Geophysical survey (TJH, DLS)

The north-east quadrant of the North Hill was selected for study because it offered an opportunity to determine whether the city grid identified by Robinson continued northwards, and to identify a house for excavation that was comparable to those he investigated. The area covered by geophysical survey in 2014 extends at two points as far as ‘Avenue B’: south-east of Block A viii, and approximately one block north of street ix (Figs. 4 and 5). This provides the Project with a clear link to identifiable features belonging to the Robinson ground plan. At the same time, earlier excavations had left this neighbourhood comparatively unexplored, with the exception of a long narrow trench, a small U-shaped trench, and several cuts across the edge of the hill where a search for urban defences had been conducted. Thus, it offered an excellent chance to gain new knowledge about the city.

Fig. 4.

North Hill twin-probe earth resistance data, after the application of a high pass filter and interpolation, shown overlaid upon the Olynthos Project contour map created from drone orthophotography (TJH, JM, HP).

Fig. 5.

North Hill magnetometer results after sensor destriping and interpolation, shown overlaid upon the Olynthos Project contour map created from drone orthophotography (TJH, JM, HP).

The geophysical survey of the North Hill in 2014 included 21 magnetometry grid squares (1.89 ha) and 15 earth resistance squares (1.35 ha). Despite lacunae in several squares due to dense vegetation cover, both techniques successfully detected a regular grid pattern of streets and blocks of houses, continuing the layout found in Robinson's main excavation area. In general, however, the resistance results provide a much clearer picture of structural elements (Fig. 4). Approximately 17 houses appear to be preserved in almost complete fashion in the resistance results. Another 15 houses appear in partial fashion around the edges of the area surveyed. Many high-resistance anomalies can be interpreted as individual walls, and compacted floor layers are suggested by distinct contrasts in the measurements obtained within buildings. Most houses appear to be units of c.290 m² with the traditional courtyard and pastas arrangement characteristic of the domestic structures excavated by Robinson on the North Hill. One example, B x 5 and B x 6, may be twice as large, possibly combining two houses joined across the stenopos (rear alleyway). It is difficult to reach a firm conclusion without excavation, but such a union would not be unprecedented, as A v 6 and A v 8 were united at some point during their occupation.

In the northern part of the survey area, close to the edge of the hill, the regular grid pattern appears to end in Block B xii. It seems that only the southern half of this block was built. To the north, in the 50 m or so between houses B xii 8 or B xii 10 and the north-eastern corner of the North Hill, there lay an open zone with only a few detectable features. On the edge of this zone Robinson looked for defensive perimeter walls. A fortification wall is visible as a dark high-resistance line close to the edge of the hill, with Robinson's trenches (white, low resistance features) cross-cutting it. (The wall was also investigated in the Olynthos Project's trench [TT]04 – see below.) The largest feature within the open zone is possibly a second perimeter wall, stretching for 50 m at an angle of roughly 70 degrees, that is also indicated by a dark high-resistance line. A series of small walls also may be identified within the open zone, but their alignments are not as regular as those on the grid plan. Some of these walls were investigated in TT03.

The magnetometer provided results that complemented, but frequently did not duplicate, the resistance data (Fig. 5): this is because magnetic signals very likely corresponded not to walls but to dense deposits of collapsed or abandoned ceramic materials (tile, pottery) between walls. The most intense magnetic signals are consistent with those produced by the remains of kilns, furnaces and large, intact storage jars (pithoi). Two of these signals are visible in the open area at the north end of the North Hill; they are dark oval features ringed by a white halo. Magnetic signals also clearly indicated roads and stenopoi (alleys) in the southern region of the geophysical survey zone. The magnetic features appear as a grid, or tic-tac-toe board. They may be caused by the use of broken ceramic materials as fill in the makeup of road surfaces or alleyways between houses. The collapse of tile roofs into these spaces is a phenomenon noted in the excavation of house B ix 6 (see below).

One area where the geophysical responses are seen to vary is on the eastern side of the southernmost area of the survey. As work proceeded here it was noted that there were significantly more rocks in the topsoil than had been encountered elsewhere, and this made collecting resistance data quite difficult. The magnetometer data also reveal fewer anomalies in this area, although the pattern of roads is still visible in both magnetometer and resistance results. The results of the 1.0 m resistance data suggest that the shallow layers are more disturbed in this area; however, the reasons for this are not clear at this stage.Footnote ⁷

To the north of the North Hill, where only magnetometry was conducted, the results are dominated by negative responses due to the recent irrigation channels cut in advance of new tree planting. Broader underlying variations are likely to be associated with geological variations, but few anomalies of archaeological origin are visible. A cluster of discrete positive responses towards the eastern end of this area may indicate a small number of intact sub-plough zone anthropogenic features, but nothing like the density of occupation on top of the hill is represented in the data. An intriguing linear negative response is visible immediately to the north of the northern tip of the hill. It is possible that this is natural, perhaps due to more shallow bedrock in this area; however, an anthropogenic origin cannot be ruled out. Such a negative anomaly could indicate the looser fill of a trench, perhaps an unreported excavation, or the presence of different soil material brought in, possibly to make a ramp.

