Counterweight trebuchet, Cardiff castle. (Michael Fulton)
Of all the siege weapons employed during the Middle Ages, artillery is perhaps the least understood and most misjudged. The rapid mechanization, digitization and continual advancement of modern technology has led us to praise comparable historical innovations, at times overemphasizing their true impact. Focusing on a selection of sensational anecdotes can lead to the impression that artillery must have been responsible for shortening sieges and inspiring the development of much more sophisticated fortifications; however, when all of the historical and archaeological evidence is considered, and synthesized with a sound appreciation of the governing physics, a different picture emerges. Although artillery was an important siege weapon and trebuchet technology developed considerably through the twelfth and thirteenth centuries, these were far from wonder weapons.
Mechanical stone-throwing engines – artillery – were used throughout the period by all parties. These were not the torsion engines of the classical period, which fell from use during the Early Middle Ages, but rather swing-beam or levered machines, better known by the later term ‘trebuchets’. Making use of a beam that rotated around an off-centre axle, these engines operated according to the principle of mechanical advantage: when force was applied at the end of the short arm, the end of the long arm rose at a greater rate, proportional to its greater distance from the axle. By attaching a sling to the end of the long arm, the projectile within could be accelerated at an even faster rate, and thrown at a higher velocity.
The vocabulary used to identify these engines varied. Authors writing in Latin often employed classical terms, such as tormentum (pl. tormenta) and manganum (pl. mangana), while petraria (pl. petrariae), a term first appearing in the Early Middle Ages, appears with increasing frequency through the twelfth century. Often, however, more ambiguous machinae are mentioned. Authors may have used certain terms to identify particular types of engine, but there appears to have been little standardization or consistency among sources before the late twelfth or early thirteenth century, at which point various forms of the term ‘trebuchet’ begin to appear, rendered into Latin as trebuchetum. In Arabic sources, the term manjanīq (pl. majānīq) appears most frequently to identify artillery. Like manganum, this was derived from the Greek manganon. The term‘arrada (pl. ʿarradat) was at times used to refer to lighter engines.
The first trebuchets were operated by traction power: a small group of people applied the necessary force by pulling on ropes attached to the end of the short arm. This technology seems to have originated in China and had made it to the eastern Mediterranean by the late sixth century. Both attackers and defenders employed these engines during the First Crusade, confirming that knowledge of this technology had spread throughout the Middle East as well as southern and western Europe by the end of the eleventh century.
Traction trebuchets were light engines, used in an antipersonnel capacity. Although they required more operators than the single person that could work a bow, they harnessed more energy, throwing heavier projectiles with greater force and inflicting more carnage. The lightest engines might be operated by only one or two pullers and a loader; more efficient engines had crews of around a dozen pullers. Due to the finite distance that a person can pull a rope in a single tug, the size of these engines was restricted, limiting the number of people that could effectively help operate them. References to engines that made use of crews significantly larger than this, such as the Frankish engine that supposedly required a crew of 600 at the siege of Damietta, are clear exaggerations.
Counterweight trebuchet, Cardiff castle. (Michael Fulton)
The strengths of the traction trebuchet were its simplicity and potential shooting rate. During the siege of Lisbon, undertaken in 1147 by a contingent of the Second Crusade on its way to the Holy Land, an eyewitness account states that two traction trebuchets, operated in shifts, were able to shoot 5,000 stones in ten hours. The author almost certainly did not sit and count the stones that were thrown, but the shooting rate of these engines appears to have impressed him. Coincidence or otherwise, trials with reconstructed engines have shown that maintaining a shooting rate of four shots per minute is not unreasonable – maintaining an adequate supply of ammunition is the greater challenge.
Traction trebuchet, Caerphilly castle. (Michael Fulton)
Despite what we now recognize as the meagre power of these engines, contemporary descriptions can give the impression that they were significantly more powerful, even capable of destroying walls. It is important to contextualize these remarks, bearing in mind that these were the most powerful ballistic engines until the late twelfth century. At some sieges protection was added to battlements in the form of sacks and mattresses stuffed with soft padding or lengths of wood. This was done at the siege of Jerusalem in 1099 as well as at a number of earlier sieges, such as the Avars’ siege of Thessalonica and Byzantines’ siege of an unnamed Persian stronghold, both in the late sixth century, and the ʿAbbāsids’ siege of Amorium in the ninth century.
The counterweight trebuchet was invented or introduced to the area around the late twelfth century. This was essentially an adaption of the existing technology, which saw the pullers replaced by a dead load. No longer limited by the distance a person could pull a rope, these engines could be built as large as the skill of their designers and strengths of their building materials allowed. The sources provide few indications of where this technology was first employed, suggesting there may have been a period of gradual development before the potential advantages were fully appreciated.
In a work on weaponry presented to Saladin around 1180, Marḍī ibn ʿAlī al-Ṭarsūsī included the earliest surviving illustration of a counterweight trebuchet. The image, and accompanying description, reveal an early stage of development. The axle was so low that it required a hole to be dug for the counterweight to fall into, while the horizontal starting position of the beam and vertical alignment of the sling, which was quite short, were better suited to a traction model and would have restricted the potential power and efficiency of the engine. Evidence of artillery damage dating to the 1170s and 1180s confirms the limited power of contemporary engines, whether traction–or counterweight-powered. Although some accounts of the siege of Acre (1189–91) give the impression that the trebuchets employed there were significantly more powerful than those found at earlier sieges, none of the eyewitness sources, men from as far away as England and Persia, suggests there were any differences between the crusaders’ artillery and that of the Muslim defenders.
Al-Ṭarsūsī’s Persian trebuchet. (The author)
By the Fifth Crusade, the counterweight trebuchet had developed sufficiently to warrant new terminology. It is at the siege of Damietta (1218–19) that the term ‘trebuchet’ first appears in the Near East. In the following decades, Arabic sources would come to develop terms to distinguish the heaviest, presumably counterweight- powered, type of trebuchet. The manjanīq maghribī (‘western’ or ‘north African’ trebuchet) is mentioned by contemporary sources at sieges such as Ḥarrān and Edessa (both 1136), often in the singular, while Ibn al-Dawādārī, writing around the early fourteenth century, retrospectively placed engines of this type at the siege of Damietta. A manjanīq maghribī was also one of fourteen trebuchets deployed against Homs through the winter of 1248/49; it is distinguished from the others by its power, said to have been able to throw stones weighing 140 artal (61kg if using the Egyptian ratl, or an unlikely 259kg if using the Damascene measure).
Less than two decades later, another terminological shift appears to have taken place. At the siege of Caesarea in 1265, the heaviest trebuchets were designated as maghribī and ifranjī (Frankish). Maghribī trebuchets are found at Safed the following year, but references to these engines disappear thereafter. They seem to have been replaced by ifranjī trebuchets, which were subsequently used at al-Bīra (1275), Margat (1285), Saone (1287), Tripoli (1289) and Acre (1291). It is possible that there was a sudden technological development, warranting new terminology; however, it seems more likely that the terminology changed independently – a new term came to be applied to the existing technology, which had developed significantly, but gradually, since the old term was first used.
The counterweight trebuchet was probably not the great wall-smashing engine that it is often believed to have been, at least not before the end of the thirteenth century. Despite the significant use of artillery at Acre during the Third Crusade, where it was employed for at least a year, most of the damage to the town’s defences was inflicted through mining. Three decades later more powerful engines were evidently used against Damietta, damaging at least one of the town’s landward towers. Yet after a bombardment of around six months, no breaches sizeable enough to be stormed had been opened. Even at Acre in 1291, where artillery damaged sections of the city’s soft sandstone defences, it was mining that opened the breach that led to the city’s capture. Evidence from the sieges of Crac and Montfort (both 1271) reveal that contemporary engines could throw stones of around 70kg, and perhaps even heavier, to distances of about 200m; however, these stones were typically thrown along great arcing trajectories and posed little threat to fortified walls 2–3m thick. Although these were the most powerful ballistic engines of the day, they were nowhere near as destructive as later bombards and siege guns.
The persistent reliance on mining ensured that traction trebuchets, which were ideally suited to support sapping operations, remained in use. Although these seem to have been operated in the shadows of their much larger cousins, their impact would have been felt by defenders. Of the thousands of stones thrown against Arsūf in 1265 (a siege where, unsurprisingly, mining efforts are emphasized by the sources) more than 90 per cent would have come from traction trebuchets.
In addition to the counterweight trebuchet, which he named the Persian manjanīq, al-Ṭarsūsī included descriptions of three types of traction trebuchet in his famous treatise. These he named the ‘Frankish’, ‘Arab’ and ‘Turkish’ varieties of manjanīq, all of which made use of triangular trestle frameworks, while the light lu‘ab, which seems to have been a pole-trebuchet that required only one puller, he regarded as sufficiently common that it deserved no description. The proper adjectives used by al-Ṭarsūsī are not found elsewhere, but new terms came into use in the thirteenth century.
The manjanīq shaytānī (‘satanic trebuchet’) is found at a number of sieges from the mid-thirteenth century and was clearly a light traction trebuchet. The manjanīq qarābughā was another type of traction trebuchet, often found alongside those of the shaytānī variety – qarābughā trebuchets were instrumental in protecting the advancing sappers at Acre in 1291. The manjanīq shaytānī may have been slightly lighter than the manjanīq qarābughā, but it is far from clear what distinguished these two types of engine. The Franks adopted the latter term, if not the technology, by the late thirteenth century, rendering it as carabohas in accounts of the Mamlūk sieges of Tripoli (1289) and Acre (1291).
