We present a method for deducing paleoaltitudes that incorporates basic physical principles of atmospheric science and inferences of paleoclimates from plant leaf physiognomy. We exploit the average west-to-east flow of the atmosphere at mid-latitudes and a thermodynamically conserved variable in the atmosphere—moist static energy (the combined internal, latent heat, and gravitational potential energy of moist air)—to develop a method that relies on a parameter that varies with height in the atmosphere in a predictable fashion. Because the surface distribution of moist static energy is constrained by atmospheric dynamics and thermodynamics, the combined internal and latent heat energies, also known as the moist enthalpy, should only vary with altitude provided we know the distribution of moist static energy. Thus, we avoid having to make assumptions about the mean annual temperature lapse rate, which varies spatially and temporally owing to unpredictable variations in atmospheric water vapor. To estimate a paleoaltitude, therefore, we require (1) a priori knowledge of the spatial distribution of moist static energy for the paleoclimate and (2) the ability to estimate paleoenthalpy for two isochronous locations: one at sea level, the other at some unknown elevation. To achieve this, we investigated the spatial distribution of moist static energy for the present-day climate of North America to estimate the deviations in moist static energy from zonal invariance. In parallel, we quantified the relation between moist enthalpy and plant leaf physiognomy of modern forests. Assuming that such deviations from zonal invariance and such relationships between physiognomy and enthalpy apply to ancient climates and fossil leaves, these investigations yield an uncertainty estimate of ±910 m in the paleoaltitude difference between two isochronous fossil assemblage locations.