3o6 



NA TURE 



{July 29, 1880 



results as I had got before. I then gradually raised the teai- 

 perature till p.irt of the wire was bright red, and finally allowed 

 it to cool. As the temperature rose, the original fizzing sound 

 died out then the high note became inaudible, then there was a 

 short interval of almost complete silence ; after that a high note 

 came in, then the fizzing sound again, which very quickly 

 changed into a deep buzz, accompanied by a very low note like 

 that of the tnning-fork, a note of medium pitch, and a high note 

 (and possibly others), then the buzz died out, and at last the high 

 note was left. When the wire cooled, the phenomena recurred 

 in the corresponding order. First the buzz came back along 

 with the low and medium notes, then it died away, and the high 

 note alone was left ; then there was silence, then the high note 

 acain, and lastly the fizzing sound. 



Most of the notes heard, certainly the most prominent of 

 them, appear to have little relation to the tuning-fork. They 

 seemed to be affected to some extent by the tension of the wire. 



When a magnet was brought up to the wire the deep note, ob- 

 tained in a similar way with w ires of other metals, was heard 

 along with those peculiar to iron and steel. 



These experiments with the steel wire appear to me to settle 

 the question as to the cause of the sound in thick iron wires. The 

 fact that the wire can be put into a condition in which no sound 

 is produced, and then made to sound by magnetising it, show s 

 that the action is due to the magnetism of the wire, and is aisij 

 an additional proof that the earth's magnetism had nothing to 

 do with it. 



This view is still further confirmed by the effect of heat on the 

 tempered steel wire. The first effect of heat is to destroy the 

 permanent magnetism of the wire, hence the initial diminution 

 of the sound. At a temperature of about 250° C. the permanent 

 magnetism is much reduced. 



On heating farther, the magnetic susceptibility of the steel 

 begins to increase rapidly, until it reaches a maximum about dull 

 red, and then it falls off again very rapidly ; hence the increase 

 of the sound to a maximum and the final falling off. 



The reason why the minimum cannot be observed with iron, 

 and not always with soft steel, is that with them the permanent 

 magnetism is less and the magnetic susceptibility in general 

 greater at ordinary temperatures, so that the increase of the latter 

 begins sooner and masks the decrease of the former. At a dull 

 red all kinds of iron or steel are much on a par as to suscepti- 

 bility ; hence in the ca'^e of hard steel, whose susceptibility 

 begins to increase rapidly only at a pretty high temperature, the 

 phenomena are much more striking, as well as more varied than 

 in the case of soft iron. 



On cooling the sound came and went again as usual, leaving, 

 however, a permanent sound of considerable loudness, which 

 was increased by repeated operations of this kind. 



As a test of the soundness of the above conclusions I was 

 anxious to examine the behaviour of the other strongly magnetic 

 metals, and Prof. Tait kindly put several pieces of nickel and 

 cobalt at my disposal. 



The piece of nickel used was 3 cm. long, 2 mm. broad, and 

 about '6 mm. thick. It was hard soldered to platinum 

 terminals, and mounted in the usual way, after being heated 

 red hot and dipped in water at dull red. 



At first it gave a very feeble high note, accompanied by a 

 gentle fizzing sound. One stroke with a m.agnet caused it to 

 emit a loud buzzing sound. On heating gently this sound was 

 somewhat reduced, and on heating farther the hissing sound 

 died away, and a high note was left, but it too was extinguished 

 before the nickel was visibly hot. 



I made some temperature determinations by means of an air- 

 bath and a mercury-thermometer, and found that at 200° C. the 

 buzzing noise first began to be softened down. After 250° C. 

 the diminntion appeared rather more rapid, but at 350° C. the 

 sound was still quite loud ; after that the falling off was very 

 rapid, and somewhere (say 400° C.) beyond the range of the 

 thermometer, the mercury in which just boiled at the end of the 

 experiment, the sound died out rather suddenly. 



The behaviour of nickel is therefore exactly what we should 

 expect from its magnetic properties, for it loses its magnetic 

 susceptibility, according to Faraday and others, somewhere 

 between 350° C. and 400° C. 



I found with nickel, as with iron, that the current itself at a 

 certain high temperature could produce much the same effect as 

 1 got by magnetising. On testing a piece of nickel after being 

 magnetised by the current I found it to be transversely magne- 

 tised. This induced me to try magnetising my nickel strip 



transversely, but although I got results this way they >vere not 

 so good as I had got by magnetising longitudinally. 



