AMERICAN INSTITUTE OF ELECTRICALENGINEERS.

NEW YORK, April 28th, 1897.The 115th meeting of the INSTITUTE was held April 21st, at 12

West 31st Street, and was called to order by Vice-PresidentSteinmetz at 8.15 P. M.The following associate members were elected by the Executive

Committee April 28th.Address.

President, Calculagraph Co., 2Maiden Lane, New York City,residence, 82 So. Clinton Street,East Orainge, N. J.

Endorsed byF. A. Pickernell.Stephen D. Field.A. N. Mansfield.

BROWNE, SIDNEY HAND Conisulting Electrical Engineer, Louis Duncan.809 Equitable Bldg., Baltimore, H. A. Rowland.residence, Ruxton, Md. Hermann S. Hering.

CARTER, HENRY W. Attorney andl Expert in Patent F. S. Hunting.Causes, Carter & Graves, 810 Thomas Duncan.Reaper Block, Chicago, Ill. A. L. Searles.

GREENWOOD, FRED. A. Secretary California Electric W. F. 0. Hasson.Works, 409 Market St., San Wynn Meredith.Francisco, Cal. F. F. Barbour,

HOAG, GEO. M. City Eleetrician, City of Cleveland, M. C. Canfield.116 City Hall, residence, 3 Chas. W. Wason.Dorchester Ave,, Cleveland, 0. E. P. Roberts.

KAMMERER, JACOB A. General Agent, The Royal Elec- Fred. A. Bowman.tric Co., Toronto, residence, 97 A. A. Dion.Macdonell Ave., Toronto, Ont. Robert A. Ross.

LOVEJOY, D. R. Assistant in Electrical Engineer- F. B. Crocker.ing, Columbia University, resi- Wm. A. Anthony.dence, 222 East 49th St., New Max Osterberg.York. N. Y.

MATHEPR, EUGENE HOLMES Superintendent and ElectricalEngineer, New Haven StreetRailway Co., Exchange Build-ing, New Haven, Conn.

RALSTON, Louis C. Graduate Student, Cornell Uni-versity, 1170 Madison St., Oak-land, Cal.

91

Theo. Stebbins.Geo. F. Sever.Charles Hewitt.

Fredk. Bedell.Edw. L. Nichols.Harris J. Ryan.

Name.ABBOTT, HENRY

92 AS8OC. IEIBERS ELECTED AND TRANSFERRED.

SHAW, HOWARD BURTON

WELLS, DANA CLEMMER

Total 11.

Assistant Professor ElectricalEngineering, Missouri StateUniversity, ColuInbia, Mo.Assistant in Physics, ColumbiaUniversity, New York, resi-dence, 109 Willow St., Brooklyn,N. Y.

TRANSFERRED FROM ASSOCIATE TO FULL MEMBERSHIP.

Approved by Board of Examiners, March 10th, 1897.

HUMPHREY, HENRY H.RICE, CALVIN WINSOR,

Total 2.

Consulting Engineer, St. Louis, Mo.

Consulting Electrical Engineer, Winchester, Mass.

The following paper on "The Synchronograph" was then readby Dr. Albert C. Crehore and Lieut. George 0. Squier, the dis-cussion of which was deferred to May 18th. The paper wasillustrated with lantern slides. The paper was also read atChicago by Dr. Frederick Bedell.

E. H. Hall.C. A. Adams.C. L. Cory.M. I. Pupin.F. B. Crocker,H. C. Parker.

A Paier presented at the xi5th Meeting of theAmerican Znstitute of Electrical Engineers,New York and Chicago, Aril2r, 1897.

THE SYNCHRONOGRAPH.

A NEW METHOD OF RAPIDLY TRANSMITTING INTELLIGENCE BYTHE ALTERNATING CURRENT.

BY ALBERT CUSHING CREHORE AND GEORGE OWEN SQUIER.

In a general view of the technical history of the art oftelegraphy, statistics show that at tlle present time, more than fiftyyears since the introduction of the telegraph, nine-tenths of thetelegraph business of the world is transmitted by hand, in sub-stantially the same manner as then. From an electrical point ofview one naturally asks why it is, that during this periodwhich represents more electrical progress than all time previous,the rapid transmission of intelligence has not made more advance.

It is to experiments upon a new electrical system of rapid in-telligence transmission and its possibilities, that your attention andconsideration are iinvited. It is not intended to enter into a dis-cussion here of the physical causes which have limited the speedand efficiency of the telegraph, but to acknowledge the greatwork of Wheatstone, Hughes, Edison, Delany and others. whohave brought rapid telegraphy to its present state of efficiency,and proceed to an explanation of the principles involved in thenew system, and an account of the experiments already carriedout in developing it. These experiments were conducted at theElectrical Laboratory of the United States Artillery School, FortMonroe, Va., where the land telegraph and telephone lines wereavailable for the actuial trials described.

PRINCIPLES OF THE TRANSMITTER.

It is difficult to treat the subject of transmitters apart fromtheir receivers, as any particular transmitter should be considered

93

CREHORE A-ND SQ UIER

in connection with the limitations of its receiving instrument. Ifwe could have a receiver sensitive enough to make a distinct andpermanent record of every change in current transmitted overthe line, provided the line were so situated as to be free from thedisturbing influences induced by external causes, it wouild beideal; and the discussion of transinitters would be simplified byreducing the elements to the line and transmitting instrumentalone. The qualities of receiving instruments include two princi-pal elements. They all require a certain amount of energy tooperate them, and in addition, most of them have inertia in themoving parts. A distinct advance is made, other things beingequal, in the receiver wlich dispenses entirely with the inertia ofmoving parts. This is accomplislhed by electrolysis in the chemi-cal receiver of Bain, which has recently reached great perfectionin the hands of Mr. Delany. It is also accomplished in thepolarizing receiver which was used in experimnents describedlater.

Transmitters for sending intelligence over electrical circuits are,in every case, instruments which operate to change the strengthof the current employed in the line. This includes the tele-phone, in which the current is a suiecession of waves differingnot only in the frequency with which they occur, corresponldingto the pitch of the tone, and in the amplitude corresponding tothe loudness, but also as to the shape of the waves correspondingto the timbre or quality. The human ear is such a delicate andwonderfully constructed receiver, that it readily translates thiscomplex wave into intelligence. If a physical instrument couldbe found which would write out in visible form the exact shapeof these telephone waves received, the eye might also be educatedto translate them. A perfectly trained eye could detect the dif-ference between the same words spoken by different individualsas the ear niow does. Even though tlle waves might be accuratelyreproduced, the simpler the waves the less the difficuilty of trans-lating them.The inherent distinction between teleplhony and telegraphy is

mainly that, whereas the telephone utilizes both the frequency ofthe waves and their forin, telegraphy relies entirely upon theduration, number, and order of arrangement of these waves, andnot their form. The art of telegraphy is practically limited inthis respect to three elemnents, or their combinations, namely,varying the duration of the waves or pulses, the direction of them,

94 [Akpr. 21t,

ON THE SYNCHRONOURAPH.

their order of arrangement, or the different combinations of these.Considering these elements separately, the first one, using wavesof different duration alone for each character upon the litie isnot at present used. The last method, a combination of variableduration and order of arrangement of waves, comprises the systemof Morse and others so universally used, and includes the morerapid system of machine telegraphy due to Wheatstone.

There are reasons why any system using waves of differentduration is not as simple as one which uses waves of equal dura-tion, when any arrangernenit of make-and-break transmitter, usinga constant source of electromotive force is employed. Some of thechief of these are found in the electrical properties of the line carry-ing the currents. The difficulties become apparent onily when itis attempted to send these waves at a very rapid rate, which isdesirable in rnachine telegraphy. The current requires time tobecome established at the receiving end of the line after theelectromotive force is introduced at the sending end. The currentwave which is sent over the line is a function of the time duringwhich the electromotive force remains applied at the transmitter.There is evidently a practical limit to the shortness of the timewhich the electromotive force must remain applied, determinedby the smallest wave which the receiver is capable of recording.

