H2 PHYSICAL CHEMISTRY 



Bernstein ("Elektrobiologie," p. 162) supposes electroendosmose 

 to be the explanation of his observations on negative osmose 

 through copper ferroeyanide membranes, although he records 

 no observations on the electric charge of the membranes or the 

 emf. On filling the osmometer with K 4 FeCy 6 solution and im- 

 mersing it in CuS0 4 solution of slightly higher osmotic pressure 

 (calculated from the freezing point) he observed a rise in the 

 manometer, in some cases overflowing at 34 cm. In order to 

 eliminate any possibility of error from the unequal temperature 

 coefficients for the osmotic pressures of the two solutions, he 

 repeated the experiments at 0°, and yet the manometer over- 

 flowed. The emf of these membranes as measured by Briinings 

 are given below. 



Comparison of the Polarisation of Various Types of Membranes 

 The behavior, mentioned above, of the copper ferroeyanide 

 membrane may seem surprising, because, of all dead membranes 

 that have been investigated, it has proved to be the most truly 

 semipermeable in osmotic experiments. Briinings (1907) using 

 isotonic solutions of CuS0 4 and K 4 FeCy 6 observed that the 

 CuS0 4 side is electropositive. He supposed the emf to be due 

 to the permeability of the membrane to K ions, and its being im- 

 permeable to the other ions. If this were the case, the emf should 

 be proportional to the logarithm of the concentration of K ions 

 (which is not true, as the emf remained very near 100 mv with 

 large variations of K ion concentration). It has been shown by 

 H. N. Morse (1914) that these membranes rapidly deteriorate 

 when in contact with other ions than those from which they are 

 formed, and for this reason it may be impossible to decide this 

 question. When we remember, however, that Morse observed 

 that the membrane may be greatly thickened by the passage of 

 an electric current, it becomes clear that the membrane cannot 

 be absolutely impermeable to copper and ferroeyanide ions until 

 its electric conductivity becomes practically zero. Evidently, 

 during the formation of the membrane (without the aid of the 

 current) negative charges are removed from the ferroeyanide 

 solution and positive charges are removed from the copper solu- 

 tion, which would produce an emf opposite to that found by 

 Briinings if the membrane is impermeable to all ions except 

 copper and ferroeyanide ions. 



