Castles around Ascalon.
The Christians perceived that the bold incursions of the enemy showed no signs of ceasing; their forces were constantly renewed, and, like the hydra, they gained increased strength by the death of their citizens. Hence, after long deliberation, our people resolved to erect fortresses around [Ascalon]. These would serve as defences against this monster, which ever increased by the loss of its heads and, as often as it was destroyed, was reborn to our exceeding peril. Within these strongholds forces could be easily assembled which, from their very proximity, could more readily check the enemy’s forays. Such fortresses would also serve as bases from which to make frequent attacks upon the city itself . . .
This experiment convinced the nobles of the realm that by establishing the two strongholds, Bethgibelin and Ibelin, they had made decided progress in checking the audacious raids of the Ascalonites. In large measure through this course the insolence of the latter had been repressed, their attacks lessened, and their projects defeated. Accordingly, it was resolved to build another fortress the following spring. By increasing the number of fortified places, they could harass the people of Ascalon by more extensive attacks and more often cause them terror, attended by sudden danger as of siege . . .
Accordingly, when winter was over and spring approached, the king and his nobles, together with the patriarch and the prelates of the church, well satisfied with the idea, assembled as with one accord at that place. Workmen were called, the people were furnished all necessary materials, and a stronghold of hewn stone, resting on solid foundations, was built. It was adorned with four towers of suitable height. From the top of this there was an unobstructed view as far as the enemy’s city, and it proved to be a most troublesome obstacle and a veritable source of danger to the Ascalonites when they wished to go forth to ravage the country. It was called in the vernacular Blanchegarde . . .
The result was that those who dwelt in the surrounding country began to place great reliance on this castle as well as on the other strongholds, and a great many suburban places grew up around it. Numerous families established themselves there, and tillers of the fields as well. The whole district became much more secure, because the locality was occupied and a more abundant supply of food for the surrounding country was made possible.
(Adapted from William of Tyre 14.22, 15.25, trans. Babcock and Krey, 2:81, 131–2.)
Raids from Ascalon were less frequent by the time Fulk came to the throne in 1131, owing to both issues in Egypt and the increasing strength of the Franks. Major raids aiming to take towns or create more general political instability had given way to minor profit-driven enterprises and small-scale efforts to ensure there remained a zone of Fāṭimid influence beyond the city’s walls. Muslim raiders were forced to target the settlements closer to the blockading Frankish castles rather than the larger wealthier communities beyond, as they risked being surprised by the garrisons of these strongholds on their return, laden with booty. Usāma ibn Munqidh took part in these raids in the early 1150s, attacking the settlements around both Bethgibelin and Ibelin during the four months that he spent at Ascalon. He found himself there as part of a Fāṭimid delegation, which had recently solicited Nūr al-Dīn’s help to disrupt Frankish efforts to fortify Gaza. After Usāma returned to Egypt, his brother was killed in a raid against this newest Frankish castle.
Castles around Ascalon.
Work at Gaza, 19km southwest of Ascalon, had begun in 1149 under the supervision of Baldwin III, the patriarch and a group of nobles. Rather than rebuild the abandoned town’s old and derelict walls, a castle was built. Nothing remains of the Frankish stronghold, but it was probably designed with a quadrangular plan, similar to the others built against Ascalon. When complete, it was given to the Templars, who used it as a base to launch more aggressive raids against the Fāṭimid city and it in turn was attacked soon after its completion and again in 1152.
The fall of Ascalon in 1153 meant that Bethgibelin, Ibelin and Blanchegarde were no longer any more exposed to attack than most other settlements in southern Palestine. This added security allowed the communities around these castles to grow and, by 1160, thirty-two families lived in the settlement outside Bethgibelin, sufficient to encourage the construction of a parish church. South of Ascalon, Gaza also flourished and, by 1170, a town wall was built to surround the community that had developed next to the castle.
Sometime in the 1160s, Amalric commissioned another quadrangular castle at Dārūm, about 14km southwest of Gaza and less than 2km inland from the coast. It is unclear whether the primary motive behind the castle’s construction was to facilitate Frankish actions in Egypt, to obstruct a potential Zankid attack from Egypt, or simply to push the frontier of Frankish control further south. Regardless, a prosperous town quickly grew up around the castle. Perhaps funded by revenues that were collected from traffic passing along the coast between Palestine and Egypt, town defences had been added by 1170, and these may have been strengthened by 1191, when Dārūm was besieged by Richard I during the Third Crusade.
The construction of a stronghold was a way of claiming authority over a region – it was a means through which rival challenges were resisted. While Muslim leaders reigned from their citadels, enforcing their rule through their sizeable armies, the Franks expanded their power and influence through the acquisition and construction of castles. Although common to all Frankish lands, this can be seen quite clearly in the southernmost and northernmost lordships of the kingdom of Jerusalem, and the regions that came under the military orders.
The lordship of Transjordan occupied the southernmost part of the kingdom. It extended from the Gulf of Aqaba up the east side of Wādī ʿAraba as far north as the Nahr al-Zarqāʾ (the Blue River), south of ʿAjlūn and Jerash, its frontiers extending into the deserts to the east and west. Montreal, commissioned by Baldwin I in 1115, was the first castle built in the region. It was originally retained by the monarchy, but appears to have become the seat of a large lordship during the reign of Baldwin II. From about 1142, the castle of Kerak was constructed under Pagan the Butler. Located 74km to the north of Montreal, Kerak became the new practical seat of power, bringing the influence of the lordship further north. While this might indicate a desire to be closer to the court at Jerusalem, it may also speak to the potential wealth and opportunity to be found east of the Dead Sea, rather than further south. With the onset of the contest for Egypt in the 1160s, the importance of the lordship, and its dominance of the desert road between Damascus and Cairo, increased.
Kerak. (Courtesy of APAAME)
Kerak (with topography).
The lordship of Transjordan appears to have been enlarged when it passed to Philip of Milly in 1161. Philip was a proven and loyal knight who traded his rather modest lands around Nablus for the right to inherit the wealthy lordship. When Philip joined the Templars a few years later, the lordship passed to the successive husbands of his daughter, Stephanie. Reynald of Châtillon, the former prince-regent of Antioch and Stephanie’s third husband from 1176, embraced his role as a marcher lord, dominating traffic along the desert road, claiming lordship over the Sinai Peninsula and threatening Arabia to the south. Kerak and Montreal anchored Frankish influence in the region and their strength – the result of successive refortification efforts over the years– was sufficient to resist multiple sieges launched by Saladin through the 1170s and 1180s. Besieged from 1187, both castles held out for more than a year and a half before they finally fell from Frankish hands.
To the north, the principality of Galilee was probably the kingdom’s earliest and most prestigious lordship. It was created by Tancred and recognized by Godfrey in the immediate aftermath of the First Crusade. With its seat of power at Tiberias, the princes (or lords) of Galilee held the lands through which most invasions from Damascus passed as they entered the kingdom. The lords’ authority reached across the Jordan to include areas east of the Sea of Galilee, where the Franks established a presence for periods during the twelfth century. When Tancred went north to assume the regency of Antioch, following Bohemond’s capture in 1100, the principality was bestowed on Hugh of St Omer, who commissioned Toron and possibly also Chastel Neuf. When Hugh died without an heir in 1106, the principality passed to the monarchy. Baldwin I seems to have divided it around this point, detaching the northern part to create a distinct lordship centred on Toron. In so doing, Baldwin not only placed a second strong figure in the region, but ensured that the lordship of Galilee would not turn into a fifth independent Frankish principality – it was clearly a lordship of the kingdom but its lords continued to style themselves as ‘princes’.
The capture of Sidon in 1110 led to the creation of another significant lordship, which was given to Eustace Garnier. During the reign of Fulk, Beaufort was added to this lordship, extending the reach, and interests, of the lords of Sidon inland and providing them with a strongpoint from which traffic moving between the Hula and Biqāʿ Valleys could be observed. As was done elsewhere, territory and strongholds were distributed to a number of strong figures, preventing any one from becoming too powerful but ensuring each was able to hold and defend his lands.
Despite the original ambitions of Hugh of St Omer, when Tyre was captured in 1124, it was not added to the lordship of Toron; the wealthy port was instead retained by the monarchy. In 1140, when Bānyās returned to Frankish hands, Fulk granted it not to the Bures family, which had inherited the principality of Galilee, but rather to Renier Brus, who appears to have held it of the lord of Beirut. When Renier died around the time of the Second Crusade, Bānyās passed to Humphrey II of Toron, who had married Renier’s daughter. This seems to have been the point at which Humphrey also gained Chastel Neuf, which overlooks the north end of the Hula Valley from the west. It is possible that the castle had become, or remained, a part of the lordship of Toron before this point; however, it seems more likely that it came into the hands of Renier Brus, perhaps when he was granted Bānyās, and passed to Humphrey when he died.
The kingdom’s northern lords ruled fairly large regions but probably also depended on wealth gathered from their Muslim neighbours to the east, either through tribute or raiding. Nūr al-Dīn’s presence in the area and eventual acquisition of Damascus would have frustrated this revenue stream, while the greater threat posed by his army drove up the costs of maintaining the lordships’ castles. With the lords unable to finance the defence of these strongholds, many were progressively given to the military orders, which came to replace the secular baronage as the primary guardians of this frontier of the kingdom. The Templars acquired Safed and stood to gain the castle at Jacob’s Ford upon its completion, while the Hospitallers assumed shared ownership of Bānyās and constructed Belvoir to the south. Chastel Neuf and Beaufort remained in secular hands until they were lost following the battle of Hattin, but both would eventually be granted to military orders in the thirteenth century.
In the principality of Antioch, it seems a system of marcher lordships had been established by about 1130, each with its seat of power at a notable stronghold. The same probably took place on a smaller scale in the county of Tripoli and to an even looser degree in the county of Edessa, where Frankish rule was never as centralized nor as ‘Frankish’ as in the principalities to the south. These ‘marcher’ lords were capable of handling small threats, but their resources were insufficient to contend with the growing power of the Zankids on their own – the same forces that led to the collapse of the county of Edessa contributed to the rapid spread of the military orders’ role and influence.
