We investigated whether halothane (HAL), administered via cerebral cortical suffusion at concentrations of 1, 2, and 3%, could induce cerebral microvascular dilatation in vivo and whether the vasodilatory response was dependent on nitric oxide (NO) synthesis. The studies were performed using N2O/fentanyl-anesthetized, paralyzed, and mechanically ventilated rats. A closed cranial window and an intravital microscopy technique were employed. This system permitted the controlled delivery of various vasoactive agents in an artificial cerebrospinal fluid (aCSF) solution and the measurement of diameters of pial arterioles and venules. Each experiment included evaluations of (a) the direct smooth muscle relaxing action of NO, using sodium nitroprusside (SNP), and (b) the capacity for generation and release of endogenous NO, using adenosine diphosphate (ADP). Following confirmation of an intact NO-relaxing and generating capacity, HAL (in aCSF) was suffused at increasing concentrations. Nitric oxide synthase (NOS) inhibition was established with topical nitro-L-arginine (L-NA) or its methyl ester (L-NAME) and the above sequence repeated. The results for rats treated with L-NA (n = 5) or L-NAME (n = 5) were analyzed separately and as a combined group. No significant differences in vascular responses were observed when comparing the two groups. Initially, both SNP and ADP produced significant diameter increases (all groupings) in arterioles (14-28% change) and venules (14-25% change). For all groups, suffusions of 1 to 3% HAL produced arteriolar dilation, ranging from a 10 to 25% increase over baseline diameter. A statistically significant dose dependency was only observed with the combined data.(ABSTRACT TRUNCATED AT 250 WORDS)