Late Roman Combat Tactics

Chapter One

The Military Manuals and Combat Tactics in ca. 211–641¹

1.1. Graeco-Roman heritage

The most important sources for reconstructing the period combat methods are the military manuals and in this we are fortunate because we possess several dating from the era under investigation. However, we still need to begin our search with the treatises that date from the period before our era because these combat systems represent the stage in development that existed at the beginning of our era. Furthermore, as will be made clear, most of the earlier treatises retained their importance at least until the advent of the firearms, and in some cases even beyond, so that we still see men fighting in ranks and files during the nineteenth century. It was because of this that the East Romans (and others) kept on producing new copies of these just like they did of their newer treatises. For example, John Lydus (Mag. 1.47), who wrote during the reign of Justinian, names the following military theorists: Celsus (1st cent.

BC

, not extant); Paterus (2nd cent.

AD

, not extant); Catiline (not extant); Cato (3rd–2nd cent.

BC

, not extant); Frontinus (1st–2nd cent.

AD

; a fragment of De Officio Militari in Lydus and Stratagems); Renatus (Vegetius); Aelian (2nd cent.

AD)

; Arrian (1st–2nd cent.

AD)

; Aeneas (4th cent.

BC)

; Patro (not extant); Apollodoros (2nd cent.

AD)

; and the emperor Julian (Military Engines, not extant). As can be seen, quite a few of these are no longer extant, but the list given by Lydus is by no means complete, because we know that there existed treatises which are no longer extant by Pyrrhus, Alexandros, Clearchos, Pausanias, Evangelos, Polybios, Eupolemos, Ifikrates, Posidonios, Dorylaos, Theodoridas, and Bryon. In addition to which, there are treatises which are still extant by: Biton (3rd cent.

BC)

; Philon of Byzantion (3rd cent.

BC)

; Athenaios (1st cent.

BC)

; Heron of Alexandria (2nd–1st cent.

BC)

; Asklepiodotos/Asclepiodotus (1st cent.

BC)

; Onesandros/Onasander (2nd cent.

AD)

; Polyainos/Polyaenus (2nd cent.

AD)

; Julius Africanus (3rd cent.

AD)

; and other fragments. Of these the treatises of the Greek texts of Onasander, Arrian, Aelian, Polyaenus and Julius Africanus were to have great influence on the development of the East Roman combat methods, while the Latin texts of Frontinus and Renatus Vegetius were to have greater influence on the Latin-speaking west. The military treatises of Aelian and Arrian in particular had great influence on infantry tactics, because both were good sources for the various tactical uses of the Macedonian phalanx. The only real difference between the Roman and Macedonian phalanx was that the Romans usually used shorter spears (hasta, kontarion, a spear usually ca. 2.5–3.74m in length) than the Macedonians (sarissa, a pike usually ca. 5–6m in length) had used. This is already in evidence in the Kestoi of Julianus Africanus and Arrian’s Ektaxis kata Alanon.

1.2. Arrian and Frontinus

²

The texts of Arrian (mainly Ektaxis kata Alanon and Taktika) and Frontinus (Stratagems and a fragment of De Officio Militari in Lydus Mag. 3.3) are particularly valuable because both were practising soldiers who described actual Roman combat practices. When their information is combined with other sources we get a good picture of how the Roman armed forces fought during the second century ad. The picture that emerges from these treatises closely resembles the Late Roman armed forces of the third and fourth centuries.

Both authors demonstrate in no uncertain terms that the Romans deployed their footmen as phalanxes, even if Frontinus’s text proves that the Romans were also using earlier methods (manipular and cohortal arrays) alongside it. It is quite probable that these methods could also be used on a smaller scale during the Late Roman period when the Romans employed smaller units in local conflicts. The infantry formations in Arrian’s Tactics consisted of open, close (pyknosis) and locked shields (synaspismos) orders. Arrian (11.4–12.6) equated the synaspismos (interlocked-shields) with the Roman testudo-formation which he stated could be square, circular or oblong in shape. Arrian’s description of the infantry tactics included most of the variants that we can find not only in the Hellenistic military treatises and in practical use among the Roman armies and also in the military manuals of Frontinus and Vegetius. These included the lateral oblong (plagia) phalanx; column (orthia) phalanx (= ‘tower formation’, called in Greek pyrgos and in Latin as turris);³ oblique (loxe) phalanx; two-fronted (amfistomos) phalanx; two-fronted double phalanx (amfistomos difalaggia); each-other-facing (antistomos) phalanx; each-other-facing-double-phalanx (antistomos difalaggia); marching column of two phalanxes side by side or behind each other in which the file leaders were on the same side (homoiostomos difalaggia), phalanx with files of light-armed placed between heavy infantry files; hypotaxisformation, in which the light-armed were placed slightly behind or slightly in front of the edges of the phalanx in an epikampios-formation (half-square); interjection of men from the rear to the front (parembole); marching formations of one, two, three or four fronted formations; wedge (embolon); hollow wedge (koilembolos); rectangular-hollow-formation (plaision, hollow oblong in which the light infantry was placed in the middle of the heavy); hollow square ‘brick’ (plinthion; square); double outflanking (hyperfalangesis); outflanking (hyperkerasis).

