Einstein's Clocks and Poincare's Maps

by Peter Galison

  But moon culminations were notoriously elusive. Even the most skilled astronomers seemed unable to fix the moon’s zenith against the stars without significant errors and that was a huge problem. Here is why. The earth spins on its axis once per day, making the stars appear to rotate every twenty-four hours. The moon travels much more slowly against the background of the stars, about once every thirty days. So in the time the moon crosses a given angle, the stars have moved thirty times that angle, thereby multiplying any error by thirty. Piloting a ship through distant shoals with that kind of uncertainty killed sailors.73

  Because of the difficulty of “shooting the moon,” astronomer-surveyors hunting their longitude reached for other, more easily measured, heavenly events. Navigators had been using total eclipses for centuries; Columbus used one to fix his longitude on his transatlantic explorations. So when American map makers wanted to fix the longitudinal relation between Washington and Greenwich, observing the moon-darkened sun seemed promising. The United States sent a steamer to Labrador on 18 July 1860, hoping to compare the results to that same event as viewed from Spain.

  Another longstanding practice of the astronomers was to watch the moon as it arced across the night sky until a star winked out behind it. That moment of disappearance (occultation) could also be exploited to set simultaneity between distant points. Measure the time of occultation locally, look on a chart (or wait for a report by post) to find what time that same event was observed in Greenwich or Paris, then subtract one from the other. So astronomers both in the United States and Europe watched and measured with enormous care as four stars of the Pleiades dipped under and then out from behind the moon. An occultation of the planet Venus took place on 24 April 1860; astronomers huddled in wait for it in the Fredericton, New Brunswick, and Liverpool observatories. Finding longitude was part of the job description of French, British, and American astronomers, and they deployed every conceivable method to find it. But the quest to link the United States to the “well determined European observatories” remained unfulfilled. Every expedition produced a new number.74

  Mapping offered both symbolic and practical mastery over space. In the great land grab of the mid-nineteenth century, fixing positions was critical for trade, for military conquest, for laying railroads.75 When the United States plunged into civil war, the Coast Survey became a strategic asset. Congress had long before demanded that the surveys cover rivers as needed by commerce or defense. Now map makers hoped to deliver on that assignment, working closely with Union admirals in North Carolina and on the Mississippi. At first, the telegraphic surveyors, including George Dean, set about reducing data already in hand to provide precise positions of key southern military sites, determining longitude differences between them.76 Watching, measuring, calculating: the surveyors plotted Rebel emplacements around Charleston and Savannah, while a small group of reconnaissance topographers joined General Sherman on his march from Savannah to Golds-boro, Georgia. When the war ended, the surveyors began plotting a future using their wartime telegraphic surveys. They had new and better measurements of almost every major town from Calais, Maine, at the northeastern edge of the United States, down to New Orleans. Benjamin Gould (who headed the survey’s telegraphic longitude effort) and his team looked eastward to fill the one critical gap—New York City to Washington—that remained in the quest for a complete electrical map of the United States east of the Mississippi.77

  The surveyors looked to the ocean. They did so in some desperation, because no matter how hard the survey had tried, convergence on a longitude difference between Europe and the United States continued to prove infuriatingly elusive. They reviewed moon culminations, restudied data on the occultations of stars and planets, and pored over old chronometer results. But reanalyzing old data was simply not enough: “The discordance of results, which individually would have appeared entitled to full reliance, is thus seen to exceed four seconds; the most recent determinations, and those which would be most relied upon, being among the most discordant. No amount of labor, effort, or expense had been spared by the Coast Survey for its chronometric expeditions, inasmuch as the most accurate possible determination of the transatlantic longitude was specially required by law.” In addition, the latest chronometer studies differed from the best astronomical studies by an embarrassing and irreducible time difference of three and a half seconds.78 No one had any idea which result to believe.

