Résumé
Direct estimates of larval retention and connectivity are essential to parameterize marine metapopulation models and to optimize the design of marine protected area networks. For coral reef fishes, while there are some estimates of self-recruitment at isolated populations, there are no direct estimates of exchange between adjacent sub-populations or between protected and unprotected areas. Here we use a likelihood-based parentage analysis to assess the relative magnitude of self-recruitment and exchange among five sub-populations of the panda clownfish Amphiprion polymnus in Bootless Bay, Papua New Guinea. These small sub-populations varied in size between 28 and 84 individuals. Each was confined to a discrete ~1ha patch of shallow sand that was separated from other subpopulations by 2 to 5km. All adults were finclipped and all juveniles that settled over a 7 month period were collected for genetic analyses. We used nine microsatellite loci and the FAMOZ software to assign settlers to putative parents, with unassigned juveniles considered to be immigrants from other unknown populations. The overall level of self-recruitment for the combined 5 sub-populations was ~25%. This figure was apportioned to an average of 10% self-recruitment within subpopulations and ~15% exchange among the sub-populations. The data suggest that both fine-scale self-recruitment and exchange among populations 2-5km apart contribute to the persistence of this clownfish metapopulation, along with substantial larger-scale dispersal. We hypothesize that A. polymnus is characterised by a multimodal distribution of larval dispersal distances that varies according to the spacing of sub-populations.