Late Roman and early Germanic Iron Age grave finds from Jutland. Tendencies in social development

Abstract

Late Roman and early Germanic Iron Age grave finds from Jutland. Tendencies in social development

Introduction

Much new material has been adduced over the last two decades to illuminate the period late Roman Iron Age and early Germanic Iron Age (c. 50-550 AD) (1).

Farm and village have thus become prominent in the finds, and the number of graves has been much increased (2-5). Re-examination of booty sacrifices is also beginning to give us an indication of how war was waged at that time and of where the enemy came from (6). The last find group comprises votive finds and hoards. The new finds in this category are from Gudme in southern Funen and Sorte Muld on Bornholm (7). Both localities must have been political centres with sacral functions. On the coast at Gudme a unique find of a trading centre is also currently being excavated. This locality near Lundeborg has so far revealed various workshop activities (8, 9).

Archaeologically, there is thus now a considerably better chance of illuminating the important social changes that i.a. written sources indicate (10-13).

Based on the grave finds, aspects of the social and political development are evaluated in the following (14). The primary investigation confines itself to Jutland. Within this area, the grave finds are divided into groups on the basis of their content of artefacts. The grave goods occur in standard sets, and these equipment categories are seen as expressions of the social position of the dead. In this manner the way is paved for an evaluation of the section of population which occupies the graves and of social structure in general. Prior to the separation of categories, the regional variation and chronological development of the area must be illuminated, however.

Regional groups

In Jutland there are in the late Roman and early Germanic periods of the Iron Age two major areas which differ in a number of ways. These two groups are found in north, and in south and central Jutland respectively (henceforth termed "northern" and "southern" groups). They can be distinguished on the following points:

1. a) the incidence of plank and bole coffins and stone cists (Fig. 1),
2. b) the incidence of cremation graves and in particular urn graves with urns of types D, K, and B (pottery types, see above, p. 113) (Fig. 2),
3. c) the presence of pot types F, H, and J1+2 in the graves (Fig. 3),
4. d) the occurrence of pot types G and C in inhumation graves; in the late Roman period, type C is most frequent in the graves of north Jutland; in the early Germanic period, C and G often occur in pairs in the graves of north Jutland, and singly in the southern group,
5. e) the distribution of fibulas of Almgren group VII, series 1 and 2, and Almgren group VII, series 3, with or without octagonal bow section (16) (Fig. 4),
6. f) differences in the form of the long-houses (Fig. 5): in south and central Jutland, the dwelling section comprises the span between five western pairs of posts, whilst this does not as a rule seem to be the case in north Jutland (17, 18),
7. g) deposition of hoards and single finds in wet and dry areas respectively (19) (Fig. 6). The duration of the above-mentioned differences in artefacts, and structures, in find circumstances, or in find combinations is shown in Fig. 7. If the distributions of the regionally disposed features are collated, a picture like Fig. 8 is obtained. Hereafter it is possible to distinguish between a northern and a southern group in relation to a line like the one indicated along the district boundaries.

As a final support for the regional groupings, variations in pottery shape and decoration can be adduced. The type G pottery is to this end divided into four phases, anticipating the chronological analysis (20). Phases a to c cover the late Roman Iron Age, phase d the early Germanic Iron Age (Fig. 10).

The appearance of the pottery is compared by means of a correspondence analysis for a series of ornamental features and pattern combinations (21). The subject area has been divided up as shown in Fig. 9. The correspondence analysis compares the similarities and differences within the total amount of pottery from one area to another. The pots and variables used are shown in appendix A. The geographical distribution of the pots and a plot of the results of the correspondence analysis for the individual phases are seen in Fig. 11.

The correspondence analysis assigns numerical values to the similarities between areas. The eight actual values are shown on the maps, and isolines are drawn between them. The most important result is that in all phases the greatest dissimilarity is between a northern and a southern group as separated on the O-lines and the close cluster of isolines around these. In fig. 12 a mean O-line has been constructed. This is almost congruent with the division of the area carried out on the basis of the various structure and artefact types, c. fig. 8.

For examples of northern and southern group pottery from phases b and d respectively, cf. figs. 13 and 14.

The further geographical definition of the local groups has not been subjected here to any independent analysis (22). However, there is apparently a sharp archaeological dividing line close to the present Danish-German border (23, 24).

