PHENOMENA OF MOTION 57 



cate and is torn by the slightest movement. If we add gelatin to 

 each solution and permit the two salts to diffuse towards each other 

 in the jelly, there is formed at the layer of contact a very resistant 

 membrane supported by the jelly. Expressed generally, if we per- 

 mit two substances which form a precipitate together, to diffuse 

 towards each other within a colloid medium which serves as a sup- 

 port, a membrane is formed at the surface of contact which may 

 have, depending upon the nature of the reacting substances, very 

 different degrees of permeability. 



Since the time of MORITZ TRAUBE,* such membranes have been 

 studied, especially by G. TAMMANN,* W. PFEFFER,** ADIE,* P. 

 WALDEN* and N. PRINGSHEIM.* The chief interest, however, cen- 

 tered in the osmotic phenomena of salt solutions which could be in- 

 vestigated with the aid of such precipitation membranes, whereas the 

 properties of the membranes themselves, with few exceptions, received 

 but secondary attention. For investigations of osmosis the following 

 substances are especially suitable: ferrocyanid of copper and ferro- 

 cyanid of zinc; indeed all ferrocyanid-metal compounds are suitable 

 since they are completely impermeable to many salts. They are 

 briefly described as semipermeable membranes, because they are per- 

 meable to water though impermeable to most crystalloids. If we 

 permit a zinc ferrocyanid membrane to develop in a gelatin jelly 

 and by the addition of potassium ferrocyanid exercise a very great 

 osmotic pressure, the membrane will break in spite of the jelly sup- 

 port, but it will not permit any potassium ferrocyanid solution to 

 diffuse through. 



Besides this extreme case there are membranes of the most differ- 

 ent permeabilities. Following up the work of the botanist N. 

 PRINGSHEIM,* H. BECHHOLD and F. ZiEGLER* 1 exhaustively studied 

 such membranes. They impregnated gelatin with silver nitrate or 

 barium chlorid, and poured the molten solution into test tubes con- 

 taining sodium chlorid or sodium sulphate. At times a layer of 

 pure gelatin was interposed. At the surface of contact membranes 

 of silver chlorid or barium sulphate were formed, which, however, 

 were permeable for the salt solution on either side, because the mem- 

 branes grew in the direction of the greater osmotic pressure, i.e., into the 

 solution with the smaller osmotic pressure. 



If, for example, the silver nitrate solution was more concentrated 

 it diffused through the membrane so that the latter grew into the 

 sodium chlorid gelatin; but if the latter was more concentrated the 

 reverse occurred. If both sides had the same osmotic pressure a very 

 thin membrane formed which was sufficient however to arrest com- 

 pletely the diffusion of both salts. Evidently the meshes of the net- 



