In a prospective, randomized trial, 161 patients with duodenal, pyloric, or prepyloric ulcer underwent selective proximal vagotomy. Randomization was then performed to determine if the operation was finished (52 patients), if a pyloroplasty should be added (56 patients), or in addition, if the nerves of Latarjet should be divided (53 patients). Prepyloric and secondary gastric ulcers were excised for microscopy; all were benign. Sex, age, site of ulcer, and duration and incidence of complications of the ulcer disease were similar for the three groups. There was one operative death. The postoperative complications did not differ for the three groups. Four patients were lost to follow-up. The average follow-up for the 156 patients was 3 years (range 1 to 8 years). Recurrent ulcer was detected up to 5 years after surgery in 4 of 53 patients who had selective vagotomy with pyloroplasty, in 4 of 53 who had selective proximal vagotomy with pyloroplasty, and in 5 of 50 who had selective proximal vagotomy. Diarrhea was rare and mild or absent. Dumping was twice as common after selective vagotomy or selective proximal vagotomy with pyloroplasty than after selective proximal vagotomy only, but dumping resistant to treatment was recorded in only two or three patients in each group. The overall results (modified Visick scale) were unsatisfactory in 7 patients after selective vagotomy with pyloroplasty, in 4 after selective proximal vagotomy with pyloroplasty, and in 10 after selective proximal vagotomy, mainly because of epigastric pain with or without recurrent ulcer. We conclude that pyloroplasty may cause mild dumping without nuisance to the patient. The rates of recurrent ulcer in long-term follow-up trials are essential for final evaluation of the operations.