Halothane anesthesia administered to enzyme-induced animals in a hypoxic atmosphere consistently produced hepatic necrosis. Rats pretreated with phenobarbital were exposed to hypoxia at varying intervals after administration of halothane, enflurane, or isoflurane anesthesia. Anesthetics were administered at 1 MAC for 2 h. For each agent, hypoxia consisting of 8 per cent oxygen-balance nitrogen for 1 h was imposed at the end of anesthesia. In other groups of rats, we also used a 15-, 30-, 60-, and 120-min interval of 100 per cent oxygen between 2 h of halothane anesthesia and the imposition of hypoxia. Controls included enzyme-induced animals with and without hypoxia, hypoxia alone, and cage controls. Hepatic injury was graded by histologic examination of the livers. Injury was greater when hypoxia followed halothane anesthesia than when it followed enflurane, isoflurane, or enzyme-induction alone. A difference in injury score existed between control animals and those anesthetized with halothane who received a 15-min interval of oxygen before hypoxia. Combined results from the 15- and 30-min delay groups also were different from control. There was no difference between control and halothane groups when the oxygen interval was 60 or 120 min. The injury score of the enflurane and isoflurane groups were comparable to that of controls. We conclude that the potential for hypoxia-induced liver injury during recovery exists after halothane anesthesia. Neither enflurane nor isoflurane anesthesia produced significant hepatic injury in this model.