e ends
of the primary pp of a second induction coil. This primary pp has a
small air gap ab.

The secondary s of this coil is provided with knobs or spheres KK of
the proper size and set at a distance suitable for the experiment.

A long arc is established between the terminals AB of the first
induction coil. MM are the mica plates.

Each time the arc is broken between A and B the jars are quickly
charged and discharged through the primary pp, producing a snapping
spark between the knobs KK. Upon the arc forming between A and B the
potential falls, and the jars cannot be charged to such high potential
as to break through the air gap ab until the arc is again broken by
the draught.

In this manner sudden impulses, at long intervals, are produced in the
primary pp, which in the secondary s give a corresponding number of
impulses of great intensity. If the secondary knobs or spheres, KK,
are of the proper size, the sparks show much resemblance to those of a
Holtz machine.

But these two effects, which to the eye appear so very different, are
only two of the many discharge phenomena. We only need to change the
conditions of the test, and again we make other observations of
interest.

When, instead of operating the induction coil as in the last two
experiments, we operate it from a high frequency alternator, as in the
next experiment, a systematic study of the phenomena is rendered much
more easy. In such case, in varying the strength and frequency of the
currents through the primary, we may observe five distinct forms of
discharge, which I have described in my former paper on the subject[A]
before the American Institute of Electrical Engineers, May 20, 1891.

[Footnote A: See THE ELECTRICAL WORLD, July 11, 1891.]

It would take too much time, and it would lead us too far from the
subject presented this evening, to reproduce all these forms, but it
seems to me desirable to show you one of them. It is a brush
discharge, which is interesting in more than one respect. Viewed from
a near position it resembles much a jet of gas escaping under great
pressure. We know that the phenomenon is due to the agitation of the
molecules near the terminal, and we anticipate that some heat must be
developed by the impact of the molecules against the terminal or
against each other. Indeed, we find that the brush is hot, and only a
little thought leads us to the conclusion that, could we but reach
sufficiently high frequencies, we co