The authors used a naturalistic design to replicate studies which validated the familial subtyping of primary depression. Ninety-three inpatients with unipolar depression had received the DST within 1 week of admission and were free of confounding medical problems. A blind rater assigned diagnoses based on chart material recorded before DST results were known. The results supported the earlier conclusions of Schlesser et al. With 8 AM sampling, 53% of familial pure depressive disease patients and 14% of depression spectrum disease patients were nonsuppressors. Differences, though smaller, remained significant with multiple sampling. Only 1 of 24 patients with secondary depression had an abnormal DST. Patterns of nonsuppression among patients with psychotic features resembled those of the larger group.