Published in Hesperia 61 (1992), 265-289. 

EXCAVATIONS AT HALAI, 1990-1991

The earlier excavations were carried out from 1911-1935 by Hetty Goldman and Alice Walker (later Alice Walker Kosmopoulos). Annual campaigns took place from 1911 until 1914; these were described in a preliminary report and the inscriptions discovered were published separately. Goldman and Walker also both completed Ph. D. dissertations on finds from the tombs excavated in the first four campaigns. Further campaigns took place in 1921, 1923, 1931 and 1935. In 1931 Goldman and Walker were assisted by Virginia Grace. No preliminary reports for these campaigns were issued, although Goldman published a few finds from the acropolis. 

Although the work of 1911-1935 was never published in full detail, Goldman eventually produced a long and invaluable article summarizing the Greek and Roman finds on the acropolis and co-authored another on the terracotta sculpture from the tombs. The Greek and Roman finds from the earlier excavations were deposited in the Archaeological Museum in Thebes and the Neolithic finds in the National Archaeological Museum in Athens. 

In recent years Dr. Robert Bridges, current Secretary of the American School of Classical Studies at Athens, has been studying the Neolithic finds from the earlier excavations and he expects soon to publish them. He has kindly put the results of his study to date at the disposal of our expedition. In 1988 and 1989 the Greek Archaeological Service excavated several dozen further tombs to the east of the acropolis (at G in Fig. 2). 

Before the current excavations began, preliminary activities were carried out for several years, including the establishment of a surface survey area of about 185 sq. km. centering on Halai (Fig. 1), intensive surface survey in and around Halai (Fig. 2) and the preparation of a new plan of the acropolis of Halai at 1:50 scale (which formed the basis for Fig. 3). 

The surface survey of 1988-89 suggests that an additional outer ring of fortifications was built around the town of Halai in the fourth century B.C., or in Hellenistic times. The wall would have followed the outer line of the stippled area in Figure 2. The line of blocks extending SE-NW along the present beach to the northwest of the acropolis (Fig. 2 at B) is more plausibly identified as part of this outer fortification wall than as the remains of ship-sheds, as has often been proposed. 

These preliminary activities, as well as the actual excavations at Halai, are part of an interdisciplinary project of survey and excavation in East Lokris (the Cornell Halai and East Lokris Project) directed by the undersigned. The excavations themselves have been supported in part by grants from the National Endowment for the Humanities, the Townsend Fund of the Department of Classics, Cornell University, and the Institute for Aegean Prehistory as well as by private donations. 

The aims of the excavations are broad. Since Neolithic occupation is still poorly documented in Central Greece, we are devoting about one-half of our time and effort to the Neolithic remains. Our goal is to recover as much information as possible about the Neolithic town, especially evidence pertaining to everyday life and relationships with other towns and regions of Greece. 

Classical and later Halai offers unique opportunities to study a small town typical of those in which the majority of the ancient population of Greece may have lived. The acropolis, with a length (NW-SE) of ca. 160 m. and a width of ca. 70 m. (Fig. 3; Pl. 1), is small enough to be thoroughly examined, even with painstaking modern techniques. Furthermore, it is almost completely accessible and the earlier excavations have already revealed many of the major architectural elements. Although ancient references are few, there is sufficient information to reconstruct the main lines of its history. In Archaic and Classical times Halai was a member of the East Lokrian league and subsequently, probably in the late 4th century B.C., it joined the Boiotian league. Destroyed by Sulla in 85 B.C., it was soon resettled and it evidently continued as a town throughout Roman times until the 6th century. In its later years it became Christianized and after the abandonment of the town a church probably continued to exist on the spot until the 12th or 13th century. A charming inscription of the 3rd century B.C., found virtually complete by the earlier excavators in a well near the temple area, gives a complete roster of the titles and names of the 29 current local officials of the town. Two producers of comedies are included, as well as three women identified as "weavers of hangings," perhaps used to adorn a local statue of Athena. The information given by this inscription is an almost ideal complement to that we hope to gain from actual excavation and study of the Hellenistic town. 

