The beginning of the Upper Paleolithic in the Iranian Zagros. A taphonomic approach and techno-economic comparison of Early Baradostian assemblages from Warwasi and Yafteh (Iran)

Journal of Human Evolution 65 (2013) 39e64

Contents lists available at SciVerse ScienceDirect

Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol

The beginning of the Upper Paleolithic in the Iranian Zagros. A taphonomic approach and techno-economic comparison of Early Baradostian assemblages from Warwasi and Yafteh (Iran) Tsenka Tsanova Max Planck Institute for Evolutionary Anthropology, Deutscher Platz 6, 04103 Leipzig, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Received 18 July 2012 Accepted 16 April 2013 Available online 03 June 2013

Southwest Asia is a key region in current debates surrounding the appearance of the first cultures attributed to anatomically modern humans, particularly the Aurignacian and preceding cultural units of the Iranian Zagros, Levant, and the Balkans (Baradostian, Ahmarien, Kozarnikien, etc.). The Zagros mountain range encompasses an immense territory that remains understudied with regard to the Upper Paleolithic as well as the first bladelet industries traditionally presumed to be the work of anatomically modern humans. Concerning the emergence of the Aurignacian, the sites of Warwasi rockshelter and Yafteh cave in the central Zagros are considered to show evidence of in situ evolution of the Upper Paleolithic from the local Mousterian. This hypothesis is tested by way of a taphonomic, technotypological and economic approach applied to the Upper Paleolithic levels of Warwasi (spits LLeAA) and Yafteh (the series from the lower part of the sequence). A comparison of the techno-economic features of both assemblages demonstrates a conceptual bond with contemporaneous technocomplexes from Levant and Europe (Ahmarian, Protoaurignacian, etc.). The techno-typological Middle Paleolithic character of the Warwasi lithic assemblage permits a discussion of a possible in situ dependence/continuum from the Mousterian or perhaps particular activities linked to the type of the occupation of the site. However, bladelet technology cannot be considered as rooted in the Zagros Mousterian. Consequently the origin of the Aurignacian sensu stricto has to be reconsidered. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Aurignacian Baradostian Mousterian Warwasi Yafteh Zagros

Introduction Europe and Western Asia witnessed significant cultural and biological change in the early phase of the Upper Paleolithic (UP) period, between 40 and 35 ky BP (thousand years before present). The final Neanderthal populations associated with the Mousterian became extinct with the arrival of anatomically modern human (AMH) populations from Africa via the Levant, according to the archeological data (Goring-Morris and Belfer-Cohen, 2003; Mellars, 2006), or via a southern route through Arabia according to new genetic studies (Mélé et al., 2012). In Europe, the Aurignacian1 is

E-mail address: [email protected]. Paleolithic culture or tradition from the beginning of the Upper Paleolithic first discovered and defined in Southwest Europe and related to the arrival and development of anatomically modern humans in Europe. The earliest chronological phases of the Aurignacian in Southwest Europe, the Protoaurignacian and Early Aurignacian, are characterized by different stone technologies (rectangular versus curved bladelets) produced according to distinct chaîne opératoire along with different bone industries. They have different typo-technological definitions and chronology (e.g., Bon, 2002; Teyssandier et al., 2010). The Aurignacian phenomenon is no longer considered as a homogeneous culture across Europe. 1

0047-2484/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jhevol.2013.04.005

associated with the full development of art objects, personal ornaments, and graphic representations, which are original expressions of the new AMH population and indicate an ability to store symbolic information outside of the human brain, a behavior associated with behavioral or cultural modernity (e.g., Klein, 2000; McBrearty and Brooks, 2000; d’Errico, 2003; Henshilwood and Marean, 2003). In Western Asia and also in the central Zagros, the UP is associated with the appearance of new projectile bladelet technology at Warwasi (Olszewski, 1993) and at Yafteh and also with the appearance of personal ornaments (Otte et al., 2007). The original name given to the UP of the Zagros, the ‘Baradostian’, is named after Mount Baradost (Iraq) where R. Solecki excavated Shanidar cave (Solecki, 1958). Later, Olszewski and Dibble (1994) renamed the Baradostian assemblages in the Zagros as Aurignacian based on a typological comparison between European and Levantine Aurignacian assemblages. Based on the studies of Olszewski and Otte, the central hypothesis for the origin of the Baradostian is that it emerged from the local Middle Paleolithic (MP) (Olszewski and Dibble, 1994; Otte and Kozlowski, 2004). The sequence from Warwasi, a rare case where a Mousterian is succeeded by rich Baradostian levels, is the primary reference for this hypothesis (Olszewski, 1999).

40

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

The importance of the Zagros for debates concerning the emergence of the UP and cultural modernity is thus based on the following: a) It is where Africa meets Eurasia and is one of the probable sources for the colonization of Eurasia by AMH (Fig. 1). The Zagros has been considered as a hub for the formation of new technologies that continually functioned as an ethnic ‘reservoir’ (Otte and Kozlowski, 2007). b) One of the principle lines of support for the replacement model is anchored in the Zagros, this region having been put forward as the origin of the Aurignacian. This has been the case ever since the first research was undertaken in the region (Garrod, 1937). c) This interpretation has been refuted (Garrod,1953; Smith, 1986) or reaffirmed (Olszewski and Dibble, 1994, 2006; Olszewski, 1999, 2001; Kozlowski and Otte, 2000; Otte and Kozlowski, 2004) on numerous occasions. d) Finally, the possibility of mechanical mixing of the Warwasi sequence rather than secure stratification (Braidwood et al., 1961), given the lack of taphonomic studies (Soriano, 2007), has been well noted by Otte and Kozlowski (2007). However, these authors note that a mixing of Mousterian and Early Baradostian levels will be difficult to explain by sloped layers and bioturbation because of the thickness of these ‘transitional’ Baradostian levels of between 1.5 and 1.8 m (Otte and Kozlowski, 2007).

The main purpose of this paper is to describe the archeological context and to characterize the lithic collections from the earliest Upper Paleolithic assemblages from Warwasi and Yafteh. What are the archeological features which permit us to discuss the appearance of behavioral modernity in the central Zagros and when did these occur? How is the regional diversity of the Early Baradostian bladelet tradition expressed? Is there in situ evolution from the Zagros Mousterian? Finally, could the central Zagros be the origin of the Aurignacian? Inter-regional comparisons allow us to discuss the spread or influence of the Baradostian tradition toward the north in the Caucasus. History of research D. Garrod conducted the first Paleolithic research in the Middle East (Iranian Kurdistan and coastal Levant) during the 1920s. Garrod’s investigations led her to comment on the origin and in situ development of the Aurignacian with diffusion from east to west (Garrod, 1930, 1937). After excavating the Levantine sites of Ksar ’Aqil and El-Wad, Garrod modified her position on the origin of the Aurignacian. She discussed the latest appearance of the Aurignacian in the Levant and a possible European origin of this culture with diffusion from west to east (Garrod, 1953). Garrod’s work in Iraqi Kurdistan was continued in the 1950s by Ralph and Rose Solecki with the excavation of Shanidar Cave (Solecki, 1955, 1998). Neanderthal remains discovered there were associated with Mousterian industries thought to date to ca. 44 ky 14C BP (Solecki and Solecki, 1993). The Baradostian, dated to ca. 34 ky 14C BP in

Shanidar

Warkaini

Mar Tarik Malaverd Warwasi Bisotun Ghar-e Khar Pa Sangar Yafteh Kunji Gar Arjeneh

Garm Roud 2 Moghanak Sefid Ab Qaleh Gusheh Qaleh Bozi

Eshkaft-e Gavi Ghadi Barmeh Shur

Figure 1. Geographical position of the main Baradostian and Mousterian sites in Iranian Zagros.

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

Shanidar Cave, is separated by a 10-ky chronological hiatus from the underlying Mousterian levels (Solecki and Solecki, 1993). In 1951, C. Coon excavated the Middle Paleolithic cave of Bistun (Coon, 1951). R. Braidwood and B. Howe worked at Warwasi (Braidwood et al., 1961) and F. Hole with K. Flannery excavated Yafteh cave in 1960s and the rockshelter of Arjeneh (Hole and Flannery, 1967).2 In the 1970s, M. Rosenberg excavated the cave Eshkaft-e Gavi, located in the southern Zagros Mountain (Fig. 1). This site preserves both Middle and Upper Paleolithic occupations (Rosenberg, 1985). Human bones from the UP were recently published (Scott and Marean, 2009). A molar tooth occurs at the base of the UP levels and near the boundary with the Middle Paleolithic, and appears to represent AHM. Beginning in 2000, a BelgianeIranian team headed by M. Otte excavated a new test pit at Yafteh to verify the previous dates for the Baradostian (ca. 40 ky 14C BP; Hole and Flannery, 1967) and obtained a date of ca. 35 ky 14C BP, which is considered to represent a minimum age (Otte et al., 2011). Two perforated Cervidae canines and two pierced shells were found along with a hematite pendant (Otte et al., 2007). These objects constitute the first known manifestations of this type of activity from a Baradostian context and evoke the Aurignacian tradition as it is defined in Europe (ibid.) Recently in northern Iran (central Albroz) three open-area Paleolithic sites from the UP have been excavated: Garm Roud, Moghanak and Otchounak (Fig. 1). As at Shanidar Cave, the Baradostian in these sites shows a discontinuity with the underlying Mousterian. They also lack the earliest phases of the UP (Berillon et al., 2007). The lithic assemblage from Garm Roud 2 is reminiscent of the Baradostian in the Zagros (Berillon et al., 2009). In 2005, a team from the University of Tübingen and the Iranian Center for Archeological research conducted field-work in four different areas of the Zagros (Conard et al., 2006). They discovered new Paleolithic localities in the eastern Karkas region (eastern part of the central Zagros) in Bardina, and they identified a number of Arjeneh points and small number of ‘rods’ typical of the Baradostian (Fig. 1). Previous work on the collections from Warwasi and Yafteh Since the 1980s, the material from Warwasi and Yafteh has been the subject of several studies. Olszewski, working with the Warwasi assemblages, described two phases of UP. The lower levels (spits AAeLL) are characterized by blade(let) production on prismatic and carinated cores but also by a high percentage of flakes and MP tool types such as retouched points and scrapers. This industry, which she re-defined as Early Zagros Aurignacian, was compared with the Bachokirian (Tsanova and Bordes, 2003) of Southeast Europe (Olszewski and Dibble, 1994), which has been interpreted as the first dispersal of the Aurignacian (e.g., Kozlowski and Otte, 2000). In this view, the Early Zagros Aurignacian is the unique example in the region for an in situ technological and typological transition from the Mousterian (Olszewski, 2001). The re-defined Early and Late Zagros Aurignacian after Olszewski’s work are part of the Early Upper Paleolithic (EUP) complex, which include blade and bladelet assemblages and have a similar or synchronous development with EUP sequences in the LevanteEarly Ahmarian (Olszewski, 2001). The Late UP from Warwasi (spits PeZ) is compared with the Levantine Aurignacian and European Late Aurignacian with typical production of twisted bladelets on burin cores (Olszewski, 2001). Otte and Kozlowski published a monograph on the UP collections from the central Zagros (Otte and Kozlowski, 2007). They

2 The majority of the recovered material was sent to different museums in the USA, with the exception of half the Yafteh collection housed in Yale Museum (Connecticut, Yale collection). The second half of Yafteh collection has been housed in Iran (Tehran collection).

