Version classiqueVersion mobile

The Human Face of Radiocarbon

 | 
Zoï Tsirtsoni

Greek Eastern Macedonia

Chapter 15. The Late Neolithic II (Chalcolithic)-Early Bronze Age transition at the tell of Dikili Tash1

Zoï Tsirtsoni

Texte intégral

Introduction

  • 1 I wish to thank the other co-directors of the current research project, Pascal Darcque, Chaido Kouk (...)
  • 2 Research conducted under the auspices of the Archaeological Society at Athens and the French School (...)

1Dikili Tash (fig. 1) is located in the south-eastern part of the Drama plain (41° 00’ 37’’ N, 24° 18’ 30’’ E), which was occupied by a marsh until the early 20th century. It is one of the biggest tells in the area (present ground surface ca 4.5 ha, height 17 m), and one of those whose development is known best, thanks to the systematic excavations of the last fifty years2.

  • 3 See especially Koukouli-Chrysanthaki 1994, p. 125‑129; Koukouli-Chrysanthaki et al. 1996a, p. 689‑6 (...)

2The research carried out from 1961 to 1996 established that the site was occupied from at least the Late Neolithic to the Late Bronze Age (ca 5400‑1100 BC). Stratigraphic and material evidence collected from the ca 13 m of archaeological deposits excavated at different parts of the mound suggested that the occupation had been more or less continuous during these millennia, and that the settlement played a quite important role in the regional networks, especially in the earlier periods (Neolithic to Early Bronze Age). Well-preserved architectural features, or entire buildings, were brought to light in levels of almost all periods. Among the most impressive was a group of four houses destroyed by fire at the end of the Late Neolithic II period3.

  • 4 The corresponding layers were reached through a series of core-samplings, but have not yet been exc (...)
  • 5 Darcque et al. 2008, p. 533‑535; Darcque et al. 2009, p. 536‑539.

3A new cycle of research started in 2008, at roughly the same time as the “Balkans 4000” project: investigations pushed back the start of the settlement by almost one thousand years, to the Early Neolithic period (ca 6400/6200 BC), and confirmed the persistence of occupation at this spot during the whole 6th millennium BC (Middle-early Late Neolithic)4. They also showed that the settlement was still important in later periods, namely the Late Bronze Age, during which we now have evidence of repeated building episodes5. Thus the site’s already long stratigraphic and chronological sequence has been further expanded, reaching now almost 8000 years (ca 6400 BC-1900 AD), part of which went on uninterruptedly. The new research also revealed further evidence about the economy and organization of the settlement during the final stages of the Late Neolithic II (Chalcolithic) period and the beginning of the Early Bronze Age, thus allowing a better understanding of the conditions that prevailed at the settlement around the “crucial” date of 4000 BC.

Fig. 1 – Aerial view of the Dikili Tash tell, in the early 1990s (photo: Ch. Tsouroukidis).

4In this report, we only deal with this last issue. First, we summarize our state of knowledge about the Late Neolithic II (Chalcolithic) and Early Bronze Age occupation at Dikili Tash prior to 2008; then we discuss in some detail the results of the 2008‑2010 fieldwork and of the analyses carried out on them, including 14C dates.

The Late Neolithic II and the EBA periods: state of research before 20086

The stratigraphy

  • 7 This designation should change in the future, since we know now that the occupation starts before t (...)
  • 8 The presence of LN II at the basis of sector I/1961 has not been proven: Koukouli-Chrysanthaki & Ro (...)
  • 9 This is due to the excavation strategy, which never proceeded by deep stratigraphic trenches (like (...)
  • 10 All measures are given in absolute altitudes.

5Layers of the Late Neolithic II (Chalcolithic) period, labelled so far “phase Dikili Tash II”7, were unearthed in almost all the parts of the tell excavated between 1961 and 1996 (fig. 2): close to the top (sector A2), in the southern and eastern slopes (sector B2, V, and VI), and at the southern, and possibly the eastern, feet (sector II and I/19618). They were found directly underneath Early Bronze Age layers (in sectors A2, VI, and I/1961), or directly above Late Neolithic I layers (in sectors B2 and V), although never both9. They thus formed the median part of the tell’s archaeological deposits, situated between ca 60.50/61 and 64.60 m above sea level10. In addition, part of the LN II settlement obviously extended also to the periphery of the tell, at an altitude comprised between 54 and 55 m. No stratigraphic hiatus (i.e. abandonment layer) had been observed, neither at the transition from one period to the next, nor between the layers of a single period.

  • 11 The most impressive being the discovery of evidence for wine-making: Valamoti et al. 2007.

6Substantial architectural remains of the LN II period have been unearthed at various parts and in various levels, but they were usually isolated and could rarely be followed over more than a few squares meters. The discovery, between 1989 and 1996, of the group of houses in sector VI (supra), with their content practically intact, radically changed this picture: it was now possible to observe in detail the layout, the construction techniques and internal arrangement of a number of truly contemporaneous buildings. The potential of this assemblage has not been fully exploited yet (see below), but its study has already added an astonishing quantity of information about the household economy at LN II Dikili Tash11.

Fig. 2 – Topographical plan of the works conducted until 2010; are shown the areas where LN II and/or EBA deposits have been investigated.

  • 12 Treuil 1992, p. 19‑20.
  • 13 This was indeed the basic assumption behind J. Deshayes’ decision to display his trenches in a “lad (...)
  • 14 Séfériadès 1983; Demoule 2004, p. 98‑99, 102‑177, table 1.4, 3.25 and 4.15.

7Although no straightforward connections could be established between layers distant from one another by several meters12, it was admitted that the deposits in the median part of the tell were ordered in a more-or-less regular way, following simple cumulative processes13. Pottery analysis from the two sectors that were closer to the tell’s centre (the western part of A2, comprising squares R-T-U/24‑25, and sector B2, comprising squares W29‑30 and X29‑30) and which displayed the longer succession of layers, allowed ten “horizons” (I‑X) to be distinguished, presumably corresponding to chronologically distinct stages of the settlement’s overall evolution. They were grouped into three sub-phases (IIA-C), which have been further correlated to analogous stages from other sites in the broader area14.

  • 15 This pottery style actually seems to appear already in the middle stages of the sequence. One of th (...)
  • 16 This type of vessel is also well-represented in the material from sector II/1967: Koukouli-Chrysant (...)
  • 17 Koukouli-Chrysanthaki & Romiopoulou 1992, p. 232‑233, expressed a different opinion when discussing (...)

8According to this scheme, the earliest layers (sub-phase IIA, horizons I‑III, corresponding to the lower levels in B2) would be characterized by the presence of incised pottery of the “Maritsa style” and of graphite-decorated pottery with fine, or alternating fine and medium, lines. These features allowed a parallelism with an early stage of the Maritsa sequence, further synchronized with the beginning of Karanovo V. The replacement of the Maritsa pottery by the “linear style” incised pottery, together with a few changes in the graphite-pottery repertoire, would mark the next stage (sub-phase IIB, ceramic horizons IVA-VIB, represented by the middle levels in B2), which fell entirely into the Karanovo V period. This stage would also witness the explosion of the local black-on-red painted pottery, especially the variant with the geometric well-organized pattern, termed “Dikili Tash style”. The last sub-phase (IIC, horizons VII to X, represented in the upper layers of sector B2 and in sector A2) is presumably characterized by further changes in the graphite-painted pottery (“negative” motifs appearing, changes in shapes, thickness of lines...), as well as in the black-on-red ware, which now takes on a more free-style mode of display (“Galepsos style”)15. The linear incised pottery would also decline significantly, although it is now that the category combining incised, graphite-painted and eventually red-crusted decoration flourishes16. This stage was thought to be more or less synchronous with the start of the Karanovo VI phase. No material assigned to “mature” Karanovo VI was retrieved, nor was material assigned to later stages, such as the one presumably represented by the phase IIIC in neighbouring Sitagroi, or in the “Final Chalcolithic” layer at Kastri on Thasos17.

  • 18 In the two southernmost squares of sector B2 (X29 and X30), the levels 12 and 10, respectively, lie (...)
  • 19 One of the elements that seemed to need revision was the evolution of the black-on-red painted pott (...)

9Subsequent evidence from the other excavated sectors suggested that the main lines of this scheme were altogether valid, but several points required adjustments. Indeed, neither the level V/East/1, cut directly into the previous LN I levels in the southern slope and lying at approximately the same altitude as the lowest LN II levels in the neighbouring sector B2, provided a typical IIA phase assemblage18, nor did the material from the houses excavated in sector VI correspond exactly to Demoule’s definition of the IIC sub-phase19. Yet, from a stratigraphic point of view they represent, respectively, the beginning and the end of the LN II sequence in this part of the tell.

