Stenhøj

A dolmen by Horsens Fjord

Abstract

Stenhøj

A dolmen by Horsens Fjord

The excavation

High on the slopes north of Horsens Fjord in eastern Jutland stands a dolmen called “Stenhøj” – Stone Barrow. In 1978, I excavated the chamber and the area in front of it. The excavation was part of an investigation of megalithic tombs near the causewayed enclosure at Toftum.1-2

The chamber resides within an 18 m long and 6.8 m wide barrow, once bordered by kerbstones. Originally, around 15 stones were present along the sides of the barrow and five or six stones at either end. Dry-stone walls of sandstone slabs filled the spaces between the stones.

The first stage of the excavation revealed a heavy packing of stones and crushed flint around the chamber, and charcoal-blackened fill packed with small stones inside it (fig. 1). The fill contained Late Bronze Age potsherds and a pressure-flaked flint dagger of type VA from a disturbed Late Neolithic burial (fig. 4a).

As the excavation progressed, the form and construction of the chamber became evident (fig. 2). It had originally five uprights, two on each side and one at the rear. The side stones, set at a slight angle to each other, enclosed an oval floor with a length of 2.6 m and a width of 1.7 m. The chamber opened to the south across a 0.6 m high entrance stone, outside which stood two uprights, marking a 1.2 m long and 0.7 m wide passage. Of the uprights in the chamber, one stood undisturbed, one had fallen into the chamber and fragments of two others stood in place, while the fifth was missing. The entrance stone and one of the uprights in the passage stood undisturbed. The chamber had been 1.7 m high and probably had one large capstone covering it. The height of the passage was 1 m, though it was probably without a capstone.

Removal of the fill in the chamber revealed an intact floor of thin stone slabs abutting and integrating the fallen upright (figs. 2 and 3a). In the middle of the floor was a thin patch of reddish-brown soil, 1.8 m long and 0.8 m wide, in the middle of which lay the broken-off tip of a pressure-flaked flint dagger of type I. Outside this layer to the west lay two pressure-flaked flint daggers of types IA and IB (figs. 3b and 4b-c).

Below the stone slabs of the Late Neolithic floor came a 15-35 cm thick layer of clay-mixed sand containing some crushed flint and small stones, and below this was a floor made of smaller stones, with a fill of crushed, burnt flint between them (figs. 5 and 6). To the west in the chamber, against a dry-stone wall and in direct contact with the floor, lay a crushed funnel-necked beaker on its side (no. 1 on figs. 5, 7 and 8). Dispersed in the fill above the floor were three amber beads and 27 sherds from five different pots (nos. 2-6 on figs. 5, 7 and 8 and a, b on fig. 8).

The fallen upright lay with its southern edge in contact with the lower floor, while its northern edge was slightly higher. When the stone was re-erected, it became evident that the floor did not continue beneath it (fig. 9a). The stone had therefore fallen before the dolmen was completed and prior to the monument’s initial use.

In the passage, level with the top of the entrance stone, was a layer of grey to yellow sandy clay (figs. 6 and 9b). Below this came a packing of stones that continued a couple of metres out in front of the passage and a little to each side of it. Below the stones was a layer of yellow sand containing pebbles and crushed flint. This was thickest in the passage (10-20 cm), while outside the layer thinned out 2 m in front and 0.8 m to each side of the opening. Below it, in front of the passage, came an up to 15 cm thick layer of greyish-yellow clay containing small stone slabs and fragments of stones.

A total of 165 sherds from 11 different pots were found embedded in the yellow sand and also partly into the underlying clay layer (nos. 7-17 on figs. 5, 10 and 11). Two metres east of the passage, in front of the kerbstone line, sherds from a small, undecorated collared flask were encountered (no. 18 on figs. 5 and 10).

The Late Neolithic burial

The burial on the upper floor in the chamber dates from an early part of the Late Neolithic period, around 2300 BC. I perceive the two daggers as grave goods, whereas the broken-off tip of a pressure flaked flint dagger, found in the middle of burial area, I consider to be the cause of death, even though similar finds in other graves have been interpreted as grave goods.6

The origin of Danish pressure-flaked flint daggers goes back to the Bell Beaker phenomenon in western and central Europe. This includes “archer burials” containing members of a male elite, who were buried with copper/flint daggers, bows, arrows and wrist guards. These archer’s graves are also common in northwestern Denmark, where the Bell Beaker phenomenon had its greatest impact.3

Clearly, the bow and arrow constituted the deadliest element in the armoury of these men. The ultimate hunting weapon, also when the prey was other humans, as shown by two Danish Neolithic skeletons with flint arrowheads embedded in them.4 In close combat, however, the bow is of little use. A study of 261 skulls from Danish megalithic tombs, mostly of Middle Neolithic or Late Neolithic origin, showed that 44 (16.9%) of these have lesions, i.e. injuries, and that in 12 cases (4.6%) these proved fatal.5 Could daggers have played a role here? A copper dagger possibly, but not the large, thin-bladed flint daggers. Being difficult to handle and easy to break, these would have been no match for striking weapons.

