XO. 2 



SOLAR CHANGES ABBOT 



figure 3. It is this : The resiihs given in table 2 show a strongly 

 marked periodicity of 25:^ months. I mentioned this discovery to 

 Dr. Dayton C. Miller. x\t his invitation, I submitted, for harmonic 

 analysis and synthesis by his celebrated machines, yj successive 

 months of solar constant results from June, 1920, to October, 1926. 

 These he used as they are given in table 2, except that smoothed 

 curves are drawn in some months of few observations. Dr. Miller's 

 work is graphically shown in figure 5. The dotted curve is that which 



Fig. 3. — Monthly mean solar constant values, August, 1918, to December, 

 1926; sun-spot numbers; and indications of approximate 26-months regular 

 periodicity in solar radiation. 



I supplied. The full curve above it is synthesized from the first 30 

 harmonic components of it as determined by means of his machine. 

 The first and second components are of little interest, as they give 

 merely the efifort of the machine to represent the ii-year sun-spot 

 cycle with only 67 years of data. Periods of yj /z, 77/S' 77 /^^ 77/7' 

 77/9> 77 1'^^' 77 1^'^' 77 1 ^A a-"cl 77 1'i-S months, however, stand out with 

 more or less distinctness. By far the strongest of them is the one of 

 77 /Z or 25fi months, but it seems to be associated with " overtones " 

 (to borrow an expression from sound) of \, \, \, and \ its period. 

 This fundamental period of nearly 2 years and 2 months has been 



