---
title: "Antikythera: Thirty gears and the sky"
publication: Awescroll
number: "003"
url: https://awescroll.com/003-antikythera
published: 2026-09-28
updated: 2026-10-05
reading_time: 9 min
language: en
---

# Antikythera

Thirty gears and the sky.

*No. 003 · 9 min read · 28 September 2026*

In the spring of 1900, sponge divers sheltering from a storm off a small Greek island found a shipwreck full of statues. Among what they brought up the next year was a lump of corroded bronze and wood about the size of a large dictionary. Nobody looked at it twice.

In 1902 it broke apart in a museum in Athens. Inside were gearwheels, bronze, precise, the size of coins. It took most of a century, and in the end a CT scanner, to say what they did. They computed the sky.

Nothing as complex is known from the thousand years that followed it.


---

*02 · The storm*

*EASTER 1900 · OFF ANTIKYTHERA*

## A heap of dead naked people.

Captain Dimitrios Kontos and his crew of sponge divers from the island of Symi had sheltered from a storm near Antikythera, a scrap of rock between Crete and the Greek mainland. When the sea calmed, they dived. The first man down, Elias Stadiatis, came up shaking. As the researcher Tony Freeth tells it, he was mumbling about a heap of dead naked people.

They were statues, marble and bronze, lying on the seabed between 50 and 64 metres down. Stadiatis brought up the arm of a bronze statue to prove it.

> **THE DEPTH**: The National Archaeological Museum in Athens gives 50 to 64 metres; other estimates are shallower. This story uses the museum&\#39;s. The dead-bodies story is Freeth&\#39;s retelling; no document from 1900 with it was found.


---

*03 · The salvage*

*NOVEMBER 1900 TO SEPTEMBER 1901*

## The first great dig under the sea.

The Greek state sent its navy. With the warship Mykali standing by, the same sponge divers went down in canvas suits and copper helmets, breathing air pumped from the surface, and for ten months they brought its cargo up piece by piece: about 36 marble statues, the bronze youth now called the Antikythera Ephebe, bronze furniture, glass, gold jewellery, amphorae. One diver, Georgios Kritikos, lost his life.

![Black and white photograph: men in caps and suits crowd the deck and rails of a steamship; in the foreground a large find, crusted from the sea, hangs in a rope net](https://awescroll.com/photos/antikythera-salvage-1901.jpg)
*Officials, sailors and sponge divers on the deck of the navy ship Mykali during the salvage, winter 1900 to 1901, with a find still crusted from the sea hanging in the net. (Photographer unknown, 1900 to 1901 · Wikimedia Commons, public domain)*

The ship was a freighter of about 300 tons, most probably bound for Italy. Silver coins found on the site put the sinking around 60 BC. Some of what it carried was already centuries old.

Somewhere in that year, probably in July 1901, a shapeless lump came up too. It went to the museum with everything else.

> **THE DATES**: The salvage months follow the timeline of the Antikythera Mechanism Research Project. The museum dates the wreck to 60 to 50 BC on one page and 75 to 50 BC on another; its coin study points to about 60 BC.


---

*04 · The lump*

*MAY 1902 · NATIONAL ARCHAEOLOGICAL MUSEUM, ATHENS*

## The lump breaks open.

Months later, in the museum, the lump split. Inside were wheels with triangular teeth, about a millimetre and a half apart. The first newspaper reports, in May 1902, credit Spyridon Stais, the former minister who had overseen the salvage, with noticing them.

Noticing was the easy part. The first researchers took it for an astrolabe, an instrument for sighting stars. It was something else, and saying what would take most of the century.

![A flat, green-black lump of corroded bronze on a clear stand; a large wheel with four spokes shows in its face](https://awescroll.com/photos/antikythera-fragment-a.jpg)
*Fragment A today, about 18 by 15 centimetres. The cross is the four spokes of the largest wheel. Almost no metal is left: what you see is corrosion that has kept the shapes. (Marsyas, 2005 · Wikimedia Commons, CC BY 2.5)*

> **WHO, AND WHICH DAY**: Accounts differ. The earliest newspapers, of 21 May 1902 in the Julian calendar Greece still used, credit Spyridon Stais; a later account names the archaeologist Valerios Stais and 17 May. This story says May 1902.


