The rate of gene flow between populations depends both on the dispersal ability of the organism and on the probability that migrants can successfully mate with residents. The degree to which this reproductive isolation can impede genetic exchange can be assessed by studying both migration and gene flow. We carried out such a study in Panamanian populations of Excirolana, an intertidal isopod with no planktonic larvae that inhabits both coasts of Central and South America. In Panama Excirolana is represented by three morphs of E. braziliensis (two closely related and found on opposite sides of Central America, a third only in the eastern Pacific) and by E. chamensis. Because populations of Excirolana from each beach in Panama, regardless of species or morphotype affiliation, are characterized by a unique combination of alleles, this isopod provides the opportunity to use genetic data to determine the number of immigrants that land on a given beach as well as the number of individuals that transfer alleles between populations. We studied isozyme patterns of 10 populations of E. braziliensis and one population of E. chamensis. Rates of gene flow were estimated both from FST statistics and from private alleles. In contrast to the private allele method, FST statistics gave results that were internally consistent in that they produced lower estimates of apparent gene flow between oceans and recognized species than they did between populations of the same morph. Rate of gene flow between populations belonging to the two morphs of E. braziliensis in the eastern Pacific was as low as apparent gene flow across the Central American isthmus and no larger than the rate of genetic exchange between E. braziliensis and E. chamensis. One locality that contained both morphs in sufficiently large numbers showed heterozygote deficits relative to Hardy-Weinberg expectations. Reproductive isolation is, therefore, of sufficient strength to justify the conclusion that the two Pacific morphs of E. braziliensis should be assigned to different species. Gene flow rate between populations of the same morph was an order of magnitude larger than rate of gene exchange between morphs but still less than one propagule per generation. Rate of immigration into every beach, calculated from the number of homozygotes of rare alleles (in excess of what would be expected from random matings among the residents), was surprisingly high for an organism with apparently limited means of dispersal and with fixed genetic differences between adjacent populations. Individuals of E. braziliensis evidently move between beaches much more freely than alleles move between populations. Therefore, some degree of reproductive isolation exists even between genotypes that belong to the same morph.Keywords: diagnostic loci, dispersal, FST statistics, geographic variation, isozymes, private alleles. tion genetics the terms 'migration' and 'gene flow' are Introduction often used interchangeably (thus implying that migrantsThe rate of gene flow between populations depends ca...