Halothane anesthesia (1%) administered in 21% oxygen for 4 hr to an outbred strain of guinea pig in the absence of enzyme induction resulted in liver damage in 40 of the 65 animals studied. Necrosis was either confluent around the central veins or in scattered foci throughout the lobules. Damage was present on the second and third days after anesthesia. By day 7 the livers had recovered, evidenced by lack of histological changes and normal serum alanine aminotransferase activity. Administration of halothane in 14 or 80% inspired oxygen did not alter the extent or incidence of liver damage. Major end-metabolites of halothane biotransformation (2-chloro-1,1-difluoroethylene, 2-chloro-1,1,1-trifluoroethane, inorganic fluoride and trifluoroacetic acid) were identified at each oxygen concentration. The metabolic inhibitor SKF-525A significantly decreased the amounts of the volatile metabolites 2-chloro-1,1,1-trifluoroethane and 2-chloro-1,1-difluoroethylene. SKF-525A also decreased the incidence and severity of hepatic damage. Both halothane (1%) and isoflurane (1.1%) anesthesia caused similar reductions in mean arterial blood pressure. However, in contrast to halothane, isoflurane was not hepatotoxic. The results indicate that liver necrosis is unlikely to be caused by anesthesia per se, but rather by hepatotoxic metabolites of halothane. This model offers the opportunity to study the pathogenesis of halothane hepatotoxicity after the administration of halothane alone.