that an instrument to measure one of the conditions of terrestrial magnetic force might be constructed on such a principle. 2450. If one might speculate upon the effect of the whole system of curves upon very large masses, and these masses were in plates or rings, then they would, according to analogy with the magnetic field, place themselves equatorially. If Saturn were a magnet as the earth is, and his ring composed of diamagnetic substances, the tendency of the magnetic forces would be to place it in the position which it actually has. 2451. It is a curious sight to see a piece of wood, or of beef, or an apple, or a bottle of water repelled by a magnet, or taking the leaf of a tree and hanging it up between the poles, to observe it take an equatorial position. Whether any similar effects occur in nature among the myriads of forms which, upon all parts of its surface, are surrounded by air, and are subjected to the action of lines of magnetic force, is a question which can only be answered by future observation. 2452. Of the interior of the earth we know nothing, but there are many reasons for believing that it is of a high temperature. On this supposition I have recently remarked, that at a certain distance from the surface downwards, magnetic substances must be entirely destitute, either of the power of retaining magnetism, or becoming magnetic by induction from currents in the crust or otherwise1. This is evidently an error; that the iron, &c. can retain no magnetic condition of itself, is very probably true, but that the magnetic metals and all their compounds retain a certain power of becoming magnetic by induction, whatever their temperature, has now been proved (2344. &c.). The deep magnetic contents of the earth, therefore, though they probably do not constitute of themselves a central magnet, are just in the condition to act as a very weak iron core to the currents around them, or other inducing actions, and very likely are highly important in this respect. What the effect of the diamagnetic part may be under the influence of such inductive forces, we are not prepared to state; but as far as I have been able to observe, such bodies have not their power diminished by heat (2397.). 2453. If the sun have anything to be with the magnetism of the globe, then it is probable that part of its effect is due to the 'Philosophical Magazine, 1845, vol. xxvii. p. 3. VOL. III. G action of the light that comes to us from it; and in that expectation the air seems most strikingly placed round our sphere, investing it with a transparent diamagnetic, which therefore, is permeable to his rays, and at the same time moving with great velocity across them. Such conditions seem to suggest the possibility of magnetism being there generated; but I shall do better to refrain from giving expression to these vague thoughts (though they will press in upon the mind), and first submitting them to rigid investigation by experiment, if they prove worthy, then present them hereafter to the Royal Society. Royal Institution, Feb. 2, 1846.—I add the following notes and references to these Researches. Brugmans first observed the repulsion of bismuth by a magnet in 1778. Antoni Brugmans Magnetismus seu de affinitatibus magneticis observationes magneticæ. Lugd. Batav. 1778, § 41. M. le Baillif on the Repulsion of a Magnet by Bismuth and Antimony, Bulletin Universel, 1827, vol. vii. p. 371; vol. viii. pp. 87, 91, 94. Saigey on the Magnetism of certain natural combinations of Iron, and on the mutual repulsions of Bodies in general. Ibid. 1828, vol. ix. pp. 89, 167, 239. Seebeck on the Magnetic Polarity of different Metals, Alloys and Oxides. Ibid. 1828, vol. ix. p. 175. 83 TWENTY-SECOND SERIES1. § 28. On the Crystalline polarity of bismuth (and other bodies), and on its relation to the magnetic form of force. i. Crystalline polarity of bismuth. ¶ii. Crystalline polarity of antimony. iii. Crystalline polarity of arsenic. Received October 4,—Read December 7, 1848. 2454. Many results obtained by subjecting bismuth to the action of the magnet have at various times embarrassed me, and I have either been contented with an imperfect explanation, or have left them for a future examination: that examination I have now taken up, and it has led to the discovery of the following results. I cannot, however, better enter upon the subject than by a brief description of the anomalies which occurred, and which may be obtained at pleasure. 2455. If a small open glass tube have a bulb formed in its middle part and some clean good bismuth be placed in the bulb and melted by a spirit-lamp, it is easy afterward, by turning the metal into the tubular part of the arrangement, to cast it into long cylinders: these are very clean, and when broken are seen to be crystallized, usually giving cleavage planes, which run across the metal. I prepare them from 0.05 to 0.1 of an inch in diameter, and, if the glass be thin, usually break both it and the bismuth together, and then keep the little cylinders in their vitreous cases. 