]\gt ; INDEX SLIP .
JONES , E. TAYLOB .
the Relation between Magnetic Stress and Magnetic Deformation in Nickel .
Phil. Trans. , 1897 , vol. 189 , pp. 189-200 .
Magnetic Stress , effect of , in Wires .
Jones , E. Taylor .
Phil. Trans. , 1897 , vol. , pp. 189-200 .
Magnetisation in Nickel , change of length due to\mdash ; influence of tension on .
Jones , E. Taylor .
Phil. Trans. A. 1897 , vol. 189 , pp. ) .
Nickel , Magnetic Stress and Deformation in .
Jones , E. Taylor .
PhiL Trans. , 1897 , vol. 189 , pp. 189-200 .
VIII .
On the Relation between Stress and Magnetic Jrmation in Nickel .
E. TAYLOR JONES , nicated by Professor ANDREW GRAY , F.R. Received January 13 , \mdash ; Read February 25 , 1897 .
THE object of the experiments described below was to determine how much of the contraction , which occurs in a nickel wire when magnetised , is due to stresses brought into play by magnetisation .
This has been attempted by several experimenters , but it has been assumed by some that the magnetic stress concerned is a contracting stress of magnitude , an assumption which has only recently been shown to be unjustifiable .
* That the system of stresses in a magnetic field , described by MAXWELL Electricity and Magnetism Art .
642 ) as explaining the observed forces between magnetic bodies , is not sufficient to account for the observed deformation of bodies in the field , is clear from the fact that these stresses increase valley with the lllagnetisation and field-strength , while the deformation in iron and cobalt becomes reversed when the field reaches a certain value .
It can also bs shown that MAXWELL'S stresses would produce no change in the dimensions of a ring magnetised by a uniform circumferential field , whereas BIDWELL has found considerable change both in the diameter and the volume of of iron .
The theory of magnetic stress has , however , been extended , chiefly by .
HELMHOLTZ , KIRCHHOFF , S Professor J. J. THOMSON , and HERTZ , 11 who have shown that , in addition to MAXWELL'S distribution of stress , the.re are other stresses in a magnetised body due to the fact that the magnetisation depends upon the strain in the body .
Hitherto no direct experiments have been made with the object of ascertaining whether these stresses , in addition to those of MAXWELL , are sufficient to account for the observed deformation of bodies placed in a magnetic field .
The facts that " " Villari reversals\ldquo ; of opposite kinds exist in iron and cobalt , and also opposite reversals of magnetic deformation , and that in nickel ( at any rate at moderate field* C. CnREX , ' Nature , ' January 23 , 1896 ; H. NAGAOKA and E. T. JONES , ' Phil. Mag May , 1896 .
S. , ' Proc. Roy .
Soc vol. 55 , p. 228 ; vol. 56 , p. 94 .
'Wied .
Ann vol. 13 , p. S 'Wied .
Ann vol. 24 , p. 52 , 1885 .
' Applications of Dynamics to Physics and Chemistry , ' p. 48 , 1888 .
$[ Ausbreitung.der Elektrischen Kraft , ' p. 275 , 1892 .
10.6.97 strengths ) there is neither Villari reversal nor reversal of deformation , suggest that there is a close connexion between the two phenomena , and this general similarity between the of stress on magnetisation , on the one hand , and magnetic deformation on the other , has led some experimenters to assume that magnetic deformation is completely accounted for by the above stresses , and to calculate from the observed deformation the general effect of stress on magnetisation .
Before this assumption can be legitimately made , however , it is necessary to make direct experiments to test its truth , and it was with this object that the following experiments were instituted .
The experiments were made in the Physical Laboratory of the University College of North Wales , and , before proceeding to describe them , I wish here to express my great indebtedness to Professor A. GRAY for kindly placing at my disposal all the necessary apparatus , and for many valuable suggestions .
Thaeory .
