5) The sources: Summary.
To sum up, there are nearly twenty sources which can be used as evidence for the battle itself. The number of the sources where one can trace the construction of the image of the enemy is much smaller. In the following two chapters I will reconstruct the events of the battle and then attempt to describe/analyze how the image of the enemy on the battlefield was reflected in the narrative sources described here.
1) Historical background
1161 was a year of triumph for Constantinople. For the first time in the history of the Eastern Empire the sultan of Ikonion came to pay homage to the Byzantine emperor.52 This event is one of the crucial points in the Byzantino-Seljuk relation in the twelfth century, which pointed to a change in the relations between the empire and the state of the Seldjukids of Rum. Now it was an open peace and union: Kilic Arslan II became one of the “sons” of Byzantine emperor and the recognized suzerain of the Seljuk part of Asia Minor. The sultan understood this literally and began to widen his domain, to which he added the principalities of the Danishmendids (1168).53 The latter tried to defend themselves by forming an alliance with Constantinople and with the powerful atabeg of Aleppo, Nur-ad-Din, whose land bordered the Sultanate of Ikonion on the east. The armies of this talented warlord prevented Kilic Arslan II from any serious actions, but only until 1174. In this year the master of Aleppo died. His followers were much more interested in seizing power than in the defense of the small principalities of Paflagonia. This was a chance for Kilic Arslan II, who initiated aggressive actions against them.54 One of the scions of Danishmend ran to Constantinople.
He came in time: Emperor Manuel was preparing a huge expedition to the east. The idea of it was present already at the time of Alexios Komnenos. His son John made an expedition to northern Syria, where he secured Byzantine suzerainty over Antiochia. Manuel himself visited Antioch and supported good relations with the kings of Jerusalem . The balance of forces in Palestine was not good for the Crusader States; at the times of Moslem advance they needed help. The Byzantine emperor was one of the few persons who could provide the needed support. More than that, he wanted to do it. A successful expedition would have raised his prestige in the eyes of the Western leaders and thus helped in the hard geopolitical game which Manuel was waging against Frederick Barbarossa in Italy. During the conflict he supported friendly relations with Alexander III, who was (by the very character of his status) an active supporter of the Crusades.
The main problem was that emperor could not simply start a war in the Near East. First he needed to establish a balance between two centers of Seljuk power – that of the Danishmendids and that of the Sultanate of Ikonion. Such balances allowed him to play on the rivalry between them and by this secure the Byzantine provinces of Asia Minor, which were left without the main imperial forces. In the beginning of the 1170s the balance was obviously shaken: Kilic Arslan captured most of the lands of the scions of Danishmend.
This is more or less the historical background; what follows is an attempt to reconstruct the events of the battle of Myriokephalon. In the research process I will try to avoid common points; the question of the results of the battle will not be discussed in great detail. I also dedicate relatively short parts to the military analysis of the event – the latter deserves a special work, which I hope to write in the near future. I have tried to avoid references to general books on the history of Byzantium.
2) Objectives, preparations, and the battle-plan.
a. The Byzantines
The actual date of the emergence of the project of the 1176 expedition is once again under debate. The preparation for the main expedition began not earlier than the spring of 1175, when both the Byzantine55 and the non-Byzantine sources place the gathering of the army, the intense recruitment of new warriors, the calling of the most distant allies, such as the Latins of Antiocheia, Hungarians,56 Serbians and Seljuks,57 the recruiting of Cumans,58 and the assembling of a huge siege-train in the military camp on the Asian side of the Bosporus.59 Some scholars have argued that this was clearly an anti-Seljuk project, the only aim of which was the occupation of Ikonion.60 I think, however, that this is only partly true.
First, all of Byzantine politics of the twelfth century was never aimed at the total annihilation of the Seljuk states. Komnenian emperors sometimes supported the Danishmendides against Ikonion, sometime allied with Ikonion against the Danishmendides, and regularly captured strategic points on the border. They were slowly moving into the peninsula – but in reality never seriously tried to annihilate the nomads. The idea here was probably the same which American forces use in the modern Iraq – it is easier to deal with two or three rival chieftains, than with one strong leader. Manuel was an exception; he believed that Kilic Arslan will be a good vassal. As the events proved, this was a mistake.
