The connections for an experimental syntonic wireless transmitting station are shown in the diagram Fig. 4. A is the aerial; T, the inductance; E, earth; L, hot-wire ammeter. The closed oscillatory circuit consists of an inductance F, spark-gap G, and a block condenser C. H is a spark-coil for supplying the energy, the secondary J being connected to the spark-gap. A mercury break N and a battery B are placed in the primary circuit of the coil. The Morse key K is for completing the battery circuit for signalling purposes. When the key K is depressed, the battery circuit is completed, and a spark passes between the balls of the spark-gap G producing oscillations in the closed circuit, which are transposed to the aerial circuit by induction. For signalling purposes it is only necessary for the operator by means of the key K to send out a long or short train of waves in some pre-arranged order, to enable the operator at the receiving station to understand the message that is being transmitted.

If a photograph could be prepared in such a manner that it would serve the purpose of the key K, and could so arrange matters that a minute portion of the photograph could be transmitted separately but in succession, and that each portion of the photograph having the same density could be given the same signal, then it would only be necessary to have apparatus at the receiving station capable of arranging the signals in proper sequence (each signal recorded being the same size and having the same density as the transmitted portion of the photograph) in order to receive a facsimile of the picture transmitted.

The following method of preparing the photograph^[3] is one that has been adopted in several systems of photo-telegraphy, and is the only one at all suitable for wireless transmission. The photograph or picture which is to be transmitted is fastened out perfectly flat upon a copying-board. A strong light is placed on either side of this copying board, and is concentrated upon the picture by means of reflectors. The camera which is used for copying has a single line screen interposed between the lens and sensitised plate, and the effect of this screen is to break the picture up into parallel lines. Thus a white portion of the photograph would consist of very narrow lines wide apart, while the dark portion would be made up of wide lines close together; a black part would appear solid and show no lines at all. From this line negative it will be necessary to take off a print upon a specially prepared sheet of metal. This consists of a sheet of thick lead- or tinfoil, coated upon one side with a thin film of glue to which bichromate of potash has been added; the bichromate possessing the property of rendering the glue waterproof when acted upon by light. The print can be taken off by artificial light (arc lamps being generally used), but the exact time to allow for printing can only be found by experiment, as it varies considerably according to the thickness of the film. The printing finished, the metal print is washed under running water, when all those parts not acted upon by light, i.e. the parts between the lines, are washed away, leaving the bare metal. We have now an image composed of numerous bands of insulating material (each band varying in width according to the density of the photograph at any point from which it is prepared) attached to a metal base, so that each band of insulating material is separated by a band of conducting material. It is, of course, obvious that the lines on the print cannot be wider apart, centre to centre, than the lines of the screen used in preparing it. A good screen to use is one having 50 lines to the inch, but one is perhaps more suitable for experimental work a little coarser, say 35 lines to the inch. To use a screen having 50 or more lines to the inch, the transmitting apparatus, as will be evident later on, will require to be very nearly perfect.

Before proceeding further it will perhaps be as well to make an experiment. If we take one of the metal prints or, more simple, draw a sketch in insulating ink upon a sheet of metal A, Fig. 5, and connect a battery B and the galvanometer D as shown, we shall find on drawing the free end of the wire across the metal plate that all the time the wire is in contact with the lines of insulating material the needle of the galvanometer will remain at zero, but where it is in contact with the metal plate the needle is deflected.

From this experiment it will be seen that we have in our metal line print, which consists of alternate lines of insulating and conducting material, a method by which an electric circuit can be very easily made and broken. It is, of course, necessary to have some arrangement whereby the whole of the surface of the metal print is utilised for this purpose to the best advantage. One type of transmitting machine used for this purpose is represented by the diagram, Fig. 6. The cylinder A is fastened to the steel shaft B, which runs in the two bearings D and D', the bearing D' having an internal thread corresponding to that on the shaft. The stylus in this class of machine is a fixture, the cylinder being given a lateral as well as a revolving movement. As it is impossible to use a rigid drive, a flexible coupling F is employed between the shaft B and the motor.

