[This is a written-up version of the paper I presented some years ago at the Staffordshire Hoard Symposium at the British Museum. These somewhat experimental ideas have had a bit of a checkered history. The blog-post of that paper was dubiously and unprofessionally hacked for an article in National Geographic. When written up I sent this to the Journal of Medieval Military History, which proceeded to sit on it for a year or more before sending some equivocal, vague and probably empirically-erroneous 'feedback' before I pulled it. None of that improved my never-very-high opinion of military historians... And by then further work on the Hoard and in the site of its discovery had rendered much of the detailed site-specific elements of this piece rather out-of-date. Anyway, I thought - on some prompting from Asa Mittman - that I might just put this 'out there' as a blog post. Thanks are due to Asa for tracking down a copy of the final 'publication version' submitted to the journal, which had disappeared somewhere into the bowels of some data stick or other. Also to Drs Jon Jarrett and Tom Rodway for input into discussions that shaped some of my thinking on this: they ought to have been included in the final acknowledgement footnote. Mea culpa.]
This paper assesses the contribution of the Staffordshire Hoard1 to the study of early medieval warfare and, conversely, how our understanding of early medieval warfare helps us think about the Staffordshire Hoard. While, at the time of its discovery it was politically necessary to say so, the Hoard does not revolutionise our knowledge of early medieval warfare. What it does do is no less important; interestingly, even startlingly, it confirms what might otherwise have remained merely plausible hypotheses.
The study of the Hoard itself has to deal with two interpretive aspects. First, one can study it in the terms of its constituent elements and, second, one should analyse it as a material cultural artefact in itself. It is considerably more perplexing in the second sense than in the first, and therefore this paper is principally concerned with the import of the hoard as broken down into a collection of artefacts.
Before proceeding further, I should say that I have not been convinced by the idea that the Hoard as a ‘trophy hoard’, a collection of material stripped from defeated enemies on the battlefield. For an individual to have slain perhaps 100 sword-bearing warriors even during a long career seems extremely unlikely. In an early medieval context, the idea that small elements of sword and scabbard decoration should serve as trophies as such seems unlikely; I discuss some reasons for this scepticism below. However, if we abandon his use of the problematic term ‘trophy’, there would be many more points of contact between my reading and that of Kevin Leahy, heroic cataloguer of the Hoard itself.
Numbers: Army sizes
The first topic to concern us is numbers, primarily the hoary debate on the size of early medieval armies. If it does nothing else, the Staffordshire Hoard should rule the notion of the thirty-six-man Anglo-Saxon army out of court for good. I have long argued that this theory was based on a misunderstanding of Ine’s Laws, clause 13.1,6 but it has continued to be referred to as supporting the notion of very small armies, even by scholars as formidable as the late Tim Reuter.
There are various ways of extrapolating from the amount of weaponry in the Hoard itself to the possible numbers of men in an army, which I will explore below, but each method results in the conclusion that the minimalist position can no longer be held. Otherwise the Staffordshire Hoard, containing elements of over eighty swords, would represent the remains of enough equipment to arm over two whole armies, which must surely be nonsense! The question at the ‘bottom line’ must be: how is it, if armies numbered between a few dozen and a couple of hundred, that fragments from nearly 100 swords could be found together in a single deposit?
All answers to this question must proceed from the fact that the swords manifested by the Hoard are principally revealed by elaborate, gold-and-garnet or otherwise finely decorated fittings and pommels. They therefore represent the upper end of the military equipment scale. Items such as these could relate to the equipment of a noble retinue. This suggestion is not unproblematic but for now it might be retained as a working hypothesis. In all cases it leaves unresolved the problem of the ratio of warriors with less elaborate equipment to those bearing the sort of weaponry represented by the fragments in the Hoard.
The first approach begins from the proposition that (in origin) the Hoard represented the collection of the finery that an aristocrat used to reward his followers. In his bid for the West Saxon throne in the 780s Cyneheard’s followers – it does not seem unreasonable to interpret these as his retinue – supposedly numbered eighty-four, about the same order of magnitude as the number of pommels in the hoard. Maybe this is a useful index of the size of an ealdorman’s personal following. It seems reasonable, moreover, to assume that the Hoard was not the only collection of such matériel in Mercia, implying much more weaponry like this in circulation.
How much more is a difficult question to answer but, to develop this line of thinking, we can play a numbers game; no weight is attached to the details. Let us assume, taking a fairly maximalist line of what the Hoard represents, that the weapons of which it contained elements amounted to a quarter of all such arms circulating in Mercia at the time. That would imply about 400-500 swords like this at The Mercian king’s disposal. Wealthy warriors, as we know, had more than one sword we might still assume that this number of weapons could equip 300 men or more. If an ealdorman’s retinue was the sort of size represented by the hoard, it is interesting to consider the Mercian king Ludeca’s defeat in East Anglia in 827. He had at least five ealdormen with him, because five were killed (one needs to fill in some mental caveats about judging the scale of a Mercian defeat from a West Saxon chronicle, but the general point can stand). If there were over five ealdormen and their retinues, plus the king and his (surely larger) this would imply that – according to this line of approach – the number of well-equipped warriors in a significant Mercian army might be higher still.
Furthermore, these weapons, as noted earlier, must represent the equipment of the army’s upper echelons. What sort of percentage of the whole did these men represent? A half? A third? A tenth? Ultimately we cannot know but the last ratio is, in my view, more likely than the first. According to this means of calculation, the implication must be that Mercian armies numbered in the region of at least the low thousands. The only way of avoiding that conclusion would be to assume that the weapons contained in the Hoard represented pretty much all of the expensive sword-fittings of the whole Mercian army…Were that the case, the Hoard’s loss would have been an event of such considerable seriousness that we might expect mention of it in the surviving written sources![This next bit made perfect sense to me at some point but now I can barely follow it at all. I think the point was to ask how one person could end up, after 4 generations, with 80 swords still in their possession, if armies were only 3 dozen strong. I think the argument was that if one random descendant of a small ‘Sawyer-sized’ army could be in the possession, via accumulated inheritance over a century, of 80 swords, then - assuming equal division of spoils among the 36 aristos and equal inheritance between sons from one generation to the next - there must have been several dozen other such collections in existence at least and that would imply a colossal rate of destruction of other similarly-sized small armies (over the previous century), which seems extremely unlikely. But now I no longer understand why I made the assumptions I did make, or why I thought they were logical or reasonable. Ho hum.]