One method of experime.nting , suggested by Mr. NAGAOKA and the present writer , would be to measure both the effect of hydrostatic pressure on magnetisation and the magnetic deformation of a ring of soft magnetic material .
From these measurements a direct comparison could be made of the observed deformation and the value calculated from the theory of KIRCHHOFF .
There is however another method which , though theoreticaIJy not so simple , is probably easier to carry out experimentally .
It has been shown by CANTON that the elongat.ion St of an ellipsoid of revolution of great eccentricity and soft magneiic material , when placed in a uniform longitudinal field is , on KIRCHHOFF'S theory , given by ( 1 ) , where length of e lipsoid , magnetisation , are the coefficients of change of magnetic susceptibility with change of density and of elongationS respectively , Modulus for the material , and is defined by \mdash ; \mdash ; \mdash ; rigidity * CANTON , ' Mem. R. Acc .
Linc ser. 4 , vol. 6 , 1890 ; WINKELMANN , ' Handbuch der Phynk , ' Bd .
3 , Part II .
, p. 260 , 1895 . .
cit. , p. 461 .
S The equation defining and is the dilatations parallel to the direction of maguetisation , and to two at right to it .
MAGNETIC STRESS AND MAGNETIC DEFORMATION IN NICKEL .
It can also be shown that if a tension per unit area be applied along uniform wire of the and produce a small increase in the magnetisation , then ' ( 2 ) .
Between ( 1 ) and ( 2 ) the quantity can be eliminated , and the resulting equation is ( 3 ) .
In order to compare this result with experiment a long cylindrical specimen should first be used to determine I and SI/ SP for several constant field-strengths .
Then the specimen should be turned down to the form of an ellipsoid of revolution and the elongation measured with the specimen in the same magnetic state as that in I and 8I/ SP were determined .
At first it was intended to do this , but afterwards it was thought sufficient to use a wire of the material ( oflength very great in comparison with its thickness ) in both parts of the experiment , thus assuming that the elongation of a long thin wire is the same as that of an ellipsoid of revolution having the same dimensional ratio .
The general agreement between the results of different experimenters , who in measuring magnetic elongation have used wires , rips and ellipsoids of various shapes and degrees of purity , shows that this is at least approximately the case .
It can also be proved independently that in a uniformly magnetised wire the term , which is generally by far the greatest on the right-hand side of ( 3 ) , represents the elongation due to those stresses which arise in consequence of the fact that magnetisation depends upon strain .
For the potential energy , which unit volume has in consequence of the magnetisation , is \mdash ; HI .
Hence , if the body has dilatations , parallel to axes of , respectively , the part of the Lagrangian Function depending on magnetisation and strain coordinates is ( supposing the quantities to vary infinitely slowly ) * The symbol is here used for the observed elongation of the specimen , i.e. the ratio of the obsorved increase of length to the whole length .
A change in the state of the surface of the specimen would be caused by the turning down , which would make that process inadvisable .
, " " Electricity and Magnetism vol. 2 , S 632 .
S J. J. Tnonsor , .
cit. Professor J. J. THONSON has shown that the strains due to the of magnetisation upon strain are given by the equations i.e. , O. In case of a long cylindrical wire with axis along ox , we have , and , hence solving for \mdash ; \mdash ; .
( 4 ) .
Now if a tension per unit area be applied to the wire and produce an increase in the netisation and sirains , then But if the elastic properties of the material are not altered by magnetisation , .
therefore Hence ( 4 ) becomes therefore This strain will be the same at all points of the wire , which is supposed uniformly magnetised , and will therefore be that part of which is due to these stresses .
NAGNETIC STRESS AND MAGNETIC DEFORMATION IN NICKEL .
In the experiments nickel was used because the quantities to be measured are both greater and of a simpler nature than in the other magnetic metals .
The specimen was an annealed wire of length centims. and mean diameter millims. , containing about 98 per cent. of nickel and traces of iron and cobalt .
