A decline in plasma insulin and an increase in glucagon are known to occur during intense and/or prolonged exercise. However, it is not established whether changes in insulin and glucagon secretion are involved in the precise matching of hepatic glucose production to the enhanced glucose uptake by muscle during brief, low intensity exercise. We studied the effects of 30-min cycle exercise at 40% of maximal aerobic capacity in healthy subjects and C-peptide-deficient subjects with type 1 diabetes (IDDM) using [3-3H]glucose to estimate glucose turnover. Diabetic subjects were studied during continuous iv insulin infusion, which normalized glucose kinetics before experimental perturbations. In control (saline-infused) experiments, endogenous glucose appearance (Ra) increased by 80-90% above baseline to match the increase in glucose disappearance in both normal and IDDM subjects, even though the latter exercised at fixed levels of plasma free insulin, averaging 203 +/- 19 pmol/L. In other experiments, somatostatin was infused, and glucagon (1.0 ng/kg.min) and insulin (at two different rates) were maintained at constant levels. Infusion of insulin in normal subjects at doses sufficient to maintain constant peripheral plasma insulin was associated with no apparent effect on glucose turnover (plasma insulin, 80 +/- 21 pmol/L, compared to 52 +/- 5 pmol/L during saline; P = NS). However, insulin infusion at doses that normalized the portal insulin concentration (approximately 208 pmol/L) together with glucagon replacement inhibited the rise in glucose production in both normal and IDDM subjects. There were similar 45-55% reductions (P less than 0.03) of the increase in Ra seen with exercise in control experiments. When peripheral plasma free insulin (and presumably portal levels as well) were increased by about 20% in this experimental setting in IDDM (278 +/- 43 pmol/L), the suppression of Ra was even more profound, and Ra failed to increase at all with exercise. We conclude that the hormonal regulation of Ra in brief duration exercise in man does not necessitate the decrements in portal venous insulin observed under more intense exercise conditions as long as an exercise-induced glucagon secretory response can occur. Glucagon secretion alone cannot prevent hypoglycemia when portal venous insulin concentrations are increased by minimal amounts, such as in insulin-treated diabetics.