58 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 3 5 



of a station in the Old World was initiated under a grant from the 

 National Geographic Society. The outfit at Mount St. Katherine, 

 originally employed at Mount Brukkaros, Southwest Africa, is the 

 gift of the National Geographic Society. 



Mr. Roebling also made it possible to send W. H. Hoover with 

 supplies to inspect the stations at Montezuma and Mount St. Kath- 

 erine, and to install there improved pyrheliometric apparatus. The 

 apparatus referred to is a specially constructed pyrheliometer of the 

 Angstrom type. It is to be read as often as possible during bolomet- 

 ric observations, and is to be calibrated daily, or nearly so, against 

 the Abbot silver-disk pyrheliometers heretofore used for daily solar 

 radiation observing. In this way the advantage of the smaller acci- 

 dental error of the Angstrom type instrument will be combined with 

 the long-continued stability of scale of the silver-disk pyrheliometer. 

 It is believed that the accuracy of the daily values of the solar con- 

 stant will be decidedly enhanced by this improved apparatus, and by 

 the revised short method tables. Mr. Hoover visited the Montezuma 

 station in February and March 1934, and will go out to Egypt in 

 September or October 1935. During his stay at Montezuma all parts 

 of the apparatus and methods were rechecked, and several improve- 

 ments were made. 



The new apparatus above referred to was prepared by the observa- 

 tory instrument maker, A. Kramer, and the fine electrical devices 

 therein by L. B. Aldrich. 



PERIODICITIES IN SOLAR VARIATION AND WEATHER 



Studies of the periodicities which superposed make up the varia- 

 tion of the solar radiation were continued by Dr. Abbot, with the 

 assistance, as computer, of Miss L. B. Simpson, under a grant from 

 Mr. Roebling. Using the best available monthly mean values of the 

 solar constant from 1920 to 1934, inclusive, additional periodicities 

 of 9%, 34, 391/2, 92, and 276 months were found in the variation of 

 solar radiation besides the seven formerly discovered of 7, 8, 11, 21, 

 25, 46 and 68 months respectively. All 12 are approximately integral 

 submultiples of 23 years. A synthesis of these 12 periodic variations 

 in the solar radiation was made. The synthesis represents the origi- 

 nal values to within an average deviation of ^%oo of 1 percent. 



Two 2-year forecasts of solar variation were prepared in 1930 and 

 1932, and were approximately verified by the event. The maxima 

 and minima were nearly correctly forecasted as to time, but the 

 curve of observation separated toward the end, as well as in 1932, 

 from the curve of forecast. These defects seem likely to be corrected 

 by the new analysis, and a forecast for 3 years in advance has been 

 ventured. 



