Many Roman campaigns involved siege operations. Sometimes this involved attacking enemy forts or military strongholds, but more often than not it was a case of attacking fortified villages, towns and cities. Such places might possess sufficient political importance to warrant their capture in any case, and when they were strongly defended by nature or artifice their loss was an even greater blow to enemy pride. Where the enemy field army could not be forced into a decisive battle the capture of one fortified place after another weakened the prestige of their leaders, so that either their support base collapsed, or they were forced to risk a battle. In other circumstances, the defeat of the enemy field army in open battle was followed up by a drive on, and the capture of, his capital or main cities. Such concerted pressure was frequently sufficient to force his capitulation.
Sieges were therefore often important components of a war, serving either to win a victory, to complete one, or to force the enemy to alter his strategy and fight on Rome’s terms. As with battles they could vary tremendously in scale. Wealthier, and more politically centralized peoples were more likely to possess large fortified towns, but a less united opponent could well possess large numbers of small strongholds. In 51 BC Cicero spent 57 days besieging the walled village of Pindenissus during his punitive expedition against the warlike tribes of Mount Amanus of eastern Cilicia (southern Turkey). The siege of a substantial town or city often took even longer and its outcome was uncertain. To conduct the siege the Romans needed to remain in the same place for months on end, which inevitably created serious supply problems. Sources of food and fodder from the area around a city would quickly be exhausted, even if the defenders had not already removed or destroyed everything. Other important supplies could also run short, and Josephus tells us that in AD 70 the Roman besiegers of Jerusalem stripped the surrounding land for miles around completely bare of trees as they gathered timber for construction and firewood. A besieging army had to devote effort to protecting its supply lines, especially if the enemy possessed a field army.
The defences surrounding Roman army bases in the 1st and 2nd centuries ad were relatively modest. This was not through any lack of engineering skill, but a reflection of their function as barracks rather than strongholds. Roman military doctrine was to leave their defences and fight the enemy in the open wherever possible. On those occasions when a fort or fortress was attacked, their simple ramparts, towers and ditches proved formidable enough to enemies lacking knowledge of siegecraft, as long as a sufficient number of defenders were present. Excavation at several army bases, notably in Britain, has revealed a layer of burnt material amongst the occupation levels. It is largely a question of academic fashion whether this is interpreted as a sign of deliberate destruction when the army decided to abandon the site or the result of enemy attack.
(Far Above) The fortress at Masada that came under siege in AD 73 (see p. 190).
(Above) Trajan's wars in Dacia (ad 101-02,105-06) were dominated by sieges Dacian strongholds like this one at Blidaru combined Hellenistic, Roman and native methods of fortification with the natural strength of the Carpathians.
Hellenistic styles of fortification already existed in much of the eastern Mediterranean and tended to be employed throughout the area until well into the Roman period. Client kings such as Herod the Great in Judaea constructed massive lines of fortifications around important cities such as Jerusalem, where the steep terrain added to their strength. Herod also constructed luxurious places of refuge such as Herodium, Machaerus and especially Masada, where a naturally strong position was made apparently impregnable. In the larger projects the walls were solid and composed of very large blocks of dressed stone.
A Hellenistic influence was also apparent in some of the hilltop towns (oppida) of Transalpine Gaul, such as Entremont where evidence has been found of a successful Roman siege in the late 2nd century вс. However, in the main the strongholds of northern Europe were constructed according to local patterns. Caesar described the strength of the native mums Gallicus, a type of stone wall given greater resilience by a box framework of wooden beams. The fortifications built by the Kingdom of Dacia suggest a range of influences, ranging from Greek to Gallic and Roman. Walls of essentially Hellenistic pattern, consisting of two facing walls and a loose rubble core, were strengthened by long wooden beams joining the two facing walls. In Britain native stone fortifications were rare, and most strongholds had earth and timber defences. The grandest were the great multi-banked hillforts of the Durotriges in the South West, where ditches and walls were arranged to permit the defenders to make maximum use of the sling. This was especially true of the approaches to their gateways, where a complex system of ramparts and ditches created a winding path along which it was easy for any attacker to lose his way, and where he would be constantly exposed to missiles against his unshielded side.
