8 



SMITHSONIAN MISCELLANEOUS COLLECTIONS 



VOL. 66 



tions of 1907 to the wave lengths employed in 191 3 and 1914. This 

 was accomplished by plotting the tabular values on a* very large scale 

 and interpolating for the proper wave length. The 67 observations 

 of 1914 which are here represented were obtained between June 12 

 and October 5. Other observations of 1914 are not as yet reduced. 



Table 6. — Comparison of distribution values of other years with results of 1913 



The results just given are indicated graphically in figure 4. Ordi- 

 nates are ratios of the mean ordinates of drift curves for 1907 and 

 1914 compared to those of 191 3 taken as standards. Abscissae are 

 positions along the solar radius, starting from zero at the center of 

 the disk and running to i.ooo at the limb. 



It appears that greater contrast prevailed in the years 1907 and 

 1914 than that which prevailed at sun-spot minimum in 1913 ; because 

 when all the curves are reduced to equality at the center of the solar 

 disk, they indicate lower values at the limb in the years 1907 and 

 1914 than in 1913. The differences are exceedingly small, although 

 so unmistakable. Thus the average departure at 92 per cent on the 

 radius is but i per cent in 1907 and but 0.36 per cent in 1914. 



We have considered whether these dift'erences are due to experi- 

 mental error, but three considerations seem to us to oppose this view. 

 First, no significant change in the apparatus or methods of experi- 

 ment occurred between 1913 and 1914. Second, the fluctuations of 

 contrast from day to day are correlated with fluctuations of the 

 solar constant from day to day, as will appear below from very 

 numerous cases occurring in 1913 and 1914. This being so, it is 

 reasonable to expect also a fluctuation of contrast from year to year, 



