Thirty Gears, One Shipwreck, No Siblings

What We Know

In the spring of 1900, a crew of Greek sponge divers under Captain Dimitrios Kontos took shelter from a storm off the tiny island of Antikythera, between Crete and the Peloponnese. One of them, Elias Stadiatis, went down to look around and came up talking about a heap of “rotting corpses and horses” on the seafloor. Kontos went down himself to check. What Stadiatis had seen wasn’t bodies. It was bronze and marble statuary, corroded and half-buried, from a Roman-era cargo ship that had gone down roughly two thousand years earlier.

Divers spent the next two seasons hauling up wonders: bronze figures, glassware, pottery, coins. Among the wreckage, unremarked at the time, was a corroded, unassuming lump of bronze and rotted wood. It sat in storage at Greece’s National Archaeological Museum for months before anyone noticed it had split apart on its own to reveal gear teeth underneath the crust.

It took decades, and eventually imaging technology that wouldn’t exist for another seventy years, to work out what the lump actually was: a hand-cranked bronze machine, known today from 82 surviving fragments holding at least 30 gearwheels, built to track the sun, the moon, and the five visible planets, and to predict eclipses years ahead of time. Its inscribed eclipse calendar is calibrated to begin in 205 BCE. The device itself is dated later, though — judged by its lettering and the Rhodian and Corinthian calendar references cut into its dials, usually to decades afterward, roughly 150 to 100 BCE. Coins recovered from the wreck date the ship’s sinking later still, to around 70–60 BCE: the mechanism was already old when it went into the water.

Nothing else like it has ever turned up. Not from the centuries before it was built. The nearest thing to a rival is a far simpler geared sundial-calendar from the Byzantine world, five or six centuries later — nothing approaching the Antikythera mechanism’s complexity survives until medieval European astronomical clocks, roughly a thousand years on. One device, this sophisticated, with no known ancestor and no known descendant.

The Roadblock

A machine this complex doesn’t get invented in an afternoon. Building it required someone who understood differential gearing and could cut dozens of bronze teeth precisely enough to mesh without slipping. Ordinarily, it would also require someone who had already built simpler versions and learned from their mistakes. That is what makes the mechanism’s isolation so strange. Sophistication like this is usually the visible tip of a tradition: workshops, teachers, apprentices, failed attempts. There should be cruder earlier versions at a minimum. There aren’t any.

Part of the problem is what doesn’t survive at all. Ancient bronze was valuable enough that it rarely stayed idle — it got melted down and recast into something newer, generation after generation, at a scale wide enough that bronze objects of any kind are far scarcer in the archaeological record than their ancient abundance implies. If devices like this existed elsewhere, they may just have been unmade, long before anyone thought to set “priceless calculating machine” apart from “scrap.” The mechanism itself only escaped that fate by accident. It sank, rather than aging out through ordinary reuse on land.

That leaves investigators working from a sample of one. Every theory about where it came from, and why nothing else did, has to explain both the object and the silence around it. No single piece of evidence settles which.

Best Guesses

Unique, rare commission

A handful of devices from one tradition, every single one vanishing without a trace, still strains belief — that’s the problem sitting on top of historian Alexander Jones’s own theory. His case rests on rarity, not uniqueness: a custom, high-end commission, built in Rhodes for a buyer connected to Epirus, its dials citing both a Rhodian festival and an Epirote one, designed as a teaching instrument rather than a research tool. He isn’t claiming it came from nowhere, either. He points to Cicero’s own account of similar, probably simpler, devices built by Posidonius of Rhodes around the same time.

Lost tradition, bronze recycled

The most commonly cited explanation starts with bronze, not with the mechanism itself. Ancient metal was too valuable to leave lying around; statues, tools, and instruments were routinely melted down and recast, which is why bronze objects of any kind survive far less often than ancient sources suggest they once existed. Other devices like it may have existed and simply been unmade. This one survived only because it sank instead of aging out on land. No fragment, no workshop scrap, no inventory line has ever turned up to back the wider tradition up.

Undiscovered wrecks hold more

The most literal answer: keep looking. The Return to Antikythera project has revisited the wreck nearly every year since 2012, recovering new statuary, timbers, and cargo each season. In 2017, divers raised a corroded bronze disc stamped with a bull, and for a moment it looked like a second mechanism’s gear. The dig’s own co-director was candid: “It is maybe decoration for furniture or maybe a seal, or it could be an instrument.” Closer study found the disc wasn’t cut precisely enough to function as a gear — probably decorative, though never conclusively identified. Hope keeps outrunning evidence here, and it won’t be the last time it does.

Ancient texts, underexamined

This theory needs no new artifact. It argues the answer may already be written down. Cicero, who studied in Rhodes under Posidonius, twice describes bronze devices modeling the sun, moon, and planets: one built by Archimedes and seized as war plunder when Syracuse fell in 212 BCE, another built by Posidonius himself. Centuries later, Pappus of Alexandria references a lost Archimedes text titled, simply, On Sphere-Making — testimony, from named people, that machines like this were built more than once. None of it mentions gearing, the mechanism’s real innovation. Cicero may be describing something far cruder than what came out of the sea.

The Verdict

The Return to Antikythera excavations are still working the same wreck, every season, and divers have not exhausted what it’s holding. A second wreck holding recognizable gear fragments would settle the “how rare” question outright; a documented ancient inventory or workshop scrap listing a device like this would do almost as much. Nobody has found either yet.

Absent that, the case comes down to how much weight two-thousand-year-old testimony can bear against two-thousand-year-old silence. Cicero and Pappus both describe machines that model the heavens; neither describes gears. Until a second mechanism turns up, or a text precise enough to prove one existed is found, the honest answer isn’t which theory wins. It’s that one object is currently standing in for an entire tradition that may or may not have existed.

One gear-cut disc, still down there or still unread, could end this by itself.