Fifty-four children aged 1-10 years, scheduled for day-case anaesthesia were prescribed either oral midazolam 0.5 mg.kg-1 or a placebo preparation 30-60 min pre-operatively on a double-blind basis. On arrival at the induction room, anxiolysis was satisfactory in 23 out of 24 (96%) children who received midazolam compared with 12 out of 27 (44%) of those who received placebo (p < 0.001); at induction of anaesthesia these proportions were 21 out of 24 (88%) and nine out of 27 (33%) respectively (p < 0.001). The time to early recovery from anaesthesia was somewhat longer in children premedicated with midazolam compared with controls (28.2 vs 21.9 min) (p < 0.05). Similarly, the time to hospital discharge was longer in the midazolam group (244 vs 185 min) (p < 0.01). Analysis of behavioural questionnaires completed by parents 2 weeks after hospitalisation indicated that there were fewer postoperative behavioural disturbances in children premedicated with midazolam compared with controls (p < 0.05).