The princes of Antioch and Tripoli more actively sought the assistance of the military orders, by granting them lands and castles, than did the kings of Jerusalem. From 1135, Zankī began taking control of the Syrian plateau around Aleppo, up to the Coastal Mountains, and extending his authority southward until he took Baalbek in 1139. A significant part of the county of Tripoli was taken in 1137, leading Raymond II to donate considerable lands, most of which were now held by Zankī, to the Hospitallers in 1142, promising his dispossessed barons cash compensation. The donation included Rafaniyya, Montferrand and the originally Kurdish-built castle of Ḥiṣn al-Akrād, which the order would rebuild as Crac des Chevaliers. The donation, Ray mond hoped, would essentially outsource the defence of the county’s eastern frontier, while the relative autonomy he gave them, explicitly laid out in their grant, reveals his desperation. The Hospitallers proved to be ideal marcher lords: their resources allowed them to actively raid their Muslim neighbours until the region fell from their hands in 1271, while peace agreements concluded by the counts of Tripoli did not apply to them – the counts could enjoy peace while watching the Hospitallers continue to weaken their Muslim rivals.
In 1152, Nūr al-Dīn led an army through the Homs–Tripoli gap all the way to the coast, where he took Tortosa. Although the town was quickly reacquired, the estimated costs of refortifying it led Raymond II to offer the town to the bishop of Tortosa, who promptly passed it on to the Templars. Raymond’s approval of this grant may have been part of a process of balancing the initial favour shown towards the Hospitallers. In addition to Tortosa, the Templars came to hold Chastel Blanc and Arima, while the Hospitallers, who also held Chastel Rouge (Yaḥmūr), Coliath (al-Qulay‘āt) and ʿAkkār, bought Margat from the Mazoir family in the months leading up to the battle of Hattin.
A similar trend is evident in the northern regions of the principality of Antioch. North of the Antioch Basin, the Templars acquired the castles of Baghrās and Trapessac, on the east side of the Belen Pass, and La Roche de Roussel (Chilvan Kale), deeper in the Amanus Mountains overlooking the secondary Hajar Shuglan Pass to the north. These strongholds gave the Templars conspicuous influence over the routes between Antioch and Cilicia. Further north, Baldwin of Marash appears to have solicited the support of the Hospitallers in the early 1140s, granting them a certain place provided they fortify it within a year; unfortunately, this place has not yet been identified.
Like the early blockading strongholds, frontier castles played an important role in both launching raids and intercepting hostile raiders. Strongholds that were captured, and could be held, deep in an opponent’s territory were particularly useful in this capacity, as played out in the early twelfth century between Aleppo and Antioch. Following the battle of Hattin, certain captured Frankish castles became important Muslim frontier bases. Toron, for example, once more opposed a hostile Tyre: in July 1189, the garrison that Saladin had established there two years earlier attacked a group of Franks who had left Tyre to gather provisions. Almost eighty years later, in October 1266, the garrison of Safed, which Baybars had installed after capturing the castle only a few months earlier, intercepted and defeated a small Frankish force that had set out on a retaliatory raid under Hugh of Lusignan, the future king of both Cyprus and Jerusalem.
As a means of securing territory, fortifications were a critical part of conquest. When a ruler sought to expand his realm, he had to either capture or construct a stronghold, or field an army capable of removing any such obstacles that might be built by an opponent to challenge his authority. It is in the area around Palestine, a region with comparatively few significant strongholds, but detailed sources, that the use of castles as tools of expansion may be seen most clearly.
Almost immediately after the First Crusade’s capture of Jerusalem, certain Franks set their sights on the lands east of the Jordan. Godfrey had joined Tancred at Tiberias and together they raided this region in 1100, exacting a tribute before withdrawing back across the river. The Golan Heights and the westernmost part of the Ḥawrān, known as the Sawād, were a particularly attractive region. Although the area has few trees, its volcanic soil is quite fertile, fed by numerous springs. The area was regularly raided by Frankish forces through the twelfth century, but gaining a permanent presence proved difficult.
In an early effort to annex the Golan, the Franks attempted to build a castle, which Ibn al-Qalānisī called [Al‘āl, east of the Sea of Galilee in 1105. Construction, however, was interrupted by Ṭughtakīn of Damascus, who arrived and defeated the Frankish force overseeing the castle’s construction. This may have been the episode in which Hugh of St Omer, prince of Galilee, was killed. The engagement, which probably took place in 1106, may appear twice in Ibn al-Qalānisī’s account. If so, a combination of luck and planning assisted the Muslims, as Baldwin I, who was then at Tiberias, was forced to confront a party of Tyrian raiders, who had invaded from the west and sacked the town outside the castle of Toron. Toron was one of Hugh’s principal strongholds, but as the new castle beyond the Jordan would also become a part of his lordship, Baldwin left to address the raid, leaving Hugh to oversee activities east of the Jordan. With the Frankish army divided, Ṭughtakīn, who had assembled his army in the Sawād, was able to defeat Hugh and destroy the incomplete castle.
Baldwin I made another attempt to establish a fortified outpost east of the Jordan a few years later, developing a cave in a sheer section of wall along the canyon carved by the Yarmūk River. Known as al-Ḥabis Jaldak (Cava de Suet), the Frankish outpost was taken by Ṭughtakīn near the end of 1111 or start of 1112, while the army of Jerusalem was busy besieging Tyre. The stronghold was taken by the sword and the entire garrison executed. Around the time Joscelin of Courtenay became prince of Galilee, Ibn al-Qalānisī suggests that Baldwin I proposed Joscelin trade certain lands to Ṭughtakīn in return for al-Ḥabis and half of the Sawād – the implication being that these were the lands to be administered from the castle. Ṭughtakīn, however, declined. The cave castle returned to Frankish hands around the time of Baldwin I’s death in 1118 and Frankish influence once more stretched into the Ḥawrān. A few years later, Baldwin II briefly considered occupying and garrisoning Jerash, 48km to the south of al-Ḥabis Jaldak. He had taken a stronghold there as part of a retaliatory raid in 1121, but decided it would be too isolated and expensive to defend.
Around 1139, Thierry of Flanders, who was then visiting the Holy Land, took part in an expedition that captured a cave stronghold east of the Jordan. It is unclear if this was al-Ḥabis Jaldak, but the stronghold had returned to Frankish hands by 1158, when Thierry, once more in the Levant, accompanied Baldwin III to relieve the castle. The Franks successfully interrupted Nūr al-Dīn’s siege, forcing him to withdraw, although Shīrkūh brought the cave castle back under Muslim control in 1165. It was back in Frankish hands by 1182, when a contingent of Saladin’s army captured the stronghold following a broader invasion of northern Palestine. According to William of Tyre, the castle fell after a siege of five days; either its defenders capitulated, after being bribed, or it was taken by force once Muslim sappers mined into the first level from the side. Later the same year, the Franks retook the cave castle on their return from a raid towards Damascus, taking advantage of Saladin’s absence in the Jazīra. The Franks mined down towards the cavern from the top until, after about three weeks, the garrison sued for peace and were granted free passage to Bosra without their arms.
One of the more famous, if least impressive, castles built by the Franks was that at Jacob’s Ford. In 1178, Baldwin IV commissioned the castle, alternatively known as Chastellet or Qaṣr al-ʿAtra, at the southern crossing over the northern stretch of the Jordan, between Bānyās and the Sea of Galilee. It was probably hoped that this would be a base from which the Templars could exert influence east of the Jordan and raid the environs of Damascus, 84km to the northeast. When Baldwin refused a large sum, perhaps 100,000 dinars, to cease construction, Saladin attacked the still-incomplete castle in August 1179, taking it before the relief force mustering at Tiberias was ready.
Jacob’s Ford, northeastern corner of the castle and the river to the right. (Michael Fulton)
Crac des Chevaliers and Jacob’s Ford in the twelfth century.
Later events have coloured many interpretations of this stronghold. There is little evidence that Jacob’s Ford was originally planned as a much larger concentric castle. Instead, its design was fairly typical of its time: an elongated enclosure, following the shape of the hill, with a single tower. It is precisely because certain other castles that shared a similar plan in the late twelfth century, notably Crac, became much grander structures that there seems to be a desire for Jacob’s Ford to have been something greater than it was. Unlike Crac or Belvoir, Jacob’s Ford did not occupy a particularly strong position topographically: it sat on a small hill overlooked by the surrounding landscape on all sides. Despite certain suggestions to the contrary, the destruction of the castle was not the opening of a floodgate that precipitated the fall of the kingdom of Jerusalem. Aside from a raid a few months later, Saladin’s forces opted to cross the Jordan to the south of the Sea of Galilee during their subsequent incursions into northern Palestine – although Jacob’s Ford was unobstructed, Saladin evidently had little desire, or need, to use the crossing.
Built from about 1168, Belvoir had by far the most complex and developed plan of any single-phase castle built in the kingdom during the twelfth century. Despite the castle’s strength, its garrison could do little as Saladin’s forces repeatedly invaded the kingdom not far from its walls through the 1180s. Its presence, however, prevented Saladin from claiming and attempting to rule the surrounding region. Those in the castle were left to watch as Saladin’s forces passed by once more in 1187, ahead of the battle of Hattin; however, they would exploit Saladin’s decision to leave the castle to his rear following the battle. Although subject to a loose blockade, the garrison launched a raid in early January 1188, seizing two caravans, one of which contained arms and provisions from the captured Templar stronghold of La Fève.
Owing to the protection they provided and their geographical positioning, strongholds were often used as staging points for campaigns. The largest armies usually collected outside cities, including Cairo, Damascus and Acre, where there were markets large enough to support them. Moving out, they might pause at a secondary stronghold to wait for further reinforcements or to determine what the enemy was doing. When invading Galilee or the Jezreel Valley, Damascene forces often paused at Bānyās, the only significant town between Damascus and northern Palestine. This is where Ṭughtakīn based his efforts to disrupt the Frankish sieges of Tyre in 1111–12 and 1124, and Saladin similarly used Bānyās as his staging point in 1179 when moving against Jacob’s Ford. During Baybars’ campaign against Beaufort in 1268, he had Jamāl al-Dīn al-Najībī, his governor of Syria (nāʾib al-Shām), take the Damascene element of the Mamlūk army to Bānyās, from where it then proceeded to Beaufort.
When countering invasions into northern Palestine, the army of Jerusalem typically positioned itself at Ṣaffūriyya, where a tower dominated the community and nearby springs. It was the availability of water, the site’s central location and the road networks leading away from the settlement, rather than the security provided by the tower, that made this an attractive position. When Ayyūbid forces assembled in the region in the thirteenth century, they preferred Nazareth, 6km to the south, which had more to offer their larger armies. When countering Fāṭimid invasions into southern Palestine in the early twelfth century, the Franks often assembled at Ramla, which, like Ṣaffūriyya, was defended by a tower, but was more importantly in a central position at a significant crossroads. Fāṭimid incursions from Egypt were naturally launched from Ascalon. More than most other sites, the defences of Ascalon were strategically important, given the stronghold’s isolation and distance from support. While large armies often arrived after marching overland, smaller raids could be conducted by forces that arrived by sea. The defensibility of the site ensured it remained an effective Fāṭimid beachhead until 1153. Later in the thirteenth century, stronger Ayyūbid and Mamlūk forces moving into Palestine from Egypt preferred to pause at Gaza, which, like Nazareth, had no significant defences in the thirteenth century, but the larger community could better cater to these sizeable armies.