Naturally, Arrian includes also a discussion of the cavalry unit formations. The cavalry unit orders were the open, close, tortoise (horses brought side-to-side so far forward and so close that each rider in the front could cover with his shield the head of the horse behind him) and ranks partially interjected into ranks (rhomboid). Arrian (16.13–4) opposed the tightening of the cavalry formation (pyknosis) too much by bringing the rear rank horses partially between those in front of them, because, in his opinion, when the horses were massed and pressed together it only panicked the horses – this means that he did not approve the same practice with the smaller 28 horsemen rhomboid. The unit formations for combat were: square (tetragônos), oblong (heteromêkês), rhomboid (28 horsemen rhomboiedês), wedge (embolos) and depthless one rank (abathês taxis). In addition to this, it is known that the Romans used the larger 128 horsemen rank-and-file rhomboids, because these were used by the Armenian and Parthian mounted archers that served as auxiliaries. Most of this followed the traditional Hellenistic material which can be found in Asclepiodotus (7.1–11) and Aelian (Matthew ed. 18–21), but Arrian’s referral to the depthless, one-rank formation (17.5) is unique to him and probably reflects contemporary Roman practices or comes from some unknown earlier treatise. The one-rank formation was to be used only when raiding or trampling enemy underfoot because it was very disadvantageous for battles. We learn from Arrian that from the reign of Hadrian onwards, Roman cavalry was expected to be equally adept at long range combat (mounted archery with crossbows and composite bows), medium range with javelins, and at close quarters with spears (lancea, lancea pugnatoria, xyston, one-handed Gallic contus, and the two-handed Sarmatian contus, which is a pike with a length of ca. 3–4m), javelins, swords and axes. Arrian also makes it clear that the cavalry could vary the amount of protective equipment according to the situation, so that at times the horses were armoured while at other times they were not. Frontinus’s text (Stratagems 2.3.23) also makes clear that the Romans expected their horsemen to be ready to dismount and fight as infantry. The cavalry was therefore purpose-built for each task so that the equipment of each unit and individual in combat reflected either the intended use of the unit or the position of each individual trooper in the combat formation. In other words, it was not the Huns who caused the Romans to become mounted archers – the Sarmatians, Alans and Parthians had a far greater role in this. We also learn from Arrian’s texts that the Roman infantry was expected to use the contus (in the Strategikon known as kontarion) and kneeling version of the foulkon/testudo/chelone (tortoise, one version of the synaspismos, interlocked-shield order) and the phalanx array against cavalry. However, it should be noted that Arrian’s influence on ancient and medieval warfare was profound, not only because of his military treatises but even more so because of his Campaigns of Alexander the Great. The military campaigns of Alexander as described by Arrian were idolized by generation after generation of military leaders.

In his De Officio Militari, Frontinus positioned the overall commander of the army (either Kaisar/Caesar/autokrator/emperor, or hyparchos/praefectus) in the centre of the army, the hipparchos (cavalry commander) or hyparchos (praefectus) on the left wing, while the right wing was placed under the so-called praitôres (praetores, praetors, governors) and lêgatoi (legati, legates). The praetores were considered strategoi (generals) and the legati their presbeutai (lieutenants, second-in-command, hypostrategoi, vicars). Even if the commander could always choose his place as the situation required, the positioning of the overall commander in the centre was also the standard positioning pattern of commanders during the Late Roman era. Vegetius placed the overall commander on the right wing, but it is clear that he had derived this instruction from some Republican-era treatise.

The combat formations listed by Frontinus in his Stratagems include the triple line of either maniples or cohorts for the Romans, and a triple line of phalanxes for their enemies (triplex acies, a variation of the Greek trifalaggia, but the latter of which could also mean epikampios); the crescent (menoeides); the convex (kurte/kyrte); the epikampios emprosthia (forward-angled half-square, or rather its variant, which was to send wings forward against the enemy’s flanks); the epikampios opisthia (rearward-angled half-square, trifalaggia); the double line either of legionaries, or with light-armed (auxiliaries) in front and legionaries behind (difalaggia); the cavalry in front and infantry behind; left and right oblique formations (oblique acie, a variation of the loxê falagx/column); the hollow square/oblong (plinthion/square, plaision/oblong); the phalanx (plagia falagx); and the modification of the arrays according to the terrain (e.g. array placed against a river, or in a defile, in which case some of the men could be placed on the heights). Most of these formations can also be found in the military treatises depicting Macedonian tactics (Asclepiodotus, Arrian and Aelian).

Italian Drill Formation in the Column of Trajan

Alan and Illyrikian Drills in the Column of Trajan

In short, it is clear that the texts of Arrian and Frontinus already contain most of the combat tactics that we will see in use during the Late Roman period. The reason for this is simple. By the turn of the second century the Romans had already come into contact with all of the different types of enemies that they were to face in the coming centuries. The Romans were fast learners, so they adopted their tactics immediately. The only details missing from the military treatises of Arrian and Frontinus are referrals to the cavalry formations, but we learn from Arrian’s Ektaxis kata Alanôn, and from the narrative sources and works of art, that the Romans typically deployed their cavalry with reserves early on. We also learn from these sources that the Romans were already using cavalry tactics that we can find in the Strategikon (6).

These include for example the African, Illyrikian, Alan, Scythian and Italian drill formations, so that the Romans were definitely using the African, Illyrian and Alan cavalry drill by the mid-second century ad. Furthermore, the Romans could also use a separate cavalry army without any infantry support, as happened for example during the German campaigns of Germanicus in 13–16 ad, even if it was more typical that the cavalry operated independently only as a vanguard for the combined army that included the infantry element. This development reached its apogee during the reigns of Philip the Arab, Decius and Gallienus, so that we see all three emperors using a cavalry formation that came to be known as the Italian Drill Formation in the Strategikon. The basic formation of the array was in use by the turn of the first century ad, as can be seen from the attached analyses of two reliefs from the Column of Trajan, but it is probable that the array reached its final shape (three divisions: flank guards and outflankers in the first line; four divisions with fill-up banda (flags of cavalry that presumably had the paper strength of ca. 500 horsemen, but which consisted of 200 to 400 men at the time the Strategikon was written) between them; a third line with two divisions) only during the reign of Gallienus, hence the name Italian Drill Formation.⁴

1.3. Modestus ca. 275 and Vegetius’s Epitoma Rei Militaris ca. 390–450

⁵

The military treatises of Modestus and Vegetius pose a problem for any analysis. It has been claimed that Modestus’s treatise is a fifteenth century forgery that has used Vegetius as its source. This, however, is by no means certain, as I have already demonstrated in my monograph Aurelian and Probus. Modestus presents a better version of the Roman infantry phalanx than Vegetius, with the implication that he may indeed have written the text for the emperor Tacitus in 275 as the text claims. It claims to be a summary of the period military practices. By and large this is true, because it depicts a pila-armed (pila is a plural of pilum heavy javelin) infantry phalanx that had apparently been introduced by Septimius Severus alongside the traditional legionary array consisting of two lines of cohorts. The only thing missing is a description of the spear-armed infantry phalanx, which we know that the Romans used alongside the pila-armed phalanx. Vegetius wrote his treatise in stages for some unknown emperor after the death of Gratian but before 450 (a copy attested to exist in Constantinople), and it describes even earlier Roman practices. Both texts appear to have been based on the texts of Cato the Censor, Cornelius Celsus, Frontinus, Paternus and the constitutions of Augustus, Trajan and Hadrian, but in such manner that both refer also to these in period context. The information provided by Vegetius (ca. 390–450) is discussed here together with Modestus (275) because its information has clearly been derived from the same source(s) as Modestus’s material – hence its data is relevant also for the late third century.

