had been formed by either of the two preceding processes (34. 36.). Being
left in, the needle resumed its first position, and then the magnet being
withdrawn the needle was deflected in the opposite direction. These effects
were not great; but by introducing and withdrawing the magnet, so that the
impulse each time should be added to those previously communicated to the
needle, the latter could be made to vibrate through an arc of 180° or more.

40. In this experiment the magnet must not be passed entirely through the
helix, for then a second action occurs. When the magnet is introduced, the
needle at the galvanometer is deflected in a certain direction; but being
in, whether it be pushed quite through or withdrawn, the needle is
deflected in a direction the reverse of that previously produced. When the
magnet is passed in and through at one continuous motion, the needle moves
one way, is then suddenly stopped, and finally moves the other way.

41. If such a hollow helix as that described (34.) be laid east and west
(or in any other constant position), and a magnet be retained east and
west, its marked pole always being one way; then whichever end of the helix
the magnet goes in at, and consequently whichever pole of the magnet enters
first, still the needle is deflected the same way: on the other hand,
whichever direction is followed in withdrawing the magnet, the deflection
is constant, but contrary to that due to its entrance.

42. These effects are simple consequences of the _law_ hereafter to be
described (114).

43. When the eight elementary helices were made one long helix, the effect
was not so great as in the arrangement described. When only one of the
eight helices was used, the effect was also much diminished. All care was
taken to guard against tiny direct action of the inducing magnet upon the