The authors used a large (n = 437) sample of depressed inpatients to determine if DSM-III criteria applied in a clinical setting defined distinct subgroups. Patients with diagnoses of depression, depression with melancholia, and depression with psychotic features were compared by age, sex, diagnoses (Axes I-V), treatment received, and treatment outcome. These data failed to support the DSM-III distinction between a melancholic and psychotic subtype. However, compared to other patients with major depression, patients with melancholic and/or psychotic features were: (a) older, (b) more likely female, and (c) more likely to be treated with somatic therapy.