Haemodynamic changes and side-effects of induction of anaesthesia with etomidate were evaluated in 60 ASA Class I or II patients. The objective was to find an optimal pre-induction dose of fentanyl which eliminated haemodynamic changes and side-effects during induction and intubation without introducing other problems. Patients were randomly assigned to four groups according to the pretreatment dose of fentanyl (Group I = 2 ml normal saline; Group II = 100 micrograms of fentanyl; Group III = 250 micrograms of fentanyl; Group IV = 500 micrograms of fentanyl) administered intravenously five minutes prior to induction of anaesthesia with etomidate, 0.3 mg/kg. There was an increasing incidence of apnoea (53, 87, 87 and 100% in Groups I-IV respectively) and a decreasing incidence of myoclonus (60, 33, 13 and 0% in Groups I-IV respectively) and injection pain (53, 13, 7 and 0% in Groups I-IV respectively), P less than 0.002 chi-square test for linear trends, with increasing fentanyl dosage. The incidences of postoperative nausea and vomiting were similar in the four groups. There were also significant linear regression relationships (P less than 0.01 ANOVA for linear regression) between increasing doses of fentanyl administered before etomidate and the prevention of increases in systolic blood pressure and heart rate during the induction-intubation sequence. The data demonstrate that increasing pre-induction doses of fentanyl are more effective at minimising side-effects and preventing increases in systolic arterial blood pressure and heart rate but also increase the incidence of apnoea during induction. The results suggest that 500 micrograms of fentanyl is an ideal pretreatment dose in fit patients prior to anaesthetic induction with etomidate.