January 6, 1898] 



NATURE 



^ZZ 



Until comparatively recently, however, accurate magnetic 

 measurements, even at ordinary temperatures, could alone be 

 made on peniianeiit magnets and the forces acting on their poles. 

 The capability of acquiring temporary or induced magnetism 

 when near a magnet, which is the characteristic property of soft 

 iron, could not be subject to strict measurement until it was 

 shown, firstly, how a given specimen of iron could be uniformly 

 magnetised in a uniform magnetic field ; and secondly, how both 

 the magnetising force of that field, and the consequent magneti- 

 sation of the iron could be measured. We know of only one 

 way of completely satisfying the first condition : namely, by 

 covering a ring-formed specimen (whose outer diameter but 

 slightly exceeds the inner) with a uniform layer of insulated wire 

 carrying an electric current, thus forming a ring-magnet with no 

 disturbing poles. And Faraday's researches on electro-mag- 

 netic induction have furnished us with a method, at once 

 accurate and convenient, of determining the magnetisation of 

 the apparently unmagnetised ring magnet, i.e. by the use of a 

 "secondary winding " or outer layer of insulated wire, connected 

 with a " ballistic " galvanometer. Knowing then the intensity 

 of magnetisation called up by a given magnetising force, we can, 

 from their ratio, express the facility with which the iron takes 

 up magnetic induction, or, in other words, its magnetic perme- 

 ability. 



To apply this method to the measurement of permeability at 

 high temperatures, both the magnetising and the secondary 

 winding must be so insulated as to be uninjured by the heating ; 

 and further, the thermometer or pyrometer, which measures 

 the temperature, must be placed inside the ring, so as to 

 measure the actual temperature of the iron. 



Among the earlier important researches on magnetic properties 

 at high temperatures, that of Baur of Zurich, in 1879, should be 

 mentioned. He experimented simply on an iron bar heated in 

 a furnace, and thence rapidly transferred to the interior of a 

 straight magnetising coil. Not, however, till 1889 was it 

 shown by Dr. Hopkinson, in his classical researches, that the 

 ring-magnet method could be successfully applied to the 

 measurement of permeability at high temperatures. The 

 windings of the ring-magnets were in this case of copper wire 

 insulated with asbestos ; the heating was carried out in a gas 

 furnace ; and the rise of temperature of the iron core was de- 

 duced from the increase of the electrical resistance of the 

 secondary winding. 



Of the more recent researches, the most remarkable is that of 

 M. Curie, published in 1895, describing experiments on the 

 magnetic behaviour of a great variety of substances, at 

 temperatures ranging up to a white heat. The method he 

 adopted — that of finding how strongly the specimens were at- | 

 tracted, when placed near a powerful magnet, was well adapted [ 

 to determine permeability in intense magnetic fields, but it is 

 much inferior to the ring-magnet method, where the per- 

 meability varies much with the magnetising force, as is the case 

 whenever the magnetic field is not intense. 



In the course of some experiments on the subject of this 

 article, the results of which have been recently published, I have 

 endeavoured to approximate, where possible, more closely than 

 previous experimenteis to the ideal method which would be the 

 outcome of the principles laid down above. The ring-magnet, 

 whose core was the iron .specimen to be tested at high tem- 

 peratures, was made very small, measuring about one inch 

 across. The temperature of this core was determined by means 

 of an electrical thermometer embedded in it, consisting of a 

 wire of pure platinum whose electrical resistance at any tem- 

 perature had been previously determined, and whose resistance, 

 therefore, if subsequently measured, gave its temperature, and 

 hence also that of the iron core in which it was laid. Asbestos 

 paper insulation, as had been used by former experimenters, 

 was found to be very imperfect at high temperatures, owing 

 largely to carbon deposited from the materials used in its 

 manufacture. This difficulty was, after some trouble, overcome ; 

 but wherever a high degree of insulation was wanted, as in the 

 case of the thermometer wire and secondary winding, it was 

 found necessary to employ mica, though, as may well be 

 imagined, the use of such an untractable material for such a 

 purpose is beset with considerable mechanical difficulties. 



Now, to find the magnetic condition of a sample of iron at a 

 given temperature with any completeness, it is not sufficient 

 merely to measure its permeability in various magnetic fields ; 

 the behaviour of the iron when subjected to what is called a 

 "cyclic" process of magnetisation must be studied — the 



NO. 1 47 I, VOL. 57] 



" hysteresis," or energy, absorbed in one double reversal of the 

 magnetisation of each cubic centimetre of the iron, must be 

 measured. 



