THE ELECTRIC AND LUMINIFEROUS MEDIUM. 
273 
supplement on mechanical principles an explanation like the present one is not so 
much any accession of logical rigour on ordinary thermodynamic premisses, as some 
precise notion of what is involved, as regards detailed molecular dynamics, in equality 
of temperature. In the present differential procedure no assumption has been made 
on that head ; and no inference that osmotic pressure is, like gaseous pressure, due to 
simple molecular bombardment is warranted. When a theoretical basis is thus found 
for VAR T Hoffs principle, the laws of the molecular influence of dissolved substances 
on the freezing point and vapour tension of very dilute solutions of course go along 
with it. 
58. It may be objected that the ajDpiication of the principle to ionized solutions 
would compel us to admit the possible theoretical existence of a gas consisting of 
ions : but that is not really so, because the argument only compares one state of 
dilution with another. Yet on the other hand there is the hypothesis, supported bv 
Beuhl s work on optical equivalents, that under certain circumstances oxygen is a 
tetrad element, so that the molecule II 3 O can take up sufficient ions to form another 
saturated molecule of type H^ = 0 = X, and that therein lies the cause of the 
regular ionization cm rent produced by solution in water (the 10 ns X being free only 
when passing from one such combination to another), as contrasted with the irregular 
ionization of free gases. Changes of valency in an element remain unexplained, but 
theii occurience is now usually accepted as matter of fact.The function of the 
osmotic diaphragm is merely passive, to prevent mixture by diffusion and consequent 
loss of mechanical availability. In a mutual solution of two substances, it is usually 
only the one that is present in large excess that gets through the diaphragm in 
purity : if it should prove to be a general law that the dialyzing action is only 
complete when there is such large excess, it would be strong evidence for the view 
that the molecules dissolved in it form the nuclei of loose molecular complexes which 
are too large and permanent to get through, while the free solvent in which the 
molecules are not thus grouped is not so hindered. 
When a solution is made more and more dilute, there comes a stage when it would 
The connexion between tlie various pbenomena may be pictured in neutral terms, as Poyxting lias 
recently done (‘ Phil. Mag.,’ Oct. 1896), starting from a hypothesis that pressure increases the 
molecular mobility ” of a fluid according to an assumed law equivalent to van’x Hoff’s jn’iuciple. In 
order to evade the hypothesis of partial dissociation in salt solutions, he restricts the sphere of action of 
a- dissolved undissociated molecule to one or tw’o or three definite molecules of the solvent, leading’ to 
correspondingly different amounts of osmotic pressure; thus a temporary chemical combination is dealt 
with instead of, or it may be in addition to, an extended sphere of influence. But the considerations 
given above show that it is the number of spheres of influence that is really effective, so that if there is 
chemical combination it must be in part with dissociated ions as in Bruhl’s view. On the special 
assumptions iuA'olved in the extension of the methods of gas-theory to liquids, Boltzmaxx (‘ Zeits. fiir 
Phys. Chemie ’ vi, p. 478) has offered a demonstration (approved by Lorektz) of the law of osmotic ‘ 
pressure, which seems to refer it to molecular bombardment, and require that the mean energy of 
translation of a molecule shall be the same in the liquid state as in the gaseous state at the same 
temperature. 
VOL. CXC.—A. 2 R 
