Although the traditional horse-archery of the Central Asian nomads had an influence on the Muslim world from the 8th to 11th centuries, it was relatively minor. Such a process was merely a continuation of those mutual influences that the cultures of settled Iran and nomadic Turkistān had upon one another for a thousand years. Turks entered the armies of Islam, either as mamlūks to be trained in Muslim traditions of warfare, or as mercenaries to fight alongside other warriors who used different tactics. Not since the expansion of the Parthians did a people from within the true nomadic zone have a more widespread influence. Even the Huns, despite their victories, had in reality been contained on the fringes of the settled area.
       The Saljūqs, who were part of the Ghuzz Turkish people, broke this pattern by first defeating their Ghaznawid employers and then going on to take over huge areas of eastern Islam. They were not merely conquerors, however, for they were occupiers as well. With the Saljūq tribal warriors came Saljūq clans, families and flocks. Spreading across much of the land, these Turcoman nomads brought not only their own culture but also the means whereby they could replenish their own ranks over numerous generations. The Turks became, in fact, a new and self-sufficient element within the world of Islam, rather than remaining a source of recruitment from beyond the Muslim frontier.
       Although the Saljūqs were to have a profound cultural and political impact on eastern Islam in the 11th century, they do

442

not at first appear to have had much effect on pictorial sources. This continued to portray a mixed cavalry armed with spears, shields, swords and bows (Figs. 363, 387 and 419). The one source that can be linked to a specific early Saljūq sultan, Malik Shāh I, shows only horses and not riders (Fig. 278).
       At this stage those Turcomans who were true nomadic horse-archers still formed the bulk of Saljūq armies, and the effectiveness of their tactics is well recorded. Their arrows clearly carried a great distance.¹ This, for example, enabled them to indulge in "zone shooting", or the dropping of arrows within a designated area such as the interior of a castle. ² A willingness on the part of these warriors to vary the style of their shooting according to circumstances probably accounts for some otherwise contradictory evidence. The Byzantine writer, Anna Comnena, claimed that Turkish arrows had extraordinary penetrating power,³ presumably when shot from close range, whereas various Crusader chroniclers maintained that they lacked penetrating power,⁴ presumably when used from long range to harass a marching foe or wound his horses. Unlike the later European long-bow, the Turkish composite bow may have relied on more powerful and regular tension, rather than the weight of an arrow, for its effect. The rate of shooting achieved by those nomad horse-archers was, however, agreed by all observers.⁵
       Another important feature of nomad horse-archery in the 11th
------------------
1. Anon., Gesta Francorum, p. 49.
2. Faris and Elmer, op. cit., p. 132.v 3. Buckler, op. cit., p. 428.
4. Smail, op. cit., pp. 80-82.
5. Ibid.

443

and 12th centuries was the widespread use of a nāwak or majra arrow-guide to shoot short arrows known as ḥusbān in Arabic or jawāldūz in Persian. Such a device probably shot those "darts" which, distinguished from javelins by the anonymous author of the Gesta Francorum, were noted at the battle of Dorylaeum.⁶ Such darts were again recorded by another anonymous Frankish chronicler at the battle of Araūf a century later. Those were also almost certainly ḥusbān.⁷ Interestingly enough, such short arrows are shown being used by a demon on one of the Nazareth capitals. Here the Christian artist, presumably lacking knowledge of just how such ḥusbān were shot, seems to give his demon either magical powers or suicidal tendencies (Fig. 269).
       It may also be worth noting that these Seljūq Turks seem to have been regarded by their Byzantine foes as more chivalrous and civilized than were west-European Crusaders.⁸ Although the Saljūq use of nomad horse-archery was very successful in battle, it rarely brought victory on its own and a final charge into close-combat was generally also needed. This would similarly be the case if the Turcomans were themselves defeated or trapped. Muslim sources concerning the battle of Manzikert in 1071 AD, which was won primarily by the professional rather than Turcoman army of the Great Saljūq sultan, describe that ruler as putting aside his bow, taking mace and sword and putting on coif and helmet before such a final charge.⁹ Crusader chroniclers agree that before coming
------------------
6. Anon., Gesta Francorum, loc. cit.
7. W. F. Paterson, "The battle of Arsuf," Journal of the Society of Archer-Antiquaries, VIII (1965), p. 22.
8. Buckler, op. cit., p. 421.
9. C. Cahen, "La Campagne de Mantzikert, d'après les Sources Musulmanes," Byzantion, IX (1934), p. 634.

444

to close combat the Turks put their bows aside, though by hanging these weapons on their shoulders rather than using bow-cases they seem to identify themselves as mamlūks instead of nomadic Turcomans. Such warriors would then take mace and swords perhaps also throwing their light lances as javelins before a final shock.¹⁰
       By this time, of course, the nomadic Turcomans had largely been relegated to the frontier zones because they had proved as politically turbulent to their Saljūq leaders as had the Arabs to the ʿAbbāsid Caliphs before them. In such frontier regions, which included Anatolian Rūm and at least the mountainous Taurus stretch of the new Crusader frontier, these Turcomans continued to use their martial energies as ghāzis.¹¹
       The state established by the Saljūqs of Rūm in Anatolia was originally such a Turcoman ghāzi province. Soon, however, the Saljūqs of Rūm threw off any real allegiance to the Great Saljūqs of Iran and established a dynasty of their own which long outlived that of their eastern cousins. They in their turn then tried to push the true Turcomans into a troubled frontier belt even further west, between Byzantine and Saljūq territories. Meanwhile, they themselves built a professional army similar to that of other Middle Eastern Muslim states. Nevertheless, Turcoman warriors remained more important in Anatolia than elsewhere, although the Sultan's force of Greek slaves and captives converted to Islam, plus foreign mercenaries from east and west and local Muslim ikdīsh militias of mixed ethnic origin, grew in importance throughout the 12th and early 13th centuries.¹² The varied
------------------
10. Smail, op. cit., pp. 82-83.
11. Cahen, "Djays," loc. cit.
12. T. T. Rice, The Seljuks, (London 1961), p. 81.

