The first enemy: the Turks of Asia Minor

On 6 May 1097 elements of the crusader army appeared before the city of Nicaea held by Kilij I Arslan, the Seljuk Sultan of Rhüm (1092–1107). The city lies in a fertile basin bounded to the west by the Ascanian Lake (Iznik Gölü). From the south gate (Yenişehir Gate) the land rises sharply into the 800-metre-high Avdan Daglari. From the north gate (Istanbul Gate) the rise into the much higher Naldökan range which the crusaders had crossed is much gentler and only becomes appreciable after three kilometres. To the east (Lefke Gate), a wide and gently sloping valley rises to a watershed then slopes mildly down to the valley of the Sangarius (Sakarya Nehri) and the military roads to Ankara and the Anatolian plateau (see fig. 5). Raymond of Aguilers and Albert of Aachen were much impressed by Nicaea’s fortifications which the leaders examined carefully. Fulcher remarked on the determination and cruelty of its garrison.1 Its fortifications were Roman, dating from the fourth century, but they had been modified and kept in repair under the Byzantine empire. A great wall, pierced at the points of the compass by four main gates, surrounded the city. It was probably about ten metres high and studded with 114 round or square towers rising to seventeen metres, and its circuit measured 4,970 metres. There was a double ditch around the outside.2 These fortifications were made the more formidable because the garrison needed to defend only half their circuit. From the north to south gates the western wall of the city followed the Ascanian Lake whose huge size, forty kilometres long, made it impossible to blockade unless the attacker had boats. It is very important to recognise that until the crusaders brought up boats they faced an enemy who had only to defend half the circuit, and this is the key to understanding the course of military events at Nicaea.3

Amongst the crusader contingents arriving for the siege was, as we have noticed, a force of Byzantine troops under the command of Tatikios. They were later reinforced by more soldiers and some small boats under Boutoumites who blockaded the Ascanian lake on the western perimeter of the city. This was a comparatively small force for Alexius to send in support of what he regarded as his men, almost his hirelings. This is especially true because, as Anna Comnena says, he had twice before attacked the city which seems to have fallen into Turkish hands in 1078, in 1081 and 1086.4Here was an ancient city whose loss to the empire was deeply felt. The Seljuks were converting it from an outpost into a real capital, thus threatening to stabilise their régime, yet the emperor would hazard only a small force, and, above all, would not come himself. Anna stresses that her father was anxious to regain Nicaea, but offers only the feeblest of excuses for his refusal to join the siege – that he feared the enormous numbers of the Franks.5 Anna’s insistence on the bad faith and untrustworthiness of all Latins is intimately connected with her case that they had broken their oath to Alexius in the matter of Antioch and owes much to hindsight.6 It is, in fact, a revelation of the extent to which much modern writing about the crusades has been from a pro-Byzantine standpoint that her statement has passed unchallenged.7 Alexius had reasons to distrust Latins – the expedition of Guiscard is clear evidence – but he distrusted almost everybody else as well and he had actively sought them as mercenaries. His caution on this occasion probably owed much to his attitude to the Seljuks of Rhüm and the curious process by which they conquered Anatolia.

The Turks are part of a vast family of steppe peoples who include the Mongols. They first appear in western history in the guise of the Huns and later as the Hungarians of the middle Danube. The Patzinacks and Uzes, who were such a scourge of the Byzantine Balkans, are also of the same people. It is, however, with the people of Turkestan – the nomadic tribes occupying a vast area from the Black Sea to Central Asia – that we are concerned, and in particular with the Oghuz who pressed on the frontiers of Islamicised Asia and Persia. Amongst them a pre-eminent family were the descendants of the legendary Seljuk.8 The expansion of Islam into Transoxania brought these Turks into intimate conflict with Islam along the frontiers where ghazis, Islamic volunteers, and nomads waged war. But the Shamanist Turks began to be converted in large numbers to Islam, and in the tenth century we see the creation of Turkish Islamic powers like the Karakhanids of Bukhara and the Ghaznavids who ruled on the borders of India.9 Thus the border between Islam and the Turks became porous to Islamicised Turks, some of whom were already established in the Islamic heartlands by a different process. The Arabs who destroyed the Roman and Persian empires in the seventh century were a warrior aristocracy ruling over diverse peoples and sheer military need forced them to incorporate those peoples into their armies.10 This diversification facilitated the recruitment of peoples with special skills – the Daylamis of the Caspian area provided good infantry until the crusader period. Kurds provided light cavalry and infantry. For a long time Khorasanian horse archers and cavalrymen were important but they tended to be replaced by Turks. At the same time, rulers favoured such processes which lessened their dependence on the tribal elements to which they owed their power.11 The destruction of the Umaiyad Caliphate of Damascus in 750 and the rise of the Abbasids of Baghdad marked a political transformation which favoured Iranian groups, especially a military élite associated with Khorasan, and Mesopotamian groups at the expense of Arabs.12 The Caliphate developed a complex military organisation and placed more and more emphasis on mamlūk troops, slave-soldiers who were often recruited from the Iranian and the Eurasian steppe. Under the Caliph Al Mu’tasim (892–902) Turkish troops became well established in the Islamic armies and he created the great palace complex of Samarra near Bagdad to house these élite forces.13 By the eleventh century Turks were a powerful element in almost all Islamic forces, even as far afield as Egypt, and at this very time political developments on the frontier made them more important. In 1025 a group of Oghuz, having been settled in Khorasan, rebelled and were ejected by the Ghaznavids, but they were followed by another in 1035, pre-eminent amongst whom were Tughril and Chagri who by 1040 dominated all of Khorasan including Merv and Nishapur and drove out the Ghaznavids. While Chagri consolidated their position in the east Tughril turned west. He and his people had imbibed a fierce Islamic orthodoxy and the domination of the Caliph at Baghdad by Shi’ite Buwaihids and others was a scandal upon which he capitalised. His entry into Baghdad in 1055 was a peaceful one facilitated by contacts with Turkish elements around the Caliph, and though he had to fight later, by 1059 Tughril was master of the Caliphate and enjoyed the title of Shah.14 In the process of constructing his power in the heart of Islam Tughril and his successor, Alp Arslan, were happy to adopt the composite armies of their predecessors in which the tribal element of the Turks was only a part, though Turkish enlistment as slave-soldiers, mamlūks, continued to be important. This was a vital element in the stabilisation of their dynasty and as a corollary they encouraged the nomadic Turks to attack the Fatimids of Palestine and the Byzantine frontier. Patronage of such a holy war would give the Shah prestige and allow the tribes to plunder, while providing a reservoir for recruitment. The scale of their success was remarkable. In 1057 they sacked Melitene (modern Malatya), in 1059 Sebasteia (modern Sivas) and in 1064 Ani and by the late 1060s they were virtually raiding at will in eastern Asia Minor, even devastating the land behind the advance of imperial armies under Romanus IV Diogenes (1067–71) when he campaigned against them in 1069.15

Such success owed something to their tactics. Traditionally steppe people ride light ponies, perhaps ten to twelve hands on average, and depend on strings of them to provide speed and endurance in battle. We have very little information about Turkish horses, though in contrast to the Mongol armies we do not hear of large strings of spare horses. When Alp Arslan fought at Manzikert in 1071 he took with him 15,000 picked cavalry, and the fact that all had a spare horse is remarked upon by the sources suggesting that it was unusual and a mark of their elite status.16 In fact, for nomads the difficulties implicit in the raising of heavier breeds with the need for stall-feeding, stud farms and the isolation of dams were overwhelming.17 Their special forte was mounted archery and they provided extraordinary fire-power and accuracy combined with speed of manoeuvre. A ninth-century Arab writer remarked of the Turkish troops then becoming common: ‘The Turk can shoot at beasts, birds, hoops, men, sitting quarry, dummies and birds on the wing, and do so at full gallop to fore or to rear, to left or to right, upwards or downwards, loosing ten arrows before the Kharijite [Arab tribesman hostile to the Abassids] can nock one.18

This combination of speed of manoeuvre with the range of the bow was extremely difficult to combat. They used the composite bow – wood with horn which reacts to compression bound on the belly side, and sinew on the back to increase elasticity. Thus, despite its shortness, the nomad’s bow had great strength and considerable range. Albert of Aachen salts his chronicle with poetic cliches; his Franks seize helms and armour as they fly to battle, but the Turks seize bows, often described as being of horn and sinew. Such vivid word pictures convey sharply the difference in style of warfare.19 In a famous passage after his account of the battle of Dorylaeum, the Anonymous praises the skill of the Turks as soldiers, but Albert gives a very much more specific example of their archery. During the pursuit after Dorylaeum, he says, the enemy remained dangerous and often turned at bay. Gerard of Quiersy, spotting a Turk on the brow of a hill, drew across him and attacked with his lance; his intended victim, however, fired an arrow which went through the shield and struck him between the liver and lungs, and while he lay dying the Turk made off with his horse. As the Franks forced the crossing of the Iron Bridge he again mentions Turkish arrows piercing a hauberk.20 If we think of a Turkish horseman drawing a bow with a pull of between twenty-seven and thirty-six kilograms, he might have an effective range of well over sixty metres so the Anonymous’s comment on their ‘astonishing range’ makes sense. Consider the effect of many such individuals firing together and it is possible to understand the frequency with which the western accounts mention the sleets of arrows which the Turks produced in battle. Fulcher speaks of the clouds of arrows which overcame the army at Dorylaeum, and Albert describes hails of arrows as the Turks fought back in the pursuit after the battle and which destroyed Swein of Denmark’s reinforcements and Renaud of Toul’s force in the final battle at Antioch.21 Such language is too frequent to be mere extravagance, as are references to the fast horses of the Turks. Raymond of Aguilers describes their hit-and-run tactics in which speed was essential to avoid crusader retribution. Ralph shows them manoeuvring outside ‘Artāh to lure Franks into ambush, while according to Albert warriors on speedy mounts opposed the crossing of the fords near the Iron Bridge and later lured Roger of Barneville to his death.22 It is possible that they had developed light tubular crossbows, throwing darts to augment their firepower.23 It was by throwing a hail of arrows that the Turks demoralised their enemies, isolated and broke up their formations before charging in for the kill at close quarters with sword and spear. This was how they had destroyed the forces of the People’s Crusade. Once at close quarters their primary weapon was the sword, at this time a straight edged weapon rather narrower and more sharply pointed than the Frankish weapon, but otherwise little different.24

It is difficult to comment on the armour of the nomads. As early as 1037/8, Bayhaqi says that when Tughril Bey entered Nishapur his 3,000 cavalry were mostly armoured. Much of our evidence about the Seljuks comes from Byzantine sources of the twelfth century which show mail shirts and poncho-like garments of mail. By this time Greek influence upon their protective equipment and style of war was becoming very strong25. However, there is ample evidence for the use of armour, scale, chain-mail and lamellar armour throughout the Middle East. Strips or scales fastened to cloth or leather seem to have been outmoded in the West, but such lighter equipment would have suited the warfare of the area very well and so have remained in use. There is no reason to think that the nomads of Asia Minor were ignorant of armour. Ralph of Caen gives a very vivid description of the fighting at Dorylaeum and tells us that in the press of battle the Turks ‘trusted in their numbers, we in our armour’ which implies that they did not have armour. However, he does not explicitly say that they had none and it would be very surprising if they were prepared to close with the Franks with no protection at all. In much of the Middle East armour was worn under the cover of other materials, most often in the form of a Hazagand, a leather jerkin with mail or lamellar within and there were variations on this like the later western brigandine. Felt or fur caps were often worn by Turks at this time.26 We can think, therefore, of the Turkish horsemen wearing rather lighter armour than the western knights, and, above all, carrying a much smaller and lighter shield. However, their skilful horsemanship and tactics were not in themselves a sufficient advantage to account for the victories of the Turks in Asia Minor – they owed far more to Byzantine weakness.27

The death of Basil II (976–1025) saw the Byzantine empire in a position of strength, extending from Mesopotamia to Bulgaria.28 But Basil’s success, and the manner in which it was achieved, left considerable problems for the Byzantine state. He significantly changed Byzantine military organisation. In the seventh century the Byzantine army had been reorganised with the division of the empire into Themes, each of which was defended by its own military-force locally recruited, housed and financed. When an imperial expedition was mounted, the army of the thematacombined with the central army based at Constantinople, the tagmata, at selected aplèkta, camps on the military highways such as that at Dorylaeum which was the gateway to the Anatolian plateau.29 However, the settlements of peasant soldiers, which were the basis of this army, were being absorbed into the estates of the aristocracy of Asia Minor who thus were able to control the armies of the Themes. In 987 a group of these families revolted against Basil and were only defeated in 989 with the aid of a corps of 6,000 Russians, the basis of the later Varangian guard. From this time onwards Basil found it prudent to rely more and more on mercenaries for the regular army which was the basis of his successful expansion of the empire in the Balkans. The increasing wealth of the empire, evidenced by the growth of cities, facilitated the replacement of territorial forces by mercenaries.30 As a result, the old armies of the Themes were somewhat neglected, except on exposed frontiers where they continued to serve a useful purpose. At the same time, he revived old legislation to prevent the great families of Asia Minor from absorbing peasant holdings into their estates in order to protect the tax base which was even more necessary to pay his professional armies. On the basis of his victory of 989, Basil established a harsh government which dominated the great noble houses and at the same time demanded a crushing taxation to support a professional army and a policy of expansion. The strains that this imposed on the empire meant that when Basil died there was bound to be change, and this was complicated by the lack of a clear line of succession on the death of his brother Constantine VIII (1025–8) whose daughters’ marriages determined the succession for the next twenty years without ever producing heirs to prolong the Macedonian house. This failure inevitably produced uncertainty and promoted the emergence of aristocratic factions who competed for power; there were thirty rebellions in the period 1028–57, such as that of Tornikios in 1047–8 which was only defeated by stripping the frontiers of troops.

