]\gt ; .
R. Shida .
" " Experimental Determinations of Magnetic Susceptibility and of Maximum Magnetisation in Absolute Measure By B. SHIDA , Thomson Experimental Scholar , University .
Glasgow .
Communicated by Sir W. THOMSON , F.R.S. Received October 10 , 1882 .
Read November 23 , 1882 .
The fact that there exists a limit to the magnetisation of a soft iron bar was first demonstrated by Joule , who , in made a number experiments on the sustaining power of an electro-magnet , and showed that when the current in the exciting coil is made stronger and stronger , that power tends to a certain definite value , or in other words , the magnetisatiou of the iron core attains a maximum .
In 1861 , an interesting research on the magnetic properties of irou was made by Thalen , who determined , among other things , magnetic susceptibility of differe1lt specimens of soft iron in absolute measure for the first time .
The units of length , mass , and time employed by Thalen were respecbively a millimetre , a milli amme , and a second .
Joule and Thalen were followed by several , most of whom , however , made experiments without giving the results in absolute units ; bnt amongst the few who have not overlooked the importance of such a system of units , Rowland made by far the most important investigations upon the subject .
He determined not only the magnetic permeability or susceptibility of certain so-called magnetic bodies , but the maximum magnetisation of those bodies in absoIute units , using the metre , the gramme , and the second as the units of length , mass , and time .
The method of Thalen and that of Rowland are essentially the same , inasmuch as they depend upon the same electrodynamic principle , that an electric current induced in a closed circuit due to sudden creation or disappearance of magnetic lines of force , is proportional to the number of lines of force tbus introduced or withdrawn , cutting the circuit .
But one notable difference of the two methods lies in the fact that the one used ellipsoids or cylindrical rods of great length , while the other chiefly used or endless rods to experiment upon .
The chief advantage of an electromagnetic method such as the above , is , as been remarked by Sir William Thomson in his paper on the " " Electrodynamic Qualities of Metals , Part \ldquo ; the ease and rapidity * Joule 's Collected Papers , page 84 , from " " Sturgeon 's Annals vol. , page 187 Sir William Thomson , ' Papers on Electricity and Magnetism p. 472 .
" " Phil. Trans 1876 , p. 693 .
S Determinations of Magnetic usceptibility .
405 which the results can be obtained ; while its disadvantage is revealed in the fact that it does not show either slow changes of magnetisation or the distribution of magnetism .
The following resuIts of the experiments which have been made at the Physical aboratory of the University of Glasgow , are given in absolute measure in which a centimetre , a gramme , and a second are taken as the units of length , mass , and time respectively , and were arrived at by means of the direct magnetometric method given to me by Sir William Thomson ( who described and the method at the recent meeting of the British Association at Southampton in the Section A ) , as founded upon a method originatel by Coulomb and discussed mathematically by Green .
This method possesses some imporant advantages over the eIectromagnetic method ; for instance , it shows at any moment any change of magnetisation of the body experimellted on hich is of great practical utility in investigat ons this kind ) ; it affords an excellent means of illustrating the distribution of magnetism in the body .
and it enables us to experiment upon a long thin bar , subjecting it to different strengths of magnetising forces , and to various amounts of longitudinal stress , and at the sftme time to determine in absolute measure , the magnetisation and magnetic susceptibiliby of the bar under these varied circumstances , which is an origina ] feature of the investigations I am going to describe .
These advantages , however , do not exist without disadvantages .
That the execution of careful investigations involves a considerable amount of time , is a serious disadvantage of this method .
After some preliminary studies , the orderly experiments were commenced about the middle of February last , and have since been carried on from day to day without intermission up to the end of May .
A number of thin wires and of thick bars of iron and steel were experimented upon .
The accompa1lying sketch ( Plate 9 ) shows the arrangement of the apparatus employed in on bhin wires .
A reflecting magnetometer , which consists of a mirror carrying at its back three small magnets and suspended by a single silk fibre about 5 centims. long , was placed on a convenient stand nearly 2 metres above the floor of the laboratory .
is a while paper screen divided into half millimetres , and bent into a circular arc of a metre radius .
It is fixed at a distance of exactly 1 metre from the magnetometer , and was used to observe the deflections of the magnetometer needle , which were read by the image of a fine wire fixed vertically in front of a paraffin lamp , , secured just behind the scale as iu a Thomson reflecting galvanometer .
is a magnet of semicircular shape meant to control the strength of the field at the point where the magnetometer needle is suspended .
It was mounted on a suitable stem in front of the magnetometer needlo , with its length in the plane of the magnetic meridian , and at a certain distance The wire to be experimented upon is represented by AA ' .
It is hung vertically a distance of 10 centims. from , and due magnetio east of the magnetometer needle , by means of an arrangement of pulleys , and weights , , each weighing about half a ramme and attached to one end of the cords , , respectively as shown .
In ordor that the vire may easily be detached from the I cords , the other end of each cord , instead of being fastened direcbly to the end of the wire , is merely hooked , by a small brass hook nich it bring iatter pafter hgmode oastening tithout twisti oendi tatter icenti 1ning towed aheavy weight tcentims.ntwisted ortion , over which ontwisted , tring wgreat number oimes T degree .
BB ' is the magnetising coil hung in such , manner from the string that boffi its centre and axis coincide those of the wire AA ' .
