Arginine vasopressin, a potent vasoconstrictor, does not raise arterial pressure in normal humans even at pathophysiological plasma levels. To examine whether the pressor effect of vasopressin in humans is buffered by baroreceptor reflex inhibition of sympathetic nerve activity, we recorded postganglionic muscle sympathetic nerve activity directly from the peroneal nerve in 12 normal men before, during, and after a 20-minute intravenous infusion of vasopressin, 4 ng/kg/min, that increased mean plasma concentrations from 6.2 ± 0.6 to 320 ± 68 (SE) pg/ml. During the first 5 minutes (n = 8), mean arterial pressure increased from 91 ± 3 to 97 ± 4 mm Hg (p<0.05) and integrated sympathetic nerve activity decreased from 271 ± 45 to 156 ± 33 units (p<0.05). At 15 minutes (n = 12), arterial pressure dit not differ from control values whereas forearm vascular resistance fell (p<0.05) and central venous pressure and heart rate increased (p<0.05). Sympathetic nerve activity remained below control levels throughout the infusion (202 ± 31 vs 254 ± 40 units before infusion; p<0.05). An effect of vasopressin on ganglionic transmission was excluded, since the sympathoexcitatory response to apnea was not attenuated during vasopressin. Thus, pathophysiologic levels of vasopressin in humans cause inhibition of muscle sympathetic nerve activity that is not due to a ganglionic blocking action. The sympathoinhibition may be caused in part by the modest increases in mean arterial and central venous pressures and attendant stimulation of arterial and cardiac baroreceptors. The reflex decrease in sympathetic nerve activity would be expected to buffer the direct vasoconstrictor effects of vasopressin.