acuum in it when
raised, it could be lifted so high that the mercury in reservoir R_1
would stand a little above stopcock C_1; and when this stopcock was
closed and the reservoir R_2 descended, so as to form a Torricellian
vacuum in reservoir R_1, it could be lowered so far as to completely
empty the latter, the mercury filling the reservoir R_2 up to a little
above stopcock C_2.

The capacity of the pump and of the connections was taken as small as
possible relatively to the volume of reservoir R_1, since, of course,
the degree of exhaustion depended upon the ratio of these quantities.

With this apparatus I combined the usual means indicated by former
experiments for the production of very high vacua. In most of the
experiments it was convenient to use caustic potash. I may venture to
say, in regard to its use, that much time is saved and a more perfect
action of the pump insured by fusing and boiling the potash as soon
as, or even before, the pump settles down. If this course is not
followed the sticks, as ordinarily employed, may give moisture off at
a certain very slow rate, and the pump may work for many hours without
reaching a very high vacuum. The potash was heated either by a spirit
lamp or by passing a discharge through it, or by passing a current
through a wire contained in it. The advantage in the latter case was
that the heating could be more rapidly repeated.

Generally the process of exhaustion was the following:--At the start,
the stop-cocks C and C_1 being open, and all other connections closed,
the reservoir R_2 was raised so far that the mercury filled the
reservoir R_1 and a part of the narrow connecting U-shaped tube. When
the pump was set to work, the mercury would, of course, quickly rise
in the tube, and reservoir R_2 was lowered, the experimenter keeping
the mercury at about the same level. The reservoir R_2 was balanced
by a long spring which facilitated the operation, and the friction of
the parts was generally sufficient to keep it almost in any position.
When the Sprengel pump had done its work, the reservoir R_2 was
further lowered and the mercury descended in R_1 and filled R_2,
whereupon stopcock C_2 was closed. The air adhering to the walls of
R_1 and that absorbed by the mercury was carried off, and to free the
mercury of all air the reservoir R_2 was for a long time worked up and
down. During this process some air, which would gather below stopcock
C_2, was expelled from R_2 by low