The Alternate Current Transformer Vol. 1: The Induction of Electric Currents (1896) — part 15 of 35

Blaserna drew from his observations the deduction that there is an interval of delay in the starting of the secondary currents, and that a small but measurable time elapses between the instant of making or breaking the primary circuit and the beginning of the secondary current. From this he made a calculation as to the velocity of electromagnetic induction, and he also stated that the interposition of dielectric substances such as glass or shellac between the coils reduced the so- calculated velocity.

Bernstein (Pogg. Ann., Bd. CXLIL, 1871, p. 72) repeated these observations of Blaserna, but did not confirm these last results. He found that the first oscillation always began at the instant of breaking or making the primary circuit, and he

• In The Electrician for June 1, 1888, a curve is given by Mr. F. Higgins, showing the rise of current in the magnets of type-printing telegraphs, and the oscillatory character of the current at starting is well marked. Mr. Higgins's curve gives the results of actual observations.

252

MUTUAL AND SELF INDUCTION.

found no effect produced by the interposition of dielectric media.

Helmholtz has carefully examined these results of Blaserna and criticised them.* He remarks that Blaserna used for his coils flat spirals of wire with many turns, and also he used the current from several Bunsen cells to create the primary current. Not only do the spirals act like a condenser, giving the whole apparatus a sensible electrostatic capacity, but the use of a battery of high electromotive force causes a considerable spark at the break, which spark has a very sensible and rather irregular duration. Also in Blaserna's experiments, the two circuits were placed at various distances apart. If a current

FIG. 93.

is started in a primary coil the effect of the induced current created in the secondary by its re-acfcion on the primary is to hasten the rise of the primary current, and at the break to accelerate its decay. As the secondary circuit is moved further off this effect is less marked. Hence, the rise and fall of the primary is more gradual and the arrival of the secondary current at its maximum value is more delayed. From this results, then, an apparent retardation of the time of the arrival of the maximum of the induced current.

• Helmholtz, " On the Velocity of the Propagation of Electrodynaraic Effects," Phil. May., Ser. 4, VoL XLIL, 1871, p. 232.

MUTUAL AND SELF INDUCTION. 053

Helmholtz conducted a series of experiments by means of his pendulum chronoscope. A heavy iron pendulum P (see Fig. 94), 'the lower end of which carried two plates of agate, could be made to execute one swing and then be caught °by a detent. These plates of agate in the course of the swing were caused to strike against and tip over two little levers I, I'. One of these levers was fixed, and the other could be moved forward so as to separate the blows. One was made to break the circuit of a primary coil Pi; when tipped over, and the

Se,

FIG. 94.

other by its movement separated a connection between a con- denser and the ends of a secondary coil, Sec, attached to it.

These being arranged, the fall of the pendulum executed these two "breaks" successively, separated by an interval of time capable of being calculated from the known motion of the pendulum. The two circuits were placed 170 centimetres apart. The primary consisted of 12 turns of thick wire, and the secondary of 560 turns of fine wire, The current was sent

