I n order to elucidate the role of insulin and glucagon during strenuous exercise (100 m/min, slope 10-12 degrees), we have determined the rates of production (Ra), utilization (Rd), and metabolic clearance (M) of glucose in normal dogs before pancreatectomy and 2 wk after total pancreatectomy (a) when they were being maintained on constant intraportal basal insulin infusion, (245 muU/kg-min) and (b) when insulin supply had been withheld before and during exercise. Such an intense exercise induced in normal dogs a prompt decrease in mean immunoreactive serum insulin (IRI) from 20 +/- 3 to 11 +/- 2 muU/ml. In depancreatized insulin-infused dogs serum IRI during rest and exercise was between 14 +/- 1 and 12 +/- 2 muU/ml. In the third group, after cessation of insulin infusion, IRI decreased by 76% (from 17 +/- 5 to 4 +/- 1) and did not decrease futher during exercise. During exercise, serum immunoreactive glucagon (IRG) increased threefold in normal dogs. In depancreatized dogs serum IRG was the same as in normal resting dogs (indicating a nonpancreatic source of the hormone) but it did not increase during exercise. In normal dogs exercise induced proportional increases in Ra, Rd, and M (threefold) and normoglycemia was maintained. Changes in glucose turnover in depancreatized insulin-infused dogs were similar to those seen in normal dogs suggesting that a decrease in insulin secretion and a rise in IRG are not essential to prevent hypoglycemia in diabetic dogs. With the cessation of insulin infusion in resting depancreatized dogs, Ra increased, M decreased, and hyperglycemia ensued. During exercise, Ra continued to rise, but M did not increase significantly. Conclusions: (a) Regulation of glucose production by liver during exercise is multifactorial. A decrease in IRI and an increase in IRG are not the only factors which can promote delivery of glucose to the peripheral tissues. The insulin glucagon molar ratio was found not to be an essential metabolic functional unit in regulating glucose metabolism during exercise. (b) It is hypothesized that increases in blood flow and capillary surface area can lead to an increase in the amount of insulin delivered to the muscle even when serum levels of IRI are reduced during exercies. It is suggested that small, but adequate amounts of insulin (as found in normal and depancreatized insulin-infused dogs) are essential in regulating glucose uptake in the working muscle. (c) Since totally depancreatized dogs had normal serum levels of IRG (originating presumably from the gastrointestinal tract), the question of essentiality of basal glucagon activity in glucose homeostasis during exercise could not be resolved by these experiments. It appears, however, that regulation of secretion of nonpancreatic glucagon differs from that of pancreatic glucagon.