Fifty patients with ruptured aneurysms were operated on within 72 hours after the first subarachnoid hemorrhage (SAH). To prevent symptomatic vasospasm, the patients were given the calcium channel blocker, nimodipine, intravenously (2 mg/hr) for 14 days and orally (60 mg four times daily) for another 7 days. At short intervals (at least every 3rd day) the blood flow velocity in the different segments of the circle of Willis was measured with a noninvasive transcranial Doppler ultrasonography method. Within the first 72 hours after SAH, the velocity was normal in the large branches of the circle of Willis and angiography revealed no signs of vasospasm. The Doppler frequency changes that relate to blood flow accelerated between Days 3 and 10, and maximum blood flow velocities were recorded between Days 11 and 20, with normalization occurring within the following 4 weeks. The changes showed a significant relationship to the source of SAH, the side of the operative approach, and the method of nimodipine administration. A comparison between the angiographically proven diameter of spastic arteries and the Doppler-measured blood flow velocity showed an inverse relationship in flow of the middle cerebral artery and the internal carotid artery that was statistically highly significant (p less than 0.001) while this correlation was only slightly significant in the A1 segment of the anterior cerebral artery (p = 0.054). Seven patients (14%) developed delayed ischemic deficits (DID's), which were all functionally reversible. One patient (2%) died as a result of decompensated vasospasm. Based on the information provided by Doppler measurement of the individual blood flow velocity changes due to vasospasm, preventive hypertensive treatment was introduced to improve the perfusion pressure while patients were still in an asymptomatic stage. Among the last 40 patients who were treated according to this regimen, reversible DID's were observed in only three patients (7.5%) and postoperative angiography to detect vasospasm was not necessary.