Leben auf dem Tell als soziale Praxis EURASIEN-ABTEILUNG, BERLIN RÖMISCH-GERMANISCHE KOMMISSION, FRANKFURT A. M. des Deutschen Archäologischen Instituts Kolloquien zur Vor- und Frühgeschichte Band 14 Dr. Rudolf Habelt GmbH ! Bonn 2010 EURASIEN-ABTEILUNG DES DEUTSCHEN ARCHÄOLOGISCHEN INSTITUTS Leben auf dem Tell als soziale Praxis Beiträge des Internationalen Symposiums in Berlin vom 26.–27. Februar 2007 herausgegeben von Svend Hansen Dr. Rudolf Habelt GmbH ! Bonn 2010 213 Seiten mit 169 Abbildungen und 12 Tabellen Titelbild Măgura Gorgana bei Pietrele, Rumänien Redaktion: Kirsten Hellström, Emily Schalk Gestaltung des Umschlages: Anke Reuter Bibliografische Informationen der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.d-nb.de abrufbar © 2010 by Eurasien-Abteilung des Deutschen Archäologischen Instituts Berlin Satz, Druck und Bindung: Druckhaus „Thomas Müntzer“, Bad Langensalza ISBN 978-3-7749-3710-9 Florin Draşovean and Wolfram Schier The Neolithic tell sites of Parţa and Uivar (Romanian Banat). A comparison of their architectural sequence and organization of social space Fig. 1. Map of late Neolithic tell sites in Romanian Banat. Tells are anthropogenic creations whose genesis, development and features are determined by certain pedo-climatic and relief conditions, by certain relations of production and social relations, by the character of the architecture and by the type of the construction materials.1 They are the material expression of an orderly use of a well-defined geographic space. From a morphological point of view tells are the result of a cumulative process, in which habitation traces accumulate vertically in a relatively consistent and regular way within a well defined area. In the greater Carpathian area tells appear starting with the Middle Neolithic period and have their broadest distribution during the late Neolithic, when many of the large settlements take on the features of a tell. In the Romanian Banat Neolithic tell-type habitations have been identified at Parţa I and II2, Bucovăţ3, Chişoda Veche4, Sânandrei5, Foeni6, Uivar7 and Liubcova8 (Fig. 1). Chron- ologically they have been traced back to the latter half of the 6th calibrated millennium and to the first half of the fifth millennium BC. 1 S HERRATT 1983; C HAPMAN 1994; CHAPMAN 1997; H ALSTEAD 1999; KOTSAKIS 1999; BAILEY et al. 2002; G OGÂLTAN 2003. 2 LAZAROVICI 1972; L AZAROVICI 1979, 204–205; D RAŞOVEAN 1994, 412; DRAŞOVEAN 1996, 32–33. LAZAROVICI et al. 2001; L AZAROVICI / LAZAROVICI 2003. 3 LAZAROVICI 1979, 143; 188; LAZAROVICI 1991a; LAZAROVICI / LAZAROVICI 2006, 375–387. 4 RADU 1979; DRAŞOVEAN 1991; D RAŞOVEAN 1994, 409– 410; D RAŞOVEAN 1996, 30; 39. 5 D RAŞOVEAN 1994, 413; DRAŞOVEAN 1996, 33 6 D RAŞOVEAN 1994a, 141–149; D RAŞOVEAN 1997, 55–63; D RAŞOVEAN 1999, 6–11. 7 S CHIER / DRAŞOVEAN 2004. 8 C OMŞA 1969; L UCA 1997. 166 Florin Draşovean and Wolfram Schier Geographic setting The two most important Banat tells, Parţa and Uivar, are the focus of this paper. These lie in the Banat, in the low southern plain of Timiş, on the banks of the present-day Timiş and Bega rivers, at an elevation of 85 m (Parţa) and 78 m (Uivar). In this region the plain had been cut by earlier meanders of the two rivers, which the regularisation works carried out later in the 18th century corrected; they gave rise to bogs and marshes that covered a great deal of the interfluve between the Timiş and Bega rivers. The Parţa tell, located about 18 km southwest of the Uivar tell, was situated on the bank of a major branch of the Timiş river, as shown by the 19th century maps drawn before the complete regularisation of the river (Fig. 2). Prior to canalisation, the site was surrounded by a series of meandering branches to the west and south and by marshy land to the east. Site excavations and tests carried out on the sediments of cliffs along the Timis river have revealed that originally the settlement was located upon a higher river bank of one of the temporary meanders south of the habitation. This river arm continued to flow and was sufficiently deep throughout the development of the tell, so that the digging of a ditch in that area was not necessary.9 Moreover, the anthropogenic strata of the settlement extend into the area of the former arm of the Timiş river. From the sedimentological point of view, the first level of habitation lies upon a brown clay layer which, in turn, covers a thick layer of yellowish sandy sediments. The Uivar tell, as could be shown by (still ongoing) sedimentological studies, was probably surrounded on the East, North and West by a former meandering branch of the river, this time the Bega river (Fig. 3). South of the tell a low terrace separated the tell from a further fossil river channel in the southeast. As in the case of the Parţa tell, the anthropogenic deposits cover a brown fossil alluvial soil and a thick alluvial sandy layer rich in mica, which stem from the periodical flooding of the Bega river.10 From the pedological viewpoint, both tells lie in areas where the common soil types are vertisol, chernozem and degraded chernozem.11 The relatively close location within the same geographic area, emphasised by identical relief and pedo-climatic conditions, suggests a homogeneous environmental background within which the Parţa and Uivar communities developed. Thus, the differences between the sites must depend upon other factors rather than their environment. Fig. 2. The tell site of Parţa on the banks of the river Timiş, shown on a historical map of the 19th century. Fig. 3. The tell Uivar “Gomila” on a historical map of the 19th century. It gives some impression of the alluvial landscape mosaic which survived the canalization of the river Bega a few decades earlier. The genesis and development of the Parţa and Uivar tells The tell of Parţa The Parţa dwellings were first observed in level 7a (Fig. 4), which from an architectural point of view is characterised by pits that have been interpreted as hovels and surface dwellings. The houses in this level are small and usually grouped in pairs. This layout presumably was determined by the organisation of the family and by certain property rela- 9 LAZAROVICI et al. 2001, 197. K ADEREIT et al. 2006, 20 fig. 9,28. 11 T ĂRĂU / LUCA 2002, 213. 10 The Neolithic tell sites of Parţa and Uivar (Romanian Banat) tions. At the core of the settlement, an isolated house has been uncovered, which, due to its central position and the fact that it was built above the sanctuaries 1 and 2, has been regarded as a possible building for worship.12 In this first stage the Parţa settlement was not fortified.13 Starting with level 7b, we can speak of a settlement planning, in which the dwellings were grouped around two structures of similar size, erected in the middle of the settlement (Fig. 5). The houses around these two structures lie at a certain distance, forming a small square. With only two exceptions, houses P17a and P43a, all of the houses are aligned along an east-west axis;14 the orientation is dictated, we believe, by the strong gusts of the summer storms blowing from the west. Research carried out thus far has shown that the two structures in the small square played a special role in the community life of Parţa. One of the structures, 12.6 long and 7 m wide, has been interpreted as a sanctuary (Fig. 6), inside of which altars for offerings, monumental statues and columns once stood. On the northern wall was an altar table, A, oriented towards the south. On a clay socle stood an idol bust. In the western part, guarded by two posts, stood altar table B. Inside the altar several flint blades were discovered, which may have