  • 20 Treuil 1983, p. 98‑99, actually proposes three subdivisions; see also Malamidou, forthcoming.

10The only sector so far to display a relatively long sequence of Early Bronze Age levels is A2: several superimposed floors and individual structures (ovens and hearths) have been discovered, but no complete house plan. The study of the associated pottery allowed two sub-phases (Dikili Tash phases IIIA and IIIB) to be distinguished, corresponding respectively to the phases IV and V of neighbouring Sitagroi, and representing the EBA I (pre-Troy I) and EBA II (Troy I) periods according to the local terminology. The phase IIB could probably be subdivided further, as is the case with the phases Va and Vb in Sitagroi20.

  • 21 Koukouli-Chrysanthaki 1994, p. 124‑125.
  • 22 Their precise nature is not known (occupation level in situ, colluvial deposits, disturbed layer?), (...)
  • 23 Koukouli-Chrysanthaki & Romiopoulou 1992, p. 236‑245.
  • 24 Ibid., p. 245‑247, mention the mixing with Neolithic sherds, but not with positive EBA I material.

11Although poorly preserved due to the proximity of the surface, the EBA level installed directly above the destruction layer of the houses in sector VI (“building horizon A”, towards 64/64.15 m), seemed to belong to the earliest stage of the period. It was characterized by an ash-grey colour, which contrasted strongly with the vivid red and yellow tones of the underlying fired debris, and by the presence of many pits. Yet, it was not excluded that this level might actually correspond to more than one event, and that part of it stretched down to the EBA II period21. The situation was different in sector I/1961. Here, the LN II deposits22 were surmounted by a series of layers including the remains of successive rectangular buildings (the only buildings so far to provide an almost complete ground plan for this period), all of which date to the EBA II period23. In other words, the EBA I settlement either never extended into this area or, if this was the case, its traces have been wiped out by later events (human activity or erosion)24. The very fact of a direct LN II-EBA II succession, at least in one area on the tell, shows (if this was needed) that the stratigraphic contact is by no means proof of chronological continuity. However, what could be easily understood at the scale of a peripheral trench seemed difficult to demonstrate at the scale of the whole settlement.

The absolute chronology

  • 25 The 14C dates from the first research program (sectors A2 and B2) have been published by Treuil 199 (...)

12The 14C dates available up to 2008 (25 measurements altogether for the LN II layers25: fig. 3) fell into the period between 5400 and 3800 cal BC, calibrated at 2σ (values between 6100 to 5296 BP, with statistical errors ranging from 20 to 200 years BP).

Fig. 3 – Diagram OxCal with the available 14C dates for LN II levels until 2008 (calibrated at 2σ).

  • 26 At first, J. Deshayes saw at this point the effects of a possible invasion from the lower Danube ar (...)

13The start of the LN II was fixed some time around 4800 cal BC. This assumption was based, on one hand, on the dates obtained for the last LN I levels in the tell, which cluster towards 4900 cal BC, and on the other hand, on the fact that the two periods seemed separated by a short gap, visible in the pottery typology26. Unfortunately, no good 14C dates were obtained from levels V/East/1 and V/East/2. But the unique date from a IIA level in sector B2 (Gif-1736: 5850 ± 160 BP = 5205‑4358 cal BC) fits the general sequence well and its calibrated result does not contradict this option.

14Two dates (Gif-1424: 5750 ± 140 BP = 4936‑4341 cal BC; Gif-1423: 5650 ± 140 BP = 4841‑4233 cal BC) refer to IIB levels, again from sector B2. The dates Gif-1736 and Gif-1424 appear particularly coherent, for they come from two successive layers of the same square (respectively levels 12 and 9 in X29). Furthermore, it is quite reassuring to see that the three results are slightly earlier than those assigned to phase IIC, or roughly contemporary with the earliest of them.

15Unfortunately there are no dates from the following levels in the same sector. The dates that refer to phase IIC – or, broadly speaking, to the later stages of the LN II period – come from two other sectors: a) A2, in the centre of the tell; the samples come from two distinct areas, respectively squares R24 and T24/U24, distant from each other by ca 15 m in the north-south direction; b) VI, lower on the eastern slope.

  • 27 Treuil 1992, p. 34. There could also be another explanation: that their stratigraphic position is n (...)

16The two dates from the southern part of sector A2 (Ly-1062 and Ly-1064) look problematic, for they are very high (respectively 6100 ± 200 and 6040 ± 120 BP), even higher than those of phases IIA and B. Calibrated at 2σ, they give results that could fit the time span of phase II (5473‑4585 and 5296‑4689 cal BC), but definitely not its last stage. This anomaly has been attributed to the small size of the samples27.

  • 28 Treuil 1992, p. 34‑35, suggested considering the seven dates together, and concluded that the end o (...)

17The other two dates, from the northern part of the sector (square R24, the nearest to the top), are more plausible and perfectly coherent with those from sector VI. Like the latter, they refer to the last LN II event (level 16, surmounted directly by level 14, of EBA date), which they place at 5750 ± 140 BP = 4936‑4341 cal BC (Gif-1425), or 5600 ± 150 BP = 4789‑4060 cal BC (Gif-1738)28.

  • 29 In two laboratories: Demokritos in Athens, under the supervision of Y. Maniatis, and Heidelberg, un (...)
  • 30 Actually the first comprehensive analysis was undertaken in 2010, as part of a regional synthesis i (...)

18The 18 dates from sector VI are, in all respects, the most reliable group of evidence. They come from secure house contexts, where it was possible to distinguish between the mass of destruction layer, on the one hand, and specific features in situ (ovens, platforms, floor deposits), on the other. In addition, the houses contained substantial quantities of charred grains: thus it was possible to date not only big charcoal fragments (as was the case in the other sectors), but also a few short-lived samples. Finally, these were processed more recently29, that is, at a time when the 14C methods were much more advanced: statistical errors are therefore much smaller (20 to 50 years BP in general, against 120 to 200 for the earlier dates), thus allowing a precision of 100 to 300 years on calibrated results, versus 600 or 900 years for the previous series. No full fine analysis of these dates was made before 200830, but their overall examination suggested a time span between 4450 and 4250 cal BC. Five dates had a lower limit in the years around 4000 cal BC (HD-20445, HD-20854, HD-20986, DEM-179 and DEM-553, with values between 5440 and 5300 BP, although sometimes with substantial errors), but they could also fit well with the main cluster. This interval corresponded to the lifetime (building and destruction) of this particular group of houses, and by no means to the whole IIC phase, whose definition is, as we said, problematic.

  • 31 The dates referring to the EBA levels in sector A2 were all judged invalid: see discussion in Treui (...)

19In any case, no other LN II level covered this destruction layer. The spot would have been unoccupied until the beginning of the EBA, which according to the unique 14C date obtained from a piece of charcoal collected in a pit (DEM-552, 4419 ± 30 BP), would have started around 3300‑2900 cal BC31.

  • 32 TL-dates were made on samples from two different ovens, one in House 3 (two measurements) and one i (...)

20It remained to explain why there was no visible trace of this 1000 year long gap in the occupation, and to check whether this was indeed the case in the other parts of the tell, or if the settlement had moved horizontally to another spot. The possibility that the gap was (in part or as a whole) an artifice of the 14C method was abandoned in the late 1990s, when a series of thermoluminescence (TL) dates was performed on the site: their results coincide, indeed, perfectly with those from the radiocarbon and, in spite of their bigger statistical deviations, confirm that the houses of sector VI were destroyed before 4200 BC32.

The limits of the settlement

  • 33 Tsountas 1908, p. 31‑49; Papaefthymiou-Papanthimou & Pilali-Papasteriou 1988, fig. 1, 7.
  • 34 Darcque et al. 1990, p. 877‑878, fig. 1; Darcque et al. 1992, p. 715, fig. 1.

