 8 128 7 019 Putnam 22 44 92 40 020 Dixie 10 5 24 7 021 Alachua 2 1 10 1 023 Alachua 4 2 22 3 Totals: 110 472 cum. total 965 130

PAGE 283

267 Table 6-1 is a list of the 19 WPT where the butterfly was seen. All of these were assumed to have at least ephemeral colonies except for WPT 021, located across the street from my home, which provided cedar for my captive colonies and only had one perching individual on one day in May, in each of two flight seasons, four years apart. The “Times pr esent” column (Table 6-1) is the number of surveys in which a site harbored the hairstreak. Dixie Co. WPT 002-005 and 007 are all within an approxim ately 9-mile stretch of County Road (CR) 361. The “Total number of trees” in a transect, and “marked trees” columns show that even small stands of cedar can support stable co lonies year after year. However, the times present column does not bear co mparison for WPT 003-005 in the Dixie Co. cluster because they were destroyed at the beginning of the 2001 season. Several of the colonies listed with lo w numbers and low times present were ephemeral in that the butterf lies were there, then absent for several field trips and present again in another flight season. The Taylor Co. site was probably productive, but the 37 cedars were sit uated around a fishing cabin where I was welcome, and in the front yards of th ree private rural homes. One of the homeowners was alarmed to find me wit h two field assistants thumping trees along his fence and the site was dropped from the survey. Waypoint 018 is, I think, a natural grove of 128 cedars in Y ankeetown, Levy Co. It has a long history of records from several collectors repres ented in the Florida St ate Collection of Arthropods (FSCA), Emmel’s records, and those of several members of the Southern Lepidopterists’ So ciety, but it apparently produced sporadic broods during my study.

PAGE 284

268 Three well-researched sites from the list were selected for more detailed comparison. Two were very productive sites in very different habitats: one a naturally occurring cedar grove, and the ot her comprised of a series of planted shelterbelts. The third is repres entative of the type locale of Mitoura sweadneri . Dixie Co.: WPT 002 On one of my first explor atory fieldtrips to Dixie Co., 5 July 1998, seven months before my field study began in earnest, I collected two female M. g. sweadneri to try captive culture of the bu tterfly. The site was a cedar grove located on the west side of CR 361 about 7. 5 mi south of the intersection with CR 358 near Jena. The first tree in which I sa w SweadnerÂ’s Hairstreak is shown in Figure 6-1 A; the lower picture, Figure 6-1 B, is of the cedars mixed with deciduous species typical of roadside trees on this stretch of CR 361. I began my dissertation field research on 6 Februar y 1999 and the first survey started with this tree. Though no hairstreaks were obser ved that day, this site was recorded as WPT 002 on the strength of my prev ious find, and I thumped 33 cedars, including the line shown in Figure 61 B, and 12 on the out side edge of a dense cedar grove on the east side of the road. Over the next fe w visits to Dixie Co., no SweadnerÂ’s Hairstreaks were seen; the firs t flight did not eclose until after 20 February that year and my first recorded si ghting in this area was on 27 February at what was to become WPT 003, 0.3 mi south of 002. On 18 February 1999, a controlle d burn was set on land of the Steinhatchee Wildlife Management Area a few miles north of WPT 002 on the west side of CR 361. This fire, fanned by unexpected winds, escaped control and quickly burned, from about Sink Creek, four miles south along the west side of

PAGE 285

269 Figure 6-1. Original tree A), and habitat B), WPT 002 CR 361, scorching or killing 80-90% of the cedars visible from the road. My wife, Linda, and I discovered the destruction while stumps were still smoldering on 20 February and photographed the original WP T 002 tree and a stand of nearby cedars (Figure 6-2 A and B). This south reach of CR 361 is bordered on the west by Pine Log Swamp and on the east by Rocky Creek Swamp. Cedars grow along the edges and on spits of slightly higher ground, and the driest land above the water supports pine plantations and Gulf c oast maritime forest. The cedars which

PAGE 286

270 Figure 6-2. Burned original tree A), and habitat B), WPT 002 were burned as badly as the ones in these pictures died, though some individuals at waterÂ’s edge were not as extensively damaged, and began to grow back during the course of the study. In Figure 6-2 B, bare sand substrate can be seen between burned clumps of grass and showing through the completely incinerated duff beneath these charred S outhern Red Cedars. Though the fire jumped the county road in several places, the destruction on the east side was minimal and restricted to pat ches of less than 0.1 hectare (0.25 acre). The dense

PAGE 287

271 grove on the east side of the road at WPT 002 was spared and became the focus of my study at this site. During the remainder of the 1999 fli ght season, WPT 002 was extremely difficult to survey. It consisted of about 97 cedars that covered an irregular rectangle 13 m east of the pavement, spanning 78 m north to south along CR 361, and 56 m deep (0.43 ha). The eastern and southern edges of the grove were bordered by Rocky Creek Swam p and the space between the lower expanse of the cedar branc hes and the water was grown up in thick Black Needlerush. The cedar canopy was almost complete with 6 or 7 taller pines, and a palmetto protruding above, and some oaks intermixed at the margins. The shade under the cedars was deep, so li ttle understory flora survived. Lower branches of the cedars interlaced so tightly that it was not possible to enter the grove to survey all the trees, so th e first transect included only 20 cedars accessible at the edge of the grove, and I found the hairstreak in four of these on subsequent visits during the 1999 and 2000 seasons. A second anthropocentric interventi on at WPT 002 opened up this tightly spaced grove, providing access for the re mainder of the study and a glimpse into the life of the colony. Sometime during the second week of February 2001, the tall pines were cut from inside the canopy of the cedars. Access to the bases of the pines was achieved by driving a tracto r directly through the grove. The pines were felled and then dragged out to the edge of the road for loading. During this process, half of the cedars were pus hed down or broken off at the ground and left in a disordered jumble. Figure 6-3 A was representative of my first view of the

PAGE 288

272 Figure 6-3. Grove after pine harves t, A) exterior, B) inside WPT 002 site on 16 February 2001, the first day of the 2001 flight season in Dixie Co. In the center right of Figur e 6-3 A, two pushed-down cedars can be seen leaning against trees on the south edge of the grov e. One of them was a marked tree from the previous year. Figure 6-3 B wa s a typical perspective from several places within the grove. In the center of this carnage, a tall cedar whose lower limbs had died years before, as a result of the dark shade of the previously solid

PAGE 289

273 cedar canopy, was marked that day as a lekking tree. Despite the disastrous intrusion, four M. g. sweadneri were seen on this visit. The path of the tractor opened a transect which was surveyed on every succeeding visit. Before the pine harvest, I had not been able to see the tops of all the trees and could only pass into a few spaces between about 12 of t hem before the mass and density of their lower branches and that of the next nearest tree became a tight squeeze difficult to a ccomplish with even minimal equipment. From inside the grove, I could tell that the trees around the perimeter were not the only ones with lower limbs still green. However, the perimeter trees and outliers from the groveÂ’s margins were mo st likely to have the conical shape with full green leaves on branches close to the ground, which I eventually recognized as the major factor required for a c ontinuously occupied, stable SweadnerÂ’s Hairstreak colony. The gr ound not scraped and torn by the tractor was covered in a layer of cedar duff and fallen pine stra w. Standing next to a once-conical cedar damaged by the tractor, I could see inside the insulating cover of leaves at the ends of branches in my first cut-a-away view of the physica l structure of the cedarÂ’s pupa-friendly microclimate. Numbers The total number of trees surveyed along the transect in its final form (Table 6-1) was 53 (35 ; 18 ). Of these, 26 (18 ; 4 ) had the butterflies present on at least one survey date. Waypoint 002 was visited 34 times over five flight seasons and the hairstreaks were present on 22 occasions. The maximum number seen on any day was 11. This popul ation was present consistently in

PAGE 290

274 every brood surveyed at Dixie, Taylor, and Levy County sites. The next most consistently occupied site on the Gulf coast was WPT 007, located 6.6 mi north on higher ground. Disposition After the pine harvest, the site was productive for sweadneri during the next three flight seasons and the hairstreak numbers seen we re definitely improved with improved access. By the end of 2001, pioneer plants had settled the open spaces and during 2002 and 2003, I carried pr uning shears on almost every visit to clip back shrubs and keep the transect open. At the same time in February 2001, pines at WPT 003, 004, and 005 were cut, but most of the rest of the trees were eventually cl eared in what appeared to be preparation for planting pines (Figure 6-4). I sampled the few remaining Figure 6-4. Cleared after the pine harvest, WPT 003 cedars on the west side of CR 361 across from these sites 12 times each, but only WPT 004 had the butterfly and only on a few occasions. During the highly active hurricane seas on of 2004, the Gulf coast of Florida was subject to heavy rain and several st orm surges from t he major hurricanes.

