Assiros Toumba in the 8th and 7th centuries BC:

The apsidal structures of Phase 1 and their function.

(This paper was prepared in 2004 for publication in a volume to accompany the new exhibition in the Thessaloniki Archaeological Museum entitled 'Towards the Birth of Cities'. For various reasons funding was not found for this volume and it seems appropriate to make it available here. Illustrations will be added in due course.)

Introduction

The settlement at Assiros Toumba, in the western end of the Langadas basin some 23 km north of Thessaloniki beside the road to Serres, was first established around 2000 BC and continued to be an important centre in the district until c 900 BC when it was abandoned for well over a hundred years. During this time the timber-framed mud brick houses were reconstructed again and again and the defensive banks which surrounded the settlement were regularly raised and reinforced so that eventually the Toumba reached a height of some 14 m above the surrounding countryside (Fig. 1). Throughout this long period of occupation there are signs of contact with southern Greece – Mycenaean and then Protogeometric – while in the 13th century BC it was clearly a central place providing granaries for a much larger community than that which occupied the relatively restricted summit. By the beginning of the Iron Age the area available on the summit of the mound had been reduced to some 70 x 40 m and it is likely that it was abandoned for a more spacious site after one of the many fires which devastated the settlement at intervals.

Around 750 BC, new buildings were constructed on the summit, but these were of a very different kind to their predecessors. Instead of the regularly arranged rectangular houses of the village settlements of previous centuries, two large apsidal structures were erected across the southern end of the mound while a number of smaller ancillary structures were built in the east and north, leaving an open space – perhaps a courtyard – in the centre. Among the finds were large storage jars – pithoi – and numbers of jugs and drinking vessels as well as fragments of large kraters – mixing bowls – capable of holding nearly 25 litres. The buildings and their context suggest a specialised function within a wider community who lived in the district, either on the large defensive ‘table’ sites such as that at Agia Anna a few km away, or in the small scattered farmsteads or hamlets which can be identified for the first time in the Iron Age. The exact function may never be certain but among the possibilities discussed below are those of a chief’s residence, a focus for social communal activity or a nodal position in the newly developed Iron Age trade networks while the absence of any objects with specific cult associations seems to rule out any primarily religious use.

Whichever is the correct interpretation, these structures reflect the many social changes which were taking place in Central Macedonia from the middle of the 8th century BC onwards, a period which sees the development of the first urban scale settlements as well as the formal colonisation of numerous coastal sites reported in the historical sources.

            Excavation at this site was conducted by the authors for the British School at Athens and the University of Birmingham between 1975 and 1989 with the wholehearted support of the ΙΑ Εphoria of Prehistoric and Classical Antiquities, and its successive directors, Dr. Katerina Rhomiopoulou and the late Dr Ioulia Vokotopoulou and the sponsorship of the British Academy, Alexander von Humboldt Stiftung, Institute for Aegean Prehistory and National Geographic Society of America among many others. Numerous articles have reported different aspects of the finds and the final report on the Iron Age levels will be published shortly by the British School at Athens.[1]

The Phase 1 deposits

The reoccupation level on the summit of the mound was clearly distinguished from those which underlay it both by the character of the deposit and by the different quality of the pottery[2]. All the remains of this period were found in a generally compacted level of dark brown colour immediately below the modern surface (Fig. 2). This level seemed to be composed of decayed daub with, occasionally, the impressions of branches and twigs rather than mud brick. Some of this material seemed to have been hardened by burning, but there were only a few pieces of charcoal or the fire-reddened fragments which might be expected from a widespread destruction fire such as encountered in the lower levels. Undoubtedly humic action, ground water and weathering will have altered the character of this deposit, since it lay immediately below the surface, and in many places formed the surface level with no sign of any disturbance attributable to cultivation. The upper few cm of the stratum were treated as ‘unstratified topsoil’, and finds recorded separately from the remainder, but the division was often an arbitrary one. In places, ‘modern’ finds—cartridge cases and a coin—or finds of later periods such as Hellenistic pottery fragments—had found their way to some depth, presumably as a result of burrowing animals which have made the mound their home since its abandonment, but their presence does not materially affect the basic chronology of this final Iron Age stratum.

The pottery clearly continued the traditions of the earlier Iron Age in shape and decoration, but its distinctive character made it unmistakable and pits and other Phase 1 disturbances which cut into the underlying levels were usually easy to pick up when the sherds were strewn, even if not observed during excavation. The fabric was thinner and harder fired with a porous, even sandy surface. The decorative features such as jug necks, twisted handles and incision were usually much more sharply defined and angular. This pottery development must have taken place during the period when the Toumba was abandoned, but there is little to tell us how long this period might have been.

Although this level covered almost the whole of the summit of the mound, even in those areas where the Phase 2 strata had been eroded away during the preceding period of abandonment, it had been disturbed in many places. At some point the stone footings of the buildings had been largely quarried for use away from the Toumba, leaving only fragments of walling from which it is hard to reconstruct any coherent plan. Even so, it is possible to see that the character of the community was very different from any that preceded it. In the first stage of reoccupation a single large apsidal building dominated the summit of the mound. After an interval it was rebuilt and a second similar building constructed close beside it to the south.

The narrow gap between the two apsidal buildings at first suggested that the summit of the mound must have been crowded with buildings[3], further excavation showed that much was open yard space with smaller, apparently rectangular, structures with no consistent orientation, to the north east. Shattered pithoi were frequent, particularly in the northern apsidal building and in the area of buildings to the north east, suggesting extensive provision for storage.

Whilst two phases of construction could be distinguished in the walls and in the floors of the northern apsidal building, it was rarely possible to observe these elsewhere and it is uncertain whether any of the minor structures belonged to the first phase. It was not possible, even where two levels were found, to see any distinction between the pottery of each. The following account will therefore treat the phase as a single one, except where the evidence for two is clear. There is no obvious variation in the functional character of the pottery from one part of the site or another apart from the concentration of complete vessels in the northern apsidal building.

