Last week, I asked for an extension on a blog post on the ceramics from the site of Pyla-Koutsopetri on Cyprus. My generous readers granted my the extension and, believe, I hope that you’ll find that you’ve been rewarded for your wait.
This is the final section in the first effort to prepare a draft of our work at the site of Koutsopetria in Cyprus which we excavated in 2009 and Dr. Maria Hadjicosti excavated in the 1990s. This excavation produced a significant assemblage of ceramic material that could be compared to a similar assemblage of material produced through intensive pedestrian survey of the plain. This comparison allowed us both to consider the excavated area in a larger context, but also to speak to the relationship between material below the plow zone and material on the surface.
My earlier posts focused on the architecture and history of the site, so here is what we can say about the pottery:
Despite being dominated by a Late Roman period building, the excavations at Koutsoeptria produced a robust assemblage of ceramics that speak to the long history of activity at this site. In this way, the excavation produced an assemblage that provides us with a useful comparative perspective on the data collected from the intensive pedestrian survey of this area and published in 2014. Among the most persistent critiques of intensive survey is that the relationship of the objects on the surface and those outside the plow-zone remains ambiguous hindering our ability to make functional arguments on the basis of artifact scatters (e.g. Sanders 2004). The formation processes and depositional history of assemblages in long-lived, multi-period sites set amid active and dynamic landscapes compound this further. At Koutsopetria excavations revealed how the persistence of residual material used in construction and floor packing, the cutting into earlier layers by later building and activity at the site, and hint at the effects of erosion and plow smear across the site created a diachronic surface assemblage. At the same time, the excavated assemblage revealed complexity that our sampling of the surface did not recognize. This complexity allows us to add meaningful detail both to our understanding of our survey assemblage and to an emerging ceramic signature present at historical period sites in the eastern part of the island.
Our discussion of the assemblage from Koutsopetria excavations relies upon two different excavation teams who sampled and analyzed ceramics based on two different strategies. During the 2009 excavations, we collected and analyzed all ceramics that were not tiles and sampled the tiles by type and extant part. It is unclear whether and how the excavation in the 1990s sampled artifacts from excavated contexts, but after excluding roof tiles from the samples, the excavation produced approximately the same number of artifacts (in 2009 we collected 3063 whereas in the 1990s they collected 3127) but much more artifacts by weight (2009 = 27778 and 1990s=82879) suggesting a more selective method of collecting ceramic material for analysis focusing on larger, presumably more diagnostic artifacts. Despite the disparity between the character of the two assemblages and the way in which they were produced, they are remarkably similar. From 2009, 68% of our material could only be assigned to the broadest possible category: Ancient Historic; from the 1990s this category of material was amounted to 59% of the assemblage by count.
The excavated area produced two discernible groups of pre-Roman material. There was a small assemblage of ceramics of Iron Age, Cypro-Archaic-Classical, and Cypro-Classical date which included coarse, medium coarse, and fine wares. These made up only a small percentage (far less than <1% by both number and weight) of the material from the excavated area and coincided with a similarly small number of artifacts associated with this period from the survey area generally. Most of this material is in secondary context and the fragments are quite small. The material likely entered into an excavated area from either Classical period activities along the base of the Vigla height where the survey documented a small concentration of Cypro-Classical age pottery perhaps from near an earlier findspot of the large, inscribed Cypro-Classical to Hellenistic period settling basin dedicated to Apollo Karaiates (Hadjisavvas 1993: 75–76, 83). Another possible location for Iron Age material is the site on the nearby Kazamas ridge or the earlier phases of activity at the fortified site of Vigla which may have been quarried for building material. During the Hellenistic period, the coastal plain saw greater activity, and this is reflected in the residual pottery from the Koutsopetria assemblage. Unlike Iron Age material which tended to be small fragments of fine wares, the material dated to either the Hellenistic period or one of the broader, related periods (Hellenistic-Early Roman or Hellenistic-Roman) tended to be larger and represent a more functionally diverse assemblage with the full range of coarse and medium coarse utility wares, amphora, kitchen wares, and fine ware. Of particular note was the long-lived (Archaic-Hellenistic) basket-handled amphora that appeared in excavated contexts and appeared both on Vigla as well as on the coastal plain. The link between these vessels and settling basin may hint at the importance of olive oil production in the area. The fine ware present was evenly split between Black-Glaze (21) and Color Coated wares (23), and this followed closely the division in the Hellenistic fine ware assemblage from the survey area suggesting that these may reflect the supply to the area during this period. The excavated assemblages did not produce kitchen or medium coarse wares that appeared in the survey although these artifacts did not appear in the immediate vicinity of the excavated area. The broader Hellenistic-Early Roman period, however, did produce a more robust assemblage. The challenge with more broadly dated material is that they tend to straddle the overlap between the Hellenistic and Roman periods.
The Hellenistic-Roman and Hellenistic-Early Roman assemblage from Koutsopetria made up just over 5% of the total assemblage from Koutsopetria. The assemblage is diverse and includes coarse and medium coarse utility wares, amphora, kitchen, and fine wares. The comprehensive character of this assemblage is consistent with finds from the survey area, but likely reflects the slow spread of settlement on the coastal plain over the course of the Hellenistic and Early Roman periods. Material from these long periods includes long-lived Rhodian type amphora, cooking pots, and fine wares types that persisted even Eastern and Cypriot Sigillatas replaced color-coated wares on local tables.
