A series of 103 African and Indian patients received extradural lumbar blockade for a variety of elective and emergency operations. Bupivacaine 0.25 per cent was used in 51 patients and 0.5 per cent in the remaining 52. Both solutions contained adrenaline 1:200,000. Results of latency of onset, duration of action, and motor blockade are presented. Operative and postoperative complications are also noted and discussed. Clinically, 0.5 per cent bupivacaine proved more satisfactory than 0.25 per cent in the operating theatre. It is concluded that bupivacaine, in this series, proved effective and apparently safe, the duration of action being prolonged.