        The date of these deposits is difficult to fix precisely, since the character of the Iron Age pottery of Central Macedonia is very conservative. There are only two fragments of painted wheel-made pottery and these suggest an eighth or seventh century date (Late Geometric or Early Archaic)[4] while the use of the apsidal building form is typical of the period. No significant pieces of charcoal were found in the deposits of this phase so unlike Phases 2, 3 & 6 where C¹⁴ and dendrochronology have provided accurate dates[5], dating independent of pottery parallels is unavailable.

The structures and finds

Northern area

The narrow trench at the west of this area uncovered a length of wall, a single course high and probably two wide, which seemed to form the corner of a building lying further to the west, with a few stones, probably fallen, to the east of it. Finds from the area around the stones included three spindle whorls and a chert flake. These finds presumably represent a yard surface. In the north east corner of the Toumba a scatter of stones and sherds marks a similar level which was perhaps part of the same open space. Finds from this area included a small cut-away-neck jug (P244), two spindle whorls, a fragment of stone pot lid and two joining pieces of worked stone. Other finds included a quern stone, an almost complete whetstone and a piece of chert resembling a threshing insert. Modern civil war disturbance was suggested by two spent cartridge cases while a live one was found in the topsoil.

The northwestern area

The Phase 1 stratum was found at a lower level on the west of the Toumba than towards the centre, because of the slope of the underlying Phase 2 buildings. Here it was less compacted than elsewhere and the original surface of occupation was indicated by only a few scattered groups of stones, except for a short length of rough stone wall, two courses high and approximately two courses wide. The finds from this open area are a typical mixture of complete and fragmentary objects with no obvious concentrations. They include much of a coarse two handled jar, three spindle whorls, two and a half sherd discs, two fragments of stone pot lids, a piece of a whetstone and of a stone axe, and a fragment of a robust bone awl. The only complete vessel was a miniature coarse pot.

The northern yard

The brown earth was deeper (up to 40 cm) and more compacted in the west of this area than elsewhere. It sloped down from east to west where an irregular row of stones, running north south, indicated a rough (or disturbed) terrace wall with a drop to the west. A second ‘row’ of stones was found at a higher level, suggesting a rebuild, while to the east of it scatters of stones and sherds on two different levels indicated an open area with two phases of activity. In approximately the centre of the area, a roughly rectangular enclosure of small stones belonged to the upper level. The western edge was well defined, but not the eastern. Measuring c. 3.5 m north–south and 1.5 m east–west, it at first suggested the outline of a burial pit, an impression reinforced by the discovery of a bronze pin with a flat disc head (SF 1208) within the rectangle at the same level[6]. Careful examination of the area revealed no sign of any pit, however, nor any bones or objects at a deeper level which might have belonged to a grave. If the rectangle is supposed to have marked the base of a totally eroded grave pit, we have to assume an eroded depth of 50–75 cm on the highest part of the Toumba, which seems rather improbable. No other clue was provided for the function of this structure.

Further east the deposit was thinner. Here there was little sign of the two levels and relatively few stones. The open area was limited to the east by fragments of stone wall belonging to the structures in the north east of the mound. In the south, towards the walls of the apsidal building, the sherd and stone scatter was rather denser and in places helped to define the outer edge of the robbed-out wall.

Finds from the lower level (where it could be identified separately) included one and a half spindle whorls, two fragments of stone pot lid, two fragmentary bone points, a piece of worked chert and one of the rare manufactured pierced clay disc. A similar assortment of oddments came from the upper level: a large portion of a shallow bowl with two horizontal handles (P175), a large part half a stone pot lid, five whole and two fragmentary spindle whorls, two sherd discs and a fragment of another, a pierced stone object and a chert tool. Metal finds included a coil of bronze wire and a fragment, a complete iron ring of uncertain date and a modern cartridge case. Objects from the topsoil included half a spindle whorl, a fragment of pot lid and a rubbing stone.

The northeastern area

The area to the north east of the apsidal buildings may have contained some structures, to judge by the fragments of stone walls and the smashed pithoi, in addition to scattered stones (Fig 3). Only one level was found in this area, either because the structures were erected as the same time as the first apsidal building and remained unaltered, or because later disturbance, root action etc., had removed any traces of the division. The walls appear to have three different alignments and do not suggest a coherent plan for the area, but may have been out-houses or animal sheds erected in a haphazard manner as required.

In the north were two short lengths of north–south wall, a single course in height and formed of small rough stones to a width of no more than 30 cm. Six metres to the east a pithos had been set in the floor but beyond this the stratum petered out towards the Toumba edge. Further south were three more fragments of stone wall, one course high and between 30 and 40 cm wide. Two met at a right angle, while the third, a little to the south west, could once have formed another right angle. All were oriented about 30° away from the alignment of the group to the north west. Immediately to the south east of these walls was a spread of pithos sherds perhaps from a pithos which had once stood in the corner, while a second scatter some 2 m further south included a base. To the south and east the scatter of sherds and stones became less dense, with a short length of wall with yet another alignment and the neck of a pithos.

Clay finds included seven spindle whorls and a fragment, a bead, five complete sherd discs, three halves and two fragments, most of a manufactured disc with two holes and a piece of a pierced object as well as a loomweight fragment and a coarse miniature vessel. There was a single piece of worked bone, a possible whetstone, and another worked stone as well as most of two querns and fragments of two pot lids. There were no large fragments of pottery vessels apart from the pithoi.

The northern apsidal building

The first traces of this building were found in 1977 when floor levels of pebbles and sherds, pithoi shattered on the floor, and the stones of the curved walls which formed its south east corner[7] were found (Figs. 4&5). However, it was only with the discovery of the south wall and south west corner nine years later that it became apparent that these were the remains of a large two-roomed apsidal building measuring nearly fifteen metres in length from east to west and over eight north to south (Fig. 6). These remains however were scanty since erosion and stone robbing over two and a half thousand years have taken their toll.

