muth plate, the recession from the pole and the taking up of the equatorial position is not due to repulsion, but to the endeavour the bismuth makes to establish the parallelism beforementioned. Leaving attraction and repulsion out of the question, we find it extremely difficult to affix a definite meaning to the words 'tends' and ' endeavour.' "The force is due,' says Mr. Faraday, to that power of the particles which makes them cohere in regular order, and gives the mass its crystalline aggregation, which we call at times the attraction of aggregation, and so often speak of as acting at insensible distances.' We are not sure that we fully grasp the meaning of the philosopher in the present instance; for the difficulty of supposing that what is here called the attraction of aggregation, considered apart from magnetic attraction or repulsion, can possibly cause the rotation of the entire mass round an axis, and the taking up of a fixed position by the mass, with regard to surrounding objects, appears to us insurmountable. We have endeavoured to illustrate the matter, to our own minds, by the action of a piece of leather brought near a red-hot coal. The leather will curl and motion will be caused, without the intervention of either attraction or repulsion, in the present sense of these terms; but this motion exhibits itself in an alteration of shape, which is not at all the case with the crystal.* Even if the direct attraction or repulsion of the poles be rejected, we do not see how the expressed relation between the magne-crystallic axis and the direction of the magnetic resultant is possible, without including the idea of lateral attraction between these lines, and consequently of the mass associated with the former. In the case of flat poles, the magnetic resultant lies in a straight line from pole to pole across the magnetic field. Let us suppose, at any given moment, this line and the magne-crystallic axis of a properly suspended crystal to cross each other at an oblique angle; let the crystal be forgotten for a moment, and the attention fixed on those two lines. Let us suppose the former line fixed, and the latter free to rotate, the point of intersection being regarded as a kind of pivot round which it can turn. On the evolution of the magnetic force, the magne-crystallic axis will turn and set itself alongside the magnetic resultant. The matter may be rendered very clear by taking a pair of scissors,

The subsequent reasoning of this paragraph might be omitted.-J. T. 1870.

partly open, in the hand, holding one side fast, and then closing them. The two lines close in a manner exactly similar; and all that is required to make the illustration perfect, is to suppose this power of closing suddenly developed in the scissors themselves. How should we name a power resident in the scissors and capable of thus drawing the blades together? It may be called a tendency,' or an endeavour,' but the word attraction seems to be as suitable as either.

The symmetry of crystalline arrrngement is annihilated by reducing the mass to powder. That force among the particles which makes them cohere in regular order' is here ineffective. The magne-crystallic force, in short, is reduced to nothing, but we have the same results. If, then, the principle of elective polarity, the mere modification of magnetism or diamagnetism by mechanical arrangement, be sufficient to explain the entire series of crystalline phenomena in the magnetic field, why assume the existence of this new force, the very conception of which is attended with so many difficulties ?*

APPLICATION OF THE PRINCIPLE OF ELECTIVE POLARITY TO

CRYSTALS.

We shall now endeavour to apply the general principle of elective polarity to the case of crystals. This principle may be briefly enunciated as follows:—

If the arrangement of the component particles of any body be such as to present different degrees of proximity in different directions, then the line of closest proximity, other circumstances being equal, will be that chosen by the respective forces for the exhibition of their greatest energy. If the mass be magnetic, this line will stand axial; if diamagnetic, equatorial.

From this point of view, the deportment of the two classes of crystals, represented by Iceland spar and carbonate of iron, presents no difficulty. This crystalline form is the same; and as to the arrangement of the particles, what is true of one will be true of the other. Supposing, then, the line of closest proximity to coincide with the optic axis; this line, according to the principle expressed, will stand axial or equatorial, according as

* 'Perhaps,' says Mr. Faraday, in a short note referring to 'the strange and striking character' of these force, 'these points may find their explication hereafter in the action of contiguous particles.'

the mass is magnetic or diamagnetic, which is precisely what the experiments with these crystals exhibit.

Analogy, as we have seen, justifies the assumption here made. It will, however, be of interest to inquire, whether any discoverable circumstance connected with crystalline structure exists, upon which the difference of proximity depends; and, knowing which, we can pronounce with tolerable certainty, as to the position which the crystal will take up in the magnetic field. The following experiments will perhaps suggest a reply.

If a prism of sulphate of magnesia be suspended between the poles with its axis horizontal, on exciting the magnet the axis will take up the equatorial position. This is not entirely due to the form of the crystal; for even when its axial dimension is shortest, the axis will assert the equatorial position; thus behaving like a magnetic body, setting its longest dimension from pole to pole.

Suspended from its end with its axis vertical, the prism will take up a determinate oblique position. When the crystal has come to rest, let that line through the mass which stands exactly equatorial be carefully marked. Lay a knife-edge along this line, and press it in the direction of the axis. The crystal will split before the pressure, disclosing shining surfaces of cleavage. This is the only cleavage the crystal possesses, and it stands equatorial.

