merely necessary to make a convenient bar of steel and to stroke it repeatedly from end to end with the lodestone, taking care that the strokes are alike in all respects. We shall henceforth employ such steel magnets instead of lodestone.

§ 3. Polarity. If we examine our magnetised bars of steel we shall find that, though the lodestone passed along every portion of them equally, the ends always possess far greater magnetic powers than do the middle portions. Further than this, we find that in each magnet these two ends are different. If we magnetise a bar A B, and another A' B', in such a way that A and A', B and B' are respectively corresponding ends (and we can do this by stroking the one bar from A to B, the other from A' to B' with the same portion of one piece of lodestone), then we find that A repels A', B repels B', but that A attracts B' and A' attracts B. We find also that if such bars be suspended they will take up a definite position with respect to the earth, setting themselves in a direction called magnetic north and south, the similarly treated ends A and A', B and B', lying in similar directions.

We may say that the ends of the bar have acquired remarkable properties, these properties being, in a sense, of opposite natures at the two ends respectively. It is usual to call the ends where the magnetic properties are displayed poles, and the bar is said to have acquired magnetic polarity.

We may here point out a possible source of error. We have seen that like poles repel, and unlike poles attract, one another; and further that all needles or bars, that are magnetised and suspended, turn so as to point in a northerly and southerly direction. That end of the bar that points towards the geographical north pole of the earth is usually called a north pole, or north-seeking pole; the other end is called a south pole, or south-seeking pole. But from what we have seen it is evident that the north-seeking pole of the bar must be of opposite polarity to the north pole' of the earth towards which it turns; and so with the south-seeking pole of the bar.

It is better, therefore, to speak of the 'north pole of the earth, but the 'north-seeking' pole of a magnetised bar.

§ 4. Constitution of a long thin Magnet.-Much light is thrown upon the nature of this Polarity by the following simple experiment:

A strongly magnetised knitting-needle is taken, and the

polarity of the ends, and neutrality of the middle, tested by use of filings and compass-needle.

It is then broken in half.

We now find each half a complete magnetic needle; two opposite poles having apparently started into existence in the previously neutral middle.

Each half is then again broken, and so on.

The original needle, and the condition when it is broken, are shown in the accompanying figure.

b a

It would seem that all these intermediate poles were existing in the original needle, but that they neutralised each other as far as any external action went.

As far as experiment goes there is nothing against the supposition that if we had a magnetic bar very thin, in fact, a single row of molecules of steel, we might continue this process of breaking up until we should find each molecule a complete little magnet with poles at its ends (if we may use the term ends with respect to a molecule which may be spherical), and a neutral region at the middle.

Experiments.--(i.) Magnetising a bar of steel with lodestone, we find that iron filings will cluster chiefly at the ends.

(ii.) Take a series of unmagnetised knitting-needles; testing them by seeing that either end acts the same on the north' end of an ordinary compassneedle.

Lay them down side by side, and magnetise them successively with a piece of lodestone or with a steel magnet; marking with gummed paper the ends that must, by the process of magnetisation, be similar.

It will then be found that (a) similar poles repel, dissimilar attract, one another; (b) the suspended needles will all set themselves with the marked ends turned in the same direction.

§ 5. Molecular theory of Magnetism.-The above experiment suggested, very early, that probably magnetism is molecular. The view adopted was that when a long thin needle was magnetised, its molecular constitution was somewhat as represented in the accompanying figure. NS is the magnetic needle, supposed to

consist of a single line of molecules. Each of these molecules is a small magnet, and they are arranged so that north-seeking poles are opposed to south-seeking poles all along: there being left at the one end a north-seeking pole, and at the other end a southseeking pole, uncompensated.

N

n

FIG. i.

It is assumed, as justified by such experiments as that alluded to in § 4, that the south-seeking pole of one molecule can neutralise, as far as external manifestation of magnetic properties goes, the opposed north-seeking pole of another molecule, while the northseeking and south-seeking poles of the same molecule, though very close to one another, do not so neutralise one another.

The state of neutrality was formerly considered to be a 'higgledy-piggledy' arrangement, or rather absence of arrangement, producing on the whole external neutrality.

