Facing a defended stronghold, an attacker had three tactical options: to attempt to overcome the defenders and their fortifications by force; to establish a blockade and wait until the besieged ran out of provisions; or to seek control through negotiations. The third option might include discussions with the stronghold’s commander regarding mutually acceptable terms of surrender, or more clandestine talks with a person of lower status, who might be willing to help the besiegers gain entrance in exchange for suitable compensation. Regardless of where a stronghold was located, how large it was or what it looked like, each besieger was faced with this same set of options, and siege strategies more often than not incorporated measures to gain success through two of them, if not all three.
When examining the tactics and siege methods exhibited by various Frankish and Muslim forces, it is important to keep in mind the context and broader strategic aims of each campaign, as well as the differences between most Frankish and Muslim armies. Very generally, Frankish forces, at least at the highest echelons, were more heavily armoured, giving them an advantage in close-quarter fighting. This influenced the Franks’ preference for siege towers during the early twelfth century, which allowed them to bring their forces to a particular point along the top of a wall, compelling the defenders to fight man-to-man. Muslim and Mongol forces, by comparison, tended to be more mobile and more numerous, especially in the thirteenth century, employing proportionally greater numbers of bows. This allowed these forces to commit more men to attacks along multiple fronts, supporting direct actions with showers of arrows.
The essence of power in Frankish society carried an obligation of combat: boys born into the nobility typically became knights or clergymen – they would fight for their lord or for God. Accordingly, although the knightly class, which fought as heavy cavalry, rarely if ever made up the majority of an army, as a section of society, it was disproportionately represented. This core fighting force was supplemented by sergeants, who fought on horseback but are known to have also fought on foot in some instances. To counter the Muslims’ lighter cavalry, and disrupt their horse archers in particular, the Franks made use of their own irregular cavalry, known as turcopoles, who were recruited from local Christian communities or were of mixed Frankish-Syrian ethnicity.
Frankish infantry was raised from among the burgesses and almost certainly included significant numbers of local Christians. As in Europe at the time, the poor and most ‘common folk’ were typically excluded from the battlefield, due to the cost of weapons and personal equipment – the infantry was far from a body of conscripted peasants. Instead, many of these individuals would have been land owners, some would have been tradesmen, and others labourers. Regardless of their occupation, many would have had previous fighting experience. For men wealthy enough to afford arms, joining the army (a temporary commitment) was a means of helping secure Frankish interests, which in turn benefited their own safety and livelihood. Most who took part, however, would have been drawn by the prospect of wealth – they could expect both pay and plunder, if they survived.
Mercenaries, the private security contractors of their day, made up another part of many armies. They were typically recruited as preformed units, rather than individually, and as such often hailed from a common region and tended to specialize in a certain style of warfare, equipping themselves accordingly. These were individuals who fought primarily for cash, rather than any socio-political incentive. Some, who regarded war as a lucrative profession, came from wealthier elements of the common class, while others were members of the nobility, who saw more profit to be gained by fighting for cash rather than taking their position in the traditional political structure.
Turkish armies, which had come to dominate the Near and Middle East with the spread of the Seljuks, were typically more lightly armed than the Franks. They often relied on large numbers of horse archers, who could engage and disengage rapidly, waiting for an opportune moment at which to commit their forces to a close fight. The core of these forces was the ʿaskar, a regular force of cavalry that fought for a certain potentate. During the twelfth century, considerable numbers of Turkomans were also employed. These were recruited from regions like Khurasan, where seminomadic groups from the Eurasian Steppe, many of whom had only recently or loosely embraced Islam, were migrating into the Middle East. These men were often used as irregular cavalry and fought largely for plunder, making them exceptional raiders but less than ideal siege troops, prone to restlessness when plunder was not forthcoming. Bedouin forces were also recruited by both Egyptian and Syrian figures. Like their Turkoman counterparts, Bedouins often fought as irregular forces and were motivated by prospects of plunder. When the army they joined was defeated, it was not unheard of for Turkoman and Bedouin forces to loot their own side’s camp, securing what they could from an otherwise disastrous campaign. Many Bedouin groups were also not opposed to allying with Frankish forces if it served their interests. During certain periods, Muslim armies also benefited from significant numbers of volunteers, who answered the call of jihad. Not unlike European crusaders, most seem to have been motivated foremost by spiritual factors, travelling to take part in the fight against the Franks and then returning to their homes.
By all accounts, Fāṭimid armies were fairly diverse during the twelfth century, employing various units of cavalry and infantry. Like the Seljuks, the Fāṭimids made use of considerable numbers of ‘foreigners’ recruited from the fringes of their territory. The regiments of Africans from the Sudan and Ethiopia frequently caught the attention of contemporary Frankish historians, while the influence of high-level Armenian figures led to large numbers of individuals whose ethnic homelands were around the Caucasus. This led Fāṭimid armies to include both Shiite and Sunni forces, as well as regiments of openly Christian Armenians. Fāṭimid forces were so religiously and ethnically diverse because, unlike their Frankish and Turkish neighbours, these ‘foreign’ elements made up the core of their army.
Although most Muslim powers in the region depended on mamlūks (slave-soldiers) to at least some extent in the twelfth century, these formed a more significant part of the Fāṭimid army than those of contemporary Turks. Through the Ayyūbid period, Muslim rulers became increasingly reliant on bodies of mamlūks, who were bought as children and raised as Muslims to become soldiers in the service of their owner. From the late twelfth century through the thirteenth, these individuals often came from Christian, Cuman and Kipchak populations around the Black Sea. With the ascendancy of the Mamlūk dynasty, mamlūks remained the core of the army, often enjoying higher preference than the free-born sons of other mamlūks.
Before a besieger attempted to overcome the fortifications of a castle or town, considerable attention might be devoted to developing siege works or overcoming resistance from the defenders who ventured out beyond the stronghold’s walls. Almost all sieges involved some kind of effort to impose a blockade or surround the besieged community, even if this involved only scouting the defences for weaknesses or intercepting appeals for aid sent by the garrison. For defenders, keeping an open link to the outside world could be critical to their continued resistance, often leading them to take steps to inhibit the attackers’ efforts to cut them off.
If a siege were expected to drag on for a considerable period of time, or it was believed that the defenders would put up a particularly active defence, besieging forces might reinforce their position with a ditch, dug between their camp and the stronghold, to limit the threat posed by sallies. This was done by Tancred at al-Athārib (1110–11), and more famously by the Frankish besiegers of Tyre (1111 and 1124) and Acre (1189–91). At each of the latter three sieges, an additional ditch and defences were excavated beyond the Franks’ position to provide protection against field forces to their rear. During the siege of Damietta (1218–19), the crusaders entrenched themselves in a similar manner, providing a degree of protection against the Egyptian forces that attempted to break the siege. Elaborate siege works like these required considerable investments of time and effort, which besiegers rarely appear to have been willing to risk – the costs associated with developing such works left besiegers with more to lose if the siege failed.
Nevertheless, the incentive to build siege works becomes clear when considering the references to the fighting that went on outside some strongholds, especially urban centres, during the early phases of certain sieges. The Franks first encountered this at the siege of Antioch in 1097–98, where the defenders were able to launch effective sallies, and significant engagements appear to have been fought in front of the single landward gate of Tyre in 1124 and 1187. It is perhaps at the numerous ‘sieges’ of Damascus during the middle decades of the twelfth century that this fighting practice can be seen most clearly. During the Frankish attacks in 1129 and 1148, as well as during the many pushes against the city made by Zankī and Nūr al-Dīn before 1154, at no point did a ladder hit the top of a wall or a sapper begin to undermine a tower. Zankī and Nūr al-Dīn may have shown restraint, hoping to win over the Damascene population they meant to rule and facilitate a smooth transition of power, but this cannot be said of the Frankish campaigns. In the autumn of 1129, the Damascenes’ ability to rally and intercept a large Frankish army, led by Baldwin II, who was joined by Pons of Tripoli, Bohemond II of Antioch, Joscelin I of Edessa and Fulk of Anjou (later Fulk of Jerusalem), prevented the Franks from establishing a siege. In 1148, during the Second Crusade, led by Louis VII of France and Conrad III of Germany, the city’s defenders again managed to skirmish successfully in the gardens of the city, delaying any formal siege efforts until conflict among the various besieging parties led to the complete collapse of the expedition.
