1819.] and on the Laws of the Communication of Heat. 12] 
tion, the lute gets hot, and would at last be detached, unless 
care were taken to cool it perpetually by a current of water. 
The box thus constructed was solidly fixed on a furnace, 
supported on all sides by iron bars. This furnace, in the figure, 
is supposed to be cut in two, that we may see the pieces in its 
inside. 
We shall terminate this preliminary description by saying, that 
the copper cylinder is filled with a fixed oil, which is gradually 
heated till it reach the requisite temperature. Then all the 
mouths of the furnace are shut; the heat then spreads itself 
uniformly through the whole mass, and the temperature remains 
stationary during a time sufficient to take all the requisite 
measures. But that mothing may alter the exactness of these 
determinations, it is necessary that the copper be always com- 
pletely filled with oil, and that the hot ‘column of mercury 
terminate at a very small height above the cover. We easily 
fulfil this second determination by adding, or withdrawing by 
means of a sucker (pipette), the requisite quantity of mercury, 
some instants before the observation. As to the first, it is 
obtained by filling the vessel with oil, when cold, and by putting 
at the top of the vessel,.a tube, L Q, whose orifice, Q, is on a 
level with the under side of the cover. Through this tube the 
oil flows outias it dilates. 
Let us now proceed to the measurement of the temperatures, 
and of the heights of the columns. 
. The oil bath contains two thermometers, the one mercurial 
and analogous to that which we have had occasion to describe 
already, and in which the temperature is calculated by compar- 
ing the weight of the mercury which has made its escape from 
the instrument, with that which it contains at zero. Such is the 
sensibility of that which we employed, that an increase of tem- 
erature of one degree made about one decigramme of mercury 
issue out. Its reservoir, DE, is every where of the same 
diameter, and is plunged into the oil to the same depth as the 
column, A’ B’. Of course, it gives the exact mean temperature 
of the column. 
The second is an air thermometer, whose cylindrical reser- 
voir, D’ E’, placed like that of the preceding, is terminated by 
a very fine tube, E’ G’H’, curved horizontally beyond the 
furnace. This tube is united at H’, with a vertical tube, a little 
larger, and well calibred, which is plunged into the mercurial 
bath, K’. To regulate this thermometer, the bath was in the 
first place heated nearly to the boiling point of the oil, while the 
extremity, K’, of the tube remained open. When the whole 
‘excess of air had been driven out by the heat, the orifice, K’, 
was plunged into the mercury, and by the cooling of the oil, the 
mercury rose gradually in the tube. It is by measuring the 
height of this column, at the maximum of temperature, and that 
pf the barometer, that the augmentation of the elasticity of the 
