]\gt ; Magnetic Induction and its Reversat in Sphericat Iron Shells .
* By J. W. NICHOLSON , M.A. , D.Sc .
, Professor of Mathematics in the University of London ; ERNEST WILSON , M.Inst .
C.E. , M.Inst .
E.E. , Siemens Professor of Electrical , in the University of London , King 's College .
( Communicated by Prof. J. A. Fleming , F.R.S. Received November 30 , 1916 .
) In a previous communication to the Society , an account was given of am investigation of the magnetic of a large space , by the use of a series of spherical iron shells .
The basis of this investigation , which is still in progress , is strictly quantitative , and the work has been retarded , to a great extent , by the fact that several processes adopted commonly by various investigators of magnetic phenomena have never been examined accurately on the theoretical side , and in several cases even on the experimental side , to an order of precision which is sufficient for delicate measurements .
One of the main objects ultimately in view , in the work described already , is the production , hout the volume of a large iron specimen , of a magnetic field so small , in comparison with that of the earth , for example , as to render possible some accurate conclusions regarding the behaviour of iron under indefinitely small magnetic forces .
This behaviour is at present a matter of doubt , and it remains to be shown that he indications of its real nature , obtained in earlier papers by one of us , continue to point in the same direction when the force is decreased below any values which have been employed hitherto .
In the experiments described in the former paper , field so low as C.G.S. unit was obtained in a spherical space of radius cm .
by the use of four concentric shells .
It was accompanied , however , by a leakage field of C.G.S. unit , and the shells still showed some traces of permanent polarisation after continued treatment with many reversals of a slowly decreasing current .
This method of demagnetisation is used extensively , and in its application coils have sometimes been wound somewhat loosely on the iron to be demagnetised , and the constant field given by Maxwell 's formula for a spherical internal space surrounded by a very closely wound coil assumed to exist .
But if a thin shell is wound in this manner , although the more internal parts may be subject to this field , the results must be widely different in the outer material of the shell .
Further experimental work on * We wish again to acknowledge a grant for the purpose of researches on shielding , which was voted to us by the Council of the Society out of the Gore Fund .
'Roy .
Soc. Proc , vol. 92 , p. 629 ( 1916 ) .
Profs .
J. W. Nicholson and E. Wilson .
Magnetic the improvement of magnetic shielding beyond the order obtained already\mdash ; in fact by such methods as the employment of a very small netising current to remove , in the internal space , the small part of the earth 's field which may remain after the operation of the shielding effect of a set of shells \mdash ; requires a precise knowledge of the value of magnetic induction in a shell , or one member of a set of shells , wound by a spherical helical coil .
the condition of indefinitely close packing is found to be essential to the success of Maxwell 's formula , which\mdash ; but only from this particular point of view\mdash ; does not appear , either from an a priori theoretical consideration , or from the experimental results described in the present paper , to have the large of validity which has been assumed for it hitherto by practical workers on magnetic phenomena .
It is noteworthy that Maxwell himself did not claim any validity for his expression of the force inside a spherical helix when that helix was not in fact in the ideal state of close winding .
The use of such a helix for the purpose of demagnetisation of a hollow shell\mdash ; where importance attaches more to the material of the shell than to the air space within\mdash ; was not adopted at that time , and the practical problem had given no indication of its ultimate importance .
Evidently the spacing of the helix constituting an ordinary cylindrically wound solenoid , if small in comparison with the radius of the cylinder , would not seriously disturb the uniform distribution of magnetic induction in a thin iron bar placed along its axis , but at the outer surface of a thick bar filling the helix the distribution would be widely different .
We may argue from this by analogy to the spherically wound thin iron shells whose thickness is of the same order as the space between consecutive turns of the coil .
Apart from the case of a closely wound coil on a shell of non-magnetisable .
material , Maxwell did not consider the problem of the spherical helix in further detail .
No subsequent writer has taken up the question , and it has been necessary to include in this paper some preliminary theoretical discussion of the problems whose experimental investigation has been required .
The calculations relate to the ideal case of indefinitely close winding\mdash ; no simple mathematical treatment of a more practical case has yet been found\mdash ; and the results are compared with those found for the actual coils which were , described in the earlier paper .
