Midazolam was given intravenously to 100 children over the age of 8 years to induce anaesthesia and a similar control group was anaesthetised with diazepam. The actions of midazolam in children were found to be similar to those of diazepam. Midazolam acted more quickly in children than has been reported in adults. The effects on the circulation and respiration, and the incidence of postoperative amnesia were similar in children to that which has been reported in adults. The incidence of postoperative vomiting after midazolam was 1% in the recovery room and 32% during the first two postoperative days. Laboratory investigations did not demonstrate any adverse effects. Midazolam has a considerable advantage over diazepam being water soluble and injectable safely without dilution. This clinical trial suggests that midazolam is a satisfactory induction agent for children.