Featured post

Dead Blog?

Because of serial spam attacks which the Blogger platform seems unable to deal with (yes - people warned me about Blogger), I have moved the...

Friday, 28 March 2014

The Peoples of Northern Europe (Part 3): Discussion and Conclusions

[Here is the concluding part, of my chapter for the Cambridge Archaeology of Late Antiquity.  Parts 1 and 2 are here and here.]

Part 3: Discussion

Developments in the Late Roman period

It is clear that, as in the Roman Empire, the third century brought big changes to Germanic-speaking barbaricum.  Many features have been noted as distinguishing the Late Roman Iron Age from its predecessor.  One of the most important, from the perspective of a ‘banded barbaricum’, is the increase in contact between Scandinavia and the Roman Empire, above all the eastern Roman Empire.  Glassware replaces bronze vessels as the most frequent import and such objects are found more widely throughout Scandinavia than before.[1]  This may reveal an increasing importance of the (possibly misleadingly-named) ‘amber routes’ from the Baltic to the Mediterranean.  The distribution of many objects reveals the importance of this artery (or cluster of arteries) of communication and, in turn the political value of controlling such a route.  The spread of political authority and identity up and down these routes is an important element of the patterns of migration in the Late Roman Iron Age.[2]  One effect of this increased power may be a more secure local power of the regional élites.  Note that the distribution of ‘lavish burials’ in Germanic-speaking barbaricum shifts noticeably westwards from the valleys of the key ‘amber route’ rivers to the ‘intermediate’ zone behind the western ‘frontier zone’ in the Late Roman period.[3]

Fourth-century archaeology generally reveals steadily increasing socio-political complexity in barbaricum, especially in the ‘frontier’ zone and Scandinavia, where it is visible across a wide range of data.  Settlements manifest a trend towards greater organisation and hierarchy.  Several features can be noted, including a degree of planning in settlement lay-out, as at the famous North Sea Coastal ‘terp’ of Feddersen Wierde, whose farms apparently follow a radial lay-out.  Feddersen Wierde also shows another trend: the appearance of the Herrenhof (see above).  Whether this relates to more rigid stratification has been questioned.[4]  Might it have represented a more communal settlement focus, for example?  Whether greater building size necessarily correlates directly with the wealth of the owners can be debated.  Nonetheless, set alongside other features of the archaeology of Germania Magna, it seems most plausible to see these buildings in terms of changing social hierarchy.  Possibly in line with this and the indications of organised planning, the late Roman period saw more frequent indications of the fencing off of individual farmsteads, plausibly demonstrating enhanced ideas of private, inheritable property.  Increasingly effective political organisation can also be seen, as noted, in the Danish bog-finds.  The rare and scattered fourth-century use of funerary ritual as a focus for local competition also suggests more settled social structures.  Relations with the Roman Empire were doubtless central to this, as noted earlier.  Although it would be as mistaken to view these as uniformly harmonious (or to downplay the seriousness of warlike interactions) as it is to see them as constantly confrontational, it is likely that peaceful social, political and economic relations were, proportionately, more normal than military.  A fairly tightly organised pattern of relationships and interactions with Rome is probably the most important element to emphasise in explaining the steady increase in trans-Rhenan barbarian socio-political complexity.

A similar picture might be posited for northern Britain.  Although the archaeological evidence is less plentiful and varied than in Germania, written indices of new confederations, the concentration of Roman imports in particular points and the emergence of large high-status sites suggests some parallels.  Close study has suggested the Empire’s ability to build up and knock down powerful groups here, just as east of the Rhine.[5]  Again, the northern frontier was, more often than not, fairly calm and the existence of an established order within which relationships could be structured must be underlined.

Ireland was drawn more tightly into the Roman orbit in the fourth century as archaeological evidence makes clear.  One element was doubtless the raiding referred to by contemporaries.  This should probably be seen alongside the possible élite distribution of late Roman imports.  Whether these come from such attacks, or from Roman payments, or from Irish leaders being able to organise exchange relations with Roman traders, they underline the increasingly important contacts across the Irish Sea.  As noted several times already, though, raiding was rarely the sole form of political relationship between the Empire and its neighbours.  Other links might have included recruitment as federate troops or into the élite auxilia palatina, two regiments of which are named ‘Attecotti’.[6]  Who the Attecotti were and where they came from are mysterious but they were certainly associated in some way with the Iris h.  Irish settlement in western Britain may have begun within the Roman period, although the evidence is nebulous.  Another form of Roman influence was, of course, Christianity.  Attempts to evangelise the Irish were made during the fifth century, but possibly began earlier.[7] This thickening network of connections doubtless lies at the root of the changes beginning in this period.  Here, however, such change may have involved the break-up of old, loose but extensive kingdoms.  It certainly seems reasonable to envisage more political stress in Ireland than in northern Britain or Germania in this period.

Fifth-Century Change

The fifth century is obviously crucial to considering the ‘late antique problematic’ beyond the limes.  As has been stressed, barbaricum was not a separate world, but the imperium’s increasingly intimately-connected periphery.[8]  The crisis into which the Western Empire was plunged from the 380s, after Magnus Maximus’ usurpation, lasting until c.420, and – more so – the failure to weather that crisis inevitably and profoundly affected the territories beyond the frontiers.  Closest to the limes, where local kings were apparently propped up by further Roman gifts and payments during the civil wars, this crisis had no immediate effect.  Franks and Alamans were only minimally involved in this period’s incursions.  Archaeologically, in some areas the period continues fourth-century developments, especially in settlements around the lower Rhine frontier.[9] 

Real problems arose further into barbaricum, however.  Here, contact with Rome was more precarious, although every bit as important in underpinning political authority and stability.  The late fourth- and early fifth-century civil wars ended the carefully managed frontier system.  This had ensured a rough parity between groups, and barbarian leaders in the interior had often been paid to counterbalance the frontier peoples.  When the managed system of diplomatic relations ended, and especially as the distracted Romans simply shored up their allies on the frontier, political stress was inevitable.  As mentioned, it had long been the case that the losing parties in barbarian political conflict took themselves to Roman territory and accordingly it is very significant that the barbarians who invaded Gaul in the early fifth century were from the ‘intermediate band’ of barbaricum: Sueves, Vandals and Burgundians.  The political stress in these regions may have led some factions to ask for support from the newly-hegemonic Hunnic leaders north of the Danube, which may have been decisive, propelling defeated elements towards the Rhine.  Nonetheless this crisis arose from internal Roman political difficulties from the 380s onwards and we should be very careful before assuming a general, exogenous Hunnic ‘push’ factor.

The early fifth-century crisis is most visible archaeologically in the North Sea Coastal regions.  The ‘Saxon Homelands’, although a frontier zone in some regards, in others have features of the interior band.  The socio-economic crises affecting the north-western Roman provinces at this time, very clearly visible in settlement abandonment, economic decline and changes in burial, doubtless impacted seriously upon the closely connected Saxon regions.  We noted this above in discussion of settlement change and abandonment and transformations in burial rites.  Saxon archaeology shares numerous features with the archaeology of Britain and northern Gaul, underlining the analytical usefulness of the concept of a North Sea cultural zone.  This period saw the re-emergence of Frisian, Anglian and Jutish identities, suggesting a break-up of the Saxon confederacy.  Migration to Britain was a crucial product of these developments.  Related to these changes are those mentioned in the heart of the interior band of Germania.  The Elbe Valley had been a crucial artery linking barbaricum and the Empire and the Roman crisis around 400 doubtless had a knock-on effect there.  By the end of the century the new Thuringian kingdom had established control of the river and spread its authority along it.

