Marine Biology (1999) 134: 753760

Springer-Verlag 1999

M. Adjeroud M. Tsuchiya

Genetic variation and clonal structure in the scleractinian coral Pocillopora damicornis in the Ryukyu Archipelago, southern Japan

Received: 7 August 1998 / Accepted: 18 May 1999

Abstract Samples of the scleractinian coral Pocillopora damicornis were collected from six sites located around four islands in the Ryukyu Archipelago, southern Japan, and subjected to allozyme electrophoresis. Seven polymorphic loci were examined for their allelic patterns. The ratio of observed to expected genotypic diversity (0.30 < Go:Ge < 0.64), the ratio of the observed number of genotypes to the number of individuals (0.47 < Ng:Ni < 0.75), and deviations from Hardy Weinberg equilibrium indicated that asexual reproduction plays a major role in the maintenance of established populations. However, populations were not completely dominated by a single or a few clones, and most clones were represented by only a few individual samples. The high frequency of typhoons in the region suggests that, in P. damicornis, fragmentation caused through occasional exposure to powerful waves is a major mode of asexual reproduction, but asexual production of planulae may also be contributing to the maintenance of populations. A signicant genetic dierentiation (FST) was found between the six populations examined (0.027 < FST < 0.092, average FST = 0.056). The

moderate gene ow is discussed according to characteristics of the larval stage of the species, and to circulation patterns in the region.

Introduction
Scleractinian corals have several modes of reproduction: sexual reproduction by broadcast of gametes or planulae, or asexual reproduction by fragmentation, budding, broadcast-spawning of asexual planulae, or polyp expulsion (Highsmith 1982; Fadlallah 1983; Stoddart 1983; Harrison and Wallace 1990; Kramarsky-Winter et al. 1997). Recently, examination of the genetic structure of populations by allozyme electrophoresis has allowed estimates of the relative importance of sexual and asexual reproduction in coral populations (Stoddart 1983; Willis and Ayre 1985; Stoddart et al. 1988; Ayre and Dufty 1994; Benzie et al. 1995). The results of these studies have revealed great variation in the relative contributions of sexual and asexual reproductions, not only among dierent species but also within single species. Stoddart (1984a) found that asexual reproduction by means of brooded larvae was particularly important in populations of Pocillopora damicornis in southwestern Australia and Hawaii, whereas Benzie et al. (1995) and Ayre et al. (1997) found that sexual reproduction was predominant in populations on the Great Barrier Reef (GBR). Sexual reproduction is followed by broadcast spawning of pelagic gametes in oviparous species and of planulae in viviparous, or brooding, species. The characteristics of the pelagic phase, such as its duration and its spatial dispersion, are poorly known because spawning is a temporally restricted phenomenon, and larvae are inaccessible by conventional methodology. Laboratory experiments, nevertheless, have shown that larval life is highly variable among species, ranging from a few minutes to nearly 300 d (Fadlallah 1983; Harrison and Wallace 1990; Carlon and Olson 1993). Therefore, dispersal of coral larvae has been mainly assessed by

Communicated by S.A. Poulet, Rosco M. Adjeroud (8)1 M. Tsuchiya University of the Ryukus, Department of Chemistry, Biology, and Marine Science, Nishihara, Okinawa 90301 Japan M. Adjeroud M. Tsuchiya Tropical Biosphere Research Center, Sesoko Station, University of the Ryukus, 2422 Sesoko, Motobu, Okinawa 905-02 Japan Present address: diterrane enne, Centre de Biologie et d'Ecologie Tropicale et Me Ecole Pratique des Hautes Etudes, URA CNRS 1453, de Perpignan, Universite F-66860 Perpignan Cedex, France
1

