Now let us see how the “coherer,” as the filings tube is called, is used in actual wireless telegraphy. Fig. 33a shows a simple arrangement for the purpose. A is an induction coil, and B the battery supplying the current. The coil is fitted with a spark gap, consisting of two highly polished brass balls CC, one of these balls being connected to a vertical wire supported by a pole, and the other to earth. D is a Morse key for starting and stopping the current. When the key is pressed down, current flows from the battery to the coil, and in passing through the coil it is raised to a very high voltage, as described in Chapter VIII. This high tension current is sent into the aerial wire, which quickly becomes charged up to its utmost limits. But more current continues to arrive, and so the electricity in the aerial, unable to bear any longer the enormous pressure, takes the only path of escape and bursts violently across the air gap separating the brass balls. Surging oscillations are then produced in the aerial, the ether is violently disturbed, and electric waves are set in motion. This is the transmitting part of the apparatus.

If a stone is dropped into a pond, little waves are set in motion, and these spread outwards in ever-widening rings. Electric waves also are propagated outwards in widening rings, but instead of travelling in one plane only, like the water waves, they proceed in every plane; and when they arrive at the receiving aerial they set up in it oscillations of the same nature as those which produced the waves. Let us suppose electric waves to reach the aerial wire of Fig. 33b. The resistance of the coherer H is at once lowered so that current from battery N flows and operates the relay F, which closes the circuit of battery M. This battery has a twofold task. It operates the sounder E, and it energizes the electro-magnet of the de-coherer K, as shown by the dotted lines. This de-coherer is simply an electric bell without the gong, arranged so that the hammer strikes the coherer tube; and its purpose is to tap the tube automatically and much more rapidly than is possible by hand. The sounder therefore gives a click, and the de-coherer taps the tube, restoring the resistance of the filings. The circuit of battery N is then broken, and the relay therefore interrupts the circuit of battery M. If waves continue to arrive, the circuits are again closed, another click is given, and again the hammer taps the tube. As long as waves are falling upon the aerial, the alternate makings and breakings of the circuits follow one another very rapidly and the sounder goes on working. When the waves cease, the hammer of the de-coherer has the last word, and the circuits of both batteries remain broken. To confine the electric waves to their proper sphere two coils of wire, LL, called choking coils, are inserted as shown.

In this simple apparatus we have all the really essential features of a wireless installation for short distances. For long distance work various modifications are necessary, but the principle remains exactly the same. In land wireless stations the single vertical aerial wire becomes an elaborate arrangement of wires carried on huge masts and towers. The distance over which signals can be transmitted and received depends to a considerable extent upon the height of the aerial, and consequently land stations have the supporting masts or towers from one to several hundred feet in height, according to the range over which it is desired to work. As a rule the same aerial is used both for transmitting and receiving, but some stations have a separate aerial for each purpose. A good idea of the appearance of commercial aerials for long distance working may be obtained from the frontispiece, which shows the Marconi station at Glace Bay, Nova Scotia, from which wireless communication is held with the Marconi station at Clifden, in Galway, Ireland.

In the first wireless stations what is called a “plain aerial” transmitter was used, and this was almost the same as the transmitting apparatus in Fig. 33a, except, of course, that it was on a larger scale. This arrangement had many serious drawbacks, including that of a very limited range, and it has been abandoned in favour of the “coupled” transmitter, a sketch of which is shown in Fig. 34. In this transmitter there are two separate circuits, having the same rate of oscillation. A is an induction coil, supplied with current from the battery B, and C is a condenser. A condenser is simply an apparatus for storing up charges of electricity. It may take a variety of forms, but in every case it must consist of two conducting layers separated by a non-conducting layer, the latter being called the “dielectric.” The Leyden jar is a condenser, with conducting layers of tinfoil and a dielectric of glass, but the condensers used for wireless purposes generally consist of a number of parallel sheets of metal separated by glass or mica, or in some cases by air only. The induction coil charges up the condenser with high tension electricity, until the pressure becomes so great that the electricity is discharged in the form of a spark between the brass balls of the spark gap D. The accumulated electric energy in the condenser then surges violently backwards and forwards, and by induction corresponding surgings are produced in the aerial circuit, these latter surgings setting up electric waves in the ether.

