Summary Nineteen mongrel dogs survived chronic occlusion of the left circumflex and of the right coronary artery without infarction due to the timely development of a collateral circulation. Only 38 per cent of the conductance of the arteries before occlusion was restored by collateral vessels. In these animals and in 15 control dogs with normal coronary arteries myocardial contractility, contractility reserve, and myocardial blood flow were studied. The same was done in dogs with chronic coronary artery occlusion after aortocoronary bypass. Myocardial blood flow was determined with the tracer microsphere technique. Contractility reserve was tested and defined as isovolumetric left ventricular pressure and dp/dt max with norepinephrine infusion and cross-clamping of the aorta. Contractile reserve was not significantly different between normal dogs and dogs with chronic coronary artery occlusion before and after aortocoronary bypass. Myocardial blood flow during control conditions was homogeneously distributed in all three groups studied. The ratio of blood flow to the endocardium and the epicardium was not significantly diddrent from unity. Coronary reserve was determined at peak reactive hyperemia following a 20 second period of coronary artery occlusion, with ongoing norepinephrine infusion. Under these conditions subendocardial flow in normal dogs rose by a factor of 7.9 while subepicardial flow increased 7.4 times. In dogs with chronic occlusion of two coronary arteries the increase of myocardial flow was nonhomogeneous; subendocardial flow to areas supplied by a normal coronary artery rose by a factor of 7.0 while subepicardial flow increased 5.7 times control. Subendocardial collateral flow rose by a factor of 2.4 and subepicardial collateral flow increased 3.5 times control. In normal dogs norepinephrine alone did not result in maximal coronary flow but only 57 per cent thereof. Dogs with chronic coronary occlusion, however, required the entire coronary reserve in areas that were supplied by a normal coronary artery, whereas areas supplied by collaterals became ischemic. Opening of an aortocoronary bypass restored normal flow to previously ischemic areas, and reduced the flow to areas supplied by a normal artery. With the bypass open no differences existed between normal dogs and those with two occluded coronary arteries. We conclude that the norepinephrine-stimulated contractile reserve of hearts with chronic coronary occlusion was comparable to that of normal hearts; however, norepinephrine forced these hearts to spend the entire flow reserve of the remaining normal artery while producing ischemia in collateral-dependent areas. The same dose of norepinephrine did not require the entire flow reserve of normal dogs.