Recent work suggests the existence of a dual corticosteroid feedback mechanism of stress-induced ACTH secretion in the rat. This possibility led us to study the kinetics of suppression of ACTH levels by corticosteroid administration in patients with nonstress ACTH hypersecretion secondary to hypoadrenocorticism. Cortisol was administered according to different protocols, which were chosen to provide extreme variations of the input signal. By this means, two phases of suppression of ACTH levels could be differentiated. A first decrease occurred without latency whenever, and as long as, plasma cortisol levels were rising. There was a linear regression between the logarithm of the increments in cortisol concentrations and the decrease in ACTH levels per minute (r = 0.951) (differential or rate-sensitive feedback mechanism). Neither the absolute doses of cortisol, nor plasma cortisol concentrations were closely correlated with the degree of suppression of ACTH by this rapid mechanism. A second decrease in ACTH levels began congruent with30 min after corticosteroid administration. In this case there was a significant linear regression between the degree of inhibition of ACTH levels and the cortisol doses (r = 0.997) (integral or dose-sensitive feedback mechanism). The dose-sensitive feedback effects of dexamethasone were less than might have been predicted from its relative anti-inflammatory potency. No rate-sensitive effects were seen with dexamethasone doses of 1.0 or 1.25 mg.