22 THEORY OF COLLOIDAL BEHAVIOR 



The writer has tested this consequence of Donnan's theory 

 for solutions of protein salts separated from water by a collodion 

 membrane, with the result that the theory was completely 

 confirmed. Through these measurements of the membrane 

 potentials the correctness of Donnan's theory was proved beyond 

 doubt. 



It may be pointed out that it is not necessary that the non- 

 diffusible ion be a colloid; it is only necessary that there be a 

 membrane which prevents one ion from diffusing; it is immaterial 

 whether or not this latter ion be a crystalloid or a colloid. If we 

 had a membrane impermeable for a SO 4 ion but permeable for 

 Na and Cl ions, solutions of NaCl and Na 2 SO 4 separated by the 

 membrane would give rise to the Donnan equilibrium, and the 

 Na 2 SO 4 solution would probably resemble a solution of Na pro- 

 teinate in regard to certain features of colloidal behavior, e.g., 

 osmotic pressure and P.D. against water. 



Donnan and his collaborators proved the existence of the 

 inequality of the concentration of the diffusible ions of two salt 

 solutions on the opposite sides of a membrane when one of the 

 ions was not able to diffuse through the membrane. Thus 

 Donnan and Allmand investigated 



"the distribution of potassium chloride between two compartments 

 separated by a copper ferrocyanide diaphragm, one compartment of 

 which contained potassium ferrocyanide (the membrane being imper- 

 meable to the Fe(CN) 6 ion). The higher concentration of potassium 

 chloride on the side free from potassium ferrocyanide, and the relation 

 of this unequal distribution to the concentration of the chloride and 

 ferrocyanide, were experimentally established. The results obtained 

 agreed, in general, with the view of membrane equilibria proposed 

 by Donnan, but a discussion of the distribution data combined with 

 electromotive-force measurements appeared to show that, at all events 

 in the case of a copper ferrocyanide membrane and potassium ferrocy- 

 anide solutions, the phenomena are not so simple as supposed in the 

 theory." 1 



More recently Donnan and Garner 2 investigated the equilib- 

 rium concentration of solutions of Na and K ferrocyanides and 

 of Na and Ca ferrocyanides across a copper ferrocyanide mem- 



1 DONNAN, F. G. and ALLMAND, A. J., J. Chem. Soc., vol. 105, p. 1963, 

 1914. 



2 DONNAN, F. G. and GARNER, W. E., J. Chem. Soc., vol. 115, p. 1313, 1919. 



