INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI

films the text directly from the original or copy submitted. Thus, some

thesis and dissertation copies are in typewriter face, while others may

be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of thecopy submitted. Broken or indistinct print, colored or poor qualityillustrations and photographs, print bleedthrough, substandard margins,

and improper alignment can adversely affect reproduction.

In the unlikely. event that the author did not send UMI a complete

manuscript and there are missing pages, these will be noted. Also, ifunauthorized copyright material had to be removed, a note will indicate

the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by

sectioning the original, beginning at the upper left-hand comer and

continuing from left to right in equal sections with small overlaps. Each

original is also photographed in one exposure and is included in

reduced form at the back of the book.

Photographs included in the original manuscript have been reproducedxerographically in this copy. Higher quality 6" x 9" black and white

photographic prints are available for any photographs or illustrations

appearing in this copy for an additional charge. Contact UMI directly

to order.

University Microfilms InternationalA Bell & Howell Information Company

300 North Zeeb Road. Ann Arbor. Mr48106·1346 USA313,761·4700 800:521-0600

Order Number 9506224

Speciation, species boundaries, and the population biology ofIndo-west Pacific butterflyfishes (Chaetodontidae)

McMillan, William Owen, Ph.D.

University of Hawaii, 1994

V-M·I300 N. Zeeb Rd.Ann Arbor, MI 48106

SPECIATION, SPECIES BOUNDARIES, AND THE POPULATION BIOLOGY

OF INDO-WEST PACIFIC BUTTERFLYFISHES (CHAETODONTIDAE)

A DISSERTATION SUBMITTED TO THE GRADUATE DNISION OF THEUNIVERSITY OF HAWAIl IN PARTIAL FULFILLMENT OF THE

REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

IN

ZOOLOGY

AUGUST 1994

By

W. Owen McMillan

Dissertation Committee:

Stephen R. Palumbi, ChairpersonE. Alison Kay

Leonard A. FreedChris Simon

Rebecca L. Cann

© Copyright 1994

by W. Owen McMillan

All Rights Reserved

iii

This dissertation is dedicated to Valerie W. McMillan - my wife and travel

companion. Off we go. Cheers .

iv

ACKNOWLEDGMENTS

This work would not have been possible without the help of many

people. I would especially like to acknowledge the support of Valerie

McMillan, without whom this research would not have been nearly as

enjoyable. Members of my dissertation committee, Becky Cann, Lenny Freed,

Alison Kay, Chris Simon, and especially my chairman and mentor, Steve

Palumbi, provided enlightened direction and discussions throughout the

evolution of this work. My lab mates, Bailey Kessing, Andrew Martin, Ed

Metz, Sandra Romano, Paul Armstrong, Frank Cipriano, Gail Grabowski, and

Tom Duda, always kept things fun. Added thanks go to Frank Cipriano, Ed

Metz, Carol Reeb, George Roderick and Sandra Romano for somehow

managing to read through my many spelling and grammatical errors to

provide thoughtful comments on earlier drafts of some of these chapters.

Richard Pyle proved an invaluable resource both as collector and

photographer and as someone who was always willing to talk about fishes,

IWP biogeography, and speciation. Thanks for collecting specimens go to C.

V. Dinar, Jane Culp, John Godwin, Randy Kosaki, Friedhelm Krupp, Jeff

Mahon, Tony Nahacky, and Larry Sharron. Lastly, Lee Weigt helped greatly

with gathering and analyzing the allozyme data. Financial support from the

Research Corporation of the University of Hawaii, Achievement Rewards for

College Scientists, Sigma Xi, The American Museum of Natural History, and

NSF (#DEB-9106870 and #DEB-92-24071) is greatly appreciated.

v

ABSTRACT

This study uses the pattern of nucleotide substitutions within two regions

of the mitochondrial genome to construct a phylog~netic and demographic

framework to explore species formation in allopatric groups of Indo-west Pacific

(IWP) butterflyfishes (Chaetodontidae). Strong parallels in the evolutionary

history of these groups, as inferred from substitutions within a 495 base pair (bp)

section of the mitochondrial cytochrome b gene (cyt b), suggest a link between

their formation and Pleistocene climatic fluctuations, and highlight a recent

period of significant demographic and evolutionary change across the IWP.

Variation within the hypervariable proline tRNA (tRNApro) end of

the mitochondrial control-region within the three Pacific species of one of

these groups reveals a rich history of demographic change marked by the

rapid coalesced of mtDNA variation into three major lineages followed by a

wave of demographic expansion. However, only one of these lineages shows

any species specificity. This lineage is confined to the Hawaiian

archipelago/Johnston Island endemic Chaetodon multicinctus. The

remaining two lineages are geographically widespread and are composed of

similar numbers of C. pelewensis and C. punctatofasciatus. This patterning of

mtDNA variation supports the genetic distinctiveness of the Hawaiian

species but argues for extraordinarily high levels of hybridization between C.

pelewensis and C. punctatofasciatus. Similar conclusions are suggested by the

distribution of allozyme variation among species.

The greater genetic differentiation of C. multicinctus is paralleled by

marked behavioral isolation. By contrast, in controlled pairing experiments,

C. pelewensis and C. punctatofasciatus do not pair assortatively. Field

VI

observations suggest that a broad hybrid zone joins the two species. In

localities across this zone, intermediately colored individuals are common

and pairing is random with respect to color pattern.

Despite being tightly coupled reproductively and evolutionary, C.

pelewensis and C. punctatofasciatus remain uncoupled phenotypically. Color

pattern differences change over a much narrower range than expected given the

complete homogenization of mtDNA and allozyme variation. Selection directly

on color pattern must be maintaining phenotypic integrity of these two color

variants while most other genes move freely across "species" boundaries.

However the nature of this selections remains enigmatic.

vii

TABLE OF CONTENTS

Acknowledgments v

Abstract vi

List of Tables x

List of Figures .. xi

Introduction 1

Chapter 1: Concordant Evolutionary Paterns Among Indo-west PacificButterflyfishes............................................................................................ 9

1.1 Abstract 9

1.2 Introduction 11

1.3 Materials and Methods 15

1.4 Results 20

1.5 Discussion......... 25

Chapter 2: Contrasting Patterns of Phenotypic and MtDNA VariationAmong Recently Differentiated Butterflyfishes (FamilyChaetodontidae).. 48

2.1 Abstract 48

2.2 Introduction 50

2.3 Materials and Methods 54

2.4 Results : 60

2.5 Discussion.. 69

Chapter 3: Random Mating and Species Boundaries in AllopatricCoral Reef Fishes................................................................................... 108

3.1 Abstract 108

3.2 Introduction 110

3.3 Materials and Methods 114

3.4 Results 120

3.5 Discussion 126

viii

Appendix A 163

Appendix B 167

Appendix C 182

References 202

ix

Table1.1

1.2

1.3

1.4

2.1

2.2

2.3

2.4

2.5

3.1

3.2

3.3

3.4

LIST OF TABLES

Number of transitions and transversions in themitochondrial cyt b gene 34

Polymorphic nucleotide positions in the cyt b gene 35

Average within- and between-species levels ofmtDNA variation in species of butterflyfishes 37

Hierarchical structure within cyt b trees 39

Pattern of nucleotide substitution along thetRNAPro end of the control-region butterflyfishes 81

Pattern of variation in C. multicinctus, C.pelewensis and C. punctatofasciatus using a 500 bppiece of the cyt b gene and a 195 bp piece of thetRNAPro end of the control-region 82

Observed and expected number of basesubstitutions under a Poisson and negativebinomial model of substitution in the tRNAProend of the control-region of butterflyfishes 83

Average percent difference within and among thethree control-region lineages in C. multicinctus,C. pelewensis and C. punctatofasciatus 84

Number of individuals of C. multicinctus, C.pelewensis and C. punctatofasciatus with each ofthe three major mtDNA lineages 85

List of museum specimens 139

Number of different pairing experiments 140

Number of hetero- and conspecific pairs 141

Summary of controlled pairing experiments 142

x

Figure.

1.1

1.2

1.3

2.1

2.2

2.3

2.4

2.5

2.6

2.7

2.8

UST OF FIGURES

Geographic distribution of members of the a)"punctatofasciatus" and b) "rhombochaetodon"species-groups ~ 40

Cytochrome b phylogeny of some subgenera withinChaetodon 43

Cytochrome b phylogeny for individuals within a)the "punctatofasciatus" and b)"rhombochaetodon" species-groups 45

Plate of Chaetodon punctatofasciatus, C. pelewensisand C. multicinctus 87

PCR primers used to amplify the control-region ofbutterflyfishes 89

Neighbor-joining tree for 38 individuals of C.punctatofasciatus, C. pelewensis and C.multicinctus using a) a 195 bp portion of thecontrol-region and b) a 500 bp portion of the cyt Bgene 91

Rate of change in the tRNAPro end of the control­region relative to the rate of change within the cytb gene in butterflyfishes 94

Average number of changes along a 195 bp segmentof the tRNAPro end of the control-region 96

Neighboring-joining cartoon of the three majorcontrol-region lineages within C.punctatofasciatus, C. ~lewensis and C.multicinctus 98

Pairwise differences among individuals in clades Aand B 103

Pairwise differences among individuals in each ofthe three major control-region lineages 105

xi

2.9 The proportion of the three major control-regionlineages across the Pacific 106

2.10 Hetero-specific and conspecific branch lengthcomparison 107

3.1 The three hybrid phenotypic classes 143

3.2 Color plates of the fishes collected in the SolomonIslands and Papua New Guinea 145

3.3 The geographic distribution of the five phenotypicclasses 156

3.4 Results of conspecific pairing experiments 158

3.5 Proportion of the five phenotype classes that werepaired and unpaired in the Solomon Islands 159

3.6 The observed and expected distribution pairs in theSolomon Islands 160

3.7 Pairing of "pure" C. punctatofasciatus typeindividuals in Papua New Guinea 161

3.8 Change in the proportion of "pure" C. pelewensisphenotypes across the Pacific relative to thechange in the proportion of mtDNA variation 162

xii

INTRODUCTION

Background

Understanding the process by which new species arise remains a

central question in biology. A key to untangling the origins of biodiversity

lies in gaining a solid understanding of the evolutionary history of groups of

closely related taxa. Only when viewed within a strong historical context can

factors associated with speciation become clear.

In this respect, there is a growing consensus that genetic data,

particularly DNA sequence data, may provide some of the keenest insights

into the history and mechanisms of speciation (Avise et. al., 1989). DNA

variation can be decomposed into neutral and non-neutral sites and

probabilistic models of nucleotide substitution can be used to determine the

phylogenetic relationships among organisms (Hillis and Moritz, 1990). This

approach can be applied at several different levels. Above the species level,

differences and similarities along a portion of DNA can be used to reconstruct

the pattern of branching events that gave rise to extant taxa. A similar

approach can be applied below the species-level. In this case, genetic changes

can be used to reconstruct the history of individuals that make up a species or

group of closely related species (Avise et aI., 1987). Examination of the

patterning of molecular variation at both levels, particularly when focusing

on newly formed groups, can yield rich insights into the history of species

formation.

Molecular analysis of genetic variation focusing at these two levels has

already greatly contributed to our current understanding of the process of

speciation (see Otte and Endler, 1989). The historical perspective gained from

these studies, coupled with an understanding of behavior, ecology, and

1

biogeography, has helped solidify many or our current paradigms of

speciation. However, most of the paradigms have been generated from

research on terrestrial organisms possessing restricted dispersal capabilities

(OUe and Endler, 1989). Relatively little critical focus has been directed at

speciation in organisms with high dispersal potential. This is especially true

in the marine realm, where taxa often possess vast ranges, huge population

sizes and extended dispersal capabilities due to long-lived planktonic larval

stages (Palumbi, 1992; Knowlton and Jackson, 1993). Classical allopatric

models predict that organisms with these types of population dynamics

should to be buffered from speciation, yet recent paleontological and genetic

research has exposed the speed at which evolutionary change can occur in

these groups (Kohn, 1990; Palumbi and Metz, 1991; Allmon et al., 1993;

Jackson et al., 1993; Kay, 1994). As a result, it is important to begin to augment

studies of speciation in low dispersal species with similar studies on high

dispersal animals (Palumbi, 1992).

Examination of speciation in the Indo-west Pacific Butterflyfishes

The above rational provides the intellectual foundation upon which

this dissertation research is built. By using the pattern of DNA sequence

variation within regions of mitochondrial DNA (mtDNA), I construct an

evolutionary and demographic framework within which to explore

speciation in Indo-west Pacific (IWP) marine taxa. I have taken a hierarchical

approach to this subject, beginning first with phylogenetic analysis above the

species level. I use the conclusions drawn from this initial phylogenetic work

to frame further finer-scale genetic analysis of newly formed species.

The IWP is an extraordinary area in which to blend information on

biogeography, ecology, and behavior with analysis of genetic variation to

2

explore the process of marine speciation. This region, spanning from the East

coast of Africa across the Indian ocean and into the Pacific to the archipelagoes

of Oceania, is the most species-rich of the marine biogeographic provinces.

Despite long standing interest in this diversity, there is little consensus as to

the evolutionary mechanisms that have given rise to the extraordinary

number of species across this region (Connell, 1978; Rosen, 1988). At the heart

of this problem is the difficulty in understanding how populations with long­

lived planktonic larval stages phases can become separated long enough to

diverge into new species. In this respect, the challenges facing biogeographers

working in the IWP can be generalized to all biologist attempting to

understand diversification in taxa, such as wind pollinated plants, migratory

birds, or even ballooning spiders, that have the potential for the wide

dispersal of gametes or offspring.

I have focused on speciation in IWP butterflyfishes (Chaetodontidae).

This family of vividly colored fishes is common on coral reefs around the

world and has been the subject of a wealth of ecological, systematic,

biogeographic, and phylogenetic research (Reese, 1975; 1991; Burgess, 1978;

Anderson et al. 1981; Hourigan, 1987;Tricus, 1987; Blum, 1988, 1989; Leis,

1989). Thus, there exist a solid biological background upon which to frame

genetic results. Larvae of these fishes spend between 40 and 60 days feeding in

the plankton before settling and metamorphosing (Hourigan and Reese, 1987;

Leis, 1989). The ecology of juveniles and adults is intricately tied to coral

reefs, with many butterflyfishes feeding exclusively on corals or coral

associate invertebrates.

My research primarily targets two "species-groups" of the more than 90

IWP butterflyfishes. Both groups are composed of morphologically similar

3

species that differ primarily by color pattern. Each partitions the IWP among

four largely allopatric species. Similar allopatric distributions are common

among many apparently closely-related IWP marine taxa.

Whether or not the generalized biogeographic pattern of closely

related, yet allopatric, species reflects chance or has some underlying

deterministic explanation is the subject of the first part of my dissertation

research. Chapter 1 begins with an overview of the challenges facing

researchers attempting to understand the origins of high IWP diversity and

suggests that genetic analysis of newly formed groups may provide additional

insights into evolution in this region. To support this claim, I use sequence

variation within a 500 base pair (bp) portion of the mitochondrial cytochrome

B gene (cyt B) to reconstruct the evolutionary history of these two

butterflyfish groups. The two groups show remarkable parallels in the

pattern and the timing of differentiation. Because it is unlikely that such a

strong similarity would arise by chance in independently evolving lineages,

this concordance argues that speciation in these groups was influenced by

common environmental factors. In this case, the very low levels of genetic

differences suggest a link between differentiation and Pleistocene sea level

fluctuations.

Clearly, much additional work on the tempo of evolution is necessary

to gauge the effects of Pleistocene changes on biodiversity in this region.

However, the congruent pattern of rapid and recent diversification in these

butterflyfish groups joins a growing body of paleontological and genetic work

highlighting a recent period of significant demographic and evolutionary

change in this region (Crarne, 1987; Pauley, 1990; Kahn, 1990; Palumbi and

Metz, 1991; Kay, 1994). These results challenge earlier notions linking the

4

extraordinary diversity of this region to long-term stable environmental and

physical conditions (Ekman, 1953; Briggs, 1974). Pleistocene climatic changes

are thought to have profoundly affected patterns of biodiversity in other

marine faunal provinces (Allmon et al., 1993; Jackson et al., 1993). Future

ecological and biogeographic work within the IWP must be tempered with

the understanding that the very recent history of this region has, likewise,

been extremely turbulent.

Chapter 2 details the finer scale distribution of mtDNA variation

within the three Pacific species of one of these buttertlyfish species-groups.

The previous chapter ended with the suggestion that the recent history of the

Pacific has been notably dynamic. The goal of further research was to provide

a high resolution demographic framework in which to explore speciation in

this region in much greater detail. Indeed, the genealogical pattern of

mtDNA variation within Chaetodon multicinctus, C. punctatofasciatus, and

C. pelewensis traces a rich history of demographic change. MtDNA variation

coalesced into three distinct lineages, or groups of individuals that share a

common ancestor, early in the history of this group. This period of loss of

mtDNA variation was followed by a period of population expansion in which

mtDNA variation within each of the three lineages was buffered from

extinction.

Yet, only one of the three lineages shows any species-specificity with

the other two composed of individuals from different species. For this group,

the failure of the mlDNA gene tree to reflect species boundaries argues for

hybridization and introgression of mtDNA. For the Hawaiian endemic, C.

multicinctus, levels of hybridization suggested by the mtDNA data are low.

By contrast, between the south Pacific, C. pelewensis and the western Pacific,

5

C. punctatofasciatus, the mtDNA genealogy suggests extraordinary high levels

of hybridization associated with the complete homogenization of mtDNA

variation.

The conclusions about the evolutionary and genetic cohesion of these

species based on patterning of variation in the cytoplasmically inherited

mitochondrial genome extend to nuclear encoded loci. The distribution of

allozyme variation among these butterflyfishes mirrors the pattern of .

mtDNA variation. Chaetodon multicinctus is again genetically the most

distinct. Populations of C. pelewensis and C. punctatofasciatus separated by

7500 kilometers show no significant differences in the distribution of

allozyme variation.

My dissertation ends by more thoroughly dissecting the nature of

species boundaries and species-defining color pattern differences in this

group. Controlled pairing experiments indicate that the degree of genetic

differentiation parallels the degree of behavioral, and presumably

reproductive, isolation among species. In these experiments C. multicinctus

shows both a higher degree of assortative pairing and a significant reduction

of intra-sexual aggression relative to conspecific trials. In contrast, pairing

between C. pelewensis and C. punctatofasciatus appears to be random and

behavioral interactions among males of the two species are identical to the

interaction among conspecific males. There is no compelling evidence to

suggest that these two species have achieved the evolutionary cohesion to

merit their species-level distinction.

Observations in zones of sympatry between these two species reaffirm

this conclusion. A broad, approximately 3000 kilometer, hybrid zone joins

the range of these two species. In this zone, a large proportion of the

6

population is composed of individuals with color patterns intermediate

between the two parental forms. Moreover, in Papua New Guinea and in the

Solomon Islands pairing among pure parental and hybrid phenotypes is

random with respect to color pattern.

Despite being tightly coupled reproductively and genetically, C.

pelewensis and C. punctatofasciatus, none-the-less, remain uncoupled

phenotypically. Color pattern changes much more abruptly than expected

given the random distribution mtDNA and allozyme variation between the

two species. The obvious implication of the strong geographic structure to

phenotypic variation in the absence mtDNA and allozyme differentiation is

that selection is preventing the erosion of color pattern in the presence of

potentially homogenizing levels of gene flow.

However, the nature of selection on color pattern in these fishes

remains enigmatic. The evolutionary and ecological significance of the vivid,

non-cryptic coloration in butterflyfishes, as in most reef fishes, is poorly

understood (Lorenz, 1967; Peterman, 1971; Brockman, 1973; Potts, 1973;

Ehrlich et al., 1977; Allen, 1980; Kelly and Hourigan, 1983; Nuedecker, 1989).

The higher degree of behavioral and genetic isolation in C. multicinctus

suggests that divergence in color pattern can be paralleled by the evolution of

assortative pairing either through species recognition or mate selection. Yet,

the random pairing among C. punctatofasciatus and C. pelewensis argues that

color pattern differences are not providing similar cues in these two species.

Though neither the field observations nor controlled pairing experiments

reveal all the forces underlying pair-formation in these species, they clearly

suggest that mate choice or species recognition based on color pattern alone

are not playing a strong role. It therefore seems unlikely that the original

7

divergence of color pattern in this group was coupled with strong inter-sexual

selection or pressure for species recognition.

It is possible that color pattern in these butterflyfishes evolved under

intra-sexual social selection (sensu, West-Eberhard, 1983) or as an aposomatic

warning. Whatever the reason, selection preventing the breakdown of color

pattern between these two species must be strong to explain the discrepancies

between the distribution of genetic and phenotypic variation. Future

ecological work should help to distinguish between these alternatives.

8

CHAPTER 1

Concordant Evolutionary Patterns Among Indo-west Pacific Butterflyfishes

1.1. ABSTRACT

The tropical Indo-West Pacific (IWP) is the most diverse of the marine

biogeographic provinces. Despite considerable interest in the mechanisms

generating this extraordinary diversity, the lack of a good fossil record and little

information on the phylogenetic relationships among closely related species

have made it difficult to distinguish among competing scenarios for the origins

of new species in this region. A molecular phylogenetic approach to

biogeography may help illuminate the evolutionary process in the IWP by

providing a more detailed picture of the evolutionary relationships among

closely related species, induding an approximate timing of divergence, than

provided by earlier studies. In this study, we report strong parallels in the

evolutionary history of two monophyletic species groups of IWP butterflyfishes

(Family Chaetodontidae) as inferred from a 495 base pair (bp) section of the

mitochondrial cytochrome b gene (cyt b). An approximately 2.0% genetic break

within both groups clearly partitions individuals between the Indian Ocean and

Pacific Ocean species. The Pacific Ocean individuals of both complexes fall

within well-defined monophyletic clades. Genetic differences within the Pacific

clade are low, on average less than 1.0%, and fail to cluster by species boundaries.

Individuals from different species, separated by thousands of kilometers, often

9

possess identical cyt b sequences, whereas conspecifics from the same reefs can

show up to 1.5% difference. Failure of the mtDNA gene tree and species tree to

match could be the result of recent hybridization or may simply reflect shared

ancestral variation.

The striking concordance in phylogenetic patterns of two independent

species groups suggests that genetic differentiation was influenced by common

environmental factors. The very low levels of within- and between-species

genetic differences imply a recent divergence time and suggest a link between the

differentiation of these groups and Pleistocene climatic fluctuations. These

results paint a turbulent picture of the recent evolutionary history of the IWP.

Keywords: Indo-West Pacific, marine fishes, historical biogeography

Chaetodontidae, cytochrome b, mtDNA, PCR, speciation.

10

1.2. INTRODUCTION

The explorer-naturalists of the 19th century were awed by the

extraordinary richness of the tropical Indo-West Pacific (IWP) marine

environment. This region, extending from East Africa across the Indian ocean to

the archipelagoes of the central Pacific, is home to over 500 species of reef

building corals, 4000 species of gastropods, and 4000 species of fishes (Abbott, 1982;

Springer, 1982; Veron, 1986). The diversity of coral reef communities across this

region is rivaled only by the diversity of tropical rain forests (Connell, 1978).

Despite over a century of description and maps of the distribution of

species, the evolutionary mechanisms underlying the origins of high IWP

diversity remain unclear (Rosen, 1988). The long-lived planktonic larval stage of

many IWP marine organisms injects a potential for gene flow and dispersal that

makes it difficult to conceptualize the process by which marine populations

become isolated and ultimately diverge into new species (Valentine and

Jablonski, 1983). Indeed, the two patterns that most characterize IWP

biogeography highlight the difficulties facing researchers attempting to

understand species formation in this region. On one hand, there is a striking

attenuation of species diversity as one moves from the shallow seas surrounding

Indonesia, New Guinea, and the Philippines outward across the Pacific (and to a

lesser extent the Indian ocean), suggesting that long-lived planktonic larval

stages cannot completely overcome strong barriers to dispersal (Verrneij, 1987).

On the other hand, many IWP species' are distributed continuously from East

Africa to Polynesia and Hawaii, over a quarter of the earth's circumference,

underscoring the capacity of larval dispersal to achieve vast geographic ranges

(Ekman, 1953). Overlaid on these broad biogeographic patterns is the observation

that for angelfishes (Pisces: Pomacanthidae), damselfishes (Pisces:

11

Pomacentridae) and cone shells (Mollusca: Gastropoda), the geographic range of a

species is not clearly correlated with the length of larval life (Thresher and

Brothers, 1985; Perron and Kohn, 1985; Thresher et al., 1989; however see Reaka,

1980).

Attempts to understand the origins of IWP biodive~sity are further

hampered by the scarcity of information on the evolutionary history of closely

related species in this region (Rosen, 1988; however see Kohn, 1990; Kay, 1994).

The fossil record is at best patchy, and for most taxa is non-existent. Those

groups that do fossilize well often lack good taxonomic characters making

phylogenetic inference difficult (Rosen, 1988). A molecular-phylogenetic

approach to biogeography should help illuminate evolutionary processes in this

region by providing homologous characters that can be used in reconstructing

phylogenetic relationships among closely related species and, given assumptions

about the rate of change of these characters over time, provide a rough

approximation of the timing of speciation.

Here we examine genetic distances and phylogenetic relationships within

species groups of IWP butterflyfishes (family Chaetodontidae). Butterflyfishes

are common and conspicuous members of coral reefs around the world (Burgess,

1978). After a planktonic larval stage of 40-60 days, larvae settle on coral reefs

where they remain for the duration of their lives (Hourigan and Reese, 1987;

Leis, 1989). Their adult ecology is intricately tied to coral reef ecosystems, with

many species feeding exclusively on coral tissue or coral-associated invertebrates.

They have a long history of association with coral reefs stretching at least as far

back as the Eocene (Choat and Bellwood; 1991). Presently the family consists of

120 species, over 90% of which are restricted to the IWP (Blum, 1989).

12

We have specifically targeted two, the "punctatofasciatus" and

"rhombochaetodon", species groups within the subgenus Exornator. Although

Blum's (1988) morphological analysis failed to resolve the relationships among

22 species that comprise this subgenus, the monophyly of each group was

supported in osteological and color pattern features. These species groups have a

remarkably similar biogeographic pattern. Each is composed of four species,

differing primarily by color pattern, that carve the IWP into large non­

overlapping allopatric ranges (Burgess, 1978; Allen, 1980; Blum, 1989) (figures

1.1a and 1.1b). This biogeographic pattern, in which related species or species

pairs partition the IWP, is evident in many IWP taxa (see Springer, 1982;

Donaldson, 1986; McManus, 1985; Blum, 1989). For example, Blum (1989)

identified 31 monophyletic species groups of IWP Butterflyfishes, of which 18

were allopatric or marginally sympatric,

In terrestrial and freshwater habitats, concordance in biogeographic

patterns among independent taxonomic groups is thought to reflect the action of

past vicariant events (Croizat et al., 1974; Platnick and Nelson, 1978; Cracraft and

Prum, 1988). However, within marine environments, the broad dispersal

capabilities of many organisms can obscure areas of endemism and make past

biogeographic barriers less obvious. As a result, the relationship between

biogeographic patterns and historical processes are much less clear (Reid, 1990).

For example, although species in both the "rhombochaetodon" and

"punctatofasciatus" species groups have a similar patterns of large non­

overlapping geographic ranges, only the species-break at the Indonesian

archipelago is concordant between the two groups.

The two species groups partition the Pacific and Indian Oceans differently.

The "punctatofasciatus" group divides the Pacific among three species: a

13

Hawaiian/Johnson Island endemic, a western-Pacific species whose range

extends across Micronesia as far east as the Marshall Islands, and a south-Pacific

species whose range extends from the Solomon Islands to the Tuamotu and

Marquesas island groups of the eastern South Pacific. In contrast, the

"rhombochaetodon" complex partitions the tropical Pacific among two species.

Again there is a western-Pacific representative; however, the range of this species

does not extend beyond the Philippines. In this group, the range of the "South

Pacific" representative extends over much of the tropical Pacific north of the

equator (see figures l.la and 1.1b).

It is unclear if the differences in biogeographic patterns reflect underlying

differences in the history of speciation or indicate differences in the history of

dispersal following speciation in these groups. This study uses genetic

differences in a large section of the mitochondrial cytochrome b (cyt b) gene to

answer two questions about the biogeographic and phylogenetic patterns between

these groups. First, did both species groups arise contemporaneously or does the

pattern of divergence among species in each group suggest that they arose over

different time scales? Secondly, if they did arise contemporaneously, did they

form in parallel or are the phylogenetic patterns random with respect to

biogeographic patterns?

14

1.3. MATERIALS AND METHODS

To gauge both within- and between-species levels of genetic differences,

we determined the DNA sequence of a 495 base pair (bp) portion of the

mitochondrial cyt b gene from 6-15 individuals of each of the four species of the

"rhombochaetodon" and "punctatofasciatus" species groups. When possible, we

chose individuals to reflect a broad sample of a species across its geographic

range. Species and collection localities were as follows: for the

"rhombochaetodon" species group, C. paucifasciatus (Eilat, Israel, n=7), C.

madagascariensis (Mauritius Islands, n=7), C. xanthurus (Bali, Indonesia, n=3;

Philippines, n=4) and C. mertensii (Guam, n=4: American Samoa, n=4; Tahiti,

n=I), and for the "punctatofasciatus" species group, C. muIticinctus (Hawaii,

n=15); C. punctatofasciatus (Bali, Indonesia, n=7; Philippines, n=3; Guam, n=6), C.

pelewensis (Moorea, Tahiti, n=9; Paga Pago, American Samoa, n=2) and C.

guttatissimus (Mauritius Islands, n=6). In addition, this cyt b region was

sequenced for a single C. auriga, C. miliaris, C. tinkeri, C. burgessi. C. declivis, and

C. flavocoronatus, and four C. argentatus.

All fishes were collected between 1990 and 1992. Individuals were either

frozen at _700

C or preserved in 95% ethanol. With the exception of the

Philippine samples, all collections were made by WOM or by a biologist working

in the area. Individuals from the Philippines were acquired through the tropical

fish trade directly from a distributor.

Genomic preparations, peR amplification, and sequencing

For each fish, approximately 0.5 grams of gill or muscle tissue was

homogenized in 2 ml of cold grinding buffer (O.2M NaCl/0.05M EDTA, pH 8.0).

Tissue stored in 95% ethanol was soaked in 5 ml of grinding buffer on ice for

approximately 1 hour before grinding. Approximately 0.5 ml of this homogenate

15

was transferred into a 1.5 ml microcentrifuge tube. Sodium dodecyl sulfate (SDS)

and proteinase K (Sigma Chemicals) were added to the homogenate to a final

concentration of 1%(v iv) and 20 ug/rnl, respectively. Following an overnight

incubation at 50° C, this solution was extracted twice with equal volumes of

buffered phenol, once with an equal volume of phenol/chloroform/isoamyl

alcohol (25:24:1) solution, and once with an equal volume of

chloroform/isoamyl alcohol (24:1). DNA was recovered by cold-ethanol

precipitation (Sambrook et al., 1989). The resulting pellet was dissolved in 100­

200 III of IX TE, and stored at _20° C until further use.

Approximately 100 ng of total cellular DNA was symmetrically amplified

during 40 cycles in a Perkin Elmer Cetus DNA Thermal Cycler in a 100111 reaction

volume with 1.5 units of Taq DNA polymerase (Perkin Elmer Cetus). DNA

primers used to initiate amplification and their position relative to the human

mtDNA sequence were (5'-ATTCTAACTGGACTATTCCTTGCC-3') (14881) and

(5'-ATTATCTGGGTCTCCGAA(C/T)AGGTT-3') (15490) (Anderson et al., 1981).

The standard thermal cycle profile was: 1) 45 seconds at 94° C, 2) 45 seconds at 55°

C and 3) 45 seconds at 72° C. Following amplification, the double stranded

product was collected by ethanol precipitation, resuspended in 10111 sterile water,

and electrophoresed through a 1% agarose/0.5X TAE gel. The resulting 609 bp

band was cut from the 1.0% agarose/O.5X TAE gel, Gene Cleaned(© biolab 101),

and resuspended in 7-30 III TE. Seven microliters of this product was used as

template for double-stranded sequencing (Palumbi et al., 1991). Both PCR

primers and an internal primer, (5'-GTAGTACTTCTCCT(C/T)CTAGTTAT-3'),

were used to initiate sequencing reactions. These reactions were carried out

using reagents and DNA polymerase enzyme provided in the Sequenase® 2.0

DNA sequencing kit (USB). Sequencing reactions were electrophoresed through

16

6-8% polyacrylamide/7M urea gels with a buffer density gradient, dried under

vacuum at 80° C, and exposed to autoradiographic film for 24-72 hours. Refer to

Palumbi et al. (1991) for the finer details of template preparation, sequencing, and

electrophoresis.

In general, only a single strand was sequenced per individual. However,

in the process of filling in sequence gaps, different PCR reactions were used.

Although this strategy does not eliminate sequencing artifacts produced by errors

in replication during peR amplification, it provided some internal check on the

DNA sequences obtained.

Sequence analysis

Sequences were aligned by eye to the published human cyt b sequence.

Phylogenetic analysis was performed with both parsimony and neighbor-joining

methods using the phylogenetic package PAUP and MEGA (3.03, Swofford, 1993;

1.01, Kumar et. al, 1993). The large number of taxa made parsimony analysis of

all possible topologies impossible. As a result, a subset of possible trees was

evaluated using the heuristic search setting on PAUP (Swofford, 1993). The tree

bisection and reconnection (TBR) branch swapping algorithm was chosen to

search for optimal trees within this framework. For parsimony analyses,

transversions were weighted 10 times more than transitions. This weighting

scheme was consistent with previous conclusions about the pattern of molecular

evolution in the cyt b gene (Irwin et al., 1991). The pairwise genetic distance

matrix used in neighbor-joining analyses were generated using a Kimura two­

parameter model (Kimura, 1980). Trees constructed from distance matrices

estimated by maximum-likelihood, Tamura's distance method and Tamura and

Nei's distance method were identical (Felsenstein, 1993; Tamura, 1992; Tamura

and Nei, 1993).

17

Phylogenetic trees were rooted using aligned sequences from several

additional IWP chaetodontids from the subgenus Exornator~ miliaris), as well

as subgenera Roaops~ tinkeri, C. burgessi, C. declivis, C. flavocoronatus) and

Rabdophorus~ auriga). Of these three subgenera, Blum's (1988) cladistic

analysis of Chaetodon placed Rabdophorus at the most basal position. As a

result, the cyt b sequence for C. auriga was used to root phylogenetic trees

examining the relationships among subgenera. For these trees, a single cyt b

sequence was chosen as the representative for each species. For those species in

which more than one individual was examined, the most common cyt b

sequence was chosen to represent that species. This convention did not alter our

interpretation of the phylogenetic patterns but simplified graphic representation.

The relative strengths of major nodes within phylogenetic trees were

examined by bootstrap re-sampling the original data matrix 500 times

(Felsenstein,1985). Bootstrapped data sets were examined by distance methods

using the phylogenetic packages MEGA, and where possible by parsimony

analysis using PAUP (Swofford, 1993; Kumar et al., 1993).

We used the randomization procedure described by Faith (1991) and log

likelihood difference tests to evaluate alternative phylogenetic hypotheses

(Kishino and Hasegawa, 1989; PHYLIP 3.5, Felsenstein, 1993). The log-likelihood

differences test examines the mean and variance of log-likelihood differences

between alternative trees across all sites. If the observed difference in log­

likelihoods was greater than 1.96 standard deviations then two trees were

considered significantly different. In Faith's (1991) procedure, the observed

18

differences in length between alternative phylogenies were compared to

differences obtained by randomly permuting the original data matrix 500 times

(program courtesy of G. Roderick, University of Hawaii, Honolulu, HI). In these

comparisons, the sequence of the outgroup taxa was not randomized.

19

1.4. RESULTS

Between-group variation

Over 90% of the 126 variable sites among the 15 Chaetodon species

examined occurred at the third position of a codon (appendix A). Corrected

genetic differences between species ranged from below 1.0% among members of

the "punctatofasciatus", "rhombochaetodon" and "tinkeri" species groups to

slightly greater than 17% for differences between C. auriga (subgenus

Rabdophorus) and members of both subgenera Roaops and Exornator (table 1.1).

Nearly all of the DNA substitutions occurring between species were transitions at

either 2- or 4-fold degenerate sites (appendix A). Transition/ transversion ratios

ranged from greater than 38:1 to less than 4:1 (table 1.1). The smaller ratios

occurred between the most divergent individuals and can undoubtedly be

attributed to multiple hits at the same nucleotide position (Desalle et al., 1987).

The actual transition/transversion ratio for the cyt b gene of these fishes probably

lies closer to the higher value, and similar to the 10:1 to 20:1 ratio reported in

mammals (Irwin et al., 1991). A total of only 7 replacement substitutions were

observed among the sample of Chaetodon species depicted in figure 1.2.

Phylogenetic analysis of these species clearly supports the monophyly of

both the "punctatofasciatus" and "rhombochaetodon" species groups (figure 1.2).

Branches leading to both complexes were well supported in bootstrap

replications and were defined by at least 11 substitutions using a 10:1 transition to

transversion weighting. This analysis, however, did not support the inclusion of

C. argentatus within the "rhombochaetodon" species group as previously

suggested by Blum (1989). Instead, this species fell between members of the

subgenus Roaops and the members of the "rhombochaetodon" species group.

Parsimony analysis grouped C. argentatus with members of the subgenus Roaops

20

(figure 1.2). In contrast, the neighbor-joining distance tree grouped C. argentatus

with members of the "rhombochaetodon" complex. Neither of these groupings

well supported by bootstrap analysis. This phylogenetic pattern suggests that the

subgenus Exornator, which includes both "rhombochaetodon" and

"punctatofasciatus" species complexes, is not a monophyletic group.

