GENETIC DRIFT AND COLLECTIVE DISPERSAL CAN RESULT IN CHAOTIC GENETIC PATCHINESS

Chaotic genetic patchiness denotes unexpected patterns of genetic differentiation that are observed at a fine scale and are not stable in time. These patterns have been described in marine species with free‐living larvae, but are unexpected because they occur at a scale below the dispersal range of pelagic larvae. At the scale where most larvae are immigrants, theory predicts spatially homogeneous, temporally stable genetic variation. Empirical studies have suggested that genetic drift interacts with complex dispersal patterns to create chaotic genetic patchiness. Here we use a coancestry model and individual‐based simulations to test this idea. We found that chaotic genetic patterns (qualified by global F_ST and spatio‐temporal variation in F_ST's between pairs of samples) arise from the combined effects of (1) genetic drift created by the small local effective population sizes of the sessile phase and variance in contribution among breeding groups and (2) collective dispersal of related individuals in the larval phase. Simulations show that patchiness levels qualitatively comparable to empirical results can be produced by a combination of strong variance in reproductive success and mild collective dispersal. These results call for empirical studies of the effective number of breeders producing larval cohorts, and population genetics at the larval stage.     

Chaotic genetic patchiness in the pelagic teleost fish Sardina pilchardus across the Siculo-Tunisian Strait

ABSTRACT

This investigation aims at assessing patterns of spatial genetic structure of the teleost fish Sardina pilchardus across the Siculo-Tunisian Strait (a well-known discontinuous biogeographic area) and delineating putative genetic stocks within the species. For this purpose, a total of 180 specimens, collected from 11 locations stretching across the western and eastern Mediterranean coasts of Tunisia, were analysed genetically by means of 18 nuclear allozyme loci. The outcome of this study revealed strong genetic differentiation among populations, with the marked genetic distinctiveness of the central Tunisian population at Mahdia. Despite the delineation of seven well-defined genetic groups, no significant correlation was found between genetic and geographic distances. Besides, the recorded population subdivision did not align with biogeographic boundaries, suggesting the presence of chaotic genetic patchiness. Recent genetic bottlenecks were evidenced in S. pilchardus populations. Patchy migration patterns were recorded among the examined pairs of sardine populations. Among the recorded 16 polymorphic loci, GPI-2 and SOD appeared to be subject to natural selection. Patterns of population genetic differentiation and structuring were not found to be driven by outlier loci that appeared to be under selection. Furthermore, the detected neutral GPI-1 locus was found to be responsible for most of the genetic variation among identified genetic clusters. Hence, natural selection cannot cause the detected genetic heterogeneity among sardine samples. Different explanations to the origin of chaotic genetic patterns, observed within S. pilchardus, were discussed.     

Oceanographic features and limited dispersal shape the population genetic structure of the vase sponge Ircinia campana in the Greater Caribbean

Understanding population genetic structure can help us to infer dispersal patterns, predict population resilience and design effective management strategies. For sessile species with limited dispersal, this is especially pertinent because genetic diversity and connectivity are key aspects of their resilience to environmental stressors. Here, we describe the population structure of Ircinia campana, a common Caribbean sponge subject to mass mortalities and disease. Microsatellites were used to genotype 440 individuals from 19 sites throughout the Greater Caribbean. We found strong genetic structure across the region, and significant isolation by distance across the Lesser Antilles, highlighting the influence of limited larval dispersal. We also observed spatial genetic structure patterns congruent with oceanography. This includes evidence of connectivity between sponges in the Florida Keys and the southeast coast of the United States (>700 km away) where the oceanographic environment is dominated by the strong Florida Current. Conversely, the population in southern Belize was strongly differentiated from all other sites, consistent with the presence of dispersal-limiting oceanographic features, including the Gulf of Honduras gyre. At smaller spatial scales (<100 km), sites showed heterogeneous patterns of low-level but significant genetic differentiation (chaotic genetic patchiness), indicative of temporal variability in recruitment or local selective pressures. Genetic diversity was similar across sites, but there was evidence of a genetic bottleneck at one site in Florida where past mass mortalities have occurred. These findings underscore the relationship between regional oceanography and weak larval dispersal in explaining population genetic patterns, and could inform conservation management of the species.Subject terms: Genetic variation, Ecology     

Local adaptation despite high gene flow in the waterfall‐climbing Hawaiian goby,Sicyopterus stimpsoni

