roduced was important,
since the temporary magnetism developed in the arrangement of Sturgeon
was sufficient to support a weight of several pounds, and an instrument
was thus produced of value in future research.

[Illustration: Fig. 5]

The next improvement was made by myself. After reading an account of the
galvanometer of Schweigger, the idea occurred to me that a much nearer
approximation to the requirements of the theory of Ampere could be
attained by insulating the conducting wire itself, instead of the rod to
be magnetized, and by covering the whole surface of the iron with a
series of coils in close contact. This was effected by insulating a long
wire with silk thread, and winding this around the rod of iron in close
coils from one end to the other. The same principle was extended by
employing a still longer insulated wire, and winding several strata of
this over the first, care being taken to insure the insulation between
each stratum by a covering of silk ribbon. By this arrangement the rod
was surrounded by a compound helix formed of a long wire of many coils,
instead of a single helix of a few coils, (Fig. 5).

In the arrangement of Arago and Sturgeon the several turns of wire were
not precisely at right angles to the axis of the rod, as they should be,
to produce the effect required by the theory, but slightly oblique, and
therefore each tended to develop a separate magnetism not coincident
with the axis of the bar. But in winding the wire over itself, the
obliquity of the several turns compensated each other, and the resultant
action was at right angles to the bar. The arrangement then introduced
by myself was superior to those of Arago and Sturgeon, first in the
greater multiplicity of turns of wire, and second in the better
application of these turns to the development of magnetism. The power of
the instrument with the same amount of galvanic force, was by this
arrangement several times increased.

The maximum effect, however, with this arrangement and a single battery
was not yet obtained. After a certain length of wire had been coiled
upon the iron, the power diminished with a further increase of the
number of turns. This was due to the increased resistance which the
longer wire offered to the conduction of electricity. Two methods of
improvement therefore suggested themselves. The first consisted, not in
increasing the length of the coil, but in using a number of separate
coils on the same piece o