The period from 249 until 269, just before the rise of Aurelian on the throne in 270, is among the darkest in the history of the Roman Empire. It was then that the Goths annihilated the Roman field army together with the emperor Decius at the battle of Abrittus in 251. It was during those years that the Franks, Alamanni, Sarmatians, Goths, Heruls and many other barbarians ravaged and pillaged Roman territory in North Africa, Spain, Gaul, Raetia, Noricum, the Balkans and Asia Minor. As if this would not have been enough, the Persians under their great military king of kings Shapur I repeatedly ravaged Roman Mesopotamia, Syria and Asia Minor and even took the emperor Valerian prisoner. It was under Valerian’s son Gallienus that things took a distinct turn for the worse thanks to the numerous usurpations and revolts at the same time as the empire was still threatened by the barbarians and Persians. As a result of this chaos Gaul separated itself from the Roman Empire and became the Gallic Empire led by the usurper-emperor Postumus. In the meantime, the East became the private domain of the Palmyrene royal house thanks to the fact that Gallienus was forced to rely on it to crush the Persians and the usurpers (Macriani and Ballista). At the time of his death in 268 the situation was even worse because the Goths and their allies still roamed in the eastern Mediterranean and the city of Milan was still controlled by the usurper Aureolus. In addition to this, the frantic and erratic monetary policies of Gallienus had created a situation in which the populace no longer trusted the value of their coins. However, it was also under Gallienus that the revival of the Roman Empire began, thanks to his relentless efforts to save it. Regardless, the situation was still critical at the time of his death, and it required the frantic efforts of three emperors – Claudius II, Aurelian and Probus – to restore order. All of these generals owed their initial rise to Valerian, but their rise to prominent positions was still one of the most enduring legacies of Gallienus, who clearly recognized their military talents and promoted them to higher posts.
Gallienus’ emergency ad hoc reforms included, for example: 1) the re-creation of cavalry army, 2) the creation of new offices; 3) restrictions put on senatorial careers, 4) defence-in-depth; 5) the use of large regional field armies; 6) new forms of propaganda. In most of these cases Gallienus merely followed the precedents set up by his predecessors, but he was certainly prepared to go further in all of these fields than any of them, so it was under him that many of the previous exceptions became the norm. The equestrian officers had already been placed in command of several legions by the previous emperors, but under Gallienus they became the commanders of all legions. The senators, however, retained some of the governorships, with the result that they still commanded legions, each of which now had an equestrian prefect in command. The unofficial titles for the temporary generals – for example the dux, comes, comes domesticorum, magister equitum and tribunus et magister officiorum – became permanent under him. Even though the Romans had sometimes used large cavalry armies in the past, it was under Gallienus that it became the dominant arm of service when he recreated the large cavalry army previously destroyed under Decius. The Romans had also used the defence-in-depth concept in the past whenever necessary, but under Gallienus it became more widely used than ever before thanks to the many threats that he faced. Gallienus was also forced to group his forces, which consisted largely of detachments drawn from other units, into regional field armies and one personal field army. This was not a new practice and had been used for example by his father and several of his predecessors whenever there had been a need to form a large field army to meet an emergency, but the scale of this phenomenon was unprecedented under Gallienus.
Just like any empire, the Roman Empire possessed armed forces and security apparatus. The official security apparatus consisted of: 1) the land forces (legions, auxiliaries, national numeri); 2) navy (Praetorian and Provincial); 3) vigiles (firemen/policemen); 4) urbaniciani (policemen, urban combat troops); 4) the imperial bodyguard units (praetoriani; equites singulares Augusti/germani; aulici/collegia/scholae/protectores; and evocati Augusti). Their strength was bolstered by: a) treaty-bound allied forces (foederati); and b) the paramilitary civilian citizen militias and policemen which patrolled seas and land areas. However, the division of the Roman Empire into the central Empire led by Claudius II, the forces of the usurper Aureolus at Milan, the Gallic Empire of Postumus, and the de facto Palmyran Empire in 268 means that the composition of armed forces varied in the different parts of the Empire. In the legitimate part of the Empire under Claudius II it was the cavalry – the equites of Gallienus – that formed the crème de la crème of the armed forces. In Gaul, however, it was the infantry which formed the flower of the armed forces. These consisted of those forces that had been left in the many garrisons of the area, and of the new locally recruited forces and very large numbers of German mercenaries (mainly Franks and Alamanni). The German mercenaries consisted also mainly of infantry and it was these forces that became the precursors of the late Roman auxilia palatina.
In normal circumstances the Imperial bodyguards and the Legio II Parthica (organization is given below) formed the principal mobile reserve army at the immediate disposal of the emperor. However, at the time of the death of Gallienus the situation was different because he had formed a separate imperial field army in about 257/8. The core of this force consisted of the cavalry units grouped together as equites under the cavalry commander (in the sources the hipparchos presumably meaning the magister equitum or comes domesticorum equitum) and of the infantry army consisting of the units drawn from the regular bodyguard units, legions, and auxiliaries. It is probable that the infantry had a separate commander (possibly comes domesticorum peditum or magister peditum) just like the cavalry and that the overall commander of both cavalry and infantry was tribunus et magister officiorum (Tribune and Master of Offices) who appears to have been Claudius II at the time of Gallienus’ murder.