Once the walls were exposed it was obvious that there were two phases of wall construction (as well as two successive floor levels within the building), though not until the last section of the wall was dismantled was it apparent which was the earlier wall and which the later. When the building was rebuilt and a second similar one added to the south, the orientation of the south wall was changed by about 8°, so that while the new south west corner was above and to the north of the older wall, the surviving stones of the new south east corner were lower and to the south of the older wall. The evidence for the order of construction, when it was at last found, was clear: in the east the south wall rested on the pavement which ran up to the north wall. Here it had been robbed away, to reveal the stones of the lower pavement running up to the earlier wall.

The only other place where evidence for both phases was preserved was in the cross wall closing the apse. The earlier wall, though robbed out when the later was built, was marked by the straight edge of the sherd pavement which had run up to it (Fig. 4) The later wall showed up as a rough alignment of larger stones, insufficiently preserved to allow us to be sure that the alignment of the whole building had changed, even though this is likely.

Of the whole stretch of the north wall, only four groups of large stones survive to indicate its line, and even these would be uncertain if they did not fit so neatly into the building as predicted from the external radius of the apse. The north west corner and much of the west wall is missing, but indicated by the edge of the sherd and pebble scatter. Only two and a half metres of the west wall survive. Here, the external faces of both the earlier and later walls are preserved, usually two courses in height and made with rough stones tightly packed together. In places there were irregular lumps of quartz, a type of stone not used in any other period for construction at Assiros. The inside wall faces are not clear and there was no attempt to prepare a foundation below floor level. The width of the wall is about 50 cm.

The south wall, as already mentioned, is the best preserved (Figs. 7&8), though even this is missing for part of its length in the centre, where only a few isolated stones of the earlier phase survive to show its alignment. The external face is irregular, because what survives is mainly the footing course built alongside the alley which separates the two buildings. Except for the south east part of the apse, the position of the internal face and the width of the wall can only be conjectural. Here large, relatively smooth stones had been carefully laid to form a firm base 90 cm wide which served for both walls. The building was constructed on a slope, presumably that of the ruins of the underlying Phase 2 and 1.5 structure so that the uppermost stones of the new wall in the east were nearly one metre higher than in the west. The floor of the earlier building may have sloped even more.

The cross wall of the apse divided the building into a main room with an internal length of 8 m and an apsidal room slightly longer (5 m) than the radius of the apse itself. In the earlier phase the wall had a width of 50 cm, to judge by the gap in the floor where it had been robbed out, while in the later only four groups of stones mark its course. There is nothing to show where the entrance was.

Nowhere do the stones of the walls stand more than 25 cm above the internal floor level, and the nature of the upper structure is a matter for debate. The dark brown earth, which was more than usually compacted in the area of these two buildings, was very difficult to break up and usually ended up as small lumps too large to pass through a 2 cm sieve mesh. Occasionally these lumps bore impressions of small branches, and sometimes had been reddened by fire. It is likely that at least the roof, and probably a large part of the later building, was constructed of clay-plastered branches. Although a careful search was made, no convincing post pits were found in the underlying levels which could have held the upright timbers needed for such a construction, or indeed for roof supports in the main room where a span of 8 m would otherwise have been needed. One possibility is that the stone walls were no more than low bearing walls for horizontal wooden beams into which the timber uprights were set. Some of the extra stones found beside the south wall may indicate the position of timbers of this kind.

Considering the size of this apsidal building, the remains of its walls seem very insubstantial. There was no heavily burnt destruction debris within the building and no charred wood to indicate the size of the main timbers.

Later disturbance has obviously penetrated the floors in many places since no continuous surface was found, but only small patches of fire-hardened clay or areas of laid pebbles or sherds, together with several shattered pithoi. For the same reason it was often impossible to distinguish the upper and lower floors or to be sure which patches of floor belonged to the earlier and which to the later building. As the floor in the main room alone sloped by over 50 cm from east to west, and the two floors were only separated by 10–15 cm, it is not surprising that this should be the case. The best preserved section of floor was in the north east corner of the main room, where a stretch of nearly 2 m of sherd and pebble pavement was preserved against the line of the robbed out cross wall (Fig. 4). Over this was a rougher surface principally formed of whitish stones. A pithos had been set in the floor against the west edge of the pavement. Another area of cobbled surface was found below a second pithos smashed in the centre of the west end of the room, where an entrance might be expected. Against the centre of the south wall was a hard patch of fire-reddened clay which perhaps represents a hearth area in the later floor, but close beside it was a group of small stones of an underlying wall which must have stood proud of the floor surface.

Because of the disturbed nature of the deposit and the lack of continuous floor levels some finds have been assigned to one floor level or the other on the grounds of the depth at which they were found. Nine vases were found at the level of the lower floor, either complete or sufficiently preserved to suggest they should be associated with the final use of the first apsidal building. Four of these were found in the north east corner of the main room and should probably be associated with the pavement in that area. These include a large handleless storage jar in fine ware with four rudimentary lugs on the shoulder (P177) and a small burnished one with incised decoration (P163 – in situ) a coarse one-handled jar (P390) and a plain cut-away-neck jug (P380). The others, which were scattered more or less at random, included a coarse miniature vessel (P151) and a small cut-away-neck jug (P150), found close together, as well as another coarse miniature vessel (P191) and a small one-handled jar with a lug (P422), and a cut-away-neck-jug with twisted handle (P226).

Other objects which certainly came from the lower level in the main room include, two spindle whorls, a loomweight fragment, and a stone pot lid. The only objects whose position may be significant were a group of pot handles, lying as if they were strung together when they fell to the floor, which may have served as loomweights in the same way as supposed for those found in the cupboards in Phase 2 Room 5. These were found about a metre from the wall in the south east corner of the main room.

Three pithoi had been set in the upper floor in different parts of the main room, and large parts of seven vases were recovered from the upper level which were probably in the room before the walls were robbed out (Fig. 9). These were a cut-away-neck-jug with twisted handle (P165) found with the shattered pithos in the west of the room, three more cut-away-neck jugs near the west (P154) and north (P153, P377) walls, as well as a fine ware jar with slashes around the shoulder (P152), part of a bowl with horizontal handles (P228) and a miniature vessel (P374). Other objects from the upper level were a spindle whorl, a complete sherd disc, a fragment of stone pot lid and a piece of worked antler.

