110 M. F/augergites on the Quantity of Rain, and [Ate. 



The quantity of rain which falls in autumn and winter is to the 

 quantity which falls in spring and summer nearly as 13 to 12; 

 and the quantity of rain in winter and spring is to the rain in 

 summer and autumn nearly as 7 to 10. 



But this table does not give the exact rate of the rain in each 

 month, because the months have not all the same number of 

 days ; and on that account, the longest ones ought to exhibit a 

 greater quantity of rain; and the shortest, a smaller quantity, than 

 that which results from their constitution. We must, therefore, 

 reduce the quantity of rain given for each month in the preced- 

 ing table to what it would be if each month were equal to the 

 12th part of the mean year (365 days 6 hours); that is to say, 

 equal each to 30 days 10*5 hours. When this reduction is made, 

 the preceding table will be changed into the following one : 



January 0-0716 



February 0-0541 



March 0-0557 



April 0-0802 



May 0-0847 



June 0*0765 



July 0-0544 



August 0-0679 



September 0*1236 



October 0-1370 



November 0*1250 



December 0*0693 



From this table, it appears that the months, when arranged 

 according to their quantities of rain, beginning with the most 

 rainy, assume the following order : 



October, November, September, May, April, June, January, 

 December, August, March, July, February. 



The month of May comes nearest to the mean quantity of rain, 

 which is O 0833. Four months are above this quantity, and may, 

 therefore, be considered as rainy. Eight months are below it, 

 or dry. 



That the order of the rain in a mean year may be the better 

 understood, I have expressed it graphically in tig. F (PI. XCV). 

 For this purpose, I took upon the line, M N, the axis of the 

 abscisses, 12 equal intervals to represent the 12 months of the 

 year. From the middle of each of these intervals, I have raised 

 a perpendicular ordinate, of a length proportional to the numbers 

 in the above table, beginning with the month of February, and 

 through the extremities of these ordinates I have passed a curve 

 line, A B C D E. This curve is very regular, with the exception 

 of the ordinate, corresponding to the month of January, which 

 is a little too long relatively to the ordinate which corresponds 

 to the month of December ; an irregularity occasioned by the 

 excessive rains which took place in January, 1814; but it would 

 doubtless disappear if the observations were continued for some 

 vears longer. This curve has two minima and two maxima. 

 The first minimum corresponds to the month of February, the. 

 second to the month of July. The first maximum corresponds 

 to the month of May, and the second to the month of October. 

 T,he two minima are nearly equal ; but the maximum of October 



