Rats were exposed to halothane vapour, 50 p.p.m., or air for a period of four weeks. Within each exposure group, some animals drank plain water, some received water plus phenobarbitone, while some received water plus isoniazid. Halothane exposure resulted in increased serum bromide concentrations and liver injury evidenced by increased serum alanine aminotransferase activity, focal hepatocellular necrosis and fatty change. Administration of isoniazid reduced halothane metabolism by 33% as assessed by serum bromide concentrations, and completely blocked the injurious effects of halothane on the liver, suggesting that halothane metabolism plays a role in halothane hepatotoxicity under these conditions. Administration of phenobarbitone partially prevented the increase in serum alanine aminotransferase activity and hepatocellular necrosis due to halothane. In contrast to isoniazid, phenobarbitone led to a slight increase in halothane metabolism. However, phenobarbitone also caused an increase in liver size, such that the amount of halothane metabolised per gram of liver was reduced by phenobarbitone treatment. These results suggest that metabolism of halothane is an important factor in liver injury due to prolonged, subanaesthetic halothane exposure.