Both treatises are valuable as sources of Roman training methods and also for their descriptions of basic infantry combat methods before the fourth century, but which also remained similar after. Both treatises make it clear that the standard Roman combat formations were the single line, double line, wedge and hollow square. Vegetius adds to this list the circle (orbis) formation. The text of Modestus represents a return to the traditional Roman combined arms tactic (infantry with cavalry) during the reign of Aurelian (270–5), so that the cavalry was no longer the dominant arm of service as it had been under Gallienus. The probable reason for the change is that Aurelian lost a significant portion of his cavalry forces very early during his reign at the battle of Placentia in 271. This, however, does not mean that the Romans would have completely abandoned the use of cavalry independently of infantry, only a change in the typical usage of different arms of service when the Romans fought major battles.⁶

Both treatises include two different legionary battle formations, the first for the older legions, which consisted of two lines of cohorts (with cavalry on the flanks), the second for the legionary array used by the new legions, which consisted of a shallow six deep infantry phalanx with cavalry posted on the flanks and reserves of infantry and cavalry. Modestus (18) and Vegetius (3.16) add further details to the cavalry wings by noting that the loricati and contati heavy cavalry were placed next to the infantry array and mounted archers further out.

The legionary phalanx of Modestus and Vegetius consisted of six ranks. The first two ranks consisted of the heavy-armed legionaries (shield, armour, helmet, spear/javelin, sword), but Vegetius (3.14) suggests that the second rank also included armoured archers (i.e. multipurpose heavy infantry). The third and fourth ranks consisted of the light-armed, in that the third rank consisted of archers and javelineers, while the fourth rank consisted of javelineers or men armed with martiobarbuli/mattiobarbuli/plumbatae (lead-weighted darts). The fifth and sixth ranks once again consisted of so-called heavy armed men. The fifth rank consisted of the ballistarii/balistarii (probably users of manuballista or arcuballista), funditores (slingers), tragularii (in this case probably users of the tragula-javelin),fustibalatores (staff-sling users).⁷ In other words, the fifth rankers consisted of the heavy armed men who were actually equipped as light-armed to make the array two-fronted if the enemy approached from the rear. The sixth rank, the triarii, consisted of strong soldiers who were equipped with scutum shields and every type of defensive and offensive equipment. The Roman unit structures were based on the tent group called contubernium, which consisted of ten men of whom eight were fighting men deployed as a single file for combat, who were supported by one recruit (tiro) and one servant. In a cavalry unit the elite units had fewer fighting men with a corresponding number of squires/servants so that the total remained ten per tent group.

In the example given of the legionary phalanx by Modestus and Vegetius, the eight-man file of heavy infantry (the contubernium) had been divided into two files of four men, meaning that it had later been strengthened with two light-armed footmen. The missile equipment of the second and fifth ranks of the heavy infantry reflect the fact that a quarter to a third of the heavy armed footmen had been trained as archers ever since the second century

BC

. This also means that in a unit that had four ranks of legionaries the Romans could deploy one of the ranks as archers if there were no light-armed specialists available. The file of eight men was divisible by two, which means that the Romans could vary the depth of their formation (4, 8, 16, or 32 ranks of men) in the same manner as in the Hellenistic model. When one remembers that the regular Roman cohort had 480 footmen (divided as 3 maniples of 160 men), Modestus’s cohort of 560 footmen (five centuries of 111 men plus supernumeraries), and the Hellenistic pentekosirarchia had 512 footmen (divided into two 256 syntagmata, each consisting of two 128-men taxeis, sing. taxis), the actual difference in sizes was not particularly significant when one takes into account the fact that in practice the size of the unit was not necessarily the same as its paper strength. Most importantly, all units were divisible by the numbers two and four. This means that the new unit types that were introduced in the third and fourth centuries, which had unit structures modelled after the Hellenistic model (units as follows: 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192, and 16384), could be used alongside the older units without any problems.

In addition to this, Vegetius adds other formations to this list from some other source not used by Modestus, but which was clearly an old treatise just like the rest of the sources used by both Modestus and Vegetius: 1) left and right oblique formations (loxe in Hellenistic terminology), 2) two forward-angled formations (wings sent forward; wings sent forward while light-armed advance to protect them; in Hellenistic terminology this represents the epikampios emprosthia formation); 3) outflanking the enemy by sending the right wing forward, while the rest remained behind while lengthening the width (compare with the Peri strategikes tactical sections; 10–11 below); and 4) the use of the terrain. Vegetius’s military treatise (3.23) contains the first referral to the use of the mixed formation (cavalry placed between infantry units) in a military treatise. We do not know if this reflected the material Vegetius was using or his own judgement derived from the sources he had read. Vegetius recommended the posting of the catafracti equites either in front of the footmen or between them for use against enemy infantry. What is certain is that the Romans were using this tactic by the fourth century, which implies that Vegetius may have derived these instructions from some unknown third century or earlier text. Even if there were earlier precedents for the use of this array (e.g. by the Achaemenid Persians and Seleucids), it is probable that the Romans had copied the formation from the Sasanian Persians, together with their clibanarii cavalry and extra-large cavalry wedge, in the course of the third century wars with them.⁸

As regards cavalry warfare, Vegetius’s text (2.14) stressed both the melee skills and skill with the bow equally, with the implication that it represents the middle stage in the development of the importance of mounted archery in Roman military thinking from Arrian to Maurice. The ending of Book III (3.26) shows nicely how the period cavalry had been influenced by the Persians, Huns and Alans and was not in need of reform when Vegetius wrote his treatise in about 390–450 – all three nations are likely to have only increased the importance of mounted archery, which is reflected in Vegetius’s mutual emphasis on the use of spears and bows. Vegetius is also valuable for his descriptions of siege warfare and naval warfare. Siege warfare remained much the same throughout the era, but Vegetius’s description of naval warfare held true only for as long as the Romans preferred the combination of ramming and missile fire. The ever increasing use of the ram-spur and spurs by the early-sixth century lessened the importance of ramming in naval warfare. However, until then his advice can be seen to reflect the actual naval tactics in use.