445

character of these armies is apparent in the art of both the Saljūqs of Rūm and their immediate neighbours to the east, the Dānishmandids (Figs. 252, 253, 254, 255, 256, 260, 261, 262 and 264).
       The cultural and military impact of these Turks upon their Christian neighbours is also apparent. It is seen, for example, on a coin from late 13th century Georgia (Fig. 426). Correspondingly, the influence of such long-established Christian states, including Byzantium, is equally obvious in the organization and tactics of the new army recruited by the Saljūqs of Rūm in the late 12th and early 13th centuries. A warrior's training in horsemanship, archery and fencing, added to the fact that an élite were creamed off as horse-archers while the rest continued to train in spear and sword, seems to reflect traditional Muslim customs.¹³ Tactics reverted to Middle Eastern tradition rather more slowly. In defeating the Byzantines at Myriokephalon in 1176 AD, for example, the Saljūqs of Rūm still relied primarily on Turcoman tactics. Yet at Erzinjan in 1230 AD when, in alliance with the Ayyūbids of the Jazīrah they defeated the Khwārazmshāh, and when they themselves were defeated by the Mongols at Köse Da'gh in 1243 AD, the Saljūqs of Rūm fought in a traditional, almost Byzantine and certainly non-Turcoman, style.¹⁴
       In fact, there is every reason to suppose that influences were mutual in the cultural melting pot that was Anatolia. Despite clashes, such as that at Myriokephalon, relations between the Saljūqs of Rūm and the Byzantines were generally amicable. Whereas the Saljūqs recruited many ex-Byzantine prisoners and also Greek mercenaries, the Byzantine Emperor recruited captured Turks and
------------------
13. Rice, The Seljuks, loc. cit.
14. Cahen, Pre-Ottoman Turkey, pp. 230-234.

446

Arabs, among others, plus "allied" contingents from various Saljūq and non-Saljūq rulers of Anatolia.¹⁵ Horse-archers of apparent Turkish origin were clearly commonplace to many Byzantine craftsmen in the 12th century (Figs. 233, 234, 235). They, or perhaps their distant Pecheneg cousins, were also known to the illuminators of 13th century Serbia who could place them among other heavy cavalry of obvious Byzantine inspiration (Fig. 634).
       Traditional Turcoman styles persisted among nomadic Turkish tribes in many parts of Anatolia and are well described in the probably late 13th [century] anonymous Book of Dede Korkut. Here equipment included many-coloured shields, swords, long-pointed lances, long mail hauberks, helmets, bows and arrows, quivers, daggers and, judging by their noise, iron lamellar cuirasses.¹⁶
       Saljūq influence did not, however, mean a revival of horse-archers in Byzantine armies, except for those provided by Turkish mercenary or allied corps. By contrast there was, in fact, an emphasis on lance-armed cavalry of rather European, almost Frankish, style. This perhaps reflected a loss of the Empire's main Anatolian recruiting grounds and the necessity of relying on the only type of cavalry now available to combat the various threats facing Byzantium.¹⁷ Most frontier regions, east and west, were for a variety of reasons heavily depopulated in the 12th century. The depression of much of the free peasantry into servitude also hindered internal recruitment and all these reasons conspired
------------------
15. Chalandon, Les Comnènes, Études sur l'Empire Byzantin, vo1. II, pp. 611-613 and 617.
16. Anon., The Book of Dede Korkut, pp. 92, 94, 145, 157 and 166.
17. Chalandon, Les Comnènes, Études sur l'Empire Byzantin, vo1. II, p. 237.

447

to force large numbers of foreign mercenary troops upon Byzantium.¹⁸ Cavalry, mercenary or indigenous, seem to have been divided into light and heavy, the former perhaps being horse-archers of Turcoman or Pecheneg origin. The latter certainly rode larger Arab or Hungarian horses, wore mail hauberks, helmets laced to coifs, and fought with lances, maces and swords while carrying large kite-shaped shields.¹⁹ Various 12th and 13th century illustrations from Byzantium and culturally related peripheral areas such as Bulgaria, Cyprus, Serbia and the Crusader states, show the development of a heavy cavalry whose equipment was akin to, though distinct from, that of western Europe (Figs. 226, 227, 229, 231, 232, 236, 271, 275 and 634).
       While the Turcomans of Rūm were downgraded in the face of Byzantine influence, those of Iran and the Fertile Crescent enjoyed an even briefer era of political and military supremacy. The huge realm of the Great Saljūqs started to fragment even before the end of the 11th century, before, in fact, the Crusaders appeared in the Middle East. Although this dynasty retained control of parts of Iran and Iraq until the late 12th century, larger areas fell to various successor dynasties. These did, however, generally continue the Saljūq system within their armed forces. Yet it would already be quite wrong to see such Saljūq military traditions as Turcoman. Saljūq forces were early divided into two distinct parts, a small professional ʿaskar, of mamlūks or mercenaries who owed allegiances to whichever ruler or local governor recruited them, and a larger element of tribal auxiliaries of whom the Turcomans were the most effective element. While the ʿaskars
------------------
18. Ibid., vo1. II, pp. 611-618.
19. Ibid., vo1. II, pp. 619-620.

