The choice between regional versus general anaesthesia for elderly patients undergoing hip surgery is debated. It is vitally important to see if the type of anaesthetic administered affects per- and postoperative morbidity and mortality. Seventy women more than 75 yr old suitable for spinal anaesthesia were included in this study. They were randomly divided into two groups: in one (n = 35), general anaesthesia was given with sodium thiopentone, fentanyl, enflurane and vecuronium; in the other (n = 35), spinal anaesthesia was performed with 3 ml 0.5% bupivacaine in isobaric solution. During the anaesthetic period, there were no statistically significant differences between both groups in systolic arterial pressure falls or in increases in the heart rate X systolic arterial pressure product. No serious cardiovascular collapse was encountered. Seven patients (20%) receiving general anaesthesia had mental changes against four only (14.2%) in the spinal group. In patients having general anaesthesia, 22.9% developed bronchopneumonia as opposed to 8.6% in the spinal anaesthesia group (p less than 0.05). The mortality rate at three months was rather similar in the two groups. It was concluded that, in order to reduce the incidence of postoperative central dysfunction and bronchopneumonia, spinal anaesthesia should be preferred in geriatric patients for lower limb surgery.