To examine whether different septal pacing sites could be distinguished by their epicardial activation patterns, six to eight stimulating electrodes were placed throughout the septum in seven open chest dogs. Unipolar electrograms were obtained from 52 epicardial electrodes during pacing from each stimulating electrode and isochronous epicardial maps were constructed. The location of each stimulating electrode was found by dissection, and its distance from the overlying epicardium was measured. To allow comparison among epicardial maps, the septum was conceptually subdivided into nine regions to which stimulating electrodes were assigned. Epicardial activation patterns from the same region were similar and these patterns allowed the region containing a stimulating electrode to be identified in many cases. Three other variables were found to have additional localizing value. There were: 1) the time from the stimulus to epicardial breakthrough, 2) the duration of epicardial activation, and 3) the area of epicardium activated in the first 5 ms after epicardial breakthrough. For those stimulating electrodes that could not be localized by their epicardial activation patterns, the distance of the stimulating electrode beneath the epicardium was well fit from these three variables by multiple regression (correlation coefficient [r] = 0.97). Thus, using all the previous factors, localization of septal pacing sites was possible in the noninfarcted dog heart by epicardial mapping. © 1983, American College of Cardiology Foundation. All rights reserved.