Akrotiri, the Chipped Stone Industry: Reduction Techniques and Tools of the LC I Phase

This assemblage is here presented in a synchronic approach, which examines the chipped stone artefacts from the point of view of the reconstruction of the obsidian reduction sequence, the flaking technology and the retouch techniques.

During the last decade, systematic studies of lithic industries in Greece (Torrence 1979, 1982 etc. on the Cyclades; Perlès 1981, 1984 on the Franchthi Cave and Kitsos; Van Horn 1976, 1977, 1978 and Runnels 1985 on the Argolid and Lerna; Moundrea-Agrafioti 1982, 1983 on south-east Thessaly) have shown that considerable information on the technological behaviour and its socio-economic aspects in the Bronze Age may be derived from chipped stone industries and that these constitute absolutely reliable chronological indicators. The internal organization of the chipped stone technology, resulting from the reciprocal influence of a multiplicity of factors, such as the raw materials, technical traditions and functional needs, displays regularities or multiformities related to the role of the stone industry in each society, i.e. its associations with the wider technical-economic system. The generalized changes which took place in the Aegean area during the Late and Final Neolithic period played their part in the establishing of multiple communication networks between the Cyclades and the periphery of the Aegean at the beginning of the Bronze Age. Their effect was felt on the conditions of trafficking and utilization of obsidian, the widespread and extensive exploitation of which raw material is immediately apparent from the increased quantities used in Helladic settlements, as well as from the basic technological changes in the manufacture of chipped stone tools.

We shall refer below to the internal characteristics of the chipped stone industry of the LC I phase at Akrotiri, with emphasis on a preliminary presentation of its technical-morphological composition. (We shall not consider here, as we had originally hoped, the spatial associations of the lithic industry and the location of intra-site variation, since artefact distribution in the three-dimensional archaeological space of the total has not yet been completed. A preliminary indicative distribution of major technical categories of the lithic industry by excavated sectors of the settlement has indicated, firstly, clear differentiation of frequencies of categories of chipped stone artefacts: Moundrea-Agrafioti 1987, Fig. 1). Of the diverse analytical perspectives which may be applied to prehistoric lithic industries, we shall here confine ourselves to elucidating certain basic issues which have been raised with regard to the chipped stone industry of the Aegean Bronze Age, concerning the procuring of raw materials, techniques of blank manufacture and, finally, techniques of tool manufacture. These issues are directly related to the economy, with regard to raw materials, technical specialization and the place and function of chipped stone tools in the final phase of occupation at the settlement.

The chipped stone industry at Akrotiri is represented by 2818 pieces of obsidian and flint recovered from: a) The LC I phase of occupation of the settlement, mainly found inside the buildings, on the ground floor, as well as in the upper storeys. A small proportion of the obsidian from this phase was found in outdoor areas, streets and squares, in particular in the accumulated debris from the collapse of the upper storeys during the final destruction of the settlement. b) The infills of the major building changes executed at the settlement during the Late MH / Early LC I phase. These levels have been revealed on a limited scale in relation to the overall extent of the LC I phase, in the deep pits dug for the columns supporting the roofing of the site and in trenches dug in free spaces or the floors of ground floor rooms within the buildings (S. Marinatos 1972, 8-9; Doumas 1978; Marthari 1984; Palyvou 1984). c) Locations where MC or EC occupation has been attested. These are infills which have come to light on a very limited scale, mainly in the foundation pits for the roofing columns and the region of the central torrent (the Fire) (Doumas 1978; Sotirakopoulou 1990). d) 20% of the total of chipped stone tools at Akrotiri are without any indication of their excavation provenance and so can not be attributed to a specific habitation phase (Moundrea-Agrafioti 1987).

In this paper we shall refer only to the material deriving from archaeological contexts of the first category, i.e. to chipped stone tools of the LC I phase of the settlement. A total of 1205 chipped stone obsidian tools may be attributed with certainty to this phase, most of which come from inside the buildings and were found under 'closed find' conditions.

RAW MATERIALS AND REDUCTION TECHNIQUES

The raw material of the stone industry of the LC I phase at Akrotiri is almost exclusively obsidian. Flint represents a very small proportion of the chipped raw material and there are also very few waste by-products of other volcanic stones, though these are not related to the process of producing chipped stone tools (Table 1.A).

The obsidian displays considerable diversity of colour and translucency: grey, opaque obsidian, with or without bands, is the variety predominant at Akrotiri (Table 1.B). There are lesser proportions of grey translucent obsidian and shiny black opaque obsidian, a variety which led S. Marinatos to suggest that the obsidian at Akrotiri did not come from Milos (S. Marinatos 1969, 48). However, neutron activation analysis of a small number of obsidian samples has indicated the Milian provenance of the obsidian at Akrotiri and, moreover, that both known quarries are represented, Adamas (at Nychia) and Demenegaki (Aspinall and Feather 1978). In the total of material studied, seven fragments of translucent white spotted obsidian were located, most probably indicating that obsidian from Yali was also imported to Akrotiri, though not for the manufacture of chipped stone tools. These were seemingly the by-products of the manufacture of minor objects, which have not yet been located.

