(This originally appeared in Wargames Illustrated 142 (July 1999), pp.16-21.)
So far we’ve seen that there is almost no reliable written evidence at all for this period: some names of leaders and battles, some details of kingship, federate settlement, religious and political thought, some possibly stylized poetic accounts of warfare, and that’s it. What has been the point of this, for the wargamer? Quite an important one actually. Simply enough, it allows the interested historical wargamer to assess the accuracy of the information with which he is presented, and the plausibility of the figures, lists and rules with which he is provided. Nearly all of the popularly available historical material on this period (either academic but now very out-of-date, or recent but shamelessly uncritical and populist) – and pretty much all wargames writing – is based upon uncritical assessment of the almost entirely unreliable written sources discussed in parts 2 and 3, whereas the recent source-critical academic assessments on which the past two parts were based are on the whole not easily accessible. Thus I hope that it will be seen that there has been as much point to the exercise (if in a slightly different way) as there is, for example, in the presentation of inaccessible, accurate uniform information on more recent periods (Mark Allen’s excellent work on the late seventeenth century, for example). Anyway, this month, happily, we stop ruling things out and start ruling things in to the wargames equation. Not only does the archaeological evidence give us some way of unravelling the political context (or scenario, if you prefer) of the ‘Age of Arthur’; it also gives us good hard data about such central things as the weaponry and appearance of warriors, and about scenery (always important – for terrain and scenarios) and logistics.
The End of the Infrastructure of Roman Britain[i]
Fourth-century Britain was a fairly prosperous place. Its economy was booming, perhaps because of the use of British grain to feed the Rhine army, the presence of the British garrison acting as a stimulus to the economy, and so on. Although towns were, generally, less densely occupied than they had been in the second century, and public building projects dried up, money was still lavished on private town-houses and especially, in the countryside, on villas. The fourth century is a period, on the whole, of the small-to-middling, comfortable country house. British villas tend not to be huge palatial affairs, but are quite numerous and generally without sharp differentiation in terms of size and wealth. The larger and more lavish villas are inclined to be found on the western edges of the lowland zone, in a band from the Dorset, through Gloucestershire and the central midlands and to the vale of York. Both of these points are very important. The economy saw the development of fine pottery industries and, in the lowland border zone where the big villas were, of mosaic industries.
This happy state of affairs came to a dramatic end around 400. Within a generation, the organized pottery and other industries disappeared, the villas were abandoned and the towns fell into disuse. It was probably one of the most dramatic periods of change in all British history. Isolated areas within some towns continued to function but probably as agricultural centres (as at Verulamium – St Albans), and new farms sprung up near to the old villas, but although this evidence nuances old ideas of utterly catastrophic and complete abandonment, the extent of the change, from stone-built, partly central-heated, mosaic-decorated villa to wooden thatch-roofed huts and halls, or from recognizably classical urban settlement to a handful of wooden huts sheltering amidst the ruins of decayed towns, cannot be understated. At the same time, new forms of burial appeared, to which we’ll return. One interesting point is that some of the more meaningful traces of continuity – Roman towns or villas continuing as Roman towns or villas are located, again, in that peripheral area of the lowland zone mentioned above, as, for example, at Wroxeter.[ii] All this took place between about the last decade of the fourth century and the end of the first quarter of the fifth. This means two things. First, historically, the change is too early to have had anything to do with Saxon invasions or migrations; second, for wargamers, it means that functioning Roman villas and towns are not accurate as elements of fifth- or sixth-century wargames. If you want terrain pieces, then villas and towns should be in states of ruinous decay, and farmsteads of timber, thatched-roofed halls and smaller huts should be located slightly away from the old ruined site. This could make for visually interesting and attractive elements of the post-Roman wargames table.