A second, not unrelated, method of extrapolating from the Hoard to army numbers begins by postulating that, rather than simply being a static snap-shot in time of material able to be gathered at a specific moment (thus suggesting the availability of enough material to equip large armies), the Hoard instead represents a single deposit of all of one aristocratic family's accumulated inherited military material, collected over a long period. The wide range of dates for the manufacture of the items in the Hoard, spanning the period between the mid-sixth century and the mid-seventh, would support this proposal. Initially, we might suppose that, in this way of seeing, the Hoard could refer to much smaller armies, with elements of weapons collected over a century or so. We must therefore go back three or four generations to cover the whole period covered by the dating of the objects within the Hoard. We might, for the sake of this exercise, also assume equal, partible inheritance from one generation to the next. To try and support a minimalist interpretation, we could further posit that the person who acquired the earliest objects was the member of a ‘Sawyerite’ thirty-six-man army (deriving this figure from Ine's Laws). Let us then, for the sake of simplicity, suppose that this man had two sons (or at least two that survived to adulthood to start a family), that each son had two sons, that their sons in turn also had two sons, and so on. In each generation, each son would inherit approximately a half share of their father's collection. Again, to keep towards a plausibly minimalist hypothesis we can posit a 33% extinction rate of male lines in each generation, across the families represented or founded by the original thirty-six-man army. In this reconstruction there would be sixty-four males of this class of warrior families (descendants of the original thirty-six) in the third generation and around eighty-six (coincidentally about the same number as the swords in the Hoard) in the fourth. Such a supposition might be countered by assuming the ‘ennoblement’ of one successful follower per warrior per generation (or at least of a number equal to that of the losses), but for now let us leave this out of the equation to keep the assumed numbers constantly at a minimal level. Let us also suppose an equal and standard increase in the collection over time. However, in order to come out with eighty swords in one collection at the end-point, when the hoard was deposited, you have to assume that this acquisition occurred at a rather surprising rate of about sixty to eighty swords per generation, as will become clear.
Thus, in this reconstruction, the family founder had one sword to start off with and acquired seventy-nine, leaving forty to each son. His sons acquired eighty each in addition to the inherited forty, leaving sixty to each son. Their sons in turn added eighty to their inherited sixty and left seventy to each of their sons. The member of this generation to whom the Hoard belonged before it was stolen or before he decided to deposit it had acquired a further ten before the theft or deposition, leaving the eighty or so represented in the Hoard. Assuming that there were no ennoblements since the initial formation of the thirty-six-man army, with the wastage of 25-33% per generation there would as noted be eighty-six aristocratic descendants of the initial army members alive at the time the Hoard was deposited. Each of them had (by the time of the Hoard's deposition), in accordance with our assumption of equal inheritance, sixty to eighty swords!
Now, we are thus far assuming (to keep within a Sawyer-sized army) that our thirty-six man army is fighting and taking booty from the equivalent of perhaps as many as sixty to eighty like-sized armies per generation to keep up the in-flow of swords or sword-elements, keeping the ones that it acquires in a treasury (or smashing them up, melting them down or removing the precious fittings from them) rather than using them to equip followers warbands and that the only swords in use are ones with ornate decorated pommels. If we remove any or all of those assumptions, and posit that, rather than starting with a thirty-six man army, we begin with thirty-six men of the powerful aristocratic class, each with a retinue, to equip which the swords are used, the army size by the time of the Hoard’s deposition mushrooms. Eighty-four men with even fifty swords each (less than I have been working with above, but to introduce a margin of error and allow that some swords came from inheritance from extinct lines) would mean 4,200 decorated swords in circulation available to arm warriors. If one adds to this an equal number of men with less elaborate swords (or spears only) one reaches 8,400 men. Alternatively, on the ‘Sawyerite’ model, we have to assume either that there were about 8,300 swords gathering dust in various halls, as ‘army-surplus’ or (more reasonably) a like number of spare pommels, which also seems rather unlikely. Even at the end of the first generation there could be 4,800 men well equipped for war. If (as I would) one thinks that the multiplier (warrior with decorated sword to warrior with non-decorated sword/spear only) would be higher than that, one reaches greater numbers of putatively available warriors. Therefore, according to either of our first two methods of extrapolating an army size from the number of swords in the Hoard, and assuming that not everyone went on every campaign, 5,000 men must be feasible for a serious force.[From here on I think I can still follow and accept the argument. Maybe the last bit was why the piece got rejected but some actual feedback and detailed comment would have been helpful nonetheless. And it could easily be cut. The whole article doesn’t depend on it and it’s not very long.]
You can play with these figures (as long as you remember that over three or four generations you have to end up with the traces of eighty pommels in the hands of one person/family), refine them and reduce the assumptions (especially by reducing the number of blades to pommels – assuming that more than one pommel in the Hoard had at some time or other been fitted to the same blade) but it is still difficult not to end up with army sizes in at least the low thousands unless – as I said above – one assumes the Hoard was the only such collection (or one of only very few) in existence. Even then I would argue that, if these eighty decorated swords were the only ones in the kingdom, that must imply an army of getting on for a thousand at the very least (and probably more than that).
Alternatively, were we to think that the Hoard represents loot taken at one time, this sort of speculation will lead ultimately to the same sort of conclusion, unless, again, one supposes that the Hoard represented more or less the only such collection in existence in Mercia. In this way of thinking the ratio of blades to pommels has to be 1:1. If these swords were collected at one point in time from the dead and captured of a defeated army, one must assume some sort of sharing of booty amongst the victors, their king, his aristocrats and their followers. Even if this was the King’s cut, one would surely have to multiply the swords represented by the Hoard by a multiplier of five or so to arrive at a figure for the initial quantity of booty, and then add on a multiplier for less well-equipped warriors and for those who escaped. One would, I suggest, end up again with a figure in the low thousands.
There are however other possible scenarios that might leave us with the option of very small armies in the ‘middle-Saxon’ period. One would be to assume that the Hoard, rather than being some sort of nefarious concealment (of a theft, perhaps) or the burial of treasure intended to be recovered, was more ritual in nature. Rather than being the property of a single individual or family, it might have been a collective deposition. The serious mid-seventh-century plague in England, it has been suggested, might have caused people either to deposit such a ‘gift to the gods’ in a reversion to paganism such as is documented by Bede, or in an act of Christian contrition, abandoning the vanities of worldly status. Breaking off the elaborate decorations from weapons to contribute to such a deposit would keep the swords themselves in circulation. Possibly, in this reading, eighty or so aristocrats from across the kingdom came forth with contributions from their own weapons. Spare pommels might have been brought too (removing the obligation to see one blade per pommel). Thus, even with the multiplier for less well-equipped followers and an addition for those who chose not to participate, this scenario might leave us with a notional army of fewer than a thousand men. It would nevertheless, I suggest, be closer to a thousand men than to the few dozen to a couple of hundred supposed in minimalist interpretations. We might, nevertheless, reasonably assume that if not every member of this warrior aristocracy served in every campaign an average force might number only a few hundred.
