Jættestuen Jordhøj

Resumé

The Passage-grave Jordhøj

Associated with the passage graves and contemporary polygonal dolmens of southern Scandinavia is a special group of finds mainly comprising pottery found at the foot of the mound, on both sides of, and often also in front of, the passage entrance. The largest amounts of material have been recorded in Zealand and Scania, where up to 50,000 sherds from over 1000 pots may occur at a single barrow [6]. In Jutland, it is exceptional to find as many as 100 vessels, and as a rule less than 50 are present. Finds of this nature are common, featuring frequently in the literature, but have received varying interpretations since the beginning of the century. One school of thought interpreted the finds as the result of evacuation of the chambers to make room for new interments [7], another school believed that they represented grave offerings: consecration offerings at the completion of the grave, repeated depositions at the facade in association with the individual interments, or offerings made at special festivals connected with ancestor worship [8-9]. A third school interpreted the finds as a combination of evacuation and offering [10]. If this last interpretation is correct, it will be necessary to distinguish and separate three categories of material from passage graves to enable a correct interpretation to be made in each case: the material in the chamber itself, the evacuated material in front of the passage entrance and facade, and finally the votive material.

In the hope of obtaining an immediate and objective evaluation of the chances of separating these assemblages, I made a special investigation of the passage grave Jordhøj, in 1964-65. This monument near Mariager, co. Randers was specially selected, as it seemed to provide the ideal circumstances for the investigation in mind. Its chamber and passage had already been examined in 1890, and had contained material from the Corded Ware Culture and Late Neolithic only [1-2-3]. It was also uniquely well preserved. The Corded Ware and Late Neolithic interments in the chamber had lain completely undisturbed, before excavation, and not the least amount of dust had intruded, nor apparently damp. Wood, probably from Late Neolithic coffins, was still preserved on the floor, and sheets of birch bark from the erection of the chamber were jammed among the stones of dry-walling between the orthostats of the chamber walls (Fig. 1-4). Since not a single sherd of Funnel Beaker pottery had been found in the grave, which must, like other passage graves, have been erected by people belonging to the Funnel Beaker Culture, the chamber must have been very thoroughly evacuated. As the condition of the mound indicated that it had not been subject to contemporary interference, it should be possible to distinguish stratigraphically the evacuation material, which must be present, from any votive material. Excavation concentrated on the area in front of the passage and at the foot of the mound on either side of the entrance, where experience has shown finds may be expected. In addition, part of the kerb was exposed on the eastern flank of the mound to ensure that the full extent of the material was known. A section from the foot of the mound in SSW to the SW corner of the chamber permitted the causes of the complete sealing of the chamber to be investigated. Finally, all courses of the dry-walling between the orthostats of the chamber were carefully examined for remains of the birch-bark sheets which had been found during the previous excavation. Since the earlier measurements of the grave had been carried out in relative haste and under difficult circumstances, new measurements were made, giving rise to certain corrections (Fig. 5).

The mound (Fig. 5): The passage grave was originally covered by a mound with a diameter of 17.6 m. and a maximum height of 3.5 m., and was surrounded by a kerb of about 60 stones over 1 m. high. In prehistoric times, presumably at the close of the Stone Age, the mound was increased in diameter by 3 m. and surrounded by a kerb of smaller stones lying over ½ m. above the original level. Later erosion has given it its present diameter of more than 25 m.

The chamber (Fig. 5 + 7): The chamber, the shape of which is apparent from the plan fig. 5, lay, as is usual in passage graves in Jutland, not centrally in the mound but displaced about 0.5 m. to the W and 1.5 m. to the SSE in the direction of the passage. As is also usual, all the orthostats in the chamber slope inwards, here at an angle of 70 ° -80 ° to the floor, whereby the chamber narrows by more than 25 % from floor to roof. At floor level, there is a space of 20-50 cm. between orthostats, due to their narrowing towards the base. About ½ m. above the floor they are in contact, however, although contact is lacking between stones 1 and 2 and 5 and 6. The spaces between stones are filled with flagstone dry-walling from the floor to the underside of the cap-stones. Between stones 1 and 2, the dry-walling is recessed about 40 cm. from the front of the strongly inclined orthostats to form a little niche (Fig. 6). 60 cm. above the floor, this niche measures 40 X 40 cm., and at this height, the original excavation exposed a flagstone wedged into the recess as a shelf (Fig. 4). Above this and between each course of dry-walling were found the remains of those birch-bark sheets which had been observed during the original excavation. These must have been inserted when the chamber was built and presented an opportunity for its radiocarbon dating (cf. p. 44) [11 + 18]. According to the original excavation reports, the floor of the chamber consisted of natural subsoil sand.

