The electrocardiographic diagnosis of posterior wall myocardial infarction remains elusive. To determine discriminating criteria a group of 27 patients with posterior infarction proven by biplane angiocardiography were compared to 97 controls. All patients had single-vessel obstruction of the circumflex artery or one of its major branches (greater than or equal to 75 percent area stenosis) without occlusive disease in the other coronary arteries. High-frequency, high-gain electrocardiograms and Cube and McFee vectorcardiograms were analyzed. Pathologic Q waves in the inferior leads were present in only 22 percent (six) of the patients; increased R-wave amplitude or duration in the right precordial leads was found in 17 to 26 percent, and an R/S ratio greater than or equal to I in lead V1 or greater than or equal to 1.5 in lead V2 was present in 22 percent (six) of patients. Vectorcardiographic criteria which improved the diagnostic yield were: (1) the presence of a QRS loop mostly anterior to the E point, and (2) the presence of an abnormally anterior T wave. This abnormal T-wave shift was present in over 70 percent of the patients with posterior infarctions and was clearly discernible from the 12-lead ECG, as manifested by tall T waves in lead V2 and flat T waves in lead V6. To approximate the T-wave angle in the 12-lead scalar ECG, an index was calculated by subtracting the amplitude of the T wave in lead V6 from its amplitude in lead V2 (T2-T6 index). An index of 0.38 mV or more yielded a sensitivity of 81 percent and a specificity of 75 percent; however, this was not as discriminating as the vectorcardiogram where a T angle of 60 degrees or more in the horizontal plane yielded a sensitivity of 70 percent and a specificity of 97 percent.