Proceedings of the Prehistoric Society

http://journals.cambridge.org/PPR

Additional services for Proceedings of the Prehistoric Society:

Email alerts: Click here

Subscriptions: Click here
Commercial reprints: Click here
Terms of use : Click here

Ritual Hoarding in Migration-Period Scandinavia: A Review of Recent

Interpretations.

John Hines

Proceedings of the Prehistoric Society / Volume 55 / January 1989, pp 193 - 206

DOI: 10.1017/S0079497X00005399, Published online: 12 June 2014

Link to this article: http://journals.cambridge.org/abstract_S0079497X00005399

How to cite this article:

John Hines (1989). Ritual Hoarding in Migration-Period Scandinavia: A Review of Recent Interpretations. . Proceedings
of the Prehistoric Society, 55, pp 193-206 doi:10.1017/S0079497X00005399

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/PPR, IP address: 150.216.68.200 on 01 Jul 2015

Proceedings of the Prehistoric Society 55,1989, pp. 193-205

Ritual Hoarding in Migration-Period Scandinavia:

A Review of Recent Interpretations
By JOHN HINES 1

CONTENTS

MAJOR CONTEMPORARY WORK ON THE SUBJECT Page 193

ASPECTS OF THE SCANDINAVIAN MIGRATION PERIOD 194

INTERPRETATIONS OF THE SCANDINAVIAN HOARDS, ESPECIALLY WEAPON DEPOSITS 194

MIGRATION-PERIOD HOARDING, ESPECIALLY OF BRACTEATES 197

SYNTHESIS AND RE-INTERPRETATION IN SOCIAL AND ECONOMIC TERMS 199

T o THE VIKING PERIOD — T H E EXTENT OF DISCONTINUITY 202

IDEOLOGY AS A FACTOR 202

Non-funerary deposits of an apparently ritual character are a persistent, often very prominent aspect of all periods
of Scandinavian prehistory from the Neolithic to the later Iron Age (post 600 AD). The hoards of the nearer end of
this series have recently been brought within theoretical and analytical studies from a variety of modern
perspectives differing in ideology and specialism. This paper offers a critical review of those studies in the light of a
detailed case-study from the nearer end. Harmonies can be found between the shaping force of economic and social
factors posited by Richard Bradley's model and the evidence of this case, although, perhaps inevitably, the effects
of these may appear more complex and even quite different from what a description of the general model can
encompass. More cautionary conclusions reached which have important implications within the construction of
general theory are that greater care ought to be exercised in identifying the establishment of polities and
socio-economic crises from this and other contemporary categories of material, and that the specific content of
ideology, particularly religious concepts, which affects the fact and the form of ritual hoards, while probably
incapable of being built into a general model like Bradley's on the same implicative basis as specific types of
economic and social structures, merits a more prominent place in studies of the topic.

MAJOR CONTEMPORARY WORK ON THE SUBJECT ing _ i n t h e l a t e r millennia of prehistory (Bradley

In a series of recent publications Richard Bradley has 1984a; 1985; 1987a; 1987b). In its most recent form,
developed a set of propositions which amount to an this theory proposes three progressive stages of hoard
extensive theory which attempts to account for the development. The initial stage, focused upon the Neo-
long-term development of non-funerary sacrificial lithic, sees a largely 'unitary' system of deliberate depo-
deposition — commonly called votive or ritual hoard- sition, without significant, non-random differentiation
in the association of particular hoarded objects with
particular locations. In the Bronze Age long-distance
• School of English Studies, Prifysgol Cymru, Coleg Caerdydd, P.O. t r a d e i n prestigious materials appears as an important
Box 94, Cardiff, Wales, CFI 3XE economic factor. Central and liminal zones may now be

193
 THE PREHISTORIC SOCIETY

perceived in the distribution of hoards, and it is argued tion of material is probably sacrificial or ritual where
that objects of symbolic value within these zones revert the context makes retrieval difficult, e.g. water or wet-
to a utilitarian, economic value in the boundary areas. lands. This criterion establishes a class of find-groups
Moving into the Iron Age 'the long distance movement regularly subject to votive deposition which by implica-
of metalwork lost some of its significance' (1987b, tion identifies a range of further topographical contexts
357). The differentiation of central and liminal zones used for such practices. The provision of grave goods is
through hoard distribution however does not disap- a particular type of ritual deposit which is not con-
pear, but at this stage is interpreted rather in terms of sidered per se here although it cannot be entirely separ-
political entities: 'As major political territories ated from ritual hoarding. The sacrificial hoards of
developed, their boundaries were strengthened by the Migration-period Scandinavia usually contain material
provision of such deposits . . . ' (1987b, 360). Con- which may be attributed with high material or symbolic
comitant developments in the material hoarded are value, such as precious metal and sword- or horse-
noted, in general terms the emergence of material pro- associated objects.
duced especially for votive purposes. The use of coinage Germanic Migration-Period chronology is quite
is considered in this light, with a 'tantalizing possibility' strong compared to both earlier and later periods, in
of the conceptualization of a contract and payment terms both of internal seriation and absolute dating.
between man and gods, and an increasing resemblance Scholarly controversy can sparkle, if not exactly rage,
between religious and commercial transactions being over a matter of a decade (Hines 1984, 16-32.; Welch
noted (1987b, 361). 1987, 2,55-59). The period covers about two centuries
This theory thus correlates the composition and dis- from the later 4th to the later 6th centuries AD and is
tribution of such hoards with macro-developments in bounded by clear material-cultural watersheds. Inter-
economic and social structure — the former appearing related hoard- and grave-finds produce a lot of material
as an essential factor before the latter — and tentatively from much of northern Europe; settlement-site evi-
hints at cognitive consequences concerning man's per- dence lags quite some way behind, but is improving.
ception of his relationship with the gods as the Roman One point that is readily evident is that Migration-
Period approaches. Period culture is far from uniform chronologically or
The near end of the chronological sequence is the regionally even within Scandinavia. In and around the
post-Roman Scandinavian Iron Age. In all we have a Migration Period, from the 4th to 7th centuries, there is
prehistorian's long view of a sequence running into the extensive change and variability in hoarding practice in
partly-lit zone of early Medieval archaeology and his- Scandinavia. The position in the early 4th century is
tory. Alongside the same trend in contemporary unusually simple and regular, with the great bog de-
archaeological thought as that which Bradley repre- posits of weapons like Torsbjerg and Nydam dominat-
sents, recent finds on the Danish island of Fyn have ing this mode of expression (Brandsted i960,180-236;
focused another group of archaeologists' attention on Hines 1989). The objects hoarded and types of site used
thisfindmaterial and its possible interpretation in terms diversify considerably in the 5 th century and the
of economic, social and religious history (Thrane diversification continues in the 6th. There appears to be
1985). Predictably, a variety of perspectives emerges, a rapid decline in the practice of depositing such hoards
some attributable to the difference of 'prehistorian' and in the late 6th century, and no certain evidence of
'medievalist', some attributable to contemporary ideo- continuity to the Viking Period (9th-ioth century) with
logical allegiance. It is hoped that a critical view back its great silver hoards (Graham-Campbell 1982).
along Bradley's model, based upon a detailed case- Periods of marked change seem to lie close to the
study from one of the later periods it seeks to cover, will boundaries of the Migration Period although they are
be of interest to readers of this journal, revealing both not perfectly correlated with them, and within the
points of agreement and points for modification. Migration-Period practices are diverse and shifting.

