Uhm Phd 9506218 R.Pdf

INFORMATION TO USERS

Thismanuscript has been reproduced from the microfilm. master. liMI films the text directly from the original or copysubmitted. 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 the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adverselyaffect 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, if unauthorized 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 ecual sectionswith 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 reproduced xerographically 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.

U·M·I University Microfilms International A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. M148106-1346 USA 313!761-4700 800:521-0600

Order Number 9506218

Hybridization and polyploidy in the coral genus Acropora

Kenyon, Jean Carol, Ph.D.

University of Hawaii, 1994

V·M·I 300N. ZeebRd AnnArbor, MI 48106

HYBRIDIZATION AND POLYPLOIDY

IN THE CORAL GENUS ACROPORA

A DISSERTATION SUBMITTED TO THE GRADUATE DIVISION OF THE UNIVERSITY OF HAWAII IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

IN

ZOOLOGY

AUGUST 1994

By

Jean C. Kenyon

Dissertation Committee:

Robert A. Kinzie, III, Chairperson John S. Stimson Samuel R. Haley Robert H. Richmond Gerald D. Carr ©Copyright 1994 by Jean Carol Kenyon

iii This work is dedicated to the memory of

Dr. Barbara Shuler Mayo,

1945-1988

friend and mentor,

whose clarity and vitality

enriched everything she touched

iv ACKNOWLEDGEMENTS

No one completes a Ph.D. without being indebted to a great many people. As my research has been conducted in diverse geographic locales, perhaps more than many candidates I have relied upon the helpfulness and good will of others.

My field work in the Northwest Hawaiian Islands was made possible with the logistical support of the Pacific

Islands Office of the U.S. Fish & Wildlife Service. I especially thank T. Michael Moser, Mitch Craig, and Robert

Cummins for assistance with collecting coral samples under ofttimes hazardous conditions. I am extremely grateful to

Bob Richmond for making available the facilities of the

University of Guam Marine Lab, and to Kiyoshi Yamazato for extending the invitation to work at Sesoko Marine Science

Center in Okinawa. Bette Willis and Carden Wallace very generously shared their expertise and facilities at Magnetic

Island, Australia, for which I am deeply indebted. As representative of The Nature Conservancy, Chuck Cook expedited my work in Palau, as did Noah Idechong, Chief of

Marine Resources of the Republic of Palau.

Frank Te and Paul Chirichetti gave me invaluable aid with collections in Guam, as did Ann Kitalong in Palau. Many individuals generously shared coral gametes with me on spawning nights, including Dave Krupp, Evelyn (Fenny) Cox,

v Frank Stanton, Bob Richmond, Fernando Rivera, Bette Willis,

Carden Wallace, Peter Harrison, Russ Babcock, Ben Stobart, and Rob Rowen.

Special thanks are due to Cynthia Hunter for help at several critical junctions, and to Randy Haley for guiding me through the treacherous waters of microtechnique.

Parts of this research were financially supported by the Charles and Margaret Edmondson Fund, the International

Agreements Fund of the School of Hawaiian, Asian, and

Pacific Studies, the Lerner-Gray Fund of the American Museum of Natural History, and a Research Corporation of University of Hawaii research assistantship. I also gratefully acknowledge the important material assistance of E. Alison

Kay, whose commitment to education and the advancement of knowledge is exemplary.

For the members of my dissertation committee I have reserved my final thanks. Bob Kinzie, my committee chairperson, consistently offered useful suggestions and approaches, for which I am truly appreciative. John Stimson,

Randy Haley, Bob Richmond, Gerald Carr and former member

Michael Hadfield gave generously of their time, energy, and expertise. Their knowledge and spirit of inquiry are gifts which I hope to as generously pass on.

vi ABSTRACT

Six species of the scleractinian coral genus Acropora, all of which participate in natural multispecies spawning events in the central or west Pacific, were used in a series of controlled experimental hybridization crosses. Five pairwise combinations of the six parental species were experimentally crossed. No hybrids were formed in one of the interspecific combinations. Hybridization levels for four of the species combinations ranged from 4% to 30%, with hybrids raised to the planula larva stage. Hybridization, along with asexual vegetative reproduction and the production of unreduced gametes, are primary conditions conducive to the development of polyploidy. Chromosome number was determined for twenty-two species of Acropora, three species of

Montipora, and one species of Fungia, using experimentally­ generated embryos which were treated with colchicine. All of the corals had 28 chromosomes, except for six of the

Acropora species, which had chromosome numbers of 24, 30,

30, 42, 48, and 54. Several alternative mechanisms involving a combination of polyploidy and aneuploidy are proposed to account for the variations in chromosome number. Acropora is the most speciose coral genus (>150 species), with nearly twice as many species as the two next most diverse coral genera. It is proposed that sympatric speciation by polyploidy during Pliocene-Pleistocene sea level

vii transgressions may in part account for the unique proliferation of species in Acropora. If hybridization

during synchronized spawning events serves as a mechanism

for speciation in corals, then regions characterized by

multispecies spawning events are more likely to serve as

sites of speciation than those where spawning is more

asynchronous. Temporal patterns of coral spawning in Palau

and Yap indicate less synchrony at equatorial than at higher

latitudes in the Central Pacific. Asynchrony of spawning in

Palau does not support the hypothesis that the Indo-Pacific

center of high coral diversity is also the center of origin.

The three species of Acropora found in the Northwest

Hawaiian Islands produce mature gametes, but do not overlap

in time of spawning, thereby precluding opportunities for

hybrid-driven speciation. The absence of Acropora from the

main islands of the Hawaiian chain is most likely due to

unfavorable currents for larval dispersal.

viii TABLE OF CONTENTS

ACKNOWLEDGEMENTS...... v

ABSTRACT...... vii

LIST OF TABLES...... xii

LIST OF FIGURES...... xiv

CHAPTER 1. INTRODUCTION...... 1

CHAPTER 2. EXPERIMENTAL HYBRIDIZATIONS DURING MULTISPECIES SPAWNING EVENTS...... 9 Introduction...... 9 Materials and Methods.. 11 Study sites and identification of corals...... 11 Collection of ripe colonies...... 12 Collection of gametes...... 13 Experimental crosses and embryo culture...... 15 Determining percent fertilization... 18 Results...... 19 Discussion...... 22 Ecological and evolutionary fates of hybrids...... 22 "Species groups" and molecular data. 27 Asymmetry of fertilization...... 28 Species concepts.. 30

CHAPTER 3. DETERMINATION OF CHROMOSOME NUMBER.... 36 Introduction...... 36 Materials and Methods...... 38 Study sites and identification of corals...... 38 Collection of ripe colonies...... 40 Collection and preparation of gametes...... 41 Embryo culture..... 45 Embryo treatment... 45 Staining and squash preparation..... 46 Photography...... 47 Results...... 48 Discussion...... 64 General mechanisms of development of polyploidy...... 70 Conditions conducive to polyploidy.. 87 Asexual reproduction..... 87 Production of unreduced gametes... 93

ix TABLE OF CONTENTS (continued)

CHAPTER 3 (continued) General mechanisms of aneuploidy.... 95 Specific mechanisms of somatic chromosome number variation..... 100 Modell...... 102 Model 2...... 105 Model 3...... 111 Model 4...... 115 Model 5...... 120 Combinations of models... 123 Predictions of the models.. 127 Genome size...... 132 Molecular data...... 135 Chromosome pairing at meiosis..... 140 Experimental hybridizations... 141 Ecological factors. 144 Sympatric speciation... 150 Biogeography...... 154 Parthenogenesis and animal polyploidy.. 162 Why is polyploidy not more common in animals?...... 169

CHAPTER 4. REPRODUCTIVE STATUS OF SOME CORALS FROM PALAU AND YAP... 173 Introduction...... 173 Materials and Methods.... 174 Palau...... 174 yap...... 176 Results...... 177 Palau...... 177 Yap...... 181 Discussion...... 182 Lunar periodicity and tidal regime.. 182 Intraspecific synchrony.... 182 Interspecific synchrony.... 183 Biogeographic implications of spawning synchrony. 185

CHAPTER 5. SEXUAL REPRODUCTION IN HAWAIIAN ACROPORA...... •...... •... 197 Introduction...... 197 Materials and Methods.... 201 Results...... 203 Acropora valida. 203 Acropora cytherea...... 213 Acropora humilis...... 213

x TABLE OF CONTENTS (continued)

CHAPTER 5 (continued) Discussion . 217 Reproductive maturity and degree of intraspecific synchrony. 217 Lunar periodicity . 220 Interspecific asynchrony . 220 Sexual vs. asexual reproduction . 222 Post-Pleistocene colonization . 225

APPENDIX Treatment parameters for A-I experimentally-generated embryos . ... 228

APPENDIX Spawning characteristics for corals A-2 from Guam, Okinawa, Australia, Hawaii, and Palau . 232

LITERATURE CITED...... 238

xi LIST OF TABLES

Table

1.1. Number of species per genus for the 109 extant genera of hermatypic corals...... 4

2.1. Spawning characteristics of corals from Guam, Australia, and Palau which were used in hybridization experiments...... 20

2.2 Percent fertilization in controlled intraspecific and interspecific crosses. ... 21

3.1. Geographic areas from which species of coral used in determining chromosome number were collected...... 39

3.2. Summary of results of methods used in culturing experimentally-generated coral embryos...... 43

3.3. Number of fixed embryos which were examined, and the number of these which were useful for determining chromosome number...... 60

3.4. Somatic chromosome number established for twenty-two species of Acropora...... 61

3.5. Summary of major features of Models 2-5...... 126

3.6. Summary of predictions of Models 2-5 based upon genome size and molecular data.. 133

3.7. Summary of predictions of Models 2-5 based upon experimental hybridizations.. ... 143

4.1. Reproductive status of corals sampled in Palau and examined with a dissecting microscope for developing eggs or larvae between May 6 and June 11, 1993...... 178

4.2. Summary of 1993 spawning dates and egg size for ten fertile coral species in Palau.. ... 179

4.3. Summary of number of corals sampled and examined by eye for developing eggs on the reef flat adjacent to Guufnuw Channel, Yap, on May 27, 1993...... 181

xii LIST OF TABLES (continued)

Table

5.1. Summary of dissections of fixed polyps of Acropora valida from French Frigate Shoals. 207

5.2. Summary of dissections of fixed polyps of Acropora cytherea from French Frigate Shoals , . 216

5.3. Comparative egg size and color for Acropora valida, A. cytherea, and A. humilis . 218

xiii LIST OF FIGURES

Figure

2.1. Design of hybridization experiments. 16

3.1. Photomicrographs of somatic chromosomes from seven species of Acropora.. 52

3.2. Somatic chromosome counts for twenty-two species of Acropora based on metaphase spreads of colchicine-treated embryonic cells...... 55

3.3. Somatic chromosome counts for three species of Montipora and one species of Fungia based on metaphase spreads of colchicine- treated embryonic cells... 62

3.4. Somatic chromosome counts in four interspecific hybrids of Acropora based on metaphase spreads of colchicine-treated embryonic cells...... 65

3.5. Two mechanisms of autotetraploidy.... 71

3.6. Two mechanisms of autotriploidy...... 74

3.7. Two mechanisms of allotetraploidy which involve formation of an intermediate, adult hybrid...... 77

3.8. Two mechanisms of allotetraploidy which do not involve formation of an intermediate, adult hybrid...... 80

3.9. Chart illustrating the possible origin of a polyploid complex.... 84

3.10. The general relationships of species according to Modell...... 103

3.11. The general relationships of species according to Model 2...... 107

3.12. The general relationships of species according to Model 3.. 113

3.13. Family tree of scleractinian corals with time frame to accommodate Models 2 and 3... 116

xiv LIST OF FIGURES (continued) Figure

3.14. The general relationships of species according to Model 4...... 118

3.15. The general relationships of species according to Model 5...... 121

3.16. Family tree of scleractinian corals with time frame to accommodate Models 4 and 5... 124

3.17. Models 2, 3, 4, and 5 aligned for ease of comparison...... 129

3.18. Chromosome number by morphologically similar groups...... 137

3.19. Distribution of Acropora danai, A.divaricata, A. gemmifera, A. valida, and A. elseyi..... 159

5.1. The Hawaiian Archipelago.... 198

5.2. Distribution of Acropora valida, A. cytherea, and A. humilis in Hawaii... 199

5.3. Dissections showing mature gonads in Acropora species at French Frigate Shoals...... 204

5.4. Oocyte growth in Acropora valida at French Frigate Shoals..... 209

5.5. Histology of Acropora valida polyps in the week prior to spawning...... 212

5.6. Colony sizes of Acropora valida on the reef flat and at La Perouse pinnacle...... 214

5.7. Colony sizes of Acropora cytherea at La Perouse Pinnacle.... 224

xv CHAPTER 1

INTRODUCTION

In applying the system of Linnaean nomenclature to scleractinian corals, early coral taxonomists worked on single specimens or on very small series which represented fragments collected from different habitats and geographical locations (Veron and Pichon, 1976). These early workers gave a different name to nearly every specimen (Sheppard and

Sheppard, 1991), thereby reflecting the historical climate of biological thought, in which naming and cataloging organisms was of prime importance. The shortcomings of this approach were compounded by the lack of appreciation of morphological variability (Veron and Pichon, 1976), as well as the absence of underwater observations (Wallace, 1978).

The inheritance of modern coral taxonomists has consequently been a tangled one, resulting in numerous taxonomic revisions (Vaughan and Wells, 1943; Dinesen, 1980; Veron,

1986; Veron and Pichon, 1976, 1980, 1982; Veron et al.,

1977; Wallace, 1978; Scheer and Pillai, 1983; Veron and

Wallace, 1984; Hoeksema, 1989) which have redescribed, revised, and synonymized past treatments. Taxonomic assignment within the Order Scleractinia at present remains based on skeletal criteria. Intraspecific skeletal variation demonstrated by "ecomorphs" is attributed to phenotypic or genotypic responses to specific ecological conditions (Veron

1 and Pichon, 1976). While some authors argue that the basic taxonomic designation of a species can be largely decoupled from the various problems that surround the species concept

(Veron, 1981; Mishler and Donoghue, 1982), others (de

Queiroz and Gauthier, 1992) argue that modern comparative biology requires a taxonomic system based on evolutionary principles which define species according to the principle of descent. At the present time, biologists still have not developed a phylogenetic system of taxonomy (de Queiroz and

Gauthier, 1992), with principles and rules formulated in terms of the central tenet of evolution. By contrast, taxonomic identification of many organisms remains deeply rooted in the idea of the biological validity of morphological species.

The coral genus Acropora is widespread throughout the tropical Indian, Pacific and west Atlantic Oceans, where colonies are typically a dominant component of the shallow reef assemblage (Wells, 1954; Rosen, 1971; Randall, 1973;

Wallace, 1978; Done, 1982; Veron, 1986). A unique feature of the genus is corallites of two types, axial and radial

(Wallace et al. 1991). Radial corallites bud off from axial corallites during growth. While all scleractinians are capable of broad morphological variation, the axial method of growth in Acropora enhances this variation in colony form, allowing a continuum of shape possibilities (Wallace,

1978). Indeed, the genus has been described as "protean"

2 (Wells, 1955, in Veron, 1986), referring to the mythical

Greek figure Proteus, who eluded those who sought knowledge of him by constantly changing his form.

Acropora is the largest of all extant coral genera

(Wells, 1987), with 368 nominal species (Veron 1986). The number of valid species probably numbers at least 150

(Veron, 1993). The speciose nature of the genus is demonstrated by comparing the number of species presently recognized in Acropora with the number of species per genus for the other 109 extant genera of hermatypic corals (Veron,

1993) (Table 1.1). Eighty-three percent of extant coral genera have fewer than ten species per genus, while the two most diverse genera aside from Acropora each have 80 species

(Montipora and Porites) .

The dominance of Acropora on shallow Indo-Pacific reefs may be attributed not only to high species diversity but also to competitive dominance by overgrowing or shading out other, slower-growing coral species (Randall, 1973;

Buddemeier and Kinzie, 1976; Wallace, 1985; Richmond, 1988).

Considerable variation exists among species in their tendency to asexually propagate by fragmentation (Randall,

1973; Tunnicliffe, 1981; Highsmith et al. 1980; Knowlton et al. 1981; Bothwell, 1981; Highsmith, 1982; Kobayashi, 1984;

Wallace, 1985). Recementing of fragments to the substrate and continued growth of the recemented fragments can result in proliferation of a large number of physiologically

3 TABLE 1.l. Number of species per genus for the 109 extant genera of hermatypic corals. Data are compiled from Veron (1993) .

Number of species Frequency Genus in genus of genera

1 41

2 14

3 13

4 5

5 6

6 3

7 3

8 2

9 3

10 1

12 3

13 1

14 1

15 5

16 1

22 1 Pavona

30 2 Favia, Goniopora

33 1 Fungia

80 2 Montipora r Porites

150 1 Acropora

4 independent colonies which represent a single clone

(Bothwell, 1981; Stoddart, 1984a; Hunter, 1988).

Reproduction in Acropora occurs by sexual as well as asexual means. Members of this genus are a chief component in multispecies spawning events on the Great Barrier Reef

(Harrison et al., 1984; willis et al., 1985; Babcock et al.

1986), western Australia (Simpson, 1985), Guam (Richmond and

Hunter, 1990), and Okinawa (Heywa~d et al., 1987), during which colonies release gametes on one to several nights during late spring or summer each year. In other regions of the Indo-West Pacific where spawning has been documented, timing of intra-and inter-specific spawning is less tightly synchronized (Shlesinger and Loya, 1985; Oliver et al.,

1988) .

Despite the simplifying effect of recent synonymies which reduces the number of recognized Acropora species, considerable taxonomic confusion still exists. This is evidenced in part by the number of unidentified acroporids recorded on field surveys by even the most knowledgable and experienced investigators (Birkeland et al. 1976; Randall et al. 1978; Randall, 1990; Maragos, 1991; Sheppard and

Sheppard, 1991; Veron, 1993). Scleractinian coral taxonomy and ecology are presently based on the assumption that coral species do not commonly hybridize, if at all (Hodgson,

1988). Recurring problems of identification based on skeletal morphology prompted the question which I address in

5 Chapter 2, "Can species of the genus Acropora form interspecific hybrids during multispecies spawning events?"

The family Acroporidae appeared in the fossil record by the early Tertiary, along with 16 ~f the 19 extant Indo-

Pacific families of hermatypic corals (Wells, 1956). The genus Acropora appeared during the Eocene (ca. 60

MYA) (Wells, 1956), but the ages of living species are not known (Potts, 1985). While some authors have suggested that some species of Acropora are recently evolved (Veron, 1986) and that Acropora may speciate rapidly relative to other genera (Hodgson, 1989), no authors have formulated or tested hypotheses which address the proliferation of species unique to this genus. In Chapter 3' I test my hypothesis that polyploidy may be a factor in the diversification of

Acropora, by establishing chromosome number for 22 species in this genus as well as for several species from other, less speciose genera and families. Alternative models are proposed to suggest how the demonstrated series of chromosome numbers found in Acropora may have arisen.

In the period since the documentation of mass coral spawning on the Great Barrier Reef (Harrison et al., 1984;

Willis et al., 1985; Babcock et al., 1986), studies of coral reproduction have sought to discover the degree to which spawning synchrony occurs in other geographic regions. In

'Portions of the work presented in Chapter 3 have been accepted for publication in the Proceedings of the Seventh International Coral Reef Symposium (1992).

6 the Central Pacific, timing of intra- and inter-specific spawning at equatorial latitudes appears to be less tightly synchronized than at higher latitudes (Oliver et al., 1988)

In Chapter 42 I examine the reproductive status of some corals in Palau and Yap, the results of which reinforce this apparent trend. The opportunities for hybridization are in part a function of interspecific spawning synchrony. In

Chapter 4 I also argue that if hybridization serves as a mechanism for speciation in corals, then regions characterized by mass spawning events are more likely to serve as sites of speciation than those where spawning is more asynchronous. This argument has relevance for a persistent controversy in coral biogeography which debates whether the Indo-Pacific region of high diversity is a

"cradle" in which species have arisen (Stehli and Wells,

1971; Potts, 1984a; McManus, 1985) or a "museum" in which species have accumulated (Ladd, 1960; Heck and McCoy, 1978;

Rosen, 1984; Jokiel, 199Gb; Wallace et al., 1991).

Only three species of Acropora are found in Hawaii.

Their distribution is restricted to the Northwest Hawaiian

Islands between Kauai and Laysan (Grigg et al., 1981). The only previous work which examined these species in Hawaiian waters (Grigg, 1981; Grigg et al., 1981) found no evidence of sexual reproduction in any of the three species, and from

2Po r t i on s of the work presented in Chapter 4 have been accepted for publication in Pacific Science

7 these data the authors hypothesized that larval recruits are of allochthonous origin. In Chapter 53 I present evidence showing that Acropora valida, A. cytherea, and A. humilis are sexually reproductive at French Frigate Shoals. Although these three species participate in multispecies spawning events in other areas of the Pacific and, therefore, may potentially form hybrids in those locations, interspecific spawning asynchrony in the Northwestern Hawaiian Islands precludes the possibility of forming intrageneric hybrids in

Hawaii.

3Ma j o r portions of the work presented in Chapter 5 have been published in Kenyon (1992)

8 CHAPTER 2

EXPERIMENTAL HYBRIDIZATIONS DURING

MULTISPECIES SPAWNING EVENTS

INTRODUCTION

Members of the scleractinian coral genus Acropora dominate multispecies spawning events on the Great Barrier

Reef (Harrison et al., 1984; Willis et al., 1985; Babcock et al., 1986), western Australia (Simpson, 1985), Guam

(Richmond and Hunter, 1990), and Okinawa (Heyward et al.,

1987), during which colonies release gametes on one to several nights following a full moon of late spring or early summer each year. Many species spawn on the same night of the year within hours of each other, releasing an enormous number of buoyant eggs and sperm from different species into the water simultaneously (Oliver and Willis, 1987; Richmond,

1988). Coral reproductive products may accumulate into extensive surface slicks (Babcock et al., 1986; Oliver and

Willis, 1987; Atkinson and Atkinson, 1992). Synchronized mass spawning events thus present many opportunities for hybridization between species.

Only one published study has explored the capacity for hybridization in scleractinian corals. Working in Hawaii,

Hodgson (1988) performed intrageneric crosses using the eggs of Montipora dilitata and sperm from Montipora patula or

Montipora verrucosa. Nine to fourteen percent of eggs showed

9 cleavage but none survived to the blastula stage. Reciprocal crosses were not performed, however, and the author noted that asymmetrical interspecific cross fertilization has been demonstrated in sea urchins (Strathmann, 1981) and hydrozoans (Miller, 1982). No cleavage was found in intergeneric crosses using eggs from Porites lobata and sperm from Montipora dilitata. Ultrastructural comparison of spermatozoa shows that substantial cytological diversity exists among different coral genera (Harrison, 1985;

Steiner, 1993).

Scleractinian coral taxonomy and ecology are presently based on the assumption that coral species do not commonly hybridize (Hodgson, 1988). Despite the simplifying effect of recent synonymies which reduces the number of recognized

Acropora species (Veron and Wallace, 1984; Veron, 1986), considerable taxonomic uncertainty and problems of identification still exist. In this Chapter I test my hypothesis that currently recognized species of Acropora may be able to form viable interspecific hybrids during multispecies spawning events, by performing a series of controlled intraspecific and interspecific crosses.

Moreover, recent unpublished studies (Willis et al.,

1992; B. Stobart, pers. comm.) suggest that Montipora digitata, which is presently considered a single species, may have a complex pattern of breeding compatibility in which different morphs are unable to successfully cross-

10 fertilize. Controlled experimental intra-and intermorphic crosses are also conducted for Montipora digitata.

MATERIALS AND METHODS

STUDY SITES AND IDENTIFICATION OF CORALS

Acropora danai and A. valida were collected from Pago

Bay, Guam and maintained at the University of Guam Marine

Laboratory. Acropora formosa, A. millepora, A. pulchra, A. cytherea, and Montipora digitata were collected from

Geoffrey Bay, Magnetic Island, an inshore fringing reef in the central Great Barrier Reef Province. Acropora clathrata and A. florida were collected in Palau, from the upper slope of the barrier reef outside Ulong Channel, and maintained at the Micronesian Mariculture Demonstration Center. Care was taken to gather widely separated colonies, to minimize collection of clonemates.

Corals were identified by reference to field guides

(Wallace, 1978; Randall and Myers, 1983; Veron, 1986), museum specimens (University of Guam Marine Laboratory,

Mangilao; Bishop Museum, Honolulu; Museum of Tropical

Queensland, Townsville, Australia), and with the assistance of taxonomic experts (R. Randall, C. Wallace). Based on features of colony morphology and pigmentation, two morphs of Montipora digitata were categorized as "fat fingers" and

"yellow spatulate" by B. Stobart. Voucher specimens of corals from Guam were deposited in the collections of the

11 Bishop Museum in Honolulu. Voucher specimens of corals from

Magnetic Island were deposited in the collections of the

Museum of Tropical Queensland in Townsville, Australia.

COLLECTION OF RIPE COLONIES

Records of previous spawning events (C. Hunter, pers. comm.; R. Richmond, pers. comm.; Harrison et ai., 1984;

Willis et ai., 1985; Babcock et ai., 1986; B. Willis, pers. comm.) and presence of colored eggs in fertile colonies

(Harrison et ai., 1984; Babcock et ai., 1986; Heyward et ai., 1987; Heyward, 1988) were used to predict spawning dates in Guam and Australia. No coral reproduction data were available for Palau, so preliminary sampling to determine the time of maturation and spawning was undertaken for twenty Acropora species (Chapter 4). Colored eggs, an indicator of maturation in Acropora and Montipora, can be readily seen by breaking open a small, distal portion of a colony in the field and examining the broken ends. Although the presence of motile spermatozoa in testes squashes has been used to refine predictions of spawning readiness

(Harrison et ai., 1984; Babcock et ai., 1986; Heyward,

1988), this method did not yield uniform results and was not generally used in this study.

In Guam and Palau, several ripe colonies or portions of ripe colonies of each species were collected several days before a predicted spawning date and maintained in large,

12 concrete flow-though aquaria. As the date of predicted spawning approached, colonies were checked nightly for indications of spawning activity. No established laboratory facilities were available at the Australian site, so colonies were brought ashore shortly before sunset, placed in individual containers on the beach, and returned to the reef if signs of spawning did not occur by 11 PM.

COLLECTION OF GAMETES

Acropora and Montipora are simultaneous hermaphrodites, with male and female gonads in each fertile polyp. Spawning. consists of the release of eggs and sperm, which have been packaged together into discrete bundles, from the mouth of fertile polyps. Bundles can be seen protruding beneath the polyp mouths from one to several hours before they are released. When egg/sperm bundles were seen protruding from the polyp mouths, conspecifics were placed in individual buckets or separate Plexiglas aquaria with no water flow and allowed to spawn. Seawater for buckets or Plexiglas aquaria in which corals were allowed to spawn was collected during the morning or early afternoon of predicted spawning nights.

This was done to preclude contamination from gametes in freshly-obtained seawater, as eggs and sperm are viable for about seven hours (Heyward and Babcock, 1986).

Packaged egg and sperm bundles are buoyant, and float to the surface upon release from the polyp mouth. At the

13 surface they dissociate into separate eggs and sperm, generally within 10-30 minutes. Spawned bundles were allowed to dissociate naturally on the surface of the containers in which they spawned. Eggs were collected from the water surface by skimming with a small beaker or gathering with a wide-mouth pipette, and were gently rinsed in a stream of seawater on a Nytex mesh screen. Acropora eggs were washed using a 320 ~m mesh size, while smaller Montipora eggs were washed using a 160 ~m mesh. Care was taken to avoid contamination by foreign sperm by using separate equipment

(pipettes, bulbs, beakers, jars) for each colony of each species, or, in the case of sieves used for washing eggs, by immersing in fresh water at least 30 minutes between uses.

The tips of all glass pipettes used to collect and transfer eggs and embryos were fire-polished, to reduce damage to the egg membrane (C. Hunter, pers. comm.). Seawater used for washing gametes and culturing embryos was either filtered with a 0.45 ~m Millipore filter (Guam) or collected earlier during the morning of spawning nights (Australia and Palau) .

Diameters of 30-50 freshly-spawned and washed eggs were measured for each Acropora species, using a dissecting microscope with a calibrated ocular micrometer.

Independent observations of spawning in the field were made by trained observers using SCUBA equipment.

14 EXPERIMENTAL CROSSES AND EMBRYO CULTURE

For each species, gametes from the two most fecund colonies were used for a series of controlled intraspecific and interspecific crosses (or, for M. digitata, intramorphic and intermorphic crosses) (Figure 2.1). The two colonies of each species or morph were randomly assigned as the first or second experimental colony. The following interspecific and intermorphic crosses were experimentally tested:

A. valida x A. danai (Guam)

A. pulchra x A. millepora (Australia)

A. pulchra x A. cytherea (Australia)

A. millepora x A. formosa (Australia)

A. clathrata x A. florida (Palau)

Montipora digitata "fat fingers" x Montipora digitata

"yellow spatulate" (Australia)

With the exception of control cultures for sperm contamination or parthenogenesis (Figure 2.1), rinsed eggs from the experimental colony or colonies were pipetted into

250-ml or 1L culture containers, and several drops of sperm effluent from the rinsing procedure(s) were added. More thorough washing of eggs was required for the sperm contamination/parthenogenesis controls, to remove excess sperm still adhering from the egg/sperm bundle. For these controls, eggs were washed on a Nytex mesh screen in several

15 Figure 2.1. Design of hybridization experiments. A and B represent two different Acropora species, or, in the case of Montipora digitata, two different morphs. Eggs are represented by circles and sperm are represented by ellipses with "tails". For species or morph A, solid and open symbols represent the gametes from the first and second experimental colonies, respectively. For species or morph B, lined- and I-' hatched symbols represent the gametes from the first and second experimental m colonies, respectively. DESIGN OF HYBRIDIZATION EXPERIMENTS Species • 0 or Morph A sa SO

Intraspecific • 0 Crosses OS­sa. I-' -.J Interspecific Cross (Hybridization)

Selfing ~ ~ ~\iiS:l Controls ~ 'L9J ~~

Contamination/ fl} Parthenogenesis -.­ \~~~/ \~~./ Controls \bO/ changes of gamete-free seawater and examined with a dissecting microscope for signs of residual sperm before being pipetted into a separate culture container. The experimental design does not allow the direction of

fertilization to be known in any hybrid embryos, nor does it allow for the possibility that self-fertilization rates may be enhanced in the presence of foreign sperm.

Crosses cultured in Guam were gently agitated with a mechanical wrist agitator. One-liter plastic culture bottles used in Australia and Palau were firmly capped, placed in a mesh bag, and attached to a buoyed line on the reef.

DETERMINING PERCENT FERTILIZATION

First cleavage occurs 1.5 - 2 hours after gametes are

combined. Percent fertilization was determined by counting

the number of unfertilized eggs and developing embryos in

culture samples between 8.75 and 12 hours after ~ombining

gametes. A portion of the embryo cultures was subsequently

treated for determination of chromosome number (Chapter 3),

while the remaining portion of cultures with developing

embryos was allowed to continue development. All controls

were cultured for 11-12 hours after being initiated.

Cultures showing no embryogenesis after 11-12 hours were

discarded.

18 RESULTS

Collected colonies from a given geographic area released gametes on the same night (Table 2.1), either synchronously (Guam and Australia) or within a maximum of several hours of each other (Palau). Interspecific synchrony in the field was determined either through direct observation (P. Chirichetti, pers. comm.; B. Stobart, pers. comm.) or inferred by examining flagged, fertile colonies in the field the following day for presence or absence of gonads (Y. Golbuu, pers. comm.).

The mean diameter of eggs (Table 2.1) from species paired in interspecific crosses did not significantly differ

(p<0.05, Student1s t test) . All intraspecific or intramorphic crosses yielded levels of fertilization better than 80% (Table 2.2), with the exception of the "fat fingers" morph of Montipora digitata, which yielded slightly less than 70% fertilization. No interspecific hybrids were formed by combining gametes from Acropora valida and A. danai. Other interspecific crosses involving Acropora yielded low to moderate proportions (4% - 30.6%) of embryos

(Table 2.2). There were no signs of embryogenesis in cultures derived from gametes of the two morphs of Montipora digitata. None of the controls for selfing and for sperm contamination or parthenogenesis showed any signs of fertilization.

19 TABLE 2.1. Spawning characteristics of corals from Guam, Australia, and Palau which were used in hybridization experiments. Beginning and end of observed time of spawning are accurate to + or - 10 minutes. Absence of precise time observations is indicated by "?".Freshly-spawned, washed eggs are from a single colony for each species, and are measured across one diameter.

Egg Size (Jlm)

Species/ Dates Spawning n Mean sd Location Spawned Time Span Diameter

GENUS ACROPORA

A. valida 8/5/90 2100-2230 50 598 63.0 Guam

A. danai 8/5/90 2100-2230 50 626 88.0 Guam

A. formosa 10/16/92 2100-? 50 489 30.5 Australia

A. pulchra 10/16/92 2100-? 50 505 55.3 Australia

A. millepora 10/16/92 2100-? 50 529 38.0 Australia

A. cytherea Australia 10/16/92 2100-? 50 556 55.4

A. florida 6/11/93 1900

A. clathrata 6/11/93 1900

GENUS MONTIPORA

M. digitata 10/12/92 2015-2025 Australia "fat fingers II

lIyellow 10/12/92 2015-1025 spatulate"

20 TABLE 2.2. Percent fertilization in controlled intraspecific and interspecific crosses. For each cross, counts were determined from a single 8.75-12-hour-old culture. N = total number of unfertilized eggs and embryos counted.

Percent N fertilization

GENUS ACROPORA Intraspecific Crosses A. valida x A. valida 80.8 240 A. danai x A. danai 87.3 353 A. formosa x A. formosa 9.7.7 388 A. pulchra x A. pulchra 96.5 399 A. cytherea x A. cytherea 89.7 397 A. millepora x A. millepora 84.6 415 A. clathrata x A. clathrata 91. 7 168 A. florida x A. florida 82.2 129

Interspecific Crosses A. valida x A. danai 0.0 302 A. pulchra x A. millepora 20.0 628 A. pulchra x A. cytherea 20.0 350 A. millepora x A. formosa 4.0 325 A. clathrata x A. florida 30.6 306

GENUS MONTIPORA Montipora digitata Intramorph crosses "fat fingers" x "fat fingers" 69.4 238 "yellow spatulate II x 91. 9 309 "yellow spatulate"

Intermorph crosses "fat fingers ll x 0.0 294 "yellow spatulate"

21 In all cultures with developing embryos, embryogenesis was allowed to continue for an additional 15-24 hours before

the cultures were discarded. There was no apparent difference in appearance or viability between interspecific

and intraspecific (or intramorphic) larvae.

DISCUSSION

ECOLOGICAL AND EVOLUTIONARY FATES OF HYBRIDS

Experimental animal hybrids frequently do not develop beyond the gastrula stage (Moore, 1961j Denis and Brachet,

1969a,bj Hodgson, 1988). The development of echinoderm and

amphibian eggs up to the blastula stage is regulated chiefly

or entirely by gene products transcribed from the maternal

genes and present as messenger RNA in the cytoplasm of the

unfertilized eggs. Massive activation of genes in the

embryonic nuclei begins only at gastrulation (Denis and

Brachet, 1969a,bj Whiteley and Whiteley, 1975). Hybrid coral

embryos in the present study were cultured to the early

planula larva stage with no apparent reduction in viability

relative to intraspecific embryos. Hybrid larvae from

parallel crosses conducted with Acropora millepora, A.

pulchra, A. formosa, and A. cytherea subsequently settled

and metamorphosed, and are presently being raised to

maturity (3-4 yearsj Wallace, 1985) by Australian colleagues

under controlled conditions (B. Willis, pers. comm.), to

examine details of skeletal morphology and assess fertility.

22 Naturally produced hybrids may successfully develop but nonetheless fail to survive in nature. The greatest opportunity for the success of hybrids and hybrid derivatives is in habitats which are intermediate to those of the parent species (Anderson, 1949), or open habitats in which competition with parent species is reduced (Grant,

1981). Closed, stable communities may provide no such habitats, but where the natural community has been disrupted by disturbance, hybrids can become established (Grant,

1981). Natural hybridization between plant groups has most frequently been documented under conditions in which the environment has been disturbed by man (Riley, 1938; Briggs,

1962; Randolph, 1966), but open terrestrial habitats are also generated by natural abiotic phenomena such as landslides, floods, fires, and glaciation, as well as biotic factors such as immigration of herbivores or new disease organisms (Grant, 1981). In this study, corals which formed experimental hybrids frequently occur as neighbors in the same densely populated habitat, where intense competition for space may be disadvantageous to survival of any naturally formed hybrid larvae. However, episodic events such as storms, hurricanes, freshwater kill, tidal extremes, disease, and predation can disrupt existing communities, opening available space and otherwise altering the physical and biotic parameters of the habitat. The various sources and effects of disturbance on coral reefs of relevance to

23 the present topic receive more integrated treatment in

Chapter 3.

