We devised a pulsed Doppler echocardiographic method of separately calculating early diastolic filling volume (EDFV) and late diastolic filling volume during atrial contraction (LDFV) and observed a relationship between diastolic filling volume and velocity in thirty patients with coronary artery disease. By analysing the transmitral flow velocity curve and mitral valve motion, EDFV and LDFV were measured on the basis of the equality of left ventricular inflow and outflow volumes. The Doppler-determined EDFV and LDFV correlated well with those obtained from the left ventricular filling curve produced by left ventriculography. Angiographic EDFV and LDFV were measured from the time (t)-volume (V) curve, using the t-dV/dt curve to define early and late diastolic phases. A good correlation was found between Doppler and angiographic EDFV (y=-3.0+1.0 x, r=0.98, p=0.0001, n=20), Doppler and angiographic LDFV (y=1.6+1.0 x, r=0.86, p=0.0001), and also between Doppler and angiographic EDFV/LDFV (y=0.05+0.9 x, r=0.93, p=0.0001). EDFV and the peak early diastolic filling velocity were significantly correlated (E velocity; y=25+0.51 x, r=0.48, p=0.0068), while LDFV and the peak late diastolic filling velocity during atrial contraction (A velocity) were not. Our results validate the method of calculating EDFV and LDFV separately and suggest that early diastole in the left ventricle has flow volume dependency, but that the late diastole filling velocity during atrial contraction may be regulated by other factors such as increased left atrial contraction.