228 
Proceedings of Royal Society of Edinburgh. [may 16 , 
for the apparatus did not admit of great accuracy, the galvanometer 
being shunted by a wire of very small resistance.) In two or three 
days the resistance was about 200 ohms. Once the plates were left 
for eleven days, when a resistance of about 250 ohms was observed. 
The process, therefore, goes on slowly for a long time. 
Origin of the Resistance. 
Most probably the resistance is due to the condensation, on 
the surface of the electrodes, of gases dissolved in the liquid. To 
test this, I left the plates in air instead of placing them in the liquid. 
The resistance in this case was of the same magnitude as before, 
and was evidently due to the condensation of atmospheric gases. 
To obtain a stronger proof, however, I placed the plates, after being 
heated, in an atmosphere of oxygen for about two hours. A resist- 
ance was found in this case about equal to that which was caused 
by leaving the plates in air for twenty hours. Leaving the plates 
in air for two hours made no appreciable change in resistance. 
Next, I connected the plates to the positive pole of a battery of 
two Bunsen cells for two minutes, and decomposed water with 
them, so that oxygen was developed on them. They were then 
connected together for a quarter of an hour to get rid of polarization, 
and then a resistance was observed equal to that got by leaving them 
in oxygen for two hours. Nearly, but not quite, the same resistance 
occurred if the plates were joined to the negative pole of the 
Bunsens, so as to develop hydrogen on them. There can be no 
doubt, then, but that the resistance is due to condensed films of gas. 
Firmness of the Gaseous Deposit 
That the deposit clings with excessive firmness to the surface of 
the metal, is evident from the fact that nothing but heating to a red 
heat destroyed the resistance. No amount of rubbing of the plates, 
however hard, made any observable diminution. 
Specific Resistance of the Films of Gas. 
In order to determine the specific resistance of the deposit, a 
knowledge of its thickness is necessary. Or, if one can assume 
that the thickness is the same in different specimens of platinum, 
it might be obtained from the above experiments (which can give 