The coil used in the first part of the experiments was composed of only one layer of silk-covered copper wire wound on a straight brass tube of about 6 millims. in its internal diameter ; the length of coil was 108 centims. , its radius was 34 of a centimetre , and the number of turns of wire on the coil was 1,795 , and the resistance of the coil including the electrodes was , at C. , ohms .
By means of this coil were obtained the results given in the columns headed 1 to 5 of the Table I. It was soon found that the coil just described was quite unsuibable for producing high magnetising forces , and thaI a modification was necessary .
The coil , when modified , was 110 centims. long , and consisted of five layers of silk-covered copper wire laid on one above another ; the radius of the innermost layer was 840 of a centimetre , and that of the outermost was 660 of a centimetl.e , and , efore , the mean radius of the coil , and the mean distance between any two adjacent layers 08 of a centimetre ; the resistance of the coil , the electrodes included , was ohms .
at C. The ends of the electrodes of the coil were permanently connected to the two terminals of a reversing key , , the other two terminals of which were in connexion with the two electrodes of a Thomson tray battery so disposed that any desired number of cells , from 1 to 60 inclusive , could be placed in the circuit .
A into galvanometer , , was inserted in one of the connecting wires as shown in the sketch , so that whenever a current is passed throngh the Determinations of Jfagnetic Susceptibility .
it was read and measured by this galvanometer .
The weight was simply used to balance the weight of the coil .
It will easily be seen from the angements of cords , pulleys , \amp ; c. , that the wire , besitdes being kept straight , can be raised or lowered hrough any desired distance within a range of about 4 metres ; and further , that when the wire is moved uP or down the coil follows the movements , keeping its position with reference to the formel .
unered .
For the purpose of observ1ng the position of the wire or the coil with great facility at any instant with reference to the line on a level with the magnetometer needle , there is provided , alongside the wire and coil , a scale divided into , and fixed to a wooden upright .
The orderly and systematic way in which the experiments were may be described generally thus : weight , the amount of which was different for specimens of the wire as will be presently stated , was put on and take1 ] off the wire , whilst the magnetising force was in action , about ten times success on ( this operation of successive application and removals of a we a wei ht will hereafter called , for brevity , " " ons and offs half a kilogramme being always on ; then the vire , having been first placed so bigh up that its effect and that of the coil on the magnetonleter was scarcely visible , was lowered 2 centims. by 2 centims. , until it was so low down that little or no eflect of the wire and coil was observable on the magnetometer , while the deflections of the magnetometer needle were noted for all the positions of the wire and coil .
This process was followed in the case of all the wires , except the hard-tempered wire , and the magnetising forces used , unless otherwise stated .
It will be needless to enter into the discussion of the details of the object of subjecting the wire to the operations of " " ons and ofls as they will , I hope , shortly communicated to the Royal Society elsewhere ; suffice it to point out that on commencing ffie prellmlna.ry xperiments , it was soon discovered that in the first instance the wire was very irregularly magnetised , but that the effect of subjecting the wire , while under the influence of the vertical force , to the application and removal of , pull a certain number of times in succession , was to remove all the irregularities as to magnetisation , besides producing an enormous augmentation of its magnetism .
The results are given in the Tables I to .
The general explanation of these and other accompanying tables is , that the " " Distances mean the distances of the centre of the wire trom the level of the magnetometer needle , those distances measured from their level upwards being reckoned positive , and those measured wards .
negative ; while the " " Deflections\ldquo ; mean the correspoIldingdeflectlons of the needle in the sions \mdash ; those deflections indicating the repulsion of the north-seeking pole .
or end of the needle , The first wire tried was a very soft iron ( pure ) wire , *supplied by Johnson and Nephew , Manchester , and is named in the table " " Dark Wire from its appearance .
It was of No. 10 B.W.G. , its breakiug ! .
stress being about 15 kilogs .
The piece experimented on was a metre long ; its radius , when carefully calculated from its weight and specific ] gravity , was 0374 of a centimetre , and therefore its sectional area was 00439 square centim. The weight which was used for the operation of ' ' Ons and Offs\ldquo ; was 8 , only with this exception , that at the beginning , while the force magnetising the wire was that due to the vertical component of the earth 's magnetism alone , a weight of 10 kilogs .
was put on once or twice .
The wire underwent an olongation of per cent. of its original length , so that it was now 1029 centims. , and its sectional area 00425 square centim. ; the tion was permanent and constant , that is , the quent application of 8 kilogs .
produced no more effect as to elongation .
The results for this are shown in the Table I. In this table , the results under the headin numbered 1 , which are those for the Glasgow vertical force , it must be mentioned , were obtained after the wire had been treated in the following manner : operation of " " Ons and Offs of a weight of 8 kilogs .
, having been performed while the wire was hanging one way , say , bhe end A up , its magrletisation was observed in the manner explained before ; the wire was then inverted , and the operation of " " Ons and Offs\ldquo ; was again performed while it was hanging with the end up , that is while the vertical force was acting in the opposite direction with respect to the wire , and its magnetisation was observed ; this process was repeated until the magnelisation of the lvire in the two cases was equal , or nearly so , in intensity , but opposite in polarity .
