Gastric acid and serum gastrin, pancreatic polypeptide, and insulin responses to cephalic vagal stimulation were studied in eight patients with duodenal ulcer using modified sham-feeding for periods varying from four to 30 minutes. In addition, the maximal acid response to sham-feeding was compared with that induced by pentagastrin in 10 healthy subjects and 14 patients with duodenal ulcer. It was found that the gastric acid response to modified sham-feeding reached the maximal value after 15 minutes of sham-feeding and amounted to about 68% of the pentagastrin maximum. The serum pancreatic polypeptide response was also increased after modified sham-feeding and depended on the duration of this procedure, whereas gastrin and insulin responses were not significantly affected by modified sham-feeding. When the peak acid output induced by modified sham-feeding was normalised as percentage of the peak response to pentagastrin, it was similar in healthy subjects and in patients with duodenal ulcer; this indicates that the increased peak acid response to modified sham-feeding observed in patients with duodenal ulcer corresponded with their greater parietal cell mass rather than with increased vagal tone.