The Victorian Global Positioning System by Trudy E

The Victorian Global Positioning System by Trudy E. Bell

TARS GLEAMED OVERHEAD in a velvet-black With this vivid account, published immediately in the sky on a crisp October night in 1848 as six astrono- Cincinnati Observatory’s periodical The Sidereal Messen- mers entered the darkened dome of the three- ger, Mitchel affords an eyewitness’s glimpse into one of the year-old Cincinnati Observatory. Seating them- great untold stories in the histories of telegraphy, as- s selves around a table laden with several large tronomy, and geodesy: the telegraphic method of deter- batteries and clocks, Ormsby McKnight Mitchel, the mining longitudes. observatory’s director, was keenly aware of the moment’s Within a few short years, the telegraph had transformed historic significance. About 9 p.m., the men were joined by a both positional astronomy and geodesy. The telegraphic telegraph operator who, Mitchel wrote, method of determining longitudes held sway for eight decades not only in the . . . perfected the necessary arrange- United States but also in Europe—be- ments for communicating directly by ing replaced only in the 1920s by radio telegraph between the Cincinnati The telegraph (wireless) positioning techniques. Yet Observatory and the Philadelphia its history appears to have been largely [Central High School] Observatory. was as revolutionary overlooked, including in two recent At ten o’clock the way offices along for determining popular bestsellers on the measure- the line were closed, and the line of ment of longitude and on the telegraph.1 the telegraph given up to the use of longitude on land the astronomers. The novelty of the as the marine WHY A DEMAND FOR operations excited the deepest inter- chronometer was LONGITUDES? est among all who were present. … In the fewer than four years between At a quarter past ten, the com- for finding longitude the installation of the first electromag- pliments of the astronomers in the at sea. Along the netic telegraph (May 1844) and the cry two observatories were passed. Cin- of “Gold!” from Sutter’s Mill (January cinnati then asked Philadelphia, way, the telegraph 1848), the already-vast continental “Are you all ready?” The answer also reformed United States grew another 50 percent came back instantly, “Aye! Aye! by several fast, enormous gulps of terri- O.K., go ahead.” Cincinnati then di- observational tory. By the time of the Cincinnati-Phila- rected Philadelphia to send her clock astronomy and delphia longitude determination, the na- signals. The answer was—“All right, nationwide tion spanned from Atlantic to Pacific look out for gong signal”—the sig- with essentially the same boundaries it nal well understood by both parties, time-keeping. has today—most of it trackless wilder- and which precedes each important ness. communication. … All was instantly Now, it’s tough to develop, defend, quiet. In about half a minute, the well mine, govern, or tax millions of square known gong signal sounded, and after the lapse of miles of uncharted land. So the U.S. Army, state governmen- half a minute more, we heard the swing of tal agencies, and commercial concerns hired astronomers, to- [Philadelphia’s] sidereal clock pendulum, every vi- pographers, and geologists to mark legal boundaries, map bration giving us a tick. And thus did it swing, beat- natural resources, ascertain military routes, divide arable ing seconds for fifteen consecutive minutes. lands into salable township squares, and survey routes for … Now, Philadelphia did not undertake to trans- the tracks of the new-fangled “iron horse” locomotives. More- port her sidereal clock to Cincinnati, that its beats over, with all the new territories came thousands of miles of might be compared with our mean solar chronom- uncharted coastlines, whose collective hazards annually eter, but with the potent (I almost said omnipotent) wrecked scores of ships with uncounted losses of cargo and aid of magnetism, this comparison is virtually and human life. Accurate maps were badly needed of harbors, absolutely effected at the distance of seven-hundred shoals, and inland waterways—the mission of the U.S. Coast miles. Survey, chartered by President Thomas Jefferson in 1807.