Attacks like these were in essence large coordinated sallies, intended to interfere with the besiegers’ ability to organize their siege efforts or to catch them off guard, weakening their strength and compromising their morale. Describing the protracted siege of Ascalon in 1153, Ibn al-Athīr remarked, ‘The inhabitants held firm and fought fiercely, some days even fighting outside the city wall. They repulsed the first group of Franks and drove them defeated back to their tents, pursuing them all the way. At that stage the Franks despaired of taking the city.’22 During the early phase of the siege of Acre in 1189, the numerical weakness of the Franks allowed the defenders to leave the city gates open, facilitating rapid sallies and forcing the Franks to keep a diligent watch lest they be surprised. Acre’s defenders were assisted by the presence of Saladin’s army, which had arrived only days after the Franks and compelled the besiegers to divide their forces in order to guard against attacks made by the field army. The city’s gates remained open from the start of the siege at the end of August until 16 October, when Saladin pulled back his army to al-Kharruba, about 12km from Acre’s walls. The Frankish defenders of Jaffa mounted a similar defence in 1192, engaging Saladin’s besieging forces beyond the town walls during the first three days of the siege. This disruption may have bought precious time, delaying Saladin’s ability to challenge the town’s defences and allowing Richard I to arrive just in time to prevent the citadel’s capture. Actions like these continued throughout the thirteenth century.
Reminiscent of the manoeuvres executed in 1129 by the defenders of Kafarṭāb and Damascus, Bohemond VI of Antioch-Tripoli intercepted Bertrand of Gibelet (Jubayl) as the latter moved against Tripoli in the late 1250s. Although Bohemond’s force was defeated and the prince was wounded, the action was sufficient to discourage a siege.
When Baybars attacked Acre in 1263, he did so with only part of his army, making a surprise appearance before the city one morning. He spent the first day surveying Acre’s defences while a contingent began undermining Doc, a fortified mill belonging to the Templars. The following day, Baybars awoke to find Frankish forces deployed in a prepared position outside the city. Although the Franks were eventually pushed back into Acre, they successfully defended its gates during the attacks that followed. This was probably a test of strength and show of force, rather than an earnest effort on the part of Baybars to take the city. The sultan was wary of committing his forces to a potentially lengthy siege with the Mongol threat still looming, but, like his later attacks in 1267 and 1271, he may have hoped to surprise Acre’s defenders and took the opportunity to thoroughly raid the surrounding region.
The Muslims remained for eight days before Acre, doing nothing besides engaging in the occasional clash between our forces and theirs, in which a few were killed on either side.
At the end of the eight days, they brought up and emplaced their siege engines . . .
They set up great barricades and wicker screens, ringing the walls with them the first night, and the second night they moved them further in, and the third night further still, and they brought them so far forward that they came up to the lip of the fosse. Behind these screens the armed men dismounted from their horses, bows in hand. And if you are wondering how they were allowed to draw so near, the answer is that they could not be stopped, as I shall now explain.
These people had their horsemen fully armed, on armoured horses, and they stretched from one side of the city to the other, that is to say, from the beach on one side to the beach on the other. There were more than 15,000 of them, and they worked in four shifts a day, so that no one was overworked. None of our men went out against those who were behind the screens, for if they had, those who were being [the first enemy line] would have defended them and barred the way, and so if it had happened that our men had gone out against them, the men on horseback would have defended them.
So in the end, the Muslims advanced to the edge of the fosse, as I have told you, and the men on horseback each carried four or five buches* on the necks of their horses, and threw them down behind the screens. And when night came, they put them in front of the screens, and bound a cord on top, and the pile became like a wall that no engine could harm, though some of our medium engines shot and battered at it without effect. The stones merely rebounded into the fosse.
After this, the enemy brought up their carabohas, small hand-operated Turkish devices with a high shooting rate which did more damage to our men than the larger engines did, since in the places where the carabohas were shooting, no one dared to come out into the open. In front of the carabohas they had made the rampart so strong and so high that no one could strike or shoot at those who were shooting. And this situation lasted as long as they were mining . . .
* Bundles of wood and other materials used by besiegers to fortify their siege works.
(Adapted from Templar of Tyre 490–1 [254–5], trans. Crawford, pp. 105–7.)
In 1290, Qalāwūn decided to bring the full weight of his army against Acre, but his untimely death meant that it was his son, al-Ashraf Khalīl, who would lead the Mamlūks against the city the following year. The siege saw a considerable period of fighting beyond Acre’s walls before siege efforts progressed and the defenders were pushed back into the city. The besiegers then entrenched their positions beyond the town ditch and steadily undermined the city’s defences.
If a besieging army opted for an aggressive strategy, this would often include the use of siege engines. Before discussing what these looked like and how various types were employed, it is important to point out that contemporary descriptions of siege machinery should be viewed critically. Very rarely do independent accounts give similar descriptions of the particular engines used at a certain siege; often, different engines are mentioned by different sources. To complicate matters, some sources had a tendency to include engines they believed ‘might’ or ‘should’ have been present at a siege when their information or understanding was incomplete. Most Frankish sources were clerics, typically strangers to the battlefield, and some were not above providing descriptions influenced by classical authors, believing this would give their work greater prestige or credence. In some accounts, regardless of the author’s ethnicity, it is possible to discern a learning curve as the author, who might otherwise have little exposure to siege warfare, gained greater insights through the compilation of his history. The initial or continued ignorance of others is often revealed through what are evident additions, containing obvious mistakes or misinterpretations, to the original account they were provided. At times, even the most reliable chroniclers and eyewitnesses with military experience recounted things with tunnel vision, focusing only on their part, that of their patron, or some other aspect during an engagement that they wished to emphasize, doing so at the expense of a more complete picture of events.
Such issues can be seen in the various accounts of Tancred’s siege of al-Athārib, which ran from October 1110 to January 1111. Matthew of Edessa reports that the siege involved Tancred camping against the town for a number of days before he took it by assault ‘without harming the garrison’. In Damascus, Ibn al-Qalānisī notes that the garrison was spared when the town fell and those who wished to remain were free to do so. Writing from Europe, Albert of Aachen provides perhaps the most detailed account of the siege’s progression. He describes the construction of siege engines over a period of days and the digging of a defensive ditch, but the breakthrough came on a Sunday after Christmas when a section of the citadel, battered relentlessly by artillery, supposedly fell and destroyed two towers as it collapsed down the slope. The besiegers were able to approach the damaged defences under the protection of a penthouse, but the rubble prevented them from storming the gap. Before this obstacle could be overcome, the defenders surrendered while favourable terms were still available. In Ibn al-Athīr’s account, the town was taken by force, but the focus is an episode of intrigue, in which the defenders planned to tunnel out of the citadel and murder Tancred in his sleep; their plan, however, was betrayed by a young Armenian. Finally, there is Ibn al-‘Adim’s rendition to consider. Composed in Aleppo in the thirteenth century, it is perhaps the best informed. Like Albert of Aachen, he notes the presence of artillery, but adds the use of a ram, the blows of which were said to have been heard great distances away. In this version of events, Tancred intercepted a pigeon carrying a note from the defenders to Riḍwān of Aleppo, which described the desperate situation within the town. This strengthened Tancred’s hand, who was at this point negotiating the town’s surrender, or the price of his own withdrawal, with Riḍwān – Tancred rejected a payoff of 20,000 dinars, demanding 30,000 and the release of prisoners. Some kind of an agreement involving a payoff seems to have been concluded between the two princes; but in a final twist, the terms of their arrangement were not met because the defenders threw open the gates of al-Athārib as soon as they received assurances of their safety, suggesting Tancred was simultaneously negotiating directly with the besieged.
Piecing together the various accounts, it seems a breakthrough in the siege occurred when Frankish forces broke into the town by force. Considering their dire position, the defenders, who had withdrawn to the citadel, opted to surrender before this was formally ordered by Riḍwān, for which they received the generous terms noted by Matthew of Edessa, Ibnal-Qalānisī and Albert of Aachen. Although we are left with a fairly clear picture of events, taking any one account on its own can leave a very different impression: something to be considered when we have just one or even two accounts of a given siege.
The simplest battering weapon was the ram, and a slight modification known as the bore. These engines had been used for millennia before the time of the crusades, but were rarely employed by Latin and Muslim forces in the Near East. Both were essentially a large beam, often capped with a heavy iron head, that was driven against a gate or wall. Rams made use of fairly blunt heads, relying on their weight to deliver a crushing blow, whereas bores used a head that came to a point, focusing the force of each strike. If the beam was light, it could be carried by its operators and swung by hand, otherwise it might be mounted on a cart or some other kind of wheeled framework, allowing it to be rolled back and forth. The most effective rams, however, were typically suspended inside a sheltered framework, allowing the beam to be swung against the masonry or gate ahead of it. This design allowed the operators to work the beam most efficiently and the surrounding structure, often a timber a-frame, could be used as a shelter for both the ram and those propelling it.
