A PENDULUM FOR THE REDUCTION TO A VACUUM. 
447 
The values above given are nearly (although not exactly) in proportion to the 
cubes of the diameters: but, it is possible that some other element, at present 
unknown, may affect the results; and indeed some portion of the air may adhere 
to, or be dragged by the sides of the sphere. As the exact measure of these 
three brass spheres was, however, a matter of importance in this inquiry, I exa¬ 
mined them more minutely, and found them to be 1*465, 2*065, and 3*030 inches 
respectively. So that the weight of adhesive air for the last tw*o spheres will 
be almost exactly as the cubes of their diameters ; and, for the first two, not 
materially differing therefrom. In fact, if the weights of air were *387, 1*084, 
and 3*422 grains respectively, the whole would agree precisely with this hypo¬ 
thesis. It is worthy of remark that, in the case of the spheres No. 1 to 7? sus¬ 
pended by a wire (see the Table in page 433), and No. 66 in the following set, 
if we consider the weight of air, dragged by the wire alone, as equal to 0*10 
grain, and deduct this value successively from the mean weight of air dragged 
by the 1*46 and the 2*06 inch spheres respectively, as there given, and by the 
3*03 inch brass sphere as given in the following set of experiments, we shall 
have 0*373, 1*040, and 3*444 grains for the weight of air dragged by the spheres 
alone. So that, on the whole, I consider the hypothesis adduced as not far 
from the truth ; and that the general expression for the quantity of air dragged 
by a pendulum consisting of a sphere suspended by a rod or wire, will be as 
follows: viz. 
grains. 
r + 0*123 X d? 
where d denotes the diameter of the sphere in inches, and r the quantity of air 
dragged by the rod or wire. And if, in the case of a sphere suspended by a 
fine wire, of the length of the seconds pendulum, we suppose r to be (as already 
stated) equal to 0*10 grain, this formula will become 
*002564 1 + *123 d 3 
where l denotes the length of the wire, in inches. 
These values do not differ materially from those obtained by the same 
spheres attached to the ends of the long rods, as given in the 14th set of expe¬ 
riments : but I have already stated that those results were obtained under less 
favourable circumstances, and are not to be relied on with the same degree of 
confidence as the present set. They will be found however to accord more 
3 m 2 