14 SPRING 2002 T H E B ENT OF TAU BETA PI In short, in the burgeoning westward- local time as determined by meridian tran- expanding nation, latitude and longitude sits of the sun or stars. This was not just determinations were big business. a matter of packing an additional pocket Determining latitude was straightfor- watch. Chronometers were bulky and ex- ward. At the instant a star of known ce- pensive precision instruments, carefully lestial position transited (crossed) the lo- cushioned in bread-loaf-sized waterproof cal meridian (imaginary north-south line wooden boxes that had to be hand-carried passing through the observer’s zenith), an when on the backs of spirited horses. As observer measured its altitude above the insurance, field astronomers, surveyors, local southern horizon—or, after the intro- and topographers routinely toted three or Figure 1 duction of the Talcott method in the 1830s, more chronometers—some actually took A typical 19th-century portable its angular distance down from the local more than a dozen. Still, despite every astronomical transit telescope zenith. A lone observer could do with sat- pain, they often found their chronometers for measuring latitudes and isfactory accuracy even in the remote had “tripped” because of temperature longitudes from the sun and wilds with a portable transit telescope (a changes or jostling, jeopardizing the pre- stars was mounted so as to confine its observations to the telescope constrained to move only in the cision of painstaking longitude calcula- meridian—the imaginary line plane of the observer’s meridian [Fig. 1]). tions. running from due north to due Determining longitude, however, was The telegraph transformed all that. south through an observer’s challenging. Because of the rotation of the The seemingly-instantaneous speed of zenith. The main telescope tube earth, a difference in longitude is equiva- what Mitchel heralded as the telegraph’s AA was supported on a lent to a difference in local time: that is, “omnipotent magnetism” effectively an- perpendicular axis BB that the difference between the instant either nihilated distance, allowing two clocks 700 rested in Y’s fixed to the the sun or a star crossed the local merid- miles apart to be compared virtually as if uprights CC as part of a single- ian of an unknown place and the instant it they were sitting side by side, in what to- piece stand of cast iron. Near crossed the local meridian of a reference day would be called “real time.” the right end of the axis was a graduated circle G that turned location. Lured by that beckoning promise, with the axis; the circle’s Now, recall that when the telegraph whither the telegraph went, field astrono- angular measurements, was invented, time zones and standard mers followed, from the first land line to indicating the altitude to which time were still 40 years off in an unknown the first transatlantic cable—just as the telescope was elevated, future. In the 1840s, clocks were set to quickly as arrangements could be made were read by a stationary each town’s own individual local solar and transportation allowed. vernier H. time—that is, local noon was the moment At the telescope’s focus the real sun transited the local meridian MANUAL TIME-COMPARISON were crosshairs [see Fig. 7] that of some astronomical observatory or sig- TECHNIQUES at night were faintly illuminated by the lamp M to make them nificant landmark in town.2 Thus, longi- The telegraphic method of determining visible against the dark sky. tude differences were equivalent to differ- differences in longitudes was actually sev- ences in the time on local clocks separated eral methods, all of which evolved within [Elias Loomis, An Introduction to by some distance east and west. the telegraph’s first five years. There were Practical Astronomy (New York: Until the advent of the telegraph, how- three techniques for measuring differences Harper & Bros., 5th edition, ever, there was no direct, instantaneous in local times, which subsequently became 1863), p. 40] way to compare clocks at places hundreds standard to use in tandem. of miles apart. The only method was indi- rect, cumbersome, and subject to lengthy EXCHANGE OF CLOCK SIGNALS delays: physically transporting a chro- The earliest technique was a direct, nometer reading the local time of a refer- manual comparison of two chronometers ence longitude to a place of unknown set to the local solar time of the cities at longitude and comparing its time with the each end of the telegraph line. It relied on

