between Anode and Kathode Spectra. 347 



the dark space reaches to the anode, the second spectrum of 

 hydrogen, and the closely set yellow bands of nitrogen, are seen 

 only on the anode. These spectra are both extremely complicated, 

 whence I infer that the molecules in which they originate are of 

 complicated structure and comparatively massive. The continuous 

 spectra of the halogens at the anode emphasises this point, they 

 are, in a sense, still more complex, and it is noticeable that the 

 brightness of the continuous spectrum is greatest in iodine and 

 least in chlorine, in fact follows the molecular masses. In con- 

 nexion with this- it seems probable that oxygen may . emit some 

 continuous spectrum at the anode, feebler than that of chlorine, 

 and not , easily detected, because not lying in the part of the 

 spectrum most easily seen. 



The kathode glow follows so closely the course of the kathode 

 rays, and the gases behave to them so precisely in the way that a 

 fluorescent liquid behaves to ultra-violet light, that the inference 

 immediately suggests itself that the excitement of luminescence 

 is of the same sort in these two cases, and also in the case of the 

 glass which fluoresces under the rays when the density of the gas 

 is insufficient to absorb them. 



The essential character of fluorescence, as I understand it, is 

 that the fluorescing substance absorbs certain kinds of radiant 

 energy and forthwith emits it again in a more or less modified 

 form as light, and continues to do this as long as the exciting 

 cause continues and for no appreciable time longer, and is not 

 itself permanently changed in constitution or material by the 

 process. 



Percival Lewis' observations 1 have shewn that various metallic 

 vapours exposed to kathode rays, without being in the course of 

 the electric discharge, or forming part of the conducting gas, 

 emit light, of which he has examined the spectra. 



In the case of each of the four metals that I have examined, 

 the rays that he found in its spectrum are all to be seen in the 

 kathode glow of the same metal, though in sodium and cadmium 

 the glow gives a greater number of rays than Lewis observed. 

 But the most convincing argument that the molecules of the gas 

 emitting the kathode glow are unaltered in constitution is that in 

 many cases the spectra are reversible. This could not be unless 

 the molecules which produce the reversed spectra by absorption, 

 were the same as those which under the stimulus from the 

 kathode emit the same rays bright. The line spectrum of 

 hydrogen is well known to be reversed in the sun, and the reversal 

 of G and F has often been observed in looking at a spark in dense 

 hydrogen. Most of the metallic lines also which I have seen in 

 the glow, as well as those Lewis has observed in the kathodo- 



1 Astro-Phys. J., xvi. 31, 



23—2 



