Genetic differentiation of bicolor damselfish (Eupomacentrus partitus) populations in the Florida Keys

Electrophoretic variation in proteins encoded by 23 loci revealed substantial genetic differentiation among populations of bicolor damselfish (Eupomacentrus partitus) collected from four coral reefs in the Florida Keys, USA, during 1986–1988. Genetic differentiation was concentrated between a sample collected from Little Grecian Rocks Reef (LGR) and the remaining samples, including fish from a reef only 600 m distant (Grecian Rocks Reef). Genetic distinction of the LGR sample derived from significantly heterogenous allelic frequencies at six of eight polymorphic loci. Aco-1 (aconitase); Ada (adenosine deaminase); Gpi-2 (glucosephosphate isomerase); Ldh-2 and Ldh-3 (lactate dehydrogenase); and Me-1 (malic enzyme); nevertheless, differentiation at cytosolic aconitase (Aco-1) far exceeded that observed for other loci (fixation index, F _ST=0.482), and differences in Aco-1 allele frequencies were largely responsible for large genetic distances (0.20) between LGR and the other reefs. Paradoxically, estimates of numbers of migrants exchanged between reefs per generation (mN _e=17.47) indicated the potential for extensive gene flow. The extent of genetic differentiation among these populations is evaluated relative to models of population genetic structure based on equilibrium between gene flow and natural selection or genetic drift.     

Genetic differentiation among populations of a broadcast spawning soft coral, Sinularia flexibilis, on the Great Barrier Reef

The genetic structure of 12 reef populations of the soft coral Sinularia flexibilis (Octocorallia, Alcyoniidae) was studied along the Great Barrier Reef (GBR) at a maximum separation of 1,300 km to investigate the relative importance of sexual and asexual reproduction, genetic differentiation and gene flow among these populations. S. flexibilis is a widely distributed Indo-Pacific species and a gamete broadcaster that can form large aggregations of colonies on near-shore reefs of the GBR. Up to 60 individuals per reef were collected at a minimum sampling scale of 5 m at two sites per reef, from December 1998 to February 2000. Electrophoretic analyses of nine polymorphic allozymes indicated that genotypic frequencies in most populations and loci did not differ significantly from those expected from Hardy–Weinberg predictions. Analysis of multi-locus genotypes indicated a high number of unique genotypes (N _go) relative to the number of individuals sampled (N) in each reef population (range of 0.69–0.95). The maximum number of individuals likely to have been produced sexually (N*) was similar to the number of individuals sampled (i.e. N*:N ˜ 1), suggesting that even repeated genotypes may have been produced sexually. These results demonstrated a dominant role of sexual reproduction in these populations at the scale sampled. Significant genetic differentiation between some populations indicated that gene flow is restricted between some reefs (F _ST=0.026, 95% CI= 0.011 − 0.045) and even between sites within reefs (F _ST=0.041, 95% CI=0.027 − 0.055). Nevertheless, there was no relationship between geographic separation and genetic differentiation. Analyses comparing groups of populations showed no significant differentiation on a north-south gradient in the GBR. The pattern in the number of significant differences in gene frequencies in pairwise population comparisons, however, suggested that gene flow may be more restricted among inner-shelf reef populations near to the coast than among mid/outer-shelf populations further from the coast. Received: 10 July 2000 / Accepted: 5 October 2000     

Protection of Genetic Diversity and Maintenance of Connectivity among Reef Corals within Marine Protected Areas

