From the data presented above we believe that there are two types of depressive illnesses. One of these is depression spectrum disease, the prototype of which is the early onset female who has considerable alcoholism among primary male relatives and more depression in female than male relatives. The other type is pure depressive disease which has as its prototype the late onset male. The late onset male has equal amounts of depression in male and female relatives and no familial increase in alcoholism over ordinary expectation. Depression spectrum disease is characterized also by having a higher degree of total familial illness (alcoholism + depression) than pure depressive disease. It is entirely possible that there are considerable differences in response to treatment in these groups. For further study the presence of alcoholism in the family might well be used as a marker for the illness and a group of patients who have alcoholism in primary relatives compared to a group of patients in whom no alcoholism is found. These findings present a differentiation based on family background, which quite possibly is genetic. Not only may there be a differential response to various treatments between depression spectrum disease and pure depressive disease but there may be considerable differences in symptomatology and course of illness.