Testing the Geophysical Results: Trial Trenches TT02, TT03, TT04 and TT06 (ZHA)

Six trial trenches were opened in order to verify anomalies identified in the geophysical survey and to explore selected areas in more detail. Two (TT01, TT05) were opened in the south-western and northern parts of one house (B ix 6, described further below). The remaining four were intended to explore other significant anomalies.

Excavation in house B ix 6 (LCN with BAA, KML, ES, SMS and ST)

The geophysical survey and the excavation of trial trenches facilitated selection of a single building for detailed investigation. The criteria for the selection were that (according to the geophysical results, interpreted in the light of the trial trenches and legacy data) it should appear to be a house, and that it was laid out in a manner that seemed reasonably typical of those around it; that it should be relatively well preserved; and that it should be similar in layout to the houses excavated by Robinson, so that the detailed information collected about the use of space within would have the greatest chance of being relevant to understanding the legacy data. For this last reason it seemed desirable to select a house on the southern side of an insula, as these tend to be organised in a more standardised manner than those on the north. Following the excavation of the two trial trenches in 2014 in a building corresponding, using Robinson's numbering system, to B ix 6, this structure was selected for complete excavation. One of those trial trenches (TT05) yielded well-preserved destruction deposits, even though the stratigraphy was relatively shallow.

B ix 6, and the trenches in it, are still in the process of excavation (Fig. 9 and Fig. 10). The discussion below briefly summarises the state of our knowledge of the building following the July 2016 field season and demonstrates the way in which our approach is producing a fine-grained picture of various aspects of the house as a structure and as a socio-economic unit. Since in most instances only the most recent occupation phase has been reached, the discussion proceeds spatially rather than chronologically.

Fig. 9.

Schematic plan of the walls, spaces and selected contexts in house B ix 6. Room numbers referred to in the text are shown in bold lower-case letters. Numbers prefaced by ‘TT’ indicate Olynthos Project trenches (Filippos Stefanou and HP).

Fig. 10.

Orthophotograph of house B ix 6. Wall socles, courtyard, stenopos and other features are visible as stone cobbles within the square excavation trenches (Filippos Stefanou).

In excavating B ix 6 our intention has been to study the patterns of domestic activity in and around the building, combining a variety of independent lines of evidence to reconstruct them in as much detail as possible. Questions we aim to address include: what foodstuffs and consumer goods were available to the occupants, and from which sources were they obtained? How and where were different domestic tasks performed? What uses were made of upper-storey space (which has rarely been identified and studied in Greek contexts)? How much did the use of space change over time, and in what ways? To what extent can the activities of different groups (such as men, women or children) can be detected through the archaeological evidence? A first step towards addressing these questions is to achieve an integrated picture of the different components of the archaeological record in each part of the structure, looking at architecture alongside artefacts, ecofacts and the results of scientific analysis. For this reason, rather than discussing the different specialist studies separately, an attempt is made below to bring together these different strands of evidence, albeit in a preliminary and summary form.

In general the soil cover in this part of the site is shallow, ranging from about 20 cm in depth over TT05 and TT11, to about 40 cm in the northern part of TT09 (see Fig. 9 for the layout of the trenches). At various times in its recent history the North Hill has been used for agricultural purposes: for example, Robinson (Reference Robinson1930, xi) reports ploughing in the early twentieth century, and this may have continued following his project, as the site reverted to agricultural use before eventually becoming the property of the Greek State. The topsoil above the house has been disturbed by ploughing to about 20 cm and in places (for example TT05 and TT11) the stone socles show signs of what might be plough damage in the form of scratches across the upper surfaces of the stones. Within the plough zone displaced cobbles were found together with fragments of tile and pottery which had been fragmented and moved about. Nevertheless, below this layer, the archaeological deposits were surprisingly well preserved, given the shallow soil cover.

Specialist studies of various samples including flots, phytoliths and starches, together with additional geochemical and micromorphology samples are still in progress. Some preliminary comments can tentatively be made about the house as a whole based on the preliminary findings of the micro-debris study, which is also ongoing. Snail shells are common and are evidence of bioturbation throughout the site. Pottery and plaster are, likewise, ubiquitous. What is more surprising is the scarcity of certain types of find despite a careful screening for them. There is relatively little bone, probably because of the acidity of the soil. (As noted below, the bones that have been recovered have almost all come from ashy deposits, where the pH of the soil has been altered.) Coins are similarly rare, with only one recovered from the samples so far.

