We evaluated the role of cortisol as a feedback inhibitor of the vasopressin response to arterial hypotension. Conscious dogs (n = 5) were each studied on 10 different days. There were five different pretreatments: 1) isotonic saline (control), 2) 5.5 or 3) 11 nmol cortisol.kg-1.min-1 iv for 30 min, 4) 6.8 pmol ACTH-(1-24).kg-1.min-1 iv for 30 min, or 5) 1.5 mg dexamethasone im given the night before experimentation. These pretreatments were followed by a stimulus period (30 min) during which mean arterial pressure was decreased a moderate (-19 +/- 1 mmHg) or severe (-29 +/- 1 mmHg) degree with a controlled infusion of intravenous sodium nitroprusside. The vasopressin response to moderate hypotension was not consistently inhibited by any of the pretreatments. In contrast, the large vasopressin response to severe hypotension in the control experiments (from 3.1 +/- 0.3 to 270 +/- 113 pg/ml) was significantly attenuated by cortisol infusion in a dose-dependent manner. Adrenocorticotropin infusion was more effective than dexamethasone as an inhibitor of the vasopressin response to severe hypotension. The data suggest that physiological increases in cortisol inhibit the large increase in vasopressin in response to severe arterial hypotension.