Decoding the Heavens
Page 17
For the time being, the Athens museum had turned Freeth and Edmunds down, because Wright was still at work on his data. But with the number of eminent astronomers on their team Wright feared it would not be long before they succeeded in imaging the fragments, and if they did so before he solved the mechanism, all of his years of effort would have been in vain. He had to get his work published, and fast.
After 500 hours work – exactly 100 days of weekend working and sleepless nights – Wright finished the planetary display of his reconstruction, and published it in a specialist clockmakers’ journal in May 2002.
But things at the Science Museum soon went from bad to worse. As part of the ongoing process of modernisation, a team of management consultants had been called in to streamline the way the institution was run. All of the curators had to apply for new jobs, and it was clear that academic research would not be a priority. The museum was under pressure to justify the money it was spending by reaching out to the public and getting as many people as possible in through its doors. That meant thinking up new ways to make science glamorous and exciting, with interactive exhibits, snappy soundbites and multimedia displays. Old-fashioned curators hunched over dusty instruments in glass cases were not part of the vision.
To be interviewed as if he were a stranger after more than 30 years of service at the museum was not something that Wright felt able to swallow. When asked to give presentations by his interview panels in early 2003, he lectured them on ‘How to Run a Museum’ and ‘The Importance of Saying What You Mean’. Still, the redundancy package gave him enough to live on for a while, and now he had more time to work on the Antikythera mechanism.
By this time Bromley had passed away and his widow had sent Wright the remaining images that she could find. So he began trying to solve the rest of the device, with nearly 700 radiographs to work from. He couldn’t see any way to publish all the images, and even if he could they wouldn’t make any sense to anyone unless he could come up with a suggested reconstruction. Although his tomography cradle had ensured that in each image only features in the desired plane were sharp, the others were all still there as blurred grey streaks, so interpreting the details was a time-consuming and specialised task. He studied the radiographs over a lightbox one by one with a magnifier, as Emily Karakalos had done, until his eyes streamed and his head hurt.
By this time, however, Wright’s son Gabriel was studying for a doctorate in medical imaging at Oxford University. Gabriel’s lab had the necessary equipment to scan in the radiographs as high-resolution digital files. He patiently scanned in all 700 radiographs and set them up on his father’s computer with some basic image manipulation software. By the end of 2003 Wright could zoom in on his images, alter their brightness and contrast, and flick effortlessly from one image to another. Before he had been estimating tooth counts by cutting out transparent circles of different sizes and divisions and laying them over the film to compare with the wheels beneath, but now he could measure them accurately with the click of a mouse. He found himself agreeing closely with Karakalos’s counts, and not with the numbers accepted by Price.
Things were really starting to move now and he published a string of papers, one for each step he made, all in specialist publications that dealt with clockwork and scientific instruments. When he measured the dials on the back of the device, he realised that although the rings were concentric, the two halves of each dial were drawn around a different centre. In other words, each dial was not made up of several separate rings, but a single spiral. The upper spiral had five turns and by measuring the marks on it he calculated that each revolution of the pointer represented 47 divisions, making 235 in all. He realised that the spiral must have displayed the 235 months of the Metonic 19-year cycle, as calculated by the gear train under the front dial.
And when he looked at the little dial next to this spiral, Wright saw that it was divided into four. The Greeks had used another period, called the Callippic period, made up of four Metonic periods so that it was 4 x 19 = 76 years long. The year was known to be 365 ¼ days long, and this longer period got rid of the awkward quarter days. This period was even mentioned in one of the fragmentary inscriptions from the mechanism, so it made sense that it would have been shown on one of the dials.
The gear train leading to this dial was lost, but by adding in three extra wheels Wright came up with a simple train in which the little dial turned once in every 20 turns of the main pointer, so that it would show where you were in the Callippic cycle. It could have been used to help keep track of long intervals of time when the mechanism was being wound forwards or backwards, and would also have been useful for converting dates given in the Egyptian solar calendar (as displayed on the front dial) with any of the various local lunar calendars.
On the front of the mechanism Wright also made sense of a strange circular arrangement that seemed to be stuck on to the front dial. Price had seen it and thought it might be the remains of a folding crank handle, but from knowledge of later astronomical clocks Wright recognised it as a Moon phase display. In his radiographs he could clearly see the remains of a little wheel at right angles to the others, designed to pick up the relative motion between the Moon pointer and the Sun pointer, to turn a little Moon ball on its axis. There was a perfectly round hollow in the crumbling fragment, showing that the ball itself had not corroded and had fallen out only after everything around it had done so. Wright suggested that it might have been made of ivory, with one half blackened with ink. The ball was set into the central boss of the Moon pointer so that only the front half showed, and it would have spun in time with the phase of the Moon, showing all black when the Moon wasn’t visible, through a sliver of light when the new crescent appeared, to a whole white face at full Moon. Wright was stunned to see it in such an ancient machine.
