Psychiatrists used a semi-structured Standardized Psychiatric Examination method to examine 810 adults drawn from a probability sample of eastern Baltimore residents in 1981. Of the population, 5·9% was found to be significantly depressed. DSM-III major depression (MD) had a prevalence of 1·1% and ‘non-major depression’ (nMD), our collective term for the other depressive disorder categories in DSM-III, had a prevalence of 3·4%. The two types of depression differed by sex ratio, age-specific prevalence, symptom severity, symptom profiles, and family history of suicide. Analyses using a multiple logistic regression model discerned that both types of depression were influenced by adverse life events, and that nMD was influenced strongly by gender, marital status, and lack of employment outside the home. Neither type of depression was influenced by income, education, or race. This study validates the concept of major depression as a clinical entity. Future studies of the aetiology, mechanism, and treatment of depression should distinguish between these two types of depression.