Introduction3

1The earliest evidence of human presence in Thasos (fig. 1) comes from the area of Limenaria, in the ochre mines of “Tzines”, which, according to radiocarbon dating, were exploited during the last phase of the Paleolithic period, around 20000 BC4. However, the first indications of permanent habitation do not appear until the middle of the 6^th millennium BC: the coastal settlement of Limenaria is so far the oldest permanent settlement on the island, with occupation attested at least since the end of the Middle Neolithic5.

2The Neolithic settlements of Thasos are located either on low hills near the sea, such as the settlements in Limenaria6, Agios Antonios in Potos7 and in the ancient city of Thasos8, or in fortified sites at some distance from the coast, as is the case of Kastri in the Theologos area9.

3Human presence has also been detected in caves beginning in the Neolithic period, in the north (“Drakotrypa” near the village of Panagia)10 and the south of the island (“Atspas” near Skala Maries11, and “Trypes” near Maries12). The limited scope of the excavations, however, does not allow any secure conclusions about the cultural character or the chronological range of the cave use. According to present knowledge, the beginning of cave use dates to within the Late Neolithic period and continues through the Bronze Age, the Early Iron Age and into historical times13.

From the Middle to the Final Neolithic Period (middle 6^th-early 4^th millennium BC)

The prehistoric settlement of Limenaria

4The earliest permanent settlement of prehistoric Thasos is located in the southern part of the island, in the modern village of Limenaria, very close to the coast. Archaeological layers from the end of the Middle Neolithic were excavated in the southwestern part of the site during a series of rescue excavations in the years 1993, 1994, 1997 and 2000. On the whole, the Neolithic layers cover 350 m², and are 1 to 2.5 m thick.

5The Lioudas plot, where a large underground chamber and some open hearths were excavated (fig. 3), provided the earliest radiocarbon date so far: according to it, the beginning of the Limenaria settlement dates from the Middle Neolithic, between 5780‑5510 cal BC14.

6The archaeological layers excavated in the Sidiropoulos plot date from the same period. Excavation here attested the presence of a habitation area at the periphery of the settlement, comprising many garbage pits15. A small part of the settlement’s interior has also been located; it comprised remains of open-air hearths and ovens, storage pits and traces of light post-framed structures. Large “benches”, made of a thick layer of clay and framed by rows of stones, constitute one of the basic architectural elements. Such benches are present in the open space of the settlement, where they certainly catered to multiple household activities.

7The excavation of a 150 m² area in the Konstantinidis plot in 2000 was the opportunity to investigate part of the settlement’s interior and reveal further elements of its intra-community organization16. Parts of three successive buildings with identical orientation (NE-SW) were revealed in the eastern zone of the plot, in trenches Φ, Χ and Ψ. This seems to suggest some stability in occupation, and perhaps even the existence of established property rules concerning the use of space. The post-framed buildings are more than 8 m long, while their width remains unknown so far (fig. 2). Inside the houses there are hearths, ovens, storage pits coated with thick layers of clay and rectangular clay constructions also used for storage purposes. A well (diameter 2.30 m) and an astonishing number of storage pits cut into the conglomerate (fig. 3) are among the characteristic features of the earliest layers. Radiocarbon dating of samples from trench Φ provided dates of the Middle Neolithic, between 5540‑5470 cal BC17. The layers above these in the same plot were dated by radiocarbon between 5320‑5210 cal BC18. They appear to be a cultural continuation of the previous phases and can be assigned to the Late Neolithic I period, based on the presence of characteristic painted and black-topped pottery (fig. 4)19.

8That the occupation of the settlement continued until the end of the 5^th millennium BC is confirmed by the presence of Late Neolithic II pottery, but out of context. It comprises decorated wares common in Eastern Macedonia, such as graphite-painted and black-on-red. Additional evidence comes from a radiocarbon date between 4330‑4070 cal BC20.

9In the Markoulis plot, deposits of the Early Bronze Age were excavated in 1995 and 1996. Final Neolithic Pottery was found as well, though not in secure habitation layers (fig. 5). This pottery is characterized by the use of raw (crusted) pigments and coatings on a brown surface, combined with the limited, “degenerated”, use of graphite and minimal use of impressed motifs (fig. 6, 7)21. Some fragments of litharge found on the same spot are the earliest evidence of lead processing for the extraction of silver in the Northern Aegean area22. The single radiocarbon date from this disturbed deposit dates the presence of this pottery tradition in Limenaria between 3970‑3790 cal BC23. This is one of the very few Final Neolithic dates available from Eastern Macedonia, to which were recently added the two radiocarbon dates from Kastri Theologos and Agios Antonios in Potos (infra).