Within-group variation

Genetic differences among individuals were low and similar in both

"punctatofasciatus" and "rhombochaetodon" species groups. A total of 42

polymorphic nucleotide positions were observed within the "punctatofasciatus"

group, and 27 within the "rhombochaetodon" group, yielding a total of 32 and 18

unique cyt b sequences in each, respectively (table 1.2a and 1.2b). All save five of

these substitutions were silent transitions at the .third position of the codon. Of

these five, four were guanine to adenine transitions occurring at the first

position of the codon resulting in conservative valine to isoleucine amino acid

substitutions (French and Robson, 1983). The only transversion occurred at the

third position of a four-fold degenerate codon for a single C. pelewensis. The

maximum genetic difference observed in the 48 individuals of the

"punctatofasciatus" group was 3.2%, slightly lower than the 3.9% among the 31

individuals of the "rhombochaetodon" group.

Within-group variation was partitioned similarly in the two species

groups. In both, the majority of the within-group variation was partitioned

between the Pacific and Indian ocean basins. Both groups showed an

approximately 2.0% (calculated over all sites and corrected for within-species

variation) genetic break between species from each basin (tables l.3a and 1.3b).

For example, C. punctatofasciatus, C. pelewensis and C. multicinctus differ from

the Indian Ocean C. guttatissirnus by 1.7-2.0%. Similarly, within the

21

"rhombochaetodon" species group, the Pacific species, C. mertensii and C.

xanthurus, differed from the Indian Ocean C. madagascariensis by 2.3-2.6% and

from the Red Sea endemic, C. paucifasciatus, by 1.7-1.9% corrected sequence

divergence (table 1.3b).

The Indian/Pacific Ocean genetic break is clearly evident from the

phylogenetic relationships among individuals within each complex (figures 1.3a

and 1.3b). In these trees, the Pacific individuals of both groups cluster together

into a monophyletic group supported at the 0.005 level by topological dependent

tests (table 1.4) (Faith, 1991). Within the "punctatofasciatus" complex, there was a

distinct Indian Ocean/Pacific bifurcation. Individuals in each basin formed

monophyletic clusters supported in 70% and 81% of bootstrap replicates (figure

1.3a). The phylogenetic pattern within the "rhombochaetodon" species group

was slightly different. Within this group, the monophyletic Pacific clade was

rooted within the Indian Ocean group. Thus, the Indian Ocean species, C.

madagascariensis, possessed the most ancestral mitochondrial genotypes,

followed by the Red Sea endemic, C. paucifasciatus, and lastly the Pacific clade

containing C. mertensii and C. xanthurus (figure 1.3b). It was unclear if this

topology reflects the true evolutionary history of this group, the limited number

of phylogenetically informative sites, or stochastic variation caused by the

random assortment of mtDNA variation since differentiation. However, the

branching order of these groups was poorly supported in bootstrapped

replications. Neither log-likelihood nor randomization tests could reject the

hypothesis that, as in the "punctatofasciatus" species group, an initial bifurcation

divided this group into Indian and Pacific forms with subsequent differentiation

further partitioning each ocean (Kishino and Hasegawa, 1989; Faith, 1991;

Felsenstein, 1993) .

22

Genetic variation within ocean basins was much smaller than the 2%

difference observed between ocean basins. Chaetodon madagascariensis and C.

paucifasciatus , the Indian Ocean and Red Sea members of the

"rhombochaetodon" group differed by only 0.85% (corrected sequence

divergence) (table 1.3b). Despite the small genetic difference, these two species

were clearly genetically distinct. Levels of genetic variation within each species

were less than half of the corrected genetic differences between species (table

1.3b). In addition, with the exception of one C. madagascariensis possessing what

appears to be an ancestral haplotype, individuals of the two species grouped

together within the same region of the phylogenetic tree (figure 1.3b).

Similarly slight levels of genetic differences define the Pacific species of

both complexes. In these cases, however, the relationships among species were

more complicated. Average pairwise genetic diversity among all Pacific

individuals was only 0.89% in the "punctatofasciatus" species group and only

0.61 % among Pacific individuals of the "rhombochaetodon" group. Levels of

within-species genetic variation were similar to the corrected genetic differences

between species (tables 1.3a and 1.3b). Indeed, some individuals from different

species, collected thousands of kilometers apart, had identical cyt b sequences.

For example, the most common sequence, found in 6 of the 17 Pacific

individuals of the "rhombochaetodon" complex, occurred in individuals

sampled from Indonesia to American Samoa. Within the "punctatofasciatus"

complex, individuals possessing identical cyt b sequences were found in Hawaii,

Tahiti, Guam, and Indonesia, effectively across the entire West-Pacific Ocean.

Furthermore, the phylogenetic relationships among the Pacific individuals

examined were distinctly polyphyletic (figures 1.3a and 1.3b). This phylogenetic

pattern was not a function of the limited resolution of the cyt b data. Distinct

23

lineages are defined, albeit by few substitutions, within the Pacific. However,

with the exception of a single C. multicinctus specific clustering, these lineages

contain individuals from different species (figures l.3a and 1.3b). Trees

constrained to be monophyletic were significantly longer than the shortest

length non-constrained trees (p<O.OOl)(table 1.4).

24

1.5. DISCUSSION

The long-lived planktonic larval stage of many marine organisms allows

for the wide geographic dispersal of gametes that make it difficult to

conceptualize the processes by which populations become isolated and diverge

into new species (Mayr, 1954; Valentine and Jablonski, 1983). A central question

in the evolution of high IWP diversity, and marine diversity in general, is the

role that chance versus deterministic forces has played in species formation (see

Valentine and Jablonski, 1983). The nature of the IWP, with its many

archipelagoes separated by large expanses of open ocean, has led to the suggestion

that species form most often when dispersal events isolate small founding

populations (Mayr, 1954; Ladd, 1960; Valentine and Jablonski, 1983; [okiel, 1990).

Other biogeographers have argued that species formation is much less random.

Populations become isolated when geological or environmental changes erect

barriers that interrupt the pattern of gene flow across the range of a previously

well mixed ancestral population (Springer, 1982, 1988; Woodland, 1983, 1986;

McManus, 1985; Donaldson, 1986; Hocutt, 1987). The rise of the Isthmus of

Panama 3.5 million years ago is the most striking example of such an extrinsic

barrier in the marine environment (Bermingham and Lessios, 1993). However,

biogeographic barriers need not be so obvious. Changes in the spatial

distribution of habitat or alterations in current patterns can erect more subtle, yet

equally profound, divisions within marine populations (Reeb et al., 1990; Avise,

1993).

Similar levels of within-group cyt b sequence differences suggest that both

the "punctatofasciatus" and the "rhombochaetodon" species-complexes arose

contemporaneously. Moreover, the two groups show a remarkable degree of

phylogenetic concordance which includes a 2.0% genetic break between the

25

Indian and Pacific ocean and evidence for more recent differentiation within

ocean basins (figures 1.3a and 1.3b). This is especially true within the Pacific,

where very small differences among individuals highlight a recent period of

rapid species formation. In-so-much as it is unlikely that such a strong temporal

and phylogenetic concordance between independent species groups would arise

by chance, this result argues that similar historical events may underlie the

formation of both groups. This concordance provides the most compelling

support to date for the role that vicariant events have played in the evolutionary

history of this region (Croizat et al., 1974; Platnick and Nelson, 1978;

Bermingham and Avise, 1986; Cracraft and Prum, 1988; however see also, Endler,

1982).

The Indo-West Pacific has had a complicated geological and climatic

history marked by tectonic plate movement and collisions, volcanism, and

periodic sea level changes (Audley-Charles, 1981; Charlton, 1986; Chesner, 1991;

Acharyya and Basu, 1993). The central Indo-Malay region of highest species

diversity is formed by the confluence of 4 of the earth's tectonic plates (Audley­

Charles, 1981). Several vicariant hypotheses have linked the generation of new

species within this region to the movement of these plates over time (Rotondo

et al., 1981; Springer, 1982, 1988;Woodland, 1986;Hocutt, 1987). Others, by

contrast, suggest that speciation has been much more recent, and argue that

changes in sea level associated with Pleistocene periods of global cooling isolated

regions across the IWP sparking a recent wave of species formation (Woodland,

1983, 1986; McManus, 1985; Donaldson, 1986).

The timing of speciation

Because the actual rate of cyt b evolution in butterflyfishes is unclear, we

can only suggest a rough temporal framework for speciation based on rate

26

estimates in taxa with good fossil records. These estimates have varied from

approximately 1.0% per million years (averaged over all sites) in sharks to

approximately 2.5% per million years in primates and ungulates (Brown et. aI.,

198~; Irwin et aI., 1991; Martin et al., 1992). The observed correlation between

metabolic rate and rates of molecular evolution suggests that the rate of cyt b

evolution in these fishes falls somewhere between the two (Martin and Palumbi,

1993). The oxygen consumption of 10-gram coral-feeding damselfish

Plectotroglyphidodon johnstonianus at rest is less than that an adult human at

rest (0.35 mg02/g/hr versus 0.31 mg02/g/hr), but still greater than the oxygen

consumption of a swimming lemon shark (0.28 mg02/g/hr) (Altmann and

Dittmer, 1968; Bushnell et al., 1989; Gochfeld, 1991).

Using this range for the rate of cyt b evolution suggests that butterflyfish

species groups diverged from each other 5-12 mya. Evolution within each group

has been much more recent, beginning with an initial differentiation between

the Indian and Pacific Oceans basins that occurred between 0.8-2.0 mya.

However, placing the divergence of these species within a clear temporal

framework is complicated. Random genetic drift of ancestral mtDNA variation

may be obscuring the timing of species formation, or the true branching patterns

in these species groups. For example individuals of the Red Sea endemic, c.

paucifasciatus, are on average 1.20% different from individuals of C.

madagascariensis (table 1.3). Correcting for within-species levels of genetic

differences suggest that these two species may have diverged between 340,000 and

850,000 thousand years ago (Wilson et al., 1985). However, this time scale may be

an over-estimation. This is because a mitochondrial gene genealogy does not

necessarily trace a species' history, which begins with the origin of barriers to

reproduction (Neigel and Avise, 1986; Palmilo and Nei, 1988). It is possible that

27

emerging species may have been fixed for very different ancestral mitochondrial

lineages. Because these lineages predate the isolation event our estimate of the

timing of divergence will be inflated. In addition, because lineages may not

reflect the actual pattern of species branching, the assortment of ancestral

variation may obscure the phylogenetic re.lationships among species. This is

especially true when a number of species diverge from each other over a short

period of time. Indeed, the odd phylogenetic position of the Red Sea endemic, C.

paucifasciatus, intermediate between Pacific and Indian Ocean species, argues

that random assortment of ancestral lineages may be obscuring the "true"

evolutionary history of these newly formed species groups.

Ancestral mitochondrial variation may be similarly obscuring the timing

of speciation and evolutionary relationships among the Pacific species of both

complexes. Within the Pacific, identical cyt b sequences are distributed broadly

and shared between recognized species (figures 1.3a and l.3b). Presently, it is

unclear if this phylogenetic pattern reflects the history of speciation or a history

of hybridization among species that have not evolved strong reproductive

barriers. The discrepancy between the mtDNA gene tree and the species

designations may reflect the incomplete sorting of ancestral variation in very

recently formed species (Neigel and Avise, 1986; Palmilo and Nei, 1988; Morin

and Kornberg, 1993). For neutral genes, polyphyly is predicted when the time

since speciation is short relative to the long-term effective population size

(Neigel and Avise, 1986). Assuming a generation time of 2 years in these

butterflyfishes and strict maternal inheritance of the mitochondrial genome, the

observed levels of within-ocean genetic variation suggests that the three species

of the "punctatofasciatus" complex shared a common maternal ancestor between

263,000 and 868,000 years (Tajima, 1983; Tricas , 1986). A similar time to common

28

ancestry is suggested for the two Pacific species of the "rhombochaetodon"

complex. Thus, populations maintaining a historical effective female

population size of only 132,000 to 434,000 individuals would be expected to share

mitochondrial lineages over this time period (see Neigel and Avise, 1986).

Given the demographic properties of these fishes, including enormous ranges,

high fecundities, and large present-day population sizes, even greater historical

effective population sizes would seem reasonable (however see Avise et aI.,

1988). For example, C. multicinctus is one of the most common butterflyfishes

on reefs around Hawaii and on Johnston Atoll. Females spawn year round, with

a peak from mid-May through mid-October, producing an estimated 200,000­

300,000eggs annually (Reese, 1991).

If this scenario is correct, then the timing of speciation would be much

more recent than the 263,000-868,000 years suggested by the coalescence of

mtDNA variation. Traits, such as color pattern in butterflyfishes, that are

important in the social ecology of a species can coalesce rapidly under strong

social or sexual selection and are more likely to reflect the actual species

genealogy (Reese, 1975;West-Eberhard, 1983).

Alternatively, hybridization and the introgression of mtDNA genes could

potentially generate the observed discrepancy between the mtDNA gene tree and

species designations in the Pacific species of both groups (Takahata and Slatkin,

1984). The presence of a distinct C. guttatissimus sequence in a C.

punctatofasciatus collected on the Indian Ocean side of Indonesia suggests that

hybridization between species may take place in areas where they overlap (also

see Randall et al., 1977). Distinguishing between these alternatives, especially

when species have formed recently, is notoriously difficult (Slatkin, 1989).

None-the-less, the conclusion that the Pacific species of both groups formed

29

recently is supported by the low intra-ocean levels of mtDNA variation and by

slight nuclear genetic differences (Nei's D values ranging between 0.001 and 0.04)

among these species at allozyme loci (Appendix C).

Despite the difficult in reconstructing the evolutionary history of these

species groups from the history of mtDNA variation, the low levels of mtDNA

variation clearly place the origins of both species groups within the Pleistocene.

This temporal framework is an order of magnitude more recent than expected

from models linking IWP diversification to the movement of tectonic plates.

However, this timing is predicted by vicariant models focusing on recent

physical changes in this region associated wi~h Pleistocene sea level fluctuations.

Pleistocene changes and the origins of these species groups

As many as seven times during the past two million years, drops in sea

level of 100-150 meters stranded much of the shallow water shelf environment

of the IWP, restricting connections among the numerous ocean basins that

comprise this region (Porter, 1989). Although only the Red Sea was completely

isolated during lowered sea stands, concomitant changes in coral reef habitat and

ocean circulation patterns are speculated to have generally disrupted gene flow

patterns across the IWP. This is particularly true between the Indian and Pacific

ocean basins, where increased temperature and decreased salinity in the Banda

Sea are envisioned to have aided the separation of these two basins (Fleminger,

1986). Today, the ranges of a number of species and subspecies pairs from a broad

array of marine taxa abut at or near the Indonesian archipelago (Allen, 1980;

Woodland, 1983; McManus, 1985; Blum, 1989). For example, within

butterflyfishes,9 of the 31 species complexes show species or sub-species

distinctions in this area (Blum, 1989). A similar biogeographic pattern is evident

within rabbitfishes (Family Siganidae; Woodland.. 1983), copepods (Labidocers:

30

Undinula, Fleminger, 1986), anemonefishes (Amphiprion; Allen, 1972), and

hermit crabs (Uca; Crane, 1975). These patterns led Woodland (1983) to suggest

two focal centers of diversification: an Indian Ocean focus and another in the

Pacific Ocean. Unlike previous biogeographers (see Ekman, 1953; Briggs, 1974),

Woodland attributed the extraordinary high species diversity within the shallow

waters surrounding Indonesia, the Philippines and New Guinea to the

confluence of these separate biotas. The phylogenetic patterns within these two

butterflyfish complexes are consistent with this explanation. A similar 2.0% total

mtDNA genetic break between Indian and Pacific ocean populations of the

coconut crab, Birgus latro, underscores the broad role that this barrier played in

the historical biogeography of this region (Lavery, 1993).

It is much less clear what role sea level fluctuations may have played in

the rapid diversification within the Pacific. Unlike the physical barriers created

between the Red Sea, and the Indian Ocean and the Indian and Pacific Oceans,

drops in sea level do not obviously partition this region. It is possible that

Pleistocene sea level regressions may have acted subtly to subdivide the Pacific.

Today, the predominately westward-flowing surface currents of the tropical

Pacific are frequently interrupted by El Nino/Southern Oscillation (ENSO)

events. These current anomalies profoundly alter the direction and speed of

surface currents across this region, providing a potent mechanism for the

eastward transport of planktonic larvae (Richmond, 1990; Vermeij, 1990).

However, paleoclimate data suggest that wind patterns were more intense and

equatorial productivity higher during glacial maximums than during interglacial

periods (Molina-Cruz, 1977; Pedersen, 1983; Janecek and Rea, 1985; Chuey et al.,

1987; Rea, 1991; but see Rea, 1990). These data have led to the suggestion that the

westward flowing surface currents were stronger and ENSO events less frequent

31

during glacial maximums (Quinn, 1971). Therefore during glacial maximums, it

is possible that genetic exchange between the western Pacific and archipelagoes of

the eastern Pacific may have been interrupted long enough for new species to

form. Pollock (1992) also focused on possible alterations and intensification of

surface current patterns during glacial maximums to explain the apparent post­

Pliocene speciation in the spiny lobster genus Panuliris.

Similar patterns of recent and rapid diversification within the Pacific is

emerging from genetic examination of other groups. The four Pacific members

of a third species group of Butterflyfishes, the "tinkeri" complex, likewise, show

slight differences «1.0%) at the 495 bp portion of the mitochondrial cyt b (figure

1.2). Similarly, the sea urchin Echinometra mathaei, previously thought to be a

single IWP species (see Mayr, 1954), has recently been shown to be composed of at

least four species which have all apparently arisen within the last 2.0 million

years (Uehara et al., 1986; Palumbi and Metz, 1991; Palumbi, 1992). These

emerging patterns suggest that whatever events led to the rapid differentiation of

these two butterflyfish groups in the Pacific, may have acted more broadly

affecting many taxa across this region. Additional work on paleo-oceanographic

and paleoclimatic conditions can only help clarify the relationship between

physical and oceanographic changes across the Pacific and diversification.

Conclusions

The evolutionary mechanisms underlying the extraordinary diversity of

the IWP are complex, reflecting an interplay between life history, adult ecology,

and historical events. Much additional work on the tempo and patterns of

evolution across the IWP is necessary to gauge the effects of recent historical

changes on biodiversity in this region. At the very least, the congruent patterns

of rapid and recent diversification in these groups highlights a turbulent period

32

in the history of this region. These results challenge the notion that the

extraordinary diversity of this region reflects long-term stable environmental

and physical conditions across this region (Ekman, 1953; Briggs, 1974). Profound

evolutionary changes associated with Pleistocene climate fluctuations have

recently been proposed for the temperate Atlantic and the Caribbean marine

faunas (Stanley, 1981; Jackson et al., 1993; Allmon et al., 1993). However, unlike

these other regions, extinction has apparently not played a significant role in the

recent history of the IWP (Vermeij, 1987). Indeed, the phylogenetic pattern

within these butterflyfishes, inferred from mtDNA sequences, suggests that this

region may have been characterized by extraordinary bursts of species formation.

A similar conclusion is emerging from the fossil record, which shows a rapid

increase in diversity of a number of IWP mollusks since the Miocene and

Pliocene, much of which may well have occurred within the Pleistocene (Kohn,

1990; Kay, 1994).

Significant demographic changes in IWP taxa, including both local

extinction and speciation, were probably associated with the most recent low

sea level stand which ended less than 10,000 years ago (Crame, 1987; Porter,

1989; Pauley, 1990; Springer, 1992). Future ecological and biogeographic work

in this region must be tempered with the understanding that the recent

history of this region has been tempestuous.

33

Table 1.1: Number of transitions/transversion (above diagonal) and genetic distances (below diagonal) DNA

substitutions in a 495 pb segment of the mitochondrial cytochrome b gene. Genetic distances were corrected for

multiple hits at the same nucleotide position using the Kimura 2-parameter model (Kimura, 1980) in which

tranversions were weighted 10 times transitions. See Appendix A for sequences.

2 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

~

1. C.tinklll

2. TYPE1

3. TYPE2

4. TYPE3

5. TYPE4

6. C.mer/C.xan

7. C.mer/C.xan

8. C.mertensii

9. C.madaagas.

10. C.paucifas.

11. C.argentatus

12. C.miliaris

13. C.multi.

14. cmu,Cpu,Cpe

15. C.punctato.

16. C.pelewensis

17. C.guttat.

18. C.auriga

•1.2

0.8

1.2

0.9

8.5

8 0

8 0

8.3

8 7

8.4

12.5

11.2

12.0

11.5

12.0

12.2

17 3

5/0

•0.2

0.5

0.7

8.7

8.4

7.8

8.1

8.7

7.2

11.7

10.7

11.5

10.7

11.2

11.2

16.9

4/0

1/0

0.2

0.5

8.1

7.6

7.5

7.8

8.3

8.0

11.5

10.8

11.5

11.0

11.5

11.8

16.2

5/0

2/0

I/O

•0.7

8.7

8.4

7.8

8.1

8.7

7.2

11.7

10.7

11.5

10.7

11.2

11.2

16.2

4/0

3/0

2/0

3/0

•9.0

8.7

8.0

8.4

8.9

7.5

12.0

10.4

11.2

10.4

11.0

11. 0

17.1

37/2

33/2

35/2

33/2

34/2

•0.4

1.7

2.7

1.9

8.7

14.1

12.2

13 .0

12.5

13.0

13.7

15.6

35/2

32/2

33/2

32/2

33/2

2/0

•1.7

2.7

1.9

8.7

13.6

12.2

13.0

12.5

13 .0

13.7

15.6

34/2

29/2

32/2

29/2

30/2

8/0

8/0

•1.9

1.5

8.1

14.5

12.5

13.3

12.8

13.3

13.5

16.2

35/2

30/2

33/2

30/2

31/2

12/0

12/0

8/0

1.2

7.5

13.9

11.9

13 .5

12.7

13.3

13 .5

15.9

38/2

33/2

36/2

33/2

34/2

9/0

9/0

7/0

5/0

7.5

13 .9

12.5

13.7

13.2

13.7

14.0

16.1

37/0

28/0

35/0

28/0

29/0

36/2

36/2

33/2

30/2

31/2

•12.4

13 .4

14.5

14.0

14.5

13.7

16.5

45/7

39/7

41/7

39/6

40/6

51/7

49/7

51/7

49/7

50/7

42/7

•10.9

10.9

10.9

10.4

10.4

17 .4

43/7

35/7

41/7

35/7

34/7

47/7

47/7

47/7

45/7

48/7

49/7

40/6

•1.4

0.6

1.0

2.5

17.6

46/7 44/7

38/7 35/7

44/7 42/7

38/7 35/7

37/7 34/7

50/7 48/7

50/7 48/7

50/7 48/7

50/7 48/7

53/7 51/7

53/7 51/7

40/6 40/6

7/0 3/0

• 4/0

0.8 •

0.4 0.4

2.3' 1.9

17.9 17.3

46/7

37/7

44/7

37/7

36/7

50/7

50/7

50/7

50/7

53/7

53/7

38/6

5/0

2/0

2/0

•1.9

17.9

47/7

37/7

45/7

37/7

36/7

53/7

53/7

51/7

51/7

54/7

50/7

38/6

12/0

11/0

9/0

9/0

•17.6

57/15

49/13

53/15

47113

50/13

53/13

53/13

54/13

53/13

55/13

51/15

53/15

57/16

58/16

56/16

58/16

57/16

•

Table 1.2: Species designation, collection site, and polymorphic nucleotide positions within a 495 bp portion of the mtDNA

cytochrome b gene region from a) the 33 individuals representing the four species that comprise the "punctatofasciatus" complex and

b) the 48 individuals representing the four species that comprise the "rhombochaetodon" species complex. In this table, dots indicate

positions that are identical to the most common sequence within each complex. Polymorphic sites highlighted in bold-face occur at

the first position of the codon. Missing nucleotide information is denoted by a"?".

A. The "punctatofasciatus" complex:1 1 1 1 111 122 2 2 222 2 ~ 2 3 3 3 3 3 3 3 3 3 444 4 4 4 4 4

2 4 6 8 9 9 2 2 3 567 8 901 3 4 677 8 B 9 0 0 122 4 7 8 9 0 1 1 3 4 5596 7 8 9 1 0 3 0 9 5 625 3 3 164 3 709 2 3 103 8 1 422 1 6 2 1 7 8 7 092

Species LocationC. rnul, C. pel, C. pun HI, Ta., Sa., Guam, Indo.(4) C C T Tee A CAT A A a T 0 A G A A T T Teo eTC C A A T T eGA eGA T C G CC. rnulticinctus Hawaii · · · · · · • A • • G • · · · · · · · · · · • G • • T • · · · • CC. mullicinetus Hawaii(2) · · · · · · · · · · · · · • C • · • T • · · · · · · · · · · · · • TC. multicinetus Hawaii(5) · · · · · · · · · · · • G • · · · · · · · · • G • · • T • · · CC. multieinelus Hawaii · · · · · · · · · · · • G • · · · · · · • G • · • T • · · • G CC. multicinelus Hawaii · · · · · · · · · ·· · · ? · · · · · · · · • G • · • T • · · · • C •

~ c. multicinctus Hawaii · · · · · · · · · · • A G • · · · · · · · · · • G • • T • · · · • CC. multicinctus Hawaii · · · · · · · · · · • G • • G • · · · · · · ? G • · • T • · · · • CC. rnult'cinctus Hawaii · · · · · · · · · · · • G • · · · · · · · · • G • · • T • · • A • CC. multicinctus Hawaii · · · · · · • G • · · · · · · · · • G • · • TC. punctatofasciatus Guam · · · · · · · · · · · · · · · · · · · · · · · · · • C • · · · · · · · ?C. punctatofasciatus Guam · · · · · · · · · · · · • G • · · · · • T • · · · · · · • T •C r\;netalofascialus Guam · • C • • T G • G • G • · · • G • · · · · · • T • · · · • A •C. punelalofascialus Guam · · · • T • · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ?C. punctatofasciatus Guam · · · • T · · · · · · · · · · · · • C • • T • · · · · ·C. punelatofasciatus Philippines · · · · · · · • GA· · · · · · · · · · · · • T • · · · · · · · ?C. punctatofasciatus Philippines · · · · · · · · · · · • C • · • T • · · · · · • TC. punctatofasciatus Philippines · · · · · · · · · · · · · · · · · · · • GC. punctatofasciatus Indonesia T • · · · · • C • G • · • A • • C • · · · · · · · · · · • T A • T ?C. punelalofasciatus Indonesia • T • · · · · · · · · • G •C. punelalofascialus Indonesia · · · · · • T • · · · · · · · · · · · · · · · · · · · · · · · ? ?C. pelewensis Samoa · · · · · · · · · · · • C • · • T • · · · • CC. pelewensis Tahiti · ? · · · · · · · · · · · · · · · · · · · · • T • · · · · · · · · · · · ?C. pelewensis Tahili · · · · · · · · · · · · · · · · ··· · · · · · · · ·· · · · · • AC. pelewensis Tahiti(2) · · · · · · · · · • G • · · · · · · · · • TC. pelewensis Tahiti · · · · · · · · · • G • · · · · · · · · • T • GC. pelewensis Tahiti · · · · · · · • C • G • · · · · · · · · · · · · · • TC. pelewensis Tahiti · · · · · · · · • GA· · · · • A • · · · · • TC. pelewensis Tahiti · · · · · · · • G • · · · · · · · · · • T • · • AC. guttatissimus Mauritius T • · · · · · • C • G • · • A • • C • · · · • C · · · · • T A • T · · • AC. guttatissimus Maurilius(3), Indonesia T • · · · · • C • G • · · • A • • c • · · · · · · · · · • T A • T • · · • AC. gutlalissimus Indonesia T • • C • T • · • C • G • · · • A • , C • · · • C · · · · • T A • T • · • A

Table 1.2: Species designation, collection site, and polymorphic nucleotide positions within a 495 bp portion of the mtDNA cyt b gene

(continued).

B. The "rhombochaeiodon" complex:

Species

e. xan, e. mer.e. mer, e. xan,e. mertensiie. rnertensiiC. xanthuruse. xanthuruse. xanthurusC. madagascariensisC. madagascariensisC. madagascariensisC. madagascariensisC. paucifasciatus

~ C. paucifasciatusC. paucifascialus

Location

Indo., PhiL(3), Guam(2)Guam(2), Samoa(3), Indo.

TahitiSamoa

PhilippinesIndonesiaIndonesiaMauritiusMauritius

Mauritius(4)MauritiusRed SeatS)

Red SeaRed Sea

1 1 1 1 2 2 2 2 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 42 7 9 3 5 8 8 1 5 7 900 1 123 5 6 8 0 0 1 1 2 6 7:l 5 9 5 3 0 3 9 504 0 6 5 840 1 6 4 2 8 1 4 020

aCCTTATCCTCCAATGTGTCAACTATTT • . · · · . · · · · · . . . · . • CT • . · · • T • · · · · · · · • T G G T CT • C

A • · · · · · • GT • · . • G? · · · · · • C

T C • G C • T CT. · · · · · • CT. G • C G • CT C • ·C?TCT· · • CA' ACT • G • C G • CT C • • C • T CT' · · · • CT. G • C G • CT C • • C • T C • · · · · · · • CT. G • C G • CT C • · · • T CT' G • · · · · • T • • C • • C

• T T C • · · ? T CT. · · · · • CT. . • C • C• T T C • · · • T C T T • · · · • CT· • C • • C

Table 1.3: Average percent genetic variation within species (boldface, along

diagonal), percent genetic differences between species (d~) uncorrected (below

diagonal) and percent genetic differences between species corrected for within­

species genetic variation (above diagonal) of butterflyfishes in the a)

"punctatofasciatus" species complex and b) "rhombochaetodon" species complex.

Uncorrected genetic difference values were calculated from by averaging all

pairwise differences, corrected for multiple substitutions at the same nucleotide

position, among individuals of each species. For corrected values, the average

level of within-species genetic variation was subtracted from the uncorrected

genetic difference values (Wilson et al., 1985). Numbers in parenthesis represent

the number of individuals sampled from each of the eight species.

A. The "punctatofasciatus" species complex(n=48):

PACIFIC INDIANC.multi. C. pmletata. C.pelewensis C. guttatissimus

C. multicinctus (n=15) 0.0060 0.0036 0.0024 0.0201

C. punctatofas. (n=16) 0.0110 0.0089 0.0004 0.0167

C.~Iewensis (ne l l ) 0.0084 0.0078 0.0060 0.0171

C. guttatissimus (n=6) 0.0250 0.0230 0.0220 0.0038

37

Table 1.3: Average percent genetic variation (continued).

B. The "rhombochaetodon" complex (n=31):

PACIFICc. mertensii C xanthurus

INDIANC. rnadagas. C. paucifas.

k mertensii (n=9) 0.0062

C. xanthurus (n=8) 0.0066

C. madagascar. (n=7) 0.0280

C. paucifasciatus (n=7) 0.0220

0.00165

0.0039

0.0300

0.0230

38

0.0234

0.0265

0.0031

.00120

0.0170

0.0191

0.0085

0.0039

Table 1.4: Hierarchical structure within the phylogenetic tree of members of the

"rhombochaetodon" and "punctatofasciatus" complexes. *,.,. represents

significant differences at the 0.005 level between the length of the observed

phylogeny and the length of the shortest length constrained phylogeny.

Significance values were obtained by comparing the observed differences

between two topologies to the difference obtained by randomly permuting the

original data matrix 500 times. In these comparisons the sequence of the

outgroup taxa, indicated by parenthesis, was not randomized.

A. "punctatofasciatus" complex:Indian Ocean species as outgroup to Pacific species <h:.miliaris ): .

Shortest length tree (figure 1.3a):Shortest length non-Indian Ocean outgroup:

Monophyly within Pacific <h:. guttatissimus)Shortest length monophyletic grouping:Shortest length non-monophyletic grouping (figure 1.3a):

B. "rhombochaetodon" complex:Indian Ocean/Red Sea species as outgroup to Pacificspecies <k tinkeri):

Shortest length tree (figure 1.3b):Shortest length non-Indian Ocean/Red Sea outgroup:

Monophyly within Pacific <h:. paucifasciatus ):Shortest length monophyletic grouping:Shortest length non-monophyletic grouping(figure 1.3b):

39

Length

8992***

55***46

6466***

24***18

+N

JohnstonAtoll

~waiianIS.IIdUl1IMarshall Is.

IlitH!I

:••. M~[S'~•...g~~~~~ "~'''~ Pitcairn

Scale (Kilometers) ~l~~~-~~~~=. . ----,

o 2001

"'"o

Figure 1.1a. The geographic distribution of the species within the" punctatofasciatus " species group (Q

; Chaetodon guttatissimus;§ I C. punctatofasciatus £2J ,C. pelewensis; andlITIl , C. multicinctus ).

Shading should be interpreted only as approximate distributions. For site specific distribution see

Blum (989). The @'s represent sampling localities.

+N... ,jiawaiian Is.

"".Johnston Atoll

~~~M~'qeusas Is.

~~~~.F;j;~. sooe~~~Pk<.;,n

Scale (Kilometers)I I Io 200

*'">-'

Figure Llb. The geographic distribution of the species within the" rhoinbochaetodon "species group (Illll, Chaetodon paucifasciatus fSJ ,..k madagascariensis § ,C. xanthurus ; andE1, c. mertensii ).

Shading should be interpreted only as approximate distributions. For site specific distribution see

Blum (1989). The @'s represent sampling localities.

Figure 1.2: Phylogenetic relationships among 15 species from three subgenera

within the large genus Chaetodon. This tree was constructed using parsimony

with C. auriga defined as the outgroup (see Blum, 1988). The neighbor-joining

tree differed only in the placement of C. argentatus, which grouped with the

members of the "rhombochaetodon" species group. The relative strength of

major nodes within this phylogeny is given by the percentage agreement in a

consensus of trees produced from a 500 bootstrap simulations of the data set

(above) and the number informative nucleotide changes (below). The

number of informative nucleotide substitutions is given in both transitional

(listed first) and transversional (listed second) changes .

42

C.declivis wilderi

C.dec1ivis wilderi Roaops100

18C. burgessi

693/2 C. fIavocoronatus

1

C.argentatus I Exornator17

C. mertensii85 1 C.xanthurus

17/2 C.mertensii1 C.xanthurus

C.mertensii Exornator99 C.paucifasciatus11 1

C. madagascar.1

C.miliaris17/3

92

17/2

100

18/3 C. pelewensisa

Exornator

4

~ Rabdophorus29/11L. ~ c. auriga

43

Figure 1.3a: Major phylogenetic relationships within the "punctatofasciatus"

species complex. The tree was constructed using the neighbor-joining

method from pairwise distance matrices estimated by the Kimura two­

parameter model (Kimura, 1980). Support for major nodes within these trees

is shown as the percentage agreement in a consensus of trees produced from

500 bootstrap simulations of the data set (above) and the number informative

nucleotide changes (below). Shaded boxes represent clusters of individuals of

the same species with one exception. A C. punctatofasciatus collected from

the Indian Ocean side of Indonesia had the distinct C. guttatissimus genotype

(see table 1.1). Species symbols are as follows: C. punctatofasciatus (@), c.pelewensis (~ ) and C. multicinctus(tI). The tree was rooted with sequences

from C. miliaris and C. tinkeri. Phylogenetic trees constructed using

parsimony were nearly identical. Consistency and retention indexes for these

trees were 0.78 and 0.86. Since all phylogenetically informative substitutions

within groups were transitions, different transition/ transversion weighting

schemes did not alter the branching pattern.

44

(/)

(/) fU ~

::s .... =.... (/) ....- ~ ~ fU....

A. The" punctatofasciatus "complex.... r:: ~ 6 .... ....S 0 - fU0 .... ....- lU e ~fU "'0 .... ::s ..s::~ r:: ..s:: fU lU fU- c.. o 00 f- :t

81C. guttatissimus 5 2

3

Indian

100Pacific

f @ 1Ii () 1

18~ 1

78 @ 15 <t 1

ct 1

o 2

12C. multicinctus

@ 12

11

11

4 1 1 1 11

11

1 ..J.

I I1

0.0 1.0% sequence divergence

45

Figure 1.3b: Major phylogenetic relationships within the

"rhombochaetodon" species complex. This tree was constructed using the

neighbor-joining method from pairwise distance matrices estimated by the

Kimura two-parameter model (Kimura, 1980). Support for major nodes

within these trees is shown as the percentage agreement in a consensus of

trees produced from 500 bootstrap simulations of the data set (above) and the

number informative nucleotide changes (below). Shaded boxes represent

clusters of individuals of the same species. Species symbols are as follows: C.

xanthurus Q), and C. mertensii «(1]). This tree was rooted with sequences

from C. miliaris and C. tinkeri. Trees constructed using parsimony were

identical. Consistency and retention indexes for parsimony trees were 0.76

and 0.84. As in the "punctatofasciatus" species-complex, all phylogenetically

informative substitutions within groups were transitions.

46

lUeo

CIJ ~

lU ~ 0CIJ .... ~

~lU .E CIJ ....QJ ~ Q.,

C/'}....

~ Q., S 0 ....... -::J 0 ....S ....

"'t:S - lUlU "0 .... ::J ..c:

QJ

:E ~ ..c: lU lU

B. The" rhombochaetodon " complex ~ - c.. Co' Cf} E-o

C. madagascariensis 1

99 C. madagascariensis 6~

11

63

3 C. paucifasciatus 7

47

CHAPTER 2

Contrasting Patterns of Phenotypic and MtDNA Variation Among Recently

Differentiated Butterflyfishes (Family Chaetodontidae)

2.1. ABSTRACT

Three very closely related butterflyfishes (Family Chaetodontidae),

differing principally in color pattern, partition the tropical west-Pacific into large

allopatric ranges. Partial sequences from the hypervariable proline tRNA

(tRNAPro) end of the mitochondrial control-region in 138 individuals of these

three species reveal a rich history of demographic change. From a common

ancestor, MtDNA variation coalesced into three major lineages. This period of

rapid coalescence appears to have been followed by a wave of demographic

expansion in which newly arising variation accumulated in all three lineages.

