Not by chemistry alone: Tephrostratigraphy, correlation, and context at two Palaeolithic sites in the southern Caucasus and Armenian Highlands

Understanding the timing of Late Pleistocene population expansions and interactions is hindered by myriad factors among which the imprecision of absolute dating is of particular concern. The identification and geochemical characterization of cryptically preserved volcanic ash layers (cryptotephra) in archaeological sites can alleviate this problem, and in the southern Caucasus and Armenian Highlands has allowed for the correlation of lithostratigraphic units at Palaeolithic sites in the two geographic regions. Compositionally distinct tephras (isochrons) allow specific stratigraphic layers to be linked temporally at the sites of Ortvale Klde (OK, Georgia) and Lusakert-1 (LKT-1, Armenia). The ∼30 ka V-18/Nemrut Formation tephra from the Nemrut volcano, eastern Türkiye is correlated with an Upper Palaeolithic layer at OK and a Late Middle Palaeolithic layer at LKT-1. This isochron suggests the long-term regional coexistence of technologically distinct populations, and the late survival of Late Middle Palaeolithic hominins. These results appear to capture a critical and until now elusive moment in hominin evolution, a period when two distinct tool-making groups, generally assumed to be Homo sapiens and Neanderthals, occupied the same region and likely interacted. However, careful comparison of results with site-specific sedimentological, taphonomic, archaeological, and chronometric data throw these correlations into question. Closer consideration of newly available glass geochemistry for Nemrut suggests further correlations: OK Layer 4c correlates to the V-30 tephra in Lake Van (∼42–37 ka); LKT-1 Lithostratigraphic Unit (LU) 4 correlates to both the ∼60 ka V-45/Çekmece and ∼62 ka V-51/Ahlat Pumice 6; and the younger LKT-1 LU 3 also correlates to the V-51, but with a different compositional range to those seen in the older LU 4. The only way to reconcile the range of compositions observed is that there are additional Nemrut eruptions that have similar glass chemistries to the deposits characterized, and that our current understanding of Nemrut's eruption stratigraphy is incomplete thus hampering our ability to resolve the age(s) of these distal tephra deposits with confidence. Therefore, while the analysis of cryptotephra holds great promise for understanding the timing of key evolutionary events, its application in the southern Caucasus and Armenian Highlands, and elsewhere, must be predicated on the careful consideration of site-specific contextual data rather than chemistry alone.