The passage is 5.7 m. long and where it enters the chamber has a width of just under 0.9 m. It narrows gradually to only 0.6 m. at the entrance between kerbstone IØ and lv.

Also its height diminishes from the chamber, where it is 1.6 m. high, to the entrance, where it measures only 0.9 m. The passage floor consisted of subsoil gravel and was broken by two thresholds, one in front of the outermost cap-stone and the other at the chamber door. On the outer threshold, which was missing at the latest excavation, stood in 1890 a large stone slab closing the passage. At the inner threshold, the cap-stone has been dropped to form a lintel and the orthostats placed so that they act as jambs separating the passage from the chamber itself.

The chamber is covered by two cap-stones resting approximately at right-angles to the rear wall, in the east on four orthostats and in the west on three. The triangular gap thus created over the passage opening is covered by a large, sloping slab, which rests on the cap-stones proper. All gaps between the cap-stones are otherwise covered by closely fitting stabs, completely sealing the roof. As far as the outer threshold 1 m. from its exit, the passage is covered by 5 stones.

The construction of the mound: As shown by the section fig. 8, the cap-stones over the chamber were completely covered by several layers of slabs, laid on one another like roofing tiles. This covering had a distinct outward inclination and extended up to 2 m. beyond the chamber. The gaps between slabs were filled with burnt flint. Above this, a dark streak indicated a covering of sods or ling turves with the grass side down. Outside the area of the chamber, the stabs were covered by a layer of stones coalescing at its periphery with a pile of stones which must be the packing around the chamber itself. Above this pile, the mound was built up of a number of sand layers interspersed with occasional thin layers of stone. Especially noteworthy is stratum 7 immediately behind the kerb, which consisted of a compact mass of white, burnt, cracked, crushed flint which originally formed the surface of the passage grave mound. The overlying layers must stem from the later extension of the mound, which is limited 1½ m. outside the passage grave kerb by a chain of smaller stones (fig. 9). This secondary kerb rested directly on stratum 3, which mainly consisted of material eroded from the mound: gravel and crushed, burnt flint like that constituting stratum 7 behind the passage grave kerb, besides a few stones, including a large slab which must have fallen from the top of the inner kerb. This flinty layer covered a dark grey, stoneless, very fine-grained sand layer with humus (stratum 4), resting directly on the sand of the subsoil.

Between the secondary kerb and the top of the passage grave kerb was a very regular, close cladding of the flat stones, which formed the periphery of the secondary mound. This mound was built after the passage grave mound had fallen into disrepair: mound fill of flint and gravel lay in ½ m. high heaps along the base of the kerb, many of the slabs covering the kerb had slid down, and the kerbstones themselves were inclined more or less outwards (fig. 10-11).

In the excavated area, comprising about one third of the mound periphery, all 22 kerb­stones were in their original positions. They were 0.8-1.2 m. high. Resting on kerbstone no. 1 just east of the passage entrance was a heavy, flat slab, the plane front of which had been flush with the front of the kerb before it had been displaced. Corresponding slabs were found along the whole of the excavated mound periphery, but had fallen from their position over the kerb. Several stood erect or at an angle in front of the kerb with one end pressed into the ancient vegetation layer, while others lay flat in the subsoil or in scree. A few others were found behind the top of the kerb (fig. 11). Originally these stabs must have formed a continuous kerb-covering right around the mound.

When newly erected, the facade, and probably the whole periphery, must have appeared roughly as in the restored elevation fig. 12. The kerb was free-standing, presumably with a vegetation-covered surface in the foreground. The stone cladding and stone heaps are the result both of disrepair and of evacuation from the chamber and of the erection of the secondary mound above the passage grave.

The finds: The excavated material has been divided into three categories according to context. The first comprises objects from the chamber, recovered in the 1890 excavation. This is the most recent material and contains pots and daggers from the Corded Ware Culture and Late Neolithic. The second comprises material from in front of the kerb at each side of the passage entrance, resting directly on the humus layer over the subsoil, and from behind the kerb. This group comprises pottery of the Funnel Beaker Culture only, and must represent offerings made at the grave facade. The third group was found mainly in front of the passage entrance, partly in a black humus layer covering group 2, and partly in a succeeding layer of stony gravel. It comprises, besides pottery of the Funnel Beaker and Corded Ware Cultures, flint weapons and tools and amber beads, and must stem from two or more evacuations of burial goods from the chamber.