The first two trenches excavated in the house were small test trenches dug in 2014 to investigate specific features. The strategy adopted subsequently has been to divide the building into a series of squares measuring 5 m × 5 m, following the orientation of the 30 m grid originally set out for the geophysical survey. This means that the trenches are at an angle of 45 degrees to the walls of the house, avoiding the possibility that walls may be hidden until the final stages of excavation. Trenches measuring 4 m × 4 m have been laid out leaving 1 m baulks between to provide sections for micromorphology sampling and to ensure good stratigraphic control. (These will eventually be removed when final samples have been taken.)

TT01 measures 5 m × 3 m. It was laid out over the western wall of the house in order to test the identification of that wall in the geophysical results and to investigate patches of low and high resistance on either side. Work was concentrated in the western and southern areas. The boundary wall of the house was located, as expected, in the western corner, its socle of cobbles (C1005) surviving to at least five courses. The top began to emerge at a depth of about 15 cm below the surface of the topsoil. West of the wall, in the corner of the trench, a cobbled surface (C1012), of which only a small part is exposed, belongs to the neighbouring house, B ix 4, and explains the high-resistance anomaly here. The top of that surface is at the same height as the top of C1005. These features were the only architecture located in the trench. On both sides of the wall the fill contained broken tile and sherds. Inside B ix 6 a series of contexts is likely to represent mud-brick collapse.

The floor level of this house lies well below the cobbled surface in its neighbour. A layer of pebbles and gravel (C1010) was encountered at a depth of about 60 cm in the south corner of the trench, with an edge running parallel to the line of C1005. Based on micromorphological analysis, C1010 seems to have been a metalled outdoor surface composed of sand and gravel with some weathered sherds. Other micromorphology samples revealed a lens of white lime plaster which had been applied to the inside wall surface and extended horizontally to form a plaster floor 1–2 cm thick, which lacked the fine finish found on the red plaster floor in TT06. The preservation was variable, but in those samples from nearest the wall it was excellent.

TT05 is a square measuring 2 m × 2 m which was laid out across an anomaly interpreted as a north–south interior wall dividing two rooms in the northern part of the house. The top surface of the socle for such a wall (C5007) was encountered immediately below the topsoil at a depth of 10–15 cm. It is composed of stone cobbles and had some plaster adhering to its west face at the northern end. No features or floor level were reached on the eastern side of the wall, although some pottery and a bronze coin were recovered from this area. On the western side of the wall, underlying a thick layer of collapsed tile, a lens was located which may perhaps represent some kind of surface (C5004), although it is not found consistently (Fig. 11). This consisted of a patchy, thin film of clay overlain by large amounts (3.7 kg) of fragmentary pottery which was clustered into somewhat discrete (though in some cases overlapping) groups, varying in density. Many of the vessels followed similar orientations, tilting towards the north and west, suggesting that they may have fallen, either from a shelf or (perhaps) from an upper storey. Although they are in fragmentary condition, a large number are at least partially reconstructable; they include both fine-ware (especially skyphoi and other cups) and medium-ware forms (mostly jugs and basins). This context, with its associated deposit of pottery, continued into the baulks on the north-west and south-west sides of the trench, where it was detected in neighbouring trenches in subsequent years, as discussed below.

Fig. 11.

Photograph of TT05 during excavation (Kate Larson).

Five additional trenches, laid out in 2015, explored the northern range of rooms further. TT07 extended TT05 to the south-west. The north–south boundary wall between spaces continued in this trench as C7007. The clusters of vessels noted in TT05 to the west of this feature continued westwards and southwards in TT07 as C7004, away from the wall first identified in 2014. This deposit yielded a significant number of fine- and medium-ware vessels which appear to have been crushed during the collapse of the building. They include a fine lekythos with an unusual profile (C7004PT01: Fig. 12, discussed below under ‘Ceramics’) and fragments from at least eight different medium-ware vessels. These were identified as resting on a surface consisting of a compact deposit of yellow/brown/grey clay with small pebbles (C7015). The space in which this material was found, (d), was delimited to its south by the stone socle for a wall running east–west (C7008), which formed part of the north wall of the pastas of the house (c). South of this wall there were fewer ceramics than in room (d), and the vessels were less complete. Nevertheless, a range of vessels was found here, including the handles for two lekanis lids. Again, these appeared to rest on a surface (C7014) comparable to that identified in (d).

Fig. 12.

C7004PT01: lekythos with an unusual profile (scale in cm) (Irene Liesk).

TT08 lies north-east of TT07. Most of the trench is occupied by parts of two spaces continuing from TT07 (d) and (e), separated by a continuation of the north–south dividing wall (C5007/C7007), which is here identified as C8003. In the north-western part of the trench this meets a substantial wall running east–west, C8005, which constitutes the rear wall of the house. Beyond is a section of the stenopos, the rear alley. The fills in both (d) and (e) contained fragments of tile and of white and red plaster. Tile and ceramic fragments seem to have been deposited in two distinct layers within the fill. Finds in (d) include fragments of a Corinthian-type skyphos (C8004PT01).