2456. Taking some of these cylinders at random and suspending them horizontally between the poles of the electro-magnet (2247.), they presented the following phænomena. The first pointed axially; the second, equatorially; the third, equatorial in one position, and obliquely equatorial if turned round on its axis 50° or 60°; the fourth, equatorially and axially under the same treatment; and all of them, if suspended perpendicularly, pointed well, vibrating about a final fixed position which 'Philosophical Transactions, 1849, p. 1. The Bakerian Lecture. seemed to have no reference to the form of the cylinders. In all these cases the bismuth was strongly diamagnetic (2295. &c.), being repelled by a single magnetic pole, or passing off on either side from the axial line between two poles. A similar piece of finely grained or granular bismuth was, under the same circumstances and at the same time, affected in a perfectly regular manner, taking up the equatorial position (2253.), as a body simply diamagnetic ought to do. The cause of these variations was finally traced to the regularly crystalline condition of the metallic cylinders. Ti. Crystalline polarity of bismuth. 2457. Some bismuth was crystallized in the usual manner by melting it in a clean iron ladle, allowing it partly to congeal, and then pouring away the internal fluid portion. Pieces so obtained were then broken up by copper hammers and tools, and groups of the crystals separated, each group or piece consisting only of those crystals which were symmetrically arranged, and therefore likely to act in one direction. If any part of the fragments had been in contact with the iron ladle, it was cleared away by rubbing on sandstone and sand-paper. Pieces weighing from 18 grains to 100 grains were thus easily obtained. 2458. The electro-magnet employed in the first instance was that already described (2247.), having moveable terminations which supplied either conical, round, or flatfaced poles. That the suspension of the bismuth might be readily effected and unobjectionable as to magnetic influence, the following arrangement was generally adopted. A single fibre of cocoon silk, from 12 to 24 inches in length, was attached to a fit support above, and made fast below to the end of a piece of fine, straight, well-cleaned copper wire, about 2 inches in length; the lower end of this wire was twisted up into a little head, and then furnished with a pellet of cement, made by melting together a portion of pure white wax, with about one-fourth its weight of Canada balsam. The cement was soft enough to adhere by pressure to any dry substance, and sufficiently hard to sustain weights up to 300 grains, or even more. When prepared, the suspender was subjected by itself to the action of the magnet, to ascertain that it was free from any tendency to point, or be affected; without which pre caution no confidence could be reposed in the results of the experiments. 2459. A piece of selected bismuth (2457.), weighing 25 grains, was hung up between the poles of the magnet, and moved with great freedom. The constituent cubes were associated in the usual manner, being attached to each other chiefly in the line joining two opposite solid angles; and this line was in the greatest length of the piece. The instant that the magnetic force was on, the bismuth vibrated strongly about a given line, in which, at last, it settled; and if moved out of that position, it returned, when at liberty, into it; pointing with considerable force, and having its greatest length axial. 2460. Another piece was then selected, having a flatter form, which when subjected to the magnetic power, pointed with the same facility and force, but its greatest length was equatorial : still the line according to which the cubes tended to associate diametrically, was, as before, in the axial direction. Other pieces were then taken of different forms, or shaped into various forms by rubbing them down on stone, but they all pointed well; and took up a final position, which had no reference to the shape, but was manifestly dependent on the crystalline condition of the substance. 2461. In all these cases the bismuth was diamagnetic, and strongly repelled by either magnetic pole, or from the axial line. It was affected only whilst the magnetic force was present. It set in a given constant position perfectly determinate; and, if moved, always returned to it, unless the extent of motion was above 90°, and then the piece moved further round and took up a new position diametrically opposed to the former, which it then retained with equal force, and in the same manner. This phenomenon is general in all the results I have to refer to, and I will express it by the word diametral :—diametral set or position. 2462. The effect occurs with a single magnetic pole, and it is then striking to observe a long piece of a substance, so diamagnetic as bismuth, repelled, and yet at the same moment set round with force, axially or end on, as a piece of magnetic substance would do. 2463. Whether the magnetic poles employed (2458.) are pointed, round, or flatfaced, still the effect on the bismuth is the same nevertheless, the form of the poles has an important in |