An estimate of the effect of spacing in the helix is thus obtained , and the calculations are of importance in themselves , as indicating some quantitative relations , inherent in problems of this nature , which it would be difficult to obtain by general reasoning .
The distribution of magnetic induction in a set of shells , all of which may possess magnetisimg helices internally and externally , follows certain very curious laws .
Induction and its Reversal Iron Shells .
Into the experimental problem of reversal of the magnetic induction of a shell , two main factors en.ter .
The first the nature of the magnetising coil and the necessary current it must contain .
As stated already , the investigation of this problem is one of the main objects of the present paper .
But in addition , demagnetisation of a mass of iron is much retarded by eddy currents set up in the iron by any change applied to the magnetising forces .
These currents , while flowing , shield the inner shells , in the case of a set , from the effect of the changed field , and if the reversals of current in coils are made too quickly , the process is very inefficient as a demagnetising agency , for , in the time occupied by two reversals , the effect does not penetrate to the inner part of the iron .
The latter part of the paper contains a further contribution to the study of this phenomenon , as it is necessary to obtain some estimate of the time required between reversals , for small .
values of the magnetic induction .
Previous work on eddy current phenomena has been almost completely restricted to cases in which the average magnetic indu ction is of a much higher order of magnitude than in the present experiments .
Before proceeding to a description of the experimental work , it is desirable to give some .
account of the solution of the theoretical problems which are involved , and which have not hitherto received attention .
The analysis required throughout , however , is a fairly simple extension of a well-known method of treating a set of shells without magnetising coils , and for this reason the work is set out somewhat briefly , and , in certain of the cases , only the final results are quoted when there is no novelty in the mode of treatment or when the complete ation is somewhat long , and only interesting from the point of view of its fiual results .
No attem is made to deal with other than the ideal cases in which the magnetising heliees are wound extremely closely .
In this manner , two simple surface conditions become applicable at every coil\mdash ; the continuity of normal magnetic force , and a discontinuity of a prescribed amount in the magnetic potential\mdash ; as shown in the following section .
But the of magnetic induction in the actual set of shells similar , in a general manner , to that calculated for these ideal cases , whose consideration has , in fact , served to direct the experimental work into the most fruitful channels .
The Spherically ing Coil .
The fact that a coil wound closely over a sphere of unit permeabihty , with all its turns perpendicular to the same diameter , produces a uniform internal field is well known .
* For a diameter , with turns each carrying a current * Maxwell 's ' Electricity and Magnetism , ' vol. 2 , para .
676 ( ed. 1873 ) .
VOL xcm.\mdash ; A. Profs .
J. W. Nicholson and E. Wilson .
Magnetic the strength of the field is parallel to the diameter .
Maxwell quotes also the formula for the magnetic potential outside such a region in the same circumstances , as the result of a investigation into the theory of current sheets .
It can , however , be obtained at once by direct integration and , as the need for such a proof appears to be marked , the investigation follows .
If OA is the diameter , and there are turns per unit length of diameter , the number in any section of centre X is , each carrying a current FIG. 1 .
where .
The solid angle subtended at by these currents ia , where is the angle in the figure , and , for a radius of the sphere , the netic potential of the whole system at is in where , if OP and thus rrino : The integral may be evaluated eadily to which may be generalised at once after the usual manner of potential theory to at an external point referred to the centre and diameter .
The external effect is therefore that of a simple magnetic doublet of moment , situated at the centre of the sphere .
The normal magnetic intensity on the surface of the sphere is with , or , and is continuous with that derived from the In.duetion its Reversal in Spherical Iron Shells .
internal potential .
The potential itself , on the other hand , is discontinuous at , for ( inSin)cos a so that the amount of the discontinuity is .
These constitute the two surface conditions mentioned already as applicable at the surface of the coil , and they admit of an instructive comparison.with the corresponding conditions for a spherical shell magnetised normally at every point\mdash ; the discontinuity of potential is in this case , where is the strength of the shell .