ASPECTS OF THE SCANDINAVIAN MIGRATION PERIOD INTERPRETATIONS OF THE SCANDINAVIAN HOARDS,

There is no essential difference in the problems of ESPECIALLY WEAPON DEPOSITS
identifying thefind-categoryin question for this period The most comprehensive published survey of this Scan-
relative to earlier ones. Non-funerary deliberate deposi- dinavian hoard-material — identified by less restrictive

194
 8./. Hines. RITUAL HOARDING IN MIGRATION-PERIOD SCANDINAVIA

criteria than would be accepted now — is that of support. A major reflex in human cognition of an
Geisslinger (1967). Geisslinger was particularly extension of this system to the relations between gods
anxious to emphasize the permissibility of a diversity of and men is to anthropomorphize the gods; to counter-
explanations for formally diverse material remains even act the grotesque, bizarre, mystifying and terrifying
within this single broad category. But individual prac- image they may present, as when Ooinn in Norse
tices not only require their own particular explanations; mythology sacrifices his eye and deposits it in a well in
at a deeper level the phenomena of change and diversity order to see into the future (Turville-Petre 1964,
within a limited area need explanation. Explaining, or 63-64).
at least seeking some way of comprehending diversity is The proposition that these weapon deposits are sig-
quite different from reducing diverse phenomena to a nificantly located in the border zones of emerging
single explanation. A dominant current in more polities is one fraught with difficulties. Bradley (1987b,
abstract analyses of Migration-Period culture in north- fig. 5) supports this view by plotting distribution against
ern Europe at present is to look for patterns or symp- settlement areas mapped by Parker Pearson (1984), the
toms of change anticipating the watershed which was extent and concept of which is largely traceable back to
the end of the Migration Period, in particular to try to Jankuhn. To say the least there must be doubts about
trace processes of state formation or the emergence of the identification of these areas and political units. The
stratified or aristocratic societies (e.g. Hedeager 1980; restriction of the areas of settlement in Parker Pearson's
Sa'rlvik 1982; Arnold 1988; Ringtved 1989). It is no maps is exaggerated, as a comparison with Genrich's
coincidence that the primary feature that Bradley cemetery distribution maps shows (Genrich 1954,
emphasizes in the later phases of his sequence of Karten 1,5,6 and 8). Parker Pearson was committed to
development is the location of hoards in the border a view of economic decline and population loss in this
zones of emerging polities. The regularization and con- period in southern Scandinavia which is only supported
trol of material resources and commerce is seen as an by the negative evidence of absence of settlement and
inseparable concomitant of this social development. cemetery sites. Attempts to reconstruct polities in this
All the scepticism that the last thirty years of archae- period work primarily with three factors: recorded
ology can muster has failed to dislodge the conclusion identities of ethnic groups, which of course need not be
that the great weapon deposits of southern Scandinavia coterminous with political units, archaeologically sep-
of the later Roman Iron Age — at Torsbjerg, Nydam, arable material culture zones, which in some cases can
Kragehul, Ejsbol and Illerup — are the equipment of be linked to named ethnic groups, and fragmentary
defeated armies sacrificed in sacred pools to war deities evidence of political organization, such as the palisaded
by the victors in battle (0rsnes/Engelhardt 1969-70; ditch and rampart Olgerdiget which arguably marks
Ilkjaer and L,0nstrup 198Z; Hines 1989). It may be too the northern limit of the Angles' territory in the later
simple to infer in the minds of those who made these Roman Iron Age. These factors achieve prirnacy by
deposits no more than a concept of fulfilling a contract, coinciding nicely to provide us with a cultural zone and
a retrospective payment or thank-offering for a victory a bounded political unit centring on and not greatly
granted. It is tolerably clear that gift-exchange, particu- overspilling the homeland of the Angles of historical
larly of the spoils and rewards of warfare, was perceived sources of the 1st century AD onwards (Hines 1984,
as one of the fundamental bonds of the relevant warrior 1-14). 'Anglian' culture spreads also onto the neigh-
societies both before and after, and therefore plausibly bouring island of Fyn, but whether those who used it
through, this period. In early Medieval literature the there were called, by themselves or others, Angles, and
practice is presented as being elevated by and into an shared any political affiliation with the mainlanders, is
ideology, as a practice laudable in itself and not simply a unknown. Three major weapon deposits, Torsbjerg,
socially functional device (Tacitus, Germania, 13-15; Nydam and the smaller Kragehul, lie within this cul-
Shippey 1978,18-24). The primary axis of exchange is tural zone, but in geographically distinct districts. It
between comitatus and lord/chief, the mutual giving takes some imagination to regard Torsbjerg Mose at
and receiving of gifts placing an obligation upon the Siider-Brarup as conspicuously liminal. This site is 6 km
receiver of future loyalty and support to the giver. In as the crowfliesfrom the shore of the Schlei, and is the
this respect the gods are brought within the social location of a market centre of recent times on a nexus of
system, but the emphasis is as much if not more upon communication channels. A large cremation cemetery
their continued support for the group as upon their past contemporary with the hoards has been found here

13*
 THE PREHISTORIC SOCIETY

(Bantelmann 1988). More convincingly distributed

around the margins of the settled area are the 'bog
people', human victims of execution or sacrifice (Glob
1969, 144-92; Munksgaard 1984). The deposition of
these, however, begins at a much earlier phase than the
weapon deposits, at a stage at which the concept of
political border zones is not so appropriate. Their
placement may better represent an embodiment of the
boundary between known and unknown, life and
death, or used/productive and unused land. North of
the Angles, somewhere, were Jutes. The Jutish penin-
sula can be divided into a number of cultural zones, but
whether any particular one, or more, or all of these were
produced by people who were called Jutes we do not
know.
Some case can be made for correlation between such
culture zones and the great weapon deposits which may
support the inference of integrated social organization.
While three prominent sites lie within the Anglian
culture zone, Ejsbol, Illerup and Vimose each lie in
separate zones outside it (fig. 1). As is apparent from the
map there is no case that these sites are specially corre-
lated with the borders between the perceived culture
zones. It is clear that the major hoard sites were re-used
for deposits of war-gear; typically there is evidence of
three or four discrete deposits over a period of up to
three centuries between the late 2nd and early 6th
centuries. They were therefore not randomly selected
locations, conveniently close to battlefields. Jankuhn
(1977) has therefore suggested that Torsbjerg at least
was a zentrales Stammesbeiligtum, a central national/
tribal holy place. Three major sites within the Anglian
culture zone, and nearly a dozen minor sites around the
major ones, give an impression of a divided war zone in
eastern Jutland and Fyn, in spite of claims for the
relatively distant origin of material deposited in these
hoards. There is a case to be made that the creation of
these deposits and particularly the re-use of sites reflects
Fig. I social cohesion at a level below the ethnic and/or politi-
Ritual hoards of weapons, and cultural zones, in cal unit, perhaps rival groups within these entities, with
Schleswig, Jutland and Fyn, of the later Roman Iron Age home districts (Hines 1989).
and Migration Period. Not all deposits are known to be Military equipment dominates the range of material
of battle-used weapons. Deposits of detached fittings,
such as pommels, scabbard chapes, buckles, etc. are hoarded in the 3rd and 4th centuries. There is a little
excluded. Zonal boundaries based on Neumann (198Z, evidence of the marginal contemporary admixture of
fig. 36b) and Genrich (1954, Karte 8 for the Anglian other categories of material: women's dress accessories
zone on Fyn in the beginning of the Migration Period). and agricultural tools. In the 5 th century substantial
Genrich's Anglian zone on the mainland does not changes in hoarding practice appear, although most of
extend to the North Sea, but is bounded to the west by a the major sites remain in use. At Ejsb0l two deposits are
hybrid and rather sparse Nordseekustengruppe. Major
weapon deposits: i,Torsbjerg; 2,Nydam; 3,Kragehul; known: one of our most comprehensive and coherent
4,Ejsb0l; 5, Vimose; 6, Illerup weapon deposits, of the mid-4th century, Ejsb0l-North,

