312 J. Trowbridge — Spectra of Hydrogen 



line spectrum of hydrogen alone seemed to be present. 

 When, however, the invisible portion in the violet was photo- 

 graphed, the bands at wave lengths 3900 and 4315 were 

 invariably present, unless the tube had been maintained, during 

 the process of filling, to a temperature of more than 350° 0. 

 After such a process of heating the spectrum became that 

 represented in Plate VI, fig. 2, while before heating it was that 

 shown in Plate YI, fig. 1. In both plates the normal spectrum 

 is above the gaseous spectra. Further toward the ultra-violet, 

 under all conditions, there were also nitrogen bands. Long heat- 

 ing diminished the strength of these bands. This process of 

 experimentation shows that mere eye inspection of glass tubes 

 filled with rarified gases is generally fallacious ; we might con- 

 clude from this eye study that the presence alone of the four 

 lines of hydrogen denoted that we had this gas in a pure 

 state ; whereas the photography of the invisible portion would 

 show that this w T as far from the truth. 



When the glass tubes filled with rarified hydrogen were 

 submitted to the influence of a steady current of electricity, it 

 was found that perfectly pure copper was deposited in a lus- 

 trous state on the glass walls of the tube which surrounded the 

 negative terminal, while an olive-green oxide of copper 

 covered the walls around the positive terminal. The same 

 tube was excited by a Puhmkorf coil, and no difference could 

 be detected in the deposits around both terminals ; they were 

 both rusty green with here and there, it may be, streaks of 

 pure copper. The mirrors produced by a strong steady current 

 at the negative terminal were very lustrous and showed no 

 trace of an oxide of copper. It was evident that the current 

 had dissociated water vapor in the presence of an excess of 

 hydrogen, and had reduced the copper at the negative pole 

 and had set free oxygen at the positive pole, which had, in 

 turn, combined with copper. The rarified gases thus acted 

 like a voltaic cell. 



When we examine the photograph of the discharge repre- 

 sented on Plate VI we see an interesting exhibition of ionization 

 and molinization. The hydrocarbon bands at wave length 4315 

 show a series gradually decreasing in length of w T aves, while 

 another band beginning at wave length 3900, due probably to 

 water vapor, shows a series increasing in length of waves. It 

 would seem that the carbon in one case endeavored to throw 

 off the hydrogen from the hydrocarbon molecule ■; and in the 

 other case the hydrogen became loaded with oxygen molecules. 

 This to and fro ionization and molinization continues until the 

 oxide of copper at the positive terminal has taken up a large 

 share of the oxygen of the water vapor present. There is thus 

 a critical point in the tube at which a sudden increase of resist- 