Only one of the three major mtDNA lineages that formed during the early

history of this group shows any species or geographic specificity. This lineage is

confined to the Hawaiian archipelago/Johnston Island endemic Chaetodon

multicinctus. The remaining two lineages are geographically widespread and are

composed of similar numbers of C. pelewensis and C. punctatofasciatus with

occasional individuals of C. multicinctus. However, significant differences in

the distribution of these lineages in populations across the south and western

Pacific defines a broad cline with one lineage dominating populations in the

48

western Pacific and the other lineage more likely to be found in the eastern south

Pacific.

The present genealogical and geographic patterning of mtDNA variation

suggests that color pattern differences that define species likely emerged at the

same time as the three major mtDNA lineages and that hybridization accounts

for the present discrepancy between the three major mtDNA lineages and

taxonomic boundaries. Within this framework, the genetic distinctiveness of

the Hawaiian species indicates very low levels of hybridization with either C.

pelewensis or C. punctatofasciatus, possibly due to greater reproductive or

geographical isolation. Between C. pelewensis and C. punctatofasciatus, the

genealogy of mitochondrial variation argues for extraordinarily high levels of

hybridization (Nm values> 100 individuals per generation) and associated

introgression. Yet, characteristics which define these species, primarily color

pattern, change abruptly at borders of allopatric species ranges. These contrasting

patterns suggest that strong selection is maintaining the cohesion of species­

specific color pattern in the presence of potentially homogenizing levels of gene

flow.

Keywords: marine fishes, mitochondrial DNA, control-region, gene tree, species

tree, hybridization, lineage sorting, Tropical Pacific.

49

2.2. INTRODUCTION

Because of its rapid rate of evolution, lack of recombination, and apparent

neutrality, the mitochondrial genome has become an important marker for

tracing the history of closely related species (Wilson et al., 1985; Whittam et al.,

1986; Avise et al., 1987; Harrison, 1989), However, in a number of instances,

discrepancies between the mtDNA gene genealogy and species boundaries

defined by phenotypic characters have been reported (Powell, 1983; Ferris et al.,

1983; Spolsky and Uzzell, 1984, 1986; Lamb and Avise, 1986; Carr et al., 1986;

Ovendon et al., 1987; Baker et al., 1989; Dowling, 1989; Moran and Kornfield,

1993). In some cases, these discrepancies can be attributed to the plasticity of the

characters used to infer taxonomic distinctions. For example, Ball et al., (1988)

attribute much of the observed morphological variation within Red-winged

blackbirds to environmental rather than genetically based differences in plumage

and size characteristics. In other instances, however, discrepancies between

mtDNA genes and "species-defining" phenotypic traits reflect underlying

differences in selective forces operating on different loci. For example,

discrepancies between mtDNA gene trees and species boundaries can arise

through hybridization even when there is strong selection against hybrid

offspring (see Takahata and Slatkin, 1984). Alternatively, such discrepancies can

occur when species have formed very recently. In this case, differences between a

mtDNA gene tree and species-defining traits may reflect differences in the

selective pressures acting during species formation (see Neigel and Avise, 1986;

Palmilo and Nei, 1988). For example, Moran and Kornfield (1993) attributed the

polyphyly of species in the rock-dwelling cichlid fishes of Lake Malawi to the

retention of ancestral mtDNA variation in very recently formed species. In these

fishes, strong selection on morphological and color variation during

50

differentiation apparently caused these traits to coalesce more rapidly than

mtDNA genes.

Rather than being viewed as a limitation of the phylogenetic resolving

power of mtDNA, discrepancies between gene trees and species boundaries can

yield rich insights into the evolutionary process. This is because the distribution

of neutral variants within and between closely related species provides a

powerful demographic framework for interpreting the nature and importance of

inter-specific differences (Neigel and Avise, 1986).

Here we focus on the discrepancy in the pattern of phenotypic and

mtDNA variation in a group of three closely related Pacific Butterflyfishes.

Members of this group of tropical marine fishes are common and conspicuous

inhabitants of coral reef communities around the world (Burgess, 1978; Allen,

1980). The three species, Chaetodon punctatofasciatus, C. pelewensis, and C.

multicinctus, partition the tropical west Pacific into large, nearly allopatric ranges

(see Blum, 1989). Chaetodon punctatofasciatus is found on coral reefs from the

Indonesian archipelago eastwards to the Marshall Islands and New Guinea

where it is replaced by C. pelewensis, whose range continues to the Society

Islands some 5000 kilometers to the east. Chaetodon multicinctus has the most

limited distribution of the three, and is restricted to reefs along the 1500

kilometer Hawaiian archipelago and at Johnson Island (see figure 2.9).

Previous phylogenetic and taxonomic evaluation of this group failed to

find any morphological differences (Burgess, 1978; Blum, 1988). However, these

species differ distinctly from one another in color patterning, Chaetodon

F!!nctatofasciatus and C. pelewensis are the most similar, differing primarily in

the striations along the body, which run vertically in C. punctatofasciatus but

diagonally in C. pelewensis (figure 2.1). Chaetodon multicinctus is more distinct.

51

Its underlying body stripes are similar to C. punctatofasciatus, but it lacks the

striking yellow wash, bright orange splash on the caudal fin, and the yellow eye­

band that characterize it's congeners. These color pattern differences are

preserved across the broad geographic range of each species with little overlap

and no evidence for a broad gradient between species patterns (Burgess, 1978;

Allen, 1980; Blum, 1989;). For example, individuals of C. punctatofasciatus

collected from coral reefs around Guam strongly resembled the individuals

collected from reefs in Indonesia some 4000 kilometers to the west. The same

pattern is true for C. pelewensis where our Tahitian, Cook Island and Fijian

collections showed only minor variation around the phenotype depicted in

figure 2.1.

In contrast, mtDNA variation is not partitioned nearly so neatly among

the three species (Chapter 1). The morphological similarity of this group is

paralleled by only slight differences at the mitochondrial cytochrome b gene (cyt

b) which suggest that these species have diverged recently. Levels of genetic

differences in this gene region suggest that all three species diverged from their

Indian ocean sibling, C. guttatissimus, between 1-2 million years ago and from

each other within the last 300-800,000 years. However, within our sample of 41

individuals, cyt b sequences failed to cluster into species-specific lineages and it

was common to find individuals of different species from widely scattered

locations across the Pacific with identical cyt b sequences.

The small sample size and slight genetic differences in the cyt b sequences

limited insights into the reasons for the observed discordance between mtDNA

and phenotypic variation. As a result, we have expanded our original data set to

include sequences from a hypervariable portion of the mitochondrial control­

region in 138 individuals from 3-4 populations across the range of each of the

52

three species. The high rate of evolution in the control-region allows us to

distinguish very closely related mtDNA haplotypes and makes it possible to

refine our picture of the genealogical and geographic patterning of mtDNA

variation among these species (Meyer et al., 1990; Vigilant et al., 1991; Wenink et

al, 1993). This high resolution mtDNA gene genealogy provides insights into

the evolutionary and demographic history of this group that are important for

evaluating the selective framework shaping the striking geographical and

phenotypic patterns that define these species.

53

2.3. MATERIALS AND METHODS

Fishes were collected from eight widely scattered localities across the

Pacific between 1990 and 1993. The species and populations sampled were as

follows: for Chaetodon multicinctus, Oahu, Hawaii (n=15), Midway Island

(n=ll), and Johnston Island (n=6); for C. pelewensis, Moorea, Society Islands

(n=9), Rarotonga, Cook Islands (n=18), Pago Pago, American Samoa (n=2), Viti­

Levu, Fiji (n=21); and for C. punctatofasciatus, Guam (n=8), Palau (n=20),

Philippines (n=21) and Bali, Indonesia (n=7). In addition, 3 C. guttatissimus, the

Indian Ocean member of this species-group, were also included as an outgroup

(see Blum, 1989). Individuals were either frozen at _700

C or preserved in 95%

ethanol. With the exception of the Philippine and Fiji samples, all collections

were made by WOM or by a biologist working in the area. Individuals from the

Philippines and Fiji were acquired through the tropical fish trade directly from

collectors working in these areas. The collection is currently stored in the _700

C

freezer at the Pacific Biomedical Research Center, Kewalo Marine Lab, Honolulu,

Hawaii and is available upon request. Voucher specimens for each population

are deposited at the Bernice P. Bishop Museum, Honolulu, Hawaii.

Genomic preparations and PCR

For each fish, approximately 0.5 grams of gill or muscle tissue was

homogenized in 2 ml of cold grinding buffer (0.2M NaCl/0.05M EDTA, pH 8.0).

Tissue stored in 95% ethanol was soaked in 5 ml of grinding buffer on ice for

approximately 1 hour before grinding. Approximately 0.5 ml of this homogenate

was transferred into a 1.5 ml micro centrifuge tube. Sodium dodecyl sulfate (SDS)

and proteinase K (Sigma Chemicals) were added to the homogenate to a final

concentration of 1% (v /v) and 20 ~g/ml, respectively. Following an overnight

incubation at 50° C, this solution was extracted twice with equal volumes of

54

buffered phenol, once with an equal volume of phenol/chloroform/isoamyl

alcohol (25:24:1) solution, and lastly with an equal volume of a

chloroform/isoamyl alcohol solution (24:1). DNA was recovered by cold-ethanol

precipitation in the presence of sodium acetate (see Sambrook et aI., 1989). The

resulting pellet was dissolved in 100-200 JlI of sterile IX TE solution.

Approximately 100 ng of total cellular DNA was subjected to 40 cycles in a

Perkin Elmer Cetus DNA Thermal Cycler in a 100JlI reaction volume with 1.5

units of Taq DNA polymerase (Perkin Elmer Cetus). We used a 12s rRNA

primer (12sar) (5'-CATATTAAACCCGAATGATAITf-3') and a Chaetodon

specific control-region (C.R.-I) (5'-ACCATATTATGTACTAGGCAC-3') to

amplify most of the control-region of these fishes (figure 2.2). This primer was

designed from initial sequences obtained from PCR amplifications using the 12s

rRNA primer and a proline tRNA (tRNAPro) primer, (5'­

CTACCTCCAACTCCCAAAGC-3') (see Palumbi et al., 1991). The standard

thermal cycle profile was: 1) 45 seconds at 94° C, 2) 45 seconds at 55° C and 3) 45

seconds at 72° C. Following amplification, the double stranded product was

collected by ethanol precipitation, resuspended in 10Jll sterile water, and

electrophoresed through a 1% agarose/O.5X TAE gel. The resulting

approximately 1200 base pair (bp) band was cut from the 1.0% agarose/D.Sx TAE

gel, Gene cleaned (© biolab 101), and resuspended in 7-30 JlI TE. Seven

microliters of this product were used as the template for double-stranded

sequencing reactions (Palumbi et aI., 1991). These reactions were carried out

using the reagents and DNA polymerase enzyme provided in the Sequenase®

2.0 DNA sequencing kit (USB). Sequencing reactions were electrophoresed

through 6-8% polyacrylamide/7M urea gels with a buffer density gradient, dried

under vacuum at 80° C and exposed to autoradiographic film for 24-72 hours.

55

Refer to Palumbi et al. (1991) for the details of template preparation, sequencing,

and electrophoresis.

Phylogenetic analysis

Sequences were aligned by eye and phylogenetic analysis was performed

using both maximum parsimony and Neighbor-joining distance methods

(Saitou and Nei, 1987; Swofford, 1993). For parsimony analysis the large number

of taxa made the evaluation of all possible topologies impractical. As a result, a

subset of possible trees was evaluated using the phylogenetic package PAUP (3.0s,

Swofford, 1993) under the heuristic search setting. The tree bisection and

reconnection (TBR) branch swapping algorithm was chosen to search for optimal

trees within this framework. To determine the effect that

transition/transversion weighting had on parsimony trees, we weighted

transversions 0, 10, or 20 times more than transitions. The genetic distances

between individuals used in the construction of Neighbor-joining trees were

corrected for multiple hits at the same nucleotide position in two ways. For the

cyt b data, we used Kimura's two parameter model (Kimura, 1980). For the

control-region data, we used the gamma correction on Tamura and Nei's (1993)

model to estimate ilij (MEGA, Kumar et al., 1993). This model was developed

specifically for the analysis of sequence data within the control-region where

strong transitional biases, differences in the types of transitional substitutions,

and site to site variation in the rate of substitution have all been observed

(Kocher and Wilson, 1991; Tamura and Nei, 1993; Kumar et al., 1993). For these

distance calculations we used k=0.6. We choose this value based on the fit of

initial control-region sequences to a negative binomial distribution with a

k=O.60. This parameter was estimated by iteration using equation 2.15 of

Southwood (1978):

56

klog(1+x/k)=log(NIno)

where ~ is the mean number of substitutions per site, N is the number of sites,

and llQ is the number of positions where no substitutions occurred. Our

estimated value of Is. was similar to values estimated from the tRNApro end of

the human control-regionfwakeley, 1993). All phylogenetic trees were rooted

using aligned sequences from C. guttatissimus. This species differed by

approximately 2.0% in a 500 bp portion of the mitochondrial cyt b and previous

phylogenetic analysis of a number of Indo-west Pacific butterflyfishes within the

subgenus Exornator identified C. guttatissimus as the most closely related

outgroup (Chapter 1).

The relative strengths of major nodes within our Neighbor-joining tree

were examined by bootstrap resampling of the original data matrix 100 times

using the computer program MEGA (Felsenstein, 1985; Kumar et al., 1993; Hillis

and Bull, 1993). In addition, we used the randomization procedure described by

Faith (1991) to evaluate alternative phylogenetic hypotheses. In this procedure,

the observed differences in length between alternative phylogenies were

compared to differences obtained by randomly permuting the original data

matrix 500 times (program courtesy of G. Roderick, University of Hawaii,

Honolulu, HI). In these comparisons, the sequence of the outgroup taxa was not

randomized.

Population analysis

We explored the spatial distribution of mitochondrial variation using

an analysis of molecular variance (AMOVA) (Excoffier et. aI., 1992) and FST

analyses (Hudson et al., 1992). In our AMOVA analysis we used the absolute

number of sequence differences between individuals as our Euclidean

57

distance. Three components of genetic variation were calculated: among

species, among populations within species, and within populations. These

components were considered to be significantly different from random if the

observed value was less than 95% of the values produced by 500 random

permutations of the squared Euclidean distance matrix (program courtesy of

L. Excoffier, University of Geneva, Carouge, Switzerland). FST values were

calculated as (1-Hw/Hb), where Hw is the average of all pairwise difference

among individuals from the same subpopulation and HQ. is the average of all

pairwise differences among individuals from different subpopulations.

Pairwise differences used in this analysis were corrected for multiple hits at

the same nucleotide position using a Jukes-Cantor model weighted for

differences in transition and transversion rates (Lynch and Crease, 1990). To

gauge the significance of the observed FST value it was compared to 500

randomly generated FST values (program available from S. R. Palumbi).

When the observed value was greater than 95% of the randomly generated

FST values, we concluded that there was significant genetic heterogeneity

between regions. The FST approach is based on an explicit island model of

population genetics and thus permits an estimate of migration between

geographic regions based on the relationship, FST =1/(1 + 2Nm), where N is

the effective number of breeding females within each locality and m is the

migration rate per generation.

We also utilized the phylogeographic method outlined in Slatkin and

Maddison (1989) to estimate the minimum number of migration events, §.,

consistent with a given phylogeny. In our data set, there were many equally

most parsimonious trees and §. was determined by averaging values obtained

from a subset of 500 to 1000 of these trees (see Slatkin and Maddison, 1989).

58

The effective migration rate consistent with our estimated value of §. and the

number of individuals examined was estimated using a computer program

provided by M. Slatkin. To determine if this value represented significant

population differentiation, it was compared to values of §. obtained from 500

randomly generated trees using MacClade (Maddison and Maddison, 1992). If

the observed value of §. was greater than 95% of the values in random trees,

then significant population differentiation between sampling locations was

assumed.

59

2.4. RESULTS

The DNA primers used to initiate PCR reactions routinely amplified an

approximately 1200 bp portion of the mitochondrial genome spanning most of

the control-region (figure 2.2). The 3' end of this piece aligned well with other

published vertebrate 115 rRNA sequences. However, of the 195 bp portion of the

5' end of this piece that we sequenced, we could only confidently align two small

segments to the tRNAPro end of control-region of Cichlids (figure 2.2). The

failure of most of our control-region segment to align was not surprising, the

tRNAPro end has been shown to be the most variable section of the control­

region of a variety of vertebrates (Saccone et al., 1991; Meyer et al., 1991; Vigilant

et aI., 1991; Pesole et aI., 1992; Brown et aI., 1993; Wenink et aI., 1993). Cichlids,

which are in the same suborder as butterflyfishes (Percoidei), are more closely

related to Chaetodontids than any other taxa for which the mitochondrial

control-region has been sequenced. Never-the-less, the lineage that gave rise to

the Cichlids and the lineage that gave rise to the Chaetodontids likely diverged

over 80 million years ago (Choat and Bellwood, 1991).

The tRNApro end of the control-region of C. multicinctus, C. pelewensis

and C. punctatofasciatus proved to be hypervariable. One hundred twenty-six of

the 195 nucleotide positions of this A-T rich region were variable, yielding a total

of 130 unique sequences or mitochondrial haplotypes from our sample of 141

individuals. Four of these changes were single base-pair additions and the rest

were single nucleotide substitutions (table 1.1). Genetic differences between

individuals chosen at random, without replacement ranged from 1.6 to over 50%

corrected sequence difference. In these 70 pairwise comparisons, transitions

always out-numbered transversions. However, as many as five transversions

occurred between individuals. To reduce the effects of multiple substitution at

60

the same nucleotide position, we calculated the transition/transversion ratio

using only those randomly chosen pairs in which a single transversion was

observed. This ratio varied from 5:1 to 37:1 in our randomly chosen subsample

and we used the average of these comparisons, 20:1, as our estimation of the

transition/ transversion ratio. Similar high transitional bias has been reported in

the control-region of other vertebrates, and it appears to be a characteristic of

mitochondrial DNA evolution (Aquadro and Greenberg, 1983; Desalle et. al.

1987; Vigilant et al., 1991; Brown et al., 1992).

Cytochrome b/control-region comparison

For 38 of these 141 individuals we had previously sampled a 500 bp

segment of the mitochondrial cyt b gene (Chapter 1). The absolute number of

unique DNA sequences among this subset of individuals was similar in the

two gene regions, with nearly every individual sampled possessing a unique

cyt b and a unique control-region sequence. As a result, diversity measures,

such as the genotype diversity index (Ball et al., 1988), that estimate the

probability that two individuals sampled from a population have different

mitochondrial sequences, were high (~ 0.90) and similar for the two gene

regions (table 1.2). By contrast, absolute levels of sequence differences among

individuals were over an order of magnitude higher in the control-region

than in the cyt b region (table 1.2). For example, average pairwise difference

among these 38 individuals was only 0.85% in the cyt b portion compared to

over 25% within the much smaller control-region segment.

The phylogenetic relationships among individuals inferred from these

two portions of the mitochondrial genome were nearly identical. Both

regions identified three major mitochondrial lineages or clades (figures 2.3a

and 2.3b). In both regions, only clade C was species-specific, containing

61

individuals of the Hawaiian endemic, C. multicinctus. Clade A and B was

composed of individuals from two or more recognized species. For example,

clade B contained multiple C. multicinctus, C. pelewensis and C.

punctatofasciatus. Although these patterns were clear in both the cyt band

control-region data, the higher level of genetic variation within the control­

region greatly extends the confidence that can be placed in these phylogenetic

groupings. In the tree constructed with the control-region sequences, lineages

were identified by at least 3, and as many as 13, substitutions and were

supported in over 70% of bootstrap replications (see Hillis and Bull, 1993).

Trees in which species were constrained into monophyletic groups were

nearly 50 steps longer. In contrast, these same groupings were distinguished

by at most a single substitution in the cyt b segment and were supported in

fewer than 40% of bootstrapped data sets. Topological dependent tests of trees

constructed using the cyt b data supported the polyphyly of these species

(p<O.005)(Faith, 1991); however monophyletic alternatives were only three

steps longer.

Although lineages A, B, and C appeared distinct, the phylogenetic

relationships among them was unclear. The particular branching pattern

depicted in figure 2.3a was not well supported in bootstrapped data sets and

was sensitive to the order in which sequences were added and the program

used to construct trees. As a result, these lineages are best regarded as equally

divergent clusters.

Relative rate of control-region evolution

The rate of evolution within this small segment of the control-region

. was extraordinarily high relative to evolution within the larger cyt b segment.

Plotting the difference among pairs of individuals in the control-region

62

against the differences among the same individuals within the cyt b gene

reveals the extraordinary rate at which change accumulates within this small

segment of the control-region (figure 2.4). Genetic differences in the control­

region segment rose rapidly relative to changes in the cyt b gene region but

leveled off with increasing divergence suggesting that, despite distance

corrections, multiple substitutions at the same nucleotide position obscured

the true number of substitutions between distantly related pairs. Because

multiple hits were less likely to affect more closely related pairs we used these

pairs to estimate the relationship between change in the control-region and

change in the cyt b gene. For example, those individuals that differed by only

0.2% of their cyt b sequence (1 change out of 500 sampled) differed by, on

average, 8.1% of their control-region sequence. Similarly, those individuals

that differed by 0.4% of their cyt b sequence differed by 9.8% in the control­

region, suggesting a relative control-region/cyt b rate of between 25:1 and 40:1.

None-the-less, we used the average of these values, 33:1, as a very crude

approximation of the relative rate of change between the two regions.

Assuming that the cyt b gene in butterflyfishes evolves between the 1.0% per

million years estimated in sharks and 2.5% per million years estimated in

mammals (Martin et al., 1992; Irwin et al., 1991), we estimate a rate of

evolution in this small segment of the control-region of between 33% and

82.5% per million years. This range was considerably higher than the 11.5­

20% reported in other vertebrates (Vigilant et al., 1991; Brown et al., 1993).

However, given the heterogeneity in rates across the different domains of the

control-region, a difference in the rate of evolution of this magnitude was not

completely unexpected (Saccone et al., 1987; Saccone et al., 1991). For example,

Ward et al., (1991) suggested that the tRNAPro end of the human control-

63

region (hypervariable domain I) evolves at nearly 30% per million years,

whereas the tRNAphe hypervariable region (domain II) evolves at a rate

closer to 15% per million years. Rate heterogeneity over smaller spatial scales

was clearly evident within the control-region of Cichlids where much of the

variation in a 442 bp segment of the tRNAPro end of the control-region was

clustered in a small (150 bp) central domain (see figure 2.3, Sturmbauer and

Meyer, 1992). The 3' end of the 195 bp portion of the control-region that we

sequenced aligned well with a conservative area just 3' from this

hypervariable domain.

Phylogenetic relationships among the total sample of individuals

Extracting the phylogenetic relationships from the larger sample of

individuals was complicated by multiple substitutions at the same nucleotide

position. Among a subset of the many thousands of most parsimonious

trees, there were on average 3.1 changes per position along this short control­

region segment. Indeed, it was not uncommon to have positions change in

excess of 10 times along shortest length phylogenetic trees (figure 2.5). As in

the human control-region, the observed number of base substitutions fit a

negative binomial much better than a Poisson model of evolution suggesting

significant constraints on base substitutions within this region (table 1.3)

(Kocher and Wilson, 1991; Wakeley, 1993). Many of the invariant positions

occurred singularly or in small blocks of 2-4 nucleic acids; however, there was

a region of approximately 20 nucleic acids beginning at position 150 that was

highly conserved. This region aligned well with the Cichlid control-region

sequence (figure 2.2). The other small region that we could confidently

aligned with the cichlid control-region sequence (l04-127) also showed a high

degree of base conservation. The distribution of transitional substitutions

64

and deletions were cluster in 3-5 bp units along this 195 bp region (figure 2.5),

Nearly all transversions occurred at positions inferred to have had more than

5 substitutions, possibly reflecting mutational hot-spots along this sequence.

Despite this extensive level of superimposed change, the three major

phylogenetic divisions identified earlier were reasonably well preserved in

the larger sample of 137 individuals (figures 2.6a-d. Both Clade A and C

continued to be strongly supported in bootstrapped replications of Neighbor­

joining trees. Clade B, which contained over 57% of the individuals sampled,

however, was supported in fewer than 30% of these replications. In this case,

multiple changes at the same nucleotide position probably obscured the

phylogenetic signal. Average pairwise difference among individuals within

this group was over 11 %, and the two most divergent individuals differed by

over 38% (table 1.4). In addition, after eliminating the 73 invariant sites

identified from the examination of the entire data set, a nucleotide position

changed, on average, 2.8 times along most parsimonious arrangements of

individuals within this clade. Moreover, the distribution of positions with 0,

1,2,3, ., ., " n substitutions continued to depart from a random Poisson model

of change (not pictured). This pattern suggested more complicated constraints

on the nucleotide substitution in this region than expected from a simple

two-rate Poisson mixture model (Wakeley, 1993).

Support for the evolutionary distinctiveness of clade B, however, was

evident from the pattern of pairwise differences among individuals in this

lineage and individuals of the other two lineages. The frequency distribution

of these comparisons were notably bimodal suggesting the presence of two

distinct clades (Hudson and Slatkin, 1991). Comparisons of the absolute

number substitutions within and between individuals of clade A and B

65

showed two distinct peaks: a larger one at 15 substitutions and a smaller one

at 27 substitutions (figure 2.7). The larger peak at 15 substitutions represented

comparison among individuals within each clade which were distinctly

unimodal (figure 2.8). The peak at 27 substitutions represented between clade

comparisons. This pattern is also reflected in the much smaller average

within-clade genetic differences, which ranged from 1.8% to 11.1%, relative to

between-clade differences, which ranged from approximately 30% to over 45%

(table 1.4).

Relationships among individuals within lineages

A remarkable difference in the phylogenetic patterning among these 3

lineages was the five-fold reduction of variation within lineage C relative to

lineages A and B (table 1.4). As a result, the branch leading to this clade was

over three times longer than the branch leading to either clade A or B (figures

2.6a-c). However, the phylogenetic patterns within each of the three mtDNA

lineages showed a similar lack of phylogenetic structure. The more basal

nodes were short relative to the more terminal branches giving the

phylogeny of these lineages a distinct star-like pattern with little internal

clade structure (figures 2.6a-c). This lack of phylogenetic structure was best

visualized by the unimodal and Poisson-like distribution of pairwise

comparisons among individuals within each lineage (figures 2.8) (Ball et al.,

1990; Slatkin and Hudson, 1991).

Partitioning of genetic variation

Clade C was the only one of the three major lineages that showed any

striking geographic or species specificity. This clade was only found within

the geographic range of C. multicinctus (figure 2.6c). Nearly all (85%) of the 32

C. multicinctus examined fell within this shallow lineage, making this

66

species the most genetically distinct of the three. The remaining 5 C.

multicinctus fell along three terminal branches within clade B (figures 2.6b-c).

In contrast, individuals in clade A and B were scattered throughout the

Pacific and showed no clear species specificity (table 1.5). It was common for

individuals of different species from widely scattered locations across the

Pacific to possess very similar control-region sequences. For example, within

clade A, our Neighbor-joining tree contained 7 terminal branch pairs

composed of a C. punctatofasciatus and a C. pelewensis (the "*" in figure 2.6a).

In 2 of these 7 pairs, individuals were collected from Tahiti and the

Philippines, roughly 10,000 kilometers apart. A similar pattern is evident

within the large, highly diverse lineage B (figure 2.6b). In this case,

individuals collected as far apart as Hawaii and the Philippines had identical

control-region sequences. .

At the species level, mitochondrial variation was distributed randomly

between C. pelewensis and C. punctatofasciatus. Neither MANOVA

(variance among species =8.98%, p=0.1782), FST (FST=0.124, p=0.28), chi­

squared [(chi-squared=1.514, p=0.212)(see Roff and Bentzen, 1989)], nor

phylogenetic approaches (s=34; p>0.05 ) could detect differences in distribution

of mtDNA variation between these two species. Despite this observation, .

mtDNA variation was not distributed randomly across the geographic range

of each species. A significant amount of the genetic variation between these

two species, nearly 24%, was partitioned among populations (MANOVA;

p=O.Ol). In particular, the proportion of lineages A and B within the seven

western and south Pacific populations examined differed significantly from

random (chi-square=13.8; p=0.035). The distribution of clade A and Bin

populations across the south and western Pacific suggests a broad cline, with

67

clade B dominating populations of the Western Pacific and clade A found in

higher frequencies within populations of the eastern South Pacific (figure 2.9).

The significant differences in the distribution of mtDNA variation among

populations of these two species could be attributed largely to the higher than

expected portion of clade B in the Tahitian samples and clade A in the

Indonesian samples. When these two peripheral populations were

eliminated there was little evidence for genetic structuring of populations

across the range of C. pelewensis and C. punctatofasciatus (chi square=4.72;

p=O.325).

68

2.5. DISCUSSION

Assuming that most variation in the mitochondrial genome is neutral,

then the pattern of branching events provides a powerful glimpse into the

demographic history of a species or group of closely related species (Tajima, 1983;

Hudson, 1990; Slatkin and Hudson, 1991; Rogers and Harpending, 1992;

Harpending et al., 1993). For these butterflyfishes, the mtDNA genealogy traces a

rich history of demographic change. The early history of this group was

apparently marked by the coalescence of mtDNA diversity into three major

mtDNA lineages. However, following the coalescence of mtDNA variation into

these lineages there appears to have been a shift in the demographic landscape of

these species. The Poisson-like pattern of pairwise differences among

individuals within each of the three major lineages argues for a recent period in

which mtDNA variation was effectively buffered from extinction (figure 2.7)

(Slatkin and Hudson, 1991; Rogers and Harpending, 1992; Harpending et al.,

1993). Although factors such as a purifying selection or extreme rate

heterogeneity [(k<O.l, if the underlying distribution of mutations in the control­

region fits a negative binomial (Rogers, 1992)] can generate a similar unimodal

peak in the pairwise differences among extant individuals, this genealogical

pattern is most consistent with a wave of population expansion following a

population bottleneck (Di Rienzo and Wilson, 1991; Rogers and Harpending,

1992; Marjoram and Donnelly, 1994). In clade A and B, the high and similar

intra-clade levels of variation suggest that a period of rapid population growth

followed soon after the initial formation of these lineages. In this case,

population expansion resulted in the widespread geographic distribution of both

lineages. Clade C, likewise, shows a Poisson-like peak in the pairwise

comparisons among extant lineages (figure 2.7). However, the levels of within-

69

lineage variation are notably smaller (figure 2.6c). This lineage is the most

geographically restricted of the three and it is possible that this pattern reflects a

recovery from a recent genetic bottleneck that was not linked to the initial

formation of this lineage. Alternatively, the shallowness of this clade may

indicate that a more significant population bottleneck was associated with the

formation of this lineage.

Given the demographic history suggested by the control-region

sequences, the lack of a strong concordance between the three major mtDNA

clades and species boundaries is intriguing. Either the formation of the

species was completely decoupled from these demographic changes or

speciation occurred within this demographic backdrop and hybridization and

introgression has since eroded the initial concordance between mtDNA

lineages and species designations. These alternative scenarios paint

contrasting pictures for the recent evolutionary history of these species,

differing both with respect to the timing and demographic parameters

associated with differentiation.

Two scenarios and their evolutionary and demographic implications

Shared evolutionary history

Under certain demographic conditions, species will share ancestral

genetic variation for many generations after the origin of reproductive

isolation (Neigel and Avise, 1986; Pal milo and Nei, 1988). If the levels of

mtDNA variation are any indication, these three species have arisen recently.

Slight genetic differences among individuals in the cyt b gene region suggest a

maximum time-frame of divergence of between 300,000 and 800,000 years

(Chapter 1). Species maintaining an effective population size of only 150,000­

400,000 females through speciation would be expected to share ancestral

70

lineages over this time period (Neigel and Avise, 1986). Given the

demographic properties of these species, which include large population sizes,

high fecundity, and vast dispersal potential due to planktonic larval stages

(Reese, 1991), it is possible that the observed patterning of mtDNA variation

among these species reflects ancestral patterns of gene flow among groups

that are no longer exchanging individuals.

If this scenario is correct, then the widespread distribution of similar

control-region sequences implies that speciation has occurred very recently,

well after the formation of clades A, B, and C. In this case genetic differences

between individuals of different species falling together at the terminal

branch tips of the mtDNA genealogy, represented by the >I- in figure 2.6a, are

the most likely to reveal the timing of speciation. In our Neighbor-joining

tree, these differences range from 0% to 10.34% with an average of 3.1%,

placing the origin of these species, based on our crude estimation of the rate of

molecular evolution within the last 37,000-94,000 years.

This scenario makes similar strong predictions about the demographic

process associated with species formation. The haploid, maternally inherited

mitochondrial genome is four times more sensitive to changes in population

size than are neutral nuclear loci (Wilson et al., 1985). Thus, the presence of

shared mitochondrial variation suggests that the differentiation of this group

proceeded without a severe, protracted population bottleneck. Presumably,

the color pattern differences that presently define these species consolidated

within this demographic framework. This scenario would be more plausible

if there were strong selective pressure on the loci coding for color pattern

during speciation. Although the evolutionary or ecological significance of

vivid coloration in butterflyfishes is unknown, color pattern may play an

71

important social or sexual role (Reese, 1975; Fisher, 1980; however see Ehrlich

et al., 1977). Because socially and sexually important traits can be subject to

strong directional selection during speciation they can change very rapidly

and coalesce quickly within emerging species (Lande, 1982; West-Eberhard,

1983). In contrast, neutral markers will be shared between newly formed

species until the time since the onset of reproductive isolation approaches 2-4

times the historical effective population size, or historical female population

size for mtDNA genes (Neigel and Avise, 1986; Palmilo and Nei, 1988).

Hybridization

Alternatively, the lack of clustering of species within the three mtDNA

lineages may be caused by recent mitochondrial gene flow. Under this

scenario, species defining characteristics (i.e., color pattern) may have formed

within the same demographic framework that caused the coalescence of

mtDNA variation into three lineages. In this case, the timing of speciation

can best be approximated from the average corrected genetic differences

between the three mtDNA lineages. This average, 26.7%, places speciation

between 325,000 and 800,0000 years ago or nearly an order of magnitude more

distance than the temporal framework suggested by the shared evolutionary

history scenario.

Subsequently, mitochondrial gene flow between species may have

eroded the initial mtDNA gene and species concordance. The species and

geographic specificity of clade C argues that this lineage was the original

lineage of C. multicinctus. Thus, under this scenario, the number of

individuals with mtDNA haplotypes that do not fall within this clade reflect

the levels of hybridization. Within our sample of 32 C. multicinctus, there are

5 individuals that posses haplotypes outside of clade C. Two pairs have

72

identical control-region sequences and probably descended from a single

hybridization event, yielding support for a minimum of three hybridization

events in the history of the individuals sampled. Given our sample size, this

phylogenetic pattern suggests an effective migration rate, Nm, between C.

multicinctus and either C. pelewensis or C. punctatofasciatus of

approximately 0.4 individuals per generation (see Slatkin and Maddison,

1989). A similar value, not dependent on a phylogenetic tree, was suggested

from the patterning of variation within and between species (FST=0.5227,

Nm=0.41) (Hudson et al., 1993).

The widespread mixing of similar control-region sequences between C.

punctatofasciatus and C. pelewensis suggests extraordinary high levels of

hybridization and associated introgression of mtDNA genes. Within our

Neighbor-joining tree there were 21 instances of a C. pelewensis and a C.

punctatofasciatus falling together at the terminal branch tips, see " in figure

2.6a. Across this entire phylogeny, there was phylogenetic evidence for a

minimum of 34 separate hybridization events. Identical estimates were

obtained from maximum parsimony trees. This genealogy suggests an

effective migration rate between these two species of nearly 300 individuals

per generation. A slightly greater estimate was obtained from the observed

FST value between species (FST=0.0007; Nm=713). Although both estimates

are likely biased upward, if hybridization accounts for the present distribution

of mtDNA variation between these two species, then the present pattern of

mtDNA variation highlights a history of substantial movement of mtDNA

genes between species boundaries (Hudson et al., 1993).

Because the mitochondrial genome does not code for genes likely to be

directly involved in reproductive isolation, this cytoplasmic marker has

73

frequently been observed to introgress across species boundaries at a rate

higher than nuclear encoded loci (Powell, 1983; Ferris et al., 1983; Spolsky and

Uzzell, 1984, 1986; Lamb and Avise, 1985; Baker et aI., 1989; Dowling, 1989;

however see Szymura et al., 1985; Szymura and Barton, 1986; Nelson et al.,

1987). Theoretical considerations suggest that even with very strong selection

against hybrids, the mitochondrial genome can introgress across species

boundaries at a rate determined principally by the frequency of hybridization

and the dispersal potential of the two parental species (Takahata and Slatkin,

1984). Under this scenario, the present patterning of mtDNA variation

suggests that the frequency of hybridization among these species can be quite

high. In addition, these fishes have a planktonic larval stage of over 40 days

(Tricus, 1987; Hourigan and Reese, 1987; Leis, 1989), indicating that

widespread dispersal and gene flow in these species may be qui te common. In

marine populations, a number of population genetic studies have

documented genetic homogeneity over vast areas presumably due to long

distance dispersal of planktonic larval stages (reviewed in Palumbi, 1992).

Distinguishing ancestry from hybridization

When the time since speciation is short, discriminating between the

retention of ancestral polymorphism and hybridization as alternative

explanations for the presence of shared gene lineages is notoriously difficult

(Slatkin, 1989). Despite this conclusion several features about the genealogical

and geographic patterning of mtDNA variation among these three

butterflyfishes lead us to support hybridization as the most likely explanation

for the observed discrepancy between the mtDN A tree and species

boundaries. However, it is important to note that none of these features by

themselves provide compelling support for either scenario. In all cases, a

74

similar pattern could, in theory, be attributed to the retention of ancestral

variation in very recently evolved species. Yet, in each case, the observed

genetic and geographical patterning of mtDNA variation can be more easily

explained by recent hybridization than by the retention of ancestral

polymorphism.