Second, none of the contemporary Byzantine sources states that the aim of Manuel was the taking of Ikonion itself. Rather the opposite; Epistola clearly states that emperor decided to make war on all the Persians (not only on Ikonion). Kinnamos gives rather ambiguous information about fear of the sultan, and, at the same time, speaks about ships which were sent to Egypt.61 In this context, the hypothesis of R.J. Lilie, who clearly states that the whole expedition was a crusade, is not baseless.62
In my opinion, even stronger support can be found in Kinnamos’ account of the embassy of a certain eunuch, Thomas,63 who went to the sultan with a mission, but was not successful.64 I think that this can be interpreted as follows: Manuel tried to ask for a “safe conduct” for his troops on their way to the Armenian lands and probably further, to Palestine. The sultan refused him because he was truly afraid that the Byzantine army would turn against his own land.
Another reason to believe Lilie is the fact that the machines and the army gathered for the expedition were disproportionately large for a mediocre city with stone walls such as Ikonion. The very formulation of the letter of Manuel also suggests rather a war against “all Muslim people of the east” than against the Seljuks of Asia Minor.65
Thus, the initial aim of the whole campaign is not as clear as the pretext for the war itself. This is well described by Michael the Syrian. According to him, the reasons for the beginning of the campaign were a raid by Kilic Arslan II, which was caused by the rumors of Manuel’s death and by an appeal of Dhu l-Nun Danishmendid to Manuel Komnenos as the recognized supreme sovereign of the Seljuk land. The scion of Danishmend made an appeal in connection with the domain of his recently deceased brother, which had fallen into the hands of the master of Ikonion.66 Another problem were bordering towns, which were supposed to be handed over to Byzantines, but were not. An embassy to Kilic Arslan headed by Thomas did not yield any clear result. From this time (d.p.q. 20 July), the high officers of the imperial army began to make concrete plans for the war against the Seljuks. The emperor was already on the border; he erected two fortresses: Dorylaion and Soublaion. Some scholars have interpreted the rebuilding as part of a general campaign of refortification, while others have interpreted it as it as part of a larger crusade project which was ended at Myriokephalon.67 One fact does not totally exclude the other, but, for the reasons mentioned above, the version of a Crusade seems preferable to me.
The sultan tried to play similar politics. The official embassy met Manuel at Dorylaion, rebuilding the fortress, which he refused to leave. This was answered immediately: Turcoman tribes of the borderland carried out great devastating raids on the southern Byzantine provinces, which had been left unguarded.68 From this moment war became unavoidable: Manuel had to answer the offences by the sultan. The latter realized his unfavorable position, but the embassy, probably led by the vizier of the sultan, Hiyas-ed-din Ibn Gabras, which arrived in Constantinople in the winter of 1175-1176 was in vain.69
In the spring the war started again. A large regiment of Byzantine forces led by the above-mentioned Dhu l-Nun Danishmendide and by a nephew of the emperor, Andronicus Vatatzes, tried to attack Neokaisarea, but were turned away thanks to the cunning of the local population. Their representatives first advised the Byzantines not to believe the Danishmendid and later attacked the camp of the imperial soldiers, sent the warriors into flight, and killed the nephew of the emperor, who was one of his favorites.70
The reaction of the Byzantine emperor was predictable. Choniates draws a beautiful picture of Manuel’s wrath: “Making his army so great, he prepared to the war in the way, as if he wanted to destroy Persian nation, level Ikonion with its walls, make the Sultan submit and put his foot on his neck.”71 This image finds support in Michael the Syrian: “Quand l’empereur des Grecs Manuel apprit que son neveu avait été à la porte de Néocésarée, il partit en colère pour venir tirer vengeance des Turks.”72 The Syrian chronicler, thus, gives a short description, while his Byzantine colleague draws a more detailed (and partly ironic) picture.