Abstract: High‐latitude coral reefs (HLRs) are potentially vulnerable marine ecosystems facing well‐documented threats to tropical reefs and exposure to suboptimal temperatures and insolation. In addition, because of their geographic isolation, HLRs may have poor or erratic larval connections to tropical reefs and a reduced genetic diversity and capacity to respond to environmental change. On Australia's east coast, a system of marine protected areas (MPAs) has been established with the aim of conserving HLRs in part by providing sources of colonizing larvae. To examine the effectiveness of existing MPAs as networks for dispersal, we compared genetic diversity within and among the HLRs in MPAs and between these HLRs and tropical reefs on the southern Great Barrier Reef (GBR). The 2 coral species best represented on Australian HLRs (the brooding Pocillopora damicornis and the broadcast‐spawning Goniastrea australensis) exhibited sharply contrasting patterns of diversity and connectedness. For P. damicornis, the 8‐locus genetic and genotypic diversity declined dramatically with increasing latitude (N_a= 3.6–1.2, H_e= 0.3–0.03, N_g:N = 0.87–0.06), although population structure was consistent with recruitment derived largely from sexual reproduction (G_o:G_e= 1.28–0.55). Genetic differentiation was high among the HLRs (F_ST[SD]= 0.32 [0.08], p < 0.05) and between the GBR and the HLRs (F_ST= 0.24 [0.06], p < 0.05), which indicates these temperate populations are effectively closed. In contrast for G. australensis, 9‐locus genetic diversity was more consistent across reefs (N_a= 4.2–3.9, H_e= 0.3–0.26, N_g:N = 1–0.61), and there was no differentiation among regions (F_ST= 0.00 [0.004], p > 0.05), which implies the HLRs and the southern GBR are strongly interconnected. Our results demonstrate that although the current MPAs appear to capture most of the genetic diversity present within the HLR systems for these 2 species, their sharply contrasting patterns of connectivity indicate some taxa, such as P. damicornis, will be more vulnerable than others, and this disparity will provide challenges for future management.     

Genetic consequences of long larval life in the starfish Linckia laevigata (Echinodermata: Asteroidea) on the Great Barrier Reef

Gene flow between populations of the asteroid Linckia laevigata (Linnaeus) was investigated by examining over 1000 individuals collected from ten reefs throughout the Great Barrier Reef (GBR), Australia, for genetic variation at seven polymorphic enzyme loci. Despite geographic separations in excess of 1000 km, Nei's unbiased genetic distance (0 to 0.003) and standardised genetic variation between populations (F _ST) values (mean 0.0011) were small and not significant. Genetic homogeneity among L. laevigata populations is consistent with the long-distance dispersal capability of its 28 d planktonic larval phase, and is greater than that observed for other asteroid species, including another high-dispersal species, Acanthaster planci, which has a 14 d larval phase. Variation within populations was also higher than previously recorded for asteroids (mean heterozygosity=0.384; number of alleles per locus ranged from 5.1 to 6.0 in each population). Among asteroids, dispersal ability is positively correlated with gene flow and levels of variation, and negatively correlated with levels of differentiation.     

Genetic structure of<Emphasis Type="Italic"> Linckia laevigata</Emphasis> and<Emphasis Type="Italic"> Tridacna crocea</Emphasis> populations in the Palawan shelf and shoal reefs

Allozyme variation of 10 populations of Linckia laevigata at 8 polymorphic loci and 13 populations of Tridacna crocea at 6 polymorphic loci were analyzed to compare genetic variability and genetic affinities among reefs in Palawan, Philippines. Two to five populations were sampled from each of four regions: the shelf reefs in (1) northern Palawan and (2) southern Palawan and the offshore reefs in (3) the Kalayaan island group (KIG) in the South China Sea and (4) the Tubbataha shoals in the Sulu Sea. Heterozygosity was highest in populations of L. laevigata from the south shelf of Palawan and populations of T. crocea from the Tubbataha shoals of the Sulu Sea. The lowest heterozygosity estimates were from the reefs of the KIG in the South China Sea, for both species. Overall F _ST values for both species were significant, with an estimated average number of effective migrants per generation (N _EM) of 4.85 (~5 individuals) for L. laevigata and 3.54 (~4 individuals) for T. crocea. Within-region comparisons showed N_EM ranging from 6.29 to 92.34 for L. laevigata and from 3.40 to 6.30 for T. crocea. The higher gene flow among L. laevigata populations relative to T. crocea is consistent with the greater dispersal potential of the former species. Finer scale genetic structuring was evident in T. crocea populations. For both species, the Tubbataha reefs in the Sulu Sea have higher genetic affinity with the populations from the southern shelf of Palawan, while the reefs in the KIG had higher affinity with the northern Palawan shelf reefs. The north and south shelf populations have the least genetic affinity. Genetic patchiness among reefs within regions suggests the importance of small-scale physical factors that affect recruitment success in structuring populations in small island and shoal reef systems in Palawan.     

