Kate Larson and Jon Meyer
ASCSA Corinth Excavations
Blue Team Session 3 Final Report
N :1016.50 N, E: 263.00 E, S: 1005.70 N, W: 255.50 E
May 27 – June 18, 2013

During Session 3 2013, the Blue Team excavated in three distinct areas in the northwest section of Nezi Field (N. 1016.50 N, E. 263.00 E, S. 1005.70 N, W. 255.50 E), with the goal of reaching Late Roman (4th-6th century) occupation layers, discovering discrete Late Roman contexts to clarify an absolute ceramic chronology, and learning more about the activity in and around Nezi at that point in time. Additionally, we have been able to refine the dating and use of the space immediately north of wall 366, around well 742, in the late 11th and 12th century. Under the directorship of Guy Sanders and field directorship of Rossana Valente, the Blue Team consisted of Athanasios Notis (pickman), Christos Sakellariou (shovelman), Panayiotis Rontzokos (barrowman), and Kate Larson and Jon Meyer (recorders).

On May 27th, the Blue team conducted a minor operation in the area immediately south and east of the Giambouranis (Turkish) House (N. 1016.50 N, E. 259.80 E, S. 1014.25 N, W. 257.85 E), clarifying N-S cut 1040 and removing the skin balk between 1040 and the large tile deposit 1080 excavated in Session 1 by Katherine Harrington and Jana Mokrišová. Beginning on May 28th, the Blue team moved south to the area bounded by E-W walls 1087 and 945 on the north (1010.50 N), robbing trench 1171 on the east (263.00 E), wall 366 on the south (1005.70 N), and the western scarp of Nezi on the west (255.50 E), last excavated during Session 3 2012 by Rossana Valente and Charles George. In this area, we have identified a Frankish pit and robbing activity, several late 11th-early 12th century Byzantine structures suggesting intensive activity, and a thick Late Roman fill, mostly of earlier architectural debris, which is probably related to the construction and elevation of well 742 in the 5th or early 6th century. By the end of the season, we reached early-middle Roman strata which indicate possible domestic occupation in Nezi but will require additional investigation in future seasons before any clear conclusions can be reached.

During the final week of excavation, Rossana Valente under the auspices of the Blue Team excavated a late 11th century coin hoard just south of wall 366 and east of wall 332, north of and under structure 452 (N. 1006.08 N, E. 264.20 E, S. 1003.70 N, W. 262.70 E), an area most recently excavated in 2009. While the original intention of this limited project had been to properly excavate a large, mostly intact jar which had become exposed due to weathering over the past several years, the discovery of the coin hoard in context 1235 offers excellent corroborative dating evidence for associated Late Byzantine pottery, and further excavation should be conducted in this area in subsequent seasons.

Early-Middle Roman (1st-3rd century)
Based on stratigraphy, N-S wall 1222 was built sometime before the 4th century, by which time it had gone out of use and was covered over (see below). The wall begins at 1010.03 N (where it abuts wall 1087) and extends south into the baulk at 1006.08 N. At the preserved elevation of 85.84 masl, the wall is constructed in two distinct manners: from uncut fieldstones of 0.15-0.20 m in its northern half, and (reused?) larger cut ashlars (0.35-0.55) at the south, possibly indicating a repair or extension of wall 1222. An earlier phase of E-W wall 366, running from 261.50 E to 262.90 E, where it disappears into the east balk, probably also was built in this period. The west end of this wall was robbed or otherwise disrupted in the Late Roman period, but the large ashlar stones still contain traces of red wall plaster in situ. A third Early Roman wall may have been located 3.60 m east of wall 1222, represented by Late Byzantine robbing trench 1171. The numerous ornate Masonry style and faux marble wall painting fragments found in the Late Roman fill in this area (contexts 1227, 1210, and 1187) could well have come from these walls, and suggest a rather wealthy domestic or semi-public space.

These walls are possibly related to the pebble floor found in situ near the north end of wall 1222, 0.50 m to the east, at 85.17 masl and/or the tile floor at 85.01 masl 2.5 m east of 1222 and a bit further south. Both these surfaces should date sometime in or before the 2nd century, based upon the material excavated directly above them (contexts 1202 and 1242, respectively), but further excavation is required to determine their actual date and extent. A third floor for which we found evidence but no in situ remains was made from cut down tiles which were used as bricks, set in a limey matrix, and covered over with soft limey plaster. This floor was destroyed by the 2nd or 3rd century and dumped in sizable chunks, the largest 0.60 x 0.70, into the area between wall 1087 and wall 945, and excavated as contexts 1202 and 1250; these contexts are dated by a large piece of square lipped water pipe dated by Palinkas and Herbst to the late 2nd-early 3rd century (Hesperia 2011 80.2), but no pottery is later than the 1st century CE.

