|
Guided by the results of the geophysical prospection we laid out a long trench on the north-south axis, 30 m in length and 5 m in width, so that we encompass the western side of the structure and the possible street to its south. We then divided the trench into six 5 m squares with the exception of the two southernmost squares 6 and 4 m long, respectively. Since the area had not been cultivated for many decades and the ground was very hard, we used a tractor to break the surface crust. On the first day, prior to excavating, we conducted an intensive surface survey of two 20 m squares that encompassed the trench. The method we used in surveying these squares was similar to the one that we had used in the surface survey of the Sikyonian plateau: two five-member teams walked each square, this time collecting all artifacts from the ground. In recording the finds we used the same forms that we had used during the intensive survey of the plateau where we registered the total number of sherds and the number of tiles among everything else. This will allow us in the future to compare the surface to the excavation assemblages in an effort to check the surface representation of buried artifacts.
We started the excavation in the first, third, and fifth square, and we then extended to the second and fourth square so that by the end of the season the dimensions of the one, continuous trench was 5 x 25 m. Excavation proceeded by layers without using baulks. Every context was recorded separately using the widely employed method of single-context recording. The excavation of the three first squares began with a diagonal trial trench, 1.4 m wide, so that we estimate the depth of the archaeological horizons and the degree of their disturbance by modern activities (particularly plowing). Once we found out that the depth was shallow and modern disturbances were small scale, we continued by removing the layers across the whole extent of the square. We sieved all layers that covered possible ancient surfaces and floors, as well as those who belonged to the foundation of structures, but only samples of layers that belonged to fills (artificial or natural).
|
|
In the northern sector, the upper levels across the whole trench (Contexts 500-503) contained large quantities of pottery and tiles mixed with rubble. More precisely we recorded 10,597 sherds of various periods (from the Early Hellenistic to the Late-Roman) with the 5th and 6th century CE being most represented. From these surface levels (Context 502) we have also retrieved a triglyph and metope slab which may originate from the long stoa. The layers underneath (Contexts 505, 508, 509) represent artificial fills most likely for leveling the ground, and date to the 5th and 6th century CE. They contained large quantities of pottery (7,458 sherds) and animal bones (over 2,200 pieces) but we did not identify any floor surfaces.
The most important layer of artificial fill in the northern sector (Contexts 510 and 511) contained numerous fragments of tiles, and is defined by walls to the south, east, and west. These walls predate the fill with the exception of a flimsy wall in the southeast corner of the trench (Context 1026), made of reused material, and placed in order to retain part of the fill. The western wall (Context 504), 7.8 m long and 0.60 m wide, is built with rubble and mortar. The eastern wall (Context 506), 3.8 m of visible length, was discovered along the eastern boundary of the trench and therefore its width could not be ascertained. It is made of dissimilar stones, some of which are reused and placed on their short sides (like orthostates). The southern wall (Context 1038) is 2.4 m long, 0.50 m wide, and features mixed masonry consisting of successive layers of rubble, mortar, and tiles. This wall was built directly south and adjacent to a preexisting stone pedestal (Context 515) measuring 0.72 x 0.65 m and over 1.10 m in height. The upper surface of the pedestal has a rebate along the edges, and its eastern (vertical) side has a rectangular frame, 0.62 x 0.24 m, which once framed an inscription that was later chiseled off.
|
|
|
Northwards the fill extended beyond the boundaries of the trench, perhaps up to the northern side of the building that we shall explore next year. As it is, the extent of the fill is ca. 20 m² and its average thickness is 0.35 m (it is deeper at the center and shallower along the edges), which represents a volume of ca. 7 m³. We dug some two-thirds of this fill, and the weight of the tiles which were found in small, not joining fragments, exceeded 1 ton. Along with the tiles, we found rubble and large quantities of pottery dating for the most part to the 5th century CE.
From the next Context (512) onwards we concentrated on the western half of the trench in order to expose older layers and possible floor surfaces. Context 512 (measuring 2 x 5 m) contained mostly Roman pottery, particularly of the late 4th to early 5th century CE. In lower levels (Contexts 514, 516, 519, 520, 522, 523) the pottery dated mostly to the early Imperial period. On the southern side of the trench, at the level of Context 520, we found a possible fragment of a mortared floor. The wall – Context 1038, perpendicular to the Context 504 wall, must belong to the 1st century CE horizon.
