Late-onset depression (greater than or equal to 60 years) is believed to be less associated with a risk of depression in first-degree relatives than early-onset depression. However, family studies in elderly probands fitting the current methodological standards of family studies are not available. The reported family study in geriatric inpatients with unipolar major depression (n = 92) supported the proposed relationship between age at onset and the proposed familial loading. A comparison to families of age-matched controls (n = 33) revealed that relatives of probands with late-onset depression are still at an increased risk of depression. However, late-onset depression was not more common in families of probands with late-onset depression than in families of probands with early-onset depression. Besides the age at onset, the recurrence of depressive episodes defined distinct patterns of familial aggregation.