6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I48 



measures as given above. And as each harmonic period is the mean of 

 many determinations, their amplitudes have probable errors far too 

 small to be due to chance. 



2. PERIODIC TERRESTRIAL PRECIPITATION 



Convinced 20 years ago of existing solar periodic variation, I 

 noticed in the long monthly records of Peoria, 111., an indication of a 

 period of about 23 years in precipitation. I then tried during several 

 years to find other periods there. They appeared to exist, but to suffer 

 variable displacements associated with changes in atmospheric trans- 

 parency. Persevering, ways avoiding these difficulties were found, 

 leading at length to a method of long-range forecasting which I used 

 with surprising success on the records of St. Louis. See Smithsonian 

 Publication 4545, figures 21, 22, and 25, and explanatory text. 



Prof. Wexler, of State University of Arizona, interested his son, 

 Jonathan Wexler, in my forecasting. Jon. Wexler, a student in 

 electronic programming, saw how the computing might be greatly 

 aided. Since 1955, he has prepared for my forecasts the long rec- 

 ords of precipitation of 55 stations in all parts of the world. I have 

 described in the publication "Solar Energy," volume 1, No. 1, 1956, 

 and volume 2, No. 1, 1957, the tedious process used for my forecasts. 

 Smithsonian Publication 4390 gives forecasts, 1950 to 1967, for 

 32 stations in the United States. But while over 5,000 copies of 

 it were sold, besides 1,500 copies freely distributed by the Institu- 

 tion, professional meteorologists are still skeptical, notwithstanding 

 evidences of useful value given in pages 1 to 6 of Smithsonian Publi- 

 cation 4471. Hence it seems good to present now, in detail, evidence 

 that the identical family of harmonic periods found in solar variation 

 is also present strongly in terrestrial precipitation, and may be used 

 for long-range forecasting. 



Figure 4 plots the forms and amplitudes of 26 periods in the precipi- 

 tation of Rochester, N.Y., 1884-1955, as computed by Jon. Wexler in 



273 

 222 separate tabulations. Period —:r- months is omitted from figure 4 



273 

 because repetitions are too few for good evidence. Period —r- is 



found in evidence at Rochester only as represented by the shorter 



273 273 273 



over-riding periods —^ and -y^- Yet —r— is strong itself in St. Louis 



precipitation. See Publication 4545, page 32, and at other stations. 