Environmental heterogeneity can promote the emergence of locally adapted phenotypes among subpopulations of a species, whereas gene flow can result in phenotypic and genotypic homogenization. For organisms like amphidromous fishes that change habitats during their life history, the balance between selection and migration can shift through ontogeny, making the likelihood of local adaptation difficult to predict. In Hawaiian waterfall‐climbing gobies, it has been hypothesized that larval mixing during oceanic dispersal counters local adaptation to contrasting topographic features of streams, like slope gradient, that can select for predator avoidance or climbing ability in juvenile recruits. To test this hypothesis, we used morphological traits and neutral genetic markers to compare phenotypic and genotypic distributions in recruiting juveniles and adult subpopulations of the waterfall‐climbing amphidromous goby, Sicyopterus stimpsoni, from the islands of Hawai'i and Kaua'i. We found that body shape is significantly different between adult subpopulations from streams with contrasting slopes and that trait divergence in recruiting juveniles tracked stream topography more so than morphological measures of adult subpopulation differentiation. Although no evidence of population genetic differentiation was observed among adult subpopulations, we observed low but significant levels of spatially and temporally variable genetic differentiation among juvenile cohorts, which correlated with morphological divergence. Such a pattern of genetic differentiation is consistent with chaotic genetic patchiness arising from variable sources of recruits to different streams. Thus, at least in S. stimpsoni, the combination of variation in settlement cohorts in space and time coupled with strong postsettlement selection on juveniles as they migrate upstream to adult habitats provides the opportunity for morphological adaptation to local stream environments despite high gene flow.     

Population subdivision in Australian temperate marine invertebrates: Larval connections versus historical factors*

This paper reviews the use of genetic data, in combination with manipulative experimentation, to infer the mode of reproduction and the extent and directionality of dispersal for a range of Australian temperate marine invertebrates. Local populations of obligately sexually reproducing species have been inferred to be strongly interconnected by larval dispersal, over distances of thousands of kilometres. Their larvae may be subject to strong post-settlement selection, but this selection is independent of obvious geographic or intertidal gradients. Within local populations selection may therefore result in apparently chaotic genetic patchiness which is eliminated by the effects of sexual reproduction and the widespread dispersal and mixing of the colonizing larvae of each generation. In partial contrast, local populations of species which rely on asexual reproduction for the maintenance of populations show evidence of similar larval connections, but no recent settlement of their sexually generated larvae has been demonstrated. The apparent connectedness of these populations may reflect either historical events or a more episodic pattern of settlement by sexually generated larvae. Local populations of these species are more highly differentiated as a result of the continued asexual replication of a limited number of genotypes. In one of these species, reciprocal transplantation of the clones within and among populations has revealed that resident clones can be highly locally adapted (as reflected by much higher asexual fecundity), which implies that selection is an important determinant of the composition of local populations. Nevertheless, the failure to detect continuing sexual recruitment into these populations obscures the evolutionary significance of this finding.     

Population genetic structure of two ecologically distinct multimammate rats: the commensal Mastomys natalensis and the wild Mastomys erythroleucus in southeastern Senegal

Using the same set of microsatellite markers, we compared the population genetic structure of two Mastomys species, one being exclusively commensal in southeastern Senegal, and the other being continuously distributed outside villages in this region. Both species were sampled in the same landscape context and at the same spatial scale. According to the expectations based on the degree of habitat patchiness (which is higher for commensal populations in this rural area), genetic diversity was lower and genetic differentiation was higher in commensal populations of Mastomys natalensis than in wild populations of Mastomys erythroleucus. Contrasting estimates of effective dispersal and current migration rates corroborates previous data on differences in social structure between the two species. Isolation-by-distance analyses showed that human-mediated dispersal is not a major factor explaining the pattern of genetic differentiation for M. natalensis, and that gene flow is high and random between M. erythroleucus populations at the spatial scale considered.     

Spatial genetic differentiation correlates with species assemblage turnover across tropical reef fish lineages

Aim

Evaluating the similarity of diversity patterns across micro- to macroevolutionary scales in natural communities, such as species–genetic diversity correlations (SGDCs), may inform on processes shaping community assembly. However, whether SGDCs not only hold across communities but also across lineages has never been explored so far. Here we investigated SGDCs across co-distributed taxa for different spatial components (α, β, γ), and formally tested the influence of dispersal traits on β-SGDCs.

Location

Western Indian Ocean.

Time period

2016–2017.

Major taxa studied

Tropical reef fish species with contrasting dispersal traits.