21Another question that remained open was that of the organization of the Late Neolithic II settlement, and more specifically that of the possible existence of a stone enclosure, like those found in several other Late Neolithic settlements in Greece (for instance Dimini or Mandalo33). In sector II, at the southern foot of the tell, excavation had brought to light over a very small surface (less than 6 m2, in the northern part of the 7 x 2 m trench) the badly preserved remains of a destruction layer, apparently LN II but with no 14C dates, and right to its south, the remains of a strongly sloping massive stone structure34. Other stone beds had appeared at the same area in higher levels, which clearly looked fallen (i.e. brought down by erosion). But the stone bed on which the excavation had stopped might have been different: it was not clear, indeed, whether this was another eroded deposit, or a structure in situ (a retaining wall or enclosure), and if yes, whether it was Neolithic as well, or later.

The 2008‑2010 research35

  • 35 In addition to funding from the two main institutions, the research was funded by the French Minist (...)
  • 36 A synthesis, focusing mainly on the LN II levels, is presented in Darcque et al. 2015.

22In 2008, excavations were resumed in two of the areas where it was possible to follow the transition from the LN II to the EBA36.

23One was sector VI, now labelled 6. Investigation was undertaken in the area of House 1, the westernmost of the four houses brought to light during the previous program (fig. 4). The sector was extended to the north in order to include the whole building; this extension provided the occasion to re-examine the LN II-EBA interface more thoroughly (total investigated area 98 m2). The other was sector II (now 2) at the southern foot of the tell. Here, an important extension was made and the sector reached 150 m2 in 2010. In addition, two small drillings were made at its centre and south (fig. 2: C5 and C6), which allowed us to complete our knowledge about the lowest parts of the sequence.

Fig. 4 – Schematic plan of houses in sector 6.

Sector 6

  • 37 This study, carried by Dr Tania Valamoti (University of Thessaloniki), received extra funding in 20 (...)
  • 38 Actually the grains that provided the date Lyon-6013 were originally believed to belong to the fill (...)

24In sector 6, the new fieldwork confirmed the excellent state of preservation and the wealth of the finds in House 1 (fig. 5). Its ground plan has been better circumscribed, although we still do not know exactly where its outer limits are. The one room excavated so far is roughly square (ca 6 x 7 m), like rooms A and B in House 4, with which it shares many similarities, and contains an oven near the back western wall (locus 6‑015), probably facing the entrance. A grindstone and a concentration of bitter vetch (Vicia ervilia) [fig. 6] were found on the platform right to the south of the oven, whereas several ceramic vessels surrounded by great quantities of charred grains were found in the space in front of it (fig. 7), mostly wheat (Triticum monococcum) and grape (Vitis vinifera). Other groups of vessels and implements (grindstones, other stone and bone tools), as well as concentrations of seeds and fruits, were found further to the south and east (fig. 8). The abundance of charred plant remains allowed a fine study of the crop management strategy in this particular household37. It also provided some further samples for 14C dates: the four new dates (table 1: DEM-2031 and 2057, Lyon-6013 and 6014) add themselves to the five existing already for this house (DEM-176 to 179 and DEM-553), and confirm that it was destroyed by fire in the years around 4340‑4260 cal BC, as suggested in particular by the two short-lived samples collected in the destruction layer (Lyon-6013 and DEM-2031)38.

Table 1 – 14C dates from samples collected in sector 6 during the 2008 work.

  • 39 There has not been any anthracological analysis on these samples.
  • 40 I express my gratitude to Philippe Lanos and Philippe Dufresne, developers of the program RenDateMo (...)

25Considering the nine dates from the area of House 1 together, old and new (table 2), we note that most of the charcoal is older or contemporary to this destruction date: this is perfectly logical since they should date both the construction and the use of the house39. But two charcoal samples (DEM-553 and DEM-2057), collected from the upper parts of the layer, have the major part of their distribution around 4250‑4000 cal BC. Although they could still fit the main destruction cluster, they might well also indicate an episode of reoccupation, shortly after the destruction of the houses (compare the modelled versions with two or with three “events”: fig. 9: a-b40). Similar evidence has also been recorded in some of the other houses, especially House 3. No architectural features are connected with this possible event, which seems normal, given the erosion that would have affected the area after its abandonment. No particular group of finds has been recorded in situ either, but several more or less well-preserved fragments could be associated with it (see discussion below).

Table 2 – All available 14C dates from House 1 up to 2012.

Fig. 5 – General view of House 1, to the northeast, at the end of the 2010 campaign (photo: P. Darcque).

Fig. 6 – Grindstone and concentration of bitter-vetch to the south of oven 6‑015 (photo: P. Darcque).

Fig. 7 – Vessels in situ in front of oven 6‑015; the surrounding sediment contained great quantities of charred grape pips (photo: P. Darcque).

Fig. 8 – Group of vessels next to a fallen wall fragment to the east of the room (photo: P. Darcque).

Fig. 9 – Modelled diagram with the 14C dates from House 1, using the program RenDateModel (© CNRS-IRAMAT).
Two versions are proposed: in version (a) only two “events” are distinguished, one with all charcoals (presumably dating the construction and use of the house) and one with seeds (dating the destruction); in version (b) we introduce the possibility of a third “event” (reoccupation) represented by the two late charcoals.

  • 41 This kind of vessel (small rounded cups with raised handle, also called “scoops”, or “puisettes” in (...)
  • 42 Supra, n. 38.

26The next occupation level (or levels) is established directly on top of these remains, and actually penetrates them, as already suggested by the previous excavations. Indeed, with the exception of a possible oven (locus 6‑001) in the northern part of the sector, at 63.98 m, all the other features belong to hollow structures, that could begin at either one or at different levels that were situated at a higher altitude and are now completely lost (fig. 10). Two main groups of features are recorded: a) groups of stones that appear today between 63.74 and 63.86 m and which were originally found in pits supporting wooden posts; b) simple pits, coated with a white marly coating on the interior, probably used for storage, or uncoated, that could have been rubbish-pits; their diameter is comprised between 0.70 and 1.50 m. They rarely hold any trace of their original content (small bones, or fragmented ceramic vessels, especially “dippers”, fig. 11)41; they are more usually filled up with material coming from the surrounding (i.e. earlier) deposits. Indeed, three out of four samples collected in or around such features gave LN II dates42, and only in one case did the sample provide an EBA date (Lyon-6012: 4445 ± 30 BP), probably corresponding to its true age. This date is perfectly coherent with the 14C date obtained from another pit above the house 3 (supra), and confirms that this part of the tell was re-occupied around 3300‑3000 cal BC.

Fig. 10 – Part of the EBA remains on top of the LN II House 1 (zenithal photo: Chr. Gaston).

  • 43 A similar fragment from Sitagroi is illustrated in Renfrew et al. 1986, pl. CI: 7 (“sieve fragment” (...)
  • 44 The rest of the vase was obtained by refitting. Similar conical bowls exist at Sitagroi phase IV: S (...)

27The only finds that are clearly associated with this layer, aside from those found inside the pits, are a group of vases found in the northern part of the sector: they comprise an almost intact “incense burner” (fig. 12)43 and a conical bowl (fig. 13), whose base was found standing on what might be a floor, at ca 63.75 m44. The almost complete absence of wall-fragments, together with the dull colour of the surrounding sediment, which bears no trace of burnt clay inclusions, suggest that we might be in an open, or possibly semi-open space (courtyard?).

Fig. 11 – EBA I “dippers” (photo: F. Bourguignon; drawing: R. Douaud).

Fig. 12 – EBA I “incense burner” (photo: F. Bourguignon; drawing: S. Monnet).

Fig. 13 – EBA I conical bowl (photo: Ph. Collet-EFA©; drawing: R. Docsan).

Fig. 14 – Some of the vessels found in House 1 (photos: Ph. Collet-EFA©; drawings: R. Docsan, R. Douaud).

Fig. 15 – More vessels from House 1 (drawings: R. Douaud).

Fig. 16 – Vessels from the other houses of sector 6.
a: brown on-cream painted pot from the area of Houses 2‑3, h. 10 cm; b: globular pot with ranges of impressions, h. 15 cm; c: graphite-painted askos from House 4, h. 12.5 cm; d: incised jar from House 3, h. 74 cm (photos: a, Ph. Collet-EFA©; b and d, St. Stournaras; drawing of c: I. Vajsov).

28The pottery retrieved from this sector in 2008 and 2010 is still under study. But some preliminary observations can be made.

    • 45 Koukouli-Chrysanthaki et al. 1996a, p. 703, fig. 17; Darcque et al. 2007, fig. 6‑7.
    • 46 Koukouli-Chrysanthaki et al. 1996a, p. 693; Darcque et al. 2011, p. 196. The recent technological s (...)

