Affective disorder (AD) is one of the major forms of functional psychoses. Although the mode of transmission is uncertain, family, twin and adoption studies strongly suggest a genetic involvement. Because a basic biochemical abnormality is not known, direct analysis of the disease using a probe for the defective gene is not possible. However, a specific locus can be tested for its relevance to the aetiology of AD by genetic linkage, using restriction fragment length polymorphisms (RFLPs). Using probes for the c-Ha-ras-1 oncogene and the insulin gene, Gerhard et al. and Egeland et al. found convincing evidence for close linkage between these markers and a locus for AD in a large Old Order Amish pedigree. In an attempt to confirm this finding, we examined three bipolar pedigrees outside the Amish population. Our results indicate the absence of linkage from 0 to 15% recombination frequency between AD and the insulin gene-HRAS1 region in these pedigrees.