Roman commanders Dark Age kings
March 10th, 2002
When Roman rule ended in Britain, military units on Hadrian's Wall
seized control of local areas for themselves, writes Tony Wilmott
How did the late 4th century Roman diocese of Britannia evolve into the patchwork of kingdoms in Britain in the 6th century? Evidence for the
period - traditionally, the start of the 'Dark Ages' - is sparse. Yet archaeologists are beginning to see a few signposts in the
From being a relatively uniform Roman province, Britain began to
and change in different ways in different areas. In the South-East, for
example, there seems to have been fairly rapid Anglo-Saxon takeover. In the West, where Iron Age hillforts were reoccupied as strongholds and
settlements, Anglo-Saxon culture was slower to take hold. Much of this is as we would expect.
One of the more interesting areas to look at is the military zone of the North - Hadrian's Wall and its surrounding area. Here, archaeological
work over recent years has brought a dramatic shift in our understanding
of how society changed in the late Roman and post-Roman periods.
The principal sites to have received archaeological attention are the
forts and their associated civilian settlements or vici. Fifteen years ago it was remarked how seldom evidence for later occupation had been
found in forts. Archaeologists assumed that the forts were completely abandoned and that - in the words of Breeze and Dobson's classic
study, Hadrian's Wall - 'the army of the Wall melted back into the soil from
which it had sprung'.
That view has now been overturned by recent excavations at places such as
Birdoswald, South Shields, Vindolanda, Wallsend, Housesteads and
The static frontier troops of the late Roman army, the so-called limitanei, were supplied and paid, either in cash or in
kind, by centralised authority. Recent excavations have supported this view with
4th century barracks at Vindolanda, Wallsend and South Shields recalling those of the 2nd century, if for a smaller unit strength.
The commanding officer's quarters at South Shields, Binchester, and possibly Lanchester, Housesteads and Vindolanda are larger and more
elaborate in the 4th century than in earlier times. These residences may reflect an extension of the authority of unit commanders into civil as
well as military spheres. They may, for example, have become civil magistrates in addition to military commanders, as their contemporaries
did in provinces such as Egypt.
The principal role of these officers was to raise men and supplies, as part of the regime of local taxation - indeed, the financial and social
burden of the army's presence may partly explain why Romanisation did
not take hold in the North as elsewhere in Britain. It is likely that the army provided the only cash employment in the area, in both army
service and in contracts for supplies. So what happened when the imperial system which sustained the army ceased to exist in Britain?
Some interesting evidence exists in the structural remains at forts from
the very end of the Roman period. A number of buildings, at Vindolanda, South Shields, Housesteads and Birdoswald, have been claimed as
Christian churches, largely on account of their having semi-circular apses. Vindolanda provides the firmest evidence, where in addition to
the 'church' a number of Christian-looking artefacts have been found,
including a stone slab incised with crosses and a 5th century tombstone of Christian type. Although Christianity was the official religion of
the late Roman Empire, the apsidal buildings at Vindolanda and South Shields lay within the former military HQ buildings suggesting that they
had gone out of use.
The defences of three Hadrian's Wall forts, Birdoswald, Vindolanda, and
Housesteads, were refurbished, but in earth and timber, rather than stone. At Birdoswald and South Shields, there are identical traces of
some kind of timber rebuilding of fort gates. At South Shields this can
be dated to the very late 4th or early 5th century, as the rebuilding followed the backfilling of a defensive ditch which had itself been cut
after a coin of 388-402 found its way into the ground.
Birdoswald, which I excavated in 1987-92, has produced a sequence of activity which clearly continues beyond the time of circulation of some
of the latest Roman pottery and coins. The southern granary was reused in a way which required a large hearth at one end of a long rectangular
building. High status finds - a gold earring, a jet ring and a late 4th/early 5th century Theodosian silver coin were found around this
hearth. This suggests the type of hall building found in other sub-Roman contexts such as at South Cadbury hillfort in Somerset or Doon Hill near
Dunbar. The implication is that high status individuals enjoyed the best
seat, nearest the fire.
The reuse of the granary was followed by two phases of timber buildings,
each phase including one building of hall size, and two or three smaller structures which might be service
buildings. Unfortunately these are not dateable. Reckoning some 25 years for each timber building might give us
a date around 480 for the end of the sequence, but longer and shorter sequences are
possible. What is clear, however, is that the transformation of the late Roman into the post-Roman period witnessed a
major change in the organisation of buildings, with the evolution of theRoman commander into what seems like a Dark Age
The same evolution can be found at Binchester, where the bath-house of the commander's residence, which was built after about 360, changed
radically in function and status. Some rooms became a smithy and a
slaughterhouse. This was followed by a period of midden dumping, and finally the site became a wooden structure used for the working of
antler. Late 4th century building activity (which probably runs well
into the 5th century) can also be seen at Catterick and Carlisle - two
of the main towns of the region.
The problem in dating, and understanding, the sequence of change
revolves around the lack of 5th century finds. All of the pottery, coins, and objects found in the latest phases of activity are of late
Roman form and are 'residual' - that is, they are not necessarily contemporary with the contexts where they were
found. We cannot know how
long they continued in use, or whether they were actually deposited a
great deal later than their manufacture.