Excavation in 1990 and 1991 took place within the part of the acropolis of Halai set aside and fenced off by the Greek Archaeological Service (Figs. 3, 4). For purposes of recording, this part of the acropolis has been divided into ten Areas designated A, B, C, etc. (Figs. 4, 5) . The excavation campaigns of both years were especially concerned with the earliest and latest occupations of the site, the Neolithic and the Late Roman/Early Byzantine. The 1991 excavations also focussed on the remains of the Archaic period (Area A) and, although to a lesser extent, on those of the Hellenistic period. 

The Neolithic Period

In all trenches where Neolithic deposits have been encountered, they immediately underlie deposits of the Archaic period or later and the uppermost Neolithic layers usually contain many small worn sherds suggesting that there was a long hiatus during which the site was abandoned. To date we have found not a single recognizable Bronze Age sherd at Halai, either in our surface survey or during excavation. The report of Simpson and Dickinson of the existence of Bronze Age material is therefore almost certainly erroneous. 

The Neolithic deposits are most accessible in the temple area, our Area F, at the northwest side of the site, because the earlier excavations removed later deposits from above them. Before our excavations began, two pits from the earlier excavations were visible passing through the Neolithic deposits (Fig. 5; see Pl. 1, b): a larger one near the center of the temple area to the south of the foundations of the Archaic temple, which represents the combined trenches excavated on behalf of Goldman by Virginia Grace in 1931 and which we have not yet re-excavated, and a smaller one immediately adjacent to the SW corner of the large church, about which we have found no records and which became our trench F3. 

 By the end of the 1991 season three of our trenches (F2-F4) had been combined to form an area of ca. 44 sq. m. and two others (F5 and F6) had also been combined to form a further area of ca. 32 sq. m. (Fig. 5). 

 Trench F3, at the southernmost end of combined trenches F2-F4, has been cleared almost completely to bedrock, which was encountered at 0.74 masl. Several cuttings are visible in the bedrock (Pl. 2, a). Most of the deposit was debris that has fallen in since the earlier excavations, but in the lowest 0.20-0.30 m. some undisturbed deposit was encountered containing pottery of Early Neolithic types ("Rainbow" ware; EN Black Burnished ware). 

Elsewhere in combined trenches F2-F4 we encountered what we take to be domestic architectural remains representing several building phases and, set against the outer side of one wall, the stone built base of an oven or hearth (Pl. 2, b, indicated by arrow). Around the oven is an ashy deposit. The pottery associated with these remains can be dated not later than an early stage of the Late Neolithic period. Deeper ashy deposits that predate the extant hearth are probably to be dated to a late stage of the Middle Neolithic, although the pottery has yet to be examined in detail. 

The pottery from Late Neolithic levels in trenches F2-F4 includes Black Burnished wares like those of the Tsangli-Larissa phase in Thessaly, Matt Painted wares, and red on white painted pottery like that from Elateia. A few sherds of coarse paste with white filled incised and punctated decoration have their closest parallels with finds from a cave at Tharounia in Euboia (Adamantios Sampson, personal communication). 

Trench F5, to the north of F2 and immediately between the foundations of the Archaic temple and the Archaic altar, passed through 0.05 m. of mixed Archaic-Neolithic deposit to reach the top of a puzzling scatter of small stones (Pl. 2, c). The stones fall within a band ca. 3 m. in width extending E-W from one scarp of the trench to the other. The top of the band is generally higher toward the center than at the edges. We are as yet uncertain whether this band of stones was a deliberate structure (perhaps the base of boundary wall?) or whether it is simply the result of post-Neolithic erosion. 

Stone built walls of Neolithic date were also encountered beneath the Archaic levels in Area A, trench 2 (Fig. 5; Pl. 3, d) and beneath Hellenistic levels in Area H, trench H1 (Fig. 5; Pl. 5, b, indicated by arrow). Until the pottery from these trenches is further studied, these walls cannot yet be assigned to particular phases of the Neolithic era. 

All soil from Neolithic deposits is being water sieved under the supervision of Julie K Near. Although detailed analysis of the botanical finds has yet to be carried out, preliminary examination suggests that the crops produced by the inhabitants of Neolithic Halai were similar to those of other contemporary communities. 

