ttw.] porcelain air thermometry. 193 
haust the air from the closed auxiliary stein by means of a Sprengei 
nip. The correction follows from the measured amount of air ex- 
usted. I have used the compensator advantageously in the constant, 
lume method by connecting it with a special small manometer of its 
'n. The principle for the reduction of the measurements, by means 
this apparatus follows easily when it is remembered that the cotn- 
nsator is in the practical instance an air thermometer, the volume of 
e bulb of which is zero. Applying therefore the general equation 
) above, 1 have at once for the air thermometer 
v(HK—hk)+v'{EK , —hW)+v"{HK"—hW')=0 .... (8) 
id for the compensator 
v'(H 1 K'-h 1 ]c')+v"{H 1 K"-h 1 Tc")=0 .... (9) 
tiere Hi and h { are the tensions which correspond in the compensator 
II and h of the air thermometer, and where K,K', . . . and A;, A/, . . . 
e abbreviations for the functions 
1+aT 9 l+aT n ' 1 + otf' 1 + at 1 
qjectively. If the first equation be multiplied by H A and the second 
JET and the resulting equations be then divided by vHi there follows 
(HK^hJc)+(^R-^-h^~¥+^¥ f \==0 . . . (10) 
id since Jc , =Jc ,/ very nearly 
(HK-hl-)^(H^~h)]c fV ^±^-=0 . . . (11) 
V ±Li J V 
r which, with an error less than 0.2 per cent, at 2,000°, T may be 
duced as follows : 
T— 
a-/i 
1 * \_H_ _ v \ JI } eJ\] 
[_E l — —\-HrHj h ] 
>,re the variable temperature T is eliminated. 
By the use of the compensator the discrepancies due to porosity 
porcelain, to fissures aud other irregularities, may be partially obvi- 
h1. It may be made to subserve another important purpose, viz, to 
sasure v'+v", the true volume of the stem and capillary connec- 
ts. To do this it is merely necessary to connect the compensator 
th the manometer, as has been explained with reference to the air 
3rmometer, in such a way, however, that the variable space above the 
Bull. 54 — -13 (847) 
