magnet, which returns to its original position when the current ceases. When, on the contrary, the magnetism is permanent, the suspended magnet does not return to its original position when the current ceases. In Professor Tyndall's experiment the deviation was permanent, and it was equally so when a bismuth bar was freely suspended and the cores within the spirals were steel magnets. Had the effect been from currents induced in the mass of the bar of bismuth, division of the bar would have stopped them, but the result was the same with powdered bismuth as with the solid mass. Moreover, since the strength of induced currents depends upon the conducting power of the substance, and as the conducting power of copper is forty times as great as that of bismuth, had the polarity been induced and not real, the effect ought to have been forty times greater when copper instead of bismuth cores were put in the spirals, whereas it was scarcely sensible. Besides these proofs, Dr. Tyndall made experiments with eleven different diamagnetic substances, of which water was one, with similar results. He then determined the polarity of twelve paramagnetic bodies by the same method, whence it appeared that the same action which produced a north pole in the paramagnetic bodies produced a south pole in those that were diamagnetic, and vice versâ, whence he concludes that diamagnetic polarity is one of the most firmly established truths of science. It follows from this that, when a man is standing, his head is a north pole and his feet a south, and the top of an iron railing on which he may be leaning is a south pole and the lower end a north. Diamagnetic bodies thus possess a polarity, the same in kind but opposite in direction to that possessed by paramagnetic ones.' They are both dual powers, and the two diamagnetic forces like the two paramagnetic being coexistent, simultaneous, and mutually dependent, there can be no doubt that the diamagnetic forces also are represented, or rather consist of curved and closed lines of force passing through the interior of the substance. Dr. Tyndall has proved that the attraction of iron, and the repulsion of bismuth, are as the square of the electro-magnetic current producing them, and that diamagnetic substances are capable of induction.

*

* Professor Matteucci still expresses doubts on this subject, but has not yet finished his experiments.

The molecular structure of substances freely suspended between the poles of a magnet has a decided effect upon the position they

assume.

It has already been mentioned that the optic axis is a symmetrical line in a doubly refracting crystal in which there is no double refraction, and that in some crystals there are two such symmetrical lines. Now, Professor Plücker of Bonn discovered, when such crystals are submitted to powerful magnetic influence, that the single optic axis in the one, and the resultant or mean line between the double optic axes in the other, set diametrically or at right angles to the line of magnetic force; and so powerful did the Professor find the action of magnetism on crystalline form, that the mineral cyanite, when suspended, arranges itself so definitely with regard to terrestrial magnetism, that it might be used as a compass needle.

Dr. Faraday afterwards observed that amorphous substances, cut in the form of a sphere, have no tendency to set or be attracted or repelled in one direction in preference to any other; but if the sphere be formed of a crystallized substance, it is a general fact that, whether it be paramagnetic or diamagnetic, it is more powerfully attracted or repelled in one direction than in any other a property named by Dr. Faraday magnecrystallic action. For example, a sphere of calcareous spar, which is a diamagnetic crystal, is most strongly repelled in the direction of its principal optic axis, and least strongly in the direction of its least axis. In a sphere of carbonate of iron, which has exactly the same crystalline form and is highly paramagnetic, the line which in carbonate of lime sets equatorially, in this case sets axially, and more strongly in that direction than in any other. The law according to which the attraction of the carbonate of iron increases from the least to its greatest or principal optic axis, is precisely the same as that according to which the repulsion of the calcareous spar increases from the least to the principal optic axis. These relations are not altered by the immersion of the spheres in liquids of either magnetism. Dr. Faraday observed that a line at right angles to the planes of principal cleavage in crystals takes the axial position, and on that account he called it the magnecrystallic axis. Its position was proved by MM. Tyndall and Knoblauch to depend upon the general fact, that the mass is most strongly repelled in the direction of the planes of principal

cleavage, and that the elective position of crystals depends more upon the direction of these planes with respect to the electric force, than upon the optic axis. The planes of principal cleavage set themselves equatorially in diamagnetic, and axially in paramagnetic substances: it was thence inferred that the phenomena offered by crystals in the magnetic field is a particular case of the general law, that the superior action of magnets upon matter in a particular direction is due to the particles of the body being closer together in that direction than in any other: in short, the line of maximum density; the force exerted being attractive or repulsive according as the particles are paramagnetic or diamagnetic.