been used for sacrifices. Altar C was discovered in the eastern part, and at the centre of the sanctuary stood a clay column. To the south of sanctuary 1 lay a structure of similar size that might have served a social function. It was designated “the house of the tribe” by the excavator.15 A fortification system belonging to this level of habitation (7b) was made up of four parallel ditches, lined on their inner edges by four palisades, which protected the settlement from the west, north and east sides.16 In the north, the ditches are so close to one another17 that they may represent different stages in the development of the fortification system. To the south, as noted earlier, the settlement was protected naturally by an arm of the river Timiş. Consequently, there was no need for a ditch, so only a stouter palisade following the river bank18 was constructed. To the north, between palisades 3 and 4, deep postholes were uncovered, suggesting some wooden constructions, possibly watch towers.19 Inside the fortification system a zone extending about 15–20 m from the palisade was devoid of houses.20 The defence system was constantly maintained, since traces of removing silt and maintenance works have been found along ditch 3, which caused an alteration of the slope. The fortification system remained in use by the time of phase 6b and was subsequently abandoned following the sudden appearance of the Vinča C culture.21 167 In the next level, 7c, new houses were built next to existing ones, sustaining the architectural organisation of the previous level22 in a well-defined settlement plan, which continued for several generations (Fig. 7). The architecture of this level, and of level 6, is characterised by massiveness and, we might say, by the professional touch with which the constructions were erected. For their construction massive logs (with diameters exceeding 30 cm) were used for the loadbearing structures, and the walls were of wattle structure covered by a layer of clay. The massive posts for some structures had a technical rationale, as they had to bear the additional weight of a second story. As a rule the houses were grouped in fours,23 which suggests a specific social organisation structure. At the centre of the settlement the two presumed community buildings continued to be in use: the sanctuary and the “house of the tribe”, which stood in the small central square. Around this square, the houses have been erected on the same location as the ones on level 7a. They are oriented along an east-west axis and built very close together; in some cases they even share a common wall. Very narrow spaces between the houses form alleys running east to the west and intersecting others, thus forming a “network structure”. The obviously cramped living space may be the result of demographic growth and the continuance of the fortification system, which did not allow a horizontal spread of the settlement. Two-storied constructions are first seen in this level, of which the most well-designed examples are houses P136B and P40–43 (Fig. 8). Sanctuary 2 was built on the same location as sanctuary 1 and is the most significant example of a building for ritual purposes in 12 L AZAROVICI / LAZAROVICI 2003, 385; LAZAROVICI / LAZAROVICI 2006, 219ff. 13 L AZAROVICI et al. 2001; LAZAROVICI / L AZAROVICI 2003, 385. 14 L AZAROVICI et al. 2001; LAZAROVICI / L AZAROVICI 2006, 228–229. 15 L AZAROVICI et al. 2001, 243; LAZAROVICI / LAZAROVICI 2003, 459 fig. 53; LAZAROVICI / LAZAROVICI 2006, 228. 16 L AZAROVICI et al. 2001, 197–202; LAZAROVICI / LAZAROVICI 2006, 228. 17 L AZAROVICI et al. 2001, 197 fig. 161. 18 L AZAROVICI et al. 2001, 197; LAZAROVICI / LAZAROVICI 2006, 366–367. 19 L AZAROVICI et al. 2001, 245; 202–203; LAZAROVICI / LAZAROVICI 2006, 366. 20 L AZAROVICI / LAZAROVICI 2003, 197; LAZAROVICI / LAZAROVICI 2006, 366. 21 L AZAROVICI et al. 2001, 201; LAZAROVICI / LAZAROVICI 2006, 369–370. 22 L AZAROVICI / LAZAROVICI 2006, 232 ff. 23 L AZAROVICI / LAZAROVICI 2006, 232. Fig. 5. Florin Draşovean and Wolfram Schier Fig. 4. 168 The Neolithic tell sites of Parţa and Uivar (Romanian Banat) 169 Fig. 6. Parţa. Reconstruction of “sanctuary 1” (LAZAROVICI et al. 2001, fig. 166b). the entire Neolithic of South-East Europe.24 This place of worship functioned until phase IIC of the Banat culture, when level 6 of the settlement was completely and systematically destroyed under circumstances that might be linked to the sudden appearance of Vinča C culture in the area. After the destruction of Tell 1 a new settlement, Tell 2, was established a few hundred yards to the west, on the opposite river bank.25 Despite the violent aspects, probably suggesting immigrant groups rather than local conflicts, shortly after the event, an acculturation or integration process might have occurred, as evidenced by the Banat culture IIC imports from the deep levels of Tell 226 and the Vinča C material from level 5 of the settlement of Tell 1.27 After the destruction of habitation level 6, the architecture becomes more rural. The houses decrease in size, although the general orientation of the houses is maintained (Fig. 9). At this time the fortification system was abandoned, and the habitation expanded westwards beyond the ditches. This level represents the end of the tell-type dwelling of Parţa 1. It is covered by level 4, which belongs to the Eneolithic Tiszapolgár culture,28 after a hiatus during the Vinča C culture and the cultural group of Foeni. From a chronological point of view, the radiocarbon dates show that the Parţa tell developed over three centuries, between 5300 and 4950 calibrated BC.29 ◄ Fig. 4. Parţa. Architecture of the lowermost habitation layer 7a (L AZAROVICI et al. 2001, fig. 58). ◄ Fig. 5. Parţa. Architecture of habitation layer 7b (LAZAROVICI et al. 2001, fig. 77). The tell of Uivar The late Neolithic-early Eneolithic tell site ”Gomila” near Uivar (jud. Timiş, Romania) is situated on the flat alluvial plain of the rivers Bega and Timiş, both tributaries of the river Tisza, about 35 km west-southwest of the city of Timişoara. Scientific exploration at the tell started in 1998 with a comprehensive survey and has been continued in annual excavation campaigns from 1999 to 2009.30 Until now 18 trenches with a total surface of 1600 m2 have been excavated, comprising more than 2500 m3 of cultural layers with a mostly high find-density. In five larger trenches house struc- 24 L AZAROVICI 1988; LAZAROVICI et al. 1991; LAZAROVICI / LAZAROVICI 2003; L AZAROVICI / LAZAROVICI 2006, 241–250. 25 L AZAROVICI et al. 2001, 201; LAZAROVICI / LAZAROVICI 2006, 370. 26 Unpublished materials deposited in the Banat Museum from Timişoara. 27 D RAŞOVEAN 1996, 32 pl. CIII; LAZAROVICI / LAZAROVICI 2006, 370. 28 L AZAROVICI et al. 2001, 181 ff. 29 M ANTU 2000, 79 Tab. 1; LAZAROVICI et al. 2005. 30 The project is organised as cooperation between the Muzeul Banatului Timişoara and the Institut für Prähistorische Archäologie der Freien Universität Berlin. It is funded by the German Research Foundation (until 2009) and the Ministeriul Culturii şi cultelor, Bucarest. For results of the first four excavation campaigns as well as various other aspects of the project, see S CHIER / D RAŞOVEAN 2004; S CHIER / D RAŞOVEAN 2005; S CHIER 2005; K ADEREIT et al. 2006; S CHIER 2006; DRAŞOVEAN 2007. 