The possible interpretation of the whole group of events seems to me to be the following. The emperor was preparing an expedition against his enemies, not only the sultan, with whom he had tried to reach an agreement about the passage of his troops (this was probably the aim of the embassy of Thomas). He was already having a border conflict with the sultan; the death of an important relative and friend at the hands of the Seljuks made the revenge unavoidable. Nevertheless, it started later than was planned. The “official version” of Kinnamos states that at first the whole enterprise was designed for the spring, but was delayed because of the low speed of some auxiliary troops and problems with Pannonian horsemen.73 Another problem for the army may have been the solar eclipse which occurred in the spring of 1176. It might have instilled some fear into the emperor, who, being a great admirer of astrology, postponed the whole enterprise for a more suitable time.74 When the war began in the summer without great success for the Byzantines and the expedition to Neokaisarea failed, he led the army to the south while a large naval force set sail to Palestine.75
The route taken by the Byzantine army, according to Choniates, was the following: from the camp of Lopadion, where the whole mass of horses and people was gathered, the army took the road to Chonae, then through the plains of middle Anatolia to the newly-built Laodicea of Pisidia (see map 1).76 This was one of the possible solutions for this trip; another route, repeating the path of the first crusade, went through Dorylaeum-Philomelion and then by the road which goes on the northern side of the modern Sultan Daglari to the valley of Ikonion, and from there further to any other possible aim of this expedition.77 Why did the emperor’s strategists select this road? Kinnamos does not offer a clear justification for the route chosen, while Choniates, as a good writer, spared his readers a long description of the way. Michel the Syrian also offers no clue, but from his account it can be inferred that this was the expedition of a huge army over a devastated land, where the enemy had destroyed and poisoned all the resources that could have been used by the imperial troops.78 I think that the selection of the route was an attempt to kill two hares with a single bullet; on one hand, by moving southwards the Byzantine army prevented the enemy from attacking the flourishing cities on the Meander, as other enemies had done some time before, while the northern corner of border was defended by Dorylaion.79 At the same time, the imperial troops from the southern provinces could easily join the army on its way to the enemy borders.
Sometime in August the army began to move in the designated direction, but the way proved to be not so easy. Warfare in the dry steppe lands of the border zone80 has extensive logistical requirements, such as water and grazing land. There were problems with water (which was probably stored for the springtime expedition) and food for the horses (which was destroyed by enemy raiders).81 In addition to this, the road was not safe because a short time before hostile nomads had devastated this land and no akritai or border guards are mentioned in any source.82 Probably these are the problems to which Manuel refers in his letter to Henry Plantagenet: “…although much of our preparations were made, not as we should have wished, nor as appeared best suited to our object…”83 Despite these problems, the first clashes with small groups of Turcomans were successful. After leaving behind its last base, the city of Laodicea, on the 16th of September,84 the huge multinational force reached the mountain pass where it was destined to meet its doom.
According to Choniates, before the battle Manuel summoned a military council. The ambassador of the sultan arrived in the camp with the final suggestion for peace.85 This time conditions seem to have been regarded as at least possible by the Byzantine side, because they were discussed by the council. “Experienced men” in the council, probably thinking about an army greatly harmed/afflicted by dysentery and lack of food, advised the emperor to agree with these conditions, but because of some unnamed youngsters, who were very influential, the ambassador left without success.86 Manuel wanted war and was ready for it. The next question was practical, how to pass a long defile and the mountain plateau that came after it before taking the straight road that led to Ikonion (map 2). The question was well-known in the Byzantine theoretical military tradition. Unfortunately there is no direct evidence to tell whether Manuel ever read any of these treatises, each of which clearly advised against the course of action that he eventually took.87 This earned him some blame from Choniates, who states that the emperor moved his army without any measures of reconnaissance, clear knowledge of the enemy, or an exact plan.88 All these things were repeated by different authors in the secondary literature.89 One of the few exceptions in this respect is John Haldon, who suggests that Manuel probably had an idea that he would be attacked at Myriokephalon by the Seljuks, which, as the events of the next day show, was what actually happened. He just did not manage to predict the direction of the attack.90
