Excavation and publication of the ancient fort at Phylla Vrachos, on Euboea, is a most welcome contribution to scholarship on fortifications in the Greek countryside. This otherwise undistinguished, rubble-walled hill fort possesses a remarkable feature: a long, twenty-room stoa-like building dominating the center of the site. Such a large and orderly structure is uncharacteristic of the rough-and-ready constructions found at most rural forts and encampments in Greece. What is the date of this structure? What was its purpose? Thanks to the collaboration of a Greek-British team, headed by E. Sapouna Sakellaraki on the Greek side and by J.J. Coulton on the British side, and with the pottery expertise of I.R. Metzger of the Swiss School at Athens, we are now certain of its general date (late archaic), we know more about its structural form, and we can recognize intriguing possibilities as we consider its historical context and purpose.
The fort is situated some 9 km east of Chalkis, on an eminence along the Vrachos ridge above the village of Phylla, overlooking the lower end of the Lelantine plain and a wide stretch of the Euboean coastline between Chalkis and Eretria. The site is directly north and 3 km inland from the prehistoric settlement at Lefkandi. Before the present investigation, which began with a survey of the site in 1994 leading to excavations in 1996 and 1999, the fort was known chiefly from the brief notice in A.W. Lawrence’s Greek Aims in Fortification (Oxford 1979), pp. 176-77, reproducing a plan published by its earliest investigator, G.A. Papavasileiou, in Archaeologiki Ephemeris 1903, pp. 131-34. The fort at Phylla also figures in a discussion of the historical topography of central Euboea in an article on the Euboean campaign of the Athenian general, Phocion, by L. Tritle in Klio 74 (1992), pp. 141-46. The present study was the first to identify the full range of periods of activity on the Vrachos site represented by surface finds of the Early Helladic, Geometric, Archaic, Hellenistic, and Byzantine periods, and the first to identify through excavation the major periods of construction and occupation at the fort, namely, the late eighth century, and late sixth to the/early fifth century BCE.
After a brief introductory chapter, two chapters describe the excavations and discuss the structural form of the features revealed in excavation (all three chapters authored by Coulton). Two following chapters (in German), by I.R. Metzger, describe the fragments of pottery vessels and lamps recovered in the three seasons of surface survey and excavation. Chapter 4 is a narrative description of these finds as they relate to their excavated contexts, and chapter 5 is the descriptive catalogue of many of the diagnostic pieces (222 sherds), nearly all of which are represented in profile drawings, accompanied by a small selection of photographs. Chapter 6, by Coulton, presents a short catalogue of stone and metal artifacts. Chapter 7, co-authored by Coulton, Metzger, A. Sarpaki (University of Crete), and S. Wall-Crowther (Oxford), summarizes the slight botanical remains (a single grape pip identified from the flotation of six soil samples), the animal bone and shell remains (out of 43 bone and 16 shell fragments, elements of cattle, pigs, sheep/goat, fish, and marine invertebrates were identified), and discusses diet, food preparation, and drinking customs that can be inferred for this site, based on the recovered pottery, in comparison to other excavated ceramic assemblages, and on general studies of dietary practices in ancient Greece. Chapter 8, by E. Sapouna Sakellaraki, offers a useful overview of the present understanding of prehistoric and classical sites in the region of Phylla Vrachos, covering most of the lower Lelantine plain and the coastal region eastward to the Kakorrema valley, some 4-5 km west of Eretria. The final chapter, by Sapouna Sakellaraki, Coulton, and Metzger, draws together all of the foregoing into a discussion of the chronology, nature, and historical role of the fort at Phylla Vrachos, focusing chiefly on its evident function as a garrison post at the end of the sixth and beginning of the fifth centuries. A two-page summary in Modern Greek is provided at the end.
The site is defined by a ruinous wall of unworked boulders enclosing an area of 230 x 80 meters on three side (cliffs protect the summit on the north side). Despite the efforts of the excavators in two small trenches against this wall, no certain evidence of its date was discovered. A more carefully built cross-wall of roughly-worked stones, preserving remains of a gateway and a postern, reduced the enclosure to approximately three-quarters of its original area. Within this reduced enclosure, traces of three smaller buildings were noted in addition to the long building (designated Building 3 by the excavators). Excavation was carried out within three of the twenty rooms of Building 3, and within the single room of Building 4. With their foundations visible on the surface, excavated deposits within these rooms were shallow (usually well under 1 meter before reaching bedrock). Nevertheless, floors and deposits of sherds above and below floor level were discernable in several sectors, as were deposits of tiles and clay from the collapsed roofs. Buildings 3 and 4 prove to have been built and occupied during the late archaic phase of the site, and the same is likely to be true of the more ruinous small Buildings 1 and 2 as well. Slight traces of walls from other buildings were noted, and these, it is suggested, belonged to buildings ruined and abandoned by the time Buildings 1-4 were built (presumably, therefore, belonging to the eighth-century phase).
