ON MKTEOEIC DUST. 93 



In addition to this we also hear of larger meteors passing through the 

 air more slowly, but leaving behind trains of luminous clouds remaining 

 visible for a considerable length of time. What becomes of all this 

 matter, and how should we expect it to look after it has settled down on 

 the surface of our earth ? Tissandier has examined microscopically some 

 powder which he has detached from the surface of a meteoric stone found 

 in Bohemia, and it was found to resemble in appearance the magnetic 

 particles found in different places to which he had attributed a meteoric 

 origin. 



4. An interesting question arises in connection with the iron particles 

 which are found in the metallic state in the sand of the desert of Sahara 

 and in other localities. How did they escape oxidation, either when first 

 detached from the molten mass of the meteoric dust, or subsequently. 

 Several explanations of this fact may be offered. If the particles are 

 really meteoric they would contain a considerable proportion of nickel, 

 and such iron is able to resist oxidation to a high degree. But the pre- 

 sence of nickel would not prevent their transformation into magnetic 

 oxide when red hot, as they must have been when they separated from 

 the meteor. I wish, however, to point out that possibly, and even pro- 

 bably, the separation has taken place at a height at which the atmosphere 

 contains only comparatively small quantities of oxygen. It is known by 

 the laws of diffusion that, assuming everything to be in a state of equili- 

 brium, each gas will form an atmosphere round the terrestrial globe 

 independently of any other gas which may be present. It follows that at 

 great heights the lighter gases will be present in preponderating propor- 

 tions, as compared with the lower regions. Calculating, for instance, the 

 proportion of oxygen which we should expect at different heights if the 

 temperature is the same throughout, we find as follows : — 



At a height of kilometres 21 per cent, of oxygen. 

 ,, 5 ,, 19'5 ,, ,, 



„ 10 „ 18 „ „ 



„ 100 „ 4-2 



,, 150 „ 2 ,, „ 



„ 200 „ 0-8 



Convection currents would, no doubt, especially within the lower 

 regions, tend to equalise the difference between the higher and lower 

 parts, but the fact must remain that, at a height at which luminous, 

 meteors have been seen luminous, the oxygen can only form a small 

 fraction of the atmosphere. But matters become still more striking if 

 we consider that probably some other and lighter gas is present in ad- 

 dition to the oxygen and nitrogen. 



Spectroscopic observations of the Aurora Borealis show the presence 

 of a green line which no one has as yet obtained from any known constituent 

 of the atmosphere. I have myself observed nitrogen, and oxygen and 

 some of the carbon compounds, under so many different conditions that 

 I am fully convinced that the line is not due to them, but must be due to 

 the presence of some unknown gaseous body. As we cannot detect it 

 near the surface of the earth, this gas must be very light. Supposing, for 

 the sake of argument, that it is as dense as hydrogen, and that at the surface 

 of the earth its quantity per cubic centimetre is only the millionth part 

 of the oxygen present in the same space, it would certainly escape all our 

 methods of analysis. But from these suppositions we can calculate that, 



