Myrhøj, 3 hustomter med klokkebægerkeramik


  • Jens Aarup Jensen



myrhøj, klokkebægerkeramik, bell beaker pottery, ceramic, house, hus, jutland, jylland, late neolithicum, sen neolitikum, artifacts, genstande


Myrhøj. Three houses with Bell Beaker pottery

Myrhøj is situated in northern Jutland, l½ km N.E. of the Mesolithic kitchen mid­den of Ertebølle on the coast of the Limfjord.

The article primarily deals with a settlement consisting of three houses with Late Neolithic flint implements and Beaker pottery. This settlement was dis­covered in 1968 during the excavation of a row of stones from the Late Bronze Age, by Oscar Marseen of Aalborg Historiske Museum, and was investigated in 1968-1972 by Forhistorisk Museum, Moesgård, under the direction of the author.

The area is ancient moorland, cultivated anew in the present century and marked by ancient sand dunes. The light sand still blows in spring and autumn, when the fields are bare.

Many stones are known to have been removed from the row during cultivation and therefore manifested themselves in Marseen's investigation only as holes filled with topsoil and often containing supporting stones. Of the nine original stones left only seven were found in situ. They varied in height between 50 and 80 cm. The row ran N.W.-S.E. and could be followed for 156 m. To the east, soil erosion had destroyed all traces, but to the west the row ended in a cup-stone in the outer kerb of a Bronze Age round barrow. One of the nine original stones also bears cup-marks. The most recent grave in the barrow is known to have been a small cist containing burnt bones and a bronze stud from the Late Bronze Age. With the cup-stones this is the most important basis for dating the stone row.

A stone cist containing a burial of the 1st century A.D. was placed close to the stone row, 35 m S.E. of the foot of the barrow.

A low sand dune around the row served to seal the underlying structures, until recent cultivation disturbed it. The stone row overlies an ancient topsoil layer which covers the houses and contains only artefacts of the same types as. those found in these. The houses (folding plate) are designated D, EAB, GAB. An investigated area west of house D is designated VfD. The three houses have a number of features in common. Most striking is the fact that the floor is below the original ground surface. The central or the eastern half is about ½ m deep while the western end is inclined. In house D, a series of parallel dark stripes was. observed in the loose sand, fig. 3, and may possibly be traces of floor joists. The houses are oriented E.-W. with slight deviations N.W.-S.E., of 13°, 8° and 24° respectively. The houses revealed themselves on excavation as rounded rectangular areas of grey soil, the largest measuring 14 x 7 m. They were divided down the middle by a few central posts; a number of wall posts were also present. The width of the house between the wall posts was nowhere less than 6 m. The distance between the wall posts in house GAB varied between 1¾ and 2½ m. This. house gives the clearest picture since there seems to have been only one occupation. Within the sunken area of house D it was only possible to demonstrate two unambiguous post holes to the west, whereas post holes were found in area VfD in direct alignment with the centre and wall lines of the house. These should be regarded as part of the house, so that house D attains a length of a least 20 m. It is not possible to decide whether the other two houses were of similar length, since excavation was confined to the sunken area.

Entrances and hearths could not be demonstrated, but all three houses contained a number of burnt stone fragments which in houses D and GAB were concentrated in a central layer. Immediately north of house D, fig. 5, and of house EAB, fig. 7, was a pit containing af few scattered artefacts corresponding to those found in the houses. South of house GAB were two undated stone-lined fireplaces, fig. 12.

The excavations were carried out in a traditional manner. The topsoil was removed mechanically and the occupation layers removed in horizontal layers supplemented by vertical sectioning. The contours of houses D and EAB were drawn at each 10 cm and of house GAB at about every 20 cm.

The occupation layer at the sunken end of house GAB was covered by a sterile layer of sand and gravel. In the two other houses the occupation layer was also covered by a sterile sand layer, but above this was another thick occupation layer. House GAB thus saw one occupation and the other houses two; house D possibly three, a layer of charcoal separating the upper layer into two halves. It was not possible to decide in the field whether the sterile sand layers represent very brief or longer breaks in occupation, no breaks in the artefact content being found. The material is therefore presented for each house, and for the whole settlement when division is not considered necessary.