The line dividing the local groups in Jutland follows an extensive tract of wer areas, but it is argued that the consequent reduced possibility of contact is too simple an explanation of the differences. Nor do I believe that different production centres and their markets lie behind the distribution. If so, these areas are also subordinate to the overall economic, social and political conditions in the region, cf. Helgö in the Mälar Valley of Sweden (25).

The regional groups of Jutland are on the other hand an expression of differences in the cultural values and norms of the different societies. Thus it is possible that there are economic differences between the areas, and quite clearly there are religious, social and political differences (26). Especially with respect to the social and political differences, the grave finds are informative, and in the following the difference in the cultural values and norms of the different regions will be illuminated.

Chronology

The coming classification of the grave finds builds on a typological-chronological analysis of the pottery. In the graves, a wide selection of pottery types is represented (27) (Figs. 15 and 16).

The chronological analysis has been performed primarily on the type G and partly on the type C vessels. Correspondence analysis is again used. The results will be plotted in scattergrams with two axes. Each vessel is an independent entity, and its shape and decoration the variables on which the comparison is based. If the diagrams obtained are to be interpreted as showing a perfect seriation, entities and variables must be distributed in a parabola around the centre of the first two axes (29).

Chronological division of grave finds of north Jutland

The data matrix with the list of the 105 vessels and their shape and decoration elements is seen in appendix C (30). The first two axes of the correspondence analysis are shown in Fig. 17. A neat arced distribution representing the chronological development is seen. The type G vessels of north Jutland, which did not find a place in the primary analysis, can be dated on the basis of their form. In Fig. 18, the two first axes of the correspondence analysis are again plotted, but with only the units (vessels) represented. There is an indication in the diagram of which shape each pot can be referred to. The following shapes are employed (31, 32) (for letter designations of definitions, see Appendix B).

The vessel shapes gradually change in the diagram Fig. 18 starting with shape B in the top left corner and ending with H1 and M1 at the top right corner of the curve. On this basis, all pots can be dated relatively on the basis of shape. There is merely one shape – N1 - which was not represented among the vessels selected for analysis. On the basis of artefact combinations in the graves, shapes B, C, V, M, and N1 are seen to belong to the late Roman Iron Age, F1 to be transitional to the early Germanic period, and H1 and M1 to be pure early Germanic. Henceforth, the following phase designations will be employed: phase a (vessel form B), phase b (D), phase c (V, M, N1) and phase d (H1, M1).

Type C vessels, which are also relatively frequent, change markedly from the late Roman Iron Age to the early Germanic Iron Age (Fig. 19) (32).

The chronological development is supported by horizontal stratigraphy at the Sejlflod cemetery (Fig. 20). This burial ground has its oldest graves centrally in grave group 1. After this, grave group 1 expands and the central part of grave group 2 arises. Finally, the east and west wings of grave group 2 are established.

Chronological division of grave finds of central and south Jutland

The data matrix with a list of the 105 pots used, with their form and decoration elements, is seen in Appendix D. The first two axes of the correspondence analysis are plotted in Fig. 23. The plot here, too, shows a nice arced distribution representing the chronological development (33). The southern group's type G vessels, which were not selected for the primary analysis, will be dated on the basis of their neck shape. In Fig. 24, the two first axes of the correspondence analysis are shown with only the units (vessels) marked. The diagram indicates which neck shape each vessel exhibits. The following neck shapes are employed.

The variations in the material with respect to the shape of vessel necks are apparent from Figs. 21-22. Neck shape changes gradually from form A at the top left corner of the diagram through forms J, U, and E I, to end with H1 at the top right corner (35). As will be seen later, forms A, J, and U belong to the late Roman period (hereafter designated phases a, b, and c), E1 and H1 to the early Germanic period (designated phase d).

Type C pots become in the early Germanic period in shape and decoration like the type G vessels - only the presence of the handle distinguished the types. The younger vessels thus show a strongly reflexed neck, cf. the form groups E1 and H1 (28).

The chronological course is supported by the horizontal disposition of groups within the cemeteries. In Fig. 25 it is shown which phases (pottery groups) are represented in the localities treated. At Møllebakken, Næsbjerg, Stenderup and Hjemsted, marked chronological positions are seen (34) (Figs. 26-29).

Pottery phases of the regional groups and the absolute chronology

A large part of the graves of both the northern and the southern group can thus be dated on their content of vessels, and the analysis results can be illustrated with the vessels shown in Fig. 30.