448

formed a standing army of cavalry and infantry, garrison troops and personal guards, the auxiliaries would only be called upon for a specific campaign.²⁰ That this process started soon in the history of the Saljūq state may be inferred from the fact that the Great Saljūq Sultan, Ṭughril Beg, was using siege engines as early as 1054 AD during his first unsuccessful attack on Manzikert.
       The varied equipment of a Saljūq regional ʿaskar is probably illustrated on an early 13th century carved gateway from Sinjār (Fig. 280). Other sources which may also illustrate professional ʿaskarīs are to be found in 12th century Iran (Figs. 367, 368, 370, 371, 374 and 381-383). The best description of Saljūq equipment, in this case probably referring to 11th or 12th century Turcomans rather than to an ʿaskar,²¹ is to be found in the romance of Warka wa Gulshāh. Here the weapons include javelin, spear, sword, bow, mace and lasoo, while armour is comparably heavy, including helmet, coif or aventail, and full hauberk.²² The late 12th century illustrations of this Warka wa Gulshāh manuscript correspond in most points to its slightly earlier text (Fig. 422), Such equipment is mirrored, either exactly or with minor variations, in other Iranian and neighbouring sources (Figs. 377, 378, 305, 424 and 425).
       Of all those atābeg or "senior officer", successor states that inherited so much of the crumbling Saljūq empire, that of the 12th and 13th century Zangids of Syria and the Jazīrah was perhaps the most active. In the history of the Crusades such names as ʿImād al Dīn Zangī and his son Nūr al Dīn Maḥmūd loom very
------------------
20. Smail op. cit., pp. 66-67.
21. Ayyūqī, op. cit., pp. 5-6, 21-23 and 29.
22. Ibid., pp. 38-39 and passim.

449

large indeed. The Zangids, of course, had to recruit from a more limited area than had their Saljūq predecessors. The same may be said of their Būrid and Artuqid rivals in Damascus and Diyār Bakr respectively. This was certainly reflected in their armies. In 1126 AD a force from Damascus even used the old ʿAbbāsid tactic of having each cavalrymen carry a foot soldier into battle on his horse's crupper.²³ Some decades later the army of Nūr al Dīn would, judging from the origins of its senior men, have largely consisted of Turcomans and Kurds, horse-archers and heavy cavalry respectively, plus ghulāms trained in traditional style.²⁴
       Large numbers of auxiliary cavalry were also recruited from the Arab bedouin tribes.²⁵ Similarly, many Turcomans such as those of the Yaruqī tribe, who were invited to the Aleppo region in the mid-1120s, could be regarded as auxiliaries.²⁶ At this time such Turcoman tribes were, to all intents and purposes, outside Muslim civilization though living within the world of Islam. Like those German barbarians who earlier took over the western regions of the Roman Empire, these Turcomans retained a separate legal system of customary law. This yāsa, as it was known, was not officially abandoned until Nūr el Dīn obliged his Turkish military élite to adhere to Muslim law,²⁷ and such a situation was bound to help preserve the Turcomans' separate
------------------
23. Smail, op. cit., p. 76.
24. Elisséaff op. cit., pp. 722 and 729.
25. Ibid., p. 730.
26. Ibid., p. 722.
27. A. N. Poliak, "The Influence of Chingiz-Khān's Yāsa upon the General Organization of the Mamlūk State," Bulletin of the School of Oriental and African Studies, X (1942), pp. 862-863.

450

identity in military, as in others matters.
       Certain Turcoman tactics similarly persisted even in the professional ʿaskars of various minor dynasties around the Fertile Crescent. When facing regular armies not comparably trained to use nomadic horse-archery techniques, this meant a constant harassment of the foe until he was so disorganized or demoralized that a final decisive charge could conclude the struggle. This the Zangids and others used against the Crusaders, although they were also occasionally prepared to meet their foes in a set-piece battle of organized ranks and, of course, to engage in siege warfare.²⁸ Another fundamental tactical change that might betray Turkish Saljūq influence was the placing of cavalry ahead of infantry as an army marched through hostile territory.²⁹ In earlier centuries the reverse appears to have been the case.
       As for as the general equipment of cavalry in atābeg Syria and the Jazīrah was concerned, the poetry of the period may, because of an inherent conservatism, tend to reflect surviving Arab and perhaps Kurdish elements rather than the newly arrived Turcomans, Certainly there is much more emphasis on swords, turbans, tark helmets, rumḥ and qanāh spears, than upon bows in such verses, though the latter are, for example, mentioned in connection with Nūr al Dīn's army.³⁰ The same is very much the case in the art of atābeg Syria and the Jazīrah (Figs. 270 and 281-283).
       However real such a traditional, non-Turcoman, character might have been of certain atābeg states, it was certainly true
------------------
28. Smail, op. cit., pp. 78-80; Elisséeff, op. cit., pp, 740-743.
29. Elisséeff op. cit., p. 738.
30. Ghaith, op. cit., pp, 34, 43, 89, 106, 125-126 and 137.