Quern-stones, fig. 15. 60 stones with wear facets are believed to be quern-stones, 23 being indeterminate fragments, 25 small rubbers, and 8 large rubbers for trough or saddle querns. Only 4 actual quern fragments could be identified. The material can unfortunately not be immediately compared with quern-stones from Danish Funnel Beaker settlements, since only about 25 querns are cursorily mentioned in the literature. It is possible that the big trough and saddle querns were first employed in the course of the Late Neolithic.

Striker-stones, fig. 15. 64 pieces were found, evenly distributed throughout the four areas. The majority are small quartzite pebbles with limited crushing, but a few have strong battering marks over large parts of the surface. This group also includes pieces of grey flint. Sharp pressers of grey flint are also assigned to this group.

Bracer, fig. 16. A flat piece of stone measuring 34 x 19 x 4 mm has been identified as a bracer fragment. It was found in house D, cf. fig. 5. All edges and corners, including the break, are ground. The three perforations were made with the same drill. The material is hornfels. Despite the three perforations, the nearest parallels seem to be narrow bracers of Sangmeister's form 1 (3). Narrow bracers belong according to Sangmeister to the later Bell Beaker groups.

Animal teeth. The very sandy soil at Myrhøj has normally prevented the preservation of bone material. However, the enamel remains of three teeth, found in house D, area VfD, and house GAB, could be identified by U. Møhl, Quaternary Zoological Laboratory, University of Copenhagen, as upper unerupted molars of young domestic oxen (m 1 and m 3). The third fragment is probably also a molar of domestic ox.

Burnt bone. 12 small fragments of burnt bone, as yet unidentified, were found.

Shells. Each house contained one or two depositions consisting of a handful of Cardium shells with a few oyster and mussel shells. It is tempting to regard the shells as the remains of a single meal, but not impossible that they were destined to be used for decorative purposes.

Amber, fig. 1 7. Two disc-shaped amber beads or buttons and a small piece of partly worked amber were found in house D. None of the other houses contained amber.

Corn impressions. 3 sherds from house D and 7 from house EAB show distinct impressions of as yet unidentified corn.

Flint chippings comprise 14,187 items distributed as follows: house D, 41.3 %; area VfD, 8.5 %; house EAB, 37.2 %; house GAB, 13 %. The 13 % in house GAB constitutes about one third of the amount in the other two houses and thus confirms the impression of a single occupation phase.

The presence of an unusually large number of broad flakes, i.e. flakes whose width perpendicular to the direction of striking exceeds the maximum length in the direction of striking, cf. fig. 18, was noted during excavation. These flakes, which occur, for example, in the preliminary preparation of axe rough-outs, have been found to form a constant percentage of the chips as a whole: house D, 26.8 %; area VfD, 26.1 %; house EAB, 29.1 %; house GAB, 28.3 %.

Blade technique is not represented at all in the material. Only 3 % of the flakes are burnt, which shows that flint was deliberately kept from the fireplaces. The raw material was mainly small blocks of locally available waterworn flint. This is shown by the presence of varying amounts of cortex in more than a third of all flakes. Two types of flint were used, both the common glossy flint of grey to grey-black colour and the mat calcareous grey flint. The glossy flint seems to have been preferred for cutting tools in particular, while the calcareous flint is used for axes, among other things, presumably because it is a tougher material. The distribution of the two types of flint is as follows:


house D

area VfD

house EAB

house GAB

glossy flint





grey calcareous flint






Cores, fig. 18. 61 pieces of flint have been classified as cores. They are 5-8 cm long and 4-6 cm wide. 63.9 % are of glossy flint and 36.1 % of mat grey flint. 70 % have retained part of the cortex.

Scrapers, fig. 18, are the commonest implements, comprising 239 specimens. 23 % (57 pieces) are fragments, 22 being edge chips. 91.6 % of the scrapers are fashioned from glossy flint. Morphologically, they must be classified as flake scrapers, with a length of 35-80 mm and a width of 30-60 mm.

Retouched flakes. 266 irregular flint flakes have been retouched, in most cases finely.

Awls, fig. 19. Only 7 awl points were found, 2 of them being very slender and thin.

Burins, fig. 19. Burins were found only in house EAB, being 5 side burins, one of which has been fashioned from a surface-flaked flint sickle. 2 polygonal burins were also found.

