The circulatory effects of iv injections of hyperosmolar solutions were studied both in intact dogs with aortic flow probes and in dogs using a standard right heart-bypass preparation. Serial iv injections of 20 ml of 10% NaCl or 50 ml of 25% mannitol produced reproducible episodic vasodilation characterized by falls in mean aortic pressure from 99 +/- 10 (SE) to 61 +/- 6 Torr and increases in aortic flow from 2.20 +/- 0.06 1/min to 3.12 +/- 0.28 (P less than 0.01). Systemic vascular resistance decreased (P less than 0.01) with each injection and serum osmolarity increased (P less than 0.01); however, there was a poor correlation between these two variables (r = -0.24). Because the mechanism of these physiological changes is unclear, the following experiments were performed to determine whether they were due to the release of vasoactive chemical mediators. We measured arterial and venous plasma histamine, a mediator released systemically in IgE-mediated anaphylactic reactions, but found no changes in histamine levels. Furthermore, pretreatment with both H1 and H2 blockers (diphenhydramine and cimetidine), agents that blocked histamine-induced hypotension, did not prevent hyperosmolar vasodilation. Also, indomethacin (a cyclooxygenase pathway inhibitor of prostaglandin synthesis) did not affect hyperosmolar vasodilation or the fall in systemic vascular resistance. Therefore, hyperosmolar vasodilation is not caused by the systemic release of histamine or by the effects of prostaglandins. The mechanism of these reactions is unknown, but it may be due to direct local effects of hyperosmolar solutions on vascular smooth muscle, perhaps mediated by local fluid and electrolyte shifts.