Myocardial infarction has been noted as a frequent complication of coronary artery surgery in many review series, although its causes are uncertain. Follow-up of 100 patients at 19.7 months (mean) after coronary bypass surgery identified 15 patients with perioperative myocardial infarction as judged by new, significant Q waves after surgery. There were no significnat differences in age, preoperative anginal class, previous infarction, presence of hypertension, hyperlipidemia, or frank glucose intolerance compared with the 85 patients without infarction. Nor was there a significnat difference in coronary artery score, left ventricular end-diastolic pressure, cardiac index, or presence of collaterals. Cardiopulmonary bypass time, duration of anoxic arrest, and number of vessels grafted did not differ. Perioperative infarction always occurred in the territory of a grafted vessel and not in comparably compromised, nongrafted areas. In 13 cases new Q waves appeared in the first 24 hours, and myocardial infarction was not clinically suspected. Eight of the 15 grafts at risk were patent at late follow-up. Mean ejection fraction was not significantly changed postoperatively, but affected segmental wall motion declined in most cases. Five patients with perioperative infarction but no patent grafts improved by only 1 NYHA class (mean), but ten patients with infarction and one or more grafts patent improved by 2.9 NYHA classes (mean). Perioperative infarction could not be correlated with currently recognized patient and operative risk factors. The consistent anatomical relationship suggested that the grafting procedure itself was critical to the occurrence of distal segmental infarction.