We sought to determine in overnight-fasted dogs whether basal or 2-deoxy-D-glucose (2-DG)-stimulated levels of pancreatic polypeptide (PP) could be reliably used as an index of cholinergic activity at the pancreas and thereby determine the effect of pentobarbital anesthesia on this cholinergic outflow. At low basal PP levels, either atropine or pentobarbital had a small effect on PP levels; at higher basal levels, both atropine and pentobarbital had a larger effect. Thus both drugs decreased PP in proportion to its initial basal level, suggesting that basal PP levels have a variable cholinergic component. Atropine abolished the PP response to intravenous 2-DG, confirming in our animal model that the PP response to neuroglucopenia is entirely cholinergically mediated. Pentobarbital also abolished the PP response to 2-DG, suggesting that anesthesia either suppresses cholinergic outflow to the pancreas or the response of the pancreatic F-cell to it. To test the latter hypothesis, the acetylcholine analogue bethanechol was administered before and during pentobarbital anesthesia. The PP response to direct cholinergic stimulation was not abolished by pentobarbital, although it was reduced modestly. We conclude that only part of the basal level of PP is under cholinergic control; all of the PP response to 2-DG is cholinergically mediated; pentobarbital anesthesia abolishes the cholinergic input to the pancreas; and if the endogenous cholinergic input influences certain pancreatic endocrine and exocrine responses, then its contribution would be seriously underestimated when studied in pentobarbital-anesthetized animals.