PAGE 291

275 Inland rivers flooded and Pine Log Swamp and Rocky Creek Swamp rose to levels not seen in any of the rain ev ents of 1999-2003. I visited WPT 002 at one point when the remaining cedars were inundated and water flowed through the transect destroying the duff layer. I think this site will not support butterflies for at least several years, if ever again. Natural succession by faster-growing deciduous species will probably crowd and shade the cedars. Waypoint 007 will undoubtedly be the nearest studied colony to survive in the short run. St. Augustine Lighthouse: WPT 001 The habitat at St. Augustine Lighthouse sp rawls over 2.4 hectares (6 acres) on Anastasia Island, with a boat ramp into Salt Run and access to the Atlantic Ocean. It comprises 1.4 hectares of Light house Park situated in front of the St. Augustine Lighthouse State Museum, and an undeveloped square of about a hectare owned by the St. Augustine School District. The City of St. Augustine manages its park for picnicking and parki ng for boaters who use the public boat ramp. The school districtÂ’s holding is a remnant dune hammock of oak and saw palmetto bordered on four sides with cedar trees. This square is not actively managed. A well-maintained baseball field behi nd R. B. Hunt Elementary School is part of the districtÂ’s property. I was given a key to the ball field gate for research access because it was locked in early 2002 to exclude dog walkers from the playground. The whole transect at WPT 001 comp rises 152 cedars, 43 of which are berry-bearing females and 36 of which prod uce pollen. I think the sex ratio is 50/50 but could not confirm this. Many of the cedars surrounding the park and undeveloped square are probably quite old, judging by trunk diameters, and may

PAGE 292

276 be remnants of a natural grove. But at least half of the transect trees were planted in the past 50 years and none is over 7 m tall; most are 4.2-6 m and the average height is 5.6 m. The shape of these trees is characteristic of the windswept habitat close to the sea. T he trees stretch along both sides of Lew Blvd. between the ball field and Light house Park, and partway along Santa Monica Ave east toward Salt Run. Eleven are along the outfield fence of the baseball field and in adjoining backyards of the homes on Lew Blvd. Figure 6-5 A shows four trees where I first saw the butterf ly on my fourth survey of the site, on 3 April 1999. All four are female cedars on the edge of the left field fence of the baseball diamond. In 24 visits to WPT 001 over five years, these were the only ones to ever have perching sweadneri butterflies, called St. Augustine Hairstreaks in the local vernacular. Fi gure 6-5 B shows part of the extensive picnic area of Lighthouse Park and depicts several old cedars with all lower limbs removed to accommodate public use. Ther e is no duff layer here, and much of the well-trod ground is bare s and. On the lawn of St . Augustine grass around the restaurant (not pictured) are several c edars that lack lower limbs and each has a circle of cleared earth surrounding its tr unk to facilitate ea sy mowing. Figure 6-6 A shows one of many shady parking pl aces on the perimeter of the school districtÂ’s square and the cedar limb extend ed in the left foreground is covered in juniper berries. None of the tr ees on either street in th is figure has lower limbs on the street side. Figure 6-6 B depicts two old cedars in the boaterÂ’s parking lot of Lighthouse Park. This image was chosen to show the mass of Muscadine

PAGE 293

277 Figure 6-5. St. Augustine Lighthouse Park A) occupied trees, B) picnic area, WPT 001 Grape vines covering these two trees of undetermined sex. Several of the trees around the parking lot and the school distri ct tract are vine-covered and some on the Lew Blvd. side are wrapped in Dodder, or Love-vine, Cuscuta gronovii . Unchecked this vine forms a mat which can kill some vegetation and though I have not seen cedars killed by it some had lower limbs completely wrapped on the side nearest the street.

PAGE 294

278 Figure 6-6. Lighthouse A) shady parki ng, B) with Muscadine vines, WPT 001 Numbers Mitoura g. sweadneri was only observed in this habitat on four occasions between 3 April and 22 June 1999 over fi ve years of surv eys. Only eight sightings were recorded and five of those were on one date. This and nearby cedar sites were evidently productive in the 1980s, based on label data of available specimens, and WPT 001 is possibly the type locale where Dr. Sweadner collected his long series in 1940 (Chermock, 1944). Some of the landscape improvements to Lighthouse Park are more recent than when Emmel and his crew worked there in the 1970s to late 1980s. Based on the limited

PAGE 295

279 recorded history of the site and my ow n research, I think the low numbers and absence of St. Augustine Hairstreak is due to habitat degradation through deliberate management practices, lack of management, and rampant development of St. Johns Co. Severa l of the sites recorded in EmmelÂ’s chromosome work no longer have cedars at all, and one is the site of a McDonaldÂ’s. Loss of natural area between widely spread colonies could make travel or dispersion difficult for this sm all butterfly, especially when replaced with tidy lawns and pavement with heavy tourist traffic. History and Disposition This site was chosen as representative of an urban habitat and because cedar trees are protected in the city of St. Augustine by an ordinance, number 86-61-A, enacted by its citizens on 23 F ebruary 1987 forbidding the cutting or destruction of any Southern Red Cedar without a permit issued by the City Building Official. This mo vement by the well-intent ioned people of St. Augustine was one of the earliest effort s by any municipality in th e United States to protect a butterfly. Though the colony seems to be missing from WPT 001 proper, and the butterfly was not found more than once in nearby Anastasia State Park (WPT 013), I think there is a possi bility of the butterflyÂ’s presence at WPT 013 because cedars have been planted along the state parkÂ’s entrance road and near the picnic area within the last ten years which may soon grow to support a future colony. Cedars like those along Lew Blvd . in Figure 6-7 could possibly be restored as butterfly habi tat since most have lower limbs intact, but the encroaching deciduous trees and vines would need to be carefully removed and

PAGE 296

280 a long term maintenance policy instit uted which would prohibit mowing beneath the cedars in order to a llow duff build-up in the c onical cedar microclimate. Figure 6-7. Possible habitat, WPT 001, east side of Lew Blvd. Cedars around the municipal offices of the city of St. Augustine Beach, south of St. Augustine Lighthouse, once s upported a well collected colony but the cedar trees in the landscaping have had t heir lower limbs removed to make way for flower beds in recent years and w ould no longer be useful for the butterfly except as lekking trees. Sites which I ex plored along the St. Johns River in this area have large old cedars, many without lower limbs, wh ich would not qualify as present or potential habitat.

PAGE 297

281 Hollister, Putnam Co.: WPT 019 The study site at the Division of Forest ry on State Highway 20 in Hollister, Putnam Co., is included in this discussi on as an example of perfect habitat for Sweadner’s Hairstreak. In this site, Southern Red Cedars were planted as a shelterbelt to provide privacy from t he busy highway for the foresters who live there. The transect at WPT 019 comprised 92 cedars in three sub-transects separated by dry, open sandy yard with spar se grass. A row of 16 was planted in 1972 along Frank Cone’s east side yard oriented north-south, beginning 16.3 m south of the pavement’s edge. There were an additional 14 trees from an earlier planting which are scattered eas t of these, including the “limbless tree” monitored in the microclimate experiments (Chapter 2). A second row of 20 trees was planted in 1967 on Mr. Cone’s western boundar y with the forest tower property, 66.7 m west of, and parallel to the first ro w (Figure 6-8 A). The third transect ran east-west, parallel to, and 27.2 m south of the pavement edge, and comprised 32 cedars also planted in 1967. An additional ten smaller trees planted 20-25 years later filled in the screen of 42 trees (Figure 6-8 B) near its west end. Fifty of these trees were female; 38 were male, and 12 were not sexually mature. The average height of the three transects was 9.23 m. Figure 6-8 C is an inside view on the forester’s side of the east-west shelte rbelt, looking west. These 92 trees were spaced at distances of 2.76.5 m apart. Many of them had interlacing branches. Cone’s east side yard trees had the lowe r limbs cut within 2-2.5 m from the ground on all but six, and many of the limbs at that height were removed from 14 of 20 trees in the west border transect. All of the cedars in the forester’s lane

PAGE 298

282 transect had all their lower limbs intact except for two at the entrance to the facility which were trimmed to improve visibility and safe egress with highway 20 (Figure 6-8 A). An auxiliary grove of 22 trees planted at the end of the Forest Figure 6-8. Hollister, Putnam Co., WP T 019: A) west border transect, B) eastwest transect, C) inside east-west transect

