We performed a prospective, randomized, double-blind study of continuous epidural analgesia for 72 hours after major abdominal procedures. Patients were randomly assigned to one of five treatment groups: epidural morphine, epidural bupivacaine, a combination of morphine and bupivacaine, epidural saline solution, and no epidural catheter. All patients received supplemental morphine sulfate or meperidine hydrochloride, intramuscularly or intravenously, as needed. Epidural infusion was begun at 2 to 4 ml/hr, depending on age and height, with two increments of 1 ml/hr allowed if pain relief was insufficient. All pain management decisions were made by nurses, who also monitored epidural function. Performance was measured four ways: pain as measured at regular intervals in the 72-hour period with a visual analog, pain as measured after 72 hours with the McGill Pain Questionnaire, amount of supplemental narcotics needed, and recovery of respiratory function and ambulation as percent of preoperative levels. The group that received the combination of morphine and bupivacaine did best on all measures; in most instances the difference between the results seen with the combination regimen and those seen with saline solution or no catheter were significant at the 0.05 level. With the exception of pruritus, complications were evenly distributed among all treatment groups, including noncatheterized controls. We conclude that epidural analgesia with the combination of morphine and bupivacaine is safe, is easily managed, and gives pain relief superior to that provided by traditional, systemic administration of narcotics.