Plasma cortisol responses to the intravenous administration of clonidine hydrochloride and placebo were evaluated in depressed patients and controls. Depressed patients had higher mean baseline cortisol levels than controls. Cortisol levels decreased during the morning study period following both placebo and 2 micrograms/kg of clonidine hydrochloride in the depressed patients, but the cortisol decrease was sixfold greater on the day of clonidine administration; these placebo-clonidine differences were statistically significant, whether calculated on an absolute decrement basis or as a percent change. In contrast, controls responded to clonidine with only a 1.5-fold greater cortisol reduction than that found after placebo, a nonsignificant difference from the day of placebo administration. Reductions in the concentration of plasma 3-methoxy-4-hydroxyphenylglycol following clonidine administration were significantly negatively correlated with baseline plasma cortisol levels, raising the possibility that abnormalities in the responsiveness of the alpha 2-noradrenergic system may be associated with the hypothalamo-pituitary-adrenal (HPA) axis dysfunction found in depressed patients.