5 66 Sir george shuckburgh’s Obfervations 
heat, any where within the limits above-mentioned, 
would be about one .part in five thoufand from what I 
t take it at a mean. I ihould not have infilled fo long on 
this circumltance, but in refpeft to the known accuracy 
of the author of this hypothefis. Neither do I find any 
reafon to believe, ' that the expanfion of air varies with 
its denfity. I have tried air whofe denfity or preflure 
was equal to 23f inches, and alfo to forty inches; but the 
dilatation, with equal volumes and equal degrees of heat, 
was very nearly the fame in both cafes. I might add a 
great deal more on thefe manometrical experiments, but 
I am afraid it would be more tedious than ufeful. I pro- 
ceed therefore to the expanfion of quickfilver. 
This experiment was made with a tube, fomething 
like a thermometer, but confiderably larger than the or- 
dinary fize, and open at one end; it was filled with 
quickfilver to a certain height, and then expofed to the 
temperatures of freezing and boiling repeatedly, the ba- 
rometer being at 30 inches : the difference of the volume 
in each inftance was determined afterwards by accu- 
rately -weighing the contents. I thus found, that if the 
quickfilver at freezing be fuppofed to be divided into 
1 31 19 parts, the increafe of volume by a heat of boil- 
ing water became equal to 208 of thefe parts = and 
~ x ~ - TT ~~ ; and fuch would be the expanfion for 
each 
2 