SuLppose on the other hand that the electromotive force has actedlong enough for the current at the receiver to reach its steady valtue,and then the circuit is suddenly broken at the transmitter. Atime will elapse before the current in the receiver is reduiced tozero. This case is not as simple as the former, because the man-ner in which the break is made must be considered. A slowbreak is different from a rapid one, when there is any arc, that is,a spark formed. The whole line has been charged to the limitof the electromotive force used, and must become sufficiently dis-charged before the next wave can be received. This produces theeffect commonly known as " tailing " which means that a signalbecomes so drawn out at tlle receiver that it interferes with thefollowing signal.

If waves of equal duiration are used, evidently more of themmay be received in a given time, than of any other combination ofwaves, for the shortest wave may be used which will operate thereceiver. With this plan, the effect of "tailing" is reduced.The use of equal waves is adopted by Mr. Delany, who also indi-cates by the chemical receiver the directions, whether positive ornegative, of these equal waves.

1897.] 95

CREHORE AND SQ UIER

The alternating culrrent is at present successfully einployed fortransmitting considerable amounts of power over long distances,and the whole system is periodically subjected to a regular anduniforin succession of waves rising gradually from zero to a maxi-

FIG. 1.

inum, and then gradually decreasing, reversing, and increasing toa negative maximum. Recognizing these facts, it seemed prob-able that it would constitute a good means for the rapid trans-mission of intelligence, if the characters of a telegraphic codecould be impressed upon such a current without seriously affectingits regular operation. It is to the consideration of a system ofrapid transmnission of intelligence by the use of the alternatingcurrent that we invite your attention.

Let the sine curve, Fig. 1, represent a regular succession ofsimple harmonic current waves given to the line by an alternatingcurrent generator. If the current passes through a key whichmay be opened or closed at pleasure, then, provided the keypreviously closed is opened at a time corresponding to the pointr of the wave uipon the horizontal axis, it is known that the cur-rent which was zero at the instant the key was opened, will remainzero thereafter, in circulits which have resistance and inductance

FIG. 2.

alone. Again, if the key could be closed exactly at a time corres-ponding to the point Q on the culrve also upon the axis, the currentwill resumne its flow undisturbed according to the sine curve. Thetrue current obtained by opening the key at P and closing it at q

[Apr. 21,96

ON THE SYNCHRONOGRAPH.

is shown in Fig. 2, where the current remains zero between thesetwo points. If the key had been closed at any other point than Q,as at R, the current would not have resumed its flow according tothe simple sine wave; but, it can be shown, would follow tlle

FIG. 3.

heavy curve of Fig. 3, and give a succession of waves alternatelysmaller and larger than the normal sine wave until after a very fewalternations, when it practically coincides with the sine wave. Inlike manner if the key is opened at some other point than P, whentherefore the current is not zero, a spark mav be observed at thebreak, and it requires time for the current to fall to zero.

Let us consider the advantages of thus operating upon an alter-nating current. It is evident that the adva.ntages above men-tioned of using a system subjected to a perfectly regular alterna-ting electromotive force, and capable if necessary of transmittingconsiderable amounts of power, is by this method made available.In addition, no spark is made in a transmitter adjusted to breakthe circuit at the exact times indicated by the curve above, whenthe current is naturally zero. This makes it possible, if it is founddesirable, to use comparatively large electromotive forces andcurrents on the line, for no matter what the maximum value of

z T E N

FIG. 4.

the current, it is made and broken by thlis plan with no sparking.It is also possible to employ waves of high frequency upon theline, the upper limit obtainable from an alternator being probablymuch higher than can be transmitted over the line.

1897.1 97

CREBORK AND SQUIERA

If a receiver were used which could reproduce an exact traceof the actual waves sent over the line, it might resemble such acombination as that represented by the heavy curve in Fig. 4.The sine wave continues uninterrupted to the point P when thekey is opened, and it is held open for one complete wave-lengthP Q, when it is again closed for a wave-length Q R; then openedfor one-half a wave-length R S; closed for a wave-length ST;opened for a wave-length T U; closed for half a wave-length u v;opened for half a wave-length v w and finally closed. By this planit is possible to use the ordinary Continental code in telegraphy,a dash being indicated when two successive waves, a positive anda negative one, are omitted by keeping the key open, and a dotmeaning where a single half-wave is omitted. The space betweenparts of a letter, as between the dash and dot of the letter "n"is indicated by the presence of one-half a wave-length, and thespace between letters, as between " t " and "e "in the word " ten,"by the presence of two half-waves, while the space betweenwords may be represented by the presence of three half-waves,and between sentences of four half-waves or more. The aboveis a single example. of which there are many, of a method bywhich the usefulness of so operating upon an alternating currentis iiade apparent, because it shows how these signals may be in-terpreted by a fixed code. It need not be said that there areother ways easily devised of interpreting the possible combina-tions of waves which may be sent in accordance with any code,and it is not our present object to present a method which isdeemed superior to others, but merely to show that the aboveplan becomes operative.A consideration of the time required to send the word "ten" by

the above plan shows that it corresponds to the time of elevenhalf-waves of current. If we suppose that the frequency is anordinary one used in alternating current work, viz., 140 completewaves per second, the time required to send the word " ten" is .0394of a second, or, by allowing three additional half-waves for thespace between the words, the word " ten" would be sent just 1200times in one minute. There is no difficulty in using over somelines, a frequency four times as great as that ordinarily used,namely, as high as 560 or even 600 periods per second. Thiswould correspond to speeds of 4800 and 5143 times sending theword " ten" per minute. The limit in each instance is only deter-mined by tile particular line ulsed.

98 Apr.] 21,

ON THIE SYNCHRONOGRAPH.

Hitherto it has not beeii pointed out how it is possible to man-ipulate a key at the high speed mentioned, so as to open and closethe circuit hundreds of times per second as desired at the exactinstants when the current is naturally zero. Evidently tlle properplace to manipulate such a current controller where the circuitmnst be miade and broken at distinct points of phase, is at thegenerator itself, or in connection with any inotor runninig syn-chronously therewithl.

It will be sufficient for purposes of illustration to show by aspecial example lhow any single half-wave may thus be controlledat the geinerator; for obviously any word or sentence may beformed by a repetition of this operation.

In Fig. 5, s represents the shaft of an ordinary 10-pole alterna-ting current generator which drives through the gears -M and N,the wheel w. The circumference of this wheel is one continuous

mAt

N W

FIG. 5,

conductor presenting a smooth surface for brushes to bear upon.If the periphery of the wheel is divided for example into 40equal parts, and it is geared to run at one-fourth the speed ofthe armature, each division thereof corresponds to one semicycleof the electromotive force produced by the generator. Upon thewheel w bear two brushes A and B carried by a bruslh-holder whichis capable of adjustment. These two brulshes are connected inseries withl the line, so that the current which passes in at onebrush, is conducted through the wheel to the other brush, andthence to the linie. The current used may be obtained from thegenerator, the shaft of which is represented at s, either before orafter it has passed through any number of transformers, since itis the frequency alone with which we are concerned.The line current is brought to the wheel w to be synchronously

operated upon. If both brushes remain continually in contactwith the wheel, the current transmitted would have the regular

1897.] 99

C0REHORE AND SQUIER

sine form represented in Fig. 1, and for each revolution of thewheel there would be 40 semi-waves or 20 complete waves trans-mitted. If one-fortieth of the circumference of the wheel iscovered by paper or other inisulating material as indicated at iFig. 5, and the brush A adjusted to ride on to and off from thisinsulation just as the current is changing from one semicycle tathe next, that is, changing sign, while the brush B is in continuousengagement with the wheel, the semicycle represented by thesection covered will be suppressed, and without any sparking,even if thLe potential used is high. In practice, the brush A iSeasily adjusted to this point by moving it slightly, backward orforward around the circumference of the wheel until the sparkingceases. This adjustment once made, the brush is fixed in positionand so remains. In each succeeding revolution of the wlheel, thiscycle of operations is exactly repeated, and the current sent overthe line would resemble that shown in Fig. 2, having every for-tieth semicycle omitted. It is only necessary to cover othersimilar sections of the circumference of the wheel in a predeter-mined order according to a code, to transmnit intelligence over the-line. The above illustration of the operation of a transmitter onthis principle is given for simplicity only, and is evidently farfrom a practical form of transmitter.The wheel w in. the above example, may have different speeds

with respect to the generator shaft, the essential condition beingthat its circumference shall contain some integer number of a unit,which is the arc upon the circuinference of the wheel if geared tothe armature, that a point fixed witlh respect to the field woulddescribe uponi it during one semni-period of the current. This.wheel therefore might be coinnected to any shaft which runs insynchronism with that of the genlerator, as for instance that of asynchronous motor if the power were obtained from a distance.

