The polar caps of Mars have long been acknowledged to be composed of unknown proportions of water ice, solid CO2 (dry ice), and dust. Gravity and topography data are here analyzed over the southern cap to place constraints on its density, and hence composition. Using a localized spectral analysis combined with a lithospheric flexure model of ice cap loading, the best fit density of the volatile-rich south polar layered deposits is found to be 1271 kg m−3 with 1-σ limits of 1166 and 1391 kg m−3. The best fit elastic thickness of this geologically young deposit is 140 km, though any value greater than 102 km can fit the observations. The best fit density implies that about 55% dry ice by volume could be sequestered in these deposits if they were completely dust free. Alternatively, if these deposits were completely free of solid CO2, the dust content would be constrained to lie between about 14 and 28% by volume. The bulk thermal conductivity of the polar cap is not significantly affected by these maximum allowable concentrations of dust. However, even if a moderate quantity of solid CO2 were present as horizontal layers, the bulk thermal conductivity of the polar cap would be significantly reduced. Reasonable estimates of the present day heat flow of Mars predict that dry ice beneath the thicker portions of the south polar cap would have melted. Depending on the quantity of solid CO2 in these deposits today, it is even possible that water ice could melt where the cap is thickest. If independent estimates for either the dust or CO2 content of the south polar cap could be obtained, and if radar sounding data could determine whether this polar cap is presently experiencing basal melting or not, it would be possible to use these observations to place tight constraints on the present day heat flow of Mars.