    The pottery assemblage from the interior of House 1 displays a great variety of wares and techniques, as well as of shapes and sizes. The range of vessels goes from tiny cups and small rectangular boxes to large jars, with all kinds of shapes in between (collared pots, bowls, amphorae…), decorated or not (fig. 14; 15). The same was observed also in the rooms of the other houses in sector 645. This variety reflects the broad range of activities performed in these spaces, suggested also by the other finds, and supports the idea that each one of them represented an individual household46. But the exact composition of the vessels’ inventory is not the same everywhere; in other words, there is no standard battery, no typical “table set”. This is not really surprising: like in our present households, people probably respected some common imperatives (i.e. have a minimum number of cooking-vessels, also have some vessels for drinking or eating), but the choice of the exact utensils that they introduced into their home was made according to their specific needs, habits or aesthetic values, and according to opportunities as well.

    • 47 Demoule 2004, pl. 98: 11167 (horizon VIII).
    • 48 Ibid., pl. XIV: 6 (DT 23, horizon VII).
    • 49 Supra, n. 15.
    • 50 From sector II/1967, i.e. the present sector 6, around the Houses 2‑3. Many thanks to Ch. Koukouli- (...)
    • 51 J. Deshayes is clearly describing it in one of the preliminary reports, as a variant of the black-o (...)

    Most of the categories and shapes found in House 1 appear in Demoule’s typology for the later stages of the LN II period (supra), but others do not. Thus, composite graphite-painted vessels have parallels, more or less close, among the material from the upper levels in the neighbouring sector B2 (compare the two-handled amphora shown here [fig. 15: a] with a specimen from level 2 in square W2947, the pedestal [Fig. 14: e] with another from level 2 in the square X2948). Furthermore, a fragmented oval-mouthed cup (fig. 14: c) is almost identical to a cup with black-on-red painted decoration (“Galepsos style”) from a lower level in the same sector49. On the contrary, brown-on-cream painted vessels like the one shown here (fig. 14: d), are not at all discussed by Demoule, although at least one parallel existed from Theocharis’ excavations (fig. 16: a)50, and possibly more sherds among the material from Deshayes’ works in sector A251. The discovery in situ of complete (or almost complete) vessels from all these categories proves that all of them are contemporaneous, at least at some point in their existence.

    • 52 Demoule 2004, p. 161. See also Boyadzhiev et al. 1993, p. 80.
    • 53 For instance, at Karanovo itself or Yunatsite: see chapters 7 and 9 in this volume.
    • 54 As was presumed by J.‑P. Demoule, see supra, p. 274 and n. 14.

    Some of the categories and shapes attested in the other houses of sector 6, for example the collared vessels with densely arranged impressions on the body (fig. 16: b), the “askoi” (fig. 16: c), or the big incised jars (fig. 16: d), are not found here. Both the stratigraphy and the radiocarbon dates prove the contemporaneous character of the four houses, so we should attribute such discrepancies to contextual rather than to chronological differences. Further, this is important for regional comparisons. Vessels with impressed (or barbotine) decoration, as well as “askoi”, are considered to be characteristic of the latest stages of the Late Chalcolithic (the end of the KGK VI “culture”)52, and it seems, in the light of the new 14C dates from a number of sites53, that this is indeed so. But the opposite is not true: their absence from a ceramic assemblage is by no means an argument that the latter is earlier in date54. The actual synchronism of all four houses from sector 6 at Dikili Tash with the “late KGK VI” levels in those same sites shows the limits of an exercise based on material evidence alone.

    • 55 Of course, other explanations are possible: presence of a shelf or upper-storey, or the erosion of (...)
    • 56 Demoule 2004, pl. C: 3‑4, D: 4.
    • 57 Todorova, chapter 10 in this volume, p. 181, fig. 12 (sp. 12: 3). None of these vessels can be take (...)

    Keeping this in mind, we should however make a comment about some pottery “types” that are apparently found only in the upper parts of the house’s destruction layer, but not within its floor deposit. If our assumption about a reoccupation of the area after the destruction of the houses is correct (supra), these vessels could go with this last episode, and not with the houses themselves55. The first is a two-handled bowl with a globular body, decorated with incised spirals; the surface between the incisions is covered by red-crusted paint (fig. 17). It should not be confused with the well-known type described by Demoule of big open vessels combining incised, impressed, graphite-painted and red-crusted decoration56. The present vessel is altogether finer (both in terms of fabric and wall-thickness), the shape is different, and the decoration is arranged freely on the whole surface and not in zones. It recalls some of the vessels found in Dolno Dryanovo, whose occupation, as we know now, also dates to the years after 4200 cal BC57. Not far from this vessel, a graphite-painted plate with an upright rim and two small handles at its base (“Kritsana bowl”) was found in a practically horizontal position (fig. 18). The presence of such a vessel, typical of the “normal” late stages of the LN II at Dikili Tash, in a presumably slightly posterior context is not surprising: it is very likely indeed that much of the material of this level (if it truly exists) would be identical to that of the underlying house-level and therefore undistinguishable in typological terms.

Fig. 17 – Two-handled globular bowl with incised and red-crusted decoration (photo: P. Darcque; drawing: R. Douaud).

Fig. 18 – Graphite-painted “Kritsana bowl” (photo: P. Darcque; drawing: R. Douaud).

Fig. 19 – Fragments of a gray-polished globular bowl with beaded rim (photo: Ph. Collet-EFA©; drawing: R. Douaud).

Fig. 20 – Fragment of a two-handled bowl with incised decoration (photo: Ph. Collet-EFA©; drawing: R. Docsan).

  • 58 I am grateful to N. Todorova for drawing my attention to these fragments, whose importance might ha (...)

29Two other possible new “types” have been recorded so far, both of which again have parallels in the Rhodopes. One is represented by an undecorated (gray-polished) globular bowl with beaded rim (fig. 19), the other by a small two-handled bowl with a sinuous profile and incised decoration (fig. 20)58. The latter is further attested by several fragments in sector 2, from contexts that also may belong to an ultimate LN II stage (see below).

  • 59 Koukouli-Chrysanthaki & Romiopoulou 1992, p. 233, and pl. 42b.
  • 60 Evans 1986, pl. XCIV: 5‑6, 10‑12, and colour pl. D: 8; Koukouli-Chrysanthaki 1972, p. 525‑526 and f (...)
  • 61 Koukouli-Chrysanthaki & Papadopoulos, chapter 18 in this volume, p. 347‑348.
  • 62 Koukouli-Chrysanthaki & Romiopoulou 1992, p. 233, with references to Séfériadès 1983, fig. 44‑46, a (...)

30We could add to these another kind of ware, the dark-burnished vessels with grooved decoration. No such fragments were collected in the area of House 1, but a well-preserved example was retrieved from the first year of excavation (i.e. the upper part of the destruction layer) in the area of the future Houses 2‑359. The suggested parallelism with vessels from Sitagroi IIIC and from Kastri60 is now supported by the new 14C dates, which place the end of the Chalcolithic at Kastri at the interval 4250‑4050 cal BC, i.e. the same as the presumed last LN II event at Dikili Tash61. One should be careful however, because this type is also attested in the EBA I period62.

Sector 2

  • 63 The results have been discussed also with some detail in Darcque et al. 2014; Darcque et al. 2015.

31In sector 2, the 2008‑2010 research confirmed and specified the LN II date of the occupation in the northern part (supra), and clarified the nature of both the preceding and subsequent events63.

  • 64 Lespez et al. 2013, p. 36‑37.

32According to the new evidence, the area was settled in this period for the first time. Indeed, the deposits lying underneath the LN II level (more than 2 m of sediments down to the natural soil) consist exclusively of colluvia, formed by the erosion of the tell itself. At the base of the sequence the remains of a dig were detected, which was made directly in the natural soil, probably during the Early Neolithic period, and filled up in the following centuries64. The 14C dates showed that at least 1000 years of erosion, from ca 6000 to 5000 cal BC (table 3), are represented in these deposits.

Table 3 – 14C dates from samples collected in Sector 2 (including cores C5 and C6).

  • 65 Both the fine macroscopic observation of layers in the field and the micromorphological analysis ha (...)

33The LN II is thus a period of expansion (or re-arrangement) of the settlement over an area that was previously uninhabited. But it appears that what was seen originally as a single occupation layer is actually a series of very closely spaced episodes, all more or less affected by the erosion. Their succession, visible in the trench’s northern and eastern section (fig. 21), has been confirmed by the micromorphological analysis of sediments from various parts of the sector65.

