188 MEASUREMENT OF HIGH TEMPERATURES. [bull. 54. 
shows the position of the one horse-power gas engine, to the left of 
which is the small centrifugal blower and to the right the bellows. The 
latter are worked by a short crank attached to the axle of the fly-wheel, 
and the blast obtained is small in quantity but high in pressure. The 
furnace-table completely surmounts the engine and carries on the left 
side (under D) a series of three or more small furnaces, resting on a 
special table of their own. These furnaces, fed by gas and the pressure- 
blast of air from the bellows, contain crucibles of the form (Figs. 14 and 
15) above, and ai»e used for boiling-point measurements.* Thermoele- 
ments are in serted from below and held in position by a suitable clamp. 
Observations and manipulations are made in the way carefully detailed 
in Chapter II. On the right side of the furnace-table (under G) is fixed 
the revolving muffle-furnace, the parts of which have just been de- 
scribed, and are therefore easily recognized. 
Quite to the right of the engine (under B), and screened from it by a 
wide board, is the manometric apparatus of the air thermometer (cf. 
page 209, below). The bulb and manometer are connected by a platinum 
capillary tube. A similar platinum tube connects the compensator with 
its manometer. Above the furnaces is an iron tube, through which the 
small furnaces may either be supplied with hydrogen or may be ex- 
hausted, as desired, the other end of this tube being simply attached to . 
appropriate gasometric apparatus. 
A set of air-thermometer bulbs are shown on a little shelf under A. 
The two vertical bulbs are of porcelain, being the forms Figs. 32 and 33 
above. The oblique bulb is a re-entrant air thermometer of glass. 
CONSTANT-VOLUME THERMOMETER — METHOD OF CALCULATION. 
The general equation. — It has already been suggested that it is expe- 
dient to choose such methods of thermometry in which the degree of 
constancy of the zero reading before and after the high-temperature cali- 
bration may afford a guaranty of the reliable character of the results 
obtained. This test can well be applied when the constant volume 
method is used. If the work be done with low pressures (<1 atmos- 
phere) it is also possible to adjust the quantity of air that at such tem- 
peratures at which porcelain tends to become viscous the tension of the 
inclosed heated gas may approach that of the atmosphere. 
The most general expression for the variables involved in any case of 
air thermometry may be rigorously put 
^"('SS-'SSXH <" 
where F, V, Y" . . . are zero volumes whose special temperatures are 
T, T', T" . . . for the pressure IT, and £, t', t" . . . for the pressure h, 
where A is proportional to the excess of volume at Tover that of the 
fixed mass of gas at t, and where a is the coefficient of expansion of the 
(842) 
