BY B. C. NOWELL. 



51 



I have thought it better not to go beyond the seventh day, 

 since it is only occasionally that the interval between one 

 change of the moon and another extends beyond that day. 



By way of testing the accuracy of my own observations, I 

 constructed another table (A) on the same principles, from 

 the data furnished by the late Mr. F. Abbott, and published 

 in the papers of this Society, fating in the previous ten 

 years, from 1870 to 1879 inclusively. If the two tables be 

 thrown into one, they will embrace a period of twenty years, 

 from 1870 to 1889. In Mr. Abbott's tables, as printed, I 

 found two gaps, one of a month, and another of a fortnight, 

 but was fortunately able to supply the omissions from the 

 meteorological journal kept by Mr. W. E. Shoobridge, which 

 he was kind enough to leave in my hands. 



My observations, as thus tabulated, give the following 

 results for the ten years from 1880 to 1889. 



Taking first the totals opposite to each day, without 

 reference to the nature of the moon's changes, we find that 

 the number of times on which rain commenced in the 

 decennium as a whole was as follows : — 

 On the day of the change 



second day 



third 



fourth 



fifth 



sixth 



seventh „ 



103 

 98 

 92 

 88 

 73 

 82 



It will be seen that the number on the day of change is 

 considerably in excess of the number on any of the other 

 days, being at the rate of 18 per cent, nearly of the whole 

 number (650) on the seven days, while the next highest is 

 only about 16 per cent. 



The corresponding numbers for the period 1870-79 were : — 



135 



second day „ 



... 80 



third „ ,, 



... 97 



fourth ,, „ 



... 81 



fifth „ „ 



... 93 



sixth ,, „ 



... 83 



seventh „ „ 



... 84 



Here again we find the largest number on the day of the 

 change. But in regard to the other days, the two periods 

 vary. In the first decennium (1870-79) the next largest was 

 on the third day ; in the second decennium (1880-89), on the 

 second. 



The proportion which the numbers on the day of change 

 tore to the total (653) on all the seven days in the first 

 period was nearly 21 per cent. 



