Clinical and epidemiological studies have consistently revealed an association between alcohol use disorders and both bipolar and nonbipolar mood disorders. However, the evidence regarding the nature of these associations is unclear. The familial patterns of alcohol and affective disorders were examined using data from a controlled family study of probands with alcohol and anxiety disorders who were sampled from treatment settings and the community. The substantial degree of comorbidity between mood and anxiety disorders among probands allowed for the examination of comorbidity and familial aggregation of alcohol and mood disorders. The major findings are that (1) alcoholism was associated with bipolar and nonbipolar mood disorders in the relatives; (2) there was a strong degree of familial aggregation of alcohol dependence and both types of mood disorders were observed; and (3) there was no evidence of cross-aggregation (i.e., increase in mood disorders among probands with alcohol dependence, and vice versa) between alcoholism and mood disorders. The independent familial aggregation of bipolar disorder and alcoholism and the finding that the onset of bipolar disorder tended to precede that of alcoholism are compatible with a self-medication hypothesis as the explanation for the frequent co-occurrence of these disorders. In contrast, the independent familial aggregation and the tendency of an earlier onset of alcoholism than that of nonbipolar depression suggest that unipolar mood disorders are frequently secondary to alcoholism