Instead of using insulating papers situated upon a single cir-cumference of the wheel, two or more similar lines may be usedeither upon different circumferences of the same wheel or upondifferent wheels, and separate brushes ride upon thie differentcircumferences. The same frequencies of current may be em-ployed to operate all lines of brushes, or currents having differentfrequencies may be employed uipon the separate circuits, all ofwhich use the same line for transmission. These arrangemnentsmake it possible either to send different messages simultaneouslyover the same line employing a single cycle as the unit, or to send

too [Apr. 21,

ON THE SYNCHRONOGRAPI.

a message employing different frequencies to represent the dif-ferent characters of a code, or many combinations of these.The employment of alternating waves of different frequencies

upon the same line by the method shown, does not have the sameobjections which exist when a constant electromotive force is used.Since the circuit is by this system always interrupted when thecurrent is naturally zero, the frequency employed is within certainlihuits a matter of indifference as the line is in the same conditionwhetlher a long or a short wave is used.

It is seen that by this simple method of operating uponi thealternating current, according to the above principles, there iscomplete control of the individual semi-waves of the current,which may be clhanging direction thousands of times in a second,far beyond the range of possible manipulation by hand. In otherwords it is easy to obtain a record of any pre-selected order orsuccession of semi-waves desired. It is evident that it is as im-portant to be able to control the semni-waves retained, as it is thosesuppressed, since they are of equal value in interpretation.Furthermore there is great utility in being able to control eachsingle semi-wave of the current, for this permits the mnaximumspeed of transmission of signals with a given frequency.A transmitter which operates upon the current at intervals

comparable with the duration of a semi-wave, but which does notact in synchronism with the current, would necessarily makeand break the circuit at timnes when the current is not naturallyzero. If this were done there would not only be sparking, butin additior, the current would be interfered with in suhc amanner as to miake it pr4bable that the record received couldnot be interpreted; for the current at each make would followsuch a curve as that shown in Fig. 3.

PRINCIPLES OF RECEIVERS.

As used throughout this paper, the term "1 receiver " will beunderstood to mean that mechanism which uses time energy trans-mitted over an electric circuit, and transforms it so as to make apermanent record whiclm miay be translated into intelligence.The term receiver is here restricted to mean instruments whichmakew a permanent record, since this is a necessary condition forthe rapid transmission of intelligence, with which we are atpresent concerned. All receivers require a certain amount of

1897.] 101

CREHORE AND SQUIER

power to operate them, and the power required affords one basisfor their classification.Another method divides receivers into two classes, those which

have inertia in the moving parts and those wlhich have none-There is no fundamental reason why any one of these generalclasses should contain all of the most rapid receivers. Any of theabove classes might incluide receivers which are very rapid. If,for instance, a receiver has inertia in the moving parts, for rapiditythe amouint of inertia shouild be small, and its natural period high,or a large amount of energy would be required to operate it. Ifa receiver has no inertia in the recording inechanism, then thepossible rapidity is limited by the power supplied.

In deciding upon the relative merits of receivers from the pointof view of rapidity, the cost of the power required offers no reasonwhy considerable amounts of power should not be transmittedover certain land lines for purposes of telegraphy. Using areceiver which possesses much inertia in its moving parts, it doesnot follow that even thouigh considerable energy reaches the recei-ver over the line, that it will be rapid in its action.The Wheatstone receiver may be taken as representative of a

type of receivers possessing inertia in the moving parts, whichhas come into successful operation. The record is made in thisinstrument by a small wheel which vibrates back and forthbetween an ink surface and the recording tape. The energy whichis essential, is that required to move this little wheel aind the partsconinected with it back and forth. Although considerable energymiay possibly be sent over the line and be expended in the instru-ment, it seems impossible to concentra1* more than a certain partof it upon the moving mechanismi. This suggests two methods ofimproving the speed of the system; either to increase the powerreceived by the moving parts, or diminish their moment of inertia.One factor which limits the Wheatstone type of receiver is thatthe moving parts are required to do the work of making therecord. This is not a necessity, since light may be employed asthe agent to make the record under the control of the movingparts, as is evidently accomplished in a form of galvanometerhaving a very minute needle with mirror attached, the slightestmotion of which is greatly magnified by the reflected beam oflight.As a type of instrument having no inertia in the recording

mechanism, may be mentioned the various forms of chemical

102 [Apr. 21

.ON YIHE SYNCHRONOGRAPE.

receivers acting by electrolysis. This type of receivers possessesmany advantages, perhaps chief among which is the fact that alarge part of the energy received is brought directly to bear uponmaking the record. Another feature is the simplicity of theessential mechanism involved, as no intermediate steps are em-ployed after the impuilse is received from the line before therecord is made. These qualities alone imply rapidity, and thisreceiver is one of the most rapid known. The limiit of rapiditywith this receiver is the power received from the line. If thepotential between the terminals of the receiver is increased, thetime required to make a given record is correspondingly reduced.The use of the alternating current permits of greater potentialsbeing realized in the receiver with less disturbing influence fromthe line than would be the case if a constant direct electromotiveforce was employed.A new type of receiver having no inertia in the recording

mechanism was used in developing the transmitter described inthese experiments. This instrument has already proved of valueas a chronograph for the measurement of minute intervals of time,and for the study of any kind of variable electric currents.Although its application as a practical telegraph receiver is not atpresent advocated, yet the realization of a mnasless receiver upondifferent lines mierits description. This receiver is based uponFaraday's discovery of a direct relation between light andelectricity.

This discovery was, that if a beam of polarized light is passedthrough somne substance in the direction of the lines of magneti-zation within that substance, there is a rotation of the plane ofpolarization in a direction which is the same as the direction ofthe cuirrent required to produce such a, magnetic field. Thedirection of rotation is unaltered, therefore, whether the lightbeam advances in the same or in the opposite direction to themiagnetization, so that a beam reflected back and forth through thesubstance several times, has its rotation increased by eqqual amountseach time. If the direction of the ray of light is at right anlglesto the lines of magnetization, there is no rotation produced. Theamount of this rotation has been investigated by Verdet, who an-nounced laws by which it may be expressed. They are summedup in the following statement:

" The rotation of the plane of polarization for monochromatic" light is in any given substance proportional to the difference in

1897.] 103

CREHOR?E AND SQUIER

"magnetic potential between the points of entrance and ernergence"of the ray"; that is, it is equal to a constant times this differeniceof potential and is expressed by the formula

=v V,

where 0 = angle of rotation, V = difference in magnetic poten-tial, and v for a given wave-length is constant in any one sub-stance and is kniown as Verdet's constant.The following example will iiake mtiore evident the application

of Faraday's discovery to this receiver. Admit a beam of ordi-nary light through a smnall aperture into a dark room anid let itfall upon a white screen. Suppose that the aperture which admitsthe beam is provided with a shutter which may be opened or closedat will. We have in this simple arrangement all the essentials ofa transmitter and receiver of intelligence. A person opening andclosing the shutter miight communicate with a second person ob-serving the screen, which would become light and dark at inter-vals in accordance with a pre-arranged code. Substitute for thefirst person in direct control of the shutter an electromagneticdevice. operated from any desired distance through an electriccircuit, anld the effect upon the screen is the same as before. Forrapid transmission it would be necessary to substitute a mechanicaltransmitter which would operate faster than a person can send byhand. There would be no particular difficulty in thus movingthe shutter more rapidly than any observer couild read fromn thescreen. It then becomes necessary to make a permanent record,which may be accomplished by substituting for the screen a self-recording surface having a relative motion across the beam. Thisis afforded by using any surface sufficiently sensitive to light,many varieties of which are available. In fact a surface is avail-ab]e which is so sensitive that it will record much faster than thematertial shutter can be moved back and forth so as to open andclose the aperture.The next step in increasing the speed, provided the limits of

the transmitter lhave not already been reached, is to secure a morerapid shutter. It was with this object in view, to obtain a mass-Zess shutter that Faraday's discovery is used. Instead of passingthe light directly througfh the aperture, it is first passed througha Nicol prism in order to obtain a beamri of plane polarized light,or it may be polarized in any other suitable manner. Supposethat a second Nicol prism like the first is placed in the path of

104 [Apr. 21.

