Bronze Age Settlement and Land-Use in Thy, Northwest Denmark Vol. II Bronze Age Settlement and Land-Use in Thy, Northwest Denmark Vol. II Edited by Jens-Henrik Bech, Berit Valentin Eriksen & Kristian Kristiansen Museum Thy Jutland Archaeological Society Bronze Age Settlement and Land-Use in Thy, Northwest Denmark, Vol. II © The authors and Jutland Archaeological Society 2018 Layout and cover: Jens Nygaard and Ea Rasmussen Translation, language revision and proofreading: Anne Bloch and David Earle Robinson, HSLS, Ebeltoft Graphics: Lars Foged Thomsen Excavation photos: Museum Thy Printed by Narayana Press, Gylling Type: ITC New Baskerville Paper: Hello Silk, 130 g Binding: Buchwerk, Darmstadt Jutland Archaeological Society Publications 102 ISBN: 978-87-93423-22-0 ISSN: 0107-2854 Jutland Archaeological Society Moesgård DK-8270 Højbjerg Distribution: Aarhus University Press Finlandsgade 29 DK-8200 Århus N Published with the support of: Farumgaard-Fonden Stiftelsen Riksbankens Jubileumsfond Front cover: Trial trench 15 at Bjerre Enge, spring 1990. Photo: J.-H. Bech. Contents Chapter 11 Bjerre Enge – landscape, environment and settlement during the Bronze Age . . . . . . . . 11 Between land and sea Bjerre Sø and Bjerre Enge The timeframe for settlement at Bjerre Enge Stratigraphy and sand drift Peat formation and paludiication at Bjerre Enge around 800 BC Bronze Age settlement at Bjerre Enge Traces of cultivation at Bjerre Enge Bronze Age settlement at Bjerre Enge: habitation area, territory and resource areas Burial mounds and settlement at Bjerre Enge Bjerre in a regional perspective Conclusion 11 13 17 17 21 24 33 36 37 41 47 Chapter 12 Bjerre 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Houses and associated structures Post structures and ditches at Bjerre 2 Ard marks and cattle hoof prints Settlement structure 57 71 75 76 Chapter 13 Bjerre 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 History Topography and stratigraphy State of preservation. Excavation strategy Houses Other culture layers at Bjerre 3 Other structures at Bjerre 3 Flint workshops at Bjerre 3 79 79 80 80 81 87 87 87 Chapter 14 Bjerre 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Introduction Houses and other structures Midden area Features outside the houses Settlement structure 89 93 106 107 107 Chapter 15 Bjerre 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Introduction Ard marks at the top of the culture layer Structures below upper ard marks Ard marks beneath the settlement Settlement structure 111 117 118 129 130 Chapter 16 Bjerre 4 – settlement, cemetery and ield system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 Introduction Bjerre 4A Bjerre 4B The ield system Conclusion 133 134 138 138 148 Chapter 17 Results from trial excavations in Bjerre Enge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Bjerre 1 Bjerre 5A, 5B and 5C Bjerre 8 and 9 Bjerre 10A-B and 11 Bjerre 12 Bjerre 13 Bjerre 14 Bjerre 15 151 155 157 158 159 159 160 160 Chapter 18 Pottery from the Early and Late Bronze Age, Bjerre 1-13 . . . . . . . . . . . . . . . . . . . . . . . . 161 The Early and Late Bronze Age pottery from Bjerre Technical observations, Early Bronze Age and Late Bronze Age Vessel forms Vessel forms from the Early Bronze Age Vessel forms in the Late Bronze Age Distribution of Early and Late Bronze Age vessels Use of vessels in the Early Bronze Age 167 171 176 178 205 221 235 Conclusion Catalogue of pottery from Bjerre illustrated by drawings in chapter 18 Bjerre 1 Bjerre 2 Bjerre 3 Bjerre 4A Bjerre 6 Bjerre 7 Bjerre 8 Bjerre miscellaneous 240 245 245 245 247 248 249 251 253 253 Chapter 19 On the provenance of the pottery from Bjerre . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255 Introduction Sample preparation and measuring technique Comparison with regional sites – Sundby, Legaard and Egshvile 255 256 262 Chapter 20 The use of pottery vessels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 Introduction Sampling Analytical techniques Methods Interpretation Results Discussion 265 266 267 267 269 269 270 Chapter 21 Bronze Age lint-working at Bjerre, Thy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281 Introduction Typological proile Non-lint stone tool inventory Technological proile – the Early Bronze Age inventories Technological proile – Late Bronze Age Bjerre 7 Raw material economy Find context and spatial patterning Discussion 281 284 303 307 320 322 336 338 Chapter 22 Functional analysis of stone tools from Early Bronze Age Bjerre . . . . . . . . . . . . . . . . . . . 349 Introduction The archaeological assemblage Methodology Results of the analyses and discussion Conclusions and applications 349 349 353 354 361 Chapter 23 Blade knives and strike-a-lights from Bjerre 7: A functional study of two special implements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365 Background Retouched blade knives Strike-a-lights Concluding remarks 365 365 369 371 Chapter 24 Amber inds from the Bronze Age of Thy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375 Introduction Bjerre 6, an Early Bronze Age house Bjerre 7, a Late Bronze Age house Other Bronze Age sites with amber inds Amber collection and export during the Bronze Age of Thy Conclusion How much amber? 375 375 378 381 382 383 385 Chapter 25 Bronze Age agriculture, land use and vegetation at Bjerre Enge based on the results of archaeobotanical analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387 Introduction Plant macrofossil analysis A comparison of the Bjerre sites Conclusion Tables 387 388 403 408 412 Chapter 26 Were the Bronze Age ields at Bjerre 4 manured? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459 Introduction Fieldwork and analytical methods Results Discussion Conclusion 459 459 462 462 467 Chapter 27 Bronze Age animal husbandry: The faunal remains from Bjerre Enge . . . . . . . . . . . . . . 469 Introduction The material Cattle bones Bones of other domesticated animals Game mammals Summary and discussion 469 469 470 471 472 472 Chapter 28 The Bronze Age Settlement at Aas, eastern Thy, Denmark . . . . . . . . . . . . . . . . . . . . . . . 477 Background Topography Archaeological history The settlement structure at Aas From Aas to eastern Thy – an attempt to test the models from Aas 477 477 477 478 495 Chapter 29 Legaard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505 Introduction History Topography Research strategy The Bronze Age settlement Discussion Conclusion Legaard house III Sønderhå 5 – THY 2788 505 505 506 506 508 518 528 531 534 Chapter 30 Early Bronze Age pottery from Legaard. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 539 Diagnostic sherds – Early Bronze Age Vessel forms, sizes and type identiications Dating with the aid of sherds Vessel usage Conclusion 539 541 544 545 546 Chapter 31 Archaeobotanical investigations based on charred material from Legaard . . . . . . . . . . 549 Introduction Results Conclusions Summary 549 554 556 557 Chapter 32 Klostergård . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 559 Two-aisled houses Three-aisled houses Parallels to the houses at Klostergård Animal pens and enclosures Ditches and dykes Stone pile N501 The settlement at Klostergård 560 563 568 570 574 575 575 Chapter 29 Legaard Martin Mikkelsen & Kristian Kristiansen With contributions by Bente Draiby and Timothy Earle Introduction At the site of Legaard, situated centrally in Thy (ig. 29.1) c. 25 km southwest of Bjerre, parts of a Bronze Age settlement area were investigated by Thy Archaeological Project (TAP)1 over the course of several successive excavation campaigns. Some other features, including Late Iron Age pithouses, were also excavated, but these are not relevant here. Although the site’s three-aisled Bronze Age longhouses and some other features are similar to those seen at Bjerre, there are also a number of differences. Legaard is, like Klostergård (Olsen vol. II, chap. 32), situated on high moraine ground as far inland as is possible in Thy. All the longhouses were found within a relatively limited area of c. 4000 m2 and the settlement was inhabited until at least the inal part of the Bronze Age. Most importantly, the site was found to include two relatively large longhouses, houses III and IX. The best preserved of these was house III, which immediately attracted attention because it had a byre in its central part, prompting discussions on house construction, the organisation of interior activities, accommodation of animals and the collection of manure and, inally, social status. As a consequence of these discussions, the main focus here will be on longhouse III, while the other features will only be dealt with in brief. History N Bjerre LEGAARD 0 20 40 km Figure 29.1. The location of Legaard in relation to Bjerre. In 1990, TAP began research in selected parts of Thy (Bech 2003; Earle et al. 1998; Kristiansen 1998). One of the main study areas was Sønderhå parish, where a number of research methods were employed. In 1994, large areas in the northern and northeastern parts of the parish were investigated by a number of detailed surveys. In one of these, a number of inds from both the Early and Late Bronze Age were located on the slopes of a small hilltop at a site named Legaard2 (ig. 29.2). These were not the only inds recorded from the area. Historical maps show two barrows marked at the edge of the hilltop at Lægaard, and in 1912 these were recorded by the National Museum. At that time, it was noted that several Late Bronze Age urns had been found in the eastern barrow.3 Before TAP began work there was no other information on inds from the hilltop itself, but just to the east of it a settlement from the Late Pre-Roman and the Early Roman Iron Age was recorded. Until 1990, no other sites were recorded in the Danish Agency for Culture’s register Sites and Monuments (in Danish: Fund og Fortidsminder) within a 400 m radius of the hill. LEGAARD 505 N Legaard Sønderhå 5 0 500 m Figure 29.2. Topography around Legaard. Topography Legaard is situated in the northeastern corner of Sønderhå parish. To the north of the hill is a quite large low-lying meadow, through which a stream runs from east to west, ending in the lake Ove Sø, from which a second stream runs into the North Sea. There is another relatively low-lying area about 500 m south of Legaard, and c. 1.4 km east of Legaard is the larger hilly area Bjergene, which rises to 56 m a.m.s.l. (ig. 29.2). Looking at the immediate vicinity of Legaard, the top of the small moraine hill is rather lat and has an area of c. 100 x 100 m. Its maximum height is c. 24 m a.m.s.l. The subsoil mainly consists of clayey gravel but on the northwestern part of the hill it becomes sandier. Relatively steep slopes run down towards meadows to the northeast. Research strategy With research ongoing between 1994 and 1997 and with a number of different research aspects, it is relevant to clarify the various steps in the research and 506 describe the choices and priorities in more detail. Figure 29.3 illustrates these various methodological aspects and stages in the research. Survey The irst step in the research strategy at Legaard was an intensive survey of a 5 ha area in 1994. This was conducted by TAP and followed the normal ield-walking procedure of an initial line-walking followed by block-walking (Westphal 2000). The area covered by the survey extended from the northern part of the hill and a terrace on the slope, turning northwards down towards the eastern end of the lake Nørhå. Finds at the site included worked lint, and some burnt lint and pottery were also recorded. Only a small number of the inds could be dated with certainty. A few blades are probably Mesolithic or Early Neolithic in date and one or two lint axes and a transverse arrowhead are Neolithic. A presumed asymmetrical sickle is from the Early Bronze Age, while another sickle or dagger is from the Late Neolithic or Early Bronze Age. A retouched blade knife is from the Late Bronze Age (see Eriksen vol. MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN N Line-walking Block-walking Ploughzone-testing 1995 1996 1997 Figure 29.3. Legaard. The study area. II, chap. 21; Jensen vol. II, chap. 23), as are a few potsherds. At least one ind – a whetstone – is from the Late Iron Age. Most of the remaining worked lints are lakes, but the inds also include 15-20 scrapers, 8-11 borers and three cores. These inds were dated within the time frame of the Neolithic/ Bronze Age. In general, the inds recovered during the survey indicated activities on and around the rather small hill in the Neolithic, Early and Late Bronze Age and Late Iron Age. Based on the Bronze Age inds at Legaard, the site was selected for further research. Plough-zone sampling In 1995, the initial survey was followed up by ploughzone sampling carried out by TAP (Steinberg 1997).4 From an area of more than 4 ha, 34 samples (out of a total of 63) were screened. In the course of the sampling a culture layer was found below the modern plough soil to the north and, especially, to the southwest of the hill (Steinberg 1997, 421, ig. H.6). The sampling showed that most worked lint could be found to the northwest and north of the hill, with a maximum in an area c. 100 m to the north (Steinberg 1997, 420, ig. H.5). The area where longhouses and other features LEGAARD 507 from both the Early and the Late Bronze Age were later found did not show high levels of worked lint. In his analysis J.M. Steinberg concluded that “there are no concentrations of lakes associated with these houses” but also that “Overall THY 3414 is a multi-component site with a large lake bearing sub-plough-zone deposit that makes the site unreliable for plough-zone testing” (Steinberg 1997, 418).5 N Trial excavation and feature excavations 1995-97 On basis of the survey results and the topography, an exploratory trench was cut on top of the hill. This revealed a number of features, some of which seemed to be from houses. Modern ploughing had reached the subsoil almost all over the hilltop and future ploughing threatened the site. Due to this situation, funds for further research were granted by the Rigsantikvar – the Keeper of National Antiquities. This led to conirmation of the initial site