Method of estimating the standing stock of Trochus niloticus incorporating Landsat satellite data,with application to the trochus resources of the Bourke Isles,Torres Strait,Australia

The habitat of Trochus niloticus in the Bourke Isles, Torres Strait, was classified into areas of a Landsat image, using high-ratio values of green (Band 2) to red (Band 3) light, along the windward reef margins. These shallow-water (< 15 m) areas have a coral and rubble/algal pavement cover, which constitutes the optimal habitat for this gastropod. The habitat was sampled to estimate the abundance of T. niloticus. The proportion of commercial-sized individuals was estimated by measuring the basal width of all individuals in a sample. A multistage sample design incorporating three spatial scales -100 m² (transect), 1500 m² (site) and 1 km² (reef) — was used to provide variance estimates for sample-design optimisation and to provide data on the spatial variation of abundance. Most variation (68%) in abundance was within reefs and was attributable to differences in reef cover. Variations in abundance and time costs for sampling 2 and 4 m transects were compared; the 2 m transect was more efficient than the 4 m transect. The abundance estimates were combined with habitat-area estimates and the proportion of commerical-sized individuals was estimated at a standing stock of 186000 (24% precision), or 14 t of commerical-sized T. niloticus.     

Population differentiation of the Atlantic spotted dolphin (Stenella frontalis) in the western North Atlantic, including the Gulf of Mexico

Information about the genetic population structure of the Atlantic spotted dolphin [Stenella frontalis (G. Cuvier 1829)] in the western North Atlantic would greatly improve conservation and management of this species in USA waters. To this end, mitochondrial control region sequences and five nuclear microsatellite loci were used to test for genetic differentiation of Atlantic spotted dolphins in the western North Atlantic, including the Gulf of Mexico (n=199). Skin tissue samples were collected from 1994–2000. Significant heterozygote deficiencies in three microsatellite loci within samples collected off the eastern USA coast prompted investigation of a possible Wahlund effect, resulting in evidence for previously unsuspected population subdivision in this region. In subsequent analyses including three putative populations, two in the western North Atlantic (n=38, n=85) and one in the Gulf of Mexico (n=76), significant genetic differentiation was detected for both nuclear DNA (R _ST=0.096, P≤0.0001) and mitochondrial DNA (Φ _ST=0.215, P≤0.0001), as well as for all pair-wise population comparisons for both markers. This genetic evidence for population differentiation coupled to known biogeographic transition zones at Cape Hatteras, North Carolina and Cape Canaveral, Florida, USA, evidence of female philopatry, and preliminary support for significant genetic differences between previously documented morphotypes of Atlantic spotted dolphins in coastal and offshore waters all indicate that the biology and life history of this species is more complex than previously assumed. Assumptions of large, panmictic populations might not be accurate in other areas where S. frontalis is continuously distributed (e.g., eastern Atlantic), and could have a detrimental effect on long-term viability and maintenance of genetic diversity in this species in regions where incidental human-induced mortality occurs.

---

Genetic population structure in two tropical sponge-dwelling shrimps that differ in dispersal potential

The spatial context in which host races of parasitic animals originate is a central issue in the controversial theory of sympatric speciation. Sponge-dwelling shrimps in the genus Synalpheus provide a good system for evaluating the possibility of resource-associated divergence in sympatry. I used allozyme electrophoresis to assess the genetic population structure of two Caribbean Synalpheus species sampled in 1988 to 1990 at a hierarchy of spatial scales. S. brooksi Coutière is a host-generalist, using several sponge species in an area, and develops directly, with no planktonic larval stage. G-tests and estimates of F _ST revealed highly structured populations in this species, with significant differentiation among samples from individual reefs within a region, and strong divergence among regions (Panama, Belize, Florida). Moreover, samples of S. brooksi taken from the two sponges Spheciospongia vesparium (Lamarck) and Agelas clathrodes (Schmidt) in Panama, and separated by 3 km, showed significant differentiation at both of the loci that were polymorphic in these populations. Genetic distances between these host-associated populations averaged >60% greater than distances between samples from the same host species and were comparable to, or greater tha, those for some inter-regional comparisons. These genetic data corroborate a previous finding of demographic differences between the same populations. The second species, S. pectiniger Coutière, occurs only in Spheciospongia vesparium, and produces swimming larvae. Although allele frequencies in this species differed significantly among the three regions, S. pectiniger showed no differentiation within regions, and significantly lower differentiation (F _ST) among regions than its direct-developing congener. These data suggest that genetic population structure in these two commensal crustaceans is related to dispersal potential, and that restricted dispersal may allow the divergence of host-associated populations on a local scale.     