The prehistoric settlement of Kastri

10The prehistoric settlement of Kastri is situated in the mountainous hinterland of the island, between the villages of Theologos and Potos (fig. 8). Systematic excavations from 1969 to 1980 focused on the Late Bronze and Early Iron Age settlement and on the extended cemeteries of the same periods24. The main excavation area (Sector I, fig. 9) was extended to the northern part of the flat top of the hill, to the south of the huge stone-heap that covered the buildings of the last habitation phase (Late Bronze/Early Iron Age) and also Neolithic remains.

11Sector I extended over 612 m². The archaeological deposits, which contain all occupation phases, are 1.5‑1.75 m thick. The presence of rock graffiti in the southeastern part of the flat top25, however, confirms that the natural rock was visible at that spot during prehistoric times.

12The Neolithic phase of the settlement has only been investigated occasionally and over a very small surface. In some cases, the Neolithic occupation layers appeared immediately under the surface layer, as in trench 326, or at a very shallow depth under the occupation layers of the Late Bronze Age, as in trench 227. Deep soundings inside some of the squares, as well as two small stratigraphic trenches within28 and outside29 the excavation grid, followed the sequence of the Neolithic layers down to the natural rock (phases Kastri Ia and Ib), below the occupation layers of the Early Iron Age (Kastri III) and the Late Bronze Age (Kastri II).

13The earliest occupation layer (Kastri Ia) was located directly on the natural rock. Very few architectural elements of this phase remained in situ. In a small deep sounding inside square 1, on a floor with a stone substructure, part of a hearth and groups of upright stones that supported the bases of wooden posts came to light30. A similar floor, with a stone substructure and the socket of a wooden post supported by stones, was excavated in the stratigraphic trench in 1975 also within the first occupation layer, just above the natural soil31. This first occupation layer was dated to the Late Neolithic I, on the basis of pottery, among which was a black-topped cup found in situ in square 1 (fig. 10: a, b). The decorated wares collected from this archaeological layer32 or dispersed among later floors33, also include painted pottery of the “Akropotamos” type (fig. 11), and pottery with painted decoration of large reddish-brown bands on a light background. These have parallels in the LN I settlements of Eastern Macedonia, such as Dikili Tash (phase I)34, Sitagroi (phase II)35 and Promachon-Topolnitsa (habitation period 2, building phases IIIA and B)36.

14The Late Neolithic II layers (phase Kastri Ib) have been excavated on a larger scale. Once again the stratified architectural remains consist of fragmentary successive floors over sparse stone substructures, with partially preserved hearths and groups of upright stones that supported wooden posts. It has been proposed that some remains of stone buildings and the earliest phase of a stone enclosure located on the most accessible, northwestern edge of the Kastri peak, could also date to the same period (fig. 12)37.

15Many types of decorated pottery occurred in this phase. Types that are widespread in Eastern Macedonia are found in small proportion such as an LN II black-on-red pottery (fig. 13) and the polychrome (actually trichrome) pottery (fig. 14) of the so-called “Dimitra type”38. Graphite-painted pottery appears in much larger proportions (fig. 15). The most common decorative patterns are groups of many parallel thin lines, different from the Dikili Tash and Sitagroi motifs, which usually consist of groups of only two or three parallel lines39. It is also worth noting the presence of patterns consisting of large graphite lines, often combined with an added (crusted) colour (fig. 16). Another widespread type of pottery is decorated with crusted red, yellow or white colour, used either alone, with motifs similar to those of the graphite (fig. 17), or in combination with incised decoration (fig. 18); the latter is also found alone (fig. 19). Similar wares are known from other sites in Eastern Macedonia and the Eastern Balkans40. Pottery with grooved or channelled decoration also occurs in substantial percentages, sometimes combined with graphite or crusted paint (fig. 20). Several types of plastic (impressed, barbotine) and relief decoration are present as well (fig. 21).

