In a prospective clinical trial, vagotomy for duodenal ulcer (DU) and prepyloric ulcer (PPU) was performed in 748 patients, 353 of whom were randomly allocated to selective gastric vagotomy and drainage (SGV + D), 54 to SGV + antrectomy (A), 273 to parietal cell vagotomy (PCV), and 68 to PCV + D. By 3 months postoperatively, basal acid secretion (BAO) had not stabilized. During the following year patients with SGV + A showed a decrease, while those with the other operations showed a rise in BAO, significant for SGV + D. One year after operation the level of BAO was the same after the 3 operations that did not remove the antrum. Peak acid output after pentagastrin stimulation (PAOPg) continued to decrease from 3 months to 1 year after SGV + A, while the other operations were followed by an increase, statistically significant for PCV. After 1 year the postoperative reduction in PAOPg was 90% for SGV + A, 45% for PCV, and approximately 60% for SGV + D and PCV + D. Overall clinical grading showed more failures following PCV than after SGV. Since failures after PCV were mainly ulcer recurrences, the final grading (after treatment of the failures) showed an equal number of failures for the 2 operations. Calculation of the probability of ulcer recurrence suggested a 6% rate after SGV + D and an 11% rate after PCV. However, when calculations took into account the location of the primary ulcer, the recurrence rate was the same after SGV + D for DU and PPU, while PCV showed a similar rate when used for DU but an incidence of 22% when used for PPU. The risk of recurrence was found to be constant from month to month during the first 2 1/2 years, after which no new recurrent ulcers were observed. It is suggested that for DU, PCV is preferable to SGV + D because the recurrence rate is the same but the incidence of sequelae is lower. When PCV is used for PPU, a higher ulcer recurrence rate may be expected.