---

*05 · What survives*

*WHAT THE SEA LEFT*

## A third of a machine, in 82 pieces.

- 82 (FRAGMENTS)
- 30 (BRONZE GEARS)
- 27 (OF THEM IN FRAGMENT A)

Seven large fragments, labelled A to G, and 75 small ones, not all of them certain to belong. Thirty gearwheels survive, 27 in Fragment A, which weighs 369 grams. About a dozen show on the surface; the rest were found by X-ray.

What is left is mostly corrosion: bronze turned to mineral, holding the shape of what was there. By the researchers&\#39; estimate, about a third of the machine survives.

![The reverse of the same fragment: a crust of green, brown and grey corrosion, with rings of fine gear teeth fused into it](https://awescroll.com/photos/antikythera-fragment-a-back.jpg)
*The back of Fragment A: rings of gear teeth, fused into the crust. (Marsyas, 2005 · Wikimedia Commons, CC BY 2.5, resized)*


---

*06 · The whole*

*A RECONSTRUCTION*

## A box with a handle, and the sky inside.

Put the pieces back together and the scans, the teeth and the inscriptions agree on the outline: a wooden case about 34 by 18 by 9 centimetres, bronze plates front and back covered in dials and writing, and at the right-hand side a handle. You turned the handle, and time went by on the dials.

> **THIS MODEL**: The machine in this story is a reconstruction drawn for it. Its gears have the published tooth counts, so every wheel turns at the rate the original was built to; where each axle sits inside the case is simplified to fit. The case size is the 2006 team&\#39;s estimate, which they call uncertain.


---

*07 · The front*

*THE FRONT DIAL*

## One turn, one year.

The biggest wheel, called b1, has four spokes and 223 or 224 teeth: they are too damaged to be sure, and nobody knows why the maker chose that number. One turn of it is one year. In front of it the dial carries two rings: the twelve signs of the zodiac, marked in 360 degrees, and the Egyptian calendar of 365 days, which could be moved to make up for the quarter day the calendar lost every year.

A golden pointer carried the Sun round the zodiac. Another carried the Moon, and on it a small ball, half black and half white, turned to show the phase: black when the two pointers met, white when they stood opposite.

As you scroll through this chapter, the machine turns one year. The Sun goes round once; the Moon a little over thirteen times.


---

*08 · Inside*

*INSIDE*

## The arithmetic is in the teeth.

Open the case and the plates come apart. Behind the dial, the gears do sums. The Moon&\#39;s train starts with a wheel of 38 teeth, twice 19, and runs to one of 127, half of 254. Greek astronomers knew that in 19 years the Moon goes round the sky 254 times. The teeth make the pointer do the same.

Every other cycle on the machine is built the same way: a ratio of whole numbers, cut into bronze.


---

*09 · Turn the handle*

*NINETEEN YEARS AT THE HANDLE*

## Turn the handle.

Nineteen years of turning: the handle goes round about 88 times (223 teeth on the main wheel, 48 on the handle&\#39;s gear), the Sun pointer 19 times, the Moon pointer 254 times, and the upper back spiral through all 235 of its months, while the eclipse spiral passes the end of its 223 and starts again.

In order: YEAR 1 · YEAR 10 · YEAR 19: THE CYCLE CLOSES.

> **Note**: Every reading follows from the tooth ratios, from a start with every pointer at zero. The eclipse months are marked from the eclipse cycle, not copied from the fragments.

> **THE HANDLE**: Only the keyway for the handle survives, at the right-hand side; the handle itself is assumed. The drive went in through a crown gear, a1, with 48 teeth.


---

*10 · The Moon's limp*

*HIPPARCHOS'S MOON, 146 TO 128 BC*

## The Moon does not keep time.

The real Moon speeds up when it is near the Earth and slows down when it is far. Hipparchos worked out a theory for it in the second century BC. The machine builds that theory in bronze: two gears of 50 teeth, k1 and k2, on axles about 1.1 millimetres apart. A pin on one runs in a slot on the other, so the second is pushed a little ahead, then held a little back: up to 6.5 degrees either way.

Both ride on the big wheel e3, which carries them round once in about nine years, the time the Moon&\#39;s orbit itself takes to turn. Michael Wright saw the pin and slot first, and rejected it as a lunar device. In 2006, two gears of 53 teeth, an awkward number to cut, confirmed it.

It is difficult to understand how this superbly economical mechanical design was conceived. (FREETH AND JONES, ISAW PAPERS 4, 2012)


---

*11 · Eclipses*

*THE BACK · LOWER DIAL*

## It predicted eclipses.

Turn the machine round. The lower back dial is a spiral of four turns divided into 223 months: the Saros, the cycle after which eclipses repeat. A follower rode in the groove to show which turn to read. In a month when an eclipse was due, a small inscription: Σ for the Moon, from ΣΕΛΗΝΗ, H for the Sun, from ΗΛΙΟΣ.