16These types of decorated pottery generally belong to the cultural horizon of the Late Neolithic II period in Eastern Macedonia, as represented by Dikili Tash phase II41 or Sitagroi phase III42, but also in other settlements of the Balkan mainland from the cultural horizons of Karanovo VI, Kodzhadermen-Gumelniţa A-B, Sălcuţa-Krivodol II‑IV43.

17A closed find from excavation square 2 is characteristic of the pottery types of phase Kastri Ib. It comprises three vessels, found next to sherds from a large storage jar and a slate slab that is probably the bottom of a hearth or an oven44. One is a closed vessel with a narrow neck, flat base and horizontal handle-lugs on the belly, with linear decoration of off-white graphite paint on a brown background. The neck is surrounded by groups of parallel lines and similar groups of lines inscribed into contiguous triangular frames decorate the body (fig. 22). The second is a large fragment of a closed vessel of similar shape with graphite decoration on black burnished background; groups of parallel lines surround the neck and free spirals decorate the body of the vessel. The third vessel is an unslipped one-handled cup.

18The chronological duration and internal evolution of the Late Neolithic II period at Kastri have not been precisely determined yet. However, even though the quantitative and statistical analysis of the stratified decorated pottery have not been completed, it is evident that some of the pottery types mentioned above go beyond both the upper and the lower chronological limits known so far from Dikili Tash II and Sitagroi III. For example, the presence of the polychrome “Dimitra type” pottery points to a chronological horizon earlier than Dikili Tash IIA and Sitagroi IIIA. Excavations and radiocarbon dates have identified this horizon in Promachon-Topolnitsa phase IIIA45, as well as in the earliest phase of the Neolithic settlement at Agio Pnevma (Serres district)46, and in contemporary phases from Neolithic settlements in the middle Struma valley47. Concerning the lower chronological limits of the destruction and abandonment layers of the Kastri settlement, a date later than Sitagroi IIIC48 and Dikili Tash IIC49 has been proposed, again on the basis of pottery.

19The first series of radiocarbon dates from Kastri were conducted in 2010, within the context of the “Balkans 4000” project that aimed to establish the absolute chronology of the end of the Neolithic and the transition to the Early Bronze Age in the Northern Aegean. The bone samples chosen came from the floors of the later occupation layers of the Neolithic settlement, in squares 1, 2, 3 and 12, which, according to the archaeological data, belong to the end of LN II.

20The earliest dates were provided by the bone samples D47a and D27, and fall in the second quarter of the 5^th millennium BC (table 1)50.

21These samples, which obviously belong to the same chronological horizon, come from layers above the earliest Neolithic layer, and from floors that are undoubtedly earlier than the destruction and abandonment layer of the Neolithic settlement. However, their stratigraphic context is not completely secure, as these are stone-paved floors, into which bone fragments from earlier phases could easily have penetrated.

22No radiocarbon dates are available as yet from the earliest occupation layer in the settlement (phase Ia), which, as we have already mentioned, was dated by pottery in the latest stage of Late Neolithic I. In the absence of a full series of radiocarbon dates that would follow closely the sequence of habitation layers, it remains unsettled whether the chronological horizon of the second quarter of the 5^th millennium BC should be attributed to the end of Kastri Ia or to the beginning of Kastri Ib. Consequently, the dating problem of the earliest LN II phases and of the latest LN I phases in Kastri, and generally in Thasos, remains open.

23The archaeological data and radiocarbon dates more securely define the cultural and chronological horizon of the end of the Neolithic settlement. Radiocarbon dating of two bone samples, D12 and D11, from the last destruction layer in square 3, provided absolute dates between 4257 and 3987 BC. A third sample (D26), from the last Neolithic layer of square 1, yielded an even later date, definitely from the 4^th millennium BC (table 2).

24These results support the suggestion that the Neolithic period in Thasos should be extended to include the beginning of the 4^th millennium BC51. The same has been suggested for Eastern Macedonia52 and the Balkan hinterland53, in settlements of the so-called Chalcolithic period assigned to the cultural groups Karanovo VI, Varna and Krivodol-Sălcuţa-Bubanj Hum54, with the closest chronological and cultural parallels found among the latter (Krivodol II, Liga 3, Galatin I)55.

25Occupation in Kastri is interrupted after this last Neolithic phase. There is no evidence of occupation during the Early Bronze Age. The site is settled again in the Late Bronze Age and seems to have flourished particularly during the Early Iron Age56.