This was obtained from Messrs. JOHNSON , MATTHEY and Co. The magnetising coil was over a metre long , and consisted of seven layers of No. 18 copper wire wound on a hollow brass core provided with a water-jacket through which water could be made to flow steadily .
The coil was mounted on a stand ( fig. 1 ) in a vertical position .
Fig. 1 .
The ends of the nickel wire were soldered into two brass pieces , the upper of which , , was a long cylindrical bar , and the other , , a short block of brass prolonged downwards by a tube closed at the lower end .
Both bar and tube fitted loosely in the core of the coil , the bar projecting and being fixed to a rigid It was at first intended to measure Modnlus for the nickel wire , a measnrement for which the ve apparatus was not very suitable , and which was moreover afterwards rendered unnecessary .
DR. E. TAYLOR JONES ON THE RBLATlOfi support above the coil , and the tube projecting slightly from the coil below .
A thin brass rod was screwed into the block , and passing downwards through an opening in the lower end of the tube , was atbached to a seatepan:below. .
The method adopted for observing changes in the length of the nickel wire was essentially the same as has been used by several experimenters .
At one end of a long light lever , of wood , a sharp gun-metal point pressed upwards against the lower end of the tube , and the fulcrumF consisted of two gun-metal points resting on two plates of gJass fixed on a support which was rigidly connected with the upper support ; the connexion is nob shown in the diagram ( fig. 1 ) .
At the end of the other ( longer ) arm of the lever , a small metal plate was attached , at one point of which a slight depression was made ; in this depression rested one of the three needle-point feet of a small light table on which a plane mirror was mounted .
The two other feet rested on a fixed horizontal metallic support .
Deflections of this mirror were observed by a telescope and vertical scale at a distance of about 3 metres .
Levers of various lengths were tried ; the arms of the lever ultimately chosen were centims. and centims. long , measured from the line joining the fulcrum points to the point at one end , and the depression at the other , respectively .
The distance of the moving foot of the mirror table from the line joining the two fixed feet was millims. The magnifying power of this arrangement was 12,300 .
The nickel wire was thus suspended inside but independently of the coil .
The effect of any " " suction\ldquo ; between the coil and the wire was , however , quite on account of the great rigidity of the support , and because this support was rigidly connected with the fulcrum support of the lever , so that any displacement of the upper support involved a similar displacement of the fulcrum support , and therefore no turning of the lever .
This was tested by observing the scale deflections caused by placing weights on the upper support S. Even if the wire had only one pole , the error due to this cause would be less than per cent. of the deflection observed current was made .
The error due to the slight non-uniformity of the field of the coil towards the ends of the nickel wire was also of the same order .
The magnetising current was generally measured by a Kelvin graded galvanometer , standardised by electrolysis of copper , the smaller currents by a Kelvin centiampere balance .
For measuring the magnetisation , which was done in separate experiments , 600 turns of No. 40 double-silk covered and shellacked copper wire were wound in on .
layer near the middle of the nickel wire .
These were connected to a balistic galvanometer which was wound with wire calculated to give the.greatest sensitiveness .
* A long solenoid with secondary coil was used to standardize this galvanometer in the usual way .
See A. GRAY , " " Absolute measurements in Electricity and Magnetism vol. an p. NAGNETIO STRESS AND MAGNETIC DEFORMATION IN NICKEL .
Measurement of I and The magnetisation was measured by observing the deflection of the galvanometer needle when the magnetising current was reversed .
The sectional areas of the nickel wire and of the coil wound on it being known , and the galvanometer frequently standardised , the induction in the wire , and hence the magnetisation , were calculated .
In order that equation ( 3 ) should hold , it was necessary that the nickel wire should be in the same magnetic state in the magnetisation and in the elongation experiments .
Hence , both experiments were made with increasing reversals of magnetisation .
Also , since during the reversals of magnetisation the magnetic is alternately applied and removed , the used in measuring SI/ SP was added and removed several times before readiugs were taken .
