I'JOG.] 



on Electric Discharge and Spectroscopy. 



105 



which the hiiuinosity round the cathode appears. We see tliat by a very 

 small increase in the potential difference the dischar«je passes from a 

 state in which no Inminosity can be detected, even in a dark room, to one 

 where the Inminosity can plainly be seen in a bright light ; thns the 

 molecnles of the gas in the tnbe just when the luminous discharge 

 is on the point of appearing, are in a state in which a very small 

 change in the electrical conditions of the tube make the molecules 

 pass from a state in which they are not giving out an appreciable 

 amount of light to one where they are brightly luminous ; and, as the 

 great increase of the current when the luminosity appears shows, this 

 change in state is accompanied by an emission of corpuscles. From 

 this and other considerations, I have come to the conclusion that 

 what takes place when the gas becomes luminous is that the internal 

 energy in the atom in consequence of its bombardment by the 

 corpuscles increases, and when it gets up to a certain critical value the 

 equilibrium of the atom becomes unstable : an explosion occurs, result- 

 ing in an expulsion of corpuscles and such a shaking up of those left 

 in the atom that these vibrate so vigorously that the energy radiated 

 is sufficient to produce luminosity. Thus, I regard the ionization of 



Discharge 



Fig. 2. 



50 Mh 



the gas as being due, not to the corpuscles in the atom being dragged 

 out by the direct action of the electric forces in the field, or as being 

 knocked out by a rapidly moving corpuscle striking against them, but 

 to an explosion due to the atom having absorbed so much internal 



2 