In normal circumstances the imperial bodyguards consisted at least of the praetoriani (10,240 foot, 1,920 horse?), 300 cavalry speculatores, a numerus of statores Augusti, equites singulares Augusti (2,048 horse), frumentarii (spies), and of the peregrini (spies). There are also reasons to believe that the staff of the Imperial Stables (stratores/equerries and grooms, the stablesiani, under tribunus stabuli) and the evocati Augusti also belonged to the bodyguards of the emperor. It is unfortunate that we do not possess any reliable numbers for the frumentarii, peregrini, stablesiani and evocati Augusti. Similarly we do not know the size of the bodyguard unit variously called the aulici/collegia/scholae/ protectores/protectores domestici. The problem is accentuated by the fact that the meaning of the names scholae, protectores and domestici changed in the course of the third to fifth centuries; but a good educated guess is that the protectores/domestici would have encompassed at least roughly the same number of soldiers as the scholae of one half of the Empire in the fifth century – a minimum of 3,500 horsemen. The reason for this conclusion is that the scholae appear to have been commanded by the protectores/domestici and at the beginning of the fourth century all those forces that were commanded by them were called protectores. The problem with this is that even the protectores proper seem to have included far more men than this before the reign of Julian.1 The other units, which were in normal circumstances at the immediate disposal of the emperor, consisted of the Urban Cohorts (cohortes urbanae/urbaniciani, 4,500 policemen usable also as a military force), the vigiles (7,000 firemen and policemen) and of the detached naval cohorts from the two Praetorian Fleets (Misenum and Ravenna) all located in Rome and of the Legio II Parthica located at Alba (Albanum). The urbaniciani operated throughout the Empire to secure supplies for the capital on behalf of the Urban Prefect, and it was because of this that we find separate urban cohorts at least at Ostia, Puteoli, Lyon and Carthage. The soldiers on leave in Rome could also be used for military purposes, but obviously these did not count in the making of any military plans.2
At the end of the third century the equipment used by the legionaries and units of bodyguards was undergoing a change in fashions which then became the standard equipment of the so-called late Roman period. What remained the same was that the legionaries and bodyguards were taught how to fight with empty-handed (martial arts) techniques and with weapons such as the short sword (semispatha, gladius), medium to long sword (spatha), various types of javelins and spears, shield, and the throwing of stones by hand or sling. In addition to this, at least a third of the infantry were also taught how to use bows, presumably including in this figure the crossbows. In contrast, all horsemen were required to able to use bows and crossbows while mounted.
It is probable that the change in fashions resulted from the dominance of Illyrian soldiers in the Roman armed forces. This was the result of their better availability during the reign of Gallienus whose principal recruiting ground was in Illyricum. The main result of this was that ridge and segmented helmets started to replace the other types of helmets, while it became fashionable to use the pilleus Pannonicus (round Pannonian hat, often made of fur). Typical protective equipment of this period consisted of the: 1) flat (or slightly curved) oval and round shields; 2) scale and mail armour; 3) metal- made muscular armour; 4) muscular leather armour and other types of ersatz armour; 5) ridge and segmented helmets. Just like before and later, it was also possible for men to be unarmoured to increase the mobility of the unit. In short, ever fewer men used the older lorica segmentata type of armour and the rectangular scutum. However, this does not mean that the older types of equipment would not have continued in use, because we find these still in use at least until the turn of the fifth century and in some cases (e.g. pseudo- Corinthian helmets) much later still, but it does mean that these were by then relatively rare. Similarly, just like before and later, legionaries could also be equipped either lightly (no armour, small shield and light javelins or bows or slings) or with heavier equipment (armour, heavy shield, and heavy javelins or long spears) as required by the situation, which means that we should not draw too drastic conclusions from the evidence. The men fought basically in the same manner as before even if military fashions had changed. The main difference between the typical legionary and the bodyguard was that the latter often wore more expensive equipment than the former.
The number of legions, auxiliary units and numeri at the disposal of the various rulers is not known, but what we know is that their internal organization was based on the following models. The legions followed three basic models: 1) the old standard legions; 2) the legions based on the Parthian legions of Septimius Severus; 3) the temporary legions built out of the detachments drawn from the legions. The Parthian legions may have been based on a different unit structure and their tactics with unit depths and equipment may have been different from the standard units (see Appendix 1).
The old standard legion consisted of about 5,120 heavy infantrymen plus recruits, servants, horsemen and specialists. Such legions consisted of cohorts (quingenary 480 men and milliary 800 men), maniples of 160 men, centuria of 80 men, and contubernia of eight men. Each contubernium (tent group/file in rank and file array) consisted of eight men, one green recruit and one servant so that it in truth consisted of ten men under a commander of ten, called a decanus. Each of the legions had also a cavalry contingent which consisted of 500 to 600 horsemen (but typically a 512 horsemen ala) all of whom were taught how to use spears, javelins, swords, crossbows, shields and composite bows so that these could be used for a great variety of missions. Severus’ Parthian legions differed from the above in that they had ca. 6,000 footmen and 732 horsemen. The attached diagrams show the basic structures of the various types of legions.
STANDARD LEGION
Probable command structure of the regular legion c. AD 90–260
- 1 Legate (S) until the reign of Gallienus who abolished the office; or Prefect (E) for the Egyptian and Parthian legions. After Gallienus the commanders were prefects (E); commander of the legion.
- 1 Laticlavian tribune (S) changed by Gallienus into tribunus maior (E); in charge of one cohort and second-in-command of the legion.
- 1 Praefectus Castrorum (camp, medics, siege equipment etc.) (E)
- 1 Praefectus Fabrorum (workmen, construction etc.) (E)
- 5 tribunes (E) each in charge of one cohort 480 men.
