COLLOIDAL SOLUTIONS 213
and the structure gradually collapses. But showing the proper-
ties of a true solid, it remains under tension, and when placed
again in water, it swells to approximately its former size, but not
indefinitely, as shown by Bancroft.
Increase in concentration of the internal phase very naturally
increases the viscosity of the colloidal solution. The addition
of non-electrolytes generally affects the viscosity in the way that
we would expect from the change produced in the fluidity of
the external phase. Since the colloid may unite with the water
to form hydrates or with the non-electrolyte, we should expect
exceptions to the quantitative application of this rule. Electro-
lytes have a similar effect on the viscosity of emulsion colloids,
potassium nitrate, ammonium nitrate, and potassium chloride
which increase the fluidity of water also increase the fluidity of
gelatine solution according to the measurements of Schroeder
(1903). Sodium sulphate, ammonium sulphate, magnesium sul-
phate and lithium chloride depress the fluidity. Acids and
alkalies however first lower the fluidity and then raise it. For a
more adequate account of this complicated subject the reader is
referred to the original papers, Schroeder, Pauli, etc.
It has often been a cause for wonder that a gel which has con-
siderable rigidity offers hardly more resistance to diffusion than
does pure water. We merely cite the names of Graham (1862),
Tietzen-Henning (1888), Voightlander (1889), and Henry and
Calugareanu (1901), giving a single observation from Voight-
lander to the effect that a 1 per cent solution of sodium chloride
in a 1, 2, and 3 per cent solution of agar gave a diffusion constant
of 1.04, 1.03, and 1.03 respectively. Similarly Ludeking (1889),
Whetham (1896), Levi (1900) Garrett (1903) and Hardy (1907)
have found that the conductivity of solutions remains constant
during gelatinization.
To understand these peculiarities, it is necessary to consider
the phenomenon of seepage of a fluid through a porous material.
Suppose, for example, that we consider a single pore; we must
assume that since it is a tube of capillary dimensions, the flow
must follow the law of Poiseuille and be proportional to the fourth
power of the radius of the pore. The question arises, "What will
be the effect upon the volume of flow of substituting for the single
pore a number of smaller pores whose total pore opening is the