196
 8./. Hines. RITUAL HOARDING IN MIGRATION-PERIOD SCANDINAVIA

and, just metres away, Ejsb0l-South, a 5th-century other hoards which for simplicity's sake one may sweep
deposit of quite precious, well-crafted bronze- and into a miscellaneous category, but which include such
silver-gilt fittings of a predominantly military character, extraordinary objects as the two gold horns from Galle-
including many scabbard fittings, belt-buckles and hus, Jutland (Brandsted i960, 2.87-94; Stenberger
clasps (0rsnes 1963). The two deposits are probably 1964, 472-73 and 488-97; Hagberg 1967, 67-78). As
separated in date by between 50 and 100 years. Ejsbel- diverse as the range of hoard-types is the range of
South represents a new type of hoard. A contemporary explanations that has been offered for them, hoard by
deposit of scabbard fittings, with one belt-buckle, is hoard or group by group. A group of currency hoards
known from Nydam (Nydam II); and from later in the on the island of Bornholm, for instance, has been taken
5th century a similar hoard comes from Sjorup in as evidence of a catastrophic period of instability, pre-
Skane, recorded as being from the shore of a lake rather sumably a period of war, on that island in the later 5th
than necessarily having been deposited within it. Simple century. Ulf-Erik Hagberg, in discussing about 100 or
commentary on such finds of the 5th century has so horse skeletons found at Skedemose on Oland
implied that they represent no more than a shift from together with a smaller number of bridle-fittings, is able
comprehensive weapon deposits to pars pro toto offer- to put together from various sources a compelling
ings, representing the same material. But these particu- picture of seasonal fertility and horse-mating rituals
lar 5th-century hoards show no evidence of battle use, involving horse races and sacrifices at cult sites (Hag-
and further tend to be distinguished by the remarkable berg 1967, 79-84).
stylistic unity of the material deposited, so much so that The deposits of gold pendants designed for suspen-
major Migration-Period art styles take their names sion from the necklace, known as bracteates, are
from the Nydam and Sjorup finds. Geisslinger conse- arguably the best material to concentrate on in a selec-
quently suggested that they represent some group offer- tive study because they form such a widespread and
ing made by a comitatus at its home site. Alongside well-populated set. It is beyond reasonable doubt that
these finds, comprehensive deposits of battle-used these artefacts were objects endowed with high sym-
weapons may still have been made as late as the early bolic significance, although it is hard to state exactly
6th century at Illerup and Kragehul. A continuing habit what that significance is. They carry symbolic motifs,
of depositing separated military fittings such as sword such as the swastika and triskele, which are separable
or scabbard mounts, which are increasingly found from the undoubtedly decorative motifs used on these
singly, seems to continue through the 7th century, objects, and not infrequently carry runic inscriptions
finally represented by the Bildso sword, datable by the interpretable in a broadly amuletic sense. Recurrent
Style D ornament on the pommel, which places it in the scenes which are quite untypical in composition of
late 7th or first half of the 8th centuries (Brandsted contemporary Germanic art on other artefact-types
i960, 300; 0rsnes 1966, 207-41). distinguish some classes of bracteate, and there is a
reasonable probability that these are iconographic
although their interpretation is controversial. They are
MIGRATION-PERIOD HOARDING, ESPECIALLY OF more than mere trinkets; all of the nearly 700 examples
BRACTEATES now known from Scandinavia are of gold (Mackeprang
Military material however has a less conspicuous place 1952; Axboe 1982; Hines 1984,199-203 and 236-40;
within a very diverse range of material deposited in Hauck 1985).
hoards in the Migration Period, particularly from about The symbolic character of these artefacts extends to
the middle of the 5th century onwards. The material their use. All the Scandinavian bracteates were dress-
within these deposits can be divided into a number of accessories, with loops by which they could be suspen-
major categories. There are dress-accessories, such as ded from a necklace. There is widespread evidence for
brooches, neckrings and a particular type of pendant, the use of the strictly non-utilitarian necklace as a
the bracteate. There is what is believed to be currency, location for amulets, not just in early Germanic society
either in the form of Byzantine imperial coinage or but far back into prehistory. Relatively few bracteates
hacksilver and ingots. There are remains of various have been recovered from the dress of the deceased in
beasts, sheep, cattle and especially horses, besides graves. Their deposition in graves is only regular in
which we may note another category of hoards of western and southern Norway and in Anglo-Saxon
horse-gear such as bridle-fittings. There are also various England, besides scattered examples in Sweden and

