Jelling. Settlements from the Iron Age and the Viking Age

Abstract

Jelling

Settlements from the Iron Age and the Viking Age

The name Jelling is inextricably bound up with perhaps the most prominent Viking monuments in Scandinavia, the mounds of King Gorm and his queen Thyra, the two rune stones, the remains of a supposed large ship-shaped stone setting, and the church that has predecessors from the Viking Period. How and why did Jelling reach the status mirrored in these monuments? Moreover, is it possible to find the preconditions for it in the Iron Age?

During the latter years, comprehensive investigations into settlements from the Iron Age and the Viking Period have been undertaken in the Jelling area (fig. 1). The purpose of this article is to give a review of the sites known from these periods in Jelling and its near vicinity.

At Haugsted 1, Four house sites were excavated, three of which date from the early Roman Iron Age, and the fourth from the Viking Period (figs. 2 and 3). The early Roman Iron Age settlement consists of a 20-m long and 5-m wide longhouse and two small houses of several phases. The Viking house probably measured c. 15 X 6-6.5 metres. It must be the annexe of a farm, the main building of which must be situated close by. Observations in a wire ditch running towards Jelling show remains of a fairly large Viking Age settlement.

The settlement at Hvesager dates to the period B2 of the early Roman Iron Age. The excavator interpreted the settlement as two adjoining farms surrounded by a common fence. The western farm covers an area of 2120 m2, which makes it the largest farm from this period in Denmark. An inhumation grave with weapons, spurs, a silver ring, a silver fibula etc. was found less than 50 metres from the farm fence. The grave find supported the interpretation of the farm complex as one belonging to a wealthy man. An alternative interpretation resting on what is thought to be fence remains running right through the presumably largest farm is that we are dealing with three farms. The fact that the site was not totally excavated makes an interpretation difficult. However, several things indicate that the last interpretation is the correct one (fig. 4). The pairs of posts between the two longhouses were originally interpreted as two outhouses. If we interpret them as one longhouse instead, we are dealing with a farm complex flanked by two farms of approximately the same size. The position of the longhouse is the same in the farms. All farms have a larger outhouse to the North, and all three farms lie on an axe that passes through the southern entrance in the fence, the entrance of the longhouse and the entrance of the northernmost building. The whole complex is almost symmetrical around the axe of the middle farm. This suggests three large farms of almost identical size within a fenced-in area of 900-1000 m2, surrounded by a common palisade. At the southern side of the middle farm, we have two outhouses, flanking a double entrance to the yard. This entrance suggests either a coach gate or a stately gateway built onto the houses.

At Hammelev Nørremark near Haderslev, a contemporary farm complex was investigated and interpreted as a chieftain's farm (fig. 5.1). However, several conditions show that the northernmost house at this site should be interpreted as two houses flanking an entrance (fig. 5.2). The farm closely resembles the middle farm of Hvesager, except for the building by the fence and the different direction. In addition, the parcel sizes are almost identical. Perhaps an extraordinary entrance characterised the farm of a chieftain -a phenomenon well established in historical times. The farm complex is rather large compared with others from this period in the Vejle area. Farms of a similar size are known from several locations in Southern Jutland.

Hammelev Nørremark was mentioned above. At Galsted in central Southern Jutland, a village from the second century AD was investigated. It consisted of eight farms, and the fenced-in areas are thought to cover an area of 1100-1200 m?. The results of this investigation suggest that farms of this size should not necessarily be interpreted as farms of a chieftain or wealthy farmer.

A few years ago, a fragment of a rotating quern of granite was found on the same field that held the Hvesager settlement. Until recently, querns of this type were attributed to the late Roman and the Germanic Iron Age.

However, the rotating quern from Galsted in Southern Jutland dates from the second century AD. Thorough trial excavation in the Hvesager settlement did not reveal any traces of a late Roman or Germanic Iron Age settlement. The rotating quern therefore probably belongs to the Hvesager settlement. Another fragment of a rotating quern was found in a positively early Roman Iron Age context on a settlement site near Gl. Sole, just north of Vejle. Several finds thus show that the rotating quern was used in Denmark already in the early Roman Iron Age, at least in Southern Jutland.

A house from the early pre-Roman Iron Age was found at the northwest corner of the farm complex at Hvesager (fig. 16.1). Several adjoining pits were dated to the same period. Archaeological investigation has not been carried out at the Haughus 2 site. The recording is based on finds gathered on the surface by a local, and on aerial photos showing house sites in a drought-stricken cornfield (fig. 6). The drawing of the settlement is based on the photo (fig. 7).

