Two groups of 20 patients scheduled for major abdominal surgery or thoracic surgery received analgesia by thoracic peridural route: 850 and 837 micrograms.kg-1 of bupivacaine, plus 4.30 and 4.20 micrograms.kg-1 of fentanyl. Anaesthesia was induced and maintained with either infusion of propofol 0.2% or infusion of methohexital 0.1% and patients were intubated and ventilated. The quality of induction was good in the two groups (3.48 mg.kg-1 in 4 min with propofol and 2.76 mg.kg-1 in 5 min with methohexital). But only 4 of propofol group needed vecuronium for intubation; they were 19 in the other group. Maintenance was extremely smooth with propofol (0.088 mg.kg-1.min-1) in 18 cases. In contrast, poor anaesthetic control was obtained with 14 methohexital patients and had to be abandoned. Hemodynamic data show fc, Pa and Ppa decreased with propofol but increased with methohexital in response to laryngoscopy and intubation. In this study, the onset of 5 sinusal bradycardias was noted with propofol. Two of them were severe 39 and 38 b.min-1 with major decrease of LVSWI to 32.45 and 24.47 g.mm-2. The role of hypovolemia (Bainbridge reflex) or vagomimetic effect of propofol is discussed. Nevertheless, this study shows that propofol given by infusion can better achieve adequate anaesthesia than methohexital.