Introduction2

1The settlement of Palioskala (53.66 m above sea level at the highest point) is situated on the eastern edge of the enclosed Pleiocene Thessalian basin, in the foothills of Mt Mavrovouni, which like Mt Ossa to the north, separates the plain of Thessaly from the Aegean Sea (fig. 1). The most important natural feature in this part of the basin was Lake Karla, known in ancient sources as Boebeis. This lake was the result of a Pleistocene tectonic subsidence, which had its greatest depth in the Thessalian basin3. Before the lake was drained in the 1960s, the Palioskala settlement stood on its eastern edge and most probably was a lake-side site during its lifetime (fig. 2). The presence of the lake and the marshy region surrounding it, could explain the limited number of prehistoric settlements in this region in comparison to the particularly dense distribution of Neolithic and Bronze Age settlements in the rest of the Thessalian plain4, even if the existence of settlements with pile-dwellings within the lake cannot be excluded.

2The continuous fluctuations in the level of the lake until recent years resulted from its particular hydrology and character. For this reason there are different opinions on its extent during the prehistoric period5. According to some opinions the level of the lake exceeded 50 m after the Middle Neolithic, receded in the Early Bronze Age, but rose considerably towards the end of the Mycenaean period and reached approximately 64 m6; according to others, it always fluctuated only to a maximum level of 44 m7. The second view fits better with the discoveries at the settlement of Palioskala, the lowest edges of which are at around 46‑47 m. After the construction of anti-flood barriers in the northeastern part of the Thessalian basin which obstructed the floodwaters of river Peneios from the lake, the level stabilised at a maximum height of 47.65 m8 and the lowest parts of the settlement of Palioskala were at times underwater. The lake is enclosed by mountains only to the east and the south, while the Thessalian basin extends out to the north and west. At periods when the level of the lake was low, it is likely that around it there were short-lived prehistoric settlements, which are not visible today as a result of alluviation9.

3The settlement of Palioskala is situated on the edge of the lake’s alluvial plain. Immediately to the northeast is a small rocky outcrop around which traces of a settlement dating to the Early Christian period have been discovered10. About 800 m away to the north is an active stream, while to the east a smaller stream flows in periods of heavy rain and drains into the lake, creating a shallow stream bed through the middle of the southern edges of the settlement.

The prehistoric settlement: circuit walls, buildings and structures

4The settlement (fig. 3) was excavated between 2001 and 2002 over an area of around 3500 m², in preparation for the construction of a drainage channel that ends in a large artificial lake in the southern part of the drained lake. The settlement has been prepared for public display and the channel has been rerouted 100 m further west11.

5The settlement is a circular and gentle magoula (tell) with a diameter of 100‑120 m, a maximum height of 6 m and an area of around 9500 m², though sherds are scattered over an area of around 15000 m²12. Earlier levelling works and tree-planting pits on the summit of the mound caused serious damage to the buildings in the upper levels and to the archaeological deposits. As a result it has been difficult at the present stage of research to distinguish clearly chronological horizons and building phases. Excavation took place over a large area, corresponding more or less to the eastern half of the settlement, but in many cases has not been completed inside the buildings. In many parts the walls of the buildings show significant deviations from their original line, probably as a result of subsidence or tectonic activity, a hypothesis to be further tested; indeed, several fault lines from the Middle Pleistocene-Holocene traverse the broad region13.

6At the summit of the mound traces of Hellenistic/Roman and modern activity have been discovered. Traces of a rectangular building, whose foundations lie over prehistoric structures (fig. 3: 12)14, probably belong to the former period, as do also some tombs, cut into prehistoric deposits in some parts of the settlement.