Methods

Using double-digest restriction-site associated DNA sequencing (ddRADseq) Single Nucleotide Polymorphism data for 20 tropical reef fishes and distribution data of 2,446 species belonging to 12 families, we analysed the correlations between within-species genetic diversity and within-family species diversity (i.e., lineage diversity) for the three spatial components (α, β, γ-SGDCs). We then related the strength of β-SGDCs per species to proxies of larval dispersal abilities.

Results

We detected positive and significant lineage-based SGDC only for the β component, that is, the families showing the greatest level of species turnover among sites contain the species with the greatest levels of genetic differentiation. We showed that the Monsoon Drift mainly explained the β-diversity patterns at both intraspecific and interspecific levels. Higher β-SGDCs were found for species with short pelagic larval duration and weak larval swimming capacity.

Main conclusions

Our study reveals a strong correlation between genetic and species β-diversity, a result explained by the presence of a ‘soft’ barrier and mediated by larval dispersal processes. This suggests that vicariance and dispersal limitation are major processes shaping β-diversity patterns from microevolutionary to macroevolutionary scales in tropical reef fishes.     

Cryptic lineage divergence in marine environments: genetic differentiation at multiple spatial and temporal scales in the widespread intertidal goby Gobiosoma bosc

The adaptive radiation of the seven‐spined gobies (Gobiidae: Gobiosomatini) represents a classic example of how ecological specialization and larval retention can drive speciation through local adaptation. However, geographically widespread and phenotypically uniform species also do occur within Gobiosomatini. This lack of phenotypic variation across large geographic areas could be due to recent colonization, widespread gene flow, or stabilizing selection acting across environmental gradients. We use a phylogeographic approach to test these alternative hypotheses in the naked goby Gobiosoma bosc, a widespread and phenotypically invariable intertidal fish found along the Atlantic Coast of North America. Using DNA sequence from 218 individuals sampled at 15 localities, we document marked intraspecific genetic structure in mitochondrial and nuclear genes at three main geographic scales: (i) between Gulf of Mexico and Atlantic Coast, (ii) between the west coast of the Florida peninsula and adjacent Gulf of Mexico across the Apalachicola Bay, and (iii) at local scales of a few hundred kilometers. Clades on either side of Florida diverged about 8 million years ago, whereas some populations along the East Cost show divergent phylogroups that have differentiated within the last 200,000 years. The absence of noticeable phenotypic or ecological differentiation among lineages suggests the role of stabilizing selection on ancestral phenotypes, together with isolation in allopatry due to reduced dispersal and restricted gene flow, as the most likely explanation for their divergence. Haplotype phylogenies and spatial patterns of genetic diversity reveal frequent population bottlenecks followed by rapid population growth, particularly along the Gulf of Mexico. The magnitude of the genetic divergence among intraspecific lineages suggests the existence of cryptic species within Gobiosoma and indicates that modes of speciation can vary among lineages within Gobiidae.     

Kin-Aggregations Explain Chaotic Genetic Patchiness,a Commonly Observed Genetic Pattern,in a Marine Fish

The phenomenon of chaotic genetic patchiness is a pattern commonly seen in marine organisms, particularly those with demersal adults and pelagic larvae. This pattern is usually associated with sweepstakes recruitment and variable reproductive success. Here we investigate the biological underpinnings of this pattern in a species of marine goby Coryphopterus personatus. We find that populations of this species show tell-tale signs of chaotic genetic patchiness including: small, but significant, differences in genetic structure over short distances; a non-equilibrium or “chaotic” pattern of differentiation among locations in space; and within locus, within population deviations from the expectations of Hardy-Weinberg equilibrium (HWE). We show that despite having a pelagic larval stage, and a wide distribution across Caribbean coral reefs, this species forms groups of highly related individuals at small spatial scales (<10 metres). These spatially clustered family groups cause the observed deviations from HWE and local population differentiation, a finding that is rarely demonstrated, but could be more common than previously thought.     