The reasons for this dramatic collapse are very complex, and I will only propose, briefly, my own theory; many others have been put forward.[iii] Local society in Britain, as we’ve seen, was not sharply divided into massively wealthy aristocrats lording it over a depressed and oppressed servile peasantry. Instead we seem to have comfortably-off ‘squirearchy’, roughly equal in economic terms within this class. Local power, and the pecking order within this ‘squirearchy’, was probably legitimized by involvement in the Roman state, office-holding which brought authority and control over resources, and social status. The regular redistribution of these offices allowed different families to share power and to compete with each other. However, after the suppression of Magnus Maximus in 388, effective Roman government was never really restored north of the river Loire in Gaul. First there was the usurpation of Eugenius (392-4), leading to more civil war and troop movements away from the north; then the death of the emperor Theodosius (395) leaving the empire to his two young sons – this situation producing more Roman politicking and bickering between Milan (the western capital) and Constantinople. These events were then overtaken by the wars with Alaric and his Goths in Italy and the Balkans, an invasion of Italy by another Gothic leader, Radagaisus, and eventually the ‘Great Invasion’ of Gaul on 31 December 406, when Sueves, Vandals and Alans from central Germany crossed the Rhine. Britain and Northern Gaul (in many ways very similar to Britain) were neglected. In this context the situation described above came to an end. The end of effective state presence meant the end of the regulated system of patronage and office-holding which ordered British society. Competition for local power became more intense, and less regulated. For this reason, in 407, the British army raised a series of usurpers, culminating in Constantine ‘III’, who set off to Gaul with the remnants of the field army in an attempt to reunify Britain with the rest of the west. The failure of that usurpation by 410-11 made things even worse. In this ‘stateless’ situation local landlords had more difficulty maintaining their authority over neighbouring free peasants and tenants and so retired to the countryside. This difficulty led to less secure control over surplus and the need to spend what there was on establishing local support. This lack of control over surplus meant that the aristocracy did not have the resources to support craft-specialization and industry; stone villas became too difficult to maintain with no stone-working or tile making industry and thus too dangerous to live in. The aristocrats were no longer secure enough in their position to leave their rural power bases to live in towns. In all this, the monetary economy collapsed. The end result was the total ‘melt-down’ of Roman-British civilization by the middle decades of the fifth century at the latest. A political vacuum was created into which new powers were sucked, and out of which new structures were created, as we’ll see. All this is straight socio-economic history and archaeology but, if you think about it briefly, it all has fundamental implications for politics, logistics, the raising of armies and the nature of warfare; this situation also provides contexts for interesting forms of wargame – campaigns, discussion wargames and skirmishes. We’ll return to all this.
The Army in Late Roman Britain: History and Archaeology
We can say quite a lot about the Roman army in Britain from the middle of the fourth century until the earlier fifth century, mostly from archaeological sources. We also have a document called the Notitia Dignitatum (The List of Dignities) which lists most (it is incomplete) of the Roman civil and military commands (including the disposition of regiments). It is a rather tedious and difficult document, and its date and context have been the subject of much debate. For the wargamer the most important point is that it gives the shield designs of most of the regiments of the late Roman army.[iv] The authenticity of some of these have been questioned but this need not worry us unduly – on the whole they do seem to be broadly reliable, as Phil Barker argues (though the counter-argument deserves more credence than he’d have us believe). However, there are big problems in the Notitia‘s list of the British garrison. For one thing most of the regiments stationed in Britain are also listed as stationed elsewhere. We shall return to this problem but essentially it seems clear that the Notitia‘s information on Britain was hopelessly out of date when the document was compiled in the early fifth century.
We know a fair amount about the appearance and equipment of the late Roman garrison, from archaeological sources. Several books on the late Roman army have appeared recently, and I do not intend to spend much time on it as here; for once, the interested reader can be referred to fairly readily available, reliable secondary works.[v] A couple of points are worth making, though. Firstly, contrary to practically all illustrations in modern books, it is very unlikely that the rank and file of the late Roman army wore brightly coloured embroidered discs and other tablets on their tunics (on the thigh and shoulder). This was indeed a late Roman fashion but all the surviving examples are civilian (and Egyptian). As far as I’m aware there are only one or two contemporary depictions which show this fashion in a military context, and then in high-ranking situations). Depictions of late Roman soldiers almost invariably show plain tunics. I would limit the embroidered tablets to officers, standard bearers and musicians, and have the remainder in plain linen tunics and trousers. Most of the troops stationed in Britain by the fifth century were limitanei (border troops) who by the fifth century had become rather second-class. They were paid less, and more subject to abuses from their officers (such as the latter pocketing their pay and uniform allowances) than the field army regiments (comitatenses). One could plausibly indicate higher-quality field-army regiments by painting their tunics in brighter colours (sometimes red for legionaries, or bleached white linen) with perhaps a heavier sprinkling of the embroidered tablets and discs. Very late Roman limitanei should look a bit rough, their clothes unbleached, drab earth and brown shades, torn and patched-up. Nevertheless all should count as regular up to about 410, and then they should become progressively irregular as they transmuted from army regiments into the personal followings of individual war-leaders.
