]\gt ; 486 ] XVIII .
Bxperiments on the Directive Power of large Steel Magnets , of Bars of magnetized Soft Iron , and of Galvanic Coik , in their Action on external small llfagnets .
GEORGB BIDDELL AIRY , Astronomer Royal , C.B. , P.R.S.\mdash ; With Appendix , containing an Investigation of the Attraction of a Galvanic Coil on a small Magnetic Mass .
JAMES SrUART , Esq. , M.A. , Fellow of Trinity College , Cambridge .
Received January 6 , \mdash ; Read February 8 , 1872 .
THE only experiments with which I am acquainted tending to throw light up the distribution of magnetic power in the different parts of a steel magnet are some very imperfect ones by COULOMB in the French Memoirs for 1789 and other years .
It appeared to me that it might be desirable to make experiments of a rather more extensive character , and to add some measures of the magnetic effect of galvanic currents , both directly by their immediate action , and indirectly by the amount of magnetic power which they produce inductively in soft iron .
For the measure of permanent magnetism I selected a bar magnet 14 inches long , inch broad , inch thick ; it has not been touched by a magnet for several years , and is likely to be in a state of very permanent magnetism .
For the galvanic currents a cylindrical coil was used inches long , inch in external diameter , and about inch in internal diameter ; it has , I believe , four layers of wire , each layer having 160 revolutions of the wire .
The battery used with it consisted of three cells , with sulphuric acid diluted in the proportion of 1 to 6 ; the plates were of zinc and graphite , each exposing on each side about 8 square inches ; the circuit was always completed about half an hour before the experiments were begun , and a delicate galvanometerwas placed in circuit by which the steadiness of the current was established .
A core of iron inch in diameter and of the same length as the coil , removable at pleasure , fits well in the inside of the coil ; the iron is quite soft , and can with ease be entirely freed from any subpermanent magnetism .
The first step in the experiment was to neutralize terrestrial magnetism within the area of magnetic experiment .
For this purpose two powerful 2-feet magnets were placed below the table on which the experiments were , with their red or north- seeking ends directed to the magnetic north , at a distance ( determined by trial ) such that the experimental compass was sensibly uninfluenced by terrestrial magnetism .
It is possible that some small residual magnetism was perceptible in the comparison with the feeble galvanic action , but none could be certainly discovered in the other experiments .
The compass used for register of the magnetic action is a small and verylivelypocketcompass , with needle inch long , not loaded with a card .
The box of this compass one with radius at inch , at whose extremity and.bisection points were taken for the compass-centre ; and one with radius inches , which was twice bisected , and ofwhich the extreme point and the three bisection-points were taken for the compass-centre .
These points were used for the magnet both with its and with its flat side towards the compass .
A similar process was adopted in using the galvanic coil , with this difference only , that the longitudinal separation of the points taken for compass-centre only inch .
A solid piece of wood was provided , in which was cut a concave channel , less than half a cylinder , such that when the galvanic coil , or the large magnet with its flat side towards the compass , was laid in the channel , its axis was sensibly at the same height as the needle of the small compass .
With the magnet 's edge towards the compass , that condition was sufficiently secured by merely laying its flat side upon the board .
The paper with station-points , being laid in proper position upon the board and secured by nails , was cut along the middle of the channel and crosswise at its ends , so that it could be bent down into the channel to permit the magnet or coil to take its proper position : when observations were finished , the paper was detached from the board , and the edges which had been cut were re-united by cementing a piece of paper behind .
The observation ( as will be seen ) consisted , in every case , of observation of the direction taken by the small needle .
And this observation was made solely by the eye .
The observer , looking endways of the small needle , made two pencil dots upon the paper , corresponding to the line of the needle-axis produced as it appeared to his eye .
If , from erroneous position of the eye , a parallactic error is produced in the position of the two pencil dots , this error is detected as soon as the compass has been removed and an attempt has been made to draw a line of direction ough the station-point of the compass ; and , to correct it , all that is required is , to draw through the station-point a line parallel to the line joining the two dots .
The whole of this operation is extremely accurate .
