Abstract: | An accurate representation of the thoracic surface is required to solve the inverse problem of electrocardiography. For the purpose we have searched the second degree surface which fits 55-88 positions distributed around the trunk. Over a group of 17 young men, we have found that the positions always scattered around an ellipsoid the cephalo-caudal limits of which extended far away from the zero potential gradient levels. This discrepancy is discussed and a proposal is formulated. |