Molecular genetic structure suggests limited larval dispersal in grass rockfish,<Emphasis Type="Italic"> Sebastes rastrelliger</Emphasis>

The grass rockfish (Sebastes rastrelliger Jordan and Gilbert, 1880) is a non-migratory, benthic, near shore species distributed along rocky reefs and sea grass beds. It occurs from Baja California, Mexico, to Oregon, USA, spanning the Oregonian and Californian biogeographic provinces. In California this fish receives intense fishing pressure from an expanding and loosely regulated live-fish fishery. It is not known whether or by what mechanism larvae and juveniles are retained locally or dispersed widely during the early life-history phase. Tissue samples of S. rastrelliger were collected between 1996 and 2001 from 405 adult fish at eight sites (42.70°N; 124.50°W to 32.67°N; 117.25°W) spanning the species range. Individuals were surveyed for polymorphism at six microsatellite loci. Allele frequency heterogeneity was not significant among all sites (F_ST=0.001; P=0.18), nor in pairwise comparisons, but a clear correlation of genetic and geographic distance was detected (P=0.019). Fit of genetic and geographic distance was stronger within biogeographic provinces than at a range-wide scale, suggesting that populations north and south of the Point Conception biogeographic boundary are not in equilibrium with respect to migration and genetic drift. Estimates of mean coastal dispersal distances associated with the isolation-by-distance relationship are on the order of 10 km generation^–1. Such limited dispersal in a species with a pelagic early life history suggests active retention mechanisms near the shore. This has important implications for coastal management zones and design of marine reserves.Communicated by J.P. Grassle, New Brunswick     

No apparent genetic basis to demographic differences in scarid fishes across continental shelf of the Great Barrier Reef

 Two species of parrot fish, Scarus frenatus and Chlorurus sordidus, are known to exhibit demographic and life-history differences across the continental shelf of the northern Great Barrier Reef (GBR). DNA sequences from the mitochondrial (mt) control region were analysed to determine whether there were genetic differences between the populations from the mid- and outer-shelf reefs. Analysis of molecular variance (AMOVA) indicated high levels of gene exchange for both species at a local scale between reefs on mid- and outer continental shelf positions (20 km apart) and at a broader scale along the length of the GBR province (>1000 km apart). There was no evidence to suggest that local differences in scarid life-history characteristics on the northern GBR have a genetic basis. Rather it appears more likely that phenotypically plastic responses to prevailing social and environmental conditions explain differences in the life-history characteristics of both taxa. However, analysis of genetic variation and historical demography revealed striking differences between the two species. S. frenatus haplotypes differed from one another at relatively few nucleotide sites (mean = 3.30), and the pairwise mismatch distribution suggested this species has undergone a population expansion within the limit of the resolution of the marker. C. sordidus haplotypes, however, differed from one another at a number of sites (mean = 7.67). Mismatch distribution analysis suggested that the population size of this species has remained at equilibrium over time. These patterns could also reflect differences in the metapopulation sizes or generation times between taxa. Some of the implications for fisheries management are discussed. Received: 28 January 2000 / Accepted: 9 July 2000     

Biosystematic studies of the red algal genus Eucheuma. I. Electrophoretic variation among Florida populations

Electrophoretic techniques were used to assess the degree of genetic differentiation among 4 species of the seaweed Eucheuma; E. isiforme, E. nudum, E. gelidium and E. acanthocladum. Nine Floridian populations were surveyed for 5 enzymes (14 loci), and compared using both phenotypic identity (I _p )and Nei's genetic identity (I) statistics. Results suggest there is a very high degree of genetic similarity between E. isiforme and E. nudum (=0.904), and between E. gelidium and E. acanthocladum (=0.857), but a very low degree of similarity between these two pairs of species (=0.372). These findings are in agreement with an earlier report which recommended the removal of E. gelidium and E. acanthocladum from the genus based on differences in reproductive anatomy. Greater genetic similarity was observed among the E. isiforme populations than among the E. nudum populations, the latter being ecologically more diverse and geographically or hydrographically more isolated. The proportion of polymorphic loci per population ranged from 0.25 to 0.36. Heterozygote deficiencies were observed at two acid phosphatase loci in offshore populations of E. nudum which showed a propensity for vegetative reproduction.     

