﻿M. F. Zollner on the Spectrum of the Aurora borealis. 123 



to observe with the same spectroscope the spectra of hydrogen, 

 nitrogen, oxygen, and carbonic acid in Geissler's tubes. Yet 

 I believe that I have ascertained, by simultaneous observation 

 of the sodium and lithium spectra, that the red aurora-line does 

 not agree with any maximum of brightness in the spectra of the 

 four gases investigated. It is more refrangible than the red 

 hydrogen-line C ; and its situation would be nearest to the place 

 of the group of dark atmospheric lines a, between C and D in 

 the solar spectrum, to which corresponds a mean wave-length of 

 0*0006279 millim. It may therefore be maintained that the 

 spectrum of the aurora borealis corresponds, in its principal lines, 

 with none of those of known terrestrial substances hitherto ob- 

 served. On the contrary, the observations of Winlock and 

 Young* make it very probable that three green lines of the spec- 

 trum of the protuberances observed during a total eclipse of the 

 sun coincide with three lines in the spectrum of the corona and 

 in that of the aurora borealis. According to the measurements 

 hitherto made, the most refrangible of these lines coincides with 

 one which, according to Kirchhoff's scale, is numbered 1474, 

 and, according to Angstrom, has a wave-length of 0*0005323 

 millim. Kirchhoff and Angstrom agree in designating this line 

 as one belonging to the spectrum of iron. 



All these circumstances seem only to increase the difficulties 

 opposed to a satisfactory explanation of the aurora borealis ; so 

 that Angstrom, in the work above cited, declares the hitherto 

 assumed analogy between the phenomena of the light of the 

 aurora and those presented by the passage of electricity through 

 rarefied air to be disproved by the non-coincidence of the aurora- 

 spectrum with any known spectrum of the atmospheric gases f. 

 I nevertheless believe that, by the following considerations, I 

 shall be able to make the assumption very probable, that, if the 

 luminous developments exhibited in the aurora are indeed, after 

 the analogy of the rarefied gases brought to incandescence in 

 air-exhausted spaces, of an electrical nature, they must belong to 

 so low a temperature that it is impossible at the same temperature 

 to observe the spectra of mean descent gases in Geissler's tubes. 

 This would render possible and, as very simple, also probable 

 the following explanation : — That the spectrum of the aurora 

 borealis does not correspond with any known spectrum of the atmo- 

 spheric gases is only because, though a spectrum of our atmosphere, 



* Silliman's American Journal, Nos, 142 and 143, 1869. 



t Recherches sur le Spectre solaire, p. 41 : — " Moreover, the two phe- 

 nomena (the aurora borealis and terrestrial magnetism) being so intimately 

 connected that the appearance of the former is always accompanied by per- 

 turbations acting on the magnetic needle, it was possible to suppose that 

 the aurora borealis was only a play of electricity, analogous to that produced 

 by the rarefied air in the electric egg ; this, however, is not the case." 



