1891 - 92 .] Prof. Tait on the Kinetic Theory of Gases. 
33 
constants in my formula, and thence to improve my determinations 
of the critical temperature, pressure, and volume. 
In § 71 of Part IV. I arrived at the conclusion that “in a 
liquid the temperature is no longer measured by E [the part of 
the kinetic energy which is independent of the molecular forces], 
but by E + c, where c is a quantity whose value increases steadily, 
as the temperature is lowered, from the value zero at the critical 
point.” For numerical data to test this conclusion, I study a cycle 
formed from the critical isothermal and any lower one, and two 
lines of equal volume, corresponding to those of the liquid and the 
saturated vapour when in equilibrium at that lower temperature. 
The change of energy in passing from one of these limits of volume 
to the other is found to be less for the critical isothermal than for 
any lower one. Thus the mean specific heat at constant volume, 
for the range of temperature employed, is less in the vapour than in 
the liquid. But from the equation, which is found to satisfy very 
closely the data for the isothermals of the gas for some 70 degrees 
above the critical point and of the vapour for 30 degrees below 
that point, it aj^pears that the specific heat at constant volume is 
sensibly constant within these limits. [At 100° C. and upwards, 
cLiG iJjT) 
it appears that falls off ; so that — | is negative, and the specific 
Civ Clt^ 
heat at constant volume is therefore, even in the gas, greater for 
smaller volumes. But this does not seriously affect the above 
statement.] Hence, at any volume less than the critical volume, 
more heat is required to raise the temperature 1 degree when the 
substance is wholly liquid than when it is gaseous. This completely 
justifies the statement quoted above, provided that we assume the 
properties of the liquid and gas to merge continuously into one 
another at the critical temperature ; but, unfortunately, the data are 
not sufficient to give more than very rough estimates of the value 
of the quantity c there spoken of. 
I am at present engaged in endeavouring to obtain more exact 
values of the constants in my equation, in order to improve my 
estimates. Thus the numbers which follow may have to undergo 
some modifications, but there seems to be no reason for thinking 
that these are likely to be serious. 
If Vp V .2 be the respective volumes of the saturated vapour, and 
17 / 3 / 92 . 
VOL. XIX. 
c 