I was thus led to try the following experiment, the result of 

 which is at least curious. Instead of passing the current through 

 the nickel itself as before, I passed it round two flat pieces of 

 iron electro-magnet-wise. These were placed with their ends 

 pretty close together, and the nickel was stretched between them 

 so that it lay in a nearly uniform field of magnetic force, whose 

 strength varied in unison with the interrupter. 



I found that with this arrangement the nickel sounded very 

 much as it did when the current was passed directly through it. 

 The sound was not so loud, but its quality appeared to be the 

 same. The sound, however, was loudest when the plane of the 

 nickel strip was parallel to the lines of force, being very feeble 

 when the plane of the strip was perpendicular to the lines of 

 force. 



A piece of watch-spring was tried in the same way, with 

 exactly similar results. 



This experiment is of course very nearly the same as some of 

 those by which the sounds due to the magnetisation and demag- 

 netisation of iron are usually demonstrated. A very full account 

 of these sounds will be found in Wiedemann's "Galvanismus," 

 Bd. ii. p. 565 ct scqq. 



I tried a piece of cobalt 6 cm. long and 6 mm. broad, ^^ mm. 

 thick, in the ordinary way. In its original state it gave no sound 

 whatever. After being magnetised longitudinally by a large 

 number of strokes it gave a sound, very feeble, however, com- 

 pared w ith that got in the same way from iron and nickel, or 

 even from hard steel ; it was, moreover, more of a pure note 

 and less of a hissing noise. Heating in the first place diminished 

 this initial sound, so that there came an interval of comparative 

 silence, then the sound rose again, and by and by the familiar 

 buzz came in ; but up to a bright red heat no maximum was 

 reached. On cooling, the phenomena reappeared in the proper 

 order. 



Cobalt behaves, therefore, just as we should expect from its 

 refractory magnetic nature. 



I may mention one curious phenomenon that appeared once ot 

 twice with cobalt and once or twice with a piece of steel. On 

 cooling, after the maximum was past, the buzz had died away, 

 and a period of comparative silence had come, strong beats 

 began to be heard, which lasted for a considerable time, and 

 then died away as the temperature fell. Various causes for 

 these might be assigned. It might have happened that two 

 parts of the metal were at different temperatures, and gave notes 

 nearly in unison. It may very well have been interference 

 between notes due to permanent and temporary magnetism ; for 

 in cobalt generally, and with the particular piece of steel in 

 question, the minimum was not marked by the absolute silence 

 which probably indicates cessation of the sound due to 

 permanent magnetism before that due to temporary magnetism 

 begins. 



Relation to Thermo-electric Properties. — As it seemed to be of 

 some interest to connect these magnetic sounds with the curious 

 thermo-electric peculiarities of iron and nickel brought to light 

 by the recent researches of Prof. Tait,^ I asked the help of his 

 assistant. Dr. C. G. Knott, who has had great experience in 

 work of this kind. 



The sounding-wire, a short piece of which was always used in 

 order to get the phenomenon pure, was inclosed along with a 

 double or triple thermo-electric junction in a small tube made by 

 rolling up a piece of sheet-copper. The tube was then heated 

 up in the blowpipe flame. This was a rough way of setting to 

 work, but it was sufficient for our purpose. 



The diagrams (Figs, i, 2, 3), made by Mr. Knott, with the ap- 

 pended notes, will show the results. I have given the observations 

 made during heating, as being on the whole probably nearest 

 the mark. The cooling, except in the case of nickel, which was 

 inclosed in a wide iron box, and did not require to be raised to a 

 very high temperature, was much more rapid than the heating, 

 and consequently inequalities of temperature due to the different 

 positions of the sounding-wire and the junction would have 

 been' more apparent In point of fact the discrepancy was not 

 greaf"^. 



The abscissa in the case of nickel is the temperature m centi- 

 grade degrees, in the other cases it is the electromotive force of a 

 junction formed of a certain pair of platinum iridium alloys 

 (called M and N) much used by Tait in his thermoelectric 

 researches, because their lines on his thermoelectric diagram are 

 ■ Trans. R. S. E., 1S72, 3, vol. x.tvii. p. 134, &c. 



