What Did the Real Antikythera Mechanism Do And Who Actually Made It?

In 2023’s Indiana Jones and the Dial of Destiny, the latest entry in the iconic adventure film series, everyone’s favourite swashbuckling archaeologist/grave robber hunts after the titular dial, a mechanism invented by Ancient Greek mathematician Archimedes to predict the appearance of fissures in time, allowing the user to travel between the present and the past…because, sure, why not? But while this fantastical plot element might sound like the product of a particularly drunken session of antiquities-themed Mad Libs, amazingly, it is actually based on a real-life artifact called the Antikythera Mechanism, funny enough. Dating from the 1st century B.C.E, this incredibly sophisticated assembly of bronze gears has baffled archaeologists for over a century, predating the earliest known mechanisms of its kind by more than a millennium. Only in recent years has its true function been determined, revealed to be an ancient form of analogue computer – the oldest on record. This is the story of the most incredible example of ancient mechanical and mathematical genius ever discovered. So let’s dive into it, shall we?

Speaking of diving, in the spring of 1900, a team of Greek sponge divers from the island of Symi were returning home from a fishing expedition off the North African coast when a violent storm struck. Blown off course, they were forced to take shelter in a cove on the small island of Antikythera, northwest of Crete. The following morning, the leader of the expedition, Captain Dimitrios Kontos, decided to try his luck in the waters off the island, and sent diver Elia Stadiatos down to search for sponges. Minutes later, a visibly shaken Stadiatos suddenly resurfaced, frantically exclaiming that the seafloor was strewn with dead, naked bodies. When his fellow divers investigated, they discovered that the “bodies” were in fact large bronze and marble statues, the cargo of an ancient shipwreck lying 45 metres below the surface. Upon reaching their home port, the crew reported their discovery to the Greek Archaeological Service, who in 1901 launched an expedition to Antikythera. Over the course of eight months, the team recovered thousands of artifacts, including the magnificent statues, pottery, jewelry, and glassware. From these finds, the wreck was determined to be a Roman ship from the 2nd Century B.C.E, its cargo of fine and expensive art objects likely destined for a wealthy buyer in Rome.

Then, on May 17, 1902, archaeologist Valerios Stais was cleaning some of the many corroded lumps of metal from the wreck when he discovered something completely unexpected: a fused mass of complex, intermeshing bronze gears and wheels, like a large mechanical clock. While the Ancient Greeks were known to have used simple peg-toothed gears for machines like water wheels and windmills, the precision of these gears was unlike anything yet discovered from the period, sparking furious debate among archaeologists. Many refused to believe that the Ancient Greeks could have created such a sophisticated device, declaring the mechanism to be a forgery or a later device that conveniently sank in the same area as the wreck. Others argued it was the remains of a particularly complex astrolabe – a common ancient device for making astronomical measurements and calculations – while still others believed it to be a planetarium or orrery – a mechanical model showing the motion of the planets.

Meanwhile, in 1905 a German philologist named Arthur Rehm conducted the first in-depth analysis of the mysterious device. Observing that many of the gears and the remains of the wooden box that held them were covered in Greek writing, Rehm attempted to translate as much of this text as possible. What he discovered was part of an ancient almanac known as a parapegma, which predicted various astronomical phenomena such as equinoxes, solstices, and the rising and setting of various constellations. He also discovered the names of the five planets known to the Ancient Greeks – Mercury, Venus, Mars, Jupiter, and Saturn – and that certain key numbers – particularly 19, 76, and 223 – appeared several times throughout the mechanism. But perhaps Rehm’s greatest observation was that not all the gears in the mechanism were simply fixed to the wooden base; instead, some rotated on or within other gears, forming epicyclic gearsets. This was a startling discovery, for epicyclic gears would not appear again in the West for another 1,500 years. All this led Rehm to speculate that the Antikythera mechanism was indeed a planetarium or other kind of calculator for predicting astronomical events. However, unable to penetrate the thick corrosion encasing the mechanism and more closely examine its components, he made no further progress, and the Antikythera mechanism soon faded into obscurity.