TABLE 1. The raw materials

 Table 1A: Quantities of obsidian and non-obsidian artefacts

Table 1B: Quantities of obsidian

Flint is light-coloured, semi-translucent or opaque and occurs only in the form of a thin tablet covered externally by an irregular, calcareous cortex (Fig. 3. 11-13). No flint beds have been located on Thera. The complete absence of waste by-products from the working of this material probably indicates its importation in the form of retouched tools. Flint of comparable form was used at Phylakopi (Bosanquet and Welch 1904, Fig. XL) and at Ayia Irini on Kea (Cummer and Schofield 1984, Pl. 44), as well as in Troy VI, where this particular flint, of unknown provenance, was used for the first time (Blegen et al. 1953, 25). There is white nodular chert on the north-east coast of Milos (Bosanquet and Welch 1904, 194), though it is not known whether it also occurs in tabular form. Cherry and Torrence consider Cycladic flint to be a form of altered rhyolite, possibly of Milian provenance (1984, 22). Tabular flint also occurs in Bronze Age settlements of mainland Greece (Van Horn 1976, 304), though we do not know whether this material is macroscopically comparable with that of the Cyclades. The limited yet specialized use of tabular flint in the Aegean area perhaps constitutes evidence of yet another raw material circulating via exchange mechanisms, perhaps similar to those operating for obsidian, for which reason the locating of the geological source deposits is of particular interest.

The chipped stone technology at Akrotiri is based entirely on raw materials imported from elsewhere. Since the pioneering distance-decay models of exchange proposed by C. Renfrew (1977), the conditions of circulation of obsidian in the prehistoric Aegean have preoccupied researchers, for they comprise part of the wider economic system of inter-regional contacts and spectrum of economic exchanges defining the emergence of the complex social formations of the Aegean Bronze Age.

Obsidian is an essential raw material which was in circulation in the Aegean from the 7th millennium BC but was used systematically only from the beginning of the Neolithic period. It was always supplementary to the native flint and chert in Neolithic settlements on the periphery of the supply zone, vis-à-vis the primary sources, the quarries of Milos. The beginning of the Bronze Age is marked by a distinct increase in the frequency of use of obsidian in the settlements of mainland Greece, while at the same time a widespread change in the mode of production of obsidian blades is observed, associated with the technique of pressure flaking. An inevitable consequence of the appearance of mechanisms controlling access to the sources, the establishing of a trade monopoly and the intervention of specialized craftsmen in extracting and processing the obsidian would have been a change in the manner of supplying the rock, as well as in the technology of exploiting a raw material of limited provenance. From as early as the beginning of the century the development of the city of Phylakopi had been linked with the control of the obsidian quarries of Milos (Bosanquet 1904). Recent research by Torrence in the Milian quarries has, however, challenged this hypothesis. According to Torrence, the lack of task-specific mining tools, defensive and workshop installations or organized obsidian manufacturing areas in the quarries on Milos, as well as the minimal standardization of the by-products of shaping obsidian macrocores, and of the techniques of the initial stages of blade core reduction by pressure flaking, cast doubt on the control of the obsidian quarries during the Bronze Age and on the employment of specialist craftsmen in the manufacture of macrocores or prismatic blades. The data from the quarries seem to support Renfrew's hypothesis that obsidian was not a sufficiently precious raw material to support an exchange trading system and that its procurement by direct access to the sources, in conjunction with balanced reciprocal exchange for those users at a distance from the sources (down-the-line model), must have been the sole system of supplying the Aegean region with obsidian during the Bronze Age (Torrence 1982, 1984b, 1986). Certainly the circulation of obsidian should not be isolated from the wider spectrum of products, technological knowledge, services and symbolic and socio-political relations linking the societies of the Aegean area (Barber 1987, 117-119). Nevertheless, it is necessary to define precisely the parameters which will connect the material evidence with the proposed interpretative schemata.

Consequently, it is interesting to see to what extent the stone industry at Akrotiri is able to contribute to the issues outlined above, though from the perspective of a Cycladic settlement which perhaps differs with regard to the mechanisms of supply from the settlements of the wider Aegean region, as well as from the perspective of a settlement of the beginning of the Late Bronze Age, a period during which a reduction in the circulation and use of obsidian is observed, apparently related to an increase in the use of metal tools (Renfrew 1982, 225; Barber 1987, 155).