The energy of any battering weapon can be expressed by the basic equation for kinetic energy:
Because the beam of a ram could be propelled to only a meagre velocity, whether it was pushed or swung into action, its mass had to be enormous, limiting the use of such engines to situations where there was a relatively easy and flat approach. If a ram was heavy enough and its head sufficiently hard, the percussive force generated by repeated blows would steadily begin to crush the stones struck by the head. Rather than punching a hole in the wall, which was almost certainly much thicker than the head of the ram was long, by breaking up or dislodging a few ashlars, men with picks could begin to pry out these and neighbouring stones, allowing them to then chip away at the core of the wall. Accordingly, it is perhaps best to see rams as mining tools – a means of helping remove outer facing stones.
Although rams were relatively cheap and simple to construct, their cumbersome nature and relative ineffectiveness against sturdy masonry defences meant their use was in decline by the time of the First Crusade. In earlier centuries, rams had been used effectively in Europe against lighter fortifications, often employed against softer sections of defences, such as wooden gates. Fulcher of Chartres and Albert of Aachen include battering rams in the lists of siege engines that they claim were built by the Franks after arriving at Nicaea in 1097. Casual references to the construction of rams can also be found in the slightly later accounts of Guibert of Nogent and Robert the Monk. Critically, Fulcher was the only one of these sources at the siege of Nicaea, and the party that he was travelling with missed most of the siege, arriving only days before it ended. The author of the Gesta Francorum and a fellow eyewitness, Raymond of Aguilers, recount that more ambiguous ‘machines’ were constructed, which appear to have been traction trebuchets and penthouses to shelter miners, rather than rams. Robert the Monk subsequently mentions rams at the sieges of Antioch and Maʿarrat al-Nuʿmān, but these are not supported by any of the other contemporary accounts.
In the late summer of 1099, a ram may have been constructed following the crusaders’ arrival at Jerusalem. According to some accounts, this was used to destroy a portion of the outer wall along the north side of the city in the final phase of the siege, allowing the siege tower behind to approach the main wall. Neither Raymond of Aguilers nor the anonymous author of the Gesta Francorum mention the ram; however, Albert of Aachen, who composed a detailed account of the First Crusade and the early Latin presence in the Levant without ever leaving Europe, provides a vivid description of its use. Albert’s account is followed closely by Ralph of Caen, who would arrive in the East years later. Fulcher of Chartres, who would come to reside in Jerusalem in the following years, but was at this point almost 700km away in Edessa, and Guibert of Nogent, who never left Europe, strangely refer to the use of multiple rams.
The apparent success of the crusaders’ ram speaks to the weakness of the outer wall, perhaps no more than an undefended forewall. The butting head would have created a noticeable weak spot: as shockwaves reverberated outwards and the masonry was forced to absorb the energy of each percussive blow, some of the bonds between mortar and stone further from the point of impact would have begun to break as well. Sappers might then have been able to work outwards from this point, driving their picks into the cracks to pull out stones and eventually open and enlarge a breach down to the ground and wide enough for the siege tower to pass through. Despite the weakness of the forewall, it was still necessary to use a very large ram, so great that the crusaders opted to burn it in place rather than attempt to drag it out of the way so the siege tower behind could be pushed up to the main wall.
Twelve years later, the Franks besieged Tyre over the winter of 1111/12. As at Jerusalem, the besiegers focused their strategy on the use of two siege towers. While the Frankish sources make no reference to the use of rams, Ibnal-Qalānisī gives a detailed account of the siege and asserts that rams were key components of the Frankish towers. Two and a half months into the siege, the towers were ready. The smaller of the two was burnt in a sally by the defenders but the larger proved a more menacing threat. A leading sailor from Tripoli, who had some experience in war and happened to be in Tyre during the siege, constructed grappling hooks that were used to grab the head or neck of the large tower’s ram as it was swung forward to butt the wall. People in the city then hauled on the ropes attached to the hooks, pulling the exposed head of the ram upwards until it threatened to topple the whole tower, forcing the Franks to break the ram in order to save the tower. Undeterred, the Franks replaced their broken ram, and those that followed, as this played out a number of times. When the hooks failed, the defenders would drop down two rocks tied together on the exposed neck of the ram, breaking the beam behind its head. But each time the Muslims destroyed a ram, the Franks brought in a new one. Each was 60 cubits (32m) long, with an iron head that weighed 20 pounds, and was hung inside the tower with ropes.
They called it a ‘ram’ because it is pushed backwards and forwards with repeated and frequent blows like a ram, and demolishes walls, no matter how solidly built. The ram was strongly covered all round with iron plates, and when it was finished the archbishop [of Besancon] intended to use it to destroy the wall . . .
A day was set when they would all attack the wall with the devices they made. The archbishop moved his ram forward to shatter the wall. It was roofed over like a house. Inside it had a long ship’s mast, with an iron-covered head. Many hands drove it against the wall, drew it back and aimed it again with even greater force. So with repeated blows they tried to undermine the face of the wall and break it down. As they shook the wall with repeated blows, the roof of the ram kept its operators safe from all danger of attack from above.
The Muslims on the walls defended themselves manfully. They collected a huge heap of old dry wood on top of the machine, which of course they could easily set on fire. At the same time, their stone-throwers were continually hurling enormous boulders at it. At last they dropped Greek fire on top. As it ignited the wood, those inside the machine found the growing heat of the fire unbearable. Realizing that the whole machine was going to be destroyed, they left it and pressed on with the attack using what other devices they could. The Turks kept on tirelessly hurling missiles at the ram, hoping either to crush it with enormous lumps of rock or to burn it with incendiary oil.
(Adapted from Itinerarium 1.59, trans. Nicholson, pp. 111–13.)
Ibn al-Qalānisī’s remarks sound almost instructional, echoing similar thoughts expressed by classical commentators, such as Vitruvius and Procopius, concerning ways of defending against rams. While it may be that it was not uncommon for Frankish siege towers to contain rams, making them unworthy of mention by Frankish authors, there are issues with Ibn al-Qalānisī’s account. He claims that the hooks caught the rams as they swung against the wall, but goes on to state that the siege tower was subsequently advanced, at which point it was destroyed when a boom arm, 40 cubits (22m) long, was extended out over the tower to drop incendiaries on it, burning the tower. This, and the large figure he gives for the length of the rams, may reflect a belief that they were being used from a significant distance, perhaps reaching 10–20m from the tower to the town wall. This would leave the necks of the rams unnecessarily exposed and unsupported, and may reveal the author’s lack of familiarity with the practical use of such engines. Critically, Ibn al-Qalānisī’s home city of Damascus had not been besieged since it was captured by Atsiz ibn Uvaq in the 1070s, at which point he was a young child. He went on to become a leading figure in the Damascene chancery (dīwān al-rasāʾil ), which may have kept him in the city and away from the battlefield until Damascus came under pressure from the Zankids from the 1130s.
Neither Frankish nor Muslim accounts support Ibn al-Qalānisī’s remarks regarding the use of rams at Tyre, but they are similar to those found in one of Saladin’s letters relating to the brief Sicilian siege of Alexandria in 1174. It states that among the siege engines employed by the besiegers were three siege towers, each equipped with a ram. Frankish accounts of the siege do not mention the rams, leaving it unclear whether Saladin included them to exaggerate the circumstances of the siege, something he was known to do in such letters, or the Franks did not consider the rams noteworthy.
Although references to rams are quite rare and often uncorroborated, it can be said with a fair degree of certainty that at least one ram was used against Acre in 1190 – Muslim and Frankish accounts provide quite similar descriptions of this engine and the attack in which it took part. Among the various crusading contingents that had joined the siege since it had begun in the late summer of 1189, was one led by the archbishop of Besancon, and it was he who commissioned the ram. There is a reference to Count Henry of Champagne building another at the same time, but nothing more is said about this mysterious and probably fictitious second ram. According to Ambroise, a participant of the siege, as well as the slightly later Itinerarium peregrinorum et gesta regis Ricardi, composed by another eyewitness who closely followed Ambroise’s poetic account, the roof of the ram was solid enough to protect it from rocks dropped from the wall above:
It was of such rich workmanship that it should not justly nor reasonably fear any creature. It was made as if a house. A great ship’s mast, straight and without knots, was in the middle, tipped with iron at both ends. Underneath [the roof of] the ram were those who would strike against the walls, having no fear there.23
Unfortunately for the besiegers, it was October by this point and probably many months since it had last rained. This allowed the defenders to burn the ram fairly easily by dropping dry wood on and in front of it, then setting this tinder alight with Greek fire, a naphtha-based flammable liquid resilient to water.
