MAZI ARCHAEOLOGICAL PROJECT 2016: REGIONAL SURVEY AND SETTLEMENT INVESTIGATIONS IN NORTHWEST ATTICA

Sylvian Fachard, Alex R. Knodell, Kalliopi Papangeli Introduction The main goals of the third season of the Mazi Archaeological Project were (1) to extend and complete the field survey of the Mazi Plain (chiefly in Areas e and d), (2) to document in greater detail the sites of Kato Kastanava and Eleutherai, and (3) to conduct geophysical investigations in and around the settlement of Ancient Oinoe. The campaign took place between June 13 and July 15, under the direction of S. Fachard, A.R. Knodell and K. Papangeli. The team involved some 35 individuals, including senior collaborators, graduate and undergraduate students, and specialists, mostly from Greece, Switzerland, and the United States.1 The co-directors are grateful to the Ministry of Culture for its confidence and support over the course of the project (2014-2016). We also express our gratitude to S. Chrysoulaki (Ephor of West Attica, Piraeus, and the Islands) and to K. Reber (Director of the Swiss School of Archaeology in Greece), as well as to the institutions that provide support in the form of financial and other resources: the Swiss National Science Foundation, the Loeb Classical Foundation, the Institute for Aegean Prehistory, Carleton College, University of Geneva and its Fond Général, University of Nebraska at Lincoln, the Swiss School of Archaeology in Greece, and the Ephorate of West Attica, Piraeus, and the Islands. The 2016 field season of MAP involved multiple components, simultaneously conducting intensive and extensive pedestrian survey, digital and traditional methods for

1 These individuals represent seventeen institutions in five different countries: Sue Alcock (University of Michigan), Brandon Baker (University of North Carolina–Chapel Hill), Michelle Berenfeld (Pitzer College), Chloe Bergstrand (Carleton College), Maura Brennan (University of Cincinnati), John Cherry (Brown University), Alex Claman (Carleton College), Christian Cloke (University of Maryland), Sarah Craft (Florida State University), Liza Davis (University College London), Marie Drielsma (University of Geneva), Lindsey Fine (Cambridge University), Mitch Groninger (University of Nebraska-Lincoln), Jean-Quentin Haefliger (University of Geneva), Christine Hunziker (University of Geneva), Thomas Kerboul (University of Geneva), Tobias Krapf (Swiss School of Archaeology in Greece), Evan Levine (Brown University), Xavier Mabillard (University of Lausanne), Maeve McHugh (University College Dublin), Julia Miller (Carleton College), Jacob Morton (University of Pennsylvania), Sarah Murray (University of Nebraska–Lincoln), Brian Niedert (University of Nebraska–Lincoln), Nefeli Piree Iliou (University of St. Andrews), Tim Poenitz (University of Geneva), Miriam Rothenberg (Brown University), Rebecca Salem (University of Nebraska–Lincoln) Catie Steidl (Brown University), Eirini Svana (Ephorate of Antiquities of West Attika, Piraeus, and Islands), Evgenia Tsalkou (Ephorate of Antiquities of West Attika, Piraeus, and Islands), Polytimi Valta (Ephorate of Antiquities of West Attika, Piraeus, and Islands), Jackson Vaughn (Texas Tech University). documenting archaeological features, cleaning operations at sites of particular significance, geophysical survey, and artifact analysis and study.2 Most significantly, this season saw the completion of the fieldwalking component of the project, including documentation of all archaeological features within this zone and (at a minimum) preliminary analysis of all ceramic and lithic materials collected (overseen by C. Cloke and J.F. Cherry, respectively) (Figure 1). The intensive survey expanded upon the 2014-2015 work of the project to focus on the middle of the plain (Area d) and the Kastanava valley, yielding new information regarding the main periods of occupation and completing the coverage of this phase of the project.3 The extensive survey explored large areas of the wooded slopes overlooking the plain to the north and south, as well as the surroundings of the Kastanava valley. Digital initiatives in high-resolution mapping and three-dimensional recording of archaeological features (led by Sarah Murray) continued through the investigation of several sites, especially at Kato Kastanava (discovered in 2015) and Eleutherai. Our permit in 2016 allowed us to clear vegetation and surface soil at locations of particular interest, in order to better reveal the form and function of the exposed architecture. Targeted cleaning at the prehistoric site of Kato Kastanava yielded ambiguous results, especially in terms of the architectural remains, some of which are clearly modern; however, the analysis of the pottery and the lithics confirmed the presence of a Late Neolithic/Early Helladic settlement at the site. Cleaning conducted at the Eleutherai fortress facilitated the exhaustive documentation of the walls, allowing for the production of the first comprehensive plan of the site. Further cleaning on the slopes of the hill confirmed the existence of Mycenaean graves and a Classical dam, and allowed for a stone-by-stone drawing of one of the two Late Roman basilicas at the site. Finally, geophysical exploration (led by G. Tsokas) was conducted at the ancient Athenian deme site of Oinoe, mainly in the lower town and on the supposed path of

2 The following individuals were responsible for organizing various aspects of the project and produced detailed reports that were invaluable contributions to the preparation of this overall report: Christian F. Cloke (pottery study), Sarah Murray (DGPS mapping and photogrammetry), Tobias Krapf and Tim Pönitz (cleaning operations at Kato Kastanava), Rebecca Salem (architectural tile); further pottery study and drawings were undertaken by Christine Hunziker and Maura Brennan. Reports on the work of each of the three survey teams were prepared by team leaders: Maeve McHugh and Evan Levine (Team 1), Catie Steidl and Thomas Kerboul (Team 2), and Sarah Craft and Tim Pönitz (Team 3). The geophysical survey team from the University of Thessaloniki was led by Gregorios N. Tsokas, working with George Vargemezis, Elias Fikos, Aristeides Nivorlis, and Panagiotis Tsourlos. 3 On 2014 and 2015 field seasons see, respectively: Fachard, Knodell, Banou 2015; Knodell, Fachard, Papangeli 2016.

2 the western fortification wall, as well as at a Mycenaean site north of Oinoe, discovered in 2014. Priorities for future work include specialist study of certain artifact assemblages, further architectural documentation and geophysical survey at sites and features of particular interest, and, if possible, a limited program of test trenches to supplement the work of the survey. We hope to undertake most of this work during a study season in 2017 and from there move swiftly toward final publication.

Intensive Survey One of the most important achievements of the 2016 field season was the completion of the intensive surface survey, in which all “walkable” and accessible territory in the study area has now been subject to intensive fieldwalking in survey units covered by transects at 10 m spacing (Table 1). Not only is this the first survey in Attica to achieve this level of coverage, it also makes the Mazi Plain one of the most comprehensively documented regions in all of Greece, since our sampling strategy was to aim for as near to comprehensive coverage as possible. We are already able to show broad patterns of ceramic and lithic distribution across the entire landscape (Figures 1 and 2). Fieldwalking in 2016 was undertaken by three field teams working in different areas: in Area d we surveyed the middle of the Mazi Plain, including the vicinity of Modern Oinoe; in Area e work in the southwest corner of the plain focused on the Kastanava valley; a third team worked to close gaps in parts of the survey area left uncovered in previous years, especially in Areas a and c, and undertook targeted gridded collection at sites of particular interest (Aghios Dimitrios and Kato Kastanava).

# of Survey Total Area Average SU Survey Area # of Features Units covered (ha) Size (ha) Area a (2014) 400 204 .51 77 Area b (2015) 441 169 .38 82 Area c (2015) 626 257 .41 48 Area d (2016) 695 222 .32 115

Area e (2015) 800 308 .39 2304 and 2016) Total 2962 1160 .40 552

4 This number includes 37 features from Kato Kastanava, which were given the designation kk instead of e, as in F_kk001.

3

Area d The center of the Mazi Plain (designated as Area d) was the principal focus of the intensive survey in 2016. The majority of this area was surveyed by Team 2, with some work in the southern part by Team 1 and in the northwest by Team 3 (Figure 3). The main feature of note in the survey area is the modern town of Oinoe, which occupies the southeastern part of Area d, and includes some dispersed neighborhoods to the north.5 Finds across this zone were relatively few, with more survey units with zero finds than anywhere else in the survey area. However, several significant concentrations areas and discoveries deserve attention.

Aghios Dimitrios In the northeast corner of Area d, just north of the Modern church of Aghios Dimitrios, a new Byzantine-period site was discovered, consisting of several built structures and a dense pottery scatter (Figure 4). The structures and terraces are arranged along a hillside, west and across a ravine from the apparently contemporaneous settlement at Kondita (investigated in 2015). The two settlements are intervisible and less than .5 km apart. Intensive survey was conducted in the few clearings in the vicinity that allowed for it. The bulk of work at this site consisted of feature mapping and gridded collection. Feature mapping revealed numerous terrace walls and enclosures, and several single-room structures built into terrace walls.6 At least one multi-room structure is also present, but precise mapping and documentation was hampered by thick vegetation and overgrowth (part of the reason for employing a gridded survey strategy, rather than transect-based survey units). The best preserved building (F_d027) was selected for drawing (Figure 5). Gridded collection took place over 72 20x20 m squares, covering 2.88 hectares. In each grid square all artifacts were counted, collected, and weighed, and diagnostic pottery and tile were brought back to base (along with all lithics). All ceramics were collected

5 The name of the village officially switched from Mazi to Oinoe in 1919 (ΦΕΚ 1919/61, see http://pandektis.ekt.gr/dspace/handle/10442/172399). In the 19th century, the Karten von Attika mapped the village under the name of “Masi” (In the Ubersichtkarte and not on a particular sheet, whose coverage ends at the southeastern corner of the Mazi plain). At the same time (1882), however, the village also appears in the Inscriptiones Graecae as Oinoe (IG III 2696 = IG II² 9636). The foundation date of the village is unknown, but the toponym of Mazi is evidently Albanian and should be linked with the dominant Arvanites population of the region from the 17th century AD onwards. 6 This pattern is also observed at the site of Palaeochori Villion, near Villia (see Papakonstandinou 2006).

