108 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 3 9 



The author mentions various considerations constantly in his mind but over- 

 looked by critics. He discusses the procedures used in holding the solar 

 constant values to a fixed scale for long-range comparisons. 



The author points out that critics' disparagement of solar variation by the 

 tabular use of standard deviations proves nothing and is merely reasoning in a 

 circle, since the increase of standard deviations may quite as well be caused 

 by real increase in solar variation as by increase of experimental error. Simi- 

 larly, increase of correlation coeflScients between two stations may be due to 

 increased solar variation rather than experimental interdependence as sug- 

 gested by critics, for thereby the competitive effects of accidental error are 

 made relatively less considerable. 



The author explains in detail the discovery of 12 long periods in solar 

 variation. Critics having claimed on the basis of an equation of Brunt's, that 

 the author's periodicities have impossibly great amplitudes, the author tabulates 

 these periodicities completely, and then synthesizes them for the years 1920 to 

 1934 and compares the synthetic curve with observed monthly mean values. 

 With a total range of 2 percent, the synthesized periodicities reproduce the 

 original observations to within an average deviation of 0.2 percent, hence 

 periodicities cannot have excessive amplitudes as claimed. 



The author reiterates his conviction that solar variation is the principal 

 cause of weather changes, and that with annual expenditure of $300,000 for 

 solar observation, principal details of weather might be predicted all over 

 the world from study of solar variation for two weeks in advance. 



Since the publication of Abbot's reply the variations of solar 

 radiation have been discussed from the standpoint of terrestrial mag- 

 netism by F. E. Dixon (Quart. Journ. Koy. Met. Soc, July 1939) 

 who finds: 



There is evidence of a connection between changes of magnetic activity and 

 the regular rises and falls in the solar constant noted by Abbot. The changes 

 are appreciable only on the first two days after alterations of the solar con- 

 stant commence, and, hence, continue only as the solar constant is changing. 



Dixon found in studying the averages for magnetic character fig- 

 ures for each month of the year, over a period of 12 years, that op- 

 posite changes of terrestrial mag-netism occur, according as the solar 

 constant begins to rise or to fall according to Smithsonian 

 observations. 



PUBLICATION 



With N. M. McCandlish, the Director prepared a paper on "The 

 Weekly Period in Washington Precipitation." ^ A discussion of the 

 daily precipitation at Washington, 1924 to 1939, indicates a perio- 

 dicity of 6 days 18 hours, or approximately one- fourth the period of 

 the sun's rotation. The period is such that the day of maximum 

 precipitation in the monthly averages comes one day of the week 

 earlier on each successive month. Unexpected changes of phase by 



* Published shortly after the close of the fiscal year as Smithsonian Misc. Coll., vol. 98, 

 No. 21. 



