4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8o 



These changes in scale of solar constant values have been deter- 

 mined by averaging the differences (given in italics in table i) 

 throughout such intervals as they remained similar in magnitude. 

 These intervals were found to coincide closely with intervals between 

 known changes of procedure, which might have affected the scale 

 of solar constant values. Based upon these facts, and allowing also 

 for small watch errors, the corrections of table 2, in percentages, 

 and in calories per square centimeter per minute, have been deter- 

 mined to reduce all ^Montezuma observations to the scale which pre- 

 vailed from August, 1920, to June, 1921, and which is believed to 

 accord closely ^ with the Mount Wilson scale of 1905 to 1920. 



These corrections of scale depending on changes of procedure, 

 and whose respective influences extend continuously for definite in- 

 tervals of many months, having been applied to the percentage 

 changes of solar constant values given in table i, the resulting curves 

 of solar variation for the twelve months of the year are given in 

 figure 2, both as depending on selected pyrheliometry, and as depend- 

 ing on corrected solar constant values. The agreement between these 

 curves is really extraordinary. It will be seen that a general simi- 

 larity of the curves to those of the sun-spot variation is found, but 

 it is not for all months as close as was found for the months of 

 July, 1910, to 1920. 



As meteorologists in various parts of the country are interested 

 in theoretical and practical studies of the variation of the sun. I 

 have thought best to furnish the following table 2 of monthly mean 

 solar constant values as originally published and now corrected by 

 taking account of the aforesaid changes of scale and of eccentricity 

 of the watches of observers. The table begins with August, 191 8, 

 and ends with December, 1926. These values are not the final defin- 

 itive ones which we shall publish soon, when the laborious recompu- 

 tation of all recent solar constant results is completed, but they will 

 probably differ very little from the final values. They lead to the 

 curve given at the top of the accompanying figure 3. The \A'olfer 

 monthly mean sun-spot numbers for the same interval of years are 

 plotted in the second line. It will be seen that while there is a general 

 tendency for higher solar radiation when sun spots are numerous. 



^ Thus from Monthly Weather Review, May, 1925, it appears that the method 

 of selected pyrheliometry verifies the corrections of solar constant at Mount 

 Wilson of 1919 and 1920. as. given on pages 177 to 180 of Annals of the 

 Astrophysical Observatory, Vol. IV. These citations indicate for Mount Wilson 

 (mean of 100 values of 1918 to 1920, excluding July, 1918, and September, 

 1920) 1.950 calories. Correspondingly, table 2 gives 1.946 calories. 