Blades, fig. 19. The only regular blades are a strike-a-light found in house GAB and a sickle blade from house D. The sickle blade has fine wear retouch on one edge and very distinct gloss.

Axe blanks, fig. 21. 9 specimens were recovered, 4 in house D and area VfD, 4 in house EAB, and 1 in house GAB.

Vertical axes. A solitary thick-butted vertical axe has been reconstructed from three fragments found separately in house EAB, fig. 20. Small vertical axes with pointed butts corresponding to the type Danske Oldsager II, fig. 387, were represented by 2 specimens in house EAB (fig. 21). Neither is an unfinished dagger and there is therefore good reason to reintroduce this type of axe as an artefact type in its own right with greater chronological and cultural value than Becker has been willing to accord to it (4).

Thick-butted, broad-edged hollow adzes are represented by 3 certain specimens from house D, only one of them complete (fig. 20). This adze was found with the edge upwards in the north wall of house D. The circumstances do not require a votive interpretation, however. At least half the axe blanks are found to be rough­outs for broad-edged hollow adzes and this type must be regarded as most characteristic of the Myrhøj complex.

Chisels, fig. 21. The material consists of an edge fragment of a thick-butted vertical chisel found in area VfD and an unfinished piece of corresponding form found in house EAB.

Flint daggers of early lanceolate types are represented by 14 fragments, fig. 22. One of them has very deliberate parallel flaking. 4 pieces derive from house D, 6 from area VfD, 3 from house EAB and 1 from house GAB.

Surface-flaked arrowheads, fig. 23. Triangular arrowheads with a concave base occur as 15 entire or fragmented specimens in all four find areas. They mainly seem to be older types corresponding to Danske Oldsager II, no. 574 and 575. 10 unfinished arrowheads of this type have been found, fig. 23.

Transverse points. 4 specimens were found in house EAB and 1 specimen in area VfD, fig. 23. The specimen of Mesolithic appearance with concave sides was found in the middle of house EAB in the lower habitation layer. The presence of transverse points in so late a context is not unexpected, but the Myrhøj specimens are some of the best-documented.

Loom weights. A loom weight of burnt clay with distinct traces of wear around the perforation was found in house D, fig. 24. 6 other lumps of clay from houses D and EAB are classified as loom weight fragments, one of them having remains of the central perforation.

Potsherds. The pottery material consists of 4,173 potsherds: 1,817 from house D, 412 from area VfD, 1,575 from house EAB and 339 from house GAB. The percentage occurrence of side-, rim- and base-sherds is given on p. 90, 1,350 sherds, 32 % of the material, are ornamented. In house GAB this figure attains 50 % of the total.

The sherds were found, like the other artefacts, in all parts of the occupation layers, but there were concentrations in houses EAB and GAB, cf. fig. 7 and fig. 12.

The ornament falls into two groups, according to pattern. One of these consists of horizontal, circumferential grooves, covering mainly the upper part of the pot, fig. 38-43. This is by far the commonest decoration, comprising the following percentages of the total decorated sherds: house D, 63. 7 %; ares VfD, 85.2 %; house EAB, 69.4 %; house GAB, 56.1 %. The other main group consists of pure Bell Beaker patterns, incised and notch-stamped and Cardium-impressed, fig. 30- 33. Undecorated vessels occur, just as there are a few examples of stamp ornament, especially circle-stamp. An undecorated vessel from house D has a cordoned rim, fig. 27.

On 6 ornamented sherds from both main groups, traces of white inlay were found.

Bell Beaker ornament consists partly of pure zoned patterns fig. 54 and 57, and partly of zones combined with a broader »picture area« in the middle of the pot, fig. 48. Both chevron filling and »metopic« division of this area occur, fig. 33. The zone designs belong to the basic elements of Bell Beaker Culture, being Clarke's motifs no. 1, 2, 4 and 5 (p. 93) (5).

The analysis of the sherds has not yet revealed the exact minimum number of pots, but at least 125 different pots seem to be represented in the settlement as a whole. On the basis of the more of less restored vessels, the material can be divided into 5 main form groups, fig. 37.

Form 1, fig. 38-45, comprises pots of Glob's B3 type (6), which has long been established as Late Neolithic (7-8). The horizontal groove ornament is with a few exceptions confined to this type, which seems to occur with equal frequency in all three houses. The vessels vary in size from small beakers to large storage vessels.

