REPORT OF THE SECRETARY 109 



or more of the stations, as dictated by the blue-violet spectrum. He 

 then took the general mean for each day, not only of the untreated 

 results, taking into account only the grades assigned by Messrs. Aid- 

 rich and Hoover for the separate stations, but also an "improved pre- 

 ferred" mean for perhaps one-fourth of all the days. These new 

 means were the results preferred after considering the blue-violet spec- 

 trum. Both of these daily means were entered in the great table, so 

 that when it is published, readers may use either the preferred general 

 mean or the "improved preferred" general mean, as they please. 



As the great table was thus being finished in manuscript, it was being 

 typewritten by Miss M. A. Neill in preparation for the printer. By 

 the end of the fiscal year it was almost finished for publication. In 

 the meantime the rest of the manuscript for volume 6 of the Amials 

 had been finished as far as possible by Dr. Abbot and typed by Miss 

 Neill. But some changes and additions will be made after the inspec- 

 tion of the great table is completed. There appears every reason to 

 hope that the entire manuscript of volume 6, including the great table 

 and its subsidiaries, tables of 10-day and monthly means, will be in 

 the printer's hands before New Year's Day. 



The study of the great table led Dr. Abbot to reconsider whether 

 the sun's variation might not be more effectively followed by observa- 

 tions limited to the blue-violet region of spectriun. He was at length 

 able to devise a method which appears promising, and which has been 

 introduced just at the end of the fiscal year at all three field stations. 

 In brief, the method contemplates inserting in front of the spectro- 

 bolometer slit a glass filter which restricts the radiation to the desired 

 blue-violet region. An exactly similar glass filter is inserted before 

 the aperture of the pyrheliometer. Knowing from the usual solar- 

 constant work of the day the atmospheric transmission coeflScients for 

 blue-violet rays, it is possible to compute the extra-atmospheric energy 

 spectrum of the restricted blue-violet spectrum given by the screened 

 spectrobolometer. A comparison of the blue-violet energy spectra at 

 the station and as computed for outside the atmosphere gives a factor 

 to multiply the screened pyrheliometer reading to what it would be 

 outside the atmosphere. 



In this way we restrict the observations to the most variable part 

 of the observed solar spectrum, and avoid those spectral regions where 

 ozone, water vapor, and extreme short and long wave lengths introduce 

 great errors. We greatly hope that this new method will yield more 

 reliable daily indications of the solar variation. 



The necessary instrumental changes for introducing the new method 

 were done by A. Kramer. He has also prepared special apparatus 

 for solar distillation of sea water after Dr. Abbot's design, and many 

 other required small jobs for the Observatory. 