In (e) a layer of small stones (C8010) was located at about 25 cm below the surface. This spanned the entire exposed extent of the space, and is assumed to represent either a floor or a subfloor. This interpretation is supported by the fact that, once exposed, both faces of wall C8003 were seen to preserve significant amounts of white plaster which ceases at the level of C8010. A number of artefacts also seemed to rest at this level, including 4 kg of medium ware comprising rims of chytrai, a hydria rim and parts of other closed vessels. It was therefore sampled on a 50 cm grid for phytoliths, starch and ICP analysis (results of which are still pending). Ceramic finds from a level above this one (C8007) included medium wares as well as fragments of black-slipped vessels. Among these were body fragments and a red-figure lid decorated with Dionysos and maenads from a lekanis (C8007PT01: Fig. 13), the knob of which was found in TT07 (C7006), suggesting that the vessel may have been located in an upper storey at the time the house collapsed. C8007 may therefore represent upper-storey collapse. Removal of C8010 over a limited area revealed a second layer about 15 cm below the first one which was similar to it but composed of larger, denser, stones (C8012).

Fig. 13.

C8007PT01: red-figure lekanis lid decorated with Dionysos and maenads (scale in cm) (Irene Liesk).

To the north of C8005, the stenopos behind the house had cobble paving (C8011). Along the two sides were narrow, raised platforms, also of cobbles (C8008 and C8009), which apparently do not bond either with the house walls or with the deeper cobbled surface of the stenopos itself. Sitting within the stenopos was a thick tile layer (C8006) beneath which lay some ceramic material including several body fragments of a beehive, resting on C8011.

TT09 lies to the north-west of TT08 and was located to investigate the stenopos and sample the southern part of the house to the north (house B ix 5). The fill in this trench was deeper than in any of the other trenches excavated so far (at its deepest, 40 cm in the northern corner of the trench). Beneath the topsoil and destruction fill, the continuation of the rear wall of house B ix 6 was located as C9005, along with the stenopos to its north and parts of three rooms of B ix 5. As in TT08, the stenopos contained a very thick layer of well-preserved (in some cases almost complete) tiles (more than 135 kg were removed from this area). Underneath was an almost complete transport amphora, as well as fragments of fine-ware vessels which may either represent refuse deposits, or alternatively may have fallen from the upper storey either of B ix 6 or B ix 5 as the two houses collapsed. This section of the stenopos lacked well-defined cobble surfaces either side of the central channel.

The stenopos is separated from the northern house, B ix 5, by the stone socle of the house's back wall (C9004). Excavations in this neighbouring property provide a valuable parallel for some of the finds from B ix 6. Work focused on the central and eastern spaces, (bb) and (cc). Space (bb) contained a variety of well-preserved ceramic vessels, including a small table amphora (C9012PT01: Fig. 14). Space (cc) was notable for its distinctive ashy deposit (C9013) containing broken ceramics and terracottas. These were found together with fragments of unburnt bone from cattle, pigs and sheep or goats, along with further burnt fragments of indeterminate species (see the section on faunal material, below, for a more detailed discussion). Interestingly, the cattle and pig bones show no sign of heat damage, and the cattle bones are likely to have been discarded during butchery. It therefore seems likely that processing, as well as cooking and consumption of meat, took place within the house. Based on these finds, (cc) seems to represent what Robinson and Graham termed the ‘flue’, and it provides a useful comparison for space (b) in B ix 6 (TT10 – discussed below). The material contained within seems to represent, at least in part, refuse.

Fig. 14.

C9012PT01: medium-ware table amphora (scale in cm) (Irene Liesk).

TT10 is located to the south-west of TT09. In the north corner of the trench the rear wall of the house continues east–west as C10005. On the east side of the trench three courses of cobble socle remain from a north–south wall (C10016) separating the western side of (d) from two other spaces (a) and (b). Excavation focused on the latter two spaces. In (a) the remains of a terracotta bathtub were located, the walls broken off just above the base, and incomplete on its south side (Fig. 15). The bath is set into a plaster or mortar surface adjacent to the rear wall of the house, to which it was connected by a thick cobble-and-mortar feature (C10009). This part of the house had a complex history which included at least one episode of significant remodelling, probably involving the rearrangement of the walls on the room's western side and the re-laying of a floor over a disused millstone. A small slot trench was cut adjacent to the bathtub in order to establish the underlying stratigraphy. It showed that a series of layers of waterproof plaster had been laid prior to the tub's installation and that the tub itself must have been placed here only during the last phase of use of this space.

Fig. 15.

C10018 showing the remains of the bathtub (foreground) and millstone (Marine Garcia).