The interpretation of Building 3, the stoa-like building, is central to an understanding of the site in the late archaic period. In size (approximately 112 x 10 meters overall, including the terrace in front) it approaches the dimensions of the Stoa of Attalos in the Athenian Agora. It is constructed in four blocks of five rooms each, each room 4.5 x 5.9 meters internally, with narrow passages separating the blocks. A terrace 2.5 meters wide runs the length of the building along its south side, and doorways open from each room onto this terrace. The doorways appear to be slightly off-center, but their poor state of preservation does not allow any regular pattern to be determined. Coulton estimates that each room could accommodate between nine and twelve couches or sleeping pallets (depending on their arrangement), so that the entire structure would accommodate about 200 men.
One of the excavated rooms contained a small, rectangular hearth-like structure, composed of re-used roof tile fragments, in its center. No ash or other evidence of burning was associated with this feature, however, and no other internal structural features were preserved within the three rooms excavated in Building 3. All three rooms contained fragments of Corinthian-type roof tiles, and in one case also fragments of Laconian-type tiles. The Corinthian tile fragments were present in sufficient quantity and sizes to reconstruct the roof system, which Coulton does in chapter 4. Composed presumably of local fabric, the tiles include regular pan tiles and eaves pan tiles, two varieties of regular cover tiles as well as eaves cover tiles, and examples of a system of ridge pan and cover tiles. The Laconian tiles are assumed to belong to repairs to this original Corinthian tile system.
In a reconstruction drawing, Coulton depicts the building as four separate five-room structures adjacent to a common walk-way along the terrace. If, however, a series of posts along the terrace supported a continuous roof in front of the line of rooms, the appearance would, in fact, be that of a large stoa. By Coulton’s own calculations, the dimensions of the roof tiles make a roofed colonnade rather more likely than the unroofed terrace (8 rows of tiles would fit on either side of the ridge tiles if the roof covered the terrace, as opposed to 5.5 rows of tiles on either side if only the rooms were roofed).
The three excavated rooms of Building 3 each produced a similar array of pottery (comparable material also came from Building 4). Drinking cups predominate (chiefly glazed or banded stemless cups and skyphoi), after which, by this reviewer’s estimate, sherds of kitchen and cooking ware abound (chytrai, unglazed or banded jugs and table amphoras, and semi-glazed or banded lekanai—quantitative information is imprecise, see below). Lamps are regularly present. Smaller numbers of glazed jugs (oinochoai, olpai) and small bowls are found, as are unglazed storage vessels or small pithoi. The vast majority of these are identified by Metzger as Euboean in manufacture, comparable in form and fabric to examples she has published from Eretria (see Metzger 19851). Fragments of at least two cup-skyphoi and a skyphos are Attic, and two black-figured cup sherds are Boeotian in manufacture. One uncatalogued skyphos sherd is identified as possibly Corinthian.
Drinking and dining clearly went on in these buildings, and cooking was done on the site (although no cooking hearths have been identified). It is noteworthy that no grinding implements (mortars or querns) were identified, although a few stones may have been used for grinding. Also noteworthy is the absence of diagnostic transport amphora sherds (the excavators suggest that wine and water was brought to the site in skins). At the lowest levels, and mixed in with the late archaic sherds, are also sherds of Late Geometric cups (at least 24 represented in the catalogue), and a few fragments identifiable as Protogeometric and Sub-protogeometric (at least three catalogued). Five Early Helladic sherds from surface collections are catalogued, as are three Byzantine/modern sherds.
