During a period of 12 months, 88 patients with severe haemorrhage from gastric or duodenal ulcers or from erosive gastritis completed a double-blind trial of either cimetidine or placebo. Only patients needing immediate blood transfusion were admitted to the trial. It was found that in patients with severe bleeding from gastric or duodenal ulcers neither the severity of bleeding nor the incidence of emergency surgery was reduced by cimetidine. Furthermore, the treatment did not improve the mortality rate. It is concluded that patients with severe bleeding from gastric or duodenal ulcers will not benefit from immediate treatment with intravenous cimetidine.