The extent of old posterodiaphragmatic myocardial infarction has been assessed by vector and scalar measurements of the Frank orthogonal ECG. 121 men angiographically proven to have coronary artery disease were selected on the basis of a Q/R amplitude ratio of greater than or equal to 0.25 and T wave inversion in lead aVF. They were grouped according to varying degrees of severity and the extent of wall motion disorders of the diaphragmatic and posterobasal segment of the left ventricle: (1) hypokinesis, n = 24; (2) akinesis or dyskinesis of one segment, n = 60; (3) akinesis and/or dyskinesis of two segments, n = 34. Discriminant analysis was performed by entering 11 QRS measurements. Rao's selection method permitted the 5 most decisive variables to be used for the actual analysis. They were the sum of elevation angles of the 5 initial vectors obtained at 0.01 sec intervals, the sum of magnitudes of the 3 terminal vectors at 0.01 sec intervals, the magnitude of the -0.03 sec terminal vector, the magnitude of the -0.04 sec terminal vectors, and the sum of duration of R in lead Y + R in lead Z. The classification routine could then identify correctly 70.6% of group 3, but only 50% and 46.7% of groups 1 and 2, respectively. 20.6% and 8.8% of group 3 were falsely predicted to be members of groups 2 and 1, respectively. Only 8.3% of group 1 were falsely classified as members of group 3. Thus, the electrocardiographic measurements enable transmural myocardial infarction to be predicted with confidence when the patient is classified as member of group 3.(ABSTRACT TRUNCATED AT 250 WORDS)