Arsūf (with topography).
Arsūf: the outer southern tower, featuring artillery damage inflicted during the siege of 1265 before the tower was destroyed by the Mamlūk sappers; artillery projectiles from the siege. (Michael Fulton)
In addition to the physical effects of artillery, these engines also had an unmeasurable psychological impact during a siege. Both traction and counterweight trebuchets were most often used to target the parapets of walls, both the battlements and the defenders behind them, creating a treacherous environment for defenders. Unfortified structures within a stronghold might also be targeted, adding to the chaos and carnage behind the defensive perimeter. As the level of destruction mounted, organization and communication may have begun to break down, while morale might begin to decline as buildings threatened to collapse and the uncertain danger of incoming projectiles continued. Another terror tactic was the throwing of heads. Both attackers and defenders used their artillery to throw the heads of dead opponents during the First Crusade, but this practice appears to have declined thereafter.
Through most of the twelfth century, Frankish forces appear to have preferred to build their trebuchets on site before each siege. This involved relying on nearby natural sources of timber in western Syria, while wood sourced from ships would have been used to build many employed against the cities of coastal Palestine. As the Franks gained control of the Levantine coast, so too did Frankish naval power come to dominate that of the Fāṭimids and their later successors, denying these forces the same opportunities to source wood from ships. In a rare exception, the Muslim forces besieging Ascalon in 1247 were provided with an abundance of timber, which they used to construct siege machines, when a storm dashed their fleet on some rocks. With the development of the counterweight trebuchet, engines grew larger and there was a greater demand for tall and straight trees. This was highlighted in a letter Baybars sent to Bohemond VI after capturing Antioch, in which he bragged how the Mamlūk army had overrun the region but had preserved the trees suitable for constructing artillery – this resource was now his.
Muslim armies appear to have adopted the practice of storing and transporting artillery at an earlier date than their Frankish neighbours. In 1123, Fāṭimid forces brought artillery with them to Jaffa and by the 1140s there is evidence that artillery was being stored in Damascus. The process through which the Franks began to depend on prefabricated engines was a result of their links with Europe. This appears to have begun with the Italians who helped the Franks conquer the coast: in 1123 the Venetians brought timber to construct the engines that were used against Tyre the following year. In 1174, the Norman-Sicilian force that attacked Alexandria brought prefabricated trebuchets, along with other engines, and Richard I of England built artillery while wintering on Sicily before joining the siege of Acre in 1191. In addition to trebuchets, both of these latter forces also imported ammunition, bringing black stones from Sicily that were harder than the stones found along the southern coast of the Levant.
The practice of storing prefabricated artillery components was well established by the Mamlūk period, but Baybars exploited this to a new degree. A logistical master, he used the cities and large castles of the interior as arsenals. Before besieging Caesarea in 1265, he rode ahead to do some scouting, arranging for timber and stone ammunition to be gathered while he was away. Upon his return, trebuchets of the maghribī and ifranjī types had been constructed, as well as some smaller ones. When the attack came, the Mamlūks overwhelmed Caesarea’s town defences on the first day and the citadel was captured after only a week; it fell so quickly that Baybars was supervising the destruction of its defences by the time engines he had sent for began to arrive from Ṣubayba and Damascus. These prefabricated machines, as well as those that had been built ahead of the siege, were then put to use against Arsūf. The following year, while his army was encamped outside Acre, Baybars had trebuchets constructed in the field once more before moving against Safed, where additional engines from Damascus arrived after the siege had begun. Unlike Caesarea and Arsūf, which were both destroyed, Safed became a regional administrative centre and was integrated into the Mamlūk network of armouries – the engines Baybars had used to take the castle became the first of its new stockpile. Trebuchets from Safed were later among those taken to Beaufort in 1268 and Montfort in 1271.
Ibn ʿAbd al-Zā hir: Baybars prepares artillery, 1265
Baybars went out to reconnoitre Ibn Arsūf and Caesarea and when he returned to his royal tent he found that wood for the trebuchets had been brought with the ammunition and supplies by Emir ʿAlam al-Dīn, deputy of the amīr jāndār. Baybars summoned the amīr jāndār, Emir ʿIzz al-Dīn, and instructed him to set up several trebuchets made from this wood to maghribī and ifranjī designs. Early before dawn the next day, Baybars went out of his tent to sit in person with the craftsmen, so that they might work their hardest, and on the same day four large trebuchets and some small ones were made. He [also] wrote to the Muslim castles in order to collect [additional] trebuchets and stonecutters, and gave orders to the soldiers to make ladders, assigning to each emir the transportation of a certain number. He then departed for the neighbourhood of ʿUyūn al-Asāwir. At midnight, he ordered the soldiers to put on their arms; and before dawn he mounted and marched on Caesarea, whose people were taken by surprise . . .
(Adapted from Ibn ʿAbd al-Zahir, trans. al-Khowayter, 2:554-5, 562.)
The engines employed in 1271 against Crac des Chevaliers and ʿAkkār were also probably prefabricated, although this is not explicitly stated. In another of his letters bragging of his success to Bohemond VI, Baybars emphasized the difficulties his army overcame in moving these engines across the rough terrain between Crac and ʿAkkār:
Count Bohemond . . . has heard of our arrival at Ḥiṣn ʿAkkār, after leaving Ḥiṣn al-Akrād; how we brought trebuchets there, up the mountains where even birds find it difficult to nest; how when dragging them we endured trials from the mud and rain; how we erected them in places where even ants would slip when they walk there; how we descended into valleys where if the sun were to shine through the clouds it would show no way out except the precipitous mountains . . .30
The network that Baybars developed would be exploited by his successors, ultimately allowing al-Ashraf Khalīl to gather at Acre in 1291 the largest assortment of trebu- chets employed in the region during the thirteenth century. The components of one engine allegedly took up 100 carts; the task of moving this machine was so laborious amidst the rains and snow of the winter season that it took a month to move the engine from Crac, a Mamlūk arsenal since 1271, to Acre – a journey that normally took a rider eight days.
Artillery was employed almost as frequently by defenders as it was by besiegers, often being used to target troop concentrations and light siege defences, such as mantlets, as well as other siege machines. The threats posed by both offensive and defensive trebuchets sometimes led to artillery battles, as each side attempted to destroy the other’s engines. Siege towers, however, were particularly attractive targets for the operators of defensive trebuchets. During the First Crusade, the artillery deployed by the defenders of Jerusalem successfully disabled Raymond of St Gilles’ siege tower and considerable damage was dealt to Godfrey’s tower by the time it reached the wall. In 1110, the first siege tower erected by the Frankish besiegers of Beirut was destroyed by stones thrown from the defenders’ artillery. The two towers built after this fared better, allowing the besiegers to step from them to the city wall. Later the same year, the Franks appear to have taken particular care when constructing a siege tower at Sidon, cladding it with padding and damp ox-hides as protection against incendiaries.
Over time, siege towers appear to have become more resistant to the stones cast by traction trebuchets, which could be thrown only so hard; however, the pots of inflammable liquids and combustible projectiles tossed by these engines, which would stick to or hook on to siege towers, remained a significant threat. In May 1190, the defenders of Acre struggled in vain to destroy the three siege towers built by the crusaders. In desperation, the task was entrusted to a Damascene man, the son of a cauldron maker, who brewed a special mixture that was then thrown by the defenders’ trebuchets against the towers, destroying them each in turn before they reached the city’s walls. Larger counterweight trebuchets had a more devastating effect against siege towers, as Louis IX discovered at Manṣūra. The Muslims used traction engines to harass the workers constructing the French king’s causeway across the river, but it was a counterweight trebuchet that was brought up to destroy the two towers that the crusaders had built. Rather than stones, this engine used incendiary projectiles to destroy the towers. The stockpiling of these larger machines, making them available to defenders, may have contributed to the declining use of siege towers from the end of the twelfth century.
Unlike heavy counterweight trebuchets, defensive traction engines could be placed on top of town and castle towers, providing them with a good field of view and adding slightly to their range. Counterweight trebuchets were set up on the ground when used in defence; their weight and the dynamic forces unleashed through the shooting sequence posed a threat to most towers they might otherwise be placed upon. Setting them up on the ground also provided these valuable machines with the protection of the curtain wall and saved the trouble of manoeuvring their great components, ammunition and massive counterweights up to the top of a tower. In 1191, during the siege of Acre, a certain defensive trebuchet, dubbed Bad Relation by the Franks, appears to have been set up on the ground behind the town’s wall, from where it repeatedly damaged a significant trebuchet erected by Philip II of France, nicknamed Bad Neighbour. In 1291, some of the Franks’ largest counterweight trebuchets were deployed in a similar manner; that of the Pisans, which was placed in an open area near the church of San Romano, near the Accursed Tower, may have been quite near to where Bad Relation had been positioned a century earlier.
