[ ] 
If the greateft expanfion, from freezing to boiling 
water, of the metal bars to be made ule of in an 
inftrument of this kind, is certainly known ; it may 
be made to mark any minute degree intended. But 
to anfwer a calculation of this kind ; the levers, and 
pullies, muft be made with the greateft exadtnefs; 
which I could not prevail on the workman that made 
mine to be fo careful of as I could wifh. It is eafy, 
however, to come to a certain knowlege of the ex- 
panfion, mechanically. I took a thick book, and 
prefling the leaves between my finger and thumb 
gently 5 meafured off an inch exactly, with a pair of 
compafles, from the 1 ft page, which made 368 pages. 
I cut off feveral flips from different leaves, each of 
which I placed between two bits of watch-fpring, 
which were marked, in order to place them in the 
fame pofition exactly, on each trial. On examining 
thefe with the inftrument, I found but few of an 
exadl thicknefs j but that, on a medium, each leaf 
raifed the minute index 1 30 divifions. So that the 
number of pages, each leaf of which is the 284th 
part of an inch, multiplied by the number of divi- 
fions it raifes the index ; fliews, that each divifion 
marks the 73,840th part of an inch expanfion per 
foot of the bar A, which is 2 feet long. 
This bar is made of fpelter, which is of fo brittle 
a nature, as makes it difficult to file one of that 
length without breaking. If it is not well caft, it 
will be extremely hollow j as was the cafe of a bar 
I made ufe of, which did not expand near as much, 
nor as uniformly, as another of the fame metal, and 
contracted much more readily y which I could not 
well account for, until it broke by accident ; when 
y O 2 it 