Only one floor was preserved in the apsidal end. No pithos was found set in the floor though pithos sherds were numerous. Large pieces of a wishbone-handled bowl (P193) probably represent a vessel in use, while a body sherd with an exaggerated conical knob (P204) and a fragment of a funnel (P382) like the complete one from Phase 2 Room 13[8] came from the same area. Other finds included one plain and two decorated spindle whorls, as well as an unusual wheel-shaped one, a fragment of a large carefully pierced stone disc which could have served as the ‘flywheel’ for a potter’s wheel, most of a pierced stone disc, half a pierced sherd disc and a fragmentary bone point. Another pot lid fragment was incorporated in the floor.

Sherds from small fine ware bowls, often with wishbone handles, were frequent in the fill associated with both buildings but no example other than P193 was beyond doubt in situ. Large fragments of pottery included a wishbone-handled bowl with peak on the rim (P350) and two fine ware jars (P331, P345).

Other finds from the immediate area of the building, include two half sherd discs, an almost complete bone spatula and two fragments of worked bone, a stone pot lid fragment, a chert blade and a fragment of iron nail. A modern thimble, and a rubbing stone were found in the topsoil in this area.

A neat stone wall was found abutting the west wall of the earlier apsidal building at an angle of about 45°. It was set at a slightly lower level and enclosed a small area of trodden floor surface, which also had two phases. Presumably it represents an addition to this end of the building, perhaps a storeroom. There were pithos sherds on the lower floor, but no other finds of special interest.

The cobbled yard and the alley

At the same time as the first of the north apsidal buildings was constructed, a cobbled pavement was laid to the south and east (Figs. 5&7). Formed with relatively small stones, it incorporated large amounts of sherd. In places this surface clearly underlay the walls of the south apsidal building and the new north one, but elsewhere the new construction had disturbed it and it did not form a continuous surface. An area of cobbling found within the apse of the south building may be part of this surface, rather than of an internal floor, but it is impossible to be certain. The pavement was removed when the northern building was constructed but the cobbling was certainly patched and renewed on many separate occasions, while bone and sherd and other debris collected on its surface in large quantities. Excavation ceased a metre to the east of the building so we do not know how far these cobbled pavements may have extended. They were not found to the north east in the first season of excavation and may not have extended very far from the building.

Only the latest pavement surface could be traced the whole length of the alley which by now also served as a drain gully between the two apsidal buildings (Fig. 7&8). It was set below the level of the walls and sloped even more steeply from the open area in the east to the end of the south building in the west. At the west end it sloped away even more sharply and seems to have been laid on a surface which cut away the underlying destruction levels of Phase 1.5 and 2. At the very edge of the trench it was difficult to separate the stones of the pavement from those of walls below and it seems to have petered out in this direction.

Finds from the make up of the lower floor included a spindle whorl, a partly pierced sherd disc, half a pierced stone disc and a chert flake. Stone objects found between the pavements and from the make up of the upper one included a broken mould, a fragment of whetstone, an axe and a piece of a rubber. A fragment of bronze pin is probably ancient, while a Turkish pipe end is clearly intrusive. In addition to the usual range of clay objects, a pierced weight or pendant, a spindle whorl, and half a sherd disc, there were two fragments of what seems to be the pierced floor of some kind of kiln, though the nearest parallel known to me for the thickness of the pieces and arrangement of the holes is to be found in the curious Late Bronze Age cooking stands of found at Thermon[9]. Another curious item was a leg fragment from some kind of tripod bowl (P454).

Finds from the surface of the upper pavement or immediately above included two spindle whorls, a sherd disc with double piercing, a piece of stone pot lid, a chert flake, and a lead pot-mend. The function of a banana-shaped clay object is uncertain while the 1966 drachma coin was perhaps dropped when the Toumba still provided a venue for May-day picnics for the Assiros villagers. The most striking find, however, was an exceptionally large wishbone handle bowl with horse-head protomes, or rather the majority of its fragments, shattered on the surface and trodden into it (P178). The same area produced one of the very few sherds which can be related to the south Greek sequence—a linear cup rim of a shape not used before the eighth century BC (P192)—as well as two ordinary cut-away jug necks with twisted handles (P227, P367).

Objects from the cobbling within the apsidal end of the south building, but probably predating it, included a spindle whorl, one whole and two half sherd discs, fragments of a bone point, a loomweight and a stone pot lid as well as three stone tools.

The southern apsidal building

Much less of this building was preserved. The apsidal end is represented by less than 2 m of curved ‘wall’ made of quite small stones. Without the pavement running up to it, we could not have been confident that this too had been an apsidal building. The central length of the north wall was represented by a single course of large blocks, and often by only single stones, although extra stones found to the south of this line may indicate the position of timber uprights which supported the roof. The north west corner was the best preserved with two courses of stones to the outer face beside the alley with a second single course at the higher level for the inside face. The west wall was represented by just over a metre of small stones forming its outside face. Like the building to the north, this also sloped nearly a metre from east to west. Of the remainder of the building there is very little trace. Moderate sized stones were scattered across the area, perhaps the remains of the other wall, while at a slightly lower level were two post-holes in the line of the west wall. One was about 3 m from the north west corner, and a second, with three stones set neatly around it, 2 m further south. If they belong to the building, which their position suggests, they could define a doorway in the centre of the west wall. Considerable care was taken to search for any other post holes in the underlying level which might also have belonged to the building. Two probable post settings found at this level in the east end of the building need not belong to its construction. No trace of any cross wall was detected.

No floor was found within the building and it had probably been at a slightly higher level than the preserved foundations. Two pithos bases found in what should have been the east end of the main room suggest that the function of this building may have been similar to its north partner. There were no whole or nearly complete pots from the Phase 1 stratum in this area and none of the few finds can definitely be associated with the building. The sum total is one spindle whorl, a large piece of sherd disc, most of a whetstone, and a fragment of stone pot lid.

