Mohammed Bhatti and Daniel P. Diffendale
Team Pink, Session 1
Nezi Field Excavations Final Report
N 1006 – 1018.00; E 261.01 – 271.28
1–25 April 2013

This is a final summary of the first season of excavation (1-25 April) in 2013 in the northeast area of Nezi field. Guy Sanders (director) and Rossana Valente (field director) supervised.The Pink excavation team consisted of Mohammed Bhatti and Daniel P. Diffendale (recorders), Panos and Tasos Kakouros (pickmen), Vassilis Kollias (shovelman), and Tasos Tsogas (barrowman).The area is bounded to the north by the scarp of H. Robinson’s 1960s excavations, to the east by the excavation scarp immediately east of Wall 866, to the south by Cuts C193 and C848, more or less, and to the west by the varying eastward extent of the Blue team’s area of operation.
The area of excavation was effectively bisected along an east-west axis comprising Wall 945, Wall 807, and the robbing trench Cut 995. Of the contexts excavated this session, the areas north and south of this axis were physically related only by Deposit 978 (to the south) overlying Deposit 994, which filled Cut 995, which in turn cut Deposit 1003 (to the north). To the north, Wall 918, Wall 1033, and, until its removal, Wall 992, all acted as another east-west axis separating the deposits to north and south. For the purposes of this report, the contexts south of Wall 945-Wall 807-Cut 995 will be referred to as comprising the South Area; contexts north of Walls 918-992-1033 as the North Area, and contexts between the two axes as the Central Area.
The east-west walls are the most prominent positive features in the area, and are constant points of reference; less easy to understand are the mostly negative traces of north-south features, for instance the fill of 1050, laid above an irregular N-S cut (no context number), possibly related to the NE-SW cut 1082; or the robbing of a N-S wall represented by the fill of 1009.

Precontextual material, ca. 9th c. BCE – 3rd c. CE
No contexts excavated this session can be dated before the Middle to Late Roman period. However, earlier human activity in the area is suggestively documented by the contents of later fills. Scattered Geometric sherds (in contexts 978, 1056, 1076, 1084, 1085) suggest the presence of Iron Age burials somewhere below. A pebble pavement clearly must have existed nearby in the Classical or Hellenistic period, later broken up, fragments of which made their way into many of the later fills (998, 1003, 1009, 1015, 1050, 1076, 1081, 1083, 1084), but which were most thickly concentrated in 1085. Of Roman date must be the numerous pieces of marble revetment of many colors and sizes, fragments of which also made their way into the majority of fills excavated in this area during this session. This marble revetment, coupled with octagonal marble or schisty limestone pavers (1009 and 1085) and chunks of preparation “brown layer” mortar for a hydraulic installation (1085), suggests that there may have been a bath facility in the neighborhood. If there were a water source capable of supplying such a bath, it perhaps could also have supplied a pre-Roman bath that was paved with pebbles. Other finds indicative of earlier Roman buildings include mostly white marble tesserae and small chunks of painted wall plaster, both found in numerous deposits.