Methods of attack
There were essentially three ways of capturing an enemy stronghold:
1 Through starvation: This involved the blockade of the stronghold, preventing men or supplies from leaving or entering. In time it was hoped that the defenders would consume all their provisions, or run out of an essential such as water, and be faced with a choice between surrender or death. Sometimes such desperate situations led to extremely brutal behaviour. During the siege of Alesia in central Gaul in 52 вс (p. 192), Vercingetorix, the Gallic rebel leader, expelled all the civilians from the town so that the remaining food would only have to meet the needs of his warriors. Caesar refused to let the refugees through his siege lines, so that these unfortunates were left to starve to death within sight of the two armies. There were often rumours of cannibalism amongst garrisons running out of food, for instance at Numantia in Spain in 134 вс (p. 32). Starving a garrison into submission required time, especially if the defenders had prepared for the attack or possessed a port and so could bring in supplies by sea. Some cities were able to survive for years on end. The warehouses and cisterns cut into the rock at Masada (p. 190) to gather water from the very occasional rainfall of the area were capable of supporting its garrison almost indefinitely. Maintaining a sufficiently close blockade required a large number of troops to remain around the stronghold for a very long time.
2 By stealth: Some fortifications could be taken by a surprise attack. If a small force was able to approach by stealth and get into a town, it could then seize control of key positions, most notably a gateway, and admit the main body. Yet this was risky, for if the defenders became aware of their presence then the small force could easily be annihilated. At Amida in Mesopotamia in AD 359, a captured civilian led a group of 70 Persians by a secret entrance into the city. The Persians occupied a tower and were able to support a major assault from outside, but when this failed they were destroyed. Such attacks were most likely to be successful if the besiegers possessed detailed information about the defences and defenders, or were assisted by traitors within the walls. Rarely could this be guaranteed, and surprise attacks required a good deal of luck. Capsa in Numidia was taken when a Ligurian auxiliary hunting for snails discovered a way up the cliffs behind the town and was able to lead a party of legionaries up there. The Romans were ready to seize any opportunity for a surprise attack, but did not rely on this method.
3 By storm: The final method of capturing a stronghold was by direct attack, sending men over, through or under the wall to overwhelm the defenders. This was by far the most costly method, for in an assault all the advantages lay with the defender and it was likely that casualties would be high. Penetrating the main line of fortifications was by no means certain, and even if the army managed to break in, there remained the possibility that the defenders would rally and drive them out. Street fighting was comparatively rare in the ancient world, but when it did occur displayed all the brutality of more modern periods. So traumatic was a failed assault that it was usually impossible to make another attempt for days or even weeks. A large city like Jerusalem, divided into several districts and quarters each protected by its own perimeter wall, required the attackers to storm successive walls. Attacking a city directly was difficult and required not simply considerable engineering skill, but a good deal of raw courage and savagery. This was the most common method employed by the professional army.
The Roman army occasionally made use of light artillery on the battlefield. Some, known as carroballistae, were mounted upon small mule-drawn carts to improve their mobility. However, the chief use of artillery came in the attack and defence of fortifications.
Roman torsion artillery consisted of two basic types - single and double armed. The single armed, a pattern which would remain essentially unchanged throughout the Middle Ages, was rare before the 4th century AD. It had a single upright throwing arm and could lob or shoot a stone with considerable force. The Romans called it the onager, or wild ass, because of the engine’s powerful kick.