The southern area

To the south of the remains of the apsidal buildings, the Phase 1 stratum was featureless and it is impossible to tell whether any of the brown earth deposits containing the distinctive pottery of this phase were in situ, or whether they resulted from the erosion of the remains further up the slope. Only at the very edge of the area of excavation was a short length of stone wall, a single course in height, discovered, which may represent another structure of this phase, although the pottery found in the vicinity was undiagnostic.

Pottery

            The pottery from the Phase 1 levels clearly continued the traditions of the earlier Iron Age in shape and decoration, but its distinctive fabric and character made it unmistakable[10]. Pits and other Phase 1 disturbances which cut into the underlying levels were usually easy to pick up when the sherds were strewn, even if not observed during excavation. The pottery was almost entirely hand-made and was thinner and harder fired. Fine burnished wares are rather less lustrous than previously. The coarse gritty ware found in every earlier level was largely replaced by a thinner fabric, often with a porous, even sandy surface. The decorative features such as jug necks, twisted handles and incision were usually much more sharply defined and angular. The impression is gained of a greater use of a knife on the clay rather than the smoothing rag. The edges of cut-outs are raw, not rounded, while twisted handles and jugs rims have sharp, angular profiles. This pottery development must have taken place during the period when the Toumba was abandoned, but without other excavated sites in the immediate vicinity there is nothing to tell us when the change took place.

            At least fifteen restorable vessels, apart from pithoi, represent pottery in use in the northern apsidal building at the time of either the earlier or the later destruction. Another three came from the cobbled surface of the eastern courtyard and the alley while only three more came from the whole of the northern area. Although this is a small sample in comparison to that from Phase 2 it seems sufficient to indicate the popularity of specific shapes at this period - though clearly this should relate to the specific function of the apsidal buildings. Cut-away-neck jugs (9) and wishbone-handled bowls (6) remain popular shapes, the former perhaps even more popular than earlier, to judge by the number of sherds. Jars with collar neck are still standard but the four-handled jar is missing. One-handled jars are present but in a variety of forms and miniature vessels are perhaps more common, relatively, than in Phase 2 (Cf. Fig. 10).

Fine ware

For the first time at Assiros, fine ware was equally common as or even more common than coarse. The predominant fine ware shapes of the Phase 1 assemblage remain wishbone-handled bowls and cut-away-neck jugs, just as they had been in the Bronze Age and the earlier Iron Age. Two varieties of wishbone-handled bowl can be distinguished. There are large examples with a single handle or a pair of handles rising high above the rim and smaller bowls with a pair of handles which extend horizontally from the rim. The first type is clearly the descendant of the Bronze Age form and the largest probably have two handles, as does the enormous example found shattered on the pavement to the east of the northern apsidal building (P178) which measures nearly 50 cm across the rim and has a capacity of nearly 25 litres. The majority probably had only one handle (P193) which can have a simple flat or hollowed (P282) end. Other handles were elaborated into schematic horse heads, of smaller size as known from Phase 2 (P25) or larger (P178). The handles on the second form are angular with triangular or circular cut-outs. In some cases the wishbone form is scarcely recognisable. Unfortunately three of the bowls which were probably in use at the time of the destruction have no indication of the handle form, whether vertical or horizontal, they once had (P295, P375, P301). Three more bowls can also be associated with the Phase 1 floors (P178, P193, P350).

            Decoration on these forms of wishbone-handled bowl usually takes the form of slashing on the rim and either side of the handle(s) or the shallow grooving well illustrated on P178. One fragment of a large and elaborate handle (P452) has fine herring bone incisions in addition to the normal slashing and another has a row of neat jabs along the edge of rim and handle (P175) and the unusual feature of vertical pattern burnish on the interior. Other bowl types can be recognised. One is similar in shape to the wishbone-handled bowl with angular horizontal handles, but has a more or less rounded handle along and level with the rim (P228 from the upper floor of the northern apsidal building,). A second, represented only by small sherds, (and easy to confuse with Early Bronze Age forms) has a small pierced or unpierced ledge lug at the rim of a shallow splaying bowl. The capacity of these different bowl types ranges from 1 to 3 litres (P193, P228).

Sherds from other varieties of bowl include one with an attached spout and hole pierced through the rim and one with thickened everted rim. One fragment represents a shallow dish, with a diameter as reconstructed of 60 cm. Even though distortion of the rim may have exaggerated this, it was still an exceptionally large example of a type not recorded so far in the Macedonian Iron Age. Another unique piece is probably from a tripod bowl (P454). This has a single ? leg attached to the side of a conical bowl with flat-topped rim. The leg is carried up to the rim where a slight lug has been formed. One possible dipper was recognised (P379). This had a flat-topped rim and round handle rising above it.

            The cut-away-neck jug is well represented in this phase with large parts of nine examples surviving on the floor levels. They range in size from P244 which is c 10 cm high with a capacity of 0.2 litres and a simple sloping rim to others which must have been c 35-40 cm in height with a capacity of c 3-4 litres when complete (eg. P165). The majority have sharply angled rims which in some cases are pinched in to form a trefoil shape.

            Although nine in ten of the vertical handles in the sherd sample are plain, only three of the ‘complete’ vessels have plain handles (P244, P377, P380). Not enough of the handle of a fourth is preserve to determine its form. Two more (P165, P367) and two large neck fragments which probably represent vessels in use at the time of the destruction (P154, P227) have a very distinctive facetted angular twisting applied to the upper part of the handle. This is regularly found on fragments (P11) and some of the handles reused as weights which were found on the lower floor of the northern apsidal building. One large neck and handle fragment (P79) has twisting which extends down the whole length of the handle in the manner more typical of Phase 2, and a small knob at the highest part. Similar smooth twisting can be seen on a small fragment while another has a handle with square section which tapers upwards. One handle found in the handle group could belong to a jug but unusually it is set below rather than at the rim. The bases of these cut-away neck jugs may be flat or slightly hollowed (P165, P379), a feature not seen in Phase 2 levels.