Things changed in 1951 when the mechanism was rediscovered by Professor Derek de Solla Price, a British physicist and historian of technology working at Yale University. Price immediately became obsessed with this ancient enigma, and devoted much of the next 20 years to cracking its many mysteries. He was aided in his quest by two key developments. Though the Antikythera mechanism had been recovered in 1901 as one solid, corroded lump, over the intervening half century it had been broken into around 20 smaller fragments, revealing more of its inner workings. Advancements in x-ray technology also made it possible to look inside the larger concretions without damaging them further. de Solla Price therefore teamed up with Greek radiologist Charalambos Karakalos to take the first detailed 2-D x-rays of the mechanism. These not only revealed ten previously unknown gears, but also hundreds more Greek characters, providing vital clues to deciphering the mechanism.

For example, on the surviving fragments of the front outer dial, de Solla Price discovered the words Pachon, Payni, and Epiphi – the 9th, 10th, and 11th months of the Ancient Egyptian lunar calendar. Meanwhile, the inner front dial was divided into 360 degrees as well as 12 segments marked with the signs of the zodiac. The dial also featured 20 individual letters, which de Solla Price realized corresponded to Arthur Rehm’s parapegma almanac. From these clues, de Solla Price determined that the front dial of the mechanism tracked the sidereal motion of the sun and the moon – that is, their motion relative to the so-called “fixed stars.” The two pointers that once travelled across the dial, however, had long ago corroded away. Also lost to history was a hand crank on the side of the case, which de Solla Price determined through context must have driven the entire mechanism. Another key feature of the front dial was the rotating outer ring, which allowed the operator to account for the effects of leap years and the intercalary month – the 4-5 days added to the end of the synodic year to synchronize it with the tropical year.

Examining the x-rays, de Solla Price and Karakalos next set about counting the teeth on each of the gears, which was vital for determining exactly which astronomical cycles the mechanism was intended to model. This proved more challenging than expected, as many of the gears were only partially preserved. And since a difference of just one tooth could entirely throw off the analysis of the mechanism, the analysis ran into many frustrating dead ends. For example, Karakalos believed that one particular gear had 128 teeth while de Solla Price believed it had 127. de Solla Price turned out to be correct, for 127 is half of 254 – the number of tropical or solar months in the Metonic Cycle.

First discovered by the Ancient Babylonians but named after 5th Century B.C.E. Greek astronomer Meton of Athens, the Metonic Cycle is a period of 19 solar years or 235 synodic or lunar months after which the phases of the moon occur at the same time of year – the same number 19 that Arthur Rehm had discovered repeated throughout the mechanism. The Metonic cycle is still used today by many religions to determine the dates of certain holidays – including Easter, Rosh Hashanah, and Ramadan. This particular 254-tooth gear drove a pointer on the rear upper face of the mechanism, which indicated all the lunar months in a single 235-month Metonic Cycle. Finally, building from Arthur Rehm’s insights about epicyclic gearsets, de Solla Price concluded that these mechanisms were used to calculate the phases of the moon. In 1974, he published his 20 years of findings in a landmark paper titled Gears From the Greeks.

de Solla Price’s paper shook the world of archaeology, causing many to rethink their assumptions about the mathematical and engineering abilities of the Ancient Greeks. It also inspired many researchers to tackle the puzzle of the Antikythera mechanism. Among these was Michael Wright, a curator at the Science Museum in London. One of Wright’s key insights was that the Science Museum had in its collection a similar artifact to the Antikythera mechanism: a 6th-Century C.E. device known as the Byzantine Sundial Calendar. When the mechanism is spun, the name of the month appears in one window while a graphical representation of the moon’s phases appears in another. While the mechanism is relatively simple – containing only 8 gears compared to the Antikythera Mechanism’s 30 – the similarities between the two designs suggests that the Byzantine Sundial Calendar was part of a long engineering tradition stretching back to the Antikythera mechanism and even beyond. This, along with the discovery of a new mechanism on the front dial, led Wright to hypothesize that the moon pointer on this dial included a rolling spherical indicator to display the phases of the moon. Incredibly, this mechanism made use of a differential gear train, a technology not previously thought to have been invented until the 16th century.