Fax: 0033 (0)4 6850 36-86 e-mail: adjeroud@univ-perp.fr

754

examination of its indirect eects on the genetic dierentiation and the gene ow among populations (Ayre and Dufty 1994). The relative importance of sexual and asexual reproduction and the spatial scale of larval dispersal have profound eects on the genetic structure of populations (Jackson 1986), as well as on fundamental evolutionary phenomena such as speciation, selection, and extinction (Cook 1978). If asexual reproduction predominates, populations are generally dominated by a small number of genotypes that are supposedly particularly well adapted to local conditions (Shick et al. 1979; Sebens 1982; Ayre 1983; Neigel and Avise 1983; Stoddart 1984a; Ayre et al. 1991). A clonal life should also accelerate evolutionary change within, and divergence among, local populations (Wright 1978). In contrast, high rates of gene ow and recruitment in strictly sexually-reproducing species reduce habitat selection and stochastic contributions to genetic dierentiation among populations (founder events and genetic drift: Williams 1975; Hedgecock 1986; Underwood and Fairweather 1989; Benzie 1994). Although early evidence suggested a negative correlation between larval dispersal and genetic dierentiation among geographically isolated populations, critical reviews (Burton and Feldman 1982; Hedgecock 1986) have shown the relationship to be less than clear-cut. Recent models have proposed that within species with mixed modes of reproduction, sexually-derived larvae provide the long-distance colonists and a constant gene ow between distant populations, whereas asexual reproduction appears to be the major source of recruitment into local populations (Williams 1975; Bell 1982). The present paper aimed to examine the variation in the genetic composition of six populations of the viviparous scleractinian coral Pocillopora damicornis from four islands separated by up to 650 km in the Ryukyu Archipelago, southern Japan. The results were used to evaluate the relative contributions of sexual and asexual reproduction, and to assess the apparent scale of larval dispersal (gene ow) between coral populations in this region. The Ryukyu Archipelago, located between 24 and 30N, comprises more than 140 islands. The coral reefs bordering these islands are among the northernmost in the world. However, benthic communities, notably scleractinian corals, are highly diverse and abundant. Previous observations by Ayre et al. (1997) showed that life-history traits of Pocillopora damicornis vary with geographical location, which is expressed by the greater importance of asexual reproduction at the margins of its geographical range. Thus, the Ryukyu Islands, roughly aligned in a NESW direction, oer an ideal eld site for studying variation in the genetic composition of corals on a large spatial scale and to assess the relative contribution of sexual and asexual reproduction. The results of this study are expected to be not only fundamental in a purely biological sense, but also of great importance in devising eective methods for the management and

preservation of the biodiversity of coral reef ecosystems. Considering the fact that coral reefs in this region have been disturbed by various natural and man-induced perturbations (Nishihira 1987), the second issue should deserve particular attention.

Materials and methods

Collection of specimens and electrophoresis Pocillopora damicornis is a branching coral, common throughout the tropical and subtropical Indo-Pacic region. Three major modes of reproduction have been documented for this species. (1) Sexual reproduction, followed by broadcast-spawning of planulae; the periodicity of planulae release varies with geographic location (Harriott 1983; Richmond and Jokiel 1984; Ward 1992); P. damicornis is a simultaneous hermaphrodite, with gonads attached to the mesenteries by stalks (Harriott 1983). (2) Fragmentation, which is the major mode of asexual reproduction in the eastern Pacic (Richmond 1987a). (3) Asexual production of planulae; this has been surmised by Stoddart (1983) on the basis of data from allozyme electrophoresis of populations from southwestern Australia and Hawaii. Pocillopora damicornis were collected from six sites around four islands in the Ryukyu Archipelago, two sites around Okinawa Island (Hentona and Abu), two sites around Aka Island (Kushibaru and Nishihama), one site around Sesoko Island, and one site around Ishigaki Island (Fig. 1). Coral reefs in the Ryukyus are relatively narrow, and consist of a shallow reef at (<2 m depth) and a reef slope which generally gives way to a sandy bottom at 920 m depth. Colonies of P. damicornis are observed in a great variety of habitats; however, large populations are restricted to the shallow reef at, and collections were therefore made in this biotope. Each sample consisted of fragments from 48 colonies separated by at least 1 m but over an area of <500 m2. Fragments were kept in dry ice or ice during transport from the eld site to the laboratory, where they were stored at A80 C until analysis. Coral fragments were ground in a solution of 0.04% B mercaptoethanol and 0.01% sucrose. A square of tissue paper (1 cm 1 cm) was placed over the sample to act as a crude lter. Whatman No. 3 chromatography paper wicks (3 mm 10 mm) were used to soak up the exudate. The wicks were placed vertically in a horizontal 12% (w/v) starch gel. Seven polymorphic loci were scored from six enzyme systems: glucose phosphate isomerase (GPI, EC 5.3.1.9.), phosphoglucomutase (PGM, EC 5.4.2.2.), hexokinase (HK, EC 2.7.1.1.), and peptidase (EC 3.4.11.) using leucyl glycil glycil as a substrate (LGG), peptidase using leucyl tyrosine (LT), and peptidase using leucyl proline (LP). GPI, PGM, and HK were run using Tris EDTA citrate (pH 7.9) buer, LGG, LT, and LP were run using Tris EDTA borate (pH 8.4) buer. All gels were run at 200 V and 30 mA for 6 h at 3 C air temperature. Statistical analyses Measures of genetic variability included the mean number of alleles per locus, the percentage of polymorphic loci (95% criterion), the observed heterozygosity (the proportion of heterozygous individuals in the sample), and the heterozygosity expected under conditions of HardyWeinberg equilibrium (gene diversity, H). Values of observed genotypic diversity (Go) were calculated from Go = 1/Sgi2, where gi is the relative frequency of the ith multilocus genotype. Expected genotypic diversity under the conditions of random mating and free recombination (Ge) was calculated following the formula of Stoddart and Taylor (1988). Departure of Go:Ge from unity provides an index of the combined eects of departures from single-locus HardyWeinberg equilibrium and

