Evidence has accumulated that opioids can produce potent antinociceptive effects by interacting with opioid receptors in peripheral tissues. This study sought to compare the effects of morphine with those of bupivacaine administered intraarticularly upon pain following arthroscopic knee surgery. In a double-blind, randomized manner, 33 patients received either morphine (1 mg in 20 ml NaCl; n = 11), bupivacaine (20 ml, 0.25%; n = 11), or a combination of the two (n = 11) intraarticularly at the completion of surgery. After 1, 2, 3, and 4 h and at the end of the 1st and 2nd postoperative days, pain was assessed by a visual analogue scale, and supplemental analgesic requirements were recorded. Pain scores were significantly greater in the morphine group than in the other two groups at 1 h. There were no significant differences at 2 and 3 h. From 4 h until the end of the study period, pain scores were significantly greater in the bupivacaine group than in the other two groups. Analgesic requirements were significantly greater in the morphine group than in the other groups at 1 h but were significantly greater in the bupivacaine group than in the other groups throughout the remainder of the study period. We conclude that intraarticular morphine produces an analgesic effect of delayed onset but of remarkably long duration. The combination of these two drugs results in satisfactory analgesia throughout the entire observation period.