502 MR. M. CHRISTY ON THE SEASONAL VARIATIONS OF 
branches of which are often greatly bent downwards, and some- 
times broken, by the weight of the fruit they bear. This, however, 
is not quite the same thing as that herein discussed. In the 
first place, in the case of these fruit-trees, the conditions are 
somewhat abnormal; for the weight of the fruit they bear is 
generally very much greater, in proportion to the strength of the 
branches, than that ordinarily borne by trees growing in a state of 
nature. In the second place, the conditions, beside being somewhat 
abnormal, are also in another way essentially different; for, in 
the case of these fruit-trees, the effect in question is mainly 
produced by the weight of the fruit and not (as in the case 
under consideration) mainly by the weight of the leaves. 
The Horse-Chestnut tree producing the branch upon which 
my observations have been made is a small one, about 40 feet in 
total height, growing in a position where it is a good deal shaded 
and overcrowded by other, and larger, trees. The branch in 
question is the lowermost upon the tree. It leaves the bole 
about four feet above the ground, where the bole is about 
53 inches in circumference. The branch, at the point where 
it leaves the bole, is about 26 inches in circumference, and 
it is about 28 fect 6 inches in total length, excluding the 
leaves. 
In order to render easy the carrying on of my observations, 
I drove a large nail into the side of the branch at a point 
16 feet 3 inches from the bole (and, consequently, 12 feet 
3 inches from the extremity) ; while, immediately beneath te 
nail, I buried a brick so that one end of it just appeared above 
the surface of the ground. lt was then an easy matter to 
measure with a piece of cord the periodical variations in the 
distance between the nail and the top of the brick. , 
My observations were continued at short intervals from 
April 21st, 1895, till April 28rd, 1898—a period (as already 
stated) of exactly three years. During this period 1 measured 
the distance between the nail and the brick no fewer than 42 
times (12 times in 1895; 9 times in 1896; 14 times in 1897 ; 
and 7 times in 1898)—that is, on an average, about once ® 
month. The results are stated in detail in Table A, which 
appears on the following page and requires no explanation. 
