KYTHERA ISLAND PROJECT - 2010
Type of Operation
British School at Athens
Kythera Island Project (KIP). E. Kiriatzi (BSA) and C. Broodbank (London) report on the 2010 study season (with contributions by A.J. Johnston [Geometric – Classical], K. Slane [Roman] and J. Vroom [Byzantine – Recent]).
KIP has at present 198 designated sites: study of 144 is complete, as is that of all tract material (Fig. 1). Provisional period distributions of diagnostic sherds are as follows. Some 46.3% of the tract pottery is dated to the Recent (19th-20th century) phase. Another 12.6% date to a single pre-Recent period, as follows (Fig. 2): Final Neolithic – Early Bronze Age I 8; EB II 121; First Minoanising (EM II-MM IA) 23; First Palace 0; Second Palace 517; Third Palace 19; Postpalatial to Geometric 0; Archaic 4; Classical 149; Hellenistic 12; Early Roman 14; Middle Roman 23; Late Roman 140; Early Byzantine 1; Middle Byzantine 8; Early Venetian 2; Middle Venetian 59; Late Venetian 38. The peaks of activity in EB II, Second Palace, Classical and Late Roman stand out, noting especially the Second Palace phase (Fig. 3). Second Palace tract material was mainly identified by sherds in a distinctive red micaceous fabric (some 95% of all Second Palace diagnostic sherds) including handles, cooking pot legs, rims of pithoi, jars and other shapes, decorated body sherds (mostly plastic and/or impressed, but a few dark-on-light painted sherds in buff non-micaceous Neogene fabrics). The four major chronological spikes reflect real peaks of presence in the landscape, although they are also easiest to recognize due to a strong suite of period-specific diagnostics. The inclusion of sherds that could also date to one of the surrounding phases (well over a quarter of the remaining pottery) fills some gaps. For example, a further 59 sherds must be either First Minoanising or First Palace, and 167 sherds either First or Second Palace. Other extended periods are boosted in this way: an additional 398 sherds definitely belong within in the Roman period, and the 865 sherds attributed a firm Late Venetian to Recent date better reflect the surge in island demography from the 18th century onwards than the mere 38 confirmed Late Venetian sherds.
Study of material from Ag. Georgios (KIP Site 111) was completed. In addition to the standard tract-walking of the entire mountain, a more closely spaced (5m) tract-based walking collection was undertaken in 2000 up to the limits of the area excavated by Ioannis Sakellerakis and his team (Fig. 4). Although this sample cannot match the quantity or quality of the excavated material, it provides insights into the overall phases and fluctuating intensity of activity at the site, from the transition to the 2nd millennium BC to the present day. 665 diagnostic sherds predominantly date to the Roman (ca. 1/3) and Middle Byzantine-Recent (just above 1/3) periods, while the prehistoric material comprises almost 1/6 of the assemblage. Prehistoric material is concentrated mainly on the south and south-west slope (tracts 9517, 9518 and 9523), immediately below the terraces excavated (the reserved area in Fig. 2). Second Palace material is abundant in these tracts, consisting mainly of cooking pots (tripod legs and rims) as well as trays in the local red micaceous fabric (Fig. 5 left), cups (plain conical and dark slipped straight-sided cups in fine local fabrics, Fig. 5 right), and medium to large vessels such as pithoi, basins, and ewers in the local mudstone-tempered fabric (Fig. 6 left). A fragment with applied curvilinear bands on its concave side may reflect a clay model of landscape representation, as those found at the excavation and studied by Emilia Banou (Fig. 6 right).
Imports are very rare and seem to relate to fabrics known on Antikythera and west Crete. All of this pottery seems to date to the Second Palace period, mainly LMI, although there is an earlier component. For example, pithoi in mudstone-tempered fabric are a feature of MM III-LM IA deposits at Kastri and a few of the cup bases may belong to carinated cups that can be even earlier. Among the large number of tripod cooking pot legs recovered from tract 9517, one is in mudstone-tempered fabric, a typical feature of the First Palace period based on the Kastri fabric typology.
It therefore seems that there is evidence further downslope from the main area of the peak sanctuary (covered by the excavation) to confirm activity from the First Palace period, a possibility already raised by the excavators and by Iphigeneia Tournavitou who has studied the pottery. Further significant finds came from tracts to the west and north of the Ag. Georgios peak where the density of prehistoric pottery is very low. Four sherds in sand-tempered fabric, with parallels in Kastri deposits Β (EM IIB-MM IA), Γ (MM IA) and less possibly Δ (MM IB-MM II), were collected from tracts 9515, 9520, 9522, 9523 (Figs 4). They raise the possibility that late prepalatial (local Minoanising) pottery exists on the site. The tiny amounts of prehistoric pottery collected from the west slopes of the peak include a possible kylix stem (tract 9521, see Figs 4) which, together with other possible Third Palace period sherds, may testify to continuing activity on the peak for some time after its acme.
Thereafter, no material securely predates the Classical period. A small number of possible Classical-Hellenistic sherds, including tiles, a kantharos and a few basins, cooking pots and amphora, indicate some form of activity during this period. As noted, Roman pottery forms the second most important component of the assemblage, with the vast majority dating late in Late Roman. Only some body sherds of Early Roman African amphorae have been identified in tracts 9519 and 9523, and there is no identifiable Middle Roman pottery. The Late Roman pottery is similar to that at Kastri, with ARS forms 104 and 105, LRC form 10C, many sherds of LR Amphora 1 and 2. The few Late Roman-Early Byzantine pieces include fragments of Kastri - Omega 330 type amphora and a possibly eighth-century cooking pot with simple concave rim. After another long gap in the material evidence, pottery of the 12th, 13th and 14th centuries AD appears frequent, mostly comprising amphorae, cooking pots and a few sgraffito bowls. The high density of Middle to Late Venetian and Recent ceramics reveals continuing activity on the site.
