It is now recognized that a sizable portion of patients who exhibit symptoms of congestive heart failure have relatively well-preserved systolic function, but have significantly elevated LV filling pressures. This syndrome, termed "diastolic heart failure," is associated with various conditions such as aging, anatomic abnormalities, hypertension, ischemic disease, tachycardia, and atrial fibrillation. Advances in the proper medical and surgical management of these patients will depend on the continued delineation of the basic physiologic mechanisms that account for normal and pathologic cardiac diastolic function. This goal can only be achieved by the integration of information acquired from basic science investigations conducted in vitro and in vivo, mathematic modeling simulation studies, and prospective, community-based investigations that characterize the incidence, prevalence, and natural history of the disease. In addition, randomized clinical trials will be needed to determine the optimal treatment strategies for this group of patients--strategy choices undoubtably complicated by a disease whose treatment is influenced to a large extent by its origin. The future therapies evaluated in these randomized clinical trials will most likely range from medical therapies that target either the heart directly or the peripheral vascular system, to surgical interventions such as direct myocardial revascularization, to gene therapy. Finally, it is worth mentioning one more unresolved issue that is of general practical concern not only to the physiologist studying diastolic function, but also to the clinician: whether or not it is even feasible to develop a single, sensitive, specific, clinically relevant index of diastolic function that is free from the contaminating influences of rate, contractility, and load. As observed by Glantz 20 years ago, developing indexes with the hope that one might fully delineate the left ventricle's diastolic properties, rather than concentrating on discovering the physiologic significance of such indexes, is probably counterproductive. More recently, in a related article, Slinker implied that an operational definition of any aspect of cardiac function must allow for the measurement of that function over an adequate range of essential variables. Therefore, as previously mentioned, the physiologist studying cardiac function has the daunting task of trying to understand, in a precise way, how the processes and mechanisms of the various phases of the cardiac cycle couple together to produce either a normal or abnormal functioning heart. It seems clear that because of the complex weave of factors that control overall cardiac diastolic function, the derivation of any single index that adequately describes LV diastolic function in vivo may not be possible.