755 Fig. 1 Locations of six sampling sites of Pocillopora damicornis around four islands in Ryukyu Archipelago, southern Japan (dotted lines approximate extent of developed coral reefs)

linkage disequilibria. Such departures are a predictable consequence of asexual reproduction, but may also result from other factors such as population subdivision. A genetically variable population with high levels of sexually-derived recruitment will display a high ratio of Go:Ge. The number of genotypes detected (Ng) is also an estimate of the minimum number of clones that are present within a sample. The ratio Ng:Ni (where Ni = number of individuals) provides the simplest index of the eects of asexual reproduction on genotypic diversity. FST, a measure of genetic dierentiation among populations, and FIS, a measure of genetic

variation within populations, were calculated using the formula of Weir and Cockerham (1984). The eective number of migrants per generation (Nem) between populations was estimated using Nem = (1/FST A 1)/4. The overall genetic dierentiation between pairs of populations was expressed by Nei's measure of unbiased genetic distance (Nei 1978). Trees were drawn using the neighborjoining method contained in the programme PHYLIP 3.5 c (Felsenstein 1995). The robustness of each node was evaluated by bootstrapping allele frequencies 100 times. Allele frequencies, measures of genetic variability, and genetic distance were

756 calculated using the programme GENETIX 3.3 (Belkhir et al. 1996). Statistical signicance of the values were calculated with permutation tests, and the Bonferroni adjustment of the signicance level for multiple tests was applied using GENETIX 3.3 (Belkhir et al. 1996).

Results
Allele frequencies of the seven loci calculated from the six populations of Pocillopora damicornis are presented in Table 1. Genetic variability within each of the six populations is summarized in Table 2. The seven loci were all polymorphic at the 95% criterion. The mean number of alleles per locus was relatively similar among

populations, ranging between 3.1 and 3.8. The observed heterozygosities were more variable among the six populations, ranging between 0.244 and 0.366. Except for Hentona and Nishihama samples, observed heterozygosities were signicantly lower than those expected under HardyWeinberg equilibrium (permutation tests with Bonferroni adjustments, p < 0.05). Despite some cases of heterozygote excess detected in some populations at GPI, PGM, HK, LT-2, and LP, a consistent decit of heterozygotes, compared with expectations from HardyWeinberg equilibrium, was found for HK, LGG, LT-1, and LP (p < 0.05; Table 3). The ratio of observed genotypic diversity to that expected under a random mating (Go:Ge) ranged between 0.30 and 0.64 (Table 2). The ratio Ng:Ni ranged between 0.47 and 0.75

Table 1 Pocillopora damicornis. Allele frequencies of six populations collected from four islands in Ryukyu Archipelago (Ni number of individuals; allele absent) Locus, allele GPI 1 2 3 4 5 6 PGM 1 2 3 4 5 6 HK 1 2 3 4 5 6 LGG 1 2 3 4 5 LT-1 1 2 3 4 LT-2 1 2 3 4 LP 1 2 3 Ni Sesoko, Sesoko Island 0.010 0.084 0.021 0.875 0.010 0.177 0.688 0.135 0.094 0.063 0.167 0.572 0.094 0.010 0.771 0.187 0.042 0.854 0.115 0.031 0.094 0.844 0.062 0.937 0.063 48 Hentona, Okinawa Island 0.188 0.812 0.094 0.094 0.167 0.583 0.062 0.250 0.167 0.135 0.427 0.021 0.146 0.083 0.719 0.021 0.031 0.896 0.031 0.063 0.010 0.083 0.031 0.865 0.021 0.979 0.021 48 Abu, Okinawa Island 0.146 0.771 0.083 0.115 0.208 0.542 0.115 0.020 0.156 0.209 0.469 0.031 0.135 0.021 0.646 0.021 0.312 0.875 0.042 0.083 0.917 0.083 0.063 0.937 48 Kushibaru, Aka Island 0.167 0.521 0.312 0.490 0.333 0.177 0.031 0.187 0.615 0.167 0.062 0.792 0.073 0.073 0.906 0.042 0.052 0.010 0.834 0.156 0.948 0.052 48 Nishihama, Aka Island 0.042 0.635 0.042 0.281 0.552 0.250 0.198 0.198 0.490 0.062 0.250 0.083 0.782 0.083 0.052 0.948 0.052 0.083 0.875 0.042 0.917 0.083 48 Ishigaki, Ishigaki Island 0.146 0.156 0.052 0.604 0.042 0.209 0.510 0.146 0.135 0.760 0.208 0.032 0.063 0.770 0.063 0.104 0.875 0.052 0.073 0.135 0.844 0.021 0.948 0.052 48

