This prospective, randomised multicentre study was designed to determine the clinical profile of 'Diprifusor' target controlled infusion compared with manually controlled infusion of propofol in 562 patients, aged 18-85 years; in a range of surgical procedures in 29 centres. The dose of propofol required for loss of consciousness was statistically significantly lower in the target controlled infusion group [1.69 (0.50) vs. 2.31 (0.75) mg.kg−1, p < 0.001] but the overall rate of propofol administration was slightly, but significantly, higher [12.1 (5.1) vs. 11.0 (6.0) mg.kg−1.h−1, p < 0.05]. The target concentration (CT) required for induction decreased with increasing age and ASA class, with premedication and with the administration of an opioid before induction. However, the amount of opioid given did not influence the CT required for induction, but enhanced the haemodynamic effects of propofol induction in both groups. Most investigators expressed an overall preference for target controlled infusion (93%) and found it easier to use (76%). Despite the lack of experience of most investigators in using target controlled infusion, the clinical profiles of both propofol administrations were similar. Data suggest that the clinical profile of target controlled infusion may be improved with experience, for example by more active titration of CT to effect. Target controlled infusion may well become the preferred choice for anaesthetists.