Full clinical and laboratory details of 203 patients with postoperative jaundice were submitted to a panel of hepatologists. All patients whose jaundice may have had an identifiable cause were excluded, which left 76 patients with unexplained hepatitis following halothane anaesthesia (UHFH). Hepatitis in 95% of these cases followed multiple exposure to halothane, with repeated exposure within four weeks in 55% of cases. Twenty-nine patients were obese, 52 were aged 41-70, and 53 were women. Thirteen patients died in acute hepatic failure. Rapid onset of jaundice after anaesthesia, male sex, and obesity in either sex were poor prognostic signs. Of the clinical stigmata of hypersensitivity, only eosinophilia was impressive. The UHFH group had a much greater incidence of liver kidney microsomal (LKM) and thyroid antibodies and autoimmune complement fixation than those patients whose jaundice related to identifiable factors. Thirteen of the 19 patients with LKM antibodies also had thyroid antibodies. In six patients retested two to three years later LKM antibodies had disappeared, although thyroid antibodies persisted. Rapidly repeated exposure to halothane may cause hepatitis, but such a complication is probably rare. Possibly obese women with a tendency to organ-specific autoimmunity may be more at risk. Nevertheless, the comparative risks of rapidly repeated halothane or non-halothane anaesthesia cannot be determined from the present data. If alternative satisfactory agents are available halothane should be avoided in patients with unexplained hepatitis after previous exposure, although in three to five patients with UHFH who were re-exposed to halothane jaundice did not recur.