The first and second columns under any of the headings numbered 1 to 8 give the result obtained in the two cases respectively : ( 1 ) while a weight of 8 kilogs .
was actualIy hanging on the wire a case to be hereafter denoted by " " On and ( 2 ) the weight was off case to be hereafter denoted by " " Off ; and the third column , if any , ontains the result obtained for the effect of the coil alone carrying a current .
The first column under 12 and 13 contains the resuIt obtained ( in the case " " Off immediately after reversing the current in the coil , the operation of " " Ons and Offs\ldquo ; having been of course erformed before the current was reversed ; while the second and * This wire is of the same kind as that used in the experiments described in Sir William Thomson 's paper , " " On the odynamic Qualities of Metals , Part VII section was 004462 square centim. A weight of 16 kilogs .
-as always used for " " Ons and Offs No elongation of the wire was : observed.irst aecond cumns under areceding third chere bneings iemarks ahose ingives tempering which wdopted ierhaps worthy opassing noticeA convenient 1ength wfwhich axhibited iodeThe 1ires experimented o $ Further particulars regarding the elaeticity , \amp ; c. , of this wire are found in Mr. , T. Bottomley 's interesting paper on the " " Effects of Long-continued on the Elasticity of \ldquo ; Proc. ) .
Soc vol. xxix ( 1879 ) , p. 221 .
wire ; the tempering was , of course , as uniform as it could be all over the length of the wire , perha , ps , with the exception of the ends it was held .
When short pieces were cut off from the extremities , the wire was centims. long ; the area of its cross-section was now 004326 square centim. , the wire having lost nearly 2 per cent. of its veight bythe process of tempering .
This wire was , of course , so exceed .
ingly brittle that the operation of " " Ons and Offs\ldquo ; of a heavy weight was an impossibility .
and consequently no weight was put on the wire at all , except those used to keep it vertically straight .
With reference to the explanation of the Table , the first column in 1 , 2 , 3 , \amp ; c. , refers to the result arrived at when the magnetising force was kept acting on the wire ; and second column , if thers be one , refers to the result arrived at directly after the withdrawal of all magnetising force , except that due to the vertical component of the earbh magnetism .
Somewhat thick bars of cast iron , hard-tempered steel , and soft iron , were then procured and experimented upon , with a view to determine approximately the law of magnetisalion of those bars , and to compare the results with each other and with those for the wires .
The bars were nearly equal in their dimensions ; they were a1161 centims. in length and very nearly square in section ; the sectional area of the cast-iron bar , when calculated from its weight and specific gravity , was approximately 950 square centim. , that of the steel bar 948 square centim. , and that of the soft iron bar square centim. With regard to the mode of experimenting in the case of these bars , though it remained the same in principle as before , it necessarily diflered in details , which I proceed to describe thus:\mdash ; In first , the coil employed for magnetising the bars was 68 centims. long , and consisted of three layers of insulated copper wire on a tube of copper nearly square , each layer containing 620 turns ; the whole area enclosed by all the turns of wire per unit length was 89 .
centims. approximately , though not very accurately on account of the difficulty of measuring exactly the dimensions of the coil , as it was not specially made for the purpose ; and the resistance of the coil was about ohms when cool .
The bar to be experimented on was placed inside the coil , with its centre and axis coincident with those of the latter ; and the whole arrangement thus fitted up was hung vertically in the same way as before by means of cords , pulleys , \amp ; c. , with the common axis of the coil and the bar at a distance of 22 centims. from , and due magnetic east of , the magnetometer ; the conDetemn tations of Magnetic Susceptibility .
nexions of the electrodes of the coil , the galvanometer , \amp ; c. , being precisely the same as before .
The same procedure in experimenting as before was followed as far as possible ; that is to say , the bar and the coil , having been placed high up to begin with , were lowered 2 centims. by 2 ce , ntims .
until they were low down , while the deflections of the magnetometer needle were read for all the positions of the bar or the coil .
In the case , however , where this procedure was hardly possible , or , at any rate , hardly worth going through , on account of the rapid variation of the current in the coil , arising partly from the heating up of the coil and partly from the polarisation of the battery ( which consisted either of the Thomson tray , Daniell 's , or of the Faure accumulators , the latter being chiefly used to obtain very high magnetising forces ) , the experiment was made in the following manner:\mdash ; A point of the bar , 28 centims. distant from its centre , having been placed on a level with the magnetometer needle ( as this position of the bar was such as to give the needle a maximum deflection for a high magnetising force ) , a strong current was allowed to pass through the coil , and as soon as the deflection of the needle was readable with a tolerable accuracy it was read off at a certain moment by one observe , while the rength of the current was measured by taking the reading of the galvanometer at the same moment by another observer on word of command from the former ; the data thns obtained will , as we shall , afford the means of determining approximately the magnetisation of the bar .
The resuIts of experiments on the bar of cast iron , steel , and malleable iron , are given in the Tables respectively , the general explanation of which has been already given in dealing with the other tables .
The first column undel .
any of the headings 1 , 2 , 3 , \amp ; c. , in each of the Tables , contains the results obtained while the netising force was in action ; while the second column , if there be one , contains the result obtained directly after the withdrawal of the force .
Now the best way to study the results given in all the Tables I to VII , is to plot curves in such a manner that the ordinates represent the " " Distances\ldquo ; of the of the wire or bar from the datum line , the level of the magnetometer needle , and the abscissae represent the " " Deflections\ldquo ; of the needle in the scale divisions .
