104 FLUIDITY AND PLASTICITY
If for simplicity we imagine the solid particles of a suspension
to be all united into sheets parallel to the direction of flow, as the
shaded areas in Fig. 33 then it is evident that the flow will be the
sum of the flows of the unshaded areas, i.e., the fluidities will be
strictly additive or
since <pi is practically zero.
But if these solid sheets were broken up into fragments, one of
which is shown as the cross-section of a cube at F in Fig. 40,
the deformation of the liquid would tend to change the form of
the cube into that of a parallelepiped as shown at G, but as the
solid is rigid, this cannot take place; so that the shearing force
can only rotate the cube around its center as shown at H. But
the failure of the solid to change its
*-----D' p" pW—^---------^ shape with the flow of the liquid
will necessitate transverse motions
in the liquid by way of readjust-
ment, hence the viscosity of a
suspension will always be greater
than it would be were the fluidities
strictly additive. If, as we believe can be proved to be the
case, the amount of transverse motion in a suspension is pro-
portional to the number of suspended particles of a given size,
and for each particle the amount of transverse motion bears a
constant ratio to the amount of shear, it will follow that the
fluidity curves of suspensions must be linear, as has already
been shown to be generally true (qf. p. 55 and Fig. 19).
Enough evidence has been given to indicate that the theory of
the subject and the most diverse sorts of experimental data are
in accord in supporting the fundamental hypothesis that fluidities
are normally additive in homogeneous mixtures and fine sus-
pensions, but not in heterogeneous mixtures. Much additional
evidence could be given, but not without taking up subjects out
of their natural order. This evidence will appear as we proceed.
For further confirmation of these views cf. Drucker and Kassel
(1911), White (1912).
James Kendall (1913), working in the Nobel Institute of Phys-
ical Chemistry under the direction of Professor Arrhenius,
concluded that "the logarithmic viscosity (or fluidity) of a solu-