Sibling species in interstitial flatworms: a case study using<Emphasis Type="Italic"> Monocelis lineata</Emphasis> (Proseriata: Monocelididae)

The genetic relationships between morphologically indistinguishable marine and brackish-water populations of Monocelis lineata (O.F. Müller, 1774) (Proseriata: Monocelididae) were analysed by means of allozyme electrophoresis. Fifteen samples of M. lineata (13 from the Mediterranean and two from the Atlantic) from coastal marine and brackish-water habitats were examined for variation at 18 loci. Eleven loci were polymorphic in at least one population of M. lineata. Low levels of within-population genetic variability were found, with average observed and expected heterozygosity values ranging from H_o=0.015±0.015 to 0.113±0.044, and from H_e=0.028±0.028 to 0.138±0.054, respectively. The occurrence of a number of private alleles indicated a marked genetic divergence among populations of M. lineata, with Rogers genetic distances ranging from D_R=0.003 to 0.676 and a highly significant F_ST value (0.918±0.012, P<0.001). UPGMA (unweighted pair-group method using arithmetic average) cluster analysis and multidimensional scaling showed a clear genetic divergence between marine and brackish-water populations. Moreover, Atlantic and Mediterranean populations were sharply separated. Our results suggest that M. lineata is a complex of sibling species.Communicated by R. Cattaneo-Vietti, Genova     

Local and regional genetic connectivity in a Caribbean coral reef fish

Coupled bio-physical models of larval dispersal predict that the Costa Rica–Panama (CR–PAN) reefs should constitute a demographically isolated region in the western Caribbean. We tested the hypothesis that CR–PAN coral reef fish populations would be isolated from Mesoamerican Barrier Reef System (MBRS) populations. To test that, we assessed population genetic structure in bicolor damselfish (Stegastes partitus) from both regions. Adult fish were genotyped from five reefs in CR–PAN and from four reefs along the MBRS at 12 microsatellite loci. Between-region F _ST (F _ST = 0.0030, P < 0.005) and exact test (x ² = 74.34, df = 18, P < 0.0001) results indicated that there is weak but significant genetic differentiation between regions, suggesting some restriction in connectivity along the Central American coastline, as predicted by bio-oceanographic models. Additionally, there is among-site genetic structure in the CR–PAN region, relative to the MBRS and between regions, suggesting higher self-recruitment within CR–PAN. This finding may be explained by differences in habitat characteristics.     

Genetic population structure of the recently introduced Asian clam,Potamocorbula amurensis,in San Francisco Bay

The genetic population structure of the recently introduced Asian clam, Potamocorbula amurensis, in San Francisco Bay was described using starch gel electrophoresis at eight presumptive loci. Specimens were taken from five environmentally distinct sites located throughout the bay. The population maintains a high degree of genetic variation, with a mean heterozygosity of 0.295, a mean polymorphism of 0.75, and an average of 3.70 alleles per locus. The population is genetically homogeneous, as evidenced from genetic distance values and F-statistics. However, heterogeneity of populations was indicated from a contingency chi-square test. Significant deviations from Hardy-Weinberg equilibrium and heterozygote deficiencies were found at the Lap-1 locus for all populations and at the Lap-2 locus for a single population. High levels of variability could represent a universal characteristic of invading species, the levels of variability in the source population(s), and/or the dynamics of the introduction. Lack of differentiation between subpopulations may be due to the immaturity of the San Francisco Bay population, the general purpose phenotype genetic strategy of the species, high rates of gene flow in the population, and/or the selective neutrality of the loci investigated.     