Material from the chamber: When the chamber was opened in 1890, it was clear of earth, but to the north and south the artefact layer was covered by a yellow, powdery deposit. Only a strip of floor 33-64 cm. wide from the passage opening to the rear wall was completely bare. In and on top of the artefact layer, which was about 30 cm. thich, lay on both sides of this strip pieces of tree-trunk, branches and hewn planks, some of which were over 1 m. long. The planks were of birch wood, whilst the trunks and branches also included aspen and hazel. The excavator did not venture to determine the function of these remains, but their location in the chamber makes it likely that they are the remains of coffins. If this is the case, there must have been at least two coffins, separated by the bare strip.

The burial goods in the northeastern part of the chamber occurred on and under the planks, but this does not necessarily imply that they were deposited simultaneously. They comprise 1: A 9683, fragment of lanceolate dagger, 2: A 9690, base sherd of coarse­bodied pot, 3: A 9684, flint strike-a-light, 4: A 9691, undecorated, conical beaker (fig. 13: 1, 3 and fig. 14: 4). The dagger was found among the planks, whilst 2, 3 and 4 were found at their southwestern corner.

In the southwestern part of the chamber, under similar conditions were found 5: A 9692, decorated beaker with flared neck (fig. 14: 5) and 6: A 9693, sherd of an undecorated pot. Both items lay near the passage opening.

In the northwestern part of the chamber were found three daggers, 7: A 9685, burnt fragment of lanceolate dagger, directly on the floor; 8: A 9686, resharpened dagger with rhombic hilt, about 15 cm. above the floor; 9: A 9687, lanceolate dagger, indeterminate depth (fig. 13: 7-9).

Finds 10-12 stood or lay in a group by the most northwesterly orthostat, but the depths at which they were recovered were not recorded, 10: A 9694, straight-walled, decorated beaker, 11: A 9695, piece of decorated beaker with flared neck, 12: A 9688, lanceolate dagger (fig. 14: 10-11, fig. 13: 12).

Only one object was recovered in the passage, find 14: A 9689, dagger with flared hilt fig. 13: 14).

Dating: The two beakers nos. 5 and 11 belong to the Corded Ware Culture and are assigned to Glob's type B 3, which is the latest of the B-group and can be of Late Neolithic age. The two other beakers are of Glob's type L 2, which is assigned to the Upper Grave period [14-16]. The strike-a-light and daggers are all Late Neolithic. The lanceolate daggers 1, 7, 9 and 12 are the earliest, from the beginning of the Late Neolithic, and may be contemporaneous with the vessels of B 3 type, and probably are, since only a few centimetres separated vessel 11 and dagger 12. Daggers 8 and 14 are later, and must both be assigned to the close of the Late Neolithic [17].

These finds are secondary in the chamber, and it must have been in connection with the earliest interments represented there that the final evacuation occurred. The grave goods removed on that occasion make up only a small proportion of the material found in front of the grave facade. The rest of this material in the form of numerous pots deposited by the kerb, stems from earlier evacuations and grave offerings but first and foremost from of the latter.

Evacuation and offerings: In the present investigation, a clear difference with regard to stratigraphic context was observed between votive pots and evacuated material. The sherds of the votive pots were concentrated along the first 5 kerbstones on either side of the passage entrance. The majority lay directly on a thin sandy humus layer which covered the subsoil, others in small heaps of mound fill at the foot of the kerb, and a few behind the displaced kerbstones at a level slightly below their top (fig. 15-16).

The lowest evacuation layer was concentrated in a 10-15 cm. thick, greyish black layer of humus, containing some charcoal fragments, 10-15 cm. above the layer containing the votive material. The material was largely confined to the area in front of the passage entrance. In a stony layer at a higher level, artefacts stemming from a later evacuation were found, and at the extreme periphery of the excavation, potsherds from a third evacuation.

The material as a whole contains 7,000 sherds. 44 pots could be identified, 10 of which are represented by more than 100 sherds, 4 by 50-100 and the remainder by less than 50. The large mass of undecorated sherds, comprising half the total, must, however, conceal other vessels. Besides pottery, the evacuation layer contained a number of flint tools and amber ornaments.