34The first building episode is characterized by an oven (locus 2‑002, base at 54.73 m), whose badly preserved remains were already seen in 1989 (supra). Its date has now been established both in relative and absolute terms: the presence of several graphite- and black-on-red painted pottery fragments (fig. 22), lying on the ground almost horizontally, suggests a date in an advanced stage of the Late Neolithic II. This agrees well with the 14C date obtained from charcoal collected inside the oven, which falls at 4520‑4360 cal BC (Lyon-6011: 5620 ± 35 BP). The remains of another oven (locus 2‑014), discovered several meters to the west in a similar stratigraphic position, probably belong to the same level. No radiocarbon dating was conducted, but a four-legged vessel with incised decoration and traces of red-crusted paint was found among the debris (fig. 23). The original position of the habitations to which these structures belonged probably lay a few meters to the north, from where they would have slid slightly.

35The next episode is identifiable only through micromorphological analysis. The north-eastern part of the sector now seems to be an outdoor space, possibly a barn, marked by the presence of calcite spherulites typical of sheep or goat excrement; alternatively, the excrement could have been added to the earth used for re-plastering the house-floor. This change in the use of space, if confirmed, could suggest the abandonment of building activity at this particular spot, but not necessarily a shrinking of the settlement in general. In any case, this seems to be the last in situ activity in this part of the site.

36Indeed, immediately after that, the area is again exposed to erosion. The successive colluvia dismantle and cover the nearby structures. One of these colluvia brings down the stones that form the bed brought to light in 1989‑1993 toward 54.43 m (labelled now locus 2‑001: fig. 24) and its continuation to the west, revealed by the new excavations (locus 2‑008, up to ca 55.10/55.20 m). The stones are therefore not in their original place, but re-deposited here as part of an important erosive episode. Three charcoal samples collected in this layer, from the area directly to the south of the oven 2‑014, gave dates at ca 4200‑4000 cal BC (table 3: Lyon-7618, DEM-2125, DEM-2144), i.e. slightly after the destruction of the houses in sector 6 (supra). Evidence from the pottery collected between the stones of locus 2‑001 is in good agreement with this date (fig. 25: a-c). This means that the structure (or the structures) to which these stones originally belonged was Neolithic: it consisted of a wall, or system of walls, that stood a few meters higher on the slope. Given its position at the periphery of the built space, it would have had a retaining or defensive function.

  • 66 We did not attempt to date the total organic matter in the layer, fearing that it would not possibl (...)
  • 67 A third sample from the same area gave a date at the early 6th millennium BC (table 3: Lyon-7620). (...)

37Erosion continued a little longer after this collapse event, meaning that the area was still exploited, although not inhabited anymore. Then all activity stopped: vegetation gained ground completely and sediments were transformed pedologically. A palaeosol separates, indeed, the loci 2‑001 and 2‑008 from the next similar event (locus 2‑007), situated between 55.58 m and 54.60 m (fig. 26); it is thicker as we get further down the slope (up to 80 cm). The new stone bed marks a re-activation of the erosion process after a long period of stability. It has not been possible to measure directly the duration of this abandonment period, for the palaeosol layer contained practically no fragments of organic materials (charcoal or bones)66. But we established that, when erosion starts anew, it affects layers that are slightly earlier in date (cf. dates Lyon-7621 and Lyon-7619), which had been exposed in the meantime67. This is attested also by the presence of some earlier sherds (e.g. LN I painted or black-topped), all severely eroded.

Fig. 21 – Northern (CC’) and eastern (FF’) profile of sector 2 (drawing: V. Anagnostopoulos, C. Germain-Vallée; “inking” R. Douaud).

Fig. 22 – Fragment of an S-profile bowl from sector 2, locus 2‑002, with graphite-painted decoration and remains of red paint applied after firing (“crust”) on the rim (drawing: R. Docsan).

Fig. 23 – Four-legged vessel with incised and red-crusted decoration (photo: Ph. Collet-EFA©; drawing: R. Docsan).

Fig. 24 – The stones of locus 2‑001, view to the east (photo P. Darcque).

Fig. 25 – Fragments of graphite-painted vessels from locus 2‑001 (drawings a, c: R. Docsan; b: S. Eliès).

Fig. 26 – The successive stone-beds (loci) 2‑007 and 2‑008 (photo P. Darcque).

38It is highly tempting to associate the re-activation of the erosion process with the reoccupation of the settlement in the Early Bronze Age. The arrival of new inhabitants would justify the extensive clearing of the area, which would have put an end to the pedogenesis. This date is also supported by the pottery found in the colluvia that are closer to locus 2‑007: apart from Neolithic, the only other diagnostic sherds are EBA (for example some “dipper” fragments), in very small proportions. Later sherds appear much higher in the deposits; their presence confirms that, once triggered, the erosion process has never really stopped in this part of the tell.

  • 68 Due to the excavation procedure (use of mechanical means down to a depth of ca 1.50 m, excavation i (...)
  • 69 Illustrated in Deshayes 1972, p. 202, bottom right; the provenance indicated is “square A, level 17 (...)
  • 70 Dolno Dryanovo (Todorova, chapter 10 in this volume, p. 179, fig. 10: 7 and 11: 7); Ilinden (Todoro (...)
  • 71 A fragment is also found in Sitagroi: Evans 1986, p. 402, pl. XCIV, top: 6 (phase III, provenance Z (...)
  • 72 Tatul (Leshtakov, Todorova and Petrova, chapter 11 in this volume, p. 197, fig. 7: 4, 6‑9); Ilinden (...)

39A closer look at the pottery assemblages from the interface between the last LN II deposits and the first deposits assigned to the EBA (i.e. the upper part of the earlier colluvia, including the palaeosol) provides some interesting information68. Several sherds are from the same type of bowl with a globular body and upright concave rim (sinuous profile), decorated with opposed groups of oblique incisions on the belly (which could actually belong to festoons) and a range of more or less well-done circular impressions at the rim’s base (fig. 27: a-c). Similar vessels are not found within the material of House 1, but a fragment was collected, as we saw, in the upper part of the destruction layer that could actually belong to the presumed “post-House 1” occupation level. They are also absent from Demoule’s typology, although they are represented within the material collected from the last LN II level in sector A2, square R2469. Parallels from a number of Final Chalcolithic sites in the Rhodopes70 seem to confirm that these kinds of vessels characterize the last stages of the Late Chalcolithic in this area71. Another singular fragment with parallels in the Rhodopes is a tiny rim with a flat prominence on it (fig. 27: d)72; it might belong to a rounded bowl with a beaded rim, unknown so far (at least bibliographically) from the sites of the Drama plain.

Fig. 27 – Bowl fragments from the interface between the last LN II and the first EBA deposits in sector 2.
a-c: with incised decoration on the belly; d: with flat prominence on the rim (drawings: R. Docsan).

Conclusions

40In light of recent research, the following things can be said about the Late Neolithic II-Early Bronze Age transition at Dikili Tash.

  1. The presence of a gap in the occupation on several parts of the tell suggests that this is a change which probably affected the whole settlement. The formation of a palaeosol at the foot of the tell, implying the cessation not only of building but also of all other activity in this area (cultivating, breeding), also points in the same direction.

  2. In spite of the definite and somehow abrupt character of this statement, the cessation of activities does not appear to have been an abrupt event, but rather the outcome of a longer process, probably spanning several generations. This is particularly clear in sector 2, where successive episodes of colluvia indicate the progressive disengagement from this area, between the third quarter and the end of the 5th millennium BC.

  3. The two sectors studied (2 and 6) illustrate the kind of contrasting evidence that can be collected in the different parts of a tell. Actually, the two are complementary: the absence of any visible stratigraphic hiatus in sector 6 is the echo of the accumulated pedogenized deposits in sector 2; the part that is missing from the former dissolved in what filled the latter; and it is probably the rehabilitation of the former that triggers a new collapse event in the latter.

    • 73 Similar signs (pitfalls created by the raindrops, water streaming, stagnated wares) are observable (...)

    The collapse of the Neolithic stonewall in sector 2, which marks the end of the LN II activities in this area, is accompanied by strong rainfall, but this does not necessarily reflect some particularly intense meteorological phenomena73. It could just be that the wall was no longer maintained as before, maybe because the inhabitants did not use it or need it anymore. The decline of the activities seems thus to come before the destructive events, and not the inverse.

    • 74 In fact, the existence of such an intermediary layer has been confirmed in the field by the 2013 ex (...)

