6 ' L. vegard: auroral spectra etc. M.-N.Kl. 1923. No. 9. 



The comparison between spectra from different altitudes thus con- 

 firms the view put forward in my first publication, that the green line 

 is due to Nitrog-en and that this gas is predominant to the very limit of 

 the atmosphere. 



Excluding the assumption of a very high temperature above 80 — 100 

 km., our results, together with the fact that auroral rays can be observed 

 reaching a height of 500 — 600 km., involve that in the auroral region 

 the atpiosphere is mainly formed by positive Nitrogen particles. 



A consequence of our results would be that the green line should l)e 

 emitted from positively charged N-ions. Such conditions, however, are 

 difficult to reproduce in laboratory experiments, and as we can only 

 expect to bring a small fraction of the molecules at a time into the 

 ionised state and in this state hit them with electric rays, we can only 

 expect in the laboratory to produce the green line with a very small 

 intensity as compared with that of the other N-lines. At the present 

 time I am making laboratory experiments in this direction. 



Although (|uite formallv we might get a sufficient density of the 

 highest strata of the atmosphere by supposing a proper distribution of 

 temperature, we should still — apart from the improbability of this 

 hypothesis — meet with the difficulty of explaining, why Nitrogen in 

 the auroral region, when bombard by electric rays, should give a 

 spectrum which is so difficult to reproduce in laboratory experiments. 



Indeed we may say that even if we did assume a high temperature 

 we should still have to suppose the gas to be in some abnormal state in 

 order to explain the appearence of new strong lines in the N-spectrum. 



The hypothesis that Nitrogen exists in a highly ionised state, how- 

 ever, at the same time explains the high density and the appearence of 

 lines in the auroral spectrum so difficult to reproduce in laboratory 

 experiments. 