For measuring the intensity of the magnetic force exerted ou the compass-needle , I determined , after consideration , to adopt the statical method ; that is , to place a constant magnet in a definite position above the compass-needle , with its magnetic axis transversal to the direction which the compass-needle had take , before the constant magne.t THE DIRECTIVE POWEII OF introduced , and to observe the deflection produced .
The measure of the force of the large magnet was then the cotangent of the angle of deviation .
The observation of the deflected needle by dots \amp ; c. was the same as before ; but the angle of deviation was never measured by degrees .
Instead of that measure , a circle upon semitransparent paper was graduated by cotangents , and thus the measure of the force of the large magnet was read off at once .
The arrangements in this state were confided to .
CARPENTER , Assistant of the Royal Observatory , by whom all the subsequent arrangements were planned and all the observations were made .
I need not say that they were made with the utmost skill and delicacy .
A small frame was constructed , carrying a floor at a definite position about inch above the compass-needle .
As it was my object to make the observations at small distances from the great magnet , where its power is great , it was necessary to use a large power in the deflecting magnet .
Mr. CARPENTER selected a horse-shoe magnet about 4 inches long , consisting of sixteen plates , each inch thick ; these were retouched a few days before they were used .
From the consistency of the results obtained at the beginning and end of each circuit of the great magnet , I am entitled to conclude that no sensible change took place in the netism of the horse-shoe magnet .
The magnet was placed in a vertical position , its two poles resting on the raised floor above mentioned .
In all cases the deflecting magnet was used in the two positions to produce deflection right and deflection left .
These arrangements sufficed for observation of the powers of the great magnet in both positions , and also for observation of the galvanic coil carrying the soft iron core , the intensity of the battery having been in some measure adjusted to make the power of the coil with core not very different from that of the magnet .
But when the coil was used without core , the force was so enormously reduced that the arrangement which applied well in the other cases failed totally in this .
A small magnet was then used , inch long , not very strongly magnetized ; its deflecting power was compared with that of the horse-shoe magnet in the following small compass being under the influence of the earth 's magnetism , the horse-shoe magnet and the small magnet were successively placed on the raised floor above mentioned , then inch higher , then inch higher , and the cotangents of deflection were compared .
Thus the following proportions were obtained:\mdash ; ets upon the raised floor Magnets inch above the raised floor Magnets inch above the raised floor With so great inequality the results are necessarily irregular .
I use as the profigure .
The results , however , for magnitude of force were obtained in numbers .
The means of these were taken in an analogous order\mdash ; first taking the sums of those on opposite sides of the magnet or coil , then taking the sums of the last-found sums for opposite ends .
The division by 4 was omitted ; and thus the numbels in the Table below give the value of 400cotangent of deviation .
At two stations the proximity of the als made it difficult to obtain actual observations ; but there was no difficulty in supplying them conjecturally , with great confidence in their accuracy .
The diagram below was drawn carefully to the positions taken by the small magnet when the edge of the large magnet is presented to the small magnet .
The same diagram will serve , almost without perceptible error , for the case when the flat side of the large magnet is presented to the smaU magnet , or for the case of a galvanic coil enclosing an iron core ; but it will not apply to the case of a galvanic coil without an iron core : for that case the axis of the small magnet in the positions numbered 35 , 5 , ON THE POWER OF MAGNETS .
489 12 , 46 , and all to the right of it must be directed almost exactly to the centre of the right-hand end of the magnet , and a similar direction must be made at 65 , 27 , 20 , 54 in respect of the left end , with corresponding changes for intermediate stations .
The angles of position were never measured , but they are fully taken into account in the subsequent resolution of forces into longitudinal and transversal components .
The following Table exhibits the orce at each station , as ascertained by the operations described above .
It will be rernembered that the numbers for the " " Galvanic coil with iron core\ldquo ; are not necessarily on the same scale as those for the " " Large magnet and that the numbers for the " " Galvanic coil without core\ldquo ; must be divided by 120 to make them comparable with those for the " " Galvanic coil with iron core Total force acting on the small magnet at each station .
Perhaps the following points are worthy of notice:\mdash ; 1 .
Remarking that , in the experiment in which the edge of the large magnet is presented to the small magnet , the distance of the small magnet is in each circuit the same at every station , it will be seen that the greatest directive force is not longitudinal at the end of the magnet , but transversal , at about .
part ( or probably less ) of the length from the end of the magnet .
