EXCAVATION SUMMARY
We, Josh Gieske and Laurie Kilker, from April 11 to April 23 2008 excavated in the courtyard of the Byzantine house uncovered by the 1961 excavations in Agora SW-G, overseen by Steven Lattimore (notebooks 230 and 235), which is now part of the area currently referred to as “North of Nezi.” We worked under the supervision of the Corinth director Guy Sanders and assistant field director Alicia Carter. We excavated with workmen Kleomenis Didaskalou (pickman) and Vasilis Kollias (shovelman and barrowman).
Our goals were to first remove the martyr, which was left on the edge of a trench excavated in the southeastern corner of the courtyard (presumably by the 1961 excavators, although we have not been able to find reference to the excavation of this roughly 1.5 m deep trench in the notebooks from that year) and then to explore the floor surfaces of the courtyard under the predominantly Frankish levels removed by the 1961 excavations. We spent approximately one week on each goal.
The internal space of the courtyard runs from 1035.70 N, 1030.60 S, 266.30 W, 274.10 E, but we excavated only within the following coordinates: for the martyr: 1035.19 N, 1030.68 S, 271.25 W, 273.53 E; for the entire courtyard: 1035.58 N, 1030.61 S, 266.53 W, 272.77 E.


FILL OF THE 1961 MARTYR (1100-1200 AD)
The martyr was located immediately west and north of an exploration trench presumably excavated in 1961, although we have not been able to find reference to excavations in this area by any of the 1961 excavators. The martyr surrounded this trench on two sides; the trench is bounded by a wall to the east and a martyr that may have supported a later wall at a higher level (now eroded) to the south. The excavations by Steven Lattimore in Agora SW-G seemed to have as their eastern boundary a later wall, which according to Lattimore’s drawings, rested immediately on top of the martyr we excavated, while the excavations in Agora SW-F by William Berg stop at the eastern and southern walls of the courtyard (i.e., in the rooms with the staircases). Handwritten notes in Lattimore’s final report suggest that excavations may have continued in the courtyard in 1963, but if so, we have not been able to find any notes or reports from these excavations. The martyr then seems to lie under the unlabelled wall, which runs north from wall 61, that was removed by Lattimore. This raised section of the courtyard was probably left around the edge of the deep trench in the southwest corner to prevent the walls of the trench from collapsing. If so, then the trench in the southwest corner was probably excavated at the same time as the rest of the courtyard in 1961.
Our excavation of the martyr began with the removal of two pi-shaped terracotta drains, which were laying on the northern edge of the martyr. Unfortunately, since the soil around these drains had been eroded practically down to the base of each drain segment, we did not have sufficient material to date the construction of this drain. However, we have observed in several plans of the area south of the South Stoa that the drain appears to have once continued approximately 2 m further east, passing through the short wall that blocks up the entrance corridor of the courtyard, and that its course was lined with stones much like the two we found in situ next to the drain. In fact, the placement of the stones in this wall, blocking the entranceway, seems to have a purposeful gap to allow passage of the drain. It should be noted that stones have been removed from this wall so that the gap is no longer readily apparent.
Excavation of the martyr itself began with the cleaning pass (5191: El. 85.25), which removed the top layer of soil that had been exposed to the elements for nearly 50 years, and concluded with context 5264, after which we had reached the level of the courtyard in all areas of the martyr (lowest elevation 84.91). Due to the relatively small area covered by the martyr, many of the 13 contexts we excavated are rather difficult to interpret since their eastern and western extents were almost universally unknown. In particular, all contexts in the southern half of the martyr below N 1033.5 could be at most 0.75m wide, thus we can only speak of these in very broad terms. Additionally, within the martyr itself, we found much evidence of root activity that both created false boundaries between contexts and produced tunnels and patches of loamy soil. For example, context 5228 may be entirely the result of plant action. The final stratum of the martyr was an approximately 15cm thick accumulation of debris (contexts 5252 and 5258). This layer was dated by pottery to the early 12th century.
