1923. No. lO. DISTRIBUTION OF MATTER IN THE ATMOSPHERE. T7 



height increased upwards in such a way that there should be a fairly 

 moderate fall of density between 120 and 600 km. 



If this hypothsis were at all possible, we might e.xpect the temperature 

 to increase towards the equator and thus perhaps get some increase of the 

 height of the atmosphere towards lower latitudes, which the auroral height 

 measurements indicate. 



Without entering further into details with regard to the calculation of 

 the pressure at various height intervals, we can conclude that the amount 

 of Nitrogon above 100 km. would be very much increased as compared 

 with the density calculated on the assumption of a constant temperature of 

 220" above 10 km.' This "high temperature hypothesis" would then imply 

 that only electron rays could combine a sufficient poietrating power with a 

 sufficient magnetic deflcctibility to explain the auroral forms. — // would 

 further lead to the green line having to be ascribed to Nitrogen because its 

 intensity as compared with that of the Nitrogen lines changes so little with 

 the height. 



On the hypothesis of an electric atmosphere we are left \erv much 

 more uncertain regarding thise points. — In the auroral region the matter 

 does not follow the gas laws and if the Nitrogen is forming clusters neither 

 the dift'erential equation (5) nor the equations 111 and (3l can be applied 

 for this part of the atmosphere. There is no equilibrium at all, but the 

 distribution of matter is determined by a continual process. — /// this case, 

 however, it seems Just possible that the a/nount of matter to be traversed by 

 the cosmic electric rays might be so small that also rays with carriers of 

 atomic order might be possible. 



It seems that we should have to choose between these two possibilities, 

 either that there must above a certain level be an increasing temperature 

 upwards, or that the upper laver of the atmosphere must be electrically 

 charged and most likely consist of electrically charged dust particles or 

 clusters, which at any rate to a very large extent, consist of Nitrogen. 



The latter assumption seems to me at the present time to be the most 

 probable and the most fertile one, because it gives us a possibility of 

 explaining a number of facts in connection with the light emitted from the 

 auroræ (type of spectrum, change of colour), and to bring a number of 

 cosmic phenomena under the same point of view (Northern lights. Zodiacal 

 light, variation of the height of the atmosphere, the blue sky, and perhaps 

 stellar neb u læ). 



If we assume a high temperature, the auroral light should be emitted 

 when ordinary Nitrogen or air at a temperature of say 800 Centigrade is 

 bombarded bv cathode ravs. — Now the spectrum of Nitrogen and air 

 has not been examined under exactly these conditions, but similar conditions 

 might well be reproduced in laboratory experiments, but if Nitrogen or air 



' See the first communication. 



