Introduction:
This is the final report for the second session of excavations of the 2015 season for Room 3, Unit II in the area of Temple E SE. In the 1990’s, Room 3 was divided by the grid system and, consequently, excavated in two different sequences. The south side was excavated during April 1992 (NB848: 1-152), by C. Keesling, while the north side was excavated during the second and third session of the 1992 season (NB853: 52-83) by M.R. Scottan. During the 2015 season, this area was unified and the excavation area was bounded by walls on all four sides. Based on the few stones still visible from the lowest courses of the walls, their original outline was traced (North Wall: 1066.80-1065.30N, 109.20-114.50E; South Wall: 1058.50-1056.90N, 109.40-115.70E; East Wall: 1065.00-1057.90N, 114.50-115.70E; 1065.20-1058.50N, 108.50-110.10E. All measurements restored).
It is important to note, however, that all four walls that delimited the room where heavily reconstructed during April 2015. Only the S section of the W wall presents a stretch of original wall longer than 1.5 m (1063.00-1058.40N, 108.90-110.40E), which turned up to be essential for the dating of the room, as illustrated below. Even when some traces of the wall were visible, blocks were covered by new courses of stone, making it impossible to assess the dimensions of these original sections. The original traces of walls, as represented in the plans of the excavated context, were traced from an orthophoto of the room that postdates the consolidation works and later compared to the visible traces of original masonry on the sides of the walls. Emilio Rodríguez-Álvarez and Kurtis T. Tanaka supervised and recorded the excavations of Room 3, with the assistance of workmen Panos Stamatis, Hecuran Ҫoli, Michalis Vathis, and Memos Karvouniaris. Dr. Guy Sanders (Director) and Larkin Kennedy (Field Director) supervised.


Goals of excavation:
The aim of the excavation of this area was to resume the work left by Keesling and Scottan during the 1992 season and to obtain a clearer picture of the chronological sequence of the use of this room, in preparation for the backfilling of this area and its subsequent opening to visitors. Current excavations were complicated by the room’s exposure since the 1992 excavations as well as the conservation works mentioned above, leaving both the upper layers of the room dry and compacted. It was only with the removal of these topmost layers that the stratigraphy of the room became clearer.


Frankish Period (AD 1210-1450)
The Mid 13th Century
Context 771, a dump fill, constitutes the earliest deposit excavated in the area during the 2015 season, dating to the mid 13th century based on the pottery. This deposit helped to provide the chronology, based on stratigraphic relationships, to contexts 778, 780, and 782, dated as well to mid 13th century. These contexts together compose a compact brown surface with charcoal inclusions that were cut by the later large pit created by cut 777 (113.90-109.90E/1058.90-1065.00N). Though this surface was noticeably compact, it does not seem to have been an actual floor as it is likely that at this time this area was an open space west of the church. This surface did not extend into the NW corner of the room. Rather, at this time (dated by the pottery), fills associated to the reconstructed N wall (contexts 730, 111.30-109.50E/1062.65-1065.20N and 732, 112.70-111.30E/1065.30-1065.90N) were deposited in this portion of the room. No coins were recovered from the contexts of this period.