4 initially for each square, then analyzed in the field, most of it by Christian Cloke and Rebecca Salem (pottery and tile specialists, respectively). Non-diagnostic material was left in the square from which it was collected, while a broad sample of diagnostic material was collected. Vast quantities of tile were noted and collected. Some grid squares contained over 600 fragments, indicating the presence of at least one large roofed structure, and possibly several. The gridded collection at F_d025 also collected a large amount of Byzantine pottery, including many glazed fine wares. This material could form the basis of a substantial study in its own right, and will be especially interesting when viewed in light of material collected from the neighboring site of Kondita (F_c025). The pottery assemblage from F_d025 will be the subject of a more detailed study by Fotini Kondyli in the 2017 season, in combination with further study of the tile assemblage by Rebecca Salem. Just north of F_d025 are two small clearings with wells, the more substantial of which remains an important locus of activity today. This well (F_d022) was inscribed with a variety of graffiti (Figure 6). While the precise dating and transcription are not yet clear, 3D RTI modeling has revealed much of the texts, indicating several visits and engravings by individuals, some likely recording dates and languages other than Greek (Arvanitika?). An inscribed cross appears late antique or Byzantine, although firm evidence for any period besides modern remains lacking, pending further expert analysis. Finally, it is worth underlining that Aghios Dimitrios is the sacred patron of the community of Mazi/Oinoe, which still celebrates its festival on October 26. The Byzantine site, given its importance, could have been the main Byzantine predecessor of the community that later settled the site of Oinoe/Mazi.

Thekeristra - Xylotyrthi This toponym designates a large sector situated immediately northwest of the village of Oinoe/Mazi. It is composed of elongated fields on sloping ground formed by a major deposit cone. In the eastern part of the sector, due north of the village, is a fairly widely dispersed group of high-density survey units with pottery and tile dating to the Early Roman, Late Roman and Byzantine periods. There were also several features present in this zone, including terrace walls, circular structures, and a cistern. This occupation seems essentially agricultural in nature.

Panaghia (Chonika?)

5 A small concentration of Byzantine pottery found north of Thekeristra (SUs d500-507), yet discrete, deserves further attention. It is in this sector that the Choni stream, whose spring emerges below the Petrogeraki summit on Mt Pastra, runs into the plain forming a U-shaped valley, before turning towards the S-E to the village. The concentration of Byzantine pottery (SUs d503 and 507) was found around the apparently modern chapel of the Panaghia, precisely situated on the course of the Choni stream on the 1:5000 maps, at a point where the stream fades away. The name “Choni” bears a strong toponymic link with “Choni-ka,” a place-name recorded in the Mazi plain by the Abbé M. Fourmont in the 18th century AD, and it is reasonable to place this toponym along the course of the stream. This rapprochement is also reinforced by the presence of an “Aghia Panaghia at Chonika,”7 most certainly a church or chapel, whose successors could well be the current Panaghia. The toponym has previously been identified with Oinoe/Mazi by Sironen. However, this identification seems overly hasty. Indeed, had this been the case, it is unclear why Fourmont would not have named the village as Mazi (unless the village was named Chonika in the 18th century, but this would be the only attestation for it); moreover, the old church of the village is dedicated to Aghia Triada and not to the Panaghia.8 Without definitely rejecting the identification of Chonika with Mazi/Oinoe, a location in the sector of the Panaghia is also possible, given the presence of a Panaghia chapel there and the toponymic continuity provided by the name of the stream on the Greek 1:5000 map. The current chapel appears to be modern, but the presence of an earlier shrine cannot be ruled out, especially since Byzantine pottery has been found in the sector. The inscription copied by Fourmont is of particular interest for the history of the Mazi Plain as a step or station on the road between Attica and Boeotia: it is an inscribed miliarium dated to the Tetrarchy (AD 293-305), long forgotten until Sironen published it in 1997.9 Its original position must have been on the ancient Oinoe road (mapped in the Table of Peutinger), attesting that it was a major and official imperial road. The trace of the road might have taken a direct course between Oinoe and the town of Eleutherai, following perhaps the

7 “dans l’ --- ἁγία παναγία στὰ chonika”. Sironen (2008, 32, n°13297=Sironen 1997, 103-104, n°32 bis) copied Fourmont’s Ms. Suppl. Gr. 854, folio 92, mo. 167. 8 An epitaph was found by Fourmont and recopied by L. Ross, in ecclesia Oenes vici, literally “in the church of the village of Oinoe”: IG III 2696 = IG II-III2, 3.2, 9632, where it is said to have been found in sacello vici Ninoi (“in the shrine [roofless?] of the village of Ninoi”). Sironen (2007, 103-104) assumes that Fourmont copied both stones at the same place, but this is contradicted by the toponym of sta Chonika. 9 See note 5. Sironen did not see the inscription, but copied it from Fourmont’s Ms. Suppl. Gr. 854, folio 92, mo. 167. A photograph is provided by Sironen (2008, tab. IX). The stone is perhaps lost.

6 track mapped on the Karten von Attika and passing north of the village of Oinoe/Mazi. It is reasonable to assume that it was the same path followed by Fourmont, who may have seen the chapel of the Panaghia from the road and found the miliarum in this sector, perhaps not too far from its original location.

Rachi Stratonos (Kazarma) This elongated limestone hill (356 m), characterized by a flat summit plateau oriented SW- NE, forms the northernmost extension of the Mt Makron range into the Mazi valley. It creates a natural obstacle, approximately situated in the middle of the plain, and has been chosen by topographers as a probable border landmark separating Boeotia from Attica (Camp 1991; Fachard 2013). It is now covered by a thick maquis presenting very limited opportunities for thorough investigation. However, a concentration of tiles and pottery was found below its eastern tip and included black-glazed pottery (F_d017). Moreover, the extensive survey revealed a complex of architectural features that may comprise a small Byzantine settlement (F_d019). The most prominent building is a church located on an elongated terrace south of the crest of the hill, overlooking the Old National Road (Figure 7). It may be part of the group of 22-24 attested paralavria belonging to the monastery of Osios Meletios, and of which four have been identified so far in the MAP survey area. It could also be the church of a yet unknown hamlet, although, given its position on the western outskirts of the modern village, it seems reasonable to associate it the latter’s first occupation phase.

Area e Area e is the largest of the MAP survey areas and can be broadly split into three zones. The first is the western end of the Mazi Plain proper, with Eleutherai and the Kaza Pass on the western end, and the Kaza stream forming a boundary running from the northwest to the southeast; most of this area was surveyed in 2015. Southwest of this boundary are located two valleys, the Profitis Ilias valley, with the eponymous monastery overlooking a group of small agricultural fields and modern houses (also surveyed in 2015). The Kastanava valley, which the Kastanava ridge separates from the Profitis Ilias valley, was the focus of fieldwork in Area e in 2016 (Figure 8). This valley is a narrow strip of agricultural land occupied by mostly unused and grain fields including a few modern farms. At its western end it forms a bottleneck and then continues, with a drivable track that probably also has a premodern antecedent, out of the survey area to the church of Aghios Georgios and Byzantine settlement of Palaiochori (also Palaiovillia). This site is recorded in the Karten von Attika, and noted as

7 a ruin in the 1835 French Staff Map drawn during the Mission de Morée. Local tradition says that the village burned down by pirates and abandoned in the Ottoman period, more precisely in the 17th century.10 From here, the route joins the “thoroughfare” and ancient tower of Kryo Pigadi, where two important ancient roads meet: the route linking the Megarid with Plataea, also known as “Hammond’s road,” and the road to Aigosthena.11 The Kastanava valley was therefore a natural corridor linking the Mazi Plain to Aigosthena and the Gulf of Corinth. The sparse modern occupation of the Kastanava valley seems to have been characteristic also in the past, with very low densities scattered across the area as a whole. Much of the area, especially south of the road that runs through it, is characterized by very low visibility, which may partially obscure artifacts observed by the survey team. Archaeological features, however, were recorded in some abundance along the northern edge of the Kastanava Valley, often along the ridge immediately outside the area of intensive survey. These features, in combination with others recorded on the Kastanava ridge (see below) suggest the presence of an agricultural community in the area, at least during the Classical, Late Roman, Byzantine, and Modern periods. Features such as terrace walls, small structures, lime kilns, and perhaps a small farmstead (Figure 9) line the lower slopes of the Kastanava ridge, and a few isolated features were also noted on the slopes just south of the Kastanava valley. The most interesting complex of features is located in the southwest corner of the Kastanava valley, which also forms the southwest corner of the survey area as a whole. A small settlement was located here, with its focal point at the church of Aghios Konstantinos (F_e162), currently under restoration by the Ministry of Culture (Figure 10). The wall paintings here bear a remarkable resemblance to those at the church of Aghios Vasileos (Figure 11), located at Kryo Pigadi, near where the continuation of the Kastanava valley path meets the ancient and modern roads to Aigosthena and the Megarid.

Closing Gaps In addition to working in previously unsurveyed parts of Areas d and e, an important goal in this final year of the intensive survey was to fill gaps in Areas a, b, and c, which for various reasons were not surveyed in 2014 and 2015. This included a section in the southwest corner of Area c, as well as small “holes” in other zones.

10 Papakonstandinou 2006.

11 Hammond 1954; Ober 1985, 124–125.

8 The most significant of these was a set of survey units to the west of Ancient Oinoe, across the modern road. These fenced areas are located on privately owned land and were accessed in 2016 through consultation with local land owners, facilitated by Eirini Svana and Polytimi Valta. This zone was especially important to investigate, given its proximity, on the one hand, to ancient Oinoe and, on the other, to the Mycenaean site and Medieval tower located at the entry to the Mazi Plain on its eastern edge. Finds were present in all fields (SUs a383-388), although not in especially high quantities. The owner of SU_a383 also provided a bag of pottery he had collected over the years in SU_a100 and SU_a101, immediately to the north; this grab sample was recorded as F_a061.

