We have conducted a randomized, double-blind comparison of 4% and 8% sevoflurane for induction of anaesthesia in unpremedicated patients aged more than 60 yr. Sevoflurane was inhaled in 50% nitrous oxide using a vital capacity breath technique, and mean, systolic and diastolic arterial pressures and heart rate were monitored continuously using a Finapres cuff. In the 8% sevoflurane group, time to successful laryngeal mask insertion was significantly shorter (mean 168 (SD 34) s vs 226 (62) s; P < 0.01) and achieved more often at the first attempt than in the 4% sevoflurane group. Arterial pressures were lower in the 8% group, but this was not significant. No patient had apnoea lasting longer than 1 min. A total of 69% of patients described induction as pleasant and 85% would choose to have it again. We conclude that compared with 8% sevoflurane, the use of 4% sevoflurane in the elderly resulted in greater cardiovascular stability but at the cost of prolonged and occasionally unsuccessful induction.