170 Florin Draşovean and Wolfram Schier ◄ Fig. 7. Parţa. Architecture of habitation layer 7c (LAZAROVICI et al. 2001, fig. 82) tures in varying states of preservation have been excavated, but only two trenches (I and XI) allowed us to record building sequences over several phases. The most complete sequence could be documented in trench I, where 3.8 m of cultural and building layers have been recorded on top of buried soil. Hence, this trench offers the opportunity for a first tentative reconstruction of the development of domestic architecture. Architectural sequence The stratigraphic sequence, which will be discussed in greater detail elsewhere, starts with a basal burnt layer 4. Here in 2008 a large house of 6.5 m by more than 10 m was discovered; it has not been excavated completely yet. This house consists of three rooms, two of which had an upper floor with paving of loam supported by massive wooden substructions. The house suffered an intensive disastrous fire; its ruins contained remains of dozens of pots and other domestic equipment (Fig. 10). Several radiocarbon dates obtained from this – at present – oldest building in the tell of Uivar lie consistently between 5250 and 5060 cal BC (1 σ range). On top of this 0.6–0.8 m thick burnt layer, which can probably be divided into two subphases, starts a sequence of almost 2 m of continuous building phases, uninterrupted by any traces of destructive fires. The lowermost of these unburnt layers, layer 3.8, showed over a total surface of 120 m2 deep oval foundation pits that were oriented in parallel lines and presumably belonged to four houses (Fig. 11). None lay wholly within the excavation trench; thus, only their width of 4.5–5 m could be recorded. Farther up, in the unburnt layers 3.7–3.5, a sequence of intensive building activities could be observed. Four or five houses with parallel orientation were rebuilt two times, leading to a dense sequence of overlapping foundation ditches. The general pattern, however, is quite different from layer 3.8: Within the long and shallow continuous ditches many small to medium size postholes could be observed; thus, the wattle-and-daub-wall had started already below the ancient floor level (Fig. 12). In the layers 3.4–3.3 for the first time unburnt house structures were recognised. They showed continuous shallow foundation ditches like those in the layers below. A great surprise, however, was the survival of wooden remains. The foundation ditches of two houses were dug along a frame of The Neolithic tell sites of Parţa and Uivar (Romanian Banat) 171 Fig. 8. Parţa. Artist’s reconstruction of houses P136B and P40–43 (L AZAROVICI et al. 2001, back cover). Fig. 9. Parţa. Architecture of habitation layer 5 (D RAŞOVEAN 2007, 28). split planks, laid out and connected by chiselled joints on a greyish ashy subsoil (Fig. 13). On top of the unburnt layers with wooden remains were several floors, but no clear structural remnants could be observed (layers 3.2–3.1). All in all the sequence of unburnt houses continues for more than two meters, in which the houses keep their general orientation and position with only slight deviations. At the upper end of this se- quence the last house burned down, preserving in its massive daub residues many negatives of the wooden construction. This house (trench I, feature 373) with three rooms and an upper floor in the westernmost room was virtually reconstructed, based on the recorded evidence (Fig. 14). While we have precise evidence for the development of domestic architecture, our insight into the diachronic structure of the settlement on a larger 172 Florin Draşovean and Wolfram Schier Fortification system Fig. 10. Uivar, trench I. Burnt house H4b-1, penetrated by later foundation pits (photograph: Institut für Prähistorische Archäologie, FU Berlin 2008). Fig. 11. Uivar, trench I. Foundations of construction layer 3.8. The large foundation pits were dug deep into the burnt layer 4.1 (photograph: Institut für Prähistorische Archäologie, FU Berlin 2007). scale is limited by the relatively small percentage of the whole site excavated as yet. Due to these limitations, the settlement layout and the correlation of the fortification system with the architectural sequence of the tell in its core area will remain a goal for future research. The great surprise in 2000, when the first geophysical survey was conducted on and around the tell of Uivar by H. BECKER, was the existence of a complex, multi-phase surrounding ditch system, enclosing not only the visible tell, but also an area as far as 90 m from the tell’s margin. The surveyed area, however, was not large enough to include the outer ditch system completely, whose existence and size was unexpected. Only in 2007 was it possible to extend the magnetic survey to the southwest and to fill in some gaps to the east of the tell (Fig. 15). Eventually our hypothesis31 could be confirmed: The outermost ditch in the northwest is clearly connected to the outermost and rather straight ditch in the southeast of the tell. For the first time we now can determine the dimensions of this enclosed tell site: It measures 400 m at a right angle of 300 m from the southwest to the northeast. The total enclosed surface is almost 9 hectares, of which the tell itself occupies only 3 hectares. The fortification system at Uivar is one of the most elaborate and complex in this part of Europe. Although built in several stages, it is the result of a uniform defence conception, made up of a doubleditch ring that protected the central area and additional concentric ditches laid out beyond the central ring. The excavated materials from the ditches and the intersection of some ditches with other site features partly allow the development stages of the defence system of Uivar to be partly determined. Thus, excavations in trench IV, which was intended to study the entrance gate to the inner precincts, revealed that in a previous phase there was only one rather shallow ditch that protected the core area. According to radiocarbon dates, the oldest phase of this ditch can be dated to the transition from the 6th to the 5th millennium cal. BC. Only in the third phase was the innermost ditch with its impressive depth of 4 m and width of almost 7 m accomplished as double ditch fortification by adding the concentric outer ditch, lined on its inner edge by a wall of horizontal wooden planks (Fig. 16a, b). This fortification, according to the radiocarbon dates, was in function for 200 years, continuing into the phase of Vinča C2. Research in trench IV has also provided evidence of the chronological relationship between the settlement and the fortification. Here the post holes from some surface dwellings are cut by the two ditches of the third-phase ring, and some storage pits were cut by the outer ditch that defended the core area. During a later stage, during the first 31 S CHIER / D RAŞOVEAN 2004, 150–154. The Neolithic tell sites of Parţa and Uivar (Romanian Banat) 173 Fig. 12. Uivar, trench I. Shallow foundation ditches of construction layer 3.4 (photograph: Institut für Prähistorische Archäologie, FU Berlin 2005). Fig. 13. Uivar, trench I. Wooden substruction below house floor, layer 3.5 (photograph: Institut für Prähistorische Archäologie, FU Berlin 2005). 