254 MUTUAL AND SELF INDUCTION.

from one Daniell cell. The two ends of the secondary were connected to the two plates of the condenser, and when the pendulum fell it broke the primary current and started in the secondary circuit an oscillatory current reverberating to and fro in the secondary wire, the condenser acting as a resonator. At a definite interval after rupture of the primary, the condenser was separated and examined by a quadrant electro- meter. The charge in the condenser showed the phase of the electrical oscillation existing at the instant of such separation. In one case Helmholtz observed 35 oscillations in J^th of a second. In order to discover if any retardation took place with increased distance of the coil, it was necessary to fix attention upon some phase in the oscillations. The successive zero points of the current were very sharply defined, and suitable for this purpose. Helmholtz found that alteration of the distance between the primary and secondary coils made no perceptible difference in the position of the zero points, and that, as far as the apparatus he was using could detect, the velocity of the electro -magnetic impulse must be greater than 195 miles per second. He pointed out in this Paper that the commencement of the secondary current is not a sharply marked thing. The spark which takes place at break of the primary lasts an appreciable time, and all this time the primary is dying gradually, and the induced current therefore is increasing. The period of duration of the break spark may be something like y-giy^th to ^thn^k °f a second, and is, therefore, a large fraction of the duration of a single electrical oscillation, which amounted to about -^Vyth of a second. The duration of the break spark can be found by observation of the time which elapses from beginning of break up to the first zero point of the secondary current oscillations, as com- pared with the mean value of the duration of an oscillation. The interval up to the first zero point is the duration of the break spark plus the time of half a complete oscillation. The duration of the spark is never constant, and depends a good deal on the amount of platinum thrown off from the contacts each time. The average duration of the spark in Helmholtz's experiments was found to be about one-tenth of the whole period of an oscillation. Helmholtz also noticed in some earlier observations evidence of electrical oscillations set up in

MUTUAL AND SELF INDUCTION. 255

a flat spiral, one end of which was insulated. In this case some 45 oscillations were detected in the space of ^th of a second. Henry also noticed that the time of subsidence of the current, when the circuit is broken by means of a surface of mercury, is very small, and probably does not much exceed the ten-thousandth part of a second. It has, however, a quite appreciable duration, for Henry found that the spark at ending presents the appearance of a band of light of considerable length when viewed in a mirror revolving at the rate of six hundred revolutions per second.

Bernstein, with the aid of a contact break somewhat different from that used by Blaserna, also examined the duration of the oscillations set up in a secondary coil. He found that the duration of the first oscillation at breaking primary was longer than that of the subsequent ones. The mean duration when using a single Grove cell in the primary circuit was '0005 second, and when using a Daniell cell only •0001 second. We shall return later to consider more recent researches on these electrical oscillations in inductive circuits and point out that they can only occur when some part of the circuit possesses sensible electrical capacity. In the case of a coil or bobbin of wire we have not only resistance and inductance, but measurable capacity present in the conductor.

§ 8. Magnetic Screening and the Action of Metallic Masses in Induction Coils. — At one stage* of his investi- gations Henry made the important discovery that, if a primary and secondary coil are separated by a metallic sheet, a notable decrease takes place in the intensity of the shock taken from the secondary circuit when a sudden discharge is passed through the primary, or continuous current started or stopped in the primary circuits. A thick copper plate was found more effective than a thin one in thus preventing the inductive effect of the primary upon the secondary coil. If a radial slit was cut in a circular metallic plate the annulling effect was altogether stopped. If the two edges of the gap (see Fig. 95) were furnished with wires leading to a magnetising spiral, Henry found he could in this way make evident the existence in the plate of a current induced by the action of the primary.

• ^hoTMag., Vol. XVI.7l840, p. 257.

256

MUTUAL AND SELF INDUCTION.

A flat coil of insulated wire was substituted for the metal plate, and it was found that the screening action of this coil was only sensible when the two ends were joined so as to com- plete the circuit. This action, by which the induction of a primary coil on a secondary is prevented by the interposition of a metallic plate, cylinder, or closed circuit of insulated wire, is called magnetic screening. The elementary explanation of this effect is not difficult to arrive at. Suppose a small con- ducting circuit of resistance B to be placed in a magnetic field so that it is traversed normally by N lines of magnetic induction. Let the constant coefficient of self-induction of this circuit be L. If, then, in any small time d t a variation of the lines of induction traversing this circuit takes place, the impressed

FIG. 95.

d N

electromotive force on that circuit will be represented by

d t '

and if at that instant the current in the circuit is i, by the principles laid down in the last chapter the current equation will be

~dt' -±-(Li + NUR»-0.