7A large part of the excavated settlement was enclosed by stone-built concentric circuit walls. In some cases the way of their construction and their date and function in relation to the overall intra-settlement spatial organisation is not yet clear15. The outer walls to the northeast, east and south, as well as some of the walls in the central part of the settlement were essentially terrace walls since they only have an outer face, which slopes towards the inner, higher part of the settlement to support it. They are constructed using small- and medium-sized blocks of limestone bonded with a small quantity of clay. There are indications that two or three circuit walls of this type existed to the southeast and south, distant from each other by up to 80 cm (fig. 3: 1), but it is not yet evident if these are from different phases or were in use simultaneously, creating a series of stepped walls. The outer circuit was excavated to its foundations: it is preserved to a height of 2.30 m (fig. 4). In the south part of the settlement and at a distance often less than 1 m from this circuit, there was possibly a further stone-built circuit which has been completely destroyed, as is indicated by heaps of stones forming a rough line (fig. 3: 2). Traces of stone foundations of buildings and other structures belonging to at least two different building phases have been identified outside the circuit wall to the southeast (fig. 3: 3).

8A further enclosure, circular in shape, has been revealed inside these terrace walls and parallel with them (fig. 3: 4). This structure has a 2 m wide opening in its east side leading to the central part of the settlement, which was later blocked up by a wall (fig. 3: 5; 5). The circuit wall to the south of this opening is 1.5‑2 m wide and both the inner and outer faces were built with small- and medium-sized, unworked limestone blocks and clay. Close to the entrance and to the north, only the outer face of this circuit wall was revealed. It sloped towards the interior of the settlement. A narrower and less well-constructed stone circuit wall, belonging to a later phase, was built on top of the previous one with the same alignment. On the wide terrace formed inside these walls a series of buildings and structures were found, in varied states of preservation due to recent levelling work and agricultural activity. They developed around the central part of the settlement, which was itself encircled by separate stone walls (fig. 3: 6). Several rectangular buildings of varying dimensions, and of different building phases, with clay floors, hearths and other internal structures, were excavated in the eastern area (fig. 3: 7; 6). A long rectangular building, with internal dimensions of around 10.50 m (east-west) by 7.90 m (north-south) [fig. 3: 8; 7], had successive phases of use, including two hearths and storage vessels (fig. 12). Circular hearths with clay floors and sturdy substructures of small flat stones and sherds were also common in the eastern area of the settlement (fig. 8).

9The central part of the settlement near the summit of the hill is encircled by two more concentric circuit walls 0.70‑1.30 m apart, built using small- to medium-sized limestone blocks and sloping toward the interior; they had only an external face of a single row of stones and were terrace walls (fig. 3: 9). Inside the innermost circle of these walls is a long rectangular building (the “central building”) with dimensions approximately 9 m (east-west) by 8.25 m (north-south), which was covered with massive rubble from its destruction (fig. 3: 10; 9). Its excavation has not continued further and its plan is elusive at present, but the lines of some of the interior cross-walls can be made out amongst the rubble. It is not yet clear if the building and the terrace walls around it were in use simultaneously but there are indications that the former was constructed earlier.

10A series of three square stone-built buildings was uncovered to the south-southeast of the above mentioned building (fig. 3: 11), on top of which there were architectural remains belonging to at least one later building phase showing a different alignment. The one in the middle with dimensions approximatively 4.85 x 3.80 m has an entrance 0.80 m wide at its northwest corner with a stone threshold, and a clay “platform” in the southwest corner slightly raised above the earth floor (fig. 10). The building was probably destroyed by fire. Finds included storage vessels (fig. 11) and three anthropomorphic figurines, including a clay torso and a marble head from an acrolithic figurine decorated with red colour (fig. 19). Two further heads from this type of figurine were also found in the upper levels of the building; among the finds there were three clay objects in the form of “bucrania” (krateutes)16 [fig. 21].

11Beneath a large pile of stones to the south of this building, two further stone-built concentric circuit walls were uncovered about 0.90 m apart. These walls form the inner boundary of the wide terrace with the buildings mentioned above, and probably encircled the central part of the settlement (fig. 3: 6). The outer wall appears to have been built with a single row of stones, sloping toward the central part of the settlement and being itself another terrace wall with a semi-circular projection 4 x 1.5 m on the face. A large rectangular building in the central and higher part of the settlement, built on the prehistoric walls on an entirely different orientation (northwest-southeast) with dimensions around 22.40 x 7.80 m (fig. 3: 12), probably dates to the Hellenistic/Roman period, if not later, as mentioned above.

