L. VEGARD. 



H.-F. Kl. 



The heights //' corresponding to 300 ' are given in the second column 

 of table VI. 



Table VI, 



If, now, in the same wa}' as before, we apply the results from our 

 auroral spectral analysis on the distribution of Nitrogen, which corresponds 

 to T' = 300 we get the somewhat, higher values for the possibly pressures 

 of H and He given in table VI. 



Also in this case, however, H and He would predominate above a 

 height of 120 140 km. 



As the ordinary Hydrogen and Helium lines are not to be found in 

 the auroral spectrum, I think we can safely conclude that the green line 

 (5578) cannot originate from any of these two gases, for this line remains 

 the most prominent to the very bottom edge of all ordinary aurorae going 

 down to a height of 95 — 100 km. 



But at this height — as we saw — the pressure of H and He is very 

 small as compared with that of Nitrogen; and remembering that the N-spectrum 

 is very easily excited, it can hardly be assumed that a gas which is only 

 present with a few per cent in the mixture shall give the strongest line. 



On the other hand I have found by spectral observations that the 

 green line is seen to the very top of the auroral ray streamers, and the 

 gas which emits this line must be a prominent component of the atmos- 

 phere up to its extreme upper limits. 



Hence we conclude that the Hydrogen and Helium layer, which has 

 earlier been supposed to dominate at the top of our atmosphere, does 

 not exist. 



For the same reason we can hardly assume the existence of some 

 light unknown gas (Geocoronium) ; for this gas had to be present at a height 

 of 100 km., with a pressure of the same order of magnitude as that of 