Suppose the conductivity of this circuit to be perfect, and E therefore zero, we have, by integration of the above, equation, the result

L i + N = const. ;

MUTUAL AND SELF INDUCTION. 257

in other words, the lines of induction L i, linked to the circuit at any instant due to the induced current generated in it, are opposite in direction to those whose variation is producing the current, and together with them make up a constant number. Hence, if the variation of N is such as to take lines of induction out of the circuit, the action of the current thereby induced is to add or increase them in the circuit at an equal rate. If we suppose our circuit to be a perfectly conducting metal plate, and just behind this metal plate there is another small closed circuit, then any variation of lines of induction passing through this plate will not take effect in producing any induced current in the small circuit, because the inductive action of the current induced in the plate nullifies, as far as the small circuit is concerned, any vari- ation of the external field. It is clear that these conclusions would apply to any surface of finite extent which possessed perfect conductivity ; the induced currents which any vari- ation of the external field would produce in this surface would always be such that the induction through each portion would be kept constant — in other words, that the perpendicular component of the magnetic induction at each point on the surface would retain a fixed value. It follows that a closed surface of zero resistance is a complete screen for all points in the interior against the effects of variation of the field on conductors on the outside of the surface ; these effects reduce to the production of surface currents in the shielding conductor, which keep the resultant field in the interior constant or at zero.

Faraday describes (" Exp. Researches," Vol. L, §1720 et seq.) an experiment which at first sight seems to disprove the fact of magnetic screening. He placed a flat copper wire spiral, which was in connection with a battery and key, between two other flat spirals which were respectively connected with the two coils of a differential galvanometer. The coils were so joined up that the inductive effect of a break and make of the battery circuit produced no movement of the galvanometer needle because it was subjected to two equal and opposite impulses from the two coils. When an exact balance was obtained a flat plate of copper, nearly three-quarters of an inch thick, was interposed between the primary spiral and one of

s

258 MUTUAL AND SELF INDUCTION.

the secondaries. The galvanometer needle was not, however, any more affected than if the copper was absent. To under- stand this we must bear in mind that the break or make of the primary current produces in the copper a secondary current, but as the effect of the primary coil on the secondary coil on that side is balanced by the other one we may regard the secondary coil next the copper plate as free to receive any inductive effect it can from the eddy current induced in the copper block. This secondary current induced in the copper generates a tertiary current in the secondary spiral, and this tertiary current consists, as we have seen, of a double short flux of electricity equal in quantity and opposite in sign. The galvanometer is then traversed by two small equal quantities of electricity in opposite directions, and as this does not sensibly affect a not very sensitive galvanometer no movement of the needle is seen. If, however, instead of the differential galvanometer, Faraday had used a differential telephone, he would have found distinct evidence of a screening action. Again, suppose that, instead of a simple make or break, Faraday had employed a steadily periodic or alternate cur- rent in the primary, this would have set up a steady periodic secondary current of equal frequency in the copper plate, and this again would have set up in the secondary coil on that side a steadily periodic tertiary current of equal period, and this might have been detected by the use of a sensitive differential electro-dynamometer or a soft iron needle galvanometer.

Henry found that a sheet of tinfoil afforded a very small amount of screening for shock, but a thick sheet of copper a very considerable one in the case of induction by battery cur- rents, and in the case of induction by Ley den jar discharges the same phenomenon was apparent. In the case of an iron screen there is an additional effect, due to the fact that the iron, by its small magnetic resistance, conducts away the lines of induction somewhat through its mass, and prevents them from extending to the space on the other side. In this case also a considerable thickness of metal is necessary to bring about the effect of annulment. When we are limited to the use, as we are in practice, of materials whose conductivity is far from being perfect, it is found that a thin screen of metal hardly affords any sensible protection from inductive effect.

