Box 1. Raised from the depths

From the following article:

In search of lost time

Jo Marchant

Nature 444, 534-538(30 November 2006)



In 1900 a party of Greek sponge divers sought shelter from a storm in the lee of the barren, rocky islet of Antikythera. Once the winds had eased, Elias Stadiatis dived 42 metres to a rocky shelf to look for late additions to his hard-earned haul. Instead of sponges nestled on the sea bed, the shape of a great ship loomed out of the blue. After grabbing the larger-than-life arm of a bronze figure as proof of his find, he returned to the surface to inform his companions. The Antikythera wreck was to yield a stunning collection of bronze and marble statues, pottery, glassware, jewellery and coins; it was also to claim the life of one of the divers, not yet aware of the risk of the bends when diving with an oxygen hose.

As busy museum staff struggled to piece together statues and vases, a formless, corroded lump of bronze and wood lay unnoticed. But as the wood dried and shrivelled, the lump cracked open, and on 17 May 1902, archaeologist Valerios Stais noticed that there were gear-wheels inside.

The gears elicited interest, but it was not until investigations delved beneath the surface that the box started to yield its secrets. The British science historian Derek de Solla Price and the Greek nuclear physicist Charalampos Karakalos made X- and gamma-ray images of the fragments in 1971. Karakalos and his wife Emily painstakingly counted the visible teeth; in 1974 Price published a heroic 70-page account of the machine (D. de S. Price Trans. Am. Phil. Soc., New Ser.64, 1–70; 1974).

"Price really put the mechanism on the map," says Tony Freeth, co-author of a new reconstruction of the device (see page 587). "He understood the essence of what it was — an astronomical computer." But Price massaged some of the data (much to the annoyance of Karakalos and his wife), and his reconstruction was unnecessarily complicated — perhaps too complicated for historians and archaeologists. They largely ignored Price's work, and he died in 1983.

That same year, a Lebanese man walked into the Science Museum in London with the pieces of another ancient mechanism in his pocket. Curator Michael Wright realized the device was a Byzantine sundial from the sixth century AD, which also contained a simple geared mechanism that drove pointers showing the position of the Moon and Sun in the sky. Studying the astronomically enhanced sundial led Wright to Price's treatment of the Antikythera Mechanism, in which he saw serious holes.

In search of lost time 

Wright ended up working with Allan Bromley, a computer scientist at Sydney University in Australia who had become interested in the Antikythera Mechanism at around the same time. Bromley wanted to study the machine with X-ray tomography, which assembles a sheaf of cross-sections of its subject. As the fragments could not be moved from the museum, and Bromley didn't have the money to ship a tomography machine to Athens, Wright used his tool-making skills to build a crude tomograph in situ. The two researchers took around 700 images of the fragments, and Wright has been working on a reconstruction that supercedes Price's ever since.

In the meantime, Mike Edmunds, an astrophysicist at Cardiff University, UK, and his friend Tony Freeth, a mathematician-turned-film-maker living in London, decided the mechanism would make a fantastic subject for a documentary. But their efforts soon turned to discovering more about how the device worked. They contacted Hewlett-Packard, which had developed a method for reading eroded cuneiform tablets that involved building up a composite computer image from pictures taken under light from a wide variety of directions, to reveal more of the inscriptions. They enlisted experts in computer-assisted tomography from British firm XTEC, which developed a new machine just for the Antikythera project.

In search of lost time M. KIRK

Derek de Solla Price tried to undo the Antikythera Mechanism's secrets.

In autumn 2005, the Hewlett-Packard equipment and all 12 tonnes of XTEC's machinery were shipped to the museum. The results have allowed the team to confirm many of Wright's ideas, and extend them. "My main fear initially was that we'd throw all this technology at it and we wouldn't do more than dot the i's and cross the t's," says Freeth. "But we got more out of it than I dared hope."

One major new result came as much from chance as from technology; a key section of a dial found sitting unnoticed in the museum's store room helped reveal that one of the dials was used to predict eclipses. Another big discovery was the identification of a 'pin and slot' mechanism to model the varying speed of the Moon through the sky (see main story).

The inscriptions are also revealing novelties, although deciphering them is hard work: some of them are less than 2 millimetres high, and there are no spaces to show where each word starts and finishes. Agamemnon Tselikas, director of the Centre for History and Palaeography in Athens, spent a concentrated three months trying to decipher the wording, working from late at night into the early hours of the morning: "I needed the silence."

So far Tselikas and his colleague Yanis Bitsakis have more than doubled the number of legible characters on the mechanism, which seem to form a manual that explains how the mechanism was to be used. It takes "imagination and intuition" to decipher the inscriptions, says Tselikas. "We are just starting to penetrate the mentality of the user of this machine." Jo Marchant