To illustrate this , the results shown in the second and third columns under 7 , Table I , are exhibited by the curves 1 and 2 respectively , Plate 10 , in which those distances measured upward from the datum line are reckoned positive and those measured downwards negative ; while those deflections indicating the repulsion of the red end of the needle are reckoned positive , and those indicating the attraction negative , according to the couvention already adopted .
412 Mr. R. Shida .
Determinations of Magnetic Susceptibility .
Mr. R. Shida .
Determinations of Wagnetic Susceptibility .
Mr. R. Shida . .
Determinations of Susceptibility .
417 : eterminations of jeptibility .
Mr. R. Shida .
\amp ; S Deteeminations of usceptibility .
Determinations ibilit .
Table Glass-hard-tempered Steel Wire .
Mr. R. Shida .
Table ( continued ) .
Glass-hard-tempered Steel Wire .
Determinations of Oagnetic tusceptibility .
Mr. R. Shida .
Table \mdash ; Steel Bar , Hard-tempered .
Mr. R. Shida .
Let\mdash ; area contained by the curve 1 , the axis or , and the line or the line area contained by the curve 2 , the axis or OY ' , and line or the ] half the length of the or bar .
half the length of the coil .
distance of the middle line of the wire or bar from the magnetometer needle .
sum of all the netic matter , northern or southern , on either side of the centre of the wire or bar .
of the solenoid or coil .
strength of the field at the point where the magnetometer needle hangs .
angle of deflection of the needle , in radian measure , corre sponding to the division of the scale .
tensity of the magnetisation of the wire at any cross .
section , or intensity of magnetisation of the bar at its centre .
ising force .
magnetising susceptibility .
area of the cross-section of the wire or bar .
Then it can easily be proved , provided that the angles of deflections are so as to be proportional to their tangents , as in the case we are considering , that .
S. is the integral sum of all the normal components of forces over the whole surface of a cylinder whose height is the of the wire or bar , and whose radius is , due to the netic matter , situated at the centre of the cylinder , provided the length of the wire or bar be infinitely great ; the correction for this length being tead of infinite , is such that must be added to the above quantity to get the integral in question , , however , tlIe sum of all the normal components due to , situated in the axis of the cylinder at a distance from its centre , over that end of the cylinder which is farthest from .
But the integral of the normal force over any closed surface due to magnetic matter inside is , ; hence .
S. ; and hence .
S. \mdash ; Let now .
S. , and then . . . . .
( 1 ) .
Determinations of gnetic Susceptibility .
435 Similarly , .
S. and hence , say ( 2 ) , therefore or , ( :3 ) .
Also , . . .
( 4 ) , and hence , in the case of thin wires , . . . . . . . .
( 5 ) .
The equation ( 2 ) gives us a means of ascertaining the value of if we know that of , as in the case of 7 or 8 , Table I. In the case where was not directly obtained by observation , was calculated from the following formula , . . . . . .
( 6 ) , where A is the area contained by all the turns of coil per unit length , and is the current strength in the coil .
In the case of a cylindrical coil , ; in which is the number of turns of wire per unit length of the coil , the mean radius of the coil , the number of layers , and the mean distance between any two adjacent layers .
As to the evaluation of the magnetising force F. Let be the mean radius of the cylindrical coil , or what is equivalent to it if the coil be not cylindrical ; theu the magnetising force at a point in the axlq of the coil at a distance from the centre , , and the same signification as before , is , ( 7 ) .
At the centre of the coil , if be very great compared wibh . . .
( 8 ) .
Now it will be observed , as the equation ( 7 ) will show , that in my *Papers on " " icity and tism William Thomson , p. 472 ; or Maxwell 's " " Electricity and Magnetism vol. VOL. XXXY .
The current strength was always measured on a Thomson galvanometer , , except when it was so weak that a small error in the galvanometer reading will produce a considerable error in the : resuIt , in which case the current was estimated from the force of the battery and the resistance of this circuit .
The strength , , of the field was calculated in terms of , the horizontal component of the terrestrial magnetism , simply by compal'ing the deflections of the magnetometer needle acted upon by a magnet ( placed behind and at a convenient distance from the needle , and with its length in the line at right angles to plane of magnetic meridian ) in the two cases : ( 1 ) When the field was due to the horizontal component alone ; and ( 2 ) , when it was due to both the controlling magnet and the horizontal force H. Since evidently the value of seriously affects the resuIts , it was thought desirable to make a fresh experiment to determine at the very spot where the magnetometer needle is suspended .
This was eflected indirectly by counting the periods of a magnetic needle at the point in question , and at another point where the exact value of was knolvn from a direct experimental determination made after the manner described in my paper on " " The Number of Electrostatic Units in the Electromagnetic Unit\ldquo ; Phil. Mag for December , 1880 ) , or more fully explained in Mr. Thomas Gray 's paper on " " The Experimental Determination of Magnetic in Absolute Measure\ldquo ; Phil. Mag for November , 1878 ) ; the value of at the point where the magnetometer hangs was found to be 1590 .
The value of V , the earth 's vertical force , is of by far the less moment , considering that the only results the accuracy of which depends greatly upon that of the value of , are those for for that particular magnetising force only ; so that it was deemed unnecessary to find by a new experiment , and consequently it was deduced from the value of and that of the dip , being taken ffi the latter according to the determination made some three years ago .