Obsidian is abundant in Neolithic levels and was clearly worked at the site; chert also occurs in smaller quantity (see Appendix 1). Other finds include fragments of terracotta figurines (Pl. 2, d: A seated female; B face, possibly of an animal; C, D human lower legs); pierced shells probably from a necklace (Pl. 2, d, E); and stone celts (Pl. 2, d, F, G). The heavy fraction from flotation includes many tiny stone beads in a variety of materials and shapes. To date, no prehistoric material has been recognized in survey or excavation which we would date later than the LN I phase. We expect to collect several samples for radiocarbon dating in the 1992 season. 

The Archaic Period

Structures of the Archaic period were revealed in all four of these trenches. In A3 (Pl. 3, a), which measured 4x4 m., a stone wall-socle appeared with one course of mudbricks still in situ. The wall is probably that of a house. To the SW, between this wall and the probable line of the street, i.e. outside the house, four round platform-like structures were encountered. That to the left in Plate 3, a is partly surrounded by flat stones creating a shallow channel. The use of the platform-like structures is not yet clear, although it is possible that they were granaries, given their resemblance to similar structures found in Geometric levels at Lefkandi. 

 Trench A4 revealed part of another Archaic house, perhaps of megaron plan, and a building adjacent to it, the plan of which is difficult to interpret (Pl. 4, a). 

Trenches A1 and A2 were tests beneath the floors of the rooms of the westernmost of the two large "North gate buildings" constructed probably in the 4th century B.C. (Pl. 3, b). In trench A1 a pavement, a wall, a probable drain and part of a mudbrick construction came to light (Pl. 3, c). In the nearby trench A2 the corner of a lightly built structure was excavated (Pl. 3, d). Its walls are made of unworked stones, set on edge to form the inner face. A rectangular slab placed on end was found within the structure and probably dates to the period of its use. The character of these remains is not yet clear. The pottery associated with them, while not plentiful, appears to date to the 6th century B.C. 

 It is notable that all the Archaic structures excavated in Area A are similarly aligned and that this alignment corresponds generally to that of the fortification walls and the main NW-WE street. The Archaic temple in Area F, on the other hand, is oriented almost exactly E-W and is thus on a slightly different alignment, although the altar to its east more nearly lines up with the rest of the Archaic structures. The evidence so far available suggests that there was a substantial community in the Archaic period within the acropolis at Halai. How densely packed the acropolis was with buildings, whether it was laid out on a strict grid plan, and whether it contained public buildings other than the temple, are questions we hope to examine in future seasons. 

In the upper levels of trench F6 scattered debris was found corresponding to levels ascribed by the earlier excavators to the "first temple area." Pieces of bronze jewelry and votive pottery were found within this debris, including a black-figured skyphos adorned with a variety of animals and with potter's and painter's signatures (Pl. 4, b) and a fragment of about one-half of a capital from the earliest Archaic temple which joined with its other half, which had been excavated by the Goldman expedition and left at the site. 

Undisturbed levels of the Classical period (480-323 B.C.) have not yet been distinguished by our expedition anywhere on the acropolis at Halai. Although we know from tombs and other evidence that the town was occupied in this period, it is possible that the acropolis itself was devoted to official activities. Or possibly the remains of Classical habitation on the acropolis were almost completely obliterated by later buildings. It is hoped that future seasons will produce more evidence on this point. 

The Hellenistic Period

A major effort is being made to excavate a substantial area of the ancient town untouched by the earlier excavators. We have chosen to concentrate at a point near the center of the site within our Areas C and H. These Areas straddle the main ancient NW-SE street which bisected the site. Four initial 4x4 m. trenches in each Area will, when the baulks are completely removed, form larger squares 9 m. on a side and the grid pattern will eventually be extended to excavate the street itself (Fig. 5; Pl. 4, d). 

The upper levels of most of the structures so far cleared in Areas C and H belong to the Roman period. Lower levels, on the other hand, can be dated by pottery and lamps to the Hellenistic era. Little can yet be said about the Hellenistic structures, since they have not yet been extensively investigated. In general, however, they appear to be of the same character and to have the same orientation as those of the Archaic period (see above) and the Roman period (see below). In some instances the lower parts of Hellenistic walls probably served as foundations for Roman buildings. The evidence thus suggests that the Hellenistic town was essentially similar to the Roman one. The general alignment in both periods continues that of the Archaic period. 