757 Table 2 Pocillopora damicornis. Genetic variation within six populations collected from four islands in Ryukyu Archipelago (Go observed genotypic diversity; Ge expected genotypic diversity; Ng observed no. of genotypes; Ni no. of individuals) Statistical Population Mean no. alleles/locus 3.5 3.8 3.4 3.1 3.1 3.4 Heterozygosity (SD) Direct count 0.244 0.363 0.303 0.273 0.366 0.285 (0.132)* (0.251)NS (0.175)** (0.174)* (0.255)NS (0.162)* Expected 0.334 0.365 0.383 0.384 0.373 0.371 (0.169) (0.237) (0.230) (0.212) (0.221) (0.197) 0.30 0.64 0.41 0.37 0.41 0.43 0.47 0.75 0.58 0.62 0.64 0.62 dierences between direct count and expected heterozygoties were calculated with permutation tests followed by Bonferroni adjustments (NS p > 0.05; *p < 0.05; **p < 0.01) Go:Ge Ng:Ni

Sesoko Hentona Abu Kushibaru Nishihama Ishigaki

Table 3 Pocillopora damicornis. Weir and Cockerham (1984) F-statistics by locus for six populations from Ryukyu Archipelago. FST standardised genetic variance between populations; FIS standardised genetic variance within populations. Statistical signicances of values were calculated with permutation tests followed by Bonferroni adjustments. NS p > 0.05; *p < 0.05; **p < 0.01 Locus GPI PGM HK LGG LT-1 LT-2 LP Average FST 0.111** 0.048** 0.069** 0.040** 0.010NS 0.018NS 0.004NS 0.056** FIS 0.116NS 0.099NS 0.174** 0.288** 0.299** 0.161NS 0.343** 0.180**

(Table 2), and in each population the number of observed similar genotypes was signicantly higher than the number of similar genotypes that can be obtained by chance (permutation tests with Bonferroni adjustments, p < 0.05). Both of these measures (Go:Ge and Ng:Ni) suggested a high degree of asexual reproduction within the six populations examined. Populations were not, however, dominated by a single or a small number of clones.

FST values were signicantly dierent from 0 at all loci except LT-1 and LP, and FST calculated for the seven loci and the six populations indicated a signicant genetic dierentiation (permutation tests with Bonferroni adjustments, p < 0.05; Table 3). FST calculated between pairs of populations indicated a signicant genetic differentiation (Table 4). The highest FST values were found between the Kushibaru sample and samples collected around Sesoko Island and Okinawa (Hentona and Abu), whereas the lowest value was found between Sesoko and Hentona samples. The number of migrants per generation (Nem) calculated from FST values is valid only at equilibrium but, making this assumption, our estimations ranged between 2.4 and 8.9 migrants per generation (Table 4). The genetic distances between populations ranged between 0.016 and 0.066 (average 0.037; Table 4). As shown in the neighbor-joining tree in Fig. 2, the highest similarities were found between populations from Okinawa Island (Hentona and Abu), and between populations from Aka Island (Kushibaru and Nishihama). The population from Sesoko Island was close to the populations from Okinawa Island, whereas the population from Ishigaki Island was clearly distinct from the other populations. However, there was no signicant correlation between genetic and geographical distances between populations (r2 = 0.006, p > 0.05).