SPRING 2002 T H E B ENT OF TAU BETA PI 15 skilled observers’ eyes, ears, hands . . . and dard for quickly determining longitude patience. differences to within a second of time. On June 9, 1844—just two weeks after Samuel F.B. Morse’s inaugural Biblical EXCHANGE OF STAR SIGNALS quotation “What hath God wrought?” After Wilkes’ pioneering proof of the con- flashed 40 miles along the first experimen- cept for the U.S. Navy, the U.S. Coast tal telegraph wire between Washington, Survey took the lead. Within months of DC, and Baltimore—Commodore (later Wilkes’ experiment, the survey’s super- Admiral) Charles Wilkes and an associate intendent, Alexander Dallas Bache—who stepped into the same two telegraph of- grew to be a towering figure in 19th-cen- fices. Wilkes, famous for recently leading tury American science—officially indi- the U.S. Exploring Expedition 87,000 cated his interest in the possibilities of miles around the Pacific Ocean (1838-42), the method. But it was another two years was one of the nation’s foremost chrono- before enough commercial telegraph wire metric experts; as former superintendent had been installed to offer actual oppor- of the U.S. Navy’s Depot of Charts and tunities. From then on, as soon as com- Instruments (predecessor of the U.S. mercial telegraph lines linked offices in Naval Observatory and the Department major cities, the Coast Survey paid to of Hydrography), he had established rig- have additional wires extended to each orous, standardized procedures for rating city’s astronomical observatory; it then marine chronometers. Now, in this first- contracted with local astronomers (or Figure 2 ever telegraphic-longitude experiment, sent its own trained observers equipped A Morse register was the Wilkes set one man’s chronometer accu- with the survey’s own transit telescopes) standard apparatus at 19th- rately to the local time of the Battle Monu- to determine differences in longitude. century telegraph offices for ment in Baltimore and the other to the The first opportunity arose in the fall recording messages. When a local time of the Capitol building in Wash- of 1846. The Coast Survey arranged to telegraph key was depressed, ington. an electromagnet mm link the U.S. Naval Observatory in Wash- attracted the armature a, For three days, Wilkes and his associ- ington with the Philadelphia Central causing lever l to pop ate took turns. At one end of the telegraph High School Observatory and a tempo- upward and press steel pen line, one watched his chronometer and rary astronomical observatory in Jersey point s against the fillet tapped the telegraph’s battery key every City, NJ, on the western bank of the (paper tape) p running 10 seconds in time to the beats of its pen- Hudson River. For this experiment, the uniformly through the dulum. Forty miles away at the other end aim was to determine differences in lon- register at about an inch a of the line, the second observer noted on gitude not only by comparing chronom- second. When the telegraph the face of his own chronometer the in- eters, but also by using stars themselves key was released, the spring- stant he heard the corresponding click of loaded lever popped back as keepers of each site’s local time. Al- down to its resting position, the armature of his receiving electromag- though technical difficulties prevented releasing the pen point from net, doing his best to estimate fractions Jersey City from making it into the loop, the fillet. In this way, dashes of a second. At the end of several minutes on October 10 and 22, 1846, Philadelphia or dots and spaces could be of 10-second signals, the first observer and Washington were finally in success- recorded quite rapidly on would communicate the exact local hour ful communication, and clock signals were the moving fillet. (In the 19th and minutes at which the beats had been exchanged to compare the approximate century, the design of the sent. Assuming an instantaneous trans- local times (just as Wilkes had done two Morse register went through mission time, subtracting the two local years before). many variations to improve times would give the minutes and seconds Then for the first time, on the evening speed, accuracy, and difference of longitude between the two reliability, so several different of October 10, an astronomer tapped a designs may actually have locations. Then it was the turn of the sec- telegraph key each time a pre-selected been used for early longitude ond observer to send similar clock signals star near the zenith was seen to cross determinations.) back to the first, ad infinitum. The mean each of the seven vertical wires of all the observations was taken to be the (crosshairs) in the eyepiece of a transit [Charles T. Chester, “On the exact longitude. After three days of ob- telescope. The local solar time of each sig- Electric Telegraph of Prof. servations, Wilkes calculated that the nal was noted at both observatories, the Morse,” American Journal of Battle Monument of Baltimore was one time difference yielding the longitude dif- Science and Arts, 2nd series, minute 34.87 seconds east of the Capitol ference. vol. 5, no. 13 (January, 1848), in Washington.3 p. 56] That technique of “sending star sig- Although some later 19th-century writ- nals,” however, was subject to any errors ers dismissed Wilkes’ experiment as in the celestial position determined for “crude,” his technique—dubbed the “ex- the star. To eliminate such errors, the change of clock signals”—became stan- star-signals technique was refined two