174 Florin Draşovean and Wolfram Schier Fig. 14. Uivar, trench I. Reconstruction of burnt house 373, layer 2.2 (3D reconstruction: S. Suhrbier, Institut für Prähistorische Archäologie, FU Berlin 2003). two inner ring development phases, it was filled with a succession of ash and charcoal layers, upon which houses were built, demonstrating the need for more habitation space. Unfortunately, the relatively limited excavation area, compared with the vastness of the tell, does not allow us to connect the three successive stages of the fortification from the south-western part of the settlement to the other ditches in the system. Despite this limitation, and with the necessary measure of caution, the materials in their excavated fill suggest that the ditches constituting the last phase of the Vinča fortification system were concomitantly dug following a well-thought-out plan. Thus, at about 140–165 m distance from the tell’s centre was a ring ditch which, as was shown by a test in the south-eastern part of the tell, varied in depth and width between 2–2.5 m and 4–6 m respectively (showing larger dimensions in trench VI than in trench X). It was lined inside by a palisade or by sturdy wooden watch-towers, whose details had been noted in trench VI that was intended to study this ditch along its north-west side. At about 10 m inside, the moat was doubled by another one that had a crowning length of 3.3 m and a depth of 3.5 m and did not have a concentric palisade. The same defence strategy is found on the inner defence ring. Here, too, behind a ditch that, due to its relatively modest size and depth might have given potential raiders the impression of vulnerability, there was an impressive ditch behind the enclosure that could stem off any invasion – if we assume a contemporaneity of both ditches. Magnetic prospection has revealed the presence of several branches deriving from the main ditches (Fig. 15) and that represent different construction stages.32 Unfortunately, at this point, it is hard to determine to which development stages these belong. It seems very likely that several branches belonging to the Vinča system were dug during the third phase of the defence system (as it was differentiated in trench IV). As was previously noted, the fortification system operated for almost 200 years. From field observations of defence ditches in trench IV, it has been observed that, after only three years, an almost 1 m thick layer of mud was washed off the 32 G ERLING et al. 2005. The Neolithic tell sites of Parţa and Uivar (Romanian Banat) 175 Fig. 15. Uivar. Magnetometric plan and excavation trenches (H. Becker, Beuerberg, and S. Suhrbier, Institut für Prähistorische Archäologie, FU Berlin 2000–2010). slopes by rainwater and silted up the ditch due to natural weathering. This phenomenon must also have occurred in the past and would have forced the Neolithic community to carry out regular maintenance and removal of silt. These works have been observed archeologically in the trench IX excavation, where they caused an alteration of the slope and gradient angle of the ditch. In addition to protecting a certain area of the site, the fortification ditches were possibly used to delimit several areas for specific functions within the social space. While the core area was the realm of habitation, social and spiritual life, the inner ring space and the outer ring space may have served different purposes. Spatial organisation inside the inner ditch Returning to the core area of the tell, if we eliminate the small anomalies, it is apparent that most houses were laid out in two concentric circles surrounding a possible square at the heart of the tell (Fig. 17). At the same time, most houses are perpendicular to the inner fortification ring that belongs to the final, third stage of the defence system. All of this information supports the idea of a well-conceived settlement plan that was followed by the Neolithic inhabitants of Uivar. This plan underwent alterations over time due to the necessity to enlarge the habitation area beyond the core of the tell, under the permanent struggle for space be- 176 Florin Draşovean and Wolfram Schier Fig. 16. Uivar, trench IV. Reconstruction of innermost double ditch. Approach from the northwest and view from inside (3D reconstruction: S. Suhrbier, 2003). tween the area confined by the fortification and a continuously growing population. This conflict was resolved very efficiently by means of the already mentioned southward enlargement of the area allotted for dwelling. Among the most important magnetic anomalies in the core area is one found in the southern part of the core area (Fig. 15, trench XI). Research carried out here has revealed a structure, which, although damaged by later Eneolithic and mediaeval complexes, has provided enough clues to suggest an architecture and furnishings for ritual purposes. With its long axis running north-east to southwest, the structure has imposing dimensions (10 m × 5 m.) and was made up of a ground floor and an upper floor. Research has shown that the upper level flooring did not entirely cover the lower floor area, thus leaving the third part from the south uncovered. The ground floor was divided into three rooms (Fig. 18). To the south was one room of 5 × 4 m in size, which had a rectangular hearth with rounded corners and raised edges in its north-western part (Fig. 19). The hearth lay next to the dividing wall that separated this room from the central room. On the west and east corners of the room the remnants of two large pots were discovered. At the centre of the building lay a slightly trapezoidal room of 3–4.2 × 5 m, in which most of the recovered ceramic items in the building, including several whole earthen pots, were found. In each of the four corners of the room was a small fire hearth with slightly raised edges. One of these hearths, in the south-eastern corner, bore a median bulging rib which split into two parts. Towards the northern wall lay an oval-shaped grinding stone. The last room of the building, the northern room, was divided into four compartments by low walls (Fig. 20), which did not rise above one metre. Inside the compartments, in the north-east and north-west corners of the building, stood two additional large-sized earthen pots. At the centre of the northernmost room, a well-burnt bovidae horn and the very poorly preserved fragments of a clayshaped ox head (bukranion) were found lying upon the floor (Fig. 21). The largest of the four compartments, the second from the northern corner, contained a complete turtle shell and a burnt stone adze (Fig. 22). In view of the four fireplaces in the central room, the four large-sized earthen pots in the four corners and the ox head in the north room, this building obviously had a purpose beyond a mundane use, thus taking on a ritual role, presumably that of a sanctuary. Settlement evidence outside of the tell Since the first geophysical survey in 2000 various hypotheses have been discussed to understand the large enclosed, but apparently uninhabited outer ring zone. The ring space might have been used as The Neolithic tell sites of Parţa and Uivar (Romanian Banat) a safe haven and protected area for herds from wild animals and from raids by other communities. We also assumed that the potters’ kilns stood there, which, due to the threat of fire that they posed, were not permitted in the crowded space of the core area. In other areas, the rarity of magnetic 177 anomalies might suggest that part of the land was cultivated. In the southern part of the ring zone, adjoining the foot of the visible mound, a rather dense concentration of anomalies can be observed (Fig. 15). It has been interpreted as later extension of the settle- Fig. 17. Uivar. Enlargement of the magnetometric plan (Fig. 15), showing the arrangement of burnt houses and empty space in the core area of the tell. 178 Florin Draşovean and Wolfram Schier Fig. 18. Uivar, trench XI. Burnt house H2b-11 with special (ritual?) features (isometric reconstruction by P. Kunz, Institut für Prähistorische Archäologie, FU Berlin 2007). Fig. 19. Uivar, trench XI. Burnt house H2b-11: large fire place with rectangular frame (photograph: F. Draşovean, 2005). Fig. 20. Uivar, trench XI. Burnt house H2b-11: northern room with four compartments (photograph: F. Draşovean, 2005). The Neolithic