Although in situ end-systolic pressure-volume relations (ESPVRs) are approximately linear throughout a limited load range, they often yield seemingly "negative" volume axis intercepts (V0) and V0 shifts with inotropic interventions. We tested whether or not these findings could stem from in situ ESPVR nonlinearity, and we examined the physiologic meaning and limitations of linearized ESPVR variables frequently used for assessing contractile state. Continuous left ventricular pressures and volumes were obtained by micromanometer and conductance (volume) catheters in six open-chest dogs. Left ventricular loading was varied throughout a wide range by rapid left atrial hemorrhage into a reservoir. Propranolol and verapamil were administered to reduce inotropic state, with heart rate maintained by atrioventricular sequential pacing. ESPVRs were fit to nonlinear [Pes = a(Ves-V'0)2 + b(Ves-V'0)] and linear (Pes = Ees (Ves-V0)] models. Contractile state was assessed by the slope of the ESPVR at V'0 (b, of nonlinear model) and by two other ESPVR model-independent measures: the slope of the dP/dtmax and end-diastolic volume relation, and the slope of the stroke work and end-diastolic volume relation. ESPVR was frequently curvilinear, and a significant correlation existed between the extent of nonlinearity (a) and contractile state. Volume intercepts derived from linear fits to the high load ESPVR range were mostly negative and were dependent on changes in Ees. V0 estimates derived from the low load portion were positive and relatively insensitive to Ees. Thus, in situ ESPVR displays contractility-dependent curvilinearity. The contractility range throughout which ESPVRs are essentially linear is typical for isolated hearts, but the range represents low values for in situ ventricles. Despite curvilinearity, Ees determined in situ throughout limited load ranges can accurately assess inotropic state; however, comparisons between ESPVRs should consider potential nonlinearity, and if possible, they should be made within similar end-systolic pressure ranges.