Command structures of the Roman auxilia in the early Empire

Roman auxiliary infantry crossing a river, from Hadrian's Wall.

I’m currently studying a FutureLearn course on Hadrian’s Wall, brushing up my knowledge in hopeful anticipation of being able to teach Roman Britain again in the Fall semester. This week, however, I haven’t got much done, as I have fallen down a rabbit hole of the Roman auxilia, the auxiliary troops that sat beside the legions in the Roman military. The following essay is the result of that. I should make clear that, though I have taught the Roman army as part of various courses for twenty years now, and read quite a bit on the subject. I do not consider myself an expert in this field, so all of this has to be provisional, pending the arrival of the better-informed.

There are two misconceptions about the auxilia that I have come across over the years. The first is easily dealt with; that the auxiliaries were primarily a light-armed force, providing troops that the legions did not. Whilst the Auxilia undoubtedly included light-armed troops and skirmishers, as well as cavalry and specialist units such as archers, many, if not the majority of auxiliary infantry units, such as the Tungrians and Batavians that won the battle of Mons Graupius for the Romans,[1] fought as close-order heavy infantry.[2] They were equipped in a similar fashion to legionaries, though with some differences, such as the use of the oval, flat clipeus as a shield instead of the semi-cylindrical scutum.[3] This would limit their ability to carry out certain manoeuvres, such as the famous testudo (tortoise) formation. The use of the thrusting hasta (spear) rather than the thrown pilum (javelin) would also impose slightly different tactics. At the siege of Placentia in 69 CE (a prelude to the first Battle of Cremona), Tacitus talks about the ‘legions in a dense line, the auxiliaries spread out’.[4] Again, the use of the clipeus rather than the scutum might prevent the auxiliaries fighting in as close order as the legions, being less able to interlock their shields. But Tacitus is only talking about one battle here, not making a general statement.

It is also possible that the auxiliaries were less intensively trained than the legions, though the evidence here is unclear.[5] It is certain that, as (predominantly, at least until the mid-second century CE) non-citizens, the auxiliaries had less status that the legions.

The other idea that persists in a lot of modern sources is the assumption that somehow the auxiliary units were administratively part of the legions. This idea is particularly embraced by writers of fiction, and is expressed in designations such as ‘Fourth Gaulish Auxiliaries of the Second Legion’ (in Rosemary Sutcliff’s Eagle of the Ninth).[6] I know of no inscription in which an individual auxiliary unit is described in these terms.

There are, however, a few bits of evidence that might suggest links with the legions. Simon James in Rome & the Sword cites two passages of Tacitus.[7] The first comes from Annals 13.35, where Tacitus talks of the movement of a legion from Germany to Armenia with (in James’ translation) ‘its alae of cavalry and cohortes of foot.’[8] Unfortunately, Tacitus’ text is a little more equivocal here than James’ translation makes it. The Latin is ex Germania legio cum equitibus alariis et peditatu cohortium. There is no possessive pronoun in there. This doesn’t make James’ translation wrong – Latin doesn’t use possessive pronouns as much as English, and often prefers to indicate possession through implication. But the ambiguity means that this passage could be translated as ‘a legion from Germany, along with cavalry alae and infantry cohorts’, and nothing is explicit is being said about a link between the legion and the auxiliaries, other than that they have all come from Germany. This makes it difficult to base an argument on this excerpt.

James’ second passage, Tacitus Histories 1.59, is more significant. Here Tacitus speaks of ‘eight Batavian cohorts, the auxilia of XIV Legion, but which had, due to the strife of the time, separated from the legion.’[9] Additionally, in Gaius Caligula’s German campaign of 40 CE, legates (presumably of the legions) had responsibility for summoning auxiliary units.[10] There is also epigraphic evidence in which the formula legio … et auxilia eius appears, ‘x Legion and its auxilia’.[11] To this may be added inscriptions showing legionary centurions in (acting) command of auxiliary units.[12]

