thermometrical barometer for measuring altitudes . 191 
is then enveloped with the bulb downward in the folds of the 
tent inverted. The lamp is put first into the case, and the 
other parts being thrust down afterwards, are kept very 
steady by the linen of the tent. The whole goes into a tin 
cylindrical box, two inches diameter and ten deep, and weighs 
lib. 4 jOz. 
The scale of an inch to a degree is. chosen, because on try- 
ing various threads, I have found that when extremely fine, 
it is almost impossible to give such strength to the bulb as to 
force the column of mercury accurately to the same height 
on repetition of the boiling, by reason of the resistance 
from friction in the tube. With an inch scale, the variations 
of the barometrical thermometer will be to those of the com- 
mon barometer as 5 to 3, and the sensibility in this instru- 
ment is so fully sufficient, that the difference of temperature 
arising from the height of a common table is immediately 
perceptible. If more were wanted, either the thread must 
be finer, which would endanger the precision ; or the bulb 
larger, the objections to which are obvious. Adhesion to the 
glass and friction must always have some small effect : and 
for the same reason that a common barometer is shaken on 
observation, this instrument, when boiled, should be tapped 
gently two or three times on the side, to free the mercury ; 
when that is done, whether on the rising of the column to its 
height, or the falling of it after a forced expansion, it will 
with this sized thread come to the same place precisely. 
Upon trying this thermometer when mounted, I found that 
a variation of 0,589 barometer, which was before ascertained 
= 1 Fahrenheit, would be equal to 233 parts on my scale 
= 0,97 inch ; and an inch on barometer would produce a 
mdcccxvii. C c 