Late Roman (5th-6th century)
Well 742 may have been constructed and in use as early as the fourth century, although further excavation around it and removal of the well structure itself will have to confirm this date. Beginning around 85.60 masl (excavated as context 1227), the walls of the well shaft were reinforced with uncut cobbles and fieldstones, built into and laid against the sides of the well shaft; the soil is completely undifferentiated from the surrounding fill. This rudimentary construction, which was clearly not intended to be seen, ceases after about half a meter, at 86.10 masl. The top of the preserved well shaft was made from three courses of larger rectangular stones roughly hewn on the exterior and more nicely cut into a curvilinear shape for the interior of the well shaft start (one course removed this season as context 1185). We have identified no external cut into which the well was dug; before it opens into a larger cistern or chamber, the well shaft is composed of the same hard packed soil we are finding throughout the excavation area which seems to have been packed or cut inside the well. Unlike at higher elevations, we have found no evidence below 85.60 masl or so of the reinforcing cobbles around the well structure which indicate built construction. The only soil we excavated below these cobbles, which seems possibly to have been cut rather than built as part of the well, contained no pottery dating later than the 1st century CE, although the deposit did contain a coin of Hadrian (coin 2013-251; context 1246).

We suggest that a series of superimposed dump fills represents a construction or raising of well 742 and the surrounding area, possibly to correspond to the street level of Late Roman Nezi. These fills cover the entire area bounded by later walls 1087 on the north, 366 on the south, robbing trench 1171 on the east, and the scarp of Nezi excavations on the west (N: 1010.08 N, E: 262.10 E, S: 1007.00 N, W: 255.50 E). The earliest, 1227 and its associated contexts (1239, 1240, and 1248, plus 1233 excavated by the Pink team this session), was laid down in the 4th century, possibly early 5th (as dated by the pottery from 1233). This filling operation was dense and over half a meter thick. Context 1210, immediately over 1227, was very similar in composition, differentiated only by the inclusion of small quantities of pottery dating to the 6th century. Similarly, context 1187, immediately overlying 1227, contained almost all 4th century and earlier material redeposited with two rims of 6th century cookware and a tubular foot goblet which dates to the second half of the 5th century or later. These fills were built up against and covered over the early-middle Roman walls 1222 and the earlier phase of 366, as discussed above.

Most of the cultural material from these fills is architectural in origin, including abundant tiles, bricks, painted and unpainted wall plaster, architectural moldings, and assorted floor tesserae. These fills likely also contained a large quantity of mudbricks which have since disintegrated, leaving behind small pieces of poorly preserved Archaic, Classical, Hellenistic, and Early Roman (1st century) pottery which we found in much greater quantities in these contexts than we did pottery which is contemporaneous with the deposit itself. Although truncated at the north by wall 1087 and the trenches to either side and the east by robbing trench 1171, this fill clearly extends south into the baulk under wall 366 and west into the baulk of Nezi.

Because there is so little material throughout these fills which is contemporary with the filling operation itself, it is unclear whether these three large deposits all represent a single construction event or the raising of the well gradually over time from the 4th-6th centuries. That these fills represent a single event is supported by the continuity in the construction method of the well, which was uncut fieldstones, cobbles, and soil (as discussed above) up to about 86.10 masl (exposed as contexts 1187 and 1210), and by the lack of evidence for floors or surfaces below surface 1157 at 86.46 masl.

A sequence of surfaces built above the huge dump fills are likely related to the use of the well in the 6th century or later. No trace of these floors and their subfloors was found west of well 742, but the high degree of disruption by later Byzantine and Frankish occupation may have destroyed them. Surface 1157 (86.46 masl) is a rock and tile floor bound together with plaster, covering the entire area east of well 742 up to robbing trench 1171. It was supported by a 0.25 m thick sandy and rocky fill containing 4th-5th NPD pottery (deposit 1162, 86.28 masl), which may have been some sort of bedding or leveling layer for the surface; during its excavation, Guy Sanders noted that this sandy limestone is the local bedrock and suggested this deposit in particular may originate from construction (or expansion?) of well 742 or well 902 to the northeast. Surface 1157 was replaced by surface 911, a reddish limey floor (N: 1009.42 N, E: 262.04 E, S: 1007.37 N, W: 259.22 E, 86.59 masl). This floor was quite well constructed, with a plastered cobble subfloor (structure 1153, 86.54 masl) and a highly compact soil bedding layer (structure 1156, 86.46 masl). By contrast, dirt surface 912 above it (N: 1009.33 N, E: 261.96 E, S: 1007.69 N, W: 259.63 E, 86.62 masl) was quite patchy and, having been exposed for a year, rather difficult to identify the boundaries of; had it not been identified as a surface last season, we probably would not have recognized it as such. All of these contexts were nearly sterile of cultural material other than tile, but they overlay context 1210 which contained 6th century material.