  Only underwater electrical telegraph cables could break the impasse. Beginning in August 1857, missions had struggled against the North Atlantic to lay a telegraph line. Cables broke again and again; in June 1858, the fleet once more launched from Plymouth, England, with the vast cable. Just three days out at sea, a gale tore at the ships ceaselessly for nine days. Despite significant damage to one ship (and one sailor so terrified that he lost his reason), cable-laying continued. On 6 August 1858, the first signals finally passed through the cable, followed by a cable break and the cessation of submarine cabling during the Civil War. Pulling out from Valentia Island in southwestern Ireland in July 1865, the Great Eastern, a mammoth ship five times larger than any other, began hauling cable toward Newfoundland. Twelve hundred miles later, both the cable and the ship’s lifting gear sank into the sea. The mission was aborted.79 In 1866 another crew set out, this time aiming to run a much-improved cable from the tiny fishing hamlet at Heart’s Content, Newfoundland (on the eastern side of Trinity Bay, about ninety miles from St. John’s) to Valentia. This time: success. Communication started on 27 July 1866, and the surveyors immediately began sending time signals. Gould dispatched his longitude team in small groups to staff the dilapidated way stations up the East Coast. To inspect the telegraph line from Calais to Newfoundland, the expedition chartered a schooner, plying from Nova Scotia’s Cape Breton Island to their destination at Heart’s Content, Newfoundland. Every mile, repeaters (relaying the signal) needed inspection; they had to crash-train dozens of isolated telegraphers.80

  Gould himself set sail for Liverpool and London on 12 September 1866 by the British mail steamer Asia, first to confer with the British cable company officers and then to bring his measuring instruments to the Irish terminus. Bumping through the Irish countryside on a jury-rigged spring-cushioned cart, the astronomers hauled their precarious tower of crates forty-two miles, then ferried their goods to Valentia across the straits from Killarney. Conditions at cable’s end were discouraging. The British company refused the Americans permission to pull land lines into electrical contact with the undersea cable for fear a lightning strike would damage their fragile leash on the New World. That meant the Americans had to set their observatory adjacent to the telegraph company building at Foilhommerum Bay, “remote,” as the time men put it, “from any other dwelling-house except the unattractive cabins of the peasantry.” On this rugged terrain, the time men banged together their makeshift observatory, 11' x 23', bolted to six heavy stones buried in earth, protected from southwest gales by the adjacent telegraph house, and shielded by rising ground from northwest weather. The larger room was their observatory; the smaller, on the eastern end, became their dwelling. Their lab was simple: a rigid mount for the transit instrument, nooks for clock and chronograph; a spot for the relay magnet that would pass the signal along toward Greenwich, the Morse register, and a recording table.81

  From what Gould dubbed Valentia’s “peculiarly unastronomical sky” came rain. Bucketsfull. Clouds allowed but one or two glances of the noontime sun. When the sun did poke through, the astronomers, ever vigilant, fixed their meridian. Finally, on 14 October 1866 at 3:00 A.M., the Americans glimpsed a few stars through the haze and took transit readings. Locals reported that for the eight weeks prior to the surveyors’ arrival it had rained at Valentia every day without exception. During the seven weeks that the survey crew lived and worked in their Irish shack by the sea, they saw four days without rain and only a single clear night. “The observations were, in general, made during the intervals of showers; and it was an event of frequen
t occurrence for the observer to be disturbed by a copious fall of rain while actually engaged in noting the transit of a star.” Time sentries at the other end of the line in Newfoundland had it worse. Telegraphic Surveyor George Dean spied nothing, not a single glimpse of the sun, moon, or stars.

  Here was Victorian high technology on the ragged edge of Britain, powering their transoceanic signal with a battery composed of a percussion gun cap with a morsel of zinc and a drop of acidulated water. Weather gods permitting, the Irish station would cable “GOULD” in Morse code; Newfoundland would respond “DEAN,” followed by time signals—half-second pulses punctuated by five-second intervals. At both ends of the cable, teams hovered over instruments. After crossing the Atlantic, the signal was too weak to drive the drum recorder, so they used the mirror galvanometer, a vastly more sensitive device invented by British physicist William Thomson. A delicately suspended mirror with a tiny magnet glued to its back reflected light from a kerosene lamp. Nearby was a coil wired to the undersea cable. When a signal current coursed through the cable, the coil became an electromagnet, gently twisting the little permanent magnet with its attached mirror, which shifted the reflected light of the kerosene lamp against a mounted sheet of white paper. Even the weakest of transoceanic time signals became visible. Anticipating the signal, observers would focus the bright light of their kerosene lamp on the mirror. And wait, hour after hour, in the cold, damp night, hoping that an electric current crossing under 4,320 miles of ocean would dance a tiny spot of reflected light across a soggy paper sheet.