Table 4 Pocillopora damicornis. Weir and Cockerham (1984) FST (above diagonal), average number of migrants per generation (Nem, in parentheses above diagonal ), and Nei's unbiaised genetic distance (below diagonal) between six populations collected from four Sesoko Sesoko Hentona Abu Kushibaru Nishihama Ishigaki 0.016 0.022 0.046 0.034 0.026 Hentona 0.027** (8.9) 0.022 0.066 0.050 0.044 Abu 0.035* (6.7) 0.033** (7.3) 0.057 0.036 0.039

islands in Ryukyu Archipelago. Statistical signicances of FST values were calculated with permutation tests followed by Bonferroni adjustments (NS p > 0.05; *p < 0.05; **p < 0.01) Kushibaru 0.070** (3.2) 0.092** (2.4) 0.078** (2.9) 0.030 0.031 Nishihama 0.056** (4.1) 0.076** (3.0) 0.052** (4.5) 0.044** (5.3) 0.040 Ishigaki 0.041** (5.7) 0.066** (3.5) 0.056* (4.1) 0.043** (5.4) 0.059** (3.9)

758

Fig. 2 Pocillopora damicornis. Neighbor-joining tree showing pattern of similarity expressed by Nei's unbiased genetic distance among six populations collected around four islands. Bootstrap values are from a consensus tree

Discussion
Genotypic diversity, the number of observed genotypes, and deviations from HardyWeinberg equilibrium, which include both excesses and decits of heterozygotes, clearly demonstrated that asexual reproduction plays a major role in the maintenance of established populations of Pocillopora damicornis in the Ryukyu Islands. These results are concordant with those obtained for 25 populations of P. damicornis from southwestern Australia (average Go:Ge = 0.40; Stoddart 1984a, b). In contrast, Benzie et al. (1995) and Ayre et al. (1997) found a predominance of sexual reproduction (average Go:Ge = 0.91 and 0.88, respectively) in populations of P. damicornis from One Tree Island (Central GBR), and along the GBR, respectively. Therefore, our results conrm that the relative importance of asexual reproduction in P. damicornis varies geographically, and support the assumption that asexual reproduction has a higher importance at the margin of the geographical range of the species. The relative importance of sexual and asexual reproduction is probably related to the physical and ecological characteristics of the reef considered. Despite the high contribution of asexual reproduction, the populations surveyed in the present study were not completely dominated by a single or a few clones, and most clones were represented by only a few individual samples, as also found for the zoanthid Zoanthus coppingeri by Burnett et al. (1995). In contrast, populations of P. damicornis from southwestern Australia (Stoddart 1984a) were dominated by few clones, as found for the sea anemone Actinia tenebrosa by Black and Johnson (1979). Fragmentation (Highsmith 1982) and asexual production of planulae (Stoddart 1983) are the two major modes of asexual reproduction reported for Pocillopora damicornis. A colony produced by fragmentation is genetically indistinguisable from a colony derived from asexually-produced larvae, and therefore it is uncertain which mode is predominant. However, two factors suggest that fragmentation is likely to be a major mode of asexual reproduction in populations of the Ryukyu Islands, as was found in the eastern Pacic (Richmond

1987a), and for other scleractinians such as Pavona cactus (Willis and Ayre 1985; Ayre and Willis 1988), Montipora digitata (Jokiel et al. 1983), Acropora cervicornis (Neigel and Avise 1983), and A. aspera, A. pulchra, A. robusta (Bothwell 1981): (1) All populations were collected on the shallow reef at (<2 m depth), a habitat which is temporary exposed to wave action and where fragmentation is likely to be facilitated; (2) The area from the Ryukyus to the northern Philippines is subject to the highest frequency of typhoons in the world, with generally more than three typhoons occuring every year between June and October. In August 1996, we observed in Hentona and Abu (Okinawa Island) a high number of coral fragments among foliaceous and branching species such as Pocillopora damicornis one week after a severe typhoon. Despite the fact that previous observations indicated that fragments generated on shallow, wave-swept reefs rarely re-attach and survive (Ayre et al. 1997), we believe that fragmentation is occurring in the Ryukyus because periods of high disturbance are generally followed by calm periods that we suppose to be consistent with the successful establishment of new colonies. However, the relationship between rate of disturbances and clonal diversity is far from clear in coral reef ecosystems. For the coral Porites compressa in Hawaii, the lowest amount of clonal proliferation was found in populations that had been intensely or recently disturbed (Hunter 1993), because these populations were in early stages of recolonization, and space was not limited. Karlson (1991) also found a decrease in clonal propagation with increasing levels of disturbances in zoanthid species. The important contribution of fragmentation that we inferred in the present study contrasts with the results of Stoddart (1984a), who suggested that asexually derived planulae were responsible for the major part of the local asexual proliferation of clones in populations from southwestern Australia. However, our results do not mean that asexual production of planula is not eective in these populations. A further study is necessary to determine the relative contribution of the two modes of asexual reproduction in populations of the Ryukyus. Signicant genetic dierentiation was found between populations of Pocillopora damicornis separated by up to 650 km in the Ryukyu Archipelago (mean FST = 0.056, p < 0.01). Despite the fact that the population of Ishigaki Island was clearly distinct in the neighbor-joining tree, the genetic distance among our samples did not correlate with geographical distance, suggesting that there was no signicant isolation by distance among populations of P. damicornis in the Ryukyu Islands over the spatial scale relevant to the present study. The level of genetic dierentiation that we found was lower than that found in populations of the same species from southwestern Australia (mean FST = 0.213; Stoddart 1984b), or in Seriatopora hystrix from the Great Barrier Reef (mean FST = 0.43; Ayre and Dufty 1994). The level of genetic dierentiation in populations from the Ryukyus was fairly similar to