In 1147, the young Emperor Manuel Komnenos had led an expedition against the Sultan of Ikonion, Mas’ud I (1118-1152). It was nearly successful: the Seljuk forces were destroyed in two battles in central Anatolia, the Byzantine forces captured Philomelion and, for a significant time, Byzantine regiments even besieged Ikonion, which had not seen warriors from Constantinople in such large numbers since the time of Mantzikert (1071). The sultan retreated to the north, where he asked for the help of the Danishmendids. Finally, after long maneuvers of near Ikonion, Manuel decided to retreat through a mountain pass which lay to the west of the city and was called Cyblicymani by the Seljuks. The Byzantine army built a camp on the eastern end of the defile, while the emperor, moved by his bravery, tried to find a good opportunity for a show of valor to impress his Latin wife. This nearly led to disaster, and if the help of his brave entourage had not come just in time, the reign of the emperor would have been much shorter. Manuel was almost surrounded, but managed to escape and counterattack the forces of the sultanate and the Danishmendids, who, at that time were allied with Ikonion. After a hard battle, the imperial forces put the enemy to flight and began their march through the mountains. The main Seljuk forces, after several fruitless attacks, went back to their capital, while Manuel, trailed at some distance by enemy scouts, successfully reached the plain of Lake Pousgusa.91
I think it possible to elaborate on the statement of John Haldon. The experience of 1146 was probably one that Manuel (and his generals) used in planning the battle.92 From their point of view, the most vulnerable point of the defile, as in 1146,was the exit to the east. The unfortunate experience acquired during another expedition to Sarapata Miloniensis, when the Byzantine regiments were attacked by archers who occupied the peaks of low mountains, was not taken into consideration.93 The messages of the reconnaissance forces (if there were any) did not reach the ears of the emperor.94 As a result, the next morning, a vanguard of the Byzantine army led by a most experienced commander, John Doukas Vatatzes, I infer, received an order to go quickly and occupy the exits of the valley.95 The emperor and/or his advisors expected trouble to come from the east. Reality was a bit different. They were attacked from above.
b. The Seljuks.
The text above is meant to give an idea of the Byzantines before the battle. What was on the other side? Did the sultan want the war? Probably he did not. A full-scale war with the empire, whose army was much bigger than all the possible forces of the sultanate, was not at all his main aim. He knew the empire well: his mother was an Orthodox princess, probably of Russian origin.96 From the very beginning of his reign, his ways and methods were aggressive, but quite cunning and sophisticated. In his politics he had an experienced ally; during one of his wars, a scion of the Byzantine family of Gabrades changed sides and joined the sultan’s court, where he probably converted to Islam and, probably some time later, became a vizier of the country under the name of Hasan Ibn Gabras.97 He was at least the head of the diplomacy of the sultanate and an important ambassador before, during, and after battle. His connections were great; one of his relatives had been killed by imperial troops in the first battle at Myriokephalon in 1146, and another one was imprisoned in Constantinople just before the second battle at the same place.98 Not much is known about how this person affected the decisions of his sovereign, but one thing that is clear is that his relation with Byzantium was, to put it mildly, not very friendly.
The aim of this alliance of a Seljuk prince and a Byzantine aristocrat during the war of 1176 was to secure the results of recent conquests, namely, Mardin, which was captured at the end of the July of 1175, and to avoid a large-scale war with the empire.99 At first they tried to negotiate terms for peace, but Manuel was angry; the embassy of Gabras to Constantinople did not yield any results. At the same time, the border war seemed to be successful. The only mistake was probably the spilling of the blood of young John Vatatzes, which seems to have provoked the anger of the Byzantine emperor and led him to start the expedition.
The sultan answered with a whole complex of actions. Kavus-name, a source from the end of the eleventh century which was most probably read by the Seljuk sultans100 and their entourage, gives a leader the following advice:
You should be informed about the kings of neighboring lands; it is necessary that in all the countries your friend and you enemy should not drink a sip of water without their people notifying you about it. About their kingdom you should be informed in the same way as about the kingdom of your king.101
This was probably the way Kilic Arslan II and his advisers went about the business of war. Their preparations are described by Michael the Syrian; he states that the sultan ordered his troops to go in front the army of the emperor in small groups and to destroy all possible crops. Special attention was dedicated to poisoning the wells with the bodies of the dead oxen and dogs.102 Another order was dedicated to fortresses; the Seljuks were supposed to defend them for a short time and then to retreat, setting the forts on fire.