In short, at the northern sector of the excavated area (Trench 1) we dug approximately a meter below ground level and came upon the 1st century CE level. The western wall of the trench (Context 504) same as the pedestal continue deeper, onto earlier horizons that we shall explore next year.
|
To the south of Context 1038, in a space 4 x 3 m that extends southwards to Contexts 1022 and 1524, we removed layers of fill (Contexts 1012, 1016, 1027, 517, 521) with abundant pottery that dates predominantly to the 5th and 6th century CE. The last two layers that we excavated (Contexts 517 and 521), of an overall thickness of 0.50 m, are not level but slope eastwards. Most likely these layers were formed by the disintegration of the western wall of the complex (Context 504) and are not artificial fills. If this hypothesis is valid, then we can explain the artificial fills of the upper levels as an effort to level the area in later periods.
The use of this space in the Late-Roman period is not clear, and the excavation failed to locate floor traces. It is likely then that it was not roofed and related to the industrial installation that we revealed directly to its south. At the inner edge of the western wall – Context 504 we observed three holes which were likely used for tethering domestic animals. These holes, 6-7 x 4 cm, did not exist from the start but were pierced at a later date.
In the central sector of the excavated area, we began by removing the surface layers (Contexts 1000, 1001, 1007, 1012) filled with rubble and pottery mostly dating to the Late-Roman period. The rubble layer spreads across the whole length of the trench covering an area of approximately 5 x 8 m. The layer underneath contained large quantities of tiles and field stones (5,696 tile fragments weighing 918 kg), that manifestly came from the crumbling of the masonry, as suggested by their concentrations around the walls (Contexts 1003, 1013, 1016-1019, 1021). The pottery from these contexts (we recorded 9,775 sherds) belongs mostly to the 5th and 6th century CE.
|
After the removal of these layers, the structures of this sector were clearly demarcated: the western wall, part of the eastern wall, and the rectangular wine vat. The western wall (Context 1009), 4.17 m long and 0.39 m wide, is aligned with Context 504 to the north but differs significantly in style and construction. It is made mostly of reused ashlar blocks that are placed on their short sides. The inner surface of this wall is the western side of the industrial installation consisting of a wine vat and a workspace directly north of it. The three other sides of the vat are built with rubble, tiles, and a few reused blocks set in mortar. The northern wall of the vat, 2.87 m long and 0.37 m wide, is preserved along its whole length and was excavated to a depth of 0.47 m. Two small Ionic column drums are visible in its masonry. The western wall of the vat, 2.08 m long and 0.18 m wide, is fragmentarily preserved while only the western half, 1.08 m long, of the south wall is still in place. The eastern half was removed at a later time for unknown reasons. At its cut (Context 1045), 2.00 x 0.8 m, we found late antique material, the latest dating to the 7th century CE (Context 1043). The floor of the vat, 2.67 x 1.87 m, consists of rectangular limestone slabs of various sizes, 0.16 m thick. Thick mortar covers the joins of the slabs, the corners of the blocks, and the base of the side-walls up to a height of 9 cm. In some places we discerned up to three layers of mortar. The floor slopes towards the southeastern corner of the vat, where the liquid was concentrated. On the outer side, where the cut lies, we did not find any traces of stone or terracotta receptacle for storing the product. On the basis of the dimensions and the shape of the structure we can be fairly certain that it is a wine vat. We hope that the flotation of the soil that we collected from the floor of the vat (some 85 liters) will produce organic remains of the process. In this thin layer overlying the floor, 8 cm in average thickness (Context 1023), we found only a few datable sherds, including two thin-walled mugs and four African Red Slipped plates of production C and D, which date to the late Imperial period and up to the middle of the 5th century CE.
In the space north of the vat we excavated layers (Contexts 1037 and 1041) with plenty of pottery (we recorded 4,506 sherds) dating to the Late-Roman period and up to the second half of the 7th century CE. By sieving these contexts, we found 41 olive seeds. Perhaps more seeds will be discovered in the future by water sieving the soil samples we have collected. At the southwestern corner of this space, against the western wall, we found two stone slabs set upright, originally parallel to each other. They measure 0.36 and 0.38 (L.) x 0.11 (W.) x 0.12 m (Ht.). The northern slab was found in situ but the southern slab has been knocked off, however we can estimate that the original distance between the two slabs was ca. 20 cm. We are not certain about the purpose of this apparatus, yet the immediate proximity to the wine vat makes us believe that it is somehow associated with the wine making process. From the earth between the two slabs (Context 1042) that we collected for flotation we found a few sherds of the 5th to 6th century CE.
