Left ventricular (LV) diastolic pressure-volume (P-V) relations arise from a complex interplay of active decay of force (i.e., relaxation), passive elastic myocardial properties, and time-varying inflow across the mitral orifice. This study was designed to quantify the passive properties of the intact ventricle and the effects of elastic recoil by separating filling from relaxation with a method of LV volume clamping with a remote-controlled mitral valve. Eleven open-chest fentanyl-anesthetized dogs were instrumented with aortic and mitral flow probes, LV and left atrium micromanometers, and a remote-controlled mitral valve. We prevented complete (end-systolic volume clamping) or partial filling at different times in diastole. The ventricle thus relaxed completely at different volumes, and we generated P-V coordinates for the passive ventricle that included negative, as well as positive, values of pressure. We then estimated ventricular volumes from ventricular weight in eight dogs, using regression equations based on data in the literature, to determine the equilibrium volume (V0), that is, volume at zero transmural pressure, in the working ventricle. We abandoned the traditional exponential approach and characterized by the P-V relation with a logarithmic approach that included maximum LV volume (Vm), minimum volume (Vd), and stiffness parameters (Sp and Sn) for the positive (p) and negative (n) phases: Pp = -Sp In[(Vm - V)/(Vm - V0)] and Pn = Sn In[(V - Vd)/(V0 - Vd)]. With this formulation, the chamber compliance, dP/dV, is normalized by the LV operating volume, and Sp and Sn are size-independent chamber stiffness parameters with the units of stress. In eight ventricles with LV weight = 131 +/- 20 g, Vm = 116 +/- 18 ml, V0 = 37 +/- 6 ml, and Vd = 13 +/- 2 ml, stiffness Sp = 14.6 mm Hg and Sn = 5.1 mm Hg were determined from the slopes of the log-linearized equations. Also, the duration of LV relaxation is increased by the process of ventricular filling (161 +/- 31 msec, filling versus 108 +/- 36 msec, nonfilling, measured from dP/dtmin, p less than 0.0001). We conclude that volume clamping is a useful method of studying restoring forces and that the logarithmic approach is conceptually and quantitatively useful in characterizing the passive properties of the intact ventricle.