The pots can be divided into the foliowing types:

2 can be identified with certainty (HAA-HAB) and 2 more tentatively (HAC-HAD) as pedestalled bowls. The first two are assigned to the Middle Neolithic, Period IB [19-20], the other two to Period II and III respectively [21-23]. A spoon, HAE, presumably belongs to one of the earlier pedestalled bowls (fig. 17).

5 shouldered bowls occur, all with a medium high neck, but somewhat variable shoulder profile. Vessel HAF (fig 18) has a gently protruding, rounded shoulder and slightly conical neck. On the basis of its shape and ornamentation it is assigned to an early part of Period II [24]. The four other bowls belonging to this group, HAG, HAH, HAJ and HAK (fig. 18-19), have a very narrow, rounded shoulder, which is much less prominent than in pot HAF. They are all dated to the end of Period II and must be considered transitional in shape between the previous pot and the following group [26].

To this latter group are assigned 10 vessels, HAL, HAM, HAN, HAO, HAP, HAQ, HAR, HAS, HAT, and HAU, all assigned to Period III [27-28]. The shoulder is here rudimentary, existing merely as a narrow deviation in the profile, a slight curve at the base of the neck marking the transition to the body through an angular or rounded deflection. These bowls also differ from the former group in ornamentation, for instance by the vertical ornament having moved down from the shoulder to the upper part of the body. The shelf is covered by horizontal lines or a simple band which may extend onto the body (fig. 20, 21).

Sherds of a lid with a vertical edge and curved top (HAV) and the corresponding vessel, HAW, belong to the same period (fig: 22) [29].

4 vessels, HAX, HAY, HAZ, HBA (fig. 23), are large, coarse storage pots of funnel beaker shape with cylindrical or slightly funnel-shaped neck and rounded or convex conical body. They are all of coarsely tempered clay with thick walls. They are assigned with some uncertainty to Period IB or II [30]. A further 2 vessels are classified as storage pots, HBD with a wavy edge over a massive lug, and HBC (fig. 24) with concave cylindrical neck and plastic ornament. The latter vessel must presumably be assigned to Period III [31].

To Period IV is assigned the coarse vessel HBD (fig. 23) with cylindrical neck and plastic curves under the lip. Vessel HBE is a coarse, undecorated, open bowl.

3 vessels are classified as suspension bowls, HBF, HBG and HBH. The neck is in all cases of medium height, straight and cylindrical or slightly funnel-shaped. The transition to the body is almost flat and the body convex conical. These vessels have either perforations under the lip or a vertical hole through the shoulder, or both (fig. 25). They are assigned to Period II.

Vessel HBJ (fig. 25) is biconical, with 4 perforated lugs at the junction of neck and body, presumably arranged in pairs. On the criterion of shape, it is assigned to Period III [32-33].

The shape of 9 vessels, HBK, HBL, HBM, HBN, HBO, HBP, HBQ, HBS, HBT, cannot be properly determined (fig. 25).

One vessel, HBU (fig. 25), belongs to the Corded Ware Culture. It is a straight-walled beaker with swollen neck and must be assigned to the Ground Grave or Upper Grave period [34].

The flint material contains 5 axes (fig. 26). They are all very fragmentary, but can be classified with considerable accuracy. Three, AR, KO, KP, have a rectangular butt section and belong to Becker's Lindø-type which is dated to Period IV. Two fragments, G and HY, of the same axe must have had a nearly square butt section, although the butt is now missing, and belong to Becker's Store Valby-type from Period V [35]. The axe U is of the thin-bladed type, a type which has not yet been sufficiently studied for it to be placed in the detailed periodic system. This also applies to the sole chisel of the find, the small fragment AAE.

Arrow-heads are represented by 8 transversal points, fashioned from blades or flakes, L, M, AD, AE, BY, CN, KM, OQ, and a tanged point of type B, UR (fig. 27). The latter does not belong to the Funnel Beaker Culture, but to the Pitted Ware Culture. Its chronological position is uncertain, but it must probably be assigned to Period IV or V [36]. Of other flint tools, a cylindrical block, V, belonging to the same culture as the tanged point, may be mentioned.

