1829.] 
Expansion of Mercury. 
361 
sary, that the bar M N should pass through this box , and that the space left between 
the bar, and the edges of the nperture, should be filled with a lute, impermeable to 
water. In order to fullfil all these conditions, we had a cylinder formed of copper, the 
bottom of which might be detached at pleasure. At the upper eud, the edge is turned 
over at right angles, forming a support on which the cover fits ; towards the bottom 
it carries two appendages in the shape of horizontal half cylinders, in which lies the 
bar M N — Fig. 2, a section of it made on a vertical plane, parelle! to the direction 
of these appendages, will give an exact idea of this arrangement. The bottom piece 
is represented in Fig. 3, — it can be fastened to the box by means of steel screws, 
which are lightened as much as possible. Even this force is inadequate to prevent 
leakage ; so that, small pieces of card have been necessarily interposed between the 
two copper surfaces. 
The use of these appendages is to allow of the lute being applied at a distance 
from the source of heat. Still, the lute becomes heated, and would eventually give 
way, if it were not kept cool by a constant current of water. 
The box thus constructed, is enclosed in a furnace, supported every way by iron 
bars. This furnace is shown in the sketch as if cut down through the middle, in 
order that all the interior arrangement may be seen. 
We shall conclude our description by saying, that the copper cylinder or box is 
filled with fixed oil, which can be heated gradually up to the temperature of obser- 
vation. The openings of the furnace are then all of them shut, and the heat disper- 
sing itself equally throughout the inass, the temperature remains stationary for 
an interval sufficient to allow of all the observations being made. But that nothing 
may interfere with the exactness of these determinations, it is necessary that the 
copper cylinder should he full of oil, and that the heated column of mercury 
should rise but a very little above the lid. This latter condition, was easily 
brought to hear, by adding or subtracting, a short time before the reading was 
to he made, a very small quantity of mercury, by the aid of a dropping tube. The 
former again was effected, by filling the cylinder in the first instance, and by ap- 
plying a pipe L Q the orifice of which U was level with the lower surface of the 
lid, for the extra quantity occasioned by expansion to flow off- 
Let us now turn to the determination of the temperatures, and of the heights of 
the columns. 
The oil bath contains two thermometers, the one mercurial, similar to that we 
have already had occasion to describe, and in which the temperature is estimated 
by comparing the weight of mercury lost by the instrument, to that which filled it 
at zero. Such is the sensibility of that we used, that an increase of 1° in the tem- 
perature occasioned the loss of a decigramme of mercury. The reservoir D E of 
the same diameter throughout, being immersed to the same depth as the column 
A' B' indicated the exact mean temperature of the, column. 
The second is an air thermometer, the cylindrical reservoir of which D' E' situa- 
ted like the preceding, is terminated by avery fine tube EG H bent into a horizon- 
tal direction on quitting the furnace. This tube is joined at H' to a vertical tube 
a little larger, of a uniform diameter throughout, and which dips into a vessel full 
of mercury K', In order to regulate this thermometer, the oil bath is heated near- 
ly to the boiling point, the emlof the tube K' beinn; left open. When the surplus air 
has been driven out by expansion, the extremity K is immersed into a saucer of dry 
mercury. During the cooling of the oil, the mercury rises a little in the tube. It is 
by measuring, during the maximum of temperature, the height of this column and of 
the barometer, that we discover the increase of elasticity in the air; whence, by a 
simple enough calculation, we can deduce the temperature as shown by the air 
thermometer. It is hardly necessary to add, that the tube had been carefully dried, 
and that in each measurement an allowance was made for capillary depression. 
The indications of this thermometer add in no ways to the exactness of those 
afforded by the mercurial thermometer. But we thought it advisable not to lose 
the opportunity of again comparing the scales of the two instruments. The results 
deduced from this comparison are included in the determination of the mean re- 
sults inserted in table 1. _ 
We have now only to describe the kind of micrometer, which we used to measure 
the heights of the columns. This instrument (Fig. 4.) is composed of a thick ruler 
of brass A B on which slides loosely a piece of the same metal M N P R S carry- 
ing at its two extremities R and S two supports or Ys, in which turns the micro- 
meter O O' having a horizontal wire in the focus. To the micrometer is hung a very 
sensible spirit level, the graduated scale of which serves to adjust the optical axis. 
The piece of copper MNP R Sil capable of the two motions ; a quick one, by the 
hand, when the side screw C is loosened, and a slow one, by means of the regulating 