Extensive Survey Exploration of the landscape outside of zones subject to intensive survey was an especially important component of the 2016 field season. As in previous years, this work focused on areas not accessible by or practical for side-by-side fieldwalking. This consisted mostly of exploring the hilltops and slopes surrounding the Mazi Plain, as well as exploring crucial routes in and out of it. Areas of particular focus were (1) the ridge along the top of Mt. Pastra, which overlooks the Mazi Plain to the south and Boeotia to the north; (2) the wooded slopes surrounding Oinoe, particularly to the south; and (3) the Kaza Pass that connects the Mazi Plain with Boeotia. Methods consisted of a combination of ground-truthing the presence (or not) of features detected in multi-spectral satellite imagery, systematic (as much as environs allowed) inspection of zones of interest, and walking routes of particular interest while exploring also their surroundings. In addition to the three large areas of interest discussed below, extensive survey was also conducted piecemeal in zones throughout the survey area in order to round out coverage and supplement the work of the intensive survey (Figure 12).

Mt. Pastra A particular target for extensive survey in 2016 was the ridge of Mt. Pastra that forms the geographical boundary between the Mazi and Boeotian plains. The most striking topographical feature of this ridge is the line of modern wind turbines across the top of it. Several features of interest were noted along the ridge in the systematic analysis of satellite imagery that occurred in 2015 and 2016, which we aimed to ground truth through targeted

9 extensive survey.12 This involved driving to the telecommunication antenna station located on the top of Revista (alt. 807.90 m), on the west side of the ridge, and hiking up to the wind turbines road and along its course, inspecting areas of interest identified in the satellite imagery and other loci of note. Ground truthing revealed that all of the round or oblong features identified in the satellite imagery were rubble-built mandria (sheepfolds). These (as well as the many others observed throughout the survey area) could come from a range of periods and obviously relate to pastoral practices, most likely seasonal transhumance. The middle and lower slopes on the south side of Mt. Pastra (facing the Mazi Plain) also seemed to bear evidence for such pastoral practices. Several clearings contained remains of mandria, strounges and wells.13 An encounter with one shepherd, Tasos from Mandra, at the well F_d026 near Aghios Dimitrios, revealed several placenames and a description of contemporary and recent historical patterns of land-use. One location of interest is situated just below the cave of Petrogeraki, at an altitude of 700 m asl. It is composed of a central habitation area, surrounded by smaller buildings used for storing cheese, strounges for milking the animals, as well as a walled well of good construction. The mandri is connected to the plain thanks to a well-built path, some stretches of which are supported by coarse rubble walls. Situated at high altitude, the site was typically used during the summer by a clan of shepherds from Mandra. They would remain there with their families from late April to late August. The slopes of Mt. Pastra were used for grazing and cheeses were produced during the entire period. At the end of the season, the cheeses were brought down to the plain with the help of horses, along the above-mentioned path. Besides grazing, the families of the shepherds (mainly the older men) were also involved in resin gathering. While these activities are difficult to date precisely, we were told that the mandri was in used well before WWII, and perhaps even in the 19th century.

12 Alex Knodell and Liza Davis undertook a program of systematic analsysis of WorldView-2 satellite imagery in the winter of 2015 and 2016 (courtesy of a Digital Globe Foundation Imagery Grant and Carleton College Faculty-Student Research Grant); Sylvian Fachard also contributed to the identification of archaeological features and areas of interest in the 2015 and 2016 field seasons. Extensive survey teams along the Pastra Ridge consisted of Sarah Craft, Davis, Fachard, and Knodell for the 16 June “ridge walk”; Craft, Davis, Piree Iliou, and Teitz continued ground trothing AOIs on 17 June. Extensive survey in the vicinity of Aghios Dimitrios and Kondita was conducted in July, in conjunction with intensive survey and gridded collection in the same area. 13 Mandri can refer generally to a built sheepfold common in rural Greece. The low walls would often be supplemented by branches in which sheep or goats would become tangled and frightened if they tried to jump or climb out. A strounga is also a kind of sheepfold or pen, distinguished from a mandri by its specific link to cheese production, where milked and unmilked animals are separated and let in and out through different entrances.

10 The final location of interest on the slopes of Mt. Pastra (at least in terms of 2016 findings) was a hill-top site of apparently prehistoric date (F_c064), located up-slope and to the north of Kondita and Aghios Dimitrios. This bedrock outcrop has remains of a few walls in situ, along with substantial collapse down the southern slopes. Sherds of handles and cups collected there are apparently prehistoric in date; other sherds, while not particularly diagnostic, could also date to prehistoric periods (Figure 13).

The Kato Maskari Ridge This ridge is an arm of Mt. Makron and forms the southern limit of the Mazi valley, stretching approximately from Kastanava to Modern Oinoe. Given the thick maquis and pine forests that cover its slopes, we investigated this region through a form of systematic extensive survey. Tracts at 25-35 m intervals were traversed through thick pine forest on the north slopes of the hills of Voreino (503 m) and Koryphi (457 m). This inquiry revealed the presence of a limited number of features (a few terrace walls and a limekiln), mostly concentrated on the lower slopes, while most of the higher sloping ground appears to have been left unoccupied, except for a few occasional rubble constructions related to the exploitation of resin. Further exploration of the saddle between the hills of Voreino and Koryphi revealed the presence of a large limestone plateau occupied by a walled mandri of 0.3 ha (f_d081), enclosing several buildings, including cheese huts and strounges. Exploration was extended towards the summit of Koryphi and led to the exciting discovery of a fortified rubble enclosure of 0.28 ha, composed of a double-faced rubble wall, pierced by a probable gate to the southwest. The surface finds included mostly Prehistoric pottery and obsidian, concentrated along the north and south walls. Around the Greek army topographic column at the summit there is an accumulation of rubble, marking a sort of destroyed round structure with Classical-Hellenistic red glazed tiles incorporated in the rubble. The view from the summit is panoramic, covering the entire Mazi Plain from Eleutherai to Oinoe; the site obviously had a strategic significance, being particularly well-suited for observation. Since this important site was situated outside our formal survey zone for 2016, we postponed further investigation to 2017.

Kaza Pass On the western end of the survey area the Kaza Pass runs northwest from Eleutherai to Erythres (Kriekouki), Plataia, and the Boeotian plain. The Old National Road follows this

11 path, overlaying earlier roads, some of which are visible along the course of the current road. This is certainly the main link between the Mazi Plain and Boeotia, although a secondary route through the Portes Pass enters from the east side of the plain, via Osios Meletios. There is a small clearing about 1 km up the pass from Eleutherai, where the Ephorate excavated a building dated to the Roman period based on a coin find (F_e200).14 Further inspection of the architecture and pottery collected during intensive survey of the surrounding fields suggests that an earlier phase of this building was built in the Classical/Hellenistic period, and might have included a tower, suggested by the presence of large limestone corner blocks with drafted edges. Further analysis of the MAP survey and Ephoreia excavation material, along with revisitation of the site, is planned for 2017. Exploration between this area and Eleutherai revealed little aside from a Modern farmstead and a handful of stone piles across a hill slope west of Eleutherai.15 Due north of F_e200, on the left bank of the Kaza stream, we located a long terrace that might have supported the ancient road leading to the pass of Dryos Kephalai (F_e215). Further exploration of the Kaza Pass followed the modern road along the length of the pass from Eleutherai to the Karoumbalo towers, a set of two Classical towers that were clearly built as a pair.16 Located only c. 20 m apart, one overlooks the Kaza Pass to the south, providing direct intervisibility with Eleutherai, while the other overlooks the remainder of the pass as it enters Boeotia, maintaining intervisibility with Plataia, Erythres (Kriekouki), and even Thebes (Figure 14). While outside of the MAP survey area, these surveillance structures are crucial for understanding the relationship of the Mazi Plain to its surroundings, and we believe that this evidence further solidifies that claim that Eleutherai was well tied into a broader Boeotian network of communication and border control.17

East Side of Plain (Bozari) Further extensive survey was conducted in the eastern part of the plain, east of the sector of Stanes Pepas, where an important Classical to Early Hellenistic and Late Roman hamlet was discovered in 2014. This area, known as Bozari (also Isoma), marks a bottleneck-passage

14 Deltion XX 15 This exploration was done by Sylvian Fachard, Alex Knodell, Kalliopi Papangeli, and Eirini Svana on 15 June 2016. 16 Ober 1985, 163–164.

17 See Fachard 2013; Knodell, Fachard, and Papangeli 2016, 150.

12 between the Mazi valley and the Kouloumbi plain. Areas not subject to side-by-side survey were investigated by systematic extensive survey, in which a team of eight surveyors, situated about 35 m apart, followed a series of east-west tracks on the maquis-covered Hill 324, over a distance of 700 m. A mandri-complex (F_b075-77) and two scatters of obsidian were discovered (F_b080 and F_b081). Farther north, the southern slopes of the Mesonychi Hill were investigated, revealing the existence of a multiroom complex, most likely a farmstead (F_a084). Carved blocks were used for the construction of the walls, and a potential basin was located in the southeast room. The very large glazed tile and pottery assemblage throughout is Classical to Hellenistic. Terrace walls were built below, forming a large farm complex. This feature has preserved a complete plan and is one the few examples of farmsteads found in our survey area. It will be drawn and studied in more detail in 2017.

Connection to Skourta A final area of interest for extensive survey is in a location analogous to the Kaza Pass, where the Mazi Plain connects (eventually) to the Skourta Plain from the northeast corner of the survey area. While this zone falls outside of our formal survey area, and no new formal documentation or collection were undertaken, visitation of certain sites and features documented by the Stanford Skourta Plain Project provided new insights into our own datasets.18 Remains of an ancient road are visible about 30 meters northeast of a bend in the modern road that connects the Mazi and Skourta plains, just above a small spring located at the same bend, approximately 1 km due west of the modern village of Panakto (Figure 15). The switchbacks present in this road are somewhat similar to those observed in the Panakton engineered path between Kokkini and Prasino.19 The use of switchbacks is also present in the Eleusis-Oinoe road documented by MAP in 2014 (F_b010).20 We also visited several sites documented by the Stanford Skourta Plain Project.21 In a small valley south-southwest of the road remains we visited the small church of Panaghia

18 We are grateful to Mark and Mary Lou Munn for guiding this venture (27 June 2016), including Sylvian Fachard, Alex Knodell, Kalliopi Papangeli, Eirini Svana, and Polytimi Valta. Formal work in this area is not possible under our current permit, since the Skourta Plain is part of Viotia, Sterea Ellada, not Attica.