tell sites of Parţa and Uivar (Romanian Banat) Fig. 21. Uivar, trench XI. Burnt house H2b-11: fragments of clay bucranion (photograph: F. Draşovean, 2005). Fig. 22. Uivar, trench XI. Burnt house H2b-11: northeastern compartment with turtle shell and complete stone adze (photograph: F. Draşovean, 2005). ment,33 due to the overcrowding of the core area space. It seems likely that this area was selected because the land here rose a bit in elevation and was less exposed to floods. In the light of recent research, however, this southern extension of the core settlement is no longer attributed to the ProtoTiszapolgár phase; a later Vinča C2 (?) date seems much more probable. In 2005 a small test trench was opened in the flat alluvial plain 80 m northwest of the tell margin, where the magnetic prospection showed a cluster 179 of very strong anomalies (Fig. 15, trench XIV). The fragmentary and rather thin daub observed suggests the presence of a building with wooden walls and only a thin coating of loam. Among the few objects found in the burnt ruins were five polishing pebbles. We consider these pebbles to be potters’ tools, used to produce the characteristic lustrous appearance of Vinča fine ware and, consequently, we suggest that the building was used for pottery manufacture. In 2007 a larger trench (XV) was opened immediately northwest of the trial trench in order to excavate the remaining part of the large magnetic anomaly. In a depth of 0.8 m below the surface a complete and almost undisturbed large burnt house was discovered (Fig. 23), measuring 11.9 × 5.3 m. Unlike the neighbouring structure unearthed in 2005, the massive daub remains corresponded to the burnt house ruins in the tell. On three sides the position of the walls could be ascertained either by in situ wattle impressions or by straight stripes of daub limiting the floor remains. The burnt floor consisted of a thick basal layer of loam, which had been renewed once. When well preserved parts of the floor were carefully turned upside down and refitted, they showed impressions of very well split massive wooden planks with an average width of 28 cm. The crack pattern observable in the central part of the floor thus followed a very carefully worked wooden substructure of split wooden beams laid out side by side across the ground plan of the house (Fig. 24). In the house’s interior two dividing walls were visible, separating three rooms of almost equal size. A great amount of secondarily burnt pottery, mostly large storage vessels and other domestic forms, was found on the floor and between the collapsed wall remains. In the central room a row of conical clay weights indicated the in situ position of a loom. This house in trench XV is not only the best preserved and one of the largest burnt houses discovered so far in Uivar, but the single one that is founded upon a carefully carpentered wooden support, thus indicating a special function. Further insight into the house’s construction was gained after the removal of the floor and the recording of the post holes (Fig. 25). A rather regular array of short foundation ditches was observed, most of which contained two post holes. Along the southeast long wall ran an almost continuous foundation ditch. All together six rows of ditch segments / post holes were recorded. Since no earlier or later large building structures were found in 33 S CHIER / D RAŞOVEAN 2004, 160. 180 Florin Draşovean and Wolfram Schier down, but it matches with an original elevated position. Our suggested reconstruction (Fig. 26) shows an almost 12 m long rectangular house consisting of three rooms. It was built upon a massive wooden platform broader than the house. The doubled posts in the southeastern post row assumedly carried a protruding roof that protected the wooden platform. While this reconstruction is hypothetical, it takes into account all of the observed details. If it is correct, it would be the first example of an elevated house in the Vinča culture. But it is also one of the very few complete and well preserved houses excavated outside a tell stratigraphy. The lighter cottage-like building excavated in 2005 in neighbouring trench XIV, however, did not show any evidence of an elevated construction. If it is contemporary with the house in trench XV, it may have been a subsidiary building used for handicrafts or storage, but not for permanent living. Unlike in the densely settled tell, we may thus Fig. 23. Uivar, trench XV. Burnt house of the tell (photograph: Overhead Fotosystems, Timişoara 2007). trench XV, all of the observed posts must belong to the house – a very fortunate situation compared to the stratigraphic excavation in the tell, where many post holes cannot be attributed to a specific building phase with certainty. The preserved southeast long wall corresponds with the second row of posts, while the outermost post row inside the continuous foundation ditch was clearly outside the house. The ground plan of the house should consist of five parallel rows of posts, indicating a four-aisled rectangular building, whose central post row carried the ridge. The abundance of post rows might be explained by their reconstruction as wooden platform carrying the house. A seventh post row in a continuous foundation ditch may not have been observed due to the restricted size of the trench. Another strong argument for the reconstruction as platform is the massiveness and careful carpentry of the split planks carrying the loam floor. Also the crack pattern observed in the burnt floor supports this hypothesis: While the central part of the floor may have collapsed as compact block, thereby preserving its structure, in the southwestern and northeastern part fragments of the floor were moved apart several centimetres and deposited in slightly tilted orientations. The described appearance would be difficult to explain, if the construction lay upon solid even ground while it burned Fig. 24. Uivar, trench XV. Burnt house floor: negatives of split plank substruction (platform) (photograph: W. Schier, 2007). The Neolithic tell sites of Parţa and Uivar (Romanian Banat) Fig. 25. Uivar, trench XV. Plan of posts below the burnt house floor (plan: S. Suhrbier, 2007). 181 182 Florin Draşovean and Wolfram Schier Fig. 26. Uivar, trench XV. Tentative isometric reconstruction of house, constructed on an elevated platform (reconstruction: P. Kunz, 2007). have evidence of a farmstead consisting of several buildings. Core and periphery: hypotheses about tell formation Ever since the first evidence appeared of a large, flat but fortified surrounding zone around the tell of Uivar, the possible functional, economic and social meaning of this outer zone has been discussed.34 In the light of the new unambiguous evidence of a permanent settlement in this outer zone, the discussion is focussed on the question of how dense or dispersed the flat settlement around the tell might have been. Looking more closely at the magnetic survey, it is possible to locate more clusters of magnetic anomalies like the one excavated in trenches XIV and XV (Fig. 27). They are loosely spread over the western and southern area of the outer enclosed zone, and even outside the outermost ditch some burnt house-like structures (coordinates 250/420, 350/330) can be recognised. In contrast, the interior of the tell shows a spotted pattern of burnt houses, suggesting frequent burning in a dense agglomeration of buildings. However, the excavated sequence in trench I underlines the relative rareness of conflagration: Only about 10 % of the houses recorded had been destroyed by fire, while most houses at the end of their lifespan