furnished with three levelling screws. Through a sufficiently wide glass
tube, _a a_, runs a copper rod covered with insulating mastic, having a
little plate or cylindrical cavity of gilded brass at the top (Figs. 2
and 3), with two arms _d d_, _d' d_. In the plate a disc of aluminium,
_m_, is suspended by means of two silk fibres, and to the disc a very
fine aluminium wire is attached, _f f'_, bent a little at the ends, as
are the arms, _d d_, _d' d_. The disc has about three millimetres less
diameter than the plate. The diameter of the plate may vary within
certain limits, but I have found it convenient to make it eighteen
millimetres. The glass tube, _a a_ (Fig. 1), should descend below the
base as much as it rises above it, that is three to four centimetres.
The length of the index is about one decimetre.

The upper ends of the two silk fibres, by which the disc and index are
suspended, are attached to the top of the glass tube, _C_, by a
contrivance which permits a change in the distance between the two
points of suspension, and a screw, _p_, is provided to raise and lower
the disc with the index. At _n_, at the lower part of the tube, _C_,
there is a kind of torsion micrometer, arranged so as to bring the index
to the zero of the scale engraved on the graduated ring, _B_, which is
formed of a strip of good paper pasted on the rim of a glass disc. The
index must be placed at the zero of the scale, and must be some distance
from the ends of the arms of the plate with which it is parallel. The
plate is about three millimetres deep.

Having levelled the instrument, so as to render the disc concentric with
the plate, and placed the index at zero, it is obvious that if an
electric charge through the wire, _h_, reach the plate with the arms, it
will electrify the disc and index: the disc will have the opposite
electricity, and the extremities of the index will take the same
electricity as the arms, and consequently the index will describe an arc
more or less great. The motion of the index is sufficiently slow to
allow the eye conveniently to follow it. Having traversed the first arc,
which I call the _impulsive_ one, the index returns, and, after only two
oscillations, comes to rest at what I shall call the _definite_ arc.

When the electric charges are of very brief duration, the impulsive arcs
are within certain limits proportional to the tensions, and the ratio
between the impulsive and definite arcs is expressed by