The Earliest History of Akrotiri: The Late Neolithic and Early Bronze Age Phases

Furthermore, the overwhelming majority of sherds belonging to vessels of everyday use gives a clue to the undoubted interpretation of the character of the early occupation of the site so far testified by archaeological evidence as one of a quite sizeable settlement.

INTRODUCTION

The recently completed thorough research on the copious pottery sherds from the site of Akrotiri on Thera has not only documented the long-known references of the excavators Marinatos and Doumas to finds dating to the Early Cycladic or even Neolithic period and to indications of the existence there of a quite sizeable settlement since at least the middle of the 3rd millennium BC (Marinatos 1968, 11, 26; 1969, 8, 13, 26, 28; 1971, 9, 23, 44; Doumas 1978, 777-778, 781) but has also placed the beginning of life at Akrotiri more than a millennium earlier and produced indisputable evidence of the unbroken occupation of the site since the Late Neolithic period.

On the basis of the ceramic evidence, it is the purpose of this paper to outline the earliest history of Akrotiri and offer some interpretation of the character of the early occupation, its beginning and its development.

THE CERAMIC EVIDENCE

From a total of about 8700 fragments, including a small number of restored complete vases, dated to the Early Cycladic or, in a few cases, the Late Neolithic period on the basis of shape, technique and decoration, about 1600 were chosen for cataloguing, either because they give indications of the shapes to which they belonged, or because they bear decoration characteristic of the earliest periods, or because, as in the case of the Talc ware for example, they are of a distinctive fabric. The rest certainly look early but they are body sherds, belonging for the most part to rather coarse closed vessels but giving no clue to the shape which they come from.

When dealing with this pottery, one is faced with two major disadvantages:

1. The lack of stratification: this is a drawback from both the typological and chronological point of view. With the exception of the vases and sherds found in the same context in the rock-cut structure of the pillar pit 6 and in the 'Fire', the large majority of the early pottery was found unstratified.
2. The fragmentary condition in which it has for the most part been preserved. In certain cases it was possible for a number of vases to be restored but these constitute only a small part of the whole.

• Stratigraphy:    

It was until quite recently believed that the Early Cycladic sherds which had occasionally been noticed in the course of the excavation were coming from the lower deposits of the site, especially from the pits dug down to the bedrock for the pillars supporting the modern roof (Marinatos 1969, 8, 13; 1971, 44; Doumas 1978, 778; Sotirakopoulou 1986, 297-298). The great number, however, of early sherds picked up among the numerous fragments of late pottery of the so-called destruction layer of the Late Cycladic city and the detailed study of the excavation diaries proved this impression to have been erroneous.

The overwhelming majority of the Neolithic and Early Cycladic pottery, more than 75% of the total, comes from the layer of the final catastrophe of the city; it was found either in the rooms of the houses, both in the upper and ground floors, or on the roads and squares around the buildings, where their upper floors fell in ruins due to the repeated earthquakes.

A plausible explanation for this is that the earlier pottery, together with bones, shells and broken obsidian blades, was contained in fragmentary condition in the pre-eruption soil of the island, which was used by the builders of the Late Cycladic city for the construction of the walls, floors, roofs (Marinatos 1973, 13-14) and thin partition walls of unfired bricks inside the rooms of their houses (e.g. rooms B1 and B1a, West House rooms 4 and 4a, Xeste 3 rooms 3b and 10). Such an explanation is further corroborated by the fact that a number of Early Cycladic sherds (about 250) were found in floor layers of rooms (e.g. room B6, Xeste 3 rooms 3a and 9, West House rooms 3a, 3c, 4, 5, and 6). Moreover, there is evidence that some at least of the basements of the late Middle Cycladic or the early Late Cycladic buildings had been founded directly on the bedrock (e.g. rooms B1, B2, D16 and West House room 4; Marinatos 1971, 22; Palyvou 1984, 135, 136, 146 and Fig. 2, 5,9), which means that in certain areas of the site the constructions that possibly existed earlier were destroyed from the foundations down to the bedrock and their remains were used as building material for the erection of the new houses.

About 10% of the total of sherds of the earliest phases were contained in the layer of the earlier catastrophe of the city, which seems to have occurred at the beginning of the Late Cycladic I period. This layer was found in the form of debris tightly packed together over the earlier floors of rooms, roads and squares with the purpose of forming a new raised level (Marinatos 1970, 8; 1971, 9, 22, 27-28, 44; Marthari 1984, 119-133; Palyvou 1984, 134-147).

As regards the pillar pits, a considerable amount of early pottery has been found in them (14% of the total), though not in the form of a more or less unmixed deposit lying just above bedrock, as might be expected. Here again the picture is similar to that of the other areas of the site: depending on the specific point where each of the pits happened to be sunk, whether in rooms of house (e.g. pits 2, 20, 21, 25, 28, 37, 39, 40, 41, 47, 48, 68) or in the open areas of the settlement, that is roads and squares (e.g. pits 3, 4, 11, 12, 14, 15, 16, 19, 27, 29, 31, 35), this pottery has come in the former case from the layer of the final destruction of the city and in the latter case from the debris layer of the earlier seismic catastrophe.