451

of Arab Shayzar in the early and mid-12th century. In his memoirs, centred upon this castle, Usāmah ibn Munqidh listed a horseman's equipment as a kazāghand fabric-covered mail haubergeon, a khūdhah helmet, a sword slung from a baldric, a rumḥ spear and a turs shield.³¹ The importance of the lance is also made clear by Usāmah elsewhere. For example, he devoted several pages to notable lance thrusts and explains that horsemen particularly feared to turn their backs on a foe armed with a spear. Horsemen also feared a lance if they themselves had only swords. Other warriors he recorded as using compound lances of exceptional length.³² Usāmah further indicated that the Frankish couched-lance technique was known, though not necessarily widely used, by the Muslims.³³ A somewhat garbled description of such lance technique, using a relatively short qunṭārīyah weapon with a broad blade, appears in al Ṭarsūsī's treatise on military matters. This was written for Ṣalāḥ al Dīn about the same time as Usāmah wrote his memoirs.³⁴
       Ṣalāḥ al Dīn began his career as Nūr al Dīn's governor in Egypt. With the death of the last Fāṭimid Caliph of Cairo in 1171 AD, however, he not only changed the official faith of the country from shīʿī to sunnī Islam, but set about recruiting a new army loyal to himself rather than to the memory of the Fāṭimids or to Nūr al Dīn and the Zangids. Such considerations necessitated Ṣalāḥ al Dīn recruiting from an even wider spectrum than was normal. He had inherited a Fāṭimid force that included many thousands of Armenians, Sudanese and Arabs, both regular and
------------------
31. Usāmah, op. cit., p. 85.
32. Ibid., pp. 41, 48-50, 70-78, 87, 90, 92 and 131.
33. Ibid., pp. 41-42.
34. Al Ṭarsūsī, op. cit., p. 113.

452

auxiliary, plus Kurds, ghulāms and Turcomans brought to Egypt by Ṣalāḥ al Dīn's uncle during the Zangid takeover.³⁵ As his power grew, however, Ṣalāḥ al Dīn downgraded, destroyed or disbanded most of the Fāṭimid forces but retained those Zangid troops who were willing to be loyal to him rather than to Nūr al Dīn. He also continued to recruit an increasing number of free Kurdish tawāshī heavy cavalry, Turcoman horse-archers and mamlūk ghulāms. As Ṣalāḥ al Dīn gradually took control of most of Syria and the Jazīrah, he incorporated the mixed forces of these areas into his own loosely knit army, to which the aḥdāth militias, muṭṭawiʿah volunteers and Arab bedouin auxiliaries of these regions could also be added.³⁶
       Crusader chronicles, or at least those who have interpreted them, tend to overemphasize the admittedly picturesque if not spectacular role of Turcoman horse-archers in the armies of Ṣalāḥ al Dīn and his Ayyūbid successors. These troops seem, however, to have played a relatively minor role in the warfare of Egypt and the Fertile Crescent in the late 12th and early 13th centuries. In fact, they zoom to have degenerated into one of two sources of auxiliary cavalry, the other being the bedouin Arabs. The most successful role for such Turcomans may now have been as raiding troops, riding ahead of an invasion force as they did during the conquest of Palestine after the battle of Ḥaṭṭīn in 1187 AD.³⁷
       The most effective horse-archers, and in fact most of the
------------------
35. Gibb, "The Armies of Saladin," loc. cit.; A. S. Ehrenkreutz, Saladin, (New York 1972), pp. 51 and 73-74.
36. Ibid.
37. ʿImād al Dīn, op. cit., p. 33.

453

troops, in Ṣalāḥ al Dīn's army were now the more disciplined mamlūks. They, however, seem to have used their bows in much the same way as had long been traditional in the Muslim east, though, perhaps, with a greater tendency to shoot on the move it not at the gallop. Armed head to foot, such mamlūk horse-archers could be sent ahead of each battalion as an advance guard of scouts.³⁸ Others were trained to dismount and shoot at a foe while drawn up in ranks to achieve greater range and accuracy³⁹ If unhorsed in battle they would continue to fight first with their bows and at the last with swords.⁴⁰
       Such behaviour would be in line with their training as reflected in the Ayyūbid military treatise of al Ṭarsūsī and, indeed, in later furūsīyya manuals. Al Ṭarsūsī, for example, advised a horse-archer to aim at the horse of an armoured foe but wait until an enemy cavalryman with a sword got very close before shooting as one could not afford to miss with one's first shot. If, however, this mounted foe were charging with a lance, or with a nāwak arrow-guide and ḥusān short arrow, the horse-archer should maintain his distance or at least have sword and shield ready to defend himself. Generally speaking, the foeman with a lance was considered the most dangerous and the one who should be dealt with first.⁴¹
       Oddly enough there seem to be more illustrations of mamlūk horse-archers from the later Ayyūbid era, in the 13th century,
------------------
38. Ibid., p. 19.
39. Anon., Itinerarium Peregrinorum, p. 92.
40. Bahā al Dīn, op. cit., pp. 81-82.
41. Al Ṭarsūsī, A. Boudot-Lamotte trans., Contribution à l'Etude de l'Archerie Musulmane, (Damascus 1968), pp. 142-144.