We may now consider a spherical helix wound closely on a shell of non-magnetic material , enclosing a concentric shell , of inner and outer radii and , and of permeability .
The radius of the coil is , and it is wound with turns per unit length of diameter , each carrying a current .
With the origin at the centre , and the diameter perpendicular to the windings as initial line of polar co-ordinates , the field in the hollow of the inner shell may be defined by a potential , whereas in the magnetic material of the inner shell the potential is of the form Capital letters denote constants as usual .
Between this shell and the coil , and outside the coil , the potentials are respectively The usual conditions are fulfilled at the two surfaces of the shell of permeability , thus These four equations may be supplemented by the conditions at the coil , where .
These are and Accordingly whence from the solution of the six equations .
These values are , as might be expected , independent of the radius of the magnetising coil , which remains Profs .
J. W. Nicholson and E. Wilson .
gnetic equivalent , in its effect on an inside iron shell , to a uniform magnetic field external to the shell .
The most convenient magnitude for measurement in such a thin shell is the maximum transverse induction ( transverse to the radius vector at when the use of an exploring coil is adopted , as in succeeding experiments .
Now , the magnetic potential in the material is and the transverse magnetic induction is or It is greatest in the plane , and the average value in that plane , denoted subsequently by , is For a thin shell , the variation across the section , or any other section , is so small that may be used for , so that , quoting the values of and , we obtain If , where , in the subsequent experiments , is about to a good approximation .
It is , perhaps , desirable at this point to lay some emphasis on the smallness of the values of .
produced in a single shell by the demagnetising currents adopted in the experiments ibed in the for.mer paper .
For , , and , a current of 1 ampere , or C.G.S. unit\mdash ; which was the order of the magnitude of the current used\mdash ; only produces an average magnetic induction , in a single shell , of amount C.G.S. units , at .
The actual value will be even smaller , on account of the loose winding of the coil , which corresponds to Consider now a magnetising coil , of radius , wound internally to the shell of radii and , perhaps on a shell of non-magnetic material in the interior .
The fields in the various regions are represented very simply .
In the space internal to the coil the field is uniform , whereas between the coil and shell , and in the material of the shell , it can in each case be represented by the superposition of a uniform field and that of a central magnetic doublet of suitable moment .
Outside the shell , being finite at infinity , it is the field of a central magnetic doublet only .
There is no necessity to record the Induction .
and its Reversal in lron Shells .
analysis , which follows the same lines as before .
The final magnetic potential in the material of the shell is found to be where The average value of , obtained as before , is , or , to a close approximation for a thin shell , directly proportional to the cube of the radius of the coil .
If the coil is effectively wound on the inner surface of the thin shell of magnetic material , we may put , and comparison with the previous case indicates an interesting theorem:\mdash ; The maximum transverse magnetic inductions obtained by winding a thin iron shell ( 1 ) internally , and ( 2 ) externally , with the same coil calrying the same current are in the ratio 1 No conclusion , however , can be drawn from this result as legards the relative maxima induced simultaneously in two thin shells of nearly equal radii with a coil wound in the small air space between them .
Their mutual effect is then of great importance and is considered in the next section .
Two Concentric Shells enclosing a Magnetising Coil .
The of the arrangement is somewhat long , but it follows the previous lines so closely that we only present a bare summary .
Let and be the two radii of the inner and outer shells , each of permeabihty , and let be the radius of the coil situated between them , so that are in ascending order of magnitude .
Between and and and , and and , two constants are in each case required to define the magnetic potential when a current is passed through the coil .
When the field is uniform , and beyond it is that of a central magnetic doublet .
Each of these two regions requires one constant , so that , for a complete specification of the magnetic distribution produced by the current , 10 constants must be determined .
The usual conditions at the four surfaces of the shells lead to eight equations , and two other equations are derived from the continuity of normal force , and the discontinuity of potential , , at the ooil .
The equations are as usual linear , and their Anal solution is expressed most conveniently as follows:\mdash ; Profs .