Alternatively, we could suppose (again) that the Hoard represents the loot captured from a single defeated army in one engagement, marrying this with the previous suggestion by assuming that the decorated elements were stripped from the weapons taken and deposited in a single, collective deposit. In this reconstruction, then, even though we must assume a 1:1 ratio of blades to pommels, only eighty or so men in an army had equipment of this quality. That would still leave the unknown modifier, of how many less well equipped men we suppose existed to each lavishly attired warrior, but the total number could easily come in at well under a thousand men. We would, however, also have to account for the men who got away, and the minimalist line of reasoning in this scenario further assumes that no decorated weapons were taken away whole rather than being stripped of their ornate fittings – which would be unlikely. Thus even here we might end up with an initial force of between one and two thousand men.
The only way that I can think of to reduce the numbers further would be to assume either that none of the enemy got away and that they were all butchered or bought their escape by giving the valuable elements of their swords to the victors as a ransom (an attractive interpretation because it makes sense of the hasty snapping off of pommels from hilts). The latter assumption is attractive too because it allows us to suppose that the iron elements were taken away by the defeated rather than assuming the incremental generation-upon-generation proliferation of swords demanded by the other ways of seeing. Such a practice is nowhere referred to in the written sources but it would make sense. Having to pay for the re-decoration of their weaponry (demanded by the culture of martial display, discussed below) would divert the vanquished’s resources away from other uses, such as producing new blades. In this reading we might assume that eighty decorated swords was the totality of such weapons in a beaten army and simply postulate an appropriate multiplier to account for less well-equipped warriors. Again I am assuming a collective ‘ritual’ deposit of the finery. The final figure for the defeated army might in this scenario be a few hundred. The problem, once again, is that we have to assume that, as well as the broken off ‘ransom’ pieces, the weapons of the fallen were similarly deprived of their fittings (but for different reasons), and that no weapons taken from the dead were kept whole. The entirety of the captured loot was deposited. To my mind, however, these minimalist readings make less sense of the composition of the Hoard as a whole, which seems to be much more selective than that. For this reason, although the other readings cannot definitively be discounted, I prefer the models based on the equipment of a noble family’s retinue.
With some scaling down, based on a probable ratio of pommels to blades greater than 1:1, these still imply higher numbers in a ‘middle Saxon’ army. Even the other readings, though, tend – as we have seen – to produce totals rather higher than the thirty-six to one or two hundred that proponents of very small armies have been wont to posit. Nicholas Brooks’ heavyweight rebuttal of Peter Sawyer’s minimalist thesis proposed, for ninth-century armies, the sort of size that I am suggesting, a conclusion which I have supported, at least for the period before the ninth century. Overall the Hoard makes it difficult to argue that significantly smaller armies, numbered in only hundreds at most, were the norm in major outbursts of warfare.
Numbers: Armament and equipment
My second topic, under the general heading of numbers, concerns armament. The preceding discussion implies that swords were not uncommon. If the weapons implied by the Hoard represent only the costlier upper end of the range of swords in existence, then, as we have just seen, we should surely envisage at least a couple of thousand swords circulating among the Mercian warrior class. It has occasionally been argued that swords were rare because they were expensive. The rarity of these items, though not the fact that they were expensive, is not a conclusion that really emerges from the evidence and the Staffordshire Hoard implicitly supports those who would question it. The implications of the Hoard are, overall, that the Mercian army was larger and better equipped than has sometimes been thought. One can only avoid accepting both of those points by emphasising one of them further. In other words, to counter the idea that the Hoard implies well-equipped troops you must argue that the armies were even bigger than has been suggested above (in order to reduce the proportion, if not the number of well-equipped troops within the army), and the idea of smaller armies can only be defended by maintaining that Mercian warriors were typically even better equipped than has generally been thought (by distributing the quantity and quality of the weaponry evidenced by the Hoard amongst a smaller number of individuals). This is far from being mere military trivia. These conclusions are important in determining how we should think about warfare’s scale, its social and political role, the resources mobilised for its support, and ultimately about the nature of Mercia and its kingship.
The Hoard and the Changes of c.600
A related issue concerns the precise moment of the deposition of the Staffordshire Hoard, within the chronology of early medieval warfare. Although the Hoard contains objects spanning a century or so, it appears to have been deposited in the third or fourth quarter of the seventh century, placing it later than a quite revolutionary phase of earlier medieval European history. To the decades on either side of 600 can be ascribed a series of changes related to the organisation and practice of warfare, though this has not hitherto been very clearly recognised and the precise nature of the changes and their interconnection still need to be worked out. Within an Anglo-Saxon context, this period is less securely understood because of the absence of written evidence from before transformations. The insular Anglo-Saxon concentration on the Augustinian mission and the conversion to Christianity has further obscured the possibility (indeed in my view the probability) that the changes observable within the archaeological record might reveal similar transformations to those that we can trace across the Channel in Northern Gaul and elsewhere. The changes in the archaeological record are very similar in both regions.
It may even be guardedly possible to characterise the changes around 600 as something of a military revolution, as they seem to encompass the raising of armies, their equipment and therefore also strategy and battlefield tactics. With equal hesitation, we could place these changes under a general heading of the end of the late antique state in the West. In the sixth-century West, armies were raised within a framework that had recognisably developed from that within which the last western Roman armies were organised in the fifth century. Warriors seem generally to have been levied from free men claiming a particular (usually barbarian) ethnic identity, living within the administrative units (such as civitates) according to which the realm was organised, and commanded by royally-appointed officers (counts and dukes). These warriors held some lands at least in return for this service and enjoyed a tax-exemption on such possessions. Other free men of ‘Roman’ ethnic identity seem to have paid taxes instead, and it is possible that such revenues financed participation in campaigns. Royal officers appear, as in late Roman times, to have been paid in drafts on taxation, and this might have been true of serving warriors too. It certainly seems to be the case that the logistics of these forces centred on the collection and stockpiling of supplies as royal dues. All of these elements (payment via the ‘delegation’ of taxation, at least partial tax-exemption, hereditary status, reward in lands) can be argued to grow out of systems of organisation that were appearing in the western Empire by c.400.