interpretation when house I was uncovered and excavated in 1995. Over the next two years the excavated area was gradually extended and further exploratory trenches were cut. The excavation methods involved removal of the plough soil by machine, after which all features were mapped and sectioned. Unlike most other rescue excavations, the participation of TAP made it possible to remove the ill from most features excavated in 1996-97. The soil was therefore not only taken from postholes associated with the longhouses and a number of pits, but also from most of the other features that could not immediately be interpreted during the excavation. Subsequently, TAP processed all the soil samples by lotation (see Stika vol. II, chap. 31). TAP also took a number of subsoil samples in the excavated area for phosphate analysis (Lienemann 1999)(ig. 29.11). The Bronze Age settlement A great number of features were found in the excavated area (ig. 29.4). Most can be dated to the Early or Late Bronze Age, but some are from the Late Iron Age and others may be from the Late Neolithic.6 The Early Bronze Age features (ig. 29.5) include at least three to ive longhouses (I, III, IX, II? and IV?). The Late Bronze Age features include six to nine longhouses (II?, IV?, V, VI, VII, XII, XIII, XIV and X?), a four-poster structure (XI), a ditch (N307), a presumed storage pit (N768) and a number of other pits. Another ditch (N618) and several ire features (ire pits and cooking pits) and other pits can only be dated broadly to the Bronze Age. The primary focus of this article will be on the Early Bronze Age, especially house III, with only a general account being given of the Late Bronze Age features. Longhouses from the Early Bronze Age 0 10 m Figure 29.4. Legaard. Plan of the excavated area. 508 20 House III This house was found in the western part of the rather lat hilltop (ig. 29.5). The house site was generally well preserved (ig. 29.6-7), although the postholes in MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN N House I Ditch N618 House III Ditch N307 House IX House II House IV House X House V House XIV House VI House XIII House VII Four-post feature XI House XII Storage pit N773 0 25 m Figure 29.5. Legaard. Main features from the Early and Late Bronze Age. the westernmost, and now most low-lying, part of the house had been almost totally ploughed away. Most of the features in the area are associated with house III, but a few features are either earlier or later. House III was 33.5 m long and 7.5-8 m wide and thereby covered an area of about 260 m2. Looking at its general layout, the postholes reveal that it had 12 sets of roof-bearing posts, two of which were found beneath the base of a later pit (N202). The distance between the postholes in each set varies between 3.3 and 3.7 m and their longitudinal span was, respectively, 2.6, 3.1, 2.1, 2.7, 2.4, 1.9, 1.8, 2.9, 2.4, 2.4 and 4.5 m. Wall posts and entrances Almost all the wall postholes were preserved. Only where later pit N202 had disturbed part of the northern side, they were absent. A total of 42 wall postholes LEGAARD 509 N N213 N214 N215 N212 N216 N219 N222 N223 N224 N247 N297 N244 N356-357 N267 N266 N261 N279 N210 N248 N278 N257 N259 N277 Ditch N307 N250 N252 N347 N286 N317 N285 N348 N284 N258 N308 N260 N204 House III N203 N243 N320 N227 N265 N322 N228 N319 N229 N282 N207 N226 N202 N251 N321 N230 N242 N239 N238 N232 N231 N236 N341 N340 House IX N877 N855 N870 N868 N886 N866 N869 N853 N867 N858 0 5 N850 N880 N887 N859 N871 N879 N881 N882 N864 N887 N860 N862 N890 10 m Figure 29.6. Legaard. Plan of houses III and IX showing features mentioned in the text. Legend for ire features: See vol. I, appendix A. were found, including one beneath the base of pit N202. The house had straight walls and rounded gables. The average distance between the wall postholes was c. 1.8 m. Although this is not clearly visible, it is suggested that the house had an entrance in each long side: between N222 and N223 in the northern wall and between N242 and N243 in the southern. The distance between the postholes at the entrances is c. 2.2-2.3 m and thereby greater than between the other wall postholes. This interpretation appears to be supported by the absence of traces of stall partitions inside the house at the suggested positions of the entrances (see below). Room partitions The house was divided into three rooms by two partition walls. The western partition was located between the third and the fourth sets of roof posts from the west and the eastern partition was between the third and fourth sets of roof posts from the east. The west- 510 ern partition was marked by seven postholes and the slightly greater distance between postholes N277 and N278 in the middle of the house suggests an opening or a ‘door’ here. Although pit N202 and ditch N307 have disturbed the picture, the eastern partition seems to have been built in exactly the same way, with the slightly greater distance between postholes N284 and N285 also representing an opening or a ‘door’ in the partition. Posts and postpipes In most of the postholes for the roof posts and wall posts, and in some of those for the partition posts, there were more or less clear traces of the post, i.e. a postpipe. This makes it possible to shed light on some constructional details that rarely are seen (see also Draiby this chap. p. 531 f). Many of these post traces probably relect the actual size and form of the post.7 Whole tree trunks (e.g. N248, N251 and N252) and cloven timbers (e.g. N244, N247 and N250) appear to have been used for both MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN A B C D Figure 29.7. Legaard excavation. A: House III, part of the western end uncovered, seen from the south. The partition wall is clearly seen with two roof posts N248 and N250, to the right and left of the ranging rod, respectively. In the foreground two wall posts. B: House III, with a pair of stall dividers at the southern side of the byre with wall posts N239 (left) and N238 (right). C: House III, seen from the east. Wall post and posts in partition walls are marked with paper plates. The slightly darker area in the middle of the house is due to moist soil arising from the use of a garden sprinkler. D: House III, seen from the west. Wall post and posts in partition walls are marked with paper plates. roof and wall posts. Although it was often dificult to determine the form with certainty, the cloven posts appear to have been both crescent-shaped in crosssection (N210, N236, N244, N257 and N279), i.e. taken from the outer part of trees, as well as more rectangular (N219, N222, N224 and N259), i.e. cloven from the central part. As for the size of the posts, one of the most clearly visible postpipes from a roof post was in the southwesternmost roof posthole, N244 (ig. 29.8): This was crescent shaped and had been cloven from a tree with a diameter of at least 0.53 m. A few other roof posts were also possibly from trees of at least this diameter, while the diameter of the trees producing the posts in the other roof postholes was probably between 0.4 and 0.5 m. The crescentshaped posts in wall postholes N236 and N210 were probably cloven from trees with a diameter of c. 0.5 m, while the other wall posts, such as N229, appear LEGAARD 511 to have been rectangular with a width of c. 0.35 m. In general, the wall postholes and the wall posts appear to have been rather smaller than the roof postholes and the roof posts. The posts had mostly been placed in the central part of the roof- and the wall postholes, but this was not always the case. For instance, the post was placed along the outside of wall postholes N226-N231 in the east gable and the same seems to have been the case in postholes N212-N216 in the northern wall at the western end. In the wall construction, the crescent-shaped posts (N203-204, N210, N227, N236 and N242-243) all appear to have had their lat side turned outwards, suggesting that the wall planks were attached to their outer surface. It is also noteworthy that posthole N216 contained traces of both a post for the wall and a post for the inner wall section. The many clear traces of the posts in the postholes may indicate that the posts were not pulled up after the house was abandoned, although this is dificult to say with certainty (see note 7). Benches and ire features Turning to the functional aspects of house III, some presumed postholes indicated a ‘connection’ between the wall and the two western sets of roof posts in the western room. These postholes were 0.04-0.20 m deep and shallower than the roof- and wall postholes. Most of the ‘connecting’ postholes were rounded rectangular and contained no visible traces of postpipes either in plan or in section. They can be interpreted as being associated with possible ‘benches’, perhaps supporting plank beds placed along the walls of the house. Similar postholes were not found in the better preserved eastern part of the house. At the centre of the western room was pit N297, dug in two or three phases. A black charcoal-coloured layer containing small pieces of ire-cracked stones shows that it had been used as a cooking pit. Cooking pit N322 was found at the centre of the eastern room and the base of cooking pit N319 was located between roof postholes N263-264. While N297 was 0.2 m deep, N322 was only 0.14 m deep; both had a slightly rounded bottom (ig. 29.9). In the southeastern corner of the house was a latbottomed pit, N321, the ill of which mostly seemed to be ash (ig. 29.9). In the southwestern corner were one or two ‘pits’ (N356-357) that were very shallow – a maximum of 0.06 m deep – with no traces of ash. These were found to contain about 30 pieces of worked lint and 12 potsherds. Whether or not the pits found in the southwestern and southeastern corners should be assigned to the house is uncertain, but it seems remarkable that they occupied a similar position in the western and eastern rooms. It seems unlikely that the pits in the southwestern corner could have been in use at the same time as postholes N266-267 supported a possible ‘bench’. The byre In the central room were several more or less straight ‘ditches’ running out from the walls towards the postholes for the roof-bearing posts (subsequently referred to below as roof posts). These ditches have N244 N257 N265 N210 N230 N238 0 1 m Figure 29.8. Legaard. House III, section through selected postholes for roof posts (N244, N257 and N265) and wall posts (N210, N230 and N238). 512 MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN N322 N297 N328 N321 0 1 m Figure 29.9. Legaard. House III, sections through cooking pits N297 and N322, pit N321 and stall partition N328. been interpreted as remains of stall dividers or partitions. Along the northern side they appear to be arranged in pairs, made up of one relatively dark and one relatively light ditch. At the southern side every second ditch was also slightly darker than the rest. These differences in colour suggest that there were two phases of stall dividers, with those in the earlier phase having a slightly lighter ill than those in the later. The number of stall dividers suggests there were spaces for seven or perhaps eight animals at the northern side, with probably the same number along the southern side, giving accommodation for a total for 14 to 16 animals in the byre in the central part of house III. When sectioning some of the stall dividers, it proved possible to observe a number of stake holes below the shallow ditches (ig. 29.9). The ill in these stake holes was much lighter than that in the ditches and due to this, and the clayey subsoil containing a lot of small stones, the stake holes were dificult to see and therefore to section. If we had not detected the stall dividers through the presence of the upper ditches, we would not have seen the stake holes in the subsoil. Consequently, if modern ploughing had gone c. 10 cm deeper, we probably would not have discovered that the house had stall partitions in its central room, even though stake holes would still have been present. It has been asked why there are so few examples of byres, given the several hundred house plans available from the Bronze Age (Rasmussen 1999, 283). Based on observations at Legaard houses III and IX, traces of stall dividers, in the form of stake holes, may well have been present without being recognised. However, in our opinion, this cannot explain the rarity of deinite traces of byres in Early Bronze Age longhouses (see also Bech & Rasmussen vol. I, chap. 2). Other features It is possible that other features inside house III date from its lifetime, but there are several features that cannot be contemporaneous with the house. Fire pit N347 could be seen to be stratigraphically earlier than stall dividers N340-341. If the central room had stall dividers when house III was built, then ire pit N347 demonstrates activity in the area prior to its construction. But if stall dividers were not established from the beginning, ire pit N347 could have been contemporaneous with an early phase in the lifetime of house III. Pit N308 may also be stratigraphically older than posthole N258 for a roof post, but the overlap between them was minor and therefore not certain. A number of features are clearly stratigraphically later than house III. Pit N348 cuts posthole N260, for a roof post, and ditch N307 cuts both pit N348 and three postholes from house III. The same deinitely also applies to the large Iron Age pit N202. Pottery Some pottery was found in features belonging to house III – in roof postholes (N257, N258?, N259, and N317), wall postholes (N239 and N232-233) and partition postholes (N283, N286 and N316). Unfortunately, none of the pottery was found and recorded by sectioning of the postholes and in no cases is it certain whether the pottery is from the beginning or the end of the house’s lifetime. The distribution of pottery in house III reveals, remarkably, that it was all found in the eastern half of the longhouse, both to the west and the east of, and within, the eastern partition wall. It has been considered whether the concentration of pottery in LEGAARD 513 postholes in and on both sides of the eastern partition wall may indicate that it