NO. 4 PERIODIC SOLAR VARIATION — ABBOT II 



years; (b) with the prevalence of sunspots; (c) with the locaHty; 

 and (d) with changes in the occupation of the land surface. Such 

 phase-alterable agencies in the atmosphere could not, it seems to me, 

 possibly produce periods of unalterable phases in solar-constant 

 measures. 



The fact remains, as yet unexplained, that both the sun's radiation 

 and the elements of the weather are subject to variations in identical 

 families of numerous periods. All are to within i percent exact 

 aliquot parts of 273 months. It would be incredible if they are not 

 associated in their cause. Furthermore, though over 20 of the solar 

 variations range only from 0.05 to 0.21 percent of the solar constant, 

 the corresponding periods in precipitation at St. Louis, Peoria, and 

 Albany range from 5 to 25 percent of normal precipitation. So the 

 percentage range in precipitation is a hundred times the range in solar 

 variation. Here is indeed a paradox worthy of explanation. 



I now give in table 3 a summary of all the periods discovered thus 

 far in solar-constant measures, their ranges, and the range of corre- 

 sponding periods in precipitation at St. Louis. Over 20 of the solar 

 periods, identical with those I have used in weather studies, have 

 been cleared of superriding periods, as shown among the total of 26 

 assembled in figure 3, and are well determined to be real. The other 

 solar periods given in table 3 are the superriding periods which have 

 been cleared away from the forms of periods shown in figure 3. They 

 also are real, but their ranges and forms are only roughly estimated 

 from the tabulations. In figure 3 the numbers attached indicate how 

 many times the period shown will repeat in 273 months. 



In table 3, line i gives the number of times periods of length given 

 in line 2 will repeat in 273 months. In line 3 the values, called S, 

 give the ranges of these periods in percentage of 1.946 calories per 

 square centimeter per minute. Line 4, called L, gives the ranges of 

 identical periods in percentages of normal precipitation at St. Louis, 

 Mo. Line 5 and the following lines give the ratio numbers and ap- 

 proximate percentages of 1.946 calories of superriding periodicities 

 found associated with periods in line 2. 



Details regarding the period of 273 months are omitted in table 3 

 for lack of repeating data. I tested every submultiple of it up to 21, 

 then kept on with only those periods that were used in weather studies. 

 In all, 31 periods were investigated. Those marked with asterisks 

 were found to differ slightly from exact integral submultiples of 273 

 months, when their lengths were adjusted to best fit the observations 

 of the solar constant. The periods with ratio values, 13, 16, 17, 19, 20, 

 and 26 proved to have such small amplitudes as to be negligible. 



