One hundred and one patients were studied in a double-blind controlled trial to assess the role of oral cimetidine in preventing the continuation or recurrence of acute upper gastrointestinal haemorrhage from various sources, chiefly peptic ulcer. The dose of cimetidine was 800 mg on entering the study followed by 400 mg six hourly. The source of bleeding was identified endoscopically in 96% of patients, peptic ulcer comprising 70%. Bleeding continued or recurred in 11 of 51 (21.5%) of patients on cimetidine and in 12 of 50 (24%) of patients on placebo. Analysis of the effect of cimetidine according to age or severity of bleeding showed no significant advantage for the drug.