The changes around 400 had effects on the archaeology of Scandinavia.  Certain forms of import began to dry up for example.  By the sixth century something of an archaeological ‘Dark Age’ is noted in some parts of the region.  It does not, however, seem to be the case that this necessarily implied social or economic decline.  Many specialists believe that the relative archaeological invisibility of ‘Early Germanic Iron Age’ Denmark may attest more to a slow consolidation of power and social hierarchies.  It may be better to think of a longer term readjustment in response to Roman political change, rather than the short-lived but dramatic crises seen elsewhere.  Settlement patterns may intensify rather than decline, and new forms of agriculture were introduced.[10]  Nonetheless, bursts of larger or more lavish inhumations around 400 and 500 imply some crises during the period in some regions.[11]

Similar dynamics are visible in Britain.  Possible changes in Roman Britain’s governance meant that the frontier band, north of Hadrian’s Wall, became more like an interior zone.  This produced some archaeologically visible signs of crisis, such as the abandonment of hillforts, like Traprain Law,[12] and changes in burial rite, wherein funerals became important in local community politics.  As with the Saxons, fifth-century crisis apparently led to a break-up of the Pictish confederacies and, while one group retained (by the seventh century at least) the name of Picts, other identities reasserted themselves.  Earlier tribal names like the Votadini and Maetae resurfaced.  In Ireland, the break-down of the imperial links that had been developing earlier, and which had probably lain behind a certain amount of political change and stress, doubtless only emphasised the latter and, as elsewhere, produced migration into former Roman territory.

Along the Rhine frontier, crisis came later, with the Roman government’s failure to re-establish its authority along the limes and thus continue to back frontier kings.  This, as noted, is detectable in the archaeology of the Frankish and Alamannic areas.  The political stress produced led to incursions into northern Gaul and elsewhere.  Eventually the situation was resolved when the Frankish faction that controlled the Roman army on the Loire established its dominance first over the Paris Basin and then over its northern rivals.  This group, the Merovingians, extended its hegemony over the Alamanni, Thuringians and Saxons, as well as, by the 530s, removing the other barbarian kingdoms from Gaul.  Frankish overlordship extended well beyond the Rhine, though, and down the Danube.  The Empire’s demise had completely changed the relationships between the trans-Rhenan peoples and the territories west of the Rhine.  Although the Merovingian kingdom in some ways inherited the Empire’s role in what had been Germania Magna, the situation differed significantly.  Its relationships with Saxons, Thuringians, Hessians, Alamans, Bavarians were unlike those between the Empire and the barbarians, not least because Frankish territories straddled the old frontier.  Leaders of non-Frankish groups were often closely involved in Frankish politics, having marriage and other ties with Frankish aristocrats.  The Merovingian realm, although able to maintain a more or less effective trans-Rhenan hegemony during the sixth century, lacked the Empire’s prestige.  Consequently, the relationships that had cyclically produced migration from barbaricum to imperial territory ceased with the Empire’s collapse.  The fall of the Roman Empire ended the ‘Barbarian Migrations’.

Changes around 600

Change in the latter part of the sixth century and in the period around 600 has been mentioned in almost all the areas from Scandinavia to the Rhine and from Ireland to Bavaria.  We have seen changes in burials in several regions, with the increasing importance of above-ground monuments in many – northern Britain, Ireland, the Rhineland and Scandinavia.  If the archaeology of some areas, like Denmark, becomes less visible, this too must be seen as an important change.  New high-status settlements appeared in northern Britain whereas the number of small fortified farmsteads (raths, cashels and crannogs) in Ireland underwent something of an explosion.  Trading patterns changed, connecting Ireland and northern Britain with France.  This hardly exhausts the transformations of this important period.  Their explanation is complex and regionally varied but one element may, not insignificantly, have been internal political crises in the lands west of the Rhine, in Gaul.  Another, again significantly, was probably a further alteration in the nature of the relationships between the surviving, Eastern Roman Empire and the West.

The Merovingian kingdom experienced a profound political crisis from the 570s to the 620s, with royal minorities and civil wars.  As with the imperial civil wars 200 years previously, this produced faction-fighting and a slackening of control over peripheral peoples.  Within Gaul, the circumstances produced an increase in local aristocratic power and more rigid social stratification.  The very analogous archaeologically-visible changes in southern Germany suggest similar developments.  By the mid-seventh century Merovingian hegemony east of the Rhine was in tatters and this must have affected local social structures. 

These events surely cannot lie behind the changes in Scandinavia, northern Britain and Ireland, however.  Indeed they are unlikely entirely to explain the Gallic changes.  We should perceive some broader shifts under way, doubtless connected to the fall-out from the Emperor Justinian’s wars of reconquest, launched in the mid-sixth century.  These terrible and destructive conflicts failed to restore imperial hegemony and had effects far beyond the areas fought over.  They were, furthermore, accompanied by a terrible outbreak of plague, adding to the period’s generally apocalyptic feel.  These changes, which did much to rupture long-standing patterns of life in the Mediterranean, doubtless played a significant role in producing the change in economic patterns mentioned earlier, leading to closer links between Ireland and northern Britain and mainland Europe.  Those shifts in long-distance trade patterns were probably an important element in political change in northern Britain and Ireland, perhaps producing, as elsewhere, more intensive local authority and a break-up of earlier, looser hegemonies.  These Mediterranean crises may even have affected Scandinavia, where the Eastern Empire had been an important source of precious metals and other prestigious imports. 

A shift in ideas may have been as important as any of this.  The Roman Empire had been an overwhelming presence for the people beyond the frontiers, moulding all sorts of ideas about power and authority.  Imperial frontier policies had formed the system within which kings and other leaders interacted.  Diplomatic payments and gifts carried enormous importance, because of their connection with the emperor at least as much as for their intrinsic worth.  Some bracteates derived their models from depictions of the Emperor on much earlier, fourth-century Roman coins.[13]  Therefore, even after the Western Empire’s collapse, ideas continued to be shaped by notions of Rome and the emperor.  In the former imperial provinces, the fiction endured for some time that the new rulers were still encompassed within the imperium, deriving their authority from official Roman political and administrative titles.  Connections with a Frankish king’s imperially-bestowed honours, possibly even including a consulate of some sort, with a Burgundian king who was a patricius or with an Ostrogothic king who was a magister militum had much the same cachet as earlier relations with a governor, vicarius or Praetorian Prefect.  The Justinianic wars changed this.  Justinian based his wars on a strident proclamation that the Western Empire had been ‘lost’ to barbarian invasion and thus needed to be reconquered.  The ultimate failure to reintegrate all the western territories resulted in a formal boundary being drawn around the imperial territories in southern Spain and Italy.  It is thus no surprise that on the Rhine frontier the centuries-old dynamic, whereby losing political factions headed for and crossed the river into Gallic territory, came to an end.  A new, more integrated zone with what might loosely be called ‘inward-looking’ relationships and political dynamics developed within Germania Magna.   From this, eventually, the polity of ‘Germany’ emerged.[14]

Awareness that the Roman Empire no longer existed in western Europe produced a profound crisis in the former imperial territories there.  No more could legitimacy be based on an allegedly official position in imperial bureaucracy or a claim to represent the Emperor.  The Emperor himself had made it clear that his writ no longer ran in the West.  ‘Barbarian’ territory’s integration within the imperial orbit made this crisis as visible beyond the old limes as within them.  New ideological underpinnings were sought.  In the former provinces these largely came from the Old Testament and it may be no accident that this was a period when Christian (and again Old Testament) ideology became more influential beyond the old frontier – most obviously in Ireland but also in northern Britain.  Christian foundations spread into Germania Magna and, further away, shifts in the ideological bases of power apparently occurred.

