314 Proceedings of Royal Society of Edinburgh. [sess. 
begins to expand. We shall thus call coefficient of absorption 
“ the rate at which the mass of the gas is absorbed by unit mass of 
the liquid per unit partial pressure ” ; by partial pressure is meant 
the pressure of the liquid mixture diminished by the vapour - 
pressure of the pure solvent. Up to a small distance from the 
critical point there is no harm in substituting for “ mass absorbed 
per unit pressure ” the ratio of mass absorbed and pressure, if only 
small pressures are considered. 
In considering the value of the coefficient of absorption in a 
special case, I shall use the vapour-pressure temperature diagram, 
for a complete discussion of which I must refer to former papers. 1 
The figure gives the general appearance of the diagram for two 
substances of widely different critical temperatures and vapour- 
pressures, in this case methyl chloride (solvent) and carbon dioxide 
(gas dissolved). It contains in addition to the two vapour-pressure 
curves of the pure constituents, ending at Cj and C 2 , the two 
critical points, condensation-curves for some of the mixtures. 
Each one of these belongs to a mixture of given composition ; the 
lower branch of the loop gives the pressures and corresponding 
temperatures at which the mixture in its lighter condition ( i.e ., as 
vapour) is in equilibrium with a liquid mixture of different com- 
position, the upper branch contains the points at which the 
mixture as a liquid is in equilibrium with a vapour. 
In our problem we have to deal with the latter, the upper 
branch : its vertical distance from the vapour-pressure curve of the 
solvent is what we have called the partial pressure of the gas, and 
the quotient of the (constant) mass of the gas which the particular 
mixture say of the lowest loop shown in the diagram contains and 
this partial pressure is the coefficient of absorption. Obviously then 
the coefficient of absorption is inversely proportional to the vertical 
distance of the upper branch of the loop and the methylchloride- 
curve. 
Owing to the peculiar way in which the upper branch of the 
loop bends round on approaching the critical curve, C 2 P C lt it 
will be seen that the partial pressure referred to will necessarily in 
the end diminish and therefore the coefficient of absorption in- 
crease. At low temperatures the partial pressure is low and the 
1 Thil. Mag., 40, p.> 175. 
