During the period 1977-1984, 53 patients with peptic ulcer resistant to H2-blockers (29 gastric and 24 duodenal ulcers) were submitted to distal gastrectomy with Roux-en-Y gastroenteroanastomosis not associated with vagotomy. The indication was used as an alternative to proximal vagotomy in cases with delayed gastric emptying, high acid output, perforation or bleeding. The study plan consisted of serial clinical and instrumental controls including determination of basal (BAO) or maximal (MAO) acid output. Operative mortality was nil. At various times after the operation, 4 patients died for unrelated reasons and 3 were lost to follow-up. Median follow-up was 84 months with an interval of from 4 to 137 months. Fifty-one patients were followed up for at least one year and 49 for at least two. Five patients (2 gastric and 3 duodenal ulcers) developed peptic recurrences (Visick IV, 9.8%) by the end of the first postoperative year and, in all cases but one, healed stably by the second year after medical (3 cases) or surgical (1 case) therapy. In the remaining patients, Visick was grade III in 6 cases (4 gastric and 2 duodenal ulcers), grade II in 6 and grade I in 34. Before operation, mean values (+/- SD) of BAO and MAO were respectively 5.84 +/- 5.03 and 29.6 +/- 18.6 mEq/h. In the immediate postoperative period there was a considerable reduction in BAO (p less than 0.02) and MAO (p less than 0.03) which continued up to the third postoperative year with a tendency to further progressive reduction in MAO. In spite of a considerable individual variability in dimensions and temporal evolution, the phenomenon occurred qualitatively in all cases. No significant difference was observed in the behaviour of BAO and MAO in gastric ulcers by comparison with duodenal ulcers and in cases with recurrence and Visick III compared to those with a favourable clinical result.