The 2024 expedition

The first season of the expedition focused on three areas (Figure 2a): the trenches within the cave (area A; Figure 2b), a lateral section to the east (area B; Figure 2c) and runoff deposits, which were disturbed by illegal excavations, where the path leading to the cave entrance merges with the modern road in the south (Figure 2a).

Figure 2.

Excavation areas in 2024: a) overview; b) area A; c) area B (figure by authors).

A total area of 16m² was excavated in the main trench in area A (Figure 3a). Large ceiling-fall blocks were encountered just below the surface, particularly in the west and south. An upper ashy deposit (0.20–0.60m thick) consists of stony debris mixed with organic material, including dung and modern refuse, and contains a disturbed assemblage combining Epipalaeolithic and Middle Palaeolithic (c. 200 000–45 000 BC) artefacts, likely due to post-depositional processes and recent disturbance (Figure 3b).

Figure 3.

a) Plan of the 2024 excavation trench with the positions of recorded features; b) north section of square L19–M19; c) north section of square N19–O19; d) stratigraphy of the northern profile (figure by authors).

Below this mixed layer, a second deposit (0.20–0.40m thick) represents the earliest intact Epipalaeolithic occupation. It consists of fine grit, ash and carbonised material, and includes burnt sediment and thermally altered flints (Figure 3c & d), suggesting an in situ fireplace (FP1). Radiocarbon determinations obtained from bone and charcoal samples place this layer at the end of the thirteenth millennium cal BC (TÜBİTAK-4157: 12 916–12 183 cal BC at 95.4% confidence; TÜBİTAK-4075: 12 413–12 070 cal BC at 88.4% confidence; C-1/2024; Figure 4a & b). The lithic assemblage supports these dates, although ongoing excavations should provide a better understanding of the cultural context.

Figure 4.

Radiocarbon determinations for a) charcoal sample TÜBİTAK-4157; b) bone sample TÜBİTAK-4075 (figure by authors).

In area B, excavations targeted a stratigraphic cross-section in compact sediments; only a few isolated lithics were recovered and no cultural layers were identified (Figure 2c). In the southern runoff deposit, systematic scraping and sieving yielded diagnostic Epipalaeolithic artefacts (lunates, backed blades, cores and points) and faunal remains.

The finds

The 2024 lithic assemblage comprises flint in various colours (predominantly red, green and yellow), obsidian and a small number of blades made from dacitic rocks. Flint and obsidian sources have not yet been identified. Preliminary analysis indicates a blade- and bladelet-based industry. The abundance of cores, crested blades and knapping debris suggests on-site production of tools. Cores are predominantly unipolar prismatic blade and bladelet cores (narrow and broad fronted), with single-platform and opposed-platform variants, as well as sub-pyramidal cores (Figures 5a–b & 6a–b).

Figure 5.

Selection of artefacts: a & b) cores; c & d) end-scrapers; e) microgravette point; f) pointed backed blade; g) lunate; h) lunate with bipolar backing; i & j) lunates with Helwan-type retouch; k) grooved antler pendant; l) animal-tooth pendant (figure by authors).

Figure 6.

Drawings of selected artefacts: a & b) cores; c) microgravette point; d) pointed backed blade; e) lunate; f) lunate with bipolar backing; g & h) lunates with Helwan-type retouch (figure by authors).

The chipped-stone tool assemblage includes end-scrapers (Figure 5c & d), notched and denticulated pieces, backed blades and bladelets, microgravette points (Figures 5e & 6c), pointed backed blades (Figures 5f & 6d), perforators, burins and lunates (Figure 5g–j & 6e–h). Some lunates show bipolar backing (Figures 5h & 6f), while four specimens display Helwan-type invasive bifacial retouch (Figures 5i–j & 6g–h). Overall, the lithic material is technologically homogeneous, indicating a consistent approach to blank production and tool manufacture. On technological and typological grounds, the assemblage can tentatively be attributed to the Late Epipalaeolithic (c. 13 000–10 000 BC), and the presence of lunates provides additional support for this assignment. In comparative terms, the blade/bladelet-based reduction sequence and microlithic components, including lunates and backed pieces with occasional Helwan-type retouch, show close parallels with Epipalaeolithic industries in Anatolia (e.g. Öküzini, Direkli, Eşek Deresi) and the Levant (e.g. Kebara, El-Wad, Hayonim) (Bar-Yosef Reference Bar-Yosef1998; Erek Reference Erek2012; Kartal Reference Kartal2009; Makarewicz Reference Makarewicz and Eren2012; Yalçınkaya Reference Yalçınkaya and Kartal2023; Gopher et al. Reference Gopher2025). Large-scale production and use of microlithic tools, including lunates, also occur in Late Pleistocene Caucasus sequences (e.g. Dzudzuana layer B, Apiancha, Satanai; Bar-Yosef et al. Reference Bar-Yosef2011; Golovanova et al. Reference Golovanova2014; Korkia Reference Korkia1990).

Identification of a grooved antler fragment and a perforated animal tooth (Figure 5k & l) indicate the use of personal adornment. Faunal remains and mollusc shells, likely of terrestrial origin, were recovered and await species level identification. Comprehensive faunal and botanic analyses are planned for forthcoming seasons.