As alluded to above, some historians have suggested that rams might have been so common that contemporary Frankish and European authors regarded their use as obvious. It seems very unlikely, however, that the various sources, who hailed from very different cultural backgrounds, would all have taken this view. Likewise, it is hard to explain why artillery and miners, which were used at many if not most sieges, were considered deserving of regular mention but rams were not. Instead, the seldom appearance of these engines, which are scarce in the twelfth century but even harder to find in the thirteenth, is more likely due to their general inefficacy. It seems the construction of rams was simply not worth the trouble when compared to using sappers alone.
Mining – more accurately undermining or sapping – was a practice that involved the weakening of the base of a wall, causing the mass of the unsupported masonry above to collapse under its own weight. This was done by excavating a cavity below the foundations of a wall, or by removing its lowest courses, temporarily propping up the structure above with wooden supports. When a sufficient stretch of the targeted wall was completely reliant on these supports, combustibles were placed around them and then set on fire, causing the wall to fall among clouds of smoke and dust. Rather than tunnelling under or even through a wall, which provided only a small passage for a limited number of troops to attack at any one time, undermining created a much larger breach, allowing for a massed frontal attack. Sappers had been used in this way since antiquity, and while both Muslim and European forces had turned to mining at sieges during the Early Middle Ages, this practice is found most often in connection with Byzantine forces in the centuries leading up to the First Crusade.
When considering the period of the crusades, sappers were employed at the very first siege and the very last – those of Nicaea in 1097 and Acre in 1291 – and countless sieges in between. Although mining was practised by all parties, it is often viewed as a Muslim siege tactic. The infrequent use of siege towers and other such large engines by Muslim forces has led to assumptions that they were less technologically savvy or possessed a superior mining tradition; in reality, their apparent partiality for sapping reflected broader tactical preferences. Whereas Frankish forces regarded the siege tower as the best tool of attack in some scenarios, this was ill-suited to most Muslim forces, which were more mobile and may have preferred means of bringing more men into an attack at the same time than was allowed by a siege tower. Likewise, many Muslim armies were from regions where timber was scarce, so developed siege tactics that avoided the need for machines constructed from large trees.
From the opposite perspective, Frankish forces were just as likely to employ sappers at a given siege as were their Muslim counterparts. Between 1097 and 1186, sappers are found at twice as many sieges undertaken by Muslim armies in the Levant than Frankish ones; however, sieges prosecuted by Muslims outnumber those of the Franks by almost exactly 2:1. Similarly, from 1187 to 1291, sieges initiated by Muslim forces outnumber those undertaken by Franks by about 5:1, but so too do references to mining. Accordingly, when considered proportionally, Frankish besiegers were as likely to employ sappers as were their various Muslim neighbours.
When considering the siege of Nicaea, it seems that at least some groups of crusaders arrived in the Near East with sufficient skill to undermine the region’s formidable strongholds. Sapping efforts were initiated by the first contingent to arrive and were under way by the time Raymond of St Gilles and the second group showed up. According to Raymond of Aguilers, who accompanied this latter contingent of Provencal forces, as well as the Gesta Francorum, composed by another eyewitness, Raymond’s sappers went on to successfully bring down the wall of a tower. Unfortunately for the crusaders, this occurred late in the day and the defenders were able to fortify the breach sufficiently to repel an assault the following morning. Albert of Aachen adds two more episodes of mining. In the first, he describes how a ‘fox’, capable of sheltering twenty men, was commissioned by two lords, one from Lorraine and the other from Swabia, but the structure was improperly built and collapsed, killing those below it. Later in the siege, a Lombard master builder, who was experienced in constructing siege works, offered to build a better shelter that would allow the Franks to take the city. He was promised 15 pounds of the coinage of Chartres and duly oversaw the construction of a penthouse, which had a steeply sloping roof and was large enough to accommodate a number of armed defenders. According to Albert, the master builder and the sappers with him undermined the wall of a tower, leading the defenders to surrender before the breach could be stormed. William of Tyre repeats these anecdotes in his later history. When the accounts are read together, it may be significant that the most successful miners apparently hailed from southern France and Italy, regions with stronger traditions of building with stone than those further north.
Albert’s descriptions of mining efforts at Nicaea reveal his clear understanding of this practice; however, many other sources, from various backgrounds, were equally familiar with mining, confirming its popular use. Roger of Antioch’s chancellor, Walter, is known to have campaigned at times with the army of Antioch and provides a vivid description of īlghāzī’s sapping efforts against al-Athārib in 1119:
Since īlghāzī was unable to take the castle by storm, he sent men from different sides to dig out a cave made underground, and he prepared fuses by grafting together dry pieces of wood so that when they reached the towers and put in that same kindling they would collapse, being supported by posts.24
An anonymous Syriac account dating to the thirteenth century, which made use of an earlier source, describes how Roger of Antioch’s forces captured ʿAzāz the previous year:
He dug tunnels in the ground under the wall, put beams under it, and then set them on fire. The wall tottered and fell; the Franks leapt in through the breach, took the fort, and slaughtered all the Muslims in it.25
A very similar description is provided by Fulcher of Chartres when describing Balak’s siege of Kharpūt, which had been seized by the Frankish prisoners Balak had been holding there:
He immediately ordered the rock on which the castle was situated to be undermined and props to be placed along the tunnel to support the works above. Then he had wood carried in and fire introduced. When the props were burned the excavation suddenly fell in, and the tower which was nearest to the fire collapsed with a loud noise.26
Descriptions of tunnelling, rather than undermining, are comparatively rare and lack corroboration. At Shams al-Mulūk’s siege of Frankish Bānyās, which began in December 1132, Ibn al-Qalānisī suggests that the Muslim miners, working under cover of vaulted shields, bored through the outer wall. Another rare example can be found in the Eracles continuation of William of Tyre’s chronicle, where it is stated that the Egyptian forces sent against Ascalon in the spring of 1247 dug under the castle’s walls and emerged on the other side.
The problem with tunnelling, and why these accounts warrant a degree of scepticism, is that it takes away the advantage of the besiegers’ numerical superiority, as each man emerging from the tunnel would presumably face a far larger number of defenders. When used effectively, tunnelling was instead a means of introducing forces clandestinely, usually with hopes that they would open a gate to allow the army waiting outside to rush in. Ammianus describes how Roman miners, at the siege of Maogamalcha in 363, reached the foundations of the wall, but, rather than expanding the cavity and bringing down the wall, a night attack was launched on multiple fronts, creating a distraction that allowed picked men to emerge from the tunnel during the confusion and open the town gates. It is unclear what influenced the author of this portion of the Eracles account, but there is a noticeable trend in Ibn al-Qalānisī’s chronicle that indicates his understanding of sapping improved through the following decades of his life.
The most famous description of mining from this period is that provided by Usāma ibn Munqidh, a native of Shayzar. He describes the first time he ventured into a mine, while taking part in Bursuq’s siege of Frankish Kafarṭāb in 1115, at which point he was about twenty years old. Here and, about two decades later, at the siege of Masurra (a journey of six days from Mosul), Usāma notes that the miners were from Khurasan. Ibn al-Qalānisī similarly notes that the sappers who helped Shams al-Mulūk capture Bānyās around the end of 1132 were Khurasanian, while those who undermined the defences of Edessa for Zankī in 1144 were from Khurasan and Aleppo. Aleppan miners are also found in the service of Richard I of England at Dārūm in the spring of 1192. Muslim sources assert that these men, who found themselves in the service of the English king, had been defenders of Acre who fell into Frankish hands when the city was captured. Some historians have postulated that these individuals may instead have been among the thirty-five people whom Richard spared when he intercepted a Muslim supply vessel off the coast of Beirut on his way to Acre in 1191.
Usāma ibn Munqidh: mining at the siege of Kafarṭāb, 1115
I descended into the fosse, with arrows and stones falling on us like rain, and entered the tunnel. I saw there a very clever thing: they had tunnelled from the fosse to the outer defences, and on either side of the tunnel they had set up posts over which stretched a plank to prevent the earth above it from falling in. They extended the tunnel along in this way using timbers right up to the base of the outer defences. Then they tunnelled under the walls of the outer defences, keeping them supported, and reached as far as the foundations of the tower. The tunnel here was narrow, as it was only intended as a way to get to the tower. As soon as they reached the tower, they widened the tunnel along the wall of the tower, supported it on timbers, and, a bit at a time, they started carrying out the pieces of chipped-away stone . . .
