Late Roman Unit Sizes

[Flavivs Aetivs]

Yeah it's him.

This is my current passage on 5th Century Roman Unit Sizes. In it I suggest an organization of units of 1500, with Iuniores and Seniores being 3000 strong.

The size of the late Roman army, and of the late Roman legion itself, is extremely controversial and relies on a few scraps of evidence. Claudian and Orosius mention a force of 2 legions and 5 Auxilia palatina numbering 5000 men, Zozimus mentions 5 Dalmatian Tagmatae numbering 6000 and an Ile numbering 600, and Sozomen mentions 6 Arithmoi of 4000 men. Ammianus mentions seven Legions and Numeri numbering less than 20,000 at Amida, and two Alae totaling 700 men, as well as several references to detachments. Finally are the controversial Panopolis Papyrii, which make no mention of unit sizes themselves, but by utilizing Byzantine sources Jones and Duncan-Jones extrapolated potential unit sizes, including approximately 1100 men for Legio II Traiana, 1800 for Legio III Diocletiana, 160 for the Equites promoti, 120 for a unit of Equites sagittarii, and 24 for a unit of Dromedarii. Another papyrus from Justinan’s time records an Arithmos of 508 men as well.

Anything that is seemingly simple about these numbers is not, especially when cross-referenced with the Notitia. The first thing about this evidence is to look at the records of Arithmoi and Numerii, which are the same kind of unit. The six Arithmoi of Sozomen number 666 each, and of Justinian numbers 508, which would show a potential strength of the Numerus at 600 men. This is fascinating, as the Strategicon makes mention of Numeri numbering 300 men. Many Auxilia Palatina units, which are recorded as Numeri or Cohortes, were often divided between Seniores and Iuniores and put at separate posts: by reconciling the Strategicon with both Sozomen and the Papyrus, one gets a “complete” Numerus of 600 men, which is comprised of an Iuniores unit of 300 and a Seniores unit of 300. This would also solve some of the problems with the Ducenarius, which will be discussed later.

Using this system, it provides some clarity to Claudian’s record of the 2 Legions and 5 Auxilia palatina units that were led by Macezel into Africa in 398. If the Palatina units, which can be equated with the Augustei, Sagittarii nervi, the Felices iuniores or seniores, the Invicti seniores, and the Leones iuniores or seniores, are assigned numbers of 600, 600, 300, 300, and 300 respectively, then that would leave another 2900 men for the two Legions to fill, making them number about 1450 men each. Zozimus’ record of the 5 units from Dalmatia would suggest 5 legions of 1200, but actually the Dalmatian Comitatenses consists of 4 Legions and one Numerus, suggesting a Numerus of 300 and four Legions of 1425. These both closely coincide with John Lydus’ units of Mattiarii and Lanciarii serving under Julian in 363, which each supposedly numbered 1500 men. Certainly Ammianus’ mention of detachment sizes of 300, 400, 500, and 750 men would make sense with a larger Legion. This presents the fascinating possibility of a 1500 man Legion, and possibly a 3000 man Legion if one takes into account the division of some units into Iuniores and Seniores. In fact, this would also allow the combined forces of the old Legio Septima Gemina, still in operation in Spain, Gaul, and Anatolia, to number a total of 4500 men!

Finally are the references to cavalry. One of the most fascinating things to notice about the Notitia, is that the newest cavalry units are called Cunei, which seems like an adoption of a Hunnic cavalry structure. Ammianus’ mention of 700 men in two Alae, and Zozimus’ Ile or Ala of 600 would seem to coincide with possibly both qunigenary and milliary strength Alae. Units listed as Iuniores and Seniores may have numbered 350 each to total 700, while there were also quingenary Ala of 500 or so. Another record is by Synesius, who mentions Unnigardae, which will be detailed later. He records them operating in units of 40, and 200, all under the same leader which would suggest a Hunnic Cuneus could have numbered 240 men; providing the Romans adopted the organization of their Cuneus from the Huns, this could suggest an organization of the Roman style Cuneus. The evidence for this, however, is circumstantial: there could be a number of other reasons for why Cunei are almost exclusive to the Middle and Lower Danubian Limes.

Thoughts? I used sources and discussion from this thread (I haven't included it in the text yet but I wrote them all down.)