56 SUidies of Diffusion through Rubber Membrancs [Sept. 



lence, and isolated with the least possible alteration of their qualities, 

 if they cotild be removed by dialysis.^ When this idea first came to 

 mind, however, execution of its essential feature appeared to be im- 

 possible. I believed that the diffusion of a solute depends very 

 largely on chemical affinity between the separating membrane and 

 the solvents on both sides of the partition. In that view, it seemed 

 highly improbable that any of the ordinary membranes, except 

 possibly collodion, could be of Service in the dialysis of lipins under 

 any circumstances. Collodion appeared to possess favorable quali- 

 ties because of its solubility in common lipin solvents and its pos- 

 sible affinity for the latter under conditions of dialysis.^ 



Collodion is the only one of the available membranes which, 

 while soluble in ether-alcohol Solutions, freely permits the passage 

 of salins, extractives, carbohydrates, and proteins from aqueous 

 Solutions to water, or to aqueous Solutions, outside, and vice versa, 

 At first thought this suggested special availability of collodion for 

 the work in mind. On the other hand lipins could not be expected 

 to dialyze through collodion in the presence of much water and, as 

 preliminary dehydration seemed an inevitable necessity for the 

 dialysis of lipins from cellular matter, the permeability of collodion 

 membranes to zoater-soluble substances did not appear, after all, to 

 imply any practical advantages for the diffusion of lipins. I also 

 recalled the fact that, in some experiments in another relation, we 

 found that collodion was occasionally rendered defective by ether 

 when the latter was used as a preservative of aqueous Solutions 

 undergoing dialysis.^ 



Continuing actively to consider these matters from one view- 

 point and then another, I thought of rubber as a possible choice of 

 membrane. Recalling the well-known fact that rubber swells very 

 markedly in ether and even in ether vapor, I assumed that the rub- 

 ber expands in ether under such conditions because ether dissolves 

 in the rubber or combines with it. This was but the prelude to the 



' After preliminary desiccation by treatment with anhydrous sodium sulfate 

 or other suitable process. 



° Collodion is a serviceable membrane for such purposes. See page 70. 



* In some experiments which Professor Welker has conducted at my request, 

 we have found that the disintegrative effect of ether on collodion membranes 

 may be due to contained alcohol and other impurities. (See page 70.) 