An initial weight of 1 kilogram was always kept in the scalepan and the effect observed at several fields of adding a few kilograms , the magnetisation being measured before and after the additional load was applied .
An increase of tension a.lways caused a diminution of magnetisation , which was not in general proportional to the weight added .
* [ Added May 15.\mdash ; This depends on the field-strength .
At low fields , within the range of tensions used in the present experiments , the effect of ension in reducing magnetisation diminishes slightly as the tension is increased , at stronger fields the opposite is th ecase ; in other words , at low fields is positive , at high fields negative .
This may partly explain the fact observed by BIDWELL that at low fields increase of tension diminishes the magnetic contraction in nickel , at high fields increases it .
For if part of the contraction is represented by then which is negative at low , positive at high fields .
] It could , however , be assumed that the change of magnetisation , divided by the increase of tension per unit area of section of the wire , gave the value of SI/ SP for the mean load .
The curve of incl.easing reversals for the mean load was then determined .
Several loads were tried , but the effects of adding 2 and 7 kilograms respectively sufficiently show the nature of the results .
Thus , the curves of increasing reversals of , and the COlTesponding values of I/ SP , were determined with 2 and kilograms in the scalepau , the total tensions being and kilograms .
Since the dimensional ratio of the nickel wire was about 500 , the mean demagnetising forcet was 00018 I. This was never more than about 0.1 per cent. of the magnetising force due to the coil .
See EWING , ' Magnetic Induction in Iron and other Metals , ' p. 196 , 1893 .
DU BOIS , ' The Magnetic Circuit in Theory and Practice , ' p. 41 , 1896 . . .
DR. E. TAYLOR JONES ON THE RELATIOW All ballistic measurements were repeated a number of times , and mean values taken , to eliminate , as far as possible , small errors of observation .
Measurement of Magnetic Change of Length .
It was necessary to determine the change of length of the nickel wire under of and kilograms , for a series of increasing fields , the current being reversed a few times at each step .
Instead of doing this directly , by first demagnetising the wire and then applying the current , it was found more convenient to measure the temporary and residual changes of length separately .
The temporary change of length was determined by first making and reversing several times a measured current , and then observing the scale deflection caused by breaking the current .
The reversed current was then made , the deflection ' observed on breaking it , and the mean of the two deflections .
taken .
The wire always lengthened when the current was broken .
No complete hysteresis loops were obtained , but the opportunity was taken of noting how much of the residual contraction could be removed by applying a reversed field .
Thus each time the current was broken a small reversed current was applied , causing an increase of length , and this current was gradually increased until the length reached a maximum , and at-higher fields again diminished .
Next a series of residual contractions was determined , the wire being ' by reversals before each reading and a current made for a short time , the scale reading being aken before the current was made and after it was broken .
The residual contraction was always found to be greater than the contraction which can be removed by applying an increasing reversed current .
This has also been observed by Mr. NAGAOKA .
* As the magnetising field is increased both the residual and the " " removable\ldquo ; contraction at first increase and then become nearly constant , the former being then about and the latter of the length of the wire .
reversed field corresponding to minimum contraction was about 16 C.G.S. The temporary contraction was , however , still increasing in the highest fields employed , value at being about 33 X , with a tension of kilograms .
The curve representing the total contraction was obtained by adding the ordinates of the " " temporary\ldquo ; and ' ' residual\ldquo ; curves .
-The difference between the tensions and kilogs .
weight was not sufficient ' Phil. Mag Jan. , 1894 , p. 131 .
Since about a centimetre of the niok wire at eaoh end was soldered in pieoa A and the length of wire whose obanges were meaeured } oentims , This length was used in oalouluting NAGNETIC ANB-NAGN DEFORMATION IN NICKEL .
to cause much difference in curves .
At the- highest fields used the contractions were , however , rather greater with the greater tension .
