A bstract. Somatostatin (ST)-induced glucagon suppression results in hypoglycemia during rest and exercise. To further delineate the role of glucagon and interactions between glucagon and the catecholamines during exercise, we compensated for the counterregulatory responses to hypoglycemia with glucose replacement. Five dogs were run (100 m/min, 120) during exercise alone, exercise plus ST infusion (0.5 gg/kg-min), or exercise plus ST plus glucose replacement (3.5 mg/kgmin) to maintain euglycemia. During exercise alone there was a maximum increase in immunoreactive glucagon (IRG), epinephrine (E), norepinephrine (NE), FFA, and lactate (L) of 306±147 pg/ml, 360±80 pg/ml, 443±140 pg/ml, 541±173 geq/liter, and 6.3±0.7 mg/dl, respectively. Immunoreactive insulin (IRI) decreased by 10.2±4 ttU/ml and cortisol (C) increased only slightly (2.1±0.3 ug/dl). The rates of glucose production (Ra) and glucose uptake (Rd) rose markedly by 6.6±2.2 mg/ kg-min and 6.2±1.5 mg/kg-min. In contrast, when ST was given during exercise, IRG fell transiently by 130±20 pg/ml, Ra rose by only 3.6±0.5 mg/kg-min, and plasma glucose decreased by 29±6 mg/dl. The decrease in IRI was no different than with exercise alone (10.2±2.0 ,U/ ml). As plasma glucose fell, C, FFA, and L rose excessively to peaks of 5.4±1.3 ig/dl, 1,166±182 geq/liter and 15.5±7.0 mg/dl. The peak increment in E (765±287 pg/ml) coincided with the nadir in plasma glucose and was four times greater than during normoglycemic exercise. Hypoglycemia did not affect the rise in NE. The increase in Rd was attenuated and reached a peak of only 3.7±0.8 mg/kg-min. During glucose replacement, IRG decreased by 109±30 pg/ml and the IRI response did not differ from the response to normal exercise. Ra rose minimally by 1.5±0.3 mg/kg-min. The changes in E, C, Rd, and L were restored to normal, whereas the FFA response remained excessive. In all protocols increments in Ra were directly correlated to the IRG/IRI molar ratio while no correlation could be demonstrated between epinephrine or norepinephrine and Ra. In conclusion, (a) glucagon controlled -70% of the increase of Ra during exercise. This became evident when counterregulatory responses to hypoglycemia (E and C) were obviated by glucose replacement; (b) increments in Ra were strongly correlated to the IRG/IRI molar ratio but not to the plasma catecholamine concentration; (c) the main role of E in hypoglycemia was to limit glucose uptake by the muscle; (d) with glucagon suppression, glucose production was deficient but a further decline of glucose was prevented through the peripheral effects of E; (e) the hypoglycemic stimulus for E secretion was facilitated by exercise; and (f) we hypothesize that an important role of glucagon during exercise could be to spare muscle glycogen by stimulating glucose production by the liver.