be quite irregular. All these variations I have witnessed;
and when it is considered that the air in the vessel, and also the glass of
the vessel, can take a momentary charge, it is easy to comprehend their
general nature and cause.

* * * * *

1506. Similar experiments to those in air (1485. 1493.) were made in
different gases, the results of which I will describe as briefly as
possible. The apparatus is represented fig. 131, consisting of a bell-glass
eleven inches in diameter at the widest part, and ten and a half inches
high up to the bottom of the neck. The balls are lettered, as in fig. 130,
and are in the same relation to each other; but A and B were on separate
sliding wires, which, however, were generally joined by a cross wire, _w_,
above, and that connected with the brass conductor, which received its
positive or negative charge from the machine. The rods of A and B were
graduated at the part moving through the stuffing-box, so that the
application of a diagonal scale applied there, told what was the distance
between these balls and those beneath them. As to the position of the balls
in the jar, and their relation to each other, C and D were three and a
quarter inches apart, their height above the pump plate five inches, and
the distance between any of the balls and the glass of the jar one inch and
three quarters at least, and generally more. The balls A and D were two
inches in diameter, as before (1493.); the balls B and C only 0.15 of an
inch in diameter.

Another apparatus was occasionally used in connection with that just
described, being an open discharger (fig. 132.), by which a comparison of
the discharge in air and that in gases could be obtained. The balls E and
F, each 0.6 of an inch in diameter, were connected with sliding rods and
other balls, and were insulated. When used for comparison, the brass
conductor was associated at the same time with the balls A and B of figure
131 and ball E of this apparatus (fig. 132.); whilst the balls C, D and F
were connected with the discharging train.

1507. I will first tabulate the results as to the _restraining power_ of
the gases over discharge. The balls A and C (fig. 131.) were thrown out of
action by distance, and the effects at B and D, or the interval _n_ in the
gas, compared with those at the interval _p_ in the air, between E and F
(fig. 132.). The Table sufficiently explains itself. It will be understood
that all