During the Early Roman period, the diversity and quantity of material from the site expands and this parallels neatly the expansion of material from this period in the survey area. The most significant distinction between the assemblage produced from excavation and survey does not appear to the be presence of Early Roman and Roman material, but the assemblage produced from excavation proved significantly more diverse. The excavated assemblage produced no examples of cooking pots or utility wares save a handful of Koan-type amphora, which were likely produced on the island. Some of this is the result of certain artifact types being shifted into broader categories. For example Rhodian amphoras which we identified as predominantly Early Roman in the survey, were dated Hellenistic-Early Roman in the excavation. The appears to be also the case for kitchen wares which were more commonly dated to the broader Roman, Hellenistic-Roman, or Hellenistic-Early Roman periods. As a result, fine ware represented the Early Roman period in the excavation. The most striking difference between the survey assembalge and the excavation assemblage is that Cypriot Sigillata comprised 28% (n=21) of the Early Roman fine wares from the survey, but only 4% (n=3) from the excavation. Other Early Roman fine wares – largely less diagnostic fragment of red slips – consisted of 27% of Early Roman fine wares from the survey (including a fragment of Arretine ware and Eastern Sigillata B) and 55% from the excavation. The remaining sherds were the common Eastern Sigillata A, but the excavation revealed six subforms (Form 19, 37, 38, 44, 65, and a lagynos) whereas the survey only produced a single recognizable subtype Atlante Form 4. It is worth noting that the 2009 excavations produced a small piece of Roman glazed pottery likely dating to the Early Roman period, but quite unusual and without parallel at sites in the region. The absence of Cypriot Sigillata from the excavation is consistent with relatively rarity of this type of Early Roman fine ware. At the nearby site of Panayia-Ematousa, near the modern village of Aradipou, Cypriot Sigillata accounted for only 8.8% of the total fine ware from the site. The absence of CS from the western part of the island may reflect the flow of ceramic materials from east to west with Eastern Sigillata entering the eastern part of the island from Levantine ports and CS circulating from the western production area. The majority of this material appears in secondary contexts, particularly in floor packing or fills, that reflect early patterns of activity in the area.
The broadly defined Roman period at Koutsopetria captures some of the transition from Early to Late Roman activity at the site. Like many places on Cyprus, the 3rd and 4th centuries are poorly represented in both the survey and excavation assemblage at Koutsopetria. The excavation, for example, produced no “pinched-handled” amphoras or forms of CRS or ARS with well-established 3rd-4th century dates. . There are a number of long-lived types of pottery that appear in the broadly dated Roman assemblage that might hint at at “middle Roman” activity at the site. For example, there are African Red Slip sherds that can be assigned to no specific type which makes it impossible to exclude the possibility of early forms existing at Koutsopetria, but no specific evidence for those early forms appeared. Among the range of undiagnostic coarse and medium coarse wares in Roman fabrics, the presence of a small number of long-lived micaceous water jars (Middle Roman 3 amphora) which appear from 1st to 6th century AD offer a glimpse of the middle Roman centuries. The presence of Roman lamps and cooking wares make clear that the coastal plain of Koustopetria was a settlement during the Roman period.
The Late Roman period is the most abundant from both the survey and excavation. The utility wares and amphoras from the excavated contexts are largely identical to those found in the survey. Late Roman 1 amphoras are predictably common in both contexts. The excavation also produced a small number (n=10) of Late Roman 2 amphora from the Aegean and Palestinian amphora (n=2 [check this]). The assemblage produced a significant quantity of kitchen ware sherds including a small number of rather late Dhiorios ware cooking pots that are likely the latest artifacts from the excavation and have comparanda from the survey of the coastal plain. As with most other periods, the fine ware from the Late Roman period provides the best opportunity to reflect on the diversity of material from our site. The two dominant categories of Late Roman fine ware were African Red Slip and Cypriot Red Slip with the former accounting for 48% of the Late Roman fine wares by count and 38% by weight and the latter being 44% by count and 53% by weight. The remaining 10% is made up of Phocaean ware and other rather less diagnostic Late Roman fine ware. It is notable that African Red Slip is significantly better represented in the excavated assemblage than in the survey assemblage. In the survey, ARS accounted for 17.4% of the Roman period fine ware whereas CRS accounted for 42.5% of the same total. The diversity of the two assemblages, however, speaks to their fundamental similarity. There are no ARS forms present in the excavated material that were not also present in the survey with ARS Forms 61, 67, and 105 appearing in both contexts. Likewise the CRS forms reflect the more common types CRS9 and CRS11 as well as the less common CRS8. Phocaean ware appeared in two forms PWH 10 and 5 and the very common PHW 3 was largely absent with only 1 possible example of that form. The presence of substantial quantities of African Red Slip pottery in the excavation assemblage supports two general impression from our survey. First, our local Late Roman fine ware assemblage was dominated by African Red Slip and Cypriot Red Slip suggesting that the site had ties both to regional production centers and Mediterranean wide trade networks. The small quantities of PHW in the excavated area does little to challenge the distribution of this type of pottery at the base of Mavrospilos and Kokkinokremos along the Late Roman coastline and coastal road. We have argued elsewhere that this concentration may mark the presence of warehouses associated with the site’s role as a emporion (Caraher et al. 2014, 295).
There is no compelling evidence for post-Roman material from the site aside from 2 fragments of early modern roof tiles. This is consistent with the distribution of the small quantities of later material in the survey which tend to be concentrated in units adjacent to the small Ottoman/Venetian coastal battery some 300 m to the east of the excavated area. The two tiles are likely the result of plow smearing, local road building, or even intruded during the excavation process rather than a reflecting evidence for a distinct later activity at the site. While it remains possible that some of the assemblage datable to nothing more narrow than Ancient Historic could include later material, it seems more likely that post-Roman activity on the coastal plain was limited and did not directly involve the collapsed church building.