MUTUAL AND SELF INDUCTION. 259

In other words, the field on the other side of the screen is " very far from constant. This has been well demonstrated in certain investigations by Prof. D. E. Hughes in carrying on some highly valuable experimental researches into the means of preventing induction upon lateral telegraph wires.* It has many times been proposed to annul mutual induction between telegraph and telephone wires by covering them over with thin metal covering, which covering is kept " to earth." It is now known, and well exemplified in Prof. Hughes's experiments, that this shielding affords no protec- tion when the covering is not very thick and when the rate ol change of the currents is not very rapid. A gutta-percha wire was enclosed in ten coverings of tinfoil, and such arrangement was not found to afford protection to induction, as detected by a telephonic wire stretched alongside. Even when twenty coatings of thin charcoal iron were put round the wire, not only was there found to be a very sensible per- manent field outside the iron, but changes of field were made manifest also. It is not to be taken that these experiments disprove the fact of magnetic screening, but only that the low conductivity of the envelopes used is ineffective at the speed of current change employed to render visible the effect of magnetic screening. It is different, however, if the induc- tive effects are being produced by a very rapid rate of change of field. For suppose that a small circuit, as before, is placed in a uniform field, and is traversed by q lines of induction due to this external field. Suppose q varies according to a simple periodic law, so that q = Q cosp t, where p = 2?r nt n being the frequency of the alternations. Then we have

but - — is the value of the impressed electromotive force in

d t

the circuit, and if we call the current at any instant i, then, by the principles in Chap. III., we have

sm(pt-6),

• See a Paper by Prof. Hughes " On Lateral Induction in Telegraph Wires," read before the Society of Telegraph Engineers, March 12, 1879. Published in The Electrician, March 22, 1879.

s 2

260 MUTUAL AND SELF INDUCTION.

in which R is the resistance and L the inductance of the- circuit, and

Suppose that R is very small compared with L^, whichi is the case when n or the frequency of alternation is made very great, then R vanishes compared with ~Lp, and if we call i' the value towards which i approximates in this case, we have

0

• — cosp t, L

and l =

d t, L

. dt dt

Hence L^'--li,

dt dt

or L i' + q = constant.

Hence the field due to the current in the circuit, together with the external field, is a constant quantity, and we get the- condition of perfect shielding. We may sum up the fore- going by saying that, if a screen of absolutely no electrical resistance is interposed between a primary and secondary coil, it- effects a perfect magnetic screening, whatever may be its thick- ness. If, on the other hand, the screen has a finite conductivity, then the screening will be very imperfect, unless a very great. thickness of material is used, and the above will be true when the change of field or the change of primary current is a simple- "make" and "break" or a slowly periodic change. When,. however, the change of current in the primary is very rapidly periodic, then the screening effects of even imperfect conductors will make themselves felt, and a comparatively thin screen of metal will effect a nearly perfect shielding for induction. This theory is strikingly confirmed by some very beautiful experi- ments of Mr. Willoughby Smith, which are described in the- Journal of the Society of Telegraph Engineers (November 8, 1883, Vol. XII., p. 458),* and entitled " Experiments on Volta-

• See also The Electrician, November 17, 1883, p. 18.

MUTUAL AND SELF INDUCTION. 261

Electric Induction." Mr. Willoughby Smith's apparatus con- sisted of two flat coils A and B (see Fig. 96), placed a certain •distance apart. One of these was a primary coil connected with -a battery, and the other was connected with a sensitive galvano- meter. In the circuit of both were current reversers, which reversed the galvanometer and battery alternately, and hence made the opposite induced currents both affect the galvano- meter in the same direction. This being arranged, the commu- tator was started so as to reverse the currents very slowly, and a sheet of copper interposed between the spirals. Under these •circumstances the interposition of the copper produced but little effect. If, however, the commutator was driven at a very rapid rate the copper plate caused a marked diminution in the galvanometric deflection, and this diminution was greater in proportion as the speed was greater. In the original Paper

FIG. 96.

a curve is given (Fig. 97) which shows the decrease in the galvanometer deflection, expressed as a percentage of the •original undiminished deflection, corresponding to various •speeds of reversal. It will be seen that the less the conduc- tivity of the metal the greater must be the speed in order that the magnetic screening may approach perfection. Iron, of •course, occupies an exceptional position. It cuts off, even at very low speed reversals, a large portion of the field, not by a true screening action, but by conducting away the lines of magnetic force and preventing their access to the secondary <j}il. It will be seen that at any given speed the order in which the metals reduce the deflection is the order of their electric conductivity, and that as far as the diagram goes the lines all (except iron) slope upward, indicating that at very

-202

MUTUAL AND SELF INDUCTION.