            Decoration apart from the twisting is rare and usually carried out with fine incision or slashing on the rim handle or shoulder. Patterns formed include pendant triangles. Grooving is used to emphasise the rim of P79 while one handle fragment has punched dots along the rim and another neat grooves forming a tree pattern on the outer surface.

Very few sherds can be assigned with certainty to the four handled jar which was typical of both Bronze Age and earlier Iron Age levels. One large fragment with thickened rim, nipple on the shoulder and a row of neat jabs just above it, could belong to this type, as could also a thickened rim sherd, a flattened vertical handle in the ‘handle group’ and a large body sherd with a nipple.

            Jars with collar necks and two horizontal handles are represented by three more or less complete examples. The neck may be splaying (P345), everted (P152) or sharply angled (P331). A large fragment, probably from this kind of vessel, has a slashed rim and a flattened knob on the shoulder (P196). One of the reused handles with a very sharp-edged twist is from a large closed vessel, perhaps of this type. An exaggerated conical knob (P204) is also most likely to belong to a large jar of this type.

            No example of a cantharos was recovered from this Phase, but the presence of several handle fragments suggested that it was an usual form. Three of these represent the pistol grip type, either plain or grooved, while two plain high swung handles could either belong to this shape or to high handled mugs.

            Funnels similar to that found complete in Phase 2 were identified amongst the sherd material. One piece was large enough for this to be certain (P382) while two more in fine ware could also be from pedestal bases or lids as could a coarse ware sherd. Another fragment was probably from a small lid.

Coarse ware

            Although about half of the sherds counted in the sample (below) were of coarse ware, only four complete vessels were recovered, and only two others which can be restored on paper. Thus it is difficult to know how far the coarse ware shapes familiar in Phase 2 have gone out of fashion, though certainly new shapes were introduced.

            Three vessels from the northern apsidal building can be classed as one-handled jars but all are quite small and lack the everted rim typical of Phases 4-2. P380 has upright sides and only a slightly everted rim. P422 has a relatively narrow mouth and an angular nipple just below the rim opposite the handle. The smallest of the three (P163), with a relatively well burnished surface has slight grooving on the handle to simulate twisting and a row of jabs around the base of the handle. There are similar jabs around the shoulder with an irregular horizontal groove on either side. A rim fragment which could be from this type has neat impressions on the rim and an additional row lower on the neck. The two-handled jar with everted rim and handles set on the shoulder is represented by fragments only.

Several fragments from a cooking pot with stand were found in the northern apsidal building, but it is not clear whether this was part of its furniture. It differs little from its Bronze Age predecessors and may reflect a tradition continuing from that period, although no complete example was found in any of the earlier Iron Age phases. A large necked jar with four knobs on the shoulder but apparently no handle (P177) was found on the lower floor in the same building. This was made in a coarse but well burnished fabric and has quite a deep narrow neck and a simple flat base. Two pierced sherds represent strainers. One is an upright rim with a diameter of c 24cm and the other a flattened base.

            Several miniature vessels were found in deposits of this phase, all in a rather coarse clay. Two are complete miniature jugs (P150, P374), another is an accurate version of a one-handled jar, complete with slashed rim (P151) Only the lower body of the fourth, a deep jug or jar, is preserved (P191). A large fragment of another of these tiny vessels (P14) comes from a straight sided-jar, but need not have been complete at the time of the destruction.

            A number of sherds with a distinct shoulder seem to belong to the ‘Thasos’ class[11], though it is not clear whether these belong to this phase or are survivals from the underlying levels. One is a rim but the others are only body sherds from necked bowls or jugs. Channelled ware is not found in this level and presumably no longer made[12].

Wheel-made grey ware

Wheel-made pottery is nowhere common in Iron Age Macedonia, but at Assiros it seems much rarer than in the Axios valley, for example[13]. 236 feature and body sherds in grey ware were recovered in total from the Phase 1 levels, more than for all the previous phases together. Most of these sherds were from carinated one-handled cups with horizontal grooving though a few came from closed vessels.

Pithos

            Pithos sherds are common, especially in the main room of the northern apsidal building and in the yard area to the north, but only a few bases were noted in addition to the two pithoi in situ on the lower floor and the three on the upper floor. This makes it unlikely that that bulk storage was a primary function of this complex. The pithoi have the tall narrow ovoid shape typical of the period with peg bases and heavy, flat-topped rims while handles and decoration are both very rare.

Sherd sample

The following observations are based on an assessment of a sample of the sherds from 88 baskets from 7 different contexts in this phase. Well over 10000 sherds were counted, of which 12% were features or carried decoration. The areas selected for this sample include three well defined contexts: the interior of the northern apsidal building, the abundant material trodden into the surfaces of the alley to the south and the cobbled yard to the east. The Phase 1 levels immediately to the south where the southern building had once stood and to the west and part of the yard area to the north have also been selected as part of this study. Although these units must contain some earlier pottery which is not immediately recognisable (unlike Mycenaean, for example) no attempt has been made to exclude burnished wares which seem to have more in common with those used in earlier periods.

            As already noted, in contrast to the earlier Iron Age, the pottery of this period, both coarse and fine, is thinner walled and harder fired often with a rather sandy fabric. As a result it is often brittle and fragments into many small pieces. Coarse ware is less frequent than in earlier periods so that coarse and fine body sherds are present in approximately equal numbers in each of the contexts. The relatively large proportion of pithos sherds from the interior of the northern building reflects the fact that two pithoi had been set in the lower floor and three in the upper. Crushed and shattered, their fragments were spread throughout the building.