Competition for power amongst aristocratic houses was a feature of Byzantine history, and this period produced able, as well as feeble, emperors, but this came at a time when the empire had severe problems. The rise of a mercenary army imposed massive financial burdens and it is hardly surprising that, with the less expansionist stance after 1025, efforts were made to cut back on military expenditure. Constantine IX Monomachus dismantled the army of the theme of Iberia. The policy was logical for the local army duplicated the mercenary one, but, to its great discontent, this area now had to pay taxes and it was exposed to hostile attack. In the East, the Turks began to press on the frontier, while in the Balkans the Patzinacks were a potential menace and in Italy the Normans became a major force in the 1040s. Thus, economy in military matters came at a time of increasing external threat. Moreover, in the east, there were a number of populations whose religious affiliations made them suspect to the Byzantine authorities (most notably the Jacobite Syrians and the Armenians), and the efforts of Constantine X Doukas (1059–67) to settle problems with these churches, intended to strengthen the frontiers, simply aroused hatred. Thus, every effort of the imperial government worsened the security situation in the eastern provinces and, by the 1060s, the Turks were raiding deep into Asia Minor with Iconium and Chonae falling victim towards the end of the decade. Even more importantly, the vacuum of power at the centre continued to be filled only intermittently and the autocracy was the subject of bitter factional conflict which prevented consistent policies from emerging. The factional struggle allowed other forces to emerge within the state – the church especially under the Patriarch Michael Keroularios (1043–58) became a major factor in the state and not a mere arm of the autocracy, while the mob of Constantinople had also to be considered. This was the background to the reign of Romanus IV Diogenes (1067–71).

When Constantine Doukas died he had invested imperial power in his wife Eudocia, to exercise on behalf of his young sons, but there was widely felt to be a need for a strong ruler and as a result, she married Romanus IV Diogenes, a successful soldier who had defeated the Patzinacks in the Balkans. He consolidated his power and produced two sons, but inevitably this aroused the hostility of the Doukas.

Romanus was determined to end the Turkish attacks on the eastern provinces, but there was considerable uncertainty about how best to do this. A considerable body of thought suggested that he should make the border provinces, with their disloyal populations, a desert and crush the Turks in central Asia Minor. He preferred to try and oust the Turks by a strategy of large-scale expeditions to the eastern frontier which would put pressure upon the Sultan Alp Arslan. But the Turks were steppe horsemen to whom the Anatolian plateau presented a congenial and familiar habitat across which they could move quickly and they melted before the lumbering Byzantine army, returning to isolate the fortresses once they had gone.31 In 1068 Romanus attacked Aleppo, and the campaign of 1069 was directed against the upper Euphrates although the rebellion of the Norman mercenary Crispin was a considerable diversion. These expensive forays produced no results and it is likely that the emperor’s enemies, led by the Caesar John Doukas were becoming a threat. In 1071, Romanus led a huge Byzantine army with the intention of bringing the Sultan Alp Arslan, who was preparing for an attack on Fatimid Egypt, to battle: Romanus needed a victory. The army was overwhelmingly mercenary with contingents of Greeks, Russians, Khazars, Alans, Georgians, Armenians, Turks and Franks; it was the Germans, amongst these latter, who attacked the emperor at Cryapege when he tried to curb their excesses. Some of the units were very good but there was enormous variation in quality and they were not used to working together. As they moved eastwards there was friction with the Armenians. It was a huge army, though the 300,000 suggested by some Arab sources is a gross exaggeration: it was probably of the order of 40,000–60,000. Alp Arslan was surprised by their coming – he had negotiated a truce with Romanus in the previous year, but the Byzantines saw continued Turkish raiding, over which the Sultan had no control, as a breach. Manzikert (east of modern Erzerum (see fig. 3)), a fortress recently captured by the Turks, was quickly recaptured, but a large section of the imperial army was dispatched to take Chliat. When Alp Arslan arrived, the Norman mercenary leader, Roussel of Bailleul, and Tarchaniotes who commanded many of the Turks, simply fled. In the crisis of the battle Andronicus Doukas seems deliberately to have betrayed Romanus who was captured in the ensuing rout.32

Manzikert was not quite the overwhelming victory that has been supposed, for much of the army was never engaged and many units escaped intact, but the emperor’s guard was slaughtered.33 Alp Arslan concluded a very merciful treaty with Romanus, for he was preoccupied elsewhere and had no wish to see a major Turkish power in Anatolia. It was the Byzantine reaction which turned the situation into a disaster. Romanus’s enemies, led by Michael VII Doukas (1071–8), denounced the treaty and blinded Romanus. The Byzantine state now dissolved in a series of civil wars, in which the numerous contenders were all prepared to call in the Turks. This was why the cities of Asia Minor which could have resisted a nomadic people with no experience of siege warfare did not do so. Instead, the keys of their gates were handed over to the Turks by the contending magnates. In the chaos, the Norman mercenary leader Roussel of Bailleul attempted to create a new state in Asia Minor and his success brought 3,000 Franks into his following. In an effort to divide his enemies he championed the imperial pretensions of John Doukas. He was only defeated in Bithynia when the imperial authorities brought in the Turkish Emir, Artuk, whose activities so far to the west are a revelation of Byzantine weakness. Even then, he escaped and in the end was betrayed to Alexius Comnenus.34 In Cappadocia and Cilicia, independent Armenian princes were happy to see the back of Byzantine rule and to take over the cities of the area. Amongst them was a former Curopalate and Domestic of Romanus IV who had fought with him at Manzikert, Philaretus Brachamius, who refused to recognise Michael VII and created a principality based on Marasch, Ahlistha and Melitene (see figs. 2 and 4). In 1074 he defeated Isaac Comnenus, duke of Antioch, and by 1078 he had acquired this major city. Nicephorous Botaneiates, who rebelled against Michael in 1078, was a former comrade in arms of Philaretus and recognised his independent principality to which he added Edessa in 1083–4. When Sulayman, leader of the Anatolian Turks, was peacefully admitted to Antioch in 1085, it is not impossible that local factionalism within the city was encouraged to this end by Alexius I Comnenus – in order to get rid of an adherent of the overthrown Botaneiates.35 When Nicephorous Botaneiates began his rebellion in 1077 he was able to gather very few troops but the government, beset by the rebellion of Nicephorous Bryennius in the west which allowed the Patzinacks in once again, called upon Sulayman, leader of the Anatolian Turks. He changed sides, however, and Botaneiates used his forces to hold down many of the cities of western Asia Minor, including Nicaea which appears to have fallen into Turkish hands at this time. Much of the army which Botaneiates used against Bryennius in the west under Alexius Comnenus was Turkish, and it was turned against Botaneiates whom Alexius succeeded in 1081, only to face the rebellion of his brother-in-law, Melissenus, who turned over more of the cities of western Anatolia to Sulayman. There can be no doubt that for the Byzantine magnates and generals, civil war in their own interests was far more important than defence of Anatolia. It seems almost as if these great nobles were quite happy to loan a vast and rich province to barbarians for whom they had the greatest disdain and whom they never seem to have regarded as rivals, like the Arabs against whom Nicephorous Phocas (963–9) and John Tzimiskes (969–76) had waged a holy war.36 It was the Emperor, Alexius Comnenus (1081–95), who finally brought to an end the chaos in the Byzantine state and established himself as the head of a group of aristocratic families who dominated the machinery of state. He was not a great soldier: he would suffer heavy defeats, such as that at Dyrrachium, and win few real victories. He was as prepared as any of his rivals to make arrangements with the Turks. He was, in the end, a skilful and cautious politician anxious to nurse his deeply wounded empire. The spirit of the crusade was deeply alien to this cautious politician. He ‘came to power as the head of a powerful aristocratic network’, so it is hardly surprising that he offered no revival of that spirit of holy war which a century before had enabled the warrior emperors, Nicephorous Phocas and John Tzimiskes, to drive the Byzantine frontier deep into Syria.37 But the Comneni and the families allied to him were all from Asia Minor which they regarded as the heartland of the empire, hence the importance which he attached to regaining it, and hence the appeal of 1095.

Anna Comnena clearly understood the importance her father attached to Asia Minor. She portrays him as a Byzantine hero, but his relations with the Turks were complex. There was never any question of holy war, on either side. It was during his reign that Byzantine sources began to refer to Sulayman of the Seljuk house and leader of the Turks of Anatolia as Shah – a title not conferred by either the Sultan or the Caliph at Bagdad.38 Sulayman was killed in 1085, and it was not until his son Kilij Arslan escaped from the Sultan’s custody in 1092 that the Sultanate of Rhüm, as it came to be called, could rise again. The Turks remained a largely nomadic people and their dominion consisted of garrisons in the cities and control of the key routes. Kilij Arslan held the important cities of Nicaea and Iconium, but on the Aegean coast were emirs like Chaka at Smyrna, and Tangripermes at Ephesus, who were at best his allies and at worst his rivals. Cyzicus, on the Propontis, was held by another emir, as were some of the ports along the Black Sea coast. At Erzerum the Saltukid Turks had established a dominion as had the Menguchekids at Erzinjan. Further south, the Danishmends had carved out a great principality based on Sivas, Kayseri and Ankara. Then there were lesser powers like Baldaji or Hasan, who ruled an enclave including much Armenian territory in Cappadocia.39 Underlining the precarious nature of these conquests were the independent Armenian princes in the Taurus range, prominent amongst whom were Constantine son of Roupen, Pazouni and Oschin, Gabriel a former associate of Philaretus who held Melitene and Thoros of Edessa.40 The Turkish powers enjoyed a very uncertain relationship with the centre of power at Baghdad where Malik Shah (1072–92) ruled. When Sulayman had threatened to become a power in Syria after his conquest of Antioch, he was defeated and killed by the Sultan’s brother Tutush in 1085 and Antioch was absorbed into the lands of the Sultan Malik Shah. Faced with this complex situation, Alexius proceeded carefully. He sent forces against Nicaea in 1081 but the attack of Guiscard in the west forced withdrawal, and he was unable to take real advantage of the weakness of the Sultanate of Rhüm on Sulayman’s death because of threats in the Balkans from the Patzinacks and others. In 1086 Malik Shah had been prepared to consider an alliance against the Turks of Asia Minor, which would have cleared the western part of the peninsula, and at the time of his death, was negotiating with Alexius for an imperial marriage which might have opened the way for an alliance against Rhüm.41 In the rivalries of the rulers of Asia Minor and the tensions between them and the Sultan at Baghdad Alexius could see opportunity which was, if anything, increased by the deaths of Malik Shah in 1092 and his brother Tutush in 1095, leaving Syria divided between the latter’s bickering sons, Duqaq of Damascus and Ridwan of Aleppo.42 It was presumably in an effort to exploit this complex situation that Alexius had asked Urban II for aid. He was presented with an independent force, the crusade, which, despite his best efforts, was not entirely within his control. His refusal to join the army as it marched against Nicaea, and indeed his whole policy towards the Franks, has to be seen in the light of this situation. To back the Franks so unequivocally as to join them in person would make relations with Rhüm difficult, should they fail. And failure might trigger internal unrest. Better by far to leave them to fight, so that, if they failed, other means could be pursued and his relations with Kilij Arslan could remain. Alexius had not attacked Nicaea in person before – he had used his generals and now he continued this policy. Alexius would support the Franks as long as they succeeded. With this equivocal ally a comfortable distance behind them, the crusader army prepared for battle against the first of its enemies – the Seljuk Sultanate of Rhüm and its principal city of Nicaea.

It is very difficult to estimate the military capacity of the Sultanate of Rhüm in this period. The Turks had never been numerous – Cahen suggests that only about 20,000–30,000 warriors entered the Caliphate at the time of its conquest.43 As steppe people like the Huns, Avars and Magyars before them, they relied on mobility and a training in horsemanship from birth. It has been suggested of earlier nomad peoples who entered Europe that to maintain their speed of assault each rider needed a string often horses. Therefore, nomad forces needed huge open ranges to graze their ponies. The Hungarian plain, where the Huns, Avars and Magyars settled, could only support 150,000 horses, an army of 15,000 men.44 We simply do not know if the Turks used such vast strings of spare horses but is likely that they did not. Much of the once rich agriculture of Asia Minor had been destroyed by the Turkish invasions, but there was still a large native population with settled cultivation, while a lot of the land was wild and mountainous. It certainly could not support horses on anything like the scale of the Syrian and Mesopotamian plains where later the Mongols had difficulty in maintaining their huge horse trains.45 It is only possible to guess, but it seems unlikely that Kilij Arslan could find as many as 10,000 Turks, even with the allies he brought to the field of Dorylaeum. It was not the vast numbers of the Turks that made them dangerous, but their sheer courage, ruthlessness and daring tactics which the crusaders themselves recognised. At the siege of Nicaea, Fulcher testifies to their savagery while the Anonymous author of the Gesta is lavish in his praise of their valour46. Their battle tactics are very well attested as we have noted – their reliance on mobility, seeking to surround their enemies and bombard them with arrows, to draw them into ambush by feigned retreat, breaking up their cohesion before venturing to take them on at close range. A passage from Nicephorous Bryennius describing what he believed happened at Manzikert nicely illustrates Turkish methods:

Taranges divided the Turkish army into many groups and devised ambushes and traps and ordered his men to surround the Byzantines and to discharge a rain of arrows against them from all sides. The Byzantines, seeing their horses struck by arrows, were forced to pursue the Turks. They followed the Turks who pretended to flee. But they suffered heavily when they fell into ambushes and traps. The emperor, having resolved to accept a general engagement, slowly advanced hoping to find an army of Turks, attack it and decide the battle, but the Turks scattered. But wheeling, with great strength and shouting, they attacked the Byzantines and routed their right wing. Immediately the rear guard withdrew. The Turks encircled the emperor and shot from all directions. They prevented the left wing from coming to the rescue for they got in its rear and forced it to flee. The emperor, completely deserted and cut off from aid, drew his sword against the enemy and killed many and compelled them to flee. But encircled by the mass of the enemy, he was struck in the hand and recognised and surrounded on all sides. His horse was hit by an arrow, slipped and fell, and threw down his rider. And in this manner the Byzantine emperor was made prisoner.47

This is a description that the crusaders would soon come to recognise. The Anonymous describes their surprise and dismay as the Turks, making an enormous and frightening noise, surrounded the army of Bohemond and poured arrows into it. Raymond of Aguilers commented of a later conflict: ‘The Turks have this custom in fighting, even though they are few in number, they always strive to encircle their enemy.’ Albert of Aachen emphasises, in an attack at Nicaea, that 10,000 mounted bowmen appeared and that just such men surrounded and broke into Bohemond’s camp at Dorylaeum.48 Turks formed an important element in all the armies which the crusaders faced, and indeed in the Byzantine army, but in Anatolia the crusaders were confronted by forces which were entirely Turkish, and they showed themselves keenly aware of the difference between them and other armies. Raymond of Aguilers speaks of Turks and Arabs, and Albert refers to Turks and Saracens, present in the army of Duqaq of Damascus, but both writers speak only of Turks in Asia Minor.49 The composite armies of the Sultan and his emirs were quite different from the forces of the Anatolian Turks who were the crusaders first enemy. In Anatolia, the Turks remained largely nomadic and their Seljuk Sultanate had not yet developed, as far as we can tell from inadequate sources, the kind of central administration which could control a composite army.50 This was the wild frontier of Islam and what confronted the crusaders was a brave, dangerous but not very numerous enemy. However, there must have been some diversity in the Turkish army, for Nicaea was strongly garrisoned and resisted bitterly. It is unlikely that these were simple mounted nomads – presumably some more specialised forces had been recruited to defend this important but exposed city, which the Byzantines had attacked more than once. In this connection it is perhaps important that Nicaea was the only city of Asia Minor to hold out. Iconium was not defended; at Hereclea the Turks tried to ambush the crusaders as they approached, then fled.51 Defence of the cities of Asia Minor was no easy matter for the Turks because these were still populated by Christians: the Turks had not captured them, they had been admitted, as garrisons, by feuding Byzantine lords or after a long period of isolation before a Turkish dominion outside their walls.52 The countryside was in the hands of the nomads but the cities were different. As soon as the crusaders began to win victories, the peoples of the cities along their route began to eject their Turkish garrisons. The Anonymous says that after his defeat at Dorylaeum Kilij Arslan had to pretend to have been victorious in order to gain admission.53 The attitude of the native population was to have an important influence on the crusade, as we shall see. It was certainly to have a great influence on the siege of Nicaea for, at the very moment that the army attacked it, Kilij Arslan was preoccupied with far-off Melitene.