Ammianus describes an occasion when a misfiring onager killed one of its crewmen in a gruesome manner. It was aimed by pointing the entire machine at the target and was reasonably accurate. Far more common for most of the period were the more sophisticated two-armed engines or ballistae. These look something like a crossbow, although working on different principles, for the force was not derived from the tension in the arms, but, like the onager, the tightness of the coils of sinew rope holding them. Ballistae varied considerably in size, some being small enough to be man portable whilst others were massive, and could fire either bolts or stones with great accuracy and force. The lightest ballistae were often referred to as ‘scorpions’ (scorpiones).
(Above) Re-enactors demonstrating a reconstruction of a twoarmed engine or ballista A machine of this sort could fire bolts, but was more normally employed to shoot stone balls. Ballistae could be much larger than this but retained the same basic design.
(Above left) A reconstruction of a light ballista of the type normally employed as an antipersonnel weapon. The bolt fired by such an engine could travel with great accuracy further and with greater force than the arrow of any archer.
(Above) A reconstruction of a single-armed engine showing the very different shape of such weapons. Known as the onager, or ‘wild ass’, from its violent kick, catapults of essentially the same design would remain common throughout the Middle Ages
Although powerful, even the largest torsion engines were incapable of breaching a well-built and substantial stone or earth wall. They could knock down thinner parapets, or temporary structures added on or around permanent fortifications to increase their height or strengthen them in some other way. However, their primary role was as antipersonnel weapons. The attacker used his artillery in an attempt to drive the defenders from the walls and so prevent them, and especially their own artillery if they possessed any, from hindering the progress of his siegeworks or attacks by battering rams and siege-towers. Artillery missiles possessed such momentum that armour or other protective gear was of little value. The impact of a stone from a ballista on a human body was especially appalling. Josephus tells the story of a man being decapitated and his head flung hundreds of yards from the body. Even more gruesome is his tale of the pregnant woman hit by a stone which tore her apart and threw the unborn child some distance away. At Jerusalem in AD 70 the rebels began keeping a careful watch for incoming missiles and calling out ‘baby coming’ as a warning. Observing this the Romans, who had probably been carving ammunition from the local light-coloured stone, began to paint the projectiles a darker shade. This made them far less visible as they approached and once again increased the rebels’ losses to artillery fire.
The Siege of Masada
In AD 66 the province of Judaea rebelled against Roman rule. One extremist faction of political assassins - known as the Sicarii from the curved flick- knife they used - infiltrated and captured the fortress of Masada, built by Herod the Great as a luxurious place of refuge. In spite of initial successes, the Jewish rebellion was stamped out in several years of heavy fighting culminating in the storming of Jerusalem in AD 70. Only a few strongholds continued to hold out, and the last of these to resist was Masada. In AD 73 the governor of Judaea, Flavius Silva, led an army against the fortress.
1 The Romans: Legio X Fretensis and an unknown number of auxiliary units. Both the legion and auxiliaries were probably significantly under strength after years of campaigning and may have together mustered little more than 5,000 men.
2 The Sicarii (often, though mistakenly, referred to as Zealots): about 960 people including a significant proportion of non-combatants such as the elderly, women and children. The defenders were led by Eleazar Ben Yair, who came from a family with a tradition of militant resistance.
Masada is situated on a steep-sided, rocky hill, accessible only by a winding and difficult pathway on the eastern side. It was amply provided with storerooms - there was also space for some cultivation of crops on the summit - and had deep cisterns cut in the rock to gather and hold water from the rare rainstorms of this area. There was therefore sufficient food and water to supply the garrison for several years and the Romans realized that they could not hope to starve the enemy into submission. Since a direct attack up the eastern path was also unlikely to succeed, Silva ordered his men to begin the construction of a massive siege ramp against the sheer western side of the hill. To prevent any escape, and also as a stark reminder that they were now surrounded, the Romans built a line of circumvallation around the hill, strengthening it with six fortlets and a number of towers. Two larger camps were built behind the line. When the ramp was completed, a siege tower carrying a battering ram ascended it and battered a breach in the fortress’s wall. Before the final assault could go in on the next day, the Sicarii killed their families and committed suicide.
