«« SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 94 



The result of this investigation indicates that the percentage change 

 of terrestrial temperatures is from 33 to 96 percent of the percentage 

 change of solar radiation involved in corresponding periodicities. 



It might have been supposed that since the earth radiates approxi- 

 mately as a " black body," the relationship would be governed by 



Stefan's law, R = aT^. In that case -^ = ^^, and we should expect 



the percentage temperature ranges to be only 25 percent of the per- 

 centage solar ranges. The actual figures deviate from this in the sense 

 of showing larger temperature ranges than would be expected. Yel: 

 the discrepancy is not so great that one cannot entertain as an explana- 

 tion the contributing influence of indirect causes, such as cloudiness, 

 which might produce changes quite as great as the primary direct 

 cause, variation in solar radiation. 



SUMMARY 



In the foregoing paper I have tried to present within moderate com- 

 pass a general view of an investigation started by Dr. Langley more 

 than half a century ago, carried on in recent years with the indis- 

 pensable financial, intellectual, and moral assistance of Mr. John A. 

 Roebling, the National Geographic Society, and others, and now ap- 

 parently reaching definite conclusions as to the dependence of weather 

 on the variation of the sun. 



I am painfully aware that the limitations of space and funds, the 

 extensive mass of evidence on which I base conclusions, my own inept- 

 ness in its presentation, and the preoccupation of readers with other 

 concerns must all combine to prevent even the most interested of 

 readers from deriving that vivid conviction of the truth and importance 

 of these conclusions which is shared with me by those of my colleagues 

 and friends who are most conversant with the evidence. Nevertheless, 

 I hope I shall not have failed to convince the reader of the following 

 propositions : 



1. The output of radiation of the sun varies, as proved by simul- 

 taneous observations at three stations remote from each other. 



2. The solar variation, seemingly irregular, really comprises 12 or 

 more regular periodicities, which support successful predictions of 

 solar changes for years in advance. 



3. The periodicities in solar variation are integral submultiples of 22, 

 years. 



4. These same and other periodicities which are all integral sub- 

 multiples of 23 years occur in departures from normal temperatures 