Hybrid plants and animals that do survive often fail to reproduce, either because they are sterile or have inappropriate behavior. A hybrid may be highly fertile

(Randolph et al., 1967) or may demonstrate complete to partial sterility (Stebbins, 1950; Grant, 1981). In an organism with high fecundity, however, even a very low level of fertility can generate a large number of viable gametes

(Stebbins, 1959). A partially fertile hybrid may reproduce sexually by sib crossing with sister hybrids, or backcrossing to one or both parental species. The resulting second-generation progeny can then continue to cross with one another and with previous generations. The result is a hybrid swarm, a variable mixture of species, hybrids, backcrosses, and later-generation recombinational types.

Hybrid swarms are known in many plant groups (Palmer, 1948;

Tucker, 1952; Lewis and Bloom, 1972; Grant and Grant, 1979;

Grant, 1981) and some animals (Sibley, 1954; Wilde and

Echelle, 1992).

Other workers have proposed the idea that hybridization among corals may operate on both the ecological and

evolutionary time scales (Hodgson, 1988; Richmond, 1988;

Willis et al., 1992). Hybridization between populations

having different ecological norms can have several kinds of

effects. One is the enrichment of the gene pool of a

24 species. Often a hybrid zone is quite narrow, so that each parental population has remained free of the genetic effects of hybridization; a narrow hybrid zone may nonetheless persist for long periods of time (Turner, 1971). In other cases some characteristics may filter from one partially isolated population into the range of the other by repeated backcrossing of a natural hybrid to one or both parental populations, a phenomenon known as introgressive hybridization (Anderson, 1949). Gene pool enrichment is most effective when hybridization is followed by introgression

(Grant, 1981; Shaw, 1981). Botanists have long recognized introgressive hybridization as an important mechanism in plant evolution (Anderson, 1949; Stace, 1980; Grant, 1981)

Though few zoologists have considered hybridization among animals as evolutionarily important (Mayr, 1963), recent genetic studies of invertebrates (Solignac and Monnerot,

1986; Spence, 1990; Spence and Gooding, 1990; Kaneshiro,

1990) and vertebrates (Dowling and Brown, 1989; Lehman et al., 1991; Dowling and DeMarais, 1993) suggest that introgressive hybridization has played a significant role in generating morphological diversity by providing additional genetic variation on which selectlon and drift may operate.

A second potential effect of natural hybridization is the production of hybrid swarms so extensive that boundaries between formerly distinct species are erased. Such hybridization is prevalent in disturbed habitats where the

25 ecological component of reproductive isolation has broken down. This effect has most frequently been documented as a result of habitat disturbance by man (Riley, 1938; Grant,

1981; Wilde and Echelle, 1992).

The third possible result of natural hybridization is hybrid speciation (Grant, 1981), the origin of a new species directly from a natural hybrid. If appropriate habitats are available, a small proportion of the descendants of a partly fertile species hybrid may become stabilized. Grant (1981) suggested several mechanisms that may act to reduce the disruption of hybrid gene complexes and give rise to new entities that are treated as separate species, including asexual reproduction, permanent translocation heterozygosity, allopolyploidy, recombinational speciation, and the production of a hybrid type that is isolated by external barriers. Recent genetic studies indicate hybridization has resulted in the origin of a new species in

Iris (Arnold et al., 1990 ), Helianthus sunflowers

(Riesenberg et al., 1990), and cyprinid fishes of the genus

Gila (DeMarais et al., 1992).

The influence of recent or ancient natural hybridization has been inferred for numerous animal and plant species complexes through morphological, demographic, and genetic data (Arnold et al., 1990; Arnold, 1992). Natural hybridization involves interbreeding between populations which have a previous history of divergence and which are

26 separated by partial ecological or reproductive isolation

(Grant, 1981). The low (4%) to moderate (30.6%) levels of interspecific hybridization in corals demonstrated by the present study (Table 2.2) suggest a previous history of divergence among the species involved in pairwise combinations, during which only partial reproductive isolation developed. Numerous episodes of separation of populations may have occurred during low sea level stands accompanying Pleistocene glaciations. Eustatic sea level change may have created a mosaic of ocean basins in which coral populations were isolated (McManus, 1985) and partially diverged, only to be united during subsequent transgressions. The subject of speciation processes during

Pleistocene sea level changes receives further treatment in

Chapter 3.

"SPECIES GROUPS" AND MOLECULAR DATA

Veron and Wallace (1984) arranged the 70 species within the subgenus Acropora (Acropora) which occur off eastern

Australia into fourteen "species groups" based on morphological similarities. The phylogenetic validity of these groups is unknown. The groupings can be viewed as taxonomic hypotheses, in which members of a group may be more closely related to each other than to members of outside groups, and the closer the groups in the series the closer their relatedness may be (McMillan and Miller, 1990).

27 According to this scheme, Acropora millepora and A. pulchra are placed in the same group, while A. formosa is placed in a different but nearby group.

McMillan and Miller (1990) used hybridization of cloned repeated sequences of DNA that are present throughout the subgenus Acropora, but not present in other genera in the family Acroporidae, to test the relatedness of species within and between groups. They found the relatedness series based on molecular data differed from that based on morphological criteria. Cloned repeated DNA sequences from

A. formosa, used as a molecular probe, hybridized strongly with genomic DNA from A. pulchra (McMillan et al., 1988) and

A. millepora, implying a close taxonomic relationship among these species (McMillan and Miller, 1990). Subsequent work sequencing nucleotides in this highly-repeated 118 base-pair sequence (McMillan et al., 1991) confirmed a close relationship among A. formosa, A. pulchra, and A. millepora.

The low (4%) to moderate (20%) levels of interspecific hybridization demonstrated by crosses involving these three species (Table 2.2) are in agreement with their close genetic affinity as revealed by molecular data.

ASYMMETRY OF FERTILIZATION

The hybridization experiments described here used both

female and male gametes from each colony, and could not

therefore discriminate for asymmetrical fertilization

28 relationships. More detailed experiments in which eggs and sperm spawned by individual colonies were separated prior to being combined with the unisexual ·gametes of another species demonstrate that fertilization relationships may not be symmetrical. For example, eggs of Acropora miIIepora were fertilized by sperm of A. formosa, but cultures which combined eggs of A. formosa and sperm of A. miIIepora showed no indications of fertilization (C. Wallace and B. Willis, pers. comm.). Similar asymmetrical relationships have been noted using the gametes of Acropora digitifera and A. gemmifera (R. Richmond, pers. comm.). If congruent with natural patterns of gamete interaction, these asymmetrical fertilization reactions have relevance for expected patterns of cytoplasmic genetic introgression. Spawned gametes are viable for about seven hours (Heyward and Babcock, 1986), while larvae may survive for several days (Babcock and

Heyward, 1986), weeks (Harrison et al., 1984; Richmond,

1988), and possibly as long as several months (Harrison et al., 1984) before settling. Larvae may travel considerable distances before settling (Oliver and Willis, 1987; Sammarco and Andrews, 1988). Maternally inherited cytoplasmic genetic elements (i.e., mtDNA) will therefore more readily introgress than cytoplasmic genetic elements carried by the paternal species.

In recent studies in plants, nonconcordance between phylogenies based on ribosomal DNA and those based on

29 chloroplast DNA has been interpreted as evidence of past introgressive hybridization (Arnold, 1992). Discordance between mtDNA-based phylogenies and those derived through other phylogenetic techniques have similarly been used to infer past introgressive hybridization in Drosophila

(Solignac and Monnerot, 1986), wolves and coyotes (Lehman et al., 1991), and cyprinid minnows (Dowling and Brown, 1989;

Dowling and DeMarais, 1993). Where introgression of cytoplasmic DNA has occurred due to past hybridization events, relationships based on mtDNA alone may not adequately reflect organismal phylogeny (Dowling and

DeMarais, 1993). The effect of asymmetric fertilization reactions on biasing mtDNA-derived phylogenies, and its usefulness in inferring past hybridization events, are of relevance as mtDNA phylogenies are increasingly refined for corals.

SPECIES CONCEPTS

The relationship between "species" as the fundamental unit of classification and "species" as a unit of evolution has occupied generations of biologists and generated an extensive body of literature (e.g., Mayr, 1942; Simpson,

1961; Van Valen, 1976; Paterson, 1982; Mishler and Donoghue,

1982). The Linnaean system of taxonomy was predicated on the objective reality of species, created as unchanging units.

Variation within species represented an imperfection in

30 conformity to an ideal, a concept which was codified in the working principle of the holotype, or type specimen.

"Typological" (Mayr, 1963) thinking dominated classical taxonomic practice, which either failed to recognize or sought to compartmentalize a continuum of morphological variability. More modern appreciation of the range of variation has led to numerous taxonomic revisions in many plant and animal groups, including corals (Vaughan and

Wells, 1943; Dinesen, 1980; Veron, 1986; Veron and Pichon,

1976, 1980, 1982; Veron et al., 1977; Wallace, 1978; Scheer and Pillai, 1983; Veron and Wallace, 1984; Hoeksema, 1989).

These revisions generally extended former boundaries and therefore had a simplifying effect, yet remained based on the paradigm that species are cohesive units with definable boundaries.

Contemporary thinking has been dominated by the biological species concept. Mayr (1942) viewed species as

"groups of actually or potentially interbreeding natural populations, which are reproductively isolated from other such groups." This concept does not embrace many organisms, however, including those in which sexual reproduction is not known, those abutting along hybrid zones, and morphologically disjunct types in fossil records. The evolutionary species concept (Wiley, 1978) attempted to broaden the biological species concept, viewing species as separate ancestor-descendent lineages with their own

31 evolutionary roles and fates. When discontinuities in the ability to interbreed are relatively complete, and coincide well with discontinuities in morphological and ecological variation, the application of biological or evolutionary species concepts can be adequately translated through existing taxonomic practice (Mishler and Donoghue, 1982).

Where observed morphological, reproductive, and ecological discontinuities do not correspond, however, as in morphologically discrete but interbreeding groups of plants, challenges to the biological or evolutionary species concepts have arisen.

Grant (1957, 1981) developed the concept of the syngameon, lithe sum total of species or semispecies linked by frequent or occasional hybridization in nature; [hence] a hybridizing group of species... ". The notion of the syngameon seeks to amend rather than abandon the biological species concept. A more inclusive unit of interbreeding than the species, the syngameon may vary from two or three semispecies to vast networks composed of numerous taxonomic species which form viable hybrids. Some semispecies may be central in a syngameon and have numerous linkages, while others are peripheral with only one linkage. Some semispecies may hybridize extensively, some to a moderate extent, and some only slightly. The components are usually good taxonomic species as treated in formal systematics.

Syngameons are developed in many plant genera, including

32 Betula, Nothofagus, Iris, Pinus, Juniperus, and Quercus

(Grant, 1981). The concept has been recently applied to a morphologically diverse group of minnows with a pervasive influence of hybridization throughout their evolutionary histories (Dowling and DeMarais, 1993). This concept could be similarly applied to the network of linkages demonstrated by the Australian acroporids examined in the present study.

Acropora pulchra showed moderate (,20%) levels of hybridization with both A. cytherea and A. millepora, which in turn hybridized at low (4%) levels with A. formosa.

Ongoing and future experiments which examine various pairwise combinations of species promise to extend and refine the structure of hybridization relationships among

Acropora.

Debates over species concepts have received renewed momentum in recent years, fueled by a need for a concept consistent with phylogenetic principles (Mishler and

Donoghue, 1982; Cracraft, 1983; Nixon and Wheeler, 1990).

Many different concepts, often with different consequences for the description and explanation of temporal and spatial patterns of diversity, have been proposed. From the proliferation of concepts has emerged a "phylogenetic species concept", by which species are defined as "the smallest aggregation of populations (sexual) or lineages

(asexual) diagnosable by a unique combination of character states in comparable individuals" (Nixon and Wheeler, 1990).

33 Phylogenetic species are terminal taxa which cannot, on the basis of available character evidence, be subdivided into other taxa (Cracraft, 1992). This definition removes the process-orientation, which is implicit in the biological species concept, from application to taxonomic practice, as pattern, rather than process, is what taxonomists can observe (Nixon and Wheeler, 1990). The phylogenetic species concept has recently been usefully applied to diverse and complex groups including birds-of-paradise (Cracraft, 1992), rodents (Engstrom et al., 1992) and grasses (Davis and

Manos, 1991). Interspecific hybridization in Acropora challenges the appropriateness of the biological species concept, and suggests that other approaches such as the phylogenetic species concept might be more suitable.

Despite recent successful applications of the phylogenetic species concept, however, a phylogenetic system of taxonomy, starting from evolutionary first principles, has yet to be developed (de Queiroz and Gauthier, 1992). By contrast, a modified Linnaean system, based on a pre­

Darwinian world view, remains in conventional usage. While some authors have argued that "species" as a basic taxonomic unit can be decoupled from "species" as a basic evolutionary unit (Veron, 1981; Mishler and Donoghue, 1982), others seek

to develop a phylogenetic taxonomy in which a particular

system of names represents entities that derive their

existence from a particular set of phylogenetic

34 relationships (de Queiroz and Gauthier, 1992). While the latter authors do not suggest the term "species" necessarily be used to describe terminal taxa, the inference is that a basic concordance must exist between evolutionary principles and a taxonomic system. A phylogenetic system of taxonomy must, in essence, be based upon a phylogenetic species concept.

The phylogenetic species concept recognizes terminal taxa ("species", Cracraft, 1992) as the units of evolutionary diversity, that is, those units produced by the various processes of differentiat{on. In analyzing the birds-of-paradise (Paradisaeidae) according to the phylogenetic species concept, Cracraft (1992) postulates 90 species which have diversified within Australasia compared to 40-42 postulated by previous workers using the biological species concept. Species boundaries in corals are perhaps best defined by multiple independent methods, including behavioral, biochemical, and micromorphological criteria

(Lang, 1984; Willis, 1990). Should the phylogenetic species concept and its incipient taxonomic interpretation (de

Queiroz and Gauthier, 1992) gain usage among scleractinian corals, the number of diagnosablydistinct end products of

Acropora diversification promises to be reinterpreted as additional reproductive, molecular, ultrastructural, and cytogenetic information accumulates.

35 CHAPTER 3

DETERMINATION OF CHROMOSOME NUMBER

INTRODUCTION

As calcifying organisms which harbor endosymbiotic zooxanthellae, reef corals have been viewed over the centuries as mineral, vegetable, or animal (Jokiel, 1987).

The sYmbiotic relationship between corals and intracellular dinoflagellate algae renders hermatypic corals functionally autotrophic, with rates of carbon fixation comparable to many marine plants (Jokiel and Morrissey, 1986). Asexual propagation through fragmentation or fission is an important reproductive strategy demonstrated both by plants (Stebbins,

1950; Grant, 1981; Silander, 1985) and corals (Jackson,

1985; Hughes and Cancino, 1985; Jackson and Hughes, 1985;

Hughes et al., 1992). Parallels between plants and corals may be further drawn for the genus Acropora, in which the axial polyp is reminiscent of the apical meristem of plants.

The main feature defining Acropora is an axial polyp at the growing tip, from which radial polyps bud off. Radial polyps have the potential to take up an axial role (Wallace, 1978).

Although the genus Acropora has received the taxonomic attention of two recent revisions (Wallace, 1978; Veron and

Wallace, 1984) which have clarified many previous problems of identification, no scientific attention has been given to genetic, ecological, or evolutionary mechanisms which

36 contribute to the uniquely speciose nature of this coral genus. Nearly twice as many species are presently recognized in Acropora (150) as are found in the next most prolific genera, Montipora and Porites, each with 80 species (Veron,

1993) (Table 1.1). The majority (83%) of the 109 extant hermatypic coral genera have fewer than ten species per genus.

Polyploidy has long been recognized as an important mechanism of speciation in the plant kingdom, but is thought to be rare in animals. Stebbins (1950) estimated 30-35% of modern angiosperms are polyploids, while more recent estimates suggest the frequency of polyploids in angiosperms may be as high as 70-80% (Goldblatt, 1980; Lewis, 1980b).

Might polyploidy be a cytogenetic mechanism which has also operated in the evolution of Acropora?

Only two previous studies have attempted to examine scleractinian coral chromosomes. Wijsman and Wijsman-Best

(1973) used somatic tissue from adult colonies, but were unable to establish chromosome number for any species.

Heyward (1985b,c) used rapidly dividing cells of externally­ developing coral embryos, and established a diploid number of 28 for four species from three families. In this chapter

I present chromosome counts for 22 species of Acropora, using mitotic cells of externally-developing embryos.

Chromosome numbers were established to test my hypothesis that polyploidy has contributed to the diversification of

37 this speciose genus. For comparison with less speciose genera, chromosome numbers were also established for three

species of Montipora and one Fungia species. Included among

the 22 species of Acropora are five species which form

experimentally-generated hybrids (Chapter 2). Chromosome

numbers were additionally determined for hybrid embryos.

MATERIALS AND METHODS

STUDY SITES AND IDENTIFICATION OF CORALS

Study areas for Acropora were chosen on the basis of

having abundant populations within access of established marine laboratories (Table 3.1). Corals in Guam were

collected from Pago Bay and maintained at the University of

Guam Marine Laboratory. Corals in Okinawa were collected

from fringing reefs off Motobu-cho and Sesoko Island, and

maintained at the Sesoko Marine Science Center. Corals in

Australia were collected from Geoffrey Bay, Magnetic Island,

15 km off Townsville. Corals in Palau were collected from a

shallow fringing reef off Malakal Island (A. carduus) or the

upper slope of the barrier reef outside Ulong Channel, and

maintained at the Micronesian Mariculture Demonstration

Center. Corals in Hawaii were collected from Kaneohe Bay and

maintained at the Hawaii Institute of Marine Biology.

38 TABLE 3.1. Geographic areas from which species of coral used in determining chromosome number were collected.

Guam Okinawa Australia Palau Hawaii

GENUS Acropora danai digitifera tenuis carduus valida monticulosa samoensis clathrata surculosa nobilis formosa florida lutkeni divaricata pulchra ocellata millepora squarrosa cytherea digitifera elseyi robusta gemmifera

GENUS Mon tipora spumosa verrucosa digitata GENUS Fungia scutaria

Corals were identified by reference to field guides

(Wallace, 1978; Randall and Myers, 1983; Veron, 1986;

Nishihira, 1988; Uchida and Fukuda, 1989), museum specimens

(University of Guam Marine Laboratory, Mangilao; Bishop

Museum, Honolulu; Museum of Tropical Queensland, Townsville,

Australia), and to taxonomic experts (R. Randall; C.

Wallace). Voucher specimens of corals from Guam, Okinawa, and Palau were deposited in the collections of the Bishop

Museum in Honolulu. Voucher specimens of corals from

Magnetic Island were deposited in the collections of the

Museum of Tropical Queensland in Townsville, Australia.

39 COLLECTION OF RIPE COLONIES

Records of previous spawning events (C. Hunter, pers. comm.; R. Richmond, pers. comm.; Harrison et al., 1984;

Willis et al., 1985; Babcock et al., 1986; Heyward et al.,

1987; Richmond and Hunter, 1990; B. Willis, pers. comm.; D.

Krupp, pers. comm.; E. Cox, pers. comm.) and presence of colored eggs in fertile colonies (Harrison et al., 1984;

Babcock et al., 1986; Heyward et al., 1987; Heyward, 1988) were used to predict spawning dates in Guam, Okinawa,

Australia, and Hawaii. No coral reproductive data were available for Palau, so preliminary sampling to determine the time of maturation and spawning was undertaken for twenty Acropora species (Chapter 4). Colored eggs, an indicator of maturation in Acropora and Montipora, can be readily seen by breaking open a small, distal portion of a colony in the field and examining the broken ends. Although the presence of motile spermatozoa in testes squashes has been used to refine predictions of spawning readiness

(Harrison et al., 1984; Babcock et al., 1986; Heyward,

1988), this method did not yield uniform results and was not generally used in this study.

For each colonial species from Guam, Okinawa, Palau, and Hawaii, several ripe colonies or viable portions of colonies were collected several days before a predicted spawning date and maintained in large, concrete flow-through aquaria. As the date of predicted spawning approached,

40 colonies were checked nightly for indications of spawning activity. No laboratory facilities were available at the

Australian site, so colonies were brought ashore shortly before sunset, placed in individual containers on the beach, and returned to the reef if signs of spawning did not occur by 11 PM. Numerous colonies of the solitary coral Fungia scutaria have been maintained for several years in shallow flow-through aquaria at the Hawaii Institute of Marine

Biology, and their annual spawning schedule was accurately known (D. Krupp, pers. comm.).

COLLECTION AND PREPARATION OF GAMETES

Acropora and Montipora are simultaneous hermaphrodites, and spawning consists of the release of eggs and testes, which have been packaged together into discrete bundles, from the mouths of fertile polyps. Bundles can be seen protruding beneath the polyp mouths from one to several hours before they are released. Packaged egg and sperm bundles are buoyant, and upon release from the polyp mouth they float to the surface. There they dissociate into separate eggs and sperm, generally within 10-30 minutes.

Dissociation can be hastened by mild agitation. Spawned gametes were collected from the water surface by skimming with a small beaker or gathering with a wide-mouth glass pipette. The tips of all glass pipettes used to collect and transfer eggs and embryos were fire-polished, to reduce

41 damage to the egg membrane (C. Hunter, pers. comm.). In some experimental procedures, eggs were separated from sperm in spawned gamete bundles by rinsing in a gentle stream of 0.45

~m Millipore-filtered seawater (MFS) on a Nytex mesh screen

(R. Richmond, pers. comm.). Acropora eggs were washed using a 320 ~m mesh size, while smaller Montipora eggs were washed using a 160 ~m mesh. Diameters of 25-100 freshly-spawned and washed eggs were measured for most species, using a dissecting microscope with a calibrated ocular micrometer.

Several methods of experimentally controlling gamete encounter were tried:

(1) When egg/sperm bundles were seen protruding from the polyp mouths, conspecifics were placed in a single, separate Plexiglas aquarium with no water flow and allowed to spawn together. Thirty to 60 minutes after spawning ceased, eggs were collected and pipetted into a culture container (Table 3.2) filled with MFS.

(2) Conspecifics were placed in separate Plexiglas aquaria and allowed to spawn. Spawned gametes were collected from individual colonies and rinsed on a Nytex mesh screen with MFS. The effluent of each washing procedure, containing sperm from the dissociated egg/sperm bundles, was collected in a beaker beneath the mesh screen. Washed eggs from each of several colonies were pipetted into a culture container

42 TABLE 3.2. Summary of results of methods used in culturing experimentally-generated coral embryos.

Culture Method of Agitation Results Container

1 L plastic bottles placed in mesh dive bag Poor with Nytex mesh and attached to buoyed across top line on reef

500 ml plastic airlifted-droplet Poor beakers stirrer*

1 L wide-mouth glass swinging paddles on a Fair jars rack*

250 ml glass mechanical wrist Good Erlenmeyer flasks agitator beakers, aquaria no agitation Variable

1 L plastic screw-cap placed in mesh dive bag Excellent bottles with closed and attached to buoyed tops line on reef

1 L plastic screw-cap placed in mesh dive bag Excellent bottles with closed and suspended from tree tops branch in high wind

1 L plastic screw-cap Wheaton modular cell Excellent bottles with closed production roller tops apparatus

*Strathmann, (1987)

43 filled with MFS (Table 3.2), and several drops of diluted sperm from each of the washing procedures were added.

(3) Conspecifics were placed in separate Plexiglas aquaria and allowed to spawn. Spawned gametes were collected from individual colonies and then combined in a culture container filled with MFS (Table 3.2). After 15-30 minutes, combined gametes were rinsed on a Nytex mesh screen with MFS in order to reduce sperm concentration in the culturing medium. Rinsed eggs were returned to a fresh culture container with fresh MFS.

Eggs and sperm are viable for about seven hours (Heyward and Babcock, 1986). To preclude contamination from gametes in seawater pumped into flow-through aquaria, seawater for

Plexiglas aquaria in which spawning corals were placed was collected during the morning or early afternoon of predicted spawning nights. Additional care was taken to avoid contamination by using separate equipment (pipettes, bulbs, beakers, jars) for each species on any given spawning night.

Equipment was thoroughly rinsed in fresh water before subsequent usage on another spawning night.

Corallites of the solitary coral Fungia scutaria were isolated in individual finger bowls several hours before predicted spawning times. The small «100 ~m), negatively­ buoyant eggs of this gonochoric species were collected from

the bottom of the finger bowl with a glass pipette. Unwashed

44 eggs from several colonies were pipetted into a culture container filled with MFS, and several drops of seawater into which a male colony had spawned were added from each of several male colonies.

EMBRYO CULTURE

Several combinations of culture containers and agitation regimes were tried, depending on available equipment and sea conditions (Table 3.2). While embryonic development was generally facilitated by mild agitation, in some cases large numbers of embryos developed with no external agitation of the containers in which gametes were mixed. Excessive agitation, such as that produced by incubating culture containers in rough seas, gave inferior results. Mild agitation sufficient to keep embryos suspended below the surface of the water in the culture container gave the most consistent and best results.

EMBRYO TREATMENT

Embryos were treated with colchicine, which disrupts formation of the mitotic spindle, followed by a hypotonic solution to spread the chromosomes, and a fixative for preservation. In treating embryos for chromosome analysis, the manipulated variables were: (1) time between mixing gametes and beginning of colchicine treatment [age of embryo], (2) concentration of colchicine, (3) duration of

45 treatment with colchicine, (4) concentration of hypotonic solution, and (5) duration of treatment with hypotonic solution. Various combinations of treatment parameters were tried (Appendix A-1) in order to optimize results. The ranges of values of manipulated variables included:

(1) time between mixing of gametes and beginning of colchicine treatment: 4.75 - 12 hours;

(2) concentration of colchicine-seawater: 0.02 or 0.05%

(w/v) ;

(3) duration of treatment with colchicine: 30-180 minutes;

(4) concentration of the hypotonic solution: 80:20, 70:30, or 65:35 seawater-tapwater;

(5) duration of treatment with hypotonic solution: 15-30 minutes.

Treated embryos were fixed in three changes of freshly­ mixed absolute ethanol-50% glacial acetic acid (1:1, v/v) and refrigerated during storage.

STAINING AND SQUASH PREPARATION

To remove lipids prior to staining, fixed embryos were soaked in diethyl ether in small stainless steel cups for 4­

6 hours, then briefly returned to fixative. Several embryos were placed on a glass slide cleaned in 95% ethanol, stained with 2% lacto-aceto-orcein or acetocarmine for 15 minutes,

46 gently rinsed with tapwater, and squashed under a cover slip. Dehydration of squash preparations was retarded for several days by sealing the cover slip edges with clear nail polish. Giemsa staining according to the cell preparation and staining techniques of MacGregor and Varley (1983) and

Weinberg et ai. (1990) was also tried.

Squashes were scanned for cells in which the chromosome complement was clearly distinct from that of other cells and the chromosomes were well spread. Chromosomes were counted in each of 45 metaphase spreads for each species, with the exception of Acropora gemmifera embryos, for which only 25 useful metaphase spreads could be found. Twenty-five metaphase spreads were also counted for the hybrid A. miiiepora x A. formosa, for which only a small number of embryos was available. For each spread, a lower and an upper limit for the number of chromosomes present was recorded, and a frequency distribution was constructed (Rahat et ai.,

1985) .

PHOTOGRAPHY

Representative complements and their individual chromosomes were photographed at 40X and 100X, respectively, using phase contrast optics, a green filter, and Kodak

Technical Pan film 2415 developed for maximum contrast.

47 RESULTS

In Guam, spawning from lab-maintained colonies of

Acropora generally occurred between 2030 and midnight, four to eight nights after the July full moon. In Okinawa,

Acropora spawning occurred between 2200 and 2330 on the full moon of late May and one night later. Major Acropora spawning at Magnetic Island, Australia occurred four nights after the October full moon, commencing at 1800 and peaking around 2100. Spawning from Montipora and some Acropora occurred over the preceeding three evenings. Species of

Acropora which were observed spawning in Palau did so four to seven nights after the Mayor June full moon. The mean size of freshly-spawned, washed eggs from 25 species of

Acropora ranged from 424 to 722 ~m (Appendix A-2) .

First cleavage occurred 1.5-2 hours after gametes were combined. During the first 12 hours of embryonic development, embryonic size remained similar to that of the unfertilized egg. Development proceeded by sequential cell divisions accompanied by a decrease in individual cell volume. The most useful chromosome preparations were those in which 10-II-hour-old embryos were cultured with colchicine for 2 hours, followed by a 65:35 seawater:tapwater hypotonic treatment. Younger embryos had fewer cells of larger volume, in which the chromosome complement was more difficult to locate. Embryos older than

11 hours had more cells available for viewing the mitotic

48 nucleus, but the smaller size of cells and closer proximity of nuclei made it more difficult to distinguish separate chromosome complements with confidence. Ten- to Il-hour-old embryos had flattened into concave discs, making them very suitable for squashes. Embryos began gastrulation about 11 hours after fertilization (Heyward, 1987), which reduced the proportion of cells undergoing mitosis.

Both lacto-aceto-orcein and acetocarmine provided useful differential staining of chromosomes. Attempts to view banding patterns by Giemsa staining in order to match chromosome pairs were not successful. In this method, embryonic cells must first be dissociated into a cell suspension by repeated flushing in 60% acetic acid with a

Pasteur pipette. Fixed embryonic cells tend to stick to glass, which resulted in massive cell loss on the interior of the pipette. Additionally, lengthy fixation hardens the embryo, such that cells resist dissociation into a suspension even with stronger acid concentrations.

The two different colchicine concentrations did not result in visible differences in chromosome appearance.

Lengthy treatment of mitotic tissue with colchicine can induce polyploidy (Dewey, 1980), as chromosomes replicate but do not move to opposite poles in the absence of a mitotic spindle. In some embryos of some species subjected to colchicine treatment times of 180 minutes, indications of

induced polyploidy could be seen. This consisted of a few

49 cells with a large number of very small chromosomes. The small size of the chromosomes, the infrequency of such duplicated complements, and comparison with cultures treated for shorter periods of time made it easy to distinguish cells in which polyploidy had been induced by the prolonged colchicine treatment.

Embryos of different species frequently yielded chromosome preparations of very different quantity and quality under the same treatment parameters, suggesting subtle variations in the timing of the mitotic cycle. For example, Acropora digitifera embryos contained many metaphase cells while most A. nobilis embryonic cells were in interphase, despite identical embryo age, culture, and treatment conditions. There was considerable inter- and intra-specific variation in the degree of chromatin condensation. For example, A. digitifera typically had embryos with rod-shaped chromosomes, while A. surculosa cultured under identical conditions had embryos with few mitotic cells and tightly-condensed chromosomes. Even within a single treatment culture, chromosomes varied from tightly condensed "dumbbells" to more elongate rods to incompletely condensed threads. Reliable, comparative chromosome morphologies could not be established due to this inconsistent treatment effect among and within species. In all species in which complements with compact, rod-shaped

50 chromosomes were found, chromosomes ranged from 1-5 ~m in length (Figure 3.1).

Most metaphase spreads had some degree of chromosome overlap, such that each individual chromosome could not always be uniquely resolved and counted. Additionally, the hypotonic solution which spreads the chromosomes may cause an artificial loss or gain of some chromosomes from an adjacent complement (Rahat et ai., 1985). Consequently, a range of chromosome counts was obtained for all species. For each species, however, the modal number in the frequency distribution of metaphase spread chromosome counts was readily evident (Figure 3.2) and was chosen as representing the number of chromosomes for that species (Rahat et ai.,

1985). There was no association, for any of the species studied, between variation from the modal number of chromosomes and individual embryos from which counts were obtained. The number of embryos which were examined and used to establish chromosome number for each studied species of

Acropora is shown in Table 3.3.

Somatic chromosome numbers in Acropora ranged from 24 to

54 (Table 3.4). Sixteen species had 28 chromosomes.

Chromosome numbers of 24, 30, 30, 42, 48, and 54 were established for the six other species.

Fungia scutaria, Montipora verrucosa, M. spumosa, and both morphs ("fat fingers'! and "yellow spatulate") of M. digitata all had 28 chromosomes (Figure 3.3). Montipora eggs

51 Figure 3.1. Photomicrographs of somatic chromosomes from seven species of Acropora. Due to overlapping, not all chromosomes in a complement may be clearly visible in a single focal plane. Fine focal adjustments are made while making counts. All embryos from which complements were photographed were treated with 0.02% colchicine. Age of embryo (hours)/duration of colchicine treatment (hours) is noted in brackets for each photo. A. Acropora cytherea, 2n=28 [11.0/2.25]. Chromosomes appear as condensed rods, many of which are metacentric. B. Acropora cytherea, 2n=28 [11.0/2.25]. Chromosomes ll appear as tightly-compacted IIdumbbells • Complement is from the same embryo as in photo IIAII. C. Acropora surculosa, 2n=28 [10.5/3.0]. Chromosomes have intermediate ll appearance between condensed rods and compacted IIdumbbells • D. Acropora danai, 2n=24 [8.25/1.5]. Chromosomes appear as condensed rods, with several areas of overlap. E. Acropora gemmifera, 2n=30 [10.5/3.0]. Chromosomes appear as short rods and dots lJl l\) (arrows), with some overlap. F. Acropora valida, 2n=42 [10.75/1.5]. Chromosomes ll appear as tightly-condensed IIdumbbells • G. Acropora ocellata, 2n=48 [8.0/2.0]. Chromosomes appear as incompletely-condensed rods, with several areas of overlap. H. Acropora elseyi, 2n=54 [10.5/2.25]. Chromosomes appear as tightly-condensed rods with several areas of overlap. 53

Figure 3.2. Somatic chromosome counts for twenty-two species of Acropora based on metaphase spreads of colchicine-treated embryonic cells. Each vertical bar represents one spread. For some spreads more than one interpretation of chromosome number was possible.

55 Acropora dana/ Acropora digitifera

illIG ~ :: III !ae I !: 111 11 11"111" II II 1111111111 II II """1 I: 111111111111111111111111111111" /1111 j8 : 1"" B.. l..--.-I _ ~ 24 ~ =" 23 '------GUAM Ol

Acropora dig/tlfera Acropora monticu/osa

illIG 8o lIG !:1111111111111111111111111111111111111111111 f: 111111111111111111/1111/111111 II H Rae o 11111 ~ae i: ~ 25 II 8 ~ 24 ~ 24 I 23 L--- _ '------: 23 GUAM OKINAM

Acropora samoensls Acropora robusta r------.------illIG 8 30

~ ~ 1 1111" 111111111111111111111111111111111 !: II 1111" 11111111111111111111111111111 ~2111 ~ : 1"1/ II II ~ 24 ~ 24

~ 23 ~ 23 '------'------GREAT BARRIER REEF GUAM

56 Acropora nobilis Acropora formosa

S 30 o

~~ III 11111I111111111I 1I1l111111U1111~ ~ ~llIIIIIIIII,lIlIulUnllllll;;Il;[IIIII. I !H R 26 ~ ~ 26 26 11 ... S S S 2. S 2' ; ~3 L------' ~ 23 L------'

OKINAWA GREAT BARRIER REEF

Acropora pulchra Acropora millepora

S 30 !:: 1IIIIIIIII11111111111111111111110lL UII I. is::2. ~ 23 L-- ~

GREAT BARRIER REEF GREAT BARRIER REEF

Acropora tenuis Acropora cytherea

S 30 o

""1/1111"""11I"""11111111 II gtlllllllll~I~IlIIII1IIlIllllllli111IIInoI !::H R 26 o ~ 26\ .. ~ 25 S ~ 26 ...... •...... •...... 1 S 2' 24 .d...... _... . . S l.--- __ ~ 23 lL- _ ~ 23

GREAT BARRIER REEF GREf>:r BARRIER REEF

Figure 3.2. (continued) 57 Acropora lutkeni Acropora surculosa

S 30 o III ~I ~~II III WIUIIIII IIl1J 1111 lUll II u. . ! 1IIIIIIIIIIIIIIIJlJIIIHIIIIIIIIIII I M ~ 25 o 25 S - - ..-- - - -.-- -. S 3 24 2 4 e 3 _ ------__ ._------23 ~ 23 L..------..I s L------.J

GUAM GUAM

Acropora clathrata Acropora carduus -.-.------n S 30 S 30

11111111111111111111I11111111111111111111111 ~~ !:: i II 11111111111111111I111111111I11111111111 ~ 26/ ~ 25 . ~ 25 S ~ 24 3 24 ~ ~ 23 '------.J 23 l-- --J

PALAU PALAU

Acropora squarrosa Acropora florida

S 30 ;~ I ! 111I1I1111111I111I1111Il1I11111\\111 E111111111111111111111111111111111 lililI 26 11 ~ 1 ~ 26/1 26 11 ~ 25 S ~2' ~ S 23 '------I::1 I

GUAM PALAU

Figure 3.2. (continued) 58 Acropora gemmifera Acropora divaricata

~ 33 M ~i-~IIIIJI[ I 1I111111l1111 ~ ar II _ 5 26 '------' GUAM OKINAWA

Acropora valida Acropora ocel/ata

S 45 ~ ::1------·------ll

T I 61 C l~: 60 III 1111111111111111111111111111111111 ~11I~1I~1!1~11.1 a39 111m ..l.. I_ , E !::46 s S 38 U---__---' WJ..L1.l--_~_J

GUAM GUAM

Acropora elseyi

67 g ... M " 66 ~~ ! II 1I11 11I1111111111111111 s a61 E S 60 l.LL1--__--'

GREAT BARRIER REEF

Figure 3.2. (continued) 59 TABLE 3.3. Number of fixed embryos which were examined, and the number of these which were useful for determining chromosome number. Except where noted, chromosome counts from 45 metaphase spreads were made for each species or hybrid. Cultures are listed in the approximate order in which they were examined.

