In an animal experiment the role of bile reflux in development of anastomotic ulcer after partial gastrectomy was investigated. On 45 pigs 2/3 partial gastrectomies and on 5 animals only laparotomies were carried out. The ulcer risk after reconstruction procedures with possible bile reflux was compared with reflux-free techniques. Anastomotic ulcer developed after maximal exposure of the gastric remnant to bile as well as after reflux-free reconstructions: B-II-Roux, 40 cm jejunum loop (n = 5/10); B-II-Roux, 20 cm jejunum loop (n = 3/5); B-II-Roux, 40 cm jejunum loop and truncal vagotomy (n = 1/5); B-I and choledochojejunostomy (n = 2/10); B-II-Roux, 40 cm jejunum loop and cholecystogastrotomy (n = 4/5). After B-I resection with medium duodeno-gastric reflux (n = 0/10) and in the control group (n = 0/5, laparotomy only) no ulcer occurred. Because basal acid output after B-I and B-II-Roux resection was reduced to O mval/h, stimulated gastric acid seems to be important in development of anastomotic ulcer. The experimental results prove the protective role of postresectional duodenogastric reflux. After partial gastrectomy for primary ulcer surgery reconstruction procedures guaranteeing duodenogastric reflux should be preferred to Roux-en-Y gastrojejunostomy.