In late October 1096 Kilij Arslan had totally destroyed the armies of the People’s Crusade. Leaving his family at Nicaea he set out to intervene in Melitene. This city was a vital communications centre on the roads from Anatolia to Mesopotamia and Iran. It was held by Gabriel, a former officer of Philaretus, who claimed the Byzantine title of Curopalate yet nominally held it of the Caliph.54 It was important for Kilij Arslan that Melitene should not be in the power of the Sultan. In 1097 the divisions of Syria and the rivalries with Baghdad, where the Seljuk Sultan Berkyaruk (1094–1105) was preoccupied with events in the East, offered a splendid opportunity to intervene, but he almost immediately found himself in competition with the Danishmends who also wished to control Melitene. Conflict was avoided for the moment because news reached the Sultan of the new threat to Nicaea, and he hastened westwards. It was perhaps easier for these nomads to move quickly than a conventional army, but this journey of not much less than 1,000 kilometres must have been very tiring. The crusaders reached Nicaea on 6 May 1097 but with only a part of their army. Bohemond and the Normans took up a position along the north wall of the city, with Robert of Flanders and Godfrey to the east. The south gate was left open for Raymond of Toulouse, whose delay at Constantinople we have noted: the North French had not yet arrived at Constantinople. The piecemeal nature of the siege underlines the lack of unity in the crusader force; it was a huge host made up of a number of major armies grouped around important leaders, but there was no overall command. In fact, they approached Nicaea from the north and simply fanned out in order of arrival, probably arranging details by consultations amongst the princes. Kilij Arslan arrived in the general area of his capital shortly before 16 May when his attack precipitated a major battle.

We have two versions of the nature of this attack. According to Raymond of Aguilers it was two-pronged: one force fell upon the Germans on the east side of the city, while the other attempted to enter the city through the vacant south gate, with the intention of sallying out against Godfrey while he was distracted. According to this version, the Provençals happened to come before the south gate and were pitching camp when the enemy arrived; they fought off the southern attack, thereby enabling the Germans to fight off the other force. This account gives the South French a beau rôle indeed, and one wonders just how Raymond could have known of the intentions of the enemy. The Anonymous makes little of the affair, saying that Count Raymond fought off an initial attack which was renewed but defeated ‘by our men’. Albert says that, alerted by the capture of an enemy messenger trying to reach the garrison, the leaders asked Count Raymond to hasten his march, but agrees with Raymond of Aguilers that the Provençals were attacked just as they were making camp. He says that 10,000 enemy archers fell upon the southerners and that the Germans, supported by the Normans of Bohemond, then attacked the enemy who were put to flight.55 This version is much the more convincing. The enemy attack clearly came from the south; the Anonymous explicitly states that the enemy came down from the hills, and Albert confirms this.56 From this location the Turks would have had a magnificent view of Nicaea and the basin which surrounds it and so could not have missed the slow progress of the Provençal forces round the city to the south gate. Clearly the Turks chose to attack when they were most vulnerable, as they prepared their camp after the forced march to the city. Kilij Arslan hoped to brush them aside, and at the least reinforce Nicaea, at the most inflict a discouraging defeat on the westerners (see fig. 5). The attempt failed because the Provençals put up a stiff resistance (and to Raymond of Aguilers they must have seemed to have been at the very centre of the affair), drawing the Turks into a close quarter battle and so giving time for Godfrey’s attack from the east on Kilij Arslan’s right flank. The sheer numbers of the crusader army were decisive in the narrow area between the wooded hills and the city wall because the Turks had little room for manoeuvre. The Anonymous, who gives the impression of a skirmish, was probably with the Normans to the north of the city. Albert makes it clear that it was a savage and close-fought battle with heavy losses on both sides. There was no overall command on the crusader side but, nonetheless, we can see generalship of a very high order at work. The count of Toulouse held his troops together at a difficult moment as they were making camp, while Godfrey seems to have rallied his forces to their relief quickly. These are not small achievements, especially when one considers the looseness of command and the uneven quality of the western forces. For many this must have been the first experience of battle, and for others their first of anything on a large scale. It was essentially the mass of the crusader army operating in a confined space which frustrated Turkish tactics and drove off Kilij Arslan, but in the circumstances, the cohesion of what must have been pretty green troops in the individual armies within the host was remarkable. Anna Comnena is quite right to speak of the Franks winning a ‘glorious victory’.57 Afterwards, the crusaders stuck the heads of the enemy dead on lances, and sent others to Alexius as tokens of victory. They were now free to besiege the city as Kilij Arslan fell back to rally more troops.

Most of the accounts we have of the siege of Nicaea are quite brief. The Anonymous says that when the crusaders first arrived, and even before the coming of the Provençals, they built siege machinery including towers and undermined the wall, but this was interrupted by the Turkish attack. After the defeat of Kilij Arslan he tells us that the count of Toulouse and Adhémar of Le Puy set troops protected by crossbowmen and archers to undermine a tower, which duly fell, but so late in the evening that the enemy were able to refortify the gap. Thereafter, it was the boats sent by the emperor to blockade the Ascanian lake at the west end of the city which forced a surrender. Raymond of Aguilers mentions fruitless efforts to storm the walls and the building of unspecified machines. He reports the same story of the undermining of a tower by the Provençals which came to nothing and stresses the importance of the boats which brought the siege to an end. This is very much the story told by Anna Comnena who says that the Count of Toulouse built a wooden tower on whose upper stories men engaged the enemy, while others below undermined what she calls the Gonatas tower, but in her account this simply has no outcome. She praises her father for providing the Franks with designs for machines and the boats on the lake, adding much detail on the negotiations for the surrender of the city.58 Because so many latin sources based themselves on Raymond and the Anonymous they tend to add little. Baudry of Dol gives a few names of participants and stresses losses in the army. The Historia Belli Sacri says, after the story of the Provençal tower, that all the leaders made machines and Robert the Monk mentions the building of wooden towers. Fulcher of Chartres gives a generally vague account but includes a list of the many siege machines used.59 The reasons for this brevity are clear; Raymond highlights the doings of his Count, while the Anonymous’s master Bohemond does not seem to have had a lot to do. But there are hints of a much more intensive siege and the account of Albert of Aachen makes it clear that the Franks went to great lengths to assault Nicaea with elaborate machinery.

Albert does not mention the early assaults on the city before Kilij Arslan’s attack and his dating is obscure. He says that it was only after seven weeks of siege that the leaders set in train the construction of catapults and assault equipment. The primary element in the assaults seems to have been the penthouse, a wooden structure with an armoured sloping roof within which attackers could undermine the wall in relative safety.60 Albert mentions an assault in which Baldwin Calderin and Baldwin of Ghent were killed, and another in which the count of Forez and a knight called Guy died. Then, on a day during which the walls were under attack by crusader machines, two men in the force of Godfrey, Henry of Esch and Count Herman, built a penthouse which they called ‘the Fox’, which was brought up against the wall with enormous labour, but it collapsed killing all twenty knights within it, though not the originators of the project who refrained from trusting their lives to the device. Such machinery required careful design and construction skills which were evidently rare, as we have noted from Ordericus’s story about Robert of Bellême.61 The next major assault which Albert mentions was launched by the count of Toulouse whose forces, covered by the fire of mangonels, crossed the ditch protected by a testudo, the same word as used by Raymond of Aguilers, and assaulted a tower. However, the enemy built a wall of stone within the tower, frustrating the attack which had to be broken off. Albert goes on to tell how boats were brought up to blockade the Ascanian lake and says that Raymond then renewed the attack. This time the Turks burned the equipment which brought the wooden penthouse and other instruments forward, and then repaired the wall which had been breached during the night. When the attack was resumed the next day only a single Norman knight could be found to press it; he was killed and his body dragged up the walls by the defenders and left hanging there. This account broadly corroborates that of Raymond of Aguilers and makes it clear that Raymond’s testudo was a penthouse.62 All these assaults were causing heavy losses which worried the leaders, especially as the catapults were having no effect on the walls. Then a Lombard engineer offered to build a machine if the leaders would finance him; they agreed to pay him fifteen pounds in the money of Chartres (where in the twelfth century thirty would buy a fine house) from their common fund.63 This first mention of the common fund points to the development of rudimentary organisation to sustain the siege. In fact, a properly built penthouse was constructed and pushed across the ditch up to the wall which was undermined and propped with wood. These props were fired and in the middle of the night the upper part of the tower fell. This frightened Kilij Arslan’s wife who attempted to flee across the lake but was captured, while the garrison of Nicaea decided to surrender.64 Albert’s account of the siege fills out considerably the rather schematic view given by the other sources, though it is chronologically confused and it is likely that he was attempting to conflate the stories of several individuals. What it does not make clear is the importance of the boats provided by Alexius, which is very evident in the other schematic accounts. This new attack from the lake, coming shortly after the arrival of the North French on 14 June, effectively doubled the length of the walls which needed to be defended as well as completely isolating the garrison, and was probably the decisive factor in precipitating their surrender on 19 June after another Frankish assault, under cover of which the Byzantines implemented the secretly-agreed surrender arrangements.65

The army had a considerable knowledge of siegecraft and we can discount Anna’s view that Alexius invented machines for them. They prosecuted the siege vigorously and suffered heavy casualties which worried the leaders; of the thirteen dead named by Anselm of Ribemont, two died in battle and three of disease during the siege of Nicaea. Albert’s mention of a common fund indicates that, although the armies in the host were grouped round several leaders, the need to cooperate was forcing organisation.66 The army must have relied on the Byzantines for supplies – wood, clamps, nails etc. and certainly it was Alexius who provided the boats which closed the Ascanian lake.67 The major problem of a besieging army, especially one this size, was food, and both Albert and Fulcher stress that Alexius sent this in good quantities, although the Anonymous remarks that some of the poor died of starvation.68 By and large the alliance had worked well in a military sense. The surrender of the city came as a surprise to the crusaders who must have sensed the intrigue from which they were excluded, but the emperor seems to have been reasonably generous in distributing the spoils of war to the westerners. Stephen of Blois tells us that Alexius sent food for the poor during the siege and agrees with Anselm of Ribemont and the Anonymous that he was subsequently very generous to the knights and princes. Only Raymond of Aguilers complains about this and his general attitude is deeply hostile to Alexius. The freeing of the Turkish garrison, however, deeply disturbed the Anonymous who feared they would later attack the Franks.69 The military value of the Byzantine alliance had been clearly demonstrated, and they prepared to march into Anatolia with an imperial contingent commanded by the Turk Tatikios.

However, this was not the only military assistance which they received. Anna reports that Alexius warned them of Turkish tactics, but he seems to have provided them with other information and ideas.70 During the siege Alexius had observed events from nearby Pelekanum and after the fall of the city he met most of the leaders, presumably to discuss strategy. It may well be that this followed up earlier discussions, of which we hear nothing. According to the Historia Belli Sacri he suggested that they send an embassy to Egypt seeking the friendship of the ‘Emir of Babylon’. It was as a consequence of this that an Egyptian embassy came to the siege of Antioch, happily at the very moment when they inflicted a heavy defeat upon the Turks at the Lake Battle in early February 1098.71 The encouraging noises made by these envoys probably exercised a considerable influence over the leaders in the summer and autumn of 1098. It was a skillful piece of diplomacy, reflecting Alexius’s intimate knowledge of the politics of the Middle East. The decay of the Abassid Caliphate in the later ninth century and the ensuing disorders enabled the dissident Shi’ites to establish a Caliphate of their own in Tunisia in 909, and from there they grasped Cairo in 969 where they set up the Fatimid Caliphate. The Fatimids sought to expand their control over Syria, but the restoration of Abassid power under the implacably Sunnite Seljuks after 1055 threatened these new conquests. In 1060 serious internal conflict broke out in Egypt amongst the diverse elements of the army which, on the pattern of the other Islamic powers, was a composite of peoples, in this case Berbers, Sudanese, Africans and Turks. By 1077 an Armenian general, Badr al-Jamali, was able to restore order but revolt and Seljuk intervention meant that Egyptian power in Syria and Palestine was confined to the cities of the coast, and by 1079 Malik Shah’s brother Tutush held Damascus and was overlord of Jerusalem, held of him by Artuk. The fragmentation of the Seljuk Sultanate after the death of Malik Shah in 1095, offered the Egyptians an opportunity to recover their lost dominion in Syria and Palestine (see fig. 3).72 Badr al-Jamali’s son al-Afdal saw the crusade as offering golden opportunities. The proliferation of tribes and powers in the Middle East meant that the precise nature of the crusaders’ interests were not perceived by the Islamic powers – they were simply another factor in a complex game, to be used, allied with as self-interest dictated. Alexius shared this mentality, and the crusader leaders were eager to capitalise.

In their discussions with Alexius at Pelekanum and before, the princes must have discussed the coming journey. Stephen of Blois did not join the other leaders at Pelekanum but he must have known of their discussions and in a letter to his wife written from Nicaea he refers to Antioch as their next target. The crusaders must have been aware from pilgrim days of the importance of this city which lay firmly across the road to Jerusalem. But they did not intend to conquer all the cities between Nicaea and Jerusalem, and Stephen’s letter holds out the possibility that, Antioch might not resist.73 This surely reflects knowledge of the situation in Syria. The death of Malik Shah in 1092 precipitated a bitter succession conflict between his brother Tutush, who held Syria, and his son Berkyaruk. When Tutush was killed in 1095 Syria was divided between his sons, Ridwan of Aleppo and Duqaq of Damascus. Malik Shah’s governor of Antioch, Yaghisiyan, was able to achieve much independence (see fig. 3). For Alexius the reconquest of Antioch was an alluring possibility. Sulayman of Nicaea had attempted to seize the lands of Philaretus in 1086 and had died at the hands of Malik Shah for his pains. The old duchy of Antioch stood between Anatolia and Syria and within striking distance of the great route centre at Melitene. It offered considerable opportunities to any power of Asia Minor. To the crusaders it was important to have it in friendly hands as they entered Syria and Palestine, the real object of their quest. Between the Byzantines and the crusaders there was a considerable community of interest.

