SUMMARY Using an analytical mathematical model, we studied and contrasted the effects of variations in geometry and volume conductor properties of the torso on epicardial and body surface potentials. The model consists of a spherical heart (blood cavity bounded by a spherical muscle shell that includes a double layer source, and pericardium) eccentrically placed in a spherical torso flung region bounded by muscle and fat layers). The effects of the following parameters on body surface and epicardial potentials were studied: (1) separation of the cardiac sources; (2) location of the heart within the torso; (3) combined effects of all torso inhomogeneities, (4) "internal" inhomogeneities (intracavitary blood, pericardium); (5) "external" inhomogeneities flung region, skeletal muscle, subcutaneous fat), and (6) hypertrophy and dilation. It was determined that, although internal inhomogeneities affect both epicardial and surface potentials similarly, the effect of external inhomogeneities on body surface potentials is different from their effect on epicardial potentials. The effects of hypertrophy and dilation are seen to depend on specific details regarding alterations in size and shape of blood cavity, heart, and activation surface. The most important conclusion of the study is that epicardial potential maps accurately reflect the underlying source configuration, are free of the effects of body shape and size, and are affected significantly by only one extracardiac inhomogeneity-namely, the lung region. Such maps, therefore, can enhance our capability to interpret and diagnose electrophysiological events within the heart. Ore Res 46: 283-291, 1980AS a consequence of the electrical activity of the heart, electrical potentials appear throughout the volume conductor in which the heart is embedded. In addition to body surface potentials which constitute the data in electrocardiography, potential distributions over the epicardium are also of great interest since they may mirror events within the heart that are not distinctly reflected at the body surface. Studies conducted on closed (intact) animals in which epicardial and intramural electrodes have previously been implanted (Spach et al., 1969;Spach and Barr, 1975a, 1975b demonstrate that epicardial potential maps contain a considerable amount of information about the underlying intramural electrophysiological events.A very important property, arising from potential theory, is that, in principle, epicardial potentials can be recovered from body surface data (Martin and Pilkington, 1972;Martin et al., 1975). Based on this capability, there is hope that epicardial data may be available noninvasively through computations based on body surface potentials and body geometry. This inverse solution, in contrast with a solution which represents the activity of the heart in terms of dipoles, can be evaluated by a direct