Epidemiological studies have consistently found women to be at greater risk than men for affective disorders. This sex effect may help clarify genetic transmission and heterogeneity. Data from eight family studies of unipolar and eight family studies of bipolar probands were used to calculate family resemblance sex ratios. These observed sex ratios were then compared to sex ratios predicted by X-linked and nonfamilial effects models. Maximum likelihood estimation of competing models revealed that X linkage was not a good fit to the unipolar data. The bipolar studies were not consistent with either the X-linked or the nonfamilial effects model.