Updated December 3, 2000
The origin of megalithic chambers
The elaborate burial rites of the Ertebølle culture included multiple interments and complex burial structures (Coles and Coles 1989:65-72, Larsson 1989) that may, at least in part, have given rise to the TRB's burial customs (Baldia 1995, Häusler 1992, 1994; Hoika 1990a, b, Persson and Sjögren 1996). These practices may have led first to the TRB North Group's non- and sub-megalithic graves and ultimately to urdolmen.[4] The coexistence of graves and urdolmen is suggested by the fact that they often contain similar EN pottery (Baldia 1995). Among the earliest datable graves containing such pottery is Konens Høj. Belonging to the Volling Group, the grave had a wooden superstructure and is dated K-919: 3645±116 B.C./2900±100 b.c. T. Madsen (1972) equated it with the graves of Vedsted and Brøndum. Vedsted contained a collared flask (ibid. Fig. 3) with decorations similar to the lugged beaker with wide furrows of the Thorsø flat-grave (ibid. Fig. 7,1). Thorsø also contained a collared flask with intersecting fringes (ibid. Fig. 7, 3). An identical design is found on the collared flask from Urdolmen A of Bogø By (Fig. 6, m). Related pottery comes from the Ølstykke urdolmen (Virum Group, K-2356: 3480±131 B.C./2760±65 b.c., Midgley 1992:497 No. 104). Similar pottery stems from the Barkær long-mounds and included a four-lug ribbed flask and a ribbed cup (Knöll 1976 No. 41B Plate 8, 5a and b). Midgley (1992:497 No. 88) lists a date of µ3961±97 B.C./µ3230±80 b.c. for Barkær, however P. O. Nielsen (personal communication 1996) has pointed out that "none of the dates concern the graves, and none turned out to be Neolithic (see Liversage 1992:102)."
The northern Barkær long-mound accommodated two side by side Konens Høj type graves (e.g. T. Madsen 1979 No. 12, Fig. 5a). This arrangement is similar to the two side by side urdolmen in the Bogø By long-mound (Fig. 6), perhaps suggesting a related, rare and ancient construction custom. In addition, the southern Barkær mound contained an axially oriented stone-cist, identified by virtue of its location as the primary burial structure, plausibly dated to the EN (Liversage 1992). Similarly, a pair of axially aligned cists were found in the Rude I long-mound, possibly dating to K3124: 3711±99 B.C./2960±90 b.c. (Madsen 1980).[5] Such linear layouts also exist in non-megalithic graves, including the two long-mounds with non-megalithic EN C ceramics of Østergård in northern Jylland (T. Madsen 1979 No. 10., Fig. 5c) and the Konens Høj type chambers within the primary long-mound of Flintbek LA 3, Schleswig-Holstein (Zich 1992, 1994a, b, 1995). Thus long-mounds may have up to five non-megalithic graves or up to four urdolmen in a row. For example Bildsø, Slagelse h., Sorø a. contained at least three axially aligned urdolmen (A. P. Madsen 1868 Fig. 48). Such sequential dolmen arrangements may be only slightly more common than the side by side urdolmen arrangement of Bogø By's two dolmen.[6] Thus the tendency to cover side by side and linear arrangements of urdolmen as well as non- and sub-megalithic graves under one long-mound that grows by accretion must be part of an early tradition.
The switch from timber to stone structures may not have been a radical change, especially if small stone-cists, such as the one of Barkær and those of Rude I, were already in use. After all, even the Konens Høj type chambers often employed large fieldstones as part of their construction. But the popularity of dolmen may be due to a depletion of easily attainable large trees suitable for building houses, palisades, boats and even single-piece wagon wheels. Increasing village size, more numerous fields, and larger herds of domesticated grazing animals would have reduced the nearby forests containing such timbers, making stone construction more cost effective. That this was indeed the case is shown by the decline in the tree trunk diameter used in the TRB architecture from the EN to the MN (T. Madsen 1990:29-30). Pollen analysis indicates tree felling in the EN and clearing of secondary forests by swidden agriculture during the MN, suggesting that the tombs were located in (large?) tracts of cleared land (S. Andersen 1988, 1992; Hansen 1993:68-74; Hoika 1990b:82 Fig. 18). Therefore, stone chambers should have occurred first in areas where a large population would have been confined to limited land, such as the Danish islands. Aner (1963), for one, sought the origin of dolmen on Sjælland. In fact urdolmen appear to be sparse in parts of Jylland, where wooden chambers were relatively common. Even in the Flintbek LA3 long-mound the Konens Høj type chambers were not replaced by urdolmen, but by later extended-dolmen.