- 1 tribunus sexmenstris (in charge of cavalry?) (E).
- 5 centurions of the 1st Cohort (incl. primus pilus who could act as praepositus for the cohort).
- 54 centurions (called centenarii by the end of the 3rd century):
- 5 unattached centurions who could be detailed for variety of purposes; these could be used e.g. as acting praepositi (commanders for the cohorts (à 480 men)).
- 9 x 1 centurion each in charge of two centuries (2 x 80).
- 9 x 4 centurions each in charge of one century (80 men).
- 4 cavalry centurions each with 128 horsemen.
- 64 infantry decani one of whom was optio/second-in-command to centurion (each decanus part of and in charge of their 8 man file/contubernium, in addition to which came a tiro/recruit and one servant used for the guarding of the camp).
- 16 cavalry decurions (each in charge of their 32 horsemen turma).
- 1st cohort 800 men (5 centuries à 160 men) plus 100 recruits and 100 servants.
- cohorts 2–10 = 9 x 420 footmen (including the decani 480) plus 60 recruits and 60 servants per cohort.
- 496 horsemen (with the decurions 512; Vegetius may have been wrong in adding the decurions to the strength of the turma, because the Roman cavalry organization was based on the Greek one; however, if Vegetius is correct then these should be added to the total for a total of 512 + 16 decurions plus about 128 servants/squires.
- at least 715 artillerymen in charge of the 55 carroballistae (cart-mounted bolt/arrow shooters) and 10 onagri (single-armed stone-throwers).
- 10 speculatores (formerly scouts), but now couriers, police officers and executioners.
- proculcatores and exploratores scouted the roads. It is not known whether these counted as part of the cavalry or were separate from it. In practice the mensores could also act as scouts.
- unknown numbers of military police with the title of stator, and unknown numbers of guard dogs. Inside each camp there was also a police station called statio under a tribune. Some of the soldiers were also used as sentinels (excubitores) and there were also other specific guards for various things.
- in addition there were unknown numbers of other specialists and bureaucrats consisting of surveyors, campidoctor (Chief Instructor), haruspex (who read the entrails prepared by victimarius), pullarius, actuarii, librarii (librarius a rationibus worked also for the state post and could act as a spy), notarii (could act as spies on the activities of the commander), commentariensis (archivist under head curator), heralds, standard-bearers, draconarii, cape-bearers, trumpeters, drummers, engineers, workmen, artisans, hunters, carters and cartwrights, doctors, medics etc.
- the legates/prefects were also guarded by a unit of singulares (both inf. and cav.), which consisted of detached auxiliaries (confusingly the staff officers in training could also be called singulares). These bodyguards were replaced by protectores detached by the emperor from his staff at the latest during the reign of Gallienus as a safety measure against usurpations.
- the legion also included beasts of burden (which, depending on the units, could be horses, asses, mules, camels, oxen).
(S) = senatorial office; (E) = equestrian office
THE PARTHIAN LEGIONS CREATED BY SEPTIMIUS SEVERUS (HOWEVER, SEE ALSO APPENDIX 1).
Vegetius’s Ancient Legion (Epit. 2.6ff.) with additional comments in brackets.
- 1 praefectus legionis formerly legatus; commander of the legion.
- 1 tribunus maior; appointed by the emperor in charge of one cohort (probably the 1st; second-in- command of the legion).
- 1 Praefectus Castrorum (camp, medics, siege equipment etc.)
- 1 Praefectus Fabrorum (workmen, construction etc.)
- tribuni minores from the ranks (6 tribunes? put in charge of the cohorts and cavalry alongside with the praepositi).
- 5 centurions of the 1st Cohort (Vegetius’ list differs from the other known lists of officers and is also 100 men short of the 1,100 men he gives for the 1st Cohort)
primus pilus in charge of 4 centuries/400 men (this probably means that there were 440 men
which consisted of 4 centuries each with 110 men)
primus hastatus ‘now called ducenarius’ in charge of two centuries/200 men (probably 220 men)
princeps 1.5 centuries/150 men (probably 165 men)
secundus hastatus 1.5 centuries/150 men (probably 165 men)
triarius prior 100 men (probably 110 men)
- 5 centurions for the cavalry.
- 45 centurions of the 2nd to 10th COs each in charge of 110 men ‘now’ called centenarii.
- 1st Cohort: 1,105 footmen (this probably means that there were 720 heavy infantry deployed four deep, two in the front and two in the back of a six-rank-array, and 360 light infantry deployed two deep between the heavy infantry ranks + 10 optiones, 10 standard-bearers, and 5 centurions). 132 horsemen (128 horsemen and 4 decurions; in truth the decurions may have been part of the 128 horsemen in addition to which came one centurion, 2 musicians and one standard-bearer; when trained to do so the 128 horsemen could form up a rhombus so that at each apex stood one decurion).
- 2nd to 10th Cohorts: 9 x 555 footmen (this probably means that there were 360 heavy infantry deployed four deep so that these formed the two front ranks and two rear ranks in a six deep formation, and 180 light infantry deployed two deep between the heavy infantry ranks + 5 optiones, 5 standard-bearers, 5 centurions).
9 x 66 horsemen (64 horsemen and 2 decurions; as noted above the decurions should possibly be included as part of the 64 horsemen; the 64 men could be formed either as a wedge or two rank-and-file oblongs).
