Recent studies suggest that duodenal ulcers may develop because of increased drive to secrete acid and decreased effectiveness of feedback mechanisms that inhibit acid output. This study was designed to compare gastric acid, gastrin, gastric inhibitory peptide (GIP) and secretin responses to meals (varying in pH) in 12 normal subjects and nine duodenal ulcer patients. Acid secretion was measured by an intragastric titration method which allows actual measurement of acid response to food within the stomach (ten per cent amino acid meal (AAM) adjusted to various pH levels, 7-1.5). Blood samples were collected at each pH level for radioimmunoassay of gastrin, secretin and GIP. Gastric acid and gastrin responses to AAM were found to be significantly greater in duodenal ulcer patients than in normal subjects. In duodenal ulcer patients, acid response to AAM at pH 7 or 5.5 reached 82% of Histalog maximum. Decreasing the pH of the meal resulted in a stepwise reduction in both acid secretion and gastrin in normal subjects and duodenal ulcer patients. At pH 1.5, acid inhibition was complete, but gastrin inhibition was partial. Secretin increased significantly at pH 1.5; there was no difference in secretin release between the groups. Plasma GIP was highest at pH 7 in all individuals. Use of a marker substance showed 80% recovery of AAM at pH 7-4; below pH 4, recovery rose to about 90%. We conclude that gastric acid and gastrin release are pH-dependent in normal and duodenal ulcer subjects. Inhibition of gastric secretion by acidified meals is associated with a pH-dependent suppession of gastrin and GIP levels and elevation of plasma secretin. This study confirms increased acid and gastrin responses in duodenal ulcer patients but shows no evidence of defective feedback inhibition of gastric secretion and gastrin release.