Cohesive energies of crystals

We comment on the controversies that seem to exist about the experimental value of the cohesive energy for silicon. We argue that this can be safely taken to be 4.62±0.08 eV/atom. We conclude that, whereas in the case of diamond the cohesive energy predicted by recent variational quantum Monte Carlo pseudopotential calculation is in excellent agreement with experiment, for silicon there remains a discrepancy of 0.19 eV/atom. To clarify these points, we mention briefly some fundamental principles of the measurement of cohesive energies and explain how they can be obtained from the thermodynamic tables.