the dotted arrow, this current also circuits the auxiliary wire, but in the direction indicated by the arrows on the dotted curve, which is the same direction in which it circuits the main wire. Thus the needle is deflected, and a signal received. When the direction of the chief current at the transmitting station is reversed, so also is the direction of the artificial current, so that again the needle is balanced. Similarly, if the direction of the current from the distant station is reversed, so also is the direction in which this current traverses the auxiliary wire, so that again both effects conspire to deflect the needle. There is, however, another way in which an auxiliary wire may be made to work. It may be so arranged that, when a message is transmitted, the divided current flowing equally in opposite directions, the instrument at the sending station is not affected; but that when the operator at the distant station sends a current along the main wire, this neutralizes the current coming towards him, which current had before balanced the artificial current. The latter, being no longer counterbalanced, deflects the needle; so that, in point of fact, by this arrangement, the signal received at a station is produced by the artificial current at that station, though of course the real cause of the signal is the transmission of the neutralizing current from the distant station. The great value of duplex telegraphy is manifest. Not only can messages be sent simultaneously in both directions along the wire—a circumstance which of itself would double the work which the wire is capable of doing—but all loss of time in arranging about the order of outward and homeward messages is prevented. The saving of time is especially important on long lines, and in submarine telegraphy. It is also here that the chief difficulties of duplex telegraphy have been encountered. The chief current and the artificial current must exactly balance each other. For this purpose the flow along each must be equal. In passing through the long wire, the current has to encounter a greater resistance than in traversing the short wire; to compensate for this difference, the short wire must be much finer than the long one. The longer the main wire, the more delicate is the task of effecting an exact balance. But in the case of submarine wires, another and a much more serious difficulty has to be overcome. A land wire is well insulated. A submarine wire is separated by but a relatively moderate thickness of gutta-percha from water, an excellent conductor, communicating directly with the earth, and is, moreover, surrounded by a protecting sheathing of iron wires, laid spirally round the core, within which lies the copper conductor. Such a cable, as Faraday long since showed, acts precisely as an enormous Leyden jar; or rather, Faraday showed that such a cable, without the wire sheathing, would act when submerged as a Leyden jar, the conducting wire acting as