Seventy-two patients with endoscopically proven gastric ulcers were entered into a prospective controlled trial to assess the efficacy of ranitidine and cimetidine in ulcer healing. All patients were biopsied on entry and at subsequent endoscopies. After exclusion of 7 patients during the first month of treatment, the remaining 65 patients, 47 males and 18 females, mean age 48.2 +/-1.5 years, at 1 month had a healing rate of 47% and 52% respectively. The non-healers continued their treatment for a further 4 weeks. This increased the healing rate to 77% and 76% respectively. If the defaulters and poor compliers are withdrawn the healing rate rises to 58% and 57% at 4 weeks and to 91% and 79% at 8 weeks respectively. There was no significant difference between the two groups as regarded initial ulcer size and severity of dyspepsia. Antacid tablet consumption during the study was comparable. The initial size of the ulcers which failed to heal after 4 weeks of treatment tended to be larger than those which healed (P less than 0.05), but smoking did not appear to influence ulcer healing. No obvious side-effects or evidence of dysplasia were found. The study shows that ranitidine is at least as effective as cimetidine in gastric ulcer healing.