Form 2, fig. 46-52, comprises low, broad vessels with rounded belly decorated with Bell Beaker designs (9). Apart from a solitary miniature beaker from area VfD, this form seems to occur only in house GAB. The morphological differences between form 1 and 2 are given in tables on p. 96 and p. 98, where the diameters of rim, waist, belly and bottom are given in percentages of the vessel height for 5 completely reconstructed vessels.

Form 3 conforms completely in its dimensions to form 2, but differs from it in having a carinated belly. It corresponds completely to the regular Danish bell beakers, Glob K1-2 (10). The form is represented by only two certain specimens from houses D and EAB, fig. 33: 11 and 54, but it is likely that a number of indeterminate sherds from these two houses derive from vessels of this type.

Form 4, fig. 55-61, straight-walled beakers, corresponds by and large to Glob's types K4-6 (11). They are as a rule slightly conical and may show a tendency towards curved sides. Many examples of this form were found in houses D and EAB, whereas only one doubtful rim-sherd, fig. 56, was found in house GAB. The ornament consists of Bell Beaker patterns, but horizontal grooves similar to those in the form 1 vessels are also found, fig. 60. House D seems to have contained an especially large number of sherds of small straight-walled vessels with Cardium zone pattern, fig. 57-58.

Form 5, fig. 62-65, conical bowls, is represented by several specimens in houses D and EAB. Apart from one specimen with horizontal zone ornament, fig. 66, they are undecorated. The same applies to a bowl from house GAB, with curved sides, form 5 a.

Differences between the houses

While only quantitative differences can be observed in the stone and flint inventory from the three houses, qualitative differences are apparent in the pottery. House GAB in particular differs from the others in containing form 2 and form 5 a vessels, but no form 4 or 5. Form 1 is, as mentioned above, common to all three houses. There is no archaeological basis for deciding whether these differences are due to the individual preference of their occupants or are chronologically determined. I do not believe, however, that house GAB is exactly contemporaneous with the other two, and this is also suggested by its situation; the difference need be no greater than a decade, however.

Bell Beaker elements are predominant in the Myrhøj pottery, comprising both vessel shape and decoration type and technique. Single-Grave elements are the straight-walled vessels and the confinement of the groove ornament to the upper half of form 1 vessels.


The few unambiguous indications of livelihood consist of quern-stones, corn impressions, corn sickles, loom weights and animal teeth. Myrhøj thus represents an ordinary Neolithic agricultural milieu, with both arable and stock farming represented.


A number of trenches have established that there are no further houses with sunken floors at the Myrhøj settlement, cf. folding plate. It is therefore not a question of village settlement and the houses cannot have been occupied for very long, or the sherds would be considerably more fragmented than they are, apart from the fact that the deposits would be thicker with considerably greater amounts of material. This accords closely with Therkel Mathiassen's conclusions, on the basis of a total registration of finds in N.W. Jutland, that at the relevant period of the Stone Age, people must have lived apart in transient settlements. The construction of the Myrhøj houses testifies, however, to a much more permanent form of occupation than Therkel Mathiassen presumably had in mind.

It is difficult to imagine a Myrhøj house reconstructed. There can by no means have been solid earth or wood walls. Personally, I imagine a light wooden framework covered with mats of some kind.

In a wider perspective, the Myrhøj find has significance for the conception of cultural development in the Bronze Age, which is regarded as a direct continuation of the Late Neolithic. At some time or another, possibly during the transition between the Early and the Late Bronze Age, true villages consisting of solid, timber-built, double-aisled longhouses emerged. Since house building and occupation forms are very tenacious phenomena, this episode, whenever it occurred, must have been of just as great importance as, for instance, the emergence of the Single Grave Culture after the floruit of the Funnel Beaker Culture.

It is difficult to find relevant material for comparison with the Myrhøj house type. One may refer in general to Simpson's survey of Beaker Culture settlements (13-14), and as far as Scandinavia is concerned to the remains of a house published by Strømberg (15). At several places in Europe, houses with sunken floors and a central row of posts seem to occur, although none as distinct as at Myrhøj have been published.

