The purpose of this study was to determine the role of gastrin and cholecystokinin in the cephalic phase of gastrin release and gastric and pancreatic secretion in conscious dogs. Sham feeding in intact dogs increased gastric acid output to about 65% of histamine maximum and pancreatic protein to 23% of caerulein maximum. Significant increases in plasma gastrin and pancreatic polypeptide but not cholecystokinin occurred. Similar effects were obtained using insulin hypoglycemia or 2-deoxy-D-glucose glucocytopemia. Atropine eliminated gastric acid response to sham feeding, insulin, or 2-deoxy-D-glucose, significantly reduced the pancreatic protein response by about 60%, and abolished plasma pancreatic polypeptide but not plasma gastrin. Blocking of cholecystokinin receptors by L-364,715 did not affect gastric or pancreatic secretory responses to sham feeding, insulin, or 2-deoxy-D-glucose and failed to influence the accompanying increments in plasma gastrin and pancreatic polypeptide. In antral-mucosectomized dogs, sham feeding-induced acid output reached only 17% of histamine maximum but the increase in pancreatic protein output was similar to that in intact dogs. In these animals, background stimulation with G17I (62 pmol/kg per h) potentiated the gastric acid response to sham feeding but had little effect on pancreatic protein output. This study provides evidence that unlike gastric acid, the pancreatic protein response to physiological or pharmacological cephalic stimulation does not depend on vagally released gastrin but probably on direct vagal stimulation of the pancreas.