In the early 1990s, Wright, along with Australian historian of technology Allan George Bromley, performed linear x-ray tomography of the Antikythera mechanism, obtaining even more detailed, easier-to-interpret 3D scans. From these scans, the pair were able to determine that much of de Solla Price’s analysis of the mechanism was fundamentally flawed, with 17 out of his 20 gear tooth counts proving to be incorrect. Based on his more accurate counts as well as inscriptions found by Arthur Rehm, Wright hypothesized that the front dial also featured pointers – now lost – representing the orbits of the five known planets – the motion of which was informed by the theories of ancient Greek astronomers Apollonius of Perga and Hipparchus of Rhodes. The Ancient Greeks knew that the planets sped up or slowed down and sometimes even reversed direction as they orbited – a phenomenon known as retrograde motion caused by the earth overtaking the planets as it orbits around the sun. Indeed, the word “planet”, from the Greek planetes or “wanderer”, is derived from this behaviour. However, the Ancient Greeks believed in a geocentric model of the solar system wherein the sun, moon, and planets orbited the earth. To account for retrograde motion, Apollonius and Hipparchus theorized that the planets not only travelled in circular orbits around the earth, but also in smaller “epicycles” attached to said orbits. While not an accurate literal representation of the solar system, this model was mathematically accurate enough to perform practical astronomical calculations – and to be mechanically modelled by epicyclic gearsets like in the Antikythera mechanism. The epicyclic system would remain the dominant model of the solar system for more than a millennium and a half until it was finally overturned by Copernicus, Galileo, and others in the 16th and 17th Centuries.

Wright and Bromley’s 3D scans led to a number of other key discoveries, such as the fact that the two dials on the rear of the mechanism were not concentric circles as had previously been assumed but rather spirals. Furthermore, the pointers on these dials were telescopic and featured a peg that rode in a groove on the spiral dial, allowing them to expand and contract as they travelled around the dials. This clever arrangement allowed much longer and accurate measurement scales to be compressed into a more compact space. Wright and Bromley also discovered a smaller dial within the lunar dial, which appeared to measure the 76-year Callipic cycle. Named after Greek astronomer Callippus, this cycle represents the common multiple of the tropical or solar year and the synodic or lunar month, and is a more accurate improvement on the Babylonian 19-year Metonic cycle. It also corresponded with the number 76 which Arthur Rehm found repeated throughout the mechanism. Based on all these discoveries, in 1997 Wright constructed the first practical working model of the Antikythera mechanism.

Another researcher who was skeptical of de Solla Price’s initial findings was Dr. Tony Freeth, a British mathematician and documentary filmmaker. Upon reading de Solla Price’s 1974 paper, Freeth, citing Occam’s Razor, found the notion of using epicyclic and differential gears to determine the phases of the moon too unnecessarily complex to be true. After all, there were much simpler and efficient mechanical means of obtaining these values, which any competent Greek mathematician would have known.

In any event, in 2000, Freeth and a team of British and Greek researchers formed the Antikythera Mechanism Research Project and launched a fresh assault on the enigmatic mechanism’s secrets. This effort involved the use of a state-of-the-art digital 3D imaging system developed by Hewlett-Packard and an 8-ton x-ray tomography machine built by UK Firm X-Tek Systems – both of which had to be transported to the National Archeological Museum in Athens since the Antikythera Mechanism is too fragile to travel. But the team obtained much more than high-resolution photos and x-rays; while working at the Museum, they were approached by a curator who had discovered a box labelled “Antikythera” in a storage room. The box turned out to contain 72 extra fragments of the mechanism, increasing the total to 82. The largest 7 of these fragments are now designated by the letters A-G; the remaining pieces by the numbers 1-75.

The team’s high-resolution scans revealed more than 2,000 new text characters in the mechanism, providing new clues to its function. For example, a second smaller dial inside the rear lunar dial is divided into four quadrants marked with names Nemea, Naa, Isthmia, and Olympia – all sites of Ancient Greek athletic games. This dial thus likely used the lunar calendar to determine the appropriate opening day of these games, which took place every 2-4 years.