PAGE 299

283 ServiceÂ’s propagation program in 1992 is be hind the small office and the forest tower (Figure 6-8 A). Numbers Waypoint 019 was surveyed 37 times, including its discovery date, 9 March 2000, until 28 September 2003. SweadnerÂ’s Hairstreak was flying on 22 of those visits. Of the 472 total sightings of the bu tterfly in my field study, 208 of them were at WPT 019. On my exploratory visi t I found five butterflies, and record days of 32 and 44 individuals were recorded here. The site was occupied in every seas onal brood in which it was sought and only absent during inter-brood periods, or the overwintering season, November to February. The productivity of this planted grove lay in the full conical shape of so many female host trees, pr oviding ideal microclimate fo r the immature stages of the insect and great mating success on t he leks. A dozen pines rose above the foresterÂ’s lane (east-west) transect, prov iding fallen needles to blend with the thick cedar duff beneath these trees. Though the grounds were sometimes mowed, these cedars were not disturbed and native wildflowers grew between the cedars and the mown right-of-way of hi ghway 20 to the north of this transect. Disposition After a record-breaking survey day of 2 September 2002, when 32 hairstreaks were counted, a ranger who liv ed on-site told me of plans by the Florida Department of Transportation (D OT) to begin work on widening highway 20 to four lanes in May of 2003. I c ontacted DOT supervisors and asked how many of the cedars would be cut and found t hat the whole transect parallel to the highway would be removed, in addition to the first six to eight trees of the

PAGE 300

284 perpendicular transects south of the curr ent pavement, to make way for two new lanes to be added on the south side of t he present roadway. I explained my research project and asked to be kept in formed of scheduling and the actual start date of the project. The final flight of the 2002 season ended soon after 6 October and I began the microclimate experiments three weeks later when I confirm ed the end of the flight. I started the duff search after the cool weather began, expecting to lose my field lab in the spring. I called before each deployment of the data loggers for each run after April 2003. The DOT pers onnel were very cooperative and sent me e-mail notices of meetings with contra ctors. As I continued work through the summer, I contacted the department about the schedule dates every few weeks. They finally settled on October 2003 and pr omised me a start date when they knew. My last full field day of 2003 was 28 September when Linda and I rescued wild orchids from the future right-of-way and only two very worn individuals were flying. I called again to ask for start date confirmation on the morning of 15 October and found that the work had started that day and someone had forgotten to call me. When I arrived, the crew was at lunch, the tractor was silent, and 33 of the 41-cedar east-west transect trees were on the ground (Figure 6-9 A). Before the crew returned from lunc h and began the afternoonÂ’s work, I discovered two leking males in ConeÂ’s east side yard transect trees as yet unscathed by the tractor. These were the last two hairstreaks of the 2003 season. I met the supervisors and t he backhoe operator and photographed the

PAGE 301

285 Figure 6-9. Destruction WP T 019: A) east-west transec t, B) west border, C) inside east-west transect removal of the last eight cedars on forest ry service property (Figure 6-9 B) and the shortening of ConeÂ’s west border transect. Figure 6-9 C was shot a few meters west along the foresterÂ’s lane of the perspective of t he Figure 6-8 C. I

PAGE 302

286 returned on 16 October to document the des truction of the cedars on the eastern transect. On 7 March 2004 I returned to survey 14 remaining cedars of the former eastern transect, 13 standing in the west border row, and 61 smaller auxiliary cedars spread about the property. Out of 88 mostly immature cedars, I found two fresh males in a tree near the former west end of the longer transect in the yard of one of the foresterÂ’s homes. I also discovered that 20 cedars on private property fronted on the nor th side of highway 20 had been taken even though they were not part of the original contract. A year and a half after the cedars were cut, two new lanes of highway less than a half mile long had been paved (Figur e 6-10 A) but were not as yet open for travel. Figure 6-10 B shows six pot ential lekking trees on the left and the auxiliary grove of young cedars behind the office in the right of the panel. If no more are damaged, these remaining trees and the other immature cedars spread about the property could foster the continuat ion of the colony of the Putnam Co. phenotype of M. g. sweadneri . Periodic follow-up surveys at WPT 019 could provide a valuable record of the successful continuation of this colony or of its demise. A valuable factor in the success of th is planted habitat was the exclusion of deciduous trees in the shelterbelt and la ck of degradation of the habitat through natural succession observed in the st udies of urban habitat, WPT 001, and the natural grove at WPT 002. The added alti tude above sea level in this part of interior Florida makes for generally dr ier conditions and provides a possible

PAGE 303

287 hedge against the threats of flooding and inundation perio dically experienced by low-lying Gulf coast habitat. Figure 6-10. New paved highway, A) WPT 019, B) auxiliary grove Specific Threats to Habitat and Conservation Concerns Threats to habitat for SweadnerÂ’s Hairst reak and, therefore, to localized colonies and potentially to whole populati ons extend from several quarters. They almost fall neatly into two broad cat egories, natural succession and abiotic events, or direct anthr opocentric manipulation of the environment or its unintended inadvertent effects. My dete rmination that viable populations of the

PAGE 304

288 butterfly can only be maintained in cedar groves with trees of critical size and conical shape, capable of providing the microclimate essential for pupal development and diapause in the duff layer below them, points to the vulnerability of the habitat. Anthropocentric Threats Changes in land use, such as the del iberate destruction of WPT 003, 004, and 005 for silviculture, are prime examples of direct manipulation. Several old cedars in a row of viable habitat were cut and bulldozed to provide a new driveway for a cement fabric ation plant in Flagler Co. at WPT 008 in March 2002. The hairstreak was not seen there in s ubsequent visits during the study. The near total ruin of the WPT 019 habitat to facilitate the travel needs of an increasing human population in north Flor ida also falls into this category. Of even greater concern is the current styl e of development of large tracts of land for housing and location of retirement communi ties, golf courses, and vacation homes in the butterfly’s range. It is cheaper and more convenient for the developer to strip all the vegetation from a building site to create a blank slate for construction of buildings and carefully designed and manicured landscaping than it would be to build in the established “messy” native environment. Figure 6-11 is an example of this development practice . The deep ruts in the foreground are the result of root-raking the pr operty after the trees and brush were removed. This step helps insure that no unplanned regrow th of natural vegetation will occur in a new homeowner’s front yard after he has moved in. Not only does this type of development destroy natural habitat, but the landscaping that replaces it may

PAGE 305

289 Figure 6-11. Palm Coast, Flagler Co. 1999 be filled with exotic plants less likely to be used by native insects and vertebrates, which further degrades local native species diversity. An example of more inadvertent habita t destruction is the controlled burn set to improve habitat in the Steinhat chee Wildlife Management Area discussed above. A second example is the destruction of WPT 020. This site was posted property of the Cyrus Ridge Hunt Club in Dixie Co. It consisted of a core of 12 perfectly conical trees arrayed around the edge of an old borrow pit that had become a deep pond, and another 12 outlying c edars. I took females from this site to start captive colonies and it was consistently occupied in 10 out 13 site visits. Dismantling of this site begin in January 2002 when large mounds of sand and clay were excavated around the pond and piled about the bases of two of the perfect trees. In the proc ess, two other marked trees were run over by a large tractor. Butterflies still flew in the firs t brood eclosion of 1 February 2002, but the destruction continued, too. Butterflies we re perching in the third brood of the

PAGE 306

290 season. Further destruction as a result of earthmoving equipment was noted in March of 2003, and the last remaining cedars were downed later in the summer of 2003. Natural Ruin Slow succession in which cedars are ov ertaken by deciduous species is more subtle than the dramatic loss caused by human activity, but it is just as devastating for the hairstreak. Waypoint 017 was a partial row of cedars at the edge of deciduous woods by the Gulf Hamm ock Post Office in Levy Co. During the study, the woods grew out to surround the cedars. The advance of Muscadine Grape vi nes over otherwise healthy cedars observed at the St. Augustine Lighthouse also occurred over a lekking tree at WPT 017, and on a tree where oviposition behavior was observed at WPT 007. Neither of the trees harbored the butterfly after the br oad-leafed vines grew out the top of the cedar. In all my field experie nce, I never saw the butterfly in a tree even partially covered by these vines. Other natural disasters such as fire s started by light ening can kill the fire-intolerant Southern Red Cedar. As demonstrated by the heavy rains, flooding, and storm surges of the 2004 hurricane season, the cedars can survive inundation that is ruinous to the butte rfly, and cedar hammocks can survive some salinity in the mean high water zone alon g the Gulf coast where the butterfly does not flourish, as at WPT 908 in the high salt marsh. The most ominous threat to the l ong term survival of the species and indeed, to the prosperity of human kind is in the creeping rise of the planetÂ’s sea level. This phenomenon is well documented in the work of several researchers at