A first basic question is whether obsidian was imported to Akrotiri in the form of unworked raw material or as pre-formed cores, which would be the case if mechanisms of control were implicated in the distribution of obsidian from the quarries. The procurement of pre-formed cores may be ascertained from the degree to which the initial phases of the reduction sequence may be reconstructed. The knapping process subsumes a series of technical actions, in temporal succession, producing characteristic products and by-products, forming in general outline a 'technical chain'. The reduction sequence for prismatic blades includes, in its complete form, the stages and phases intervening between the unworked raw material and the finished artefact. It is obvious that if obsidian was imported to Akrotiri in the form of unworked nodules then the initial phases of the sequence should be represented at the site and, consequently, it is not possible to investigate the degree of intervention at the quarries in controlling the flow of obsidian out of Milos. Conversely, if circulation of obsidian was controlled by groups of specialist artisans which intervened in the quarrying and pre-forming of the raw material, then the first stages of the reduction sequence should be absent from the settlements which imported macrocores or obsidian products through economic and trading transactions.

However, the degree of reconstruction of the reduction sequence may also be influenced by intra-settlement factors, e.g. the existence of special areas or workshops for working the raw material. When there is a distinction between the areas in which the artefacts are prepared and those in which they are used, the reasons for incomplete distribution of the initial phases of the technical chain of obsidian blade manufacture may be attributed to the differentiated use of space and not to the modality of procurement of the raw material. In this case the intervention of specialist craftsmen in the manufacture of chipped stone artefacts may be inferred, though it is difficult to determine the role of local artisans with regard to the economy of procuring raw materials. The degree of reconstitution of the reduction sequence for obsidian, as well as the degree of standardization of products, may give some indications concerning the aforementioned points.

If we assume that pre-formed cores were imported to Akrotiri from Milos, the first phases of the reduction sequence, attested by the frequency of totally corticated flakes and characteristic waste by-products of the shaping of macro cores for blade production, should be absent. Likewise, if the manufacture of pressure blades, and the rejuvenation and repair of cores was executed by specialist craftsmen in areas outside the sectors of the settlement excavated to date, the by-products of these activities should be absent from the areas of tool usage, while, on the contrary, prismatic blades, the desired product of the pressure flaking technology, should predominate numerically.

Table 2 presents the frequencies of debitage categories arranged according to blank morphology (flakes, blades, indeterminate) and the technical categories referred to in the three initial stages of the reduction sequence: a) unworked raw material, by-products of shaping the nodules and shaping the crests (Table 2. 1-3); b) removal of crests in order to shape dorsal ridges-guides for detaching the first blades (Table 2. 4-6); c) cores and products of manufacture or rejuvenation of the striking platform of the core (Table 2. 7-10); d) tertiary products of knapping and amorphous debris (Table 2. 11-12).

Concerning the unworked raw material, the starting point of the reduction sequence, only a small number of natural obsidian nodules have been found in the LC I deposits at Akrotiri and these of such small dimensions as to be considered unsuitable for knapping (Table 2.1). There are no reserves of raw material, as would be the case in workshops or storage areas, nor is there a quantity of unworked raw material which could be regarded as corresponding to small-scale domestic supply.

Table 2: Technological categories of the chipped stone industry

Of the assemblage of chipped stone artefacts, 20.9% of the debitage are related to cortex removal (Table 2. 2-3). Primary, totally corticated flakes are few in number (3.6% of the chipped stone assemblage), while 17.3% of the total of debitage bear partial cortex (secondary debitage) mainly flakes, but also blades. The majority of debitage flakes at Akrotiri must correspond to the first phases of cortex removal from the nodules. However, it is known that decortication of the core was not always complete, since corticated surfaces remain even in the final phases of core reduction, as six residual cores from the settlement attest (Fig. 1. 8, 9). The frequency of cortical pieces at Akrotiri appears low in comparison with e.g. the data from the Shrine and Sector PLa at Phylakopi where the frequency of cortical debitage varies between 57.2% and 74.8% (Torrence 1985, Table C1, C2). It is, however, quite comparable with the frequency of cortical debitage from settlements further away from the source of obsidian, e.g. EH and MH Lerna and even the Neolithic settlements of mainland Greece (Runnels 1985, Table 6).

The technique of detaching blades by pressure produces the characteristic by-products of the shaping of the core, crested blades of first or second series (Fig. 1. 1-4, Fig. 2. 8, Fig. 4. 4, 6) and rejuvenation flakes of the pressure platform (Fig. 1. 6-7, Fig. 2. 6). Of course, some of the crested blades may also be produced in phases of advanced debit age or be related to repairs to the morphology of the core during the reduction sequence. 10.4% of the chipped stone industry at Akrotiri is associated with the primary manufacture and rejuvenation of cores (Table 2. 4-10). The frequent rejuvenation of the pressure platform of the blade cores is attested by the numerous pressure rejuvenation flakes with centripetal preparation (Fig. 1. 7), as well as by a few core tablets (Fig. 1. 6). It is difficult to estimate the technological significance of these frequencies of debitage categories. The fact that a proportion of primary debitage products contributes to the tools, reduces its exclusive link with the sequence of flaking. One fact is, however, indisputable: the existence of by-products of the initial stages of blade core manufacture attests the application of the primary stages of blade core exploitation at the settlement.

