106 



Prof. J. J. Thomson. 



[Mar. 9, 



tory, which transforms up in the ratio of 400 to 1 ; an alternating 

 current of about 35 amperes, making 80 alternations per second, was 

 sent through the primary of this. A current of the gas under examina- 

 tion entered the discharge tube through the tube C, placed underneath 

 the arc, and blew the gas in the neighbourhood of the arc against the 

 platinum electrode E,- which was connected to one quadrant of an 

 electrometer, the other quadrant of which was connected to earth. 

 To screen the electrode E from external electrical influence, it was 

 enclosed in a platinum tube, D, the end of which was made of fine 

 platinum wire gauze, which, though it served as a screen for electro- 

 static action, yet allowed the gases in the neighbourhood of the arc to- 

 pass through it. This tube was connected to earth. After passing 

 out of the tube, the electrode E was attached to one end of a gutta- 

 percha-covered wire wound round with tin-foil connected to the 

 earth ; the other end of this wire was connected to the electrometer. 



The experiments were of the following kind : — The quadrants of 

 the electrometer were charged up by a battery ; the connexion with 

 the battery was then broken, and the rate of leak observed. When 

 the arc was not passing the insulation was practically perfect, the 

 spot of light reflected from the mirror of the electrometer hardly 

 moving appreciably in the course of three minutes. As soon, how- 

 ever, as the arc was started, and for as long as it continued, the insu- 

 lation of the gas surrounding E in many cases completely gave way. 

 There were, however, some remarkable exceptions to this, which we 

 now proceed to consider. 



Oxygen. 



We shall begin by considering the case when a well-developed arc 

 passes through oxygen : — 



1. When the electrode E is charged negatively. In this case it lose& 



its charge very rapidly, it does not, however, remain uncharged, 

 but acquires a positive charge, increasing until the electrode E 

 has acquired a potential Y. Y depends greatly on the size of 

 the arc and the proximity of the electrode ; in many of my ex- 

 periments it was from 10 to 12 volts. 



2. When the electrode is charged positively. If the potential is 



very high, the electrode leaks until the potential sinks to Y \ 

 after reaching this potential the leak stops, and the gas seems 

 to insulate as well as when no discharge is passing through 

 it. If the potential to which E is initially raised is less than Y 

 (a particular case being when the electrode is entirely with- 

 out charge to begin with), the positive charge increases until 

 the potential of E rises to Y. 



Thus we see (x) that an electrode immersed in the arc oxygen can 