Branch tips and divergence

If shared history accounts for the widespread distribution of similar

mtDNA haplotypes among species, one would expect that any evolution

occurring after the onset of reproductive isolation would accumulate at the

branch tips first. For example, a panmictic ancestral population divided into

two emerging species will initially share many gene lineages and closely

related haplotypes. However, over time, new mutations will cause these

closely related haplotypes and gene lineages to diverge. At the same time,

mutation will create a new generation of very similar, but not identical

haplotypes. This new generation of similar haplotypes will only be found in

individuals of the same species. Thus, assuming that any time at all has

elapsed since speciation, divergence among species will be most evident at the

terminal branches of a gene genealogy. Specifically, one might expect that

branch tips containing individuals of the same species would be significantly

shorter than branch tips containing individuals of different species. In our

Neighbor-joining tree there are 19 cases in which individuals of different

species fell together at the terminal branches, represented by the * in figure

2.6a, and 14 instances in which conspecific individuals fell together, represent

by the ** in this figure. Figure 2.10 plots the frequency distribution of genetic

differences between individuals in these two types of terminal nodes.

Although the mean genetic difference is slightly higher in the heterospecific

75

versus conspecific comparisons, 3.1% versus 2.4%, we can detect no branch tip

divergence among these species (t=0.922; p=0.36). The shared ancestry model

is compatible with this observation only if speciation has occurred so recently

that no control-region divergence has occurred. By contrast, recent and high

levels of hybridization easily explains the lack of divergence between

individuals of the different species that fall together at the terminal tips of

our mtDNA genealogy.

Geographic patterning of mtDNA variation:

Additionally, the random distribution of mtDNA variation between C.

pelewensis and C. punctatofasciatus at the species level but not at the

population levels seems curious under the shared history scenario. The

shared history scenario requires that the observed cline in mtDNA haplotypes

predated speciation and that species boundaries consolidated around it. This

might be reasonable if there is strong differential selection on these mtDNA

lineages in different parts of the Pacific (see Endler, 1977). However, there

does not appear to be any strong association between these lineages and

obvious environmental or physical characteristics. In fact, differences in the

distributions of these lineages among populations are so minor, with the

proportion of lineage B never rising above 67%, as to almost preclude any

selective explanation. By contrast, if one assumes that mtDNA variation is

neutral, then any underlying structuring of mtDNA haplotypes would likely

be accentuated during speciation (Neigel and Avise, 1986). Thus, one would

expect genetic differences between species to build faster than genetic

differences between populations.

The present geographical distribution of mtDNA variation between

species, however, is easily incorporated into the hybridization scenario.

76

Under this scenario, lineage B, the dominant lineage in the Western Pacific,

was the original lineage of C. punctatofasciatus and lineage A, the dominant

lineage within the eastern South Pacific, was the original mtDNA lineage of

C. pelewensis. It is possible that lineages coalesced at speciation and were

initially restricted geographically in the central Malay region and in the

archipelagoes of the South Pacific, respectively. However, in the 325,000­

800,0000 years since speciation, the high fecundity and wide dispersal

capabilities of these species have permitted each lineage to expand its

geographic range. To date, lineage B has been able to spread across the entire

range of C. pelewensis. Lineage A has done nearly as well, spreading as far

east as the Philippines (figure 2.9). Under the hybridization scenario, the non­

random distribution of mtDNA among populations of these two species

merely suggests that expansion has not completely eroded the historical

pattern. Indeed, it is populations at the endpoints of each species range, the

Tahitian and Indonesian populations, that generate the observed genetic

differences among populations.

Genealogical patterning of mtDNA variation

Lastly, the presence of three distinct species and three distinct mtDNA

lineages is difficult to reconcile under the shared history scenario. This

concordance may have arisen by chance; however, the strong geographic and

species specificity of clade C, as well as, the non-random distribution of clade

A and B detailed above argue against this possibility. It seems more probable

that phenotypic and mtDNA evolution were contemporary and that the

species-specific color patterns formed within a demographic framework in

which populations were sufficiently small to cause the coalescence of

mtDNA variation.

77

Demographic changes, speciation, and hybridization

Interpreted as a whole, the patterning of mtDNA variation begins to

build a substantial case for hybridization as the most plausible explanation for

the widespread distribution of similar mtDNA sequences among these

species. This framework suggests that following phenotypic and mtDNA

differentiation, the enormous fecundity and the high dispersal potential

linked to long-lived planktonic larval stages allowed newly formed species to

expand population size and extend their geographic range. In the case of C.

pelewensis and C. punctatofasciatus, the newly emergent mtDNA lineages

and phenotypic lineages met. Reproductive barriers were apparently

insufficient to prevent these species from hybridizing and there appears to

have been only slight impediments to the geographic expansions of the

mitochondrial lineages of these two species.

In the presence of such pervasive mixing of mtDNA variation between

C. pelewensis and C. punctatofasciatus the geographic and species-specificity

of lineage C within C. multicinctus. is notable. On the one hand, this pattern

may merely reflect the extreme geographic isolation of this species. The

Hawaiian Islands (and Johnson Atoll approximately 1300 kilometers to the

southwest) are the most isolated group of islands within the Pacific. The

nearest known source population for C. punctatofasciatus lies in the Marshall

Islands nearly 2300 kilometers to the southwest and the source population of

C. pelewensis is nearly twice as far to the south. Thus, it is possible that few

individuals of C. punctatofasciatus or C. pelewensis ever reach Hawaii,

thereby limiting the possibility for hybridization. Previous allozyme stu.dies

of marine populations across this region have revealed slight but significant

genetic differences, generally associated with lower levels of heterozygozity,

78

between Hawaiian and other Pacific populations (Winans, 1980; Nishida and

Lucas, 1988). In addition, mitochondrial sequences of populations of

Hawaiian sea urchins show a similar pattern of reduced genetic diversity

relative to populations nearer the western Pacific (Palumbi and Metz, 1991;

Palumbi, 1994).

On the other hand, the genetic isolation of the Hawaiian endemic, C.

multicinctus, might be attributed to a greater degree of reproductive isolation.

Chaetodon multicinctus is the largest (Appendix C) and most phenotypically

distinct of the three species (see figure 2.1). Although the underlying body stripes

are similar to C. punctatofasciatus, C. multicinctus is almost totally white, lacking

the striking yellow and orange coloration that characterizes C. punctatofasciatus

and C. pelewensis. Chaetodon punctatofasciatus and C. pelewensis are more

similar to one another in size, differing primarily in the striations along the body

which run vertically in C. punctatofasciatus but diagonally in C. pelewensis

(figure 2.1). These color patterns change abruptly at the suture zone between

species (Appendix C). Thus, levels of hybridization and gene flow between

species capable of generating a nearly random pattern of mtDNA variation have

failed to erode species-defining color patterns. The contrasting patterns of

mtDNA and phenotypic variation suggest that strong selection is acting to

maintain color pattern differences. Positive assortative mating among species,

similar to that observed among different colored morphospecies of Western

Atlantic hamlets (Hypoplectrus: Serranidae) (Fisher, 1980), may account for the

sharp demarcation of phenotypic differences among these species. Future work

on the degree of reproductive isolation among species, as well as on the

patterning of variation at other genetic markers, will help interpret the cohesion

79

of phenotypic variation in these species within a background of potentially

homogenizing levels of gene flow.

80

Table 2.1: Pattern of nucleotide substitution in a 195 base portion of the

tRNAPro end of the mitochondrial control-region of Butterflyfishes.

NucleotideNucleotide Proportion Introduced'[Replaced of Sequencet G A T C Del

G 0.138(0.103-0.179) 82 3 2 0

A 0.396(0.349-0.426) 157 - 6 4 0

T 0.258(0.103-0.179) 0 2 116 0

C 0.180(0.221-0.282) 6 1 106 0

Insertion 0.021(0.153-0.021) 1 2 1 0t Averaged over a11141 sequences. Numbers in parenthesizes are the

minimum and maximum percentages observed.t Average number of unambiguous changes from a suite of 500 equally most

parsimonious trees generated from a heuristic search using the phylogenticprogram PAUP (Swofford, 1993). For this search transversions wereweighted 10 times transitions.

81

Table 2.2: Comparison of within and between species mtDNA variation

using a 500 bp segment of the cyt b gene and a 195 bp segment of the tRNAPro

end of the mitochondrial control-region.

GenotypeNumber of types diversity indext % seq. difference'[

CYfB CONTROl... CYfB CONfROL CYfB CONTROL

All individuals (n=38) 30 34 0.97 0.98 0.85 25.40

C. multicinctus (n=15) 10 12 0.90 0.97 0.67 13.9

C. punctatofasciatus (n=12) 11 12 0.98 1.00 0.99 17.4

C. pelewensis (n=l1) 9 11 0.97 1.00 0.71 18.5

t Genotype diversity is give by h = (1-~xi2)n/(n-1), whereXi is the frequency of the ith type and n is the sample size.:j: percentage sequence difference is the average pairwise divergence amongindividuals. For the cyt b data set, genetic differences between individualswere calculated using the Kimura 2-parameter model (Kimura, 1980). For thecontrol-region, distances among individuals were corrected using a Tamura­Nei model with a gamma correction (a=0.6) (Kumar et al., 1993).

82

Table 2.3: Observed and expected number of base substitutions under both a

Poisson and negative binomial model of nucleotide substitution in the

tRNAPro end of the mitochondrial control-region of butterflyfishes.

Estimated NUMBER OF SITESnumber of changes

per sitet Observed Poisson Negative binomialo 69 8.17 63.271 28 25.70 31.892 16 40.44 21.433 16 42.42 15.604 4 33.37 11.795 14 21.01 9.116 11 11.02 7.15

>7 32 7.88 29.57t The number of changes at each of the 195 sites was inferred from arandomly chosen most parsimonious tree generated during a heuristic searchusing the phylogenetic program PAUP (Swofford, 1992). The mean numberof changes per site along this tree (m=3.15) was used to compute the expectednumber under both the Poisson and negative binomial distribution (k=O.6).The negative binomial fit the observed distribution of changes per site farbetter than a Poisson model (chi-squared=12.426, p=O.09 and chi­squared=586.616, p<O.0001, respectively).

83

Table 2.4: Average percent difference of all pairwise comparisons of genetic

difference between individuals within (along diagonal), between uncorrected

(below diagonal) and between corrected (above diagonal) for the three

mitochondrial lineages identified in this study. Corrected differences were

calculated by subtracting the average within-clade variation from the average

between clade variation (Wilson et al., 1985). Genetic distances between

individuals were corrected for multiple hits at the same nucleotide position

using the Tamura and Nei distance method with a gamma correction (a=0.6)

(Kumar et al., 1993).

A B C

Clade A 9.3 17.9 21.7

Clade B 28.1 11.1 40.5

CladeC 46.1 34.5 1.8

84

Table 2.5: Distribution of individuals within each of the three major mtDNA

lineages from populations across the Pacific.

Numberof Individuals Clade A Clade B CladeC

C. multicinctusMidway Island 11 0 2 9

Oahu Island 15 0 3 12Johnson Island 6 0 0 6

C. punctatofasciatusGuam 8 2 6 0Palau 20 3 17 0

Philippines 21 8 18 0Indonesia 7 0 7 0

C. pelewensisTahiti 9 6 3 0

Am. Samoa 2 0 2 0Cook Islands 18 7 11 0

Fiji 21 1 17 QTotals 138 81 30 27

85

Figure 2.1. Color plate of a) Chaetodon punctatofasciatus, b) C. pelewensis,

and c) C. multicinctus.

86

a)

b)

c)

&l

Figure 2.2: peR primers used to amplify an approximately 1200 bp portion of the mitochondrial

genome spanning the control-region of butterflyfishes. The 3' end of this 1200 bp piece aligned well

with the 12s rRNA gene of other vertebrates. However, we were only able to align two small regions of

this piece with the tRNApro end of the control-region of Cichlids. Alignment of the Chaetodon

multicinctus sequence to both Cyathopharynx furcifer (Pisces: Cichlidae) and Crossostoma lacustre

(Pisces: Homalopteridae) sequence was done by eye. Numbers at the beginning of each sequence are the

position of the aligned regions relative to the 5' end of the 12s rRNA gene of C.lacustre and the

tRNApro end of the control-region of C. furcifer (see Tzeng, et al, 1992;Sturmbauer and Meyer, 1993).

Chaetodon mtDNA

tRNA-Pro[S, ~~C.R.-l iIRNA.Phe :;lIRNA-lhr 1 ~ 72J12sar

00\0

Control region:

ryathopl1drynx t urc i fer

Cbeet.cdon multicinctus

Cyat!lcpharynx turc i t or

Ctiee t ccton mv t c i c i nct us:

125 rRNA:crcsac sc cr-s Lacus r r-e

Ctiaet cdcn multicinctus

cros.scs r cr-e lacustre

Chaet cdcn mul r f c I nc r us

5' (111) 21 -CATA.AAGC-TAAGGGGTACAT- - - -AAACCATAGCTGTAATATTCAACTAACTATTTACTGAAAGCTAAACGATAGTTT-AAGACCGATCACACCTCTCACATAGTT

II 1111 I II I I11111 II III I I I II III III I II II I III II I I I I II I I5 ' - TGT ATTAGAT AAGAAG - ACTT f.CT AAAACC AAAGTAAT AAGAAGACCTTGA - TACTTATTGAGTAATGM - GAAAC1"TCTAACTeAGCT AAA TTTCATCCCGTCA

OJ' -AAGATATACCAAGTACCCACCATCCT-AT7MTTAA-TJ..ATA-TTTAATGTAGTAAGAGCCCACCATCAGTTGATTTC7CAATGATAACGGTTCTTGA

111111111111111111111 I I I III I II 11111 II 111111 111111 III 1111111111111 11115' - AAGACATACCAAGTACC-ACCATTTCTGTATAGGAAATAACl\ACCCM- - - - -TAAGAACCTACCATCGGTTGATATCTGAATGATAACGGTTATTGA

5' II 221) -AGGGTCGGTAAAACTCGTGCCAGCCACCGCGGTTATACGAr.AGACTCAAGTTGACAGTCACTCGGCGTAAAGT- -GGTTAAGATGAACMAAACTAAAGCCAAAT- r.CCTTCAAA

111111111111111111111111111111111111111111 I I 111111 II 1111111111 1111111 1111I1 II 111111111 11115 ' - AGGGCCGGTAAAACTCGTGCCAGCCACCGCGGTT1\T ACGAGAGGCCCT AGTTGATAGGTG - - CGGCGTAAAGGGTGGTTAAGGAGAGCAAGAAITAAAGCT AAGGGM"( "reTT· G

S' -ACTGTTATACGTTCTCGMGGTGAGAAGCCCAA

I 11 III I I I II I 11II1 IIS· -GCCGTCATACGCTTCTGAGTATC~AAAGCCMA

Figure 2.3a: Neighbor-joining tree for 38 individuals from three species of

butterflyfishes [Chaetodon multicinctus (n=15), c. pelewensis (nel l) and C.

punctatofasciatus (n=16)] inferred from substitutions within a 195 base

portion of the 5'-end of the mitochondrial control region. This tree was

constructed from pairwise distance matrices corrected for multiple

substitutions using a Tarnura-Nei distance with a gamma correction (a = 0.6)

(Kumar et aI., 1993). The relative strength of major nodes within this

phylogeny is given by the percentage agreement in a consensus of trees

produced from 100 bootstrap simulations of the data set (above each branch)

and the number of informative nucleotide changes (above each branch).

Chaetodon guttatissimus was used as an outgroup to root changes along this

tree. Trees constructed using parsimony with a 20:1, 10:1, and 1:1

transition/ transversion bias were nearly identical, differing only in the

position of individuals within each group. Consistency and retention indexes

for these trees generated under a 10:1 weighting scheme were 0.49 and 0.79 .

Collection localities are as follows: Ha- Oahu, Hawaii; Phil- Philippines;

Bali- Bali, Indonesia; AS- Pago Pago, American Samoa; TA- Moorea, Society

Islands.

90

B

c

13100

..------- c. pun Guam 4r---- C. multi Ha 2,Ha 9, C. pun Phil. 31..-_ C. pele AS 2

C. multi Ha 11C. pun Guam 3a

C. pun Phil. 4C. peleTH8

C. pun Guam 2C. pun Bali 1

1"---- C. pun Guam 4a'---- C. pele AS 1

IIL--- C. pele TH 2'-tL.----- C. pele TH 1

C. pun Bali 3C. pun Bali 5

C. pun Bali 4

C. multi Ha 1, Ha 6C. muitiHa5

C. multi HasC. multi Ha 7

C. multi Ha 10

C. multi Ha 12

C. multi Ha 4, Ha 14C. multi Ha 15

C. multi Ha 3C. multi Ha 13

70

5

4

62

a)

84

3

iii iii

C. pele TH5C. peleTH3

C. peleTH9..------ C. pele TH 4

C. peleTH7C. pun Guam 3

C. peleTH6C. pun Phil. 2

A

a 5% sequence difference

91

Figure 2.3b: Neighbor-joining tree for 38 individuals from three species of

butterflyfishes [Chaetodon multicinctus (n=15), c. pelewensis (n=l1) and C.

punctatofasciatus (n=16)] inferred from substitutions a 500 base portion of the

cyt B gene region (Note: this tree is not drawn to the same scale as 3.3a). This

phylogeny was constructed from pairwise distance matrices corrected for

multiple substitutions at the same nucleotide position using a Kimura two­

parameter model (Kimura, 1980). The relative strength of major nodes

within this phylogeny is given by the percentage agreement in a consensus of.trees produced from 100 bootstrap simulations of the data set (above each

branch) and the number of informative nucleotide changes (above each

branch). Chaetodon guttatissimus was used as an outgroup to root changes

along this trees. Trees constructed using parsimony were nearly identical,

differing only in the position of individuals within each group. Consistency

and retention indexes for these trees were 0.85 and 0.85. Collection localities

are as follows: Ha- Oahu, Hawaii; Phil- Philippines; Bali- Bali, Indonesia;

AS- Pago Pago, American Samoa; TA- Moorea, Society Islands.

92

B

5

c

A

r-------- C.pun Guam 4

C. multi Ha 2,Ha 9

C. peleAS 2

C. pun Phil 3

C. pun Guam 3a

C. pun Guam 4a

C. multi Ha 11; C. pun Bali 3, Bali 1; C. pele TH 8, AS 1

C. peleTH2

c.punGuam2

C. pun Bali5

C. pun Phil 4

C. pun Bali 4

C. peleTH 1

C. multi Ha 1C. multi Ha 3, Ha 5,Ha lO,Ha 14

C. multi Ha4

C. multi Ha 7

C. multi Ha 8

% sequence difference

40 C. multi Ha 121 C. multi Ha 6

C. multi Ha 13

1..---C. pun Guam 3

C. pele TH 3, TH 6

C. peleTH4

C. peleTH5C. peleTH9

C. punPhi12 ~Ac. peleTH7 ~

I I

21

51

o

b)

93

......."0(l) 20.....u(l)

'"' 18'"'0u

"-"

l:: 16 .~ ..0 •.-~ 14'"'- •8 12

f t.....l:: •8 10 0C1J 0u 8l::(l)

'"' 6 4'~.-"0 4 .(l)

:g ......

20(l)

U::s 0l::~ 0 .2 .4 .6 .8 1.0 1.2 1.4 1.6 1.8 2.0 2.2

% nucleotide difference, cytochrome b (uncorrected)

Figure 2.4: Mean percentage nucleotide difference with one standard

deviation error bars in a 195 base-pair portion of the tRNAPro end of the

mtDNA control region relative to changes in a 500 bp segment of the

mitochondrial cyt B gene for all pairwise comparisons of individuals within

clade B (see figure 2.2a). Percent differences within the control region were

corrected for multiple hits at the same nucleotide position using the Tamura­

Nei distance method with a gamma correction (a=0.6).

94

Figure 2.5: Average number of substitutions per position along a 195 bp portion of the tRNApro end of

the control-region estimated from 500 equally most parsimonious trees generated from a single

heuristic search on the computer program PAUP (Swofford, 1991). These trees were constructed

assuming that transitions were 10 times more likely to occur than transversions. The two large shaded

bars represent the two regions that aligned well with the control-region of the Cichlid, Cyathopharynx

\0(Jl

furcifer (Pisces: Cichlidae) (see figure 1). White bars represent those positions where trans versions

occur. A "d" represents a position where a sequence addition was observed..

suouruusqns JO # a~huaAV

Q)us::Q)

~Q)ens::o'SoQ)

""'.-4e....s::oubes::o

.-4!tl

s::o..............enoc,

o-\J"')......oC"l

\J"')N

96

Figure 2.6: Neighbor-joining cartoon of the three major mtDNA lineages.

The size of each shaded box indicates the average pairwise genetic distances

(width) and the number of individuals (height) within each. Support for

each clade is given by the percentage agreement in a consensus of trees

produced from 100 bootstrap simulations of the data set. Chaetodon

guttatissimus was used as an outgroup to root this tree. The telescoped

portion represents the relationships among a) the 30 individuals of c.

punctatofasciatus (blacken boxes) and C. pelewensis (shaded boxes) that fell

within clade A, b) the 81 individuals of C. punctatofasciatus (blacken boxes),

C. pelewensis (shaded boxes), and C. multicinctus (shaded circle) within clade

B and c) the 27 individuals of C. multicinctus (shaded circle) within clade C.

Trees constructed using parsimony with a 10:1 transition/transversion

bias were similar, but not identical. In these trees, members of clade Band

clade C fell within distinct lineages. However, individuals in clade A

clustered at the base of these trees, as in the neighbor-joining tree, but failed to

group within a defined lineage. Forcing these individuals into a

monophyletic lineage produced trees that were only 7 steps longer.

Consistency and retention indexes for parsimony trees were extremely low at

0.25 and 0.77, respectively.

97

100

a)

72

36

1~111o 5

% sequence difference

11IIII, Tahi~!.Cook 1.1**L.-_ 1IiII, FIJI

iii, Tahiti I., Philippines *

iii , Cook Islands

II, Fiji I.,Palau *

L..-__II , Fiji

III, Tahiti., Philippines It.L

• , Philippines r-*., Philippines kl.*

.,Palau ~

11m, Tahiti ~_

• , Philippines~11II, Cook Islands I

• , Philippines *m,Cook Islands

Ill, Tahiti

11m, Fiji I.,Guam *

11III, Tahiti

1Dl, Cook Islands I*.,Guam

IL-_ 16,Cook Islands

m, Cook Islands III, Philippines *

• , PalauL.-__ ., Philippines

I I I I "o 5% sequence difference

98

Figure 2.6: Collection location and phylogenetic relationships and among

individuals within clade B (continued),

99

I " " Ia 5% sequence difference

Indonesiac::J Fiji

c::J Fiji_ Palau

Fiji_ Guam\----;;;;;;;; _ Indonesia

'------I _ Palau_ PalauI::::J Cook Islands_ Guam

_ Palau'------..:; r:::J Cook Islands

---- 0 Hawaii (2)c::J Cook Islands

.J ..--- r:::J Cook Islands.--. ~F~ ..

I::J Fiji Jl

L~==:;;-~ - Philippines_ Guam- Philip.pines_ Parau

_ Philippinesr.::::J Tahiti

I::::::J Fiji_ Guam

_ Palau_ Indonesia

c::::J Fijic::J Fiji

CO -Hawaii (2), Philippines_ Palau

l'::J '& Hawaii_ Palau

L... Philippines

_ Philippines.. Philippines

II ,...---- t::J Am. Samoa_ Palau~---r--_ Philippines

.. Palau

I ....--c:. F~hiliPPines.....--- c:::::J Cook Islands

111 ~Fi~alaua::::J £iii

- Iiiiii Philippines_ Palau

.. Palau

t::J Cook Islands_ Philippines

r:::::2 Fiji_ Palauc::::l Am. Samoa

L- Philippines--z-- c::J Tahiti

c::::::I Cook IslandsE::::] Cook Islands~Fiji,,--- c::II Cook Islands

_ Palaut::J Fiji

_-_..:- _ Philippines_ Indonesia

r---- c::J Fiji_ Indonesia

t::::::J Cook Islandsr::J Cook Islands

r---L=~-~ PalauG - Guam'---- .. uam

1.- Indonesia..------- _ Tahiti

I I II I Ia 5

% sequence ditference

c\

36

100

72

b)

100

Figure 2.6: Collection location and phylogenetic relationships and among

individuals within clade C (continued).

101

c)72

36

100

I I I I I Ia 5

% sequence difference

102

\

@ , Hawaii (2)@,Hawaii@,Hawaii® ' Ha., Johnston (2)@,Midway@,Hawaii

® ' Johnston Atoll@,Midway@,Hawaii

®,Hawaii

@,Hawaii@ , Ha (2), Johnston~ ,Hawaii

@ , Johnston Atoll@,Midway@,Midway

@,Midway~, Midway@,Midway@ , Johnston Atoll

I I I I I Ia 5

% sequence difference

350

300(/)

~0

250(/).-1-0resS- 2000u..... 15001-0QJ

..0 100E;j

Z 50

00 5 10 15 20 25 30 35 40

Number of pairwise differences

Figure 2.7: Frequency distribution of pairwise differences among individuals

in clades A and B (see figure 2.6).

103

Figure 2.8: Frequency distribution of pairwise differences among individuals

in a) clade A, b) clade B, and c) clade C (see figure 2.6).

104

a) 70

en 60~ Clade A0

50en'crt:l0.. 40E0u

30......01-0

~ 20

E;j 10Z

00 5 10 15 20 25 30 35 40

Number of pairwise differences

b) 450

en 400~0 350 Clade Ben.-1-0rt:l 3000..

§ 250u

...... 2000

aJ 150..0E 100;j

Z 50

00 5 10 15 20 25 30 35 40

c)Number of pairwise differences

90

80en~ 70 Clade C0en

'C 60rt:l0..E 500u 40......01-0 30Q)

..020E

;:l

Z 10

5 10 15 20 25 30 35 40Number of pairwise differences

105

I I IKilometers

tN

2000o

Oahu, n=15

' .. C. pelewensis

~ .{~tl:n~,

Cook I I . ~ ..s anus, 1l=~8 . " ,

' ..

C. multicinctus

'itFiji, n=21

..- \

......,-~ ..., ..

C. punctatofasciatus

~

aQ\

Figure 2.9: Proportion of the three major mtDNA lineages in the populations sampled: 0, clade A; _,

clade B; ~, Clade C. Larger squares represent the approximate geographic range of each species. See Blum

(1989) fo site specific distributions of each species

......o""-J

4

(/)

~ 3.9(/).....I-<ro

~8 2

......o!n

"S::l 1Z

o<0.5 0.5-1.0' '1.0-1.5' '1.5-2.0

• heterospecific pairs

II conspecific pairs

5.0-5.5 5.5~.0

% sequence difference between terminal pairs

Figure 2.10: Frequency distribution of the branch length of terminal nodes containing individuals of the

same species (shaded bars) versus those containing individuals of different species (blacken bars).

CHAPTER 3

Random Mating and Species Boundaries in Allopatric Coral Reef Fishes

3.1. ABSTRACT

Three closely related butterflyfishes (Chaetodontidae) that differ in

color pattern, partition the tropical west Pacific into large allopatric ranges.

Previous mtDNA and allozyrne analysis supported the evolutionary

distinctiveness of only the Hawaiian endemic, Chaetodon rnulticinctus.

MtDNA and allozyme variation was distributed randomly between C.

pelewensis and C. punctatofasciatus. Pairing experiments with C. pelewensis

and C. punctatofasciatus suggest that the greater genetic differentiation in C.

multicinctus is paralleled by greater behavioral and presumably reproductive

isolation. Chaetodon multicinctus pairs assortatively and heterospecific trials

show significantly lower levels of intra-sexual aggression relative to that

observed in conspecific trials. By contrast, pairing between C. pelewensis and

C. punctatofasciatus appears to be random with respect to color pattern.

Additionally, levels of intra-sexual aggression between these two species are

similar to that observed in conspecific trials. Observations of C. pelewensis

and C. punctatofasciatus in the zones of sympatry yield similar conclusions.

A broad, approximately 3000 kilometer, hybrid zone joins the geographic

ranges of the two species. In localities across this zone, individuals with color

patterns intermediate between the two parental forms are common.

108

Moreover, in both Papua New Guinea and the Solomon Islands, pairing

among "pure" parental and "hybrid" phenotypes is random with respect to

color pattern.

Despite being tightly coupled reproductively and evolutionary, C.

pelewensis and C. punctatofasciatus remain uncoupled phenotypically. Color

pattern difference between the two species change over a much narrower

range than expected given the complete homogenization of variation at

presumably neutral loci. These results suggest that selection directly on color

pattern, and probably little else, is maintaining phenotypic integrity of these

two color variants while most other genes move freely across "species"

boundaries.

109

3.2. INfRODUcnON

Phenotypically similar groups that have allopatric or parapatric

distributions have always posed interesting challenges in evolutionary

biology. Because many of the traits that differ in these groups are similar to

characteristics used to define species, these forms (e.g., species, subspecies,

races) are thought to represent the first stage of diversification and thus

provide a context within which to explore the nature and timing of barriers to

gene flow (Mayr, 1942). Moreover, in many instances, allopatric forms meet

and form hybrid zones which reveal the extent to which reproductive

isolation and ecological coexistance can occur (Harrison, 1990). Information

on the levels and patterning of cytological, biochemical, morphological, and

genetic variation across hybrid zones coupled with behavioral and ecological

studies have provided insightful views into the nature of species distinctions.

When multiple data sets are available, it is the concordance or discrepancies

between different types of data that have provided the best clues into the

origins of reproductive isolation, the maintenance of genetic cohesion, and

the ecological and evolutionary significance of inter-specific differences

(Avise and Saunders, 1984; Littlejohn and Watson, 1985; Barton and Hewitt,

1985; Harrison, 1990; Hewitt, 1989; Moore and Price, 1993; Mallet, 1993; Parsons

et al., 1993).

Nearly all studies examining the genetic, reproductive and

evolutionary cohesion of parapatric or allopatric taxa have focused on

terrestrial or freshwater groups with limited dispersal potential (Barton and

Hewitt, 1981; Avise and Saunders, 1984; Littlejohn and Watson, 1985;

Szymura and Barton, 1986; Rand and Harrison, 1989; Moore and Price, 1993;

Mallet, 1993; Parsons et al., 1993). There have been few examinations of

110

hybrid zone complexes in the marine environment, where organisms often

posses extended dispersal capabilities due to a long-lived planktonic larval

stage (however see Schopf and Murphy, 1973; Bert and Harrison, 1988). Yet,

the stability of a hybrid zone and the flow of genes and gene complexes across

it is thought to rely on a balance between selection and gene flow (Barton and

Hewitt, 1985; Moore and Price, 1993). In terrestrial systems, where dispersal

distance per generation is often measured in meters or kilometers, tension

zones between taxa can become trapped and stabilized in areas of low

population density (Barton and Hewitt, 1985; Hewitt, 1989). However, in

marine organisms with long-lived planktonic larval stages, dispersal is a

much more potent force and hybrid zones between taxa are likely to be much

more dynamic.

Here we focus on the genetic and reproductive cohesion in a group of

three closely related allopatric West Pacific butterflyfishes, Chaetodon

multicinctus, C. pelewensis and£ punctatofasciatus. Species differ from one

another primarily in color pattern (figure 2.1). In these bright non-cryptically

colored reef fishes, color pattern distinctions are often the only diagnostic trait

between closely related Butterflyfishes, the differences among these species

are greater than among some broadly sympatric species (Burgess, 1978; Blum,

1988; Blum, 1989). As a result, early taxonomist gave each color variant a

species level distinct. Subsequent collections showed little phenotypic

variation and recent revisions of this family have not challenged the original

distinctions (Burgess, 1978; Allen, 1980; Blum, 1988).

However, mtDNA and allozyme variation is partitioned far less neatly

among the three species. Sequences of the mtDNA control-region suggest

that species diverged in allopatry 300,000-800,000 years ago. Yet, these data

111

confirm the evolutionary distinctiveness of only C. multicinctus. MtDNA

variation is distributed randomly between C. pelewensis and C.

punctatofasciatus and argues for extraordinary high levels of hybridization

and introgression of mtDNA variation (Chapter 2). Allozyme variation is

partitioned similarly. Genetic differences between C. multicinctus and either

C. pelewensis or C. punctatofasciatus suggest very low levels of gene flow

(Nm<l). By contrast, allozyme variation is distributed randomly between

populations of C. pelewensis and C. punctatofasciatus separated by over 7500

kilometers consistent with high levels of hybridization and introgression of

many nuclear-encoded genes.

Because color pattern may be important in the reproductive and social

ecology of these brightly colored reef fishes, it may be under strong selection

(Lorenz, 1966; Peterman, 1971; Reese, 1975; Ehrlich, 1977; Kelly and Hourigan,

1983). Adults of all three species form long-lasting male/female pairs that

defend contiguous, non-overlapping, feeding territories and color pattern

may be an important component of mate choice or species recognition in this

group.

We examine the degree of assortative pairing among these species

using both controlled pairing experiments and field observations in zones of

sympatry. In the controlled pairing experiments, pair formation and

behavior of adults was examined using individuals of each species collected

from phenotypically homogeneous populations. In these areas, social

interactions are restricted to phenotypically similar individuals and the

pairing behaviors under controlled experimental conditions probably reflect

innate rather than conditioned discrimination of phenotypic differences. For

C. pelewensis and C. punctatofasciatus we also examined pairs in a narrow

112

zone bounded by Papua New Guinea (PNG), the Solomon Islands (SI) and the

northern portion of the Great Barrier Reef (GBR) where both species co-occur

(Blum, 1989). Several authors have reported seeing fishes with intermediate

color patterns in this area (Randall et al., 1977; Allen, 1980; Blum, 1989; Allen

and Swainston, 1992). Yet the abundance and fate of these reputed hybrids, as

well as behavioral interactions among the two siblings remained unclear.

113

3.3. MATERIALS AND METHODS

Characterization of phenotypic variation

Of the three species, C. multicinctus, the Hawaiian archipelago and

Johnston Island endemic, is phenotypically the most distinct. Although the

underlying vertical body stripes are similar to C. punctatofasciatus, it lacks the

striking yellow wash, the bright orange coloration on the caudal fin, and a

yellow eye band. Chaetodon pelewensis and C. punctatofasciatus are more

similar, possessing nearly identical yellow and orange patterning but differing

in the orientation of the underlying black body stripes which run diagonally

in C. pelewensis and vertically in C. punctatofasciatus.

We have assembled collections of all three species from 13 widely

distributed locations across the western Pacific between 1990 and 1993. Our C.

multicinctus individuals were collected from Midway Island (n=14), Oahu

(n=20) and Molokai (n=70) in the high Hawaiian Islands and from Johnston

Atoll (n=25). Chaetodon pelewensis were collected from Moorea, Society

Islands (n=9), Rarotonga, Cook Islands (n=21), Viti-Levu, Fiji (n=87).

Chaetodon punctatofasciatus were collected from Bali, Indonesia (n=12), the

Philippines (n=28), Palau (n=82), Guam (n=8), and the Marshall Islands (n=l).

In addition, we visually sampled three locations, two in the 51 and one

within PNG, in the zone of overlap between C. pelewensis and C.o

punctatofasciatus. The collection is currently stored in the -70 C freezer at

the Pacific Biomedical Research Center, Kewalo Marine Lab, Honolulu,

Hawaii and is available upon request. Voucher specimens for each

population were deposited at the Bernice P. Bishop Museum, Honolulu,

Hawaii.

114

We augmented our collections with previous collections housed in the

Bernice P. Bishop Museum in Honolulu, Hawaii and the California Academy

of Sciences in San Francisco, California. Most specimens in these museums

were collected within the last forty years and included individuals from

Pitcairn Island (n=8), the Marquesas (n=3), the Society Islands and Tuamotus

archipelago (n=35), the Philippines (n=7), the Caroline Islands (n=5), Guam

(n=3), Palau (n=6), and the Marshall Islands (n=20) (see table 3.1).

We did not observe any striking variation in the phenotype of the C.

multicinctus across its geographic range. However, C. pelewensis and C.

punctatofasciatus showed considerable phenotypic variation. In these two

species, individuals were given a phenotype score between 1 and 5 based on

deviations from the "pure" parental phenotype. Type 1 individuals possessed

distinct and sharp diagonal bands characteristic of C. pelewensis. whereas the

black bars of type 5 individuals ran vertically, typical of C. punctatofasciatus.

Type 2 and 4 individuals were clearly allied with one of the "pure"

phenotypes yet showed slight variations. For example, in a type 4 individual,

some vertical bands bifurcate, muddling the otherwise sharp pattern of the

"pure" C. punctatofasciatus phenotype. Type 2 individuals are similarly

muddled but none-the-less could be clearly classed with C. pelewensis. Type 3

individuals, however, were strikingly intermediate and could not be reliably

classed into either C. punctatofasciatus or C. pelewensis (see figure 3.1). It is

important to note that these phenotype classes did not reflect the underlying

genetics of color pattern, which are unknown in these fishes, but only our

interpretation of pattern differences.

Because these designations reflected differences in the underlying dark

banding pattern they did not fade with preservations and we were able to

115

clearly phenotype fishes fixed in formalin and preserved in alcohol for over

ninety years.

Random mating and pair-bonding experiments

Behavioral experiments

An initial pilot study indicated that strong pairs formed between adult

C. multicinctus collected from different reefs around Oahu within hours after

being released into a large, 2500 gallon, donut-shaped aquarium at the Hawaii

Institute of Marine Biology. The dimensions of this tank were approximately:

6 meters outer diameter, 4.3 inner diameter and a height of 1 meter. This

tank, which lacks corners, allows for the continual forward movement of

individuals. To further approximate natural conditions the bottom of this

tank was lined with small coral heads of Porites compressaJ P. lobada,

Montipora verrucossa, and Pocillopora meanadrina collected from patch reefs

in Kaneohe Bay, Oahu, Hawaii. All four species of coral are common within

the range of C. multicinctus and closely related species, if not the same

species, are common within the range of C. punctatofasciatus and C.

pelewensis (Hourigan, 1987; Hourigan, 1989; O. McMillan, personal

observation). The behavior of pairs in tank experiments was similar to the

behavior of pairs in nature with individuals spending the majority of their

time feeding and swimming in very close proximity to each other.

