The hypothesis that exercise-induced coronary vasodilation is a result of sympathetic activation of coronary smooth muscle beta-adrenoceptors was tested. Ten dogs were chronically instrumented with a flow transducer on the circumflex coronary artery and catheters in the aorta and coronary sinus. During treadmill exercise, coronary venous oxygen tension decreased with increasing myocardial oxygen consumption, indicating an imperfect match between myocardial blood flow and oxygen consumption. This match was improved after alpha-adrenoceptor blockade with phentolamine but was significantly worse than control after alpha + beta-adrenoceptor blockade with phentolamine plus propranolol. The response after alpha-adrenoceptor blockade included local metabolic vasodilation plus a beta-adrenoceptor vasodilator component, whereas the response after alpha + beta-adrenoceptor blockade contained only the local metabolic vasodilator component. The large difference in coronary venous oxygen tensions during exercise between alpha-adrenoceptor blockade and alpha + beta-adrenoceptor blockade indicates that there is significant feedforward beta-adrenoceptor coronary vasodilation in exercising dogs. Coronary venous and estimated myocardial interstitial adenosine concentrations did not increase during exercise before or after alpha + beta-adrenoceptor blockade, indicating that adenosine levels did not increase to compensate for the loss of feedforward beta-adrenoceptor-mediated coronary vasodilation. These results indicate a meaningful role for feedforward beta-receptor-mediated sympathetic coronary vasodilation during exercise.