(the former conferring them by association upon the latter), then distinguishes magnetic action into four kinds or modes,namely, the ordinary, the diamagnetic, the induction of currents, and the rotation of a ray; and points out that any acceptable hypotheses ought to account for the four modes of action, and, it may be added, ought to agree with, if not account for, the phenomena of electro-chemical action also. De la Rive conceives that as regards these modes of action this hypothetical result may be obtained, and both Ampère's and Weber's views also retained, in the following manner. All the atoms of matter are supposed to be endowed with electrical currents of a like kind, which move about them for ever, without diminution of their force or velocity, being essentially a part of their nature. The direction of these currents for each atom is through one determinate diameter, which may therefore be considered as the axis. Where they emerge from the body of the atom they divide in all directions, and running over every part of the surface converge towards the opposite end of the axis diameter, and there re-enter the atom to run ever through the same course. The converging and diverging points are as it were poles of force. Where the atoms of matter are close or numerous in a given space (and chemical considerations lead to the admission of such cases), the hypothesis then admits that several atoms may conjoin into a ring, so that their central or axial currents may run one into the other, and not return as before over the surface of each atom: these form the molecules of magnetic matter, and represent Ampère's hypothesis of molecular currents. Where the atoms, being fewer in a given space, are farther apart, or where, being good conductors, the current runs as freely over the surface as through the axis, then they do not form like groups to the molecules of magnetic matter, but are still considered subject to a species of induction by the action of external magnets and currents; and so give rise to Weber's reverse currents. The induction of momentary currents and the rotation of a ray are considered by De la Rive as in conformity with such a supposition of the electric state of the atoms and particles of matter.

The Lecturer seemed to think that the great variety of these hypotheses and their rapid succession were rather a proof of weakness in this department of physical knowledge than of

strength, and that the large assumptions which were made in turn for each should ever be present to the mind. Even in the most perfect of them, i. e. De la Rive's, these assumptions are very considerable; for it is necessary to conceive of the molecules as being flat or disc-like bodies, however numerous the atoms of each may be; also that the atoms of one molecule do not interfere with or break up the disposition of those of another molecule; also that electro-chemical action may consist with such a constituted molecule; also that the motive force of each atom current is resident in the axis; and on the other hand that the passage of the current over the surface offers resistance; for unless there were a difference between the axial and the surface force in one direction or the other, the atoms would have no tendency to congregate in molecules. In making these remarks, however, the speaker had no thought of depreciating hypothesis or objecting to its right use. No discoverer could advance without it; and such exertions as those made by De la Rive, to bring into harmony thoughts which in their earlier forms were adverse to each other, were of the more value, because they were the exertions of a man who knew the value both of hypothesis and of laws, of theory and of fact, and had given proofs of the power of each by the productions of his own mind. Still the speaker advocated that mental reservation which kept hypothesis in its right place and which was ready to abandon it when it failed; and as examples referred to Newton, who (as is shown by his Letters to Bentley) had very strong convictions of the physical nature of the lines of gravitating force, yet in what he publicly advanced stopped short at the law of action of the force, and thence deduced his great results;-and also to Arago, who, discovering the phenomena of magnetic rotation, yet not perceiving their physical cause, had that philosophic power of mind which enabled him to refrain from suggesting one.

IX.-ON SOME MECHANICAL EFFECTS OF
MAGNETISATION.

WISHING in 1855 to make the comparison of magnetic and diamagnetic phenomena as thorough as possible, I sought to determine whether the act of magnetisation produces any change of dimensions in the case of bismuth, as it is known to do in the case of iron. The action, if any, was sure to be infinitesimal, and I therefore cast about for a means of magnifying it. The idea which appeared most promising was to augment in the first instance by a lever the small amount of change expected, and to employ the augmented effect to turn the axis of a rotating mirror. By making the axis small enough it was plain that an infinitesimal amount of rectilinear motion might be caused to produce a considerable amount of angular motion. This I proposed to observe by a telescope and scale after the method of Gauss. I consulted Mr. Becker, and, thanks to his great intelligence and refined mechanical skill, I became the possessor of the apparatus now to be described.

A B (fig. 3) is the upper surface of a massive block of Portland stone. It is 21 inches wide, 13 inches deep, and 29 inches high. In it are firmly fixed two cylindrical brass pillars, C C, 1 inch in diameter and 35 inches in height. Over the pillars pass the two clamps, O O', and from the one to the other passes a cylindrical cross bar, 11 inches long and of an inch wide. This cross bar is capable of two motions; the first up and down the two pillars C C, parallel to itself; the second being a motion round its own axis. To this cross piece is attached the magnifying apparatus A.

The bar to be examined is set upright between the two pillars; being fixed firmly into a leaded screw imbedded in the Portland stone. It is surrrounded by an electro-magnetic helix B. On the top of the bar I rests one end of a small cylindrical brass rod, with pointed steel ends. This rod fits

accurately into a brass collar, moving up and down in it with the least possible friction. The other point of the rod presses against a plate of agate very close to a pivot round which the plate can turn. The agate plate is attached to a brass lever 2.1 inches long, whose fulcrum is the pivot just mentioned. Any motion of the point against which the rod presses is magnified about fifty times at the end of the lever. From this end passes a piece of fine steel fibre round the axis of a rotating mirror, which turns as the end of the lever moves. The mirror rotates with its axis. For accurate experiments an illuminated vertical scale is placed at a distance of about twelve feet from the mirror, which is observed through a telescope placed beside the scale. The magnifying apparatus is shown in detail in fig. 2, where M is the mirror; S and S' two centre-screws, whose points constitute the pivot round which the lever turns; E is a small counterweight; T T is the cross-piece to which the magnifying apparatus is attached. A naked section of the magnifying apparatus is given in fig. 1. I is the bar to be magnetised, F the brass rod with the pointed steel ends, divested of its collar, one of its ends pressing against the plate of agate near the pivot x, and the other resting upon the bar of iron at y. From the end L of the lever the steel fibre passes round the axis a of the mirror M. When the bar I changes its length, the motion at L turns the mirror; and when I resumes its primitive length, the mirror is brought back to its first position by the spiral hair-spring shown in the figure.

In a lecture, of which the following is an abstract, the instrument just described was employed to show the elongation of a bar of iron by magnetism. It is the instrument referred to in Heat as a Mode of Motion,' 3rd edition, p. 85. Merely breathing against an iron bar produces a visible expansion. By squirting warm water from a syringe-bottle against the bar, and by employing ether or alcohol in the same way for cooling, the luminous beam which forms the index may, in a few seconds, be caused to pass through a distance of twenty or thirty feet.

ON A MAGNETIC EXPERIMENT.*

SOME years ago I devised an apparatus to enable me to investigate certain mechanical effects which accompany the act of

* Proceedings of Royal Institution, vol. iv. p. 317.

R