You are on page 1of 16

3

The Square “Fighting March” of the Crusaders


at the Battle of Ascalon (1099)

Georgios Theotokis

On 12 August 1099 the Latin knights and footsoldiers of the First Crusade left
Jerusalem to meet the Fatimid army of the grand vizier Al-Afdal which, at that
time, had invaded Judaea and had encamped close to the coastal city of Ascalon.
The army was estimated to be around twenty thousand strong, including both
infantry and cavalry.1 This would be the irst of several major expeditions by
the Egyptians launched against the Crusader states in Palestine, all entering
through Ascalon and its coastal plain. The Latin leaders were irst alerted about
a possible large enemy force approaching from the south on 9 August and on
the next day the Crusader armies began their forty-kilometer march south to the
city of Ascalon where the enemy was last reported to have camped. According
to Raymond of Aguilers, one of the main chroniclers of the First Crusade and
an eye witness of the events, the Latins numbered 1,200 knights and no more
than 9,000 footsoldiers and “they marched in nine ranks, three to the rear, three
to the front, and three in the middle so that attack would be met in three ranks
with the middle one always available to bolster the others.”2
In this paper I will examine a number of theories about the origin of this
particular marching formation, based on the manuals attributed to the Byzantine
Emperors Maurice (582–602), Leo VI (886–912) and Nicephoros Phocas (963–
69) and several anonymous Byzantine military treatises of the sixth and tenth
centuries. Also, I will seek to understand why the Latins adopted this particular
formation and how effective it would have been against the two enemies of the
Crusaders in the Holy Land, the Seljuks and the Fatimids. The question whether
we have any other examples of the use of this “ighting march” after the battle
of Ascalon will also be considered.

1 John France, Victory in the East, A Military History of the First Crusade (Cambridge, 1994),
pp. 359–60; Hans Delbrück, History of the Art of War, 4 vols. (London, 1990), 3: 407.
2 Le “Liber” de Raymond d’Aguilers, ed. John H. Hill and Laurita L. Hill (Paris, 1969), pp.
156–57; Raymond d’Aguilers, Historia Francorum qui ceperunt Iherusalem, trans. J. H. Hill
and L. Hill (Philadelphia, 1968), pp. 133–34. For the battle of Ascalon, see: France, Victory in
the East, pp. 356–66; Richard C. Smail, Crusading Warfare, 1097–1193, (Cambridge, 1995),
pp. 174–75; Charles W. C. Oman, A History of the Art of War in the Middle Ages AD 378–1485,
2 vols. (London, 1991), 1: 288–91; Delbrück, History of the Art of War, 3: 406–07.

Journal of Medieval Military History 11.indb 57 18/03/2013 15:31:46


58 Georgios Theotokis

In general, Western European armies did not always adopt the same marching
formations, but a rather similar pattern was followed depending on a number of
factors like the terrain, the number of men and divisions, the danger of being
attacked en route and the ratio between infantry and cavalry units.3 Usually we
would have had a column of units in which the cavalry marches ahead of the
infantry for better protection, keeping the baggage train to the rear, while lightly
armed mounted squadrons of lank-guards and rear-guards would be posted by
a prudent commander to prevent surprise attacks. But there is no evidence that
the marching formation described before the battle at Ascalon, a square forma-
tion of three divisions on each side, had ever been applied by a Latin army
before, either during the Crusade or by any other Western European army.4 And
the question that immediately arises is what would have caused this change in
battle-tactics?
Undoubtedly, this “ighting march” would have been the answer to the
threat posed by the Seljuks, an enemy that the Latins had faced at Doryleum
in July 1097 and then again at Antioch in the summer of 1098.5 The Seljuk
Turks were typical mounted steppe warriors who based their battle tactics on
four major principles. First, they took advantage of their speed and mobility
in the battle-ield which was largely due to their equipment, armor and horses
that were, according to a number of studies, much lighter that the Frankish.
Their principal weapon was the bow, but they also carried a shield, lance and a
sword. As for their armor, it is very dificult to know what they were wearing
as defensive equipment in the later decades of the eleventh century. But the
inluence from the Byzantines was becoming strong and from the First Crusade
the Latin chroniclers mention heavily armed knights with hauberks.6 They used
their mobility to attack their enemies on the march and while on rough terrain

3 For a good introductory study on this topic, see: David Nicolle, Crusader Warfare, vol. 1
(London, 2007), pp. 60–70; to compare with the Byzantine marching formations of the tenth
century, see E. McGeer, Sowing the Dragon’s Teeth: Byzantine Warfare in the Tenth Century
(Washington, D.C., 1995), pp. 329–41; J. Haldon, Warfare, State and Society in the Byzantine
World, 565–1204 (London, 1999), pp. 154–66.
4 I am grateful to Professor B. S. Bachrach for pointing out to me the possible classical inluence
of the Latin terms used by the chroniclers to provide information about the invasions of Lombard
Italy by Pippin (754 and 756) and Charlemagne (774), and the latter’s retreat from Spain in 778.
See: B. S. Bachrach, Early Carolingian Warfare: Prelude to Empire (Philadelphia, 2001), pp.
160ff; idem, “Charlemagne and the Carolingian General Staff,” The Journal of Military History
66 (2002), 313–57.
5 France, Victory in the East, pp. 170–85; Oman, The Middle Ages AD 378–1485, 1: 273–79;
Smail, Crusading Warfare, pp. 168–70; Delbrück, History of the Art of War, 3: 403.
6 There is signiicant evidence to show that all these were much lighter than the western European
ones: Smail, Crusading Warfare, pp. 77–8; France, Victory in the East, pp. 149, 204–5; Andrew
Ayton, “Arms, Armour, and Horses,” in Medieval Warfare, ed. Maurice Keen (Oxford, 1999),
pp. 190, 192–93; David Nicolle, “The Impact of European Couched Lance on Muslim Military
Tradition,” Journal of the Arms & Armour Society 10 (1980), 13. Nicolle also mentions the very
useful illustrations of a mid-eleventh-century Turkish manuscript named Warqa-wa-Gulshah. It
is also possible that the Turks might have developed a kind of light crossbow to throw darts at
their enemies to maximize their casualties: France, Victory in the East, p. 148, especially n. 23.

