Previous studies have examined whether calcium antagonists given sublingually, intravenously, or orally affect an experimentally induced vasospasm. This study was designed primarily to determine the efficacy of nimodipine given intracisternally in reversing the cerebral vasospasm produced in dogs by the injection of 4 ml of autogenous blood into the cisterna magna. The magnitude of the vasospasm and the vascular response to nimodipine were determined arteriographically. One group of animals was studied acutely to determine the dose and pharmacodynamic effects of intrathecal nimodipine, a second group was studied 1 to 4 days after the intrathecal injection of blood, and a third group was studied 3 to 6 days after successive injections of cisternal blood given 48 hours apart. Some animals were studied further 24 hours after the nimodipine treatment. Intrathecal administration of 4 ml of 10(-3) M nimodipine promptly and completely reversed the cerebral vasospasm in all groups. This effect lasted at least 4 hours but had disappeared by 24 hours. The intrathecal administration of nimodipine produced a transient drop in blood pressure. In contrast, nimodipine given sublingually (0.28 to 0.58 mg/kg) or intravenously (0.1 mg/kg) produced a persistent hypotensive effect without affecting the vasospasm. The results show that the presence of nimodipine in the cerebrospinal fluid will reverse the blood-induced vasospasm, and suggest that this calcium antagonist administered intrathecally would be an effective emergency treatment of cerebral arterial spasm.