In the decades around 600 this system of army-organisation, which essentially left the army as an independent royal coercive force and, alongside the continuation of royal taxation permits us to refer to post-imperial western polities unproblematically as states, seems to have given way to a system in which warriors were raised mainly from the retinues of wealthy land-owning aristocrats. This is seen best in Francia. Privileges and immunities left some aristocrats to raise forces from their lands, into which royal officials could not enter. As royally-bestowed titles began to be held more often within particular families and aristocratic dynasties become more visible in the written sources. Archaeological evidence equally reveals the appearance of a more secure local and regional élite. In the past I have characterised the shift in army-raising as a move to a more selective levy. This was an unfortunate choice of term. What appears to be the case is more of a change in the ways in which those who served were selected. There are other ways in which too stark a contrast ought to be moderated but on the whole it does not seem unreasonable to describe the transformation that took place as a ‘privatisation’ of the army, one which set in place the outlines of a general template for the raising of armies that persisted until the introduction of Compagnies d’Ordonnance and similar forces in the late Middle Ages. This shift parallels a decline in the other key element of royal intervention in local society and surplus-extraction, taxation, which similarly seems largely to have withered in this period. The linkage between the two ought not to be surprising given what has been said about the sixth-century situation. As a result, most of the western European kingdoms after c.600 are more difficult to classify as states. The ability of central government to impose its will on local communities without significant negotiation with the landholding élite was greatly circumscribed.
The extent to which this Frankish narrative is generally applicable across western Europe is of course debatable. The evidence is best from Francia and that has a tendency to skew discussions towards the Merovingian realms. Nevertheless, general features of this reconstruction can be traced in other areas, notably Spain. In England, of course, the absence of written data from the sixth century makes detailed comparison impossible. Nonetheless it is significant that the very fragmentary evidence relating to military service in the seventh century resembles that from across the Channel in Frankish Gaul. The general similarity between the changes observable in the archaeological record, in cemeteries and settlements, in lowland Britain (Anglo-Saxon England) and northern Gaul also suggests that social transformations in both areas proceeded on similar (though probably not identical) lines. The implication of analogous socio-political developments and attendant military change in England and Francia is furthered by looking at the evidence of weaponry.
Around 600, a change took place in the forms of weaponry that are visible archaeologically. In the fifth and sixth centuries, the archaeological record in this sphere (furnished principally by thousands of furnished inhumation burials) centres on armament suggestive of a more open and fluid form of fighting. Throwing axes (franciscae), heavy, barbed javelins descended from the classical pilum (angones), and other types of throwing spear are attested. Scramasaxes are small and narrow and research on the Anglo-Saxon evidence has suggested that shields in this period were relatively small. Their bosses, characterised by a terminal disc, underline the possibility that they were more like bucklers than heavy shields, able to be moved about to block thrusts and catch an opponent’s blade in a ‘fencing’ style of fighting. The similarity of boss designs suggests that the same may be true of Frankish shields, which are less well attested in the cemetery evidence.
By contrast, after the decades of transformation weaponry seems to be much more suited to heavy, close-fighting battle. Shields are larger and spearheads bigger and heavier. The missile weapons of the sixth-century (francisca, ango, etc.) drop out of the record and scramasaxes develop into broad bladed chopping weapons (the Breitsax), seemingly combining some of the features of the sword (less common in the funerary evidence) and the single-handed battle axe (which disappears from this record until the Viking age). However, the Staffordshire Hoard immediately suggests some grounds for caution. The number of swords in the Hoard suggests that the relative frequency of such weapons in the burial record results from the reduction in the lavishness of grave-furnishing in that period rather than, as I have mooted in the past, its replacement by the heavier scramasax – better suited to close-packed ‘shield-wall’ fighting. The Staffordshire Hoard provides good evidence that the élite elements at least of an Anglo-Saxon army continued to use the long, slashing sword as a primary weapon. It might be the case that such weapons were popular because they could be used on foot or (especially) from horseback. The idea that Anglo-Saxon armies might contain mounted and dismounted elements should not be rejected. Alternatively, the crucial variable of the type of warfare needs to be introduced. Warriors might fight with mounted in some levels of warfare but dismount and in others.
There are, of course, problems involved in extrapolating tactical practice from the nature of weaponry but it does not seem unreasonable to propose that the implication of the changes in armament is, overall, towards battlefield combat that relied more heavily upon close-fighting. Some support for this notion might be found in comparing late Roman tactics with those of the later part of the early Middle Ages. The fourth-century Roman infantry continued to focus its battlefield practice upon volleys of missiles (heavy and light javelins, lead-weighted darts, arrows and slingshot), using significant numbers of light troops, before contact. By contrast, insofar as they can be reconstructed from jejune source material, tactics in the ninth century involved a slow and steady advance in dense close-packed masses (whether on foot or horseback). Some spears seem to have been thrown immediately before contact but hand-to-hand fighting appears to have been the main means of resolving the issue. This suggests, in this area, a significant shift in emphasis rather than fundamental change.
The relationship between these proposed tactical modifications and the changes in the ways in which the army was raised is difficult to ascertain. Some sixth-century weaponry, such as the francisca, was difficult to use effectively and must have required training so the shift to a more ‘retinue-based’ army should not be seen as equating crudely with a move towards more ‘professional’ warriors. On the other hand, if the suggestion of a greater emphasis on close fighting is accepted, such forms of combat might require greater cohesion within bodies of men. Mutual trust and practice in moving together in a tightly-packed body might well be better suited to warriors trained together in a household than to irregularly-assembled regional levies. The social and economic changes around 600 might furthermore have enabled the raising of slightly larger forces. The later period’s ‘shield-wall’ tactics seem better suited towards controlling more numerous warriors. If there was also, in larger campaigning forces, a greater distinction between the aristocratic warriors and retinues, forming the core of the army, and larger numbers of less regularly mobilised men, the former may have dismounted to provide a stiffening for the latter. The greater wealth concentrated in the hands of Frankish and Anglo-Saxon aristocrats in the later period probably also permitted the acquisition of more armour for their retinues, although this proposition cannot be verified. If accepted, the supposition would further enable a concentration upon close fighting.
Like most issues relating to early medieval warfare, none of the above can be definitively established but the argument is sufficiently grounded in our fragmentary evidence to be retained as a working hypothesis. The Staffordshire Hoard, dated securely to after the period of change, fits this schema. Although the Hoard’s selective composition prevents its use to discuss tactics and armament in detail, it can support the idea suggest that élite retinues were both numerous and well equipped. On balance it also seems to imply the existence of the larger armies that would be more feasible after the economic changes that occurred around 600. Whether one accepts these suggestions, it certainly illustrates the wealth that could now be concentrated in aristocratic hands and diverted towards the equipment and adornment of warriors.
The Nature of Warfare
This last point is underlined when we consider how the Hoard speaks to the nature of early medieval warfare itself. I have previously argued, against a common view which claims that harrying and sieges were the main means of deciding the outcome of warfare, that the period 450-900 was one when pitched battles were unusually common. I drew attention not just to the frequency with which battles are mentioned in narrative sources but also to the need to see warfare in a broader, socio-economic context. There were factors concerning social structure and the élite’s dependence upon the army and military affairs for its legitimacy that pushed towards more frequent battlefield confrontation. Moreover, highly-developed fortification is notably absent between the Roman Empire’s fragmentation and the ninth century and siege-warfare was not the elaborate science it would be later. This is not surprising. Settlements, other than churches and monasteries, were not the foci for wealth that they would become. If loot and booty oiled the cogs of early medieval politics, which they did (though not to the extent, or in quite the way, that is often surmised), it was not going to be yielded through the seizure of sixth-to eighth-century towns or other settlements.
