VISCOSITY AND FLUIDITY 91

ing observation "As a rule, the greater the difference between the
viscosities of the pure liquids, the greater is the difference between
the calculated and the observed values of the mixtures." The
second conclusion is confirmed by the sudden drop in the fluidity
of a mixture as it is cooled below its critical-solution tempera-
ture. This has often heen noted and commented upon, and will
be discussed more fully at a later point. In xindercooled liquids

.01 00

80

100

Weight percentage of second component of mixture.
FIG. 35.—1. Fluidity curve of nitrobenzene and etfcyl acetate at 25°; 2.
Fluidity curve of ethyl alcohol and acetone at 25°; 3. Fluidity curve of benzene
and ethyl acetate at 25°; 4. Fluidity curve of benzene and ethyl ether at 25°;
5. Fluidity curve of carbon bisulphide and ethyl ether at 25°; 277. Viscosity curve
(dotted) of ethyl alcohol and acetone at 25°. Were viscosities additive, this
curve would be linear (dashes).
and other Tery viscous substances it has been often noted that the
viscosity curves have a, very high degree of curvature, at least
during a part of their course. This is in haraion7 with the fifth
and sixth conclusions.
Were the fluidity-volume concentration curves invariably
linear, it would constitute an experimental verification of our
fourth conclusion. Unfortunately for this purpose, the fluidity
concentration curves are rarely, if ever, perfectly linear, for the
reason that has been, indicated; viz.- that there is perhaps nearly
always some molecular change on mixing, even though very
feeble, and to this change fluidity is very sensitive. These