For the sake of simplicity we have represented the apparatus as using an induction coil, but in all stations of any size the coil is replaced by a step-up transformer, and the current is supplied either from an electric light power station at some town near by, or from a power house specially built for the purpose. Alternating current is generally used, and if the current supplied is continuous, it is converted into alternating current. This may be done by making the continuous current drive an electric motor, which in turn drives a dynamo generating alternating current. In any case, the original current is too low in voltage to be used directly, but in passing through the transformer it is raised to the required high pressure. The transmitting key, which is inserted between the dynamo and the transformer, is specially constructed to prevent the operator from receiving accidental shocks, and the spark gap is enclosed in a sort of sound-proof box, to deaden the miniature thunders of the discharge.

During the time that signals are being transmitted, sparks follow one another across the spark gap in rapid succession, a thousand sparks per second being by no means an uncommon rate. The violence of these rapid discharges raises the brass balls of the gap to a great heat. This has the effect of making the sparking spasmodic and uncertain, with the result that the signals at the receiving station are unsatisfactory. To get over this difficulty Marconi introduced a rotary spark gap. This is a wheel with projecting knobs or studs, mounted on the shaft of the dynamo supplying the current, so that it rotates rapidly. Two stationary knobs are fixed so that the wheel rotates between them, and the sparks are produced between these fixed knobs and those of the wheel, a double spark gap thus being formed. Overheating is prevented by the currents of air set up by the rapid movement of the wheel, and the sparking is always regular.

In the receiving apparatus already described a filings coherer was used to detect the ether waves, and, by means of a local battery, to translate them into audible signals with a sounder, or printed signals with a Morse inker. This coherer however is unsuitable for commercial working. It is not sufficiently sensitive, and it can be used only for comparatively short distances; while its action is so slow that the maximum speed of signalling is not more than about seventeen or eighteen words a minute. A number of different detectors of much greater speed and sensitiveness have been devised. The most reliable of these, though not the most sensitive, is the Marconi magnetic detector, Plate XIII.b. This consists of a moving band made of several soft iron wires twisted together, and passing close to the poles of two horse-shoe magnets. As the band passes from the influence of one magnet to that of the other its magnetism becomes reversed, but the change takes a certain amount of time to complete owing to the fact that the iron has some magnetic retaining power, so that it resists slightly the efforts of one magnet to reverse the effect of the other. The moving band passes through two small coils of wire, one connected with the aerial, and the other with a specially sensitive telephone receiver. When the electric waves from the transmitting station fall upon the aerial of the receiving station, small, rapidly oscillating currents pass through the first coil, and these have the effect of making the band reverse its magnetism instantly. The sudden moving of the lines of magnetic force induces a current in the second coil, and produces a click in the telephone. As long as the waves continue, the clicks follow one another rapidly, and they are broken up into the long and short signals of the Morse code according to the manipulation of the Morse key at the sending station. Except for winding up at intervals the clockwork mechanism which drives the moving band, this detector requires no attention, and it is always ready for work.

Another form of detector makes use of the peculiar power possessed by certain crystals to rectify the oscillatory currents received from the aerial, converting them into uni-directional currents. At every discharge of the condenser at the sending station a number of complete waves, forming what is called a “train” of waves, is set in motion. From each train of waves the crystal detector produces one uni-directional pulsation of current, and this causes a click in the telephone receiver. If these single pulsations follow one another rapidly and regularly, a musical note is heard in the receiver. Various combinations of crystals, and crystals and metal points, are used, but all work in the same way. Some combinations work without assistance, but others require to have a small current passed through them from a local battery. The crystals are held in small cups of brass or copper, mounted so that they can be adjusted by means of set-screws. Crystal detectors are extremely sensitive, but they require very accurate adjustment, and any vibration quickly throws them out of order.

