Proximal gastric vagotomy (PGV) is the operation of choice in duodenal ulcer (DU), in view of its minimal morbidity-mortality. We discuss the number of recurrences (R) involved in this technique; most are the result of insufficient denervation, whereas a few ‘new gastric ulcers’ are due to emptying problems. Only 1 year after our first PGV (1970), we attempted the incorporation of postoperative tests (pH of gastric juice samples) peroperatively. Later it became possible to situate the electrode on the gastric mucosa after passing it down the esophagus. Finally, the electrode was pressed against the mucosa, and a minimum pH of 6.4 was found to define complete denervation. By this criterion, only two Rs (0.4%) were observed in 519 PGVs. An area resisting pH rise was established: in one fifth of cases the preantral region only rises to pH 7 after gastroepiploic nerve sectioning. We carried out a prospective randomized study of PGV + sectioning of the gastroepiploic nerve (Sge), PGV + pH test and PGV alone. The study involved 293 cases (1980), including 269 followed up for an average of 57 months; Rs were observed in 1.4, 1.7 and 10.1% of cases with the first, second and third method, respectively. In our opinion, PGV + Sge with cardiohiatofundolysis may reduce R to under 4 %.