Siege Warfare in the Roman World

An extract from ‘Siege Warfare in the late Republic.’

Caesar’s Gallic sieges, 57–51 BC

By the time of Caesar, the legions had long been noted for their skills in field engineering, best illustrated by the camp they traditionally entrenched after each day’s march. Besieging armies are often mentioned building such a camp, or sometimes a pair of camps as Scipio had done at Numantia. However, the German scholar Willy Liebenam believed that he could discern a particular style of siegecraft that dispensed with all preparations in order to deliver a sudden and unexpected attack. Ironically, his inspiration came from the siege of Gomphi, a town in Greece which Caesar subjected to repentina oppugnatio (‘violent assault’) in 48 BC, when it shut its gates against him. But even here, the legionaries’ first act was to build a camp outside the town, and their second was to construct ladders, shelters and screens (Caes., BCiv. 3.80); the assault, when it came, was certainly swift, but Caesar’s preparations had been thorough. The situation at Cenabum (modern Orléans in France) four years earlier was very similar. Having arrived too late in the day to organise an attack, Caesar’s troops settled down and pitched camp. However, when the townsfolk attempted to flee in the dark, the legionaries sprang into action; firing the town gates, no doubt to illuminate the chaotic scene, they set about looting and burning the place (Caes., BGall. 7.11).

No fewer than 17 sieges are known to have been prosecuted by Caesar himself, and many involved the constructional skills of his soldiers. Nowhere is this clearer than at Avaricum (modern Bourges), a town almost entirely surrounded by marshes, except to the south, where the only approach route was obstructed by a deep gully. When he besieged the town in 52 BC, Caesar had to construct a great embankment so that he could bring overwhelming numbers of men across the gully and up to the walls; in 25 days, the massive structure, 330ft wide (98m) and 80ft high (24m), was complete. A similarly breathtaking feat of engineering was accomplished in the following year at Uxellodunum (Puy d’Issolu), where Caesar ordered the construction of a 60ft (18m) embankment, from which a ten-storey artillery-armed siege tower could target the fresh-water spring that was sustaining the inhabitants and prolonging the siege.

More usually, embankments served as runways along which heavy siege machinery could approach the walls of the besieged town. Sulla’s embankment at Piraeus had fulfilled this function, as had Lucullus’ at Themyscira and Pompey’s at Jerusalem. Caesar’s emban-kment at Noviodunum in 57 BC was of this sort: ‘after the shelters were speedily brought up to the town, an embankment thrown up, and towers erected, the Gauls were amazed by the size of the works, whose like they had neither seen nor heard of before, and, perturbed by the speed of the Romans, they sent representatives to Caesar to discuss surrender’ (Caes., BGall. 2.12). Similarly, ‘when [the Atuatuci] saw a siege tower erected in the distance, after shelters had been brought up and an embankment constructed, they at first jeered from their walls and ridiculed why such a machine had been built so far away’ (Caes., BGall. 2.30); but their scorn turned to alarm when the tower began its steady progress towards their walls, and they promptly sued for peace.

In all of these cases, for the chosen strategy to succeed, certain topographical features, such as the gully at Avaricum, made an embankment essential. Under different circumstances, an assault could be accomplished without one. For example, in 52 BC at Gergovia, atop a formidable hill accessible only from the south, Caesar decided to creep forward across the difficult terrain, consolidating ground as he went. From his initial encampment below and to the east of the hill, he seized the Roche Blanche, a small hill to the west, and ‘carried a 12ft double ditch from the larger camp to the smaller, so that even individuals could pass back and forth, safe from a sudden attack of the enemy’ (Caes., BGall. 7.36). Unfortunately, his plans were botched by the impetuosity of his troops, who were caught on disadvantageous terrain and repulsed; during the fighting withdrawal, no fewer than 46 centurions fell. In 1862, the archaeological remains of Caesar’s earthworks were uncovered by Colonel Eugène Stoffel, during a programme of archaeological excavations sponsored by Napoléon III to provide information for his Histoire de Jules César. More recent work by the Association pour la recherche sur l’Age du Fer en Auvergne (ARAFA) has confirmed the existence of Caesar’s two camps; but, at several points along the presumed course of the double ditch, only a single ditch was found, 1.70m wide and 1m deep, raising the possibility that the earthwork was not of uniform character over its entire length.

The earthworks at Gergovia were on a fairly small scale, and have more in common with field fortifications (for example, the ditches and artillery positions supporting the battle line at the Aisne in 57 BC; BGall. 2.8), than with siege-works. However, Caesar’s general readiness to throw a rampart around an enemy town is surprising, for the technique of periteichismos practised by Scipio at Numantia had not been used (as far as we know) for 25 years. Its last proponent had been Sulla, at the siege of Praeneste, when he ‘cut off the town at a great distance with a ditch and a wall’ (App., BCiv. 1.88); four years earlier, at Athens, he had ‘commanded the army to surround the town with a ditch, so that no one might secretly escape’ (App., Mith. 38). The tactic presumably appealed to Caesar in the 50s, as it had to Sulla in the 80s. Perhaps such large-scale earth-moving exercises helped maintain discipline amid the tedium which sometimes accompanied siege warfare. Certainly, Plutarch claims that, when M. Licinius Crassus confined Spartacus’ slave army in the toe of Italy in 71 BC by walling off the peninsula, part of his reasoning was ‘in order to keep the soldiers busy’ (Plut., Crass. 10.7).

Equally, experienced soldiers like Sulla and Caesar must have appreciated the demoralising effect that encirclement had on an enemy. In 52 BC, after Caesar spent two days surrounding Vellaunodunum, ‘on the third day, ambassadors were sent from the town to surrender’ (Caes., BGall. 7.11). If they had not, it is likely that Caesar would have launched an assault. This was certainly the case in the following year at Uxellodunum. Prior to Caesar’s arrival, his legate C. Caninius Rebilus planted three camps on the surrounding hills and ‘proceeded to carry a rampart around the town’ (BGall. 8.33); but it was Caesar’s attack on their water supply that led to the townsfolk’s surrender. Years earlier, in order to keep the Atuatuci within their walls while his embankment was under construction, Caesar had surrounded the town with ‘a rampart 15,000 feet [4.4km] in circumference, with closely-spaced forts’ (Caes., BGall. 2.30). Here, the investment was simply a prelude to an aggressive assault. Caesar’s contemporary, the prolific letter-writer Cicero, claimed to have used a similarly aggressive technique when he besieged Pindenissus in 51 BC; summarising the whole operation in a letter to his friend M. Porcius Cato, he wrote: ‘I surrounded the town with a rampart and ditch; I penned it in with six forts and large camps; I attacked it with embankments, shelters and siege towers’ (Ad fam. 15.4.10).