PAGE 307

291 the University of Florida who studied its e ffect displayed in the retreat of coastal forests along the Gulf coast (Williams et al ., 1999). So much of Florida is at, or barely above, current sea-le vel that the good coastal habitat for the butterfly will be seriously reduced during the present cent ury, according to current predictions. If SweadnerÂ’s Hairstreak survives into the 22nd century, it will be on the much higher ground of central interior Florida c ounties in cedars that may not as yet have been planted. Implications for Conservation of Biodiversity This entire discussion has centered on Mitoura gryneus sweadneri and its larval host tree, but this is not the only butterfly that uses Southern Red Cedar. Three other hairstreaks (Figure 6-12) were frequently seen at Dixie Co. sites and at WPT 019 in Putnam Co. The Red-banded Hairstreak, Calycopis cecrops , was often recorded flying with M. g. sweadneri and perching near the tops of some of the same lekking trees. The Gray Hairstreak, Strymon melinus , while seldom seen in the tree tops, was often encountered within 2.7 m of the ground in cedar groves at the approximate height where ovipositing M. g. sweadneri were Figure 6-12. Other hairstreaks: A) Calycopis cecrops , B) Strymon melinus , C) Atlides halesus

PAGE 308

292 most often encountered. The specific attr action of these butterflies to cedar has not been studied, though neither is known to utilize cedar as larval host. The Great Purple Hairstreak, Atlides halesus , was observed seeking mates and attempting copulation in a tall female cedar at WPT 007 on two occasions in different years of the st udy and was flushed from the cedars at WPT 019 at least three times. Florida danaids were seen sheltering from the wind in Dixie Co. cedars and sometimes roosting in the trees. On seve ral occasions, I flushed Swallowtails from cedars on cloudy days. Hugo Kons accompanied me on several excursions to hunt for SweadnerÂ’s Hairstreak and he collected voucher specimens of all Lepidoptera encountered. I kept field notes on the other butterflies I saw all during my surveys. In unpublished work by Hugo Kons and Robert Borth, they listed 75 butterfly species captured along CR 361 near my Dixie Co. sites, including representative butterflies of fi ve families. I found the single species representing the Riodi nidae in Florida, Calephelis virginiensis , on Frog Fruit along the road at WPT 002, increasing t he list to 76 butterfly species. Gopher Tortoises were encountered at th ree of my sites and Otters were observed near WPT 004 in the first year of my study. I have also seen the Crestless Plume Orchid, Pteroglossaspis ecristata (state listed as threatened, proposed endangered), in the dry habitat at WPT 019. Though loss of the cedar habitat essential to Mitoura g. sweadneri might not be the death knell for any of these other species, they are only a fe w of the myriad inhabitants comprising the biodiversity associated wit h Southern Red Cedar in nor thern and central Florida.

PAGE 309

293 CHAPTER 7 DISCUSSION AND CONCLUSIONS I began five years of research on t he conservation biology of SweadnerÂ’s Hairstreak with two broad objectives. O ne was to determine whether the Florida taxon, Mitoura g. sweadneri , was significantly different from the mo re northern subspecies, Mitoura g. gryneus , which occupies a broad range across the eastern United States. If so , this could warrant its taxonomic elevation from subspecies to full species status as indica ted in its original description by F. M. Chermock in 1944. Before the conser vation status of an organism can be evaluated, there needs to be a definitive determination of its taxonomy. A second broad objective was to study the relationship of the Florida butterfly to its habitat for clues to its highly localized distribut ion and reputation of rarity based on its frequent occurrence in low numbers of individuals. Though I pursued these two avenues of investigation concurrently , they are discussed individually. Taxonomic Distinctiveness In my search for convincing evidence of the species level distinctiveness of Mitoura g. gryneus and Mitoura gryneus sweadneri , I compared several aspects of the life history, phenotypic alar c haracters, and characters of the genitalic structures of both sexes. My final test of species distinctiveness was looking for evidence of divergence in characters that could pose possible reproductive isolating mechanisms that would prec lude rejoining the spatially separated populations I compared in hy bridization experiments.

PAGE 310

294 Life History In side-by-side culture of captive co lonies of the two butterflies, no distinctions were found in size or featur es of the eggs, larv ae in any stadium, or pupae. Study of the annuli of the eggs with scanning electron microscope (SEM) micrographs showed an equal range of va riation down to the numbers of petals in the rosettes around the micropyle. Mi crographs of the head capsules of each instar revealed identical size r anges and characters for both taxa. Though the butterflies were thought to be separated by their choice, or obligate use of larval host plants, recent botanical taxonomic investigations have determined that Eastern Red Cedar used by M. g. gryneus and Southern Red Cedar used by the Florida butterfly were the same species, Juniperus virginiana . The southern cedar which grows in Florida is currently considered a variety of the more northern cedar tree. This possible distinction was tested by giving gravid females of each butterfly alternating spri gs of either cedar during oviposition. When the larvae of either butterfly were fed the same plant that their eggs hatched on, both grew to adults with no major observable differences in development, though some larvae would not adapt to a change in host partway through their development. The seeds of th ese cedars are distributed by birds, so trees growing in Florida could have spr outed from seeds carried south in winter migration. I concluded that larval host choice was not a species distinction between the butterflies. The single feature of the lif e histories of these butte rflies that was different was in eclosion strategy. The northern butte rfly is bivoltine from New England to Georgia. The first flight of spring produc es pupae, some of which eclose in a

PAGE 311

295 limited second brood soon after developm ent is complete, and the remaining pupae diapause to eclose in small cohor ts of an extended second brood. The Florida butterfly produces at least three and usually four consecutive broods over an extended flight season of 220 or more days. Each brood ecloses within 12-15 days of pupation in a shotgun strategy that increases the chances of mating for each flight of adults. The pr ogeny of the last flight of both adults overwinter as pupae, to eclose in spring. I think that t he difference in eclosion strategy between the two is directly related to differences of climate and length of growing season between Florida and the more temperate eastern U. S. Phenotype Distinctiveness Distinctive differences in alar charac ters were thought by several authors to be significant only at the subspecies level. Characters that in the literature were attributed to M. g. sweadneri were compared in phenotypes from different counties across its range east to west al ong 29° north latitude. I determined that the ranges of variation in mo st of these characters were too broad to be distinct because more western populations in Flor ida showed character states drifting toward similar states in the northern taxon. The alar characters that did seem to show significant differences were a lar ger black spot on the tornal lobe of the hindwing and longer tails of Mitoura g. sweadneri. The Thecla spot of the nominate subspecies, Mitoura g. gryneus , usually has more orange around a smaller black eye, whereas the Thecla spot of the hindwings of eastern Florida butterflies had no orange, or a small amount of brown co loration, while those of some western Florida populations did hav e orange scales. The straighter post median line of the Fl orida butterfly is usually, but not always, more jagged in M.

PAGE 312

296 g. gryneus . Finally, the dorsal wing surfaces of the nominate subspecies have orange or gold discal areas in first-flight individuals and more solid brown to black in individuals from the ex tended partial second brood. In the Florida taxon, the orange-gold discal character is usually r educed to a dusting of color in the eastern and central Florida phenotypes, while it is sometimes expressed as solid yellow-gold discs in western counties. Th is dorsal distinction is a matter of degree that changes with the season, locati on, and sex of both of the butterflies. Some specimens from Liberty Co. and Jackson Co. of northern Florida appeared to have a blend of alar characters of the Florida taxon and the northern nominate subspecies. This suggests that these are intergrades of the two taxa and further blurs the distinctiv eness of alar characters. My genitalic dissections of a limited num ber of specimens of both butterflies also showed a range of variation in se veral structures not mentioned in the literature. I determined that the few slight differ ences I saw were probably not significant barriers to reproduction. Hybridization I tested this hypothesis in hybridizat ion experiments and found that crosses between the two taxa could produce sexu ally viable offspring which mated to produce F2 adults. A male F1 hybrid mated with an M. g. gryneus female and a female F1 hybrid mated with an M. g. sweadneri male to produce backcrosses. Though these experiments were conducted on a small scale with little replication, the fact that the hybrids were possible, though in relatively small numbers, suggests these taxa are not entirely r eproductively isolated even though my test subjects were from temporally and spatially separated populations. The only

PAGE 313

297 matings that totally failed to produce offs pring were between sibling F2 hybrids, which could indicate incompatible gamet es, or inbreeding depression. I prefer to keep the Florida phenotype as subspecies Mitoura gryneus sweadneri to provide stability, to preserve its i dentifiable distinctive history, and to give attention to differences in the observed phenom enon of its eclosion strategy. Habitat Requirements Through extensive study of cedar habi tats that supported colonies of Mitoura g. sweadneri in seven Florida counties, compared with cedar sites that never harbored the butterfly, I determined the essential habitat requirements of the butterfly. Last instars crawl down from their host trees to pupate in the duff below them. Repeated samples of temper ature data from different seasons showed that the duff layer beneath typical c onical-shaped cedars, in combination with the cover provided by a cedar with lowe r limbs intact just above the ground, provide a microclimate in which the tem peratures are lower in the warm months and higher in winter than the ambient temperature outside the tree. This moderation of temperatur e ranges may provide signals that activate the developed butterfly to eclose when c onditions are right for finding mates and nectar sources, and keep the mature pupae in diapause through the lean winter months when premature eclo sion would mean a short adult life and lost chances for reproductive success. With these criteria, a given stand of cedars can be evaluated as good or inferior habitat for the butterfly . I think that part of the reason M. g. sweadneri usually fly in small numbers is that these criteria are not easily met in many cedar habitats. The butterflies have a reputation fo r being rather sessile creatures that