Another type of machine is shown in Fig. 7. The drum in this case is stationary, the table T moving laterally by reason of the screwed shaft Fig. 7.Fig. 7. and half nut F. The table, shown separate in Fig. 8, carries a stiff brass spring A, to which is attached a holder B made to take a hardened steel point. The holder is provided with a set screw P for securing the steel point Z. The spring and needle are insulated from the rest of the machine, as shown in the drawing. In working, the metal print is wrapped tightly round the cylinder of the machine, the glue image being, of course, uppermost. To fasten the print a little seccotine should be applied to one edge, and the joint carefully smoothed down with the fingers. Fig. 8.Fig. 8. If there is any tendency on the part of the print to slip round on the drum, a couple of small spring clips placed over the ends of the drum will act as a preventive. It is necessary to place the print upon the drum in such a manner that the stylus draws away from the edge of the lap and not towards it, and the metal prints should be of such a size that when placed round the drum of the machine a lap of about ³/₁₆ths of an inch is allowed.

The steel point Z (ordinary gramophone needles may be used and will be found to answer the purpose admirably) is made to press lightly upon the metal print, and while the pressure should be sufficient to make good electrical contact, it should not be sufficient to cause the needle to scratch the surface of the foil. The pressure is regulated by means of the milled nut H. The electrical connections are given in Fig. 9. One wire from the battery M is taken to the terminal T, and the other wires from M and F lead to the relay R. The current flows from the battery M through the spring Y, through the drum and metal print, the stylus Z, spring A, down to the relay R, and from R back to the battery M. As the drum carrying the single line half-tone print is revolved, the stylus, by reason of the lateral movement given to the table or cylinder as the case may be, will trace a spiral path over the entire surface of the print. As the stylus traces over a conducting strip the circuit is completed, and the tongue of the relay R is attracted, making contact with the stop S. On passing over a strip of insulation the circuit is broken and the tongue of the relay R returns to its normal position.

As already stated, the conducting and insulating bands on the print vary in width according to the density of the photograph from which it is prepared, so that the length of time that the tongue of the relay R is held against the stop S, is in proportion to the width of the conducting strip which is passing under the stylus at any instant. The function of the transmitter is therefore to send to the relay R an intermittent current of varying duration.

The two photographs Figs. 10 and 10a are of a machine designed and used by the writer in his experiments. In this machine the drum is 3.5 inches long and 1.5 inches in diameter. The lead screw has 30 threads to the inch, and the reduction between it and the drum is 3:1, so that the table has a movement of ¹/₉₀th inch per revolution of the drum.

From the brief description of the various types of machines that have been given it will be apparent that in the design of the machine proper there is nothing very complicated, although the addition of the driving and synchronising apparatus complicates matters rather considerably. The questions of driving and synchronising the machines at the two stations is fully dealt with in Chapter IV.

Although the design of the machines is rather simple great attention must be paid both to accuracy of construction and accuracy of working, and this applies, not only to the machines (whether for transmitting or receiving) but for all the various pieces of apparatus that are used. Too much care cannot be bestowed upon this point, as in the wireless transmission of photographs there is a large number of instruments all requiring careful adjustment, and which have to work together in perfect unison at a high speed.

The machine shown in Figs. 10 and 10a was designed and used by the writer solely for experimental work. It will be noticed in the description given in the appendix of the method of preparing the metal prints that a 5" × 4" camera is recommended, while the machine, Fig. 10, is designed to take a print procured from a quarter-plate negative. This size of drum was adopted for several reasons, and although it will be found quite large enough for general experimental work the writer has come to the conclusion that for practical commercial work a drum to take a print 5" × 4" will give better results.

In making a negative of a picture that is required for reproduction purposes, the line screen in the camera is replaced by a "cross screen," i.e. two single line screens placed with their lines at an angle of 90° to one another, and this breaks the image up into small squares instead of lines. By looking at any ordinary newspaper or book illustration through a powerful magnifying glass the effects of a cross screen will readily be seen. With a cross screen a certain amount of detail is necessarily lost, but with a single line screen the amount lost is much greater. If there is any very small detail in the picture most of this would be lost in a coarse screen, hence the necessity of employing as fine a line screen as practicable in order to get as much detail in as possible. It is mainly on this account that a 5" × 4" print is recommended, as, if fairly bold subjects are used for copying, the small detail (this is, of course, a very vague and indefinable term) will not be too fine, and the time required for transmitting reasonable. For obvious reasons it is a great advantage to put the print under pressure to cause the glue image to sink into the soft metal base and leave a perfectly flat and smooth surface. It is essential that the bands on the print lie along the axis of the cylinder, so that the stylus traces its path across them, and not with them.

