A great deal of confusion exists concerning the reactions of the coronary blood vessels to many products of metabolic origin and to numerous drugs. The prevailing contradictions regarding the reaction of the coronary vessels arise from the circumstances under which the measurements have been made, the indirectness of the methods used in some studies and the abnormality of the conditions in others. We have become concerned with this problem and have developed a method by which the action of various drugs and other substances directly on the smooth muscle of the coronary vessels can be determined without at the same time being altered by the other factors which indirectly affect the rate of coronary flow. Our method, which employs a normal denervated dog's coronary vessels perfused in situ with defibrinated dog's blood, permits a direct determination of whether a substance has a vasodilator or a vasoconstrictor action on the coronary