484 Meteorology. 
of the usual length, and between 1 and 2 
tenths of an inch internal diameter; it was 
carefully filled with dry mercury, and inver- 
ted into a cylindrical cup containing mer- 
cury; the diameter of the cup was such as 
to require scarcely any sensible allowance for 
the rise and fall of the mercury in the tube. 
It suffered no material change, as may be 
inferred indeed from the annual means, till 
near the end of the period. During my 
absence in August 1798, it had been in un- 
skilful hands; a part of the mercury had 
been lost out of the cup, and probably a few 
atoms of air had got into the tube. Not 
being at that time particularly interested in 
meteorology, I contented myself with noting 
the daily observations as usual, without sum- 
ming up the monthly means, or making any 
comparative observations. The consequence 
was, that some years elapsed before I was 
struck with the depression that had taken 
place in the mercury, which upon examina- 
tion, appears to have been about 2 js of an 
inch. It was not till after another period. of 
5 years that I determined to renew the ba- 
rometer. A tube was taken of- about -+ of 
an inch internal diameter, having a large 
bulb at the lower extremity; this was filled. 
with mercury that had a few minutes’ before 
