One hundred and nineteen patients with endoscopically confirmed prepyloric (n = 59) or duodenal (n = 60) ulcer were stratified for ulcer location before entering a randomized double-blind trial comparing ranitidine (150 mg twice daily) and a potent liquid antacid (Novaluzid®: 10 ml seven times daily, with a neutralizing capacity of 600mmol H). Fifty-four patients with prepyloric (26 receiving ranitidine) and 53 patients with duodenal ulcer (28 receiving ranitidine) completed the trial in accordance with the protocol. The 4 and 6 weeks' healing rates for prepyloric ulcers were 54% 68% and 61% versus 69% 79% and 74% for the ranitidine, the antacid, and whole groups, respectively. For duodenal ulcers these figures were 89% 84% and 87% versus 100% 96% and 98% for the ranitidine, antacid, and whole groups. respectively. Differences in healing rates between treatments were statistically insignificant within strata for ulcer type, but healing rates for prepyloric ulcers were significantly lower than for duodenal ulcers (p < 0.002). A significant early pain relief was found in all groups, and side effects, including diarrhoea, were rare. In conclusion. these two ulcer treatment modalities appear to be equally effective in the short term. In addition, the data emphasize the need for proper stratification of prepyloric and duodenal ulcers in clinical trials of ulcer healing.