284 I'HILOSOPHICAL TRANSACTIONS. [ANNO ] 76'2. 



unavoidably arbitrary ; but the method here proposed appears sufficiently com- 

 modious and easy of execution, the divisions being adjusted by measures every- 

 where known, and at all times obtainable: for however the inches used in dif- 

 ferent countries may differ in length, this cannot affect the accuracy of the scale, 

 provided the proportions between the wider and narrower end of the gage are 

 exactly as -fa of those inches to -fa, and the length 240 of the same lOths; 

 and that the pieces in their perfectly dry state, before firing, fit precisely to the 

 wider end. When one gage is accurately adjusted to these proportional measures 

 2 pieces of brass should be made, one fitting exactly into one end, and the other 

 into the other; these will serve as standards for the ready adjustment of other 

 gages to the dimensions of the original. By this simple method we may be 

 assured that thermometers on this principle, though made by different persons 

 and in different countries, will all be equally affected by equal degrees of heat 

 and all speak the same language: the utility of this last circumtance is now too 

 well known to need being insisted on. 



8. If a scale 2 feet in length should be reckoned inconvenient, it may be 

 divided into 2, of 1 foot each, by having 3 pieces of brass fixed on the same 

 plate; the 1st and 2d, -fa of an inch apart at one end, and -fa at the other; the 

 2d and 3d, -fa at one end, and -fa at the other; so that the first reaches to the 

 120th division, and the 2d from thence to the 240th. 



g. As this thermometer, like all others, can express only the heat felt by 

 itself, the operator must be careful to expose the pieces to an equal action of the 

 fire with the body whose heat he wants to measure by them. In kilns, ovens, 

 reverberatories, under a muffle, and wherever the heat is pretty steady and 

 uniform, the means of doing this are too obvious to need being mentioned. 

 But in a naked fire, where the heat is necessarily more fluctuating, and unequal 

 in different parts of the fuel, some precaution will be required. The thermo- 

 meter-piece may generally be put into the crucible, along with the subject- 

 matter of the experiment. But where the matter is of such a kind as to melt 

 and stick to it, the piece may be previously inclosed in a little case made of cru- 

 cible clay. The smallness of the pieces will admit of this being done without 

 inconvenience, at least in any but the smallest crucibles, as the pieces themselves 

 may be diminished to any size that may be found proper, provided only that one 

 of the dimensions, -^ u of an inch, be preserved as mentioned in obs. 4. For the 

 very smallest sort of crucibles, the case may be put in close to the crucible, so 

 as to form as it were an addition to its bulk on the outside. If it be asked, why 

 the case is not always thus put in by the side of the crucible? it is answered, 

 that in judging of the heat of large crucibles from a thermometer- piece placed 

 on the outside of them, we may sometimes be deceived, as the piece in its little 



