Following the stratigraphic sequence described by D. Garrod in her pioneering excavations at Tabun Cave, the Levantine Mousterian is traditionally divided into three sub-stages (B, C, D) according to changes in modal forms of Levallois production. Despite the ubiquity of this terminology, none of these three stages from Tabun has been described in detail. Furthermore, our knowledge of variability in technology and chronology within the Levantine Middle Paleolithic has increased markedly in the last three decades. This variation is especially apparent in the Early Middle Paleolithic, when both Levallois and non-Levallois laminar methods appear in most sites but with different frequencies. In this paper we provide a reconstruction of the Levallois technology in the early Middle Paleolithic of Tabun Cave using assemblages from Jelinek's excavations in Unit IX (Garrod's layer D). In contrast to many other contemporary sites, where non-Levallois laminar production is strongly represented, the assemblage of Unit IX is dominated by a recurrent, unipolar Levallois technology. Understanding the specific character of the "prototype" of Tabun D is an important step towards assessing inter-and intra-assemblage variability within the Levantine Early Middle Paleolithic. Appreciating this variation is in turn essential for explaining the apparently sudden appearance of these early Mousterian assemblages between 200 and 250 kya. The method used at Tabun Unit IX is both flexible and efficient, yielding both large numbers of blanks and a range of products while reducing the waste of material involved in shaping and maintaining of the core. In these assemblages, blades, flakes and Levallois points, and a variety of other products, were produced through systematic exploitation of different parts of the core's surface. In this respect the technology of Tabun Unit IX differs from the 'laminar method' known from other early Levantine assemblages in the organization of production and in its economic features and the range of products.
INTRODUCTIONA rcheological studies at Tabun Cave over the past 75 years have fundamentally shaped perceptions of the Middle and Lower Paleolithic in the eastern Mediterranean (e.g., Garrod 1956;Garrod and Bate 1937;Grün and Stringer 2000;Jelinek et al. 1973;Jelinek 1975Jelinek , 1977Jelinek , 1982a Jelinek , b, 1990McCown and Keith 1939;Mercier and Valladas 2003;McPherron 2006). The first phase of work at the site resulted in the standard tripartite division of the Levantine Middle Paleolithic, in which the earliest phase ('Tabun D') is characterized by the abundance of blades and elongated points (Copeland 1975; Garrod and Bate 1937: 77-78; Jelinek 1982a,b; Shea 2003 and references therein). Despite the site's significance, and despite important subsequent work by Jelinek et al. (1973), Meignen (1994 and Monigal (2002), the character of lithic production in the Early Middle Paleolithic (henceforth EMP) of Tabun has not been fully explored. In apparent contrast to many other EMP sites in this r...