Sulphate of zinc is of the same form as sulphate of magnesia, and its cleavage is discoverable by a process exactly similar to that just described. Both crystals set their planes of cleavage equatorial. Both are diamagnetic.

Let us now examine a magnetic crystal of similar form. Sulphate of nickel is, perhaps, as good an example as we can choose. Suspended in the magnetic field with its axis horizontal, on exciting the magnet the axis will set itself from pole to pole; and this position will be persisted in, even when the axial dimension is shortest. Suspended from its end, the crystalline prism will take up an oblique position with considerable energy. When the crystal thus suspended has come to rest, mark the line along its end which stands axial. Let a knifeedge be laid on this line, and pressed in a direction parallel to the axis of the prism. The crystal will yield before the edge, and discover a perfectly clean plane of cleavage.

These facts are suggestive. The crystals here experimented with are of the same outward form; each has but one cleavage; and the position of this cleavage, with regard to the form of the crystal, is the same in all. The magnetic force, however, at once discovers a difference of action. The cleavages of the diamagnetic specimens stand equatorial; of the magnetic, axial.

A cube cut from a prism of scapolite, the axis of the prism being perpendicular to two of the parallel faces of the cube, suspended in the magnetic field, sets itself with the axis of the prism from pole to pole.

A cube of beryl, of the same dimensions, with the axis of the prism from which it was taken also perpendicular to two of the faces, suspended as in the former case, sets itself with the axis equatorial. Both these crystals are magnetic.

The former experiments showed a dissimilarity of action between magnetic and diamagnetic crystals. In the present instances both are magnetic, but still there is a difference; the axis of the one prism stands axial, the axis of the other equatorial. With regard to the explanation of this, the following fact is significant. Scapolite cleaves parallel to its axis, while beryl cleaves perpendicular to its axis; the cleavages in both cases, therefore, stand axial, thus agreeing with sulphate of nickel. The cleavages hence appear to take up a determinate position, regardless of outward form, and they seem to exercise a ruling power over the deportment of the crystal.

A cube of saltpetre, suspended with the crystallographic axis horizontal, sets itself between the poles with this axis equatorial.

A cube of topaz, suspended with the crystallographic axis horizontal, sets itself with this axis from pole to pole.

We have here a kind of complementary case to the former. Both these crystals are diamagnetic. Saltpetre cleaves parallel to its axis; topaz perpendicular to its axis. The planes of cleavage, therefore, stand in both cases equatorial, thus agreeing with sulphate of zinc and sulphate of magnesia.*

Where do these facts point? A moment's speculation will perhaps be allowed us here. May we not suppose these crystals

* Topaz possesses other cleavages, but for the sake of simplicity we have not introduced them; more especially as they do not appear to vitiate the action of the one introduced, which is by far the most complete.

to be composed of layers indefinitely thin, laid side by side, within the range of cohesion, which holds them together, but yet not in absolute contact? This seems to be no strained idea; for expansion and contraction by heat and cold compel us to assume that the particles of matter in general do not touch each other; that there are unfilled spaces between them. In such crystals as we have described, these spaces may be considered as alternating with the plates which compose the crystal. From this point of view it seems very natural that the magnetic laminae should set themselves axial, and the diamagnetic equatorial.*

We have a very fine description of sand-paper here. The sand or emery on the surface is magnetic, while the paper itself is comparatively indifferent. By cutting a number of strips of this paper, an inch long and a quarter of an inch wide, and gumming them together so as to form a parallelopiped, we obtain a model of magnetic crystals which cleave parallel to their axis; the layers of sand representing the magnetic crystalline plates, and the paper the intermediate space between two plates. For such a model one position only is possible between the poles, the axial. If, however, the parallelopiped be built up of squares, equal in area to the cross section of the model just described, by laying square upon square until the pile reaches the height of an inch, we obtain a model of those magnetic crystals which cleave perpendicular to their axes. Such a model, although its length is four times its thickness, and the whole strongly magnetic, will, on closing the circuit, recede from the poles as if repelled, and take up the equatorial position with great energy. The deportment of the first model is that of scapolite; of the second, that of beryl. By using a thin layer of bismuth paste instead of the magnetic sand, the deportment of saltpetre and topaz will be accurately imitated.

Our fundamental idea is, that crystals of one cleavage are

* In these speculations we have made use of the commonly received notion of matter. Mr. Faraday, for reasons derived from electric conductibility, and from certain anomalies with regard to the combinations of potassium and other bodies, considers this notion erroneous. Nothing, however, could be easier than to translate the above into a language agreeing with the views of Mr. Faraday. The interval of space between the lamina would then become intervals of weaker force, and the result of our reasoning would be the same as before.