There is some evidence of an antagonism between peptic ulcer and diabetes mellitus. This antagonism is possibly related to anomalous insulin metabolism in duodenal ulcer patients. To evaluate this issue, 471 Rochester, Minnesota residents who had surgery for peptic ulcer were examined and then followed for up to 34 years. Their experience of diabetes mellitus was compared with that of Rochester residents generally, for whom data on diabetes incidence and prevalence were available. At surgery, the prevalence of diabetes was somewhat increased for gastric and slightly diminished for duodenal ulcer patients. Of the 445 patients who were free of diabetes at surgery, 27 developed it. By 20 years, the cumulative incidence of diabetes (9.1%) was less than expected (10.3%), but over the 20 years the difference was not significant for either gastric or duodenal ulcer patients. The risk of diabetes was not influenced by age, gender, or smoking status but analysis for relative weight at surgery indicated that obese patients had a 2.6-fold increase in risk of subsequent diabetes, whereas those with less than 1.2 relative weight had reduced risk. Relative weight was the only significant predictor of the time to subsequent diabetes in a proportional hazards regression analysis. The proposition of an antagonism between peptic ulcer and diabetes is not supported by the data.