Two hundred and twenty-five Polynesians were selected from a larger study population for the evaluation of potential genetic influences on the susceptibility to bancroftian filariasis. Analysis showed that there was significantly familial clustering of patients with filariasis and that this clustering was most compatible with genetic transmission of disease susceptibility. The data best fit a model in which the hypothetical gene for filariasis was recessive with a frequency of 0.82+/-0.15 in the population and a penetrance of 0.62+/-0.14. The alternative hypothesis that susceptibility was environmentally (ie, not genetically) determined was also compatible with the data but was estimated to be 1.9 times less likely to account for the observed findings than the genetic hypothesis. Extensive evaluation of HLA-A and-B locus specificities failed to detect significant linkage either between particular antigen specificities and the clinical manifestations of filariasis or between individual haplotypes (indicated by HLA markers in studies of large families) and the predisposition to filarial infection or disease.