[A] Quarterly Journal of Science, vol. xii. pp. 74. 186. 416. 283.

122. If the rotation of the disc be reversed, the electric currents are
reversed (91.), and the pole therefore moves to the left hand. If the
contrary pole be employed, the effects are the same, i.e. in the same
direction, because currents of electricity, the reverse of those described,
are produced, and by reversing both poles and currents, the visible effects
remain unchanged. In whatever position the axis of the magnet be placed,
provided the same pole be applied to the same side of the plate, the
electric current produced is in the same direction, in consistency with the
law already stated (114, &c.); and thus every circumstance regarding the
direction of the motion may be explained.

123. These currents are _discharged or return_ in the parts of the plate on
each side of and more distant from the place of the pole, where, of course,
the magnetic induction is weaker; and when the collectors are applied, and
a current of electricity is carried away to the galvanometer (88.), the
deflection there is merely a repetition, by the same current or part of it,
of the effect of rotation in the magnet over the plate itself.

124. It is under the point of view just put forth that I have ventured to
say it is not necessary that the plate should acquire and lose its state in
a finite time (120.); for if it were possible for the current to be fully
developed the instant _before_ it arrived at its state of nearest
approximation to the vertical pole of the magnet, instead of opposite to or
a little beyond it, still the relative motion of the pole and plate would
be the same, the resulting force being in fact tangential instead of
direct.

125. But it is possible (though not necessary for the rotation) that _time_