A GROUP OF SOLAR CHANGES 

 Bv C. G. ABBOT 



The author recently pubhshed ^ a new method of testing the varia- 

 bihty of the sun. Heretofore such variation has been indicated only 

 by successive observational values of the solar constant of radiation. 

 The new method depends on the selection of moments when the sun 

 is equally high above the horizon, the atmosphere equally clear, the 

 quantity of atmospheric water vapor identical, and the month of the 

 year the same, so that the temperature conditions will be substantially 

 comparable, both around the recording instrument and in the atmos- 

 phere itself. 



Under such circumstances, if they could be met ideally, the atmos- 

 phere, although it reduces the intensity of the sun's radiation, reduces 

 it in the same proportion on every chosen occasion. Accordingly, the 

 pyrheliometric measurements of total solar radiation made at such 

 moments should show the same percentage variations of the sun as 

 the solar constant observations, in which atmospheric influences are, 

 as we suppose, eliminated. Since the most critical selection must 

 admit some inequality in sky conditions, the new method is not appli- 

 cable to individual days, but gives good results only for means of 

 fairly numerous groups of days, such as occur in the course of a 

 month of observdng. 



The new test of comparative pyrheliometry on selected days was 

 applied to the observations of the months of July from, 1910 to 

 1920,^ omitting the years 191 2 and 191 3 when the great volcanic 

 eruption of Mt. Katmai, Alaska, rendered the atmosphere so hazy 

 that no suitable days could he found for comjDarison. The results 

 are shown in figvire i, in which the single smooth curve represents 

 the selected pyrheliometric observations, the dotted curve represents 

 the hitherto published solar constant work, and the double full curve 

 represents the variation of the Wolfer sun-spot numbers. It will be 

 seen that, except for the year 191 4, the new test is closely confirm- 

 atory of the solar variation shown b\- the published solar constant 

 work, and that there is an exceedingly close correspondence between 



^ See Monthly Weather Review, May. 1926. 



Smithsonian Miscellaneous Collections, Vol. 80. No. 2 