But the crusaders appear to have been aware of other factors in the political situation of the lands they were entering. In the Taurus area there were a number of independent Armenian princes amongst whom Thoros of Edessa was very important. Oschin, who claimed to be descended from the Arsacids, held the castle of Lampron and was loyal to Alexius whom he served as governor. He seized part of Adana as the crusaders approached.74 Constantine, son of Roupen, claimed to be a descendant of the old Armenian ruling family of the Bagratids and held the fortress of Partzapert near Sis. Tatoul was at Marasch, Kogh Vasil at Raban and Gabriel at Melitene.75 These princelings throve in the complex politics of the area, playing off the Turkish emirs of neighbouring cities. It is clear that the crusader leaders had heard about them, for Matthew of Edessa says that they wrote letters to Thoros and to Constantine son of Roupen.76 It is probable that such matters were discussed with Alexius who perhaps suggested a course of action to take advantage of the situation. The Armenians had a tradition of hostility to Byzantium, as we have noted. Constantine, son of Roupen, was particularly hostile but on the other hand Oschin was friendly. Furthermore, both Gabriel of Melitene and Thoros of Edessa claimed to be imperial officials – Curopalatoi; the latter, we are told, was ‘expecting to hand it [Edessa] over to the emperor’.77 It is possible that the Armenians themselves made contact with the Franks but were this the case Alexius would surely have wanted to control subsequent events. When Tancred entered Cilicia and appeared before Tarsus in late September 1097 he was met by an Armenian, who was already known to him and had resided with him, who offered to attempt to negotiate the surrender of the city. At Nicaea Baldwin of Boulogne, Godfrey de Bouillon’s younger brother, had made the acquaintance of Bagrat, brother of Kogh Vasil of Raban. Baldwin also entered Cilicia, but on Bagrat’s urging left for Ravendan and the great adventure which eventually made him lord of Edessa.78 The hope of support from such eastern Christians was probably fed by the uprisings in the cities of Anatolia after the crusader victory at Dorylaeum, and it seems likely that it had a profound effect on crusader policy. What we have to see at this stage is that the crusaders probably knew a great deal about the lands into which they were venturing. Norman and Frankish mercenaries had long served in the Byzantine armies. Roussel of Bailleul, Crispin and, before them, Hervey had held land in the Armenian theme. William of Apulia wrote his Gesta Roberti Wiscardi as the crusaders left for the East and could give a good account of the battle of Manzikert, presumably from Norman veterans of the Byzantine service.79 So Alexius had no monopoly of information, but for the moment he and the Frankish leaders enjoyed a community of interest, but it was not one to which he was willing to contribute more than a few troops under Tatikios, charged with guiding the crusaders and taking over any cities they might capture. Anna’s caveat, ‘if indeed God granted them that favour’, probably reflected Alexius’s thinking. He would take what profit he could without heavy commitment, for there was much danger ahead for the expedition. They were not venturing into the unknown, merely into a dangerous hinterland that had been Turkish now for a generation.

The crusader leaders acted quickly. Nicaea fell on 19 June. On 26 June the first contingents left Nicaea, amongst them the Normans of South Italy. Various groups left subsequently, the last being the Provençals on 28 June and the army gathered at a place where there was a bridge, which Anna Comnena identifies as Lefke, about twenty-five kilometres east of Nicaea. A number of crusaders had stayed behind at Nicaea and took service with the emperor, while Anselm of Ribemont was sent to the imperial court by the leaders in order to settle outstanding business.80 They had already decided to go to Antioch, so necessarily they had to direct their path towards the old Byzantine fortress at Dorylaeum (Eskişehir) which was the gateway to the Anatolian plateau. The sources are quite clear that in the two days of march after the concentration of the army they broke into two groups, a vanguard and a main force. Raymond of Aguilers says that this happened after one day’s march, which suggests that the Provençals had left Nicaea a day later than the first contingents. We know how they divided; the vanguard was led by Bohemond, Tancred, Robert of Normandy and Stephen of Blois, probably fewer than 20,000 in all. The second, larger force, comprising the rest of the army was under Robert of Flanders, Hugh of Vermandois, Godfrey de Bouillon and Raymond of Toulouse, – rather more than 30,000 strong.81 It is more difficult to suggest why this happened. Fulcher, who was in the vanguard, simply confesses that he does not know; the Anonymous says there was confusion in the dark as the army left its place of concentration, while Raymond of Aguilers says that it was the fault of Bohemond and his companions who rushed on rashly (temere). Albert of Aix says that it was the result of a deliberate decision of the princes who after two days of marching the army together, now felt the need to divide it for foraging. Ralph of Caen tells us that some thought the division deliberate, and specifically denies this, which suggests that even after the crusade the matter was still being debated. It is likely that sheer size and the lack of any overall commander were the real reasons. The army of Frederick Barbarossa on the Third Crusade was 100,000 strong and seems to have taken three days to pass any single point. The sources for the battle of Dorylaeum make clear that most of the casualties were suffered by stragglers between the two forces, which would suggest that the host became strung out simply as a result of the natural frictions of the march.82 The disagreements and uncertainty of the three eyewitnesses – Raymond with the main force, Fulcher and the Anonymous with the vanguard, support this view. It also reflects the incoherence of the crusade’s command arrangements. It is worth remembering that the baggage train of Peter the Hermit’s much smaller force straggled a mile along the road and that the crusader army at its maximum strength was well over twice that size. But perhaps the leaders conferred at some point and gave their blessing to a division already becoming apparent. At the time of the battle Raymond of Aguilers says quite clearly that the two parts of the army were two miles apart – over five kilometres.83

The crusaders had now begun a march which would result in what is conventionally called the battle of Dorylaeum, for Anna Comnena says that it took place when Kilij Arslan ambushed Bohemond and the vanguard ‘on the plain of Dorylaeum’. In a letter of the leaders to the West on 11 September 1098, they referred to the battle at ‘Dorotilla’ which sounds very like the same place. One manuscript of the chronicle of Raymond of Aguilers refers to the battle ‘in campo florido’. Albert says that the battle took place ‘in vallem Degorganhi’, now called the Orellis, but later has Bohemond’s messenger to the other leaders say that the enemy attacked down the Orellis into the Degorganhi: neither of these place names can be identified and Albert does later use the name Orellis to mean somewhere quite different.84 However, there are grave difficulties about the idea that the battle was fought at or near Dorylaeum. The Anonymous says that the army marched one day from Nicaea and encamped for two days by a bridge while all the contingents gathered, then marched for two days until the battle on the third day. Raymond of Aguilers says that on the third day after the concentration of the army they met the enemy. Anselm says that after a two day march they encountered the enemy on the morning of the third day which was ‘kal. Iulii’, 1 July; Fulcher confirms the date and confirms that the battle began in the morning.85 Thus the crusade began to leave Nicaea on 26 June and concentrated at a river crossing, from which it departed on 29 June. It then marched for two days and fought the enemy in the morning of 1 July. When we examine the distances and the likely rates of march of the crusader army it is evident that they could not have reached the close vicinity of Dorylaeum in this time. Anna Comnena says that the army concentrated at the bridge of Lefke, which probably means the bridge over the Göksu, a western tributary of the Sakarya Nehri. Nicaea to Lefke on the Roman road is twenty-five kilometres, and Dorylaeum another ninety kilometres. If, as has been suggested, the army marched south to the Göksu and crossed it in the vicinity of Yenişhehir (a distance of thirty kilometres) they still had to cover roughly the same distance to Dorylaeum. A study of the rates of march of the individual armies across Europe to Constantinople suggests that, in the most favourable circumstances, the forces of Godfrey and Peter the Hermit never did more than twenty-nine kilometres per day. The army which left Nicaea was much larger and lacked a clear overall command and is likely to have progressed much more slowly. Barbarossa’s army probably managed about twenty-nine kilometres per day in Europe.86 Even at these rates the army would have been about thirty kilometres short of Dorylaeum after two days of marching, but they were probably moving much more slowly for they were in the presence of the enemy and encumbered with a heavy baggage-train. We can reasonably accurately date the departure of the army from Dorylaeum and its arrival at Antioch as being 4 July to 20 October. In 105 days of marching (with fifteen days of rest) they travelled 1180 kilometres, an average of thirteen kilometres per day which the Chronologie of Hagenmeyer suggests varied between eight and eighteen kilometres. There is no point in seeking comparison with events after Antioch when the army was much smaller. Furthermore, the crusaders knew the enemy were about and this would have restricted their speed, even if the vanguard did push on somewhat. All this suggests that the battle could not have taken place more than forty kilometres, or just conceivably fifty kilometres, south of Lefke or the Göksu crossing. Hagenmeyer recognised the problem and suggested Bozüyük just over fifty kilometres south of Lefke and about the same from Yenişhehir. This is probably as far as the army could conceivably have reached and it certainly could be regarded as being in the valley of Dorylaeum, as suggested by the letter of the leaders. Runciman points out that a Byzantine road runs further north through Sögüt and enters the plain ten kilometres short of Dorylaeum, where he thinks the battle took place. However, as Runciman admits, although this road does cross rivers, the countryside was very steep indeed and this probably rules out any of these crossings. But more simply, this was most certainly further than the army could have reached.87 What is clear is that the battle took place in a wide valley, for Albert says that Bohemond’s force was well to the right of the main force as well as ahead of it. Moreover, there was a river, for Albert mentions streams and Ralph of Caen, whose description is detailed, says that it was fought after a river crossing. William of Tyre follows Albert for the most part but with some variations. He says that the army followed a river in the valley of Gorgoni, and that the main force was to the right of Bohemond’s, reversing Albert’s statement.88 Albert’s account of a battle fought where two valleys join, taken together with Raymond’s mention of the ‘flowered field’ and the general description of the battle, suggests that it was fought in open land on the road towards Dorylaeum, and the comments of Albert and Ralph indicate not far from a river crossing or crossings, although these played no role in the major action. In fact to understand the battle we need to understand fully the circumstances in which the army found itself, the country and its road system.

After the capture of Nicaea it is clear from Stephen’s letter that the leaders had decided to march to Antioch, and evidently they had decided not to take the coastal route. They also rejected the ‘Pilgrim Road’ due east from Nicaea via Iuliopolis (near the modern village of Çayirbano) and Ancyra (Ankara) down through the heart of Asia Minor and across the Cilician Gates to Tarsus.89 Instead they decided to mount the Anatolian plateau towards the Byzantine military station at Dorylaeum (modern Eskişehir) which, at 800 metres commands the obvious point of entry to the plateau via a broad valley the sides of which rise to 1,200 metres and beyond (see fig. 2). Because Anna Comnena mentions the bridge at Lefke it has been assumed that the host marched east from Nicaea up the gently sloping plain, over the watershed and into the valley of the Sakarya and then up that of its southern tributary, the Kara Su, to its upper reaches just north of Bozüyük, where the land opens out into the wide valley which leads to Dorylaeum. But it is difficult to believe that the army would have taken this route, for the valley of the Kara Su, even in its lower reaches, is very steep and difficult and at Bilecik enters a spectacular gorge before narrowing even further into a grim steep defile which would have formed a perfect ambush site. The Byzantine road forked at Bilecik providing a road via modern Sögüt to Dorylaeum, but this road too is dangerously scenic and offers no open sites until it is very close to Dorylaeum. It is far more likely that the crusaders marched south from Nicaea. The first stage of this journey over the Avdan Dagi, whose peaks rise to 835 metres would have been quite difficult but thereafter they could cross the Göksu in the vicinity of modern Yenişehehir. From there a Roman road crossed the Ahl Dag, which rise to 1000 metres and emerged into the broad valley above Bozüyük, roughly where the modern Ego road from Bursa meets route 650 from Bilecik, just south of the narrow gorges of the Kara Su and some three to five kilometres north of Bozüyük. While by no means easy this route is no longer and offered a much more open approach to the high plateau.90 It is very likely that it was at this junction of roads in the plain that the battle of Dorylaeum took place (see figs. 2 and 6). Albert clearly indicates that the site was where two valleys meet, and the open ground here is about the right distance from the crossing of the Göksu. Moreover, the Anonymous says that when the crusader force came it formed up to the right of Bohemond’s trapped vanguard – it was, therefore, from the right that the attack came. This is also the force of Albert’s insistence on telling us that the vanguard moved to the right of the main force and William of Tyre’s careful correction that they were to the left, which fits with the Anonymous’s account. Both are explaining the subsequent alignment of the battle.91This would fit with the suggestion made here that the crusaders approached along the gentle valley from the west and were ambushed by the Turkish army lying in the southern valley to their right. The logic of the battle is clear. Kilij Arslan and his Turks were returning to the fray. This time he had concluded an alliance with the Danishmend Emir and together they were ready to attack the Franks. They chose to do so on the approaches to the high plateau and at a point of maximum advantage where they could lay an ambush and destroy an isolated part of the crusader force before its main weight could be brought to bear. It was the strategy of the Nicaea attack, but this time in less confined ground where Turkish speed of manoeuvre could be maximised.92 The Turkish army was probably much smaller than the total force of the crusaders and so had to avoid direct conflict with the main force and defeat their enemy in detail. Fulcher’s 360,000, though supported by the Anonymous, is sheer fantasy. In the accounts of the Crusade of 1101 we hear of the 700 knights in the rearguard of the main Lombard army being savaged by 500 Turks, while the army which destroyed the Bavarian and Aquitainian army was only 4,000 in all.93 The Turkish force was entirely mounted and was probably roughly equal to the knights in the whole crusader host. Therefore, a battle of movement involving the cavalry element would nullify the huge numeric advantage of the western forces and, in the attack on the crusader vanguard, Kilij Arslan would actually outnumber the western knights. If the Franks had marched up the gorge of the Kara Su they would surely have attacked them there, just as they would later destroy the Byzantine army at Myriokephalon in 1176.94

