by John France
The arrival of the army before Antioch on 21 October triggered a debate on strategy. The suggestion was made in the council of leaders on 21 October that the army should mount a distant blockade of Antioch. Those who favoured the idea pointed out that the army was tired – they had after all been marching for some four months – and much of it was dispersed amongst captured strong-points. Better to sit out the winter in comfort, they urged, and wait for the arrival of reinforcements from the emperor and the West. It was an intelligent idea and evidently was supported, and perhaps even conceived of, by Tatikios who may well have known that this was the method by which the Byzantines had recaptured Antioch in 969 when Baghras was an important base (see fig. 7).85 Tatikios revived the idea when the army was starving in January and February 1098. The count of Toulouse, however, urged his comrades to trust in God and pressed successfully for an immediate and close siege, and this was the course of action taken.86 The great virtue of the close siege was that it kept the army together under the control of its leaders; a distant blockade could have had a very adverse effect on the sense of purpose of the Christian army. However, many crusaders were on garrison duties away from Antioch and Raymond of Aguilers says that knights were all too eager for such work.87 These captured fortresses and cities were the fruits of the Armenian strategy which the crusaders had followed, and despite Raymond’s fulminations it now served them well. As already noted, the Armenian population sent aid and their merchants sold food, albeit at a high price, but the availability of these lands as a source of food was very valuable. The presence of many knights in fortresses facilitated this and reduced the food problem to a degree. In effect, the crusaders imposed a close siege upon Antioch and enjoyed some of the benefits of a distant blockade and their situation was later improved when Baldwin of Boulogne took over Edessa.
Fig. 8 Siege of Antioch, October 1097 – February 1098
The reason for this debate was quite simply the scale of the problem which they now faced. Antioch was no longer the magnificent city of late Roman times, when its population had reached 300,000 and it rejoiced in its position as capital of the East. Its prosperity was ruined by a series of disasters – the fire of 525, the earthquakes of 526, 528 and 588, the Persian sack of 540, the plague of 542 and the Arab capture of 638. However, though the inhabited city was much reduced, it still sheltered within the walls of Justinian’s rebuilding completed by 560 and as modified by earthquake, war and the ravages of time.88 The crusaders were much struck by the splendour and strength of the place, as was Ibn Butlân in 1051 when he described its walls as having 360 towers: archaeological investigation has found evidence of over sixty (we are not sure of the original total).89 Antioch was built on the eastern bank of the Orontes river, though its walls touched the stream only at the Bridge Gate (see figs. 7 and 8). The fortified area was about three kilometres long and two kilometres deep extending up the eastern wall of the Orontes valley formed by the northern extension of the Jebel al-Ansariye. The mass of Mount Silpius included in the enceinte rises to a height of 512 metres and about 700 metres north of its highest point stood the citadel, rebuilt after the Byzantine reconquest of 969 and dominating the whole enclosure. The wall then dropped into the deep gully of the Parmenian torrent where the Iron Gate restrained this dangerous stream before climbing onto the southern flank of Mount Staurin and then descending sharply to the plain by the Orontes at the northern edge of the city.90 The crusaders approached the city from the north, and here in the valley bottom the wall was pierced by three gates which the crusaders called the St Paul Gate at the very foot of the mountain, the Dog Gate further along and then the Gate of the Duke nearer the river. This northern wall of the city was a double wall, for Albert of Aix mentions an outer wall in connection with operations outside the Dog Gate and tells us that Tancred lurked in the space between the main wall and the barbican before launching a surprise attack on the besiegers during the second siege.91 Beyond the Dog Gate the wall angled towards the river bank, though it is not clear that the outer wall continued at this point, which it met at the Bridge Gate giving access to the plain on the west bank of the Orontes and the roads to St Symeon Port, Alexandretta and Marasch. After that the wall followed the river fairly closely then turned away from it to the St George Gate, from which a road ran to the ancient suburb of Daphne and beyond down to Laodicea and