In the tank experiments, two males of different species were placed in

the tank together with a female that was the same species as one of the males.

The adults used in these experiments were collected from five widely

scattered locations across the Pacific and shipped to Oahu, Hawaii. Chaetodon

punctatofasciatus were collected from Palau or the Philippines, C. pelewensis

were collected from either Fiji or the Cook Islands and C. multicinctus were

116

collected from reefs around Molokai, Hawaii. Each of these regions were well

entrenched within the geographic range of a given species and, with the

possible exception of the Fijian sample, all individuals shared a consistent

phenotype pattern. Within Fijian populations, a small proportion of the

population (less than 10%) showed patterning of stripes along the body wall

that were slightly muddled (type 2) relative to the "pure" C. pelewensis type

(figure 2.1).

Fishes were received via overnight air cargo from July 1992-April 1993.

After clearing U'S. customs, fishes were unpacked and 15-25 conspecifics were

housed together in large, 600 gallon, saltwater holding tanks at Kewalo

Marine Lab, Honolulu, Hawaii. In order to prevent the outbreak of disease,

all fishes were subjected to an eleven day incubation and treatment period.

During this period, individuals were bathed three times for 30 minutes in

1ml of formalin per gallon of sea water. This concentration of formalin was

fatal to the egg, juvenile, and adult stages of many of the common

ectoparasites that threaten these fishes in captivity but did no apparent harm

to the fishes (Stoskopf, 1993). Each of the three formalin treatments was

followed by three non-treatment days. During all time in captivity, with the

exception of the actual pairing experiments, fishes were fed a specialized diet

designed by the Waikiki aquarium.

Following the treatment and quarantine period, fishes were sexed by

catheterization (Ross, 1984). The soft portion of the dorsal fin of females was

marked with a small hole to facilitate identification in pairing experiments.

Fishes were placed into individual 5 gallon holding tanks where they were

allowed to recover for 2-3 days. After this recovery period, two males and a

female of approximately the same size were transported to Hawaii Institute of

117

Marine Biology, released into the donut tank and allowed to acclimate over

night.

Behavioral observations were made by WOM from a blind within the

center of the tank. One-way glass allowed WOM to monitor the behavior of

the group without the fish being aware of his presence. Fishes were observed

during 4 to 5, 10-minute observational periods in which 1) the amount of

time the female spent with each male 2) the amount of time the two males

were together and 3) the aggressive interactions among individuals were

recorded. For these observations, "together" was defined as within 0.25

meters of each other. Aggressive interactions included attacks, chases, and

dorsal fin flexing displays. Each of these observational periods were separated

by one hour. All data were recorded directly using a computer-based events

recorder (BEAST version 3.0, Losey, 1988). Following the experiment, fishes

were collected, weighted, and the standard length ~as measured. Individuals

were only used in one experiment.

Between 5 and 9 trials of all possible combinations of two males and a

female among the three species were examined (table 3.2). These

heterospecific trials were compared to 12 conspecific trials (four per species) in

which two males and a female of the same species were placed in the tank

together. For statistical analysis, the observational periods were grouped and

the mean time for each event was used to characterize a trial. A pair was

considered to have formed if a male and female spent more than 50% of the

observational period within close proximity to one another. The paired-male

was designated as the focal male. In the cases (n=5) when a female associated

with both males more than 50% of the time, the male that she spent the most

time with was declared the focal male.

118

Field observations

In addition to the controlled pair-forming experiments, we examined

natural pairs of C. pelewensis and C. punctatofasciatus in areas where the

ranges of the two species overlap. We established three sites in a transect

across this zone. Two of these sites were from adjacent island groups within

the 51 separated by roughly 150 kilometers. The third was along the north

coast of PNG near Madang, some 1600 kilometers to the west.

The phenotypes of both paired and unpaired individuals were

examined during transect dives. In these dives, the observer (WOM) swam

in one direction maintaining a depth of between 10 and 20 meters below sea

level. He recorded the phenotype of every individual and whether or not

that individual was paired. As noted previously, paired individuals often

swim in very close proximity to each other. In addition, non-overlapping

territories were generally small, 50-100 square meters in areas of high coral

cover, making it possible to positively identify the members of a pair (Reese,

1975; Tricas, 1989). In instances when only a single individual was initially

observed, we continued to monitor that individual for one minute. If no pair

was located then that individual was scored as a solitary individual. During

these dives, we also scored the phenotype of any juvenile individuals that

were encountered. A transect lasted roughly 70 minutes and covered a

distance of approximately 300 meters.

119

3.4. RESULTS

Patterning of phenotypic variation

With the exception of PNG and the SI, our collections were composed

predominately of individuals exhibiting one or the other "pure" phenotype

(figure 3.3). The highest frequency of "non-pure" phenotypic classes occurred

in Fiji, where approximately 10% of individuals showed slight variation from

the "pure" C. pelewensis patterning. Museum specimens from the

Philippines, Palau, and Tahiti collected over 40 years ago showed little

deviation from those collected between 1990 and 1993.

By contrast, within the two locations in the 51 [the New Georgia Islands

(n=64) and the Russell Islands (n=250)] individuals showed almost every

conceivable phenotypic pattern (figure 3.2). Nearly 70% of the individuals

examined had phenotypes that were intermediate between parental types.

Roughly a third of these intermediate phenotypes could not easily be allied

with either parental phenotype (i.e. type 3) (figure 3.1).

Within PNG, individuals showed a distinct sift towards C.

punctatofasciatus-like phenotypes (types 4 and 5). Over half of the 86 adults

observed along the barrier reef off Madang lagoon had a "pure" C.

punctatofasciatus pattern. Of the remaining half, 22 individuals possessed the

similar type 4 phenotype. There were only four individuals with a "pure" C.

pelewensis phenotype.

In both PNG and the 51 deviations within any of the "hybrid"

phenotype classes were extreme and it was clear that our five phenotypic

classes did not adequately categorize color pattern variation. Nor did these

classes take into consideration striking asymmetry in color pattern on either

side of an individual. For example, one individual (not collected) had the

120

pattern of a C. pelewensis on one side and the pattern of C. punctatofasciatus

on the other. We collected and photographed 37 individuals from the 51 and

PNG. From this collection, 8 of the 10 individuals given a type 3 designation

showed very different patterning on either side. The remaining type 3

individuals showed a moderate degree of asymmetry (figure 3.2). Color

pattern asymmetry was less pronounced in type 2 and type 4 individuals. For

these two classes, most (8 of 13) individuals showed a moderate degree of

color pattern asymmetry, 1 showed a high degree of asymmetry, and 4 were

nearly symmetrical. By contrast all"pure" phenotypes had identical, or

nearly identical, color patterns on either side.

Behavioral experiments

Conspecific trials

The interactions among individuals in the trials in which two males

and a female of the same species were placed into the donut tank together

were identical in all three species. In these trials, a single male-female pair

always formed and this pair fed and swam together for on average 82% of the

observational period. In contrast, the female spent less than 1% of the

observational period with the remaining unpaired male (figure 3.4a). This

was because whenever the unpaired male came within visual range, it was

attacked by the paired male. Aggression between males could be quite violent

involving both chases, bites, and spears with extended dorsal spines. The

unpaired male was clearly subordinate. He actively avoided contact with the

pair and it was not uncommon for him to spend most of the observational

period hiding beneath a coral head as far away as possible from the pair. In all

controls, females exhibited very little aggression towards either male (figure

3.4b).

121

In controls there was no indication that size, measured in either

standard length or weight, was a good predictor of which male paired with

the female. In half of the controls the larger male formed a pair with the

female and in the other half it was the smaller male (p>O.25, binomial test).

Heterospecific trials

The results of the 18 trials between individuals of C. punctatofasciatus

and C. pelewensis were qualitatively and quantitatively similar to the

conspecific controls. In these trials, a single male-female pair always formed

and this pair spent nearly 85% of the observational period in close proximity

to each other and away from the remaining unpaired male. There was no

evidence for positive assortative pairing among males and females of C.

punctatofasciatus and C. pelewensis under these experimental conditions

(table 3.3). Nearly as many heterospecific as conspecific pairs formed in these

trials suggesting few pre-mating barriers to reproduction in these two species

(chi-square=O.277, p=0.27 based on 500 randomizations, Roff and Benson,

1989). The behavioral interactions among males were identical to the

interactions among conspecific males (chi-square=1.42, p=0.464; z=-0.934

Mann-Whitney U-test, p=0.3498). In all but one trial, when the two males

were observed to come in contact, the paired male was extremely aggressive

towards the unpaired male (table 3.4). This high level of intra-male

aggression was particularly telling because it suggested that the two species

failed to distinguish between hetero- and conspecific color patterning.

Six 'out of 8 heterospecific pairs occurred between a male C.

punctatofasciatus and a female C. pelewensis (Fisher's exact test, p=O.OOl).

However, it is unclear if this result was a function of chance, was attributable

to a better physical condition of C. punctatofasciatus males or reflected some

122

innate dominance of C. punctatofasciatus over C. pelewensis. Neither size

nor a simple measure of condition (weight/length ratio) played an obvious

role (p>O.25, binomial test).

By contrast, the Hawaiian endemic, C. multicinctus, exhibited a much

higher degree of behavioral isolation. In trials between C. multicinctus and C.

pelewensis only a single heterospecific pair formed, indicating a high degree

of assortative pairing (table 3.3) (chi-square=7.543, p=O.001 based on 500

randomizations). This pair was between a male C. pelewensis and a female C.

multicinctus. In addition, males were often less aggressive towards

heterospecific males than they were to conspecific males in the control trials

(chi-square = 8.896, p=0.004) (table 3.4). This effect was variable, however, and

was only seen in 6 of the 10 trials where contact between males was observed.

In these trials, less than 2.0% of the time that males were in contact was spent

in aggressive posturing. In the remaining 4 trials, levels of aggression were

high and identical to what was observed in conspecific trials. The two types of

reactions suggest that, although there was some degree of

behavioral/reproductive differentiation between these two species,

differentiation was not absolute.

A similar, but slightly more complex, pattern was evident in the

behavioral trials between C. multicinctus and C. punctatofasciatus. In 4 of the

13 trials, a heterospecific pair formed and assortative pairing was not clear

(chi-square=3.343; p=0.06 based on 500 randomizations). In each case, the

heterospecific pair was between a female C. punctatofasciatus and a much

larger male C. multicinctus. As in the C. pelewensis and C. multicinctus trials

there were significant differences in the levels of male/male aggression

compared to conspecific controls (table 3.4) (chi squares=9.976, p=0.03), though

123

again, there was a clear bimodality to the responses of males towards one

another. Either they were highly aggressive or they very nearly completely

ignored each another.

Behavior of pairs in zones of overlap

Theories for the reinforcement of pre-mating isolation predict a higher

degree of species discrimination in sympatric relative to allopatric

populations (see Butlin, 1989). Thus it becomes important to verify the

pattern of random pairing among C. pelewensis and C. punctatofasciatus

observed among allopatric individuals with observations of the pattern of

pair formation in areas where both species co-occur.

Nearly eighty-five percent of the 306 adults scored in the 51 were

observed as pairs. There was no significant difference in the phenotypes of

individuals that were paired versus individuals that were not paired (figure

3.5) (chi-squared=3.038 and 5.04, p=0.55). In addition, in the 130 pairs

examined there was no compelling evidence for positive assortative pairing

among similarly patterned fishes (figure 3.6) (chi-squared=5.040, p=0.98). In

other words, an individual's phenotype was not correlated with whether or

not it was paired and, if paired, with the phenotype of it's partner.

We might expect any pattern of assortative mating to be strongest for

"pure" phenotypes. This is because, assuming color pattern is polygenetic,

individuals with "pure" phenotypes are more likely to be recruits from

"pure" parental populations. Thus, any innate preferences for assortative

mating may not have broken down by repeated hybridization and

backcrossing. There were a total of fifty-one pairs containing a "pure" C.

pelewensis phenotype. Within this group there were 10 instances in which

two "pure" individuals were paired, 13 instances in which a "pure" type 1

124

individual paired with a phenotypically similar type 2 individual, and 29

cases in which a "pure" C. pelewensis phenotype paired with types 3-5. This

distribution was not significantly different from that expected under a model

of random pairing (chi-square=1.4, p=0.49) suggesting few barriers to

reproductive isolation. There were too few individuals of "pure" C.

punctatofasciatus phenotypes (type 5) in these two populations to do a parallel

type of analysis.

Within PNG, although the majority of individuals possessed C.

punctatofasciatus-like phenotypes (type 4 and 5), there was no compelling

evidence for positive assortative mating. Individuals with "pure" C.

punctatofasciatus phenotypes were just as likely to be paired with a similar

phenotype (type 4 or 5) as with an individual with a hybrid or C. pelewensis­

like pattern (chi square=0.42, p=O.81) (figure 3.7).

125

3.5. DISCUSSION

Among the three similar species studied, only C. multicinctus appears

to have achieved the evolutionary, genetic, and reproductive cohesiveness

consistent with its species-level designation. Chaetodon pelewensis and C.

punctatofasciatus, show only superficial differences at both allozyme and

mtDNA loci (Chapter 2 and Appendix C). Coupled with the lack of

differentiation at presumably neutral genetic markers is a lack of pre-mating

barriers to reproduction. Pairing, the basis for mating in these fishes (Reese,

1991; but see Lobel, 1989) appears to be random among individuals collected

in allopatry and those observed in sympatry. These genetic and behavioral

results suggest that there is little inhibiting the movement of genes between

these two species, and provide no compelling evidence to suggest that either

has achieved the evolutionary cohesion to merit their species-level

distinction.

Despite this conclusion, color pattern distinctions remain true across

the vast majority of each species' range (figure 3.3). Levels of gene flow

between Philippine populations of C. punctatofasciatus and Fijian

populations of C. pelewensis are far higher than levels expected to prevent

the accumulation of neutral differences by genetic drift (Appendix C). Yet,

these populations show completely different distributions of phenotypic

variation, with all 30 individuals collected from reefs around the Philippines

possessing the "pure" C. punctatofasciatus phenotype and all individuals

(n=88) from Fiji possessing a C. pelewensis-like phenotype. Plotting the

proportion of "pure" C. pelewensis-like color pattern in localities along a

roughly linear transect across the tropical Pacific yields a "sharp" change in

the distribution of phenotypes between Fiji and the 51 (figure 3.8). By

126

contrast, the proportion of one of the two main mitochondrial lineages in

these same populations changes only slightly over 10,000 kilometers. These

incongruent patterns suggest two different possibilities. Either color pattern is

not genetically determined in these fishes and thus does not reflect

evolutionary history, or selection directly on color pattern, and probably little

else, is maintaining phenotypic integrity while other markers move freely

between "species."

Environmental cline

In many fishes, color pattern is known to be highly heritable and can be

controlled by a single gene or by a suite of many autosomal and sex-linked

genes (McPhail, 1969; Endler, 1980; McAndrew et al., 1988; Teve et al., 1989).

As a result, it has always been assumed that color pattern in butterflyfishes is,

likewise, under genetic control. However, little is known about the genetics

of color pattern in this family, raising the possibility that color pattern is at

least partially determined by environmental conditions. It is difficult to

imagine an environmental gradient that would maintain tight control over

color pattern from the Pitcairn to Fiji (some 5000 kilometers), change so

abruptly, and then be constant again from PNG to the Marshall Islands and

Indonesia. Temperature, which has been shown to have an effect on pattern

formation of butterfly wings, shows little systematic change across the tropical

Indo-west Pacific (Reynolds, 1988).

The only obvious environmental difference among reefs across this

area is a strong west-east attenuation of species diversity (Stehli and Wells,

1971; Rosen, 1988; Belasky and Runnegar, 1993). Coral diversity drops from

over 68 genera on reefs within the shallow seas surrounding the Philippines

to less than 34 genera on reefs in Tahiti (Rosen, 1988). The "species" are

127

obligate corallivores, and diet may shift in different parts of the Pacific. Diet

has been shown to effect color pattern in many aquarium kept fishes, albeit

subtlety, and it is possible that some food item, or lack thereof, is contributing

to the patterning of phenotypic variation (B. Carlson, pers. comm.).

Several observations argue against diet shifts as a major factor

contributing to the strong color pattern gradient. First, a number of closely

related coral feeding butterflyfishes show no variation in color pattern across

this region. Second, these "species" are coral feeding generalists, feeding on

most corals within their territory in roughly the proportion of their

abundance (Hourigan, 1987). Many of the most abundant corals in outer reef

habitats where these species are found are distributed broadly from Indonesia

through French Polynesia (Veron, 1986). Lastly, individuals with very

different phenotypes often occupy the same territory and presumably feed on

the same suite of coral species. Thus, it seems unlikely that adult diet

influences color pattern in these fishes. However, color pattern crystallizes

quickly upon settlement and newly recruited juveniles possess adult

coloration, suggesting that color patterns may be fixed at or prior to settlement

(McMillan, personal observation). As a result, differences in larval and very

early juvenile diet may be contributing to the observed phenotypic gradient.

Secondary contact and hybrid zone dynamics

On the other hand, color pattern in these butterflyfishes may indeed be

under genetic control. The genealogical pattern of extant mtDNA variation

in this group is most consistent with a scenario in which the color pattern

differences that define species arose in allopatry between 300,000-800,000 years

ago (Chapter 2). Thus, the extreme phenotypic variation in the SI and PNG

may represent hybridization following secondary contact. In this case, the

128

obvious implication of the strong geographic structure to phenotypic

variation in the absence of mtDNA and allozyme differentiation is that

selection is preventing the erosion of color pattern outside the zone of

secondary contact and in the presence of potentially homogenizing levels of

gene flow.

Although it is unclear how color pattern is inherited in these fishes,

the extreme phenotypic variation, including the observation of asymmetrical

color patterns on either side of many intermediate classed individuals, argues

that color pattern is probably under complicated genetic control involving

multiple gene loci and epigenetic developmental pathways. The asymmetry

in color pattern is particularly interesting because it is often associated with

developmental instability resulting from both environmental and genetic

stress (Palmer and Strobeck, 1986; Leary and Allendorf, 1989). Hybrids often

show increased morphological asymmetry relative to parental populations,

presumably due to the break-up of intrinsic co-adaptive gene complexes

(Leary et al., 1985; Zakharov, 1981, Graham and Felley, 1985; Ross and

Robertson, 1990; Markow and Ricker, 1991; but see Lamb et al., 1990).

Asymmetry is more prevalent in hybrids from genetically distinct parental

taxa. In these cases, the asymmetry is evident in a many morphological traits

and is often associated with reduced viability (Leary and Allendorf, 1989).

Presently, it is unclear if the pattern of asymmetry extends to other

morphological characters as well, or is restricted to color pattern in these

butterflyfishes. However, the shear number of hybrids, coupled with the

observation that individuals with hybrid phenotypes are paired and the high

levels of "interspecific" gene flow suggested by the pattern of mtDNA and

allozyme markers argue that hybrids between the two color forms are viable.

129

In many respects, the behavioral, genetic, and phenotypic patterns in

these fishes parallel the patterns reported in well studied hybrid zones in

neotropical butterflies and North American birds (Moore, 1987; Moore and

Buchanan, 1985; Mallet and Barton, 1989a and 1989b; Moore et al., 1991;

Mallet, 1993; Moore and Price, 1993). In zones of sympatry, mating between

races of Heliconius erato butterflies and the red and yellow-shafted flickers,

Colaptes auratus, is random and hybrids are viable (Mallet, 1993; Moore and

Price, 1993). In addition, with the exception of wing coloration in races of

Heloconius and plumage coloration in northern flickers, parental

populations show only slight differences at presumably neutral genetic

markers (Grudzien and Moore, 1986; Grudzien et al., 1987; Mallet, 1993).

In both Heloconius butterflies and northern flickers, strong frequency

dependent selection is envisioned to be contributing to the discrepancy

between phenotypic and genetic variation. For Heliconius, differential

predation on oddly patterned butterflies is thought to be maintaining the

distinction between races. This selection has been measured and shown to be

quite high, nearly 52% against foreign rnorphs (Mallet and Barton, 1989b). For

northern flickers the mechanism of frequency dependent selection is less

clear. Although red- and yellow-shafted flickers differ primarily by plumage

characteristics, the hybrid zone very neatly follows several ecological

gradients (Moore and Price, 1993). Moore and Price (1993) proposed that both

ecological selection against hybrids and frequency dependent social selection

on plumage traits contribute to the narrowness of the cline between

subspecies. In either case, average selection against loci involved in or closely

linked to plumage traits must be high to explain the width of the hybrid zone

(Moore and Price, 1993).

130

Similar frequency-dependent selection regimes may operate in these

butterflyfishes. Under this model, the ability of a fish to survive and

reproduce must depend on its color pattern relative to the color pattern of

other conspecifics on the reef where it settles. In this case, two obvious

questions need to be explored. First, how strong is selection on color pattern,

and second, how is it acting? For the first question, previous theoretical

models can provide a reasonable estimate of selection preventing the erosion

of color pattern given the width of the hybrid zone and some idea of dispersal

distance in these fishes. As for as the second question, these results only

begin to dissect the nature of selection operating on color patterns in these

fishes.

How strong is selection on color pattern?

For the dynamics of a wide variety of hybrid zones, the width, defined

as the inverse of the maximum slope of change between alleles or polygenetic

traits, is a function of dispersal potential and the strength of selection (Barton

and Hewitt, 1981; Szyrnura and Barton, 1986; Barton and Gale, 1993; Moore

and Price, 1993). For a hybrid zone maintained by frequency dependent

selection, simple mathematical models suggest that effective selection, ~

acting on a typical gene involved in the determination of a polygenetic trait is

roughly equivalent to (2.8cr/w)2, where q is the average dispersal, defined as

the standard deviation in parent to offspring distance, and w is the width of

the cline (Barton, 1983; Mallet and Barton, 1989a; Moore and Price, 1993).

Although our collections are widely scattered, the change in proportion of

"pure" C. pelewensis-like phenotypes across this transect suggests a hybrid

zone width of approximately 3000 kilometers (figure 3.8).

131

This spatial scale is at least an order of magnitude broader than most

terrestrial hybrid zones (Harrison, 1990). For example, the width of hybrid

zones in races of Heliconius butterflies are 10-100 kilometers (Mallet, 1993).

Similarly, in flickers, the hybrid zone between the red-and yellow-shafted

subspecies seldom exceeds 300 kilometers. However, the dispersal potential

of these fishes is likely orders of magnitude greater than these terrestrial

examples. For example, in Heliconius butterflies, estimates of dispersal

distance range between 0.7 and 2.3 kilometers per generation (Mallet, 1993).

In flickers, although some juveniles are known to disperse up to two

hundred kilometers, most fledglings nest near the parents and adults show

strong breeding site fidelity. These birds, likewise, show marked population

structuring at both mitochondrial genes and allozymes (Moore and

Buchanan, 1985; Grudzien and Moore, 1986; Grudzien et al., 1987; Moore et

al., 1991). In contrast, butterflyfishes have a larval stage that drifts in the

plankton 40-60 days before settling (Tricas, 1987; Hourigan and Reese, 1987;

Leis, 1989). Although the details of larval dispersal in these fishes are

unknown, this stage is known to facilitate transport of larvae of other marine

species with similar long-lived planktonic larval stages thousands of

kilometers. For example, Caribbean reef fishes are commonly transported as

far north as New York, some 2000 kilometers from the nearest reproductive

populations (Allen, 1980). Similarly, during el Nino years, waves of Indo­

west Pacific marine organisms settle in the Galapagos nearly 5000 kilometers

from the nearest source population (Richmond, 1990). Strong directional

currents certainly aid dispersal in these cases; however, these examples serve

to underscore the scale of dispersal in marine organisms with planktonic

larval stages.

132

Assuming a moderate dispersal distance of 200-500 kilometers per

generation for these fishes, we estimate a per locus selection coefficient acting

against homogenization of color pattern genes of between 0.04 and 0.22. If

only three loci are involved in the determination of color pattern in these

fishes then total selection acting against the introgression of color pattern may

be as high as 66%. Clearly our estimation of selection on color pattern genes

is sensitive to our assumption about dispersal in these fishes. However,

given complete homogenization of allozyme and mtDNA variation and high

estimated levels of migration between locales separated by 7500 kilometers,

we believe that this estimation is conservative.

The nature of selection on color pattern in butterflyfishes

If our estimation of dispersal distance is correct then selection

maintaining the observed gradient in color pattern is quite strong. The

evolutionary and ecological significance of the vivid, non-cryptic coloration

in butterflyfishes, as in most reef fishes, is poorly understood (Lorenz, 1967;

Peterman, 1971; Brockman, 1973; Potts, 1973; Ehrlich et al., 1977; Allen, 1980;

Kelly and Hourigan, 1983; Nuedecker, 1989). Butterflyfishes have a well

developed visual cortex and visual cues are thought to be important in both

inter- and intra-specific recognition (Peterman, 1971; Reese, 1975, 1989; Snow

and Rylander, 1982; Bauchot et al., 1989). Many teleosts can neurally or

hormonally alter color pattern and these changes have been demonstrated to

be important in intra-specific communication (Barlow, 1963; Lanzing and

Bower, 1974; Nelissen, 1976; Hailman, 1979; Rowland, 1979). Although

butterflyfishes are sexually monomorphic and capable of only slight

modifications in color, bright coloration may similarly be an important intra­

specific social signal for either species-recognition, mate attraction (epigamic

133

sexual selection), or intra-sexual selection (Zumpe, 1965; Lorenz, 1966;

Peterman, 1971; Reese, 1975). Alternatively, bright conspicuous coloration in

butterflyfishes may have evolved as aposomatic warning, or disruptive

signals directed towards potential predators (Gosline, 1965; Neudecker, 1989).

What do the behavioral and genetic patterns reveal about the nature of

color pattern in this group? First, the higher degree of behavioral and genetic

isolation in C. multicinctus suggests that divergence in color pattern can be

paralleled by the evolution of assortative pairing either through species

recognition or mate selection. However, the random pairing among C.

punctatofasciatus and C. pelewensis argues that color pattern differences are

not providing similar cues in these two species. Our behavioral experiments

extend this conclusion to phenotypically homogenous allopatric populations.

Though neither the field observations nor our controlled pairing

experiments reveal all the forces underlying pair-formation in these species,

they clearly suggest that mate choice or species recognition cues based on color

pattern are not playing a strong role. It therefore seems unlikely that the

original divergence of color pattern in this group was coupled with strong

inter-sexual selection or pressure for species recognition.

It is possible that intra-sexual social selection, similar to that proposed

to maintain differences between red and yellow-shafted flickers, may be

maintaining the color pattern differences in these butterflyfishes. In this case,

strong selection among individuals of the same sex for territorial space

provides a likely mechanism. For C. multicinctus, the establishment of a

territory is thought to be a prerequisite for successful reproduction (Tricas,

1987; Hourigan, 1987). Removal experiments suggest that competition for

open territorial space in C. multicinctus is fierce. When one individual of a

134

pair of C. multicinctus is removed, the remaining individual quickly

establishes a new mate (Hourigan, 1987). Similarly, after a pair is removed, a

new pair quickly establishes itself within the bounds of the original territory

(Hourigan, 1987). These results indicate a substantial population of "floater"

individuals waiting for space to become available and suggest that intra­

sexual competition for space may be quite strong (Hourigan, 1987; 1989).

Such strong intra-sexual competition was evident in our pairing

experiments among all three species. The outcome of these experiments was

determined essentially by male-male encounters. One male is clearly

dominant and the female always associates with him. In a number of

experiments (n=5) when two heterospecific males were used in a second

experiment, this time with a female from a different species, the original

dominance hierarchy was maintained and the new female paired with the

dominate male. Likewise, in experiments in which two females and a male

were added to the tank, there was strong female/ female aggression and the

pair formed between the male and dominate female (n=3) (McMillan,

unpublished data). Similar evidence for strong intra-sexual competition

comes from field observations of C. multicinctus. In this species, territorial

defense is directed primarily towards members of the same sex (Hourigan,

1989). Because both males and females compete for open space in these

species, intra-sexual social pressures on color pattern would extend to both

sexes.

Thus, it is possible that color pattern in these territorial reef fishes

evolved under strong intra-sexual social pressure (sensu, West-Eberhard,

1983) and has failed to erode because individuals that settle outside the range

of their conspecifics fail to establish a territory. In the one-on-one social

135

situation set up in the pairing experiments, one male was able to dominate

the space independent of color pattern. In the more complex social network

of a coral reef, many individuals are likely competing for a limited number of

open territories. Under these situations, oddly marked individuals may not

provide the necessary social signals to keep other fishes competing for

territorial space from continually attacking and eventually displacing them

from territories. Color pattern is thought to be an important social signal in

many avian species and changes in social status often follow experimental

manipulation of plumage characters (Butcher and Rower, 1987). For example,

the red epaulet of the male red wing blackbird is important in territorial

defense. In this species, males set up and defend breeding territories from

other males. If the epaulet is blackened, intrusion into a male's territory

increases and he is frequently displaced (Smith, 1972).

Observations in the 51, where phenotypic variation is extreme and

unrelated to an individual's ability to hold a territory, clearly indicate that the

frequency-dependent selection on color pattern can break down. But, it is the

rapidity at which a break down occurs that is the critical dynamic of this

theory. In short-lived species, where adult turn-over is high, the social

dynamics of a community may change quickly. However, in a long-lived

species, like these butterflyfishes, where pairs are stable and can maintain the

same territory for more than four years, the social dynamics of a community

may change much slower (Hourigan, 1989; Reese, 1991). Under these social

dynamics, frequency-dependent selection against the odd phenotype may

erode more slowly relative to the flow of neutral genes.

In PNG, C. punctatofasciatus-like phenotypes (type 4 and 5) dominate

the population but adults with odd phenotypes hold territories and mating

136

appears to be random. It is possible that the proportion of odd-colored

individuals in this population, roughly 27%, has exceeded some threshold

beyond which frequency-dependent selection is no long acting. Alternatively,

it is possible that recruitment of the juvenile fish and their growth into the

adult population is what is color-type dependent. Predation, which has

shown to be a powerful force in maintaining a narrow hybrid zone among

races of Heliconius butterflies, may be playing a role in the pattern of

phenotypic variation in these fishes, especially among juveniles.

Butterflyfishes are underrepresented in the gut content of piscivorous fishes

(Hiatt and Strasburg, 1960; Hobson, 1974; Hourigan, 1987). Their laterally

compressed bodies and sharp dorsal spines are thought to make them difficult

to eat and has lead to the suggestion that the bright, vivid colors serve as an

aposomatic advertisement to potential predators (Gosline, 1965; Neudecker,

1989). Indeed, chaetodontids do not flee when large predators are nearby.

Instead, they turn sideways and display their lateral body markings

(Neudecker, 1989; McMillan, personal observation). Thus, it is possible that

oddly patterned juveniles are more heavily preyed upon, inhibiting the

erosion of color pattern differences.

Conclusions

It is clear that despite being tightly coupled reproductively and

genetically, C. pelewensis and C. punctatofasciatus, none-the-less. remain

uncoupled phenotypically. Is the hybrid zone between the two forms

stationary, as the hybrid zone between red- and yellow-shafted flickers (Moore

and Buchanan, 1993), or are the color forms in the process of homogenizing?

The patterning of phenotypic and genetic variation provided by this study

represents but a single point in the history these forms that likely emerged

137

hundreds of thousands of years ago (Chapter 2). Given that the dynamics of

hybrid zones are tightly coupled to dispersal, this hybrid zone between these

high dispersal forms is likely to be quite turbulent. Further study can only

help elucidate the forces shaping morphological and genetic variation in

these butterflyfish.

138

Table 3.1: Specimens of Chaetodon punctatofasciatus and C. pelewensis

examined from collections at the California Academy of Science, San

Francisco, California and the Bernice Pauley Bishop Museum, Honolulu,

Hawaii.

Location no. Collection Dates Catalogue numbersMarshall Is. 20 1967-1970,1986,1972 BPBM 1776, 6245, 6298, 8230, 8244,

10782-10783; CAS 58959, 42230

Palau 6 1966-1970;1955-1956 BPBM 6841, 8294,10192; CAS 37896-97

Guam 3 1968;1922 BPMB 4189, 8451

Caroline Is. 5 1953, 1959 CAS 37898

Philippines 7 1901, 1931-33, 1949 SU 9609, 25791, 29899;CAS 52775

Society Is. / 35 1956-58;1967-1971 BPBM 6053, 6070, 6120,8357, 13569,

Tuamotus 17264; CAS 37899-902

Pitcairn I 8 1970-1971 BPBM 13245, 16555, 16937

Marquesas 3 1971 BPBM 11670, 11769

139

Table 3.2: Number of pair-formation experiments matching a single female

with a conspecific and a heterospecific male.

,C. multicinctus C. peJewensis C. punctatQfasciahls

6C. multi./C. pele.

fII' C. multi.ZC. punct,

C. pele. / c. punct.

5

6

140

9

7

9

Table 3.3: Results of controlled pairing experiments matching a single female

with a conspecific and a heterospecific male for each of the three possible

hetero-specific comparisons. ,a)

C. pelewensis

~ C. punetatofasciatus

C. pelewensis C. punctato.

3 2

6 7

,b)

C. pelewensis

ti'c I"-=. mu ticmctus

C. pelewensis C. multicinctus

5 1

0 5

,c)

C. multicinctus

ti' c. punetatofasciatus

C. multicinctus C. punctato.

6 3

0 Aq

141

Table 3.4: Summary of the results of pair-bonding experiment among the

three species. For this table, levels ~f male/male aggression within

heterospecific and control trials was categorized into strong (>50% of the time

males were within visual contact they were involved in aggressive displays)

and slight «30% of the time that males were in visual contact were they

involved in aggressive displays). "nc" signifies that no male/male contact

was recorded.

# hetero. MALE/MALE AGGRESSION# trials pairs Strong nc slight

Con trols 12 9 0 3e. punctato.ZC, pele. 18 8 12 1 5e. pele./ e. multi. 11 1 . 4 1 6**e. punctato'/c. multi. 13 4 3 3 7**** denotes a significant difference in levels of male/male aggression at the0.01 level between heterospecific and conspecific trials based on a chi-squaredtest. For this test, the probability density curve of possible chi-squared valuesfor a given sample size was estimated by 500 randomizations (see Rolf andBenson, 1989). If the observed chi-square value was greater than 495 of therandomly generated values then a significant difference was assumed. Mann­Whitney U-tests in which the actual proportion of time males were involvedin aggressive displays yielded identical results.

142

Figure 3.1: Plate of the three hybrid phenotypic classes. Type 2 individuals (a)

resemble the phenotype of C. pelewensis but with definite deviations from

the typical diagonal banding pattern. Similarly, type 4 (b) individuals

resemble C. punctatofasciatus but with slight muddling relative to normally

distinct vertical bands. Type 3 (c) individuals are striking in their variation

and can not be confidently allied with either of the two parental types.

143

a)

b)

c)

144

Figure 3.2: Color plate of individuals collected in a) the Russell Islands, SI b) New

Georgia Islands, SI and c) Madang, PNG.

145

146

Figure 3.2: Color plate of individuals collected in a) the Russell Islands, 51 b) New

Georgia Islands, 51 and c) Madang, PNG (continued).

147

a)

148

Figure 3.2: Color plate of individuals collected in a) the Russell Islands, 51 b) New

Georgia Islands, 51 and c) Madang, PNG (continued).

149

150

Figure 3.2: Color plate of individuals collected in a) the Russell Islands, 51 b) New

Georgia Islands, 51 and c) Madang, PNG (continued).

151

152

Figure 3.2: Color plate of individuals collected in a) the Russell Islands, 51b) New

Georgia Islands, 51 and c) Madang, PNG (continued).

153

c)

154

......UlUl

Figure 3.3: The geographic distribution of the five phenotype classes of C. punctatofasciatus and C.

pelewensis. The wheel represents the proportion of each of the five classes observed within a given

location: ., "pure" C. pelewensis (type 1);~ , type 2; 1m, type 3; II , type 4 and; D, "pure" c.

punctatofasciatus (type 5) (see figures 2.1 and 3.1). Locations with an asterisk (*) beside the sample size were

augmented with museum specimens (see table 3.1). The smaller circles represent regions where 5 or fewer

individuals were examined.

2

3

4

C. pelewensis

.~

C. multicinctus

Marshall Is.(n::21")

Kilometers

·f'!' ".(n=87) . ...•. I . ta.e .• Pitcairn (n=11 )'~I., .

sOlom~n':i~n"31"1' , , Society. ,.(n~4~'i N"': Cook Is.(n=21) .

r-----. •

o 2000

c. punctatofasciatus

.....~

Figure 3.4: Results of pairing experiments among conspecifics. Figure 3.4a

depicts the amount of time female spent with each male during conspecific

pair-forming trials. In 3.4b, levels of aggression, defined as the proportion of

time two individuals were in visual contact and involved in aggressive

displays, are plotted for each of three possible pairwise interactions.

157

a) 1.0'"00.-I-<OJ 0.8P..-res~

.90.6.....

res:>I-<OJen. ..00 0.4.......0~0.- 0.2.....I-<0P..0I-<~ 0.0

Female with Female withfocal male non-focal male

b) 1.0-r----------------------,.....ures'E 0.88s::.-OJ 0.6.9............os:: 0.4.9.....I-<oP..o 0.2t

o.o~--Aggression

towardsnon-focal maleby focal male

Aggressiontowards

non-focal maleby female

158

Intra-pairazzrcssionbb... - .............. v ..

30-.-----------------r-----,paired

unpaired

25

20s::0..........C'a

.-<;j0.. 1500......0

~10

5

o1 2 3

Phenotype class

4 5

Figure 3.5: Histogram of the proportion of the 5 different phenotype classes

among individuals that were paired (blacken bars) versus the proportion of

the 5 different phenotype classes among individuals that were not paired

(shaded bars).

159

• observed-~., II expected-

:ll~lmn

0;:;::

~I~ lI!'..WI ~r ::;:;:

~H: ;..