12Five Neolithic graves with contracted burials were found in the southeast section of the excavation and outside the circuit walls, along with two tile graves, three pit graves and a single cist tomb found in different parts of the settlement and dating to the Hellenistic/Roman period, or to an even later period.

Relative and absolute chronology of the settlement17

13At the present stage of research a series of observations may be made to define at least a broad spectrum of the settlement’s chronology18, based on the preliminary study of the pottery and the available absolute dates. One should take into account however the destruction and disturbance of the upper levels by ploughing and levelling work, the incomplete excavation of the interiors of the buildings, and the eroded nature of the pottery.

141. A few eroded Late Neolithic sherds were found in different trenches. These include pieces of decorated bowls of the Classic Dimini style, and decorated and monochrome sherds probably belonging to the pre-Dimini phase Arapi. Occupation in the Classic Dimini phase is supported by the date of the sample Lyon-7640 (calibrated age: 4690‑4456 BC19): this sample comes from a level between two of the stone circuit walls that encircle the central building in the upper part of the settlement (trench E21)20.

152. The rest of the pottery can be placed in the Final Neolithic, in the transitional phase from the Final Neolithic to the Early Bronze Age, and in the Early Bronze Age as well, though a detailed classification of the settlement’s pottery is beyond the scope of this article.

16The monochrome medium and coarse brown-light brown pottery is more common than the grey and the black, while the quantity of fine ware is limited. In the coarse ware, but also frequently in the medium ware, the clay contains many inclusions, while in the very coarse vessels there are large pieces of grit and fragments of schist. The fine and medium slipped and burnished ware usually has surfaces that range from brown to dark brown and red-brown, and from dark grey to black. On the light-coloured surfaces, whether slipped or not, black mottling can be seen, caused by firing. In general, firing was poor: the fired clay is soft and the vessels fragment quite easily.

17The storage vessels are usually of brown or orange-brown clay and frequently have a brown to red-brown slip. Sometimes they are decorated with applied finger-impressed and “rope-like” coils (fig. 13; 23A: a)21. The common shapes are pithoid vessels, or vessels with an incurved upper body and a conical lower body that can have strap handles on the belly (fig. 11)22. Storage vessels with a high or low conical neck (fig. 13; 23A: a-c)23, a funnel neck, sometimes having vertical strap handles on the belly (fig. 23A: d-f)24, or bucket-shaped also exist (fig. 23A: g). A storage vessel from the later phases of the settlement has a narrow circular mouth with a low rim, semi-globular body and conical lower part, which ends in a narrow base, and was probably meant to be placed in a stand (fig. 12)25.

18“Baking pans” with thick low walls, sometimes of irregular height, forming a vessel with a flat base and very low rim, are quite common. A few similar vessels have a row of circular holes below the rim26. The clay of these vessels is frequently full of grit and the surfaces are only roughly smoothed. Cooking vessels are also made from coarse clay as well as other vessels such as a lid27.

19Tableware includes bowls, with moderately thick walls and a smooth, curving profile and slightly in-turned rim, which usually have a light-red to brown-red slip, sometimes with mottling (fig. 14; 24B: a). Deeper or shallower open bowls with straight walls are also common, although less than the previous type (fig. 24B: g). S-profile bowls (fig. 24B: h), shallow “dishes”, sometimes with a slightly concave base (fig. 24B: b)28, and open bowls with rolled rims, are moderately common and may have a black or deep red-brown burnished slip (fig. 15; 24B: d-f)29. Carinated bowls or bowls with a biconical body and sharply everted rim which is sometimes separated from the body by a shallow groove (fig. 24B: i)30, cups or small bowls with a more or less well defined S-profile (fig. 24C: a, b)31, basins with a slightly or sharply out-turned rim (fig. 25D: a, b)32, and “pyxides” (fig. 16; 25E: a-c)33, are also represented.