inland to Syria via the Jisr ash-Shogur. The wall then followed the line of a gorge, the Wadi Zoiba, rising up onto Mount Silpius. The inhabited part of the city nestled down on the lower slopes of Mounts Staurin and Silpius and the narrow strip of plain by the Orontes. In ancient times the city’s dominating feature had been the great colonnaded Street of Herod and Tiberias, running from north to south. It must have existed in some form in crusader times, for its line has now survived in the modern Kurtulus Caddessi. But there was a huge and barely inhabited area within the circuit of some twelve kilometres, amply studded with towers. In Byzantine times the city was held by a garrison of 4,000, such was the passive strength of the defences and their inaccessibility. Yaghisiyan seems to have had forces of about the same size. Raymond of Aguilers, as we have noted, suggests a garrison of 2,000 first-rate knights, 4,000–5,000 other mounted men and more than 10,000 foot. Stephen of Blois gives a total number of 5,000 enemy troops in the city which was probably much nearer the mark for the evidence of events suggests that the garrison was no more than adequate for its task. At the end of December 1097 the crusaders sent a major foraging expedition out into Syria at a time when Duqaq of Damascus was approaching with a relief army; on 9 February 1098 the crusaders fought against another relief force under Ridwan of Aleppo. On both occasions Yaghisiyan’s garrison mounted major sorties in support of the expected relief but although these punished the crusaders they were not strong enough to inflict a major defeat upon them nor to burn the Bridge of Boats which enabled the crusaders to cross the Orontes and threaten the Bridge Gate or travel to St Symeon. When the crusaders threatened to build a counterfort outside the Bridge Gate, Yaghisiyan made a desperate effort to prevent it by ambushing the supply column coming up from the coast, but despite initial success he was unable to prevent the construction.92
Once they had resolved upon a close siege the crusader leaders were forced into a cautious strategy. Because Antioch was so large the risk to the attacker was dispersal of his strength. There could be no question of surrounding this huge area; the crusader army was not big enough and any effort to invest a considerable section or sections would open the way for the enemy to sally out and defeat the crusaders in detail. Nor was there much chance of a sudden assault being successful. In the flatter areas the defences were formidable and the enemy could move their forces about quickly. In the mountains it would be very difficult to move large bodies of troops across the jumble of rocks, ravines and harsh slopes and to sustain them while they tackled the walls which rise to twelve metres and are sited to take advantage of the very rough terrain. Of course this cut both ways – the enemy could not seriously assault them from this direction, but with their knowledge of the pathways of the area the Turks could and did mount raids which sapped crusader strength. There is a back road to the high eastern defences of the city; today it leaves the Aleppo road some two kilometres north of what was the St Paul Gate and winds around the city passing via the modem Altinözü into the upper Orontes valley near the Roman bridges at Jisr ash-Shugur which give access to the Syrian plain. There was certainly at least a path here, for its use was vital in the final crusader assault on the city and it probably led to the Roman road from Antioch via Delphi to the Jisr ash-Shugur (see figs. 4 and 7).93 It was presumably used by both sides for raiding. But the dominating factor which shaped the actions of the army, once they had decided on a close siege, was the need to avoid dispersal of effort. This explains the very slow and very systematic extension of the siege, so that as late as March 1098 Bruno of Lucca could tell his fellow citizens that the army ‘had surrounded the city in siege, though not very well’.94 The crusader
army had approached from the north and it was from here that they systematically extended their grasp over Antioch. It was not perhaps the ideal position but they were well placed to fend off enemy attack coming down the Orontes valley onto their rear and to have attempted to establish major forces on the west bank of the Orontes opposite the Bridge Gate immediately would have strained the army’s resources (see fig. 8). The sheer size of Antioch enjoined upon the crusaders a cautious strategy of building up the blockade; Kerbogah’s failure to appreciate this problem was later a major factor in his defeat. In the first phase of the siege, which would last until Christmas 1097, the crusader army steadily extended its grip on the city, though at a great cost in lives.