Albert of Aachen: defensive artillery at Jerusalem, 1099
At length the duke [Godfrey] and his men were oppressed and in difficulties from the constant blows of these five mangonels [the defenders had arrayed against them], and they used the strength of the Christians to place the siege tower close to the rampart and walls, so in this way it might withstand the mangonels more safely and, as the mangonels could not be brought into a large space because of the buildings, houses and towers, they were less able to throw and batter the engine. Once the tower had been brought up next to the walls, and the five mangonels could not find a sufficient distance to withdraw from it, a stone loaded and ejected by force from them flew over the too-close engine or sometimes, being too short in flight, fell next to the walls and crushed Saracens.
(Adapted from Albert of Aachen 6.17, trans. Edgington, p. 425.)
Ibn ʿAbd al-Ẓāhir provides a detailed account of how artillery was employed by both the Mamlūk besiegers and Frankish defenders of Margat in 1285. Ahead of the siege, Qalāwūn sent out mobilization orders from Egypt, ordering the artillery stored in Damascus to be readied, while arms and provisions were prepared and assigned to various emirs. Marching with the army by this point in time were a number of individuals identified as experts in the art of sieges and blockades. The army moved north, collecting artillery from other strongholds before making a final rapid advance on Margat, proceeding by forced marches to take the defenders off guard. In the opening phase of the siege, the castle was surrounded and the artillery, which was brought up on men’s shoulders, was deployed against the castle. Qalāwūn had at his disposal three ifranjī trebuchets, three of the qarābughā type and four of the shaytānī variety. These began to bombard the castle, suppressing the defenders, while sappers started to undermine the castle’s walls. The defenders’ artillery was destroyed by the ifranjī engines, allowing the Mamlūks to advance some of their trebuchets. This turned out to be a mistake: once the Franks repaired their trebuchets, these smashed those the Mamlūks had brought forward and killed some of their operators. The sappers eventually brought down a section of the outer wall and the defenders sued for terms shortly thereafter, fearing the miners had worked their way under additional sections of the castle’s defences. Qalāwūn agreed to allow the defenders to go free, provided their property was left behind. The castle then became a regional administrative centre and was incorporated into the Mamlūks’ network of armouries – the trebuchets Qalāwūn had just used were deposited in the stronghold. Along with the garrison of fighters entrusted with the castle’s defence, a group of men dedicated to the use of artillery was also stationed within.
The engines deployed at Margat were probably as large and powerful as any that would be used in the Levant during the remaining years of the thirteenth century. Although the archaeological evidence suggests that these machines were capable of throwing large stones considerable distances, the ultimate power of these engines is indicated by the pockmarks they left at Margat – little more than cosmetic blemishes on some of the castle’s southfacing basalt walls. Few artillery stones weighing more than about 100kg have been discovered, suggesting references to projectiles significantly larger than this are exaggerations.
Little is known about who designed and built these machines. In an age when basic carpentry skills were a necessary part of life for most men, simply seeing a trebuchet, whether traction- or counterweight-powered, would have allowed most to construct a functioning engine of modest scale. Ideal proportions and dimensions were probably passed along between builders, each refining his machines through a gradual process of trial and error. Although counterweight trebuchets could be directly scaled up – the design of a big engine was no more complicated than that of a small one – greater skill was necessary to move larger components and mitigate the dangers that accompanied any kind of failure, which grew proportionally.
Some Frankish lords seem to have had a certain skill in either designing or overseeing such projects. Robert of Belleûme appears to have had a knack for building siege towers, doing so at least twice for Robert of Normandy at the end of the eleventh century, while construction of the siege towers at Jerusalem, the most demanding engines built by the besiegers in 1099, was entrusted to William Embriaco, leader of the Genoese party rescued from Jaffa, and Gaston of Béarn, a lord from the Pyrenees. Later, figures are more closely linked with artillery. In an uncorroborated anecdote, William of Tyre asserts that an Armenian, named Havedic, was summoned to help manage the Franks’ artillery during the siege of Tyre in 1124. During the Seventh Crusade, Joscelin of Cornant, an engingneur, was responsible for building the artillery and various engines used to support the construction of Louis IX’s causeway at Manṣūra. John the Armenian, whom Joinville called ‘the king’s artillerist’, was sent to Damascus to buy horn and glue for crossbows after the crusaders returned to Palestine, suggesting he was responsible for such smaller ballistic weapons.
Who oversaw or designed most of the large siege engines constructed by the Franks and Ayyūbids is unclear. Under Baybars, however, ʿIzz al-Dīn Aybak al-Afram, amīr jāndār, appears to have been responsible for the sultan’s artillery. He was praised by Baybars for overseeing the preparation and use of the Mamlūks’ artillery at Caesarea in 1265 and then at Safed the following year. He seems to have been assisted to a degree by a certain ʿalam al-Dīn (‘flag of the Faith’, a not uncommon laqab – an honorific cognomen), identified as nāʾib amīr jāndār (deputy to the amīr jāndār), who oversaw the gathering of timber and ammunition ahead of the siege of Caesarea. Sayf al-Dīn Balabān al-Zaynī, amīr ʿalam, brought the trebuchet from Ṣubayba to Caesarea, arriving days after the siege had ended, and it was he who collected the trebuchets that were at Damascus and brought them to Safed the following year. In the lead up to the siege of Acre in 1291, al-Afram was once again called upon, sent ahead to Damascus to prepare the artillery there and orchestrate the assembly of engines deposited in various Syrian strongholds before the main army arrived in Palestine from Egypt.
Besieging forces that were able, or willing, to devote considerable time to a siege might opt to rely on a blockade rather than more active measures. This option was best suited to smaller forces confronting proportionally impressive defences. It was costly to keep a besieging army in the field for a considerable length of time; however, a passive blockade might avoid sacrificing forces unnecessarily while waiting for the defenders to eat their way through their supplies. To be successful, the besiegers had to be well provisioned and face little threat from relief forces, who would be given ample time to assemble and disrupt the siege. The protracted sieges of Kerak (1187–88) and Montreal (1187–89), are clear examples of this. With the army of Jerusalem crushed at Hattin, the fairly small detachments entrusted with reducing these mighty castles had no need to launch frontal attacks, opting instead to simply wear down their defenders.
Siege forts might also be built to strengthen a blockade. These were any fortified structure that helped inhibit communication, disrupt the flow of supplies into the besieged stronghold and impair the defenders’ ability to launch sallies. Their use was common in western Europe, especially in Norman spheres of influence. In the three decades leading up to the First Crusade, siege forts had been employed in southern Italy at the sieges of Bari (1068) and Troia (1071), in France at Ste Suzanne (1083), and in England at Rochester (1088) and Bamburgh (1095).
Antioch, town defences and siege forts of the First Crusade (with topography).
When the contingents of the First Crusade arrived at Antioch in 1097, they recognized that they had little chance of overcoming the city’s mighty defences by force. Although there were periods of severe famine during the siege, the crusaders were able to sustain themselves and defeat the relief forces from Damascus, Aleppo and Mosul. To contain the defenders, three siege forts were constructed: Bohemond built the first in late November 1097 to the north of the city, outside the St Paul Gate; Raymond of St Gilles built the second in early March 1098 to the west of the city, near the Muslim cemetery outside the Bridge Gate; and Tancred saw to the third in early April, most likely strengthening a monastery to the south of the city outside the St George Gate. It is perhaps surprising, given the regularity with which siege forts were built in Europe through the late eleventh century and early twelfth, that these were essentially the only ones constructed by the Franks. The shortage of timber and greater threat of significant relief forces may have discouraged their use. Although a number of castles were constructed in the early twelfth century to help impose distant blockades against certain cities, these were often built more than a few kilometres away and were rarely part of active siege operations – an exception was Raymond of St Gilles’ castle outside Tripoli.
In his account of Godfrey’s siege of Arsūf in 1099, William of Tyre makes a point of highlighting that a critical factor behind the failure of the siege was the Franks’ lack of naval support. He returns to this theme when discussing the unsuccessful siege of Acre in 1103 and emphasizes the meagre size of the Frankish ‘fleet, such as it was,’ employed against Tyre in 1111–12.31 William was most concerned with the involvement of naval forces because of the assistance they could provide in blockading a coastal town; however, Italian and crusader fleets could also supply provisions, manpower, expertise and building materials.
In the late eleventh century, Egyptian authority along the coast of the Levant had been reliant on its fleet. Following the First Crusade, Fāṭimid naval influence declined as increasing Italian commercial interests saw European vessels come to dominate these waters. This meant that relief by sea for most coastal towns under Fāṭimid control was less reliable, leading ports to the north, including Tripoli and Tyre, to look with increasing focus to Damascus for support. The waters off the coast of the Levant remained under European control for the remainder of the period, providing a vital lifeline for the Latin principalities as Frankish power peaked and then declined.
Acting in conjunction with land forces, a fleet could monitor the maritime approaches to a coastal stronghold and blockade its harbour during a siege. Frankish naval support was often provided by Italian forces, who offered their assistance in exchange for significant commercial rights in the town once it was captured. On some occasions, however, crusaders from northern Europe fulfilled this role: an English fleet that joined the First Crusade assisted with the siege of Antioch and Norwegian naval forces took part in the siege of Sidon in 1110. During the Franks’ initial conquest of the coast between the summers of 1099 and 1124, naval elements were present at 68 per cent of these sieges: 85 per cent of those where a considerable fleet was present were successful; two-thirds of those undertaken without significant naval support ended in failure. Although a fleet could offer support, it was rarely sufficient to capture a stronghold on its own. For example, Fāṭimid naval forces attacked but failed to take Jaffa in 1101, 1105, 1123 and 1151.
