The Soldier Who Killed Archimedes

In the late spring of 212 BC, after two full years of siege, the Roman general Marcus Claudius Marcellus broke into the city of Syracuse through the Hexapylon gate on its northern wall. The opening had been engineered by a Roman officer who had noticed, during a festival of Artemis, that the Syracusan defenders were drunk on the ramparts. The garrison was put to the sword. The civilian population was, by Marcellus’s specific orders, to be spared. Marcellus had given strict instructions on one other point as well — that the elderly mathematician living in a house near the harbour was, on no account, to be harmed.

The instructions were not transmitted to every Roman soldier in the assault force. Sometime within the first hours after the breach, a junior soldier — never named in the historical record — pushed open the door of a small house in the western part of the city and found inside an elderly man, perhaps seventy-five years old, kneeling on the floor and drawing geometrical figures in a tray of sand.

The soldier, in the standard version of the story, ordered the man to come with him. The man, absorbed in his diagram, asked the soldier not to disturb his circles. The soldier — irritated, possibly mistaking the old man for a fugitive, certainly unaware of the older man’s identity — drew his sword and killed him.

The dead man was Archimedes of Syracuse, the mathematician whose work would shape Western science for the next two thousand years. The soldier was — when his identity later became clear — flogged and discharged from the Roman army on Marcellus’s personal orders. The Syracusan campaign that had killed the man who anticipated calculus was, by every available account, the most successful Roman siege of the Second Punic War.

It was also, in the older man’s death, the single most consequential Roman military error before the destruction of Carthage forty-six years later.

Why Rome was at Syracuse

Syracuse had been allied with Rome through most of the third century BC. Its long-reigning king Hiero II had supported Rome through the First Punic War and the early years of the Second, and the alliance had been a stable feature of the western Mediterranean political order. Hiero died in 215 BC, however, and the throne passed to his fifteen-year-old grandson Hieronymus, who almost immediately broke the Roman alliance and signed a treaty with Hannibal’s Carthage.

The political reasoning was complicated. Carthage had been winning the Second Punic War in spectacular fashion since Hannibal’s invasion of Italy in 218 BC. Cannae, the catastrophic Roman defeat in southern Italy, had happened in 216 BC. To a fifteen-year-old king watching from Syracuse, joining the apparent winning side made sense. Within a year, Hieronymus had been assassinated by an internal Syracusan conspiracy, but his successors — including the dual leaders Hippocrates and Epicydes — continued the Carthaginian alliance. Rome could not allow a hostile Syracuse to dominate the central Mediterranean. Marcellus was dispatched in 214 BC to bring the city back into the Roman orbit by force.

He arrived with a substantial army (the modern estimate is around twenty-five thousand troops) and a substantial fleet (roughly sixty quinqueremes). He expected the city to fall in months. It took two and a quarter years.

What Archimedes built for the defense

The reason for the extended siege was, by every contemporary account, Archimedes. He was approximately seventy-three years old in 214 BC and had been the leading mathematician and military engineer in Syracuse for at least four decades. King Hieronymus’s successors put him in charge of the city’s defensive engineering. He delivered.

Plutarch, Polybius, and Livy each give substantially overlapping descriptions of the war engines Archimedes designed. The key devices were:

• Massive cranes mounted on the harbour walls, fitted with iron grappling hooks. These could be lowered onto Roman ships approaching the seawall, locked into the ship’s bow, and then raised — lifting the ship by its prow until water poured into the stern and the ship capsized. Polybius, who interviewed survivors of the siege a generation later, gives the most precise description of these devices, which were operated by counterweights and could be deployed within seconds of a ship coming into range.
• Catapults of unusual range and accuracy, including one type that could fire stones weighing approximately ten talents (roughly 250 kilograms) at the Roman siege towers. Marcellus’s largest siege tower, the sambuca, was destroyed by one of these.
• The famous “claws” (manus ferrea), heavy hooked beams that could be swung out from the walls to drop down onto the decks of approaching ships, then withdrawn rapidly to drag the ships against the rocks.
• Possibly burning mirrors that could focus sunlight on Roman ships at close range and set the sails on fire. The mirror story has been contested for centuries — Plutarch does not mention it, and the geometry of focusing enough sunlight at useful range is borderline impossible — but it was repeated by later Byzantine and Arab writers and may have a kernel of truth in some other incendiary device.

