at a factory weld, and the end was lost in 82 fathoms of water. The ship then returned to the buoy and tried to underrun the wire, but it soon broke again, and for the moment further attempts were abandoned.

Previous to this two unsuccessful attempts had already been made to connect Great Britain and Ireland by cables made on the lines of the Dover-Calais cable of 1851-one, undertaken by Messrs Newall & Co., between Holyhead and Howth, June 1, 1852, which failed three days after; and the other, a heavy six-wired cable, undertaken by the same firm, between Portpatrick and Donaghadee, October 9, 1852, which broke in a gale after sixteen miles had been paid out.

In June 1854 Messrs Newall recovered the whole of this sixteen miles of cable, and completed the laying to Portpatrick, thus rendering another attempt at a bare wire cable unnecessary, if, indeed, it was still thought desirable.

Mr Dering's faith in the soundness of his views is still unshaken, for he goes on to say: "Instead of a single wire, as in 1853, I should now advocate the use of a bare strand of wires for each of the conductors. And I must add, considering the craving there is at present for Wireless Telegraphs, that it seems to me not altogether improbable that the less ambitious but (for, at all events, long distances) far more feasible plan of using bare wires will yet have its innings." And who, in these days of electrical marvels, will dare to say him nay? I, for my part, will not, for I have seen more unlikely things come to pass. The dream of to-day, "idle and ridiculous as it may seem, has been so often realised on the morrow, that the cautious historian of science must not look for finality in any of its applications.1

1 For recent applications of the bare-wire principle, see Melhuish, p. 111, infra.

JOHN HAWORTH-1862.

On March 27, 1862, Mr Haworth patented "An improved method of conveying electric signals without the intervention of any continuous artificial conductor," in reference to which a lecturer of the period said:1 "I have not met one single gentleman connected with the science of telegraphy who could understand his process, or its probability of success. I applied to him for some information, but he is unwilling to communicate any particulars until experiment has sufficiently demonstrated the practicability of his plans."

In the discussion which followed, Mr Cromwell Varley, electrician of the old Electric and International Telegraph, and the old Atlantic Telegraph, Companies, said: "Being informed that Sir Fitzroy Kelly and the learned chairman (Mr Grove) had seen Haworth's system in operation, and that the latter gentleman was a believer in it, he had tried the experiment upon a very small scale in his own garden, with apparatus constructed according to the instructions of Mr Haworth. His two stations were only 8 yards apart, and, although he used a very sensitive reflecting galvanometer, and twelve cells of Grove's nitric acid battery, he could not get any signals, although the experiments were varied in every conceivable way."

Under these circumstances it will not be surprising if I, too, after a careful study of the specification, and with the light thrown upon it by a further patent of October 30, 1863, have failed to understand the author's method. Indeed, I feel in much the same mental condition towards it as Tristram Shandy's connoisseurs, who, "by long friction, incumbition, and electrical assimilation, have the happiness,

1 T. A. Masey, Society of Arts, January 28, 1863.

at length, to get all be-virtu'd, be-pictured, be-butterflied, and be-fuddled." However, I will do my best to translate the terrible phraseology of the letters patent into plain English; and if after this my readers cannot divine the mode of action I will not blame them-nor must they blame me! My description of the apparatus is based on the complete specification and drawings of the second patent, which were lodged in the Patent Office on April 30, 1864, and which must therefore be supposed to contain the inventor's last word on the subject.

A, Z (fig. 4) are copper and zinc plates respectively, curved as shown, and buried in the earth about 3 feet

apart. The superficies varies according to distance and other circumstances: thus, for distances up to 75 miles plates 1 foot square suffice; over 75 and up to 440 miles, plates 24 by 16 inches are required. G, F are copper cylinders, 24 by 4 inches, buried in earth, which is always moist. At a point distant about 3 feet from the centres of A and z a wooden box J is buried, containing a coil of insulated copper wire, No. 16 gauge, wound upon a wooden reel. The ends of the coil are attached to binding screws shown on top of the box. B is a wooden box containing a wooden reel divided into three compartments, x, y, z (fig. 5). x is filled with fine covered-copper wire, the

ends of which are brought together and secured on the outside of the reel. y is filled with thicker covered-copper wire, wound in the same direction as x, and the ends are severally connected to binding-screws, shown on the outside. z is half filled with insulated iron wire, wound in the same direction as x and y; the ends are fastened together on the outside of the reel as with coil x. The compartment is then filled with more of the same iron wire, wound double, and in the reverse direction to the coil below it. These double wires are not twisted, nor bound together, nor allowed to cross one another, but are wound evenly in layers side by side; and the ends of each coil are secured together on the outside of the reel as in the case of the lower coil, and adjacent thereto. Usually the wire of coil x is No. 32 gauge; y, No. 16; and z, No. 20; but the sizes and quantities required must vary according to distance and other circumstances.

c is any suitable telegraph instrument of the needle pattern.

D is a condenser of a kind which an electrical Dominie Sampson would call prodigious! A wooden box divided lengthwise into two compartments well coated with shellac. In each compartment is placed a band of stout gold-foilboth well insulated, and connected at their ends to the binding-screws a, g, and b, h, respectively (fig. 6). Each compartment is filled with sixty rectangular plates of guttapercha, on which insulated copper wire, No. 32 gauge, is wound in one continuous length from the first plate to the last, and the ends are attached to the binding-screws a, g, and b, h, respectively. "I fix binding-screws c, d, e, f, k, and in the positions shown, and connect them with the wire upon the plates in its passage through the box. I then pass from end to end of each compartment over the plates, and lying on them, but well insulated from them,