The final results tabulated at the end of the paper , namely , in the Tables , \amp ; c. , were derived from the mathematical considerations above discussed , and from the results given in the col.responding Tables of Deflection I , II , III , , \amp ; c. , with the exception of the results given in 7 , 8 , 9 , and 10 of the Table , and 2 , 3 , and 4 of the Table F. The intensity , I , in these exceptional cases was obtained by assuming that the deflectiions of the magnetometer needle due to the magnetism of the bar alone that is to say , the Determinations of Magnetic Susceptibility .
deflections due to the coil being taken into consideration ) , correponding to the distance 28 centims. ( a distance approximately corresponding to a maximum deflection for high magnetising forces ) are proportional to the intensity I. The deflections due to the coil alone were calculated from bhe strength of the current in the coil after a manner to be discussed later on .
Just a few words are perhaps necessary to explain the details of the Tables , \amp ; c. In the first place the resuIts given in the first , econd , third , \amp ; c. , horizontal lines along with the numbers 1 , 2 , \amp ; c. , in Tables \amp ; c. , correspond to the first , second , third , \amp ; c. , vertical columns under the headings 1 , 2 , \amp ; c. , in the corresponding Tables of Deflection I , II , III , \amp ; c. No sign or a negative one is pretixed to the numbers , according as the polarities of the wire or bar or coil were similar or dissimilar to those induced in the wire by the earth 's vertical force alone , if the numbers refer to quantities indicating the ernetisation ; or as the netising forces were in a similar or dissimilar direction to the vertical force , if the numbers refer to the quantities representing the magnetising forces either directly or indirectly .
Again , it will be observed that in 7 and 8 of the Table , and of tlJe Table , there were obtained two values for , one calculated and the other observed ; the object of this was twofold : ( 1 ) To that the calculated value was within the erl'ors of observations in the measurements of the current strengths , the dimensions of the coil , \amp ; c. ; and ( 2 ) , to render the resuIts for this maximllm magnetisation of the wires corresponding to these tables independent of .
the accuracy or inaccuracy of the measurement of the current strengths ; the observed value for was used , in these cases , to evaluate the quantities I , , \amp ; c. The rest of what is given in the Tables , \amp ; c. , will , I hope , explain itself .
But by far the readiest mode of studying the whole results , is to refer to the graphical representation shown in the Plates 11 , 12 , 13 , 14 , 15 , the first three and the fourth of which conbain the curves representing the " " intensity of magnetisation\ldquo ; and the " " magnetic susceptibility\ldquo ; respectively of the wires , and the last contaillS the curves representing the " " intensity of magnetisation ' ' of the bars .
In other words , the curves in the Plates 11 , 12 , 13 , and 15 are so drawn that the abscissae are proportional to the magnetising force , and the ordinates to the intensity I ; whereas in the curves in the Plate 14 the abscissae and the ordinates are proportional to the force and the susceptibility respectively .
As regards , first of all , the Plates 11 , 12 , 13 .
The curves in the Diagram I correspond to the " " Dark Wire those in the Diagram II to the " " Bright Wire and those in the Diagram III to the " " Steel Pianoforte magnetising force was acting ; the curve ( c ) is one showing the effeot of suddenly reversing current in the coil ; the curves ( d ) and are those showing the effect of " " Ons and Offs\ldquo ; while ths reversed current was circulating through the coil , the former corresponding to the case " " On ' ' and the latter to " " Off\ldquo ; ; while the curve ( f ) is one so drawn that the ordinate at every point of it is half the algebraical difference of the ordinates of the curves ( b ) and , and benoe exhibits approximately a which should have been obtained had the wire been experimented on without being subjected to the action of a .
Had it not been for the sake of convenience of comparison , therefore , the curves , ( d ) , and should have been drawn on the negative side of the origin .
Exactly the same explanation applies to the curves in the Diagram II as to the corresponding curves in the Diagram I. In the Diagram III , the curves and ( b ) show the results for " " Steel Pianoforte Wire and are subject to the same explanation as the corresponding curves ( a ) and ( b ) in the Diagram I or ; while the curves ( c ) and ( d ) refer to the " " Glass-hard-tempered Wire\ldquo ; the former representing the resuIt obtained when the magnetising force was in action , and the latter that obtaiued immediately after it was withdrawn .
Glancing at the curves in the Diagram I , we see something very striking .
In the first place , we cannot help being struck with the remarkable eifect of " " Ons and Offs\ldquo ; on the magnetisation of the dark wire , when we compare the curve ( a ) or ( b ) with the curve ( f ) .
But a still more remarkable resuIt is revealed in the fact that there is a surprising difference , as the curves ( a ) and ( b ) show , between the intensity of magnetisation of this in the case of ' ' On and that in the case of ' ' Off\ldquo ; for low magnetising forces ; and that the difference gets less and less remarkable as the magnetising force is more and more increased , becoming nothing at 15 units of the force , then changing into a negative quantity for still higher magnetising forces , and uItimately attaining a constant negative value .
In other words , tho intensity of magnetisation of the wire is greater or less ] it is actually under the action of a constant pull than while it is free from it , according as the magnetising force to which the wire is subjected is below or above a certain value\mdash ; a value which might , therefore , be called critical .
