The early middle palaeolithic blade industry from Hummal, Central Syria

Abstract

The Hummal site, situated in the El-Kowm area of central Syria, is characterised by the presence of many artesian springs related to faults in the substratum, and by high-quality Lower Eocene flint outcrops. The Hummal site is in direct contact with the old artesian spring that supplied water to a pool of variable size. Attracted by the water, animals and raw material, humans settled continuously in the immediate vicinity of the source from the Lower to the Upper Palaeolithic, as attested by an archaeological record more than 20 metres deep. Systematic excavations in Hummal began in 1999 under the direction of J.-M. Le Tensorer and S. Muhesen (Le Tensorer 2000). More than 20 archaeological layers from Upper to Lower Palaeolithic were recognised and thousands of artefacts gathered. This in situ sequence, containing layers 6a, 6b, 6c and 7a, 7c, integrated the Hummalian. A blade industry was additionally discovered in a massive sand deposit, subsequently labelled alpha h. From 2001 systematic excavation of the upper sequence of the Hummalian was undertaken under the direction of the author. The excavation area reached 26m2, and more than 7000 lithic objects and more than a hundred faunal remains were collected. The excavated area was divided into two distinct parts: West and East. In 2009 the new Sondage S1 was opened in the southern part of the site and a surface of about 2m2 was excavated. The Hummalian blade industry excavated in all three sectors is subdivided into stratified archaeological layers and is clearly positioned between the Yabrudian and Mousterian complexes. Taphonomic factors such as erosion, diagenesis and trampling, alongside the probable lack of sedimentation, had a destructive effect on a number of the archaeological remains. The common flaking technique is direct percussion with a hard hammer. The unidirectional flaking system dominates in all layers, but bidirectional is also well represented, especially in Sand alpha h and layers 6c-2 and 7c. The goal of production was elongated blanks regardless of their size. The blank blades encompass a number of specimens with different morphologies. These blanks, although looking morphologically different – either prismatic or Levallois-like – seem to be the result of a single reduction strategy involving different kinds of core volume management. It seems that the flint knapper moved from Laminar debitage to Levallois-like debitage when the volume of cores decreased, with the core becoming flatter and requiring more preparations to control the manufacture of blanks. But many times the morphology of cores seems to have remained constant despite the diminishing size, showing that the core volume management was maintained from the early stage through to exhaustion. As blank production was carried out until exhaustion of the core, the assemblage includes blanks with a size scale ranging from elongated blades to small bladelets. But there was also a separate production of bladelets from burin-cores, and of bladelet cores and small flakes from truncated-faceted pieces. Their presence cannot be solely interpreted as being aimed at maximising the productivity of the flint. The end-products obtained during their flaking must have represented a desired supplementary element next to implements manufactured by the main reduction strategy. All these elements indicate a level of complexity in blank production. The importance of recycling in the Hummalian is demonstrated by double patinated tools, the reuse of broken items and debris for bladelet manufacturing, and Yabrudian scrapers as cores. The retouched tools made on flakes and blades seem to be quite standardised in their metrical and non-metrical attributes, both between the assemblages and the tools categories. The estimated TL age for Hummalian is approximately 200 ka (Richter 2006, Richter et al. 2011) and is comparable to those of the Laminar phenomenon highlighted at Hayonim layer ‘F top’ and ‘F base’, which have mean TL-dates on heated flint of 210 ± 28 ka and 221 ± 21 ka, respectively (Mercier et al., 2007), or with Tabun’s unit IX (Tabun D-type), with its mean TL dates of 256 ± 26 ka, and Rosh Ein Mor, dated 200 ka (Rink et al. 2003). These assemblages were discovered at different site types that varied in their use of Laminar and Levallois reduction strategies and in their production of diverse tools. The collections from Tabun and Rosh Ein Mor, in contrast to the Hummalian, seem to be dominated by the Levallois method (Meignen 1994, 143, Hauck 2010, 200). At present it seems that the lithic industries from Hayonim layers F and E (Meignen 1998, 2000) and the undated Abu Sif layers B and C (Neuville 1951) show the greatest resemblance to the Hummalian industry.