J. W. Nicholson and E. Wilson .
Magnetic If both being positive , the transverse magnetic induction at is , at the inner surface of the inner shell , and at the outer surface of the outer shell , These may be taken as the average values across the sections of maximum , in the case of the shells to which our experiments relate , with an error of not more than about 3 per cent. In the special cases , whence and , whence , one shell is abolished , and the two previous solutions for a single shell are obtained .
This solution is exact when the coil is wound indefinitely closely .
It is of interest to calculate the numerical values in one special case for future comparison , and for this purpose we select the two largest shells used in the experiments , and referred to in the former paper as Nos. 3 and 4 .
Their permeabilities were practically identical , and .
The radii of the inner ( No. 3 ) were and cm .
, and of the outer ( No. 4 ) and Let the coil be wound on the inner so that cm .
, then whence and therefore , when excited by a current in turns per unit diameter of a coil wound on the inner shell , .
in the outer should be in on reduction , and in the inner , in , their ratio being innel. .
The mutual action of the shells has thus changed the ratio from 1 : 2 to 3 : 4 .
The maxima are therefore more nearly equal than might have been anticipated , and , in the earlier experiments , a current used to demagnetise any one shell must have been almost equally effective in the case of the shell on the other side of the air Induction and its Reversal in Spherical on Shells .
space containing the coil .
Moreover , in the absence of the outer shell , the value of .
produced in the inner may be shown to be in , so that its value is increased in the inner shell , by the presence of the outer , in the ratio .
Thsse results may be summarised as follows for this pair of shells : If a coil wound round the inner , in the absence of the outer , produces a value of .
equal to unity , then ( 1 ) if the inner is removed , and the outer placed round the coil , .
in the outer is , and ( 2 ) if both are present simultaneously , .
in the inner is about , and in the outer is .
We can deduce that in the demagnetisation of the entire set of shells in the original experiments , the application of currents of alternate signs simultaneously in all the coils must be at least six times as effective , for the second and third internal shells , as the successive application of these currents\mdash ; a result which could hardly have been foreseen from general principles .
Certain minor features in the curves then obtained can be interpreted on this basis .
These ratios necessarily become larger in the experiments described below , when all four shells are present simultaneously , and the current is passed through the coil on No. 3 .
But we do not include a detailed ] ation of this case , which would be tedious .
A discussion of one or two particular problems can be obtained somewhat simply , and , by comparison with experiment , will enable a determination to be made of the effect to which the theory may be vitiated by loose winding of the coils .
We shall accordingly consider only special problem to which the more interesting experiments relate .
Magnetising applied to the Outcrrnost of a Set of Four Shells .
The most important with reference to the testing of the theory in its ultimate application to the further shielding of the internal hollow of a set of four shells by the use of a magnetising current\mdash ; is perhaps that in which the current is applied to a coil wound round the outer shell , whose value of .
is tested by an exploring coil , as described later .
The necessary calculation for this case is greatly facilitated by the use of the recurrence formulae developed in the earlier paper , * whose notation will be adopted without further description .
The constant field in the hollow being , calculation gives the following successive values of the functions and there described .
The mean radii of the four shells and air space form the sequence , , cm .
, 'Roy .
Soc. Proc , vol. 92 , p. 635 ( 1916 ) .
Profs .
J. W. Nicholson and E. Wikon .
Magnetic and the permeabilities , from inner to outer , are , 163 , 148 , where fairly large error in the first , and a smaller error in the second , are of no importance .
We then find , the recurrence formula without a rougher approximation than 1 per cent. , It is desirable that these intermediate functions should be placed on record for reference in future work on magnetic shielding , which is now in progress with these specimens of iron .
Again , and finally relating to the air space just inside the } shell , on which the coil is wound .
If the potential in the fourth shell is and its inner and outer radii are and , accordingly , giving the ratio of A to , where is its permeability .
Application of the conditions at and at the coil shows , moreover , that so that A and may be obtained .
Thus in the outer shell , with If the current is in amperes , division by 10 is necessary in order to obtain the value in C.G.S. units .
If , or half a turn per centimetre of diameter , were a sufficiently Induction and its in Iron Shells .
close winding to render the formula exact , a current of 1 ampere should give the experimental value C.G.S. units for the outer shell when examined by an exploring coil .
