1878.] the Spectrum of the Electric Discharge. 



183 



and negative portions of the gas. The first of the above conditions 

 can only result from difference of heat, which is known to produce 

 different spectra from the same gas. I do not think the effects are due 

 to difference of temperature. It is true that the negative electrode is 

 more heated than the positive in the electric discharge in vacuo, but 

 the heat disseminated by it throughout the negative glow produces in 

 its totality but a slight rise in temperature throughout the volume of 

 the negative glow. 



1st. If it be the effect of heat it must be what may be termed mole- 

 cular heat, as the change in the character of the spectrum being com- 

 paratively sudden between the negative and positive light is against 

 the phenomena being caused by change of temperature throughout. 



2ndly. Is it caused by chemical decomposition ? This is possible, 

 but a different chemical effect pervading two definite portions of the 

 electric discharge is a new effect and not to be hastily assumed. I 

 have shown (" Phil. Trans.," 1852) that the electric discharge has an 

 electro-chemical polarity when acting on attenuated gases, the positive 

 terminal producing an oxidating, and the negative a deoxidating effect ; 

 but this effect in my experiments only manifested itself at the terminals, 

 although it may molecularly pervade the gas. 



3rdly. Is it due to electric polarity ? I incline to think it is, but to 

 a polarity so affecting the molecules of the gas, that, if not actually 

 decomposed, they have something like a chemical polarity impressed 

 upon them. This would to some extent favour Mr. Lockyer's view, 

 though not supporting it to its full extent. 



The results may help to explain the phenomena observed in some 

 stars where one or more lines belonging to the spectrum of a given 

 substance is observed, while others are wanting ; and if stars have their 

 atmospheres in a state of electric polarity, as is to some extent the case 

 with this earth, or of electric discharge, as is the case with this earth 

 when the Aurora Borealis or Australis is visible, the spectra would differ 

 more or less from those normally observed here. If the spectrum of 

 the negative light were examined through a series of prisms, there can, 

 I think, be little doubt that the very faint intermediate lines would 

 be obliterated by absorption in passing through the glasses, while the 

 bright lines would remain, and thus the spectrum of a nebula would 

 be presented ; but it would be but a partial representation of the true 

 spectrum, and the line spectrum seen in the nebulas may thus be a 

 partial spectrum. 



P.S. December 23. 



My attention has been called to Mr. De La Rue's paper recently 

 printed in the " Phil. Trans.," which, although he kindly sent me a 

 copy, I had not read when I made the above communication. He finds 