"high speeds the screening of even the worst conductors will approach perfection. It would no doubt be found that, if the

-telephone were used as a detector, the magnetic screening of a»

"copper plate or thin tinfoil sheet would become very manifest for high notes when not in anyway marked or distinguishable

-for notes or sounds of low frequency of vibration.*

As far back as 1840 Dove had made experiments! on the

• effect of the introduction of cores of various materials into the primary circuit of an induction coil. His apparatus consisted of two similar primary bobbins wound on tubes of non-metallic substance and connected in series (Fig. 98). Over each primary bobbin was wound a secondary circuit, and these secondary circuits were connected in series, but so that the induction of

90

80

I70 s53

r

« 40 30 20 IO 0

1 ^^

Copper Zinc

Tin Iron

Lfead

\ /^

^*^~

\ /

^~^~~

JL- 3

^s^

/

/

^s'

/

/

^^

/ 2

^^> —

/ /

7 / ^5

„

100 £00 1000 1500 200C!

Rsversals Per Minute

FIG. 97.

the two primary bobbins operated in opposite directions and nullified on the whole secondary circuit each other's effect. Exact neutralisation was obtained by adjusting one of the secondaries. When this was the case, various cores of iron rods of different kinds were inserted in one primary bobbin, and it was found that the induction balance was destroyed.

• The above explanation of the cause of the difference between the screening of the different metals is not that given by the distinguished investigator, but it is the explanation which to the author seems most in accordance with known principles.

t Dove, Poggendorffs Annalen, Vol. XLIX., 1840.

MUTUAL AND SELF INDUCTION.

263

By inserting iron wires of a certain size in the other core, balance could be again obtained, but not simultaneously, as estimated by the galvanometer and by the shock. Thus, with a bar of forged iron, 110 wires had to be inserted in the other coil to obtain an equilibrium, as estimated by the galvano- meter; but, as far as could be judged by the shock, 15 wires

were sufficient. With regard to different kinds of iron, experi- ment shows that if we class them according to galvanometrie effect we obtain a different series to that at which we arrive when classifying them in the order in which they create sensation by shock. Thus grey rough cast-iron is the kind

264 MUTUAL AND SELF INDUCTION.

which approached nearest to bundles of soft iron wire in respect of increasing the shock. Enclosing iron wires in a brass tube reduced the action of the wires in disturbing the indictive balance and rendered them very little better than a bar of solid iron. When the primary current was a discharge from a Leyden jar, Dove found that the physiological effect (shock) of the secondary current, as estimated, was reduced by the introduction into the primary bobbin of non-magnetic conducting cores ; in other words, the introduction of a core of non-magnetic but highly conducting material into the primary bobbin reduced the power of a primary discharge to create a secondary discharge. These last results may be obtained in a more modern form by the substitution of a Bell telephone to detect the tertiary currents generated by the metal core.

Let a Bell telephone be connected in series with the secondary coil of a small induction coil, of which the primary is wound on a hollow bobbin and the frames are wholly of wood or non-metallic substance. A convenient form is that known as Du Bois-Beymond's sliding coils. Let an interrupter in the primary circuit make and break the circuit rapidly. This being so, the telephone emits a steady rattle or hum. If a massive copper rod is introduced into the primary bobbin as a core, the telephonic rattle is more or less suppressed ; if a core of soft iron wire is introduced the noise is increased ; if a core of solid iron or steel is used the noise may be increased, but not so much as when the divided iron is used. The explanation of the exalting effect of the soft iron wire is simple. The presence of the iron reduces the reluctance of the magnetic circuit. More lines of induction therefore flow through the secondary circuit, and hence the strength of the secondary current is increased, and the mean rate of change of induction through it is also increased. The diminishing effect of the copper core is explicable in the light of the knowledge that in such a conducting core the primary current generates induced currents, and these in their turn re-act upon the secondary circuit, inducing in it a tertiary current. The directions of the currents induced by the primary in the solid core and in the secondary circuit are the same. The direction, however, of the first half of the tertiary current developed in the secondary by the current in the copper core is

