Forty-two patients with endogenomorphic depression (ED) and 42 patients with other psychiatric disorders received an overnight dexamethasone test of hypothalamopituitary-adrenal (HPA) suppressibility. Plasma and urinary cortisol measures showed that the ED patients had significantly greater HPA activity before dexamethasone and less complete HPA suppression after dexamethasone. High cortisol vlaues after dexamethasone correlated strongly with spontaneous HPA disinhibition, as indicated by high baseline midnight plasma cortisol levels. Criteria for defining normal suppression responses were developed. All patients with depressive neuroses and most patients with other nondepressive disorders had completely normal responses to dexamethasone. About half of the ED patients had abnormal responses, whether or not they were receiving other drugs at the time of the test. Drug-free patients with depressive neuroses or other disorders showed no abnormal responses to dexamethasone. The effects of psychotropic drugs on the test require further study. Patients with two or more abnormal cortisol values after administration of dexamethasone were identified correctly as ED at confidence levels close to 100%. The dexamethasone suppression test may be of value as a laboratory aid in the diagnosis of "endogenous" depression.