Atracurium infusion requirements were determined in 28 children anesthetized with N2O:O2 narcotic, N2O:O2 halothane (1% inspired), and N2O:O2 enflurane (2% inspired). When the patient was recovering from a bolus dose of 0.4 mg/kg atracurium, a continuous infusion of atracurium was started and the rate was adjusted to maintain 90-99% muscle twitch depression. Patients receiving enflurane anesthesia required atracurium at an infusion rate of 4.9 +/- 0.3 micrograms X kg-1 X min-1 which was a significantly lower rate (P = 0.0001) than those anesthetized with halothane (8.3 +/- 0.4 micrograms X kg-1 X min-1) or with N2O:O2 and narcotic (9.3 +/- 0.5 micrograms X kg-1 X min-1). At the onset of neuromuscular blockade, the twitch response disappeared faster after train-of-four stimulation repeated every 10 s than after single twitch rates of stimulation at 0.1 Hz. In children, during halothane anesthesia after 0.4 mg/kg atracurium, the response of the adductor of the thumb was ablated in 2.0 +/- 0.3 min with train-of-four stimulation, and in 3.7 +/- 0.4 min with single twitch stimulation. The authors recommend the use of a nerve stimulator during continuous infusion of atracurium because of the marked interpatient differences in infusion-rate requirements.