The effect of early operation on cerebral vasospasm was studied in 150 patients with aneurysmal subarachnoid hemorrhages who fulfilled all of the following criteria: admission by day 2 after subarachnoid hemorrhage, no rebleeding, clinical grades I to IV on admission, subarachnoid hemorrhage alone on computed tomography scan, not operated on between days 4 and 20, and availability of bilateral carotid angiograms done by day 2 and redone between days 7 and 9. The patients were divided into two groups: those operated on by day 3 (group 1: 116 patients) and those operated on after day 20 or not operated on (group 2: 34 patients). Severity of both subarachnoid hemorrhage on computed tomography scan and angiographic vasospasm were graded into 0-IV. Angiographic vasospasm was observed in 95% of group 1 and in 88% of group 2 patients. A significant difference could not be found between groups 1 and 2 in the angiographic vasospasm grades. The incidence of symptomatic vasospasm in group 1 was 18%, which was significantly lower than the 44% in group 2. In group 1 patients with subarachnoid hemorrhage grades II to III, the incidences of symptomatic vasospasm and low density area on computed tomography scan were 13% and 10%, respectively. Both of these rates were significantly lower than those in group 2, which were 50% and 36%, respectively. However, in patients with subarachnoid hemorrhage grade IV, no differences could be found between groups 1 and 2. There was a close correlation between the angiographic vasospasm grades and the incidence of symptomatic vasospasm in group 1. However, in group 1, no correlation could be observed between the site of ruptured aneurysms or the timing of operations and vasospasm. Although there is still a limit to the effect of early operation on cerebral vasospasm in patients with subarachnoid hemorrhage grade IV, symptomatic vasospasm after subarachnoid hemorrhage may be ameliorated by early operation in patients with subarachnoid hemorrhage grades II to III.