* The fact that the two pairs of curves ( d ) and ( e ) *This confirms the result given on page 62 of Sir Willi Thomson 's paper on the " " Electrodynamic Quahties of Metals , Part VII\ldquo ; Phil. Trans 1879 ) , in which he calls this value Yillari Critica , l Value as having been previously obtained by Villari .
Determinations of Jfagnetic and and ( b ) are symmetrically placed with respect to the horizontal axis , each to each , shows that the uItimate effect of " " Ons and Offs is to netise the wire to the same degree of intensity , under the same circumstances , whether the magnetising force be in one or in the opposite direction .
On the other hand , the curve ( c ) shows that when the magnetising force is so high as 60 units or so the wire seems to lose its retentiveness , so much so , that the reversal of the polarities of the by the reversal of the force is so complete that the operation of " " Ons and Offs\ldquo ; produced no permanent effect ; but that when the magnetising force is below that value the simple reversal of the force is not so effective as to annul the permanent effects of " " Ons and Offs or even to reverse polarities of the wire .
It is obvious that the excess of the intensity of magnetisation represented by the curve ( b ) over that represented by the curve ( f ) , corz.esponding to any magnetising force , is a measnre of the retentiveness of the wire for that magnetising force .
Remarks so very similar to thoss msde on the curves in the Diagram I apply to the corresponding curves in the Diagram II that is quite unnecessary to mention them .
The comparison of the two sets of curves in the two diagrams , however , presents many points of .
The curves ( a ) and ( b ) in these diagrams show that for some low magnetising forces the intensity of magnetisation of the " " Bright Wire\ldquo ; is than that of the " " Dark Wire ; \ldquo ; this is , perhaps , not because the .
is more susoeptible of magnetisation than the latter , but chiefiy because of the fact that there is for each wire a certain amount of pull ( used for " " Ons and Offs which would give a maximum effect on the magnetisation of the wire , and that a weight of 12 kilogs .
is nearer that value for the bright wire than of 8 kilogs .
is for the dark wire .
As regards the critical point , we see that it is about 15 units in the case of the dark wire , while it is about 10 units in the case of the bright wire ; bnt this point is no doubt , not only for different kinds of wire but also for different of the pull .
But it is in the curve ( c ) that the chief interest lies .
The comparison of the curves ( c ) and ( e ) in the two diagrams shows that the eflec of reversing the magnetising force on the change or reversal of magnetisation is considerably less in the case of the bright wire than in the case of the dark wire , both which must doubtless be accounted for by supposing that the one ( tolerably soft iron ) has a greater coercive force than the other ( exceedingly soft iron ) , as might be expected .
The comparison of the curves in the III with those in the I or is also interesting .
The most striking point is that , unlike the case of soft iron wires , there is no such thin , .
as critical point in the case of steel wire , as the curves ( a ) and ( b ) in the Diagram III point out ; for every magnetising force the intensity of iron wire ; but seeing that when the magnetising force is so high units or so ( when the permanent effect of " " Ons and Offs\ldquo ; beginS to be insignificant , that is , when retentiveness gets inconsiderable ) , the intensity of magnetisation of the steel wire is very much the same as that of the soft iron wires , I think it probable that the above diiference is , in a great measure , due to the fact that a weight of 16 kilogs .
( less than one-sixth of the breaking weight of the pianoforte wire ) used for the operation of " " Ons and Offs\ldquo ; is far too small to produce anything like full eflect on the magnetisation of the steel wire , and that this difference can be greatly diminished by using a heavier weight ( perhaps 40 or 50 kilogs .
) to operate " " Ons and Offs The difference that exists between the intensity of netisation oi the steel pianoforte wire and that of the glass-hard-tempered steel wiro , corresponding to low magnetising forces , is greatly due to a similar cause ; but observing that there subsists a considerable difference in the intensity of magnetisation of these two wires even for so high a magnetising force as 50 or .
60 units , it seems probable that the intensity of magnetisation of the glass-hard-tempered steel wire is really smaller for every magnetising force than that of the iron .
tempered steel wire , even when the eifect of stress is taken into account .
As regards the limit of the magnetisation of these wires , on compal.ing the curves ( a ) and ( b ) in these diagrams , it will be seen that that limit is attained at so low magnetising force as 80 units or so , both in the case of the soft iron wires and the non-tempered steel wire , and that the maximum magnetisation of the pianoforte wire is not lower than that of the soft iron wires in the ordinary cases\mdash ; results certainly unexpected .
On the other hand , the comparison of the curves ( b ) and ( c ) in the Diagram III requires a careful study .
It shows that at about 80 or even 100 units of the magnetising force there is a notable dilference between the netisation of the nontempered and glass-hard-tempered steel wires ; but whether this difference is due to the fact that the maximum magnebisation of the latter is not yet reached at the above-stated magnetising force , or it represents the actual difference in the maximum netisation of the two wires , it is difficult to deoide .
In whichever way this difference is accounted for , it is not unfair to say that the maximum magnetisation of the glass-hard-tempered steel wire is very nearly , if indeed not exactly , equal to that of the steel pianoforte wire or the sofb iron wires , and that the minimum magnetising force corresponding to the maximum magnetisation is somewhat higher in the case of the former 1 .
