Sixty nine patients with chronic duodenal or juxtapyloric ulceration were studied in a prospective double blind randomised trial to compare the efficacy of antacid and placebo at high (30 ml seven times daily) and low (10 ml as required) doses. After four weeks ulcers had healed in 12 out of 18 patients (67%) receiving "low dose" antacid compared with in six out of 17 patients (35%) receiving low dose placebo; ulcers had also healed in six out of 19 patients (32%) receiving "high dose" antacid compared with in two out of 15 patients (13%) receiving high dose placebo. Overall, the effect of antacid was superior to that of placebo in healing ulcers (p less than 0.05) and the effect of low dose treatment was superior to that of high dose treatment (p less than 0.01). There were no significant differences between antacid and placebo at eight weeks. Antacid was better than placebo in relieving pain, but the difference was not significant. Poor compliance and high incidence of diarrhoea made high dose antacid an impractical treatment. Low dose antacid was associated with a significantly better rate of healing than high dose antacid and was far better tolerated. This low dosage of antacid should be considered to be an active treatment in trials of ulcer healing.