Centuries
Late Roman order of battle
Until the later fourth century there is a consensus that Roman armies still consisted predominantly of infantry units, with cavalry employed chiefly in the role of scouts, flank and rear guard, to attempt, or threaten, the enemy flank, and to exploit and then follow up any weakness or retreat of units in the enemy line. Roman tactics revolved around the heavy infantry, who formed both the main battle line, and whose auxiliaries—slingers, archers, javelin-men— functioned as light-armed troops. Cavalry remained ancillary to these functions. Even after the introduction of more, and more heavily armed, mounted units from the later third century, this basic pattern hardly changes.
The relative increase in the importance of cavalry seems in fact to have been fairly late, contrary to assumptions usually made. Most of the limited evidence for the later fourth and fifth centuries shows that the proportion of cavalry to infantry did not increase dramatically as a result of the defeat at Adrianople, but on the contrary, remained much the same—at Strasbourg, Julian had some 3,000 cavalry and 10,000 infantry, and over a century later in 478 a field army in the east is reported to have consisted of 30,000 infantry and 8,000 cavalry. The proportion of mounted to foot soldiers is certainly greater than in a legionary army of the first century, but cavalry are still by no means either the dominant or the key element in late Roman armies of this period.2 Adrianople is usually named as the defining moment at which heavy cavalry decisively proved their superiority over infantry and ushered in the new age of mounted warfare.3 Yet the Gothic victory resulted not from some dramatic superiority of cavalry over infantry, but rather from the simple facts that the emperor Valens received incorrect information about the numbers of the Gothic force, and that a substantial detachment of Gothic horsemen joined the battle after the main Roman line was committed, were able to take it by surprise in the flank and roll up the whole formation upon itself. This is not to minimize the impact of the defeat (which was clearly expressed by Ammianus); nor is it to ignore both the gradual adoption of a heavier panoply by Roman cavalry units, as well as the creation of some new heavy cavalry units, during the later third, fourth and fifth centuries.
But there is plenty of evidence throughout this period and well after Adrianople that Roman infantry were able to hold off and defeat barbarian cavalry, and that the proportion of cavalry to infantry units remained approximately the same for the next century: roughly 1:3 in numbers of units, but far fewer in absolute numbers of men, since the unit sizes in the cavalry were smaller.4 Within the Roman cavalry there was an increase from the late third to the early fifth century in the numbers of very heavily armoured units, and it has been calculated that by the time of the Notitia Dignitatum, a late fourth/ early fifth century document giving the order of battle of the eastern and western armies, heavy armoured cavalry (cataphracti and clibanarii) made up some 15 per cent of the comitatenses cavalry, in comparison with lancers and other heavy cavalry (61 per cent) and light cavalry (24 per cent). Yet there were in turn more of these cavalry units assigned to the limitanei than to the field armies, which suggests strongly that cavalry were still regarded throughout the fourth and fifth centuries as most valuable in scouting and patrolling, or covering the wings and flanks of a mainly infantry army.5
But Roman infantry no longer differed substantively in their arms and armour from their barbarian counterparts. During the third century at the latest the adoption of a somewhat lighter panoply for regular infantry units (in comparison with the classic legionary equipment) seems to have been completed, corresponding to a change in infantry training and tactics. The combat engineering skills of the heavy infantry of the first centuries BC and AD are concentrated in a few specialist units, for example, while the emphasis moves away from the highly trained individual, fighting within a distinctive tactical sub-unit, to the infantryman as one of a mass, whose effectiveness depended not on individual skills so much as on unit coherence. The change corresponds in the archaeological picture to the adoption of the Germanic spatha and the greater diversity of weaponry within each tactical unit, as described, in fact, by Vegetius. The chief qualities which now distinguished Roman infantry from their foes were tactical discipline and training in close-order drill and battlefield manoeuvring, together with the heavier personal armour—mostly mail—of those soldiers selected to serve in the first ranks of the battle line. This certainly gave Roman infantry a continued advantage in most contexts over their barbarian enemies in the European and Balkan theatres. And there seems little doubt that the increasing significance of cavalry during the second half of the third century parallels and is causally related to these changes.6
