Exp. 5.-A current of 2 ampères was passed through the same wire, resulting in an elongation due to heating of 460 ten-millionths, the temperature of the wire being therefore raised about 3°3. The former observations were again made with the results given in the last column of Table III and in fig. 2.

It will be seen that with both specimens of iron wire, the effect of a current is of just the same general character. It acts oppositely to tension, heightening the curve of elongation instead of lowering it. This action is certainly not due either directly or indirectly to mere current heating. It has been shown that the thinner wire even when carrying 2 ampères was only about 10°-7 warmer than when no current was passing through it. Such a small rise of temperature would be quite incompetent by itself to account for the effect in question, for the elongation curves of a given specimen of iron have been found to be not sensibly altered when taken under widely different conditions of temperature. Nor would it exert any material influence upon the susceptibility of the iron; and, even if it did, the curves would not be affected in the manner observed.

It is hardly worth while attempting to frame an explanation until many more phenomena of the same order have been investigated. Similar experiments were afterwards made with nickel and cobalt.

Exp. 6.-A nickel wire was used, the diameter of which was 0.65 mm. The retractions which it underwent in fields of gradually increasing strength are given in the second column of Table IV.

Exp. 7.-A current of 1 ampère was passed through the nickel wire, producing a heat elongation of 340 ten-millionths. Taking the coefficient of expansion as 0-0000129, this implies a rise of temperature of 2° 6. The retractions of the wire when carrying a current are given in the third column of the table. Remembering that the figures in the second and third columns denote millionths of a centimetre, the close agreement between the two is very remarkable. I have elsewhere* fully described the method of observation adopted, but I may perhaps mention that each number as set down in the table was obtained by the subtraction of two readings, the one taken when there was no current in the magnetising coil, the other when the current was turned on. The former or zero reading was continually changing, owing to small alterations of temperature, the index rarely being absolutely at rest. All the figures were dictated, and when the second experiment was made, I had not seen the results of the first. I may add that the table contains all the observations which were taken in the two experiments.

Though at first inclined to attribute such small discrepancies as exist entirely to observational or instrumental errors and to infer that the current had no influence whatever upon the contraction, I think it appears pretty clearly from a careful inspection of the differences tabulated in the fourth column that this is not actually the case. Four pairs of observations agree exactly; once only the retraction with the

* 'Phil. Trans.,' vol. 179, A, p. 218.

current seems to be greater than without it, while in the ten remaining pairs the retraction is slightly greater without the current than with it. It may, perhaps, be fairly concluded that the current has a real but very small effect in diminishing the retraction. Now I have before remarked that the degree of retraction which nickel undergoes when magnetised is materially affected by comparatively small changes of temperature: the retraction of the same specimen has been found to be greater in a cold room than in a warm one, at least in fields up to 400 or 500. Probably this is to be explained by the influence of heat in diminishing the magnetic susceptibility of nickel, the retractions being really the same for the same intensity of magnetisation. Such small effect as appears to be produced by the action of the current may, therefore, be accounted for simply by the rise of temperature (2° 6) which it causes.

Tension has a large effect upon the magnetic retraction of nickel :" it is, therefore, the more remarkable that the action of a current, which operates so markedly upon iron, should in nickel be practically insensible.

Exp. 8.-The results with no current obtained for a strip of rolled cobalt, the length of which between the clamps was 10 cm., and the cross section 1.82 sq. mm., are given in the first two columns of Table V.

Table V.-Cobalt Strip, section 1.82 sq. mm.

Exp. 9.-A current of 2 ampères through the strip caused a heat elongation of about 600 ten-millionths, indicating, if the coefficient of expansion is taken as 0.0000125, a rise of temperature of 4°8. The retractions observed while this current was passing are set out in the third column of the table. From an inspection of the differences

Roy. Soc. Proc.,' vol. 47, p. 469.

tabulated in the fourth column, it appears that the effect of the current is to increase the retraction very slightly.

According to Rowland the susceptibility of cobalt is increased by heating. The small additional retraction indicated when the current was passing was, therefore, no doubt due to the increased susceptibility consequent upon current heating. It may be noted that tension seems to have no material effect upon the magnetic retraction of cobalt.*

Summary.

In an iron wire carrying a current, the maximum magnetic elongation is greater, and the retraction in strong fields is less, than when no current is passing. The effect of the current is opposite to that of tension.

The magnetic retractions of nickel and of cobalt are not sensibly affected by the passage of a current through the metals. (Tension considerably modifies the magnetic retraction of nickel, but not that of cobalt.)

III. "On the Measurement of the Magnetic Properties of Iron." By THOMAS GRAY, B.Sc., F.R.S.E. Communicated by LORD KELVIN, P.R.S. Received May 3, 1892.

(Abstract.)

This paper gives the method of experiment and results obtained in some investigations on the time-rate of rise of current in a circuit. having large electromagnetic inertia. The experiments were made on a circuit containing the coils of a large electromagnet having laminated cores and pole pieces. The mean length of the iron circuit was about 250 cr cm. and its cross section 320 sq. cm. The magnetising coil had 3840 turns, when all joined in series, and a resistance of 10'4 ohms. The coils were so arranged that they could be joined in a variety of ways so as to vary the resistance, inductive coefficient, &c., and also to allow the magnet to be used either as an open or a closed circuit transformer.

The electromotive force used in the experiments was obtained from a storage battery, and the method of experiment was to trace the curve, giving the relation of current to time, on a chronograph sheet.

One set of experiments shows the effect of varying the impressed E.M.F. on the time required for the current to attain any given percentage of its maximum strength. The results show that for any particular percentage there is always a particular E.M.F. which takes

* Loc. cit.