We studied the effect of several doses of atropine on the serum gastrin and pancreatic polypeptide responses to vagal stimulation in healthy human subjects. Vagal stimulation was induced by sham feeding. To eliminate the effect of gastric acidity on gastrin release, gastric pH was held constant (pH 5) and acid secretion was measured by intragastric titration. Although a small dose of atropine (2.3 mug/kg) significantly inhibited the acid secretory response and completely abolished the pancreatic polypeptide response to sham feeding, this dose of atropine significantly enhanced the gastrin response. Higher atropine doses (7.0 and 21.0 mug/kg) had effects on gastrin and pancreatic polypeptide release which were similar to the 2.3-mug/kg dose. Atropine (0.78 and 2.3 mug/kg) without sham feeding significantly inhibited basal acid secretion and also led to significant increases in serum gastrin above basal levels. The gastrin response to sham feeding with 2.3 mug/kg atropine was significantly greater than the sum of the gastrin responses to sham feeding alone and to 2.3 mug/kg atropine alone, indicating potentiation of vagal gastrin release by atropine. We conclude: (a) Unlike vagally mediated gastric acid secretion and pancreatic polypeptide release which can be blocked by atropine, vagal gastrin release is potentiated by atropine. This observation suggests the existence of a vagal-cholinergic pathway which normally (i.e., in the absence of atropine) inhibits gastrin release. (b) Because atropine (without sham feeding) increased basal gastrin levels, it is likely that the cholinergic pathway which inhibits gastrin release is active even when the vagus nerve is not stimulated by sham feeding.