Returning to the recurrence formulae , we notice that This result enables us to correct the previous formula , applicable to a coil wound between the third and fourth shells , for the presence of the other two shells in the interior .
For , in that formula , the function remains unaltered and the only effect of the two inner shells is to alter , which becomes instead of its original value in the former calculation , which was The rest of the formula is as before , and refer to the third shell .
The numerical value of in the present case is found .
to be corresponding to , and thus and , finally , for a coil carrying a current , wound round the third of a set of four shells of the present constants , .
in the outer is given by in , or in , if the current is in amperes .
For , the value per ampere of current is C.G.S. units .
A comparison with a direct experiment is made later .
In the meantime , it is instructive to notice that , in the absence of the two inner shells , .
for the fourth shell when the third is wound was calculated previously as 558 in .
The inner : shells have raised its value in the ratio 13 : much smaller mutual effect is involved , as might be expected , owing to the shielding influence of the third shell .
Experimental Determination of .
by an Exploring Coil .
The four shells remained , throughout the experiments , set up concentrically , each being wound with a magnetising coil , and the axis of all the coils was the diameter coinciding with the direction of the earth 's field .
The value of was in each case , the number of turns wound on the shells being respectively 34 , 38 , 42 , and 46 .
The various dimensions have been specified already , and the shells are formed from high permeability magnet steel .
Since the previous experiments were made , the butting surfaces of all the shells have been machined in such a manner that each of the two halves of any Profs .
J. W. Nicholson and E. Wilson .
gnetic ; original shell penetrates the other .
Preliminary tests , not described in this paper , have indicated that the originally present on account of this lack of machining has almost entirely disappeared , so that the theoretical formulae are strictly applicable if the coils are sufficiently closely wound .
The four concentric shells were placed with their junction planes at right angles to the direction of the earth 's magnetic field , and therefore parallel to the plane of any winding of a coil .
The exploring coil lay .
a plane parallel also to these junction planes , and distant about 7 cm .
from the junction plane of the fourth shell .
It surrounded the iron of the outermost shell , referred to .
as No. 4 , and consisted of one complete turn of insulated copper wire touching the outside of the shell , and one complete turn touching also the inside .
The ends of the inside turn were brought out of the shell through a small hole in the junction plane , by bending the two ends , which are now stranded at right angles to the plane of the coil and then threading them through the hole .
The inside and outside turns were then placed in series so that the coil completely surrounded the iron once .
The section of the effective part of the coil is thus a plane annulus whose radii are practically those of the outermost shell , and the coil is cut , in view of its parallelism to the magnetising coils , by the transverse magnetic induction at in the outer shell , over the whole area of section of that shell .
normal induction across the coil is thus the product of the sectional area , where and are the radii cm .
, of the outer shell , and the average value of .
for the outer shell as calculated already .
More strictly , this product .
should be multiplied by a magnitude of the order cos10o , on account of the small displacement 7 cm .
of the plane of the exploring coil from the diametral plane .
But this would only make a change not exceeding 2 per cent. The exact position of the exploring coil was so chosen that , while iving B with this degree of accuracy on one ground , it should at the same time be so far removed from the junction that the effect of any small air space should be negligible .
For although the two hemispheres constituting the whole shell were dovetailed together , the degree of fit was not sufficient to ensure perfect contact throughout .
A very small air space does in fact still exist , and we may therefore assume that the observed netic induction might on this account be slightly smaller than the calculated value even in the ideal state of packing of the magnetising coils .
The resistance of the exploring coil was quite egligible in comparison with that of the ballistic alvanometer .
This instrument , which was of the d'Arsonval type , had a resistance of 467 ohms at C. ; to this additional resistance was added in order to obtain the sensibility necessary to ensure Induaet and its Reversal in Iron Shells .
141 that the defleotion was not of a nitude capable of serious error on account of permanent set in the suspending fibre .
Let be the resistance of the circuit of galvanometer , exploring coil , and extra resistance when this was used , and the periodic time of the swinging coil of the instrument , in which a steady current of amperes produces a deflection .
Let be the ratio of two successive deflections of the galvanometer coil when on open circuit .