In 1137–38, the Byzantine emperor John Comnenus used Antioch as a staging point for his campaigns in Syria. His pretext for demanding free entrance into the city was that it was necessary to store his siege engines there over the winter. Intending to besiege Aleppo the following year, he claimed it was too far to store them in one of the towns of Cilicia. The emperor was effectively using the city’s value as a staging point to reinforce his claim to suzerainty over it and the remainder of the Frankish principality.
There were a number of ways in which strongholds could contribute to a lord’s income. In addition to supporting his efforts to collect a share of the produce from the surrounding region, they might assist with the launching of raids and the exaction of tributes, taxes and tolls. The wealth accrued in these ways helped in turn to finance the upkeep, development and defence of these structures, as well as the costs of campaigning and the numerous other expenses incurred by the nobility and elites.
Raids were perhaps the most important, or at least most common, form of warfare during this period. Not only did they provide the figure who orchestrated them, as well as those who took part, with wealth (agricultural produce, animals, captives, precious metals, etc.), they impoverished the targeted adversary, inhibiting his ability to both defend his lands and launch similar raids. Particularly successful or regular raiding could often lead to an advantageous truce, in which the victimized party would agree to pay a tribute or share certain lands in exchange for peace. Strongholds also provided positions to which raiders could withdraw, as both defeated raiding forces and those weighed down by large quantities of booty were vulnerable.
It was not uncommon in Latin Europe at the time for lords to raid their neighbours, demonstrating their authority, extending their influence and acquiring wealth; however, this was extremely rare among the Latin lords of the Levant. Aside from the brief periods of open war between the princes of Antioch and Edessa in the early twelfth century, which were part of a broader network of alliances and rivalries, raids conducted by the Frankish baronage appear to have been limited to Muslim lands. This focus is rather exceptional when compared with the political landscape these men, or their ancestors, hailed from. Hostilities among local Muslim powers were more common, although these conflicts usually revolved around gaining territory or influence, often by acquiring strongholds – raiding between Muslim powers west of the Euphrates seems to have been almost as rare as it was between Christians. Most of the raids we know about were led by the Frankish princes or powerful lords, including those of Turbessel, Galilee and Transjordan, and their leading Muslim counterparts.
Raiding was an obvious way of acquiring immediate wealth, but so too could it be a powerful threat. The arrival of the First Crusade in Palestine led most towns along the coast to placate the crusaders by selling them provisions. In the decade that followed, many would offer an annual tribute to the kingdom of Jerusalem in exchange for peace. In 1285, almost two centuries later, Bohemond VII of Tripoli (and nominal prince of Antioch, the city having been lost in 1268) agreed to destroy the island stronghold of Maraclea when prompted to do so by Mamlūk Sultan Qalāwūn, who threatened to raid the lands around Tripoli if he did not.
From a defender’s perspective, offering tribute was a way of avoiding potentially greater costs that might accompany raids, or a way of diverting a direct attack. For attackers, accepting tribute avoided the uncertainty and risks that were associated with raiding, or the costs and challenges that accompanied besieging a stronghold and then attempting to administer the surrounding region. But such a system of tributes, truces and taxes relied on an acknowledged network of regional borders – it had to be clear which areas would contribute to the tribute and would in turn be protected from further hostilities. A landscape of acknowledged landholding limits like this would have been foreign to most parts of Europe at the time.
Contemporary European boundaries were relatively fluid and frontiers shifted as the authority of one baron spread at the expense of his neighbour’s. Perhaps owing to the higher population density and the continued use of many ancient administrative boundaries, borders were fairly well defined and agreed upon in the Levant. This did not inhibit frontier warfare and raiding, or otherwise imply that these borders were obstacles (they did not limit the movement of people or armies), only that it was widely appreciated where a certain regional power’s authority, or that of a smaller landowner, ended and another’s began. Likewise, borders could be shifted if necessary. For example, the archbishopric of Tyre traditionally fell under the patriarchate of Antioch; however, when Tyre became a part of the kingdom of Jerusalem, rather than the principality of Antioch or county of Tripoli, the archbishopric was transferred to the patriarchate of Jerusalem. The archbishopric had clearly defined and agreed upon limits, making its transfer, rather than the area of its extent, the matter of dispute. This episode, which was ultimately adjudicated by Pope Innocent II, speaks to the Franks’ appreciation of the region’s existing boundaries, even as relative new comers to the area.
When passing from Damascus to Acre in 1183, the Andalusian pilgrim and traveller, Ibn Jubayr, noticed that there was a certain tree along the road that served as a border marker. He states that Frankish robbers would take captive anyone they found east of the tree but would set loose all those they found to the west of it. Although the tree was east of Bānyās, which had been held by a Muslim garrison since 1164, the Franks still lay claim to this region. Regardless of where the Muslims considered the border to have been, all acknowledged that the caravan was in Frankish territory by the time it crossed the Hula Valley, at which point it was compelled to take a detour to Toron, where each traveller was to pay one dinar and a twentieth. The caravan then set out for Acre, where the merchants would pay their customs dues.
The Sultan, al-Malik al-Ẓāhir Baybars, and a number of emirs sent a vast amount of barley and flour by se˙a from Damietta to Jaffa, which was in the hands of the Franks. When the Sultan went towards Syria, John of Ibelin, count of Jaffa, sent a messenger presenting his obedience and bringing gifts . . . When Baybars reached Damascus, a messenger came from the ruler of Acre asking for safe conduct for envoys from all the Frankish houses. The sultan wrote to the governor of Bānyās telling him to allow them to come, after which the Frankish leaders arrived and asked for peace. The sultan made many demands of them, and when they refused, he upbraided them and treated them with contempt. Muslim forces had set out to raid their lands from Baalbek, and the Franks asked that this force should be withdrawn. It happened that prices were high in Syria and the bulk of Muslim imports were coming from Frankish lands. So peace was agreed on the terms of theIbn al-Furāt: raiding and peace negotiations, 1261 status quo at the end of the days of al-Malik al-Nāṣir . . . In the same way, a truce was agreed with the lord of Jaffa and the ruler of Beirut . . .
(Adapted from Ibn al-Furāt, trans. Lyons and Lyons, 2:43–4.)
Because regional boundaries were relatively agreed upon, land typically changed hands when a power centre was taken. Rarely was part of a region alienated and absorbed by another; when it was, it was often done quite formally, by treaty rather than force. Unlike in Europe, where lordships often grew organically through force and intimidation, this was not the norm in the Levant. With more powerful figures in closer quarters, it was often wise to acknowledge boundaries, whether they were respected or not. The Franks’ efforts in the early twelfth century to extend their influence into the Sawād through the construction of castles, as discussed above, was a relatively unique exception. The Franks recognized that they would need a stronghold to control this region, but pausing to build a castle in the lands of an opponent whose military structure was based on significant and mobile field forces was a risky endeavour, and it cost Hugh of St Omer his life.
If too weak to besiege a neighbouring power centre directly, the threat of doing so, or of conducting repeated raids into its lands, might result in an offer of tribute. This quickly became a formalized practice, as a way of expressing dominance and extracting wealth. In 1109, following the Franks’ capture of Tripoli, Ṭughtakīn of Damascus agreed to give up the castle of ʿAkkār and pay an annual tribute, which was to be gathered from the castles in the Syrian Coastal Mountains west of Homs, including Maṣyāf and Ḥiṣn al-Akrād (Crac des Chevaliers). In 1110, building on his success after taking Beirut, Baldwin I compelled Sidon to increase its annual tribute from 2,000 dinars to 6,000. In the same year, Tancred launched a raid against Shayzar, leading the city to agree to an immediate tribute of 10,000 dinars, a gift of horses and the release of prisoners in exchange for peace. When Tancred died in 1112, Roger of Antioch demanded that Aleppo and Shayzar continue to make their annual tribute payments of 20,000 and 10,000 dinars respectively.
Instability among the Muslims also played into Frankish hands. In 1117, the interim leader of Aleppo, fearing the approach of Īlghāzī, bought peace by surrendering the castle of al-Qubba to the Franks. By 1123, however, the balance of power had shifted and it was the Franks who were forced to give up al-Athārib in exchange for a truce. Issues in Damascus led the Būrids to pay the kingdom of Jerusalem an annual tribute in the mid-twelfth century. In 1149, a series of back-and-forth raids resulted in a two-year truce and tribute was once more to be paid to the Franks. When the kingdom’s army moved to support the Damascenes against Nūr al-Dīn in 1151, the Franks made sure to collect their tribute before they withdrew. Even after taking Damascus, Nūr al-Dīn was inclined to buy peace with the Franks, which he did in 1155, in order to campaign in Anatolia. The following year, he arranged another truce with Jerusalem, which cost him 8,000 Tyrian dinars. Although substantial, these tributes may have been, to a degree, compensation paid by the Damascenes to the kingdom for acknowledgement of their right to the lands east of the Jordan, which the Franks had previously shared and at times still claimed.
Sometimes, rather than concluding an agreement of tribute, it was arranged for the lands of a certain region to be shared in condominium (munāṣafa). Whereas dividing territory might encourage further raiding, this was a meansof promoting peace by sharing the produce and administration of an area. With both sides taking part of the total revenue, however, some of the highest tax rates are recorded for these areas.
In 1108, the first of these arrangements was concluded between the kingdom of Jerusalem and the emirate of Damascus. In return for a truce, the Franks were to receive two-thirds of the produce from the Sawād and Jabal ʿAwf, roughly the region from the Golan Heights to the mountains of ʿAjlun. In 1109, a similar agreement saw the Franks receive one-third of the produce of the Biqāʿ. It may have been at this point that the Franks were also given ʿAkkār and Munayṭira, in return for not raiding the lands of Damascus, Maṣyāf, Ḥiṣn al-Akrād and Ḥiṣn al-Tūfān, from which a tribute would also be paid.The agreement was either confirmed later the same year or renewed the following year. In 1111, Baldwin I and Ṭughtakīn renegotiated their agreement and it was concluded they would each receive a half of the Sawād, Jabal ʿAwf and al-Jabāniyya. Although these arrangements were less common or infrequently documented in the north, the Franks were compelled to share Jabal al- Summāq (the mountainous region southeast of Ḥārim) with al-Bursuqī at one point in the 1120s. In this instance, the Franks, who retained most of the administration of the area, reportedly caused problems for the Muslim revenue collectors.
