during the puesuge of Electric Discharge. 1 7 



increasing effect upon the spectrum of the oxygen in proportion 

 as the latter gas is removed : for this reason two tubes had to 

 be rejected for accurate determinations. In the first one, to 

 which the above remarks have reference, the previously described 

 spectrum of carbonic acid became continually more and more 

 prominent. For this fact, which at first seemed very surprising, 

 a sufficient explanation was soon found. The same exhaustion- 

 apparatus, namely (constructed with mercury), had been employed 

 just before to exhaust carbonic oxide, which here has the same 

 action as carbonic acid (115). In the second tube, traces of ni- 

 trogen were recognizable ; this tube had also to be rejected, be- 

 cause it evidently contained air mixed with the oxygen gas. 



110. In the extinction of the electric current in oxygen gas, I 

 have already (73) recognized a proof that no current can exist in an 

 absolute vacuo, to which we can only approximate practically to 

 a greater or less extent. I may here add that the electric cur- 

 rent in an empty space (if one can exist there) certainly cannot 

 be luminous ; for if this were the case, the spectrum for empty 

 space would necessarily reappear in the different gas-spectra. The 

 latter, however, have nothing in common. 



111. I next endeavoured to determine the spectrum of the 

 oxide of nitrogen (NO^). The attempt, however, was unsuccess- 

 ful ; a tube of binoxide of nitrogen gave the spectrum for ni- 

 trogen itself with a modification evidently attributable to pure 

 oxygen, which was proved by the existence of a bright band near 

 the red, and because the brownish red became a bright red at 

 the boundary of the orange. The narrow characteristic lines in 

 the red, orange, and yellow retained their entire sharpness. The 

 extreme ones alone, which the bright band partly covered, had 

 disappeared, so that only six such lines remained between the 

 new bright band and the boundary of the orange. The brilliant 

 red band was gradually extinguished ; and the result was the 

 formation of the pure spectrum of nitrogen gas of a splendour 

 which I had never before observed. 



Binoxide of nitrogen, present in so small a quantity as to be 

 scai'cely recognizable by the most sensitive balance, was thus che- 

 mically analysed. It was instantly decomposed by the passage 

 of the current into its constituents ; nitrogen and oxygen, and the 

 latter disappeared by gradually forming oxide of platinum with 

 the metal of the negative electrode. 



112. In consequence of the above-observed behaviour, it was 

 scarcely to be doubted that the higher and less stable stages of 

 oxidation of nitrogen would immediately separate into their simple 

 constituents. For confirmation, a tube was filled with nitrous 

 acid (formed by mixing binoxide of nitrogen and oxygen in the 

 right proportion^s) and then exhausted. The spcctrmii was the 



Phil. Mag. S. 4. Vol. 18. No. 1 17. Juhj \Kid. C 



