The effects of truncal vagotomy and parietal cell vagotomy on gastric acid secretion and plasma gastrin and pancreatic polypeptide release were studied in 28 duodenal ulcer patients under basal conditions and after modified sham feeding and infusion of pentagastrin (2 micrograms/kg/h). Before vagotomy gastric acid output in response to modified sham feeding was significantly higher than basal acid secretion in all subjects tested and reached about 45% of the pentagastrin maximum. No difference in the increase in acid response, or in the pancreatic polypeptide response to modified sham feeding was found between patients with high and low basal secretion. Plasma gastrin concentration was unaltered by modified sham feeding before and after truncal vagotomy or parietal cell vagotomy, although after vagotomy it tended to reach higher values than before this procedure. After truncal vagotomy, basal pancreatic polypeptide concentration was decreased and modified sham feeding-induced pancreatic polypeptide increment was completely eliminated. Four weeks after parietal cell vagotomy, the modified sham feeding-induced increment in plasma pancreatic polypeptide was significantly decreased and observed only in seven of 12 patients. Four to five years after parietal cell vagotomy all subjects responded to modified sham feeding with pancreatic polypeptide increment similar to that before vagotomy and in three of 12 patients acid response to modified sham feeding was seen. This study indicates that truncal vagotomy eliminates gastric acid and plasma pancreatic polypeptide responses to vagal excitation while parietal cell vagotomy abolishes gastric acid response and reduces temporarily the pancreatic polypeptide response to modified sham feeding (possibly because of transient impairment of the vagal innervation of the pancreas). (2) A high ratio of basal to maximal acid output in non-operated duodenal ulcer patients is not associated with a low acid response to modified sham feeding, nor with a high pancreatic polypeptide concentration, and (3) Restitution of the pancreatic polypeptide response to modified sham feeding five years after parietal cell vagotomy does not indication ineffective denervation of the parietal cells.