Sixty patients with endoscopically confirmed gastric ulceration took part in a single-blind trial of cimetidine 1 g daily compared with conventional treatment--namely, carbenoxolone in patients aged under 60 and Caved-(S) in those over 60. Twenty-nine patients received cimetidine: in 12 (41%) ulcer healing was complete after one month, in 26 (90%) healing was complete after two months, and all ulcers were healed after three months of treatment. In the under-60s, ulcers were healed in a greater proportion of patients given cimetidine than in those given carbenoxolone. The difference, however, was significant only at the 5% level, which owing to small numbers was of doubtful clinical validity. In the over-60s cimetidine and Caved-(S) were of similar efficacy. The 54 patients with healed ulcers are being followed up for two years; so far there have been 16 recurrences (30%). Further trials are needed, which should include maintenance treatment aimed at lowering the unacceptably high recurrence rate.