16 SPRING 2002 T H E B ENT OF TAU BETA PI years later. In the summer of 1848, the How did the two different astronomi- telegraph was used on seven nights to de- cal clocks increase the precision of longi- termine the difference in longitude be- tude determinations? An astronomer at tween the Harvard College Observatory one end of the line tapped a telegraph key in Cambridge, MA, and the brand new pri- each second in time to his sidereal clock, vate observatory of Lewis M. Rutherford while a second astronomer at the other in New York City. This time, pre-selected end of the telegraph line listened to the pairs of zenith stars were used, with both clicks of the armature magnet along with observatories noting the local times of the beats of his solar chronometer. When each pair of signals—in this way, differ- the two ticked in exact coincidence, the ences in local times were independent of listener noted his local solar time. After the stars’ absolute celestial positions. More- several such coincidences, the second as- over, the pairs were selected so that both tronomer began beating, and the first as- stars transited Cambridge’s meridian be- tronomer began listening. fore the first star reached New York’s me- In half an hour of monotonous beating and ridian about 12 minutes later. listening, this so-called “method of coinci- This final refinement, known as the dences” allowed the two stations to compare “exchange of star signals,” immediately their local times, and thus their longitudes, became a standard technique. accurate to within the clock errors. All three techniques—exchange of METHOD OF COINCIDENCES clock signals, exchange of star signals, and Meanwhile, a third timing technique us- method of coincidences—were first tried Figure 3 ing better telegraphic apparatus was de- in the summer of 1848 to determine the In the revolving-disk chronograph (top) designed veloped in July and August of 1847 in a differences in longitude between New by Ormsby McKnight Mitchel, this-time-satisfactory repetition of the York City and Cambridge. That October Cincinnati Observatory Washington-Philadelphia-Jersey City ex- 1848, they were used to determine the dif- director, a make-circuit clock periment. After a few nights of exchang- ference in longitude between Philadelphia marked every other second ing clock signals, the parties hit upon an and the Cincinnati Observatory 700 miles with a tiny dot. At the end of ingenious method of removing the error west, setting a new distance record. every revolution, the disk’s intrinsic in estimating fractions of a sec- By the end of the year, Bache declared position was shifted by 0.07 ond by ear: they would compare two types in his annual report of the superintendent inch. Two hours of observa- of chronometers beating at slightly differ- of the Coast Survey that the telegraphic tions could be recorded on each circular sheet, on which ent rates. method of determining longitude “may be alternate seconds appeared Astronomers use two types of clocks. considered to have passed into one of the as radial dotted lines and A solar clock counts 24 hours for the pe- regular methods of geodesy.” observations as dots riod of the earth’s rotation measured by irregularly in between the sun (essentially, the time between me- (bottom). ridian transits from one noon to the next). A sidereal clock, on the other hand, counts AUTOMATED RECORDING [Annals of the Dudley 24 hours for the period of the earth’s ro- TECHNIQUES Observatory, tation measured by the stars (essentially, But Bache’s deputy in charge of tele- vol. 1 (Albany, N.Y.: Weed, from a meridian transit of a star one night graphic longitude determinations, Sears Parsons & Co., 1866), p. 33 and opposite p. 40.] to its transit the next). Cook Walker, was troubled. He knew that But the two 24-hour days are not the all three telegraphic techniques suffered same length. In fact, the sidereal day is from two sources of human error: the three minutes 56 seconds shorter than the sender’s imperfect imitation of a clock beat solar day, because during a day the earth and the recipient’s imperfect noting of has also progressed along its orbit, chang- the beat’s arrival. Moreover, Walker was ing the position of the sun with respect to uncomfortable with there being no perma- the background stars. Thus, a sidereal nent record of a longitude determina- clock gains upon a mean solar clock one tion—a concern he mentioned to any as- second in about six minutes. So if a side- tronomer who would listen. real clock and a solar clock are placed side Now, the history of recording devices by side, they tick exactly together once for telegraphic longitudes is murky with about every six minutes, at a moment nasty controversy over credit and prior- whose time can be accurately calculated. ity of invention, which are beyond the (Actually, since most 19th-century solar scope of this brief history of measurement chronometers beat twice a second, the co- techniques. But the central technical is- incidences happened every three minutes.) sues and their major resolutions are clear.