19 See Vanderpool 1978; Ober 1985; Fachard and Pirisino 2015.

20 Fachard, Knodell, and Banou 2015. 21 Munn and Munn 1989, 1990.

13 Polonias (site A29 in Skourta terms), with a large Byzantine pottery and tile scatter; a site recorded as a Classical-Hellenistic Farmhouse (A28), which was located based on black and red gloss pottery, as well as pithos and amphora fragments, but the lack of architectures suggests it was probably bulldozed. The most useful analogue in our exploration of Skourta was site B19, associated with the toponym of Patima. This site was noteworthy for its rubble-built masonry, apparently prehistoric pottery, ground-stone celts, and abundance of obsidian both at the site and in the surrounding area, including several blades. This is the closest case for comparison we have for Kato Kastanava, and further work is needed in terms of comparing the Skourta documentation and finds with those documented by MAP at Kato Kastanava.

Feature Documentation Feature documentation in 2016 expanded upon efforts begun in 2014 and 2015. As in previous field seasons, all features encountered in the course of intensive and extensive survey within the designated survey area were mapped, described, photographed, sketched, and catalogued in the project database. A total of 552 archaeological features and feature complexes have now been documented in this way, with 225 new features recorded in 2016. A major focus of the 2016 field season was further work at locations or sites of special interest. Features of special interest were therefore subject to a variety of other modes of documentation, including DGPS mapping and photogrammetric modeling, as well as aerial photography and traditional architectural drawing. Our permit also allowed for cleaning operations at the sites of Eleutherai and Kato Kastanava in order to clarify the architectural plans and (where possible) phasing at these sites (see below). A special program of geophysical survey was implemented in 2016 at and around the site of Oinoe.22 See Table 2 for locations of special interest and the methods employed at each.

22 Differential GPS mapping and Photogrammetry at all locations was supervised by Sarah Murray using, for mapping: a Leica CS25 RTK GNSS dGPS survey unit running corrections broadcast through local cell phone networks through a SmartNet subscription; and for photography: a Nikon D7100 camera with a Zeiss Distagon T* 3.5/18 ZF.2 lens. Aerial photography was conducted by Giannis Asvestas (Ephorate of Antiquities of Attika, Piraeus, and the Islands) using a Phantom 2 drone with a GoPro HERO4 digital camera. Architectural drawings were done by Michelle Berenfeld. The geophysical survey was led by Gregory Tsokas.

14 Specific Location General Type of and Feature Notes Location Documentation Number Photogrammetry; Two models were created in 2016, Southwest Gate DGPS; Aerial before and after cleaning. Photography North curtain walls Photogrammetric and towers (1-6), Model; DGPS interior and Mapping exterior Southeast Gate and Photogrammetric Southeast Gate Model; DGPS Tower Eleutherai Photogrammetry; Fortress Tower 13 DGPS; Aerial Photography Photogrammetric Elevations of Parts Mortared reconstruction and Model; DGPS of Curtain Wall 10 conglomerate on limestone socle Mapping Photogrammetric Eleutherai Dam Model; DGPS Mapping Photogrammetry; Basilica A Architectural Drawing All features with DGPS Mapping F_kk designation Classical? Cleaning; “Farmstead” Architectural F_kk032 drawing Enclosures Gridded F_kk015 and collection F_kk016 Gridded Enclosure F_kk019 collection Kato Kastanava Expanded gridded collection of Gridded F_e088 from 2015; cleaning in NE Enclosure F_kk012 collection; corner and at SW end, revealing an cleaning opening. Rectangular enclosure; cleaning along all four walls with expanded Enclosure F_kk009 Cleaning efforts on east and west sides, and in southwest corner. Cleaning of outline of structure; Enclosure F_kk007 Cleaning surface cleaning in one corner Magnetometry; Oinoe “Lower Town” Ground Penetrating Radar

15 Eastern Slope Magnetometry Eastern Circuit Tomography Mycenaean Site Magnetometry north of Oinoe Mazi Tower F_a073 Photogrammetry; Multi-method documentation DGPS; Gridded team. Architectural drawing in and spot section, accompanied by Velatouri Tower F_b010 collection; perspective view from the north. cleaning; On site analysis of gridded architectural collection of tile, supplemented by drawing grab sample of pottery. DGPS Mapping of all structures; F_d025 Gridded Collection across site as a whole Aghios Dimitrios Photogrammetry, RTI imaging, Well F_d022 Architectural Drawing Architectural Building F_d027 Drawing Architectural drawing here was based on a photogrammetric model created in 2015, Area b Architectural supplemented by a site visit, in Tumulus F_b037 (Karaiskaki?) Drawing order to test the utility and transferability of photogrammetric models to a more traditional medium. Table 2: Summary of Feature documentation at locations of particular interest

Kato Kastanava Kato Kastanava is the toponym associated with what appears to be a major prehistoric site discovered in the course of extensive survey during the 2015 season of MAP. Standing rubble built walls were revealed by a forest fire that affected the eastern edge of a low ridge that divides the Profitis Ilias Valley and the Kastanava Valley in the southwest corner of the survey area. The discovery of obsidian and prehistoric pottery, both in initial grab samples and in subsequent gridded collections, indicated the presence of a substantial prehistoric site, and the large stone enclosures suggested that this site was at a scale unknown in the area. However, the discovery near the end of the 2015 season left many unanswered questions concerning the full extent of the site and the chronology of the architecture.

16 Work in 2016 aimed to complete the total exploration of the site through detailed architectural survey and mapping with DGPS and to clarify the chronology of the stone enclosures through cleaning of vegetation and surface soil.23 The survey revealed several further enclosures and other structures in the wider area, while cleaning clarified the plan, construction, and potential use of structures of several different types (Figure 16). Chronological information remains somewhat fleeting, however. It now seems that at least one of the enclosures is a modern mandri (sheepfold) (F_kk007),24 but further finds associated with other structures, most notably (F_kk009), and throughout the area, continue to suggest the presence of a substantial prehistoric settlement. Gridded collection in the main part of the site was expanded in 2016 to cover much of the eastern complex of enclosures. Finds were similar to those collected in 2015: lithic and pottery remains of a Neolithic to Early Helladic date ranges. Diagnostic materials of the FN and EH I periods were especially noteworthy, according to ceramic specialists T. Krapf and M. Nazou.25 Gridded collection was also undertaken in two further locations, feature complexes noted in 2015 but documented in detail in 2016: first comprises F_kk014-016; the second is F_kk019. Both are groups of enclosures located up a small valley from Kato Kastanava, each occupying one of two sheltered clearings on the south side of the Kastanava ridge. Each also has a small building built into it, and at F_kk019 pottery and tile were found in association with it, apparently early modern. Nothing further was found in these survey grids, although vegetation cover was fairly thick. Cleaning operations were carried out in four locations. The first was the round enclosure designated F_kk007. This was the most substantial of the structures discovered in 2015, although upon clearing heavy brush it appears to be a mandri or perhaps a strounga, directly comparable to several others found throughout the survey area (see above). No finds were found in the course of surface cleaning, which was done at a detailed level in the northeast corner. A single lithic was found within it in 2015, and others in the near vicinity.

23 Sarah Murray conducted the mapping and site survey of Kato Kastanava, along with limited photogrammetry. Cleaning operations were overseen by Tobias Krapf and Tim Pönitz. 24 Because of the number of features at Kato Kastanava, we altered our recording system, designating features here kk###, rather than eXXX as previously. So the feature designations e087 and e088 reported in 2015 (Knodell, Fachard, Papangeli 2016) no longer obtain. 25 Study in summer 2016 confirmed and refined dates described more tentatively in Knodell, Fachard, Papangeli 2016.

17 The second cleaning operation was undertaken at F_kk009, the large rectangular enclosure in the center of the complex. Here walls were cleared of vegetation, while more detailed work and surface cleaning was done on the east and west walls. This served to clarify the plan considerably, revealing a very large enclosure with an opening in the northeast corner. High densities of lithic and ceramic finds were recorded here in 2015 and further finds collected in the course of cleaning in 2016. This may have been a locus of prehistoric habitation, although the stone structural remains are far too large to be a building. More likely the formed an enclosure in which there may have been a habitation location built from more ephemeral materials. The largest enclosure recorded at Kato Kastanava is located some 60 m to the south (F_kk012). Cleaning was undertaken here and at a smaller enclosure immediately adjacent to it (F_kk011). The clearing of vegetation revealed that F_kk012 is indeed a continuous enclosure with openings on either side, including a sort of corridor-style entrance in the northwest corner. F_kk011 is similar in appearance to F_kk007 and may have served a similar function, perhaps as a smaller holding pen for milking adjacent to the much larger animal enclosure. The highest overall number and several high-density grid squares of lithic finds were located in F_kk012. Aerial photography following cleaning documents the site as a whole in much greater detail than we had previously (Figure 17). This complex of features is curious, to say the least. Architecturally, their rubble-built forms have most in common with structures normally designated as early modern farming installations, but the concentration of prehistoric surface finds around them indicates earlier activity as well. This can be explained either in terms of multi-period occupation, with artifacts coming first and structural remains later, or multi-period incidents of use and re-use, quite likely involving rebuilding. At this point the latter seems the most likely explanation, but we still have more questions than answers regarding this intriguing site. After cleaning operations and more detailed study this year we expect that they can only be answered through test excavation in order to obtain stratigraphic information and, perhaps, finds associated with initial phases of occupation or construction. We suggest this as a priority for rescue excavation for the Ephorate, perhaps in 2017.26

26 We note that a plan of terrestrial photogrammetry was also planned for Kato Kastanava in 2016. Unfortunately this was impossible due to the company we planned to work with being unable to the necessary equipment. This methodological approach remains of interest for future work.

18 A final feature of interest in the area is F_kk032, a small building, also rubble built, on the south side of the Kastanava ridge (Figure 18). This has been tentatively interpreted as a small farmstead, due to apparently Classical/Hellenistic pottery and tile finds.