At low and stren ths the contraction is less , but at strong fields greater , than the value obtained by Mr. NAGAOKA for an ellipsoid of nickel .
* Water was kept flowing through the coil for a considerable time before readings were taken , and all temperature changes took place so slowly that their effects were easily distinguishable from the magnetic effect .
All the readings were repeated several times , and the points representing the results always lay very near the curve ( fig. 3 ) .
Final Results .
On the right-hand side of the equation ( 3 ) third term is much greater than the other two .
Hence it was not necessary to know and with very great , accuracy .
The value of YOUNG'S Modulus for a per cent. nickel wire has recently been given by as C.G.S. value of was used , and the rigidity was measured by observing the torsional oscillabions of the wire , to which a cylindrical vibrator was attached .
This gave approximately .
Hence The greatest numerical value of obtained was about while the greatest value of the first two terms in , representing the elongation due to MAXWELL'S stresses , was but Tables I and II contain the values of , the change SI of magnetisation caused by increasing the tension by per unit area of section , the values of Sl/ l calculated from equation ( 3 ) , and the observed total contraction TABLE I. Load kilogs .
section of wire kilogs .
Temperature C. .
'Wied .
Ann vol. 53 , p. 487 , 1894 .
'Wied .
Ann vol. 59 , p. 668 , 1896 .
DR. E. TAYLOB JONES ON THB TABLE II .
Load kilogs .
SP section of wire kilogs .
Temperature C. Fig. 2 .
BTIO STRESS AND MAGNETIC DEFORMATION IN NICKEL .
In fig. 2 , Curv A is the curve of increasing reversals of magnetisation with a load of kilogs .
The difference of the ordinates of Curves A and represents the MAGNETIC AND MAGNETIC NlOKBL diminution of magnetisation caused by adding two kilogs .
to the load , and the rence of ordinates of A and represents the effect of adding seven kilogs .
to the load .
In fig. 3 curve represents the observed total contraction , curve the residual contraction after the field has been removed , the least contraction atlainable by reversing the current after the field has been removed , the value of ca , lculated from equation ( 3 ) , and the difference , all as functions of the field , the load being kilogs .
The curve therefore represents the contraction in nickel corrected for the effects of known stresses .
In the last columns of Tables I and II , the corresponding values of are given .
It will be seen that these numbers are approximately proportional to the numbers in the preceding column representing the values of This is also shown in fig. 4 , where the abscissae of the points . . .
proportional to , and the ordinates to the corrected contraction - for the load kilog .
The points all lie , to within about 5 per cent. , on a certain straight line passing through the origin .
An error of 5 per cent. in the values of and might be caused by much smaller errors in the measurement of I. Especially was this the case with the other load kilog .
, for the change of magnetisation in this case ( Table I ) was smaller , and therefore more difficult to measure accurately , though even in this case there is no regular deviation from proportionality between and With the load kilog .
, however , the change of length observed in the nickel wire is closely represented by the equation where has the value at the temperature C. , and where the second and third terms on the right-hand side are numerically very small in comparison with the first and fourth .
[ Added May -The result that the magnetic contraction in nickel , corrected in the manner described above for the effects of KIRCHHOFF'S stresses , is under certain conditions approximately proportional to , is at present to be regarded as purely empirical , and without further experiments it cannot be said to be generaly true .
lt is , therefore , proposed to continue the investig tion by repeating the experiments on the nickel wire under different conditions , especially as regards temperature .
It should be borne in mind that in the deduction of the theoretical value of the magnetic contraction , the material is supposed to be perfectly " " soft and no account is taken of hysteresis .
Some experiments made by NAGAOKA on an ellipsoid of nickel ( ' Wied .
Ann 53 , p. 496,1894 ) seem to show , however , that the contraction in this metal depends only on the value of the magnetisation , being almost independent of the manner in which that .
value has been reached .
It is , therefore , unlike1-..ythat:any considerable pancy can arise in consequence ]