SPRING 2002 T H E B ENT OF TAU BETA PI 17 One major question was how to con- may all be explained by the hy- nect an astronomical clock to a recording pothesis that the time of propaga- device without degrading the precise tion of the galvanic wave from the time-keeping of the clock. Harvard Col- place of the clock or star signal sta- lege Observatory director and instru- tions, to that of the receiving reg- ment-maker William Cranch Bond, Cin- ister, though small, is not quite in- cinnati Observatory director Mitchel, Cin- sensible. 5 cinnati instrument-maker John Locke, Coast Survey instrument-maker James In other words, Walker had stumbled Saxton, and various other people experi- onto the discovery that an electromag- Figure 4 mented with methods ranging from fine netic signal was not instantaneous, as In the cylindrical chronograph wires, human hairs, spider silk, and everyone had supposed, but had a finite (above) designed by Harvard sweeps of the pendulum through globs of and measurable velocity through a cir- College Observatory director William Cranch mercury to electromagnetic induction (the cuit—which he calculated to be 18,800 Bond, a cylinder covered with ultimate winner). miles per second (a speed later much a large rectangular sheet of The other primary difficulty was get- measured, debated, and revised). paper revolved once per ting the recording device itself to function Although the Morse register served minute; an ink-filled glass pen with precise uniformity. as either a primary or backup recorder marked time as a continuous Locke was the first to try using the fil- in several early longitude determina- line that was offset let (paper tape) of a Morse register stan- tions, astronomers quickly rejected it as momentarily by each tick of dard at telegraph offices for recording the ultimate solution. For one thing, the an astronomical clock. Both messages [Fig. 2]. In a seminal experi- fillet ran irregularly, depending on the cylinder’s uniform speed ment on the night of January 23, 1849, a whether or not the pen was writing. of revolution as well the paper’s horizontal advance clock that Locke designed was set up in Worse, since the fillet ran out of the reg- were controlled by a novel the Philadelphia observatory and con- ister at an inch a second, a mere hour of break-circuit machine the nected by telegraph to the respective ob- longitude determinations spewed out Bonds called a spring servatories in Cambridge, New York City, more than 300 feet of paper tape and a governor, which consisted of and Washington, DC. When an astrono- night’s work close to half a mile—imprac- a train of clockwork mer at any one observatory tapped a tele- tical for useful storage or subsequent connected with the axis of a graph key as a star transited his local analysis. flywheel (housed between meridian, the star signals were telegraphi- Cincinnati Observatory director the trapezoidal panels). The cally recorded at all four observatories. Mitchel developed a flat disk 22 inches in marked sheet (below), when taken off the cylinder, had the The result was four long rolls of paper diameter made by pasting a damp sheet minute columns nearly tape—one each at Cambridge, New York, of paper over a circular wooden hoop, vertical (being slightly spiral Philadelphia, and Washington, DC. All which dried to become as taut as a drum- on the cylinder) with the were graduated into equal parts by the head. The disk revolved horizontally once seconds marked off ticking of the Philadelphia clock and were per minute, anticipating the form of an horizontally on each minute printed with the instants the star was seen oversized 20 th-century phonograph scale. As with Mitchel’s disk, to pass each wire of the transit telescope record [Fig. 3]. At least two of Mitchel’s each 13-inch-long rectangular at each observatory. The position of each revolving-disk chronographs were placed sheet contained two hours of printed mark permanently recorded the in actual operation, one at the Cincinnati observations. fraction of a second. Observatory itself and the other at the [Annals of the Harvard College Not only was this experiment an effec- Dudley Observatory in Albany, NY. Observatory, (Cambridge: tive proof of the viability of telegraphically Meanwhile, Harvard College Metcalf & Co., 1856) vol. 1, registering astronomical observations— Observatory’s director Bond at Cam- part 1, pp. l and lii.] it also turned out to be a classic case of bridge, along with his two instrument- serendipity. When Walker measured the making sons, pursued designs for a cy- four fillets for the times of the star sig- lindrical chronograph [Fig. 4]. Eventually nals relative to the clock signals from they succeeded in overcoming two chal- Philadelphia, he was surprised to note lenges—making both the cylinder re- “small, but appreciable, differences . . . in volve and the paper advance with uni- the respective readings of the apparent form motions—using a novel break-cir- date of the same event as recorded at the cuit machine they called a spring gover- different stations.” The farther an obser- nor, which consisted of a train of clock- vatory was from the graduating clock at work connected with the axis of a fly- Philadelphia, the greater was the discrep- wheel. Ultimately, the Bonds’ cylindrical ancy. These differences, Walker reported, spring-governor design became the standard for chronographs adopted by most 19th-century observatories.

