Previous observations (Garcia et al. 1998; Sohn et al. 1997; Nagueh et al. 1997) indicate that mitral annulus velocity may be viewed as a "load-independent" index of filling and that wedge pressure is linearly related to the transmitral flow velocity (V(E)) to mitral annulus velocity (V(MA)) ratio (V(E)/V(MA)) measured at maximum velocity. In healthy subjects, the mean value observed for [V(E)](max)/[V(MA)](max) was 1:0.23 approximately 4. No prior physiologic or physical explanation for the basis of, or for the magnitude of, the ratio has been proposed. We propose a physiologic, model-based, quantitative explanation for these observations and test our simplified model's prediction in an invasive (n = 30) and noninvasive (n = 34) test groups of subjects. The simplified geometric model is based on the known constant volume (within a few percentage points) attribute of the four-chambered heart. Accordingly, left-atrial and left-ventricular volumes reciprocate so that their sum is constant throughout the cardiac cycle. The model predicts that: 1. the ratio (V(E)/V(MA)) is a constant approximately 3 in healthy hearts; and 2. V(E)/V(MA) should be linearly proportional to left ventricular end-diastolic pressure (LVEDP). Model prediction was tested using V(E) and V(MA) echocardiographic data from 34 subjects (noninvasive group), and simultaneous echocardiographic and high-fidelity hemodynamic (LVEDP) data in 30 subjects (invasive group). Excellent agreement was observed between model prediction and observed data. For the noninvasive (healthy) group, [V(E)](max)/[V(MA)](max) = 4.20 +/- 1.11. For the invasive group, [V(E)](max)/[V(MA)](max) was observed to be linearly related to LVEDP, [V(E)](max)/[V(MA)](max) = 0.19 (LVEDP) + 1.54, r = 0.92. Hence, [V(E)](max)/[V(MA)](max) is a legitimate flow-derived index of diastolic function because it is a derivable consequence of the heart's constant-volume pump attribute.