The Saros is not a whole number of days but 6,585 and a third, so each cycle shifts the eclipses by eight hours. A small dial inside the spiral counts three cycles, 54 years, to put that right. The inscriptions went further than dates: direction, magnitude, even colour.

Above it, the upper spiral has five turns and 235 months: nineteen years, after which the Moon&\#39;s phases return to the same dates. The inscription names it: spiral divided into 235 sections.

> **THE GLYPHS**: On this model the eclipse months are marked from the eclipse cycle, not copied from the fragments. The spiral&\#39;s pointer runs through eighteen years as you scroll: one Saros.


---

*12 · The Games*

*THE BACK · UPPER DIAL*

## And when the Games were.

Inside the calendar spiral sits a small dial in four quarters, one turn every four years. It names athletic festivals: the four great crown games and lesser ones.

- YEAR 1 · Isthmia, Olympia: The Isthmian Games at Corinth, and the Olympic Games.
- YEAR 2 · Nemea, Naa: The Nemean Games, and the Naa at Dodona.
- YEAR 3 · Isthmia, Pythia: Corinth again, and the Pythian Games at Delphi.
- YEAR 4 · Nemea, Halieia: Unreadable in 2008; since read by Paul Iversen as the Halieia of Rhodes, games for the sun god.

The twelve month names on the calendar are Corinthian. In 2008 the team pointed to Corinth&\#39;s colonies in northwest Greece, or to Syracuse, the city of Archimedes. Iversen argues for the calendar of Epirus, and for a maker on Rhodes.


---

*13 · How we know*

*1905 TO 2016*

## A century of looking inside.

- 1905 · Albert Rehm: The first to understand it as a calculating machine. In his notes: the numbers 19, 76 and 223.
- 1958 · 793 letters: Derek de Solla Price, with the epigraphist George Stamires, raises the letters read to 793.
- 1971 · Karakalos and Price: Charalambos Karakalos X-rays the fragments; with Derek de Solla Price they count thirty gears and estimate their teeth. Price publishes Gears from the Greeks in 1974.
- 1990 · Wright and Bromley: Michael Wright and Alan Bromley see inside with linear tomography, an early 3D X-ray method.
- 2005 · The Research Project: A microfocus CT scanner, the BladeRunner, is brought to the museum in Athens; Hewlett-Packard's imaging relights the surfaces in software. Nature, 2006.
- 2016 · The inscriptions: The full corpus of the inscriptions is published.

The scans read text sealed inside the fragments for two thousand years. By 2016 about 3,500 characters had been read, which the Associated Press reported as a quarter of the original. The back cover explains what the front shows: Freeth calls it a user&\#39;s manual; Alexander Jones told the AP it was more like a long museum label.

Now we have texts that you can actually read as ancient Greek; what we had before was like something on the radio with a lot of static. (ALEXANDER JONES, TO THE ASSOCIATED PRESS, 2016)


---

*14 · The planets*

*2021 · THE FRONT, AGAIN*

## Venus 462, Saturn 442.

The front is where most of the machine is lost. The words survive better: the front cover lists the planets. In 2016 Alexander Jones read two numbers in it, 462 in the Venus section and 442 in the Saturn section. They are cycles: 289 of Venus&\#39;s in 462 years, 427 of Saturn&\#39;s in 442. Freeth writes that the makers seem to have found them themselves.

In 2021 a team at University College London used them to rebuild the front: nested rings for the Moon, the planets and the Sun, and the date, with the planets marked by stones. A lone gear of 63 teeth in Fragment D, unexplained until then, fits Venus: 63 shares the factors 3 and 7 with 462.

Because of the loss of evidence, we cannot claim that our model is a replica of the original. (FREETH AND COLLEAGUES, SCIENTIFIC REPORTS, 2021)

> **ON THIS MODEL**: The rings here carry each planet at its average rate, to show the layout; they are not the 2021 team&\#39;s gearing.


---

*15 · Who made it*

*NO SIGNATURE*

## Nobody signed it.

There is no name. The lettering dates it to about 150 to 100 BC. Working back from the eclipse dial, Carman and Evans in 2014, and separately Freeth, found a starting date in 205 BC: the date the dials count from, not the day it was built. Rhodes is the growing favourite for where.

