| Abstract: | We examined the hypothesis that secondary contact generates an allele-frequency cline at the aminopeptidase-I locus (Lap) in the marine mussel, Mytilus edulis. It has been proposed that variation at the Lap locus is neutral and that the cline results from secondary contact between differentiated oceanic and estuarine populations (Levinton, 1980). We tested this hypothesis by comparing the genotypic distributions in samples from the cline to distributions that incorporate mixing effects. We employed a statistical model that determines the degree of contact using a maximum likelihood estimator and then incorporates the mixing estimates into an expected distribution of genotypes. Wahlund effects resulting from possible admixture are thereby incorporated into the expected distribution. Failure of the model to reconcile the observed with the expected distribution of genotypes indicates that the observed population structure does not result from admixture. The null hypothesis of mixing was unable to explain about 33% of the samples. Combined tests demonstrated the general departure from the mixing model to be highly significant. The distribution of heterozygote discrepancies across the cline was inconsistent with the expectations of a mixing model. Therefore we reject explanations for the structure of the Lap cline that involve secondary contact. Selection directed at the Lap locus appears necessary to explain the genotypic structure of clinal populations. |
