The prevalence of chronic peptic ulcer among 152 out-patients with rheumatoid arthritis attending a district general hospital was found to be 16.4% (25 patients), as against 6.6% (10 patients) among an equal number of matched control out-patients with other diagnoses. There was no evidence that any form of anti-inflammatory treatment increased ulcer risk, but smoking was significantly associated with a higher prevalence of ulcer. This relationship was shown to be quantitative and to depend upon an increased frequency of duodenal ulcer in women who smoked. It is possible that greater numbers would have shown the effect of smoking to apply to both sexes, and that for all patients with rheumatoid arthritis abstention from smoking could diminish the high risk of peptic ulceration associated with this disease.