The development of heavily armoured cavalry was a response to the use of similar mounted shock troops in the east, especially following the defeats suffered at the hands of Sassanid armies from the middle of the third century. But such armoured mounted units were not intended to replace, but rather to supplement the usual cavalry formations, generally on the flanks, and to stiffen the main battle line of the Roman army, still composed of disciplined and well-trained infantry, and thus served to neutralize the equivalent forces on the enemy side. They certainly neither replaced the basic infantry order of battle, nor did they function as the main battle arm. Indeed, the mass of both the Parthian and Sassanid armies were light horse archers or infantry, the heavy cavalry representing a noble elite which was relatively limited in numbers. And even against the Sassanid heavy cavalry (and once the lessons of the third-century defeats had been learned), disciplined Roman infantry formations, correctly handled, held their own, and could on occasion move out against cavalry which had not charged in order to reduce the effects of the enemy archery.7
The sixth century
That Roman battlefield tactics continued well into the sixth century to regard well-trained and disciplined infantry as essential—whether or not they were the majority of the troops involved—is clear from several sources. First, there was great concern voiced by a number of authorities precisely on the question of discipline in the ranks—this is a major issue raised in the introduction to the late sixth-century Strategikon; it recurs in comments of the slightly earlier chroniclers Menander Protector and John of Ephesus. Indiscipline seems to have become increasingly common from the 530s onwards, a reflection of irregular payments as well as alienation between officers and men. The sources provide many instances.8 We read frequently of officers reintroducing strict order and discipline; and Roman commanders themselves referred to Roman discipline as the quality that set them apart from and gave them an advantage over their adversaries. Both Belisarius and Narses, in speeches attributed to them, referred to the traditional high standards of Roman discipline and efficiency.9 Attention to drill and manoeuvres is also referred to in the narrative sources, whether on the eastern or western fronts, although a distinction between infantry and cavalry is rarely drawn.10
Second, the few detailed descriptions we have of major battles show infantry continuing to play an important role. At the Roman defeat at Kallinikon in 531, where infantry drawn up on the wing formed a less numerous element than the cavalry in the battle line, they defeated a frontal heavy cavalry assault. At Taginae (Busta Gallorum) in 551/2 infantry formed the centre of the main Roman line, and Procopius emphasizes the discipline and order of the Roman troops, drawn up in their ranks and columns by squadron or regiment, both in this battle and in the battle of Mons Lactarius the following year. At the battle on the Casilinus river, Agathias describes the classical Roman battle line of the late Roman period, as discussed also in Vegetius: the infantry formed the main battle line, drawn up in serried ranks, the front rank consisting of the most heavily armed and armoured soldiers with the lighter troops behind them. The discipline and training of the Roman formation is illustrated by the fact that, while the Frankish charge succeeds in pushing through the line, the troops do not break but fall back around the gap, allowing the reserve infantry—made up of allied Heruli—to counter-charge and the line to reform, while at the same time, the Roman mounted units on the flanks, equipped with both spears and bows, are able to outflank the enemy wedge and break it up with concentrated archery. A similar description occurs for the Roman infantry line in an engagement during the eastern wars in 556/7 (well-armoured Roman troops advancing with linked shields).11
Discipline and order were key components of the Roman infantry formation frequently singled out by commentators. And while allowing for a degree of rhetoric and ideological bias, there are enough references to the contrast between Roman orderliness and the disorder of their foes (and sometimes their allies) to suggest that there was indeed a real difference. In the fighting against both Moors and Vandals in North Africa, against the Goths in Italy, as well as against the formidable Sassanid armies, contemporaries refer on several occasions to the discipline and order of the Roman forces, both infantry and cavalry, in contrast with their opponents. Theophylact Simocatta compares the Roman and Persian forces cooperating in the Persian civil war in 591: the Romans were disciplined, ordered, calm; they displayed “ordered cohesion”, and saved their foolhardy allies from a defeat when the latter, having been routed by the enemy, were compelled to withdraw behind the Roman line which protected them and threw the enemy back. Theophylact recounts a similar tale, in which the steady infantry line saves the retreating Roman cavalry during the wars against the Avars.12 Procopius records an incident from the Gothic war in which Belisarius is taken to task by members of his own retinue for not trusting his infantry. Although too few in numbers to be drawn up in the main battle line, it is nevertheless pointed out that their poor fighting record immediately beforehand was due to their bad officers who, as the only mounted soldiers in their formation, tended to run away before battle was joined, thus quite naturally totally demoralizing the men who broke easily when attacked.13 In terms of numbers, too, infantry remain a substantial element: in the invasion of Vandal Africa, for example, Belisarius’ army was made up of 10,000 foot and 5,000 horse, even though the infantry played for the most part a secondary role in the two main battles.
