at Low Pressures through the Passage of Electricity. 177 



To return to the tables, if the values of — , — r^ are taken 



p pAQ 



as coordinates, each table determines a series of points lying 



close to a curve which in each case is nearly rectilinear. 



V * A 



Further, if the values of-As, - A/ (i\, given in the last two 

 pB pAQD' & 



columns of the tables, are taken as abscissas and ordinates 



respectively, the observations can be represented by points 



lying near the same straight line. (See fig. 3.) 



This shows that, within the limits of experimental error, 



— r^Tr^ is a function, linear or very nearly linear, of-^c 

 yj>AQD J J pV' 



In order to ascertain what part the electrodes played in 

 these effects, the zinc plates hitherto used were removed from 

 the bell-jar, and were replaced by commercially pure silver 

 plates. Pure silver is non-oxidizable ; and therefore if any 

 appreciable fraction of the combination of the gases was due 

 to the substance of the electrodes or to oxidation at one or 

 both of them, different results would be obtained when the 

 experiments were repeated with silver plates. > 



These were specially made by Messrs. Johnson & Matthey. 

 The proportion of pure silver in their composition was 

 998 : 999. The upper plate was 3 cm. in diameter, and was 

 surrounded by an ebonite guard-ring flush with the metal. 

 The lower one had a diameter of 8*2 cm. ; but it too was 

 supplied with an ebonite guard-ring fastened on it, and 

 leaving a circular aperture of 3 cm. diameter. Thus the 

 discharge was confined to pass between two circular areas, 

 each of 3 cm. diameter, and unsymmetrical discharges were 

 prevented. 



Tables V., VI., VII. (p. 178) give the results of the experi- 

 ments made in the same manner as before, the silver plates 

 being at the respective distances *4, *8, 1 cm. ! 



As in the previous tables, so also in thesp, the values of 



-ry. lie within a very restricted range of variation. The 



limits may here be taken to be 1*6 and 2. The substantial 



constancy of 4 -r4 is thus, as already said, independent of the 



electrodes. 



As before, if «-^r, *,,— ft are taken as coordinates, each 

 pD p.AQ.D 



of these experiments contributes a point near to the same 



line which was determined by the observations made with the 



zinc electrodes. 



Phil. Mag. S. G. Vol. 9. No. 49. Jan. 1905. N 