were just levelled and rebuilt on top. A gross volumetric calculation leads to similar results. The overall volume of the settlement mound can be estimated at 70,000 m3, plus an unknown amount of soil redeposited as colluvium around the tell. The volume of loam necessary to build one three-roomed house of 10 m by 4.5 m will certainly not exceed 20 m3, probably less. Since most of the tell’s volume consists of building material, a minimum of 3500 houses would be necessary to form such a mound. Taking into account the eroded soil redeposited in the colluvium, it is more likely that the total sum of houses accumu34 S CHIER / D RAŞOVEAN 2004, 159; 225. The Neolithic tell sites of Parţa and Uivar (Romanian Banat) 183 Fig. 27. Uivar. Magnetometric plan with indications of definite (trench XIV/XV) and suspected burnt houses or farmsteads in the flat periphery of the tell. lated over the lifespan of the settlement would be around 5,000. Even when counting carefully, no more than 200–250 burnt house structures are visible on the magnetic survey chart (Fig. 10). Again only 4–7 % of all houses seem to have been destroyed by fire, be it by accident or deliberately.35 If we apply the same frequency estimate for destructive fires to the outer flat settlement zone, it is realistic to assume 10–20 times more unburnt houses (farmsteads?) between the half dozen magnetic anomaly clusters. Unburnt houses, however, could not be detected by cesium magnetometry in Uivar until now. Under these preconditions the 6 hectares of tell’s environs, enclosed by the oval outer ditch and palisade, may well represent a large dispersed settlement area, where houses were destroyed by fire only occasionally. The late Neolithic settlement at Uivar, thus, appears as a complex spatially and / or chronologically differentiated phenomenon. From a bird’s-eye perspective it resembles a huge fried egg: An almost circular central mound is surrounded by an oval flat settlement area, both enclosed by an impressive ditch accompanied by a palisade along its inner edge. This appearance would have been the result of several centuries of continuous habitation, building and decay / destruction of houses requiring a continuous input of building material (earth and wood). The digging of defence ditches at Uivar led to the removal of more than 10,000 m3 of soil, of which more than half is yellow clay. In historic times, this surplus soil would have been used to build the rampart’s berm. Research carried out in the areas near the ditches has not confirmed such deposits, nor has it found yellow clay strata that might derive from the ditches. Hence, one of the sources of the building loam used for the houses might have been the ditches – but even now it is clear that the volume of the tell exceeds the volume of the ditches several times. Substantial loam pits must have existed either within the enclosed flat area (invisible to cesium magnetometry?) or outside the outermost ditch, as yet undetected. 35 The evidence for deliberate ritual burning of houses (cf. CHAPMAN 1999, TRINGHAM 2005) is still ambiguous in Uivar: Some daub fragments show a slag-like porous structure indicating very high temperatures that are unlikely to occur in an accidental fire without additional fuel. Most of the burnt houses contain very few remains of their contents, while two (in trench XI and XV) are full of different classes of pottery, including large storage vessels. So perhaps house fires occurred for different reasons. If a ritual of destroying a house by fire existed, this ritual was performed only occasionally. 184 Florin Draşovean and Wolfram Schier As opposed to loam, the consumption of wood occurred in some kind of dynamic equilibrium to its naturally regenerating supply – the forests around the settlement. As described above the changing construction techniques and wood-usage patterns indicate some kind of forest management in order to safeguard a sustained wood supply. As the settlement grew, more and more of the surrounding forest might have been turned into secondary forests or even into open bush and grassland, mainly in areas under additional ecological pressure such as grazing herds of livestock (predominantly cattle). Whether a decreasing wood supply might even have been a limiting factor for settlement growth is still an open question. The crucial question, however, is the relation between the central tell and the outer settlement zone. Concerning their temporal relationship three versions are possible: a) The outer zone may have been settled only during the later phases of the tell, thus representing an expansion of the settlement with decreasing density and / or time of the building activities. b) Both the tell and the outer settlement zone might have been coexistent over the total time span of the settlement. In this case a functional, economic or social differentiation of both settlement compounds should be assumed in order to explain their differences in density and stratification. c) The outermost ditch may have confined the extension of an early flat dispersed settlement. Denser spacing of houses, more frequent rebuilding and/or socially determined building site continuity might have led to the continuous accumulation of sediment in the centre, while the settlement activity in the outer zone gradually decreased and eventually ended, resulting in a contraction process of the settlement. At present, none of the three models can be verified. Since there is some evidence that the large house outside the core zone (trench XV) might date to an early phase of the settlement, the expansion hypothesis does not seem very likely. A decision between hypothesis b) and c), however, would require more extensive excavation in the outer zone and, thus, future research. Conclusions Drawing on the observations summarized above, the following conclusions can be offered: 1. The Parţa and Uivar tells span a 600-year development, between 5300 and 4700 calibrated BC,36 over the transition between the 6th and 5th millennia and contemporary for a time that can be estimated as a few decades. From a historical point of view, the Parţa tell belongs to the Middle Neolithic period and experienced its florescence at the end of this time span, in the second phase of the Banat culture, which is contemporary with phase B2 of the Vinca culture.37 Its development appears to have had a violent end, possibly caused by new population groups with a new material culture, Vinča C, the same people who through successive habitation waves gave rise to the Parţa II and Uivar tells. 2. Both tells have well-developed defence systems, whose building called for the concentration of significant manpower that could not be mobilised without the involvement of an authority able to speed up and coordinate the community social energy. In both cases, the defence systems also betoken a remarkable civil engineering design that presupposes the organisational commitment of the elites for such wide-scale public works. Similarly, one can find a long-range strategy for the use and maintenance of the system and a uniform outlook concerning the relationship of the settlement-defence system. For example, even when the settlement on the Parţa tell became overcrowded, the building of houses near the defence ditches was avoided, as these might have been easily set on fire by a potential aggressor, and the fire might have easily spread to all buildings in the site. As such, we find the existence of the concept of global safety for all inhabitants, of the common wellbeing, which is more important and which prevails over the ideas of expedience and individual benefit. At Uivar this view is present too, and, thus, constructions were not permitted outside the defence system. Despite this planning, when the core area became overcrowded, the construction of houses in the southern part of the tell, between the central ring and the outer ditches, was accepted. In connection with the possibility of the simultaneous existence of building phases of defence systems and of the houses in the settlement, we can assert that in the case of both tells these can be indirectly proven, since there was a continuous need for the extraction of loam or earth due to the need for fortification (see above). The situation is identical at Parţa and Chişoda Veche, and we believe that only one answer is acceptable: Namely, as the tell population dug out the defence ditches, they were also intensely engaged building houses. If this hypothesis is true, we shall have to rethink our assumptions about the organisation and social structure of Late Neolithic society. 