Finally, about 5% of the total came from the still unexplained complex of the 'Fire' and the area of the huge stone pithos close to it, as well as from the rock-cut vaulted structure encountered while opening pillar pit 6 (Marinatos 1969, 19-24, 27). These two areas can be considered as sealed contexts, the only ones up to this moment known from Akrotiri, dating to a late phase of Early Cycladic III and Early Cycladic II / III respectively (Sotirakopoulou 1986, 309; 1987). Their pottery contents, therefore, which also included whole vases, are useful for typological and chronological correlations.

• Frequency of shapes:    

The closed shapes predominate over the open, and the coarse or semi-coarse over the relatively fine. Among the 1600 selected pieces, the shapes appearing most frequently are the bowl, representing 36%, and the pithoid jar of various types, amounting to about 26%. There follow the pans/hearths, the cups, the wide-mouthed kraters with more or less cylindrical or two-stage neck profile, the pyxides, the jugs and the askoid vessels, numbering in turn from 55 down to 20 examples each. Other recognizable shapes are represented by even fewer examples.

• Decoration:    

All the known types of decoration and quite a few of the motifs of the Early Cycladic repertoire are present at Akrotiri: both incised rectilinear and curvilinear, pointillé, impressed, painted dark-on-light and white-on-red, as well as plastic, including ribbed, rope-pattern and impressed fingertip decoration. The number of the decorated sherds, amounting to only 4% of the total and 20% of the catalogued pieces, is comparatively small. The types of decoration most usually appearing are the incised rectilinear and painted dark-on-light, whereas the incised curvilinear, rope-pattern and impressed do not number more than a maximum of 25 examples each. Other types are represented by only a few fragments.

• Distinctive fabrics:

A number of Talc ware sherds, about 215 pieces, were noticed among the thousands of Akrotiri pottery fragments and were all collected. Some of them bear incised decoration consisting of simple rectilinear motifs or rope pattern. This fabric is well known from the Early Bronze Age period of Ayia Irini on Kea and has also been noticed in phases A2 and B at Phylakopi on Milos (Caskey 1972, 373; Wilson and Eliot 1984, 81; Evans and Renfrew 1984, 64, 66). Almost equal in number are the micaceous sherds noticed among the catalogued pieces. There are also quite a few fragments of coarse to semi-coarse reddish, brown or red-brown fabric, as well as some of fine red fabric.

All these, looking imported rather than local (Doumas 1983, 110), may, after answering the question of their provenance by analysis, give clues to the commercial and cultural relations between Thera and the other islands or even the Mainland during the earliest periods of Akrotiri.

• Typology and chronology:

Despite the fragmentary state of the earliest pottery, a variety of shapes has been recognized. Because of the lack of stratification, their dating has for the most part been based on parallels. Both the shapes and the number of examples of each increase as we move from the earlier to the later stages of the period. The number may in certain cases be very small indeed but, even so, the examples are indicative of the existence of the phase to which they typologically belong. In this respect, all the known phases of the Early Cycladic period are represented at Akrotiri. (For a full bibliography referring to most of the shapes to be listed below see Sotirakopoulou 1987. In this paper references are given only for the ones not mentioned there.)

The Late Neolithic period, to start with, is represented by 16 sherds almost all of open vessels, probably cups, bearing decoration of simple white rectilinear motifs on a dark surface (Fig. 1).

A few sherds belonging to EC I funnel-necked jars/kraters and spherical pyxides (Fig. 2) are representative of the 'Pelos group'.

The pear-shaped vase ('bottle') and the incised curvilinear decoration, considered to be characteristic of the transitional stage between the Early Cycladic I and Early Cycladic II periods ('Kampos group'), are also present.

Of the Early Cycladic II period ('Syros group') there are more examples: funnel-necked jars/ kraters (Tsountas 1899, 88, Pl. 8.1; Zervos 1957, Fig. 201-203, 214-217), black-coated kylikes (Tsountas 1898, 174, Pl. 9: 15; MacGillivray 1980, 16, Fig. 4: 417), lentoid pyxides, jugs with painted dark-on-light decoration (Fig. 3a), sauceboats, both with Urfirnis coating and painted dark-on-light, and open, rather coarse vessels or deep bowls with flat rims decorated with impressed triangles (Kerbschnitt).

The so-called 'Amorgos group' is found in the form of its characteristic jars/kraters and wide-mouthed jugs.

Of the Early Cycladic III 'Kastri group' we have straight-sided open bowls, tankards, bell-shaped cups, spherical pyxides (Tsountas 1899, 93, 122, Pl. 9: 15, 20; Rubensohn 1917, 44, Abb. 45; Bossert 1967, Fig. 4: 2; MacGillivray 1980, 18-19, Fig. 5: 56, 419, 433; Wilson and Eliot 1984, 78, Fig. 1: f, g), kraters and jugs with two-stage neck profile (Caskey 1972, 366, B41 a-c, 372, C14-C16, C17), one-handled pedestalled cups decorated with dark-on-light cross-hatching, as well as jugs with leaf-shaped mouth and decoration of vertical ribs on the body, which may also be classed with the forms of the 'Phylakopi I group'.

