THE AURORA BOREALIS 213 



Lines of auroral and cathode Lines of 



nitrogen spectrum. krypton. Intensity. 



I. II. 



11 4060 4060 4057 8 



12 4260 4260 4274 8 



13 5570 Absent. 5570 10 



The last is the characteristic line of the aurora, and is one 

 of the two brilliant lines of the krypton spectrum; the 

 other brilliant krypton line is in the yellow, and cannot 

 be easily photographed when the light is not bright but 

 flickering, as the auroral light is. 



I ain not able to decide yet whether the lines are all 

 due to krypton or to the cathode spectrum of nitrogen. 

 Certainly there is a striking similarity between the 

 nitrogen spectrum and that of the aurora ; and, on the 

 other hand, the lines of krypton, though sufficiently 

 coincident with those of the aurora to satisfy criticism, 

 leave other bright lines of the krypton spectrum un- 

 accounted for. Yet the cathode spectrum of nitrogen 

 does not contain the line 5570, the most brilliant of the 

 auroral spectrum, and the one most easily discovered by 

 aid of a pocket spectroscope. Experiments on this matter 

 are not yet decisive. 



Moreover, it appears improbable that the aurora should 

 always exhibit only one spectrum. The discharge of 

 electricity through a mixture of gases reveals more or less 

 completely the spectrum of each. Those gases which are 

 present in smallest amount have, as a rule, their spectra 

 proportionately enfeebled. But it does not always happen 

 that all the lines of the spectrum of any one gas are 

 proportionately enfeebled ; sometimes the character of the 

 spectrum itself is altered. The interposition of a Leiden 

 jar and a spark-gap often causes a radical alteration in 

 the spectrum of a gas. This can be well seen with argon ; 

 when the discharge is altered, many of the red lines of the 