With a length of 22-24 metres, the eastern longhouse is rather large for this period. The house belongs to the most ordinary type of the early Roman Iron Age. It has six pairs of roof bearing posts, placed close to the walls, thus creating a spatial central nave. The walls are constructed from post put closely together -a wall type that does not otherwise appear until the late Roman Iron Age. The other supposed longhouse cannot be followed in the whole of its extent. Only four pairs of roof bearing posts and part of a wall ditch are detectable. The eastern gable seems to be rounded, and the roof bearing posts seem to be a little closer to the middle than in the first house. Both features suggest a late Roman Iron Age date. However, whether we are dealing with concurrent early Roman Iron Age houses or with the two phases of a farm belonging to the transition between the early and the late Roman Iron Age cannot be finally concluded.

At Skovgade, at least ten houses from the Germanic Iron Age and the Viking Period were excavated (fig. 8). Northernmost in the excavated area was a fence and two phases of an approximately 29-metre long Viking Age house. The latest house is typologically dated to the beginning or middle of the 10th century.

Around 60 metres south of these houses lay another three houses and a pit house. Judging from the finds from the adjoining pits, the houses belong to the end of the late Germanic Iron Age or the early Viking Period. The pit house seems to belong to the same complex.

A longhouse was found in the southernmost part of the area. Immediately south of this, a small house was overlapping an older house. The Longhouse, Va, had been extended lengthways towards the West from 28 to 32 metres. The house type and the surrounding finds date the farm to the sixth or the seventh century. The situation of the small house, VC, suggests that it dates from the same time as house Va.

It is possible that the Viking Age houses were the main buildings in a farm complex. However, they could also be large outhouses or even have had a third function. The southern longhouse, Va, was no doubt the main building of a Germanic Iron Age farm. Still, we are not dealing with a well-defined farm complex.

Both the Germanic Iron Age phase and the Viking Age phase are probably the eastern part of a larger settlement. During road construction, the remains of two house sites were found in the road track at Søndergård 2 (fig. 9). The houses were c. 9 and 10 metres long, respectively. Potsherds from adjoining pits date solely from the early pre-Roman Iron Age. The houses must be interpreted as two independent economic units from this period. A little east of the settlement, an urn grave dating from the late Roman Iron Age was found.

The remains of fifteen houses from the late Roman Iron Age and early Germanic Iron Age were found in an 11.000-1112 large area at Søndergård 1 (fig. 10). The settlement consists primarily of four longhouses, A, B, D, and M. The houses are 25-34-m long and 5.8-6.5-m wide. Each longhouse must represent a farm.

House D is the main building in the only clearly defined farm complex on the site. The fence on the northern side surrounds an area of 850-900 m2. The dating of the houses rests largely on the house typology, as there were only few finds. However, pottery dating from the third century AD was found in several postholes in house B (fig. 11). This dating is in keeping with the house type. The structures of the house A attribute it to the third century AD. The house type of house D dates from the fourth or the fifth century AD. House M has several parallels from the same period.

L is a 16 or 17-m long and c. 6-m wide house, which separates itself from the other houses by the size, the wall ditch, and the placing of the roof bearing posts. The closest parallels to this house are from the third century AD.

The connection between the longhouses and the smaller houses is difficult to establish. Two pit houses were found, one in the western and one in the southeastern part of the excavated area. On the highest part of the area, a little east of house M, was a 2.75-m deep well that could have been used during several phases of the settlement. A pit in house F contained small fragments of a basalt type well known from the time just before the birth of Christ. Recent finds from Southern Jutland have documented the use of basalt in Denmark in the early Roman Iron Age. Thus, the basalt fragment was probably used in the farm dating from the fourth or the fifth century AD in which it was found.

The site also contained several pits from the early pre-Roman Iron Age and an urn grave from the late Bronze Age.

In a road track at Skinbjerg, just west of Søndergård 1, a 25-m long and 5.5-m wide longhouse and two small houses were excavated (fig. 12). The longhouse type dates from the third or fourth centuries AD. Trial excavations just north of the road track showed traces of at least two more longhouses. This settlement is no doubt part of the same settlement complex as Søndergård 1.

