[This is a series of articles that I started writing for Wargames Illustrated or another of the ‘glossies’ but changes in the nature of the magazines and the fact that it turned out to be rather historical, as opposed to wargames, in content, I shelved it. Then I decided that I’d post it up here – or at least the historical background pieces. Maybe I’ll still submit the parts with solid wargames content to the magazines in future, with a reference to this on-line background.]
A millennium and a half after Odoacer deposed Emperor Romulus, the end of the Western Roman Empire continues to be the subject of intense debate. The Late Roman army has long been popular among wargamers, as are encounters between these forces and ‘invading barbarians’, often called ‘German’, or at least ‘Germanic’. This series brings some results of recent academic research to wargamers’ attention, principally by presenting a series of tabletop scenarios: battles in western Europe between 411 and 471. Before moving on to those, however, I will (begging your indulgence) look at some general points and commonly held but (in my view) erroneous ideas about fifth-century history (in Parts 2 and 3) and discuss the nature and appearance of fifth-century armies (in Part 4). It is generally thought that there was little real difference between ‘Roman’ and ‘barbarian’ forces in the fifth-century West. This is undoubtedly correct but I will argue that, rather than being resulting from the recruitment of large numbers of ‘tribal’ non-Romans, this was principally because ‘barbarian’ forces were, in effect, late Roman armies. This in turn was an outgrowth of fourth-century changes in the Roman army. The ‘barbarization’ of the Roman army was a more complex and subtle process than the simple matter of enlisting barbarians. We must look at some background to this first and, in so doing, I shall present a consciously ‘heretical’ response to a deliberately provocative question:
Just how good was the Late Roman Army?
There are big problems with asking this question, mainly concerning the sources available. All the documents are, of course, Roman – there are no ‘barbarian’ written accounts. Roman narratives are mired in stereotypes; their authors were expected to ape the great practitioners of classical history (Polybius, Tacitus, Sallust) and they were especially soaked in classical ethnography. Their readers knew what to expect when they were told about barbarians, and ‘truth’ as we would understand it was not part of that; history, as Cicero said, was about moral lessons – not reliable details of what actually happened. Roman ethnography, since early Republican days, dwelt on certain ideas such as that what differentiated the barbarian from the civilised man was law, the preserve of people who could see both sides of a problem. In turn (so the Romans thought) the reason why they had this moderation was climatic and biological: they lived just the right distance from the sun. Barbarians who lived too far away from the sun, in the north, were big and fierce but stupid, although very fecund. Thus northern barbarians are always wild, mercurial and massively numerous. This makes it extremely difficult to see through the stereotypes in which an account like Ammianus Marcellinus’ of the battle of Strasbourg (357) is embedded (thousands of wild, ferocious Alamanni are beaten off after exhausting themselves against the stolid, rational disciplined Roman legions) to any reality that lay behind.
Other sources have different axes to grind, often about the wickedness and corruption of the imperial state, which, we might imagine, play up all the army’s short-comings and inefficiencies. By way of comparison, I recently had to endure a train journey from York to Birmingham in a carriage dominated by five ever more drunken, loutish, bullying, racist and misogynist ‘squaddies’ on their way back from a training course and making as much of possible of their military status. Clearly (one hopes!), an aberration, but the sort of ammunition gratefully accepted by those with a particular case to argue. In the same way, one has to be careful about generalising from the tales of scandalous goings-on in the Roman army: similarly ‘true’, perhaps, but – equally – how typical? Then there are laws, official lists and treatises which we cannot be sure had much of a direct link with reality, being concerned with how things ought to be rather than how they were.
Trying to get some purchase on the late Roman army is then a difficult exercise involving a lot of unravelling and trying to ‘read against’ the sources. We also have archaeological data, of course, giving us an idea of both sides’ technological competencies but it cannot answer some other important questions before us.