Initially, the role of the pressure-flaked flint daggers was symbolic, but they may have had a role as weapons within the limits of the social framework of which they were part. They may have been used to settle “disputes” between “men of honour”, so to speak, and this could be why we find broken-off dagger tips in the middle of burial areas.

The dolmen and its primary use

The rectangular chamber with access from one end across a high entrance stone is a typical form in late EN II. The angled position of the uprights at the sides of the chamber, creating an oval floor, and the two passage stones in front of the entrance, however, point forward in time. The form of the chamber therefore fits well with a date at the very beginning of the Middle Neolithic.

Stylistically, the small funnel-necked beaker, which lay crushed and flattened on its side on the floor, may date to late in the Early Neolithic or the beginning of the Middle Neolithic. It probably originally stood upright, and it was possibly the only artefact placed on the floor in connection with the burial.

Only a very few sherds from each of the five pots were found embedded in the layer of sand between the two floors (nos. 2-6 on fig. 5). These pots (figs. 7 and 8) consist of a small open bowl (no. 2), a small shouldered cup (no. 3), a shouldered vessel (no. 4), the bowl from a pedestalled bowl (no. 5) and a clay ladle (no. 6). Stylistically, the decoration of the pots points unanimously to a date early in MN I. Three amber beads and an unused flint flake also lay imbedded in the sand. Two crucial questions relate to this sand layer: When was it deposited, and how are we to interpret its content of potsherds and amber beads? I will return to these questions in a separate section below.

The small shouldered cup (no. 3) is especially noteworthy. Only a third of its upper part and nothing of its lower part is preserved. Its rim diameter is 9 cm, and the height of its out-turned neck is only 1.5 cm. The cup has a 4 cm high and 2 cm wide band-shaped handle drawn in a high arc from its rim to the upper part of its belly. Both the surface and edges of the remaining sherds are heavily eroded.

I am not aware of any direct parallels to this cup, either in Denmark or in adjacent regions with Funnel Beaker culture pottery. Band-shaped handles that are raised high above the rim are similarly not known in Denmark. They do occur, however, on single-handled shouldered vessels with a conical neck from eastern Niedersachsen in association with “Tiefstich” pottery: There are good examples from Oldendorf (fig. 12a), Barskamp (fig. 12b) and Közlin.13 The most northerly example I know of is from Schuby near Slesvig.14 None of these shouldered vessels constitutes a direct parallel to the shouldered cup from Stenhøj, and we cannot prove that the cup is an import, though I tend to believe this to be the case. The temper of the sherds appears different from that of the other pots, and the obvious erosion shows that the vessel was broken into fragments long before it ended up in the chamber, perhaps being kept as a rarity in its own right. Furthermore, a handful of small pedestalled bowls from Jutland with a solid middle part emphasise possible links to the south. For example, there is a parallel to these in the same grave at Oldendorf that contained a shouldered vessel with a raised handle.15

The pottery found in front of the dolmen passage consists of 11 pots, each represented by a limited number of sherds constituting from a few percent to 25% of the individual vessels (nos. 5-17 on figs. 5, 10 and 11). The pots comprise a shouldered vessel, two bowls, four funnel-necked beakers and four pedestalled bowls. According to normal stylistic standards, the pots decorated with imprints of whipped cord date very early in MN I, while those decorated with arc-stab lines date somewhat later in the same period. In the Horsens Fjord area, however, the use of whipped cord continued later than in other areas, and there are no grounds for assuming that the pots found in front of the entrance are not contemporaneous.11

These pots represent depositional activity, as was normal in front of megalithic tombs in MN I. I have recently discussed the pottery deposition phenomenon in connection with a passage grave at Nørremarksgård, Bygholm, with extraordinarily good conditions for observation.12

Depositions in front of megalithic tombs were events limited in time that took place in connection with burials in the chambers. At Nørremarksgård, eight separate events occurred in the undisturbed part of the area in front of the tomb over a period of 150-200 years, or one deposition every 20-25 years on average. A covering of sand was added immediately after one deposition, and a survey of the literature shows that this behaviour happened frequently. After the final deposition at Nørremarksgård, a layer of stones sealed off the entire area where depositions had occurred. Again, a survey of the literature shows that this was a frequent occurrence.

