Grim devastation: Roman Republican sieges in Hispania and Gaul

This article is from World Archaeology issue 137

Subscribe now for full access and no adverts

The terror of our Roman name means that all will know that the exhaustion of a long siege or a bitterly harsh winter cannot dislodge a Roman army from any city it has surrounded. Only victory ends the army’s fight, and it will make war more by perseverance than speed. Such perseverance is necessary in every kind of warfare, but especially when conducting sieges, for the fortifications and natural defences of many cities make them impregnable, so only time itself can defeat them through hunger and thirst.

Livy (History of Rome 5.6.8-9)

So says Claudius Crassus in his speech supporting the siege of Veii, a city near Rome, traditionally dated to c.396 BC. Although the siege is largely mythical and Livy probably crafted the speech when writing about it some 400 years after the supposed event, Crassus/Livy could well reflect the Romans’ idealised view of sieges and the steadfastness of their army in conducting them successfully. Livy’s Roman readers who had been involved in sieges would know, however, that grim reality hides behind the simple ‘exhaustion of a long siege’ and a besieging Roman army could fail.

As the Roman Republic progressed, the Roman army steadily developed into a highly efficient military machine, becoming one of the most ‘professional’ amateur militias in the ancient world. As well as the Romans evolving their military capability through their own initiatives, they embraced the benefits of incorporating successful features of the armies and strategies they encountered. This included learning the art of siege from Greek, Hellenistic warfare. Accounts of the Peloponnesian War indicate that siegecraft had been developing there from at least the 5th century BC. Rome was using such Hellenistic-based siegecraft by the time of the First Punic War (264-241 BC), against Carthaginian-held cities in Sicily.

Hellenistic origins

The Hellenistic siege system involved establishing a series of camps on strategically favourable positions around the enemy city. There were generally two main camps, often on opposite sides of the city, and a number of intervening smaller forts, all linked by a line of fortifications facing the besieged city and forming a blockading ring. If attack by an enemy relief force threatened from behind, a second, rearwards-facing line of fortifications would be constructed.

While there was still time before total blockade, the besieged may send ‘useless mouths’ (the elderly, women, and children) away to reduce the threat of famine. But these would have to run the gauntlet between enemy forces, and so may have come to a quick end.

To protect the army while siegeworks were constructed, preliminary defences such as a simple palisade and ditch could be created in front.

Siegeworks were not always continuous around a city, but could be interrupted where topography or natural obstacles made them superfluous, or alternatively when linear defences were only required in particularly vulnerable areas.

Fortifications

A rampart formed the main element of all Roman fortifications. It was constructed from locally available turf, stone, or earth, varying according to local conditions. A timber palisade or chevaux-de-frise was usually placed on top.

Towers could be erected along the rampart to improve observation and strengthen vulnerable locations with extra firing positions. They also provided platforms for artillery. These were important siege and assault weapons because of their range and accuracy. Iron artillery bolts are often found at the sites of such engagements.

There would usually be a ditch running along the front of the rampart, ground conditions permitting. This both increased the strength of the defences and provided material for the rampart. Additional obstacles and ditches could be provided in front.

Comparatively well-preserved stone structures in Spain at the battlefield of Pedrosillo and at Numantia indicate two of the forms such fortifications could take. Alesia in France is an especially good example of how elaborate siegeworks could be.

The huge timber requirements for siege fortifications would have had immense environmental impact. The siege of Jerusalem (AD 70), for example, saw trees cleared up to nine miles away. The struggle for Massilia (Marseille) (49 BC) between Caesar and Pompey saw the loss of extensive woodland, made worse by Pompey’s forces felling and removing trees to deprive Caesar’s besieging army of timber. The Pompeians repeated this for Caesar’s siege of Urso (Osuna, Spain) in 45 BC, clearing trees within six miles of the settlement and forcing Caesar to transport timber from his recently taken city of Munda.

Accommodation for the besieging troops would have followed normal Roman army practice and taken the form of leather tents, which could have walls and thatch added for better weatherproofing and warmth. Sometimes timber, mud-brick, or stone structures replaced tents, presumably for longer sieges or in particularly bad weather.

Sieges in Hispania

Sieges and direct assault of large indigenous settlements, referred to in the Roman literature as oppida, featured significantly in Rome’s conquest of the Iberian Peninsula, Roman Hispania, which began during the Second Punic War (218-201 BC), but took until the late 1st century BC to complete.

One of the best-known examples of such conflicts is Numantia (near Soria). This oppidum was a centre of Celtiberian resistance and managed to resist at least six Roman assaults and one siege between 152 and 135 BC. Its final defeat required a complex siege by Scipio Aemilianus in 133 BC, which concluded with the inhabitants famously choosing suicide over capture. Thanks largely to Adolf Schulten’s influential archaeological work here in the early 20th century, and Appian’s ancient account of the siege (Spanish Wars 90-98), it is known that Scipio’s siege followed the classic Hellenistic model. An initial defensive palisade and ditch were created around the city. Two main camps were established opposite each other (probably Castillejo and Peña Redonda), Scipio commanding one and his brother the other, with seven intervening forts. The encirclement was completed by a linking rampart (no evidence of a ditch), interval towers, and fortlets blocking the river Duero. Clearly, this required immense physical labour, but the large number of troops available, the known construction speeds of other Roman and ancient sieges, and modern time-and-motion studies indicate the work may surprisingly have taken only a few days.