36 S CHIER / D RAŞOVEAN 2004, 201–203; LAZAROVICI et al. 2005. 37 L AZAROVICI 1991, 34; LAZAROVICI et al. 2005. The Neolithic tell sites of Parţa and Uivar (Romanian Banat) 3. The structure of the living space was determined by family organisation, by property relations, and it evolved in close connection with the demographic development of society and the possibility of enlarging dwelling space. At Parţa, the dwellings in the first level are made up of maximally two rooms; they do not have a second storey and are grouped two by two, only to be grouped four by four at levels 7c-6. At these levels, where the possibility of enlargement was limited by the defence system, population growth triggered the construction of two-storey buildings that offered the necessary living space to the more numerous family members. At Uivar, with the example of building 373, we find a similar development under the restrictive circumstances of the defence ditches: the house growing from two rooms, which it had at the time of its construction, to three rooms and yet another floor level in its last functioning phase. Thereby, alterations were carried out in order to ensure living space for a greater number of family members. 4. The family space is much better outlined following the Parţa research. Here, the final development phase of buildings P. 40–43 underlines the fact that the ground floor of these constructions saw the development of productive activities such as weaving, bean grinding and meal preparation, whereas the upper floor space, as shown by the presence of different pots and small-sized ovens, was meant for sleeping and social activities. 5. Both communities had specialized buildings for ritual purposes, which, although they were not identical, have many features in common. The Parţa sanctuary, through its monumental statues, ox heads and the ritual-related grinding and spinning, may fall into the category of an agrarian-pastoral ritual, closely connected with nature’s yearly cycles. At Uivar, with the exception of the one ox head present, all of the other elements are not so clear at first sight. On closer inspection, we can find the presence of four large-sized earthen food vessels, one each lying in the four corners of the building, and the presence of four fireplaces in the four corners of the central room and of four compartments in the north-western room. This repetition of the number 4, we believe, is not a random coincidence. It may be connected to the four natural cycles and, in association with the symbol assigned to the ox head, it may also have been associated with fertility and fecundity. 6. The impressive size of an enclosed zone around the tell proper, as could be demonstrated in Uivar, raises many new questions about the temporal, economic and social relation between “core” and ”periphery” and about the mechanisms of tell formation. Generalising from the evidence in Uivar, it is quite likely that late Neolithic tells in the Carpathian basin are nothing but the visible 185 parts of much larger and more complex central settlements, whose structural and functional principles still remain to be explored. These are only some of the preliminary conclusions that might be drawn from the study of the Parţa and Uivar settlements. We are sure, however, that the ongoing research at the two tells, combined with work carried out at similar sites from this geographical area, will enable us to complete a string of conclusions, which at present have only a preliminary status and might seem to present only the most fragile ground for discussion. Bibliography BAILEY et al. 2002 D. BAILEY / R. ANDREESCU / A. J. H OWARD / M. J. M ACKLIN / S. M ILLS , Alluvial landscapes in the temperate Balkan Neolithic: transition to tells. Antiquity 76, 2002, 349–355. C HAPMAN 1994 J. C HAPMAN, The origins of farming in south-east Europe. Prehistoire Européenne 6 (1994), 133–156. C HAPMAN 1997 J. C HAPMAN , The origins of tells in Eastern Hungary. In: P. Topping, Neolithic Landscapes (Oxford 1997) 139–164. C HAPMAN 1999 J. C HAPMAN, Deliberate house burning in the prehistory of Central and Eastern Europe. In: A. Gustafsson / H. Karlsson (ed.), Glyfer og arkeologiska rum – en vänbok till Jarl Nordbladh (Göteborg 1999) 113–126. C OMŞA 1969 E. C OMŞA, Données concernant la civilisation de Vinča du sud-ouest de la Roumanie. Dacia N.S. XIII, 1969, 11–44. D RAŞOVEAN 1991 F. DRAŞOVEAN , Aşezarea neolitică de la Chişoda Veche (jud. Timiş). In: G. Lazarovici / F. Draşovean, Cultura Vinča în România (Timişoara 1991) 71–72. D RAŞOVEAN 1994 F. DRAŞOVEAN , Die Stufe Vinča C im Banat. Germania, 72, 1994, 409–425. D RAŞOVEAN 1994a F. DRAŞOVEAN , The Petreşti culture in Banat. Analele Banatului 3, 1994, 139–170. D RAŞOVEAN 1996 F. DRAŞOVEAN , Cultura Vinča târzie (faza C) în Banat. Biblioteca Historica et Archaeologica Banatica 1. Editura Mirton (Timişoara 1996). D RAŞOVEAN 1997 F. DRAŞOVEAN , Die Petreşti-Kultur im Banat. Prähistorische Zeitschrift, 72, H. 1, 1997, 54–80. Draşovean 1999 F. DRAŞOVEAN , Cultura Petreşti în Banat. Studii privind aşezările preistorice în arealul Mureşului inferior (Timişoara 1999). 186 Florin Draşovean and Wolfram Schier D RAŞOVEAN 2007 F. DRAŞOVEAN , The Neolithic tells from Parţa and Uivar (South-west Romania). Similarities and differences of the organization of the social space. Analele Banatului, Seria nouă 15, 2007, 19–31. G ERLING et al. 2005 C. G ERLING / M. REHFELD / M. W OIDICH, Gräben, Gruben, Häuser-Siedlungswesen und Architektur der späten Vinča-Kultur im Banat. In: W. Schier 2005, 35–40. G OGÂLTAN 2003 F. G OGÂLTAN, Die neolithischen Tellsiedlungen im Karpatenbecken. Ein Überblick. In: E. Jerem / P. Raczky, Morgenrot der Kulturen. Frühe Etappen der Menschheitsgeschichte in Mittel- und Südosteuropa. Festschrift für Nandor Kalicz zum 75. Geburtstag (Budapest 2003) 223–262. H ALSTEAD 1999 P. H ALSTEAD, Neighbours from hell? The household in Neolithic Greece. In: P. Halstead, Neolithic society in Greece. Sheffield Studies in Aegean Archaeology 2 (Sheffield 1999) 67–95. H ORVÁTH 1987 F. H ORVÁTH , Hódmezövásárhely-Gorzsa. A settlement of the Tisza culture. In: L. Tálas, The Late Neolithic of the Tisza Region (Budapest, Szolnok 1987) 31–46. KADEREIT et al. 2006 A. K ADEREIT / B. S PONHOLZ / M. RÖSCH / W. S CHIER / B. KROMER / G. WAGNER, Chronology of Holocene environmental changes at the tell site of Uivar, Romania, and its significance for late Neolithic tell evolution in the temperate Balkans. Zeitschrift für Geomorphologie N.F. 142, 2006, 19–45. KALICZ / RACZKY 1984 N. K ALICZ / P. RACZKY, Preliminary report on the 1977– 1982 excavations at the Neolithic and Bronze Age tell settlement of