Pots deposited in front of megalithic tombs are seldom fully preserved and never complete. The obvious explanation for this is secondary disturbance, but the investigations at Nørremarksgård revealed otherwise. In one of the depositions, the pots were apparently first deposited as complete vessels, but then subsequently deliberately broken, followed by removal of parts of them. In another event the pottery had been broken somewhere else, followed by deposition of a batch of selected sherds in a pit in front of the kerbstone line.

At Stenhøj we also found the sherds embedded in a layer of sand and ultimately covered by a layer of stones. Furthermore, only a small part of each pot was present and there are no signs of secondary disturbance within the main distribution area of these sherds. Figure 5 shows the recorded positions of the sherds; the size of the dots reflects the number of sherds in each record. The sherds formed two clusters 1 m in front of the entrance, with a third more scattered group directly by it. Figure 5 also shows the “position” of the individual pots, as deduced from an average of the positions of their sherds. The positions of the pots fall between the sherd clusters, showing that sherds from each of the individual vessels were found in different clusters. The limited number of sherds from each pot, together with the distribution of the sherds in different clusters, makes deposition of individual pots in front of the tomb an unlikely scenario. It seems rather that there was deposition of selected sherds from pots broken elsewhere.

The collared flask that lay isolated, 2 m away from the other pots, was clearly not part of the depositions. It is typologically earlier than the rest of the pottery found at Stenhøj, and it may indicate that the barrow contains an earlier grave. But as we only excavated the chamber area, we are unable to verify this.

The fill between the two floors in the chamber

When was the layer of sand containing crushed flint and small stones formed, and how can we explain its content of amber beads and a few sherds from five different pots? It is unlikely to result from disturbance associated with the Late Neolithic burial, leading to artefacts from the original floor becoming embedded in the sand laid down as a foundation for a new floor. Firstly, if the sherds came from pots standing on the floor, then why let one pot stand undamaged, but not all of them? Secondly, if the pots were disturbed and became mixed up in the sand, what happened to the missing sherds? Thirdly, if the sherd from the shouldered cup represents the remains of a pot placed on the floor, what caused the erosion of its edges?

An alternative and more logical explanation is that the sand was added as a sealing cover in connection with the primary activities in the chamber. To make room for the Late Neolithic burial, people began emptying the chamber, but stopped at the top of the fallen upright and chose to place the new floor here, leaving parts of the original fill untouched.

If a deliberately added fill layer preserved the funnel-necked beaker lying on the floor, what happened in other tombs where complete pots are preserved? In the early types of chamber, with no entrance for access, it appears that the chambers, completely hidden within barrows and hermetically sealed, have everything preserved as it was at the time of burial, i.e. they are devoid of fill when excavated. We see examples of this in the dolmen chambers at Grøfte and Ølstykke and possibly also at Kellerød and Stenvad.16, 17

However, chambers with no access are one thing, while chambers with an entrance are something quite different. The latter were constructed with the aim of allowing repeated access, either for rituals associated with existing burials or for new burials. Being visible, they were also exposed and subjected to reuse in later periods, often resulting in the destruction of evidence associated with their initial use. As a result, the information on them is limited, but some is available.

A handful of dolmen chambers have evidence indicating that preservation was due to the addition of a layer of sand after the burial(s) and before any secondary use (Klokkehøj 18 and Barendorf 19), or it is explicitly stated that a layer of sand was added following the initial use of the chamber (Oldendorf, 20 Aldersro 21 and Ettrup 22).

Since its publication, the Grønhøj passage grave has been crucial to our understanding of depositions in front of megalithic tombs, but it has also provided us with very detailed information on the primary use of a passage grave. 23 A 25 cm thick layer of sand covered the floor and above this was a darker layer that contained disturbed Late Neolithic burials. The lower sand layer contained six complete pots – four unbroken and two in the form of sherds (fig. 14)24 together with a thick-butted flint axe and a couple of amber beads. The pots date to MN I and MN II, while the axe is from the end of the Funnel Beaker culture. Two of the pots, both dating to MN I, lay in the middle of the chamber, high up in the sand, while the four others stood or lay on the floor close to uprights (fig. 13).

The two pots in the middle of the chamber are unlikely to have been in a primary position, since they are some of the earliest vessels in the chamber. They must have been moved in connection with a new burial, presumably from a position on the floor close to an upright. This indicates that the layer of sand was already present when the later burials were added in MN II, or to put it more succinctly: People were not only buried in a chamber, they were buried in sand in a chamber. Unfortunately, the acid soil in the area where Grønhøj is situated has resulted in no traces of bones being preserved.

