For Dieter Seebach: They that go down to the sea in ships, that do business in great waters:These see the works of the Lord, and his wonders in the deep 1 ).Electrostatic and polarization energies for the three known polymorphic crystal structures of 1,4-dichlorobenzene, as well as for one particularly stable virtual crystal structure generated by computer search, were calculated by a new accurate numerical integration method over static molecular charge densities obtained from high level ab initio molecular-orbital calculations. Results are compared with those from standard empirical atom-atom force fields. The new electrostatic energies, which include charge density overlap (penetration) effects, are seen to be much larger than and sometimes of opposite sign to those derived from point-charge models. None of the four polymorphs is substantially more stable than the others on electrostaticenergy grounds. Molecule-molecule electrostatic energies have been calculated for the more important molecular pairs in each of the four structures; trends are found to be very different from those indicated by point-charge energies or by total energies estimated with a parametric atom-atom force field. Conclusions based exclusively on analysis of intermolecular atom contacts and point-charge electrostatics may need to be modified in the light of the new kind of calculation.Introduction. ± Every crystal structure determined by X-ray analysis provides the answer to a question ± what is the molecular structure of the compound? ± and simultaneously poses a new question ± why did the compound choose to crystallize in that particular crystal structure rather than in another? We can describe in precise metrical detail the intermolecular packing that is found to occur in the new crystal structure (or crystal structures, for polymorphism is turning out to be far more common than previously supposed [1]), and there is usually no shortage of ways to −explain× the new structure in terms of local intermolecular interactions. By means of energy calculations with empirical-potential functions, we can usually confirm that the new structure corresponds to a minimum in the calculated energy hypersurface, but that is about as far as our present capabilities can go.For a given molecule of known structure, we still have no good theory that enables us to predict its crystal structure with confidence [2]. There is no deep mystery about this. As far as we can tell, the fundamental physics of supramolecular chemistry are understood. The main problem is that, when molecules are brought into contact, the interaction energy is a complicated function of mutual orientation and distance, with the result that many different packing arrangements of a given molecule can have packing energies within a range of only a few kJ mol À1 . For example, for benzene, 30