A hundred and ten women admitted to a psychiatric hospital within 90 days of childbirth were individually matched for age, psychiatric syndrome, and year of admission with women admitted to the same hospital with illnesses unrelated to childbirth. Both groups were followed up after a mean interval of nine years, and 72 matched pairs of patients for whom adequate information was obtained were then compared. The previous and subsequent psychiatric morbidity of these two groups, their subsequent obstetric careers, and the psychiatric morbidity of their first-degree relatives were all very similar. However, the puerperal women had significantly fewer relapses in the follow-up period, fewer committed suicide, and the psychiatric morbidity of their relatives tended to be lower. This better outcome was most marked in puerperal subjects with major depressions; those with manic disorders fared no better than controls. These results suggest that puerperal psychoses are basically the same as affective illnesses occurring at other times but, because childbirth is a uniquely potent precipitant of affective illness, some of those who develop puerperal episodes have a lesser genetic predisposition to affective illness than the generality of women with affective disorders.