PAGE 314

298 never stray far from their cedar hosts. Th is could explain their highly localized populations. In years with great reproducti ve success, some potential colonists disperse and may be found in previously unoccupied new cedar habitat (or join existing colonies). This could explain why individuals are sometimes found alone on cedars far from established colony sites. Because colonies typically exist in low numbers in complicated habitat struct ure, the butterfly has evolved lekking behavior to find mates to ensure reproduc tive success. Newly eclosed males meet in specific lek trees in consecut ive broods season after season, and year after year, to await the arri val of receptive females. Conservation Concerns The butterflies are constrained to a narrowly specified niche in cedar habitat. When cedars are damaged by naturally occurring fires, floods, or gradual degradation through natural succession of plant communities, they no longer provide the niche requirements and colo nies fail. Likewise, when good cedar habitat is damaged or destroyed by human activity, once-productive colonies wink out. As more and more natural l andscape is modified by the burgeoning human population of Florida, habitat for this butterfly will be lost. The butterfly is already scarce in counties along the r apidly developing Atlantic coast. As sea-levels rise in the coming century, coastal habitat will be lost and the last refuge of SweadnerÂ’s Hairstr eak will be in interior count ies of higher elevation. Since the butterfly has been document ed from colonies in 27 Florida counties, it is not ever likely to be listed as an endangered species. To confirm the absence of a butterfly that usually o ccurs in low numbers but is, or has been so widely distributed would be a truly Herculean task. The best hedge against the

PAGE 315

299 loss of this species and the biodiversity associated with cedar groves would be to reinstitute the former program of massi ve cedar seedling production by the Florida Division of Forestry to make inexpensive tr ees available to Florida citizens for planting in shelterbelts and landscaping of large landholdings. If each purchaser were given a single-page handout with a brief description of viable habitat for the butterfly, the cedar-plant ing public would bec ome aware of its existence and conscious of this part of our natural heritage. To help ensure the long-term survival of SweadnerÂ’s Hairst reak, managers of FloridaÂ’s publicly-held natural lands and parks need to be educated about the connection between this endemic butterfly and uncut conical-shaped cedars so that maintenance regimes planned to protect the hairstr eak and its niche can be instituted and passed on to succeeding generations.

PAGE 316

300 LIST OF REFERENCES Adams, R. P. 1986. Geographic variation in Juniperus silicicola and J. virginiana of the southeastern United States: mu ltivariate analyses of morphology and terpenoids. Taxon . 35 (1):61-75. 1993. Juniperus. In Flora of North Am erica Editorial Co mmittee; N. R. Morin, convening editor, eds. Flora of North Amer ica North of Mexico Vol. II Pteridophytes and Gymnosperms. New York: Oxford University Press. 475pp., figs. 412-420. Alcock, J. 1983. Territoriality by Hilltopping males of the great purple hairstreak, Atlides halesus (Lepidoptera, Lycaenidae): convergent evolution with a pompilid wasp. Behavioral Ecology and Sociobiology . 13 (1):57-62. 1987. Leks and hilltopping in insects. Journal of Natural History . 21 (2):319328. Alcock, J., and K. M. O’Neill. 1986. Density-dependent mating tactics in the Grey hairstreak, Strymon melinus (Lepidoptera: Lycaenidae). Journal of Zoology . 209 (1):105-113. Allen, T. J. 1997. The Butterflies of West Virginia and Their Caterpillars . Pittsburgh, Pennsylvania: University of Pittsburgh Press. 388pp., 50pp. of pl., ill. Barrick, W. E. 1979. Salt Tolerant plants for Florida landscapes . University of Florida Sea Grant College, Report 28. Ga inesville, Florida: University of Florida. 71pp., ill. Baylis, M., and N. E. Pierce. 1991. The effect of host-plant quality on the survival of larvae and oviposition by adults of an ant-tended lycaenid butterfly, Jalmenus evagoras . Ecological Entomology . 16 (1):1-9. Bird, C. D., G. J. Hilchie, N. G. Kondla, E. M. Pike, and F. A. H. Sperling. 1995. Alberta Butterflies . Edmonton, Alberta, Canad a: Provincial Museum of Alberta. 349pp., 879pl. Boisduval, J. B. A. D., and J. E. LeConte. 1833. Histoire Générale et Iconographie des Lépidoptéres et C henilles de l’Amerique Septentrionale . Vol. 1.Paris: Librairie En cyclopedique De Roret. 228pp., 78pl.

PAGE 317

301 Bradbury, J. W. 1981. The evolution of leks. In R. D. Alexander and D. W. Tinkle, eds. Natural Selection and Social Behavior: Recent Research and Theory . New York: Chiron Pr ess. 532pp. 138-169. 1985. Contrasts between insects and vertebr ates in the evolution of male display, female choice, and lek mati ng. In B. Hölldobler and M. Lindauer, eds. Experimental Behavioral Ecology and Sociobiology: in memoriam Karl von Frisch . Sunderland, Massachusetts: Sinauer Associates. 488pp. 273289. Bridges, C. A. 1994. Catalogue of Family-Gr oup, Genus-Group and SpeciesGroup Names of the Riodinidae and Lycaen idae (Lepidoptera) of the World . Urbana, Illinois: Char les A. Bridges. 1156pp. Brown, J. W. 1982(83). A new species of Mitoura Scudder from southern California (Lepidoptera: Lycaenidae). Journal of Research on the Lepidoptera . 21 (4):245-254. Brown, J. W., and D. K. Faulkner. 1988(89). The Butterflies of Isla de Cedros, Baja California Norte, Mexico. Journal of Research on the Lepidoptera . 27 (3-4):233-256. Calhoun, J. V. 1996. Sweadner’s Haristreak (sic), Florida’s rarest butterfly? News of the Lepidopterists’ Society . 38 (6):220 & 224. 1997. Updated list of the butterflies and skippers of Florida (Lepidoptera: Papilionoidea and Hesperioidea). Holarctic Lepidoptera . 4 (2):39-50. Cassie, B., J. Glassberg, P. A. Ople r, R. K. Robbins, and G. Tudor. 1995. The North American Butterfly Association (NABA) Checklist & English Names of North American Butterflies . Morristown, New Jersey: North American Butterfly Association, Inc. 43pp. Chapman, R. F. 1969. The Insects Structure and Function . New York: American Elsevier Publishing Company, Inc. 819pp., 509figs. Chermock, F. H. 1944. Some new North American Lycaenidae. The Canadian Entomologist . 76 (11):213-216. Clench, H. K. 1961. Family Lycaenidae. In P. R. Ehrlich and A. H. Ehrlich, How to Know the Butterflies . Dubuque, Iowa: Wm. C. Brown Company Publishers. 262pp., 525figs. 1975. Introduction. In W. H. Howe, The Butterflies of North America . Garden City, New York: Doubleday & Company, Inc. 633pp, 97pl., 71figs. 1981. New Callophrys (Lycaenidae) from Nort h and Middle America. Bulletin of the Allyn Museum . 64 :31pp., 48figs.

PAGE 318

302 Comstock, J. A., and C. M. Dammers. 1932. Metamorphoses of five California diurnals (Lepidoptera). Bulletin of the Southern California Academy of Sciences . 31 :part 2:33-45. 1938. Notes on the metamorphosis of Mitoura Spinetorum Hew. (Lepidoptera, Theclinae). Bulletin of the Souther n California Academy of Sciences . 37 :part 1:30-32. Coolidge, K. R. 1924. The life-history of Mit oura loki Skinner (Lepid.: Lycaenidae). Entomological News . 35 :199-204. Correll, D. S., and M. C. Johnston. 1970. Manual of the Vascular Plants of Texas . Renner, Texas: Texas Res earch Foundation. 1881pp. Cottrell, C. B. 1984. Aphytophagy in butterflies: its relationship to myrmecophily. Zoological Journal of the Linnean Society . 80 :1-57. Cramer, P. 1782. De Uitlandsche Kapellen voorko mende in de drie WaereldDeelen Asia, Africa en America . Vol. 4. Amsterdam and Utrecht: S. J. Baalde and Barthelemy Wild. 208 pl., 390 figs. Cushman, J. H., and D. D. Murphy. 1993. Conservation of North American lycaenids – an overview. In T. R. New, ed. Conservation Biology of Lycaenidae (Butterflies) . Occasional Paper of the IUCN Species Survival Commission No. 8. Ox ford: Information Press. 173pp. 37-44. Daniels, J. C. 2003. Butterflies of Florida Field Guide . Cambridge, MN: Adventure Publications. 256pp., ill. Devries, P. J. 1990. Enhancement of symbioses between butterfly caterpillars and ants by vibrational communication. Science . 248 :1104-1106. 1991a. Call production by myrmecophilous riodinid and lycaenid butterfly caterpillars (Lepidoptera): morphologi cal, acoustical, functional, and evolutionary patterns. American Museum Novitates . 3025 :1-23. 1991b. Detecting and recording the calls produced by butterfly caterpillars and ants. Journal of Research on the Lepidoptera . 28 :258-262. DeVries, P. J., and G. O. Poinar. 1997. Ancient butterfly-ant symbiosis: direct evidence from Dominican amber. Proceedings of the Royal Society of London Series BBiological Sciences . 264 (1385):1137-1140. DeVries, P. J., J. A. Thomas, and R. Cocroft. 1993. A comparison of acoustical signals between Maculinea butterfly ca terpillars and their obligate host ant species. Biological Journal of the Linnean Society. 49 :229-238.

