AbstractIn order to know whether the exact determination of the corpus–antrum border is of importance in highly selective vagotomy (HSV) the operation was performed on two matched groups of patients with duodenal ulcers. In group 1 (35 patients) HSV was performed using anatomical landmarks to define the distal extent of dissection. In group 2 (35 patients) HSV was performed using the same technique and then a gastrotomy was made and any remaining acid‐secreting mucosa was identified using pH‐metry under pentagastrin stimulation. In 22 cases secretion of acid was observed distal to the anatomically defined corpus‐antrum border. In these patients afurther distal division of vagal branches was done until no acid secretion was observed. Two months and 1 year after the operation, basal acid output (BAO), pentagastrin‐stimulated acid output (PAOpg), insulin‐stimulated acid output (PAO1) and basal serum gastrin determinations were done and compared with the corresponding preoperative data. Our results show that despite the more extensive division of the antral vagal branches in group 2, BAO and PAOpgwere similar in both groups postoperatively. Postoperative PAOIwas also similar. Differences were observed, however, in postoperative serum gastrin levels–group 1 42.5 ± 6.0 pmol/l and group 281.7 ± 7.4 pmol/l 1 year after the operation. In no cases did the more extensive dissection of antral vagal branches lead to clinical or radiological gastric stasis. Thus, if the branches of the nerves of Latarjet are used as landmarks in performing HSV, in many cases vagally innervated, acid‐secreting mucosa will be left in the stomach, but apparently this is quantitatively of minor importance. On the other hand, a more extensive distal vagal dissection leads to disturbingly high postoperative ser