;:::;:: :~:::: ;::::;: :\~- ~?~ {=:= :.::; :;::::::j::;: .:. :;::::: ::::. ::;:: ~~r

'.' ;:;::: .;::;)(if] :;:;:: :;:{ ;jf :::::;

- t=: .:::. ....:: .::;:;,-.:: ::::::. "::::: f\]~:t w~:: :;::::: :::;:;:

;:::;:; :::. :::::::~~r:;::;: :::::: .::::> :~f : ;

- ::::;: ::::.:. ;:;:;:: : ; ; ::t!1!1!l:

:;:;:!: ;:;:: '.:.:

h ::.::'~ j:~: \W

:~~~i~:f:· :;:::::

I:.::::: f".. \!j ;~i~ii~

~.- ::::;::;/

::;:::;t1~ t:~ ~~t "'.jf= ::::::; )j:

'.' [~j~~:~ ::;:;:; :;:;: .:.::: :;:;:; {j] ;:;::j:!.f:~.

.:.'::r

i::~::r }j~

::::;: r:j~ .:. ::::.- :':..- :::.,: ::::::: ".'::: :tj: fj: :::::: "'.

I

:::::: -x. j:j;::; :f(j :ff .:.; :.::=- I",

lilll:i,:'

:f: .:;::.: :::;:: :;:;:: :j:::; ::::~: ;::;: ~:~::

i'.'::

f:~: :!!t!:lll

if!!;:::::j;:

~:t (:: :::~~~ i~j~~- !!!i]t::::::: :::.:. ~t~::::::: .:.:.:;

~:f'.' :::;::. ::::::: :;:;~ I:

.;.:.;

:~irr:::::;.

I.;. :;:::::

2 r~: uma ?r a .:::::: a 2 II,Jj ::::::- - ~ .... l- I- l- I- - ::::: l- I-I

1/1 1/2 1/3 1/4 1/5 2/2 2/3 2/4 2/5 3/3 3/4 3/5 4/4 4/5 5/5

Phenotype of individuals in pairs

18

16

14trJ1-4"; 12c,'010

1-4Q) 8

164

2

o

Figure 3.6: The observed (blacken bars) and expected (shaded bars)

distribution under a model of random mating of the 15 possible types of pairs

between individuals of C. pelewensis, C. punctatofasciatus, and hybrids in the

Solomon Islands. Expected distributions for each of the 15 possible were

calculated assuming that the observed proportion of each phenotypic class

within the Solomon Islands represented the underlying pool of adults. For

this estimation, we grouped adults from both locations within the Solomon

Islands. We felt justified in pooling these samples because there were no

significant differences in the distribution of phenotypes between the two

Island groups (chi-square=6.359, p=0.17).

160

• Observed

[ill] Expected

24,-----------;=:============::1,22

20

rJ) 18J-<.....cU 16

0...'0 14

aJ 12

S 10

Z 8

6

4

2

o

Phenotype of individuals in pairs

Figure 3.7: The observed (blacken bars) and expected (shaded bars)

distribution of combinations of "pure" C punctatofasciatus phenotypes paired

with other "pure" C punctatofasciatus phenotypes, type 4 phenotypes, or

hybrid/C pelewensis -like phenotypes in New Guinea. The expected

distribution under random mating was calculated as explained in figure 3.6.

161

1

.9Q)

~g .8Q)

...t:~ .7....tilt::~ .6Q).....Q)

P!.

rJl .5

e~ .4P-.

.......o .3t::o

1:o .2P-.e

P-. .1

o

Palau

2000 4000 6000

Cook Islands Pitcairn Island

1Tahiti

.9

.8

-e.7 ~

t"ClClJs::

.6 ~Zc

.5 S......0

.4 s::.9.....I-<0

.3 P..0I-<c,

.2

.1

08000 10000 12000

Distance from Bali,Indonesia(kilometers)

Figure 3.8: Change in the proportion of "pure" C. pelewensis phenotypes

(blacken circles) (type 1) plotted against the distance (in kilometers) from Bali,

Indonesia compared to change in the proportion of one of the two distinct

mtDNA lineages (shaded circles) revealed in our recent analysis of mtDNA

variation in these two species (see Chapter 1).

162

APPENDIX AA 495 bp portion of the mitochondrial cytochrome b gene for 18 Chaetodonspecies. C. tinkeri was arbitrarily chosen as the representative sequence.Sequences are from 5'-3' and begin at the first codon position. Numbering isaccording to the human mitochondrial sequence (Anderson et al., 1981). A"." represents sequence identity. Missing nucleotide information is denotedby a "?".

Species Sequence ...

•• T

.. T

• .T· .T

· .T.. T

• .A• .A• .A· .A· .A

• .A• .A• .C

.. C• .C

• .C• .C

??

• .C

• .C

• .C

· .C.. C

• .C

??? ??? ??? t .,

•• C

•• T•• T

[149661GTT GCC ACA GCA TTT TCT TCT GTC GCC CAT ATC TGC CGA GAC GTC[45]

.??

.??

.??

.??

C.tinkeriC.declivis wilderiC.flavocoronatusC.burgessiC.declivis wilderiC.mer/C.xanC.mer/C.xanC.mertensiC.madagascariensisC.paucifasciatusC.ar:gentatusC.miliarisC.multicinctusC.mult/C.pun/C.peleC.punctatofasciatusC.pelewensisC.gutatissimusC.auriga

AAC TAC GGC TGA TTA ATC CGC AATC.tinkeriC.declivis wilderiC.flavocoronatusC.burgessiC.declivis wilderiC.mer/C.xanC.mer/C.xanC.mertensiC.madagascariensisC.paucifasciatusC.argentatusC.miliarisC.multicinctusC.mult/C.pun/C.peleC.punctatofasciatusC.pele~lensis

C.gutatissimusC.auriga

.. T

.. T

.. T

.. T

.. T

.. T

.. A• .A.. A• .A• .A

• .T.. T .. A .. C

.. T .. A

.. T .. A

.. T .. A

.. T .. A

.. T .. AC.. ..A .. C

[15011]CTT CAC GCC AAC GGT GCC TCC[90]· .C

• .C.. C• .C.. C

• .C

• .C

· .C

· .C.. C.. c ??? ???.. C.. C

· .C

• .C.. C• .C

163

APPENDIX A (continued)A 495 bp portion of the mitochondrial cytochrome b gene for 18 Chaetodonspecies.

[15056JC.tinkeri ATA TTC TTT ATC TGT ATT TAC ATG CAC ATC GGC CGA GGA CTT TAT[135]C.declivis wilderi ., TC.flavocoronatus ., TC.burgessi .. TC.declivis wilderi .. TC.mer/C.xan · .C · .T · .TC.mer/C.xan .. T ooTC.mertensi · .T ooTC.madagascariensis · .T · .T · .CC.paucifasciatus · .T · .T · .CC.argentatus .. C · . T · .CC.miliaris · .T · .A · .T · .? · .CC.multicinctus Goo .. T ooA · .TC.mult/C.pun/C.pele G.. .. T · .A · .TC.punctatofasciatus G.. ooT · .A · .TC.pelewensis G.• ooT · .A ooTC.gutatissimus G.. ., T · .A · .T · .CC.auriga · .C · .C ooC · .T C.T · .T · .T · .C · .C

[15101)C.tinkeri TAC GGC TCA TAC CTT TAT AAA GAA ACA TGA AAC GTC GGC GTA GTA[180]C.declivis wilderi · .TC.flavocoronatus · .TC.burgessi · .TC.declivis wilderiC.mer/C.xan · .G ooT · .G .. TC.mer/C.xan · .G · .T · .G .. TC.mertensi · .G · .T ooG " TC.madagascariensis · .G · .T ooG .. TC.paucifasciatus · .G .. T · .G .. TC.argentatus · .T · .G · .GC.miliaris ooT · .C · .G .. T Aoo .. TC.multicinctus · .T · .C · .C Aoo · .T .. TC.mult/C.pun/C.pele · .T · .C · .C A.. · .T · .TC.punctatofasciatus ooT · .C · .C Aoo .. T .. TC.pelewensis .. T · .C ooC Aoo .. T .. TC.gutatissimus ooT · .C · .C · .G Aoo .. T .. TC.auriga ooT .. G ooT · .C · .G A.. .. T · .C

[15146]C.tinkeri CTC CTC CTT CTA GTC ATG ATA ACC GCC TTC GTA GGG TAT GTC CTC[225jC.declivis wilderiC.flavocoronatusC.burgessiC.declivis wilderiC.mer!C.xan ooT · .T ooA • .1\ ooC .. TC.mer/C.xan · .T · .T ooA • .1\ .. C · . TC.mertensi ooT ooT ooA • .1\ · ,C · . TC.madagascariensis · .T · .A · .A .. C · .TC.paucifasciatus ooT · .T · .A , ,1\ , ,C .. TC.argentatus · .C · .T • .1\ · .A · .CC.miliaris · .T ooT ooA • .G · .A · .TC. mu It icinct us .. T ooC .. T · .A GC.C.mult/C.pun!C.pele · .T · .C · .T • .1\ GCGC.punctatofasciatus · .T · .C · .T ooA GC.C.pelewensis · .T · .C ooT ooA GCGC.gutatissimus · .T · .C · .T · .A GC. · .AC.auriga " T ooA GC. ooT .. T .. T · ./\

164

APPENDIX A (continued)A 495 bp portion of the mitochondrial cytochrome b gene for 18 Chaetodonspecies.

[15191JC.tinkeri CCC TGA GGA CAA ATG TCA 1'1'1' TGA GGG GCT ACT GTA ATT Ace AAT[270JC.declivis wilderi .. GC.flavocoronatuse.burgessi · .GC.declivis wilderie.mer/e.xan · .G · .e •• TC.mer/C.xan .. G · .C •• Te.mertensi · .G · .1'e.madagascariensis · .G · .1' · .ee.paucifasciatus · .G •. T · .ee.argentatus .. C .. C .• 1' · .ee.miliaris • .1' · .e .. C •. T · .Ce.multicinctus · .1' · .A · .1' · .Ce.mult/e.pun/e.pele · .1' .. A · .1'e.punctatofasciatus • .1' .. A · .1'e.pelewensis .. 1' · .A •. Te.gutatissimus · .1' · .Ae.auriga · .A · .G · .A · .e · .A · .e •. T •• T

[ 15236Je.tinkeri CTe CTC TeT Gee GTe eCT TAT ATT GGA AAC Ace TTA GTA CAA TGA1315)C.declivis wilderie.flavocoronatuse.burgessiC.declivis wilderie.mer/C.xan · .e · .C C..e.mer/C.xan · .e · .C e ..e.mertensi · .C · .e c ..e.madagascariensis · .e e ..e.paucifasciatus · .e C.GC.argentatus ??? ??? ??? ??? ??? ??? ???

e.miliaris • .1' .. 1' · .e · .C · .e .. Ge.multicinctus · .C • .1' •. T

e.mult/C.pun/e.pele · .e · .e •• T •• T

e.punctatofasciatus · .C .. C •• T .• TC.pelewensis · .C · .C •• 'I' .• Te.gutatissimus · .C · .C •• T .• T

e.auriga · .1' .. 1' · .? · .C · .C · .1' C.?

[15281]e.tinkeri ATe TGA GGA GGC 1'1'1' Tce GTA GAC AAC GCC ACC TTA ACT eGA TTe[360Je.declivis wilderiC.flavocoronatuse.burgessie.declivis wilderie.mer/C.xan .. 1' · .G .. 1' · .T e.Ge.mer/e.xan · .1' · .G · .1' •• T e.Ge.mertensi · .1' .. G •• T .• T e.GC.madagascariensis · .1' · .G •• T .• T e.Ge.paucifasciatus •• T · .G · .1' •• T e.Ge.argentatus · .G · .1' C .. · .ee. miliaris .. G · .e .. 1' · .eC.multicinctus · .G .. 1' •• T .. 1' .. 1'e.mult/e.pun/C.pele .. G · .G • .1' .. 1' • .1' •. 1'C.punctatofasciatus · .G •• T .. 1' · .1' .• 1'e.pelewensis · .G •• T •. 1' .. 1' .• 1'C.gutatissimus .. G •• 'I' • .1' • .1' .• 1'C.auriga · .G · .G · .e · .A .. 1' e .. · .e

165

APPENDIX A (continued)A 495 bp portion of the mitochondrial cytochrome b gene for 18 Chaetodonspecies.

[15326]C.tinkeri TTC GCT TTC CAC TTC CTC CTC CCT TTC ATC ATT GCA GCC GTA ACC[405]C.declivis wilderiC.flavocoronatusC.burgessiC.declivis wilderiC.mer/C.xan · .C · .T · .CC.mer/C.xan · .C · .T · .CC.mertensi · .T · .C · .GC.madagascariensis · .C · .T · .CC. paucifasciat us .. T · .CC.argentatus · .C .. T · .TC.miliaris · .T · .C · .T .. T .. TC.multicinctus · .C · .T · .T · .GC.mult/C.pun/C.pele · .C .. T · .T · .T · .GC.punctatofasciatus · .C · .T · .T · .Gc.pelewensis · .C .. T " T · .T · .GC.gutatissimus · .C · .T · .T · .TC.auriga · .C · .C · .T · .G · .T .. T

[15371]C.tinkeri ATG GTT CAT CTA ATG TTC CTC CAC CM ACG GGC TCT AAC AAC CCC[450]C.declivis wilderi · .AC.flavocoronatus · .AC.burgessi · .AC.declivis wilderi · .AC.mer/C.xan · .A · .C · .A · .AC.mer/C.xan · .A · .C · .A · .AC.mertensi · .C · .A · .AC.madagascariensis · .C · .C · .AC.paucifasciatus · .A · .C · .C · .A · .AC.argentatus · .C · .C · .A · .T · .AC. miliaris · .A · .T .. T · .A · .G .. TC.multicinctus · .C · .G .. T · .A · .A .. T .. TC.mult/C.pun/C.pele · .C · .G .. T · .A · .A .. TC.punctatofasciatus · .e .. G .. T · .A .. A .. T .. Te.pelewensis · .e · .G .. T · .A · .A .. T .. Te.gutatissimus · .G .. T · .A · .A .. T '".. ,e. auriga · .A A.. · .e · .A · .e

[154161e.tinkeri eTA GGC eTG AAe TeG GAT ATA GAC AAG ATe TCA TTe CAT eCT TAC!495je.declivis wilderiC.flavocoronatusC.burgessiC.declivis wilderiC.mer/C.xan .. T · .A · .CC.mer/C.xan · .A · .CC.mertensi .. T · .C · .A · .CC.madagascariensis .. T · .C · .A · .'-

C.paucifasciatus .. T · .C · .A · .CC.argentatus T .. · .A .. T .. A · .C · .A .. T · .CC.miliaris .. T · .A C.. ??? ??? ??? ??? ??? ??? ???C.multicinctus .. T · .A · .G .. T · .CC.mult/C.pun/C.pele .. T · .A .. G .. T · .C · .CC.punctatofasciatus .. T · .A · .G .. T · .C · .CC.pelewensis .. T · .A · .G .. T · .C · .CC.gutatissimus · .A .. T · .A · .G " T · .C · .CC.auriga .. T · .C · .T · .A · .C · .C

166

APPENDIX B

A 195 bp portion of the tRNAPro end of the mitochondrial control-region for141 individuals of C. multicinctus (n=32),c. pelewensis (n=501 C.punctatofasciatus (n=561 and C. guttatissimus (n=3). Sequnces are from 5' to3' with a ''." and "_" represents sequence identity and a gap, respectively.Missing nucleotide information is denoted by a"?".

Species Location Seguence

• •• TT ••••••••••• -C • C •••••••••••• A ••••••••••••

•.•. T. ..••••.•.. - ...• '" ••......•.•..••....•.•••••••••••••••• - ••••••••••••••••••• C ••••••••

••••••••• G•••••• - ••••••••••••••••••••••••••••

•••• T••••••••••• - ••••••••••••••••••••••••••• ,

................ - .

................ - .

•••• T••••••••••• - ••••••••••••••••••••••••••••

•••• T••••••••••• - ••••••••••••••••••••••••••••

................ - AC .

45GTACCAGATAAGAAGA-ATTACTAAAACCAAAGTAATAAGAAGAC... TT -C .•••• T••••••••••• - ••••••••••••••••••••••••••••

•••• T•.••••••••• - ••••••••••••..••••.•••••.•••••••••••• G•••••• - ••••••••••••••••••••••••••••

••• TT•••••.•.••• -C •••..•••..•••••.•••••.••.•••••• T••••••••••• - ••••••••••••••••••••••••••••

... T? G•••••• -G •••••••••.•••...•••••..•••.

· .. TT C -C A ••• C .· .. TTG -C A G.• •• TT ••••••••••• -C •••••••••••••• A •••••••••• G•· .. TT C ••••••• -G •••••••••••••• A ••••••••••••

Is TT - AC. GC G .???TT -G.C AC.. C .· .. TT C A. -G AC.. C G .· .. TT C -G C.. C G T????????????????-G AC.. C G o

?. TT ••••••••••• -C •••••••••••••. A ••.••••.•••.• •• TT ••••••••••• -G • C •••••••••••• AC.. C ••• G ••••

· .. TT. A A. -G A .••• TT ••••••••••• -T ••••••.••••.•• A •••.•..•..•.

· .. TT... C ..•..•• -G •.••.•••••••.• AC •• C .•• G ••••

• •• TT ••••••••••• -C •••••••••••••• A ••••••••••••

••• TT •••••••••• G-G •••••••••••••• A ••••••••••••

· .. TT G-G •••••••••••••• A ••••••••••••

• •• TT • A ••.•••••• -C •..••••••••••• A ••••••••••••

• •• TT ••• C ••••••• -G •••••••••••••• A ••••••• G••••

•••• T••••••••••• -T •••••••••••••• A ••• C ••••••••

· .. TT -G . C AC.. C G ••• T• •• TT ••• C ••••••• - ••••••••••••••• A ••••••••••• T

· .. TT -G AC.. C G ••••

• •• TT ••••••••••• -G •••••••••••••• A ••••••••••••

· .. TT -G ••••.•.••..••. AC G•••.A T -G AC.. C GG .

Hawaii(2)Hawaii(2l,PhilHawaiiHa(2),JI,MI(2)HawaiiHa(l), JI(2)HawaiiHawaiiHa, MI (2)

HawaiiHawaiiHawaiiJohnson IslandJohnson Island,Johnson IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandAmerican SamoaAmerican SamoaTahitiTahitiTahiti, CookTahitiTahitiTahitiTahitiTahitiTahitiCook IslandsCook IslandsCook IslandsCook .IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook Islands

C. multicinctusC. multi,C. puncta.C. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisc. pelewensisC. pelewensis

167

C. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatcfasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatus

Cook IslandsCook IslandsFijiFijiFi jiFijiFijiFijiFijiFijiFijiFi jiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiGuamGuamGuamGuamGuamGuamGuamGuamPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesIndonesiaIndonesiaIndonesiaIndonesiaIndonesiaIndonesiaIndonesiaPalauPalauPalauPalauPalau

... TT C -C .• • • . • • • . . . • • • A ,.· .. TT 0 • ; ••••• GC A•••••• A•••••... TT o' '" -C •. . • • 0 •••• 0" .AC 0 •••• ,.

'" TT.......• 0" -C •• • • • • • • • • • • • • A••••••••••••

· .. TT .. 0 •••••••• - ••••••••••••••• AC .. C... G....... TT o' •• -G •••••••••••••• A ••••••••••.•

... TT 0" •••• -C •• o ••••••••••• A ••••••••••••· .. TT . A ••••••••• -C •••••••••••••• A ••••••••• 0 ••

... TT -C •• • • • • • o' 0 •••• A ••••••••••••

... TT -C .•..•.•••••••.•••••••......

'" TT '" -C •••••••••••••• A •••••••••• ,.

A •• TT -G •••••••••••••• A ••••••••••••

· .. TT . A -G A T· .GTT 0 • - ••••••••••••••• AC . . C • • • G .· T -T AC .. C .· .. TT -C T A .... TT o' -C •• • • • • • • • • • • . • A ••••••••••••••• T ••••• 0 ••••• • -C •••••••.•.••. . A •••••••.•••.

• ••• T ••••••• 0 ••• -G ••••••••..•••• AC G ••.•

••• TT ••• 0 ••••••• -G ••••.•••••.••• A •••.••••.•••

• •• TT ••••••••• A. -G •••••••••••••. A ••••••.•.•. T

· .. TT -G •• 0 ••••••••••• A •••••••• G •••

... TT -G.C A G T

.... T C •••••• • - •••••••••••• o' .A ••••••••••..

... TT.. o' 0" ••• G-G 0 T AC .. C G .· •• TT •••••••• 0 •• -C 0 ••••••••••••• A ••••• G •• G •••

• •• TT • A ••••••••• -C •••••••••••••• A ••••••••••••

... TT -Co •• • • • • . • • • • • • A ••••••••••••

... TT C ••••••• -G •••••••••••••• AC .. C ••• G ••••

• .•• T •• • C ••••• • • - ••••• 0 •••••• 0 •• A ••••••••••• T

????????????????-? A.....•......? ? ? ? ? ? •• C -G AC .. C G .••• TT •• o' •••••• • -C •• 0 •••••••••• GA " ..• ••• T ••••••••••• -T G • A ••• C ••••••••

· .. TT -C • • • • • • • • . . T A G .••• TT •••••.••••• -C •••• 0 ••••••••• A ••••••••••••... TT -C ••• . . . . . • • . . • • A ••••••••••••

• .• TT .••••.••••• -G •.••••••.••••. AC G.••.••• TT .••••••••• • -C ••• • . . . . • • . . . . A ••••••••••••••• TT •• o ••••••• • -C ..• • • • . . • . . • • • A ••••••••••••

••• TT •. • C •••••• • -G •••••••••••••• AC . . C •• •G ••••

... TT -G •••••••••••••• AC G ••••

· .. TT -G . C AC .. C G .· .. TT C -G AC.. C G .... TT -G •••.•...••••.• A •••••• " .• "

• .• TT ••• C ••.•••• -G •••••••••••••• A ••• C ••••••••

· .. TT C -G AC .. C G T

••• TT •.•••.•••• . -C ••.. • • • G•••••• A ••••••••••••· .. TT C - .. A ACG . C G .••• TT •••••••••• • -C •.•••.•••••...• C•..•••..•..

•.. TT ••••••••••. -C ••••••• G•••••• A ••••••••••••

••• TT •••• , .••• o. -C .••••.••.•.•.• A •••••...•••.

· .. T G-G AC .. C ••• G .... TT C ••••••• - ••••••••••••••• A ••••••••••••... TT G-C A ••••••••••••••• TT •••• , •••••• -G •••••.•••.•..• A •.••••...• "

• •• TT ••••••••••• - ••••••••••••••• A ••••••••• AGT

•.• TT .•••••••••• -G •.•.••.••••••• A .•....•.•.••

••• TT •••• , •••••• -C •••••••• , •••••••••••••••• G.

••. TT •••••••..•. -G •.••..••.•••.. A .....•••.••.

· •. TT .•••.•.•••• -G ••••••.••••.•• AC G.•.•

... TT -C T· .. TT . A -C A T

168

C. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punc~atofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus Palau'-. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. guttatissimus Mauritius Is.C. guttatissimus Mauritius Is.C. guttatissimus Mauritius Is.

••• TT •• G•••••••• -C •••••• G •.••••• A •••.••••••••••• TT??? •• , .??G-C •.•••••••••••• A .•••••••••••

••• TT •• • C •••••• • - ••••••••••••••• A••••••••••••••• TT •••••••••• • -C .• . . • • • . . • • . , .A ••...•.••. , .••• TT ••••••••••• -G.C •••.•••••••. A ••••••• G •• GT• •• TT ••••.•••••• -C •••••••••••••• A •••••••••• G •••• TT •• • C •••••• • - ••••••••••••••• AC • • C •• •G ••••

••• TT •.••••••• A.-C ••••••.••••••• A .•••••••••••••• TT ••••••••••• -C •• • • . • • • • . . • . . A ••••••••••••• ••• T ••••••••••• -T •••••••••••••• A •• • C •••• • • • •

• •• TT • A••••••••• -G •••••••••••••• A ••••••••••• T• •• TT • A •••••••• G-G •••••••••.•••• A .••••.••••• T• •• TT ••••••••••• -C •••••••••••••• A •••••••••• G •

••• TT •••••••••• • -C •••..•...••.. •A •••••..•••••••• TT •••••••••• • -C ••..••....•.• . A ••••••••••••• •• T •••••••••••• - •••••••• G •••••• AC •• C ••••••••• •• T ••••• ; •••••• - ••••.•••••••••. AC •• C ••• G•••." .TT.A ••• G••••• - •••••••••••.••• AC•••••. G ••.•

169

APPENDIX B (continued)Control region sequences.

Species Location Seguence

•..••.. G.- ••• - ••..•••••.••...••••...•••......

......... - - .

....•.... - - .•. G G.. G .••••••••• - ••• - •••••••••••.••.•••••••.• C •.••••

· .... - .. C- .. A- .... G.. A.... G... C.... G.C. C.. C..

90CTTAAATAT-TTG-TTGAATAGTGAAAAAATTTCTAGGTTAGTTA· .. G. - .. C- .. A- ...• G.. A.... G... C....• AC. C.. C..

••••••••• - ••• - ••••••••••• G•••••••••••••••.•••· .. GT- .•. - .. A- .. A. G.• A G GAC.C.. C..

· .. G.- - .. A? . A. G.. A G C AC.C........G.- - .. A- .... G.. A G C AT.C .. C..· .. G. - -C. A- .. AGG .. A GG C T .· .. G.-G .. - .. A- G.. A G C ATCC .. C..

Is .. CCG .-C .C- .. A- G.. A.. G.G C.. T. GACC C..· CC.. -C .. - .. A- G.. A. AG . G C.. T... CC C..TCC .. - .•. - .. A- G.. A.. G.G C.. T.. ACC C..TCC .. -C .C- .. A- G.. A.. G.G C.. T.. ATC C..TCC .. -C .C- •. A- G.. A.. G.G C.. T.. ACC C..· .. G.- - .. A- .. A. G.. A G AC.C.. C..· CC.. -C. C- .. A- .... G.. A.. G. G C.. T. GAC .... C..... G.-C •. - .. A- .. A.G .. A G C AT .C .. C....... - ... - .. A- .. A.G .. A G , GAC.C.. C..TCC .. -C .C- .. A- ... GG .. A.. G.G C.. T.. ATC ... C.... C.. - ... - .. A- .. A.G .. A G.. GC AC .C .. C..· .. G. -C .. - .. A- GGC . A G C AT . C.. C..· .. G. -C .. - .. A- GG .. A G AT. C.. C..· .. G. -C .. - .. A- G.. A G C GACCC .. C..TCC .. -C . C- .. A- G.. A.. G. G.. GC .. T.. ATC C..· .. G. -C .. - .. AA G.. A.. G. G AC.C ..CC .. -C.C- .. A- G AG .G C.. T.. ATC C..· .CG. - .•. - .. A- GG .. A G C AT.C .. C...CC .. -C .C- .. A- .. AGG .. A.. G.G T.. ACCC .. C..... G.-C .. - .. A- .... G.. A G C AT.C .. C...CC .. -C.C- .. A-C .AGG.. A G ,. T.. ACC C...CC.. -C. C- .. A- G.. A.. G.G C GACC C..· - - .. A- G.. A.. G. G C ACC C..· - - .. A- G.. A G C AC.C .· .. G. - - .. A- .. A. G.• A G AC. C.. C..... G.- T.. A- G.. A G C AC .C .. C..TCC .. -C .C- .. A- G.. A.. G.G C.. TC.ACC... C..· .. G. -C .. - .. A- GGC.A G C AT. C.. C....... - ... - .. A- .. A.G .. A G C GAC ....C..

Hawaii (2)Hawaii (2) ,PhilHawaiiHa(2),JI,MI(2)HawaiiHa(1),JI(2)HawaiiHawaiiHa,MI(2)HawaiiHawaiiHawaiiJohnson IslandJohnson IslandJohnson Island~lidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandAmerican SamoaAmerican SamoaTahitiTahitiTahiti, CookTahitiTahitiTahitiTahitiTahitiTahitiCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsFi jiFijiFi jiFijiFiji

C. multicinctusC. mUlti,C. puncta.C. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensis

170

C. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusc. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatus

FijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiGuamGuamGuamGuamGuamGuamGuamGuamPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesIndonesiaIndonesiaIndonesiaIndonesiaIndonesiaIndonesiaIndonesiaPalauPalauPalauPalauPalauPalauPalauPalauPalauPalauPalauPalau

· .. G. - - .. A- .. A.G.. A G AC. C.. C..... G.-C .. - .. A- .. A.G.. A•... G.....•... AC.C .. C..... G.- .. C- .. A- G.• A.. G.G C AC.C .. C..... G.-C .. - .. A- G.. A.... G C AC .C .. C..... G. -C .. - .. A- GG .. A•... G C AT. C.. C..· .... - ... - .. A-C .AGG .. A•. G.G. G.C GAC.C.. C...CCG.-C .C- .. A- .• A.G.. A.. G.G C.. T.. ACC C..· .. G.-C .• - .. AA G.• A.. GGG AC. C ...... - ..• - •• A- .. A.G .. A•. G. G AC.C .. C.." . G. - ... - .. A- .• A. G.. A•. G. G•........ AC.C .. C..· .• G.- ..• - .. A- .• A. G.• A.... G C • • • • • AT. C.. C...CCT .-C .C- .CA- ••.. G.• A.AG. G C.. T.. ACC ... C..... G. - A.• A- .. AGGC. A G C AC.C .. C....... - - .. A-C.AGG.. A G C AC.C .. C..... G.-C .. - .. A- ... GG .. A G.G.C AT.C .. C...CC.. -G.C- .. A- .. A.G .. A.AG. G C.. T.. ACC ... C..· .. G.-C .. - .. A- .• AGG .. A.. G.G C AT.C .. C..TCC .. -C.C- .. A- .•.. G.. A•. G.G C ACC ... C..... G.-G .. - .. A- .. A.G .. A•.. GG AC.C .. C..... G.- ..• - •. A- .•.. G.. A G C GACCC .. C..... G.- - .. A- .. A.G .. A G AC.C .. C..TCC .. -C . C- .. A- .•.. G.. A.. G. G C.. TC. ACC ... C..· .CG. -.G. - .. A- ... GG .. A.. G.G C AT.C .. C..... G.- ... - .. A-..• AGGC. A•... G C AC. C.. C..TCC .. -C.C- .. A- .... G.. A.. G.G C.. T.. ATC ... C..· .. G. - ..• - .. A- .. A.G .. A•..• G AC.C .. C..... G.-C .• - •. AA .•.. G.. A•. G AC. C .... G.- - .. A- .. A.GC.A .. G.G AC.C .. C..... G.- .. C- .. A- G.. A.. G. G C AC. C.. C..... G.- ... - .. A- G.. A G C •• • • • AC.C .. C...CC.. -C .C- •. A-C .. GG .. A G C • • T.. ACC ... C..· .. G. -C .. - .. A- .... G.. A G C AC . C.. C..... G.-C .. - .. A- .. A.G .. A G C AC.C .. C..TCC .. - .. C- .. A- .... G.. A.. G. G C.. T.. ACC C..· .CG.-C. C- .. A-C. A. G.. A.. G. G C.. T.. ACC C..TCC .. - .. C- .. A- G.. A. AG.G C.. T... CC C..TCC •. - .. C- .. A- G.. A..G.G C.. T.. ACC C..... GG-C.. - .. A- GG .. A G C •• • • • AT.C .. C..... G.- ... - .. A- .. A.G .. A G C • • • • • ATC ,..CC.. -CGC- .. A- T.. A.. G.G C.. T.. ATC C..· .. G. - ... - .. A- G.. A... GG c ..... AC . C.. C..TCC •. - .. C- .. A- G.. A..G.G C • • T.. ACC ... C..... G.- .. C- .. A- .. A.G .. A G C AC.C .. C..· .... - .. C- .. A- .•..... A G C AC. C.. C..· .. G. - - .. A- .. A. G.. A G AC. C.. C...CC.. -C .. -.CA- ..... C.T G CC.. T.C.C .... G.-C .. - .. A-.CAGG.. A G C AC C..T.. G. - ... - .. A- GG .. A.. G. G C AC. C.. C..... G.-C .. - .. A- GG .. A G C •• • • • AT.C .. C..· C... -C .. - .. A- G.. A G C AT. C.. C..· .. GG-C.. - .. A- GG .. A G C GAT.C.. C..· ... G... C- .. A- G.. A.•.. G C AC. C.. C..... G.-C .. - .. A- GG .. A G T.. AT .C .. C...CC •• -C .C- .. A-C •. GG .. A G C • • T.. ACC ... C..· .. G. - .. C- .. A- G.. A G C • • • • • AC.C .. C..· .. G. - - .. A- G.. A G C GACCC .. C..· .. G. - - .. A- G.. A G C AC. C.. C..... G.- - .. A- .. A.G .. A G C AC.C .. C..· .CG.- - .. A- ... GG G. G C AT.C .. C..••• G.- .•• - •• A- •••• G.. A •••• G ••. C •.••• AC .C •. C ..

.CC.. -C.C- .. A- G.. A.AG.G C.. T.. ACC ... C..· .ATG-C.. - .. A- G.. A.... G C..... AC.C .. C..TCC .. - .. C- .. A- G.. A.. G. G C.. T.. ACC ... C..

171

C. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. guttatissimusC. guttatissimusC. guttatissimus

PalauPalauPalauPalauPalauPalauPalauPalauMauritius Is.Mauritius Is.Mauritius Is.

• •••• -C •• - •. A- •• A. G •• A •••. G••••••••• AC • C •• C .•

••• G.- ••• - •• A- •••• G •• A •••• G••• C ••••• AC.C .• C ••

• •• G .-C •• - •• AA•••• G •• A •• G .G ••••••••• AC. C •••••• •••• - ••• - •• A-C •• GG •• A •••• G••• C ••••• AC • C •• C ••

• •••• - ••• - •• A-C. AGG •• A ••• GG.G.C •• T .GAT. C •• C ••T.ATG-C •• - •• A- •••• G •• A •••• G.•. C •.••• AC.C •. C ••

• •• G.- ••• - •• A- .• A .GC .A •••• G••••••••. AC.C •• C ••

••• G.- ••• - •• A- •••• G •• A •••• T ••• C ••••• AC ••.• C ••.CC.G-G .• -.CA- ••••• C.C.AG.G •.•. C ••• CAC.C •••••.CC •• -A.C- .CA- ••••• C. T ••••••••...•• CAC .C •••••

..C •• -C •• - .CA- ••••• C.C ••••••••. C ••• CAC .C •••••

172

APPENDIX B (continued)Control region sequences.

Species Location Sequence

•••••• A•••••••••••••••••••••••••••••••••••••••••••••••.•.•••••••••• T ••.••••.••••••.• T •••....

••••••••••••••••••••••••••••••••• C ••••••••••••

.AT .. CA T C.. C TT A•••••.••.•.••••••••••••••••••.•••.•.•• T ••••...

.A .• .G •••••••••••••••••••••••••••••••••••••••••••

.AT A T C.. C TT A

136AGCTTTGTCCCGTCAAAGATACACCAAGTATCATCATCCCTGTATG.AT ..CA C. T C .. C TT A

.AT A T C.. C TT A

.A CA T C.. C TT" A· TTC.. A T C.. C T A.AT A T C .. C TT A

Is .. A •••• A•••••••••••••••••••••••••• C••• T ••••••••· AT A C T CGC..AT A•••••••••••••••••••••••••••••• T ••••••• A.A •••• A.T C .•• T .•.•...•

.A •••• A•••••••••••••••••••••.•••• C••• T ••••••••

.AT A T C.. C T .

.A •••• A•••••••••••••••••••••••••• C••• T ••••••• A

.AT A T C.. C T A

.AT A T C.. C TT A

.A A.T C T .

.AT A.T T C.. C TT .

.AT A T C.. C TT A

.. T A T C.. C TT , .CAGAT .. CA T C.. C •.•••.•.... A.A A.T C T .· .T A T T C.. C T A· AT A C T GC.· .TC .. A.T •..•.••..••••••...••. C •• C •.• TT ••••.. A.AT A C T ..AT A T C.. C TT A.AT A•••••••••••••••••••••••••• C••••••••••••

.p.,••• • A••••••.••••••••••••••••••• C••• T •.••••••

.AT A T C .. C T A

.ATC.CA T TT A

.ATC .. A T C.. C T •••.•.•

.AT .. CA T C TT A

.AT A C C.. C T .· . T A T C.. C TT A.AT A T C.. C T A

Hawaii (2)Hawaii (2), PhilHawaiiHa(2) ,JI,MI (2)HawaiiHa(1),JI(2)HawaiiHawaiiHa,MI(2)HawaiiHawaiiHawaiiJohnson IslandJohnson IslandJohnson IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway Islar.dMidway IslandAmerican SamoaAmerican SamoaTahitiTahitiTahiti, CookTahitiTahitiTahitiTahitiTahitiTahitiCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsFijiFijiFi jiFijiFiji

C. multicinctusC. mUlti,C. puncta.C. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. mul ticinct usC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pe Lcwen s i sC. pelewensisC. pelewensisC. pelewensisC. pe.lewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisc. pelewensisC. pelewensisC. pelewensis

173

C. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusc. punctatofasciatusc. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatus

FijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiFijiGuamGuamGuamGuamGuamGuamGuamGuamPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesPhilippinesIndonesiaIndonesiaIndonesiaIndonesiaIndonesiaIndonesiaIndonesiaPalauPalauPalauPalauPalauPalauPalauPalauPalauPalauPalauPalau

.AT ••• A ••••••.••••••• T •••.•••• C •• C ••• TT •••.•••

.AT ••• A •••••••••••••• T •••••••• C •• C ••• T ••••••••

.AT •• CA •••••••••••• C. T •••••••• C •• C ••• 'IT •••••• A

• AT •• CA •••••••••••••• 'I ••••••••••• C ••• 'IT •••••. A

•• T ••• A•••••••••••••• T •••••••• C •• C ••• TT •••••• A

• A •••• A•••••••••••••• 'I ••••••••••• C ••• 'IT ••.••• A

.A •••• A •••••• G ••••••••••••••••••• C ••• 'I ••••• G •.