The “electrolytic” detector rectifies the oscillating currents in a different manner. One form consists of a thin platinum wire passing down into a vessel made of lead, and containing a weak solution of sulphuric acid. The two terminals of a battery are connected to the wire and the vessel respectively. As long as no oscillations are received from the aerial the current is unable to flow between the wire and the vessel, but when the oscillations reach the detector the current at once passes, and operates the telephone receiver. The action of this detector is not thoroughly understood, and the way in which the point of the platinum wire prevents the passing of the current until the oscillations arrive from the aerial is something of a mystery.

The last detector that need be described is the Fleming valve receiver. This consists of an electric incandescent lamp, with either carbon or tungsten filament, into which is sealed a plate of platinum connected with a terminal outside the lamp. The plate and the filament do not touch one another, but when the lamp is lighted up a current can be passed from the plate to the filament, but not from filament to plate. This receiver acts in a similar way to the crystal detector, making the oscillating currents into uni-directional currents. It has proved a great success for transatlantic wireless communication between the Marconi stations at Clifden and Glace Bay, and is extensively used.

The electric waves set in motion by the transmitting apparatus of a wireless station spread outwards through the ether in all directions, and so instead of reaching only the aerial of the particular station with which it is desired to communicate, they affect the aerials of all stations within a certain range. So long as only one station is sending messages this causes no trouble; but when, as is actually the case, large numbers of stations are hard at work transmitting different messages at the same time, it is evident that unless something can be done to prevent it, each of these messages will be received at the same moment by every station within range, thus producing a hopeless confusion of signals from which not a single message can be read. Fortunately this chaos can be avoided by what is called “tuning.”

Wireless tuning consists in adjusting the aerial of the receiving station so that it has the same natural rate of oscillation as that of the transmitting station. A simple experiment will make clearer the meaning of this. If we strike a tuning-fork, so that it sounds its note, and while it is sounding strongly place near it another fork of the same pitch and one of a different pitch, we find that the fork of similar pitch also begins to sound faintly, whereas the third fork remains silent. The explanation is that the two forks of similar pitch have the same natural rate of vibration, while the other fork vibrates at a different rate. When the first fork is struck, it vibrates at a certain rate, and sets in motion air waves of a certain length. These waves reach both the other forks, but their effect is different in each case. On reaching the fork of similar pitch the first wave sets it vibrating, but not sufficiently to give out a sound. But following this wave come others, and as the fork has the same rate of vibration as the fork which produced the waves, each wave arrives just at the right moment to add its impulse to that of the preceding wave, so that the effect accumulates and the fork sounds. In the case of the third fork of different pitch, the first wave sets it also vibrating, but as this fork cannot vibrate at the same rate as the one producing the waves, the latter arrive at wrong intervals; and instead of adding together their impulses they interfere with one another, each upsetting the work of the one before it, and the fork does not sound. The same thing may be illustrated with a pendulum. If we give a pendulum a gentle push at intervals corresponding to its natural rate of swing, the effects of all these pushes are added together, and the pendulum is made to swing vigorously. If, on the other hand, we give the pushes at longer or shorter intervals, they will not correspond with the pendulum’s rate of swing, so that while some pushes will help the pendulum, others will hinder it, and the final result will be that the pendulum is brought almost to a standstill, instead of being made to swing strongly and regularly. The same principle holds good with wireless aerials. Any aerial will respond readily to all other aerials having the same rate of oscillation, because the waves in each case are of the same length; that is to say, they follow one another at the same intervals. On the other hand, an aerial will not respond readily to waves from another aerial having a different rate of oscillation, because these do not follow each other at intervals to suit it.

If each station could receive signals only from stations having aerials similar to its own, its usefulness would be very limited, and so all stations are provided with means of altering the rate of oscillation of their aerials. The actual tuning apparatus by which this is accomplished need not be described, as it is complicated, but what happens in practice is this: The operator, wearing telephone receivers fixed over his ears by means of a head band, sits at a desk upon which are placed his various instruments. He adjusts the tuning apparatus to a position in which signals from stations of widely different wave-lengths are received fairly well, and keeps a general look out over passing signals. Presently he hears his own call-signal, and knows that some station wishes to communicate with him. Immediately he alters the adjustment of his tuner until his aerial responds freely to the waves from this station, but not to waves from other stations, and in this way he is able to cut out signals from other stations and to listen to the message without interruption.

