We measured plasma ACTH and cortisol at 20-min intervals for 24 h in depressed patients and healthy control subjects. The data were analyzed by the PULSAR program to quantitate the number of hormone pulses, their amplitude, length, maximum and interval. We found that in both healthy and depressed subjects the circadian pattern of pituitary-adrenal activity is the result of significant time-related changes in pulse amplitude with no change in pulse frequency. Depressed patients who had an abnormal response to dexamethasone also had changes in pituitary-adrenal rhythm in the unmedicated state. These included ACTH and cortisol pulses whose amplitude, maximum and duration were greater than in the controls as well as a phase advance in the cortisol circadian rhythm. Some of those features were shared by patients who responded normally to dexamethasone suggesting that rhythm-related indices of pituitary-adrenal function may be a more sensitive index of disturbed pituitary adrenal regulation than the Dexamethasone Suppression Test.