22 



SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I46 



times as great as the amplitude of sine curve line F (fig. 11), is in 

 percentage of the solar constant. 



I will now give in figures 19 and 20 two more examples from our 

 study of the period 273/9, or 30-1/3 months, in the precipitation of 

 Buenos Aires, Argentina. Tables 2 and 3 of Figure 19 are from a 

 direct photograph of Jonathan Wexler's electronic computation of 

 Category 1, Divisions 1 and 2, for the period of 30-1/3 months.** 

 Taking first Division 1, Mrs. Hill's graph (fig. 20) from Wexler's 

 means discloses an obvious overriding period of 273/18. Removing 

 this as usual, the curve remaining discloses 273/27. After this I 

 removed 273/45, then 273/63. This left a nearly smooth sine curve, 

 273/9, of about 30 percent amplitude, with minimum at the fifth, and 

 maximum at the twentieth month. 



* As shown on figure 19 we arbitrarily cut down precipitation when very high 

 to 20O percent normal. 



PRECIPITATION 

 BUENOS AIRES, ARG. FRACTION TAELE 30 AND 1/ 3 MO. 



DIVISJOM 1 CATEGORV , 



TABLE 2 



DIVISION 2_. CATEGORY 1 



TABLE 3 



Fig. 19. — Electronic tabulation of precipitation at Buenos Aires, 30J-month period. 