A synchronization of the pottery phases of the northern and southern groups and a determination of the phases in relation to the primary chronological system can occur on the basis of the fibulas in the graves. If these are looked at for a moment without considering the regional differences, it is seen that there is a change in the commonest types throughout the period as shown in Figs. 31 and 32. At the beginning of the late Roman period, the fibulas of Almgren group VII, series 2 and 3, are the most commonly occurring (36-38). They are replaced by fibulas of type Gudumholm, Haraldsted and Nydam. The designation "Gudumholm fibula" is used here of crossbow fibulas with short spiral (maximum 6 turns) and a wire bow. The bow continues onto the foot, which tapers to the front. The catch is low and square, perhaps rhombic/trapezoid. The latest type of the simple fibulas comprises the cruciform fibulas.

As far as the display fibulas are concerned, the oldest part of the sequence is characterized by rosette fibulas, the later part by sheet-metal fibulas. Around the transition from phase b to c, ornamental fibulas with double spiral like the slightly later sheet fibulas, but without the head plate (39, 40) occur (Fig. 33). The sheet fibulas proper are seen in a form with rectangular or semicircular head plate in most of phase c. The same type continues in phase d but now only with rectangular head plate and with animal heads in profile at the top of the foot plate. The late development within fibulas of phase d and the appearance of relief fibulas cannot be illuminated on the basis of the grave finds from Jutland.

Combinations of pots and fibulas in the graves of Jutland can warrant a synchronization of the phase divisions of the regional groups and a correlation between these and the primary chronological system (41, 42) (Fig. 34). It is uncertain whether the entire framework of the later Roman and early Germanic Iron Age is filled out with the registered material (43, 45). The late early Germanic period seems, however, at any rate to be represented in the southern group, cf. the youngest graves of the Hjemsted cemetery with peripheral placing and with typologically late vessels (Fig. 35). Parallels to this are known from Onsvala in southwestern Scania in combination with a relief fibula (44) (Fig. 36).

The local chronological division does not coincide with the primary chronological system's phases C1, C2, C3, and D, cf. diagram p. 141 (46). Phase a is a part of C1 , but whether C1a is represented is uncertain. Phase b also contains C1b material, but the main part belongs to C2. Phase c starts in late C2 and covers all of C3 until the cruciform fibulas appear. Phase d corresponds to D.

Social analysis

It remains to illuminate the variation in grave goods and its relation to the social position of the dead. To this end, the actual graves have been divided from the outset according to sex, and the social groupings it has been possible to distinguish testify in particular to status differences within the graves. Features are also seen in certain of the men's categories which must be associated with the political leadership.

The precise significance of the different categories of grave furnishings must be evaluated on the basis of the grave goods themselves, of the geographical distribution of the graves, and of their relative positions within the cemeteries. The social meaning of the different categories is confirmed, for example, by the circumstance that the cemeteries are established with regard to these. The actual interpretation must likewise contain an evaluation of the graves' further importance in society and the concrete, historical context.

This approach builds on a long tradition of social studies based on grave content and make-up. In particular it has its roots in the principles formulated by A. Saxe and L. R. Binford at the beginning of the 1970s (47). They believed that the graves reflected both some of the social relationships of the deceased, and the size and type of society which recognized the position of the dead person. A large number of studies up through the 1970s and early 1980s followed the same principles (48). More recent investigations focus, however, also on how graves (and material culture seen in broad perspective) have a symbolic value and are actively employed by different population groups in continuation of social ideology (49).

The number of graves and their representativity

The geographical distribution of the grave finds of Jutland is shown on the map Fig. 37 with the designations for the larger cemeteries (more than ten graves) (50). The grave finds whose pottery or other furnishings can be phase-dated comprise 379 finds or 48.7 % of the registered number (51). This figure can be increased by dating a number of graves from the Sejlflod, Stenderup and Næsbjerg cemeteries on the basis of horizontal stratigraphy, cf. the maps Figs. 20, 27, and 28 (52).

Categories of grave goods

In the following treatment, an attempt has been made to evaluate the integrity of the graves. In the diagrams, Sand (SJ thus respectively designate grave finds that contain the complete furnishings minus the part that is missing due to natural decomposition, and finds where only a part of the normal inventory is thought to be present. A grave designated US can contain objects from several graves, and a parenthesis around the entire grave designation indicates that the find is under conservation and the information therefore deficient.

