then interrupted; and then the needle is withdrawn (disruptive discharge).

The repetition of each of these processes furnishes an increment of magnetism to the needle; and, provided that all the operations effected are of the same sort and the conditions identical, the results of the experiments are well represented by a hyperbolic formula such as formula (1). The limit A appears to be the same for the passages and the interruptions, but less for the establishments.

TABLE IV.-Needle 2 millims. in diameter, magnetized by 1 Grove's element.

After the tenth passage, interruptions having been produced unawares, the moment of the needle suddenly rose to 55.96, and was not carried beyond that limit by 50 establishments. Ten passages with the spiral then raised the magnetic moment to 57.56, and twenty more passages to 57.88.

TABLE V.-Needle 2 millims. in diameter, magnetized by 1 Grove's element.

that the permanent magnetism produced by a current A becomes more considerable when the needle, after magnetization, has been submitted to the action of a current B, feebler and in the opposite direction. If B is made to tend to 0, the phenomenon still continues; and this is not at all surprising, since partial demagnetization (corresponding to loss of the temporary magnetism) is the consequence of the cessation of the current A.

Thirty passages to the spiral did not raise much the magnetic moment of this needle*. The results obtained by the disruptive discharges are less regular than the preceding, although the empiric formula still represents them. The irregularities doubtless proceed from the difficulty of working these discharges in perfectly identical conditions.

The preceding experiments already establish that the extra currents are without magnetic action within the coil which produces them. If it were otherwise, interruption would be a more efficacious process of magnetization than passing the needle to the spiral; now augmentation of the magnetic moment has never been observed when needles magnetized by a great number of passages have been submitted to repeated breakings of the circuit.

II. The circuit comprises, besides the pile, two coils P and Q. The phenomena observed on account of slow passages are the same as in the case of a single coil; but the effect of the extra currents complicates the phenomena arising from interruption. We will suppose the coil P much more powerful than Q. If the two coils are placed one after the other, two needles p and q, magnetized, each in the corresponding coil, to the limit relative to the passages, acquire a greater magnetic moment through the interruptions; but the relative increment is much greater in the less powerful coil. Example:

TABLE VI.-Needles 142 millims. long and 2 millims.

The extreme magnetization corresponding to the passages, from the first two observations of each column, would be, according to our empirical formula, 18-61 for p, and 4.04 for q. The increment of the magnetic moment produced by the interruption is, for p 0.21 in absolute value, and about 4 in relative value; for the needle q the values are 1.46 and 4.

*The two preceding experiments were made immediately after those of Table I., with needles almost identical and in the interior of the same spiral. The establishments and interruptions are effected by means of a cup containing mercury, into which the extremities of the conducting wires dipped.

Therefore, conformably to the observations we have previously given, the extra current proceeding from P is sensible in Q, that from Q in P, but the magnetic effect of each is nil in the coil from which it emanates.

If the coils P and Q are placed as derivations with respect to one another, the direct extra current of each coil traverses the other in a direction contrary to the current from the pile. In this case experiment shows that the interruptions occasion a diminution of the magnetic moment, and the establishments an augmentation, as might have been foreseen from what precedes. III. The circuit comprises a coil and a condenser.

This case is realized with a Ruhmkorff coil in the following manner: the condenser C (fig. 4) is fixed in a derivation destitute of resistance, on which the interruptions are practised. The coil B may also be placed in the derivation, and the interruption be produced in a point of the principal circuit (fig. 5).

In each of these arrangements the condenser forms with the coil a closed circuit apart from the interruption, either directly or by the intermediation of the pile. It would not be the same if the interruption were made in the principal circuit with the arrangement of fig. 4, or in the derivation in the case of fig. 5: these last two arrangements are useless, as experiment has shown; but if one of the former be employed, the rupture of the circuit is observed to produce a diminution of the magnetic moment of the needles magnetized by being passed to the coil.

The effect observed cannot be attributed to the extra currents of the condenser; for these act in a direction contrary to the result obtained. The condenser employed behaves, with respect to the production of extra currents, like a coil of negligible power.

* That of a Ruhmkorff coil, of which the explosion-distance is 3 centims.

But it must be remarked that the intensity of the extra current of the coil is superior to that of the principal current. The direct extra current strongly charges the condenser, which must consequently discharge itself after the extra current has ceased. The coil thus receives, by reflection, the extra current which it has produced, and is traversed by it in the opposite direction to that of the principal current. The reflex effect is, for a given condenser, the more intense the more powerful the coil employed; and experiment shows that with a very feeble coil the effect of the interruption vanishes.

IV. Magnetization by induced currents.—We have studied only the following case :

When the circuit includes two coils, the slow introduction of a core of soft iron into one of them, or its extraction, is without effect upon the magnetism of a needle placed in the other coil. But if the core be introduced slowly and withdrawn suddenly, the direct induced current augments the magnetic moment of the needle placed in the second coil. Repetition of the same operation causes the moment of the needle to tend towards a limit, which it rapidly approaches. The formula

in which A, B, and a are constants, appears very well adapted to represent the magnetic moment y after a passages. The following Table permits the appreciation of the degree of accuracy of the formula ::

The differences between calculation and observation are a little more considerable in this Table than in the preceding ones; but account must be taken of the difficulty of always withdrawing the soft iron in identically the same manner, in order to produced induced currents of the same intensity. Regard being

had to this consideration, the agreement is satisfactory. On the contrary, hyperbolic formula are not at all suitable for the representation of experiments of this sort; they correspond to a much slower increase, sensible even after twenty operations of the same kind, while here the augmentation ceases to be appreciable after seven or eight operations. Quetelet represents by a formula of the form y=B(1-e)

(3)

the magnetic moment acquired by a steel bar magnetizea by one, two, .. a frictions. This formula represents also, as we have just seen, the increment of the magnetic moment produced in a needle by equal induced currents; it does not suit for that produced by the interruptions of one and the same continuous current, acting on a needle innocent of all anterior magnetization. A complete theory of magnetism should account for this differ

ence.

To avoid the intermissions of the principal current in the preceding experiment, the needle should have been placed in a fixed position in its spiral, and its total magnetic moment measured by the method of deviations. The variations observed are related to the permanent magnetism, which may alone be altered by the passage of the induced currents.

We have seen that the induced currents produced by wresting from contact an electromagnet placed in the circuit produce absolutely similar effects.

V. Effects of piles the current of which is not constant.—If the current of the pile employed is not rigorously constant, the effect of the polarization of the electrodes modifies profoundly the phenomena. The following results were obtained with a pile containing bichromate of potash, prepared several days previously.

If the circuit comprises, besides the pile, only one coil, into which a needle magnetized by a great number of passages is introduced, the establishment of the current augments the magnetic moment by a quantity more or less considerable, often enormous. The current therefore possesses a much greater intensity at the moment of its closing than it retains a moment afterwards. When the resistance of the spiral is augmented, the polarization is less strong, and consequently the proper effect of the establishment of the current tends to disappear.

With respect to the interruptions, they have no very marked effect upon the needles, at least while the resistance of the coil is not very strong; but in the latter case, if we introduce into the coil a strongly magnetized needle with its south pole to the left of the principal current, we always obtain by the interrupPhil. Mag. S. 4. Vol. 49. No. 323. Feb. 1875.

H