(Above Left) The main centre of the Jewish rebellion had been Jersualem, but a few strongholds like Masada on the west bank of the Dead Sea continued to resist after its fall
(Above Right) Camp F may have housed as much as half of Legio X Fretensis. It is built on a steep slope near the foot of the ramp.
(Above) The massive siege ramp built by the Romans at Masada was originally even higher, for the top was crowned by a section constructed from timber.
(Centre left) Camp C in the Roman siegeworks at Masada shows some traces of internal buildings. Note the bases of towers on the line of circumvallation.
(Above) A view to the west from Masada itself, looking down at Camp A, В - the larger structure in the centre of a sue comparable to F- and C. Note the line of circumvallation.
The Jewish Rebellion was over, and the Roman army had delivered a stark warning to their other provincial subjects of the relentless punishment which awaited any rebels. A small Roman garrison was stationed at Masada for some time after the siege.
The Siege of Jotapata, AD 67
‘One of the men standing on the wall beside Josephus had his head carried away by a stone, and his skull was shot, as from a sling, to a distance of three furlongs; and a woman with child was struck on the belly just as she was leaving her house at daybreak, and the babe in her womb was flung half a furlong away. So mighty was the force of these stone-throwers.
More alarming even than the engines was their whirring drone, more frightful than the missiles was the crash.’
Josephus, The Jewish War 3. 245-6 (Loeb translation).
Josephus’ description of the terrible power of missiles hurled by the Roman catapults at the siege of Jotapata in AD 67.
According to Vegetius, each century of a legion was supposed to have been equipped with a mobile scorpion (or carroballista), whilst every cohort possessed a larger ballista. Whether or not there was a fixed complement of artillery for a legion, it is probable that the number taken on campaign varied according to the local need. Artillerymen were drawn from the ranks of the cohorts rather than forming a separate unit. The size of engine and skill of their crews produced by each legion probably varied, and Josephus tells us that amongst the army suppressing the Jewish Rebellion those of Legio X Fretensis were thought to be the best. There is no direct evidence for auxiliary troops ever using artillery, but it remains possible that at times some units did.
Surrounding the enemy
In 52 вс, Caesar pursued the army of Vercingetorix throughout most of central Gaul before finally cornering it outside the hill-top town of Alesia. During the previous few months he had successfully besieged and stormed the strongly fortified town of Avaricum and failed to take Gergovia in a surprise attack. The Gallic army was large, making a direct attack on its camp and the town of Alesia itself highly risky. Instead Caesar surrounded the hill with a circuit wall 10 Roman miles long (known as a line of circumvallation), incorporating eight camps for his troops and 23 forts to strengthen the position. Towers were built along the line of the main earth and timber rampart at 119-m (130-yard) intervals to provide observation points and platforms for throwing or shooting missiles. The wall itself was some 12 Roman feet high (just under 4 m), had sharpened stakes sticking out from it to hinder anyone trying to climb over, and was protected by two ditches, the inner one filled with water. In front of the ditches were rows of stakes and a checkerboard pattern of individual stakes concealed in round pits, a trap which the soldiers nicknamed ‘lilies’ because of their shape. Caesar was aware that Vercingetorix had sent messengers to gather forces from all of the Gallic tribes to march to his relief. Therefore, once the line of circumvallation was complete, Caesar set his men to building another line of fortifications, this time facing outwards (and therefore known as a line of contravallation), extending 14 Roman miles. Thus, whilst Caesar blockaded Vercingetorix inside Alesia, his own army was effectively under siege from the much larger Gallic force attempting to come to the rebel leader’s aid. Fighting was fierce as the Romans struggled simultaneously to prevent the Alesia garrison from breaking out and the relieving army from breaking in. At times it looked as if the defences would be breached, but Caesar and his officers led an aggressive defence, rushing reinforcements to threatened sectors and at times leading counterattacks against the enemy flanks outside the Roman lines. When food finally ran out, Vercingetorix was forced to surrender.
