[ * 5 * ] 
n:able air mu ft be 5490 times lighter than water, or 
near 7 times lighter than common air. But if the 
denfity of common air was 850 times lefs than that 
of water, then would inflammable air be 9200 times 
having weighed his bottle under water, both when full of air and 
when exhaufted, he fuppofes the difference of weight to be equal 
to the weight of the air exhaufted ; whereas in reality it is not fo 
much : for the bottle, when exhaufted, mull neceflarily be com- 
preffed, and on that account weigh heavier in water than it would 
otherwife do. Suppofe, for example, that air is really 800 times 
lighter than water, and that the bottle is comprefled 4 o"g P art 
of its bulk ; which feems no improbable fuppofition : the weight 
of the bottle in water will thereby be increafed by’ T ~o^ of the 
weight of a quantity of water of the fame bulk, or more than 
vt of the weight of the air exhaufted : whence the difference of 
weight will be not fo much as °f t ^ ie weight of the air ex- 
haufted : and therefore the air will appear lighter than it really is in 
the proportion of more than 15 to 14, /. e. more than 857 times 
lighter than water : whereas, if the ball had been weighed in air 
in both circumftances, the error arrifing from the compteffion 
would have been very trifling. 
It appears, from fome experiments that have been made by 
weighing a ball in air, while exhaufted, and alfo after the air was 
let in, that air, when the thermometer is at 50°, and the barometer 
at 29I, is about 800 times lighter than water. Though the 
weight of the air exhaufted was little more than 50 grains, no error 
could well arife near fufEcient to make it agree with Hawkfbee’s 
experiment. Air feems to expand about o part by i° of heat, 
whence its denfity in any other ftate of the atmofphere is eafily 
determined. The denfity here affumed agrees very well with the 
rule given by the gentlemen, who meafured the length of a degree 
in Peru, for finding the height of mountains barometrically, and 
which is given in the Connoiffance des mouvemens celeftes, annee 
1762. To make that rule agree accurately with obfervation, the 
denfity of air, whofe heat is the fame as that of the places where 
thefe obfervations were made, and which I imagine we may 
ellimate at about 45 0 , fhould be 798 times lefs than that of water, 
when the barometer ftands at 294-. 
6 lighter 