Beretyóújfalu-Herpály. Part I. Neolithic. Acta Archaeologica Hungarica 36, 1984, 85–136. KALICZ / RACZKY 1987 N. K ALICZ / P. RACZKY, The Late Neolithic of the Tisza Region. A survey of recent archaeological research. In: L. Tálas, The Late Neolithic of the Tisza Region (Budapest, Szolnok 1987) 11–30. KALICZ / RACZKY 1987a N. KALICZ / P. RACZKY, Berettyóújfalu-Herpály. A settlement of the Herpálz culture. In: L. Tálas, The Late Neolithic of the Tisza Region (Budapest, Szolnok 1987) 105–125. KOTSAKIS 1999 K. KOTSAKIS, What tells can tell: Social space and settlement in the Greek Neolithic. In: P. Halstead, Neolithic society in Greece. Sheffield Studies in Aegean Archaeology 2 (Sheffield 1999) 66–76. LAZAROVICI 1972 G. LAZAROVICI, Aşezarea neolitică de la Parţa. Tibiscus 2, 1972, 3–26. LAZAROVICI 1979 G. LAZAROVICI, Neoliticul Banatului. Biblioteca Musei Napocensis IV (Cluj-Napoca 1979). LAZAROVICI 1991 G. LAZAROVICI, Cultura Banatului. In: G. Lazarovici / F. Draşovean, Cultura Vinča în România (Timişoara 1991) 32–40. LAZAROVICI 1991a G. LAZAROVICI, Bucovăţ, Cremeniş (jud. Timiş). In: G. Lazarovici / F. Draşovean, Cultura Vinča în România (Timişoara 1991) 54–58. LAZAROVICI et al. 1991 G. LAZAROVICI / F. D RAŞOVEAN / L. TULBURE, Sanctuarul neolitic de la Parţa (Reşiţa 1991). LAZAROVICI 1989 G. LAZAROVICI, Das neolithische Heiligtum von Parţa. S. Bököny (ed.), Neolithic of Southeasern Europe and its near eastern connections. Varia Archaeologica Hungarica II (Budapest 1989) 149–174. LAZAROVICI et al. 2001 G. LAZAROVICI / F. D RAŞOVEAN / Z. M AXIM, Parţa. Biblioteca Historica et Archaeologica Banatica XIII (Timişoara 2001). LAZAROVICI / LAZAROVICi 2003 G. LAZAROVICI / M. LAZAROVICI, The Neo-Eneolithic architecture in Banat, Transylvania and Moldavia. In: D. Grammenos, Recent research in the prehistory of the Balkans (Thessaloniki 2003) 369–486. LAZAROVICI et al. 2005 G. LAZAROVICI / C.-M. LAZAROVICI / E. JILOT / Z. M AXIM, Absolute chronology of the Banat Culture. In: V. Spinei / C.-M. Lazarovici / D. Monah, Scripta praehistorica. Miscellanea in honorem nonagenarii magistri Mircea Petrescu-Dîmboviţa oblata (Iaşi 2005) 179–191. LAZAROVICI / LAZAROVICI 2006 M. LAZAROVICI / G. LAZAROVICI, Arhitectura neoliticului şi epocii cuprului din România. I. Neoliticul (Iaşi 2006). LINK 2006 T. L INK, Das Ende der neolithischen Tellsiedlingen. Ein kulturgeschichtliches Phänomenon des 5. Jahrtausends v. Chr. im Karpatenbecken (Bonn 2006). LUCA 1998 S. A. LUCA, Liubcova-Orniţa. Monografie arheologică. Editura Macarie (Târgovişte1997). M ANTU 2000 C. M. M ANTU, Relative and absolute chronology of the Romanian Neolithic. Analele Banatului 7–8, 1999– 2000, 75–105. RADU 1979 O. RADU, Plastica neolitică de la Chişoda Veche şi câteva probleme ale neoliticului târziu din nordul Banatului. Tibiscus 5, 1979, 67–76. S CHIER / DRAŞOVEAN 2004 W. SCHIER / F. DRAŞOVEAN , Vorbericht über die rumänisch-deutschen Prospektionen und Ausgrabungen in der befestigten Tellsiedlung von Uivar, jud. Timiş, Rumänien (1998–2002). Prähistorische Zeitschrift 79, 2004, 145–230. The Neolithic tell sites of Parţa and Uivar (Romanian Banat) S CHIER / DRAŞOVEAN 2005 W. SCHIER / F. DRAŞOVEAN , Masca rituală descoperită în tellul neolitic de la Uivar (jud. Timiş). Analele Banatului, NF. 12–13, 2005, 41–56. S CHIER 2005 W. S CHIER (ed.), Masken – Menschen – Rituale. Alltag und Kult vor 7000 Jahren in der prähistorischen Siedlung von Uivar, Rumänien (Würzburg 2005). S CHIER 2006 W. S CHIER, Neolithic house building and ritual in the late Vinča tell site of Uivar, Romania. In: N. Tasić / C. Grozdanov, Homage to Milutin Garašanin (Beograd 2006) 325–339. SCHIER 2008 W. SCHIER, Uivar: a late Neolithic-early Eneolithic fortified tell site in western Romania. In: D. Bailey / A. Whittle / D. Hofmann, Living well together? Settlement and materiality in the Neolithic of south-east and central Europe (Oxford 2008) 54–67. S HERRATT 1983 A. S HERRATT, The Neolithic period in Bulgaria in its European context. In: A. Poulter, Ancient Bulgaria: Papers presented to the International Symposium on the Ancient History and Archaeology of Bulgaria, University of Nottingham, 1981 (Nothingham 1983) 188–198. TĂRĂU / LUCA 2002 D. TĂRĂU / M. L UCA, Panoptic al comunelor bănăţene din perspectivă pedologică (Timişoara 2002). TRINGHAM 2005 R. TRINGHAM , Weaving house life and death into places: a blueprint for a hypermedia narrative. In: D. W. Bailey / A. Whittle / V. Cummings (ed.), (Un)settling the Neolithic (Oxford 2005) 98–111. Florin Draşovean Muzeul Banatului, P-ţa Huniade nr. 1 RO-300002 Timişoara fdrasovean2000@yahoo.com Wolfram Schier Institut für Prähistorische Archäologie der Freien Universität Berlin Altensteinstr. 15 D-14195 Berlin wschier@zedat.fu-berlin.de Zusammenfassung Der Beitrag vergleicht die von den 1970er bis Anfang der 90er Jahren gegrabene mittelneolithische Tellsiedlung von Parţa und die 1999-2009 untersuchte Tellsiedlung von Uivar, beide im rumänischen Banat gelegen und nur etwa 20 km von- 187 einander entfernt. Beide liegen in einer flachen Alluviallandschaft in unmittelbarer Nähe eines noch heute existierenden (Parţa) bzw. fossilen (Uivar) Flusslaufes. In Parţa konnte, auf Grund der langen Forschungsdauer, ein größerer zusammenhängender Siedlungsausschnitt in seiner Bebauungsabfolge untersucht werden. In Uivar wurde in mehreren kleineren Ausschnitten im zentralen und randlichen Bereich des Tells gegraben. Erstmals in Rumänien konnte dort durch eine großflächige geophysikalische Prospektion ein komplexes System an konzentrischen Befestigungsgräben sowie Siedlungsspuren außerhalb des sichtbaren Siedlungshügels nachgewiesen werden. Ein einphasiges Haus mit singulärer, offenbar abgehobener Bauweise konnte 2005 rund 80 m außerhalb des Tells unter kolluvialer Überdeckung vollständig untersucht werden. Die Bebauungsstruktur in Parţa weist einen über die Zeit zunehmenden Organisationsgrad auf, besonders spektakulär sind zwei als „Heiligtum“ interpretierte Sondergebäude. In Uivar lässt ein besonders großes Gebäude mit zahlreichen Feuerstellen und weiteren Besonderheiten ebenfalls eine spezielle, wohl rituelle Funktion vermuten. In beiden Tellsiedlungen bestehen die regulären Häuser aus mehreren Räumen und weisen oft ein Obergeschoss auf. Die Entwicklung in Parţa setzt früher ein als in Uivar und endet mit einem Zerstörungshorizont zu Beginn von Vinča C. Die Besiedlung wird dann auf einem benachbarten Hügel, dem Tell Parţa II, fortgesetzt, der allerdings nicht Gegenstand des Beitrags ist. In Uivar beginnt die Siedlungstätigkeit wohl etwas vor ihrem Ende in Parţa. Die rund 4 m mächtige Stratigraphie gehört mehrheitlich den Stufen Vinča C1 und C2 an. Die Hausarchitektur weist über die Zeit charakteristische Veränderungen auf, doch erstaunt die hohe Lagekontinuität der Gebäude. Das Befestigungssystem konnte nur durch kleine Sondagen untersucht werden, die Funde lassen bis jetzt keine feinchronologische Differenzierung zu. Es spricht jedoch viel für ein dynamisches Baugeschehen und eine ständige Anpassung an eine sich vergrößernde Siedlungsfläche. Schließlich wird die These vertreten, dass der Grabenaushub nicht zur Anlage von Wällen (nicht nachgewiesen), sondern zur Gewinnung von Baulehm für den Hausbau verwendet wurde. Abschließend wird für Uivar auf der Basis des Tellvolumens die Anzahl insgesamt errichteter Häuser hochgerechnet. Über eine Lebensdauer von rund 500 Jahren lassen sich 3.500-5.000 Häuser schätzen, von denen nur ein geringer Teil durch Brand zerstört wurde. Das Verhältnis zwischen verdichteter Kernbesiedlung, die zur Tellakkumulation führte und lockerer Außensiedlung kann noch nicht abschließend interpretiert werden.