Species/Hybrid No. embryos No. embryos % examined useful useful

Acropora danai 60 28 46.6 A. valida 32 12 37.5 A. digitifera (Okinawa) 34 18 52.9 A. monticulosa 72 18 25.0 A. divaricata 56 16 28.6 A. nobilis 54 24 44.4 A. surculosa 40 20 50.0 A. lutkeni 104 30 28.8 A. ocellata 108 36 33.3 A. squarrosa 64 28 43.8 A. gemmifera* 142 14 9.9 A. digitifera (Guam) 25 12 48.0 A. robusta 48 12 25.0 Fungia scutaria 44 25 56.9 Montipora verrucosa 16 12 75.0 Montipora spumosa 96 21 21. 9 Acropora formosa 114 18 15.8 A. millepora 60 13 21. 7 A. millepora x A. formosa** 13 5 38.5 A. pulchra 56 22 39.3 A. millepora x A. pulchra 59 17 28.9 A. cytherea 31 15 48.4 A. millepora x A. cytherea 46 18 39.1 A. tenuis 108 24 22.2 A. elseyi 102 36 35.3 A. samoensis 72 16 22.2 A. carduus 55 20 36.4 A. clathrata 32 17 53.1 Montipora digitata "fat fingers" morph 36 21 58.3 "yellow spatulate" morph 35 18 51.4 Acropora florida 30 17 56.6 A. clathrata x A. florida 22 15 68.2 *Chromosome counts from only 25 metaphase spreads were determined for this culture **Only 13 embryos were available from this culture, providing chromosome counts from 25 metaphase spreads

60 TABLE 3.4. Somatic chromosome number established for twenty­ two species of Acropora.

Locality Species Number of Somatic Chromosomes

Guam danai 24 Guam, Okinawa digitifera 28 Okinawa monticulosa 28 Great Barrier Reef samoensis 28 Guam robusta 28 Okinawa nobilis 28 Great Barrier Reef formosa 28 Great Barrier Reef pulchra 28 Great Barrier Reef millepora 28 Great Barrier Reef tenuis 28 Great Barrier Reef cytherea 28 Guam surculosa 28 Guam lutkeni 28 Palau clathrata 28 Palau carduus 28 Guam squarrosa 28 Palau florida 28 Guam gemmifera 30 Okinawa divaricata 30 Guam valida 42 Guam ocellata 48 Great Barrier Reef elseyi 54

61 Figure 3.3. Somatic chromosome counts for three species of Montipora and one species of Fungia based on metaphase spreads of colchicine-treated embryonic cells. Each vertical bar represents one spread. For some spreads more than one interpretation of chromosome number was possible.

62 Montipora verrucosa Montipora spumosa

So 30 ~ 211 T I 2S C f: =-=-mlrr~~~llU[Ulll~~ C 27 H ! S 28 . __ ._ __._.._.._..._ _ _. .. s 211 11111. .__ ~ 26 ~215 5 ------..-.- - --.-.... . ------s·-----··--·-·------··------·----- ~ 24 __ . .._._._ _ __.._ _ __ ._._.__ . ~ 24_. . ._. _ ~ 23 L..------' ~ 23 '------' HAWAII GREAT BARRIER REEF

Montipora digitata Montipora digitata ·fat fingers· morph ·yellow spatulate· morph !~mli[lrnl,lllnJ;JI;i,[-,-,I[11 mf ]]~~~ g=~=-I-I··I··l-I--I-I-~I··llllill'ui[[-[;llmulnm~ 11 111 R 211 H .11 ------.- o -..- ---- _ _.._._ R 211 ~ o .. ------··------t 26 ~ 26 _ g 24 ~ M ·~ __·,····_·_··,,···.... M ••••• ····_··M'·····_.. ·· ....._·._,._•• _ ••• "'_'M" _ •. "M'" _"...... __ ,•••••••••••_ •• __.... 24 M -_.- ..-. -.-.----...-.••-----.---.--.-..-----. E 23 S L..-- ---' SE 23 '-- ---'

GREAT BARRIER REEF GREAT BARRIER REEF

Fungia scutaria !~~ ~-lllllllHlI;IU[JIIUUlIII[II]tllU oR 211 ~ 211 S ~ 24 1···_·--_·······_·_···_······················· ~ 23 L..-- ---'

HAWAII

63 contain zooxanthellae when they are spawned (Babcock and

Heyward, 1986). Zooxanthellae in Montipora embryos did not appear disrupted by the colchicine and hypotonic treatments, and did not obscure clear views of coral chromosome complements. Acropora and Fungia embryos have not yet acquired the symbiotic zooxanthellae which live inside adult tissue (Krupp, 1983; Babcock and Heyward, 1986; Richmond,

1988). All of the experimentally-generated hybrids (Chapter

2) had 28 chromosomes, as did their parent species (Figure

3.4) .

DISCUSSION

The somatic chromosome number of 28 established here for sixteen species of Acropora, three species of Montipora, and one species of Fungia is the same as that reported by

Heyward (1985b,c) for three other species of scleractinian coral (Goniopora lobata, Lobophyllia hemprichii, and

Montipora dilitata). Together these represent four different families from three suborders. Twenty-eight appears to be a highly conserved somatic chromosome number amongst the corals. Against this background, the disparate somatic numbers of 24, 30, 42, 48 and 54 found for six species of

Acropora stand out in sharp contrast. It is proposed that this variation in somatic chromosome number arose through the combined processes of polyploidy and aneuploidy.

64 Figure 3.4. Somatic chromosome counts in four interspecific hybrids of Acropora based on metaphase spreads of colchicine-treated embryonic cells. Each vertical bar represents one spread. For some spreads more than one interpretation of chromosome number was possible.

S 30 Mo ------~~ ]Jf 111111IJIIIII1II1II~IIIll1nlll.ll~ ~ 11._ .-. A. millepora x A. formosa i g" --- ._------

S 30 ~ 24 o ------_.--~--_...._--- --_..---- E S 23 M 29 ! I A T - .. -lr-- I 28 _-[~~jl1minnmI ._:==~___:- ..: GREAT BARRIER REEF . ... _.-. C (J) --- --.... --_.- (J) C 27 • ___4 ____ Acropora formosa H ------._------R 26 o ------g 30 ~ --- 25 ~ ------.------29 S T o 24 I 28 ___~_ 4_ C -11[rnIIJrnil[[IJ]ill~IIIU M ------...--.-.. :]IIJ II C 27 ~ 23 H - -_.-.__._-- ...._.~. __._- Ro 28 M IC~~~~=~------_·:::=-~:::.::.~·- o 25 GREAT BARRIER REEF S - -_. __._------24 3 ------E ------S 23

GREAT BARRIER REEF Acropora millepora

S 30 ~o 29 .-.------

A. millepora x A. pulchra ;:: •. nul [IIUIIIlIIIIUJ "11111[11 [IIIU ....

~R25 26 I. .. -...... -- .--.- -- s ... - .... S 30 o ~ --.-----.-- _.-- .-.------.------.--... -- n· 3 " . 29 E···· T S 23 I 28 ~. ~ ~ C .. .-- . IllHIIUrnUIHIHrHIHI l. 0"1 GREAT BARRIER REEF ...,J C 27 H ~_~_-~~~_. oA 26 .1 ... .-:-:: .Acropora pulchra . ~ 25 S __4 • ••__•• ,_ •• _ .•• , •• .. ._. .. • __ " "_' _._._._. _._ •••• ,. __._ ._.__••• __ •••.. _ •• __ ._. ._. _.__._ 30 -- .. _ .._-.- --.- -.---.-.-- _11 24 g o ~ 29 M -----,-----_...__ .- --'---'--'-'-"~'~-- -_.__.-._------_.. _.,---._,..... -_._._...- ...... _-"-- - ~ 23 ~ 28 =-:-=,.IIIIIlUUIIJIIJJIIITUIIUllllll. _..... c 27 H -·-11·· ..-....-.... ----- .-.--. --_._-... --.- .-...-.. -... -.-. R 28 . GREAT BARRIER REEF o -'.1-.. -..-.-.. - ...------..--.--.-- .------..-..- .... o 26 . M8 _1.. ------.-.------.--._. Mo 24 . ._. . . ._...... _.. . ~ 23

GREAT BARRIER REEF Figure 3.4 (continued) Acropora pulchra

S 30 o ---.-. - -- ..-- -.-. -.-.--.. -.. -.---..---. --.-...-.. ------... - -. M 29 -----·---·.--··---1 A A. cytherea x A. pulchra r 28~-~~~~1I11 UlUII] In 111 11111 II IJIIIIJIIII ._ c 27

S 30 ln~:~" =~ ..- -_.- I:: o ------M o 24 A 29 M 1.1------._ T ~ 23 I 28 C GREAT BARRIER REEF (J) co C 27 H R 26 Acropora cytherea

.0 ---- -_.-----_ ..- _.'... ~-~-- M o 25 g 30 ... _._---- ..__ .-...... - _. _._----- _.__..._--_.__._...--,. ---.---+- _._~----~------S ------.---- _------~--- ~ o 24 29 T M ------~ .. ------E ~ 28 S 23 ]~][]]]]~J~II~~~!l~l~JI~~I_ C 27 lfillJll H R 28 o - -.------_.._-_._------_._---_.__ ._-_. __. M GREAT BARRIER REEF o 25 S ._-_._._,------_.------3 24 -----_.- ._------_..- E S 23

GREAT BARRIER REEF Figure 3.4 (continued) Acropora clathrata

A. clathrata X A. florida

oS 30 _.._.~.,- ~ 29 T I 28 ~~- ·=-11······ ····~Ilfnlrlnllfllrll rrr~ ·1/·1 ,11 C

0'1 C 27 PALAU ~ H R 26 ~=._=== ~~ o •iJ •Ii. J_ .•.... •••.••••••••• .•-•••••••. =•...••• Acropora florida ~ 25 - -_._------_.-. __._- --_.---.------_.•. _.._-._. __..._------~-- -_ ..-----_._.. - ..... ------_.. _.__ .._. ,"- -.----~---- S 30 S o o 24 ~ M - ----_._-~_ ..-..__ ._._~------_._------_ ..- '--'------_..-.._._- . __ ._._._--- 29 T ~ 23 I 28 C C 27 H PALAU R 26 o M o 26 SS,.24 . ------SE 23 1------~------.---. _._-_._------1

PALAU Figure 3.4 (continued) GENERAL MECHANISMS OF DEVELOPMENT OF POLYPLOIDY

Polyploidy is a chromosomal alteration in which an organism possesses more than two complete chromosome sets.

For example, in the angiosperm genus Festuca there are species in which the somatic 2n = 14, 28, 42, 56, and 70, where n represents the number of chromosomes in the gametes.

Such species are known as diploids, tetraploids, hexaploids, octoploids, and decaploids, respectively. These numbers are based upon 7, the gametic chromosome number of the diploid species. This number is known as the basic chromosome number

(x), which may be viewed as the most fundamental manifestation of the genome, the primary set of genetic information carried by the organism. In a diploid species n

= x, but in a polyploid species n is a multiple of x. In the case of the octoploid above, for example, 2n = 8x = 56.

Two extreme categories of polyploidy can be recognized, autopolyploidy and allopolyploidy. Autopolyploids derive extra chromosome sets from a single parent species, either through self- or cross- fertilization. Autotetraploids can arise through the union of two unreduced (2n) gametes

(Figure 3.5A) or through chromosome doubling in the zygote or other somatic cells (Figure 3.5B) . Autotetraploids often have reduced fertility, because not all quadrivalents will

separate by pairs during the first meiotic division, and a high proportion of gametes with an unbalanced complement of

chromosomes will consequently be produced. However, meiosis

70 Figure 3.5. Two mechanisms of autotetraploidy

71 Two Mechanisms of Autotetraploidy

Unreduced A Gametes

Chromosome »» Doubling & Non-disjunction III ~lll

Parent species Autotetraploid zygote or other somatic cell B

72 can be controlled by genes or gene combinations that suppress multivalent formation and elevate pairing, which enhances the production of balanced gametes (White, 1978; deWet, 1980; Lewis, 1980a). Autotetraploids producing balanced gametes can fertilize themselves, combine with the balanced gametes of other autotetraploids, or backcross with diploids. Autotriploids (3n) arise through the union of an unreduced, diploid (2n) gamete and a reduced (n) gamete

(Figure 3.6A), or by hybridization between diploid and autotetraploid (Figure 3.6B). Triploids are sterile as a result of their gametes having an unbalanced complement of chromosomes, two from some trivalents and one from other trivalents. Nonetheless, the sterile triploid may propagate itself by asexual reproduction. The diploid species that give rise to autopolyploids are usually ones that can self­ fertilize, as the probability of two different diploid parents producing diploid gametes contemporaneously in the same population and then mating is slight. Of naturally­ occurring plant polyploid species, fewer than 10% are autopolyploids (Dewey, 1980). Evidence from artificially produced examples indicates that autopolyploidy is rarely successful at higher levels than tetraploidy (Stebbins,

1950; Dewey, 1980).

Allopolyploids arise through hybridization between two chromosomally differentiated taxa. One of several mechanisms is involved, which mayor may not include the production of

73 Figure 3.6. Two mechanisms of autotriploidy.

74 Two Mechanisms of Autotriploidy

Unreduced A Gamete

Autotetraploid

Reduced Diploid Parent Gametes 8 Species

75 an intermediary hybrid (Stebbins, 1950; White, 1973, 1978;

Stace, 1980; Grant, 1981).

1. Normal, reduced gametes from both parent species may unite to produce a hybrid. The hybrid is usually sterile because its non-homologous chromosomes cannot pair during meiosis. Though infertile, the hybrid may actually be more vigorous than its parents and may be able to propagate itself asexually. Fertility can be restored at some point in the history of the hybrid by mitotic non-disjunction in the germ cell line. This doubles the number of chromosomes, which gives each chromosome a homologue to pair with and so allows the production of gametes through normal meiosis.

When two such gametes combine, either through self-or cross­ fertilization, the result is a viable, fertile organism with a chromosome number which is the combined somatic number of its parents (Stebbins, 1950) (Figure 3.7A). Polyploids may arise either from parents with the same chromosome number

(monobasic allopolyploids) or different chromosome numbers

(dibasic allopolyploids) .

2. A normal, reduced gamete from one parent species may unite with an unreduced gamete from the second parent species to produce a triploid hybrid. Unbalanced gametes render the hybrid sterile, but asexual reproduction may enable the hybrid to propagate itself. Unreduced, triploid

76 Figure 3.7. Two mechanisms of allotetraploidy which involve formation of an intermediate, adult hybrid.

77 Two Mechanisms of Allotetraploidy

Gametes

A Species B

Species A

Normal Gamete Normal B Species B Gamete

78 gametes may be produced by meiotic non-disjunction. These may backcross with normal, reduced gametes from the first parent species to form a viable, fertile tetraploid with homologous pairs of chromosomes. The allotetraploid has a somatic chromosome number equal to the sum of the somatic chromosome numbers of the two parent species (Figure 3.7B)

3. Normal, reduced gametes from both parent species combine, but mitotic non-disjunction in the zygote transforms the embryo and thus the adult into an "instant" allotetraploid with the combined somatic chromosome number of its parents.

There is no intermediary hybrid (Figure 3.8A).

4. Unreduced gametes from each of two diploid parents combine to form an allotetraploid in one step. Again there is no intermediary hybrid involved (Figure 3.8B).

Stebbins (1950) distinguished between genomic allopolyploids, produced by distantly-related parents, and segmental allopolyploids, produced by more closely-related parents. The chromosomes of parents of genomic allopolyploids are so different that they are unable to pair in the diploid hybrid, but form bivalents regularly in the tetraploid derivative. Genomic hybrids are therefore sterile, but their tetraploid derivatives are typically fully fertile. Chromosomes of parents of segmental

79 Figure 3.8. Two mechanisms of allotetraploidy which do not involve formation of an intermediate, adult hybrid.

80 Two Mechanisms of Allotetraploidy

Mitotic Non-disjunction

Zygote

Allotetraploid Zygote

A Species B

Species A Unred~ced > Gamete ~LL ..r>?) ~» LL« I III <

81 allopolyploids share a considerable number of homologous chromosomal segments or even whole chromosomes. The degree of sterility in the intermediate hybrid depends upon the extent to which the two parental genomes are differentiated.

The more closely related the parents, the more fertile the hybrids will be due to bivalent formation during meiosis, and the more the derived tetraploid will behave like an autopolyploid with four homologous chromosome sets. In general there is an inverse correlation between the fertility of a diploid hybrid and that of its tetraploid derivative: fertile hybrids produce sterile tetraploids, and sterile hybrids produce fertile tetraploids. However, chromosome pairing may also be genetically influenced, such that segmental allopolyploids and even autopolyploids may have high fertility (Dawson, 1962; deWet, 1980; Lewis, 1980;

Grant, 1981). In contrast to a segmental allopolyploid or autopolyploid, a genomic allopolyploid is strongly isolated from and usually morphologically discontinuous with its nearest relatives. Stebbins (1950) notes that, since genomic allopolyploids are more successful than segmental allopolyploids, the development of allopolyploidy in a genus

is greatly favored by the presence in it of many species which are sufficiently closely related to each other so as

to produce vigorous F, hybrids, but strongly enough

differentiated so that their chromosomes are nearly or

entirely incapable of pairing with each other.

82 Between the extremes of auto- and allopolyploidy there are many gradations in plants, because there are many gradations between inter-varietal and inter-specific differences (Darlington, 1973). Furthermore, two autotetraploids of different species might cross to produce an allotetraploid (Darlington, 1973). Limited interbreeding among the diploid and tetraploid forms, or between different tetraploids, may continue to build the ploidy level

(Stebbins, 1950; Dawson, 1962; Lewis, 1980; Stace, 1980;

Grant, 1981) (Figure 3.9). Although the majority of allopolyploids are either tetraploids or hexaploids, and so combine the genomes of only two or three different ancestral species, a number of higher polyploids of the genomic type are known (Stebbins, 1950). Factors which promote the formation of one new tetraploid species from its diploid ancestors often operate to promote the origin of additional new tetraploids from other diploids in the same species group (Grant, 1981). The tetraploids can then produce hexaploids and octoploids. The result of continued polyploidization in different anastomosing lineages is a polyploid complex (Figure 3.9). Such polyploid complexes, which may involve autopolyploids, allopolyploids, and autoallopolyploids, have been described for many plant genera (Stebbins, 1950; White, 1978; Grant, 1981). Extensive polyploid complexes are produced only after some of the speciation processes in a genus on the diploid level have

83 Figure 3.9. Chart illustrating the possible origin of a polyploid complex commencing from five diploid species AA, BB, CC, D1D1 , and D?D2 • Genomes A, B, C and Dare strongly differentiated structurally so that thelr chromosomes usually fail to pair in the interspecific hybrids. Genomes D1 and D2 are differentiated from each other structurally to an intermediate degree, so that partial bivalent formation takes co If:> place in both the diploid hybrid and the polyploid derivative. The boxed taxa have each genome represented twice and are expected to have full fertility. The other taxa are expected to have varying degrees of sterility. The chromosome numbers on the left are based on a complex with a basic number x=7. Modified from Stebbins (1950) and Stace (1980). OCTOPLOIDS AAAABBBB IAABBCCD1D11 2N.aX.56 autoallo-

AAAAAAAA.\ gen:~i:m,iC allo- HEPTAPLOIDS 2N"7X·49 allo- autoallo- HEXAPLOIDS IAABBCcl CCD1D1D1D1 2N·6X-42 co U1 PENTAPLOIDS 2N·5X·35 genomic genomic segmental auto- allo- allo- allo- TETRAPLOIDS I~ AAAA IAABB ABCD1.'CCD1D1I °1°1°2°2 2N·4X·28 I \ -, , TRIPLOIDS 2N·3X·21 is A, A, ICD,

DIPLOIDS ~AB+rnID ~CD11D1D1~D1D21D2D21 2N·2X·14 Species A Species B Species C Species 0 1 Species O2 reached a certain stage of maturity, so that at least some of the species have diverged widely from the others with respect to their chromosomes (Stebbins, 1950). Most of the polyploid species complexes which are difficult problems for the systematist contain segmental allopolyploids, autopolyploids, or both; genomic allopolyploids most often stand out as clearly marked species (Stebbins, 1950)

Newly arisen autotetraploids and segmental allopolyploids tend to have reduced fertility due to multivalent association during meiosis. Over evolutionary time, however, the duplicate genes that a tetraploid has derived from its diploid ancestors can experience different mutations, so that initially identical loci can diverge and the tetraploid becomes diploid not only in chromosomal behavior but in gene structure (Dawson, 1962; Stebbins,

1950; Lewis, 1980; Levin, 1983). This divergence of duplicate loci, termed diploidization, has been documented in several plants (Dawson, 1962; deWet, 1980, Grant, 1981).

Natural or artificial selection for fertility will favor genes that promote bivalent formation at the expense of multivalent formation. This concept has been supported by experimental work in which known autoploids of the grass genus Zea were selected for fertility, and the frequency of multivalent formation significantly decreased over ten generations (deWet, 1980; Grant, 1981).

86 CONDITIONS CONDUCIVE TO POLYPLOIDY

Based on the mechanisms described above, three primary conditions are conducive to the development of polyploidy:

(1) for allopolyploids, the ability to reproduce asexually, so that an intermediate hybrid might propagate, and the rare events of meiotic and mitotic non-disjunction might have increased chances of occurring, (2) for allopolyploids, the ability to form viable interspecific hybrids, and (3) for both auto- and allopolyploids, the production of unreduced gametes. The subject of hybridization in Acropora has been treated in Chapter 2. Evidence concerning the other two factors as they may operate in the development of polyploidy in both plants and corals is the subject of the following section.

Asexual Reproduction

Polyploidy is the most widespread cytogenetic process affecting higher plant evolution (Averett, 1980). Its occurrence is well docmented within mosses (Crosby, 1980), ferns (Wagner and Wagner, 1980), and angiosperms (Goldblatt,

1980; Lewis, 1980b), but rare among gymnosperms (Delevoryas,

1980). Estimates of polyploid frequency among angiosperms vary from 30% to 80% (White, 1942; Stebbins, 1950; Grant,

1963; Goldblatt, 1980; Lewis, 1980b). The highest frequency occurs among herbaceous perennials, with lower frequencies in annuals and woody plants (Stebbins, 1950, 1970; White,

87 1973, 1978; deWet, 1980). Among perennial angiosperms, polyploidy is especially prevalent in those species with effective means of vegetative reproduction such as stolons, rhizomes, bulbs, and suckers (subordinate shoots from a bud on the root or stem) (Dawson, 1962; Stebbins, 1950, 1970;

White, 1973, 1978; Grant, 1981). Extensive hybridization in a group of vegetatively reproducing perennial angiosperms is expected to yield polyploid complexes (Grant, 1981). By contrast, although natural hybridization occurs quite freely in oaks (Quercus) and Eucalyptus, vegetative reproduction does not occur and there are no polyploid species in these genera (White, 1978).

The association between polyploidy and vegetative reproduction holds true within other taxa as well. Sequoia, one of the very few conifers in which polyploids occur naturally, regenerates asexually by suckers (Grant, 1981).

Unlike liverwort bryophytes, which show neither appreciable vegetative reproduction nor polyploidy (Crosby, 1980), vegetative regeneration is possible from almost any portion of the moss plant (Averett, 1980). Polyploidy in ferns is associated with apomixis (loss of sexual reproduction)

(Wagner and Wagner, 1980). The green algae (Chlorophyta) are the most widely studied algae with respect to ploidy levels.

Filamentous green algal genera such as Stigeoclonium and

Chaetomorpha contain polyploid series similar to those in mosses and vascular plants (Godward, 1966; Stebbins and

88 Hill, 1980), as do several species of Cladophora (Godward,

1966j Nichols, 1980). Among the brown algae (Phaeophyta), the Dictyotales and Fucales show evidence of polyploidy

(Godward, 1966). The Dictyotales are exceptional among advanced orders of brown algae in their capacity for significant asexual reproduction (Clayton, 1988), while the sporophyte thalli in Fucales are perennial and long-lived

(Clayton, 1988).

Perennials with efficient means of vegetative reproduction are better equipped than annuals to pass through periods of sexual sterility before chromosome doubling or the union of unreduced gametes produce fertile polyploids. Such plants are also better able to compete by occupying available space or by spreading into new habitats

(deWet, 1980). The temporal and spatial opportunities for the chance events needed to restore fertility are thereby enhanced by the perennial, vegetatively-reproducing habit.

Vegetative reproduction is one mode by which plants and protists may clone, that is, produce an assemblage of individuals that are genetically identical. Vegetative reproduction results in a number of genotypically identical replicates, or ramets, which collectively comprise the clone, or genet. Certain genetic and ecological properties are enhanced by the reproductive and morphological aspects of cloning. By avoiding recombination and segregation, cloning enables successful genomes to be inherited intact.

89 Clonal individuals can reproduce in complete isolation, so reproduction is still possible when encounters with other individuals are rare. Even one individual entering a new habitat can start a new population. By spreading the risk of localized mortality among scattered individuals, cloning increases the survivorship of the genome. Cloning thereby gives a genome great potential longevity, effectively delaying or even circumventing senescence. Because of potential immortality and continued clonal growth, genets can differ enormously in size, with disparate contributions to the sexual gene pool (Cook, 1983, 1985; Hughes, 1985;

Silander, 1985; Hughes, 1989).

Parallels between clonal plants and benthic modular animals are numerous (Jackson, 1977, 1985; Buss 1979;

Harper, 1981; Coates and Jackson, 1985; Hughes and Cancino,

1985; Hughes, 1989; Hughes et al., 1992). Both are typically sedentary, and must acquire resources from a fixed point.

Both must adjust morphologically or behaviorally to their immediate surroundings, and use defenses which are not predicated on escape. Spatially-scattered ramets are likely to encounter different microenvironments, and therefore to experience different selective regimes. Because genets often

survive despite considerable mortality among their ramets,

severe reduction in population size does not cause genetic depletion in clonal animals to the same extent as among

unitary, aclonal animals. Modular iteration within ramets,

90 such as found in corals, bryozoans, and colonial ascidians, allows great plasticity of growth form. Modules which escape localized damage to the ramet can repair tissue loss through continued asexual reproduction. Through the amplification of the genet, both as iterated modules within the ramet and spatially-dispersed ramets comprising the genet, longevity is prolonged and senescence is effectively delayed. Modular animals often resemble plants in form, having broadly analogous reproductive, defensive, and competitive mechanisms (Hughes, 1989).

Colonial corals are modular benthic animals which exemplify the salient features of a clonal life history.

Individual ramets (colonies) increase in size through modular (polyp) iteration, and additional ramets are established when a colony becomes subdivided. Ramet separation and dispersal occur through fragmentation or fission, and may be effected by agents including storms

(Stoddart, 1963; Maragos et ai., 1973; Randall and Eldredge

1977; Highsmith et ai., 1980; Knowlton et ai., 1981; Dollar,

1982), bioerosion (Highsmith, 1981; Jackson and Hughes,

1985; Hughes, 1988), and vertebrates (Hunter, 1988).

Additional processes such as 'polyp balls' (Rosen and

Taylor, 1969), 'polyp bail-out' (Goreau and Goreau, 1959;

Sammarco, 1982; Richmond, 1985) and asexual production of

larvae (Stoddart, 1983; Ayre and Resing, 1986) propagate

parental genotypes intact. Thus, coral populations often

91 have few genotypes represented by many independent colonies

(Niegel and Avise, 1983; Stoddart, 1983, 1984a,b, 1985;

Hunter, 1988; Hughes et al., 1992), which may be dispersed among diverse habitats. Recently or intensely disturbed habitats appear to have the greatest clonal diversity

(Hunter, 1988, 1993). Because of genet longevity and annual sexual reproduction in many corals (Richmond and Hunter,

1990), generations strongly overlap and can interbreed

(Potts and Garthwaite, 1991; Hughes et al., 1992).

As dominant members of the shallow reef assemblage, members of the genus Acropora are particularly susceptible to fragmentation during storms (Tunnicliffe, 1981;

Highsmith, 1980, 1982). The degree of success in establishing new colonies from fragments varies according to the size of fragments (Highsmith et al., 1980; Knowlton,

1981; Kobayashi, 1984) and the species (Randall, 1973;

Birkeland et al., 1979; Bothwell, 1981; Wallace, 1985). In some species of Acropora, larval recruitment may contribute little to the establishment of new colonies, and asexual propagation may be the primary mode of propagation (Bak and

Engel, 1979; Tunnicliffe, 1981; Rylaarsdam 1983; Hughes,

1985; Wallace, 1985).

Thus, the clonal life history traits which are closely associated with the development of polyploidy in protists, mosses, ferns, and higher plants also characterize modular corals in general and members of the genus Acropora in

92 particular. The spatial replication and temporal longevity of clonemate genomes increases the opportunities for the rare cytogenetic events that are necessary for polyploidy to arise. Development of a polyploid complex is facilitated by the coexistence of several overlapping generations, which enhances the possibility of contact between gametes of different ploidy levels produced in different generations.

Additionally, the tendency towards philopatric settlement

(Stoddart, 1983; Hughes, 1985; Jackson, 1986) or limited dispersal of larvae (Oliver and Willis, 1987; Sammarco and

Andrews, 1988) suggests that, at least over short ecological time spans, generations with different ploidy levels may co­ exist within a proximity which allows limited backcrossing between generations, a requisite condition for the development of higher ploidy levels (Figure 3.9).

Production of Unreduced Gametes

Unreduced gametes can potentially function at all levels of polyploid complex development, including the parental, hybrid, and polyploid generations. Union of unreduced gametes is one of the most important mechanisms whereby polyploid individuals arise in sexually reproducing populations (Lewis, 1980a; Grant, 1981). In most animals and plants, unreduced yet functional gametes are formed at a low frequency by a failure of meiosis (Dawson, 1962; deWet,

1980; Lewis, 1980a). Genotype as well as phenotype appear to

93 determine the frequency with which unreduced gametes are produced (deWet, 1980). In Zea mays strains, a particular gene greatly increases the number of unreduced gametes produced (Lewis, 1980aj deWet, 1980). Adverse growing conditions have also been observed to increase the frequency of unreduced gametes in populations of Gilia, and this is probably true for many plant species (Lewis, 1980aj deWet,

1980). The frequency of unreduced diploid eggs in amphibians can be experimentally increased by pressure or temperature shock (Bogart, 1980).

Modular animals, including corals, lack a germ line in which cells responsible for gametogenesis are segregated at the outset of embryological development (Hughes, 1989)

Instead, in anthozoans, somatic gastrodermis remains competent for differentiation into primary gonial cells after migration into the mesoglea (Campbell, 1974). Somatic mutations in mobile undifferentiated cells can eventually become part of the germ line (Hughes, 1989j Hughes et al.,

1992). Endopolyploidy, in which somatic cells become polyploid due to dysfunction of the mitotic spindle, is normal in most animal groups (White, 1973j Brodsky and

Uryvaeva, 1985). Migration of polyploid somatic cells into the germ line, followed by normal meiosis, would result in the production of balanced gametes with an unreduced number of chromosomes. The absence of a specific germ cell line sequestered early in development therefore effectively

94 augments the proportion of cells in which endopolyploidy may occur and result in the production of balanced gametes.

The probability of unreduced gametes being produced in clonal corals is therefore heightened by the modular nature of the clone whereby each polyp can produce many gametes, the spatial repetition of ramets, the temporal longevity of genets, and the chances of somatic mutations becoming part of the reproductive cells. Even if only a small proportion of the gametes are unreduced, the high fecundity of individual colonies ensures that a large number of such gametes are regularly produced. Moreover, if some clones were genetically programmed to produce a high proportion of unreduced gametes, as is the case in Zea mays, the longevity of genets would ensure their recurrent contribution to the gamete pool.

GENERAL MECHANISMS OF ANEUPLOIDY

In any critical study of the possible occurrence of polyploidy it is important to bear in mind that an altered number of chromosomes can arise by fusion or subdivision of

existing chromosomes, rather than by replication (Dawson,

1962). Whereas polyploidy represents numerical differences

with respect to whole sets of chromosomes, aneuploidy refers

to differences in number of individual chromosomes. Several

categories of aneuploidy and the mechanisms by which they

are generated are relevant to the present discussion,

95 including dysploidy (descending and ascending), nullisomy, and tetrasomy. A brief description of these mechanisms and their importance in plant cytogenetics follows, to aid in understanding the models which are proposed to account for the variation in chromosome number determined by this study.

In higher plants, changes in haploid number, which may involve an increase or decrease by one chromosome at a time, can occur on the diploid level or on higher ploidy levels.

Increases or decreases in the gametic chromosome number which occur on the diploid level are known as dysploidy.

Descending dysploidy reduces the basic chromosome number x.

This process is well documented in the plant cytological literature (Stebbins, 1950; White, 1978; Grant, 1981).

Reduction of the haploid number by one chromosome can occur with relatively little loss of chromosomal material by reciprocal translocation of unequal chromosome segments

(Stebbins, 1950; Stace, 1980; Grant, 1981). Portions of chromosomes on either side of the centromere are often inert. The active distal portion of an acrocentric chromosome "A" may be reciprocally translocated to another, non-homologous chromosome "B" in the ancestral complement.

All of the genetically active material on "A" becomes translocated onto "B", while the centromere of "A" receives in exchange a small inert segment. The now inert chromosome

"A" will then be lost during meiotic non-disjunction, resulting in reduction of the gametic number. Sequential

96 aneuploid reduction at the diploid level results in several basic numbers (e.g., x=8, 7, and 6) being represented within a related taxonomic group.

Aneuploid increase in gametic chromosome number can occur by misdivision of the centromere. For example, some plant genera have chromosomes with more than a single spindle fiber attachment at the centromeres. The centromere may divide transversely, producing two halves each of which retains the ability to move normally to the spindle poles at meiotic anaphase and be retained in the genotype. One metacentric chromosome thereby gives rise to two telocentric chromosomes (Stebbins, 1970; Grant, 1981). Ascending dysploidy may have played a large role in the evolution of the extensive series of basic chromosome numbers found in some plant genera such as Carex (Cyperaceae) (x= 5, 6, 7, 8, and 9) (Stebbins, 1950; Grant, 1981).

Descending dysploidy is known in a number of plant groups. Stebbins (1950) lists 25 such groups and Grant

(1981) updates the list with an additional five. Ascending dysploidy is known in a smaller number of plant groups. Six such groups are listed by Stebbins (1950) and an additional three by Grant (1981). A given taxon may demonstrate both descending and ascending dysploidy. For example, the original basic number in the angiosperm genus Clarkia

(Onagraceae) is x=7 (Lewis, 1953), but descendent basic numbers of x=6 and x=5, as well as ascendent basic numbers

97 of x=8 and x=9, are also found. Some aneuploid series are considered to be ascending by some scholars and descending by others. Because documented changes in basic number proceed more frequently in the direction of decrease than of increase in plants, when experimental evidence is lacking the most primitive species is generally assumed to have the highest basic number in an aneuploid series, and the trend has been toward reduction (Stebbins, 1950; Grant, 1981).

Another type of aneuploidy which simulates a progressive increase in chromosome number can be derived by a combination of successive aneuploid increase or decrease in basic number followed by genomic allopolyploidy

(Stebbins, 1950; Grant, 1981). For example, if a genus with an initial basic number of x=7 produces by successive aneuploid reduction derivatives with x=6 and x=5, then various monobasic and dibasic polyploid combinations from these species can produce every haploid number from n=10 upward (Stebbins, 1950). Stebbins (1950) lists fourteen angiosperm genera in which polyploidy has been superimposed upon several basic numbers generated by dysploidy. The most extensive aneuploid series in the plant kingdom is that in the genus Carex (Cyperaceae), in which the basic numbers x=6, 7, 8, and 9 may have derived from an original basic number x=5 (Stebbins, 1950; Grant, 1981). There are polyploid series in Carex derived from each of these basic numbers (monobasic), as well as polyploids derived from

98 hybridization between lineages with different basic numbers

(dibasic). Thus, different hybrid combinations of 7-paired and 8-paired diploids will produce a series of tetraploids with n=14, 15, and 16. Polyploid drops and gains, in which aneuploidy reduces or increases chromosome number at the polyploid level (Darlington, 1973), may extend the series in both directions. Stebbins (1950) terms such a system

"aneuploid amphidiploidy", and attributes it to many of the higher chromosome numbers in angiosperms.

The loss of one chromosome pair is called "nullisomy".