On the evening of 30 June Fulcher and Ralph of Caen both say that the vanguard saw Turkish forces, substantiating intelligence which had already suggested that they were in the vicinity; this last comment suggests that Tatikios was with the vanguard, although no chronicler mentions him. Clearly at least, the vanguard, more than five kilometres ahead of the main force, were aware of the enemy presence.95 Albert of Aachen places the battle in the evening – starting as the army camped at the ninth hour, late afternoon. However, Albert here seems to be trying to make sense of his sources, hence perhaps his error on which side of the valley the vanguard was following, for his suggestion of an evening battle is connected with the act of making camp. But the Anonymous says that the battle raged from the third to ninth hour, and Fulcher suggests that the vanguard was on its own from the first to sixth hour (6–7am–noon). As these writers were actually with the front force they should be preferred, particularly as Ralph of Caen confirms their story that contact was made with the enemy on the evening before the battle and that the march was resumed the next morning when the crusaders were forced to pitch camp when it became apparent that a large enemy army was present. It was probably making sense of this sequence of events which confused Albert whose account, however, contains much valuable information.96 Fulcher’s account is peculiarly vivid for he was in the camp where: ‘We were all indeed huddled together like sheep in a fold, trembling and frightened, surrounded on all sides by enemies so that we could not turn in any direction’, while the Anonymous was with the knights of the vanguard who were outside the camp from which the women brought water.97 Ralph says that after an anxious night the army moved on and forced the passage of a river after which the appearance of the enemy compelled them to pitch their camp; Fulcher says they camped by a marsh which gave them some protection from the enemy and that later the enemy broke across the marsh. His account of murderous fighting in the camp is supported by Albert, who says that Robert of Paris died there trying to help the rank and file and adds the picturesque detail that young women tried to make themselves look beautiful so that they would be spared the sword. Ralph of Caen shows the knights depressed by their inability to save the others.98 Crusader sources therefore suggest two distinct actions within the battle. Fulcher speaks of the leaders fighting while those like him in the camp desperately resisted. Albert says that at the sight of the enemy Bohemond and the knights rode forward but were unable to prevent the Turks getting into the camp. Ralph tells us that when the camp was pitched the knights attacked the enemy, but were driven back in disorder and saved only by Robert of Normandy who rallied them with scornful words – subsequently they were involved in heavy fighting in which Tancred’s brother William was killed. The Anonymous says that when the enemy were sighted Bohemond ordered the foot to pitch camp and the knights to attack the enemy, and then makes it clear that the cavalry were driven back on the camp, for he says that in the subsequent fighting the women brought water to them. Raymond of Aguilers suggests that the camp was sacked by the enemy. Ralph says that thereafter the knights fought hard, commanded separately by Bohemond and Robert of Normandy, and appears to show these men imposing solid discipline upon their followers.99 The Anonymous tells us that from the first the vanguard was surrounded – ‘we are encircled’ he has Bohemond say – yet Fulcher speaks of a marsh on one side of the camp protecting them and the subsequent development of the battle was to the vanguard’s right. This can be explained by reference to the lie of the land. The convergence of the two valleys forms a natural basin against the northern rim of which Bohemond was pinned by the Turkish main force, but smaller troops of the enemy probably menaced from the surrounding hills, for the Anonymous mentions the enemy presence there.


Fig. 6a Bohemond is 5 km ahead of the main army in company with Robert of Normandy and the Counts of Blois and Flanders together with the Byzantines; having descended from Nicaea to the northwest they enter the main valley leading to Dorylaeum and see the Turks. Bohemond orders his foot to make camp quickly and throws forward his cavalry to protect them.


Fig. 6b The Franco-Norman cavalry is driven back on the camp, rallied by its leaders, and forms the outer shell of resistance in a ‘wearing-out fight’. The crusader army is surrounded, though partially protected by a marsh (location conjectural). They cling on, relying on their compact mass hoping for help from the main force.


Fig. 6c Godfrey and the Provençals of the main army arrive forcing the Turks to break off their attack and turn to meet the new threat to their left. The new arrivals form up to the RIGHT of Bohemond’s beleaguered force


Fig. 6d The Count of Toulouse enters the main valley through the Drumlins which mark its western shoulder, and his attack on their rear and left forces the Turks to flee leaving victory to the Crusaders.

Throughout the morning there was heavy and unpleasant fighting at close quarters. The western knights seem to have been pinned against the southern side of their camp holding off the Turks who, however, were able to penetrate from other sides despite the difficulties presented by a marsh on one side and the considerable resistance of the crusader footmen. About noon, after five to six hours of this bitter fighting, the knights of the main force came up to relieve their comrades. The Anonymous describes the formation of a battle line, but this is the tidiness of hindsight (see fig. 6c).100 The main force was probably well out of sight of the battle in the western valley and, although messages seem to have been sent back early, it was not until about noon that they appeared. This is not surprising, for the main army’s knights had to prepare themselves for battle and then to ride five kilometres along a road which was probably choked with transport and stragglers. It is unlikely that they had much time to form into line. Far to the right, the bishop of Le Puy seems to have charged behind a small hill and come upon the enemy now turning to face the new threat on their left, from the rear. At the convergence of the two valleys there are a number of glacial drumlins and one of these was probably the hill to which reference is made.101 There is no reason to believe that this was planned; rather a pell-mell battle developed in which skirmishes such as that in which Godfrey with 50 sodales attacked what they believed to be Kilij Arslan and his household on a low hill were the rule.102 A running fight ensued in which the enemy often turned to fight causing casualties like Gerard of Quiersy. The enemy’s camp was sacked and the nomads were pursued along the road so that, for two or three days after, the army passed enemy soldiers and horses fallen by the wayside.103 Casualties appear to have been heavy although how far we can regard Albert’s 4,000 Christians and 3,000 Turks as precise figures is a different matter. They do, however, sound small enough to be credible and large enough to suggest heavy fighting. Large numbers of the main force, the foot, the non-combatants generally and presumably some knights, were never engaged at all. It is interesting that Fulcher says that most of the casualties were those caught straggling between the two crusader armies, a comment substantiated by Raymond of Aguilers.104

Dorylaeum was a nasty experience for the crusaders. They were not caught totally by surprise in that they knew the enemy were near, but it is odd that the leaders in the vanguard did not warn the main force behind them. Presumably, they simply took it for granted that the enemy was around but could not guess that his main force was so close. It is unlikely that Kilij Arslan was ignorant of the whereabouts of the crusader main force. He attempted to destroy their smaller element in favourable circumstances, counting on numeric superiority to bring victory in a mobile battle over the knights in the vanguard. The crusaders were alert and their foot prepared to pitch camp while an element of the knights confronted the enemy and were put to flight, falling back on the camp where their solid formation, and the fact that the site was confined by the edge of the plain and a marsh, enabled them to resist the Turks. The Turks were drawn into close quarter fighting both against the knights and in amongst the tents and baggage. ‘The enemy were helped by numbers’, says Ralph, referring to the knights, ‘we by our armour’, which suggests that the knights adopted a solid formation and refused to be broken up by the enemy’s attacks with arrows and missiles. The stall-fed horses of the western knights may have been larger than the ponies of the Turks, and this weight advantage may have helped to solidify their resistance but, in general it was of no more use to them than it had been to the Byzantines. The western knights in the vanguard must have been quite helpless and the progress of the Turks in the camp would have destroyed their entire position, but relief came. Both sides seem to have been surprised by the enemy. The crusaders were appalled by the enemy tactics which struck the Anonymous as menacing and daring and Fulcher as totally new: ‘to all of us such warfare was unknown’. He was also struck by the fact that the enemy were entirely mounted: ‘All were mounted. On the other hand we had both footmen and bowmen.’ Albert of Aix remarks time after time in his account on Turkish use of the bow which clearly struck the crusaders as novel.105 But the leaders had been warned by Alexius and Frankish contact with the east, and even those in the vanguard managed to keep control of their forces – though luck played its part in this. Furthermore, they seem to have made sure that all were alert, for although the timing of the attack was a surprise, as probably was its direction, when it came, camp of a sort was made quickly. From the viewpoint of the crusaders, what is striking is that the battle evolved and was never directed. Although only a fraction of the crusader army was engaged, their advantage in numbers had much to do with their victory – just as it had at Nicaea. For Kilij Arslan seems to have repeated the error made at Nicaea; he counted on the enemy panicking under a surprise attack. When they resisted he was drawn into a bloody close-quarter battle in which the crusader footsoldiers in the camp made stiff resistance, partly because of their very numbers. As at Nicaea the appearance of a relief force, in this case one part of which under Adhémar came from an unexpected direction, drove his men from the field. That this was a pell-mell affair with no evidence of overall command (which led to the division in the crusader ranks in the first place) should not be allowed to detract from the quality of the crusader leadership.106 The army was alert and when the surprise attack came managed to establish a camp which subsequently formed a fortress. Robert of Normandy rallied knights alarmed by the novel methods of the enemy and subsequently he and Bohemond imposed a discipline upon them. The enemy broke into the camp and did much destruction, but the foot evidently fought hard, otherwise the camp which anchored the cavalry in their struggle would have been swept away. All of this suggests a formidable coherence in the crusader army and a considerable will to fight. It must be remembered that the terror which they inspired had served the Turks well in their fights with the Byzantines and others who found their missile tactics difficult to counter. Above all, the sense of isolation created by encirclement panicked large forces time after time. At Dorylaeum some of the knights did panic – those under Bohemond – but they were rallied by Robert of Normandy. Once discipline and solidity of formation was reimposed, partly because they simply couldn’t do anything else pinned against their own tents, the knights found that they could resist – though fairly passively. It was a lesson Nicephorous Botaneiates had learned as a general under Constantine IX during a retreat in the presence of the Patzinacks:

[Botaneiates] ordered his men not to spread out as the rest of the men were seen to be doing and not to turn their backs to the enemy making themselves into a target for Pecheneg arrows. … The Pechenegs on seeing a small group which advanced information and in battle order, made a violent sortie against them. … retired when they saw it was impossible to disperse the Byzantines…. They were unable to engage the Byzantines in hand-to-hand combat for having made a trial of close fighting, they had many times lost a great number of men.107

In any case, there was a limit to the losses the Turks were prepared to take. The loss of Nicaea was a blow to the Seljuk Kilij Arslan for like his father he aspired to be something more than a ruler of nomads – hence the acquisition of Nicaea as a capital and the effort to seize Antioch under Sulayman. But he was a lord of nomads and for them murderous casualties were simply not worthwhile before an enemy who could be evaded and whose departure would allow them to return to their pasture-lands. If Albert’s figure of 3,000 is in any way to be believed they had suffered badly enough for their leader’s ambitions. Only once again would they stand and fight – at Heraclea where an ambush was attempted and failed but it seems to have been so feeble that most of the sources do not mention it.108 But if the Turks were now in no position to check the crusaders, they did not know that and Fulcher says that from this time the army proceeded very carefully, while Albert says they resolved not to break up again.109 The Turks of Anatolia had been defeated, in so far as that means anything when speaking of a nomadic people who had clearly not been driven out of Asia Minor. Their ruling house had suffered a severe blow. They had lost a capital which gave them prestige, access and control over the emirates of western Asia Minor who were now at the mercy of the Byzantines. It opened the way, as we shall see, for a Byzantine reconquest in western Asia Minor. It was a stunning triumph for the crusaders for hitherto the onward march of the Turks had been unstoppable, as they themselves recognised for, as the Anonymous says, ‘the Turks… thought that they would strike terror into the Franks, as they had done the Arabs and Saracens, Armenians, Syrians and Greeks by the menace of their arrows’.110

In part they had been defeated by luck. Kilij Arslan had mistaken the People’s Crusade for the totality of the western effort and had to return from Melitene when they besieged Nicaea. His attack on the Provençals at Nicaea was mistimed, as was that against the vanguard near Bozüyük. But the victors made their own luck. It was their solid resistance that Kilij Arslan underestimated, hence their victory and his defeat. This rested on their manner of war in the west, which called for disciplined close-quarter fighting in which heavily armoured men played a key role. Ultimately, however, they differed from earlier enemies of the Turks by their motivation, their religious fanaticism which underpinned their fighting style. In the crisis of the battle at Dorylaeum that zeal showed in their password, ‘Stand fast altogether, trusting in Christ and in the victory of the Holy Cross. Today, please God, you will all gain much booty’.111 And so of course they did, and their spoils were much more than merely the pickings of the nomad camp. For the defeat at Dorylaeum seems to have sparked off revolts in some of the cities along the crusader line of march. The Anonymous says that as the Sultan fled he had to trick his way into the cities which his forces then looted. By contrast, the Christian army was welcomed in the vicinity of Iconium and this reception would become even warmer in east. These were truly the fruits of victory, for as a later eastern source commented, ‘The land was shaken before them.’112