20Bases are generally flat, although there are a few examples of pedestals and a few low ring bases. On coarse ware there are strap handles, sometimes with a small conical knob on one end (fig. 23A: g), but also tongue-shaped or bilobate handles. Bowls can have pierced simple conical or roughly squared knob lugs (fig. 24B: g), and there are also twin lugs (fig. 25G: a), as well as a few conical horn-shaped lugs from deep bowls. Tubular lugs are also known; lugs of this type disposed horizontally below the rim are usually seen as markers for the transitional period between the Final Neolithic and the Early Bronze Age (fig. 25F: a-c)34.

21Decorated pottery is very limited and it seems that the symbolic role of pottery was transferred at this period to other artefacts, such as those of metal35. Crusted decoration with light red to pink, and rarely white paint or the combination of the two, is found in small quantities, eventually with spiral motifs36. Incised decoration is also limited (fig. 25H: a, b). Sherds with the “Bratislava-type” decoration, which is well represented at Petromagoula37, Mikrothives38 and Voulokaliva39, seem to be very rare. Channelled decoration on sherds with black or brown-black surfaces is also very rare40. As mentioned above, finger-impressed and “rope-like” cordons are almost exclusively found on storage vessels. Shallow impressions are very occasionally found on the rim of small tableware vessels.

223. Apart from the Late Neolithic charcoal sample Lyon-7640 (trench E21) mentioned above, four more charcoal samples were submitted for radiocarbon dating. They come from the following contexts:

1. Sample Lyon-7641: trench H26, charcoal sample 1, unit 9. It comes from a possible working surface, clay-coated, with small pebbles and numerous fragments of charcoal. This surface lies in contact with the west face of the circuit wall, which was founded on an earlier circuit wall of the same alignment.

2. Sample DEM-2157: trench I22, charcoal samples 11‑19, units 31, 33. It comes from a level outside the older circuit wall, which continues into trench H26 where the previous sample was found. This level was close to the lowest exposed part of the wall and can confidently be distinguished from the overlying post-depositional level. The calibrated age of the previous sample Lyon-7641 corresponds to the stratigraphic sequence of the two structures since it is younger by several decades than the present sample.

3. Sample DEM-2156: trench K16, charcoal samples 4, 5, 7, 8, 9, 46, units 30, 31, 40. It comes from a level below deep secondary deposits close to the foundation of the external terrace circuit.

4. Sample Lyon-7642: trench Θ13, charcoal samples 1‑6, units 29‑30. It comes from a level beside a stone-built wall which has not yet been completely excavated.

23The levels from which DEM-2157 and DEM-2156 come contained finds which either were related to human activities taking place outside the circuit walls, or are likely to have been rubbish, thrown out beyond the circuit walls of the settlement at a time when they were still in use. In both cases, the date of the samples should correspond to part, at least, of the period of use of the circuit walls.

244. The ¹⁴C determinations of these samples (see table below) give calibrated ages between 4460 and 3803 BC.

25Two phases can be distinguished, one a little after the middle of the 5^th millennium BC, represented by sample DEM-2157, and one after 4235 BC, as indicated by the remaining three determinations (without excluding the possibility of other subdivisions). These dates partly agree with the chronology derived from the pottery. There are however pottery features mentioned above which strongly suggest the continuation of habitation at the settlement in the Early Bronze Age.

26For the time being it appears that Palioskala is a multiphase settlement, but there is need for a more thorough chronological resolution. It correlates partly with the Rachmani levels at Rachmani itself and at Pefkakia, while the radiocarbon determinations suggest that it should also correlate partly with the Thessalian settlements at Prodromos in Karditsa (Western Thessaly)41 and Profitis Ilias at Mandra (between the eastern and western Thessalian plains)42. It also appears that Palioskala shares some features with Coleman’s “Petromagoula-Doliana Group” (rolled rims, plastic lugs, finger-impressed plastic coils with a “rope” like appearance on storage vessels, and some S-profile bowls), which intervenes between the Final Neolithic and the Early Bronze Age43. On the other hand, other features of the “Petromagoula-Doliana Group” like the characteristic “Bratislava-type” decoration, are represented hitherto by few and rather dubious sherds whereas the tall conical beakers are seemingly absent. To the early stages of the Early Bronze Age may belong the “pyxis” and some shapes mentioned above, though some types of vessels like the numerous “baking pans” were already known in the previous period.

