In order to localize the site of its insulinogenic stimulus, glucagon was infused individually into the pancreas, liver, and peripheral circulation of normal dogs. Blood levels of insulin, glucose and nonesterified fatty acids were monitored in the femoral artery, portal and hepatic veins. The pancreatic artery infusions provoked a more rapid and marked insulin release initially than the same amount of glucagon given into the portal or femoral vein. The degree of hyperglycemia was comparable in the three groups, although the pancreatic group had more suppression of fatty acids probably related to the enhanced insulin secretion.Intrapancreatic infusion of glucagon in amounts sufficient to release insulin always produced concomitant hyperglycemia. In comparison, small amounts of glucose administered by the same route released insulin without elevating the blood glucose level. Larger quantities of glucose, calculated to produce insulin levels similar to those attained with glucagon, provoked hyperglycemia of lesser magnitude and shorter in duration.The effect of epinephrine on glucagon-stimulated insulin release was markedly different from its effect on a glucose infusion. A quantity of intrapancreatic epinephrine which normally inhibits the glucose stimulus had no effect on delaying the insulin response until the amount of epinephrine was tripled. Simultaneous measurements of insulin and glucose in the hepatic venous blood during the intrapancreatic infusion resulted in response curves similar to those found in the portal vein. Even though the hepatic venous glucose values were almost identical during the infusions at the three different sites, insulin values were two to three times lower with portal or femoral than with pancreatic infusions.We conclude from the data that glucagon does play an independent role in the release of insulin from the pancreas, but that its effectiveness is less than glucose under normal circumstances. The glucagon stimulus is effective despite the presence of ordinary amounts of epinephrine. Since the magnitude of the insulin released from intraportal glucagon is limited and delayed, the importance of its action as a gut factor in response to an oral glucose load is challenged. DIABETES 19:429-37, June, 1970.The insulinogenic property of glucagon has challenged the classical role of glucose in controlling insulin release. Glucagon-like activity is implicated more by its presence in the pancreas and gut than by its glycogenolytic action. Its effectiveness in eliciting a marked insulin response when administered intravenously 1 and again during oral glucose loading 2 casts doubt on the function of the pancreatic source. 3 The following experiments were performed in order to compare the effectiveness of glucagon from both these locations.Glucagon* or glucose administered via the pancreatic artery or portal vein simulated local concentration of these agents respectively in the alimentary circulation. Femoral venous infusion represented the type of periphFrom the