At Fårupvej, a village from the early Roman Iron Age was investigated (fig. 13). The preservation conditions in the area were generally bad. Ploughing had almost destroyed the roof bearing posts of the longhouses in the northern part of the settlement, and probably was responsible for the striking lack of outhouses. No fence traces were preserved. The interpretation of the individual farm complexes therefore depends on the mutual positioning of the houses.

The basic structure of the settlement is two almost east-west lying farm sequences, probably representing two phases of the village. One phase was situated northerly in the excavated area and comprises four or perhaps five farms. The longhouses of three of the farms, BL, BM, and BN, all lay on a rulerstraight line. Farm BO lay a little to the Northeast. The fifth farm with an uncertain affiliation is farm BP southeast of the others. The main buildings of the farms measure between 15 and 18 metres lengthwise, except for the longhouse of farm BP, which has a length of only c.10 metres. This farm is untypical for the site and the interpretation is uncertain.

The other phase of the village in the southern part of the excavated area comprises five farms, BF, BG, BH, BJ, and BK. Here, too, the longhouses are situated along an almost straight line with all outhouses placed on the north ern side. In this phase, the longhouses are between 17 and 20 metres long. In spite of the missing fences, the farm complexes are quite easy to distinguish. However, the number of outhouses belonging to each longhouse at any given time is uncertain. The outhouses, the ground plans of which do not overlap, may be from the same time. On the other hand, they may also reflect a situation where one or more new outhouses were built, before the old ones were pulled down. However, new houses were often built on top of the old ones, as seen on the abovementioned settlement of Haugsted 1. Useful criteria for clarifying this question are not at hand, as neither finds nor the filling of the postholes etc. can separate the individual settlement phases. Yet, farm complexes as BF and BK seem integrated and the situation of the houses seems well considered.

Farm BK yielded a miniature pot in each of the four entrance posts in the northernmost house. Miniature pots or fragments of such were also found in fifteen of the sixteen postholes from the longhouse (fig. 14). Miniature pots are a well-known phenomenon in houses from this period. However, this is the first time they were found in such big numbers in a single house. Usually, the pots are considered small sacrifices or parts of such.

They are preserved as whole pots or just as a few fragments. This probably mirrors the fact that the pots themselves were less important in the ritual. Their situation at the bottom of the postholes suggests them being placed there when the house was built, as secondary holes were not established. However, the pots could also have been destroyed later, if the posts were dug up. If so, this operation has left no traces.

Very little pottery was found in the houses at Fårupvej (fig. 15), but the pits were full of ceramics. Groove ornamentation occurs very often, most often as three horizontal grooves, sometimes combined with zigzag bands. The pottery from the settlement at Hvesager is very similar.

The northern row of houses at Fårupvej is interpreted as the earliest phase. There is no certain documentation of two village phases or a relative time sequence between the phases. The longhouses of the southern row are more regular and of more identical structure than the longhouses of the northern row of farms. The number of farms agrees with the number of farms in the younger village at Plantagevej (see below).

The area also has some house sites that cannot immediately be attached to the phases described above. A single house a little northeast of the rest of the settlement, house AS, is dated by ceramics to the early pre-Roman Iron Age (fig. 16.2). It is 14-15 metres long and 6 metres wide and has traces of a wall ditch north and south of the eastern part. Towards the west, a ditch-like filling seems to follow the gable on the outside.

House B in the open area between the two phases of the early Roman Iron Age village has a completely different form to that of the longhouses of the village (fig. 16.3). The type does not give any clues as to the date. The frail wall posts and roof bearing posts resemble those of the house AS from the early pre-Roman Iron Age. It is interesting, that the area around this house has a concentration of pits with ceramics from this period. Much therefore points at this house being from the early pre-Roman Iron Age. If this dating is correct, we are dealing with an unusually large house for the period.

The Plantagevej settlement is from the pre-Roman and early Roman Iron Age (fig. 17-18). The settlement consisted of several separated built-up areas or farms. Generally, the remains were better preserved than the ones at the Fårupvej settlement, as traces of fencing were sporadically preserved. As at Fårupvej, the rest of the farms were mainly sorted out according to house types and the location of the longhouses and the small houses. The farms BL, BQ, BR, BU, BV, and BW are probably the older ones. The longhouses here are ten to twelve metres long. A small outhouse is situated close to several of the longhouses. The farms BL and BU are situated closely together and probably date from different periods.

