Left ventricles of control dog hearts and dog hearts failing due to chronic tachycardia were examined in vivo by echocardiography for systolic function and size, then subsequently studied with an isolated-heart system (artificial perfusion, artificial loading). During 3 weeks of tachycardia (250 bt/min), area ejection fraction fell by 58%, while end-diastolic transverse area increased by 56% (measurements at 120 bt/min). Judging from post-perfusion left-ventricular weights, the dilation occurred with no hypertrophy, raising the question whether the failure model may be associated with anabolic dysfunction. End-diastolic pressure-volume (P-V) relations occurred at higher volumes in failing chambers than in controls, and this was marked by increases in two indices of chamber size (candidate reference volumes): the volume resulting in a diastolic stress of 16 g/cm2, and the volume at which the nearly straight, low-stiffness segment of the end-diastolic P-V relation meets the upward bending, high-stiffness segment. Developed P-V relations of failing chambers were shifted to higher volumes and to lower pressures, the lower pressures being due more to reduced stress-developing ability (contractility) than to reduced wall/cavity ratio (pressure/stress ratio). On average, shortening ability (normalized difference between reference volume and extrapolated volume-axis intercept, i.e., apparent ejection fraction from reference volume in absence of afterload) was not different from that of controls. Isovolumic pressure waves of the failing and dilated chambers were of almost normal duration and shape, extending further the range of conditions where isovolumic pressure can be predicted by fitting a model isovolumic wave function to the isovolumic phases of ejecting beats.