The eastern wall of the area (Context 1010) was only partially exposed as it extends beyond the limits of the trench. It is oriented north-south, that is parallel to Context 1009, and its length comes to 4.65 m. It is built of irregular stones, a layer of tiles laid flat on the upper surface of the wall, and with mortar as bonding agent. Although we have not uncovered yet the foundation level of this wall, therefore we cannot date its construction, its contact with Context 1524 to the north shows that it postdates that wall. At the northwestern corner of the wall, and in contact with it, two ashlar blocks were found in a row, forming a bench-like feature. The northern one was found in situ, the southern one slightly moved. Their width comes to ca. 0.40 m and the length to 1.20 m. Based on the discovery of many olive seeds in this area and on examples of ancient olive-pressing installations from elsewhere, it is conceivable that these stones were used as a base for the pressing of the stacks of olives via a lever and weights press. Since, however, we have not found yet traces of a receptacle for storing the output, our interpretation of these stones as press beds should be considered as tentative.
In the southern sector of the excavated area (south of the wine vat) the upper levels below the surface layer (Contexts 1514, 1034, 1516, 1036, some 20 cm thick) contained a lot of rubble and tile fragments. The pottery that we recorded (9,434 sherds in all) dates mostly from the 5th and 6th century CE as do the 13 coins retrieved from these horizons (12 in Context 1036). In addition, here we found more animal bones than in other trenches. Noteworthy is the discovery of two Syrian amphoras (Late-Roman 5) that were found in large, joining pieces. In Context 1516 many of the tile fragments join, which suggest that we are dealing with a destruction layer rather than an artificial fill as in the northern sector of the excavated area. The space is defined by walls to the west, south and east. The western wall (Context 1523), 6.25 m long, 0.50 m wide, and 0.51 high, is oriented north-south and built with rubble and mortar. It is aligned with Contexts 1009 and 504 to the north, but its masonry style and its contact with Context 1009 wall shows that it is a later addition. At its southern end the wall turns at a 90 degrees angle towards the east and ends at an ashlar block that manifestly served as the base of the door jamb. The original opening of the door on the south side is unknown, but it must have been smaller than the surviving gap of 1.7 m which is probably due to the robbing of the eastern door jamb. The eastern part of the south side (Context 1525) is preserved to a length of 1.27 m, while the total length of this side comes to 4.5 m. The wall along the eastern side (Context 1541), oriented north-south, is 4.8 m long but has been only partially exposed because it extends beyond the limits of the trench. Only the upper layer of the wall has so far been revealed, built of half-dressed stones and tiles.
|
|
|
Near the southeastern corner of this space we found the end of an installation consisting of two converging stone slabs set in a thick layer of mortar and clay mixed with ash. The installation continues eastwards beyond the limits of the trench and will be fully exposed in 2014. On the basis of the remains and of the composition of this layer, it is likely that we are dealing with the mouth of a small lime-kiln. If so, we can account for the presence of the four limestone blocks lying next to it. These were probably singled out for calcination which for whatever reason never took place.
To the south of this area, we dug part of a street with a drain along its axis and a retaining wall on its southern side. The upper levels of the trench (Contexts 1500-1503, 1508, 1510), ca. 20 cm thick, contained much pottery (we recorded 4,405 sherds) mainly of the Late-Roman period (5th to 7th century CE). With the removal of these layers, we exposed the upper part of a retaining wall (Contexts 1526 and 1527), of east-west orientation, showing two distinct phases. The eastern part (Context 1526), 2.25 long and 0.40 m wide, is made of limestone orthostates. The upper surface of two adjacent blocks preserves a dove-tail clamp cutting (the ancient πελεκῖνος). The height of the orthostates is 0.74 m and they seat on a second row of stones which we barely exposed this year by excavating deeper along the external face of the retaining wall (Context 1507). We have not discovered yet the base of this wall, therefore we do not know its date. Still, the depth of the wall, its fine masonry and the use of the dove-tail clamp argue for a date in the Early Roman or even the Hellenistic period. Unlike this section, the western part of the retaining wall (Context 1527), 2.7 m long, is clearly a later addition. It is made of reused blocks including three column drums, tiles, and rubble.