Dorylaeum was a great Roman way-station and the key to the route system of the Anatolian plateau.113 From there they had a choice of routes to Antioch (see fig. 2). From Dorylaeum ran the great military road through Ancyra to Sebasteia and the far frontiers, towards Lake Van and the Caucasus. At Ancyra the traveller could turn south to the ‘Pilgrims’ Road’ to Tyana, the Cilician Gates and on to Antioch. This road forked east for Caeserea-in-Cappadocia, whence it led down to Comana, Germanicea Caeserea (Marasch) and thence to Antioch. It was along this road network that the Byzantine emperors had gathered the forces of the provinces on their way to the frontiers. They could have taken this route, which they would have known from earlier pilgrimage, direct from Nicaea; that they did not reflects serious political considerations. The Byzantines were, above all, interested in the south and west of Anatolia, and it can hardly be a coincidence that the route chosen facilitated the campaign by Alexius and his generals which would carry them to Philomelium by June of 1098.114 However, there was a choice of routes south from Dorylaeum: the quickest lay via Pessinus (near modern Ballihisar), Archelais (modern Aksaray), Tyana (Kemerhisar, south-west of modern Nigde) and the Cilician Gates, but this would have taken the army across the arid heart of Anatolia with all the problems of watering and the extremes of temperatures which we have noted. It was possible to fork south and east at Pessinus and descend via Philomelium (modern Akşehir) towards Iconium (modern Konya), or south and west via Amorium to the vicinity of modern Afyon. Another road ran due south via Nacolia (modern Seyitgazi) to join the route to Iconium just north of Afyon, while further west was another route via Cotiaeum (Kütahaya) to Afyon.115 The sources are very vague about this early part of the journey: they all wrote long afterwards when the memory of hard marching had been eclipsed by much later doses of the same thing, and many more spectacular events. There is, however, some indication that they took the route via Nacolia. Albert of Aachen says that on the fourth day of their march, having suffered terrible thirst, they rested in the Malabranias valley, which cannot be certainly identified, where many died of drinking too much. Nacolia (Seyitgazi), on the river Seydi, is eighty kilometres from the battlefield of Dorylaeum, very roughly four days march, and could thus be Malabranias – though there could be no certainty. This tale of hardship and suffering is confirmed by Fulcher and the Anonymous who was very worried by the heavy loss of horses.116 Here in high summer with temperatures around the 30° centigrade mark, the crusaders were crossing the Anatolian plateau; this is not flat land, but highly scenic, scarred by deep scarps and dry valleys, and almost waterless. It is a majestic, rather frightening landscape, and a harsh environment for a large force to traverse. Albert tells us that the army divided after a while, with Godfrey’s brother, Baldwin, and Bohemond’s nephew, Tancred, setting off on a different route from the main army. Baldwin took a difficult road into the valley of the Orellis, while Tancred went to Philomelium and thence to Iconium and Heraclea (modern Erégli), and the main army proceeded to Antioch-in-Pisidia (Antiochetta, now modern Yalvaç) which lies to the south of the Sultan Daglari. At this point, however, Albert’s account is at its worst. Antiochetta is described as being next to Heraclea, which is listed on Tancred’s journey as coming before Iconium. Moreover, there is no further mention of the journey of the main army until it reaches Marasch, presumably because Albert’s informants were with Baldwin on his diversions to Cilicia and Edessa. Fulcher confirms that the army went to Antiochetta, but offers no information on the route taken.117 In fact the army could have taken any of the routes from Dorylaeum. However, no source describes anything remotely resembling the crossing of the Sultan Daglari mountain range which rises suddenly and sharply out of the steppe to over 2,600 metres; the accounts of suffering reflect the passage across the dry steppe, not that over a formidable mountain barrier. Therefore, the likelihood is that they took a western route, probably via Nacolia approaching Antioch-in-Pisidia roughly via the modern Afyon and passing to the south of the Sultan Daglari via their western foothills, which are relatively gentle (see fig. 2). Tancred and Baldwin probably left the main army in the vicinity of Afyon and pushed along the more direct route to Iconium north of the Sultan Daglari via Philomelium, presumably watching for enemy attack; perhaps one took the road via the ancient Hadrianopolis (south-east of Akşehir) and the other that through Laodicea (modern village of Halici, east of Akşehir).

The really interesting question is why the army went to Antioch-in-Pisidia at all, for the road from Dorylaeum via Polybotus (modern Bolvadin) and Philomelium (modern Akşehir) to Iconium is shorter (by at least three days march) and more direct. Albert attributes the splitting of the army to the needs for supply. He and Fulcher stress that Pisidia was a fertile and pleasant land, where the army enjoyed a brief rest and Godfrey was injured by a bear while out hunting. After his account of the hardships and want on the dry steppe the Anonymous mentions a ‘fertile country, full of good and delicious things to eat’ which may well be Pisidia.118 Indeed, Pisidia is a fertile rolling country, a great contrast with the steppe to the north of the Sultan Daglari, and this must have been a real consideration in planning the route of the army. At the same time, the Anonymous indicates that the populations of the cities of Asia Minor rose against the Turks and that, for the Byzantines, Antioch, the chief city of Pisidia, was a desirable prize. The foraging needs of the army, together with the cooperation with Byzantium, probably combined to draw the army along this route. Hagenmeyer suggests that they left Antioch about 5 August, arriving at Iconium on 15 August, a rate of march of about twelve to thirteen kilometres per day through this relatively flat country, though the last forty kilometres into Iconium pass through harsh and waterless hills.

Tancred and Baldwin seem to have rejoined the main army at Iconium which the Turks made no effort to defend, although its Byzantine defences were probably still intact. The local population welcomed the crusader army and advised them to carry much water because the land to the east was dry. The road to Heraclea passes over a featureless plain, probably then something of a salt desert, but now brought back to life by irrigation. We do not know which of three possible routes they took from Iconium to Heraclea, which vary in distance between 140 and 170 kilometres for the only clue is that they spent two days resting at a river after two days march eastwards. This must refer to the Çarasamba which, however, cuts all the routes, but there is no reason to believe that they did not take the shortest route.119 At Heraclea the Turkish garrison attempted to ambush them but their scouts had warned them and the enemy were brushed aside easily and the city captured.120 The army rested there for four days. They now faced a very important choice of route, for east of Heraclea lay the Taurus mountains, in a great arc from south-west to north-east, dividing Anatolia from Syria. They could either journey south-east on the ‘Pilgrim Road’ via the Cilician Gates, Tarsus, Adana and Alexandretta (Iskenderun) to Antioch which was the more direct route, or they could take the road to Caeserea-in-Cappadocia (Kayseri) across the Taurus and down via Coxon (Göksun) and Marasch (Kahramanmaraş). The difference between these two routes was considerable: Heraclea to Antioch via the Cilician Gates is a journey of some 350 kilometres, but via Caeserea over 630 kilometres. It was extraordinary that they chose the latter route for the main army, while dispatching Tancred and Baldwin into Cilicia. Why was this strange choice made?

It needs to be stressed how difficult travelling overland was in this period. Although the road system of Asia Minor was basically that of the Romans, it is unlikely that the roads were in good condition after thirty years of political chaos and then Turkish domination.121 Though sometimes the journey was relatively easy there were other occasions, as in the pass south of Göksun, when every step was a calvary. For most of the time it must have been simply very unpleasant and dangerous, even without considering the possibility of enemy attack. The death of horses and pack animals must have been appalling and militarily disastrous; just after Dorylaeum, we hear of knights mounted on oxen, their horses having perished on the dry steppe.122 Only the strongest of motives could have led the army to march northwards to Caeserea, deliberately ignoring a much shorter route. Historians have been strangely slow to grasp the scale and importance of the diversion via Caeserea. It has been suggested that the narrowness of the famous Cilician Gates – only twenty-five metres at one point, and the hostile climate of Cilicia explain the decision. However, although the road to Caeserea is less abrupt than that over the Cilician Gates, the long sustained climb (Caeserea is at 1,254 metres) would have been sapping, while the road down to Marasch offers going every bit as difficult and narrow as either the Cilician Gates or the Belen pass from Cilicia to Antioch over the Ammanus Mountains, often called the ‘Syrian Gates’. Further the road rises to a maximum of over 1,700 metres, while the Cilician Gates never rise above 1,000 metres. The real military risk of the direct route was that the garrison of Antioch might challenge their crossing of the Belen Pass but from their perspective at Heraclea there were unknown risks of a similar kind facing them in the mountains. Moreover, the season was quite advanced and, while the army was now hardened, the loss of animals must have slowed it down. This opened the risk of being caught by the snows which can come as early as October in the high passes, for the road to which they were committed rises to 1,700 metres.

It is likely that what we see is the development of an Armenian strategy which had been discussed with Alexius, either at Constantinople or at Pelekanum after the fall of Nicaea.123 As the Crusade advanced many of the cities in their path ejected their Turkish garrison and welcomed the crusaders. In addition, they had contact with Armenians as we have noted and, at Iconium, Christians gave them intelligence about local conditions. The Christian population of Asia Minor had suffered badly at the hands of the nomadic Turks, whose violent and arbitrary dominion was resented. Raymond of Aguilers knew that Antioch had only fallen to the Turks some fourteen years before, and he catalogues the sufferings of its Christian people.124 When the Emperor Alexius retreated from Philomelium, about 20 June 1098, most of the local population chose to leave with him rather than again face their Turkish masters.125 In a passage which has received surprisingly little attention, Stephen of Blois says that in Cappadocia the army directed its march against a powerful local emir, Hasan, who is probably more correctly called Baldajii. His brother, Abu’l-Qasim, had ruled at Nicaea after the death of Sulayman, whose son Kilij Arslan was held captive by Malik Shah (1086–92). Hasan himself briefly held power at Nicaea after his brother, but Kilij Arslan escaped from prison on the death of Malik Shah in 1092 and resumed power at Nicaea.126The crusaders, therefore, were prepared to confront real opposition in pursuit of what we may call their Armenian strategy, and they drove into his lands as they advanced towards Caeserea and then turned south to Antioch.

The long uphill march took the army past the area of modern Nigde over a series of dramatic scarps into wide upland plains, often watered by great lakes. Towards Caeserea they captured a strong place which was given to Simeon, a local man whose presence in the army points to forethought. Beyond Caeserea, which they reached about 21 and left about 24 September, they travelled through steep and broken country for some eighty-six kilometres to a city which had held out for three weeks against Turkish siege; there Peter d’Aups, a westerner in the service of Alexius, was given control (see fig. 2 and 4). This place has been identified as ‘Plastencia’, on the authority of Bauldry of Dol, and recent research identifies the Greek place of that name with Elbistan, a city well off the crusaders’ path to the east on the road to Melitene. The likelihood is that this was Comana where the army seems to have arrived about 30 September.127 The army left Bohemond to pursue the besiegers of Comana and went on to Coxon (Göksun) on 4 October, which the local Christians promptly surrendered to them. There a false rumour that the enemy were deserting Antioch led Raymond of Toulouse to send a force of 500 knights, under Peter of Castillon, to seize the city; at a settlement of heretic Christians near to Antioch they were informed that the rumour was false whereupon some of them under the command of Peter de Roaix, went on to establish a Provençal base in the valley of Ruj, parallel to the Orontes valley on the eastern side of that river. Rugia was about seven kilometres from Rusa to the south of Antioch.128 The main army followed along down the bitter and painful pass near what is now called the Püren Geçidi, which rises to 1,630 metres, the downward slope of which is a penance even in modern transport. The Anonymous records that horses and animals died in falls and knights sold off their arms at any price rather than carry them across this ‘damnable mountain’.

At Marasch the Turkish garrison had fled and the army was welcomed by its Armenian ruler, Tatoul, who, as a supporter of Alexius, continued to hold the place.129 The army had now emerged from the mountain passes and stood at the head of a great flat valley, the Amouk, which stretches down to Antioch and the coast beyond, between the Ammanus range to the west and the Kartal Daglari range to the east on the edge of Syria. The success of their Armenian strategy had delivered the mountain cities over to them, and now the army was able to set out on the last leg of the journey down the Amouk. But before they set out, local inhabitants told the leaders that ‘Artāh, which the crusader sources call Artasia, would welcome them but had a strong Turkish garrison. The leaders sent Robert of Flanders ahead with 1,000 knights, on whose arrival the Armenian population butchered the Turkish garrison and opened the gates. Ralph of Caen suggests that Baldwin and Tancred commanded this expedition and never mentions Robert, but his account confirms that of Albert in its main outlines. Once the Franks were installed they were besieged by a force which Albert numbers at 20,000. They provided a lesson in tactics for the crusaders. A small number of lightly armed Turkish horsemen trailed their coats outside the walls and when a lot of Franks, foot and horse, rushed out they fled drawing their enemies into an ambush which cut them off from the city. Robert of Flanders rescued them by a charge from the city, but Christian losses in men and horses were heavy. Ralph of Caen also tells us that many Franks were lured out of the city and suffered heavy losses in close-quarter combat with the Turks. The survivors retreated into the city where the depleted garrison now had to face a close siege. The siege was lifted with the arrival of 1,500 reinforcements and the city was given a Frankish garrison, which Ralph says was in the control of Baldwin. The bitter fight underlines the importance of ‘Artāh to the crusaders and the fact that it later changed hands, for it was captured by Kerbogah, strengthens the point.130 From this it would appear that the main army had marched down the Amouk until it was just north of the great lake to the north of Antioch. There the road forked; to the west it passed the Belen Pass and arrived before the Bridge Gate on the west side of Antioch. The eastern fork led the army to ‘Artāh, which Ralph of Caen would later describe as the ‘shield of Antioch’.131 It stood close to the modern Reyhanli across the road to ’Azāz, and just north of its junction with the Antioch-Aleppo road to the east of the Iron Bridge, which controlled the crossing of the Orontes to the north of Antioch. The capture of ’Artāh helped to secure the eastern approaches to Antioch as a prelude to a siege, thus isolating the city from its obvious source of support. The Armenian strategy provided a friendly hinterland and a springboard for this isolation of Antioch, which was increased by the expedition of Tancred and Baldwin to Cilicia. Albert emphasises that all this was done with the agreement and consent of the leaders of the army and this must include the Byzantine representative, Tatikios, whose man took over Comana and, presumably, at least some of the other cities. This was much more than mere individual opportunism, the reason usually given for the expedition to Cilicia.

Tancred and Baldwin of Boulogne’s expedition to Cilicia is very well known and has generally been treated as a private enterprise affair (see fig. 2).132 The sources are often not very informative on how it came about. Raymond of Aguilers never discusses this event, perhaps because the Provençals were not involved; Fulcher was much more concerned about the expedition to Edessa, in which he participated, and says that Baldwin took his own men into Cilicia, while the Anonymous, as so often, simply reports the events without explanation. Ralph of Caen, who likes to present Tancred as an emerging leader, tells us that Tancred chose to undertake this expedition.133 Albert of Aix reports that, probably in the region of Afyon, Tancred and Baldwin were sent along the northern road to Iconium, but that Tancred was ahead after Heraclea and went down to the coast through Cilicia, leaving Baldwin who got somewhat lost following behind.134 This presents events in a different light and it should be noted that each of these young men seems to have had substantial forces at his disposal. When they came to blows at Mamistra Tancred attacked with 500 men but was defeated by the larger force of Baldwin. Earlier, at Tarsus, Tancred had been reinforced by 300 men from Bohemond, and in the quarrel over this city both young men claimed that they were acting in the name of their superiors, Bohemond or Godfrey, in passages which smack of the ‘my big brother is bigger than yours’ syndrome.135 The impression is of an expedition in which the ardour and greed of two young men got out of hand. It is interesting that friendly locals once more appear in a notable role. Tancred had with him an Armenian whom he had known earlier and it was perhaps this influence, and their fear of Bohemond, that led the Armenian population of Tarsus to prefer his rule – though they eventually submitted to Baldwin.136 At Adana Tancred found a city already half-liberated by the local Armenian prince Oschin and partly occupied by a Burgundian, Welf; given Oschin’s good relations with Alexius it would seem likely that Welf was another westerner in imperial service.137 At the end of the Cilician adventure Baldwin was persuaded by Bagrat, an Armenian whom he had got to know at Nicaea and who was the lord of Ravendan, to strike east into the Armenian territories towards Edessa to Tell-Bashir, but we know from Fulcher that he first returned to the main army.138 Baldwin then became embroiled in the complex politics of the Armenian princes and in February of 1098 received a request from Thoros, the Armenian ruler of Edessa, to go to that city which after many adventures he reached on 20 February.139 By 8 March 1098 Baldwin had intrigued with disaffected citizens to overthrow Thoros and was in effective control of the city.140 Local Christians, as we have already noted, delivered over many key cities as far south as Ruj to the Franks and this is corroborated by the Damascus Chronicle which specifically mentions the fighting at ‘Artāh.141 It was no wonder that Anselm of Ribemont would boast in a letter to the west that the army held 200 forts and cities, while Stephen put the figure at 160.142 This should be seen as the fruits of a deliberate policy of which the Cilician expedition was a part.