The military balance of power
Estimates of the size of the late Roman army are fraught with problems and vary between 400,000 and 600,000 men. I tend to prefer minimalist views, so let’s work with the lower figure. Obviously this enormous number was spread pretty thinly around the long imperial frontiers and between the various field armies. Nevertheless, this was the pool of military manpower upon which the emperors could draw and the Empire’s famous road system meant that reinforcements could be transferred with, by any pre-industrial standards, remarkable efficiency. That said, as far as we can discern, most operational western Roman armies were pretty small. Keeping large forces in the field was a major undertaking. Emperor Valens (364-78) maintained a large (20,000?) army on the lower Danube for three years when he was fighting the Tervingian confederacy but this put such a strain on the Balkan provinces that he was eventually glad to be able to use this and a squabble with Persia to excuse a fairly unsatisfying peace with the Goths (alas for Valens, even without a great triumph, his war had thoroughly destabilised the trans-Danubian region). When Julian invaded Persia (disastrously as it turned out) with an army of 60,000 men the logistical efforts needed were legendary. These points are worth remembering.
What of the barbarians? Romans usually describe barbarian forces as many tens of thousands strong. 80,000 is a particular favourite; there are supposed to have been 35,000 Alamans at Strasbourg; and so on. These numbers are habitually accepted by wargames writers and by very good historians, who really ought to know better. The idea that even a whole confederacy could field an army tens of thousands strong is a patent absurdity. Where were the Alamans to get 35,000 men? In barbaricum east of the Rhine there were only scattered, small agricultural settlements, based mainly (though not entirely) on subsistence agriculture using (as was generally the case in the first millennium) a fairly knife’s edge seed-to-yield ratio. A high percentage of the yield had, furthermore, to be set aside for sowing next year’s crop rather than being used as food. There were no towns. The largest barbarian settlements known to us in fourth-century ‘Greater Germania’ sheltered a few hundred souls. This statement has two important consequences. The agriculture of trans-Rhenan barbaricum could not support large agglomerations of human beings not engaged in agriculture (towns, or armies). Nor was sufficient surplus generated to allow barbarian leaders to live in lavish élite palaces or similar settlements. Furthermore, recruitment would (beyond the households of the chiefs and kings) have to be in terms of small numbers of individuals from hundreds – thousands – of little settlements. Those groups would not be able to take much of their settlement’s food reserves to feed themselves when on campaign without endangering their village (knife’s edge economy remember – and men were needed at harvest) during the difficult winter months. The upshot is that a large army of the sort usually envisaged would take a long time to summon and assemble and by the time it had all been brought together, much if not most of its portable food supplies would have been consumed. Once our hypothesised 35,000-man army set off (especially if it had eaten up much of the food it brought) it would eat its way across friendly territory like a massive, deadly swarm of locusts. How long would the winter food reserve of a rural community of, say, 100 people, feed 35,000 men for? Barely one meal. The situation does not improve (indeed I would argue that it gets worse) if one assumes that barbarian armies drove their food with them, on the hoof. There would be hungry soldiers and famine before our proposed army had even reached the Rhine. Of course barbarian warriors served in the hope of loot and perhaps feeding themselves from Roman stocks. However, the countryside of late Roman northern Gaul would not feed large bodies of men with much more facility than that in barbaricum (towns were declining in size and there was a significant shrinkage in the rural settlement pattern). Surplus grain stocks would be walled up in towns or forts, requiring risk and bloodshed for their acquisition. With an army of tens of thousands there would be little food or loot to go round; desertion would, one imagines, rapidly soar.
