Some studies on a difunctional, diglycidyl ether of Bisphenol-A (DGEBA) type epoxy resin, viz., Epon 828 of shell cured with different amounts of a tetrafunctional curing agent, viz., metaphenylene diamine (mPDA) [Fig. l(b)] had that the sample cured with stoichiometric amount of curing agent, i.e., 14.5 parts of mPDA by weight per hundred parts of resin (phr), had the highest crosslink density and the highest glass transition temperature (Tg). However, this sample had the smallest bulk density at room temperature compared to the densities of the nonstoichiometric samples studied. It was suggested that, among other reasons, this could arise because the crosslink site did nokprovide a suitable environment for close packing of molecules and, therefore, when a sample was cooled from the final curing temperature to room temperature, the molecular packing in the sample with the highest crosslink density was relatively poor. Support for this observation has also come from wide angle X-ray scattering studies on these sample^,^ which showed that the average intermolecular distance was the largest in the most highly crosslinked samples. Poor molecular packing in highly crosslinked samples cured with stoichiometric amounts of curing agent in relation to those cured with nonstoichiometric amounts has been inferred from their low bulk density in other epoxide systems Recently, Bellenger et a1.6 have published the results of their calculations of the Van der Waals Volumes (V,) and packing coefficients for one of the nonstoichiometric samples reported in the earlier studies,'.' viz., the sample cured with 10 phr of mPDA and also for the stoichiometric sample (14.5 phr). They found on analysis of the data for these two samples that the packing density increases with crosslink density and concluded that the bulk density variations cannot therefore be directly assimilated by variations in the degree of molecular packing as they could arise from the changes in V, as a result of the epoxide-amine addition reaction. The purpose of the present communication is to show that the above analysis is very sensitive to the values of V , used by the authors.V , values for the network were calculated by Bellenger et al.6 from the group contribution values given by Bondi in his book: which suggests a value of 3.7 cm3/mol for oxygen. In an earlier publication,8 Bondi discussed the approximations involved in calculating V, which result from the assumptions that have to be made in obtaining bond distances, bond angles, and contact distances. As regards the group contribution of oxygen, Bondi has emphasized the lack of good contact distance data. He has further stated that as a result of being too deeply buried within the molecule for frequent (or any) collision with neighboring molecules, the "effective volume" occupied by ether oxygen differs from molecule to molecule. In ethylene oxide, on the other hand, the ether oxygen is sufficiently exposed to yield a value of V, = 5.5 m3/mol. In alkyl ethers, especially polyethers, V, = 3.7 cm3/mol and i...