Agios Antonios Potos

26The existence of a Bronze Age settlement on the hill of Agios Antonios, near the village of Potos in Southern Thasos, has been known since the beginning of the 1970s57. The 2009‑2010 excavations58 added a new Early Bronze Age site to those already known on the island and established that the first occupation of the site took place around 4000 BC, during the Final Neolithic.

27On the edge of the eastern slope of the hill, a stratigraphic trench was opened (trench ΓΚ) in order to investigate the existence of phases prior to the Bronze Age. A layer of red soil containing a few artefacts that revealed the presence of a Final Neolithic phase, probably the earliest occupation on the Agios Antonios hill. Two bone samples from successive excavation units provided absolute dates, the first between 3927‑3644 cal BC (Lyon-7650/SacA-22611) and the second between 3888‑3662 cal BC (Lyon-7911/SacA-23514)59. Such dates are rare for Northern Greece, and also for the entire Balkan Peninsula60. They provide Thasos with two more dates from the first half of the 4^th millennium BC, along with the aforementioned date from Limenaria (supra, p. 342). The pottery found in the same context as the samples presents, as expected, parallels to types of the Final Neolithic from Limenaria and Kastri: carinated bowls decorated with simple graphite patterns, crusted paints and coatings (fig. 23)61.

The Early Bronze Age I (2^nd half of the 4^th millennium BC)

28At the beginning of the 4^th millennium BC, highland settlements, like the one at Kastri, are abandoned. Occupation is concentrated along the coastline, either in new locations, as in Agios Ioannis62, or in sites occupied already during the Neolithic, as at Limenaria63 and Agios Antonios Potos64. A settlement must also have existed at Skala Sotiros since the beginning of the EBA. Indeed, it seems that the anthropomorphic stelae used as building material in the fortified settlement of the advanced Early Bronze Age II (infra, p. 353) would have come from there65.

Agios Ioannis

29On the bay of Agios Ioannis, in Southeast Thasos, a small settlement was excavated between 1998 and 2005, which belonged to the transitional phase from the Final Neolithic to the Early Bronze Age66. The architectural remains comprise mainly clay structures, hearths and ovens that are densely concentrated over an area of about 200 m². There were also stone benches, storage pits (often very large), and circular gneiss slabs with a diameter of up to 1.50 m, used during the preparation of food or for drying fruit and fish. In the northwestern part of the excavated area, circular low walls with one or two rows of stone came to light. We assume that they belong to two small huts of elliptical shape. At a distance of about 30 m north of the excavation trenches, the research conducted by the Geophysics Department of the University of Caen pointed out the possible existence of a refuse ditch, but this has not been confirmed so far by excavation67.

30The absolute dates from Agios Ioannis of Theologos range between 3700 and 3000 cal BC68. Nevertheless, the average calibrated date of the samples falls in the years between 3500 and 3140 cal BC, which seem to correspond more convincingly to the period of occupation of the site. Furthermore, if we consider that animal bones represent the dated events more securely – as they are not affected by aging phenomena, as is the case for wood, and their life ranges are narrower – the result is an average date between 3360 and 3100 BC. We therefore have here (as in the Lemonidis plot in Limenaria, infra) the earliest Early Bronze Age: the EBA I (fig. 24). The presence of this same phase is also likely certain in the part of the settlement excavated outside the enclosure of Skala Sotiros (infra, p. 353) but this has not been confirmed by absolute dates.

Limenaria

31In Limenaria there are no radiocarbon dates between 3800 and 3350 BC. On the other hand, the dating of two samples from layers containing EBA I together with typical EBA II incised pottery (fig. 25) in the Lemonidis plot, excavated in 200269, provided dates between 3350 and 3100 BC70. The excavated area in this plot does not exceed 40 m². Two parallel stone walls have come to light, possibly from a complex building with an external staircase (fig. 26). The abundance of spindle whorls found during the excavation confirms the same impressive intensification of the spinning industry that has already been observed in Skala Sotiros and in Agios Antonios Potos71. A clay crucible proves that copper processing took place in the settlement72.