Beneath fallen roof tiles (Pl. 5, c) at the SE side of Room 3 in Area H (trenches H1 and H5, Pl. 5, a) a deposit of pottery and objects was found which may be dated to the late 2nd or early 1st century B.C. The deposit includes terracotta lamps (Pl. 6, a, top row), a side-spouted vessel (Pl. 6, b) and many terracotta loom weights. A fall of similar tiles in Area C may be about the same date. Perhaps the collapses of roofs that these situations attest may be connected with the destruction of the town by Sulla in 85 B.C. mentioned by Plutarch (Sulla, 26), although this can only be a very tentative hypothesis until further excavation has taken place. 

The Roman Period

The existence of the main NW-SE street is attested in trench C2, where the street was bounded by a line of stones probably representing a curb (Pl. 5, d). The cross street is best attested by similar lines of stones passing through trenches C3 and C4 (Pl. 4, d) and through trench H5 and the area of the former baulk between H5 and H6 (Pl. 5, a, indicated by arrow; a later wall is built over part of this curb in H5). That this cross street had at least partly gone out of use before the end of occupation in Area H is attested by the existence of the corner of a building in trench H6, at a higher level than and blocking the cross street (Pl. 5, a, lower left). 

The houses of the Roman period were divided into several rooms and are built from a great variety of material, including many ashlar blocks reused from earlier structures. The finds, although scanty, attest to their domestic character. A basin, mortar and grindstone were found in situ in a good state of preservation in trench C3 ext. (Pl. 4, d, during excavation; Pl. 4, c after cleaning). 

Pottery and lamps (e.g., Pl. 6, a, bottom row) associated with the houses may be dated to all three Roman stages, Early (31 B.C. to 150 A.D.), Middle (150-350) and Late (350-650), although remains of Middle Roman are as yet scanty. The latest datable material firmly associated with the Roman houses is probably of the 6th century, after which the town may have been abandoned. 

Although no excavation has of yet taken place in the larger "square" church in Area G, it is likely to have been in existence during the time of the latest Roman occupation of the town. How long it continued in use is not clear---perhaps it was abandoned at the same time as the town---, but it was eventually replaced by a smaller church built over its ruins, probably in the 12th or 13th century. Two tile covered graves excavated in the upper levels of trenches A4 and H5, and others like them cleared by the earlier expedition, cannot be closely dated, but were probably dug after the abandonment of the town. 

APPENDIX 1: NEOLITHIC CHIPPED STONE, by Kerill O'Neill and Melanie Pomeroy

During the first two seasons of excavation, we have collected 891 pieces, of which 312 were found in the course of digging, and 579 were separated from the heavy fraction left after the excavated soil had been passed through a flotation system. Thus 65% of the Halai chipped stone assemblage was recovered only by means of water-sieving the excavated soil. All the excavated soil from Area F, and control units from the other areas, were water-sieved. This process has made possible the recovery of tiny pieces of chipped stone, as well as flake and blade fragments and, occasionally, complete artifacts. 

94% of the pieces are obsidian. Chert makes up the remaining 6%. While none of the obsidian has as yet been subjected to characterization studies, all the pieces we have examined so far conform with the visual characteristics of Melian obsidian. The chert pieces vary widely in color, texture and quality. We have yet to identify any local sources for workable chert. 92 pieces, 10.3% of the collection, are blades or tools. The presence of two cores (Fig. 6, A), and four crested blades (Fig. 6, B), gives evidence that nodules were worked at this site. There are also three platform removals, and a number of cortical flakes. 

The great majority of diagnostic chipped stone pieces examined so far appear to be of Late Neolithic character. This preponderance of Late Neolithic artifacts would appear to tally with the preliminary analyses of the pottery. There is only one percussion blade (Fig. 6, C). The rest of the blades are pressure-flaked and parallel-sided. Proximal dorsal trim is very common (Fig. 6, D, E). There are many retouched pieces, i.e. blades and flakes modified by discontinuous or continuous retouch, including those modified by use-retouch, or use-wear, only (Figs. 6, F, G, H; 7, A, B, C). One artefact appears to be a pointed piece; both ends come to points, and one in particular appears to have traces of crushing and rounding from use as a drill or piercing tool (Fig. 7, D). Only one piece which resembles a sickle element has been identified (Fig. 7, E). No microscopic analysis, however, has yet been performed on this or any other piece to detect traces of silica gloss. The dearth of sickle elements is perhaps due to the location of the excavations within the Neolithic settlement and not out in the fields where these tools were used. There are two notched pieces in the assemblage, one made from a blade (Fig. 7, F), and the other from a flake (Fig. 7, G). 

