Anaesthesia with continuous i.v. ketamine and 65% nitrous oxide in oxygen was given to a total of 49 patients undergoing major abdominal surgery. A control group was premedicated with atropine and other groups received in addition rectal diazepam or clorazepate i.v. For further patients had been on oral diazepam or barbiturates for 1-14 years; as premedication they received atropine alone. The anaesthetic technique gave good operative conditions in the 4 groups of patients. The haemodynamic stimulation of ketamine was significantly reduced in patients premedicated with diazepam. Psychotomimetic side effects were not prominent in any of the groups. Patients premedicated with diazepam required a lower rate of ketamine infusion as compared to controls during the initial 30 min of anaesthesia. The patients in the other groups did not differ from the control group in this respect. There were large differences in metabolic pattern between the groups. As compared to the controls, the patients on long-term diazepam or barbiturates had high concentrations of hydroxylated metabolites, with levels higher than that of norketamine. The patients pretreated with diazepam had very low plasma levels of hydroxylated metabolites. Clorazepate premedication did not significantly affect the metabolism of ketamine. The biological half-life of ketamine was significantly increased in the diazepam-treated group, and it was shortened in those on long term treatment with barbiturates or diazepam.