We discovered that the coupling beat interval from a slow to a tachycardiac pacing period considerably affected the pattern of the beat-to-beat alternation of the tachycardia-induced sustained contractile alternans. We analyzed the relationship between the coupling interval and the pattern and amplitude of the alternans in the isovolumic left ventricle of canine blood-perfused hearts. The alternans pattern and amplitude varied transiently over the first 30-50 beats and became gradually stable over the first minute in all 12 hearts. We discovered that stable alternans, even under the same tachycardiac pacing, had three different strong-weak beat patterns depending on the coupling interval. A relatively short coupling interval produced a representative sustained alternans of the strong and weak beats. A relatively long coupling interval produced a similar sustained alternans but in a reversed order of even- and odd-numbered beats counted from the coupling interval. However, sustained alternans disappeared after 1-3 specific coupling intervals. We conclude that ventricular pacing rate does not solely determine the pattern and amplitude of sustained contractile alternans induced by tachycardia.