Eleutherai Mapping and documentation at Eleutherai focused on the digital and traditional documentation of several features, including some previously known and some identified in 2015. Cleaning operations were undertaken at the fortress itself, namely in the creation of a comprehensive plan, extensive work on the two main gates of the site, the rediscovery of a Mycenaean grave below the fortress, the cleaning and drawing one of the basilicas located between the fortress and settlement, and documentation of a large dam in the ravine located northeast of the fortress. These operations have dramatically enhanced our knowledge of the site and will result in a detailed architectural and archaeological restudy of the site.27

Fortress At the site of the Eleutherai fortress, the main goal of the 2016 season was to complete the stone plan begun in 2015 (Figure 19). Sections of the walls were cleaned, including both gates and a part of the ancient road. The entire fortress was mapped and documented using photogrammetric modeling and RTK DGPS mapping.28 The existence of at least three phases of construction can now be described: To the first phase (I) should be associated several sections of polygonal retaining walls found midway between the summit and the south wall of the fortress (see figure 19). This first phase includes a large bastion-like structure with drafted edges. East of this point the wall is poorly preserved, but its trace can be followed towards the NE for another 70 meters, where it probably met the north wall, later reused by the walls of phase B (curtain C7). The west limit of this ensemble cannot be determined, but it might have enclosed the flat plateau west of the Polygonal building at the summit. The masonry of the wall, consisting of polygonal limestone blocks carved locally, is similar (although less careful) to the masonry of the Polygonal building excavated by Stikas.29 This construction has a different orientation

27 IN prep, to be submitted to Hesperia.

28 This work was undertaken by S.C. Murray.

29 Stikas 1938.

19 than the fortress wall but was respected and saved when the fourth-century curtain was constructed. According to Chandler, it was a tower, built “considerably earlier in date than the main fortification, and must have belonged to the old Boeotian town of Eleutherai.”30 Based on the masonry and the earliest Iron Age surface pottery, it seems reasonable to assign a date in the second half of the 5th or early 4th century. The extended documentation of this structure and associated remains demonstrate that this building was not isolated, but part of a larger fort defending the summit plateau of the hill. To the second phase (II) can be assigned the spectacular walls of the much-admired fortress, most often dated to the 4th century BC (see below). The fortress had a maximum length and width of 237x117 m, enclosing a surface of some 3 ha. It consisted of 13 towers and 14 curtain walls. Three posterns were built on the north and east sides, while two monumental gates allowing the passage of carts were positioned on the southeast and southwest sectors (Figures 20 and 21). The masonry is dominantly trapezoidal on the west, north and east fronts (with polygonal and ashlar blocks), and all blocks are carved in the local hard limestone of the hill. Limestone was used on the south wall as well, but here the constructors also relied on conglomerate, extracted in the stream beds surrounding the hill. Some curtains of the south wall have a limestone socle with a levelling course, intended to support an elevation made of a different material: the final choice was conglomerate, but this does not exclude the possibility of a mud brick elevation. It is worth noting that conglomerate blocks were easier and cheaper to carve and could be almost uniformly prepared in the quarry. It was therefore quicker to build part of the wall with regular conglomerate blocks than to carve individually trapezoidal and polygonal limestone ones. The use of conglomerate on the south wall, naturally protected by vertiginous cliffs, would have provided an economic relief in the overall construction budget. It might also be interpreted as a sign of haste in construction, a change of plans, the presence of a team of stone masons skilled in conglomerate walls, or the eroding of funds. A stretch of the ancient road crossing the fortress was cleaned and mapped precisely in 2016 (Figure 19). Its existence was first noted by Ober during fieldwork at the site.31 The road was functional by Phase II at the latest, and would have been used during the

30 Chandler 1926, 12. It is certainly too large and elongated to be a proper “tower”: instead, its internal plan, divided into rooms, favors an identification with a strong house or building. It is generally accepted that this building is older than the fortress (phase II).

31 Ober 1987, 213

20 construction of the south wall to bring the conglomerate blocks to the wall from the stream bed east of the Basilicas. An earlier date is possible and would explain the will to build the fort of Phase I just above it. We believe that this was part of the northern branch of the road, while a south branch followed the stream in the ravine. The presence of two branches of the same road is not uncommon. The Sacred Road to Eleusis had two different branches after the sanctuary of Aphrodite. Some branches were steeper and quicker, while others were more suitable to carts. It seems likely that ascending carts would have taken the north branch while descending vehicles used the ravine. At the Eleutherai bottleneck, traffic in the ravine might have been disturbed by falling rocks (still very common today) and winter torrential floods continuously eroding the terrace walls supporting the ancient road. Taking the north branch avoided these obstacles and offered a gentle and regular ascension. A date of the fortress in the 4th century BC is commonly held. More precise dating points to a date in the second quarter of the fourth century BC,32 but a date in the last third of the century have been suggested as well.33 The fortress has been mostly associated with an Athenian military presence, but recent evidence has suggested a Boeotian construction and related with the period of Theban hegemony under the lead of Epameinondas.34 It is important to stress here that this period offers a valid historical context for the construction of the fortress, but that a later date cannot be excluded based on the masonry and the architecture. The walls of New Halos, built at the very end of the 4th century BC, offer good comparison; moreover, Greek stonemasons did not stop carving trapezoidal blocks after the 4th century, and we have excellent and well-dated parallels for such masonry in the very end of the 3rd century as well (New Pleuron). The surface survey, in particular, SU_e428 and SU_e429, brought to light an excellent assemblage of Late Classical to Hellenistic Pottery. The inscription at the gate appears to date to the second half of the 3rd century and indicates that the fortress was occupied at that time, which in turn reminds us that several repairs might have taken place on the walls in the Hellenistic period. Although a construction date between c. 375 and 300 BCE seems preferable, only an excavation could resolve this debated issue. The third phase (III) is represented by repairs on the south wall as well as the modification of both the Oinoe and Plataea gates. These repairs involve the use of spolia, mostly blocks from Phase II reinserted in the walls (often in a position in which they were not

32 Ober 1985; Camp 1991, 2000, 46.

33 Beschi 1968, 19; Winter 1971, 162 n. 44.

34 Camp 1991, 2000; Fachard 2013. See also Adam 1982; Ober 1985, 1987; Harding 1988, 1990.

21 originally meant to be), some capstones from the crenellation, fragments of tiles, and, more important, mortar. The chinking of stones and tiles between reused blocks are characteristic of Justinianic walls.35 The abundant LR pottery from the site makes a Justinianic phase of repairs possible, without excluding a Byzantine occupation, as indicated by the presence of pottery from this period.36 Finally, the walls of the fortress were damaged in the Early Modern and Modern periods. A large limekiln (F_e121), 5 m in diameter and over 3.6 m high, was found west of the fortress and is likely responsible for much of the damage suffered by the SW section of the fortress. This lime kiln was exploited by people from Villia and the lime was sold on the main road to Thebes (from the Kaza tavernas).37 Further damage to the walls took place during WWII, and many German cartridge shells attest to the combat that took place in this strategic zone. At the end of the war, the fortress was used by the ELAS forces, who initially resisted a British assault before beating a retreat. The assault included tanks and Spitfires, whose firepower might have resulted in considerable damage.38 The rubble (including barb wire) below the destroyed curtain C6 and the towers T6 and T9 could have resulted from direct hits. A monument commemorating events from WWII is located on a ridge immediately across the Kaza Pass from the Eleutherai fortress.

Mycenaean Graves A Mycenaean cist-grave (F_e210), excavated by the Ephorate in 1984,39 was situated on the south slope of the hill, just above the modern road. The cist grave is made of two parallel slabs on the north and south sides, with rubble walls forming the east and west limits (Figure 22). It reportedly contained LHII pottery. The presence of graves in this ravine is best explained by the existence of a road or path at that time. This conclusion seems to confirm the existence of communication axis on the Late Bronze Age through the Kaza pass (see conclusions). Euripides (Suppliant Women 757-760) mentions the location of the burials of the common soldiers of the Seven Against Thebes episode as being located “under the shady rock of Eleutherai” and Plutarch (Theseus 29) mentions that these graves were visible at

35 Daly 1942, 506 36 Edmonson 1966, 58; Ober 1987, 215

37 Information provided by the owner of the Kaza taverna, who held it from his grandfather.

38 We thank Dr Roderick Bailey (University of Oxford) for sharing this information with us.

39 E. Baziotopoulou-Valavani, ADelt 40, 1985, 46. See AntK 59 (2016) 147.

22 Eleutherai during his own time. It seems reasonable to assume that some prehistoric graves were visible here in the historical period.

Basilica A The two Early Christina basilicas excavated in 1939 by Stikas are situated on the lower eastern slopes of the hill, where the ancient road exiting through the Oinoe Gate of the fortress reached a natural terrace overlooking the stream.40 Cleaning operations and documentation focused in 2016 on Basilica A (Figure 23). The masonry of the building uses blocks from the fortress exclusively, including window blocks. The narthex is entered from the southwest through a monumental gate; there are subsidiary rooms to the west of the narthex that do not appear to open through the south, north or west wall. The apsidal room in the northeast corner may have been a baptistery; the masonry is composed of reused blocks, stones, and mortar. It would be interesting in the future to compare this mortar with that used at the Eleutherai fortress, mostly at the monumental gates, the south walls, and in the sector of the cistern, directly facing the basilicas. Stikas describes the Basilica as a typical modest country church due to the lack of any sculpture or decoration. Interestingly, he suggests that the settlement would have moved to this area in Early Christian times; this is not unusual in later Roman settlements, in particular with the increasing dominance and appeal of churches that originally stood outside settlements, especially those that were constructed on burial grounds over the remains of saints and martyrs. In 2017 we hope to clean and map Basilica B, located immediately adjacent to Basilica A. Further work in this area should investigate the nature of the settlement (if any) around the Basilicas and its precise connection with the fortress, which might have played the role of refuge for the population of LR Eleutherai.

