The authors performed a randomized, prospective trial comparing enflurane, halothane, and isoflurane (each administered with nitrous oxide) to establish which inhaled anesthetic produced the fewest complications and the most rapid induction of anesthesia for children undergoing general anesthesia for diagnostic procedures as oncology outpatients. Sixty-six children, ranging from 8 months to 18 years, underwent a total of 124 anesthetics. Induction of anesthesia (time from placement of facemask to beginning of skin preparation) was faster with halothane (2.7 +/- 1.0 min, mean +/- SD, n = 46) than with enflurane (3.2 +/- 0.8 min, n = 43) or isoflurane (3.3 +/- 1.2 min, n = 35). Emergence from anesthesia (time from completion of the procedure to spontaneous eye opening) was more rapid with enflurane (4.7 +/- 4.4 min) than with halothane (6.2 +/- 4.5 min) or isoflurane (6.2 +/- 3.9 min). Total time from the start of procedure until discharge was longer with isoflurane (25.1 +/- 6.8 min) than with enflurane (21.5 +/- 8.6 min) or halothane (22.3 +/- 7.6 min). During induction, the incidence of laryngospasm was greatest with isoflurane (23%) and the incidence of excitement least with halothane (13%). During the maintenance of, emergence from, and recovery from anesthesia, coughing occurred most frequently with isoflurane. During the recovery period, headache occurred most frequently with halothane (9%); there were no significant differences in the incidence of nausea, vomiting, hunger, or depressed effect. The authors conclude that the rapid induction and minimal airway-related complications associated with halothane anesthesia make it an excellent anesthetic agent for pediatric patients undergoing short diagnostic procedures.