The late Roman army is fairly archaeologically visible not so much from its buildings and fortifications but from its metalwork. Rank and status within the army and the civil service were manifested by particular forms of brooch and by large ‘chip-carved’ belt-sets, illustrated adequately in the two Osprey ‘Warrior’ books on the late Roman army. These are found both in burials, where a display was made of the deceased’s standing by burying him with these badges of rank, and from chance finds on other sites. The distribution pattern of finds of these items is very interesting and important, as we’ll see. Archaeology also gives us an idea of the other equipment of the late Roman soldier in Britain. A few fragments of helmets have been found, mostly from the Saxon Shore. Some seem to be fairly typical of the simple, bowl-shaped late Roman helmets; another seems to be a modified 3rd-century type kept in use; most interestingly of all, there is evidence that some helmets were leather but with some metal fittings.[vi] I am not aware of any finds of late body armour, but mail and lamellar are both known archaeologically just over the channel in Gaul and I do not see it as unlikely that the higher-status units in Britain may equally have been armoured. Shields were round or oval. A hole was left in the centre of the board, across which was laid a bar,[vii] which the soldier gripped. There was no loop for fore-arm; this is therefore a single fist-gripped shield. The hole was covered, and the fist protected, by a sharply conical, pointed boss, allowing the soldier to punch with the shield-boss. Weaponry included weighted lead darts (plumbata), as have been found at Wroxeter. These were carried on the inside of the shield and hurled during the charge. Spearheads are common; the late Roman garrison of Britain seems commonly to have been armed with stout thrusting spears (though they too could have been thrown at close range); in very late Roman contexts these seem more common than heavy throwing spears of the old pilum tradition, though some probably fourth-century examples have been found on the line of Hadrian’s Wall.
A sword was found at Richborough, but, to me at least, more interesting is the fact that a large number of axes have been found associated with late Roman military contexts – at Richmond, Housesteads, Coldham, Caistor-by-Norwich, Brancaster, Caerleon, etc. These axes are of a type often associated with the Merovingian francisca.[viii] However, the Frankish throwing axe is never mentioned by anyone in a Frankish context before the mid fifth century, and no archaeological examples are known from the Frankish homelands. On the other hand, there is a whole series of references to axes in the late Roman army, as well as the archaeological instances just mentioned (to which can be added many more from the continent). It is clear to me that the late Roman army on the frontiers made much use of axes, perhaps as a cheap side-arm, and perhaps as heavy throwing weapons, alongside the lead-weighted plumbata, at short range (an axe would be cheaper and more reusable than a pilum). The Franks would then have picked up the use of this weapon from the late Roman army in the fifth century. Another interesting find is the composite bow discovered in fragments at Caerleon. Finds of arrowheads also demonstrate the importance of archery in the late Roman army.