Naval forces were equally critical to the defence of coastal strongholds. Fāṭimid vessels from Tyre and Sidon brought assistance to Acre when it was besieged by the Franks in 1103; unfortunately for the defenders, no help came by sea the following year when the Franks renewed their siege efforts, contributing to the city’s capture. In 1108, the arrival of a Fāṭimid fleet was instrumental in relieving Sidon, defeating the Italian flotilla blockading the city and bringing more defenders; the Franks were finally persuaded to lift the siege when they learned a relief army from Damascus was on its way. Tripoli was less fortunate the following year, falling to its Frankish besiegers while awaiting relief. Unbeknown to the city’s garrison, Fāṭimid vessels were headed their way with provisions to last a year – they arrived at Tyre eight days after Tripoli was captured. During the siege of Beirut in 1110, a fleet of nineteen Fāṭimid warships overcame the naval force blockading the city, bringing the defenders supplies and a brief moment of respite before the city eventually fell. Later the same year, the Norwegian fleet assisting with the siege of Sidon successfully prevented the Fāṭimid fleet, anchored at Tyre, from providing assistance to Sidon’s defenders.
Symbolic of the changing tide of maritime power, Frankish besiegers imported supplies by sea during their siege of Tyre through the winter of 1111/12. This allowed the Franks to avoid Ṭughtakīn’s raiders, who were attempting to distract the besiegers and intercept supplies brought by land. Egyptian influence dipped so low during the following years that the Fāṭimid governor of Tyre was replaced for a period by a Damascene representative, Sayf al-Dīn Masʿūd, who ruled until additional supplies were brought from Egypt. Revealingly, it was Ṭughtakīn, rather than an Egyptian figure, who negotiated the final surrender of the city in 1124. Further south, Ascalon remained a Fāṭimid outpost for another three decades. The construction of the castle at Gaza around 1150 forced the Fāṭimids to rely exclusively on their fleet to send reinforcements and supplies to the town, which, according to William of Tyre, they did diligently four times per year. When Ascalon was besieged in 1153, an Egyptian fleet was able to break through the Frankish blockade, but the town ultimately fell after its landward defences were compromised.
Following the decline of the Fāṭimids, Saladin’s siege of Tyre in 1187 was a rare example of a Muslim siege supported by a fleet. In the aftermath of the battle of Hattin, Saladin’s army passed by Tyre’s mighty defences twice: first as it marched up the coast to Beirut, and then as it moved back down to Ascalon. Only after he had taken Jerusalem did Saladin return to Tyre, making camp on a hill outside the city on 22 November, having paused to wait for his siege engines to arrive and his son, al-Ẓāhir, to join him. The Franks were able to use their fleet to harass the besiegers, shooting them from both the north and south with crossbows as the Muslims advanced along the isthmus between the city and the mainland. Saladin summoned his fleet of galleys from Egypt, which forced the Frankish vessels to seek refuge in Tyre’s harbour, completing the blockade. The size of the Muslim fleet was probably around a dozen vessels; Ibn al-Athīr places the number of galleys at ten, while Ernoul gives the figure of fourteen. Regardless, the siege was eventually broken when the Franks made a successful naval sally on 30 December, seizing five of the Egyptian galleys anchored opposite the harbour. When he saw the engagement, Saladin ordered a frontal attack, but this was countered by a sally led by Hugh of Tiberias. Defeated on both fronts, Saladin sent what remained of his fleet to Beirut. The Franks gave chase, forcing the Muslim sailors to ground their vessels and flee overland. With the blockade no longer in effect, his forces making little progress against the city’s formidable defences, and winter weather setting in, Saladin withdrew on 3 January, disbanding most of his army.
Ibn al-Athīr, who owed no personal loyalty to Saladin, summed up the siege fairly condemningly: ‘When Saladin saw that the Tyre operation would be a long one, he departed. This was his practice. Whenever a city held firm against him he tired of it and its siege and therefore left it.’32 To be fair, Saladin’s decision was a sound one: his army had been campaigning since the start of the summer, the strength of Tyre’s fortifications was renowned and the city was swelling with defenders, refugees who had gathered there as other strongholds fell. Looking ahead, Saladin may also have expected an attack from Europe in response to his capture of Jerusalem. Rather than continue a difficult siege through the winter, Saladin cut his losses and appeased his forces who longed to return home. While his fleet had been bloodied, additional vessels from Egypt would prove instrumental in supporting the defenders of Acre at times through the arduous siege of 1189–91.
A fleet could also facilitate operations by providing logistical support. In 1115, Fāṭimid land and sea forces made a coordinated assault against Jaffa, making a rapid attack and then withdrawing only to strike again ten days later. During this operation, the fleet appears to have transported the ladders and artillery used by the land forces. Decades later, in a final gasp of Fāṭimid aggression against the Franks, a large fleet, said to have included as many as seventy vessels, raided Jaffa and other towns as far north as Tripoli in 1151. Nūr al-Dīn may have hoped to coordinate an attack with the arrival of these raiders, but was preoccupied with his efforts to acquire Damascus.
In the second half of the twelfth century, Christian forces exploited their control of the sea to import siege engines along with large marine forces. The Byzantines appear to have brought siege equipment to Damietta in 1169, much as the Sicilians did when attacking Alexandria in 1174. During the Third Crusade, Richard I of England used his fleet to transport artillery he had commissioned while on Sicily to Acre along with his army. After Acre had been captured, Richard disassembled these engines and stowed them on his ships, allowing him to move them rapidly by sea, as was done when he attacked Dārūm in 1192. This type of manoeuvre would have been appreciated by Baybars, who, despite his aspirations, was never able to outfit a successful fleet.
Although Egypt retained some of its naval traditions, and the Levantine coast gradually fell from Frankish hands, the Mamlūks launched only one significant naval operation between 1260 and 1291: a disastrous attack against Cyprus in 1271 that ended with the fleet wrecked on the rocks around Limassol harbour. Despite this failure, the Mamlūks had little need of a fleet given the dominance of their army. The only real threat they faced came from the Mongols, who had no naval presence in the region. Leaning on the overwhelming strength of their armies, neither Baybars nor Qalāwūn initiated a siege that relied on blockading forces – the Mamlūks intended nearly all of their sieges to be concluded through aggressive action.
As discussed earlier, the involvement of maritime forces might influence what engines were employed at a siege. The arrival of a Genoese squadron and then an English one probably facilitated the construction of the siege forts at Antioch during the First Crusade. Raymond’s siege fort was built soon after he and Bohemond set off to escort elements of the latter to the city in March 1098, suggesting they returned with building materials. The following year, it was Raymond who once more championed the expedition to escort the Genoese sailors trapped in Jaffa to Jerusalem, which returned with the timber used to build his siege tower. Over the following decades, similar links can be drawn between the use of significant siege engines and the presence of naval elements.
Some historians have suggested that the frequency with which siege engines were built when Italian elements were present had more to do with the carpentry or engineering skills of the sailors than simply the supplies they brought. The experience of such men handling rigging and manipulating large beams should not be downplayed; however, the early twelfth century was a time when the nobility took an active role in the design of castles and siege engines, while many below this rank would have built their own houses, often using few if any metal fasteners. In such an age, when men would have been far more skilled with an axe than the average person today, the abilities of the rank and file, let alone those with marginally more skill, should not be underestimated. Although some were impressively large, siege engines of this period were structurally simple and the lifting and manoeuvring of large beams into place would have been by far the greatest challenge associated with their construction.
Perhaps the most obvious link between certain nautical traditions and siege engines can be seen at the siege of Lisbon, a sideshow during the Second Crusade. While Louis VII of France and Conrad III of Germany took the overland route via Anatolia, a group of Anglo-Norman and German crusaders made their way east by sea, sailing around the Iberian Peninsula. On their way, this group took Lisbon in 1147 on behalf of Alfonso I of Portugal. During the siege, the Anglo-Normans, who attacked the city from the west while the Germans besieged it from the east, erected two traction trebuchets to support their mining efforts against a section of the western wall, between the Iron Gate and banks of the Tagus River. One of the trebuchets, operated by the knights and their table companions, was set up opposite the Iron Gate, while another, placed near the riverbank, was manned by sailors. Although this supports the theory that sailors had certain talents that were transferrable to the construction and operation of siege engines, this was evidently not a limiting factor, as the knights do not appear to have faced any trouble working their engine.
While the benefits of a fleet are fairly obvious, broader developments led to an increasing reliance on naval transportation. Whereas most participants of the First and Second Crusades had travelled overland, the only contingent to follow this route during the Third Crusade was that led by Emperor Frederick I; all others opted to travel by sea for at least the final leg of the journey, setting sail from southern France or Italy, if not the northern ports of Europe. Subsequent crusading armies would similarly arrive in the East by sea. This was paralleled by events in the Levant: whereas the majority of Frankish forces marched overland when attacking Egypt in the twelfth century, those who joined the Fifth and Seventh Crusades instead travelled by sea.
