BACKGROUND Electrical and mechanical alternans are thought to result from a common cellular mechanism. To confirm this phenomenon in vivo and extend our understanding, we investigated the effects of temperature, verapamil, and caffeine on alternans in intact beating hearts. RESULTS We recorded surface ECG, monophasic action potential (MAP) using suction electrodes, and left ventricular pressure (LVP). Alternans of MAP configurations and of LVP were evaluated in 20 dogs. MAPs were recorded from the apex, base, and midportion of the left ventricle with LVP. The hearts were driven from the left ventricular apex at a basic cycle length of 1000 milliseconds, and alternans was induced with an abrupt shortening of the cycle length to 400 milliseconds. MAPD30/100, the ratio between the MAP durations (MAPD) at 30% and 100% repolarization levels, was measured as an index of MAP configuration. The magnitude of MAP or LVP alternans was defined as the difference in MAPD30/100 or in LVP between the fifth and sixth paced beats. The magnitude of MAP alternans differed by recording site but correlated with activation time. Lowering the temperature increased the magnitude. MAP alternans was always associated with LVP alternans under control conditions. Verapamil significantly attenuated the magnitude of MAP alternans but did not change that of LVP alternans. Caffeine attenuated the magnitude of both MAP and LVP alternans. T wave alternans was suppressed by verapamil or caffeine. CONCLUSIONS Activation time is one of the factors modifying electrical alternans. Delayed intracellular Ca2+ cycling plays a role in the concomitant occurrence of electrical and mechanical alternans.