275. Among the other finds from Palioskala should be mentioned chipped stone tools, particularly the bifacial leaf-shaped points and arrow heads of flint (fig. 17). The polished stone tools are chiefly axes and chisels, as well as quernstones and polishers. Bone tools are also present. There are clay zoomorphic and anthropomorphic figurines, including a two-headed figure (“kourotrophos” type44), and acrolithic figurines, the heads of which sometimes preserve traces of red paint (fig. 19)45. Jewellery includes a small “ring-pendant” of shell (fig. 20), thus adding a new material for the manufacture of these objects, in addition to gold, silver, copper, clay and stone known already from other sites46. There are also fragments of rings, beads and “buttons” from Spondylus gaederopus (fig. 18: a-e), or stone (fig. 18: f-i). There are also clay spindle whorls and “sling bullets”. Finally, one of a number of copper objects47 is a small flat copper axe with a rectangular cross-section. It has a square butt and convex cutting edge with a strong curve, splaying slightly beyond the long sides of the shaft (fig. 22)48. In Thessaly copper axes and chisels attributed to the Final Neolithic period have been found at Sesklo49, Pefkakia50, and Profitis Ilias in Mandra (Late/Final Neolithic)51, while similar finds of this period are known from other settlements in Greece52.

Discussion

The settlement’s plan

28Hitherto both the available calibrated dates for the charcoal samples and the pottery suggest that the concentric organisation of the settlement, with its circuit walls, the “central building” and the houses on terraces between the circuit walls, was the result of long building activity at the site. Palioskala, therefore, must be seen as a multiple phase settlement of which only the upper and most fragmentary parts have been excavated. This is further supported by the results of small test pits in the central part of the settlement, where stone-built walls were revealed below the dense horizon of the buildings and circuit walls unearthed so far. Whether the settlement was smaller in the Late Neolithic period, but was almost completely covered by the buildings of the Final Neolithic – Early Bronze Age in which settlements were generally larger in extent (see below), is difficult to determine at this stage. The fact that, so far, no traces of Late Neolithic deposits have been found in the southeastern part of the excavation outside the terrace circuits, lends force to this hypothesis. It is also clear that the houses, structures and circuit walls of the settlement underwent several alterations of their original plan. The earlier settlement plan was clearly significantly altered in some places, as shown among others by the fact that the eastern entrance went out of use, and by the construction of slighter circuit walls at a higher level on the same alignment as the early circuits. To what extent the original buildings in the centre of the settlement continued to exist and what their function was, are topics which will be clarified with the continuation of excavation.

29The results of the excavation show a concentric settlement plan, with densely packed buildings in the central part53. The relationship of the structures which developed beyond the magoula (in the southeast sector of the excavation) to the part of the settlement surrounded by walls, is still to be explored. Already from the Middle Neolithic a similar double arrangement, with distinctive features for each part of the settlement, has been recognised at Sesklo54. The concentric arrangement is frequent in the settlements of the Eastern Aegean in the Early Bronze Age55, but the closest and earlier example is that of neighbouring Late Neolithic Dimini on the Pagasitic gulf56. Even if comparisons between Dimini and Palioskala are still premature due to their probable temporal difference, it should be noted that at Dimini, the central area was occupied by a central courtyard with a number of buildings around its perimeter, while at Palioskala there appears to be an “isolated” building57, provided that the latter was simultaneously in use with the circuit walls around it, which, as mentioned above, is under discussion. A dense arrangement of houses sharing a common orientation was found in the upper level of the Rachmani phase at Pefkakia close to the summit of the settlement mound58. In contrast, at the settlement of Profitis Ilias at Mandra, which was surrounded in its latest phase by a circular stone-built enclosure, the buildings appear to be limited in number, with the exception of a megaroid house and surviving fragments of other buildings close to it, but also along the length of the internal side of the circuit59.