The Latin principalities along the coast relied on their maritime links to Europe for their economic success and the manpower that sustained them. The county of Edessa and principality of Antioch also profited from alliances with certain Armenians. Although the kingdom of Jerusalem benefited from these links through the accession of the first two counts of Edessa to the throne of Jerusalem, Cilician Armenia was distant and never strong enough to send help as far south as Palestine. Of the four Latin principalities, the county of Edessa was the only one that did not control a significant stretch of the Mediterranean coast, weakening its ties to Europe. Revealingly, the Second Crusade was launched in response to Zankī’s capture of Edessa in 1144, but the crusaders quickly reoriented their focus southwards. The decision to attack Damascus has drawn considerable criticism ever since the attack was made in 1148, but Damascus was much closer to the Franks’ regional centre of power and a more appealing target for Europeans. By comparison, Edessa was three times as far from the coast, lacked the religious and political significance of Damascus and was controlled by Nūr al-Dīn, who had shown his resolve to retain the city in 1146.
Although the Franks had little use for naval siege forces by the time stronger Muslim powers began to appear in Aleppo and then Damascus in the mid-twelfth century, the importance of the Italian merchant communities remained. They provided the conduit through which trade flowed between Europe and the Near East and were often the ones responsible for ferrying crusaders to and from the Levant. This service became critical as the Franks placed increasing reliance on the support offered by European crusaders in the second half of the twelfth century, something that became a dependence following the battle of Hattin and Saladin’s conquests. Although the Franks would look to recover the lands they had once held further inland, the seats and focus of Frankish lordship were already firmly rooted along the coast by the time participants of the Third Crusade set sail for Europe. When Louis IX later thought to fortify a position on the road between Jaffa and Jerusalem in the early 1250s, the local baronage dissuaded him, pointing to the site’s distance from the sea and the difficulties associated with resupplying an inland site. Their suggestion that he instead refortify Sidon was accepted – the Franks had become a coastal power.
A stronghold might be taken through less aggressive means if parties were willing to talk. This could result in a negotiated surrender; alternatively, it might involve deception. In 1098, during the siege of Antioch, Bohemond made contact with an Antiochene within the city, named Firuz, who commanded one of the towers along the southeastern section of the city’s walls. Different accounts of the event provide different details, but all accept that Firuz agreed to help Bohemond and a party of men into his tower on the night of 2/3 June 1098. Having gained control of the tower and adjoining walls, the Franks opened a gate and the town was quickly occupied, leaving only the citadel beyond their control twenty-four hours after the operation began. When Nūr al-Dīn finally acquired Damascus fifty-six years later, he had a similar group of betrayers to thank. Other strongholds changed hands through less sensational means.
Treaty
As discussed in Chapter 2, some treaties between Frankish and Muslim parties involved the sharing of territory, others stipulated the transfer of strongholds. Envoys exchanged by Richard I and Saladin through 1191–92 negotiated terms of a peaceful division of Palestine. The final agreement recognized the status quo, including a provision that Ascalon not be fortified in the immediate future, and gave Christian pilgrims access to Jerusalem. More dramatically, al-Kāmil’s talks with Frederick II resulted in Jerusalem’s peaceful return to Frankish hands in 1229, and negotiations with al-Ṣāliḥ Ismāʿīl during the Barons’ Crusade brought Safed, Beaufort and eastern Galilee back under Frankish control in 1240. The results of these negotiations were considerable, but so too were they rare; a large crusading force was present in each case, giving the Franks considerable, if only temporary, leverage. While great potentates might trade castles and territory through treaties, lesser figures could have an equally profound impact on the outcome of individual sieges.
Betrayal
Acts of apparent treachery were typically motivated by prospects of wealth and security. While anyone in a stronghold might be a potential betrayer, chroniclers often identify treacherous figures as members of marginal or minority communities. Firuz, for example, may have been an Armenian convert to Islam, who was motivated to help the crusaders because the ruler of Antioch, Yaghī Siyān, had confiscated his wealth, or simply by Bohemond’s promise of a reward. Despite the mixed populations of most cities in the region, examples of Muslims or Christians attempting to help their besieging coreligionists are relatively rare. Members of different ethnic groups of the same religion, who were often able to gain positions of higher power, are more often identified or implicated in acts of betrayal during sieges.
Although some predominantly Christian communities around Antioch and Cilicia used the arrival of the First Crusade to rise up and expel their Muslim garrisons, these episodes are unique to this specific occasion. Reports of treachery or disloyalty, let alone insurrection or rebellion, are extremely rare, casting doubt on the traditional theory that the Franks ruled over a hostile population of locals. Those living on the land appear to have been far more interested in feeding their families than the religious persuasion of the figure who claimed authority over the region, while merchants were often drawn to lucrative markets, regardless of whose flag they were under. Nevertheless, religion was central to polemic arguments and it seems some contemporary authors took advantage of this. In a strange episode recounted by Ibn al-Athīr, the ruler of coastal Jabala, Ibn Ṣulayha, conceived a number of plots to divert Frankish attacks in the early twelfth century. In one, he had local Christians reach out to the Franks and promise to help them gain one of the city’s towers; the Franks who subsequently climbed up were then cut down one by one as they reached the top of the wall. The episode appears to be a complete fabrication, inspired by events that took place at Antioch a few years earlier, and is suspiciously absent from Ibn al-Qalānisī’s account, from which Ibn al-Athīr drew extensively.
The Franks relied quite heavily on the support of local Christians, nowhere more so than in the county of Edessa, but it was members of these communities who were accused most of treachery during sieges of Frankish strongholds. Eastern Christians may have been convenient scapegoats for Latin chroniclers who regarded them with distrust, yet charges of betrayal were not limited to Frankish authors. During Mawdūd’s attack on Edessa in 1112, Eastern Christian accounts describe how a group of Armenians helped a party of Muslims up into their tower. The act of treachery was quickly discovered and Joscelin of Courtenay, who was at that point commanding the city’s defence, led an attack that retook the betrayed section of defences.
When dealing with Muslim figures, Kurds are accused of committing acts of betrayal with alarming frequency. When Mawdūd besieged Frankish Turbessel in 1111, Joscelin was able to break the siege by bribing one of the Kurdish emirs. Later, in the 1190s, the Kurdish governor of Jubayl was said to have been bribed by the Franks to hand over the town in the wake of the Third Crusade. Such acts were by no means limited to sieges involving a Christian party: Usāma ibn Munqidh relates a story in which a Kurdish emir led a plot to betray Amida (Ā mid) to the Artuqid army besieging the city at the time, and it was a Kurdish emir who was accused of betraying the barbican of Sinjār to Saladin in December 1182, leading to the fall of the city.
Kurdish fighters and emirs, including Saladin and his uncle, Shīrkūh, were an important part of many of the twelfth-century Muslim armies raised in the Jazīra to fight in western Syria; however, the fraction of these armies they accounted for is unclear. Accordingly, determining if Kurds were proportionately responsible for more acts of betrayal than members of other Muslim ethnic groups is hard to say. In all likelihood, there was at least a degree of racial prejudice in some of these accusations, which were launched predominantly by Arab historians. It is also worth noting that Usāma praises various Kurdish figures elsewhere in his work.
When Damascus came under threat from the Zankids in the mid-twelfth century, loyalties were put to the test. During Nūr al-Dīn’s blockade of the city in 1154, Mujīr al-Dīn attempted to buy the support of the Franks, reportedly offering them Baalbek in exchange for their help, much as Muʿīn al-Dīn had offered Bānyās to Baldwin III in 1140 when it had been Zankī, Nūr al-Dīn’s father, who had threatened the city. Meanwhile, others who looked favourably towards Nūr al-Dīn made different arrangements. Ibn al-Athīr claims it was the town militia that rose up in support of Nūr al-Dīn; however, Ibn al-Qalānisī, a resident of the city at the time, points out that it was a Jewish woman who was responsible for the fall of the city – only after she let down a rope and helped a number of soldiers over the wall did the locals rally and open the city gates.
Damascus was similarly contested at the end of the century, following Saladin’s death. When al-ʿAzīz and Al-ʿĀdil besieged the city in 1196, one of al-Afḍal’s emirs betrayed the eastern gate, Bāb Sharqī, to Al-ʿĀdil, forcing al-Afḍal to surrender by the end of the day. Three years later, when Al-ʿĀdil was then in possession of Damascus, al-Afḍal and al-Ẓāhir Ghāzī besieged the city. One of al-Afḍal’s emirs, Majd al-Dīn, arrang ed for thecentral northern gate, Bāb al-Salām, to be opened for him. Seeking glory for himself, the emir did not tell al-Afḍal or any of the other emirs of his plan. Rushing into the city unsupported, his fairly small force was confronted and cut down as it passed the Umayyad Mosque.
In a very different scenario, the defenders of the Cave of Tyron, a Frankish stronghold in the mountains east of Sidon, were accused of having been bribed by Shīrkūh into surrendering their stronghold in 1165. In what appears to have been a rare example of mass defection, the defenders are said to have fled to Muslim territory afterwards; their leader was the only one who was later apprehended.