Unfortunately wireless tuning is yet far from perfect in certain respects. For instance, if two stations are transmitting at the same time on the same wave-length, it is clearly impossible for a receiving operator to cut one out by wave-tuning, and to listen to the other only. In such a case, however, it generally happens that although the wave-frequency is the same, the frequency of the wave groups or trains is different, so that there is a difference in the notes heard in the telephones; and a skilful operator can distinguish between the two sufficiently well to read whichever message is intended for him. The stations which produce a clear, medium-pitched note are the easiest to receive from, and in many cases it is possible to identify a station at once by its characteristic note. Tuning is also unable to prevent signals from a powerful station close at hand from swamping to some extent signals from another station at a great distance, the nearer station making the receiving aerial respond to it as it were by brute force, tuning or no tuning.

Another source of trouble lies in interference by atmospheric electricity. Thunderstorms, especially in the tropics, interfere greatly with the reception of signals, the lightning discharges giving rise to violent, irregular groups of waves which produce loud noises in the telephones. There are also silent electrical disturbances in the atmosphere, and these too produce less strong but equally weird effects. Atmospheric discharges are very irregular, without any real wave-length, so that an operator cannot cut them out by wave-tuning pure and simple in the way just described, as they defy him by affecting equally all adjustments. Fortunately, the irregularity of the atmospherics produces correspondingly irregular sounds in the telephones, quite unlike the clear steady note of a wireless station; and unless the atmospherics are unusually strong this note pierces through them, so that the signals can be read. The effects of lightning discharges are too violent to be got rid of satisfactorily, and practically all that can be done is to reduce the loudness of the noises in the telephones, so that the operator is not temporarily deafened. During violent storms in the near neighbourhood of a station it is usual to connect the aerial directly to earth, so that in the event of its being struck by a flash the electricity passes harmlessly away, instead of injuring the instruments, and possibly also the operators. Marconi stations are always fitted with lightning-arresters.

The methods and apparatus we have described so far are those of the Marconi system, and although in practice additional complicated and delicate pieces of apparatus are used, the description given represents the main features of the system. Although Marconi was not the discoverer of the principles of wireless telegraphy, he was the first to produce a practical working system. In 1896 Marconi came from Italy to England, bringing with him his apparatus, and after a number of successful demonstrations of its working, he succeeded in convincing even the most sceptical experts that his system was thoroughly sound. Commencing with a distance of about 100 yards, Marconi rapidly increased the range of his experiments, and by the end of 1897 he succeeded in transmitting signals from Alum Bay, in the Isle of Wight, to a steamer 18 miles away. In 1899 messages were exchanged between British warships 85 miles apart, and the crowning achievement was reached in 1901, when Marconi received readable signals at St. John’s, Newfoundland, from Poldhu in Cornwall, a distance of about 1800 miles. In 1907 the Marconi stations at Clifden and Glace Bay were opened for public service, and by the following year transatlantic wireless communication was in full swing. The sending of wireless signals across the Atlantic was a remarkable accomplishment, but it did not represent by any means the limits of the system, as was shown in 1910. In that year Marconi sailed for Buenos Ayres, and wireless communication with Clifden was maintained up to the almost incredible distance of 4000 miles by day, and 6735 miles by night. The Marconi system has had many formidable rivals, but it still holds the proud position of the most successful commercial wireless system in the world.