Understanding the population genetic structure of Gleditsia triacanthos L.: seed dispersal and variation in female reproductive success

Fine-scale genetic structuring is influenced by a variety of ecological factors and can directly affect the evolutionary dynamics of plant populations by influencing effective population size and patterns of viability selection. In many plant species, genetic structuring within populations may result from highly localized patterns of seed dispersal around maternal plants or by the correlated dispersal and recruitment of siblings from the same fruit. This fine-scale genetic structuring may be enhanced if female parents vary significantly in their reproductive success. To test these hypotheses, we used genetic data from 17 allozyme loci and a maximum-likelihood, ‘maternity-analysis’ model to estimate individual female fertilities for maternal trees across a large number of naturally established seedlings and saplings in two populations of Gleditsia triacanthos L. (Leguminosae). Maximum-likelihood fertility estimates showed that the three highest fertility females accounted for 58% of the 313 progeny at the first site and 46% of the 651 progeny at the second site, whereas 18 of 35 and 16 of 34 females, respectively, had fertility estimates that did not exceed 1%. Additional analyses of the second site found individual female fertility to vary significantly both within and among juvenile age classes. Female fertility at the first site was weakly correlated with maternal tree size and spatial location relative to the open, old-field portions of the population, where the great majority of seedlings and saplings were growing, but no such correlations were found at the second site. Estimates of realized seed dispersal distances indicated that dispersal was highly localized at the first site, but was nearly random at the second site, possibly reflecting differences between the two sites in the behaviour of animal dispersers. The combined estimates of seed dispersal patterns and fertility variation are sufficient to explain previously described patterns of significant fine-scale spatial genetic structure in these two populations. In general, our results demonstrate that effective seed dispersal distributions may vary significantly from population to population of a species due to the unpredictable behaviour of secondary dispersers. Consequently, the effects of seed dispersal on realized fine-scale genetic structure may also be relatively unpredictable.     

Within‐generation consequences of postsettlement mortality for trait composition in wild populations: An experimental test

There is a critical need to understand patterns and causes of intraspecific variation in physiological performance in order to predict the distribution and dynamics of wild populations under natural and human‐induced environmental change. However, the usual explanation for trait differences, local adaptation, fails to account for the small‐scale phenotypic and genetic divergence observed in fishes and other species with dispersive early life stages. We tested the hypothesis that local‐scale variation in the strength of selective mortality in early life mediates the trait composition in later life stages. Through in situ experiments, we manipulated exposure to predators in the coral reef damselfish Dascyllus aruanus and examined consequences for subsequent growth performance under common garden conditions. Groups of 20 recently settled D. aruanus were outplanted to experimental coral colonies in Moorea lagoon and either exposed to natural predation mortality (52% mortality in three days) or protected from predators with cages for three days. After postsettlement mortality, predator‐exposed groups were shorter than predator‐protected ones, while groups with lower survival were in better condition, suggesting that predators removed the longer, thinner individuals. Growth of both treatment groups was subsequently compared under common conditions. We did not detect consequences of predator exposure for subsequent growth performance: Growth over the following 37 days was not affected by the prior predator treatment or survival. Genotyping at 10 microsatellite loci did indicate, however, that predator exposure significantly influenced the genetic composition of groups. We conclude that postsettlement mortality did not have carryover effects on the subsequent growth performance of cohorts in this instance, despite evidence for directional selection during the initial mortality phase.     

Genomic analysis of a cardinalfish with larval homing potential reveals genetic admixture in the Okinawa Islands

Discrepancies between potential and observed dispersal distances of reef fish indicate the need for a better understanding of the influence of larval behaviour on recruitment and dispersal. Population genetic studies can provide insight on the degree to which populations are connected, and the development of restriction site‐associated sequencing (RAD‐Seq) methods has made such studies of nonmodel organisms more accessible. We applied double‐digest RAD‐Seq methods to test for population differentiation in the coral reef‐dwelling cardinalfish, Siphamia tubifer, which based on behavioural studies, have the potential to use navigational cues to return to natal reefs. Analysis of 11,836 SNPs from fish collected at coral reefs in Okinawa, Japan, from eleven locations over 3 years reveals little genetic differentiation between groups of S. tubifer at spatial scales from 2 to 140 km and between years at one location: pairwise F_ST values were between 0.0116 and 0.0214. These results suggest that the Kuroshio Current largely influences larval dispersal in the region, and in contrast to expectations based on studies of other cardinalfishes, there is no evidence of population structure for S. tubifer at the spatial scales examined. However, analyses of outlier loci putatively under selection reveal patterns of temporal differentiation that indicate high population turnover and variable larval supply from divergent source populations between years. These findings highlight the need for more studies of fishes across various geographic regions that also examine temporal patterns of genetic differentiation to better understand the potential connections between early life‐history traits and connectivity of reef fish populations.     