We have now an arrangement that is capable of taking the place of the key K, Fig. 4, and the diagram, Fig. 11, gives the connections for the complete transmitter. A is the aerial, E earth, T inductance, L ammeter. The closed oscillatory circuit consists of a spark-gap G, inductance F, Fig. 11.Fig. 11. and a condenser C. The secondary J of the coil H is connected to the spark-gap, and the primary P is in circuit with the mercury break N, the battery B, and the local contacts of the relay R. The action is as follows. When contact is made between the stylus Z and the drum V by means of the conducting bands on the line print, the circuit of the relay R and the battery M is completed. The closing of the local circuit of the relay R actuates the second relay R', allowing the primary circuit of the coil H to be closed. As soon as the primary circuit of the coil is completed sparks pass between the electrodes of the spark-gap G, causing waves to radiate from the aerial. The duration of the wave-trains radiated depends upon the duration of contact made by the relays R and R', and this in turn depends upon the width of the conducting strip that is passing under the stylus. The battery M should be about 4 volts, and the battery D about 2 volts. The two-way switch X is connected up so that the relay R' can be thrown out and the key K switched in for ordinary signalling purposes. If any sparking takes place at the point of the stylus, a small condenser C' (about 1 microfarad capacity) should be connected as shown. In the present instance the condenser should be used more as a preventive than as a cure, as in all probability the voltage from M will not be sufficient to cause destructive (if any) sparking; but, as most wireless workers know, anything in the nature of a spark occurring in the neighbourhood of a detector (this, of course, only applies when the receiving apparatus is placed in close proximity to the transmitter) is liable to destroy the adjustment.

In transmitting over ordinary conductors where the initial voltage is fairly high and the self-induction of the circuit very great, the use of the condenser will be found to be absolutely essential. It has also been noted that the angle which the stylus presents to the drum has a marked effect upon the sparking, an angle of about 60° being found to give very good results.

If the size of the single line print used is 5 inches by 4 inches, and a screen having 50 lines to the inch is used for preparing it, then the stylus will have to make 250 contacts during one revolution of the drum. Assuming the drum to make one revolution in three seconds, then the time taken to transmit the complete photograph can be found from the equation T = w × t × s, where w is the width of the print, t the travel of the stylus during one revolution of the drum, and s the time required for one revolution of the drum. In the present instance this will be T = 4 × 90 × 3 = 1080 seconds = 18 minutes. The number of contacts made by the stylus per minute is 5000, and in working at this speed the first difficulty is encountered in the use of the two relays. The relay R is lightly built, and capable of working at a fairly high speed, but R' is a heavier pattern, and consequently works at a slightly lower rate. This relay must necessarily be heavier, as more substantial contacts are needed in order to pass the heavy current taken by the spark-coil.

Relays sensitive and accurate enough to work at this speed will in all probability be beyond the reach of the majority of workers, but there are several types of relays on the market very reasonable in price that will answer very well for experimental work, although the speed of working will no doubt be slower.

For the best results the duration of the wave-trains sent out should be of the same duration as the contact made by R, and therefore equal to the time taken by the stylus to trace over a conducting strip; but if the duration of the contact made by R is t, then that made by R' and consequently the duration of the groups of wave-trains would be t - v where v equals the extra time required by R' to complete its local circuit. The difference in time made by the two relays, although very slight, will be found to affect very considerably the quality of the received pictures. Renewing the platinum contacts is also a great expense, as they are soon burnt out where a heavy current is passed. If the distance experimented over is short so that the power required to operate the spark-coil is not very heavy, one relay will be sufficient providing the contacts are massive enough to carry the current safely. It is useless to expect any of the ordinary relays in general use to work satisfactorily at such a high speed, and in order to compensate for this we must either increase the time of transmitting, or, as already suggested, make use of a coarser line screen in preparing the photographs.

For reasons already explained, all points of make and break should be shunted by a condenser. The effective working speed of an ordinary type of relay may be anything from 1000 to 2500 dots a minute, depending upon accuracy of design and construction.

