To establish anesthesia with minimal respiratory and cardiovascular depression using propofol, the effects of varying the rate of delivery on anesthetic induction dose requirements and hemodynamic changes were studied in four groups of 20 patients each undergoing body surface surgery. All patients were premedicated with temazepam and received 1.5 micrograms/kg fentanyl 5 min before induction. Propofol was delivered at 50, 100, or 200 mg/min by the Ohmeda 9000 infusion pump (groups 1, 2, and 3, respectively) or by bolus of 2 mg/kg (group 4) until loss of verbal contact. Anesthesia was maintained thereafter with propofol infused at 6 mg.kg-1.h-1. Using slower infusion rates, induction took significantly longer (124, 92, 62, and 32 s in groups 1, 2, 3, and 4, respectively) and was achieved with significantly smaller doses of propofol (1.40, 1.96, 2.61, and 2.15 mg/kg in groups 1, 2, 3, and 4, respectively). Slow infusion (groups 1 and 2) caused less depression of systolic and diastolic blood pressure than rapid infusion (groups 3 and 4), but the differences were not statistically significant. Patients in groups 3 and 4 had significantly greater decreases in pulse rate and a greater incidence of apnea than did patients in group 1. There was no correlation between the size of the induction dose and subsequent maintenance requirements of propofol. The finding that the sleep dose of propofol is reduced at slower infusion rates has important practical and theoretical implications when considering the relative potencies of intravenous anesthetics.