In Brief Reactive oxygen species (ROS) mediate volatile anesthetic preconditioning. We tested the hypothesis that isoflurane (ISO) generates ROS from electron transport chain complexes I and III. Rabbits (n = 55) underwent 30 min coronary artery occlusion followed by 3 h reperfusion and received 0.9% saline, the complex I inhibitor diphenyleneiodonium (DPI; 1.5 mg/kg bolus followed by 1.5 mg/kg over 1 h), or the complex III inhibitor myxothiazol (MYX; 0.1 mg/kg bolus followed by 0.3 mg/kg over 1 h) in the absence and presence of 1.0 minimum alveolar concentration ISO. ISO was administered for 30 min and discontinued 15 min before coronary occlusion. Infarct size and ROS production (n = 32) were determined using triphenyltetrazolium staining and ethidium-DNA fluorescence, respectively. Adenosine triphosphate (ATP) synthesis in mitochondria obtained from rabbit hearts (n = 24) subjected to drug interventions was measured by luciferin-luciferase luminometry. ISO significantly (P < 0.05) reduced infarct size (19% ± 4%) as compared with control (39% ± 4%). MYX (35% ± 4%), but not DPI (24% ± 2%), abolished this protection. ISO increased ethidium-DNA fluorescence (83 ± 11 U) as compared with control (40 ± 12 U). MYX (35 ± 3 U), but not DPI (78 ± 9 U), abolished ROS generation. DPI and MYX selectively reduced complex I- and complex III-mediated ATP synthesis, respectively. ROS generated from electron transport chain complex III mediate ISO-induced cardioprotection. IMPLICATIONS: The electron transport chain complex III inhibitor myxothiazol, but not the complex I inhibitor diphenyleneiodonium, abolished isoflurane-induced protection against ischemia-reperfusion injury and reactive oxygen species production in rabbits. Reactive oxygen species generated from electron transport chain complex III act in the signal transduction process mediating preconditioning by isoflurane in vivo.