The area lies along the national water divide, draining westwards to the Mediterranean and eastwards to the Dead Sea. It is located in the Mishash formation geological unit that is composed of thick layers of flint and chalk overlain by clayey soil (Picard 1956). In a previous excavation in 2006, the edge of a large Palaeolithic stone-knapping site was discovered in the neighborhood (Barzilai, Birkenfeld and Crouvi 2009
). Since the excavation was small and most of the surrounding area was built-up, it was difficult to estimate the size of the site. It was considered that the center of the site probably lay adjacent to one of the brecciated Campanian flint outcrops in the vicinity.
In the present excavation, two areas (Areas B and C) were excavated on the saddle (c. 790 m asl), located c. 150 m north of the 2006 excavation (Fig. 1). Area B (32 sq m) was opened on high ground that had not been damaged by construction work, and Area C (12 sq m) was opened on a lower level that had been cut and damaged by construction work. The excavation yielded soil layers, Middle Palaeolithic flint items and Byzantine pottery. Four trenches were mechanically dug next to Area C to try to determine the outer limits of the archaeological remains. A section cut by the construction work between Areas B and C was documented and sampled for geological analysis.
Area B (Fig. 2)
Stratum 4. The excavation reached the bedrock, comprised of flint rock, weathered as a result of tectonic activity.
Stratum 3. Above the bedrock was an 0.6–0.8 m thick layer of flint pebbles and pale clayey sediment on a moderate north–south slope, bounded on the west by a large hollow (Fig. 3). Several concentrations of flint pebbles and clay sediments were observed in the area, differentiated by the color of the sediments, the size of the flint pebbles and the ratio between the knapped items and the natural flints. For example, in the upper part of the stratum in the southern part of the area, there was a concentration of knapping debris characteristic of the Mousterian industry from the Middle Paleolithic period (250,000–50,000 years BP). By contrast, near the seam with the bedrock in the northern part of the area, a concentration of Levallois cores and flakes was found together with tools of the Mousterian industry (Fig. 4).
Stratum 2. A layer of dark, compacted brown clay was found. The upper part contained natural flints and knapped flints, as well as a few Byzantine potsherds and tesserae, while the lower part, near the seam with Stratum 3, contained large flint chunks and knapped flint items.
Stratum 1. The surface layer contained a gray soil fill intermixed with a few natural and knapped flints and modern finds. This layer was probably formed by mechanical surface leveling.
The area was characterized by a brown, compacted clay soil layer, reminiscent of that in Stratum 2 in Area B. It yielded natural flints, knapped flints and a few Byzantine potsherds and tesserae. In the northern part of the area, a soil heap, piled up during the construction work, yielded diagnostic flint items.
The many flint items collected in the excavation were formed on brecciated flint from the Mishash formation; some are fresh and sharp, indicating that they were knapped at the site. The flint assemblage is typical of the Middle Palaeolithic Mousterian culture. The assemblage consists mainly of cores and knapping debitage, characteristic of the initial stages of the knapping process, and large flint nodules that indicate that this was a knapping site. These features are very similar to those of the flint assemblage from the 2006 excavation (Barzilai, Birkenfeld and Crouvi 2009
). A few tools were found, including two trihedral adze tools with a triangular cross-section (Fig. 5) that came from the soil heap in Area C, and that were probably used to quarry flint nodules. Similar tools are known from Palaeolithic quarrying and knapping sites in northern Israel (Barkai, Gopher and La Porta
Area B was probably one of the spots where flint quarrying and tool production took place. The discovery of knapped items together with trihedrals suggests that the hollow in Area B was manmade and was the result of flint quarrying. The hollow was probably filled in with nodules and knapping debris, both when the flint chunks were quarried and later. The knapped flint items are characteristic of the Mousterian industry, and are similar to those discovered in the area excavated in 2006, indicating that both the areas are probably part of a single site.
Flint items in secondary deposition were recently discovered on Meqor Hayyim Street in Talpiyot, Jerusalem (Barzilai 2011
). Since these items were also formed on brecciated flints from the Mishash formation, and include items from the Mousterian culture, they may well have come from the same knapping site. The raw flint from the Mishash formation in the Arnona neighborhood was evidently used by the inhabitants of the southern Jerusalem region during the Palaeolithic period. The main focus of settlement in this period was in wadis on the fringes of the desert, whose natural caves served as dwellings. One such example is the Umm Zif Cave that was excavated in the 1930s, and that contained tools knapped on Turanian flint, as well as on brecciated Campanian flint that probably came from the Arnona area (Neuville 1934; Barzilai et al. 2020). It is therefore probable that the cave dwellers of the Middle Palaeolithic period, who lived in small nomadic hunter-gatherer groups, exploited the flint resources at the site for tool production.