In the wireless transmission of photographs it is absolutely essential to use some form of rotary spark-gap, as where sparks are passed in rapid succession the ordinary type of gap is worse than useless. When a spark passes between the electrodes of an ordinary spark-gap, Fig. 12, we find that for a fraction of a second after the first spark has passed, the normally high resistance of the gap has been lowered to less than one ohm. If the column of hot gas which constitutes the spark is not instantly dispersed, but remains between the electrodes, it will provide an easy path for any further discharges, and if sparks are passed at all rapidly, what was at first a disruptive and oscillatory discharge will degenerate into a hot, non-oscillatory arc.^[4]

Two forms of rotating spark-gaps are shown in Figs. 13 and 14, and are known as "synchronous" and "non-synchronous" gaps respectively. In the synchronous gap the cog-wheel is mounted on the shaft of the alternator, and a cog comes opposite the fixed electrode when the maximum of potential is reached in the condenser, thus ensuring a discharge at every alternation of current. With this type of gap a spark of pure tone is obtained which Fig. 13.Fig. 13. Fig. 14.Fig. 14. is of great value where the signals are received by means of a telephone, but where the signals are to be mechanically recorded the tone of the spark is of little consequence. In a non-synchronous gap a separate motor is used for driving the toothed wheel, and can either be mounted on the motor shaft or driven by means of a band, there being no regard given to synchronism with the alternator. The fixed electrode is best made long enough to cover about two of the teeth, as this ensures regular sparking and a uniform sparking distance; the spark length is double the length of the spark-gap. The toothed wheel should revolve at a high speed, anything from 5000 to 8000 revolutions per minute, or even more being required. The shaft of the toothed wheel is preferably mounted in ball-bearings.

Owing to the large number of sparks that are required per minute in order to transmit a photograph at even an ordinary speed, it is necessary that the contact breaker be capable of working at a very high speed indeed. The best break to use is what is known as a "mercury jet" interrupter, the frequency of the interruptions being in some cases as high as 70,000 per second. No description of these breaks will be given, as the working of them is generally well understood.

In some cases an alternator is used in place of the battery B, Fig. 4, and when this is done the break M can be dispensed with. In larger stations the coil H is replaced with a special transformer.

The writer has designed an improved relay which will respond to currents lasting only ¹/₁₀₀th part of a second, and capable of dealing with rather large currents in the local circuit.^[5] This relay has not yet been tried, but if it is successful the two relays R and R' can be dispensed with, and the result will be more accurate and effective transmission.

The connections for a complete experimental station, transmitting and receiving apparatus combined, are given in Fig. 15. The terminals W, W are for connecting to the photo-telegraphic receiving apparatus Q, being a double pole two-way switch for throwing either the transmitting or receiving apparatus in circuit. There is another system of transmitting devised by Professor Korn, which employs an entirely different method from the foregoing. By using the apparatus just described, the waves generated are what are known as "damped waves," and by using these damped waves, tuning, which is so essential to good commercial working, can be made to reach a fairly high degree of efficiency.

The question of damped versus undamped waves is a somewhat burning one, and no attempt will be made here to deal with the merits or demerits of the claims made for the respective systems. A series of articles describing the production of undamped waves and their efficiency in working compared with damped waves will be found in the Wireless World, Nos. 3 and 4, 1913, and are well worth reading by any one interested in the subject.

A diagrammatic representation of the apparatus as arranged by Professor Korn is given in Fig. 16. The undamped or "continuous" waves are generated by means of a high-frequency alternator or Poulsen arc. In Fig. 16, X is the generator, F inductance, C condenser; the aerial inductance T is connected by the aerial A and earth E. By this means the waves are tuned to a certain period. A metal print, similar to what has already been described, is wrapped round the drum D of the machine, and when the stylus Z traces over an insulating strip the waves generated are in tune with the receiving station, but when it traces over a conducting strip, a portion of the inductance T is short-circuited, the period of the oscillations is altered, and the two stations are thrown out of tune.

The receiving station is provided with an aperiodic circuit, which consists of an inductance F', condenser C', and a thermodetector N. A string galvanometer H (described in Chapter III.), and the self-induction coils B, B' are connected as shown, the coils B, B' preventing the high-frequency currents, which change their direction, from flowing through the galvanometer. The manner in which the string galvanometer is arranged to reproduce a transmitted picture is shown in Fig. 24.

The connections adopted by the Poulsen Company for photographically recording wireless messages are given in Fig. 17, a string galvanometer of the Einthoven type being used. The two self-induction coils S and S' are in circuit with the detector D and the galvanometer G. The condenser C' prevents the continuous current produced by the detector from flowing through the high frequency circuit; P is the primary of the aerial inductance and F the secondary. The method of transmitting adopted by Professor Korn appears to be a simple and reliable arrangement, provided that an equally reliable method of producing the undamped waves can be found. Owing to the absence of mechanical inertia it should be capable of working at a good speed, while the absence of a number of pieces of delicate apparatus all requiring careful adjustment add greatly to its reliability.