Further to the south, part of space (b) was separated from (a) by a further wall socle (C10015). Space (b) contained ashy deposits similar to those located in space (cc) in TT09 and seems likely to have been a second flue, an interpretation supported by its location adjacent to the bathing area – an arrangement frequently found by Robinson. The deposits in (b) contained part of the unburnt jaw of a juvenile pig (probably slaughtered for meat), along with an extensive collection of pottery. Among the vessels present were a nearly complete lekythos with an unusual white-slipped interior (C10021PT01), a lamp of Howland type 25B (C1017PT01) and a one-handler with full profile (C10017PT02), along with several medium-ware vessels, some of which were for cooking. Given the unburnt state of the bone material found in both (b) and (cc), it seems likely that in each case the ‘flue’ was used for secondary discard of ash and other household refuse, and perhaps as an outlet for smoke from a hearth located elsewhere, rather than as a fireplace. In comparable spaces in the houses they excavated, Robinson and Graham (Reference Robinson and Graham1938, 189–97) identify examples of this practice as well as other cases where a fire was lit directly in the space itself.

TT11 lies to the south-west of TT10. The socle for an east–west wall (C11015) runs across the northern part of the trench, separating the pastas (c) from an interior room to its north. A section of the central and western part of the wall is missing before a fragmentary continuation of it is found in the western part of the trench (C11016). It is unclear whether there was originally a stretch of wall here which was destroyed, or whether this was a doorway. C11016 is clearly damaged, with stones missing, and C11015 is apparently twisted and buckled, its upper surface pointing southwards. Curiously, the face of the wall is slightly out of alignment with the face of what appears to be the continuation of the same wall, in TT07. The reason for this may become clearer when the intervening baulk is removed.

Excavation in TT11 focused on southern space (c), the pastas. Below the topsoil and a layer of tile collapse was found a large, well-dressed ashlar block which is possibly reused from an earlier structure. It was laid perpendicular to C11015 and probably served as the base for a set of wooden stairs leading to a balcony and rooms in an upper storey above. On top of this block, beneath a layer of tile, lay a collection of nearly complete, though smashed, ceramic vessels (Fig. 16). These included a handmade, miniature, unslipped, square tray (C11014PT04), a miniature salt cellar (C11014PT02), and at least seven black-slipped drinking vessels. An iron nail was also found among them. To the east, south and west of the block the tile layer continued at a lower level and further ceramic vessels were found beneath, including additional fragments of C11014PT04. These rested on a pebble layer which may represent a floor or subfloor, which was about 20 cm below the top of the upper surface of the stair base. The southern corner of TT11 grazed the face of the southern wall of the pastas (C11024), which is neatly built of small angular stones in a very different style from the typical cobble socles found elsewhere. Resting against its face was a small, unslipped jug (C11022PT01) which had lost its neck, rim and handle in antiquity, but was otherwise intact.

Fig. 16.

C11014: mid-excavation photograph showing the ashlar block with ceramic vessels emerging on top (Patrik Klingborg).

Although further cultural deposits remained in most of these trenches, new areas were opened in 2016 with the aim of exposing the most recent occupation levels over the whole of the house, before returning in a future season to the old trenches to address specific questions about earlier phases or about the house's construction. A further eight trenches were therefore laid out, improving our knowledge of the overall organisation of interior space and of the context of the house within the broader fabric of the block in which it is located.

TT12 sits over the southern boundary of the house, revealing part of the street to its south, together with a length of the socle belonging to the facade and the area just inside it. Excavation concentrated on the facade itself (C12007) and interior spaces behind. The western extent of this socle is faced with cut stone blocks (visible in Fig. 10 in the southernmost excavation square, to the left of the tree), including a long but relatively narrow threshold block which disappears into the south-western baulk. Running along the interior face and continuing eastwards inside the other blocks are two lines of unworked stone cobbles. It is unclear whether C12007 forms a continuous feature right across the trench, because the tree interrupting its course was not removed in 2016. Two further wall socles abut C12007, running north–south, and are roughly parallel. The westernmost of these, C12006, is faced at its southern end with a further two ashlar blocks and further north is composed solely of cobbles, with a few tile fragments also incorporated into it. It disappears into the baulk at the northern corner of the trench. This north–south wall separates a space (h) with a gravelly deposit probably representing a surface (C12023) to the west from a second space (i) which had a cobbled surface at the north end (C12019), but no discernible surface further to the south. The second socle, to the east (C12008), is more roughly built, consisting of larger cobbles which are today less precisely placed and lack any evidence of facing. All three of the socles in this trench preserved evidence of mortar between the stones.

A major question to be resolved in a future season is whether the threshold block marks the location of a doorway which was in use at the time the house was destroyed, or whether it was being reused as building material for the house facade. From the perspective of the plan, the internal space to the east of C12006 looks like an entrance, leading into a cobbled courtyard to the north (see discussion of TT13 below) and with cut blocks accenting the socle at the corner. It is possible that there were two entrances leading into different spaces, although in the houses excavated by Robinson, where there were two entrances one typically led into a space adjacent to the party wall.

