Class 3 antiarrhythmic agents exhibit reverse rate-dependent lengthening of the action potential duration (APD), i.e. changes in APD are greater at longer than at shorter cycle lengths. In spite of the several theories developed to explain this reverse rate-dependency, its mechanism has been clarified only recently. The aim of the present study is to elucidate the mechanisms responsible for reverse ratedependency in mammalian ventricular myocardium. Action potentials were recorded using conventional sharp microelectrodes from human, canine, rabbit, guinea pig, and rat ventricular myocardium in a rate-dependent manner. Rate-dependent drug-effects of various origin were studied using agents known to lengthen or shorten action potentials allowing thus to determine the drug-induced changes in APD as a function of the cycle length. Both drug-induced lengthening and shortening of action potentials displayed reverse ratedependency in human, canine, and guinea pig preparations, but not in rabbit and rat myocardium. Similar results were obtained when repolarization was modified by injection of inward or outward current pulses in isolated canine cardiomyocytes. In contrast to reverse ratedependence, drug-induced changes in APD well correlated with baseline APD values (i.e. that measured before the superfusion of drug or injection of current) in all of the preparations studied. Since the net membrane current (Inet), determined from the action potential waveform at the middle of the plateau, was inversely proportional to APD, and consequently to cycle length, it is concluded that that reverse rate-dependency may simply reflect the inverse relationship linking Inet to APD. In summary, reverse rate-dependency is an intrinsic property of drug action in the hearts of species showing positive APD - cycle length relationship, including humans. This implies that development of a pure K+ channel blocking agent without reverse rate-dependent effects is not likely to be successful.