ef the 'temperature of boiling Water. 365 
greater than by the theory alluded to !t i. But thefe were 
but grofs experiments, and perhaps unworthy of fuch a 
competition. They induced me, however, to make the 
following. In the beginning of laft year (1778) with 
the afliftance of Mr. ramsden, I procured a moll excel- 
lent thermometer, every way adapted for this purpofe. 
It was about fourteen inches long, but the interval 
between freezing and boiling only 8^ inches'^, and 
though every degree was fomething lefs than the 
-jtjth of an inch, yet, by means of a femi-tranfparent piece 
of ivory, which applied * itlelf clofe behind the glafs 
tube, Hiding up and down in a groove cut in the brafs 
fcale for that purpofe, carrying a hair-line diviiion, at 
the extremity of which was a vernier dividing each de- 
(b ) The fame inftrument immerfed in fnovv juft melting at the top of Mount 
Cenis fell to 32°, the point of freezing obferved at the level of the fea. 
(c) It may poftibly be fuggefted, that if this interval had been greater, viz * 
20, 30, or 40 inches, I fhould have had a much larger fcale and more convex 
nient inftrument; but in this, as in moft other mechanical contrivances, our 
progrefs beyond certain limits is prevented; for if the perpendicular height of 
the column of quickftiver be much increafed, the weight of it will be fuch as 
to diftend the ball, and the inftrument may differ from itfelf in a vertical and 
horizontal pofttion by half a degree, as I have feen in a tube, only fifteen inches 
long; and if this circumftance be endeavoured to be corre&ed by making the 
bulb of the thermometer thicker, its fenfibility will be proportionably dimi- 
ni fifed. If my experience were to lead me to conclude any thing, 1 fhould 
conftder a tube of a foot long as a maximum , and the bore of fuch a diameter 
as to admit a ball of a quarter or one fifth of an inch. 
Vo l* LXIX. C c c greo 