Following the Third Crusade, it was agreed that Sidon, Ramla and certain other regions would be shared. Al-ʿĀdil appears to have given up his rights to these places, along with Nazareth, in a subsequent peace that was concluded when elements of the Fourth Crusade began to land at Acre. According to al-Maqrīzī, half of Sidon still belonged to the Muslims in 1228, when the Ayyūbid administrators were thrown out by crusaders who had come to the Levant with Frederick II.
In the early thirteenth century, the Hospitallers, based in their castles in the northern part of the county of Tripoli and southern part of the principality of Antioch, found themselves in a particularly strong position. Having been bypassed by Saladin in 1188, they were able to launch pestering raids over the following decades into the lands of Homs and Hama, the rulers of which were often preoccupied with the greater threats posed by their fellow Ayyūbids. These Hospitaller castles also surrounded the Assassin enclave in the southern section of the Syrian Coastal Mountains. The order was probably extracting a tribute from the Assassins in the second half of the twelfth century and Wilbrand of Oldenburg valued this tribute at 2,000 marks when he visited in 1211. The burden was such that the Assassin legation sent to Louis IX at Acre, following the collapse of the Seventh Crusade, was willing to waive the customary tribute they demanded of rulers if Louis could release them from the payments they owed to the Templars and Hospitallers. The appeal failed and the Assassins found themselves the ones sending the king a tribute following the intervention of the orders, such was their influence.
With the rise of the Mamlūks, roles were reversed. After Baybars captured the Templar stronghold of Safed in 1266, the Hospitallers scrambled to extend their peace that was in place in the region of Homs. Baybars agreed to this provided the Hospitallers relinquish the tributes they were receiving from Hama (4,000 dinars), Abū Qubays (800 dinars) and the Nizārī Assassins (1,200 dinars), and pay further amounts of both wheat and barley. The treaty was ratified the following year, after Baybars had destroyed the order’s mill outside Acre. The truce was to last for ten years, ten months and ten days and the final version saw the Hospitallers give up not only their tributes from Hama and the Assassins, but also Shayzar, Apamea and elsewhere. As it turned out, the Assassins simply paid Baybars the same tribute they had once paid the Hospitallers. This was not an insignificant development, as the Assassins had previously exacted protection money from the distant Ayyūbid sultans of Egypt.
In addition to tributes, Baybars’ treaties often included condominia. In his 1266–67 treaty with the Hospitallers, the region from the Orontes to a certain western boundary was to be held in condominium, as were the order’s lands around Acre. Around the same time, Baybars negotiated a truce with Acre, which would have converted Sidon into a condominium once more had it been ratified. Another, negotiated in 1268 but also never ratified, would have turned considerable parts of the territory of Haifa, the Carmel and ʿAtlit into condominia, while Sidon was to pass entirely to the Mamlūks. In 1271, the Hospitallers gave up their rights to previous condominia, while the lands around the castle of Margat were to be ruled as such. A later treaty between Qalāwūn and Lady Margaret of Tyre saw the ninety-three settlements of the lordship of Tyre divided between the two parties: five to Qalāwūn, ten to Margaret, and the remainder to be shared. This agreement seems to have been negotiated in 1271 between Baybars and John of Montfort, Margaret’s husband, but it was not ratified until 1285, after both Baybars and John had died.
Besides controlling land, and the revenue drawn from it, fortifications could also dominate trade and impose tolls. While the Templars customarily built their towers and smaller strongholds along roads and Christian pilgrim routes, Roger of Antioch had evidently discovered how profitable the revenue from the hajj (the annual Muslim pilgrimage to Mecca) could be. In 1117, a clause in a treaty with Aleppo granted him not only a certain stronghold and tribute, but the right to collect the customary tax levied on those making the pilgrimage from Aleppo to Arabia. With an eye to similar revenue streams, Baldwin I had built the castle of Montreal two years earlier along the main caravan (and hajj) route through Transjordan. In the words of Albert of Aachen, ‘he established a new fortress, so that in this way he might more powerfully subdue the land of the Arabians, and passage to and fro would no longer be available to merchants except by the king’s favour and licence, nor would any ambushes or enemy forces suddenly appear, but would quickly be apparent to the king’s faithful stationed in the citadel’.18 Prior to this, the Franks had prevented a Turkish emir from fortifying Wādī Mūsā (near Petra) in 1107, and preyed on caravan traffic in the region in 1108 and again in 1112 or 1113. Accordingly, they would have been well aware of the profit to be gained by controlling this region.
For powerful Muslim rulers, whose strength lay in the size of their large armies, exposed strongholds could be a liability. Rulers such as Īlghāzī, al-Bursuqī and Zankī, who brought forces from the Jazīra and made considerable use of seasonal Turkoman forces, showed a willingness to slight the fortifications of some smaller towns they captured in western Syria. The defences of more significant urban centres were typically retained and even developed, as these were often more defensible and protected wealthier communities. Nūr al-Dīn continued this policy to a degree even as his power grew. For example, when he took Chastel Neuf in 1167, he decided to destroy it, rather than attempt to hold it, much as his father had opted to do after capturing Arima in 1142. Both were relatively deep in Frankish territory and it would have been assumed that they would have come under immediate and concerted pressure after Muslim forces withdrew. This willingness to destroy fortifications, in order to deprive the Franks of them, was demonstrated by Saladin during the Third Crusade and by his nephews during the Fifth Crusade.
In 1190, as the siege of Acre dragged on and increasing numbers of European crusaders arrived in the Levant, Saladin issued orders to slight many of the fortifications he had captured along the coast, reissuing these orders in 1191 following Acre’s capture. If left intact, these towns could serve as fortified bridgeheads for future crusaders. Less than three decades later, with large numbers of crusaders in Egypt, al-Muʿaẓẓam ʿĪsā similarly recognized that the Franks’ ability to hold inland territory was based on their possession of strongholds. Fearing that his brother, al-Kāmil, might offer the crusaders Palestine in exchange for leaving Egypt, which he did, or that the Franks might attempt to recapture the Holy Land, he ordered an extensive slighting programme that included almost all castles in Palestine.
The group most opposed to the fortification of the Levant, however, may have been the Mongols. Relying on large armies of semi-nomadic fighters raised in Central Asia, strongholds were regarded as bases of potential resistance. When they invaded the Levant at the start of 1260, the Mongols pursued a policy of supporting local allies, who recognized their suzerainty, while destroying all strongholds held by those who had not submitted with sufficient zeal. Where there was resistance, as at Aleppo, . Ṣubayba, ʿAjlun, Baalbek and Damascus, destruction followed. Hülagü may have intended to extend this destruction to the Assassins: in a letter sent to Louis IX of France in 1262, he asserted that he had charged Kitbugha with reducing their strongholds before leaving the region in 1260.
Strongholds slighted by Saladin (1190–91) and al-Muʿaẓẓam ʿĪsā (1218–19)
As the Mamlūks began to conquer Frankish lands in the 1260s, they instituted a policy of slighting coastal fortifications while retaining strategic inland castles. Like Saladin, Baybars and his successors recognized the threat posed by coastal strongholds and the protection they might offer European crusaders should they be captured. As defences were torn down, whole communities were disrupted, devastating certain industries and trade networks. Inland, large castles, such as Safed and Crac des Chevaliers, were developed and used as administrative centres, while the strategic value of others, such as Beaufort and Margat, ensured they too were preserved as secondary power bases. Both Baybars and Qalāwūn were aware of the benefits strongholds provided opponents in the region: al-Mughīth ‘Umar, Baybars’ final Ayyūbid opponent, was based in Kerak, while Qalāwūn was forced to confront Sunqur al-Ashqar, a rival to the Mamlūk throne, who held Saone. Both dissidents were overcome but neither stronghold was destroyed, as each was the administrative centre for the surrounding region.
Strongholds preserved and slighted by the Mamlūks, 1260–91.
In 1270, Baybars was alarmed by reports that Louis IX and other leading European nobles were sailing eastward. He had bridges built and the slighted defences of Ascalon reduced further, arranging for the stones of the stronghold to be thrown into the sea. The feared attack on Egypt never materialized, as the Eighth Crusade ended in disaster outside Tunis with Louis IX among the dead.
Back in 1250, Louis had been a prisoner in the Egyptian camp when al-Muʿaẓẓam Tūrānshāh was murdered by a group of mamlūks, including Baybars, initiating the ascent of the regime that followed. Fear of the French king’s return had led Baybars to take preparatory measures shortly after assuming power in Cairo: he saw to the defences of Alexandria; ordered the mouth of the Nile narrowed at Damietta, to prevent any large ships from sailing up the river; and arranged for the construction of a watch tower at Rosetta, where the other main branch of the Nile emptied into the Mediterranean. In the years following, reports of planned crusades led Baybars to ensure that the Alexandrian fleet was in a good state of repair and that the city was provided with sufficient artillery around its walls, reportedly numbering a hundred engines.
While a crusade was feared in Egypt, the Mongols were instead the greatest threat to Mamlūk rule in Syria. Accordingly, while coastal fortifications were destroyed, Baybars rebuilt and developed not only the interior castles that he took from the Franks, but also those that the Mongols had slighted in 1260. It was hoped that these strongholds would counteract the numerical strength of the Mongols, tying up an invading army until Egyptian forces could be assembled and moved into Syria.
As William of Tyre remarked in the twelfth century, it did not matter how strong defences were if there was no one to defend them. Garrisons were often composed of professional fighters, men whose social rank or occupational choice made war their profession. Because sieges were relatively rare, these experienced fighters were often relied upon to enforce local authority, take part in raids and even serve in the army at times.
When the Mongols (may God defeat them) occupied Syria, they began to destroy the castles and walls. They demolished the walls of the citadel of Damascus, and the castles of Salt, ʿAjlun, Ṣarkhad, Bosra, Baalbek, Ṣubayba, Shayzar and Shmemis. When the sultan [Baybars] took charge of affairs and God established him as the support of the Faith, he took an interest in the reconstruction of these castles and the completion of the destroyed buildings, because these were the strongholds of Islam. All these were repaired during his time; their fosses were cleared out, the flanks of their walls were broadened, equipment was transported to them, and he sent mamluks and soldiers to [garrison] them.
(Adapted from Ibn ʿAbd al-Ẓāhir, trans. Sadaque, pp. 117–18.)