Cicero, writing between the 50s BC and 45 BC, describes two machines like it. One was a bronze sphere by Archimedes, carried off from Syracuse when the city fell; in his dialogue a Roman named Gallus turns it and it shows the Sun, the Moon and the planets.

The invention of Archimedes deserved special admiration because he had thought out a way to represent accurately by a single device for turning the globe those various and divergent movements with their different rates of speed. (CICERO, ON THE REPUBLIC 1.22, TRANSLATED BY C. W. KEYES (1928))

The other, says a speaker in a later dialogue, was made recently by "our friend Posidonius", who taught on Rhodes. The researchers take Cicero as a good witness that such machines existed in his time, and are wary of following him further back.


---

*16 · Still there*

*ANTIKYTHERA · NOW*

## The wreck is still there.

Nothing as complex is known from at least a thousand years after it. Bronze was worth melting down, so if others were made, they went back into the pot. The earliest known device of comparable complexity is Giovanni de&\#39; Dondi&\#39;s astronomical clock: he spent sixteen years on it and finished it in 1364.

The fragments are in the National Archaeological Museum in Athens. The ship is still off Antikythera. Jacques Cousteau dived on it in the 1950s and again in 1976; the present excavation began in 2012, and in 2016 it found the skeleton of someone who sailed on the ship. In 2022 they raised a marble head of Herakles that belongs to a statue brought up in 1900. In 2025, the 125th year since the sponge divers, they lifted the first intact section of hull planking, perhaps from this ship, perhaps from a second, smaller one.

No new piece of the machine has been found. Two thirds of it are still missing.

