Nitric oxide (NO), also known as endothelium-derived relaxing factor, is produced by endothelial nitric oxide synthase (eNOS) in the intima and by neuronal nitric oxide synthase (nNOS) in the adventitia of cerebral vessels. It dilates the arteries in response to shear stress, metabolic demands, pterygopalatine ganglion stimulation and chemoregulation. Subarachnoid hemorrhage (SAH) interrupts this regulation of cerebral blood flow. Hemoglobin, gradually released from erythrocytes in the subarachnoid space, destroys nNOS-containing neurons in the conductive arteries. This deprives the arteries of NO, leading to initiation of delayed vasospasm. But such vessel narrowing increases shear stress, which stimulates eNOS. This mechanism normally would lead to increased production of NO and dilation of arteries. However, a transient eNOS dysfunction evoked by an increase in the endogenous competitive NOS inhibitor, asymmetric dimethylarginine (ADMA), prevents this vasodilation. eNOS dysfunction has been recently shown to be evoked by increased levels of ADMA in cerebrospinal fluid (CSF) in response to the presence of bilirubin-oxidized fragments (BOXes). A direct cause of the increased ADMA CSF level is most likely decreased ADMA elimination owing to disappearance of ADMA-hydrolyzing enzyme [dimethylarginine dimethylaminohydrolase II (DDAH II)] immunoreactivity in the arteries in spasm. This eNOS dysfunction sustains vasospasm. CSF ADMA levels are closely associated with the degree and time course of vasospasm; when CSF ADMA levels decrease, vasospasm resolves. Thus, exogenous delivery of NO, inhibiting the L-arginine-methylating enzyme or stimulating DDAH II, may provide new therapeutic modalities to prevent and treat vasospasm. This paper will present results of pre-clinical studies supporting the NO-based hypothesis of delayed cerebral vasospasm development and its prevention by increased NO availability.