&NA;Standard microelectrode techniques were used to study the effects of a free radical generating system on action potentials recorded from guinea pig ventricular myocardium. Free radicals were generated by mixing xanthine oxidase (0.02‐0.04 μ/ml) with the superfusate‐modified Locke's solution containing purine 2.3 mM. The system was validated by demonstrating that it could reduce cytochrome C at a rate of 15.9 ± 1.5 mol/l/min. This rate was decreased to 3.0 ± 0.3 (p<0.001) in the presence of superoxide dismutase (12 mg/100 ml), and the reaction was absent if xanthine oxidase and purine were premixed for 60 minutes prior to adding cytochrome C. Superfusion of guinea pig ventricular strips with the free radical generating system (20‐30 minutes) resulted in a highly significant reduction in resting potential from − 79.3 ± 1.8 mV to −70.9 ± 1.4 mV (p<0.0001,n= 6) and in action potential amplitude from 110.9 ± 2.2 mV to 101.7 ± 4.0 mV (p<0.0001). There was an accompanying fall in maximum rate of depolarization (Vmax) from 254.1 ± 17.7 V/sec to 207.1 ± 18.6 V/sec (p<0.01) and no significant change in action potential duration. These changes were accompanied by spontaneous activity in 3 of 6 preparations and reversed after 20‐30 minutes washing in Locke's solution. They were largely abolished by adding superoxide dismutase (12 mg/100 ml) to the superfusate and completely absent if the xanthine oxidase and purine were premixed for 60 minutes before superfusing the myocardium. We conclude that the phenomena observed may contribute to the genesis of reperfusion arrhythmias. (Circulation Research1987;61:50‐54)