Transient hypotension has been observed in patients after rapid intravenous administration of mannitol, 25 per cent, in clinical doses. These studies were conducted to determine the mechanism for the hypotension, to determine dose and rate of injection response curves in rabbits, and to determine which vascular beds were most reactive. Studies in six patients showed mean decreases in blood pressure of 23 +/- 6.0 per cent (+/-SE) and in total peripheral resistance of 38 +/- 7 per cent after infusion of mannitol. Studies in 18 patients during cardiopulmonary bypass with mechanically fixed cardiac output demonstrated decreases in mean blood pressure of 30 +/- 5 to 40 +/- 3 per cent, depending on dose and rate of administration of mannitol. Patients not on bypass compensated for large decreases in total peripheral resistance by increases in cardiac output (3.6 +/- .4 at baseline to 4.4 +/- .4 l/min) during mannitol-induced hypotension with no change in heart rate. Serum osmolality increased as blood pressure decreased. Significant but clinically unimportant decreases in sodium and potassium ions, hemoglobin, pH, and base excess values were observed. Studies in 18 rabbits showed that the greater the dose or rate of injection of mannitol the greater the decrease in blood pressure. Injection of radiolabeled microspheres in rabbits demonstrated a near doubling of blood flow to skeletal muscle tissue during the hypotension. This occurred with both equiosmotic hypertonic glucose (17 +/- 3 to 32 +/- 7 per cent) and mannitol (17 +/- 1 to 31 +/- 5 per cent), but not after isotonic saline solution. Changes in blood flow to other organ beds were variable and unimportant. The results suggest that hypotension following the intravenous administration of hyperosmotic solutions is due primarily to vasodilation in skeletal muscle.