On the basis of these orders one can draw several conclusions. At first, Kilic Arslan II’s estimation of the enemy seems to have been much clearer than his opponent’s. From the beginning of the expedition he knew the strong and the weak points of the enemy army. He knew well the possible problems with food and water and ordered these resources to be destroyed, but at the same time he realized that the enormous siege machinery which was carried by Manuel’s warriors could easily destroy any of his fortresses. There is no indication from which direction he awaited the attack, but Byzantine sources clearly show that Turks destroyed crops and poisoned wells along the whole route of the army--and this means that they knew it very well. The chosen strategy was quite traditional Seljuk hit-and-run--in the same way the army of Alexius Komnenos near Philomelion was accompanied by small groups of scouts under the Emir Monolikos.103 In some parts the description of Michael the Syrian is reminiscent of Byzantine military treatises, especially one of them, De velitatione bellica, which recommends a similar way of supporting constant contact with an enemy army, which enables the general to know many things about the enemy and to control him at the same time.104 This was the case here as well; Kilic Arslan II was well-aware of the size of the enemy army and called for help from eastern lands.105
Was there a moment of fear for the sultan? It certainly seems so. He tried to start negotiations three times. Despite all his military talents, Kilic Arslan II clearly understood that the army which was led by the Byzantine emperor was formidable. The battle plan suggested by the Seljuks was for them probably the only solution: to block the heavy Byzantine cavalry and the many wagons in a narrow place, attack them with a few gulyams [heavily armored elite horsemen] and a multitude of less well-protected Turcomans, while the archers would shower them with a rain of arrows.106 The same stratagem had already been used when mobile regiments, headed by the emperor himself, were attacked in a narrow defile near Sarapata Miloniensis, but there the Byzantine cavalry managed to regroup and smash through the enemy which blocked the way.107 Despite the absence of direct evidence about the battle plan and the method of coordination between the different Seljuk regiments on the next day, it seems clear to me that there was a plan of some sort.
As the next day proved, the Seljuks seemed to know nearly everything about the enemy: the quality and number of its warriors, the ranks of the leaders, and, it seems, even the order in which the Byzantine army was going to enter the defile. The sources of their information could have been many: the Turcomans, who harassed Manuel’s army on its way, the survivors of the minor battles which had taken place on the borders of the sultanate, and probably even spies, who could have been present in the large and loosely organized mass of imperial warriors.108 From the very beginning of the campaign and well before the battle Kilic Arslan II had a much clearer idea of the enemy than Manuel--and this led him to a victory that, it seems to me, he did not expect.
3) The Battle109
a. The Byzantines
On the morning of 17 September 1176, the imperial army began to enter the long defile of Myriokephalon. The marching order, according to Choniates, was the following: the first regiment of the army, headed by John and Andronicus Angelos, then the second headed by Constantine Macroducas110 and Andronicus La(m)parda,111 which probably consisted of experienced eastern border guards. The right wing was headed by the brother of the emperor’s wife, and the left wing by Theodor Mavrosomes.112 They were followed by the wagons, the supplementary units, then by the emperor with his personal guards, and, finally, the Byzantine order was completed by the forces of the rearguard under the command of Andronicus Condostephanos113 (figure1) The first two regiments successfully passed through the difficult places, and, according to the special order, which I think was given to them by Manuel, moved forward at high speed, to occupy positions at the end of the defile.114 They were attacked by the Seljuks. Nevertheless, the imperial archers defeated the enemy who “withdrew” back to their first positions. The next unit of the Byzantine army was attacked immediately after its passage; its inexperienced commander, Baldouin of Antioch, exhibited miracles of bravery, but without great success. According to Choniates, this was a key moment in the battle; the next phase was actually a catastrophe. The Turks attacked the core of the Byzantine army, namely, the left and right wings, the baggage, and personal guards of the emperor from the left, right, and above. The arrows poured down from the skies like rain, killing horses, people, and, what was more important, and the oxen that drew the wagons with siege weaponry.115 The situation was extremely dangerous; the enemy seemed to be everywhere and this provoked some understandable fear. At the same time, the way was also blocked from behind. Byzantine sources are silent concerning the moves Andronicus Condostephanos made, but from the chronicle of Michael the Syrian it can clearly be seen that the rearguard of the army was also attacked by a great multitude of Turcomans, who can be expected to have paid special attention to the carriages with food and water.116 The Byzantine army was blocked from all sides (figure 2).
