Previous in vivo studies of the coronary vascular effects of halothane (HAL) were complicated by varying hemodynamic conditions and global cardiac work demands. Accordingly, the current study evaluated changes in coronary blood flow (CBF) and associated variables during selective intracoronary administrations of HAL in in situ canine hearts using an extracorporeal-controlled pressure perfusion system. Findings during HAL were compared to those during isoflurane (ISO). The left anterior descending coronary artery (LAD) of 8 open-chest dogs anesthetized with fentanyl and midazolam was perfused at constant pressure (109 +/- 2 mm Hg) with HAL-free arterial blood or with blood equilibrated in an extracorporeal oxygenator with HAL (0.5%, 1.0%, 2.0% in 95% O-2-5.0% CO2). In the LAD bed, measurements of CBF were obtained with an electromagnetic flowmeter and used to calculate myocardial oxygen consumption (MVo(2)). Percent segmental shortening (%SS) was measured with ultrasonic crystals. Changes in CBF by HAL were compared to those dur ing maximal vasodilation with adenosine. Separate studies (n = 5) were performed using 1.4% [1 minimum alveolar anesthetic concentration (MAC)] ISO and the findings compared to those during an equianesthetic (1.0%) concentration of HAL. HAL caused concentration dependent increases in CBF, and decreases in MVo(2) and %SS. With 2.0% HAL, the level of CBF was 50% of the maximal adenosine-induced response. At equianesthetic concentrations, HAL caused increases in CBF that were one-third of those caused by ISO, while the decreases in MVo(2) and %SS caused by the drugs were not significantly different. We conclude that HAL has a direct concentration-dependent relaxing action on vascular smooth muscle in the coronary circulation of the in situ canine heart. The ability of HAL to increase CBF significantly while it was reducing local myocardial O-2 requirements by a direct negative inotropic effect attests to the potency of this vasodilator action. HAL was a less potent direct coronary vasodilator than ISO, whereas it had a comparable direct negative inotropic effect.