NO. 4 PERIODIC SOLAR VARIATION — ABBOT 1 5 



group 8 to 12 employs for months the same scale that is used for the 

 decades in groups 24 to 30 and 14 to 21. All the remaining curves, 

 2 to 7, are plotted on a closer scale in abscissae as indicated in figure 3. 



All the curves in figure 3 are based on September 1923 as zero. 

 The decade lo-day means begin with September 5, and the monthly 

 means, used in plots 2 to 12, begin with September 15, 1923. Inas- 

 much as all the curves in figure 3 depend for their forms and ranges 

 on all the Smithsonian solar-constant work from September 1923 

 to December 1952, a synthesis of them all should represent the total 

 monthly variation of the sun during that epoch, if the monthly solar 

 variation is altogether composed of the family of periodicities 

 integrally related to 273 months.* 



I have made such a synthesis from September 1923 to December 

 1935, and present it in figure 4. Except for periods 2 and 3, I used 

 ragged curves of figure 3, not the smooth ones. The dotted curve 

 in figure 4 is my synthesis, and the heavy full curve represents the 

 monthly solar-constant means observed. Except in 1924 and 1925, 

 there is as fair an agreement as the probable error of observation 

 would lead to expect. 



Regarding 1924 and 1925, in chapter 3, pages 33 to 42, volume 6 

 of the Annals of the Astrophysical Observatory, there are detailed 

 the difficulties found in preserving a sound and constant scale of 

 solar-constant measurements. On pages 40 to 42 is told in detail 

 how the results of other cooperating observatories were made to fit 

 with the scale of results of Montezuma, which we considered the 

 best. Several changes of the scale of measurement at Harqua Hala 

 were made in 1924 and 1925 to harmonize with Montezuma. Besides 

 these more or less arbitrary changes, it is related on pages 33 to 35 

 that a defective process of reduction of observations was introduced 

 in 1923. Its erroneous character was not discovered till 1936. A 

 new and correct method was then devised. In order to correct the 

 results obtained from 1923 to 1936, the observers at the field stations 

 were required to remeasure the holographic plates of that interval in 

 order to obtain the data needed in the new method devised in 1936. 

 But parts of the earlier years 1924 and 1925 could not be thus sal- 

 vaged. Some results for those years were therefore recomputed from 

 the observations in the best way possible with data available, but they 

 may not be as good as the work preceding and following. Accord- 

 ingly I do not regard the heavy curve in 1924 and 1925 as being of 



* It is, of course, well known that day-to-day short ups and downs occur 

 in solar variation, and are associated with weather, and with ionospheric 

 fluctuations. 