The loss of one or more chromosome pairs in a polyploid can often be tolerated because of the presence of duplicate genetic material in the homologous or homeologous chromosomes (Stace, 1980; Grant, 1981). Darlington (1973) cites examples of polyploid drop in twenty plant groups.

The final mechanism of aneuploid increase of interest

in the present study involves the incorporation of an extra pair of chromosomes of the regular complement as a result of

lagging during cell division. This results in a tetrasomic

type (2n + 2). The extra tetrasomic chromosomes can then

diverge from their homologues through interchanges with non­

homologous chromosomes and by divergences in gene function

(Grant, 1981). Extra chromosome pairs arise from meiotic

irregularities in many polyploids (Grant, 1981). Aneuploid

increase in the angiosperm genus Clarkia has taken place in

99 some cases by the incorporation of tetrasomic chromosomes which have undergone translocations (Grant, 1981).

SPECIFIC MECHANISMS OF SOMATIC CHROMOSOME NUMBER VARIATION

Five alternative models which invoke combinations of polyploidy and aneuploidy are proposed to account for the variation in Acropora chromosome number demonstrated in this study. Wallace et al. (1991) consider Acropora to be monophyletic because the axial corallite is a feature shared by all of its species, but by no other scleractinian genera.

The following models are based on the assumption of the monophyletic origin of Acropora. With regard to the general mechanisms involved, both autopolyploidy and allopolyploidy may have been operative in chromosome doubling. However, allopolyploidy is a more likely mode of genome duplication in Acropora than autopolyploidy, due to (1) the demonstrated capacity for hybridization (Chapter 2); (2) very low levels of experimentally-demonstrated self-fertilization (Heyward and Babcock, 1986; Richmond and Hunter, 1990; Chapter 2); and (3) lack of evidence for parthenogenesis in Acropora, a demonstrated method of autopolyploidy in animals. The cytogenetic mechanisms may involve production of an intermediate hybrid (Figure 3.7) or instant polyploidization through the combination of two unreduced gametes or zygotic chromosome doubling (Figure 3.8).

100 The inference of a basic number in a polyploid series is an important step in fitting chromosome numbers to an evolutionary hypothesis. Sometimes the diploid members of a series have disappeared from a genus or even a family, erasing evidence of the lower numbers (Stebbins, 1950;

Darlington, 1973; White, 1978; Grant, 1981). Because the gametic numbers n=7, 8, and 9 occur together in many angiosperm families, several authors have suggested that the original basic number of the angiosperms lies in the range x=7-9 (Ehrendorfer, 1964; White, 1978; Grant, 1981). Many genera and some families of woody angiosperms have high apparent basic numbers in the range x=12 to x=21. Some authors regard these not as truly basic numbers in the phylogenetic sense but as old polyploids from an earlier cycle of evolution (Stebbins, 1950, 1971; Tischler, 1954;

Darlington, 1956; Grant, 1981). Goldblatt (1980) and Lewis

(1980b) based their analysis of the frequency of polyploids in monocots and dicots, respectively, on the premise that species with n ~ 11 have polyploidy in their ancestry. As an alternative, Grant (1981) suggested that high basic numbers in the range x=12-16 are products of ascending dysploidy from an original angiosperm basic number in the range x=7-9.

In this study, corals from four families had a somatic chromosome number of 28. This fact suggests that the basic scleractinian chromosome number x=14, an assumption on which one of the following models is based (Model 5). However,

101 somatic chromosome number has been established for less than

5% of Indo-West Pacific scleractinian corals. Given the lack of information concerning chromosome number in so many species of corals, an alternative possibility is that the basic scleractinian chromosome number x=7, and that 28 somatic chromosomes represents a tetraploid level of cytogenetic development (2n = 4x = 28). This assumption underlies four of the following five models (Models 1-4). It is itself an hypothesis that can be tested by further chromosome counts in additional genera and families. Common to Models 1-4 is the proposal that the ancestral basic number x=7 characterized a pool of ancestral diploids, 2n

2x = 14. Through tetraploidy a stable ancestral Acropora species 2n = 4x = 28 developed, from which many other

Acropora species have radiated through classic allopatric speciation since the Eocene (Potts, 1983, 1984a, 1985;

Rosen, 1984; McManus, 1985).

Modell

A second basic number x=6 derived from an original basic number x=7 through descending dysploidy (Figure 3.10).

The somatic chromosome numbers 24, 30, 42, 48, 54 are all related as part of a polyploid series through this derived basic number x=6, i.e., 4x, 5x, 7x, 8x, 9x. Although diagrammed as a linear sequence in Figure 3.10 for the sake of simplicity, the scheme involves a number of sequential

102 Figure 3.10. The general relationships of species according to Modell. The original basic number is x=7, from which x=6 is derived by descending dysploidy. Though the species with somatic chromosome numbers 24, 30, 42, 48, and 54 are depicted in a linear relationship for the sake of simplicity, their relationships are generated by hybridizations, polyploidy, and backcrossing as diagrammed in Figure 3.9.

I-' o w DECAPLOID 2n·10x·60 • ? ~ NANOPLOID 2n·9x·54 I Aelseyi t OCTOPLOID .... • 2n·8x·48 I Aocellata J HEPTAPLOID II 2n·7x·42 I A valida t ? f-.J HEXAPLOID 36 0 I ~ It 2n"r Agemmifera PENTAPLOID II 2n.Sx.'30 I t Adivaricata TETRAPLOID 2n-4x·28 •I 2n.4x.24 Adana;

TRIPLOID 11\ If\ 2n 1l2x·12 DIPLOIDS 2n-2x·14 ~X -6 ANCESTRAL POOL X=7---- Original basic number episodes of backcrossing and polyploidy (Figure 3.9). For example, a somatic chromosome number of 24 would represent tetraploidy, from which an octoploid with 48 chromosomes could develop, involving either one or two parental species.

Limited backcrossing between octoploid and tetraploid would produce a hexaploid with 36 chromosomes, which in turn might interbreed with previous generations, eventually producing the diagrammed sequence of numbers.

Though the even multiples in this series are expected to be fully fertile, the odd-numbered multiples are expected to be sexually sterile because of meiotic irregularities

(White, 1978; Grant, 1981). As all of the species in this series (A. danai, A. gemmifera, A. divaricata, A. valida, A. ocellata, A. elseyi) were observed to be fertile and have high fecundity, a polyploid series with odd and even multiples based upon x=6 is not a viable explanation for this sequence, and consequently will not be considered further.

Model 2

Relative to an original basic number x=7, the chromosome numbers 28 and 42 represent tetraploidy and hexaploidy, respectively. The chromosome number 54 is nullisomic, generated through loss of a pair of chromosomes from an octoploid (8x - 2). Through the mechanism of descending dysploidy a second basic number x=6 was derived. A pool of

105 ancestral diploids 2n = 2x = 12 based on this derivative basic number was then generated. Relative to x=6 the somatic chromosome numbers 24 and 48 are tetraploids and octoploids, respectively. The two species each with a somatic chromosome number of 30 would be independent cases of tetrasomic aneuploidy based upon the tetraploid somatic number 28

(Figure 3.11).

By this model somatic chromosome numbers of 24, 3D, 42,

48, and 54 therefore represent tetraploids, hexaploids, or octoploids, in some cases accompanied by nullisomy or tetrasomy. All species are theoretically expected to be fully fertile, which is consistent with observations. Within the outline of this model a number of alternative scenarios are possible, each of which is discussed with reference to the particular species involved.

Acropora valida

With 42 somatic chromosomes, A. valida represents a hexaploid level of cytogenetic development with respect to the assumed original basic number x=7. As such it could combine two or three ancestral genomes, depending on its mode of origin (Figure 3.9). A corollary example from angiosperms is the cultivated wheat Triticum aestivum, a genomic hexaploid (2n=42) which combines three ancestral genomes (Dawson, 1962; White, 1973, 1978; Grant, 1981).

106 Figure 3.11. Chart depicting the general relationships of species according to Model 2. The original basic number is x=7, from which x=6 is derived by descending dysploidy. An octoploid with 56 somatic chromosomes is reduced by nullisomy to 54 somatic chromosomes. Thirty somatic chromosomes originate through tetrasomy operating on a tetraploid with 28 somatic chromosomes.For the sake of simplicity, pathways invoking autopolyploidy are not included. See text for additional discussion of ~ alternative pathways and ancestral relationships. o ~ OCTOPLOID 2n-8x-56 2n-8x-48 2n-2-S4.. A. oce/lata A.e/seyi

HEXAPLOID 2n-6x-42 A. va/ida r-' o co

TETRAPLOID 2n+2-30 ~~~ _ A.danai 2n-4x-28 A.gemmifera 2n-4x-24 A.divaricafa TRIPLOID (sterile) "2n-3x-21 DIPLOIDS, /' 2n-2x-14 2n-2x-12 Ancestral Pool X=7 • X=6 Original basic number Derived basic number Acropora elseyi

With 54 somatic chromosomes, A. elseyi represents an octoploid level of cytogenetic development with respect to the assumed original basic number x=7, followed by nullisomy. The original octoploid complement may have arisen in one of several ways (Figure 3.9) :

1. as an autooctoploid, followed by gene divergence and diploidization, thereby incorporating only one ancestral genome;

2. as an autoallooctoploid based on an extant or an extinct genomic tetraploid with 28 somatic chromosomes, followed by gene divergence and diploidization of the genome. By this route of formation two ancestral genomes would be incorporated.

3. as a monobasic genomic allooctoploid of two parents, each with 28 somatic chromosomes. By this mode four ancestral genomes would be incorporated.

4. as a dibasic genomic allooctoploid of two parents with

42 and 14 somatic chromosomes, respectively. An example of this alternative is the artificially produced intergeneric angiosperm Triticale. Its parents Triticum aestivum (2n

42) and Secale cereale (2n = 14) form a sterile hybrid, but when the hybrid somatic chromosome number 28 is doubled the resulting polyploid Triticale, with 56 somatic chromosomes, is fertile (Dawson, 1962; Grant, 1981).

109 Acropora danai

With 24 somatic chromosomes, A. danai represents a tetraploid of a derived basic number x=6. A. danai could be an autotetraploid generated from an ancestral diploid (2n =

2x = 12), or a monobasic allotetraploid produced by hybridization between two such ancestral parents.

Acropora ocellata

With 48 somatic chromosomes, A. ocellata represents an octoploid derivative of the basic number x=6. Similar to the options presented for A. elseyi, A. ocellata could have been produced as:

1. an autooctoploid of a single parental genome, followed by gene divergence and diploidization of the genome. Only one ancestral genome would be involved through this route.

2. an autoallooctoploid based on an extant or an extinct genomic tetraploid with 24 somatic chromosomes, followed by gene divergence and diploidization of the genome. By this route of formation two ancestral genomes would be incorporated.

3. a monobasic genomic allopolyploid in which both parents had 24 somatic chromosomes, thereby incorporating four ancestral genomes. The angiosperm Nicotiana tabacum

(2n=48) is believed to be an amphidiploid between two species, each of which has 24 somatic chromosomes in its somatic cells (Stebbins, 1950j Dawson, 1962).

110 The extreme morphological dissimilarity of A. ocellata to A. danai render it unlikely that the latter is directly involved in the parentage of A. ocellata.

Acropora gemmifera

An extra pair of chromosomes of the regular tetraploid complement of 28 would result in a tetrasomic type (2n + 2) with 30 somatic chromosomes. Genetic factors promoting formation of bivalents rather than multivalents during meiosis, as well as divergence of duplicate gene loci over time, would give rise to a stable type.

Acropora divaricata

With 30 somatic chromosomes, mechanisms similar to those suggested for A. gemmifera could be operative. Given their marked morphological dissimilarity, different ancestral species are implicated.

Model 3

This model shares many features in common with Model 2.

However, a third basic number x=8 is derived from x=7 through ascending dysploidy, from which a pool of ancestral diploids 2n = 2x = 16 is generated. Hybridization at the diploid level between ancestral types with 14 and 16 somatic chromosomes generates sterile hybrids with 15 somatic chromosomes. Chromosome doubling then converts the hybrid to

111 a dibasic allotetraploid with 30 somatic chromosomes (Figure

3.12). As suggested in Model 2, given the marked morphological dissimilarity of A. gemmifera and A. divaricata, different parental species are likely involved in their ancestries.

Parallel aspects of Models 2 and 3 can be found in the angiosperm genera Vicia and Claytonia. In the genus Vicia, somatic chromosome numbers of 10, 12, 14, 24 and 28 consist of diploids and tetraploids derived from two basic numbers

(x=6 and 7), accompanied by aneuploidy. The tetraploids with

2n=24 are most likely derived from diploid ancestors with

2n=12, and those with 2n=28 from ancestors with 2n=14

(Stace, 1980). The polyploid complex in the North American genus Claytonia (Portulacaceae) is structured around three basic numbers (x=6,7, and 8), each of which has given rise to separate polyploid lineages, interlineage polyploids, and aneuploids clustering around the ploidy levels (White, 1973j

Grant, 1981).

Models 2 and 3 require the fairly recent coexistence of diploid and tetraploid forms. Many of the sixteen extant families of scleractinian corals first appear in the fossil record during the Cretaceous, with the family Acroporidae appearing about 100 MYA (Wells, 1956j Veron, 1986). The genus Acropora emerged during the Eocene, about 60 MYA

(Wells, 1956), but the ages of living species are not known

112 Figure 3.12. The general relationships of species according to Model 3. The original basic number is x=7, from which x=6 is derived by descending dysploidy and x=8 is derived by ascending dysploidy. An octoploid with 56 somatic chromosomes is reduced by nullisomy to 54 chromosome. Allotetraploids with 30 chromosomes are formed via hybridization between diploids with two different basic numbers. For the sake of simplicity, pathways invoking autopolyploidy are not included. See text for f-' additional discussion of alternative pathways and ancestral relationships. f-' W OCTOPLOID 2n-8x-56 2n-8x-48 2n-2-54.. A.ocellata A.elseyi

HEXAPLOID 2n-6x-42 A.valida

I-' I-' t/>.

. A.gemmifera TETRAPLOID A.divaricata A.danai 2n-30 2n-4x-28 2n-4x-24

sterile hybrid or triploid 2n-15t "'-2n-3x-21 2n-2~ DIPLOIDS, "<, / 2n-2x-12 Ancestral X-a .. 2n-2x-14 • X-6 Pool DerIved .... X-7. DerIved baa'e numb.r OrIgInal basIc number baa'e numb.r (Potts, 1985). Despite the antiquity of the genus, it appears not to have acquired its modern ecological prominence until several million years ago during the Plio­

Pleistocene (Rosen, 1993). A general time frame of events to accommodate Models 2 and 3 is diagrammed in Figure 3.13.

Fourteen somatic chromosomes would still typify corals when modern families began to emerge in the Cretaceous.

Independent episodes of tetraploidy might subsequently occur within separate family lines, resulting in the somatic number of 28 found in species from three additional families. The diploid taxa with 14 somatic chromosomes from which tetraploids emerged may have become extinct, or they may still exist but have not yet been found.

Model 4

The fourth model preserves several components of the preceeding two models. Fifty-four chromosomes arises through nullisomy acting on an octoploid, as in Models 2 and 3, and

30 chromosomes represents tetrasomy acting on a tetraploid, as in Model 2. In this model, however, the progressive loss of chromosomes in the gametes operates not at the diploid level, but at the tetraploid level (Figure 3.14). The gametic chromosome number is progressively reduced in the tetraploids from 14 to 13 to 12 in a stepwise fashion over time, as genetic material becomes translocated to other chromosomes. Corals based on the gametic chromosome number

115 Figure 3.13. Family tree of the sixteen extant families of scleractinian corals (modified from Wells, 1956), showing a time frame of changes in somatic chromosome number to accommodate Models 2 and 3.

I-' I-' CT\ FAMILY TREE OF SCLERACTINIA Jurassic Cretaceous ,Tertiary ouaternarv ,I ,;I' i~ Pocilloporidae ~~ Astroco~niidae ~~ ~ : Acroporldae 14) .! @~!. ~ ~ Poritidae f28\ j " V : j Siderastreidae I-' , f-l ~ : Agariciidae : ~~ -~::-:~ Fungiidae@

.--; ~~. - , ':

_ , .~, Oculiniidae ~ ,.., ~I i : ;Pectiniidae : : UMUSSidae.@ , , , , , Meandrinldae

j Trachyphyllidae -. sa: , i Favlldae Merulinidae Caryophyllidae Dendrophyllidae MYA 195 140 65 5 Figure 3.14. Chart depicting the general relationships of species according to Model 4. The original basic number is x=7. An octoploid with 56 somatic chromosomes is reduced by nullisomy to 54 somatic chromosomes. A form with 42 chromosomes originates through backcrossing between octoploid and tetraploid. Thirty somatic chromosomes originate through tetrasomy operating on a tetraploid with 28 somatic chromosomes. Sequential aneuploid reduction in tetraploid gametes produces gametes with 12 ...... chromosomes. The intermediary gametic chromosome number of 13 is shown in parentheses co because no corals have been found which are based on this number. For the sake of simplicity, pathways invoking autopolyploidy are not included. See text for additional discussion of alternative pathways and ancestral relationships. OCTOPLOID 2n-8x-56 2n-8x-48 2n-2-54.. Aoce/lata A elseyt

HEXAPLOID 2n-6x-42 Ava/ida I--' I--' \D

TETRAPLOID 2n+2-30 ~.~ ..... A.dana; 2n-4x-28 ..... (n-13) -. 2n-4x-24 Agemmifera A.divaricata TRIPLOID (sterile)

DIPLOIDS, 2n1J2x-14 Ancestral Pool X=7 Original basic number of 13 have not yet been found. Corals with 24 somatic chromosomes are generated by the combination of gametes with

12 chromosomes. These in turn undergo polyploidy and create an octoploid with 48 chromosomes. A form with 42 somatic chromosomes represents a hexaploid level of development, generated by backcrossing between octoploid and tetraploid.

Because the basic chromosome number x is defined as the gametic number of the diploid species, x remains the same for all taxa in this model. Aneuploid variation in the gametes of taxa at higher ploidy levels is known in the

Claytonia virginica (Portulacaceael group in angiosperms

(Rothwell, 1959).

Model 5

This model differs from the preceeding four models in that the basic chromosome number x is assumed to be 14 rather than 7. Twenty-eight somatic chromosomes therefore represents a diploid rather than a tetraploid level of development (Figure 3.15). Pathways identical to those proposed in Model 4 function in Model 5, but operate at different ploidy levels. Because reduction in gametic chromosome number occurs at the diploid level, the basic chromosome number x is sequentially reduced from 14 to 13 to

12. A form with 42 somatic chromosomes represents a triploid

level of development, generated by backcrossing between

tetraploid and diploid.

120 Figure 3.15. Chart depicting the general relationships of species according to Model 5. The original basic number is x=14. A tetraploid with 56 somatic chromosomes is reduced by nullisomy to 54 somatic chromosomes. A form with 42 somatic chromosomes originate through backcrossing between tetraploid and diploid forms, but the triploid is expected to have low fertility. Thirty somatic chromosomes originate through tetrasomy operating on a diploid with 28 somatic chromosomes. A second basic number x=12 is generated through sequential aneuploid reduction in diploid gametes. The intermediary gametic chromosome number of 13 is shown in parentheses because no corals have been found which are based on this number. For the sake of simplicity, 1-' pathways invoking autopolyploidy are not included. See text for additional discussion l'V I-' of alternative pathways and ancestral relationships. TETRAPLOID 2n=4x=56 .. 2n=4x=48 2n-2=54 A.ocellata A.elseyi

TRIPLOID 2n=3x=42 A.valida f-4 N N

DIPLOID 2n+2=30 ~.~ __ A.danai A.gemmifera 2n=2x=28 --+- (0=13) ...... 2n=2x=24 A.divaricata X=12 Derived basic number Gametes, Ancestral Pool X=14

Original basic number Unlike the previous models, Models 4 and 5 are not predicated on the co-existence of diploid and tetraploid forms during the radiation of Acropora. They can be accommodated by the time frame of events shown in Figure

3.16. In these models, the tetraploid somatic chromosome number 28 is of ancient origin, and may have involved an ancestor common to one or more families. In this case, the original diploid forms with 14 chromosomes probably no longer exist.

Combinations of Models

The two species with 30 somatic chromosomes (A. gemmifera and A. divaricata) may have acquired this number through separate pathways; i.e., one species may be tetrasomic (as in Models 2, 4, and 5) and the other a dibasic allopolyploid (as in Model 3).

Major features of Models 2-5 are summarized in Table

3.5. Models 2, 3, and 5 are consistent with the conclusion drawn from angiosperms that the major lines of evolution have been at the diploid level (Stace, 1980). Models 2, 3, and 4 each include allopolyploid pathways by which all species are expected to have high fertility, which is consistent with observations. In Model 5, triploid A. valida would be expected to have very low fertility, which is not consistent with observations (Table 2.2, Chapter 2). For

123 Figure 3.16. Family tree of the sixteen extant families of scleractinian corals f-' (modified from Wells, 1956), showing a time frame of changes in somatic chromosome tv ,p. number to accommodate Models 4 and 5. FAMILY TREE OF SCLERACTINIA Jurassic Cretaceous Tertiary Ouaternary o, ,• ,0 i i : Pocilloporidae : ~ Aslrocoeniidae , ~~ Acroporidae 28) i @i @ : ~ Poritidae 28 '~ ~ :

o i__! Siderastreidae , 0 ::: Agariciidae f-I ~~:;:~~::::3c:::::::::::,:J: N Fungiidae@ Ul o 0 ----: :3: Oculiniidae ; :~pectiniidae r~ ~ : i:MUSSidae@ : : : o~ Meandrinidae

: Trachyphyllidae - s::::c::::\ . Faviidae Merulinidae Caryophyllidae Dendrophyllidae MYA 195 140 65 5 TABLE 3.5. Summary of major features of Models 2-5. Included are the six Acropora species with somatic chromosome numbers that differ from 28. Somatic chromosome number for each species is listed in parentheses. Ploidy level, aneuploid variation, basic number(s), and expected fertility according to each Model are noted.

Species Model 2 Model 3 Model 4 Model 5

Acropora danai (24) tetraploid tetraploid tetraploid diploid x=6 x=6 x=7 x=12

Acropora gemmifera (30) tetrasomic dibasic tetrasomic tetrasomic tetraploid tetraploid tetraploid diploid x=7 x=7,8 x=7 x=14

Acropora divaricata (30) tetrasomic dibasic tetrasomic tetrasomic f-' tv tetraploid tetraploid tetraploid diploid

Acropora valida (42) hexaploid hexaploid hexaploid triploid* x=7 x=7 x=7 x=14

Acropora ocellata (48) octoploid octoploid octoploid tetraploid x=6 x=6 x=7 x=12

Acropora elseyi (54) nullisomic nullisomic nullisomic nullisomic octoploid octoploid octoploid tetraploid x=7 x=7 x=7 x=14

*low fertility is expected. In all other cases, high fertility is expected. reasons explained previously, where alternative pathways invoking autopolyploidy or allopolyploidy are possible, the allopolyploid route is the preferred one. All models are consistent with relative ploidy levels attained within the comparative evolutionary time scales of angiosperms and the family Acroporidae. Angiosperms radiated in the early

Cretaceous, ca. 130 MYA. Though a level as high as 38-ploid with a polyploid drop is known in the dicots (Hair and

Beuzenberg, 1961), most angiosperm polyploid lineages have not proceeded beyond the octoploid level (Stebbins, 1950;

White, 1978; Grant, 1981). The family Acroporidae emerged during the late Cretaceous, ca. 100 MYA (Wells, 1956; Veron,

1986), and the genus Acropora during the Eocene, ca. 60 MYA

(Wells, 1956). Octoploid levels of cytogenetic development within the Acropora are therefore possible within the time scales involved. In the course of time the diploid species of a polyploid complex may die out, and the lower polyploids may become extinct as well, leaving the higher polyploids behind as the only living representatives of the complex

(Stebbins, 1950; Darlington, 1973; White, 1978; Grant,

1981). Thus the ancestral diploids invoked in Models 2-4 may no longer exist.

PREDICTIONS OF THE MODELS

Model 1 is not included in the following discussion, as predictions of fertility in four species (Acropora

127 gemmifera, A. divaricata, A. valida, and A. elseyi) are inconsistent with observations. Consequently, it is not considered a plausible model. In Model 5, predictions of fertility in Acropora valida are also inconsistent with observations, but this model is included in the following discussion because it is the only model which does not invoke somatic chromosome numbers that have not yet been found. The remaining three models are consistent with observations of species fertility. Models 2-5 are diagrammed together (Figure 3.17) for ease of comparison. Of the four models, Model 5 is the most parsimonious, that is, invokes the fewest numbers of steps, followed in order by Models 4,

2, and 3. Given that the genus Acropora includes at least

150 species (Veron, 1993), chromosome number has yet to be determined for roughly 85% of its species. It would be premature, therefore, to prefer any of these models based upon the criterion of parsimony.

While recognizing the options provided by combining models, for the sake of brevity these model combinations will not be independently treated in the following predictions. Their precise predictions, however, are consistent with the logic of the methods discussed in each of the following sections.

128 Figure 3.17. Models 2, 3, 4, and 5 aligned for ease of comparison.

129 OCTOPLOID 20·8x-56 20-8x-48 ..--- ~ A.ocs/lata 2n-2·54 ~ A.elseyl ..

HEXAPLOID 20·6x-42 A.valida

A.danal TETRAPLOID 20·2-30 ~.~...... 20·4x-28 A.gemmiferll A.divaricata TRIPLOID (sterile) 1\)·21 1.\'2 DIPLOIDS, 20·2x-14 .. X.6 Ancestral X·7_-_------rD.rlwd Pool Original baele number b••le number MODEL 2

OCTOPLOID 20·8x·66 2n-8x-48 . A.oce/lata 20-2-54..--- ' A.elseyi

HEXAPLOID 20-6x-42 A.valida

A.gemmifera • A.divaricata A.danal TETRAPLOID 20.30 20·4x-28 20-4x-24

sterile hybrid /J~ or triploid / \"/20.3X_21 02x·16 1\ 20 " DIPLOIDS, Ancestral Pool Orlglna' baa Ie number MODEL 3

130 OCTOPLOID 2n"8x"56 2n"8x"48 ..--- ~ A.ocel/ata 2n-2"64 ~~ A.e/sey;

HEXAPLOID \2n"6x-42 A.va/ida

A.danal TETRAPLOID ~.~------­ / 2n"4x-24 2n+2"30 A.gemmifera A.d;var;cata TRIPLOID (sterile)

DIPLOIDS, Ancestral Pool Original basic number MODEL 4 TETRAPLOID 2n"4x-56 2n-4x-48 ..--- ~ ~ A.oeel/ata 2n-2"54 ~~ A.e/ssy;

TRIPLOID \2n"3x"42 A.valida

DIPLOID 2n+2-30 ~.~-----­ A.gemmifera A.diver/csta Derived baalc number

Gametes, Ancestral Pool Orlgina' baaic number MODEL 5

Figure 3.17. (continued) 131 Genome size

Determination of chromosome numbers and assessment of their numerical relationship within a taxon is a first line of evidence in establishing the possible presence of polyploidy. Successive levels of polyploidy should be accompanied by stepwise increases in nuclear DNA content

(Schultz, 1980). Fusions or fragmentations of chromosomes, rather than duplication of sets, may otherwise account for differences in chromosome number. Along with chromosome number, measurement of genome size has been used as evidence of polyploidy in freshwater suckers and salmonid fishes

(Shultz, 1980). No genome size data are available for corals

(McMillan and Miller, 1989; McMillan et ai., 1991).

Predictions of the models based upon genome size are summarized in Table 3.6. Although determination of DNA content would not clearly discriminate among the various models, it would be useful in supporting or opposing the overall polyploid concept suggested by this study. Aneuploid reduction in the gametes proceeds by reciprocal translocation of chromosome arms of unequal length, followed by loss of a centromere with a small amount of DNA.

Consequently, the decrease in gametic chromosome number is not accompanied by a substantial decrease in the amount of

DNA. Common to all models, then, is the expectation of no substantial difference in nuclear DNA content between

Acropora species having 28 chromosomes and A. danai with 24

132 TABLE 3.6. Summary of predictions of Models 2-5 based upon genome size and molecular data. Under genome size, the amount of somatic nuclear DNA expected for each taxon is listed. "D" represents the average amount of somatic nuclear DNA (units: 10-9 mg) of acroporids with 28 somatic chromosomes. Under molecular data, the closest relative expected among those taxa examined in this study is listed; (28) designates an acroporid with 28 somatic chromosomes, but no specific taxon is implied. See text for additional explanation.

GENOME SIZE MOLECULAR DATA

Model(s) Model(s)

Species 2,4,5 3 2,4,5 3

Acropora danai (24) D D A. ocellata A. ocellata

A. gerrunifera (30) 1.07D D (28) A. divaricata

I-' w A. divaricata (30) 1. 07D D (28 ) A. gerrunifera w A. valida (42 ) 1. 5D 1. 5D (28 ) (28 ) A. ocellata (48 ) 2D 2D A. danai A. danai

A. elseyi (54) 1.93D 1. 93D (28 ) (28 ) chromosomes. If 48 chromosomes represents an octoploid

(Models 2-4) or tetraploid (Model 5) condition, then A. ocellata would have twice as much DNA as the taxa with 24 or

28 chromosomes. If 42 chromosomes represents a hexaploid

(Models 2-4) or triploid (Model 5) state, then A. valida would have 1.5 times as much nuclear DNA as the taxa with 28 or 24 chromosomes. Assuming that each chromosome in a complement contains the same amount of DNA, A. elseyi would have 3.6% less DNA than a full octoploid (Models 2-4) or tetraploid (Model 5) complement.

In Model 3, derivation of the basic number x=8 from x=7 proceeds by division of a chromosome at the centromere, and no substantial change in the amount of DNA is expected.

According to this model, A. gemmifera and A. divaricata are dibasic allotetraploids, and would have the same amount of

DNA as taxa with 28 or 24 chromosomes. In Models 2, 4, and

5, A. gemmifera and A. divaricata, each with 30 chromosomes, originated as tetrasomics of taxa with 28 chromosomes.

Assuming equal division of DNA among chromosomes, they would be expected to have roughly 7% more nuclear DNA than taxa with 28 somatic chromosomes. However, increases in DNA can be brought about by random duplications, wherein genes or segments of chromatids are duplicated through unequal crossing over, or by redundant duplication of DNA (Keyl,

1966). Consequently, the origin of small differences in genome size may be difficult to interpret.

134 Methods for the isolation of high molecular weight

(nuclear) DNA from scleractinian corals have been described, using spawned spermatozoa as source material (McMillan et ai., 1988). A similar methodology using spawned Acropora oocytes, which can be more precisely quantified than spermatozoa, might be useful in calculating comparative genome sizes. Alternatively, a method for isolating nuclear

DNA from somatic tissue of soft corals (ten Lohuis et ai.,

1990) might be useful in determining genome size if applied to quantified batches of Acropora oocytes. The latter method avoids problems associated with the presence of pigments in somatic tissue of soft corals, and might therefore be gainfully used with pigmented Acropora oocytes.

Molecular Data

Veron and Wallace (1984) arranged the 70 species within the subgenus Acropora (Acropora) which occur off eastern

Australia into 14 "species groups" based on morphological similarity. The groupings can be viewed as a taxonomic hypothesis, in which members of a group are more closely related to each other than to members of outside groups, and the closer the groups in the series the closer their relatedness (McMillan and Miller, 1990). Though A. oceiiata is not found in east Australian waters and consequently was not included in the groupings of Veron and Wallace (1984), its morphological affinities are with members of the A.

135 humilis group (Wells, 1954; Veron and Wallace, 1984), with which it was later grouped (Veron, 1990, 1993).

The phylogenetic validity of these groups is unknown.

The addition of somatic chromosome number to species arranged in their groups (Figure 3.18) raises two discrepancies between the models proposed in this study and the groups of Veron and Wallace (1984). First according to all the proposed models (Fig. 3.17), A. ocellata is more closely related to A. danai than to other members of the A. humilis group. Second, according to Model 3, Acropora gemmifera and A. divaricata are more closely related to each other than to corals with 28 somatic chromosomes or to any of the other acroporids for which somatic chromosome number was determined in this study. The other models are consistent with the other groupings and the sequence of their arrangement, and in fact could be useful for refining hypotheses regarding phylogenetic relationships. It might be hypothesized, for example, that A. carduus is one of the progenitors of A. elseyi.

McMillan and Miller (1990) used hybridization of cloned repeated sequences of DNA that are present throughout the subgenus Acropora, but not present in other genera in the family Acroporidae, to test the relatedness of species within and between groups. They found the relatedness series based on molecular data differed from that based on morphological criteria. Their data suggested a more distant

136 Figure 3.18. Chromosome number by morphologically similar groups. Species of Acropora for which chromosome number has been determined are arranged according to "species groups" of Veron and Wallace (l984). Number of chromosomes is shown in parentheses.

I-' W -...] CHROMOSOME NUMBER BY "SPECIES GROUPS" OF VERON AND WALLACE (1984)

A. humilis ~H·OYQ A. aspera group A. nasuta group A. gemmifera (30) A. pu/chra (28) A. valida (42) A. monticu/osa (28) A. mil/epora (28) A. /utkeni (28) A. samoensis (28) A. digitifera (28) A. se/ago grouQ A. echinata groug A. ocel/ata (48) A. tenuis (28) A. carduus (28) A. e/seyi (54) I-' LV ex> A. robusta group A. hyacinthus group A. robusta (28) A. cytherea (28) A. /oripes group A. danai (24) A. hyacinthus (28) A. loripes (28) A. nobilis (28) (=A. surculosei- (=A. squerrosei-

A. formosa groug A. divaricata grouQ A. florida grouQ A. formosa (28) A. c/athrata (28) A. florida (28) A. divaricata (30)

-synonymlzed by Veron & Wallace (1984) relationship between A. formosa and A. nobilis than is indicated by the adjacent position of their groups according to Veron and Wallace (1984). It also implied a closer relationship between A. formosa and both A. hyacinthus and

A. valida than is suggested by morphological data.

Subsequent work sequencing nucleotides in this highly­ repeated 118 base-pair sequence from seven Acropora species

(McMillan et al., 1991) confirmed a close relationship among

A. formosa, A. pulchra, and A. millepora. Additional areas of disagreement arose between phylogenetic relationships implied by nucleotide sequencing and those implied by morphological similarity. By molecular data, Acropora digitifera appears more closely related to a member of the

A. echinata group (A. longicyathus) than is suggested by morphological criteria. Similarly, by molecular data A. tenuis is more distant from A. pulchra/A. millepora than is suggested by their adjacent group positions according to

Veron and Wallace (1984).

Either technique--hybridization of cloned repeated sequences or direct nucleotide sequencing of this highly­ repeated 118 base-pair sequence unique to Acropora--would be useful in testing the relationships outlined in the proposed models. Of the results generated to date using molecular criteria (McMillan et al., 1988, 1991; McMillan and Miller,

1988, 1989, 1990), none are contrary to relationships proposed by the models based on somatic chromosome numbers.

139 Table 3.6 summarizes predictions of the models based on molecular data, expressed for each species as the taxon (or taxa) with which it is expected to have the closest relationship.

If any of the proposed models are correct, then the application of these techniques would be expected to indicate that A. danai and A. ocellata are more closely related to each other than to other taxa examined in this study. Acropora valida and A. elseyi would be expected to be more closely related to an unidentified acroporid with 28 somatic chromosomes than to the other studied taxa.

According to Models 2, 4, and 5, A. divaricata would most likely be more closely related to an unidentified acroporid with 28 somatic chromosomes than to A. gemmifera. In contrast, according to Model 3, A. divaricata would be expected to be more closely related to A. gemmifera than to an acroporid with 28 somatic chromosomes.

Chromosome Pairing at Meiosis

The ability of homologues to pair at meiosis has been extensively used as a test of chromosome homology, and therefore in testing the ancestry of suspected polyploids

(Stebbins, 1950; Dawson, 1962; White, 1973; Grant, 1981).

The present state of knowledge of coral gametogenesis does not lend itself to application of this technique, as the

140 timing of meiosis has not been clearly determined for either male or female gametes. The sole observation germane to this point comes from Babcock and Heyward (1986), who observed a single polar body within the mucous egg coat of eggs of

Goniastrea favulus collected less than ten minutes after spawning, and a second polar body being extruded adjacent to the first approximately 15-30 minutes after spawning.

However, polar body release was not observed in two other species (Montipora digitata and Goniastrea aspera) which were specifically monitored for its occurrence.

Experimental Hybridizations

All of the parent species involved in successful interspecific hybridizations (Chapter 2) had 28 chromosomes.

Possession of equal chromosome number, however, is no assurance of forming successful intrageneric hybrids, as evidenced by the complete failure of hybridization between two morphs of Montipora digitata (Chapter 2), each with 28 chromosomes. The development of allopolyploidy in angiosperms frequently involves the initial formation of hybrids between parents which have two different somatic chromosome numbers (dibasic polyploidy). There are no

intrinsic barriers, therefore, against coral parents with different somatic chromosome numbers forming successful

interspecific hybrids. If any of the proposed models are

correct, however, it might be expected that species which

141 are derived from different basic numbers have sufficiently diverged so as to entirely preclude hybridization or reduce the frequency of its success.