This evidence offers valuable points of comparison and contrast with the history of Kastri as recently reconstructed through the combination of old excavation and new survey data. In the same way, the current survey data on Ag. Georgios complement the main body of material from the Sakellarakis excavations, and the in-depth studies of his collaborators which are shortly to be published.
Seven further new sites were studied in 2010. Sites 119, 120A, 120B, 124 (A and B), 126 and 128 are shoreline scatters around the small creeks in and around the promontory of Avlemonas, below the Ag. Georgios mountain. They are mostly generated by passing shipping, with little evidence for habitation (see Fig. 8). They represent the greatest concentration of such sites in the survey area: other examples are 137A and 137B further north around the headland on Diakofti islet and Site 103 around the Venetian fort at Avlemonas itself. As a group, these sites can be compared to more heavily settled harbour sites, primarily Kastri, but also its neighbour Site 068 and others in southern Paliopoli around the now alluviated embayment. Characteristic of these shoreline scatters are poor preservation due to high levels of surface attrition and alteration, very high proportions of amphorae (including a dramatically higher incidence of long-range imports from, variously, Africa, the western Mediterranean, the Black Sea and the Levant), and the presence of small numbers of decorated bowls of types otherwise rare or unknown in the survey material, which may be cast-off ships’ equipment rather than normal imports. However, the chronology of these sites varies greatly, suggesting shifting preferences possibly dependent on tiny shifts in local coastal morphology, ship technology, other coastal activities such as salt gathering, and the location or absence of nearby settlements. For example, 120A and 120B face each other on the opposite flanks of a small and now remote inlet. 120A has a little Second Palace pottery, possibly connected to a substantial prehistoric focus (Site 121) further inland, as well as Venetian and Recent material, yet on 120B all datable material is Early Roman. 124A and 124B share a history of uncertain Roman presence followed by definite Venetian and Recent activity, but their earlier histories are markedly different and distinguishable spatially, with a peculiarly dense but tiny EB II scatter slightly further inland on 124A and Classical and Hellenistic black glaze eroding down the talus near a coastal cave on 124B, and possibly the remnants of a ritual site matching that excavated by Tsaravopoulos on the Dragonera islets opposite. Of the remaining shoreline sites, 119 is a large scatter around the light beacon at the tip of the Avlemonas anchorage, first regularly used in Hellenistic and/or Early Roman times, with a marked spike in Late Roman, uncertain continuity through Byzantine and, like Avlemonas itself, a resurgence in Venetian times. Nearer to Paliopolis, Sites 126 and 128 saw more activity in Second Palace, with variable histories of use and abatement from the Classical period onwards.
Among work to complete dating of other partially-studied sites, attention is drawn to Sites 137A and 137B, which between them cover the inner coastline of Diakofti island, now Kythera’s principal harbour and a potential anchorage since antiquity. Sites 137A and 137B saw sporadic activity from Late Neolithic (attested by an obsidian arrowhead), with a slight spike in the Second Palace period (possibly continued into the Third?) that may relate to the export or coastal circulation of the island’s red micaceous pottery. Classical finds are strikingly few compared to Kastri, but this entry/exit point flourished during Kastri’s relative eclipse in the Hellenistic to Early Roman periods, with a slight dip in Middle to Late Roman times as Kastri reasserted its primacy, and a second surge in Middle Byzantine to Venetian times.
Among observations from continuing pottery study, a stem fragment from a Protogeometric kylix was identified from the inland sector of the Kastri grid (Fig. 9). This is of obvious interest given Homeric references to Kastri (already known as Skandeia), and is so far the only sherd from the site definitely dated between the 13th and 8th centuries BC.
Unpublished field report, British School at Athens (E. Kiriatzi and C. Broodbank)
Date of creation
Fig. 1/ Tract map of the survey area with ceramic density (red=high density, blue=low density) and the subsequently defined sites. The bordered area is presented in more detail in Fig. 8.
Fig. 2/ Bar chart showing the chronological distribution of tract sherds that are safely dated to a single period.
Fig. 3/ Second Palace period: distribution of sites (red squares: fully studied, pink squares: preliminarily studied) vs tract pottery (dark blue circles: definite, light blue circles: probable).
Fig. 4/ Tract map of the Agios Georgios mountain showing high density of pottery (red) at the peak in contrast to the low density (blue) on the lower slopes (extracted from map of Fig. 1).
Fig. 5/ Surface finds from Agios Georgios (KSS 111) dated to the Second Palace period: a. cooking pot legs and tray fragment in red micaceous fabric (left), b. bases of conical and straight-sided cups in fine local fabric (right).
Fig. 6/ Surface finds from Agios Georgios (KSS 111) dated to the Second Palace period: a. pithos rims in mudstone-tempered fabric (left), b. fragment with applied pieces of clay on its concave side (landscape model?) (right).
Fig. 7/ Surface finds from Agios Georgios (KSS 111): a. sherds in sand-tempered fabric dated to First Minoanising (to First Palace period) (left), b. fragment of a kylix stem dated to Third Palace period (right).
Fig. 8/ Tract map of the Kastri-Avlemonas-Diakofti area, showing original ceramic densities and subsequently defined sites.