The influence of premedication on induction characteristics and respiration was studied in 40 children breathing spontaneously during enflurane-nitrous oxide anaesthesia. Two different premedications were used. Twenty children (Group DA) received a rectal solution containing diazepam 0.25 mg kg-1 and atropine 0.015 mg kg-1 and 20 (Group DMS) received a rectal solution of diazepam 0.5 mg kg-1, morphine 0.15 mg kg-1 and scopolamine 0.01 mg kg-1. The children in Group DMS had a significantly higher preoperative sedative score (P less than 0.01), faster induction of anaesthesia (P less than 0.01), lower occurrence of airway problems during induction and a smoother intubation (P less than 0.05) than the children in Group DA. However, the end-tidal carbon dioxide tensions were higher and the occurrence of apnoea was more common in Group DMS than in Group DA. Thus it was concluded that if the more sedative premedication (DMS) is to be used for enflurane anaesthesia in children, controlled ventilation would be preferable.