What is more, we are asked, in the usual hypotheses, to accept that confederacies like the Alamanni could habitually put in the field larger armies than the late medieval kingdom of England (largest reliably documented army: 33,000 – and then possibly not all in one place), which was larger, much more populous, had large urban settlements, a complex monetary economy, state credit, sophisticated taxation and systems of government, and agricultural techniques that generated much greater surpluses. Armies of the size glibly accepted for the barbarians east of the Rhine are not reliably attested in western Europe until the latter years of Louis XIV. This cannot be right. I have already pointed out that the Roman Empire, with its vastly more complex economic and governmental systems had great difficulty in putting armies of 20,000 or more into the field for any length of time. Instead, some campaigns were fought with forces of only a few thousand, sometimes only about 2,000 men. John Haldon’s detailed examination of middle Byzantine logistics suggests that Byzantine field forces were generally of about 2-5,000 men. Again, that state had considerably greater complexity and fiscal, economic and logistical resources than were available to barbarian rulers. Finally, sometimes the Romans dropped their guard. Ammianus Marcellinus lets slip that his hero, Julian, had enormous difficulty dealing with a raiding force of 600 Franks, which he eventually starved out. 600 men could starve in northern Gaul. Imagine the fate of 35,000… All this leads me to propose that barbarian armies were not – indeed cannot have been – large. If the estimate for Julian’s army at Strasbourg is correct (13,000 men) then I strongly suspect that Chnodomer’s Alamans would have been at a numerical disadvantage. If, as seems likely, that figure of 13,000 is a ‘paper strength’, numbers might have been about equal but the balance of probabilities would still suggest to me that a numerical advantage on the Roman rather than the barbarian side is the more plausible. There are signs that Alamannic rulers were becoming more sophisticated in their government. They harnessed manpower to construct some impressive hillforts. Yet even so I would regard their ability to put even 10,000 men in the field as a real achievement (and not especially likely). Even the forces of the huge, and more sophisticated, Carolingian Empire, 500 years later, rarely reached that mark.
So, the numerical advantage lay with the Romans. The same was true in equipment. The Empire turned
its enormous resources to stamping out helmets and producing mail and lamellar armour in numerous state factories. The archaeological record shows fairly clearly that the late Roman army remained well armoured. The barbarians’ capacity to produce good equipment has probably been underestimated. There were large, organised iron-workings in various parts of barbaricum. At least one elaborate mail shirt was thrown into one of the Danish bog deposits (Thorsbjerg), suggesting that whoever organised the ritual ‘sacrifice’ of captured matériel felt it could be disposed of (I suspect that, contrary to usual readings, not everything was thrown into the marsh, more valuable items being retained to be disposed of as gifts to a leader’s followers). That said, I still think it would be erroneous to view the barbarians as anything like as well armoured, on average, as the Romans. In addition, the Romans had a network of well-built fortifications, artillery and a logistical system that, while far from faultless, was about as effective as anything known in the pre-industrial world (or later!).
In training and cohesion, the better Roman regulars, at least, must also have had a clear advantage over the barbarians. They were well drilled and trained and the prestige of the army attracted career soldiers from beyond the Empire, which must have been an effective means of creaming off much home-grown barbarian military talent. Barbarian veterans, though, were also surely well trained – if in a rather different way – and experienced. Such warriors, one imagines, will have been more than a match for many of the Roman conscripted troops in the less élite units. This in turn might be why the Romans invested so much in fortification – to negate such advantages (as well as hindering barbarian foraging by keeping large quantities of grain securely walled up).
However one looks at it, in any sort of straight fight or prolonged war, the barbarians were on a hiding to nothing. The idea that they were out to conquer the Roman Empire is a fallacy. Only a mad barbarian would have tried and it is no surprise that (in those terms) none of them ever did try. The aims of barbarian warfare against the Romans were much, much more subtle than that, usually aimed at securing better, more favourable relationships with the Empire and the gifts, subsidies and above all prestige that came with them. By the fourth century, society east of the Rhine and north of the Danube was saturated with Roman influences. As far as we can tell, by then all power in barbaricum was expressed using Roman symbols and idioms. Attacking the fountain of ideas of legitimate authority was no straightforward matter.
The barbarian ‘threat’
So, you’re thinking, if the balance of power was so unequal, why did the Romans themselves keep banging on about the barbarian threat and investing large amounts of money in border defences? The answer is that, as with the promotion of so many ‘bogey men’ throughout history, this was more about internal politics than external danger. The Roman Empire – obviously – was a big place and difficult to govern. As with all states, it could only work effectively if the local élite (or the major part of it) chose to side with the government. By the third century earlier means of binding the Empire together had ceased to be effective and severe crisis resulted. To be as brief as possible, the reconstituted Empire of Diocletian and his colleagues and successors (essentially) introduced a huge imperial bureaucracy. Participation in this brought the same sorts of rewards in local politics, in the fourth century, that involvement in municipal government had in the first. The attractiveness to local élites of participation in imperial administration was essentially what bound the Empire together – much though the workings of that might look like ‘bribery and corruption’ to modern British eyes. But the bureaucracy needed justifying. It was enormous (25-35,000 men) and expensive.