ON TEE SYNJCHRONOGRAPH.

the polarized beam. If the second prism known as the "analyzer"is turned so that its plane is perpendicular to that of the first prismknown as the "polarizer," all the vibrations not sorted out by thepolarizer will be by the analyzer. In this position wlhen the planesare perpendicular to each other, the prisms are said to be "crossed,"and an observer looking through the analyzer finds tlle lighttotally extinguished as thouigh a shutter interrupted the beam. Byturning the analyzer ever so little from the crossed position, lightpasses through it, and its intensity increases until the planes ofthe prisnms are parallel, and if one of the prisms is rotated, therewill be darkness twice every revolution. In order to accomplishthe same end that is obtained by rotating the analyzer withoutactually doing so, the following plan is adopted: Between thepolarizer and the analyzer is placed a transparent medium whichcan rotate the plane of polarization of the light, subject to thecontrol of an electric current, without moving any material thing.The mediumiused in this receiver is liquid carbon bisulphide con-tained in a glass tube with plane glass ends. There are manyother substaiices which will answer the purpose, some better thanothers. This was selected because it is very clear and colorless,and possesses the necessary rotatory property to a considerable ex-tent. It only possesses this property, lhowever, when situated ina magnetic field of force, and the rotatory power is proportionalto the intensity of the magnetic field. To produce a magneticfield in the carbon bisulphide, a coil of wire in series with the linefrom the transmitter is wound around the glass tube. When thecurrent ceases, the carbon bisulphide instantly loses its rotatorypower. The operation is as follows: First the polarizer andanalyzer are permanently set in the crossed position, so that nolight emerges from the analvzer. A current is sent through thecoil around tlle tube. The plane of polarization is inmmediatelyrotated. This is equivalenit to rotating the polarizer through acertain angle, and hence light now emerges from the analyzer.Interrupt the current, the medium loses its rotatory power, andthere is again complete darkness. The arrangement makes aneffectual shutter for the beam without moving any mass of matter.

This illustrates how Faraday's discovery may be utilized to re-place the electromagnetic shutter in the above example by amassless shutter, which enables the current waves sent over theline to be recorded upoIl the sensitive surface without movingany material thing. An advantage of this receiver is that the

1897J] 105

CREHORE AND SQUIER

speed is not limited by the receiver but only by the naturalproperties of the line or of the transmlitter. Used in coilnectionwith the transmitter already described, the real limit is found tabe in the line itself.An analysis of this receiver shows that the energy received

over the line is not used directly in making the record, but theagent whichl makes the record is the beam of light which derivesits energy from a local source. The energy received from theline merely controls this local energy which may have considerablepower behind it. This controlling phenomenon is one of the fewknown cases where electricity acts directly upon light. Themnechanism by which this action is effected is not at present known,and any experimental evidence upon it would increase our know-ledge of the connection between ether and ordinary matter, aswell as the constituition of matter itself. The use of this directinfluence of electricity on light makes the speed of transmissionthrough the receiver comparable with the velocities of theseagents.

DESCRIPTION OF THE TRANSMITTER USED.

In these experiments, the operation upon the alternating cur-rent according to the principles already stated, was accoinplishedby means of a prepared perforated paper tape, which was causedto move by the generator itself. A view of this tape, showing amethod of operating upon the current is given in Fig. 6.The line current is brought througlh the wires ww to two

brushes BB' not shown in the view, held by the adjustable sup-port s. The plan of the brushes is shown in Fig. 7. One bruslhB bears upon the tape from above, and the other brush B' bearsfrom below immediately opposite the other brush so that theywill meet through the perforations in the transmitting tape T.When tlhe brushes meet through the perforations in the tape, theline circuit is closed, and when paper passes between themn, separa-ting the brushes, the line circuit is broken.

It is so arranged that the brushes pass off from, and on to thepaper, thus making and breaking the circuit, at the instant cor-responding to the points in the current wave, Fig. 1, when thealternating current is naturally zero. The tape T passes over awheel r geared to the generator shaft, so that for one revolutionof the armature, the tape advances a fixed distance. If the gen-

106 [Apr. 21,

]ON THE SYNCHRONOGRAPII.

erator has ten poles, this fixed distance on the tape correspondsto five complete waves or ten alternations or semi-cycles of thegenerator current. One-tenth of this fixed distaince correspondsto one alternation or semi-cycle of the current, and may be takenas the unit of distance in perforating the tape. If therefore ahole is made in the tape equal in length to this unit, and thebrushes B and B' happen to pass off from the paper so as to meetthrough the hole at the instant the current is naturally zero, thenthey will pass on to the paper again, breaking the circuit at thenext following instant when the cuirreint is also naturally zero,since the length of the hole corresponds to oine seini-cycle of thecurrent.

Suppose that a succession of these unit holes is made, the tapebetween the holes being also of unlit length, then the circuit willbe made and broken as by the first hole at the points of zero cur-rent. In practice it would probably not happen that the brusheswere at first so situated as to pass off fromn and on to the paper at

T B

B

FiG. 7.

the instant the current is zero. In this case a suiecession of sparksappears, one each time the brushes pass on to the tape, and bymoving the brushes along the tape it will be observed that thisspark eitlher increases or decreases in intensity, according to thedirection mnoved; but at regular intervals, equal to the unit men-tioned above it disappears. This position of the brushesfor no sparkiing is easily found by trial, and once obtained renmainsfixed. By this simple practical operation wlhiel experience showsrequires but a moment to accomplish, the essential condition ofsynchronouslv operating upon the current in the manner describedis secured. The brushes once adjusted always so remain, andsince there is no sparking, it is possible to use high electromotiveforces upon the line without injurious effect upon the brushesand tape. It is also plain that this method of operating upon thecurrent is not affected by the speed of the generator, since thetransmnitting device is always in synchronism with the generator,

1897.] 107

CREBORK AND SQUIER[

whatever the speed. The speed of the generator, and conse-quently the rate of sending, as far as the transmitter is concerned,can be varied at pleasure between wide limits, without any effectupon the synchronous operation described.An example, giving the data from an early experiment, will

-illustrate how this is accomplished. The generator was a FortWayne 10-pole alternator giving a potential at its terminals of1,000 volts. This was transformed to about 300 volts, beinig con-venient to handle, and sufficiently high for the purpose. The endof the shaft E, Fig. 6, of the generator carries a small pinion,which engages the gearing G, and revolves the wheel P once inevery 18.4 revollutionis of the arnature. This makes the t.1 partof the circunmference of the wheel P correspond to one semi-cy-cle of the current. The circumnference of the wheel was about100 ems., and the length of a unit, therefore, . 8w of this, or.54 cms.For convenience, a tape mnade of ordinary paper, had its two

ends joined so as to make a continuous belt, which inade it pos-sible to use it repeatedly. The tape passed from the large wheelP to the loose pulley Q mounted upon a base-board A, and underthe guiding pulley attached to the support s which controlled thetape, immediately before passing th- brushes B B'.In preparing the tape, the Continental code was employed as

described, the omission of two half-waves meaning a dash, andone half-wave a dot. Ilaving obtained the length of a unit onthe wheel, the tape is first divided into these equal units, and thenthe proper units are cut out to form a message. The units whichare not cut out form the dots and dashes. To use a continuoustape it is necessary for its length to be some exact multiple of theunit, in order that it may start on the seconid revolution exactlyas it did the first.The generator current of high potential passed directly through

the primary of a transformer, and the secondary was used as thesource of electromiiotive force for the line. This secondary cir-cuit which includes the line was brought to the transmitter to beoperated upon as described.A diagram of the electric circnits as employed in this experi-

ment is shown in Fig. 8, where A represents the alternator, T thetransformner, B the brush bearing upon the transmitting wheel w,and L the line. Another diagram illustrating how the mnethodmay be used with currents-of the same or different frequencies,

-108 [Apr. 21,

FIG. 6.

