The variability of electrographic data used up to now is too large to discriminate easily between normal and abnormal. 2. This variability is partly caused by errors in the measuring process and its consequences. A sepcial difficulty is found in determining the beginning and the end of QRS, so that instantaneous vector magnitudes related to the starting or ending moment of QRS will have a high variance and therefore a low comparability. Displacements of the heart will play an important role as a source of errors. 3. It therefore seems to be more appropriate to describe structure and changes in time of the vector loop. To this end 3 new functions v, w and u are described, all three invariant towards positional changes of the heart in space. The term v means a time-dependent function describing the first derivative of the length of the instantaneous resulting vector of all electrical sources. The term w indicates the projection of the first derivative of the instantaneous vector into a plane orthogonal to the resulting vector. Therefore, v is dependent on the change of all resulting electrical forces, whereas w is the velocity of the peak of the resulting vector on its way along the vector loop. The interpretation of u is complicated; u is a measure of how much the vector leaves a plane, in which it normally more or less remains; u is therefore a measure for the edgeside view of the vector loop. 4 These data are combined with the well-known vectorial data to form a system of vectorial description. 5 The interpretability of these data in physiological terms is exemplified. The description presented here allows to go back to cardiac processes in the proper sense and to largely elimenate external influences like those of the thoracic shape on the ecg and vcg. 6. This way of describing an ecg is specially adapted to the discrimination of normal and abnormal in preventive tests.