A Frankish castle or citadel was ultimately the possession of a lord, who might rule from it directly or delegate command of the stronghold to a castellan, or viscount if it were a crown possession. This individual often received the support of a group of retainers, a body of knights and sergeants drawn from the gentry, an entourage of capable fighting men who were supported for their services. Depending on the size of the stronghold, and various other factors, this well born body of fighters might make up 10–25 per cent of the total number of defenders, far fewer if townspeople were conscripted into the defence of a castle or town during a period of crisis. The remainder typically consisted of a semi-permanent fighting force, which had various other responsibilities, and perhaps a body of paid mercenaries, who fought purely for cash rather than any socio-political motivation.
As Frankish castles grew in the thirteenth century, mercenaries were increasingly used to defend them, funded from the same deep coffers with which the military orders had financed their great construction projects. Companies of crossbowmen were particularly valuable to the defence of strongholds. Although the number of mercenaries paid to provide garrison services probably increased in the thirteenth century, using paid soldiers in this way would not have been uncommon in the twelfth – mercenaries were among the Frankish defenders of Bānyās when it was taken in 1132.
Muslim defensive arrangements were similar. The smaller number of strongholds typically meant that, under the Seljuks and Ayyūbids, the ruler of an iqṭāʿ, or occasionally a deputy, was in command of a fairly professional body of defenders, with whose support rule was enforced. During the Mamlūk period, with more centralized rule, strongholds were entrusted to administrative governors, while a distinct garrison commander was also appointed by the sultan. Many of these garrison forces would have been mamlūks, raised from childhood to be soldiers.
In times of need, everyone inside a stronghold might be pulled into service – the consequences of defeat were death or slavery if the defences fell by force. Accordingly, it was typically in the interest of non-military figures, such as townspeople, to assist the garrison. In some rare instances, where a considerable portion of the general population supported the cause of the besiegers, assistance was minimal.
The environment of desperation that might accompany a siege provided an opportunity for women, generally excluded from both Frankish and Muslim military systems, to take a more active role, even if they were rarely involved in direct combat. Women played an instrumental role in retaking Shayzar following the Assassins’ brief seizure of the citadel in 1109. The ruling Banū Munqidh had left to attend a festival hosted by the local Christians, permitting a group of Assassins to sneak into the citadel and surprise its remaining guards. Women in the barbican helped pull men up into the stronghold, allowing the Assassins in the citadel and town to be hunted down and slaughtered. During the Fāṭimid attack on Jaffa in May 1123, a portion of the garrison may have been absent, having gone north with Baldwin II, as the defenders are described as being few in number. However, ‘the women of Jaffa were constantly ready with generous aid for the citizens who were struggling mightily. Some supplied stones, and others water to drink.’19 Fifteen years later, in April 1138, an unnamed woman commanded Buzāʿa when it was besieged by Byzantine and Frankish forces.
In a less direct capacity, women played an important part in ensuring dynastic stability, which limited the vulnerability of strongholds in this period when male rulers often died young. Melisende, daughter of Baldwin II and widow of Fulk, served as regent for her son Baldwin III during his minority, fighting to retain her position even after her son came of age. Melisende’s sister, Alice of Jerusalem, similarly sought to control the regency of her daughter, Constance of Antioch, following the death of her husband, Bohemond II. Like her mother, Constance would fight to assert her own rule later in life. Although all three were strong women and had an impact on the governance of their principality, the conduct of military matters, including sieges, remained the preserve of male councillors and deputies, as it was in Muslim spheres.
The marriage of Ḍayfa Khātūn, daughter of al-ʿĀdil, to al-Ẓāhir Ghāzī in 1212 ended the first phase of intra-Ayyūbid conflict. Her husband died in 1216 and their son followed him to the grave in 1236, at which point she assumed the regency of her grandson, al-Nāṣir Yūsuf, ruling through a council of men. Military actions were still undertaken by male figures, but Ḍayfa Khātū n was an influential figure in guiding policy. Although she initially sided with al-Ashraf against their brother al-Kāmil, she kept Aleppo relatively aloof of the continuing Ayyūbid power struggle following their deaths in 1236 and 1238 respectively.
Town and castle garrisons were an important source of manpower for field armies. It was in the interest of rulers to retain experienced and capable fighting men, many of whom were used to garrison their strongholds when not otherwise needed. It was natural for a lord or muqṭaʿ to summon these forces when joining his prince’s army, but doing so came with risks.
When Frankish Bānyās was besieged in December 1132, the town and citadel were captured after only a few days. As had been the case at Jaffa in 1123, the defenders of Bānyās are noted as being unprepared and without enough men, suggesting a portion of the garrison had gone north with Fulk to address the regency of Antioch. By comparison, the town resisted for around four weeks when besieged by Frankish and Damascene forces in 1140, but only for about a week when attacked by Nūr al-Dīn in 1164, when a number of its defenders were probably in Egypt with Amalric and the army of Jerusalem. The size of these forces that joined the army, or the proportion of Bānyās’s defenders that they accounted for, is unclear. Later, on 30 September 1183, a force from Transjordan, almost certainly members of the garrisons of Kerak and Montreal, encountered a Muslim scouting party while heading north to join the army of Jerusalem at Ṣaffūriyya. If Bahāʾ al-Dīn is to be trusted, this was not a small force, as a hundred of the Franks found themselves prisoners after the engagement.
When garrison forces were similarly drawn into the kingdom’s army in 1187, ahead of the battle of Hattin, the result was disastrous. As Smail remarked, ‘The existence of Christian Syria depended ultimately on the simultaneous existence of an adequate field army and garrisons.’20 It was reasonable to weaken the garrison of a strong castle or town provided the threat posed by the larger field army was sufficient to discourage or break a siege; however, this tactic relied entirely on the strength and survival of that army. This appears to have been impressed upon the Franks in 1179: defeated in the field, they were unable to muster sufficient numbers in time to relieve Jacob’s Ford when Saladin invaded again months later. Over the following years, as demonstrated during Saladin’s invasion of northern Palestine in 1183, the Franks declined to risk battle unless it was on their terms. With the army of Jerusalem intact, Saladin was unable to invest a significant stronghold, from which he could then make an attempt to hold the surrounding region. Although this was a prudent policy, it was not a popular one among certain influential figures.
In 1187, under mounting criticism and with the Frankish ruling elite bitterly divided, Guy of Lusignan made a series of decisions that led to the battle of Hattin. The resulting defeat decimated not only the Frankish army, and any potential relief force, but also significant components of many garrisons, contributing to the loss of many strongholds and the near-collapse of the kingdom of Jerusalem during the months that followed.
It is difficult to determine the size of most garrisons, let alone the total number of defenders during a given siege. Aside from the challenges associated with estimating the number of people in any large group, some sources acknowledged the tendency of contemporaries to exaggerate. For instance, Ibn al-Athīr explicitly calls out ʿImād al-Dīn al-Isfahānī, Saladin’s secretary and biographer, for misrepresenting Saladin’s victory in 1176 over Sayf al-Dīn II of Mosul at Tell al-Sultān, about 45km south of Aleppo. Whereas ʿImād al-Dīn claims that Saladin’s 6,000 cavalry triumphed over a force of 20,000 Mosuli cavalry, Ibn al-Athīr, who was in a position to review the Mosuli register, composed by no less a figure than his own brother, clarifies that the Mosuli force was instead between 6,000 and 6,500.
In the absence of detailed registers and financial records, we are left to rely on the remarks of contemporary observers and historians. For example, when al-Muʿaẓẓam ʿĪsā besieged the new Templar castle of ʿAtlit in 1220, it is said to have been defended by 4,000 men, including 300 crossbowmen. The bulk of the kingdom’s fighting force, which probably included part of ʿAtlit’s garrison, was in Egypt taking part in the Fifth Crusade, as was Oliver of Paderborn, who provided these figures. The threat to ʿAtlit was deemed sufficient that the master of the order returned to Palestine with a tested force of Templars and a number of barons. The regular strength of the garrison is unclear, as is how many Templar knights and sergeants may have been among the defenders, while the 300 crossbowmen were probably mercenaries. In a later episode, Baybars’ secretary, Ibn ʿAbd al-Ẓāhir, noted that 480 fighting men defended the Templar castle of Beaufort when it was besieged in 1268. When the castle surrendered on 15 April, these men were led away into captivity, having arranged for the women and children in the castle to be escorted to Tyre as a term of their surrender. This would have provided an opportunity to count the male defenders, but there is no indication of their station. As at ʿAtlit, a minority of these fighters would have been knights; the majority were more likely professionals of a lower class, supplemented by every able-bodied man during such a time of emergency.
The size and make-up of Muslim garrisons is perhaps even less clear, although a small window is provided by the accounts that describe the garrisons installed by Qalāwūn in the 1280s. In 1285, after taking the Hospitaller castle of Margat, a considerable force was entrusted with its defence. This included 1,000 aqjiyya soldiers, 400 craftsmen, 150 emirs of the rank of ṭablkāna, and additional contingents of certain mamlūk corps. In 1290, a year after the capture and destruction of Tripoli, a force of 600 cavalry, under Sayf al-Dīn Taqwī, was left to defend the new settlement that was established inland around the castle first built by Raymond of St Gilles. Although large, these figures align with those reported when Margat was under Frankish rule.
Wilbrand of Oldenburg visited Margat in 1211 and left a description of what he saw. He observed that it was guarded by 4 Hospitaller brothers (knights of the order), supplemented by twenty-eight watchmen. A force of 1,000 people, other than the general citizens of the castle and the attached castle-town, were retained in peacetime to defend the stronghold and presumably contribute to the Hospitallers’ offensive capacity. Supplies were laid in to sustain the castle for five years. Elsewhere, charter evidence reveals that, by 1259, the Hospitallers had arranged to garrison Mount Tabor with forty knights and Crac des Chevaliers with sixty knights, although the number of others from lower stations is unknown.
The most detailed breakdown we have relates to Safed. Bishop Benedict of Marseilles, who led the effort to rebuild Safed when it returned to the Templars in 1240, claims that the defenders were provided with victuals for 1,700 people in times of peace and for 2,200 in times of war. The peacetime garrison consisted of fifty knights and twenty sergeant brothers, indicating that this was the Templars’ main interior base south of the Lebanon. These men were joined by fifty turcopoles and 300 crossbowmen. A labour force of 820, augmented with 400 slaves (probably captured Muslims not valuable enough to ransom) fulfilled the day-to-day operations and construction work at the castle. Collectively, these individuals consumed 12,000 cartloads of barley and grain as well as other victuals each year. Besides people, animals were needed to help work the land and horses were required for the knights and other cavalry elements, as was specie to pay the mercenaries and labourers intrinsic to the castle’s function.