197
 THE PREHISTORIC SOCIETY

continental Europe. Almost uniformly the grave finds fertile land. Several of these examples, however, are
identify the bracteate as a female dress-accessory. Evi- clearly of newly ploughed land, previously agricul-
dence of wear, especially on the loops, and of repair, turally marginal because it was boggy, or heath, or
implies that the bracteates were actually used as pen- stonefield. The latter two categories occasionally figure
dants over their whole geographic range. independently in the list of recorded categories. A sur-
Very rare indeed is the confirmed recovery of brac- prising proportion, nearly 4% of finds, are recorded in
teates from settlement sites. The point is worth stressing woodland or having been made upon the removal of
since a tendency is detectable in contemporary Danish trees; these too may fall into the category of agricul-
archaeology to track the elusive Migration-Period turally marginal land, although on more than one
settlement pattern via treasure hoards (Axboe 1987, occasion the trees involved were in an orchard. A final
76; Thrane 1987; 1989, 39-41). Four Frisian terpen recurrent group of contexts is a small percentage of
have produced such objects. There are two examples finds from banks of sand, gravel or earth and from
worth serious consideration from Scandinavia: two gravel pits.
fragments from the well-known production site at The numbers of instances recorded under most of
Helgo, central Sweden, and a hoard from Gudme on these categories are too small to form any realistic
Fyn (Axboe 1982, 72.-73; Poulsen 1987). At the latter, model of developments in the use of bracteates as items
recent excavations (Axboe, pers. comm.) have con- for deposition, although certain features, such as the
firmed that nine bracteates and other valuables appear easterly distribution of known coastal deposits, are
to have come from one post-hole of a timber-framed striking. It may however be affirmed that the reliably
building on a site in the area called Gudme II. identified bogfinds of bracteates form a different group
Graves and settlements account for less than 10% of from the ploughfind group in terms of distribution,
the 350 or so bracteate locations in Scandinavia. A date, and the composition of the hoards in which they
survey of the remaining findspots reveals about 29% of occur. Although bracteates are apparently in pro-
the whole with no recorded details and a further 15% duction and use from the middle of the 5 th century,
where there are merely records of the finding of the perhaps earlier, none of the bogfinds contains early
objects in the ground. Amongst the remaining finds, types, and it is possible to argue that this category
nearly half of the total, it is remarkable how frequently belongs to the 6th century. The mean number of brac-
certain categories of topographical context recur in the teates found in each of the bogfinds is fractionally less
reports of the finds (Appendix). Since these reports are than five; the mean number amongst the ploughfinds,
mostly verbal, it is obvious that one can only get general which include several particularly early pieces, is less
impressions from them; it is also the case that the than two. It is easy to see how the circumstances of
topographical contexts do not form a series of discrete, retrieval could encourage such discrepancies but hard
mutually exclusive contexts and thus somefindsmay be to believe that this alone governs them. We also find
listed under more than one category. A group of con- amongst the bogfinds interesting recurrent combina-
texts in or beside water stands out; some 10% of tions of bracteates with other classes of material
bracteate findspots seem likely to have been bog depos- (table 1).
its, the objects reported as being found in wet or boggy The combination of bracteates, beads and one
hollows, during peat or marl digging, or during drain- brooch is strikingly similar to the typical inventory of a
age work. Three finds each are reported in or beside woman's grave in this period. This phenomenon finds
rivers and lakes, and one find tentatively attributed to an echo in what Bradley describes as hoards used as
the course of a lost river. A surprising seven finds have surrogate burials in Bronze Age Wessex (1984b,
been made on or just off the coast; a possible eighth in 110-13). Presumably the explanation of this develop-
this category may be a bracteate recently reported as ment lies in the symbolic significance to the living of
found at Jaywick Sands, near Clacton, Essex. A further being able to dispose of a particular sumptuary set
20% of all recorded Scandinavian findspots form a (Levy 1982, 69-70), normally with the dead; if the need
category of land in agricultural use. This generally for such display is particularly urgent it may be incon-
means little more than that the objects were found venient to depend upon an appropriate person to die. In
during ploughing, or sometimes when lifting veg- fact, the appearance of these 'surrogate burials' in the
etables, but attracts attention as it has been suggested 6th century coincides with a marked decrease in the
that such items could be special offerings made on rich same areas in the numbers of richly-furnished graves

198
 8./. Hines. RITUAL HOARDING IN MIGRATION-PERIOD SCANDINAVIA

(Ringtved 1989, 137-40). A group offivehoards from for religious rituals (Hauck 1988). Research has shown
Jutland and south-eastern Norway which contain brac- a concentration of gold finds, including bracteates, on
teates together with beads, other pendants or finger- the coast nearest to Gudme where natural harbours
rings and also scabbard mouthpieces of a distinctive were found in the 5 th and 6th centuries. The Gudme
6th-century type is attached to this group of surrogate gold concentration obviously suggests some structured
burials. These are of interest in combining typically relationship between these sites and Gudme, some 4 km
male and female items when used as grave goods. inland (Crumlin-Pedersen 1987). Whatever sort of
Double burials are not unfamiliar in Migration-Period centre Gudme was it was certainly not unique; the
Germanic cemeteries, so we could regard these as sur- clustering of similarfindsgives evidence of several more
rogate double burials. Four of the bogfinds combine such centres of deposition (fig. 2). Little more than
bracteates only with coin or hackgold, with material 30 km away from Gudme at Odense, another pagan
that one would more readily expect to find in other place-name, Odins ve, 'Othin's shrine', still on Fyn,
archaeological contexts as currency rather than grave there is one such concentration. It is possible to pick out
goods. The combination of currency and personal items a series in Jutland with extensive and intriguing impli-
is not unusual. The distinction between the bogfinds cations: one around Ars in Alborg amt just south of
and other contexts is quite marked. While the mean Limfjorden, an area which comprises Borremose and
number of bracteates per bogfind is nearlyfive,only one Razvemose where the Gundestrup cauldron was found,
grave with as many as three bracteates is recorded, five and another about 50 km south of here in Viborg amt,
with two, and nearly twenty with one. Only in Vaster- surrounding another pair of pagan place-names, Vium
gotland, Sweden, do we find two bracteate hoards < at veum, 'the place of the shrines' and Gudum < gob
interpretable as surrogate burials from ploughland heim. There is no obvious connection between this area
without reported features suggesting a bog-deposit. and any natural harbour. In terms of the emergence of
Here, as in the bogfinds, surrogate burial appears to be a ports it is interesting to note that a concentration on the
delayed, 6th-century development within bracteate use. North Sea coast, further south in Ribe amt, does not
As yet I can think of no sure criterion whereby one centre upon the location where the port of Ribe
might test whether the differentiation of bogfinds and emerged in the 8th century but rather around the
finds in other contexts reflects the difference between natural anchorage which the present port of Esbjerg
sacrificial hoards and the safe-deposit type of hoard. makes use of.
In the most recent Danish and German research on
bracteates increasing attention has been paid to their
geographical distribution. This has been directed by the SYNTHESIS AND RE-INTERPRETATION IN SOCIAL AND
larger question of the emergence or existence of ECONOMIC TERMS
organized polities with a centre of power. A particular Present perceptions of the Migration Period in southern
concentration of gold finds around Gudme on eastern Scandinavia thus see trading routes, cult centres and
Fyn, which has been perceived since the 19th century settled zones coinciding in a coherent system. There
and recently increased by intensive metal-detector work seems to be little, room within this system for the
and some excavation, has been explained through phenomenon which Bradley postulates from the evi-
Gudme being a religious and social centre of power in dence of both Bronze Age Britain and Viking Age
the late 5th and 6th centuries. It is noted that the name Scandinavia of symbolic artefacts losing their symbolic
Gudme is derived from an Old Norse god heim, 'home character and reverting to bullion value in trading
of gods', and studies of the distribution of some twelve zones, at least as far as the bracteates are concerned.
examples of this name throughout Scandinavia support The only find that might fit this theory is the bracteate
the view that it represents some sacred cult site. How- fragments found at Helgo. It may however be a permis-
ever the place-name itself need be no earlier than the sible digression to suggest that a slightly later find-type,
Viking Period (Kousgard Sorensen 1985). the so-called guldgubber of Scandinavia, uncertainly
Karl Hauck has drawn connections between the dated from the later 6th century through into the Viking
coastal find-places of many bracteates and suggested Age, was used in this way. These are small, thin gold
that these represent established sea routes with depend- foils, shaped and/or impressed to present a series of
able anchorages, implying not just that such anchorages usually recurrent figural scenes. Attention is newly
could develop into gateway trading sites but also as sites being redirected at these because of the numbers

