Thirty-six patients in whom an asymptomatic duodenal ulcer had been detected endoscopically were followed up clinically and endoscopically during the following months. Ten of the patients were receiving either no treatment or placebo, and 26 patients were receiving maintenance treatment with either 150 mg ranitidine or 400 mg cimetidine at night. Treatment remained unchanged during the follow-up period. The cumulative annual rate of spontaneous healing was approximately one quarter, whether or not patients were receiving active maintenance therapy. However, the likelihood of developing symptoms during the year after detection of the asymptomatic ulcer was significantly greater in those patients receiving no treatment or placebo (77%) than in those receiving active maintenance therapy (27%). One patient, receiving no treatment, bled from his ulcer during the follow-up period. We conclude that routine endoscopic reexamination, to detect asymptomatic ulcer recurrences in patients receiving maintenance treatment for duodenal ulceration, is probably unnecessary, since the recurrences rarely cause clinical problems, despite prolonged failure to heal.