The other lesson that any Roman politician worth his salt would have learnt from imperial history up to the mid-third century was that serious threats to power came from Roman rivals, especially those far away from the political centre, out on the frontier with the army. The late Roman solution, in brief, involved several things. Reducing the size of the provinces and separating military and civilian office were important. So, and above all else, was the creation of the ‘inside-out’ Empire. Fourth-century emperors governed from the frontier or close to it (late Roman capitals are all placed to be close to particular stretches of frontier: York [at times], Trier [the usual western capital], Milan, Sirmium [mod. Sremska Mitrovica, Serbia], Constantinople [mod. İstanbul], and Antioch [mod. Antakya). This meant that they could involve the provincial aristocracies of strategically crucial frontier provinces in their government (Pannonians [loosely, from the south of modern Hungary] were much involved in Valentinian I’s government, for example, and Gauls in that of his son, Gratian). For several reasons, but mainly to ensure that they had sizeable bodies of the best troops under the immediate control, the emperors also created large field armies. The army grew importantly in the later Empire. This, also, cost money and required justification. The raison d’être of the ‘praesental’ field armies (those ‘in the presence’ of the emperor) was – clearly – to do with Roman politics, but no Roman emperor could openly say that he needed a huge and expensive army to overawe his Roman rivals (or last long if he did so): that would be the most obvious tyranny.
Thus the importance of the barbarian bogeyman. Emperors had a duty to be victorious over barbarians; no emperor felt secure until he could claim such a victory (often less than easy to secure, given the barbarians’ usual sensible avoidance of pitched battle). This ideology weighed heavily on imperial shoulders and caused the Empire many problems. Defending the citizens from barbarian ravages was the sign of good imperial management. Hence the lavish spending on border frontiers and the justification for large numbers of soldiers up near the frontier (archaeology suggests that much of northern Gaul was harnessed to the logistical support of the troops and the thousands of civil servants in the region). All this required money. That money was raised by taxation, taxation was the primary function of the bureaucracy, and the bureaucracy, as I said earlier, was the glue that bound the Empire together. The whole Roman state, you might say, was built on the notion of the ‘barbarian threat’. I am not saying that the people beyond the frontier posed no threat, or that they could not cause very serious damage to the provinces in their periodic irruptions. Nor was it all pretence; as I said, the demands of the rhetoric could be quite a burden. The ‘barbarian threat’, as it looms in official sources, was, however, much more a Roman artefact than a military-political reality. The proof of this can be found as, when things moved beyond the sphere of political rhetoric, when the chips were really down, the Romans almost always prioritised dealing with Roman rivals over confronting barbarians, even ones streaming through the interior provinces.
“The emperor’s coming: look busy!” The army ‘on the ground’
‘When I came back to your sacred court, I led recruits from the Thebaïd to Hierapolis [in Cilicia]. It was then that Your Clemency, after awarding me some leave, judged it well to promote me prefect of the Dionisiada ala [cavalry regiment] in the province of Egypt. But when the sacred letter [i.e. from the emperor] was communicated to Count Valacius his office replied that other men had produced letters of the same type. Now, it appears that the latter had been promoted on recommendation [of their patrons] whereas I had been promoted by imperial decision, in recognition of the services I had rendered, listed above…’
This extract comes from a remarkable papyrus letter written, about 341, to Emperor Constantius II (337-61) by a cavalry officer, Flavius Abinnaeus. It survived quite by chance along with a number of other letters from Abinnaeus, which shed invaluable light on everyday life in the late Roman army in Egypt. Abinnaeus goes on to ask that Constantius confirm his appointment as tribunus of the ala and rescind those of his rivals, who had been given the title by lesser degrees of patronage. Happily, perhaps, he seems to have got his way for he is attested with the title of prefect of the ala in four later papyri. I have excerpted it here to reintroduce the important topic of ‘corruption’ in the late Roman army. How efficient, one is entitled to ask, was an army in which several different people could be promoted to the same post in a unit by different patrons, and run around claiming to be the commanding officer?