Ancient divergences and recent connections in two tropical Atlantic reef fishes<Emphasis Type="Italic"> Epinephelus adscensionis</Emphasis> and<Emphasis Type="Italic"> Rypticus saponaceous</Emphasis> (Percoidei: Serranidae)

Reef habitats of the tropical Atlantic are separated by river outflows and oceanic expanses that may preclude larval dispersal or other population connections in shorefishes. To examine the impact of these habitat discontinuities on the intraspecific phylogeography of reef-associated species we conducted range-wide surveys of two amphi-Atlantic reef fishes that have dispersive pelagic larval stages. Based on 593 bp of mtDNA cytochrome b from the rock hind Epinephelus adscensionis and 682 bp from the greater soapfish Rypticus saponaceous (n=109 and 86, respectively), we found evidence of relatively ancient separations as well as recent surmounting of biogeographic barriers by dispersal or colonization. Rock hind showed slight but significant population genetic differentiation across much of the tropical Atlantic Ocean (_ST=0.056), but deep divergence between the southeastern United States and seven other localities from the Bahamas to the south, central and east Atlantic (mean pairwise d=0.040, overall _ST=0.867). The geographic distribution of the two rock hind lineages is highly unusual in genetic studies of Caribbean Sea reef fishes, because those lineages are separated by less than 250 km of open water within a major biogeographic region. In contrast, highly significant population genetic structure was observed among greater soapfish from the SW Caribbean, Brazil, and mid-Atlantic ridge (_ST=0.372), with a deep evolutionary separation distinguishing putative R. saponaceous from West Africa (mean pairwise d=0.044, overall _ST=0.929). Both species show evidence for a potential connection between the Caribbean and Brazilian provinces. While widespread haplotype sharing in rock hind indicates that larvae of this species cross oceanic expanses of as much as 2000 km, such a situation is difficult to reconcile with the isolation of populations in Florida and the Bahamas separated by only 250 km. These findings indicate that populations of some species in disjunct biogeographic zones may be isolated for long periods, perhaps sufficient for allopatric speciation, but rare gene flow between zones may preclude such evolutionary divergence in other species.Communicated by P.W. Sammarco, Chauvin     

Genetic studies in marine fishes

Nine polymorphic loci were found among 42 presumptive protein-coding gene loci surveyed among 474 red drum (Sciaenops ocellatus) sampled in 1987 from 13 nearshore and 1 offshore localities from the Atlantic coast of the southeastern USA and the northern Gulf of Mexico. The mean number of alleles over the polymorphic loci was 3.8, and the average heterozygosity over all loci examined was estimated as 0.047. These data indicate that red drum have normal levels of genetic variability. Wright'sF1 _ST values (the standardized variance of allele frequencies between samples) over all polymorphic loci ranged from 0.009 to 0.027 (meanF1 _ST =0.019), and estimates of the effective number of migrants (N _e m) per generation using Wright's island model ranged from 9.0 to 27.5. High levels of gene flow among the red drum samples were also indicated by Slatkin's qualitative analysis using conditional average allele frequencies. Nei's estimates of genetic distance between pairs of samples ranged from 0.000 to 0.009, indicating a high degree of nuclear gene similarity among all samples. Highly significant heterogeneity in allele frequencies at the locus for adenosine deaminase was detected between red drum sampled from the Atlantic and those sampled from the Gulf and among red drum sampled from the Gulf.     

Gene flow among giant clam (Tridacna gigas) populations in Pacific does not parallel ocean circulation

Tridacna gigas (L.) populations were sampled from 13 locations throughout the west Pacific in 1990–1991 and screened for allozyme variation at seven polymorphic loci. Significant genetic differences were found between the few large stocks of T. gigas remaining in the world. Cluster analysis identified two major regional groups of populations: a Central Pacific group consisting of Kiribati and the Marshall Islands, and a West Pacific group comprising all other populations. Within the latter group, the Great Barrier Reef populations were significantly differentiated from those in the Solomon Islands. The main routes of gene flow did not run parallel with known major surface currents as might be expected, but crossed them. Gene exchange was limited east-west between the Central and West Pacific groups and the GBR and the Solomon Islands. The lack of correspondence between the major surface currents and the routes of gene flow suggests that the genetic structure of T. gigas reflects historical patterns of migration that no longer occur, rather than presentday dispersal. These findings raise fundamental questions concerning the origin and maintenance of genetic diversity in Pacific faunas, and have important implications for the mariculture and conservation of T. gigas.     