Journal of Medieval Military History 11.indb 58 18/03/2013 15:31:46


The “Fighting March” of the Crusaders at Ascalon (1099) 59

and their primary objective was to encircle or outlank the main body of the
enemy units using their numerical superiority – although this was not always
the case, with their battle-tactics giving the impression they were numerically
superior. Their attack was directed against the lanks and/or rear-guard of the
column, forcing the units of their enemy to ight while walking backwards and
being in great danger of being cut off from the rest of the army, in an attempt to
break the cohesion of the Latins – a tactic which, of course, developed gradually
during the twelfth century. Third, being expert mounted archers they were able
to apply the tactic of feigned retreat. Releasing constant showers of arrows from
a distance and falling back when their enemies charged forward to neutralize
them, but then pretending to retreat, they would make a sudden turn and come
back to harass them. Their aim was to confuse and demoralize the enemy, and
to isolate and break up their formations before charging in with their swords
and lances.7 The tactic of feigned retreat, however, was also frequently used in
Western Europe as well in that period and before – Hastings (1066) and Messina
(1061) can be seen as two characteristic examples from the eleventh century –
and it was to be applied by the Latins in the Middle East as well. The signiicant
difference, however, from the tactics used by the Turks was that the latter were
maneuvering constantly and in great numbers, with their retreat lasting for the
entire day or even days in order to weary their opponents and draw them to a
well prepared ambush, like the disaster at Harran (1104) where Bohemond and
Baldwin followed the Seljuk retreat into the sandy and hilly terrains east of
Harran – a serious tactical error considering their experience in Middle Eastern
warfare.
According to our Latin sources, these steppe tactics were completely unknown
to the majority of the Crusaders who certainly cannot have fought against any
large Seljuk force before8 – with the most notable exception being Bohemond
and men from his contingent.9 Thus, based on their experience in the previous

7 The most expert works on Seljuk warfare are: Nicolle, Crusader Warfare, 2: 107–69, especially
pp. 139ff; Reuven Amitai-Preiss, Mongols and Mamluks: The Mamluk-Ilkhanid War, 1260–
1281 (Cambridge, 2005); Smail, Crusading Warfare, pp. 75–83; Oman, The Art of War in the
Middle Ages, 1: 206–19, 273–74; France, Victory in the East, pp. 147–49; J. Waterson, The
Knights of Islam, the Wars of the Mamluks, (London, 2007), pp. 33–53, especially pp. 37–44.
The most detailed and accurate chronicle account of the mid-twelfth century concerning the
Seljuk ighting tactics: Anna Comnena, The Alexiad, trans. Elizabeth R. A. Sewter (London,
2003), XV. iii, pp. 479–80 (Alexiad).
8 This is attested by: Gesta Francorum et aliorum Hierosolimitanorum. The Deeds of the Franks
and the Other Pilgrims to Jerusalem, ed. Rosalind Hill (London, 1962), ix, p. 19; Fulcherius
Carnotensis, Historia Hierosolymitana, ed. Heinrich Hagenmeyer (Heidelberg, 1913), Book I:
XI: 6, p. 194; William of Tyre, A History of the Deeds Done Beyond the Sea, trans. and ed.
Emily A. Babcock and Augustus C. Krey (New York, 1976), vol. I, Book III, XIV, pp. 170–1;
Albert of Aachen, Historia Ierosolimitana, History of the Journey to Jerusalem, ed. and trans.
Susan B. Edgington (Oxford, 2007), Book II, chapter 39, p. 130.
9 Norman forces led by Robert Guiscard and his son Bohemond had defeated two thousand
Turcopole troops in battle on the eve of the battle of Dyrrachium (October 1081). They had
also faced Turcopole and Patzinak troops during their siege of Larisa (April–May 1083) in the

Journal of Medieval Military History 11.indb 59 18/03/2013 15:31:46


60 Georgios Theotokis

two years they decided to adopt a square formation of three divisions – not
columns but most likely, lines10 – on each side, and mixed11 their cavalry and
infantry units, probably placing an average of about 130 knights and 1,000 foot-
soldiers in each division. This, indeed, would have been very effective against
the Seljuk battle tactics and would have allowed the Latins to avoid a surprise
attack from the lanks or rear and form a solid mass formation of soldiers, with
the infantry forming into the perimeter to protect the cavalry units in the center
of the formation until the latter would have had the chance to break through and
deliver the decisive blow to the Turks.
However, the Crusaders were not about to face any Seljuk forces on their
march to Ascalon but rather a Fatimid invasion army. Thus, we have to ask
ourselves what was the make-up of the Egyptian armies in the late eleventh
century and what similarities – if any – do we may detect between the ighting
tactics of the latter and those of the Seljuks The armies of the Fatimids in
Egypt were a composite and multinational force that based its power on merce-
naries and large bodies of slave-soldiers like the Daylami infantrymen from
northern Iran, Turkish and Kurdish cavalrymen, while they also included signii-
cant numbers of Bedouins, Sudanese foot-archers and non-Muslim elements
like Armenians, Greeks and Slavs.12 Even though the Turkish element in the
Fatimid armies had formed a powerful elite that was favored by the govern-
ment of Cairo due to its specialization as mounted cavalry, they never played a
dominant role. The black slave-soldiers from Sudan were much more numerous
and their rivalry with the Turks in terms of ethnic origin, pay and prestige