- artillerymen (55 carroballistae each with 11 men and 10 onagri per legion), ‘squires’, servants,
various kinds of standard-bearers and musicians, and other specialists like clerks, medics, wood- workers, masons, carpenters, blacksmiths, painters, siege-equipment builders, armourers etc. (aquiliferi, imaginarii/imaginiferi, signiferi/draconarii, tesserarii, optiones, metatores, librarii, tubicines, cornicines, buccinators, mensores, lignarios, structores, ferrarios, carpentarios, pictores etc.)
- On the basis of my above hypothesis regarding the organization behind Vegetius’ figures, a possible overall fighting strength of Vegetius’ legion may have been: 6,100 footmen (3,960 heavy infantry; 1,980 light infantry) plus supernumeraries; 726 cavalry; at least 660 artillerymen with 55 carroballistae and 10 onagri; at least 510 recruits left to defend the marching camp together with the servants and workmen. The extra men on top of the older paper strengths may actually represent the recruits not normally included in armed strengths, but one cannot be entirely sure of that. It is possible that the cohorts were really 555 men strong because Dio’s reference to 550 men does suggest this. The obvious problem with Vegetius’ information and my reconstruction based on it is that we have practically no evidence to corroborate it, but at least if one presents the information in this manner it does make sense and is therefore plausible. Vegetius notes that the legion could also include several milliary cohorts, which probably refers to the Praetorians which had milliary cohorts after Septimius’s reign, or it refers to the practice of Vegetius’s own day to group together different units to form ‘temporary legions’ that were later called mere by the East Romans (sing. meros/division).
The equipment of the auxiliary forces which belonged to the ‘line infantry and line cavalry’ was basically the same as in the legions so that the auxiliary units typically used the flat (or slightly curved) oval or round shields, spatha-longsword, spears, and javelins, but other types of equipment was also used. The auxiliary units were either quingenary or milliary in size. Some of the auxiliary units specialized in a particular type of combat (e.g. cavalry cataphracts, mounted archers, foot archers, slingers), but the vast majority were required to be able to perform all types of combat either as line infantry/cavalry or as skirmishing forces. There also existed a special type of auxiliary force called numeri. There was no uniformity in their unit size or use in combat. In addition to this there were barbarian forces called laeti (defeated enemies) which were settled on Roman territory with the duty of providing soldiers when required to do so. Despite being called auxiliary forces the above-mentioned auxiliaries were actually part of the Roman army. At this time the real auxiliary forces were the allies who were called foederati (treaty-bound allies). These forces consisted of whatever type and size of force the ally could contribute for their Roman ally. The following list (based on my earlier studies) summarizes the information.
Approximate size and organization of auxiliary units:
The attached drawings and the images in the plates depict some of the types of forces employed by contemporary Romans. One of the drawings in the text (page 14) is a three-panel drawing of some of the scenes from the Arch of Constantine by Bellori. It gives a good overview of the different types of forces employed by the Romans from the Germanic auxiliaries (unarmoured spearmen, the auxilia palatina of the late Roman period) to the Meroitic foot archers (archers with arrows placed in the hair) and from the shieldless unarmoured horsemen of Constantine to the scale-armoured Praetorians of Maxentius. Note the resemblance of the latter to the Sarmatian cavalry in Trajan’s column. The only differences are that the horses are unarmoured and the helmet is the so-called pseudo-Attic helmet. Readers should however be aware that Bellori has taken some artistic liberties and has drawn most of the helmets inaccurately. For a more accurate picture, see the photos in the plates. Bellori’s drawing gives only a general image.
Roman elite horseman equipped with a ridge helmet and scale armour (source Arch of Galerius)
Soldiers wearing coifs in the Vatican Manuscript (drawn after Bishop & Coulston)
Aurelius Sudecentius, legionary of the Legio XI Claudia (first half of the fourth century AD)
Third-century tombstones drawn after Bishop and Coulston. Severus Acceptus, legio VIIIAugusta (Istanbul). The “armour” in this image is actually the so-called subarmalis, which was worn underneath the armour, but it could also be worn as the only “armour’' as in this case. The subarmalis or other ersatz annour could be made out of linen, felt or leather.
Third-century tombstones drawn after Bishop and Coulston. Julius Aufidius, legio AVI Claudia (Veria)
A sketch/drawing of a soldier depicted in a shield found at Dura Europos. The date is therefore before its capture in 253 or 256. Note that the Romans continued to use also this archaic style equipment at least until the seventh century when updated versions are depictcd in the so-called David Plates. However, there is quite secure evidence for the persistence of these styles much later in East Rome.
A sketch/drawing of a soldier depicted in the Synagogue mural depicting the battle of Eben-Ezer at Dura Europos. Dale before 256. Note the scale annour coif and the six-sided shield. These continued in use at least until the fourth century.
The above funerary relief shows a Roman auxiliary charging over the enemy footman. The relief is early for this period, but it show's nicely how the Romans taught their cavalry to run over the die enemy infantry formations like lawnmovers. If the morale of the enemy held and they w'ere in close order, this tactic was not necessarily successful because even when the horses were taught to do this, it was possible that the horse’s instincts prevailed and the horse refused to charge into the solid looking enemy formation. The source of the draw ing: Duruy.
A so-called ADLUCTIO coin of Probus depicts him addressing the soldiers. Source: Beger 1696. The important point is that it depicts the soldiers with the rectangular curved/ cylindrical scutum shield. This is of course a 17, h century drawing, but its depiction of equipment is still accurate. See e.g. the wildwinds coins website RIC 581/Cohen 19. The works of art continue to depict this type of shield at least until the turn of the sixth century and it is very likely that it was also used at least until then. One possible reason for the continued use of this type of shield would have been the idolisation of the gladiators, and vice versa it is possible to think that it fell out from use thanks to the ending of the gladiatorial games in the fifth century.