Then there was the problem of Price’s differential gear. Price had described two linked inputs that sat on a turntable, driving it around at a speed that was half the sum of their two rotations. But when Wright looked closer, he could see only one input. So it wasn’t a differential gear at all, but appeared to be an epicyclic one, similar to the ones he had already suggested for the planets. It wasn’t in the right place to be modelling any of the planets they would need to be on a turntable concentric with the front dial. But there is another use for epicyclic gearing: to calculate gear ratios that are too complicated to achieve by normal fixed gears. It was commonly used for this purpose in the elaborate astronomical clocks of Renaissance Europe.
Wright drew up spreadsheets of all the possible numbers of teeth for the gears in the train, but couldn’t see what it was meant to calculate. And he noticed a couple of other strange features that were hard to explain. The first was that the turntable had 223 teeth around its edge, which didn’t seem to engage with anything. That was odd – 223 is a prime number and you’d only bother to make such a wheel if you needed the prime number for a particular gear ratio. It didn’t make sense to cut such a wheel and then just use it for a turntable that didn’t need any teeth at all.
And on the turntable he saw a double wheel system in which one small wheel sat almost, but not quite, on top of the other. The bottom wheel had a pin sticking up from it, which engaged with a slot in the top one, so that as the wheel with the pin rotated it drove the other wheel around. Because the wheels rotated around slightly different centres, the pin from the bottom wheel would slide up and down in the slot, towards and away from the centre of the upper wheel, introducing a wobble into its speed of rotation.
Wright had seen such mechanisms in astronomical clocks. They were used to model the fact that planetary orbits are ellipses, not perfect circles, so their apparent speed varies. No one had worked out the maths to model the planets that way at the time the Antikythera mechanism was made. But the ancient Greek astronomer Hipparchus did use such a wobble in equations to describe the motion of the Sun and Moon.
It was a wonderful discovery – the earliest example of such a pin-and-slot me
chanism by close to 1,500 years. And it gave valuable support to Wright’s idea that pins and slotted levers were used in epicyclic gearing at the front of the device to model variations in the motion of the Sun, Moon and planets. But positioned as it was towards the back of the mechanism and on a mysterious turntable, he was stumped about the purpose of this particular pin-and-slot. Although he realised the similarity to Hipparchus’s theory, he couldn’t see how it could possibly have anything to do with modelling the Sun or Moon.
By this time, Tony Freeth and Mike Edmunds had the go-ahead for their project, and Wright knew that they would be visiting the Athens National Archaeological Museum in October 2005 to image the fragments. Wright was due in Athens that same week, to present his results at a conference. It would be his last chance to claim the solution to the Antikythera mechanism for his own before the new boys muscled in on his territory. He had to get his reconstruction finished in time.
So he rushed it. With the weeks running out, the best answer he could come up with was that the lower back spiral had shown a period of four ‘draconitic’ months, split into 218 half days. Draconitic months follow the Moon as it crosses the plane of the Sun’s orbit and are useful for predicting eclipses. He didn’t know why the maker of the device would have wanted to display half-days on that dial, but this was the only astronomical period that made any sense with the gearwheels he had measured. The same result could have been calculated more easily with a fixed gear train – there was no need for epicyclic gearing – but he figured that maybe the designer’s technical talents weren’t matched by his mathematical skills.
Wright still couldn’t see how the pin-and-slot or the 223-tooth wheel could have worked within the mechanism, so he concluded that these must have been spare parts, recycled from other devices. After all, his own model was made from old brass door plates. The exciting thing about this idea was that it would be proof that the Antikythera mechanism wasn’t a one-off. Within the surviving fragments we would have the remains not just of one unique device, but two or three.
Meanwhile, Wright overlooked a broken-off shaft sitting just next to the 223-tooth turntable. He had noticed it early on and always meant to explore the idea that it might have carried a missing gearwheel. Later, Tony Freeth would place a wheel here, solving all of Wright’s problems at once.
October comes and Wright arrives in Athens with his finished model, grimly triumphant as Freeth’s team completes its imaging. On the day of his talk he demonstrates the workings of his device to a small but captivated audience. He turns the handle on the side like a magician and there’s a hush as time passes before everyone’s eyes, just a soft clicking sound as the Moon traces undulating circles through a miniature sky, cycling from black to silver as the golden Sun glides slowly round and the planets meander back and forth, their seemingly random paths guided by a hidden clockwork order.
Wright sees three decades of his life passing as the heavenly cycles run their course, from the young curator who was once captivated by Price’s work and wished it were his own, to the man he is now, standing here with the Antikythera mechanism finally recreated and working again for the first time in 2,000 years.
His presentation was meant to be the high point of the conference. But there is a late addition to the programme. Freeth, after finishing his imaging of the fragments and visiting the conference exhibition to glance at Wright’s model, has gone home to London. But now his collaborators take the stage. In particular, Mairi Zafeiropoulou – a sturdy and rather formidable archaeologist who works with the Athens museum’s bronze collection – triumphantly shows a new piece of the mechanism that she has recently found in the museum stores. It is a substantial piece of the lower back dial.
8
The New Boys
Vision is the art of seeing what is invisible to others.
— JONATHAN SWIFT
THERE WAS AN ear splitting crack, as a dazzling ten-inch spark tore across the room. The cable’s free end, bursting with a quarter of a million volts, stripped a violent path through the air and discharged its devastating load of electrons into every computer in the vicinity.