It appears, however, that the set of crystals with regard to the line of magnetic force does not depend solely upon their density in particular directions. Professor Matteucci, of Pisa, has proved that the diamagnetic force is inversely as the conducting power of substances for electricity, that the conducting power is a maximum in the planes of principal cleavage, and that a needle of crystallized bismuth, in which the planes of cleavage are parallel to its length, places itself equatorially with more force when these planes are vertical, or at right angles to the force, than when they are horizontal or parallel to it. Experiments had hitherto been made only with diamagnetic or slightly paramagnetic bodies, which induced M. de Roux to try the effect of magnetism on pulverized iron compressed by the hydraulic press, which reduced the grains of iron to lamellæ equivalent to planes of cleavage. Cubes of this substance, suspended by a thread over a horseshoe magnet, oscillated for a longer time when the lamellæ were perpendicular than when they were horizontal; that is, the force was stronger when the lamella were equatorial than when they were axial, exactly the same result as in Professor Matteucci's experiment with the needle of bismuth. Thus the vertical position of the cleavages, which increases the diamagnetism of the bismuth, increases also the paramagnetism of the iron. M. le Roux observes that these results are independent of the influence of the currents of electricity induced in the oscillating body, for the fundamental character of the phenomena of Arago's discovery of rotation by induction is, that the oscillations diminish rapidly in extent without any sensible diminution in their duration, while in his experiments the time of the oscillations varied.

He concludes that the arrangement of the molecules must be intimately connected with paramagnetism or diamagnetism itself, since the effect of that arrangement is equally sensible in bismuth and iron, although the diamagnetism of the former is 25,000 times weaker than the paramagnetism of the latter.

The diamagnetism of conducting substances and metals, such as gold, silver, and copper, is augmented by division. Compression has also a great effect on magnetic action. For example, a bar of soft iron sets with its longest dimensions from pole to pole of a magnet, but a bar of compressed carbonate of iron-dust, whose shortest dimensions coincide with the line of pressure, sets equatorially. A bar of bismuth whose plane of principal cleavage is parallel to its length sets equatorially, but a bar of compressed bismuth dust, whose shortest dimensions coincide with the line of pressure, or a bar of bismuth whose principal planes of cleavage are transverse to its length, sets with its length axially. The antithesis is perfect whether the bars are under the influence of a magnet or electro-magnet. For since the diamagnetic force is inversely as the conducting power of a body for electricity, and that the latter is a maximum in the direction of the planes of principal cleavage, therefore when these planes are parallel to the axis of the bismuth bar it sets equatorially; but as the conducting power is augmented when the bismuth dust is compressed in the direction of the force, the diamagnetic power is diminished, and the bar sets axially. Again, since the paramagnetic force augments with the conducting power, the action of the magnet on the iron is antithetic to that on the bismuth.

The action of an electro-magnet on copper is strongly contrasted with that which it exerts on iron or bismuth. For when a copper bar suspended by a thread revolves before its pole, it is brought to a dead halt as soon as the electric current acts upon it, and maintains its position with considerable tenacity, for it does not return when pushed out of it, but keeps its new place with stiffness; however, as soon as the electric current ceases, there is a strong revulsion, the bar revolving the contrary way. Even when swinging with considerable force it may be caught and retained in any position at pleasure, but there is no revulsion when it is arrested either in the axial or equatorial position; at any angle between these two, but especially midway, the electricity will make it move towards the axis, but it is arrested

before it comes to it. The action depends much on the form and dimensions of the bar and the magnetic pole, which ought to be flat. The phenomena are due to the high electro-conducting power of the copper, and are met with in some of the other pure metals, though in a far inferior degree.

Great magnetic power is requisite for all these experiments. Dr. Faraday employed a magnet that could sustain a weight of 450 lbs. at each pole, and the poles were either pointed or flat surfaces at pleasure, as the kind of experiment required.

Heat strongly affects the magnetic properties of bodies. Dr. Faraday found that, when the temperature of nickel is increased, its magnetic force diminishes; when that of iron is increased its magnetic force remains the same, while that of cobalt increases; which seems to indicate that there is a temperature at which the magnetic force is a maximum, above and below which it diminishes. Nickel loses its magnetism at the temperature of boiling oil, iron at a red heat, and cobalt near the temperature at which copper melts. Calcareous spar retains its magnetic character at a very high temperature; but the same substance when it contains iron, and also oxide of iron, loses it entirely at a dull red heat. A crystal of the ferrocarbonate of lime was absolutely reversed by change of temperature, for at a low heat the optic axis pointed axially, and at a high temperature equatorially. With the exception of these substances, magnecrystals, whether paramagnetic or diamagnetic, are generally all affected alike by heat. The difference between the forces in any two different directions, as for instance the greatest and least principal axes, diminishes as the temperature is raised, increases as the temperature is lowered, and is constant for a given temperature. No unmixed or pure substance has as yet passed by heat from the paramagnetic to the diamagnetic state. No simple magnecrystal has shown any inversion of this kind, nor have any of the chief axes of power changed their characters or relations to one another.

It appears that, as the molecules of crystals and compressed bodies affect magnetism, so magnetism acts upon the molecules of matter, for torsion diminishes the magnetic force, and the elasticity of iron and steel is altered by magnetism. M. Matteucci has found that the mechanical compression of glass alters the rotatory power of a polarized ray of light transmitted through it, and that