74 SUidies of Diffusion throngh Rubber Membranes [Sept. 



reacting material in the diffusate was not due to perforations of the 

 bags. The diffusion of biuret-reacting material was always greatest 

 in degree through the bags containing the largest proportion of 

 protein. 



The results of tests 1-15 show that biuret-reacting material dif- 

 fiised through the rubber membranes under the conditions imposed. 

 In Order to determine more definitely, however, whether protein dif- 

 fused through the rubber, we repeated the essential features of tests 

 1-15, but applied additional tests to the diffusates [Table i (16- 

 21)]. 



The results of tests 16-21 (Table i) confirm the findings of tests 

 1-15, and also show definitely that the dififusible biuret-reacting ma- 

 terial was triie protein. 



The data of tests 1-2 1 suggest that osmosis depends upon affin- 

 ities between the membrane, and the solvent or solute, or both. We 

 made a direct test of this mattter in a preliminary way by adding 

 water to the solvent and thus disturbing its affinities with the mem- 

 brane without decreasing the solubility of the solute. The findings 

 are given in the summary pertaining to the eighth series (Table i ). 



The results of tests 22-24 show that water exerted a disturbing 

 osmotic influence and that diffusion of the protein was entirely pre- 

 vented by the water. We extended these experiments to determina- 

 tions of the influence of associated, readily diffusible lipins, in the 

 presence or absence of water. The results are given in tests 25-34. 



In the tenth series the oil in the diffusates was emulsified with 

 a little soap Solution and then repeatedly extracted with ether until 

 all the fat was removed. The water containing the soap, and the 

 aqueous extract of the oil, were evaporated to dryness and the 

 biuret test applied to a concentrated Solution of the residue. 



Experiments with parchment-paper bags. The foregoing re- 

 sults with rubber membranes naturally increased our desire to make 

 comparative observations with bags of parchment and collodion. 

 The results of the tests with parchment are given in tests 35-42, 

 Table 2. That osmosis depends upon accord between the solvent 

 and the membrane is obvious from these results also, for the protein 

 substance, which is readily diffusible through parchment from 

 aqueous Solution, does not dialyze through such a membrane from 

 an alcohol-ether Solution. 