Cores at Akrotiri are exceptionally few (1.3%), characterized by polymorphism and all have been found in the final phase of reduction. They are exhausted and cannot be connected with an active process of blank manufacture. There are six examples, representing the two known types of prismatic, pressure blade cores: the cylindrical, with peripheral blade removal (Fig. 1. 10-11) or with only frontal blade removals (Fig. 1. 9), and the tabular with two opposing detachment faces. Certain debitage techniques attest that the original dimensions of some blade cores must have been quite large (Fig. 1. 4, Fig. 2. 8) while from the lengths of the complete blades the average length of core seems to have been 40 mm (Table 7.B, Fig. 1. 1-3). There are also three globular flake cores with multiple percussion platforms, also in the final phase of reduction (Fig. 1.8). We do not know what proportion of the flakes from the settlement are by-products of the manufacture of blade cores or were produced in an independent technical process, i.e. which are intentionally manufactured products of a flake manufacture sequence, as certain examples would suggest.

The tertiary debitage products, flakes, blades and blanks of indeterminate form, account for 61.7% of the chipped stone industry assemblage at Akrotiri (Table 2.11). This category comprises: 1) products of small dimensions which may be considered as by-products of core reduction; 2)blanks, which bear no macroscopic traces of use or retouch (about 2/3 of the total of tertiary blanks); 3) blanks which have been transformed into or used as tools, and finally 4) debris, angular by-products of knapping (4.9%) connected with the phases of rejuvenation of the core overhang, the final reduction of the cores or the re-cycling of raw material.

In all, 1/3 of the artefacts at Akrotiri is related in some way to the initial phases of the debitage process, from which proportion it is surmised that there are concrete indications of the existence both of the first phases of cortex removal and shaping out of cores and the phases of rejuvenation and regular core reduction in the excavated sector of the settlement. These phases of the technical chain would be missing if pre-formed cores were imported or if there was a clear intra-settlement differentiation between areas of usage and areas of manufacture of chipped obsidian artefacts. These observations support the hypothesis that at least part of the raw material was imported to the settlement in its natural state as nodules. However, the low proportion of by-products of the first phases of the technical chain again raises the issue of the importation of macrocores, as well and the possible existence of disposal areas for the residue from knapping. Bearing in mind the urban organization of the built area revealed at Akrotiri (Doumas 1983, 130), the existence of such areas in the unexcavated part of the settlement cannot be dismissed. The picture presented by the by-products of the debitage sequence at Akrotiri is rather different from that expected, of a systematic and specialized production of pressure blades. We shall now seek possible elucidation of this issue in the morphometric features of the debitage.

DEBITAGE PRODUCTS

A diacritical trait of Bronze Age stone industries is the manufacture of prismatic blades with standardized morphometric features, a regularity which is attributed to the application of pressure technique, a debitage technique which permits control of blade size and optimal reduction of the raw material. The conditions in which this new technique appeared are not known exactly. Perhaps it constitutes a continuance of the generalized changes in modes of obsidian blade manufacture which commenced in the Helladic area in the Late Neolithic (Perlès 1981, 1984, 135; Moundrea 1982, 102-103). At the beginning of the Bronze Age considerable improvements were made in the technique of blade removal and the mode of application of force for removal of the blade changed. Pressure is distinguished from indirect percussion by the preparation techniques of the pressure platform and from the different morphology of the blade cores (Van Horn 1976; Torrence 1979, 1984a; Cherry and Torrence 1984; Runnels 1985). The economic consequences of this new technology are obvious, since pressure is able to ensure significant economy of raw material and virtual mass manufacture. It is not yet clear whether this technical innovation, which demands particular skill, was incorporated in the common substratum of basic technological knowledge of the social unit or whether it should be linked to the emergence of craft specialization. The techno-morphological variety of pressure cores and the lack of standardization in core reduction strategies raise doubts as to the existence of specialist craftsmen and there is a contradiction between two phenomena which make a parallel appearance around the beginning of the Bronze Age: the greater ease of procurement of obsidian - which is attributed to improvements in communications and means of transport which permit the export of increased quantities of raw material from Milos - and the strategic optimization of techniques of obsidian blade manufacture. The expediency of technological change which ensures economic use of the raw material and a small number of waste by-products can only be understood if it is also accompanied by changes in labour organization within the settlements importing or 'consuming' obsidian. In terms of this construct, certain points may be distinguished on the basis of the lithic industry at Akrotiri, principally concerning the expediency of debitage techniques and conditions of product usage, flakes and blades, at the beginning of the Late Bronze Age.