This is a large and very strong castle, fortified with a double wall, displaying in itself many towers, which seem more apt for sustaining heaven than for defence. For the mountain on which the castle is sited is extremely high, such that it holds up the high heaven on its shoulders like Atlas. Very broad at the base and rising gradually on high, it liberally furnishes to its masters each year 509 cartloads of render, which the efforts of its enemies cannot prevent, however often they have tried.
This castle belongs to the Hospitallers and is the greatest support of all that land. For it is opposed by many strong castles of the Old Man of the Mountains and the sultan of Aleppo, whose tyranny and assaults it has held in check to such an extent that it receives from them each year for keeping the peace the equivalent of the value of 2,000 marks. And because it is on guard lest any treason should occur, as can happen, each night it is guarded by 4 knights who are brothers of the Hospital, and by another 28 watchmen. For in time of peace in their outlay for defending the castle the Hospitallers maintain 1,000 people over and above the other citizens of the castle, in such a way that they provide them with every convenience and necessity, [sufficient to supply] the castle with the necessities of life for five years. At the foot of that mountain is a city called Valenia, which although it was at one time larger – so it is said and as may be seen – through divine punishment is now desolate and destroyed. Its episcopal seat has been translated into the castle of Margat, this being on account of fear of the Saracens. This castle is 6 short miles distant from Tortosa.
(Adapted from Wilbrand of Oldenburg 11, trans. Pringle, pp. 69–70.)
Margat. (Courtesy of Denys Pringle)
In times of siege, the numbers of defenders would have swelled as more were hired, recruited or sent from elsewhere in support. According to the Templar of Tyre, more than a thousand Hospitaller and secular knights and sergeants defended Arsūf during Baybars’ siege operations in 1265. Although this is almost certainly an exaggeration, Ibn ʿAbd al-Ẓāhir, who was present, wrote that thousands were taken away, bound with ropes, after the castle surrendered.
Castles like Margat, Safed and Arsūf were among the most impressive in the region – the Templars reportedly paid 1,100,000 Saracen bezants on building costs during the first two and a half years of Safed’s reconstruction and around 40,000 bezants annually thereafter. We know less about the defenders of smaller strongholds. Baldwin, the castellan (and later lord) of Ramla, had eight knights under his command in the first decade of the twelfth century, a force sufficient to defend the town’s tower; although it was an important administrative centre in the early twelfth century, Ramla lacked town walls. To the north, when Bohemond II joined Baldwin II’s push against Damascus in 1129, he left 100 men to defend Kafarṭāb, a significant town between Maʿarrat al-Nuʿmān and Shayzar. This force marched out to confront a party of Muslim invaders and, although defeated, discouraged an attack on Kafarṭāb. Later, in the 1180s, the small Templar stronghold of La Fève seems to have been garrisoned by 80 knights, although this may be an estimate of the total number of defenders.
The castellan and most other retainers who were entrusted with the defence of a large Frankish castle would have been drawn from the knightly class; however, the majority of defenders would have been of lower birth – mercenaries, Frankish burgesses or landowners, and members of the local Syrian population. Many Frankish strongholds were built in regions with a pre-existing Eastern Christian community, where a natural symbiotic relationship formed: the settlement provided goods and services needed by the stronghold, while the stronghold provided a market for such; likewise, a castle might discourage raids in the region, while it might in turn benefit from locals taking part in its defence.
Although Frankish castles are often regarded as enclaves of Latin Europe in an otherwise Muslim Middle East, this was hardly the case. When the kingdom of Jerusalem was established, a slight majority of people living in Palestine were probably adherents of Islam, although the cities along the coast would become predominantly Christian as they fell into Frankish hands. To the north, the princes of Antioch came to rule over diverse populations of Christians and Muslims of various sects, while the county of Edessa may have contained predominantly Armenian and Greek Christians.
Armenians probably made up the bulk of the defenders in the county of Edessa, some holding strongholds as lords and many more serving in garrisons, but local Christians also filled important positions to the south, at least in the twelfth century. In the kingdom of Jerusalem, Baras the Armenian is found in charters of the 1120s and George the Armenian appears in the service of the lord of Caesarea in 1145. The progenitors of most of these families probably accompanied Baldwin I or Baldwin II south when each assumed the kingship of Jerusalem, or came in the service of the Courtenay family, either when Joscelin I became prince of Galilee or following the collapse of the county of Edessa. To these can be added a number of Syrian Christians, such as Arnulf, son of Bertrand the Syrian, and David the Syrian, who held a cave in the Lebanon. In the principality of Antioch, the knights of Margat included the lords Zacharias, Georgius and Theodorus, whose names suggest they were Greeks. The post of marshal of Jerusalem, one of the kingdom’s highest military appointments, may have been occupied by a local Christian during the second quarter of the twelfth century, a certain ‘Sado’, which might correspond with the Arabic Saʿd or Saʿīd. Although it seems unlikely that Sado was a Muslim, a charter from 1155 indicates that two Arab knights held villages in the kingdom.
If local Christians were able to hold lordships, it seems figures of lower rank would have been relied upon to provide some of the manpower required to garrison certain strongholds. Explicit references to garrisons of local Christians are rare south of the county of Edessa; however, one can be found in William of Tyre’s history, where it is noted that the constable of Tiberias entrusted command and the defence of al-Ḥabis Jaldak to local Christian Syrians. This earned the constable of Tiberias William’s criticism when the stronghold was taken in 1182.
In the far south, there were significant Christian populations in Transjordan. Both Montreal and Kerak were founded in areas occupied by local Christian communities, which, according to William of Tyre, meant that the locals could be relied upon. Local Christians helped Reynald of Châtillon transport his prefabricated fleet from Kerak to Aqaba in late 1182, and the townspeople of Kerak were permitted to seek refuge in the castle when Saladin besieged the stronghold the following summer. After the region fell under Muslim rule in the late 1180s, its Christian communities remained. The pilgrim Thietmar, who visited Montreal in 1217 on his way to St Catherine’s Monastery in the Sinai, even encountered a Frankish widow still living in the community outside the castle.
Coreligionists were similarly relied upon to garrison most Muslim strongholds. Around the time of the First Crusade, Turks probably made up the core of most garrisons in Syria; however, a number of ethnic Armenians, who embraced Islam, are known to have been a part of some. For most Seljuk figures, theirʿaskar probably consisted mostly of Turks, while the majority of the defenders of certain Arab strongholds, like Shayzar, may have been Arabs. Inevitably, there would have been a mix of ethnicities represented in many garrisons, which came to include growing numbers of men born around the Black Sea as Ayyūbid rulers became increasingly reliant on regiments of mamlūks.
Beliefs that coreligionists were more trustworthy inspired efforts to populate urban centres with such individuals. During the First Crusade, a portion of the Christian population living in Jerusalem may have been expelled ahead of the crusaders’ arrival. This policy was reversed following the city’s capture: Christian communities from Transjordan were transplanted to Jerusalem, in part to repopulate the city following the massacres that had accompanied its fall.
The populations of captured towns were treated in a range of ways; however, faith does not appear to have been a leading consideration. When cities were taken by force, as occurred at Jerusalem (1099), Caesarea (1101) and Beirut (1110), a sack typically followed and any survivors could expect to be enslaved. Where terms of surrender were arranged, different scenarios might play out. At Acre (1104), the townspeople were permitted to remain for a fee; at al-Athārib (1110), the local population was allowed to stay if it so chose; and at Tyre (1124), most of the population was escorted out of the city, although a significant number of locals probably chose to remain. Along the coast, cities came to be inhabited primarily by Franks, due in large part to the interest shown by the Italian merchant communities and other Europeans. Further inland, some towns that had surrendered on good terms probably remained predominantly Muslim.
Muslim policies were similar when taking over a town. The predominantly Christian population of Edessa was largely respected when the city fell to Zankī at the end of 1144; although he captured the town by force, the local population and captives were quickly set free. Nūr al-Dīn was less accommodating when he took the city in 1146, following its brief recapture by Joscelin II and Baldwin of Marash. During Saladin’s conquest of Palestine in 1187, Frankish urban populations were evicted, which seems to have suited those affected. While most appear to have been free to leave, the conditions of Jerusalem’s surrender placed a small ransom on everyone who wished to depart: 10 dinars (or bezants according to Frankish sources) for a man, 5 for a woman and 1 or 2 for a child. The terms imply that those who could not afford the tax would be enslaved. The treatment of Syrian Christians and Muslims living under Frankish rule at this time was of less interest to both Frankish and Muslim commentators, so we know less about their fate.
From the payments made at Jerusalem, Ibn al-Athīr reports that 60,000 men had been in the city when it surrendered, its population having swelled as people from the countryside and refugees from other cities fled there in the weeks leading up to Saladin’s arrival. These people then joined other refugees behind the walls of Tyre. This enormous body of potential defenders was a strength, so long as they could be fed, and probably contributed to Tyre’s stiff resistance in December 1187. Critical to Tyre’s defence, the Franks were able to maintain their dominance of the sea and a supply route remained open. Earlier, in 1183, Kerak was similarly swollen with extra people, guests at the wedding of Humphrey IV of Toron and Isabelle of Jerusalem. Although these were exceptional cases, involving larger groups of refugees than was the norm, providing asylum could be a critical function of many strongholds.
The groups of people that strongholds were designed to shelter varied considerably, from town walls that secured whole urban communities to rural towers that were occupied by no more than their defenders. While farmers and burgesses might hope to find refuge in a local tower, large strongholds might provide security for defeated armies in times of need.
At the battle of Ascalon, fought to the north of the city on 12 August 1099, remnants of the First Crusade defeated a large Egyptian army. As the battle turned against the Fāṭimids, the city’s defenders closed the northern Jaffa Gate, preventing the crusaders from entering the town but pinning hundreds between the town’s walls and the Christian army. To the east, the right wing of the Muslim army was able to escape into the town through the eastern Jerusalem Gate. Despite their considerable losses, the Fāṭimids’ ability to secure and then defend Ascalon, thanks in large part to the squabbling that broke out among the crusaders after the battle, ensured the city remained an Egyptian possession for another fifty-four years. Circumstances were reversed a few years later, when in May 1102, following the Franks’ defeat at the second battle of Ramla, Baldwin I was saved by the network of fortifications that the Franks had recently acquired and begun to expand in southern Palestine. Fleeing the victorious Fāṭimid army, the king escaped first to Ramla, taking refuge in the tower that acted as the town’s citadel, where he was briefly besieged before escaping and making his way across country to Arsūf and thence to Jaffa.
