32 
On the Measure of Temperature. 
was immersed in the oil with every care, so that the whole column of mer- 
curv should be covered by it : at this instant the fine point ot the air tube 
was sealed by a blow pipe, and the height of the barometer noted. The tube was 
now withdrawn, and removed to a different room, the temperature ot which 
was almost stationary. It was there fixed in a vertical position, in such a 
way that the extremity dipped into a bath of mercury perfectly dry. 1 he point 
being then broken off, the mercury rose till the equilibrium with the outward pres- 
sure was restored, the tube being left in this situation till it acquired the tempera, 
tlire of the room, indicated by a very sensible thermometer hung at a short distance. 
When this effect took place,’ the height of the column in the tube was measured by 
means of a vertical scale furnished with a vender. The height of the barometer 
w as noted, and the difference of these gave the elasticity of the cooled air. The 
tube was then withdrawn, every precaution being taken to retain the whole of the 
mercury that had risen into the' tube. The tube and mercury were weighed ; after- 
wards the empty tube; and lastly the tube filled with mercury- Deducting now from 
this last result,' the former two] we have the weights of two volumes of mercury, 
the one equal to that of the heated air, the other to that of the cooled,. From these 
weights, we conclude w-hat are the volumes, which are afterwards reduced to what 
they would have been under the same pressure ; as the elasticity of the cooled air was 
known, having been measured as above indicated, and that of the heated air was 
equal to the pressure of the atmosphere at the moment the tube was scaled*. 
In order to give a clear idea of the degree of confidence to be placed in the results 
which we have eatablished,it may not he without its use to enter into a little detail, 
as to the precautions adopted to ensure accuracy in each experiment. 
One of the principal difficulties -we have to contend with in such experiments, is 
that of establishing a perfect uniformity of temperature in a considerable mass of 
liquid, heated to two or three hundred degrees ulwve the temperature of the sur- 
rounding air This condition however may he perfectly attained, when the tempera- 
ture of experiment is that of ebullition of the liquid employed, as in that case it is 
necessarily fixed. But in every other case, the process of heating or cooling affect- 
ing different parts of the liqaid unequally, is opposed to the establishment of that 
uniformity. We are, however, of opinion,tliat the arrangement of our apparatus does 
in a great measure,obviate this particular objection, and for these reasons : 1st, the 
vessel of copper being fixed within the furnace, the whole forms a considerable 
mass which must cool very gradually, particularly when close upon Its maximum 
of temperature; and 2dly, the liquid being kept in a constant state of agitation, 
the heat must distribute itself pretty equally throughout it. But to remove every 
doubt that could possibly arise on this head, we have inserted two thermom eters 
in a horizontal position in the vessel, both being at the same height, ami then con- 
ducting the experiment in the usual manner, we never had occasion to observe 
more than two or three tenths of a degree between the two thermometers. 
Even if we suppose that all the points of the liquid stratum which surrounds 
the air tube were not exactly of the same temperature, the error would be less than 
might on a first view lie supposed ; for in the arrangement of the apparatus it ap- 
pears, that the bulb of the thermometer corresponds nearly to the middle point of the 
length of the tube, and consequently this instrument ought, in everv case to indi 
cate a temperature, differing hut little from the mean of the temperatures of the 
different parts of the tube. It was, in fact, this consideration that induced us to 
adopt a cylindrical tube, in preference to a bulb of anv other form. We mav 
mention here also as connected with the present subject, the necessity of having the 
* AH our experiments made in the manner just described, have been calculated 
after the following formula. a c 
Calling P the weight of a volume of mercury equal to that of the heated air T the 
temperature of that air reckoned on the mercurial thermometer Hi, !i lr \ • •» 
P'T'H'the same quantities for the cooled air . also tak ug l n ’he the v aln S 
a,r at 0= which has been heated to T, but ui’ider uniform pressute ar d d ,hl 
mean expansion of glass between the temperatures T and l', we have' ^ ^ ^ 
w_ P H ( 1+d (T — T') ) (1 + 0.00375 T'l 
P' H' — 
from which expression we mav see, that an air thermometo,. . , - 
pansiou of glass, would show for any temperature T of thf 00rre ° te , <1 for the ex- 
ter a number of degrees which may be thus expressed s ' mercurial thermome- 
p J£ 
*" (1+dT—T') ) (266.6/ +T')_266.67. 
