AMPLIFICATION OF THE LAW OF POISEVILLE

21

To the present writer it seems probable that the kinetic energy
correction is truly constant for all tubes which are perfect cylin-
ders. Irregularities in the bore of the tubes will, however, have
very great influence in altering the amount of the correction,
since the correction, cf. Equation (7), depends upon the fourth
power of the radius of the tube. The shape of the ends of the
capillary has already been referred to in this connection, but it
seems preferable to consider the effect of the shape of the ends
of the tube as quite distinct from the kinetic energy correction.
There has been a tendency among many recent experimenters
to overlook the kinetic energy correction altogether, which is
quite unjustifiable. We have indicated that it is not practicable
to make the correction negligible. The only course open seems
therefore to be to select a capillary which has as nearly as possi-
ble a uniform cylindrical (or elliptical) cross-section, to assume
that m for such a tube has the constant value of 1.12, but to
arrange the conditions of each experiment so that the kinetic
energy correction will not exceed 1 or 2 per cent of the viscosity
being measured. In this case an error of several per cent in the
value of the constant will not affect the result, unless an accuracy
is desired which is higher than has yet been attained. If such an
accuracy is desired the value of m should be found for each tube
by the method of Knibbs which has been discussed above, or by
the method employed by Bingham and White (1912), which will
be described below in dis-
cussing the alteration in the
lines of flow at the ends of
the tube.

Correction for Phenomena
of the Flow Peculiar to the
Ends of the Tube.—If two
tubes of large diameter are
connected by a short capillary, the lines of flow will be as
represented in Fig. 3, the direction of flow being readily visible
in emulsions, suspensions, or when a strongly colored liquid is
allowed to flow out from a fine tube in the body of colorless
liquid near the entrance to the capillary, as was done by
Reynolds (1883). In the reservoir at the entrance A there is
apparently no disturbance until the opening of the capillary is

FIG. 3.-

-Diagram to illustrate viscous
flow.