96 



ANNUAL REPORT SMITHSONIAN INSTITUTION, 193 5 



fitted within the error of determination with the lengths of the 

 periods given in figure 2. It is not to be supposed that the period of 

 the fundamental solar fluctuation, whatever its cause, has any exact 

 relationship to the year or the month, but it is probably a period of 

 time lying somewhere between 272 and 276 months. 



It is interesting to remember that nearly 30 years ago Dr. George 

 E. Hale discovered magnetism in sun spots. It has been observed 

 ever since. The curious fact has been observed that the magnetism 



I 2 



193.1 



>' 5 rf J? 



o < < < 



2 -) Z 2 



1932 



>^ 2 5 > 



O < < < 



z -) Z 2 



1933 



a > 2 



UJ O < 

 <0 2 -) 



1934 



1935 



PiGURB 3. — Predicted and observed solar variation. The maxima and minima occur in 

 the two curves at nearlj' identical phases. The observed curve may be faulty in 1932 

 owing to the Chilean volcanic eruption. The separation of the curves toward the 

 end is due to a 23-year periodicity not taken account of. 



of sun spots reverses its sign approximately each lll^ years, so that 

 approximately 23 years are required to carry the sun through a com- 

 plete magnetic cycle. Also, it has been well known for over a cen- 

 tury that sun spots fluctuate in prevalence with a period which is 

 irregular, but averages about Wy^ years. The irregularity of the 

 sun-spot cycle is, however, large, ranging from 8 to 16 years, so that 

 the discrepancy between the usually preferred period, 11% years, 

 and the half of 23 years is not perhaps significant. 



As we well know, a violin string, for instance the A string above 

 middle C of the scale, vibrates in a fundamental and many har- 