    The last event visible through ordinary excavation (i.e. the house destruction layer in sector 6 and neighbouring sectors) is actually not the true final event in the life of the settlement. The 4200‑4000 cal BC episode has been clearly identified in sector 2 through the analysis of sediments, and confirmed by 14C dates. Its existence in the area of sector 6, although not proven, seems very probable on the basis of contextual data and 14C dates74.

  4. The rather discrete presence, in both sectors, of a few ceramic types possibly connected with this last LN II episode prolongs the existent pottery sequence and attests a number of interesting affinities with neighbouring sites, beyond those already known from previous periods (other sites in the Drama plain, Strymon valley). It would be inexact to speak about a “transitional” material, since no transition, properly speaking, is attested on the site itself. On the other hand, the fact that some of these types are found almost identically in the EBA I period, starting some 700‑1000 years later, could indicate that life went on somewhere near to the actual site of Dikili Tash, and that some of the ingredients of the new lifestyle had their roots in the late 5th millennium developments.

Notes

1 I wish to thank the other co-directors of the current research project, Pascal Darcque, Chaido Koukouli-Chrysanthaki, and Dimitra Malamidou, as well as René Treuil, director with Ch. Koukouli-Chrysanthaki of the previous programme on the site, for accepting that I sign alone this synthesis, and also for providing their useful comments. I also thank all the collaborators of the present program, especially Cécile Germain-Vallée, for her work on the stratigraphy and sediments analysis from sector 2 and for being so patient in explaining to me the subtle aspects of erosion and pedogenesis processes in this area (see also infra, p. 291 and n. 65).

2 Research conducted under the auspices of the Archaeological Society at Athens and the French School at Athens. See Koukouli-Chrysanthaki & Romiopoulou 1992; Treuil 1992 and 2004; Darcque et al. 2007; Koukouli-Chrysanthaki & Treuil 2008, Darcque & Tsirtsoni 2010, and also the website: www.dikili-tash.fr (with a full up-to-date bibliography).

3 See especially Koukouli-Chrysanthaki 1994, p. 125‑129; Koukouli-Chrysanthaki et al. 1996a, p. 689‑696; Darcque et al. 2007, p. 252‑254.

4 The corresponding layers were reached through a series of core-samplings, but have not yet been excavated; see Lespez et al. 2013.

5 Darcque et al. 2008, p. 533‑535; Darcque et al. 2009, p. 536‑539.

6 See also Darcque et al. 2011.

7 This designation should change in the future, since we know now that the occupation starts before the Late Neolithic I.

8 The presence of LN II at the basis of sector I/1961 has not been proven: Koukouli-Chrysanthaki & Romiopoulou 1992, p. 230. See also infra (n. 22).

9 This is due to the excavation strategy, which never proceeded by deep stratigraphic trenches (like for instance at neighbouring Sitagroi), but by independent squares at different parts of the tell that provided slightly “overlapping” sequences: see Deshayes 1968, p. 1062‑1064; Treuil 1992, p. 19‑20.

10 All measures are given in absolute altitudes.

11 The most impressive being the discovery of evidence for wine-making: Valamoti et al. 2007.

12 Treuil 1992, p. 19‑20.

13 This was indeed the basic assumption behind J. Deshayes’ decision to display his trenches in a “ladder-like” manner and to stop excavation when he reached the layer that made the junction with the next phase: Deshayes 1968, p. 1062‑1064. The idea was again defended by Koukouli-Chrysanthaki & Romiopoulou 1992, p. 230‑231, discussing the stratigraphic position of the LN II finds from Theocharis’ sector II/1967: “The discovery of an analogous destruction layer in the 1967 French sector [squares .. W29, X29, W30, X30] and in the level 16 of the square R24 at the top of the mound, confirms the existence of a uniform destruction layer in the settlement at the end of the Neolithic period. This is further confirmed by the discovery of the same destruction layer in the sectors of the new program [= V and VI]”. See also Demoule 1989, p. 688, and Demoule 2004, p. 64‑67, and table 1.1. In the latter, we read (p. 64): “…les niveaux du Néolithique Récent de Dikili Tash ne sont pas une accumulation désordonnée de remblais et de deblais, mais une sédimentation progressive et ordonnée [..] les couches y sont parfaitement horizontales, avec une succession de ‘niveaux’ tous les 0.25‑0.30 cm en moyenne”.

14 Séfériadès 1983; Demoule 2004, p. 98‑99, 102‑177, table 1.4, 3.25 and 4.15.

15 This pottery style actually seems to appear already in the middle stages of the sequence. One of the most emblematic and best-preserved specimens (no. inv. DT 150: Demoule 2004, pl. 18: 2, XII: 2, pl. E: 1) probably comes from a level assigned to horizon V (and not from horizon VII, as marked erroneously in the caption of colour pl. E: 1). Surprisingly, it is not included in Demoule’s discussion about the stratigraphic evolution of this pottery (ibid., p. 95, and table 3.13), where it is said however that the roots of this style could go back to earlier phases.

16 This type of vessel is also well-represented in the material from sector II/1967: Koukouli-Chrysanthaki & Romiopoulou 1992, pl. 42a (no. inv. Philippi Museum A532); shown also in colour in Demoule 2004, pl. C: 2.

17 Koukouli-Chrysanthaki & Romiopoulou 1992, p. 232‑233, expressed a different opinion when discussing the material from Theocharis’ sector II/1967, which is actually part of the future sector VI. They proposed synchronizing the destruction level discovered there (which would prove to be the destruction layer of houses 2 and 3 further investigated in 1989‑1996) with the last sub-phase of Sitagroi III (IIIc) and the end stages of the KGK VI culture (namely Sălcuţa IV). Based on the absence of black-on-red vessels among the floor deposits, they even risked the hypothesis that this level might be later than the last Chalcolithic layer at Kastri. Radiocarbon dates from several Sălcuţa IV sites and from Kastri itself, however (see in this volume chapter 2, p. 56 and chapter 18 p. 347‑348), clearly show today that this is not true.

18 In the two southernmost squares of sector B2 (X29 and X30), the levels 12 and 10, respectively, lie between 60.75‑61.10 m; the floor of level V/East/1 lies at 60.45 m, its summit towards 61 m. The argument of horizontality, although rejected by Treuil, was retained as we saw by Demoule (supra, n. 13).

19 One of the elements that seemed to need revision was the evolution of the black-on-red painted pottery. Indeed, “Galepsos style” pottery is already found in level V/East/1 together with “Dikili Tash style” fragments, whereas it was absent from the material collected in the houses of sector VI (Koukouli-Chrysanthaki et al. 1996a, p. 694). See also discussion in Darcque et al. 2011, p. 195‑196; and supra, n. 15.

20 Treuil 1983, p. 98‑99, actually proposes three subdivisions; see also Malamidou, forthcoming.

21 Koukouli-Chrysanthaki 1994, p. 124‑125.

22 Their precise nature is not known (occupation level in situ, colluvial deposits, disturbed layer?), for they were only seen at the bottom of a 1 x 2 m trench, but their date is clearly indicated by characteristic pottery finds (Theocharis & Romiopoulou 1961, p. 84‑87; Daux 1962, p. 917‑919, fig. 7‑8). Theocharis associated these pottery finds with the remains of a relatively large (0.75‑0.80 m) stone-built wall that appeared at approximately 55 m. Interestingly, one of the arguments for considering the wall as Neolithic was its solid construction, which contrasted, according to the excavators, “with the cheap buildings of the Bronze Age”. However, no other stone-built walls are known so far in the LN II period. See also Koukouli-Chrysanthaki & Romiopoulou 1992, p. 234.

23 Koukouli-Chrysanthaki & Romiopoulou 1992, p. 236‑245.

24 Ibid., p. 245‑247, mention the mixing with Neolithic sherds, but not with positive EBA I material.

25 The 14C dates from the first research program (sectors A2 and B2) have been published by Treuil 1992, p. 33‑36. Those of the second research program (sectors V and VI) are only partly discussed in a number of preliminary reports; see also Darcque et al. 2011, p. 198‑199, and fig. 13.

26 At first, J. Deshayes saw at this point the effects of a possible invasion from the lower Danube area (Deshayes 1972, p. 205), but this scenario has since been abandoned. The existence of a gap has been confirmed by Demoule’s pottery analysis (Demoule 2004, p. 149, 184), as well as by the study conducted by myself on the material from sector V. The missing stage would correspond to the phase III of Promachon-Topolnitsa and the “Early Eneolithic” at Strumsko: see Darcque et al. 2011, p. 194‑195.