The circumstances encountered in the passage grave Wangels LA 69, in northeastern Holstein, provide important supplementary information to that from Grønhøj. In the chamber were ten complete vessels, nine that were almost complete and parts of a further 38, ranging in date over the entire Middle Neolithic. Some of the complete and almost complete pots stood upright, while others lay on their sides and some stood upside down. All lay embedded in sand at floor level, and all the complete pots, including those standing upside down, were filled with sand. The conclusion reached was that the sand was present in and around the pots before they were moved and turned over.27 Many flint and stone tools lay scattered between the pots, and even though only a few bone fragments were preserved, there is no doubt that numerous burials were present in the chamber. The sand clearly led to some of the pottery surviving unbroken.

In addition to the above examples, complete pots are known from a number of other megalithic tombs where we lack information about the finds circumstances. These include a dolmen at Stendis, passage graves at Hagebrogård, Mogenstrup and Troldhøj, and megalithic tombs of unknown type at Kollund, Lønt, Odder and Skarpsalling.²⁸

With one exception, all the above examples lie west of the Great Belt or in northern Germany. They are not necessarily representative, but we should be aware of possible regional differences. For example, the chamber in the Trekroner dolmen near Copenhagen had a black, greasy burial layer containing bones from at least ten individuals that were ¹⁴C-dated to the first part of the Middle Neolithic. The only artefact encountered was a Neolithic potsherd.29 Similarly, there was a burial layer containing the bones of several individuals and datable to the first part of the Middle Neolithic in the Aldersro passage grave. Also present in this burial layer were a few amber beads, a fragment of a battle-axe and individual sherds from at least 30 different pots, though there were no indications that complete pots had been present.30 The Frälsegården passage grave in Falbygden, Sweden, had a burial layer containing remains of at least 51 individuals who collectively covered all of the Funnel Beaker culture. In addition to well-preserved bones, the layer contained 27 potsherds, 167 amber beads, 28 pieces of flint, 12 tooth beads and 12 bone beads.31 The earlier burials at Frälsegården, and the condition of the bones at Trekroner and Aldersro, indicate that defleshing and dismemberment of the corpses took place prior to burial.

If we return to Stenhøj, we can conclude that the layer of sand above the primary floor in the chamber is not unique. Sand layers are evident in other tombs in western Denmark and northern Germany, where they apparently formed an integral part of the burial rites. The pottery content of the Stenhøj layer is, however, different. How should this be interpreted? If we compare it with the pottery depositions outside the chamber, we find that the vessels are of the same types and dates and, furthermore, that both of the layers in which they are embedded consist of yellow sand with an admixture of crushed flint and pebbles. Consequently, if the pots outside the tomb represent a deposition, then those inside may well do so, too – or …?

The pottery depositions took place “by definition” in front of the tombs, while inside them we find the remains of the dead accompanied by grave goods. However, given we invented this rule, we may also have to revise it. Anne Birgitte Gebauer recently told me that the passage and the chamber of the Tustrup passage grave contained numerous potsherds that come from pots deposited outside the tomb. Selected sherds were apparently removed from the pottery deposits outside the tomb and redeposited inside it.34 The occurrence of individual sherds from at least 30 different pots in the Aldersro passage grave mentioned above may also be of relevance here. The deposition of pottery at megalithic tombs may indeed be a much more complex activity than we have so far imagined.

Stenhøj in a local context

Stenhøj lies 600 m north of the causewayed enclosure at Toftum (fig. 15.2). I have excavated two more dolmens to the south of the Toftum site (fig. 15.4),36 while an investigation of the location of further two possible megalithic tombs proved negative (fig. 15.3), even though historical sources document the existence of a chamber at one of these sites.35 A kilo­metre southwest of Toftum, two further megalithic tombs are known (fig. 15.5),³⁷ and within the limits of the map, we have three more localities where megalithic tombs may have been present.38

Clusters of megalithic tombs in association with causewayed enclosures are well documented.39 We should view the megalithic tombs located immediately south of the Toftum causewayed enclosure in the light of this phenomenon, but uncharacteristically we can hardly speak of a cluster. The reason is that activities at the causewayed enclosure, contrary to what was normal at these sites, probably did not continue into the Middle Neolithic. Whatever the cause of this, it meant that the building of megalithic tombs in association with the causewayed enclosure came to a halt. Given its location in the opposite direction from the other tombs, and at a greater distance to the causewayed enclosure, it is uncertain whether we should see Stenhøj as being associated with the latter monument. The answer is no, if the pottery found in the chamber and in front of the entrance represents the construction and earliest use of the monument. On the other hand, if there was an earlier phase, as suggested by the collared flask, there may well have been a connection. As for the burial in the chamber, and the depositions outside it, we currently know of no contemporaneous settlement in the vicinity from where the people involved might have come.

Torsten Madsen
Galten