There’s no space for much detail but it is important, I think, to make clear that the Roman army was not the marvellous, clinically ultra-efficient fighting machine that many people seem to think. Indeed it appears (as, on reflection, one might expect) to have been very much like the earlier standing armies of the early modern period (the French army before Louvois, for example). If we move away from ‘official’ sources or the classicising histories we see a rather different side of the army. Individually, the typicality or even reality of some of the episodes narrated can be questioned, as says Doug Lee in his valuable chapter on the late Roman army in the Cambridge Ancient History, but cumulatively they present an interesting picture. Many abuses attested in early modern armies to be are found. For example, officers drew ‘dead pays’ (in this case claiming rations for soldiers who are dead) and pocketed their men’s uniform allowance. Synesius of Cyrene, a Libyan aristocrat and later bishop, writes a letter in which he tells of the arrival of a unit of horse archers in his town – except that now, says he, they are just plain archers, for the officers have sold off their horses. (In the seventeenth century, the French army was once prevented from moving because the artillery commander had rented out his draft horses to local farmers for the day.) A letter of the Antiochene rhetorician, Libanius, tells how troops stationed in a region have started running a protection racket on the local peasantry. The Theodosian Code (the mid-fifth-century compilation of imperial edicts) hints at the same sort of thing when it repeatedly prohibits the soldiery from demanding exorbitant ‘cash alternatives’ (with force one imagines) from the civilians charged with providing them with pay in kind (i.e. foodstuffs and clothes). Eventually the Emperors gave up and said that the troops must just exchange at a fair price! The 376 Gothic rebellion was essentially brought on by local Roman officers trying to ‘make a quick buck’ out of the situation. And so on.
The demilitarisation of the Roman aristocracy has been much overstated. The army continued to provide a career path for provincial aristocrats. Ammianus Marcellinus himself was such a ‘gentleman officer’. This did not mean that such men allowed military service to get in the way of their traditional lifestyle. Roman aristocrats mixed negotium (‘business’: the service of the state in some function or other) with longer periods of otium (leisure) on their estates. This was quite accepted at the highest levels of governors, vicarii and praetorian prefects but it seems to have been expected, by office-holders, to apply much further down the chain of command. The beleaguered emperors legislated to stop civil servants spending years (years!) away from their jobs. What is really surprising is that they had to pass exactly the same law for military officers. Those who stayed away for between one and three years were to be demoted ten to thirty places on the list of seniority (and thus promotion). Only an officer who absented himself from the colours for more than four years was to be dismissed. Analysis of Ammianus Marcellinus’ career suggests periods with the army interspersed between spells away ‘at leisure’. I have a strong suspicion that any cohesion or efficiency in late Roman units may, as in many another army, have been down to the ‘NCOs’ rather than the officer corps.
Like most things in the Empire, the closer you were to the emperor, the more likely it was that things ran as they were supposed to. The sorts of stories we hear about may or may not be typical but they have a very human ring and one can easily envisage that they became all the more frequent the further away one went from the emperor and his chief officers. As I said, the opportunities represented by the sorts of scam listed above were the things that led people to get involved with the Empire; they were what held it together and stopped it springing apart into myriad regional units. Nevertheless one does have to ask – seriously – what effect such abuses had on the fighting efficiency of many of the units defending Rome’s frontiers.
All in all, the late Roman army was a bit like the United States army (as I envisage it). It was huge, generally well equipped by the standards of the day, usually well supplied and, broadly, well led. Its strategy was based upon the deployment of overwhelming numbers and ‘firepower’ and ruthless demonstrations of absolute military capability. Unsurprisingly it has remained popular with American military historians, strategic advisers and politicians of a particular bent. However, on a man-for-man basis, away from the elite formations the troops were not particularly well trained or of especially high combat value, winning by virtue of numbers, basic training and ‘hardware’. Thus, like the USA, the Empire, for all that it built up these threats for the purposes of internal politics, had enormous problems in dealing with low-level politico-military threats, such as those presented by the barbarian polities. Cracking a nut with a sledgehammer, as with Valens’ Gothic war of 367-9 only served to destabilise regions beyond the frontier and cause much more serious headaches further down the line. Alternatively, as with the Alamannic campaign of Valens’ elder brother Valentinian I at about the same time, they ran the risk of increasing the prestige of the object of their attack. This seems to have been the case with Alamannic kings like Macrianus on the receiving end of Roman attacks and dirty tricks. The parallels with recent history are too obvious to need pointing out.
