Our goal was to better understand the mechanisms underlying muscarinic receptor actions on the ventricle in vivo. Therefore, we studied the effects of vagal stimulation on ventricular repolarization and of vagal tone on lethal arrhythmias induced by 30 minutes of left anterior descending coronary artery ligation in anesthetized cats. Experimental groups included normal control cats subjected only to coronary ligation and cats pretreated with atropine, pertussis toxin (PTX), or propranolol. All cats received bilateral cervical vagal stimulation (Vstim) at 1, 3, and 5 Hz for 1 minute at 10-minute intervals. Before coronary ligation, Vstim slowed sinus rate, prolonged the PR interval, and lowered blood pressure. Most important from the point of view of electrophysiological function was a vagally induced acceleration of ventricular repolarization in paced and unpaced hearts, which could be explained by the effects of acetylcholine (ie, shortening the subepicardial muscle action potentials). The effect on repolarization was blocked by atropine or PTX but not by propranolol. The extent of sinus slowing and acceleration of repolarization was directly related to the level of functional PTX-sensitive G protein (P < .05). Coronary occlusion was performed during atrial pacing such that the heart rate in all groups was equal. The incidence of ventricular fibrillation (VF) was 10% in the control group and 50% and 54% in atropine and PTX groups, respectively (P < .05). During atrial pacing before coronary occlusion, a vagal index was calculated as percent QTc shortening during Vstim. When the vagal index was 13% to 26%, the incidence of VF during occlusion was zero. When the vagal index was 0% to 12%, VF was 52% (P < .01). Conclusions are as follows: (1) Vstim accelerates ventricular repolarization in cats via a pathway that incorporates a PTX-sensitive G protein and involves an altered gradient between epicardium and endocardium. (2) Removal of vagal tone during ischemia favors VF, as predicted by a vagal index.