Constantine the Great’s Elite Cavalry
The drawing on the left depicts a horseman in the Arch of Constantine. What is striking about the cavalry images is that the Constantinian cavalry has no armour or shields. Some of them do not even have a helmet, like this one, but when they do, and most do, they have the pseudo-Attic helmet which I have depicted here separately. What is also notable about these images is that many of them show both hands to make it clear to the viewer that the horseman does not carry any shield. This suggests that these men did not carry shields so that they could demonstrate their manhood in combat. Does this mean that the cavalry in question would have consisted of Herul horsemen who did not use shields or armour in combat before they had proved their manhood by killing an enemy, or does this mean that Constantine’s Roman soldiers would have emulated them? Both are possible because the Arch also gives a prominent role for the Cornuti auxiliaries and Meroitic archers. If this is the case, then it would have been foreigners who decided the war in favour of Constantine. However, in light of the use of Pseudo-Attic helmets it is more likely that the soldiers are Roman elite cavalry who just demonstrated their superior fighting skill and spirit by not using armour and shields. Whatever the truth, the inclusion of these images implies in no uncertain terms that Constantine employed a cavalry of this type and it was this cavalry that had a decisive role in the battle of the Milvian Bridge in 312. In short, the image suggests that the cavalry of this type and the Meroitic foot archers, which are also in that image, both had a very significant role in this battle. Did the Meroitic infantry ride pillion to the bridge? The Greco-Roman military theory (e.g. Arrian, Techne Taktike 4.4, Modestus 18, Vegetius 3.16) and other period evidence (see e.g. the attached tombstone and Dura Europos painting on page 10) recognized a type of cavalry that did not carry shields so it is clear that the images in the Arch must represent reality and that Constantine employed this type of cavalry to great effect. Its main advantage was speed and my assumption is that Constantine unleashed these cavalry forces in pursuit of the defeated armoured enemy cavalry at the battle of the Milvian Bridge with the result that they were unable to flee. It is difficult to know with certainty what type of spear was employed by these men because the size of the reliefs dictated the length shown, but in light of the fact that the spears are used like javelins in the images the spears cannot have been longer than about 2.5 metres or at most 3.74 metres so that the latter would have already been a lanciarius or contarius. The continuity in the types of cavalry (Arrian, tombstone, Arch etc.) suggests that the same type of cavalry forces were also used in the latter half of the third century.
These Roman soldiers drawn after the mid-third century Ludovisi Sarchophagus depict some of the more unusual types of shields used by the Romans at this time. In addition to this, they continued to use the sexagonal and hexagonal shields.
A scene in the Arch of Constantine depicting the siege of Verona in 312. Bellori has drawn the helmets incorrectly, but the drawing still gives a good overall image of how the Roman soldiers were equipped.
A scene in the Arch of Constantine depicting the battle of the Milvian Bridge in 312. Note the horsemen without armour and shields with Pseudo-Attic helmets and the Meroitic archers.
Constantine’s army marching.
Roman military doctrine was very practical. It was based on the expectation that the generals possessed accurate intelligence of the enemy and its activities and that the logistical network provided adequate supplies for the Roman armed forces. This combination would then enable the commander to pick and choose the best time and place and how to engage the enemy.
The Romans also took safety precautions very seriously so that they employed layers of spies, scouts and patrols for the protection of their armies, and always built fortified marching camps. Combat doctrine also required that marching of the armies was done by the book so that soldiers were similarly protected by layers of spies, scouts and patrols. Typical marching formations were the hollow square/oblong array (baggage in the middle), epikampios opisthia (baggage behind or in the hollow), epikampios opisthia with the baggage train and rearguard following so that it could be formed into a hollow square/oblong, and the use of marching columns where the above were not possible. In addition to this, the Romans typically posted vanguards and rearguards that could be used both defensively and offensively. In certain conditions the vanguard could consist solely of cavalry so that it was used separately from its infantry support.
Roman combat doctrine had always been based on the combined arms concept, but from the reign of Alexander Severus onwards the Romans had depended increasingly on their cavalry to win their battles, to such an extent that we see Decius fighting only with cavalry against the Goths and Gallienus fighting mainly with his cavalry. At the time of the death of Gallienus, Roman combat doctrine was based on the use of cavalry when this was advantageous and on the use of the combined arms concept when this was advantageous, but in such a manner that the cavalry was clearly the favoured arms of service.3
The combined arms combat doctrine was based on the advantageous use of all arms of service, but in such a manner that the infantry formed its core. Land tactics were traditional so that the Romans could deploy their infantry in cohorts (one to four lines) or as phalanxes so that the battle line was usually divided into left, centre and right. Standard combat formations were the lateral phalanx with baggage train (with ballistae-carts) posted behind for its protection and the double phalanx (or two lines of cohorts) if the baggage train did not follow. Standard depths for these lines were 6 (4 ranks of heavy infantry and 2 of light infantry), 12 (8 heavy, 4 light), and 24 (16 heavy and 8 light), but in practice this could vary according to the quality of the unit, place and amount of light infantry, the number of men fit for service, and the type of unit (see Appendix 1). The cavalry was usually posted on the flanks and in reserve to protect the infantry, to outflank the enemy, or to pursue the defeated foe. The light infantry was placed where considered the most advantageous.
