SOLAR RADIATION AND ATMOSPHERE — STETSON 



155 



in and year out, his prognostications would be "clear and cool," and 

 his forecasts would be 100 percent correct. At a height of 60 kilome- 

 ters or some 40 miles, the temperature would begin to rise again. 

 Recent investigations give some evidence that at extreme heights, 

 up where the auroral fires plaj^, temperatures of 1,000° C. have 

 to be postulated to account for the presence of the ionized oxygen 

 that is there. The extremely rarefied condition of this upper at- 

 mosphere, however, calls for perhaps a quite different interpretation 

 of temperature than that to which we are ordinarily accustomed 

 when determining temperatures by the thermometer at the earth's 

 surface. 



Ascending through the cross sections of the atmosphere, we find 

 there is a rapid decrease in the amount of atmospheric pressure. 

 Within the first 3 miles from the earth's surface, half the total 



o 10 



5 4 3 

 YEARS BEFORE 





 SUNSPOT 

 MAXIMA 



2 3 4 

 YEARS AFTER 



FiGXIKE 1.- 



-Relative frequeucy of occurrences of aurorae at the Blue Hill Observa- 

 tory before and after years of sunspot maxima. 



amount of oxygen and nitrogen, the principal atmo^-pheric ingredi- 

 ents, are included. The limiting height to which the thinning 

 atmosphere extends is somewhat difficult to fix. Perhaps we should 

 place it at 200 to 300 miles, although recently Dr. Carl Stormer has 

 observed auroral streamers reaching to heights of 600 kilometers 

 or more. ^^Hiere auroral streamers go, some of the thin iiimosphere 

 must extend. 



If we make a chart of the numbers and occurrences of aurorae we 

 find there seems to be a curious connection between the frequency 

 and brightness of auroral displays and the state of the sun as marked 

 by the appearance of sunspots. Professor Brooks, director of the 

 Blue Hill Observatory, has kindly allowed me access to the records 

 made of aurorae at that station for the last 30 years. Utilizing 

 the observations of the brighter aurorae, we may make a graph 

 showing the variations in the auroral frequencies occurring in years 