Of the debitage products at Akrotiri flakes are in the majority (55.5%), while blades represent only 34.6% of the total of blanks. For approximately 10% of the debitage assemblage at Akrotiri manufacture and use renders the initial morphology of the blank indistinguishable (Table 3.A). It is known that blade frequency appears reinforced in calculations based on the arithmetic total for each category of blank since, due to their elongation and small width, blades fracture more easily than flakes along their long axis. In Table 4, which shows the degree of preservation of flakes and blades, the different behaviour of these two forms of blanks in breakage is clear: about 2/3 of flakes are preserved whole, while blades are mainly represented by their medial or proximal segments (59.7% of the total). Consequently, a better impression of the relative frequency of flakes and blades is given by estimating the minimum number of each form of blank on the basis of debitage preserving the butt (Table 3.B). By this calculation blades represent only 28.8% of the total of blanks at Akrotiri.

TABLE 3. The products of debitage  

*MINIMUM =plain and proximal pieces.

TABLE 4. Degree of conservation of flakes and blades

TABLE 5. The butt morphology

TABLE 6. Degree of platform overhang remove

The frequency of blades at the site seems low for a Bronze Age settlement in comparison, e.g., with data from Neolithic or Early Bronze Age settlements in mainland Greece, where blades exceed 2/3 of the debitage total. It is, however, strikingly higher than that known for the LC III strata of the Shrine at Phylakopi, where blade frequency varies from 6-14.3%(Torrence 1985, Table C1). (Indicatively we mention that at Dimini (LN) blades represent 77.7% of the total of blanks (Moundrea 1982) and at Kitsos (LN) 60% (Perlès 1981). On the contrary, at Saliagos (LN) blades do not exceed 10% of the debitage, a seemingly inexplicable fact for a Late Neolithic settlement. I consider it possible that this phenomenon is associated with the abundance of raw material and perhaps with different mechanisms of supply and access to the obsidian sources operating in Cycladic settlements.)

This relatively high frequency of flakes in a Bronze Age settlement may be understood if the flakes are by-products of an intensive blade manufacture process, or if they are products of an intentional debitage and connected with an independent chain of blank manufacture. From the morphology of the butt it seems that the flakes at Akrotiri were produced by different processes than those for blades. The categories of butt not associated with special preparation of the striking platform and corresponding primarily to detachment by direct or indirect percussion - cortical, linear, punctiform -are mainly encountered on flakes (Table 5). The flat butt, corresponding to lack of preparation of the striking or pressure platform is the commonest category of butt on flakes. The flakes are of small dimensions: 90% of the total are less than 30 mm long. The correlation of length with elongation index (length/width) shows that the most numerous categories of whole flakes are the short, wide flakes with elongation index equal to or less than 1.5 (Table 8). These traits may be interpreted either by the hypothesis of independent flake manufacture by percussion, mainly from small cores, or by the abandonment or utilization of a large part of the by-products of blade manufacture in the excavated sector of the settlement. Even though special obsidian working areas have not been located at LC I deposits of Akrotiri, we have seen that 39.2% of the total of flakes may be associated with the initial reduction sequence and core rejuvenation processes (Table 1). Judging from the small contribution of flakes to the tools (Table 9) and the selection of larger blanks for use, we believe the high frequency of flakes is indicative of the existence of primary phases of debitage in the excavated part of the settlement, though the manufacture of flakes by a process independent of blades cannot be ruled out.

The blades at Akrotiri display the typical morphological traits of pressure: rectilinear parallel or converging margins and dorsal ridges, direct striking angle, trapezoidal section (Fig. 4. 1-9). The careful faceted preparation of the pressure platform of the core creates the suitable perpendicular striking angle and protects the crutch from slippage of the pressure tool when force is applied to remove the blade. Characteristic of pressure, and a distinguishing trait with regard to Neolithic techniques, is the avoidance of removing the core overhang which projects markedly at the proximal end of the blade. Preparation of the pressure platform of the core was, however, not systematic at Akrotiri: only 54% of the blade assemblage has dihedral or faceted butts (Table 5). 1/3 of blades has a small, plain butt, confirming the possibility of application of pressure to a flat part of the percussion platform also (Pelegrin 1984). Independent of butt morphology, the typical platform overhang of pressure is observed on 80% of the blades in the Akrotiri assemblage, the proximal section of which is preserved (Table 6).