The Romans employed four standard ways to defeat the enemy: 1) ambushes and surprise attacks; 2) to break the enemy line with deeper formations or with a wedge array; 3) to outflank the enemy either on the left or right flank; 4) to outflank the enemy on both wings. To achieve these, the Romans employed several different grand tactical formations most of which are described by Vegetius (see the diagrams borrowed from my earlier books), but the other military treatises and narrative histories prove that the Romans used a variety of defensive arrays like the oblong and hollow square arrays that they could also use for offensive purposes. In addition to this, the Romans could use a double front (amphistomos phalanx, orbis) or double phalanx (diphalangia amphistomos, 4 duplex acies) if the enemy had managed to outflank the Romans.
Basic unit orders were: 1) the open order; 2) close order with shields placed rim-to-rim in width; 3) the defensive tortoise (testudo/foulkon) against cavalry (front rank kneeling, shields rim-to-boss in depth and almost rim-to-rim in width), which was used to bring an enemy cavalry attack to a halt with a wall-like appearance after which the front-rankers rose and attacked if the enemy cavalrymen had advanced close enough; 4) the offensive testudo (shields rim-to-boss in width and depth), which could also be used defensively by having all men kneel to receive enemy missiles, but with the expectation that the men would rise to their feet to fight.; 5) the siege testudo (rear rank kneeling to help the ‘mounting’ of the formation by other men, shields rim-to-boss in depth and almost rim-to-rim in width or alternatively rim-to-boss if the shields allowed this); 6) the irregular array (drungus/droungos).
simplified versions of the standard infantry battle formations not specifically mentioned among Vegetius's seven tactics
The attached diagrams provide an overview into the different tactical concepts and formations used when the Romans used the combined arms approach. In addition to these methods, the Romans could also employ their infantry units in ways that negated the advantages posed by the enemy (e.g. by opening up their array to allow cavalry wedges, elephants, chariots etc. to pass through) and/or they could employ specialist units like the mace/club-bearers against heavily armoured opponents like the cataphracts. A good example of the use of these tactics can be found in the narrative of the battle of Emesa in 272 and these are detailed there in greater depth. Appendix 1 also suggests a possible variation to the tactics discussed when the Romans deployed two different types of legion simultaneously.
The principal cavalry array used by the Romans was the so-called Italian drill formation of the Strategikon consisting of two combat lines each of which had three divisions with wing units in front. It had three basic versions: the largest standard cavalry array for armies in excess of 10–15,000 men had four divisions in the second line with rearguards as the fourth line; the medium sized array (15,000/12,000/10,000 and more than 5,000/6,000 men) had two divisions in the second line; the smallest (2,000–5,000/6,000 men or less) had only one division in the second line. The diagrams in the Strategikon enable us to reconstruct these in detail. It shows that the standard large cavalry array had about 31,000 horsemen, but on the basis of the other information in the treatise we know that it also allowed the use of even larger cavalry arrays. On the basis of later evidence, I have speculated that the extra men in these cases were posted as the third line. The accompanying diagrams are from my dissertation and article, which are based on the Strategikon. The Ks in the diagrams (from kaballarios) stand for ranks of regular soldiers in the array.
The diagram of the large cavalry array in the sixth century Strategikon (symbols show this array to have ca. 31–32,000 horsemen).
The 6,000–7,000 horsemen divisions of horse (mere, sing. meros) in their turn consisted of three moirai (‘regiments’, sing. moira) each with a maximum of 3,000 men. The standard way to form these moirai was to make the two flank moirai of each meros consist of the so-called koursores (runners/skirmishers). The units of these skirmishers used typically the highly mobile irregular droungos array when needed. The Romans placed between these skirmishers the moira of the defensores (defenders). The defensores were almost always deployed in the close-order so that they could protect the koursores when needed. The moirai were naturally divided into smaller units of 300–400 men (or 200–400 men; the Strategikon has two variants) which were variously called tagmata, arithmoi or banda. This means that the Strategikon did not envisage that the the cavalry units would have been close to their paper strengths (ca. 500–600 men), but I would not consider it impossible that the units could also have been deployed in full strength. The tagmata, arithmoi or banda in their turn consisted of smaller units of approximately 100 men, but in practice these seem to have been deployed side-by-side to form the above-mentioned units. However, there is still every reason to believe that other combat formations, unit structures and unit formations continued to exist side-by-side with these because the Romans had not yet completely standardized the equipment, training and fighting styles of their cavalry units of different ethnic origins.
The reconstructed large cavalry battle array with all its parts.
The reconstructed super large battle array for cavalry armies in excess of ca. 50,000 men.
It is very likely that the Romans continued to use the tactical formations described by Arrian in his Taktika and Ektaxis kata Alanôn because they continued to use the same types of cavalry units with the same types of infantry formations. In the latter case Arrian describes a cavalry formation which he used for the pursuit of the Alans. It is probable that this same type of array was also used when the cavalry was deployed separately in front of the infantry. The Armenian symmachiarii formed the flanks of the front line. These Armenian units consisted of katafraktoi (cataphracts possibly used as a front rank), kontoforoi (contarii), and of the hippotoxotai (mounted archers). Between these two wings were six lochoi of auxiliary alae. These consisted of the hippotoxotai (mounted archers) who were also equipped with javelins, swords and shields. The two wing reserves on both flanks consisted of the legionary cavalry. The legionary cavalry in its turn consisted of units of lancers (kontoforoi, machairoforoi) and of units of javelineers (logchoforoi, machairaforoi, pelekoforoi), or alternatively these units included both types within the same units so that the kontoforoi formed the two front ranks and the lonchoforoi the next two in a formation of four ranks. In addition to this it is probable that the commander followed the cavalry line in the middle with his equites singulares (elite auxiliaries), and it is probable that Aurelian deployed his cavalry in this manner at the battle of Emesa in 272. See later.
