We sought to determine the ejection variables that are principally responsible for increases in isovolumic ventricular relaxation rate observed with increases in stroke volume. In nine isolated canine hearts, left ventricular ejection was controlled by patterns specially designed to isolate the ejection parameters most critical to isovolumic relaxation rate. When stroke volume was augmented by increases in end-diastolic volume (EDV) with end-systolic volume (ESV) held constant, isovolumic ventricular relaxation rate was unchanged, as gauged by the time constants of single-exponential fits to decaying pressure. In contrast, when ESV was decreased with EDV held constant, isovolumic relaxation time constants decreased significantly, from approximately 100 to 70 msec (protocol I). The important difference in these two situations might have been that the time of end ejection was delayed in the case with faster isovolumic relaxation. To rule out other parameters that may have influenced isovolumic relaxation, ejection velocity was varied in another protocol (protocol II) by either delays in time of the onset of ejection or advances in end-ejection time, always with constant ESV and EDV. Here isovolumic relaxation was progressively slowed as end ejection occurred earlier, whereas isovolumic relaxation rate was insensitive to changes in the onset of ejection, consistent with the unique importance of end ejection to isovolumic relaxation. In fact, our analysis reveals the remarkable finding that changes in isovolumic relaxation time constant produced by either protocol I or protocol II could be related quantitatively to end ejection by a single curve.(ABSTRACT TRUNCATED AT 250 WORDS)