At Heraclea, or shortly thereafter, the princes must have decided to implement their Armenian strategy which probably aimed to reproduce the principality which Philaretus had ruled in the years before 1085, elements of which (such as Edessa) remained independent and in some sense attached to the empire. Gabriel of Melitene seems to have held aloof from the crusade. The idea of creating such a liberated zone was probably developed in discussions with Alexius – Tatikios was his man on the crusade and he seems to have aided and abetted the process – but it was made possible by the success at Dorylaeum and the reaction of the native population to it. After Heraclea the leaders decided to capitalise on their success and launched the main army into a long diversion over very difficult territory, driving back the forces of Hasan. Into the more sheltered area of Cilicia a small force led by Tancred and Baldwin was dispatched. It was a risk, but one which succeeded handsomely. The establishment of a great bastion of Byzantine power on the Syrian border was welcome to both Alexius and the crusader leaders. It would enable the Byzantines to conquer southern Asia Minor. For the crusaders liberation of the persecuted Christians of the east was one of the objectives of their journey. Furthermore, such a Byzantine bulwark would provide a secure base for the real objective of their endeavour – Jerusalem. We have to remember that they had come for Jerusalem, for Palestine, not Antioch or some North Syrian domination. It is a point which the mass of the army would make forcibly to its leaders in the later months of 1098.143 As things turned out this plan was never properly realised. Its central assumption was a common interest between the Byzantine empire and the crusaders; the stress of events undermined this. Even so, despite a heavy price in garrison troops detached from their force the conquests paid off handsomely for the First Crusade. Food, useful intelligence and supplies reached the crusade from the Armenians whose merchants frequently visited the city and Armenians helped in the routing of Turkish forces and the slaying of Yaghisiyan.144 The possession of so many bases in the general area of Antioch, the old dominion of Philaretus, gave the crusade a much needed platform for their assault on Antioch. Baldwin’s possession of Edessa enabled him to send aid and supplies to the army at Antioch. It was also a powerful distraction for local Islamic leaders. In May 1098 this factor caused Kerbogah to divert his huge relief army for a three-week siege, which was fatal for his chances of success against Antioch.145 Militarily, the policy was a striking success and the choice made on the road from Heraclea proved to be a correct one, dangerous though it must have appeared. It enabled the crusaders to confront their second enemy, the Turks of Syria, with a considerable territorial base and much assistance which was extremely valuable.

1 RA, pp. 42–3; AA, 314; FC, pp. 81–2.

2 A. M. Schcider and W. Karnapp, Die Stadtmauer von Iznik-Nicea (Berlin, 1938); S. Eyice, Iznik-Nicaea: the History and the Monuments (Istanbul, 1991). Stephen of Blois was exaggerating when he said that the city had 300 towers: Hagenmeyer, Kreuzzugsbriefe, p. 149.

3 At present the north gate is 250 metres from the lake, but the extensive marshes at this point are very much closer and point to the advance of the shoreline since the eleventh century. The same phenomenom is remarked by the south gate which at present is 350 metres from the shore, but there modern filling has taken place on a large scale. The re-entrant of the walls on this south-west corner of the city is probably to be explained by the shoreline and marsh in ancient times. The three landward gates of the city and almost all the enceinte still surround the small town of Iznik, although the walls are ruinous in places by the lake, where the watergate has long perished. There is an outer wall which terminates north and south on the shore, protecting the littoral of the city, but this is a late construction and its crudity is particularly evident at the gates in comparison with the fine work of the inner Roman gates. It is a construction of the thirteenth century.

4 Alexiad, pp. 335–6, 130, 206; on the last occasion Alexius’s commander was Tatikios.

5 Alexiad, p. 330–1.

6 France, ‘Anna Comnena’, 22–3.

7 See in particular Runciman’s declaration of faith, 1. 171: ‘he [Alexius] believed that the welfare of Christendom depended on the welfare of the historic Christian Empire. His belief was correct.’

8 C. Cahen, Pre-Ottoman Turkey (London, 1968) pp. 19–20.

9 C. Cahen, ‘The Turkish invasion: the Selchukids’, in Setton, Crusades 1. 139–40.

10 B. Lewis, The Arabs in History (London, 1958), pp. 49–63.

11 C. E. Bosworth, ‘Recruitment, muster and review in medieval Islamic armies’, in V. J. Parry and M. E. Yapp, eds., War, Technology and Society in the Middle East (London, 1975), p. 60; J. D. Latham and W. F. Paterson, Saracen Archery (London, 1970), p. xxiii; this cycle of rulers throwing off tribal dependence is a major theme of Ibn-Khaldûn, especially pp. 123–263.

12 H. A. R. Gibb, ‘The Caliphate in the Arab States’, in Setton, Crusades 1. 81–2.

13 D. Ayalon, ‘Preliminary remarks on the Mamluk military institution in Islam’, in V. G. Parry and M. E. Yapp, eds., War, Technology and Society in the Middle East (London, 1975), pp. 51–4, Bosworth, ‘Recruitment, muster and review’, pp. 62–3; O. S. A. Ismail, ‘Mu’tasum and the Turks’, Bulletin of the School of African and Oriental Studies, 29 (1966), 12–24.

14 Cahen, ‘The Turkish invasion’, 143–6.

15 S. Vryonis, The Decline of Medieval Hellenism in Asia Minor (London, 1971), pp. 87, 94–5.

16 C. Cahen, ‘Les changements techniques militaires dans le Proche Orient médiéval et leur importance historique’, in V. G. Parry and M. E. Yapp, eds., War, Technology and Society in the Middle East (London, 1975), 115 comments on the need for research on Turkish horses. Lindner, ‘Nomadism, horses and Huns’, 3–19; Bundari, Dawolat al-Saljuq (Cairo, 1900), p. 37 – I owe this reference to Professor A. K. S. Lambton.

17 A matter which Davis stresses. See above p. 73.

18 al-Jahiz quoted in Latham and Paterson, Saracen Archery, p. xxiii.

19 AA, 344, 359, 369, 400.

20 AA, 331–2, 362: La Chanson d’Antioche, ed. S. Duparc-Quioc, 2 vols. (Paris, 1977–8), 2. 1606, 8979 mentions Gerard at Nicaea and later at the battle against Kerbogah.

21 Latham and Paterson, Saracen Archery, pp. xxv, 30; AA, 334, 377, 424; FC, p. 85.

22 RA, pp. 50–1; RC, pp. 639–41; AA, 362, 408; see below pp. 192–3, 206, 271–2.

23 Latham and Paterson, Saracen Archery, pp. xxiv–xxxi; C. Cahen, ‘Un traité d’armurerie composé pour Saladin’, Bulletin d’Etudes Orientales de l’Institut Français de Damas, 12 (1947/48), 132–3 and ‘Les changements techniques militaires’ pp. 116–17, points to the development of this light crossbow which threw large numbers of darts, hence its nickname ‘hailstone’, which is described in some detail by Latham and Paterson, pp. 145–51.

24 There are some fine eleventh-century examples in the armoury of the Topkapi Palace in Istanbul.

25 D. Nicolle, ‘Early medieval Islamic arms and armour’, Gladius, special volume (1976), 53.

26 Nicolle, ‘Early medieval Islamic arms’, 26, 53–82; M. V. Gorelik, ‘Oriental armour of the Near and Middle East from the eight to the fifteenth centuries as shown in works of art’, in R. Elgood, ed. Islamic Arms and Armour (London, 1979), pp. 30–63; RC, 621.

27 On the destruction of the People’s Crusade see above p. 93–5; W. E. Kaegi, ‘The contribution of archery to the Turkish conquest of Anatolia’, Speculum, 39 (1964), 96–108 rather exaggerates the influence of archery. After all, many Turks fought for the Byzantines with the same tactics.

28 The general explanation of Byzantine decline here relies on: M. Angold, The Byzantine Empire 1025–1204, a Political History (London, 1984), pp. 1–113 and Vryonis, Hellenism, pp. 70–142. See also Charanis, in Setton, Crusades 1. 177–219; Jenkins, Byzantine Empire; Ostrogorsky, Byzantine State, pp. 280–315.

29 On the Byzantine army see J. D. Howard-Johnstone, Studies in the Organisation of the Byzantine Army in the Tenth and Eleventh Centuries, unpublished D. Phil, thesis, University of Oxford, 1971; H. Ahrweiler, ‘L’organisation des campagnes militaires à Byzance’, in V. J. Parry and M. E. Yapp, eds., War, Technology and Society in the Middle East (London, 1975), pp. 89–96 is particularly useful on its subject. See also Ostrogorsky, Byzantine State, pp. 87–90; on Dorylaeum which had baths for 1,000 troops see Vryonis, Hellenism, pp. 31–2.

30 Angold, Byzantine Empire, pp. 63–5; C. Mango, Byzantium (London, 1988), p. 57.

31 W. C. Brice, ‘The Turkish colonisation of Anatolia’, Bulletin of John Rylanad’s Library, 38 (1955), 18–44.

32 On Manzikert see C. Cahen, ‘La campagne de Mantzikert d’après les sources musulmanes’, Byzantion, 9 (1934), 613–42, and for corrections based on the Greek sources Vryonis, Hellenism, pp. 96–103; A. Friendly, The Dreadful Day. The battle of Mantzikert, 1071 (London, 1981), pp. 163–92; J. C. Cheynet, ‘Mantzikert. Un désastre militaire?’, Byzantion, 50 (1980) 410–38.

33 Cheynet, ‘Mantzikert’, p. 431 suggests losses of five to ten per cent in total much less than those of Alexius Comnenus’s, reputedly 5,000 at Dyrrachium ten years later, on which see above pp. 75–6.

34 J. Schlumberger, ‘Deux chefs normands des armées byzantines au XI siècle: sceaux de Hervé et de Raoul de Bailleul’, Revue Historique, 16 (1881), 289–303; L. Bréhier, ‘Les aventures d’un chef normand en Orient au XI siècle’, Revue des Cours et Conferences, 20 (1911–12), 99–112; see also Marquis de la Force, ‘Les conseillers latins d’Alexis Comnène’, Byzantion, 11 (1936), 153–65.

35 On Philaretus see J. Laurent, ‘Byzance et Antioche sous le curopalate Philarète’, Revue des Eludes Arméniennes, 9 (1929), 61–72; T. S. R. Boase, The Cilician Kingdom of Armenia (Edinburgh and London, 1978), pp. 3–4; Skoulatos, Anna Comnena, pp. 263–5; Cahen, Turkey, pp. 76–7; Vryonis, Hellenism, pp. 109–10. E. Sivan, L’Islam et la Croisade; Idéologie et Propagande dans les Réactions Musulmanes aux Croisades (Paris, 1968), p. 19 suggests that Sulayman was acting for the emperor in his capture of Antioch.

36 Cahen, Turkey, p. 76; Ostrogorsky, Byzantine State, p. 257.

37 A. R. Gadolin, ‘Alexius I Comnenus and the Venetian trade privileges. A new interpretation’, Byzantion, 50 (1980), 439–46, suggests that the great trading concessions, extended to the Venetians in the Golden Bull of 1082, were intended as much to stimulate the damaged Byzantine economy as to persuade them to give support against the Normans; M. Angold, ‘The Byzantine State on the eve of the Battle of Manzikert’, in A. Bryer and M. Ursinus, eds., Manzikert to Lepanto: the Byzantine World and the Turks, 1071–1571, Byzantinische Forschungen 16 (Amsterdam, 1991), 33.

38 Sivan, L’Islam et la Croisade, p. 19 stresses the lack of any spirit of jihad amongst the Seljuk leaders. In taking over Antioch, Sulayman was at pains to safeguard the Christian population; Cahen, Turkey, p. 75.

39 Vryonis, Hellenism, pp. 114–16; Cahen, Turkey, pp. 76–82.

40 Matthew, 30–33; Michael the Syrian, Chronique de Michel le Syrien, Patriarch jacobtte d’Anttoche 1166–99, ed. and tr. J. B. Chabot, 4 vols. (Brussels, 1963, reprint of 1899–1910 edition) [hereafter cited as Michael], vol. 3. 179; see below pp. 168–9, 304–7.

41 Cahen, Turkey, pp. 77–80.

42 Cahen, ‘The Turkish invasion’, 161–3.

43 Cahen, ‘The Turkish invasion’, 157.

44 Lindner, ‘Nomadism, Horses and Huns’, 8, 15.

45 A. V. S. Lambton, Continuity and Change in Medieval Persia (London, 1988), pp. 21–4 and especially 21, n. 60.

46 FC, pp. 82–83; GF, p. 21.

47 Nicephorous Bryennius, Commentarii, ed. A. Meineke (Bonn, 1836), pp. 41–2 cited and tr. Kaegi, ‘Archery’ p. 106.

48 GF, p. 19; RA, p. 52 Krey, First Crusade, p. 135; AA, 320, 328–31.

49 GF, p. 30; RA, p. 52; AA, 375.

50 On which see below pp. 200–6.

51 GF, pp. 23–4.

52 Vryonis, Hellenism, pp. 112–13; Cahen Turkey, pp. 76–7, 83–4.

53 GF, p. 22.

54 Matthew, 28; Cahen, Turkey, pp. 81–2; Michael 3. 179.

55 RA, p. 43; GF, pp. 14–15; AA, 320–1. It must be admitted, however, that some aspects of Albert’s account of the early siege are confused. A long list of those present, 315, includes Robert of Normandy who, however, is not amongst the leaders who urged Raymond of Toulouse to hurry, 319, yet is recorded as taking part in the battle, 320. In fact he did not arrive until 3 June. I would guess that Albert was trying to reconcile confused and contradictory information from his sources.

