1. The effect of atrial natriuretic peptide on osmotically stimulated thirst appreciation and consequent fluid intake was investigated in healthy man. 2. Six seated male subjects were studied on two occasions: synthetic α-human atrial natriuretic peptide (99–126) (2 pmol min −1 kg −1 ) or placebo (saline, 150 mmol/l NaCl) was infused intravenously for 105 min; 30 min after the start of atrial natriuretic peptide/placebo infusion, hypertonic saline (855 mmol/l NaCl) was infused (0.06 ml min −1 kg −1 ) for 60 min. Subjects were then allowed free access to water for the next 2 h; infusion of atrial natriuretic peptide/ placebo continued for the first 15 min of the drinking period. 3. The plasma atrial natriuretic peptide concentration did not alter significantly during infusion of hypertonic saline and placebo; it rose to a steady state of 12.7 ± 1.1 pmol/l (mean ± SEM) during the infusion of atrial natriuretic peptide and hypertonic saline, and remained at this level during the first 15 min of the drinking period. During infusion of hypertonic saline and atrial natriuretic peptide or placebo, similar increases in plasma osmolality ( P < 0.001) and plasma vasopressin concentration ( P < 0.005) occurred. During infusion of hypertonic saline and atrial natriuretic peptide or placebo, thirst increased significantly over the time course of both studies ( P <0.01), but the effect of atrial natriuretic peptide infusion compared with placebo infusion was to significantly decrease thirst at 60 min. 4. Drinking rapidly abolished thirst and vasopressin secretion before changes in plasma osmolality occurred. Subjects drank significantly less water after atrial natriuretic peptide infusion compared with after placebo infusion ( P <0.01). 5. In conclusion, physiological increases in plasma atrial natriuretic peptide concentrations blunt osmotically stimulated thirst appreciation and attenuate subsequent fluid intake in hyperosmolar man.