13th to 14th Centuries
Later in the 13th century, or perhaps in the early 14th, a deep circular pit defined by cut 763 (113.60-112.50E/1061.85-1062.85N) (filled by 746 and 755), ca. 1m in diameter and 2m deep, was cut into the surface made up by contexts 778, 780, and 782. The pit was filled rather quickly after the cut was made as joins were found throughout. The material recovered from the pit was characterized by a much higher density of pottery, bone, and charcoal, most likely the waste of food preparation that took place nearby. One sample from context 746 and two from context 755 were taken for flotation and analysis. It should be noted, however, that no evidence for such food preparation activities was found in Room 3 itself. Three coins were recovered from context 746 (nos. 2015-158, -164, and -165) and five from context 755 (nos. 2015-182, -195, -197, -209, and -211). Of these, three are roughly contemporary, 2015-164 and -182 of William Villehardouin, and 2015-197, a French issue possibly of Louis IX. The pit was also rich in non-ceramic finds, including a large amount of glass (including prunted beakers), iron fragments (including an iron handle inventoried as MF-2015-15), shell, and bronze fragments, including a bronze finger ring inventoried as MF-2015-23. The only decoration the state of preservation of the ring allows to elucidate is a small incused square in its centre.
This juncture spanning the 13th and 14th centuries also saw significant activity in the southern extent of the area that would become Room 3. At this time two levelling fills (contexts 657 and 667, dated by the pottery) were deposited in the area. Their original extents cannot be traced due to later activity in the room (namely the large pit formed by cut 777, discussed below), however, it is clear that they were deposited over the surface defined by contexts 778, 780, and 782. These two layers extended under the walls that would later define Room 3, and thus provide a terminus post quem for the walls’ construction. Indeed, it is possible that these two fills were deposited in anticipation of the South wall, with the context 667, rich in large stones, laid to provide a more stable basis for the wall’s construction. Though dated by the context pottery, seven coins were recovered from context 667 (nos. 2015-78, -79, -80, -82, -84, -85, and -86) and eight from context 657 (nos. 2015-58, -59, -60, -67, -68, -69, -72, and -73). Of these, only coin 2015-69, an issue of William Villehardouin, is roughly contemporary, though still somewhat earlier, than the pottery. The rest are earlier issues, mostly dated to Manuel I or Latin imitatives. Along with an average amount of iron, glass, and bronze recovered from the area, three bronze spindle hooks were recovered along with one bronze bead or button, inventoried as MF-2015-27.
In the northern extent of Room 3, levelling fills were deposited (contexts 677, 681, 727, and 737), each dating between the second half of the 13th to the 14th centuries, based on the pottery, likely to prepare the area for the “Frankish floor” removed above these contexts during the 1992 campaign. These were deposited directly overlying the surface defined by contexts 778, 780, and 782 and were later cut by a pit excavated in 1992 (pit 1992-1). One coin was recovered from context 677 (no. 2015-92) and two from context 727 (nos. 2015-178 and -179); two of these (nos. 2015-92 and -179) were Manuel I and dated to the Byzantine period, and thus are much earlier than the date given by the pottery. Coin 2015-178 is a Latin imitative type A coin, and dates somewhat closer to the period defined by the pottery, ca. 1204-1261 CE. One find was inventoried from these contexts, a lead weight (MF-2015-8) found in context 677.
Sometime in the late 13th to early 14th centuries a large shallow cut (context 777, 113.90-109.90E/1058.90-1065.00N, filled by 693, 697, 709, 721, 738, 744, 771, 776) extended over a large portion of Room 3. This cut truncates the pit defined by cut 693 and cuts the levelling fill for the “Frankish floor.” The earliest fills of this pit are contexts 744 and 776, dated by the pottery to the late 13th to early 14th centuries. Only one coin was recovered from these contexts, no. 2015-157 from context 744, a Latin imitative type A coin, dating slightly earlier than the context itself.
The early 14th century
The large pit (cut context 777) was filled more extensively in the early 14th century (contexts 738, 697, 709, 721, and 693). Two coins were recovered from context 738 (nos. 2015-142 and -150) and five from context 697 (nos. 2015-101, -102, -103, -107, and -108). Of these, all were earlier Byzantine issues save nos. 2015-103, a French issue of the early 13th century, and 2015-142, an issue of William Villehardouin. Contexts 697, 709, and 721 were especially rich in finds, with a large amount of iron, glass (including fragments of prunted beakers), and bronze recovered. One find was inventoried from this fill, a bone die (MF-2015-22) from context 709. Despite the numerous finds from the fills of the pit, its purpose remains enigmatic. Unlike the pit made by cut 693, it is less clear if this pit was made solely to receive the waste of food preparation. The presence of butcher marks on some of the bones, however, prompted the taking of two flotation samples from context 721, and it is hoped that the analysis of these samples will yield a clearer picture of the use of this area and the pit.
This pit was capped by levelling contexts 690 and 649. Into this last deposit, a pit (cut 653, filled by context 650, 111.80-110.90E/1061.45-1062.10N) was dug at about the same time as its initial deposition. All this activity dates to the early 14th century based on pottery (contexts 690, 650). Only one coin was recovered from these contexts, no. 2015-52 from context 649, a coin issued by William Villehardouin dating as late as 1278. Finds were comparatively few from these contexts, but include a few glass fragments of a prunted beaker and an iron double hook (none were inventoried).


Conclusions:
Work in Room 3 during the 2015 season yielded important information pertaining to the phases of use of the area that would become Room 3. One of the most significant findings of the season was yielded by contexts 657 and 667, which provide a terminus post quem for the construction of the walls of Room 3. It seems likely that these two fills were used to create a firm, level basis for the walls and thus the walls must post date the late 13th to early 14th centuries.
We are thus able to rethink the early use of Room 3 as an open area in the cluster of rooms west of the church. The two main pits in the area (cuts 763 and 777) provide ample evidence for the types of activities that might have occurred in the wider area, namely the preparation of food (though it should be reiterated that no evidence was found for such activity in the room itself) and the use of the area of Room 3 as a location to dump the detritus from those activities. It would seem, however, that the walling in of the area was not enough to change the area’s function, as the cutting of the large central pit (cut 777) appears to post date the construction of the walls. The surface associated to the definition of this area as a room would be the context excavated in 1992 as Frankish floor, a white clay surface recorded extensively in the excavation of the Frankish quarter (described in detail in NB848:150-151 and NB853: 80).


Recommendations for Future Excavation:
As the area is currently being backfilled to be made ready for tourists, it is unlikely that further exploration will be easily made in the area. A more fruitful endeavour would then be to reconcile the findings of the 2015 season with those of the 1992 campaign. In particular the general dating of the area needs further work, as pit 1992-1, dated to the mid-13th century, seems to have cut through contexts to which our work has ascribed later dates. A review of the materials recovered from the 1992 excavations will therefore be necessary to clarify the chronology of Room 3.