56 GF, p. 15; AA, 320.

57 Alexiad, p. 334.

58 GF, pp. 14–17; RA, pp. 43–4; Alexiad, pp. 335–6. The famous ‘Greek Fire’ was not included in the aid which Alexius offered to the crusaders. Much has been written on the nature of this mysterious substance; J. Bradbury, ‘Greek Fire in the West’, History Today, 29 (1979), 326–31; H. R. E. Davidson, ‘The secret weapon of Byzantium’, Byzantinische Zeitschrift, 66 (1973), 66–74; J. Harvey and M. Byrne, ‘A possible solution to the problem of Greek Fire’, Byzantinische Zeitschrift, 70 (1977), 91–9; J. R. Partington, A History of Greek Fire and Gunpowder (Cambridge, 1960); C. Zenghetin, ‘Le feu grégois et les armes à feu byzantines’, Byzantion, 7 (1932) 265–86; the crusaders claimed it was used against them at Jerusalem: see below p. 350.

59 BD, 27–9; HBS, 181; Robert the Monk, Historia Iherosolimitana, RHCOc. 3. (hereafter cited as RM), 756; FC, p. 82.

60 Rogers, Siege Warfare, studied this and other sieges and has helped to clarify my thinking considerably. Rogers prefers ‘armoured roof to my term ‘penthouse’.

61 Murray, ‘Army of Godfrey de Bouillon’, says that Henry was related to Godfrey de Bouillon (his brother Godfrey was also on the crusade) and came of a family which held the castle of Esch-sur-Sûre in the Ardennes; AA, 321–2; see above p. 105.

62 AA, 322–5; RA, p. 44.

63 AA, 325 and n.a.

64 AA, 325–8.

65 Alexiad, pp. 337–8.

66 AA, 325; Hagenmeyer, Kreuzzugsbriefe, p. 145.

67 Rogers, Siege Warfare, p. 81: Alexiad, p. 336; GF, p. 16; RA, p. 44; FC, p. 82.

68 FC, p. 82; AA, 320; GF, p. 17.

69 RA, p. 44; FC, p. 83; Hagenmeyer, Kreuzzugsbriefe, pp. 140, 144–5; GF, p. 17–18.

70 Alexiad, p. 336.

71 HBS, 181, 189–90, 212; GF, p. 37; RA, pp. 58, 109–10. On this embassy and the whole issue of diplomatic relations between the crusaders and Islamic powers on the First Crusade see M. A. Köhler, Allianzen und Verträge zwischen frankischen und islamischen Herrschern im Vorderren Orient (Berlin, 1991), pp. 1–72.

72 Gibb, ‘The Caliphate’, 85–95; P. M. Holt, The Age of the Crusades (London, 1986) pp. 9–15; Y. Lev, State and Society in Fatimid Egypt (Leiden, 1991); on the fragmentation of the Seljuk power and its impact on the crusade see below pp. 357–8.

73 Hagenmeyer, Kreuzzugsbriefe, p. 140.

74 Alexiad, pp. 372–4; Boase, Armenia, pp. 3–4.

75 Runciman, Crusades, 1. 299; Boase, Armenia, p. 4.

76 Matthew, 30.

77 Michael, pp. 173–4; Matthew, 35; ‘Anonymous Syriac Chronicle’, ed. and tr. A. S. Tritton and H. A. R. Gibb, Journal of the Royal Asiatic Society (1933), 69.

78 AA, 342–3, 350–1.

79 William of Apulia, 11. 1110. See also M. Mathieu, ‘Une source négligée de la bataille de Mantzikert: Les “Gesta Roberti Wiscardi” de Guillaume d’Apulie’, Byzantion, 20 (1950), 89–103. On the Normans in Byzantium see above p. 153, n. 34.

80 Alexiad, p. 341; Hagenmeyer, Kreuzzugsbriefe, p. 145.

81 AA, 328–9; GF, p. 18 does not mention Stephen of Blois; Hagenmeyer, Kreuzzugsbriefe, p. 145.

82 FC, p. 85; GF, p. 18; RA, p. 45; AA, 328–9; RC, 620–1; on losses amongst the stragglers see below p. 181, n. 104; Nesbitt, ‘Rate of march’, 178–80.

83 RA, p. 45; he later, p. 49, tells us that St Symeon Port was ten miles from Antioch – it is actually twenty-seven kilometres.

84 Bibliothèque Nationale 5131A, in which Raymond’s account is conflated with that of Fulcher, also represents a conflation of traditions of his own work. The life of Adhémar in the Chronicon monasterii sancti Petri Aniciensis, which is known separately as Gesta Adhemari Episcopi Podiensis Hierosolymitana, RHC Oc. 5. 354–5, refers to this battle as taking place ‘in campo florido’. This work is most certainly based on Raymond’s, but I think it was written close to the time with other recollections added, and the story that the battle was fought ‘in a flowered field’ may be one of them. It is unfortunate that the editors of Raymond of Aguilers in RHC Oc, 3. 240 capitalised the name without making clear its derivation. The latest edition by Hill and Hill, p. 45 n. 4 gives only a cryptic note; AA, 329–30.

85 GF, p. 18; RA, p. 45; Hagenmeyer, Kreuzzugsbriefe, p. 161.

86 Nesbitt, ‘Rate of march’, pp. 173–4, 178–80.

87 Hagenmeyer, Chronologie 169, p. 85; Runciman 1. 186, n. 1.

88 AA, 328–9; RC, 621; WT, p. 129; Runciman, 1. 186, n. 1 has an ingenious reconstruction of the battle.

89 On which see D. French, Roman Roads and Milestones of Asia Minor: Fasc. 1. The Pilgrims’ Road, British Institute of Archaeology at Ankara Monograph No. 3, British Archaeological Reports, International Series 105 (Oxford, 1981).

90 I would like to thank Dr David French, Director of the British Institute of Archaeology at Ankara, who told me of the existence of this road. He is currently writing an article on the routes of the crusades and very generously explained his ideas to me.

91 WT, 129; GF, p. 20.

92 Ibn al-Qalanisi, Damascus Chronicle of the Crusades, extracts ed. and tr. H. A. R. Gibb (London, 1967) [hereafter cited as Damacus Chronicle of the Crusades], pp. 41–2.

93 FC, p. 84; GF, p. 20; AA, 565; Ekkehard, p. 31; These forces may have been detached elements of a much larger allied force formed by Ridwan of Aleppo, the Danishmend Malik Ghazi and Karajan of Harran. The full size of their army which finally defeated the Franks, whose army probably started 50,000 strong, at Mersivan is unknown but the long harassment which preceded the final attack suggests that it was even smaller than the Western force.

94 Vryonis, Hellenism, pp. 123–5; on numbers see above pp. 157–8.

95 FC, pp. 83–4; RC, 621.

96 GF, p. 21; FC, p. 86.

97 FC, p. 85; GF, p. 19.

98 FC, pp. 85–6; AA, 329–30; RC, 622–3.

99 FC, p. 86; AA, 329–30; RC, 622–23; RA, p. 45.

100 GF, p. 20.

101 GF, p. 19.

102 AA, 331.

103 RA, p. 46; AA, 331.

104 AA, 330, 323; FG, p. 86; RA, p. 45.

105 GF, 19, 21; FC, p. 85; AA, 328–9.

106 AA, 330 implies that Godfrey was in command of the main force but this reflects his general prejudice in favour of his hero.

107 Attaliates quoted by Kaegi, ‘Archery’ p. 103.

108 GF, p. 23; for the rest it is authors who follow him who mention it: Tudebode, p. 30; HBS, 184; RM, 767.

109 FC, p. 87; AA, 333.

110 GF, p. 21.

111 GF, pp. 19–20.

112 GF, pp. 23–4; Tritton, ‘Anonymous Syriac Chronicler’, p. 70.

113 W. M. Ramsay, The Historical Geography of Asia Minor, Royal Geographical Society Supplementary Papers, 4 (London, 1890), pp. 212–13; K. Belke et al, eds., Tabula Imperii Byzantini, 5 vols. (Vienna, 1977–84), 4. 94.

114 On Alexius’s campaign of 1098 see below pp. 299–302.

115 Ramsay, Historical Geography, pp. 199–221. For an outline of the ancient and Roman roads of the area see W. M. Calder and G. E. Bean, A Classical Map of Asia Minor (London, 1959) and the useful map with Gazeteer in Vryonis, Hellenism, pp. 14–15 and the comments pp. 30–3; D. French, ‘A study of Roman roads in Anatolia’, Anatolian Studies, 24 (1974), 143–9, ‘Roman road system in Asia Minor’, Aufstieg und Niedergang der Romischen Welt 2. 7. 2 (1980), 698–729; Roman Roads and Milestones of Asia Minor, Fase. 2: An Interim Collection of Milestones, Pts. 1–2, British Institute of Archaeology at Ankara, Monograph 9, British Archaeological Reports International Series 392 (i) and (ii) (Oxford, 1988), p. 540 and map, The Pilgrims’ Road, especially p. 130 and map; Belke, Tabula Imperii Byzantini, 2. 32.

116 FC, p. 87; GF, p. 23. It is possible that the valley of Malabranias is that of the Porsuk near Kütahya (ancient Cotiaeum) which is about four days march south of Eskişehir (Dorylaeum) for this, I am fairly certain, is the route they took.

117 AA, 341–2; FC, pp. 87–8; GF, p. 23.

118 FC, p. 87; AA, 341–2.

119 Even if they took the longest route, however, and allowing for delay as they heard of the enemy forces, Hagenmeyer’s chronology of this part of the journey is surely wrong. He suggests that after a two day march they arrived at the Çarasamba on the 20 August where they rested for two days, then arrived at Heraclea about 10 September. This means a march of between 140 and 170 kilometres over a period of twenty-one days. A daily march rate of seven or eight kilometres per day over these flat lands seems unduly slow. By contrast, Hagenmeyer suggests that the army left Heraclea about 14 September and reached Caeserea-in-Cappadocia (modern Kayseri) on 27 September, a daily rate of seventeen kilometres up into the mountains. We can never be precisely certain of any dates other than 4 July for the departure from the field of Doylaeum and 20 October for the arrival at the Iron Bridge outside Antioch. I would suggest, however, that they must surely have reached Heraclea by the end of August but that we must allow for a slower rate of march in the mountains. All such dates must be approximate but I suggest:

(C) = Hagenmeyer’s dates

Heraclea 31 August. (10 Sept)

Heraclea–Caeserea: 240 km

4–21 September, eighteen days march at 13/14 km per day

(14–27 September, 14 days march at 17 km per day)

Caeserea–Comana: 86 km

24–30 September, seven days march at 12 km per day

(end September–3 October, four days march at 21 km per day)

Comana–Göksun: 55 km

1–4 October, four days march at 12/13 km per day

(4–6 October, three days march at 17 km per day)

Göksun–Marasch: 80 km

7–14 October, eight days march at 10 km per day

(8–13 October, six days march at 13 km per day)

Marasch–Iron Bridge: 150 km

15–20 October, six days march at 25 km per day

(14–20 October, seven days march at 20 + km per day)

This dating tries to take account of geographical differences. It is acknowledged that distances and dates can only be approximate.

120 GF, p. 23.

121 French, ‘Roman road system’, 713, points to the lack of evidence about maintenance in the Byzantine period.

122 FC, p. 88.

123 C. Cahen, La Syrie du nord à l’époque des croisades (Paris, 1940) pp. 209–10 was much struck by the ‘vaste détour’ which the army took to reach Caeserea-in-Cappadocia and thought it was the result of an ‘intérêt essentiellement byzantin’ to seize this area which controlled routes to north and east. The point made here is that at this stage Byzantine and Crusader interests were very close and, perhaps for different favoured this course of action. See above p. 187; GF, p. 23.

124 RA, p.64.

125 Vryonis, Hellenism, pp. 194–223; Alexiad, pp. 349–50.

126 Hagenmeyer, Kreuzzugsbriefe, p. 150; Cahen, Turkey, pp. 78–80.

127 Hagenmeyer, Chronologie, 188, pp. 97–8 suggested Comana but the most recent edition of the Gesta, p. 25 follows BD, 39; for Elbistan see Belke, Tabula Imperii ßyzantini, 2. 109–10. Dr David French thinks that ‘Plastencia’ may be somebody’s recollection of, quite literally, ‘a pleasant place’ whose name he had forgotten, transformed by Bauldry into a proper name. The case for Comana is that there was certainly a city there on the route and it is difficult to see where else could be intended. For the probable line of the road to Comana see French, Roman Roads and Milestones of Asia Minor, Fasc. 2. Pt. 2, p. 550 and map.

128 GF pp. 25–7; R. Dussaud, Topographie Historique de la Syrie Antique et Médiévale (Paris, 1927), pp. 163–7 identifies this area, and points out that as the Antiochenes held the Iron Bridge over the Orontes (on which see below pp. 206–8) that it must have been to the east of that river. As RA, p. 99, makes it clear that Rugia was relatively close to Albara and Ma’arra which can be identified, this is almost certainly correct. On the importance of this acquisition see below p. 224.

129 AA, 136–7; Boase, Armenia, p. 4.

130 AA, 358–61; RC, 639–41.

131 RC, 712.

132 This is the clear implication of Runciman 1. 197. Mayer, The Crusades, p. 48 remarks: ‘The two of them were almost certainly seeking their own personal gain’, while Riley-Smith, Idea of Crusading, p. 58, speaks of the two men breaking away from the main force.

133 FC, p. 89; GF, pp. 24–5; RC, p. 629.

134 AA, 340–2.

135 AA, 349–50, 347, 343–5.

136 AA, 342; E. A. Hanawalt, ‘Norman views of eastern Christendom: from the First Crusade to the Principality of Antioch’, in V. Goss and C. C. Bornstein, The Meeting of Two Worlds (Michigan, 1986), pp. 115–21, stresses Tancred’s pragmatic attitude to Eastern Christians.

137 RC, 634; AA, 345.

138 FC, p. 89.

139 FC, pp. 90–1; AA, 352–3; on Edessa see M. Amouroux-Mourad, Le Comté d’Edesse 1098–1150 (Paris, 1988) Chap. 1, ‘Fondation et Evolution du comté d’Edesse 1098–1150’, pp. 57–91; J. Laurent, ‘Des Grecs aux croisés; étude sur l’histoire d’Edesse 1071–98’, Byzantion, 1 (1924), 347–449. Tritton, ‘Anonymous Syriac Chronicle’, p. 70, says that Baldwin was sent by Godfrey who had been asked for help by Thoros.

140 Matthew, 37; Fulcher alone of the contemporary sources tries to pretend that Baldwin was not a party to the plot, pp. 91–2, and WT, 158–9 follows him; AA, 354–5.

141 Damascus Chronicle of the Crusades, pp. 42–3.

142 See above p. 133 n. 35.

143 See above, pp. 1–25 and below, pp. 310–11.

144 Matthew, 33; GF, 33, 37, 70, 48.

145 See below, pp. 261–2.

If you find an error please notify us in the comments. Thank you!