Agios Antonios Potos

32As already mentioned, the excavation of the settlement in 2009‑201073 added a new Early Bronze Age site to those already known in Thasos. The great majority of the deposits belong to the advanced phases of this period. However, on the eastern slope, there have been indications of earlier occupation phases, the most ancient of which is the Final Neolithic attested in trench ΓΚ (supra). Further to the west, in trench ΓΘ, some house floors were investigated under the EBA II layer: they belong to the period between the Final Neolithic and the advanced EBA, thus proving that occupation here continued during the EBA I period. We so far do not have any absolute dates from this layer, but the pottery matches the finds from Agios Ioannis and Limenaria, and proves that the phase corresponding to the second half of the 4^th millennium BC exists here as well.

The Early Bronze Age II‑III (3^rd millennium BC)

33More recent phases of the Early Bronze Age in Thasos (EBA II and EBA III), dating from the 3^rd millennium BC, have been attested at Limenaria74, Agios Antonios Potos75 and Skala Sotiros76. They correspond more or less to the Early Helladic II and III periods (EH II‑III) in Southern Greece, and to the Early Cycladic II and III (EC II‑III) in the Central Aegean. Evidence of occupation during the same period is also found in the caves “Drakotrypa”, “Atspas” near Skala Maries, and “Trypes” in Maries (supra, p. 339).

Limenaria

34Architectural remains from the advanced phases of the Early Bronze Age came to light on top of the natural plateau on which the settlement developed. They were investigated in 1995 and 1996 in the Markoulis plot, over an area of 70 m². The deposits were approximately 1 m thick77.

35Two successive building phases have been excavated, corresponding to the EBA II and EBA III. The EBA II habitation phase was located partly on natural soil and partly on soil fill added to smooth the irregularities of the natural ground on top of the hill. A stone wall defines the northern limit of what is probably an open space. The floor is made of pebbles, shells and compact earth. The excavation revealed a stone-built hearth on top of it, a refuse pit and the remains of a clay bench, next to which two vessels were found in situ.

36A new fill covered this first occupation level. The next building phase, belonging to the EBA III, was installed on top of it. A curved cobblestone wall defines the western limit of a stone-paved space; rows of oblong stones are vertically embedded in the pavement. Neither the purpose of this construction nor its functional or symbolic role has been clarified yet. An oval stone with shallow cavities was also found in the western part of this pavement. It belongs to the type usually called “kernos” or “offering table” and is common in habitation and grave complexes of the Minoan world78. The only date from this phase falls in the Early Bronze Age III, between 2290‑1920 cal BC79.

Agios Antonios Potos

37This settlement reached its peak, as mentioned above, at the beginning of the 3^rd millennium BC, when there was a general tendency of the populations to move toward the sea, presumably dictated by the increased importance of metal and textile-based sea trade80.

38At the southeastern limit of the excavated area, in trench ΓΘ, destruction layers of the advanced Early Bronze Age (EBA II-ΙΙΙ) have been investigated. They contained large quantities of big storage vessels and many objects relevant to the processing of grains, mainly grinders and mortars. Two bone samples from this layer dated to the years between 2800‑2500 cal BC (DEM-2133 and DEM-2169)81. Pottery that co-occurred with the samples comprises characteristic types of the first half of the 3^rd millennium BC, mostly brown or black-coloured bowls (fig. 27), jugs and two-handled cups (depata amphikypella).

39The architectural remains of the EBA II‑III settlement form a group of seven buildings with stone foundations and brick superstructures (fig. 28). Hearths, ovens, benches and refuse pits have been found in the interior of the buildings as well as in their courtyards. Most buildings were one-roomed with rectangular floor plan, and only one is apsidal. In addition to the EBA II‑III habitation remains on the eastern slope, fragmentary remains of three houses from the Middle and Late Bronze Age have been found in the southwestern and the northern part of the excavated area. They provided two radiocarbon dates: the sample from the northern part dates between 1735‑1530 cal BC, while the one from the southwest between 1926‑1751 cal BC. The latter probably pinpoints the beginning of the MBA in the site82.

Skala Sotiros

40In the coastal village of Skala Sotiros the excavations have located an Early Bronze Age settlement, part of which is surrounded by a stone enclosure with monumental stone buildings in its interior, dating to the 3^rd millennium BC83.

The building complex with the stone enclosure

41The objective of the first excavation program84 was to investigate the stone enclosure discovered on the Profitis Ilias hill under the village church. The excavation focused mainly on the southwestern part of the hill and revealed the western branch of the enclosure, part of the southern branch, and the remains of the monumental stone buildings inside it. The buildings and the enclosure show at least three successive phases of repairs or modifications (fig. 29).