same Balkan campaign. See: Georgios Theotokis, “The Campaigns of the Norman Dukes of
Southern Italy against Byzantium, in the Years between 1071 and 1108 A.D.” (unpublished
D.Phil. thesis, University of Glasgow, 2010), pp. 217, 247, 277ff. Also, western knights had
fought against Muslims in Spain who used ancient Bedouin cavalry tactics like the repeated
charges using javelins rather than bows and arrows: Nicolle, Crusader Warfare, 2: 143.
10 I ind Verbruggen’s argument that the crusaders formed three lines more convincing than
Delbrück’s columns. See: Jan F. Verbruggen, The Art of Warfare in Western Europe during the
Middle Ages, trans. Sumner Willard and Mrs Richard W. Southern (Woodbridge, 1997), pp.
208–09; Delbrück, History of the Art of War, 3: 406–7.
11 The argument that the units were mixed and that the infantry was not marching in front of
the cavalry is reinforced by the Gesta: “At daybreak on Friday [12 August], each [of the
commanding generals] drew up his own men, and foot-soldiers with archers were ordered to
precede the knights.” Even though this statement by the anonymous author of the Gesta is not
exactly clear, it probably means that during the march in the previous day the units were mixed,
which seems to me more reasonable than having foot-soldiers preceding the knights for forty
kilometres in two separate units: Gesta, p. 95. We also need to acknowledge here that putting
infantry forward to protect the cavalry in the battle-ield formations was common in that period
– see the examples at Hastings (1066) and Dyrrachium (1081).
12 For the organization, consistency and battle tactics employed by the Fatimids: McGeer, Sowing
the Dragon’s Teeth, pp. 225–46; Oman, The Art of War, 1: 208–16; B. J. Beshir, “Fatimid
Military Organization,” Der Islam, 55 (1978), 37–56; William J. Hamblin, “The Fatimid Army
during the Early Crusades,” DPhil thesis, University of Michigan, 1985; Yacov Lev, “Infantry
in Muslim Armies during the Crusades,” in Logistics of Warfare in the Age of the Crusades,
ed. John H. Pryor (Aldershot, 2006), pp. 185–206.

Journal of Medieval Military History 11.indb 60 18/03/2013 15:31:46


The “Fighting March” of the Crusaders at Ascalon (1099) 61

frequently developed into civil wars. By the end of the eleventh century, Turkish
mounted soldiers continued to be employed by the Fatimid Sultans, although
not in the same numbers as in the previous periods.13 But our attention should
be focused on the role of the Bedouin auxiliary troops in the Middle East, a
corps of soldiers that regularly accompanied the armies of the Fatimids. This
lightly armed cavalry wore little or no armor and were not armed with a bow
but rather used a short lance,14 while they are attested in several cases to have
been using very similar enveloping tactics to the Seljuks.15 In fact, according
to Fulcher of Chartres, they were the ones who opened the battle at Ascalon by
attacking the left lank of the Crusaders in an attempt to encircle them, thus the
Muslims were described as “a stag lowering his head and extending his horns
so as to encircle the aggressor with them.”16 They are identiied in the Praecepta
Militaria of Nicephoros Phocas with the term Arabitai (Αραβίται) to distinguish
them from the rest of the Muslim units (Agarenoi, Αγαρηνοί), and we read about
their battle tactics: “If the enemy proceeds in close order with their forces in
proper formation, bringing along a vast host of cavalry and infantry, and their
forces move in against one side of our units, the Arabitai will encircle our four-
sided [infantry] formation in a swarm, as they usually do, conident in their
horses.”17 Further, we read in the late tenth-century military treatise On Tactics:
“He [the leader of the rear-guard] should have them [his soldiers] ride around
a distance to the rear, so they too can be on the lookout and prevent the enemy
from making an unexpected attack from behind and to ward off the very bold
onslaughts of Arabs and Turks.”18
In what we have seen so far, the ighting march of the Crusaders was the
answer of the Latins to the threat posed by their enemies – in this case the
Fatimids, whose corps of Bedouin soldiers applied similar ighting tactics to
the Seljuks. But the major question is where we should trace their square (3x3)
marching formation? The three-line formation was not something new for the

13 McGeer, Sowing the Dragon’s Teeth, pp. 238–39, 246–48; France, Victory in the East, pp.
359–60; Lev, “Medieval Egypt”, pp. 120–22.
14 McGeer, Sowing the Dragon’s Teeth, p. 239; see also: David Nicolle, Arms and Armour of the
Crusading Era, 1050–1350: Islam, Eastern Europe and Asia (London, 1999).
15 Contrary to Smail, who says that “the Arab and Berber horsemen of the Egyptian armies did
not ight in the same way as the Turks.” But Smail is only using Maurice’s Strategikon and Leo
VI’s Tactica and completely ignores Phocas’ Praecepta, while his only Greek primary source is
the Alexiad of Anna Comnena. See: Smail, Crusading Warfare, pp. 83–7, 174–75; for a detailed
study of the Arab and Turkish encircling tactics, see: Nicolle, Crusader Warfare, 2: 122–51.
16 Fulcherius Carnotensis, Historia Hierosolymitana, Book I, XXXI. 6, p. 314; France, based on
the accounts of Raymond of Aguilers, the Gesta Francorum and the anonymous Chronicle of
Damascus, has suggested that it was Tancred’s charge in the midst of the Egyptian army’s camp
that opened the battle outside Ascalon: France, Victory in the East, pp. 363–64.
17 Presentation and Composition on Warfare of the Emperor Nicephoros, II. 101–10, in Sowing
the Dragon’s Teeth, p. 28 (“The Praecepta Militaria”).
18 The Anonymous Book on Tactics, in Three Byzantine Military Treatises, trans. George T. Dennis
(Washington, D.C. and Dumbarton Oaks, 1985/latest edition: 2008), chapter 10, 27–36, pp.
280–81 (“On Tactics”).