41

documented part of the assemblages stored in the United States from Yafteh (at Yale University) and from Warwasi (in the University of Pennsylvania Museum of Archaeology and Anthropology) and have defended the hypothesis of a local origin for the Baradostian technocomplex belonging to the Aurignacian tradition. J.-G. Bordes and S. Shidrang described the bladelet production in the Baradostian levels from Yafteh (collection from the Archeological Museum of Tehran). Based on different types of bladelet technology, they split the sequence into two parts: (1) a lower part corresponding to the Early Baradostian with long, pointed, straight and curved bladelets, with direct bilateral retouch; (2) an upper part corresponding to the Late Baradostian with small twisted bladelets obtained from burin cores, with inverse retouch on the concave edge and direct retouch on a convex edge (Shidrang, 2007; Bordes and Shidrang, 2009). Bordes and Shidrang noticed the parallel development of the Early Baradostian in the Zagros with the synchronous phases in the Levant (Early Ahmarian) and in Europe (Protoaurignacian). They pointed out the absence of a west European phase of Early Aurignacian (Aurignacien ancien sensu stricto) in Iran and the Levant, and they proposed that the origin of the Aurignacian sensu lato is not in the Zagros (Bordes and Shidrang, 2009). Materials and methods The present study concerns the entire Warwasi assemblage stored at the University of Pennsylvania Museum of Archaeology and Anthropology (Philadelphia) and the Yafteh collection from Frank Hole’s 1960 excavation, stored at Yale University in the United States. The aim of this study is to re-evaluate the hypothesis proposing an in situ origin for the Baradostian from the local Mousterian (Warwasi) and to document the Mousterian features that might support an in situ development of the Early Baradostian based on the assemblages from Warwasi and Yafteh. We apply a taphonomic approach to the Early Baradostian of Warwasi with the goal of assessing mixing between Mousterian and Baradostian levels, presumed to be a result of natural causes or excavation methods (Olszewski and Dibble, 1994). A taphonomic analysis of lithic industries consists of developing and applying certain methods designed to measure and test the validity of distinct assemblages of excavated remains. This type of analysis employs refits to address stratigraphic issues (e.g., Cahen et al., 1980; Hofman and Enloe, 1992; Bordes, 2000). The present analysis re-evaluates the Warwasi sequence by looking for conjoining blade fragments and technological refits, as well as ‘second order refits’ of raw material groups (e.g., Petraglia, 1992). A techno-typological study was also conducted for the two collections with the aim of producing a synthesis of the concepts of reduction methods and significant techno-typological characteristics (economy, desired products, and blank production goals). This study is based on the chaîne opératoire approach whose concepts and analytical tools have been made explicit in numerous publications (e.g., Tixier, 1978; Inizan et al., 1995; Pelegrin, 1995, 2000). Warwasi and Yafteh: previous excavations, description of the sites, stratigraphy and chronology Warwasi rockshelter: presentation of the site and its archeological sequence The rockshelter at Warwasi was excavated as part of Braidwood’s extensive archeological research in the eastern end of the Fertile Crescent in the 1950s and 1960s (Braidwood and Howe,

42

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

burin cores. There is stratigraphic continuity and most probably overlap between the Late and the Early Baradostian spits. The very typical regular twisted bladelet and carinated burin cores for the Late Baradostian are identified in the top of the Early Baradostian levels (spits AA and BB). The Early Baradostian levels (spits AAeLL, 1.2 m thick) consist of less standardized bladelets, larger, yet still thin, and less regular than those observed in the overlying levels. The assemblage contains numerous pointed pieces (flakes), similar to Mousterian points and small ‘truncated-faceted’ flake-cores (cf. Dibble and McPherron, 2007) also known as Kostienki knives or Nahr Ibrahim cores (Débenath and Dibble, 1994; Schroeder, 2007). Between the bottom of the Early Baradostian levels and the top of the Mousterian levels the lithic density decreases. However, looking at the distribution of the quantity of artifacts there is not a clear stratigraphic hiatus between the Mousterian and the Early Baradostian (see also Olszewski and Dibble, 1994). Hole and Flannery suggest that the industry from the Early Baradostian levels has two chronological phases (Hole and Flannery, 1967). The earlier levels (DD-LL) are strongly flake-dominated, while in levels AA-CC flakes and bladelets are more evenly distributed, and levels PeZ are dominated by bladelets (Olszewski and Dibble, 1994). c) The Middle Paleolithic deposits or Zagros Mousterian is localized at the base of the sequence in spits JJeCCC (w1.8 m thick) (according to Dibble and Holdaway, 1993) or NNeCCC (according to Otte and Kozlowski, 2007 and this study) and is characterized by the production of small Levallois flakes (1.2e 2.0 cm), as well as small ‘truncated-faceted’ cores. Retouched products consist of scrapers and/or convergent pieces identical to Mousterian points. The intense rejuvenation of numerous scraper edges is noted throughout the sequence (Dibble and Holdaway, 1993). Figure 2. Picture of Warwasi rockshelter excavation (F. Hole photo).

Yafteh cave: presentation of the site and its archaeological sequence 1960; Braidwood et al., 1961). The site is located 19 km northeast of Kermanshah in the Kurdistan region of western Iran, at an altitude of 1300 m above the Tang-I-Knisht Valley in the Zagros Mountains (Fig. 1). The excavation method within the Pleistocene deposits consisted of arbitrary 10 cm thick spits labeled from the top as AeZ, AAeZZ, and AAAeCCC. The excavated surface was 8.0 m2 to a depth of 5.6 m (Fig. 2). Small finds (less than 2 cm) were sieved and carefully recovered for each spit. No radiometric dates are currently available.3 Stratigraphy of the lithic assemblages and initial observations. Based on the existing publications (Olszewski, 1993; Otte and Kozlowski, 2007; Bordes and Shidrang, 2009) and my initial observations, Warwasi’s archeological sequence contains three major technological phases characterized as follows (from top to bottom): a) Epipaleolithic or Zarzian, 1.9 m thick; b) Late and Early Zagros Aurignacian (sensu Olszewski, 1993), or Late and Early Baradostian sensu stricto (after Otte and Kozlowski, 2007), located in spits PeLL, 2.2 m thick. The lithic material of the Late Baradostian levels (spits PeZ) consists of long, narrow, and pointed bladelets, which are very regular and standardized and essentially derive from carinated

3 Efforts have been underway since 2010 at the MPI-EVA, Leipzig, to obtain radiocarbon dates from fauna samples from Warwasi. Unfortunately the collagen in the bones is insufficiently preserved.

Yafteh cave is located about 13 km to the northwest of the present town of Khorramabad in Luristan province (western Iran), in the heart of the Zagros mountain range (Fig. 1). The cave, of karstic origin in the limestone cuesta of Kuh-I Sefid, is located at an elevation of 1278 m above the sea level. Yafteh was discovered in 1963 by Frank Hole and Kent Flannery during their investigations into the prehistory of southwestern Iran. The two test trenches excavated in the entrance of the cave cover a surface of around 16 m2 (Fig. 3). The recent excavations made by Otte and his team have confirmed and completed the observations of the stratigraphy established by Hole and Flannery (Otte et al., 2007). Twenty-one levels were excavated using the same method of 10 cm spits used at Warwasi. The separation of the levels is based on anthropogenic indicators: lithic and bone concentrations as well as fire places corresponding to successive phases of occupations to over a depth of 2.2 m (Fig. 3). Pleistocene sediments in some places show post-depositional bioturbation as a result of small rodent burrowing. However, most of the excavated areas appeared to be intact. Ash lenses and scatters to a depth of more than 190 cm have been clearly observed (Otte et al., 2007; Shidrang, 2007). As emphasized previously, the lower part of the sequence, dated to ca. 35 ky 14C BP (Otte et al., 2011), is characterized by the production of rectilinear bladelets from cores on small blocks (prismatic, sub-pyramidal) or on the edges of flakes. These bladelets were reserved for the production of Arjeneh points or large Dufour bladelets (the shortest being 3e4 cm in length; e.g., Fig. 16). The upper part of the sequence, dated to ca. 33 ky BP, sees the appearance of carinated elements and witnesses the progressive replacement of Arjeneh points by twisted Dufour bladelets. There is

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

43

Figure 3. (A) Picture of Yafteh cave; (B) plan of Yafteh cave showing the location of Hole’s 1965 test pits, the 2005 test pit and location of the studied lithic sample; (C) west profile F-G15/16 (after Biglari et al. Otte et al., 2007); (D) west profile of F. Hole’s test pit, square 4de4c (unpublished).

44

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

AA

overlying spits X, W, V, U, T, S, R, Q, P

BB CC DD EE FF GG

Intensely altered

HH

Slightly altered Fresh, none altered

II JJ KK LL NN OO QQ 0%

20%

0%

20%

40%

40%

60%

60%

80%

80%

100%

100%

AA BB CC DD Cores EE FF

Flakes Elongated flakes Blades

GG

Blade-bladelets

HH

Bladelets Burin spalls

II JJ KK

underlying spits TT, UU, VV, ZZ

LL

Figure 4. Warwasi. Top: statement of conservation of the Early Baradostian lithic assemblage from spits AA to LL and into three underlying spits of Mousterian (NN-OO). Intensely altered artifacts: covered with patina, with crushed edges or with blunted flake scar ridges; slightly altered: with patina or slight patina, with fresh edges and scar ridges; fresh artifacts: without patina, with sharp edges and fresh scar ridges (see SOM Table 1); Bottom: distribution of technological groups from spits AA to LL.

no difference in the state of preservation of the remains between the two segments of the sequence. Just as with Warwasi, the density of the lithic material progressively diminishes toward the base of the sequence. Chronology. First, the Yafteh sequence was dated in the 1960s by a series of radiocarbon dates (Hole and Flannery, 1967). The 11 charcoal samples come from ash beds from the lower, Early Baradostian, part of the sequence (below 200 cm). The minimum age of >40,000 (SI-335) 14C years BP indicates an early UP occupation of the site (Hole, 2012). However, some of the dates are relatively imprecise and inconsistent with the stratigraphy (ibid.).4 In 2005 and 2008, the Yafteh sequence was re-excavated and redated (Otte et al., 2007). The new radiocarbon dates seem to be in better agreement with the stratigraphic succession and indicate an occupation age between approximately 24,500 and 35,500 14C years BP (uncal). Moreover the ten new charcoal samples collected in the lower part of the Yafteh sequence have been interpreted as a single chronological signal around 33,500 14C years BP with the

4 Seven of the eleven dates sampled by Hole and Flannery (1967) are from square Y4e. On the bottom of the square, at 290 cm and 285 cm, the dates are >35,600 BP (GH-706) and >40,000 BP (SI-335). In the same square at 250 cm another ash bed was dated to 21,000
800 BP (SI-336).

Figure 5. Schematic reconstruction of the Warwasi sequence: EUP Baradostian spits, part of the underlying Mousterian spits and the overlying Late Baradostian spits. Distribution of 15 second order refits groups.

lithic assemblages corresponding to a single phase (Otte et al., 2011). Analysis of the Early Baradostian assemblages from Warwasi and Yafteh Taphonomic and techno-typological study of the Warwasi Early Baradostian assemblage (spits LLeAA) The interest in the taphonomic approach is to evaluate the stratigraphic position and integrity of the Early Baradostian assemblage from Warwasi. Here, surface alteration, raw material groups (second order refits), conjoins, and technological refits are all considered in assessing the integrity of the assemblage. Of particular importance is a determination of whether the conjoins and refits are inter- or intra-level. All the lithics from the Early Baradostian levels (spits LLeAA) were sorted and included in the taphonomic study. A series of diagnostic pieces (complete artifacts with preserved platforms, pieces from shaping and re-shaping cores.) from each level (more than 15%) were documented in detail for the techno-typological characterization of the Early Baradostian (Table 1, see Tables 3 and 4). Our initial observations allowed us to separate the groups of artifacts made on similar raw materials, to assess the initial size of the blocks and to assess the techno-typological integrity of the Early Baradostian assemblage.

