AMPLIFICATION OF THE LAW OF POISEUILLE 31

fluidity of the surface film, whether it be surface tension, absolute
pressure, positive or negative polarization, static electricity, or
magnetism may therefore affect the amount of flow. And these
effects when detected experimentally would undoubtedly be
attributed to slipping or to the overcoming of external friction.
So while we might expect the effect of slipping to be more pro-
nounced in cases where the liquid does not wet the surface, it is
quite possible that even when the liquid does wet the surface,
the fluidity of the liquid near the surface is not identical with
that within the body of the liquid.

On the other hand, it is important to remember that the
thickness of the layer of liquid affected by the forces of adhesion,
with which we are here chiefly concerned, is only molecular.
Even with mercury in a glass tube, the thickness of the .layer of
air seems to be of molecular dimensions. One may get an idea of
the upper limit to this thickness by the following experiment.
A thread of mercury was placed in a narrow capillary so that
the air surface would be relatively large. Taking care that no
air-bubbles were present, the length of the thread was measured
with a dividing engine, in a determined part of the tube. The
tube was exhausted from both ends simultaneously and the
thread moved back and forth in order to sweep out the supposed
layer of air. When the mercury was finally brought back
to its former position no decrease in length could be detected.
In order to have slipping under ordinary conditions of measurement
it would appear that the surface film must be of very much more
than molecular thickness or else it must have practically infinite
fluidity. In view of the strong adhesion1 between all liquids and
solids it seems improbable that the particular layer of liquid in
contact with the solid should show an amount of flow which is
comparable in amount with that of all of the other practically
infinite layers of liquid.

Nevertheless if the value deduced by Helmholtz for water
on a metal surface be correct, X = 0.23534, the effect of slipping
ought to be readily observed. According to Whetham (1890), if
we take R = 0.051, Eq. (12) becomes

V = 7 117.67 X 10~6
0776

LC/. Duclaux, 1872.