228 Clayton — Six and Seven Day Weather Periods. 



between 40° and 50° 1ST. lat. be divided into intervals of this 

 length there are formed approximate squares the sides of 

 which are from TOO to 800 miles in length. The space be- 

 tween 65° and 105° W. Ion. was divided up into three of these 

 squares ; and Mr. Arthur Sweetland, an assistant, counted the 

 number of storms central in each square on each day of the 

 six and seven day periods during the seven years 1886 to 

 1892, inclusive. The results are shown in the accompanying 

 table. 



The numbers of storms central between latitudes Jfi° and 50° N. in sections of 15° of 

 longitude during the seven years 1886-92. 





Seven 



Day Period. 







Six Day 



Period. 







Long. 



Long. 



Long. 





Long. 



Long. 



Long. 



Dav. 



90°-105° 



T5°-90° 



60°-75° 



Day. 



90°-105° 



75°-90° 



60°-75' 



i 



66 



86 



76 



6 



68 



102 



92 



2 



52 



98 



77 



1 



58 



121 



93 



3 



49 



97 



92 



2 



69 



97 



102 



4 



58 



73 



86 



3 



59 



103 



93 



5 



55 



96 



80 



4 



74 



101 



105 



6 



56 



86 



92 



5 



62 



101 



97 



7 



53 



87 



77 











It is seen from this table that there were two distinct 

 maxima of frequency in each square, and the maxima occurred 

 successively one day later for each 15° of longitude eastward. 

 This corresponds exactly with the average eastward move- 

 ments of storms in the United States, and indicates that the 

 periodicities in question arise from a regularity in the origin 

 and eastward drift of ordinary cyclones. When the time is 

 separated into two intervals, namely, from 1886-1888, and 

 1889-1893, the same double maxima are found, and also the 

 later occurrence of the maxima in the eastern squares. How- 

 ever, in the seven day period the corresponding maxima in 

 each case occur on the average - 46 day earlier in the interval 

 1886-1888 than in the interval 1889-93, showing that the 

 length of the period used was about 0*03 hr. shorter than the 

 true length which would be 7d. 6*43 hr. The two intervals in 

 tbe case of the six day period showed that the most probable 

 length of the period was 6d. 3*94 hrs. When the space be- 

 tween latitudes 30° and 40° is divided into squares correspond- 

 ing with those between latitudes 40° and 50°, the chief maxima 

 are found on the same days as in the corresponding longitude 

 farther north, showing that the displacement of the maxima 

 is one of longitude and not of latitude. 



To study the question of periodicity in the case of succeed- 

 ing individual storms, the charts of the U. S. Weather Review 

 showing the position of the storms at 8 A. M. and 8 P. M. and 

 the directions of their tracks were carefully examined. It 

 was found that during intervals of about 27 days, correspond- 



