Eighty-seven morbidly obese patients were prospectively randomized to two operations: gastric bypass was performed on 42 and gastric partition on 45. Gastric bypass proved to be more effective; gastric bypass patients lost 15% more of their original weight at 12 months and 21% more at 18 months. There were no failures in the gastric bypass group; 28 of the 45 operations failed in the gastric partition group. An additional 60 patients underwent gastric bypass since the completion of the study. In the total series of 147 patients who underwent gastric bypass or gastric partition, there was no mortality, and the surgical complication rate was 12%. Because the gastric pouches and the anastomoses were similar in the two operations, the superiority of the gastric bypass may well be due to a heretofore unexplained effect of distal gastric and duodenal exclusion.