Atrioventricular (AV) node reentrant tachycardia is now routinely cured by selective radiofrequency ablation of slow AV node pathway conduction. However, debate remains concerning the optimum method for localizing the site at which radiofrequency energy should be delivered to eliminate slow-pathway conduction. Some investigators have proposed simple anatomy-guided ablations posteriorly near the ostium of the coronary sinus, whereas others suggest an electrophysiology-guided ablation using either recorded "slow potentials" or mapping of the retrograde atrial exit site of slow AV note pathway conduction when possible. To examine these issues, we systematically studied slow potentials recorded in the AV junction of patients undergoing radiofrequency catheter ablation for medically refractory AV node reentrant tachycardia. In 67 patients with the slow-fast form of AV note reentrant tachycardia, we performed detailed atrial mapping along the tricuspid annulus within the triangle of Koch. Two types of slow potentials were identified. Low-amplitude, low-frequency potentials, found in 48% of patients, were localized to the mid to posterior portions of the triangle of Koch, whereas high-amplitude, high-frequency potentials, observed in 22% of patients, were located only posteriorly near the ostium of the coronary sinus. In response to a bolus infusion of adenosine or incremental atrial pacing-induced AV node Wenckebach periodicity, the low-amplitude, low-frequency potentials showed an increased duration and further reduction in amplitude and frequency and often totally disappeared. In contrast, in spite of these maneuvers, the high-amplitude and high-frequency potentials remained unchanged. Of the 25 (37%) of 67 patients in whom the earliest retrograde atrial activation during ventriculoatrial slow AV nodal pathway conduction could be recorded, no patient exhibited low-amplitude, low-frequency potentials, and only 7 (28%) of 25 of these patients showed high-amplitude, high-frequency potentials. High-amplitude, high-frequency potentials persisted after successful radiofrequency ablation of slow pathway conduction. Fewer applications of radiofrequency energy were required for successful elimination of slow pathway conduction in patients in whom the retrograde atrial exit site of slow-pathway conduction could be localized, compared with those patients who only exhibited retrograde fast AV nodal pathway conduction. We conclude that high-amplitude, high-frequency potentials are part of atrial activity, whereas the origin of low-amplitude, low-frequency potentials is unclear and may represent either true intranodal biophysical electrical activity or merely artifact or far-field potentials. Regardless, the recording of high-amplitude or low-amplitude potentials is not required for successful ablation of slow-pathway conduction, although the ability to localize the retrograde atrial exit of slow-pathway conduction may assist in the ablation procedure.