To evaluate nasally administered midazolam 0.2 mg · kg-1 for preinduction of anaesthesia in paediatric patients the authors studied ASA 1 patients scheduled for elective surgery. Forty-five children, ages 3 to 126 months, were randomized in three groups : group D (n = 16) received diazepam 0.33 mg · kg-1 orally, group P (placebo) (n = 13) 0.04 ml · kg-1 normal saline via the nasal route ; in group MDZ (n = 16) the children were given intranasal midazolam 0.3 mg · kg-1. The premedication was assessed on a 5-point sedation scale, modified to include the response to mask placement and the quality of the induction of general anaesthesia. The degree of sedation, heart rale, blood pressure, respiratory rate and oxygen saturation levels were recorded on the arrival in the operating room (0 min) and 3, 6, 9, 12 and 15 min (mask placement) after drug administration. With intranasal midazolam sedation was demonstrable at 6 min and was significant at 9 and 12 min. In this group all the children were calm or drowsy. The induction of anaesthesia was equivalent in group D and MDZ but easier than in those patients receiving normal saline. Vital signs did not change during the study period in any of the three groups. Intranasal midazolam was slightly more effective than oral diazepam. In children, it produces anxiolysis and sedation with rapid onset and is an attractive alternative to other routes for preanaesthetic medication. © 1990 Masson, Paris.