A later settlement phase comprises the farms BJ, BK, BN, BS, and BT. All the longhouses have six pairs of roof bearing posts and a length of 17-20 metres. Several farms of this phase are easily separated due to preserved traces of fencing. The longhouses and outhouses are larger than those of the previous period. Farms BJ and BK are lying within the same fence. The fenced-in area measures 750 m2. Along the outer side of the northern fence is a 15-m long house, which must be contemporary with the two farms. Farm BN had an almost identical predecessor in a slightly displaced situation towards the North. The charcoal in the filling of this longhouse suggests that the house burnt down. Thus, the existence of this house must have been short, as a similar phase is not seen in the other farms. The fence of farm BS shows that the farm was extended towards the North. The original farm consisting of just one longhouse was enlarged with a large outhouse placed against the northern fence. In the later phase, the fenced-in area was around 450m2.

From the older phase, ceramics were found in just five postholes from four farms. The only characteristic piece is a fragment of an X-shaped handle. However, several pits have yielded pottery from the period. Characteristic potsherds are rim fragments with a wide facet, fragments of pots with straight sides, and pots with an added moulding below the rim, ornamented with finger imprints and angled or straight insertions.

From the younger phase, pottery was found in forty-four postholes, spread over the five farm complexes (fig. 19). Pottery was also found in the pits, for instance rim sherds with a wide edge and finely facetted. Apart from the farm complexes mentioned, there are three longhouses, which do not naturally fit into the two phases. Two of the longhouses probably have a small outhouse. These "surplus" farms probably just reflect the fact that a division into phases is artificial, as the village was continuously developing and houses and especially outhouses were often rebuilt.

Roadwork just south of the settlement reveiled three cremation graves. One of these contained grave goods consisting of two iron knives and an iron fibula. In the filling around the urn was found the pin for a bronze fibula. The grave dates from the second century AD.

At Fårup, cablework revealed postholes along a 150-m long stretch. Several concentrations of strong postholes were seen in the sides of the ditch. The dimensions of these holes suggest roof-bearing posts of several houses. East and west of here was found a 10-20-cm thick culture layer. Potsherds found here date from the early Roman Iron Age. The observations suggest farms, possibly a village complex from the early Roman Iron Age. Three kilometres west of Jelling, at Trollerup, stones made brittle by fire were found in an area sloping towards a stream. A sherd of a soap stone vessel was also found here. According to the landowner, square areas, possibly pit houses, can be observed in the corn at certain times. Observations and finds leave no doubt that a large Viking settlement lies in this area (34).

Definite traces of the earliest Iron Age settlement,such as house sites, are few and scattered.

The Hvesager and Fårupvej localities contain definite house sites from the early pre-Roman Iron Age (fig. 16, 1-2). We should perhaps attribute another house at Fårupvej to this period, a solitary house south of the northernmost row of farms (fig. 16, 3). This house is surprisingly large this period. Two partly excavated houses in a road track at Søndergård 2 are probably also from the early pre-Roman Iron Age. Apart from house sites, all excavations mentioned contained pits with pottery dating from the early pre-Roman Iron Age.

Six farm complexes at Plantagevej represent the late pre-Roman Iron Age. Each of the farms consist of a longhouse with four pairs of roof bearing posts. Four of the farms have an outhouse with two pairs of roof bearing posts.

Most of the Iron Age settlements at Jelling belong to the early Roman Iron Age. Five farms at Plantagevej belong to the beginning of this period. Three of these were individually fenced, whereas a common fence surrounded two. Each farm consisted of a longhouse with six roof-bearing posts and up to three outhouses. The farms show signs of repair such as the replacement of roof bearing posts. The settlement phase is matched by pottery with much thickened and finely facetted rims. The stray positioning of the farms corresponds to the idea of a cluster settlement (35).

The successor to the cluster settlement at Plantagevej should probably be looked for at the Fårup site, which has not been investigated. The marked differences between the pottery of Plantagevej and Fårupvej respectively suggest an intervening settlement phase. If we presume that each of the three settlement phases lasted around 100 years, we still lack the settlement of roughly two generations.

This should perhaps be found at the Fårup site. At Fårupvej, we have -especially in the later phase- a strictly constructed and well-planned village with houses arranged in rows, which contrasts the loose structure of a cluster settlement. According to the pottery from the site, seen as a whole, the Fårupvej village dates from sometime in the first century AD until sometime in the second century AD. Defensive purposes might explain the denser village structure. Unfortunately, nothing was left of the presumed fencing.

