In order to test the hypothesis that age of onset of manic-depressive psychosis may be associated with family history, the authors investigated 479 clinical records of patients suffering from affective disorders and admitted to Hospital during the period 1966-1981. Given the retrospective nature of the study, standardized diagnostic criteria were not possible. A positive family history of psychopathological disorders has been found associated with an early age of onset in comparison with a negative family history; the bipolar type has been found associated with an early age of onset in comparison with the unipolar type. Nevertheless, both these variables and their interaction appeared to be insufficiently predictive. The possible influence of other variables such as sex, marital status, years of education, migration, residence have been investigated, but none of them gave any significant result. This suggests the presence of other more predictive variables, which have not considered in this paper.