This prediction is borne out by several experimental hybridizations which failed (R. Richmond, pers. comm.;

Chapter 2) :

A. ocellata (48) x A. smithii (28) (= A. robusta, Veron and Wallace, 1984)

A. ocellata (48) x A. surculosa (28)

A. danai (24) x A. valida (42)

In each case the parents have different chromosome numbers.

More relevant to Models 2, 3, and 5, however, the parents of each cross are derived from different basic numbers. In

Model 4, x=7 for all taxa, so the demonstrated failure in experimental hybridization cannot be ascribed to differences in basic numbers. For each model, Table 3.7 summarizes the experimental crosses which would not be successful in forming hybrid embryos, based upon the assumption that interspecific hybridization is more likely to proceed within

lineages derived from the same basic number, than between

lineages derived from different basic numbers.

Within lineages founded on the same basic number, varying degrees of successful experimental hybridization might be expected, based upon the further assumptions that:

142 TABLE 3.7. Summary of predictions of Models 2-5. For each of the six acroporids with a somatic chromosome number that differs from 28, the taxa with which experimental hybridizations are expected to fail (due to different basic chromosome numbers) are listed for each model; (28) designates acroporids with 28 somatic chromosomes. See text for additional explanation.

Model(s)

Species 2,5 3 4

Acropora danai (24) A. elseyi A. elseyi None A. valida A. valida A. gemmifera A. gemmifera A. divaricata A. divaricata (28 ) (28)

A. gemmifera (30) A. danai A. danai None A. ocellata A. ocellata A. elseyi A. valida (28 )

A. divaricata (30) A. danai A. danai None A. ocellata A. ocellata A. elseyi A. valida (28 )

A. valida (42) A. danai A. danai None A. ocellata A. ocellata A. gemmifera A. divaricata

A. ocellata (48) A. elseyi A. elseyi None A. valida A. valida A. gemmifera A. gemmifera A. divaricata A. divaricata (28 ) (28 )

A. elseyi (54) A. danai A. danai None A. ocellata A. ocellata A. gemmifera A. divaricata

143 (1) interspecific hybridization is not intrinsically precluded by different chromosome numbers in the parents, but, (2) parents with the same chromosome number are likely to have a higher frequency of success in forming interspecific hybrids than are parents with unequal chromosome number.

Predictions based upon genome size and molecular data are identical for Models 2, 4, and 5 (Table 3.6) .

Establishing somatic chromosome number for additional species of scleractinian corals would aid in resolving whether the original basic chromosome number x is <10

(Models 2 and 4), as is the consensus for angiosperms

(Ehrendorfer, 1964; White, 1978; Goldblatt, 1980; Lewis,

1980b; Grant, 1981), or is in a higher numerical range

(Model 5). Experimental hybridization trials (Table 3.7) would aide in discriminating between Models 2 and 4.

ECOLOGICAL FACTORS

Among the angiosperms, the development of polyploidy is favored by unstable environmental conditions on both ecological and evolutionary time scales (Stebbins, 1950,

1970; Ehrendorfer, 1980; Grant, 1981). The opening of new habitats for colonization in an ancestral territory as a result of climatic, human, or other disturbances provides the ecological opportunity in which allopolyploids can exploit their advantage of hybrid vigor (Ehrendorfer, 1980;

144 Grant, 1981). Neopolyploids, characterized by heterozygosity, are frequently better adapted for survival under adverse, extreme conditions than are their diploid progenitors (Lewis, 1980a). Chances for the origin, establishment, and expansion of neopolyploid angiosperms are best in widespread successional to subclimax communities, and in areas being settled by invasive, weedy floras. In these conditions of open communities, species may be present which previously existed in separate habitats, facilitating the production of hybrids and consequently of allopolyploids. By contrast, under more stable conditions and in well-balanced climax communities, stable diploids and older polyploids prevail, limiting the origin and expansion of neopolyploids (Ehrendorfer, 1980). The trend in which the incidence of angiosperm polyploids increases with latitude in European floras has been interpreted as an effect of

Pleistocene glacial disturbance rather than increasing severity of the present climate (Stebbins, 1950, 1970;

Dawson, 1962).

The availability of new ecological habitats is thus generally credited as the chief external factor favoring the establishment and spread of polyploidy. Unless such habitats are accessible, new polyploids will face competition with

already-established diploids. Though polyploids arise

sympatrically, they must have different ecological

requirements than their parents if they are to successfully

145 compete, or must be able to disperse to areas unoccupied by their parents (Dawson, 1962; Lewis, 1980a). As new genetic types, polyploids are likely to have a wider range of ecological tolerances than their diploid progenitors, and to be well adapted to the colonization of newly available areas

(Stebbins 1950; Levin, 1983). Consequently, polyploid taxa usually are distributed beyond the ranges of their progenitors or occupy different habitats within the ranges of their progenitors (Dawson, 1962; Grant, 1981; Levin,

1983) .

On ecological time scales, coral reef communities demonstrate many of the properties so frequently associated with the development and maintenance of polyploidy in terrestrial plant systems. While some workers (Goldman and

Talbot, 1976) have considered coral reef communities to be highly balanced systems in which species composition is stabilized and resilience to disturbance is pronounced, others (Sale, 1977; Connell, 1978) describe coral reefs as being nonequilibrium systems. Disturbance is viewed to be of sufficient frequency as to prevent competitive exclusion, and thus serves in maintaining high species diversity.

On coral reefs many types of disturbances can affect diversity. Sources of disturbance to shallow-water coral communities include storms, hurricanes, freshwater kill, tidal extremes, disease, bleaching, and predation (Huston,

1985; Richmond, 1993). They can vary in both frequency and

146 intensity, thereby creating stochastic opportunities for exploitation of opened space by organisms with good colonizing ability. Disturbances such as tidal exposure and wave action are most frequent and intense near the reef surface and decrease rapidly with depth. Sediment movement is greatest near the surface, but can affect coral at any depth. Thus, the effect of physical disturbance is concentrated at shallow depths and decreases monotonically with depth (Huston, 1985). This decreasing depth gradient of wave energy was reflected in the effects of Hurricane Allen on the reef at Discovery Bay, Jamaica. Although storm effects were visible to a depth of 30 meters, the damage was concentrated on two species (Acropora cervicornis and A. palmata) in the shallower reef zones (Huston, 1985). Depth­ differential effects of abiotic disturbance would fall more heavily upon members of Acropora, which constitute a dominant component of the shallow reef assemblage (Wallace,

1978; Veron, 1986), than on species which occur over a greater depth gradient.

Changes in the frequency and intensity of extreme events are probably more ecologically significant than moderate changes in the mean values of environmental factors

(Smith and Buddemeier, 1992). Storms provide long-term episodic control of reef community development by catastrophic pruning and substrate renewal. Changes occurring under the slow influence of continuously acting or

147 chronic factors may be reversed by sudden but rare catastrophic disturbances. For example, Hurricane Allen, the most severe Caribbean hurricane on record as of 1980, reversed a slow trend in the disproportionate reduction of rarer, competitively inferior species in Discovery Bay

(Porter et al., 1981). In contrast to mortality under ambient conditions, mortality induced by storm activity was greatest in the most abundant species.

Severe storms may nearly obliterate populations of species already prone to fragmentation, while the long-term survivors may be those which have greater resistance to fragmentation. The Caribbean species Acropora cervicornis routinely propagates by asexual fragmentation, with little apparent recruitment from larvae (Bak and Engel, 1979;

Rylaarsdam, 1983). In Discovery Bay, 98% of A. cervicornis fragments which were found 3-9 days after Hurricane Allen were dead 5 months later (Knowlton et al., 1981). Jackson and Hughes (1985) separate clonal organisms into a spectrum of space-occupying strategies, the extremes of which are termed "guerrilla" and "phalanx". Guerrilla-strategists grow rapidly and fragment readily, but have poor competitive ability. Consequently, during its lifetime a guerrilla­ strategist clone moves across the substratum, losing competitive encounters with other clonal species, but surviving by moving on. Guerrillas tend to be more abundant in environments with higher levels of disturbance. Phalanx-

148 strategists are built more robustly and do not fragment as readily. They grow and take over space slowly, and tend to inhabit environments with lower levels of disturbance. Of the six species of Acropora which may have arisen by polyploidy and/or aneuploidy, five (Acropora danai, A. gemmifera, A. valida, A. ocellata, A. elseyi) have robust morphologies which appear relatively resistant to fragmentation. They are frequently found in high-energy environments, in which their sturdy skeletons can withstand both normal background and storm levels of wave energy.

After a series of damaging storms in Guam during 1991 and

1992 (Birkeland, 1992), Acropora danai, A. gemmifera, A. valida, and A. ocellata were among the few shallow-water acroporids remaining in Pago Bay (pers. obs.). Species with more friable morphologies, such as A. squarrosa and A. cerealis, were severely "pruned" by the succession of major typhoons. Even the robust Acropora phalanges, however, will undergo fragmentation and cloning given severe levels of disturbance. A number of large, living fragments of the sturdy Acropora danai were observed strewn over the shallow­ water reef in Pago Bay during July, 1992, with various degrees of re-cementation to the substrate (pers. obs.)

After a severe disturbance such as a hurricane, then, it is the phalanx species which are more likely to survive. Colony size affects the ability of some corals to reproduce

(Szmant-Froelich, 1985). The large fragments into which

149 robust phalanges are separated may still retain their ability to sexually reproduce. Species with robust morphologies thus are better able to renew growth and produce gametes than species with more fragile morphologies after severe storms. Similarly, in terrestrial angiosperms, polyploidy has been facilitated by unstable environmental conditions which favor organisms with good colonizing ability under adverse conditions. Depth-differential effects of abiotic disturbance, ecological preference for shallow­ water habitats, and capacity to colonize even severely disturbed areas through asexual propagation of fragments are attributes of Acropora which predispose this genus to the development of polyploidy.

SYMPATRIC SPECIATION

Discussions of Indo-Pacific coral biogeography which attempt to explain modern distribution patterns are invariably couched in models of allopatric speciation.

Mechanisms including sea level change (Rosen, 1984; Potts,

1984a, 1985; McManus, 1985), land bridges (McManus, 1985), and unspecified vicariance events (Wallace et ai., 1991) have been invoked to effect physical isolation of populations. In comparing three models of the effect of

Pliocene-Quaternary sea-level fluctuations on scleractinian speciation, Potts (1984a, 1985) suggested speciation was suppressed because sea levels fluctuated so rapidly that

150 clonal coral populations did not experience enough generations in isolation to differentiate into new species.

Polyploidy, however, represents an essentially instantaneous sympatric speciation event, for in just one generation a postzygotic barrier causes reproductive isolation and interrupted gene flow between a fledgling population of polyploids and the parent populations that surround it

(Dawson, 1962; Grant, 1981; Levin, 1983; Futuyma, 1986).

Even-numbered polyploids are isolated from their ancestors because any hybrids they produce will be odd-numbered polyploids, which are largely or entirely sterile as a result of meiotic irregularities (White, 1978). For example, allotetraploids will be genetically isolated from both diploid parents, for the triploid backcross progeny will have a high proportion of aneuploid, inviable gametes

(Grant, 1981). Even when the progeny between ploidy levels are fertile, as may be the case for tetraploid crosses between a hexaploid and a diploid, gene flow is prevented by the tetraploids producing sterile triploid offspring when crossed with the diploid parent and producing sterile pentaploid offspring when crossed with the hexaploid parent

(Dawson, 1962). A biological barrier to interbreeding exists without any spatial segregation.

Although many extant coral genera appeared in the

Jurassic, the genera Acropora, Fungia, Galaxea, Pocillopora, and Seriatopora originated in the Tertiary (Achituv and

151 Dubinsky, 1990). The genus Acropora appeared during the

Eocene (ca. 60 MYA) (Wells, 1956), but the ages of living species are not known (Potts, 1985), because diagenesis in fossil reefs usually destroys the skeletal detail needed for identification at the species level (McMillan et ai., 1991).

Despite its geological record back to the early Tertiary,

Acropora does not seem to acquire its modern ecological prominence anywhere until some time in the Plio-Pleistocene

(Rosen, 1993). Successful neopolyploid species tend to originate under unstable environmental conditions and in areas of contact and hybridization between diploid ancestors

(Ehrendorfer, 1980). Rapid, sympatric speciation by polyploidy during Pliocene-Quaternary glacial periods, particularly during sea-level transgressions when new shallow-water habitats became available, may in part account for the notable diversification of Acropora.

The onset of Quaternary glaciations around 2.5 MYA intensified background levels of cyclic sea level fluctuations, giving rise to 100,OOO-year frequency cycles that dominated the late Pleistocene (Pauley, 1991). During sea level regressions, shallow shelf and lagoonal habitats were largely eliminated, and corals became restricted to predominantly oceanic, continental slope, or outer atoll slope reefs (Potts, 1984a; Pauley, 1991). The effects of glaciation on loss of coral diversity were more dramatic in the Caribbean than in the west Pacific. Thirty-two percent

152 of Pliocene hermatypic corals became extinct in the

Caribbean by the Early Pleistocene, while only 11% of

Pliocene coral species in Papua New Guinea and 10-16% of

Pliocene corals on the south Polynesian island of Niue became extinct (Pauley, 1991). As Pacific rates of coral extinction during the Pliocene were no greater than background levels (Pauley, 1989, 1991), pre-existing levels of Acropora diversity would have been preserved, serving as a parental species pool for future speciation processes.

During sea level regressions, species were driven to refugia on continental or atoll slopes (Potts, 1984a; Potts and

Garthwaite, 1991; Pauley, 1991), where limited habitat heterogeneity and competition for space with indigenous taxa may have inhibited any speciation process by polyploidy.

Populations isolated in ocean basins (McManus, 1985) may have diverged, though not enough to achieve full reproductive isolation. During sea level transgressions, however, new, heterogeneous shallow-water habitats became increasingly available (Pauley, 1991). Newly flooded, unoccupied inner reefs and lagoonal areas would be available for colonization by newly-arisen hybrids as well as by species that survived low sea level stands. Though sexually sterile, hybrids could persist and spread through vegetative cloning, in time giving rise to allopolyploids by any of the several mechanisms outlined earlier. Competition between parental and polyploid genotypes would be reduced by the

153 increased habitat diversity. The heterogeneity of reef habitats and their capacity to exert a qualitatively unique selective regime that acts to select for different kinds of individuals has been demonstrated by Potts (1984b)

Simulating environmental change, he reciprocally transplanted cloned corals among five different reef habitats (outer flat, crest, lagoon, inner flat, reef slope) and demonstrated differential growth and survival of replicate genotypes in the different habitats.

As sea level rise proceeded during transgressions, varied habitats continued to be created as the degree of communication between inner reef and oceanic water was altered (Pauley, 1991). As is the case for terrestrial angiosperms, the opening of new habitats in developing, successional reef communities would provide the ecological conditions in which polyploids with colonizing ability could exploit their hybrid advantage. During Pliocene-Pleistocene sea level transgressions, then, speciation by polyploidy would have been facilitated by the availability of new habitats for colonization and the increased habitat heterogeneity by which competition between parent and polyploid would be reduced.

BIOGEOGRAPHY

The tropical Indo-Pacific is the largest marine biogeographical province. It extends from the coasts of

154 Africa and Arabia, through the island archipelagoes of the

Far East, to the eastern shores of the Pacific. Coral diversity is not evenly spread throughout the Indo-Pacific, but is centered within the triangle demarcated by the

Philippines, Indonesia, and Papua New Guinea. Patterns of global coral distribution, and hypotheses concerning mechanisms by which these patterns have been generated, are treated in further detail in Chapter 4. Though coral biogeography has traditionally been pursued at the generic level due to problems of species identification, recent taxonomic revisions (Veron, 1986; Veron and Pichon, 1976,

1980, 1982; Veron et al., 1977; Wallace, 1978; Veron and

Wallace, 1984) are providing the means by which reef biogeographers can increasingly work at the species level.

The genus Acropora is widely distributed both longitudinally and latitudinally within the tropical Indo­

Pacific. Longitudinally the genus extends from the Red Sea to the west coast of Colombia; latitudinally its members are found from Japan in the north to Lord Howe Island in the south (Veron, 1986, 1993). The Red Sea contains nineteen species of Acropora (Sheppard and Sheppard, 1991), exemplifying a general pattern in which the number of coral species and genera in the Indian Ocean is lower than in the

Far East region, but rises abruptly again in the Red Sea.

Sheppard and Sheppard (1991) suggest that during Pleistocene low sea level stands the Red Sea became an inhospitable,

155 hypersaline lake, separated by an emergent sill from the

Indian Ocean. Salinities of >50 ppt prevailed in the Red

Sea, killing all existing corals. Contemporary reefs recommenced growing on pre-Pleistocene reef structures only in the last 6000-7000 years. Summertime upwelling of cold water in the Arabian Sea, which presently acts as a partial barrier to the passage of coral larvae from the Indian

Ocean, was presumably weaker during the Holocene transgression and thus did not hamper the historical recolonization of the Red Sea. During the winter, upwelling ceases and there is a large flux of warm water between the

Red Sea and Indian Ocean. This model, which describes historical and contemporary colonization patterns (Sheppard and Sheppard, 1991), does not address the source area of contemporary coral larvae entering the Red Sea nor the strongly seasonal (late spring or summer) nature of coral broadcast spawning in the northern hemisphere.

Only eleven species of hermatypic scleractinian coral presently exist in the eastern Pacific (Richmond, 1990a).

Two views dominate concerning the historical biogeographical development of eastern Pacific coral communities since their separation from the Caribbean at the end of the Pliocene.

The "long-distance dispersal" view argues that modern eastern Pacific coral communities are derived by larval dispersal from the central Pacific, following massive extinctions after tectonic raising of the Isthmus of Panama

156 (Dana, 1975; Zinmeister and Emerson, 1979; Richmond, 1987).

The ·'vicariance" view maintains that eastern Pacific coral communities are derived from a previously widespread, pan­

Tethyan biota that has since been modified by tectonic events, speciations, and extinctions (McCoy and Heck, 1976;

Heck and McCoy, 1978). Acropora valida is the most widespread Acropora species, its distribution extending from the Red Sea (Sheppard and Sheppard, 1991) to Gorgona Island,

Colombia (von Prahl and Mejia, 1985). The latter record consists of three colonies which have not been relocated since their documentation (P. Glynn, pers. comm.), and a viable breeding population of this species most likely does not exist in this region.

Regardless of the mechanisms involved, Pleistocene glaciations clearly altered modern coral distributions at the geographic extremes of Indo-Pacific Acropora distribution. A common pattern in both the angiosperms and the ferns is that the polyploid members of a group are geographically widespread while their diploid relatives have more narrow distributions (Stebbins, 1950; Grant, 1981), but this tendency has many exceptions (Stebbins, 1950;

Ehrendorfer, 1980). Different species of Acropora vary widely in the extent of their geographical distribution,

from the pan-Indo-Pacific A. valida to a species endemic to

the Houtman Abrolhos Islands off western Australia (Veron,

1993). All of the species in the present study which have a

157 somatic chromosome number other than 28 have wide distributions (Figure 3.19), but no general pattern relating proposed ploidy level to extent of distribution can be discerned' .

Of 99 Acropora species whose distributional patterns are described by Veron (1993), 23 have relatively narrow distributions, restricted to specific coastlines (e.g.,

Great Barrier Reef), island archipelagoes (e.g., Japan or the Philippines), atolls (e.g., Cocos [Keeling] Atoll), or islands (e.g., Houtman Abrolhos Islands). In a morphometric analysis of five morphologically similar species comprising a "species group", Wallace et al. (1991) noted that Acropora taxa showing the greatest degree of endemism display character states that are the most derived. Relative to more widespread taxa, species with restricted distributions may therefore be of more recent origin, and would be worth studying in terms of the ideas presented here. It could be hypothesized that they are relatively new taxa which have not had sufficient time to extend their geographical range.

It would be of particular interest to study their reproductive potential: are they sterile, which might

'Acropora ocellata has not been treated in recent taxonomic revisions (Veron and Wallace, 1984) and may present some uncertainties of identification. Veron (1993) records its distribution between the Red Sea and Cocos (Keeling) Atoll, and notes its absence from southern Papua New Guinea, western Australia, the Philippines, and Japan. Randall and Myers (1983) list this species from Guam and Birkeland et al. (1985) from Samoa.

158 Figure 3.19. Distribution of Acropora danai, A. divaricata, A. gemmifera, A. valida, and A. elseyi. Stars are reliable records from Veron and Wallace (1984). Boxes refer to specimens in biogeographic collection lodged in the Museum of Tropical Queensland, Townsville, Australia. Maps were compiled and provided courtesy of C.C. Wallace.

159 .,,

If.. • . • I!I ~$ ~' • -r...... Distrr-i. b'LL tion of A. e lsey1: 160 suggest a hybrid origin, or do they produce viable gametes, capable of high levels of cross-fertilization? If so, then their somatic chromosome number could be determined by using the techniques described here. Does their somatic chromosome number suggest they might be neopolyploids?

The Houtman-Abrolhos Islands endemic (Acropora sp. W.

Australia; Veron, 1993) presents an example of the type of study which might be undertaken. Forty-two other species of

Acropora have been recorded in these islands off the coast of western Australia (Veron, 1993). The endemic has close affinities with A. horrida (Veron, 1993). Mass spawning has been documented on western Australia reefs during March

(Simpson, 1985; Atkinson and Atkinson, 1992). If the endemic were sterile, then it could be hypothesized that it is a hybrid which has not yet bypassed the sterility barrier. A series of experimental hybridizations between A. horrida and morphologically similar species which spawn at the same time could be conducted to narrow down the range of probable parents. Molecular data (McMillan et ai., 1991) would be useful in establishing the degree of relatedness among the endemic, A. horrida, and any Acropora with which A. horrida formed interspecific larvae. Ideally the endemic's identity as a hybrid might be demonstrated by resynthesis. In contrast, if the endemic were fertile, it could be hypothesized that it is an allopolyploid which has overcome the sterility barrier. Somatic chromosome number could be

161 determined by using the techniques described in this chapter. A series of backcrosses could be conducted between the endemic and A. horrida, and between the endemic and other morphologically similar acroporids which spawn at the same time. Molecular data (McMillan et ai., 1991) could be used to determine the degree of relatedness between the endemic and any acroporids with which it formed interspecific larvae.

PARTHENOGENESIS AND ANIMAL POLYPLOIDY

In animals, polyploidy has frequently been associated with parthenogenesis (White, 1978; Sexton, 1980; Hughes,

1989; Orr, 1990). Several alternative cytogenetic mechanisms of parthenogenesis result in clonal propagation of the maternal genotype. Apomictic (ameiotic) parthenogenesis is a mechanism of cloning, as are forms of automictic parthenogenesis that suppress genetic mixing (Hughes, 1989).

In apomictic parthenogenesis, meiosis is suppressed, and the egg undergoes only a single mitotic maturation division. The maternal genotype is consequently preserved and passed intact to the offspring. In the more primitive automictic parthenogenesis, meiosis is involved in the production of eggs, together with either premeiotic, intrameiotic, or postmeiotic restitution of the genome. If chromosome doubling occurs before meiosis, and no crossing-over occurs during meiosis (pre-meiotic restitution), the resultant egg

162 genome will be identical to the maternal genome.

Alternatively, eggs which are identical to the maternal genome may arise through intrameiotic restitution, which occurs by suppression of the first meiotic division, accompanied by lack of crossing over. Postmeiotic restitution of the maternal genome occurs when normal meiosis with no crossing over is followed by the fusion of two pronuclei. Both apomictic and automictic parthenogenesis can consequently result in the production of unreduced eggs, the same end product which is fundamental to several mechanisms by which polyploidy arises in non-parthenogens.

Though parthenogenesis and vegetative reproduction are both mechanisms of cloning, they are usually mutually exclusive, probably due to structural constraints (Hughes,

1989). Vegetative reproduction requires a relatively simple morphology that can be subdivided without critical disruption and regenerated quickly from incomplete or rudimentary parts. In contrast, parthenogenesis is found in animals which are not as readily subdivided or which have limited regenerative capacity. Animals with intricate exoskeletons, such as arthropods, cannot be viably fissioned or fragmented. Structural complexity therefore impedes vegetative reproduction, and parthenogenesis becomes the only possible method of extensive cloning in such animals.

Vertebrates are another group in which morphological complexity is too great for vegetative reproduction to be

163 possible. Both vegetative reproduction and parthenogenesis have arisen independently among many animal lineages and repeatedly in some of them (Hughes, 1989).

Apomictic parthenogenesis is widespread in obligate parthenogens and common in cyclical parthenogens. In cyclical parthenogenesis, apomixis alternates with bisexual reproduction. Its advantages are equivalent to those gained by bisexual animals that clone by vegetative reproduction

(Hughes, 1989), such as corals. Apomictic parthenogens are frequently polyploid, and automictic parthenogens only somewhat less so (Futuyma, 1986; Hughes, 1989). The significance of the association between vegetative clonal reproduction in plants, parthenogenetic clonal reproduction in animals, and polyploidy may be the spatial replication and temporal longevity of clonemate genomes, which increases the opportunities for the rare cytogenetic events that are requisite for polyploidy to occur. By this reasoning, polyploidy might well be expected to arise in animals such as corals that clone by vegetative reproduction.

In contrast to the view that polyploidy has been insignificant in animal evolution (White, 1978), Bogart

(1980) asserts, llthere is a growing bank of data which suggests that naturally occurring animal polyploids may play an interesting and significant role in population genetics and speciation." Among non-insect invertebrates, there are a number of polyploid turbellarians, oligochaete worms, and

164 leeches. Turbellarians can reproduce by parthenogenesis, gynogenesis, or asexually by fission. In the oligochaetes, polyploidy and parthenogenesis are frequently combined, but there are also many polyploid species that reproduce sexually (White, 1973). Fifty-two percent of North American earthworms clone through apomictic parthenogenesis (Hughes,

1989), and among these polyploidy is common. In many annelid apomictic parthenogens where polyploidy has arisen, the different ploidy levels partition the habitat. The oligochaete Lumbricillus lineatus has gynogenetic triploid, tetraploid, and pentaploid clones, which arise independently from bisexual diploids in different parts of their geographical range. These polyphyletic ploidy lineages partition the intertidal gradient, with triploids predominating highest on the shore (Hughes, 1989). Though relatively more rare in pulmonate molluscs, tetraploid, hexaploid, and octoploid populations of several species are known. Among prosobranch molluscs, the freshwater snail

Potamopyrgus antipodarum is an allotetraploid apomictic parthenogen which spread rapidly throughout Europe after accidental introduction, the result of good colonizing ability and broad ecological tolerance (Hughes, 1989). Among non-insect arthropods, the brine shrimp Artemia salina consists not only of a diploid sexual form but also of asexual diploids, triploids, tetraploids, pentaploids, octaploids, and decaploids (Futuyma, 1986).

165 Fewer than 100 known insect forms are polyploid, and all well documented polyploid insects reproduce parthenogenetically (Lokki and Saura, 1980). Most polyploid insects have the more advanced type of parthenogenesis, apomixis. Polyploid parthenogenetic insects have apparently arisen from diploid bisexual ancestors by way of diploid parthenogenetic intermediates, though none of the hypotheses concerning their cytogenetic origin has been experimentally verified. In congruence with plant polyploid attributes of efficient vegetative reproduction and perenniality, polyploid insects have good colonizing ability and relatively long life spans. Despite typically being flightless or sluggish forms, polyploid insects have very wide distributions in comparison with the diploid bisexual race (Lokki and Saura, 1980; Hughes, 1989). Dispersal and colonization of parthenogenetic forms is efficient because the transport of a single individual at any stage of development suffices to establish a new population. Most polyploid insects have life cycles extending over two or more years (Lokki and Saura, 1980). In the well-studied

European lepidopteran Solenobia triquetrella and the coleopteran weevil genera Otiorhynchus and Trachyphloeus, diploid bisexual forms live in areas which were unglaciated in the most recent Ice Age, while polyploid parthenogenetic forms have spread to areas exposed by the retreating ice sheet (Lokki and Saura, 1980; Hughes, 1989). Colonization of

166 areas disturbed by Pleistocene glaciation appears to be the main factor explaining the modern distribution of different forms.

Among vertebrates, some of the more primitive fishes are tetraploids (paddlefish, sturgeon, gar), as are some members of the more modern salmonid family (trout, whitefish, grayling) (Schultz, 1980). Members of the freshwater sucker family Catostomidae are believed to have arisen as tetraploids from the closely-related minnow family

Cyprinidae (Uyeno and Smith, 1972). Some of the minnows

(barb, carp, and goldfish) are tetraploid (Ohno et al.,

1967; Wolf et al., 1969). Except for goldfish, the ploidy condition of these groups is believed to be of an antiquity that limits investigation of their origin (Schultz, 1980)

Among goldfish, as well as two genera of the livebearing fish family Poeciliidae, however, polyploidy appears to be an active process. Available data suggest that the origin of polyploidy involves, as a first step, hybridization which results in the formation of diploid gynogenic or hybridogenic populations (Schultz, 1980). Ohno (1970) has championed the view that, rather than playing a secondary role in animal evolution, polyploidy has provided the additional, uncommitted gene loci necessary for major steps in animal evolution. The marked increase of DNA in amphibians relative to fishes is believed to have arisen by polyploidy. All of the polyploid amphibians and reptiles are

167 associated with diploid species; there are no distinctly polyploid families, genera, or even species groupings.

Reptilian polyploids are all-female triploid lizards which are thought to reproduce parthenogenetically. Urodele amphibians have all female triploid populat~0ns which reproduce gynogenetically. Anuran amphibians have bisexual triploid, tetraploid, hexaploid, or octoploid populations or species (Bogart, 1980). Polyploids are thought to arise from females, as there is no direct evidence that they can be generated by polyspermy or unreduced sperm. Bogart (1980) suggests that most amphibian and reptile polyploids arose through hybridization involving genetically divergent species or populations.

With the exception of Acropora digitifera, chromosome number for all species in this study were determined from colonies collected at random within a single geographical area. In some documented cases of animal polyploidy (e.g., the oligochaete Lumbricillus lineatus and the brine shrimp

Artemia salina), different clones appear to arise independently in separate localities, or separate clones of different ploidy levels may partition a habitat gradient in a single locality. Such cases of varying ploidy level within a single morphological species are invariably coupled with the process of parthenogenetic autopolyploidy. Because parthenogenesis has not been demonstrated in Acropora, it is not expected that interspecific races varying in ploidy

168 level are to be found within the same locality or between different geographic areas. However, autopolyploidy, the result of parthenogenesis, can also arise through other mechanisms, including self-fertilization by unreduced gametes and zygotic chromosome doubling. Low levels of self­ fertilization have been documented in some species of

Acropora (Heyward and Babcock, 1986; Richmond and Hunter,

1990). Therefore, the possibility that any given morphological species of Acropora might have varying chromosome numbers within the same or separate geographical areas cannot be ruled out by this study.

WHY IS POLYPLOIDY NOT MORE COMMON IN ANIMALS?

The classic explanation for the scarcity of polyploidy in bisexual animals was that it disrupted the sex­ determining mechanism. Muller (1925) suggested that in organisms with heterogametic sex determination, doubling the chromosome number would upset the regular segregation of sex chromosomes and therefore the determination of sex. For example, if the XY gametes produced by XXYY individuals were fertilized by the xx gametes from an XXXX parent, the result would be XXXY tetraploids. Such individuals in Drosophila, on which Muller's arguments were based, are sterile intersexes. It is now known that sex determination in

Drosophila is based on the ratio of sex chromosomes to autosomes. By contrast, in many groups of both animals and

169 plants sex determination depends instead on the chromosome peculiar to the heterozygous sex (the Y or W chromosome) possessing strongly dominant sex-determining factors. In such groups where diploid males are heterogametic, XXXX are female and XXXY are male. Polyploidy is no longer theoretically precluded in such groups of bisexually­ reproducing animals. Thus, the tetraploid race of the artificially-produced stock of silkworms Bombyx has a 1:1

ZZZW (female): ZZZZ (male) sex ratio without any intersexes or sterile genotypes (White, 1978).

White (1978) subsequently suggested that polyploidy was rare among bisexual animals because cross-fertilization was obligatory, and the development of polyploidy would be enhanced by self-fertilization. Indeed, because the cytogenetic rnistakes necessary for the formation of polyploids are less likely to occur in two individuals than in a single individual, species capable of self­ fertilization would appear more prone to its development.

White recognized that the lack of information about the capacity for self-fertilization in many hermaphroditic groups made his idea difficult to evaluate. However, although there are many known species of polyploid fish, the only known species of self-fertilizing hermaphroditic fish is not polyploid (Schultz, 1980).

The apparent rarity of polyploidy in animals may be an artifact of the type of animals in which chromosome number

170 has been adequately studied. The factors which predispose

Acropora to polyploidy are life history traits commonly ascribed to plants: asexual propagation, long life span of genetic clones, and high fecundity of individual colonies.

These factors greatly amplifify the opportunities for the rare cytogenetic mistakes that are necessary for polyploidy to arise. However, these traits are also characteristic of many marine animals with a clonal life history. The principles outlined in this work should be applicable to other organisms that possess suites of life history characteristics similar to those of corals, and that are exposed to environmental changes of similar frequency and magnitude. Grant (1981) suggested that clonal life history was exceptional in the animal kingdom and was connected to the relative lack of poly~loidy. However, thirteen metazoan phyla contain organisms which clone through vegetative methods and twelve metazoan phyla contain parthenogens

(Hughes, 1989). Orr (1990) argues that polyploidy disrupts the dosage-compensation mechanism of organisms possessing a degenerate sex chromosome, and that polyploidy is rare among animals because they often possess a degenerate sex chromosome, but common among plants because they rarely possess a degenerate sex chromosome. However, hermaphroditic animals, lacking sex chromosomes, frequently do show polyploidy (White, 1978; Orr, 1990). Moreover, the majority of coral species are hermaphroditic (Richmond and Hunter,

171 1990). Invertebrate animal chromosomes have been poorly studied relative to those of plants because most marine invertebrates are neither commercially farmed nor otherwise exploited for human use. Although more than 500 species of hermatypic corals occur in the Indo-West Pacific (Veron,

1986), until this study, chromosome number had been determined for only four species (Heyward, 1985b,c). Animal chromosome studies have generally been carried out in unitary (non-clonal) organisms (White, 1973), which do not possess the life history characters conducive to polyploidy.

A cytological survey of the genomes of animal taxa might reveal higher incidences of polyploidy than are now apparent, particularly among hermaphroditic or clonal animals such as sponges, bryozoans, and ascidians. The first estimates of somatic chromosome number in actinians showed that one of the two species studied was tetraploid (Shaw et al., 1987). Electrophoretic banding patterns of three enzymes from the coral Pocillopora damicornis in Hawaii are inconsistent with diploid explanations and suggest that a tetraploid model is more likely for this species (Stoddart,

1985) .

172 CHAPTER 4

REPRODUCTIVE STATUS OF SOME CORALS IN PALAU AND YAP

INTRODUCTION

In the period since the documentation of mass coral spawning on the Great Barrier Reef (Harrison et al., 1984;

Willis et al., 1985; Babcock et al., 1986), studies of coral reproduction have sought to discover the degree to which spawning synchrony occurs in other geographic regions.

Additional information regarding coral reproduction exists for the Caribbean, eastern Pacific, Hawaii, central Pacific, southern Japan, Red Sea (reviewed in Richmond and Hunter,

1990), and western Australia (Simpson, 1985). The degree of reproductive synchrony among species in these regions ranges from annual mass spawning involving more than 140 species over several nights in late spring on the Great Barrier

Reef, to little apparent interspecific synchrony in Hawaii or the Red Sea. The general trend appears to be one of closer synchrony with increased annual seawater temperature range (Richmond and Hunter, 1990).

Differences exist among coral species with regard to sexuality (hermaphroditic vs. gonochoric), mode of reproduction (internal brooding of larvae vs. broadcast spawning of gametes), and time of reproduction (Harrison and

Wallace, 1990). The majority of species studied to date are hermaphroditic broadcast spawners (Richmond and Hunter,

173 1990). Most coral species have a lunar periodicity in spawning behavior which is consistent from year to year.

Eggs of many coral species become pigmented (yellow, orange, or pink) several weeks prior to maturation and spawning

(Harrison et al., 1984; Babcock et al., 1986; Heyward et al., 1987; Heyward, 1988).

Despite a total of 310 coral species reported from

Palau (Maragos, 1991), fewer than a dozen species had previously been examined for reproductive activity. The most recent studies date from the early 1940's (Abe, 1937;

Kawaguti, 1941). No published observations of coral reproduction exist for Yap. In Guam, the area closest to

Palau and Yap which has been reasonably well studied, spawning behavior which results in several small-scale synchronized events has been documented over the summer months June - August (Richmond and Hunter, 1990; Chapter 3) .

In this chapter I present the results of monitoring selected corals for reproductive activity in Palau during late spring and early summer 1993, with additional observations from

Yap.

