a kind of anti-magnetism! I have not succeeded in confirming this discovery by the following experiment. In the first place I considered it necessary to exclude the weakening effects, which constantly take place when a second uniting wire is introduced into a closed circuit; and it is easy to attain this by employing also another battery, which is placed in contact with the extremities of the interior helix. Now the hollow tube having been rendered magnetic by a surrounding helix, a considerable quantity of iron filings adhered to the poles. The contact of the interior helix, with its own battery, was then established. Not a single grain of the filings fell off, whatever was the direction of the current;—a proof that more experiments are requisite on the part of M. Parrot to confirm this discovery. This is also another experiment, which proves incontestably that the inner surface of a hollow cylinder is entirely deprived of magnetism (29). The two extremities of an inner helix were placed in contact with a very delicate galvanometer; the cylinder was then attached as armature to a horse-shoe strongly magnetized. The adherence was very powerful, and yet the needle was not in the least affected. But when one of the ends of a small bar of soft iron was introduced into the inner helix, the needle was greatly agitated as soon as the other extremity was placed in contact with one of the poles. The hollow cylinder had been rendered magnetic by a surrounding coil; there was in this case also no deviation of the galvanometer placed in contact with the interior coil.
37.
To these observations relating to soft iron magnetized by influence, it is necessary to add a few words with respect to steel, this being endowed with a coercitive force, supposed to be null in soft iron. Subjected to inductive currents, sufficiently energetic to overpower this resistance, steel retains the magnetic condition which the currents have caused it to adopt, even when these latter have disappeared. In fact, currents of short duration, which have no influence on soft iron, magnetize considerably tempered steel. If we refer to the soft iron wire subjected to the experiment, described in Art. 34, the wire does, it is true, become magnetic, but the magnetism cannot be developed in presence of the helix, which is endowed with a contrary magnetism. If these two portions could be separated, there would be an evident magnetization; but this cannot be effected, as the mag-
