The Clausius-Mossotti function for the dielectric constant is expanded in terms of single atom and pair polarizabilities, leading to contributions that depend on both the trace and the anisotropy of the pair-polarizability tensor. The short-range contribution of the anisotropic part to the pair polarizabilities has previously been obtained empirically from light scattering experiments, whereas the trace contribution is now empirically determined by comparison to dielectric experiments. For helium, the short-range trace part agrees well with electronic structure calculations, whereas for argon qualitative agreement is achieved.Optical and electrical experiments on fluids indicate that the polarizability of a dense system of noble gas atoms is not simply the sum of the individual atom polarizabilities, regardless of whether these experiments measure the trace or the anisotropy of the polarizability tensor. Examples of properties that depend on the trace are the dielectric constant, the index of refraction, and the isotropic light scattering; the Kerr effect and the depolarized light scattering depend on the anisotropy. All show deviations from the independent polarizability model which can be accounted for most simply by introducing a pair-wise additive pair polarizability.This pair polarizability can in principle be calculated quantum mechanically as a function of the separation of two atoms, just like the intermolecular potential (1, 2). As for the intermolecular pair potential, the only part of this pair polarizability that is rigorously known (besides the classical part) is the large separation asymptotic term. This term is of the same physical origin as the London dispersion term of the van der Waals attraction and, just like it, decays as A/r6, where A is a constant that measures the correlation between the motion of the electrons in a pair of atoms (3,4). It is a quantum mechanical term, and is labeled QM to distinguish it from the classical dipole-induced dipole (DID) interaction that asymptotically varies with the same power of r. As in the Lennard-Jones potential, we are forced at present to add to the long-range part an empirical term that can be either an exponential or a power law in order to account for the short-range (SR) distortions in the individual polarizabilities due to repulsive interactions between a pair of atoms (5). The constants in this empirical term are deduced by fitting a given type of experiment, and the resulting pair polarizability can then be used to account for other optical and electrical properties. This pair polarizability can also be compared to approximate theoretical predictions.Unfortunately, the experimental consequences of the deviation from simply additive individual polarizabilities are generally small and therefore difficult to measure accurately. This has been established in a series of calculations (6-9) of different properties in which the quantum mechanical contributions to the pair polarizability were neglected. These contributions were taken into accoun...