Then if is the number of turns in the exploring coil , A the area of cross-section of the iron in the plane of the coil , in square centimetres , and the mean of the first and second deflections obtained by reversal of the current in \amp ; e magnetising coil , we can the formula : The following values of the constants were obtained as the result of a calibration of the instrument and shell:\mdash ; sec. sq .
cm .
with .
Throughout the experiments , double de{lections were observed on a scale having mm. divisions at a distance of 1340 mm. from the mirror .
The decrement factor was used so long as it did not exceed But the resistance of the exploring coil is so small in comparison with that of the galvanometer that , when no extra resistance is inserted in the circuit , the galvanometer is short-circuited when used for deflections produced by reversal of current in a magnetising coil .
Under these circumstances the decrement is very much greater , and the spot of light , after the first throw , returns to zero and gives a relatively small deflection on the other side .
The ordinary treatment of damping is not.then applicable .
But we notice that failure to correct for a damping factor only introduces an error of about 3 per oent .
, and this occurs in practice in our experiments when a resistance of 10,000 ohms is added .
The interpretation of readings with smaller resistances added can be found by a preliminary expel.iment , which was made as follows:\mdash ; The third magnetising coil , wound between the two outermost shells , was used and a current of amperes passed through it .
Deflections were observed with varied additional resistances in the circuit of the exploring coil .
Reversal of current in the magnetising coil in every case produces the same value of .
in the fourth shell , and therefore the same induced electromotive force in the exploring circuit .
The product of deflection and total resistance should therefore be the same in every case , if the damping factors * See , e.g. , ' Phil. Trams , vol. 180 , p. 446 1889 ) .
Profs .
J. W. Nicholson and E. Wilson .
Magnetic are identical .
This product may be taken as correct with 10,000 ohms added 3 to the resistance of the galvanometer , and thence the true deflection , in the absence of damping , calculated in the other cases .
The interpretation of a reading with any added resistance may then be obtained by interpolation .
This method has the advantage of assuming no theory of the abnormal damping .
The results are shown in the annexed Table:\mdash ; Table I. Variation of Damping with Resistance .
If is multiplied by the value in the last column , we get a practical determination of the decrement factor , independently of any theory , and valid even for the enormous factor which occurs with no added resistance .
Interpolation is not in fact necessary , for all the experiments later are performed with added resistances of , 1000 , or ohms .
In the experiments of the earlier paper , large resistances were always used , so that no discussion of abnormal damping was then necessary .
Thus any first swing with , 1000 , 10,000 ohms added resistance must be multiplied by the factors , or respectively , in order to obtain the undamped reading .
We do not attempt to discuss any theory as to the origin of these values , for it constitutes an independent problem of some difficulty , and is not strictly relevant to the present paper .
Experimental ResuTts .
The values of .
have been determined in the following cases:\mdash ; The four shells were set up as described , and the readings taken in correspondence with various currents in the magnetising coil on the third shell , the exploring coil on the fourth shell .
( b ) Similar readings were taken when the system was excited instead by the magnetising coil on the fourth shell .
The values of .
were obtained in accordance with the discussion in the preceding section , and are shown in the subjoined Tables II and III .
its Reversat Iron Shells .
Table II.\mdash ; Magnetising Coil on the Third Shell .
Table III.\mdash ; Magnetising Coil on Fourth Shell .
Discussion .
The graphs shown in fig. 2 indicate also the results of these experiments .
For small values of the currents , they are very closely represented by straight lines , as would be expected .
The effect of the curve of magnetisation then comes into play , so that , although , on account of loose winding of the coil , the values of ./ current are initially smaller than predicted on theoretical grounds , they ultimately become larger than those calculated on the assumption of constant permeabilities of the shells , and the graphs beoome more parabolic .
The calculations of an earlier section indicate that , if the permeabilities were constant , current should be equal to when the third magnetising coil is used , and 74 when the current is passed through the fourth .
For the smaller values of the current , the values of .
may be expected , iu view of the curve of magnetisation , to follow the empirical law , where is the current , at least roughly , although the constant would have no simple physical meaning on account of the different magnetising forces on the various shells which shield one another .