Caesar’s army made use of extensive lines of fortifications at other times, most notably at Dyrrachium in 48 BC during the civil war. There Caesar’s army attempted to blockade Pompey’s more numerous forces, who in turn responded by building their own line of fortifications to prevent this. In this case it was Caesar who was eventually forced to abandon his aim and withdraw. Lines of circumvallation were used at other times in situations which were more clearly those of siege rather than part of the complex manoeuvrings of field armies. In 134 BC Scipio Aemilianus surrounded the Celtiberian town of Numantia, in Spain, with a wall strengthened by forts, traces of which have been found by archaeologists. Interestingly enough, he did this because he preferred to avoid battle with the Numantines and preferred to use his greatly superior numbers to impose a blockade.
Lines of circumvallation were not restricted to blockade, but were often used as part of a direct assault. At Jerusalem in AD 70, the Romans suffered some initial reverses when they tried to take the second of the three walls surrounding the city. Great siege ramps built up against the fortress of Antonia in 17 days of hard labour had just been destroyed by fire or undermining. At this point, Titus ordered his men to construct a dry stone rampart surrounding the city. This huge task was completed in just three days, the units competing with each other to finish their section first. After this the Romans returned to preparations for an assault on Antonia and over the next weeks captured the remaining parts of the city piece by piece. The line of circumvallation served the practical purpose of hindering any attempts by the defenders to break out. In addition it was also a strong visual statement of the Romans’ intent. The defenders had taken refuge behind a wall for their own protection. Now the Romans had encircled them with another wall to make certain of their destruction. The point was even more clear at Masada, where the most complete remains of a Roman siege can be seen on the ground today. At the beginning of the siege in AD 73 Legio X Fretensis completely surrounded the hilltop fortress with a wall strengthened by five forts and a number of towers. In many sections the wall ran along the top of the cliff on the opposite side of the valley, against which it was inconceivable that the rebels would try to break out. The wall was a constant reminder of the defenders’ isolation and hopeless position. They were trapped, and every day the siege ramp which the Romans were constructing against the hillside grew a little bit higher. The psychological pressure on the besieged was massive. At Machaerus, and probably also Narbata, where traces of Roman siege lines have also been found dating to the Jewish Rebellion, this grew too much and prompted surrender before the Romans launched their final assault.
A view of the reconstruction of Caesar's lines of circumvallation at Alesia. Wherever possible at least one of the ditches was filled with water. Light ballistae or scorpions were mounted in the towers to shoot down on any attackers.
At Alesia the Romans dug pits, fixed a sharpened stake in the centre, and then covered them over with vegetation. Caesar tells us that the legionaries with their macabre sense of humour nicknamed these traps ‘lilies’ (lilia) from their circular shape. The technique was often used to protect other army bases as is shown by this complex of lilies from Rough Castle in Scotland.
Breaking into a fortress
Since artillery was rarely capable of breaching a strong wall, other means had to be employed. Some walls could be undermined by men using picks and crowbars, as long as these were able to reach and work at the foot of a defended wall. It was in situations like this, far more than open battle, that the famous Roman testudo or tortoise formation was most often employed, with a roof of overlapping shields protecting the legionaries from enemy missiles. During one siege in the civil war of AD 69, the defenders grew so desperate at their inability to break the combined strength of the testudo with ordinary arrows and stones that they actually pushed over the wall one of their own great catapults. This smashed the interlocking shields and inflicted heavy loss, but also took with it much of the rampart and opened a breach into the town.
