Previous electrocardiographic models of myocardial ischemic injury have assumed that transmembrane potential changes are uniform throughout a region of ischemia such that injury currents arise exclusively at the boundary between normal and ischemic myocardium. In such models, the distribution and amplitude of ST segment deflections are considered to arise from a polarized surface interfacing normal and ischemic myocardium. This concept in modeling ischemic injury was derived from the application of principles of electric field theory which had been successfully applied previously to ventricular activation in which QRS potentials are considered to arise from polarized surfaces representing the relatively narrow interfaces between depolarized and nondepolarized myocardium. The present paper outlines the limitations of modeling ischemic injury as a polarized surface in terms of the failure of the predictions of such a model to be supported by the experimentally observed: 1) distribution and relative amplitude of epicardial ST segment elevation overlying a region of ischemia; 2) directional changes in epicardial ST segment elevation that occur with changes in the size of an ischemic region; and 3) nonuniform distribution of transmembrane potential changes which occur within a region of ischemia. A new electrocardiographic model of ischemic injury is formulated which accounts for the nonuniform distribution of transmembrane potential changes which occur throughout a region of ischemia. The model accurately describes experimental observations regarding ST segment deflections which had remained inconsistent with previous models.