They then set about cutting up dry wood and stuffing the tunnel with it. Early the next morning they set it ablaze. We had put on our armour and marched to the fosse, under a great shower of stones and arrows, to launch an assault on the citadel once the tower collapsed. As soon as the fire began to do its work, the layers of mortar between the stones of the wall began to fall out, then the wall cracked, the crack widened and the tower fell.
(Adapted from Usāma ibn Munqidh, trans. Cobb, pp. 85–6.)
References like these have led to the conclusion that there was a regional tradition of expertise dealing with stone in the areas of Khurasan and Aleppo. Although certainly possible, this was probably also true of many other regions that would have been represented in large forces like Bursuq’s caliphal army. Accordingly, describing miners as Khurasanian, those who hailed from a region more than 2,000km from Damascus, today consisting of northeastern Iran, northern Afghanistan, eastern Tajikistan and Uzbekistan, and Turkmenistan, may simply have been a way of identifying ‘Easterners’, outsiders to an individual living in western Syria. Ibn al-Qalānisī may have used ‘Aleppan’ as a similar slur, notably absent in the work of Usāma, who grew up less than 120km from Aleppo and more in its sphere of influence than that of rival Damascus.
Mining required skilled labour, but the regularity with which sappers were employed, by both Frankish and Muslim forces, reveals that experienced sappers were readily available. The speed with which these men could work is also astounding. Al-Bursuqī took Kafarṭāb on 9 May 1125, before moving on to take Zardanā and then besiege ʿAzāz. Although it is not stated when the siege of ʿAzāz began, al-Bursuqī’s sappers reportedly compromised the stronghold’s two lines of outer walls before a Frankish relief force arrived under Baldwin II on 22 May. In 1132, catching the Franks off guard, the forces of Shams al-Mulūk of Damascus are supposed to have undermined the defences of Bānyās in less than five days. When Saladin besieged Jaffa in 1192, his miners targeted a section of the town’s defences that had recently been slighted, but which had been refortified when the Franks reoccupied the town in 1191. They were able to bring down a section of wall on the third or fifth full day of the siege; however, the defenders were initially able to defend the breach from behind a line of improvised defences. During Baybars’ siege of Crac des Chevaliers in 1271, the southwestern corner of the castle’s outer defences appears to have been successfully undermined in a period of eight days. At Montfort, the mines opened by Baybars’ sappers later that year are still clearly visible to visitors. A large cavity was excavated from the southwestern outer tower in no more than three days, while the exceptionally wide mine dug into the southern wall of the upper castle may have been developed in the eleven days between the fall of the outer defences and surrender of the upper castle.
Despite the extremely hard stone that was used at some castles, miners appear to have worked incredibly efficiently if allowed to toil away unhindered. In this light, it is easy to understand the effort and importance Saladin placed on filling the northern ditch at Kerak in 1183 and 1184, which would have allowed him to set his miners to work against the castle’s northern wall. Where such obstacles were not a problem, mining efforts could proceed incredibly quickly. According to William of Tyre, while Baldwin IV was confronting Saladin at Kerak one year, a Damascene force, which raided the villages around Mount Tabor, was able to undermine a tower in just four hours. Similarly, when Baybars encamped outside Acre in April 1263, his sappers undermined a tower, possibly the Templar’s fortified mill of Doc, which was defended by four knights and thirty infantrymen, in about thirty-six hours.
Montfort: mine in the outer southwestern tower, and mine in the southern wall of the upper ward. (Michael Fulton)
The speed at which medieval miners could work is even more impressive when considering that their primary tool was a single-beaked pick, unlike a modern pickaxe, which has two heads or beaks, one tapering to a point and the other splaying into a narrow shovel. Besides providing more options to the user, the second head of a modern pickaxe helps to balance the tool when swung, keeping its centre of mass in line with the shaft and the operator’s hands. This helps prevent the tool from twisting with each strike – any twist or roll results in a less efficient and less predictable stroke. These same factors led to the development of the two-headed North American frontier axe, which far outperformed traditional European designs. Although more efficient, two-headed picks required around 30–40 per cent more metal. In an era when tools often stayed in a family through multiple generations, using almost twice as much iron than was minimally necessary was probably regarded as a frivolous expense.
The value of miners and the service they provided was certainly appreciated and at times a bonus was paid as an added incentive. In 1105, Ṭughtakīn reportedly offered 5 dinars for every stone his men pulled from the stronghold Baldwin I had recently erected in the Golan. Richard I of England offered first 2, then 3 and finally 4 bezants to whoever would extract a stone from the tower opposite his position during the siege of Acre in 1191. In 1266, Baybars is said to have offered 100 dinars to each of the men who removed one of the first ten stones from Safed, in addition to the 300 dinars each sapper was paid for his work – almost certainly inflated amounts.
Mining was dangerous. Like a lumberjack felling a tree, determining how a structure would fall once the mine was lit and how large a cavity was needed were considerations that no amount of experience could predict with absolute certainty. Joscelin I of Edessa was seriously wounded, almost buried alive, when a tower made of brick suddenly collapsed while his men were sapping it. In 1108, Jāwulī successfully took Bālis after a siege of five days, but the breach that led to its capture was caused when an undermined tower fell prematurely, killing many of the sappers working below it. In a similar incident in 1184, a number of Saladin’s miners working against the small stronghold at Jinīn were killed when it collapsed on them. A group of sappers employed by Philip II, who were working to undermine one of Acre’s towers in 1191, found themselves in danger when they lit their mine and the tower above began to lean, threatening to crush them. However, there was more than just the structure above to worry about.
Artillery was commonly used to support mining efforts and, if positioned incorrectly, sappers might be subject to falling projectiles that had hit the defences directly above them. A greater threat was posed by the defenders along the parapet, who were in ideal positions to drop things on top of them. When working at the base of a wall, miners often made use of a penthouse for protection. These timber shelters typically had sloping sides, which deflected stones and other things dropped by the defenders, and perhaps even the odd friendly artillery projectile. To protect the men they sheltered from incendiaries, penthouses were often coated with hides, at times soaked in vinegar, and in rare instances were covered with metal plating. These shelters received various labels, often referred to by animal names used to identify classical engines, including ‘tortoise’, ‘sow’ and ‘cat’. There seems to have been little consistency in the use of these terms and some were at times used to identify quite different engines, even siege towers in rare examples.
While other siege technologies developed, especially artillery, the basic practice of mining remained the same. Nevertheless, it remained the most physically destructive siege weapon used by any army during this period. It was the pick and the shovel, rather than the ram or trebuchet, that ultimately compromised the fortifications of the most impressive castles and formidable towns, as occurred at Acre in both 1191 and 1291.
The primary function of a siege tower was to allow a group of besiegers to dominate a section of defences. By elevating a number of attackers above the level of the parapet, defenders lost some of the protection provided by their battlements, exposing them to the arrows and stones cast from the tower. These might also have provided a distraction, drawing the defenders’ attention away from other forces who might be trying to undermine or scale nearby walls.
Some siege towers were stationary, typically erected at a distance to provide a vantage point and shooting platform; others were mobile, allowing them to be constructed at a safe distance and then advanced to better dominate the besieged and impair their ability to defend a certain stretch of their defences. The latter were more common and could have a greater impact; however, they were more complicated to build, requiring some kind of carriage, and the ground ahead of them needed to be prepared before they could be moved. A small cluster of rocks, an area of soft ground or a dip in the topography could halt the advance of such a tower or threaten to tip it over. Where the tower was equipped with a bridge, allowing besiegers to access the top of the wall, any ditches or moats would need to be filled and the fill adequately tamped so that the tower could move over it without sinking or tipping. Although these bridging towers have come to be the most iconic, most siege towers built by Frankish and Muslim forces were not used to provide assailing troops access to the besieged parapet.
Medieval siege towers typically had at least three levels. The lowest level contained a staging area and was probably often designed to allow space for men to help push the tower forwards – towers were rarely drawn by draught animals in the Near East. The middle level(s) provided internal support and a sheltered space for water and men preparing to engage in the fight. The top level, reached by internal ladders, was where most of the action took place. It needed to be higher than the top of the besieged wall, allowing the archers and those throwing stones from this level to subject the defenders to plunging fire. Where the tower was designed with a bridge, this was typically at the highest of the middle levels, which would hopefully align with the parapet, allowing assault forces to enjoy the support of the troops at the highest level, while keeping them out of their way.
Siege tower (from Viollet-le-Duc).
