Changes in intra-SA nodal pacemaker localization were produced through stimulation of the decentralized cervical vagi and stellate ganglia in the anesthetized dog. Shifts in pacemaker to the rostral, middle, or caudal regions of the SA node produced a change in the timing as well as a change in the sequence of activation of recording sites overlying the AV node. Epicardial pacing with a plaque electrode from either the rostral, middle, or caudal regions of the SA node produced the same activation sequence of the AV nodal electrodes irrespective of the epicardial SA nodal pacing site. The inability of epicardial SA nodal pacing to precisely reproduce the activation pattern of the atrial septum overlying the AV node observed with a natural SA nodal pacemaker can be explained by the geographic relationship of the pacemaker cells within the node to the preferential internodal pathways and the area of atrial tissue stimulated by pacing. Pacing activates a large mass of tissue, whereas an intrinsic pacemaker probably acts as a more localized focus. The inability of pacing to reproduce the activation pattern seen with spontaneous rhythm may be a determinant in the varied P wave morphology seen with coronary sinus or AV nodal junctional rhythms, as compared with the more consistent morphology seen with pacing.