[ 77 ] 
has not been in the leaf): out of order, in nine months 
I have had it ; nor the other in nine years. 
The exadtnefs and facility with which an account 
of the variations in the weight of the atmofphere 
may be kept, with the help of a barometer of this 
kind, mufl be very evident. I have frequently found, 
by extraordinary variations that have happened in 
the night, when the wind has rifen conliderably, 
how little the obfervations made with common baro- 
meters are to be depended on ; and have feveral 
times found by the regifters, that the mercury had 
funk 50 or 60 divifions ; and one night particularly 
had funk 117 degrees, and returned within a degree 
and half of the place I had marked it on going to 
bed. When a ftrong guff of wind riles, one may 
very plainly perceive the index of this latter baro- 
meter to link feveral divifions, and rife again as it 
abates. 
Befides the fatisfadtion that a barometer of this 
kind might afford to a curious obferver, I fhould 
imagine it might alfo be ufefully applied to the find- 
ing the height of the atmofphere ; with a much 
greater degree of exadtnefs, at leaf!:, than can well 
be afforded by any other. 
It is generally allowed from experiments, that a 
column of air 72 feet high is equal in weight to 
I inch of water of the fame bafe ; fo that, if the air 
were of equal denfity throughout, the atmolphere 
could be little more than 5 miles high. But as the 
denfity is found to decreafe by the difference of pref- 
fure, and the air to be more rarined or expanded 
in proportion to its diftance from the earth, it feems 
reafonable to conclude, that if, by accurate experi- 
, ments, 