No cash, no wages
By the late Roman period, military pay had become inconsequential in value, and the main cash receipts of the soldiers were precious metal
donatives, such as gold coins minted on imperial occasions such as accessions and
jubilees. Coinage becomes rare on military sites before
the end of the 4th century, with coins of 388-402 very uncommon indeed.
It is possible, as Richard Brickstock of Durham University has suggested, that the civilian population of the area hoarded small
coinage to buy gold, against a future requirement to pay taxes in gold.
Possible evidence of this was found in the form of a gold solidus in the
concrete floor of a town house at Carlisle (the coin seems to have been
accidentally dropped into the wet concrete - a huge financial loss).
Some archaeologists, such as Martin Millett of Cambridge University,
have argued that the end of Roman administration in Britain was the
result of a revolt of an elite who resented paying taxes for an imperial system which no longer supported their
needs. As the provincial landowners and city governments took the reigns of power, the money
economy which existed in large part to provide the cash resources to pay
taxes ceased to exist. We must assume that the army, the most visible consumer of the fruits of
taxation, ceased to be paid and supplied.
Pottery production and use ends at the same time as coinage. By the end
of the 4th century the major producers of Crambeck and calcite-gritted
wares supplied some 90 per cent of the pottery needs of the military
zone, possibly through a system of military contracts. There is no stylistic change in the material after about 370 and, as pottery
specialist Jerry Evans has pointed out, no pottery has been found in
contexts post-dating the latest coins.
The specialised and centralised pottery industries needed a market
economy within which coin circulated. After the soldiers ceased to be paid in cash such an economy no longer
existed, and the end of the pottery industries followed inevitably. The surviving
had relied (as we do today) on buying their pottery from mass-producers, were simply unable to make pottery
themselves. No doubt they reverted to
using organic materials, such as wood and leather, for their domestic
implements, which have not survived.
Evidence suggests that the late 4th/early 5th centuries saw radical change in personal ornaments
too, in clothing fashions, and in tools.
Finds specialist Hilary Cool has examined assemblages of objects of this
period from across Britain, concluding that broadly similar changes took
place across the country. The number of hair-pins, for example, becomes
very small. It has been suggested that this might reflect the influence of Christianity in its recorded disapproval of elaborate
hairstyles. The use of bracelets changes, with copper alloy dominating for most of the
4th century but bone bracelets coming in towards the end of the century
A particular type of green cylindrical glass bead also appears more often in the later 4th
century, being found in association with bone
bracelets in several graves (most of these in the South). At the same
time, there seems to be a rise in the use of jet finger rings.
Intriguingly, a glass ring from Birdoswald is dark enough to appear black. There is a trend towards a greater number of small penannular
brooches during this period than previously. In short, people began to
wear different things from 20 years or so before, preferring simpler means of personal decoration and dark coloured
jewellery. To a visitor
from abroad, people in Britain would have looked different after the end
of the Roman period compared to before.
Among other classes of find, knives seem to outnumber more specialised tools by the end of the 4th
century, a reversal of the mid-4th century situation. This might reflect either a reduction in the variety of craft
skills available, or an increase in the value of iron following the end
of mass production. When iron was in short supply, it probably made more
sense to turn it into knives - which can be used for a wide variety of
tasks - than into chisels, awls, adzes and other more specialised items of kit.
People also began to show a greater urge for recycling. Counters used for
gaming, and spindle whorls used in textile manufacture began to be
made increasingly from reused (and much earlier) red Samian pottery, trimmed and pierced to fit their new
Although hard to date precisely, the changes seem to be taking place in the last quarter - even the last half - of the 4th
century. At Birdoswald, the large fort granaries went out of use about 350 and were
never rebuilt. The implication may be that a system of regular local
supply in kind had replaced supply by military contracts, and that large
stockpiles were no longer acquired.
We are beginning to see that it was a long transition from late Roman to post-Roman - which was not entirely bound up
with, but was hastened by,
the occurrences of the first decade of the 5th century when the Romans officially withdrew from
At Birdoswald, I would argue that the only change in the early 5th century was that the troops of the fort were no longer paid or supplied
by central authority. The unit was still there, however, and - drawing on the ideas of John Casey - I suggest that the old system of official
coercion might 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.
It may be that the kind of commander-patronus attested by the large
commanders' houses in the late forts continued to be an important figure as the 5th century went on. These men may have been of sufficient
influence to become imperceptibly more like chieftains in control of warbands than Roman
commanders. Such an idea would explain the use of
the hall as a centre to the settlement. Birdoswald may have become the centre of a small petty kingdom indistinguishable from others with
totally different antecedents north of the Wall, or to the west of Britain.
Some of these successor settlements perhaps had a much longer life than
others. Catterick may be a candidate for longevity, as may Corbridge - both also the probable centres of petty kingdoms. South
Shields, however, shows no evidence at all for 5th century continuity. Early 5th
century burials within the elaborate courtyard house in the fort may in
fact point to a violent end. It is probable that this was related to the
position of South Shields on the coast, vulnerable to the threat of
attack from the eastern ocean.
Tony Wilmott is a Senior Archaeologist at English Heritage's Centre for
Archaeology . His book, 'The Late Roman Transition in the North',
written with Pete Wilson, was recently published by BAR at £25.00.