It must be noted that this sort of evidence is rare, and on that basis it seems safest to conclude with the likes of George Cheesman, Adrian Goldsworthy and Ian Haynes, that the auxilia were in general independent units.[13] The evidence from Tacitus and inscriptions would suggest that at times some auxiliary units were attached to legions, but it seems better to assume with Cheesman that these were ad hoc and impermanent arrangements, rather than the regular system.[14]

But this raises another question, one which isn’t as thoroughly discussed in the literature as might be hoped – to whom were the auxiliary units responsible? Obviously, those units specifically attached to a legion would be commanded by the legate of that legion. But more generally on campaign, the auxiliaries went alongside the legions and under the same commander, rather than be directed by legionary legates. In Britain in 55 BCE Julius Caesar seems to have treated his legates and auxiliary prefects on an equal footing.[15] At the battle of Mons Graupius, Agricola seems to have directed his auxiliary units without reference to any intermediate command structure.[16]

But what of peacetime? To whom did the prefect of a fort on Hadrian’s Wall report? Who decided it was time for I Tungrians to move on from Vindolanda and be replaced by IX Batavians? It has to be said that there is not a great deal of evidence. There are inscriptions that indicate auxiliary units could be brigaded together under a single officer.[17] But these arrangements may, like the attachments to legions, have been uncommon and temporary. It has been suggested that Hadrian’s Wall was controlled from the fort at Stanwix, on the grounds that this was the largest fort on the Wall, and housed the most prestigious auxiliary unit on the wall, the 1,000-strong Ala Petriana, the only ala millaria (‘thousand-strong ala’) in the whole of the province.[18] It is certainly true that the commander of that unit would be the most prestigious commander of an auxiliary unit on the Wall – indeed, in the province. But, as Breeze and Dobson point out, it doesn’t necessarily follow that he had overall command in the area. In any case, that only defers the question – to whom did the brigade commanders report?

Ultimately, of course, the auxiliary commanders were responsible to the governor in the province, who commanded all troops. But was there any intermediate level of command? In a modern army, we would expect so. As a consequence it has been argued, for instance by Laurence Keppie, that even if the auxiliary units were not formally attached to the administrative structure of the legions, nonetheless the legions had a degree of control over the auxiliary units around them.[19] 

The problem is, as Breeze and Dobson point out, there’s precious little evidence for legionary legates exercising such authority. If anything, the evidence is the other way, suggesting that auxiliary commanders communicated with provincial governors directly. A papyrus dated to 103 CE features the governor of Egypt writing directly to the commander of an auxiliary cohort.[20] One might also note an inscription from Rough Castle on the Antonine Wall in Scotland that reveals a legionary centurion in (presumably temporary) command of an auxiliary cohort.[21] The centurion has not come from the closest legion, VI Victrix at York, but from XX Legion at Chester.

In one of the Vindolanda tablets, Flavius Cerialis, commander of IX Batavians, writes to Crispinus to ask the latter to put in a good word for Cerialis with Marcellus, the provincial governor.[22] Unfortunately, apart from his being in a position to do Cerialis a favour, we know nothing more about Crispinus, or the position he held.

It seems likely to me that the command structure of an army on campaign, where the legions and auxiliaries were all independently responsible to the overall army commander, was replicated in peacetime, with the auxiliary prefects directly responsible to the provincial governor, in his role as commander of the army of the province.[23] The governor’s staff must have been considered sufficient for the bureaucratic and administrative needs of the auxiliary units in the province.

This does not mean that the legate of a legion could not order the commander of an auxiliary unit around. The legionary legate was a senator, of a higher social rank to the equestrians from which the auxiliary commanders were drawn, and appointed directly by the emperor.[24] It would be a brave auxiliary prefect who would refuse without very good reason a legionary legate’s reasonable request, e.g. to send a vexillation of troops. But this authority of the legionary legates over the prefects was rather more informal than it has sometimes been conceived as.

Of course, everything changed with the army and provincial reforms of the third and fourth centuries, but that is beyond the scope of this essay.