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

45

Figure 6. Warwasi. Top: examples of selection of second order refit groups 1 and 5, first step for finding blade and debitage refits. Down: artifacts refitting 2, 3, 6, 11, 16 and 8 debitage refitting of three overshoots.

46

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64 overlying spits X, W, V, U, T, S, R, Q, P

underlying spits TT, UU, VV, ZZ

Figure 7. Schematic reconstruction of the EUP Warwasi sequence and distribution of the blade fragment conjoins and debitage refits. Example of debitage refitting 4a: core and its overshoots blades distributed between spits AA, BB and GG.

Raw material. The raw material used at Warwasi is of unknown provenance. This raw material is a cryptocrystalline radiolarite chert that most often has a red-brown color with a fine, opaque, smooth and homogeneous grain. Similar and probably identical red-brown chert, smooth and sometimes striped, has been discovered in the massif of Bisotun (Jaubert et al., 2004). What little is preserved of the cortex in the Early Baradostian assemblage is fresh and indicates collection from a primary deposit. There are few indications of the original nature of the block. Some preserved cortex on the cores and the flakes indicates that the nodules have an original size of 4e6 cm (see below). All the color variations (gray, blue-green, and blackbrown, yellow-orange) are observable within the same piece. A recent survey in the Kermanshah plain (south and northwest from Warwasi) has confirmed rich sources of radiolarite nodules and rectangular shaped raw blanks used in the Zagros Mousterian assemblages (Biglari, 2004, 2007). Surface alterations. The entire lithic assemblage from the Early Baradostian levels and four spits from the underlying Mousterian were counted and classified according to their surface state (N ¼ 8685). Three different degrees of alteration were used for classification: (1) intensely altered (covered with patina, with crushed edges or with blunted flake scar ridges); (2) slightly altered (patinated or slightly patinated, fresh edges and flake scar ridges); and (3) fresh (no patina, sharp edges and fresh flake scar ridges). The criteria for this sub-division were based on macroscopic observations of the patina and freshness of the edges and ridges. The Mousterian assemblages contain a large percentage of altered pieces. Beginning with spit LL and continuing to the base of the sequence (to the Mousterian), the number of strongly altered pieces grows considerably (Fig. 4A). The Early Baradostian material (levels LLeAA) seems to be less altered than the Mousterian. The intense alteration concerns more than 7% (76 of 1068) of the UP artifacts (bladelet cores and bladelets). The surface alteration seems independent of the tool type and technology. In the lower parts of the

Early Baradostian spits both MP and UP types can have similar surface alteration. Most of the Early Baradostian artifacts are slightly altered (Fig. 4). The density of artifacts increases from the bottom toward the top of the sequence. The artifacts decrease between the Mousterian and the Early Baradostian levels (spits NN and LL). Moreover the distribution of the techno-typological groups in the Early Baradostian spits seems to be relatively consistent (Fig. 4B). We point out the decreasing frequency of flakes toward the upper levels of the Early Baradostian, with the parallel increase of bladelets (Fig. 4B). Between the Early Baradostian and Late Baradostian levels the quantity of artifacts increases considerably (spits AAeZ). In spit AA, fresh material becomes more common and altered artifacts less so. The Late Baradostian material has a low degree of alteration. Based only on the surface alteration and the density distribution of the artifacts, it seems impossible to clearly outline the limits of the Baradostian phases. Stratigraphic continuity and possible relations between the Early and the Late Baradostian levels are more likely. If the occupations phases of the Early and the Late Baradostian are chronologically distinct (by hundreds or thousands of years), the supposed contact between them should be interpreted as admixture. In any case, both Baradostian phases share typological and economical features and variability that could probably represent an in situ development from Early to Late Baradostian tradition. Nevertheless, the Mousterian levels appear to be more distinct from the Early Baradostian. This separation of the MP and Early Baradostian levels is suggested by the low density of material at the limit spit (LL), as well as by taphonomic and technotypological differences of both assemblages. Second order refits. The unique aspects of some raw material allowed for the isolation of several groups of ‘second order refits’.5 After observing all material from the Mousterian and Early Baradostian levels, many objects appear to be made on raw material with unique aspects: specific inclusions, rust, striped texture, color, etc. Taking into account the combination of all these aspects of the specific raw material, we could isolate by macroscopic observation many groups of artifacts that likely came from one block of flint. The techno-typological features of the artifacts in each group were considered as well. Numerous objects were refit and classified into 22 groups. However, the second order refits method has less confidence than the refits of débitage but at the same time it is the first step of the refitting analysis. Finally, taking into account the uniformity of the red chert, only the most reliable 15 groups of artifacts were retained and analyzed (Fig. 5). This reliability is confirmed by the unique aspect of these materials (as ‘rusted’ green flint, etc.), the technotypological coherence between the artifacts within each group, and the refits of débitage found intra group (Fig. 6). Blade fragment conjoins and débitage refits. The taphonomic analysis included all blades and elongated flakes with a break length greater than 2.5 cm, which represents 426 pieces. There are 63 fragments from the Late Baradostian (spit ZZ), 246 fragments from the Early Baradostian (spits LLeAA) and 117 fragments from the Mousterian (spits NNeQQ). Overall, 31 pieces were refit with a success rate of 7.3% (Table 2). Blade fragments conjoin and débitage refits demonstrate no connection between the Mousterian and Early Baradostian (Fig. 7). Nevertheless looking at the distribution of the second order refits, the upper levels from the Mousterian and the levels from the Early Baradostian appear connected (Fig. 5) but at the same time the absence of blades or débitage

5 These are pieces with specific inclusions in the raw material or cortex or whose texture or identical color permit them to be refit or considered as coming from the same block of raw material.

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

47

Figure 8. Warwasi, Early Baradostian. Cores: (1) unidirectional rectangular core, the last extracted bladelet with triangular pointed shape; (2) bidirectional prismatic bladelet core; (3) truncated-faceted core on flakes, with burin; (4) pyramidal bidirectional bladelet core; (5) unidirectional bladelet core with faceted platform; (6) burin core on blade, typical for Late Baradostian levels; (7) carinated core.

refits do not allow us to suggest a certain connection between the Mousterian and the Early Baradostian. Between the Early and Late Baradostian were observed mostly second order refits and only one blade fragment refit (Fig. 7). The Early Baradostian assemblage can be considered as relatively well preserved and coherent, which is confirmed by the

techno-typological and economic analyses below. However, the lack of refits between the Mousterian and the Early Baradostian does not necessarily mean that no mechanical inter-level mixture occurred. There are a few bladelets in the Mousterian levels similar to the UP technology that are probably intrusive and resulted from the suggested bioturbation activity. These UP elements have

48

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

Figure 9. Warwasi, Early Baradostian levels. (1e6) Pointed bladelets, with straight or slightly curved profiles, direct bilateral fine retouch; (7) pointed bladelet, with straight profile, unilateral fine retouch; (8) frontal under crest pointed bladelet, inverse retouch; (9e12, 21) Dufour bladelets with inverse retouch, associated with direct retouch (on examples 9 and 11), straight, slightly curved and twisted profiles (12, 21); (13e15) intercalated overshot blades extracted during the bladelet production; (16) naturally pointed bladelets; (17e 20) narrow twisted bladelets from carinated or burin cores typical for the Late Baradostian level. Notice the regularity between the edges and the ridges.

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

49

Blades 40

Blade_bladelets Bladelets

35

30

30 25

20 thick (mm)

width (mm)

Blades Blade_bladelets Bladelets

25

20 15

15

10 10 5

5 0 0

20

40

60

80

100

length(mm)

0 0

5

10

15

20

25

30

35

40

45

width (mm)

Figure 10. Dimensions of the bladeebladelet production from Early Baradostian levels in Warwasi. Left: length and width; right: width and thickness. Notice the continuity between blades and bladelets.

relatively fresh surfaces that differ from the levels of alteration in the Mousterian assemblage. In the upper spits of the Early Baradostian levels, some of the lithics have techno-typological features common for the Late Baradostian. These include twisted regular bladelets (e.g., Fig. 9: 17e19), burin cores (Fig. 8: 6) and carinated cores (Fig. 8: 7), which appear in spits AA, BB and CC and usually have fresh surfaces. Such elements differ from the main bladelet component observed in the Early Baradostian levels and can be considered as linked to the Late Baradostian phase. This is most probably caused by the immediate contact of the levels or mixture between Early and Late Baradostian phases. The faunal study suggests the presence of rodents in the upper levels of Warwasi, and they probably represent recent intrusions. The activities of Ellobius may have reworked part of the deposits soon after deposition and more recently at depths 3e5 m (Turnbull, 1975). An additional argument for the suggested relative integrity of the MP and UP assemblages at Warwasi comes from the fauna. The typical Mousterian fauna, such a Cervus elaphus (e.g., Table 14), is not present in the Baradostian levels. The absence of true refits of artifacts between the Mousterian and Early Baradostian levels at Warwasi is considered here a stronger argument against admixture than the fact that some raw materials are found in both contexts. In summary, the taphonomic analysis demonstrates that the Warwasi sequence is mostly devoid of significant inter-level mixture. This is especially true for the Early Baradostian assemblage from the middle part of the sequence (spits IIeCC), which yielded a coherent lithic production (Fig. 6).

AA BB CC DD EE rectangular

FF

slightly curved and curved

GG

twisted

HH II JJ KK

Techno-typological characterization of the Early Baradostian (spits LLeAA) The Early Baradostian assemblage from Warwasi (spits LLeAA) is characterized by the production of small blades, bladelets and flakes. The study here concerns mostly tools and cores that are 15.6% of diagnostic artifacts (Table 1). The assemblage is composed of bladelet cores, bladelet tools, retouched points, domestic tools (end-scrapers, burins, etc.), and non- or lightly retouched blanks (Table 3). A set of small flakes is associated with truncated-faceted and other types of cores on flakes (e.g., Table 9). From an economical point of view, the assemblage is characterized by a high percentage of fragmented tools. Almost half of the studied tools (45.1%) are fragments. They seem to be well reduced by retouch and secondary exploitation, as are the cores for small flakes. Almost 22% of the studied sample are tools and blanks, which are re-flaked (redébitage process)6 as small cores according different approaches (e.g., Table 9). This particular aspect of the economy is probably due to the small size, 4e6 cm length of the pebbles and blocks, of the raw material and some particular activity using small flakes. Blade and bladelet production. Most significant for the reconstruction of the blank production methods and techniques are the least reduced cores showing traces of core shaping, including crested elements. Six prismatic and three pyramidal bladelet cores show a unidirectional and bidirectional reduction method on relatively small blanks (abandoned core length of 3.5e5.0 cm). Many of these cores blanks have a flat, rectangular shape and show an intention to produce long, rectangular, pointed blanks (Fig. 8: 1). The flaking surface is most often placed on a flat, narrow edge of the blank (Fig. 8: 1). The longitudinal convexity of the flaking surface is preserved by the removal of overshot blades or flakes. A considerable number of overshot products resulting from these bladelet cores attest to this typical UP method and knapping technique (Table 4). Similar pattern of bladelet production, i.e., using deliberate detachment of overshot blades and bladelets (typical for pyramidal cores), is quite common in early UP assemblages in the Levant and in Europe (e.g., Fox, 2003; Tsanova, 2008). These products, as well as the crested elements and a few ‘centering’ flakes (e.g., Inizan et al., 1999), are typical of a unidirectional convergent flaking method. The bladelets

LL 0%

20%

40%

60%

80%

100%

Figure 11. Distribution of the bladelets after the profiles in the Early Baradostian levels in Warwasi. The number of the projected sample is 350 bladelets: 46 are with straight profiles, 182 are slightly curved and curved and 122 are twisted (see SOM Table 2).