MUTUAL AND SELF INDUCTION. 265

opposite to the direction of the current developed in the secon- dary by the action of the primary. Hence it results that the •current in the secondary circuit is more or less wiped out by the opposing inductions due to the primary circuit and the •currents induced in the copper core. Otherwise the operation might be regarded thus : — Suppose the primary circuit to be traversed by a periodic current creating a simple periodic flux of induction through the copper core. As we have seen, under the head of magnetic screening, this variation of induction would induce currents in the copper core, which would them- selves generate a flux of induction which would, if the con- ductivity of the core were perfect, or the rapidity of change of induction infinite, be exactly equal and opposite at each instant to the flux of induction producing those currents.

If the conductivity is not quite perfect, or the rate of -variation not very great, yet nevertheless the direction of

FIG. 99.

the field of magnetic force inside the copper, due to the cur- rents induced in its mass, will more or less oppose the field of force at every instant which is by its fluctuations generating those currents. If the thick line III in Fig. 99 represents the sinusoidal or simple periodic change of induction or magnetic field in the interior of the copper, due to the primary helix, :and if the dotted line 22 represents roughly the changing field due to the eddy currents generated in the core, which are nearly 180° behind the primary in phase, the integral or sum of both superimposed fields represented by 3 3 at any instant is less than the original one due to the primary alone at the corresponding instant. Also the mean rate of cliamje of the resultant field is less, and the secondary circuit experiences at every instant a less inductive electromotive force. The same reasoning which we have employed in the case of magnetic

266 MUTUAL AND SELF INDUCTION.

shielding applies here, and the differences in the reducing effect of cores of various rnetals would be found to be less at high speeds of alternation than at low. In some small induction coils used for medical purposes the strength of the secondary current is graduated by drawing in or out of the primary coil a copper tube which slips over the bundle of fine iron wires used as a core. The rationale of the action of this copper tube in so operating is in a general way to be found in the principles laid down above.

When Henry obtained possession of the " Experimental Researches " of Faraday, as detailed in the fourteenth series of his " Experimental Researches," he was exercised in his mind to reconcile the results obtained by Faraday on the interposi- tion of metallic screens between inducing and induced circuits with his own. Faraday had found that when the galvanometer was used as a current finder " it makes not the least differ- ence " whether the space between the primary and secondary coils was air, sulphur, shellac, or such conducting bodies as copper and other non-magnetic metals. On the other hand, Henry found that a shock from a secondary coil which, would paralyse the arms was so much reduced by the inter- position of a metallic plate as hardly to be sensible on the tongue. Here was evidently something to be explained, and in a long memoir (Phil. Mag., Series 3, Vol. XVHL, 1841, p. 492 ; also Transactions of the American Philosophical Society, Vol. VHL, 1840) Henry examined this and other matters. He first verified Faraday's experience by attaching the ends of a secondary coil to a galvanometer and bringing up suddenly towards it a permanent magnet, or a coil traversed by a steady current. The swing of the galva- nometer was found to be quite unaffected in extent by the interposition of a plate of copper. Again, in place of the copper plate, a closed metallic conductor (an endless coil) was employed, but whether the circuit of this coil was open or closed it made not the slightest difference on the galvano- meter deflection.