            Among the coarse feature sherds were a surprisingly large number of fragments of cooking stands (15%). As usual, simple rim fragments form the majority of the features (37%), and another 8% were clearly everted. The use of slashing or finger-impressed decoration was slightly more common than in Phase 2, accounting for 1 in 8 of the coarse jar rims. Few rims could be specifically assigned to the one-handled jar and only one from the two-handled version. The frequency of detached handles, principally vertical, indicates, however, that coarse jars with vertical handles are still a regular part of the repertoire. Base fragments of the simple flat form are well represented, while flattened and raised flat types are less common. Nipples and finger-impressed decoration (apart from that used at the rim) are perhaps a little more common than in Phase 2 though the numbers of examples are very small (6 and 16 respectively)

            In the fine ware, bowl rims remain the most frequent type (>27%) and most of these should come from wishbone-handled bowls. Only a few carried the handle scars on the rim but handles themselves were a little more common than in Phase 2 and may reflect the resurgence in popularity of the two handled form. Occasional examples of bowls with basket handles were also noted. Cut-away-necked jugs remain a popular closed shape (8% of features) while collar necked jars and 4-handled jars are still be standard (>5% each). There are, naturally, quite large numbers (6%) of tiny rim fragments from jars of different types but few are large enough to measure their diameter and thus gain a fair indication of the shape to which they belong. There are also three examples of cantharos jars with high handles. Detached handles are well represented and as usual vertical types predominate. One in ten are twisted, usually with a well defined angle to each facet of the twist. Bases are particularly common (18% of features). These are predominantly flat but there are also smaller numbers of flattened and ‘raised flat’ types as well as a single example of a pedestal base, though curiously none of the hollowed type represented in the whole vessels.

            Pithos sherds in this sample (excluding the sherds of vessels broken in situ) accounted for about 10% of the total number of sherds, but included only three bases. Stray Mycenaean and comb-incised Bronze Age sherds are still surprisingly frequent (four in a thousand and one in a thousand respectively), but this reflects the number brought up to this level through pit digging. The wheel-made grey class is still rare but perhaps a little more common than previously (> four in a thousand). Iron Age incised sherds remain very rare.

Relationship to other sites

The range of pottery in this phase at Assiros, unsurprisingly, matches that from Saratse (Perivolaki)[14] and is broadly comparable with that from Schichten 7-4 at Kastanas, assigned to the 8th and 7th centuries by the excavator[15], but no precise match seems possible. In the same way there are similarities to the pottery from the upper stratum of the Toumba and the lowest of the Table at Vardaroftsa (Axiochori)[16] and other central Macedonian sites. The pottery recovered from the Pre-Persian levels at Olynthos has rather different characteristics, notably long knobbed handles[17]. At Vergina cut-away-neck jugs and bowls with rim handles are both characteristic[18] but other features there are unknown at Assiros. At all these sites, regional preferences are just as important as chronological differences in determining the character of the late Iron Age pottery at each. The limited number of wheel-made grey sherds match the whole vessels published from the Chauchitsa cemetery[19] and found at other sites in Central Macedonia still unpublished.

Other finds

Among the most common finds in this phase, as in every other at Assiros, were the ubiquitous clay spindle whorls, mostly biconical and without decoration. Sherd discs were also frequent and many have served a similar function. Apart from the group of reused handles found in the lower level of the northern building only a few loomweight fragments were found. There were fragments of stone pot lids, although only one was a much as 50% complete, together with a few examples of stone tools, such as axes, whetstones and chipped chert blades, which are probably residual. There were a number of items of utilised stone but only a few fragments of quernstones. Antler handles as well as broken bone tools were found including several robust awls but bone from these superficial levels was much less common than in previous phases.

This range of finds is similar to that found in earlier Iron Age levels at Assiros and published sites such as Vardaroftsa (Axiochori)[20] and Kastanas[21]. It seems consistent with ‘domestic’ activities, but has nothing in common with the range which might be expected at a sanctuary of this period. The only unusual items are the bronze pin which is difficult to parallel, and the stone ‘potters’ wheel and fragments of kiln floor, (if correctly identified). As often, however, no trace of either the kiln or the atelier has been found. Good clay is available locally but wheel-made pottery (the grey ware described above) was only made in very limited quantities during the Iron Age.

Architectural Parallels

The apsidal house form is now quite well known from Iron Age contexts (as well as earlier Bronze Age ones)[22], but debate still continues as to whether it had any specific function, or was simply the most convenient way of constructing a timber building with a large single main room. Some examples have been assumed, with some justification, to have a specifically religious character, such as the ‘Bay Hut’ at Eretria[23] or the apsidal building at Perachora[24] which match closely the temple models found at Perachora[25]. The location of another apsidal building of ? early 7th century date in the area of the Artemision at Thasos[26] may indicate a cult use although there is no specifically votive deposit associated with it. At Poseidi near ancient Mende in Kassandra, Ioulia Vokotopoulou discovered an apsidal, unroofed enclosure containing votive deposits with pottery dating back to the Mycenaean period, which was succeeded by an apsidal temple[27]. In contrast recent excavation at Thermon has shown that it is unlikely that the ‘apsidal peristyle’ first detected by Sotiriades, formed part of an apsidal precursor to the Archaic temple, though its date probably falls within the later Iron Age[28]. Examples of this form of building at Nichoria from the 10th and 9th centuries BC[29] do not seem to be for cult use while that found at Antissa on Lesbos is equally unconvincing as a temple[30]. The most discussed building of this form is the ‘Heroon’ at Lefkandi where the picture is especially complicated because of the difficulty of relating the use of the building to the burials within it[31]. At Assiros there is a fragment of another, smaller apsidal building from earlier in the Iron Age (Phase 3)[32], in addition to those described above while at Toumba Thessalonikis the principal building in Phase 4 is apsidal in form[33]. Recent finds of Late Geometric oval buildings at Skala Oropos associated with iron working and other manufacturing processes[34] remind us how little we know of standard architectural forms during the ‘Dark Ages’. They find parallels in Epirus at Vitsa[35] and in Anatolia at Old Smyrna[36].

The Assiros apsidal buildings (Fig. 11) are large (almost 15 m in length) and it is difficult to make any precise architectural comparisons. No certain trace was found (despite a careful search) of the internal or external posts observed at both Nichoria and Lefkandi. The internal cross walls of the northern building at Assiros are as natural a feature of the structure as they were in the Middle Helladic period.