Dam Northeast of the basilicas, the large polygonal wall discovered by the stream in 2015 (F_e148) was entirely cleared of vegetation and documented with photogrammetry. The masonry consists of massive limestone blocks precisely fit together. Only the east face of the wall, 12 m long, is visible; its profile is inclined, indicating that it acted as a retaining wall of some sort. Given its position in the bed and right bank of the stream, it is best interpreted as a dam. Its function would have been to protect the town of Eleutherai and its surroundings from floods, possibly providing a balancing reservoir as well. This dam is a remarkable feature,

40 Stikas 1939.

23 relatively rare in the countryside, whose most sophisticated parallel is the Inopos reservoir in Delos (Fincker and Moretti 2007).

Geophysical Survey at Oinoe Geophysical investigations were carried out at the ancient settlement and deme center of Oinoe. The goals of this operation were to locate potential streets and buildings and to gather evidence regarding the position of the Late Roman and possibly Classical fortification walls of the lower town. Magnetometry was employed, using a fluxgate gradiometer to investigate two sectors of the lower town, totaling an area of c. 6800 sq meters. One sector of 1600 sq meters was selected for Ground Penetrating Radar (GPR), focusing on the potential presence of several streets and a buried structures. Moreover, 27 electrical resistivity tomographies (ERTs) were carried out in various sectors of the site. Most of them were positioned on the supposed line of the west wall in order to locate its trace. An anomaly was detected in sectors W1, W2, and W3, suggesting the existence of a large wall approximately oriented north- south which can tentatively be associated with the Classical west fortification wall of the settlement running down from the upper town (Figure 24). In W3, another wall oriented NW- SE seems to join this section: based on its orientation, it could link up with the Late Roman defensive line visible further north. At this point, it seems that two different traces should be distinguished: a Classical-Hellenistic wall to the east, and a Late Roman extension to the west. Last, a limited magnetic survey, using the same methods described above, was conducted at the Mycenaean site located northwest of Oinoe, covering a surface of 2400 sq meters. These preliminary investigations were aimed to assess the applicability of various methods of investigations and to assess the general profile of this sector. They will be followed by more systematic operations in the winter of 2016, most notably in the form of an electric resistivity survey, a method much better suited to the particular environment of the Mazi Plain.

Velatouri Tower The final location of special interest in 2016 was the Velatouri tower,41 situated on the summit of Hill 532, also known as Pyramida. The difficulty of access of this particular site

41 Ober 1985, 157–158.

24 (first visited by a survey team in 2014), requiring a 1.5 hour hike each way from the nearest drivable location, led to the design of a special expedition in order to document the site as thoroughly as possible in a single day (see Table 2), including perspective and section drawings, as well as photogrammetric modeling (Figures 25, 26, 27).42 The topographical significance of this location is paramount, as the tower commands a panoramic view from the Bay of Eleusis to the Mazi Plain, including several of the main land routes through the mountains. Other features of particular significance visible from the Velatouri tower are: the Trikerata, the Kandili Pass, the valley leading to Paleochori, parts of the Thriasian Plain, Panakton, Eleutherai, Oinoe, the Portes Pass, and the Monasteries of Osios Meletios and Aghion Nipion. In terms of finds, an informal gridded collection was conducted on site by dividing the area surrounding the tower (in and around the remains of a structure tentatively identified as a baracks) into eight roughly equal units. 811 tiles were collected weighing 80.3 kg. Two different sets of tiles are represented, dating from the Classical to Hellenistic periods. This date range corresponds with the pottery collected at the site, and more detailed study may reveal more about the phasing, pottery use, and roof construction of the buildings. One 4th century lekane fragment collected from the opening in the top of the tower joins with another collected outside the tower in 2014. In the course of this expedition Hill 513 was also explored, intervisible with Velatouri and located to the northeast across a small valley. A tower was reported in the general area by Baziotopoulou-Valanaki in the Deltion, but we now believe this to be a reference to the Velatouri tower.43 No tower remains were found on the summit, but there was a curious, long, linear feature of stone rubble running across the top of this hill for some distance. Its function, if any, is unclear, but at times it runs parallel to a dirt track, suggesting perhaps it is the result of bulldozing for the track. A track is mapped along this course on several maps (even where none is visible on the ground or in satellite imagery) and leads eventually to Aghios Vlassios. The course of the ancient carriageable road between Eleusis and Oinoe crosses this path at the saddle between both hill, before descending westwards to Sterna

42 This outing (6 July 2016) was coordinated by Sylvian Fachard and Alex Knodell, accompanied by Ephoreia representatives Eirini Svana and Polytimi Valta, who supervised limited cleaning and pottery collection, Sarah Murray and Jacob Morton for photogrammetry, Michelle Berenfeld for section drawing, Chloe Bergstrand from perspective drawing, and Rebecca Salem for tile analysis.

43 ADelt 45 (1990) 68.

25 Oikonomou, in the eastern section of the Kouloumbi plain. This section of the ancient road was mapped in 2014.

Diachronic Discussion and Conclusions In terms of coverage, the 2016 season brought to a conclusion the fieldwalking survey of the entire Mazi Plain. In the span of three seasons, 11.6 sq km were investigated using an intensive method, while the main landmarks, hills, and summit surroundings the plain were explored using extensive methods. In total, 2962 survey units were laid out in the plain: 63569 pottery sherds, 39452 tile fragments, and 377 lithics were counted. Perhaps most importantly, all artifacts collected in the course of the project have now been subject to (at a minimum) preliminary analysis, which makes the diachronic discussion below possible.44 A summary of our main results is provided here, highlighting some issues of particular interest and objectives for future work.

Prehistory The first observable occupation of the Mazi Plain seems to have occurred in the Late Neolithic and Early Helladic I periods. The main settlement was situated at Kato Kastanava, at the confluence of several streams, loosely stretching over several hectares on the first limestone slopes of the hill. Some large rubble enclosures were built, perhaps grouping together dwelling(s) and animal sheds. Another site, perhaps fortified, existed on the summit of the Pournari hill (alt. 457), and will be investigated in 2017; it is located 900 m southwest of the Neolithic hunting site discovered in 2014 (F_a017). Overall, it seems that the early occupation of the plain is best characterized as small-scale communities of agro-pastoralists who also undertook limited hunting activities. Based on the quality of the pottery, the community of Kato Kastanava cannot be compared with the large sites found in the rich farmland of the south bank of the Asopos valley in Boeotia, or with those of East Attica. The

44 The overall program of artifact analysis was organized and overseen by Christian F. Cloke. Christine Hunziker took the lead on organization in 2015 (when Cloke was not present) and continued to assist in 2016. John F. Cherry is responsible for lithic analysis, which he undertook in 2014 and 2016. Rebecca Salem began a study of the architectural tiles collected by the project in 2016. Fotini Kondyli contributed to ceramic analysis in 2015 as a specialist in Byzantine material culture, and Aude-Linne Pradervand and Marc Duret also contributed to pottery analysis in 2015. Guy Ackerman and John Papadopoulos have been invaluable consultants on Hellenistic and Archaic and Prehistoric materials, respectively. In 2016 Margarita Nazou and Tobias Krapf provided especially important insights into material from Kato Kastanava, while Lindsey Fine examined other prehistoric material. Specialist study of material will continue in the 2017 study season.

26 Mazi Plain was perhaps used for summer pasture by communities farming the rich agricultural surfaces of the areas of Plataea and Kriekouki, and parallels are perhaps to be found in the communities occupying the Skourta plain at these times. Following the EH I period, no traces of occupation have been found until the Late Bronze Age. Such a gap has also been recorded in the Skourta plain, which again offers an interesting parallel case. In the Mycenaean period, the region seems to be reoccupied on a larger scale. One site has been located in the eastern part of the plain, just northwest of the ancient site of Oinoe. It is positioned on a sloping ground descending towards a small stream, at the junction of Pleistocene and Holocene deposits. Abundant ceramic evidence points toward an LH IIIA/B occupation, stretching over some 4-6 ha. It is hoped that geophysical investigations will reveal further information. Another LBA site seems to have acquired considerable importance at that time: situated north of Aghios Dimitrios, in the north-central area or the plain, a hilltop (F_c064) produced a small number of very striking ceramic finds: these included several probable prehistoric coarse ware sherds, short stems that may belong to LBA cups, and other atypical pieces. The high quality of this material and the topographic prominence of its findspot suggest that this hill was a place of local importance in the Bronze Age, and it should certainly be investigated further in 2017. Other findspots include Hill 302, just above the Tzamas stream, east of the modern village of Mazi, and the acropolis hill of Eleutherai, where two cist-graves were excavated by the Ephorate on its southern slopes and a kylix stem was found at the summit. The nature and extent of the Mycenaean occupation in the Mazi plain is still difficult to assess, but the survey has brought new light to this region, especially in terms of its role as a crossroads, with the discovery of new sites at the major entry points on each side of the plain. This fresh evidence, combined with the study of communication networks in Attica, suggests that the Mazi plain was situated on the main route linking Eleusis to the Mycenaean palace of Thebes.45 Cosmopoulos has suggested that Eleusis could have been a dependent of Thebes, “given the economic significance that Eleusis had for Boeotian trade.”46 If Eleusis was indeed Thebes’ main hub to the Saronic Gulf and Aegina,47 then the Mazi plain was a crucial point on this route, and the Eleutherai bottleneck might have acquired its strategic importance at that time. Following the collapse of

45 Fachard and Knodell, forthcoming. 46 Cosmopoulos 2015, 125.

47 Cosmopoulos 2015, 103.

27 palatial Thebes and the decline of Eleusis, the Mazi Plain does not seem to have any detectable settlement, but the route must have remained in use.

