NO. I WEATHER DOMINATED BY SOLAR CHANGES ABBOT 1 3 



curve A of figure 3. Inasmuch as three of the five periodicities which, 

 combined, yield curve B are determined entirely from the work of 

 1924 to 1930, and the other two are to a large extent thus deter- 

 mined, the part of curve B from 1918 to 1923 may be regarded as 

 if it were a forecast. Its good fit ^ encourages us to expect to see 

 these five jieriodicities continue to hold until 1933, producing the 

 general march of solar variation forecasted in curve I of figure 3. 



In former publications dealing with possible solar periodicities, I 

 was indebted to Dr. D. C. Miller for the use of his harmonic analyzing 

 machine. Two of the periods which I then thought real, namely of 

 about 25 months and 1 1 months, are re-discovered by my present 

 method. I feel better satisfied, however, this time, because there is 

 nothing arbitrary about my present analysis. It does not assume 

 periods not indicated by the observations as does the ordinary method 

 of harmonic analysis, which deals with submultiples of some arbi- 

 trarily assumed period. 



I propose soon to apply a similar method to the individual daily 

 observations, in the hope of discovering shorter periodicities. Thus 

 far I have not gone very far in this line, and will reserve it for a later 

 paper. At present, I will only mention that in the year 1924 there 

 appeared to be continuing periodicities of 45 days and of the eighth 

 part thereof. 5.6 days. These are illustrated in curve H of figure 3. 

 Other periodicities seemed to hold from 2 to 4 months and then 

 disappear. 



So far, I have disclosed in solar radiation continuing periods of 

 approximately | and -'j of the ii^-year sun-spot cycle, and of 1/16, 

 1/36, and 1/50 of the Bruckner cycle of 33 years. Besides these there 

 were periodicities approximating 45 and 5.6 days in the year 1924, of 

 which it is uncertain whether they belong to these families, though 

 they approximate to 1/90 and 1/720 of the ii^-year cycle. 



WEATHER PERIODICITIES 



If, as suggested by the title, weather is governed by solar varia- 

 tion, and if, as has just been shown, the solar variation from 1918 

 to 1930 comprises five definite continuing periodicities, we should 

 expect to find these same periodicities in the weather. 



For data to investigate this point, I took from '" World Weather 

 Records " ' the Washington monthly mean temperatures from 1918 



* Regarding discrepancies of igi8 to 1920, see footnote on page 9. 

 " Smithsonian Misc. Coll., Vol. 79, 1927. 