But the function of one particular component remained frustratingly elusive: a large 223-tooth gear behind the lower rear dial. This is connected to an epicyclic set of 4 smaller gears, which de Solla Price had theorized calculated the phases of the moon. However, not only does this not make sense from a complexity standpoint, but Tony Freeth discovered all four gears in the set have the same number of teeth. This would make the output the same as the input – rendering the whole mechanism pointless. However, Freeth then discovered an observation made by Michael Wright that one of the epicyclic gears features a pin that engages in a slot in another. Its rotation axis is also mounted at a slightly different angle, meaning that the rotation transferred by one gear to another will periodically slow down or speed up. Upon reading this, Freeth had a Eureka moment, for this mechanism perfectly modelled the orbit of the moon. We now know that the moon’s orbit is not perfectly circular but rather elliptical, causing its motion across the sky to periodically speed up and slow down – a phenomenon known as the anomalistic cycle. Not knowing about elliptical orbits, however, Ancient Greek astronomers like Hipparchus modelled this behaviour via epicycles.

But while this ancient theory neatly explained the function of the offset epicyclic gears, there was a further wrinkle: the moon’s orbit is constantly shifting, tracing a flower-petal-like path around the earth like a giant spirograph. The time it takes for the moon to return to perigee – its farthest distance from the earth – is slightly longer than the time it takes to return to the same point in the sky – a difference of just 0.112579655 turns per year. Based on this, Freeth found that if the input gear had 27 teeth, the rotation of the output gear was slightly too fast; if, by contrast, it had 26 teeth, the rotation was slightly too slow. But if the input gear had 26 and a half teeth, the output ratio was exactly 0.112579655 – accurate to 9 decimal places. Though a gear can’t physically have 26 and a half teeth, Freeth quickly realized that 26 and a half times two is 53 – the exact number of teeth on the remaining gear in the set. By using precise gear ratios and slightly offset gear axes, the designer of the Antikythera mechanism had succeeded in mechanically modelling the elliptical orbit of the moon to a high degree of precision – a staggering intellectual achievement in any era, let alone the first century B.C.E.

But there was more to come, for Freeth soon realized that the number 223 Arthur Frehm had found repeated all over the mechanism corresponded to the 223 lunar month Saros Cycle that governs solar and lunar eclipses. Furthermore, the lower back dial of the mechanism was covered in short letter groups, nearly all of which featured the letters Sigma and Eta. Realizing that these stood for Selene and Helios, the Ancient Greek gods of the Moon and Sun, Freeth determined that the lower dial was, in fact, a sophisticated eclipse predictor. With this final breakthrough, the Antikythera Mechanism Research Project was able to build a new working model of the mechanism with all components accounted for save one – a small 63-tooth gear called R1 whose function remains a mystery to this day. Some believe it to be the last remaining component of Michael Wright’s hypothetical planet-tracking mechanism, the rest of which either corroded away long ago or remains hidden somewhere beneath the Aegean Sea, waiting to be discovered. Others, however, doubt the existence of such a mechanism, arguing that Wright’s hypothetical reconstruction is overly complicated and lacks the ingenious elegance of the rest of the mechanism. Time will tell whether the truth will ever be discovered.

But the most tantalizing mystery of all still remains: who actually built this mechanical marvel – and when? Unfortunately, hard evidence is rather thin on the ground, though there are some tantalizing clues. Though the 1901 expedition that first excavated the wreck dated it to the 2nd century B.C.E, excavations conducted by explorer Jacques Cousteau in the 1950s and 70s uncovered coins minted in the Greek city of Pergamon dated to 86 BC, reducing the age of the wreck by a full century. And while nothing resembling the Antikythera Mechanism has yet been discovered from this period, there is evidence that devices of this type were widely known in the ancient mediterranean. For example, just a few years after the Antikythera ship went down, the Roman statesman and writer Cicero wrote that his colleague, the philosopher Poseidonius of Rhodes:

“…recently made a glove which in its revolutions shows the movements of the sun and stars and planets, by day and night, just as they appear in the sky.”