PAGE 319

303 dos Passos, C. F. 1964. A synonymic list of the Nearctic Rhopalocera. The LepidopteristsÂ’ Society Memoir No. 1 . 145pp. 1970. A revised synonymic catalogue with taxonomic notes on some Nearctic Lycaenidae. Journal of the LepidopteristsÂ’ Society . 24 (1):26-38. Downey, J. C. 1962a. Host-plant relations as data for butterfly classification. Systematic Zoology . 11 (4):150-159. 1962b. Myrmecophily in Plebejus (Icaricia) icarioides (Lepidoptera: Lycaenidae). Entomological News . 73 (3):57-66. 1966. Sound production in pupae of Lycaenidae. Journal of the LepidopteristsÂ’ Society . 20 :129-155. 1987. Lycaenidae (Papilionoidea). In F. W. Stehr, ed. Immature Insects . Dubuque, Iowa: Kendall/Hunt Publishing Company. 754pp. 443-445. Downey, J. C., and A. C. Allyn. 1978. Sounds produced in pupae of Lycaenidae. Bulletin of the Allyn Museum . 48 :1-14. 1979. Morphology and biology of the imma ture stages of Leptotes Cassius Theonus (Lucas) (Lepid.: Lycaenidae). Bulletin of the Allyn Museum . 55 :27pp., 29figs. 1980. Eggs of Riodinidae. Journal of the Lepi dopteristsÂ’ Society . 34 (2):133145. 1981. Chorionic sculpturing in eggs of Lycaenidae. Part I. Bulletin of the Allyn Museum . 61 :29pp., 67figs. 1984. Chorionic sculpturing in eggs of Lycaenidae. Part II. Bulletin of the Allyn Museum . 84 :44pp., 157figs. Dyar, H. G. 1890. The number of molts of lepidopterous larvae. Psyche . 5 :420422. Edwards, H. 1889. Bibliographical Catalogue of the Described Transformations of North American Lepidoptera . (Bulletin of the United States National Museum, No. 35) Washington, D. C.: Government Printing Office. 147pp. Ehrlich, A. H., and P. R Ehrlich. 1978. Reproductive strategies in the butterflies. I. Mating frequency, plugging and egg number. Journal of the Kansas Entomological Society . 51 (4):666-697. Ehrlich, P. R., and A. H. Ehrlich. 1961. How to Know the Butterflies . Dubuque, Iowa: Wm. C. Brown Company Publishers. 262pp., 525figs.

PAGE 320

304 Ehrlich, P. R., and D. D. Murphy. 1981(82). Butterfly nomenclature: a critique. Journal of Research on the Lepidoptera . 20 (1):1-11. Ehrlich, P. R., and P. H. Raven. 1965. Butterflies and plants: a study in coevolution. Evolution . 18 (4):586-608. Elias, T. S. 1980. The Complete trees of North Am erica: field guide and natural history . New York: Outdoor Life/Nature Book: Van Nostrand Reinhold Co. 949pp. ill. Eliot, J. N. 1973. The higher classification of the Lycaenidae (Lepidoptera): a tentative arrangement. Bulletin of the British Museum (Natural History) Entomology . 28 (6):371-505. Emlen, S. T., and L. W. Oring. 1977. Ecology, sexual selection, and the evolution of mating systems. Science . 197 :215-223. Emmel, T. C. 1993. SweadnerÂ’s Hairstreak, Mitoura gryneus sweadneri (Chermock). In T. R. New, ed. Conservation Biology of Lycaenidae (Butterflies) . Occasional Paper of the IUCN Species Survival Commission No. 8. Oxford: Information Press. 173pp. 124-125. 1997. Butterfly Gardening: Creating a Butterfly Haven in Your Garden . New York: Friedman/Fairfax Publishers. 112pp., ill. Ewel, K. C . 1990. Swamps. In R. L. Myers, and J. J. Ewel, eds. Ecosystems of Florida . Orlando, Florida: University of Central Florida Press. 765pp., ill. Eyre, F. H. 1980. Forest Cover Types of the United States and Canada . Washington, D.C.: Society of American Foresters. 148pp. Fiedler, K. 1991. Systematic, evolutionary, and ecological implications of myrmecophily within the Lycaenidae (Ins ecta: Lepidoptera: Papilionoidea). Bonn: Zoologisches Forschungsinstit ut und Museum Alexander Koenig, Bonner Zoologische Monographien, Nr. 31 :1-210. Fiedler, K., B. Holldobler, and P. Seufert. 1996. Butterflies and ants: the communicate domain. Experientia . 52 :14-24. Field, W. D., and J. F. Gates Clarke. 1972. Introduction to H. M. Tietz. An Index to the Described Life Histories, Early Stages and Hosts of the Macrolepidoptera of the Cont inental United States and Canada. Sarasota, Florida: published by A. C. Allyn for The Allyn Museum of Entomology. 536pp. Forister, M. L. 2004. Oviposition preference and larval performance within a diverging lineage of lycaenid butterflies. Ecological Entomology . 29 :264272.

PAGE 321

305 2005a. Influence of host plant phenology on Mitoura nelsoni (Lepidoptera: Lycaenidae). Annals of the Entomological Society of America . 98 (3):295301. 2005b. Independent inheritance of preference and performance in hybrids between host races of Mitoura butterflies (Lepidoptera: Lycaenidae). Evolution . 59 (5):1149-1155. Foster, S., and J. A. Duke. 1990. A Field Guide to Medicinal Plants: Eastern and Central North America . Boston: Houghton Mifflin Company. 366pp., 48pl., ills. Gerberg, E. J., and R. H. Arnett, Jr. 1989. Florida Butterflies . Baltimore, Maryland: Natural Science Publications, Inc. 90pp., ill. Gilpin, M., and I. Hanski. 1991. eds. Metapopulation Dynamics: Empirical and Theoretical Investigations . London: Academic Press. 336pp. 1997. Metapopulation Biology: Ecology , Genetics, and Evolution . San Diego California: Academic Press. 512pp. Glassberg, J. 1993. Butterflies Through Binoculars: a Field Guide to Butterflies in the Boston-New York-Washington Region . New York: Oxford University Press. 160pp., 40pl. 1999. Butterflies Through Binoculars: t he East: a Field Guide to the Butterflies of Eastern North America . New York: Oxford University Press. 242pp., 71pl. 2001. Butterflies Through Binoculars: t he West: a Field Guide to the Butterflies of Western North America . New York: Oxford University Press. 374pp., 127pl. Glassberg, J., M. C. Minno, and J. V. Calhoun. 2000. Butterflies Through Binoculars: a Field, Finding, and Gar dening Guide to Butterflies in Florida . New York: Oxford University Press. 242pp., 44pl. Gordh, G., and D. H. Headrick. 2001. A Dictionary of Entomology . New York: CABI Publishing. 1032pp. Grossbeck, J. A. 1917. Insects of Florida. IV. Lepi doptera. F. E. Watson (ed.) Bulletin of the American Museum of Natural History . 37 (Art. I):1-147. Hanski, I., and C. D. Thomas. 1994. Metapopulation dynamics and conservation: a spatially explic it model applied to butterflies. Biological Conservation . 68 :167-180.

