134 Dr. J. Erskine-Murray. 



tested, and the difference found to be O18. The difference had been 

 tested from time to time since that day, the plates having been left 

 in the condenser undisturbed in the intervals. The following table 

 shows the whole series of these results : 



Electric difference between surfaces of a 

 platinum plate in natural condition, 

 and a platinum plate after 45 mins. 

 Time. exposure to dry oxygen gas. 



Dec. 24, 12.30 P.M 0'33 of a volt. 



24,12.47 0-29 



25, noon k 0'18 



27, 0-116 



28, 11.20 A.M 0-097 



31, noon 0-047 



Jan. 4, 11 A.M 0042 



11, 11.40 A.M 0-020 



After detailing some experiments in which the plates "were coated 

 with the gases by electrolysis, Lord Kelvin concludes : " Thus in the 

 case of polarization by oxygen, as well as in the case of polarization 

 by hydrogen, the effect of exposure to the dry gas was considerably 

 greater than the effect of electroplating the platinum with the gas by 

 the electromotive force of one volt." 



The large effects on contact-potential produced by films formed 

 from gases are clearly shown in these experiments of Lord Kelvin's. 

 It is well known that platinum and other metals have the property 

 of occlnding large quantities of gas in their surface layers, and that 

 the condensed gas is possibly in the liquid state, which would account 

 for the similarity between the effects of liquids and of gases on 

 metals. 



56. The plates with which my first experiments on this subject 

 were made were of " high conductivity " copper. They had been 

 polished on fine glass-paper 5 hours before, and their mutual 

 potential had remained constant at 



0-02 volt 



for 4 hours. The lower plate was then put into a glass vessel, into 

 which oxygen gas was admitted from a cylinder, and the oxygen, 

 which was of Messrs. Brin's manufacture, guaranteed 93 95 per cent, 

 oxygen (nitrogen is usually the only impurity), was allowed to stream 

 through the glass vessel containing the plate for some minutes, and 

 the exit and inlet of the vessel were then closed. Forty-five minutes 

 later the plate was taken out and its potential again measured with 

 the other plate, which had remained in air during the interval. 

 Counting time from the moment at which the plate was taken out of 



