94 EEPOBT— 1882. 



at a height of 200 kilometres, it must exceed the oxygen in proportion of 

 170,000 to 1 — that is to say, at that height the atmosphere would practi- 

 cally contain no oxygen. If the gas is less dense than hydrogen this 

 proportion would still further be increased. Even those who do not feel 

 inclined to assume the presence of an unknown gas will not deny the 

 possibility that free hydrogen may be present in our atmosphere in such 

 small quantities as we have assumed. Electric discharges which are 

 • constantly going on partially decompose the aqueous vapour of the 

 atmosphere, and some of it will escape recombination. If then free 

 hydrogen exists even in quantities only which are very small on the 

 surface of the earth, it may be hi preponderating proportion in the upper 

 regions. 



After having considered the above explanation of the fact that some 

 of the meteoric iron may be in the metallic state, I have made a few 

 experiments which tend to show that iron dust may separate from a red- 

 hot meteor even in atmosphere containing considerable quantities of 

 oxygen, and yet escape oxidation. Tissandier has shown how by burning 

 an iron wire in oxygen we may often obtain iron spherules of exactly 

 the same nature as those floating in our atmosphere. I have obtained 

 similar spherules by using an iron file as one pole of a dynamo-machine, 

 and pas.sing the file over a copper wire connected with the other pole. 

 The sparks flying off in all directions are found to consist chiefly of iron 

 globules like those to which Tissandier ascribes a meteoric origin ; but 

 in addition we have small spongy masses which are metallic, and present 

 exactly the same appearance as the metallic iron found in the Saharah 

 desert. The most remarkable fact, however, is this, that we find even a 

 few globules of iron which are metallic. These globules must have been 

 in a state of f asion, and yet they did not oxidise at the contact of the air, 

 no doubt owing to the fact that a large number of particles used up the 

 free oxygen of the air in the neighbourhood of the few particles which 

 thus escaped, 



5. The question of meteoric dust suggests another interesting reflec- 

 tion. Mr. Aitken has recently shown how a condensation of aqueous 

 vapour only takes place round a nucleus of solid matter. It is no doubt 

 one of the most interesting questions to decide what forms in different 

 localities the most common nucleus for fogs, rain, snow, or hail. We 

 conclude this Report by mentioning the very suggestive fact, that 

 Nordenskjold has found iron particles as a nucleus to hailstones at Stock- 

 holm, and that observations of the same kind have been made in Spain, 

 where also hailstones were found to contain iron particles as a nucleus. 

 Other observations of the same nature seem to exist. 



There is every reason to believe that the blue colour of the sky is due 

 to minute particles scattering the light. These particles must be much 

 more minute than any which the Committee has at present investigated ; 

 they must, in fact, be beyond the limits of microscopic power. It might, 

 however, be possible ultimately to find out the nature of these small 

 particles. It is interesting to record the observation made by the author 

 of this Report, that in the valleys of the Himalayas which are cultivated, 

 the colour of the sky is much whiter than in the valleys which are 

 barren and devoid of any vegetation. 