We have not space to give a description of the other commercial systems, but a few words on some of the chief points in which they differ from the Marconi system may be of interest. We have seen that an ordinary spark gap, formed by two metal balls a short distance apart, becomes overheated by the rapid succession of discharges, with the result that the sparking is irregular. What actually happens is that the violent discharge tears off and vaporizes minute particles of the metal. This intensely heated vapour forms a conducting path which the current is able to cross, so that an arc is produced just in the same way as in the arc lamp. This arc is liable to be formed by each discharge, and it lasts long enough to prevent the sparks from following one another promptly. In the Marconi system this trouble is avoided by means of a rotating spark gap, but in the German “Telefunken” system, so named from Greek tele, far off, and German Funke, a spark, a fixed compound spark gap is used for the same purpose. This consists of a row of metal discs about 1/100 inch apart, and the spark leaps these tiny gaps one after the other. The discs are about 3 inches in diameter, and their effect is to conduct away quickly the heat of the discharge. By this means the formation of an arc is prevented, and the effect of each discharge is over immediately, the sparks being said to be “quenched.” The short discharges enable more energy to be radiated from the aerial into the ether, and very high rates of sparking are obtained, producing a high-pitched musical note.

The “Lepel” system also uses a quenched spark. The gap consists of two metal discs clamped together and separated only by a sheet of paper. The paper has a hole through its centre, and through this hole the discharge takes place, the discs being kept cool by water in constant circulation. The discharge is much less noisy than in the Marconi and Telefunken systems, and the musical note produced is so sensitive that by varying the adjustments simple tunes can be played, and these can be heard quite distinctly in the telephone at the receiving stations.

In the three systems already mentioned spark discharges are used to set up oscillatory currents in the aerial, which in turn set up waves in the ether. Each discharge sets in motion a certain number of waves, forming what is known as a train of waves. The discharges follow one another very rapidly, but still there is a minute interval between them, and consequently there is a corresponding interval between the wave-trains. In the “Goldschmidt” system the waves are not sent out in groups of this kind, but in one long continuous stream. The oscillatory currents which produce ether waves are really alternating currents which flow backwards and forwards at an enormous speed. The alternating current produced at an ordinary power station is of no use for wireless purposes, because its “frequency,” or rate of flow backwards and forwards, is far too low. It has been found possible however to construct special dynamos capable of producing alternating current of the necessary high frequency, and such dynamos are used in the Goldschmidt system. The dynamos are connected directly to the aerial, so that the oscillatory currents in the latter are continuous, and the ether waves produced are continuous also.

The “Poulsen” system produces continuous waves in an altogether different manner, by means of the electric arc. The arc is formed between a fixed copper electrode and a carbon electrode kept in constant rotation, and it is enclosed in a kind of box filled with methylated spirit vapour, hydrogen, or coal gas. A powerful electro-magnet is placed close to the arc, so that the latter is surrounded by a strong magnetic field. Connected with the terminals of the arc is a circuit consisting of a condenser and a coil of wire, and the arc sets up in this circuit oscillatory currents which are communicated to the aerial. These currents are continuous, and so also are the resulting waves.

The method of signalling employed in these two continuous-wave systems is quite different from that used in the Marconi and other spark systems. It is practically impossible to signal by starting and stopping the alternating-current dynamos or the arc at long or short intervals to represent dashes or dots. In one case the sudden changes from full load to zero would cause the dynamo to vary its speed, and consequently the wave-length would be irregular; besides which the dynamo would be injured by the sudden strains. In the other case it would be extremely difficult to persuade the arc to start promptly each time. On this account the dynamo and the arc are kept going continuously while a message is being transmitted, and the signals are given by altering the wave-length. In other words, the transmitting aerial is thrown in and out of tune alternately at the required long or short intervals, and the receiving aerial responds only during the “in-tune” intervals.

The various receiving detectors previously described are arranged to work with dis-continuous waves, producing a separate current impulse from each group or train of waves. In continuous wave systems there are of course no separate groups, and for this reason these detectors are of no use, and a different arrangement is required. The oscillatory currents set up in the aerial are allowed to charge up a condenser, and this condenser is automatically disconnected from the aerial and connected to the operator’s telephones at regular intervals of about 1/1000 second. Each time the condenser is connected to the telephones it is discharged, and a click is produced. These clicks continue only as long as the waves are striking the aerial, and as the transmitting operator interrupts the waves at long or short intervals the clicks are split up into groups of corresponding length.