Early Iron Age and Archaic The total absence of pottery evidence between the Mycenaean and Archaic periods is surprising, as it is difficult to believe that no human presence tried to exploit this valuable land for so many centuries. The absence of archaeological footprints might not reflect the reality, as it is notoriously difficult to recognize Proto-Geometric and Geometric pottery on surveys. Excavation would likely be useful here, and it may be worthwhile in the future to lay out exploratory trenches in the settlements of Oinoe and Eleutherai to put this supposed absence to the test. The presence of Geometric pottery has been claimed at Antiope’s Cave,48 but we have none in our current assemblage. In our current state of knowledge, only Archaic pottery is attested in the Cave of Antiope and goes back to the late 7th-early 6th century BC. It is plausible that the settlement of Eleutherai was also occupied at this time, but no firm evidence exists. Recent and remarkable epigraphical discoveries from Thebes indicate that Eleutherai and Oinoe existed in the last quarter of the 6th century,49 and Oinoe became a Cleisthenic deme after 509/8. Therefore, the earliest archaeological evidence for both settlements should go back the last quarter of the 6th century at the latest. Athenian military involvement beyond the Kaza Pass is well attested in our sources, as early as 519 BC, when they fix the boundaries of Thebes on the Asopos. Plataia and Hysiai are then under Athenian protection and influence, as is Eleutherai. The first political boundaries between Thebes and Athens seem to appear at this time. In 506, the Boeotians attack Oinoe and temporarily occupy Eleusis (SEG 56, 521). But the Athenians strike back and seem to push the borders of Attica back to the Kaza pass. Eleutherai may have been under Athenian control at that time, and our earliest sources already attest to the border dynamics that make this landscape so interesting.

Classical The Classical period sees a dramatic increase of occupation and exploitation in the Mazi plain. The site of Oinoe was a deme, fortified before 431 BC and used as an Athenian

48 E. Baziotopoulou-Valavani reports the presence of Geometric pottery at the Cave of Kissos above Eleutherai (ADelt 45 [1990] 68), which we take as the cave of Antiope.

49 Aravantinos 2006; Matthaiou 2014.

28 strongpoint during the Peloponnesian War. Good ceramic evidence for the 5th century has been found at the settlement, which stretched over some 4-6 ha. The town was accompanied by at least one hamlet at Stanes Pepas, with Attic fine wares present. In the western part of the plain, habitation was centralized at the large settlement of Eleutherai. The latter was not fortified, but we now have enough evidence to suggest that a fort was built above Eleutherai on the acropolis hill in the second half of the 5th century. Activty at the Cave of Antiope continues throughout the Classical period into the Roman period. The political affiliation of Eleutherai in the 5th century is a complex historical matter. Thucydides shows that Oinoe – and not Eleutherai – was a border town of Attica in 431, which suggests that Eleutherai was then part of Boeotia. Oinoe fell to the Boeotians in the later years of the War but was handed back to Attica after the defeat of 404 BC. These were years of conflict, with renewed military operations, large movements of armies, repeated depredations in Attica, including the flight of slaves. According to the literary sources, most of the rural population of Attica was supposed to have been living inside the city-walls of Athens, but this has yet to best tested by archaeological data. MAP is adding an interesting perspective to this historical period in terms of settlement patterns and occupation, and is already revealing more information about the communities living in this strategic and repeatedly attacked Attic border deme. The 4th century marks a peak in occupation and exploitation. In terms of settlement patterns, the central hubs of habitation in the Mazi Plain remain Eleutherai and Oinoe, but a series of hamlets emerge around them, following a trend that started in the 5th century. Around Oinoe we have hamlets at Stanes Pepas, the Mazi tower, and the Kouloumbi plain. Several farmsteads have been found (at least three, possibly four). Near Eleutherai, a substantive hamlet existed east of the town (around SUs e348, e349, e350, and e352). Two imported transport containers were found here: the double handle of a Koan-type amphora (e350_P005) and the toe of a Knidian amphora (e350_P006). The simultaneous study of finds from Eleutherai and this hamlet may help to delineate settlement hierarchy during this period and to assess inhabitants’ contacts and commerce with the wider Mediterranean world. A farmstead with a possible tower (F_e200) seems to have existed 1.7 km to the northwest of Eleutherai, just south of the ancient road leading to the Kaza pass.50 Given the small number

50 Originally excavated by the Ephorate and interpreted as a Late Roman villa, this feature makes use of abundant building material from an earlier date, including several large limestone blocks with drafted edges. Some of the earlier foundations are in situ and it seems that a Classical/Hellenistic compound, perhaps with a tower, was later enlarged in the Roman period. Pottery (e700 and e701) is Classical, Hellenistic, and Roman.

29 of recognized Classical farmsteads in the survey area, it seems that this form of inhabitation and exploitation was not widespread in the Mazi Plain. Instead, the evidence suggests that a double-nucleic model prevailed, with secondary settlements gravitating around two main hubs. The fourth century is a period of extreme political tension and economic competition between Attica and Boeotia. These tensions leave a signature in the landscape in the form of large-scale fortifications at Eleutherai and Oinoe. The presence of two massive fortifications in the same plain and only 6 km apart is awkward and bears no parallels in Boeotia and Attica. It is best explained by the existence of a political border in the middle of the plain. The position of this border has been tentatively placed on the Rachi Stratonos (Camp 1991; Fachard 2013); it might have continued north, following the streams that cut the plain in its center. Ceramic densities are among the lowest in the plain (excluding the high densities around the Byzantine settlement of Aghios Dimitrios) and this may relate to the existence of a border space that could have been exploited for agriculture but not for permanent habitation. The Eleutherai fortress can be interpreted as a demonstration of Theban force, mostly upsetting Athens in the second and third quarters of the 4th century as the dominant power of mainland Greece. The walls of the fortress have no parallels in Attica, but many in Boeotia (Camp 2000). In reaction, the Athenians were probably compelled to strengthen and update their defenses at Oinoe: a new circuit is built on the “acropolis,” well-defended by several massive towers, including the largest found in Attica outside Athens. It certainly hosted a garrison, under the authority of the strategos ths choras stationed at Eleusis. This fortification phase likely dates to the second half of the 4th century BC and bears many parallels with the Long Walls rebuilt at that time (conglomerate ashlar blocks, headers and stretchers, etc.). The walls at Oinoe offered a secure refuge for the rural population of the deme, scattered throughout the eastern half of the plain and in its surroundings valleys and hills.

Hellenistic At the end of the 4th century, occupation in the Mazi Plain might have suffered from the military operations at Panakton, Eleusis, and Phyle between Kassander and Demetrios.51 In the early 3rd century BC, robbery, and murder are mentioned in the Kythaeron, and it is said that “the region of Eleutherai was all-desolate because of the wars” (Lucian, Dialogues of the

51 Fachard 2013: 85

30 Dead, 27.2). These troubled times perhaps forced the Athenians and Boeotians to remove their garrisons from Oinoe and Eleutherai, resulting in some security vacuum. However, Early Hellenistic pottery is still found throughout the plain, at Eleutherai (fortress and settlement), Oinoe, as well as most of the hamlets occupied in the 4th century. Based on the pottery evidence, it would be hard to note any difference between the end of the 4th century and the beginning of the third. However, pottery evidence does increasingly become scarce after the period 250-200 BC, with the exception of the Cave of Antiope.52 More precisely, its seems that no clearly identified 2nd-century pottery has been found at Oinoe or in the other hamlets of the deme. This does not mean that these sites are abandoned, but the pottery evidence stands in sharp contrast with earlier periods and indicates a very discrete human presence at these locations. Throughout the entire survey area, we have very limited evidence for the 2nd and 1st centuries BC. This trend, in sharp contrast with earlier and later periods, corresponds to what has been observed in other field surveys in Greece, including Boeotia and southern Attica, where most of the farms in the deme of Atene seem to have been abandoned in the mid-Hellenistic period.53 Although closer study of the pottery is necessary, we temporarily conclude that occupation in the Mazi plain progressively decayed in the 3rd century BC and reached a low point in the 2nd and 1st centuries BC.

Roman This declining curve seems to continue through the ER period, but it is certainly reversed in the MR and LR periods, with abundant evidence of revival throughout the entire plain. This, of course, has much to do with LR pottery recognition, but the overall distribution of LR pottery, in particular, suggests a level of inhabitation comparable to the Late Classical period. The town of Oinoe is occupied and perhaps extended with a new fortification wall in the lower town. The previous defenses on the acropolis are also repaired and modified at that time, possibly for the needs of a garrison. These building operations show that an important community is seeking security behind new walls in the 5th and 6th centuries AD. At the Eleutherai fortress, walls are repaired, and the plans of the monumental gates are modified, testifying to the large-scale operations taking place in the plain at this time. Several hamlets

52 Ritual activity at the cave was apparently long-lived, as shown by the button base of a 2nd century BC moldmade bowl and several other later Hellenistic pieces. 53 For Greece generally, see Alcock 1993; on Boeotia: Binliff and Snodgrass 1985; Bintliff 2012:313; on Atene: Lohmann 1993.

31 and even farmsteads of the Classical-Early Hellenistic periods are also reoccupied. One of the largest dense scatters of pottery discovered during intensive fieldwalking was in the north central part of the Mazi Plain. Recovered in these SUs were many sherds from Late Roman vessels, suggesting a strong Late Antique in this stretch of the valley. The considerable size of this scatter indicates a site of larger scale, perhaps a hamlet or small town, and further study of this pottery should help to fine-tune our understanding of the site’s chronology and the variety of activities taking place in this vicinity. Borders were not an issue anymore for some time: in the 2nd century AD, Pausanias noted that Eleutherai was part of Attica and it most likely remained so until Late Antiquity. The disappearance of the border zone at that time could provide an explanation for the progressive occupation and settlement of the north- central section of the plain in the MR to LR periods (2nd-7th centuries AD).