Guile
If a besieging army could not persuade any of the defenders to help them, they could try to trick them. One of the ways to do this was by intercepting communications. During the siege of ʿAzāz in 1125, a pigeon carrying a letter to the defenders landed in al-Bursuqī’s siege camp. The letter had been sent by Baldwin II, who was in Antioch, in response to the garrison’s appeal for help, and informed the defenders that a force was preparing to move to their relief. Al-Bursuqī changed the letter to say that no relief was coming, signing the new note with Joscelin of Courtenay’s name. Unfortunately for al-Bursuqī, the trick did not work and the garrison stubbornly continued to hold out until a relief force arrived and defeated the besiegers. In 1127, Zankī, who had recently succeeded al-Bursuqī in Mosul, similarly intercepted a pigeon when besieging Nisibis, a possession of Ḥusām al-Dīn Timurtāsh of Mardin. The letter instructed the garrison to hold out for another five days, at which point relief would arrive; however, Zankī altered the note to read twenty days, betting that this was too much to ask of the defenders. The gamble worked and Nisibis was quickly surrendered.
Baybars was also known to engage in deception. When he arrived to besiege Beaufort in 1268, the Frankish defenders sent word of their situation to Acre. The returning messenger was intercepted and a forged letter was sent into the castle to sow discord among the Franks, informing the castellan that he should beware of certain figures under his command. According to Ibn ʿAbd al-Ẓāhir, Baybars’ secretary and later biographer, Baybars forged another letter when besieging Crac des Chevaliers in 1271. On this occasion, the new letter ordered the castle’s commander to surrender the stronghold, which he duly did. Later the same year, during his siege of Montfort, Baybars is said to have shot a pigeon carrying a note from a spy in his army. The message apparently contained little of value as Baybars, taking advantage of a theatrical opportunity, handed the note to the Frankish envoys who were at that time visiting his siege camp, remarking that it was addressed to them. When he set his sights on Cursat in 1275, Baybars had the castellan of the patriarch’s castle lured out and detained. Although a letter was sent ordering the garrison to surrender, they refused and held out for seven months. Meanwhile, the castellan joined his father in the dungeons of Damascus, where he died.
One of the more controversial events of the period occurred when Baybars summarily executed the Templar garrison of Safed in 1266. During the siege, he tried to sow discord between the Franks and the Syrian Christians by offering the latter, who accounted for some of the archers and light cavalry in the castle, safe-conduct to leave if they chose. In the negotiations of surrender that followed, the sources offer different accounts of events, obscuring a clear understanding of what actually happened. An agreement of some kind was concluded and the Franks came down from the castle believing they were free to go. A number of explanations are given to explain the massacre that followed. From a Frankish perspective, the Templar of Tyre suggests that the massacre was an act of revenge carried out against the defenders for throwing back a customary gift that they had been offered at the start of the siege. Muslim sources provide a number of explanations, each containing a reason why the agreement of surrender had not been formalized or was otherwise invalid. Whatever the circumstances may have been, Baybars was successful in luring the Frankish defenders out of the castle and securing their arms and wealth before carrying out what was likely an act of more general revenge, perhaps intending it to also serve as an example for others who defied him.
The Nizārī Assassins were masters of infiltration and became famous for the daring high-profile murders carried out by their members. In either 1109 or 1114, a group infiltrated Shayzar while the ruling family was taking in the Christian Easter celebrations, briefly occupying the citadel before they were expelled. About three decades later, they again turned to trickery when taking the castle of Maṣyāf, which was held by a mamlūk of the ruler of Shayzar. On another occasion, they bribed a friend of the gatekeeper of an isolated strongpoint known as Ḥiṣn al-Khurayba (‘fortress of the stone heap’). In return for money and an iqṭāʿ, the m an visited the castle and, upon gaining entrance, murdered his friend, the gatekeeper, along with the other three occupants of the castle before handing it over to the Assassins.
Attempts to infiltrate a stronghold and take it from the inside were far less likely to work than is presented in the movies and efforts to do so were correspondingly rare. In one such instance, a group of Armenians from Edessa sneaked into Kharpūt in disguise to rescue Baldwin II in 1123. The party successfully killed the jailers and freed the king, who promptly led a takeover of the castle. When news reached Balak, he returned and led a more traditional siege, setting his sappers to work undermining the stronghold while his artillery harassed the defenders. Joscelin I of Edessa, who had been among the Frankish prisoners, had left to get help but was unable to return with a relief force before the mines were lit and the defenders were compelled to surrender the stronghold that had been their prison. In an act of swift revenge, the Armenians and lesser Frankish prisoners were all executed; only Baldwin and a few other high-ranking figures were spared.
Maṣyāf. (monumentsofsyria.com)
On 2 May 1267, Baybars led a disguised force against Acre, carrying captured Templar and Hospitaller banners. The Franks were not fooled as the Mamlūks approached, forcing Baybars to content himself with some fighting beyond the city’s walls. He returned and raided around the city two weeks later, remaining in the region for four days while his troops laid waste the surrounding landscape and destroyed Ricordane, the Hospitallers’ fortified mill. A few years later, the Mamlūks tried to use the same trick, this time at sea. The fleet sent against Cyprus in 1271 sailed under banners with crosses on them, but they again fooled no one. Although elaborate schemes involving deception or treachery carried the chance of significant rewards, strongholds more often changed hands through negotiations.
Negotiations typically preceded any surrender – unconditional submission was rare. In some cases, however, the agreed-upon terms were so generous that they might be more accurately considered a trade. This was a practice most common among senior Muslim figures, a way of weakening a rival or rebellious subordinate while preserving his dignity and mitigating the likelihood that a more deep-seated rivalry would develop. For example, in 1110, Ṭughtakīn gained the surrender of the rebellious governor of Baalbek, following a siege of thirty-five days, by offering him Ṣarkhad in exchange. Although such terms were usually offered or requested after a siege had started, sometimes talks were opened before hostilities began, reducing or even avoiding open conflict between closely associated forces.
As part of Zankī’s efforts to consolidate his power in western Syria, he besieged Homs in late 1135. The city’s ruler, Khīr Khān, had been an ally, but Zankī had him arrested shortly before siege operations began, compelling him to order the defenders to surrender, which they did not. Although Khīr Khān’s sons successfully defended the city, they were so weakened by the affair that they offered Homs to Shihāb al-Dīn Maḥmūd of Damascus in return for Palmyra. In 1146, following Zankī’s death, Muʿīn al-Dīn, regent of Damascus, set out to acquire Baalbek. The city was administered by Najm al-Dīn Ayyūb, Saladin’s father, who, along with his brother, Shīrkūh, had joined Zankī’s service in Mesopotamia and followed him to western Syria, where they went on to serve his son, Nūr al-Dīn. Ayyūb recognized that Nūr al-Dīn was preoccupied with affairs elsewhere, so quickly surrendered Baalbek in return for a different iqtā[. In 1154, after Nūr al-Dīn had established himself and finally gained entrance to Damascus, he benevolently granted Muʿīn al-Dīn a significant iqtā[, which included Homs, in exchange for the formal surrender of the citadel of Damascus. The lord ofQalʿat Jaʿbar, Shihāb al-Dīn, was less inclined to give up his mighty castle. Having been captured by the Banū Kilāb and brought before Nūr al-Dīn in 1168, he refused to exchange his castle for another. The mighty stronghold resisted two forces that Nūr al-Dīn sent against it, but eventually the castle was given up in return for Sarūj and al-Mallāha, as well as a sum of 20,000 dinars, a lucrative iqṭāʿ but one without a significant castle.
The practice of exchanging strongholds remained common throughout the Ayyūbid period. Two years after Saladin’s death, ʿIzz al-Dīn Jūrdīk al-Nūrī surrendered Jerusalem to al-Afḍal and al-Ādil, for which he received Bānyās and the surrounding region, including Belvoir and the Golan. A generation later, al-Sā lih Ismāʿīl and al-Kāmil struggled for control of Damascus following the death of their brother, al-Ashraf, in 1237. Although al-Ṣāliḥ Ismāʿīl initially took control of the city, he was quickly forced to accept Baalbek and Bosra in exchange when confronted by al-Kāmil. In 1239, al-Kāmil’s son and successor in Egypt and Damascus, Al-ʿĀdil II, became wary of the increasing independence shown by his cousin, al-Jawād Yūnus, to whom the administration of Damascus had been entrusted. When al-Jawād rejected an offer of Alexandria and Montreal as an iqtā’ in return for relinquishing Damascus, Al-ʿĀdil II mobilized his Egyptian forces. This led al-Jawād to offer Damascus to al-Ṣāliḥ Ayyūb, Al-ʿĀdil II’s brother and to whom al-Kāmil’s northeastern lands had passed, for which he would receive Sinjār and Raqqa.
Trading or offering alternative strongholds as a condition of surrender was not unique to the Frankish period. For example, when Malikshāh took Aleppo in 1086, he compensated its dispossessed ruler with the iqṭāʿ of Qalʿat Jaʿbar. Nor, it seems, were the Franks excluded from this practice. According to Ibn al-‘Adim, the Muslim ruler of ʿAzāz offered Tancred his town in exchange for another in 1107 or 1108. Among Frankish parties, however, strongholds rarely changed hands in this manner.
