180 
MR. H. L. CALLENDAR ON THE PRACTICAL 
A somewhat similar Table is given by Thomson and Joule (‘ Phil. Trans./ 1854 ), 
calculated from a more complicated, empirical formula, which does not represent the 
results of experiment nearly so w T ell. It will be seen that the corrections are so 
small that it is not worth while to apply them, except in investigations of a higher 
degree of accuracy than any yet attempted. 
They are rather larger for a C0. 2 air thermometer, but still so small that a CO, air 
thermometer is practically correct up to 300°* (Pegnault) ; that is, the corrections are 
smaller than the uncertainty due to the expansion of the envelope or other causes; 
the magnitude of this correction varies directly as the initial pressure p 0 . 
Assuming equation (E), we ought to have for the compressibility of air at O' C. 
1 - r ■ = <£—0 o = -OO1173(--l), 
V oA \Po ) 
or, at 5° C., 
P v f P \ 
1-= ‘00111-1 approximately 
Wo \Po ) 
From Regnault’s formula,! if p >=2 metres and _p 0 = 1 metre of mercury, we find, for 
air at 4 '7 5° C., 
1—— = -00109. 
Po V 0 
But the equation (E) does not agree so well with his results throughout the whole 
range of his experiments, because it is only a rough first approximation to the 
departure of air from the law of a perfect gas, and is only applicable within moderate 
limits of pressure and temperature. At high temperatures dissociation would occur, 
and the formula would no longer apply. In the case of C0 2 the dissociation is quite 
measurable, even at 1000 ° C. 
The Sources of Error to which the Air Thermometer is liable. 
The first and most important source of error, for which no correction can be applied, 
is impurity in the dry air. Water vapour and C0 2 are most to be avoided, and the 
bulb of the air thermometer must be very thoroughly cleaned. The apparatus used 
for supplying pure dry air, used in the earlier experiments, appeared perfect, but has 
since been proved faulty. See Appendix, p. 222 . 
The air thermometers were dried by repeated exhaustion and re-admission of dry 
air at temperatures between 300° and 500° C. 
* ‘ Memoires de l’lnstitnt,’ tom. 21, p. 187. 
t ‘Memoires de 1’Institnt,’ tom. 21, p. 421. 