  The mobile astronomers strictly followed their disciplined hard-won procedures, pried from longitude campaigns that both Gould and Dean had run during the Civil War. Gould’s team received its first signal on 24 October 1866; in the weeks that followed, Gould and his telegraphers eked out four more exchanges in tiny windows of astronomical weather. An ocean from Gould, in Newfoundland, Dean’s struggle to relay the signal down to Boston was much less successful. As the cable snaked along the eleven hundred rough miles from Heart’s Content to the entry point into the United States at Calais, Maine, communication broke down “day after day, and week after week.” Nothing worked. Suddenly, on 11 December, a sharp frost gripped the defective land line coming into Calais. All at once ice admirably insulated the wire and the pulse shot like lightning from Heart’s Content to Calais. Just before New Year’s Day 1867, the Newfoundland team steamed into Boston Harbor, having bagged the longitude difference between Harvard and Greenwich observatories.82

  With the Atlantic wired for simultaneity, the tempo of electric map-making only increased. Immediately following the British-American collaboration, the French hauled a line from Brest through St. Pierre off Newfoundland to Duxbury, Massachusetts. American astronomer-surveyor Dean and his peripatetic team of surveyors deployed again, almost immediately, to begin planning with the French naval authorities for a time check. Once the Brest-Paris line was secure, the surveyors had their dream, a series of triangles that would serve as a double check of their longitude results. For example, the longitude differences ought to (and did) add to zero if one went from Brest to Paris to Greenwich and back to Brest:

  (Brest-Paris) plus (Paris-Greenwich) plus (Greenwich-Brest) = 0.83

  The French Naval Lieutenant de Bernardières was all too aware that his American and British counterparts were on the longitudinal march. In the spring of 1873, at his superiors’ instigation, U.S. Naval Lieutenant Commander Francis Green had begun sending suboceanic time signals to map the West Indies and Central America. Traveling the seas in the side-wheel steamer The Gettysburg, Green’s team brought precision to the longitude of Panama, Cuba, Jamaica, Puerto Rico, and many other islands.84 On his return in 1877, the authorities set the lieutenant commander to work again, this time to exploit the newly laid telegraphic cables crossing the Atlantic: London to Lisbon to Recife in northeast Brazil. For the first time the American Navy could ply its hydrographic trade all along the eastern coast of South America from Para to Buenos Aires. At last, the Navy boasted, it would settle the much-disputed locations of places like Fort Villegagnon (in the Bay of Rio de Janeiro), where previous longitude men had clashed over an astonishing thirty-second time discrepancy, an eight-mile uncertainty over where one’s ship would hit the eastern edge of South America.

  With the help of the French, who shot their Parisian signals down through the wires to Lisbon, the Portugese too joined the Greenwich-based map-making scheme.85 At Porto Grande off the Brazilian coast, the longitude team halted, waiting for the “sick season” to subside before they dared land at Pernambuco: “A less interesting place than Porto Grande to pass two or three weeks could not be found,” groused the Americans. “The island of St. Vincent is merely a heap of cinders.”86 Finally, the astronomers hopped a royal mail-steamer into Pernambuco and set the machine into motion: a pulse from Greenwich Observatory to Land’s End, down through 828 miles of undersea cable, out from the instrument room of the Eastern Telegraph Company near the lighthouse at Carcavellos, then back out from the Royal Observatory of Lisbon across the Atlantic, into the grounds of the Brazilian Naval Arsenal, up insulated wire over the arsenal walls, along the roof of the port captain’s office, through the esplanade, by balconies, and down to one of Thomson’s mirror galvanometers, where it would sway its beam of white light.