TT13 lies to the north-east of TT12. No wall socles were found in this trench. Instead, the major feature was a cobbled surface (C13009) belonging to what must have been an unroofed courtyard (k), which surely represents a continuation of C12019. At the south and east corners of TT13, although there is apparently no dividing wall, there is a change of surface (an absence of cobbles), suggesting perhaps roofed but unwalled spaces (Fig. 10). In the southernmost part of the unpaved area a cut stone block (possibly a base playing some kind of structural role) lies at the boundary between the two surfaces, close to the south-western baulk. A deposit of at least 23 loomweights was also found here clustered closely together and disappearing into the baulk. Overlying the cobbles of (k) was a thick destruction deposit composed (in places) of as many as six layers of fallen roof tiles (7963 tile fragments were recovered, weighing over 602 kg). A total of 19 kg of smashed fine- and medium-ware pottery lay beneath (Fig. 17). In addition, 28.7 kg of amphora sherds, estimated to comprise at least 12 separate vessels, were recovered from here. The sheer quantity of roof tile suggests that as the house collapsed the pitched tiled roofs on the surrounding rooms slid down into the courtyard.

Fig. 17.

C13007 showing crushed ceramics overlying the cobbled surface of the courtyard (k) (Marine Garcia).

TT14 lies to the northeast of TT13. Most of the trench is occupied by the eastern end of the pastas (c). In the eastern corner of the trench a wall (C14010) perpendicular to the southern wall of the pastas represents the boundary between house B ix 6 and its eastern neighbour, B ix 8. Like the western boundary wall exposed in TT01, this seems to have been a retaining wall, although, to preserve its stability, only its top two courses were exposed on its eastern face, while soil was left in place below as excavation progressed deeper. At about 0.5 m below the preserved top of C14010, beneath a very thick layer of tile, a small area of a surface from house B ix 8 was exposed, consisting of clay and pebbles. Inside B ix 6, south of the southern wall of the pastas (C14004) and abutting against it, is an interesting structural feature in the form of what appears to be the stone setting for a wooden post which may have helped to stabilise or support the wall and the tiled roof above.

TT15, on the western side of B ix 6, provides additional evidence of its boundary with B ix 4, as well as revealing the western end of the northern range of rooms in B ix 6. C15015, the socle belonging to the north–south boundary wall between the two houses, came to light about 20 cm beneath the surface. In the western corner of the trench, a small area of house B ix 4 was exposed. Flecks of red plaster in the fill against the western face of C15015 suggested that it once had a red plaster surface. At about 10 cm below its upper surface a substantial floor came to light (C15012). This has a mortar surface with a neatly moulded curb, forming a border for a black and white pebble mosaic beyond, which appears not to have had any deliberate design (Fig. 10, visible at the north-western edge). This floor suggests the space may have been of the type identified by Robinson and Graham as an andron – a room associated in Athenian literary sources with use for the symposium. Its presence within the northern range of rooms of the house is unusual, but not unprecedented (compare house A v 8, where an andron occupies an identical position: Robinson and Graham Reference Robinson and Graham1938, 95–6). A fragment, probably from a bell krater, with a red-figure figurative design, was found in this space in B ix 4.

Abutting C15015 in the northern corner of the trench, a small part of the rear wall of the house came to light (C15020). In the southern corner of the trench, and also abutting C15015, the northern wall of the pastas was identified as C15011. The majority of TT15 is occupied by room (f), to the north of this wall, which, together with (a) and (b) detected in TT10, probably makes up a complex of a type referred to in the context of Robinson's excavation as the oikos unit (or ‘oecus unit’: Mylonas Reference Mylonas and Robinson1946). This group of rooms was interpreted as a living space which consisted of a main room, occasionally furnished with a stone-curbed hearth, and two smaller spaces, one apparently used for bathing and the other (the ‘flue’) containing ashes. In B ix 6, room (f) would equate with the larger living space, (a) and (b) with the bathing area and flue, respectively. A layer of small pebbles similar to those identified as floors or subfloors in TT07 and TT08 (C7015, C8010, C8012) was located here (C15013) and probably represents a floor level, although further cultural levels lie beneath it.

TT16 lies to the south of TT15. The trench is bisected north–south by the socles for two walls (C16004 and C16009) which are of different thicknesses and constructions, and follow slightly different alignments. Each has at least two courses and continues below the level excavated in 2016. Work concentrated on the west side of these walls, where a deposit (C16007) was located within which there were multiple ceramic vessels lying crushed in small pieces. These included sherds of amphora along with fine- and medium-ware vessels such as a large, coil-built tub or basin which was found partly resting on top of the wall socle. Further excavation will be necessary here to determine the relationship between this deposit and any floor level, which may lie beneath. Curiously, despite its proximity to TT01, no trace of this deposit was found in that trench, which was excavated to a deeper level.