•• T ••• A••• T •••••••••• T •.•••••• C •. C .•• T ••••••. A

.ATC .• A•••••••••••••• 'I •••••••• C •• C ••• 'IT •••••••

.ATC •• A•••••••••••••• T •••••••• C •• C ••• T ••.•••• A

.AT •• CA ••• T •..••••••• T •.••••••••• C .•• TT •••••• A

•••••• A. 'I ••••••••••••••.•.••••••• C ••• T ••••••••

· .TC •• A•••••••••••••• T •.•••••• C •• C ••• TT •••••. A

.A •••• A •••••••••••• C.T ••••.•••••• C .••• T .••••••

• • 'I ••• A••••••••••••••••••••.•• C •• C ••• 'IT •••.•• A

.AT ••• A••••.••••••••••..•••••• " .C ••• 'I •..•• GC.

.AT ••• A•••••••••••••• T •••••••••.• C ••• TT •••••• A

.A •••• A •••••••••••.•••.•••••.•••• C ••• T •..•••••• ATC •• A. T •••.•••••••• 'I •••••••• C •• C ••• 'IT •.•..•.

GAT ••• A•••••••••••••• 'I •••••.•. C •• C ••••••••••• A

.AT ••• A••••.•••••••.• T ...••••• C •• C ••• TT ••.••.•

.AT ••• A •••••••••••••••••••••••••••••• 'I •..•••••

· • 'I ••• A. T •••••••.••••••••••••• C •• C .•• 'IT ••.••• A

• • TC •• A•••••••.•••.•• 'I ..••••.• C •• C ••• 'IT ..•••• A

.A •••• A. 'I ••••••••••• G •••.•••••••• C .•• 'I ••.•.•••

.AT ••• A •••••.•••••••• T •••••••• C .• C ••• TT ••.••••

• .T ••• A ••. T •••••••••• T •••••.•• C •• C ••• T •..••.• A

.ATC •• A.••••••••••. C. TG .•.•••. C •• C .•. 'IT •..••.•

• ATC • CA •••••••.•••••• 'I •.•.•••. C •• C ••• 'IT •..••.•

.AT •. CA ••.•.•••..•••• T ••..••.•.•• C ••• TT.C ...• A

.AT ••• A.•••••••••••••••.••••••••• C .•.••.••• G ..

.ATC .CA ••••.••••••••• 'I ••.•.•••.•• C ••• 'IT ••.••. A

.ATC •• A•.••••••.••••• T .••••••• C .• C ••• 'IT ..•••••

.A •••• A•••••••••••••••••••••••••• C •••••••• C •• A

.AT •.• A••••••..••••••••....•...•.••.• 'I ••••..•.

.AT ••• A•.••..•••.•.•••..•••.••••• C ••• T •... CGC.

.AT •.• A.••••.•.•••.••••....••.... C •.• T ••••••. A

.AT ••• A.•••.•.••••.•••••..••.• C •• C •.• TTTC •.•• A

••••.• A••• T •.•••.•••• T ••••.••• C .• C ••• TT .•••.. A

.A •.•• A.T .••••••••. C •••....••.••• C .•• T •.•••••.

.AT •• CA ••••••.•.••••• T •.••...• C •• C ••• TT .••.•• A

.A •••• A •••• A ••••••••••••••••••••• C ••• T •••• C •• A

.AT •• CA •••• A.•••••. C. T •....••. C •. C •.• TT •••.•. A

.AT .• CA .•.••••••••. C •••..••••. C •• C ..• TT •••.•• A

.AT ••• A.•.•••••.••..• T ••..••.• C •• C .•• TT .•••.• A

.A •. C.A •••••.•••••.••••••••.••••. C •.• T .••.•• C.

· . T. C. A••..••.••••.•• T •..•.••. C •• C ••• 'IT ••••.. A

• • TC •• A••.••••••••..• 'I ••...•.• C .• C ••• 'IT .••.•• A

• .T •.• A•••••.•••..••• T •••••••• C •• C •.• TT •••••• A

.AT •.• A ••• 'I ••.•••••.• T ••...••• C •• A ••• TT ••••.• A

.AT ••• A.•.•..••.••••• T •..•.••• C .• C ••• TT •••••. A

.AT .. CA .•••••••••.. C. T •..•.••. C .• C •.• TT ••••.. A

•. 'I •.• A••••..•••••.•. T •••...•. C •• C •.• TT •••... A

.A •••• A•••••••••••••••••••••••••• C ••• ? ..•. G ••

.AT •. CA. 'I C. 'I •..•...• C .• C ••• TT .••••• A

GAT •.. A •••••.•••..•.• T •...•••• C •• C •••••.••.•• A

.AT .• CA.G ••...•••••.. 'I ••••••.••.•••.• TT •••••. A

• ATC . CA ••••••••.••••• 'I ••••.•• GC •• C ••• 'IT .•••.. A

• • TC •• A. 'I .••...•.••.•••....••• C •. C ••• 'IT .•••.• A

.AT •• CA •••••••..••.•. T ••••.••..•• C •.. TT ....•• A

.ATC .. A•••••.••••••.••••••••.••.• C •.• T •.••.• C.

•• T .•• A••••.••••••••• 'I •••••••• C •• C ••• TT •••••• A

.AT •.• A•.•••...••....•.....•••••• C •.• T .••••. AT

174

C. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. guttatissimusC. guttatissimusC. guttatissimus

PalauPalauPalauPalauPalauPalauPalauPalauMauritius Is.Mauritius Is.Mauritius Is.

.AT ••• A•••••••••••••• T •••••••• C •• C .•• TT .•••..•

.AT •• CA•••••••••••••• T ••••••••••• C ••• TT •...•• A•• T ••• A ••• T •••••••••• T •••••••• C •• C ••• T ••••••• A

• A•.•• A••• T •••••••• C • T •••••••.••• C•.• TT •••••• A.A.C •• A•••••••••••• C. T ••••••••••• C ••• T •• C •••• A•• T••• A•••••••••••••• T •••••••• C •. C ••• TT •••••. A.AT ••• A•••••••••••••••.••••••• C •• C ••• TT ••.••••.AT •• CA•••••••••••••• T •••••••• C •• C ••• TT •••..• A.A •• C.A. T •••••••••••••••••••••••• C ••• TT •••.•• A.A .• C .A ••••. C ••••••••••••..•••••• C ••• T •••... CA.A •• C .A •••••••••• , •••••••..•••••• C•.• T .••... CA

175

APPENDIX B (continued)Control region sequences.

Species Location Sequence

• .- ••••••• C •••••.•.•••••••••••.••.•••.•.••••••

• .- •••••••••••••••••••••••••••••••••••••••• G .•

.G- •.•••.•••••.•.••••.••••.•••••..••••••.•••..• .- •••••••••••••••••.•••••••••••••••••• G••••••• .- ..•••••••••••.••••••••.•.•.•••..••. T ••••...

· .- •••••••••••••••.••.••.••••••.•..••.• G••.•..

GG-.A.... TM .....•.............. G..... T.... GA.

GG-.A M T A..G- .A M ........................•. T GA.G. -. A AA T .C.. GA.G.-.A G.M •......................... T.C A.

Is .. G-.A •... TM T T.GG-.A TAA T.C .. GT.CG- AA T.C.. GT ..G-.A TM .....•.................... T.C ... T..G-GA AA T.C .. GT.GG-. A A A T GG.GG-.A TM T.C T.GG-. A••• G • M •••••••••••••••••••••••••• T •••.• A.GG-. A. C A A T G..G-.A.C .. TM T.C .. GT.GG-.A.C A A T A.GG-. A G. AA T A.GG-. A G. AA T .C AGGG- .A •..•. M •••••••••..•.•..•.• , ....•• T ..•.. A..G-.A.C .. TM T. ??????

GG-.A TAA T A..G-.A TM T .CT .GT.GG-. A CA TG GA.GG-.A TAA T.C .. GT.GG- • A••• G •M ••••.••••••••••••••••••••• T •C •.• A.GG-.A TA T.C .GG-.A TA T.C T.GG-.A GTM T TA.G.-.A A T A.GG-.A.C A A T AGG. - . A A GA T A..G- TAA G T .C • . CT.GG-.A G.AA T G.GG-. A. C ••• A •• , •••••••..••••.••..•.•••• T A.

182Hawaii(2) AA-AGATAACGGCCCAATAAGAACCTACCATCAGTTGACATCTACAHawaii (2), Phil GG-.A...•. M G T GA.Hawaii .. - A .Ha(2) ,JI (1) ,MI (2 .. - .Hawaii .. - .Ha (1), J. I • (2), •• • •••••••••..••••..•••••.•••..•••••••......•

Hawaii .. - ..............•............................Hawaii .. - .....•..........•..........................Ha , M.I. (2), GG-.A.C A A.. ooT GG.Hawaii .. - A .Hawaii .. - G.GHawaii .. - .....•.....................................Johnson Island .. - T.Johnson IslandJohnson IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandAmerican SamoaAmerican SamoaTahitiTahitiTahiti, CookTahitiTahitiTahitiTahitiTahitiTahitiCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsFijiFijiFijiF'ijiFiji

C. multicinctusC. multi,C. puncta.C. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. m'ulticinct usC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewens.1.sC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisc. pelewensisC. pelewensisC. pelewensisc. pelewensisc. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensis

176

C. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. pelewensis FijiC. punctatofasciatus GuamC. punctatofasciatus GuamC. punctatofasciatus GuamC. punctatofasciatus GuamC. punctatofasciatus GuamC. punctatofasciatus GuamC. punctatofasciatus GuamC. punctatofasciatus GuamC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctato fasciat lIS PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus PhilippinesC. punctatofasciatus IndonesiaC. punctatofasciatus IndonesiaC. punctatofasciatus IndonesiaC. punctatofasciatus IndonesiaC. punctatofasciatus IndonesiaC. punctatofasciatus IndonesiaC. punctatofasciatus IndonesiaC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus PalauC. punctatofasciatus Palau

GG-GA A A T GG.GG-GA.C A A T AGGG-.A AA T GA.GG-.A A T A.GG- .A G. M T.C A.GG-.A A T.C .. GA..G-.A TM T.C .. CT.GG-.A M T A.GG- .A.C A A T GA.GG-.A A A T, A.GG- . A•••.• A ••••••••••••••••••••••••••• T •••.. A..G-.A TM T.C .. GG..G- .A.C M T A.GG- .A A T.C .. GA..G-.A G.A G T GA.GG-.A TM T,C .. GT.GG- .A G.M T A.GG-.A A T.CT .. T.GG- .A.C A A T GA.GG-.A M T A..G-.A A A T ??

GG- T Po.A • • • • • • • • • • • • • • • • • • • • • • • • • • T • C •• CT.

G .-.A .•••• CA T A..G-.A.C M T A...G- .A.C .. TM T.C T.G. - .A.C A A T.C .. GA.GG-.A TM T A.GG- .A. C A A T GG.GG-.A A , T A.GG-.A CAG T A.GG-.A TA T.C .GG- .A A T GAGGG- .A. C A A T GA.GG- •••.••. M ...••.•••••••....••••••.•. T •C ..• T .GG- T.A T.C .. G..GG-.A TM T.C .. GT.GG- M T.C .. GT.G. - .A G.M G T.C AGG.-.A M T A..G-.A.C .. TAA T.C T.GG- .A M G T A.GG- TM T.C T.GG-.A M G T A.GG-.A M T A..G- .A A A T GA.G.-.A TM T TT.GG-.A G.A T G.GG-.A TA T.. T .. A..G-.A G.A T A.G. - . A M T . C.. GA ..G-.A G.AA T .C A.GG-.A TM T GA.GG- .A G.M T.C A.GG-.A TM T.C A.GG-·.A M T GA.

GG-.A AA G T A.GG-.A A T A.GG-.A A T GA.GG- . A CA T GA..G-.A A T A.GG-.A TM T .C .. GT.GG-.A ••.• TM ••.•..•••..•.•...••....•.. T .•••. G.GG- M T.C .. GT.

177

C. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. punctatofasciatusC. guttatissimusC. guttatissimusC. guttatissimus

PalauPalauPalauPalauPalauPalauPalauPalauMauritius Is.Mauritius Is.Mauritius Is.

GG- .A.C •• TA •••••••••••••••••••••• A •••• T •••. GA.

GG- .A.C ••• A•••••••••••••••••.••••••••• T •..•• A.

GG-.A ••••• M •••••••••••••••••••••••••. T .•••. A.GGG• A•••• TA•••••••••.•••••.••.•••••••• T •C •. GA•

•• G.A ••••• M •••••••••••.•••••••••••.•• T.C •• GA.

GG-.A •••• TM •••••••••••••••••.•••••••• T •••.• G..G- .A •••• TA ••••••.••••••••••.•••• A•••• T •••. GA.

.G-.A ••••• M ••••••••••••••.••••••••••• T ••••• A.GG- . A•••• TTA •••••••••••••••••••••••••• T .••. TT .G.-.A ••••• TA••••••••.••••••••••••••••• T ••.. TT.G.-.A •••• TM ••••••••.•••••.••.•••.•••. T •.•• TT.

178

APPENDIX B <Continued)Control region sequences.

Species

C. multicinct usC. mUlti,C. puncta.C. multicinctusC. multicinctusC. multicinct usC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinct usC. mu Lt icinct usC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. multicinctusC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC.' pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensisC. pelewensis

Location

Hawaii (2)Hawaii (2), PhilHawaiiHa(2), JI (1) ,MI (2)HawaiiHa (1), J. I • (2) ,

HawaiiHawaiiHa, M.I.(2),HawaiiHawaiiHawaiiJohnson IslandJohnson IslandJohnson IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandMidway IslandAmerican SamoaAmerican SamoaTahitiTahitiTahiti, Cook Is.TahitiTahitiTahitiTahitiTahitiTahitiCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook IslandsCook I s landsCook IslandsCook IslandsCook IslandsCook IslandsCook Ls LandsCook IslandsFi jiFi jiFijiFijiFiji

Sequence

195TGATAACGGTTAT

... C .

... C .

... C .

... C ..

••• C •••••••••

... C ..

... C ..••• C ..

••• C •••••••••

••• C .

••• C ..

••• C ..••• C ..

... C ..

••• C .

... C ..

....... A.....

?????????????

?????????????

..... .. .. C ...

..... G.......

C .

179

C. pelewensis Fiji .. .......... 0 •••

C. pelewensis Fiji . .... G •...•..

C. pelewensis Fiji . ............C. pelewensis Fiji . ........ c ...C. pelewensis Fiji . ............C. pelewensis Fiji ..0 ..........C. pelewensis Fiji . ............C. pelewensis Fiji . ............C. pelewensis Fiji . ............C. pelewensis Fiji ..G ..........

C. pelewensis Fiji . ............C. pelewensis Fiji • •••••••• 0 •••

C. pelewensis Fiji . .....................C. pelewensis Fiji .. ........................C. pelewensis Fiji .. .......................C. pelewensis Fiji .. ......................C. punctatofasciatus Guam .....................C. punctatofasciatus Guam .. ......................C. punctatofasciatus Guam .. .......................C. punctatofasciatus Guam . ...................C. punctatofasciatus Guam ......................C. punctatofasciatus Guam ........................C. punctatofasciatus Guam .......................C. punctatofasciatus Guam ...................C. punctatofasciatus Philippines .. .....................C. punctatofasciatus Philippines .. ......................C. punctatofasciatus Philippines .. .....................C. punctatofasciatus Philippines ....................C. punctatofasciatus Philippines .. ....................C. punctatofasciatus Philippines .. .......................C. punctatofasciatus Philippines .. .......................C. punctatofasciatus Philippines ?????????????C. punctatofasciatus Philippines .. ...... A ....C. punctatofasciatus Philippines .. ...................C. punctatofasciatus Philippines .. ....................C. punctatofasciatus Philippines . ....................C. punctatofasciatus Philippines .. ...................C. punctatofasciatus Philippines ..?? .................C. punctatofasciatus Philippines . ..................C. punctatofasciatus Philippines .. .....................C. punctatofasciatus Philippines .. .....................C. punctatofasciatus Philippines .. .....................C. punctatofasciatus Philippines .. ..................C. punctatofasciatus Philippines .. ....................C. punctatofasciatus Indonesia .. .....................C. punctatofasciatus Indonesia .. ............C. punctatofasciatus Indonesia . ............C. punctatofasciatus Indonesia . ............C. punctatofasciatus Indonesia .............C. punctatofasciatus Indonesia . ............C. punctatofasciatus Indonesia .............C. punctatofasciatus Palau . ............C. punctatofasciatus Palau . ............C. punctatofasciatus Palau .............C. punctatofasciatus Palau . ............C. punctatofasciatus Palau .............C. punctatofasciatus Palau .............C. punctatofasciatus Palau . ............\.,. punctatofasciatus Palau . ............C. punctatofasciatus Palau .............C. punctatofasciatus Palau . ............C. punctatofasciatus Palau .............C. punctatofasciatus Palau . ............

180

C. punctatofasciatus Palau .. ........................C. punctatofasciatus Palau •.••• Co 0 •••••

C. punctatofasciatus Palau .. ........................C. punctatofasciatus Palau .. ........................C. punctatofasciatus Palau ••••• 0 oA .••••C. punctatofasciatus Palau .. ........................C. punctatofasciatus Palau .. ........................C. punctatofasciatus Palau .. ........................C. guttatissimus Mauritius Is. 0 ........................

C. guttatissimus Mauritius Is. .. ........................C. guttatissimus Mauritius Is. .. ........................

181

APPENDIXC

Draft 9 May 1994

Allozyme variation among recently formed Pacific Butterflyfishes

(Chaetodontidae)

W. Owen McMillant, Lee Weigt+§, and Stephen R. Palumbi"

tDepartment of Zoology and

Kewalo Marine Laboratory

University of Hawaii

Honolulu, Hawaii 96822

(808)539-7311

(808)599-4817(fax)

owenm@uhunix.uhcc.hawaiLedu

palumbi@uhunix.uhcc.hawaiLedu

tSmithsonian Tropical Research Institute

Apartado 2072

Balboa, Republic of Panama

(507)28-0980

(507)28-0516(fax)

§Alternate Mailing Address: Smithsonian Tropical Research Institute

Unit 0948, APO, AA 34002-0948 USA

182

ABSTRACT

Three closely related butterflyfishes (Chaetodontidae), that differ in

color pattern, partition the tropical west Pacific into large allopatric ranges.

Variation among species at 31 allozyme loci mirrors the pattern of variation

observed in mitochondrial DNA. Genetic differences among the three species

are slight, consistent with their recent differentiation. Similar to the pattern

in mtDNA only the Hawaiian and Johnston Island endemic, Chaetodon

multicinctus, shows significant of genetic differentiation at allozyme loci

(Nei's D=0.04). Genetic differences between C. pelewensis and C.

punctatofasciatus are slight (Nei's D=0.002), consistent with very high

migration rates (Nm>14) between these two species at locales separated by

over 7500 kilometers.

183

INTRODUCTION

Three allopatric butterflyfishes, Chaetodon multicinctus, C. pelewensis,

and C. punctatofasciatus differ primarily with respect to color pattern.

Previous mitochondrial DNA (mtDNA) analysis suggested that all three

species diverged from the Indian Ocean, C. guttatissimus, between 800,000 and

1,600,000 years ago and from each other between 300,000 and 800,000 years ago

(McMillan and Palumbi, submitted). Yet, these data confirm the evolutionary

distinctiveness of only C. multicinctus. MtDNA variation is distributed

randomly between C. pelewensis and C. punctatofasciatus with little

indication of significant subdivision across the combined range of both

species.

Although much of the observed mtDNA variation could have

predated speciation, the genealogical and geographic distribution of mtDNA

variation argues that hybridization and inter-specific introgression is a more

probable explanation for the failure of the mtDNA gene tree to reflect species

boundaries (McMillan and Palumbi, submitted). The mtDNA has been

observed to move across species boundaries even when there are strong

reproductive barriers (Powell, 1983; Ferris et al., 1983;Spolsky and Uzzell,

1984; Lamb and Avise, 1986; Carr et al., 1986; Baker et al., 1989; Dowling et al.,

1989). Theoretical work suggests that the rate of introgression is determined

primarily by the level of gene flow (Takahata and Slatkin, 1984). These

butterfishes have a 40-60 day planktonic larval stage are probably capable of

dispersing gametes over hundreds, if not thousands, of kilometers (Tricas,

1987; Hourigan and Reese, 1987; Leis 1989).

To augment our previous mtDNA work, we provide a description of

the variation in this group at allozyme loci. Allozymes provide a rough

184

meter with which to gauge the extent of gene flow among species at

presumably neutral nuclear markers (however see Karl and Avise, 1992).

Because the markers are more likely to be linked to loci and co-adaptive gene

complexes under strong selective pressure they may show a very different

pattern of introgress relative to mtDNA (Barton, 1983; Takahata and Slatkin,

1984). The extent to which these two different genetic markers show similar

or different patterns of introgression sheds light on the evolutionary and

genetic cohesion of these species, and provides a better foundation upon

which to examine the significance of interspecific color pattern differences in

this group.

185

MATERIALS AND METHODS

Approximately 30 individuals from each of the three species were

examined for genetic differences at allozyme loci using standard starch gel

electrophoresis (Murphy, et al., 1990). All thirty individuals were taken from

phenotypically homogeneous populations at the approximate center of each

species' geographic distribution. The sampling locations were as follows: for

C. multicinctus - Molokai, Hawaii (n=32); for C. punctatofasciatus - the

Philippines (n=30); for C. pelewensis - Viti-Levu, Fiji (n=27).

The heart, liver, eyes, brain and a small portion of skeletal muscle were

removed from each individual and homogenized in two volumes of

grinding buffer (0.25M sucrose, 2% phenoxyethanol) and centrifuged at 14,000

rpm for 10 min at 4° C. The resulting supernatant was removed, absorbed

onto filter wicks, and electrophoresed through slabs of 11% polymerized

starch using four different running buffers (table 1). Thin slices of starch gels

were stained for 19 different enzyme systems that yielded a total of 31 distinct

loci (table 1). Histological staining methods followed Aebersold et al. (1987).

Following staining, gels were photographed and allelic variation was scored

alphabetically in order of decreasing anodal mobility of the protein produced.

Allele frequencies at polymorphic loci were determined and genotype

frequencies at polymorphic loci were tested for conformity to Hardy­

Weinberg expectations. Nei's (1978) unbiased genetic distances (D) .were

calculated between all species. The significance of the observed differences

between jackknifed average genetic differences was determined as outlined by

Muller and Ayala (1982), using a program provided by Lessios (1990).

In addition, F-statistics were computed between the three species.

Because the FST approach is based on an explicit island model of population

186

genetics, it permits an estimate of migration between geographic regions

based on the relationship, Nm=(1-FST)/4FST where Nm is the effective

number of individuals within each locality and m is the migration rate per

generation (Wright, 1951).

187

RESULTS

Of the thirty-one loci examined, 12 were monomorphic across all three

species and 9 showed only slight variation. Significant polymorphism (major

allele at less than 0.95 frequency) was evident at the remaining ten loci (table

2). Significant deviation (p<0.05) of genotype frequencies from Hardy­

Weinberg equilibrium within each of the three localities was evident at only

the PEP A locus in Fijian populations of C. pelewensis. However, because at

least one deviation at the 0.05% level was expected to occur by chance in our

39 chi-squared tests, we attributed this observation to sampling artifact rather

than a biologically meaningful pattern.

Overall genetic differences among species at allozyme loci were very

small and similar to levels previously reported among conspecific

populations of marine fishes across the Pacific (Winans, 1980; Rosenblatt and

Waples, 1986; Planes, 1993). There were no fixed allelic differences among the

three species and only slight shifts in the frequencies of alternate alleles at

different loci (table 2). Of the three, C. multicinctus was genetically the most

distinct with Nei's D of 0.040 and 0.046 from C. punctatofasciatus and C.

pelewensis, respectively (table 3). At two loci, GPI 2 and MDH 2, the

dominant allele in Hawaiian populations of C. multicinctus was present in

very low frequency within C. punctatofasciatus and C. pelewensis. At other

polymorphic loci, C. multicinctus showed much smaller levels of variation

and had significantly lower levels of heterozygozity relative to it's two

siblings (table 4). Estimates of effective migration rates between C.

multicinctus and its two siblings were less than an Nm value of 1.0, a level

thought to prevent the accumulation of genetic differences by genetic drift

(Slatkin, 1987) (table 5).

188

By contrast, C. punctatofasciatus and C. pelewensis were nearly

identical at allozyme loci. Nei's D between these two species was very slight

(ca., 0.002) and not significantly different from zero. The largest difference in

the allele frequency occurred at the MDH2 loci. Within C. pelewensis, the A

allele occurred at a frequency of nearly 34% within Fijian populations. In

contrast, in Philippine populations of C. punctatofasciatus that same allele

reached a frequency of less than 10% (table 3). At other polymorphic loci,

allele frequencies were nearly identical despite the nearly 7500 kilometers

separating the two species. As a result, FST values between these two

locations were very small, 0.019 (0.023,0.009: 95% confidence interval),

suggesting a very high effective per generation migration rate (Nm=13)

between the eastern and western Pacific (table 5).

189

DISCUSSION

The patterning of variation at allozyme loci among these similar

butterflyfishes mirrors the pattern of mtDNA variation. Genetic

differentiation is slight among species consistent with a recent formation of

the group. None-the-less, similar to the pattern of mtDNA variation, the

Hawaiian/Johnston Island endemic, C. multicinctus, exhibits clear genetic

differences at allozyme loci suggesting low levels of intra-specific gene flow

(table 5) (McMillan and Palumbi, submitted). In addition, the lower levels of

heterozygozity in this species at allozyme loci are paralleled by much lower

levels of mtDNA variation in C. multicinctus relative to C. pelewensis or C.

punctatofasciatus (McMillan and Palumbi, submitted).

By contrast, both allozyme and mtDNA variation is distributed

randomly between Fijian populations of C. pelewensis and Philippine

populations of C. punctatofasciatus. The genealogy of mtDNA variation

suggests that these two species diverged in allopatry between 300,000 and

800,000 years ago and the present homogenization of mtDNA variation is the

result of high levels of high levels of hybridization and associated

introgression of mtDNA. Because allozyme differences accumulate much

slower than mtDNA differences, some of the biochemical similarity between

these two species may reflect historical association (Wilson et al., 1985; Reeb

and Avise, 1990). However, the strong concordance between estimates of

migration between Fiji and Philippine populations based on allozymes and

those based on mtDNA argues for contemporary movement of nuclear, as

well as mtDNA, variation across species boundaries (table 5). Long distance

dispersal during the 40-60 planktonic feeding larval stage probably accounts

for the widespread homogenization of genetic variation in these two species.

190

This genetic pattern indicates that, despite their species-level distinction, C.

punctatofa.sciatus and C. pelewensis are tightly coupled evolutionary.

Despite this conclusion, color pattern distinctions remain true across

the vast majority of each species' range (Blum, 1989). Levels of gene flow

between Philippine populations of C. punctatofasciatus and Fijian

populations of C. pelewensis are far higher than levels expected to prevent

the accumulation of neutral differences by genetic drift (Slatkin, 1987). Yet,

these populations show completely different distributions of phenotypic

variation, with all 30 individuals collected from reefs around the Philippines

possessing a typical C. punctatofasciatus color pattern and all individuals

examined from the Fiji possessing a typical C. pelewensis color pattern. The

obvious implication of this discrepancy is that strong selection on color

pattern is preventing the breakdown of species-specific color patterns in the

background of potentially homogenizing levels of gene flow.

191

LITERATURE CITED

Baker, R J., S. K. Davis, R. D. Bradley, M. J. Hamilton, and R. A. Van der

Bussche. 1989. Ribosomal-DNA, mitochondrial DNA, chromosomal,

and allozymic studies on a contact zone in the pocket gopher, Geomys.

Evolution 43:63-75.

Barton, N. H. 1983. Multilocus clines. Evolution 37:454-471.

Blum, S. D. 1989. Biogeography of the Chaetodontidae: an analysis of

allopatry among closely related species. Envir. Biol. Fish. 25: 9-31.

Carr, S. M., S. W. Ballinger, J. N. Deer, L. H. Blankenship, and J. W. Bickham.

1986. Mitochondrial DNA analysis of hybridization between sympatric

white-tailed deer and mule deer in west Texas. Proc. Natl. Acad. Sci. USA

83:9576-9580.

Dowling, T. E., G. R. Smith, and W. M. Brown. 1989. Reproductive isolation

and introgression between Notropis cornutus and Notropis

chrysocephalus (family Cyprinidae): comparison of morphology,

allozymes, and mitochondrial DNA. Evolution 43:620-643.

Ferris, S. D., R. D. Sage, C.-H. Huang, J. T. Nielsen, U. Ritte, and A. C. Wilson.

1983. Flow of mitochondrial DNA across a species boundary. Proc. Natl.

Acad. Sci. USA 80:2290-2294.

Hourigan, T. F. and E. S. Reese. 1987. Mid-ocean isolation and the evolution

of Hawaiian reef fishes. Trends Ecol. Evol. 2:187-191.

192

Lamb and Avise, 1986. Directional introgression of mitochondrial DNA in a

hybrid population of tree frogs: the inference of mating behavior. Proc.

Nat!. Acad. Sci. USA 83:2526-2530.

Leis, J. M. 1989. Larval biology of butterflyfishes (Pisces, Chaetodontidae):

what do we really know? Envir. BioI. Fish. 25:87-100.

Lessios, H. A. 1990. A program for calculating Neils genetic distances and

their jackknifed confidence intervals. J. Hered. 81:490.

Mueller, L. D., and F. J. Ayala. 1982. Estimation and interpretation of genetic

distance in empirical studies. Genet. Res. Camb. 40:127-137.

Nei, M. 1978. Estimation of average heterozygosity and genetic distance from

a small number of individuals. Genetics 89:583-590.

Nishida, M. and S. J. Lucas. 1988. Genetic differences between geographic

populations of the crown-of-thorn starfish throughout the Pacific region.

Marine Biology 98: 359-368.

Karl, S. A. and J. C. Avise. 1992. Balancing selection at allozyme loci in

oysters: implications from nuclear RFLPSs. Science 256:100-102.

Planes, S. 1993. Genetic differentiation and restricted larval dispersal of the

convect surgeonfish Acanthurus triostegus in French Polynesia. Mar.

Ecol. Prog. Ser. 98:2237-246.

Powell, J. R. 1983. Interspecific cytoplasmic gene flow in the absence of

nuclear gene flow: evidence from Drosophila. Proc. Natl. Acad. Sci. USA

80:492-495.

193

Reeb, c. A. and J. C. Avise. 1990. A genetic discontinuity in a continuously

distributed species: mitochondiral DNA in the American Oyster,

Crassostrea virginica. Genetics 124(2):397-406.

Rosenblatt, R. H. and R. S. Waples. 1986. A genetic comparison of allopatric

populaitons of shore fish species from the eastern and central Pacific

Ocean: dispersal of vicariance? Copeia 1986:275-284.

Spolsky, C. and T. Uzzell. 1984. Natural interspecies transfer of

mitochondrial DNA in amphibians. Proc. Natl. Acad. Sci. USA 81:5802­

5805.

Slatkin, M. 1987. Gene flow and the geographic structure of natural

populations. Science 236:787-792.

Takahata, N. and M. Slatkin. 1984. Mitochondrial gene flow. Proc. Natl.

Acad, Sci. USA 81:1764-1767.

Tricas, T. C. 1987. Life history and behavioral ecology of Chaetodon

multicinctus. Ph. D. dissertation, University of Hawaii, Honolulu.

Wilson, A. c, R. L. Cann, S. M. Carr, M. George, U. B. Gyllensten, K. M.

Helm-Bychowski, R. G. Higuchi, S. R. Palumbi, E. M. Pranger, R. D. Sage,

and M. Stoneking. 1985. Mitochondrial DNA and two perspectives on

evolutionary genetics. Biol. J. Linn. Sci. 26:375-400.

'Alright, S. 1951. The genetical structure of populations. Ann. Eugen. 15:322­

354.

194

Winans, G. A. 1980. Geographic variation in the milkfish, Chanos chanos. 1.

Biochemical evidence. Evolution 34:558-574.

195

Table 1: Enzymes, enzyme abbreviations, Enzyme Commission numbers,

number of loci, and gel buffers used in our electrophoretic analysis of

Chaetodon punctatofasciatus, C. pelewensis, and C. multicinctus.

Enzyme Abbr. EC No Loci BuffertAspartate aminotransferase AAT 2.6.1.1 2 TC,CTCreatine kinase CK 2.7.3.2 4 R WDiaphorase DIA 1.6.2.2 1 R WEsterase EST 3.1.1.1 3 R WGlucose-6-phosphate dehydrogenase G6PD 1.1.1.49 1 TVDGlucose phosphate isomerase GPI 5.3.1.9 2 CTlsocitrate dehydrogenase ICD 1.1.1.42 2 TCLactate dehydrogenase LDH 1.1.1.27 2 TVDMalate dehydrogenase MDH 1.1.1.37 2 CTMalate dehydrogenase (NADP+) MDPH 1.1.1.40 2 TCMannose phosphateisomerase MPI 5.3.1.8 1 TVDpeptidase B PEPB 3.4.13.,. 2 R Wpeptidase 5 PEPS 3.4.13." 1 R Wpeptidase A PEPA 3.4.13." 1 R WPhosphoglucomutase PGM 2.7.5.1 1 TC, CTPhosphogluconate dehydrogenase PGDH 1.1.1.44 1 CTSuperoxide dismutase SOD 1.15.1.1 1 TVDTriose phosphate isomerase TPI 5.3.1.1 2 R Wt CT: amine-citrate, pH 6.0 (Clayton and Tretiak, 1972); TC: tris-citrate, pH 8.0(Selander, et al., 1971); RW: tris-citrate, pH 8.5 (Ridgway, et al., 1970); TVB: tris­versene-borate.. pH 8.0 (Selander, et al., 1971).

196

Table 2: Allele frequencies for the 19 variable loci in Chaetodon multicinctus.

C. punctatofasciatus, and C. pelewensis.

SpeciesLocus allele C. multi. C. puncta. C. pele.AAT1 A 0.043

B 0.020C 0.980 1.000 0.957

AAT2 A 0.018B 1.000 0.982 1.000

DIA A 0.909 1.000 0.978B 0.091 0.022

EST1 A 0.045 0.161 0.177B 0.955 0.732 0.694C 0.089 0.097D 0.018 0.016E 0.016

EST2 A 0.076B 0.924 0.982 1.000C 0.018

EST3 A 1.000 0.963 0.968B 0.037 0.032

G6PD A 0.280 0.278 0.261B 0.700 0.667 0.696C 0.020 0.056 0.043

GPIl A 0.985 0.946 0.935B 0.015 0.032C 0.016D 0.054 0.016

GPI2 A 0.036B 0.607 0.758C 0.048D 1.000 . 0.357 0.194

197

Table 2 (continued)

ICD2 A 0.879 1.000 1.000B 0.121

MEl A 0.020B 0.980 0.875 0.957C 0.100 0.043D 0.025

ME2 A 0.980 0.975 0.978B 0.020 0.025 0.022

MDHI A 0.016B 1.000 1.000 0.984

MDH2 A 0.985 0.089 0.339B 0.015 0.911 0.661

PEPB2 A 1.000 0.964 1.000B 0.018C 0.018

PEPS A 0.030 0.036 0.016B 0.970 0.964 0.984

PEPA A 0.271 0.273B 1.000 0.708 0.727C 0.021

PGM3 A 0.016B 1.000 0.982 0.984C 0.018

SOD A 0.018B 1.000 0.982 1.000

198

Table 3: Matrix of Nei (1978) unbiased genetic distance coefficients (D) based

on allelic variation at 31 presumptive loci.

C. punctatofasciatusPhilippines

C. punctatofasciatus

C. pelewensis 0.046

C. pelewensisFiii

C. multicinctusHawaii

C. multicinctus 0.040 0.002

199

Table 4: Summary of genetic variability at 31 loci in C. multicinctus, C.

pelewensis, and C. punctatofasciatus (standard errors in parentheses).

C. multicinctus(Hawaii)

0.042(0.016)

0.041(0.015)

Direct­count

Mean heterozygosityHardy­

Weinbergexpectedt

35.5

Percentageof loci

polymorphict1.4

(0.1)

Mean no.of allelesper locus

31.2(0.7)

Mean samplesize perlocusSpecies

C. punctato. 26.1 1.7(Philippines) (0.6) (0.2)

C. pelewensis 29.4 1.7(Fiji) (0.6) (0.2)

48.4

45.2

0.094(0.30)

0.080(0.026)

0.090(0.028)

0.088(0.029)

t A locus was considered polymorphic if more than one allele was detected.:I: Unbiased estimate (see Nei, 1978)

200

Table 5: FST values and values of Nm (in parentheses) among species

estimated from both allozyme variation (below diagonal) and mitochondrial

control-region sequences (above diagonal). For mtDNA data see McMillan

and Palumbi, submitted.

C. punctatofasciatusPhilippines

C. punctatofasciatus

C. pelewensisFiji

0.018 (27)

C. rnulticinctusHawaii

0.479 «1)

C. pelewensis

C. multicinctus

0.019 (13)

0.253 «1) 0.230 (<1)

201

0.480 «1)

REFERENCES

Abbott, R T. 1982. Kingdom of the seashell. CrownPress, New York.

Acharyya, K. S. and P. K. Basu. 1993. Toba ash on the Indian subcontinent

and its implications for correlation of late Pleistocene alluvium.

Quaternary Research 40:10-19.

Allen, G. R 1972. The anemonefishes, their classification and biology. T.F.H.

publications, Neptune City.

Allen, G. R 1980. Butterfly and angelfishes of the world, vol 2. Wiley­

Interscience, New York.

Allen, G. R, and R Swainston. 1992. The fishes of New Guinea. Kristen

Press, Inc., Madang.

Allmon, W. D., G. Rosenberg, R. W. Portell, K. S. Schindler. 1993. Diversity

of Atlantic coastal plain mollusks since the Pliocene. Science 260:1626­

1229.

