We investigated the effects of alpha and/or beta adrenergic blockade (with phentolamine and/or propranolol) on glucose homeostasis during exercise in six normal subjects and in seven Type I diabetic subjects. The diabetics received a low dose insulin infusion (0.07 mU/kg X min) designed to maintain plasma glucose at approximately 150 mg/dl. In normals, neither alpha, beta, nor combined alpha and beta adrenergic blockade altered glucose production, glucose uptake, or plasma glucose concentration during exercise. In diabetics, exercise alone produced a decline in glucose concentration from 144 to 116 mg/dl. This was due to a slightly diminished rise in hepatic glucose production in association with a normal increase in glucose uptake. When exercise was performed during beta adrenergic blockade, the decline in plasma glucose was accentuated. An exogenous glucose infusion (2.58 mg/kg X min) was required to prevent glucose levels from falling below 90 mg/dl. The effect of beta blockade was accounted for by a blunted rise in hepatic glucose production and an augmented rise in glucose utilization. These alterations were unrelated to changes in plasma insulin and glucagon levels, which were similar in the presence and absence of propranolol. In contrast, when the diabetics exercised during alpha adrenergic blockade, plasma glucose concentration rose from 150 to 164 mg/dl. This was due to a significant increase in hepatic glucose production and a small decline in exercise-induced glucose utilization. These alterations also could not be explained by differences in insulin and glucagon levels. We conclude that the glucose homeostatic response to exercise in insulin-dependent diabetics, in contrast to healthy controls, is critically dependent on the adrenergic nervous system.