Three further trenches, TT17, TT21 and TT22, have exposed more of the boundary walls of the house. TT17 removed part of the baulk between TT08 and TT09, revealing that the platforms on either side of the stenopos (C17007 and C17008) continued from TT08, before terminating in curved ends. These may have served to canalise the flow of water here, directing it to the centre of the channel. A half-preserved beehive lid recovered from the cobbled surface of the stenopos is composed of orange fabric. It originally measured some 0.7 m in diameter, with a maximum thickness of nearly 0.03 m, and was finished with two inscribed concentric circles near its central area. In TT21, on the eastern side of the northern range of rooms, only the western corner of the trench lies within house B ix 6, one corner of space (e). In contrast with the part of the space excavated in TT08, however, there was a notable absence of pottery here.

Bridging Between Different Analytical Scales: The ceramic assemblage from the North Hill (BAA, KML, AP)

While the geophysical survey and trial trenches together offer a large-scale impression of the character and layout of a whole neighbourhood in the north-east part of the North Hill, the excavation of house B ix 6 offers a very detailed case study of the organisation of a single house and, ultimately, some of the activities taking place within. Bridging between such different analytical scales is a challenge: the use of the GIS enables the data to be nested and contained within a single system, but it does not provide a conceptual framework through which to link these very different analytical units. Such a link is provided, however, by artefacts and their distribution. These can be studied using a range of different ways of understanding ‘context’ to provide a web of associations encompassing both excavated material and surface finds. The ceramic study from the North Hill provides an example of this approach. It currently incorporates material from all of the North Hill trenches and will expand to include surface material following the July 2017 season. Together this material serves as a basis for re-evaluating the conclusions reached by Robinson, and for modelling various cultural practices of the inhabitants of the North Hill. Insights gained from excavating a single house can be applied in the interpretation of surface assemblages from larger areas such as neighbourhoods or districts, or smaller ones – such as those from a trial trench. Similarly, the overall assemblage can be used as a reference in evaluating how typical the material from a more localised context might be.

Fabrics and Forms

Some provisional comments can be made about the ceramic assemblage as a whole. The relative quantities of different fabrics found by the Project correspond poorly to Robinson's finds. In the course of processing and recording the pottery, the Project has found that most non-fine-ware pottery shapes fall into the medium-ware category. Robinson presented only a few non-fine-ware types in the publications, so it is not fair to compare our proportion of non-fine wares with his; however, he found many more fine wares than the Olynthos Project has. It seems likely that this was because his fine wares were culled from large-scale excavation, which included cemeteries as well as residential contexts, and because the majority of medium and coarse wares were discarded unless they were almost complete. However, the uneven distribution of ceramics between different spaces in a single house, which is also a striking feature of Robinson's data, is also to some extent characteristic of house B ix 6, suggesting it is not simply a result of his collection and discard practices.

Chronology

As noted in the Introduction, the close dating of the destruction of Olynthos in historical sources has led to the assumption that its ceramic assemblage offers a snapshot of the forms in use in the mid-fourth century bc. Nonetheless, as Nicholas Cahill (Reference Cahill2002, 53–7) showed, the scatter of late coins from the northern part of the North Hill gives good reason for scholars to test this hypothesis independently. As noted above, a number of contexts of B ix 6 seem to contain objects in use or storage at the time of the abandonment and collapse of the house. More research is needed to be able to suggest dates with confidence. On the whole, however, the dating of all ceramics thus far recovered from the North Hill excavations is in line with the traditional fourth-century bc phase of occupation, destruction and abandonment. There is almost no material dating earlier than the second quarter of the fourth century bc. The earliest datable fragment noted so far from the North Hill is from C9002, an upper-level fill, which yielded an Attic fabric Type C cup foot fragment (c.500–480 bc). This is much worn, so that it seems possible that it had been included in the fabric of the mud-brick superstructure of the house, rather than being in use in the house itself. Of the common shapes, such as Attic-type skyphoi, there may be a range of dates represented. This form develops a more flaring out-turned rim and contracted foot with a cyma profile to the wall. There were some fragmentary bases and rims that looked like earlier fourth-century bc forms. There are also very few pieces that seem likely to date later than the mid-fourth century.Footnote ⁸

Expansion of this study, adding surface material from across the North Hill, will provide a larger sample with which to address questions of dating, as well as of variation between neighbourhoods in their use of different vessel types. It will also provide a context against which to evaluate the assemblage from B ix 6, and the houses excavated by Robinson.

Geophysical survey (DLS)

Geophysical survey was undertaken in 2016 with resistance and electromagnetic induction.Footnote ¹⁰ The aims of geophysical survey on the South Hill were, first, to understand the layout of this district of the city, about which specialists disagreed,Footnote ¹¹ and second, to determine potential locations suitable for excavation. A 30 m × 30 m grid at 45 degrees was used, a continuation of that employed on the North Hill. The majority of the survey was undertaken in the middle and the southern side of the South Hill. Most of the northernmost and southernmost ends of the hill were avoided, since the presence of earlier excavation trenches in these areas interfered with the ability to survey in contiguous blocks there.