The dark wire correspouding to .
The bright wire . . . . . . . . . .
, , .
The steel pianoforte . .
, , eredwi .
, , ' Off The curve ( a ) in the Diagram shows that the maximum residual magnetism of the tempered steel wire is considerably greater than three-fourths of the total netism of which it is a residue ; whereas in the case of the soft iron wires the ximum residual magnetism is only a small fraction of the total magnetism .
Passing now on to the curves in the Plate 14 , no more words are perhaps necessary to explain them , because the explanations given of ths curves in the Plates 11 , 12 , 13 will exactly apply to the corresponding curves in the Plate 14 , if we substitute the words " " Magnetic Susceptibility\ldquo ; for " " Intensity of Magnetisation By the corresponding curves is meant the curves which are marked by the same letters , such as , ( b ) , \amp ; c. , in the diagrams designated by the same numbers , such as I , II , \amp ; c. With regard to results for the netic susceptibility , it may be remarked that the results of the preliminary experiments not glven the paper , showed that the susceptibility of any one of the wires is diflerent according to dilferent circumstances under which it is placed , that is to say , that here is , for each magnetising force , an infinite number of values for the susceptibility corresponding to all infinite number of amounts of pull to the applications and removals of which the wire might have been subjected ( though this appears to cease to the case when the magnetising force exceeds acertain value , that when the wire begins to lose its retentiveness ) , not to speak at al.l of the different values for the susceptibility the wire at any glven stage of its history , according to the different amounts of a permanent pull to which the wire may be subjected .
Hence it is evident that we should have precise knowledge of the history , past and present , of the body whose susceptibility we wish to determine ; and this is the very reason why the experiments were made on the wires under definite circumstances .
The two sets of the values for the susceptibility of each wire , one for the case " " and the other for " " Off given in the corresponding table and represented by the curves , are , therefore , those corresponding to that particular circumstance under which the wire was experimented on .
The gnetic susceptibility of ; Mr. R. SLida .
the soft iron wires when retentiveness is disregarded , can be calou .
lated , if required , from the magnetisation represented curves ( f ) , Plates 11 , 12 , 13 .
The greatest value for the magnetic susceptibility I obtained of soft iron wire is about 730 , the corresponding magnetising force being the Glasgow vertical force , and it is probably still greater for smaller magnetising forces ; while the magnetic susceptibility of the same wire for so high a magnetising force as 100 units , is only about 13 , and still smaller , no doubt , for higher magnetising forces .
Thess results are truly surprising , and will dispel any doubt as to the old view that the value of is constant or nearly so for all or a certain range of magnetising force .
I will now proceed to explain the curves in the Plate 16 which represent the results for the bars .
The " " direct curves\ldquo ; show the results obtained by commencing with a small magnetising force which was gradually increased until it is so high as to netise the bars very strongly , if not to saburation ; while the " " return curves\ldquo ; represent the results obtained by coming down from a high magnetising force to lower and lower magnetising forces , passing through the zero and going up gradually to a high magnebising force on the negative side of the zero .
It may be mentioned that the reason why for the steel bar the direct curve was not obtained is because the bar , which was one of those originally intended to be used for Sir William Thomson 's new Siphon Recorder , was previously magnetised strongly , and , therefore , the experiment on it was commenced by using a high magnetising force to start with ; and that there is evory reason to believe that the direcb curve for the steel bar is something like that for the cast-iron bar .
On the " " direct curves\ldquo ; in the Plate 15 , we see that the magnetisation of the bar is somewhat less for high magnetising forces than that of the steel bar , and is much less for every gnetising force than that of the soft iron bar ; and that maximum magnetisation of the soft iron bar is about 1340 , that of the steel bar is about 860 , and that of the cast-iron bar is only about 770 , while the corresponding least magnetising force in the case of the first is only about 190 units , and in the case of the second and third , it is roughly 450 and 400 units or more .
Of course , it is not quite right to assume that the above results represent accurate comparisons of the magnetisable qualities of those different kinds of iron and steel , because the bars are not the same in dimensions , which have very considerable effects on the intensity of magnetisabion .
Still considering that the difference in dimensions veen the soft iron bar and the other bars is very small , while in both the maximum intensity of magnetisation and the minimum magnetising force corresponding to it they differ greatly from each obher , it is certain Determinations of Magnetic Susceptibility .
443 that both the cast-iron bar and the steel bar are greatly inferior to the soft iron bar in respect to magnetisability .
This is indeed unexpected , and in some measure astonishing , rememberiug that the steel pianofor forte wire was not at all in nferior in this respect to the soft iron wires , at least for higher magnetising forces .
The difference that is found in the maximum intensity of magnetisation and the mlnlmum magnetising force corresponding to that between the soft iron bar and the wires is , however , no doubt , chiefly due to the effects of the dimensions of the bar .
Another point of interest lies in the " " return curves .
They show that in the case of each bar the magnetisation of the bar did not reverse until the magnetising force exceeded a value on the negative side of the zero , and that this value is conslderable even the case of the soft iron bar , considerably greater cas the case of the cast iron , and still greater\mdash ; greater by a vast amount\mdash ; in the case of the steel bar .
