consists for each line of two hollow galvanometer coils i i, with a needle h, pivoted so as to present its poles within their interiors. Four electro-magnets jj jj, which, however, might be permanent magnets, are placed underneath to control the zero of the needle. From one end of this needle a fibre leads to the writing pen, which is a fine glass siphon k, drawing off aniline ink from a small well. This fibre 02, is kept taut by being tied to the cross fibre o1, which is stretched by the spring 0. For the other circuit there is a similar galvanometer arrangement with a thread o., kept taut by the thread o' and the spring o1. Now when the current from each line passes through its proper galvanometer, by the wire shown, the needles move to degrees corresponding to the strength of the currents flowing, and, therefore, approximately to the ranges of the writing pencil at the sending station. By the cross fibres these two motions are combined on the siphon, and the latter inscribes the letters of the message on a moving slip of paper p. This copy may be either larger, smaller, or of the same size as the original writing. Cowper's telegraph will of course be useful under certain circumstances for confidential correspondence. It has been employed in railway work, and operated with success over a line thirty miles long.

III. SUBMARINE TELEGRAPHS.

We come now to the important branch of submarine telegraphy, which has developed so largely since 1851, when the first cable was submerged from Dover to

Calais, that there are now about 100,000 miles of cable in working order, and representing property to the amount of forty millions sterling. A submarine cable consists, like a land line, of three parts-1st, the wire or conductor conveying the current; 2nd, the insulator keeping it on the wire, and 3rd, the outer sheathing which protects it from injury. The conductor is usually a seven-wire strand of the best copper, offering a low resistance to the passage of the current. The insulator, which takes the place of air in the overland line, and keeps the electricity from flowing into the sea or earth, is generally of gutta-percha, applied in a hot molten state to the wire in three coatings. Sometimes, however, indiarubber wound in tapes about the wire is used, especially if the cable is intended for tepid seas. The outer sheathing consists of jute or hemp-yarn and galvanised iron wires. These give strength to the cable in order that it may be safely laid and mended, as well as guarded from damage on the bottom from ships' anchors, wrecks, or shingle.

Owing to the proximity of the copper wire to the sea water and the ground, the effect of "induction" is far more strongly felt upon it than on a land line, and the result is that in telegraphing through it the signals are sensibly delayed. Each signal current "induces " opposite electricity in the sea water around, and there is an attraction between them which holds the signal current back, as it were. This effect is hardly noticeable on long land lines because the telegraph wire is so high above the ground; but in deep-sea cables, or subterranean lines, it reduces the speed of signalling very much, and also makes the separate signal currents "run together." The instruments, therefore, which

operate successfully on long aerial lines are perfectly useless for working long cables. Sir William Thomson was among the first to grasp this fact, and by his invention of the Mirror Galvanometer, which is specially adapted for cables, he rendered ocean telegraphy practicable and the first Atlantic cables a commercial

success.

The submarine circuit is illustrated in Fig. 26,

B

K2

R

FIG. 26.

where L is the cable lying on the sea-bottom, в the sending battery, к the signalling key (which ought to have double-current levers), and E the earth-plate at the sending station. At the receiving station, R is the receiving instrument, and E the earth-plate as before. It will be seen, however, on comparison with Fig. 16, which represents the aerial circuit, that two strange apparatus, c c, are introduced. These are "condensers," and their use for sharpening the signals through a cable was suggested by Mr. C. F. Varley. The telegraph condenser is composed of a series of tinfoil sheets, each separated from the next by a layer of paraffined paper, and all the alternate plates connected to form one pole of the condenser. Thus the 1, 3, 5, etc., plates are joined to one pole a, and the 2, 4, 6,

etc., to the other pole b. Now when this device is inserted in the circuit at each station, the running of the signals together in the cable is less noticeable on the receiver, and sharper, clearer signals are presented to the operator reading the message.

The mirror galvanometer, employed as a receiver on many cables, is represented in Fig. 27, where c is a

H

FIG. 27.

coil of fine insulated wire, surrounding a small magnetic needle hung by a silk fibre, and carrying a tiny mirror attached to it. The details of this needle are shown in the lower figure, where c c are sections through the coil, м is the magnet-needle, carrying in front of it a small mirror. This needle is enclosed in a small chamber, glazed by a lens G, and inserted in the hollow of the coil c. A curving magnet H is supported over the coil to adjust the position of the smaller magnet in the chamber. Now a ray of light from a lamp L in front of the galvanometer is thrown

upon the tiny mirror and reflected back upon a white screen or scale s. The coil c is connected between the end of the conductor of the cable and the earthplate, as in the land circuit; a condenser, however, being usually interposed between the cable and the galvanometer.

Then the signal currents in passing through the coil deflect the tiny magnet hung within it, and the mirror, being carried by the magnet, throws the beam of light off in a different direction. Positive, or "dot," currents are arranged to throw the spot of light toward the left side of the scale; and negative, or "dash," currents throw it to the right side. Thus the wandering of the spot of light on the screen, watchfully followed by the eye of the clerk, is interpreted by him as the message. Letter by letter he spells it out, and a fellow-clerk writes it down word for word.

This receiver, however, like the sounder, has the disadvantage of leaving no permanent record, and Sir William Thomson has therefore introduced his Siphon Recorder on several long cables-for instance, the Eastern Telegraph Company's lines to India, and the AngloAmerican Company's cables across the Atlantic. It is undoubtedly one of the finest inventions of the age, although it is not very well known, because only a few are required, and these are only to be seen abroad, unless one should be specially exhibited. A general view of the instrument is given in Fig. 28, and the leading parts are detailed in Fig. 29. The principle of its action is just the reverse of the mirror galvanometer. In that instrument a tiny magnet moved within a fixed coil of wire; in the siphon recorder a light coil of wire moves between the poles of a powerful magnet. The