In a double-blind, randomised controlled trial, we studied 40 patients who received one of four intra-articular injections at the end of arthroscopic surgery. Each group contained ten patients. The patients in Group 1 received normal saline 25 ml; those in Group 2 received bupivacaine 0.25% 25 ml; those in Group 3 received morphine 5 mg in normal saline 25 ml; and those in Group 4 received a combination of bupivacaine 0.5% 12.5 ml and 5 mg of morphine made up to 25 ml with normal saline to produce the same bupivacaine concentration as Group 2. At the time the patient awoke, and 30min, 1hr, 1hr 30min, 2hr, 4hr, 8hr, 12hr, and 24hr postoperatively, pain was assessed using a visual analogue scale. The need for supplementary analgesic agents in the first 24 hours was recorded. All pain scores were significantly lower (P < 0.05) in Groups 2, 3 and 4 compared with the control group with the exception of Group 2 at 24 hours. Pain scores were significantly lower (P < 0.05) for Group 2 compared with Group 3 for the first 90 minutes postoperatively. At 4, 8, 12 and 24 hours postoperatively the pain scores were significantly lower (P < 0.05%) for Group 3 compared with Group 2. Group 4 had the lowest pain scores over the recorded period compared with the other groups. The need for supplemental analgesia was significantly lower (P < 0.05) in the treatment Groups 2, 3 and 4 compared to the control Group 1. There was no significant difference in supplemental analgesic requirements between Groups 2, 3 and 4.(ABSTRACT TRUNCATED AT 250 WORDS)