It is now clear that reactive oxygen species (ROS) can act as signalling molecules in the cerebral circulation under both physiological and pathological conditions. Some major products of superoxide (O(2)(.)(-)) metabolism, such as hydrogen peroxide (H(2)O(2)) and hydroxyl radical (OH(.)), appear to be particularly good cerebral vasodilators and may, surprisingly, represent important molecules for increasing local cerebral blood flow. A major determinant of overall ROS levels in the cerebral circulation is the rate of generation of the parent molecule, O(2)(.)(-). Although the major enzymatic source of O(2)(.)(-) in cerebral arteries is yet to be conclusively established, the two most likely candidates are cyclo-oxygenase and nicotinamide adenine dinucleotide phosphate (reduced form) [NADPH] oxidase. The activity of endogenous superoxide dismutases (SODs) play a vital role in determining levels and effects of all individual ROS derived from metabolism of O(2)(.)(-). The term 'oxidative stress' may be an over-simplification that hides the complexity and diversity of the ROS family in cerebrovascular health and disease. Although a generalised increase in ROS levels seems to occur during several vascular disease states, the consequences of this for cerebrovascular function are still unclear. Because enhanced breakdown of O(2)(.)(-) by SOD will increase the generation of the powerful cerebral vasodilator H(2)O(2), this latter molecule could conceivably act as a compensatory vasodilator mechanism in the cerebral circulation under conditions of elevated O(2)(.)(-) production. Some recent clinical data support the concept of a protective role for cerebrovascular NADPH oxidase activity. Although it is quite speculative at present, if NADPH oxidase were to emerge as a major source of beneficial vasodilator ROS in the cerebral circulation, this may represent a significant dilemma for treatment of ischaemic cerebrovascular conditions, as excessive NADPH oxidase activity is associated with the progression of several systemic vascular disease states, including hypertension and atherosclerosis. Despite data suggesting that antioxidant vitamins can have beneficial effects on vascular function and that their plasma levels are inversely correlated with risk of cardiovascular disease and stroke, the results of several recent large-scale clinical trials of antioxidant supplementation have been disappointing. Future work must establish whether or not increased ROS generation is necessarily detrimental to cerebral vascular function, as has been generally assumed, or whether localised increases in ROS in the vicinity of the arterial wall could be beneficial in disease states for the maintenance of cerebral blood flow.