A considerable number of sherds belong to types dating either to Early Cycladic II or to Early Cycladic III or to both phases: bowls with incurving rims, pans/hearths, sherds from the bodies of pithoid jars with various types of lugs or handles (tubular, horizontal radially incised, crescent-shaped vertically pierced), funnel-mouthed pithoi (Fig. 4) and pyxides, wide-mouthed kraters with more or less cylindrical neck and outflaring rim.

The 'Phylakopi I group' of the Early Cycladic III period is represented by incised conical pyxides, askoid vessels (Fig. 5), jugs (Fig. 3b, 6b) and cups (Fig. 6 a, c, d) with painted dark-on-light decoration, plain cups with a pair of bosses diametrically opposed a little below rim, miniature amphoroid vessels, barrel jars (Fig. 7) and neckless pithoid jars with more or less spherical body.

Finally, there are quite a few Milian bowls, dating from the final stage of the Early Cycladic III or the transition to the Middle Cycladic period.

EARLY CYCLADIC AKROTIRI: A SETTLEMENT OR A CEMETERY?

The impression one gets when looking through the considerable amount of Early Cycladic pottery already dug up at Akrotiri is that of domestic wares.

The types known to have come almost exclusively from cemeteries, such as the Early Cycladic I spherical pyxis, the 'bottle', the Early Cycladic II funnel-necked jars/kraters, the pyxides and jugs with painted dark-on-light decoration, and the kraters of the 'Amorgos group' are few indeed and their presence is of little value in interpretation when compared to the strikingly predominating coarse domestic pottery.

Our scanty knowledge of Early Cycladic settlements, as well as the fact that fragments of painted jugs and pyxides have also been found in early deposits of domestic pottery at the sites of Ayia Irini on Kea and Phylakopi on Milos (Caskey 1972, 363: B12 and B14, 373: C28; Wilson and Eliot 1984, 80; Atkinson et al. 1904, 86; Evans and Renfrew 1984, 64, phase A2) corroborate the indisputable evidence of the existence at Akrotiri of a sizeable Early Cycladic settlement, at least as far as the excavated part of the site is concerned.

No walls dating from this early period have yet been identified. It is possible that remains of Early Cycladic walls may be hidden below the Late Cycladic ruins, at least at the parts of the city where the later buildings were not founded on the bedrock. It has, however, been suggested that the three rock-cut vaulted structures found up till now at the site (the ones in the cellar of the room D3 and in pillar pits 17 and 6, the last of which was full of household wares, mainly pithoid jars and pans/hearths) may in fact have served as dwellings during the 3rd millennium BC (Sotirakopoulou 1987).

THE DISTRIBUTION (Fig. 8)    

The Early Cycladic pottery of Akrotiri was found distributed all over the excavated area of the Late Cycladic city; its distribution however, is so uniformly uneven that it deserves some further comment.

The tendency of the earliest pottery to concentrate south-westward and become all the sparser as we move north-eastward is striking. It is interesting to note that the number of sherds dug up in the area defined by the north end of Building B to the north and by the room with the curved wall east of B7 to the east amounts to more than 75% of the whole. The densest concentration is seen in the south-westernmost part of the site, that is Xeste 3 and the areas just to the north and south of it, which alone gave 34% of the total of Neolithic and Early Cycladic pottery; the thinnest concentration, on the other hand, appears in the north-easternmost part, that is the area occupied by the Sector A, the 'kitchen' (Marinatos 1973, Topographical Plan), the House of the Ladies and the pits in the immediate vicinity of these buildings (pits 25, 26, 27, 28, 29, 31), which produced only 1.05% of the total. Furthermore, the Early Cycladic sherds coming from the West House, the most completely excavated building of the site up to this moment, are two and a half times fewer than those of the incompletely excavated Building G and even somewhat fewer than the ones coming from the narrow passage of the Telchines Road between the Buildings B and G.

These observations, taken together with the facts that (1) 12 of the 16 certainly Neolithic sherds have come from Xeste 3 and its immediate vicinity; (2) 10 more body sherds looking Neolithic were also found in the final destruction layer of the rooms 7 and 14 of Xeste 3; (3) the sherds of the Early Cycladic I spherical pyxides and funnel-necked jars were again found mainly in Xeste 3 and near it (pillar pit 2) constitute a considerable body of evidence in support of the hypothesis (Sotirakopoulou 1987) that the southwesternmost area of the site was the place of the first installation in Late Neolithic times and continued to be the nucleus of the settlement throughout the Early Cydadic period, in the course of which it gradually extended east and north-eastward.

It remains to express the hope that further future digging at the site, while illuminating more aspects of life of the flourishing Late Cycladic city and the tragic circumstances of its destruction, will at the same time shed more light on its earliest history, which, however fragmentary, appears to be far from insignificant.

--------------------------------------------

----------------------------------------