The narrative sources confirm the use of the plagia falagx (lateral phalanx), antistomos difalaggia, hyperfalaggêsis, epikampios emprosthia (forward-angled halfsquare), menoeides (crescent), plinthion/plaision (hollow square/oblong), embolos (wedge), koilembolos (hollow wedge), triplex acies/trifalaggia (triple phalanx/line), kyrte (convex), and the continued use of the cavalry formations with reserves (i.e. the different size variants of the Italian Drill Formation) during the period 284 until 363.⁹ This means that the Romans continued to use all of the tactics described in the previous treatises by Arrian, Frontinus, and Modestus, and also by those that used earlier material like Vegetius did or which can be found in such Hellenistic treatises as those by Asclepiodotus and Aelian.

1.4. De Rebus Bellicis by anonymous ca. 365–366

¹⁰

The military treatise known with the title De Rebus Bellicis was written for two emperors (brothers Valentinian and Valens) by an anonymous author in about 365–6. The terminus ante quam for the date of the treatise is the year 367/8, because it is known that it was then that Valens put into effect policies advocated by the treatise. It is also probable that the treatise was written by someone belonging to the staff of Valentinian for Valens. The best piece of evidence for this is that the author considered formal education unnecessary (Preface 4). What counted in his opinion were intellectual abilities. Valens lacked good education and could not even understand Greek.¹¹ The anonymous author also states (Preface 10) that many of the remedies suggested by him were not unknown to the closest friends of the emperors, and that the author had just collected and assembled those from all sources. This implies that many of the suggested reforms had indeed been discussed in the imperial circles and that it had been the author who had collected them while adding his ‘own inventions’ so that the emperors (esp. Valens) could consult those more easily. In this context it is of particular note that Ammianus (30.9.4) claims that the emperor Valentinian was himself a painter and modeller and had himself invented many kinds of new arms – it is therefore not too farfetched to suggest that some of the inventions presented in the DRB were actually the handiwork of Valentinian himself and that these had been collected together by the anonymous author and then sent to Valens as a set of instructions from his brother. It is possible that Valentinian had painted at least some of the original illustrations himself.¹²

The stated aim of the treatise was to improve the military capabilities of the Roman Empire through financial and military reforms and through inventions. The author claimed that the financial problems facing the emperors had been caused by: 1) Constantine’s release of the temple treasuries into circulation that had made the rich richer and the poor poorer (the rich violated the rights of the poor with impunity); 2) fraud (debasement of coins), committed by mint workers that had resulted in inflation; and 3) corruption of governors that had exhausted the taxpayers. The author proposed to remedy these and other problems by: 1) limiting public grants to the administrators; 2) placing all workers of the mint on an island; 3) appointing upright governors; 4) forcing the soldiers to retire when they were still capable of tilling the land; 5) introducing military reserves in the form of reservists who received lower salaries; 6) reducing the taxes paid in kind by the farmers by 50 per cent; 7) refortifying the frontiers to provide security for the farmers; 8) and by introducing various new military machines and equipment meant to make the army more efficient.¹³

Many of these suggested reforms were actually implemented by Valens and Valentinian, which lends further credence to the claims that the unknown author’s suggestions reflected the ideas presented in the imperial courts by both the emperors themselves and by their courtiers, and that he merely collected these ideas for the emperors’ use.¹⁴ Similarly, it is clear that the fortifying of the frontiers with forts and towers about a mile apart against the barbarians suggested by the De Rebus Bellicis reflected the aspirations of both emperors. The imperial policy under the brothers was to isolate the Roman Empire from its neighbours through legislation, by fortifying and/or refortifying the borders, and by stopping the tribute payments in return for treaties when possible.¹⁵

As noted above, the DRB (7.1ff.) also claims that the effectiveness of the Roman armed forces could be improved by introducing improvements to the military equipment and machines and tactics. These suggestions are simultaneously either very impractical, or traditional and highly practical. For example, the ballista quadrotis is a wheeled vehicle drawn by two horses, manned by two men, powered by a windlass instead of torsion, and turned easily around to face any of the four directions by means of slots. It is difficult to see what the novelty of this machine was, unless it was the turning mechanism itself. Whatever the claimed novelty, it is certain that it was very practical and various kinds of wagon/cart-mounted ballistae continued to be used until the advent of gunpowder.

The following list of military equipment found in the treatise follows closely my analysis in MHLR 361–395 (Vol. 2):

The tichodifrus was a mobile protective cover used to protect advancing ballistae and men in formation. It could also be used in sieges as a cover. It is impossible to detect what was the novelty about this type of movable cover (possibly the nails, spears and tridents covering it), but similar equipment was certainly used before and after, at least in sieges.

The clipeocentrus was a small round shield (clipeus/parma) studded with nails. The author envisaged two uses for it: 1) by soldiers; 2) as an attachment to a tichodifrus or another device. Shields like this had been used for ages so it is very difficult to see any innovation in this.

The tribulata (plumbata-dart, with a caltrop attached) and plumbata mamillata (a dart with lead weight attached to increase its penetrating power) appear similar to standard darts used by the Romans, so it is unknown what was unique about these (possibly a different type of shaft and/or head?).

The author advocated the use of various types of cataphracted (scale-armour) scythed chariots (currodrepanus, currodrepanus singularis, currodrepanus clipeatus) for use in pursuit of the enemy. This smacks of theory, but not conclusively, because of the periodical referrals to the use of scythed chariots in warfare. Their inclusion in, for example, the works of Vegetius and Leonardo da Vinci suggest the ancient, medieval and renaissance militaries experimented with these kinds of chariots. What is notable about the intended uses of the chariots and scythed chariots is that all treatises (Julius Africanus’s Kestoi, the DRB and Vegetius) intended these as counter measures against the Persians. On the basis of Julius Africanus’s Kestoi and Vegetius’s Epitoma rei militaris (also known as De re militari) it is clear that scythed chariots were not only used for pursuit but also as anti-elephant devices. None of these experiments appear to have been long-lasting, because war chariots required very level terrain which made them impractical (a waste of resources).