The best way to make a significant profit from warfare was to defeat the enemy army in battle, because early medieval armies took their wealth with them. More and more the lesson is underlined that, particularly in the immediately post-imperial centuries, people wore their wealth, and that was nowhere truer than with warriors, except that they also rode theirs. The price of a warhorse remained fairly constant at about 10 solidi across western Europe between 450 and 900. It is difficult to know what that really meant in practice, given that the solidus was usually a somewhat abstract unit of account; suffice it to say that people exchanged reasonably sized parcels of land for horses. These were then given lavishly decorated harnesses, bits, bridles and saddles – some of this seems to be represented in the Hoard. This, it is worth pointing out, was a hugely risky investment; horses die distressingly easily on campaign. Looting the average Anglo-Saxon village –as we currently understand it – was not going to recoup such a loss and sacking a monastery was not usually an option within the norms of warfare. Even with the economic changes that had occurred by the end of the eighth century, this opposition further explains the Vikings’ choices of target and the fear that they instilled in their enemies.
An early medieval warrior’s own accoutrements did not cost the equivalent of a year’s income from a whole village, as has sometimes been implied, but they certainly did not come cheap. They were adorned and decorated as much as possible. The sometimes-seen notion that things like the Sutton Hoo helmet represent ‘parade armour’ is misconceived. The early medieval warrior was a frightening and imposing, a glittering and plumed figure. I do not doubt that in their own way these were every bit the dangerous strutting dandies that were their descendants in the Hussar regiments of a millennium later. The Staffordshire Hoard’s items emphasise this with the almost casual gilding and ornamentation of just about every object or surface that could be so decorated. This is one instance where the Hoard fascinates and intrigues, but does not surprise. What is perhaps more important, following on from some of the arguments made earlier, is just how much of the surplus from the agriculture of seventh-century Anglo-Saxon England was being invested in what we might call the ‘dandification’ of warriors. When that is taken into account, the lack of impressive settlements in earlier ‘Middle-Saxon’ England becomes easier to understand.
Battles were a huge risk – early medieval people knew that – but if victory was gained the rewards were enormous. Taking an enemy army’s horses, let alone its weaponry and armour, would represent a major windfall. Furthermore we know of the lavish tents that kings took on campaign. Kings took their treasuries with them. Charles the Bald, one of the most interesting, if also the least lucky, of early medieval commanders lost his royal finery once to the Bretons and once to his nephew Louis the Younger, as well as temporarily losing three crowns and fine jewellery on a Viking campaign. By the ninth century, and probably earlier, merchants like the shield-sellers who followed Charles the Bald’s army in 876, accompanied armies and their wares too were looted by victorious troops. Finally there were the elaborate banners that armies carried, of whose capture, as later, especial mention and celebration were made. Most meaningful early medieval discussions of trophies concern banners. After his 892 victory over the Vikings on the river Dyle, King Arnulf sent sixteen captured banners to be paraded through Bavaria as proof of his success. The one thing in the Hoard that might have served as a trophy is the cross, and that has been smashed up.
The extent of risk involved in battles is probably one reason why mechanisms seem to have existed to restrict, normalise, and even ritualise the conduct of endemic early medieval warfare. For warfare was endemic. Mercia fought forty-two recorded wars between 600 and 850, and that is almost certainly only a record of the serious outbreaks that their enemies thought worth remembering. Even with these limitations on our knowledge, half of the Mercian kings in this period fought such a major war within two years of their accession, most of the rest within four (figure). Such was the frequency of warfare that if you were defeated one year, you might expect to take back some or even most of what you lost a year or so later. Serious wars came when this sort of tit for tat failed to keep things flowing. That is the most important thing about loot and booty: the point was not to hang on to it but to keep it circulating. John, a nobleman on the Spanish march, sent the pick of the loot he took in a minor victory over the Moors – a fine horse, a jewelled sword and a mail shirt – to Louis, Charlemagne’s son (the future Louis the Pious), in return for which Charlemagne granted him lands on the frontier. Arguably, the reason for taking material booty was to give it away, often but, as we have just seen, not always to one’s followers. As everyone knows, the dragon in Beowulf is the figure of a bad ruler because he hoards, he literally sits on, his treasure. This is one reason why I cannot see the Staffordshire Hoard as a ‘trophy hoard’.
Another is the Hoard’s composition. It might be salutary to remind ourselves that for all the intrinsic interest and importance – to us – of the items recovered, to early medieval people the important bit of a sword was not the bright, shiny pommel and scabbard fittings but what we see as the rusty length of iron: what they saw as the long, sharp instrument for killing people. When people talk about important gifts or possessions that might, in a way, have functioned as trophies, it is as swords, not bits of decoration. Let us remember what an impressive feat of blacksmithing a sword-blade was, and how much it was valued. Frans Theuws, who has thought hard and seriously about sword chronologies, has said that, because blade-design stayed more or less the same, an early medieval sword itself is pretty much undatable; all the things that we can pin a date to are in most important regards the ephemera, the things you can change: scabbard, hilt and pommel. For the purposes of the chronological association of its elements, a sword is not a sealed context. Furthermore, one adds a new pommel and hilt to a good blade; one does not have a new blade made to fit to a nice pommel.
To think about what the Hoard might represent then let us first return to the idea of a retinue. In the light of what I have discussed so far, an ealdorman with a sizeable military following, such as I have mentioned, might not want simply to equip it functionally. He would want it to look impressive too. When Wilfrid formed his own retinue he did not just provide it with horses and weapons but with clothes too. When I started thinking about the Staffordshire Hoard, one idea that occurred to me was that this might similarly be thought of in terms of the ‘ornamentation’ of a medieval retinue’s appearance that would reflect on their lord – possibly even identify him: horse-harness decoration, pommels and scabbard fittings, elements of helmets, shield-fittings. It would make an impressive show, one aimed at competing with the appearance of other retinues.
The problem with this theory, though – and others – is why it would all be in one place. It does not explain the Hoard as ‘artefact-in-itself’. Given the importance of belts in the early middle ages, as markers of office and of service, the absence of buckles is also, in my view, fatal for this idea. Hypothetically, it could nevertheless represent what a lord might have to so furnish new recruits or further reward old ones. It could be booty collected by his retinue and returned to the lord for those purposes. It might be a ‘treasury’ of – literally – political capital collected, deposited and not retrieved in the course of Mercia’s turbulent internal political history. Perhaps. These are no more than wild stabs in the dark.