FIG. 10.

e g r a_ D~~~~~1

FI.G. 11.

I I - d o -1tj

*-0

0

Fmu. 12. Polarizing Receiver Record. Message sent at Fort Monroe,Va., Aug. 1(1, 1896, at the rate of about 1,200 words per miinite.This particular message sent in one-half a second. The numiiberof ldots Perl seeondl is 337. Continental ('ode is use(l.

. S

ON THE SYNCIHRONOGRAP.H.

is given in Fig. 9, where several generators A1, A2, A3, etc., are re-presented upon the same shaft, and each is connected to a separatebrush B1, B2, B8, etc., bearinig upon the wheels wl, w2, W., etc.,upon a common shaft and connected to the line. By placing theinsulating papers in the proper positions upon the wheels, itbecomes possible to transmit in succession, first a eurrent of onefrequency and then of another.

L

FIG. 8.

If any error is made in laying off the units upon the tape, or ifthe length of the tape changes in any way after they are accuratelylaid off, the effect of this error is cumulative from period to period,and although at any particular time the tape might be in phase,sometime later it woruld not be so, and sparking would occur.This would also be the case if there were aniy slipping of the tape

FiG. 9.

around the wheel P. To overcome these difficulties it is onlynecessary to have holes punched at regular intervals in the tapewhich engage in pins at corresponding intervals on a wheel madeto receive it.A simple experimental mnethod which does away with the

necessity of making pins to feed the tape, is to glue strips of thinpaper, seen at c, Fig. 6, having lengths corresponding to the paper

-1897.] 109,

CREHORE AND SQ UIER

intervals of the tape, that is, one unit for a dot and two units fora dash, upon the circumference of the wheel P, which has a smoothpolished metal surface. One brush is continually in contact witlthe wheel, and the other rides on to and off from these paperstrips, making anid breaking the circuit at the zero phase of thecurrent. The length of message permitted, is limiited by thenumber of units in the circumference of the wlheel, which in theexample taken was 184.

Instead of using any gearing, as in the example given, to reducethe speed of revolution of the wheel P, the tape might be rundirectly from a small wheel upon the armature shaft. The uniton this small wheel is one-tenth of its circumference, if there areten poles to the generator, so that any mressage sent by fasteningpapers upon this wheel would be limiited to ten semi-cycles. Ifa single unit of this small wheel is covered by paper, and thebrush adjusted for no sparking, one semi-cycle in every ten willbe omitted. A record obtained in this case with the polarizingreceiver is shown in the circle at A, Fig. 10, in which each lightspot corresponds to one semi-cycle of current, and it is seen thatone in every ten is omnitted. The record in the eirele B of thesame figture was obtained by using two units of paper on the samewheel, having two units between them, and shows that two semi-cycles are omitted in every tein.

Records obtained by the use of paper fastened upon the largewheel P, Fig. 6, are shown in Fig. 11, where it is seen that theword " telegraph" was transmitted. A record obtained by theuse of tape is shown in Fig. 12, where the sentence "one wirewill do work of ten" was transmitted onl August 10, 1896. Thismnessage was sent at the rate of 337 semi-cycles of current persecond, thus requiring about half a second altogether.

Since the speed of transmission depeinds upon the frequency ofthe alternating currenit, the limit of speed is determined by theparticular alternator used. In the above example tthe alternatoravailable was designed to runi at a speed of abouLt 1600 revolutionsof the armature per minute, corresponding to a frequency of 133,or 266 alternations per second. To increase the speed of trans-mission, this alternator was run as high as 2400 revolutions perminute, beyond which it was thought dangerous to go. Thiscorresponds to a frequency of about 200 complete cycles or 400alternations per second. Throuigh the kindness of 'Dr. M. I.Pupini of Columbia College, a special high-frequency alternator

[Apr. 21.,10

IuT

| u. 14. 1T'he 1PoltarizinIg P'linto-Chroliogratph.

XON TIlE SYNCHRONOGRAPE.

was loaned for the purpose of testing this system at higher speedsof transmission. This alternator, shown in Fig. 13, is in fact fouralternators upon the same shaft, having 18, 22, 26, and 30 polesrespectively. To obtain the highest speed, the 30-pole machinewas used, and the transmitting wheel geared to the shaft as withthe ten pole alternator. The speed of armature uised was 2180revolutions per minute, corresponding to 1090 semi-cycles persecond, or 65,400 per minute, or a frequency of 545. No difficultywas experienced in sending and recording messages at this rapidrate which corresponds to between three and four thousand wordsper minute.

THE POLARIZING RECEIVER.

The statnment of the general principles employed in this receiverhas previously been given, and it remains to describe the actualform. This instrument was designed for a military chironographto measuire the velocity of projectiles, and is known as the Polar-izing Photo-Chronograph1. A view of this is shiown in Fig. 14.Without giving a complete description of the instrument, which

may be found in the reference cited, it will suffice to describe itsessential elements. A sensitive photographic plate 12 x 12 inchessquiare is carried in a metal plate holder, which revolves in thewheel w driven by the mnotor Mr. A powerful beam of light fromthe arc lamp, A, situated upon an inverted T-rail, o', serving as anoptical bench, is condensed by the lens L, and passes through thepolarizer P, a Nicol prism, thence through the glass tube T, con-taining liquid carbon bisulphide, and surrounded by a coil of wire,through the analyzer A, a second Nicol prism. The light receivedthrough the anialyzer is finally passed tlhrough a second lens L' tofocuis the beam uipon the horizontal radial-slit in front of the mov-ing sensitive plate. In its operation, the analyzer A is rotateduntil the liglht is completely extinguished, when no current is pass-ing around the tube T. The coil upon the tube is in circuit withthe line from the transmitter, aind the closing of the circuit at thetransinitter thus sends a current around the tube, and light im-nediately appears upon the camera slit. This is accomplished in-stantly upon closing the circuit, without involving the motion ofany mnaterial thing. Upon breaking the circuit the light immne-

1. " The New Polarizinig Photo-Chronograph," Crehore and Squier. JohnWiley & Sons, New York, 1897. Chapman and Hall Ltd., London.

1897.] ill

CREHORE AND SQUIER

diately disappears, and by observing the light conme and go, it iseasy to read with the eye as rapidly as can be sent by hand. Toproduce a permanent record it is only necessary to rotate the pho-tographic plate in the wheel w. The time required by the pho-tographic plate to make a clear record, depends largely upon theintensity of the light; but the intensity of light which it is practic-able to obtain allows the time of exposure to be much shorterthan is required for the purpose of a telegraph receiver. For in-stance, suppose the width of slit is one millimetre at a distanceof 150 millimetres from the centre of revolution, and the platerotates 1000 times per minute, the velocity of a poinit on the plateis 1570.8 ems. per second, and the exposure is therefore about.000063 second; for the point crosses the millimetre slit in thistime. Tlle above figures are those actually used with the chrono-graph in measLuring the velocity of projectiles inside the bore of agun and the records obtained are perfectly clear. The rapiditv ofthis receiver is illuistrated by stating that as many as seven ob-servations upon a projectile inside the bore of a U. S. 3.2-inchbreech-loading field rifle have been recorded in the first 57 centi-metres (1 foot 102 iniches) of its travel, and observations as ineartogether as 3.8 cms. (11 inches) have been obtairned. These cor-respond in time to intervals less than a thousandth of a second,or they bear about the same relation to a second, as a seconddoes to a third of ail hour.In chronography as applied to gunnery, since the agent which

operates upon the transmitter circulit is the projectile itself mak-ing and breaking the circuit by passing through screens, evidentlyif the screens are properlv placed according to a code, a messagecould be transnmitted to the receiver by a projectile in its flight.

