20 



SMITHSONIAN MISCELLANEOUS COLLECTIONS 



VOL. "JJ 



Figure 12 shows a direct comparison of JMontezuma and Harqua 

 Hala from 1920 to 1924, Over 300 days common to both stations 

 have been arranged in 16 groups of graduaJly increasing mean solar- 

 constant vahies, as indicated by Montezuma observations. These 

 identical groups of days' results were also averaged for Harqua 

 Hala. Of course, in this way the range shown at Harqua Hala must 



\S^0 1.91 1.92 \.93 1.94 1.95 1.96 



Covnymri^o-n of 339 c/di/5 ./920 -/924. 



Fig. 12. — Correlation between Montezuma and Harqua Hala on solar 



variation. 



necessarily be less than that shown at Montezuma, because some of 

 the extreme Montezuma values will be extreme on account of error 

 of observation, and will not be extreme at Harqua Hala. I have, 

 therefore, given Harqua Hala a more open scale so as to incline the 

 line at 45°. The correlation is obvious. 



Thesis (c). — Observed changes in solar radiation are clearly asso- 

 ciated with visible changes in the sun. 



Figure 13 shows a comparison of Wolf sun-spot numbers with 

 all of our thousands of solar-constant values obtained from 1905 to 