Siege towers had been used across Europe and the Near East by the Romans, but they were more commonly employed in Europe than the Middle East during the Early Middle Ages. In the century preceding the First Crusade, Norman forces, some of whom would later take part in the crusade, seem to have used siege towers with particular regularity. In 1091, Robert of Belleûme built one for Robert of Normandy at Courcy, and another for the duke at the siege of Bréval the following year. In southern Italy, Norman forces employed siege towers against Trani (1042), Bari (1068–71), and just across the Adriatic at Durazzo (Dyrrachium, mod. Durreäs) (1081).
Although appearing less frequently in the Near East, siege towers were used at the sieges of Amorium (ʿAmmūriyya) (830) and Edessa (1070/71), among others. The towers at Amorium may have been quite small, being described as large enough to shelter ten men, while that built for Alp Arslān at Edessa may have been significantly larger, as it was erected on top of ten carts. In both instances, the besiegers were unable to push the towers close to the defenders’ walls: at Amorium, one of the towers became stuck in the material that had been used to fill the fosse; at Edessa, Alp Arslān’s tower collapsed.
During the First Crusade, siege towers were employed with effect at Maʿarrat al-Nuʿmān in 1098 and Jerusalem the following year. The Gesta Francorum and Fulcher of Chartres also mention the construction of ‘wooden towers’ at the earlier siege of Nicaea. It is possible that the crusaders had begun to build one or more siege towers, which remained incomplete when Nicaea surrendered to the Byzantines, who had made contact with the defenders by sailing across the Ascanian Lake. Alternatively, the original reference may have been included by the author of the Gesta to add to the grandeur of the siege, a detail Fulcher later included when compiling his account in the years following the crusade.
In late 1098, Raymond of St Gilles led his forces against Maʿarrat al-Nuʿmān. He ordered the construction of a siege tower and had the section of the town ditch ahead of it filled. Although it was equipped with four wheels and was pushed up almost to the wall, the tower was not used to convey men to the parapet. Instead, it dominated a section of the town’s defences, allowing nearby attackers to use ladders to climb up to the top of the wall. The towers at Jerusalem were also designed to command the walls ahead of them, but it is debatable whether they were also designed to transfer men to the opposing parapet – the eyewitnesses are silent on this point.
Two siege towers were constructed during the crusaders’ siege of Jerusalem. Raymond of St Gilles commissioned one to the south of the city, where his Provencal forces had established themselves on Mount Zion, and construction of another was ordered by Godfrey of Bouillon to the north of the city, where the remainder of the army was positioned. Albert of Aachen describes the northern tower as having three levels, with the men who pushed it at the bottom, a group of fighters in the middle and Godfrey and his brother Eustace with their men at the top. After a period of intense fighting, planks were pushed out to span the gap between the second level of the tower and the wall, creating a bridge to the parapet. Fulcher of Chartres provides an alternative version, describing how a pair of beams that the defenders had tied to the battlements as additional protection were cut free by men in the tower and used as a makeshift bridge. Planned or improvised, it seems the tower’s role as a bridge to the parapet was secondary. Whatever the case may have been, as soon as the first Franks stepped foot on the wall, others used ladders to climb up elsewhere as confusion and panic began to spread among the defenders.
Siege towers were the largest terrestrial engines of war built during this period and they required the finest available timber. Although suitable trees were rare in Palestine and many other regions, small forests of tall and straight trees grew around Beirut and in other pockets through the Lebanon and mountainous regions of what became the principality of Antioch. The significance of building materials was apparent at the siege of Jerusalem in 1099. As fate would have it, a group of Genoese vessels became trapped in Jaffa by a Fāṭimid squadron during the siege. A party of crusaders was sent to collect the sailors and they brought back timber from the Genoese vessels, sinking what remained of the ships. These materials were used to construct Raymond of St Gilles’ siege tower to the south of the city. Despite its strength, Raymond’s tower was disabled by the defenders’ artillery, having been built and brought forward on the restricted plateau outside the Zion Gate.
North of the city, Godfrey’s tower was built in a safer position further from the town wall; however, its builders were forced to use materials scavenged from the surrounding countryside. Ironically, this inferior timber permitted Godfrey to build his tower in sections and then move it the better part of a kilometre one night, where it was reassembled over the course of the following days. This allowed it to oppose a less defended part of the northern wall and forced the defenders to reposition their artillery in the tight urban constraints of the city. Although the tower was shot at for only one day, the final day of the siege (15 July 1099), it was already limping to one side before it reached the wall.
Jerusalem, siege of 1099, final deployments (with topography).
Following the capture of Jerusalem and the crusaders’ victory at the battle of Ascalon, Godfrey besieged Arsūf, spending six weeks constructing artillery and two siege towers. Unfortunately for the Franks, one of the siege towers collapsed under the weight of the men trying to use it to reach the besieged parapet. Although the Franks had considerable time to build the tower, their lack of naval support was probably to blame for the tower’s shoddy construction. Many Italian sailors would have been accustomed to rigging large beams and their experience might have helped the Franks; however, their real contribution would have been the timber that could be harvested from their ships. The sacrifice of one or more of their vessels was typically a sound investment for the Italian merchants, as they could expect a considerable share of the town and quite lucrative commercial rights in exchange for their support if the siege were successful.
The stripping of ships for materials to construct engines was probably a common practice. The masts of these vessels were particularly useful, ideal for the vertical corner supports around which the rest of the framework of a siege tower was built. Prepared planks from the decks and even hulls could be used to enclose the towers and subdivide their levels. Direct references to the use of masts and oars to build siege towers are found at the sieges of Caesarea (1101), Tyre (1111–12), Ascalon (1153) and Alexandria (1167). It was common practice to beach at least part of a fleet during a siege, providing a natural opportunity to strip materials if they were needed.
At the siege of Ascalon, a naval force of fifteen vessels was commanded by Gerald of Sidon, who was charged with intercepting the Egyptian fleet that was expected to arrive to relieve the city. Baldwin III was compelled to wait until Easter to build a siege tower, at which point he could purchase and impress pilgrim vessels, using their masts and other components for the tower; artillery and penthouses were built with the leftovers. In Egypt, a regime change in April 1153 ended plans to send a relieving field force, although a fleet still managed to break the Frankish naval blockade, extending the siege until August.
The Italians who contributed to sieges along the coast quickly came to appreciate the value of certain building materials. In 1123, a Venetian fleet of crusaders arrived in the Levant with the intention of assisting with the siege of a coastal town. According to Fulcher of Chartres, then living in Jerusalem, it brought ‘very long timbers, which when skilfully made into siege machinery by carpenters, could be used for scaling and seizing the high walls of cities’ – in other words, wood for building siege towers.27 This is the only cargo Fulcher mentions. After spending Christmas and then Easter in the Holy Land, the Venetians set out with the army of Jerusalem to besiege Tyre, where two siege towers were built using the materials they had brought. Although the king, Baldwin II, was at that time a captive of Balak in Ḥarrān, the besiegers successfully compelled the surrender of Tyre after a lengthy siege. Timber was brought from an equally exotic, if quite different, source a decade and a half later during the siege of Bānyās.
In 1140, the army of Jerusalem assembled in response to Zankī’s aggression against Damascus. In return for mobilizing and compelling Zankī to withdraw from the region, Mu’īn al-Dīn had agreed to help take Bānyās for the Franks – the governor of which had recently rebelled and declared his support for Zankī. The combined army besieged the town for a while before it was decided to send for large beams from a stockpile in Damascus; these were then used by the Franks to construct a siege tower. After the ground ahead of the engine had been levelled, the tower was pushed forwards, allowing the Franks to inflict considerable harm from their vantage point and dominate the defences opposite them. This compelled the garrison to negotiate terms of surrender with Muʿīn al-Dīn, who dutifully turned the town over to the Franks.
In Egypt, the siege tower built to facilitate efforts against Damietta in 1169 was probably constructed using timber brought on board or stripped from Byzantine ships, or Frankish vessels that had joined the fleet. Five years later, the Sicilian force that attacked Alexandria almost certainly brought specially prepared timber, if not partially assembled engine components. According to a letter that Saladin sent out after the siege, the Sicilians erected three siege towers, complete with rams, which were brought into action the day after they landed. Saladin tended to exaggerate events in such letters – it would have been quite an achievement if the Sicilians had been able to deploy any towers, let alone artillery to support them, given the siege lasted less than a week.