PAGE 322

306 Harris, L., Jr. 1972. Butterflies of Georgia. Norman, Oklahoma: University of Oklahoma Press. 326pp., 24pls. Harvey, J. D. 1987. Riodinidae (Papilionoidea). In F. W. Stehr, ed. Immature Insects . Dubuque, Iowa: Kendall/Hunt Publishing Company. 754pp. 446447. Hodges, R. W., D. R. Davis, T. Domini ck, D. C. Ferguson, J. G. Franclemont, E. G. Munroe, and J. A. Powell. 1983. Check List of the Lepidoptera of America North of Mexico including Greenland . London: E. W. Classey Limited and Wedge Entomologica l Research Foundation. 284pp. Holland, W. J. 1940. The Butterfly Book . New York: Doubleday, Doran, & Company, Inc. 424pp., 77pl., 198figs. Howe, W. H. 1975. The Butterflies of North America . Garden City, New York: Doubleday & Company, Inc. 633pp., 97pl., 71figs. Jaeger, E. C. 1955. A Source-Book of Biological Names and Terms . Springfield, Illinois: Charles C Thomas Publisher. 317pp. Johnson, K. D. 1976a. Three new Nearctic species of Callophrys (Mitoura) , with diagnostis (sic) of all Nearctic consubgeners (Lepidoptera: Lycaenidae). Bulletin of the Allyn Museum . 38 :30pp., 17figs. 1976b. A new species of Callophrys ( Mitoura ) from Mexico (Lepidoptera: Lycaenidae). Pan-Pacific Entomologist . 52 (1):60-62. 1978. Specificity, geographic distributions , and foodplant diversity in four Callophrys ( Mitoura ) (Lycaenidae). Journal of the Lepidopterists’ Society . 32 (1):3-19. 1980. Revision of the Callophryina of the World with Phylogenetic and Biogeographic Analyses (Lepidoptera: Lycaenidae) . City University of New York. (Ph.D. Dissertation) 902pp. (UMI No. 8112361) 1983. The suitability of Juniperus (Cupressaceae) for larvae of Callophrys hesseli (Rawson and Ziegler) (Lycaenidae). Journal of the Lepidopterists’ Society . 37 (1):78-79. Johnson, K. D., and P. M. Borgo. 1976. Patterned perching behavior in two Callophrys ( Mitoura ) (Lycaenidae). Journal of the Lepidopterists’ Society . 30 (3):169-183. Johnson, K. D., and J. J. Kruse. 1997. Colombian species of tailless “ Cyanophrys ” sensu lato and their sister taxa: priorities for biological study. Review of Colombian Theclinae . 1 (5):1-33. Manizales, Colombia: Museo de Historia Natural de la Universidad de Caldas

PAGE 323

307 Johnson, K. D., and E. L. Quinter. 1982(83). Commentary on Miller and Brown vs. Ehrlich and Murphy et al .: Pluralism in systematics and the worldwide nature of kinship groups. Journal of Research on the Lepidoptera . 21 (4):255-269. Kimbell, C. P. 1965. The Lepidoptera of Florida, an annotated checklist. In Arthropods of Florida and Neighboring Land Areas Vol. 1. Gainesville, Florida: Division of Plant Industry, Florida Department of Agriculture and Consumer Services, 363pp., 26pl. Kitching, R. L. 1987. Aspects of the natural histor y of the lycaenid butterfly Allotinus major in Sulawesi. Journal of Natural History . 21 (3):535-544. Klots, A. B. 1951. A Field Guide to the Butterflies of North America, East of the Great Plains . Boston: Houghton Mifflin Co mpany. 349pp., 40pl., 10figs. 1970. Lepidoptera. In S. L. Tuxen, ed. TaxonomistÂ’s Glossary of Genitalia in Insects . Copenhagen: Munksgaard. 359pp., 248 figs. Lederhouse, R. C. 1982. Territorial defense and lek behavior of the Black Swallowtail butterfly, Papilio polyxenes . Behavioral Ecology and Sociobiology . 10 (2):109-118. Little, E. L. Jr. 1971. Atlas of the United States trees . Vol. 1 Conifers and important hardwoods. Washington: Unit ed States Department of Agriculture Forest Service Miscellaneous Publication No. 1146. 200 figs., 12 ind. Mattoni, R. H. T. 1979. The Scolitantidini II. The worl dÂ’s smallest butterfly? Notes on Turanana , and a new Genus and species from Afghanistan (Lycaenidae). The Journal of Research on the Lepidoptera . 18 (4):256-264. Mayr, E. 1963. Animal Species and Evolution . Cambridge: Belknap Press of Harvard University Press. 797pp. McDunnough, J. 1938. Check List of the Lepidopt era of Canada and the United States of America. Pa rt 1 Macrolepidoptera. Memoirs of the Southern California Academy of Sciences Vol. 1 . 275pp. Miller, L. D., and F. M. Brown. 1981. A Catalogue/Checklist of the Butterflies of America North of Mexico. The LepidopteristsÂ’ Society Memoir No. 2 . 280pp. 1983. Lycaenidae. In R. W. Hodges, D. R. Davis, T. Dominick, D. C. Ferguson, J. G. Franclemont, E. G. Munroe, and J. A. Powell, eds. Check List of the Lepidoptera of America North of Me xico including Greenland. London: E. W. Classey Limited and Wedge Entomological Research Foundation. 284pp.

PAGE 324

308 Minno, M. C. 1994. Order Lepidoptera: Family L ycaenidae. In M. Deyrup and R. Franz, eds. Rare and Endangered Biota of Florida. Vol. IV. Invertebrates. Gainesville, Florida: Universi ty Press of Florida. 798pp. Minno, M. C., and T. C. Emmel. 1993. Butterflies of the Florida Keys . Gainesville, Florida: Scientific Pu blishers, Inc. 1 68pp., 29pl., 53figs. Moerman, D. E. 1998. Native American Ethnobotany . Portland, Oregon: Timber Press. 927pp. Mosher, E. 1916. A classification of the Lepidopt era based on characters of the pupa. Bulletin of the Illinois Stat e Laboratory of Natural History . 323pp., 36pl. Murray, E. 1983. Notae spermatophytae no.2. Kalmia . 13 :3-11. Nation, J. L. 1983. A new method using hexamethyl disilazane for preparation of soft insect tissues for sc anning electron microscopy. Stain Technology . 58 (6):347-351. New, T. R. 1993. Part 1. Introduction. In T. R. New, ed. Conservation Biology of Lycaenidae (Butterflies) . Occasional Paper of the IUCN Species Survival Commission No. 8. Oxford: Information Press. 173pp. New, T. R. 1997. Butterfly Conservation . Auckland: Oxford University Press Australia. 248pp. New, T. R., M. B. Bush, I. W. B. Thornton, and H. K. Sudarman. 1988. The butterfly fauna of the Krakatau Islands after a century of colonization. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences. 322 (1211):445-457. Nice, C. C., and A. M. Shapiro. 2001. Population genetic evidence of restricted gene flow between host races in the butterfly genus Mitoura (Lepidoptera: Lycaenidae). Annals of the Entomological Society of America . 94 (2):257267. Opler, P. A., and G. O. Krizek. 1984. Butterflies East of t he Great Plains. An Illustrated Natural History . Baltimore: The Johns Hopkins University Press. 294pp., 54pl., 362figs. Opler, P. A., and V. Malikul (illustrator) . 1992. A Field Guide to Eastern Butterflies . Boston: Houghton Mifflin Co mpany. 396pp., 48pl., 7figs. Petrides, G. A. 1988. A Field Guide to Eastern Tr ees, Eastern United Stastes and Canada . Boston: Houghton Mifflin Company. 272pp., 58pl.

PAGE 325

309 Pierce, N. E. 1985. Lycaenid butterflies and ants: se lection for nitrogen-fixing and other protein-rich foodplants. American Naturalist . 125 (6):888-895. Pierce, N. E., M. F. Brab y, A. Heath, D. J. Lohman , J. Mathew, D. B. Rand, and M. A. Travassos. 2002. The ecology and evolut ion of ant association in the Lycaenidae (Lepidoptera). Annual Review of Entomology . 47 (1):733771. Pinheiro, C. E. G. 1990(91). Territorial hilltopping beha vior of three swallowtail butterflies (Lepidoptera, Pap ilionidae) in western Brazil. Journal of Research on the Lepidoptera . 29 (1-2):134-142. Powell, J. A. 1975. Family Riodinidae, the Metalm arks. In W. H. Howe, ed. The Butterflies of North America . Garden City, New York: Doubleday & Company, Inc. 633pp., 97pl., 32figs. 259-272. Pyle, R. M. 1981. The Audubon Society Field Guide to North American Butterflies . New York: Alfred A. Knopf, Inc. 924pp., 759pl. Rawson, G. W., and J. B. Ziegler. 1950. A new species of Mitoura Scudder from the Pine Barrens of New Je rsey (Lepidoptera, Lycaenidae). Journal of the New York Entomological Society . 58 (2):69-82. Remington, C. L., and R. W. Pease, Jr. 1955. Studies in foodplant specificity. 1. the suitability of swamp white cedar for Mitoura gryneus (Lycaenidae). The LepidopteristsÂ’ News . 9 (1):4-6. Robbins, R. K. 1982. How many butterfly species? News of the LepidopteristsÂ’ Society . 3 :40-41. 1988. Male foretarsal variation in Lycaenidae and Riodinidae, and the systematic placement of Styx infernalis (Lepidoptera). Proceedings of the Entomological Society of Washington . 90 (3):356-368. 1990. The Mitoura spinetorum complex in New Mexico and the validity of M. millerorum (Lycaenidae: Theclinae). Journal of the Lepi dopteristsÂ’ Society . 44 (2):63-76. Robbins, R. K., and O. Aiello. 1982. Foodplant and oviposition records for Panamanian Lycaeni dae and Riodinidae. Journal of the LepidopteristsÂ’ Society . 36 (2):65-75. Robbins, R. K., and G. B. Small. 1981. Wind dispersal of Panamanian hairstreak butterflies (Lepidoptera: Lycaenidae) and its evolutionary significance. Biotropica . 13 (4):308-315. Rutowski, R. L., J. Al cock, and M. Carey. 1989. Hilltopping in the Pipevine Swallowtail butterfly ( Battus philenor ). Ethology . 82 (3):244-254.