A curve for the residual magnetism was only obtained , or at least only shown , for the cast iron ; but the fact that those points in the curves corresponding to the zero magnetising force represent the maximum residual magnetism of the corl.esponding bars , will us a rough indication of what might be the residual magnetism curves for the other bars .
I have now given the general explanations and discussions of all the results of the experiments , and as 1 fear space does not permit me to enter into a fuller discussion of all the details of the results and of the inferences that can be drawn from them , I am obliged to them untouched .
There is , however , one very estlng and important conclusion which .
can be derived from the results and which I cannot help noting specially , as it illustrates the beauty of this magnetometric method , and that is , in regard to the change in the distribution of magnetism of the wires or bars due to the corresponding change in the magnetising orce ey force to which they are subjected .
It has already been said that one way to sbudy the results gven the Tables I , II , \amp ; c. , is to trace curves in the manner explained .
Now , 1 is easy to get two such curves as ( 1 ) and ( 2 ) of the Plate 10 for each set of the resuIts , one representing the effect due to both the magnet sm of the wire or bar , and the coil carrying a current , and the representing the same due to the coil alone .
If we draw another curve such that its abscissa at every point of it is the difference th abscissae at the same point of the curves , we obtain a curve representing the effect dne to the magnetism of the wire or bar alone .
The curve representing the effect the coil aIone can be easily *Compare these curves with those given in Sir William Ihomson 's paper referred to before , " " Phil. Trans 1879 , Plates 8 and 9 Mr. R Shida .
obtained , if necessary , from the value of , because evidently the curve represented by the equation , ( 9 ) , in which , \amp ; c. , retain the same meaning as before , will be the one required , namely , one in which the ordinates are proportional to the vertical distances of the magnetometer needle from the centre of the coil , and the abscissae to the deflections of the needle due to the coil .
A theoretical curve representing the effect due to the magnetism of the wire or bar solenoidally distributed , that is to say , with a certain quantity of free magnetic matter of northern polarity at one extremity and the same quantity of free magnetic matter of southern polarity at the other extremity of the wire or bar , can be obtained in a similar way ; iu fact , the equation ( 9 ) will such a curve , if we substitute the quantity of the free magnetic matter at either end of the wire or bar for and half the length of the wire or bar for Now the curves ( 1 ) , ( 2 ) , ( 3 ) , and ( 4 ) , in the Plate 16 , were obtained in the way just explained from the results given in 1 , 2 , 4 , and 7 , and the Tables I and ( that is , the results for the " " Dark Wire ; they represent .
the curves showing the effects due to the magnetism of the wire alone , and correspond respectively to 545 ( in vertical force ) , , and units of the magnetising force , while ( 5 ) is a theoretical curve repre , senting the effect which should have been obtained had the same wire been magnetised solenoidally , so as to contain 8 units of the quantity of free magnetic matter of one polarity at one end of it , and the same quantity of matter of opposite polarity at the other end .
These curves form the true comparisons of the magnetisations of the wire in the different cases , because they are all reduced to the same standard , that is to say , they are all so drawn that their abscissae represent the deflections of the magnetometer needle which should have been obtained had the field been one and the same , namely , units in all cases .
The comparisons of the curves ( 1 ) , ( 2 ) , ( 3 ) , and ( 4 ) show that the greater the magnetising force the greater is the distance from the centre or origin of the points of the ordinates corresponding to the maximum deflections of the magnetometer needle , while the parison of the curves ( 4 ) and ( 5 ) shows that these points in the case of the curve ( 4 ) are almost , if not exactly , coinciding with those in the case of ( 5 ) ; showing quite distinctly that the magnetisation of the wire for a low magnetising force is far from being solenoidal , but stronger at the central parts of the wire than in the other parts ; but that as this force is made stronger and stronger , the magnetism of the wire becomes more and more equally disbributed to the ends until the dis* Determinations of Magnetic Snsceptibility .
tribution becomes nearly , if not altogether , solenoidal , when the force is made so high as give the wire the maximum magnetisation .
More or less similar facts can be arrived at from the results for other wires , and also those for the bars .
These facts are truly interesting , seeing that they entirely agree with theoretical considerations .
Indeed they have been pointed out theoretically by Sir William Thomson , * and indicated by Rowland .
But I believe my experiments are the hrst , the results of which have brought out those facts so clearly as not only to leave no room for doubt , but also to enable us to see the law by which the change in the distributlono mag ma netism in a cylindrical rod due to the chang.e of magnetising force to which it is subjecte , govern is overned ; and I hope they will be of service in guiding the future investigatol .
R of electro-magnetism or otherwise .
It is impossible for me to conclude this paper without my most grateful thanks to Sir William Thomson for the very kind guidance and instruction he has given me in the course of experiments .
*Papers on " " and netism , \ldquo ; S Ma , 1873 , 4 t6 Mr. R. Shida .
; Determinations of Magnetic usceptibility .
I I I $ : : : : Mr. R. Shida .
: \mdash ; \mdash ; : : : ; Determinations of Magnetic ptibility .
: .
' Mr Determinations of agnetic Snsceptibility .
VOL. XXCI .
Mr. R. Shida .
Determinations of Magnetic Susceptibility .
fvoc .
Soc of West ) Soc of , xethz I AGRA II .
) No \amp ; .
Soc. .
f3 .
DIA GRA IIJi .
West ' .
: / ' / - .
/ ' / / I I- 1