14 amber beads were recovered, but only 6 are so well preserved that their shape can be determined (fig. 28). Three are tear-shaped, flat pendants with a hole through the pointed end, DD, DX, YZ. Two are axe-shaped, one of which has preserved a spatular blade and neck, while the other is a fragment of the same kind. On the basis of their similarity to the stone battle-axe prototypes, they must be assigned to the end of the Funnel Beaker Culture, Period IV or V [3 7].

The total material from the chamber and from the facade thus shows that the passage grave was in use throughout the Middle Neolithic portion of the Funnel Beaker Culture, the closing stages of the Corded Ware Culture comprising the Ground Grave and Upper Grave periods and through a large part of the Late Neolithic. The amounts and kinds of material from the different periods are very different, however, the early Middle Neolithic periods predominating, in particular Periods II and III. These differences are directly determined by the changing burial rituals, the custom of making offerings, which is responsible for the largest part of the pottery, being abandoned here in Period III.

The primacy of votive material is immediately apparent from the plan fig. 29 with associated histograms, even though this plan also includes pottery from the evacuation layers. Most of the pottery is concentrated immediately in front of the kerb in row D, whereafter the density decreases steadily towards the southern limit of excavation. Within this area of falling density, the sherds were concentrated in rows VI and IX in front of kerbstones 2 v and 1-2 ø. Eastwards to kerbstone 5 ø the density of potsherds falls gradually, whilst to the west the decrease is more sudden, the number of sherds in the squares before kerbstone 3 being only one third of the number in front of stone 2. In the area of greatest concentration in front of the kerb, most of the sherds lay in heaps, forming an almost continuous bank east and west of the passage entrance (fig. 30). Near the foot of the kerb, the sherds were found in or on a thin layer of yellow sand, associated with some burnt flint which must be fallen mound-fill. A little further out they lay on the sandy humus which covers the subsoil, and some large sherds had been pressed into this layer by slabs which had fallen from the top of the kerb. It is in fact symptomatic that the largest concentrations of sherds were found in front of these slabs. The largest heaps contained more than 500 sherds, and several of them contained the remains of several pots. The large heap in front of kerbstone 2 v in square DV thus contained large fragments of 4 pots and smaller parts of 3 others (fig. 31). In the heap in front of kerbstones 2-3 ø in square DX were found large pieces of 5 vessels and smaller parts of a further 3. Only exceptionally were all the sherds of one vessel found together in one heap. A few sherds may occur behind the kerbstones, others spread over several squares, but with a few exceptions always in the same stratum. This dispersion suggests that a very large proportion of the vessels originally stood on top of the kerb, which accords with the circumstances for votive finds at many other passage graves [38].

Altogether, it has been possible to identify 36 votive vessels, deposited over a period of time extending from the construction of the passage grave in Period IB till Period III. 19 of these vessels were placed at the facade east of the passage entrance, and 17 west of it. The original position may in most cases be determined with a fair degree of accuracy, either from the dispersion pattern or from the position of single sherds behind the kerb. In the plan fig. 29, the accession number of each vessel is encased in the square from which the largest number of its sherds -if more than 10- was recovered, corresponding to the italicized squares in the list, p. 28 seq. The proviso is ignored, in the case of bowls HAA and HAB, and the spoon HAE: only a few sherds were found, but the original location is fairly certain. In their case the positions marked on the plan and in the list are those of the sherds found on the kerb. The encased accession number will, however, only occasionally mark the original position of the vessel -most vessels, as mentioned above, probably stood on the kerb, and the sherds recovered near the facade will therefore be nearest the place where the vessel was originally placed. The plan is much simplified in relation to the excavation records, as the position of every find, also of single sherds, was marked accurately, and where several sherds lay together in a heap, they were exposed and removed as an entity, their exact context being noted. Where the original position of vessels is discussed, it has therefore been possible to give due consideration not only to the location of the largest concentration of sherds, but to the dispersion as a whole, with special weight being given to sherds found on top of the kerb and in the compact heaps in front of them. When the original position of the individual vessels had been determined, typologically closely related vessels were found to occur in close association. A distinct difference was also noticed between the assemblages east and west of the passage entrance. Thus at the eastern part of the facade a group of 4 large storage vessels with pits under the lips was found, while nothing similar was found west of the passage. Three of these vessels were originally placed on or near kerbstones 5 ø and 6 ø, HAX, HA Y and HAZ. The sherds of the fourth vessel, HBA, were too sporadic in occurrence for the original location to be determinated. In contrast, both the early pedestalled bowls, HAA and HAB, and the spoon which belongs to them, HAF, were found west of the entrance, at stones 1-2 and 4 respectively. The relatively early shouldered vessel, HAE, had likewise stood west of the passage on stone 2. In the later material from the end of Period II, a group of 4 very uniform vessels, HAG, HAJ, HAH, HAK, was found near stones 2-3 ø. In immediate contact with the bowls were found the suspension bowl HBF and vessel HBL, which must both be contemporary with them. The remaining 6 identifiable vessels from the eastern facade are all later and must be assigned to Period III. They are all large, coarse vessels, only three of which are ornamented, HAR, HAS, HBC, HBD, HAO and HAT.