THE CHEMICAL RECEIVER.

In a practical form of receiver, it is an advantage to have theinessages received in suclh form that they are ready for immediateuse, and this is the case with the chemical receiver to whichreference has already been made.Through the kindness of Mr. Delany, some of the sensitive

paper tape used in his system of machine telegraphy was ob-tained for experiments with the synchronous transmitter. Asimple method of obtaining records of currents with this tape,which is certain in its action and does not involve aniv special

[Apr. 21,112

ON IYHE SYNCHRONOGRAPH.

apparatus, is to place the tape upon a smooth metal surface,which serves as one electrode, and to draw a steel needle, servingas the other electrode, along it guided by the straight edge of aruler. If a direct current is used, no record appears when thecurrent is in one direction, and it does appear when the currentis reversed. If a second needle is substituted for the plate elec-trode, the record appears on one side of the tape for a directcurrent anid on the other side for the reversed cuirrent.

If the two nieedle electrodes are placed side by side uipon thetape, a record will appear at one needle for a direct current, andat the other for the reversed curren-t. Employing the alternatingcurrent with the single needle and plate as electrodes, the recordshows a regular succession of distinct miarks, separated from eachother by equal intervals. Each mark exhibits an intensitv vary-ing approximately accordinig to the siine curve. Since by thisarrangement the current makes its record in one direction only,the result is that alternate semi-cycles of the current are sup-pressed, and alternate ones are recorded.By receiving with two needles side by side, all the alternations

are recorded, those that were suppressed before now appearing atthe second needle. The record then appears as two parallel linesof marks having the maxiintiti intensities in one line oppositethe spaces in the other. Using the tranismnitter as already de-scribed with a semi-cycle as a unit in preparing the tape. andreceiving in two lines, it is founid that somrse of the marks areomitted in one line and some in the other, and to facilitate trans-latinig it is simpler to bring the two lines into coincidence toobserve the dots and dashes of the message. A message was th-enprepared upon tlle transmitter wheel, using a complete cycle as aunit, instead of a semi-cycle. Wlhen received in a single line thismnessage is coinplete, no matter to which terminal of the cireuitthe receiving needle is connected, because eaclh unit now containsboth a direct and a reversed current, one of which will record.The same message was then received in two lines, and one line

gave the complete message as before, wlhile in the other line thereappeared a record for each complete unit in which the current was,nmade. The papers of either the first or the second lhalf of eachcomplete cycle composing the message upon the wheel were next-removed, and the message received in two lines as before. Theresult showed the inessage complete in one line, while in the otherline appeared an uninterrupted succession of marks just as givenby the simple alternating current received in one line.

1897.1 113

CRREHORE AND SQ UJER

If then an uniniterrupted line of marks can be received in oneline at the sanme time that a message is being received in theother, this uninterrupted line can be used for a second messageentirely independent of the first. The next experiment accom-plislhed this, and it is now possible to use the same line to sendtwo entirely iindependent mnessages in the same direction at thesame time at a high rate of speed. The preparation of the trans-mitting tape to accomplish tlhis, simply requires that the twomessages, each prepared with a douible unit, shall be displaceda semi-cycle with respect to each otlher as they pass tlhrough thetransmitter.The advantages of duplexing the line, that is, sending two

independent messages in opposite directions over one wire 'atthe same time seem more inmportant than those of diplexing theline. An arrangement of circuits whieh accomplishes thisproves to be very simple. Moreover it permits entirely differentfrequencies to be employed by the transmitters at the two endsof the line, and as before involves no synchronoius receiver ateither end. By duplexing the line the spee(d of transmissionover a single wire is practically doubled; for example a line thatcarries 3000 words simplex can carry 6000 words per minuteduplex.

It is desirable in many cases to manifold the original copies ofthe message received, and experiments were ii(ade to accomplishthis. All that is necessary is to attach to one terminal, insteadof a single needle, as mnany needles as the nmumber of copies de-sired, having each inake its record upon a sensitive surface. Themanifolding process evidently applies to either simplex or duplexreceiving. Manifold copies of inessages may be received inwidely different localities at the same time fromn one and thesame transmitter, by connecting the receivers in series or inparallel.The alternating current is adapted to use with condensers in

series with the linie wlhere a direct current cannot ordinarily beemployed. An experiment was carried out to send a messagethrough a condenser having a capacity of 9.57 microfarads in serieswith the line, and it was found that the message was transmittedcorrectly. One object of this experiment was to establish the pos-sibility of using a set of Morse instruments upon the line at thesame time that the messages were being transmitted at a highrate of speed by the alternating current. By shuilting condensers

[Apr. 21.114

19]ON THE SYNUICRONOGRAP1.

arounid the set of Morse instruments it was found that the opera-tion of either system did not affect the working of the other, sothat it becomes possible to use the same high speed line for acomplete system of quadruplex telegraphy at the same time.Indeed it seems possible that the present Wheatstone systemcould be operated over the line in conjunction with the alternatingcurrent messages. The experiments with the chemical tapewhich have been outlined above, together with otlhers not heregiven, demonstrate the flexibility of a svstem of intelligencetransmission employing the alternating current.The use of the alternating current as a means of sending intel-

ligence in connection witlh the fact that a message can be sentthrough condensers, suggests the possibility of using the prin-ciples of electrical resonance employing circuits having naturalperiods of their own which will pick out and respond to currentsfrom the line having their own frequency.Although the above illustrations have em-lployed for the most

part the Continental code representing a dot and a dash in aparticular manner by the oimission of certain waves, and thespaces between letters and words by the presence of waves, yetit is evident that this is but one of many corrmbinations which thissystem permits, and that mentioned above is not to be underst.zodas representing the most desirable one.A characteristic of the records inade by electrolysis is the

natural separation of the positive and negative waves of current,which is an advantage in interpretation. This separation is alsoaccomplished in the polarizing receiver by emiploying two receivertubes. Instead of setting the polarizer and analyzer for extinc-tion they are so placed that some light is normally transmittedthrough each tube. The tube coils are so connected that a positivecurrent produces approximiiate extinctioni in one tube, and amaxinmumn transmission of liglht througlh the other. A negativecurient transmits a maximum of lig,ht tlhrough the first tube, andproduces approximnate extinction in the second. An alternatingcurrent therefore causes a record of the positive waves throughone tube, and the negative waves through the other, and thusaccomplishes all in this respect that the clhemical receiver does.

THE LTINE.

It is generally understood that the line limits the speed oftelegraphy. The limnit is usually reached because of the dis-

1897.] 115

CREIIORE AND SQUIER

tributed electrostatic capacity of the line rather than its resistance.The influence of the distributed capacity is to ehanige the formof the wave as well as reduce its amplitude. With a given lengthof line having a certain static capacity, there exist limiits to thespeed obtainable with any given set of instruments which wouldbe a difficult matheinatical problem to predetermnine. The diffi-culty in making this calculation is in the influence exerted by theparticular instruments used. With different instruments theupper limit of speed is very different with the same line. Ittherefore seems that the only way to determine this question isby submitting the systeml to actual trial over a long line.

In order to test this system over as long a line as was available,the land telegraph and telephone lines upon the military reserva-tion at Fort Mlonroe were joined in series, making about thirteenmiles of iron wire having a resistance of 320 olhms. Not onlywas no difficultv experienced in transinitting and receiving ines-sages over this line, but resistance was introduced muaking about1,500 ohms total including tlhe polarizing receiver coil of 390ohms. This trial was at a frequency of about 200 conmpleteperiods per second. With the chemical receiver a coil of 10,900ohms was used in the laboratory and the record was plainlyreceived at a frequency of about 545 complete periods per second.