‘Shit,’ said Roger Hadland.
It was summer 2005 and he was standing in the crowded development lab of X-Tek, the company he had founded 20 years earlier to build sophisticated X-ray imaging systems for industry. From humble beginnings the company had grown explosively, carrying him out of his garage and into smart buildings just outside the picturesque town of Tring, in Hertfordshire. The firm’s success was mainly down to Hadland’s passion for engineering – his speciality was thinking of new ways to build machines that could see better than anyone had seen before. He pushed his team to come up with ever smaller and more powerful X-ray sources, capable of imaging in minute detail the insides of anything from tiny electronic components to barrel-sized nuclear-waste drums.
X-Tek’s precision-imaging equipment was particularly popular with microelectronics companies, who used it for quality control on their computer chips. But in 2001, just as the company had leased an extra site in order to expand still further, the dot com market crashed, taking many of X-Tek’s customers with it. In the following years the company had struggled – Hadland was forced to lay off more than a third of his staff and innovation had slowed to a trickle. He planned to sell, but couldn’t bear to hand over his company to callous investors who would break it apart for profit, and no one else was interested.
Then the Antikythera mechanism had come into his life, in fact taken it over, and somehow he had found himself in a daunting race against time. Inspired as never before, he was diverting company resources and turning away precious customers in an attempt to build a machine that would take the ultimate images of the ancient fragments. He wasn’t sure whether to be glad or sorry – there were certainly others in the company who thought he had gone mad. On days like today, he feared they might be right.
The project had its roots back in 1998, seven years before Hadland had ever heard the fateful word ‘Antikythera’. Mike Edmunds, head of astronomy at Cardiff University, was sipping tea and looking for a history project for one of his students when he came across the work of Derek de Solla Price.
Edmunds is a genial man with red cheeks and white hair, and an office with a huge, sunny window that looks out over a leafy Cardiff side street. He works on questions that go beyond the Earth, Moon and Sun, beyond anything that the ancient Greeks could have imagined: the evolution of stars, galaxies, elements and the cosmic dust that ultimately makes up planets and everything on them. On this day, however, he reined in his imagination from the edges of the universe and instead sent it 2,000 years back in time.
Reading Price’s words, Edmunds was surprised that he hadn’t heard about the Antikythera mechanism before, and that it didn’t have a recognised place in the history of astronomy. Reviewing what was known about it would make a perfect student project.
Edmunds also described the mechanism to an old friend of his, Tony Freeth. A mathematician by training, Freeth earned a PhD as a young man exploring the strange, abstract lands of set theory. But equations weren’t enough for him and he soon moved away from academia to set up a business making documentary films, working out of his book-filled home in west London. Freeth – pale and balding with a grey moustache – is more serious than the jovial Edmunds. But he’s capable of quite passionate intensity when the moment seizes him. He called his company Images First, because he felt that pictures were the way to tell any story that mattered. The resulting films were earnest accounts of topics such as Alzheimer’s disease, apartheid and African agriculture.
The unsung Antikythera mechanism would make a unique subject for a documentary, urged Edmunds. Freeth, too, was amazed that it was so scarcely known, when surely it should be an icon of the ancient world. The complexity of its gearing appealed to his love of maths and logic, while he realised that making a film about the mystery might at last give him a subject that would appeal to a wider audience. He mentioned t
he idea to a few possible buyers, but with Price’s work still standing as the last published study of the mechanism, he always got the same response: ‘There’s nothing new.’ If he was going to sell a film on the Antikythera mechanism, he would need some fresh results.
He started to research the device, turning his attention first to Price’s triumphant Gears from the Greeks. Like Michael Wright before him, he soon saw details that didn’t add up. For Freeth, the first red flag was Price’s description of the sponge divers who salvaged the Antikythera wreck: ‘Only six divers were available, and because of the water depth they could not remain on the bottom for more than five minutes, which together with four minutes for ascent and descent entailed about nine minutes of submersion without air-tanks or tubes to help them.’
Nine minutes of submersion without air-tanks or tubes? Surely that was impossible. A quick search on the Internet revealed that the world record for freediving at the time was significantly shorter than that, even for lying completely still in a pool,2 and that sponge divers at the turn of the century would have worn full diving suits and helmets, with breathing tubes.
Freeth applied similar scrutiny to the rest of Price’s paper and was soon questioning every part of his reconstruction. Like Wright, he saw through the technical bluffs and realised that Price had massaged Charalambos Karakalos’s numbers to fit his theories. Freeth’s fundamental concern was that Price’s model involved complicated combinations of gears to calculate what were ultimately quite simple numbers. He even came up with his own ‘Minikythera’ design that could do exactly what Price’s could, using far fewer gears. Anyone sophisticated enough to have built the mechanism wouldn’t have been so wasteful.
Tony Freeth became fascinated, obsessed even, by the device. And along the way, his motivations changed. His primary aim was no longer to make a film about the Antikythera mechanism. Instead, he resolved that he would be the man to solve its long-standing mystery. The Antikythera bug had infected its next victim.