Although there is no persuasive indication that infantry had declined in importance by the end of the fifth century, this evidence does suggest that infantry discipline and order was a frequent cause of concern in the middle of the sixth century and afterwards. Belisarius was clearly sceptical of the steadiness of his infantry on two occasions in different theatres, suggesting that there was more than just a local issue here, while the speeches put into the mouths of both Belisarius and the Persian leader Firuz at Daras in 530 allude to the usual weakness and lack of order among the Roman infantry. In addition, there is no doubt that the emphasis does swing towards cavalry by the later sixth century, and there appear to be several reasons for this.
In the first place, while the empire had relied on allied mounted troops for a number of specialized roles—especially horse archers—during the fifth and first half of the sixth centuries, it is clear from Procopius’ account of the contemporary Roman cavalry soldier that Roman traditions and styles of mounted fighting were beginning to alter as a result of such contacts, and that the Romans were themselves training their mounted troops in such skills. That the numbers of such composite archer-lancer units was probably quite small is suggested by the fact that, while the advantage accruing to the Romans from their archery in respect of the Goths in Italy is emphasized, they do not seem to achieve parity with the Persians, even if Roman archery was, bow for bow, more effective than the Persian, as Procopius claims. Indeed, massed Persian archery remains a problem for Roman forces throughout.14
In the second place, the nature of much of the warfare of the middle years of the sixth century—the reign of Justinian—demanded armies that could move rapidly, confront equally mobile enemies, bring them to battle or harass them, and then move again to take up new dispositions elsewhere. This is especially true of the Italian and North African wars where, although infantry continued to play a key role—obviously in respect of garrison and related duties, but also in the line of battle—warfare assumed a guerrilla aspect to which rapidly moving cavalry were well suited. The accounts of battles in both Procopius and Agathias often give cavalry the main role (particularly in view of the versatility as both shock—and missile-troops ascribed to them by Procopius), and the reinforcements which arrive from time to time are frequently disproportionately of cavalry. This was also true of the later sixth-century warfare against the Avars in the Balkans, where again infantry continue to play a role, but where the account of Theophylact, the main narrator for these events, frequently implies that cavalry played the dominant role (as one might expect in a highly mobile war against a mounted nomadic people).15 It seems often to have been the case that either cavalry forces fought with minimal or no infantry support, or that infantry acted merely as a reserve and a protective wall for retreating cavalry. This was certainly so at Ad Decimum and Tricamerum (535) in Africa and at battles in the Gothic war already noted. It seems to have been even more the case by the 570s and afterwards: at Melitene in 575, at the battle on the Nymphios in 583, at Solachon in 586, on the Araxes in 589, as well as in a series of victories won by the general Priscus along the Danube in 600 and in Heraclius’ campaigns against the Persians in the years 622—6.16
In the third place, there is a good deal of evidence to suggest that the empire recruited substantial numbers of new cavalry units during the sixth century, thus altering the overall balance between the two arms. Evagrius reports, for example, that Tiberius Constantine recruited large numbers of cavalry from among various barbarian and indigenous peoples in 575/6 (“squadrons of excellent cavalry”), and it is probably at this point that the new formation of the Optimates (see above) was established. If so, this suggests a substantial increase in mounted units, and given the composition of the praesental armies which later made up most of the forces in the Opsikion region, cavalry now began to dominate the field armies of the empire: indeed, the latter division included also the Bucellarii, an elite brigade established also during the second half of the sixth century also consisting of mounted units.17 Apart from these (they may have numbered from 2,000 to 3,000), other cavalry units were raised at the same time, while already for Justinian’s reign there is good evidence for the establishment of new mounted units along traditional lines—five units of Vandal cavalry (presumably lancers) were established from among the prisoners taken during Belisarius’ African campaign; similarly, units of Persian and Armenian cavalry are found, as well as of Ostrogothic cavalry (posted to the eastern front). Other units, of heavy cavalry and clihanarii as well as of light cavalry/ archers, seem to have been formed during the fifth and early sixth century.18
