Mr. F. S. Spiers on Contact Electricity. 8t 



tube. When the pressure was -±- mm. the P.D. was 



-0*33 volt. 



I then again let in a little hydrogen, 1 mm. merely ; the P.D. 

 immediately became relatively more positive, it being 



-0-54 volt. (Cp. § 12 above.) 



On again exhausting, the P.D. fell ; when the pressure was 

 T ^— mm. the reading was 



-0-42 volt, 

 and at -J— mm. 



300 



-0-35 volt, 



just as before, at about the same pressure. I again let in a 

 little hydrogen, this time 3 mm. ; the P.D. became 



-0-48 volt. 



I then let in dry air, until the pressure was 1 atmosphere. 

 The P.D. then moved slowly in the positive direction. The 

 air was admitted at 6.35 p.m. 



At 6.47 the P.D. was -0'43 volt. 



At 7.0 it was -0'40 volt. 



At 7.20 -0-37 volt. 



The next morning at 10.30 it was 



-0-17 volt. 



This very slow return to the positive value, or, rather, in 

 the positive direction, is just, what one would have expected 

 on the assumption that the Yolta effect is a function of the 

 medium surrounding the gases. Consider for a moment the 

 state of the surfaces of the metals just at this juncture. The 

 whole object of the experiment was to remove, to the utmost 

 extent possible, the original condensed surface-layers of air 

 from the plates, and to replace them by layers of hydrogen. 

 From the method that was adopted, aud from the results that 

 followed, it is almost certain that this state was actually 

 arrived' at. Then the true contact P.D. between iron and 

 platinum in an atmosphere of hydrogen and hydrogen only 

 (neglecting for the moment the probably inactive nitrogen) is 

 about —0*60 volt (platinum being positive to iron). The 

 fact that this value was only reached long after the metals 

 had cooled from a red heat in an atmosphere of hydrogen 

 is easily explicable. The whole or greater part of the Volta 

 effect is due to the layers of gas in immediate and close contact 

 with the metal surfaces. Naturally, when the metals were hot, 



