4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. Ill 



raised the temperatures at the predicted low dates. If so the error 

 would be compensated. But one is inclined to think that this would 

 not always be the case, and that the higher intervening values are 

 higher in the mean than they should be in comparison with the values 

 occurring on the predicted dates. 



I sought a method to estimate the average increases of tempera- 

 ture between the predicted dates which would give too low a mean 

 value of the excess. For this purpose I plotted the daily departures 

 from the normal Washington temperatures from January i to April 

 I, 1946, as shown in the accompanying figure. Then with a scale 

 divided into intervals of 6.6456 days, I marked, as shown by short, 

 heavy vertical lines in the figure, the dates which seemed best to 

 represent the minima of the 6.6456-day periodicity at Washington.^ 

 Inspection indicated that January 17.0000 would be a proper date 

 for the zero phase. From this date I calculated all corresponding 

 phases from January i, 1940, to December 31, 1948. Assuming 

 these 495 dates to be exact dates of minimum temperatures at Wash- 

 ington, I computed yearly tables of temperature departures as ar- 

 ranged in groups averaging 6.6456 days in length. In these yearly 

 tables of 55 lines each, approximately two out of every three lines 

 were 7 days long, and the others 6. To arrange the short lines sym- 

 metrically I wrote 45 lines with the first space unfilled, and the rest 

 with the seventh space unfilled. 



Table i gives the mean yearly marches of the 6.6456-day periodicity 



Table i.—Mcan march of 6.6436-day temperature groups at Washington 

 in the years ig40-ig48 



Mean temperature marches 



Year 



1940 — o?o2 — o?ii — o?8o — 1?05 — o?20 +i?09 +o?34 



1 2.63 2.31 1.76 3.65 3.31 3.18 1.72 



2 0.22 2.16 2.98 2.96 3.49 2.27 2.29 



3 1.84 2.62 2.78 2.80 1.38 1. 00 0.76 



4 0.61 1.49 2.69 2.38 2.85 2. II 1.02 



5 1-86 1.84 2.09 2.13 2.51 2.49 2.24 



(J 0.72 2.16 4.62 4.31 3.36 2.78 2.64 



7 —0.69 2.89 2. II 2.00 2.47 2.38 1.93 



I94S 1.49 2.25 1.93 1.45 1.33 3.44 3.02 



Mean 1.08 2.20 2.62 2.71 2.46 2.59 1.99 



s It used to be the view of most meteorologists, and perhaps still is, that 

 the brief fluctuations of temperature which we associate with the term "weather" 

 are caused by terrestrial complexities. I think the figure shows plainly that 

 they are mainly of solar origin. 