entered these postholes at the same time, probably when the posts were removed, and that the use of earthenware jars, at least at the end of the lifetime of house III, was restricted to its eastern part.8 The pottery from house III is all of a coarse ware and, although exact dating is dificult (Rasmussen 1993), it is certainly of Early Bronze Age character (for further discussion of the pottery from house III, see Kristensen below). Radiocarbon dating In order to date house III, ive samples were submitted for 14C analysis (see vol. I, appendix A, house no. 38 and ig. 29.10).9 A sample of charcoal (Alnus sp.) from roof posthole N250 was dated to 3195 ± 50 BP (AAR-6565; vol. I, appendix B). This calibrates at 1 σ to 1511-1416 BC, i.e. period II of the Early Bronze Age. A charred grain of barley (Hordeum vulgare) from wall posthole N236 was also analysed, giving a date to 3065 ± 50 BP (AAR-6564; vol. I, appendix B). At 1 σ, this gives a calibrated date of 1398-1268 BC, i.e. the second half of period II and the very irst part of period III. Finally, a grain of naked barley (Hordeum vulgare var. nudum) from wall posthole N222 was dated to 3015 ± 50 BP (LuS-6109; vol. I, appendix B). Calibrated at 1 σ, this results in a date of 1381-1133 BC, i.e. at the end of period II and/or in period III. In combination, the two grain samples give a date of 3040 ± 35 BP, which calibrates at 1 σ to 1384-1231 BC. The average date for all three samples is 3092 ± 29, which calibrates at 1 σ to 1411-1303 BC (ig. 29.10 and vol. I, appendix B). The use of alder (Alnus sp.) for fuel may seem unusual as it is a tree that grows well in damp and wet conditions. It cannot compete with birch in terms of heat output, but it burns slowly and is therefore good to keep the ire going in the hearth. It has been managed by coppicing and also been used for making charcoal in historical times. At Legaard, it was most probably cut near bogs where peat could also be cut for fuel. If it had been coppiced, the age of the twigs would be close to the radiocarbon date. We cannot determine the stratigraphic age of the 14C samples as they were all taken from postholes, but the time interval between the dates suggests that the alder date could refer to the early period of the house’s use, while the charred grain, with its later dates, is from a later period of use. As alder does not develop into a large tree in Northern Europe – it matures after about 30 years and can reach an age of 150 years – the fact that the date for the charcoal sample falls in the 15th century BC indicates that the house construction cannot be much later than 1400 BC. Depending on the biological age of the sample (twigs or tree trunk, ranging from 30 to 150 years of age), 514 the evidence suggests that house III was constructed during the middle part of period II (with the beginning of period II placed at between 1500 and 1450 BC (see discussion in Randsborg 2006, chap. III-IV)). The later dates for the charred grain may relate to the end of house III’s lifetime. Based on the radiocarbon dates, it seems reasonable to suggest that house III’s lifespan lay during period II, perhaps ending shortly before or in the early part of period III (beginning around 1330 BC)(Olsen et al. 2011). In contrast to the above, two samples from different parts of cooking pit N297, in the western end of house III, dated to the Mesolithic Late Ertebølle culture and the Late Neolithic, respectively (K-6906-6907; vol. I, appendix B). An analysis of the material recovered from pit N297 shows a high content of sedges (Carex sp.), indicating that the charcoal from this pit derives from peat (Stika vol. II, chap. 31, ig. 31.1). Given that the 14C analysis provides a true date for the charcoal, the unexpectedly early date may be the result of peat of Mesolithic age being burnt in the Bronze Age cooking pit (Bech 2003). It thereby constitutes another example of the use of peat for fuel in the Early Bronze Age, as previously demonstrated at a barrow at Damsgård in Sønderhå parish, only 3 km south of Legaard (Olsen & Bech 1996). These samples were therefore not suitable for dating house III. House IX House IX was found parallel with and just 4 m south of house III (ig. 29.6) and we believe that the two houses were similar in layout and use, one probably replacing the other over time. The western end of house IX was badly preserved, probably due to modern ploughing, while the eastern part had fared better but was dificult to distinguish from one or two overlapping houses and other features in that area.10 Due to the poor preservation of the western end of the house its length is uncertain, but we suggest that it was c. 33-34 m long. Its width in the eastern part was 7.6 m and was probably the same through rest of the house. House IX therefore had an area of about 255 m2. It apparently had nine sets of roof posts, but it is uncertain whether the westernmost and poorly preserved presumed posthole N855 was a posthole for a roof- or a wall post. One roof posthole is missing from the central part due to the later ‘pit’ N850, and the absence of another from the southeastern part is probably because we did not recognise it during the excavation in 1996. The base of a roof posthole was found beneath the presumed cooking pit N862. The distance between the roof posts in each pair was 3.43.6 m and the longitudinal spans were 2.9, 4.4, 3.1, 3.9, 3.3, 3.2, 4 and 3.9 m. MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN OxCal v4.2.3 Bronk Ramsey (2013); r:5 IntCal13 atmospheric curve (Reimer et al. 2013) LN II EBA I EBA II EBA III LBA IV LBA V Legaard House I LuS-6112 Legaard House III K-6906 K-6907 AAR-6565 AAR-6564 LuS-6109 Av. AAR 6564-65, LuS 6109 Legaard House IX AAR-6553 AAR-6552 LuS-6111 LuS-6110 Av. AAR 6552, LuS 6110-11 4500 4000 3500 3000 2500 2000 1500 1000 Calibrated date (calBC) Figure 29.10. Radiocarbon dates from Legaard. Traces of the wall postholes were preserved in the northern side and in the eastern end. The average distance between the wall postholes in the northern side was 2.4 m. They were generally between 0.35 and 0.45 m in diameter. The building might have been divided up into three rooms. In the western part, the two presumed postholes (N886 and N887) could be interpreted as postholes for partition posts at a door opening. Similarly, there may have been partition postholes between the roof postholes in the third set from the east. Unfortunately, the gable of house II was placed here, making interpretation dificult. Traces of the post – the postpipe – could only be seen in a few of the postholes. The clearest traces could be seen in the third set of roof postholes from the west (N858-859). The posts in this set were crescent shaped and the roof post had been cleaved from a tree with a diameter of at least 0.5 m. Most of the other roof postholes were smaller than those LEGAARD 515 in the third set, so the size of the posts in this set may not be representative of the general situation. Only in the wall posthole N877 could traces of the wall post be seen. In the presumed central room were two ‘ditches’ (N879-N880), thought to be stall dividers along the northern side; they were similar to those found in house III. However, we were unable to ind stake holes either under these ditches or along both sides of the central room. The central room also had some irregular shallow features (N853, N864, N881, N882 and N887) with a rather pale deposit interpreted as remains of the loor level. 0 4 8 metres Figure 29.11. Legaard. Phosphate concentrations in the area of houses III and IX. At Legaard, samples for phosphate analysis were taken every 2 m within an area of c. 1100 m2 covering house III, the central and western parts of house IX and their immediate surroundings. Samples were not taken around the eastern part of house IX, because several other houses had been placed in the same area. The phosphate values lie within the range 178-698 ppm. Before looking at the results it should be noted that the westernmost parts of houses III and IX were not as well preserved as the central and eastern parts. Looking at the results, it must irst be noted that the byre with two construction phases in house III does not show high levels of phosphate. According to J. Lienemann (1999), the p-concentrations inside the house are similar to those in houses only used for living and storage, but with no cattle or other animals accommodated inside. This means that the stable could only have been used temporarily (see also Bech & Rasmussen vol. I, chap. 2). Alternatively, the low values might indicate that dung was continuously removed from the byre. In house IX, part of the byre shows relative high values, maybe because slight remains of the loor level were preserved in this part of the house. Relatively high values were also found just outside the entrance in the northern wall of house III and along its northeastern and southeastern walls. This could indicate that dung was placed along the outside of the walls in the easternmost part of house III. Slightly higher values are evident just outside the entrance in the southern wall of house III, while the same appears not to be the case outside the entrances to house IX if they occupied the same positions as in house III 516 MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN Although a number of other features were found ‘within’ house IX, none of them can with reasonable certainty be related to the house. This is true of several shallow cooking pits (N866-871) in the western part, while in the eastern part it has been impossible to separate features relating to house IX from features relating to houses II and IV. Potsherds were found in four, or probably ive, of the roof postholes and in two of the shallow features in the central part of the house. Nearly all of these were of a coarse ware (Kristensen vol. II, chap. 30). Radiocarbon dating Four samples from house IX have been subjected to 14 C analysis (see vol. I, appendix A, house no. 39, and ig. 29.10). A piece of unidentiied charcoal (from roof posthole N859) has been dated to 3080 ± 55 BP (AAR-6552; vol. I, appendix B) and a charred grain of barley (Hordeum vulgare) from roof posthole N890 produced an almost identical date of 3055 ± 50 BP (LuS-6111; vol. I, appendix B). A charred grain of barley (Hordeum vulgare) from roof posthole N860 was dated to 2970 ± 50 BP (LuS-6110; vol. I, appendix B). Combining these three dates results in a date of 3032 ± 30 BP (vol. I, appendix B). At 1 σ, this dates the samples from the house to 1375-1228 BC. The concurrence between the three 14C dates, including the charcoal date, at the transition between periods II and III, and thereby slightly later than the inal date for house III, suggests that house IX was constructed after house III had been abandoned. This conclusion is also supported by the use of smaller-sized posts in the construction of house IX, something that will be discussed further in the conclusion. A weighted statistical analysis of the 14C dates for charred grain from houses III and IX indicates more than 60% probability that the samples dated from house III are older than those from house IX (J. Heinemeier pers. comm.). If we include the date for the alder charcoal from house III, this interpretation receives further support. It should be added that the question of an overlap between the two houses is also related to their presumed life span. If we assume the latter to be around 60 years, a succession of houses is much more likely than if we assume the lifespan to have been 100 or more years. As there were repairs to the houses, our evaluation is that a 60-year lifespan is the more likely. An unidentiied charred grain from roof posthole N490 has been dated to 3520 ± 45 BP (AAR-6553; vol. I, appendix B). Calibrated at 1 σ, this gives a date of 1908-1772 BC, which is somewhat earlier than the dates from the other three samples and also earlier than seems reasonable on basis of the house type.11 House I House I was found in the northern area on the lat hilltop (ig. 29.5). It was relatively well preserved but a number of other features in the area makes interpretation of some of the elements dificult. In the proposed interpretation, house I was c. 17.5 m long and 5.3-5.5 m wide (ig. 29.12) and therefore covered an area of about 95 m2. It seems to have had ive pairs of roof posts in two almost parallel rows, 1.6-1.9 m apart.12 The distance between the roof posts in the easternmost pair seems to have been slightly less than in the three sets to the west. The longitudinal span between the sets was c. 3.8, 3.3, 3 and 2.9 m. The ‘room’ in the western part of the house was thereby somewhat larger than the other ‘rooms’. The walls were straight with rounded corners to the gables and at the southern side the wall posts were spaced at an average interval of 1 m. It has not been possible to localise entrances, not N N97 N84 House I 0 5 m Figure 29.12. Legaard. House I. Legend for ire features: See vol. I, appendix A. LEGAARD 517 even in the southern wall where all wall postholes appear to have been found. Only one small potsherd was found, in wall posthole N97. This was of a coarse ware and dated to the Early Bronze Age. A large number of features were recorded within the house, especially in the central and eastern parts, but it is dificult to relate these to the house with reasonable certainty. Most could be from activities that were not contemporaneous with the house. Finds recovered from features inside and outside the house document activities during both the Early and the Late Bronze Age. One sample has been 14C dated (see vol. I, appendix A, house no. 37, and ig. 29.10): A charred grain of barley (Hordeum vulgare) from the roof posthole N84 gave a date of 2990 ± 50 BP (LuS-6112; vol. I, appendix B), calibrated at 1 σ to 1286-1126 BC. Although a single 14C date from an area with activities during several periods does not provide a secure date for house I, on the basis of the house type and the 14C date it seems reasonable to suggest that house I is from the Bronze Age period III. Longhouses and other features from the Late Bronze Age Houses II and IV Houses II and IV were found in the eastern part of the excavated area (ig. 29.5), in the same area as the eastern end of house IX (ig. 29.13). The overlapping of these houses and some constructional elements raise a number of questions with respect to interpretations and dating.13 In our interpretation, four pairs of roof postholes have been assigned to houses II and IV. The missing roof posthole in the southeastern part of house IV could be due to poor preservation conditions in this area. The distance between the roof posts appears to have been c. 3.2 m in both houses. The longitudinal spans in house II may have been 4.6, 4.4 and 2.5 m, while those in house IV may have been 5.6, 4.3 and 2.8 m, and the span at the western end of this house is therefore remarkably long. The size of the two houses is uncertain due to the poor preservation of their eastern ends. If they had four sets of roof postholes, they would have an area of c. 95 m2. Based on their positions relative to the roof postholes, a number of cooking pits and other pits might be connected with the western and central parts of house II and the western part of house IV. But it is dificult to assign speciic pits to either of the two houses. Pit N510 must be mentioned because it contained a number of potsherds, some of which can probably be dated to Bronze Age period IV. 518 Based on their uniformity in size, orientation, spacing of the roof postholes and the presumed concentrations of cooking pits, we suggest that house IV followed house II or vice versa. The two houses are dated primarily to the Late Bronze Age period IV, but one of them could have begun in period III. The remaining features from the Late Bronze Age will only be presented in general. Two groups of overlapping longhouses (ig. 29.14) were found in the southern part of the excavated area (ig. 29.5). Houses V-XIV Houses V, VI and VII appear to represent a series of longhouses replacing each other through time. On basis of pottery and their construction, they can be dated to periods IV-(V). If the succession was V, VI and VII, then a slight reduction in size took place through time. One other Late Bronze Age longhouse may also have been located here. The presumed western end of house X was found just east of longhouses II and IV and may represent a further longhouse in the sequence comprising longhouses II-IV and V-VI-VII. Furthermore, it is suggested that houses XII, XIII and XIV comprise another series of longhouses replacing each other, which, based on house type and pottery, can be dated to periods V-VI. These longhouses are smaller than longhouses V-VI-VII, demonstrating a further reduction in size. If longhouse XII was the inal building at the Legaard site, then there seems to have been a general reduction in the size of longhouses throughout the Late Bronze Age. A number of other features found in the southern part of the excavated area can be dated to the Late Bronze Age on basis of the pottery they contained. Unfortunately, it is impossible to relate these features to speciic longhouses and they therefore reinforce the impression of considerable activity having taken place outside and nearby contemporaneous longhouses. Discussion Before discussing various aspects of the Legaard site, a few problems should be pointed out. Firstly, we do not know how much of the total settlement area was actually excavated. Given the large number of features along the eastern side of the excavation and some features found in the easternmost trial trench (ig. 29.4), probably further features existed at the site, including remains of several longhouses to the east and perhaps also to the northeast of the excavated areas. Secondly, the survey and ploughzone testing revealed that, at least to the north of the excavated MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN N 859 N510 House II House IX House IV 0 5 m Figure 29.13. Legaard. Houses II and IV. Legend for ire features: See vol. I, appendix A. House X N House V House VII House XIV House VI House XIII House XII Four-poster feature XI 0 5 Storage pit N768 m Figure 29.14. Legaard. Main features from the Late Bronze Age. Legend for ire features: See vol. I, appendix A. area, there was quite a lot of worked lint in the plough soil (Steinberg 1997, 420, ig. H.5). Although only a few diagnostic lint tools were found, probably some of the worked lint, and perhaps most of it, should be related to activities in the Bronze Age; typical Mesolithic and Neolithic lint tools are virtually absent. Whether or not the area containing worked lint was also used for settlement remains unknown and the majority of lint tool production may have been located at some distance from the contemporaneous settlement site at Legaard, as indicated at Aas in eastern Thy (Mikkelsen vol. II, chap. 28), but contrary to what was observed at Bjerre 3 (Bech vol. II, chap. 13). Thirdly, the presence of a Late Iron Age settlement may have disturbed the picture of Bronze Age activities. Some of the many features that cannot be dated with reasonable certainty may well belong to the Late Iron Age. LEGAARD 519 Settlement continuity at Legaard The three-aisled longhouses III, IX and I can all be dated to the Early Bronze Age periods II-III. These three buildings were all located on the virtually lat top of the hill, although the eastern part of house IX was placed where the terrain begins to slope to the southeast. If we include the 14C dates for charcoal, then the weighted dates from house III are earlier than those for house IX. The dates for grain from house III probably relate to the inal phase of the lifespan of the building, while that for charcoal relates rather to its construction. Based on this, house III could have been built during the early or middle part of period II, and abandoned about 60 years later when house IX was constructed as a replacement parallel to it. House IX then lasted into period III. The two houses had approximately the same size and orientation and both appear to have had three separate rooms with a stall in the central part. However, house III had larger roof posts and wall posts than house IX. It also seems unlikely that two houses would be sited so close to each other at the same time. All in all, it seems reasonable to suggest that house IX replaced house III. Based on the 14C dating evidence, house I is probably from the Early Bronze Age period III, although it would have been better to have had more dates. It also had a completely different orientation to houses III and IX, but was aligned roughly similarly to the other longhouses at the site. A new longhouse orientation might therefore have been introduced with the construction of house I at the Legaard site. We conclude that house I is later than house III and, based on the radiocarbon dates, house I could be contemporaneous with, or somewhat later than, house IX. The latter possibility is indicated by the building’s different orientation. As an alternative, partial contemporaneity between houses I and IX is discussed further below. Based on the house type, with a large number of cooking pits in the western part, and some pottery evidence, it is possible that the two houses II and IV overlap house IX and follow on from each other and should be dated to periods III-IV. Furthermore, based on similar evidence relating to the house types and pottery, the presumed sequence comprising the three-aisled longhouses V, VI, VII, and perhaps longhouse X, can be dated to periods IV-V, and another sequence comprising longhouses XII, XIII and XIV can be dated to periods V-VI. All the longhouses thought to be of Late Bronze Age date were found on the slopes to the south and southeast. This means that the earlier houses were placed on top of the hill, while those that came later were sited on the south- and southwest-facing slopes. In this respect, the Legaard site shows similarities to the Vilhøj site (Mikkelsen vol. II, chap. 28). 520 The question is whether the longhouses excavated at Legaard represent an unbroken sequence extending from period II to at least period VI, as is presumed to be the case for the settlement at Vilhøj near Aas (Mikkelsen vol. II, chap. 28). Continuous settlement in the Legaard area from the Bronze Age period II to the Early Pre-Roman Iron Age is possible if the average lifetime of three-aisled longhouses was c. 100 years and if there was only one farmstead with one longhouse at the site throughout this time. Although a lifetime of c. 60 years might apply to the large longhouses III and IX, it is more doubtful in the case of the smaller longhouses (Mikkelsen vol. II, chap. 28; Bech vol. II, chap. 11). At Legaard, we may be ‘lacking’ longhouses from periods III and V and there may also be missing longhouses from late period VI if the settlement was continuous. Taking the unexcavated areas at Legaard into consideration, it seems reasonable to suggest an unbroken sequence of longhouses extending from period II until the end of the Bronze Age. The character of the settlement at Legaard Let us now turn to the character of the settlement at Legaard: Did it consist of a single farmstead with a single longhouse or was there a single farm with one or two contemporaneous and complementary buildings? As previously outlined, we are hampered by the fact that the settlement area was only partially excavated and we will also consider below whether nearby settlements can be related to the Legaard site. Firstly, however, some possible scenarios can be proposed. During period II and part of period III, two large farmhouses of almost the same proportions and size followed on from one another. They also had the same orientation and both may have been divided into three, with a stable in the central room. They quite evidently represent the same household(s) that wanted to maintain their standing. But the two houses do differ slightly with respect to various elements: the number and arrangement of the pairs of roof posts and the distance between the wall posts and the positions of the room partitions. Furthermore, the postholes, and therefore probably also the posts, in house III were larger than those in house IX. Consequently, the timber in the earlier house was substantially sturdier than in its successor. In conjunction with the 14C dates, these various differences suggest that house III was earlier than house IX, supporting the evidence from the environmental data discussed below. In period III, house IX could have been replaced by house I or even house II, which both have a quite different orientation. This marked a new phase in the settlement layout. If there had been only one farmstead with a single longhouse, the possible change from MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN a 260 m2 longhouse to one of only approximately 95 m2 appears to be a very dramatic reduction in size. However, if houses I or II had no byre, then the reduction in living space could be said to be from c. 155 m2 to 95 m2. K. Kristiansen suggests an alternative interpretation whereby house IX was partly contemporaneous with house I and that this additional or subsidiary house marked a transition to a changed orientation at the site. This conclusion is supported by the absence of ire features in house I. According to this scenario, we would have to assume that, when house IX was abandoned sometime in period III, house II was added to house I as the new residential quarters and the two buildings formed an economic unit with a distance of c. 40 m between them. This then constituted the third occupation phase. The two longhouses had the same orientation, were of about the same size and had their walls apparently constructed in the same way, all of which speaks in favour of them being at least partly contemporaneous. But while house II has a large number of ire-/ cooking pits in its western room, no such pits can be related to house I. Furthermore, the cross span in house II is somewhat greater than in house I. It can be suggested that these differences relect functional variations, with house II being used as residential quarters and house I as a subsidiary building for other activities, maybe even periodically as a byre. This would give further support to the conclusion that they belonged to the same household. If this were the case, the reduction in indoor space following the demise of house IX would be from c. 255 m2 to c. 190 m2, i.e. a minor reduction in farm size compared to the irst proposed model. However, there is a slight chronological problem. House I has been 14C dated to period III. Houses II and IV can be dated on the basis of pottery to an early part of period IV, although a later part of period III cannot be excluded. Assuming that the pottery had a short lifetime, thereby marking the inal phase of the building, it is reasonable to suggest that the two houses overlapped in time, but we need further 14C dates in order to establish a more secure conclusion. In this scenario, only the actual excavated houses at the site are considered. According to the second scenario, we possibly lack a subsidiary building from the earliest farmstead on the site, house III. However, there were also several other settlements in the neighbourhood; these are briely described below and then discussed in relation to the Legaard site. In conclusion, the excavated clusters of Late Bronze Age longhouses appear to form a succession in time, but due to the unexcavated areas to the east of the site, it is impossible at the moment to extend the analysis further. The Legaard settlement in relation to other settlements in the vicinity Looking at the settlements in the near vicinity of Legaard and employing the model of the relationship between settlements and barrows from Aas (Mikkelsen vol. II, chap. 28), it seems possible that there were least seven, and perhaps as many as ten, contemporaneous settlements from periods II-III in an area of c. 9 km2 and within a maximum radius of 2 km of Legaard (ig. 29.15). The evidence supports this scenario if we include barrow groups as indicators of settlements. These groups comprise: Pengshøje. A group of four barrows has been recorded 700-800 m to the west of Legaard. Although no inds are known from these barrows, it seems a reasonable possibility that they date from the Early Bronze Age. We suggest