The study of late antique barbaricum has very important points to make.  The experience of the peoples of northern Europe contradicts the view of Late Antiquity as a period of continuity or steady, uniform development in a particular direction.  There was constant change, and, frequently, periods of considerable upheaval.  North-western European archaeology shows, furthermore, that the fifth-century demise of the western Empire was a dramatic series of events producing crisis throughout barbaricum, as well as within the western provinces.  However, one thing that remained constant between c.300 and c.550 at least was the Roman Empire’s dominant influence in these regions.  In that sphere, the collapse of the West made little immediate difference and there the traditional framework of Late Antiquity would seem to be underlined.  It was the mid-sixth-century dramas and their fall-out in the century or so afterwards that made a huge difference, perhaps as much in these far northern and western regions as in the Mediterranean itself.  One reason for this was the awareness, finally, of living ‘after Rome’.  In that sense the period c.300-c.650 has a unity in barbaricum, very much a Late Antique unity, based perhaps ironically around the enduring influence of the Roman Empire.

[1] E.g. Hårdh (2003).
[2] Halsall (2007), p.131.
[3] Quast (2009).  Compare Gebühr (2009) and Becker (2009).

[4] Von Carnap-Bornheim (2015); Memorial Colloquium Haarnagel (2010). Nicolay (2014) sees Herrenhöfe as indicative of hierarchy, reasonably enough.
[5] Hunter (2007)
[6] Not Dig. occ. 5.197, 200.
[7] Ó Croínín (1995) pp.23-27
[8] Halsall (2014)
[9] Theuws & Hiddink (1997), pp.77-80.
[10] Robinson & Siemen (1988); Jensen, J., (2003).
[11] Fischer (2014).
[12] Feachem (1955-56).
[13] Magnus (1997), p.196; Pesch (2007), esp. her Formularfamilien A and C. On the chronology of bracteates, see Axboe (2004).
[14] Halsall (2014)

Thursday, 27 March 2014

The Peoples of Northern Europe (Part 2): Description of Data

[This is the second installment of the piece I am writing for the Cambridge Archaeology of Late Antiquity.  Part 1 is here. This segment is a rough and ready, but I hope useful, description of data.]



Throughout this period, the people of Germanic-speaking barbaricum generally lived in longhouses.[1]  Many of these were divided into aisles by internal posts, but single-aisled houses were known in some areas.  The distribution of different building traditions makes little sense in terms of recorded political or ethnic groupings and straddles the frontier.  Whether or not houses contained byres for cattle or horses (Wohnstahlhäuser) was an important cultural feature with implications for world-views: were animals kept separate from humans, or could beasts and people share a roof?  Subtle but interesting changes in the length and organisation of longhouses must relate to shifts in ideas as well as to socio-political and economic factors.  In addition to long-houses there were – where internal byres were lacking – separate stables, granaries and other ancillary buildings.  Granaries were usually raised on a platform of stout posts, although ‘lean-to’ granaries are also known.  Byre-less houses appear alongside Wohnstallhäuser in some areas, in the North Sea coastal areas especially in the fifth and sixth centuries.  Long barns are also known (identified by pollen remains), frequently aligned differently from the long-houses.

The best-known ancillary buildings are the semi-subterranean Grubenhäuser (known in British archaeology as ‘SFBs’ – Sunken-Featured Buildings – and in French as ‘fonds de cabane’).[2]  This was one of the first building-types to be recognised archaeologically, as the soil ‘fill’ in the original sunken floor was often discoloured, compared with surrounding areas, by the inclusion of organic and other debris after abandonment, making these structures easier to spot.  Various forms exist, with different numbers and arrangements of post-holes.  The sunken space in some British examples may have been floored over (hence the adjective ‘sunken-featured’, replacing ‘sunken-floored’ in the 1980s), although evidence suggests that this was not always the case in mainland Europe.  It was once thought that people lived in these huts and, although some, like slaves, may have done so, this interpretation stemmed largely from the fact that Grubenhäuser were recognised earlier than associated halls.  Grubenhäuser had diverse purposes but many have yielded evidence of weaving.  The common ethnic interpretation of the Grubenhaus as characteristically ‘Germanic’ makes increasingly little sense.  Similar sunken-featured buildings appeared in third-century Gaul and are now known throughout the late antique West, though in diverse forms.[3]  By c.650 they are found on the British east coast as far north as the Mounth (Scotland), where there was no ‘Germanic’ settlement.[4]  Although very common in Germania, it is more helpful to think of SFBs as part of an architectural ‘vocabulary’ (especially, within the Empire, where stone buildings were falling into disrepair), with influences spreading across the North Sea and other frontiers in all directions.  Given Rome’s cultural dominance, we should think very carefully before denying any provincial Roman cultural input into this supposedly ‘barbarian’ building-form.

In the north-western coastal regions of Germany and the Netherlands the ‘terp’ (in the Netherlands), ‘Würt’ (in Germany) or ‘wierde’ (in Frisian) is a distinctive settlement type, formed by the gradual accumulation of occupation layers, one on top of another.  Each major reorganisation involved raising the surface, burying the previous levels of occupation.  Over time the site became something of an island above the surrounding fields.  In antiquity this raising was necessary to avoid flooding in coastal areas.[5]  In areas occupied by the Alamanni after the Roman withdrawal from the Agri Decumates, some former villa sites were occupied, a well-documented example being Wurmlingen (D).[6]  Although the stone buildings were not maintained as such, elements of the Roman settlement and its associated estates remained important.  Otherwise Alamannic settlement outside the Höhensiedlungen (see below) is not especially well documented, though an excavated farmstead at Sontheim (D) suggests that habitations were not radically dissimilar from those in more fully explored areas.[7]

The ratio of settlements housing several families to individual farmsteads varies from place to place and time to time and has been the subject of much archaeological debate.  Some settlements in Germania could be quite large, sheltering two or three hundred people in the largest phases of Feddersen Wierde (D), Wijster (NL) and Vorbasse (Dk) for example.[8]  The economies of some could be diverse or specialised in different ways.  Wijster might have grown on the basis of trade with the limes, for example.  Nonetheless, although one might hypothesise that social developments pointed the way towards urbanisation, there were no real towns in Germanic-speaking barbaricum.