There is one piece of equipment, however, that the Polish Hussars found very practical (probably a defence against lassoes) which are the wings attached to the riders’ back, as shown in the illustration currodrepanus clipeatus.¹⁶ On the basis of the actual usage of the other tactics (wagon laager together with hollow square/oblong formations and carroballistae) depicted in the DRB during the period from the late fourth century until the early sixth century, and on the basis of the inclusion of scythed chariots both in the DRB and Vegetius, I would actually suggest that there is a very high probability that the Romans did indeed experiment with these during their wars against the Persians during the very same period. The unfortunate fact is that this is not visible from our narrative sources because these give us actual details of combat very sparingly.

In the case of the thoracomachus (a defensive padded undergarment worn underneath lorica or cliuanus known as subarmalis), the author actually admitted that it was a very old invention. He just advocated the wearing of garments on top of the thoracomachus to protect it from rain. In the opinion of the DRB, a foot soldier was ready for combat when he wore the thoracomachus, socci-boots, ocreae-metal greaves, galea-helmet, scutum (presumably the regular large scutum; the author (19.1) recommended the use of a clipeus i.e. the clipeocentrus for use in wooded terrain), gladius, and lanceae-lances (the extant illustrations have just a single lancea/hasta). This was all that was needed to fight against miles (footman or soldier in general). The only unique feature in this is that the author appears to prefer the use of lighter equipment (no armour on top of the padded garment) over heavy. This indeed appears to have been the preferred type of equipment after the writing of the DRB because Vegetius (1.20) criticizes this. Vegetius claims that the Roman infantry had been protected by armour and helmets until the reign of Gratian and that it was because of this lack of armour that the Goths had defeated the Romans with multitudes of archers.

As already noted, the DBR appears to have reflected the policies adopted by Valentinian and Valens. They were short of cash, so they had to economize by not equipping all of the new recruits with armour. It was not really necessary for all the men in the infantry phalanx to be armoured. If the front, rear and flanks of the infantry formation consisted of fully equipped men, the men in the centre could wear only padded coats. Moreover, the Romans actually preferred to equip their men without armour if the army fought in the woods, forests or difficult terrain. The DRB also specifically advocated the use of the small round clipeus in wooded terrain because it was easier to use in dense woods, but this view was not shared by the Strategikon of Maurice (Str. 12.B.20.7–8) in the sixth century. Maurice advocated the use of larger than regular shields in difficult terrain, presumably because the nonarmoured soldiers required extra protection. It is therefore clear that the military theorists had different views of the practicability of the various types of shields.

The ascogeryfi were calf-skins inflated with air for use as pontoon bridges with the addition of iron stakes, ropes, and goat-hair mats. The author stated that only about 50 pack-horses and some men were all that was needed to carry these. The use of inflated calf-skins to build a pontoon bridge was an old invention, and all one can say is that the author just favoured their use presumably because these were easy to carry.

The liburna with wheels on both sides of the ship, powered by oxen yoked on machines in the hull or hold, may have been a new invention by the author. He claimed that this construction technique enabled the Romans to build so massive a warship that the speed created by the wheels would make it easy to crush enemy warships. Modern era experiments conducted between the seventeenth and nineteenth centuries have proven that this was possible, if the Romans had used horses instead of oxen. For example, Prince Rupert’s wheeled vessel outstripped the Royal Barge rowed by 16 men in 1683. However, I would not preclude the possibility that oxen could have achieved the same, if the power transmission included some features that increased the speed of the wheels progressively from one wheel to another. Nevertheless, we still do not know for certain if the Romans ever used this type of ship in combat.¹⁷ All that the modern experiments prove is that it would certainly have performed as explained.

The most important section of the treatise (19) concerns the tactics that the author envisaged for Roman armies when they were to use the above-mentioned pieces of equipment, and instructions for the use of the civilian paramilitary forces together with the fortifications for the protection of the frontiers and borders. The author stated that the dux (general) was to array his soldiers (presumably the 6,000- man legion) in three agmina with intervals in between so that each had no more than 2,000 troops. Notably, this is the way how Theodosius I the Great deployed his army at Poetovio. It should also be noted that it was possible for each of the separate 2,000 to 6,000 strong units in a single large hollow square/oblong to be deployed as separate hollow oblongs or double-fronted phalanxes to give each unit an allround defensive perimeter if the enemy managed to break inside the massive hollow square/oblong.¹⁸

On the basis of the DRB, it is therefore likely that after ca. 366 the Roman infantry phalanxes consisted of the 2,000 men moirai (each deployed as a hollow oblong) that were grouped together to form 6,000 men mere (legions). The author advised that when the army was operating in icy regions, with the implication that he expected the army to operate also during winter (note how a similar order by the emperor Maurice led to his downfall in 602), the soldiers were to wear the thoracomachus with the other relevant pieces of equipment (see above) so that they could withstand both the cold and missiles. The soldiers were also to use the clipei (small round shields), because the dense woods would make a larger shield cumbersome. The extant works of art from the fourth century confirm that the Romans indeed increasingly used the small and medium sized round shields rather than the large oval/rectangular scutum, which does suggest the probability that the above instructions reflected reality. The ascogefyri were to be used for the crossings of rivers. In short, the extant evidence suggests that the instructions of the DRB (lighter equipment etc.) reflected reality.

On the basis of the analysis of the Persian war under Julian and Jovian in 363, the DRB instructed the Romans to oppose the Persians with the infantry square formations and a greater quantity of military equipment. The tichodifrus, each manned by two men, was to be used to protect the ballistae carts/wagons and the scutati (shield-bearers). In other words, the author thought that the Roman heavy infantry scutati (= skoutatoi = shield-bearers) and carroballistae required additional protective measures against the Persians. After all, during the 363 war with Persia, the Persians had managed to penetrate inside the Roman hollow square marching formation, which had left the carroballistae which were inside vulnerable. The tichodifrus protected the scutati in two different manners: the shield protected them against missiles, while its spikes protected them against cavalry and elephants. We do not know whether the Romans used these movable shields, but what we do know is that they used caltrops against enemy cavalry and that Urbicius (at the turn of the sixth century)¹⁹ also recommended the use of a sort of Spanish Riders to protect the Roman hollow square similarly from the effects of the cavalry charge. It is also of note that the anonymous tenth century De re militari (G.T. Dennis’s Campaign Tactics) includes a childlike drawing of a wagon that may represent a device similar to the tichodifrus in the DRB. See image 4 below.