that a settlement from this time can be found nearby. Damsgaard. About 1.5 km southwest of Legaard is a larger group of barrows located south of the farm of Bakkegården. One of these barrows (site no. 110112314) was found to contain a stone cist from a late part of the Early Bronze Age or an early part of the Late Bronze Age and an urn burial was found in the northernmost barrow (site no. 110112-308). We suggest that there was a settlement to the north or northwest of this group of barrows. Sønderhå Vest. About 2 km southwest of Legaard and south of the aforementioned group of barrows, a number of lint lakes and lint tools, including an asymmetrical sickle, were found during survey within an area of at least 70 x 100 m (site no. 110112-277). It is suggested that a settlement from the Late Neolithic and the Early Bronze Age was located here. Sønderhå Nord. About 1.2 km to the south of Legaard, and just northeast of the village of Sønderhå, stray inds, including at least three asymmetrical sickles and some other lint tools (site no. 110112317) in a private collection, indicate the existence of a settlement from the Early Bronze Age in or near this area. There are also records of scattered barrows with no known inds. Hørsted NV. About 0.8-1.2 km to the southeast of Legaard, three barrows (site nos. 110112-201, 110107-15-16) have been recorded on the edge of a rather long hill. There are no known inds from these barrows, but at least two asymmetrical sickles have been found at the western end of the hill. A lint scatter was excavated here and a very small dwelling house from period II/III was uncovered (THY 2788, Sønderhå 5 house I, site no. 110112-313)(see Earle this chap., p. 534 f). The evidence suggests that larger households, such as those in houses III and IX at Legaard, could have had smaller supporting settlements in their LEGAARD 521 N Settlement area Undated barrow Burial from BA Aarup Settlement from BA Hoard 216 9 Bjergene Nord Pengshøje 199 11 Legaard 13 Haagaard 207 Bjergene Vest 14/15 205 17 204 Hørsted NV Damsgaard 313 Sønderhaa Nord 308 317 Kildegård 314 277 0 Sønderhaa Vest 500 m Figure 29.15. Settlements and possible settlement areas around Legaard. vicinity. Conversely, taking the distance between Legaard and this site into consideration, the small building can also be seen as part of an independent settlement. Bjergene Vest. Some 0.7-0.8 km to the east of Legaard, ive barrows have been recorded on a hilltop. A relatively large barrow (site no. 110112-205) was found to contain a grave from period II, and in a long barrow (site no. 110112-204) was a burial from the Early Bronze Age (Aner et al. 2001, Ke 5040-5041). Furthermore, a number of asymmetrical sickles have been recorded in a private collection from this area (Olsen 1991). Consequently, a settlement may have been located in the vicinity of these barrows. Bjergene Nord. About 1.5 km northeast of Legaard and just north of a larger group of barrows containing a number of burials (site nos. 110107-17, 110110-9, -11, -13 -14/15, and maybe 110112-207) from the 522 Early Bronze Age (see below) a survey located a few lint tools from the Late Neolithic and/or the Early Bronze Age and a considerable amount of pottery, some of which is from the Late Bronze Age (site no. 110110-199). Further to the north, settlement traces in the form of postholes and pits are presumed to date from the Late Bronze Age (site no. 110110-216). We suggest that this area was settled at least in the Late Bronze Age. Below the southernmost member of the aforementioned larger group of Bronze Age barrows (site no. 110107-17), a cooking pit was found that produced a radiocarbon date of 3370 ± 155 BP (K-6461) which proves earlier activity in the Bjergene area west of Legaard at the end of the Late Neolithic or during period I of the Early Bronze Age. However, the settlement below and around the barrow, site no. 11010717, dates irst and foremost from the LN I (Earle et MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN al. 1998; Prieto-Martinez 2008)(THY 2758, site no. 110112-259), and excavations did not provide clear evidence of houses contemporaneous with the Legaard site, perhaps due to the limited extent of the excavated area. Based on current evidence, however, the area between Bjergene Nord and Bjergene Vest, which is the higher part of the Bjergene area, is mainly to be perceived as a barrow landscape during the Early Bronze Age (Kristiansen 1998 and below). Aarup, Kildegaard and Haagard. Finally, we can point out these three areas where the topography and nearby barrows indicate that a settlement could have been located here in the Early Bronze Age. The scattered settlement evidence from the microregion around the Legaard site suggests that the model from Aas (Mikkelsen vol. II, chap. 28), i.e. one settlement from Early Bronze Age periods II-III per km2 and ive barrows located nearby, also applies in the vicinity of Legaard. The Legaard settlement in relation to barrows and graves Looking now at the Legaard site, there are two recorded barrows (site nos. 110112-211-212) at the edge of the hilltop (ig. 29.15) and one presumed barrow (site no. 110112-319) just north of the area.14 We have no information on inds from the barrow, site no. 110112-212, but it was probably at least 15 m in diameter and 1.5 m high. A number of urns are stated to have been found in the barrow, site no. 110112-211. This barrow, at least, was probably in use during the Late Bronze Age, given the number of longhouses found at Legaard. The barrow, site no. 110112-319 was located by TAP and there is no information on its size and no known inds from it. Although it is not possible to date these three barrows securely, there are no indications of earlier settlements in the vicinity, for example passage graves or other activities dating from the Funnel Beaker Culture, in the excavated part of the Legaard site. Furthermore, the presumed size of site no. 110111212 appears larger than normal Single Grave barrows and maybe also be larger than average Late Neolithic barrows. A date in the Early Bronze Age therefore seems reasonable, in which case at least two or three barrows were built in relation to the Early Bronze Age settlement at Leegaard, conirming the close relationship between farmstead and barrows seen elsewhere. However, if three to ive barrows were the ‘norm’ for an ordinary settlement, then only two or perhaps three barrows associated with a settlement such as at Legaard, with large houses, seem to be rather few. We suggest that some of the barrows in the nearby barrow groups could have belonged to the Legaard farm, most probably the group on the hilltop Bjergene, which has barrows and rich burials enough for more than one settlement. The evidence therefore suggests that, within the micro-region of Legaard, a complex spatial relationship existed between large and small farms and barrows. Such relationships could have changed over time, as may be indicated by the reduction in house size seen during the Late Bronze Age at the Legaard site (see also discussion in Bech & Olsen vol. I, chap. 4). Houses III and IX and their economic foundation Both the size and the unusual tripartite construction of houses III and IX, with cattle stalls in their central byre and two living compartments, deserve fuller discussion. How should these large farmhouses be understood in relation to other farms and to the economic resources needed to support them, and what is the signiicance of the two habitation sections? Timber – a declining resource The large Legaard farmhouses III and IX were built in an area devoid of large forests (Andersen 1995a, 1995b; Stika vol. II, chap. 31). Despite this, nothing was spared in terms of the stout timbers use in their construction (see also Draiby’s reconstruction drawing and calculation of timber consumption this chap., p. 531 f). This included the use of bole walls and a large roof span; the width of the house was c. 8 m, compared with the width of the normal farmhouse of 6-7 m. When we add to this the fact that peat was apparently used for fuel in the cooking pits in house III, then it becomes abundantly clear that timber was a scarce resource. This tells us two things: Timber resources were controlled and not available to all in equal measure and the architecture of a grand farm building must have been governed by strong social and economic conventions. It therefore makes sense that most farmhouses employed wattle-and-daub wall constructions, which saved timber. Another indication of the scarcity of large forest trees/oak trees is the use of stone cists for burials, rather than the oak cofins that were common in southern Jutland where primeval forest still survived (Aaby 1986; Bech & Rasmussen vol. I, chap. 2). If we compare the timber consumption in the large Legaard farms, even though there was a decline in timber dimensions from house III to house IX, as also seen in the Bjerre houses from the same period, the difference is remarkable. Some farms obviously had access to and controlled supplies of good timber, and LEGAARD 523 as timber was a scarce resource, this meant that others were not able to obtain good building timber, as seen at Bjerre, which probably represents the normal situation in Thy. The origin of wealth: Grazing land and cattle In all literary sources relating to the late prehistoric societies of Celtic and Germanic peoples, cattle constitute the single most valued resource. The wealth of high-ranking people was measured in the number of cattle they owned or controlled, together with other factors such as the number of clients, i.e. the people who contributed in one way or another to their household. The most famous Irish saga, the Táin, is about a battle over a prized and renowned bull (Kinsella 1969), and cattle raids igure prominently in the old Indo-European literature, from Homer to the Irish sagas (Lincoln 1981). However, the need for grazing land for large herds of cattle and sheep also led to a drastic reduction in forests. Initially, during the earlier 3rd millennium, woodland was deliberately burnt to create pastures, but from the Early Bronze Age onwards the pollen record reveals a new and inal decline of the remaining forests, corresponding to the period of occupation of the two large Legaard farms, although perhaps beginning in the previous period (period I) or even earlier, in the Late Neolithic. The construction of many new farms throughout Thy at this time, combined with the importance of cattle and the need to accommodate the most valuable animals in the larger farms, goes a long way to explaining the decimation of Thy’s last forests. Pollen data from 12 Bronze Age barrows (Andersen 1999) demonstrates that the areas around the barrows were heavily grazed. In one case, where there were two successive phases of barrow construction, the pollen record demonstrates how overgrazing around the primary barrow resulted in greatly elevated levels of plantain (Plantago lanceolata) pollen, reaching an astonishing level of 40%. Consequently, grazing pressure was high during the Early Bronze Age in Thy and this, in turn, prevented forest regeneration. In addition, the construction of barrows using turves depleted the good grazing land around these monuments, as each barrow required on average turf from 2-3 ha of grazing land. This situation was unsustainable in the long term and by the end of period III barrow construction had come to a halt, farmhouses were downsized and only wattle-and-daub was used for house wall construction. There is even some suggestion of some emigration from Thy during period III to Jæren in southwest Norway as possibly indicated by the occurrence of ornaments of typical Thy type in this area (Randsborg 1972, map 20). The Legaard settlement is a prime example of this development at one of the largest farmsteads in the region. 524 The social and economic signiicance of interior space The single most important factor in creating interior monumentality is the width of a house, rather than its length, although some relation exists between them. The addition of extra width to the Legaard farms III and IX created an imposing inner space, as shown by B. Draiby’s reconstruction drawing (this chap., p. 531 f). Anyone who has experienced a house of this width can testify to the feeling of huge interior space and monumentality it conveys. Width also made higher demands on the roof construction and on the sturdiness of the supporting posts. We are therefore inclined to believe that limits existed as to who was able to build farmhouses of extra width and with bole walls, both of which demanded access to high-quality timber. This seems to have been a prerogative of families of high standing and it certainly demanded the control of timber resources and skilled labour. It is probably no coincidence that these farms were built to house cattle. Consequently, the introduction of the three-aisled farmhouses around 1500 BC was most probably a response to a social need to house cattle, and farmhouses with a width of c. 7-8 m were evidently better suited to accommodating two rows of stall dividers and also had more space for storage of winter fodder above the byre. There were therefore both social and economic considerations associated with the construction of such large farmhouses. However, there was also the possibility of subdividing the living sections, as is suggested at least for the western section. There were, accordingly, functional divisions within the residential sections of the farm, suggesting that different activities were performed within each household. Single and/or double households The plan of Legaard house III has been presented previously (e.g. Mikkelsen & Kristiansen 1996, 168; Kristiansen 1998, 286, 1999, 545; Earle et al. 1998, 20; Rasmussen 1999, 282; Bech & Mikkelsen 1999, 75; Bech 2003, 54; Jensen 2002, 121; Earle 2004, 121; Mikkelsen 2012, 56, 2013, 62).15 Legaard houses III and IX were soon classiied as chiely houses and chiely halls (Earle et al. 1998, 19-20),16 based on their size. K. Kristiansen later related Legaard house III to twin rulers, a ritual chief and a warrior chief: “The typical chiely farm hall of this period is 30-40 m long and 8 m wide, and appears together with the new institution of warrior aristocracies and twin rulers. In these big, three-aisled halls, we ind a bipartite architecture: there are two identical living quarters, each with a hearth. In the house … [Legaard house III] we ind the central hall occupied by the most costly, prestige good – stables for the cattle. The house has two entrances, one on each side of the house, close to the living quarters. So this is a farm hall MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN for two families with their cattle. Even then, they had nearly 100 m2 in the living quarters, which were further subdivided, a truly chiely compartment.” (Kristiansen 1999, 545). Later, J.