MATERIALS AND METHODS

PALAU

The Palau Islands are the westernmost group of the western Pacific Caroline Islands. The islands form a gently curving arc which centers around 7°N latitude and 134°E

174 longitude. Eight different sites representing a variety of habitats were sampled, including shallow fringing reefs off volcanic (Malakal) and lagoonal limestone islands (Rock

Islands, Ulong Island), upper channel slopes (K-B Bridge,

Ngesaol, and Ngeremdiu), and outer barrier forereefs

(Lighthouse and Ulong Channels) .

Twenty-seven species from five families were sampled and examined. Of these, 26 species are known to broadcast spawn gametes and one (Acropora brueggemanni) internally broods larvae (Atoda, 1951). Twenty-four species are hermaphrodites, while three (Porites cylindrica, Porites lutea, and Turbinaria peltata) are gonochoric. The genus

Acropora was prioritized for study because its members are major components of the shallow reef assemblage, and because maturing eggs are easily visible. Species were identified using Veron (1986) and a list compiled by Maragos (1991) of corals reported from Palau since 1975.

All corals sampled measured at least 30 cm diameter, to avoid sampling young colonies which had not yet reached sexual maturity. Some species of Acropora are well known to propagate through asexual fragmentation (Randall, 1973;

Birkeland et al., 1979; Bothwell, 1981; Highsmith, 1982;

Wallace, 1985), and care was taken to minimize sampling from clonemates by choosing widely separated or morphologically distinct colonies.

175 Branch pieces 5-7 cm in length were removed by hand and both broken ends were examined by eye for conspicuous eggs.

Collected samples were fragmented with hammer and chisel

into at least five pieces, and the polyp calices along the

periphery of each fragmented piece were examined with a

dissecting microscope for eggs and testes. For each gravid

sample, the maximum length of 10-50 eggs was measured and

color of eggs was noted. General, non-quantitative

assessment of sample fecundity was also recorded.

Viable portions of colonies with pigmented eggs were

placed in a shallow flow-through tank and examined several

times each evening for signs of spawning. When spawning was

observed, the remaining part of the colony in the field was

subsequently examined for presence or absence of gonads.

YAP

Yap Island is located in the western Caroline Islands,

roughly 250 miles northeast of Palau and 550 miles southwest

of Guam. Partway through the study, when it became apparent

that the temporal pattern of reproduction in Palau differed

from that characteristic of Guam, a trip was made to Yap to

assess the fertility of selected corals at an intermediate

latitude. Two to eight colonies of each of ten species were

1 sampled on the shallow (10-15 ) reef flat adjacent to

Guufnuw Channel and examined by eye for presence or absence

of eggs.

176 RESULTS

PALAU

Of 27 species examined between May 6 and June 11, 1993, thirteen had gravid colonies (Table 4.1). Of these thirteen species, twelve are known to broadcast spawn gametes and one

(Acropora brueggemanni) broods larvae. Of the twelve broadcast-spawning species, colonies of six species

(Acropora carduus, A. clathrata, A. florida, A. secale,

Acropora sp. [arborescent], and Turbinaria peltata) were maintained in flow-through tanks and were observed spawning

(Table 4.2). Spawning by all six species occurred during the week between full moon and last quarter, between 7 PM and 10

PM. Acropora carduus spawned four and five nights after the

May full moon. Acropora clathrata, A. secale, A. florida, and Turbinaria peltata spawned six nights after the June full moon, and Acropora sp. (arborescent) spawned one night later. Subsequent observations of gravid colonies which had been flagged for relocation on the reef confirmed that spawning had occurred in these colonies as well.

Seven of the thirteen gravid species (Acropora florida,

A. secale, Acropora sp. [arborescent], A. granulosa, A. formosa, A. horrida, A. paniculata) still had immature white eggs at the end of the study period (June 11). Of these, three species had some colonies which had spawned after the

June 4 full moon (Acropora florida, A. secale, Acropora sp.

[arborescent]). The size of eggs still developing in gravid

177 TABLE 4.1. Reproductive status of corals sampled in Palau and examined with a dissecting microscope for developing eggs or larvae between May 6 and June 11, 1993

Species Colonies Colonies Sampling sampled gravid interval

Acropora carduus 6 6 May 6-May 8 A. hyacinthus 10 o May 6-June 2 A. formosa 39 12 May 6-June 9 A. humilis 20 o May 6-June 9 A. nasuta 17 o May 6-June 9 A. tenuis 22 o May 6-June 9 A. echinata 18 o May 6-June 9 A. digitifera 24 o May 6-June 9 A. secale 12 2 May 6-June 11 A. divaricata 2 2 May 8 A. pulchra 4 o May 8-May 16 A. cytherea 13 o May 8-June 2 A. austera 5 o May 8-June 2 A. granulosa 5 5 May 8-June 8 A. clathrata 16 16 May 8-June 11 A. florida 19 15 May 8-June 11 A. valenciennesi 4 1 May 12-May 24 A. horrida 4 4 May 14-June A. paniculata 1 1 June 4-June 11 Acropora sp. 2 2 June 4-June 11 (arborescent) A. brueggemanni 12 8 May 6-May 22 Porites lutea 3 o May 6 Porites cylindrica 8 o May 6-May 8 Echinopora lamellosa 5 o May 6-May 22 Montipora aequituberculata 13 o May 8-May 22 Turbinaria peltata 1 1 May 21-June 11 Pectinia paeonia 4 o May 22

178 TABLE 4.2. Summary of 1993 spawning dates and egg size for ten fertile coral species in Palau. Numbers after spawning dates refer to number of nights following the full moon. Sizes of immature, white eggs were measured shortly after the June full moon. Between 30 and 150 eggs were measured for each species. NM= not measured.

Species Spawning Mean diameter Mean diameter dates spawned eggs white eggs (jlm) (jlm)

Acropora carduus May 10,11 (4,5) 531 A. formosa 470 A. secale June 11 (6 ) NM 674 A. granulosa* ?June? (698 ) 540 A. clathrata June 11 (6) 566 A. florida June 11 (6) 562 A. horrida 567 A. paniculata 610 Acropora sp. June 12 (7) NM 555 (arborescent) Turbinaria peltata June 11 (6) 643

*Though not observed, spawning probably occurred in colonies with darkly pigmented eggs

179 colonies of these species (Table 4.2) suggests spawning probably also occurred in July.

Although spawning was not documented in Acropora granulosa, the presence of colonies with either darkly pigmented or white eggs shortly after the June full moon suggests spawning is probably "split" over at least two months for this species as well. The size of developing eggs in Acropora formosa, A. horrida, and A. paniculata when last sampled around June full moon (Table 4.2) suggests maturation and spawning occurred in July, although continued development into August cannot be ruled out. Developing eggs in Acropora divaricata, A. valenciennesi, and A. brueggemanni were not monitored long enough to determine time of maturation and release (Table 4.1).

The degree of intraspecific synchrony varied greatly among species and sites. All examined colonies of Acropora carduus, A. horrida, A. clathrata, and A. granulosa were gravid when first sampled (Table 4.1). In contrast, even at the same study site some colonies of A. florida, A. secale, and A. brueggemanni were gravid while others were not. Some species when sampled from different sites showed similar degrees of reproductive development, while other species showed disparities. For example, ripe colonies of Acropora clathrata were found on both the east and west barrier reef slope (Lighthouse Channel and Ulong Channel area, respectively). By contrast, the fertility of colonies of A.

180 formosa varied between study sites. Either all the colonies of this species examined from a given site were gravid

(fringing reefs off Malakal Island and a site in the Rock

Islands) or none were gravid (e.g., off Ulong Island and the

K-B Bridge) .

YAP

Of the ten species examined, all had colonies with darkly pigmented eggs (Table 4.3). Every colony of every species examined had conspicuous eggs. Two species (A. danai and A. formosa) had some colonies with colored, ripe eggs while other co10nies had immature white eggs. No colony had eggs in both developmental stages. All sampled colonies of the remaining eight species had colored, ripe eggs.

TABLE 4.3. Summary of number of corals sampled and examined by eye for developing eggs on the reef flat adjacent to Guufnuw Channel, Yap, on May 27, 1993, 8 days before the June 4 full moon.

Species Colonies Colonies with Colonies with sampled colored eggs white eggs

Acropora danai 8 6 2 A. formosa 6 3 3 A. valida 6 6 a A. digitifera 3 3 a A. tenuis 3 3 a A. cytherea 6 6 a A. hyacinthus 6 6 a A. florida 6 6 a A. surculosa 2 2 a Montipora digitata 4 4 a

181 DISCUSSION

LUNAR PERIODICITY AND TIDAL REGIME

The six coral species which were observed spawning in

Palau all did so between full moon and lunar last quarter.

These observations are in agreement with studies from other regions, including the Great Barrier Reef, Okinawa, and Guam

(Richmond and Hunter, 1990).

On any given spawning night, the time of spawning occurred roughly midway between low and high tide, a period of maximum incoming tidal current flow. Timing of gamete release to coincide with peak water movement may serve to facilitate mixing of gametes and dispersal of developing embryos.

INTRASPECIFIC SYNCHRONY

Gravid corals sampled in Palau .showed substantial variation in intraspecific synchrony, while those in Yap were well synchronized. The degree of intraspecific synchrony may be characteristic of a particular species. For example, spawning in Acropora danai in Guam typically occurs in two consecutive summer months (C. Hunter, F. Te, pers. comm.). The presence of colonies with either white or pigmented eggs suggests the same is true for this species in

Yap. Similarly, A. formosa in Yap probably demonstrates

"split" spawning, as some colonies had ripe eggs shortly before the June full moon while others contained immature

182 white eggs. Variability in egg size among gravid colonies of

A. formosa in Palau suggests spawning may occur in both July and August.

INTERSPECIFIC SYNCHRONY

Substantial interspecific synchrony has been documented among corals on the Great Barrier Reef and, to a lesser extent, in Okinawa and Guam (Heyward et al., 1987j Richmond and Hunter, 1990j Appendix A-2). The presence of colored eggs in nearly every colony of ten species examined in Yap one week before the June full moon suggests a major spawning event occurred in those waters shortly thereafter. Local residents are aware of the mass synchronous spawning of corals from four to eight nights after the full moon just prior to the wet season (B. Goldman, pers. comm.). In contrast, the pattern in Palau indicated by this study is less synchronous. The gravid corals identified during this study were spread out over at least several months in their spawning periodicity. Spawning was directly observed or inferred in May (Acropora carduus) and June (A. clathrata,

A. secale, A. florida, Acropora spp. [arborescent],

Turbinaria peltata), while other species still had immature eggs developing (A. formosa, A. horrida, A. paniculata, A. granulosa). The size of developing eggs in these species

(Table 4.2) suggested spawning probably occurred in July and possibly in August. Some species clearly spawn over at least

183 two months, e.g. A. florida, A. secale, Acropora sp.

(arborescent) .

Additionally, a number of species which are important components of the shallow reef assemblage, and which participate in synchronized mass spawnings elsewhere, were already devoid of gonads by the start of this study on May 6

(Acropora digitifera, A. humilis, A. tenuis, A. echinata, A. nasuta, A. cytherea and others, see Table 4.1). There is at present no evidence to choose between two alternatives describing their reproductive behavior in Palau: (I) these and other species are well-synchronized, releasing gametes earlier in the year, or (2) these and other species are asynchronous, with spawning spread out over several months.

Only future research will discriminate between these alternatives. Even given the most conservative premise of synchronous spawning in April, the time range of coral reproduction in Palau would be a minimum of four months, from April to July.

The present study lends weight to the trend noted by

Richmond and Hunter (1990) that reproductive seasonality and synchrony diminish at lower latitudes, accompanied by decreased temperature range (Oliver et al., 1988). In Palau the seawater temperature is stable year-round, ranging between 28°C and 31°C, with an average of around 29°C (Hatai,

1937). In this regard it is interesting to note that giant clams (Tridacna and Hippopus spp.) in Palau have been found

184 in spawning condition throughout the year, whereas in more seasonal climates of Australia and Fiji spawning is more likely to occur in the warmer summer months (Heslinga et al., 1990).

BIOGEOGRAPHIC IMPLICATIONS OF SPAWNING SYNCHRONY

The latitudinal diversity gradient describes the attentuation of species richness which is found in progressing from low tropical latitudes to higher temperate and polar latitudes. The evolutionary origin of this pattern has recently been re-explored by Jablonski (1993). Working primarily with echinoderms, he restores to their original palaeolatitudes twenty-six orders considered to have good fossil preservation potential, and then groups them into 10° latitudinal belts based on their first fossil occurrence.

Significantly more first appearances of these orders are found in tropical waters than expected from sampling effort alone. Jablonski concludes that tropical regions have been a major source of evolutionary novelty, rather than refuges which have simply accumulated diversity. The latitudinal diversity gradient of scleractinian corals is characterized by the greatest generic diversity lying somewhat north of the equator in both the Atlantic and Indo-Pacific (Stehli and Wells, 1971). Water temperature and levels of solar energy necessary to sustain symbiotic zooxanthellae are

185 generally considered to be the primary factors which limit corals to tropical and subtropical waters (Veron, 1986).

General patterns in the worldwide distribution of hermatypic corals can be described by longitudinal as well as latitudinal gradients. The longitudinal attenuation of diversity in both terrestrial and marine flora and fauna from west to east across the Pacific is a salient feature of

Pacific basin biogeography (Kay, 1979). For marine organisms, this pattern of eastward attenuation has been described for corals (Wells, 1954), molluscan cones and cowries, asteroid echinoderms, and pomacentrid fish (Kay,

1979). In examining the worldwide distribution of coral genera, Stehli and Wells (1971) noted two separate foci of high diversity, one in the Indonesian region of the Indo­

Pacific and another in the Caribbean Sea of the Atlantic.

The Indo-Pacific center of high coral generic diversity lies within the triangle demarcated by the northern Philippines,

Borneo, and Papua New Guinea, and extends along a line south to the southern Great Barrier Reef (Veron, 1985, 1986).

Coral diversity gradually diminishes eastward across the

Pacific. Approximately 350 hermatypic scleractinian species are recorded from the Philippines, 275 from the Mariana

Islands, 240 from the Marshall Islands, 70 from the Line

Islands, 42 from Hawaii, and 11 from the west coast of

Panama (Richmond, 1990a). In a westerly direction diversity remains high from northwest Australia to the central Red Sea

186 (Veron, 1986). In regards to the pattern of attenuation away from the center, Wells (1954) noted that the same genera seem to drop out in the same sequence regardless of direction away from the diversity center.

A persistent problem in coral biogeography concerns the processes by which this longitudinal pattern has developed and been maintained. Coral biogeographers have long debated to what degree the Indo-Pacific center of high diversity is a "cradle" in which species have arisen (Stehli and Wells,

1971j Potts, 1984aj McManus, 1985) or a "museum" in which species have accumulated (Ladd, 1960j Heck and McCoy, 1978j

Rosen, 1984j Jokiel, 1990b; Wallace et al., 1991). Stehli and Wells (1971) interpreted generic diversity data to suggest that centers of diversity are also evolutionary centers of origin where new forms evolve and from where they spread to more marginal areas. If this model is correct, they predicted that areas of high diversity must exhibit many young genera, while areas of low diversity must show the presence of many old genera. These authors used the time of first known occurrence in the fossil record anywhere in the world to determine the age of each genus. They subsequently noted that young average generic ages were concentrated in the IndO-Pacific area of high diversity, with a progressive increase in average generic age in all directions away from the diversity center.

187 An alternative model proposes that shallow water marine invertebrates which lived among the islands and banks of the

Pacific during and after the Cretaceous could have migrated with the prevailing winds and currents westward toward

Indonesia (Ladd, 1960). These same currents are used by

Stehli and Wells (1971) in support of their own model. These authors note that, except in the Equatorial Counter

Currents, the movement of surface waters within the region most favorable for hermatypic coral growth is principally east to west, which "doubtless" hampers the eastward spread of planktonic coral larvae from their area of origin.

Taking exception to the view that a taxon's place of origin is the region in which it is most diverse, McCoy and

H~ck (1976) note that present centers of diversity may not have always been so. They propose an alternative, vicariance model in which the present-day global distribution and diversity patterns of corals are explained by the existence of widely-distributed ancestral biotas which have since been modified by tectonic events, speciation, and extinction.

They specifically suggest that the older groups of genera in the Indo-Pacific diversity center represent a formerly widespread biota which became established by a series of small range extensions throughout the continuous Tethys

Seaway. They further submit that genera and species are now accumulating in the Indo-Pacific high diversity area, due to

188 its greater shoreline area available for colonization and increased chances for isolation and diversification.

Combining features of the center-of-origin and the vicariance models, McManus (1985) accepts that the differential survival of many families and genera of pan­

Tethyan organisms explains a major component of the diversity of organisms in Southeast Asia, especially at the generic level. However, the proliferation of species in certain coral genera (Acropora and Montipora) , the centrality of Southeast Asia within the ranges of many species, and high regional species endemism are used as evidence to argue for enhanced spe~iation processes within

Southeast Asian seas. Combining inference from contemporary terrestrial biogeographic data with direct geological indications, McManus posits cycles of land bridge construction and destruction and sea level changes to account for the alternate appearance and disappearance of isolated sea basins in Southeast Asia. In this view, a shifting network of land bridges occurring between the mid­

Miocene and Pleistocene resulted in a mosaic of isolated ocean basins. Allopatric speciation thereupon produced heightened levels of diversity superimposed on the ancient

Tethyan pattern. McManus rejects the possibility of allopatric speciation under present conditions in Southeast

Asia, assuming that isolation of any single population for tens or hundreds of generations is necessary for speciation

189 to occur. In support of this idea McManus states that 1) reefs off high islands of the Indonesian and Philippine archipelagoes are sufficiently common to provide continuity between coral populations, 2) most currents which interconnect insular regions change direction between the northeast and southwest monsoon seasons, and 3) typhoons and tectonic events reduce the possibility of isolation.

McManus notes that the effect of isolation on speciation rates would vary between groups. Organisms such as corals with long generation times would be less likely to speciate during short periods of isolation such as might have occurred during Pleistocene sea-level changes (Potts,

1984ai see Chapter 3 for a discussion of this model). Potts

(1983) also championed the ideas that the most likely sites for coral speciation were the shallow continental shelves from Southeast Asia to northern Australia, and that oceanic coral faunas in the Pacific have been maintained by repeated, long distance dispersal from the central continental shelves of the western Pacific.

Rosen (1984) proposes a different model to account for the center of high coral diversity, one in which the net effect is species accumulation in the refuge of Australasian continental shelves. He maintains that even during low sea levels along continental margins, isolation and speciation cannot occur on shallow continental shelves because numerous local populations will always be interconnected by gene

190 flow. Instead, speciation during high sea levels is envisioned along the archipelagoes of the western Pacific from the Philippines to Fiji. During periods of high sea level stand, the distance between .emergent high islands is maximized, exceeding the dispersal potential of coral larvae. Allopatric speciation is facilitated by the resultant vicariant isolation of populations during high sea levels. During lowered sea levels, new species disperse westward with the prevailing currents, becoming at least temporarily established on shallow banks and shoals which have become newly available habitat for corals. Species eventually accumulate on the continental shelves of

Australasia where their populations are stabilized by localized gene flow, but they become extinct on the small island reefs on which they originated.

Models of Indo-West Pacific biogeography typically invoke the capacity for long-distance dispersal by coral planula larvae. Under laboratory conditions, Harrison et al.

(1984) noted a survival period of 91 days for a planula of

Acropora hyacinthus, and Richmond (1987) found that planulae of Pocillopora damicornis remained competent (able to successfully settle and metamorphose) for over 100 days.

These periods of time are sufficient to allow dispersal over large distances (Richmond, 1987, 1988). In plankton tows at

100 stations throughout the "east Pacific barrier", an area of tropical sea between the Central and East Pacific devoid

191 of oceanic islands, Scheltema (1988) found coelenterate larvae at 66 stations. Among these were planula larvae

(Scheltema, 1988), several of which were induced to settle and were identified as pocilloporids from their calices

(Richmond, pers. comm.). Jokiel (1990a) documents rafting of adult coral colonies on volcanic pumice and other floating debris. Coral larvae can settle on drifting material such as pumice, and rafted colonies may drift for years before establishing new populations in distant areas. Dispersal potential is therefore not limited by longevity and competency of planktonic larvae. Calculations of massive transport of corals by rafting on drift pumice into the

Great Barrier Reef from regions of lower diversity lying to the east and southeast support the hypothesis that centers of diversity may be sites of species accumulation rather than sites of species origin (Jokiel, 1990b). Working with the family Fungiidae, Hoeksema (1989) concludes that Indo­

West Pacific distribution patterns may be explained by a combination of dispersal, vicariance and refugiality.

Due to problems with species-level taxonomy and distribution records, coral biogeographic hypotheses have until recently been based exclusively upon generic data.

Using both quantitative and qualitative morphometric characters, Wallace et al. (1991) construct a phylogenetic tree for the five species in the Acropora selago group. When this tree is compared with biogeographical distributional

192 data, a pattern appears in which species displaying the most derived character states are most restricted in their distribution. These authors take this pattern as evidence against the center-of-origin hypothesis proposed in Stehli

and Wells (1971).

Considerations of the degree of spawning synchrony and

the opportunities for hybridization suggest a new approach

to this continuing problem. Should hybridization and

polyploidy serve as mechanisms of speciation (see Chapters 2

and 3), then the degree to which coral spawning is

synchronized among species influences the opportunities for

hybridization, and therefore the biogeography of speciation.

Hybridization is in part a function of spawning synchrony.

Geographic regions with high species diversity and

multispecies spawning, such as the Great Barrier Reef,

provide more opportunities for hybridization than regions

with lower species diversity and greater asynchrony of

spawning, such as the Red Sea (Shlesinger and Loya, 1985) or

Hawaii (Heyward, 1985a; Hunter, 1989; Kenyon, 1992). Aside

from two sites on the northern coast of Papua New Guinea

(Oliver et al., 1988), no published data exist regarding

temporal patterns of coral reproduction in the high

diversity triangle demarcated by Indonesia, Borneo, and

Papua New Guinea. The total of 69 genera presently

documented in Palau (Maragos, 1991) is comparable to the

generic totals for the Great Barrier Reef, and places Palau

193 within the high coral biodiversity center of the Indo­

Pacific. To the extent that hybridization may serve as a mechanism for speciation, the trend towards spawning asynchony in both Palau and northern Papua New Guinea supports the view that the center of diversity reflects the process of species accumulation more than the process of species origin.

By contrast, the most recent species inventory in Yap

(Orcutt et al., 1989) documents 102 fewer species and 12 fewer genera than are present in Palau. Generic data place

Yap outside the 70-genera-contour by which the high diversity center is delimited (Veron, 1985). Despite the fewer number of species in Yap, the high degree of reproductive synchrony documented by this study and corroborated by local residents implies there are more opportunities for hybridization. Coral populations in equatorial latitudes, as exemplified by Papua New Guinea and

Palau, are less likely to function as sites of hybrid-driven speciation than those in higher latitude regions such as

Yap, Guam (Appendix A-2), and Australia, which are characterized by synchronized interspecific spawning.

In this light it is interesting to consider the rate of coral evolution and degree of spawning synchrony in the

Atlantic. Stehli and Wells (1971) note that the oldest average generic age observed in the Indo-Pacific (44 million years) is younger than the youngest average generic age

194 encountered in the Atlantic (49 million years). Furthermore, the youngest average generic age encountered in the Indo­

Pacific (23-24 million years) is about 50~ that of the youngest in the Atlantic. These authors conclude that the evolution of hermatypic coral genera has been proceeding about twice as fast in the Indo-Pacific as in the Atlantic.

Although rates of evolution are contingent upon many complex and interacting factors, a pattern of more rapid evolution in the Pacific, characterized by synchronous spawning and enhanced opportunities for hybridization, is consistent with what is known of the mode and temporal patterns of reproduction in the Atlantic. Broadcast spawning is the primary mode of reproduction in the Pacific region and Red

Sea, whereas the present data suggest brooding may be the chief mode of reproduction in the Caribbean (Richmond and

Hunter, 1990). This trend is exemplified by the fact that

Caribbean Porites tend to be brooders while Pacific Porites broadcast spawn gametes (Szmant, 1986). Internal fertilization would rely less upon chance events of gamete contact than would external fertilization resulting from the mixture of gametes that occurs during broadcast spawning

(Oliver and Willis, 1987). Also, brooding is usually associated with lower fecundity than spawning, and produces

fewer, larger larvae (Fadlallah, 1983). The brooding mode is therefore less conducive to interspecific hybridization than

is the spawning mode. Among Caribbean genera which spawn

195 gametes, including Acropora, Montastrea, and Diploria, reproductive maturity occurs during the late summer but spawning does not appear to be synchronized between species

(Szmant, 1986; Soong, 1991). The two Atlantic species of

Acropora (A. palmata and A. cervicornis) appear to recruit primarily from asexually-produced fragments rather than sexually-derived larvae (Bak and Engel, 1979; Rylaarsdam,

1983). Predominance of the brooding mode, relaxed temporal patterns of spawning, and recruitment into the population by asexual propagules would all contribute to lower potential levels of hybridization and slower rates of hybrid-driven evolution in the Atlantic relative to the Pacific.

196 CHAPTER 5

SEXUAL REPRODUCTION IN HAWAIIAN ACROPORA

INTRODUCTION

In Hawaii, Acropora first appears in the fossil record in the Miocene but disappears during the Pleistocene (Wells, in Grigg, 1981). The existence of Acropora in Hawaii during the Recent had long been doubted until the late 1970's, when three species were reported from the Northwest Hawaiian

Islands (Grigg et al., 1981; Figure 5.1). Unlike other

Hawaiian corals which are found throughout the Archipelago from Hawaii to Kure (Grigg, 1983), Acropora has a restricted distribution. Acropora cytherea occurs from Kauai to Laysan, while Acropora valida and Acropora humilis have been found only at French Frigate Shoals and Maro Reef. All three species are most abundant at French Frigate Shoals (Grigg et al., 1981; Figure 5.2).

Previous limited sampling in June, September, and

November gave no evidence of sexual maturity in any of the three Acropora species (Grigg et al., 1981). It was hypothesized that larval recruits originate from Johnston

Atoll, 720 km to the southwest. The absence of Acropora from other islands of the Hawaiian chain had been attributed in part to lack of sexual reproduction by these corals in

Hawaii (Hourigan and Reese, 1987).

197 0 1790 1750 170 1650 1600 re I. N ;;-) .....Midway I. "J ':~~I Peart and Hermes Reel

G;\ Laysan I. Usianski C- ::.:~:. :~ ~ .,~O ~r= I \":,. \Gardnl!/' Pinnacles II co \) French Frigate Shoals I~ker I. NORTHWESTERN ~);~) HAWAllAN ISLANDS

Oahu r"", Molakai miles ·.a,.~ , {~.:~" ~S.~Moui (anCl' "-V"" o 400 -200 MAIN

OCEAN I I HAWAIIAN ISLA'NDS PACIFIC ~ -- 1 0 I ~~I I I ~1 _ 18 Figure 5.1. The Hawaiian Archipelago Figure 5.2. Distribution of Acropora valida, A. cytherea, and A. humilis in Hawaii. Adapted from Grigg et al. (1981).

I-' ~ ~ Distribution and Abundance of Acropora in the Hawaiian Archipelago

10000 =f""------=------:------" French # Frigate _ A. humilis Shoals c ~ A. valida o 1000 I o A. cytherea o n .. i abundant Maro e 100 t tv s o common o p '+ e r 10 ....j uncommon t Layson 5 rare a Kauai Nihoa a t a 2 M 0 m MK 0.1 ' I, I ,I ,'! ," 20 21 22 23 24 25 26 27 28 29 Latitude N Source: Grigg et al.,1981 Numerous studies undertaken since the discovery of

Acropora in the Northwest Hawaiian Islands more than a decade ago indicate the dominant sexual pattern in tropical

Pacific corals is broadcast spawning of gametes during late spring or summer (Richmond and Hunter, 1990). All species of the genus Acropora (subgenus Acropora) studied to date are simultaneous hermaphrodites, with an annual gametogenic cycle in which oocytes develop over a period of about nine months and testes over about ten weeks (Wallace, 1985;

Babcock et al., 1986; Szmant, 1986; Soong, 1991). Several weeks prior to spawning, oocytes of many coral species become colored (Harrison et al., 1984; Babcock et al., 1986;

Heyward et al., 1987; Heyward, 1988). In this Chapter I present evidence showing that Acropora valida, A. cytherea, and A. humilis are sexually reproductive at French Frigate

Shoals.

MATERIALS AND METHODS

All corals were collected from French Frigate Shoals

{24°45'N, 166OW) , an atoll with a small vestige of emergent basalt (La Perouse Pinnacle), located 830 km northwest of

Honolulu (Figure 5.1). All three species were sampled from

7-9 m depth around La Perouse Pinnacle at roughly biweekly intervals from late May through July in 1989. Full moons occurred on May 20, June 18, and July 18. In 1990, Acropora valida and A. cytherea were sampled on July 11 and August

201 12. Full moons occurred on July 7 and August 6. With the exception of A. humilis, 6 to 12 colonies of each species were sampled on each collection date. Sampled colonies measured> 20 cm in at least one dimension. Efforts were made to sample from widely scattered colonies to minimize collections from clonemates (Hunter, 1988). Additional samples of Acropora valida were collected from the shallow

«1m) reef flat off Tern Island, 12 km to the north of La

Perouse Pinnacle. Samples from 5-15 colonies were collected at roughly weekly intervals between May 26 and July 16 in

1989, and several times in July 1990. Branch pieces 7-12 cm in length were fixed in 10% formalin-seawater. Several unfixed fragments were saved for microscope examination.

Fixed specimens were decalcified in 4% formic acid­ seawater. For each sample piece, 5 fertile polyps from a 1 cm length of tissue behind the sterile growing tip were randomly chosen for dissection (Wallace, 1985). Maximum and median diameter of each oocyte were measured, and oocyte size was computed as the geometric mean. Comparison of sizes of fixed and fresh oocytes obtained from the same colony indicated fixation and storage in formalin for several months had no significant effect on oocyte size (at p<0.05,

Student's t test for 6 colonies). The length of each individual testis in each polyp was also recorded.

Histological sections (7-9 ~m thick) of polyps, prepared by

202 standard techniques and stained with hematoxylin and eosin, were examined to determine maturity of gonads.

To examine the distribution of size classes of A. valida, 100 colonies on the reef flat were haphazardly chosen and measured to the nearest centimeter across 2 colony widths at right angles. Colony size was computed as the geometric mean. At La Perouse, 30 colonies were similarly chosen and measured. To measure the distribution of size classes of A. cytherea, 34 colonies were haphazardly chosen at La Perouse. Their dimensions were measured and sizes were computed as for A. valida.

RESULTS

All three species were simultaneous hermaphrodites with the same arrangement of gonads in fertile polyps (Figure

5.3). Strings of oocytes developed within the two lateral pairs of mesenteries, while testes developed within the ventral and dorsal directive mesenteries. Testes developing in the ventral directive mesenteries were longer than those in the dorsal directive mesenteries. The dorsal testes were absent from some polyps. All fertile samples had a sterile tip 2-5 cm in length in which fertile polyps were not found.

A~OPOAA ~LIDA

Colonies from both sites contained white oocytes and light brown testes when first sampled in late May, 1989.

203 Figure 5.3. Dissections showing mature gonads in Acropora species at French Frigate Shoals. A Acropora valida B Acropora cytherea C Acropora humilis. 00, oocyte; LT, long testes; ST, short testes. Scale bars represent 500 ~m.

N o >Po 205 Orange-pink pigmentation was first noted in oocytes of La

Perouse samples on June 14 and in oocytes of reef flat samples on June 16. In 1990, all colonies from both sites had orange-pink oocytes and light brown testes when first sampled in early July. The colors gradually deepened with time.

Spawning was inferred to have occurred in reef flat colonies in 1989 between July 1 and July 9, on or between new moon (July 2) and first quarter (July 10) (Table 5.1).

In colonies at La Perouse in 1989, spawning occurred between

July 11 and July 27, a possible range of 3-26 days after those from the reef flat. The lunar phase of spawning at La

Perouse cannot be determined with more precision than the period between first quarter (July 10) and last quarter

(July 25) due to the long interval between samples. Time of spawning at either site in 1990 is not known because later summer samples, devoid of fertile polyps, were not obtained.

However, comparable mean sizes of oocytes were found at both sites one month later in 1990 than in 1989 (e.g., for reef flat samples, mean oocyte size is 667 ~m on July I, 1989 and

669 ~m on July 31, 1990) (Table 5.1). At both sites in 1989, ambient water temperature on the last sample date before spawning was 29.50C.

Of 165 samples collected from both sites during 1989 and 1990 before spawning, only two samples lacked fertile polyps, and seven samples had such a low number of fertile

206 TABLE 5.1. Summary of dissections of fixed polyps of Acropora valida. Only latter sampling periods are shown for each year.

No. Colonies % Fertile Mean Size Max. Size Mean No. Max. Testis Date Sampled Colonies of Oocytes of Oocytes Oocytesl Length (jlm) (jlm) Polyp (jlm)

Reef flat

1989 July 1 12 100 667 806 6.4 2500

July 9 15 7 605 725 4.6 2000 July 16 12 0 tv 0 1990 -...J July 31 8 100 669 812 6.5 2000

La Perouse

1989 July 11 12 100 642 812 5.8 2900

July 27 15 0

1990 August 12 7 86 635 793 7.6 1750 polyps that five polyps could not be found in the entire sample for dissection. Polyp fecundity (number of oocytes/polyp) ranged from 3-15, with a mean of 7.3 (n=780 polyps). The mean number of polyps cm-2 was 26.0 (counted on a 1 em x 1 em square of decalcified tissue, mean of 2 counts per sample, n=147 samples). Although the proportion of polyps which were fertile was not quantified for each sample piece, generally about half of all polyps behind the sterile tip were fertile. Using an estimate of 13 fertile polyps cm-2 , a mean fecundity of 94.9 oocytes cm-2 was calculated.

On any given sample day, there was significant variability among colonies in the mean number of oocytes/polyp (Model II one-way ANOVAs, p<0.05), an index of colony fecundity. Although for each of the four site/year combinations there was significant variability among sample days in mean number of oocytes/polyp (one-way ANOVAs, p<0.05), there was no trend between the number of oocytes/polyp and time, indicating there was neither degeneration nor sequential release of oocytes over the time periods studied. On any given sample day, there was no significant association (Pearson R, p<0.05) between the mean size of oocytes and the number of oocytes produced by a colony.

Latter stages of oocyte growth from both the shallow reef flat and La Perouse Pinnacle sites in 1989 and 1990

(Figure 5.4) are well described by linear models of oocyte

208 Figure 5.4. Oocyte growth in Acropora valida at French Frigate Shoals. Each point represents a colony mean oocyte size calculated from 5 fixed, dissected polyps. Only the slopes of the lines describing La Perouse 1989 and reef flat 1990 data are not significantly different. tv o 1.0 j'

700 I 8 • ~ I 650 f-

~ .~ E ~ :::J 600 ~~ "--..-/ A Q) N . - 550 L ~v (f) - • • N ...... Q) 0 -+-' .>, u 0 0 Reef flat 1989 0 500 I 0 ~~ 0 La Perouse 1989 •/':;, Reef flat 1990 450 L F'I - ... La Perouse 1990

400 I \..1 I I I I May June July August size vs. time of year (r2 = 0.85, 0.81, 0.85, and 0.78 for

1989 reef flat, 1989 La Perouse, 1990 reef flat, and 1990 La

Perouse data, respectively.) There were significant differences (p

La Perouse and 1990 reef flat lines.

The maximum individual testis length was 3200 ~m. For any given sample day, there was significant variability in testes size among colonies (Model II one-way ANOVAs, p

The maturity of gonads before spawning was confirmed by histological sections of reef flat samples collected July I,

1989 (Figure 5.5). The oocyte germinal vesicle had an asymmetric position adjacent to the cell membrane. The periphery of the oocytes stained blue with hematoxylin, due to the presence of cortical vesicles just beneath the cell membrane (Szmant-Froelich et al., 1980). Spermatozoa had condensed heads about 1 ~m in diameter, with "bouquets"

211 'l:... . • l· ~ ,-

C.:?

.100- .... '" '17. •

Figure 5.5. Histology of Acropora valida polyps in the week prior to spawning. A cross section B mature testis with bouquets of sperm tails. 00, oocytej TS, testisj GV, germinal vesiclej NU, nucleolusj TL, sperm tails. Scale unit = /lm

212 (Wyers, 1985) of sperm tails. Tails were absent from testicular sections of colonies sampled one week earlier.

The distributions of colony sizes from the reef flat and La Perouse (Figure 5.6) are not significantly different

(Kolmogorov-Smirnov two-sample test and Mann-Whitney-U test). However, the proportion of reef flat colonies in small size classes «9.5 cm) is substantially greater than the proportion at La Perouse (46% vs. 23.3%, respectively).

ACROPORA CYTHEREA

Oocytes and testes attained sizes which characterize mature conspecifics in Okinawa, O"apan (Appendix A-2) .

