STUDYING THE SUN 



By C. G. ABBOT, 



Secretary of the Smithsonian Institution, and Director of the 

 Smithsonian Astrophysical Observatory 



For many years we have operated stations on high mountains in 

 distant desert lands making daily observations of the intensity of the 

 sun's rays, on which all life and weather depend. This tedious cam- 

 paign is just reaching its victory. 



As shown in figure i, short-interval changes of solar radiation, 

 taking four or five days in which to produce a rising or a falling 

 sequence of solar change averaging only 0.8 per cent, obviously cause 

 changes in the temperature and barometric pressure at Washington. 1 

 Opposite causes in solar change plainly produce opposite effects in 

 weather. Some of the effects are simultaneous with their solar causes. 

 Others are delayed 10 or more days, probably drifting down in waves 

 from distant centers of direct solar action. These delayed effects, as 

 well as the direct ones, are often large, equalling or exceeding 5 

 Fahrenheit in their average values. 



The conclusion is surprising. Hitherto it has commonly been sup- 

 posed that weather is merely the fluctuation effected by local terres- 

 trial conditions in the orderly periodic march of climate. These new 

 results indicate, on the contrary, that weather is principally caused by 

 frequent interventions of variations of the sun, affecting terrestrial 

 affairs. 



The delayed effects just pointed out give promise of long-range 

 forecasting values. It would be still more promising if regular perio- 

 dicities should be found in the solar variations. This is the case. 

 Figure 2 shows that from 1918 to 1930 the principal variations in the 

 solar radiation, as given by monthly mean values, are well represented 

 as the sum of five periodicities, and five only. They are of 68, 45, 25, 

 11, and 8 months, respectively, and are all closely related to the 1 i^-year 

 sun-spot cycle and the 33-year Bruckner cycle. The search for shorter 

 periods is beginning, and, as curve H shows, reveals a period of 45 days 

 and another one-eighth thereof as having continued through the 



year 1924. 



The expected march of solar variation in the years 1931 and 1932 



is shown in curve I. 



1 And other stations as well. 



