Seventy-two patients with duodenal (54) or prepyloric (18) ulcers have taken part in a 12-week double-blind trial. Twenty-four patients received cimetidine, 1 g/day; 24 patients received 10 ml of an antacid suspension (buffering 85 mmol acid) 1 and 3 h after every meal and at bedtime and 0.6 mg L-hyoscyamine in sustained-release tablets twice a day; and 24 patients received placebo. The healing rate after 3 weeks' treatment was 67% (p less than 0.005 compared with placebo) with cimetidine, 50% (p less than 0.01) with antacid/anticholinergic, and 13% with placebo. After 6 weeks' treatment 83% were healed with cimetidine (p less than 0.005 compared with placebo), 96% with antacid/anticholinergic (p less than 0.005), and 33% with placebo. A further 6 weeks' treatment gave healing rates of 96% for cimetidine, of 100% for antacid/anticholinergic, and of 50% for placebo. Compared with placebo, cimetidine but not antacid/anticholinergic caused a faster relief of night-time ulcer pain (p less than 0.05). There was a significant correlation between healed ulcers and complete relief of ulcer symptoms (p less than 0.05). In the placebo group the ulcers of nonsmokers healed to a higher extent than those of smokers (p less than 0.05). During 1 year of follow-up there was no difference between the two actively treated groups in number or severity of symptomatic relapses. Time to relapse was, however, significantly shorter after treatment with cimetidine than after antacid/anticholinergic (p less than 0.05). Recurrences occurred more often (p less than 0.05) after slow healing (6-12 weeks) than after fast healing (3 weeks).