From the dimensions of the blades at Akrotiri it is difficult to estimate the distribution of length categories since whole blades are few and deviation of the sample from the mean great (CV =37.1) (Table 7.B). On the contrary, distributions of width and thickness are based on a larger number of artefacts and are perhaps more representative. The distributions of blade width show a distinct trend of products towards low values: 68.5% of the blade total are of width 7-12 mm (average = 11 mm), that is they are bladelets. Do we have, however, a standardized manufacture which enables us to assume the existence of specialist craftsmen? It is difficult to assess the meaning of basic statistical distributions of dimensions with regard to the issue of technical specialization. The average blade width and thickness at Akrotiri, and especially the standard deviation for each dimension, are more or less analogous with those given by Torrence for a series of Cycladic settlements and for Knossos (Torrence 1979, 81, Table 2, 3). In these settlements the coefficient of variation from the mean for width and thickness is considered to be high and is interpreted as proof of the lack of the required degree of standardization expected when tool knappers are specialist craftsmen. If the variety of manufacture techniques for prismatic blades is also taken into consideration, we should perhaps maintain that at Akrotiri we do not have indications of specialized and serial blade production.

RETOUCH TECHNIQUES

Bronze Age lithic industries are characterized by the low degree of blank shaping by retouch, as a consequence of which the typological range of tools is small. We have tried to investigate the morphological composition of the tools at Akrotiri in an effort to elucidate the issue of the degree of retouch in lithic industries at the beginning of the Late Bronze Age. Within the framework of a technological approach, what is of interest is the degree and kind of secondary intervention. which is linked not only to functional but also to cultural behaviour.

On the basis of the existence or not of retouch or use scars, the debitage products at Akrotiri were divided into the following categories: 1) unretouched products, 2) blanks with peripheral use retouch, 3) splintered pieces (pièces esquillées) and 4) retouched blanks (Table 9). From the distribution of the lithic industry within the above categories, we observed that at Akrotiri the degree of blank shaping is especially low: only 10.6% of the total is shaped out by retouch. The a posteriori tools - blanks with use scars and splintered pieces - represent 37% of the lithic industry. The proportion of blanks left unretouched and without obvious use retouch is over half the debitage total for the settlement.

If we look at the distribution of categories of degree of shaping according to blank form - flakes, blades, indeterminate, tabular flint - we observe the different contribution of blanks to the use and retouch groups (Table 9). We have already noted that both for flakes and blades one out of two blanks does not belong to the categories of tools or of blanks with use. Blades, however, bear traces of use-wear with greater frequency than flakes (27.6% of blades versus 19.9% of flakes). In the group of blanks with traces of use flakes predominate (52.8% of the total), while in the group of tools it is blades (46.2% of the total). It has been maintained that the morphometric standardization of blades is linked to the different functional orientation of chipped stone tools during the Bronze Age, where they were used supplementary to metal tools (Torrence 1979, 74). Of the splintered pieces flakes predominate, but the morphological features of the blank are indeterminate for 49.2% of the total of the group. Finally, the flint tablets are all shaped out as tools.

TOOLS

At Akrotiri the shaping out of blanks by retouch - secondary technical intervention which transforms the morphology of the blank at the edges and/or on the faces - is restricted to the manufacture of only certain groups of tools: the tools senso strictu (Table 10.A).

The tools with marginal retouch (Table 10.A, 1-3) constitute the most numerous group of tools sensa strictu at Akrotiri (91 pieces). We observe tools bearing continuous marginal retouch, direct or inverse, on one or both edges (Fig. 4. 4-9), or marginal retouch creating a notched and serrated outline (Fig. 3. 9-10). It is not certain whether these groups always result from the retouch process, but the peripheral retouch removals display relative regularity and, in several instances, are covered with abrasions due to use. In a few instances the tools are also shaped out with abrupt marginal retouch.

Next, in fairly low frequency (1/3 of the total of tools sensa strictu), are the tools bearing a significant degree of retouch: sickles, truncated pieces, end scrapers, piercers and projectile points (Table 10.A, 4-8). These tools are the only ones which may be considered indicative of the degree of application of retouch methods in the LC I phase at Akrotiri. These groups constitute the most interesting tool groups at Akrotiri, both from the techno-morphological point of view and for chronological parallels with Cycladic settlements outside Thera.

Sickles - a group of tools defined typologically by an a posteriori trait, the existence of silica gloss, the characteristic macroscopic smoothing and polish created by reaping - are the only tools at Akrotiri manufactured from flint. We know that obsidian was also used for reaping but only two blades with the typical dull abrasion which this creates on obsidian have been identified (Fig. 3. 9-10). Although we should be circumspect as to the frequency of use of obsidian for reaping before studying micro-traces, one fact is clear: there are no obsidian tools at Akrotiri techno-morphologically comparable with the typical denticulated flint sickles (Fig. 3. 11-13). Sickles are the only tools produced by a different technical chain than that operating for obsidian: the natural, thin flint tablets are not reduced but knapped with bifacial retouch in order to remove the cortex covering the edges. The retouch shaping out is bifacial, low angle, sometimes invasive, but never covering, as observed on sickles with analogous traits from mainland Greece. Bifacial retouch produces regular wide adjacent serrations on the cutting edge of the tool, which bear the characteristic silica gloss (Fig. 3. 11, 13). Since this type of denticulated retouch does not occur on any other chipped stone tools at Akrotiri, we believe that pre-formed sickle elements were imported to the settlement, which fact is ascertained both from the lack of by-products from flint knapping and their shaping technology which could be described as 'alien' to the predominant technique of tool production at Akrotiri.