The implications of these points for wargames ought to be revolutionary. The traditional ‘ancients’ wargame is based entirely upon Roman ideology. The Romans’ own view of their army as a well-oiled, remorselessly efficient fighting machine is accepted uncritically. The Roman army was indeed a fearsome and impressive organisation – whatever else one says about it, it was, given the constraints of operating in a pre-industrial society, always an impressive organisation – but whether it was always as finely-honed a fighting machine as it had been at some points during the late Republic and early Empire must be a moot point. The second aspect of Roman ideology that is accepted is the huge numbers of the barbarians. As shown above, the idea of the 30-80,000-man barbarian army cannot seriously be entertained (let alone wild figures such as the 200,000 men claimed for the Gothic host that invaded Italy in 405-6). As a result, therefore, Roman troops are represented as high-quality, well-equipped troops with a high points-value (and most ancients games still work to some extent on a points system to ensure a balance between the two sides, which is fair enough). To make up for this, the barbarians, less well trained, armoured and equipped and thus valued at fewer points per figure/element have to be far more numerous to make up for their shortcomings. All this is fine in that it appears to equate with the Roman writers’ accounts. But, as shown above, there are huge problems in accepting either of the premises based on those narratives.
In major set-pieces, Roman troops should be at least as numerous, and usually more so, and better-equipped than their barbarian opponents. If the Romans are deploying elite palatine or crack comitatensian regiments, these troops should also be well trained, with high morale. In such instances, the barbarians should, as in history, usually stand no chance. This is not going to make for much of a game. Any historically plausible ‘classic battle’ between Romans and barbarians ought to be about as much fun as a nineteenth-century colonial game in which the Africans (Zulus, Mtabele, Mahdists or what-have-you) have no numerical advantage over the Europeans. Most barbarian leaders recognised this too,
which is why many surrendered as soon as the Roman army appeared. Yet the fact remains that barbarian armies, as at Strasbourg, could put up a hell of a fight and drive Roman units from the field in rout. They could also defeat Roman forces, occasionally even the praesental field army – and remember that the army destroyed at Adrianopolis in 378 included most (if not all) of the best regiments in the eastern army. So to give the barbarians even the slim chance that they stood historically against an imperial field army, or the rather better odds that they had against regional forces, we have various options. The first is to upgrade the barbarians. Their depiction under most rule sets seems, however, usually about right, though there may be scope for some minor improvements here and there. The barbarians were for the most part hardened warriors with combat experience. They were not regulars though, as we would understand it, or universally heavily armoured.
A second, probably better option, is to downgrade the Romans. I would argue that even the top-ranking Roman units should not, in terms of morale or training, be better than the best barbarian units, though they should be better equipped and, as regulars, perhaps have greater unit cohesion. The bulk of Roman troops should, however, be worse, man for man, than the barbarians. Any advantages, again, should solely relate to their equipment and armour and, perhaps, to the cohesion (and maybe size) of their units.
Their true advantage, however, should lie in their numbers.
Perhaps the cheapest or most cosmetic solution relies on the wargames element. We can simply assume that the barbarian element represents far fewer men than the regular Roman one. One might even choose to portray this visually by mounting rather fewer figures on a ‘warband’ element base than on a regular element base (and/or by mounting more regulars on theirs).
The best option is probably to go the way of the scenario: ambushes, attacks on supply trains, rear-guard actions and the like. Small battles, wherein a typical, small Roman operational force of about 2,000 men is attacked by a major barbarian army (say, 3-5,000) might yield a plausible ‘straightforward’ tabletop battle but, if my understanding of the scale of many rules is right, slightly smaller than average. All of these ideas, naturally, play pretty well with Warhammer Ancient Battles.