Forty feet of copper wire, covered with silk, were wound on a short cylinder of stiff paper, and into this was inserted a hollow cylinder of sheet copper, and into this again a rod of soft iron. When the latter was rendered magnetic, by

MUTUAL AND SELF INDUCTION. 267

suddenly bringing in contact with its two ends the different poles of two magnets, a current was generated in the wire, but the strength of this current, as measured in the galvano- meter, was the same whether the copper cylinder was present or was removed. Henry then noticed that there was one element of difference between the indications of a galvano- meter and that of the magnetising spiral. If the two secondary currents at "break" and "make" of a primary were sent through a magnetising spiral and through a gal- vanometer, the arrangement might be such that the induced current at "make" of the primary was unable to give any sensible magnetisation to the steel needle enclosed in the spiral, but at " break" was able to magnetise it to saturation. Nevertheless, in both cases the "throw" of the galvanometer was the same. Similarly with the degree of shock felt, the galvanometer indications being alike for the inverse and direct induced current ; yet that induced current gave the greatest shock which was able to produce the greatest magne- tisation. The explanation of these facts became clear as soon as it was seen that the deflections of the galvanometer depended upon the whole quantity of the discharge, and must necessarily be alike for the inverse and for the direct current, but that the magnetising effect and the physiological shock depended upon the maximum value of the instantaneous discharge current, and might therefore be very different for the two induced currents. It was then evident that any -actions by which this maximum value of an induced current was decreased, whilst its duration was increased and total quantity left unaltered, would result in rendering this current less easily detectable by shock or magnetisation, but make no difference in its effect on a galvanometer. Aided by this thought, he repeated Faraday's experiment with the balanced coils referred to in § 8 ("Experimental Eesearches," Vol. L, § 1,790 et seq.). A galvanometer was provided having two equal wires of the same length and thickness wound on the same frame, and also a double magnetising spiral was pre- pared by winding two equal wires round the same piece of hollow straw. Coil No. 1, connected with a battery, was supported perpendicularly on the table, and coils Nos. 3 and 4 were placed parallel, one on each side, and each coil connected

268 MUTUAL AND SELF INDUCTION.

in series with one coil of the differential galvanometer and with one spiral of the magnetising helix. The two outside •coils were then adjusted so that when the battery circuit was made and broken, and the current started and stopped in the middle coil, no indication was given by the galvanometer, and no magnetisation produced in a steel needle placed in the double helix. A thick zinc plate was then introduced between the primary coil and one of the secondaries, and it was found that the needle of the galvanometer still remained stationary on making and breaking the primary current, but that the steel needle in the spiral became powerfully magnetic. This indicated that the two secondary currents, whilst still equal in total quantity, had been so affected that one had a less maximum value than the other, and hence a differential magnetising action was produced. A similar effect was observed when a galvanometer and magnetising spiral were together introduced into the secondary circuit of a single primary and secondary circuit. The interposition of a metal sheet considerably reduced the magnetising power or the shock, but left the galvanometer deflection unaltered. In order to increase the number of facts, this last experiment was varied by the exchange of a soft iron needle for the hard steel needle in the magnetising coil, the metal screen being interposed in each case, and it was found that whereas the metal screen cut off almost entirely the power of the secondary current to magnetise hard steel, it could yet slightly magnetise the soft iron. A screen of cast iron half an inch thick, how- ever, not only neutralised the power to magnetise hard steel, but reduced the deflection of the galvanometer as well. The general explanation of the foregoing facts, as due to Henry, is as follows : — The secondary current, as we have seen, is a brief discharge, which rises very suddenly to its maximum value and then fades gradually away. The current curve of the secondary current, due to the rupture of a primary circuit, may be represented by the thick firm line in Fig. 100. If a metal screen is interposed between the primary and the secondary circuit the screen gets a similar secondary current generated in it, and this last again acts by induction to gene- rate a tertiary current in the secondary circuit. This tertiary current consists of two portions — first, an inverse part opposite

MUTUAL AND SELF INDUCTION. 269>