Various forms of higher-status settlement were known.  Some sites, like Feddersen Wierde, reveal the existence of farmstead complexes that are larger than the others on the settlement.  These have been dubbed Herrenhöfe (‘lordly farms’).[9]  Sometimes such a higher-status farmstead may be separated from the rest of the settlement.  The possible Herrenhof at Vorbasse was slightly removed from the other farmsteads, for example.  High-status farmsteads are frequently associated with traces of craft-working, especially metallurgy and iron-working.  At Heeten (NL) the trend reaches a sort of apogee in a small fortified site overlooking an iron-production site, the production from which could be quite significant.[10]   In south-west Germany some hillforts, generally known as Höhensiedlungen, were occupied in the fourth century.   Their nature and function varied from one example to another: some may have been refuges, others cult centres.  It is not impossible that some were forward bases of the limes.  Some, however, were certainly the centres of local leaders, and produce evidence of craft-production, including the imitations of Roman official brooches mentioned above.  Whatever an individual site’s precise role, the ability to levy manpower sufficient to raise fortifications and sometimes level areas of these hilltops suggests considerable local authority.  Although several Höhensiedlungen were abandoned in the middle of the fifth century, others, such as the famous Runder Burg (D) continued in use nonetheless.[11]  Other fortified sites are the ring-forts of Scandinavia, especially well-documented on the islands of Öland and Gotland (S), the best-known of which may be Eketorp (Öland).[12] The largest ring-fort on Gotland, Torsburgen might, after fourth-century rebuilding, have stood 7m high and have required 1000 men for defence.[13]

These settlements did not remain unchanged.  The fourth-century phases of long-used sites represented the results of steady evolution since the early Roman Iron Age but in some areas the fourth-century settlement pattern was quite new.  One feature of the later Roman period, for example, was a steady lengthening of the Wohnstallhäuser.  Nonetheless there were key moments of change.  The Frisian region experienced a decline in settlement in the later fourth century.[14]  Some areas experienced a significant crisis in the very early fifth century.  In the ‘Saxon homelands’, for example, many sites were abandoned.  Others showed rather different characteristics from their fourth-century precursors, with shorter longhouses, a higher number of ancillary Grubenhäuser and a more disordered-looking layout.  Loxstedt (D) has given its name to sites of this type.[15]  Where crisis was weathered, the subsequent settlements looked broadly similar to those of the Late Roman Iron Age, although there were differences in the nature and layout of longhouses.  At the very end of the period covered a new form, with bowed long walls, began to appear, sometimes called the ‘Warendorf’ type.[16]  This eventually became common in numerous areas bordering the North Sea, as a result of the trading and other connections that bound these communities.  We should think of similar features explaining earlier similarities in the same general cultural zone, rather than reaching immediately for the concepts of ethnicity and migration to explain them.

In Norway, around the start of the Late Roman Iron Age, large houses for the sheltering of ships appeared.  These required significant manpower for their construction and for the manning of the vessels.  By the Viking Age, they became quite widespread in the Scandinavian world but in Late Antiquity they were apparently confined to Norway.  Myhre suggested, from their concentration and association with other signs of élite presence, that some are signs of political organisation.  This view has been debated and modified, although the linkage between larger boathouses and socio-political development seems sound.[17] 

A range of trading settlements existed around the North Sea and Blatic, beginning around the start of the Late Roman Iron Age.  Throughout our period, these generally seem little different from prestigious rural settlements apart from a notable concentration of Roman imports and sometimes manufacture, too.  One of these is known at Dankirke on Jutland (Dk) and the famous site of Helgö in the Mälar valley (S) originates in about this period.[18]  More spectacular are the prestigious settlements at Gudme, associated with a small port at Lundeborg, on Fyn (Dk), and Uppåkra in Scania (S).[19]  The wider distribution of the imports that reached Fyn through Lundeborg suggests that local elites used these sites to control access to prestigious Roman products.  The ability to use much sought-after imports as gifts probably underpinned rulership in the area, giving the leaders in this North Sea/Baltic area some advantages over those in central Germania.  What happened when Roman trading contacts dried up in the fifth century is interesting and instructive.  The evidence from Gudme seems to show an increased emphasis on the site’s ritual role.  Around the end of our period the famous wics or emporia around the North Sea – Quentovic, Dorestad, Hamwic, Ipswich, Lundenwic and so on – were coming into existence, but these lie beyond the scope of this chapter.[20]

Northern Britain

North of Hadrian’s Wall, much of late antiquity conforms to what has been called a ‘gap’ in the region’s rural settlement archaeology.[21]  The classic earlier Roman Iron Age ‘type sites’, the ‘brochs’ – tall stone towers – and the souterrains – settlements with semi-subterranean storage space – had died out by the third century.  Their successors are much less easy to define and locate.  Some former brochs continued to be occupied, however, with cellular structures cut into the ruins.  From the middle of the first millennium the so-called Pitcarmick type of house, with a central soak-away, is known in Scotland.

Some higher-status fortified sites existed in the Late Roman period, often larger than those founded slightly later.[22] Examples include the fortified promontory at Burghead (Moray) and the hillfort of Clatchard Craig (Perthshire).  Further south, the Traprain Law hillfort was occupied in the later fourth century.  The principal moment of change, however, came around 600, when new hillforts began to be used across northern Britain, in areas of Scottish, British and Pictish political control.  Classic examples include Dunadd, Dumbarton Rock and Dundurn (in Scottish, British and Pictish areas respectively), which share several features.  They are comparatively small, composed of several enclosures, and reveal evidence of craft-production and elite presence, comprising, in the western regions especially, pottery and other imports from France. Dunadd is associated with a crannog (see below) at Lough Glashan, which has yielded evidence of leatherworking and other craftsmanship probably associated with supplying the hillfort’s occupants.[23]  In western parts of northern Britain the dun, a small ring-fort, is known.  Some are sufficiently small to originally have been roofed over.  These are known since the Pre-Roman Iron Age but continued to be used through our period. [24]  The transformations of this period seem to have brought longer-distance exchange contacts to the north and western shores of the Irish Sea in greater density than hitherto, as the finds at Scottish hillforts and high-status Irish sites like Lagore crannog make clear.[25]


Late Roman Iron Age settlement archaeology in Ireland is exiguous.  The principal issue is the emergence of the classic early medieval rural site, the small enclosed farmstead, known as a rath or a cashel.[26]  Rath tends to be used for sites with an earthen bank and ditch, while cashel is employed to described stone-walled sites (which usually lack a ditch).  These sites are difficult to date.  Attempts to push raths and cashels before the ‘early Christian’ period have generally been unconvincing.  Nonetheless, similar, if less clearly demarcated, settlements possibly underlay the ring-forts.  The steady increase in carbon14 dates allows some provisional points.[27]  Some fourth- or fifth-century foundations existed but the numbers increase significantly in the sixth, before a dramatic upsurge in the seventh – a picture which confirms the suppositions of these sites’ earliest investigators.  Most have only a single bank and ditch, though some have more.  The earthworks’ defensive function is debatable.  Most banks would be unlikely to deter serious attack and archaeological evidence shows that some fell into disrepair or become overgrown, even during the site’s occupation.  The bank more likely served to delineate a farmstead, while also protecting cattle and other property against robbers or rustlers.  Nonetheless, larger sites apparently had more serious defensive capabilities, with possible indications of gateway-towers and complex entrance arrangements.  Some forts have the inner area built up, either gradually through the burying of previous occupation layers (as with the terps) or through deliberate raising of the interior.  Such activity would further enhance the site’s visibility.  Some sites, however, appear not to have been permanently occupied by humans, probably serving instead as corrals. 

