216 ESSAYS BIOGEAPHICAL AND CHEMICAL 



longer visible when the pressure falls below 0*000035 

 millimetre the pressure observed when, in a mixture of 

 krypton and helium, the green line of krypton became 

 very faint and almost invisible. Neglecting the influence 

 of temperature, the pressure of the atmosphere can be 

 made to give its height by the formula 



H = 18-382 (log. B-log. b) kilometres. 

 Substituting for B (barometer) its normal height, 760 

 millimetres, and for b the pressure of the krypton, 0-000035 

 millimetre, we have height = 135 kilometres, or about 84 

 miles. This number is reasonably near the figures given 

 by Cavendish and others. Professor Birkeland, the latest 

 authority, it will be remembered, thinks the altitude is 

 from 100 to 200 kilometres, 1 or from 62*5 to 125 miles. 



We may next ask Since the spectrum of krypton is 

 so persistent, why is it not visible in air ? The answer is 

 Because the presence of nitrogen renders it invisible, 

 for it is not possible to distinguish less than one part of 

 krypton by volume in 7100 parts of air. But krypton 

 does not show its spectrum in argon, which may be said 

 to constitute about 1 per cent, of the volume of air. Now 

 7100 parts of air will yield about 70 parts of argon, and it 

 should be possible to distinguish the krypton line if the 

 amount of krypton present wee O'Ol part, or 1 part in 

 7000. This would give, for the proportion of krypton 

 in air, 1 part in 700,000. Recent experiments, have 

 shown that it is possible to extract 1 part of krypton 

 from about 7,000,000 of air ; and of xenon, which, owing 

 to its lower vapour pressure, can be extracted from air 

 with more ease than krypton, there is only about 1 part 

 in 40,000,000 of air. It is therefore clear why the spec- 

 trum of krypton is not visible in that of crude argon. 



We come next to the question Is there any reason 

 to believe that krypton may concentrate in the higher 



1 Expedition Norv&gienne, Christiania, 1901, p. 28. 



