1896-97.] Lord Kelvin on Vapour Pressures of Liquids. 431 
liquid from either bottle ; and, unless explicitly mentioned, will 
not be included among the stopcocks referred to in what follows. 
It will generally be convenient to make the quantities of the two 
liquids introduced such that they stand at as nearly as may be the 
same levels in the two bottles, as indicated in the drawing. 
§ 3. Operate now on one only of the liquids until it is got into 
equilibrium, with its upper level at some point in its glass tube, 
and nothing but its own vapour between this surface and the closed 
stopcock immediately above it. To do this proceed as follows : — 
Close and keep closed the two stopcocks of the liquid not operated 
on, and work the air-pump with the other four stopcocks all open 
until an exhaustion, not quite as perfect as is possible, of the air 
over the liquid operated on is produced. 
§ 4. Then close the lower air-pump stopcock, and go on working 
the pump until the liquid in the tube ceases to rise further above 
its level in the bottle. Close the two stopcocks of this liquid. 
§ 5. Operate similarly on the other liquid. 
§ 6. Close now the lower air-pump stopcock, and equalise the 
pressures of air and vapour above the liquids in the two bottles by 
opening their neck stopcooks. If the levels of the liquids in the 
two columns are lower than convenient for observation, some air 
should be allowed very cautiously to run back from the air-pump 
into the two bottles through the lower air-pump stopcock. After 
doing this, repeat the operation of § 4 for each liquid. 
§ 7. The operation of § § 4, 5 must be continued long enough to 
di.stil out of the upper part of each liquid, in its glass tube, air or 
any foreign * volatile substance sufficiently to prevent any sensible 
pressure on the free surface other than that of the vapour of the 
solvent. 
§ 8. By proper thermal appliances, indicated by the dotted lines 
in the diagram, and the lamp under the upper bent metal tube 
(inserted merely as an indication that somehow the metal tube is 
to be always slightly warmer than the warmer of the two liquid 
surfaces, in order that there may be no condensation of vapour in 
jt), bring the upper surfaces of the liquids to any one temperature, 
or to two different temperatures. The difference of levels of the 
* Compare Ostwald, Physico-Chemical Measurements , translated by Walker 
(Macmillan, 1892), last paragraph, page 112. 
