Plasma dexamethasone concentrations and cortisol response to dexamethasone were measured in 29 normal healthy volunteers, 23 depressed patients, and 10 patients with anorexia nervosa at 4:00 PM postdexamethasone. In each of the 3 groups, nonsuppressors had lower dexamethasone concentrations than suppressors. Of the subjects with plasma dexamethasone at or below 0.7 ng/ml, a significantly higher proportion (48%) were nonsuppressors compared to the proportion above 0.7 ng/ml (14%), all of whom were patients. Plasma dexamethasone concentrations in a subgroup of depressed nonsuppressors were high (mean 1.35 ng/ml), whereas the remainder were low (0.42 ng/ml) and were similar to the normal nonsuppressors (0.35 ng/ml), suggesting different mechanisms for nonsuppression in the subgroups. Plasma dexamethasone concentrations were similar in nonendogenous and endogenous depressives, in men and women, and in medicated and drug-free patients. None of the variables of age, weight, history of weight loss, Hamilton depression rating score, predexamethasone cortisol, or postdexamethasone cortisol were significantly correlated with plasma dexamethasone, except for body weight and a history of weight loss in the depressed group only. Mean plasma dexamethasone concentrations increased significantly from week 1 to week 2 in 7 depressed patients, whereas plasma cortisol decreased; however, the relationship between dexamethasone and cortisol varied considerably for individual patients.