Albert of Aachen’s account of the order of battle of the besieging army is largely supported by that of Ralph of Caen. Neither the Anonymous nor Raymond of Aguilers gives an order of siege such as that which they provide for Nicaea.95 Albert says that Adhémar of Le Puy commanded the army for the approach to Antioch but does not make clear that Bohemond had already gone before and arrived at Antioch on 20 October.96 He gives a poetic description of the army in all its glory and numbers – 300,000 – as it took lunch on 21 October 1097 close to the city at a place called Altalon. He then records its deployment for the siege (see fig. 8). Albert seems to have been given an order of siege related to the point at which the army came up to the city on its north wall by the St Paul Gate. He says that a group of lesser figures, Tancred together with Roger of Barneville and Adam FitzMichael established themselves close to ‘Altalon’ and kept the Turks blockaded there which suggests that this group was close to Bohemond who took ‘the gate towards Persia, at the point where the mountain meets the plain’ – a good description of the location of the St Paul Gate, the natural point of entry for any force coming down the valley of the Orontes. Ralph confirms that Tancred was close to Bohemond but suggests that he was encamped with the next contingent along, the North French. Albert says that Tatikios, whom Ralph never mentions, took up station somewhat back from the city wall, but forward of him were Baldwin of Hainault, Robert of Normandy and Robert of Flanders. Beyond these were the forces of Stephen of Blois and Hugh of Vermandois. This is confirmed by Ralph, with the exception of the names of lesser leaders. Somewhat later, after a brief description of the city, Albert makes it clear that Adhémar and Raymond of Toulouse were attacking the Dog Gate, and that beyond them was the gate besieged by Godfrey opposite which a pontoon bridge was later constructed across the Orontes. Ralph’s more schematic description also appears to confirm this. This left the army with major problems which are rather well summarised by Ralph of Caen.97 St Symeon was their obvious port of supply, but via the Bridge Gate the garrison was in a position to cut the road and to raise havoc with small groups or individuals foraging in the plain to the west and south of the city. They could also attack traffic going to Laodicea and, to a lesser extent, Cilicia. Their access made it difficult to establish a force along the southern part of the wall; south of the Bridge Gate the western wall was so close to the river that an attacking force could have done nothing, while the southern wall was built on a deep gully, the Wadi Zoiba, which made it almost unapproachable. On this side the St George Gate, giving access onto the road to Laodicea, was the only point on the southern circuit worth attacking but as long as the enemy had free access over the Bridge Gate this was hazardous. In addition, the garrison could sally forth from the western side of the river and fire on the Frankish camps close to the walls, and this, as Raymond of Aguilers reports, they did.98 Control of the bridge over the Orontes in front of the Bridge Gate was therefore the vital point in the attack on the city. As long as that gate was open the garrison could take the initiative and could easily bring in supplies via the unguarded St George Gate which it covered. There was, however, a limit on what a garrison of not more than 5,000 could achieve. Yaghisiyan could only harass the crusaders in the hope that they would tire of the siege or that they would be defeated by the allies he was actively seeking. For the crusaders, it was vital that they close the Bridge Gate because of the damage the garrison could inflict upon them. This was not achieved until March of 1098, by which time the siege was six months old. In that period there were other priorities pressing upon the crusader army.
The crusaders seem to have recognised that there was little chance of taking Antioch by assault, and nothing of the sort seems ever to have been suggested. Stephen of Blois, who would later be chosen to command the crusade, remarked to his wife that Antioch was ‘a city great beyond belief, very strong and unassailable’ and this opinion was supported by Raymond of Aguilers while Fulcher suggests that the leaders adopted a Micawberish policy of sticking out a siege and seeing if something would turn up.99 All these are writing with hindsight but Fulcher was probably right in a sense. All they could do was to conserve their own forces and squeeze Antioch, disrupting the normal life of the city in the hope that something would give. They were probably aware that the Greeks had seized it in 969 after a blockade, by corrupting one of the commanders, and that treachery had opened its gates to Sulayman in 1086.100 This strategy of blockading Antioch meant that there were very few siege operations such as those we have noted at Nicaea. Early in the siege Adhémar and Raymond were faced by constant sallies from the Dog Gate. Outside it, just below the confluence of the Parmenian torrent and the stream from St Peter’s fountain, was a marshy area with a small bridge (see fig. 8). The Provençals first tried to demolish the bridge with hammers and other tools but it was too strong. Then they built a wooden penthouse covered with osier which they pushed onto the bridge; this would not only have prevented sallies but would have also acted as a base for operations against the gate. The enemy showered the machine with arrows – probably the real reason the bridge could not be demolished – and the Christians replied with bows and crossbows, but the defenders sallied forth and drove off the attackers and set the penthouse on fire. Then three mangonels were built in the hope of destroying the outer wall which defended the city at this point, but they failed. In the end the Provençals organised masses of men (Albert says 1,000) to block the bridge with huge stones and tree-trunks.101 Albert never dates this fight but it seems to have occurred early on, before the Bridge of Boats, which was certainly in place by December 1098, was built.102 Another machine was used in late March or April 1098. After the building of a fort outside the Bridge Gate, the leaders decided to built another penthouse, a talpa which they pushed onto the bridge with a view to breaking it down, thereby preventing any further enemy sallies. Although it got onto the bridge and began its work, the enemy caught its crew asleep and burned it, much to the irritation of the army as a whole.103 In neither case was the machinery used in an all-out assault – merely as a means of tightening the screw on Antioch.