A third, related form of settlement is the crannog, a defended, wholly or partly man-made island within a lake or water-course.[28]  Prehistoric Irish lakeside settlements are known but the crannog-proper seems to appear in the late sixth and seventh centuries.  Given the relative dates of crannogs in the respective areas, whether the form was introduced from Britain has been mooted.  This diffusionist explanation is unnecessary but reminds us that cultural influences could spread in both directions across the Irish Sea at a time of deepening contacts, not just in the direction of historically-attested migration – a point equally applicable to early Anglo-Saxon archaeology.  The crannogs are built up from timbers or, in areas where woodland resources were less good, from rocks, and ringed with a sometimes quite elaborate palisades.  The crannog was clearly constructed partly for defence and the necessary labour implies control over considerable manpower.  Indeed, the finds from crannog sites have been more impressive than those from other ring-forts (although the nature of the sites’ stratification may be partly responsible for this).  Seventh-century imported French pottery (E-ware) and various forms of craft-working have been revealed, alongside weapons and other items.  Lagore crannog is referred to as a royal centre by written sources, confirming these sites’ high status.

Other forms of settlement are hard to locate.  Some promontory forts were occupied.[29]  A handful of unenclosed sites are also known, the ephemeral nature of which further explains the difficulty of locating earlier unfortified settlements.  If, as is usually and reasonably supposed, the various forms of ring-fort can be related to ‘comfortably-off’ farmers and higher social strata, similar points might account for the invisibility of the dwellings of lower orders of late antique/early medieval Irish society.[30] 

Domestic buildings within the forts have been difficult to define and excavate, further explaining the archaeological invisibility of Roman Iron Age settlements.[31]   Those which have been found appear to have been roundhouses.  Rectangular buildings have also been discovered but seem to be later than the period that concerns us.  The same is probably true for the souterrains, underground store-houses, refuges and passages (sometimes to sally-ports in the defences) found on Irish sites.

The appearance of these types of site, alongside a range of other evidence, suggests profound change taking place in Ireland, coming to a head around 600.   The most important change might, of course, be the enclosure (or more substantial enclosure) of farmsteads, making them considerably more visible archaeologically, but this would still have considerable implications.  Economically, these changes are associated with a much greater dependence upon dairy farming.[32]  This was apparently a long process but was worked through around 600.  The change to dairy-farming was intricately bound up with the changes in social relationships that took place in Ireland in the Late Roman Iron Age and may have been more important in that regard than in the purely economic sense.   Wealth, tribute and other legal payments were measured in cows.  The need for pasture and the creation of corrals led to all sorts of changes to the landscape.  Some of the ‘ring forts’ of the period may have been corrals and the ability to gather and protect cattle from raiding, inside a defensible work, was doubtless a major factor in the development of raths and cashels.  The best-known early medieval Irish saga, the Táin bó Cúailnge (The Cattle Raid of Cooley) deals with a great cattle raid.  Its historical setting, in the earlier Roman Iron Age, seemingly precedes the Irish political obsession with cattle, a fact which may argue against the story being an age-old, orally-transmitted tale from long before 600.[33] 

After the changes culminating around 600, the Irish economy was complex, with much craft-specialisation focused upon high-status sites, not only secular but monastic.[34]  The Church introduced a vitally important new element into the Irish settlement pattern and economy.  New religious settlements lay at the heart of complex social and economic relationships.  Ireland was now keyed into long-distance trading contacts to a much greater degree.  Ireland does was not closely bound into the Roman World during the Early Roman Iron Age.  Archaeologically-revealed imports are few and scattered.  The Romans showed little interest in the island, which seemingly was rather poor and underdeveloped economically.  Some change seems to have occurred in the Late Roman era, when imports, if not dramatically more numerous, are apparently more concentrated and possibly associated with high-status sites.[35]  However, that trading links with the Roman Empire or the Mediterranean remained fairly sparse.  As manifested by fine table-wares imported from the Mediterranean, whatever merchants from the Roman world sought in exchange for their products was concentrated more on the eastern shore of the Irish Sea.  That might of course have been because the trade ‘piggy-backed’ on ships transporting supplies to military garrisons.  Trade patterns changed around 600, however, and French imports are found much more commonly on Irish sites, further underlining Ireland’s greater incorporation into European society at this time.[36]



The normal burial rite throughout Later Roman Iron Age Germania was cremation, varying in detail between regions.  In some regions the ashes were buried with neither urn nor accompanying goods, leaving these areas blank on many distribution maps.  Large cremation cemeteries, used for centuries, are known in the ‘Saxon Homelands’.  The remains were buried in decorated pots, sometimes accompanied by the charred remnants of the costume in which the dead were cremated and sometimes with unburnt grave-goods placed alongside the ashes.  Throughout the area, though, grave-goods were sparse.  In Germany inhumation became more common as the later Roman Iron Age progressed, although, outside the lavish burial clusters discussed below, rarely with grave-goods.  Inhumation’s appearance and spread further attests to Roman culture’s increasing influence on barbarian society.[37]

Change and, especially, crisis nevertheless show up quite well in the funerary record.  Well-furnished burials, especially inhumations, stand out against the general background of nondescript cremation.  Third-century socio-political stress is visible in clusters of lavishly-furnished inhumations.  At the end of that century in the Elbe-Saale region (D), the Haßleben-Leuna group appeared, in which the display of material revolved around imported Roman silverware.  Rich burials, also employing Roman imports, are known slightly earlier in the North Sea Coast area; others have been found, further north, in Mecklenburg.  The Zakrzów (Sakrau: Poland) group of lavishly-furnished burials dates to the end of the third century and is found in the Przeworsk Culture sometimes associated with the Vandals.[38]  These manifestations of wealth were made to an audience of – presumably – mostly local people present at the burial itself, and strove to smooth over tension in local politics following the death of a member of a prominent local family.  A lavish display, probably associated with gift-giving, attempted to ensure that the tension was eased and that family-members could accede to their forebears’ local position.  This reading of furnished burials is important throughout this chapter. 

Until about 550, Swedish and Danish graves are, with notable exceptions, usually fairly nondescript.  Inhumation (as in Zealand) and cremation (e.g. Donbaek, [Dk]) were practised, both without many grave-goods (see the site at Hjemsted, Dk).  Weapon burials have been estimated as comprising only about 1-2% of the total but are scarce throughout Germanic-speaking barbaricum at this time.  More lavish burials are known near Esbjerg (Dk), for example, some being interred in old barrows.  Danish burials are frequently grouped into family clusters.  Earlier within the period, grouping by sex apparently took place, but later disappeared.  Small boats were sometimes used as coffins (e.g. Praestestien, Dk).  Although few graves are known from sixth-century Denmark, those that have been identified are cremations.[39] 

Cremation (urned or un-urned) occurred in roughly equal proportions with inhumation in western Norway but dominated in eastern and central Norway.  As was frequently the case in Germania, inhumations were usually better furnished.  Burials were usually covered by mounds, earthen in the east, of stone and earth in the west, often bigger than they had been in earlier periods.  Further regional variation is found in the fact that burial often took place in large cemeteries in the east, whereas in the west it occurred in smaller groups associated with individual farms.  The dead were placed in cists, with the corpse and, presumably, personal belongings at one end, and other grave-goods at the other.  Weapons are again rare before about 400, but other grave-goods, including imports from the Roman Empire and other areas, could occur in quite lavish quantities with males and females.[40]  Above-ground stone funerary monuments, whether single stones decorated with images and/or runic inscriptions or placed around a grave or graves as ‘ship settings’, are common throughout Scandinavia. 