Works cited

Breeze, David J., and Dobson, Brian (2000), Hadrian’s Wall, 4^th edn, London: Penguin

Cheesman, George Leonard (1914), The Auxilia of the Roman Imperial Army, Oxford: Clarendon Press (pagination from 2018 edition, Frankfurt am Main: Outlook)

D’Amato, Raffaele (2012), Roman Centurions 31 BC–AD 500, Oxford: Osprey Publishing

Dobson, Brian, and Mann, J.C. (1973), ‘The Roman Army in Britain and Britons in the Roman Army’, Britannia 4, 191–205

Forty, Simon (2018), Hadrian’s Wall: From Construction to World Heritage Site. A Journey Along the Wall, and Back in Time, Yeovil: Haynes.

Goldsworthy, Adrian (2000 [2019]), Roman Warfare, London: Cassell & Co. (pagination from 2019 edn, New York: Basic Books)

Goldsworthy, Adrian (2003), The Complete Roman Army, London: Thames & Hudson

Haynes, Ian (2013), Blood of the Provinces: The Roman Auxilia and the Making of Provincial Society from Augustus to the Severans, Oxford: Oxford University Press

James, Simon (2011), Rome & the Sword: How Warriors & Weapons Shaped Roman History, London: Thames & Hudson

Keppie, Lawrence (1983), The Making of the Roman Army: From Republic to Empire, London: Batsford.

Sutcliff, Rosemary (1954 [1977]), The Eagle of the Ninth, Oxford: Oxford University Press (pagination from 1977 edn, London: Puffin)

---

My thanks to Eric Morse for sharing some thoughts with me on this topic.

[1] Tacitus, Agricola 36.1–2.

[2] Goldsworthy 2000 [2019], 113; Haynes 2013, 274.

[3] On auxiliary equipment, see Goldsworthy 2003, 136. I am oversimplifying here – as Haynes 2013, 239–249 argues, there would have been considerable variety in terms of equipment from auxiliary unit to auxiliary unit.

[4] Tacitus, Histories 2.22: densum legionum agmen, spara auxiliorum.

[5] The suggestion is made in Dobson and Mann 1973, 195.

[6] Sutcliff 1954 [1977], 17 (ch. 1: ‘Frontier Fort’).

[7] James 2011, 303 n. 32.

[8] Cavalry units in the auxilia were termed alae (singular ala), which can be translated as ‘wings’, though it’s more common to retain the Latin term. Infantry units were cohorts (singular cohors), usually anglicized as ‘cohorts’.

[9] octo Batavorum cohortes, quartae decimae legionis auxilia, tum discordia temporum a legione digressae.

[10] Suetonius, Life of Caligula 44.

[11] Cheesman 1914 [2018], 40 n. 146, cites three examples – CIL [Corpus Inscriptionum Latinarum] 8.2637: legio III Augusta et auxilia eius, Lambaesis, 158 CE: CIL 13.8017: legio I Minervia pia fidelis Severiana Alexandriana cum auxilis, Bonn; CIL 3.3228: vexillationes legionum Germaniciarum et Brittanniciarum cum auxilis earum.

[12] E.g. CIL 3.1918, a centurion of III Legion in command of I Belgae; RIB [Roman Inscriptions of Britain] 2144, a centurion of XX Legion in command of VI Nervians at Rough Castle (to be discussed later). For further examples see D’Amato 2012, 6.

[13] Cheesman 1914 [2018], 29-31; Goldsworthy 2000 [2019], 112; Haynes 2013, 43.

[14] Cheesman 1914 [2018], 29, 30.

[15] Julius Caesar, Gallic Wars 4.22.

[16] Tacitus, Agricola 35–36. See Haynes 2013, 272, on issues with Tacitus’ account of Mons Graupius, which may affect the reliance we can place on it.

[17] E.g. CIL 11.6344, four cohorts under a single officer in Spain.

[18] E.g. Forty 2018, 114.

[19] Keppie 1983, 190.

[20] POxy 7.1022.

[21] RIB 2144.

[22] Vindolanda Tablet 225.

[23] This appears to be the conclusion of Haynes 2013, 325, though he is not explicit.

[24] Goldsworthy 2003, 60, 64.