30.11 News Feat Antikythera MH

NEWS FEATURE NATURE|Vol 444|30 November 2006

IN SEARCH OF LOST TIME The ancient Antikythera Mechanism doesn’t just challenge our assumptions about technology transfer over the ages — it gives us fresh insights into history itself. Jo Marchant reports.

t looks like something from another world that such a sophisticated technology appears explains, shows the Metonic cycle — 235 — nothing like the classical statues and vases seemingly out of the blue is perhaps not that months fitting quite precisely into 19 years. that fill the rest of the echoing hall. Three surprising — records and artefacts from 2,000 The lower spiral, according to the research by Iflat pieces of what looks like green, flaky years ago are, after all, scarce. More surpris- Edmunds and his colleagues, was divided into pastry are supported in perspex cradles. Within ing, to an observer from the progress-obsessed 223, reflecting the 223-month period of the each fragment, layers of something that was twenty-first century, is the apparent lack of a Saros cycle, which is used to predict eclipses. once metal have been squashed together, and subsequent tradition based on the same tech- To show me what happens inside, Wright are now covered in calcareous accretions and nology — of ever better clockworks spreading opens the case and starts pulling out the various corrosions, from the whitish tin oxide out round the world. How can the capacity to wheels. There are 30 known gear-wheels in to the dark bluish green of copper chloride. This build a machine so magnificent have passed the Antikythera Mechanism, the biggest tak- thing spent 2,000 years at the bottom of the sea through history with no obvious effects? ing up nearly the entire width of the box, the before making it to the National Archaeological smallest less than a centimetre across. They Museum in Athens, and it shows. Astronomic leaps all have triangular teeth, anything from 15 to But it is the details that take my breath away. To get an idea of what the mechanism looked 223 of them, and each would have been hand Beneath the powdery deposits, tiny cramped like before it had the misfortune to find itself cut from a single sheet of bronze. Turning the writing is visible along with a spiral scale; there on a sinking ship, I went to see Michael Wright, side knob engages the big gear-wheel, which are traces of gear-wheels edged with jagged a curator at the Science Museum in London for goes around once for every year, carrying teeth. Next to the fragments an X-ray shows more than 20 years and now retired. Stepping the date hand. The other gears drive the some of the object’s internal workings. It looks into Wright’s workshop in Hammersmith is a Moon, Sun and planets and the pointers on just like the inside of a wristwatch. little like stepping into the workshop where the Metonic and Saros spirals. This is the Antikythera Mechanism. These H. G. Wells’ time machine was made. To see the model in action is to want fragments contain at least 30 interlocking Every inch of floor, wall, shelf to find out who had the ingenuity gear-wheels, along with copious astronomi- and bench space is covered to design the original. Unfor- cal inscriptions. Before its sojourn on the sea with models of old metal “It’s a popular notion tunately, none of the copious bed, it computed and displayed the movement gadgets and devices, from that technological inscriptions is a signature. of the Sun, the Moon and possibly the planets ancient Arabic astrolabes But there are other clues. around Earth, and predicted the dates of future to twentieth-century development is a simple Coins found at the site by eclipses. It’s one of the most stunning artefacts trombones. Over a cup of progression. But history Jacques Cousteau in the we have from classical antiquity. tea he shows me his model is full of surprises.” 1970s have allowed the No earlier geared mechanism of any sort of the Antikythera Mecha- shipwreck to be dated some- has ever been found. Nothing close to its tech- nism as it might have been — François Charette time shortly after 85 bc. The nological sophistication appears again for in his pomp. The model and inscriptions on the device itself well over a millennium, when astronomical the scholarship it embodies suggest it might have been in clocks appear in medieval Europe. It stands have consumed much of his life (see use for at least 15 or 20 years before that, as a strange exception, stripped of context, of ‘Raised from the depths’). according to the Edmunds paper. ancestry, of descendants. The mechanism is contained in a squar- The ship was carrying a rich cargo of luxury Considering how remarkable it is, the ish wooden case a little smaller than a shoe- goods, including statues and silver coins from Antikythera Mechanism has received com- box. On the front are two metal dials (brass, Pergamon on the coast of Asia Minor and vases paratively scant attention from archaeologists although the original was bronze), one inside in the style of Rhodes, a rich trading port at the or historians of science and technology, and the other, showing the zodiac and the days of time. It went down in the middle of a busy ship- is largely unappreciated in the wider world. the year. Metal pointers show the positions of ping route from the eastern to western Aegean, A virtual reconstruction of the device, pub- the Sun, the Moon and five planets visible to and it seems a fair bet that it was heading west lished by Mike Edmunds and his colleagues in the naked eye. I turn the wooden knob on the for Rome, which had by that time become the this week’s Nature (see page 587), may help to side of the box and time passes before my eyes: dominant power in the Mediterranean and had change that. With the help of pioneering three- the Moon makes a full revolution as the Sun a ruling class that loved Greek art, philosophy dimensional images of the fragments’ innards, inches just a twelfth of the way around the dial. and technology. the authors present something close to a com- Through a window near the centre of the dial The Rhodian vases are telling clues, because plete picture of how the device worked, which peeks a ball painted half black and half white, Rhodes was the place to be for astronomy in in turn hints at who might have been respon- spinning to show the Moon’s changing phase. the first and second centuries bc. Hipparchus, sible for building it. On the back of the box are two spiral dials, arguably the greatest Greek astronomer, is But I’m also interested in finding the answer one above the other. A pointer at the centre thought to have worked on the island from to a more perplexing question — once the of each traces its way slowly around the spiral around 140 bc until his death in around technology arose, where did it go to? The fact groove like a record stylus. The top dial, Wright 120 bc. Later the philosopher Posidonius set

