[ 642 ] 
by weighing the quantity of mercury that exa&Iy 
tilled the ball, and alfo the quantity that filled the 
whole length of the tube; I found that the mercury 
tVo an inch of the tube, was the 100000 th part 
of that contained in the ball ; and with the edge of a- 
file, I divided the tube accordingly. 
This being done, I filled the ball and part of the 
tube with water exhaufted of air ; and left the tube 
open, that the ball, whether in rarefied or condenfed 
air, might always be equally preffed within and with>- 
out, and therefore not altered in its dimenfions. Now 
by placing this ball and tube under the receiver of an 
air-pump, I could fee the degree of expanfion of the 
water, anfwering to any degree of rarefaction of the 
air ; and by putting it into a glafs receiver of a conden- 
fing engine, I could fee the degree of compreffion of 
the water, anfwering to any degree of condenfation of 
the air. But great care muft be taken in making 
thefe experiments, that the heat of the glafs ball be 
not altered, either by the coming on of moifture, or- 
it’s going off by evaporation ; which may eafily be pre- 
vented by keeping the ball under water, or by ufing 
oil only, in working the pump and condenfer. 
In this manner, I have found by repeated trials, 
when the heat of the air has been about 50 degrees, 
and the mercury at a mean height in the Barometer, 
that the water will expand and rife in the tube, by 
removing the weight of the atmofphere, 4 divihons 
andJL.; or one part in 21740; and will be as much 
comprelled under the weight of an additional atmof- 
phere. Therefore the compreffion of water by twice 
the 
