the Luminosity of Gases, 247 



and that this is subsequently reduced by the carbonaceous 

 matter of the flame. Again, it may be asked how is it that 

 copper chloride, which is known to lose its chlorine at quite 

 moderate temperatures, can not only be vaporized but be 

 made incandescent in a flame, and in a flame too which con- 

 tains reducing gases ? 



The solution of the difficulties here presented is not so 

 obvious that it should be withheld, yet I have been unable in 

 searching the literature of spectrum analysis to find any dis- 

 cussion of these fundamental questions. 



With regard to the luminosity of " non-luminous "" gas- 

 flames and the luminosity of gases in Geissler-tubes the 

 difficulty has been to some extent grappled with, and as such 

 flames are the media in which colorations are obtained it will 

 be convenient to discuss them first. 



Attempts to produce Luminosity in Gases by External 

 Heating. — The consideration of the luminosity of flames free 

 from solid matter, such as the flame of hydrogen or the 

 ordinary Bunsen gas-flame, has been held to involve the 

 question as to whether gases can be made luminous simply by 

 increasing their temperature. The luminosity of gases through 

 which the electric discharge is passed and of those in a flame 

 may be directly due, it is alleged, in the first case to the 

 electricity, and in the second to the chemical processes. 



At the end of last century Wedgewood (Phil. Trans. 1792, 

 p. 271) performed an experiment which proved that a current 

 of air sufficiently hot to raise a strip of gold to bright red- 

 ness possessed no luminosity. 



In 1879 Hittorf (Wied. Ann. vii. pp. 587, 591, and ibid. 

 xix. p. 73, 1883) found that on passing an electric current 

 from a battery consisting of 1600 elements through air at a 

 pressure above 15 millim. the platinum electrodes became 

 white-hot, yet the air in their immediate neighbourhood 

 showed no luminosity even when the platinum was melting. 

 He found the same to be the case when pure iridium electrodes 

 were employed. These when melting communicated no 

 luminosity to nitrogen, hydrogen, or carbon monoxide. He 

 concluded that the luminosity and line-spectra of gases could 

 only be obtained at much higher temperatures developed by 

 the momentary discharge of a condenser. He adds: — u The 

 light of ordinary flames, which do not contain solid particles, 

 is not determined by the temperature but by chemical pro- 

 cesses and must be regarded as phosphorescence. For if these 

 gases acquire the temperature of the flame, without being 

 involved in the chemical process, they radiate no light per- 

 ceptible to the eye. One can convince oneself of this by 



S2 