Thus the end of a magnetised bar would present to external bodies a whole set of molecular north-seeking or south-seeking poles, while the end of an unmagnetised bar would present a mixed surface of north-seeking and south-seeking poles whose total external action would be nil.

Professor D. E. Hughes has taken up the molecular theory, and has reduced it to order, much extending it. In fact, he has done so much that the chief questions left unsolved are the fundamental ones: 'What is the polarity of a molecule? and, Why do similar poles repel and dissimilar attract one another?' The question of diamagnetism also needs some further research.

The main drift of Professor Hughes's theory, each portion of which he has supported by experiment, is somewhat as follows:

(i.) Each molecule (possibly each atom) of every substance possesses to a fixed and unalterable degree a property called polarity. The poles of the molecule are probably fixed in that molecule, and can be made to lie in a changed direction only by rotation of the molecule.

(ii) The opposite poles of consecutive molecules do, when the molecules are so rotated that these 'poles 'lie one over against the other, neutralise each other with respect to external action.

(iii.) In neutrality there is always some symmetrical arrangement by which the molecules neutralise one another with respect to any external action; either by the molecules being paired off with opposite poles together, or by a whole line of molecules neutralising each other's poles in one long chain, or by some equivalent arrangement.

One of the great features of Hughes's theory is the symmetry of arrangement that is considered to exist under all circumstances. The accompanying figures indicate two possible cases of neutrality.

Since the molecule, whatever its real form, is practically for magnetic purposes a short line with poles at the ends, we have here represented the molecules as lines; indicating thereby their magnetic, though not their actual shape.

(iv.) When the bar is magnetised to the greatest possible degree, the arrangement will be such that there is a chain of molecules in which the north-seeking pole of the molecule at the one end of a chain, and the south-seeking pole of the molecule at the other end of the same chain, will be left unneutralised; while between these there is complete neutralisation. And, further, these chains must be arranged side by side so that the free northseeking poles are all at one end of the bar and the free southseeking poles at the other.

In the above two cases this might be effected in fig. ii. by a rotation of the molecules so as to form one line as in fig. i., while in fig. iii. it would be only necessary to break the chain at any point. Thus there is a limit to possible strength of magnetism. The utmost that can be done is to arrange the molecules in the

most advantageous position. The strength of magnetism will then depend on the ultimate properties of the molecules.

(v.) Usually, or perhaps in all cases, the setting of the molecules so as to give evident magnetism is not completed. The molecules are partly rotated away from their position of what we may call 'short-circuiting' or neutralisation, and we get all intermediate cases between perfect neutrality and perfect magnetisation. A piece of iron when magnetised to its fullest extent is said to be saturated.

(vi.) Arrangements of the molecules, in which they do not pair off, or in any other way neutralise one another, and so give zero external action, are unstable. In fact, if only there be given free movement to the molecules, so that they can act on each other without hindrance, they will always 'short-circuit' with one another and give us zero external magnetic action.

(vii.) The rearrangement of the molecules so as to give external magnetism is resisted both by their action on each other and by the mechanical difficulties in the way of rotation of the molecules. This mechanical rigidity, that opposes the 'magnetisation' of a bar and also opposes the molecular tendency to a return to the neutral, short-circuited, condition, is called by the inexact name of coercive force. Any action (as heating, hammering, and the like) that gives freer play to the molecules will diminish the coercive force of the bar. In general, therefore, it is useful to hammer a bar while it is being magnetised, but bad to do so when it is no longer under the magnetising influence.

(viii.) The degree of molecular rigidity depends upon the chemical nature of the metal and upon its temper. Thus, some specimens of 'soft iron' will hardly retain any trace of magnetism when the magnetising influence has been removed, while some specimens of 'hard steel' will remain unaltered for a long time.

(ix.) Temporary and Residual Magnetism.-In general, if we magnetise a bar, it will readily lose part of its magnetism, but will retain the remainder unaltered for a long time. It is supposed that the molecules can, without much difficulty, rotate part of the way back to their original positions of neutrality; but that at a certain point their freedom of movement is limited by the 'molecular rigidity' (or coercive force) spoken of above.