156 ROYAL SOCIETY OF CANADA 



and benzene containing copper oleate in solution on the other, no pene- 

 tration of tlie membrane occurred, that is, the colloid copper oleate wliich 

 passed tlirough the i-ubber septum does not pass through parchment. 



Tliese and other results of Kahlenberg's observations make it quite 

 clear that tlie membrane is not a passive element in osmosis. In the 

 •case of rubber septa Avith pyridine only those substances would pass 

 through which are soluble in hydrocarbons, whether they are crystalloids 

 or colloids. " The current view that crystalloids always pass through 

 membranes," he says, "more readily tJian colloids is evidently untenable, 

 as it has been shown that just the opposite may occur and that even 

 crj^alloids may be separated from each other by dialysis when the proper 

 septum is chosen. Whether substances can he separated hy dialysis or 

 not, does not depend at all on their crystalline or non-crystalline nature, 

 as is so commonly supposed, hut upon their affinity for the septum em- 

 ployed." 



Of course, there are in nature no membranes like in composition to 

 rubber, amy more than there are membranes like parchment, or like the 

 precipitated membranes of Traube and Pfeffer, but the one distinctive 

 point obtained from the use of rubber septa which membranes of otlier 

 composiition do not permit us to asjcertain, is that osmosis is due to the 

 solvent activity of the membrane, whatever its nature. From this we 

 may conclude that the composition of the membrane^ and the solubility 

 in that membrane of the solutes on eitlier side of it are all important 

 factors in determining whether osmosis shall obtain. 



With these generalizations and with a new view point, we are in a 

 position to appreciate and understand some problems wliich have had 

 their origin in the results of some investigations of my own on the micro- 

 chemistry of the cell and which, I believe, have a profound significanjce 

 in relation to heredity. 



It is well known that oaseinogen and egg albumen when put in loop 

 of the intestine are absorbed apparently unchanged and excreted by the 

 kidneys. When these siame proteids are intravenously injected they are 

 then also excreted by the kidneys. In the former case the mucous mem- 

 brane of the intestine itself, composed of "colloidal" material, takes up 

 the "colloidal" protedds and pass them on into the blood. In both 

 oases the reflected epitluelium of Bowmfln's capsule, itself also of "col- 

 loidal " composition, allows the foreign proteids to pa.=« through with 

 the water and salts. Again, we know tliat the normal proteids and fats 

 of the blood plasms pass through the endothelial cells of the blood 

 capillaries to provide constituents of ihe lymph. The fats in the Ivraph 

 are destined in part to penetrate tlie cytoplasm of certain connective 



