﻿226 Mr. J. A. Brown on the Potential required 



used being indicated in each case. One set was taken above 

 and one below the critical pressure, which, with the sparking 

 distance used, through the whole course of the investigation 

 was always approximately at l'l millimetre. In both cases 

 the potential to maintain a definite current will be seen to be 

 lower as the number of cells is increased. The drop is not a 

 constant one however^ as the lines tend to come closer with 

 fairly large total potentials, very little decrease being noticed 

 when the number of cells is increased beyond that indicated 

 by Curve 5 (fig. 10). From a priori reasoning it would 

 seem to follow that when external resistance and electromotive 

 force are large, the potential of the electrodes should attain a 

 definite value. This has accordingly been taken as being 

 nearest the true value for a steady current, as it is obvious 

 that a current is most stable when both R and E are large. 



This explains the reason why the sets of curves in Part I. 

 of this investigation were taken in the way that they were,, 

 i. e. keeping a large force on (1040 to 1200 volts), and 

 varying the current by changing the ohmic resistance. It 

 should be mentioned, in passing, that the sparking distance 

 of the zinc and the aluminium was not exactly the same 

 length, a fact that was only to be expected from the way in 

 which the gaps were put together. This will explain any 

 little discrepancy between tables taken with different metals 

 during the latter course of the experiment (but see below). 



B. Effect of Material of Electrodes. 



When the apparatus was first planned it was assumed, on 

 the evidence of previous investigators, that the material of 

 the electrodes would have no effect on the sparking potential, 

 and consequently on the potential required to maintain the 

 current. The electrodes were accordingly made of zinc, as 

 it happened to be the most convenient metal on hand. It 

 was found, however, that the zinc tended to form a film on 

 the glass walls of the sparking-chamber, the spark-length in 

 this case being only 0*5 centimetre, and so to give a metallic 

 path for the current. It did not, however, make a stable 

 connexion, so that definite determinations for permanent cor- 

 rection could not be taken as at first hoped. It was in trying 

 to obviate this that the effect on the sparking potential due to 

 the electrodes was first noticed. 



Another gap with aluminium electrodes was put in to do 

 away with the disintegration observed with the zinc. It 

 was soon noticed, however, that the two sets of electrodes 

 did not behave in the same way. After a fairly large current 



