Fifty adult outpatients with endoscopically proven gastric, prepyloric or duodenal ulcers were included in a prospective, randomised double-blind trial of ranitidine (40 mg X 3 daily and 80 mg at bedtime) versus placebo. After 4 weeks the ulcers had healed in 21 of 25 patients receiving ranitidine compared with 7 of 23 in patients receiving placebo (p less than 0.001). The ranitidine treated patients had fewer days of pain (p less than 0.001) and lower consumption of antacids (p less than 0.01) than placebo patients. Patients whose ulcers were not healed after 4 weeks went into an open 4 weeks trial with ranitidine. After the second 4 week period there were still 5 unhealed ulcers, all located in the prepyloric region. No serious side effects or haematological or biochemical abnormalities were observed. It is concluded that ranitidine is a very potent and safe ulcer healing substance. Patients with prepyloric ulcers may need a higher dose or a longer period of treatment.