101 
centre respectively proportional to the amount of rain which 
had fallen when the wind was in that direction ; by then con- 
necting these spaces a curve is formed, which shews at a 
glance the character of the rain for that particular period, 
The plates are only drawn for the mean results, which are the 
only ones of any importance. The period of five years is 
rather shorter for very accurate mean results than could 
have been wished, but absence from home after that period 
had elapsed put a stop to the observations. 
There are one or two points to which I wish to draw atten- 
tion. First of all, taking the average monthly rain at three 
inches, the first six months of the year are below the ave- 
rage, the other six months above it. November and July are 
by far the two wettest months in the year; and in each the 
greatest amount is from S. W. April is much the driest 
month, and there is nearly as much rain in it from the northern 
portion of the compass as from the southern. 
With regard to the gross amount which fell from each 
point in the entire year, that which fell from S. S. W. and 
W. is much above the average. From the other points it is 
below it. 
There is a curious circumstance with regard to the curve 
of the entire year (Plate 6): that if it be divided by a line 
running N. E. and S. W., then the rain on either side of this 
line is equal all but a fraction of an inch. This is the more 
remarkable, as these two points had been fixed on by Profes- 
sor Dove, in his Paper on the Winds, as being the points of 
greatest and least barometric pressure. That is to say, the 
wind supposed at S. W., any shift of it, either towards S. or 
W. produces a rise of mercury ; and also any shift on either 
side of N. E., a corresponding fall. 
Now in the rain the greatest amount is from S. W. (cor- 
responding with the lowest state of mercury). The least is 
from N. E. (where the mercury is highest), and on either side 
‘of this line it varies regularly, as an inspection of the Plate 