NATURE | Vol 444 | 30 November 2006 ©

2 0 0 6 N a t u r e

P u b l i s h i n g G r o u p NEWS FEATURE 535

J. MARCHANT/ANTIKYTHERA MECHANISM RESEARCH PROJECT NEWS FEATURE NATURE|Vol 444|30 November 2006

up an astronomy school there that continued “I’m very surprised to find a mechanical Hipparchus’ tradition; it is within this tradi- representation of this,” adds Alexander Jones, tion that Edmunds and his colleagues think a historian of astronomy at the University of the mechanism originated. Circumstantial evi- Toronto, Canada. He says the Antikythera dence is provided by Cicero, the first-century Mechanism has had little impact on the his- bc Roman lawyer and consul. Cicero studied tory of science so far. “But I think that’s about on Rhodes and wrote later that Posidonius had to change. This was absolutely state of the art made an instrument “which at each revolution in astronomy at the time.” reproduces the same motions of the Sun, the Wright believes that similar mechanisms Moon and the five planets that take place in the modelled the motions of the five known plan- heavens every day and night”. The discovery of ets, as well as of the Sun, although this part

ANTIKYTHERA MECHANISM RESEARCH PROJECT RESEARCH MECHANISM ANTIKYTHERA the Antikythera Mechanism makes it tempting of the device has been lost. As he cranks the to believe the story is true. gears of his model to demonstrate, and the And Edmunds now has another reason to days, months and years pass, each pointer think the device was made by Hipparchus or his alternately lags behind and picks up speed to followers on Rhodes. His team’s three-dimen- mimic the astronomical wanderings of the sional reconstructions of the fragments have appropriate sphere. turned up a new aspect of the mechanism that is both stunningly clever and directly linked to Greek tragedy work by Hipparchus. Inside out: computer tomography of the main Almost everyone who has studied the mecha- One of the wheels connected to the main fragment allowed for accurate modelling. nism agrees it couldn’t have been a one-off — it drive wheel moves around once every nine would have taken practice, perhaps over several years. Fixed on to it is a pair of small wheels, closely mimics the varying motion of the Moon generations, to achieve such expertise. Indeed, one of which sits almost — but not exactly around Earth, as described by Hipparchus. Cicero wrote of a similar mechanism that was — on top of the other. The bottom wheel has When the Moon is close to us it seems to move said to have been built by Archimedes. That a pin sticking up from it, which engages with faster. And the closest part of the Moon’s orbit one was purportedly stolen in 212 bc by the a slot in the wheel above. As the bottom wheel itself makes a full rotation around the Earth Roman general Marcellus when Archimedes turns, this pin pushes the top wheel round. about every nine years. Hipparchus was the was killed in the sacking of the Sicilian city of But because the two wheels aren’t centred in first to describe this motion mathematically, Syracuse. The device was kept as an heirloom the same place, the pin moves back and forth working on the idea that the Moon’s orbit, in Marcellus’ family: as a friend of the family, within the upper slot. As a result, the move- although circular, was centred on a point offset Cicero may indeed have seen it. ment of the upper wheel speeds up and slows from the centre of Earth that described a nine- So where are the other examples? A model down, depending