Regions adjacent to the Danish islands must also have participated in the development of early megalithic construction. Confirmation is found in the distribution of urdolmen in long-mounds. They are confined to the vicinity of the western Baltic Sea. This core area (Fig. 1) may have barely extended to the Elbe in the south, because only a few small chambers, such as the Barskamp 709 urdolmen in small long-mound and megalithic enclosure, appear in a confined area just west of that river, together with long-mounds having a megalithic enclosure, but no megalithic chambers (Laux 1990 Fig. 3, 1991:88 Map 1). The core area continues into Mecklenburg, but there are apparently no urdolmen east of the Oder (Schuldt 1972:250 Map 3, Sprockhoff 1967). Skåne was an integral part of this original urdolmen sphere, as evidenced by the trapezoidal long-mound of Örnakulla 49, Skabersjö s. (Bägerfeldt 1993:76 Fig. unnumbered). But here too, the urdolmen may belong to a later construction phase, when large timbers became scarce. The sphere must have continued as far north as Årstad on Sweden's west coast, where a mildly trapezoidal megalithic enclosure with an axially oriented urdolmen of Herberg (Årstad 88 produced a date of St-3310: 2850±65 b.c. (Bägerfeldt and Kihlstedt 1985 No. 93, Persson and Sjögren 1996 Appendix). Farther north the urdolmen-like chambers may be later, because the urdolmen at Skee, near the Norwegian border, appears to have a perpendicular orientation to its rectilinear (?) enclosure (ibid. 1985 No. 5). The (three?) undated dolmen-like chambers in the Falbygden area could also be younger (Bägerfeldt 1993:212 Fig. unnumbered). The northernmost dolmen Holteness III, Hurum, Norway, appears to be of the later polygonal type, implying that megalithic construction reached Oslo Bay rather late, as their oldest date of ca. 3414 B.C. (T-5828: 2710±80 b.c., Østmo 1990:20) seems to confirm.
If one accepts the partial coexistence of urdolmen and Konens Høj type mortuary structures, as Zich (1994b:2), following T. Madsen suggests, the earliest urdolmen construction could coincide with Konens Høj's construction (ca. 3752-3529 B.C./3000-2800 b.c.) But this range is rather broad and dating urdolmen too early leaves an (unreasonably?) large period for the development of subsequent dolmen types. Viewing the Konens Høj date as too early, minimizing Hoika's observation that the extended dolmen of Rastorf LA 1 could show affinities with the Satrup style, and stressing the Ølstykke urdolmen's date of 3480±131 B.C. P. O. Nielsen (personal communication 1996) points out that "earlier pottery than that belong to the Fuchsberg/Virum phase has never been found associated with dolmens. That is the key to dating the beginning of dolmens at 36003500 B.C., most likely around 3550 B.C." Keeping in mind that Nielsen bases his chronology on the analysis of some 400 C14 dates, his reasoning is certainly sound, but the C14 date of Rastorf LA 1, an extended-dolmen with half-height entrance stone, has a more complex architecture than the Ølstykke urdolmen, while exhibiting an identical C14 date.[7] Therefore, Ølstykke's date with its larger standard deviation may be closer to its upper sigma of 3611 B.C. This is only 34 years less than the Konens Høj date (3645±116 B.C). Of course, the lower sigma of Konens Høj is 3529 B.C., so the beginning of megalithic chamber construction may be sought somewhere between 3645 and 3529 B.C. Nevertheless, the Årstad 88 date of ca. 3590 B.C. could lend credence to Bägerfeldt's start date of 3600 B.C.
Related Links
Neolithic/Copper
Age Link Index: Links to News Bulletins, Articles, Site Reports,
Databases, etc. about the Neolithic/Copper Age in Europe.
Please send comments or questions to Max Baldia.