By use of a recently described method, which estimates the rate of gastric acid secretion by measuring the rate of sodium bicarbonate infusion needed to keep intragastric pH constant, gastric acid secretion rates and changes in serum gastrin were measured in five normal subjects while gastric pH was kept at 5.5, 4.0, 3.0, or 2.5. Preliminary experiments revealed that the method did not accurately measure acid secretion at a pH lower than 2.5. Stimulation of acid secretion was produced by gastric instillation of a solution of amino acids and cornstarch. The secretion rate with the amino acid meal was highest at pH 5.5 and was 60% of that produced by a steak meal at the same pH. As the pH of the amino acid meal was decreased, there was a stepwise reduction in acid secretion so that at pH 2.5 the rate was only half as great as at pH 5.5. The amino acid meal produced increases in serum gastrin that were also less marked than those produced by a steak meal. With amino acid stimulation, serum gastrin responses were similar at pH 5.5, 4.0, and 3.0, but no increase in gastrin could be measured when the meal was maintained at pH 2.5. A group of six patients with duodenal ulcers was compared with seven normal subjects at pH 5.5 and 2.5. Ulcer patients released more gastrin and secreted more acid at each time period at both pH values. More important, the degree of inhibition at pH 2.5 was significantly less in ulcer patients. For example, during the 2nd h after stimulation acid secretion was inhibited by only 30% in ulcer patients compared with 70% in normal subjects. These findings suggest a defect in autoregulation of gastrin release and gastric acid secretion at low pH in ulcer patients which may play a role in pathogenesis of this disease.