Late Roman, ca. 4th / 5th c. CE
Possibly the earliest context known so far is deposit 1000, in the South Area, with ceramic materials no later than the 3rd/4th c. CE. It is probable, however, that this is simply earlier material redeposited, perhaps in the 5th, 6th, or even 7th-8th centuries; only further excavation can say for certain.
The earliest visible structures existing are probably the parallel E-W walls 918 (to the north, with the in situ column base) and 945 (to the south), although they cannot yet be directly dated. At some time during or after the 4th c. CE, in the Central Area, a fill consisting of destruction debris (tiles, carbon, nails, slag) was dumped between them (Deposit 1084, date: ceramics); this debris was laid in part against the north face of Wall 945. Beneath 1084, but thus far unexcavated, what may be a real destruction layer is waiting to be understood; perhaps the material deposited in 1084 came from that destruction. 1084 may have been laid as a leveling fill for a floor or working surface above (1083), which was probably put into use in the 5th c. (pottery). Someone dropped an old coin of Constantius II or Julian II (coin 2013-100: 355-361 CE) onto this surface, and there was also a notable amount of glass. Associated with the surface of 1083, and dated by it stratigraphically, was hearth 1065, paved with flat slabs of non-local schist and ringed with tile. This was a substantial construction measuring some 0.90 x 0.80 m, and set against the then still-standing upper courses of Wall 945.
The surface of 1083 was analogous in composition, elevation, and date (5th c. pottery) to that of Floor 1060/1061 immediately to the north of Wall 918 in the North Area, and they likely saw contemporaneous use; deposited in both were sherds of Niederbieber 77 amphorae as well as of Eastern Aegean micaceous cooking ware—though these are not uncommon finds for the area. Floor 1060 was laid against the ‘stylobate’ of Wall 918; this part of the wall was probably still standing when 1060 was in use. The people using Floor 1060 saw the need to burn something, perhaps for cooking, and hence they constructed the small Hearth 1059 (5th c. stratigraphically) within it, close to the presumably still-standing line of Wall 918, partially demarcating the hearth with tiles (ca. 0.3 m diameter).
However, Hearth 1059 may not have been in use for long; the packed surface of deposit 1056 was laid directly above Floor 1060, probably also in the 5th c. (pottery), and this sealed over the hearth. Three coins were dropped onto this surface, including an old one of Constantius II (coin 2013-72: 346-361 CE; coins 2013-68 and 76 illegible). At some later point, perhaps already in the 5th c., a pit was cut into the surface of 1056 (no context number) and then filled (1048, 5th c. stratigraphically). However, the nature of this pit can hardly be comprehended, given its likely severe truncation by the Robinson excavations. Above the pit fill and the surface of 1056 was constructed the earthen Floor 1044/1045, probably still against the not-yet-robbed-out Wall 918. A clear pale green glass tubular ring goblet foot with knop stem deposited in the fill of Floor 1044/1045 dates to the mid 5th c. or later, while a coin of Leo I (coin 2013-59: 457-474 CE) was dropped on the floor surface.
Whatever happened in the North Area after the 5th c. will remain a mystery, since the overlying strata have been removed by the Robinson excavations—unknown unknowns. As for known unknowns, the nature of deposit 1050 is something of a mystery, especially its relationship with Wall 992.