There is , however , a diminution and then an increase in proceeding from either of these positions to the other , and the directive force opposite the middle of the magnet 's length is less than either of them ; so that , in making the entire circuit of the magnet , there are six maxima and six minima .
2 .
When the flat side is presented to the small magnets , the same statement holds for the outer circuit , but not for the inner circuit .
3 .
With increase of distance , the diminution of force is much more rapid at the end than at the side of the large magnet .
with core included gives and 6 .
The law of magnitude of forces for the coil alone differs greatly that of a steel magnet .
In the inner circuit the proportion of the force at the end to force at the middle of length is , for the steel magnet , for the coil ; in the outer circuit they are and 7 .
Still more remarkable is the difference in the law of direction of the forces near the ends .
Using the term " " pole\ldquo ; to denote that point near the extremity to which the directions of forces rudely converge , the pole of the steel magnet is within the magnet , and distant from the end by about of the magnet 's length : but the pole of the galvanic coil is absolutely at its end ; indeed some of the experimental directions of force fall a little beyond the end .
It is evident , from the remarks of Nos. 6 and 7 , that a magnet cannot in any wise be represented as a system of revolving galvanic currents , with an equal number of circuits at every part of its length .
With the view of presenting the results in the form which may probably be found most advantageous for comparison with the conclusions from any future theory , I have resolved the forces into rectangular directions , parallel and transversal to the axis of each magnet , by the .
graphical process .
Upon each mean line of direction of force ( ascertained as is described above ) I have laid down the mean measure of the force ( as found above ) , and upon this measure as hypotenuse I have constructed a right-angled triangle , the lengths of whose sides give the two forces .
From the nature of the pre- ceding operations , it is only necessary to form these numbers for one quadrant each magnet .
The results are given in the following Tables .
The centre of the large magnet or coil is in every case the origin of coordinates of the external magnetic poin on which the action of the large magnet \amp ; c. is estimated\mdash ; the axis of the longitudinal ordinate being the axis of the magnet , and the axis of the transversal ordinate being normal to it .
The powers are estimated as those of the red end of the large magnet operating on a small external mass of red magnetism .
It will be remembered that , for the galvanic coil without core , all the numbers must be divided by 120 .
ON TIIE DIRECTIVE POWER OF MAGNETS .
Large Bar Magnet .
Galvanic Coil .
It does not appear possible to infer from these numbers , by any direct analytical process , the law of distribution of magnetism in the bar .
It must be done , I believe , synthetically , by assuming a law , and computing the forces which would result that law , and then comparing these computed forces with the forces actually observed .
The only law which I have tried is the supposition that the intensity of netism is NDCCCLXXII .
3 X and the results of integration Longitudinal force force I have computed these numbers for each of the eighteen stations .
For comparison with observation , I have taken the experiments with the flat side towards the small magnets , which represents most nearly the case of a linear large magnet ; and , for facility of comparison , I have multiplied the experimental numbers by 6 .
The following is the comparison:\mdash ; The agreement is not satisfactory ; but I am unable to suggest the nature of the change that ought to be made in the assumed law .
I shall add only one remark , of a somewhat practical character .
In a paper published original by Dr. LAMONT in Annalen , ' vol. cxiii . .
239\amp ; c. , and ofwhich ON THE DIRECTIVE POWER OF MAGNETS .
493 a translation , by W. T. LYNN , Esq. , of the Royal Observatory , is printed in the Philosophical Magazine , 1861 , November , Dr. LAMONT inferred the proportion of the effects of different steel magnets from the proportion of the effects of different soft iron bars under the influence of induction .
The remark No. 4 ( above ) goes far , I think , to justify this assumption .
APPENDIX .
Remarking the singularity of the experimental result as to the apparent localization of the attractive pole of a galvanic coil at the very extremity of the coil , I commenced an investigation of the theoretical attraction of a coil , on the laws of galvanic attraction usually received .
On my mentioning the subject to my friend JAMES STUAIT , Esq. , Fellow of Trinity College , Cambridge , he kindly undertook , at my request , to prepare a complete theoretical investigation .
I am happy in being permitted by Mr. STUART to place before the Royal Society his matheinatical discussion of the attraction of the coil , which I am confident will be found to be very complete and of great elegance .