In the following classification of the various categories of grave goods, the characteristic objects do not usually comprise the full furnishings of the grave. Certain types are thus either common to all categories or occur so rarely that they are omitted from the survey. The former involve the commonest types of pottery, iron knives without special decoration of the haft, and bone combs.

Graves of the northern group, phases a and b

The categories of the women's graves are characterized by a qualitative and quantitative decline in grave goods from category 1 to category 3 (Fig. 38). It can also be remarked that cat. 1 graves do not contain pottery, and that only the grave from Hasseris Grusgrav has a container, but in the form of a wooden bucket (Fig. 39).

In the men's graves, three categories can be distinguished (56) (Fig. 40, 42). Cat. 1 comprises a single grave with a bracelet of gold, a so-called serpent ring. These rings were presumably favours received for faithful military service (54). They are always found in graves devoid of weapons but well endowed with imported bronze and glass vessels (55). The cat. 2 graves all have weapons and also often contain gold rings and imported goods (Fig. 41). Cat. 3 men's graves are characterized by axes and arrows.

In addition there is a small group of quite rich graves which due to uncertain find context unfortunately cannot be assigned to a particular sex. These graves contain gold rings and in some cases imported goods and are presumably in particular men's graves - perhaps a "civilian" counterpart to the weapon graves.

Only two small graves must be children's graves.

In relation to the amount of material, the categories of men's and women's graves have an even geographical distribution (Fig. 43).

The graves of the Lundegårde cemetery are assumed to be largely contemporaneous within the framework of phases a and b. In Fig. 44, the presence of men's and women's sections is seen. The cat. 2 women's graves keep to themselves in the southern part of the site, while children's graves are scattered. The last-named can here also be recognized from the grave goods, which comprise a pot of type H, perhaps with broken-off handle. A few beads also occur.

Graves of the northern group, phase c

Besides containing the largest amount of ornaments and costume, the cat. 1 women's graves are characterized by some special types of ornaments, viz. silver pendants, berloques and mosaic beads (Fig. 46). Very large bead necklaces are also seen. The graves also contain special pottery shapes, and wooden buckets or glass beakers, perhaps merely symbolized by glass sherds, whereas the other types have only the ordinary pottery types. Examples of a well-endowed woman's and man's grave are seen in Fig. 45. The category 3 graves are very sparsely furnished, and a few have been identified as women's graves only by the position of the body.

Men's graves cat. 1 are like the women's characterized by special types of containers such as wooden buckets, imported metal and glass beakers and special types of pots (Fig. 47). In addition arrows and axes occur, and valuable objects such as game-pieces, gold rings and a small pendant silver case. Belt mountings and accessories are held in common with cat. 2. A few cat. 2 graves contain a lance or spearhead. The cat. 3 men's graves contain very few objects, and as for the women's graves, a few can be identified only by the position of the body, Fig. 48.

Only a few men's graves can be distinguished outside the Sejlflod cemetery, but the women's categories are evenly distributed (Fig. 48).

In phase c, the Sejlflod cemetery has to be employed in an evaluation of the relative position of the different categories. The early phase c graves are found in grave group 1, the late ones in the centre of grave group 2 (Fig. 20) (57). Men's and women's graves are seen to be placed in separate sections (Fig. 49). The sex-determined graves are centrally placed on the site in relation to the contemporaneous graves as a whole. Children's graves can be separated both by size and by grave goods. The children's grave goods do not contain the sex-determined objects or pot type C, whereas tripartite miniature beakers and pot type H (as at Lundegårde) are seen, perhaps with a small knife or a couple of beads. The children's graves lie close to the sex-determined graves of the adults. In addition there is a group of adults' graves without particularly characteristic grave goods peripherally in the cemetery. The cat. 1 women's and men's graves are in grave group 1 centrally placed, while in grave group 2 they show a greater tendency to collect in the southern central part - women to the east of the open place, men to the west of it. Cat. 3 graves are as a rule placed at the edge of the sex-determined sections.

Graves of the northern group, phase d

Three categories of both men's and women's graves can be distinguished. Women's graves are, as in the earlier phases, characterized by gradually decreasing amounts of ornaments and costume accessories (Fig. 50). The particular ornament types of this phase's cat. 1 graves are peltate pendants and other peltate ornaments.

