The purpose of this study was to find whether gastric resection enhances the incidence of carcinoma in the remaining part of the stomach. 66 male Wistar rats were subjected to stomach resection according to the Billroth I or the Billroth II method. These rats, as well as control animals with intact stomachs, were fed the carcinogen N-Methyl-N'-nitro-N-nitrosoguanidine (NG). -- 25 of 66 animals developed carcinomas in the gastric remnant. Precancerous lesions were seen in 18 rats. The tumours were characterized histologically as adenocarcinomas. They were almost exclusively localized in the region of the gastroenteral anastomosis. The process of tumour formation in the resected stomach was completed within 17-31 weeks on continuous administration of NG in a concentration of 120 mg/l in the drinking water. In contrast to these findings, the development of cancer in the intact stomach required on average 41 weeks under the same conditions of NG administration. However, with regard to the incidence of malignant changes, no significant difference was observed between animals undergoing the Billroth I method and those undergoing the Billroth II method.--The results suggest that the resected stomach of the rat is more susceptible to induction of cancer than the intac one. Exposure of the resected stomach to an oral carcinogen induces carcinogenesis predominantly in the anastomotic region.