Family members of bipolar probands have been repeatedly shown to have an increased risk for mood disorders. However, a range of different syndromes in the bipolar spectrum are commonly observed in these relatives. This suggests the hypothesis that these different syndromes may be genetically related. It further suggests that bipolar disorder may be better conceptualized from a genetic standpoint as a quantitative trait. In such a model, the same genes may predispose to a variety of phenotypes ranging from schizoaffective disorder to cyclothymic temperament. Previous attempts to test such a multifactorial model have provided some limited support. However, other studies argue that some forms of bipolar disorder such as bipolar II may be genetically distinct. In this review, various quantitative and categorical models of illness are considered and the data supporting them reviewed. It is proposed that the existing data may best fit a model in which different sets of genes predispose to overlapping phenotypes that are in part both quantitative and distinct in nature.