It is worthwhile to compare the attitudes of the DRB and Vegetius towards the use of scythed chariots in warfare. The DRB recommended the use of scythed chariots for pursuit when the enemy turned to flight. Vegetius’s view (3.24) was that the sickle-bearing chariots were next to useless, because they rarely found a level field and even a single wounded or fallen horse made the chariot useless, which was easy to achieve with caltrops. This, however, was not the entire truth, because Vegetius also saw a use for scythed chariots. He considered the chariots drawn by cataphracted horses with two fully armoured men (i.e. the currodrepanus of the DRB) armed with sarisae to be very useful as a counter-measure against elephants. This may mean that Vegetius criticized the DRB-based experiment made by Valens, but at the same time it may also mean that the Romans had used scythed chariots with great success against Persian elephants. Regrettably, in the absence of detailed information about the Roman campaigns against the Persians after 366 we do not know the reason for Vegetius’s comments. It is possible that Vegetius drew these conclusions on the basis of much earlier ancient material but did not see any practical relevance for his day. However, on balance the circumstantial evidence does suggest that the inclusion of scythed chariots in both the DRB and Vegetius had a contemporary relevance, so that the sickle-bearing chariots were introduced into the Roman army thanks to the DRB and this then resulted in the comments of Vegetius in the Epitoma rei militaris/De re militari. In short, there is every reason to believe that the Romans used scythed chariots at least when operating against the Persians from ca. 366 until the fifth century.

Ballistae wagons/carts and a possible tenth century tichodifrus

Source: DRM, Vatican Library, Cod.Gr.1164, f.238v, Rome (Dennis, 1985, Campaign Tactics, 261).

The childlike drawings of military machines from the anonymous tenth century DRM demonstrate the remarkable continuity of the Roman military methods throughout the centuries. The figures 1 and 2 are clearly wagon-mounted ballistae and the figure 3 is a cart-mounted ballista, but the figure 4 is less easy to explain. However, since the figure 4 was to be used in the field together with the ballistae wagons and carts and because it bears some resemblance to the tichodi frus of the DRB, it is in my opinion likely to be the tenth century equivalent of the tichodifrus, a device used to protect the ballistae wagons and men operating these.

Another important tactical detail in the DRB refers to situations in which the army could not be deployed for combat. In that case the author instructed the Romans to array their field artillery and machines around the packed formation so that the enemy could not threaten its flanks. With this in mind the DRB instructed the general to take twice as many beasts of burden as usual to secure an adequate supply of animals to transport the tormenta (field artillery) because it was particularly important to possess a reserve of these to replace those exhausted or killed. The Romans had a long history of the use of field artillery to protect their army and these arguments were sound. The instruction to increase the numbers of beasts of burden was a particularly relevant reform in the aftermath of Julian’s failed Persian campaign, which failed in particular because Julian lost too many of his beasts of burden. In sum, the author recommended the use of a variant version of the wagon laager (carrago, karagos) which we find the Romans using from this date onwards until the sixth century, and an increasing of the number of beasts of burden as a security measure. For an analysis of these different types of hollow square/oblong and wagon laager formations, see Chapter 9 on infantry battle formations.

Since the DRB (20) reflected the policies that Valentinian and Valens adopted it is not surprising to find out that the treatise promoted the building of effective fortifications around the empire. The DRB required the building of a continuous line of forts (castella) at one mile intervals with strong walls and towers. In order to save money, the forts were to be built by local landowners, and these were to be garrisoned by the vigiles (guards = i.e. civilian burgarii) and the countryside was to be patrolled by agrarii (a paramilitary force of peasants?). The emperors implemented a watered down version of this, because the intervals between the forts and guard towers only rarely reached the recommended minimum distance. It was presumably either too costly or unnecessary to build as many forts as recommended.

In sum, the extant evidence suggests that these regulations (lighter equipment, hollow squares, mobile artillery and machines, strengthening of the frontiers with fortifications etc.) became the military doctrine followed by the Roman emperors from 366 onwards. This doctrine was an expression of the lessons learned in the course of the previous 30 years of wars, and in particular the reflection of the defeat in the war against the Persians in 363. The narrative sources together with the texts of Urbicius and Syrianus Magister (see below) prove that the different versions of the hollow square/oblong (plinthion, plaision) and wagon laager (carrago) became the standard combat methods for the period from ca. 366 until the 530s. The only other combat methods that can be found in the narrative sources are the standard plagia falagx (lateral phalanx), mixed formation (cavalry posted between footmen) and the Italian Drill Formation for cavalry. During this period the commanders appear to have varied the infantry battle formations as follows: 1) if the infantry forces consisted of veterans, the commander could use either the lateral or mixed phalanx; 2) if the infantry consisted of veterans and the situation required an allround defensive perimeter the commander used the hollow oblong/square; 3) if the quality of the infantry was poor or consisted of untested men, the commander used either the traditional wagon laager, or a wagon laager with extra safety measures, or the wagon laager described by the DRB (ballistae carts and footmen protected by the tichodifri).²⁰

1.5. Definitiones (second to third century) = Hermeneia (before sixth century)

²¹

There also exists a glossary of military terms called Hermeneia, which may date from the Late Roman period. The Hermeneia is not an original treatise, because it contains the earlier text known as the Definitiones. The Definitiones (also known as Glossarium militare or Lexicon militare) in its turn is merely a lexicon of military terms (63 entries with additional subentries) based on earlier military treatises such as Arrian, Aelian, Asclepiodotus and other Hellenistic treatises. According to Dain (1967, 332), the entries 29 and 59 of the Definitiones are unique to this lexicon. This, however, is only partially true. We find corresponding information at 29 (parembole vs.parentaxis) also in Asclepiodotus (6.1), and similarly the referral to the leptysmos (thinning of the phalanx by dividing the file of sixteen into two files of eight) at 63 is almost a word-for-word copy of Aelian’s text (Matthew ed. 49). In short, there is nothing unique in this text. Notably, the Peri strategikes (25.13–5) warned against the thinning of the depth too much in such cases, which is not found in either Definitiones/Hermeneia or Aelian. Bury dates the Hermeneia to Justinian’s reign, but Dain leaves the exact date open. After all, the Hermeneia is merely a copy of the earlier glossary in the Definitiones. The principal value of the text for this study lies solely in the preservation of the military lexicon of phalanx warfare which remained the same. The exact dating of these two similar treatises is therefore meaningless.