 By far the most highly worked tools in the assemblage are the projectile points. All four are barbed and tanged (although in two cases the tangs have been snapped off). One of the points is short and squat with sharply defined barbs (Fig. 7, H). The other three are longer and slimmer with modest barbs (Fig. 7, I, J, K). In one case, only half the point is preserved. It seems likely that the short and squat projectile point was fashioned either at a different period from the rest, or for a different purpose. 

 Detailed analysis of the Halai chipped stone assemblage is planned. Furthermore, excavation of the earlier levels, and the establishment of a chronological sequence should reveal more about the development of the lithic industry during the Neolithic period. Studies to date have shown that nodules were worked at Halai, and that the small sample examined so far, which includes numerous tiny flakes, has a wide variety of types. Five different tool types are represented: retouched pieces, pointed pieces, sickle elements, notched pieces, and projectile points (and the classification "retouched pieces" is very broad and covers pieces with a variety of forms and functions). It is too soon to characterize the lithic industry at Halai but the recovery of even the smallest chips should, after further study, make possible a detailed picture of chipped stone tool production. 

APPENDIX 2: HELLENISTIC AND ROMAN POTTERY AND SMALL FINDS, by Karen E. Carr

Hellenistic

 1. Stamped loomweight (H91-491). D. avg. 0.11, Th. 0.03. Coarse local fabric. Buff clay with large red, black and white inclusions. Flat disk with two suspension holes and dolphin stamp (Fig. 8; L. 0.054). Goldman's Type IIB (Goldman 1940, pp. 509-513), examples of which were found in contexts of "the end of the 3rd to the beginning of the 2nd century B.C." (p. 509); ours is probably later. Goldman also mentions a possible example of a dolphin stamp from Halai (p. 513, no. 39), but without illustration. 

 2. Lamp (H91-473). Pl. 6, a, B. Missing only vertical handle. L. 0.109, D. 0.06. Gray ware with few inclusions. Thin, metallic black slip. Wheelmade lamp with flat plain bottom; globular body curves to plain disk with raised collar added separately. Central filling hole. Close to O. Broneer, Corinth, IV, ii, Terracotta Lamps, Cambridge, Mass. 1930 (hereafter Broneer), Type XIX, no. 329, pp. 66-70, pl. 6 (ca. 100 B.C. - 50 A.D.), although lacking the decoration on the rim and with a rounded rather than triangular nozzle; R. H. Howland, Agora IV, Greek Lamps and Their Survivals, Princeton, 1958 (hereafter Howland), Type 37C, no. 504, p. 121, pls. 19, 45 (dated ca. 50 B.C.- 50 A.D.), is also similar, except for the ware. 

 3, 4. Lamps. Pl. 6, a, A and C. Similar to preceding. 

 5. Lamp (H91-469). Pl. 6, a, D. Missing only vertical handle. L. 0.082, D. 0.039. Badly encrusted with lime, which prevents detection of glaze. Wheelmade lamp with irregular, slightly rounded bottom; sides curve to a flat sunken top. Like Howland, Type 35D, no. 484, p. 114, pls. 17, 43 (late 1st. c. B.C. into the 1st. c. A.D.), but the type begins as early as the 3rd c. B.C.; cf. Broneer, Type XVI, no. 198, p. 56, pl. 5. 

 6. Spouted jug (EU H5b(15)92). Pl. 6, b. Whole except for the broken spout. H.0.15, D. of rim 0.12, Th. 0.003. Buff clay with gray-black slip. Badly encrusted with lime. Globular body, raised ring base, flaring neck with molded rim. Single rounded handle curves from below rim to upper body. Strainer across bottom of neck, with five holes in quincunx pattern. This shape is found commonly from the 2nd. c. B.C. onward: see, for example, G.R. Edwards, Corinth VII, iii, Corinthian Hellenistic Pottery, Princeton, N.J., 1975, no. 778, p. 149, pls. 36, 64. 

Early Roman

 7. Lamp (H91- 436). Pl. 6, a, F. Area H, topsoil. Half of discus only. D. of discus 0.074. Discus shows a pattern of tongues forming a rosette. Like Broneer Type XXI, no. 393, p. 73-76, pl. 7 (late 1st. c. B.C. to late 1st c. A.D.). 