William of Tyre: siege tower at the siege of Bānyās, 1140
It finally became evident to the Christians that no advantage could be gained unless they could build a wooden tower, move it close to the walls, and wage war upon the besieged from above. But in all that region no suitable material for such a purpose was to be found. Anar [Muʿīn al-Dīn Unur] therefore dispatched men to Damascus for tall beams of great size which long ago had been set aside especially for such a purpose. He bade them use all possible speed to accomplish their errand and return . . .
The messengers sent to Damascus returned without delay. They brought with them immense beams of the necessary size and strength. They were quickly dressed by the carpenters and workmen and put together solidly with iron nails. Soon an engine of great height towered aloft, from whose top the entire city could be surveyed. From this vantage point, arrows and missiles of every sort could be sent, while great stones hurled by hand could also help to keep the defenders back. As soon as the engine was ready, the ground between it and the walls was levelled off, and the machine was attached to the ramparts. There, as it looked down upon the whole city, it seemed as if a tower had been suddenly erected in the very midst of the place.
Now for the first time the situation of the besieged became intolerable; they were driven to the last extremity, for it was impossible to devise any remedy against the downpour of stones and missiles which fell without intermission from the movable tower . . .
In addition, they were now debarred from passing back and forth about the ramparts and could not without peril of death carry aid to their comrades who were falling. For the weapons and modes of assault used by those fighting below could be considered little or nothing in comparison with the manifold dangers to which they were exposed from the fighters in the tower. In fact, it seemed to be rather a war with gods than with men.
(Adapted from William of Tyre 15.9–10, trans. Babcock and Krey, 2:108–10.)
Siege towers are sometimes used to characterize Frankish siege operations. Their use, however, was restricted almost exclusively to engagements along the coast: of the twenty sieges where siege towers were used between the First Crusade and the end of the twelfth century, seventeen were along the shores of the Mediterranean. Unsurprisingly, the use of siege towers declined as the coast fell into Frankish hands: during the century following the arrival of the Franks, two-thirds of sieges featuring siege towers took place in the first fifteen years.
Following the Franks’ capture of Ascalon, and with it the entire Palestinian coast, siege towers were not again employed outside Egypt until the Third Crusade. During the lengthy siege of Acre (1189–91), three wheeled siege towers were employed by crusading forces in 1190: one built by Louis III, Landgrave of Thuringia; another by Conrad of Montferrat and the Genoese; and the last by Guy of Lusignan. Revealingly, the wood is described as seasoned or dried, suggesting it had been brought by sea or salvaged from ships, which Muslim sources appear to confirm. Efforts were made to prepare the ground ahead of the towers and measures were taken to fireproof them, which included coating them with hides soaked in vinegar. Despite this, all three towers were eventually burnt. Later in the year, a different kind of elevated platform was used.
Use of Frankish siege towers by decade.
With the siege making little headway through the summer of 1190, the Pisan-led fleet that controlled access to Acre by sea made an attempt against the Tower of the Flies, which dominated the entrance to the harbour. At the centre of their plan was a galley with a fortified masthead higher than the tower. It was hoped that Frankish archers in the mast-castle would be able to dominate the tower, perhaps also throwing incendiaries onto the tower’s roof, while others would climb up via two large siege ladders raised from accompanying galleys. In another version of events, the mast-castle was equipped with a bridge. By all accounts, the attack was countered by galleys from the city and was finally defeated when the mast-castle was burnt. This was not the first time that a ship had been equipped with what might be described as a fortified crow’s nest: in 1108, at the siege of Sidon, the mastheads of some ships were similarly fortified.
What might be more accurately described as a maritime siege tower was built by participants of the Fifth Crusade during the siege of Damietta in 1218. From the end of May, the crusaders, accompanied by a large contingent of the Frankish baronage under King-regent John of Brienne, established themselves on the west side of the Damietta branch of the Nile, opposite the city and the Tower of the Chain. The tower rose from the water and controlled access up the river with a chain, which ran across the main shipping route between the tower and the city. The crusaders were reliant on supplies arriving by sea, so were obliged to take the tower, and the city, before continuing upriver towards Cairo. One of the best accounts of this siege was penned by Oliver of Paderborn, and it may have been he who devised the means by which the Franks took the tower, although the humble cleric omits any hint of personal involvement in his account of events.
We realized that the tower could neither be captured by the blows of petraries or of trebuchets (for this was attempted for many days), nor by bringing the Tower [of the Chain] closer, because of the depth of the river, nor by starvation, because of the surroundings of the city, nor by undermining, because of the roughness of the water flowing about. With the Lord showing us how and providing an architect [Oliver himself], and with the Germans and the Frisians providing supplies and labour, we joined two ships which we bound together sturdily by means of beams and ropes and so prevented (by their closely connected structure) the danger of drifting. We erected four masts and the same number of yards, setting up on the summit a strong fortress joined with poles and a network of fortification. We covered it with skins about its circumference, as protection from the attacks of their machine, and over its top as a defence against the Greek fire. Under the fortress was made a ladder, hung by very strong ropes and stretching out thirty cubits beyond the prow. This task having been successfully completed in a short time, the leaders of the army were invited to see it, so that if anything was lacking that ought to be supplied by material or by human ingenuity, they would point it out.
(Adapted from Oliver of Paderborn 12, trans. Gavigan, pp. 64–5.)
It became apparent that no measure of artillery bombardment would compromise the tower, and a marine attack, involving two ships with siege ladders and another with a fortified masthead, was repulsed. The besiegers then built a more substantial engine. Two ships were joined together and a siege tower built on top of them, using the two masts of each vessel as the corner posts, while their yards provided horizontal structure. On 24 August, the floating tower was moved into place and a bridge was let down, allowing the besiegers to engage the Muslim defenders of the Tower of the Chain. After a bitter fight, the tower’s garrison surrendered. The Franks finally crossed the river in early February, enabling them to besiege Damietta’s landward defences. No siege towers were employed during this second phase of the siege, which continued until 5 November 1219.
Siege towers were seldom employed during the thirteenth century, but a few were used during the intra-Frankish conflict of the early 1230s. The Lombard (or imperial) party erected a siege tower against Beirut, and their Ibelin adversaries employed two towers when besieging the Cypriot stronghold of Kyrenia. Both were coastal towns and siege efforts probably benefited from naval support. Further inland, Louis IX of France appears to have constructed two siege towers of a sort in Egypt during the Seventh Crusade.
Although the crusaders who accompanied Louis IX to Egypt were able to take Damietta virtually unopposed in 1249, they stalled and were eventually defeated at Manṣūra, where the Fifth Crusade had collapsed three decades earlier. Manṣūra is about 60km southwest of Damietta, just south of where the Tinnis branch of the Nile diverts to the east, and it was here that the Muslim army waited for the Franks, shielding Cairo, 110km to their rear. In order to confront the Muslims and continue their way south, Louis IX ordered the construction of a causeway across the Tinnis branch of the river. The crusaders approached the task with optimism, having successfully dammed another distributary during their southward march. The earlier dam, however, had been constructed at the point of divergence, allowing the water to flow to the left without any build-up of pressure, but the causeway site was more than a kilometre from where the Tinnis stream left the main Damietta branch of the Nile. Besides the water pressure, there was also the enemy to consider; skirmishers had plagued work on the earlier dam but the crusaders were now faced by the main Egyptian army.
To protect the workers constructing the causeway, Louis commissioned two siege towers. According to John of Joinville, the seneschal of Champagne and a participant of the crusade, these were ‘called chas chastiaus, because they had two castles [or towers] in front of the cats [or penthouses] and two houses [or covers] behind the castles to protect those on guard from the blows of the engines of the Saracens, who had sixteen engines just across [the river].’28 Using the towers to command the opposite bank and the shelter provided by the penthouses in front of them, the Franks began constructing their causeway. The Muslims countered this with their artillery and by widening the river opposite the Frankish mole. Their traction trebuchets drove away those guarding the workers, while a counterweight trebuchet threw incendiaries at the towers and penthouses, eventually setting them on fire. Desperate, Louis begged his barons for timber from their ships in order to build a new penthouse, but this too was destroyed by the Muslims’ artillery.
Although Louis’ towers did not contribute to a siege in the traditional sense, they fulfilled what was essentially the same function: they allowed those at the top level to overlook and dominate their opponents. During one of his campaigns in Egypt, Amalric had built similar structures to secure a bridge he was building over the Nile. Here too, a Muslim force was able to interrupt construction by commanding the far bank. Whereas Louis had the luxury of stripping seasoned and planed timber from ships, Amalric was probably forced to rely primarily on palms.
