The exposed sediment and soil sequence consisted of three Pedo-sedimentary units (Fig. 3; Table 1). The upper Unit 1 (thickness 0.2–0.8 m) is a recently disturbed soil profile, consisting of scattered chert clasts mixed with modern waste. This soil underwent agricultural use and thus, was not analyzed. Unit 2 is a clay-textured buried vertisol (thickness 0.5–1.2 m), containing Byzantine potsherds and Paleolithic flint artifacts. An abrupt wavy unconformity (gilgai micro-relief) separates this unit from theunderlyingUnit 3, which is a clay-textured buried vertisol, with cracks, slikenslides and manganese concretions (thickness >4 m). The evidence suggests that this unit and, to a lesser extent, Unit 2, comprised a high amount of swelling clay minerals (>30%). A calcic horizon (thickness c. 1 m) composed of 20–30% calcite nodules (size 1–3 cm), appears at a depth of c. 0.5 m from the top of Unit 3. Most of the Paleolithic flint artifacts were found scattered in this unit, having no horizontal order.
Table 1: Pedo-sedimentary unit characteristics
Unit 1 2 3
Thickness (m) 0.2-0.8 0.5-1.2 +4
Soil type N/A Vertisol Vertisol
Color (dry; determined by Munsell Color Charts) N/A  Brown (7.5YR 4/2) Dark grayish brown (10YR 4/2) 
 Field texture N/A  Clay  Clay 
Grain-size fractions (%): Grain-size distribution was analyzed using the Malvern MS-2000 laser diffraction (LD) instrument. The LD underestimates the clay fraction comparing the traditional methods (e.g., pipette): 15% clay measured using the LD is equivalent to ~40% clay measured using traditional methods.
 Medium and coarse sand (260–2000µm) N/A
 Fine sand (63–260µm) N/A   6
 Silt (2–63µm) N/A 78  82 
 Clay (<2µm) N/A 16  15 
 Structure N/A Angular blocky  Angular blocky 
 Boundary Gradual and smooth  Abrupt and wavy (Gilgai micro relief)  Not exposed 
Calcite (%)Calcimeter used for analysis; error is 5% N/A 10.7  11.3 at the calcic horizon 9.1 beneath the calcic horizon
Special features Chert clasts, Modern waste Flint tools and pottery Cracks, slikenslides; common very coarse irregular calcite nodules; few fine Mn concretions; big chert clasts (0.5–1 m) at lower part; flint tools
Throughout the excavation, a total of 4073 flint items was collected, comprising 1973 chipped stone artifacts and 2100 natural flint chunks (Table 2). Most were abraded and heavily patinated, with only a few fresh items. The flint type was local brecciated Campanian flint, common to the site area. Among these were extremely large blocks (20 × 50 × 50 cm) and small chips (1 × 1 × 1 cm) that were naturally embedded in the sediment of Unit 3 (Fig. 4).
Excavation(test pit)East-section cleaningTrenches 1–8Trenches I–IVSurfaceTotal (n)Total (%)
Total artifacts628696110264273197348
Natural pieces39198219346074210052
Total Flint101916783037243474073100
The lithic assemblage comprised mainly debitage and debris, while cores and tools were extremely rare. The few burnt items (N=36) may point to fire activities at the site.
Technologically, two flake industries were identified. The first and dominant one was of unprepared flakes (Figs. 5, 6:1–3). This knapping method was relatively simple, lacking preparations. This is supported by the debitage composition of mainly flakes and primary elements, while core trimming elements that usually imply planning, were scarce. The target products were thick flakes that could be divided into two groups according to size: medium (6–8 cm) and small (3–4 cm). The size seems to be an outcome of the initial core size rather than their ordinal stage within the reduction sequence.
The second and minor industry was Levallois, represented only by five flakes (Fig. 7). As opposed to the thick flakes, the thin Levallois blanks had a curved profile and were carefully prepared. Their low quantity in the assemblage was probably related to raw material. The local brecciated flint is hard and tends to break; hence, it is unsuitable for producing thin items, such as the Levallois flakes.
The cores consisted of six types (Table 3). The most frequent (N=16) were polyhedral flake cores (Fig. 6:4), which had at least three platforms with no hierarchy among them. The central surface cores (N=10) were the second most frequent type, as exemplified by a bifacial tool made on an exhausted core (Fig. 8:1). Conceptually, they resembled Levallois technique, lacking preparations by having a preferential flaking platform. The rest were core on flakes, tested nodules, core fragments (N=8 each) and blade cores (N=3).
 Table 3. Core-type frequencies
Polyhedral cores1630
Blade cores36
Cores on flakes815
Core fragments815
The tools consisted of 108 items, with a predominance of scrapers and retouched flakes (Table 4). Rounded, side and distal sub-types were common within the scrapers (N=45; Fig. 8:2, 3). The retouched flakes (N=28) included fine and inversely retouched items (Fig. 8:4, 5). Some of the tools were quite massive, having a bifacial retouch (Fig. 9). The remainders were notches and denticulates, retouched blades, burins, perforators, pointed flakes and varia.
Table 4. Tool-type frequencies
Retouched flake2826
Retouched blade66
Pointed flake44
Bifacially retouched tool44