Sixty consecutive patients with endoscopically proven duodenal ulcers were given colloidal bismuth subcitrate tablets either as 120 mg q.d.s. or 240 mg b.d., in a randomized single-blind study. The efficacy of each regimen was determined by endoscopic examination and antral biopsy at 4 weeks; if the ulcer remained unhealed, treatment was continued and endoscopy repeated at 8 weeks. The ulcer-healing efficacy of the two regimens was identical; however, in the four times daily group only 27% remained Campylobacter pylori positive after 8 weeks of treatment compared with 58% of the twice-daily group. Similarly, only 21% of twice daily patients were free of histological gastritis compared with 42% of the four times daily patients.