NO. 7 ASTROPHYSICAL OBSERVATORY, I904-I953 — ABBOT 9 



variation, with an amplitude of about ^ percent of the solar constant, 

 appears as the mean result of four well-observed repetitions. 



This new evidence and that of 1915 (Pub. 4545, fig. 15) seems to 

 me to show decisively the 27-day rotation period in solar variation. 



I have referred to the apparent 27-day effect on Washington pre- 

 cipitation at pages 45 and 46, and table 8 of Pub. 4545.^ Using 27.0074 

 days as the exact period, I found good success in predicting the 



1955 



1.935 



Fig. 1. — The 27.006-day period in solar variation. 



175 days more apt to have rain in Washington than the other 190 or 

 191 days of the years 1942-1954. (See table 8, Pub. 4545.^) The 

 formula failed only in 1952 and 1953 (possibly because of bomb 

 explosions and fallout). Continuing to use the same formula in pre- 

 diction for the year 1963, the whole group of preferred days proved 

 displaced, recurring one day too early. But if the period is slightly 

 changed to 27.0056 days it would have made no difference in fore- 

 casts prior to 1954, but in 1963 the preferred days (one day earlier) 

 would have had 1.30 times the average precipitation falling in all 

 others. 