In 1133, Pons of Tripoli took refuge in Montferrand after he was defeated attempting to confront a large party of Turkoman raiders. The castle was besieged but held out long enough for Fulk to arrive with the army of Jerusalem. Four years later, Pons was again besieged in Montferrand, this time by Zankī, and Fulk mustered his forces once more and marched north to relieve the castle. On this occasion, however, the relief force was defeated. The loss would have been significantly greater had not Fulk and elements of his army been able to seek refuge in Montferrand, joining Pons and the besieged inside the castle. Although this provided the castle with numerous defenders, provisions were stretched extremely thin. Zankī rejected an initial offer of surrender, but was inclined to agree when a Frankish relief force approached, securing advantageous terms before the defenders learned help was at hand. In exchange for their freedom, a tribute of 50,000 dinars was arranged.
Safed played a critical secondary part in the defence of Bānyās in 1157. Nūr al-Dīn’s siege was interrupted by the sudden arrival of a Frankish army led by Baldwin III of Jerusalem, which compelled the Muslims to withdraw and regroup. Baldwin then made a critical error: as he saw to the repair of the town’s defences, he released his infantry and a number of knights, believing the main threat to have passed. Nūr al-Dīn, however, had kept his army together, allowing him to ambush what remained of Baldwin’s force near Jacob’s Ford, half way between Bānyās and Tiberias. The Franks were soundly defeated and many important nobles were taken prisoner. The Muslims searched for Baldwin III’s body among the dead, not realizing the king had been among those to escape, having fled to the safety of Safed. Nūr al-Dīn then returned to Bānyās and besieged its citadel once more. He had not yet captured it when another Frankish relief force arrived, compelling him to abandon the siege.
Baldwin III’s use of Safed as a point of refuge was not the first time he had taken advantage of a stronghold in this way. When evacuating the indefensible rump of the county of Edessa, around 1150, Baldwin skilfully moved between strongholds, frustrating Nūr al-Dīn’s efforts to disrupt the marching order of the Frankish column before it reached the principality of Antioch. A generation later, in 1176, a Frankish force from the kingdom of Jerusalem benefited from the security provided by Tyre when raiding into the Biqāʿ Valley. Intending to join up with Raymond III of Tripoli, the raiding party was intercepted and forced to flee southward.
The importance of a stronghold to fall back on is perhaps most clearly seen in 1177, during Saladin’s invasion of southern Palestine from Egypt. The episode corresponded with the absence of much of the kingdom’s army, which had joined Philip of Flanders on campaign to the north. Baldwin IV, just sixteen at the time, held most of his remaining forces in Ascalon, while the Templars who had not gone north concentrated in Gaza. This left Saladin with a predicament: although his forces far outnumbered those of the Franks, to besiege either Ascalon or Gaza would leave a considerable hostile force within striking distance. Saladin decided to march past both, allowing elements of his army to raid the area west of Jerusalem. Ramla was evacuated and subsequently sacked, nearby Lydda was assaulted and many in Jerusalem, fearing a siege, fled to the Tower of David. As the Muslim army spread out, drawn by the prospect of plunder, the Franks recognized their opportunity. The Frankish army, led by Reynald of Châtillon, included Baldwin II of Ramla and his brother Balian of Ibelin, Reynald of Sidon, and the king’s uncle, Joscelin III, titular count of Edessa and now seneschal of the kingdom, as well as Master Eudes of St Amant and his Templars. The Franks attacked the core of Saladin’s army near Montgisard, soundly defeating it. Saladin had been unable to concentrate and organize his forces in time, but the Muslims’ defeat was only the start of their troubles. With no stronghold to fall back to, the scattered Muslim army began a disorganized flight all the way to Egypt, a catastrophe made worse when local Bedouins, seeing their weakness, raided the baggage that had been left at al-Arish.
Saladin avenged this rout ten years later at the battle of Hattin. Few who fought on the Frankish side escaped following their defeat; however, a number of Templars and Hospitallers fled to Safed, 19km to the north, and Belvoir, 24km to the south, respectively. These men had a brief moment of respite to make last-minute preparations before contingents of Saladin’s army arrived; the ensuing sieges of both castles lasted more than a year.
A similar situation played out following the battle of Forbie in October 1244. Having joined forces with the army of al-Ṣāliḥ Ismāʿīl of Damascus, which was led by al-Manṣūr Ibrāhīm of Homs, the Franks suffered particularly high losses in the ensuing battle against al-Ṣāliḥ Ayyūb of Egypt and his Khwārizmian allies. According to a letter sent to Europe after the battle, only thirty-three Templars, twenty-six Hospitallers and three Teutonic Knights survived the engagement; most nobles who had taken part were killed or captured and huge losses had been sustained by the Frankish infantry and crossbowmen. The patriarch of Jerusalem, who wrote the letter, along with Philip of Montfort and others who had managed to escape, took refuge in the new castle at Ascalon, work on which had begun a few years earlier during the Barons’ Crusade. The castle was attacked a few weeks after the battle and was subjected to a loose siege that would continue in some form until 1247, when efforts were considerably intensified and it was finally taken.
While strongholds could provide shelter in the immediate aftermath of a defeat, they were also used as safe havens by figures on the run.Qalʿat Jaʿbar, for instance, was where Jāwulī Saqāo established himself around 1107, having been forced to flee Iraq. In 1121, Dubays ibn Sadaqa, former ruler of al-Hilla (96km south of Baghdad), was similarly welcomed at the castle when he fled the region. Both Jāwulī and Dubays lost little time looking to the Franks for further support.
Although designed primarily to keep people out, the impressive strength of castles and citadels made them natural detention facilities. Following his capture by the Dānishmands in 1100, Bohemond was imprisoned in Neocaesarea (mod. Niksar), while Baldwin II, during his first captivity, was held in Mosul before Jāwulī moved him to Qalʿat Jaʿbar, where Joscelin I, Joscelin II and Yvette (daughter of Baldwin II) were also held captive at points in time. In a sensational episode, Joscelin I and his cousin Galeran found themselves the captives of Balak in Kharpūt, where they were later joined by Baldwin II, during his second stint in captivity from 1123. Aided by a party of Armenians, the Franks took control of the castle, sending out Joscelin to bring help. Balak returned and retook the castle, ordering the execution of all but the most senior figures. Baldwin II was moved to Ḥarrān and eventually released, but Galeran would later be executed.
Qalʿat Jaʿbar. (monumentsofsyria.com)
The dungeons of Aleppo seem to have been particularly active, as notable figures defeated by Aleppan armies frequently found themselves its guests. Walter the Chancellor, who left a detailed account of the battle of the Field of Blood and his captivity thereafter, was among those taken to Aleppo following the Frankish defeat. Later, Joscelin II was blinded and would eventually die there, and Reynald of Châtillon spent about sixteen years in one of its cells. Further south, Gervais of Bazoches, third prince of Galilee, died a prisoner in Damascus and many notables taken captive by Nūr al-Dīn following the battle of Ḥārim in 1164 were held there for a period, as was Reynald of Sidon from 1189 to 1190. Although many castles served the function of prison at one time or another, it is perhaps important to remember that imprisonment, which was done with hopes of collecting a ransom, was reserved for the wealthy. In 1191, having captured the self-proclaimed emperor of Cyprus, Isaac Comnenus, Richard I is said to have promised him that he would never be forced to wear iron chains, so the English king had him clamped in silver shackles once he surrendered. Isaac was then given to the Hospitallers and held prisoner at Margat.
Those wealthy enough to be held for ransom could also prove useful during their captivity, occasionally acting as negotiators, a job some proved particularly effective at given their motivation to see a conclusion to hostilities that would bring about an exchange of prisoners. In 1167, Hugh of Caesarea, who had fallen prisoner to Shīrkūh in Egypt, was instrumental in negotiating the terms of Saladin’s surrender of Alexandria to Amalric. Whether as a place of refuge for a defeated army, safe haven for citizens of a town, or place of incarceration, strongholds were dynamic structures.
We are fortunate that some sources record who commissioned certain strongholds and who subsequently added to them; however, we are left to infer who was responsible for the establishment and development of most. As little as we know about who paid for these works, we know even less about the individuals who were responsible for designing them and overseeing construction, let alone the hundreds of people that provided the various skills and labour that went into the building of any stronghold. We can assume that the commissioner of a castle, citadel or line of city walls had an influence on what the finished product would look like, but the extent to which this person was involved in choosing a layout or various design features is far from clear. In exceptional cases, we hear about an expert who was charged with overseeing refortification efforts. For example, Saladin placed Bahāʾ al-Dīn Qarāqūsh al-Asadī in charge of building his new citadel at Cairo, and later summoned him to refortify Acre after its capture in 1187.
Qarāqūsh was a manumitted mamlūk of Saladin’s uncle, Shīrkūh, who went on to become one of Saladin’s most capable emirs. Saladin tasked him with making repairs to Cairo’s defences in 1170–71, before entrusting him with construction of the city’s new citadel. The work was evidently satisfactory as he was relied on thereafter for Saladin’s most important fortification works. When called to secure Acre, Qarāqūsh brought with him tools, supplies and manpower, indicating he had built up a logistical network through his work in Egypt. Qarāqūsh was part of the Turko-Kurdish military elite that had spearheaded the subjugation of Fāṭimid Egypt and his respon- sibilities at Acre included the city’s defence – it was he who commanded the garrison through the Frankish siege of 1189–91, becoming one of the Franks’ highest-ranking prisoners following the city’s surrender.
Inscriptions on some Muslim strongholds identify the year that certain elements were added and the responsible patron. Although these reveal who oversaw or paid for building works, as will be addressed below, they provide a far from complete picture. For instance, inscriptions at Cairo, Jerusalem, Damascus, Bosra, ʿAjlun and Mount Tabor credit the construction of towers to al-ʿĀdil, some of his sons and certain emirs; however, the similarities between these projects suggest a singular guiding policy was in place. Al-ʿĀdil appears to have arranged for all of these works to conform to his signature style, which included the use of large quadrangular towers and marginally drafted ashlars. Similar stylistic signatures can be seen in the works of Ayyūbid contemporaries, such as al-Ẓāhir Ghāzī of Aleppo and al-Mujāhid Asad al-Dīn Shīrkūh of Homs. While this gives a sense of how commissioners could influence the appearance of their strongholds, the role and input of the architects and master builders, who were probably used at multiple sites, remain less clear. Epigraphic and textual evidence from the Mamlūk period similarly provides the names of a number of emirs who oversaw significant refortification efforts; however, the same obscurities typically remain, while the broader assortment of building styles evident in Mamlūk work complicates things even further.
