We have investigated the effect of four doses of remifentanil on the incidence of respiratory depression and somatic response at incision. Remifentanil was administered as a loading dose of 0.125, 0.25, 0.375 or 0.5 microgram kg-1 and at a maintenance infusion rate of 0.025, 0.05, 0.075 or 0.1 microgram kg-1 min-1, respectively, with an infusion of propofol 6 mg kg-1 h-1. Responses occurred in 88% of patients with remifentanil 0.025 microgram kg-1 min-1 compared with 30-40% in the other groups. Respiratory depression after incision increased from 6% with remifentanil 0.025 microgram kg-1 min-1 to 73% with 0.1 microgram kg-1 min-1. Increases in propofol infusion rate to 7.2-8.4 mg kg-1 h-1 produced adequate maintenance of anaesthesia. Reductions in remifentanil doses to 0.025-0.05 microgram kg-1 min-1 resulted in adequate respiration at the end of surgery in 88% of patients. Maintenance infusions of the two drugs for spontaneous ventilation are likely to be in these ranges. However, the ideal loading doses and infusion rates for induction remain to be established.