1905-6.] Electrolysis through Precipitation Films. 461 
solution in the cold, and consequently contained the violet modi- 
fications of these salts. 
5. The experiment on the temperature coefficient of the cell 
containing ammonium sulphate solution and a chromic hydroxide 
film contains several points of interest. In the first place, the tem- 
perature coefficient is much greater than that of the ammonium 
sulphate solution alone, which, according to Kohlrausch,* is *0215 
for *7 equivalent solutions. 
Change of temperature must, therefore, in some way alter 
either the nature of the film or the nature of the process of 
conduction through the film. That the film is altered by changes 
of temperature is highly probable from a chemical point of view. 
The fact that the temperature coefficient is much smaller at the 
lower temperatures also supports the view that the change is in 
the film itself. 
The slow recovery to the original value for 25° C. is suggestive 
of the same change. It will be noticed that after the cell had 
been cooled to 0° C. it was kept at 25° for about sixty minutes, when 
the conductivity was found to be much higher than its original 
value at 25°. After the cell had remained at constant temperature 
of 25° until the following morning, the conductivity fell to exactly 
the original value. 
6. The experiment given in extenso in the earlier part of the 
paper is only one of many observations made of the effect of 
electrolysis with direct current. When the film had formed by 
the process of diffusion, and no further diminution of conductivity 
was observed, a direct current was sent for a few minutes through 
the cell. The direction of the current was such that the film 
would be reinforced by the meeting of the metallic ions and the 
hydroxyl ions. On disconnecting the direct current, and stirring 
the solutions on each side of the diaphragm by means of the 
electrodes, one invariably noticed a considerable diminution of 
conductivity, but that, on allowing the cell to stand for some time, 
the conductivity very slowly returned to its former value. It is 
not probable that this is simply due to polarisation, though 
further experiments are desirable on this point. It may be caused 
by a heating effect in the film, or by electric endosmosis, the water 
* Kohlrausch und Holborn, Leitvermogen, p. 151, 1898. 
