Two millennia underwater transformed the bronze into something else entirely.
Two millennia before the digital age, an unknown Greek craftsman enclosed the movements of the heavens inside a bronze box — a machine of gears meant to mirror the cosmos itself. Recovered in fragments from a Roman shipwreck, the Antikythera mechanism has long been celebrated as proof of ancient genius; yet a new Argentine simulation, built from the actual measurements of its surviving pieces, finds that those gears would have jammed within months. The discovery does not diminish the ancients so much as remind us that time and the sea are ruthless editors, and that what survives of a great idea is rarely the idea at its best.
- A digital simulation of the 2,000-year-old Antikythera mechanism — built from real fragment measurements — shows its gears jamming or disengaging, often before four months of movement could complete.
- The finding strikes at the heart of a cherished assumption: that the mechanism's sophistication meant it actually functioned as a reliable astronomical instrument.
- Two millennia submerged in seawater converted the original bronze into atacamite, a copper chloride compound that cracked and warped as it dried, potentially distorting every measurement researchers have relied upon.
- Rather than concluding the ancients were imprecise, lead researcher Esteban Szigety argues the opposite — that the original craftsmen were more skilled than the corroded remains suggest.
- The debate now turns on an unanswerable question: are scientists studying a masterpiece of ancient engineering, or only its disfigured ghost?
Two thousand years ago, someone in ancient Greece built a bronze machine to read the sky — a box of interlocking gears designed to track the sun, the moon, the planets, and predict eclipses. It sank with a Roman ship, and when divers recovered its eighty-two fragments more than a century ago, the world recognized it as evidence of a sophistication no one had credited the ancient world with possessing.
Now a simulation by Esteban Szigety and Gustavo Arenas at Argentina's National University of Mar del Plata has introduced a troubling complication. Using the actual measurements of the surviving fragments — including the V-shaped gear teeth and all their documented irregularities — the researchers ran the mechanism digitally. The gears jammed. They disengaged. The solar pointer rarely managed even four months of movement before the system failed. A device built to forecast celestial cycles across years ground to a halt in weeks.
The finding echoes a concern Michael Edmunds raised in 2006, when he identified significant gear misalignments suggesting the device would have been imprecise in practice. But Szigety draws the opposite conclusion: not that the ancients fell short, but that the damage we are measuring is not their work at all. Aristeidis Voulgaris explains that two millennia underwater transformed the bronze into atacamite, a compound that cracked and shrank as it dried, warping the geometry beyond recognition. What survives may be a shadow of the original — corroded, compressed, and misaligned by time.
Even so, Edmunds cautions that better tolerances alone would not have eliminated all mechanical error. The Antikythera mechanism remains an anomaly that resists resolution. The simulation does not solve the mystery of what it once was; it deepens it, suggesting that ancient precision may have been both more ambitious and more fragile than we have ever allowed ourselves to imagine.
Two thousand years ago, someone in ancient Greece built a machine to read the sky. It was a bronze box filled with gears—perhaps thirty or forty of them, each one cut with teeth meant to mesh with the others in perfect synchrony. The device tracked the sun, the moon, the planets. It predicted eclipses. Then it sank into the sea.
When divers found it in a Roman shipwreck more than a century ago, they recovered eighty-two fragments. Archaeologists and engineers have spent decades trying to understand how it worked, treating it as a window into a sophistication we didn't know the ancients possessed. But a new simulation, conducted by Esteban Szigety and Gustavo Arenas at Argentina's National University of Mar del Plata, suggests the machine may have had a fundamental problem: it probably didn't work at all.
The researchers built a digital model using the actual measurements of the surviving fragments. They included the V-shaped teeth on the gears and all the documented deviations from perfect geometry. Then they ran the simulation. In many tests, the gears jammed. In others, they disengaged entirely. The solar pointer—the needle that would have tracked the sun's position—rarely managed to complete even four months of simulated movement. A machine designed to predict celestial events across years or decades ground to a halt in weeks.
This finding reopens a question that Michael Edmunds raised in 2006: what if the preserved fragments are too damaged to tell us what the original mechanism could do? Edmunds had already identified significant misalignments in the gears. He suggested that such errors would have made the device imprecise as a practical instrument. But Szigety takes a different view. Rather than concluding the ancients were less skilled than we thought, he proposes they were more skilled. The deformation we see now, he argues, may not reflect the original craftsmanship at all.
Two millennia underwater transformed the bronze. Aristeidis Voulgaris, from the Culture and Tourism Directorate in Thessaloniki, explains that seawater converted the metal into atacamite, a copper chloride compound. When the fragments dried, the material cracked, shrank, and warped out of alignment. What we're measuring today may be a ghost of the original geometry—a shadow cast by corrosion and time.
Yet even this explanation has limits. Edmunds notes that even with better tolerances, the mechanical relationships between the gears would have introduced errors in the lunar pointer's movement. The Antikythera mechanism remains an anomaly: a device so advanced for its era that we still struggle to understand how it could have been made, and whether it could have functioned as intended. The simulation doesn't solve the mystery. It deepens it, suggesting that precision in the ancient world may have been both more ambitious and more fragile than we imagined.
Notable Quotes
The original craftsmen may have been more precise than the current deformed dimensions indicate— Esteban Szigety, National University of Mar del Plata
Two millennia in the sea transformed the bronze, causing it to crack, shrink, and lose alignment— Aristeidis Voulgaris, Culture and Tourism Directorate of Thessaloniki
The Hearth Conversation Another angle on the story
Why does it matter if this machine actually worked? It's two thousand years old either way.
Because it changes what we think ancient engineers understood. If it worked, they had mastered mechanical computation. If it didn't, we're looking at something else—ambition that exceeded capability, or a prototype that was never meant to be practical.
The simulation shows the gears jam. Isn't that the answer?
Not quite. The fragments are deformed by corrosion. We're simulating a broken thing, not the original. The researchers think the real machine may have been far more precise than what survived.
So we can't actually know.
Not from the fragments alone. But that's the point—two thousand years in saltwater changes everything. The bronze itself becomes something else. We're trying to read a blueprint written in rust.
What would it have taken to make it work back then?
Extraordinary skill. Gears cut by hand, fitted to tolerances we associate with modern manufacturing. And even then, the lunar pointer would have drifted. It was never going to be perfect.
Then why build it?
Because predicting eclipses was worth the effort. Because understanding the cosmos mattered enough to try.