Before leaving weaponry, it is worth stating that, as with the axes wrongly ascribed to Franks, a great deal of late Roman military equipment has been erroneously associated with barbarian federates. The belt-buckles and other fittings mentioned above were still described as ‘federate uniform’ in a British Museum exhibition in 1997, although it has been known for over twenty years now that such finds have absolutely no ‘Germanic’ or barbarian associations at all (although barbarians recruited into the army would have worn them like all other troops). The good thing about this is that we no longer have to associate certain finds, like single ‘fist-grip’ shield-bosses to ‘Germanic’ mercenaries; they are to be associated with the Roman army in general. The weapons from late Roman weapon graves no longer have to be pidgeon-holed as barbarian or federate; we can use this evidence to study the late Roman army.[ix]
The fortifications and bases of the late Roman army are enigmatic. They are less visible than their earlier predecessors, perhaps because late Roman troops were often billeted in towns and other civilian settlements for ease of supply (as the Wroxeter plumbata, and other finds, would prove). The best-known late Roman British forts are those of the ‘Saxon Shore’ a system running from the Wash to the Solent, with famous examples at Portchester and Pevensey. The use and occupation of these is enigmatic and still debated. Internally there is little substantial evidence of occupation; what there is is sporadic. Occupation frequently seems to be attested down to about 400, but is very ramshackle at Portchester by this date. There is little or no evidence of permanent occupation on any scale thereafter. At Richborough, however, there is interesting evidence of a late and rather dramatic flurry of occupation, with numerous weapon finds, some apparently damaged, and 20,000 coins of the early fifth century, more than from the rest of Britain put together. A mid fifth-century hut was found at Portchester; perhaps by this date the old fort had become the stronghold of a warband.
The other great fortified system of Roman Britain is, of course, Hadrian’s Wall. In the late Roman period the nature of occupation of its forts changed. Occupation was much less dense, and the people who lived within the walls may have been whole communities, with women and children, living in family ‘chalets’ rather than barrack blocks. There are very few coin-finds of the late fourth century on the Wall and its forts, but such as there are do go to the very end of the fourth century. After that, evidence of occupation of the old Roman style dries up. However, the wall was clearly not entirely abandoned. Ken Dark has assembled a significant amount of evidence which suggests that certain forts on the line of the Wall were used as high-status sites well into the post-Roman period.[x] The excavations at Birdoswald fort make this longevity of high-status occupation quite clear. By then the forts seem to have become the centres of local leaders, perhaps with their warbands.
This type of occupation may go back into the late Roman period. If we look at the distribution of finds of late Roman metalwork we find that it is very odd indeed – quite the opposite of that which we would expect.[xi] Apart from a find at the western end of Hadrian’s wall, the distribution of late fourth-century military metalwork is entirely restricted to the lowlands – the early Roman civil zone – with three finds north of the Humber. Early to mid fifth-century metalwork and weapons show a very similar distribution, though there are now seven find-spots north of the Humber. Now, late Roman metalwork is often found in areas where the army was not stationed, as in the barbarian homelands, when ‘Germanic’ troops in the Roman army took their old badges of service home with them, but it is never absent from areas where the army was stationed. This distribution suggests, indeed proves, to me that the Roman army had withdrawn from the Highland zone in the late fourth century. The only alternative explanation would be that the late Roman army in the north was entirely distinct, in terms of its material culture (uniform; equipment) from the troops in the south, and this idea may be ruled out by comparison with Gaul, where the frontier troops, the equivalents of those stationed on the Wall, used the same metalwork and so on as that found across northern Gaul and lowland Britain.
A context for this withdrawal would be the usurpation of Magnus Maximus, who took the British field army to Gaul in 383. To protect the productive part of his British base but at the same time make the most of his limited manpower he would appear to have withdrawn the northern and western garrisons to a line stretching from the Severn estuary to York, so slightly in advance of the old, first-century Foss Way frontier. The highland zones may have been given over to local chieftains or warleaders, or barbarian British federates, to defend. Such leaders could be those who now occupied some of the forts in Wales and Hadrian’s Wall. It is interesting that many old Welsh legends and genealogies feature Magnus Maximus as a founder figure (in the form of Macsen Gwledig). These sources have absolutely no evidential value on their own but alongside the archaeological data, they become very suggestive. Doubtless Maximus never intended this arrangement to be permanent, but that was the way it turned out. This is very important for post-Roman British politics.