The circuit walls

30At the settlement of Palioskala the outer concentric circuit walls, besides their function as terrace walls, appear to have been intended in part for the protection of the settlement from the periodically rising water levels of the nearby lake, and also from the severe aggregation activity of streams coming down the slopes of Mt Mavrovouni and drained into the lake, thus preserving a dryer environment for the development of the settlement60. The “fortifications” on the west slope of Sesklo61 are considered to serve the function of terraces, while inclined terrace walls also exist at Dimini62. At Pefkakia, a wall which probably formed part of the circuit of the lowest Rachmani level had been built with two faces and a fill between them, while in the middle level of the same phase terrace walls built in the same fashion were also discovered63.

31As in the Late Neolithic, the settlements of the Final Neolithic and the Early Bronze Age have circuit walls which define the limits of the settlements both physically and symbolically64. They more often have the form of a magoula (tell), or are flat sites, as for example the excavated Thessalian settlements of the Late and early Final Neolithic at Profitis Ilias in Mandra (upper phase)65, at Platykambos66 and at Mikrothives67. From the end of the Late Neolithic period the Thessalian settlement at Profitis Ilias in Mandra was enclosed by a stone-built circuit wall, which replaced earlier encircling ditches68. Built circuits have been discovered at settlements of the Final Neolithic period at Mandalo, in Western Macedonia69, at the settlement of the Final Neolithic-Early Helladic I at Zagani70 and probably at the Final Neolithic-Early Helladic I settlement at Houmeza in Attica71. The settlement of the Final Neolithic at Strofilas on Andros is considered to be one of the earliest examples of fortified settlements in the Cyclades, with an outer wall and strong circuit wall, which is strengthened in the vicinity of the gate and elsewhere with stout curved bastions. The wall, which is 1.6‑2.5 m wide and has an estimated height of 3.50‑4 m, is built with the technique of two-faces with rubble fill72.

32Circuit walls as boundaries of settlements could have had different roles in periods of peace or conflict73. Their functions are more generally a matter of debate, and they could differ from one settlement to another74, given that circuit walls appear over a large area at different times in the European Neolithic75. Regardless of whether one accepts that the circuit walls of Neolithic settlements in Thessaly were defensive structures76, or rejects on the contrary their defensive role as incompatible with Neolithic society77, it is clear that the circuits were planned and achieved with the participation of the members of the community. This should reinforce the perception of social coherence and identity through the repeated negotiation of social relationships of the groups which took part in their construction78.

The distribution of Neolithic and Early Bronze Age sites in the vicinity of Palioskala and in the Thessalian plain (with emphasis on the Final Neolithic period)

