Available evidence suggests that male homosexuality is both familial and somewhat heritable and that some cases may be caused by an X-linked gene. However, most studies have recruited subjects in a relatively unsystematic manner, typically via advertisements, and hence suffer from the potential methodological flaw of ascertainment bias due to volunteer self-selection. In the present study we assessed the familiality of male homosexuality using two carefully ascertained samples and attempted to replicate findings consistent with X-linkage in three samples. The percentage of siblings of the probands rated as either homosexual or bisexual, with a high degree of certainty, ranged from 7 to 10% for brothers and 3 to 4% for sisters. These estimates are higher than recent comparable population-based estimates of homosexuality, supporting the importance of familial factors for male homosexuality. Estimates of lambda s for male homosexuality ranged from 3.0 to 4.0. None of the samples showed a significantly greater proportion of maternal than paternal homosexual uncles or homosexual male maternal first cousins. Although our results differed significantly with those of some prior studies, they do not exclude the possibility of moderate X-linkage for male sexual orientation.