Trial trenches (depth 1 m) opened prior to the excavation revealed flint artifacts but no clearly identifiable archaeological layer. It was therefore decided to open two excavation squares (2 × 2 m each; depth of southern square c. 0.45 m, depth of northern square 0.1 m) and to excavate a long trial trench (length c. 12 m) using mechanical equipment (Figs. 2–4). All excavated sediments were sifted. Three strata of soil were exposed (3–1; Fig. 5), with a mixture of pebbles and chunks of flint, flint artifacts and potsherds. Most of the flint objects were worn. They date from the Middle Paleolithic and seem to indicate that the excavation area extended along the fringes of a knapping site of this period.
Remains from prehistoric periods have been known in southern Jerusalem since the beginning of prehistoric research. In 1897, Father Germer-Durand mountedan exhibition at the Notre-Dame de France of flint tools that he had collected in Jerusalem. At the beginning of the twentieth century, an Acheulian culture site was studied in Baq‘a and ‘Emeq Refa’im, about 1 km from the current excavation (Efimenko 1915). In 1933, the first prehistoric excavation was carried out in Jerusalem, on ‘Emeq Refa’im Street, by M. Stekelis and R. Neville (Stekelis 1948), and the site was subsequently excavated by O. Bar-Yosef (Arensburg and Bar-Yosef 1962). Handaxes and tools on Acheulian flakes in secondary deposition were found at the site and dated to the Lower Paleolithic. Many prehistoric remains were surveyed and documented by Stekelis (1956) elsewhere in Jerusalem. In recent years, remains from the Lower and Middle Paleolithic have been uncovered in clay layers containing flint pebbles in the Ramat Rael and in the neighborhoods of Baq‘a and Arnona (Barzilai, Birkenfeld and Crouvi 2008; Barzilai, Agha and Crouvi 2010), in Meqor Ḥayyim (Barzilai 2011) and on Israel Eldad Street (Permit Nos. A-7401, A-7118).
 
Stratum 3. A layer of brown clay was found in the lower part of the excavation, and above it were flint pebbles and large chunks of natural flint interspersed with a few flint items. This layer was almost sterile.
 
Stratum 2. Above Stratum 3 was a dark-brown clay layer (thickness 0.5–1.1 m) containing a dense mix of pebbles, large chunks of flint and numerous worked flint items (Fig. 6). The clay was very rich in calcareous concentrates, which indicate the creation of paleosol. Stratum 2 produced most of the finds in the excavation. The flint items, mostly worn, were not arranged on any horizontal layer nor in any concentration, but were rather scattered throughout the stratum together with flint pebbles. It therefore seems that they were swept here after the site was abandoned in a post-stratification process, and that they had originated at a nearby site. Most of the flint objects date to the Paleolithic period.
 
Stratum 1. Above Stratum 2 was a layer of brown clay (thickness c. 1 m) containing a great deal of silt and a few stones. It also yielded small worn sherds, dated mainly to the Roman, Byzantine and Early Islamic periods, along with flint items from the Paleolithic period and present-day finds. This layer seems to be colluvium swept to the site after the Paleolithic period.
 
Flint Items
Omri Barzilai
 
The excavation revealed 208 worked objects, which constituted 11.7% of all the flint items from the excavation (Table 1). All the objects, both worked and unworked, were made of brecciated flint of the Mishash Formation. Most were worn and coated with orange-yellowish patina, indicating that they had been rolled through water and transported to the site.
  
Table 1. Assemblage of flint objects 

Category

%

N

Primary elements

37.7

46

Flakes

52.5

64

Blades

4.1

5

CTE

5.7

7

Total debitage

100

122

 

 

Debitage

58.7

122

Chips

29.3

61

Tools

7.2

15

Cores

4.8

10

Total assemblage

100

208

 

 

Assemblage (modified)

11.7

208

Natural flints (unmodified)

88.3

1577

Total flints

100

1785

 
The worked objects are divided into four subgroups: debitage, chips, cores and tools (see Table 1). The debitage includes flakes, primary elements, blades (Fig. 7:1) and core trimming elements. The flakes, which were the most common element in the assemblage, include mainly medium-sized and small specimens, most of which had smooth striking surfaces. There was also a small group of Levallois flakes and flakes with a faceted striking surface (Fig. 7:2–5). All the cores in the assemblage are flake cores (Table 2), which include polyhedron, radial and Levallois types (Fig. 8) and tested nodules. The excavation uncovered 15 tools (Table 3), including retouched flakes, notches and denticulates, side scrapers, perforators and handaxes. Among the tools, the small handaxe stood out in particular (Fig. 9).
 
Table 2. Types of cores

Type

%

N

Polyhedral

40

4

Radial

30

3

Levallois

10

1

Tested nodule

20

2

Total

100

10

 

Table 3. Types of tools

Type

%

N

Scrapers

3

20.0

Notches and denticulates

4

26.7

Perforators

2

13.2

Retouched flakes

4

26.7

Handaxes

1

6.7

Varia

1

6.7

Total

15

100

 
In its composition and extent of preservation, the flint assemblage from the current excavation resembles that retrieved from the excavation at Meqor Ḥayyim (Barzilai 2011). Most of the assemblage consists of unworked chunks of Mishash Formation flint, whose origin has been identified along the watershed in the neighborhood of Armon Ha-Naẓiv. All the worked items are worn and coated with well-developed patina, suggesting that they were swept to the site from elsewhere.
The technological composition of the assemblage, which includes unfaceted flakes from polyhedral cores and especially Levallois items (flakes and a core), dates the assemblage to the Middle Paleolithic. A worn handaxe in the assemblage may indicate that part of the assemblage dates to the Lower Paleolithic.
The examination of the flint assemblage suggests a date in the Middle Paleolithic. The items’ wear and patina, along with the location of the site between the sources of flint in the Arnona neighborhood and the Meqor Ḥayyim site, may indicate that the assemblage was swept to the site. This probably occurred prior to the beginning of the Holocene, because the site was covered with dark clay which has been documented and dated at a number of sites in the vicinity of Arnona-Talpiyyot (O. Crouvi, pers. comm.).
 
Conclusion. The excavation revealed the continuation of a Paleolithic-period site. As in nearby excavations, here too an archaeological stratum was found within a layer of clay covering the bedrock. However, unlike previous excavations at Ramat Rael, Arnona and Eldad Street, most of the flint items are worn and are not densely distributed within the strata. It therefore seems that they were swept to the excavation area, which is probably on the fringes of a knapping site. The almost complete absence of handaxes (only one was found, on the surface) and the location of the site south of the ancient flow of the stream in ‘Emeq Refa’im, show that the current excavation area is part of the Middle Paleolithic site uncovered in Arnona, Ramat Rael and Eldad Street, and not part of the Lower Paleolithic site uncovered in ‘Emeq Refa’im. This was apparently a large knapping site, as evinced by the pebbles and large flint chunks in the clay layers, the flint outcrops discovered in the vicinity and the rather large quantity of worked flint objects.