14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 85 



to 1923. I supplemented them to 1930 by taking monthly mean values 

 of " Max." plus " Min.," as given in the " Climatological Data." ' In 

 some previous work I had prepared a plot of the average yearly march 

 of Washington mean temperatures. From this smoothed curve I took 

 values corresponding to the 15th day of each month, and subtracted 

 from my monthly mean data. Thus I obtained the temperature-depar- 

 tures which constitute weather, as freed from the average march of 

 events which constitutes climate. These results are plotted in curve A 

 of figure 4 and given in column 9 of table 5. 



I then analyzed these temperature-departure data in the manner 

 already explained regarding the solar data. I employed in my analysis 

 the same periods of 68, 45, 25, 11, and 8 months used in the solar 

 work. These were found to represent to a surprisingly close approxi- 

 mation the variation of Washington temperature-departures since 

 1918. The agreement with observed data was somewhat improved by 

 adding a sixth period of 18 months. These six periodicities are shown 

 graphically in curves C, D, E, F, G, H of figure 4, and their summa- 

 tion in curve B. The actual data from which these curves are plotted 

 are given in columns i to 8 of table 5. 



The reader, I think, will agree with me that the similarity between 

 curves A and B of figure 4 is both close and significant. Not only are 

 the main trends of the original observations fairly well reproduced in 

 the periodic summation, but many of the details also. Discrepancies, 

 indeed, occur at several times, and unfortunately a principal one is 

 found in 1930. One, therefore, hesitates to predict that the tempera- 

 ture departures of 193 1 and following years will be defined by the 

 same six periodicities without modifications of amplitudes or phases. 

 Nevertheless the discrepancy of 1930 is not much more pronounced 

 than several preceding ones, after which fair agreements returned. 



It may be objected that the five solar periodicities alone were in- 

 sufficient to give the best representation, without adding a sixth of 

 18 months not found conspicuously in solar variation. Is not this last 

 periodicity possibly of terrestrial origin? May it not be due to some 

 peculiarity of Washington surroundings which lends a predisposition 

 to a periodicity of 18 months? For analogy, consider an automobile 

 on a dirt road. It vibrates as the wheels strike the irregularities of 

 the road, in a manner depending on these outside interferences. But 



' Issued monthly by the United States Weather Bureau, Washington, D. C. 



