222 Proceedings of the Royal Society of Edinburgh. [Sess. 
There are, of course, limiting conditions to the above experiments. 
Thus, if the angle of the conical electrode be increased indefinitely until 
the cone degenerates into a plane surface, the preference of the spark 
for that electrode would naturally be expected to diminish ; and so it 
is found in actual fact. 
It may also be noted that when a discharge followed by perforation 
of the dielectric has once been induced by motion of the dielectric* the 
difficulty of provoking a second spark discharge becomes still greater than 
formerly : on examining the apparatus in the dark it is usually easy to 
discern the faint lines of an almost silent discharge threading the hole 
from electrode to electrode. 
II. Phenomena of Repulsion . — Doubrava has pointed out* that in 
general any dielectric plate set between discharging electrodes experiences 
a definite force tending to displace it in the direction of one of the 
electrodes. He declares, however, that “ the experiment only succeeds 
when the electrodes are sufficiently separated to give a crackling brush 
discharge.” This is not the case. Repulsion effects on all kinds of plates 
are observable even when the electrodes are sparking freely (through the 
plate, or round its edge) or discharging by the so-called “ silent ” discharge. 
[See § IV. below for some notes on the form of these discharges.] In observ- 
ing the effect in the present instance, dielectric plates (16 cm. x 16 cm. 
is a convenient size) were suspended midway between the electrodes 
by a fine silk bifilar suspension from insulators on the ceiling. The 
suspension was about 15 feet long, a length sufficient to permit of even 
very small forces producing sensible deflections, and the plate itself was 
screened off from draughts (fig. i.). In the case of metal plates, which were 
also found to exhibit repulsion phenomena, one thread can conveniently be 
replaced by a fine copper filament leading to an electroscope screened from 
electrical action by fine wire gauze. The behaviour of the leaves of the 
electroscope then indicates the condition of the plate as regards charge. 
Dielectric plates were removed by tongs when it was desired to investigate 
their charge. No difference in the results was ever experienced on earth- 
ing one of the electrodes. 
In general it was found that when the electrodes were of similar shape 
and size, discharge between the electrodes resulted in a strong repulsion 
of the plate in the direction from the + to the — . [If the potential was 
not high enough to permit of discharge of any nature, then the plate could 
easily be induced to perform regular vibrations from pole to pole, after the 
* Wiedemann Annalen, Bd. viii., p. 476, “ Ueber die Bewegung von Platten zwischen den 
Elektroden einer Holtz’schen Maschine.” 