I append to it a comparison of the numerical results of the theory with the numerical results of experiment ; and the agreement will be found to be so great as to justify entire confidence in the assumed law of galvanic action and the mathematical treatment of it , and a high estimate of the accuracy of the experimental observations .
Investigation of the Attraction of a Galvanic Coil on a small Magnetic Mass .
By STUART , M.A. , Fellow of Trimty College , Cambridge .
Received July 26 , \mdash ; Read December 5 , 1872* .
From investigations given by , we can deduce an expression for the potential at an external point of a closed circular galvanic current carried by a wire of indefinitely small section .
Let be the radius of the circle , let the distance of from the centre of the circle , be , and let the line make an angle with the normal to the plane of the circle : then it can be shown that when is less than and when is greater than * Abbreviated from the Appendix originally presented and read with the paper .
To calculate these quantities , we know that , .
We shall only consider the case of those points for which is greater than .
Substituting these values in the expression which in such instances holds for , we have cos which , after some reduction , we obtain . . . .
, . . . . . . . . .
( 1 ) cos - .
( 2 ) Each of these expressions consists of a series of terms in ascending powers of , which will be converging .
ON THE DIRECTIVE POWER OF MAGNETS .
495 We shall now seek to find X and for a galvanic current traversing a wire coiled into the form of a hollow cylinder , of which the internal radius is , the external radius , and the length is .
We shall suppose the individual turns of the wire to lie so close as that each may be as an exact circle .
Let A be the axis of the coil , so that A and are the centres of its two faces ; then .
Let be the middle point of A B. Let be the attracted point , its perpendicular distance from A B. Let A Let be the centre of any turn of the wire regarded as a circle of radius ; then it is readily seen that for the whole cylindrical bobbin the forces X , are given by where and stand for the expressions on the right-hand side of ( 1 ) and ( 2 ) respectively , and where depends on the strength of the current .
To perform the integrations for the length of the bobbin in these expressions , we have the formulae ; and Making these substitutions for .
and , the integrals with respect to become rals respect to , which can be easily evaluated by a continued application of the method of integration by parts , the being from to .
If we then integrate the These expressions for X and will be converging for all points situated at a greater distance than from any point of the axis AB , inasmuch as they are composed by adding together corresponding terms of series which are then all convergent .
Among other points , these expressions hold for such as are situated on the axis external to the bobbin , and not nearer A or than by the distance .
For such points , however , the expressions become illusory , assuming the form ; they may , however , be evaluated by the methods for the evaluation of vanishing fractions .
is clearly zero .
X may be more readily obtained directly from the expression for ; from that expression we find that for a single circular current the attraction on such points is Hence , in the case of a bobbin , if be the distance of the attracted point from , the middle point of the axis of the bobbin , we have which gives X for points situated on the axis for which is not less than . .
ON THB DIRECTIVE POWER OF MAGNETS .
The expressions for forces which concern us now are those given by the general formulae for and .
And a moment 's glance at these will show that they explain the apparent position of the pole at the very extremity of the coil : for , in order to ascertain the values of the forces in a plane at right angles to the axis passing through the extremity of the coil , we must make ; and if the other end of the coil be very distant , may be taken .
Substituting these values , it will be seen at once that X , the longitudinal force , , while , the transversal force , has a value , which indicates a force directed to the extremity of the coil .
In order to make a complete comparison , I have , for all the eighteen stations treated in the former Tables , taken the values of , and graphically .
For I have adopted , and for ; these numbers correspond to the internal and external surfaces of the coil , but they appear to me best to represent ( though doubtless with some inaccuracy ) the quantities used in the theoretical investigation .
Then I have ( with the kind assistance of EDWIN DUNKIN , Esq. , of the Royal Observatory ) made the complete calculation of the formulae for every station .
As the numbers first obtained were not immediately comparable , I have made them more nearly so by trebling the numbers given by theory and doubling those in the preceding .
Table .
The results are as follows : \mdash ; In spite of some discordances in the large forces ( which it was impossible to measure with accuracy ) , there is enough of agreement to show that confidence may be placed in the method of theoretically computing the attraction of the galvanic coil .