6 Re-débitage or reworked flakes and tools: a subsequent knapping operation, performed on a blank coming from the first debitage process, which becomes, in effect, ‘a core’ to be reworked and that may have different objectives: the production of small products: flakes or/and blades (Tsanova, 2008).

50

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

Figure 12. Warwasi. (1, 3, 5, 6) Truncated-faceted cores on flakes; (2, 4) small flakes produced from truncated-faceted cores; (7) burin core on flake, anvil percussion; (8) bladelet core on cortical flake edge.

with rectangular profiles were detached from the center of the flaking surface. Striking platforms are mostly unfaceted; and the preserved platforms on the artifacts show that most often the blades and the bladelets were detached from plain platforms (Table 5). Most of the platforms retain clear traces of the percussion (Table 6). Blade and bladelets detached from plain platforms with direct soft hammer percussion most often form a lip on the platform (Fig. 9: 1) and the edges of the core are abraded and prepared for each blade or bladelet detachment Fig. 8: 1. The angle between the flaking surface and platform is less than 80 (e.g., Pelegrin, 2000). The most common unidirectional volumetric blank production at Warwasi is characterized also by the production of intercalated blades and bladelets7 (Fig. 9: 13e14). Economy of the blanks. Retouched bladelets on a variety of pointed blanks are the most frequent tools in the Early Baradostian levels at

7 This is an extraction of a blade during bladelet detachments so that this intercalated blade has the negatives of the previously detached bladelets.

Warwasi. Blades and bladelets are integrated in one continuous reduction sequence, which includes a large morpho-metrical variety of elongated blanks, blades, small blades, bladelets, and microbladelets (Fig. 10). The smallest bladelets are classified as microbladelets (15.0e20.0 mm long, less than 6.0 mm width and less than 1.5 mm thick). The main objective of blank manufacture is the production of pointed bladelets designated also for production of projectiles. They were recognized for the first time in the eponymous site of Gar-Arjeneh rockshelter and regarded as counterparts to Krems, Font-Yves and El-Wad points found in Shanidar, Warwasi, Yafteh and other Levantine sites (Hole and Flannery, 1967). The blank economy is characterized by selection of bladelets and small blades for projectiles points or other bladelet tools. Flakes were transformed by lateral or bilateral convergent retouch into scrapers or retouched points. The blades and small blades were modified into various types of tools: end-scrapers, burins, etc. (Table 7). Characteristics of the bladelets used as tool blanks. Two types of blades and bladelets were detached: one from the central part of

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

51

Figure 13. Warwasi, Early Baradostian levels. Retouched points: (1, 5e7) on flakes; (2, 3) on elongated blanks; (4) on blade.

the flaking surface (e.g., Fig. 9: 1, 2, 5, 16) and the other from the lateral edge of the core (e.g., Fig. 9: 10, 11). Both were transformed into tools. The edges are parallel or convergent, and they do not always present a good parallelism with the flake scar ridges of the bladelets (e.g., Fig. 9: 1, 3, 4). The parallelism between the edges and the flake scar ridges is one of the characters that differentiate the Early and Late Baradostian bladelet productions. In the supposed zone of contact between the two phases of Baradostian, the ‘intrusive’ bladelets from spits AA and BB show good parallelism between edges and scar ridges (Fig. 9: 17e20). The rectilinear and slightly curved bladelets obtained by the described knapping method predominate all through the Early Baradostian levels (Fig. 11). Twisted bladelets, usually obtained from carinated cores or burins, increase toward

the top of the sequence, and they always outnumber the curved bladelets (Fig. 11). Retouched bladelets. Retouched bladelets appear sporadically in the deeper spit LL and increase toward the top. Most significant for the determination of the assemblage type is the pointed morphology of the retouched blade and bladelets (Fig. 9: 1e7). The second important type of bladelets are lateralized and inversely retouched, i.e., large Dufour bladelets (Fig. 9: 9e12). These two bladelet tool types are generally considered as being projectile armatures (e.g., O’Farell, 1996; Normand et al., 2008). They are considered as evidence of behavioral modernity as they probably reflect part of a broader pattern of ecological extension and subsistence among early Homo sapiens populations (Shea and Sisk, 2010). The various typologies of the bladelets in the

52

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

Figure 14. Yafteh, Early Baradostian levels (below 190 cm). (1) Prismatic blade unidirectional core; (2) unidirectional bladelet core; (3e8) pointed bladelets (Gar-Arjeneh type) representing different stages of manufacture, the largest examples are less retouched, the smaller bladelets are entirely retouched (i.e., 6, 7); (9) retouched bladelet with step fracture used most probably as projectile weapons; (10e14) Dufour bladelets with straight profile, inverse retouch associated with direct retouch (10, 13e14); (15, 18) backed bladelets-rods.

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

53

Figure 15. Yafteh, Early Baradostian levels (below 190 cm). Discs: (1) core on ventral face on flake; (2, 4) burins-core with arch-shaped opposed extremities by long sub-parallel direct retouch; (3) circular-shaped scraper, bifacial low retouch; (5) pre-shaped core or tool with bifacial retouch; (6) circular-shaped scraper, on flake, abrupt sub-parallel retouch.

Warwasi Early Baradostian levels probably correspond not only to projectile points, but also to other kinds of tools. In the second case, they could represent different stages of reduction of the tool by retouch: re-shaping and reuse (Table 8). There is, for example, typological overlap between backed bladelets and retouched bladelets. Most of the bladelet tools are broken and it is impossible to precisely determine their types. Some of them are probably bladelet points, which seem to have similar morpho-technological characteristics as an Arjeneh point. Of 87 bladelet tools, 54 are broken: 27 proximal, 11 medial and 16 distal. On 33 complete bladelets, the shape of the tool allows us to precisely classify the type. The retouch does not significantly affect the initial size and morphology of the blank. Some distal parts of the points are un-retouched or marginally retouched (e.g., Fig. 9: 1, 4, 5, 7). The technology appears to be geared to the production of pointed shapes with straight profiles (e.g., Fig. 8: 1; Fig. 9: 16). Typologically, Gar-Arjeneh points and rectilinear, or slightly curved bladelets with inverse retouch, correspond to the European types of

Font-Yves, Krems points, and Dufour bladelets sensu stricto. The methods and economy of blade-bladelet production are similar to what is seen in the Protoaurignacian from Southwest Europe (Bon, 2002; Bordes, 2006), the Fumanian from North Italy (Broglio et al., 2005), at Krems-Hundssteig in Lower Austria (Teyssandier, 2007), and in the Kozarnikian from northwest Bulgaria (Tsanova, 2008). Comparisons between prismatic blade-bladelets production and ElWad points can be extended to the Early Ahmarian levels from the Levant (Belfer-Cohen and Goring-Morris, 2007). Small reworked cores, flake production, burin tools and cores. Most of the analyzed cores from the Early Baradostian level at Warwasi are a result of secondary flaking on blanks or tools and indicate a clear intention of small flake production and also of elongated flakes for retouched points or scrapers (Table 7). These blanks were re-flaked according to a variety of methods (Table 9). The majority are ‘truncated-faceted’ cores, which are known from the local Mousterian contexts in Iran and in Europe (Dibble and McPherron, 2007) and that were geared toward the production of

54

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64 Arjeneh point

Length and width of the retouched bladelets

El Wad Dufour bladelet

14

Backed bladelet Retouched bladelet

12

Width and thickness of the retouched bladelets

Dufour bladelet Backed bladelet

4,5

Retouched bladelet

4 8 Thickness (mm)

Width (mm)

10

6

4

Arjeneh points El Wad points

5

3,5 3 2,5 2 1,5 1

2

0,5 0

0 0

10

20

30

40

50

0

5

10

15

20

Width (mm)

Length (mm)

Figure 16. Yafteh, Early Baradostian levels (below 190 cm). Comparison of the dimensions of the retouched bladelets; pointed bladelets are larger than Dufour and backed bladelets.

small flakes similar to Levallois flakes (Fig. 8: 3; Fig. 12). Some of the edges of the truncated-faceted cores from Warwasi show edge damage or shiny parts of edges, which is most probably related to traces of use as tools. The mode of exploitation is occasionally similar to Kostienki cores (Fig. 12: 3, 6). It is characterized by the production of small blanks on the flake scar ridges of the blank or tool with inverse truncation used as a flaking platform. Sometimes small flakes are produced on the ventral part of ‘big’ flakes as Kombewa cores. All of these secondary productions of small flakes are a behavior rooted in the Middle Paleolithic. Some of the small cores in Warwasi are exploited on the narrow edge as a burin technique (Fig. 12: 7). There is overlap between the different ways in which pieces were reworked: truncated-faceted with burin (Fig. 8: 3), etc. The truncated-faceted cores from Early Baradostian levels at Warwasi are close to Levallois cores by their morphology and characteristics described below. They are used for obtaining small flakes (Fig. 12: 1e6) and shaped by means of two asymmetrical convex surfaces. Each removal is prepared so that the small flakes have a faceted (N ¼ 76) or dihedral platform (N ¼ 20) like Levallois products. The dorsal flake scar patterns are mostly centripetal (N ¼ 37) or unidirectional (N ¼ 24), as well as bidirectional (N ¼ 14). The flake economy shows that the large and the small reworked flakes have different economic positions. Flakes and elongated flakes have been systematically chosen for points and scrapers (Table 7), while the small flakes are un-retouched with edge damage or probable traces of use (Fig. 12: 2, 4). Almost half of the studied tool blanks are made on flakes (Table 7). There are occasional large flakes (longer than 5 cm), which clearly replicate MP types (scrapers, Levallois flakes). Retouched points, similar to Mousterian points, were made on elongated flakes or blades. They are present throughout the Early Baradostian levels alongside various scraper types (lateral, convergent, etc.). Their technique (direct hard hammer percussion) differs from that of bladelets Table 1 Warwasi, EUP Baradostian levels: representation of the studied lithic material. Techno-typological groups (spits AAeLL) Cores and fragments Debitage Retouched tools Total:

Total number of the collection (after Olszewski, 1993)

Studied number (Tsanova, 2010)

Studied %

449 5984 997 7430

255 459 347 1161

57.0 8.0 35.0 15.6

(direct percussion with a soft organic hammer). The absence of cores and cortical flakes corresponding to the production of these elongated laminar blanks does not permit a reconstruction of the beginning of the reduction sequence. Ten blades are distinct from the main UP bladeebladelet production. They consist of rectilinear, wide blades with faceted platforms and convergent or bidirectional scar patterns. They probably result from an MP reduction method and can be intrusive. A large number of the artifacts have faceted platforms (cf. Table 5). The scar patterns of the blades show that most of them are obtained by unidirectional convergent and bidirectional reduction methods (Table 10). Most of the elongated large flakes with convergent flake scars are clearly linked to the main bladeebladelet production technology. They have the same position as the intercalated blades and were retouched into points as the previously described heavy blades with MP character (Fig. 13). The high frequency of tool fragments with retouch does not allow the precise determination of the blank type: flake or blade (Tables 5 and 10). Levallois cores (longer than 5 cm), as well a few Levallois flakes, are distinct from the small truncated-faceted cores. They differ by their size and may be intrusive in the Early Baradostian levels. The absence of large Levallois flakes and diagnostic products confirm that this technique has already been abandoned in the beginning of the UP. The ‘truncated-faceted’ approach can be considered an economic solution for small flake needs, adapted to the raw material and linked to some of the domestic activities. The production of small flakes continues in the overlying Late Baradostian levels. However, the Mousterian reminders of retouched points and scrapers persist all through the Early Baradostian with bladelet production. The burins (N ¼ 23), cores-tools (N ¼ 32) and burin cores (N ¼ 22) are produced by using the burin blow technique (e.g., Inizan et al.,

Table 2 Warwasi, stratigraphic assemblages and considered numbers of artefacts and fragments for refitting. Spits

Z AAeLL QQeNN Total:

Stratigraphic assemblages

Considered number of artifacts

Number of refitted artifacts

% of refits

Number of successful refits

Late Baradostian Early Baradostian Zagros Mousterian

63 246 117 426

5 23 3 31

7.9 9.3 2.6 7.3

10 46 6 62

Ratio of successful refits.