Byzantine The next boom in the long-term history of the plain begins in the 11th century, when the Monastery of Osios Meletios became one of the major monastic centers of Central Greece. This was founded by a Cappadocian monk then based at Thebes, and the influence of this site reorganized communication routes through the mountains, promoting in particular the route through the Portes pass, dotted with small affiliated churches such as Aghioi Theodoroi. By the 12th century, the settlement pattern in the Mazi Plain had changed dramatically. Eleutherai and Oinoe had been abandoned, and two new hubs of settlement emerge. The first is a sort of two-part zone comprised of the sites of Kondita and Aghios Dimitrios. This location high on the hills of Mt Pastra, overlooks the plain from above while being safely hidden from the main road. The second is located on the western end of Area b, at near two previously known churches connected to Osios Meletios – Aghios Georgios and Aghia Paraskevi. The settlement at Aghios Georgios is spread over several terraces. A full complement of fine and coarse wares indicates the presence of a sizable settlement. A particularly interesting find was this cross-stamped body sherd, probably from an amphora or a storage vessel. The stamp in the form of the cross can be seen in examples of other Middle Byzantine amphorae and is associated with the transport and selling of agricultural goods – especially wine, oil, and garos. The second substantial Byzantine settlement is in the vicinity of the Kondita Frankish tower, which produced large quantities of Middle and Late Byzantine pottery, including glazed decorated tablewares with well-established chronologies. The location of the site, on the northern limits of the plain and significantly above the flatter, agricultural land below, is

32 typical of Byzantine settlement patterns, in contrast to the concentration of Classical and Late Roman settlement at Oinoe to the south. The idea was to be somewhat more secluded, away from the main road, able to observe from above without being noticed from below – this is perhaps an artifact of earlier periods when the Mazi Plain was a less friendly crossroads. The substantial size of the site is also noteworthy, with a clearly defined ceramic scatter covering an area of some 30 ha. This picture was expanded even further with the discovery of the neighboring settlement of Aghios Dimitrios in 2016. The pottery suggests activity in the area both before and after the arrival of the Franks, who came after the Conquest of Constantinople in 1204. Vessels for food production, fine table wares, and a large number of storage vessels point to fairly large-scale habitation and agricultural production. The settlement at Kondita appears to be the largest Middle Byzantine and Frankish settlement of the Mazi Plain. Three main zones then stand out for the Byzantine period. Osios Meletios and the church of Aghioi Theodoroi are located in the mountains as the main monastic center resting above the plain one of the main routes north (different than in previous periods). Dependencies then emerged in the Koloumbi plain at Aghios Georgios and Aghia Paraskevi along the southern route. Kondita and Aghios Dimitrios, however, seem to represent the largest hub of settlement.

Modern The evidence of the modern period is chiefly of rural production, mostly related to pastoralism, agriculture, or resin production. These remains are less impressive, but no less significant. The arrival of Arvanites (Albanian) populations in this region is dated back to the 15-16th century. The villages were then known as the Dervenochoria and were granted some level of autonomy during the Turkish occupation. These villages again controlled the passage between central Greece and the Peloponnese, a strategic role they kept until the War of Independence—yet another example of the long-term significance of this region as a crossroads and borderland.

33 Figures

Figure 1: Map of the Mazi Archaeological Project Survey Area showing Survey Units, overall ceramic densities (pottery+tile per hectare), archaeological features, and sites or other places of special interest. (map by A. Knodell)

34

Figure 2: Lithics found in each survey unit (map by A. Knodell)

Figure 3: Area d map (map by A. Knodell)

35

Figure 4: Gridded collection and feature outlines at Aghios Dimitrios (map by S. Craft)

Figure 5: Aghios Dimitrios building drawing (drawing by M. Berenfeld)

36

Figure 6: Well above Aghios Dimitrios (F_022) and inscription (enhanced model by S. Murray)

Figure 7: F_d019 structure above Modern Oinoe (photo by M. McHugh)

37

Figure 8: Kastanava valley and ridge map (map by A. Knodell)

Figure 9: Drawing of Farmstead F_kk032 (drawing by M. Berenfeld)

38

Figure 10: Church of Aghios Konstantinos (F_e162) (photo by A. Knodell)

Figure 11: Comparison of wall paintings from the Churches of Aghios Konstantinos and Aghios Vasileos (photos by A. Knodell)

39

Figure 12: Extensive Survey tracts with Mandri and well locations (map by S. Fachard)

Figure 13: Pottery collected at F_c064 (photos by N. Piree Iliou)

Figure 14: View down Kaza Pass from Karoumbalo Towers plus photo of tower 2, taken from tower 1, looking toward Boeotia (photos by A. Knodell)

40

Figure 15: Ancient road remains in pass connecting the Mazi and Skourta plains (photo by S. Fachard)

Figure 16: Kato Kastanava area plan (entire ridge) (map by A. Knodell)

41

Figure 17: Kato Kastanava Aerial Photo (photo by G. Asvestas)

Figure 18: Farmstead(?) at Kato Kastanava (drawing by M. Berenfeld)

42

Figure 19: Eleutherai Fortress (drawing by S. Murray and S. Fachard)

Figure 20: Eleutherai East Gate (photo by G. Asvestas)

43

Figure 21: Eleutherai Plataia Gate (photo by G. Asvestas)

Figure 22: Mycenaean Grave at Eleutherai (photo by A. Knodell)

44

Figure 23: Basilica A at Eleutherai Drawing (drawing by M. Berenfeld)

Figure 24: Geophysical survey locations at Oinoe (map by G. Tsokas)

45

Figure 25: Velatouri Tower perspective drawing (drawing by C. Bergstrand)

Figure 26: Velatouri Tower section drawing (drawing by M. Berenfeld)

46

Figure 27: Velatouri Tower 3D model (model by S. Murray; see www.maziplain.org/media for the interactive version)

47 References

Adam, J.-P. 1982. L’Architecture militare greque. Paris. Alcock, S.E. 1993. Graecia Capta: The Landscapes of Roman Greece. Cambridge: Cambridge University Press. Aravantinos, V.L. 2006. “A New Inscribed Kioniskos from Thebes.” BSA 101: 369–377. Beschi, L. 1968. “La fortezza ellenica di Gyphtokastro.” In Les fortifications depuis l’antiquité jusqu’au Moyen-Age dans le monde méditerranéen, Acta of the VIIIth Scientific Congress of the International Castles Institute, Athens 25-29.4.1968, 1–19. Camp, J.McK. 1991. “Notes on the Towers and Borders of Classical Boeotia.” American Journal of Archaeology 95(2):193-202. ———. 2000. “Walls and the Polis.” In Polis & politics: studies in Ancient Greek history: presented to Mogens Herman Hansen on his sixtieth birthday, August 20, 2000, edited by, P. Flensted-Jensen, Th. Heine Nielsen, and L. Rubinstein. Copenhagen: Museum Tusculanum Press, University of Copenhagen. Chandler, L. 1926. “The North-West Frontier of Attica.” Journal of Hellenic Studies 46(1):1- 21. Cooper, F.A. 2000. “The Fortifications of Epaminondas and the Rise of the Monumental Greek City.” In City Walls: The Urban Enceinte in Global Perspective, edited by J.D. Tracy, 155-191. Cambridge: Cambridge University Press. Cosmopoulos, M.B. 2015. Bronze Age Eleusis and the origins of the Eleusinian Mysteries. New York, NY: Cambridge University Press. Daly, L.W. 1942. “Echinos and Justinian’s Fortifications in Greece.” AJA 46: 500–508. Edmonson, C. 1966. “The Topography of Northwest Attica.” Ph.D. diss., University of California, Berkeley. Fachard, S., 2013. “Eleutherai as the Gates to Boeotia.” Revue des études militaires anciennes 6: 81-106. Fachard, S., A.R. Knodell, and E. Banou. 2015. “The 2014 Season of the Mazi Archaeological Project, Northwest Attica, Greece.” Antike Kunst 58: 178-186. Fachard, S., and D. Pirisino. 2015. “Routes out of Attica.” In Autopsy in Athens: Recent Archaeological Research in Athens and Attica, edited by M. Miles, 139-153. Oxford: Oxbow. Fachard, S., and A.R. Knodell. Forthcoming. “Modeling Mobility in Mycenaean Attica.” In Athens and Attica in Prehistory, edited by J.C. Wright, N. Papadimitriou, S. Fachard, and N. Polychronakou-Sgouritsa, with E. Andrikou and E. Banou. Athens: Melissa. Fincker, M., and J.-C. Moretti. 2007. “Le barrage du Réservoir de l’Inopos à Délos.” BCH 131, 187–228. Hammond, N.G.L. 1954. “The Main Road from Boeotia to the Peloponnese through the Northern Megarid.” Annual of the British School at Athens 49:103-122. Harding, P. 1988. “Athenian Defensive Strategy in the Fourth Century.” Phoenix 42(1):61-71. ———. 1990. “Athenian Defensive Strategy Again.” Phoenix 44(4):377-380.

48 Knodell, A.R., S. Fachard, and K. Papangeli. 2016. “The 2015 Mazi Archaeological Project: Regional Survey in Northwest Attica (Greece).” Antike Kunst 59: 132-152. Matthaiou, A.P. 2014. “Four Inscribed Bronze Tablets from Thebes: Preliminary Notes.” In The epigraphy and history of Boeotia: new finds, new prospects, edited by, Nikolaos Papazarkadas, 211–222. Brill studies in Greek and Roman epigraphy. Leiden ; Boston: Brill. Munn, M.H., and M.L.Z. Munn. 1989. “Studies on the Attic-Boiotian Frontier: The Stanford Skourta Plain Project, 1985.” Boeotia Antiqua I: Papers on Recent Work in Boeotian Archaeology and History, edited by J.M. Fossey, 73-127. Amsterdam: J.C. Gieber ———. 1990. “On the Frontiers of Attica and Boiotia: The Results of the Stanford Skourta Plain Project.” In Essays in the Topography, History and Archaeology of Boiotia, edited by A. Schachter, 33-40. Teiresias Supplement 3. Montreal: Department of Classics, McGill University. Ober, J. 1985. Fortress Attica: Defense of the Athenian Land Frontier, 404-322 BC. Leiden: Brill. ———. 1987a. “Pottery and Miscellaneous Artifacts from Fortified Sites in Northern and Western Attica.” Hesperia 56:197-228. Papakonstandinou, M.A. 2006. Τα Βίλλια. Στο πέρασµα του χρόνου (1821-1949). Athens: Melissa. Stikas, E.G. 1938. “Ανασκαφή “Ελευθερών (Πανάκτου),” PAAH: 41-49. ———. 1939. “Ανασκαφή “Ελευθερών” PAAH: 44-52.

Vanderpool, E. 1978. “Roads and Forts in Northwestern Attica.” California Studies in Classical Antiquity 11:227-245.

49