Late Roman, ca. 5th/6th c. CE
A meter or two to the southeast, in the Central Area, and still in the 5th or maybe the 6th c., a large amount of mixed fill was deposited between Walls 1033, 945, and 1007; this was deposit 1085 (dated by pottery). Within the deposit there seem to have been diverse sources of fill; in its southwest extent the sediment was blackish and especially dense with fragments of marble revetment and over fifty large cobble- to small boulder-size fragments of pebble-mortar pavement. Further east, it included several cobble-sized pieces of iron slag, perhaps deposited after use as road-metal nearby. Significant numbers of fragments of glass slag in the deposit suggest a workshop in the vicinity. A notable inclusion was a fragment of a Hellenistic or Roman terracotta figurine of Baubo (MF-2013-8). At the time that 1085 was deposited, there was probably a N-S wall to its south, against which it was deposited; east of the wall was a section of pavement. This wall and pavement would later be robbed out (context 1009), sometime between the end of the 5th and the early 7th cs. (see further below). It seems that this wall would have made a narrow (ca. 1 m wide) N-S passageway with Wall 1007 to its east, and the passageway was paved. Maybe there was a staircase here, which started going up to the south between walls 1009 and 1007, turned east between the south end of wall 1007 and the wall robbed out by 995, and returned north between walls 1007 and 866. This is only speculation.
Sometime thereafter an irregular NNE-SSW trench, 1082, was cut through the Central Area between Wall 945 to the south, and probably Wall 918 to the north, though its northern extent was later truncated by Cut 1032. The trench of 1082 was cut through deposits dating to the 5th/6th cs. (1083, 1085, maybe 1076) and filled by material dating to the 5th/6th cs. (1081), indicating that it, too, should be dated within those centuries. The narrow width and odd orientation of the cut make it unlikely to have been for robbing out a wall; perhaps it was cut for the purpose of robbing out a drain. Whoever did the robbing left a single piece of whatever they were robbing behind, a squared stone block ca. 0.20 x 0.35 m, at the bottom of Cut 1082, and which shares its orientation. If this stone did form part of a drain, there is no indication of how it worked—no mortar, cement, or the like; maybe the robbers were after lead pipes? The line of cut 1082 is picked up north of Wall 918 by the irregular cut at the bottom of context 1050. To the south, there are tiles laid in Wall 945 where Cut 1082 meets it, although otherwise the wall seems to be constructed solely of stone. Perhaps these tiles were laid to replace the putative robbed-out drain. Cut 1082 was filled with debris (marble, schist, plaster, pebble pavement) sometime in or after the 5th/6th cs. (context 1081, no precise date: pottery), probably soon after the robbing of the drain.
After the filling of cut 1082, just to its east and partially overlaying it, and above 1085, the destruction-debris fill of 1076 was deposited, not before the 2nd half of the 5th c. (dated by a clear greenish glass tubular goblet foot). This fill must also have lain against the wall/pavement of 1009, but maybe not for long, as the latest material in the fill of 1009 also dated to the late 5th c. The robbing activity represented by 1009 could have happened at any time until the early 7th c. deposition of 1003 above it, however.
The relationship between the robbing trenches of 1009 and 995 makes for a vexing question. It seems likely that 995 was used to rob out the upper courses of the eastward extension of Wall 945, although this cannot yet be answered definitively. Wall 945 should predate all contexts known thus far. The contents of 994, the fill of trench 995, suggest that the latter was cut in the late 7th or early 8th c. However, this trench apparently bisects the N-S line of the robbing trench of 1009: the southward continuation of 1009 seems to be visible in the south scarp of 995; 1009 seemingly dates sometime between the construction of wall 945 and its robbing out by 995. Perhaps two walls intersected, and the robbers of the N-S 1009 left the E-W 945 alone? It is to be hoped that future excavation south of Cut 995 will clarify this vexing nexus.
In the South Area, the latest datable material in the dump fills of 996 and 998 date to the mid 6th c., but they could well have been redeposited as part of later leveling activity.

Late Roman, 6th /7th c. CE
The late 6th to early 7th centuries CE witnessed much human activity in the way of movement of fill and construction of floors. The upper courses of Walls 918 and 945, which now survive only as foundations or robbing trenches, were probably still standing and in use during this period. Not much can be said about this period in the South Area, however, without further excavation.
In the Central Area, fills 1067 and 1062 were deposited as leveling fills during the late 6th century (dated by pottery and stratigraphy, respectively), on which the floor surface of 1036/1038 was laid shortly thereafter, traced for an area of ca. 1.5 x 3 m. Still within the ambit of the late 6th century (dated stratigraphically), Pit 1032 was cut through the surface of 1036/1038 and then filled in with old rubbish (deposit 1029), including marble, plaster, part of an andesite millstone, glass, and earlier ceramics. Soon after the pit was filled, Floor 961/1022 was laid over it in an area roughly 3 x 3 m square, extending further east than the traced limits of 1036/1038. The builders of Floor 961/1022 were presumably also responsible for constructing Hearth 960/1021/1025, at the southern extent of the floor, against the still-standing upper courses of Wall 945. They used the end of a single large tile, W. 0.46 m, Th. 0.03 m, broken into six fragments, to line the edge of their hearth; these tiles showed evidence of burning on their inner faces. At least one olive and one grape pit were charred in this hearth, and an Archaic or Classical miniature terracotta quadruped was deposited here as well, though it is archaeologically impossible to say whether this was intentional; the figurine was probably already broken when deposited. Unfortunately, the people who used Floor 961/1022 were careful with their money; only a single illegible minimus (coin 2013-46) ended up in this context.
The years after 600 CE were a time of destruction. The roof that sheltered Floor 961/1022 collapsed, possibly due to fire, resulting in deposit 1003. This is dated to the first half of the 7th century by pottery, which included stewpots, Palestinian, Gaza, and Late Roman Amphorae 1 and 2.The entire context has been saved as Lot 2013-001. This context is the eastward extension of deposit 929 excavated in 2012, but ceased along an arbitrary N-S line; 929 was saved as Lot 2012-47. Other things in the deposit included vessel and window glass and iron nails, but only a single piece of marble revetment, in stark contrast to the dump fills in the area that include dozens of marble fragments.