199
 THE PREHISTORIC SOCIETY

recently discovered at Sorte Muld on Bornholm (Watt overworked models of 'crisis' and incipient catas-
1986; pers. comm.). The largest concentrations have trophe. In fairness, this may have been Bradley's
been found at or near economically special coastal sites, emphasis in criticizing a contentious study by Parker
all on islands, Helgo, Eketorp (Oland) and Sorte Muld. Pearson (1984), which attempted a Marxist explana-
One of the Sorte Muld pieces was wrapped around a tion, postulating economic factors as the determinants
few pieces of hackgold. Examples have been found in of culture and cultural change. The intensification of
graves in Smaland, Ostergotland and Uppland, Sweden, hoard formation is seen as a symptom of social insta-
and two were found in excavating Eskilstuna kloster, bility or crisis in much the way that Bradley indicates;
recalling a bracteate found beneath a stave-church in but an underlying economic crisis in 4th- to 6th-century
Gudbrandsdal, Norway (Stenberger 1964, 587-88; Jutland is postulated, which led to abandonment,
Mackeprang 1952, 140). Two Norwegian hoards of migration, economic depression and emptiness in the
guldgubber were ploughfinds, from Hauge in the Jseren 7th and 8th centuries. There can be no pretence that we
area of Rogaland and Kirkeseter, Sor-Trondelag. know very much about Migration-Period economics
Hauge is a rich Migration-Period centre, not coastal, and the economic history of southern Scandinavia. Our
and not obviously a trading site. It is possible that these knowledge of fundamental agricultural details such as
two Norwegian hoards may have been ritual hoards of the organization of labour, field and crop systems and
the Migration-Period type. As Margrethe Watt (pers. slaughtering practice is negligible, as is our knowledge
comm.) has pointed out the negligible weight and the of the status and organization of artisans, the locations
fragility of the objects renders them individually at which they worked and the mechanisms for distribut-
impracticable for use as currency, but this would not ing their products. Visible none the less in this supposed
prevent their re-employment in bulk in this way. Their period of economic crisis, between the mid-5 th and late-
deposition in such quantities at Sorte Muld would still 6th centuries, is the expansion of Scandinavian culture,
require special explanation. One bracteate has now in the form of material influence in all of Anglo-Saxon
(1988) been found at the Sorte Muld site (Axboe, pers. England and on continental Europe from the Rhineland
comm.). to Hungary (cf. Haseloff 1981). The maintenance
through several generations of gold wealth at home and
There is no reason to object fundamentally to the
expansive influence abroad are not obvious characteris-
present consensus that the later Roman and Migration
tics of economic crisis. The liberal-economic argument
Periods in Germanic Europe saw overall progress in the
that a competitive society, with a larger number of
development of less egalitarian and more stratified
broadly equal, closely neighbouring and apparently
societies. It may however need stressing that change
wealthy power centres, is a dynamic and successful one
between, say, the 3rd and 7th centuries was no uniform,
would appear to fit the data rather better.
constant progress in a single direction, but involved
local regressions and in net terms was a change in degree If changes in hoarding practice do reflect changes in
rather than an absolute change from a model egalitarian political structure then again the probably conservative
society to a model state society.There is a seductive rather than radical natures of such changes implicit in
coherency between the standard view of social change this case may be emphasized. In the 3rd and 4th cen-
in Migration-Period culture and the thesis put forward turies there is evidence of a stable and controlled struc-
in Bradley's Munro lectures (Bradley 1984a). He ture in eastern Denmark, perhaps the territory of a
argued that votive hoards were created by those with recently-created confederacy of the Danes, and by con-
appropriate wealth in order to break the circulation of trast a divided Jutland and Fyn. It is striking how the
gifts and creation of debts of loyalty. With regard to distribution of bracteates in Jutland and Fyn (fig. 2)
Iron Age Denmark the suggestion is made that the respects the division of these areas into cultural zones in
intensification of deposition reflects an intensification the Roman Iron Age which was previously considered
of instability and consequently an emphasis on the roles in relation to the weapon-hoard distribution (fig. 1).
of the living rather than monuments to the dead. The material appears scarcely, and marginally, in the
To propose here the substitution of the word 'com- Anglian zone, and the remaining zones have one major
petition' for 'instability' in Bradley's formula is more concentration each. The case for a single hierarchical
than a terminological refinement. Such instability that system in each of these zones underlying the distribu-
we may infer in this period was personal rather than tional pattern, formerly a possibility, is now stronger,
communal. 'Instability' is too closely implicated with particularly with the greater number of finds

200
 8./. Hines. RITUAL HOARDING IN MIGRATION-PERIOD SCANDINAVIA

sharpening the focus upon the centre. The boundaries

of these zones can be seen in terms of topographical
features, such as the Skjern river crossing central
Jutland, but none of these is so imposing that the
apparent maintenance of these zones, with their inter-
nal structures, can be satisfactorily attributed to this
factor alone. The shift in location of the major centres
from east to west in Jutland is intriguing. Some factors
suggest themselves, particularly a shift towards facili-
ties for communications for trade and exchange, with a
major north—south routeway in Jutland passing
through the two more northerly concentrations and
another cluster by the anchorage at Esbjerg, as the
North Sea became more of an area to exploit. It may
have been that the east coast of Jutland, nearer the
Danes, was particularly the battle zone. The contrast
between the western and eastern sides of present-day
Denmark in this material diminishes greatly at this time.
The changes in the composition of the hoards are
more likely to be emblematic of a change in social
format. The weapon deposits throw emphasis onto the
achievements of a group in terms of battles won. Those
groups, however, probably constituted a military elite,
displaying differential ranking within themselves, and
the material should not encourage exaggerated notions
of egalitarianism. In the later period, changes include an
apparent shift from emphasis on achievement to
emphasis on status; the comitatus offering may cease to
be plunder from enemies but selected finery of their own
military equipment. Across the 5th to the 7th centuries
there is a very clear move from items or collections
associable with the group towards items or collections
associable with the individual, both men and women, in
individual deposits. The 'activity' reflected by the later
deposits is the possession of wealth rather than the
fighting of battles. There is an intensification of deposi-
tion in that the individuals are represented at a greater
number of sites, but one may balk at expressing an
opinion as to whether the activities of a greater number
of individuals, or a greater expenditure of individual
.1-2 effort of whatever form, is found in the later period
• 3-5 Number of Bracteates compared with the earlier.
A ready inference is the enhanced division of the elite
> 6 «• per Find into rival individuals, individuals perhaps staking their
Fig. z claim to personal, or family, aristocratic status. Such
Distribution of provenanced bracteate-finds from hypothetical division carries no implication that a
Schleswig, Jutland and Fyn. Cultural zones asfig.i united Danish nation, even a united Jutish nation or
whatever entities formerly existed in Jutland, did not
continue to exist in the Migration Period, even if only
conceptually. More likely it was the opposite. For a time