Even so, I do not believe that this was a moment of general panic. The emperor finally got a clear idea of the situation and tried to organize some kind of opposition to the attacks of the Seljuk cavalry. When re-formed heavy regiments removed Seljuk blocking groups, the experience of Sarapata Milonienses was tried, but the carriages blocked the place which was needed for the heavily-armed knights to attack. This was the moment of truth for the whole army, and panic began to rule the mass of people which only some hours ago had been the army of the Only Emperor in the world. He was later brave enough to describe its state in very dramatic sentences. The situation became even worse when the army was shown thee head of Andronicus Vatatzes, who was killed near the gates of Neokaisareia.117 Even at this disastrous moment, the emperor, who did not lose hope, attacked some Seljuks on the slopes of the mountains, but unsuccessfully. The head of the Byzantine army and his bodyguards were stranded on a hill surrounded by deep ravines; a sandstorm blinded the warriors of both sides, who fell into the chasms.118 Finally, realizing that his situation was hopeless, the emperor gathered some soldiers and tried to break through the enemy regiments.119 The rest of the battle was rather a massacre than a defeat. In darkness and dust the Byzantines and Seljuks filled gullies, ravines, and hills with the dead bodies of their friends and enemies. Everyone was thinking about the safety of his own life.120
The problem here lies in the next episodes of Choniates. Magdalino calls them “anecdotes” and speaks accurately about “indirect criticism” of Nicetas.121 Jakov Ljubarskij is much more radical; he clearly states that pieces were invented by the author for the sake of his ideological structures.122 I incline to agree with the latter. Both episodes (with the bloody river and with the rude warrior) are built in the basis of the traditional motifs of Byzantine Kaisekritik, which will be addressed in chapter 3. Thus, in these episodes one should not believe Choniates as much as many scholars have in the past.
Manuel, who managed to break through a group of Gulyam guards (I think that these were the troopes he confronted), was not alone.123 Several groups of Byzantine warriors somehow managed to save themselves from the massacre (figure 3).124 These groups (which seem not to have been small) joined the emperor. He organized them into some likeness of order and, after several clashes with the Seljuks, who tried to capture him, reached the camp at the end of the defile, where fresh new units of the Angeloi brothers and Andronicus Lampardas waited for their emperor.125 By the evening of 17 September several other important people, including the opysthophylax Andronicus Condostephanos, managed to reach Manuel’s camp. Seljuk control over the whole defile was complete and the Turcomans began to rob the bodies. On the fortified hill, however, the situation was different; despite the panic and destruction of the main army, the Byzantine emperor still saved a notable force, which enabled him to speak from a position of dignity, if not power. Strangely enough, the position of Kilic Arslan at that time was not much better.126
Figure 1. The beginning of the battle at Myriokephalon.127
The divisions of avant-garde headed by Angeloi, Makroducas and Lampardas are passing defile on a high speed.
The rest of Byzantine army (corps of Balduin of Antioch and Mavrozomes, huge siege train ) are entering the defile.
The column is covered from behind by forces of Manuel and a regiment of Andronicos Condostephanos.
The Seljuk on the tops and slopes of the hills are preparing to start their attack.
Figure 2. The apex of the battle at Myriokephalon.
Avantgarde of Byzantine forces fortifies on the hill on the eastern end .
