The relative risk (standardized incidence ratio [SIR] for colorectal cancer after partial gastrectomy for ulcer disease was examined in a population-based cohort comprising 6459 patients operated during 1950 to 1958. Follow-up through 1983 revealed 131 cases of colorectal cancer versus 150.3 expected cases (SIR = 0.87; 95% confidence limits 0.73, 1.03). The relative risk was decreased during the first 19 years of follow-up (SIR = 0.75; 0.58, 0.96) and close to equal thereafter (SIR = 1.02; 0.79, 1.29). Sex, age at operation, type of operation, and diagnosis at operation (stomach or duodenal ulcer) did not affect the relative risk significantly. Confounding by socioeconomic status is one conceivable explanation for the decreased relative risk during the first years of follow-up, whereas the increasing risk with longer duration of follow-up might be a consequence of the surgical procedure itself.