The Staffordshire Hoard underlines just how badly, by its very nature, the archaeological record sometimes serves us. Our knowledge of early medieval helmets is based upon high-end examples deposited intact in graves but the Hoard suggests to me that ordinary helmets might often have been as much chronological mélanges as swords: one cheek piece from one helmet and the other from another, a crest-band added from a third. Similarly, it reinforces a similar point about revealing the wealth of the early Middle Ages. Although his dismissal of cemetery archaeology’s value in establishing the power of a deceased individual and his or her family was too extreme, James Campbell was right that the Sutton Hoo treasure is small beer compared to what the written sources tell us kings had in their treasuries. In weight, the Staffordshire Hoard gold, the largest find of Anglo-Saxon gold, amounts to about 800 solidi. It is not possible to move simply from the solidus-as-coin to the solidus-as-unit-of-account that I mentioned earlier but, if we suppose briefly that it is, the Staffordshire Hoard would buy about eighty horses. That makes it more than just a tidy sum, but we should remember that equipping a reasonably-sized army, in line with the suggestions made earlier, could cost in excess of fifty Staffordshire Hoards.
I wonder if the Staffordshire Hoard might be much more similar to a coin-hoard than has thus far been suggested: a collection of units of bullion that could be used in certain transactions – maybe political ones: again, a sort of treasury. I note the deliberate cutting up of pieces and I should like to know more, when the work is done, about the weights of the components. I wonder if it is a sort of ‘hack gold’. Or perhaps it is there as raw material, to be melted down and reworked into more fashionable objects and ornaments: or coins.
The Staffordshire Hoard also confirms Mercia’s wealth. We ought not to be surprised about where the biggest find of Anglo-Saxon gold was located. The centres of Mercia were located far from the coast and thus access to foreign trade, something that seems to have made tight Mercian control over places like Kent and London a perennial problem. Nevertheless the heartland of the kingdom, like that of its two biggest rivals, Deira (Northumbria) and Wessex, lay in a band of territory along the border between the lowland, or arable, zone of Britain (or, as it was fast becoming by the time of the hoard’s deposition, Anglo-Saxon England) and the more pastoral highland zone. This region had been the most economically prosperous part of late Roman Britain, containing the most elaborate villas, the mosaic industries, and the towns which seem to have prospered most in the period and which in some cases show the best evidence of high status continuing into the fifth century. The connection between the economic prosperity of this zone and the power of the Anglo-Saxon kingdoms that emerged there deserves further study, bridging the traditional division between studies of the Roman and Anglo-Saxon periods. For these reasons the location of the Hoard ought not to astonish us.
This last point serves as a useful overall conclusion to this discussion of the Hoard’s contribution to our understanding of early medieval warfare and society. The more I consider it, as with my point about swords and trophies, the more I come round to the idea that, however we might admire and value the Staffordshire Hoard – in monetary as well as intellectual terms – it might actually have been rather less of a big deal at the time it was deposited. That, to me, makes it, if anything, more – not less – interesting and important.
 For interim information on the Hoard, see above all the lovely booklet prepared by K. Leahy & R. Bland, The Staffordshire Hoard (London, 2009); See also http://www.staffordshirehoard.org.uk/ (accessed 4 Oct., 2011). The Portable Antiquities Scheme website, on which the details of the finds are posted as they are updated, is: http://finds.org.uk/ (accessed 4 Oct., 2011). Summaries of the papers given at the Staffordshire Hoard Symposium in London can also be found on-line at http://finds.org.uk/staffshoardsymposium (accessed 4 Oct., 2011).
 Leahy & Bland, The Staffordshire Hoard, p.12; K. Leahy, ‘The Staffordshire Hoard Discovery and Initial Assessment’, pdf document down-loadable from http://www.finds.org.uk/staffshoard/artefacts/ (accessed 4 Oct., 2011), p.6.
 G. Halsall, Warfare and Society in the Early Medieval West, c.450-900 (London, 2003), pp.119-33, for a summary of the debate, and references to key works.
 P.H. Sawyer, The Age of the Vikings (London, 1962), pp.118-36. Another important minimalist argument (though less so than Sawyer’s) about Anglo-Saxon military service was E. John, ‘English feudalism and the structure of Anglo-Saxon society’, in id., Orbis Britanniae and Other studies (Leicester, 1966), pp.128-53.
 Ine’s Laws 13.1: ‘We call up to seven men thieves, from seven to thirty five is a band; above that is an army (here).’ D.M. Whitelock (ed. & trans.), English Historical Documents vol.1, c.550-1042 (2nd edition; London, 1979), doc. no.32, pp.398-407, at p.400. G. Halsall, ‘Violence and society in the early medieval West: an introductory survey’, in id. (ed.) Violence and Society in the Early Medieval West (Woodbridge, 1998), pp.1-45 at p.8; Halsall, Warfare and Society, p.16, 59.
 T. Reuter, ‘The recruitment of armies in the early middle ages: what can we know?’ in A. Nørgård Jørgensen & B.L. Claussen (ed.), Military Aspects of Scandinavian Society in a European Perspective AD 1-1300, National Museum Studies in Archaeology and History 2 (Copenhagen, 1997), pp.32-37, at p.36.
 The number of different swords represented in the hoard is uncertain. The number of pommels was originally given as eighty-four but has subsequently risen to ninety-seven. Some other types of ornament could relate to an even higher number of weapons. In what follows, however, I have retained the earlier, lower figure. This is for several reasons. Kevin Leahy has counselled that further research might reduce the number of artefacts as separate fragments are recognised to come from the same original item. Second, the ratio of pommels to swords might have been higher than one-to-one, as discussed below. Third, a single warrior might own more than one sword, which could skew the figures if ‘sword’ is used synecdochically for ‘warrior’, and, fourth, underestimating the number of weapons gives the greatest strength possible to the minimalist argument that I am refuting. For Cyneheard, see Anglo-Saxon Chronicle s.a. 757, Whitelock (ed. & trans.) English Historical Documents vol. 1, doc. no.1, pp.175-6
 Famously, for example, the will of the Ætheling Æthelstan (1014/5) bequeaths no fewer than ten swords to different people: Whitelock (ed. & trans.) English Historical Documents vol. 1, doc. no.129, pp.593-6. Late Anglo-Saxon heriots (collections of weapons to be paid to a lord at the holder’s death intestate) envisage multiple swords held by earls and thegns. Laws of Cnut, c.71. Whitelock (ed. & trans.) English Historical Documents vol. 1, doc. no.49, pp.454-67, at p.465. Nicholas Brooks suggested that the Hoard itself might be related to heriot payments: http://finds.org.uk/staffshoardsymposium/papers/nicholasbrooks (accessed 7 Oct., 2011). Arrangements for the repayment of material when a follower died are attested in some parts of Europe as early as the seventh century and earlier but, as there is no evidence of the existence of heriots as such in Anglo-Saxon England before the tenth century, extreme caution is required before adopting this proposition in detail.