454

than from the days of Ṣalāḥ al Dīn himself. Their equipment would appear to have been fairly standardized, with a minority wearing heavy armour and riding horses possibly with bards and chamfrons (Figs. 129, 291, 300, 304, 307, 308 and 651). Elsewhere such horse-archers appear to use crossbows, as in early 14th century Granada, or lances. Those equipped with the latter weapon often ride heavily caparisoned or barded horses (Fig. 287).
       This varied Kurdish or mamlūk cavalry, heavy and light, seem to have fought in close cooperation. Shujʿān horsemen, perhaps including the horse-archers and crossbowmen, delivered controlled charges while their withdrawal was covered by an élite of probably armoured horsemen known as abṭāl. This seems to have been an elaboration of the earlier Arab karr wa farr tactic that itself perhaps reflected the Byzantine system of cursores shock-cavalry archers supported by defensores to protect their flanks.⁴²
       A great many Ayyūbid heavy cavalry, excluding those from the ruler's own mamlūks, seem to have been numbered among those contingents drawn from the Jazīrah area. This was, of course, close to the homeland of the free Kurdish tawāshīs. During the siege of ʿAkkah, Muʿizz al Dīn of Sinjār, one of the surviving Zangid rulers of this region, led a force of horsemen armed with qanāh lances and swords, wearing sawābigh jubbah mail hauberks and tarik helmets, possibly with crests, but with no mention of bows.⁴³ Even Ṣalāḥ al Dīn's foes noticed that the cavalry of Taqī al Dīn, the sultan's nephew, were not horse-archers.⁴⁴
------------------
42. Ṭarsūsī, op. cit., p. 126; Bivar, "Cavalry Equipment and Tactics on the Euphrates Frontier," pp. 288-290.
43. ʿImād al Dīn, op. cit., p. 254.
44. Anon., Itinerarium Peregrinorum, p. 94.

455

On the other hand, the last Artuqids of Ḥien Kayfā in the northern Jazīrah may have retained the horse-archery techniques of their Turcoman forebears.⁴⁵ The late 12th and early 13th century art of this area shows warriors and equipment as mixed as were the origins of the troops using it. Those armed with spear or sword could carry shields of purely Byzantine style (Fig. 285), wear a variety of mail and lamellar armours (Figs. 286, 288 and 306), have their forefingers over the quillons of their swords in purely Iranian style (Fig. 291) and also use their lances in many different ways (Figs. 287 and 303).
       Those of specifically Kurdish origin are, on at least one occasion, described as wearing dirʿ hauberks and carrying large, very convex leather yalab shields.⁴⁶ Elsewhere the origins of Ayyūbid heavy cavalry are not so specific, although their equipment is described in detail. The kumāh, or "veiled", horseman fought outside ʿAkkah with swords.⁴⁷ In such actions, using swords, luṭūṭ maces and spears, they were clearly quite capable of unhorsing heavy European knights.⁴⁸
       It is rare for one minor incident and one individual warrior to be recorded in both Arab and Crusader chronicles. Yet this happened with the death of the Ayyubid champion Ayāz the Tall during the battle of Qaisarīyah on 30th August 1191 AD. Ayāz had earlier been described as fully armoured⁴⁹ and when, during
------------------
45. Mubārakshāh, op. cit., p. 268.
46. ʿImād al Dīn, op. cit., p. 375.
47. Bahā al Dīn, op. cit., pp. 90-91.
48. ʿImād al Dīn, op. cit., pp. 267-270; Bahā al Dīn, trans. F. Gabrieli in Arab Historians of the Crusades, p, 202.
49. ʿImād al Dīn, op. cit., p. 304.

456

this battle, he was thrown from his horse he was struck down before being able to remount because of the weight of his iron, armour.⁵⁰ The rest of Ayāz's equipment included a bow, quiver, sword and a spear that was heavy enough to be noted with astonishment by his Frankish slayers.⁵¹ Comparable equipment, including mace and swords was still used by Ayyūbid mamlūk regiments half a century later at Manṣurah.⁵² Such heavy cavalry, mamlūk or otherwise, plus its equipment, appears in much art of the Ayyūbid era (Figs. 161, 173, 174, 300, 301 and 651).
       Ayyūbid light cavalry equipment was as varied as that of heavier troops, though naturally less abundant. Those described as jarīdah carried the lightest equipment and were employed for rapid raids into enemy territory or to hold isolated outposts where they could also act as infantry.⁵³ Such troops were often Arab auxiliaries who, noted for their speed and manoeuvrability, were very effective in ambushing enemy convoys.⁵⁴ These warriors were described by their Crusader foes as despising armour on the grounds that it was an attempt to escape one's predestined day of death.⁵⁵ Other comparable troops, Arab or Turkish, were similarly lightly equipped and fought with bow, winged or knobbed mace, sword, dagger and light spear.⁵⁶ Indeed
------------------
50. Ibid., p. 381.
51. Ibid., Anon., Itinerarium Peregrinorum, p. 84.
52. Ibn Wāsil, op. cit., pp. 289-290.
53. Gibb, "The Armies of Saladin," loc. cit., ; ʿImad al Dīn, op. cit., p. 399.
54. Bahā al Dīn, op. cit., pp. 161 and 342; Anon., Itinerarium Peregrinorum, pp. 131-132.
55. Norris, op. cit., p. 99.
56. Anon., Itinerarium Peregrinorum, p. 78.

