226 Clayton — Six and Seven Day Weather Periods. 



In the preceding are given the longest series of six and seven 

 day periods found ; but in all the numerous cases the mean 

 lengths of the periods rarely differed more than one hour from 

 6 days 4 hours, and 7 days 6 hours, respectively. 



The next step in the investigation was to ascertain whether 

 the end of one series was separated from the beginning of the 

 next by even multiples of these mean intervals. This was 

 found to be true if the series in each case were divided into 

 two groups separated from each other about half a period, 

 three days in one case and three and a half in the other. This 

 discovery enabled a more exact determination of the periods 

 to be made, and the mean lengths were found to be 7d. 6*4 hr., 

 and 6d. 4-0 hr. 



The next step was to divide the time from Jan. 1, 1885, to 

 Aug. 31, 1893, into intervals of these lengths and ascertain 

 whether the periods would show themselves in the average of 

 all observations. The first investigation in this line was to 

 count the number of times on each day of both periods that 

 barometric minima were observed at Blue Hill Observatory. 

 The results are shown for each interval of three years in the 

 following tables : 



Number of Barometric Minima at Blue Hill Observatory — Ion. 71° W. and lot. I$° N. 

 On each day of the Seven day period. 





l 



2 



3 4 5 





6 



7 



i885-'87, 



43 



38 



47 48 40 





43 



44 



18S8-'90, 



44 



37 



30 44 44 





25 



50 



1891-'93, 



34 



19 



31 46 35 





35 



39 



Sum, 



121 



94 



108 138 119 



: 



103 



133 





On eacl 



i day of the Six day period. 











l 



2 



3 4 



5 





6 



1885-'87, 



34 



55 



49 40 



53 





53 



1888-'90, 



43 



52 



45 40 



52 





43 



189l-'93,, 



35 



40 



37 34 



53 





4L 



Sum, 



112 



147 



131 114 



158 





137 



In both of these tables are included all the barometric minima 

 recorded at the Observatory. Both tables show two maxima of 

 frequency in each period which persistently remain on the same 

 days of the periods, thus proving that the periods have defi- 

 nite and constant lengths ; and that these lengths have been 

 approximately determined. A close examination of the 

 figures, however, leads to the conclusion that the mean lengths 

 as given are not absolutely correct. Thus in the first group of 

 the seven day period, 1885 to 1887, the maxima probably 

 occur between the third and fourth, and the fifth and sixth 

 days ; while in the last group, 1891 to 1893, the maxima occur 

 sharply on the fourth and seventh days, thus indicating a for- 

 ward movement of the maxima of about half a day in six 