zoi
 THE PREHISTORIC SOCIETY

this format seems to have lead to success at least in the detailed concept of the votive ritual which gave them
sense of cultural expansion. However, one form of any more ready an understanding of the earlier beha-
success in competition is the suppression of competi- viour than we have now. The literature of the later
tion, the suppression of opposition — in terms of the period, mostly written down in the 13 th century but
form of competition we can postulate here, through the preserving some sections datable in much their surviv-
agglomeration of wealth and power in restricted zones ing form as early as the later 9th century, contains
or centres, socially or geographically. Centralization references which are reminiscent of archaeological
and cultural stagnation are apparent in post Migration- hoards: a ring-hilted sword (a 6th- to 8th-century type)
period Scandinavia, which is in many areas archaeo- lying on an island (Edda: HHj 8-9), gold found strewn
logically empty. But that is not to say that this could not in the grass, lying in the heather or placed on a rock
have been a period of relative stability and prosperity. (Edda: Vsp 62, Fftn 21; Saxo Grammaticus, 5.137), the
Perhaps the most remarkable aspect of the recent dis- great river hoard of the Rhinegold (Edda, various; see
coveries at Gudme is the great quantity of fine 7th- to Vglsunga saga), and weapons and treasures (including
8th-century metalwork found distributed over the same another ring-hilted sword) at the bottom of Grendel's
area, Gudme II, as fine Migration-Period and Viking- mere in Beowulf (Beowulf, 1492-1622). Explanation
Period metalwork (Thrane 1987; Axboe 1988). of the gold deposits in this literature, mostly in the
mythological poetry, is simplistic. The gold may be the
playthings of the gods from a lost golden age of leisure;
TO THE VIKING PERIOD — THE EXTENT OF the simplicity of the gods' delight in gold creates a
DISCONTINUITY
distinct impression of naivety, forming one pole in a
balancing of an anthropomorphized image of the divine
The demise after many centuries of ritual hoarding in with a transcendental one (Edda: Vsp 8, Hvm 41-52).
southern Scandinavia therefore need not be explained Gifts are given to or by the gods for favours, particularly
through a catastrophic change in human affairs but may for oracular information (Edda: Vsp 29, Grm 3, HHj
just as-well mark fulfilment in the economic and social 4). The location of Grendel's mere in untamed land
factors which according to Bradley's theory had been corresponds to the apparent location of the bog corpses,
governing its form through those centuries. A slender but the symbolic character of the location is an obvious
thread of apparent continuity, possibly a tradition of one to re-invent (Edda: Vsp 8, 29, Hvm, 41-52, Grm 3,
practice, seems to run through to the Viking Period. HHj 4). A richer imaginative treatment of gold appears
From the later 6th and most of the 7th century we have in the literature dealing with human heroes. Most
about a dozenfindsof brooches and sword-fittings from distinctive is a tragic vision of gold. Heroes may per-
southern Sweden and Denmark which satisfy some form spectacular feats to win treasure but it is cursed
criterion for votive deposits. Most are from water or and tragic. In a sense the gold dies, or loses its power,
bogs, or marked by a stone (Stenberger 1964, 488-92; with the hero. The dragon's hoard is reburied with
Geisslinger 1967, 68—69). The Bildso sword, style- Beowulf; when Gunnarr dies the secret of the great
dated to the later 7th or 8th century, is the last known VQlsung hoard, the Rhinegold, dies with him (Beowulf,
bogfind of a weapon, but around the same time swords 3110-82; Edda, Rgm 5, Ffm 10, Atk). There is a
begin to appear from rivers in and around the old correspondence here with what appears in the Migra-
Anglian territory in Schleswig. Comparable river tion Period, particularly in the surrogate burials, with
deposits have a longer history in northern Germany gold as a measure of a person's worth, in the poetry
from the Elbe eastwards, and influence from this zone is both activated by and dying with the great man. But this
probable (Geisslinger 1967, 98—107). The deposition elementary point is the limit of the correspondences.
of swords in rivers is quite familiar from the Viking The later literature gives no hint of the non-funerary
Period, and in fact accounts for the majority of Viking character of all of the votive hoards, even of the sur-
swords recovered in England (Wilson 1965, 50—52). rogate burials, which are moreover predominantly
The latest bogfinds and earliest Scandinavian river- female in character.
finds do not overlap geographically, however, and may
represent convergent rather than connected traditions.
Even if the practice were handed down in Scandinavia IDEOLOGY AS A FACTOR
to the Viking Period rather than copied from the south, Consideration of such later interpretations of the
the people of the Viking Period did not inherit any observable remains of ritual practices in terms of belief,

202
 8./. Hines. RITUAL HOARDING IN MIGRATION-PERIOD SCANDINAVIA

thought and attitude leads finally to the vexatious changes in southern Scandinavia in the first half of the
problems of ideology or religion as cultural variables 1st millennium AD but no case has yet been made that
and determinative factors in the form of the deposits. those changes compelled exclusively, although they
The proposition concerning the liminality of the loca- may have encouraged, the choice of military spoils as a
tion of human victims suggests that an imaginative form of votive offering. Moreover the constant charac-
association of a fundamental opposition in thought, ter of religion, the sense of some divine force, in the case
known versus unknown, with the biological and geo- of votive offerings to be propitiated in some way fav-
graphical environment formed a significant constraint ourable to the giver by the gift, has never in religious
on the shape of material culture (cf. Levi-Strauss 1966, studies been satisfactorily reduced to being a product of
138 ff.). Bradley certainly moves into the realm of the economic and social factors. The potentially symbolic
mind as he suggests that 'once the votive deposits . . variables of all ritual hoards, such as location, material
came to be seen as a form of payment' (my italics), and treatment, may or may not prove to be usefully
implying a cognitive development, interpreting a prac- classifiable in terms of comparative ideological, psycho-
tice that was current, and consequently adapting future logical and religious data, a process which this and
manifestations of the practice to the interpretation, other studies show some of the scope for on an individ-
' . . it was only logical for religious transactions to ual case basis. But the effort to do so to extend our
resemble the workings of the contemporary economy' understanding of this category of archaeological data
(1987b, 360-61). The place of the idea in the formation would not be misplaced.
of votive hoards would be essentially parallel in the
suggested extension of a social gift-exchange system to
the gods through the weapon deposits. Acknowledgements. Versions of this paper have been read at
These perceptions support the view that religion has a seminars in the Department of History (Medieval Archae-
peripheral place in relation to the deep structures of ology), University College London and Department of Arch-
cultural process. In the cases noted just above it may be aeology, University of Reading. Without the interest and
encouragement of James Graham-Campbell and Richard
argued that ideology develops as the conceptual range Bradley it is unlikely that I should ever have ventured so far
of the mind and is extended by economic and social into this field. But above all thanks are due to Morten Axboe
developments. Ancient ideology can frustrate the in Copenhagen, for the ready provision of offprints and
archaeologist for being too variable and unstable, information about recent developments in Denmark, where
although one should note that this constitutes a striking the productivity of current excavation and co-operative work
on the growing archaeological data deserve to be the envy of
degree of freedom for this aspect of culture. A pattern of us all.
oppositions of known versus unknown, tamed versus
wild, may be an ad hoc imaginative correlate of an early
Scandinavian Iron Age ritual of human sacrifice. But it
APPENDIX: A SELECTIVE CLASSIFICATION OF BRACTEATE
forms a stage in no perceptible macro-development in
DEPOSITION CONTEXTS
culture like the postulated economic and social
developments nor of any tradition or constraint upon Finds are referred to by numbers in the Axboe (1982)/
Mackeprang (1952) catalogues.
cultural form shared across the millennia. Modern
religious studies hold out no prospect of a theory of a a. Water and wetland deposits (see also b to d)
system of religious development to stand on an equiva- Sjadland:7, 30, 33
lent footing with the economic and social factors in Fyn: 56, 60, 64, 64a
Bradley's model (cf. Sharpe 1975, esp. 47-71; Morris Jutland: 4, 67, 71, 73, 74, 76, 83, 84, 85, 94,95,99,100,107
1987, esp. I I I - Z Z ) .
Schleswig: 114
Norway: 127, 152
Nevertheless the role of specific ideology in the for- Southern/central Sweden: 180, 247, 269, 271a, 285, 299, 301
mation of ritual hoards can be relegated to too minor a Gotland: 201
place. Particular forms of ideology and religion may Niedersachsen: 323, 324, 325
reflect economic and social circumstances, but only Pommern: 329
necessarily so in so far as those circumstances set the b. In and beside rivers
parameters in and around which the human mind Sjaslland: 36
works. The history of the great weapon deposits ought Fyn: 62
to reflect social and probably concomitant economic SE Norway: 125