In any spark system with a properly designed aerial a coil taking ten amperes is capable of transmitting signals over a distance of thirty to fifty miles, but where the number of interruptions of the break required per second is very high, as in radio-photography, it must be remembered that a much higher voltage is needed to drive the requisite amount of current through the primary winding of the coil than would be the case if the interruptions were slower. It is possible to use platinum contacts for the relays, for currents up to ten amperes, but for heavier currents than this some arrangement where contact is made with mercury will be found to be more economical and reliable.

In the transmitter already described and given in Fig. 11, the best results would be obtained by finding the speed at which the relay R' works best, and regulating the number of contacts made by the stylus accordingly.

The method employed by De' Bernochi (see Chapter I.) of varying the intensity of a beam of light by passing it through a photographic film, which in turn alters the resistance of a selenium cell, has been very successfully employed in at least one system of photo-telegraphy. Its application has also been suggested for wireless transmission, and although with any system using continuous waves this would not be very difficult, it could hardly be adapted to work with the ordinary spark system. The apparatus for receiving from this type of transmitter would, on the other hand, necessarily be more elaborate than the methods that are described in the next chapter, and as far as the writer's experience goes, experiments along these lines would not prove very profitable, as simplicity is the keynote of success in any radio-photographic system.

It has been suggested that in order to decrease the time of transmission a cylinder capable of taking a print 7 inches by 5 inches be employed, the print being prepared from rather a coarse line screen—say 35 to the inch—and a traverse of about ¹/₅₀ inch given to the stylus, thus reducing the time of transmission to about twelve minutes. It is questionable, however, whether the increase in speed would compensate for the loss of detail, as only very bold subjects could be transmitted. As already pointed out, wireless transmission would only be employed for fairly long distances, and the extra time and expense required to receive a fairly good detailed picture is negligible when compared with the enormous time it would take to receive the original photograph by any ordinary means of transit.

The public much prefer to have passable pictorial illustrations of current events than wait several days for a more perfect picture—the original, and the advantage of any newspaper being able to publish photographs several days before its rivals is obvious. There can also be no doubt but that a system of radio-photography, if fairly reliable and capable of working over a distance of say thirty miles, would be of great military use for transmitting maps and written matter with a great saving of time and even life. Written matter could be transmitted with even greater safety than messages which are sent in the ordinary way in Morse Code, as the signals received in the receiver of an hostile installation would be but a meaningless jumble of sounds, and even were they possessed of radio-photographic apparatus the received message would be unintelligible, unless they knew the exact speed at which the machines were running and could synchronise accurately.

CHAPTER III

RECEIVING APPARATUS

There are only two methods available at present for receiving the photographs, and both have been used in ordinary photo-telegraphic work with great success. They have disadvantages when applied to wireless work, however, but these will no doubt be overcome with future improvements. The two methods are (1) by means of an ordinary photographic process, and (2) by means of an electrolytic receiver.

In several photo-telegraphic systems the machine used for transmitting has the cylinder twice the size of the receiving cylinder, thus making the area of the received picture one-quarter the area of the picture transmitted. The extra quality of the received picture does not compensate for the disadvantage of having to provide two machines at each station, and in the writer's opinion results, quite good enough for all practical purposes, can be obtained by using a moderate size cylinder so that one machine answers for both transmitting and receiving, and using as fine a line screen as possible for preparing the photographs.

The writer, when first experimenting in photo-telegraphy, endeavoured to make the receiving apparatus "self-contained," and one idea which was worked out is given in Fig. 18. The electric lamp L is about 8 c.p., and is placed just within the focus of a lens which has a focal length of ³/₄ inch. When a source of light is placed at some point between a lens and its principal focus, the light rays are not converged, but are transmitted in a parallel beam the same size as the lens. It has been found that this arrangement gives a sharper line on the drum than would be the case were the light focussed direct upon the hole in the cone A. An enlarged drawing of the cone is given in Fig. 19. The hole in the tip of the cone A is a bare ¹/₉₀ inch in diameter—the size of this hole depends upon the travel per revolution of the drum or table of the machine used—and in working, the cone is run as close as possible to the drum without being in actual contact. The magnet M is wound full with No. 40 S.C.C. wire, and the armature is made as light as possible. The spring to which the armature is attached should be of such a length that its natural period of vibration is equal to the number of contacts made by the transmitting stylus. The spring must be stiff enough to bring the armature back with a fairly crisp movement. The spring and armature is shown separate in Fig. 20.