In contrast to the vessels on the other side of the passage, those of the western facade are completely dominated by notch-stamp-ornamented vessels, first and foremost by the type with a shelf at the junction of neck and body. 7 vessels are so well preserved that identification is certain, and a large piece belonging to a vessel of indeterminable form is decorated with the same characteristic ornament. These vessels, which constitute a typological entity which must be assigned to Period III, differ markedly from those found east of the passage. 4 of them, HAL, HAM, HAO and HBN, were found by the kerb­stones 2-3 and 3 v, while one, HAP, was found by kerbstones 4-5 v, and one by stone 6. Finally, vessel HAN was found at kerbstone 2, and HBM at stone 3. In the same region were found 2 more notch-stamp-ornamented vessels, HAD and HBK, corresponding both in date and style to the others. Besides the stylistically very uniform vessels were found the very coarse storage vessel HBD, which is perhaps as late as Period IV and the undateable coarse vessel HBT, which was found as a large collection of sherds in front of stones 1-2.

A pattern of this nature, which is already manifest in the early depositions and further accentuated in the later ones, right up to the point where offerings cease, cannot be fortuitous. Only if several vessels had been deposited at each offering would such an arrangement in groups, uniform in type and date, occur. How many offerings are represented cannot be established with certainty, merely that the number of offerings is considerably smaller than the number of pots present. As the typological classification reveals, there cannot have been less than 4. The maximum number can only be guessed at, but is hardly greater than 6 or 7. The number of offerings is thus surprisingly small in relation to the long period of time, presumably 200-300 years, in which the grave was in use by the Funnel Beaker people.

The deposition of several pots at each offering is reminiscent of the custom at the contemporaneous mortuary or cult houses of Tustrup type, which are by now familiar [40]. In these houses, the number of vessels lies in three cases between 26 and 40, whilst two houses, one of which was very small, contained only one vessel each. In these wooden buildings with their earth-driven posts, which at Tustrup at least were still in good condition when the house was destroyed by fire, the pottery must have been deposited over a short period of time, or as much evidence seems to suggest, on a single occasion.

At several passage graves, the material and the context in which it was recovered also suggest that even finds with a large number of pots comprise only a few offerings. If this notion is correct-and it is at the moment based on a restricted material­offerings must surely have been made on the occasion of interments in the cham­bers. Neither is there, in the passage graves of Jutland, anything to suggest that they were originally intended for mass burial [41]. Where many skeletons are found in passage graves, as is often the case in eastern Denmark, they are as a rule much later than the time of erection, often from the close of the Neolithic, and belong to cultures other than the Funnel Beaker Culture [42].