Journal of Medieval Military History 11.indb 61 18/03/2013 15:31:47


62 Georgios Theotokis

Latins in the East, although it was applied solely as a battle formation rather
than a marching one. It had been used extensively by the Romans, who formed
their phalanx formation based on the deployment of three lines of maniples;19
hastati, principes, triarii, as we read in detail in Vegetius’ Epitome of Military
Science (c. A.D. 400).20 In fact, the 3,000-men strong Roman legion was divided
– after the early third century BC – into three sections: the 1,200-strong hastati
that formed the irst line, the 1,200-strong second line called principes and the
600-strong third line of the triarii. Two classic examples from the second half
of the eleventh century regarding the use of the three-line battle-formation are
Hastings (1066) and Dyrrachium (1081). In the irst case the Normans formed
three divisions (left-lank, right-lank and center) each of which was divided
into lines of cavalry, infantry and skirmishers, while the same formation was
probably applied to Dyrrachium as well – even though we are not entirely sure
if Guiscard had any archers with him.21 And we can ind other examples of
Crusader armies using a three-line battle formation, like the cases of Sarmin
(1115), “Ager Sanguinis” (June 1119) and Hab (August 1119).22 But the three-
line formation that we have seen so far was speciically a battle formation.
Thus, our main question remains: from where were the Crusaders inluenced in
adopting a square (3x3) marching formation? In the following part of this study,
I will look into the Byzantine military manuals of the tenth century in search of
clues regarding any links between the Latin “ighting march” and the Byzantine
order of march of the post-965 period. But before going into the examination
of these manuals, I will begin with an introduction to their role, inluence and
background.
The Byzantine military manuals formed the “legacy” of experienced and
glorious generals in the warfare in the East. The Praecepta Militaria of Nice-
phoros Phocas was composed during the latter’s reign as emperor and probably
in the year 965,23 representing the emperor’s view of the ideal battle tactics and
formations of the units of the Imperial Army in its struggles against the Muslims
in the third quarter of the tenth century. Phocas, however, almost certainly relied
in two other contemporary texts (c. 950), whose ideas he did not simply copy
but reined and expanded with a critical eye.24 Other works for this period

19 I use Delbrück’s term of maniples, as the author explains the signiicance of not having a straight
and uninterrupted line in a phalanx formation (comparing the examples of the Macedonian and
Roman phalanxes), but rather a line based on maniples which formed small intervals at regular
points (the maniples were 20 men wide) that were illed just before the enemy approach. See:
Delbrück, History of the Art of War, 1: 272–77.
20 Vegetius: Epitome of Military Science, trans. Norman P. Milner (Liverpool, 2001) II. 15,
pp. 47–49; see also Sextus Julius Frontinus’ Strategemata (irst century A.D.): Frontinus,
Stratagems, Aqueducts of Rome, trans. Charles E. Bennett, Loeb Classical Library (London,
1925) II. iii. 16.
21 Theotokis, “The Campaigns of the Norman Dukes”, p. 219.
22 Verbruggen, The Art of Warfare, pp. 208–10.
23 McGeer, Sowing the Dragon’s Teeth, p. 180.
24 Sylloge tacticorum, quae olim Inedita Leonis Tactica dicebatur, ed. André Dain (Paris, 1938);

Journal of Medieval Military History 11.indb 62 18/03/2013 15:31:47


The “Fighting March” of the Crusaders at Ascalon (1099) 63

include the treatise On Skirmishing, attributed probably to Nicephoros Phocas’


elder brother Leo – the strategos of the Cappadocian theme and later Domestic
of the armies of the West – and written in or shortly after A.D. 969. Its context
focuses on skirmishing and hit-and-run tactics applied by the Imperial forces in
the eastern frontiers of the empire, thus relecting the kind of small-scale and
defensive warfare of the pre-963 period. The Taktika of Nicephoros Ouranos,
composed by a senior general of Basil II “the Bulgarslayer” in the East and
the Balkans during his governorship of Antioch between the years 999 and
1007,25 and the treatise On Tactics which was written most likely by the same
oficer sometime between the years 991 and 995 and is set in the operational
theater of the Balkans complete our analyses. All of these texts and handbooks
of the period set rather strict guidelines on the formations and operational role
of different units of the Imperial Army, but they also allowed the commander a
great deal of discretion in the ield. They relect the practice of older and well
established strategies and tactics, along with a number of innovative ideas put
into practice, and the task of the historian is to distinguish between the two.
As the author of On Skirmishing notes: “We have acquired this knowledge not
simply from hearing about it [from the old military manuals] but also from
having been taught by a certain amount of experience.”26
To return to the theory regarding the link between the Crusaders’ “ighting
march” and the Byzantine marching and battle formations of the tenth and elev-
enth centuries, the Taktika of Nicephoros Ouranos presents a detailed account
of the standard Byzantine order of march, based on the mutual dependence of
infantry and cavalry units. We read in the Taktika:
The cavalry units should proceed in order, some to the right side, some to the left,
some ahead, and some behind as rearguard. The infantry units should proceed in good
order on the inside of the cavalry units, three chiliarchs with their contingents on the
right side, three on the left, three in front, and three behind in the rear. You [general]
must proceed in the middle so that you can offer support to whichever side the enemy
attacks.27

However, this particular account on the early eleventh-century order of


march cannot fully explain the three-by-three square formation of the Crusaders
at Ascalon. If the latter had been inluenced by Nicephoros’ Taktika then they

L’histoire du texte d’Elien le Tacticien, ed. André Dain (Paris, 1946), pp. 156–57; See also:
E. McGeer, “The Syntaxis Armatorum Quadrata: A Tenth-Century Tactical Blueprint,” Revue
des études byzantines 50 (1992), 219–29, for a detailed commentary on the work and possible
links of the Syntaxis and the Sylloge with Aelian’s Tactica, Nicephoros’ Praecepta and Ouranos’
Taktika.
25 The Taktika of Nicephoros Ouranos [chapters 56 through 65], in Sowing the Dragon’s Teeth, pp.
88–163; Jean-Adré de Foucault, “Douze chapitres inédits de la tactique de Nicephore Ouranos,”
Travaux et Memoires, 5 (1973) 281–312; Eric McGeer, “Tradition and Reality in the ‘Taktika’
of Nikephoros Ouranos,” Dumbarton Oaks Papers 45 (1991), 129–140.
26 “On Skirmishing”, 13–15, p. 146.
27 “The Taktika of Nicephoros Ouranos”, 64. 30–39, p. 148; compare with: “On Tactics”, 10.
1–73, pp. 278–82.