## Sources & notes

1. [Freeth and colleagues, "Decoding the ancient Greek astronomical calculator known as the Antikythera Mechanism", Nature 444 (2006): the case of about 340 by 180 by 90 mm, Fragment A of about 180 by 150 mm, one turn of b1 a year, the pin and slot on axes about 1.1 mm apart giving 6.5 degrees, which Wright had identified but rejected as a lunar mechanism, the two 53-tooth gears, Hipparchos between 146 and 128 BC, the lettering of 150 to 100 BC, the 6,585 and a third days of the Saros, "spiral divided into 235 sections", "more complex than any known device for at least a millennium afterwards"](https://www.nature.com/articles/nature05357) (primary)
2. [Supplementary notes to Nature 2006: b1 with "223/224" teeth and "the reason for the builder's choice of 223 or 224 teeth is not known"; a1 of 48 teeth; the weight of Fragment A, 369.1 g](https://web.archive.org/web/20130126213307/http://www.antikythera-mechanism.gr/system/files/0608_Nature-Supplementary.pdf) (primary)
3. [Freeth, Jones, Steele and Bitsakis, "Calendars with Olympiad display and eclipse prediction on the Antikythera Mechanism", Nature 454 (2008): the five-turn Metonic spiral, the four-turn Saros spiral of 223 months with its glyphs, the Games dial and its names, the Corinthian months, northwest Greece or Syracuse](https://www.nature.com/articles/nature07130) (primary)
4. [Freeth and colleagues, "A Model of the Cosmos in the ancient Greek Antikythera Mechanism", Scientific Reports 11 (2021): 82 fragments, a third surviving, 30 gears and 27 of them in Fragment A, the phase ball, 462 for Venus and 442 for Saturn, the nested rings, "we cannot claim that our model is a replica of the original"](https://www.nature.com/articles/s41598-021-84310-w) (primary)
5. [UCL News, "Experts recreate a mechanical Cosmos for the world's first computer" (12 March 2021)](https://www.ucl.ac.uk/news/2021/mar/experts-recreate-mechanical-cosmos-worlds-first-computer) (primary)
6. [Freeth and Jones, "The Cosmos in the Antikythera Mechanism", ISAW Papers 4 (2012): the movable calendar ring, the keyway for the handle at the right, about a dozen gears visible, triangular teeth, gears e5, e6, k1 and k2 of 50 teeth, "It is difficult to understand how this superbly economical mechanical design was conceived", Cicero as a witness, and the earliest newspaper reports of the find (To Asty and Skrip, 21 May 1902, Julian) crediting Spyridon Stais](http://dlib.nyu.edu/awdl/isaw/isaw-papers/4/) (primary)
7. [Freeth, "Eclipse Prediction on the Ancient Greek Astronomical Calculating Machine Known as the Antikythera Mechanism", PLOS ONE (2014): direction, magnitude and colour in the eclipse inscriptions; an epoch in 205 BC](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0103275) (primary)
8. [Carman and Evans, "On the epoch of the Antikythera mechanism and its eclipse predictor", Archive for History of Exact Sciences 68 (2014): 205 BC](https://link.springer.com/article/10.1007/s00407-014-0145-5) (primary)
9. [Voulgaris and colleagues, "The Initial Calibration Date of the Antikythera Mechanism after the Saros spiral mechanical Apokatastasis" (2022): a calibration date of December 178 BC](https://arxiv.org/abs/2203.15045) (primary)
10. [Iversen, "The Calendar on the Antikythera Mechanism and the Corinthian Family of Calendars", Hesperia 86 (2017): the Epirote calendar, a maker on Rhodes](https://www.ascsa.edu.gr/publications/hesperia/article/86/1/129-203) (primary)
11. [Iversen, "The Antikythera Mechanism, Rhodes, and Epeiros" (2020): the Halieia of Rhodes in year four](https://archive.nyu.edu/bitstream/2451/61288/47/02.%20Iversen.pdf) (primary)
12. [National Archaeological Museum, Athens, "When the sea turns into a vast watery grave" (2016): the depth of 50 to 64 m, about 36 marble statues, the coins and a sinking around 60 BC, Cousteau's dives of 1959 and 1976](https://www.namuseum.gr/en/monthly_artefact/when-the-sea-turns-into-a-vast-watery-grave/) (primary)
13. [National Archaeological Museum, Athens, "The Mysteries of the Mechanism of Antikythera": the first researchers and the astrolabe; 793 letters in 1958; linear tomography from 1990](https://antikythera-mechanism.namuseum.gr/en/science-historian/) (primary)
14. [Nikolaos Kaltsas, director of the National Archaeological Museum, "The Antikythera shipwreck and the treasures", Proceedings of Science 170 (2012): a freighter of about 300 tons, its cargo of statues, bronze furniture, gold jewellery, glass and amphorae, the trade in Greek art to Italy](https://pos.sissa.it/170/006/pdf) (primary)
15. [National Archaeological Museum, Athens, exhibition "The Antikythera Shipwreck": "the first major underwater archaeological expedition", sponge divers with the Greek Royal Navy](https://www.namuseum.gr/en/temporary_exhibition/to-nayagio-ton-antikythiron-to-ploio-oi-thisayroi-o-michanismos/) (primary)
16. [Antikythera Mechanism Research Project, timeline compiled by Alexander Jones and Yanis Bitsakis: Easter 1900, the salvage months, the mechanism out of the sea probably in July 1901, Karakalos's radiographs of 1971 to 1972, the BladeRunner in October 2005, the inscriptions in 2016, de' Dondi as the earliest device of comparable complexity](https://web.archive.org/web/20220519044743/http://www.antikythera-mechanism.gr/history) (secondary)
17. [Antikythera Mechanism Research Project FAQ, "How was it recovered?": helmet suits with air hoses](https://web.archive.org/web/2016/http://www.antikythera-mechanism.gr/faq/general-questions/how-was-it-recovered) (secondary)
18. [Antikythera Mechanism Research Project FAQ, "Where was it made?": the growing consensus for Rhodes](https://web.archive.org/web/2016/http://www.antikythera-mechanism.gr/faq/general-questions/where-was-it-made) (secondary)
19. [Antikythera Mechanism Research Project FAQ, "Have any others been found?": bronze valuable and recycled](https://web.archive.org/web/2016/http://www.antikythera-mechanism.gr/faq/general-questions/found-any-others) (secondary)
20. [Tony Freeth, "An Ancient Greek Astronomical Calculation Machine Reveals New Secrets", Scientific American (January 2022): the storm, Elias Stadiatis and the "heap of dead naked people", the lump "the size of a large dictionary", gears "the size of coins", Rehm's notes, the 38 and 127 teeth, the planetary cycles found by the makers, the 63-tooth gear and Venus](https://www.scientificamerican.com/article/an-ancient-greek-astronomical-calculation-machine-reveals-new-secrets/) (secondary)
21. [Return to Antikythera, "The sponge divers and the original crew": Captain Dimitris Kontos, Ilias Stadiatis, Georgios Kritikos who lost his life](https://antikythera.org.gr/history/1900-1901/sponge-divers/) (secondary)
22. [Associated Press (June 2016): about 3,500 characters read, "a quarter of the original"; Alexander Jones on the static](https://www.pressherald.com/2016/06/11/writing-on-ancient-machine-deciphered/) (reporting)
23. [Cicero, On the Republic 1.21 to 22, translated by C. W. Keyes (1928), at attalus.org: Archimedes' globe, Marcellus, Gallus](https://www.attalus.org/cicero/republic1.html) (archival copy)
24. [Cicero, On the Nature of the Gods 2.88, translated by H. Rackham (Loeb, 1933), at LacusCurtius: the orrery recently made by Posidonius](https://penelope.uchicago.edu/Thayer/E/Roman/Texts/Cicero/de_Natura_Deorum/2B*.html) (archival copy)
25. [Woods Hole Oceanographic Institution, "Ancient Skeleton Discovered on Antikythera Shipwreck" (19 September 2016)](https://www.whoi.edu/press-room/news-release/skeleton-antikythera-shipwreck/) (primary)
26. [Return to Antikythera, 2022 season: the marble head of Herakles belonging to the statue raised in 1900](https://antikythera.org.gr/2022/06/20/2888/) (primary)
27. [Return to Antikythera, 2025 season (in Greek): the first intact section of hull planking raised, its thinner planks perhaps from a second, smaller ship](https://antikythera.org.gr/2025/07/01/3022/) (primary)
28. [Dictionary of Scientific Biography, "Dondi, Giovanni": sixteen years, completed in 1364](https://www.encyclopedia.com/science/dictionaries-thesauruses-pictures-and-press-releases/dondi-giovanni) (secondary)
29. [Marsyas, "NAMA Machine d'Anticythère 1" (2005), Wikimedia Commons, CC BY 2.5: the front of Fragment A](https://commons.wikimedia.org/wiki/File:NAMA_Machine_d%27Anticyth%C3%A8re_1.jpg) (archival copy)
30. [Marsyas, "NAMA Machine d'Anticythère 4" (2005), Wikimedia Commons, CC BY 2.5, resized: the back of Fragment A](https://commons.wikimedia.org/wiki/File:NAMA_Machine_d%27Anticyth%C3%A8re_4.jpg) (archival copy)
31. [Unknown photographer, the salvage on the Mykali (1900 to 1901), Wikimedia Commons, public domain](https://commons.wikimedia.org/wiki/File:Antikes_Wrack_von_Antikythera_03.jpg) (archival copy)
32. [Creative Commons Attribution 2.5 Generic licence (CC BY 2.5)](https://creativecommons.org/licenses/by/2.5/) (primary)

