striking results had been obtained on actual telegraph lines, where there was no battery, and where the infinitesimal currents produced by speaking into a Bell telephone on one wire were able to induce currents in a parallel wire sufficient to render the words audible in another telephone in its circuit. Dr Channing found this to be possible "under very favourable conditions.” 1

Another striking illustration is furnished by Prof. Blake, of Brown University, U.S., who talked with a friend for some distance along a railway (using the two lines of rails for the telephonic circuit), hearing at the same time the Morse signals passing along the telegraph wires overhead.2

PROFESSOR JOHN TROWBRIDGE-1880.

Such are a few of the early instances noted of the extreme sensitiveness of the telephone, by the aid of which the problem of wireless telegraphy was now to be attacked with a fair measure of success, and advanced a long way towards a practical solution.

Mr J. Gott, then superintendent of the Anglo-American Telegraph Company at St Pierre, was, I believe, the first to suggest the employment of the telephone in this connection. In a brief communication, published in the 'Jour. Inst. Elec. Engs.' (vol. vi. p. 523), he says: "The island of

1 For a curiously similar case, the result of a wrong connection of the line wires, see the 'Telegraphic Journal,' vol. ix. p. 68.

2 The absence of insulation in this experiment recalls the fact that a telephone line using the earth for the return circuit often works better when the insulation is defective, as it is then less affected by extraneous currents. Thus, in 1882, the Evansville (Ind.) Telephone Exchange Company worked 400 miles of line without insulators of any kind (the wires being simply attached to the poles), and generally with better results than when insulators were used. (Electrician,' vol. ix. p. 481.)

St Pierre is, perhaps, better insulated than most places. Hundreds of yards from the station, if a wire be connected to earth, run some distance, and put to earth again, with a telephone in circuit, the signals passing through the cables can be heard."

There are two offices on the island,-one used for repeating the cable business on the short cables between Sydney, C.B., and Placentia, N.F., and operated by the Morse system, with a comparatively powerful battery; the other is the office at which the Brest and Duxbury cables terminate, and is furnished with very delicate instruments —the Brest cable, which is upwards of 2500 miles long, being operated by Thomson's exceedingly sensitive deadbeat mirror galvanometer; whilst on the Duxbury cable the same inventor's instrument, the siphon recorder, is used. The Brest instrument was found seriously affected by earthcurrents, which flowed in and out of the cable, interfering very much with the true currents or signals, and rendering it a difficult task for the operator to decipher them accurately. The phenomenon is not an uncommon one; and the cause being attributed to the ground used at the office, a spare insulated wire, laid across the island, a distance of nearly three miles, and a metal plate connected to it and placed in the sea, was used in lieu of the office ground. This had a good effect, but it was now found that part of the supposed earth-currents had been due to the signals sent by the Morse operator into his wire, for when the recorder was put in circuit between the ground at the cable office and the sea ground-three miles distant—the messages sent by the Morse were clearly indicated, clearly, in fact, that they were automatically recorded on the tape.

It must be clearly understood that the two offices were in no way connected, nor were they within some 200 yards of

F

each other; and yet messages sent at one office were distinctly read at the other, the only connection between the two being through the earth, and it is quite evident that they could be so read simultaneously at many offices in the same neighbourhood. The explanation is clear enough. The potential of the ground at the two offices is alternately raised and lowered by the Morse battery. The potential of the sea remains almost, if not wholly, unaffected by these, and the island thus acts like an immense Leyden jar, continually charged by the Morse battery and discharged, in part, through the short insulated line. Each time the Morse operator depressed his key he not only sent a current into his cable, but electrified the whole island, and this electrification was detected and indicated on the recorder.1

As the result of these experiences, Mr Gott gave it as his opinion that "speaking through considerable distances of earth without wires is certainly possible with Bell's telephone, with a battery and Morse signals."