The shutter C is about ¹/₄ inch square and made from thin aluminium. The hole in the centre is ¹/₁₆ × ¹/₈ inch, and the movement of the armature is limited to about ³/₃₂ inch. In all arrangements of this kind there is a tendency for the armature spring to vibrate, as it were, sinusoidally, if the coil is magnetised and demagnetised at a higher rate than the natural period of vibration of the spring. This causes an irregularity in the rate of the vibrations which affects the received image very considerably. A photographic film is wrapped round the drum of the machine, being fastened by means of a little celluloid cement smeared along one edge.

This device, although it will work well over artificial conductors, is not suitable for wireless work, as it is too coarse in its action; it can be made sensitive enough to work at a speed of 1000 to 1500 contacts per minute, with a current of .5 milliampere. It is impossible to obtain a current of this magnitude from the majority of the detectors in use, so that if any attempt is made to use this device for radio-photography it will be necessary to employ a Marconi coherer (filings), as this is practically the only coherer from which so large a current can be obtained.

There have been many attempts made to receive with an ordinary filings coherer, but as was pointed out in Chapter I. these have now been discarded in serious wireless work, being only used in small amateur stations or experimental sets. As the reasons for this are well known to the majority of wireless workers there is no need to enumerate them here.

A method whereby a filings coherer can be decohered, the act of decohering closing a local circuit which contains the photographic receiving apparatus, is given in the diagram Fig. 21.

In the figure, the coherer C is fixed in rigid supports, one support being provided with a platinum pin F. To the coherer is connected the sensitive electro-magnet M, which becomes magnetised as soon as the incoming waves act upon the coherer. To the armature B is attached a light aluminium arm S, pivoted at K, and carrying at the other end the striker G, which is fitted with a platinum contact. When the armature B is attracted the coherer is decohered by the force of the impact between the contacts F and G. To prevent damage to the coherer the force of the blow is taken off by the ability of the striker to work back through a hole in the arm S, the spring N keeping it normally in a fixed position. T and P are adjusting screws, and the terminals J are for connecting to the receiving apparatus. With this arrangement a very short wave-train causes only one tap of the contacts, so that only one mark is registered on the receiving drum for every contact made on the transmitter.

The drawing, Fig. 22, gives a diagrammatic representation of apparatus arranged for another photographic method of receiving. The machine shown in Fig. 6 is used in this case. A is the aerial, E earth, P primary of oscillation-transformer, S secondary of transformer, C variable condenser, C' block condenser, D detector, X two-way switch, T telephone.

A De' Arsonval galvanometer H is also connected to the switch X, so that either the telephone or the galvanometer can be switched in. The galvanometer can be made sensitive enough to work with a current as small as 10^-7 of an ampere, with a period of about ¹/₁₅₀th of a second. The screen J has a small hole about ¹/₈ inch diameter drilled in the centre. Under the influence of the brief currents which pass through the detector every time a group of waves is received, the mirror of the galvanometer swings to-and-fro in front of the screen J, and allows the light reflected from the source of light M to pass through the aperture in the screen, on to the lens N.

Round the drum V of the machine is wrapped a sensitive photographic film, and this records the movements of the mirror which correspond to the contacts on the half-tone print used in transmitting. Every time current passes through the galvanometer, the light that is received from M,^[6] passes through the aperture in the screen J, and is focussed by the lens N to a point upon the revolving film. As soon as the current ceases, the mirror swings back to its original position, and the film is again in darkness. Upon being developed a photograph, similar to the negative used for preparing the metal print is obtained. If desired the apparatus can be so arranged that the received picture is a positive instead of a negative.

The detector used should be a Lodge wheel-coherer or a Marconi valve-receiver, as these are the only detectors that can be used with a recording instrument. If the swing of the galvanometer mirror is too great, a small battery with a regulating resistance can be inserted in order to limit the movement of the mirror to a very short range; the current of course flowing in an opposite direction to the current flowing through the coherer.

In this, as in all other methods of receiving, the results obtained depend upon the fineness of the line screen used in preparing the metal prints; and as already shown the fineness of the screen that can be used is dependent upon the mechanical efficiency of the entire apparatus.