While the dates assigned to the pottery from the martyr ranged from Middle Byzantine to Late Byzantine, it was only the Late Byzantine pottery which afforded more specific dating by century. It seems likely that the small size of the pottery samples we were able to collect from the martyr was the reason for this lack of precision in dating many of the contexts. However, working from those which were able to be dated precisely, we have decided that most or all of the martyr should be dated to the 12th century and more precisely to the first half of that century. We do not think that the designation of some pottery as Middle Byzantine poses a problem to this conclusion because those contexts dated as Middle Byzantine produced very few sherds, and moreover, the likelihood of contamination of a small, raised section of soil left by an excavation almost 50 years ago is great.
With regard to the two coins associated with the martyr context, the first from 5191 was actually found while the workmen were enthusiastically cleaning the area around the well and was merely associated with 5191 as a convenience. It dates neither the martyr or the courtyard floor because the area around the well was excavated to an area below the general level of the courtyard by Steven Lattimore in 1961. The second coin was found in a mixed fill deposit 5252 and the numismatist was not sure whether or not the object was, in fact, a coin. Thus, the dating of the martyr must depend entirely on the pottery evidence and the notes from the 1961 excavations.


BYZANTINE COURTYARD OF THE 1961 HOUSE (1080-1120 AD)
After the removal of the martyr and the creation of a more or less uniform surface throughout the courtyard (with the exception of the trench in the southeast corner and the area around the well which were excavated in 1961), the two largest portions of which were dated by pottery to 1100 +/-20 years (context 5324) and 1110 +/-10 years (context 5327). Between them, these two floor surfaces occupied the southern two-thirds of the courtyard. Despite their dissimilar appearance—5324 was a tile and cobble floor while 5327 was a hard-packed, poor-quality pebble floor—that these two floors were in existence simultaneously is suggested by several factors. Beyond the clearly contemporaneous pottery dates, the presence of context 5315, an accumulation of silt from rain water washing loose soil to the lowest spot in the courtyard, which forms a continuous surface between the southern halves of 5324 and 5327 also demands that both floor surfaces existed simultaneously. In addition, the indefinite edges of and the diffuse relationship between the northern halves of 5324 and 5327 suggest the coexistence of the two disparate floor surfaces. Both floors become very similar in compaction and inclusions in their northern halves where 5324 supposedly overlaps 5327, and the difficulty we had determining the separation between the floors may have been due to the fact that they were at one point the same floor. Certainly their elevations support the idea that both contexts belong to a single surface, and it seems only to be their varying appearances that suggests against the idea that they are one and the same floor. Because we found traces of neither floor under the other in their best preserved sections (the western portion of 5324 and the eastern portion of 5327), it seems safe to assume that perhaps neither floor ever covered the full extent of the courtyard and thus each covered only a half of the space. Unfortunately, the excavation trench from the 1961 excavation has removed much of the SE corner of the courtyard so type of floor which existed in that corner cannot be known.
In the north there was a third floor context (5332) which appeared to be a continuation of the tile and cobble floor to the south, although a much poorer version of it. The date given by the pottery (late 11th - early 12th) is consistent with the date of ca. 1100 given to the other two floors. This northern floor ran up against the north wall of the courtyard, while to the east it was terminated by an area in the NE corner which was excavated to a lower level in 1961. To the south its edge was indefinite, in the same general area where the edges of 5324 and 5327 were also found to be indefinite. Over all these indefinite edges lies context 5309 which was excavated as a mound lying on top of the floor, but may in reality have been the continuation of floor 5324 to the north. The edges of 5309 were also indefinite and there was no clear transition into the parts of three floors (5324, 5327, 5332) which lay beneath it. Unfortunately the pottery from this context was too small to give a precise date, but in light of the likelihood that the contexts to the north and south of 5309 were all part of the same floor, 5309 itself may have belonged to the same floor (and perhaps it was disturbed or simply very uneven in this area). If this is the case, then there is a single, continuous floor surface that covers most of the remaining area of the courtyard. The small patch of good-quality pebble floor (context 5291) may also belong to this floor, as it both abuts and lies immediately on top of the same red layer (part of which was excavated as 5293 and tentatively dated to the late 11th – early 12th century) as 5332.