For currents to , the average value of A is , neglecting the reading ' .
2 .
4 .
6 .
8 account of initial slope of the magnetisation curve .
We may take as the true value of A very closely , when ampere .
The first four readings may be fitted into the formula , with errors in of magnitudes .
The later values require an extra term in the empirical formula , but for the currents of this order , only a very small experimental error not exceeding about 3 per cent. is indicated .
Small magnetic inductions can thus be measured with extreme accuracy by this method of the exploring coil .
The limiting value of the ratio with the present magnetising coil is evidently about instead of , showing that Maxwell 's formula involves a considerable error when applied to a coil even so closely wound as the present one , when the iron of the core is near the coil .
At the same time , the data are ow available for future use of these coils , and the need for a theoretical investigation of a loosely wound coil , which has been prominent for some time , without any solution , is removed .
The Table for the coil on the fourth shell exhibits tInduction aShellsfor nurtherl 4 remarks .
Perhaps one of the most interesting problems opened up by the present investigation is an enquiry into the limit to which demagnetisation of a shell can be carried .
It would be necessary , for certain types of previous magnetio history of the iron , to apply reversals of a continually diminishing force , which began with a value corresponding to maximum permeability , as shown on the ( BH ) curve of the material .
This would be difficult , on account of the necessary magnitude of the current for an arrangement of shells such as the present , which , however , as shown in the earlier paper , that of maximum shielding for a large space .
Since effective demagnetisation , or , at least , the production of uniformity in the internal field of the central hollow , is a necessary preliminary to the removal of that residual field by the application of currents , the problem is of some importance But at this point it is sufficient to indicate the necessity for a determination of what constitutes " " effective netisation Induced Currents .
The reversal of magnetisation within undivided iron magnet cores has been studied to a great extent on its experimental side , and it is known that the induced currents in the iron continue for a considerable time , and tend to maintain.the magnetisation in its condition preceding the reversal .
When a ballistic galvanometer is used in combination with an exploring coil wound on such magnets , the deflections of the instrument do not give the complete change of magnetic induction ; the effect is similar to the damping caused by connecting a resistance across the terminals of the instrument .
A continued current in one direction , caused by the slow change of magnetisation , gives rise to a first deflection considerably greater than the second\mdash ; in the opposite direction\mdash ; and the third deflection may exceed the second .
So far as are aware , the lowest forces at which the delay caused by induced currents in large masses has been examined , are and in C.G.S. units .
In the first case the magnet was of wrought iron of diameter cm .
( 4 inches ) , and in the second , of dynamo magnet steel of diameter cm .
( 12 inches ) .
Changes at the centre of the core were observed 46 seconds after reversal in the former case and 400 seconds in the latter .
The experiments showed , moreover , that , with small forces of this order of Phil. Trans.\ldquo ; Journ IProc .
Profs .
J. W. Nicholson and E. Wilson .
Magnetic magnitude , there is but a single maximum rate of change of the magnetic induction , which occurs at a comparatively early stage .
But the values of the magnetic induction corresponding to these forces were of the orders 7000 and 5000 C.G.S. .
In the experiments of present communication , the netic induction in the outermost shell is only C.G.S. units when a current of 7 amperes is used in its magnetisin coil .
Corresponding to a current of amperes in the coil wound on the third shell , the maximum transverse induction in the urth or outermost shell is 272 C.G.S. units .
These currents have not as yet been greatly exceeded in our experiments on demagnetisation .
It is interesting , and of some importance , to enquire whether the delay caused by the reversal of such small inductions is sufficiently great to affect the decrement , as ordinarily obtained , of the readings of a galvanometer whose periodic time is 8 seconds .
Some nents have been made in order to test this point .
A list of deflections due to a reversal of a current in the third coil , and its corresponding effect on an exploriIlg coil wound on the fourth shell , has been used already for another purpose .
The same arrangement , with a current of amperes reversed , and various additional resistances inserted in the galvanometer circuit , was used also in determining the ratio of first to second deflection corresponding to each resistance .