Journal of Medieval Military History 11.indb 63 18/03/2013 15:31:47


64 Georgios Theotokis

would have deployed their cavalry units as vanguard, rearguard and lank units
with some 300 knights on each side of the formation, while 12,000 footsoldiers
would have been marching on the inside of the aforementioned cross-shaped
formation – three thousand3,000 on each side. Also, the 300 knights would most
likely have been divided into six (3x2) units of 50 men each,28 accompanied by
the 3,000 footsoldiers divided into three taxiarchies (drungi) of 1,000 men each.
No square (3x3) formation is formed at any time, so I believe we must seek our
answer elsewhere.
A very tempting theory is that the Latin “ighting march” could have been
inluenced by the basic ground plan for a Byzantine camp of the late tenth
century, as described once more by the contemporary military manuals. The
historical sources contain several detailed descriptions of camps which relect
their structure and use in periods of warfare. According to the evidence from a
great number of these manuals from as early as the sixth century,29 the preferred
Byzantine camp’s ground plan was a square or rectangle with two entrances on
each side of the camp and four major roads dividing the camp into nine sectors;
other versions note three entrances on the east and west side and two on the
north and south with ive roads dividing the camp into twelve sectors. Square
camp formations were encouraged by the tacticians in that they provided more
security and they managed to counter the enemy’s numbers, since if the latter
chose to besiege all four sides its force would be dispersed.30 A great change in
the early Byzantine period came with the introduction of the centrally crossed
roads which replaced the old Roman T-shaped intersection,31 and their number
depended on how many divisions the general had.
Now, if we shift our attention from the camp’s ground plan to the Byzantine
infantry formations of this period, we will be able to detect many similarities
between them. Writing about the ideal formation of the infantry when deployed
with units of cavalry in the ield, Nicephoros Phocas writes in his Praecepta:
The formation of the infantrymen under discussion is to be a double-ribbed square,
thus called “a four-sided formation” by the ancients, which has three units on each side
so that all together there are twelve units on the four sides. In case the cavalry force

28 The 50-man cavalry unit as part of a 200-man bandum was introduced by Leo VI and it is
mentioned in “Praecepta Militaria”, IV. 1–2, p. 38.
29 Vegetius, Epitome of Military Science, III. 8, pp. 79–83; The Strategikon attributed to Emperor
Maurice (582–602): Maurice’s Strategikon: Handbook of Byzantine Military Strategy, trans.
George T. Dennis (Philadelphia, 1984), B. 22, pp. 158–62; the anonymous treatise On Strategy
written during the reign of Justinian (527–65): On Strategy, in Three Byzantine Military
Treatises, 28. 1–44, pp. 86–8; “On Tactics”, 1. 1–190, pp. 246–54.
30 “On Tactics”, 1. 20–5, p. 246.
31 The structure of a Roman camp was dominated by the via principalis, which linked the porta
principalis sinistra (or porta praetoria) with the porta principalis dextra – in the middle of
which was the general’s tent –, and the via praetoria which ran vertical to the via principalis.
For this observation see the comments and the extensive bibliography provided by: McGeer,
Sowing the Dragon’s Teeth, p. 348, n. 31.

Journal of Medieval Military History 11.indb 64 18/03/2013 15:31:47


The “Fighting March” of the Crusaders at Ascalon (1099) 65

is quite large and the enemy does not bring along a similar number of infantry, twelve
intervals should be left open.

And the more important for our case is what immediately follows: “If, on
the other hand, the cavalry force is not large and the enemy does bring infantry
along, eight intervals should be left open.32
The similarities between the Byzantine plan of a camp and the typical infantry
formation are more than obvious. In both cases we have a square formation
divided by eight or more crossroads into a number of sectors depending on the
nature of the terrain, the enemy’s forces and, more importantly, the number of
infantry and cavalry units. As Nicephoros suggests, when a Byzantine general
was short of cavalry units he could adjust his infantry formation to only eight
intervals, meaning nine sections in all. This, of course, its very well with the
low number of knights the Crusaders had brought to Ascalon, in comparison
with the 9,000 footsoldiers. But a fundamental question is how we can link the
ground plan for a tenth-century camp to the marching formation of the Byzan-
tine infantry. Once more, the answer lies with the writings of Nicephoros, in the
chapter where he examines the encampment of the army:
They [soldiers] must keep their places in the camp exactly as they set to deploy in
battle formation, so that, in the event of a sudden report of the enemy, they will be
found ready as though in battle formation.… Eight intervals must be left open in the
army’s encampment so that three chiliarchs have two intervals. These must be in the
shape of a cross on the four sides of the encampment … two roads from the east to
west and two from north to south.33

The conclusion that we can draw from this paragraph is that the square in
which the Byzantine infantry deployed for battle was patterned after the standard
ground plan for encampments. During the decades-long struggle in Cilicia and
Syria, from the 930s under the celebrated general John Curcuas,34 the Byzantines
would have faced sudden attacks by the Muslims and would have needed to stay
on high alert while camping on hostile ground. Attacks when least expected, and
especially during the night,35 were not uncommon in Byzantium’s wars in Syria,
thus what Nicephoros writes about the event of a sudden report of the enemy
and the need for a rapid mobilization of the army’s units.
The inal issue has to do with the link between the tenth-century Byzantine
army formations and the Crusader army of the late eleventh century. The trans-

32 “Praecepta Militaria”, I. 39–51, p. 14.


33 “Praecepta Militaria”, V. 23–6, p. 52.
34 For Curcuas’ achievements in the East, see: George Ostrogorsky, History of the Byzantine State
(Oxford, 1989), pp. 276–77.
35 The author of On Tactics devotes two chapters to night attacks against enemy camps. See:
“On Tactics,” 12, 25, pp. 282–83, 312–14; for what the early Abbasid military manual of the
tactician Al-Harthami has to say about night attacks and in what degree it encourages such
attacks, see: Umar Ibn Ibrahim al-Awsi al Ansari, A Muslim Manual of War: Being Tafrij al
Kurub i Tadbir al Hurub, trans. and ed. George T. Scanlon (Cairo, 1961), p. 93.