However, no clear picture emerges from the available data concerning the proportion of fertile colonies in the population or the degree of synchronization in spawning

(Table 5.2). Because taking samples from deep within a colony was avoided due to unwarranted breakage of the skeleton, it is possible that infertile samples were from sterile zones of peripheral growth (Wallace, 1985).

ACROPORA HUMILIS

This species is reported to be rare in the Northwest

Hawaiian Islands (Grigg et al., 1981). Only one colony was found during this study, at La Perouse Pinnacle. When first sampled on June 6, 1989, this colony contained conspicuous pink oocytes and light brown testes. Samples collected June

213 Figure 5.6. Colony sizes of Acropora valida on the reef flat (n=100) and at La Perouse Pinnacle (n=30).

N I-'..,.

TABLE 5.2. Summary of dissections of fixed polyps of Acropora cytherea.

No. Colonies % Fertile Mean Size Max. Size Mean No. Max. Testis Date Sampled Colonies of Oocytes of Oocytes Oocytesj Length ( /lm) (/lm) Polyp (/lm)

1989

May 30 4 75 481 704 6.1 1500

June 1 1 100 607 707 5.5 2750

June 3 5 0

June 6 7 43 546 743 3.8 2500

N f-' June 14 7 43 637 775 4.8 1350 0'\ July 27 8 25 571 778 2.3 750 1990

July 11 8 50 633 774 5.7 1500

August 12 8 12.5 622 765 3.3 No testes 14 and thereafter lacked fertile polyps. Mean oocyte size

from 10 dissected polyps was 714 ~m (range 529 - 894), with

a mean of 7.3 oocytes/polyp. Maximum testis length was 1850

~m. Behind the sterile zone of the branch tips, almost all polyps were fertile. Spawning was inferred to have occurred between June 6 and June 14, thus centering around lunar

first quarter, which fell on June 10. Ambient water

temperature on June 6 was 29.5°C.

DISCUSSION

REPRODUCTIVE MATURITY AND DEGREE OF INTRASPECIFIC SYNCHRONY

Observations in this study regarding size, color,

fecundity and histological appearance of gonads are in

agreement with observations from other regions where

spawning has been directly observed in the field or aquaria

(Table 5.3). Testes up to 3280 ~m in length are reported for

Acropora valida from the Great Barrier Reef (Wallace and

Oliver, in Grigg et al. 1981). Wallace (1985) calculated a

mean fecundity of 96 oocytes cm-2 in A. valida colonies

containing mature gonads.

217 TABLE 5.3. Comparative egg size and color for Acropora valida, A. cytherea, and A. humilis. Measurements are from freshly-spawned eggs.

Species Area Mean Color Reference Diameter (gm)

Acropora GBR 633 Wallace, 1985 valida GBR orange-red Babcock et al.1986 Guam 650 pink C.Hunter,pers.comm. Guam orange-red Heyward, 1988 Guam 598 orange-pink J.Kenyon, Chpt.3

Acropora GBR 556 pink J.Kenyon, Chpt.3 cytherea Okinawa 547 pink J.Kenyon, Chpt.3

Acropora GBR 467 pink-orange J.Kenyon, Chpt.3 humilis GBR pink Babcock et al.1986 Guam 700 dark pink C.Hunter,pers.comm.

In both years, colonies of A. valida from the two sites showed significant differences in oocyte growth rates

(Figure 5.4), with slower growth at the deeper La Perouse site than on the shallow reef flat. In 1989, spawning of colonies at La Perouse occurred several days to possibly as much as several weeks after those on the reef flat. In 1990, the slower growth of oocytes at La Perouse suggest spawning again was delayed relative to its occurrence on the reef flat. As gamete viability is greatly reduced after seven hours (Heyward and Babcock, 1986), populations at the two sites may be temporally reproductively isolated. At both sites, gonads attained mature sizes roughly one month later

218 in 1990 than in 1989 (Table 5.1, Figure 5.4). In other regions where Acropora spawning has been inferred through sequential samples or directly observed in two successive years, a similar one-month shift has been documented

(Harrison et al., 1984j Willis et al., 1985j Babcock et al.,

1986j pers. obs.).

The maximum length of testes fixed one day prior to spawning in Acropora cytherea from Okinawa was 1750 ~m

(pers. obs.). Gonads in samples of A. cytherea from La

Perouse Pinnacle thus were of mature size. However, the proportion of fertile colonies within the population, degree of synchronization among colonies, and time of spawning remain in need of clarification. Acropora cytherea spawning has been documented one or two nights after the late spring full moon in Okinawa (Heyward et al., 1987j Richmond and

Hunter, 1990j pers. obs.). At Johnston Atoll, Acropora cytherea spawning in 1988 was observed in early June, 5-6 nights after the May 31 full moon (R. Kosaki, pers. comm.)

Sampling of A. cytherea in this study may therefore have been initiated after a major spawning event in late spring, leaving many colonies devoid of gametes. Intra- and/or inter-colony asynchrony may separate the population into groups which" spawn over several months, as in Hawaiian

Montipora dilatata and Montipora verrucosa (Hodgson, 1989j

Hunter, 1989).

219 LUNAR PERIODICITY

Lunar phase of spawning in A. valida and A. humilis differed in this study from that reported for these species from other regions. In 1989, spawning of reef flat colonies of A. valida occurred on or between new moon and first quarter. On the Great Barrier Reef, Guam, and Okinawa, spawning has been documented between full moon and last quarter (Richmond and Hunter 1990; pers. obs.). For A. humilis, lunar period of spawning at La Perouse centered around first quarter. On the Great Barrier Reef, Guam, and in the Red Sea, spawning has been reported between full moon and last quarter (Willis et al., 1985; Babcock et al., 1986;

Richmond and Hunter 1990; Shlesinger and Loya, 1985; pers. obs.) .

INTERSPECIFIC ASYNCHRONY

Along with other genera on the Great Barrier Reef, species of Acropora participate in simultaneous multispecies spawnings, during which gametes are released on one or several successive nights of the year in late spring

(Harrison et al., 1984; Willis et al., 1985; Babcock et al.,

1986). Synchrony of spawning among species of Acropora is less complete in Guam (Richmond and Hunter, 1990; Appendix

A-2) and Okinawa (Heyward et al., 1987; Richmond and Hunter,

1990; Appendix A-2) yet overlap in spawning time occurs in each locality. In the northern Red Sea, species of Acropora

220 do not overlap in spawning time (Shlesinger and Loya, 1985).

Oliver et al. (1988) noted that the relationship between mass spawning and high temperature and tidal range on the

Great Barrier Reef did not apply to all other geographic locations. At French Frigate Shoals, tidal range «1 m) is smaller but temperature range (6.50C) is comparable to several stations along the Great Barrier Reef. The three species of Hawaiian Acropora appear to lack temporal overlap in spawning. Hawaiian coral species in general do not exhibit mass spawnings such as those on the Great Barrier

Reef (Heyward, 1985a; Hodgson, 1989; Hunter, 1989; Richmond and Hunter, 1990).

The ability of some species of Acropora to form interspecific hybrids during multispecies spawning has been demonstrated by recent experimental work (Willis et al.,

1992; R. Richmond, pers. comm.; Chapter 2). Acropora valida,

A. cytherea, and A. humilis participate in multispecies spawning events on the Great Barrier Reef, in Guam, and in

Okinawa (Richmond and Hunter, 1990). Of the three species present in the Northwest Hawaiian Islands, A. cytherea has formed hybrids under experimental conditions in Australia with Acropora pulchra (Chapter 2). At French Frigate

Shoals, the asynchrony of spawning demonstrated by the three

Acropora species precludes the possibility of naturally forming intrageneric hybrids.

221 SEXUAL VS. ASEXUAL REPRODUCTION

Questions concerning the contribution of sexual vs. asexual reproduction to the origin of individual colonies remain open for many coral species. In a study of eight species of Acropora on the Great Barrier Reef, Wallace

(1985) found four species primarily recruited from larvae, three species primarily recruited from fragments, and one species used both modes. Several other workers have established the importance of fragmentation in the recruitment of new Acropora colonies (Randall, 1973;

Birkeland et al., 1979; Highsmith, 1982; Bothwell, 1981). In this study, fragments of A. valida were routinely found on the surf-washed reef flat, but were uncommon in the deeper waters around La Perouse Pinnacle. The paucity of fragments at La Perouse suggests that its histogram of colony sizes

(Figure 5.6) reflects the result of larval recruitment rather than fragmentation determining the origin of individual colonies. On the reef flat, by contrast, the greater proportion of colonies in smaller size classes might be explained by fragmented colonies and those which are re­ growing from fragments. If so, then the continuity of the La

Perouse histogram for A. valida suggests that recruitment is a regular rather than a sporadic process.

Colonies of A. cytherea at La Perouse display a continuum of growth forms, from encrusting to vasiform to tabular. Encrusting forms, which measured as much as a meter

222 in diameter, appear to spread out and occupy available substrate before growing into a vertical pedestal, from which further radiating growth results in a tabular plate.

Encrusting growths conform to the relief of the substrate, and do not appear to arise from the recementing of fragments. The shape of the histogram for A. cytherea table colony sizes (Figure 5.7) suggests recruitment is a regular process, with decreased growth or survivorship as the tabular part of the colony approaches one meter in diameter.

Moreover, the occurrence of colonies of A. valida and A. cytherea on steeply-sloping or vertical surfaces, where asexually-derived fragments could not have landed and become re-attached, indicates larval recruits do contribute to the establishment of new colonies. The premise that larval recruits presently derive primarily from local populations is more plausible than an origin from a more distant source such as Johnston Atoll.

Previous suggestions of the sexual sterility of

Hawaiian Acropora populations (Grigg et al., 1981) have been invoked to support the idea that coral populations living at the extremes of their physiological limits recruit primarily by asexual means (Richmond, 1990b; Richmond and Hunter,

1990). The occurrence of sterile populations of other marine invertebrates has been described (Mileikovsky, 1971) and ascribed to peripheral environmental conditions. The absence of either mature gonads or planulae in Panamanian

223 Acropora cytherea 12

N u 10 m Perouse b e r 8

0 f 6 C 0 I 4 0 n i e 2 s

0 0 100 200 Colony Diameter (em)

Figure 5.7. Colony sizes of Acropora cytherea at La Perouse Pinnacle (n=34).

224 populations of Pocillopora damicornis during a two-year study (Richmond, 1985) additionally supported this supposition for corals. More recent studies (Glynn et al.,

1991) suggest eastern Pacific populations of P. damicornis are hermaphroditic spawners, with low but discernable rates of larval recruitment. Though recruitment through asexual means clearly operates in both east Pacific populations of

P. damicornis and Hawaiian Acropora, the idea that these peripheral populations do not produce reproductive propagules is not supported by results of recent research.

POST-PLEISTOCENE COLONIZATION

Grigg (1981) hypothesized that, after its disappearance in the Pleistocene, Acropora recolonized Hawaii from

Johnston Atoll by way of the Subtropical Countercurrent. The demonstrated survival period of 91 days for a planula of

Acropora hyacinthus (Harrison et al., 1984) is consistent with this hypothesis. An alternative possibility is suggested by the observation that the primary affinity of the inshore Hawaiian fish fauna is with the Ryukyu Islands and southern Japan (Hourigan and Reese, 1987). Although the transport time of two current drogues which reached the

Northwest Hawaiian Islands from Japan was more than a year

(McNally et al., 1983), rafting of coral colonies (Jokiel,

1990a) provides a mechanism by which such distances might be bridged. All three species of Acropora presently found in

225 Hawaii are also reported from the Ryukyu Islands (Heyward,

1987j Nishihira et al., 1987j Sakai and Yamazato, 1987).

The contemporary absence of Acropora from the main

Hawaiian Islands becomes more problematic in the light of new evidence for sexual reproductive maturity. Of the three species, only A. cytherea has been found around the most northwesterly of the main Hawaiian islands (Kauai), where its occurrence is rare (Grigg et al., 1981j Figure 5.2).

Most habitats where Acropora is found in the Northwest

Hawaiian Islands are found in the main islands as well

(Grigg, 1981j pers. obs.). Unfavorable currents for larval dispersal in late spring through summer would seem the most likely factor limiting their spread. According to an oceanographic model developed by Seckel (1962) and modified by MacDonald (1984), the Hawaiian Archipelago is bathed predominantly by the North Pacific water type, a water mass which can be envisioned as an ellipse that seasonally oscillates up and down the island chain. The ellipse is displaced southward during the winter and northward during the summer in response to the intensity and position of the

California Current Extension along its southern boundary.

According to this scheme, an ellipse of warm water encompasses the Northwest Hawaiian Islands from April-May through September-October, its margins varying from year to year but generally attenuating around Necker Island (Figure

5.1). If this model is correct, unsettled larvae produced by

226 coral gametes spawned in late spring through summer would remain entrained within this water mass until its oscillation southward to the main Hawaiian Islands in

October/ a time period beyond the longevity of most larvae.

Doppler current profile transects in early November 1990 indicated instantaneous currents running east to west in the main Hawaiian Islands but west to east in the Northwest

Hawaiian Islands/ with an apparent boundary at Necker Island

(J. Polovina/ pers. comm.). Other current measurements in waters of the Northwest Hawaiian Islands in March-June 1969 showed the presence of a northerly drift averaging 4.6 cm/sec (Patzert et al., 1970). Acropora larvae produced in late spring or early summer would not be transported in the direction of the main Hawaiian Islands under the regime of these currents.

227 APPENDIX A-I. Treatment parameters for experimentally­ generated embryos. Variables include the age of the embryo culture when exposed to colchicine, concentration of colchicine, duration of treatment with colchicine, concentration of hypotonic solution, and duration of treatment with hypotonic solution.

Embryo Spawn Embryo Age [Colchicine]/ [Hypo] / Culture Date (hr' min II ) duration duration (min) (min)

INTRASb'E.~IFIC CROSSES Acropora danai 8/5/90 8'45 11 0.05%/45 80:20/20

8/5/90 8'45 11 0.05%/90 80:20/20

8/9/90 8'15 11 0.02%/60 80:20/20

8/9/90 8'15 11 0.02%/90 80:20/20

8/9/90 8'45 11 0.02%/30 80:20/20

7/23/92 10' 0.02%/180 65:35/20

7/23/92 10'15" 0.0275%/180 65:35/20

A. valida 8/5/90 10-12' 0.05%/45 80:20/20

8/5/90 10-12' 0.05%/90 80:20/20

8/12/90 10'45" 0.02%/30 80:20/20

8/12/90 10'45" 0.02%/90 80:20/20 A.digitifera (Okinawa) 5/29/91 4'45"-5'30" 0.02%/30 70:30/15

5/29/91 6'45"-7'30" 0.02%/45 70:30/20

5/29/91 6'45"-7'30" 0.02%/135 65:35/20

5/29/91 10'-10'45" 0.015%/90 65:35/20 A.digitifera (Guam) 7/24/92 10' 0.02%/180 65:35/20

7/24/92 10'30 11 0.02%/180 65:35/20

228 APPENDIX A-1. (Continued) Treatment parameters for experimentally-generated embryos.

Embryo Spawn Embryo Age [Colchicine]/ [Hypo] / Culture Date (hr'min") duration duration (min) (min)

INTRASPECIFIC CROSSES

A.gemmifera 7/20/92 10'-10'30" 0.02%/180 65:35/20

A.nobilis 5/29/91 6'45"-7'30" 0.02%/45 70:30/20

5/29/91 6'45"-7'30" 0.02%/135 65:35/20

5/29/91 10'-10'45" 0.015%/90 65:35/20

A. divaricata 5/29/91 6'45"-7'30" 0.02%/45 70:30/20

5/29/91 6'45"-7'30" 0.02%/135 65:35/20

5/29/91 10'-10'45" 0.015/90 65:35/20

A.nasuta 5/29/91 6'30" 0.02%/45 70:30/20

5/29/91 6'30" 0.02%/120 65:35/20

A.monticulosa 5/29/91 6'30" 0.02%/45 70:30/20

A.surculosa 7/03/91 7'-8' 0.015%/120 65:35/25

7/03/91 8'30" 0.015%/110 65:35/20

7/21/92 10'-11' 0.02%/180 65:35/20

A.ocellata 7/04/91 8' 0.015%/120 65:35/25

7/04/91 8' 0.015%/120 80:20/20

7/05/91 7'30"-8' 0.015%/120 65:35/20

7/05/91 7'30"-8' 0.015%/120 65:35/60

7/06/91 8'-9' 0.015%/120 65:35/20

7/21/92 10' 0.02%/180 65:35/20

229 APPENDIX A-1. (Continued) Treatment parameters for experimentally- generated embryos.

Embryo Spawn Embryo Age [Colchicine]/ [Hypo] / Culture Date (hr' mi n ") duration duration (min) (min)

INTRASPECIFIC CROSSES

A.lutkeni 7/05/91 8' 0.015%/120 65:35/20

7/06/91 6'30 11 0.015%/120 65:35/20

7/06/91 6'30" 0.015%/120 65:35/60

A. squarrosa 7/06/91 8'30 11 0.015%/120 65:35/20

A.robusta 7/20/92 10' 0.02%/165 65:35/20

11 7/23/92 9'15 - 0.0275%/180 65:35/20 10'15 11

A.samoensis 10/15/92 10'45" 0.02%/135 65:35/20

A. tenuis 10/16/92 10'30"- 0.02%/135 65:35/20 11'30"

11 A.pulchra 10/16/92 10'15 - 0.02%/135 65:35/20 10'45 11

A.formosa 10/16/92 10'30 11 0.02%/135 65:35/20

A.millepora 10/16/92 10'30 11 0.02%/135 65:35/20

A.elseyi 10/16/92 10'30" 0.02%/135 65:35/20

A.cytherea 10/16/92 11' 0.02%/135 65:35/20

A.clathrata 6/11/93 10'30" 0.02%/120 65:35/20

A.florida 6/11/93 10'30" 0.02%/120 65:35/20

Montipora verrucosa 7/30/92 10' 0.02%/120 65:35/20

7/30/92 10' 0.02%/180 65:35/20

230 APPENDIX A-I. (Continued) Treatment parameters for experimentally- generated embryos.

Embryo Spawn Embryo Age [Colchicine] / [Hypo] / Culture Date (hr' min") duration duration (min) (min)

INTRASPECIFIC CROSSES

M. digitata "fat fingers" 10/12/92 10'30" 0.02%/135 65:35/20

10/13/92 10'45" 0.02%/135 65:35/20

M. digitata 10/12/92 10'30" 0.02%/135 65:35/20 "yellow spatulate"

M. spumosa 10/13/92 10'30" 0.02%/135 65:35/20

Fungia scutaria 8/15/92 8' 0.02%/120 65:35/20

8/15/92 8'30" 0.02%/90 65:35/20

8/15/92 9'30" 0.02%/140 65:35/20

INTERSPECIFIC CROSSES

A.millepora 10/16/92 10 '30" 0.02%/135 65:35/20 x A.pulchra 10'30" 0.02%/135 80:20/20

A. millepora 10/16/92 10'30" 0.02%/135 65:35/20 x A.formosa

A.pulchra 10/16/92 10'45" 0.02%/135 65:35/20 x A.cytherea

A.pulchra 11/14/92 II' 0.02%/120 65:35/20 x A.cytherea

Acropora 6/11/93 10'30" 0.02%/120 65:35/20 clathrata x A.florida

231 APPENDIX A-2. Spawning characteristics for corals from Guam, Okinawa, Australia, Palau, and Hawaii, including dates, time of day, egg color, and egg size . Beginning and end of observed time of spawning are accurate to + or - 10 minutes. Absence of precise time observations is indicated by II?II. Freshly-spawned, washed eggs are from a single colony for each species, and are measured across one diameter.

Egg Size (jlm)

Species/ Dates Spawning Egg Color n Mean sd Location Spawned Time Span

FAMILY ACROPORIDAE

Acropora divaricata 5/29/91 2215-2330 orange-red 25 424 54.2 Okinawa tv W tv A. nasuta 5/28/91 ? Okinawa 5/29/91 2215-2330 orange-red 25 439 45.7 5/30/91 2200-2245

A. surculosa 7/3/91 2100-? pink-orange 30 453 22.4 Guam 7/4/91 2030-? 7/6/91 2040-2100 7/21/92 2035-? 7/22/92 2030-? 50 615 49.0 7/24/92 2030-? A. digi tifera Okinawa 5/29/91 2215-2330 orange-red 25 457 39.9 Guam 7/21/92 2110-? 7/24/92 2100-? light orange 75 588 69.6 APPENDIX A-2. (Continued) Spawning characteristics for corals from Guam, Okinawa, Australia, Palau, and Hawaii. Egg Size (11 m)

Species/ Dates Spawning Egg Color n Mean sd Location Spawned Time Span

A. humilis 10/15/92 2030-? orange-pink 50 467 59.1 Australia

A.monticulosa 5/29/91 2200-2330 light orange 25 469 37.0 Okinawa

A. aspera 5/29/91 2200-? orange-red 25 489 39.6 Okinawa

A. formosa 10/16/92 2100-? light orange 50 489 30.5 Australia N w w A. nobilis 5/29/91 2215-2330 pink-orange 25 495 46.2 Okinawa

A. pulchra 10/16/92 2100-? pink-orange 50 505 55.3 Australia

A. tenuis 10/15/92 ? light orange Australia 10/16/92 1800-1845 50 514 78.7

A. millepora 10/16/92 2100-? pink 50 529 38.0 Australia

A. carduus 5/10/93 ? Palau 5/11/93 2020-2035 light orange 50 531 72.0 APPENDIX A-2. (Continued) Spawning characteristics for corals from Guam, Okinawa, Australia, Palau, and Hawaii. Egg Size ( /lm)

Species/ Dates Spawning Egg Color n Mean sd Location Spawned Time Span

A. cytherea Okinawa 5/30/91 2220-2250 pink 40 547 62.8 Australia 10/16/92 2100-7 pink 50 556 55.4

A. samoensis 10/15/92 2100-7 orange-pink 100 550 77.0 Australia

A. ocellata 7/4/91 2325-0100 light orange 30 554 47.6 Guam 7/5/91 2320-2400 7/6/91 2310-2400 7/21/92 2400-7 light orange 25 638 49.6 I\.l w A. florida 6/11/93 1900<7<2300 light orange 30 562 27.9 "" Palau A. ciathrata 6/11/93 1900<7<2300 light orange 50 566 32.0 Palau

A. secale 6/11/93 1900<7<2300 orange-red Palau

Acropora sp. 6/12/93 1900<7<2200 pink-orange (arborescent) Palau

A. robusta 7/6/91 2310-2400 orange-red 50 570 58.1 Guam 7/20/92 2340-7 red-orange 25 532 25.0 7/21/92 2340-7 7/23/92 2330-7 7/24/92 7 APPENDIX A-2. (Continued) Spawning characteristics for corals from Guam, Okinawa, Australia, Palau, and Hawaii. Egg Size ( /Lm)

Species/ Dates Spawning Egg Color n Mean sd Location Spawned Time Span

A. valida 8/5/90 2100-2230 orange-pink 50 598 63.0 Guam 8/12/90 2130-?

A. lutkeni 7/22/90 ? 30 607 29.0 Guam 7/5/91 2320-2400 red-orange 50 622 34.9 7/6/91 2310-2400

A. danai 8/4/90 2100-2215 Guam 8/5/90 2100-2230 pink-orange 50 626 88.0 8/6/90 2100-? 8/9/90 2110-2230 tv w 7/21/92 2110-? pink-orange 25 621 47.8 U1 7/22/92 2200-? 7/23/92 2210-?

A. elseyi 5/28/91 2030 red-orange 20 650 82.7 Okinawa 5/29/91 1945 Australia 10/16/92 2100-? red-orange

A. squarrosa 7/6/91 2100-2130 red 30 651 153.1 Guam

A. gemmifera 7/20/92 2320-? orange-pink 50 722 74.7 Guam 7/21/92 2310-? 7/22/92 2310-?

Montipora 7/29/92 2045-2115 light pink verrucosa 7/30/92 2045-? Hawaii APPENDIX A-2. (Continued) Spawning characteristics for corals from Guam, Okinawa, Australia, Palau, and Hawaii. Egg Size ( /lm)

Species/ Dates Spawning Egg Color n Mean sd Location Spawned Time Span

Montipora 10/14/92 2000-? spumosa Australia

Montipora digitata 10/12/92 2015-2025 Australia 10/13/92 2000-? "fat fingers" 10/14/92 2000-? "yellow 10/12/92 2015-2025 spatulate" 10/13/92 2000-?

N LV FAMILY FAVIIDAE 0'\ Goniastrea 7/22/92 2200-? pink-orange sp. Guam

FAMILY FUNGIIDAE

Fungia 8/15/92 1715-1830 no color scutaria Hawaii

FAMILY OCULINIDAE

Galaxea 7/21/92 2200-? light red 25 390 17.4 fascicularis Guam APPENDIX A-2. (Continued) Spawning characteristics for corals from Guam, Okinawa, Australia, Palau, and Hawaii. Egg Size (p.m)

Species/ Dates Spawning Egg Color n Mean sd Location Spawned Time Span

FAMILY DENDROPHYLLIDAE

Turbinaria 6/11/93 1900

tv W -..J LITERATURE CITED

ABE, N. 1937. Postlarval development of the coral Fungia actiniformis var. palawensis Doderlein. Palao Trop. Biol. Stn. Stud. 1: 73-93.

ACHITUV, Y. and Z. DUBINSKY. 1990. Evolution and zoogeography of coral reefs. In: Z. Dubinsky (Ed.). Ecosystems of the World. v. 25. Coral Reefs. Elsevier Science Publishers. pp. 1-9.

ANDERSON, E. 1949. Introgressive Hybridization. Wiley, NY.

ARNOLD, M.L. 1992. Natural hybridization as an evolutionary process. Ann. Rev. Ecol. Syst. 23: 237-261.

ARNOLD, M.L., J.L. HAMRICK, and B.D. BENNETT. 1990. Allozyme variation in Louisiana irises: a test for introgression and hybrid speciation. Heredity 65: 297-306.

ATKINSON, S. and M.J. ATKINSON. 1992. Detection of estradiol-17B during a mass coral spawn. Coral Reefs 11:33­ 35.

ATODA, K. 1951. The larvae and postlarval development of some reef-building corals. III. Acropora brueggemanni (Brook). ~. Morph. 89: 1-13.

AVERETT, J.E. 1980. Polyploidy in plant taxa: summary. In: W.H. Lewis, (Ed.). Polyploidy, Biological Relevance. Plenum Press, NY, pp. 269-271.

AYRE, D.J. and J.M. RESING. 1986. Sexual and asexual production of planulae in reef corals. Mar. Biol. 90: 187­ 190.

BABCOCK, R.C. and A.J. Heyward. 1986. Larval development of certain gamete-spawning scleractinian corals. Coral Reefs 5: 111-116.

BABCOCK, R.C., G.D. BULL, P.L. HARRISON, A.J. HEYWARD, J.K. OLIVER, C.C. WALLACE, and B.L. WILLIS. 1986. Synchronous spawning of 105 scleractinian coral species on the Great Barrier Reef. Mar. Biol. 90: 379-394.

BAK, R.P.M. and M.S. ENGEL. 1979. Distribution, abundance, and survival of juvenile hermatypic corals (Scleractinia) and the importance of life history strategies in the parent coral community. Mar. Biol. 54: 341-52.

238 BIRKELAND, C. 1992. Typhoons Omar and Brian: business as usual for Guam. Reef Encounters, Newsletter of the Int. Soc. for Reef Studies 12:11.

BIRKELAND, C., R.T. TSUDA, R.H. RANDALL, S.S. AMESBURY, and F. CUSHING. 1976. Limited current and underwater biological surveys of a proposed sewer outfall site on Malakal Island, Palau. Univ. of Guam Mar. Lab. Tech. Rept. 25, 59 pp.

BIRKELAND, C., G. GRIMM, and R.H. RANDALL. 1979. Separation of age and colony size as factors influencing the growth rates of corals. Abstract. In: E.V.Krasnov (Ed.). Use and preservation of coral reef resources. XIV Pac. Sci. Cong., Khabarovsk, p. 18.

BIRKELAND, C., R.H. RANDALL, R.C. WASS, B. SMITH, and S. WILKINS. 1985. Biological resource assessment of the Fagatele Bay National Marine Sanctuary. Development Planning and Tourism Office, Government of American Samoa, and Sanctuary Program Division, NOAA, U.S. Dept. of Commerce.

BOGART, J.P. 1980. Evolutionary implications of polyploidy in amphibians and reptiles. In: W.H. Lewis, (Ed.). Polyploidy, Biological Relevance. Plenum Press, NY, pp. 341­ 378.

BOTHWELL, A.M. 1981. Fragmentation, a means of asexual reproduction and dispersal in the coral genus Acropora - a preliminary report. Proc. Fourth Int. Coral Reef Symp., Manila. pp.137-144.

BRIGGS, B.G. 1962. Hybridization in Ranunculus. Evolution 16: 372-390.

BRODSKY, V.Y. and I.V. URYVAEVA. 1985. Genome Multiplication in Growth and Development (Development and Cell Biology Series No. 15). Cambridge University Press, Cambridge.

BUDDEMEIER, R.W. and R.A. KINZIE, III. 1976. Coral growth. Oceanogr. Mar. Biol. Ann. Rev. 14: 185-225.

BUSS, L.W. 1979. Habitat selection, directional growth and spatial refuges: why colonial animals have more hiding places. In: G. Larwood and B.R. Rosen (Eds.). Biology and Systematics of Colonial Organisms. Systematics Assn. Sp. Vol. 11: 459-497.

CAMPBELL, R.D. 1974. Cnidaria. In: A.C. Giese and J.S. Pearse (Eds.). Reproduction of Marine Invertebrates. Academic Press, New York, pp. 133-199.

239 CLAYTON, M.N. 1988. Evolution of life histories of brown algae. Bot. Mar. 31: 379-387.

COATES, A.G. and J.B.C. JACKSON. 1985. Morphological themes in the evolution of clonal and aclonal marine invertebrates. In: Population Biology and Evolution of Clonal Organisms. J.B.C. Jackson, L.W. Buss, and R.E. Cook (Eds.). Yale Univ. Press, New Haven, pp. 67-106.

CONNELL, J.H. 1978. Diversity in tropical rain forests and coral reefs. Science 199: 1302-1310.

COOK, R.E. 1983. Clonal plant populations. Amer. Sci. 71: 244-253.

COOK, R.E. 1985. Growth and development in clonal plant populations. In: Population Biology and Evolution of Clonal Organisms. J.B.C. Jackson, L.W. Buss, and R.E. Cook (Eds.). Yale Univ. Press, New Haven, pp. 259-296.

CRACRAFT, J. 1983. Species concepts and speciation analysis. Curro Ornith. 1: 159-187.

CRACRAFT, J. 1992. The species of the birds-of-paradise (Paradisaeidae): applying the phylogenetic species concept to a complex pattern of diversification. Cladistics 8: 1-43.

CROSBY, M.R. 1980. Polyploidy in bryophytes with special emphasis on mosses. In: W.H. Lewis, (Ed.). Polyploidy, Biological Relevance. Plenum Press, NY, pp. 193-198.

DANA, T.F. 1975. Development of contemporary eastern Pacific coral reefs. Mar. BioI. 33: 355-374.

DARLINGTON, C.D. 1956. Chromosome Botany. First Ed. George Allen and Unwin Ltd., London.

DARLINGTON, C.D. 1973. Chromosome Botany and the Origins of Cultivated Plants. Third (revised) Ed. George Allen and Unwin Ltd., London.

DAVIS, J.I. and P.S. MANOS. 1991. Isozyme variation and species delimitation in the Puccinellia nuttalliana complex (Poaceae): an application of the phylogenetic species concept. Syst. Bot. 16(3): 431-445.

DAWSON, G.W.P. 1962. An Introduction to the Cytogenetics of Polyploids. Blackwell Scientific Publications, Oxford, 96 pp.

240 DELEVORYAS, T. 1980. Polyploidy in gymnosperms. In: W.H. Lewis, (Ed.). Polyploidy, Biological Relevance. Plenum Press, NY, pp. 215-218.

DEMARAIS, B.D., T.E. DOWLING, M.E. DOUGLAS, W.L. MINCKLEY, AND P.C. MARSH. 1992. Hybrid origin of Gila seminuda (Pisces: Cyprinidae): implications for evolution and conservation. Proc. Natl. Acad. Sci. USA 89: 2747-51.

DENIS, H. and J. BRACHET. 1969a. Gene expression in interspecific hybrids. I. DNA synthesis in the lethal cross Arbacia lixula male x Paracentrotus lividus female. Proc. Nat. Acad. Sci. U.S.A., 62: 194-201.

DENIS, H. and J. BRACHET. 1969b. Gene expression in interspecific hybrids. II. RNA synthesis in the lethal cross Arbacia lixula male x Paracentrotus lividus female. Proc. Nat. Acad. Sci. U.S.A., 62: 438-445.

DE QUEIROZ, K. and J. GAUTHIER. 1992. Phylogenetic taxonomy. Ann. Rev. Ecol. Syst. 23: 449-480.

DEWET, J.M.J. 1980. Origins of polyploids. In: Polyploidy, Biological Relevance. W.H. Lewis, (Ed.). Plenum Press, NY, pp. 3-15.

DEWEY, D.R. 1980. Some applications and misapplications of induced polyploidy in plant breeding. In: Polyploidy, Biological Relevance. W.H. Lewis, (Ed.). Plenum Press, NY, pp. 445-470.

DINESEN, Z.D. 1980. A revision of the coral genus Leptoseris (Scleractinia: Fungiina: Agariciidae). Mem Qg. Mus. 20: 181­ 235, pls. 1-16.

DOLLAR, S.J. 1982. Wave stress and coral community structure in Hawaii. Coral Reefs 1:71-81.

DONE, T.D. 1982. Patterns in the distribution of coral communities across the central Great Barrier Reef. Coral Reefs 1: 93-107.

DOWLING, T.E. and W.M. BROWN. 1989. Allozymes, mitochondrial DNA, and levels of phylogenetic resolution among four minnow species (Notropis: Cyprinidae). Syst. Zool. 38: 126-143.

DOWLING, T.E. and B.D. DEMARAIS. 1993. Evolutionary significance of introgressive hybridization in cyprinid fishes. Nature 362: 444-446.

EHRENDORFER, F. 1964. Cytologie, taxonomie und evolution bei samenpflanzen. Vistas in Botany 4: 99-186.

241 EHRENDORFER, F. 1980. Polyploidy and evolution, In: Polyploidy, Biological Relevance. W.H. Lewis (Ed.). Plenum Press, NY, pp. 45-60.

ENGSTROM, M.D., a. SANCHEZ-HERRERA, and G. URBANO-VIDALES. 1992. Distribution, geographic variation, and systematic relationships within Nelsonia (Rodentia: Sigmodontinae). Proc. BioI. Soc. Wash. 105 (4): 867 - 881.

FADLALLAH, Y.H. 1983. Sexual reproduction, development and larval biology in scleractinian corals. A review. Coral Reefs 2: 129-150.

FUTUYMA, D.J. 1986. Evolutionary Biology. Sinauer Associates, Sunderland, MA, 600 pp.

GLYNN, P. W., N.J. GASSMAN, C.M. EAKEN, J. CORTES, D.V. SMITH, and H.M GUZMAN. 1991. Reef coral reproduction in the eastern Pacific: Costa Rica, Panama, and Galapagos Islands (Ecuador). Mar. BioI. 109: 355-368.

GODWARD, M.B.E. 1966. The Chromosomes of the Algae. St. Martin's Press, NY, 212 pp.

GOLDBLATT, P. 1980. Polyploidy in angiosperms: monocotyledons. In: W.H. Lewis, (Ed.). Polyploidy, Biological Relevance. Plenum Press, NY, pp. 219-239.

GOLDMAN, B. and F.H. TALBOT. 1976. Aspects of the ecology of coral reef fishes. In: a.A. Jones and R. Endean (Eds.). Biology and Geology of Coral Reefs. III. Biology 2. Academic Press, New York, pp. 125-154.

GOREAU, T.H. and N.E GOREAU. 1959. The physiology of skeleton formation in corals. II. Calcium deposition by hermatypic corals under various conditions in the reef. BioI. Bull. Mar. BioI. Lab., Woods Hole 117: 139-250.

GRANT, V. 1957. The plant species in theory and practice. In: E. Mayr (Ed.). The Species Problems. Amer. Assoc. Adv. Sci., Washington, DC.

GRANT, V. 1963. The Origin of Adaptations. Columbia Univ. Press, NY.

GRANT, V. 1981. Plant Speciation. Columbia University Press, New York.

GRANT, V. and A. GRANT. 1979. Hybridization and variation in the Opuntia phaeacantha group in Central Texas. Bot. Gaz. 140: 208-215.

242 GRIGG, R.W. 1981. Acropora in Hawaii. Part 2. Zoogeography. Pac. Sci. 35: 15-24.

GRIGG, R.W. 1983. Community structure, succession, and development of coral reefs in Hawaii. Mar. Ecol. Prog. Ser. 11:1-14.

GRIGG, R.W., J.W. WELLS, and C.C. WALLACE. 1981. Acropora in Hawaii. Part 1. History of the scientific record, systematics, and ecology. Pac. Sci. 35: 1-13.