Metals and altitude drive genetic diversity of chironomids in Andean streams

相似文献   

Combining hydrodynamic modelling with genetics: can passive larval drift shape the genetic structure of Baltic Mytilus populations?

While secondary contact between Mytilus edulis and Mytilus trossulus in North America results in mosaic hybrid zone formation, both species form a hybrid swarm in the Baltic. Despite pervasive gene flow, Baltic Mytilus species maintain substantial genetic and phenotypic differentiation. Exploring mechanisms underlying the contrasting genetic composition in Baltic Mytilus species will allow insights into processes such as speciation or adaptation to extremely low salinity. Previous studies in the Baltic indicated that only weak interspecific reproductive barriers exist and discussed the putative role of adaptation to environmental conditions. Using a combination of hydrodynamic modelling and multilocus genotyping, we investigate how oceanographic conditions influence passive larval dispersal and hybrid swarm formation in the Baltic. By combining our analyses with previous knowledge, we show a genetic transition of Baltic Mytilus species along longitude 12°‐13°E, that is a virtual line between Malmö (Sweden) and Stralsund (Germany). Although larval transport only occurs over short distances (10–30 km), limited larval dispersal could not explain the position of this genetic transition zone. Instead, the genetic transition zone is located at the area of maximum salinity change (15–10 psu). Thus, we argue that selection results in weak reproductive barriers and local adaptation. This scenario could maintain genetic and phenotypic differences between Baltic Mytilus species despite pervasive introgressive hybridization.     

Weak genetic structure indicates strong dispersal limits: a tale of two coral reef fish

The extent of dispersal by pelagic larvae in marine environments, including coral reefs, is central for understanding local population dynamics and designing sustainable marine reserves. We present here the first example of a clear stepping-stone genetic structure throughout the Caribbean basin for a common coral reef species, the French grunt (Haemulon flavolineatum). Analysis of microsatellite DNA markers indicated that French grunt population structure may be characterized by overlapping populations throughout the Caribbean, influenced by independent population dynamics but with no fixed geographical boundaries. In addition, different spatial genetic patterns were found in different oceanographic regions. A second species, the bluehead wrasse (Thalassoma bifasciatum), has a much longer pelagic larval duration than French grunts and showed no explicit spatial pattern of genetic variation. This finding is concordant with the hypothesis of a positive relationship between larval dispersal and duration in the plankton. While the magnitude of the genetic signal of population structure in French grunts was very low (F(ST) approximately 0.003), the pattern of isolation-by-distance throughout the Caribbean indicated considerable population structure with important ecological and conservation significance.     

Coral reef fish larvae settle close to home

Population connectivity through larval dispersal is an essential parameter in models of marine population dynamics and the optimal size and spacing of marine reserves. However, there are remarkably few direct estimates of larval dispersal for marine organisms, and the actual birth sites of successful recruits have never been located. Here, we solve the mystery of the natal origin of clownfish (Amphiprion polymnus) juveniles by mass-marking via tetracycline immersion all larvae produced in a population. In addition, we established parentage by DNA genotyping all potential adults and all new recruits arriving in the population. Although no individuals settled into the same anemone as their parents, many settled remarkably close to home. Even though this species has a 9-12 day larval duration, one-third of settled juveniles had returned to a 2 hectare natal area, with many settling <100 m from their birth site. This represents the smallest scale of dispersal known for any marine fish species with a pelagic larval phase. The degree of local retention indicates that marine reserves can provide recruitment benefits not only beyond but also within their boundaries.     

Evidence for isolation by time in the European eel (Anguilla anguilla L.)

Life history traits of highly vagile marine species, such as adult reproductive success and larval dispersal, are strongly determined by oceanographic and climatic forces. Nevertheless, marine organisms may show restricted dispersal in time and space. Patterns of isolation by distance (IBD) have been repeatedly observed in marine species. If spawning time is a function of geographical location, temporal and spatial isolation, can easily be confounded or misinterpreted. In this study, we aimed at discriminating between various forces shaping the genetic composition of recruiting juveniles of the European eel (Anguilla anguilla L.). By controlling for geographical variation, we assessed temporal variation and tested for possible isolation by time (IBT) between recruitment waves within and between years. Using 12 polymorphic allozyme and six variable microsatellite loci, we show that genetic differentiation was low (F(ST) = 0.01-0.002) and significant between temporal samples. Regression analysis between genetic and temporal distance, was consistent with a subtle interannual pattern of IBT. Our data suggest that the population dynamics of the European eel may be governed by a double pattern of temporal variance in genetic composition: (i) a broad-scale IBT of spawning cohorts, possibly as a consequence of the large migration loop in anguillids and strong variance in annual adult reproductive contribution; and (ii) a smaller-scale variance in reproductive success (genetic patchiness) within cohorts among seasonally separated spawning groups, most likely originating from fluctuating oceanic and climatic forces. The consistency of both mechanisms remains to be verified with fine-scale analyses of both spawning/migrating aged adults and their offspring to confirm the stochastic/deterministic nature of the IBT pattern in eel.     