A further complex of camps near Numantia, at Renieblas, was investigated by Schulten, too. These camps probably also relate to hostilities against Numantia, perhaps forming initial bases or supply depots. There is evidence here of forces associated with the Roman civil war that played out in Hispania c.82-72 BC. This started when the senior Roman official Sertorius claimed the title of governor, in opposition to the Sullan regime that was established in Rome. Despite initial successes against Sullan forces, including those commanded by Pompey, the tide eventually turned and the conflict fizzled out after Sertorius was assassinated in 73 BC. When considered alongside evidence of destruction in Numantia itself, as well as a possibly related camp at Peña Redonda-Caracierzo, it seems this conflict may have caused the (now Romanized) city to suffer yet another assault or siege.

If so, it was only one of the numerous city sieges and assaults carried out during the Sertorian War. One of the best known is at Valencia, where Pompey’s victorious forces carried out brutal reprisals. Many buildings were destroyed by fire and several badly mutilated skeletons have been found scattered in a public area, displaying signs of torture and violent deaths, including tied or severed hands, ropes around necks, and a man impaled on a spear. This is sadly not the only case of Roman army brutality against a city that chose to resist it – with such acts sending a warning to others.

Cabezo de Alcalá is another notable Sertorian War siege. Remains of burning, street barricades, and large amounts of Roman weaponry inside the city indicate it suffered violent destruction. Outside, a siege ramp leads from a camp to a gap in the city rampart. Recent investigations revealed a complex siege system of banks and ditches, small enclosures of uncertain purpose (fortlets?), plus 86 lead sling bullets and many other projectiles, including artillery ballista bolts.

The oppidum of Pallantia (Palenzuela) experienced several Roman attacks between 151 and 72 BC. Evidence of these has recently been identified and consists of linear features indicative of possibly more than one phase of siegeworks, as well as five camps.

There is curiously limited archaeological evidence of sieges, or military activities in general, relating to the Caesarian–Pompeian conflict in Hispania (49-45 BC). One example can be found at Ulia (Montemayor), besieged in 48 BC and 46-45 BC. Here, the density and precisely recorded findspots of hundreds of sling bullets and artillery projectiles indicate the location of siegeworks and troop movements – with the main assault lines seemingly lying to the north and south of the city – and recent LiDAR images indicate possible camps, but the features are yet to be assessed by fieldwork.

Recently found evidence for a siege at Cerro Castarreño is currently being investigated, but it may relate to a late 1st century BC context. There are inner and outer lines of siegeworks with regularly spaced camps.

Caesar and Gaul

So, when Julius Caesar led his armies into what are now France, Belgium, and the western part of Germany, collectively known to the Romans as Gaul, between 58 BC and 50 BC, Roman sieges were nothing new. Nevertheless, Caesar left a literary legacy in his Gallic Wars that promoted himself as being the master, almost an innovator, of siegecraft, even though his siege model was essentially nothing novel and Hellenistic in style.

Caesar’s sieges mainly occurred in 52 BC, with the exception of taking the oppidum of the Atuatuci in 57, the location of which is now uncertain. His capture of some oppida is dismissed in the Gallic Wars in a few sentences, but, for some reason, careful accounts are provided for sieges at Avaricum, Gergovia, and Alesia, and – to a lesser extent – for Uxellodunum. Caesar’s account of Alesia is, in fact, the most detailed we have of any siege in Greco-Latin sources. Few of these sieges are known archaeologically, the best being Alesia, followed by Gergovia, and Uxellodunum, thanks to investigations instigated by Napoléon III in the 1860s, which have since been augmented with more recent and ongoing research. As will be discussed, there is also now evidence of a little-known Roman operation at Hermeskeil, Germany.

Excavations of the siegeworks at Alesia (Alise-Sainte-Reine) revealed a complex of structures, varying in form according to local conditions. They are consequently not as uniform as Caesar would have us believe, but variously did have components he describes – a rampart with interval towers, one or more ditches, defensive obstacles and man-traps, and a front- and rear-facing line of siegeworks. These were constructed after the usual preliminary defences were created to protect the troops working on the siegeworks. There was the typical ring of several camps, and at least one ‘interval camp’ between the inner and outer siege lines. Two of the camps (B and C) can be proposed as the typical Hellenistic-style main camps.

Caesar seeks to promote his military genius by implying these siegeworks were unusually complex and extensive. The reality is they were within the normal realm of Roman military siegecraft; it would, in fact, be surprising if the Roman army could develop and confidently put into practice new techniques ‘on the spot’ and under hostile conditions.