Two morphological varieties of sickles may be distinguished, probably due to the different initial morphology of the flint tablet: broad trapezoidal elements (Fig. 3. 12) and narrow elongated elements (Fig. 3. 11, 13). The six sickles at Akrotiri were not found in spatial proximity to each other, nor are there indications as to the material or manner of hafting. They display marked techno-morphological similarity with those from Phylakopi (Bosanquet and Welch 1904, Fig. XL. 11-17), Ayia Irini on Kea (Cherry and Torrence 1984, 22-23; Cummer and Schofield 1984, Pl. 44. 488-490) as well as of Troy VI (Blegen et al. 1953, Fig. 300). These artefacts attest a form of composite sickle completely different from that of the Neolithic, consisting of regular narrow flint blades set adjacent to one another (Moundrea-Agrafioti 1983). The change in concept of the complex sickle appeared in mainland Greece towards the end of the Early Bronze Age, when the first broad denticulated flint flake elements are encountered, and continued in Middle and Late Helladic times (Van Horn 1976, 1980; Runnels 1985). In the Cyclades corresponding denticulated elements seem to have appeared during the MC period and constitute the few chipped stone artefacts which are a clear chronological indicator: at Phylakopi they appear in the MC layer but are particularly numerous in the LC I phase (Cherry and Torrence 1984, 22). All the sickles at Akrotiri are from LC I contexts. Deeper levels have so far produced no corresponding elements.

The other groups of tools sensa strictu display fewer peculiarities. The piercers are fashioned on blade flakes and are characterized by polymorphism (Fig. 3. 2, 3, 4, 8). They bear the typical traces of wear from alternating circular movement, though without the corresponding specialized points of drills of standardized dimensions. Consequently they bear witness to techniques of drilling the hard raw materials, yet neither their number nor morphology permits us to link them with specific craft activities.

End scrapers are represented by a few but typical examples and are fashioned on irregular blades (Fig. 3. 1, 4). They always bear distal abrasion and are usually composite tools with retouch or use-retouch on the edges.

Finally, in the LC I infills at Akrotiri three projectile points were found (Fig. 3. 6, 7). There are the two familiar types: point with notched base, a type which seems to have been in use mainly during the Middle Bronze Age, and tanged point, known from the Late Neolithic period.

The range of tools described above may be regarded as 'poor' as regards the number of artefacts in each category. It should not, however, be regarded as 'specialized' but as polymorphic. We have no knowledge of the factors which shaped these traits, which we interpret as indices of many, parallel and slightly specialized technical activities.

A second group of tools includes the so-called tools senso latu, or tools a posteriori, i.e. blanks which were used without any prior retouch of their active sections. Included in this group of tools are blanks with peripheral discontinuous use scars and splintered pieces. Only the study of micro-traces can elucidate the degree of participation of waste blanks in tools. Within the perspective of our study, only tools bearing retouch, splintering or use-scars on the sides were examined systematically under a stereo-microscope at low magnification (x60) to locate micro-striations or abrasion on the surface and edges. The micro-traces of use were not, however, catalogued from the use-wear perspective but only for the purpose of investigating the degree of covering retouch and especially scars on splintered pieces by use-wear traces. In the drawings abrasion is denoted by a row of dots, indicating the locus and extent of abrasion. Use causes macroscopic scars at the edges, ends and on the faces, distinguishable from retouch since they are discontinuous and dispersed, i.e. they do not display that regularity of arrangement observed on tools with marginal retouch. Marginal use retouches are observed, mainly in the form of discontinuous removals on one or both cutting edges of the blank, primarily in the form of tiny serrations (Table 10B.1; Fig. 2. 8, 9; Fig. 4. 1-3, 5-6). The removals are frequently accompanied by micro-striations on the faces and abrasions on the sides (Fig. 2. 8). It is of course difficult to determine the frequency of fortuitous removals in this group of tools, i.e. of removals not resulting from use but caused by accidents or post-depositional effects.