|
|
|
Along the northern side of the retaining wall, the strata that are associated with its disintegration (Contexts 1506 and 1509) contained rubble, roof tiles, and pottery (we recorded 1,378 sherds) of the 5th to 6th century CE. Further down, we dug a pebbly layer (Contexts 1512 and 1518) half a meter thick, that contained large numbers of tiles (we counted more than 10,000 pieces) and sherds (we recorded 15,655) of the Late-Roman period (5th to 6th century). This layer probably represents an artificial fill for raising the surface, seemingly the same action as the fill (Context 510) in the northern half of the trench. After removing it, we came upon the cover stones of a drain along the road axis. It is a channel 5 m long, 0.20 m wide, and 0.14 m high, with a terracotta floor made of tiles, and with mortared side walls 0.20 m wide. It is covered by stones of various shapes and sizes, some half-dressed and others (among which four column drums) in second use. We investigated the interior of the channel by removing four cover stones at its western end. By cleaning its interior to a length of just 1.5 m (Context 1537) we retrieved only 44 sherds which included a few thin-walled mugs of early imperial date. For a secure date of this drain we need more evidence, particularly from its foundation trench. The layer to the north of the drain (Context 1522), ca. 0.30 m thick, had many small tile fragments (we counted 679) and plenty of pottery (we recorded 2,992 sherds) of various periods with the 5th to 6th century CE being the predominant one. In the layer to the south of the drain (Context 1521), ca. 0.15 m thick, we also found large numbers of tiles (520 fragments) and sherds (3,236) which for the most part date to the 3rd to 4th century CE. Below this layer, and against the inner side of Context 1527, we found two concentrations of sherds (Contexts 1529 and 1530), many of which join to restore nearly complete examples of early imperial (1st to early 2nd century CE) thin-walled mugs. We can thus reasonably assume that these vessels were discarded along the southern side of the street, which means that the retaining wall was already in place.
Directly underneath the context where these vessels were found, we discovered a hard-packed surface with many pebbles and gravel, which extends on both sides of the drain (Contexts 1532 and 1533). Although we have not excavated this context yet, therefore its date cannot be ascertained at this point, its texture and consistency suggests that we are dealing with a road surface. At its northwestern corner two leveled stone blocks, roughly aligned, came to light that may belong to the northern curb of the street.
Although the excavation of this area has not been completed yet and most of the walls have not been uncovered to the level of their foundation, we can already identify different building phases. The dating of these phases is not firm because we had only few closed contexts, their majority showing pottery of more than one period. The “contamination” of the archaeological horizons is partly explained by their shallow depth and subsequent disturbance by later activities – plowing above all. In retrospect, the very shallow depth of the Late-Roman horizons (the floor of the wine vat was found just 20 cm below ground surface) is the result of the virtual abandonment of this area after the 7th century CE.
The structure on the northern side of the excavated area, built with dressed stones and mortar, belongs to the earlier phase, which probably dates to the early Imperial period. Thus far a section 8.5 m long of the western wall (Context 504) and part of the southern wall ca. 4.5 m long (Contexts 1022 and 1524) have come to light, whereas the northern wall must lie in the unexcavated area. The entrance to this building was from the south through a door 0.86 m wide. The stone pedestal (Context 515) inside the building probably belongs to the same phase, even though its base has not been revealed yet. The eastern wall (Context 506) was built at a later date, partly with reused blocks, but is only partially revealed for it lies along the edge of the trench. Context 1038, perpendicular to the western wall of the building, was erected against the stone pedestal with successive layers of rubble and tiles, perhaps in the Early Roman period, whereas the insubstantial southeastern "wall" (Context 1026) was placed in the Late-Roman period in order to retain the thick artificial fill to its north. The industrial installation in the central sector of the excavated area with the wine vat, the olive press and the space for tethering animals, came into being most likely in the 5th century CE. Later, but still within the late-Roman era (5th-6th century), the southern structure was added (Contexts 1523-1525) which was roofed at least in the beginning. Within it, we found small part of an installation that we interpreted as lime-kiln which extends beyond the eastern limits of the trench. The dating of the kiln is not known yet, but it must postdate the southern, roofed space since a lime kiln could hardly operate in a closed environment. Towards the south, the southern space bordered a street, some 3 m wide, with retaining wall on its southern side (Contexts 1526 and 1527). We do not know yet the date of the original phase of this street, but its width and position within the ancient town grid suggests that it was rather an alley that cut the insula into two halves. Down the spine of this street we found a built drain of small dimensions which appears to have been put later by cutting though the early imperial road surface.
The street to the south and the industrial installation in the central sector of the excavated area are the only spaces for which their function is now known. For the other excavated spaces their precise use is still wanting. We hope that the continuation of the excavation in depth and its extension to the east will help us to interpret the use of some of these rooms and to answer several questions: When was the building in the northern sector of the trench erected, for what purpose and what is its relation to the stoa at its north? What is the purpose and significance of the stone pedestal within it? What was the use of the space to the south of the wine vat and when was the lime-kiln planted here? When did the southern retaining wall come into being, and how many phases does the street have?
