196 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1935 



decreases, but more slowly than those lower down because of the 

 reduced density of the air. Shortly before sunrise on a winter day 

 a radio signal may have to go to a height of 400 kilometers before 

 it is reflected back. In such conditions signals of still shorter wave- 

 length may not find enough free electrons at any height and con- 

 sequently will not be reflected back at all. 



While much of the conductivity of the upper atmosphere is due 

 to electrons formed by the sun's ultraviolet light, the charged par- 

 ticles from the sun, which cause magnetic storms, also form free 

 electrons. The electrons formed by this means are found largely 

 at a height of a little above a hundred kilometers, i. e., at much the 

 same level as the lower layer formed by sunlight. In the middle 

 of the night when ultraviolet sunlight is cut off, a strongly conduct- 

 ing region may suddenly be formed at a height of about 100 kilome- 

 ters at times of magnetic disturbance. 



The charged particles which cause magnetic storms travel from 

 the sun to the earth much more slowly than light, hence it has been 

 possible, at the time of an eclipse, to establish definitely that ultra- 

 violet light and not charged particles is responsible for most of the 

 conductivity regularly present in both the 100 to 150 kilometer and 

 probably also the 200 to 400 kilometer region, in temperate latitudes. 

 In polar regions, as might be expected, the effect of charged parti- 

 cles is much more marked and much more frequent. 



It has been suggested that the electrical conductivity of the up- 

 per atmosphere is dependent to some extent on the weather condi- 

 tions at the earth's surface; also, that thunderstorms play an ap- 

 preciable part. We are not yet, however, in a position to state any- 

 thing very definite about this, and further observations are needed. 



Aurorae. — ^While observations of the aurorae have, as yet, pro- 

 vided little additional information about the state of the upper at- 

 mosphere, they must be mentioned since they are clearly due to the 

 same stream of charged particles which causes magnetic storms. The 

 aurora is very closely related to magnetic storms, bright aurorae 

 being usually seen at times of magnetic disturbance. 



From the work of Professor Stormer in Norway and others, we 

 now know accurately the heights of aurorae. The heights are meas- 

 ured by taking simultaneous photographs of the aurora from two 

 distant stations. It is found that while the tops of the rays may 

 go up to great heights — 400 kilometers or more — there is a much 

 sharper boundary at the bottom, at a height of about 100 kilome- 

 ters. It will be remembered that radio measurements showed that 

 the conductivity due to charged particles was most marked at just 

 about this level. 



