31 
On the Measure of Temperature . 
The comparison of the mercurial thermometer with that of air, was long ago 
made by JVI. Guy Lussac, .within the limits of temperature,, answering to the 
freezing and boiling points of water. The result of his experiments is, that within 
that interval of temperature, the scales scarcely differ in their indications. 
Mr- Dalton again thinks, that the mercurjpl thprmpmet^r marks about 1° 
in excess of the air thermometer towards the middle of the scale, where, in fact, 
it might be concluded, the difference of the two would be greatest, since by hypo- 
thesis they coincide at the two extreme points of 0° and 100 degrees. 
We see then, that if there be any difference in the rates of expansion, of air afnd 
mercury, it must be very small between the limits of the freezing and boiling 
points of water- 
We extended tin's comparison to low temperatures. Our first experiment made 
at 20°, gave a perfect agreement between the indications of the two instruments ; 
and in a great number of observations made at 30° tp 36°, we have found 
small differences, sometimes on one side, sometimes on the other, showing, iu fact, 
the mean of each series to be the same*. So that witljiu an extent of 130'*, 
the difference of the two scales is so small, as not to be appreciable amongst the 
errors of observation. 
No experiments are more easily prosecuted than these, while we keep within the 
boiling point of water, but in attempting to pursue theiuq.uiry amongst the higher 
temperatures, so many difficulties are met with, as oblige us to have recourse to me- 
thods of a more tedious as well as complicated nature. Those which we followed, 
and which we are now to describe, appeared to us to promise as much precision as 
researches of this nature will admit. 
The apparatus consists of a rectangular vessel of copper seven decimetres in 
length (in. 27,5), one decimetre in breadth (in. 3,9), and one in depth (in* 3,9). 
This vessel has on one of its sides two apertures ; in one of these a mercurial ther- 
mometer is inserted in a horizontal position, in the other a tube is similarly 
placed, well dried, and filled with perfectly dry air. 
The vessel is fixed in a furnace in such a manner that every part of it can be 
equally heated. It is filled with a fixed oil which, it isknpwn, is capable of being 
heated to upwards of 300° before it reaches the boiling point. 
The tube which contains the air, is terminated on the side of the aperture by a 
short tube of a very small diameter, which projects a little.beyond the vessel* The 
quantity of air contained in the extern; >1 part of th,s tuhp- and which is not 
heated equally with the rest, may be entirely neglected. We know that it 
never exceeded ^th of the total mass; but besides this, we took the precaution to 
heat it during each experiment, in order, as much us possible, to diminish even this 
small error. 
The vessel has a cover with several openings in it. Tn some are fixed vertically, 
thermometers which serve to indicate, whether the different part s of the liquid muss 
are at the same temperature ; others carry spindles furnished with flyers or vanes, 
the rotation of which agitates the liquid, the object of which is to establish ,a uuir 
fortuity of temperature. 
The following is the or>’er which was followed in each experiment. The vessel 
wrns first heated, to a temperature very near that which it was,tvi f shed finally to 
establish, and all tbe openings of the furnace shut. The heat having a tendency 
to be equally distributed throughout the apparatus, the temperature ,pf the oil 
still rose several degrees, when it became stationary and consequently easy to 
measure with accuracy. This was done by the horizontal thermometer, which 
* In order to give an idea of the closeness of these comparative valuations, we 
shall give the following particulars of some made between 36° and 30°. 
Air thermometer 
Mercurial thermometer. corrected for the expansion of glass. 
—36, •29 
—36,18 
—•34,72 
—34,84 
—■■(4,31 
—33,40 
— 32,27 
—32,13 
—31,63 
—31,54 
—31,26 
—31,04 
—30,46 
—30,50 
—29,68 
—29,64 
Mean — 32,452 
-32,420 
