background and objectives: the use of different mac multiples of sevoflurane and halothane, and the different effect of n2o on the mac of these agents in children, are usually argued as the reason for the same induction and recovery times of these agents which have different blood solubilities. this study evaluated the induction time to different mac multiples and the recovery time of anesthesia maintained with a fixed expired concentration (1 mac) of halothane or sevoflurane, associated or not to n2o, in children under epidural caudal anesthesia. methods: participated in this study 63 children allocated in 4 groups according to the anesthetic drug used. group 1 = halothane; group 2 = halothane plus n2o; group 3 = sevoflurane; group 4 = sevoflurane plus n2o. all children were submitted to caudal epidural blockade. mask induction was started with 1 mac of halogenate followed by 0.5 mac increments at every three respiratory movements until a maximum of 3 mac was obtained. the following parameters were recorded: heart rate, systolic and diastolic blood pressure, inhaled agent expired fraction and times for induction, recovery and response to commands. results: duration of surgery did not differ significantly among groups. the times to eyelash reflex loss, end of induction, emergence and oriented responses for groups 1 and 2 were longer than for groups 3 and 4 (p < 0.001) without significant differences between groups 1 and 2 and between groups 3 and 4. there were no differences in heart rate and systolic and diastolic blood pressure. conclusions: in the age range studied and with the anesthetic technique used, induction and recovery times were different between the halothane and the sevoflurane group, but did not differ when nitrous oxide was added to the halogenate.