Moving south again, within Germania Magna, furnished inhumation really began to be common during the fifth and sixth centuries, although cremation was practised throughout the period and beyond.  Along the North Sea Coast, clusters of well-furnished inhumations within large cremations appeared around 400.  A lavish burial at Fallward (D) contained the chair decorated with imperially-inspired ‘chip-carved’ motifs mentioned earlier, as well as a ship and other goods.[41]  These burials should be seen as responses to crisis, much like the lavish burials of the Roman Iron Age just mentioned.  That crisis deepened, however.  A large number of burial-grounds were abandoned, although the extent has probably been overestimated.  Later fifth- and sixth-century, furnished inhumation cemeteries are known from the region.  Similar to the type by then common throughout northern Europe,[42] they are nevertheless less neatly arranged and follow less clear patterns of development than those in Merovingian Gaul or the Rhineland.  Archaeologists have therefore hesitated to refer to them as Reihengräberfelder (row-grave-cemeteries) like the others.  As in the former provinces, these probably reveal local power structures more open to competition than had hitherto been the case.  Close study shows that, as in other post-imperial regions (though again with significant regional peculiarities), age and gender played important parts in structuring the grave-goods custom.  Very lavish late fifth- and early sixth-century burials are largely absent in this region, contrasting somewhat with the period around 400.  Nonetheless, the subjects of inhumations were clearly distinguished in death from the majority of the (cremated) population, adding to the relative lack of investment in lavish grave-goods to suggest that the Saxon élite was relatively secure in its position.  It still had to display its distinction in burial, but funerals were seemingly not the locus for particularly fierce competition.[43]

Further up the Elbe in, by now, probably Thuringian regions, the decades around 400 also saw social stress.  The ‘Niemberger Group’ of furnished inhumations appeared in the last quarter of the fourth century[44] and, like the Haßleben-Leuna Group in the same region a century earlier, show some families responding to social pressure by using a funerary ritual that more obviously distinguished them from their neighbours.  This stress is visible in a contemporary decline in the quality of local craftsmanship.[45]  Further significant change occurred in the mid-fifth century.  Some well-furnished burials are known, some under mounds and others accompanied by horse-burial.  The practice of skull-deformation, sometimes – albeit questionably – associated with the Huns is attested.[46]  These developments, alongside the appearance of new forms of material culture, might suggest another phase of competition and stress, possibly associated with the emergence of a powerful Thuringian kingship after the collapse of Hunnic hegemony.

Changes in burial similarly manifest crisis in the Rhineland and in the Frankish and Alamannic homelands.[47]    Unlike the North Sea regions or other areas further into Barbaricum, these crises apparently occurred slightly later in the fifth century.  In Frankish areas, furnished inhumation began to appear in the early fifth century.  After a relative hiatus in the mid-century, more visible in male than female burials, lavishly-furnished inhumation re-emerged from c.475 and, by c.525 it had become a common rite, practised by whole communities, probably encompassing several settlements, sharing large focal cemeteries.  The numbers and types of grave-goods were governed by the deceased’s age and sex.  The most lavish burials were those of people whose death would have caused stress within local community politics, requiring smoothing over with elaborate funerary ritual.  These were often younger adult women – whether recently betrothed or married or the mothers of young children – whose death would bring a marriage-alliance into question and raise issues of inheritance where young children were involved, and mature adult males – family-heads whose sons had not yet established themselves, with a household, within local politics.

This was a period of ongoing change in Alamannic territory, although a consistent picture is currently quite difficult to establish.  Some important developments took place c.400.  Changes in burial-custom occurred and, as in Saxony, some cemeteries were abandoned.  Many new sites, however, began to be used.  At Lampertheim (D) the cremation cemetery ceased to be used around 400 and an inhumation cemetery quickly took its place.  As elsewhere, inhumation increased in popularity from the mid-fifth century.  As yet quite small communal inhumations cemeteries were founded, wherein grave-goods accompanied the dead.  There were, as before, some lavish graves but these were rarer.[48]  This implies social change.  A breakdown of secure local power from the mid-fifth century may be suggested by the abandonment of many Höhensiedlungen (see above).[49]    

By the mid-sixth century, Alamannic cemeteries look rather like those known in Frankish territory, with entire communities participating in the furnished inhumation rite.  Here, however, the ritual involves more extreme displays than are found in the core of the area of Merovingian hegemony.  Alamannia is not the only area on the fringes of Merovingian power where inhumations are very elaborately furnished.  Helmets, body-armour and whole horses (sometimes more than one) accompany male burials, while women took particularly lavish assemblages of jewellery with them to the grave.  The most elaborate burials (from around 500) often took place in well-constructed chambers.  As in other regions, the form and numbers of grave-goods were closely related to deceased’s age and gender.[50]  Families competed by burying their relatives with appropriate assemblages of grave-goods, to the greatest possible extent. 

In what, by the early seventh century, had come to be thought of as Bavaria, fifth-century instability is once again revealed by the appearance of furnished inhumation cemeteries.[51]  Similar to those of northern Gaul, the Rhineland and Alamannia, they were founded in the fifth century and had evolved into large communal cemeteries by the early sixth.  It is noteworthy, however, that these cemeteries are found overwhelmingly on what had formerly been the imperial side of the frontier.[52]

Important changes occurred towards the end of the sixth century throughout Germanic-speaking barbaricum.  In Scandinavia they may have come slightly earlier.  The largest Norwegian funerary mound, at Raknehaugen, dates to 530-50.  Once believed to have been a cenotaph, the mound has been shown to have covered a poorly-furnished cremation.  This importantly highlights the fact that ostentatious above-ground commemoration need not tally with lavish investment in buried objects.  The two means of marking the dead could arise from quite different social or ritual demands.  From the mid-sixth century and into the seventh, during the transition to what is (oddly) known as the ‘Merovingian Period’, Norwegian burials changed somewhat.  Mounds became smaller and fewer graves are known in central regions.  In the seventh century, imported grave-goods became scarcer, but weapon burial more common.  Cremation increased in proportion to inhumation.  Seventh-century high-status sites are known, most notably at Borre on the Oslofjord (Vestfold, N), where, as at Raknehaugen, mounds cover poorly furnished cremations.  In terms of funerary display, a permanent, above-ground mark on the landscape had become more important than a transient display of grave-goods.[53] 

Swedish archaeology, from the mid-sixth century onwards, is dominated by the lavish boat burials of Uppland, especially those from Valsgärde and Vendel (the latter site giving its name to the pre-Viking period in Swedish archaeology).[54]  The Vendel and Valsgärde boat burials were of men but the opposite phenomenon was recorded at Tuna in Badelunda (Västermanland, S) with its lavish female boat burials.  Near Valsgärde, at Fullerö, the burials in the boats were cremations.  The Scandinavian boat burial custom was thus quite diverse.  At Gamla Uppsala is another series of large barrows, long associated with particular semi-legendary local kings (as was also the case at Raknehaugen).  We should not take such identifications too seriously, though the Uppsala barrows clearly demonstrated considerable local power.  As with some contemporaneous Norwegian mounds, the graves covered were poorly furnished.  Some mound burials of Norwegian type are found in the west coast (e.g. in Bohuslän) and northern regions of Sweden.