759

values obtained for populations of P. damicornis separated by up to 1200 km along the GBR (FST = 0.084; Ayre et al. 1997). Investigations of various invertebrates have demonstrated a broad range in the level of genetic dierentiation, from populations that are highly genetically subdivided even where they have the potential for long-distance dispersal (sea anemones: Ayre 1983, 1984, Homan 1986, Ayre et al. 1991; corals: Ayre and Dufty 1994) to populations with little genetic dierentiation even when separated by thousands of kilometers (asteroid: Williams and Benzie 1996). In contrast to Ayre et al. (1997), who concluded that populations of Pocillopara damicornis along the GBR were connected by high gene ow, we conclude that the gene ow between populations examined in the Ryukyus is relatively moderate, despite the dispersive capability of the planulae and the physical connection between the sampling sites. In the laboratory, larvae of P. damicormis were kept alive for up to 212 d (Harrigan 1972), and Richmond (1987b) demonstrated that larvae can settle and metamorphose after 103 d. Post-metamorphic rafting is also a possible means of long-distance dispersal. Colonies of Pocillopora sp. were found attached to pumice and Japanese shing oats by Jokiel (1984), and calculating from coral growth-rates, he concluded that these attached colonies could have traversed a total distance from 20 000 to 40 000 km during their lifetime. These oating colonies could also produce larvae along the route and thereby colonize outlying islands. With 140 islands separated by <300 km, and with the general southnorth circulation caused by the major oceanic current, the Kuroshio, the Ryukyu Archipelago should favour the connection of populations through the occasional movement of larvae from one island to the next in a stepping-stone fashion across generations. In addition, more localized currents and weather-dependent currents also seem to favour the connection of islands along other directions. A large number of oating plastic cards, deposited in Aka Island to estimate the dispersal of larval slick during a mass-spawning, were collected in Okinawa (T. Hayashibara unpublished data), indicating that transport of larvae perpendicular to the main oceanic current is also possible. Thus, given the apparent long-distance dispersal capabilities of the planulae, and the geography and circulation patterns of the Ryukyu Archipelago, one might predict a high gene ow and a non-signicant genetic dierentiation between populations of P. damicornis in the Ryukyu Islands. Moderate dierentiation among separated populations has also been attributed to recent geographical expansion and the subsequent expansion in population size (Palumbi and Wilson 1990). This hypothesis deserves attention, since coral reefs in the Ryukyus are relatively recent in a geological time. Our knowledge of the geological history of the area is, however, not sucient to examine this hypothesis in detail. Given the apparent similarity of habitats from which we collected samples, dierential selection pressures arising from environmental dierences probably do not play a major

role and are not sucient to generate a high level of genetic dierentiation. However, as pointed out by Stoddart (1984b), many environmental parameters can not be easily measured or indeed comprehended at all, and it is therefore dicult to say if selection pressures acting on populations are uniform or not. Our results are in accordance with previous models of dispersal and population maintenance in species with mixed modes of reproduction (Williams 1975; Bell 1982). In populations of Pocillopora damicornis of the Ryukyu Islands and southwestern Australia, in populations of Pavona cactus from the GBR (Ayre and Willis 1988), and other anthozoans such as sea anemones (Ayre et al. 1991), asexual reproduction plays a primary role in the maintenance of local populations, although sexual reproduction is likely to be the primary source of episodic and successful colonization by larvae that are genetically diverse, and the means of long-distance gene ow.
Acknowledgements We gratefully acknowledge the following persons: K. Hashimoto, T. Shibuno, O. Abe, the sta of Ishigaki Tropical Station, the sta of Akajima Marine Science Laboratory, the sta of Iriomote Station, the sta of the Department of Chemistry, Biology and Marine Science of the University of the Ryukyus, and the sta of Sesoko Station for logistic support; and J. Benzie, G. Beretta, B. Bilgre, F. Bonhomme, D. Booth, L. Cadoret, M. Hidaka, P. Lenfant, S. Nakamura, H. Ota, S. Planes, K. Sakai, B. Salvat, H. Shimajiri, A. Takemura, M. Toda, S. Williams for their help and advice; and two anonymous reviewers for valuable comments and suggestions. M. Adjeroud was supported by a postdoctoral fellowship from the Japan Society for the Promotion of Science.