Journal of Medieval Military History 11.indb 65 18/03/2013 15:31:47


66 Georgios Theotokis

mission of knowledge in military affairs from the East to the West can be attrib-
uted to the large numbers of mercenaries from countries like France, Germany
and England that were employed by the Byzantine emperors from the 1030s.36
Three hundred Normans from southern Italy took part in the empire’s expedition
against the Sicilian Arabs in 1038, being transported by units of the Imperial
Navy and marching with elite units of the Byzantine Army for three years,37
with a signiicant number of them and their descendants being identiied by
Anna Comnena as late as 1078.38 Individual Frankish mercenaries are recorded
in the Byzantine payrolls as early as 1047,39 facing a variety of enemies like
the steppe tribes of the Patzinaks and the Cumans in the Balkans and, since
the 1050s, the Seljuks in Armenia and the Taurus mountains.40 Famous oficers
include the names of Hérve, Robert Crispin and Roussel of Bailleuil; all of them
being active mostly in north-eastern Asia Minor between the 1050s and 1080s.41
Although the length of this paper does not allow me to go into more detail, there
are several examples from this period where contemporary chroniclers mention
Frankish regiments being deployed alongside units of the Byzantine army and
the famous Varangian Guard. In addition, the German regiment of the Nemitzoi
was present at Matzikert (1071),42 Anglo-Saxon mercenaries gradually replaced
the Scandinavian element in the Varangian Guard in the 1080s–90s43 while in

36 Smail seriously underestimates the signiicance of the presence of thousands of Western troops
in Imperial service from the second quarter of the eleventh century. Even more surprisingly,
when he mentions some of the battles where the Frankish troops were deployed alongside
the Imperial units, he only mentions the years 1138, 1164 and 1169. See: Smail, Crusading
Warfare, pp. 122–23.
37 Michele Amari, Storia dei Musulmani di Sicilia, 3 vols. (Catania, 1935), 3: 438–55; Jean Gay,
L’Italie méridionale et l’Empire Byzantin depuis l’avènement de Basile Ier jusqu’à la prise de
Bari par les Normands (867–1071) (Paris, 1904), pp. 450–54; Graham Loud, The Age of Robert
Guiscard: Southern Italy and the Norman Conquest (London, 2000), pp. 78–80; F. Chalandon,
Histoire de la domination normande en Italie et en Sicile, 2 vols. (Paris, 1907), 1: 88–96.
38 They were mentioned as “Maniakatoi” from the name of the general of the Sicilian expedition,
George Maniaces. Alexiad, I. iv-vi, pp. 39–48.
39 Leo Tornikius’ rebellion: Ioannes Skylitzes, Synopsis historiarum, Corpus Fontium Historiae
Byzantinae, ed. I. Thurn (Berlin, 1973), vol. I, p. 439–40; Anonymus Barenses, Chronicon,
Rerum Italicarum scriptores: raccolta degli storici Italiani dal cinquecento al millecinquecento,
ordinata da L.A. Muratori, Lapi, Città di Castello, 1900–, v. 151, s.a. 1047; The History of
Psellus, ed. with critical notes by Constantine Sathas (London, 1899), pp. 140–41.
40 Although they were mostly used to man towns and castles because of the steppe tactics of the
Patzinaks in the Balkans and the Seljuks in the East.
41 For the career of these three Franks, the most basic works are: Jonathon Shepard, “The Uses
of the Franks in Eleventh-Century Byzantium,” Anglo-Norman Studies 15 (1993), 275–305;
Alicia J. Simpson, “Three Sources of Military Unrest in Eleventh-Century Asia Minor: The
Norman Chieftains Hérve Frankopoulos, Robert Crispin and Roussel of Bailleul,” Mesogeios/
Mediterranée 9–10 (2000), 181–207.
42 For this German unit: John Haldon, The Byzantine Wars: Battles and Campaigns of the
Byzantine Era (Stroud, 2000), p. 114.
43 Alexander A. Vasiliev, “The Opening Stages of the Anglo-Saxon Immigration to Byzantium in
the Eleventh Century,” Annales de l’Institute Kondakov 9 (1937), 39–70; Jonathon Shepard,
“The English and Byzantium: A Study of Their Role in the Byzantine Army in the Later

Journal of Medieval Military History 11.indb 66 18/03/2013 15:31:47


The “Fighting March” of the Crusaders at Ascalon (1099) 67

the spring of 1090, Robert I of Flanders sent ive hundred Flemish knights to
Alexius Comnenus after being asked for military assistance the year before.44
Next year, in 1091, we ind Frankish troops ighting alongside the Imperial Army
and the Cumans against the steppe tribe of the Patzinaks at Mount Levounion.45
Also, the Norman army consisted of experienced knights who had faced Byzan-
tine troops in battle – from the Italian “expansion” (1041–71) to Dyrrachium
(1081–84) – while Bohemond had fought against Turcopoles during his irst
campaign against Dyrrachium ifteen years previously.46 Further, if we are to
believe the author of the Historia Belli Sacri, both Tancred and Richard of
Principate and possibly a number of their followers knew Arabic,47 something
rare for the armies of the First Crusade. Also, it is highly likely that Bohemond
himself spoke Greek. And last, our most important clue of this theory, we read
in the Alexiad of Anna Comnena that the Crusader leaders, while staying in
the capital before proceeding to besiege Nicaea in the spring of 1097, “were
instructed in the methods normally used by the Turks in battle; told how they
should draw up their battle-line, how to lay ambushes; advised not to pursue too
far when the enemy ran away in light.”48
If we examine the pitched battles between the Muslims and the Latin armies
of the irst generation of the Crusaders, we can identify two possible cases
where the Latins used the “square ighting march.” At the battle of Hab,49 on
14 August 1119, Latin forces clashed with the Seljuk army of the Emirs Toktagin
of Damascus and Il-Ghazi of Mardin. Certain that a battle would take place on
the morning of 14 August , the Crusaders formed their ranks in a three-by-three
line formation,50 with each division numbering about eighty to ninety knights

