Crow et al. [1993: Am J Med Genet (Neuropsychiatr Genet) 48:159-160] have reported excess sharing of alleles by male sibling pairs with schizophrenia, at a triplet repeat marker within the androgen receptor gene, indicating that mutations at or near this gene may be a risk factor for males. In this report, we describe a pair of male siblings concordant for both schizophrenia and Reifenstein syndrome, which is caused by a mutation in this gene. This provides support for the hypothesis that the androgen receptor may contribute to liability to develop schizophrenia. Because of this, we have examined a collection of 23 pedigrees multiply affected by schizophrenia for linkage to the androgen receptor. We have found no evidence for linkage by both the LOD score and affected sibling-pair methods, under a range of genetic models with a broad and narrow definition of phenotype, and when families with male-to-male transmission are excluded. However, because of the small number of informative male-male pairs in our sample, we cannot confirm or refute the excess allele sharing for males reported by Crow.