 8. Bowl rim (EU H1b(40)157). Fig. 8. Floor of Area H, Room 3. H. pres. 0.04, D.est. 0.16. Dark orange-red clay with good quality, darker, browner glaze. Wide everted level rim. No decoration. Possibly imitation Samian ware. Resembles H.S. Robinson, Agora V, Pottery of the Roman Period: Chronology, Princeton, 1959 (hereafter Robinson), no. G 21, p. 25, pls. 4, 57, 62, 73. Robinson dates this form to the first half of the 1st c. A.D. 

 9. Cup base (EU H1b(40)157). Fig. 8. Floor of Area H, Room 3. Base only. H. pres. 0.016, D. of base 0.045. Dark pinkish clay with darker orange-red glaze. Hemispherical cup. Similar to Robinson, no. G 73, p. 29, pls. 5, 66. Date as preceding. 

Middle Roman

 10. Lamp (H90-11). Pl. 6, a, H. Fill over Area H, Room 3. Only the handle preserved. H. pres. 0.041. Buff clay, unglazed. Like Broneer, Type XXVII, no. 556, pp. 90-102, pl. 11, of Corinthian manufacture. The type starts in the late 1st c. A.D. and reaches its height in the Hadrianic period. 

 11. Amphora handle (EU C1a(7)91). Not illustrated. Floor (?) of Area C, Room 1. Coarse brick-red clay with white lime inclusions. Eroded, perhaps by later use as a tool. Heavy, thick, straight handle, ovoid in section, with shallow vertical fluting. Similar to Robinson, no. K 113, p. 69, pl. 15. The Athens amphora comes from a sealed deposit of the first half of the third century. Although this shape is more common in the fourth century, and continues into the early fifth; the third century date seems more likely for our piece. None of the later examples from the Agora have fluting on the handles and no definite fourth century material has yet been discovered at Halai. 

 12. Glass beaker (H90-207). Fig. 8. Floor (?) of Area C, Room 1. Base only preserved. H.pres. 0.009, D. of base est. 0.044. Milky semi-opaque pale blue. Heavy, disk-like base, with body widening upward. Similar to Robinson, no. M 105, p. 94, pl. 55 (second half of 2nd c. A.D.). 

 13. Glass beaker (H90-181c). Fig. 8. Fill of Area H, Room 3. Fragment of rim. H. pres.0.017, D. of rim est. 0.08. Transparent, pale bluish-green. Rim rounded off, with thickening on inside. Similar to J. Hayes, Roman and PreRoman Glass in the Royal Ontario Museum: A Catalogue, Toronto, 1975 (hereafter Hayes 1975), no. 180, p. 64, fig. 5 (2nd to early third c. A.D.). 

Late Roman

 14. Lamp (H90-338). Pl. 6, b, G, and Fig. 8. Fill over Area H, Room 3. Half of the top is preserved lengthwise; handle broken off. L. pres. 0.085, W. pres. 0.035. D. of discus 0.044. Reddish-brown clay with no visible inclusions. Ovoid, tapering slightly toward the round filling hole. Circular discus holds cross with slightly flaring bars. Small double circles around and inside the cross. Rim: double raised palm branch pattern continuing around past nozzle. Base of vertical handle. Close to J. Perlzweig, Agora VII, Lamps of the Roman Period, Princeton, 1961, no. 2544, p. 182, pl. 40 (6th c. A.D.). Also like Broneer, nos. 1456 and 1489, pp. 118-119, pl. 22 (his Type XXXI, which begins in the late fourth century A.D. and does not outlast the sixth at Corinth). 

 15. Lamp (H90-70). Pl. 6, a, E. Fill over Area H, Room 3. In two pieces; missing front part and nozzle. L. pres. 0.096, W. 0.076. D. of discus 0.062. Bright red clay with no visible inclusions. Large unpierced vertical handle. Like Broneer, no. 1482, pp. 118-119, pl. 22 (his Type XXXI), but with a single filling hole at rear of discus. Row of concentric circles on the rim; an indistinct design on the discus. Raised circle with many smaller circles inside on the base. Date as above. 