The burials of the other members of society were less extravagant.  The Valsgärde boat burials were surrounded by cremations of women, with few grave-goods other than the burnt remains of dress ornaments.  There were also some robbed cist burials.  At Tuna in Badelunda, where the boat burials were of women, the surrounding cremations were apparently of men.  Some cremations nevertheless contained remnants of decorated metalwork whose quality equalled that interred in the barrow graves.  Inhumation in coffins is also recorded.   Most Swedish cemeteries are quite small and located on the best arable and pastural land.  The Mälar valley boat burials were, however, sited on the periphery of the settled area.  On the island of Gotland (S), there are no boat burials but from the mid-sixth century onwards graves were unusually well-provided with grave-goods.  75% of male burials included weaponry, representing a significant change from the region’s earlier burial customs.[55]

Further south, especially on the Rhineland and South-Western German fringes of the Frankish realm, change in burial custom also included increasing use of above-ground mounds from c.600.  Local elites apparently began to separate themselves in death from the remainder of their communities.  Sometimes their graves lay on the edge of larger cemeteries, as at Kirchheim am Ries (D), but elsewhere Separätfriedhöfe (separate cemeteries) were established, at Niederstotzingen (D), for example.  Otherwise, as in Norway, burials became more nondescript and grave-goods gradually diminished throughout the seventh century.  In general, grave-goods deposition, particularly in male burials, persisted longer in the north.  On the Rhine and in south-west Germany, grave-goods had generally fallen out of significant use by c.675, whereas in northern Germany and the northern Netherlands they continued into the eighth century and beyond.   Cremation also persisted in the latter areas.  In Alamannia and Bavaria, the decades around 600 and, especially, the seventh century saw the creation of churches, often associated with well-furnished burials, as part of the same processes as produced the construction of barrow burials.[56] 

Northern Britain

Much interesting work has recently been done on northern British burial customs but the topic remains difficult because of their generally undiagnostic character.[57]  From early in Late Antiquity, inhumation in cists (stone lined trenches) increased in frequency.  Attested as early as the third century in the Isles, it became more significant, especially during the fifth century, when more communal – if still small – inhumation cemeteries are more visible.  Without grave-goods, chronology is vague and frequently only established through Carbon14 dating.  Inhumation’s introduction has been attributed to Christian influence but the bases for the assumption are weak.  More plausible, though not entirely satisfactory, is the linkage of the rite’s growing popularity, as in parts of Germany, to the influence of the Roman Empire.  It must be said, though, that elements of the rite can be interpreted as continuations of earlier burial practices.[58] 

During late antiquity, especially in the sixth and seventh centuries, further developments occurred which, in some regards, parallel those in Germania.  Above-ground monuments – cairns and barrows of various shapes – began to be employed.  The famous symbol stones were also introduced.  Their dating remains tendentious and as yet none has been found in definite, primary association with a burial.  Nonetheless, that the earlier, ‘Phase I’ symbol stones probably originated as grave-markers some time in the later sixth century remains broadly uncontroversial.  The enigmatic symbols are much disputed.  No specific interpretation commands wide assent; most are controversial.  That said, the idea that they convey information about names, individual or familial, seems plausible.  In western Scotland, the seventh century saw the enclosure of burial sites, which might be associated with the church’s establishment in this region.[59]


Late antique Irish burial is also fairly nondescript.  Unfurnished inhumation was popular, again producing the attendant problems of establishing a chronology.  Early Iron Age burial in Ireland was fairly varied but generally archaeologically ephemeral, possibly including exposure and excarnation.  Cremation was also common, though the remains may have been deliberately dispersed.  Later in the Iron Age, burial became more archaeologically visible, under various forms of monument, and inhumation became more common.  As elsewhere this might represent Rome’s cultural influence.  The use of delineated areas for burial in Late Roman Iron Age Ireland makes Christianity’s precise impact difficult to ascertain.   Certainly, churches and monasteries employed bounded graveyards but earlier custom makes it difficult to maintain the equation, common in Ireland and northern Britain, between cemeteries with boundary banks and ditches and those without with Christian and unchristian respectively.[60] 

Another similarity with northern Britain is the use of above-ground stone markers.  In Ireland these are famously marked with inscriptions in the ogham alphabet (and later in Latin too) generally commemorating the deceased.  Doubtless originating in some way as a result of Roman influence, they probably began in the fifth century, although some are possibly earlier, and continued through the period.[61]

Hoards and Bog finds

Our final category of archaeological site is the votive or ritual deposit of material.  The most famous late antique representatives of this class are doubtless the bog-deposits from Denmark and the far north of Germany.  Nydam, Ejsbøl and Illerup (all Dk) are justly celebrated for having provided some of the best evidence for Iron Age society in the region and for the quality of their analysis.[62]  Ritual bog-deposition had a long history in Denmark and elsewhere.  The famous lavish finds of the late Roman Iron Age represent the matériel from defeated armies.  These deposits are difficult to explain.  Unlike grave-goods deposition, where the creation or maintenance of individual families’ status was at stake, it might be that the votive bog deposits represent more of a communal rite.  We should nonetheless not see this as in any way egalitarian.  That the deposition was organised is suggested by the finds themselves.  These rituals centred upon the destruction, the removal from circulation, of what, under usual circumstances, became war-booty, whose distribution was controlled by war-leaders.[63]  Here, they disposed of large quantities of potential loot, one imagines in the form of a gift to the gods.  In so doing, they demonstrated their authority, removed from circulation valuable items that other warriors might have used as gifts, and enhanced the prestige and value of the remaining items, which they could bestow upon their followers.  There is no reason to suppose that all of the defeated army’s material was thrown into the marsh.  Furthermore, mounted men escape from a battle more easily, possibly leading to an under-representation of horse-furniture.  These points should make us sceptical about attempts to deduce detailed military organisation from the bog finds,[64] from the exact numbers of specific types of weapons and other forms of equipment.  Be that as it may, that the fourth-century bog-finds clearly illustrate the northern barbarian leaders’ ability to raise substantial armed forces and thus their political power and authority is incontrovertible.  Later in our period, the custom declined, which – when taken alongside other archaeological evidence, suggests steadily increasing political authority in Denmark.