203
 THE PREHISTORIC SOCIETY

c. In and beside lakes Bradley, R. 1984a. Consumption, Change and the Archaeo-
Sjselland: 23, 28 logical Record. University of Edinburgh Occasional Papers
Southern Sweden: 279 no. 13.
Bradley, R. 1984b. The Social Foundations of Prehistoric
d. Lost watercourse? Britain. Longman.
Bradley, R. 1985. Exchange and social distance — the struc-
Jutland: 107 ture of bronze artefact distributions. Man 20, 692-704.
e. Coastal Bradley, R. 1987a. A comparative study of hoarding in the
late Bronze Age and Viking economies. In G. Burenhult, A.
Sjadland: 24a, 29a, 38 Carlsson, A. Hyenstrand and T. Sjevold (eds), Theoretical
Maribo: 46 Approaches to Artefacts, Settlement and Society, 379-87.
Fyn: 62 BAR International Series 366.
Skane: 236, 237 Bradley, R. 1987b. Stages in the chronological development
Essex: 307c of hoards and votive deposits. Proceedings of the Prehisto-
ric Society 53, 351-62.
f. Heath Brendsted, J. i960. Danmarks Oldtid. Ill — Jernalderen.
Jutland: 72, 80,101,104 Gyldendal.
Southern Sweden: 297 Crumlin-Pedersen, O. 1987. Hafen und Schiffahrt in der
Romischen Kaiserzeit sowie in der Volkerwanderungs- und
g. Stonefield Merowingerzeit Danemarks. Friihmittelalterliche Studien
SW Norway: 168 21,101-23.
Southern Sweden: 221 Edda: The Poetic Edda. Various editions, e.g. R. C. Boer, Die
Edda. Haarlem, 1922.
h. Gravel Individual poems:
Jutland: 78, 80a, 96 Atk: Atlakviba
Southern Sweden: 268, 289 Ffm: Fafnismdl
Grm: Grimnismdl
i. Banks HHj: Helgakvida Hjprvardssonar
Hvm: Hdvamdl
Sja;lland: 24a, 27, 29a Rgm: Reginsmdl
Jutland: 93, 97 Vlk: Vglundarkvida
Southern Norway: 125, 128, 141, 143 Vsp: Vgluspd
SW Sweden: 265, 268, 289, 290, 292 Geisslinger, H. 1967. Horte als Geschichtsquelle. Offa-
Mecklenburg: 327 Bucher Neue Folge 19.
Genrich, A. 1954. Formenkreise und Stammesgruppen in
j. Trees/woodland Schleswig-Holstein. Offa-Bucher Neue Folge 10.
Sjadland: 34 Glob, P. V. 1969. The Bog-People. Faber.
Fyn: 3,56,60,62 Graham-Campbell, J. 1982. Viking silver hoards: an intro-
Jutland: 75, 78,107 duction. In R.T. Farrell (ed.), The Vikings, 32-41.
SW Norway: 167 Phillimore.
Southern/central Sweden: 236, 266, 298 Hagberg, U.-E. 1967. The Archaeology of Skedemose II.
Gotland: 219 Stockholm: The Royal Swedish Academy.
Hagberg, U.-E. 1983. Ein Schatzfund der Volkerwanderungs-
zeit: Djurgardsang bei Skara, Vastergotland, Schweden.
Studien zur Sachsenforschung 4, 79—92.
Haseloff, G. 1981. Die germanische Tierornamentik der
BIBLIOGRAPHY Volkerwanderungszeit. Vorgeschichtliche Forschungen 17.
Arnold, C. J. 1988. An Archaeology of the Anglo-Saxon Hauck, K. (et al.) 1985. Die Goldbrakteaten der Volker-
Kingdoms. Routledge. wanderungszeit. Bd.i. Miinstersche Mittelalterschriften
Axboe, M. 1982. The Scandinavian gold bracteates. Acta Bd.24/i.WilhelmFink.
Archaeologica (Copenhagen) 52,1—87. Hauck, K. 1988. Volkerwanderungszeitliche Seeverkehr,
Axboe, M. 1987. Die Brakteaten von Gudme II. Fruhmittel- erhellt mit Schiffsresten und Fundorten von Goldbrak-
alterliche Studien 21, 76-81. teaten. In B. Hardh, L. Larsson, D. Olausson and R. Petre
Axboe, M. 1988. Danerkonge og Danevirke. Sonderjyske (eds), Trade and Exchange in Prehistory. Studies in Honour
Mdnedsskrift 2/88, 35-44. ofBerta Stjernqvist. Acta Archaeologica Lundensis Series in
Bantelmann, N. 1988. Siiderbrarup. Ein Grdberfeld der 8°, no. 16,197—211.
romischen Kaiserzeit und Volkerwanderungszeit in Angeln. Hedeager, L. 1980. Besiedlung, soziale Struktur und politi-
I: Archdologische Untersuchungen. Offa-Biicher Neue sche Organisation in der alteren und jungeren romischen
Folge 63. Kaiserzeit Ostdanemarks. Praehistorische Zeitschrift 55,
Beowulf, ed. 8c trans. M. J. Swanton, Manchester U.P. 1978. 38-109.

