NO. 17 WASHINGTON AND NEW YORK WEATHER ABBOT 5 



paper. They are based on the corrected period, 6.6485 days, and as- 

 suming January 17.0000, 1946, as the basic date of temperature 

 minimum at Washington. 



In checking the resuhs for 1949 I prepared a talkie of 7 cokunns 

 and 55 Hues. The departures from normal temperatures for dates 



ZSNO^ 30 DEC. S 



10 IS 20 2S 



WASHINGTON 



Fig. 2. — Temperature departures from normal at Washington, December 1949. 

 Dotted lines indicate predicted dates for minima. 



predicted to yield minima of Washington temperatures according to 

 table 3, I entered in the fourth column. In each of the 55 lines I then 

 entered the departures for the 3 preceding and 3 following days, 

 along with the central predicted date. This made up a table of 7 

 columns and 55 lines. In this way 20 days were used twice, owing 

 to overlapping. To obviate this defect, I cut off the temperature de- 

 partures on overlapping dates, alternately from the first and the 

 seventh columns, removing 10 departures from each of these columns. 

 From this table, which to save printing I do not publish, I found the 

 column in each line which carried the minimum temperature for that 



