[ 604. 1 
From a few obfervations of this kind, carefully re- 
peated, the expan lion of the bafis may be fettled 
and this once done, the making experiments upon 
other bars will become very eafy, and compendious. 
The bads of this inflrument (as well as other parts 
thereof) is brafs. I chofe this fubftance rather than 
any other whofe expan Pion was greater or lefs, becaufe 
I found, from fome grofs experiments previously 
made, that the expanfion of brafs was nearly a me- 
dium between thofe bodies, which differ mod: in their 
expanfion : A confiderable convenience arifes from 
this circumdance ; becaufe as the meafures, taken in 
common experiments, are their difference from brafs, 
the dependence upon the thermometer will be lefs, as 
thefe differences are lefs. This precaution I have 
found the more neceffary, as the greateft errors that 
experiments made with this indrument are fubjedt to, 
feem to be chiefly owing to the thermometer, tho’ 
that which I us’d was well graduated, and good in 
other refpedts ; but this muft neceffarily happen, as 
the fcale and fenfibility ©f the micrometer, when 
thofe metals were try’d which differ mod from the 
bafis, were greater than that of the thermometer. 
The bar of brafs which compofes the bafis is an 
inch broad by half an inch thick, and Hands edge- 
ways upwards j one end is continu’d of the fame 
piece at right angles, to the height of three inches and 
an half, and makes a firm fupport for the end of the 
bar to be experimented ; and the other end adts 
upon the middle of a lever of the fecond kind, 
whofe fulcrum is in the bafis ; therefore the motion 
of the extremity of the lever is double the differ- 
ence between the expanfion of the bar, and the bafis. 
This 