Since the polarizing receiver gave indications showing the ap-proximnate strength of the varying currents by the intensity ofthe light upon the plate, it was used to study the effects upon thecurrents of arbitrarily introducing capacity and inductanee in seriesinto the line, especially the effect upon the make of an alternatingcurrent at different points of phase. Fig. 15 shows the generalappearance of the simple alternating current with different expos-ures, at different speeds of the plate and the same frequency ofalternation. In Fig. 10 the innerrecord c is that of a circuit having390 olims resistance, 1.03 henrys inductance, and 4.78 microfaradscapacity at a frequeney of 137. At eaclh miake it is observed thatthe first wave is small, followed by a large one. The record at Dis for a similar eircuit in all respects except that the capacityis doubled, being 9.57 mnicrofarads. Theoretical curves' havebeen computed for these cases and they are in agreement withthe records shown.The method of neutralizing the effects of the distributed

1. "The New Polarizing Photo-Chro:iograph "-Journal of tie ur S. Arti-lery, Fort Monroe, Va., Nov.-Dec., 1896.

[Apr. 21,16

KKK~~~~ F K k L I~~~~~DLLR IN MILION

I I4j_~~

T I~~K'~ p~10

F-~~~~~~~~~~~~~~~~~~-

0. oD

0

At

I IW

Fia. 1I.

'-4

-11

!L L 1= I. . .I.I.I I I.IL I. I.I. -T

r -1------ 1-1-i+ H4..At4..E.I.SX. 1--l UO0-1-tJl-t+kkl-Lj-4 .

ELLEG ..1.-[ ±±- --.-t . -1-U, 1=± tCo11400r-i

9000 90

Fjo. 15. Chronograph Records of the Alterniating C(urrenit, unldervarying conditions of circuit and speed of plate.

I"IG. ]:;. TIW I'tlllill Iligh-UrC(luency Allernator, attache(I ti,Svnefironous Trzinsmitter.

ON lIEF SYNCHRONOGRAPH.

capacity of lines by introducing distributed indletance as is nowdone in some telephone lines, would have an especially usefulapplication in a line employing the alternating current for tele-graphy.

CONCLUSION.

When the extent of the transmission of intelligence at thepresent time is considered, and the direct influeniee which thisservice has upon the development of the world's progress, anyproposition which promises to increase its efficiency slhould bereceived with consideration.To better comprehend the volume of this service it is of inter-

est to observe the statistics on the subject. These have beenprepared for the United States mail service, the Western UInionTelegraph Comnpany, and the American Bell Telephone Com-pany of the United States, and are exhibited in graphical formin Fig. 16. The statistics for the United States AMail service forthe last few years were furnished througlh the courtesy of thePostmaster- General.

It is inoticed in general, that there is an increase inl all depart-ments of the intelligence transmission service from the earliestdates. The number of pieces of mail sent during 1896 was5,693,000,0@00 whieh is the greatest amount ever sent in a singleyear. The greatest nuinber of telephone messages on record for asingle year is 757,000,000 in 1895. The largest number of tele-graph messages was seint in 1893 anid amnounted to 66,000,000.Thus the greatest number of telegraph-messages as compared withtelephone messages is in the ratio of 1 to 111. The greatestnumber of pieces of mail is in the ratio of 86 to 1 as comparedwith telegraph messages, or in the ratio of '1- to 1 as comparedwith telephone messages. It is also seen that the cost of themail service of the United States in 1896 was $90,626,000, orabout $1.25 per capita. The greatest receipts for any year of theAmericani Bell Telephone Company were in 1895 $16,400,000,about 25 cents per capita, while the greatest receipts of theWesterni Union Telegraph Company were in 1893 $24,97T,000,about 35 cents per capita.

It appears therefore that the. people of the U,nited States payfor a telegraph service of about one-eighty-sixth the amount, aboutone-fourth of that paid for the entire mail service of the UnitedStates. It also costs one and-a-half times as much for telegraph

1897.1 117

CREIHORE AND SQUIER

service as for the telephone service, although the number of tele-phone messages is about eleven-anid-a-half times as great.A conelusion to be drawn from the above general data seems

to be that the people are willing, to pay more in proportion for akind of service like that of the telegraph than any other. Fromthe point of serving the people, as well as from a bulsiness stand-point, it appears that improvement in this class of intelligencetransrnission is at present mueh to be desired. The present stateof the art of telegraphy points to improvements along the line ofautomnatic imachine transmission.

It is of interest to inquire what effects a system of telegraphycapable of sending continuously 3,000 words a minLite would haveon the existing methods. To take a single example of the busi-ness between New York and Chicago, where about 40,000 lettersare carried daily, it would require but two lines in continuousoperation to handle the entire buisiness. At present it takes threedays to receive by mail a business reply between New York andChicago. This transmissioni by maclhine telegraphy could be ac-complished easily the same day. It is thought that an effect ofthis would be to increase business transactions to such an extentthat the total volurne of intelligence transmitted would be augmen-ted, rather than to diminish the business now done by existingmethods.The class of business which such a system would probably at

first obtain would be the less urgent telegraph business of greatervolume, such as the Associated Press dispatches and newspaperpress reports. Among the possibilities is the simultaneous pub-lication of the same newspaper in different parts of the country.For example, in an edition of a daily paper having twelve pagesand eight columns per page, mnaking ninety-six columnns in all,there are less than 185,000 words. At the rate of 3,000 wordsper minute it would only require about an lhour to transmit theentire contents of the paper. This calculation furthermoreassumes that the whole paper is uniformly printed in fine type.It would require a single operator, working bv hand and aver-aging twenty words per minute, over six days of twenty-fourhours each to send this amount.The system proposed is especially adapted to meet the de-

mands of this class of business; for the great flexibility of thealternating currenit as employed, permits if necessary con-siderable amounts of power to be transmitted over the line which

118 [Apr. 21,

ON THE SYNCHRONO(JRAPH.

may be used for nmaking simultaneous manifold copies of thesamne dispatches in each of widely separated cities. In this iian-ner each of the several newspaper conmpany subscribers in eachcity receives the identical service with the mininiuin delay, sinceeach copy received is an original. Each additional subscriber to thisservice represents no appreciable expense to the companTy, since itrequires but another receiving needle. Furtherimore, the useof the alternatingc current permits the line to be used quadruplexat very rapid speeds, that is, four entirely different dispatchesnay be sent over one wire at the same time, two in each direc-tion, and any number of copies of one or all the dispatches maybe received independently at the same time.

In addition to the above it is practicable to enmploy the line fora system of the ordinary quadruplex telegraphy at the samie time.In trial experiments in the laboratory, particular instructions weregiven to the operator of the Morse instruinent to observe if pos-sible whein messages were being, sent by the alternating current,and absolutely no effect was detected.The objection may be urged, that it is already difflicult to handle

the business at the present rate of operation of the Wheatstonesystem, and if the inistruments worked faster it could not behandled. This objectioin is undoubtedly a real one in some cases,and it is partly this fact which indicates that it may be easier toinaugurate new methods than to attempt to adapt the new rapidtransmitters to the present methods.A telegraph office of the future will probably present a differ-

ent appearance frotn that which inav now be seen in any of thelarge cities. At presenit in operating the Whieatstone system inthis country, sending to long distances at the rate of 150 to 900words per minute, both those who prepare the sending tape andthose who translate the receiving tape are employes of the tele-graph comiipany and are near the sending and receiving instru-mients. If it requires about ten men to prepare tape, and asiany imiore to translate it for a single instrunment operatinig at150 words per minlute, it will require twenty tinies this workingforce for one of the rapid machinie transrnitters. Evidentlychanges would be required in the present methods to handle thisbusiness.