The pottery is all quite fragmentary. There appear to be few or no joins, and with the exception of three lamps, no complete profiles are preserved. Metzger makes the best of this material, cataloguing a generous selection of pieces that preserve diagnostic shape (usually rims and feet, illustrated in profile drawings) or decoration. Published comparanda are cited for approximately half of the 222 catalogued sherds. Comparanda come chiefly from the excavations at the Athenian Agora,2 Olympia,3 Tocra, and from Eretria. Most of the datable comparanda (some 60 items) fall within the range of 550-475, with a definite clustering of comparanda in the vicinity of 500 BCE. The case for dating the construction and occupation of Building 3 (and with it Buildings 1, 2, and 4, and probably the secondary cross-wall of the circuit) to sometime close to the turn of the sixth to the fifth centuries is thus reasonably secure. The excavated assemblages correspond better to military mess units (discussed below) than they do to civilian households, a characteristic that also suits the barracks-like arrangement of the major building on the site. The further facts that no loom weights or spindle whorls were found, while a three-bladed bronze arrowhead and stones that could have been prepared for slinging were recovered in the rooms of Building 3, all lend strong support to the interpretation advanced by the excavators, namely, that the occupation of this phase was that of a military garrison. Satisfactory as the presentation of excavated pottery is in most respects, there are a few deficiencies. Catalogue descriptions are spare, and fabric descriptions are especially so (no Munsell soil color numbers are given, for example, to specify descriptions such as “Ton braun” or “Ton beige”). A single measurement is usually given, but sometimes no measurement is reported at all. A more serious deficiency is the fact that counts of catalogued sherds are tabulated and reported, both by Metzger in chapter 4 and by Coulton in chapter 7, as statistically representative of the material preserved on the site. The selective nature of the catalogue (and its bias in favor of glazed pottery) is evident, however, when the reader observes how often, in chapter 4, Metzger mentions sherds that are not catalogued. So, for example, there are frequent references to uncatalogued sherds like: “zahlreiche Wandfragmente ungefirnisster Küchenkeramik” (p. 48), “Boden- und Wandfragmente von chytren und Amphoren” (p. 49), and occasional lists like: “zahlreiche Fragmente von Chytren, von offenen und geschlossenen gefirnissten und ungefirnissten Gefässen, Schalenfragmente… Randstücke und Henkel von Skyphoi, ein Randstück einer Oinochoe mit leicht verdickter Mündung, Wandfragmente von Amphoren oder Hydrien, Fragmente einer niedrigen Lampe…” (p. 49). No effort is made to relate numbers of sherds to whole vessels (by counting only rims, for example). Statistical reckoning based on inventories of fragmentary pottery is always difficult to manage sensibly (as Coulton acknowledges, pp. 92-94, referring to figures from other sites; for the description of a reasonable method, see S.I. Rotroff and J.H. Oakley, Debris from a Public Dining Place in the Athenian Agora, Hesperia supplement xxv, Princeton 1992, pp. 131 and 137). In the present instance, however, the numbers and percentages offered by Metzger in table 4, p. 58, and by Coulton in tables 7.4 and 7.5 on p. 93 and in figure 7.1 on p. 94 for the finds from Phylla Vrachos cannot be considered useful.
A yet more serious problem vitiates the analysis of dining and drinking customs presented by Coulton in chapter 7. The problem arises from the manner in which conventional terminology in pottery classification has led to undue assumptions about usage and functions, compounded in this case by an unrecognized idiosyncracy of the pottery descriptions by Metzger. The authors believe that, although drinking cups were ubiquitous in the late archaic phase of the fort at Phylla Vrachos, mixing bowls for wine were not. Only one catalogued late archaic sherd, a foot fragment, is identified as possibly from a krater (K 145, probably a lekane foot). This has led Coulton to suggest that drinking at Phylla Vrachos in the late archaic period went on in “an informal setting,” where “each man may simply have mixed wine and water to his taste in his own cup, making a special mixing vessel unnecessary” (pp. 94-95). Coulton’s hypothesis about such an uncanonical drinking custom is unnecessary, however, because mixing bowls actually are relatively abundant among the sherds recovered at Phylla Vrachos. They are the bowls that Metzger has called “lekanides”. It is unfortunate that Metzger has adopted the diminutive form, “lekanis”, instead of the normative “lekane” to describe an undecorated, or simple band-decorated mixing bowl (the same practice is adopted in her publication of the material from the Eretria Thesmophorion, Metzger 1985). This has led Coulton to interpret these “lekanides” as the sort of small bowls, often lidded, that are frequently associated with women’s toiletries, but also with food service (p. 96). Yet the catalogue descriptions, dimensions, and profiles of 15 different rims and three different feet (including K 145 and K 204 in these counts), make it clear that these are what Sparkes and Talcott call “lekanai” (1970, p. 211), which previously were called “semi-glazed kraters”, variations of which Gauer calls “Kratere” and “Reifenkratere” (1975, pp. 134-41). To quote Sparkes and Talcott: “That the shape [i.e., the lekane] was used as a formal mixing-bowl for wine, a krater, as well as for all sorts of domestic purposes, is obvious from many of the representations” (1970, p. 211 n. 1).
