[ 152 ] 
probably is not in any degree affected by the moifture 
or drynefs of the air. The pulley, on which thefe 
threads are placed, is made double : that on which 
the weight e is fufpended, furrounds one part j and 
the thread on which the counterbalance weight d is 
fufpended, furrounds the other: fo that when the 
portion of the index is properly adjufted, it cannot 
eaiily be milplaced, the weight will always keep in 
its proper pofition on the furface of the mercury, 
carrying the index and regifter, as the mercury rifes 
or falls in the tube. 
The inward circle of the dial-plate is divided into 
three parts, correfponding with 3 inches generally 
allowed for the rife and fall of the mercury in com- 
mon barometers. Each inch is divided into twelve 
lines, and each line fubdivided into ten parts, on the 
outward circle. The regifters are very ilender, and 
mark very diftin&ly half of thefe diviftons, which 
is the 240th part of an inch rife of the mercury in 
the tube. 
Many fudden changes of the temperature of the 
air, and preflure of the atmofphere, have probably 
pafted unnoticed, for want of fome eafy method of 
marking the variations with fufticient precifion. It 
has been accidentally remarked, that the mercury 
has funk to a great degree, and rofe very fuddenly, 
during the fhock of an earthquake ; but, from the 
fuddennefs of the motion, the degrees could not be 
alcertained. Any luch fudden alteration, or even the 
common changes, will appear with fo much certainty 
by the regifters, that I fliould imagine, inftruments 
of this kind will greatly afiift thofe, who are obliged 
to a daily attention, in order to minute the changes 
3 that 