18 SPRING 2002 T H E B ENT OF TAU BETA PI THE “AMERICAN METHOD” With the addition of apparatus for permanently recording observations, the telegraphic method of determining longitudes was essentially complete. (One significant later modification was the practice of hav- ing observers switch sites midway through a series of observing nights to cancel most effects of their respective “personal equations” or reaction times.) The technique won immediate acco- lades from high places and adoption on the other side of the Atlantic. Before the Royal Astronomical Society on December 14, 1849, George Biddell Airy, the As- tronomer Royal of the Royal Greenwich Observatory, kicked off a speech by stat- ing: “The Americans of the United States, although late in the field of astronomical enterprise, have now taken up that science with their characteristic energy, and have already shewn their ability to in- struct their former masters.” Thereafter, the telegraphic method of determining longitudes became widely known as the “American method,” 6 one of the first major contributions of the American scientific community to worldwide astronomi- Figure 5 cal practice. all the various independent longitude de- Between 1844 and 1856, the terminations disagreed with one another U.S. Coast survey had DETERMINING GLOBAL by a good four seconds of time—a substan- telegraphically determined LONGITUDE tial fraction of a mile, and 20 to 30 times primary geodetic baselines By 1856, the Coast Survey had telegraphi- worse than any telegraphic uncertainty on between astronomical points of geodetic significance, had U.S. land. cally determined geodetic baselines ex- linked the American system of tending from Montgomery, AL, to New So it was with anticipation that on July longitudes with key British Brunswick, Canada [Fig. 5], most referred 27, 1866, a lone Coast Survey officer stood points in Canada, and had to the meridian of the U.S. Naval Obser- on the windswept beach of the fishing vil- plans for further expansion in vatory. Still lacking, however, was a de- lage Heart’s Content, Newfoundland, the south (dashed lines). For finitive reference of all U.S. longitude watching the end of a telegraph cable be- their fundamental reference measurements to the meridian of Green- ing hauled out of the ocean. Offshore was meridian, most observations wich, England, thereby linking astronomi- anchored the Great Eastern, which had used either the Harvard cal and geodetic determinations in the just completed its two-week voyage lay- College Observatory in Cambridge, MA, or the U.S. New World with those made in the Old. ing the 1852-mile single length of tele- Naval Observatory in This wasn’t for lack of trying. Astrono- graph cable from Valencia, Ireland. Just Washington, DC. mers on both sides of the Atlantic had as soon as the cable’s electrical connec- (In actuality, the baselines were logged decades of independent determi- tions were complete and the officer had not ruler-straight, but followed nations from half-a-dozen types of astro- ascertained that the signals were sharp, the routes of the telegraph nomical observations—such as lunar he telegraphed his superiors. Within days, wires—and some observa- occultations (eclipses of stars by the moon) the Coast Survey had dispatched astrono- tions were cross-checked and lunar and solar eclipses—whose tim- mers and transit telescopes to both sides over two different routes.) ings could be compared from both sides of the Atlantic as fast as steam ships and of the Atlantic. Even more ambitious, overland conveyances could travel. [Map designed by Trudy E. Bell from historical data in between 1849 and 1851, the Harvard Col- An exchange of star signals was Chapter IV of Elias Loomis, lege Observatory launched two Grand deemed impractical because of the The Recent Progress of As- Chronometric Expeditions, sending nearly three-hour wait for stars transit- tronomy; Especially in the scores of chronometers by ship back and ing the Valencia meridian to reach the United States (New York: forth 19 times across the Atlantic and com- meridian of Heart’s Content. Moreover, Harper & Brothers, third edi- paring them with reference chronometers the persistently miserable weather in tion, 1856)] in Boston and Liverpool. Despite all pains, both Newfoundland and Ireland—during