HAIR, J.B. and E.J. BEUZENBERG. 1961. High polyploidy in a New Zealand Poa. Nature 189: 160.

HARPER, J.L. 1981. The concept of population in modular organisms. In: R.M. May (Ed.). Theoretical Ecology, Sinauer Assoc., Sunderland, MD., pp 53-77.

HARRISON, P.W. 1985. Sexual characteristics of scleractinian corals: systematic and evolutionary implications. Proc. Fifth Int. Coral Reef Congo Tahiti 4: 337-342.

HARRISON, P.L. and C.C. WALLACE. 1990. Reproduction, dispersal and recruitment of scleractinian corals. In: Z. Dubinsky, (Ed.). Coral Reefs, Ecosystems of the World, pp. 133-207.

HARRISON, P.L., R.C. BABCOCK, G.D. BULL, J.K. OLIVER, C.C. WALLACE, and B.L. WILLIS. 1984. Mass spawning in tropical reef corals. Science 223: 1186-1189.

HATAI, S. 1937. The Palao Tropical Biological Station. Palao TroQ. BioI. Stn. Stud. 1: 1-15.

HECK, K.L. JR. and E.D. MCCOY. 1978. Long-distance dispersal and the reef-building corals of the eastern Pacific. Mar. BioI. 48: 349-356.

HESLINGA, G.A., T.D. WATSON, and T. ISAMU. 1990. Giant Clam Farming. Pacific Fisheries Development Foundation (NMFS/NOAA), Honolulu, Hawaii, 179pp.

HEYWARD, A.J. 1985a. Sexual reproduction in five species of the coral Montipora. Hawaii Inst. Mar. BioI. Tech. Rept. 37: 170-178.

HEYWARD, A.J. 1985b. Comparative coral karyology. Proc. Fifth Int. Coral Reef Congr., Tahiti 6: 47-51.

HEYWARD, A.J. 1985c. Chromosomes of the coral Goniopora lobata (Anthozoa: Scleractinia). Heredity 55: 269-271.

243 HEYWARD, A.J. 1987. Genetic systems and hereditary structure of reef corals. Ph.D. dissertation, James Cook University of North Queensland.

HEYWARD, A.J. 1988. Reproductive status of some Guam corals. Micronesica 21:272-274.

HEYWARD, A. and R.C. BABCOCK. 1986. Self- and cross­ fertilization in scleractinian corals. Mar. BioI. 90:191­ 195.

HEYWARD, A.J., K. YAMAZATO, T. YEMIN, and M. MINEI. 1987. Sexual reproduction of corals in Okinawa. Galaxea 6: 331­ 343.

HIGHSMITH, R.C. 1981. Coral bioerosion at Enewetak: agents and dynamics. Int. Revue Ges. Hydrobiol. 66: 335-375.

HIGHSMITH, R.C. 1982. Reproduction by fragmentation in corals. Mar. Ecol. Prog. Ser. 7: 207-226.

HIGHSMITH, R.C., A.C. RIGGS, and C.M. D'ANTONIO. 1980. Survival of hurricane-generated coral fragments and a disturbance model of reef calcification/growth rates. Oecologia 46: 322-329.

HODGSON, G. 1988. Potential gamete wastage in synchronously spawning corals due to hybrid inviability. Proc. Sixth Int. Coral Reef Symp., Townsville 2:707-714.

HODGSON, G. 1989. The effects of sedimentation on Indo­ Pacific reef corals. Ph.D. dissertation, University of Hawaii, 338 pp.

HOEKSEMA, B.W. 1989. Taxonomy, phylogeny and biogeography of mushroom corals (Scleractinia: Fungiidae) zoologische Verhandelingen, Leiden 254: 1-295.

HOURIGAN, T.F. and E.S. REESE. 1987. Mid-ocean isolation and the evolution of Hawaiian reef fishes. Trends Ecol. Evol. 2:187-191.

HUGHES, R.N. 1989. A Functional Biology of Clonal Animals. Chapman and Hall, NY, 331 pp.

HUGHES, R.N. and J.M. CANCINO. 1985. An ecological overview of cloning in Metazoa. In: Population Biology and Evolution of Clonal Organisms. J.B.C. Jackson, L.W. Buss, and R.E. Cook (Eds.). Yale Univ. Press, New Haven, pp. 153-186.

244 HUGHES, T.P. 1985. Life histories and population dynamics of early successional corals. Proc. Fifth Int. Coral Reef Cong., Tahiti 4: 101-106.

HUGHES, T.P. 1988. Long-term dynamics of coral populations: contrasting reproductive modes. Proc. Sixth Int. Coral Reef Symp., Townsville 2: 721-725.

HUGHES, T.P., D. AYRE, and J.H. CONNELL. 1992. The evolutionary ecology of corals. Trends Ecol. Evol. 7(9) 292-295.

HUNTER, C.L. 1988. Genotypic diversity and population structure of the Hawaiian reef coral, Porites cornpressa. Ph.D. dissertation, University of Hawaii, 136 pp.

HUNTER, C.L. 1989. Environmental cues controlling spawning in two Hawaiian corals: Montipora verrucosa and M. dilitata. Proc. Sixth Int. Coral Reef Symp., Townsville 2: 727-732.

HUNTER, C.L. 1993. Genotypic variation and clonal structure in coral populations with different disturbance histories. Evolution 47(4): 1213-1228.

HUSTON, M.A. 1985. Patterns of species diversity on coral reefs. Ann. Rev. Ecol. Syst. 16: 149-177.

JABLONSKI, D. 1993. The tropics as a source of evolutionary novelty through geological time. Nature 364: 142-144.

JACKSON, J.B.C. 1977. Competition on marine hard substrata: the adaptive significance of solitary and colonial strategies. Am. Nat. 111: 743-767.

JACKSON, J.B.C. 1985. Distribution and ecology of clonal and aclonal benthic invertebrates. In: Population Biology and Evolution of Clonal Organisms. J.B.C. Jackson, L.W. Buss, and R.E. Cook (Eds.). Yale Univ. Press, New Haven, pp. 297­ 356.

JACKSON, J.B.C. 1986. Modes of dispersal of clonal benthic invertebrates: consequences for species distributions and genetic structure of local populations. Bull. Mar. Sci. 39: 588-606.

JACKSON, J.B.C. and T.P. HUGHES. 1985. Adaptive strategies of coral-reef invertebrates. Am. Sci. 73: 265-274.

JOKIEL, P.L. 1987. Ecology, biogeography and evolution of corals in Hawaii. Trends Ecol. Evol. 2(7): 179-182.

245 JOKIEL, P.L. 1990a. Long-distance dispersal by rafting: reemergence of an old hypothesis. Endeavour 14:66-73.

JOKIEL, P.L. 1990b. Transport of reef corals into the Great Barrier Reef. Nature 347~ 665-667.

JOKIEL, P.L. and J.I. MORRISSEY. 1986. Influence of size on primary production in the reef coral Pocillopora damicornis and the macroalgae Acanthophora spicifera. Mar. BioI. 91: 15-26.

KANESHIRO, K.Y. 1990. Natural hybridization in Drosophila, with special reference to species from Hawaii. Canadian ~. Zool. 68(8): 1800-1805.

KAWAGUTI, S. 1941. On the physiology of reef corals. V. Tropisms of coral planulae considered as a factor of distribution on the reef. Palao Trop. BioI. Stn. Stud. 2: 319-328.

KAY, E.A. 1979. Little worlds of the Pacific. An essay on Pacific basin biogeography. Harold L. Lyon Arboretum, Univ. of Hawaii, Honolulu, 40 pp.

KENYON, J.C. 1992. Sexual reproduction in Hawaiian Acropora. Coral Reefs 11: 37-43.

KEYL, H.G. 1966. Increases of DNA in chromosomes. In: Chromosomes Today, Vol. 1. Darlington, C.D. and K.R. Lewis (Eds.). Oliver and Boyd, Edinburgh, pp. 99-101.

KNOWLTON, N., J.C. LANG, M.C. ROONEY, and P. CLIFFORD. 1981. Evidence of delayed mortality in hurricane-damaged Jamaican staghorn corals. Nature 294: 251-252.

KOBAYASHI, A. 1984. Regeneration and regrowth of fragmented colonies of the hermatypic corals Acropora formosa and Acropora nasuta. Galaxea 3: 13-23.

KRUPP, D.A. 1983. Sexual reproduction and early development of the solitary coral Fungia scutaria (Anthozoa: Scleractinia). Coral Reefs 2: 159-164.

LADD, H.S. 1960. Origin of the Pacific island Molluscan fauna. Am J. Sci. 258-A: 137-150.

LANG, J.C. 1984. Whatever works: the variable importance of skeletal and of non-skeletal characters in scleractinian taxonomy. Palaeon. Am. 54: 18-44.

246 LEHMAN, N., A. EISENHAWER, K. HANSEN, L.D. MECH, and R.O. PETERSON. 1991. Introgression of Coyote mitochondial DNA into syrnpatric North American gray wolf populations. Evolution 45: 104-119.

LEVIN, D.A. 1983. Polyploidy and novelty in flowering plants. Am. Nat. 122: 1-25.

LEWIS, H. 1953. Chromosome phylogeny and habitat preference of Clarkia. Evolution 7: 102-109.

LEWIS H. and W.L BLOOM. 1972. The loss of a species through breakdown of a chromosomal barrier. ~. BioI. Hungarica 12: 61-64.

LEWIS, W.H. 1980a. Polyploidy in species populations. In: Polyploidy, Biological Relevance. W.H. Lewis, (Ed.). Plenum Press, NY, pp. 103-144.

LEWIS, W.H. 1980b. Polyploidy in angiosperms: dicotyledons. In: W.H. Lewis, (Ed.). Polyploidy, Biological Relevance. Plenum Press, NY, pp. 241-268.

LOKKI, J. and A. SAURA. 1980. Polyploidy in insect evolution. In: W.H. Lewis, (Ed.). Polyploidy, Biological Relevance. Plenum Press, NY, pp. 277-312.

MACDONALD, C.D. 1984. Studies on recruitment in the Hawaiian spiny lobster Panulirus marginatus. Proc. Res. Inv. NWHI. UNIHI-SEAGRANT-MR-84-01 1:199-220.

MACGREGOR, H. and J. VARLEY. 1983. Working with Animal Chromosomes. John Wiley & Sons, NY, 250 pp.

MARAGOS, J.E. 1991. Surveys of stony corals off Ngeremeduu Bay, eastern Babeldaob, and the southern barrier reef of Palau. The Nature Conservancy, Honolulu, 20 pp.

MARAGOS, J.E., BAINES, G., and P. BEVERIDGE. 1973. Tropical cyclone Bebe creates a new land formation on Funafuti Atoll. Science 181: 1161-1164.

MAYR, E. 1942. Systematics and the origin of species. Columbia Univ. Press, NY.

MAYR, E. 1963. Animal species and evolution. Belknap Press of Harvard Univ. Press, Cambridge, MA.

MCCOY, E.D. and K.L. HECK, JR. 1976. Biogeography of corals, seagrasses, and mangroves: an alternative to the center of origin concept. Syst. Zool. 25: 201-210.

247 MCMANUS, J.W. 1985. Marine speciatio~, tectonics and sea­ level changes in Southeast Asia. Proc. Fifth Int. Coral Reef Cong., Tahiti 4: 133-138.

MCMILLAN, J. and D.J MILLER. 1988. Restriction analysis and DNA hybridization applied to the resolution of Acropora nobilis from Acropora formosa. Proc. Sixth Int. Coral Reef Symp., Townsville 2: 775-777.

MCMILLAN, J. and D.J. MILLER. 1989. Nucleotide sequences of highly repetitive DNA from scleractinian corals. Gene 83: 185-186.

MCMILLAN, J. and D.J. MILLER. 1990. Highly repeated DNA sequences in the scleractinian coral genus Acropora: evaluation of cloned repeats as taxonomic probes. Mar. Bioi. 104: 483-487.

MCMILLAN, J., D. YELLOWLEES, A. HEYWARD, P. HARRISON, and D.J. MILLER. 1988. Preparation of high molecular weight DNA from hermatypic corals and its use for DNA hybridization and cloning. Ma~. Bioi. 98: 271-276.

MCMILLAN, J., T. MAHONY, J.E.N. VERON, and D.J. MILLER. 1991. Nucleotide sequencing of highly repetitive DNA from seven species in the coral genus Acropora (Cnidaria: Scleractinia) implies a division contrary to morphological criteria. Mar. Biol. 110: 323-327.

MCNALLY, G.J., W.C. PATZERT, A.D. KIRWIN, and A.C. VASTANO. 1983. The near-surface circulation of the North Pacific using satellite tracked drifting buoys. ~. Geophys. Res. 88: 7507-7518.

MILEIKOVSKY, S.A. 1971. Types of larval development in marine bottom invertebrates, their distribution and ecological significance: are-evaluation. Mar. Biol. 10: 193-213.

MILLER, R.L. 1982. Identification of sibling species within the "8arsia tubulosa complex" at Friday Harbor, Washington, (Hydrozoa: Anthomedusae). ~. Exp. Mar. Biol. Ecol. 62: 153­ 172.

MISHLER, D.D. and DONOGHUE, M.J. 1982. Species concepts: a case for pluralism. Syst. Zool. 31(4): 491-503.

MOORE, J.A. 1961. A cellular basis for genetic isolation. In W.F. Blair (Ed.). Vertebrate speciation. Univ. Texas Press, Austin, pp. 62-68.

248 MULLER, H.J. 1925. Why polyploidy is rarer in animals than in plants. Amer. Nat. 59: 346-353.

NEIGEL, J.E. and J.C. AVISE. 1983. Clonal diversity and population structure in a reef-building coral Acropora cervicornis: self recognition analysis and demographic interpretation. Evolution 37(3): 437-453.

NICHOLS, H.W. 1980. Polyploidy in algae. In: W.H. Lewis, (Ed.). Polyploidy, Biological Relevance. Plenum Press, NY, pp. 151-161.

NISHIHlRA, M. 1988. Field Guide to Hermatypic Corals of Japan. Tokai University Press.

NISHIHIRA, M., K. YANAGIYA, and K. SAKAI. 1987. A preliminary list of hermatypic corals collected around Kudaka Island, Okinawa. Galaxea 6: 53-60.

NIXON, K.C. and Q.D. WHEELER. 1990. An amplification of the phylogenetic species concept. Cladistics 6: 211-223.

OHNO, S. 1970. Evolution by Gene Duplication. Springer­ Verlag, NY.

OHNO, S., MURAMOTO, J., and L. CHRISTIAN. 1967. Diploid­ tetraploid relationship among old world members of the fish family Cyprinidae. Chromosoma: 23:1-9.

OLIVER, J.K. and B.L. WILLIS. 1987. Coral-spawn slicks in the Great Barrier Reef: preliminary observations. Mar. BioI. 94: 521-529.

OLIVER, J.K, R.C. BABCOCK, P.L. HARRISON, and B.L. WILLIS. 1988. The geographic extent of mass coral spawning: clues to proximate and ultimate causal factors. Proc. Sixth Int. Coral Reef Symp., Townsville 2: 803-810.

ORCUTT, A.M., R. CORDY, P.J. RAPPA, and B.D. SMITH. 1989. Yap Proper coastal resource inventory. Prep. for U.S. Army Corps of Engineers, Pacific Ocean Division. Prep. by Sea Grant Ext. Serv., University of Hawaii, Honolulu, 36 pp.

ORR, H.A. 1990. "Why polyploidy is rarer in animals than in plants" revisited. Am. Nat. 136 ':5): 759-770.

PALMER, E. J. 1948. Hybrid oaks of North 7,merica. lI. Arnold Arboretum 29: 1-48.

PATERSON, H.E.H. 1982. Perspective on speciation by reinforcement. ~. African lI. Sci. 78: 53-57.

249 PATZERT, W.C., K. WYRTKI, and H.J. SANTAMORE. 1970. Current measurements in the central north Pacific Ocean. Hawaii Inst. Geophys HIG-70-31, 26pp.

PAULEY, G. 1989. Effects of sea level fluctuation on insular coral faunas. Am. Zool. 29: 90A.

PAULEY, G. 1991. Late Cenozoic sea level fluctuations and the diversity and species composition of insular shallow water marine faunas. In: E.C. Dudley (Ed.). The Unity of Evolutionary Biology. Proc. Fourth Int. Congo of Syst. and Evol. Biol. 1: 184-192.

PORTER, J.W., J.D. WOODLEY, G.J. SMITH, J.E. NEIGEL, J.F. BATTEY, and D.G. DALLMEYER. 1981. Population trends among Jamaican reef corals. Nature 294: 249-250.

POTTS, D.C. 1983. Evolutionary disequilibrium among Indo­ Pacific corals. Bull. Mar. Sci. 33: 619-632.

POTTS, D.C. 1984a. Generation times and the Quaternary evolution of reef-building corals. Paleobiology 10: 48-58.

POTTS, D.C. 1984b. Natural selection in experimental populations of reef-building corals (Scleractinia). Evolution 38: 1059-1078.

POTTS, D.C. 1985. Sea-level fluctuations and speciation in Scleractinia. Proc. Fifth Int. Coral Reef Cong., Tahiti 4: 127-132.

POTTS, D.C. and R.L. GARTHWAITE. 1991. Evolution of reef­ building corals during periods of rapid global change. In: E.C. Dudley (Ed.). The Unity of Evolutionary Biology. Proc. Fourth Int. Congo of Syst. and Evol. Biol. 1:170-178.

RAHAT, A., M. RAHAT, and J.B. SEARLE. 1985. A simple method for the preparation of hydra chromosome spreads: introducing chromosome counts into hydra taxonomy. Experientia 41: 282­ 283.

RANDALL, R.H. 1973. Reef physiology and distribution of corals at Tumon Bay, Guam, before crown-of-thorns starfish Acanthaster planci (L.) predation. Micronesica 9: 119-158.

RANDALL, R.H. 1990. The general geographic and geological setting, physiographic description, and community structure of the reef-building corals of Ngesacl Reef, Koror State, Republic of Palau. In Appendix J of Environmental Assessment for Ocean Development Company, Inc. Palau Resort Development, Ngesaol, Koror State, Republic of Palau. Prep.

250 for Ocean Development Co., Tokyo, by Brewer/Brandman Associates, Honolulu, 47 pp.

RANDALL, R. and L. ELDREDGE. 1977. Effects of typhoon Pamela on the coral reefs of Guam. Third Int. Coral Reef Symp. Vol.2, pp. 525-531.

RANDALL, R.H. and R.F. MYERS. 1983. Guide to the Coastal Resources of Guam. Volume II. The Corals. University of Guam Press.

RANDALL, R.H., C. BIRKELAND, S.S. AMESBURY, D. LASSUY, and J.R. EADS. 1978. Marine survey of a proposed resort site at Arakabesan Island, Palau. Univ. Guam Mar. Lab. Tech. Rept. 44, 73 pp.

RANDOLPH, L.F. 1966. Iris nelsonii, a new species of Louisiana Iris of hybrid origin. Baileya 14: 143-169.

RANDOLPH, L.F., I.S. NELSON, and R.L. PLAISTED. 1967. Negative evidence of introgression affecting the stability of Louisiana Iris species. Cornell Univ. Agr. Expt. Stat. Memoir 398: 1-56.

RICHMOND, R.H. 1985. Variations in the population biology of Pocillopora damicornis across the Pacific Ocean. Proc. Fifth Int. Coral Reef Cong., Tahiti 6: 101-106.

RICHMOND, R.H. 1987. Energetics, competency, and long­ distance dispersal of planula larvae of the coral pocillopora damicornis. Mar. BioI. 93: 527-533.

RICHMOND, R.H. 1988. Competency and dispersal potential of planula larvae of a spawning versus a brooding coral. Proc. Sixth Int. Coral Reef Symp., Townsville 2: 827-831.

RICHMOND, R.H. 1990a. The effects of the El Nino/Southern Oscillation on the dispersal of corals and other marine organisms. In: P.W. Glynn (Ed.). Global Ecological Consequences of the 1982-1983 El Nino-Southern Oscillation. Elsevier Ocean. Series 52, pp. 127-140.

RICHMOND, R.H. 1990b. Relationships among reproductive mode, biogeographic distribution patterns and evolution in scleractinian corals. In: M. Hoski and O. Yamashita, (Eds.) Advances in Invertebrate Reproduction 5, Elsevier Science Publ., pp. 317-322.

RICHMOND, R.H. 1993. Coral reefs: present problems and future concerns resulting from anthropogenic disturbance. Amer. Zool. 33: 524-536.

251 RICHMOND, R.H. and C.L. HUNTER. 1990. Reproduction and recruitment of corals: comparisons among the Caribbean, the tropical Pacific, and the Red Sea. Mar. Ecol. Prog. Ser. 60: 185-203.

RIESENBERG, L.H., R. CARTER, and S. ZONA. 1990. Molecular tests of the hypothesized hybrid origin of two diploid Helianthus species (Asteraceae). Evolution 44: 1498-1511.

RILEY, H.P. 1938. A character analysis of colonies of Iris fulva and I. hexagona var. giganticaerulea. Amer. ~. Bot. 29: 323-331.

ROSEN, B.R. 1971. Principal features of reef coral ecology in shallow water environments of Mahe, Seychelles. ~. Zool. Soc. Lond. 28: 163-183.

ROSEN, B.R. 1984. Reef coral biogeography and climate through the late Cainozoic: just islands in the sun or a critical pattern of islands? In: P. Brenchley, (Ed.). Fossils and Climate. John Wiley & Sons, NY, pp. 201-262.

ROSEN, B.R. 1993. Book review of: Modern reef development and Cenozoic evolution of an oceanic island/reef complex: Isla de Providencia (western Caribbean Sea, Colombia). Reef Encounters, Newsletter of the Int. Soc. for Reef Studies 13:20.

ROSEN, B.R. and J.D. TAYLOR. 1969. Reef coral from Aldabra: new mode of reproduction. Science 166: 119-120.

ROTHWELL, N.V. 1959. Aneuploidy in Claytonia virginica. Amer. ~. Bot. 46: 353-360.

RYLAARSDAM, K.W. 1983. Life histories and abundance patterns of colonial corals on Jamaican reefs. Mar. Ecol. Prog. Ser. 13: 249-260.

SAKAI, K. and K. YAMAZATO. 1987. Preliminary list of hermatl~ic corals around Sesoko Island, Okinawa with a note on the decrease of the species richness from 1980 to 1986. Galaxea 6:43-51.

SALE, P.F. 1977. Maintenance of high diversity in coral reef fish communities. Am. Nat. 111: 337-359.

SAMMARCO, P.W. 1982. Polyp bail-out: an escape response to environmental stress and a new means of reproduction in corals. Mar. Ecol. Prog. Ser. 10: 57-65.

252 SAMMARCO, P.W. and J.C. ANDREWS. 1988. Localized dispersal and recruitment in Great Barrier Reef corals: the Helix experiment. Science 239: 1422-1424.

SCHEER, G. and C.S.G. PILLAI. 1983. Report on the stony corals from the Red Sea. Zoologica, Stuttgart 133: 1-198.

SCHELTEMA, R.S. 1988. Initial evidence for the transport of teleplanic larvae of benthic invertebrates across the East Pacific Barrier. BioI. Bull. 174: 145-152.

SCHULTZ, R.J. 1980. Role of polyploidy in the evolution of fishes. In: W.H. Lewis, (Ed.). Polyploidy, Biological Relevance. Plenum Press, NY, pp. 313-340.

SECKEL, G.R. 1962. Atlas of the oceanographic climate of the Hawaiian Islands region. U.S. Fish Wild. Servo Fish. Bull. 61:371-427.

SEXTON, O.J. 1980. Polyploidy in animal evolution: summary. In: W.H. Lewis (Ed.) Polyploidy, Biological Relevance. Plenum Press, NY, pp. 379-381.

SHAW, D.D. 1981. Chromosomal hybrid zones in orthopteroid insects. In: W.R. Atchley and D.S. Woodruff (Eds.). Evolution and speciation: essays in honor of M.J.D. White. Cambridge Univ. Press, Cambridge, pp. 146-170.

SHAW, P.W., J.A. BEARDMORE, and J.S. RYLAND. 1987. Sagartia troglodytes (Anthozoa: Actinaria) consists of two species. Mar. Ecol. Prog. Ser. 41: 21-28.

SHEPPARD, C.R.C. and A.L.S. SHEPPARD. 1991. Corals and coral communities of Arabia. In: Fauna of Saudi Arabia 12: 1-170.

SHLESINGER, Y. and Y. LOYA. 1985. Coral community reproductive patterns: Red Sea versus the Great Barrier Reef. Science 228: 1333-1335.

SIBLEY, C.G. 1954. Hybridization in the red-eyes towhees of Mexico. Evolution 8: 252-290.

SILANDER, J.A. JR. 1985. Microevolution in clonal plants. In: Population Biology and Evolution of Clonal Organisms. J.B.C. Jackson, L.W. Buss, and R.E. Cook (Eds.). Yale Univ. Press, New Haven, pp. 107-152.

SIMPSON, G.G. 1961. Principles of animal taxonomy. Columbia Univ. Press, New York.

253 SIMPSON, C.J. 1985. Mass spawning of scleractinian corals in the Dampier Archipelago and the implications for management of coral reefs in Western Australia. Dept. Cons. and Environ., Perth, W.A., Bull. 244: 1-35.

SMITH, S.V. and R.W. BUDDEMEIER. 1992. Global change and coral reef ecosystems. Ann. Rev. Ecol. Syst. 23: 89-119.

SOLIGNAC, M. and M. MONNEROT. 1986. Race formation, speciation, and introgression within Drosophila simulans, D. mauritiana, and D. sechellia inferred from mitochondrial DNA analysis. Evolution 40: 531-539.

SOONG, K. 1991. Sexual reproductive patterns of shallow­ water reef corals in Panama. Bull. Mar. Sci. 49(3): 832-846.

SPENCE, J.R. 1990. Introgressive hybridization in Heteroptera: the example of Limnoporus Stal (Gerridae) species in western Canada. Canadian~. Zool. 68(8): 1770­ 1782.

SPENCE, J.R. and R.H. GOODING. 1990. Evolutionary significance of hybridization and introgression in insects. Canadian~. Zool. 68(8): 1699-1700.

STACE, C.A. 1980. Plant Taxonomy and Biosystematics. Edward Arnold, London, 279 pp.

STEBBINS, G.L. 1950. Variation and Evolution in Plants. Columbia University Press, NY, 643 pp.

STEBBINS, G.L. 1959. The role of hybridization in evolution. Proc. Amer. Phil. Soc. 103: 231-251.

STEBBINS, G.L. 1970. Variation and evolution in plants: progress during the past twenty years. In: M.K. Hecht and W.C. Steere (Eds.). Essays in evolution and genetics in honor of Theodosius Dobzhansky. Appleton-Centry-Crofts, NY.

STEBBINS, G.L. 1971. Chromosomal Evolution in Higher Plants. Edward Arnold, London.

STEBBINS, G.L. and G.J.C. HILL. 1980. Did multicellular plants invade the land? Am. Nat. 115: 342-353.

STEHLI, F.G. and J.W. WELLS. 1971. Diversity and age patterns in hermatypic corals. Syst. Zool. 20: 115-126.

STODDART, D.R. 1963. Effects of Hurricane Hattie on the British Honduras reefs and cays, Oct. 30-31, 1961. Atoll Res. Bull. 95: 1-142.

254 STODDART, J.A. 1983. Asexual production of planulae in the coral Pocillopora damicornis. Mar. Biol. 76: 279-284.

STODDART, J.A. 1984a. Genetical structure within populations of the coral Pocillopora damicornis. Mar. Biol. 81: 19-30.

STODDART, J.A. 1984b. Genetical differentiation amongst populations of the coral Pocillopora damicornis off southwestern Australia. Coral Reefs 3: 149-156.

STODDART, J.A. 1985. Biochemical genetics of Pocillopora damicornis in Kaneohe Bay, Oahu, Hawai.i. In: P.L. Jokiel and R.H. Richmond (Eds.). Coral Reef Population Biology. Hawaii Inst. Mar. Biol. Tech. Rept. No. 37. pp. 133-150.

STRATHMANN, R.R. 1981. On barriers to hybridization between Stronglocentrotus droebachiensis (O.F. Muller) and S. pallidus (G.O. Sars). ~. Exp. Mar. Biol. Ecol. 55: 39-47.

STRATHMANN, R.R. 1987. General Procedures. In: M.F. Strathmann (Ed.). Reproduction and Development of Marine Invertebrates of the Northern Pacific Coast. Univ. of Washington Press, pp. 3-104.

SZMANT-FROELICH, A. 1985. The effect of colony size on the reproductive ability of the Caribbean coral Montastrea annularis (Ellis and Solander). Proc. Fifth Int. Coral Reef Congr., Tahiti 2: 372.

SZMANT, A.M. 1986. Reproductive ecology of Caribbean reef corals. Coral Reefs 5:43-54.

SZMANT-FROELICH, A., P. YEVICH, and M.E.Q. Pilson. 1980. Gametogenesis and early development of the temperate coral Astrangia danae (Anthozoa: Scleractinia). Biol. Bull. 158: 257-269.

TEN LOHUIS, M., P. ALDERSLADE, and D.J. MILLER. 1990. Isolation and cloning of DNA from somatic tissue of soft corals (Cnidaria: Octocorallia). Mar. Biol. 104: 489-492.

TISCHLER, G. 1954. Das problem der basis chromosomenzahlen bei den angiospermen gattungen und familien. Cytologia 19: 1-10.

TUCKER, J.M. 1952. Evolution of the California oak Quercus alvordiana. Evolution 6: 162-180.

TUNNICLIFFE, V. 1981. Breakage and propagation of the stony coral Acropora cervicornis. Proc. Natl. Acad. Sci. 78(4): 2427-2431.

255 TURNER, J.R.G. 1971. Two thousand generations of hybridization in a Heliconius butterfly. Evolution 25: 471­ 482.

UCHIDA, H. and T. FUKUDA. 1989. Pictorial Dictionary of Okinawan Marine Life. Coral Guide. ~hinsei Tosho Press.

UYENO, T. and G.R. SMITH. 1972. Tetraploid origin of the karyotype of catostomid fishes. Science 175: 644-646.

VAN VALEN, L. 1976. Ecological species, multispecies, and oaks. Taxon 25: 233-239.

VAUGHAN, T.W. and J.W. WELLS. 1943. Revision of the suborders, families and genera of the Scleractinia. Geol. Soc. Am. ~. Pap. 44: 1-363, pIs. 1-51.

VERON, J.E.N. 1981. The species concept in 'Scleractinia of Eastern Australia'. Proc. Fourth Int. Coral Reef Symp., Manila 2: 183-186.

VERON, J.E.N. 1985. Aspects of the biogeography of hermatypic corals. Proc. Fifth Int. Coral Reef Cong., Tahiti 2: 83-88.

VERON, J.E.N. 1986. Corals of Australia and the Indo­ Pacific. Angus and Robertson, Australia, 644 pp.

VERON, J.E.N. 1990. Re-examination of the reef corals of Cocos (Keeling) Atoll. Rec. West. Aust. Mus. 14(4): 553-581.

VERON, J.E.N. 1993. A biogeographic database of hermatypic corals. Species of the Central Indo-Pacific genera of the world. Aust. Inst. Mar. Sci. Mono. Ser. 10: 433 pp.

VERON, J.E.N. and M. PICHON. 1976. Scleractinia of Eastern Australia, Part I. Families Thamnasteriidae, Astrocoeniidae, pocilloporidae. Aust. Inst. Mar. Sci. Mono. Ser. 1: 86pp.

VERON, J.E.N. AND M. PICHON. 1980. Scleractinia of Eastern Australia, Part III. Families Agariciidae, Siderastreidae, Fungiidae, Oculinidae, Merulinidae, Mussidae, Pectiniidae. Aust. Inst. Mar. Sci. Mono. Ser. 4: 422 pp.

VERON, J.E.N. and M. PICHON. 1982. Scleractinia of Eastern AustYalia, Part IV. Familiy Poritidae. Aust. Inst. Mar. Sci. Mono. Ser. 5: 159 pp.

VERON, J.E.N. AND C.C. WALLACE. 1984. Scleractinia of Eastern Australia, Part V. Family Acroporidae. Aust. Inst. Mar. Sci. Mono. Ser. 6: 485 pp.

256 VERON, J.E.N., M. PICHON, and M. WIJSMAN-BEST. 1977. Scleractinia of Eastern Australia, Part II. Families Faviidae, Trachyphyllidae. Aust. Inst. Mar. Sci. Mono. Ser. 3: 233 pp.

VON PRAHL, H. and A. MEJIA. 1985. Primer informe de un coral acroporido, Acropora valida (Dana, 1846) (Scleractinia: Astrocoeniida: Acroporidae) para el Pacifico americano. Rev. BioI. Trop. 33 (1): 39-43.

WAGNER, W.H. and F.S. WAGNER, 1980. Polyploidy in pteridophytes. In: W.H. Lewis, (Ed.). Polyploidy, Biological Relevance. Plenum Press, NY, pp. 199-214.

WALLACE, C.C. 1978. The coral genus Acropora (Scleractinia: Astrocoeniina: Acroporidae) in the Central and Southern Great Barrier Reef Province. Mem. QQ. Mus. 18(2): 273-319, pIs. 43-103.

WALLACE, C.C. 1985. Reproduction, recruitment and fragmentation in nine sympatric species of the coral genus Acropora. Mar. BioI. 88: 217-233.

WALLACE, C.C., J.M. PANDOLFI, A. YOUNG, and J. WOLSTENHOLME. 1991. Indo-Pacific coral biogeography: a case study from the Acropora selago group. Aust. Syst. Bot. 4: 199-210.

WEINBERG, J.R., V.R. STARCZAK, C. MUELLER, G.C. PESCH, and S.M. LINDSAY. 1990. Divergence between populations of a monogamous polychaete with male parental care: premating isolation and chromosome variation. Mar. BioI. 107: 205-213.

WELLS, J.W. 1954. Recent corals of the Marshall Islands. U.S.G.S. Prof. Pap. I: 285-486.

WELLS, J.W. 1955. A Survey of the Distribution of Reef Coral Genera in the Great Barrier Reef Region. Cornell University.

WELLS, J.W. 1956. Scleractinia. In: Moore, R.C. (Ed.). Treatise on Invertebrate Paleontology, Part F, Coelenterata. Geological Society of America and University of Kansas Press, Lawrence, Kansas. pp. 329-479.

WELLS, J.W. 1987. Notes on Indo-Pacific scleractinian corals. Part II. A new species of Acropora from Australia. Pac. Sci. 39: 338-339.

WHITE, M.J.D. 1942. The Chromosomes. Second Ed. Methuen, London.

WHITE, M.J.D. 1973. The Chromosomes. Sixth Edition. Chapman and Hall, London, 214 pp.

257 WHITE, M.J.D. 1978. Modes of Speciation. W.H. Freeman and Co., San Francisco, 455 pp.

WHITELEY, H.R. and A.H. WHITELEY. 1975. Changing populations of reiterated DNA transcripts during early echinoderm development. In: A.A. Moscona and A. Mamay (Eds.). Current Topics in Developmental Biology. Academic Press, NY.

WIJSMAN, H.J. and M. WIJSMAN-BEST. 1973. A note on the chromosomes of some madreporarian corals from the Mediterranean (Cnidaria, Mlthozoa, Scleractinina: Dendrophyllidae, Cryophyllidae, Faviidae). Genen. Phenen. 16(2): 61-64.

WILDE, G.R. and A.A. ECHELLE. 1992. Genetic status of pecos pupfish populations after establishment of a hybrid swarm involving an introduced congener. Trans. Am. Fish. Soc. 121(3): 277-286.

WILEY, E.O. 1978. The evolutionary species concept reconsidered. Syst. Zool. 27: 17-26.

WILLIS, B.L. 1990. Species concepts in extant scleractinian corals: considerations based on reproductive biology and genotypic population structures. Syst. Bot. 15(1): 136-149.

WILLIS, B.L., R.C. BABCOCK, P.L. HARRISON, and J.K. OLIVER. 1985. Patterns in the mass spawning of corals on the Great Barrier Reef from 1981 to 1984. Proc. Fifth Int. Coral Reef Cong., Tahiti 4: 343-348.

WILLIS, B.L., R.C. BABCOCK, P.L. HARRISON, and C.C. WALLACE. 1992. Experimental evidence of hybridization in reef corals involved in mass spawning events. Abstract, Seventh Int. Coral Reef Symp., Guam.

WOLF, U., RITTER, H., ATKIN, N.B., and S.OHNO. 1969. Polyploidization in the fish family Cyprinidae, order Cypriniformes. I. DNA content and chromosome sets in various species of Cyprinidae. Humagenetik 7: 240-244.

WYERS, S. 1985. Sexual reproduction of the coral Diploria strigosa (Scleractinia, Faviidae) in Bermuda: research in progress. Proc. Fifth Int. Coral Reef Congr., Tahiti 4: 301­ 306.

ZINMEISTER, W.J. and W.K. EMERSON. 1979. The role of passive dispersal in the distribution of hemipelagic invertebrates, with examples from the tropical Pacific Ocean. Veliger 22: 32-40.

258