1.6. Urbicius: Epigrams, Taktikon, Epitedeuma ca. 505–18

²²

A man known as Urbicius wrote two tactical treatises the Taktikon and the Epitedeuma, and two epigrams during the reign of Anastasius (491–518).²³ These three texts are of the greatest importance for the analysis of Roman combat doctrine and tactics at the turn of the sixth century. The epigrams were a preface to the Taktikon and the Epitedeuma. The Taktikon in its turn is an epitome of Arrian’s Taktikon (second century ad) while the Epitedeuma presents the author’s invention that could be used to protect the infantry against barbarian cavalry charge. My own view is that Urbicius was a magister militum per Orientem (stratēlatēs tes Anatolēs) and patricius as the later sources claim. Like so many other military appointees of Anastasius, he lacked military experience which he sought to remedy through the study of earlier military treatises. The results of this study are visible in the treatises he wrote.

Urbicius sought to improve the general standard of the period infantry with his treatise that summarized Arrian’s treatise on the Macedonian phalanx. Urbicius basically lists all of the typical Hellenistic unit names and sizes for phalangites/hoplites (16,384 heavy-armed), psiloi (8,192 light-armed) and cavalry (4,096 horsemen), which were based on the structures of 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192, and 16384; unit orders (pyknosis, shields rim-to-rim; synaspismos, shields interlocked; chelône/testudo, a formation which was either square or round in shape, the equivalent of the foulkon); commands (turning, wheeling, about turn, counter-marches, reinforcing etc.); and various types of phalanx formations. The combat formations included: 1) regular Macedonian plagia phalanx; 2) prostaxis (one stifos of 4,096 men of light-armed men placed forward on each of the flanks of the phalanx, a sort of crescent array, which was probably used by Narses at Taginae/Busta Gallorum in 552); 3) entaxis (insertion of light-armed between the files of hoplites called by Asclepiodotus parentaxis); 4 hypotaxis (the light-armed placed behind the flanks of the phalanx – Arrian’s original text likened it to the epikampios = epikampios opisthia); and 5) hollow square or round chelône/testudo (tortoise), with wagon mounted ballistae on the basis of the Epitedeuma. The baggage train, the touldon/skeuoforoi, was to be arrayed as follows: 1) if it was expected that the enemy would be in front, the baggage train was to be placed behind the army; 2) if it was expected that the enemy was to the right, then the baggage train was to be on the left of the army; 3) if it was expected that the enemy was to the left, then the baggage train was to be on the right of the army, 4) if the enemy was suspected of threatening all sides, then the baggage train was to be placed inside (i.e. inside a hollow square/oblong or circle or double phalanx).

However, Urbicius knew that inexperienced footmen were not really ready to use the complex manoeuvres represented by the Taktikon/Arrian so he recommended in the Epitedeuma the use of the new anti-cavalry devices invented by him alongside the wagon mounted ballistae (carroballistae, ballistroforoi hamaxai) and the hollow square formation to negate the moral impetus of the barbarian cavalry charge. He sought to negate the psychological impact of the impetuous barbarian cavalry charge (the impetus) in the minds of the soldiers by building around it an encircling fence (this was his new invention) outside the infantry square at such a distance that the enemy arrows would be unable to reach the soldiers. When this took place the ballistae wagons (carroballistae, ballistroforoi hamaxai) on the outer edges of the square began to shoot at the attacking enemy horses. The ballistae outranged the bows significantly (ballistae range was three bowshots) so this was expected to work well. The combination of the ballistae and fence would result in the death of the foremost barbarians so that these would pile up in front of the fence and also on the nails of the fence. The following ranks of barbarian cavalry would only pile up in the same place.

1.7. De scientia politica dialogus /Peri politikes epistemes (On Political Science) ca. 530–550

²⁴

The surviving portions of the On Political Science (Peri politikes epistemes/De scientia politica dialogus) traditionally attributed to Peter the Patrician also contain a short treatise dealing with military matters. The treatise is usually dated to a period before 532, but I prefer to date it to the period from the late 540s to the 550s because the importance of having both infantry and cavalry present would have been the target of particular attention at a time when Procopius felt it necessary to defend the use of cavalry in his Wars and Narses had achieved great advantage from the use of infantry. Furthermore, the treatise concentrates on the use of the lateral phalanx (plagia falagx) and cavalry, with the implication that its information is more relevant to the period when the infantry was expected to perform better than when Urbicius wrote his treatise. Most of the treatise deals with the relationship between the civilian population and the armed forces. The most important piece of information in the treatise concerns the infantry. The training scheme shows that the heavy armed footmen were expected to fight like ancient hoplites by using shields and doru-spears. The text stressed the value of the infantry in combat. The infantry was deployed with cavalry wings, separate flank guard cavalry next to them and with reserves of infantry and cavalry where needed. On its own the cavalry was deployed as two lines. In short, the treatise shows the Romans using infantry phalanxes in which the men were armed like hoplites and it also shows the Romans using their cavalry and infantry with reserves. The unknown author lamented the period preference for cavalry because in his opinion infantry and cavalry were complementary and equally important for success. This demonstrates nicely the sixth-century prejudices of the period military men against which the unknown author and others spoke.

1.8. Syrianus Magister: Peri strategikes/strategias, Naumachica and Rhetorica militaris (ca. 545–52)

²⁵

Modern consensus opinion among historians is that Syrianus Magister (Syrianos Magistros) is the author of three military treatises entitled Peri strategias/strategikes (also known as De re strategica), Naumachica and Rhetorica militaris. Even if we cannot know this for certain, it is quite probable that this is the case. The name Syrianus comes from the Naumachica. The other two treatises were previously credited to anonymous authors.

Syrianus’s treatise is a compilation of earlier materials and personal observations and instructions. Most of the modern authors suggest that Syrianus used the lost naval treatise of Aeneas Tacticus (ca. 350–340s

BC)

as his source for the Naumachica. This is possible, but equally well he could have used the lost naval treatise of Aelian (Devine ed. p.43, 2.1 on p.45; Matthew ed. 2), which modern research has established to be his principal source for the infantry tactics. The Rhetorica militaris was modelled after the theoretical handbook on oratory written by Hermogenes (ca. 160–230 ad), but the actual contents are the work of Syrianus himself. The Peri strategikes is a compilation of material drawn from Aelian, Apollodorus, De arcus usu and some other unknown military and political treatises which Syrianus has updated with his own material and streamlined to form up a unified text.