 ASC, s.a. 827. Whitelock (ed. & trans.) English Historical Documents, vol.1, p.186.
 This makes the rate of sword-acquisition over a generation considerably less preposterous, as it reduces the number of victorious battles that would need to occur. Followers could bring elaborate weapons to their lord. Obviously, we do not need to assume that all the swords were acquired by loot. The ‘Sawyerite’ reading makes purchasing or commissioning the manufacture of dozens of swords just to keep them in a cupboard more than slightly unlikely. An aristocrat expecting to have to equip a large following would be more likely to commission sword-manufacture on this kind of scale. Postulating retinues also allows us to reduce the number of new swords acquired per generation as we can envisage some being returned to the lord by warriors who left his service or who died intestate. Thus, to some extent, the presumption of larger armies reduces the scale of the warfare implied by the Hoard.
 One problem with the army-size debate has been a tendency to assume that one order of magnitude applied for armies ‘across the board’. I suggest that there were many variables related to the type or scale of warfare being undertaken that could affect this significantly. Major field forces in the seventh century might, for the sake of argument, have habitually numbered five or six thousand men but that would not preclude the existence of small, endemic border-raiding armies of a couple of hundred or fewer.
 This suggestion was made Simon Keynes at the Staffordshire Hoard symposium at the British Museum in April 2010. This paper can be read on-line at http://finds.org.uk/staffshoardsymposium/papers/simonkeynes (accessed 5 Oct., 2011).
 Admittedly this would make the cross and inscription difficult to account for.
 N.P. Brooks, ‘England in the ninth century: the crucible of defeat’, Transactions of the Royal Historical Society 5th ser. 29 (1979), pp.1-20, reprinted with postscript in his Communities and Warfare 700-1400 (London, 2000), pp.48-68.
 Halsall, Warfare and Society, pp.132-3.
 Implicit, e.g., in J. Campbell, ‘Bede’s Reges and Principes’, Jarrow Lecture (1979), reprinted in id., Essays in Anglo-Saxon History (London, 1986), pp.85-98, at pp.94-95; L. Alcock, Arthur’s Britain: History and Archaeology, AD 367-634 (Harmondsworth, 1971), p.329.
 The date of the Hoard’s deposition is vexed. The metalwork suggests that the Hoard was buried not much later than the end of the third quarter of the seventh century. The inscription has caused more debate. Elizabeth Okasha argues for a date in the eighth century or later (http://finds.org.uk/staffshoardsymposium/papers/elisabethokasha accessed 10 Oct., 2011) but Michelle Brown and David Ganz propose earlier dates which could easily tally with the metalwork http://finds.org.uk/staffshoardsymposium/papers/michellebrown http://finds.org.uk/staffshoardsymposium/papers/davidganz (both accessed 6 Oct., 2011). Brown’s and Ganz’s arguments seem to me to be the more solid.
 Halsall, Warfare and Society, pp.40-53.
 I have modified the discussion in my chapter in the allegedly-forthcoming Cambridge History of War, edited by Anne Curry.
 My understanding of the state, as this stood in 2003, can be found in Halsall, Warfare and Society, pp.20-25, with references. I have subsequently moderated my opinion, to stress – indeed to give a decisive role to – the possibility of exercising coercive force within a polity. Other definitions, with which I would generally accord in whole or part (although there are important differences of emphasis between them), can be found in C.J. Wickham, Framing the Early Middle Ages, Europe and the Mediterranean, 400-800 (Oxford, 2005; paperback edition, 2006), pp.56-62; M. Mann, The Sources of Social Power, vol.1: A History of Power from the Beginning to A.D. 1760 (Cambridge, )1986, p.37; John Haldon, The State and the Tributary Mode of Production (London, 1993), pp.32-33.
 T.M. Dickinson & H. Härke, Early Anglo-Saxon Shields (= Archaeologia 110; London, 1993).
 The scramasax could be used as an agricultural implement but, when deposited ritually, in a burial, it is best
to see it as weapon, as it would be this aspect of its functions that would have symbolic content.
 See Halsall, Warfare and Society, pp.180-8 and references.
 Maxims I, lines 58ff. S.A.J. Bradley (trans.), Anglo-Saxon Poetry (London, 1982), pp.344-50.
 H. Elton, Warfare in Roman Europe, 350-425 (Oxford, 1996); P. Southern, & K.R. Dixon, K.R., The Late Roman Army (London, 1996); G. Cascarino & C. Sansilvestro, L’Esercito Romano: Armamente e Organizzazione Vol.III Dal III Secolo Alla Fine dell-impero d’Occidente (Rimini, 2009).
 U. Dahmlos, ‘Francisca – bipennis – securis’: Bemerkungen zu archäologischem Befund und schriftlicher Überlieferung’, Germania 55 (1977), pp.141-65.
 Halsall, Warfare and Society, pp.156-62.
 Ibid, pp.215-27.
 Rural settlements: (above all) H. Hamerow, Early Medieval Settlements: The Archaeology of Rural Communities in North-West Europe 400-900 (Oxford, 2002); N. Christie (ed.), Landscapes of Change: Rural Evolutions in Late Antiquity and the Early Middle Ages (Aldershot, 2004); E. Zadora-Rio, “Early Medieval villages and estate centres in France (c.300-900)” in J-A. Quiros Castillo (ed.), The Archaeology of Villages in Europe (Documentos de Arqueologia e Historia 1: Abadiño, n.d), p.77-98; J. Burnouf, D. Arribet-Deroin, B. Desachy, F. Journot & A. Nissen-Jaubert, Manuelle d’Archéologie Médiévale et Moderne (Paris, 2009), p.95-153 (chapter by Anne Nissen-Jaubert); I. Catteddu, Archéologie médiévale en France: Le premier Moyen Âge (Ve-XIe siècle) (Paris, 2009), esp.p.25-87. Towns: R. Hodges & B. Hobley (eds.), The Rebirth of Towns in the West, 700-1050. C.B.A. (London, 1988); H. Clarke & B. Ambrosiani, Towns in the Viking Age (rev. ed; Leicester, 1995); N. Christie & S.T. Loseby (ed.), Towns in Transition: Urban Evolution in Late Antiquity and the Early Middle Ages (Aldershot, 1996); G.P. Brogiolo, N. Gauthier & N. Christie (ed.), Towns and their Territories Between late Antiquity and the Early Middle Ages (Leiden, 2000); C. Scull, ‘Urban structures in pre-Viking England?’, in J. Hines (ed.), The Anglo-Saxons from the Migration to the Eighth Century. An Ethnographic Perspective (Woodbridge, 1997), pp.269-98.
