Forty-five measurements of diameters of 12 human cerebral arteries were performed during 10 craniotomies under moderate changes in mean blood pressure and end tidal CO2. The mean change in blood pressure was 30 +/- 16 mm Hg (standard deviation) and that of end tidal CO2 was 14 +/- 6 mm Hg (standard deviation). These changes were induced with nitroprusside, phenylephrine, and adjustment of ventilator rate. Measurements were made through the operating microscope focused at the highest power, with meticulous attention to constant angle and distance from the artery. The mean diameter change in the large cerebral arteries (carotid, middle cerebral artery, vertebral artery) was less than 4%, but the smaller arteries (anterior cerebral artery, M2 segment of middle cerebral artery) showed diameter changes as large as 29% and 21% to end tidal CO2 and blood pressure changes, respectively. These data suggest that at the time of craniotomy, diameters of the large cerebral vessels do not significantly change during moderate variations in blood pressure and CO2, but that larger changes may occur in smaller vessels. This constancy of diameter suggests that the transcranial Doppler velocities obtained during intraoperative monitoring of craniotomies may closely reflect blood flow through the insonated artery.