 16. Dish (EU H6 b(14)96). Fig. 8. Lower street level outside Area C, Room 5. Part of base preserved. D. of base est. 0.095, Th. 0.008. African Red Slip. Orange-red clay with no visible inclusions; glossy though thin light orange slip. Large dish/bowl with sloping floor. Foot of medium height, flaring at top. J. Hayes, Late Roman Pottery, London, 1972 (hereafter Hayes 1972), Form 104, pp. 160-166 (530-580). Two grooves around floor. Stamped decoration in center shows male frontal bust (a saint?) similar to Hayes 1972, stamp type 234 (around second quarter of sixth century) but lacking body below shoulders, and with no nose or mouth. The drapery is closer to Hayes 1972, type 245. The stamp is off-center and must have been accompanied by another bust on the other side. Over the top of the head is part of another stamp, possibly Hayes 1972, type 78. 

 17. Dish or bowl (EU H5 b(8)43). Fig. 8. Lower portion of fill over Area C, Room 3. D. est. 0.28. Fragment of rim. Dark reddish clay with slightly darker, very thin slip. Late Roman C ware. Vertical rim incorporates a flange. Parallels: Hayes 1972, Form 3F, no. 25 (Athens Agora P 26088), pp. 333-338 (ca. 525-550). 

18. Amphora (EU H4 a(2)2). Fig. 8. Area H, topsoil. Only handle and part of body preserved. H. pres. 0.19, D. of handle 0.037. Gritty buff clay. Upper body plain; below the handle is closely spaced ridging. The thick, sagging handle is ovoid in section. These ridged amphorae are common at Halai in all 6th. c. contexts. Like J. Riley, "Coarse Pottery", in Supplements to Libya Antiqua V, vol. II: Excavations at Sidi Khrebish, Benghazi (Berenice), Tripoli, n.d., no. 347, Late Roman Amphora Type 1a, pp. 212-216, fig. 91. Riley dates this type between the 6th and the mid-7th centuries A.D. It has a wide distribution and is common at Athens, possibly originating in the Antioch region. 

19. Glass goblet base (H90-218b). Fig. 8. Area H, topsoil. Only base preserved. D. of base 0.034. Blown glass. Thin, transparent greenish-colorless; some small bubbles. Disk-likebase, rising steeply toward center, with hollow edge; bottom pushed up into stem. Many examples of this type have been found at Halai. Others have the thin stem preserved. Closely resembles Hayes 1975, no. 653, p. 158, fig. 11, and Robinson, no. M 342, p. 116, pl. 54 (early 6th c.). Hayes sees the narrow stem as suggesting a date in the 6th or early 7th centuries A.D. No "wine-glasses" occur at any site before the late fifth century. 

20. Copper alloy cross (H90-558). Fig. 8. Fill over street in Area H. H. 0.028, W. 0.018. Considerable corrosion. Bars flare slightly toward the ends, except the top bar which tapers to a pierced holefor hanging, perpendicular to the cross itself. The cross is decorated with three raised double circles on one side, each with a tiny inner circle. Reverse plain. 

21. Copper alloy spatula (H90-311). Fig. 8. Fill over street in Area H. L. 0.195, L. of blade 0.074, W. of blade 0.012. Relatively little corrosion. Blade U-shaped in cross-section with pointed tip. Handle is circular in section with a groove running down its length. The handle ends in an ovoid nodule. 

None of the finds post-date the sixth or very early seventh centuries A.D., except a few isolated sherds of medieval lead-glazed pottery, probably to be associated with the construction of the smaller church inside the ruin of the sixth century basilica. Therefore the smaller church was probably not associated with a resettlement of the town of Halai. Further detailed study of the ceramics is underway to compare the trenches and areas and to relate the site as a whole to other excavated sites in Greece. This study will compare the ratio of fine wares to coarse wares in each context, the proportion of imported pottery and amphorae in each period, etc., and will, I hope, lead eventually to a much-needed chronology of the local coarse wares. Preliminary results suggest a higher proportion of coarse wares in Area H than in Area C, but until further excavation and stratigraphic analysis take place this must be regarded as only a tentative hypothesis. 