Continuous waves have certain advantages over dis-continuous waves, particularly in the matter of sharp tuning, but these advantages are outweighed to a large extent by weak points in the transmitting apparatus. The dynamos used to produce the high-frequency currents in the Goldschmidt system are very expensive to construct and troublesome to keep in order; while in the Poulsen system the arc is difficult to keep going for long periods, and it is liable to fluctuations which greatly affect its working power. Although all the commercial Marconi installations make use of dis-continuous waves exclusively, Mr. Marconi is still carrying out experiments with continuous waves.

There are many points in wireless telegraphy yet to be explained satisfactorily. Our knowledge of the electric ether waves is still limited, and we do not know for certain how these waves travel from place to place, or exactly what happens to them on their journeys. For this reason we are unable to give a satisfactory explanation of the curious fact that, generally speaking, it is easier to signal over long distances at night than during the day. Still more peculiar is the fact that it is easier to signal in a north and south direction than in an east and west direction. There are also remarkable variations in the strength of the signals at certain times, particularly about sunset and sunrise. Every station has a certain normal range which does not vary much as a rule, but at odd times astonishing “freak” distances are covered, stations having for a short time ranges far beyond their usual limits. These and other problems are being attacked by many investigators, and no doubt before very long they will be solved. Wireless telegraphy already has reached remarkable perfection, but it is still a young science, and we may confidently expect developments which will enable us to send messages with all speed across vast gulfs of distance at present unconquered.

Wireless telephony, like wireless telegraphy, makes use of electric waves set up in, and transmitted through the ether. The apparatus is practically the same in each case, except in one or two important points. In wireless telegraphy either continuous or dis-continuous waves may be used, and in the latter case the spark-frequency may be as low as twenty-five per second. On the other hand, wireless telephony requires waves which are either continuous, or if dis-continuous, produced by a spark-frequency of not less than 20,000 per second. In other words, the frequency of the wave trains must be well above the limits of audibility. Although dis-continuous waves of a frequency of from 20,000 to 40,000 or more per second can be used, it has been found more convenient to use absolutely continuous waves for wireless telephony, and these may be produced by the Marconi disc generator, by the Goldschmidt alternator, or by the Poulsen arc, the last named being largely employed.

In wireless telegraphy the wave trains are split up by a transmitting key so as to form groups of signals; but in telephony the waves are not interrupted at all, but are simply varied in intensity by means of the voice. All telephony, wireless or otherwise, depends upon the production of variations in the strength of a current of electricity, these variations being produced by air vibrations set up in speaking. In ordinary telephony with connecting wires the current variations are produced by means of a microphone in the transmitter, and in wireless telephony the same principle is adopted. Here comes in the outstanding difficulty in wireless transmission of speech. The currents used in ordinary telephony are small, and the microphone works with them quite satisfactorily; but in wireless telephony very heavy currents have to be employed, and so far no microphone has proved capable of dealing effectively with these currents. Countless devices to assist the microphone have been tried. It was found that one of the causes of trouble was the overheating of the carbon granules, which caused them to stick together, so becoming insensitive. To remedy this the granules have been cooled in various ways, by water, gas, or oil, but although the results have been improved, still the microphones worked far from perfectly. Improved results have been obtained also by connecting a number of microphones in parallel. The microphone difficulty is holding back the development of wireless telephony, and at present no satisfactory solution of the problem is in sight.

The transmitting and receiving aerials are the same as in wireless telegraphy, and like them are tuned to the same frequency. The receiving apparatus too is of the ordinary wireless type, with telephones and electrolytic or other detectors.

Wireless telephony has been used with considerable success in various German collieries, and at the Dinnington Main Colliery, Yorkshire. Early last year Marconi succeeded in establishing communication by wireless telephony between Bournemouth and Chelmsford, which are about 100 miles apart; and about the same time a song sung at Laeken, in Belgium, was heard clearly at the Eiffel Tower, Paris, a distance of 225 miles. The German Telefunken Company have communicated by wireless telephony between Berlin and Vienna, 375 miles, and speech has been transmitted from Rome to Tripoli, a total distance of more than 600 miles. These distances are of course comparatively small, but if the microphone trouble can be overcome satisfactorily, transatlantic wireless telephony appears to be quite possible.