This year al-Ādil encamped with his army around the citadel of Mount Tabor. He gathered artisans from all towns and employed all the emirs of the army in constructing this place and carrying the stones. Some 500 builders were employed in its erection, not counting the labourers and stone cutters. Al-ʿĀdil did not quit the site until the citadel was completed.
(Adapted from al-Maqrīzī, trans. Broadhurst, p. 156.)
At Ṣubayba, both round and square towers were added by the Mamlūks, but what is strange is the use of different masonry styles. Analyses of the surviving epigraphic evidence point to a single reconstruction campaign, overseen by Badr al-Dīn Bīlīk al-Khaznadār, one of Baybars’ personal mamlūks. Whereas the quadrangular towers appear fairly functional, most of them built around earlier Ayyūbid towers, greater architectural grace is found in the rounded towers. The latter are quite similar to the southeast tower at Crac, which was also built under Baybars following the castle’s capture in 1271.
Under the Mamlūks, when a castle was taken at the end of a siege, it was often granted to one emir while another was charged with overseeing its repair or expansion. Following the capture of Safed in 1266, Baybars made Majd al-Dīn al-Ṭawrī governor (wālī ) of the stronghold and the surrounding region and placed ʿAlāʾ al-Dīn Aidughd in command of the castle’s garrison. When he returned to inspect the castle the following year, the sultan took a more proactive role in developing its defences, dividing reconstruction tasks among the emirs of his army.
It was not uncommon for Muslim commanders, when building or destroying a stronghold, to assign sections or certain towers to their emirs. In 1191, Saladin had destroyed the defences of Ascalon in this way and went on to refortify Jerusalem in a likewise manner, as the Third Crusade threatened to besiege the city. Al-ʿĀdil distributed the task of destroying the citadel of Dārūm among the emirs of Egypt in 1196, and he may have leaned on the Ayyūbid princes of Syria, whom he had recently come to dominate, for assistance when he set about rebuilding the citadel of Damascus. When considering the former, this seems to have been a means of mobilizing labour, while circumstances surrounding the latter suggest it had more to do with procuring financing, as al-Manṣūr Muḥammad of Homs, who was more willing to help than others, is the only Ayyūbid prince named among the inscriptions marking the construction of the citadel’s various towers. Half a century later, in 1265, Baybars distributed the work of dismantling Arsūf among his emirs in a similar fashion, forcing the defeated Hospitaller garrison to take part.
On the nineteenth [of Shaʿbān, 12 September] the Sultan [Saladin] arrived at Ascalon, which he wished to raze, being himself unable to hold it. He divided the towers among the emirs for demolition and great was the lamentation and weeping among the inhabitants in grief and sorrow at its razing. For it was one of the most beautifully constructed of towns, most strongly fortified in its walls, and most delightful to dwell in. The destruction and burning did not cease until the month of Shaʿbān had ended [21 September].
The Ḥāfiẓ ʿAbd al-‘Aẓīm al-Mundhirī, in his book al-Muʿjam al-Mutarjam, says: ‘I heard the illustrious Emir Iyāz ibn ʿAbdallah – Abū al-Manṣūr al-Bānyāsī al-Nāṣirī – saying ‘‘When we razed Ascalon, I was given the Towers of the Templars. Khutluj demolished a tower on which we found inscribed ‘Built by the hand of Khuṭluj’ which was a most strange coincidence.’’ Likewise, the illustrious Qadi Abū al-Ḥasan ʿAlī ibn Yaḥyā al-Kā tib told me: ‘‘I saw at Ascalon the Tower of the Blood while Khutluj al-Muʿizzī was demolishing it in the month of Shaʿbān. And on it I saw this inscription: ‘The construction of this tower was ordered by our illustrious master, the Emir of Armies – Badr al-Jamālī – and executed by his servant and lieutenant Khuṭluj in Shaʿbān.’’’ I marvel at the coincidence, that it should be built in Shaʿbān b y a Khuṭluj and destroyed in Shaʿbān by a Khuṭluj.’
(Adapted from al-Maqrīzī, trans. Broadhurst, pp. 93–4.)
Returning to Safed, the task of rebuilding the towers of the castle was given to Sayf al-Dīn al-Zayni, though how much creative freedom the emir had is unclear. In 1268, following the capture of Beaufort, Şārim al-Dīn Qāymāz al-Kāfirī was made deputy (nāʾib), while rebuilding efforts were entrusted to Sayf al-Dīn Balabān al-Zaynī, presumably the same man who had overseen the construction of the towers at Safed. Three years later, following the surrender of Crac in 1271, Şārim al-Dīn Qāymāz al-Kāfirī was made governor, having received Safed earlier in 1266, while the rebuilding of the castle was delegated to ʿIzz al-Dīn al-Afram and another emir. Al-Afram is again found with another ‘restoring’ the districts, or perhaps the castle, of al-[Ullaiqa, a recently surrendered Assassin stronghold, later the same year.
It is unclear whether these men were personally experienced or proven in the arts of fortification, or if this was rather part of their positions. ʿIzz al-Dīn Aybak al-Afram was amīr jāndār, the emir of arms-bearers, which carried the rank of emir of a thousand and responsibility for the Mamlūk armoury and high-level executions. He was in charge of Baybars’ artillery at the sieges of Caesarea, Arsūf and Safed and had been given the task of destroying the church of Nazareth in 1263. In 1275, he was among a small group charged with smoothing the banks of the Black River, the westernmost of three that flowed south into the Lake of Antioch, in order to help the army cross. He seems to have had some experience in such engineering projects as it was he who carried out the clearing and development work along the canals of the Nile during Baybars’ reign. Sayf al-Dīn Balabān al-Zaynī was amīr ʿalam, the emir of the banner, and another leading figure under Baybars. In 1264, he was sent to inspect the strongholds of Syria, to ensure they were sufficiently supplied, and to review the troops of Hama and Aleppo. It was also he who brought a trebuchet from Ṣubayba to Caesarea in 1265 and, in 1266, conveyed the artillery stored in Damascus to Safed. Ibn ʿAbd al-Ẓāhir does not associate him with any other rebuilding efforts. While ʿIzz al-Dīn al-Afram would appear to have had the experience to design or at least competently direct the construction of a castle, Sayf al-Dīn Balabān al-Zaynī seems to have acted in a more supervisory role, lacking direct engineering experience.
Turning to the epigraphic evidence, many Muslim building projects were commemorated with an inscription. These can be quite helpful when trying to date building phases, as they are often quite formulaic, typically including the name of the sultan, the emir who oversaw the work or whose lands the stronghold was in, and occasionally the person(s) who supervised the construction of the tower, wall or feature displaying the inscription. From these inscriptions we know that refortification efforts at Qalʿat Ṣadr, which were overseen by al-Ādil in 1182–83 on behalf of Saladin, were coordinated by one of his emirs, Ṣārim al-Dīn Barghash al-ʿĀdilī, who appears later supervising building works at Kerak. Similarly, the name of ʿIzz al-Dīn Aybak, who carried out work on behalf of al-Muʿaẓẓam ʿĪsā, can be found on a number of strongholds, including Mount Tabor (1212) and ʿAjlun (1214).
A number of inscriptions found at Ṣubayba also provide the names of lower figures. Among those dating to the early Ayyūbid building phase, which was completed under al-ʿAzīz ʿUthmān, some mention a figure named Abū Bakr, presumably the master builder or person who directly oversaw the work. A slightly later inscription on the large cistern in the southwestern part of the castle reveals that it was constructed in 1239–40, during the rule of al-Saʿīd Hasan (son of al-ʿAzīz ʿUthmān), under the supervision of Emir ʿAzīz al-Dawla Rayḥān al-ʿAzīzī, while an emir named Mubāriz al-Dīn was governor of the castle. A later Mamlūk inscription, which commemorates the reconstruction of a part of the castle during the reign of Baybars, notes that the work was carried out under Bīlīk, who held Ṣubayba as part of the iqṭāʿ of Bānyās. The governor of the castle at this time was a figure named Bektut, while building activities were directed by two civilians, ʿAbd al-Rahmān and ʿAbd al-Wahhā b. These final figures were, respectively, the muhandis and miʿmār, which might, rather inaccurately, be translated as engineer and architect – the status of the latter was more akin to a repair person than a designer of buildings.
This kind of evidence can be found elsewhere. At Montreal, an inscription on one of the towers rebuilt at the end of the thirteenth century includes the names of Sultan Ḥusām al-Dīn Lādjīn (r. 1296–99), ʿAlāʾ al-Dīn Qubrus al-Manṣūrī, the emir who oversaw the tower’s reconstruction, and a certain Muḥammad ibn ʿAbd al-Hamīd, the muhandis. Unfortunately, we know little more about this last figure or those who held this position elsewhere.
Despite the number and detail of the sources for this period, we still have few indications of who was ultimately responsible for determining the layouts of castles, the shapes of their towers and the siting of various elements. Although many lords and emirs probably provided significant input, the knowledge and experience to bring these visions to fruition was the responsibility of a master mason or muhandis, who combined the roles of architect, engineer and project coordinator. Below such experts were those who did more of the heavy lifting. There are references to the use of captive Franks during Saladin’s construction of the citadel at Cairo and Muslim prisoners are said to have been used to help reconstruct Safed, but we know little else about the workforce used to build strongholds.
* * *
Fortifications were constructed for a number of reasons and fulfilled a plethora of functions; however, all strongholds shared a defining purpose: to keep the undesired out. Even if the surrounding countryside were suddenly flooded by a hostile army, a stronghold was an island of controllable land. Without taking a region’s strongholds, an invader could hardly expect to maintain control once his army withdrew. Often built in strategic positions, in both political and geographical senses, strongholds were ideal places from which to project influence and launch raids or larger campaigns. Their garrisons could be used to boost the size of a field army and they might act as a secure place of refuge should a campaign fall subject to disaster. These capabilities, however, also made them potential bases of resistance, obstacles for powerful princes who relied more on large field armies to enforce their rule.
A vast number of factors went into the decision of where to build a stronghold. The availability of building materials and natural resources, sources of potential income and specific strategic objectives were just some of the most obvious. All strongholds, from the smallest tower to the largest castle, required the support of a local population. While many were built near existing communities, others were constructed in areas with obvious advantages, which then attracted the growth of new settlements. Local administration was often centred on these structures, as their defensibility was symbolic of the local ruler’s authority – as true of the castles and towers in the countryside as of the citadels that dominated cities.