Previous studies investigating the peripheral action of locally instilled morphine after arthroscopic knee surgery found evidence for an analgesic effect. Follow-up studies have lead to conflicting results. We used patient-controlled analgesia (PCA) to test the analgesic potency of intraarticular morphine. METHODS. Patients undergoing arthroscopic knee surgery under general anaesthesia received, after written informed consent and in double-blind and randomised manner, 1 mg morphine diluted in 10 ml saline either intraarticularly or intravenously at the end of the surgical procedure. A control injection of 10 ml saline was given at the other site. The pain intensity on a visual analogue scale (VAS) and the cumulative morphine consumption were recorded at 1, 2, 3, 4, 6, 8 and 24 h after the end of general anaesthesia. METHODS Wilcoxon rank sum test with P < 0.05. RESULTS. A total of 59 patients were included in the study; 29 received morphine intraarticularly (verum group), 30 intravenously (control group). There was no difference in gender, age, duration of arthroscopy or anaesthesia. There were more than 60% diagnostic arthroscopies in both groups; other types of surgery were comparable, with the exception of cruciate band repair procedures only in the control group. We found no difference in morphine consumption or pain intensity between the two groups throughout the study period. Median overall consumption of morphine after 24 h was 14 mg in the verum group and 15 mg in the control group, with wide interindividual variation. Pain intensities were remarkably low. The peak pain intensity of both groups was found at 1 h postoperatively, with median 16/100 on the VAS in both groups. Blinding was robust. CONCLUSION. We found no reduction in postoperative morphine supplementation after 1 mg morphine intraarticularly compared to 1 mg intravenously given at the end of knee arthroscopies. There were also no differences in pain intensities on a VAS. We conclude that titration of postoperative pain with a morphine-filled PCA pump was unable to show a difference in analgesic potency between intraarticular and intravenous morphine.