Scientist of the Day - The Antikythera Mechanism

Front of Fragment A of the Antikythera Mechanism, showing the main four-spoked gear wheel, b1, National Archaeological Museum of Athens (Wikimedia commons)

On May 17, 1902, an archaeologist named Valerios Stais was examining a small cook-book-sized rocky slab that had been brought up the previous summer from the floor of the Aegean Sea, off the coast of an island named Antikythera, between Crete and Greece. A shipwreck had been found there by sponge divers in 1900, who brought up some statuary and amphorae; the next summer, as the government took over, other objects followed, including our mystery object.  On this day in 1902, Stais noticed that the rock seemed to have a part of a gear wheel embedded in it. This was the first recognition that the object was some kind of mechanism.  Google commemorated the 115th anniversary of the discovery with a Google Doodle on May 17, 2017 (last image)

Fragments C, A, and B (left to right) of the Antikythera Mechanism, rear view, on display in the National Archaeological Museum of Athens (Wikimedia commons)

Subsequent examination revealed that the object was a metal sandwich, and the meat and cheese were gears. On the back were fragments of text, in Greek, but in reverse.  At some point, before or after May 17, the object broke into three large fragments and scores of smaller ones. The largest piece, now called Fragment A, was the piece with a large four-spoked toothed gear visible on one side, probably what caught Stais’s attention, with evidence of other gears beneath it (first image).  Other fragments had sections of dials, one of them a spiral.  There are now 82 fragments known, all in the National Archaeological Museum in Athens.

Fragment B of the Antikythera Mechanism, showing part of the Metonic Cycle spiral, National Archaeological Museum of Athens (Wikimedia commons)

Although the Antikythera Mechanism, as it came to be called, attracted some interest in the next 50 years, it was generally assumed that it was of a much later date than the shipwreck, probably constructed in the second post-Christian millennium, when people started making mechanical clocks.   It was not until 1951 that a future Yale professor, Derek de Solla Price, took a closer look at the Antikythera Mechanism and concluded that it was some sort of an astronomical computer or planetarium, and that it was indeed ancient. Even though de Solla Price wrote a widely-read Scientific American article about the Mechanism in 1959, and a monograph in 1974, and even built a replica (fifth image), the Mechanism was still considered a "wonder" for the next few decades, something outside the ordinary ancient Greek experience. But in the last decade of the 20th century, continuing right up to the present, interest has picked up considerably.  The Mechanism has now been examined by X-rays, and tomography, recording details of every layer, millimeter by millimeter, including an abundance of engraved text and instructions.  There is an international Antikythera Mechanism Research Project to which a number of eminent historians of technology and mechanical engineering have signed on, and another such project at Oxford, and the workings of the mechanism are now much better understood, although interpretations differ.

Fragment C of the Antikythera Mechanisn, with part of the two front dial rings, National Archaeological Museum of Athens (Wikimedia commons)

Basically, the Antikythera Mechanism seems to be an astronomical solar, lunar, and planetary calendar and/or planetarium, and it was probably built in the early 2nd century BCE, perhaps at Rhodes, or Pergamon. The front has two concentric dials on which were engraved the signs of the zodiac and a degree scale, marking out the days of the year.  Until recently, that was assumed to be a 365-day solar year; now it seems it might be a 354-day lunar year.  Many concluded that a number of pointers must have protruded from beneath the smaller top dial, so that when you turn a knob at the side of the Mechanism, the pointers move, the moon pointer most rapidly, the pointers for the sun and the five visible planets more slowly. The clockwork that drives these pointers is quite intricate, because the sun and moon change speed, and the planets occasionally move retrograde, or backwards, and their forward speeds are variable as well. However, if I understand developments correctly, the existence of planetary pointers may now be out of favor, since no clockwork to drive them survives, even though text recording long-scale planetary periods is engraved on parts of the Mechanism.

There are also five dials on the back of the Mechanism, two large spiral-shaped dials and three small ones. The large dials keep track of what is called the Metonic cycle, the fact that 19 solar years exactly equals 235 lunar months, and the Saros, which is an eclipse cycle of 223 months. One of the small dials even records the four-year Olympic Games cycle. Fragment B shows a part of the Metonic cycle spiral (third image).

Replica of the Antikythera Mechanism, built by Derek de Solla Price, 1970s, now in the National Archaeological Museum of Athens (Wikimedia commons)

It used to be thought that the Antikythera Mechanism was an aberration, something completely untypical of Hellenistic Greek science. And it is true, we have no other device like it from the ancient world. But scholars now stress that there is nothing embodied in the Mechanism that cannot be explained by Greek astronomical theory, and that building the mechanism required skill, but nothing beyond the capabilities of a Greek craftsman, in this case, an exceptionally gifted one.  We surmise, from the textual evidence of Cicero and others, that mechanical astronomical devices such as this were not that uncommon in the Hellenistic world. Perhaps if more of them had survived, we wouldn't be quite as astonished, as we tend to be, by the Antikythera Mechanism. Or perhaps it would still be a jaw-dropper. If you get to Athens, it is a must-see in a city of must-sees; the larger fragments are displayed in the National Archaeological Museum: and they have reconstructions as well.

Google Doodle for May 17, 2017, commemorating the 115th anniversary of the discovery of the Antikythera Mechanism (doodles.google/doodle)

At least 10 replicas of the Antikythera Mechanism have been built since de Solla Price put together the first one (fifth image).  They differ because we have only a fraction of the supposed 30 (or 37) gear wheels, requiring that much of the mechanism be imagined. One amateur machinist and clockmaker in Australia, known to me only as Chris from Clickspring, has been building a replica for 7 years now, and he has documented the entire construction process on video, condensed down into what so far are 12 videos, that vary in length from 7 to 25 minutes each.  The videos are, to me, mind-boggling, revealing not only what can be done in a modern machine shop by a detail-oriented craftsman, but also demonstrating the techniques and tools that would have been used by the ancient Greek builder. The YouTube video channel is called Clickspring, and you can find the first video here. I encourage you to watch the first two, because then you will not be able to resist the next 10 (with more to come).  Chris started this project before it was discovered that the calendar on the front probably tracks the lunar year and not the solar year, so I don’t know what kind of problems that has caused, but I am sure he will figure it out as he finishes the project.

And if you have seen the latest Indiana Jones film, Indiana Jones and the Dial of Destiny (2023), in which the Antikythera Mechanism plays a leading role, I will observe only that: 1) the mechanism did not emerge from the sea utterly corrosion-free; 2) it was not built by Archimedes; 3) you cannot put the fragments back together (only 2 of them in the movie) and make it operate; and 4) it is not a time machine.  There might be a few other nits to pick, but I will leave it at that.  I did, however, enjoy the film.

William B. Ashworth, Jr., Consultant for the History of Science, Linda Hall Library and Associate Professor emeritus, Department of History, University of Missouri-Kansas City. Comments or corrections are welcome; please direct to ashworthw@umkc.edu.