Serial preoperative and postoperative electrocardiograms and vectorcardiograms were obtained in 500 patients undergoing saphenous vein aortocoronary artery bypass graft surgery. Evidence of transmural myocardial infarction was found early postoperatively in 67 patients (13 percent). Age and sex distributions, number of vessels diseased or vessels grafted, and preoperative and postoperative New York Heart Association functional classification (mean follow-up, 26 months) did not differ in the groups with and without infarction. Increased duration of cardiopulmonary bypass time (more than 120 minutes) was slightly greater in the group with infarction ( P < 0.05). Multivariate analysis revealed that 60 percent of patients in the group with infarction were identified by a 1st day serum glutamic oxaloacetic transaminase value greater than 100 U/liter; however, for each such patient identified, there was approximately one false positive result. Use of other values (creatine phosphokinase, cardiopulmonary bypass time and total anoxic rest time) did not improve discrimination. Twenty-five percent of all transmural infarctions occurred within the zone of myocardium supplied by a diseased ungrafted artery. In 32 patients with early evidence of transmural myocardial infarction in a zone of myocardium supplied by a grafted artery, postoperative angiography showed as many with patent as with occluded grafts. Of 154 patients in the group without infarction who had early postoperative graft angiograms, 30 (19 percent) had one graft occluded and yet no evidence of transmural infarction by our criteria. Therefore, early postoperative evidence of transmural myocardial infarction as defined in this study is an unreliable indicator of the status of the graft supplying the zone of infarction.