SPRING 2002 T H E B ENT OF TAU BETA PI 19 which “it was an event of frequent occurrence for the observer to be disturbed by a copious fall of rain while actually engaged in noting the transit of a star”—essentially precluded getting clear skies at both places the same night.7 Thus, the transit telescopes at both ends of the cables were used prima- rily to nail down local times. In addition, because of the weak voltages over nearly 2,000 miles of cable, no fillet register could automatically record the signals. Instead, the astronomers at each end had to eyeball deflections of a pencil of light thrown onto a screen by a delicate mirror galvanometer to indicate when the astronomer at the opposite end beat the seconds of his local solar clock. In short, only clock signals were ex- Figure 6 changed by eye and hand—a striking parallel to Figure 7 The telegraph literally The switchboard of the Wilkes’ pioneering first experiment in comparing Harvard College Observatory, influenced telescope— first installed in 1859, was chronometers by eye and ear 22 years earlier. specifically eyepiece—design. basically an internal telegraph From October 24 through November 20, 1866, With the “American method of system used as what today clock signals were exchanged on five nights. Ex- transits” system, an astronomer would be called a local-area periments on other nights quantified the observ- tapped a telegraph key each data network. The arrange- ers’ personal equations and determined the veloc- instant a star crossed a vertical ment, here of the upgraded ity of the signals over such an unprecedented length wire in his telescope, instead of system installed in 1871, of wire. Additional clock and star signals were ex- listening to clock ticks and shows the switchboard from changed at intermediate telegraph stations for an- counting seconds and both the top (most of the other few thousand miles down both sides of the calculating the delay of looking picture) and the side (far at a clock face before manually Atlantic from Valencia to the Greenwich Observa- right). Six pairs of wires (top) writing down the time. Because could be connected through tory and from Heart’s Content to the U.S. Naval an astronomer could tap a four switches (middle plate) Observatory. telegraph key as rapidly as a to link any combination of By this experiment, the telegraph yielded the piano key, he or she could keep three instruments (bottom); first directly-measured longitude of the dome of the eye glued to eyepiece and could the connections shown make U.S. Capitol west of the Greenwich Observatory: make timings every second or a circuit between the west five hours, eight minutes, and 2.22 seconds. two instead of needing 10 or 15 equatorial (a refracting That was the good news. The bad news was that seconds between wires. Thus, telescope of four-inch the signaling method used essentially turned the instead of being restricted to the five or seven vertical wires aperture and 60-inch focal cable into a giant capacitor, whose discharge rate length) with the south clock standard in transit telescope and the east chronograph. affected the signal transmission times. And so this eyepieces of the 1840s (upper), Moreover, financed by the U.S. transatlantic measurement, while the first word, wires could be more closely Coast Survey, the switchboard was not yet the last. spaced. As a result, 19th- allowed for temporary century astronomers made electrical connections to link WIDER INFLUENCE eyepieces with 35 or more Harvard College Observatory As ambiguous as the first transatlantic determina- wires—including special- to telegraph wires around the tion of longitude by the American method was, by purpose ones with wires at an country for geodetic 1866 the telegraph had bequeathed far more to as- angle (lower) for measuring operations. The system could tronomy and geodesy than precise longitude mea- differences in declination also transmit time signals to (celestial latitude) as well as surements. regulate timepieces. right ascension (celestial Perhaps its greatest contribution was the longitude) for star catalogues. [Annals of the Harvard College “American method of transits,” widely adopted at Observatory, (Cambridge: individual observatories in the United States and [Elias Loomis, An Introduction to Metcalf & Co., 1856) vol. VIII, abroad for nightly astronomical data recording [Fig. Practical Astronomy (New York: part I, p. Plate IV (opposite 6].8 Astronomers quickly realized the telegraph al- Harper & Bros., 5th edition, page 22)] lowed them to use not just the customary seven 1863), pp. 52 and 93] widely-spaced vertical wires in a transit telescope’s

20 SPRING 2002 T H E B ENT OF TAU BETA PI field of view, but to install dozens of finely-spaced wires, The telegraph and chronograph also sparked astrono- eliminating the need for multiple nights of observation to mers’ discussions on quantifying personal equation and re- attain the same precision [Fig. 7]. In other words, the tele- action time,9 techniques that ultimately made their way graph as electromagnetic observing assistant increased the into instrumental psychology. Lastly, the telegraph pro- precision and productivity of astronomical timings—accord- vided steady income to astronomical observatories from ing to contemporary estimates—by a factor between 4 and selling accurate local time signals—and later standard time 60, meaning observers could produce star catalogues much signals—to railroads and jewelers; by the late 19th century, faster. Moreover, by taking over timekeeping and record- such time services had paved the way for the acceptance ing functions, the telegraph effectively “de-skilled” astro- of standard time zones worldwide.10 nomical timings—meaning that positional astronomy now could be farmed out to relatively inexperienced observers.

