8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I43 



This recommends my forecasts to those whose interests demand fore- 

 knowledge of wide departures from the normal. A second recom- 

 mendation arises from the fact, which may be verified by inspection of 

 figures 1, 2, 10, and n of Publication 4390, that large trends in de- 

 partures from normal precipitation, continuing over several years, 

 may be tolerably forecasted. Third, as far as time has elapsed, results 

 remain encouraging. In figure 4 and table 1 I give evidence that dur- 

 ing the 24 months of 1959 and i960, three years after November 1956, 

 (the latest record used in the forecast) a strong correlation (about 50 

 percent) between forecast and event was observed. 



In preparing temperature forecasts or precipitation forecasts, as 

 explained in my paper "A Long-Range Forecast of United States 

 Precipitation," it is necessary to compute new monthly normals from 

 the published records. These new normals relate respectively to the 

 years when the Wolf sunspot numbers are greater, and to the years 

 when they are less than 20. 



To illustrate the need for taking account of sunspot frequency in 

 normals of weather records, and to show interesting features of the 

 difference between stations in their relations to sunspot frequency, I 

 give figure 5. It shows the difference in percentage of normal temper- 

 ature between times of sunspots less and more than 20 Wolf numbers 

 for the 12 months of the year. There is a partial similarity among the 

 graphs, but Salt Lake City and Los Angeles behave rather differ- 

 ently from the other eight. It will be noted that ranges of 2° or even 

 3 F. occur for some cities, which are ignored in usual monthly nor- 

 mals where sunspot frequency is neglected. 



The following dates may be used to separate SS > 20 and SS < 20 

 groups of months: 



Table 2. — Intervals of high and lozv sunspot numbers 2 



SS>20 



2 Table 2 and remarks thereon, unfortunately omitted from Publ. 4390, are 

 here quoted from "Solar Energy," vol. 2, No. 1, June 1958. 