None of the three treatises contain any specific dating information, which means that Syrianus’s treatise has to be dated on the basis of its contents. The inclusion of the name Belisarius gives us the date post quam, and Leo VI the Wise’s use of Syrianus’s treatise as a source gives us the ante quam date.²⁶ The modern consensus opinion among the historians is that Syrianus wrote his treatise during the so-called Middle Byzantine Era, so that the ninth century is seen as the likeliest era.²⁷ However, I do not share this view. I am inclined to accept the sixth century dating that H. Köchly and W. Rüstow (1855), and G.T. Dennis (1985) gave to the Peri strategikes, which is obviously to be extended to cover the rest of the treatises now credited to Syrianus. In fact, this is also the dating scheme preferred by Constantin Zuckerman (1990, 216), who suggests that Syrianus wrote his treatise in the late sixth-early seventh century. Indeed, the latest possible dating for the treatises would be the early- to late-seventh century for the reasons given below. In short, the likeliest date for the Peri strategikes is the reign of Justinian after the Empire had recovered from the plague and now faced a naval threat (specifically the years 545–552), or the early part of the reign of Justin II.

It has also been suggested that the different parts of the Peri strategikes were written at different periods of time. This suggestion was first made by Köchly and Rüstow. On the basis of the fact that the Peri strategikes books 1–32 were longer and more detailed than books 33–47, they suggested that those were not in their original form, but had been shortened by a later epitomizer. This has caused D. Lee and J. Shepard (29–30) to suggest that these books represent an independent work by an unknown author who then added these books to the original text during the Middle Byzantine period. Neither of these arguments is valid, because the Peri strategikes is a compilation which the author has updated with his own observations. The known sources for books 1–32 are Aelian and Apollodorus while it is conjectured that the last books, 44–47, have been borrowed from an unknown archery treatise which has received the name De arcus usu in research literature.

In short, it is probable that all of the books contain at least some information that has been borrowed from some other treatise or treatises. As will be made clear below, one of the sources used was clearly a treatise which contained similar instructions for the arraying of the infantry phalanx with cavalry (PST 35), as can be found in Modestus (18) and Vegetius (3.16), and which was no longer valid after the Strategikon was published. The use of the different sources obviously influenced the length of the material contained in each book, which in its turn means that the more concise presentation of the material in the books 33–47 cannot be taken to mean that these were either later abridgments or books added to the original by some unknown author.

I will first go through the arguments in favour of Middle Byzantine dating for the Peri strategikes, together with counter arguments, after which I will add material which supports the sixth-century dating. Philip Rance (2007, 707–19) has reiterated the standard arguments in favour of the sixth century dating and against it, after which (719ff.) he added new arguments in favour of the Middle Byzantine dating. Rance has noted that the following arguments do not provide decisive evidence for the dating: 1) the referrals to triumphs (these were used at least until the eleventh century); 2) Belisarius (well-known person who could also be used as an example by later authors); 3) Arabes using ambushes (the name Arabes was more typical for the Middle Byzantine authors, but can be found e.g. in the text of Evagrius); 4) kataphraktoi (typical classification for heavy cavalry later, but can be found in John Lydus); 5) Persians as enemies (this name was also used later to mean Arabs, Turks and others); 6) current emperor inciting enemies against enemies (Justinian famous for this, but typical for most of the Roman emperors); and 7) the prominence of archery (typical also after the sixth century). In addition to this, Rance demonstrates that the absence of the elephants from the Peri strategikes is not relevant for the dating, because elephants were not mentioned in the Strategikon either. However, he notes that the inclusion of the Naumachica/Naumachia in the companion points to the post-649 date of composition, when in his opinion instructions for naval warfare would have been more relevant than in the sixth century – the Strategikon includes only a section on riverine warfare. This, however, is a mistake because the sixth century Romans certainly faced serious naval threats in the form of the Vandals and Ostrogoths, both of which could easily have led Syrianus to the conclusion that the inclusion of the naval treatise was also necessary. The need for the inclusion of the naval treatise in the compendium was particularly relevant during the years 545–52²⁸ because it was during those years that Totila built a sizable fleet. For example, the navy which he took to Sicily consisted of 400 warships and a large fleet of cargo ships. The situation had changed by the time the Strategikon was written, hence only the inclusion of the riverine warfare in that treatise.

The infantry phalanx of the Peri strategikes (16) frontal view with files ca. 63 cm wide. The spears of the first four ranks reached the front. The front rankers had extra large shields with a spike attached to the shield boss for shield bashes and all men had conical helmets with a spike on the top for head buts.

The infantry phalanx of the Peri strategikes (16). The hoplites are also depicted as multipurpose footmen with bows and arrows as expected by the Peri Strategikes (36, see also MHLR 6, 342).

Note the hasta-spear, sword, hoplite-shield with protruding shield-boss, scale-annour, pteruges, helmet and the arrow-quiver. The shield with the spike could have been used by front-rankers in phalanx formation. This suggests the use of multipurpose troops and phalanx formation later described by Syrianus Magister in the Peri Strategikes (16, 27, 36).

A ‘Trojan’ archer from the same manuscript. Note the use of scale-annour also by the archers.

SOURCE: Vergilius Romanus, origin probably western from the 4th or 5th century.

Large infantry shields in art

1) An infantryman in the Altar of Domitius Ahenobarbus (1st cent.

BC)

. Drawn after Bishop and Coulston. Note the size of the rcufum-shield.

2) A footman in the Glanum stele. 1st century

BC

. Author’s drawings. Note the baldric and homed helmets that predate the late Roman era helmets depicted in the Arch of Constantine the Great by more than three centuries.

3) Heroes duelling, a wedding chest, Breton, 12th century. Yet another example of the use of large shields in the areas fonneriy belonging to the Roman Empire. Drawn after the photo of the painting in NicoHe (1999,109). It is probable that this was the typical size also for the late Roman shields in situations in which the ancient author referred to the use of extra large shields.