92 
Proceedings of the Royal Society 
of fluid in one state throughout. He secured geometrical, hut not 
physical, continuity. For, as Clerk-Maxwell showed, one part of 
his curve makes pressure and volume increase simultaneously, a 
condition essentially unstable. The idea which occurred to me was, 
while preserving geometrical continuity, to get rid of the region of 
physical instability, not (as I had suggested in my former Note) by 
retaining Thomson’s proposed finite maximum and minimum of 
pressure in the isothermal, while bringing them infinitely close 
together so far as volume is concerned, and thus restricting the 
unstable part of the isothermal to a finite line parallel to the 
pressure axis ; but, by making both the maximum and minimum 
infinite. Geometrical continuity, of course, exists across an asymptote 
parallel to the axis of pressures; so that, from this point of view there 
is nothing to object to. On the other hand there is essentially 
physical discontinuity, in the form of an impassable barrier between 
the vaporous and liquid states, so long at least as the substance 
is considered as homogeneous throughout. 
It appeared to me that here lies the true solution of the difficulty. 
As we are dealing with a fluid mass essentially homogeneous through- 
out, it is clear that we are not concerned with cases in which there 
is a molecular surface-film. 
Suppose, then, a fluid mass, somehow maintained at a constant 
temperature (lower than its critical point), and so extensive that its 
boundaries may be regarded as everywhere infinitely distant, what 
will be the form of its isothermal in terms of pressure and volume 1 
Two prominent experimental facts help us to an answer. 
First. We know that the interior of a mass of liquid mercury can 
be subjected to hydrostatic tension of considerable amount without 
rupture. The isothermal must, in this case, cross the line of 
volumes ; — and the limit of the tension would, in ordinary language, 
be called the cohesion of the liquid. I am not aware that this result 
has been obtained with water free from air; but possibly the experi- 
ment has not been satisfactorily made. The common experiment in 
which a rough measure is obtained of the force necessary to tear a 
glass plate from the surface of water is vitiated by the instability of 
the concave molecular film formed. 
Second. Aitken has asserted, as a conclusion from the results of 
direct experiment, that even immensely supersaturated aqueous 