27 Treuil 1992, p. 34. There could also be another explanation: that their stratigraphic position is not the one presumed. Indeed, both dates presumably refer to the level 16 (between 64.18‑64.30 m, the same observed also in R24, infra). But the excavation got considerably deeper here (63.35‑62.75 m: ibid., p. 16), so maybe the original position of the samples was lower. The 1974 excavation diary mentions that samples were taken from a thick black burnt deposit lying between 63.35‑63.60 m in square T24, and in another “peat-like” deposit at ca 63.35‑63.60 m in U24 (= the same?), i.e. almost 1 m below the LN II-BA I interface. Unfortunately, there is no description of these deposits in the final publication (Treuil 1992, p. 27).

28 Treuil 1992, p. 34‑35, suggested considering the seven dates together, and concluded that the end of the LN II period at Dikili Tash could be situated as early as 4500 cal BC. At that time, there was no 14C date yet from sector VI.

29 In two laboratories: Demokritos in Athens, under the supervision of Y. Maniatis, and Heidelberg, under the supervision of B. Kromer.

30 Actually the first comprehensive analysis was undertaken in 2010, as part of a regional synthesis integrating also the first results of the “Balkans 4000” project. It has been presented as a communication at the 38th International Symposium of Archaeometry at Tampa (Florida, USA); see Maniatis et al. 2014.

31 The dates referring to the EBA levels in sector A2 were all judged invalid: see discussion in Treuil 1992, p. 36. Actually, only one is truly aberrant (Ly-1061, 6480 ± 270 BP), the other two (Ly-1305 and Ly-1602) could still enter the EBA time span (late 4th-early 3rd mill. BC), but their stratigraphic order should be reversed.

32 TL-dates were made on samples from two different ovens, one in House 3 (two measurements) and one in House 4 (one measurement). Those from House 3 (BDX-5442 and BDX-5443) overlap at the time span 4380‑4290 BC, whereas the one from House 4 (BDX-5440) has a lower limit at 4110 BC: see Roque et al. 2002.

33 Tsountas 1908, p. 31‑49; Papaefthymiou-Papanthimou & Pilali-Papasteriou 1988, fig. 1, 7.

34 Darcque et al. 1990, p. 877‑878, fig. 1; Darcque et al. 1992, p. 715, fig. 1.

35 In addition to funding from the two main institutions, the research was funded by the French Ministry of Foreign Affairs, the Institute for Aegean Prehistory (INSTAP), the CNRS, and the Philippi municipality. Excavation continued in 2012 and 2013, but there is no mention of that work here.

36 A synthesis, focusing mainly on the LN II levels, is presented in Darcque et al. 2015.

37 This study, carried by Dr Tania Valamoti (University of Thessaloniki), received extra funding in 2010 and 2012 from the National Geographic Society. For a presentation of the first results, see Darcque et al. 2010; Valamoti 2015; Valamoti et al. 2015.

38 Actually the grains that provided the date Lyon-6013 were originally believed to belong to the fill of an EBA pit (6‑011), cutting the LN II destruction layer; see also infra. The other sample (DEM-2031) was taken from a concentration of acorns. Notice the very good agreement between the results of the two laboratories.

39 There has not been any anthracological analysis on these samples.

40 I express my gratitude to Philippe Lanos and Philippe Dufresne, developers of the program RenDateModel, precursor of the more sophisticated Chronomodel (© CNRS, UMR 5060‑iramat) [available on-line: www.chronomodel.fr], for authorising its use and for helping me out with the modelling of these data. The diagrams shown here are just a sample of a bigger modelling project on the absolute chronology of Dikili Tash, also integrating the results from TL dates.

41 This kind of vessel (small rounded cups with raised handle, also called “scoops”, or “puisettes” in French) are well-attested in the EBA I levels at Dikili Tash itself (Séfériadès 1983, fig. 43) and in neighbouring Sitagroi, phase IV (Sherratt 1986, fig. 13.4: 6‑12).

42 Supra, n. 38.

43 A similar fragment from Sitagroi is illustrated in Renfrew et al. 1986, pl. CI: 7 (“sieve fragment”). Assigned originally to phase V (indistinct deposits from the upper part of sounding ZE: ibid., p. 210), it was later attributed to phase IV (Elster & Nikolaidou 2003, p. 424, fig. 11.3). “Incense burners”, but in a more developed version with handles, are also found at a number of sites in Central and Western Macedonia (Agios Mamas, Archontiko, Mandalo), in layers of the advanced EBA and the transition to the MBA: Pilali-Papasteriou & Papaefthymiou-Papanthimou 2002, p. 146, fig. 12.

44 The rest of the vase was obtained by refitting. Similar conical bowls exist at Sitagroi phase IV: Sherratt 1986, fig. 13.7: 1 (without handles).

45 Koukouli-Chrysanthaki et al. 1996a, p. 703, fig. 17; Darcque et al. 2007, fig. 6‑7.

46 Koukouli-Chrysanthaki et al. 1996a, p. 693; Darcque et al. 2011, p. 196. The recent technological study of the ceramic vessels from House 4 (César Watroba, unpublished Master dissertation, University of Paris I, 2012) suggests, however, some differences in the vessels’ possible function between the rooms A and B. The possibility of some kind of differentiation among the spaces should be further examined in the light of other studies.

47 Demoule 2004, pl. 98: 11167 (horizon VIII).

48 Ibid., pl. XIV: 6 (DT 23, horizon VII).

49 Supra, n. 15.

50 From sector II/1967, i.e. the present sector 6, around the Houses 2‑3. Many thanks to Ch. Koukouli-Chrysanthaki for confirming the provenance of this vessel and allowing me to show it here.

51 J. Deshayes is clearly describing it in one of the preliminary reports, as a variant of the black-on-red pottery, and even notices that it is found in the latest levels: Deshayes 1968, p. 1068 (“il arrive, du moins dans les niveaux les plus récents, qu’un engobe jaunâtre s’interpose entre l’argile et les motifs peints, qui sont alors de couleur brunâtre”).

52 Demoule 2004, p. 161. See also Boyadzhiev et al. 1993, p. 80.

53 For instance, at Karanovo itself or Yunatsite: see chapters 7 and 9 in this volume.

54 As was presumed by J.‑P. Demoule, see supra, p. 274 and n. 14.

55 Of course, other explanations are possible: presence of a shelf or upper-storey, or the erosion of levels situated higher on the tell. The hypothesis will be tested by further study of the material collected in the interface between the house destruction layer and the first EBA level.

56 Demoule 2004, pl. C: 3‑4, D: 4.

57 Todorova, chapter 10 in this volume, p. 181, fig. 12 (sp. 12: 3). None of these vessels can be taken as an exact parallel (notice the use of the triangular impressions technique, which is absent from the specimen from Dikili Tash), but the overall conception looks much alike. Further fragments from the site of Ilinden at ca 50 km from Dikili Tash might also be close to the present vessel (Todorova 2014, p. 285, fig. 10:14).

58 I am grateful to N. Todorova for drawing my attention to these fragments, whose importance might have gone otherwise unnoticed or misunderstood. I also wish to thank her for sharing with me many unpublished data from her own investigations in the Rhodopes.

59 Koukouli-Chrysanthaki & Romiopoulou 1992, p. 233, and pl. 42b.

60 Evans 1986, pl. XCIV: 5‑6, 10‑12, and colour pl. D: 8; Koukouli-Chrysanthaki 1972, p. 525‑526 and fig. 3.

61 Koukouli-Chrysanthaki & Papadopoulos, chapter 18 in this volume, p. 347‑348.

62 Koukouli-Chrysanthaki & Romiopoulou 1992, p. 233, with references to Séfériadès 1983, fig. 44‑46, and Sherratt 1986, pl. XCVI: 1‑11.

63 The results have been discussed also with some detail in Darcque et al. 2014; Darcque et al. 2015.

64 Lespez et al. 2013, p. 36‑37.

65 Both the fine macroscopic observation of layers in the field and the micromorphological analysis have been carried by C. Germain-Vallée, archaeologist at the District Council of Calvados (France).

66 We did not attempt to date the total organic matter in the layer, fearing that it would not possible be to connect the eventual result with a particular archaeological event.

67 A third sample from the same area gave a date at the early 6th millennium BC (table 3: Lyon-7620). Obviously this is a result of reworking, but it is not possible to say whether the charcoal was already in a secondary position before its deposition here, or if some mixing took place afterwards.

