IRoebltng jfunfc 



THE WEEKLY PERIOD IN WASHINGTON 

 PRECIPITATION 



By C. G. ABBOT and N. U. McCANDLISH 



Smithsonian Institution 



Golf and tennis players working in Washington, D. C., are some- 

 times disappointed by a succession of rainy Saturday afternoons. So 

 insistently has this forced itself on our attention that we have examined 

 the daily weather record at Washington for the years 1924 to 1939 

 to see if there is any real ground for supposing that there is a marked 

 period of seven days in precipitation. 



Other researches of this Institution have indicated that variations 

 of the sun are important causes of weather changes. As the sun's 

 rotation period is slightly less than 4 weeks, we were inclined to 

 suppose that a period of one-fourth that of solar rotation or slightly 

 less than 1 week, if found in weather, might be associated with the 

 sun's rotation. If it were really caused by a solar variation, and 

 related to the sun's rotation, then the weekly weather period, while 

 nearly constant in length, would probably shift in phase occasionally. 

 For it is well known that solar features, such as sunspots, while they 

 may be nearly fixed on the sun's surface for several rotations, soon 

 disappear and are succeeded by others at other solar longitudes. 



Statistics soon persuaded us that, so far as there is evidence for 

 this nearly weekly weather period, it seemed indeed to be a little 

 less than 7 days. As a first approximation we assumed that the day 

 of maximum precipitation shifts 1 day earlier in the week on each 

 successive month. As a rough and ready index of precipitation we 

 computed for each month separately the percentages of the total 

 monthly precipitation which fell on the seven successive days of the 

 week, Sunday to Saturday. We arranged these monthly percentages 

 in a table in which the week-day names shifted to the left by 1 day 

 each month. 



Taking sums of the seven columns of our table we were then in a 

 position to see whether the day of maximum precipitation did actually 

 remain in the same column, or nearly so, for long intervals. To illus- 

 trate these procedures we give tables 1 and 2. 



As is apparent from table 2. the same column contains the day of 

 maximum precipitation for the yearly average for 4 successive years, 



Smithsonian Miscellaneous Collections, Vol. 98. No. 21 



