112 



AlfNUAL REPORT SMITHSONIAN INSTITUTION, 193 5 



17 I show such forecasts made only one year in advance, and cor- 

 rected to the actual event at the beginning of each year. In both 

 figures the forecasts v^ere made without knowledge of the events, 

 bemg based solely on a study of what had happened in the preceding 

 46 years. 



Similar forecasts have been made for over 30 stations in all 

 parts of the United States for the years 1934, 1935, and 1936, but 

 not published. Such sensationally important disclosures require 

 as yet more long-continued verification of their success before it 

 would be prudent or wise to make them public. However, the year 



1922 1923 



NORTH PLATTE, NEB. 



DERM?TURES FROM NORMAL TEMPERATURE 



(5-MOHTH RUNNING MEANS) 



= OBSERVED = PREDICTED 



FiGUEB 17. 



1934 having elapsed, a comparison of the predictions with the events 

 has been made. The results have been divided into four classes. 

 Excellent, Good, Half and half, and Bad. To illustrate this classi- 

 fication, I show in figure 18 a fair sample of each class both as to 

 temperature and precipitation. The comparison resulted in the fol- 

 lowing classifications : 



A. Temperature. 



Excellent, 7: Eastport, Key West, Detroit, Salt Lake City, Helena, Port- 

 land, San Diego. 



Fair, 17: Albany, New York, Washington, Hatteras, Mobile, Nashville, 

 Cincinnati, Chicago, St. Paul, St, Louis, Omaha, Bismarck, Cheyenne, 

 Denver, Santa Fe, Red Bluff, Spokane. 



Half and half, 3 : New Haven, Galveston, North Platte. 



Bad, 4 : Charleston, Little Rock, Abilene, San Francisco. 



B. Precipitation. 



Excellent, 11: Eastport, Burlington, New York, Detroit, Chicago, Duluth, 

 St. Paul, St. Louis, Little Rock, North Platte, Bismarck. 