The wargamer interested in the historical plausibility of his games but who likes an evenly matched and interesting game nonetheless, should, however, not fight standard ‘even-points’ pitched battles between Roman field armies and the barbarians at all. Any such game that is even remotely historically plausible should result, on at least nine occasions out of ten, in an extremely unpleasant, straightforward and bloody thrashing for the outnumbered and outclassed barbarians. If the traditional game (Romans vs. hordes of barbarians) is really what you want, then at least come clean and admit that what you are gaming is not a ‘plausible historical reconstruction of history’ but a game representing Roman ideas of warfare.
The late Roman army won most of its battles against barbarian enemies. But then it massively outnumbered and ‘outgunned’ its opponents, so this is hardly surprising. Against a more evenly matched opponent, the Sassanid Empire, its record is less impressive. More suggestive of the late Roman army’s real quality is the fact that sometimes the barbarians won! Perhaps the Roman army had declined but the evidence is far from crystal clear on this point. Maybe, therefore, the question we should be asking – and this would be really heretical – is ‘just how good was the early Roman army?’
 Romulus ‘Augustulus’ (Little Emperor) was not technically the last western emperor. His father, Orestes, had driven out Emperor Julius Nepos in 475 but Nepos still ruled Dalmatia and continued to do so until he was murdered (apparently with the connivance of the local bishop, Glycerius, who happened to be Nepos’ own predecessor as emperor!) in 480, while planning a come-back invasion of Italy. Romulus was never acknowledged as legitimate by the Constantinopolitan court, which recognised Nepos.
 Ammianus, incidentally, was not present at Strasbourg writing his account some decades afterwards, making heavy use of the ‘official’ battle report by the then caesar (junior emperor) Julian. This is well worth remembering. In this his account differs dramatically from his eye-witness memoirs of battles against the Persians, such as the siege of Amida and his escape where his writing is much more engaging.
 I deal at length with the question of army size in my Warfare and Society in the Barbarian West, 450-900 (London, Routledge, 2003), ch.6.
 Fortified granaries are known from northern Gaul.
 R. MacMullen, Corruption and the Decline of Rome (New Haven, 1988) is excellent, and fascinating, for cataloguing instances of inefficiency and corruption but essentially misses the point that this was the glue that bound the Empire’s disparate regions into one unit. If you are interested in this, admittedly not very military, topic, see C. Kelly, Ruling the later Roman Empire (Cambridge, Mass., 2004).
 It has even been suggested, plausibly, that some massacres of ‘threatening barbarians’ were actually staged by the Romans themselves. Constantius II’s butchery of the Sarmatian Limigantes (trying to surrender but then – allegedly – making an attempt on the emperor) is a case in point.
 The previous three paragraphs owe much to two excellent articles by John Drinkwater: ‘“The Germanic threat on the Rhine frontier”: A Romano-Gallic artefact?’, in R.W. Mathisen & H.S. Sivan (ed.), Shifting Frontiers in Late Antiquity (Aldershot, 1996), pp.20-30; & ‘Julian and the Franks and Valentinian I and the Alamanni: Ammianus on Roman-German relations.’ Francia 24 (1997), pp.1-16.
 Presumably the Ala quinta Praelectorum mentioned as stationed at Dionisiada in the Notitia Dignitatum (Eastern Section, ch.28).
 My translation from the French of A. Chastagnol, Le Bas-Empire (Paris, 1990), pp.242-3.
 A.D. Lee, ‘The army’, in The Cambridge Ancient History, vol.13: The Late Empire, A.D. 337-425 ed. A.M. Cameron & P. Garnsey (Cambridge, 1998), pp.211-37.
 The situation always reminds me of the episode in Asterix and Caesar’s Gift where the drunken legionary, presented by Caesar (as a rather bad joke) with the Gaulish village as his gift of land on discharge, says “What use is a village? I can’t drink a village!”
 A.H.M. Jones, The Later Roman Empire, 284-602 (Oxford, 1964), pp.644-5. MacMullen, Corruption and the Decline of Rome, ch.3, contains most of these stories.
 Theodosian Code, 7.18.16. The Theodosian Code and Novels and the Sirmondian Constitutions, trans. C. Pharr (Princeton, NJ, 1952).
 J. F. Matthews, The Roman Empire of Ammianus Marcellinus (London, 1989).