A series of square structures was built up against the south side of well 742, presumably in the Late Roman period along with the other features associated with the well (although we found virtually no datable material in any of them, and stratigraphically they can only be associated clearly with the well itself). All covered about the same area of 1007.15-1007.90 N and 258.35-259.15 E (0.75 x 0.80 m). We are uncertain about their use and function: the most likely is that they represent a sequence of entrances or thresholds to the well from the south, but they were not associated with any floors we identified during excavation; further excavation of Late Roman strata in the Byzantine room to the south may help resolve this. The earliest (no context number; unexcavated) is a platform or base consisting of a single layer of uncut cobbles and fieldstones 0.08-0.20 m large. These stones are difficult to differentiate from the stones supporting well 742 and from the stones of wall 1222, immediately to the west. This structure seems to have been built over the southeastern half of wall 1222, and extends further south than the overlying tile and ashlar structures, into the line of wall 366 and southern baulk. Excavation of this platform and the underlying soil deposit is necessary to clarify these relationships. These stones may have been a bedding surface for a level of flat laying tiles immediately above them (context 1220, 86.12, pottery Roman NPD). These tiles were then covered with a 0.05-0.10 m thick layer of reddish clayey sand (deposit 1209) which resembles the Late Roman fill operation of context 1210 (above). A large cut ashlar (0.60x0.70x0.40), which may have come from wall 1222, was placed on this soil, and topped by another layer of flat lying tiles (context 1158, 86.54 masl).

Unfortunately, we have not been able to identify any walls or primary deposits which correspond to the Late Roman use of the well, possibly suggesting that it was in an exterior space, or roofed with a simple wooden structure. It should also be noted that we retrieved no evidence for the well head in any period (i.e. no stones with use wear rope cuttings), which was therefore either wooden or spoliated at a later date for use elsewhere.

Around the Turkish House, we provided additional dating evidence for context 1080 (Session 1 Blue). Deposit 1143 is part of the same destruction fill as deposit 1080. Like 1080, 1143 contained large quantities of tile, along with pottery, glass (especially vertical foot goblets), and three additional coins (C-2013-192, -193, and -194, all illegible minimi). According to the current ceramic chronology, the pottery indicates a date in the second half of the 5th century, but a North African style moldmade lamp in local Corinthian fabric (L-2013-4) suggests a date closer to the mid-6th century for this deposition of material based on comparanda with material from the Fountain of Lamps.

Early-Middle Byzantine (7th- 10th century)
No early Byzantine activity has been detected in the excavation area, based on the lack of ceramic material dating to this period.

Middle-Late Byzantine (11th- 12th century)
Prior excavations have clarified that the large bothros pit in the middle of NW Nezi (Cut 870) was filled over the course of the mid 10th-late 11th century (see Session 1 2013 Blue final report). The putative N-S robbing trench Cut 1040 was also filled during this period (deposit 1136), further support that the far northwest corner of Nezi field around the Turkish House served as a source of stone for new construction activities and a garbage dumping area during this century and a half. So far, we have been unable to ascertain whether Cut 1040 (and therefore a now-missing wall from a prior period) continues south of Cut 870, as suggested by Harrington and Mokrišová.

South of wall 1087/747, which Harrington and Mokrišová date to the late 11th-early 12th century, another N-S wall was robbed at some point during the 11th century, represented by rectangular cut 1171 (N-S 1009.57-1007.06 N, E-W 262.98-262.28 E) extending northward from wall 366. Both fills in this cut, 888 and 1167, contained material dated to the 11th century. Based on the flurry of activity in this vicinity dated to the late 11th century, it is likely that cuts 1040 and 1171 provided source lime and stone for these constructions. Immediately south of wall 1087 is lime pit 882, initially identified and preliminarily excavated by George and Valente in 2012, dated by them to the late 11th century (based on material excavated from context 917; we recovered no material from the removal of the lime plaster itself which dates later than two micaceous water jar bases from the 6th century in context 1176). Even though the stratigraphy is unclear as none of these structures connect, the pottery phasings and common sense suggest that the sequence of events in this area is 1) robbing of walls, 2) cutting and use of the lime pit to create mortar for the walls from the robbed stone, and 3) construction of walls 1087/747 and 366 .

