REPORT OF THE SECRETARY 



107 



(c) Solar variati^m aiid temperature changes. — Obviously the 

 weather depends on tlie sun. If the sun's emission of radiation 

 varies, then the weather must change in some measure on that ac- 

 count. Havinjjf six consecutive years of daily observations of solar 

 variation, made and reduced in the most exact \sx\y at Montezuma, 

 tlie variations have been compared with temperature changes in 

 Washington, Williston, and Yuma. 



Figure 3 shows the sohir radiation measurements at Montezuma 

 since 1924. Satisfactory, nearly satisfactory, and unsatisfactory ob- 

 servations are indicated thereon by circles, crosses, and points, re- 

 spectively. In passing, I draw attention to the facts that the results 



FiGURK 4. — Solar changes and associated teniporaluie chaiigt-s at Wasbington 



average higher in 1924 and 1925 than in 1929, but return to higher 

 values during the summer of 1930. Also the years 1924, 1926, and 

 1028, are more affected by long range variations than 1925, 1927, and 

 1929. This fact tends to verify the 2-year period of solar variation 

 to which I drew attention in last year's report. 



What I now particularly note are the numerous cases of sequences 

 of ascending and descending solar radiation values, occupying about 

 4 days per sequence. These are indicated by curved full and dotted 

 lines respectively in Figure 3. There are 98 cases of ascending and 

 91 ca.ses of descending sequences thus indicated. If it had been pos- 

 sible to observe on all days, there would probably have been nearly 

 twice as many such sequences. I have omitted cases where the 