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64 Table 3 Warwasi, Early Baradostian levels (spits AAeLL): tool typology and technotypological classification. Tool typology and technological groups

Number

End-scrapers Burins Composite tools Borer Retouched blades Truncated tools Retouched bladelets with direct retouch (and Arjeneh points) Retouched bladelets with inverse retouch (Dufour type) Retouched points Side-scrapers Scaled pieces Pieces with few retouch (considered as tools) Divers Core-tools Total retouched tools Blades (from which 45 w/little retouch) Flakes (from which 45 w/little retouch) Total blades and flakes with few retouch Bladeebladelet blanks (from which 7 w/little retouch) Bladelet blanks (from which 14 w/little retouch) Burin spalls Total blades and bladelets blanks Unidirectional cores (from which 7 convergent) Bidirectional cores Centripetal cores Others Total cores Total:

17 23 5 1 23 12 73 14 71 46 4 16 10 32 347 61 106 167 56 216 20 292 128 60 53 14 255 1103

1999), a particular retouch technique for manufacturing the tip of the burin as a tool. Very often this technique was used to obtain small blanks corresponding to burin spalls, so in this way the burins are used as cores. As the tools in the Early Baradostian level at Warwasi are heavily reduced, it was sometimes difficult to evaluate if burins were used as a tool or as cores for bladelets (e.g., Otte and Kozlowski, 2007). In the lower Baradostian levels, the typology of the burin tools is diverse. These are dihedral burin, on break, on truncation or transversal, made mostly on elongated blanks (Table 7). The burin cores and reworked blanks according to the burin technique are produced mostly on flakes (Table 9; Fig. 12: 7). Burin cores producing twisted bladelets, typical of the Late Baradostian levels, appear sporadically in spit CC (Fig. 8: 6) and in greater quantities only in upper spits AA and BB. Taking into account the regularity of the small twisted bladelets and their fresh surfaces, most of them can probably be linked to a zone of contact between the two phases or as representing mixing from the overlying Late Baradostian levels (Fig. 9: 17e20). Warwasi synthesis and discussion Given the absence of refits, there is no evidence for mixing between the Mousterian and the Baradostian levels in Warwasi. Moreover, the low density of artifacts and near absence of Levallois elements in the Baradostian suggests that the main assemblages Table 4 Warwasi, Early Baradostian levels (spits AAeLL): blades products from shaping and re-shaping out the cores. Product Blade Bladeebladelet Bladelet Flake Total:

Crest

Under crest

Overshots

Tablet

7 2

3 1

7 1 17

1

22 (from which 6 crests) 10 (from which 1 ‘under’ crest) 8 11 50

6 6

5

55

(Middle and Upper Paleolithic) are relatively well preserved and are lacking significant mixture. The Early Baradostian assemblage from spits LL to AA is characterized by: - Blade and bladelet production deriving from mostly unidirectional and less often bidirectional reduction methods on prismatic or pyramidal cores. The flaked surfaces (surface de débitage) are oriented on a flat face of the blank or on an edge (Table 3; Fig. 8: 1, 2). There is continued production from small blades toward bladelets confirmed by the intercalated bladese bladelets and with regular overshot removals (Fig. 9: 13e15). - Direct percussion with a soft hammer is suggested by the majority of the striking platform edges, as well as the platforms of the bladelets themselves. - The most typical retouched tools are pointed bladelets with direct bilateral retouch (Arjeneh points) and the rectangular Dufour bladelets with lateral and inverse retouch. - This approach to bladeebladelet production at Warwasi is comparable with the European earliest phases of the UP (Protoaurignacian, Fumanian, Early Kozarnikian, etc.) with a resulting similar morphology of rectilinear or slightly curved bladelets (Fig. 9: 1e12, 16). However, the Early Baradostian of Warwasi has a wider variability of blank production with flexible methods and techniques. - Other components of the Baradostian assemblage from Warwasi could be perceived as having a Mousterian character. Truncated-faceted and diverse types of small cores are associated with a production of small flakes by recycling/reusing some of the tools and flakes (Table 9). Large numbers of retouched points typologically close to Mousterian points are made on various types of blanks (Fig. 13; Table 7). Nevertheless, the Early Baradostian assemblage from Warwasi is typologically characterized as much by UP laminar types (endscrapers, burins, retouched blades and bladelets, bladelet points) as MP types (continuously retouched flakes reminiscent of scrapers, points on elongated and short flakes). The typological heterogeneity of the retouched tools most likely results not only from the high reduction and recycling processes of their initial shapes but also from some particular activities linked to the function of the site (see below). The small truncated-faceted cores and the pointed pieces are an integral part of the Early Baradostian assemblage and do not represent contamination from the underlying Mousterian. These MP elements coexist with the UP bladelet production. The Levallois method has been almost abandoned in the beginning of the UP in Warwasi and replaced by bladelet projectile technology. Flake frequencies decline toward the top of the sequence while retouched and un-retouched bladelet frequencies increase progressively. The various methods of reworking small flakes can be interpreted as an ‘easy’ technological solution resulting from domestic activities and adapted to the nature of the raw material. From an economic point of view, the Early Baradostian presents characteristics typical of European and Levantine transitional assemblages such as the production of tools evoking the Mousterian and the employment of techniques typical of the UP. In this sense, the Early Baradostian may be considered as a ‘transitional’ industry as suggested by Otte and Kozlowski (2004). However, the genuine production of bladelets (direct marginal soft hammer percussion) is unknown in these so-called ‘transitional’ assemblages from the Levant and Eastern Balkans. Bladelets from these types of industries are not standardized and do not form part of an organized reduction sequence, which is distinct from that of blade production (e.g., Boker Tachtit in the Negev: Marks and Kaufman, 1983; Ksar ’Akil, levels XX-XVI: Ohnuma, 1988; Umm el

56

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

Table 5 Warwasi, Early Baradostian levels (spits AAeLL): types of preserved butt for the different groups of blanks. Blank Cortical blade Blade Bladelet Crest blade ‘Under’ crest blade Crest bladelet Burin spall Cortical flake Tablet flake Elongated flake Flake Flake or blade? Total:

Cortical

Plain

Dihedral

Faceted

Punctiform

Shattered

Undetermined

Removed

79 193 7 2 6 11 1

3 2

16 9 1

2 5

1 7

12

4 13

1

2 1

2 3 1 15

1 6 2 24

1

20 58 9 386

1 2

1 1

1 1

2 2 13 3 27

4 15 50 7 103

Tlel in Syria: Boëda and Bonilauri, 2006; Kanal in Turkey: Kuhn et al., 1999; or Bacho Kiro and Temnata in Bulgaria: Tsanova, 2008; etc.). However, in most of the transitional MP/UP assemblages, small blades tending toward bladelets are produced, but they are almost never retouched. Nevertheless, Middle and Upper Paleolithic techno-typological features are known to coexist in Levantine Initial UP assemblages from Tor Sadaf in Jordan, which is characterized by a coexistence of Levallois and bladelet technologies (Fox, 2003). Techno-economic and typological characteristics of the Early Baradostian assemblage from Yafteh (spits below a depth of 190 cm) The Yafteh Early Baradostian assemblage presents a similar bladelet production technology as the Warwasi assemblage but also some typological and techno-economic differences. In particular, the Early Baradostian of Warwasi is comparable with the lower part of the Yafteh sequence as subdivided by Bordes and Shidrang (2009) and that dates to ca. 35 ky 14C BP (see above). The total number of artifacts in this collection is 9905. The presence of all techno-typological groups (cores, flakes, blades and bladelets, debris and chips) suggests that important parts of the reduction process took place on the site. The analyzed portion of the Yale collection consists of 9.2% or 910 lithic artifacts from the lower part of the UP sequence of square 4d. This sample includes retouched tools as well as un-retouched blanks and a large sample of cores (Table 11). Overall, the Yafteh assemblage is very well preserved. In comparison to the Warwasi assemblage, the Yafteh collection is less broken and less altered. The goal of blank production at Yafteh was the production of bladelets and points that were presumably to be used as part of projectile weapons. These bladelets were produced from small pebbles (5e7 cm for the largest and less than 4 cm for the smallest). The rolled nature of the cortex indicates that the raw material was collected from a nearby fluvial context. This raw material is of the

7

4 12 4 38

Total 1 105 241 8 4 10 13 1 6 44 143 26 602

same nature as that of Warwasi: a cryptocrystalline chert that is most often a red-brown or green-gray, opaque, smooth, and homogeneous in grain. Cores. The bladelet cores are mostly pebbles and flakes with unidirectional, and to a lesser extent, bidirectional flaking methods (Table 12). The cores have a prismatic volume according the natural shape of the pebble (N ¼ 80). The shaping of the pebbles is simple: the striking platform is often chosen on a natural surface or managed by a removal of core tablet to create an oblique angle with the flaking surface (Fig. 14: 1, 2). The platform angle is abraded and blunted in preparation for flaking by a tangential blow with a soft, organic hammer. The prismatic volume is shaped by crests on the face of the core opposite removals. The elongated triangular shape of the flaking surface is predetermined by shaping the distal part of the core; its morphology is pyramidal (Fig. 14: 1). This kind of shaping and reshaping of the cores maintains the convexity of the flaking surface. The flaking surfaces are mainly on the wide face and on the wide and the narrow face of the core. Cores on flakes mostly exploit the thickness of the blank following a frontal progression. This approach is called flaking on the edge of blanks (Fig. 14: 2). Some of the cores indicating different phases of reduction show turning system (progression de débitage tournant) of blank production (N ¼ 5) or on a wide face of the blanks (N ¼ 7). When we look at the state of abandonment of 142 cores, 114 present edges very well abraded and ready to be exploited. The cores on the edge of blanks are similar to the burin technique. These two groups are highly diverse and present overlapping techno-typological features. It is not possible to determine the technological position of some of them: core or tool? Clearly the burin cores (N ¼ 7) and the carinated burins (N ¼ 8) typical for the Late Baradostian levels were reduced as cores. However, their association with Early Baradostian levels and the main blade-bladelet production is uncertain. They are almost absent in the basal part of the test pit, and only appear at the top of the lower part of the sequence (spit 9: 189 cm

Table 6 Warwasi, Early Baradostian levels (spits AAeLL): preparation of the butt and employed knapping techniques. Butt type Cortical Plain Dihedral Faceted Punctiform Shattered Other (undet. þ removed) Total:

Presence of lip, abrasion on the knapping edge

Presence of lip, without abrasion

181

19 1 7

2 2

Without lip, with abrasion 1 46

Bulb and traces of hard hammer 74 18 60

Shattered-anvil

1

2 2

1 186

3 27

52

4 156

3

Total 1 321 19 69 4 2 8 424

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

57

Table 7 Warwasi, Early Baradostian levels (spits AAeLL): typology and economy of retouched tools. Tool types End-scrapers Burins Composite tools Perforators Unilateral retouched blades Bilateral retouched blades Troncatures Retouched bladelets Retouched points Side-scrapers Scaled pieces With little retouch Divers Coreetools Total: %

On blade

On blade/ bladelet

On bladelet

Blade or flake

On elongated flake

On flake

Other (undet.)