Splintered pieces (pièces esquillées) are the second most frequent a posteriori tools. They are considered to be tools created through use of blanks for stable, direct percussion, a use comparable with that of the chisel or wedge (Table 4. 2, 11). According to the extent and localization of the characteristic splintering, different stages of use of these peculiar tools may be distinguished: 1) initial stage of use: splintering confined to the ends of the blank (41.4% of total) (Fig. 4. 11, 13); 2) advanced stage of use: splintering scars cover the faces entirely, rendering the original form of the blank indeterminate (15.5%) (Fig. 4. 12, 16); 3) in the advanced stage of use of splintered pieces in the second category distinctive marginal splintering of burin type are created and the corresponding 'batonet' by-products (Fig. 4, 14). Diverse combinations of extent and localization of splintering creates variations of the basic forms (1) and (2) (43.1% of total) (Fig. 4. 15, 16). Flakes predominate among the recognizable splintered piece blanks, perhaps indicative of the selection of more resistant blanks for use with indirect percussion.

Splintered pieces account for 1/3 of the tools sensa latu at Akrotiri (Table 10.B, 2). If the splintered pieces are classed in the group of tools sensa strictu, as usually happens, they would represent 60.3% of the tools sensa strictu from the settlement. The frequency of this artefact is high in Neolithic, but especially in Bronze Age lithic industries in the Aegean. (The splintered pieces from Early Helladic and Middle Helladic Lerna are mainly of obsidian and represent from 41.8% (MH) to 78.7% (EH III) of the total of obsidian tools in the settlement: Runnels 1985, Table 11.) We are in a difficult position when we try to pin-point the technical activities with which splintered pieces may be associated, since in only a few instances there are abrasions on their dihedral angles, which are assumed to come into contact with the material being worked. We do, however, wish to stress a special kind of technical information which this tool provides and which is related to the chain of production of chipped stone tools: recycling of the raw material. It is clear that a large proportion of splintered pieces are blanks in secondary use (Fig. 4. 11) even exhausted and residual cores which are used with indirect percussion (Fig. 4. 16). We may suppose that this characteristic denotes economic behaviour, perhaps connected with the cost of procuring obsidian.

To recapitulate, we conclude that the tools of the LC I phase at Akrotiri are characterized principally by the small degree of retouch, numerical predominance of a posteriori tools and lack of standardization of typical taxonomic groups. Consequently there is agreement with the picture we have for Bronze Age tools. Tools with peripheral retouch, which may be associated with a variety of cutting tasks, predominate morphologically. The group of blanks with dispersed marginal traces of use may also be associated with similar activities. The most important difference vis-à-vis the range of tools of the Late Neolithic and Early and Middle Bronze Age of mainland Greece is the reduced presence of groups of sickles. This fact is usually attributed to the greater ease of obtaining metal ore during the Late Bronze Age, and so harvesting activities were executed with metal tools. The low frequency of chipped stone composite sickles, as well as the metal sickles found at Akrotiri (S. Marinatos 1969, Fig. 39), seems to confirm this assumption. Flint sickles constitute the most elaborate chipped stone tools in the settlement, yet are so few in number that they cannot be regarded as representative of significant harvesting activities. We note, however, that there are as yet no comparative data on the composition of the tool range of Cycladic settlements, and we do not know whether the sickles of the EC and MC phase are plentiful, as is the case in the settlements of mainland Greece, where however there is a closer association of sickles with flint. Only use-wear study is capable of distinguishing whether obsidian was used for harvesting activities in Cycladic settlements. The remaining tools sensa strictu do not furnish specific information on the existence of systematic manufacturing activities requiring a specialized and morphologically standardized tool equipment.

CONCLUSIONS

The chipped stone tools from the LC I phase at Akrotiri display the typical traits of a lithic industry of prismatic obsidian blades produced by pressure.

With regard to strategies for procuring raw materials we have noted the total dependence of the lithic industry at Akrotiri on imported raw materials. The location of the initial phases of the technical chain of chipped stone tool production in the settlement is suggestive of a mechanism of procuring the raw materials which is not included within a commercial scheme of exchange affecting the morphology of the imported raw material. With regard to techniques of blade production, the morphometric characteristics of unworked products display internal formal diversity which we believe is due to the lack of technical specialization. Even though the prismatic blades constitute specialized products embodying a significant technical investment, they are not fashioned with standardized knapping techniques which could link them to a serial production by specialized artisans. They merely comprise a part of the blank production strategy which is completed by the production of flakes, which are part of the tool repertoire, though less frequent than blades. We believe that the pressure technique may be subsumed by the populace's common fund of basic technical knowledge. The general composition of the range of tools presents a picture of an equipment of generalized and common functions not completed by tools for specialized manufacturing activities. The lithic industry at Akrotiri includes the typical denticulated sickle elements of tabular flint which confirms the temporal association of this artefact with the early phase of the Late Cydadic Bronze Age. We believe that the gaps in this general presentation of the techno-morphological traits of the lithic industry of the final phase of habitation at Akrotiri will be filled with the completion of the study of the distribution of chipped stone artefacts in the inhabited area.

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

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