References
Ayre DJ (1983) The eects of asexual reproduction and intergenotypic aggression on the genotypic structure of populations of the sea anemone Actinia tenebrosa. Oecologia 57: 158165 Ayre DJ (1984) The eects of sexual and asexual reproduction on geographic variation in the sea anemones Actinia tenebrosa. Oecologia 62: 222229 Ayre DJ, Dufty S (1994) Evidence for restricted gene ow in the viviparous coral Seriatopora hystrix on Australia's Great Barrier Reef. Evolution 48: 11831201 Ayre DJ, Hughes TP, Standish RJ (1997) Genetic dierentiation, reproductive mode, and gene ow in the brooding coral Pocillopora damicornis along the Great Barrier Reef, Australia. Mar Ecol Prog Ser 159: 175187 Ayre DJ, Read J, Wishart J (1991) Genetic subdivision within the eastern Australian population of the sea anemone Actinia tenebrosa. Mar Biol 109: 379390 Ayre DJ, Willis BL (1988) Population structure in the coral Pavona cactus: clonal genotypes show little phenotypic plasticity. Mar Biol 99: 495505 Belkhir K, Borsa P, Goudet J, Chikhi L, Bonhomme F (1996) ne tique des GENETIX, logiciel sous Windows TM pour la ge nome et Populations, CNRS UPR populations. Laboratoire Ge de Montpellier II, Montpellier (France) 9060, Universite Bell G (1982) The masterpiece of nature: the evolution and genetics of sexuality. Croon Helm, London Benzie JAH (1994) Pattern of gene ow in the Great Barrier Reef and Coral Sea. In: Beaumont AR (ed) Genetics and evolution of aquatic organisms. Chapman & Hall, London, pp 6779 Benzie JAH, Haskell A, Lehman H (1995) Variation in the genetic composition of coral (Pocillopora damicornis and Acropora

