Blood flow velocities in both middle cerebral arteries and regional cerebral blood flow in their perfusion territories were measured simultaneously in 36 healthy subjects. In 20 subjects, the measurements were first performed under basal conditions and then repeated 15-20 min after an intravenous injection of 1 g of acetazolamide. Reproducibility of simultaneous blood flow and velocity measurements was tested by examining 16 subjects on two occasions under basal conditions with an interval of 20 min. Relatively good reproducibility was found for repeated measurements of velocity and blood flow, being best when side-to-side asymmetry was assessed. The increase in blood flow velocities after acetazolamide was symmetrical (right side, 34.2%; and left side, 35.5%), and the velocity increase was significantly correlated with basal values. The increase in cerebral blood flow was also symmetrical (right side, 29.8%; left side, 30.1%) but not correlated with basal flow values. No significant relationship was found between velocity increase and increase in regional cerebral blood flow. This finding is probably not only due to methodological inaccuracies but may suggest that acetazolamide has an effect on the diameter of the middle cerebral artery or on the magnitude of this artery's perfusion territory. This study supports the use of acetazolamide for assessing cerebral vasoreactivity following the definition of lower limits for velocity and flow increase and for asymmetry of the response.