on whether the pin is a little year circle. In the Antikythera Mechanism, this of the workings of the heavens might have had farther in towards the centre or a little farther theory is beautifully translated into mechanical value to a cultivated mind. Bronze had value for out towards the tips of the teeth (see illustra- form. “It’s an unbelievably sophisticated idea,” everyone. Most bronze artefacts were eventu- tion on page 551). says Tony Freeth, a mathematician who worked ally melted down: the Athens museum has just The researchers realized that the ratios of the out most of the mechanics for Edmunds’ team. ten major bronze statues from ancient Greece, gear-wheels involved produce a motion that “I don’t know how they thought of it.” of which nine are from shipwrecks. So in terms Raised from the depths In 1900 a party of Greek sponge shrivelled, the lump cracked open, data (much to the annoyance of Curator Michael Wright realized divers sought shelter from a storm and on 17 May 1902, archaeologist Karakalos and his wife), and his the device was a Byzantine sundial in the lee of the barren, rocky islet Valerios Stais noticed that there reconstruction was unnecessarily from the sixth century ad, which of Antikythera. Once the winds were gear-wheels inside. complicated — perhaps too also contained a simple geared had eased, Elias Stadiatis dived 42 The gears elicited interest, but it complicated for historians and mechanism that drove pointers metres to a rocky shelf to look for was not until investigations delved archaeologists. They largely ignored showing the position of the Moon late additions to his hard-earned beneath the surface that the box Price’s work, and he died in 1983. and Sun in the sky. Studying the haul. Instead of sponges nestled started to yield its secrets. The That same year, a Lebanese man astronomically enhanced sundial on the sea bed, the shape of a great British science historian Derek de walked into the Science Museum in led Wright to Price’s treatment of ship loomed out of the blue. After Solla Price and the Greek nuclear London with the pieces of another the Antikythera Mechanism, in grabbing the larger-than-life arm physicist Charalampos Karakalos ancient mechanism in his pocket. which he saw serious holes. of a bronze figure as proof of his made X- and gamma-ray images find, he returned to the surface of the fragments in 1971. Karakalos to inform his companions. The and his wife Emily painstakingly Antikythera wreck was to yield a counted the visible teeth; in 1974 Rome Byzantium stunning collection of bronze and Price published a heroic 70-page Pergamon marble statues, pottery, glassware, account of the machine (D. de S. GREECE Athens jewellery and coins; it was also to Price Trans. Am. Phil. Soc., New Ser. Peloponnese claim the life of one of the divers, not 64, 1–70; 1974). Syracuse Rhodes yet aware of the risk of the bends “Price really put the mechanism when diving with an oxygen hose. on the map,” says Tony Freeth, Kythera Mediterranean Sea As busy museum staff struggled co-author of a new reconstruction Antikythera to piece together statues and of the device (see page 587). “He vases, a formless, corroded lump understood the essence of what it Crete of bronze and wood lay unnoticed. was — an astronomical computer.” But as the wood dried and But Price massaged some of the

NATURE|Vol 444|30 November 2006 NEWS FEATURE

of the mechanism, “we’re lucky we have one”, Model success: points out Wright. “We only have this because Michael Wright