The distribution of Roman military finds also impacts upon the reliability of the Notitia Dignitatum. As will be noted from the appendix, it contrasts diametrically with the Notitia‘s location of the regiments. We have already mentioned that most if not all of the British field army are listed twice in the Notitia: once in Britain, once elsewhere. It seems clear that this represents the troops of one of the British usurpers (often assumed to be Constantine III but I should prefer to see it as Magnus Maximus, especially as recent work tends to suggest that the Notitia‘s information predates Constantine’s suppression), restationed elsewhere. The Notitia‘s information on the field army was, then, out of date. Given this, it is illogical to suppose that its lists of the Dux Britanniarum‘s army and of the Saxon shore garrison, under its Count, were any more current. The evidence of the Notitia cannot therefore disprove the implication of the archaeological evidence, that the army withdrew to the lowlands in the late fourth century. I would propose, therefore, that the British army list in the Notitia predates the usurpation of Magnus Maximus, after which, as mentioned above, it was difficult for the government to get accurate information from Britain anyway.
What was the fate of the Roman-British Army? Another interesting implication of the archaeology of the period is that it shows clearly that apparently official military metalwork still made its way to Britain throughout the first quarter of the fifth century. It seems that some British units remained in some sort of existence in the lowlands perhaps as late as the 440s,[xii] though quite as what is anyone’s guess. We should remember, however, that this metalwork could have entered Britain in the first decades of the fifth century but remained in use until later. The metalwork suggests a semi-official character at least. I would suggest that gradually natural wastage and the lack of payment for the troops meant that the units slowly but surely either died out or disbanded or turned themselves into warbands lording it over their locality. Others still may have left Britain for Gaul. As late as the 460s a certain Riothamus was campaigning on the Loire with an army of ‘Britons’.[xiii] By the middle of the fifth century, though, the whole situation had changed…
Appendix: The British garrison regiments in the Notitia:
I list here the names of the regiments (and their station) which were not part of the Count of Britain’s field army. The latter are listed, and their shield designs may be found, in Barker’s Armies and Enemies of Imperial Rome. Although it is unlikely (as noted) that they were still in the forts listed, any regiments which still remained in Britain after Constantine III’s expedition to Gaul are doubtless to be found amongst those below, whereas the Count’s field army had, seemingly, long gone. I suspect, though, that Constantine III took a significant number of the remainder with him in 407.
Under the Count of the Saxon Shore:
Numerus Fortensium (Bradwell)
Milites Tungrecani (Dover)
Numerus Turnacensis (Lympne)
Equites Dalmatae Branodunenses (Brancaster)
Equites Stablesiani Garionenses (Burgh Castle)
Cohors prima Baetasiorum (Reculver)
Legio Secunda Augusta (Richborough)
Numerus Abulcorum (Pevensey)
Numerus Exploratorum (Portchester)
Under the Duke of the Britains (Dux Britanniarum):
VI Legio (York)
Equites Dalmatae (?)
Equites Crispiani (Doncaster)
Equites Catafractarii (Morbio = ?)
Numerus Barcariorum Tigrisiensium (South Shields)
Numerus Nerviorum Dictensium (Dicti = ?)
Numerus vigilum (Chester-le-Street)
Numerus Exploratorum (Bowes)
Numerus Directorum (Brough)
Numerus Defensorum (Kirkby Thore)
Numerus Solensium (Greta Bridge)
Numerus Pacensium (Piercebridge)
Numerus Longovicianorum (Lanchester)
Numerus Supervenientium Petuariensium (Brough on Humber)
… and ‘On the Line of the Wall’:
IV Cohors Lingonum (Wallsend)
I Cohors Cornoviorum (Newcastle-on-Tyne)
I Ala Asturum (Benwell)
I Cohors Frixagorum (Rudchester)
Ala Sabiniana (Halton)
II Ala Asturum (Chesters)
I Cohors Batavorum (Carrawburgh)
I Cohors Tungrorum (Housesteads)
IV Cohors Gallorum (Chesterholm)
I Cohors Asturum (Great Chesters)
II Cohors Dalmatarum (Carvoran)
I Cohors Aeliae Dacorum (?Castlesteads)
Ala Petriana (Petrianis = ?)