APPENDIX 3: BALLOON PHOTOGRAPHY by Andrew Heafitz

Balloon photography was used at Halai during both the survey of the site in 1989 and the digging seasons of 1990 and 1991. The original system used during the 1989 survey season consisted of a 35 mm camera lifted into the air by a spherical tethered balloon three meters in diameter. The camera was suspended from the rigging underneath the balloon is such a way that it always pointed straight down. The balloon was moved over its target area by tethers and pictures were taken by remote control. Although the photographs obtained with this system were satisfactory, the camera operator did not have much control over what was included in each photograph. Also, there was always doubt that the camera had actually taken a photograph once it was too high for its shutter and motor drive to be heard from the ground. 

In order to overcome these shortcomings, a new platform was constructed for the 1990 digging season. The same spherical balloon was used. Although the shape is not as aerodynamic as a blimp-shaped balloon, it is more compact, takes less gas to inflate, and is easier to handle when inflated but still on the ground. The small size also makes it possible to inflate the balloon on the deck of a ship and fly it from a small motor or row boat. The whole aerial photography system, including the balloon, can be carried on an airplane and transported to the site in a suitcase. 

 The new camera platform used in 1990 and 1991 (Pl. 6, c) represents a considerable advance over the initial system. The camera, instead of simply hanging from the balloon's rigging, is mounted inside a stabilized, two axis, powered gimbal mount constructed from nomex honeycomb-fiberglass. The outer section of the mount isolates the gimbals from the motion of the balloon. The inner gimbals are powered in the elevation and azimuth axes by two small servo motors, which are operated from the ground by radio control; hence the operator can aim the camera in any direction from perpendicular to parallel with the horizon and can rotate it a full 360 degrees in the horizontal plane. A video camera is mounted in tandem with the still camera and its signal, in addition to being recorded on videotape on board, is transmitted to the ground by a video downlink. The video signal can thus be displayed on a three inch color monitor in real time so that the operator can see exactly the same view as the still camera. Consequently, the cameras can be aimed with great precision. 

The camera control panel on the ground has two joysticks, one to aim the cameras, the other to activate the still camera's shutter, resulting in still pictures which exactly match the scene displayed on the monitor. The exposure of the picture is determined by an automatic exposure mode on the camera and the lens is focussed at infinity. Since great depths of field are not generally necessary, fast shutter speeds can be used in order to minimize the effects of any motion imparted to the camera from the motion of the balloon or its tether lines. The video downlink is invaluable; not only does it provide for aiming the cameras and indicate when they are perfectly still, but its audio track also confirms that a photograph has actually been taken by the still camera. Both the camera shutter and the sound of the camera's motor drive reaching the end of the roll of film can be heard, so that problems can quickly be identified, with a resultant savings of time and film. 

The video tape also provides an instant record of each flight. It can be reviewed either on the control panel monitor immediately after the balloon is hauled down or at any later time and it shows exactly what was in the camera's field of view when each picture was taken. Film, by contrast, requires processing and cannot be viewed without some delay.

 A prime mission for balloon photography at Halai is photographic documentation of the progress of the excavations at a finer level that is possible by means of architectural plans. During the 1991 season, the balloon was flown an average of 4 out of 5 days every week. Each trench was regularly covered from altitudes ranging from 4 m. (e.g., Pl. 4, a) to 40 m. (e.g., Pl. 4, d). Other photographs taken from altitudes of 100 meters or more cover much or all of the acropolis (e.g., Pl. 1). This versatility makes it possible to maintain a photographic record which can show features ranging in size from 1 millimeter to over 100 meters across. The video monitor also makes it easier to create complex photo mosaics covering more ground than can be captured in one picture. Balloon photography is also helping to document features in the shallow water adjacent to the acropolis. The series of aerial photographs for each trench may also be used in conjunction with regular architectural plans to create phase drawings., i.e. plan views over time. When drawn on transparent mylar, such drawings can overlie one another to show a sequence of features and artifacts over time, whether in order of their discovery or of their actual use in ancient times. Aerial photographs can also be developed and printed overnight and can be distributed the next day for use as an up-to-date reference in documenting new finds. 

 In future seasons we plan to substitute a medium format camera for the 35 millimeter camera for better photographic resolution. We also hope to put into use a system of photogrammetry, whereby plans can be generated directly from aerial photographs. It may also be possible to link such plans electronically with the computerized database of finds so that architectural features and finds can be represented graphically in three dimensions, either on a computer monitor or as drawn out by a plotter. 

Last modified 4.5.99