As far as men's graves are concerned, two cat. 1 graves without weapons and one with weapons are seen (Fig. 51). The graves have iron shears, wooden caskets and a wooden bucket with some costume accessories. Cat. 2 can still have razors and belt buckles with back-plates. whereas cat. 3 contains only graves with very simple belt buckles.

The material contains a group of children's graves. They can be distinguished by their length, but no longer by the artefacts they contain. The grave goods have become more like those of the adult's graves.

The geographical distribution of the various different grave categories is shown in Fig. 52. The distribution shows in particular that only the best endowed graves are discovered outside the large cemeteries.

In phase d it is again the Sejlflod cemetery which has to be employed when the relative position of the various categories is to be evaluated (Fig. 53). The earliest phase d graves are found to the northwest in grave group 2, the latest to the south of this. With respect to sex-segregated sectors, there is a distinct female sector in the oldest part with men's graves placed around it. In typical fashion, the best endowed men's and women's graves occupy a central position. In the younger part, however, there is no longer any clear sexual division, the best appointed men's and women's graves being found together, and the less well endowed around them. Children's graves lie, as earlier, scattered among the sexed graves. Graves without costume accessories seem, as in grave group 1, to lie peripherally to the grave group as a whole.

The development in the graves of the northern group viewed as a whole

At Sejlflod (phases c and d), graves with sex-specific burial gifts and children's graves are found in the centre of the cemetery. Peripheral to the cemetery, graves are simultaneously established without any sex marking. The latter are assumed to represent a subordinate population group, whose children are not buried in the cemetery. In the Lundegårde cemetery, which covers the earlier part of the period, a similar group of peripheral graves cannot be differentiated. This may, however, be due to the general paucity of graves, or to Lundegårde being a clan cemetery, while Sejlflod must reflect a larger community, probably a village.

We find children's graves throughout the period in the large cemeteries, in connection with men's or women's graves. Initially, they contain the special objects like vessel with broken-off handles or miniature pots (Fig. 54). In phase d, however, a selection of true adult objects can be present. The significance of this should be considered in relation to the development within men's and women's graves. If we first turn to the women's graves, the content of the various categories naturally changes in the course of the period. There is, however, the common denominator, that, for example, the best-endowed category is always marked by a comprehensive set of costume accessories and ornaments, in combination with special types of receptacle, for instance. The other categories have a less comprehensive inventory but can well contain, for example, gold rings, imported goods and similar valuable objects, only not those which characterize the best graves of the period.

The differentiated categories of men's graves differ in quite another manner as time goes by, than the women's graves. This is due to the fact that weapon graves occupy a varying position relative to civilian graves. Weapon graves clearly decline in status, and this may be seen as a shift in importance from military to civilian leaders. Arrows and axes in the later graves are perhaps even symbols of other male prestige spheres such as hunting and craftsmanship (58).

For both sexes, phase d differs in a very important manner from the preceding phases, in that there is a lower incidence of true sex indicators such as keys, resinous cement, spindle whorls, and razors, just as there seems to be a partly shared costume with fibula and agraffes (53). This can be correlated with the shift in the relative position of graves in the Sejlflod cemetery. From having previously been placed in men's and women's sectors, the graves finally seek the company of other graves of the same degree of wealth. Another kind of status is thus marked with the interment. It is tempting to see these changes in connection with changes in social organization. Leadership finally divorced itself, from a society of clans where status and rank had importance, both within and between clans. Therefore the wealth of the grave occupants is expressed with little regard for kin-related concepts like sex and age. Age is mentioned just because the development within children's graves contributes to underpinning the theory of such a social change. The feeling that a dissolution of norms was involved is strengthened in phase d by the circumstance that more scope is shown for placing the dead and their appurtenances in the grave than before. The development thus probably reflects an increased social differentiation at the cost of socially egalitarian clan structures.

Graves of the southern group, phases a and b (59)

Cat. 1 of the women's graves is characterized primarily by containing glass beakers and/or wooden buckets (Fig. 55). In addition it exhibits gold ornaments and a wide range of costume accessories. Necklaces with silver, millefiori or large polychrome beads are held in common with cat. 2 (fig. 56). In addition, iron combs should also be mentioned. The cat. 3 graves have a less comprehensive costume inventory consisting of three or four fibulas, and cat. 3 contains only one fibula and simple necklaces.