Indeed, the Antikythera ship was built in the Rhodian style, while the parapegma almanac has been shown to be most accurate at latitudes similar to Rhodes. Furthermore, Rhodes was also home to Hipparchus, on whose astronomical theories much of the mechanism is based. Others, however, point to an origin in Pergamon, which minted the coins found aboard the shipwreck and whose Library was second only to the great Library of Alexandria in terms of preserved ancient knowledge.

But perhaps the most tantalizing theory credits the mechanism’s creation to an even more legendary intellect. Among the many findings of the Antikythera Mechanism Research Project was that the names of certain months on the calendar dial were not universal across the Hellenistic world but were instead specific to the city-state of Corinth and its colonies. And one of the major colonies of Corinth was Syracuse on what is now the island of Sicily – home to none other than mathematical and mechanical genius Archimedes. Indeed, in his account of the 212 B.C.E. siege in which Archimedes was killed by the Romans, Cicero claims that the general in charge, Marcus Marcellus, made off with a sophisticated astronomical instrument designed by the great genius himself.

But while compelling enough to inspire the writers of Indiana Jones, unfortunately this origin story is almost certainly false. For one thing, Archimedes died more than 100 years before the Antikythera ship sank, while in 2017 it was determined that though the calendar style used by the mechanism was indeed specific to Corinth, it could not have come from Syracuse. Then again, a 2014 study conducted by Christian Carman and James Evans at the University of Puget Sound found that the start date on the Saros cycle eclipse predictor corresponded to around 205 B.C.E. – only 7 years after Archimedes’ death – meaning that while Archimedes may not have physically built the Antikythera Mechanism itself, he may well have invented many of its operating principles, inspiring a centuries-long tradition of mechanical computer construction. Without further evidence, we may never know for certain. But regardless of who exactly designed it, the Antikythera mechanism stands as an astonishing monument to the genius of the Ancients – an object the likes of which would not appear again until the 14th Century. It also serves as a sobering reminder of just how much ancient knowledge has been tragically lost to history. As Derek de Solla Price once wrote:

[The Antikythera Mechanism] requires us to completely rethink our attitudes toward ancient Greek technology. Men who could build this could have built almost anything they wanted to. The technology was there, and it has just not survived like the great marble buildings, statuary, and the constantly recopied literary works of high culture.”

Expand for References

James, Peter & Thorpe, Nick, Ancient Inventions, Random House Publishing Group, 2006

The World’s Oldest Computer, National Geographic, June 29, 2014, www.youtube.com/watch?v=pt4WxxduSGY

Freeth, Tony, Decoding an Ancient Computer, Scientific American, December 2009, https://www.cs.virginia.edu/~robins/Decoding_an_Ancient_Computer.pdf

The Antikythera Mechanism Explained With Dr. Tony Freeth, Event Horizon, June 1, 2023, www.youtube.com/watch?v=0y-Qxw4DLOY

Sample, Ian, Mysteries of Computer From 65 BC Are Solved, The Guardian, November 30, https://www.theguardian.com/science/2006/nov/30/uknews

Kaplan, Sarah, The World’s Oldest Computer is Still Revealing Its Secrets, The Washington Post, Hune 14, 2016, https://www.washingtonpost.com/news/speaking-of-science/wp/2016/06/14/the-worlds-oldest-computer-is-still-revealing-its-secrets/

Casselman, Bill, The Antikythera Mechanism II, https://www.ams.org/publicoutreach/feature-column/fcarc-diff4

Share the Knowledge! FacebooktwitterredditpinteresttumblrmailFacebooktwitterredditpinteresttumblrmail
Print Friendly, PDF & Email
Enjoy this article? Join over 50,000 Subscribers getting our FREE Daily Knowledge and Weekly Wrap newsletters:

Subscribe Me To:  | 

Leave a Reply

Your email address will not be published. Required fields are marked *