One would expect that settlements of Myrhøj type would be very common in Denmark, when the many graves and loose finds from the same period are considered (23). It is important to realise, however, that they will be extremely difficult to find. Imagine the Myrhøj houses situated in an ordinary ploughed field: the cover of sand dunes, stone row and ancient topsoil would have been removed and the ploughing would have removed the upper 30 cm of the sunken house floors. All the post holes would have disappeared and only a diffuse area no greater than 8 x 5 m and with a depth of 20 cm would be left. The appearance would correspond to the deepest parts of the house pits which are shown darker in the general plan, cf. folding plate.

A few structures of this type were excavated by Oscar Marseen and the author in 1963-1968, but await publication. The article very briefly discusses these settlements which all derive from the Over-grave phase. The oldest of them, from Blegind in central Jutland, exhibits a purely east Danish Single-Grave milieu, containing inter alia D-arrowheads and pots of Glob's types H and I. Common to all these settlements is the presence of sherds with Bell Beaker ornament, fig. 70-71. An important observation is that Beaker ornament is found on sherds containing crushed stone temper, which thus differ decisively from sherds in pure Single-Grave tradition at the same settlements. At Myrhøj, the clay is mainly tempered in this fashion. A closer analysis of the settlement material will therefore probably result in the demonstration of a gradually increasing Bell Beaker influence in the course of the Over-grave period in Denmark. A number of loose finds of Bell Beaker sherds, from the west coast of Jutland among other places, show that in Denmark, Bell Beaker contacts must be equally expected via the western coastal regions as via the Baltic regions.

In the presentation of the Myrhøj material it has seemed natural to emphasize the many Bell Beaker elements which characterize it: the bracer, much of the form and decoration of the pots, and the winged arrowheads and other surface-flaked flint tools.

According to the most recent investigations, the Bell Beaker Culture is thought to have originated in southern France whence it spread to the greater part of Western Europe into Central Europe and the North (25). In certain regions, especially in the Rhine area, the Bell Beaker Culture came into contact with Corded Ware groups, related to our Single Grave Culture. Mixed cultures arose containing characteristic elements from both, as in our Myrhøj find. The Bell Beaker Culture covers a period of 7-8 centuries. On the basis of Carbon 14 analyses the earliest occurrence in southern France is placed at about 2500 B.C. There are dates of 2300-2000 B.C. from N.W. Europe and finally the special group in Holland is radio-carbon dated to the period between 2000 and 1800 B.C. (26). A beaker of this group has been found in a grave with an early Over-grave axe (Grossen Bomholt near Rendsborg) (27). A radio-carbon dating from the central part of house D at Myrhøj, taken from the lower charcoal-containing stratum of the middle occupation layer (layer 6, cf. fig. 4) gives a date of 1910 ± 100 B.C. (K. 2067). This is within the range of the younger part of the Bell Beaker Culture as given above, which accords with the archaeological observations in which it has already been mentioned that the bracer from house D is one of the younger types of Bell Beaker bracers. The same applies to a certain extent to the decoration at least of the pots. From the Single Grave culture in Denmark there are a couple of datings from the late Bottom-grave period and the transition to the Over-grave period (K. 1451 and K. 1284), both of which are somewhat earlier than the Myrhøj dating (2050 and 1950 ± 100 B.C.). Radio-carbon datings of indeterminate but presumably late Late Neolithic material are later, at 1610-1500 B.C. (28).

It is obvious that the Myrhøj find should be assigned to early Late Neolithic, since the transition from Middle Neolithic to Late Neolithic is defined by the replacement of the battle axe by the surface-flaked lanceolate dagger. It is, however, just as obvious that a large part of the pottery closely conforms in shape and style to that of the Over-grave period, while other parts just as obviously bear a Bell Beaker stamp.

The opinions on the significance of the Bell Beaker Culture for cultural development in Denmark have differed. Glob accords the culture a fundamental importance in the development to Late Neolithic: The relative scarcity of artefacts from this cultural group is in inverse relation to its significance for the cultural development both of Denmark and the rest of Europe (29). Others are more reticent (30). The Myrhøj find more than hints that the Battle Axe (Single Grave) Culture in its final phase, under compact influence from the Bell Beaker Culture is transformed to that Dagger Culture which ends the Stone Age, just as it is, by and large, in the rest of Europe.

Jens Aarup Jensen





Jensen, J. A. (1972). Myrhøj, 3 hustomter med klokkebægerkeramik. Kuml, 22(22), 61–122.