-H. Bech seemed to be less certain when he characterised houses III and IX as buildings that “... belong to a group of houses of seemingly “chieftain” status built at the same time over most of Jutland” (Bech 2003, 53), but in the same article he also writes: “The big Legaard halls no doubt had chieftain status …” (Bech 2003, 57). Most recently, a new interpretation has been presented in which it is suggested that house III was the residence of a high/highest status family, with a room for the owner’s family, a room used as a byre and a room in the eastern end for a “… family of slaves or non-free workers …” (Mikkelsen 2013, 62). The fact that farmhouses of Legaard type, with two habitation units and stalls for cattle in the central part, have now been demonstrated elsewhere than in Thy, for example at Kongehøj II near Askov (Poulsen & Brønd 2008; Bech & Rasmussen vol. I, chap. 2, ig. 2.15), and that a number of other longhouses with two habitation units but no trace of a central byre are also known, indicates that we are dealing with an institutionalised construction. The question then is what was the nature of this social institution? To answer this question, we must irst consider the nature of the residents in the two habitation sections: Obvious possibilities include two separate households, a generational division or a functional division within a single large household. The generational hypothesis seems the least probable: An older generation would not be likely to reside separately and maintain a household in a large living area of nearly 100 m2, on a par with a younger generation. In most societies the older generation is downsized and is taken care of by the younger generation. They would therefore live rather within the household of the younger generation or be given a smaller separate house. In K. Kristiansen’s interpretation: The fact that we are dealing with two equally-sized habitation sections, each large enough to accommodate a full household, and that each had a separate entrance but on opposite sides of the house, suggests that these were coequal households (ig. 29.16). The opposing entrances suggest that they each owned a row of cattle on the side with the entrance. These observations make a functional division impossible and we therefore conclude that the Legaard farm was a production unit that, for some reason, required the cooperation of two households. Why was this? In periods of expansion there was a need for labour, for the construction of both houses and barrows, and this need could not be entirely met by each individual household. Moreover, a double household would surely be in a better position to cope in periods of crisis or conlict, when supplementary labour from other farms and kin was not available. We can also think in terms of the preservation of wealth and power: A double household can exchange women for marriage, if they are not too closely related, and thereby retain the wealth of the families within the households. This raises the question of the relationship between the two households. Double leadership is attested throughout the Bronze Age and Iron Age societies of the Indo-European speaking world, from Mycenaean and Spartan kingships to Germanic tribes (Parpola 2005). Such dual power functions normally resided in the social structure, where they tended to balance power. The Indo-European kinship system also practised relocation of young boys as foster sons, typically in the household of the mother’s brother (uncle). If we assume that double household farms were based on the principle of foster brothers and their families, we would have an explanation that ensured good relations with other kin groups (the mother’s group), and they could thereby play a cen- Entrance N Door Kristiansen: Mikkelsen: Religious chief Owner, high status Entrance Door Stable Stable Warrior chief Unfree workers 0 5 m Figure 29.16. Legaard. Interpretations of house III. Legend for ire features: See vol. I, appendix A. LEGAARD 525 tral role in the alliances that formed the foundation for the metal trade (Rowlands 1980; Kristiansen & Larsson 2005, 226ff). Taken together, all the resources required to construct and maintain large farmhouses such as houses III and IX at Legaard over longer periods of time indicate that they belonged to a powerful minority of leading families which controlled vital resources of timber, cattle, labour and good grazing land. This conclusion is further supported by the fact that most other houses were smaller and narrower and lack stalls to accommodate cattle. Small residential houses, such as house I at Sønderhå 5 (see below) (Earle this chap., p. 534 f), could have provided labour for the farm during periods II-III. However, the fact that the Legaard farms were built for two households also suggests that this was a form of selfsupport, creating a more stable basis for maintenance of the farm, should support from other households fail (for other examples of double or possible double households see Bech & Olsen vol. II, chap. 4). Twin or double households were an institution in Bronze Age society the signiicance of which we still do not completely understand. They also correspond to a dominating twin symbolism in Bronze Age cosmology. Twin households would therefore appear to have occupied an important social and ritual position in Bronze Age society. In M. Mikkelsen’s interpretation: M. Mikkelsen has recently suggested alternative interpretation (Mikkelsen 2012, 2013). In his view, Legaard houses III and IX constituted the residences of a family with high or even the highest status in the Bronze Age society during the lifetime of these two longhouses in period II and part of period III. He suggests that the owner’s family resided in the western end, where wooden benches and/or other constructions marked the owner’s status for those who were permitted access to this part of the house (ig. 29.16). Although the eastern room was of the same size as the western room, it was used to accommodate a household of unfree labourers, perhaps comprising one or more families. In this room there were no wooden benches. Furthermore, the difference in status implied in this interpretation was also expressed in the location of the entrances. The entrance for the owner and his family was on the bright, sunny southern side of the house, while that for the unfree labourers was on the dark northern side of the house. If both ends of the building were occupied by households of equal status, it would be expected that they both at least had entrances on the same side of the house. It is therefore suggested that houses III and IX were not inhabited by two equal households but by two very unequal households. 526 In I.K. Kristensen’s interpretation: Based on her analysis of the distribution of household pottery and, to some extent, hammerstones and quernstone fragments in Bjerre 6 house IB, which showed that these objects were most frequent in the eastern part, I.K. Kristensen proposes that the latter was used mainly for storage and initial preparation of food (Kristensen vol. II, chap. 18). She also suggests that a similar situation may have existed at Legaard, although the western part here was more heavily damaged by ploughing and the two parts of the house are therefore not on equal footing in terms of preservation (Kristensen vol. II, chap. 30). Kristensen’s observations can, however, be taken to support both Mikkelsen and Kristiansen in their interpretations. A division of some activities between the two habitation areas demonstrates their close daily interaction. However, as these indings are based on a very limited number of cases, we must await further contextual information on the function of dual habitations units in Bronze Age houses/farms. Chiely lineages and chiely farms – an ongoing debate In the preceding section we have demonstrated that many factors and a high degree of social and economic complexity is concealed behind the apparently straightforward term: ‘chief’, or ‘chiely farm’. We propose instead to talk about chiely lineages as constituting all free farmers. This deinition takes its point of departure in the link demonstrated between farms and barrows (Mikkelsen vol. II, chap. 28), and the fact that it has been calculated that probably only around 20 -25% of the living population could have been buried in barrows on their death (Holst et al. 2013). Barrows constituted the ritual and ‘heroic’ legitimisation of the lineage and soon created an ancestor genealogy that living members of the lineage could proit from (Kristiansen 2006). However, although everyone belonged to the same social category of free farmers, here termed chiely lineages, they were neither equally powerful nor wealthy. Substantial differences existed in farm size as well as in burial wealth, although smaller dwelling houses may have belonged to the less privileged group of commoners. How this variation arose and was eventually maintained, in the case of Legaard for at least three to four generations, is too complex a matter to deal with here. We will instead focus on Legaard’s relative position within the larger Early Bronze Age farms. The question is how common were the large farms of the Legaard type or even larger ones? What relative level of social and economic ranking were these associated with?17 MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN If we consider the greater area of northwest Jutland (ig. 29.17), 12 sites have longhouses of about the same size as Legaard, but at least ive houses at ive different sites (Virksund, Rosgårde, Kås Hovedgård and Lundbro) are substantially larger (ig. 29.18).18 The large house at Virksund has been shown by radiocarbon dating to be contemporaneous with Legaard house III (Bech & Olsen 2013). It should be mentioned that in two of the aforementioned very large houses (Rosgårde house VI and Lundbro house I) items were found with relations to bronze casting (Kristensen 2015) and they thereby display two of the most central elements used to deine high-status settlements. At the site of Kongehøj II, near Askov, a radiocarbondated farm identical to and of the same age as Legaard house III was found together with an even larger farm, also of the same date (Poulsen 2008 and pers. comm.). This raises the question of the role of the tripartite farmhouse in relation to even larger farms. Three out of the four known very large longhouses were found within or close to an area around Skive in Salling with an unusually large number of period III burials (Randsborg 1974, 198). But if these very large longhouses are more common in this relatively rich area, then the same should also be expected in Thy – the other area with an unusually large number of period III burials (Randsborg 1974; Kristiansen 1978). N Bjerre 6 Vestermark Øster Ørbæk Fårtoft V Legaard Stenildvad Ginnerup Båndruplund Virksund Kås Hovedgård Lundbro Hostrup Strand Glattrup V Skivevej 3 Rosgaarde Solbakken 0 Stensgaard 10 20 km Figure 29.17. Sites with large (light green) and extraordinarily large (dark green) houses in northwest Jutland. LEGAARD 527 Site no. Museum no. Site name House no. Size Area in m2 110112-279 THY 3414 Legaard House III c. 33.5 x 7.5-8 m c. 260 m2 ------------------House IX c. 33.5 x 7.6 m c. 255 m2 Period EBA II-III EBA II-III 110211-32 THY 2728 Bjerre 6 House IB c. 25.5 x 8.5-7 m (2nd phase) c. 205 m2 EBA II 110310-74 THY 5074 Vestermark House XII c. 31 x 7 m (1st phase) c. 220 m2 -----------------c. 36.5 x 7-6 m (2nd phase) c. 250 m2 EBA II 110605-128 THY 5056 Ginnerup House I c. 27.5 x 7.5 m c. 205 m2 EBA I 120703-22 VMÅ 2466 Øster Ørbæk House 1 c. 35 x 8.5 m c. 295 m2 EBA II 120814-339 VMÅ 2610 Stenildvad House 2 c. 36.6 x 8 m c. 290 m2 EBA II 130107-306 SMS 951A Skivevej 3 House X c. 36 x 7 m c. 250 m2 EBA II? 130110-110 SMS 742A Koustrupgårde House II c. 31 x 7-6.5 m c. 210 m2 -----------------House IV c. 30 x 7.4-6.6 m c. 210 m2 130110-111 SMS 731A Rosgårde House VI-VII c. 32 x 8 m (1st phase) c. 255 m2 c. 46 x 8 m (2nd phase) c. 360 m2 EBA II-III EBA II-III? EBA II-III 130117-243 SMS 933A Virksund I House K I c. 38 x 7.5 m (1st phase) c. 285 m2 c. 53 x 7.5 m (2nd phase) c. 400 m2 130201-94 SMS 960A Glattrup VI House K IV c. 33.5 x 7.75 m c. 260 m2 EBA II? 130410-150 SMS 785A Lundbro II House I c. 36 x 9 m c. 325 m2 EBA II? 130910-241 VSM G637 Båndruplund House II c. 30.6 x 7.5 m c. 225 m2 EBA II-III? 131004-119 SMS 974A Hostrup Strand House IV c. 32.5 x 7.6 m c. 245 m2 EBA II? 131005-162 SMS 788A Kås Hovedgård II House I Bevaret 28.5 x 10.5 m Over 300 m2 180511-155 HOL 20.434 Solbakken House I c. 6.7 x 33.5 m c. 225 m2 EBA II? 180804-240 HOL 20.350 Stensgård House N12 c. 46 x 8.75 m c. 400 m2 EBA II? EBA II-III? Figure 29.18. Reported sizes of large (200-300 m2) and extraordinarily large (> 300 m2) houses from the Early Bronze Age in northwest Jutland. 