Page:Scientific Memoirs, Vol. 2 (1841).djvu/14

25.

If we first consider the experiments enumerated by M. Parot, p. 123 of the Bulletin, it appears that we may deduce from them the important result, that the solid thickness of hollow tubes has but little influence on their magnetic power. In fact, the magnetism of the tube designated by $B$, of 2½ lines solid thickness, differs merely 190 in magnetic power from the tube $D$, which is but one line in thickness. Even the energy of the solid cylinder of equal surface is far below the proportion relative to its mass. These experiments are connected with the beautiful known law established by Mr. Barlow, and which relates to the action of terrestrial magnetism on solid and hollow soft iron spheres, a law which is also confirmed by the experiments of Captain Kater on sheet-iron tubes, and which has undergone the rigorous and scientific calculations of M. Poisson. Mr. Barlow fixed the limit of the solid thickness at about 130 of an inch; such is the cylinder $F$ of M. Parrot, constructed of common sheet-iron, of 245 of an inch in thickness, and the result of which differs considerably from other experiments. Its magnetic power amounted only to 32½, while that of the tube $D$ of one line in solid thickness is expressed by 89.

26.

We ought to observe extreme caution in drawing conclusions from the experiments in question. The manner of comparing magnetic forces from the adherence of the armature is itself subject to serious inconveniences, which consist, for the most part, in the complicated accidents which accompany the contact of the armature and its sudden interruption. The errors which proceed from this cause being very considerable, it is at least worth while to endeavour to separate them from others depending on the continual variation of the voltaic action, a source not less prolific of serious errors. The somewhat difficult experiments instituted on the magnetic force of a soft iron bar, bent in a horseshoe form, and which I have related in Art. 9. of this memoir, made me sensible of these inconveniences in all their extent. But they may easily be overcome, by keeping an exact account of the force of the current, by means of a galvanometer interposed in the circuit. I had constructed, purposely for such experiments, an instrument of very great service. A needle.