The paper reports results of one project from the National Institute of Mental Health Epidemiologic Catchment Area (ECA) Program: the Duke ECA study (also known as the Piedmont Health Survey). To determine if depressive symptoms are different in the depressed elderly, 46 community subjects, over 60 years of age with a current diagnosis of dysthymic disorder and/or major depressive disorder, were compared with 30 community subjects between the ages of 35 and 50. Persons in the same age categories who had no psychiatric disorder were selected as controls. The participants were 3,920 adult subjects in the community (including an elderly oversample) were interviewed. Prevalence of current dysthymia/major depression was 2.8 percent for the elderly and 4 percent for the 35-50 age group. Older depressed subjects were more likely to complain of somatic problems, lethargy, and trouble sleeping, but also were more likely to complain of thoughts about death, a depressed mood, and a sense that life was hopeless. The elderly were less likely to complain of crying spells, feelings of worthlessness, and weight gain. None of these subjects had significant cognitive impairment and the elderly depressed were no more likely to complain of memory problems or difficulty concentrating than their middle-aged counterparts. Older depressed subjects, however, had more objective evidence of increased difficulty with short-term memory. The elderly depressed were more likely than their age-matched controls and the middle-aged depressed to be taking both sedative hypnotic agents and benzodiazepines for their depressive symptoms. Tricyclic antidepressant use was equivalent for the middle-aged and elderly depressed. (Author/LLL)