 Halsall, Warfare and Society, pp.90-91, 134-7, contra T. Reuter, ‘Plunder and tribute in the Carolingian empire.’ Transactions of the Royal Historical Society, 5th series, 35 (1985), pp.75-94; id., ‘The end of Carolingian military expansion’, in P. Godman & R. Collins (ed.), Charlemagne’s Heir. New perspectives on the Reign of Louis the Pious (Oxford, 1990), pp.391-405.
 Halsall, Warfare and Society, p.175, and references.
 Ibid., pp.153-4 and references.
 See H. Hamerow, Early Medieval Settlements; P.A. Rahtz, ‘Buildings and rural settlement’ in Wilson, D.M., (ed.) The Archaeology of Anglo-Saxon England (Cambridge, 1976), pp.49-98.
 G. Halsall, ‘Playing by whose rules? A further look at Viking atrocity in the ninth century.’ Medieval History vol.2, no.2 (1992), pp.3-12.
 E.g. J.F. Verbruggen, The Art of Warfare in Western Europe During the Middle Ages, Revised and enlarged
edition, trans. S. Willard and Mrs R.W. Southern (Woodbridge 1997), pp.23-24.
 E.g. Alcock, Arthur’s Britain, p.333. The most recent publication, S. Marzinzik, The Sutton Hoo Helmet (London, 2009) has a more subtle and guarded interpretation.
 Again, the Hoard should not surprise. Its implications are entirely in line with the views of J. Campbell, ‘The sale of land and the economies of power in early England: problems and possibilities’, Haskins Society Journal 1 (1989), pp.23-37, reprinted in id., The Anglo-Saxon State (London, 2000), pp.227-45.
 Sedulius Scotus, On Christian Rulers, 3. P.E. Dutton, Carolingian Civilization: A Reader (Peterborough, Ontario, 1996), pp.402-11, at p.406.
 Continuation of Fredegar, 44. J.M. Wallace-Hadrill (ed. & trans.), The Fourth Book of the Chronicle of Fredegar and its Continuations (Oxford, 1960), pp.80-121, at p.113.
 Royal tents: E.g. Fredegar, Chronicle, 4.27. J.M. Wallace-Hadrill (ed. & trans.), The Fourth Book of the Chronicle of Fredegar and its Continuations (Oxford, 1960), pp.2-79, at p.18; Gregory of Tours, Histories, 4.14. L. Thorpe (trans.), Gregory of Tours: The History of the Franks (Penguin 1974), p.210; Regino, Chronicon, s.a. 860 (describing 851). S. MacLean (trans.), History and Politics in Late Carolingian and Ottonian Europe: The Chronicle of Regino of Prüm and Adalbert of Magdeburg (Manchester, 2009), p.137
 Regino, Chronicon, s.a. 860 (describing 851). MacLean (trans.), History and Politics in Late Carolingian and Ottonian Europe, p.137; Annals of Saint-Bertin, s.a. 865, 876. J.L. Nelson (trans.), The Annals of St Bertin. Ninth-Century Histories, Volume 1 (Manchester, 1991), p.127, 197.
 Annals of Saint-Bertin, s.a. 876. Nelson (trans.), Annals of St-Bertin, p.197.
 Annals of Fulda, s.a. 891. T. Reuter (trans.), The Annals of Fulda. Ninth-Century Histories Volume II (Manchester, 1992), p.123. Cp. Annals of Saint-Bertin, s.a. 865: Nelson (trans.), Annals of St-Bertin, p.127.
 http://www.staffordshirehoard.org.uk/staritems/the-folded-cross (accessed 5 Oct., 2011).
 Halsall, Warfare and Society, pp.2-3, pace G. Williams. G. Williams, ‘Military institutions and royal power’, in M.P. Brown & C.A. Farr (ed.), Mercia. An Anglo-Saxon Kingdom in Europe (London, 2001), pp.295-309, atp.295.
 Halsall, Warfare and Society, p.27.
 Ibid., pp.134-45.
 Monumenta Germaniae Historica. Diplomata Karolinorum I. Pippini, Carlomanni, Caroli Magni Diplomata, ed. E. Mühlbacher (Hanover, 1906), no.179.
 For similar points, see Campbell, ‘The sale of land and the economies of power in early England’.
 Leahy, ‘The Staffordshire Hoard Discovery and Initial Assessment’, p.6.
 Again, the will of the Ætheling Æthelstan, with its reference to the sword that had once belonged to Offa,
serves as a good example. Above, n..
 F. Theuws & M. Alkemade, ‘A kind of mirror for men: sword depositions in Late Antique northern Gaul’ in Theuws, F., & Nelson, J.L., (ed.), Rituals of Power. From Late Antiquity to the Early Middle Ages (Leiden, 2000), pp.401-76.
 Stephen of Ripon, Life of Wilfrid ch.2: arma, equos vestimentaque. B. Colgrave (ed. & trans.), The Life of Bishop Wilfrid by Eddius Stephanus (Cambridge 1927), pp.6-7.
 K. Böhner, ‘Die frühmittelalterlichen Spangenhelme und die nordischen Helme der Vendelzeit’, Jahrbuch des römisch-germanischen Zentralmuseums Mainz 41.2 (1994), pp.471-549.
 J. Campbell, ‘The impact of the Sutton Hoo discovery on the study of Anglo-Saxon history’, in C.B. Kendall & P.S. Wells (ed.), Voyage to the Other World: The Legacy of Sutton Hoo (Minneapolis, 1992), pp.79-101. Campbell’s article is hugely important and provides many cautionary notes. Nonetheless, he seems to me to be unnecessarily sceptical of the ability of burial archaeology to reveal unusual wealth and status. The latter must be understood in chronological, regional and even cemeterial context, as well as through the prism of contemporary ritual usage, rather than according to absolute values. The ‘wealth’ of the Sutton Hoo mound 1 burial cannot therefore be negatively compared, at least in any straightforward fashion, with the lavish possessions described being used ‘in life’ or even with the much more lavish grave of Childeric I (a comparison made by J.M. Wallace-Hadrill in ‘The graves of kings: an historical note on some archaeological evidence.’ Studi Medievali 3rd series 1.1 (1960), pp.177-94, reprinted with postscript in id., Early Medieval History (Oxford, 1975), pp.39-59.
 Halsall, Barbarian Migrations and the Roman West; R. White, Britannia Prima.
 I should like to thank the Leverhulme Foundation, whose grant of a Major Research Fellowship for the years 2009-2012 enabled me to research and write this paper. An initial version was presented to a one-day workshop on the Staffordshire Hoard at the British Museum in May 2010, organised by the Museum in conjunction with the Portable Antiquities Scheme. I am grateful to Dr Helen Geake for the invitation to speak to that workshop and to the participants for feedback and discussion.