Limited gene flow in the brooding coral Favia fragum (Esper, 1797)

Understanding population connectivity in corals is particularly important as these organisms are increasingly threatened by abiotic and biotic factors. This study examined the population genetic structure of the brooding coral Favia fragum across four locations in the Caribbean and Western Atlantic using mitochondrial and nuclear markers. Morphological features were also compared to test whether genetic diversity corresponds with skeletal morphology. When comparing across distantly related Caribbean and Bermudian locations, F _ST values were high and significant, indicating strong genetic structure. At a local scale, significant genetic structure was found among reefs in Panama, while no genetic structure was found among reefs within Barbados, Bermuda or Jamaica. Surprisingly, a single haplotype for each of the three markers examined was found in Bermuda, where samples varied significantly from all other locations in three out of four morphological features analyzed. These data indicate that gene flow of F. fragum may occur locally among reefs but is highly restricted at distant locations. Furthermore, isolated populations, such as that of Bermuda, must be self-seeding to maintain the observed genetic uniformity.     

Genetic structure of Dascyllus aruanus populations in French Polynesia

Dascyllus aruanus were collected from 13 different locations in French Polynesia between 15 November 1990 and 15 February 1991 in order to examine larval dispersal on four spatial scales: within-feef, within-island, within-archipelago, among-archipelagoes. Average polymorphism was analysed by protein electrophoresis at two levels (P₉₅=0.285 and P₉₀=0.107) from 12 and 3 polymorphic loci, respectively. Spatial genetic variation displayed a low level of differentiation between populations among archipelagoes, and homogeneity at lower spatial scales. Two hypotheses are proposed to explain the genetic structure observed. The first suggests substantial gene flow between the islands during the pelagic larval phase, the second that the absence of differentiation is the result of recent colonisation. Genetic variation amongst size classes showed a significant heterozygote deficiencies at two loci (PGM ^* and EST-2 ^*) in the smallest size class. This suggests a cyclic selection which affects larvae and adults differentially. The data revealed little differentiation among populations at the different localities, despite the short larval duration of D. aruanus; this suggests that larval duration is not the main factor presently affecting genetic structure in an insular model. Correspondence to: S. Planes at the Université de Perpignan     

Molecular genetic variation in tarpon (<Emphasis Type="Italic">Megalops atlanticus</Emphasis> Valenciennes) in the northern Atlantic Ocean

The tarpon (Megalops atlanticus) is a highly valued game fish and occasional food fish in the eastern and western Atlantic Ocean. Tarpon have a high capacity for dispersal, but some regional biological differences have been reported. In this study we used two molecular genetic techniques—protein electrophoresis of nuclear DNA loci, and restriction fragment length polymorphism analysis of the mitochondrial DNA (mtDNA)—to assess this species population genetic structure in the eastern (coastal waters off Gabon and Sierra Leone, Africa) and western (coastal waters off Florida, Caribbean Sea) Atlantic Ocean north of the equator. Genetic differentiation was observed between tarpon from Africa and tarpon from the western Atlantic Ocean. A unique allele and haplotype, significant differences in allozyme allele and mtDNA haplotype frequencies between the African and western Atlantic samples, and significant F_ST analyses suggest that levels of gene flow between tarpon from these two regions is low. Among the western Atlantic Ocean collections, genetic diversity values and allele and haplotype frequencies were similar. AMOVA analyses also showed a degree of genetic relatedness among most of the western Atlantic Ocean collections: however, some significant population structuring was detected in the allozyme data. A regional jackknifed F_ST analysis indicated the distinction of the Costa Rica population from the other western Atlantic populations and, in pairwise analyses, F_ST values tended to be higher (i.e., genetic relatedness was lower) when the Costa Rican sample was paired with any of the other western Atlantic samples. These data suggest that Costa Rican tarpon could be partially isolated from other western Atlantic tarpon populations. Ultimately, international cooperation will be essential in the management of this species in both the eastern and western Atlantic Ocean.Communicated by P.W. Sammarco, Chauvin