We compared left ventricle (LV) volume (V) simultaneously measured using the conductance catheter (VM) with volume calculated from three LV dimensions (VD) determined ultrasonically from endocardial crystals. Seven adult mongrel dogs (20-30 kg) were anesthetized and instrumented to measure micromanometer LV pressure and V. Three pairs of crystals were placed orthogonally in subendocardial positions and a conductance catheter was placed in the LV retrograde across the aortic valve. Under steady-state conditions, over the range of a single cardiac cycle, the relation between VM and VD was well described by a straight line. There was an excellent correlation of conductance and dimension volumes with r equal to 0.97 +/- 0.04 and SEE 0.8 +/- 0.5 ml. The gain (1/alpha) and parallel conductance volume (alpha VC) were constant. At lower volumes obtained during bicaval occlusion, however, the relation between VM and VD was curvilinear. 1/alpha and alpha VC both decreased as LVV fell. Thus, determination of absolute volume using the conductance catheter depended on the conditions under which the data were obtained. Under steady-state conditions, alpha VC calculated by both the saline method (mean +/- SD, 50 +/- 15 ml) and by regression of VM and VD, (45 +/- 21 ml) were similar. Consequently, absolute LV end-diastolic volumes and end-systolic volumes by the conductance and dimension methods were similar (53 +/- 14 ml and 38 +/- 14 ml vs. 56 +/- 17 ml and 44 +/- 16 ml, respectively, p = NS). When volume decreased during bicaval occlusion, there was a progressively greater decrease in VM as compared with VD. The absolute slope (EES) of the end-systolic pressure-volume relation (ESPVR) was consistently higher by the dimension method, group average, 16.3 +/- 7.6, than by the catheter, 8.5 +/- 5.9, p less than 0.05. The direction and magnitude of the change in EES at different inotropic states (autonomic blockade; dobutamine), however, was similarly measured by both the conductance catheter and dimension method. We conclude that the gain and offset of the conductance catheter are relatively constant at steady state but vary when volume is reduced by caval occlusion. Thus, the conductance catheter accurately measures absolute volumes at steady state but can underestimate the slope and position of the ESPVR when it is determined by caval occlusion. The conductance catheter does, however, accurately measure the directions and magnitude of change in contractile state.