Several studies have suggested the utility of separating the type II (BPII) from type I bipolar affective disorder (BPI). The data supporting this suggestion include differences in the course of BPII illness, poorer response of BPII to medications, the higher frequency in BPII patients of non-affective co-morbidity and an increased risk of BPII disorder among relatives of BPII patients. This report of a family in which six sisters and several other family members share the disorder further supports the suggestion that BPII disorder 'breeds true', at least in some families. Since the diagnosis of the BPII disorder is relatively unstable, further study will be needed to establish BPII as distinct from BPI and unipolar disorders. Among others, molecular genetic methods will be useful in the resolution of this issue.