But the taking of so many observations requires time ; and if, 

 during this time, the temperature of the iron be not perfectly 

 constant, all efforts at refinement in the magnetic measurements 

 are thrown away. The heating of the ring must, then, be 

 thoroughly under control. The method I adopted was an 

 electrical one. The ring-magnet was furnished with an extra 

 winding of asbestos- insulated platinum wire, so wound as to 

 have no magnetising influence ; and by passing through this 

 wire a suitable electric current, heat could be generated in the 

 ring at any desired rate. This method, however (the principle 

 of which was adopted as long ago as 1888 by M. Ledeboer) is 

 not used to full advantage unless combined with an effort to 

 thermally isolate the body to be heated. Each ring-magnet was 

 therefore thickly wrapped with asbestos, and supported in the 

 centre of a closed and partially exhausted glass vessel (oxidation 

 of the iron core was also in this way avoided). 



This method of heating proved most satisfactory. The loss 

 of heat by radiation and conduction being slight, the ring-magnet 

 could not rapidly alter its temperature ; and there is probably 

 no way in which we can supply heat-energy to a body, which 

 can compete with the electrical resistance method, either as 

 regards constancy or control. For obtaining temperatures up 

 to 1300° or 1400° C. — a white heat — this method, combined as 

 far as possible with " thermal isolation," and an electrical 

 method of measuring the temperature, should in the future 

 prove of the greatest value in all cases where the physical 

 properties of bodies at high temperatures require careful 

 investigation. 



The original intention of my experiments was to ascertain 

 exactly in what way the specific electrical resistance of iron 

 changes at and about the " critical temperature " at which the 

 magnetic properties of iron so nearly disappear, so as to throw 

 light if possible on the molecular state which we characterise by 

 the term " magnetic." With this object the iron core of the 

 ring-magnet was formed ot a long insulated iron strip, among 

 the turns of which the platinum thermometer wire was buried ; 

 and with this piece of apparatus simultaneous measurements of 

 the magnetic qualities and electrical resistance of a sample of 

 iron could be made with accuracy alike at low and high 

 temperatures. 



Let us now consider the magnetic changes which occur in soft 

 iron when heated. At ordinary temperatures iron shows a kind 

 of unwillingness, so to speak, to become slightly magnetised — 

 its permeability under small forces is not great. Beyond a certain 

 limit, however, it exhibits the greatest readiness to become 

 further magnetised, and continues to have a high permeability 

 until magnetised very strongly. But from this point it begins to 

 show signs of magnetic saturation, and ultimately refuses 

 to be further magnetised without the application of very great 

 force. 



Now, as the temperature of the iron is gradually raised, it is 

 found that practical magnetic saturation takes place sooner and 

 sooner. Iron refuses to become so strongly magnetised at higher 

 temperatures. Thus the permeability in strong magnetic fields 

 falls off as the temperature rises — very slowly at first, then more 

 rapidly, till, near the "critical temperature," the permeability 

 rapidly drops to quite a low value. 



On the other hand, in weak magnetic fields, the behaviour of 

 iron up to within a few degrees of the " critical temperature " is 

 precisely opposite. The permeability rises with the temperature 

 — at first slowly, then above 500° C. with ever increasing rapidity, 

 until at last that lack of susceptibility to small forces disappears, 

 and iron shows itself just as amenable to magnetic influence in 

 small magnetic fields, as in the larger ones, where the maximum 

 permeability occurs. 



At this temperature — say 15° below the critical temperature 

 (about 750° C. — a red heat), iron possesses all those qualities at 

 once which are sought after by the transformer maker : — Practical 

 absence both of hysteresis and of eddy currents (the latter owing 

 to the greatly increased electrical resistance), and a permeability 

 nearly four times as great as that attainable in commercial trans- 

 former iron. So magnetic, indeed, is the iron, that even the 

 earth's magnetic field, in the direction of its greatest intensity, is 

 enough to induce strong magnetisation (B = 5000), in fact almost 

 saturate the iron (for which at this temperature a relatively low 

 induction suffices). The behaviour of a compass needle near a 

 slowly cooling ingot of cast steel, should be rather interesting, 



