Depolarization of cardiac muscle is achieved by ‘fast inward current’ through channels which are inactivated within about 1 ms. When the cells are repolarized the process of inactivation of fast channels is rapidly reversed. The class 1 anti-arrhythmic drugs delay the disappearance of inactivation until long after repolarization is complete. In theory, it should be possible to produce a similar extension of refractory period by delaying the repolarization itself. Quinidine and disopyramide caused minor delays of repolarization, but both were primarily class 1 agents, and in addition had undesirable anticholinergic activity. Amiodarone, already in use for many years as an atianginal drug, prolonged action potential duration (APD) and was shown to have an anti-arrhythmic action in rabbits, dogs and man. Although prolongation of APD lengthens QT, a long QT may be caused by phenomena other than prolonged APD, such as heterogeneity of sympathetic drive. Association of long QT with arrhythmia does not, therefore, invalidate the principle that homogeneously prolonged APD should be anti-arrhythmic. In practice, amiodarone, bretylium, sotalol, thyroidectomy, and long-term beta-blockade prolong APD, and are associated with low incidence of arrhythmia. Many mechanisms controlling cardiac repolarization have been proposed, but how repolarization is delayed by individual agents is not fully elucidated.