Radioactive microspheres (15 mu) were used to measure regional cerebral blood flow during intravenous infusion of nicotine (36 micrograms/kg/min) in anesthetized, open chest dogs. Experiments were conducted with uncontrolled mean aortic pressure and intact autonomic receptors (Series I; n = 9), and in four groups of dogs with mean aortic pressure held constant (Series II); 1) with intact autonomic receptors (n = 6), 2) after beta adrenergic blockade (n = 8), 3) after alpha and beta adrenergic blockade (n = 6), 4) after alpha and beta adrenergic and cholinergic blockade (n = 4). In Series I, nicotine raised mean aortic pressure (+ 72%) and increased flow in cerebral cortex (+ 67%), cerebellum (+ 38%), pons (+ 46%), medulla (+ 39%), and spinal cord (+ 48%). In all regions, but cortex, increases in vascular resistance limited nicotine-induced increases in flow. In Series II, nicotine changed flow only in cortex. Without blockade, nicotine increased cortical flow (+ 38%); but beta blockade abolished this increase in flow. After alpha and beta blockade nicotine again raised cortical flow (+ 29%), and additional cholinergic blockade had no effect on this response. It is concluded that nicotine causes predominant beta receptor mediated vasodilation in cerebral cortex, although it also activates alpha (vasoconstrictor) receptors and a non-adrenergic, non-cholinergic vasodilator mechanism in this region of brain.