Uxellodunum was besieged in essentially the same manner as Alesia. Three camps were established on nearby hills and then siegeworks were constructed around the oppidum. One difference was that measures had to be taken to divert a spring that provided the inhabitants with water. Locating Uxellodunum proved challenging for modern scholars, but its identity was confirmed in 1992 at Puy d’Issolud, after weaponry and features relating to the spring were found. The site of the camps and the siegeworks remain largely unknown, however.

Gergovia is generally regarded as being on the Merdogne plateau near Clermont-Ferrand, where there is a suitable oppidum and nearby Caesarian fortifications. An alternative candidate in the locality is presented by the oppidum at Gondole. Caesar says a large and a small camp were established, with a double-ditch linking the two to allow protected movement between the camps. Remains of these fortifications and further ditches, possibly linked to the siege or to action against Gondole, have been found. It is perhaps surprising that Caesar gives so much attention to his siege of Gergovia as he ultimately had to withdraw after an unsuccessful assault on the oppidum.

A two-phase Roman camp and annexe at Hermeskeil is the latest Gallic War site to be identified. Excavations since 2010 have produced finds that are comparable to those from Alesia. Hermeskeil lies close to the oppidum of Hunnenring and may have been used as an assault base on that, rather like the relationship between the complex of camps at Renieblas and Numantia. Alternatively, the camp could have formed part of general offensive operations in the area. Interestingly, it has been possible to associate this fortification specifically with Labienus, one of Caesar’s legates, from petrological analysis of the stones used for the soldiers’ hand-mills. This has allowed their provenances to be linked to Labienus’ known movements during the campaigning that took place in 52 BC.

Sieges are risky and grim

Caesar certainly seems to have liked sieges, since he conducted 17 of them himself, while his legates undertook several more. This may have made him an unusual Roman commander, for, when capturing cities, most commanders would probably want the quick and efficient result of a battle outside the city or a direct assault rather than a siege, as sieges demanded significantly more effort, for longer, and potentially in vain as there was no guarantee of success. Indeed, the many negative aspects of sieges could mean they were only commenced after a failed assault, or when a siege was the only option due to the strength of a city’s defences or the local topography making an assault unviable or unwise.

Compared to a battle or assault, the length of sieges meant that not only was an army delayed from other operations, but the men were kept in prolonged danger. They were especially vulnerable while constructing the preliminary defences, and even after the main siegeworks were completed, they potentially faced the constant threat of attack by the besieged and perhaps also by an enemy relief force (as occurred at Alesia).

The strength of such attacks could cause a siege to fail, but so could non- violent developments. If food or water supplies, always logistical challenges for besieging armies, or even the provision of sufficient firewood, proved inadequate, sieges would have to be abandoned. The environmental impact of siege armies could have an impact on their success as well. One of the main such impacts was the steadily accumulating sewage generated by soldiers and their animals.

The sewage was presumably at least partially carried out of the camps, if it was safe to do so, but it would have only been viable to dump it near the defences. To visualise quantities, the men and animals at Numantia would have covered a football pitch with about 1cm of faeces and filled an Olympic swimming pool with more than 7cm of urine – every day! The resulting stench would have been horrendous, and is remarked on by ancient authors, some commenting that it could even force Roman camps to be abandoned if they were more than briefly occupied. Sieges were even worse for pervading stench than these temporary camps, as battles left decomposing soldiers and horses between the siegeworks and the city, to which can be added the smell emanating from the suffering inside the city itself.

The sewage, decaying dead, plus general rubbish, lack of hygiene, and cramped living-conditions, meant that sieges entailed general suffering and an ever-increasing threat of disease for a besieging army. The situation could easily evolve into stark choice – death by the enemy or by the siege- camps. Soldiers at Numantia experienced such grim reality. While wintering there in 153/152 BC, men died for reasons attributed to the cramped conditions and bitter cold, made worse by low supplies. Similarly, dysentery and cold killed or caused suffering for many during the winter of 141/140 BC, eventually forcing the operation to be abandoned.

Under such conditions, the besiegers could easily have felt themselves under siege. Perhaps it was not just city residents who hated sieges.

Further reading:
• This piece provides a taste of the topics discussed in a recently published volume dedicated to the archaeology of Roman siege warfare in these regions: Mike Dobson (ed.) (2026) Roman Republican Sieges: operations in Gaul and Hispania (Oxford: Archaeopress).
• Additional information on some of the sites and siege techniques that are explored here can also be found in:
Andrew Fitzpatrick and Colin Haselgrove (eds) (2019) Julius Caesar’s Battle for Gaul (Oxford: Oxbow).
Michel Reddé (2025) La Gaule devant César: ce que révèle l’archéologie (Paris: Les Belles Lettres; in French).
Alan Wilkins (2024) Roman Imperial Artillery: outranging the enemies of the Empire (Oxford: Archaeopress).
David Espinosa Espinosa et al. (eds) (2025) Military Presence and Civic Integration in Hispania Ulterior from Sertorius to Caesar (Oxford: Archaeopress).