to be accurately measured by means of any ordinary form of ballistic galvanometer joined in circuit with a second, or induction, coil surrounding the iron.
It occurred to me, somewhat more than a year ago, that the cause of the difficulty in the ordinary method of measuring magnetic quality in large masses of iron might be made an advantage if the method of experiment were changed, that, in fact, the desired information might be more easily obtained by recording, not the total value of the induction for each of several particular values of the magnetizing force, but the curve showing the rise of current with time, immediately after the circuit is closed. Preliminary experiments showed that, for many of the cases most difficult by the ordinary method, this could be done both accurately and easily. In these preliminary experiments the coils of a large electromagnet, having laminated cores and pole pieces, were joined in circuit with a storage battery and a non-inductive resistance. An automatic arrangement was devised, by means of which the circuit could be closed and the instant of closing recorded by a chronograph. The ends of the non-inductive resistance were at the same time connected to the quadrants of an electrometer. At any desired interval after the closing of the circuit the electrometer could be disconnected from the resistance, and the instant of disconnecting recorded by the chronograph. The deflection of the electrometer needle then showed the difference of potential between the ends of the non-inductive resistance at the instant when it was disconnected. By this means a sufficient number of points on the curve of rise of current could be observed to enable it to be drawn. The results obtained in these early experiments agree with those given in this paper, and the method may be found useful in cases where the rise of current is too rapid to allow of the successful operation of the method since used. It is evident that a modification of this method with a ballistic galvanometer substituted for the electrometer could be easily arranged to give satisfactory results. The ordinary ballistic galvanometer and induction coil method can also be used with a break circuit arrangement, allowing the total time integral to be taken by steps. This method of experiment was not used after the preliminary results for which it was devised had been obtained. It was abandoned, partly because of its considerable complication, but mainly because a considerable number of separate experiments are required to obtain the complete curve, thus involving a corresponding number of independent magnetizations of the magnet.
The arrangement of the apparatus used in the experiments, the results of which are here illustrated, is shown in the annexed diagram (fig. 1). In this diagram is the electromagnet experimented on; , a commutator for reversing the battery, or, when the connection to it was as shown by the dotted lines, for simultaneously cutting out the battery and short-circuiting the magnet coils; is the battery, and a resistance of about one ohm in its circuit to prevent excessive current, when, by the operation of the commutator, it is for an instant short-circuited; is a mirror galvanometer, nearly "dead beat" and of short period; , the galvanometer lamp; , a chronograph, and , a key for closing the circuit. The drum of the chronograph is