Evacuations. The number of primary graves in Jordhøj, i. e. belonging to the Funnel Beaker Culture, cannot be determined with accuracy since they may have been only partially cleared on each occasion. The amount of material occurring in front of the grave, which can be identified as dump, indicates, however, that they were not numerous. The earliest evacuation comprises the sherds of 4 vessels whose shape and type can be determind, 4 flint axes, a chisel fragment, 8 transversal arrow-heads, 14 amber beads and a small collection of cremated human bones [42 a]. This material was found mainly in the area in front of the passage mouth and a few metres to either side, embedded in a layer of greyish black sand with a little humus. Apart from two transversal arrow-heads, L and M, and one of the axe-shaped beads, the flint tools, pots and amber lay, as will be seen in fig. 29 c, in a single horizon, sloping down from the kerb to the foot of the mound. Of the 4 vessels identified, two comprise the set of lid and bowl HAV and HA W, and one the conical neck of a biconical lugged vessel, HBJ, all dated to Period III. The remaining pot, HBB, is a coarse vessel with wavy edge, of uncertain date. Of the 4 flint axes, the thin-bladed one, U, was found at the middle of the passage mouth, and one of the thick-butted ones, AR, in front of the kerbstone 2 v. The remaining two thick-butted axes, KO and KP, were found near each other and at the same level in front of kerbstone 2 ø, lying above a heap of sherds containing vessels HAJ, HAK and HBC. The thick-butted axes are assigned to Period IV. The 8 transversal arrow-heads can be assigned only to the Neolithic in general. The amber beads, the shape of four of which could be determined, were found in the region of the transversal arrow-heads and at the same level. One axe-shaped bead which lay close to the thick-butted axe AR, and the other bead of the same shape, must be considered the latest objects in this evacuation, being assigned to Period V. The stratigraphical position of these objects suggests that they were thrown out on the occasion of a thorough clearance of the chamber. This can first have occurred long after the building of the grave and long after offerings had ceased to be made at the facade. The passage grave facade was already in a strong state of disrepair, and the votive pots had fallen from their position on the slabs covering the kerb, many of these slabs having themselves also already fallen. The most recent dateable material in the evacuation layer also belongs to Period IV or V. Before then no evacuations had apparently taken place although it must be admitted that the earliest material from the evacuation layers is considerably later than the erection of the grave.

The material in the upper evacuation layer to some extent confirms the dating of the first evacuation. It contained the massive, thick-butted axe E + HY, which is assigned to Period V, and the tanged arrow-head UK, which must belong to a very late part of the Funnel Beaker Culture. This later evacuation layer almost covered the passage grave kerb. It lay uppermost in a stony layer 30-40 cm. above the lower evacuation layer, and only a few centimetres below the top of the kerb.

A final evacuation comprises a piece of a Corded Ware vessel HBQ, from the Ground Grave or the Upper Grave period. It lay in fragments, but as a separate unit, between stones at the extreme periphery of the mound at the far western end of the excavation, above vessel HBD. This evacuation was probably carried out by the people who last took possession of the chamber, and whose graves lay undisturbed for nearly 4000 years, until the first investigation of the chamber in 1890.

The link between the first evacuation and the cessation of offerings may be an indication that clearing of the chamber was not carried out for solely practical reasons. There can, however, have been many reasons. The decay which marks Jordhøj's facade between the last offering and the first evacuation suggests that the people who originally built it had relinquished it for one reason or another. The evacuation can thus have been carried out by newly arrived people who had taken the grave into use, and in that connection cleared up after their predecessors, who did not concern them, even though they belonged to the same culture. This does not explain, however, why the custom of placing votive pots at the facade ceased. On the basis of the admittedly relatively few modem investigations which have been carried out, there would seem to be a connection between the cessation of offerings and the earliest evacuations, also at other sites [43]. If this connection is general, it may mean that there was a change in the religious tradition which lay behind the building of the passage graves, involving the abandonment of offerings and a loss of piety towards the deceased [44]. Since one cannot take it for granted, however, that the earliest material in the chamber postdates any evacuations, owing to the possibility that it had by some chance escaped clearing, the problem is difficult to solve on the basis of early investigations [45].

The many votive finds at passage graves and Middle Neolithic dolmens in the south Scandinavian megalithic area are in the case of the earliest material almost without exception contemporaneous with the building of the grave, which usually occurred in Period IB [46]. They do not, however, cease everywhere at the same time. At Jordhøj, the largest and most recent part of the votive vessels belonged to Period III, which is not unusual for northern Jutland, whereas it is a rarity in southern Jutland to find vessels later than Period II [47], and the same is true of Langeland, where offerings are largely confined to Period I B and the earliest part of Period II [ 48]. In Zealand and Scania, on the other hand, the offerings apparently continued into Periods III and IV [49]. Recent excavations and publications showing the original location of pottery are still lacking for these areas, however, so some of the late material may have been evacuated from the chambers [50]. Just as at Jordhøj, the Funnel Beaker people continued to bury their dead in the passage graves after the offerings had ceased, in Jutland, though, more sporadically than before. Although both in north and south Jutland a very large number of megalithic graves nevertheless contain finds assigned to Periods IV and V, they have often, like Jordhøj, been cleared to make way for other folk, of the Corded Ware Culture or Late Neolithic, so even the latest Funnel Beaker material has been evacuated from the chamber.

Poul Kjærum