92 FLUIDITY AND PLASTICITY
changes are of no interest to us at this point, but the fact is very
important for us that the fluidity-volume concentration curves
are much more nearly linear than the viscosity concentration
curves. To become convinced of this the reader should plot
the fluidity concentration curves of several of the mixtures given
by Wijkander (1878), Linebarger (1896), Dunstan et cet. (1904)
and others and compare them with the viscosity curves given
by those authors. A few of these curves are given in Fig. 35.l
Kendall (1913) has gone over the whole range of available data
and finds that the percentage deviation between the observed
and calculated values is 11.1 for the viscosity curves and 3.4 for
the fluidity curves.
FLUIDITY AND TEMPERATURE
The conclusion that fluidities are additive has far-reaching
consequences, so that there arise tests for the conclusion which
were at first quite unsuspected. For example, it is evident that
the reasoning which has been found to hold for mixtures of fluids
must also hold for mixtures of the same fluid at different tempera-
tures; for a fluid at any temperature may be thought of as a
mixture of appropriate amounts of portions of the fluid main-
tained at the extreme temperatures. Hence, we are led to the
hypothesis that the fluidity-temperature curves of pure fluids
should normally be linear, and the viscosity-temperature curves
hyperbolic. This relation cannot hold through a change of state
because a new cause of viscosity then enters in. Furthermore the
fluidity of liquids is closely related to their volumes, as we shall
see later, and the volumes of liquids do not generally increase in
a linear manner with the temperature. Then, too, association and
dissociation may play a disturbing factor, so that as in mixtures,
a perfect verification can scarcely be expected. In Fig. 36 there
are given the fluidity-temperature and viscosity-temperature
curves of mercury and of water from 0 to 100°C. Both of the
1 We have here but a rough test of the truth of the hypothesis that
fluidites are additive in homogeneous mixtures, because the fluidites of the
components are too close together, all of the components are certainly not
inert, and volume concentrations should have been studied. More rigorous
tests of the hypothesis will be made after the law of Batschinski has been
considered.