The magnitude of the correction in the fifth column of Table III may be open to some doubt, as are all corrections of such a character, and the significance of the above experiment in the author's mind lies not so much in the actual magnitude of the values given in the last column of this table as in their comparative magnitudes. For this reason the entire experiment reported was performed in a single session using the same gelatin solution, so that, whatever the magnitude of the correction, it would be the same in all cases.
Actually the results in the case of the acid titrations are in fair agreement with those of Hitchcock (8). In the present experiment it is seen that, within the limits of experimental error, one gets the same value for the number of cc. of tenth normal acid bound by 1 gm. of gelatin whether one titrates with the acid or with the gelatin. In the case of the base there is a small difference, due probably to carbon dioxide, but this effect is in a direction opposite to that which one would expect on the assumption that it is due to appreciable adsorption.
From this it is concluded that the binding due to adsorption in the case of gelatin is not significant compared to that due to chemical neutralization. The author realizes that gelatin is a poor choice for a basis of generalizations, and similar work is at present in progress on various other proteins. He does feel, however, that the conclusions of Hoffman and Gortner from their work on the prolamines may also be too widely generalized, and that, on the whole, the acid or alkali bound by adsorption in the case of proteins will not constitute the large majority of the total amounts bound, though certainly one will expect a certain amount of such binding in all cases. It also seems that before placing undue emphasis on the conclusions of these workers the possibilities of equivocal results due to specific technique should be considered. This technique consisted in introducing weighed amounts of dry protein into a definite volume of standard acid or base at the equilibrium temperature, in general, and, "after about 15 minutes, during which time the flask was shaken several times," determining the pH of the equilibrium solution. Is it possible that the actual speed of solution of the protean is such that, even though reproducible results are obtained using identical technique, actual equilibrium conditions are approached only when comparatively high concentrations of acid or alkali are employed, in which cases the solution velocity of the protein may he expected to be greater, other factors remaining constant?