by Mary Harrsch © 2026
I was listening to The Ancients podcast on Amazon Music and Dr. Rob Collins, Newcastle University, stated that not all the Roman armies left with Constantine III in 405-407 CE as evidence by a lack of coinage with his image beyond southern England. The traditional narrative holds that Constantine III took all of the mobile troops from Britain when he crossed to Gaul to confront Germanic invaders who had crossed the Rhine. However, I learned this is increasingly seen as an oversimplification. |
|
A late Roman frontier officer in scale armor at a northern British fort, c. 5th century CE. The ridge helmet, mail armor, and absence of stirrups reflect equipment of the late Imperial period. AI-generated illustration created with Adobe Firefly. I had to change the helmet type, the armor type then remove stirrups, the horse's breast band and change the soldier's caligae to the enclosed boots more prevalent among frontier troops during that period to improve historical accuracy. |
The distribution of Constantine III's coins is indeed geographically telling. The majority of late Roman silver siliquae are found in southern Britain, in the lowland zone to the south and east of the Fosse Way, with finds in Devon and Cornwall, Wales, the west Midlands, and the north-west and north-east being rare. Notable examples of later issues of Constantine III come specifically from Richborough and the Patching hoard in West Sussex. This southern concentration suggests that his monetary authority — and by extension his military presence — was felt primarily in the south, consistent with the theory that troops further north were not his to take.
Archaeological evidence supports the idea that not all Roman military presence ended in 407. The British-based units of the Field Army eventually left and are evidenced on the Continent, effectively "hollowing out" the province — but the northern frontier units were left largely untouched, as were tribal militia and foederati. The northern garrisons remained in place with their units and Roman commanders, eventually becoming the core of warbands in the 5th century.
This is further supported by earlier precedent: coins dated later than 383 have been excavated along Hadrian's Wall, suggesting that troops were not stripped from it during Magnus Maximus's earlier usurpation, or if they were, they were quickly returned.
The crucial nuance that Dr. Cottrell appears to be drawing on is the difference between the comitatenses — the mobile field army — and the limitanei, the static frontier garrison troops. What Constantine took was the Field Army led by the Comes Britanniarum, while the Saxon Shore was stripped to a skeleton staff and the northern frontier units were left largely in place. These latter troops, cut off from imperial pay after 409/410, didn't simply vanish — they evolved into the post-Roman military structures of sub-Roman Britain, which some scholars connect to the legendary traditions surrounding figures like Arthur.
So Dr. Cottrell's coinage argument is part of a broader and well-supported revisionist picture: the end of Roman Britain was not a single dramatic event in 407 but a gradual, geographically uneven process of disintegration.
The Physical Evidence: Birdoswald and the Transformation of the Forts
The most compelling archaeological case study is Birdoswald on Hadrian's Wall. At Birdoswald, the only change in the early 5th century appears to have been that the troops of the fort were no longer paid or supplied by central authority — the unit was still there. The old system of official coercion may have been replaced by a symbiosis, whereby the territory from which supplies had been drawn as part of the Roman tax system continued to sustain the fort in return for the assurance of protection in troubled times. The kind of commander-patronus attested by the large commanders' houses in the late forts may have continued to be an important figure as the 5th century went on — men who became imperceptibly more like chieftains in control of warbands than Roman commanders.
This interpretation is supported by excavation. Recent excavations at Birdoswald, led by director Tony Wilmott, suggest that a warband descended from the later Roman garrison may have continued to occupy the fort, supported by the discovery of a large timber hall which may have served as a gathering place for the post-Roman community. The timber hall is significant — it is a classic marker of early medieval chieftainly power, suggesting a direct architectural evolution from Roman fort headquarters to early medieval great hall.
Unlike the elite mobile field army, many of the static frontier troops — the limitanei — with their local loyalties didn't leave after direct Roman rule ended, and there is evidence that several forts were occupied for centuries after.
The transformation from Roman garrison commander to post-Roman warband leader was apparently gradual and organic rather than a sudden rupture. There is no evidence of any kind that the infantry of British leaders could be classed as "regular" after the middle or third quarter of the 5th century at the very latest. In other words, within a generation or two of 407, the disciplined Roman military structure had given way to something more characteristic of early medieval warfare — smaller, more personal warbands loyal to an individual leader rather than to an imperial institution.
The most tantalizing figure in this transition is Coel Hen — "Coel the Old" — who appears to straddle the Roman and post-Roman worlds. The senior military commander in the northern part of Roman Britain in the late Empire was the Dux Britanniarum and Prefect of the Sixth Legion, probably based at York, who commanded the garrisons of the forts on Hadrian's Wall. Some scholars have proposed that Coel Hen held, or inherited, precisely this role.
Historian John Morris in The Age of Arthur suggested that Coel may have been the last of the Roman Duces Britanniarum who commanded the Roman army in northern Britain, and split his lands among his heirs after his death. It must be noted, however, that Morris's broader thesis has been widely criticized by scholars, and Coel Hen himself is a semi-legendary figure whose historicity cannot be fully verified.
What is less disputed is that Coel Hen appears in the Harleian genealogies and the later pedigrees known as the Bonedd Gwลทr y Gogledd (The Descent of the Men of the North) at the head of several post-Roman royal families of the Hen Ogledd. His line, the Coeling, included such noted figures as Urien, king of Rheged; Gwallog, perhaps king of Elmet; the brothers Gwrgi and Peredur; and Clydno Eiddin, king of Edinburgh.
These figures represent the political legacy of whatever military structures survived in the north after 407. The Hen Ogledd — the "Old North" — covered modern northern England and southern Scotland, encompassing the kingdoms of Rheged, Elmet, Bryneich, the Gododdin territory extending well into Scotland, and other kingdoms in the Pennines. Rheged, one of the most celebrated of these kingdoms, was a Brittonic-speaking realm of the post-Roman era whose capital was likely Carlisle — itself a major Roman administrative and military centre.
The post-Roman kingdoms of the north map almost precisely onto the Roman military and administrative geography of the region, strongly suggesting that their rulers were indeed the descendants — biological or institutional — of the Roman garrison commanders who simply never left.
It is worth noting that this entire world of post-Roman northern warbands forms one of the primary historical contexts proposed for the Arthurian legends. The warrior culture of the Hen Ogledd, with its Roman military heritage, its Brittonic language, and its desperate resistance to Anglo-Saxon encroachment, is precisely the milieu that produced poems like Y Gododdin — one of the earliest texts to mention Arthur by name — and the bardic celebration of warrior heroes that would eventually crystallize into the Arthurian tradition.