  Subtle, evanescent, and unreliable. Yet when that electrical pulse finally reached Rio in July 1878, it was of royal significance. Brazil’s Emperor Pedro II had been in the United States during 1876, then launched a great European tour, visiting Troy with Heinrich Schliemann, receiving communion at the Holy Sepulcher in Jerusalem, and reveling with Queen Victoria’s daughter and her husband in Vienna. In Paris the Academy of Sciences inducted Pedro as a foreign associate; Victor Hugo received him in celebration of literature. Back in London, he lunched with the Queen at Windsor Castle. Given this public life, and apparently an equally entertaining private one, Pedro returned somewhat reluctantly to Rio. Nothing, however, could keep His Majesty away from Lieutenant Commander Green’s ramshackle observatory to witness the electrical arrival of European time.87

  Today the word “observatory” may bring to mind a romantic image: a glistening hemisphere perched on the craggy peak of a mountain; the astronomer, sliding the sky slit open, swivels his great brass telescope to view the heavens, while white-coated assistants shuffle quietly at his side. Not for Green and his U.S. Naval team, who would come to town (such as it was, say, in Porto Grande) and set up shop. Finding a location with a good view of the sky near the telegraphic office was obviously key. They would then scrape a meridian (north) line onto the ground and build a cement and brick pier. On top the astroexplorers set a small slab of marble that they carried with them and, on top of that, their precious brass transit instrument, through which they could watch stars cross a spider’s thread that marked the meridian line. It would take two sailor-astronomers about an hour to assemble their portable wood observatory (8' x 8'), with a canvas top in case of rain. Into this tiny station Green crammed clocks, telegraph keys, and either a recording drum or a Thomson galvanometer to display the incoming signal. When the observer needed a telegrapher at his side, the room was, one could say, cozy.

  Having entertained Pedro II with the electromapping of Rio, the American Navy sailed on. Early in 1883, Washington ordered the team to the west coast of South America now that cables joined it too to the telegraphic system. This new expedition followed cables running from Galveston, Texas, through yellow-fever-infested Vera Cruz in southeastern Mexico. With Vera Cruz on the electric worldmap, they sailed down the western coast of South America, hoping to exploit the cable binding Salina Cruz in southeastern Mexico to Valparaiso, Chile. The team faced down quarantines at New Orleans and Galveston, only to arrive in Peru in the midst of the Chilean military occupation.

  Thanks to the intervention of the American minister to Peru, Mr. S. L. Phelps, the occupying admiral of the Chilean army promised to assist the time bearers. From Lima,
the navy men headed south on 13 October 1883, tying Valparaiso to Lima, and by early 1884 they were banging together their wooden observatory in Paita, Peru. Pounding rain drenched the desiccated town for the first time in seven years: “the soil of the place, ordinarily arid and dusty, was converted into a sickening and fetid mud . . . the whole town . . . rendered almost uninhabitable.”88 Furthermore, “operations in Peru,” the officers reported, “were at points occupied as military posts by an invading army, and the observers had to contend with the dila-toriness, and in the case of the military commandant at Arica, the indifference and stupidity, of officials,” not to speak of custom delays, fog, and broken cables. Yet on 5 April 1884, the American officers set sail for New York with their South American longitudes.89

  Figure 3.9 Portable Observatory: Bahia, Brazil. American Naval officers determining simultaneity—and therefore longitudinal differences—in order to map more precisely the Americas. Longitude expeditions by French, British, and American teams were part scientific, part military, and part exploration, the “astronomical observatories” often no more than portable wooden shacks accompanied by a few pieces of electrical telegraph equipment, some magnets, mirrors, and a surveyor’s telescope. They came at a cost: numerous surveyors died of disease and accident while trying to pin electrical simultaneity to the shores of the Americas, Asia, and Africa. SOURCE: GREEN, REPORT ON TELEGRAPHIC DETERMINATION (1877), FRONTISPIECE.