214 MESSRS. W. SPOTTISWOODE ANT) J. FLETCHER MOULTON 
individual discharges. These average effects may remain constant notwithstanding 
great variability during the individual discharges. 
XII .—On unipolar discharges. 
We have seen that, inasmuch as the discharge from the air-spark terminal produces 
its special effect without any indistinctness or confusion close up to the opposite 
terminal, it would appear that the discharges from the two terminals are so far inde¬ 
pendent that the discharge may take place from one and the free electricity pass right 
through the tube to the immediate vicinity of the other without evoking a specific 
response from the latter terminal. And if each such discharge does in any way call 
forth from the other terminal a specific response, it must be so slight that it does not 
affect materially the electrical condition of the interior of the tube, or the effect which 
that condition produces on conducting systems outside the tube. And we have also 
seen that this independence implies that the electricity leaves the terminal from which 
it starts in consequence of the electric tension within that terminal, and only in a very 
subordinate degree in consequence of the correlative action at the opposite terminal. 
Lest these should seem to be too hastily drawn conclusions we will proceed to describe 
a class of phenomena which furnish very important evidence of their truth. 
If we take two exhausted tubes of the same general type, and connect one terminal 
of each with one terminal of a large Holtz machine, and connect their other terminals 
with the other terminal of the machine, interposing an air-spark (say in the positive 
circuit) so that the electricity has two alternative paths, the one through the one 
tribe and the other through the other, the air-spark being common to both paths, a 
very remarkable phenomenon will be witnessed (Plate 19, fig. 19). If the air-spark 
be of a suitable magnitude it will be found that one of the tubes is wholly traversed 
by the discharge, but that the other is occupied only by a luminous column extending 
from the positive terminal into the tube for a considerable portion of its length, and 
gradually tapering to a point. If the air-spark do not exceed a certain limit, depending 
upon the “ resistance ” of both tubes, there will be no luminosity at the other end of 
the tube, and no discharge through it. No effect will be produced upon the luminous 
column, nor on any portion of the discharge, by breaking the connexion with the 
distant terminal, showing, what the appearance of the column itself sufficiently indi¬ 
cates, that the discharge is unipolar or incomplete. Slight indications of blue haze are 
sometimes seen at the tip of this tapering column, due probably to some negative 
electricity gathered from the neighbourhood, but not directly discharged from the 
opposite terminal. The discharge is, in fact, one which passes into the tube but not 
with sufficient force to pass through it, and which accordingly returns by the way by 
which it entered. The cause of this recall we shall examine hereafter ; for our present 
purposes it suffices to point out the fact that here we have a discharge from one pole 
which is unable to approach near enough to the other pole to get relief there, and 
