Although volatile anesthetics result in cerebral arterial dilation, the precise mechanisms underlying this effect are not known. In vitro tension recordings were used to study the vasodilating potencies of halothane and isoflurane in isolated cerebral vessels and to examine the possible role of the endothelium in modulating any effects observed. Cylindrical segments of the rabbit basilar artery and midline ear artery from the same animal were placed in a flow-through bath of 37 degrees C oxygenated (95% O2/5% CO2) physiologic salt solution and stretched to a resting tension of approximately 2,000 dynes. They were then constricted with 3.0 x 10(-2) M K+, 1.0 x 10(-3) M norepinephrine, or 5.0 x 10(-6) M serotonin and exposed to either halothane or isoflurane at concentrations of approximately 0.5, 1.0, 1.5, and 2.0 MAC in varied order for 15 min at each concentration. A 30-min period of perfusion with anesthetic-free, vasoconstrictor-containing perfusate separated successive exposures to an anesthetic. Vessels prepared in this fashion retained their responsiveness to both vasoconstrictors and volatile anesthetics for as long as 4 h. They also relaxed appropriately to acetylcholine, indicating that the endothelium was intact. Concentrations of volatile anesthetic in the tissue perfusate were directly measured using gas chromatography, and the relationship between bath concentrations (expressed as MAC fractions) and the degree of relaxation were determined. The data were analyzed by parallel line regression. Halothane was found to be a significantly more potent vasodilator of the isolated basilar artery than was isoflurane. For example, in K(+)-constricted vessels, the concentration of halothane needed to produce a 50% reduction in tension was 1.32 MAC, compared with 1.66 MAC for isoflurane. Comparable differences were found in the basilar artery in the presence of other constrictors. However, there was no significant difference between the two agents in their effects upon the ear artery. In a separate series of experiments, the endothelium of basilar artery segments was removed by drying. Removal was confirmed by observing a diminished dilator response to acetylcholine. These vessels were subsequently constricted with K+, and relaxation dose-response curves were obtained for both halothane and isoflurane. There were no differences in the dose-response curves for deendothelialized versus intact vessels, with halothane still the more potent relaxant after endothelial removal. These data demonstrate that halothane and isoflurane cause a dose-dependent relaxation of rabbit cerebral vessels, regardless of the vasoconstrictor used. Halothane was a more potent relaxant of the basilar artery when expressed on a MAC-fraction basis.(ABSTRACT TRUNCATED AT 400 WORDS)