NO. 3 



SIXTY-YEAR WEATHER FORECASTS — ABBOT 



II 



computed from all the records without regard to the sunspot fre- 

 quency prevailing. At Peoria the rainfall is about 7 percent higher 

 at sunspot maximum than at sunspot minimum. If a long term of 

 years of monthly records is being used to compute such a period as 

 7 months, mentioned above, the tabulation at Peoria would be thrown 

 into confusion, if published normals, taking no account of this, were 

 used. My first step therefore is to compute new normals. I have 

 chosen to use 20 Wolf sunspot numbers as the dividing line between 

 high and low sunspot prevalence. Table 3 gives the normals for 

 Peoria, Albany, Washington, and St. Louis according to my revisions. 



Wind OlRCcriON 

 -Syv vy NW. 



Fig. 7. — How Washington, D. C, temperature depends on wind direction. Tem- 

 peratures Fahrenheit, from about 1,500 consecutive days. 



Second, phases depend on the season of the year.^ 

 Terrestrial effects from solar causes lag behind the actuating causes. 

 Thus the hottest part of summer falls many weeks after the June 

 solstice, and the warmest part of the day occurs several hours after 

 noon, at most stations. Lags for my periods depend, in the first in- 

 stance, on the season of the year. It is not possible to fully compen- 

 sate for this. As a compromise measure, I divide the year into three 

 parts — January to April ; May to August ; September to December. 

 Third, phases depend on sunspot frequency. 



s I published this observation in the year 1944 (see Smithsonian Misc. Coll., 

 vol. 104, No. 5, p. 27). 