Late Roman, 7th / 8th c. CE
After the building comprising the Central Area lost its roof, it also lost its north and south walls. The robbing trench 1027 was cut to remove the upper courses of Wall 918, and then filled in (deposit 1015; dated stratigraphically to the early 7th c. or later). Who could have done such a thing? Two or possibly three Vandal coins were found in the fill (coins 2013-42, 43, 44), but admittedly they would have had to be rather late Vandals. To the south, the eastward extension of Wall 945 was robbed out resulting in Cut 995; the cut was topped up with fill (context 994) of the late 7th / early 8th c. (dated by pottery). Deposit 994 included earlier material, like so many deposits, but some terracotta sima fragments and a sherd of stamped Samian ware (C-2013-9) were unusual.
It will be crucial for future excavation to investigate the narrow dark deposit exposed by context 1050 along the N edge of Wall 918/1033, to see whether this is a foundation trench for that wall, a robbing trench for that wall, or something else entirely. Unfortunately, the brief pedestaling of Wall 992 and Cut 1027 have made an already complex area almost impossible to make sense of, in particular the precise stratigraphic relationship between the North Area and the Central Area.
No material from the construction of Wall 992, or from contexts excavated directly beneath it, need postdate the 5th c.; however, the narrow strip of sediment excavated as Deposit 1035 should be the same fill as that excavated immediately to the N as 990 and to the S as 991; 991 should date to the late 6th / early 7th c. (pottery), hence, Wall 992 should date to this time or later.
It is in the late 7th /early 8th century that we can begin to say something about the South Area, but it is not a very interesting something. A large amount of fill was dumped (contexts 978, 970, 979, 976) in the area south of the line of Wall 945, none of it containing material datable later than the 5th c. CE, but stratigraphically later than the late 7th/early 8th c. fill of 995. Some interesting earlier material was included in these fills, such as the Roman votive terracotta theater mask (MF-2013-1), part of a terracotta lion’s head spout, and antefix fragment, all from 978. A fragment of the ubiquitous pebble pavement is visible in the soil exposed by 978. At some point during the exposed life of 978, someone burned something on it and left behind the ash (deposit 979) as well as an illegible coin (2013-4).

Byzantine, 8th – 12th c.
Still in the South Area, deposits 964, 965, 968, and 973 were laid down in the Early Byzantine period, as attested by the limited presence of Early Byzantine pottery. These appear to have been fills, although it is possible they served as surfaces. In the Middle Byzantine period, Wall 807 was built on the foundations of the robbed out Wall 945, as attested by the fill of the foundation trench 1008 (dated by pottery to Middle Byzantine). In the Late Byzantine period, during the early 12th c., some wall robbing was carried out in the area of Wall 918, as attested by context 1014. This may have been connected with the construction of Wall 540 on a N-S line across the area, also dated to the early 12th c. stratigraphically.

Modern
In the late 19th c., a pit was dug for an orange tree (context 791, dug in 2012); this cut through the western end of the contexts excavated this session in the North Area. In the 1960s, Henry S. Robinson conducted excavations in the Nezi field; these excavations cut through the northern extent of the contexts excavated this session (2013-I) in the North Area, making necessary some cleaning of the area (cleaning context 1042).
Suggestions for Future Excavations
There are two possible future alternatives. The first would be to bring the Southern area down to the 5th-6th century levels already exposed in the Central area. Alternatively, and more immediately promising would be to excavate the Central area to the levels exposed to the West, such as context 1080. A potential destruction layer was exposed in the NW corner of context 1084, which may prove to be as fruitful as Context 1080. Moreover, open questions remaining that could be cleared by future excavations include: the potential relationship between Cut 1082 and irregular NS scarp exposed by Context 1085; the relationship of the cut exposed by context 1009 with its southward continuation South of cut 995; the relationship of Walls 1007 and 866 with the fills to the west.