Professor John Trowbridge of Harvard University, America, was, however, the first to systematically study the problem, and to revive the daring project of an Atlantic telegraph without connecting wires, and the less ambitious but equally useful project of intercommunication between ships at sea.2 In fact, Trowbridge's researches may truly be

1 See now Salvá's curious anticipation in 1795 of this phenomenon, p. 2, ante. The peculiarity, due to geological formation, is not confined to St Pierre; it is often met with in practice, though usually in lesser degrees. See some interesting cases, noted by G. K. Winter and James Graves, Jour. Inst. Elec. Engs.,' vol. i. p. 88, and vol. iv. p. 34.

2 Mr H. C. Strong of Chicago, Illinois, claims to have suggested in 1857, in a Peoria, Ill., newspaper, the possibility of communication between ships at sea by means of a wireless telegraph then recently invented by his friend Henry Nelson of Galesburg. See Mr Strong's letter in the New York 'Journal of the Telegraph,' August 15,

said to form a new starting-point in the history of our subject, for, as we shall see later on, it is chiefly to him that Messrs Preece, Bell, and probably other experimenters in this field, owe their inspirations. His investigations,

therefore, deserve to be carefully followed.

The observatory at Harvard transmits time-signals from Cambridge to Boston, a distance of about four miles, and the regular recurrence of the beats of the clock afforded a good means of studying the spreading of the electric. currents from the terminal of the battery which is grounded at the observatory. In all the telephone circuits between. Boston and Cambridge, in the neighbourhood of the observatory line, the ticking of the clock could be heard. This ticking had been attributed to induction, but this, according to Prof. Trowbridge, is an erroneous conclusion, as he shows by a mathematical analysis into which we need not enter. The result goes to show that, with telephones of the resistance usually employed, no inductive effect will be perceived by the use of even ten quart Bunsen cells between wires running parallel, a foot apart, for a distance of 30 or 40 feet.

For this and other reasons, he says, it is impossible to hear telephonic messages by induction from one wire to another, unless the two run parallel and very close to each other for a long distance. This distance generally exceeds the limit at which the ordinary Bell telephone ceases to transmit articulate speech. The effects which have usually been attributed to induction are really, he says, due to the earth connections and to imperfect insulation.

Having determined in this manner that the echoes of the

1 See pp. 92 and 137, infra. Professor Trowbridge's researches are given at length in a paper, "The Earth as a Conductor of Electricity," read before the American Academy of Arts and Sciences in 1880. See also 'Silliman's American Journal of Science,' August 1880, which I follow in the text.

time-signals observed on the telephone lines were not due to induction, but to leakage from the clock circuit, Prof. Trowbridge proceeded to study the extent of the equally electrified or equi-potential surfaces of the ground surrounding the clock battery. His method of exploration was to run a wire 500 or 600 feet long to earth at each end, including a telephone of 50 to 60 ohms resistance. Evidence of a current in this exploratory circuit was plainly shown by the ticking sound which making and breaking the circuit caused in the telephone, and the time-signals could be distinctly heard in a field 220 yards from the observatory where one earth of the time-signal wire is located. At a distance of a mile from the observatory, and not in the direct line between that place and the Boston telephone office, the time-signals were heard by connecting through a telephone the gas-pipes of one building with the water-pipes of another only 50 feet apart. In another experiment at the Fresh Pond lake in Cambridge, signals sent from Boston to Waltham (ten to twelve miles) were heard by simply dipping the terminal wires of the telephone in the lake, and some distance apart, where they must have been far away (? four miles) from the battery earth.

Prof. Trowbridge performed a large number of similar experiments, varied in every way, all going to prove (1) that a battery terminal discharging electricity to earth is the centre of waves of electrical energy, ever widening, and ever decreasing in strength or potential as they widen; and (2) that on tapping the earth in the way described at two points of different potentials (not very distant, if near the central source, and more removed the farther we recede from the source) we can obtain in the telephone evidence of their existence. Prof. Trowbridge then goes on to say:

"In a discussion on the earth as a conductor, Steinheil