the magnetic moment of the core. To imitate as nearly as possible
permanent steel magnets, it is therefore necessary to use electro
magnets, the cores of which are easily saturated. The core should be
thin and long and of the horseshoe type; the amount of wire wound round
it should be large in comparison with the size of the core.

[Illustration]

Here is a magnet partly wound which was used in one of our earliest
experiments, but which was a failure on account of having far too much
mass in the core in comparison with the amount of copper wire wound
round it. Since then we have greatly diminished the iron and increased
the copper. The cores of the instruments on the table are composed of
two or three No. 18 b.w.g. charcoal iron wires, and are wound with one
layer of 0'120 inch wire in the case of the current indicators, and
eighteen layers of 0.0139 inch wire in the case of the potential
indicator. If from the diagram, Fig. 1, we plot a curve the abscissae of
which represent exciting current, and the ordinates magnetic moment of
the soft iron core, we find that a considerable portion of the curve
is almost a straight and only slightly inclined line. If it, were a
horizontal straight line the core would be absolutely saturated, but
such as it is, it answers the purpose sufficiently well, for with a
variation of exciting current from 10 to 100 amperes the magnetic moment
varies but slightly. If a small soft iron or magnetic steel needle, _n
s_, be suspended between the poles, S N, of an electro magnet of such
proportions as described above, and the current, after exciting the
electro magnet, _e e_, be lead round the coils, DD, it will be found
that for all currents between 10 and 100 amperes the needle, _n
s_, shows a definite deflection for each current. Here we have a
galvanometer with permanent calibration. In this case the deflection of