Altman, P. L. and D. S. Dittmer. 1972. Biology data book. Fed. Am. Soc. Exp.

Biol., 2nd Ed., Bethesda.

Anderson, G. R V., A. H. Ehrlich, P. R. Ehrlich, J. D. Roughgarden, B. C.

Russel, and F. H. Talbot. 1981. The community structure of coral reef

fishes. Am. Nat. 117:476-495.

Audley-Charles, M. G. 1981. Geological history of the region of Wallace's

line. Pp. 24-35 in T. C. Whitmore, ed. Wallace's line and Plate tectonics.

Clarendon Press, Oxford.

202

Avise, J. C. 1992. Molecular population structure and the biogeographic

history of a regional fauna: a case history with lessons for conservation

biology. Oikos 63:62-76.

Avise, J. C. and N. C. Saunders. 1984. Hybridization and introgression among

species of sunfish (Lepomis): analysis by mitochondrial DNA and

allozyme markers. Genetics 108:237-255.

Avise, J. c.,J. Arnold, R. M. Ball, E. Bermingham, T. Lamb, J. E. Neigel, C. A.

Reeb, and N. C. Saunders. 1987. Intraspecific phylogeography: The

mitochondrial DNA bridge between population genetics and systematics.

Ann. Rev. Ecol. Syst. 18: 489-522.

Avise, J. c., R. M. Ball, and J. Arnold. 1988. Current versus historical

population sizes in vertebrate species with high gene flow: a comparison

based on mitochondrial DNA lineages and inbreeding theory for neutral

mutations. Mol. BioI. Evol. 5:331-344.

Baker, R. J., S. K. Davis, R. D. Bradley, M. J. Hamilton, and R. A. Van der

Bussche. 1989. Ribosomal-DNA, mitochondrial DNA, chromosomal,

and allozymic studies on a contact zone in the pocket gopher, Geomys.

Evolution 43:63-75.

Ball, R. M., J. E. Neigel, and J. C. Avise. 1990. Gene genealogies within the

organismal pedigrees of random-mating populations. Evolution 44:360­

370.

203

Ball, R. M., S. Freeman, F. C. James, E. Bermingham, and J. C. Avise. 1988.

Phylogeographic population structure of Red-winged Blackbirds assessed

by mitochondrial DNA. Proc. Nat1. Acad. Sci. USA 85:1558-1562.

Barlow, G. 1963. Ethology of the Asian teleost Badis badis. II. Motivation

and signal value of the color patterns. Anim. Behav. 11:97-105.

Barton N. H, and K. Gale. 1993. Genetic analysis of hybrid zones. Pp 13-45 in

R. Harrison, ed. Hybrid zones and the evolutionary process. Oxford

University Press, New York.

Barton, N. H. 1983. Multilocus clines. Evolution 37:454-471.

Barton, N. H. and G. M. Hewitt. 1981. The genetic basis of hybrid inviability

between two chromosomal races of the grasshopper Podisma pedestris.

Heredity 47:367-383.

Barton, N. H. and G. M. Hewitt. 1985. Analysis of hybrid zones. Ann. Rev.

Eco1. Syst. 16:113-148.

Bauchot, R., J. Ridet, and M. Bauchot. 1989. The brain organization of

butterflyfishes. Environ. Bio. Fishes. 25:205-219.

Belasky, P. and B. Runnegar. 1993. Biogeographic constraints for tectonic

reconstructions of the Pacific region. Geology 21:979-982.

Bermingham E. and H. Lessios. 1993. Rate variation of protein and

mitochondrial DNA evolution as revealed by sea urchins separated by the

Isthmus of Panama. Proc. Nat1. Acad. Sci. USA 90:2734-2738.

204

Bermingham, E., and J. C. Avise. 1986. Molecular zoogeography of

freshwater fishes in the southeastern United States. Genetics 113:939-965.

Bert, T. M. and R. G. Harrison. 1988. Hybridization in western Atlantic stone

crabs (genus Menippe): Evolutionary history and ecological context

influence species interactions. Evolution 42:528-544.

Blum, S. D. 1988. Osteology and phylogeny of the Chaetodontidae (Pisces:

Perciformes). Ph. D. Dissertation, University of Hawaii, Honolulu.

Blum, S. D. 1989. Biogeography of the Chaetodontidae: an analysis of

allopatry among closely related species. Envir. Biol. Fish. 25: 9-31.

Brockman, H. J. 1973. The function of poster coloration in the beaugregory,

Eupomacenrus leucostictus (Pomacentridae, Pisces). Z. Tierpsychol. 33:13­

34.

Brown, J. R., A. T. Beckenbach, and M. J. Smith. 1993. Intra-specific DNA

sequence variation of the mitochondrial control-region of the white

sturgeon (Acipenser transmontanus). Mol. BioI. Evol, 10:326-341.

Brown, W. M., E. M. Pranger, A. Wang, and A. C. Wilson. 1982.

Mitochondrial DNA sequences of primates: tempo and mode of

evolution. J. Mol. Evol. 18:225-239.

Burgess, W. E. 1978. Butterflyfishes of the world. T. F. H. Publications,

Neptune City.

205

Bushnell, P. G., P. L. Lutz, and S. H. Gruber. 1989. The metabolic rate of an

active, tropical elasmobranch, the lemon shark (Negaprion brevirostris).

Exp. BioI. 48:279-283.

Butcher, G. S. and S. Rohwer. 1987. The evolution of conspicuous and

distinctive coloration for communication in birds.

Butlin, R. 1989. Reinforcement of premating isolation. Pp 158-179 in D. Otte

and J. A. Endler, eds. Speciation and its consequences. Sinauer Press,

Massachusetts.

Carr, S. M., S. W. Ballinger, J. N. Deer, L. H. Blankenship, and J. W. Bickham.

1986. Mitochondrial DNA analysis of hybridization between sympatric

white-tailed deer and mule deer in west Texas. Proc. Natl. Acad. Sci. USA

83:9576-9580.

Charlton, T. R. 1986. A plate tectonic model of the eastern Indonesia

collision zone. Nature 319:394-396.

Chesner, C. A., W. 1. Rose, A. Deino, R. Drake and J. A. Westgate. 1991.

Eruptive history of earth's largest quaternary caldera (Toba, Indonesia)

clarified. Geology 19:200-302.

Cheuy, J. M., D. K. Rea, and N. G. Pisias. 1987. Late Pleistocene

paleoclimatology of the central equatorial Pacific: a quantitative record of

eolian and carbonate deposition: 28:323-339.

Choat, J. H. and Bellwood 1991. Reef fishes: their history and evolution. Pp.

39-66 in P. Sale, ed. The ecology of fishes on coral reefs. Academic Press,

Inc. San Diego.

206

Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science

199:1302-1310.

Cracraft, J. and R. O. Prum. 1988. Patterns and processes of diversification:

speciation and historical congruence in some neotropical birds.

Evolution 42:603-620.

Crame, J. A. 1986. Late Pleistocene molluscan assemblages from the coral

reefs of the Kenya coast. Coral Reefs 4:183-196.

Crane, J. 1975. Fiddler crabs of the world (Oeypodidaeg: genus Uca).

Princeton University Press, Princeton.

Croizat, L. G., G. Nelson and D. E. Rosen, 1974. Centers of origins and related

concepts. Syst. Zool. 23: 265-287.

Desalle, R. T. Freedman, E. M. Prager, and A. C. Wilson. 1987. Tempo and

mode of sequence evolution in mitochondrial DNA of Hawaiian

Drosophila. J. Mol. Evol. 26:157-164.

Di Rienzo, A. and A. C. Wilson. 1991. The pattern of mitochondrial

variation is consistent with an early expansion of the human population.

Proc. Natl. Acad. Sci. USA 88:1597-1601.

Donaldson, T. J. 1986. Distribution and species richness patterns of Indo­

West Pacific Cirrhitidae: support for Woodland's hypothesis. Pp 623-628

in T. Uyeno, R. Arai, T. Taniuchi and K. Matsuura, eds. Indo-Pacific fish

biology, proceedings of the second international conference on Indo­

Pacific fishes. Ichthyological Society of Japan, Tokyo.

207

Dowling, T. E., G. R. Smith, and W. M. Brown. 1989. Reproductive isolation

and introgression between Notropis comutus and Notropis

chrysocephalus (family Cyprinidae): comparison of morphology,

allozymes, and mitochondrial DNA. Evolution 43:620-643.

Ehrlich, P. R., F. H. Talbot, B. C. Russell, and G. R. V. Anderson. 1977. The

behavior of chaetodontid fishes with special reference to Lorenz's "poster

colouration" hypothesis. J. Zoo1. London 183:213-228.

Ekman, S. 1953.. Zoogeography of the sea. Sidgwick and Jackson, London.

Endler, J. A. 1977. Geographic variation, speciation, and clines. Princeton

University Press, Princeton, New Jersey.

Endler, J. A. 1980. Natural selection on color pattern in Poecilia reticulata.

Evolution 34:76-91.

Endler, J. A. 1982. Problems in distinguishing historical from ecological

factors in Biogeography. Amer. Zoo1. 22:441-452.

Excoffier, L., P. E. Smouse, and J. M. Quattro. 1992. Analysis of molecular

variance inferred from metric distances among DNA haplotypes:

application to human mitochondrial DNA restriction data. Genetics

131:479-491.

Faith, D. P. 1991. Cladistic permutation tests for monophyly and

nonmonophyly. Syst. Zoo1. 40:366-375.

Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the

bootstrap. Evolution 39:783-791.

208

Felsenstein, J. 1993. PHYLIP: phylogenetic inference package, version 3.5.

University of Washington, Seattle.

Ferris, S. D., R. D. Sage, C.-H. Huang, J. T. Nielsen, U. Ritte, and A. C. Wilson.

1983. Flow of mitochondrial DNA across a species boundary. Proc. Natl.

Acad. Sci. USA 80:2290-2294.

Fisher, E. A. 1980. Speciation in the Hamlets (Hypoplectrus: Serranidae)- a

continuing enigma. Copeia 1980:649-659.

Fleminger, A. 1986. The Pleistocene equatorial barrier between the Indian

and Pacific oceans and a likely cause for Wallace's line. Pp. 84-97 in A. C.

Pierrot-Bults, S. van der Spoel, B. J. Aahuraner, and R. K. Johnson, eds.

Pelagic biogeography. UNesco technical papers in marine science. No. 49.

French, S. and B. Robson. 1983. What is a conservative substitution? J. Mol.

Evo1. 19:171-175.

Gochfeld, D. J. 1991. Energetics of a predator-prey interaction: corals and

coral-feeding fishes. Pacific Science 46:246-256.

Gosline, W. A. 1965. Thoughts on systematic work in outlying areas. Syst.

Zoo1. 14:59-61.

Graham, J. H. and J. D. Felley. 1985. Genomic co-adaptation and

developmental stability within introgressed populations of Enneacanthus

glorioslls and E. obesus (Pisces, Centrarchidae). Evolution 39: 104-114.

209

Grudzien, T. A. and W. S. Moore. 1986. Genetic differentiation between the

yellow-shafted and Red-shafted subspecies of the Northern Flicker.

Biochemical Systematics and Ecology 14:451-453.

Grudzien, T. A., W. S. Moore, J. R. Cook, D. Tagle. 1987. Genetic population

structure and gene flow in the northern flicker (Colaptes auratus) hybrid

zone. Auk 104:654-664.

Hailman, J. P. 1979. Environrnentallight and conspicuous colors. Pp.289-354

in E. H. Burtt, [r., ed. The behavioral significance of color. Garland STPM

Press, New York.

Harpending, H. c., S. T. Sherry, A. R. Rogers, and Mark Stoneking. 1993. The

genetic structure of ancient human populations. Current Anthropology

4:483-496.

Harrison, R. G. 1989. Animal mitochondrial DNA as a genetic marker in

population and evolutionary biology. Trends Ecol. Evol, 4:6-11.

Harrison, R. G. 1990. Hybrid zones: windows on evolutionary process.

Oxford Surveys in Evolutionary Biology 7:69-139.

Hewitt, G. M. 1989. The subdivision of species by hybrid zones. Pp 85-110 in

D. Otte and J. A. Endler, eds. Speciation and its consequences. Sinauer

Press, Massachusetts.

Hiatt, R. W. and D. W. Strasburg. 1960. Ecological relationships of the fish

fauna on coral reefs on the Marshall Island. Ecol. Monogr. 30:65-127.

210

Hillis, D. M. and J. J. Bull. An empirical test of bootstrapping as a method for

assessing confidence in phylogenetic trees. Syst. Biol. 42:182-192.

Hillis, D. M., and C. Moritz. 1990. Molecular systematics. Sinauer Press.,

Sunderland, MA.

Hobson, E. S. 1974. Feeding 'relationships of teleostean fishes on coral reefs in

Kona, Hawaii. U. S. Fish. Bull. 77:915-1031.

Hocutt, C. H. 1987. Evolution of the Indian Ocean and the drift of India: a

vicariant event. Hydrobiol. 159:203-223.

Hourigan, T. F. 1987. The behavioral ecology of three species of

butterflyfishes (Family Chaetodontidae). Ph. D. Dissertation, University

of Hawaii, Honolulu.

Hourigan, T. F. 1989. Environmental determents of butterflyfish social

systems. Envir. Biol. Fishes 25:61-78.

Hourigan, T. F. and E. S. Reese. 1987. Mid-ocean isolation and the evolution

of Hawaiian reef fishes. Trends Ecol. Evol. 2:187-191.

Hudson, R. R. 1990. Gene genealogies and the coalescent process. Oxfor.

Surv. Evol. BioI. 7:1-44.

Hudson, R. R., M. Slatkin, and W. P. Maddison. 1992. Estimation of levels of

gene flow from DNA sequence data. Genetics 132:583-589.

Irwin, D. M., T. D. Kocher, and A. C. Wilson. 1991. Evolution of the

cytochrome b gene of mammals. J. Mol. Evol. 32: 128-144.

211

Jackson, J. B., P. Iung, A. G. Coates, L. S. Collins. 1993. Diversity and

Extinction of tropical American mollusks and emergence of the Isthmus

of Panama. Science 260:1624-1626.

Janecek, T. R. and D. K. Rea. 1985. Quaternary fluctuations in the northern

hemisphere trade winds and westerlies. Quater. Res. 24:150-163.

[okiel, P. L. 1990. Transport of reef corals into the great barrier reef. Nature

347:665-367.

Kay, E. A. 1994. Evolutionary radiations in the Cypraeidae. Journal of

Molluscan Studies (in press).

Kelly, C. D. and T. Hourigan. 1983. The function of conspicuous coloration

in chaetodontid fishes: a new hypothesis. Am. Behavior 31:

Kimura, M. 1980. A simple method for estimating evolutionary rate of base

substitution through comparative studies of nucleotide sequences. J. Mol.

Evo!. 16:111-120.

Kishino, H. and M. Hasegawa. 1989. Evaluation of the maximum likelihood

estimate of the evolutionary tress topologies form DNA sequence data,

and the branching order in Hominoidea. J. Mol. Evol. 29:170-179.

Knowlton, N. and J. B. C. Jackson. 1993. Jack-of-all-trades or master of some?

New taxonomy and niche partitioning on coral reefs. Trends Ecol. Evol.

(in press).

Kocher, T. D. and A. C. Wilson. 1991. Sequence evolution of mitochondrial

DNA in humans and chimpanzees: Control-region and a protein-coding

212

region. Pp. 391-413 in S. Osawa and T. Honjo, eds., Evolution of life.

Spring-Verlag, New York.

Kohn, A. J. 1990. Tempo and mode of evolution in Conidae. Malacologia

32:55-67.

Kumar, S., K. Tamura, and M. Nei. 1993. MEGA: molecular evolutionary

genetics analysis, version 1.01. Pennsylvania State University, University

Park.

Ladd, H. S. 1960. Origins of the Pacific island molluscan fauna. Amer. J.

Science. 258:137-150.

Lamb and Avise, 1986. Directional introgression of mitochondrial DNA in a

hybrid population of tree frogs: the inference of mating behavior. Proc.

Nat!. Acad. Sci. USA 83:2526-2530.

Lamb, T., J. M. Novak, and D. L. Mahoney. 1990. Morphological asymmetry

and interspecific hybridization: a case study using hylid frogs. J. Evol,

BioI. 3:295-309.

Lande, R. 1982. Models of speciation by sexual selection of polygenic traits.

Proc. NatI. Acad. Sci. USA 78:3721-3725.

Lansing, W. R. J. and C. C. Brower. 1974. Development of colour patterns in

relation to behavior in Tilapia mossarnbica (Peters). J. Fish BioI. 6:29-41.

Lavery, S. 1993. Ph. D. dissertation. Dept of Zoology, University of

Queensland, Brisbane, Australia

213

Leary, R. F. and F. W. Allendorf. 1989. Fluctuating asymmetry as an indicator

of stress: implications for conservation biology. Trends Ecol. Evol. 4:214­

217.

Leary, R. F., F. W. Allendorf, and K. L. Knudsen. 1985. Developmental

instability and high meristic counts in interspecific hybrids of salmonid

fishes. Evolution 39:1318-1326.

Leis, J. M. 1989. Larval biology of butterflyfishes (Pisces, Chaetodontidae):

what do we really know? Envir. BioI. Fish. 25:87-100.

Littlejohn, M. J. and G. F. Watson. 1985. Hybrid zones and homogamy in

Australian frogs. Ann. Rev. Ecol. Syst. 16:85-112.

Lobel, P. S. 1989. Spawning behavior of Chaetodon multicinctus

(Chaetodontidae); pairs and intruders. Envir. BioI. Fish. 25:125-130.

Lorenz, K. 1967. On aggression. Bantam, New York.

Losey, G. 1988. Behavior events acquisition and analysis system, "BEAST".

Windward Technology, Kaneohe, Hawaii.

Lynch, M. and T. J. Crease. 1990. The analysis of population survey data on

DNA sequence variation. Mol. BioI. Evol. 7:377-394.

Maddison, W. P., and D. R. Maddison. 1992. MacClade: analysis of phylogeny

and character evolution. Sinauer Press, Sunderland.

Mallet, J. L. B. 1993. Speciation, raciation, and color pattern evolution in

Heliconius butterflies: Evidence from hybrid zones. Pp. 226-260 in R.

214

Harrison, ed. Hybrid zones and the evolutionary process. Oxford

University Press, New York.

Mallet, J. L. B. and N. H. Barton. 1989a. Inference from clines stabilized by

frequency-dependent selection. Genetics 122:967-976.

Mallet, J. L. B. and N. H. Barton. 1989b. Strong natural selection in a

warning-color hybrid zone. Evolution 43:421-431.

Marjoram, P. and P. Donnelly. 1994. Pairwise comparisons of mitochondrial

DNA sequences in subdivided populations and implications for early

human evolution. Genetics 136:673-683.

Markow, T. A. and J. P. Ricker. 1991. Developmental stability in hybrids

between the sibling species pair, Drosophila melanogaster and Drosophila

simulans. Genetica 84:115-121.

Martin, A. P and S. R. Palumbi. 1993. Body size, metabolic rate, generation

time and the molecular clock. Proc. Natl. Acad. Sci. USA 90:4087-4091.

Martin, A. P., G. J. P. Naylor, and S. R. Palumbi. 1992. Rates of mitochondrial

DNA evolution in sharks are slow compared with mammals. Nature

357:153-155.

Mayr, 1942. Systematics and the origins of species. Columbia University

Press, New York.

Mayr, E. 1954. Geographic speciation in tropical echinoids. Evolution 8:1-18.

215

McAndrew, B. J., F. R. Roubal, R. J. Roberts, A. M. Bullock, and I. M. McEwen.

1988. The genetics and histology of red, blond, and associated colour

variants in Oreochrornis niloticus. Genetica 76:127-137.

McManus, J. W. 1985. Marine speciation, tectonics and sea-level changes in

southeast Asia. Proc. 5th Int. Coral Reef Congr. 4:133-138.

McPhail, J. D. 1969. Predation and the evolution of a stickleback

(Gasterosteus). J. Fish. Res. Bd. Canada 26:3183-3208.

Meyer, T. D. Kocher, P. Basasibwaki, and A. C. Wilson. 1990. Monophyletic

origin of Lake Victoria cichlid fishes suggested by mitochondrial DNA

sequences. Nature 347:550-553.

Molina-Cruz, A. 1977. The relation of the southern trade winds to upwelling

process during the last 75,000 years. Quater. Res. 8:324-338.

Moore, W. S. 1987. Random mating in the northern flicker hybrid zone:

implications for the evolution of bright and contrasting plumage patterns

in birds. Evolution 41:539-546.

Moore, W. S. and D. B. Buchanan. 1985. Stability of the northern flicker

hybrid zone. Evolution 39:135-151.

Moore, W. S. and J. T. Price. 1993. Nature of selection in the northern flicker

hybrid zone and its implication for speciation theory. Pp. 196-225 in R.

Harrison, ed. Hybrid zones and the evolutionary process. Oxford

University Press, New York.

216

Moore, W. S., J. H. Graham, and J. T. Price. 1991. Mitochondrial DNA

variation in the Northern Flicker (Colaptes auratus, Aves). Mol. Bio!.

Evol. 8:327-344.

Moran, P. and I. Kornfield. 1993. Retention of an ancestral polymorphism in

the Mbuna species Flock (Teleostei: Cichlidae) of Lake Malawi. Mol. Biol.

Evol. 10:1015-1029.

Neigel, J. E. and J. C. Avise. 1986. Phylogenetic relationships of

mitochondrial DNA under various demographic modes of speciation.

Pp. 515-534 in S. Karlin and E. Nevo, eds. Evolution process and theory.

Academic Press, New York.

Nelissen, M. 1976. Contribution to the ethology of Tropheus moorii

Boulenger (Pisces, Cichlidae) and a discussion of the significance of its

colour patterns. Rev. Zool. Afr. 90:17-29.

Neudecker, S. 1989. Eye camouflage and false eyespots: chaetodontid

responses to predators. Envir. BioI. Fishes 25: 143-147.

Nishida, M. and S. J. Lucas. 1988. Genetic differences between geographic

populations of the crown-of-thorn starfish throughout the Pacific region.

Marine Biology 98: 359-368.

Otte, D. and J. A. Endler. 1989. Speciation and its consequences. Sinauer

Press, Sunderland, Massachusetts.

Ovenden, J. R., A. G. MacKinlay and R. H. Crozier. 1987. Systematics and

mitochondrial genome organization of Australian rosellas (Aves:

Platycercidae). Mol. BioI. Evoi. 4:52-543.

217

Palmer, R. A. and C. Strobeck. 1986. Fluctuating asymmetry: measurement,

analysis, and patterns. Ann. Rev. Ecol. Syst. 17:391-421.

Palmilo, P. and M. Nei. 1988. Relationships between gene trees and species

trees. Mol. BioI. Evoi. 5:568-583.

Palumbi, S. R. 1992. Marine speciation on a small planet. TREE 7:114-117.

Palumbi, S. R. and E. C. Metz. 1991. Strong reproductive isolation between

closely related tropical sea urchins (genus Echinometra). Mol. Biol, and

Evol. 8:227-239.

Palumbi, S. R., A. Martin, S. Romano, W. O. McMillan, L. Stice, and G.

Grabowski. 1991. The simple fool's guide to PCR. Dept. of Zoology,

University of Hawaii, Honolulu.

Parsons, T. J., S. L. Olson, M. J. Braun. 1993. Unidirectional spread of

secondary sexual plumage traits across an avian hybrid zone. Science

260:1643-1646.

Pauley, G. 1990. Effects of late Cenozoic sea-level fluctuations on the bivalve

faunas of tropical oceanic islands. Paleobiology 16:415-434.

Pedersen, T. F. 1983. Increased productivity in the eastern equatorial Pacific

during the last glacial maximum (19,000 to 14,000 yrs B. P.). Geology 11:16­

19.

Perron, F. E. and A. J. Kohn. 1985. Larval dispersal and geographic

distribution in coral reef gastropods of the genus Conus. Proc. Fifth

Intern. Coral Reef Congr. 4:95-100.

218

Pesole, G, E. Sbisa, G Preparata, and C. Saccone. 1992. The evolution of the

mitochondrial d-loop region and the origin of modern man. Mol. Biol,

Evol. 9:587-598.

Peterman, R. M. 1971. A possible function of colour in coral reef fishes.

Copeia 2:330-331.

Platnick, N. 1. and G. Nelson. 1978. A model of analysis for historical

biogeography. Syst. Zoo1. 27:1-16.

Pollock, D. E. 1992. Paieoceanography and speciation in the spiny lobster

genus Panulirus in the Indo-Pacific. Bull. Mar. Sci. 51:135-146.

Porter, S. C. 1989. Some geological implications of average quaternary glacial

conditions. Quaternary Research 32:245-261.

Potts, G. 1973. The ethology of Labroides dimidiatus on Aldabra. Anim.

Behav. 21:250-291.

Powell, J. R. 1983. Interspecific cytoplasmic gene flow in the absence of

nuclear gene flow: evidence from Drosophila. Proc. N atl. Acad. Sci. USA

80:492-495.

Quinn, W. H. 1971. Late quaternary meteorological and oceanographic

developments in the equatorial Pacific. Nature 229:330-331.

Rand, D. M. and R. C. Harrison. 1989. Ecological genetics or a mosaic hybrid

zone: mitochondrial, nuclear, and reproductive differentiation of crickets

by soil type. Evolution 43:432-449.

219

Randall, J. E., G. R. Allen, and R. C. Steene. 1977. Five probable hybrid

butterflyfishes of the genus Chaetodon from the central and western

Pacific. Rec. West. Aust. Mus. 6:3-26.

Rea, D. K. 1990. Aspects of atmospheric circulation: the late Pleistocene (0­

950,000 yr) record of eolian deposition in the Pacific Ocean. Palaeogeo.,

Palaeodim., Palaeoecol 78:217-227.

Rea, D. K. 1991. Late Pleistocene paleoclimatology of the central equatorial

Pacific: flux patterns of biogeneic sediments. Paleoceanography 6:227-244.

Reaka, M. L. 1980. Geographic range, life history patterns, and body size in a

guild of coral-dwelling mantis shrimps, Evolution 34:1019-1030.

Reeb, C. A. and J. C. Avise. 1990. A genetic discontinuity in a continuously

distributed species: mitochondrial DNA in the American Oyster,

Crassostrea virginica. Genetics 124(2):397-406.

Reese, E. 1975. A comparative field study of the social behavior and related

ecology of reef fishes of the family Chaetodontidae. Z. Tierpsychol. 37:37­

61.

Reese, E. S. 1989. Orientation behavior of butterflyfishes (family

Chaetodontidae) on coral reefs: spatial learning of route specific

landmarks and cognitive maps. Environ. Biol. Fish. 25:79-86.

Reese, E. S. 1991. How behavior influences community structure of

butterflyfishes (Family Chaetodontidae) on Pacific coral reefs. Ecology

international Bulletin 19:29-41.

220

Reid, D. G. 1990. Trans-arctic migration and speciation induced by climate

change: the biogeography of Littorina (Mollusca: Gastropoda). Bull. Mar.

Sci. 47:35-49.

Reynolds, R. W. 1988. A real-time global sea-surface temperature analysis.

Journal of Climate 1:75-86.

Richmond, R. H. 1990. The effects of the El Nino/Southern Oscillation on

the dispersal of corals and other marine organisms. Pp. 127-140 in P. W.

Glynn, ed. Global ecological consequences of the 1982-1983 El Nino­

Southern Oscillation. Elsevier Press, Amsterdam.

Roff, D. A. and P. Bentzen. 1989. The statistical analysis of mitochondrial

DNA polymorphisms: chi-square and the problem of small samples.

Mol. BioI. Evol. 6:539-545.

Rogers, A. R. 1992. Error introduced by infinite-sites model. Mol. BioI. Evol.

9: 1181-1184.

Rogers, A. R. and H. Harpending. 1992. Population growth makes waves in

the distribution of pairwise genetic differences. Mol. BioI. Evol. 9:552-569.

Rosen, B. R. 1988. Progress, problems and patterns in biogeography of reef

corals and other tropical marine organisms. Helgolander Meeresunters

42:269-301.

Ross, K. G. and J. L. Robertson. 1990. Developmental stability, heterozygosity,

and fitness in two introduced fire ants (Solenopsis invicta and S. richteri)

and their hybrid. Heredity 64:93-103.

221

Ross, R. M. 1984. Catheterization: a non-harmful method of sex

identification for sexually monomorphic fishes. Prog. Fish. Cult. 46:151­

152.

Rotondo, G. M., V. G. Springer, G. A. J. Scott, and S. O. Schlanger. 1981. Plate

movement and island integration-a possible mechanism in the

formation of endemic biotas, with special reference to the Hawaiian

Islands. Syst. Zoo1. 30:12-21.

Rowland, W. J. 1979. The use of color in intraspecific communication. Pp.

379-421 in E. H. Burtt, [r., ed. The behavioral significance of color.

Garland STPM Press, New York.

Saccone, c., G. Pesole and E. Sbisa. 1991. The main regulatory region of

mammalian mitochondrial ONA: structure-function model and

evolutionary pattern. J. Mol. Evo1. 1991 33:83-91.

Saccone, c., M. Attimonelli and E. Sbisa. 1987. Structural elements highly

preserved during the evolution of the d-loop-containing region in

vertebrate mitochondrial DNA. J. Mol. Evo1. 25:205-211.

Saitou, N. and M. Nei. 1987. The Neighbor-joining method: a new method

for reconstructing phylogenetic trees. Mol. BioI. Evo1. 4:406-425.

Sambrook, J., E. F. Fritsch and T. Maniatis. 1989. Molecular cloning: a

laboratory manual, 2nd edition. Cold Spring Harbor Laboratory, Cold

Spring Harbor, New York.

222

Schopf, T. M. and 1. S. Murphy. 1973. Protein polymorphism of the

hybridization seas tars Asterias forbesi and Asterias vulgaris and

implications for their evolution. Biol. Bull. 145:589-597.

Slatkin, M. 1989. Detecting small amounts of gene flow from phylogenies of

alleles. Genetics 121:609-612.

Slatkin, M. and R. R. Hudson. 1991. Pairwise comparison of mitochondrial

DNA sequences in stable and exponentially growing populations.

Genetics 129:555-562.

Slatkin, M. and W. P. Maddison. 1989. A cladistic measure of gene flow

inferred from the phylogenies of alleles. Genetics 123:603-613.

Smith, D. G. 1972. The role of the epaulets in the Red-winged Blackbird,

Agelaius phoeniceus social system. Behavior 41:251-268.

Snow, J. 1. and M. K. Rylander. 1982. A quantitative study of the optic system

of butterflyfishes (Family Chaetodontidae). J. Hirf. 23:121-125.

Southwood, T. R. E. 1978. Ecological methods. Chapman and Hall, London.

Spolsky, C. and T. Uzzell. 1984. Natural interspecies transfer of

mitochondrial DNA in amphibians. Proc. Natl. Acad. Sci. USA 81:5802­

5805.

Springer, V. G. 1982. Pacific plate biogeography, with special reference to

shore fishes. Smithsonian Contributions to Zoology, Number 367.

Smithsonian Institution Press, Washington. D.C.

223

Springer, V. G. 1988. The Indo-Pacific fish genus Ecsenius. Smithsonian

Contributions to Zoology, Number 465. Smithsonian Institution Press,

Washington, D.C.

Springer, V. G. and J. T. Williams. 1990. Widely distributed pacific plate

endemics and lowered sea-level. Bull. Mar. Sci. 47: 631-640.

Stanley, S. M. 1981. Neogene mass extinction of western Atlantic mollusks.

Nature 293:457-459.

Stehli, F. G. and J. W. Wells. 1971. Diversity and age patterns in hermatypic

corals. Systematic Zoology 20:115-126.

Stoskopf, M. K. 1993. Fish Medicine. W. B. Saunders Co., Philadelphia.

Sturmbauer, C. and A. Meyer. 1992. Genetic divergence, speciation and

morphological stasis in a lineage of African cichlid fishes. Nature

358:578:581.

Sturmbauer, C. and A. Meyer. 1993. Mitochondrial phylogeny of the endemic

mouthbrooding lineages of cichlid fishes from Lake Tanganyika in

eastern Africa. Mol. BioI. Evol. 10:751-768.

Swofford, D. 1993. Phylogenetic analysis using parsimony (PAUP), Version

3.1.1. University of Illinois, Champaign.

Szymura, J. M. and N. H. Barton. 1986. Genetic analysis of a hybrid zone

between the fire-bellied toads Bombina bombina and B. variegata, near

Cracow in southern Poland. Evolution 40:1141-1459.

224

Szyrnura, J. M., C. Spolsky and T. Uzzell. 1985. Concordant change in

mitochondrial and nuclear genes in a hybrid zone between two frog

species (genus Bombina). Experentia 41:1469-1470.

Tajima, F. 1983. Evolutionary relationships of DNA sequences in finite

populations. Genetics 105:437-460.

Takahata, N. and M. Slatkin. 1984. Mitochondrial gene flow. Proc. Natl.

Acad. Sci. USA 81:1764-1767.

Tamura, K. 1992. Estimation of the number of nucleotide substitutions when

there are strong transition-transversion and G+C-content biases. Mol.

BioI. Evol. 9:678-687.

Tamura, K. and M. Nei. 1993. Estimation of the number of nucleotide

substitutions in the control region of mitochondrial DNA in humans and

chimpanzees. Mol. BioI Evol. 10:512-526.

Teve, D., M. Rezk, R. O. Smitherman. 1989. Genetics of body color in Tilapia

mossambica. Journal World Aquaculture Society 20:214-222.

Thresher, R. E. and E. B. Brothers. 1985. Reproductive ecology and

biogeography of Indo-West Pacific Angelfishes (Pisces: Pomacanthidae).

Evolution. 39:878-887.

Thresher, R. E., P. L. Colin, and L. J. Bell. 1989. Planktonic duration,

distribution, and population structure of western and central pacific

damselfishes (Pomacentridae), Copeia 1989:420-434.

225

Tricas, T. C. 1987. Life history and behavioral ecology of Chaetodon

multicinctus. Ph. D. dissertation, University of Hawaii, Honolulu.

Tricas, T. C. 1989. Determents of feeding territory size in the corallivorious

butterflyfish, Chaetodon multicinctus. Animal Behavior 37:830-841.

Tzeng, C-S, C-F Hui, S-C Shen and P. C. Huang. 1992. The complete

nucleotide sequence of the Crossostoma lacustre mitochondrial genome:

conservation and variations among vertebrates. Nucleic Acids Research

20:4853-4858.

Uehara, T., M. Shingaki and K. Taira. 1986. Taxonomic studies in the sea

urchin, genus Echinometra from Okinawa and Hawaii. Zoo1. Sci. 3:1114.

Valentine J. W. and D. Jablonski. 1983. Speciation in the shallow sea: general

patterns and biogeographic controls. Pp. 201-2.26 in R. W. Sims, J. H. Price

and P. E. S. Whalley, eds. Evolution, time and space: the emergence of

the biosphere. Academic Press, London.

Vermeij, G. J. 1987. The dispersal barrier in the tropical Pacific: implications

for molluscan speciation and extinction. Evo1. 41: 1046-1058.

Vermeij, G. J. 1990. An ecological crisis in an evolutionary context: el nino in

the eastern Pacific. Pp. 505-518 in P. W. Glynn, ed. Global ecological

consequences of the 1982-1983 EI Nino-Southern Oscillation. Elsevier

Press, Amsterdam.

Veron, J. E: N. 1986. Corals of Australia and the Indo-Pacific. Angus and

Robertson, Sydney.

226

Vigilant, L. M., M. Stoneking, H. Harpending, K. Hawkes, and A. C. Wilson.

1991. African populations and the evolution of human mitochondrial

DNA. Science 236:1503-1507.

Wakeley, J. 1993. Substitution rate variation among sites in hypervariable

region I of human mitochondrial DNA. J. Mol. Evol. 37:613-623.

Ward, F. H., B. L. Frazier, K. Dew-Iaer, and S. Paabo. 1991. Extensive

mitochondrial diversity within a single Amer-indian tribe. Proc. Natl.

Acad. Sci. USA 88:8270-8274.

Wenink, P. W. A. J. Baker, and M. G. J. Tilanus. 1993. Hypervariable-control­

region sequences reveal global population structuring in a long-distance

migrant shorebird, the Dunlin (Calidris alpina). Proc. Natl. Acad. Sci.

USA 90:94-98.

West-Eberhard, M. J. 1983. Sexual selection, social competition and

speciation. The Quarterly Review of Biology 58:155-183.

Whittam, T. S., A. G. Clark, M. Stoneking, R. L. Cann, and A. C. Wilson. 1986.

Allelic variation in human mitochondrial genes based on patterns of

restriction site polymorphisms. Proc. Natl. Acad. Sci. USA 83:9611-9615.

Wilson, A. c., R. L. Cann, S. M. Carr, M. George, U. B. Gyllensten, K. M.

Helm-Bychowski, R. G. Higuchi, S. R. Palumbi, E. M. Pranger, R. D. Sage,

and M. Stoneking. 1985. Mitochondrial DNA and two perspectives on

evolutionary genetics. BioI. J. Linn. Sci. 26:375-400.

Winans, G. A. 1980. Geographic variation in the milkfish, Chanos chanos. I.

Biochemical evidence. Evolution 34:558-574.

227

Woodland, D. J. 1983. Zoogeography of the Siganidae (Pisces): an

interpretation of distribution and richness patterns. Bull. Mar. Sci. 33:713­

717.

Woodland, D. J. 1986. Wallace's line and the distribution of marine inshore

fishes. Pp. 453-469 in T. Uyeno, R. Arai, T. Taniuchi, and K. Matsuura,

eds. Indo-Pacific fish biology, proceedings of the second international

conference on Indo-Pacific fishes. Ichthyological Society of Japan, Tokyo.

Zakharov, V. M. 1981. Fluctuating asymmetry as an index of development

homeostasis. Genetica 13:241-256.

Zumpe, D. 1965. Laboratory observations on the aggressive behavior of some

butterfly fishes (Chaetodontidae). Z. Tierpsychol. 22:226-236.

228