The static pressure-volume (PV) curve from TLC to RV of 11 human subjects was fitted by a hyperbolic-sigmoid model: P = k1/(VM--V)+k2/(Vm--V)+k3, where VM and Vm are the upper and lower asymptotes respectively, and k1, k2, k3 are shape constants. Least-squares nonlinear regression was used to evaluate the constants for the individual and mean data. Average SD of residuals was 0.57 cm H2O and average reduction of residual variance was 99.93%. In spite of substantial differences between PV curves, the latter can be modelled accurately. For the mean PV curve, values for VM, Vm and k1, k2, k3 were 110% VC, -4.34% VC, 260 cm H2O/% VC, 50.5 cm H2O/% VC and 3.13 cm H2O respectively. Unlike previously proposed models, the above includes data below FRC. It describes the truly linear portion of the PV curve at and above FRC. The lower inflection point is accomodated at different lung volumes. When used in a compartmental analysis of a homogeneous lung exposed to a constant pleural pressure gradient, it predicts sequential emptying of dependent and nondependent lung regions consistent with that observed experimentally.