PAGE 326

310 Sargent, C. S. 1902. The Silva of North America: a description of the trees which grow naturally in North America exclusive of Mexico . Vol. XIV Caricaceae – Coniferae. Cambridge, Mass.: The Riv erside Press, and Boston and New York: Houghton, Mifflin and Company. 152pp., 35pl. 1922. Manual of the Trees of North America , 2nd Edition. (1961 reprint) New York: Dover Publications. 910pp., 783figs. Scott, J. A. 1974. Mate-locating behavior of butterflies. The American Midland Naturalist . 91 (1):103-117. 1975. Genus Strymon Hübner. In W. H. Howe, ed. The Butterflies of North America . Garden City, New York: Doubl eday & Company, Inc. 633pp., 97pl., 71figs. 303-306. 1986. The Butterflies of North America . Stanford, California: Stanford University Press. 583pp., 64pl., 71figs. Scudder, S. H. 1869. A systematic revision of some of the American butterflies; with brief notes on those known to occur in Essex County, Mass. First Annual Report of the Tr ustees of the Peabody Academy of Science . 24-83 et errata . 1889a. The Butterflies of the United States and Canada with Special reference to New England . Vol. II Lycaenidae, Papilionidae, Hesperiidae. Cambridge. Published by the author. 1007pp. 1889b. The Butterflies of the United States and Canada with Special reference to New England . Vol. III Appendix, Plates. Cambridge. Published by the author. 271pp., 89pl. Seufert, P., and K. Fiedler. 1996. Life-history diversity and local co-existence of three closely related lycaenid butte rflies (Lepidoptera: Lycaenidae) in Malaysian rainforests. Zoologischer Anzeiger . 234 (4):229-239. Shapiro, A. M. 2002. Species concepts and conser vation law: why we have a problem. News of the Lepidopterists’ Society . 44 (4):124-131. Silba, J. 1984. An international census of the Coniferae I. Phytologia Memoirs 7 :1-79. Singer, M. C., P. R. Ehrlic h, and L. E. Gilbert. 1971. Butterfly feeding on lycopsid. Science . 172 (3990):1341-1342. Small, J. K. 1923. Land of the question mark. Journal of the New York Botanical Garden . 24 (277):1-23.

PAGE 327

311 1933. Manual of the Southeastern Flora . Chapel Hill, North Carolina: The University of North Carolina Press. 554pp., ill. Stoll, C. 1782. Proeve van eene rangschikkinge der Donsvleugelige Insecten, Lepidopteræ . In P. Cramer, De Uitlandsche Kapellen voorkomende in de drie Waereld-Deelen Asia, Africa en America . Vol. 4. Amsterdam and Utrecht: S. J. Baalde and Bart helemy Wild. 208 pl., 390 figs. Tietz, H. M. 1972. An Index to the Described Life Histories, Early Stages and Hosts of the Macrolepidoptera of t he Continental United States and Canada. Sarasota, Florida: published by A. C. Allyn for The Allyn Museum of Entomology. 536pp. Tilden, J. W., and A. C. Smith. 1986. A Field Guide to Western Butterflies . Boston: Houghton Mifflin Co mpany. 370pp., 48pl., 16figs. United States, Forest Service. 1974. Seeds of Woody Plants in the United States / prepared by the Forest Servic e; C. S. Schopmeyer, technical coordinator. Agricultural handbook No . 450. Washington, D. C.: Forest Service, U. S. Department of Agriculture. 883pp., 8pl. Van Haverbeke, D. F., and R. A. Read. 1976. Genetics of eastern redcedar. USDA Forest Service research paper WO-32. Washington: Forest Service, U. S. Department of Agriculture. 17pp. Ward, D. B. 1997. Big Trees: the Florida Register . Orlando, Florida: Florida Native Plant Society. 233pp. Watson, F. D., and J. E. Eckenwalder. 1993. Cupressaceae Bartlett, Redwood or Cypress Family. In Flora of No rth America Editor ial Committee, eds. Flora of North America North of Mexico . Vol. II Pteridophytes and Gymnosperms. New York: Oxford Univ ersity Press. 475pp., figs. 399-401. Weed, C. M. 1926. Butterflies Worth Knowing . Garden City, New York: Doubleday, Page, & Company. 286pp., 48pls. Weiner, M. A. 1991. Earth Medicine, Earth Food: plant remedies, drugs, and natural foods of the North American Indians . New York: Ballantine Books. 230pp., ill. Wilhite, L. P. 1990. Juniperus silicicola (Small) Bailey. In R. M. Burns, and B. H. Honkala (technical coordinators), Silvics of North America . 1 Conifers; 2 Hardwoods. Vol. II. Agricultural handbook No. 654. Washington, D. C.: Forest Service, U. S. Departm ent of Agriculture. 877pp., ill. Wickman, P. O., a nd C. Wiklund. 1983. Territorial defence and its seasonal decline in the Speckled Wood butterfly ( Parage aegeria ). Animal Behaviour . 31 (4):1206-1216.

PAGE 328

312 Willians, K, K. C. Ewel, R. P. Stumpf, F. E. Putz, and T. W. Workman. 1999. Sea-level rise and coastal forest retreat on the west coast of Florida, USA. Ecology . 80 (6): 2045-2063. Wilson, E. O. 1975. Sociobiology: the new synthesis . Cambridge, Belknap Press of Harvard University Press. 697pp., ill. Winter, W. D. 2000. Basic Techniques for Observing and Studying Moths and Butterflies. The LepidopteristsÂ’ Society Memoir No. 5 . 444pp., ill. Wunderlin, R. P., and B. F. Hausen. 2000. Flora of Florida. Vol. I: Pteridophytes and Gymnosperms. Gainesville, Florida: University Press of Florida. 366pp., 76figs. 2003. Guide to the Vascular Plants of Florida second edition. Gainesville, Florida: University Pr ess of Florida. 788pp. Yack, J. E., M. L. Smith, and P. J. Weatherhead. 2001. Caterpillar talk: acoustically mediated territoria lity in larval Lepidoptera. Proceedings of the National Academy of Sciences of the United States of America . 98 (20):11371-11375. Zanoni, T. A. 1978. The American junipers of the section Sabina ( Juniperus , Cupressaceae) a century later. Phytologia . 38 (6):433-454. Ziegler, J. B. 1960. Preliminary contribution to a redefinition of the genera of North American hairstreaks (Lyc aenidae) north of Mexico. Journal of the LepidopteristsÂ’ Society . 14(1):19-23.

PAGE 329

313 BIOGRAPHICAL SKETCH Akers Pence grew up in Arkansas and graduated from Hendrix College, with a major in English literature, in 1970. He taught ninth-grade English for several years in Arkansas. In 1978, he moved to Brooklyn, New York, and started a wallcovering-installation busine ss as a contractor and sole proprietor. Akers was diagnosed with HodgkinÂ’s Lymphoma in 1987 and after recovering from treatment changed prio rities and began to attend meetings of the New York City Butterfly Club and counted butterflies in the annual Xerces Society 4th of July counts. He had a recurrence of lymphoma in 1992 and, after surviving chemotherapy, he decided to pursue a ch ildhood dream of becoming a biologist. He returned to school at Brooklyn College as an adult undergrad, to cover the math and science prerequisites fo r graduate school. In 1996, he was a summer research intern to Dr. Jerry Ro sen at the American Museum of Natural History (AMNH) and spent the field seas on studying native bees in Arizona. He met Dr. T. C. Emmel when he visited the AMNH that summer and determined to come to the Un iversity of Florida to wo rk on the recovery of the endangered Schaus Swallowtail. He has since done field research on (and captive propagation of) t he state endangered Miami Bl ue with Jaret Daniels and tries to spend as much time in the field as possible.