528 When analysing differences in burial wealth, we ind a similar hierarchical organisation as seen in settlements in Thy and on Mors during periods II-III. Out of a total number of 143 analysed burials, K.M. Hornstrup (1998) classiied six as belonging to an elite segment. If we apply these igures to settlements, then we should expect approximately one in 24 settlement sites to yield exceptionally large houses.19 If this comparison has relevance, we should not be surprised that very large houses, bigger than those at Legaard, have not yet been found in Thy. It is more surprising that so many very large houses have been found in Salling. Returning to the local area around Legaard (ig. 29.15), the number of burial inds from the Early Bronze Age located at Bjergene, about 1 km east of Legaard, attracts attention. Swords have been found in at least three or four of these barrows (Aner & Kersten 2001, Ke 4989, 5010, 5012 and 5013), and a burial from Bronze Age period I may have contained a sword, a dagger and a spear. This relatively rich burial cannot be related to the settlement area at Legaard, where no longhouses from this period have yet been found, but it indicates that a high-status settlement was located in the vicinity of Bjergene (and Legaard?) in period I. More interesting is a burial from period II, which contained both a sword and a gold ring. This rich burial could be contemporaneous with Legaard houses III and/or IX. It is therefore possible that at least one of the males (the head of the household) from Legaard was buried at Bjergene. We still lack the very largest longhouses, like those seen in southern Jutland, but for the time being the Legaard farm is the largest farm unit both in the local area and in Thy as a whole. It therefore seems reasonable to link some of the rich weapon burials from Bjergene to this site. Conclusion The Legaard site provides the best evidence so far for the existence in the Early Bronze Age periods II-III of large farms accommodating two households, with a byre for housing cattle in their central part. Recent excavations have testiied to the existence of similar farmhouses in southern Jutland (Poulsen 2008; Bech & Rasmussen vol. I, chap. 2). Large farms such as these required access to substantial quantities of high-quality building timber. This was, however, a scarce resource in Thy at the time, as demonstrated by smaller farms, most notably at Bjerre, from the same period. This situation testiies to the existence of a hierarchy in farms and households and a social organisation in which twin households played a central role. The nature of these two households is still debated, and we will have to await new excavations, providing good contextual evidence, before the matter can be resolved. However, the presen- MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN tation of the Legaard farms, with two households and housing of cattle, has taken our understanding of the social and economic organisation of the Bronze Age a large step forward and will focus attention on the remaining unresolved problems in future excavations. Notes 1. Excavations at the Legaard site began in 1995 with a team of students from University of Gothenburg led by K. Kristiansen. J.-H. Bech from Museum Thy also participated. House I was excavated and trial trenches revealed the existence of several further houses and some very large postholes that later turned out to represent house III. As the site was complex and also under threat from ploughing, Museum Thy applied for and received funding for further rescue excavation. M. Mikkelsen was called upon to direct the excavations, in collaboration with K. Kristiansen and TAP, because of his extensive experience in excavating Bronze Age houses. 2. The Legaard site has site no. 110112-279 and reports and inds are archived at Museum Thy under the case no. THY 3414. 3. The two barrows have site nos. 110112-211 and -212. The urns were found in no. 211. 4. The plough-zone sampling was undertaken in 1995 under the direction of J. Steinberg. It was carried out in conjunction with the trial excavation and excavation of house I. 5. Finds resulting from the ploughzone sampling are at present archived in US. They have not been reviewed in connection with this publication. As with other sites in the TAP, it is possible to make a comparison between results from the survey and from the ploughzone sampling. This has not been done systematically for the Legaard site but a few comments can be made. Comparing inds from the survey with those from the ploughzone sampling, the actual numbers for scrapers are 15-20:6, borers are 8-11:3, cores are 3:8 and pottery is 3:34. Although a number of factors should be taken into consideration, it is notable that lint tools from the survey outnumber those from the ploughzone sampling, while the opposite is the case with lint cores and, especially, the pottery. 6. The features from the Late Iron Age include two pithouses and two ditches in the southwestern part of the excavated area, minor parts of one or two presumed longhouses, a possible pithouse in the northwestern trial trench and a relatively large pit that has disturbed part of the large house III (ig. 29.4). The presence of these features raises questions about the dating of some of the other features in the excavated area. 7. It is dificult to establish with absolute certainty whether the posts were removed or not. House IX has one set of postholes for roof-bearing posts that are much sturdier than the rest and have the same dimensions as those in house III. Assuming that house IX is later, this post pair could derive from house III. 8. It does not seem possible to explain this difference in distribution as being due to the preservation of the westernmost part of house III being relatively poor. Most of the western half of the house was as well preserved as the eastern half, and in the eastern half pottery was found in the lower part of postholes below pit N202. 9. When the house was excavated, it was presumed, on the basis of the house type with relatively large roof- and wall postholes, that house III belonged to the earliest ‘generation’ of three-aisled longhouses in Thy and it was therefore suggested in the excavation report that it should be dated to Bronze Age period II. 10. The eastern part was uncovered in 1996, but although all features in the area were sectioned, we unfortunately did not identify house IX before the central and western parts of the house were uncovered in 1997. 11. The result may indicate that there was activity in the area in the Late Neolithic or early in Early Bronze Age period I. 12. The spacing of the roof postholes in each set is relatively narrow compared to the other threeaisled longhouses on the site and also relative to most other Early Bronze Age longhouses. However, Bjerre 2 house III from period II also has a relatively narrow cross span and therefore this feature cannot, in itself, be used in dating longhouses as has been suggested by for example Artursson (2005). 13. On basis of the shape of the wall postholes in house II, it was discussed, both during and after the excavation, whether house II should be dated to the Late Iron Age. But given the massive concentration of cooking pits of typical Bronze Age type, and the similarities between the arrangement LEGAARD 529 of the presumed roof postholes in both house II and house IV, it seems reasonable to suggest that the two houses are from the Bronze Age. 14. Even though they were marked on the ordnance map, Engelhardt did not record these barrows in 1875, when he toured Sønderhå parish. According to information from 1912, there were still substantial remains of site no. 110112-212, so this barrow at least must have been easy to see in 1875. The suspicion is that Engelhardt unfortunately did not visit the Legaard area and he obviously did not obtain any other information on the barrows at the site. 15. Of these, Bech, Earle and Kristiansen all participated in the excavation at the Legaard site and all therefore have detailed knowledge of the results from the excavation. The interpretation of house III (and later house IX) as a chiely house has been discussed several times in TAP. 16. The term ‘hall’ (in Danish: hal) was irst used by Becker (1972), when he referred to a 27 x 8 m longhouse at Spjald and a 33 x 8m longhouse at Bjerg in western Jutland. If it is only size that distinguishes a 530 hall from other smaller longhouses, then Legaard houses III and IX might be called halls according to Becker’s deinition. But if the term is not only deined by the overall size of the building, but also or instead by the size of an internal ‘room’, then the Legaard houses, as such, or a room inside the Legaard houses, cannot be termed as halls, given the byre evident in the largest central part of these buildings. 17. The most comprehensive analyses of three-aisled longhouses from the Early and Middle Bronze Age, prior to Bech & Olsen (vol. I, chap. 4), were presented by Rasmussen (1999), Artursson (2005, 2009) and Bech & Olsen (2013). 18. Unpublished material relating to igure 29.18 has been kindly placed at our disposal by I.K. Kristensen, Skive Museum, B.H. Nielsen, Vesthimmerlands Museum, N. Terkildsen, Holstebro Museum and J.-H. Bech, Museum Thy. 19. Obviously a number of source-related problems are relevant when material from graves is compared with settlement sites. MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN Legaard house III Reconstruction by Bente Draiby This substantial, regular house site corresponds to a typical large farm from the middle of the Early Bronze Age. The 33.5 m long and 8 m wide building with semicircular, rounded gables and a loor area of 260 m2 was not one of the very biggest of Bronze Age farms. Nevertheless, its construction, with large timbers, broad wall posts and sturdy bole walls, is impressive given that it was built in a time of timber shortage (ig. 29.I.A1). Just as striking is its architecture both in plan and in section (ig. 29.I.A2). The long building is divided up into three rooms: Two solid partition walls with door posts in the middle separate the two gable rooms W E A1 A2 Ground plan; Longitudinal section, W-E N S Plan of roof construction; Cross-section through byre, N-S A3 0 5 10 m Figure 29.I.A1-3. Reconstruction of Legaard house III by B. Draiby. As part of the reconstruction, a hearth has been drawn in on the ground plan in both the western and eastern ends of the house. No evidence for the existence of these was demonstrated in the excavation, but they are a suggested reconstruction based on how hearths are normally located relative to cooking pits (see also Bech & Olsen vol. I, chap. 4). LEGAARD 531 Figure 29.I.B. The interior of the room at the western end of the house. The chieftain receives guests. Drawing: B. Draiby. used for habitation etc. from the large central byre which, with its area of 120 m2, occupies almost half of the building's total area. The house’s two entrances are located here; one in the southwest corner and the other in the northeast. Immediately to the left of each of these entrances, which have the character of gates, was possibly also a narrower entrance. The mirror symmetry of the entire structure is remarkable and is emphasised by the long, straight rows of stalls along the north and south walls being displaced by the position of the entrances, such that the byre is the irst one sees on entering the building. Together with the timberwork, the byre has represented wealth and status of this large farm. The building is situated on a west-facing hillside and, as the wall posts in the gables can be presumed to have been dug to the same depth, the loor of the west end lay rather more than 0.6 m lower than that in the east. The result of this is that the gable room to the west is the highest room in the building (ig. 29.I.B). The fact that this position- 70-80 years About 30 years 50-60 years TS a b a-a a b-b SK TS VS A b BB HR BP About 20 years 1 BP VS A TS 2 GR TB TS B VS DS BP SP BP a b a-a b-b a BP 3 b 4 Diam. at root: 70 cm 50 cm 30 cm 20 cm Logs (trees): 12 8 25 50 50 4 Figure 29.I.C. Diagram showing timber use. Logs and postpipes: A: Upper ends. B: Root ends. Postpipes in plan and section from: 1: Cloven upper end. 2: Complete upper end. 3: Root end with axe marks from felling. 4: Felled at root, possibly from wind-thrown tree. Logs used: TS: Roof post. VS: Wall post. DS: Door post. SK: Partition wall post. HR: Head. GR: End- or gable plate. TB: Cross beam. BB: Tie beam. SP: Rafters. BP: Bole planks. 532 BENTE DRAIBY ing of the building was entirely intentional is shown by the subsequent house IX, immediately south of house III, being situated according to the exactly the same pattern. Apart from at the wider entrances, the wall posts forming the 2.5 m wide side aisles are placed between 1.5 and 1.7 m apart. Into these wall posts, which were about 0.4 m wide and made of cloven timber with the silver grain (i.e. the radially-cloven surface) facing outwards, were inserted bole planks. The wall posts of the long sides stand in pairs, at right angles to the longitudinal axis of the building, where wall posts and rafters were directly connected in a locked joint. The wall posts around the gables, which form precise semicircles, are positioned according to a completely ixed pattern, determined by the roof construction (ig. 29.I.A3). The best solution is with three rafters over the end- or gable plate with the middle one slightly displaced relative to the longitudinal axis and the two others rebated into the end plate, forming an even arc. This pattern can be found in other Bronze Age houses with dispersed wall posts. A thatched roof of reeds has been selected here, but the roof could just as well have been of turf, which the solid timber construction could easily have borne. There are several plausible reasons for this choice, for example prioritisation of grassland areas for hay-making and loft room for winter fodder under a higher roof. There were reeds in abundance around the nearby lakes and bogs, but the same cannot be said of grassland, which should also be used for the barrow construction. All the posts in house III are made of whole or cloven logs. None of the cross-sections of the postpipes reveal further cleavage into quarter-section timber, and there are only a very few plank-shaped cross-sections that probably represent half-section timber from which the rounded side has been cut away. The bole planks, as wall planks, would necessarily have been cloven timber, while the other horizontal timber, tie beams, sills and plates could have been made of whole or half logs. Figure 29.I.C shows how many trees of various ages had to be felled to supply the timber for this large longhouse, in all about 150 oak trees. It is abundantly clear that the bole walls were the most resource-demanding part of the construction, both with respect to the number of trees, cubic metres of timber and labour required. LEGAARD 533 Sønderhå 5 – THY 2788 by Timothy Earle Located about 750 m SSE of Legaard on a small hill that rises to 26 m a.m.s.l., THY 2788 is apparently a small Early Bronze Age house site (ig. 29.2). Because the original inds assemblage from the site contained asymmetrical lint sickles (ig. 29.II.A) and no Dagger period diagnostics, the site was tentatively assigned an Early Bronze Age date. Subsequently, a lithic concentration was deined and preserved postholes and cooking pits were documented. In summer 1993, a full excavation took place, the goal of which was to document the preserved structural features at the site and to investigate associated economic activities. Because the culture layer had been ploughed out, virtually all the artefacts were recovered from the plough soil. Fieldwork at THY 2788 involved shovel testing, ploughzone sampling and normal procedures for feature identiication and excavation (Steinberg 1996). Across the top of the hill, shovel tests roughly deined the lithic concentration. Fifteen 2 x 2 m ploughzone samples (three per 50 x 50 m unit) were laid out in a standard off-set grid across ive blocks surrounding the area of the knoll (Steinberg 1996, 378, ig. 7). Four exploratory trenches were then cut by machine, and groups of features were exposed, including post- 0 5 cm Figure 29.II.A. One of three asymmetrical lint sickles found on the surface of the site by amateur archaeologist P. Brandt. 