Jbtne Thermometrlcal Experiments. 525 
contained, which may be made to boil by the flame of a large 
wax candle. The thermometer is fattened to the tin veflel 
in fuch a manner as that its bulb may be about one inch above 
the bottom. The fcale of this thermometer, which is of 
brafs, exhibits on one fide of the glafs tube a few degrees of 
Fahrenheit’s fcale, viz. from 200° to 216°. On the other 
fide of the tube are marked the various barometrical heights, 
at which the boiling water (hews thofe particular degrees of heat 
which are fet down in Sir g. shuckeurgh’s table. With this in- 
ftrument the barometrical height is fhewn within one-tenth of 
an inch. The degrees of this thermometer are fomewhat longer 
than one-ninth of an inch, and confequently may be fub- 
divided into many parts, efpecially if a nonius is ufed. But 
the greateft imperfection of this inftrument arifes from the 
fmallnefs of the tin veflel, which does not admit a fufficient 
quantity of water : and I find, that when a thermometer is 
kept in a fmall quantity of boiling water, the quickfilver 
in its ftem does not ftand very fteady, fometimes riling or 
falling even half a degree ; but when the quantity of water is 
fufflciently large, for inftance is ten or twelve ounces, and is 
kept boiling in a proper veflel, its degree of heat under the fame 
preflure of the atmofphere is very fettled. 