The very early part of the siege was remembered by the chroniclers as a happy time when food was plentiful and the enemy quiescent; with the resources of the Orontes valley and the Amouk at their disposal, and fairly free access out into the Syrian plain foraging was easy. But this did not last for long as enemy forces from Antioch and Harem began to harass the army. The main effort of the garrison seems to have been devoted to attacking along the west bank of the Orontes from both the Bridge Gate and the St George Gate, near which there was a ford, cutting communications with St Symeon (see fig. 8). Their activities figure large in the accounts of Raymond of Aguilers and Albert, for the forces about which they are best informed were close to this area. The Anonymous does not mention this fighting – he was preoccupied with enemy raiding down the valley of the Orontes and the mountains above Bohemond’s camp.104 The crusaders identified Harem as the main base of these raiders and sent Bohemond and Robert of Flanders to attack it (see fig. 4). Their scouts found it but were driven back to where Bohemond’s main force lay in ambush. In the subsequent fighting, Alberadus of Cagnano was killed and Herman of Cannae lost a horse, but the enemy suffered numerous casualties. Harem was not taken and remained such a grave threat that wh
en materials and skilled men arrived with the Genoese fleet on 17 November the leaders resolved to build a fort, which later rejoiced in the name Malregard, on the hill above the camp of Bohemond.105 It was well placed to check enemy raids coming down the Orontes valley or around the back of the city’s defences.
Much greater pressure was exerted upon the garrison of Antioch by the construction of the Bridge of Boats (see fig. 8). Albert of Aachen is very clear about its purpose – to check enemy raids and open the road to St Symeon, and he is quite clear that it was situated near to the Duke’s Gate which the Germans besieged.106 It enjoyed a limited protection because attackers had to cross the Wadi al Quivaisiya which flowed into the west side of the Orontes a little way to the south. Unfortunately, nobody says when it was built, but it was certainly in place by the time of the foraging expedition of December 1097.107 Ralph mentions the new bridge but his account of the siege is schematic and the author of the Gesta Francorum never mentions it. Albert’s dating at this point is very poor and he is ignorant of the expedition to Harem, the building of Bohemond’s tower and the arrival of the Genoese fleet, but the context of his account suggests that the Bridge of Boats was built in November 1097. Christmas can certainly be taken as the absolute outside date for the building of this bridge and the probability is that it was built much earlier, drawing on the materials and expertise brought by the Genoese. The pace of siege warfare was, by modern standards, intolerably slow but we can see just how active and organised the crusaders must have been. It was no small achievement to construct Malregard and the Bridge of Boats, which must have been fortified to withstand enemy attack, in a period of less than six weeks, at a time when the army was depleted by knights who were keen to serve outlying forts and when there was a good deal of continued small-scale fighting.108 This is a tribute to the smooth functioning of the collective leadership which was probably well suited to the conduct of a siege with its deliberate pace. The crusaders must have been painfully aware of how open Antioch was on its southern flank, and Albert says they awarded Tancred forty silver marks per month and sent him into the mountains to establish a camp to blockade two gates on that side of the city and observe enemy movements. It is not clear which two gates are intended but one is described as being in the mountains and the other near the Orontes, which suggests that Tancred was between the St George Gate and the posterns above it.109 Raymond of Aguilers and the Anonymous both say that Tancred was paid to establish a fortress outside the St George Gate in April 1098. It is possible that Albert has misdated the event but he appears to be telling us about some earlier initiative so perhaps the episodes of gallantry and success related by Ralph of Caen occurred at this time. Moreover, in the New Year of 1098 the Anonymous describes this area as ‘Tancred’s mountain’ and Albert says specifically that Tancred abandoned this position at the end of 1097. Indeed, the emphasis in Albert’s account of Tancred’s doings is on surveillance and fighting in the mountains, rather than the counterfort at the St George Gate of which we hear later.110 If Tancred did establish himself here for a time it must have been a blow to the city. Their garrison was aggressive in its efforts to attack the crusaders and they enjoyed the great advantage of height, for from Mt Silpius they could observe the whole crusader army and its movements. This was the other side of the coin. The crusaders had begun a war of attrition – Yaghisiyan tried the same thing. By the Gate of the Duke there were woods where some of the Franks relaxed. On one occasion a noble archdeacon of Metz was playing dice with a lady when they were ambushed and on another Arnulf of Tirs was killed going to the rescue of some people who had been ambushed. After this, Godfrey ordered the clearing of the woods.111 Albert lamented the sufferings of the army: ‘Morning, noon and night every day there were these sudden attacks, sallies, scenes of carnage and endlessly you could hear in the Christian camp always new lamentations over further losses’.112