Of the two remaining areas of the courtyard where this floor surface has yet to be identified, one is a cut (filled by 5288 and 5289) that separates floors sections 5291 and 5327. The other is the NW corner of the courtyard, the area where excavation in the courtyard should be resumed. Here, under remnants of a hard-packed surface with some tiles (context 5298) we made a preliminary exploration of a succession of layers that represent construction activity in the 1961 house. The lowest of these layers appears to be a pebble floor of decent quality, on top of which is a thin layer of plaster thrown down or mixed on the floor for use elsewhere in the house. This secluded NW corner would have been an ideal, out-of-the-way place for such an activity. Above the thin residue of plaster leftover from the remodeling was a softer layer of debris, possibly thrown down purposely on top of the plaster. On top of all these layers were the remnants of a later, poorer quality floor which were removed as context 5298. The relationship of the floors in this corner to those in the rest of the courtyard will only be clear once the NW corner of the courtyard is excavated, but it may be that the same construction project which accounts for the plaster also produced the fill (5288 and 5289) placed in the cut in NE corner mentioned above. Unfortunately, the large area removed around the well in 1961 creates an additional difficult in relating all the different patches of floor in the courtyard.
Above the potentially uniform floor surface composed primarily of contexts 5325, 5324, 5327, and 5332 with the possible addition of 5291 and 5309, were left very few remains of later levels (apart from the martyr).One of these, a small circular deposit (5296) in the SW corner of the courtyard (loosely dated by a few sherds to the 11th century), we initially thought to be obscuring the course of the N-S rubble wall. However, beneath it we found not the line of the wall but instead the same material as under the adjacent floor 5324, so that either 5296 cuts into 5324 or, more likely, is merely another portion of this somewhat uneven floor surface. The N-S rubble wall in the SW corner should be examined more closely to determine how it relates to these contexts and the courtyard floor. The other later level remaining in the courtyard was a remnant of a tile and cobble floor (5300) placed immediately on top of the one below (i.e. 5324). This second floor, most of which had been removed in 1961, was dated by pottery to the late 11th – early 12th century, so it may simply be an addition to the floor below to create a more level surface. Because 5300 continues under the western half of the south wall of the courtyard (which is later than the eastern half) into the next room to the south, excavation in this room may help clarify the stratigraphy of the courtyard.
One anonymous coin, dated to 969-1030 AD, was found in the excavation of the courtyard (in context 5332), but because such coins tend to remain in circulation for a long time before being deposited, this coin was not used for dating the floor in which it was found. In the same context was found a fragment of a twisted glass bracelet classified as Byzantine in date. The only other find of note from the courtyard was a Geometric bronze pin head, which could have come from a disturbed grave, possibly in the area of the courtyard.


CONCLUSIONS
All of the contexts we excavated this session appear to belong to a period of less than 100 years. The martyr itself probably represents deposits accumulated sometime in the mid 12th century, while in the courtyard as a whole most if not all of our context are concentrated around the year 1100. Thus, the floor surface created in the courtyard around that date appears to have been in existence for only a generation or two at most, perhaps receiving some repairs or additions where needed, before it was covered over and a new floor was created at a higher level. Some suggestion of this rather dramatic change in floor level can be seen in the entrance from the courtyard into the room to the NW. Here, a step, which very likely corresponds with the courtyard floor we excavated, once led through a 2m wide opening in the north wall of the courtyard into the room beyond. However, at a later phase, probably with the creation of a new floor sometime in the mid to late 12th century, the step must have been covered over and a new threshold created at a higher level from three reused blocks which have rectangular cuttings in the blocks on the ends for door posts. Excavation in the NW corner of the courtyard and in the room to the south of the courtyard should provide additional evidence to help reconstruct this period in the life of the 1961 Byzantine House.