The quantity most directly related to the net charge or valence of corpuscular proteins is the electrophoretic mobility. In 1932 Abramson (1) showed that in solutions of the same ionic strength the electric mobility of the same protein at different hydrogen ion activities is directly proportional to the number of hydrogen ions bound or released as determined by titration curves. His comparisons were made with gelatin, serum albumin, egg albumin, and casein. The generality of Abramson's rule has been confirmed by Daniel (2), Moyer and Abels (3), Moyer and Abramson (4), Longsworth (5), and also by Carman, Palmer, and Kibrick (6). Longsworth's data on egg albumin are especially interesting, for his comparison extends from pH 3 to 12. Inasmuch as the results of Abramson and of Moyer seem to have been misunderstood, it should be emphasized that they deal with comparisons of data on dissolved proteins. If the protein in the adsorbed state happens to agree, it is incidental to the argument.Not only did Abramson show proportionality between the electric mobility and titration curves, but he also attempted to calculate the valence of proteins from the available electrophorefic and titration data. The agreement, however, was incomplete.Although acid-base titration and membrane potential measurements on protein solutions give results that are related to the electric mobility, the actual determination of the net charge from these measurements is subject to difficulties in interpretation and questionable assumptions. On the other hand, an exact interpretation of electric mobility data in terms of net charge requires independent information concerning the size and shape of the protein molecule and the sizes of the other ions in the solution. Information regarding size and shape is not yet available for any protein. Extreme limits of variation of these quantities, however, may be obtained from sedimentation and diffusion data by the assumption, on the one hand, of spherical shape and enough hydration, or, on the other, of enough asymmetry, to account for the diffusion constant.The results presented in this communication indicate that when these factors are taken into account, values for the valence calculated from titration, electro-* This work was aided, in part, by a grant made to one of us (L. S. M.) by the