33In strong contrast with the Late Neolithic period (Classic Dimini phase), where settlements have been discovered all around Lake Karla, in the Final Neolithic and more precisely during the “Rachmani culture” the settlement of Palioskala is for the time being the only one known on the eastern side of the lake, over a distance of about 23 km. The nearest known settlements for this period are between 5.5 and 13.5 km away and located at some distance from the opposite western and southwestern shore of the lake (before its drainage)79. In the eastern Thessalian plain the number of settlements decreases significantly in the Final Neolithic phase, and the thirty three known Final Neolithic settlements are about one quarter of the number of settlements of the Late Neolithic Dimini phase, while in the Early Bronze Age the number of settlements almost triples in relation to those of the Final Neolithic80. The focus of the surface surveys on the plain alone and not on the hilly or mountainous districts, as well as the difficulty of recognising and classifying the pottery of the Final Neolithic are responsible to a degree for the small number of known settlements of this period81. In the eastern Thessalian plain and more generally in Thessaly, the settlements of the Final Neolithic, with only a few exceptions, are established on older Neolithic settlements, while settlement continues at these sites into the Early Bronze Age in some 70% of cases82. The settlements do not have the dense distribution of previous periods and they are a considerable distance apart, with the exception of a few clusters on the dry Pleistocene terrace in the northwest part of the plain and in the Revenia hills which separate the eastern from the western Thessalian plain83. Settlements of this period show a preference for the edges of the plain close to the foothills, and also in the hilly regions84, which offer a greater variety of environment and greater opportunities for subsistence85. At the settlement of Palioskala the deeply dissected slopes of Mt Mavrovouni with its large number of small valleys and the lacustrine environment particularly favoured a mixed farming economy, combined with fishing86 and hunting in the upland and in the well-watered surroundings of the settlement. The wetlands surrounding the lake offered good grazing in the summer, and the same is true of the slopes of Mt Mavrovouni in the winter. On the other hand, in recent years (before the lake was drained), the fields which became available when the lake level dropped were so fertile, that sowing could take place without ploughing in the natural channels left by the water as it retreated87.

34The obvious reduction in the number of Final Neolithic settlements in Thessaly has been related to the greater size and larger populations achieved by the settlements in this period88, a fact which marks the change from the closely-distributed Neolithic settlements of a similar size to the nucleated settlements characteristic of the Bronze Age89. According to other opinions however, the Thessalian settlements of the Final Neolithic do not differ in size from the settlements of the Middle and Late Neolithic and therefore there is no question of a systematic nucleation90. The obviously greater extent and increased population of the mainland settlements should reflect, from an economic perspective, the increased role of animal husbandry and the cultivation of larger areas around the settlement which became possible with the use of oxen for ploughing91. With these economic and demographic parameters the role of some members of the community, who showed a more decisive conduct of affairs in periods of crisis, would provide a stimulus for the appearance of social inequalities and some kind of hierarchy that would be better seen later in the Bronze Age92. However, evidence from the Final Neolithic in Thessaly shows that these social changes are not the result of uninterrupted development93. In the fertile western Thessalian plain, the gradual abandonment of settlements from the Early Neolithic until the Final Neolithic, is not to be related to demographic contraction but to the development of nomadic cattle husbandry by small mobile groups of herdsmen, whose ephemeral establishments are difficult to locate94. Already from the end of the Neolithic a preference is apparent for districts close to grazing land at a higher level, while the use of caves95 increased just as much in the territory of the Attica-Kephala culture, which corresponds to the early Final Neolithic in Central and Southern Greece96, as in Eastern Macedonia97. In Central and Southern Greece, the settlements of the Final Neolithic increase in number in contrast to Thessaly, a fact which is attributed to population increase98. Many of these sites are thought to have had seasonal character and to relate to the exercise of transhumant pastoralism at the “periphery” of permanent settlements with mixed farming economies99, a practice which has also been proposed for the mountainous region of Epirus100. In some parts of Western Macedonia, such as the Haliakmon valley, the Ptolemaida-Vegoritis basin and Kitrini Limni, the density of settlements suggests a population growth at the end of the Neolithic period and in the Final Neolithic101 particularly in the first half of the latter102. In Eastern Macedonia, 45% of the sites of the Late Neolithic and the Early Bronze Age can be related to animal husbandry or to marine resources, while the coastal settlements in Thasos appear to be related to the larger settlements of the interior103. An increase in the number of Final Neolithic settlements is also attested in the coastal regions of continental Greece as well as in the Aegean, and was correlated with the major importance given to the marine roads and the role of commerce104. Nevertheless, in the whole of Northern Greece and more generally on the Balkan peninsula a drastic reduction in the number of settlements can be observed at the end of the 5^th millennium BC, i.e. at the transition from the Final Neolithic to the Bronze Age105, whereas the “Petromagoula-Doliana Group”, placed at the end of this transitional period, has been considered as a “Proto-Bronze Age” culture and related to migration movements from northern Balkan cultures106.