The men's graves are more difficult to distinguish, and only two categories of grave furnishings can be differentiated (Figs. 57, 59). The cat. 1 graves contain sword and shield and perhaps spear and lance - probably what was originally the full complement of weapons (60). The weapons are found combined with imported glass or bronze receptacles and a mirror (Fig. 58). In both cat. 1 and cat. 2 graves there is a costume inventory consisting of belt garniture and perhaps a fibula.

Children's graves are not characterized by special types of object, but appear as (small) adult graves.

The distribution of grave categories in the Vorbasse cemetery shows that the cat. 1 women's graves form a separate group, and that there is a men's section in the centre surrounded by women's and children's graves (Fig. 60).

The geographical distribution of the various grave categories is even with the exception of the cat. 1 women's graves (Fig. 61). The reason for this poor representation can be the absence of datable pottery types in the inventory.

The graves of the southern group, phase c

Among the furnishings of the women's cat. 1 graves may be wooden buckets with bronze fittings (Fig. 62). In addition, they exhibit a comprehensive ornament and costume inventory including one or two display fibulas. The ornaments can be large necklaces with silver pendants, millefiori beads or large polychrome beads. Cat. 2 contains fewer fibulas and smaller necklaces. The other equipment consisting of belt buckles, pins, spindle whorls, etc. is found again in cat. 3, in which the graves, however, always contain only one fibula in the costume.

The cat. 1 men's graves contain imported glass receptacles (Fig. 63). In addition, wooden buckets with bronze fittings, glass game-pieces and gold finger-rings occur and parts of a set of weapons. The cat. 2 graves contain merely belt buckles and perhaps a fibula.

Phase c children's graves contain pottery and perhaps belt buckles and necklaces (Fig. 64). A couple of the children were given particularly small pots.

The weak representation of men's graves makes it difficult to evaluate the importance of the different categories in the cemeteries. In the case of Stenderup, a centrally placed men's section is seen surrounded by cat. 2/3 women's graves (Fig. 65). Children's graves can be placed near both men and women.

In Enderupskov there are at the same time two women's sections and one men's section (Fig. 66). Children's graves are rare and as usual situated among both women's and men's graves. With respect to the disposition of categories, it can be said that the wealthier graves (cat. 1) are centrally placed in relation to the sex-determined sectors and a north-south line in the cemetery. The unsexed graves are found with those without furnishings (and thus undatable) peripheral to the men's and women's sections.

At the Næsbjerg cemetery, one area can be distinguished with women's graves and one with men's graves (Fig. 67). Children's graves lie scattered among the adult's graves.

The cat. 1 graves are geographically evenly distributed, whereas the more poorly appointed graves can only rarely be distinguished outside the proper cemeteries (Fig. 68).

Graves of the southern group, phase d

The graves of this phase are few, and sparsely appointed. The following categories can be distinguished: women's graves, cat. 1, with costume accessories of 2 simple fibulas and pins (Fig. 69); women's graves, cat. 2, with beads and/or hooks and eyes; men's graves, with belt buckles and/or agraffes (for securing trousers or leg wrappings) (61) (Fig. 70).

The structure behind the disposition of the graves at the Hjemsted and Enderupskov cemeteries in phases c and d is uniform (Figs. 71-73). In both places a change occurs in phase d from a division of the cemeteries into sex-segregated sections with the grave category marked by the position of the graves to apparently more randomly disposed graves. The segregated sections are abandoned, the categories are not disposed according to any discernible pattern, and older graves are not respected. At Hjemsted the end of this development is apparently seen with a few late graves with very sparse furnishings and peripheral placement. A subordinate population group represented by peripherally placed graves without sex-specific furnishings is seen at both cemeteries in phase c and - doubtless due to the size of the material - only at the Hjemsted cemetery at the beginning of phase d. Thereafter it is not possible to distinguish any presence of graves of this kind.

The development in the graves of the southern group seen as a whole

The graves without sex-specific furnishings or entirely without equipment are seen in phases c and d to be peripherally placed in the cemeteries. The same can be the case earlier, but cannot be perceived in the oldest, small cemeteries. As for the northern group, the peripheral graves are thought to represent a subordinate population group. Children's graves are seen throughout the sequence, in decreasing amounts indeed, but with precisely the same form of expression and placement in relation to the other graves. The children have grave goods like small adults, and their graves are placed among the sex-determined graves.