it was out of reach of the scrap-metal man.” devoted his life A. WRIGHT But ideas cannot be melted down, and to decoding and although there are few examples, there is replicating the some evidence that techniques for modelling Antikythera the cycles in the sky with geared mechanisms Mechanism. persisted in the eastern Mediterranean. A sixth-century ad Byzantine sundial brought to Wright at the Science Museum has four surviving gears and would probably have used at least eight to model the positions of the Sun and Moon in the sky. The rise of Islam saw much Greek work being translated into Arabic in the eighth and ninth centuries ad, and it seems quite possible that a tradition of geared mechanisms continued in the caliphate. Around ad 1000, the Persian scholar al-Biruni described a “box of the Moon” very similar to the sixth-century device. There’s an Ara- Padua clock) the timing of eclipses. The time- them, came west at the same time. As François bic-inscribed astrolabe dating from 1221–22 telling function seems almost incidental. Charette, a historian of science at Ludwig currently in the Museum of the History of Sci- It could be argued that the similarities Maximilians University in Munich, Germany, ence in Oxford, UK, which used seven gears to between the medieval technology and that of points out, “for the translation of technology, model the motion of the Sun and Moon. classical Greece represent separate discover- you can’t rely solely on texts”. Most texts leave But to get anything close to the Antikythera ies of the same thing — a sort of convergent out vital technical details, so you need skills to Mechanism’s sophistication you have to wait clockwork evolution. Wright, though, favours be transmitted directly. until the fourteenth century, when mechanical the idea that they are linked by an unbroken But if the tradition of geared mechanisms to clockwork appeared all over western Europe. tradition: “I find it as easy to believe that this show astronomical phenomena really survived “You start to get a rash of clocks,” says Wright. technology survived unrecorded, as to believe for well over a millennium, the level of achieve- “And as soon as you get clocks, they are being that it was reinvented in so similar a form.” The ment within that tradition was at best static. used to drive astronomical displays.” Early timing of the shift to the West might well have The clockwork of medieval Europe became examples included the St Albans clock made been driven by the fall of Baghdad to the Mon- more sophisticated and more widely applied by Richard Wallingford in around 1330 and a gols in the thirteenth century, after which much fairly quickly; in the classical Mediterranean, clock built by Giovanni de’Dondi a little later of the caliphate’s knowledge spread to Europe. with the same technology available, nothing in Padua, Italy, both of which were huge astro- Shortly after that, mechanical clocks appeared remotely similar happened. Why didn’t anyone nomical display pieces with elaborate gearing in the West, although nobody knows exactly do anything more useful with it in all that time? behind the main dial to show the position of where or how. It’s tempting to think that some More specifically, why didn’t anyone work out the Sun, Moon, planets and (in the case of the mechanisms, or at least the ability to build earlier what the gift of hindsight seems to make

Wright ended up working taken under light from a wide the varying speed of the Moon with Allan Bromley, a computer variety of directions, to reveal more through the sky (see main story). scientist at Sydney University of the inscriptions. They enlisted The inscriptions are also M. KIRK in Australia who had become experts in computer-assisted revealing novelties, although interested in the Antikythera tomography from British firm XTEC, deciphering them is hard work: Mechanism at around the same which developed a new machine some of them are less than 2 time. Bromley wanted to study the just for the Antikythera project. millimetres high, and there are no machine with X-ray tomography, In autumn 2005, the Hewlett- spaces to show where each word which assembles a sheaf of cross- Packard equipment and all 12 starts and finishes. Agamemnon sections of its subject. As the tonnes of XTEC’s machinery were Tselikas, director of the Centre fragments could not be moved shipped to the museum. The for History and Palaeography in from the museum, and Bromley results have allowed the team to Athens, spent a concentrated three didn’t have the money to ship a confirm many of Wright’s ideas, months trying to decipher the tomography machine to Athens, and extend them. “My main fear wording, working from late at night Wright used his tool-making skills initially was that we’d throw all this into the early hours of the morning: to build a crude tomograph in situ. technology at it and we wouldn’t “I needed the silence.” The two researchers took around do more than dot the i’s and cross So far Tselikas and his colleague 700 images of the fragments, Derek de Solla Price tried to undo the the t’s,” says Freeth. “But we got Yanis Bitsakis have more than and Wright has been working on Antikythera Mechanism’s secrets. more out of it than I dared hope.” doubled the number of legible a reconstruction that supercedes One major new result came characters on the mechanism, Price’s ever since. documentary. But their efforts as much from chance as from which seem to form a manual that In the meantime, Mike Edmunds, soon turned to discovering more technology; a key section of a dial explains how the mechanism was an astrophysicist at Cardiff about how the device worked. They found sitting unnoticed in the to be used. It takes “imagination University, UK, and his friend Tony contacted Hewlett-Packard, which museum’s store room helped reveal and intuition” to decipher the Freeth, a mathematician-turned- had developed a method for reading that one of the dials was used inscriptions, says Tselikas. “We film-maker living in London, eroded cuneiform tablets that to predict eclipses. Another big are just starting to penetrate decided the mechanism would involved building up a composite discovery was the identification of a the mentality of the user of this make a fantastic subject for a computer image from pictures ‘pin and slot’ mechanism to model machine.” J.M.