However, a settlement surrounded by a fence was found at Hvesager. This complex, including farms and a rich man's grave, was dated to the second century AD. The Hvesager farms are therefore probably contemporary with the latest village phase at Fårupvej. The fence was strong enough to have been built for defensive purposes. The grave content of spurs and weapons reveals the military involvement of the deceased, who must have been the head of the settlement. Apart from the strong fence, two more traits make these farms differ from other farm complexes of the period: the special entrance with flanking houses in the south fence and the somewhat atypical houses on the northern side of the farm complexes. We do not know houses like these from other sites, excepting a single farm north of BH at the Fårupvej site. This house further has a noteworthy detail: the eastern gable has two post holes on the inner side, which are smaller and shallower than those of the roof bearing posts. Later, similar constructions have been found at a contemporary settlement at Taulov. However, in one house at Taulov, the posts are not inside the gable (fig. 20). Could these postholes perhaps be the traces of a gable opening -perhaps even a gateway?

The last sites from the early Roman Iron Age, Haugsted 1 and 2, are probably both single farms. The pottery from Haugsted 1 suggests a dating from the later part of the period. As mentioned, Haugsted 2 has only been documented by aerial photos and a few gathered potsherds. To clarify whether we are dealing with one or two farm complexes and whether they are from the same period would require an excavation.

Late Roman and early Germanic Iron Age settlements are found at Søndergård 1 and Skinbjerg, which belong to the same settlement complex. At the moment, five certain farm complexes have been uncovered, three of which are dated using pottery and house typology to the third or fourth century AD, whereas two date from the fourth or the fifth century AD.

We only know of a few house sites from the last decades of the prehistoric period, the later Germanic Iron Age, and the Viking Period. At Skovgade, the only farm complex so far from the late Germanic Iron Age was excavated. The house typology and surrounding finds date the farm to the sixth or the seventh century AD. Two longhouses from the Viking Period were also excavated here. One succeeded the other, and the younger of them might have been in use when the erection of the Jelling monuments began. At Haugsted, two kilometres east of Jelling, and at Trollerup, c. three kilometres west of Jelling, settlement traces from the Viking Period were established. However, it is impossible to form an idea of the character of these settlements from the material available.

During the period of 1985-1996, archaeological investigations in the area south of Jelling have revealed a little more than 150 house sites front the Iron Age and a few from the Viking Period. The house sites represent phases of prehistoric Jelling during a period of fifteen hundred years. In the early Iron Age there may be no real village formation, whereas single farms consisting of a single longhouse are scattered in the terrain. During the century before the birth of Christ, at the latest, several farms -probably five- are gathered to a village at the present Plantagevej. From this time onwards, the village moves around in the area according to the well-established pattern, apparently with an interval of two generations. The overall picture is that the village moves slowly towards the Northeast. The best-investigated period is the early Roman Iron Age. The settlement of this period has three levels, the chieftain's complex, the village, and the single farm. The five farms of the village must have been the most important economical factor, based on agriculture and animal husbandry. Two farms are lying in the fringes of the settlement. Their situation may have been decided by economic specialisation in cattle breeding or other types of animal husbandry. A comparison between the rich grave at Hvesager and a number of East Jutland graves places the owner of the Hvesager grave on the commander level in the hierarchy (36). This commander was supposedly the leader of a warrior force from a tribal area corresponding to a "herred" (the name of an old Danish administrative unit). His leader function thus extended beyond the local village level, in analogy with the conditions of the contemporary Visigoth tribe in what is now Rumania. According to a written source, the leading persons of this people lived in a farm surrounded by a palisade, separated from the ordinary village population (37). In other words, the leader function is mirrored in the settlement pattern in a way resembling that of Jelling. The commander was probably chosen because of his qualifications. If he was chosen from the local Jelling population and therefore settled here, we should not necessarily expect to find forerunners or successors of the chieftain's complex in the local area. The next leader might just as well have been chosen in another area . If, however, there is a strategic aspect in the positioning of the commander, we might expect to find one or more successors to the Hvesager complex within the area. Due to its situation in a watershed, Jelling has been a centre of land traffic in historic times. It has been argued that this was also the case in the Viking Period, and that this was the reason for the high status of Jelling at this time. Perhaps this also explains the presence of a commander in the early Roman Iron Age. However, for the time being we do not have sufficient knowledge of the intervening period to establish whether a central function was attached to Jelling for any longer period. It is noteworthy that the artefacts from the Hvesager settlement and from all the other investigated sites as a whole do not include finds representing such a central function.

Folmer Christiansen