When Frankish fiefs were exchanged, it was typically done when a baron gave up a lesser holding in order to inherit a greater one, or sold off lands he could no longer support. The former was a means through which a prince could prevent any one baron from becoming too powerful: Tancred relinquished the principality of Galilee in order to take up the regency of Antioch, and Philip of Milly surrendered Nablus in return for permission to inherit the lordship of Transjordan. When Frankish nobles were instead downgrading, they typically sold their stronghold, and associated fief, to one of the military orders. This was done by lords great and small who could no longer meet the obligations expected of them, or had no further desire to do so. At large sites, the military orders were occasionally given an interest or partial share from the mid-twelfth century. Humphrey II of Toron famously gave half of Bānyās to the Hospitallers in 1157, Maurice of Transjordan gave the same order property in both Montreal and Kerak in 1152, and defence of Safed appears to have been entrusted to the Templars before the order received outright ownership of the castle in 1168. Julian of Sidon was one of the last barons to hold a major interior castle, selling Beaufort to the Templars only after the Mongol invasion and their raid on Sidon in 1260.
The majority of strongholds that fell following a siege did so after some kind of surrender was arranged; instances of towns and castles falling to the sword were comparatively rare but far from unheard of. Town defences were more likely to be captured in this way than citadels, much as the outer defences of a castle were more likely to be taken by force than the keep or another final point of defence. While the defenders of Nicaea negotiated a surrender with the Byzantines before elements of the First Crusade broke into the town, the civic defences of both Antioch and Jerusalem were taken by force, although both citadels held out long enough for those within to negotiate a surrender. During the following decades, the towns of Haifa (1100), Caesarea (1101, 1265), Tortosa (1102), Bālis (1108), Tripoli (1109), Beirut (1110), Hama (1115), Bānyās (1132, 1157), Buzāʿa (1138), Edessa (1144, 1146), Bilbays (1168), Tiberias (1187, 1247), Jaffa (1187), Arsūf (1265), Antioch (1268) and Acre (1291) were among the towns taken by storm, although the citadels of most surrendered on terms, even if these resigned those within to a life of slavery. In a reversal of the norm, when the town of Manbij surrendered in 1176, those in the citadel continued to resist, leading it to be taken by force. The castles of al-Ḥabis Jaldak (1111), Jacob’s Ford (1179) and Ascalon (1247) were among those that fell to the sword.
It can be difficult to determine how exactly some sieges ended, as sources were at times provided with contradicting information, and some deliberately presented events to suit certain agendas. For example, Ibn al-Qalānisī reported that Ḥārim fell to the sword in 1157, while William of Tyre wrote that its defenders surrendered, securing both their freedom and property. Despite their differing reports, both men were contemporaries: Ibn al-Qalānisī was in his eighties and living in Damascus, while William of Tyre was in his late twenties. Although the latter was then living in Europe and would not begin to compose his history for another decade, a Syriac account supports his suggestion that the castle surrendered before it could be stormed. Among the events that followed the battle of Hattin, Bahāʾ al-Dīn recorded that many strongholds, including Toron (1187), Bakās (1188) and Bourzey (1188), were taken by storm, which contradicts the accounts of most other Muslim historians, who followed ʿImād al-Dīn’s version of events more closely, asserting these places were instead surrendered to Saladin’s forces.
There was more at stake than simply possessing the besieged stronghold. If a castle or town fell without an agreement of surrender in place, the customs of war provided no security for the defeated or their property. The longer a siege lasted, the more financing it required: a quick siege could be profitable but a long one costly. Accordingly, both attacking and defending forces had to weigh these considerations. If a stronghold fell quickly by force, there was plenty of plunder to go around, more so since the besieged did not have the opportunity to eat through their provisions. This was the most profitable outcome for the rank and file of a besieging army and a natural incentive. Saladin’s troops were reportedly disappointed when the sultan accepted the surrender of Jaffa in 1192, denying the army the opportunity to sack the town. When Mamlūk forces broke into Antioch in 1268, the city was plundered and those who could not flee to the citadel were slaughtered or enslaved. There was so much loot to be taken that it took two days to properly divide it all, during which period the gates were closely guarded to ensure nothing, and no one, escaped.
Unless a besieging commander was in dire need of liquid capital to pay his army or gain their support, which could be expected if a stronghold fell to the sword, a quick surrender was often in his best interests. This limited the disruption to local commerce, from which he now stood to gain, and might mitigate ill-will among the conquered by sparing them the hardships of a lengthy siege. As a siege dragged on, terms typically became less generous: they needed to be liberal enough to persuade the defenders to surrender, but sufficiently harsh to extract enough of a return to compensate the investment that had gone into the siege, a figure that grew with each day the siege ran on. Terms of surrender could range from allowing defenders to leave with only their movable property, to forbidding them from removing any arms or money, to depriving them of any property or even their freedom, reducing them to a state of slavery.
The besieged had the reverse to consider: they were forced to weigh their prospects of holding out and the likelihood of a relief force arriving against what they might be able to extract in return for their surrender. In instances where it was clear that no relief would be coming, strongholds typically fell quite quickly. Garrisons would often make a show of resistance, impressing on the besiegers the potential challenges they could expect to face, but were quick to accept or propose terms before their opponents could invest much time or patience.
While the besieged had only their immediate situation to contemplate, the besiegers also had to judge how much time and effort they were willing to invest in this component of what was naturally a broader campaign, even if ingress and egress from enemy territory were the only other elements. During more complex campaigns, or those that were part of a broader expansionist policy, besiegers were also aware of the message that their terms conveyed to others. By initially offering generous terms, it might be hoped that the defenders of other strongholds would be encouraged to surrender relatively quickly; when resistance was shown, more draconian terms sent an equally clear message.
The multitude of factors to be considered led the terms offered and agreed upon to vary from siege to siege. Some short sieges resulted in harsh terms while some long ones ended with exceptionally generous ones. In this respect, Saladin’s campaigns after the battle of Hattin can be quite enlightening. Immediately after the battle, the consequences of his victory were far less obvious than they would later become. His primary interest was to acquire certain key strongpoints deep in Frankish territory as quickly as possible. Foremost among these was Acre, evident in his rush to the coast and the generous terms that he quickly offered, allowing the defenders to freely leave with their possessions. On their part, those who were in the city were probably still dazed from the defeat at Hattin and had little faith in their ability to resist, let alone that they would be quickly relieved. Saladin was happy to offer similar terms through the remainder of 1187 as he pressed his advantage. His strategy was to acquire as many Frankish strongholds as possible before the kingdom could recover and mount any kind of organized resistance, as eventually formed in Tyre, or a significant force of crusaders arrived from Europe.
From a tactical point of view, Saladin may have made a significant error in the autumn of 1187. After joining his forces with those of Al-ʿĀdil at Ascalon, the more prudent military decision would have been to attack Tyre, rather than Jerusalem, securing the entire coast south of Tripoli and depriving the Franks of a port south of the Lebanon. A crusading army landing at Tripoli or a port further north would be compelled to march south via the narrow coastal road past Beirut, where a small force could halt the advance of a much larger army, or through the Biqāʿ Valley and past Baalbek, a region that had never fallen under Frankish rule. The decision to instead invest Jerusalem was a political one, fitting with the rhetoric of jihad that had helped Saladin gain support and legitimacy. Likewise, the harsher terms demanded at the end of the siege, which compelled every Frankish person to pay a ransom in exchange for their freedom (10 dinars/bezants for each man, 5 for a woman and 1 or 2 per child), reflected in part the considerable blood that had been spilt by the Franks in 1099 – Saladin could not be seen to retake the city without a measure of similar brutality. The capture of Jerusalem sent shockwaves through both the Muslim and Christian worlds; however, the period spent besieging the city afforded the defenders of Tyre time to prepare themselves, their numbers swelling with the arrival of those who could afford to pay the cost of their release from Jerusalem. By the time Saladin moved against Tyre, winter had arrived and the defenders exhibited considerable resolve.
Saladin’s campaign in Palestine, 1187.
During Saladin’s campaign through western Syria the following summer, the harshest terms accompanied the shorter sieges. At Bourzey and Saone, and possibly also Sarmīniyya, Muslim forces had captured the larger part of both castles before they were surrendered, suggesting they lacked a sufficient number of defenders. Others, such as the Templar castle of Baghrās, appear to have been more prepared and better defended, demonstrating this before accepting favourable terms reasonably quickly.
Throughout the broader region, the greatest castles were those that held out longest, while some, including Crac and Margat, were simply bypassed. Safed and Belvoir continued to hold out through 1188, as their garrisons may have expected to be the first to receive relief if a significant force were assembled or a group of crusaders suddenly arrived. Only 37km and 55km from the coast respectively, they would have had a reasonable impression of the events going on around them. The defiance shown by the defenders of Montreal and Kerak is harder to explain. Southern Palestine had fallen completely before the end of 1187, leaving these two castles isolated. They may have been lulled into a sense of comfort by the passive blockades imposed by their besiegers or remained hopeful that the situation to the north was less dire than that which we can assume their besiegers were describing to them. Regardless, after two of the longest sieges of the period, the defenders of both were allowed to return to Frankish lands when they surrendered, testament to their perceived ability to continue resisting and the limited resources that Saladin was willing to devote to their capture.
Saladin’s campaign in western Syria, 1188.