NEWS FEATURE NATURE|Vol 444|30 November 2006

Gearing up: this to mechanical clocks, a geared mechanism reconstruction shows, needs to be powered by something other than among other things, a person; it was not until medieval Europe that the offset wheels of clockwork driven by falling weights makes an the Moon’s nine-year appearance. cycle (lower left). A labelled diagram is on Invention’s evolution page 551. Bert Hall, a science historian at the University of Toronto in Canada, believes a final break- through towards a mechanical weight drive might have come about almost by accident, by ANTIKYTHERA MECHANISM RESEARCH PROJECT RESEARCH MECHANISM ANTIKYTHERA adapting a bell-ringing device. A water clock could have driven a hammer or weight mechanism swinging between two bells as an alarm system, until someone realized that the weight mechanism would be a more regular way of driving the clock in the first place. When the new way to drive clocks was discovered, says Hall, “the [clockwork] technology came rush- ing out of the wings into the new tradition”. Researchers would now love further mechanisms to be unearthed in the historical record. “We hope that if we can bring this to people’s attention, maybe someone poking around in their museum might find something, or at least a reference to something,” says Edmunds. Early Arabic manuscripts, only a fraction of which have so far been studied, are promising to be fertile ground for such discoveries. Charette also hopes the new Antikythera reconstruction will encourage scholars to take the device more seriously, and serve as a reminder of the messy nature of history. “It’s still a popular notion among the public, and among scientists thinking about the history of their disciplines, that technological development is a simple progression,” he says. “But obvious — that clockwork would be a good sophistication — they imported philosophers history is full of surprises.” thing to make clocks with? for centuries. In the meantime, Edmunds’ Antikythera Serafina Cuomo, a historian of science at Seen in this light, the idea that the Antiky- team plans to keep working on the mecha- Imperial College, London, thinks that it all thera Mechanism might be expected to lead nism — there are further inscriptions to be depends on what you see as ‘useful’. The Greeks to other sorts of mechanism seems less deciphered and the possibility that weren’t that interested in accurate timekeeping, obvious. If it already embodied more fragments could be found. she says. It was enough to tell the hour of the the best astronomy of the This week the researchers day, which the water-driven clocks of the time time, what more was there “I find it as easy to are hosting a conference in could already do fairly well. But they did value to do with it? And status Athens that they hope will know ledge, power and prestige. She points out symbols do not follow believe that this yield fresh leads. A few that there are various descriptions of mecha- any clearly defined arc of technology survived minutes’ walk from the nisms driven by hot air or water — and gears. progress. What’s more, unrecorded, as to National Archaeologi- But instead of developing a steam engine, say, the idea that machines cal Museum, Edmunds’ the devices were used to demonstrate philo- might do work may believe that this was colleagues from the sophical principles. The machines offered a have been quite alien to reinvented in so similar University of Athens, deeper understanding of cosmic order, says slave-owning societies a form.” — Michael Yanis Bitsakis and Xeno- David Sedley, a classicist at the University of such as those of Ancient phon Moussas, treat me to a Cambridge, UK. “There’s nothing surprising Greece and Rome. “Perhaps Wright dinner of aubergine and fried about the fact that their best technology was the realization that you could octopus, and explain why they used for demonstrating the laws of astronomy. use technology for labour-saving would one day like to devote an entire It was deep-rooted in their culture.” devices took a while to dawn,” says Sedley. museum to the story of the fragments. Another, not mutually exclusive, theory is There is also the problem of power. Water “It’s the same way that we would do things that devices such as the Antikythera Mecha- clocks are thought to have been used on occa- today, it’s like modern technology,” says Bit- nism were signifiers of social status. Cuomo sion to drive geared mechanisms that displayed sakis. “That’s why it fascinates people.” What points out that demonstrating wondrous astronomical phenomena. But dripping water fascinates me is that where we see the potential devices brought social advancement. “They only provides enough pressure to drive a small of that technology to measure time accurately were trying to impress their peers,” she says. number of gears, limiting any such display and make machines do work, the Greeks saw a “For them, that was worth doing.” And the to a much narrower scope than that of the way to demonstrate the beauty of the heavens Greek élite was not the only potential market. Antikythera Mechanism, which is assumed and get closer to the gods. ■ Rich Romans were eager for all sorts of Greek to have been handcranked. To make the leap Jo Marchant is Nature’s News Editor.