We performed a randomised double-blind controlled study in patients undergoing elective knee arthroscopy to assess the effect of intra-articular morphine on postoperative pain relief. Patients in the study group (n = 10) received 5 mg of morphine in a 25 ml dilution intra-articularly while those in the control group (n = 10) received 25 ml of saline. Postoperative pain was assessed at intervals by a visual analogue scale and the requirement for supplementary analgesia was recorded. Those in the study group had significantly lower pain scores and required less systemic analgesics than those in the control group. Plasma profiles for morphine and its metabolites were assayed and showed that they were too low to produce effective analgesia. Evidence suggests that analgesia was mediated by local action within the joint.