NOMOS. 33 



same time understanding the extremely philosophical 

 views of this great chemist upon the constitution of 

 all salts. It was also natural for Dr. Graham to 

 conclude that chemical action was essential to the 

 presence of electricity in a metal, because electricity 

 and magnetism are inseparably united ; and being so, 

 the conclusion is, not that a metal may be the seat 

 of chemical changes during the presence of electricity 

 in it, but that it must be. 



The cogency of this reasoning depends, of course, 

 upon the existence of these intimate relations be- 

 tween electricity and magnetism. The evidence is 

 only indirect, and it cannot be otherwise at present. 

 Still, enough has been said to allow us to assume, 

 with some show of reason, that the metallic parts of 

 the circuit, like the fluid parts, may be the seat of 

 chemical changes during the presence of the current. 



When air enters into the circuit there is also 

 some reason for believing that it is the seat of de- 

 finite chemical changes during the pas- 

 sage of the current. We might argue as * 

 much from the photographic powers of 

 the electric spark, though these powers 

 are only manifested collaterally, and out 

 of the immediate track of the current. 

 The fact, however, that the discharge is thus marked 

 by chemical functions is an argument that during 

 the discharge the air is the scene of active chemical 

 changes. Of itself, perhaps, the fact would be of 

 little moment, but coupled with what we know of 

 the condition of the fluid parts of the circuit, and 



D 