The Roman assault from the sea, which Marcellus had counted on as the primary axis of attack, became impossible within the first weeks of the siege. Multiple Roman ships were sunk or capsized at the harbour walls. The fleet drew back beyond the range of Archimedes’s catapults and remained there for the next two years.

The siege became a blockade. Marcellus’s land force tried to assault the city through the Epipolae plateau on the western side, but Archimedes had also defensively engineered the landward walls — with similar machinery, though less is recorded about the specific devices used. The Romans starved the city. The defensive engines, by Plutarch’s account, made the Roman troops so afraid of the city walls that they would retreat at the sight of any rope or beam appearing above the parapet, assuming it was the start of an Archimedean attack.

The breakthrough came not from an engineering breakthrough but from the festival of Artemis. The annual three-day religious festival in the spring of 212 BC, by all accounts, was observed by the Syracusan defenders with the usual amount of wine. A Roman officer named Sosis, who had been studying the defenses at close range for months, noticed that the watch on the Hexapylon gate was visibly drunk by the end of the second day of the festival. He led a small commando team up the wall by ladders that night. They opened the gate. The Roman army poured in.

Marcellus’s order

Marcellus, by the consistent account of Plutarch and Livy, gave specific orders before the assault began that Archimedes was to be captured alive and brought to him unharmed. The order was based on Marcellus’s longstanding respect for Archimedes’s reputation — both his geometrical work and his terrible effectiveness as a defensive engineer. Marcellus, who was an unusually cultured Roman general for the period, had been negotiating throughout the siege for an eventual surrender that would spare Archimedes specifically.

The order, as we have seen, did not reach every soldier in the assault force. The killing happened in the chaos of the city’s capture, in the first hours after the breach, when Roman troops were spread across multiple districts. The soldier who killed Archimedes was probably acting individually, ordering the elderly man to come with him for processing as a prisoner. The geometrical figures in the sand were probably mistaken for a sign of senility or stubbornness.

The famous quotation — Noli turbare circulos meos in the standard Latin, or μή μου τοὺς κύκλους τάραττε in the Greek of Plutarch’s account — is reported by Plutarch with the explicit caveat that the exact words are uncertain. Plutarch gives multiple variants in his Life of Marcellus. Polybius, the earlier and more reliable source, does not give a death-scene quotation at all. The quote is therefore probably a literary embellishment of the early Imperial period rather than a contemporary record of what Archimedes actually said.

What Archimedes was working on at the moment of his death is also not certainly known. Plutarch suggests it was a geometrical problem — possibly related to the sphere-and-cylinder volume theorem he had taken particular pride in proving four decades earlier. Whether the figure in the sand was a circle, a paraboloid, or something else entirely cannot be reconstructed from the sources.

What Marcellus did afterward

Marcellus, informed of the death within the first day, by all surviving accounts wept. He had the soldier identified, flogged, and discharged. He gave Archimedes a state funeral at Roman expense and had the elderly mathematician’s family — a sister and possibly a nephew, the genealogy is unclear — granted Roman protection. He commissioned a tomb in the western necropolis of Syracuse, with an engraved depiction of a sphere inscribed in a cylinder (a geometrical relationship that Archimedes had specifically requested be marked on his grave).

The tomb was rediscovered, more than a century later, by the young Roman quaestor Marcus Tullius Cicero — sent to Syracuse in 75 BC on routine fiscal duties — who found it abandoned and overgrown in a thicket of bramble outside the city walls. Cicero had the site cleared and the inscription restored. He describes the moment, with some pride, in his Tusculan Disputations. The Syracusans themselves, Cicero notes drily, had largely forgotten where Archimedes was buried.

The tomb was subsequently lost again. Modern Syracuse has identified several candidate sites in the western suburb, but the Cicero account does not give precise enough coordinates to confirm any of them. A monumental tomb known as the Tomba di Archimede in the city’s Neapolis archaeological park is shown to tourists, but archaeologists are confident it is a Roman-era Hellenistic tomb of no particular connection to the mathematician.

The soldier who killed Archimedes is the only Roman in the historical record of the Second Punic War who is remembered specifically for this one act and otherwise entirely unknown. His name is not recorded. His age is not recorded. Whether he survived his flogging and went home is not recorded. He stands in for every junior officer of every military operation in human history who has made a decision that turned out, in retrospect, to be considerably more consequential than the decisions of the senior officers commanding him.