The increased emphasis upon cavalry was, therefore, a fifth—and especially a sixth-century response to a change in the empire’s overall strategic situation, a response which may well have increased in pace from the middle years of the sixth century. Yet while the emphasis in the tactical formations and battlefield situations discussed in the Strategikon is placed upon the cavalry, the chapter dealing with combating barbarian and foreign peoples makes it quite clear that infantry continue to play an important role. The chapter on infantry formations, devoted specifically to this arm, begins by noting that infantry training and discipline have been greatly neglected in recent times, and that the purpose of the section is to redress the balance. The point is reinforced by the section on infantry tactics in the mid-sixth-century anonymous treatise on strategy, which merely paraphrases a number of ancient authorities and describes in effect a Macedonian phalanx rather than any formation employed in East Roman armies.19 Yet at the same time this treatise views cavalry primarily as a screen in front of the main infantry battle line, or as flank cover or pursuit troops.
This apparent emphasis on cavalry, which in fact takes the infantry more or less for granted, actually suggests an effort to ensure that Roman commanders take cavalry more seriously into account in the strategic context for which the Strategikon was written—against the Slavs, and more particularly the nomadic Avars. In the latter text, for example, while the emphasis on cavalry formations is clear, only one chapter (Chapter III) actually assumes that cavalry alone are involved. The other chapters deal with mixed formations, with sieges (in which infantry would normally play the central role) or with contexts in which cavalry would as a matter of course be employed against an enemy such as the Avars—surprise raids into enemy territory, ambushes and related undertakings. The author of the Strategikon also notes that mounted troops should anyway be dismounted and fight as infantry wherever the situation demands, as occurred at Solachon near Mardin in 586.20
Infantry were thus by no means an insignificant element in late Roman armies of the sixth and first half of the seventh centuries. In the wars with the Persians, whose own infantry were often quite numerous (if not particularly well-trained),21 as well as in the Balkans, infantry could not be ignored: in some situations they were essential. It was infantry units who carried out garrison duties, manning both major defensive installations as well as minor outposts and fortlets. In addition, much of the fighting in the densely wooded and hilly Balkan regions depended upon infantry, particularly when opposing the tactics of the various Slav tribes which the Strategikon describes. Ambushes were usually carried out and defiles had to be held or seized by troops on foot, and difficult tracks and pathways that could not be followed by mounted troops were accessible only to infantry. And where battles had to be fought in difficult terrain, it was advised that cavalry units be dismounted and drawn up in infantry formation to make the best use of the available resources.22 In a number of battles, infantry seem to have formed up in a solid line behind the cavalry units, acting as both a reserve and as a defensive wall behind which the cavalry could shelter. There are still examples of infantry formations forming up in solid ranks with linked shields to drive an already retreating force back,23 but the general tendency by the early seventh century would thus seem to be one in which infantry are increasingly passive and defensive, serving both as a reserve once the enemy has been repulsed or turned and a safe haven for defeated or withdrawing Roman cavalry units.