Total

11 9 1 1 9 11 1 0 5 5 2 1 2 5 63 18.2

0 1 1 0 0 0 1 3 1 0 0 1 0 0 8 2.3

0 0 0 0 0 0 1 84 0 0 0 2 2 1 90 25.9

3 6 1 0 1 1 3 0 31 6 0 1 0 5 58 16.7

2 2 0 0 1 0 3 0 11 11 0 2 1 2 35 10.1

1 4 1 0 0 0 3 0 23 24 2 9 5 18 90 25.9

0 1 1 0 0 0 0 0 0 0 0 0 0 1 3 0.8

17 23 5 1 11 12 12 87 71 46 4 16 10 32 347 99.9

deep), near to the Late Baradostian levels. At the same time, cores on edge blank (N ¼ 10) and tool-burins (N ¼ 29) are distributed through the Early Baradostian levels (spits below 190 cm). The burin cores are made on pebbles or flakes. The resulting rectilinear or slightly curved bladelets are detached from the long, narrow section of the blank. Shaping is basic if not non-existent; the flat striking platform creates an oblique angle with the flaking surface. In order to predetermine the pointed aspect of the bladelets, a partial crest was created in the distal part of the core. Cores for small flakes are sporadic. There is only one truncatedfaceted core and three others made on flakes from the middle of the lower levels. Other possible cores for small flakes are the discs that are discussed below. Some of them are probably core-tools: splintered (piece esquillée) and burin techniques, as well as retouch can represent different phases of use for the discs as cores or/and as tools. Most of the discs demonstrate shaped and very finely retouched blunted edges and that suggests that they could have been used last as tools (Fig. 15: 2e5). Almost half of the blade-bladelet cores are abandoned without any visible reason with abraded and blunted preparation of the knapping angle. This state of abandonment suggests a particular economic behavior of site provisioning of cores. At the same time, the cores in the lower levels are not so reduced. It is the small size of the nodules that gives an impression that the assemblage is reduced. A large proportion of cores are lightly reduced (3e5 products). Non-retouched blades and bladelets, economic aspects of blank production. The largest portion of the sample consists of complete blades and bladelets and proximal fragments. The same pattern is observed as well for the retouched blades and bladelets. The main morpho-metric character of the numerous bladelets is the significant variability of the blanks: from wide and long to very narrow and thin microbladelets (Fig. 16). The length of these slender blanks ranges from 2 to 7 cm. A large quantity of these un-retouched blanks are pointed, a shape that corresponds to the last negatives on the less reduced cores. Only two ‘under’ crests,8 two ‘neo’-crests9 and 18 cortical blades and bladelets were analyzed for this study, representing the beginning of this reduction sequence. The number of intercalated blades and

8 ‘Under’ crest is a term translated from French and it describes a crest blade detached after (below) the first (initial) crested blade. This under crest blade retains the negatives of the initial crested blades (Inizan et al., 1999). 9 ‘Neo’ crest blade: crested blade shaped after the removal of the initial crested blade and during debitage (Inizan et al., 1999).

bladelets shows a similar exploitation of the cores as at the Warwasi Early Baradostian levels (see above): extraction of the largest products during the bladelet production. The predominate platform type is plain. Regarding the technique of blank production, the edges of the core and the platforms of the bladelets in the majority of cases exhibit traces of abrasion and scars typically referable to marginal percussion with a soft hammer. Most of the blades and bladelets show a good regularity between the edges and the flake scar ridges. Their profiles are more often straight and slightly curved (N ¼ 216) but also curved (N ¼ 75) and twisted (N ¼ 73). The twisted bladelets are not necessarily produced on carinated or burin cores like most of the bladelets in the upper levels. Twisted and slightly twisted bladelets can also be made from the cores on edge and prismatic cores as shown by the last negatives of certain cores: prismatic, pyramidal, on edge, and on the burin cores (e.g., Fig. 12: 7, 8). Bladelet tools. The most frequent tools are made on bladelets (Table 13). These are pointed bladelets with direct, bilateral retouch (Arjeneh points), bladelets with alternate retouch (Dufour type), backed bladelets, as well a large number of retouched bladelets which overlap with the other types of bladelet tools. The blanks for the pointed bladelets are rectilinear or slightly curved with lengths varying between 3 and 5 cm (Fig. 14: 3e8). The edges are altered by direct bilateral retouch that is marginal to semi-abrupt (Fig. 14: 6, 7) and can be partial, continuous on the two edges or localized at the point (Fig. 14: 4). Significant variability can be observed between the ‘heavily retouched’ (Fig. 14: 6, 7) and ‘slightly retouched’ (Fig. 14: 3e5) points that is probably the result of the different phases of their manufacture, utilization and reutilization. This high techno-economic variability may explain why some of the points can be classified as the following: for the ‘very’ retouched as Arjeneh points (Fig. 14: 6e8); for slightly retouched as El-Wad type (Fig. 14: 4, 5); and for the finely retouched, like Font-Yves points (Fig. 14: 7). Also some of the bladelets with direct or alternate retouch are truncated on the distal part of the edge, such as the Protoaurignacian bladelets from Fumane cave (Fig. 14: 10). The Dufour bladelets are more robust and less regular in the lower part of the sequence (Fig. 14: 10e14). They are manufactured on rectilinear or slightly curved bladelets that are part of the same production sequence as the points. The rectilinear Dufour bladelets are transformed by an alternate retouch: direct on the left edge and inverse on the right edge. The shape of the Dufour bladelet is sometimes triangular, symmetrical or asymmetrical, and slightly truncated by direct retouched in the distal part (Fig. 14: 10, 13). The

58

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

Table 8 Warwasi, Early Baradostian levels (spits AAeLL): typology and retouch type of the bladelet tools. Tool type

Bilateral, direct, semi-abrupt or fine, pointed

Bilateral, direct, semi-abrupt

10

3 10

Arjeneh points Probable Arjeneh points? Dufour Backed 1 edge Backed 2 edges Semi-lune backed Retouched bladelets 1 edge Retouched bladelets 2 edges With little retouch Total:

Marginal

Direct and inverse

Inverse

1

5 1 1 6 1 1 15

3

11 1 28

10

Direct abrupt and semi-abrupt

Dufour bladelets can have rectangular shape with a truncated or broken distal section (Fig. 14: 11, 12). Some of the bladelet tools show evidence of use as projectile weapons with complex fractures classified according to Fischer et al. (1984). Two-step terminating bending fractures were identified (Fig. 14: 9) along with two possible spin-off fractures. At the same time, useewear analysis on pointed bladelets of El-Wad or Font-Yves type demonstrate that they have been used not only for hunting but also for cutting and other domestic activities (Willams, 1997; Normand et al., 2008). We believe that heavily retouched points may have been re-used for domestic activities. Backed bladelets called ‘rods’, which also exist sporadically at Warwasi (only three items in spits AA, BB and FF), represent at Yafteh a series of non-pointed blanks with parallel edges (Fig. 14: 16; Table 13). The retouch, unlike the Arjeneh points and Dufour bladelets, is abrupt, bilateral, and rectifies the edges. The retouched bladelets are mostly proximal and medial fragments with continuous bilateral or unilateral retouch. They have a similar characteristic (profiles, morphology and preparation of the platform) as the un-retouched and other types of bladelets tools. Some of them are probably point fragments because of their similar sizes. Blade and flake tools. Blades and flakes were used to manufacture different kinds of burins (on truncation, angle, and dihedral). These burins are separated from the burin cores (see above) based on their dimensions, morphology and the number of spalls. Angle burins are mostly made on broken blades, while the dihedral burins are manufactured on flakes (Table 13). End-scrapers from the lower layers of square 4d are made on various blanks. Almost half are made on large rectilinear or slightly curved blades, and small blades, which are most of the time interior blades from bladelet production, under crests, etc. The other half is manufactured on short primary or secondary flakes, which are byproducts of blade production. Apparently the large elements and by-products were selected for this type of domestic tool. The front of the long end-scraper can be thick or thin and shaped and reshaped by semi-abrupt retouch. The morphologies of the front are most often arched. Three end-scrapers on blades are clearly Aurignacian type with notable scaled and very regular retouch (e.g., Otte and Kozlowski, 2007). End-scrapers on flakes show a tendency

1 11 3 3 19

4

11

4

11

Total

14 10 15 5 4 2 17 15 5 87

toward round shaped tools with semi-abrupt continuous retouch (Fig. 15: 6) with some of them approaching carinated type with large fronts. The retouch of the round shaped end-scrapers can be long, bladelet-like and, in this case, considered to be a core for microbladelets. We do not totally exclude the possibility of a bladelet production on end-scrapers but there is no evidence for systematic production and use of such a small bladelet in the lower levels at Yafteh cave. These round end-scrapers increase in the upper levels in Yafteh but on smaller flakes than in the lower levels. They can be considered as a particular style that is not very common in the other earliest Upper Paleolithic assemblages. Discs are another particular artifact with a round shape. They are mostly found in the lower levels and in the lower part of the upper levels. They are shaped by bilateral flat retouch. Some of the visible blanks are primary flakes (Fig. 15: 4, 5). The blanks can be thick (>1 cm) or thin (<1 cm). Regarding the sharpness and the regularity of the retouched edges, four of the discs can be considered as tools: end-scrapers, which correspond to the previously described, round shaped end-scraper. The majority of the discs are possible cores, which does not exclude the possibility of also being used as a tool (Fig. 15: 2, 4). Looking at the thickness and the convexity of the larger discs, they probably were used as cores on the ventral face on a flake (Fig. 15: 1). Typologically these cores on the narrow edge of flake correspond to end-scraper-burin (Fig. 15: 2, 4). The shattered edges of some discs suggest that they have been reworked using anvil percussion (Fig. 15: 1, 2). The discs were interpreted also as pre-shaped blanks for use as cores (Bordes and Shidrang, 2009). Nevertheless, cores or tools, most of these artifacts show shaped and carefully retouched edges. These mostly blunted edges suggest a possible last utilization as a tool (Fig. 15: 2e5). Such a rounded variety of tools were found in small number in Unit D from Dzudzuana cave in southern Caucasus (Bar-Yosef et al., 2011). Techno-economic remarks From the techno-economic point of view, all bladelets are obtained by a similar reduction method on small pebbles. Furthermore, for the manufacture of points the selected bladelets are longer and wider than the blanks of the Dufour bladelets (Fig. 16). However, their sizes clearly overlap and suggest a continuous production