Well 742 potentially served as the source of water for mixing lime mortar, as the lime pit appeared to be set immediately against its north side. We identified a possible narrow cut around the top of the well, filled with a loose, dark black soil (deposit 1175, 86.40-86.16 masl) which was quite distinct from the surrounding red sandy Late Roman matrix; both this deposit and the comparable black soil around highest preserved course of the well (context 1185) contained 11th century pottery. Therefore, we suspect that well 742 was a two-phase structure, as after removal of the upper course, we have been unable to identify any type of cut around the well to indicate that the lower Late Roman fill was disturbed during the course of well construction. However, the construction method of the highest preserved course of the well – three rectangular ashlars with roughly carved exteriors, and a well carved curved interior surface around the opening for the well, held together with cobble chinking stones – is quite similar to what we can ascertain of the lower levels; on the other hand, George and Valente removed several stones from around the well in 2012, and these may represent the true 11th century construction and use phase of the well.

After construction of the walls, the area was adapted for occupational use, probably as a form of interior or exterior domestic space bounded by the wall 1087 at the north and 366 at the south. Our 2013 excavation of round, stone built hearth 910 generated ceramics dated to the late 11th century; the interior fills excavated by George and Valente in 2012 were dated to the early 12th, suggesting this hearth was only in use for a few decades at most. Construction of the hearth seems to have disrupted the western edge of lime pit 882, which was never concretely identified by either us or George and Valente; all we found were spotty bits of lime visible in the western face of the pit, quite dissimilar from the much better preserved solid surfaces of lime found to the east. Well 742 likely continued to be used as well.

Excavations in the area E of Wall 332 (N: 1006.06 N; S: 1003.75 N; E: 264.17 E; W: 262.77 E) allowed further investigation of the Byzantine strata in Nezi field.

The structure 452 (N 1005.51, S 1004.01, E 263.75, W 262.78) exposed during excavation in 2009 and removed in June 2013 has now been interpreted as the foundation of a staircase. It did not directly lay against any walls or other structures: the closest wall to this staircase, Wall 332, was separated from structure 452 by a thin layer of soil, 2 cm thick on average. Nevertheless we can argue that staircase 452 can be related at least to the western room bounded by Wall 332, Wall 366 and Wall. And it seems to be contemporary to northern Walls 540, 749 and 1087 the eastern Wall 807, even though no physical relationships are preserved. At the moment we can infer that during the Late Byzantine period the building identified in this area consisted at least of two floors.

Structure 452 laid on a hard packed earth deposit (deposit 1235). This fill was dumped South of Wall 366 and West of Wall 332 in order to have a proper surface for bedding the staircase foundation (structure 452). When deposit 1235 was dumped a hoard of bronze coins was deposited below the north side of Structure 452, few centimeters beneath the lowest course of stones. These coins belong almost all to the reigns of Nicephorus III and Alexius I, and range from 1178 to 1118 (coins 2013-221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 236, 238, 252, 254, 255, 256), except for two earlier ones: coin 2013-231 dated to 969-1030 and coin 2013-234 dated to1070 -1075, which were likely still in use at the beginning of the 12th century. The coins do not seem to have been contained in a ceramic vessel, we hypothesize that they were originally in a leather bag. The presence of a hoard beneath staircase does not seem to be unique in Nezi: in the field N of Nezi another hoard was identified in similar condition of preservation according to G. Sanders.

Unfortunately not much pottery was associated with this hoard. On the contrary, the new, wide deposit exposed beneath it and unexcavated appears to be very rich in ceramics. This new context is around the Frankish pits 527 and 528 and, according to the stratigraphy exposed in these cuts, it seems to be all the same deposit exposed beneath 1235 and to continue further south. We definitely encourage more excavation here: the numerous coins we found in 1235 can provide a terminus ante quem for this layer exposed beneath 1235.

Frankish (13th century)
The Frankish period is another phase of disruption, rather than occupation, in this area; George and Valente also identified a series of pits and wall robbings dating to the 13th century (see Final Report Session 3 Blue 2012). The western extent of wall 366 was robbed, an action described by cut 687 , and a shallow circular pit was dug below it at the presumed intersection of walls 366 and 306 (N-S 1006.91-1005.81 N, E-W 258.23-257.15 E, 86.17-85.82 masl). These cuts were then filled by deposits 1204 and 1183 around the third quarter of the 13th century.

Outstanding Issues
For the next season of excavation in Nezi Field, we recommend the following for this area:
1. Excavate the single course stone feature immediately south of well 742 and the fill beneath it in order to investigate the nature of the sequence of stone and tile structures overlaying it. More work may need to be done in the Byzantine room to the south in order to determine the full southern extend of the stone feature and the late Roman fills which pass under wall 366.
2. Remove the remainder of well 742 and the soil surface around it to resolve its initial construction and use.
3. Expose more of the pebble floor and tile floor to determine their extent, date, and relationship to walls 1222 and the earlier phase of 366, in order to learn more about Early-Middle Roman Nezi.
4. Resume work in the area around the Frankish pit 527, because the coin hoard in 1235 provides an excellent terminus ante quem for the underlying deposits