Cyclosporine administration has been associated with the development of lymphomas in human transplant patients as well as animals. Its effect on the genesis of common epithelial carcinomas is unknown. To investigate this N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) was administered in drinking water to Wistar rats. Seventy-five young healthy male animals were divided into six groups and received cyclosporine alone, cyclosporine followed by MNNG, MNNG alone, cyclosporine during MNNG administration, MNNG followed by cyclosporine, and no treatment. Cyclosporine seemed to have minimal overall health effects and no cancers were encountered in the group receiving this agent alone. Animals in all carcinogen-treated groups developed gastric and upper intestinal carcinomas by Week 39. No statistically significant differences among carcinogen-treated groups were evident with respect to tumor incidence, histology, or distribution. There appeared to be trends (not statistically significant) toward a greater incidence of small bowel carcinomas in animals receiving cyclosporine plus MNNG as compared to those receiving MNNG alone; greater multiplicity of small intestinal carcinomas in animals receiving cyclosporine after MNNG as compared to all other groups; and greater incidence of small bowel tumors greater than 1 cm3 in animals receiving cyclosporine after MNNG as compared to all other groups. The median total tumor volume in the animals receiving cyclosporine following carcinogen was significantly greater than in any other group. This study does not support a policy of aggressive surveillance for gastrointestinal carcinoma in the human population receiving cyclosporine.