Another system, and one that has been tried as a possible means of recording wireless messages, is as follows. The wireless arrangements consist of apparatus similar to that shown in Fig. 22, but instead of a Lodge coherer a Marconi valve is used, and an Einthoven galvanometer is substituted for the reflecting galvanometer. The Einthoven galvanometer consists of a very powerful electro-magnet, the pole pieces of which converge almost to points. A very fine silvered quartz thread is stretched between the pole pieces, as shown in Fig. 23, the tension being adjustable. The period of swing is about ¹/₂₅₀th of a second. A hole is bored through the poles, and one of them is fitted Fig. 23.Fig. 23. with a sliding tube which carries a short focus lens N. The light from M passes through the magnets, and a magnified image of the quartz thread is thrown upon the ebonite screen J. This screen is provided with a fine slit, and when the galvanometer is at rest the shadow of the thread just covers the slit in the screen and prevents any light from M reaching the photographic film. Upon signals being received the shadow of the thread moves to one side for a long or short period, uncovering the slit, and allowing light to pass through. The lens R concentrates the collected light to a point upon the revolving film. The connections for the complete receiver are given in Fig. 24.

The modified form of the Einthoven galvanometer, as arranged by Professor Korn for use with his selenium machines for photo-telegraphy over ordinary land lines, consists of two fine silver wires which are displaced in a lateral direction between the pole pieces when traversed by a current; the current passing through both wires in the same direction. A small shutter of aluminium foil is attached to the wires at the optical centre. The silver wires used are ¹/₁₀₀₀ inch in diameter, with a natural period of about ¹/₁₂₀th of a second; the length of wires free to swing being usually about 5 cm.

The period of the wires depends to a great extent upon their length and diameter, and also upon their tension. By using short fine wires the period can be made much smaller, but a greater current is required to produce a similar displacement. Where the current available, as in wireless telegraphy, is very small, and a definite displacement of the wires is required, it is at once apparent that with wires of a given diameter there is a limit to their length and therefore to the period. Finer wires can be used, but here again there is a practical limit to their fineness, although galvanometers have been constructed with a single silvered quartz thread ¹/₁₂₀₀₀th of an inch diameter, which, when placed in a powerful field, will give a good displacement with a current as small as 10^-8 ampere.

With the apparatus arranged by the Poulsen Company, given in the diagram, Fig. 17, for photographically recording wireless signals, the current required to operate the galvanometer for signals transmitted at the rate of 1500 a minute is 1 × 10^-6 ampere, while for signals up to 2500 a minute a current about 5 × 10^-6 ampere is necessary.

Another very sensitive instrument, employed by M. Belin, and known as Blondel's oscillograph, consists of two fine wires stretched between the poles of a powerful electro-magnet, a small and very light mirror being attached to the centre of the wires. The current passes down one wire and up the other, and the wires, together with the mirror, are twisted to a degree depending upon the strength of the received current. In order to render the instrument dead-beat the moving parts are arranged to work in oil. The light reflected from the mirror is made use of in a manner similar to that shown in Fig. 22.

In all photographic methods of receiving, the apparatus must be enclosed in some way to prevent any extraneous light from reaching the film, or better still placed in a room lighted only by means of a ruby light.

The following method is given more as a suggestion than anything else, as I do not think it has been tried for wireless receiving, although it is stated to have given some good results over ordinary land lines. It is the invention of Charbonelle, a French engineer, and is quite an original idea. His method consists of placing a sheet of carbon paper between two sheets of thin white paper, and wrapping the whole tightly round the drum of the machine. A hardened steel point is fastened to the diaphragm of a telephone receiver, and this receiver is placed so that the steel point presses against the sheets of paper. As the diaphragm and steel point vibrates under the influence of the received currents marks are made by the carbon sheet on the bottom paper.

Over a line where a fair amount of current is available at the receiver, the diaphragm would have sufficient movement to mark the paper, but the movement would be very small with the current received from a detector. This difficulty could no doubt be overcome to a certain extent by making a special telephone receiver, with a large and very flexible diaphragm, and wound for a very high resistance. The movement of an ordinary telephone diaphragm for a barely audible sound is, measured at the centre, about 10^-6 of a c.m. With a unit current the movement at the centre is about ¹/₇₀₀th of an inch. Greater movement of the diaphragm could be obtained by connecting a Telephone relay to the detector, and using the magnified current from the relay to operate the telephone.