A few errata should be noted. In figure 4.2, p. 59, the identifying numbers on the sherds labeled “(c) K 69” and “(d) K 155” should be switched. In figure 5.1, p. 77, the drawing of the rim labeled K 26 should not have a center line (the diameter is indeterminate). On pp. 91 and 96, references to “Schwartz 1995” (cited three times) correspond to no entry in the bibliography.
This review, like the book, closes with observations on the historical situation of Phylla Vrachos. The late-eighth century occupation coincides in time with the abandonment of the nearby settlement at Lefkandi, and both events call to mind the tradition of the Lelantine War, when Chalkidians and Eretrians fought over the adjacent Lelas plain. It is a reasonable guess that the walls of the main enclosure at Phylla Vrachos were built when the site was inhabited in the late eighth century. The excavators suggest that this site was a stronghold controlled by the Eretrian side in this conflict. But whether it was a military outpost or a refuge for a civilian population in a time of danger is unclear. In either event, this period of habitation was not long, to judge by the lack of seventh-century material.
The late archaic occupation, involving construction on the scale of a major public building and the activities of a garrison numbering probably some 200-odd men (or significantly more, allowing for slave attendants), was also not very long-lasting. When the site was abandoned in the early fifth century it was left in ruins and untouched, apparently, except for limited salvage operations (it is likely that many of the roof tiles were recovered), and was probably only occasionally visited by small numbers of people across the following two and a half millennia. This short-lived episode of military construction and activity on the borders of Chalkidian and Eretrian territory (the exact location of the frontier is uncertain) corresponds remarkably well with the interval between the defeat of the Chalkidians and colonization of their territory by the Athenians in 507/6 (Herodotus 5.77), and the destruction of Eretria by the Persians in 490, when the Athenian cleruchy was withdrawn (Herodotus 6.100-101). The excavators are surely right to conclude that the late archaic remains on this site were the product of publicly organized activities related to the presence of the Athenian cleruchy in Chalkidian territory. Their caution about being more specific in their interpretation is also justified. Was this a stronghold maintained by the Athenian colonists to secure the outlying portions of their new territory? Or was it an Eretrian outpost, securing a new frontier after the defeat of their Chalkidian enemies by the Athenians? The finds from Phylla Vrachos, specifically the pottery and roof tiles, suggest no remarkably close connection with Athens or Attica, so the physical evidence for Athenian involvement is inconclusive.
The most striking and definite feature to emerge from this investigation of Phylla Vrachos is the evidence for a rational organization and deployment of a military force by civic authorities (be they Athenians, Athenian cleruchs, or Eretrians). The evident division of this force into eating and drinking units calls to mind the sussitia of Lycurgan Sparta. The number of such dining units accommodated in Building 3, twenty, calls to mind the decimal system of civic and military organization that the Athenians adopted at the proposal of Cleisthenes at the very time of the Chalkidian cleruchy (as the excavators point out, Eretria had a five-fold military organization). Rational organization of soldiery in this manner, and particularly the organization of sussitia and their installation in garrison posts, was an ideal that found later expression in the thought of Plato ( Republic 416e, Critias 112b, Laws 625e, e.g.) and Aristotle ( Politics 1264a, 1331a, e.g.). Aristotle even points out that such dining groups promoted an anti-tyrannical spirit ( Politics 1313a). Discussions of the background and historical development of such political theorizing now needs to take account of the physical manifestation of such concepts at the end of the archaic period at Phylla Vrachos. For this result, and for the range of detailed findings that this project has documented, the excavators and authors of this work are to be congratulated.
1. I.R. Metzger, Eretria vii, Das Thesmophorion von Eretria: Funde und Befunde eines Heiligtums, Bern 1985.
2. B.A. Sparkes and L. Talcott, The Athenian Agora, xii, Black and Plain Pottery, Princeton 1970.
3. W. Gauer, Olympische Forschungen viii, Die Tongefässe aus den Brunnen unterm Stadion-Nordwall und im südost-Gebeit, Berlin 1975.