The unit orders of the Roman cavalry varied according to the type of unit and situation. Standard unit formations were the square array (usually four ranks and eight files, or units with depths of five or ten ranks as Arrian recommends and which was also to become the standard during the so-called Byzantine times) and oblong arrays of various sizes up to ca. 512 men. The horsemen in these could be deployed in open order for marching or in close order for combat. In addition to this, the Romans used the wedge array (64 men or the massive regimental version from ca. 300 to 1,500 horsemen), rhombus (128 men) and irregular order called drungus/droungos (varied in size).
The Romans divided their navy into two basic categories: 1) The Praetorian Fleets of Italy; 2) The Provincial Fleet in the provinces. The two Praetorian Fleets, Classis Praetoriae Misenatium/Misenatis and Classis Praetoriae Ravennatum/Ravennatis/Ravennas, formed the core of the naval forces and served as reserves in Italy. The Provincial Fleets included at least the Classis Alexandrina, Classis Syriaca, Classis Nova Libyca, Classis Germanica, Classis Pannonica, Classis Moesica, Classis Britannica, Classis Pontica, Classis Mauretanica (thirteen liburnae), Classis Nova Libyca, and Classis Africana.6 According to Vegetius (4.31), each of the praetorian fleets had a single naval legion attached to them. We do not possess similar information regarding the provincial fleets, but it is still clear that the provincial fleets had some sort of fighting contingents attached. Some of the provincial fleets were in the hands of the usurpers in 268. Besides their use in naval combat the Romans also used their fleets to support military operations, to protect the sea lanes, and to ship both men and supplies where needed.
The accompanying map shows the naval deployment pattern at the first half of the third century. The situation after this is not known with certainty. On the basis of the narrative sources, it is possible but by no means certain that Gallienus may have created a separate fleet for the Aegean Sea.
The coastal defence system employed by the Romans consisted of two separate systems: 1) the passive/defensive measures (forts, towers, fortified cities and towns along the coasts and rivers); 2) the active/offensive measures undertaken by the fleets. The civilians were responsible for the basic defence (control of harbours, collection of tolls/ taxes, prevention of smuggling and wrecking, guarding of the coasts against piratical attacks), but the fleets performed in most cases the active/offensive defence with the possible help of corvéed civilian ships. The Romans could appoint special officers to take control of larger sections of the coast. Civilians were also required to contribute ships and men whenever needed.7
Ships
The fast liburna (a fast galley, pl. liburnae) was the principal type of warship used by all fleets so that the triremes served as flagships of the provincial fleets while the fives or sixes served as flagships for the praetorian fleets. In addition to this, the praetorian fleets had significant numbers of larger ships (mostly triremes with fives and sixes serving as flagships) to secure them advantage over the provincial fleets in case of need. The fleets also possessed special types of ships, boats and rafts to enable them to transport horses, men, equipment and supplies, the numbers of which could be bolstered with corvéed civilian transports or warships.
Naval Tactics8
It was always considered risky to row or sail during the stormy winter season, which meant that this was usually avoided. However, the sources prove that the Romans sometimes took the risk to surprise their enemies. The Roman navy expected that its personnel would always possess good knowledge of the winds, tides, locale and signs of weather to enable it to fight in the most advantageous manner against enemies. Combat tactics on the seas mirrored those on land. At the heart of plans was the obtaining of information of enemy plans through spies and scouting boats, and then surprise the enemy with an ambush or surprise attack.
Naval battle formations used by the Romans were: a) the crescent formation with the best ships and crews posted on the wings to encircle the enemy; b) the convex array in which the best ships and crews were placed in the middle to break apart the enemy formation; c) the double convex array when there were enough ships to provide reserves; d) the line abreast with reserves behind; e) the double line abreast if there existed enough ships; f) the defensive circle. Individual ships and groups of ships were trained to perform:
1) the diekplous manoeuvre to penetrate into the enemy formation so that one galley ran forward into the gap between two enemy galleys where it sheared the oars of one galley after which the second galley followed it and rammed the immobile defenceless ship;
2) the outflanking of the enemy line with the periplous manoeuvre so that the outermost galleys extended their line to encircle the enemy line.
A naval battle consisted of several stages. When the Romans possessed larger ships, they started to bombard the enemy with the ballistae and bows at a range of about 300–600 metres. Next the larger ships started to employ the onagri (stone throwers), slings and staff-slings from about 150 metres. The stone throwers used either stones and/ or fire-bombs and the ballistae primarily incendiary arrows. When the Romans reached within about 30–40 metres they started to use light javelins, and then at 20–30 metres they moved on to heavy javelins. At a ramming distance, the galleys either rammed, sheared the enemy’s oars, or grappled the enemy vessel for boarding.
The principal foreign naval threats during this period were the Franks, Saxons, Borani, Goths, Heruls, Blemmyes, and Aksumites. Their ships were no match for the Roman ships, but when the Romans were hampered by civil wars they could still invade and raid with relative impunity.
The various methods employed by the Romans in siege warfare were the most advanced of the age. They can be divided into defensive and offensive techniques.
