Introduction: Although numerous theories exist for the mechanisms of defibrillation, experimental data directly relating these mechanisms to the termination of reentry in whole hearts are lacking. Methods and Results: Using video imaging technology, we recorded ∼5,000 optical action potentials simultaneously from the anterior and posterior ventricular epicardium of rabbit hearts during cardioversion of stable reentrant arrhythmias. Monophasic shocks at three strengths for each polarity were delivered between electrodes inside the right ventricle (RV) and above the left atrium. Cardioversion efficacy at the three strengths was 21%, 42%, and 92% for RV+ shocks, which primarily depolarized the epicardium, and 10%, 15%, and 33% for RV- shocks, which primarily hyperpolarized the epicardium. The mechanism of cardioversion for RV+ shocks was elimination of excitable gaps and reentry via excitation ahead of wavefronts and action potential prolongation at wavetails, both of which increased with shock strength. Partial elimination of these gaps resulted in resetting of preshock reentry and/or induction of new reentry. RV- shocks cardioverted primarily via deexcitation, which terminated reentry by creating new postshock wavefronts via break excitation that rapidly activated excitable gaps. Outcome was dependent on the preshock state for both polarities at strengths near the 50% success level. Before successful shocks, more epicardium was recovered, resulting in more excitation and longer postshock depolarization (RV+ shocks) and faster postshock elimination of excitable gaps (RV- shocks). Conclusion: These findings provide a direct mechanistic link between shock-induced changes in Vm and the effect of polarity, strength, and timing on cardioversion efficacy.