OBJECTIVE Hypocapnia causes cerebral arterial constriction, whereas volatile anaesthetics cause dilatation. The purpose of this study was to compare the direct effects of halothane, isoflurane and sevoflurane on hypocapnia-induced constriction of isolated cerebral arteries in vitro. METHODS Basilar and middle cerebral arteries of mongrel dogs (n = 11) were cut into rings and mounted for isometric tension recording in organ baths containing Krebs' bicarbonate solution, aerated with CO2 5% and O2 95% at 37 degrees C. After constriction with 20 mM KCl, hypocapnia was induced by replacing the aerating gas with CO2 2.5% and O2 97.5% in the presence or absence of anaesthetics. RESULTS Exposure of cerebroarterial rings to the hypocapnic gas produced sustained vasoconstriction (418 +/- 19 mg), reaching a plateau within 10 to 15 min. Halothane (0.5, 1, 2 MAC) attenuated the hypocapnia-induced constriction (P < 0.05). In contrast, isoflurane and sevoflurane attenuated this constriction only at 2 MAC (P < 0.05). Attenuation by halothane was greater than that by isoflurane or sevoflurane at each concentration (P < 0.05). NG-nitro-L-arginine (3 x 10(-5) M) did not alter the contractile response to hypocapnia. When a similar degree of constriction was induced by addition of 10 mM KCl, halothane (1 and 2 MAC) preferentially attenuated the constriction induced by hypocapnia to a greater extent than that induced by 10 mM KCl (P < 0.01). CONCLUSIONS Hypocapnia-induced vasoconstriction of isolated dog cerebral arteries precontracted with KCl is more susceptible to halothane than isoflurane or sevoflurane. This may account for the greater increase in cerebral blood flow during halothane than isoflurane or sevoflurane anaesthesia.