Within cat. 1 of the women's graves there is a high degree of continuity of expression. In phases a-b they are characterized by extensive ornament and dress furnishings and are endowed with special objects in the form of glass beakers, wooden buckets with bronze mountings, gold finger-rings and perhaps key or spindle whorl. In phase c, the corresponding group of graves also has rich personal equipment, but to a slightly lesser degree gold finger-rings and special kinds of receptacle. This development continues, so the best-furnished women's graves in phase d can be characterized only by their dress equipment.

As far as the men are concerned, the best endowed categories stand out throughout the sequence especially with the grave goods proper rather than the costume accessories. In phases a-b, they contain a full set of weapons, gold finger-rings, and imported goods. In phase c, only parts of the weapon set are present in these graves. In the southern group the shift from full set of weapons to parts of one seems to be a part of the general decline in grave furnishings. In phase c, men's graves still contain glass beakers, wooden buckets, gold rings and game-pieces, however, whereas in phase d, the men have retained only the costume accessories.

Within the southern group, the sex of the grave occupant is marked right from the start of the period less clearly than in northern Jutland. Thus grave goods are only rarely seen which symbolize the male and female activity areas, such as spindle worls and hooks, razors, strike-a-lights and tinder stones and arrows. Men's weapons appear to a greater degree as if they should be seen more in connection with political status than as a manifestation of male equipment as such. The concepts sex and age which were formerly associated with the social importance of the clan are also less marked in the graves of the southern group than in those of the north.

On account of the decline in grave goods of the southern group and the gradual reduction in weapons already from phases a-b to c, it is difficult to maintain that the phase d graves without weapons are civilian graves. In the southern group there seems on the contrary to have been maintained a leadership structure whose power and position is marked by the presence of military grave goods.

The structure of the cemeteries is not essentially changed in the course of the late Roman period. Not until phase d does a major break occur in development. It becomes almost impossible to distinguish categories within the sex-determined graves, and there is an almost total dissolution of the grave structure. Men's and women's sectors disappear as well as specially endowed graves with specific positions within the cemeteries. Furthermore, older graves are now respected to a very small degree along with existing men's and women's sectors. Children's graves can still to a certain extent be followed among the sex­determined graves, and peripherally placed graves are still established without indications of sex. Only in the very last part of phase d does the grave equipment fail to such a degree that no structure can be distinguished behind the furnishing and disposition of graves.

Tendencies in social development

The transition from early to late Roman Iron Age

At the beginning of the late Roman Iron Age, society had undergone major and sudden changes. Particularly noticeable is the shift in settlement which must be related to changes in the basic economy and size of household and to the introduction of new technology in a number of spheres.

With a starting point in the village of Vorbasse, it seems that the early part of the late Roman period farm had the following appearance (62): a long-house with one or two outhouses and a fenced yard. Pit-houses are likewise known in this period, and the farm is in principle as we find it so frequently in the 4th and 5th centuries. In relation to the early Roman period farm, the differences are great and the increase in area considerable. The new type of farm could house, and most of the time needed, a larger household than previously. The increase in the labour force may partly have been in the form of casual labour from a subordinate population group with limited rights, and in fact graves have been distinguished which could be interpreted in this way. When we reach the early Germanic Iron Age, a couple of instances can also be cited of small houses with fireplace, in association with the farms (63). It may be a case of separate accommodation for the farm­hands. With the change in the size of farms, a concentration of settlement on the better soils occurs. Agriculture undergoes a change and becomes more intensive, using the manure from many stabled animals, cf. the abandonment of the dyke systems. Cross-country work waggons appear (64) and other technological novelties are the rotary quern, the shaft furnace and the vertical loom (76). Apparently, a good economic basis has been created, with higher priorities being given to secondary occupations, and an improved standard of living (66).

The change in the early part of the late Roman period is marked and notable in that several of these innovations had been known for centuries in the areas just south of the Jutland peninsula (67). Not until some time in the 2nd century was society apparently ready to absorb them. Receptivity to innovations was possibly built up during the increasing contact of the previous centuries with the Central European, Germanic tribes and the Roman Empire (68). The many border disputes of the succeeding centuries must have had great importance for the development in Germania, much beyond the border areas. One effect can have been changes in access to power and prestige in domestic societies and a military activity detached from society's other institutions (70, 71). The south Scandinavian booty sacrifices can be connected to such a social restructuring. Most of these finds are from the late Roman and early Germanic periods and were particularly marked in the area in question around the year 200 (69) (Fig. 74).

