October 15, 1891] 



NATURE 



57 



molecules tumble through unstable states. This may 

 have something to do with the fact, well known to 

 engineers, that numerous repetitions of a straining 

 action, so slight as to be safe enough in itself, have a 

 dangerous effect on the structure of iron or steel. 



Another thing on which the theory throws light is the 

 phenomenon of time-lag in magnetization. When a piece 

 of iron is put into a steady magnetic field, it does not 

 take instantly all the magnetism that it will take if time 

 be allowed. There is a gradual creeping up of the mag- 

 netism, which is most noticeable when the field is weak 

 and when the iron is thick. If you will watch the manner 

 in which a group of little magnets breaks up when a mag- 

 netic force is applied to it, you will see that the process 

 is one that takes time. The first molecule to yield is 

 some outlying one which is comparatively unattached — 

 as we may take the surface molecules in the piece of iron 



Fig. 15, from one of Hopkinson's papers, shows what is 

 I observed as the temperature of a piece of steel is gradually 

 j raised. The sudden loss of magnetic quality occurs when 

 the metal has become red-hot ; the magnetic quality is re- 

 covered when it cools again sufficiently to cease to glow. 

 Now, as regards the first effect— the increase of suscepti- 

 bility with increase of temperature — I think that is a con- 

 sequence of two independent effects of heating. The 

 structure is expanded, so that the molecular centres lie 

 further apart. But the freedom with which the molecules 

 obey the direction of any applied magnetic force is in- 

 creased not by that only, but perhaps even more by their 

 j being thrown into vibration. When the field is weaV, 

 heating consequently assists magnetization, sometimes 

 very greatly, by hastening the passage from stage a to 

 stage b of the magnetizing process. And it is at least a 

 , conjecture worth consideration whether the sudden loss of 

 magnetic quality at a higher temperature is not due to the 

 I vibrations becoming so violent as to set the molecules 

 i spinning, when, of course, their polarity would be of no 

 j avail to produce magnetization. We know, at all events, 

 I that when the change from the magnetic to the non- 

 1 magnetic state occurs, there is a profound molecular 

 change, and heat is absorbed which is given out again 

 j when the reverse change takes place. In cooling from a 

 ! red heat, the iron actually extends at the moment when 

 this change takes place (as was shown by Gore), and so 

 ■ much heat is given out that (as Barrett observed) it re- 



FiG. 14. — Cycle of loading and unloading. 



to be. It falls over, and then its neighbours, weakened | 

 by the loss of its support, follow suit, and gradually the \ 

 disturbance propagates itself from molecule to molecule \ 

 throughout the group. In a very thin piece of iron — a 

 fine wire, for instance— there are so many surface mole- 

 cules, in comparison with the whole number, and con- ! 

 sequently so many points which may become origins of j 

 disturbance, that the breaking up of the molecular com- 

 munities is too soon over to allow much of this kind of 

 lagging to be noticed. I 



Effects of temperature, again, miy be interpre'.ed by 

 help of the nnlecular theory. When iron or nickel or j 

 cobalt is heated in a weak magnetic field, its susceptibility ' 

 to magnetic induction is observed to increase, until a stage 

 is reached, at a rather high temperature, when the magnetic ; 

 quality vanishes almost suddenly and almost completely. 



NO. I 146, VOL. 44] 



Fig. 13. — Relation of magnetic inductive capacity 

 steel (Hopkinson). 



I temperature in hard 



glows, becoming brightly red, though, just before the 

 change, it had cooled so far as to be quite dull. [Expeii- 

 ment, exhibiting retraction and re-glow in cooling, shown 

 by means of a long iron wire, heated to redness by the 

 electric current.] The changes which occur in iron and 

 steel about the temperature of redness are very complex, 

 and I refer to this as only one possible direction in which 

 a key to them may be sought. Perhaps the full explana- 

 tion belongs as much to chemistry as to physics. 



An interesting illustration of the use of these models 

 has reached me, only to-day, from New York. In a 

 paper just published in the Electrical World (reprinted 

 in the Electrician for May 29, 1891), Mr. Arthur Hoopes 

 supports the theory I have laid before you by giving 

 curves which show the connection, experimentally found 

 by him, between the resultant polarity of a group of little 

 pivoted magnets and the strength of the magnetic field, 

 when the field is applied, removed, reversed, and so on. 

 I shall draw these curves on the screen, and rough as 

 they are, in consequence of the limited number of 

 magnets, you see that they succeed remarkably well in 

 reproducing the features which we know the curves for 

 solid iron to possess. 



It may, perhaps, be fairly claimed that the models 

 whose behaviour we have been considering have a wider 

 application in physics than merely to elucidate magnetic 

 processes. The molecules of bodies may have polarity 

 which is not magnetic at all— polarity, for instance, due to 

 static electrification - under which they group themselves in 