It is thought that a telegraph company of the future will fulfila somewhat different function from the present ones. The com-pany will own its wires and rights of way as now, but the tend-

1897-.3] 119

CREBORE AND SQUlER

ency of the offices proper will be to transmit and receive lettersalready prepared ratlher than to undertake the preparation of theletters as well. The incomne of the company will be derivedfrom the rent of its lines at a fixed price per minute, or a fixedprice per lhundred words. The service of the telegraphl officethen becomn-es like that of the post office, its duty being to receiveand deliver letters already prepared, as the post office does. Thedifference between the two offices is in the manner in whichl thisis accomplislhed. The telegraph office becomes a post office whichiemplovs an electric current in a copper wire to carry its lettersinstead of a railroad train. The advantages in point of speed ofdelivering letters by the former method are apparent. Insteadof requiring twenty-fouir lhours to deliver letters between N,ewYork and Chicago, it will require but a few hours at most, andmake it possible to receive a replv the same day. It is probablethat sueh a system would take inore business from the presentpostal system than any otlher; for whlen telegraph letters can besent at reasonable rates comnparable with postaoe, in a few hoursinstead of mnany days, a certain amount of present post office busi-ness will be diverted. AMore than this, wlhen business can be donewith greater facilitv than at present, the total volume of businesswill undoubtedly be increased, because transactions may takeplace in a day which formnerly required a week.

It would be to the interest of such a coinpany to seek that classof less urgent business now done by correspondence, rather thanthe class handled by thie present telegraph companies, where thehiighest speed of delivery is expected. If one trunk line becomesestablislied betweeni large business centers, it will draw businessfrom a surrounding area. For instance, if a line were establishedbetween New York and Chicago, and a person in Albany desiredto coMmunicate with Chicago or points beyond, it would bequicker to send the letter to New York for transmission over thetrunk line to Chicago, and then bv rail to its destination, than tosend directly by rail fromn Albany. With a few trunk lines insuiecessful operation it would not be long before they would bemultiplied.

It is understood that these telegraph letters are sent by mail inenvelopes in the usual manner, except that the envelope containsthe prepared mnessage ready to be sent through the transmitter,and thus the telegraph office becoiies relieved of the preparationof the letters which is not strictly a part of its business. When

120 [A&pr. 21,

ON I'llE SYNCHRONOG1APtA.P12

the system comes into general use, business offices will bave theirown perforators, and it will become necessarty for the operator tolearn the telegraph alphabet as a part of his preparation as asteilographer and typewriter. The three-key perforating machineis comparatively inexpensive, but undoubtedly a machine couldb)e devised at an early date, as an attaclhment to the present type-writer, for the purpose of perforating letters at the same timethat they are being written by the typewriter in the usual way.This could be constructed to operate by the use of electro-inagnets, aild can be attached to a typewriter withouLt interferingin any way with its operation. No extra power would be re-quired, for this can be derived from an electric current whichoperates the attachment. The writing nmay be perforated at thepresent rate of speed of typewriting without the operator havingany knowledge of the telegraph alphabet as far as perforating isconcerned. This machine will cost more than the three-keyperforator, but it would in a short time more than pay for thedifference in cost on account of the great gain in speed, and also.because it prints a copy of the letter wlhich may be kept on file.Before these perforators are introduced into comnmon use it willbe necessary to establish offices in the immediate vicinity of theterminals of the trunk lines, to prepare letters for persons fur-nishing printed or written copy, as well as to furnish a printedtranslationi when desired of letters received from the centraloffice. The opportunity to obtain a cheaper rate for preparedletters will act as an inducenment to those emiploying a stenog-rapher to add a peiforator to their offices.

Coneerniing the daily correspondence of the large businesshouses between cities which are the terminals of the trunk lines,it mnight be an advantage for them to have exclusive use of theline for a certain number of mDinutes daily at a certain fixed timeof day, by subscribing and paying an annual rental to the com-pany. Knowing definitely at wlhat lhour the mail would be dis-patched daily, it would then be possible for each house to sendby messenger its daily mail already prepared for transmissionto the geineral transmission office, where it could be placed inboxes prepared for the stibscribers, to be taken out and trans-mitted when its time arrives. The distribution at the receivingend of the liine could be accomplished as now by the regular mailservice.

In the limnited uise of rapid automatic intelligence transmissionat present, the sending and receiving records are made uipon

1897.] 121

CREHORE AND SQU1ER

prepared paper in the formni of tape. fn the larger volume ofbusiness which is being considered here, it does not seem certainthat tape would be the best form for the sending and receivingpaper. It would be an advantage to have the letters receivedulpon sheets of paper with the dots and dashes arranged in parallel lines. Besides facilitatinig the readiig. this form wouild be-more convenient for mailing. lt would also easily pernmit refer-ence to any part of the letter at a glance. The amount of paperrequired by the use of sheet formi inistead of tape would be re-duced, which is an item of importance where suclh a volume oflbusiness is being handled. Sending and receiving from the sur-face of a cylinder seems entirelv practicable.Another point which must be considered is whether with these

systems, the induced currents froimi nieighboring wires along theline or from aniy other cauise will affect the legitimnate signalsmaterially, as has been at timies the case with the Wheatstonesystem. In reply to this it can be said that these receivers fortelegraphy are not necessarily more sensitive to smiall currentsbeeause they are rapid. On the contrary, thev may be made toreqnire as m-uch current as is found desirable to rid them of tlheeffects of outside influences, and at the same time retain theproperty of quick action in response to cuirrents of the propermagnitude. In this connection it may be said that the utility ofa single line wire becomnes so great that more attention will begiven in the future to the line construction and mainitenance. Ifmlillions of dollars are invested in the construction of a singlerailroad, is it not as necessary to make the telegraph liines whichcarry iinportant and profitable business as perfect in their con-struction?The telegraph line of the future will comprise substantial poles

carrying a few copper wires worked to their full capacity fortranismiiitting electric signals. The cost of maintenance of sucha line when once constructed will be little more than for an or-dinary iron wire now uised, while its carryinig capacity for intel-ligence at 3,000 words per minute simplex will be about equal to160 wires used for hand transmission simplex. By duplexingthe line, the carrving capacitv is doubled and becomes 6,000words per minute, wliich is about equal to 160 wires workedduplex, or to 80 wires worked by hand quadruplex.

It is thought that the influence whlicli the inauguration of atelegraph letter system would have uipon the existing telegraph

[AXpr. 21,N122

ON I'HE SYNCHRONOGRA PH.

and telephonie business would be to increase rather than diminishit. Each of these services has its own special field of usefulnessbut little affected by the others. A new field would be occupiedrather than an old field supplanted. The present telegraph andtelephone would still hlave their natural field of operation. eventhough the best hopes for a telegraph letter service are realized.A single line capable of sending 6,000 words per rminute

between New York and Chicago, becomes a different kind of in-vestment from a long distance telephone line where the numnberof words per mninute with the fastest rate a speaker can talk isvery slow in comparison, and the charge is $9 for five minutes'use of this line.The application under government control of a rapid system

of correspondence transmission such as has been outlined, operat-ing in conjunction with the present postal system, by supple-menting and relieving their service could hardly fail to prove ofbenefit to the people of the United States. This comes withinthe proper duty of the Post Office Department, and would beunder the direct control of the Postmaster-General. The simpli-fication in operation and expense which would result from unitingdirectly with the general post offices of large cities the telegraphletter service would soon be realized by the people and a betterservice insured.As a practical means toward ultimately assuming the direct re-

sponsibility of this new service, it would probably be easy to se-cure private companies which would be willing, to contract withthe Post Office Department to transmit telegraph letters at a fixedrate for a term of years. In this manner the Department couldgradually absorb this branch of its business and be relieved of anysudden new responsibility and radical reorganization.

It is not thought that the development of a rapid initelligencetransmnission service to the extent suggested could be accomplishedbefore many years, nor indeed that the manner or means of thisdevelopment should closely follow the lines indicated, but thatsomething analogouts to this development seems among the pos-sibilities if not the probabilities of the near futuire.The persistent efforts of Mr. Delany amid the great systein

which he has developed are well known, and tlhe ideas whichl hehas advanced in regard to the applications of rapid systemsare in the main in accordance with those stated herein.

1897.] 123