﻿the Rays producing Aurora Borealis. 235 



spectrum, however, we must remember that the aurorse are 

 produced under circumstances very different from those 

 present in the previous experiments. Thus the solar rays 

 have a much greater density and penetrate through rarified 

 gases and perhaps also fine dust particles, and these dif- 

 ferences might account for the greater number of lines 

 observed in the auroral spectrum and the appearance of the 

 auroral line A,= 5570. 



The spectrum of the phosphorescent light produced by 

 a-rays falling on certain substances has been examined by 

 E. Marsden *. Zinc sulphide gave a single band extending 

 from \ = 5920 to \ = 4250, and willemite a band between 

 \ = 5800 and \ = 4870. The auroral line falls in the middle 

 of these bands. 



The spectral analysis at its present state may indeed furnish 

 us with certain suggestive points ; but it does not give an 

 experiment am crucis for the determination of the nature of 

 those rays which produce aurora borealis. We shall have to 

 wait for further experiments ; but already the previous ex- 

 periments have shown that a- rays (compared with the other 

 radium rays) are very effective in producing luminosity in 

 the air through which they pass, a property which is very 

 essential if rays of this type are to explain the brilliancy of 

 aurorse. 



Rays consisting of atoms or molecules might not only 

 make the matter through which they pass luminous, but by 

 their impact they might themselves give out light. If so, 

 we should have a means of investigating the nature of the 

 carrier by observing the Doppler effect in the auroral 

 spectrum of those lines which are produced by the radiation 

 icselr. 



The spectrum analysis of the light produced by a-rays, 

 however, cave no trace of helium lines. Thus in this case 

 the light given out from the carrier is either wanting or 

 very feeble as compared with that produced in the surround- 

 ing air. 



The difficulty in detecting a possible spectrum from the 

 carrier is made even greater through the Doppler effect, 

 which for the very large velocities would amount to about 

 one-tenth of the total wave-length. The light from the 

 auroras entering the spectroscope will usually correspond to 

 directions forming the most different angles with the direc- 

 tion of the radiation; consequently the light corresponding 

 to a certain line from the carrier would be spread out to a 



* E. Marsden, Proc. Roy. Soc. A. Ixxxiii. p. 04S (1910), 

 R 2 



