The purpose of this study was to quantitate the uncertainty inherent in the electrophysiologic mapping of ventricular preexcitation as seen in Wolff-Parkinson-White syndrome. An ink-coated needle electrode was constructed to serve as a point source of unipolar stimulation along the atrioventricular junction. Activation times for 11 ventricular mapping sites at the atrioventricular junction were measured for each stimulation point. Maps were successfully completed for 18 right free wall and 14 posterior septal stimulation points. The activation time at the mapping site closest to the stimulation point was termed the index activation time. Activation times identical to the index activation time were noted at 1.3 +/- 0.5 mapping sites for each free wall stimulation point and 1.9 +/- 0.9 mapping sites for each septal stimulation point (p < 0.05, septal versus free wall stimulation points). Activation times falling within 5 msec of the index activation time were noted at 2.4 +/- 1.0 mapping sites for each free wall stimulation point and at 3.9 +/- 1.4 mapping sites for each septal stimulation point (p < 0.05, septal versus free wall stimulation points). The uncertainty of electrophysiologic mapping can be quantitated, and this error should be considered when making inferences regarding the anatomy of accessory pathways based on electrophysiologic data. A knowledge of the uncertainty inherent in the localization of accessory atrioventricular connections by electrophysiologic mapping can be used to plan borders of surgical dissection that will account for this uncertainty at a 95% confidence level.