How Rome Took Cities: The Art of the Siege
The Roman legion was designed for open battle, but Rome won its empire through sieges as much as through field engagements. The ability to take fortified positions — to reduce cities that refused submission, to breach walls that geography or construction made seemingly impregnable — was as central to Roman military power as the legion’s battlefield performance. Siegecraft required different skills, different equipment, and different timescales than open combat, and Rome developed all three to a level of systematic competence that its opponents rarely matched.
The simplest siege technique was circumvallation: surrounding the target city with a continuous line of fortifications that prevented supply, reinforcement, and escape, and then waiting for starvation and disease to do the work. This was the method of choice when time was available and the objective was important enough to justify the commitment of forces. Caesar’s siege of Alesia in 52 BC is the canonical example — he built two concentric rings of fortifications, one facing inward toward the besieged town and one facing outward against the Gallic relief army, and held both simultaneously. The engineering was staggering: over thirty kilometers of circumvallation completed in under a month. Vercingetorix surrendered when the relief army failed to break through. The method was expensive in time and labor but essentially guaranteed success against any city that could not be relieved from outside.
When time was not available or circumvallation was impractical, the Romans applied direct assault techniques. The testudo — the tortoise formation, in which soldiers locked shields above their heads to form a continuous overhead cover — allowed troops to approach walls under missile fire. From close range, attackers could attempt to undermine foundations, apply battering rams to gates, or scale walls using siege towers that were constructed to match or exceed the height of the defenses. These towers were elaborate wooden structures on wheels, pushed forward on leveled ground or constructed ramps, equipped with a drawbridge at the appropriate height that could be dropped onto the wall’s parapet.
The construction of assault ramps — agger — was among the most demanding Roman engineering tasks. At the siege of Masada in 73 AD, Roman forces constructed an earthen ramp up the side of the mountain on which the fortress stood, a structure that is still visible two thousand years later and represents an investment of labor that makes sense only in the context of the Roman determination to leave no resistance unpunished and no rebel stronghold standing. Masada was held by fewer than a thousand people. Rome sent an entire legion, the Tenth Flavian, and several thousand auxiliary troops, and spent months building the ramp rather than accepting that the position was inaccessible. The message was as important as the military result.
Artillery supported siege operations in both attack and defense. The Roman ballista — a large tension-powered crossbow mounted on a stand — could project bolts or stones with accuracy at ranges up to several hundred meters. The onager — a torsion catapult — threw heavier projectiles with less accuracy but greater destructive effect. These were not weapons of mass destruction by any modern standard, but they were effective against personnel exposed on walls, against wooden structures, and against siege equipment. A well-defended city could make a Roman approach costly enough to influence the tactical decisions about whether to assault or circumvallate.
The psychological dimension of Roman siegecraft was deliberate and important. Roman practice — not universal, but consistent enough to be a policy — was to offer surrender terms before assault and to punish resistance with severity if the assault was required. A city that opened its gates could expect relatively mild treatment. A city that required a full assault could expect massacre, enslavement, or destruction depending on the circumstances. The destruction of Carthage in 146 BC and the sack of Jerusalem in 70 AD were not uncontrolled violence; they were calibrated demonstrations of what Rome was willing to do to cities that resisted. The audience for these demonstrations was every other city in the region weighing its options.
Josephus, the Jewish historian who observed the Roman siege of Jerusalem as a captive and then as a collaborator, describes Roman engineering and organization with a mixture of admiration and horror that is the most detailed eyewitness account of Roman siege operations that survives. The construction of assault ramps, the deployment of artillery, the organization of the circumvallation, the systematic destruction of the city’s resources — all of it is recorded with a precision that reflects both Josephus’s access and his evident belief that this level of military competence was worth documenting regardless of which side it served.
The cities Rome failed to take are as instructive as the ones it captured. Hatra in Mesopotamia repelled Roman assaults twice, in 116 and 198 AD, under Trajan and Septimius Severus respectively, exploiting terrain, walls, and the limits of Roman supply lines at extreme distance. The Rhine and Danube frontiers were not held by siegecraft but by the logistical impossibility of sustaining operations deep into territory that Rome did not control. The limits of Roman siege capability were ultimately the limits of Roman logistics: what could be supplied, what could be garrisoned, what could be held after the army moved on. Engineering could solve tactical problems. Strategy had a harder edge.