Function

These structures stand out in terms of their commanding position on the summit of the ancient mound at Assiros, set rather surprisingly across, rather than backing into, the direction of the fierce Vardaris wind which blows so often in winter from the north west. Their size and their isolation, without any surrounding community, also suggest a special significance. Their location contrasts markedly with the typical major settlements of the period, which are usually on the spacious but low flat topped ‘tables’ with steep, readily defensible perimeters such as that at Agia Anna or Krithia a few km away or those currently under excavation at Anchialos, Toumba Thessalonikis and Karaburun for example[37].

The storage pithoi seem to suggest a domestic or ceremonial rather than a cult function (particularly as they are placed in such a way as to reduce the available area of the main room) but are not numerous enough to indicate bulk storage at this period. There is certainly no sign of any of those cult-related items such as figurines or offerings of jewellery (except the bronze pin) so typical of the sanctuary sites of Southern Greece.

It is not possible to determine if the location of these buildings is related in any way to the presence of the underlying pithos burial[38]. This was set into the ruins of the Phase 2 and 1.5 settlements at some point during the 150-200 years which intervened before the reuse of the mound for the apsidal building but its date cannot be given more precisely. The only object associated with the burial was a simple bronze button of the type frequent at Vergina. The burial does not give the impression of being that of an individual of any particular status or prestige – though of course this could have been demonstrated in ways which have left no trace.

The presence of so many smaller pottery vessels in use in the apsidal buildings and their immediate vicinity is intriguing, though they are on the whole unremarkable, with one important exception, the very large dark grey bowl (P178) found smashed on the pavement to the east of the buildings. Both the size and elaborate horse-head protomes suggest that this was a prestige item rather like the kraters which certainly occupied pride of place at archaic symposia or those of Mycenaean style introduced to Macedonia in the 13^th century BC[39]. The fact that the most common shapes are in fine ware, and include small amphorae, jugs and bowls of a suitable size for drinking, is probably significant.

One possible interpretation is that these substantial structures at Assiros were built by a local chieftain, away from the village settlement(s?) which he controlled, as a residence and ‘audience chamber’, where one may imagine a Homeric bard might have entertained the chief’s guests as they drank. Such activity would help bond the community to a leader who was able to dispense entertainment or privilege. The remaining buildings, constructed on a less grandiose scale, could certainly not have housed any large community and probably served as ancillary structures for ‘retainers’, for storage for the household or for keeping animals—or perhaps all of these at the same time. The lack of any restorable vessels from this part of the site contrasts with the number from the apsidal buildings and the cobbled yard around them.

Another interpretation could be that these buildings served the elite members of equal status in a larger community as ‘neutral’ meeting places for ceremony, symposia (as indicated by the kraters and drinking cups) and ‘political’ discussion. The size of such an early aristocratic community is, of course, hard to determine but it is of interest that the survey work conducted by the University of Thessaloniki in the Langadas Basin[40] has discovered traces of small outlying Iron Age settlements, which did not occur in the Bronze Age. These suggest a hierarchy of settlement and reinforce the impression that the Toumba at Assiros may have served as a central place for a short period during the later Iron Age. The discovery of a cemetery in the district could, of course, provide important information about the extent of social differentiation or ranking, but remains unachieved. Such cemetery evidence as is available from Vergina might well lend credibility to the idea of local ‘aristocracies’ since the tumuli there often have an original, central burial which in some cases is also the richest in terms of grave offerings. Neither here, however, nor at any of the other cemeteries whose excavation has been reported in recent years but are still awaiting publication, is there any sign of the small number of outstanding burials which could suggest a small number of preeminent individuals to whom one might accord ‘royal’ status.

A further possibility which should be considered is that this site served an economic role in the developing trading relations between neighbouring and more distant communities. When long distance transport was still dependent on pack animals, as was the case until at least the end of the nineteenth century, Assiros village regularly served as the second overnight stop on the one of the main routes inland from Thessaloniki to the interior and beyond to the Strymon valley (and, ultimately, Bulgaria and the lower Danube basin) – a Han. While the level of economic development in late Iron Age Macedonia is still largely a matter for speculation, there is no doubt that this was a period which saw, on the one hand, a population increase and, on the other, the first intensive southern Greece influence through trade and colonization. Both these indicate a level of resources which continued to attract both economic and political exploitation for several centuries. Both suggest impetus which encouraged, perhaps for the first time, an extensive network of economic links between the coastal ‘colonies’ and their rich hinterland. Such a network could well have led to the development of regular ‘trade’ routes and, with these, appropriate venues for rest and refreshment which were not always located in settlements. Assiros Toumba, located beside an almost perennial stream, would have been a logical place for watering animals even after its abandonment as a settlement and could have developed into such a prehistoric equivalent of a Han.

By the sixth century the result of this interaction was a string of prosperous urban communities along the coast of Macedonia and Thrace, whose material culture shows the extent to which southern Greek products were imported and locally imitated[41]. Excavation of parts of the Archaic settlements and cemeteries at such sites as Pydna, Anchialos, Sindos, Toumba Thessalonikis, Karaburun, Agia Paraskevi, Sani, Mende, Toroni and Akanthos for example, already clearly illustrates this cultural amalgam. It is probable too that the process of urbanization had already started in some parts of the hinterland, but this is a field for future research since little is yet known about the character of the settlements or the extent to which the social structures or material culture developed on the coast spread to the interior. Until this is done the apsidal buildings at Assiros (and the finds in them) which belong to the early stages of this process are hard to set in context. It is still too early to know whether their function was administrative, ceremonial or economic – or a combination of these. It is equally uncertain whether the complex originates in local Macedonian traditions or is a reflection of the steadily increasing influence of the colonies founded from Euboea and other southern Greek centres – an influence in terms of ideas and technologies, as well as objects, which dates back in Central Macedonia at least to beginning of the Proto-Geometric period.

K.A. Wardle, Diana Wardle, Institute of Archaeology and Antiquity, University of Birmingham

k.a.wardle@bham.ac.uk