457

the light spear, often of bamboo, was regarded even in early 13th century India as the typical weapon of the Arabs.⁵⁷ Illustrated sources of the 12th and early 13th centuries from Egypt and the Fertile Crescent show many lightly equipped troops, both in apparently Arab costume (Figs. 162, 310 and 325) or in probably Turkish dress (Figs. 290 and 295) or of no particularly identifiable origin (Fig. 329).
       A comparable decline in the importance of nomadic Turcoman horse-archery took place in the east of Islam following the fragmentation of the Saljūq empire in the mid-12th century. In the Ghaznawid state, of course, such troops had never been more than one element in a mixed army.⁵⁸ The Ghūrids who overthrew this latter dynasty in the second half of the 12th century were famed more for their infantry than for their cavalry, yet horse soldiers did have an important part to play even in the early 12th century. At this time some were clearly heavily armoured,⁵⁹ while by the end of that century at least Ghūrid leaders generally fought on horseback with spears.⁶⁰ Mounted troops from similar regions in Afghanistān, tājiks and khaljīs continued to serve the subsequent mamlūk dynasty of "Slave Kings" in northern India. Here they were apparently noted for the long shamshīr swords that they kept in scabbards beneath their saddles, a fashion that may indeed have been known in the earliest centuries of Islam.⁶¹
------------------
57. Mubārakshāh, op. cit., pp. 260-261.
58. Bosworth, The Ghaznavids, pp. 109-110.
59. Maulānā Minhāj, al Dīn, op. cit., p. 351.
60. Ibid., p. 461.
61. Mubārakshāh, op. cit., p. 259; Ashgar, op. cit., pp. 137-138; Ibn Isḥraq, op. cit., p. 153.

458

Such saddle-swords may have been illustrated in late 12th century Turkish Azarbayjān (Fig. 422) and early 13th century northern Iran (Fig. 391). Such a weapon certainly appears on a carving of uncertain, but possibly 11th to 13th century, date at Orissa in eastern India (Fig. 491). Elsewhere, the cavalry of eastern Islam seem at this time to have been similar to those of the central regions (Fig. 369).
       Apart from that uncertain carving at Orissa e, which shows a bowcase on the saddle in early, almost pre-Islamic style, there seem to be no representations of horse-archery in the sub-continent during those centuries. Yet horse-archers, mamlūks rather than nomadic Turcomans, were vital to the Delhi Sultanate. Such heavily armoured troops fought with bow, spear, javelin, mace, lassoo and sword.⁶² That archery had reached a high point of perfection is suggested by the great variety of arrowheads used by these warriors. Some were designed specifically to penetrate mail of the lamellar of jawshan. Others were used against shields of wood, cane or leather or in the long-distance archery of siege warfare, Still more were designed either to pierce a quilted khaftān or similarly mode horse-armour, to slay an unprotected foe, or be used against a man wearing heavy iron armour.⁶³ There is, in fact, no reason to suppose that such a degree of sophistication was not normal among all mamlūk archers, though not necessarily among the tribal Turcomans.
       Further north, among those originally Turkish mamlūks and still nomadic Qipchaqs and Qangalis who together formed
------------------
62. Ashgar, op. cit., p. 35; Digby, op. cit., pp. 15-16.
63. Mubārakshāh, op. cit., p. 241.

459

the backbone of the Khwārazmshāh's army, similar heavy cavalry apparently predominated in the early 13th century. These warriors were soon to fail against the Mongols, but in 1212 AD they succeeded in destroying the Buddhist Qarā Khitai dynasty that ruled over much of Muslim Turkistān. The Qarā Khitai could indeed have been regarded as a very Chinese-influenced vanguard of the Mongol horde that was so soon to follow. On this occasion, however, the Khwārazmshāh's army largely consisted of heavily armoured cavalry riding equally armoured horses.⁶⁴ Many may also have been horse-archers, although the nīzah spear of willow wood seems to have been a major weapon among such troops.⁶⁵
       The art of mid- and late 13th century Iran generally illustrated the traditional military equipment of this region rather than that of the recently arrived Mongol conquerors. They rarely appear in such sources until the mid-14th century. Such an interpretation would seem to be supported by arms and armour in early Īl-Khānid manuscript illumination.⁶⁶ This is particularly true of the so-called Red-Ground Shāhnāmahs, most of which probably date from the early or mid-14th century. Iranian sources from rather earlier, in the 13th century, seem to illustrate cavalrymen who, though variously armed, were clearly within the Saljūq tradition (Figs. 263, 390, 392, 395, 396, 397, 400, 401, 402, 403, 405, 406 and 645A & 645B). In all probability the heavier cavalry of the Khwārazmshāhs, with their characteristic use of horse-armour, appear in one unique west-Iranian, though rather later, manuscript
------------------
64. Maulānā Minhāj, al Dīn, op. cit., p. 262.
65. Mubārakshāh, op. cit., pp. 260-261.
66. D. T. Rice, Islamic Art. (London 1965), pp. 114-115 and 122; E. J. Grube, The World of Islam, (London 1966), pp. 103-104.