The exploring coil was then cut out of the galyanometer circuit , and deflections were observed corresponding to the same resistances as before , ranging to 10,000 ohms , when the coil of the galvanometer was initially set in motion by a primary cell of low voltage woYking through a very high resistance , the battery circuit being open while the readings were taken .
The ratios of first to second deflection were again calculated , and the comparison is made in Table Table When the additional resistance was zero , or in short circuit , the second deflection was so small as to render the ratio very uncertain .
But the ratios Induction and its Reversal in Spherical Iron Shells .
147 corresponding to other resistances are in such close agreement that the delay in the change due to eddy currents must be effectively negligible .
This result is of some importance , for it might have been inferred from the earlier experiments that a considerable delay was possible in a shell of thickness 2 .
The argument would be applied as follows:\mdash ; In the distribution of induced currents in its mass a plate is comparable with a wire whose diameter is twice the thickness of the plate ( loc. cit The time required to effect a complete reversal of any value of the magnetic induction in one of our shells should be of the same order as in a cylinder of diameter 4 cm .
In two cylinders of different diameters , .
each subjected to the reversal of the same magnetic force , similar magnetic events will happen , but at times proportional to the squares of the diameters .
Comparing therefore a solid shell of thickness 2 cm .
with the solid magnet of diameter 10 cm .
, mentioned in the earlier experiments , the delay in the shell would be of order 46 seconds .
Comparison with the magnet of diameter 30 cm .
gives 400 seconds .
It is evident that if the induced currents are to disappear when the magnetic induction is of the order 6000 , a considerable time must elapse between successive reversals of the magnetising force in this process of demagnetisation by reversal and diminution .
But the present experiments show that for magnetic inductions not exceeding 300 C.G.S. units this necessary time interval is greatly reduced , for the delay due to induced currents is negligible comparison with such a time as 8 seconds , the periodic time of the ballistic galvanometer employed .
This has important implications in connection with work at frequencies .
Summary .
( 1 ) The paper contains a solution of certain problems which arise in the production of an effective magnetic shield for large spaces .
These relate mainly to the effective demagnetisation of the shells of which the shield is constituted .
( 2 ) Theoretical solutions of problems relating to the effects of indefinitely closely wound coils on various shells of such a shield are given , and compared with the experimental values for an actual .
coil , as determined by an exploring coil and ballistic galvanometer .
( 3 ) The experiments supply an estimate of the deviations of Maxwell. 's formula , for the field inside a spherically wound helical from the true values , when the spacing in the iS of importance . .
XCIIL\mdash ; L Lord Rayleigh .
( 4 ) A study of the necessary interval between current reversals in the process of netisation has been made , and it is shown that the delay in reversal of magnetic phenomena in considerable masses of iron , due to eddy currents , is extremely small when the magnetic inductions are less than 300 C.G.S. units .
On the of Revotving Fluids .
By RAYLEIGH , O.M. , F.R.S. ( Received December 8 , 1916 .
) So much of depends ultimately upon the dynamics of revolving fluid that it is desirable to formulate as clearly as possible such simple conclusions as are within our reach , in the hope that they may assist our meant when an exact analysis seems impracticable .
An important contribution to this subject is that recently published by Dr. Aitken .
* It formed the starting point of part of the investigation which follows , but I perhaps to add that I do not share Dr. Aitken 's views in all respects .
His paper should be studied by all interested in these questions .
As regards the present contribution to the theory it may be well to premise that the limitation to symmetry round an axis is imposed throughout .
The motion of an inviscid fluid is governed by equations of which the first expressed by rectangular co-ordinates may be written , ( 1 ) where , ( 2 ) and is the potential of extraneous forces .
In ( 2 ) the density is either a constant , as for an incompressible fluid , or at any rate a known function of the pressure .
Referred to cylindrical co-ordinates , with velocities , reckoned respectively in the directions of increasing , these equations , ( 3 ) " " The Dynamics of Cyclones and Anticyclones.\mdash ; Part 3 'Roy .
Soc. Edin .
Proc. , vol. 36 , p. 174 ( 1916 ) .
Compare Basset 's ' Hydrodynamics , ' S19 .

