20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I46 



Fig. 17. — Solar constant and severe magnetic storms. 



weather. An early speaker said he acknowledged the results of long 

 continued studies of total solar variation, which had been made, as 

 probably sound. But the variations found appeared to be of the order 

 of 1 percent, or much less. No reasonable theory could show that 

 these might have important weather influences. He distrusted statis- 

 tical conclusions unless grounded on sound theory. Statistics might 

 show that it is dangerous to go to bed, for a great majority of de- 

 cedents die in bed." The panel appeared to agree unanimously with 

 him, for nobody supported the view that solar variation is an im- 

 portant weather element. 



In Section 4, above, I show that a family of periods, exact har- 

 monics of 273 months, exist in solar variation. In Section 3 I show 

 that the solar constant measurements are abundantly accurate enough 

 to support this fact. In figures 10 and 111 show in many members of 

 this harmonic family, how the shorter periods overlie and confuse the 

 graphs of the longer ones, and how the shorter ones are computed 

 and removed, leaving the longer ones as approximately sine forms. 



I now show in figure 18 exactly similar treatment of the 39-month 