The relative regularity with which the Franks employed siege towers at significant sieges is in contrast to the infrequency with which they were built by Muslim forces. This was not because Muslim armies lacked the knowledge or skill to do so, or even the materials, as was demonstrated by their ability to furnish such at the siege of Bānyās in 1140. Rather, these cumbersome towers were not as congruent with their style of warfare. Muslim tactics typically involved bringing more men into action than could be conveyed, or supported, by a siege tower.
Ibn al-Athīr notes the construction of siege towers during Sultan Barkyāruq’s siege of Isfahan in 1102, and the siege of Mosul, undertaken by his brother, Sultan Muḥammad, which ended when news of Barkyāruq’s death arrived in January 1105. In what appears to be a unique example in the Levant, Saladin is said to have employed moveable siege towers against Tyre in 1187. The original reference appears to be that in ʿImād al-Dīn’s convoluted versed rendition, subsequently abridged by Ibn al-Athīr and Bahāʾ al-Dīn. Conspicuously, the towers do not feature in any of the Frankish accounts, suggesting ʿImād al-Dīn included them for dramatic effect or that he meant that these were less impressive shelters, perhaps similar to the one Baybars employed at Caesarea in 1265.
Although Muslim forces rarely built siege towers, they still had to defend against them. To do so often required creativity and ingenuity, especially since no town or castle was attacked with siege towers more than twice in a generation, leaving few opportunities for defenders to learn from previous experience. By comparison, some Franks who fought with the army of Jerusalem in the early twelfth century would have been intimately familiar with their use, having taken part in numerous sieges where such towers had been built.
One of the ways to defend against siege towers was to build counter towers. These were typically temporary wooden structures, erected behind a wall or on top of a tower to reclaim the advantage of elevation. A counter tower was built by the defenders of Durazzo in 1081, which opposed the siege tower advanced by the Norman force led by Robert Giscard and his son, Bohemond. The defenders of Damietta appear to have constructed a similar tower in response to the siege tower built by the Franks in 1169. During the Fourth Crusade, the defenders of Constantinople raised the height of certain towers with a further two or three wooden storeys to confront the ship-born siege ladders of the Venetians in the spring of 1204.
Counter towers provided advantageous positions for archers and others throwing stones and incendiaries, but sometimes more elaborate contraptions were devised, as at the siege of Tyre in 1111–12. According to Fulcher of Chartres, two counter towers were raised inside the town, higher than the Franks’ siege towers; however, other accounts suggest that it was an apparatus of a different sort that was responsible for destroying the assailing towers. Albert of Aachen describes the defenders’ engine as an exceptionally high tree, raised with ropes, which suspended and then dropped a large wooden ring, covered with an inflammable mixture of materials, onto each of the two siege towers in turn. This roughly fits with Ibn al-Qalānisī’s account, which recounts how, after the smaller of the two Frankish towers had been burnt in a sally, a leading sailor from Tripoli erected a baulk on the tower opposite the Frankish siege tower. On top of this solid beam of wood was a second, set horizontally, with a rod of iron at one end and ropes attached at the other, so that it could be rotated with pulleys and a winch like the yard on a mast. The apparatus was used to drop containers of filth and then combustibles on the Franks and their tower, while pots of burning oil were thrown by hand from the town walls. Simpler means were used against the northern siege tower at Jerusalem in 1099. According to Albert of Aachen, a tree trunk, into which nails and hooks were driven, was soaked in incendiaries and thrown over the wall, threatening to burn the base of Godfrey’s siege tower until it was extinguished with vinegar.
The humble ladder was the oldest and simplest siege weapon, and responsible for the fall of more strongholds than any other. So simple yet functional was the ladder that every two knights were responsible for providing one at the siege of Jerusalem in 1099. Although the first men to break into the city did so from the northern siege tower, the majority who scaled the city’s walls, like those at most other sieges, probably did so by ladder. In 1101, impatient Frankish forces besieging Caesarea launched a frontal attack with ladders before their siege tower was complete, taking the city by force.
Most ladders would have been fairly stout wooden constructions, although some were little more than ropes. In 1265, the Mamlūk governor of al-Bīra reported that the Mongols had cast ninety iron pegs, which were attached to ropes to raise the Mongols’ ladders. When the siege was broken, the pegs were brought to Baybars as trophies. Only weeks later, Baybars’ forces scaled the walls of Caesarea on the opening day of the siege, using iron horse pegs and harnesses. The following year, the Mamlūks climbed the walls of Safed using horse pegs once more, nailing them into the lower parts of the walls. Elsewhere, more preparation was put into the construction of ladders.
Despite the importance of ladders, medieval sources rarely devote much attention to them, focusing instead on the frontal assaults of which they were a pivotal part. These attacks might consist of an all-out push, a general assault launched by all available forces; however, more measured attacks, which were often carried out in relays, were more common. The typical size and composite nature of large Muslim armies, consisting of a number of emirs, each leading a contingent of fighters supported by his iqtā’, meant that they were naturally disposed to making attacks in turns. This gave a measure of autonomy and authority to each unit as it made its attack, while allowing others to rest.
In 1137, Zankī encamped his army in the plain below Montferrand, sending his emirs up the rocky slope to attack the castle in turns. Saladin similarly sent portions of his army forwards in rotation during the sieges of Beirut (1182), Tyre (1187) and Bourzey (1188), and presumably many others. When Baybars attacked Acre in 1263, having pushed the Franks back into the city, the Mamlūk emirs attacked the city’s gates in turns.
Frankish armies were not that dissimilar; although typically smaller, they were also built around a collection of lords, each bringing a group of fighters that often fought under him. At the siege of Ḥārim in 1177, the Franks settled in to besiege the castle through the winter. According to William of Tyre, once a blockade had been established, the Franks attacked in relays, ‘according to custom’.29 Likewise, during the first phases of the siege of Acre in 1189, the Frankish army was too small to completely blockade the city, so components of the army watched the city gates in rotation. During the Seventh Crusade, Louis IX entrusted the protection of his engines at Manṣūra to his brothers in rotation, while groups of lower-ranking knights received the night shifts in between.
Adana was full of Jacobite Christians with their metropolitan John of Edessa. When the emperor [Alexius Comnenus] captured it, he left a force to guard it and moved on to Antioch. They rejoiced to be under Greek rule and freed from the severe taxes of the Franks. At dawn on Sunday, while they were quiet and unsuspecting, a Turkish army came upon them, surrounding them like a moat. They began at once a fierce attack like a wind of swords, planting ladders against the walls and swarming up them. When they pushed them down on one side they sprang up on another. The defenders were weakened by the blows of arrows and stones and the great assault that encircled them. They endured in this distress from dawn until midday. God turned his face from them and they were delivered into the hands of the enemy in a way marvellous to tell, incredible to the hearer. A Turk climbed a ladder against the wall and, when he reached the top, the wall was still above him. He gripped a stone projecting from the wall and stood on it. One of the defenders who stood on the wall above him thrust at him with his spear to throw him down. The Turk laid hold of the spear and the man on top pulled it hard to release it from his grip and in this way the Turk was pulled up on to the battlements. He brandished his sword at the man below who gave way before him and went down; fear and trembling took hold of those near and they fled. The Turks were encouraged, climbed up after the pioneer and seized the wall. In a moment it was full of Turks. They went down into the town, opened the gates, and the army entered . . . They drove out all the people, made the men kneel, and killed them with the sword; they sacked the houses, convents and churches; gathered spoil without end; and took captive boys and girls, whole groups. They took also the metropolitan, priests and young deacons, binding them with ropes and taking them into sad captivity. They destroyed the town, laid it waste, and went to their own land. When the news reached the emperor he sent an army to pursue the Turks, but it did not overtake them as they had seven days’ head start. The captives were sold in various places, especially in Malatya.
(Adapted from Anonymous Syriac Chronicle, trans. Tritton, pp. 276–7.)
These natural divisions within most armies probably facilitated the allocation of other tasks. During the Third Crusade, Acre’s defenders were divided into four groups to clear the town ditch, which the Franks were endeavouring to fill: the first went down into the ditch and cut up the material and bodies that choked the fosse; the second carried away the more manageable pieces; the third guarded these workers; the fourth worked the artillery and defended the walls. When besieging the city a century later, al-Ashraf Khalīl similarly divided the tasks of maintaining the attack and guarding the siege works into four shifts. Artillery also seems to have been regularly worked in relays, allowing for a constant barrage if desired. This most famously took place during the crusaders’ siege of Lisbon in 1147. John Comnenus may have similarly used shifts of men to operate his artillery at Shayzar in 1138, as Zankī appears to have done at Baalbek in 1139 and Saladin did at Tyre in 1187.