Recent measurements of the constant-volume specific heat of argon (and of oxygen) at critical density and close to the critical point are compared with calculations based on the three-dimensional nearest-neighbor lattice gas models. It is argued that the configurational specific-heat density $C^{*}\mathrm{}\left(T\right)=\frac{\rho C_{\mathrm{config}}\mathrm{}\left(T\right)\mathrm{}}{{\rho }_{max}}$ should be compared with the theoretical configurational specific heat per lattice site. On this basis the calculations above $T_c$ agree to within 10% with the observations for argon and are consistent with a divergence like ${(T-T_c)}^{-\frac{1}{5}}$ over the range $\frac{(T-T_c)}{T_c}={10}^{-1\mathrm{}} \mathrm{to} {10}^{-4\mathrm{}}$. Below $T_c$ the strength of the apparently logarithmic divergence of the specific heat for the fcc lattice agrees to within 10% with the experimental data but the theoretical magnitude of $C^{*}\mathrm{}\left(T\right)\mathrm{}$ is too small for $T>\mathrm{}0.9\mathrm{}{T}_c$ by an almost constant amount of $1.05k$ per site.

• Received 17 July 1964

DOI:https://doi.org/10.1103/PhysRev.136.A1599