Although an essentially simple concept, the battering ram remained one of the most effective methods of knocking a hole through a wall. Save in the simplest of fortifications, gateways were usually avoided since, although wood could be burned, the chief ingenuity of the defenders’ engineers had usually been exercised to protect these spots. The army constructed massive rams mounted either in a wheeled shed or mobile tower, and consisting of a long beam, bound round with thick ropes to stop it from splitting, with an iron tip, usually shaped like the head of a butting ram. Both rams and towers usually required the construction of huge ramps from earth, timber and rubble, to allow them to cross the defensive ditches and moats and reach the wall itself.
(Above) A scene from Trajan’s Column showing a group of legionaries adopting the famous testudo or tortoise formation. With their shields lifted to form an overlapping roof above their head, the soldiers were well protected from most missiles This formation was often used to approach and undermine a wall with crowbars
(Above Right) Re-enactors from the Ermine Street Guard form testudo during a display in the amphitheatre at Caerleon. The historian Livy claimed that the testudo had first been developed as a spectacular entertainment performed at festivals in Rome.
Throughout the siege the defender would hope to hinder the progress of this work. Sallies were made to prevent the besiegers’ working-parties from performing their task, and to demolish or burn what they had already completed. Roman rams and towers were protected from fire by a covering of hides or iron plates. Once the rams reached a wall then it was only a matter of time before breaches were made. Therefore renewed efforts would be made to burn these engines. At Jotapata in Galilee in ap 67 the rebels managed to drop a large boulder and snap the head off one of the Roman rams. Another method was to lower sacks of straw over the wall and swing them onto the spot where the ram was expected to strike, deadening the force. Sieges were often contests between the ingenuity and determination of the attacker and defender.
An alternative to breaching a wall by battering ram was to dig a mine beneath it and fill it with combustible material. The props supporting the mine could then be fired, collapsing the tunnel and bringing the wall down with it. To guard against this the defenders dug counter-mines, judging the route of enemy tunnels from sound and vibration. These could be used either to undermine the attacker’s ramps and other works, or to connect with the enemy tunnels and allow an attack upon their miners. Dura Europus’ Roman garrison was besieged by the Persians in the 3rd century AD. Archaeologists discovered a Persian mine and a Roman counter-mine joining on to it. Inside were the bodies of Roman defenders and Persian attackers, killed in a gloomy skirmish or when the entire complex collapsed, undermining a section of the town’s wall and one tower. Tunnelling often carried risks for both sides. At Jerusalem the rebels managed by mining to destroy Roman ramps approaching the fortress of Antonia, but were dismayed when hours later the same tunnelling also caused the fortress to collapse and opened a route into the Great Temple.
The only other alternative to breaching or undermining a wall was to climb over it. At its simplest this meant an attack by escalade, the assaulting troops putting ladders up against the wall and climbing them. This required careful judgment, for it was important to ensure that the ladders were long enough to allow them to reach the parapet. Soldiers climbing ladders were horribly exposed to missiles of all types, and the ladders could be overturned or broken. Even if the men managed to fight their way onto the wall they could easily be overwhelmed, for it took time for more men to climb and reinforce them. For the same reason it was very difficult to retreat, and the leaders of such an attempt were almost invariably killed if it failed. A more effective method was offered by siege towers which could lower a drawbridge onto the rampart and permit men to approach the wall in cover and climb onto it in larger numbers. Even so the fighting could be vicious as the defenders met the men crossing the bridges.
A scene from Trajan’s Column showing legionaries and auxiliaries attempting to storm the Dacian capital of Sarmizegethusa by escalade. On the extreme right an auxiliary soldier climbs a ladder and strikes at the enemy while holding a severed head in his left hand. Abo notice the ladder carrying party on the left.