Perhaps the most ingenious, and certainly most deceitful, defence was that of Beaufort. When Saladin led his army against the castle in April 1189, Reynald of Sidon came out and negotiated a three-month truce, at the end of which he promised to surrender the castle. Muslim accounts assert that Reynald intended to enter Saladin’s service, arranging to have a residence in Damascus and a salary, while the three-month grace period was required to collect his family, who were at Tyre. Pragmatically, the three-month truce allowed Saladin to turn his attention elsewhere; his peace with Bohemond III of Antioch ended in May and Frankish opposition was rallying in Tyre. When the three months were up and Saladin returned, it seems he was aware that Reynald had been acting in ill faith, having used the truce to strengthen the castle and gather supplies. Reynald visited Saladin a few days ahead of the deadline to request more time, claiming Conrad of Montferrat had not allowed his family to leave Tyre. When he returned on the due date to ask again he was detained and sent first to Bānyās and then onward to a prison cell in Damascus. Despite the episode of intrigue, the defenders obtained Reynald’s release when they finally surrendered in 1190, by which point the Frankish siege of Acre had begun. The defenders of Kerak had similarly been able to secure not only their own liberty when they capitulated in 1188, but also that of Humphrey IV of Toron, a prisoner since the battle of Hattin.
Eighty years later, the rapid fall of the strongholds around Antioch following Baybars’ capture of the city in 1268 is understandable. Unlike Saladin’s conquests in the region, when garrisons had offered at least a nominal show of resistance, most strongholds, even those held by the military orders, were abandoned to Baybars without a fight – liberty was not offered to those who compelled the Mamlūks to approach their walls before surrendering. The main difference between the campaigns of 1188 and 1268 was that Antioch had not fallen to Saladin. By comparison, it was the first stronghold that Baybars took upon invading the region, removing the natural hub from which relief forces would most likely be organized, even though Bohemond VI was based in Tripoli by this point. By the end of the spring of 1268, the only remnant of the principality of Antioch in Frankish hands north of Latakia was Cursat, the patriarch’s castle, which had surrendered half of its territory to be left alone.
There was little shame in surrendering a stronghold when the odds were almost immeasurably stacked against the garrison, but doing so when confronted by a more modest force, when resistance was judged to be quite possible, was criminal. When Humphrey II of Toron, lord of Bānyā s and constable of the kingdom of Jerusalem, accompanied Amalric to Egypt in 1164, he left the defence of Bānyās to one of his knights, Walter of Quesnoy. The rapid fall of the town when it was besieged by Nūr al-Dīn that October led to suspicion that Walter had accepted a bribe, having acted in collusion with a canon named Roger. The accusations of treachery terrified those who were implicated. The following year, the defenders of the Cave of Tyron were believed to have colluded with Shīrkūh when surrendering their stronghold. The castellan was the only one among the defectors to be apprehended, whereupon he was taken to Sidon and hanged. Around the same time, Shīrkūh took the cave castle of al-Ḥabis Jaldak, east of the Jordan. Defence of the cave castle was led by twelve Templars, whom Amalric, upon learning the stronghold had fallen when his relief force reached the Jordan, accused of surrendering too quickly. According to William of Tyre, the king had the twelve Templars hanged. Al-Ḥabis Jaldak had returned to the Franks by 1182, when it was taken by a contingent of Saladin’s army led by Farrukhshāh. This time, Fulk of Tiberias, a vassal of Raymond III of Tripoli in his capacity as prince of Galilee, was blamed when it was alleged that the leaders of the castle’s garrison, a group of local Syrian Christians whom Fulk had appointed, had been bribed to surrender. Although it is possible that there was treachery in each case, it is not hard to sympathize with the defenders of al-Ḥabis: it was the kingdom’s most easterly outpost, nestled claustrophobically into a valley wall, rather than perched on top of a peak where a relief force might be viewed from a distance.
In 1249, Louis IX and elements of the Seventh Crusade fought their way ashore near Damietta. Once they had landed, the Muslim force stationed in the city, which had initially attempted to oppose the crusaders’ landing, withdrew. Deserted by the army, many of the city’s inhabitants followed. Al-Ṣāliḥ Ayyūb’s anger at Damietta’s abandonment was not limited to Fakhr al-Dīn, who led the forces that had withdrawn from the city, but included the citizens who had fled, a number of whom he ordered executed; responsibility for Damietta’s defence, at least in the eyes of the sultan, seems to have extended to its inhabitants.
With so much interaction between the various groups of Muslims and Christians, it is not surprising that their fortifications and siege techniques bear such striking resemblances. Arms and armour, along with the make-up of armies, varied, reflecting different cultural, political and economic factors, but siege engines and techniques were fairly universal. All parties were well acquainted with mining and the various engines and tactics employed by others, so it was enough to see a ram, a siege tower, a trebuchet or a mine to appreciate how it worked and, with a little bit of trial and error, to replicate any superior attributes. The only new engine to be developed during this period was the counterweight trebuchet. The lack of any notice of its sudden appearance and similar absence of any observations that a certain group possessed a particularly impressive artillery tradition suggest that it was employed and developed by various parties as its strengths became apparent. The legendary Greek fire, which gained infamy from the seventh century, had been a particularly fearsome weapon when used at sea. Unlike other siege weapons, its make-up was not immediately identifiable to witnesses. Once a closely guarded secret, a sense of its components had spread with time and a plethora of naphtha-based incendiaries, of varying consistencies, were employed through the twelfth and thirteenth centuries.
Every army would have employed experts and men that we might now call siege engineers; however, there is a risk of placing too much emphasis on the contributions of these individuals, especially in the twelfth century. This was a period when most men would have been able to wield simple carpentry tools with a reasonable degree of comfort, precision and experience, enabling many to provide the labour necessary to build most engines. As time went on, an increasing value was placed on the expertise of those with a proven ability to supervise and design these machines. In the early twelfth century, members of the Frankish baronage are typically found directing the construction of siege towers, while the experience of sailors, accustomed to working with ropes, rigging and long beams, was also clearly valued. If William of Tyre’s account of the siege of Tyre is to be believed, the mysterious Havedic is an early example of an identified expert.
In the thirteenth century, experts are more commonly found among the retainers of significant individuals. In the service of Louis IX was John the Armenian, the king’s artillery man, so trusted and skilled that he travelled to Damascus following the disastrous Egyptian campaign to procure horn and glue for making crossbows. Also serving the French king was Joscelin of Cornant, who built the crusaders’ engines at Manṣūra. Although it is not directly stated, ʿIzz al-Dīn al-Afram appears to have had some expertise coordinating Baybars’ artillery, as he is regularly commended for his conduct overseeing these engines and their use. His experience did not end there, as he was also charged with repairing certain castles and his position appears to have made him responsible for a number of engineering projects, which included the preparing and maintaining of riverbanks. Although ʿIzz al-Dīn al-Afram appears again ahead of the siege of Acre in 1291, once more making arrangements for the sultan’s artillery, the Mamlūks were by this point employing a number of ‘siege experts’, noted among the forces Qalāwūn had assembled in Egypt before moving against Margat in 1285.
* * *
Besiegers had numerous tools at their disposal. Every siege, like every stronghold, was unique and the weapons that attackers employed reflected this. The siege engines of this period were fairly simple, easily appreciated by those with experience of war, allowing knowledge of any developments to spread quickly. Accordingly, the decision not to employ a certain engine at a given siege can rarely be blamed on ignorance; a plethora of factors from urgency to the availability of materials and local topography all played a part. But engines were only one type of siege weapon. A fleet could play an instrumental role, assisting with a blockade and providing manpower, materials and perhaps even expertise. The emotions of those within could also be manipulated, through offers of reward or compensation, threats or trickery. With so many potential weapons, sieges could end in a variety of ways. While falling to any enemy by the sword was typically the worst scenario for defenders, negotiated surrenders could involve a vast array of terms.
It is hard to say that any of the various parties or powers possessed a superior siege tradition. In the lead-up to the First Crusade, Turkish forces had overrun Arab- ʿAbbāsid strongholds, while Fāṭimid forces overcame the Seljuk defenders of Jerusalem in 1098, as the crusaders took Nicaea, Antioch and finally Jerusalem in turn. Technologically, each group from this point onwards had access to a similar set of tools. The designs of the engines employed by various besiegers inevitably varied; however, the basic assortment of technological and diplomatic options was available to all, if they had the desire and resources to employ them. Each ruler or commander used the weapons and tactics best suited to the nature of his army and the innumerable variables relating to the particular circumstances presented by the targeted stronghold and unique context of each siege. All parties were capable of conducting blockades, launching massed frontal attacks, and employing artillery and miners, but the choice to do so depended on situational factors.
When looking at broader trends, the Franks showed a preference for the use of siege towers in the early twelfth century, particularly when attacking approachable coastal strongholds with the assistance of naval elements. Muslim forces often relied more heavily on waves of frontal attacks, permitted by their typically larger manpower reserves. While the Franks regularly engaged Muslim and fellow Frankish strongholds with similar levels of aggression, some Muslim sieges of coreligionists were more tempered. On some occasions, such as Nūr al-Dīn’s ‘sieges’ of Damascus, this reflected a desire to gain the favour, or at least avoid the hostility, of the local population. Some others were little more than shows of strength, launched to extend or reinforce a ruler’s hegemony – in such instances, the neighbouring or rebellious potentate was typically left in command of the besieged stronghold or compensated with one of slightly less significance when a surrender was arranged.
A variety of considerations went into each siege, not least the ultimate objectives of the broader campaign of which the siege was a part. Often these would not only dictate what sites were targeted but also influence the tempo of operations and the tactics that were employed. While each siege was a chess match of sorts between attackers and defenders, there was a larger game afoot that often had a dramatic effect on how specific sieges played out.