Stratum 1. Two sections of a wall (W104; exposed length 12 m, width 0.75 m, preserved height 0.25–0.95 m) were exposed in the eastern part of the excavation area. The wall was aligned north–south at the bottom of a steep bedrock terrace that was specifically dressed for its construction. It was built of a row of medium-sized fieldstones, preserved a single course high; small fieldstones were placed along its eastern side. Most of W104 had collapsed and some of its stones were washed away; it was covered with alluvium and fieldstones. Fragments of jars from the Middle Bronze Age and ribbed body sherds of jars from the Roman–Byzantine periods were discovered in the alluvium.
Stratum 2. A farming terrace retaining wall (W100; width 0.65–0.90 m), aligned north–south and built of medium and large fieldstones with small stones in-between, was exposed on top of the bedrock, west of W104; it was preserved to a maximum of three courses high (0.85 m). Wall 100 delimited the western side of a farming terrace, which contained terra rossa soil (L109; width c. 5 m). The pottery discovered in that soil included a bowl fragment from Iron Age II (Fig. 4:4) and potsherds ascribed to the Roman period, including bowls (Fig. 4:8, 9), dating to the second–fourth centuries CE. A layer of alluvium (L109a; thickness c. 0.3 m) that had accumulated on the terra rossa soil in the farming terrace yielded potsherds from the Roman–Byzantine periods (third–sixth centuries CE), including bowl (Fig. 4:10) and a jar (Fig. 4:11). The accumulation of alluvium indicates that the farming terrace was used for a prolonged period. Next to the western side of W100 was a layer of small fieldstones (L115; Fig. 5) that descended steeply to the west on top of the bedrock. Within the layer of stones was a complete cooking pot dating to the Early Roman period (first century CE; Fig. 4:7) that had been deliberately placed there. The aperture of the pot was covered with a body sherd of a jar; the pot was empty. Four large elongated fieldstones arranged in two courses of headers-stretchers (L123) were placed above the pot, next to W100.
The ceramic finds from Layer 115 were mixed and included an amphoriskos (Fig. 4:1), a ledge handle (Fig. 4:2) and jar (Fig. 4:3) from Middle Bronze I, a holemouth jar (Fig. 4:5) and jar (Fig. 4:6) from Iron Age II and also several ribbed body fragments of jars from the Roman period.
Stratum 3. Three retaining walls (W101–W103) of modern farming terraces (I–III) were exposed. The walls were built on earthen fill that was leveled (thickness 0.25–0.35 m) above the bedrock. It seems that the fill was brought here from the top of the slope at the beginning of the twentieth century CE when the area was prepared for agriculture, and that it served as the foundation for the farming terraces. This fill was rich in Roman-period potsherds (second–fourth centuries CE), including bowls (Fig. 6:1–6) and jars (Fig. 6:7, 8), as well as potsherds from the Byzantine period (fourth–sixth centuries CE), including bowls (Fig. 6:9–14), a jar (Fig. 6:15) and a lamp (Fig. 6:16). It seems that these potsherds originated at a site dating to the Roman–Byzantine periods, which was located at the top of the slope. Terrace I sealed below it Farming Terrace 2 and its retaining wall (W100). Olive trees were planted on Terraces I–III. Numerous potsherds from the Roman period (second–fourth centuries CE) that had been washed down from the top of the slope were discovered on the surface of the terraces.
Remains of a dwelling (L120; 7×10 m; Figs. 7, 8) that was built at the end of a bedrock cliff were documented northwest of the excavation area. It was built of ashlars in secondary use and large fieldstones bonded with coarse mortar. The structure’s ceiling was vaulted and the floor consisted of concrete. A wooden door with an iron lock and hinges was discovered on the floor. The main opening was set in the northern wall of the building (W106). A relieving arch was built on the inside of the opening, above the lintel stone. This wall stood to almost its entire height unlike the rest of the building’s walls that had partially collapsed and their stones were scattered about. Incorporated in a niche built in the western part of W106 was a section of a steel railroad track used as a lintel, which indicates the building was not constructed earlier than the mid-nineteenth century CE. Below the southwestern corner of the building was a drainage system whose opening was composed of large masonry stones which led to a natural crevice in the bedrock.
Remains of a watchman’s hut (4.5–5.5×8.0 m; Figs. 9, 10) were documented c. 40 m south of the building; it was preserved to practically its entire height. Founded on the bedrock descending to the west, it was built of different size fieldstones, with dressed stone in the corners of the building. The hut had two rooms (L121, L122) whose openings were set in the southern side. Room 121 had a dome roof, whose top had collapsed inward, and Room 122 was a courtyard covered with a temporary tin roof. The structure was used as a shelter until recently.
It seems that the site was first inhabited at the beginning of the Middle Bronze Age. This site is probably the continuation of a site that was excavated c. 500 m north of here where a Middle Bronze II temple was exposed (Permit No. A-6087). After an extended hiatus, the area was prepared for agriculture in the Roman period. Numerous potsherds from the Roman–Byzantine periods were discovered at the site and these probably originated in a large settlement that was located at the top of the slope in these periods. A complex olive press and remains of building foundations from the Byzantine period (Permit No. A-5918) were exposed along the fringes of the site, on the saddle at the bottom of a spur that descends north of Nahal Refa’im. During the Byzantine period, this region was almost completely abandoned and this continued until the twentieth century when a village was established at the top of the slope and new farming terraces were prepared along its slopes.