The assault and sack
Whichever method the attackers employed to assault a stronghold, the actual storming was an extremely difficult and dangerous operation. Technology could make a breach or assist a soldier to climb a wall, and perhaps help to suppress the defenders, but ultimately the storming party had to climb sword in hand up into the town. Casualties were often heavy, especially amongst the boldest men who led the way, and the chance of failure was high. If the main line of defences were crossed and troops entered the town then it was easy to become lost in the maze of narrow streets composing many ancient settlements. In ad 67, attacking the hillside town of Gamala on the Golan Heights, the Roman troops had trouble making progress through the alleys, most of which were just wide enough to permit passage of a donkey with panniers. The legionaries climbed onto the roofs of houses and began pushing up the slope, by scrambling from one building to the next. However, the weight of men packed onto the houses caused roofs and buildings to collapse. As the attack lost momentum, the defenders rallied and counterattacked, chasing the Romans back out of the town. At Jerusalem in ad 70 the Romans struggled for weeks to capture the Great Temple even after they had managed to cross its outer wall and gain a lodgement. The siege of a large settlement defended by a determined opponent was a slow, gruelling business, which wore down the soldiers’ morale, as one attack was followed by another.
The town of Gamala in the Golan Heights was besieged by the Romans in AD 67. A breech in the wall is visible just to the left of centre in this picture and stones and arrow heads from Roman artillery were excavated on this site Although the first assault failed when the attackers lost momentum in the maze of narrow alleyways threading through the town, a second attack proved more successful. As at Masada, many of the defenders chose to commit suicide rather Hum face capture.
The philosopher-Emperor Marcus Aurelius commemorated his Danubian campaigns in a column that was at once more stylized and in many ways more brutal than Trajan's Column. In this scene we see Roman troops burning a native settlement. On the far right an auxiliary infantryman prepares to behead a kneeling German captive The price of fighting against Rome was always high.
When an army finally did capture a stronghold, its occupants, civilians as well as active defenders, were subjected to a brutal sack. Polybius claimed that the Romans deliberately caused as much destruction as possible, slaughtering and dismembering animals as well as people, to deter other communities from resisting Roman demands to surrender. Convention developed into a law that the defenders were allowed to surrender on favourable terms if they did so before a Roman ram touched their wall, but that otherwise they could expect little mercy. Male inhabitants were usually slaughtered, women raped, though only in exceptional circumstances killed in the initial orgy of destruction. After that, as tempers cooled and the desire for profit took over, prisoners would be taken for sale as slaves, though at times any considered to have a low market value, such as the very old, were still massacred. Looting was widespread. In theory it was organized and the army pooled all its plunder for a fair distribution, but the practice may not always have been so neat. Josephus tells us that after the capture of Jerusalem in AD 70 the price of gold was devalued throughout the eastern provinces as the soldiers returned to their bases and started disposing of their plunder. Amidst the burnt levels of aristocratic houses destroyed in the sack of Jerusalem an arm was found. At Maiden Castle in Dorset the skulls of the defenders, including several already seriously wounded, were repeatedly hacked by the Roman attackers. Even more gruesome evidence of the horror of a Roman sack was found in Valentia in Spain, and probably dates to the capture of the town during the civil war in 78 BC. Skeletons were found showing signs not just of wounds received in the fighting, but of torture. One individual appeared to have been singled out for special attention, having his arms tied behind his back and a pilum thrust up his rectum. Perhaps a civil war created stronger passions and produced such atrocities, but it is important to remember that the ancient world was often an extremely brutal and unpleasant place.
Sieges figure heavily in the propaganda of ancient cultures from Egypt and Assyria onwards as amongst the proudest achievements of ‘Great Kings’. In part this was because attacking fortified places was always extremely difficult and the balance of advantage normally lay with the defender. For the professional Roman army, almost alone amongst the fighting forces of the ancient world, the balance for a time changed. Engineering skill, combined with the army’s characteristic determination and aggression, and a frequent willingness on the part of its commanders to accept the casualties inherent in direct assault, gave the Romans the capacity to capture any fortified place. More often than not they succeeded, even against such apparently impregnable fortresses as Masada. The army’s great proficiency in siege warfare gave it a marked advantage over all its opponents.