460

known as the Kitab-i Samak Ayyār (Fig. 641). Here the equipment is probably a development of both Saljūq and earlier styles, perhaps grown heavier through long experience in warfare against Central Asian nomadic horse-archers in Turkistān.
       Interestingly enough, an almost exactly comparable process was once again occurring north of the Eurasian steppes, in response to the same threat that Turkish nomadic horse-archers posed, By now, of course, it was not only the Christians of Kievan Russia who were so threatened. The isolated Muslim state of Bulghār, with its well-established armaments industry at the confluence of the Kama and Volga rivers, was being squeezed between Russia and the still largely animist Qipchaq and Qangali nomads of the steppes. Some Bulghārs do, in fact, appear to have been converted to Christianity in the early or mid-12th century, while their primitive Finno-Ugrian forest-dwelling neighbours to the north were so converted by the 13th⁶⁷ (Fig. 635).
       In 12th and 13th century Kiev itself, heavier cavalry armour, the widespread adoption of the long sabre, helmets with rigid and life-like visors, horse-armour and in particular the chamfron, and to some extent the separation of horse-archers and heavy close-combat cavalry into separate corps, were all apparent. Once again such developments are reflected in a great deal of surviving equipment from this area (Figs. 622, 625, 636 and 637), equipment that may well mirror developments to the south, judging from surviving Russian art sources (Figs. 626-633).
       In a second Christian area that bordered the Muslim world, the Crusader states in Syria and Palestine, there seems by contrast
------------------
67. I. Hrbek: "Bulghār," Encyclopedia of Islam, second edition, vol. I, pp. 1304-1308.

461

to have been almost complete stagnation, except insofar as military changes reflected developments in western Europe.⁶⁸ During the Crusaders' first offensive phase, existing techniques proved adequate, while in the later defensive years the Franks were generally forced to rely on counter-siege warfare that left little scope for tactical innovation. Such developments as did occur, including the recruitment of turcopoles, seem to have been learned from the Byzantines rather than developing as original concepts.⁶⁹ The turcopoles were not horse-archers in the Turcoman, nomadic
tradition, though many may have carried bows in mamlūk, Arab or Byzantine style. Their primary role seemed to have been as light cavalry who fought either as scouts and skirmishers or alongside other mounted Frankish troops⁷⁰ (Figs. 272 and 275). Even the placing of cavalry outside closely packed ranks of infantry⁷¹ may be a reflection, at least in part, of comparable tactics by Zangid forces.
       While the Mamlūk state of Egypt and Syria successfully concluded the Ayyūbid offensive against the Crusaders, these Mamlūks also refined Ayyūbid developments in military technology and tactics. By the end of the 13th century, having defeated the Franks, confined the Mongols and overcome numerous lesser foes within the Middle East, they were clearly among the most successful troops in the known world. Their superior patterns of logistics, armaments, tactics and discipline were to provide the foundation of a military tradition upon which later Mamlūk
------------------
68. Smail, op. cit., pp. 113 and 116-118.
69. Ibid., pp. 111-112; Vryonis, op. cit., pp. 133-134.
70. Smail, op. cit., pp. 110-112.
71. Ibid., pp. 156-157.

462

and Ottoman successes were to be built.⁷²
       In this not new but constantly refined military tradition, the role of cavalry was clearly paramount. Although infantry were still considered important, horsemen bore the brunt of offensive warfare and large-scale manoeuvre in which their speed, striking power and the weight of their weapons were considered superior.⁷³ Even if an army had to rely on its infantry when surprised by a foe or when drawn up in set-piece battles the final outcome always depended upon Mamlūk cavalry.⁷⁴ Since this was their primary role, it is not surprising to find that most Mamlūk furūsiyya manuals laid greater emphasis on training in the use of the lance than on any other weapon, even including the bow.⁷⁵ Very much the same was true in early Mamlūk art from Syria and Egypt, where spears and swords are generally more common than bows (Figs. 175, 177, 178, 180, 647 and 648).
       Even where archery was described, such furūsiyya manuals clearly show that it was not in the Turcoman style. Rather it seemed to be a development of earlier Byzantine and ʿAbbāsid traditions. When shooting, the Mamlūk archer rolled up his presumably armoured sleeve to make it easier to bend his arm and also, perhaps, to avoid snagging the bowstring.⁷⁶ Although this was exactly paralleled in European travellers' descriptions
------------------
72. Scanlon, A Muslim Manual of War, p. 21.
73. Al Aqṣarā'ī, op. cit., p. 325; al Anṣārī, op. cit., p. 72.
74. Al Aqṣarā'ī, op. cit., p. 107.
75. N. Rabie, "The Training of the Mamlūk Fāris," in War Technology and Society in the Middie East, V. J. Parry and M. E. Yapp edits., (London 975), pp. 156-157.
76. Ibid., p. 158.

