300 
Proceedings of Royal Society of Edinburgh. [sess. 
connected at its lower end to the bent stem of the air-bnlb and 
also by a flexible tube to the reservoir C. D is the constant 
volume mark on the stem of the air-bulb. The part of the instru- 
ment which is filled with mercury when the air is at 0° C. is 
shaded in the diagram. 
§ 3. To make an observation of temperature the mercury is 
adjusted to the mark D by raising or lowering the reservoir. The 
height of the column in B above the level of the mark D now 
gives the total pressure of the enclosed air. The air-bulb and 
pressure-gauge may be cut off entirely from external pressure by 
closing the stop-cock between the pressure-gauge and the reservoir. 
If the absolute temperature corresponding to any given value of 
the pressure be known, we can calculate that corresponding to any 
other pressure by simple proportion. 
Thus, in order to graduate the thermometer in degrees of tem- 
perature, it is only necessary to find the total pressure for any one 
temperature; that of melting ice, which we may take as 273° on 
the absolute scale of temperature, is the most convenient. To fix 
this point on the scale, the bulb is immersed in melting ice and 
the mercury adjusted to the volume mark by raising or lowering 
the reservoir. The height, h, of the mercury column in B above 
the level of the mark now corresponds to 273° absolute. But, 
since pressure and temperature are in simple proportion, the de- 
grees at all parts of the scale are of equal length; thus 7^/27 3 is 
the length of one degree. 
§ 4. It will be seen that the reservoir is merely a device for sup- 
plying mercury at the proper pressure to fill up the gauge-tube, 
and that it may be replaced by any suitable mechanical device, 
such as a cylinder full of mercury in which a piston is forced 
down by a screw. It is obvious that the height of the mercury in 
B above its level in the open reservoir C always corresponds to the 
atmospheric pressure at the moment. This portion of the apparatus 
is in fact a barometer if the reservoir C be open, but it is not in the 
least essential that it should be so. 
§ 5. The atmospheric pressure is, after all, merely an accident ; 
and its constant recurrence in all calculations on the pressure 