Eleventh Century,” Traditio 29 (1973), 53–92; John Godfrey, “The Defeated Anglo-Saxons
Take Service with the Eastern Emperor,” Anglo-Norman Studies 1 (1978), 68–70.
44 France, Victory in the East, p. 81.
45 These Frankish troops were under the command of Constantine Humbertopoulos, a trusted
Imperial oficer who was a nephew of Robert Guiscard and had fought against his uncle at
Dyrrachium in 1081. See: Alexiad, VIII. v, p. 257.
46 Georgios Theotokis, “The Raids of the Norman Dukes of Southern Italy”, pp. 217, 247, 277ff;
idem, “The Norman Invasion of Sicily (1061–1072): Numbers and Military Tactics,” War in
History 17 (2010), 1–22.
47 Historia Belli Sacri, R.H.C. Oc., vol. 67, p. 198; for the spread of the Arabic culture and
language in Sicily up to the 1060s, see: A. Metcalfe, “The Muslims of Sicily under Christian
Rule,” The Society of Norman Italy, ed. Graham A. Loud and Alexander Metcalfe (Leiden,
2002), pp. 289–317, especially pp. 289–93.
48 Alexiad, X. xi, vol. II, pp. 67–8; Sewter, p. 329.
49 The most detailed of our primary sources about the battle and the Latin battle formations
is: Walter the Chancellor, The Antiochene Wars, trans. Thomas S. Asbridge and Susan B.
Edgington (Aldershot, 1999), II. 12, pp. 152–56; Other sources include: Fulcherius Carnotensis,
Historia Hierosolymitana, Book III, IV. 1–4, pp. 624–29; William of Tyre, A History of the
Deeds Done Beyond the Sea, vol. I, XII.12, pp. 532–35; the main secondary works are: Oman,
The Art of War, 1: 297–300; Smail, Crusading Warfare, pp. 180–81; Delbrück, History of the
Art of War, 3: 409; Verbruggen, The Art of Warfare, pp. 209–11.
50 Delbrück disputes Heerman’s description of linear formations: Delbrück, History of the Art of

Journal of Medieval Military History 11.indb 67 18/03/2013 15:31:47


68 Georgios Theotokis

and all the infantry placed in the front rank marching behind the cavalry. We
read in Walter the Chancellor’s account:
Nine battle-lines were drawn up … [they] set out enthusiastically on the march,
proceeding in this order: three lines [of cavalry] were placed in the vanguard and the
infantry was positioned to the rear, so that they could protect them and be protected
by them; the royal force [of Baldwin II], prepared for the protection of the former and
the latter; the line of the Count of Tripoli was placed on the right, with the barons’ [of
Antioch] lines on the left.

Verbruggen argues that there was no rear-guard and that the lank units of Robert
Fulcoy on the left and Pons of Tripoli on the right constituted the second line
which extended the front of the Latin army. Oman and Smail, however, believe
that there was another body of Antiochene barons that was put behind Bald-
win’s divisions, thus forming the third line of the Latin army. Whatever the case
though, this was certainly not the square formation that was seen at Ascalon
exactly twenty years before. It resembles more the cavalry formation dictated by
Leo VI in his Tactica of two lines of three and four divisions respectively, which
were supported by two units of νωτοφύλακες (lank-guards) that were posted on
the sides but slightly behind the second line of cavalry.51
The armies of Baldwin II and Toktagin of Damascus met for a second time
some seven years later, not far from Damascus at a place called Marj-es-Safar
on 25 January 1126. Our main Latin source, Fulcher of Chartres, is vague in his
narrative and it all depends on how we translate his comments: “Soon after, our
horse and foot were organized in twelve divisions, in such a way that each could
support the other, if neccessary.”52 The Latin alteruter means either or one of
two and Fulcher is most likely referring to the infantry supporting the cavalry in
each division and not each of the twelve divisions supporting each other. Thus,
is it possible that Baldwin put his army in the ield forming a rectangle which
was three divisions long and four deep, or the other way around? Judging by
the large number of infantry that the Turks had raised, although no estimates
are given by either Fulcher or Ibn-al-Qalanisi we understand that the atabeg of
Damascus had brought with him “a great host from the armed bands and untried
youths of Damascus”,53 and by the topography of the area close to the city of
Damascus which would certainly not have allowed the Latin army much room
to maneuver, that the most likely formation would have been the three-long and
four-deep rectangle. As we have already seen, Phocas presents this formation as

War, 3: 409; Otto G. L. Heerman, Die Gefechtsfuhrung abenlandischer Heere im Orient in der
Epoche des ersten Kreuzzugs (Marburg, 1887), pp. 89–95.
51 Leo VI, Tactica, Patrologia Cursus Completus – Series Graeca, vol. 107, ed. J. P. Migne (Paris,
1863), constitution XII, cols. 1–30.
52 “Ordinatae sunt tam militum quam peditum acies duodecim, ut ab alterutra corroboretur
caterva, si necessitas admoneret.” Fulcherius Carnotensis, Historia Hierosolymitana, Book III,
L. 7, p. 789.
53 Ibn-al-Qalanisi, The Damascus Chronicle of the Crusades, extracted and translated by Hamilton
A. R. Gibb (London, 1932), p. 175.

Journal of Medieval Military History 11.indb 68 18/03/2013 15:31:47


The “Fighting March” of the Crusaders at Ascalon (1099) 69

a different version of the square ighting march for narrow battle-ground and in
the case when the enemy brings along a similar number of infantry.
What we have concluded so far is that, although it was beyond any doubt
an innovative idea on the part of the Crusaders, we can trace the inluence for
the marching formation of the Latin army before the battle of Ascalon to the
Byzantine experience in warfare gained during the Byzantine expansionist wars
in Syria in the mid tenth century. Several tactical innovations can be identiied
in the structure and performance of the Byzantine army itself on the battleields
of this period, as relected in the writings of contemporary military tacticians.
One of them was the formation of the infantry, which was patterned after the
standard ground plan of a military camp, a square divided by four crossroads
into nine sectors. Although we have to doubt as to the direct inluence that these
Byzantine manuals had on the Latins, the latter were certainly “exposed” to the
Byzantine military practices and battle tactics. And this knowledge in military
affairs and, most importantly, the experience in ighting an enemy that had a
number of common characteristics with the Seljuks, were transmitted to the
Latin armies of the First Crusade by the thousands of mercenaries who were
employed by the Byzantine emperors from the 1030s.
It seems strange that the Crusaders did not use this marching formation on
other occasions, like in the catastrophic defeat at Harran in 1104,54 but we have
to bear in mind that other factors played a signiicant role as well. I will base
my analysis of the suitability of the “square ighting march” in the Middle East
on three major points: the topography of the region, the leadership of the Latin
armies and the degree of discipline and experience of the troops, especially
the infantry. Much attention is paid to the topography of the battle-ield by
Nicephoros Phocas who, as I have mentioned, proposes different versions of
his standard infantry formation for both lat and narrow terrain.55 The “square
ighting march” was ideal for open and relatively lat terrain and not suitable
at all for mountainous or hilly areas. Considering that the key features of the
topography of Syria and Palestine are a narrow coastal strip, broader in the south
than in the north, a belt of mountains running parallel to the coastal roads and
small inland plains and plateaux that stretch east to the desert, this particular
battle formation applied by the Crusaders en route to Ascalon – one of the few
areas that favor marching formations with a relatively broad front – was far
from ideal for the Levant.56
Since discipline and morale in an army largely depend on its command and the
abilities of its leadership, it is easy to understand the direct correlation between
the leadership in a Latin army and its performance in a battleield. It is true that

