252 



NATURE 



\yan. 1 6, 1890 



there is between the two classes no intermediate class. 

 The magnetic property of iron is exceedingly easily des- 

 troyed. If iron be alloyed with 12 per cent, of man- 

 ganese, the kick on the galvanometer which the material 

 will give, if made into a ring, is only about 25 per cent, 

 greater than is the case with the most completely non- 

 magnetic material, instead of being some hundreds of 

 times as great, as would be the case with iron. Further, 

 with this manganese steel, the kick on the galvanometer 

 is strictly proportional to the magnetizing current in the 

 primary, and the material shows no sign of hysteresis. 

 In short, all its properties would be fully accounted for if 

 we supposed that manganese steel consisted of a perfectly 

 non-magnetic material, with a small percentage of metallic 

 iron mechanically admixed therewith. Thus the property 

 of non-magnetizability of manganese steel is an excellent 

 proof of the fact — which is also shown by the non-mag- 

 netic properties of most compounds of iron — that the 

 property appertains to the molecule, and not to the atom ; 

 or. to put it in another way, suppose that we were 



to imagine manganese steel broken up into small par- 

 ticles, as these particles became smaller there would at 

 length arrive a point at which the iron and the manganese 

 would be entirely separated from each other : when this 

 point is reached the particles of iron are non-magnetic. 

 By the magnetic molecule of the substance we mean the 

 smallest part which has all the magnetic properties of the 

 mass. The magnetic molecule must be big enough to 

 contain its proportion of manganese. In iron, then, we 

 must have a collection of particles of such magnitude that 

 it would be possible for the manganese to enter into each 

 of them, to constitute an element of the magnet. Man- 

 ganese is, so far as I know, a non-magnetic element 

 Smaller proportions of manganese reduce the magnetic 

 property in a somewhat less degree, the reduction being 

 greater as the quantity of manganese is greater. It 

 appeared very possible that the non-magnetic property 

 of manganese steel was due to the coercive force being 

 very great — that, in fact, in all experiments we were still 

 on that part of the magnetization curve below the rapid 



rise, and that if the steel were submitted to greater forces 

 it would presently prove to be magnetic, like other kinds 

 of steel. Prof. Ewing, however, has submitted man- 

 ganese steel to very great forces indeed, and finds that 

 its magnetism is always proportional to the magnetizing 

 force. 



No single body is known having the property of 

 capacity for magnetism in a degree which is neither very 

 great nor very small, but intermediate between the two 

 extremes. We can, however, mix magnetic and non- 

 magnetic substances to form bodies apparently inter- 

 mediate. It is, therefore, interesting to consider what 

 the properties might be of such a mixture. It depends 

 quite as much on the way in which the magnetic part is 

 arranged in the mass, as on its actual quantity. Suppose, 

 for example, it is arranged as in Fig. 4^in threads or 

 plates having a very long axis in the direction of the 

 magnetizing force — we may at once determine the curve 

 of magnetization of the mixture from that of the magnetic 



substance by dividing the induction for any given force 

 in the ratio of the whole volume to the volume of magnetic 

 substance. If, on the other hand, it is as in Fig. 5 — with 

 a very short axis in the direction of the force, and a long 

 axis perpendicular thereto — we can equally construct the 

 curve of magnetization. This is done in Fig. 6, which 

 shows the curve when nine-tenths of the material is highly 

 magnetic iron, arranged as in Fig. 5, whilst the other curve 

 of the same figure is that when only one-tenth is magnetic, 

 but arranged as in Fig. 4. You observe how very different 

 is the character of the curve — a difference which is reduced 

 by the much less proportion of magnetic material in the 

 mixture in the one case than in the other. One peculiarity 

 of these arrangements of the two materials in relation to 

 each other is, that the resulting material is not isotropic ; 

 that is, its properties are not the same in all directions, but 

 depend upon the direction of the magnetizing force in the 

 material. Of course, this is not at all a probable arrange- 

 ment, but it is instructive in showing the character of the 



