The goal of this study was to test the hypothesis that circulating catecholamines are primarily responsible for alpha-adrenergic coronary vasoconstriction during submaximal exercise. The experimental series consisted of chronic studies in which a regional left ventricular sympathectomy was performed with phenol. Myocardial perfusion to the innervated and sympathectomized left ventricular regions was measured in these animals during (1) a control period, (2) treadmill exercise, (3) exercise during beta-adrenergic blockade, and (4) exercise during combined alpha- + beta-adrenergic blockade. We found no differences in myocardial perfusion between the innervated and sympathectomized regions or the transmural distribution of perfusion during any of these interventions. Thus, there is no evidence for neurogenic alpha-adrenergic coronary vasoconstriction. However, during exercise in the presence of alpha- and beta-blockade, coronary resistance (mmHg X min X 100 g/ml) was significantly less in both the innervated (0.65 +/- 0.07) and sympathectomized (0.68 +/- 0.07) regions than during beta-blockade, 0.90 +/- 0.17 and 0.89 +/- 0.16, respectively. This suggests that coronary alpha-adrenergic constriction was produced by circulating catecholamines. This concept of humorally mediated, alpha-adrenergic coronary vasoconstriction was strengthened by in vivo and in vitro studies that demonstrated that alpha-adrenergic supersensitivity of the coronary vasculature was not present. Myocardial beta-adrenergic supersensitivity was observed in the phenol regional sympathectomy model; however, this effect was blocked by propranolol (1 mg/kg). This indicates that alpha-adrenergic vasoconstriction in both myocardial regions during submaximal exercise is produced by circulating catecholamines. The major conclusion of this study is that, during submaximal exercise in the canine, alpha-adrenergic coronary vasoconstrictor tone is predominantly due to circulating catecholamines rather than direct neural effects.