Magnesium sulfate, reportedly clinically effective against some ventricular arrhythmias, has not been extensively studied for its effects on experimental ischemic and reperfusion arrhythmias. We evaluated the effects of three high extracellular concentrations of magnesium (Mg2+) 1.19, 2.38, and 4.76 mM in 70 isolated perfused rat hearts following coronary artery ligation and reperfusion, at each of three different perfusate ionized calcium (Ca2+) concentrations (0.9, 1.25, and, 2.5 mM), where 1.25 mM is close to physiologic. At 0.9 mM Ca2+, higher concentrations of Mg2+ increased the sinus node cycle length (p less than 0.02) and virtually abolished ischemic ventricular tachycardia (VT) and reperfusion ventricular fibrillation (VF) (p less than 0.01), otherwise consistently found in this model. At 1.25 and 2.5 mM Ca2+ increasing Mg2+ had no effect on the incidence of ischemic or reperfusion arrhythmias, although at 1.25 mM Ca2+ ischemic VT had longer cycle lengths and VT appeared after a longer delay (p less than 0.01). In the nonischemic dog heart, marked increases of serum Mg2+ progressively prolonged the A-H, H-V, and QR S intervals, the ventricular effective refractory period, and the sinus cycle length, while the arterial blood pressure fell. Because of its relatively modest electrophysiologic and hemodynamic effects, it is inferred that intravenous magnesium may be given therapeutically with relative safety.