In the fourth area of excavations, where during the past two seasons we had brought to light the northeastern corner of the large pi-shaped complex, we turned our attention to the excavation of the areas to the north and the east of the building in order to investigate its relation to earlier structures that had also appeared. To this end, we extended the trench by 5 m eastwards, and continued the excavation within the boundaries of the possible road to the north. We removed the topsoil consisting of natural fills resulting from the erosion of the slopes of the acropolis (Context 3082) with the help of the small bobcat. It also contained mixed material dating as late as the 7th century CE. With this layer removed, we discovered the eastern continuation of Walls 3045 and 3065 and the outline of Wall 3091 perpendicular to the walls above. These three walls along with Wall 3066, all built with ashlar blocks, define a rectangular room measuring internally 4.7 x 5 m. The south wall misses a section ca. 2.2 m long, where we identified the robbing trench (Context 3089) of the corresponding stones which goes as far down as the bedrock. From the fill of the trench (Context 3090) we collected limited pottery dating up to the early 5th century CE. From the layers that we excavated inside the room (Contexts 3093, 3088, 3100) we retrieved a relatively small amount of pottery (a few hundred sherds) with the latest dating to the early 7th century CE. In order to investigate the foundation of the wall that surrounded the room, we dug the northern half of the room down to bedrock (Context 3115) which appeared at a depth of just 40 to 50 cm below the surface of the walls. The pottery found at this depth (190 sherds) dates up to the 1st century CE. The two parallel walls are preserved in two courses, the bottom resting on bedrock with the upper one serving as a euthenteria, while Wall 3045, oriented east-west, includes a third course below because the bedrock slopes down towards the east. It became clear that the parallel Walls 3066 and 3091 are later than Wall 3045, given that they are built against it. With the exposure of the natural bedrock, the foundation trenches of Walls 3045 and 3091 also appeared (Contexts 3124 and 3126 respectively). As expected, the foundation trench of the more solid Wall 3045 is wider and deeper. In the fill of this trench (Context 3125) we found just eight sherds, of which three were Hellenistic and one was Early Roman. From the soil found in the trench of Wall 3091 (Context 3127) we recovered just one sherd, belonging to a thin-walled mug of Early Roman date.
To the east of the room we just excavated the topsoil (Context 3082), and we shall wait until we extend the trench eastwards next year, in order to dig deeper. In the area to the south of the room, the contexts under topsoil (Contexts 3092, 3099, 3102) produced pottery dating as late as the late antique period. Once removed, we discovered a distinct layer of rubble near the northwest corner (Context 3104) and a well (Context 3118) framed by two stone bases in the middle of the trench. The rubble deposit (Context 3104) covered a pit of irregular shape, ca. 2.4 x 2.9 m (Context 3105), dug down to bedrock and exposed to a depth of ca. 80 cm below the surface of Wall 3066. The pottery found in this deposit dates up to the late 6th to early 7th century CE. The well (Context 3118), 0.70 m of inner diameter, has a stone rim made of four unequal-sized blocks with a curved carving on their inner side. The excavation of the interior advanced to a depth of 1.80 m in arbitrary passes (Contexts 3130, 3134, 3135) but was not completed. The passes contained much rubble and tile fragments together with pottery dating up to the early 2nd century CE, which provides a terminus ante quem for the filling of the well. The upper part of the well, 1.08 m high, is lined with three courses of stones (0.35, 0.27 and 0.46 m high from top to bottom), while the rest is carved in bedrock. On the west side of the well and adjacent to it a rectangular stone base is preserved in situ, measuring 0.62 x 0.48 m. A second base, measuring 0.47 x 0.47 m, stands to the east of the well, 2.20 m from the first. The use of these bases is not certain yet, namely whether they supported wooden posts for holding the roof or a wooden lifting apparatus for drawing water from the well. The surface of these two stones is ca. 20 cm higher than the rim of the well. The distance of almost 1 m that separates the eastern base from the well goes against the second hypothesis.