534 TIMOTHY EARLE holes deining the small house on the little knoll. An intensive pattern of ploughzone samples was then laid out around where the preserved features had been identiied. In all, 37 units were excavated and screened from the top of the knoll. The goal was to recover a representative sample of artefacts from the plough soil surrounding the house (ig. 29.II.B). An area of roughly 30 x 25 m was then cleared of plough soil and exposed features were mapped and excavated (ig. 29.II.C) . The preserved western gable end, two sets of somewhat irregular placed roof-support posts and a central cooking pit deine reasonably well a threeaisled Bronze Age house. The house was apparently small, perhaps only 9.3 m long and 5.5 m wide. It was oriented east-west with a markedly rounded western end and less angled than is typical of many Bronze Age houses in northwest Jutland (Bertelsen et al. 1996; Bech & Olsen vol. I, chap. 4). The irregularity in the pattern of preserved structural features probably relects poor construction materials due to the scarcity of local wood, the lower status of the occupants and/or speciic uses of the structure. Although much more regular, its closest parallel is a small house from Early Bronze Age period III excavated at Ørum in southern Thy (Bech & Olsen 2013, vol. I, chap. 4, ig. 4.6B). This measured 9 x 6 m and had three pairs of roof-support posts. Due to associated ire features, this building is interpreted as a residential house, as also appears to have been the case with the small house at THY 2788. A radiocarbon sample of cereal grain (Hordeum vulgare) from cooking pit N38 in the house at THY 2788 was dated to 3010 ± 50 BP (LuS 6108, vol I, appendix B), calibrated at 1 σ to 1377-1132 BC, i.e. Early Bronze Age periods II-III. The house could therefore have been contemporary with the large Legaard farms and perhaps functioned as an auxiliary settlement. Outside the house were various scattered features, mainly cooking pits. Charcoal from ire feature N25, close to a cooking pit, gave a radiocarbon date in the Early Neolithic: 4765 ± 60 BP (LuS 6106, vol. I, appendix B), calibrated at 1 σ to between 3639 and 3386 BC. However, this feature is most likely from the Bronze Age, with peat containing fossil wood having been used for fuel in the same way as seen in house III at Legaard. From a later period, cremation pit N17 was dated to 2170 ± 50 BP (LuS 6107, vol. I, appendix B), calibrated at 1 σ to 357-166 BC, i.e. Early Iron Age. Perhaps this burial is related to a plougheddown barrow listed in the parish inventory of 1912 from the same knoll, but of which no trace remains today. In a number of cases, cremation burials from the Early Iron Age have been found located outside burial mounds (Hornstrup et al. 2005). Figure 29.II.B. Distribution of lint lakes at the site based on 37 ploughzone samples. The area of lithic production lies close to the house site, but since the lakes were produced by hard hammer percussion, a typical Late Bronze Age technology, the relationship with the Early Bronze Age house is open to discussion. Some retouched frost lakes and a few pressure-worked fragments of sickles were also found in the plough soil. Diagram: J. Steinberg. 300 300 200 200 100 100 63039300 63039200 4674600 4674500 63039100 Area of lithic production 4674400 63039000 4674300 Number of flakes in a plow zone sample 10 m Cooking pit Posthole Figure 29.II.C. Distribution of features at the site. Legend for ire features: See vol. I, appendix A. 0 4 8 m N38 N17 N25 N SØNDERHÅ 535 References Aaby, B. 1986. Mennesket og naturen på Abkæregnen gennem 6000 år – resultater af et forskningsprojekt. Sønderjysk Månedsskrift 1986/9, pp. 277-290. Andersen, S.T. 1995a. History of Vegetation and Agriculture at Hassing Huse Mose, Thy, Northwest Denmark, since the Ice Age. Journal of Danish Archaeology 11 (1992-93), pp. 57-79. Andersen, S.T. 1995b. Pollenanalyser fra Ove Sø. Geobotaniske Undersøgelser af Kulturlandskabets Historie. DGU Kunderapport nr. 12, pp. 36-55. Becker, C.J. 1972. Hal og hus i yngre bronzealder. Nationalmuseets Arbejdsmark 1972, pp. 5-16. Bertelsen, J.B., M. Christensen, M. Mikkelsen, P. Mikkelsen, J. Nielsen & J. Simonsen 1996. Bronzealderens bopladser i Midt- og Nordvestjylland. Skive: Skive Museum. Earle, T. 2004. Culture Matters in the Neolithic Transition and Emergence of Hierarchy in Thy, Denmark: Distinguished Lecture. American Anthropologists 106/1, pp. 111-125. Andersen, S.T. 1999. Pollen Analyses from Early Bronze Age Barrows in Thy. Journal of Danish Archaeology 13 (1996-97), pp. 7-17. Earle, T., J.-H. Bech, K. Kristiansen, P. Aperlo, K. Kelertas & J. Steinberg 1998. The Political Economy of Late Neolithic and Early Bronze Age Society: The Thy Archaeological Project. Norwegian Archaeological Review 31/1, pp. 1-28. Aner, E., K. Kersten, E. Koch & K.-H. Willroth 2001. Die Funde der älteren Bronzezeit des nordischen Kreises in Dänemark, Schleswig-Holstein und Niedersachsen. Band 11. Thisted Amt. Neumünster: Wachholtz. Holst, M.K., M. Rasmussen, J.-H. Bech & K. Kristiansen 2013. Bronze Age ‘Herostrats’: Ritual, Political, and Domestic Economies in Early Bronze Age Denmark. Proceedings of the Prehistoric Society 79, pp. 265-296. Artursson, M. 2005. Gårds- och bebyggelsesstruktur. In: P. Lagerås & B. Strömberg (eds.), Bronsåldersbygd 2300-500 f. Kr., pp. 84-159. Skånska spår. Arkeologi längs Västkustbanan 6. Lund: Riksantikvariämbetet. Hornstrup, K.M. 1998. Overgangen fra ældre til yngre Bronzealder. Et eksempel fra Nordvestjylland, Danmark. In: T. Løken (ed.), Bronsealder i Norden – Regioner og interaksjon, pp. 23-33. AmS-Varia 33. Stavanger: Arkeologisk museum i Stavanger. Artursson, M. 2009. Bebyggelse och samhällsstruktur. Södra och mellersta Skandinavien under senneolitikum och bronsålder 2300-500 f. Kr. Riksantikvarämbetet & Göteborg Universitet. Gotarc Series B, No. 52. Göteborg: Göteborg University. Stockholm: Riksantikvarieämbetet. Hornstrup, K.M., K.G. Overgaard, S. Andersen, P. Bennike, P.H. Mikkelsen & C. Malmros 2005. Hellegård – en gravplads fra omkring 500 f.Kr. Aarbøger for Nordisk Oldkyndighed og Historie (2002), pp. 83-162. Jensen, J. 2002. Danmarks Oldtid. Bronzealder 2000-500 f.Kr. København: Gyldendal. Bech, J.-H. 2003. The Thy Archaeological Project – Results and Relections from a Multinational Archaeological Project. In: H. Thrane (ed.), Diachronic Settlement Studies in the Metal Ages, pp. 45-60. Jutland Archaeological Society publications 45. Højbjerg: Jutland Archaeological Society. Kinsella, T. 1969. The Tain from the Irish Epic Tain Bó Cuailnge. Translated by Thomas Kinsella. Oxford: Oxford University Press. Bech, J.-H. & M. Mikkelsen 1999. Landscapes, settlement and subsistence in Bronze Age Thy, NW Denmark. In: C. Fabech & J. Ringtved (eds.), Settlement and Landscape. Proceedings of a conference in Århus, Denmark May 4-7 1998, pp. 69-77. Højbjerg: Jutland Archaeological Society. Kristensen, I.K. 2015. Spor efter bronzehåndværk i Salling og Fjends. In: S. Boddum, M. Mikkelsen & N. Terkildsen (eds.), Bronzestøbning i yngre bronzealders lokale kulturlandskab, pp. 107-119. Yngre bronzealders kulturlandskab 5. Viborg: Viborg Museum. Holstebro: Holstebro Museum. Bech, J.-H. & A.-L.H. Olsen 2013. Early Bronze Age houses from Thy, Northwest Denmark. In: K.-H. Willroth (ed.), Siedlungen der älteren Bronzezeit, pp. 9-32. Studien zur nordeuropäischen Bronzezeit 1. Neumünster: Wachholtz. Kristiansen, K. 1978. The Consumption of Wealth in Bronze Age Denmark. A Study in the Dynamics of Economic Processes in Tribal Societies. In: K. Kristiansen & C. Paludan-Müller (eds.), New Directions in Scandinavian Archaeology, pp. 158-190. 536 MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN Studies in Scandinavian prehistory and early history 1. Copenhagen: The National Museum of Denmark. Olsen, A.-L.H. 1991. Recording of artefacts from private collections in Sønderhå parish 1991. Museum Thy, unpublished report. Kristiansen, K. 1998. The Construction of a Bronze Age Landscape, Cosmology, Economy and Social Organisation in Thy, Northwestern Jutland. In: B. Hänsel (ed.), Mensch und Umwelt in der Bronzezeit Europas, pp. 281-291. Kiel: Oetker-Voges. Olsen, A.-L.H. & J.-H. Bech 1996. Damsgård. En overpløjet høj fra ældre bronzealder per. III med stenkiste og ligbrændingsgrube. Kuml (1993-94), pp. 155-198. Kristiansen, K. 1999. Symbolic structures and social institutions. The Twin rulers in Bronze Age Europe. In: A. Gustafsson & H. Karlsson (eds.), Glyfer och arkeologiska rum – en vänbok till Jarl Nordbladh, pp. 537-552. GOTARC series A, 3. Göteborg: University of Göteborg. Kristiansen, K. 2006. Cosmology, economy and longterm change in the Bronze Age of Northern Europe. In: K.-G. Sjögren (ed.), Ecology and Economy in Stone Age and Bronze Age Scania, pp. 171-193. Lund: National Heritage Board. Kristiansen, K. & T.B. Larsson 2005. The Rise of Bronze Age Society. Travels, transmissions and transformations. Cambridge: Cambridge University Press. Lienemann, J. 1999. P-mapping in THY 3414. ABOLA, Arbeitsgruppe für Bodenkunde Landschaftsökologie und angewandte Botanik GmbH, Oldenburg, Deutschland, unpublished report. Lincoln, B. 1981. Priests, Warriors, and Cattle. A Study in the Ecology of Religion. Hermeneutics, studies in the history of religions 10. Berkeley: University of California Press. Mikkelsen, M. 2012. “Dobbeltgårde” i yngre bronzealder. In: S. Boddum, M. Mikkelsen & N. Terkildsen (eds.), Bebyggelsen i yngre bronzealders lokale kulturlandskab, pp. 41-65. Yngre bronzealders kulturlandskab 2. Viborg: Viborg Museum. Holstebro: Holstebro Museum. Mikkelsen, M. 2013. The topographical placing of the Late Neolithic and Bronze Age settlements and an introduction to a new interpretation of the layout of the individual farms in the Bronze Age. In: K.-H. Willroth (ed.), Siedlungen der älteren Bronzezeit, pp. 33-66. Studien zur nordeuropäischen Bronzezeit 1. Neumünster: Wachholtz. Mikkelsen, M. & K. Kristiansen 1996. Legård. Arkæologiske udgravninger i Danmark 1996 (209). Det Arkæologiske Nævn. København. Olsen, J., K.M. Hornstrup, J. Heinemeier, P. Bennike & H. Thrane 2011. Chronology of the Danish Bronze Age based on 14C dating of cremated bone remains. Radiocarbon 53/2, pp. 261-275. Poulsen, M.E. 2008. Bygherrerapport for HBV j.nr. 1275 Kongehøj Etape II. Arkæologiske undersøgelser af bebyggelse fra bondestenalder og bronzealder. Sønderskov: Museet på Sønderskov. Poulsen, M.E. & J. Brønd 2008. Bondestenalder og bronzealder på Kongehøj ved Askov. Fund af hustomter og gravanlæg fra 2500-500 f.Kr. Museet på Sønderskov – Museumsavisen 28/47, pp. 17-27. Parpola, A. 2005. The Nasatyas, the Chariot and ProtoAryan Religion. Journal of Indological Studies, Nos. 16 & 17 (2004-2005), pp. 1-63. Prieto-Martinez, M.P. 2008. Bell Beaker Communities in Thy: The First Bronze Age Society in Denmark. Norwegian Archaeological Review 41/2, pp. 115-158. Randsborg, K. 1972. From Period III to Period IV. Chronological studies of the Bronze Age in Southern Scandinavia and Northern Germany. Nationalmuseets skrifter, Arkæologisk-historisk række 15. Copenhagen: The National Museum of Denmark. Randsborg, K. 1974. Befolkning og social variation i ældre bronzealders Danmark. Kuml (1973-1974), pp. 197-208. Randsborg, K. 2006. Opening the Oak-cofins. New Dates – New Perspectives. In: K. Randsborg & K. Christensen, Bronze Age Oak-cofin Graves. Archaeology & Dendro-dating, pp. 3-162. Acta Archaeologica Vol. 77, Supplementa Vol. VII. Rasmussen, M. 1993. Bopladskeramik i Ældre Bronzealder. Jysk Arkæologisk Selskabs Skrifter 29. Højbjerg: Jysk Arkæologisk Selskab. Rasmussen, M. 1999. Livestock without bones. The long-house as contributor to the interpretation of LEGAARD 537 livestock management in the Southern Scandinavian Early Bronze Age. In: C. Fabech & J. Ringtved (eds.), Settlement and Landscape. Proceedings of a conference in Århus, Denmark May 4-7 1998, pp. 281-290. Højbjerg: Jutland Archaeological Society. Rowlands, M. 1980. Kinship, alliance and exchange in the European Bronze Age. In: J. Barrett & R. Bradley (eds.), Settlement and Society in the British Later Bronze Age, pp. 15-55. BAR British Series 83. Oxford: British Archaeological Reports. 538 Steinberg, J. 1996. Ploughzone sampling in Denmark: Isolating and interpreting site signatures from disturbed contexts. Antiquity 70, pp. 368-390. Steinberg, J.M. 1997. The Economic Prehistory of Thy, Denmark. A Study of the Changing Value of Flint Based on a Methodology of the Plowzone. University of California, Los Angeles, unpublished PhD thesis. Westphal, J. 2000. Thy Archaeological Project. Report of Field Survey 1992-1997. Museum Thy, unpublished report. MARTIN MIKKELSEN & KRISTIAN KRISTIANSEN