Discussion Table I11 shows that correction for axial dispersion is required for precise determinations of MWD, since, with the correction, the heterodispersity index is within a few percent of the values measured directly by light scattering and osmometry. Table I11 also shows that p values obtained from GPC analysis even after axial dispersion correction are consistently higher than those obtained by absolute methods of measurement. This could result from diffusion of small molecular weight components of the sample across the membrane from the solution side to the solvent side during the osmotic pressure measurements, causing an apparent increase of M , values and consequently a decrease in p.The values of M , and M , determined by GPC are, on the other hand, in error by as much as f10% (and in the case of sample cis-3, by about 23%). It can be shown, however, that an error of only 1% in solvent flow rate in the GPC experiment can introduce an error of about 20% in the absolute values of molecular weight, and at the same time leave the heterodispersity index practically unaffected. We conclude that the error in M , and M , obtained by GPC is thus due to slight inconsistency in flow rate. This implies that the distributions shown in Figures 4-6 are correct distributions of molecular weight, except for a possible small shift along the x axis as a result of random error in flow rate during the GPC experiments.Figures 4-6 (and similar figures not shown) indicate that the Schulz-Zimm is an adequate representation of the MWD for this class of polymers. In fact, it was even an adequate description of the MWD for sample cis-3 which had a p = 3.3. This is especially interesting since the abnormally large p value for this polymer can be attributed to poor temperature control during the polymerization (Vandenberg, 1974).Considering the above it seems reasonable to assume that the MWD of polymers prepared by cationic polymerization can be adequately described by the Schulz-Zimm distribution if p is no greater than 3-4. Whether or not this notion can be extended to polymers with broader distributions of molecular weight remains to be seen. Nazionale delle Richerche, /-40 126 Bologna, and lstifufo "G. Ciamician", 1-40 126 Bologna, lfaly The perchlorylation reaction of benzene and its numerous side reactions is discussed in view of finding a procedure to synthesize the unknown perchlorylbenzene-d5. A method for producing the latter is given. A new apparatus which avoids contact and hence side reactions between a reagent and a product is described.