42The 2001 excavation85 traced the course of the northern branch of the enclosure to a length of 12 m, and revealed that the area surrounded by the enclosure was smaller than originally thought. Inside this area, where repairs had also taken place, a corridor with stone-paved and clay-plastered floors was located (fig. 30), similar to those unearthed inside the western branch of the enclosure over the course of the 1986‑1992 excavations.

43The earliest phase of the enclosure, detected in the western and southern branches, incorporated several entire or fragmented anthropomorphic stelae (“menhirs”) within the construction (fig. 31)86. They clearly indicate the existence of a building phase (Skala Sotiros I) prior to the construction of the enclosure, which could be dated to the EBA I, or maybe even to the Final Neolithic. The question of the provenance of these stelae remains open, despite some clues from the excavation suggesting that they could have stood originally in the area occupied later by the stone enclosure87.

44The earliest building phase of the enclosure itself (Skala Sotiros IIa), and of the stone buildings inside it, was dated by radiocarbon to a little before the middle of the 3^rd millennium BC88. During the next building phase (Skala Sotiros IIb) the gate at the south end of the western branch was closed off, and replaced by a gate at the west end of the southern branch. During the last building phase (Skala Sotiros IIc), new buildings adjoined those from the previous phase, but their floors, as well as the level of the narrow streets between them, were raised. During this phase a bastion was added to the western branch of the enclosure. Pottery from phase II is black-coloured, brown-black and sometimes red, with burnished or smoothed surfaces, either plain or with incised decoration.

45According to the radiocarbon dates acquired up to now, the phase Skala Sotiros II in the fortified settlement dates from the second half of the 3^rd millennium, between 2561 and 2138 BC89. On top of the destruction layer of building phase IIc, the well-defined layer of a new occupation phase (Skala Sotiros III) appears, surmounting the stone buildings, probably without any chronological gap.

46During this new phase, the enclosure is no longer used as a fortification and a large part of it was filled with soil. This is all the more evident as one hearth and a floor from phase III lie upon the western branch of the enclosure. On the eastern end of its northern branch, however, the enclosure maintained its former role as a space organizing element. The northeast end of the enclosure, preserved to a fair height, delimits an extended compact clay floor. A clay oval “portable” hearth on a stone slab came to light in this area, along with a large quantity of vessels, stone tools and many spindle whorls90. The floor belongs to phase III, and its destruction layer corresponds to the one excavated in the western part of the enclosure91.

47The destruction layer of this phase bore signs of strong fire and yielded rich finds in situ (fig. 32). Some characteristic types are worth noting (fig. 33): open and closed table wares, among which are many deep bowls with a spout on the rim and nipple-like appendixes or trumpet-like handles, often with one raised handle as well. The depas amphikypellon, common in Troy II and the Northern Aegean islands, was found in small numbers along with many storage vessels with smoothed or coarse surfaces and plastic decoration on the rim.

48The finds suggest intense storage and spinning activities. Absolute dates place this latest habitation phase (Skala Sotiros III) in the end of the 3^rd and the two first centuries of the 2^nd millennium BC. However, no chronological gap has been detected between the phases IIc and III92.

49Some rare potsherds suggest that the occupation of the hill might have continued during the LBA. The sporadic remains of occupation from the Historic times, between the end of the 7^th century BC and the Early Byzantine era, are much clearer, however.

The open settlement

50The excavation of two plots in the modern village of Skala Sotiros during the years 2006 and 2007 brought into light part of the settlement that was not contained in the stone enclosure93. In the first plot, 40 m southwest of the enclosure, stone building foundations came to light over a length of about 10 m (fig. 34). The excavation needs to continue in order to confirm that they belong to the Early Bronze Age.

51In the second plot, about 70 m south of the enclosure, the remains of a block of houses were found that dated to the middle of the 3^rd millennium BC. Over an area of about 200 m², we unearthed several post-framed buildings with clay floors and numerous clay structures and benches. Many grinders, mortars and storage vessels were found nearby. The earliest phase is characterized by the presence of rows of big upright stones, probably meant to retain small terraces. The study of the material from the floors of the post-framed buildings has not been completed yet, but the common presence of vessels with incised decoration is evident. Deep bowls with richly decorated bodies and one-handled cups with incised bands on their bellies are very frequent.