Whether the northern and southern groups experience an equally strong orientation towards the provincial Roman and Germanic societies in the central parts of Europe is uncertain - some differences can probably already be discerned here (72). Thus the grave finds show that objects which must be associated with the clan-based relations like sex and age are most strongly marked in the graves of the northern group. Moreover the well­furnished graves with men have a civilian counterpart, and the only true booty sacrifice known in this early part of the late Roman period is from Trinnemose, whereas the southern group has, in the same period, weapons in practically all the well-appointed men's graves and several extensive booty spoil sacrifices. The clan as a structure underpinning society is most persistent in Northern Jutland. These organisational circumstances go back perhaps to the early Roman period (73, 74).

Late Roman Period

In the late Roman period, the farms continue in the same form as before. In Vorbasse their placement becomes regular in the course of the late Roman period, as if planned (75). The individual farms and there enclosed plots can be seen to maintain their relative size through a number of phases.

This draws attention to the plot of the farm as key to division of the land. The farmsteads have also in this period individual fences with access from each farm both to the village common and to the surrounding fields. Permanent fixed boundaries must mean that the farm labourers - the presumptive subordinate population group - must be landless.

The grave finds are throughout the late Roman period still relatively well appointed, although an incipient decline can be noted in the southern group. The most important difference in the regional groups is seen within the men's graves. In those of the northern group there are thus no longer weapons in the best-appointed graves, as there are in the southern group. Once again there seems to be a connection between weapons in graves and the extent of booty sacrifices. In the late part of the late Roman period, no booty sacrifices are known from the northern group's area, while many depositions occur along the east coast of central and southern Jutland (Fig. 75). The territory of the southern group is clearly greatly affected by the many hostile incursions. Thus a major defence construction is undertaken here with the blocking of Haderslev Fiord and the construction of the Olger Dyke (76-79). The latter goes back to the beginning of the late Roman period. These structures attest a leadership structure which transcends the local area and judging by the well-equipped weapon graves must have been military in nature. In north Jutland, however, the graves indicate civilian leadership, and no invading foreign armies can be traced at this time.

Early Germanic period

In the course of the early Germanic period, the villages change. In Vorbasse, the farms with their enclosed plots thus diminish in size (80). In some parts of the country, forest gains ground again (81-83). This possibly reflects an increased importance of arable farming at the expense of animal husbandry - perhaps in connection with a change in the cultivated cereal types (84, 85).

Some crafts attain an unsurpassed standard and show the economic capacity of an elite section of the population (86, 87). The concentration of wealth is seen, for example, in a burnt house from Dankirke in southwestern Jutland with a valuable store of imported glass objects (88, 89).

The grave finds of the early Germanic period should doubtless be understood in continuation of the above. Within the southern region it should merely be remarked that the graves are no longer used to display and legitimate social position. A leadership structure divorced from the clan organization has become so well established that the burial rituals no longer need to confirm it. The leadership can be assumed still to have been of a military kind, the conflicts in this area continuing unabated through the period.

In northern Jutland, development does not follow an exactly parallel course. The changes in grave furnishing here seem to underline wealth at the expense of the clan-based relations. A primary leadership structure and status positions outside the clans is a near possibility. The disputes within the area are, however, not so fierce and the incitement to maintain a permanent leadership form not so great that the basis of leadership is institutionalized in a way that makes a marking of status at burial unnecessary. In fact leadership seems to seek further confirmation in the rituals around its religious/ritual functions (90).

An attempt has been made to sketch social development in Jutland through the periods late Roman Iron Age and early Germanic Iron Age. Both north, and south and central Jutland are judged to have attained a permanent leadership structure divorced from the clan system. North Juttland presumably does not attain this state as early as the south and central area. One reason for this may be sought in the history of the area, the southernmost regions of Jutland already in the centuries after the birth of Christ having attained a relatively closer contact with and influence from the Roman border areas. Another reason which possibly follows on the first is the great military pressure on the east coast of south and central Jutland.

Painted with a broad stroke of the brush it is undoubtedly also possible to see a permanent leadership for a large area expressed in the regionality of the period. The difference between the small and strongly marked regional groups of the Pre-Roman and early Roman periods and the less marked distributions of the following centuries can i.a. have its origin in the suppression of local conflict under a central political leadership.

Jytte Ringtved