Unspecified unit (Carlisle)
Numerus Maurorum Aurelianorum (Burgh-by-Sands)
II Cohors Lingonum (Drumburgh)
I Cohors Hispanorum (Stanwix)
II Cohors Thracum (?Moresby)
I Cohors Aeliae Classicae (Tunnocelo = ?)
I Cohors Morinorum (Ravenglass)
III Cohors Nerviorum (?Maryport)
Cuneus Sarmatorum (Ribchester)
I Ala Herculea (Old Carlisle)
VI Cohors Nerviorum (?Old Penrith)
[i] The best book on all of this, by far, is A.S. Esmonde Cleary, The Ending of Roman Britain (London, 1989).
[ii] See, recently, R. White, R., & P.A. Barker [not the wargaming Phil Barker, in case you were wondering…] Wroxeter. Life and Death of a Roman City (Stroud, 1998). The authors’ maximalist interpretation should be treated with some caution. Clearly, though, something reasonably significant was going on in Wroxeter through the fifth century and perhaps beyond.
[iii] The best by a long chalk is Esmonde Cleary’s, in The Ending of Roman Britain. Other recent explanations can be found in K.R. Dark, From Civitas to Kingdom. British Political Continuity, 300-800 (London, 1994); N. Higham, Rome, Britain and the Anglo-Saxons (London, 1992); N. Higham, The English Conquest. Gildas and Britain in the Fifth Century (Manchester, 1994); M.E. Jones, The End of Roman Britain (Ithaca, 1996).
[iv] These are given most accurately in P. Barker, Armies and Enemies of Imperial Rome, 4th edition (Worthing, 1981). Barker used a reproduction of a better manuscript than that used by the most recent editor of the document, O. Seeck (ed.), Notitia Dignitatum accedunt Notitia Urbis Constantinopolitanae et Latercula Provinciarum (Frankfurt am Main, 1876) (repr. Frankfurt 1962). Seeck’s version thus has a number of errors.
[v] P. Southern & K.R. Dixon, The Late Roman Army (London, 1996) is the best for illustration and discussion of material; see also H. Elton, Warfare in Roman Europe, 350-425 (Oxford, 1996) for discussion of organization, strategy and tactics. The two Osprey ‘Warrior’ series books by Simon MacDowall, Late Roman Infantryman 236-565 AD (Warrior 9) and Late Roman Cavalryman (Warrior 15) are broadly reliable and nicely illustrated summaries. Barker, Armies and Enemies of Imperial Rome (above, n.4), though now nearly 20 years old, is still fairly trustworthy as far as late Roman troops are concerned, and as stated has an invaluable catalogue of the army’s shield designs.
[vi] M. Lyne, ‘Late Roman helmet fragments from Richborough’, Journal of Roman Military Equipment Studies 5 (1994), pp.97-105.
[vii] This was laid along the shield’s vertical axis rather than, as with earlier legionary shields, its horizontal.
[viii] Indeed Southern & Dixon, Late Roman Army (as n.5), pp.115-6, state that this is a Frankish weapon, evidence of Barbarian federates.
[ix] On this, see G. Halsall, ‘The origins of the Reihengräberzivilisation: Forty years on’, in J.F. Drinkwater & H. Elton (ed.), Fifth-Century Gaul. A Crisis of Identity? (Cambridge, 1992), pp.196-207.
[x] K.R. Dark, ‘A sub-Roman re-defence of Hadrian’s Wall?’ Britannia 23 (1992), pp.111-20.
[xi] H.W. Böhme, ‘Das Ende der Römerherrschaft in Britannien und die angelsächsische Besiedlung Englands im 5. Jahrhundert.’ Jahrbuch der Römisch-Germanischen Zentralmuseums Mainz 33 (1986), pp.469-574, figs. 18 and 44.
[xii] This is the argument proposed by Böhme, above n.11.
[xiii] Sidonius Apollinaris, Letters III.9. Sidonius. Letters II, Trans. W.B. Anderson (London, 65). Whether Riothamus’ men were from Great or Lesser Britain (Brittany) is unclear; Breton and Briton are the same in Latin. Some of the British migration to Brittany must have included this sort of warleader, though.