2.04
 8./. Hines. RITUAL HOARDING IN MIGRATION-PERIOD SCANDINAVIA

Hines, J. 1984. The Scandinavian Character of Anglian Miller and C. Tilley (eds), Ideology, Power and Prehistory,
England in the pre-Viking Period. BAR British Series 124. 69-92. Cambridge U.P.
Hines, J. 1989. The military context of the adventus Saxo- Poulsen, P. 1987. Der Brakteaten-Fund von Gudme 1982.
num: some continental evidence. In S. C. Hawkes (ed.), Friihmittelalterliche Studien 21, 74-75.
Weapons and Warfare in Anglo-Saxon England, 2.5-48. Ringtved, J. 1989. Jyske gravfund fra yngre romertid og
Oxford University Committee for Archaeology Mono- aeldre germanertid. Tendenser i samfundsudviklingen.
graphs Series. Kuml 1986, 95-231.
Ilkjser, J. and Lonstrup, J. 1981. Interpretation of the great Sarlvik, I. 1982. Paths Towards a Stratified Society.
votive deposits of Iron Age Weapons. Journal of Danish Stockholm studies in archaeology 3.
Archaeology 1, 95-103. Saxo Grammaticus, Gesta Danorum. See History of the
Jankuhn, H. 1977. Archaologische Beobachtungen zur Relig- Danes ed. H. Ellis Davidson, trans. P. Fisher. Brewer, 1979.
ion der festlandischen Angeln. Studien zur Sachsenfors- Sharpe,E. J. 1975. Comparative Religion. Duckworth.
chung 1,2.15-34. Shippey, T. 1978. Beowulf. Studies in English Literature
Kousgard Serensen, J. 1985. Gudhem. In H. Thrane (ed.), no. 70, Edward Arnold.
GudtneProblemer, 10-17. OdenseU.P. (SeeThrane 1985.) Stenberger, M. 1964. Det forntida Sverige. Almqvist &t
Levi-Strauss, C. 1966. The Savage Mind. Weidenfeld and Wiksell.
Nicolson. Tacitus, Germania. Trans. H. Mattingly, revised by' S. A.
Levy, J. 1982. Social and Religious Organisation in Bronze Handford. Penguin Classics, 1970.
Age Denmark: An Analysis of Ritual Hoard Finds. BAR Thrane, H. (ed.) 1985. Gudme Problemer. Skrifter fra histo-
International Series IZ4. risk institut, Odense Universitet, nr 3 3. Largely republished
Mackeprang, M. B. 1952. De Nordiske Guldbrakteater. Jysk in German, with some changes, in Friihmittelalterliche
Arkseologiske Selskabs Skrifter II. Studien 21 (1987).
Morris, B. 1987. Anthropological Studies of Religion. Cam- Thrane, H. 1987. Das Gudme-Problem und die Gudme-
bridge U.P. Untersuchung. Friihmittelalterliche Studien 21,1—48.
Munksgaard, E. 1984. Bog bodies — a brief survey of inter- Thrane, H. 1989. Siedlungsarchaologische Untersuchungen
pretations. Journal of Danish Archaeology 3,120—23. in Danemark. Praehistorische Zeitschrift 64, 5-47.
Neumann, H. 1982. Olgerdiget — et bidrag til Danmarks Turville-Petre, E. O. G. 1964. Myth and Religion of the
tidligste historie. Skrifter fra Museumsradet for Sonderjyll- North. Weidenfeld and Nicolson.
ands amt, 1. Vglsunga saga. The Saga of the Vglsungs. Ed. and trans: R. G.
0rsnes, M. 1963. The weapon find in Ejsbol Mose at Hader- Finch. Nelson, 1965
slev. Ada Archaeologica (Copenhagen) 34, 232-47. Watt, M. 1986. Guldmasndernes Parade. Fra Bornholms
0rsnes, M. 1966. Form og Stil. Copenhagen: Museum 1986, 67-78.
Nationalmuseets Skrifter, Arkajologisk-historisk raskke 11. Welch, M. G. 1987. Reflections on the archaeological connec-
0rsnes, M./Engelhardt, C. 1969-1970. Sonderjyske og tions between Scandinavia and eastern England in the
Fynske Mosefund. 3 vols. (Reprint of Engelhardt's publi- Migration period. Studien zur Sachsenforschung 6,
cations of 1863-1869 with a foreword by M. 0rsnes.) 251-59.
Parker Pearson, M. 1984. Economic and ideological change: Wilson, D. M. 1965. Some neglected late Anglo-Saxon
cyclical growth in the pre-state societies of Jutland. In D. swords. Medieval Archaeology 9, 32-54.

205
 TABLE I : HOARDS COMBINING BRACTEATES AND DRESS-ACCESSORIES OF 'SURROGATE BURIAL'-TYPE, AND WITH
'CURRENCY' ITEMS IN APPARENTLY VOTIVE CONTEXTS

Cat.
Find-place I m n o Find circumstances
no. a f 8 h

24a Kitnses, Sjaelland 20 1 - Apparently washed out of bank above beach

33 Maglemose/Gummersmark, Sjadland 8 1 10 Peat-cutting in bog

62 Hesselagergards skov, Fyn 1 —— Removal of beech stump, not far from shore,
near river
80 Torning Vesterhede, Jutland 7 - 1 Peat-cutting on heath

83 Tapdrup, Jutland 14 - 27 Peat-cutting

84 Holmgards mose, Jutland 11 1 28 - - - - - 2 - Peat-cutting in bog

85 Overhornbask, Jutland 5 1 2 Peat-cutting in bog

92 Vester Nebel, Jutland 6 - 4/5 Found during work in field

95 Agerskov, Jutland 8 1 32 Peat-cutting in bog

96 Holmsland, Jutland 1 1 — Gravel digging, about 50 m from

watercourse
101 Skonager, Jutland 15 1 — In heath-peat or -turf
107 Skodborg, Jutland 4 1 5 Found during ploughing of low-lying land,
possibly a dried-out watercourse
264 Erska Hakonsgarden, Vastergotland 2 - 2 Found during ploughing and hoeing

266 Vattlosa socken, Vastergotland 4 - 16 In a wooded area

267 Grumpan, Vastergotland 3 - 8 — — — — 2 — Found after ploughing

3 24 Landeggen, Niedersachsen 4 - 2 In a bog, possibly eroded out of a Hiigel

56 Bolbro, Fyn 9 - - - 39 - In saturated level during digging of ditch

through former grove
60 Oure, Fyn 7 - - 3 i i - - - - i iS 3 In low-lying area, formerly covered with alder
thicket, ploughed seven times before discovery
72 Ars, Jutland 15 - 1 _ _ _ _ _ _ _ _ 6 - - Found during ploughing

94 Kjellers mose, Jutland 5 1 11 In peat bog

97 Norre Hvam, Jutland 10 — — 2 _ _ _ _ _ _ _ _ I 4 _ Found during ploughing of bank

269 Djurgardsang, Vastergotland 2 [diverse other objects; see Hagberg 1983] Found during ditching work

299 Soderby, Uppland 4 - 1 5 - - - - 1 - - 1 2 1 Found in marl during ditching work

329 Korlin, Pommern 6 - 1 — i-or-i 2 — — — 2 — — — In bog

64 Rynkeby, Fyn 5 - Peat-cutting in bog

71 Ars, Jutland 2 — — — In location of seasonal pond during deeper

than usual ploughing
73 Stenildvad, Jutland 4 - - 18 - Peat-cutting in bog

180 Norra Torlunda, Ostergotland — — — 1 — During first ploughing of drained bog area

99 Darum, Jutland 12 - n In bog

127 Stavijordet, Akershus 9 - - 5 - - 3 - Found during ditching in bog

128 0vre Toyen, Akershus 18 - 4 — — — I — — 2 — — — — In earthen bank

125 Sletner, 0stfold 18 - 2 Potato-planting in area recently brought into

cultivation on steep bank beside brook
157 0vre Malle, Rogaland 14 - 5 _ _ _ _ _ _ 6 - - 3 - Beneath heap of stones forming long feature
thought to have been a wall

Key (classification of artefact types largely according to Mackeprang 1952). Female dress-accessories: a,bracteate; b,brooch; c,beads;
d, pendant; e, neckring; f, armring; g, finger ring; h, clasps. Indeterminate character: i, button; j,gold spiral ring. Currency items: k, coin;
1, 'payment' ring; m, hackgold; n, electrum or silver. Male items: o, scabbard mouth-piece. Catalogue numbers refer to Mackeprang (1952) and
Axboe (1982). From the top, thefirstgroup of finds (248—324) combine bracteates with female dress-accessories alone; the second (56—329) is
of bracteates with female dress-accessories and currency items; the third (64—180) of bracteates with currency items alone; and the fourth
(99-157) of bracteates with female dress-accessories, currency items and a male item