Intra-urban survey on the South Hill (DLS)

Random collections have been performed in a total of 60 grid squares across the full length of the South Hill. Where the entire square had been previously excavated and a random sample area that had not been disturbed by prior archaeologists could not be located, only grab samples were collected across the square. Where the grid square was covered with dense vegetation and inaccessible, no collection was performed.

The material collected on the surface of the South Hill has to date been washed, sorted according to ware, fabric, and surface treatment, counted, and weighed. Finds have not yet been drawn or examined individually for detailed analysis. The results presented here are therefore preliminary, and will be reported in greater detail in the future.Footnote ¹²

The density map of pottery from the South Hill survey shows that the distribution of artefacts runs across the entire hill (Fig. 22). A clustering of higher densities is noted at the centre, but there is no area with a consistently low density of material, suggesting that occupation extended across the entire hill. A notable result of the gridded collection was the collection of a significant amount of fine ware with banded decoration dated to the Iron Age or Archaic period (see the ceramics report below). This was discovered in 34 of 60 grid squares with random collections (Fig. 23). Its significance is that it is earlier than most of the other artefacts recovered in other contexts. A large number (14) of special finds were also recovered during the survey on the South Hill. These included objects of metal, terracotta and stone.

Fig. 22.

Density of pottery collected during the South Hill survey, 2014–16 (DLS, HP).

Fig. 23.

Distribution of Archaic/Iron Age pottery collected during the South Hill survey, 2014–16 (DLS, HP).

Trial trenches on the South Hill (TT18, TT19, TT20) (EBT)

On the basis of the results of the geophysical and surface surveys we decided to excavate three trial trenches in order to verify various anomalies present in the geophysics results, to investigate the city plan, to explore overlapping walls of different orientations and their stratigraphy, and finally to distinguish earlier archaeological deposits from later fills. Three small trial trenches TT18, TT19 and TT20 were laid out within the squares of the grid established for the geophysics (for locations, see Fig 2, Fig. 19, Fig. 20). Work in them lasted for less than three weeks. All the trenches were excavated using a contextual methodology (see above under ‘Methods’) and samples for flotation, phytoliths and geochemistry were taken from each context.

The Pottery from the South Hill survey (AP)

This discussion summarises the range of pottery from the intra-urban surface survey, which complements the evidence from the geophysical survey and excavation. The assemblage as a whole contrasts with the ceramic assemblages from the North Hill excavations (discussed above) and from the field survey (discussed below) in having a much longer chronological span. The earliest diagnostic pottery dates to the Late Neolithic period. A few sherds with black-on-red decoration and a small fragment of a pithoid vessel with black-burnished surface on the interior represent this period. Only one fragment with characteristic rope decoration and oblique incisions dates to the Early Bronze Age. There is a variety of handmade diagnostic pottery dating to the Iron Age. This includes sherds of two-handled phialai with high-standing knotted handles (e.g. Fig. 27); jugs, probably with cutaway necks, with double rolled or triple handles (e.g. Fig. 28); and characteristic grey handmade pottery.

Fig. 27.

Phiale with high-standing knotted handles (scale in cm) (Irene Liesk).

Fig. 28.

Jug with triple handle (scale in cm) (Irene Liesk).

A large set of pottery dating to the Archaic period consists of vessels with banded and/or wavy decoration. The great majority of Archaic pottery is attributable to local workshops of the Chalkidiki. The hydria is the most common shape. However, there are also sherds of probable East Greek origin, including a sherd from a Chian kylix decorated with a lion on the exterior (Fig. 29), a sherd of polychrome Ionic pottery (Fig. 30), and another from a Corinthian exaleiptron, which is a rare shape in the area. Wheel-made grey monochrome pottery is considered to be of local production with eastern influence. A sherd of a lebes with horizontal rim bearing banded orange decoration indicates the influence of Aeolian bucchero.

Fig. 29.

Chian kylix: (a) exterior, showing arching tail of lion; (b) interior (scale in cm) (Irene Liesk).

Fig. 30.

Ionic polychrome sherd (scale in cm) (Irene Liesk).

The great majority of the recorded pottery dates to the Classical period. Red-figured pottery is not common. There is a sherd of a local squat lekythos with a palmette, and a sherd of a local skyphos. Black-slipped pottery with a buff fabric containing sparse mica is attributed to local workshops. As with the material of this period from the North Hill, only a small quantity can be considered to be Attic. The commonest shape is the bowl, either without handles and with incurving rim, or with one handle and an out-turned rim. These are mainly local products. There are also various types of kantharos, an askos and a stemmed cup (probably Attic). Wares of orange micaceous clay with many white inclusions were also locally produced. Jugs of many types predominate among the finds. Several transport amphora fragments belong to the well-known Mendean type. There are also black-slipped Attic and local lamps.

There is no Hellenistic or Roman pottery. From the Byzantine period there are some fragments of yellowish glazed vessels that are attributable to the last occupation of the site and the Byzantine structure that Robinson excavated here.