Table 9 Warwasi, Early Baradostian levels (spits AAeLL): techno-typological classification of the small cores, their blanks and the ‘re-debited’ small flakes. Blanks

Elongated flakes Flakes Flake or blade? Total:

Small re-debited flakes

Small cores resulting from re-debitage Truncatedfaceted

Burin

Kombewa type

Kostienki type

Splinteredanvil

Kombewa type

Truncatedfaceted

Splinteredanvil

1 14 6 21

4 9 9 22

1 10 1 12

e 1 e 1

1 8 4 13

2 31 e 33

e 14 e 14

1 e e 1

Undetermined

Other flakes

Total

4 21 4 29

37 46 25 108

51 154 49 254

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

59

Table 10 Warwasi, Early Baradostian levels (spits AAeLL): scar pattern of the different type of blanks. Blank type Blades Bladeebladelets Bladelets Elongated flakes Flakes Flakes or blades? Total:

Unidirectional

Unidirectional parallel

Convergent

Bidirectional

41 17 92 23 68 21 262

8 7 53 1 1 1 71

48 33 133 13 23 9 259

17 2 5 3 15 4 46

between blades and bladelets. Some of the blanks of the points can be considered as blades or rather small blades (4 cm long). The existence of varied types of burins indicates a distinct production for the small twisted bladelets typical for the upper levels. Carinated burins and burin cores for small twisted bladelets and microbladelets are not integral or systematically produced in the lower levels (below 190 cm). The microbladelets are straight or twisted, produced from the same cores as the bladelets. Microbladelets were also selected for tool manufacture (Fig. 16). The cores indicate that many of the last removals (15e45 mm long blades and bladelets) were twisted or curved (Fig. 17). Yafteh synthesis The stratigraphic position of the Early Yafteh assemblage is situated in 1 m thick sediments from 190 cm to the bottom of the test pit at 290 cm. The raw material consists of small pebbles 4e7 cm long of homogeneous grained chert collected from a fluvial context. The methodology of blank production follows the natural convexity of the pebbles. The pre-shaping of the core is simple: opening a knapping platform on one of the extremities of the pebble and/or preshaping a postero-lateral crest on the core. Very often the triangular shape of the bladelets is a result of shaping the distal extremity of the core. They are thin (1e3.5 mm) and slightly curved. The most regular of them, with edges parallel to scar ridges, are marginally retouched. In spite of the different states of the retouched bladelets and the resulting typological diversity, two major types of bladelet tools emerge: first, Arjeneh points on rectilinear or slightly curved bladelets with direct bilateral retouch, marginal to semi-abrupt, and second, Dufour bladelets shaped with alternate retouch (direct on the left edge and inverse on the right edge). The bladelet blanks of the different types of tools are obtained by continued exploitation of the blade-bladelet cores following the same method (unidirectional Table 11 Yafteh cave, lower levels (below 190 cm): techno-typological classification of the studied lithic sample from square 4d.

Retouched tools

Non-retouched blanks

Cores

Total:

Technological groups

Studied number

Studied %

Blades Bladelets Microbladelets Flakes Other Subtotal of the tools: Blades Bladelets Microbladelets Subtotal of the non-retouched blanks: Unidirectional Bidirectional Others Subtotal of the cores:

71 265 53 37 6 432 94 194 78 366

7.5 28.2 5.6 3.9 0.6 45.9 10.0 20.7 8.3 38.9

116 21 5 142 940

Total number of the assemblage is 9905 artifacts, 9.2% were studied here.

12.3 2.3 0.5 15.1 99.9

Centripetal

Transversal

Undetermined

4 1 1 6 34 4 50

6 3 19 4 12 7 51

2 13 15

Others

Non applicable

Total

1

2

1 3 4 1 10

3

127 63 307 55 175 63 790

5 16 26

prismatic, frontal, on the wide and narrow face, semi-turning) and technique (direct soft hammer percussion). Some of the bladelets show complex bending fractures resulting, most probably, from use as projectiles. Another techno-typological feature of the Yafteh Early Baradostian assemblage is the quantity and diversity of the burin tools and cores. Two types of end-scrapers, long ones on blades and short, round ones on flakes, characterize the assemblage. There is considerable overlap between the different types of tools and the modalities of their re-shaping. The disc types have similar shaping by bilateral flat retouch but various (e.g., tool, core or toolecore) technological states. From an economic point of view, we suggest that blanks for the bladelet tools were obtained from an integral and coherent reduction sequence. For the manufacture of the Dufour bladelets, the selected blanks are slightly shorter and narrower than the Arjeneh bladelets (Fig. 16). Many tools and cores are made on cortical or other flakes that are by-products of blade production. The upper and lower assemblages from Yafteh, corresponding to the Early and Late Baradostian traditions, appear similar in their blank production and economy, as well as the mode of tool manufacture. However, they present important differences that bear witness to progressive behavioral changes: a) the Early Baradostian assemblage (lower part of the sequence) is an industry dominated by the production of bladelets, which are larger than those of the Late Baradostian. Dufour bladelets are less standardized and regular compared with the upper part of the sequence; b) in the Late Baradostian the production of small bladelets becomes more normalized and the majority comes from burins. These pointed and twisted bladelets have a fine inverse retouch on the concave edge, which is often associated with a direct retouch on the convex edge. Comparison of Early Baradostian Warwasi and Yafteh assemblages Inter-stratigraphic and taphonomic comparison The assemblage from the lower layers of Yafteh cave is better preserved and less fragmented than the Early Baradostian assemblage from Warwasi. We noticed a different pattern in the density distribution of both assemblages: in the Warwasi assemblage there is an increase in the quantity and the state of preservation of Table 12 Yafteh cave, lower levels (below 190 cm): blanks of the cores and blank production method. Blank cores type/ Debitage method On block On pebble On flake On slab On tablet On blade Undetermined Total:

Unidirectional

Bidirectional

9 65 18 3 1 4 16 116

4 10 2 1

3 20

Undetermined

6

6

Total 13 81 20 4 1 4 19 141

60

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

Table 13 Yafteh cave, lower levels (below 190 cm): typology of the tools and their blank types. Tool types End-scraper End-scrapereburin Burin Burin with truncature Dihedral burin Angle burin Retouched blade Blade with little retouch Arjeneh point Arjeneh point? El-Wad point Backed point Dufour bladelet Backed piece Truncation bladelet Retouched bladelet Bladelet with little retouch Side-scraper Discs Scaled piece Total:

Blade Bladelet 16 2

Micro Flake Block Undet. Total bladelet

1

14 2 1 10 1

4 2 5 24 6

1

1

8

50 5 3 1 58 38 1 84 24

1 1

1 1

1 1

25 4 19 3

1

31 2 3 5 14 7 24 6 51 7 3 1 83 43 1 104 35

associated with hearths, preserved fauna, and burned artifacts, suggesting relatively intensive occupation of the sites while the context of the assemblage from Warwasi is characterized by the absence of hearths and by poorly preserved fauna. Based on those archeological features, Hole and Flannery have proposed different functions for the sites: Yafteh is interpreted as a base camp, while Warwasi as a butchering station (Hole and Flannery, 1967). In both sites the density of the material reduces toward the base of the sequence. Comparing the geo-archeological context and the thickness of the deposits, it has to be noted that the Zagros sequences (Yafteh, Warwasi) have a more extensive sedimentation (more than 1 m deposits) than the supposed contemporaneous bladelet assemblages from Europe. The earliest bladelet UP assemblages from Europe are usually found in the context of short archeological deposits: the layers are not thicker than 20e30 cm (e.g., Early Kozarnikian: Tsanova, 2008; Protoaurignacian: Le BrunRicalens et al., 2005; etc.).

Techno-economic and typological comparison 1 1 71

1 8 265

53

37

1 1

5

2 9 1 432

artifacts toward the top of the Early Baradostian levels. The Mousterian is absent in Yafteh cave. In Warwasi rockshelter, the earliest UP levels do not have a hiatus with the underlying Mousterian levels. We suggest that some Levallois elements exist in the lower level of the Early Baradostian assemblage from Warwasi but they progressively disappear toward the top of these levels. However, tools similar to Mousterian tools (retouched points and scrapers) existed and have been produced through the Early Baradostian occupation. Stratigraphically, the earliest UP assemblages from Warwasi and Yafteh are underlying and connected to Late Baradostian levels, which are noticeable for the production of small twisted bladelets on burins. The assemblage from Yafteh is Table 14 Comparison of the preserved fauna species at Warwasi (after Turnbull, 1975) and Yafteh (after Mashkour et al., 2009). Fauna Equus hemionus Bos primigenius, wild cattle Capra hircus aegagrus Ovis orientalis, wild sheep Gazella sp. Cervus elaphus Sus scrofa Canis lupus, wolf Vulpes vulpes Panthera pardus, leopard Felis sp. Mustela foina Meles meles Meriones vinogradovi Meriones libycus Meriones cf. persiscus Meriones vinogradovi Allactaga sp., jerboa Ellobius cf. fuscocapillus Ellobius cf. lutesence Chionomys cf. nivalis Ochotona cf. rufescens Lepus cf. capensis Turtle Fish Bird

Warwasi x x x x x

Yafteh x x x x x x

x x x x x x x x x x x x

x x

x x x x x x x

The raw material used in both sites is the same chert but collected from different sources: in primary deposits for Warwasi and in fluvial deposits for Yafteh. The faunal assemblage of Yafteh is rich in fish remains (Otte et al., 2007), which indicates the nearest fluvial context as a source of food and raw material. Concerning the integrity of the Early Baradostian from Warwasi (spits LLeAA), we can note that it is coherent from a technotypological and economic point of view, as well as being distinct from the underlying Mousterian. However, continuity between the Mousterian and the Early Baradostian is confirmed by the use of blank production similar to the Levallois method, as well as by the production of various blanks (blade, bladelets, flakes) retouched into various tools with UP character (bladelet tools, end-scrapers, burins), and with MP character (retouched points and scrapers). This Mousterian tradition is much less expressed at Yafteh. There are some retouched points comparable with those from Warwasi (Otte, 2012) and side-scrapers but in much smaller quantity than in Warwasi. The flakes in Yafteh are mostly used as blanks for endscrapers, burins and cores. It seems in Warwasi that flakes have a more significant place in the economy because they have been used frequently for various types of tools and as cores for small flakes. The truncated-faceted cores of the Warwasi assemblage are similar to Levallois conception and bear a strong MP character. At the same time, this is an ‘easy’ economic solution for obtaining small flakes. Some of the discs of the Early Baradostian Yafteh assemblage have a similar role, as cores for small flakes. Many discs and comparable truncated-faceted artifacts were found during the 2005e2008 excavations (Otte, 2012), and they suggest close comparisons with the Lower Baradostian assemblage from Warwasi. The round scrapers in Yafteh seem to be a particular style that is absent in Warwasi. Apart from the ‘archaic’ character, the assemblage from Warwasi strongly resembles that of Yafteh. The principle points on which these two Baradostian assemblages coincide are blank production (prismatic or pyramidal unidirectional cores and cores on the edge of blanks), the blank production economy (continuity in the production of blades and large bladelets), and the blank production techniques (direct marginal percussion with a soft hammer). The most significant comparison element is found in the concept of lithic armatures that are predetermined by the initial form of a blank with convergent edges, hence products that are already pointed upon removal. The regularity of the bladelets increases in both sites from the bottom toward the top of the sequence. Additionally, the bladelets are considerably more standardized in the Late Baradostian.