52Two finds are particularly important for our knowledge of the local copper metallurgy. Along with several lumps of ore and some slag, the excavation yielded a nozzle and an open clay mould with two casting surfaces, for a pick and a chisel94.

53All these finds certify that the area outside the south branch of the enclosure was occupied, probably at a date a little before the construction of the stone enclosure, also confirmed by the recent radiocarbon dates (supra). We thus have a relatively clear image of Skala Sotiros at that period: an open settlement, which developed a central architectural nucleus around the middle of the 3^rd millennium BC.

Conclusions

54The archaeological and archaeometric research conducted so far in the prehistoric sites of Thasos have established a stratigraphic and chronological sequence of the island’s occupation during the Neolithic and Bronze Age periods95.

55In Limenaria, the excavated Neolithic occupation layers are amidst the earliest known so far in Eastern Macedonia. The absolute dates confirm an uninterrupted occupational sequence from 5800 to 5200 BC96, and the stratified archaeological finds provide an outline of the cultural profile of Thasos during the Middle Neolithic and the beginning of the Late Neolithic I period, the only available so far. For the moment, we have no absolute dates from the first half of the 5^th millennium BC, to which the latest phase of the LN I and the earliest phase of the LN II belong. Maybe the corresponding layers should be sought in other spots of the plateau where the settlement developed.

56The chronological gap in the radiocarbon dates from Limenaria (between the beginning and the third quarter of the 5^th millennium BC) is partly covered by the recent dates from Kastri (supra, p. 348 and table 1), falling in the second quarter of the 5^th millennium BC.

57The Late Neolithic II appears in Limenaria only in some disturbed deposits97, while its presence is corroborated by only one radiocarbon date between ca 4300‑4100 BC (supra, p. 341). The period is represented by a few unstratified pieces of pottery from the caves “Drakotrypa” in Panagia98 and “Atspas” in Skala Maries99, and possibly from the ancient city of Thasos100. Thus, for the moment, the only stratified archaeological material and absolute dates from reliable layers come from Kastri, Theologos.

58The three radiocarbon dates from the last destruction layer at Kastri (supra, p. 348 and table 2), are extremely interesting. Two samples (Lyon-7914, Lyon-7915) date from the end of the 5^th millennium BC and the third (DEM-2134) reaches without doubt the beginning of the 4^th millennium BC. Along with the dates from Agios Antonios Potos and Limenaria they confirm the existence of the Final Neolithic period in Thasos.

59The characteristics of the Final Neolithic, and of the Transitional phase between the Final Neolithic and the Early Bronze Age, have been studied by Johnson for the whole Aegean and the Southern Balkans101. They have been related to the Chalcolithic phase “Rachmani” in Thessaly and to the beginning of the Early Helladic period. Phelps had previously defined the duration of the Final Neolithic in Southern Greece on the basis of pottery types102, taking into consideration the opinions expressed by Renfrew a few years before103. It was suggested that the crusted and pattern-burnished wares were characteristic of the early FN, whereas the bowls of the “Kum Tepe type” characterized the latest phases of the period. In the Southeast Balkans the widespread graphite-painted pottery, although still present, gives way to a trend of raw (crusted) pigments after the middle of the 5^th millennium BC; the latter also seem to prevail in Central Greece104.

60Johnson focused on the fact that dates between 3800 and 3300 BC are rare in Greece and in the Eastern Balkans, although this is not generally the rule in the Northwestern Balkans, where 4^th millennium dates have been recorded. He proposed two distinct phases: the Final Neolithic, from 4500 to 3700 BC, and a phase just before the Aegean EBA, from 3700 to 3300 BC, which he named Transitional. Even though all the earlier dates from Pefkakia belong to the 5^th millennium BC, the recent dates from Palioskala extend to the first two centuries of the 4^th millennium105. Recent dates from Mikrothives near Volos suggest that the beginning of the EBA could be placed much earlier, maybe just after 3600 BC106.

61Very few, ambiguous dates from the 4^th millennium BC are known from Eastern Macedonia. The earliest, from the beginning of the 4^th millennium, belong to the last phases of the Neolithic. A date between 3960‑3780 cal BC comes from Sitagroi phase III (not from its latest layers)107, and two dates between 3885‑3760 cal BC are known from Paradeisos108. Nothing comparable exists at Dikili Tash109, although the recent excavation in sector 6 corroborated the existence of a Neolithic phase in Dikili Tash immediately after the destruction of Houses 1‑4, which, according to the available absolute dates, belong to the beginning of the last quarter of 5^th millennium BC110.