54 Oman, The Art of War, 1: 321–24.


55 “Praecepta Militaria”, II. 151–91, pp. 30–3.
56 For the topography of the Middle East: Naval Intelligence Division, Syria, Geographical
Handbook Series (London, 1943); Mark Whittow, “The Political Geography of the Byzantine
World – Geographical Survey,” in The Oxford Handbook of Byzantine Studies, ed. Elizabeth
Jeffreys, John Haldon and Richard Cormack (Oxford, 2008), pp. 219–31.

Journal of Medieval Military History 11.indb 69 18/03/2013 15:31:47


70 Georgios Theotokis

there were signiicant periods in the twelfth century when the Latins were not
under a uniied command. Two characteristic examples are those of the Second
and Third Crusade with the absence of a dominant igure between the various
kings exacerbating the lack of an ambitious strategy. In addition, political strife
in the kingdom of Jerusalem could undermine the unity of the high command,
an example being the battle of Hattin (1187) that followed the “regency crisis”
in the Kingdom between Guy de Lusignan and Raymond of Tripoli in 1186–87.
If we consider the twelfth century as a whole, however, it was typical on the part
of the Crusaders to submit to a single leadership under the kings of Jerusalem.
In the case of the period before the battle of Ascalon, Bohemond had emerged as
the charismatic leader of the First Crusade that led the Latin army at Doryleum,
Harem and Antioch. Bohemond of Taranto was an experienced knight in the
ighting in the East, having faced Byzantine and Turkish troops in his campaigns
in Byzantine Illyria in 1081–84, and he showed great commanding abilities and
strategic thinking in organizing an ambush against the Seljuks of Aleppo in
1098 or proposing the battle plan of the Crusader army at Antioch in the same
year, always keeping a tactical reserve with him and using the topography of
the battleield to cover the lanks of his army.
But for troops to perform a march through enemy territory, keeping a square
formation of nine units of mixed infantry and cavalry also required a great
degree of experience and training in the ield, mostly on the part of the footsol-
diers who were less maneuverable than the horsemen. I have mentioned the fact
that the majority of the Latin armies of the First Crusade were ignorant of the
ighting tactics of the Seljuks, but their bloody experience at Doryleum and their
march through Asia Minor certainly proved extremely useful. Thus, by the early
summer of 1098, the Latin East would experience for the irst time the mixed
units of infantry and cavalry where the foot-soldiers acquired a fundamental role
in Middle East warfare. However, we should underline the fact that by this stage
of the Crusade the infantry would have evolved into a quite formidable ighting
unit – with better armor protection as well – which was most needed in the East.
This comment does not suggest that the footsoldiers of the First Crusade were
a mere rabble of untrained men when they crossed to Asia Minor in 1097, but
that it took several months of intense interaction with the Turks to develop into
a cohesive and disciplined unit.57
A inal key question that has to be analyzed, based on the aforementioned
conclusions about the suitability of the “square ighting march” in the Middle
East, is why the Crusaders did not adopt this marching formation sooner, from
the point they marched southwards from Nicaea to the Anatolian plateau? In
addition to those already highlighted, several other factors need to be consid-

57 France, Victory in the East, pp. 294–96; idem, “Technology and the Success of the First
Crusade,” in War and Society in the Eastern Mediterranean, ed. Yacov Lev, (Leiden, 1997),
pp. 174–75; Smail believes that the idea of a developing infantry force is unsatisfactory and
that the footsoldiers at Doryleum were a force it for war, although he bases his argument solely
on the Anonymous’ account: Smail, Crusading Warfare, pp. 115–20.

Journal of Medieval Military History 11.indb 70 18/03/2013 15:31:47


The “Fighting March” of the Crusaders at Ascalon (1099) 71

ered: the topography of the regions from which the Crusaders marched through
in the summer of 1097 clearly made a three-by-three square – or even three-by-
four rectangle – marching formation almost impossible. We have to add to this
the sheer numbers involved in the march to Doryleum – some 50,000 troops
divided in two large groups (vanguard and rearguard of about 20,000 and 30,000
men respectively)58 – in contrast with the 1,200 knights and 9,000 footsoldiers
marching south to Ascalon. Further, the Crusaders in 1099 had to march some
forty kilometers to reach the Egyptian camp after having received intelligence
about the exact whereabouts of their enemy. In striking contrast, the huge Latin
host marching south from Nicaea had to cover a distance that took them two
days (28–30 June) at a daily rate of no more than twenty-ive to thirty kilom-
eters.59 Even by the standards of a modern army it would have been extremely
dificult to have several units of hundreds or even thousands of soldiers each
marching through rough terrain for two days in close marching formation with
tanks and mobile infantry units. If we add to this the large numbers of civil-
ians and the baggage train that certainly would have accompanied each contin-
gent of the Latin Army,60 the lack of intelligence regarding the presence of an
enemy army – at least before the evening of 30 June – and the lack of any
overall commander as we have already seen, we can understand why the “square
ighting march” could not have been used in these early stages of the Crusade.

58 France, Victory in the East, p. 170.


59 France, Victory in the East, pp. 170–71.
60 For the civilian casualties, see: France, Victory in the East, p. 170, n. 82; p. 181, n. 104.

Journal of Medieval Military History 11.indb 71 18/03/2013 15:31:47


Journal of Medieval Military History 11.indb 72 18/03/2013 15:31:47

You might also like