Offensive sieges had some standard features, which were: 1) terms of surrender offered to the enemy at first to avoid the costly siege; 2) use of surprise attacks if possible; 3) use of a traitor; 4) if the enemy had refused to surrender, assault using testudo formation with ladders, possibly with some protective sheds; 5) if the first assault failed, siege engines might be used, mounds built, and mines excavated; 6) if the place was considered too costly to take by other means, the defenders would be starved into surrender.
Standard defensive siege techniques were as follows: 1) scorched-earth policy to make the attack difficult if there was prior information of the enemy invasion; 2) the building of sophisticated fortifications with enough provisions placed inside to withstand a siege; 3) exploitation of defensive features like walls and towers; 4) use of siege engines; 5) the sending of a relief army against the besiegers; 6) use of diversionary invasion; 7) guerrilla warfare against the besiegers; 8) if all else failed, then the Romans offered terms of surrender.
The Romans were the beneficiaries of the revolution that had taken place in siege warfare during the Hellenistic age, but they had not merely copied the equipment they had inherited, they had improved it. The siege equipment employed by the Romans included the sambuca (a hollow tube to land men on top of a wall), various kinds of siege towers, various kinds of battering rams (from simple rams all the way up to the city-taker ‘helepolis’), borers, drills, flails, the ‘fire hose’ to spread liquids, mining equipment, various types of sheds, fire bombs, cranes, and various types of artillery (small to large ballistae, steel-spring powered artillery, repeating ballista, onager and possibly also mangonels/ trebuchets if the super heavy stones used at the siege of Cremna are evidence for that). A sample of these is illustrated here.
The primary concern for all emperors was their personal safety and this dictated their approach to intelligence gathering. Most effort was directed towards potential domestic foes, but obviously they did not forget external threats. The imperial bodyguard units ensured the personal safety of the emperor and were also the principal means of obtaining intelligence of internal and external threats. The principal intelligence gathering units were the frumentarii and peregrini, but other units of bodyguards or troops posted in the city of Rome were also used as secret services. The other principal units in charge of intelligence gathering were the speculatores, protectores/scholae/aulici, evocati Augusti, praetoriani, equites singlares Augusti and the urbaniciani. With the exception of protectores, who were commanded by their comes or magister, and the urbaniciani, who were commanded by their prefect, the other units were all under one or two praetorian prefects. These units combined the functions of bodyguard with intelligence gathering and other special operations which included assassinations, police operations and interrogation. However, we should not forget that in practice emperors could use anyone they trusted (for example eunuchs) as their special operatives. Even if these were the principal intelligence gathering organizations, we should not forget that the contribution of regular military units was also important, because they were also given internal and external security missions. In fact it was the frontier units that gathered most of the intelligence from across the border. In addition to this, the Romans used civilian paramilitary police forces and other civilians for the gathering of intelligence, which reported important findings to local military commanders and/or the governor of the province.9
Thanks to the limitations posed by speed of travel (feet, horses, wagons, ships) the general peacetime deployment pattern of Roman forces was defensive. The Romans simply could not transfer their forces fast from one front to another so they had to post them close to the border. These forces were then used to interfere in politics across the border so that enemies could be kept divided. It was because of this that the forces posted near the frontier conducted active intelligence gathering operations across the border. However, there were occasions in which intelligence gathering operations, diplomacy and the forces posted on the frontier proved insufficient to meet an enemy invasion. In such situations, the Romans typically sent detachments from other legions and units (or entire legions and auxiliary units) to the threatened sector. This naturally weakened the areas from which these units had been taken, and enemies opposite those sectors often exploited this and invaded. When this happened, emperors were often required to raise new legions and other units to meet the crisis.
The above-mentioned peacetime deployment pattern was changed when the emperors conducted offensive wars or when there had been a prolonged crisis like there was after ca. 251. In these cases the emperor was accompanied by the only significant field army of the Empire, which was the case for most of the time from the reign of Gallienus onwards. He created a personal field army, which formed the core of the army which also accompanied Aurelian and Probus in their campaigns. The keeping of this major field army in the field resulted in the creation of an elite fighting force, which gave the emperor the advantage over all usurpers and enemies. The withdrawal of these forces in its turn weakened the defences along the frontiers even further. However, this was still safer than giving generals such large forces that they could try to usurp power themselves. Regardless, as we shall see, the emperors were still forced to do this, and sure enough the generals did usurp power. It was therefore possible to secure only that sector of the Empire or frontier where the emperor was present in person with his personal field army. The principal threats to the Roman Empire were not the foreign enemies but the power-hungry generals. Foreign enemies could threaten the Roman Empire only when the Romans were fighting civil wars.
The combination of civil wars with foreign invasions had caused a change in Roman strategy, which was the adoption of the defence-in-depth. It has often been claimed that Gallienus was the man who invented it, but this is not true because there are instances of this strategy before him. However, it was Gallienus and Valerian who were forced to adopt this on a greater scale than ever before. They fortified and strengthened several cities in Gaul, Italy, the Balkans and Asia Minor, both for defensive and offensive purposes. Similarly, it has been claimed that Gallienus created the first real cavalry army to enable him to engage the many foreign invaders fast, but this is also erroneus because the Romans had used large cavalry armies well before him. However, it was once again Gallienus who made use of cavalry forces more than any of his predecessors with the exception of Decius.10
As we shall see, the stabilization of the situation under Aurelian and Probus enabled the latter to attempt to return to the strategy of preclusive security along the borders. It was because of this that Probus built or rebuilt fortifications along the borders and posted forces along those frontiers. However, there was still one major difference in this system compared to that which had existed before the 250s: Probus was always accompanied by a huge field army.