REFERENCES 1. Dava Sobel’s Longitude: The True Story of a Lone Genius Who Solved 5. Sears C. Walker, “Telegraphic Operations of the Coast Survey.— the Greatest Scientific Problem of His Time (Walker Publishing Co., New Velocity of the Galvanic wave,” American Journal of Science, series 2, York, 1995) focuses on Harrison, inventor of the first reliable marine chro- volume 8, 1849, p. 143. nometer, ending 50 years before the invention of the telegraph. Tom 6. Airy, George B., “On the Method of observing and recording Tran- Standage’s The Victorian Internet: The Remarkable Story of the Telegraph sits, lately introduced in America; and on some other connected subjects,” and the Nineteenth Century’s On-line Pioneers (Walker Publishing Co., Monthly Notices of the Royal Astronomical Society, vol. 10, p. 26, 1849. I New York, 1998) covers the right 19th-century time period, but focuses am indebted to Professor Rand Evans, department of psychology, East exclusively on the telegraph as a traditional communications medium, omit- Carolina University, for pointing out the connection to the name “Ameri- ting all uses of what we would now call “data communications.” can method.” 2. For a complete history of local time vs. standard time, see Michael 7. The story of the first telegraphic determination of transatlantic lon- O’Malley, Keeping Watch: A History of American Time (Smithsonian In- gitude was detailed by Benjamin Apthorp Gould in “The Transatlantic stitution Press, Washington, D.C., 1990). Longitude, as Determined by the Coast Survey Expedition of 1866. A 3. A contemporary summary of Wilkes observations plus other impor- Report to the Superintendent of the U.S. Coast Survey.” Smithsonian tant early results appears in Section IV “Application of the Electric Tele- Contributions to Knowledge No. 223 (vol. XVI, article VII). City of Wash- graph to Astronomical Uses” of Chapter IV of The Recent Progress of ington: Smithsonian Institution, 1870. Astronomy; Especially in the United States (New York: Harper & Broth- 8. Marc Rothenberg, “Sears Cook Walker,” American National Biog- ers, third edition, 1856) by Elias Loomis, one of the astronomical pioneers raphy, 1999, vol. 23?, p. 514. of the American method. 9. Simon Schaffer, “Astronomers Mark Time: Discipline and the Per- 4. Quoted in The Coast Survey, 1807-1867 (Volume I of the History of sonal Equation,” Science in Context, vol. 2 (1988): 115-45. the Commissioned Corps of the National Oceanic and Atmospheric Ad- ministration), by Captain Albert E. Theberge, NOAA Corps (Ret.) 10. For a definitive history of the role of the telegraph in time services www.lib.noaa.gov/edocs/BACHE2.htm, p. 6 of 38. This 623-page online and time balls operated by 19th-century observatories, see Ian Bartky’s book, heavily footnoted, is a detailed and lively account of the Coast Sur- Selling the True Time: Nineteenth-Century Timekeeping in America vey under its first two directors, Ferdinand Rudolph Hassler and (Stanford University Press, 2000). See also O’Malley, op. cit. Alexander Dallas Bache.

Trudy E. Bell, an independent scholar and freelance science and technology journalist, has a master’s degree in the history of science and American intellectual history from New York University in 1978. Formerly an editor for Scientific American magazine (1971- 78) and a senior editor for IEEE Spectrum magazine (1983-97), she is the author of some 300 articles on astronomy, history of astronomy, engineering, scientific expeditions, and science and adventure travel—17 of which have won top journalism awards. She was the lead writer for the IEEE’s millennium book Engineering Tomorrow: Today’s Technology Experts Envision the Next Century (IEEE Press, 2000—see www.ieee.org/ products/tomorrow/) and is working on an oral-history book for the IEEE Microwave Theory and Techniques Society on the history of radio and radar since World War II (see http://206.166.222.61/MTT.pdf ). ([email protected] and http://home.att.net/~trudy.bell)

SPRING 2002 T H E B ENT OF TAU BETA PI 21