Electrometric titration measurements of serum albumin B (Kekwick) have been determined at several ionic strengths. These results, together with the available data in the literature for serum albumin B, egg albumin, and ß-lactoglobulin have been used to compare values for the valence calculated from measurements of titration, electrophoresis, and membrane potentials. The results indicate that the usual interpretation of titration curves is open to serious question. By extrapolation of the titration data to zero ionic strength and protein concentration, there results an "intrinsic" net charge curve describing the binding of H⁺ (OH^-) ion alone. This curve agrees closely, in each case, with values of the valence calculated from mobility data (which in turn are in close accord with those estimated from membrane potential measurements). The experimental titration curves in the presence of appreciable quantities of ions and protein deviate widely from the ideal curve. It is suggested that, under these conditions, binding of undissociated acid (base) leads to erroneous values for the net charge. This binding would not affect the electrophoretic mobility.

Values of the net charge obtained by the two extreme models from electrophoretic data are in agreement within 15 to 20 per cent. The agreement between the cylindrical model and the titration data is somewhat better in each case than with the sphere; i.e., this comparison enables a choice to be made between asymmetry and hydration in the interpretation of results from sedimentation and diffusion measurements on proteins. It is concluded that the proteins discussed here are somewhat asymmetric and also hydrated.