Emergent patterns of population genetic structure for a coral reef community

What shapes variation in genetic structure within a community of codistributed species is a central but difficult question for the field of population genetics. With a focus on the isolated coral reef ecosystem of the Hawaiian Archipelago, we assessed how life history traits influence population genetic structure for 35 reef animals. Despite the archipelago's stepping stone configuration, isolation by distance was the least common type of genetic structure, detected in four species. Regional structuring (i.e. division of sites into genetically and spatially distinct regions) was most common, detected in 20 species and nearly in all endemics and habitat specialists. Seven species displayed chaotic (spatially unordered) structuring, and all were nonendemic generalist species. Chaotic structure also associated with relatively high global F_ST. Pelagic larval duration (PLD) was not a strong predictor of variation in population structure (R² = 0.22), but accounting for higher F_ST values of chaotic and invertebrate species, compared to regionally structured and fish species, doubled the power of PLD to explain variation in global F_ST (adjusted R² = 0.50). Multivariate correlation of eight species traits to six genetic traits highlighted dispersal ability, taxonomy (i.e. fish vs. invertebrate) and habitat specialization as strongest influences on genetics, but otherwise left much variation in genetic traits unexplained. Considering that the study design controlled for many sampling and geographical factors, the extreme interspecific variation in spatial genetic patterns observed for Hawaìi marine species may be generated by demographic variability due to species‐specific abundance and migration patterns and/or seascape and historical factors.     

The use of energy stores in the puerulus of the spiny lobster Jasus edwardsii across the continental shelf of New Zealand

Nektonic pueruli of the spiny lobster, Jasus edwardsii, were caught from two locations about 20 km apart across the continental shelf on the south east of the North Island, New Zealand. The pueruli were assayed for total protein, glucose, glycogen, and lipid content. Only the lipid content differed between pueruli caught onshore and offshore (mean difference=3.1 mg or 3.4% of dry mass). The average difference in lipid content measured over this distance was used to calculate the rate of energy consumption and timing for pueruli to actively swim from the continental shelf to shore. These results confirmed previous theoretical estimates and indirect measures. Furthermore, the rate of energy consumption would allow all of the pueruli caught offshore to swim to shore based on their total measured lipids. However, some individuals with low energy stores may be energetically compromised at arrival which may affect their subsequent moulting and survival. The results of this study indicate that lipid is the primary format for energy storage of the nektonic puerulus of the spiny lobster and that these lipid reserves have sufficient energetic capacity to allow the puerulus to actively swim the distance across the shelf to settle on the coast.     

Reduced genetic diversity and increased reproductive isolation follow population‐level loss of larval dispersal in a marine gastropod

Population‐level consequences of dispersal ability remain poorly understood, especially for marine animals in which dispersal is typically considered a species‐level trait governed by oceanographic transport of microscopic larvae. Transitions from dispersive (planktotrophic) to nondispersive, aplanktonic larvae are predicted to reduce connectivity, genetic diversity within populations, and the spatial scale at which reproductive isolation evolves. However, larval dimorphism within a species is rare, precluding population‐level tests. We show the sea slug Costasiella ocellifera expresses both larval morphs in Florida and the Caribbean, regions with divergent mitochondrial lineages. Planktotrophy predominated at 11 sites, 10 of which formed a highly connected and genetically diverse Caribbean metapopulation. Four populations expressed mainly aplanktonic development and had markedly reduced connectivity, and lower genetic diversity at one mitochondrial and six nuclear loci. Aplanktonic dams showed partial postzygotic isolation in most interpopulation crosses, regardless of genetic or geographic distance to the sire's source, suggesting that outbreeding depression affects fragmented populations. Dams from genetically isolated and neighboring populations also exhibited premating isolation, consistent with reinforcement contingent on historical interaction. By increasing self‐recruitment and genetic drift, the loss of dispersal may thus initiate a feedback loop resulting in the evolution of reproductive isolation over small spatial scales in the sea.