Evaluation of patients for aotocoronary bypass (ACB) surgery is dependent on the findings at cardiac catheterization. Left ventricular angiography cannot distinguish between reversible and irreversible myocardial dysfunction. In 82 patients macroaggregates of human serum albumin, labelled with iodine-131, were injected before operation into the right coronary artery and technetium-99m labelled microspheres were injected into the left coronary artery. Two weeks after operation similar technetium-99m and iodine-131 injections were made into the native vessels and grafted vessels, respectively. Myocardial distribution of the 20 mu particles was visualized with a rectillinear scanner. Sixty-seven patients had normal myocardial perfusion scans preoperatively even though wall motion was abnormal in 34. There was substantial improvement in the left ventricular function of these 34 patients; the mean ejection fraction increased from 56 to 65%. Fifteen patients had abnormal myocardial perfusion scans; improvement in left ventricular function in this group occurred only if the bypass graft was made to an area with normal regional perfusion preoperatively. It is certain that an adequate capillary bed is present if the myocardial scan is normal, regardless of the degree of critical stenosis in either coronary artery. The patient with a completely occluded coronary artery and normal myocardial scan is a good candidate for ACB and will show improved left ventricular contraction if the ACB is patent,