Even though the calculated values for the lattice energy could not be directly compared with the experimental ones because of the limitations of the approach used, it can be illustrative to compare the values obtained in the different runs when the hydrogen bonds are omitted. When the electrostatic term is not included, values of the lattice energy for rigid molecules (no subrotations) are in the range 70-100 kJ mol -~, whereas when molecular subrotations are relaxed lattice-energy values are in the range 100-160 kJ mo1-1. For runs including the electrostatic term, the lattice energy is in the range 100-170 kJ mol -~ when rigid molecules are considered and 130-200kJ mol -~ when molecular subrotations are relaxed. So, an unexpectedly large Coulombic contribution to the lattice energy, ranging from 30 to 60% of the total estimated lattice energy, is found. In our opinion, this high contribution is a qualitatively significant feature, in spite of the approach and the potential parameters used and the method of estimating effective atomic charges. Generally, it was tacitly assumed that the Coulombic contribution to the lattice energy is negligible but some previous results indicate that this assumption is not valid in all hydrocarbon crystals, reaching a value of 29% of the total lattice energy of benzene and up to 59% in other cases (Williams & Cox, 1984).If the Coulombic contribution to the lattice energy seems to be significant for these crystals, it is true, however, that the inclusion of the electrostatic term does not significantly affect the values of the optimized structural parameters. The overall fit between the theoretical configuration and experimental structure is, in general, slightly more satisfactory for runs including the electrostatic contribution, but conclusive general results on the effects of its inclusion on molecular rotation and translation would be difficult to establish.Within the limits of the approach and the potential functions used, we think that the reported results clearly indicate that the electrostatic term makes an important contribution to the lattice energy of these compounds but has little effect on the optimized structural parameters. A major limitation for accuracy in these calculations, in our opinion, is the description of the hydrogen bonds, but further improvements should be possible by defining a realistic potential function for these interactions.