62This list of dates is completed by radiocarbon dates from the beginning of the 4^th millennium BC from Neolithic settlements in Thasos. The excavation of the Markoulis plot in Limenaria provided a date between 3970‑3790 cal BC (supra, p. 342). Although it does not come from a secure layer, but rather from a storage pit, it is important because the pottery that co-occurred with the sample is characteristic of the beginning of the 4^th millennium: carinated bowls decorated with very simple linear graphite-painted motifs, and also with raw (crusted) off-white, yellow or pink/red pigments and coatings111.

63The presence of Final Neolithic in Kastri is corroborated by the three recent dates, one of which seems to belong to the beginning of the 4^th millennium BC (supra). The recent investigation in Agios Antonios Potos contributed two more Final Neolithic dates, between 3927 and 3662 cal BC (supra, p. 349: Lyon-7911 and Lyon-7650). Thus, the archaeological data from Thasos help to fill, albeit partly, the chronological gap observed in the Southeastern Balkans between the end of the Neolithic and the beginning of the Early Bronze Age. Of course further excavations in Thasos and Eastern Macedonia, as well as more radiocarbon dates, are necessary in order to understand what happens between the two.

64For the moment, there are only a few radiocarbon dates for the beginning of the cultural phase that comes after the Final Neolithic in Eastern Macedonia. These do not cover the entire 4^th millennium BC: seven dates from Sitagroi IV fall in the years between 3500‑3200 BC112 and two from Dikili Tash fall between 3328 and 3015 BC113. The start of the EBA in the “Katarraktes” Cave in Sidirokastro dates to the same timespan between 3340 and 3000 BC114, whereas another date from Kryoneri, at the lower Strymon valley, might also date to the end of the 4^th millennium BC but, due to its big error, might date to the 3^rd millennium instead115.

65It is true that we do not yet have adequate proof, based on absolute dates, for the period between 3700 and 3300 BC from Thasos either. The upper limits of the dates from Agios Ioannis are set around 3600‑3500 BC, but a global examination suggests that the main phase of the settlement should be situated around 3300 BC, during the EBA I116. The site has been considered as Transitional, because it maintains some of the typological characteristics that Johnson ascribes to this phase, at least as this appears in the Southern Aegean.

66Concerning the 3^rd millennium BC, which includes the EBA II and EBA III, the radiocarbon dates from Limenaria, Agios Ioannis Potos and Skala Sotiros perfectly fit the chronological pattern known from the rest of the Aegean.

67The excavations in Skala Sotiros confirm that during the Early Bronze Age a new settlement type appears on Thasos. It consisted of a fortified central nucleus surrounded by an open settlement. Its appearance coincided with the emergence of the early urban centres in the Aegean world117, to which the Northern Aegean islands, Thasos and Samothrace seem to belong118. The monumentality that differentiates the stone buildings in the interior of the enclosure in Skala Sotiros from the post-framed buildings outside it, suggests that the fortified complex should be identified as a space with some special function connected to a central authority119.

68In the middle of the 2^nd millennium BC, during the Late Bronze Age, the populations return to the mountainous hinterland. The Kastri peak is inhabited again, and new settlements appear at fortified mountainous sites such as Ai-Lias120, and probably Koukoudia121. However, settlements near the coast do not disappear. Occupation layers of the Middle and the Late Bronze Age have been revealed at Agios Antonios, and possibly at Skala Sotiros. A date from the second half of the 2^nd millennium BC from Limenaria122 indicates the continuation of occupation in one more coastal site. The movement of the population to the hinterland was completed during the Early Iron Age123.

69Kastri seems to constitute, during the Early Iron Age, an important settlement for Southern Thasos. Several sites in the southwestern part of the island are connected to it, such as Ai-Lias, Paliokastro in Maries124 and Koukoudia125. Its relationship to the barely investigated Early Iron Age settlement at the site of the ancient city of Thasos in the northern part of the island126 has not been determined yet, and we look forward to the results from the new excavation program concentrating on the pre-colonization phase of the historical city127.