68 Due to the excavation procedure (use of mechanical means down to a depth of ca 1.50 m, excavation in more or less horizontal units), this interface is preserved or can be reconstituted in only a few stratigraphic units.

69 Illustrated in Deshayes 1972, p. 202, bottom right; the provenance indicated is “square A, level 17”, later becoming R24, level 16 (Treuil 1992, p. 27). This level is not devoid of mixing with EBA I material, but no such fragments are recorded from pure EBA I contexts. The combination of a range of impressions with incised lines on the belly is found on several EBA vessel types associated with the Baden and Coţofeni “cultures” (see among others Roman 1976; Roman & Nemeti 1978; and recently Horváth 2012, fig. 14; 19: 5; 32: 22; 33; 36), but this is not exactly the same decorative technique nor the same shape.

70 Dolno Dryanovo (Todorova, chapter 10 in this volume, p. 179, fig. 10: 7 and 11: 7); Ilinden (Todorova 2014, p. 285, fig. 10: 10‑16).

71 A fragment is also found in Sitagroi: Evans 1986, p. 402, pl. XCIV, top: 6 (phase III, provenance ZA 38). Evans arranges it in the “grooved” fabrics, which are said to be found in small quantities in the upper levels of phase III (and more common in phase IV).

72 Tatul (Leshtakov, Todorova and Petrova, chapter 11 in this volume, p. 197, fig. 7: 4, 6‑9); Ilinden (Todorova 2014, p. 285, fig. 10: 3, 5).

73 Similar signs (pitfalls created by the raindrops, water streaming, stagnated wares) are observable indeed in the thin sections from all the colluvial layers that precede this event. Seasonal storms are very frequent phenomena in the Mediterranean.

74 In fact, the existence of such an intermediary layer has been confirmed in the field by the 2013 excavation (see http://chronique.efa.gr/index.php/fiches/voir/4078/), but these results are not at all discussed here.

Table des illustrations

Légende Fig. 1 – Aerial view of the Dikili Tash tell, in the early 1990s (photo: Ch. Tsouroukidis).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-1.jpg
Fichier image/jpeg, 280k
Légende Fig. 2 – Topographical plan of the works conducted until 2010; are shown the areas where LN II and/or EBA deposits have been investigated.
URL http://books.openedition.org/momeditions/docannexe/image/525/img-2.jpg
Fichier image/jpeg, 2,3M
Légende Fig. 3 – Diagram OxCal with the available 14C dates for LN II levels until 2008 (calibrated at 2σ).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-3.jpg
Fichier image/jpeg, 532k
Légende Fig. 4 – Schematic plan of houses in sector 6.
URL http://books.openedition.org/momeditions/docannexe/image/525/img-4.jpg
Fichier image/jpeg, 80k
Légende Table 1 – 14C dates from samples collected in sector 6 during the 2008 work.
URL http://books.openedition.org/momeditions/docannexe/image/525/img-5.jpg
Fichier image/jpeg, 124k
Légende Table 2 – All available 14C dates from House 1 up to 2012.
URL http://books.openedition.org/momeditions/docannexe/image/525/img-6.jpg
Fichier image/jpeg, 176k
Légende Fig. 5 – General view of House 1, to the northeast, at the end of the 2010 campaign (photo: P. Darcque).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-7.jpg
Fichier image/jpeg, 120k
Légende Fig. 6 – Grindstone and concentration of bitter-vetch to the south of oven 6‑015 (photo: P. Darcque).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-8.jpg
Fichier image/jpeg, 84k
Légende Fig. 7 – Vessels in situ in front of oven 6‑015; the surrounding sediment contained great quantities of charred grape pips (photo: P. Darcque).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-9.jpg
Fichier image/jpeg, 112k
Légende Fig. 8 – Group of vessels next to a fallen wall fragment to the east of the room (photo: P. Darcque).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-10.jpg
Fichier image/jpeg, 88k
Légende Fig. 9 – Modelled diagram with the 14C dates from House 1, using the program RenDateModel (© CNRS-IRAMAT). Two versions are proposed: in version (a) only two “events” are distinguished, one with all charcoals (presumably dating the construction and use of the house) and one with seeds (dating the destruction); in version (b) we introduce the possibility of a third “event” (reoccupation) represented by the two late charcoals.
URL http://books.openedition.org/momeditions/docannexe/image/525/img-11.jpg
Fichier image/jpeg, 492k
Légende Fig. 10 – Part of the EBA remains on top of the LN II House 1 (zenithal photo: Chr. Gaston).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-12.jpg
Fichier image/jpeg, 120k
Légende Fig. 11 – EBA I “dippers” (photo: F. Bourguignon; drawing: R. Douaud).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-13.jpg
Fichier image/jpeg, 40k
Légende Fig. 12 – EBA I “incense burner” (photo: F. Bourguignon; drawing: S. Monnet).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-14.jpg
Fichier image/jpeg, 76k
Légende Fig. 13 – EBA I conical bowl (photo: Ph. Collet-EFA©; drawing: R. Docsan).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-15.jpg
Fichier image/jpeg, 84k
Légende Fig. 14 – Some of the vessels found in House 1 (photos: Ph. Collet-EFA©; drawings: R. Docsan, R. Douaud).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-16.jpg
Fichier image/jpeg, 84k
Légende Fig. 15 – More vessels from House 1 (drawings: R. Douaud).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-17.jpg
Fichier image/jpeg, 64k
Légende Fig. 16 – Vessels from the other houses of sector 6. a: brown on-cream painted pot from the area of Houses 2‑3, h. 10 cm; b: globular pot with ranges of impressions, h. 15 cm; c: graphite-painted askos from House 4, h. 12.5 cm; d: incised jar from House 3, h. 74 cm (photos: a, Ph. Collet-EFA©; b and d, St. Stournaras; drawing of c: I. Vajsov).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-18.jpg
Fichier image/jpeg, 112k
Légende Fig. 17 – Two-handled globular bowl with incised and red-crusted decoration (photo: P. Darcque; drawing: R. Douaud).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-19.jpg
Fichier image/jpeg, 24k
Légende Fig. 18 – Graphite-painted “Kritsana bowl” (photo: P. Darcque; drawing: R. Douaud).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-20.jpg
Fichier image/jpeg, 68k
Légende Fig. 19 – Fragments of a gray-polished globular bowl with beaded rim (photo: Ph. Collet-EFA©; drawing: R. Douaud).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-21.jpg
Fichier image/jpeg, 40k
Légende Fig. 20 – Fragment of a two-handled bowl with incised decoration (photo: Ph. Collet-EFA©; drawing: R. Docsan).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-22.jpg
Fichier image/jpeg, 36k
Légende Table 3 – 14C dates from samples collected in Sector 2 (including cores C5 and C6).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-23.jpg
Fichier image/jpeg, 292k
Légende Fig. 21 – Northern (CC’) and eastern (FF’) profile of sector 2 (drawing: V. Anagnostopoulos, C. Germain-Vallée; “inking” R. Douaud).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-24.jpg
Fichier image/jpeg, 708k
Légende Fig. 22 – Fragment of an S-profile bowl from sector 2, locus 2‑002, with graphite-painted decoration and remains of red paint applied after firing (“crust”) on the rim (drawing: R. Docsan).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-25.jpg
Fichier image/jpeg, 40k
Légende Fig. 23 – Four-legged vessel with incised and red-crusted decoration (photo: Ph. Collet-EFA©; drawing: R. Docsan).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-26.jpg
Fichier image/jpeg, 64k
Légende Fig. 24 – The stones of locus 2‑001, view to the east (photo P. Darcque).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-27.jpg
Fichier image/jpeg, 80k
Légende Fig. 25 – Fragments of graphite-painted vessels from locus 2‑001 (drawings a, c: R. Docsan; b: S. Eliès).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-28.jpg
Fichier image/jpeg, 72k
Légende Fig. 26 – The successive stone-beds (loci) 2‑007 and 2‑008 (photo P. Darcque).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-29.jpg
Fichier image/jpeg, 152k
Légende Fig. 27 – Bowl fragments from the interface between the last LN II and the first EBA deposits in sector 2. a-c: with incised decoration on the belly; d: with flat prominence on the rim (drawings: R. Docsan).
URL http://books.openedition.org/momeditions/docannexe/image/525/img-30.jpg
Fichier image/jpeg, 56k

CC-BY-NC-ND-4.0

Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Cette publication numérique est issue d’un traitement automatique par reconnaissance optique de caractères.
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search