Page:Scientific Memoirs, Vol. 2 (1841).djvu/28

 16 magnetic, unless the centre $$C$$ of the voltaic circuit is situated on the same side as the bar, or that this latter is placed at the interior of the circuit. The following experiment may serve to confirm this opinion.

A wooden box, $$ABCD$$ (Pl. I. fig. 3), 3 feet in length, and 1 foot broad, was surrounded with six convolutions of brass wire. The ends of this wire being placed in contact with a single voltaic pair, a small bar of soft iron, placed in $$M$$, was sufficiently magnetized to attract a quantity of iron filings. But in placing this same small bar on $$M$$, there was no trace of magnetism even on employing the powerful battery of 16 pairs. This simple experiment appears to me to be sufficiently important to render the facts intelligible. We might thus lay down the general rule; that if we considered any body magnetized by induction, there will be a development of magnetism only at the surfaces, whose rays of curvature traverse the body, or bisect it according to its thickness. If the rays are obliged to be prolonged in order to bisect the body, there will be no magnetism evident. Moreover, the currents must always be so directed, that the body shall be situated in the interior, and surrounded by the circuit. When a plate of soft iron, or of sheet-iron, is bent in any way whatsoever, and is exposed to the influence of electric currents, the curve of inflection $$A$$ and $$B$$ (Pl. I. fig. 4) are at the same time the limits of the developed magnetism. The iron filings will attach themselves only to the extremities, or convex surfaces, as indicated by the figure. I did not consider it necessary to make this experiment, for there can be no doubt of its being the case. This manner of considering these relations also explains why soft iron employed to complete a voltaic circuit, acts like all other metals, and is not magnetized transversally or according to the axis of the currents, A plate of sheet-iron $$F$$ (Pl. I. fig. 5 ), of half a line in thickness, and curved as shown in the figure, was soldered to a voltaic pair of 3 inches in length; on plunging it into acidulated water, a testing needle was slightly affected, but not more than by a current of the same intensity which had been passed through a conjunctive wire of copper.

M. Parrot seems to be of opinion, that an interior helix not only adds nothing to the development of magnetism, but that it even exercises some weakening action, whatever may be the direction of the current. In truth, this would be very remarkable!