Infantry in the line of battle c. 640—900
Delbruck asserted uncompromisingly that a disciplined infantry was nonexistent in the Byzantine world.24 Yet neither the Persian nor the Arab wars were conducive to any sudden transfer of attention away from infantry towards cavalry. Infantry continued to play an important role in the battles fought against the early Islamic armies, which were themselves made up predominantly of infantry, troops whose use of camels and horses gave them greater mobility than their foes but who fought for the most part on foot.25 But thereafter the highly mobile nature of the warfare which dominated Byzantine-Arab relations gave the Arab mounted infantry an advantage over traditionally outfitted Roman infantry. Byzantine tactics and strategy had to respond to the nature of the threat from the raiders, and the fact that those contingents from the main field armies of the second half of the seventh and the eighth centuries were referred to as kaballarika themata—“cavalry armies”—illustrates the nature of the reaction. The empire certainly continued to employ infantry, and they continue to appear in the sources—when the types of troops involved are specified at all—in their traditional role. Infantry units played an important part in the campaign against the Bulgars in 678/9, for example, and in the guerrilla warfare along the eastern frontier in the later ninth and tenth centuries (and presumably before this) were a recognized element of the provincial armies in official and semi-official sources. But here, their role was confined primarily to guard-post duties, garrisoning forts and watchtowers situated at key points, or waiting in ambush for enemy forces shepherded along before the pursuing cavalry units.26
Leo’s Tactica makes it clear that infantry—both heavy and light troops— continued to function, and they are assigned a role in the mixed tactical formation he goes on to describe. Although derived from the Strategikon, the information has nevertheless been brought up to date to a degree.27 What Leo tells us, however, does not give us much information on the real degree of participation of infantry in the Byzantine armies of the period. Infantry certainly played a role in the wars against both Bulgars and Arabs during the campaigns of the eighth and ninth centuries, but the vast majority of descriptions of battles for the period c. 650—800 which appear in the narrative historical sources make virtually no mention of them. Perhaps more significantly, the few technical references to types of provincial soldiery which date to the period before the middle of the tenth century make no mention of them either. An account drawn from ninth-century information describing provincial field armies mustered to meet the emperor on their way to campaign in Syria assumes that they are composed of cavalry; official and semi-official regulations about the minimum property required for the maintenance of soldiers refer only to regular thematic cavalry or to sailors of the provincial fleets. Infantry are not classed, in this context, as soldiers at all.28
This absence is at first sight surprising, but reflects perhaps two features of the evolution of Byzantine tactics up to the tenth century. First was the nature of the thematic armies themselves, whose increasingly seasonal campaigning, localized recruitment and physical dispersal were not conducive, were indeed antithetical, to the maintenance of line infantry discipline and order. Infantry of this type were not suited to formal battlefield formations and manoeuvres, although garrison duties and irregular skirmishing warfare in broken country, lying in wait for hostile forces, would have been within their competence.