The machine drawn in this story is a reconstruction, not a replica: its gear trains use the tooth counts published by Freeth and colleagues (b1 is drawn with 223 of its 223 or 224 teeth), so the Sun, Moon, Metonic, Saros, Exeligmos and Games pointers turn at the rates the original was built to, but the positions of the axles inside the case are simplified, the plates' engraving is illustrative Greek, not a transcription, and the eclipse months on the lower spiral are drawn from the eclipse cycle, not copied from the fragments. The readouts start from a moment with every pointer at zero; that is not a historical date. The front planet rings show average rates only. Where sources disagree the story says so: the depth of the wreck (50 to 64 metres by the museum; other estimates are shallower), the date of sinking (between about 85 and 50 BC across the sources; around 60 BC by the museum's coin study), who first noticed the gears in 1902 and on which day (Spyridon Stais in the earliest newspapers, of 21 May Julian; Valerios Stais on 17 May in a later account), which front ring was movable (the calendar ring by ISAW Papers 4, the zodiac ring by the project's older FAQ), and when it was made (lettering of 150 to 100 BC; a dial epoch of 205 BC by Carman and Evans and by Freeth; a calibration date of 178 BC by Voulgaris and colleagues, 2022). The story of the diver and the dead naked people is Tony Freeth's retelling; no document from 1900 containing it was found. The count of characters read is the Associated Press figure of 2016; other counts differ because they count different things.

*Photographs: Marsyas, Fragment A front and back (CC BY 2.5, the back resized), and an unknown photographer of the 1900 to 1901 salvage (public domain), via Wikimedia Commons; file pages and licence in the sources*

*Model: Drawn for this story from the published tooth counts of Freeth and colleagues (2006, 2008, 2021); layout simplified*

---

*Antikythera is an Awescroll, a true story built as a single scroll. Read it as it was made: https://awescroll.com/003-antikythera. This text edition is for reading, quoting and indexing; please credit Awescroll and link the story.*