760 palifera) populations from dierent reef habitats. Mar Biol 121: 731739 Black R, Johnson MS (1979) Asexual viviparity and population genetics of Actinia tenebrosa. Mar Biol 53: 2731 Bothwell AM (1981) Fragmentation, a means of asexual reproduction and dispersal in the coral genus Acropora a preliminary report. Proc 4th int coral Reef Symp 1: 137144 [Gomez ED et al. (eds) Marine Sciences Center, University of the Philippines, Quezon City, Philippines] Burnett WJ, Benzie JAH, Beardmore JA, Ryland JS (1995) Patterns of genetic subdivision in populations of a clonal cnidarian, Zoanthus coppingeri, from the Great Barrier Reef. Mar Biol 122: 665673 Burton RS, Feldman MW (1982) Population genetics of coastal and estuarine invertebrates: does larval behavior inuence population structure? In: Kennedy VS (ed) Estuarine comparisons. Academic Press, New York, pp 537551 Carlon DB, Olson RR (1993) Larval dispersal distance as an explanation for adult spatial pattern in two Caribbean reef corals. J exp mar Biol Ecol 173: 247263 Cook RE (1978) Asexual reproduction: a further consideration. Am Nat 113: 769772 Fadlallah YH (1983) Sexual reproduction, development, and larval biology in scleractinian corals a review. Coral Reefs 2: 129 150 Felsenstein J (1995) PHYLIP (Phylogeny Inference Package) Version 3.5 c. Department of Genetics, University of Washington, Seattle (Distributed by the author) Harrigan JF (1972) The planula larvae of Pocillopora damicornis: lunar periodicity of swarming and substratum selection behaviour. PhD thesis. University of Hawaii, Honolulu Harriott VJ (1983) Reproductive seasonality, settlement, and postsettlement mortality of Pocillopora damicornis (Linnaeus), at Lizard Island, Great Barrier Reef. Coral Reefs 2: 151157 Harrison PL, Wallace CC (1990) Reproduction, dispersal and recruitment of scleractinian corals. In: Dubinsky Z (ed) Ecosystems of the world. 25. Coral reefs. Elsevier, Amsterdam, pp 133196 Hedgecock D (1986) Is gene ow from pelagic larval dispersal important in the adaptation and evolution of marine invertebrates? Bull mar Sci 39: 550564 Highsmith RC (1982) Reproduction by fragmentation in corals. Mar Ecol Prog Ser 7: 207226 Homan RJ (1986) Variation in contribution of asexual reproduction to the genetic structure of populations of the sea anemone Metridium senile. Evolution 40: 357365 Hunter CL (1993) Genotypic variation and clonal structure in coral populations with dierent disturbance histories. Evolution 47: 12131228 Jackson JBC (1986) Modes of dispersal of clonal benthic invertebrates: consequences for species' distributions and genetic structure of local populations. Bull mar Sci 39: 588606 Jokiel PL (1984) Long distance dispersal of reef corals by rafting. Coral Reefs 3: 113116 Jokiel PL, Hildemann WH, Bigger CH (1983) Clonal population structure of two sympatric species of the reef coral Montipora. Bull mar Sci 33: 181187 Karlson RH (1991) Fission and the dynamics of genets and ramets in clonal cnidarian populations. Hydrobiologia 216/217: 235 240 Kramarsky-Winter E, Fine M, Loya Y (1997) Coral polyp expulsion. Nature, Lond 387: p. 137 Nei M (1978) Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics, Austin, Tex 89: 583590 Neigel JE, Avise JC (1983) Clonal diversity and population structure in a reef building coral, Acropora cervicormis: self recognition analysis and demographic interpretation. Evolution 37: 437453 Nishihira M (1987) Natural and human interference with the coral reef and coastal environments in Okinawa. Galaxea 6: 311321 Palumbi SR, Wilson AC (1990) Mitochondrial DNA diversity in the sea urchins Strongylocentrotus purpuratus and S. droebachiensis. Evolution 44: 403415 Richmond RH (1987a) Energetic relationships and biogeographical dierences among fecundity, growth and reproduction in the reef coral Pocillopora damicornis. Bull mar Sci 34: 280287 Richmond RH (1987b) Energetics, competency, and long-distance dispersal of planula larvae of the coral Pocillopora damicornis. Mar Biol 93: 527533 Richmond RH, Jokiel PL (1984) Lunar periodicity in larva release in the reef coral Pocillopora damicornis at Enewetak and Hawaii. Bull mar Sci 34: 280287 Sebens KP (1982) Asexual reproduction in Anthopleura elegantissima (Anthozoa: Actinaria) seasonality and spatial extent of clones. Ecology 63: 434444 Shick JM, Homann RJ, Lamb AN (1979) Asexual reproduction, population structure and genotypeenvironment interactions in sea anemones. Am Zool 19: 699713 Stoddart JA (1983) Asexual production of planulae in the coral Pocillopora damicornis. Mar Biol 76: 279284 Stoddart JA (1984a) Genetical structure within populations of the coral Pocillopora damicornis. Mar Biol 81: 1930 Stoddart JA (1984b) Genetic dierentiation amongst populations of the coral Pocillopora damicornis of southwestern Australia. Coral Reefs 3: 149156 Stoddart JA, Babcock RC, Heyward AJ (1988) Self-fertilization and maternal enzymes in the planulae of the coral Goniastrea favulus. Mar Biol 99: 489494 Stoddart JA, Taylor JF (1988) Genotypic diversity: estimation and prediction. Genetics, Austin, Tex 118: 705711 Underwood AJ, Fairweather PG (1989) Supply-side ecology and benthic marine assemblages. Trends Ecol Evolut 4: 1620 Ward S (1992) Evidence for broadcast spawning as well as brooding in the scleractinian coral Pocillopora damicornis. Mar Biol 112: 641646 Weir BS, Cockerham CC (1984) Estimating F-statistics for the analysis of population structure. Evolution 38: 13581370 Williams GC (1975) Sex and evolution. Princeton University Press, Princeton Williams ST, Benzie JAH (1996) Genetic uniformity of widely separated populations of the coral reef starsh Linckia laevigata from the East Indian and West Pacic Oceans, revealed by allozyme electrophoresis. Mar Biol 126: 99107 Willis BL, Ayre DJ (1985) Asexual reproduction and genetic determination of growth in the coral Pavona cactus: biochemical genetic and immunogenic evidence. Oecologia 65: 516525 Wright S (1978) Evolution and the genetics of populations, Vol. 4. Variability within and among natural populations. University of Chicago Press, Chicago