The effects of mannitol on the cerebral microcirculation were studied using isolated perfused intracerebral (parenchymal) arterioles in rats. After vessels developed spontaneous tone, increasing concentrations of mannitol were applied extraluminally. Mannitol induced vasodilation in a dose-dependent manner. In a initial series of vessels, a 2% mannitol solution produced a maximal increase of vessel diameter of 35.8 ± 2.2% (mean ± SEM; n = 5). The degree of vasodilation was comparable to that produced by a sucrose solution of nearly equal osmolarity , suggesting that mannitol produced nonspecific vasodilation by increasing osmolarity. Organ bath solution pH was varied to 6.8 and 7.6 from the control pH of 7.3 in order to examine the effects of mannitol on vessels with different degrees of vessel tone. The mannitol dose-response curves in these vessels were parallel to each other, indicating that vasodilating effects of mannitol are essentially constant irrespective of the initial tone state of the vessel. In a second series of experiments, the effects of intraluminally applied mannitol on intracerebral arterioles were examined and compared to extraluminal application. Intraluminal application of 2% mannitol solutions produced a 42.6 ± 2.0% (n = 5) increase in vessel diameter, which was not significantly different from that produced by extraluminal application (42.5 ± 2.9%; n = 5; P = 0.89). Our results indicate that the direct action of mannitol on vascular smooth muscle in intracerebral arterioles in vasodilation. Our data do not necessarily conflict with the autoregulatory vasoconstriction theory for mannitol action because a relatively small osmolarity change after the intravenous infusion of mannitol in the usual dosage (1 g/kg) would be expected to have a modest effect on vascular diameter when compared with the autoregulatory vasoconstriction induced by decreased blood viscosity after mannitol infusion.