463

of the Mongols,⁷⁷ it was surely no more than a common-sense precaution to be undertaken by all archers if they were wearing lamellar amour. Not until the mid-14th century did a Muslim treatise on archery put greater emphasis on shooting from horseback rather than while standing, kneeling, squatting or sitting. Most such works seemed to be more concerned about the combined use of shield and bow in sieges and set-piece battles.⁷⁸ In fact, it was apparently the practice of mounted Mamlūk archers in the 14th century to draw up in ranks, dismount, empty their quivers onto the ground and then shoot from a squatting or kneeling position. In this manner each rank protected the one ahead from being overrun by the foe.⁷⁹
       Such Mamlūk mounted archers were also trained to shoot from horseback, if need be, in all directions. When this was done on the move, however, it generally seems to have been done from close range as demonstrated in one furūsiyya game or exercise known as the naban.⁸⁰ An even closer-range type of horse-archery was practiced in the gīnhaj exercise. Here the target lay on the ground and was apparently shot at as the rider virtually rode over it.⁸¹ These were clearly not harassment techniques of archery. Rather they were shock tactics, and this may account for the appearance of crossbows in the hands of at least one
------------------
77. G. D. and A. M. Gaunt, "Mongol Archers in the Thirteenth Century," Journal of the Society of Archer-Antiquaries, XVI (1973), p. 21.
78. Latham and Paterson, "Horse-Archers of Islam" loc. cit. ; J. D. Latham and W. M. Paterson, Saracen Archery, an English version and exposition of a Mameluke work on archery (ca. A.D. 1368), (London 1970), p. 101.
79. Ibn Khaldūn, "Muqaddimah," in Bosworth, "The Armies of the Prophet," p. 205.
80. Ayalon, "Notes on the Furūsiyya Exercises and Games in the Mamluk Sultanate," pp. 55-56.
81. Latham and Paterson, "Horse-Archers of Islam", loc. cit.

464

such Mamlūk horseman from a Syrian source. Among the many other illustrations of late 13th and early 14th century horse-archers from Egypt, Syria and southern Anatolia, some certainly seem to be shooting at targets almost beneath their horses' hooves (Figs. 131,177A, 178B, 309, 650B and 652).
       Of course the Mamlūk state in Egypt and Syria also employed light cavalry. Some acted as scouts, in which case they would neither wear the dirʿ hauberk nor carry a turs shield but be armed only with bow and arrows.⁸² Such warriors were probably mamlūks (Figs. 177 and 179). Others, perhaps a great majority, would have been tribal auxiliaries. Here one may find the only real survival of Central Asian nomadic horse-archery, for many Turcoman as well as Kurdish tribes were paid to protect the frontiers of Syria, Palestine and Lebanon from invasion. Comparable Arab bedouin tribes were engaged to watch the borders of Syria, Sinai and Egypt.⁸³ The most readily recognizable auxiliary warriors in the art of the era appear to be Arabs, although it is also possible that some Turcoman tribes were already adopting those bedouin customs and dress that were to render them virtually indistinguishable from their fellow nomads in these regions in later centuries (Figs. 133, 311, 639 and 642).
       Finally there are the Mongols themselves. Their weapons, though strictly outside the scope of this study and owing more to Chinese than to Muslim traditions, were naturally to have a profound impact upon the Middle East. Mongols were nomad horse-archers above all else, although they were competent in all forms of warfare
------------------
82. Al Ansārī, op. cit., p. 80.
83. Poliak, Feudalism in Egypt, Syria and Lebanon, 1250-1900, pp. 9-10.

465

including sieges. Various features of equipment and tactics were noted both by European travellers and Muslim foes. Mongol warriors, for example, were said to use shields of wickerwork or cane only when on guard-duty and not while engaging in horse-archery.⁸⁴ They or their Central Asian allies also carried large bow-cases slung from their belts, which most Mamlūk horse-archers apparently did not. Mongol troops also had a habit of sliding their light lances beneath their saddle-girths when using other weapons.⁸⁵ Elsewhere Mongols are described as carrying one or two bows, no less than three quivers, plus an axe and a lassoo. Richer warriors used slightly curved single-edged swords and rode horses protected by mail or lamellar peytrals, crinets and chamfrons though not, apparently, cruppers. Other horse-armours were of leather up to three layers thick. Protection for the rider could consist of a metal helmet with a leather aventail, and a cuirass of iron lamellae.⁸⁶
       These styles would, however, not immediately appear in the art of those Muslim areas overrun by the Mongols. Nor, according to Rashīd al Dīn, would the well-established armaments industries of Iran learn to make new forms of armour for their new masters until the 14th century. Even when they did so, they also apparently continued to manufacture shields and various other items in traditional style for those local dynasties which survived under Mongol and Īl Khānid suzerainty⁸⁷ (Figs. 410, 638, 639 and 642-644).
------------------
84. Gaunt, loc. cit.
85. Al Aqṣarā'ī, op. cit., pp. 328-329.
86. M. Komroff, Contemporaries of Marco Polo, (London 1928), pp. 47-48.
87. Gorelick, "Oriental Armour of the Near and Middle East from the eighth to the fifteenth centuries as shown in works of art," loc. cit.

466

Nevertheless, in north-western Iran the presence of a new Mongol capital city later greatly encouraged the further development of an existing local arms manufacturing industry.⁸⁸ Sword-making had, for example, apparently been rudimentary in this region prior to 1300 AD.⁸⁹
------------------
88. Allan, op. cit., p. 67.
89. Ibid.