BACKGROUND: Volume loading is commonly used to adjust preload and optimise cardiac output. It is difficult to monitor preload at the bedside because filling affects ventricular diastolic function and consequently end diastolic pressure, which is the variable used to monitor preload. OBJECTIVE: To assess the effects of gradual volume loading on the different components of left ventricular diastolic function---filling velocities, relaxation, and chamber compliance---to identify how excessive loading produces diastolic dysfunction. METHODS AND RESULTS: Eight mongrel dogs, anaesthetised and mechanically ventilated with both the chest and the pericardium closed, were studied during basal conditions (B), during gradual volume loading with physiological saline---5 ml/kg (VL5), 10 ml/kg (VL10), and 15 ml/kg (VL15)---and during infusion of isosorbide dinitrate (10 g/kg/min) started after the VL15 load was achieved. Dogs were monitored haemodynamically and by transthoracic Doppler echocardiography to assess peak modal velocities of the E and A waves, E/A ratios, and the deceleration time of the E wave. M mode recordings of aligned mitral and aortic valve motion were also obtained to calculate the isovolumic relaxation time. Effects of volume loading on ventricular diastolic function seemed to occur in two phases. Small and moderate volume loads (VL5 and VL10) promoted early ventricular filling, increasing E wave velocities, improving the mean (SD) E/A ratio from 1.95 (0.3) (B) to 2.0 (0.27) (VL5) and 2.6 (0.3) (VL10) (P < 0.00005), prolonging the E wave deceleration time, and only slightly increasing ventricular diastolic pressures. These changes suggest an improvement in ventricular compliance. Extreme volume loads (VL15) produced an abrupt reduction in early ventricular filling, which was transfered to late in diastole, by decreasing E wave velocity, by increasing A wave velocity, and by decreasing E/A ratio from 2.6 (0.3) (VL10) to 0.8 (0.05) (VL15) (P < 0.00005). The E wave deceleration time was shortened and left ventricular diastolic pressures were much increased, all suggesting a deterioration in chamber compliance. All these restrictive changes were promptly reversed by the perfusion of isosorbide dinitrate. The isovolumic relaxation time steadily increased with volume loading. CONCLUSIONS: Small and moderate volume loads improved ventricular diastolic function by promoting early ventricular filling and increasing ventricular compliance. Extreme volume loads promptly induced a diastolic restrictive pattern, transferring filling to the second part of diastole (increasing dependence on atrial contraction) and reducing ventricular compliance. These changes in ventricular diastolic function were independent of simultaneously measured haemodynamic systolic performance and were promptly reversed by isosorbide dinitrate, which after extreme loading promoted early filling, myocardial relaxation, and improved chamber compliance.