bulb under the action of an enormous potential: _this_ will be the
principle of the vacuum pump of the future. For the present, we must
secure the best results we can with mechanical appliances. In this
respect, it might not be out of the way to say a few words about the
method of, and apparatus for, producing excessively high degrees of
exhaustion of which I have availed myself in the course of these
investigations. It is very probable that other experimenters have used
similar arrangements; but as it is possible that there may be an item
of interest in their description, a few remarks, which will render
this investigation more complete, might be permitted.

[Illustration: FIG. 30.--APPARATUS USED FOR OBTAINING HIGH DEGREES OF
EXHAUSTION.]

The apparatus is illustrated in a drawing shown in Fig. 30. S
represents a Sprengel pump, which has been specially constructed to
better suit the work required. The stop-cock which is usually employed
has been omitted, and instead of it a hollow stopper s has been fitted
in the neck of the reservoir R. This stopper has a small hole h,
through which the mercury descends; the size of the outlet o being
properly determined with respect to the section of the fall tube t,
which is sealed to the reservoir instead of being connected to it in
the usual manner. This arrangement overcomes the imperfections and
troubles which often arise from the use of the stopcock on the
reservoir and the connection of the latter with the fall tube.

The pump is connected through a U-shaped tube t to a very large
reservoir R_1. Especial care was taken in fitting the grinding
surfaces of the stoppers p and p_1, and both of these and the mercury
caps above them were made exceptionally long. After the U-shaped tube
was fitted and put in place, it was heated, so as to soften and take
off the strain resulting from imperfect fitting. The U-shaped tube was
provided with a stopcock C, and two ground connections g and g_1--one
for a small bulb b, usually containing caustic potash, and the other
for the receiver r, to be exhausted.

The reservoir R_1 was connected by means of a rubber tube to a
slightly larger reservoir R_2, each of the two reservoirs being
provided with a stopcock C_1 and C_2, respectively. The reservoir R_2
could be raised and lowered by a wheel and rack, and the range of its
motion was so determined that when it was filled with mercury and the
stopcock C_2 closed, so as to form a Torricellian v