Youne— Vapour-pressure of a pure Liquid at constant Temp. 93 
determinations of vapour-pressure at the lower temperatures, but 
may be expelled when the tube is strongly heated. 
When the vapour in the experimental tube has been completely 
condensed, the more complete the absence of air the greater must 
be the difficulty in inducing vaporisation again; and it is probable 
that if the liquid and glass were absolutely free from permanent 
gas, the difficulty would be insurmountable except at relatively 
very high temperatures. Great difficulty was actually experienced 
in most cases in starting vaporisation; and this may be taken as a 
proof that the removal of air was nearly complete. 
Between the extreme cases of (a) non-volatile and (6) gaseous 
impurities, we must shortly consider those of impurities, (c) less 
volatile and (d) more volatile than the liquid under examination. 
The influence of such impurities is similar to, but less marked than, 
that of non-volatile and gaseous substances respectively. ‘The less 
volatile impurities affect the pressure chiefly at the largest volumes, 
the more volatile chiefly at the smallest; the greater the difference 
between the volatility of the impurity and that of the pure 
substance, the more closely will the effect resemble that produced 
by the non-volatile and gaseous impurities respectively ; the smaller 
the difference in volatility, the more closely will the form of the 
curve representing the observed pressures approach that of a gently 
sloping straight line. 
Better results will in all cases be obtained by slow vaporisation, 
if possible without ebullition, than by condensation. 
2. Vaporisation of Mercury.—The vapour-pressure of mercury 
is negligible at low temperatures, but increases more and more 
rapidly us the temperature rises; it is only 0°27 mm. at 100° and 
16°81 at 200°, but at 800° it is 248°6 mm. Any error due to the 
vaporisation of mercury will, therefore, be of greatest importance 
at the highest temperatures. ‘That the diffusion of mercury-vapour 
through a highly compressed gas is an exceedingly slow process is 
well shown by the results obtained with isopentane (Proc. Phys. Soc., 
1895, vol. xiii., p. 621; and Zeits. Phys. Chem., vol. xxix., p. 210). 
Diffusion through a liquid must be still slower; and, considering the 
short time required for the determination of the vapour-pressure 
of a liquid at any one temperature, it is certain that the partial 
pressure of the mercury-vapour must be a very small fraction of the 