Regiments of Mavrozomes and Balsuin of Antioch are destroyed.
Siege train is under attack and blocks the movement of Manuel’s guards.
Turcomans attack the division of Andronicos Condostephanos.
Figure 3. The flight of the emperor.
1.Manuel is trying to escape.
2. Small groups of Byzantines are trying to find a road to the eastern end of the defile.
b) The Seljuks
The position of Kilic Arslan II and his warriors differed from that of Manuel in many ways. First, the sultan was not directly on the battlefield. Choniates states that the Seljuk sultan was always cautious and never put his head in danger by engaging in direct combat with the enemy.128 A friend of the sultan, Michael the Syrian, also states clearly that at the beginning of the day the Byzantine army was in view of Ikonion or at a distance of three hours from the headquarters of the sultan.129 Second, Kilic Arslan clearly knew what the enemy was doing and which part of the army should be attacked, and this was noted especially by Choniates.130 The forces of the sultanate were probably positioned thus: The main heavy cavalry forces were situated on the slopes of the hills and in the ravines between them. This was a typical Seljuk stratagem of sequential ambushes – if the enemy could manage to defend itself from the strike of the group in one ravine, the latter retreated and attacked the enemy again from the next ravine.131 The archers and stone-throwers on top of these and the whole enormous mass of Turcomans was ready to block the imperial army from the back. The whole plan worked quite well; the first regiments, headed by experienced commanders, were let pass, while the second part of the army, which was under the command of less experienced leaders, was attacked, probably by the light cavalry. A counter-attack of the Byzantine forces was stopped, not without some losses, while the main archers (probably mounted in the traditional Seljuk way) poured a rain of arrows on the horses and oxen (figure 2). At another location, where the “observer of Michael the Syrian” (this person probably was the source of information which was used by Michael the Syrian) was observing the fight, huge stones were thrown down on the enemy.132 All these actions, especially the shots aimed at the wagons, indicate a high degree of coordination between the different parts of the Seljuk forces and, once more, a clear appreciation of the most vulnerable points of the imperial army. Despite this, the Byzantines managed to crush some of the Seljuk battle groups.133 Finally, after the destruction of the divisions of Maurozomes and Balduin of Antioch (which cost the Seljuks many lives), gulyam guards faced the remnants of the Manuel’s personal warriors, who had fortified themselves on a hill. This moment, perhaps, was crucial for the battle: a sandstorm saved the emperor of Constantinople. The chasm became chaos.
The picture painted by Choniates gives the clear idea that from that moment on the battle became rather a series of clashes in which, it seems to me, the Seljuks were given the order not to kill the emperor. This hypothesis is based on several pieces of information: a) the fact that the emperor passed through the Gulyam regiment nearly alone, but was not killed; b) the fact that in one of the anecdotes told by Choniates a Seljuk warrior does not try to kill the emperor, but rather to capture him with his horse. This fits well into Seljuk ideology, which did not advise a clever ruler to kill the enemy. This attitude perhaps derived from the double position of the Seljuks of Asia Minor, who considered themselves as subjects of both the caliph of Bagdad and of the emperor of Byzantium at the same time.134 The person of the supreme ruler was almost holy, and moreover, despite the war, the relations of the Seljuk sultan with the Byzantine emperor on a symbolic level were still the relations between father and son.
While the Byzantine emperor was assembling his forces on one side of the defile, the whole territory seems to have become the ground for Turcoman plunder, examples of which can be seen in the story of Choniates.135 Were these Turkmen behaving as the loyal subjects of their sultan at this moment? I think not, and if this were so, the situation of Kilic Arslan II was no less dangerous than on the morning of that day; a horde of marauding Turcomans was probably even more dangerous for him than the remnants of the Byzantine army. Seljuk leaders, however, succeeded in organizing at least part of these forces at the eastern end of the defile. The Byzantine camp was blocked. The remnants of the army in the mountain pass were either slaughtered or taken into captivity. This was the end of the military actions at Myriokephalon in a strict sense, but merely the beginning of the armed diplomatic bargaining about the conditions of peace which took place in the next few hours.