NO. 7 MONTEZUMA SOLAR-CONSTANT VALUES — ABBOT I3 



the months December to February, when the atmospheric conditions 

 at Montezuma are less favorable, and when many days are lost to 

 observation. It cannot be claimed that the periodicities removed are 

 perfectly correct in forms. Hence the final residuals are larger than 

 they should be on this account also. We may conclude that of the 

 variations of solar radiation indicated in figure i,A, and which ex- 

 ceed I percent in range, accidental error of observation contributes 

 less than 2/10 percent, and the periodic variations nearly i percent 

 of the total range. 



In table 2 I gave the details of the 14 periodicities in the variation 

 of solar radiation which have been discovered. There may be others 

 of less than 5-months period, some of minor amplitude, and still 

 others exceeding 273 months in period, which our observations have 

 not yet continued long enough to discover. Indeed the large fluctua- 

 tions of Great Lakes levels occurring at intervals of about 45 and 

 91 years seem to indicate that the double and quadruple of the master 

 period of 273 months are of very great importance in meteorology. 

 There is also the noted Bruckner period, of about three sunspot 

 cycles, which may also be found eventually in solar-constant values 

 if they continue to be observed for some years longer. 



