A], that the spark from the end _n_, which contained both portions of
electricity, was a double spark. It could not have consisted of two sparks
taking separate courses, for such an effect would have been visible to the
eye; but it is just possible, that the spark of the first end _n_ and its
jar, passing at the smallest interval of time before that of the other _o_
had heated and expanded the air in its course, and made it so much more
favourable to discharge, that the electricity of the end _o_ preferred
leaping across to it and taking a very circuitous route, rather than the
more direct one to the ball. It must, however, be remarked, in answer to
this supposition, that the one spark between _d_ and _e_ would, by its
influence, tend to produce simultaneous discharges at _n_ and _o_, and
certainly did so, when no preponderance was given to one wire over the
other, as to the previous inductive effect (1414.).

[A] Philosophical Transactions, 1834, pp. 584, 585.

1418. The fact, however, is, that disruptive discharge is favourable to
itself. It is at the outset a case of tottering equilibrium: and if _time_
be an element in discharge, in however minute a proportion (1436.), then
the commencement of the act at any point favours its continuance and
increase there, and portions of power will be discharged by a course which
they would not otherwise have taken.

1419. The mere heating and expansion of the air itself by the first portion
of electricity which passes, must have a great influence in producing this
result.

1420. As to the result itself, we see its effect in every electric spark;
for it is not the whole quantity which passes that determines the
discharge, but merely that small portion of force which brings the deciding
molecule (1370.) up to its maximum tension; then, when its forces are
subverted and discharge begins, all the rest passes by the same course,
from the influence of the favouring circumstances just referred to; and
whether it be the electricity on a square inch, or a thousand square inches
of charged glass, the discharge is complete. Hereafter we shall find the
influence of this effect in the formation of brushes (1435.); and it is not
impossible that we may trace it producing the jagged spark and the forked
lightning.

* * * * *

1421. The characters of the electric spark in _different gases_ vary, and
the variation _may_ be due simply to the effect of the heat