Secondly, when we have evidence, this is precisely what we find the thematic infantry doing. Although the numbers of the standing forces in each district are unknown—perhaps 4,000 or more in the larger themata as Leo’s Tactica prescribes—the occasional references to the epilektoi of each thema make it clear that these are usually cavalry.29 And while this limited material shows that much of the provincial cavalry suffered similar defects to the infantry in respect of organizational discipline, the nature of warfare—raids, harassment of invaders and so forth—inevitably gave them greater prominence, significance and importance. Yet the author of the treatise on skirmishing or guerrilla warfare notes on several occasions that the enemy cannot be defeated without an adequate force of Roman infantry to press home attacks on their encampments, occupy the defiles and ambush the withdrawing enemy columns, and so on.30 This reflects the fact that the attacking forces themselves were often composed of large numbers of infantry, which made their raids the more dangerous, since they were better able to pursue the rural population to their fastnesses, pillage and ravage their villages and homesteads, and resist Roman cavalry attacks on their encampments.31
Moreover, the nature of the warfare adopted deliberately by the East Roman government in the period from c. 640 until well into the eighth century— avoiding direct confrontation wherever possible—will not have promoted battlefield confidence, tactical cohesion and discipline, especially among infantry units, a tendency which will further have reduced their relevance and effectiveness in battlefield contexts. The treatise on skirmishing warfare, although written in the second half of the tenth century, reflects quite clearly the traditional form of warfare which had dominated the eastern frontier in the second half of the ninth century and up until shortly before the time of writing. Thus it becomes quite clear that the regular thematic infantry were regarded as potentially unreliable, undisciplined and easily demoralized. They would attack an enemy camp when ordered to do so less because they were brave soldiers than because they were eager for booty; they were slow moving and might hold up the commander’s main operation; in line-of-battle, in an attack upon an enemy formation, for example, cavalry soldiers and officers were to be drawn up in their rear to ensure they pressed home the attack, maintained order and did not try to flee. The commander had to be attentive to their morale, encouraging them before any combat with harangues, promises of rewards and so forth, to keep them from melting away.32 The mass of the infantry were slow and difficult to muster in time, and relatively poorly equipped: the heavy infantry were equipped with shields and spears, light troops with bows, javelins or slings, although they could also be sent on ahead with cavalry units and were regarded as more useful in this type of warfare.33 The general impression is thus that infantry seem to have had very low status compared with mounted troops, and the loss of his horse was a social as well as a disciplinary disaster for the cavalryman, especially if he then had to serve with the infantry.34
Nevertheless, they continued to play a role, sometimes an important one. Thus we read of the involvement of the Roman heavy infantry alongside the cavalry during a campaign in southern Italy in the 880s, for example, and the campaigns led by Basil I in the 870s against both Arabs and Paulicians in eastern Anatolia involving the besieging and capture of fortresses and other strongholds cannot have been carried out without a substantial force of effective infantry.35 While Leo’s account in the Tactica of the armament of the heavy infantry is drawn from that of the Strategikon of Maurice, and is unlikely to have applied to the average peasant conscript, certain details not given in the older treatise suggest an attempt to describe contemporary arms: the large, round shields, for example, as well as the battleaxes, which do not appear in the Strategikon. More realistically, where both the Tactica and the mid-tenth-century Sylloge taktikon prescribe either mail or lamellar armour (the lamellar of either iron or horn), and if this is not possible then padded garments of cotton or coarse silk, the later treatise ascribed to the emperor Nikephros II Phocas, the Praecepta militaria, lists only the latter. Given that far greater attention was paid to the heavy infantry at this time than in Leo’s time or earlier, it seems that the infantry were expected to possess only the most basic, and least costly, protective equipment.36 It should also be borne in mind that the relationship of infantry to cavalry in Leo’s Tactica reflects the balance in the Strategikon of Maurice on which it was based, and this may not necessarily reflect the actual situation of the period after the first half of the seventh century. But Basil I (867—86) is also credited with a major effort to improve the efficiency of the army, inaugurating better training than had apparently been the case before his reign, and the possibility that he was responsible for an improvement in infantry training, effectiveness and status should not be discounted.37
The fate of the East Roman infantry after the middle of the seventh century thus reflected the empire’s overall strategic and political-military situation during the period after the first Arab conquests, and, to a degree, the shifts in social relations in the provinces that these changes stimulated. It produced, in effect, a vicious circle of declining discipline and battlefield effectiveness, on the one hand, coupled with an increasing need for an irregular infantry force in ambuscades, garrison and guard duties, and so forth, and a correspondingly decreasing ability to function effectively as battlefield troops. That this was indeed the case is suggested by the efforts made by the generals of the middle of the tenth century to revive a proper, heavy line infantry element in their armies. We will return to this below.
World History