Denmark’s precise physical geography has enabled these finds to be carefully excavated, reconstructed and analysed.  Elsewhere, analogous finds take the form of ‘hoards’.  Some such deposits might initially have taken a similar form to that of the bog-finds, but subsequent drying out of ground or changes in watercourses have left them resembling the simple burial of material.  Caution is thus required in treating these kinds of deposit as necessarily different.   One of the most famous is hoards is that from Traprain Law (southern Scotland), largely composed of late Roman silver and belonging to the decades around 400.[65]  The material’s origins are debated.  Loot from raiding has been suggested but it may be more profitable to interpret the hoard as a diplomatic payment from Roman authorities to a frontier leader, especially in the turbulent era of Roman civil wars between 383 and 420, which involved several withdrawals of troops from Britannia.  Indeed the distribution of later Roman finds in northern Britain and Ireland suggests concentration on fewer sites, and frequently in hoards.  The interpretation of hoards varies between the ‘functional’ – the burial of wealth for later retrieval – and the votive – gifts to the gods.  A single explanation is unlikely to work everywhere.  One might hazard an intermediate interpretation, as suggested for the Danish bog-deposits, wherein wealth was ritually taken out of circulation but in a very ‘practical’ way, to enhance the value, as gifts or rewards, of what remained.

Part 3 is here.

Notes (see Bibliography)

[1] Waterbolk (1999); Mejdahl & Siemen (2000); Hamerow (2002), pp.12-38; Brather (2009), pp.30-39; Dijkstra (2011), pp.191-222.
[2] Hamerow (2002), pp.31-34; Tipper (2004)
[3] van Ossel & Ouzoulias (2000).
[4] Lowe (1999), pp.18-21.
[5] Thasing (2013). For a recently-excavated example, see Bestemann et al. (1999).
[6] Reuter (2003); 2005).
[7] Steuer (2005).
[8] Wijster: Van Es (1967); Feddersen Wierde: Haarnagel (1979); Vorbasse: Hvass (1983)
[9] Memorial Colloquium Haarnagel (2010); von Carnap-Bornheim (2015).
[10] Verlind & Erdrich (1998).
[11] For Höhensiedlungen see now Eger (2014)
[12] Borg, Näsman & Wegraus (ed.) (1976); Näsman & Wegraus (ed.) (1979).
[13] Myrhe (2003), p.75. Recently, Viberg et al. (2014).
[14] Nieuwhof (2011)
[15] Hamerow (2002), p.94.
[16] Hamerow (2002), pp.15-18.
[17] Myrhe (1985); (1997); Stylegar & Grimm (2005); Wickler & Nilsen (2012).
[18] Dankirke: Hansen (1989); Jensen, S., (1991); Helgö: Fischer & Victor (2011)
[19] Gudme/Lundeborg: Nielsen, Randsborg & Thrane (ed.) (1994); Stilborg (1997). Uppåkra: (2003) Larsson & Hårdh (ed.) (2003).
[20] Welch (2000) for a brief, useful survey.
[21] For Scottish rural settlement see the recent overview in Foster (2014); Dunwell & Ralston (2008), pp.133-40. Other recent overviews of Northern British archaeology include Driscoll, Geddes & Hall (ed) (2011); Clarke, Blackwell & Goldberg (2012).
[22] Again, a convenient recent overview can be found in Foster (2014). Ralston (2004); Driscoll (2011), pp.264-6.
[23] Crone & Campbell (2005).
[24] Harding (1997); Campbell (1999), pp.23-25.
[25] Campbell (2007)
[26] On Irish ringforts, see, most recently, O’Sullivan et al. (2013), pp.50-53. Also Edwards (1996), pp.6-33; Stout (1997);
[27] Stout (1997), pp.22-31.
[28] O’Sullivan et al. (2013), pp.58-62; Edwards (1996), pp.35-41
[29] O’Sullivan et al. (2013), pp.62-64.
[30] Edwards (1996), pp.41-48.
[31] O’Sullivan et al. (2013), pp.88-101.
[32] O’Sullivan et al. (2013), pp.321-3.
[33] Táin Bó Cuailnge (The Cattle-Raid of Cooley)
[34] On the Irish economy see the excellent accounts in O’Sullivan et al. (2013), pp.101-12 (settlement economy), 179-214 (farming), 215-46 (craftsmanship), & 247-81 (trade).
[35] O’Sullivan et al. (2013), p.251. See also Freeman (2001).
[36] O’Sullivan et al. (2013), pp.255-66.
[37] In English, Todd (1987) remains a useful descriptive summary of Roman Iron Age archaeology but many of the interpretations are now dated. There is as yet no detailed book-length overview in English to replace and update this. Todd (1997) for a briefer version.
[38] On these prestigious burials, see, in English, Todd (1987), pp. 46-47, 57, 71.  Map: Todd (1987), p.40. Generally: Becker (2009); Gebuhr (2009); Quast (2009); Abegg-Wigg & Lau (ed.) (2014); Zakrzów/Sakrau: Quast (2014) Mecklenberg: Voss (2009).
[39] Donbaek: http://slks.dk/fileadmin/user_upload/KUAS_Donbaek_JA_low.pdf (accessed 10/08/2016). Hjemsted: Ethelberg et al. (1986); Praestestien: Halsall (2000), p.101; Fischer (2014).
[40] Norwegian burials: Solberg (2009); (2015); Stylegar (2014)
[41] Schön (1999).
[42] Todd (1987), pp.88-89, and refs.
[43] Capelle (1998), pp.33-50; Siegmund (2003)
[44] Schmidt (1983), pp.515, 518.
[45] Ibid., p.536.
[46] Ibid., p. 544. Crubézy (1990) for reservations about the Hunnic connection.
[47] Böhme (1974); Theuws & Hiddink (1996), p.78. 
[48] Alamannic burial: Quast (1997) is a useful survey. Christlein (1991), pp.50-62.
[49] See above, n.59.
[50] Analysis of later Alamannic burials: Jørgensen, Alt & Vach (1997); Donié (1999); Brather (2004); Obertová (2008); Hausmair (2015).
[51] Koch (1968).
[52] Menghin (1990), p.80, fig.65.
[53] See above, n.88. Raknehaugen: Myhre (2003) p.87; Borre: Myhre (1992).
[54] Lamm & Nordstrom (ed.) (1983). Klevnäs (2015).
[55] Rundqvist (2003).
[56]Kirchheim: Neuffer-Müller (1983). Niederstotzingen: Paulsen (1967); end of lavish burial: Stein (1967); an example of a founder burial under a church: Burnell (1998).
[57] Maldonado (2013) is essential. Windlow (2011).
[58] Dunbar & Maldonado (2012) for case study and discussion.  See also Fraser (2009), pp. 36-37.
[59] On barrows and cairns, in addition to Maldonado (2103), see Ashmore (1978-80).  On symbol stones see Henderson & Henderson (2004); Gondek & Noble (2011).
[60] O’Sullivan et al (2013), pp.283-317
[61] Swift (1997).
[62] For good recent studies of the phenomenon see Ilkjaer (2002);  Jørgensen, Storgaard & Gebauer Thomsen (ed.) (2003); Abegg-Wigg & Rau (ed.) (2008); Carnap-Bornheim (2014).
[63] See Gregory of Tours Historiae 2.27 for a late fifth-century Frankish analogy.
[64] E.g.von Carnap-Bornheim (2015).
[65] Hunter & Painter (2013).