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64 25

Straight negative Slightly curved Curved Twisted

Width (mm)

20

15

61

of the fauna: mid-size herbivores (Equus and Bos) in Warwasi and small size of herbivores in Yafteh cave. This perhaps suggests seasonal occupation and exploitation of horses at the site of Warwasi. The differences between the fauna assemblages can potentially be related to the distinct functions of both sites: a base camp (Yafteh) versus a butchering station (Warwasi). Paleoenvironment

10

5

0 0

10

20

30

40

50

Length (mm)

Figure 17. Yafteh, Early Baradostian levels (below 190 cm). Length and width of the last bladelet removals on the cores; profile types of the last extracted bladelets.

The absence of radiocarbon dates for Warwasi as well the precise fauna and other analysis does not allow us, for the moment, to discuss the contemporary of the earliest UP occupations in Warwasi and Yafteh. Both assemblages mark the beginning of a new technological tradition in the central Zagros, the Early Baradostian. The new series of radiocarbon dates indicate that earlier occupation of Yafteh cave took place around 35,450
600 14C BP. This interpretation of a single chronological signal around 33,500 14C BP is based on a Bayesian model (Otte et al., 2007, 2011-uncalibrated). The dates themselves are between ca. 36,000 and 30,000 BP. Faunal assemblages and subsistence behavior The bone assemblages from Early Baradostian levels at Warwasi and Yafteh present significant taphonomic differences. The Warwasi fauna assemblage is poorly preserved and consists primarily of teeth. The larger mammalian fauna is restricted mostly to Equus hemionus, Capra aegagrus, Ovis orientalis, and Bos primigenius (Turnbull, 1975). Most likely the shelter served as a lookout point for hunters tracking onagers and was used briefly but frequently over the years (ibid.). Only certain parts of the hunted animals were brought back to the site. The inhabitants were probably able to selectively target prime age adult onagers (Bakken, 2000). The function of the Warwasi rockshelter is explained as a butchering station (Hole and Flannery, 1967). The bone assemblage from the lower levels of Yafteh cave contain a high number of identified faunal remains, with better preservation and variability than the Warwasi assemblage (Table 14). Here, herbivore teeth are extremely fragmented, and the most preserved skeletal parts are the extremities: phalanges (Mashkour et al., 2009). At Yafteh, horse is completely absent and the most typical fauna accumulated by humans are small herbivores such as ovi-caprids and gazelles. The frequency of young animals at Yafteh suggests a spring/summer occupation (ibid.). Birds, turtles and fish probably related to human subsistence are typical for the Baradostian in Yafteh (ibid.), but they are not present in Warwasi. According to Hole and Flannery (1967), Yafteh cave served as a base camp. At Warwasi, except for the presence of C. elaphus, which is typical for the Mousterian, there are more similarities between the Mousterian and Early Baradostian fauna composition than between the Early Baradostian levels at Warwasi and Yafteh. This statement is confirmed by Turnbull (1975): “geologically older Mousterian and Baradostian hemiones from Warwasi, which inhabited higher, colder, drier land, apparently possessed larger teeth.”. The main differences between the sites are the presence of horse and the size

Two earlier glacial episodes characterize the transitional Middle to Upper Paleolithic period. The first occurs during the Mousterian at ca. 75e70 ky BP (MIS 4) and the second between 59 and 24 ky BP (MIS 3), regarded as a moderately warm interval (e.g., Klein, 1999), separated by an interstadial at 60 ky BP (Wright, 1961; Skinner, 1965 cited in; Shidrang, 2007). The last pleniglacial cycle is characterized by a colder and dry climate and corresponds to the appearance of the Early Baradostian and the bladelet technology between 40 and 33 ky 14C BP. The paleoenvironment during this period appears semiarid with a small quantity of trees and bushes (Roustaei et al., 2004). At Warwasi, the large size of the molars of the Mousterian and Baradostian E. hemionus suggests that they inhabited higher, colder, drier lands, while the rodents indicate that Kermanshah valley had a sparse bush cover, as today (Turnbull, 1975). According to the species distribution of Yafteh fauna, the cave might have been surrounded by several ecological niches: steppe lowland, the piedmont, and probably forested spots (Mashkour et al., 2009). According to the faunal and paleoenvironmental analysis, the Warwasi Early Baradostian assemblage appears to be associated with different paleoenvironment conditions than those recorded for Yafteh cave. These differences may be observed through the lithic technology as adaptations to specific paleoenvironmental conditions. The Early Baradostian assemblage from Warwasi is by its recorded behavior, technology, and economy close to the Middle Paleolithic assemblages. The various pointed blanks (flakes, blades and bladelets) can be used for hunting and domestic tools, and they may appear to be adapted to the size of the exploited fauna mostly represented by horse. The assemblage from Yafteh has a mostly Upper Paleolithic character, and the variability of blank production is restricted mainly to bladelets and blades. The projectile bladelet technology and the Arjeneh points seem more adapted to the exploited fauna represented by small sized herbivores. Inter-regional comparisons Northwest of Warwasi and Yafteh, Shanidar cave contained a Baradostian assemblage with comparatively late dates of 34 ky 14C BP (e.g., Otte and Kozlowski, 2007). In the Caucasus, the earliest UP is marked by assemblages from Dzudzuana cave (Unit D) and Ortvale Klde rockshelter (Layer 4) that are comparable with the Zagros assemblages with similar bladelet production, finely retouched bladelets as well as a rounded variety of end-scrapers, endscrapers-burin on flakes and blades (Bar-Yosef et al., 2011). One of the differences with Baradostian assemblages is the absence of inversely retouched bladelets (Dufour). The Caucasian earliest UP assemblages, like those from the Zagros, are characterized by numerous burins. These technological and stylistically common features suggest a spread of this tradition toward the North. The original radiocarbon dates place the beginning of the UP from Dzudzuana and Ortvale Klde at ca. 34e32 ky 14C BP (Bar-Yosef et al., 2006, 2011) with the oldest possible boundary between the MP and UP at around 38 ky 14C BP (Adler et al., 2008). Moreover, the industry of Unit D from Dzudzuana cave resembles the early UP assemblages from Mezmaiskaya cave dated by radiocarbon and electron spin resonance to the same period (35e34 ky 14C BP) (Adler et al., 2006; Golovanova et al., 2006).

62

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

Discussion and conclusion Thinking about the origin of this bladelet technology that characterizes AMH in Europe (Protoaurignacian) and in the Levant (Ahmarian), the earliest UP Warwasi assemblage can be seen as linked to the local Mousterian technology but with the addition and the development of a new element: projectile bladelet technology. The Early Baradostian from Warwasi seems to represent a behavioral and typologically dependent continuity with the underlying Mousterian. This continuity in the Warwasi sequence is found also in the distribution of faunal species. The persistence of MP features such a pointed flakes and blades similar to Mousterian points, scrapers, and truncated-faceted artifacts could be interpreted as resulting from a particular set of activities linked to the function of the site and perhaps seasonal occupation. This supposition could explain the differences with the Yafteh Baradostian assemblage and the coexistence of the UP bladelet technology with the distinct socalled ‘MP’ components. On a regional scale, Warwasi and Yafteh are local phases of one bladelet technology tradition indicating probably different levels of variability of the Early Baradostian. The new chronological data from Yafteh cave indicates an intermediate chronological position of the Early Baradostian levels, partly overlapping with the Levantine Ahmarian tradition and preceding most of the 14C ages associated with Levantine Aurignacian (Otte et al., 2011; Zwyns et al., 2012). Keeping in mind the absence of precise 14C ages and any other possibilities to date the assemblages from Warwasi, some chronological scenarios can be hypothesized. A first proposal is that Baradostian assemblages from both sites are contemporaneous and the differences they present result from different types of site occupations (long versus shortterm occupation, seasonal occupation, different activities, mobility patterns and exploitation of the environment). A second scenario is that the Yafteh assemblage represents a later, more ‘modern’ phase of the Early Baradostian, while the Warwasi assemblage is an older, more ‘archaic’ phase of Baradostian representing the ‘transition’ or the technological transformation toward the UP expressed in the coexistence of various pointed tools, which were used as hunting weapons. Bladelet projectile technology is an element associated par excellence with behavioral modernity. Moreover, personal ornaments, such as perforated canines and fossil shells, hematite pendants and the bone industry reinforce the observed modern and developed character of the Early Baradostian in Yafteh. The Warwasi UP assemblage did not yield such personal ornaments. The beginning of the Iranian Zagros UP is marked, as in the Levant and in Europe, by the appearance of new projectile bladelet technology. We note the inertia of MP technological behaviors at Warwasi paralleling the emergence of traits characteristic of the UP such as bladelet production. Concerning the previous work on the Baradostian collections: 1. This work agrees with the former hypotheses proposed by Olszewski (1999, 2001) and supported by Otte and Kozlowski (2007). The Warwasi UP assemblage presents a probable in situ development from the MP techno-typological and economical elements such as truncated-faceted pieces, retouched points and scrapers. This opens the possibility that these behaviors could have been transmitted over time from the MP into the EUP, and they could also result from particular activities and site function. 2. Bordes and Shidrang (2009) observe distinct parallelism between the earliest UP sequences from the Zagros and Levant, but this is only for the initial bladelet UP phases (Early Ahmarian and Early Baradostian, followed by industries of twisted bladelets). The Baradostian tradition follows a different path than those

from the Levant or Western Europe. The EUP blade industries from the Levant (Boker Tachtit, Ksar ’Akil levels XXV-XXI, etc.) lack comparable evidence in the central Zagros. The UP industries with ‘carinated’ cores or/and Dufour bladelets with inverse retouch are confused very often with the Aurignacian sensu stricto. Such elements were found in assemblages from the Caucasus region (Golovanova et al., 2009; Bar-Yosef et al., 2011), in the eastern Balkans (Sirakov et al., 2007; Tsanova, 2008) but they are not directly comparable with the Western European ‘classical’ Aurignacian (Aurignacien ancien). The term ‘Aurignacian’ as created in Western Europe is sometimes loosely applied to assemblages distant in time and space (e.g., Golovanova et al., 2009). It is still employed for archeological entities in the Levant and elsewhere, and its use is problematic (Belfer-Cohen and GoringMorris, 2012). Various assemblages from the beginning of the UP (e.g., EUP assemblages) from Europe and Western Asia called ‘Aurignacian’ have different techno-typological and economical characteristics (e.g., Tsanova, 2008; Teyssandier et al., 2010; BelferCohen and Goring-Morris, 2012). Consequently in this paper the discussion concerns the origin of the Baradostian bladelet technology rather than the Aurignacian. The variability in lithic production is a reflection of cultural traditions, but also of dynamic development: testing, adapting and using new technological solutions to better accommodate and exploit the environment, etc. (e.g., Phillips and Saca, 2003). This variability might have been stimulated by the arrival of new ideas through inter-regional (inter-cultural) contacts of human groups. In this light, the persistence of Mousterian features in the Early Baradostian may suggest cultural admixture/acculturation of the local Mousterian technology with a progressive replacement by a bladelet technology, presumably made by modern humans who moved into the region. Others may see that the Mousterian character is rather an expression of economical behavior linked to particular activities and the site function. In the southern Levant, the UP and modern projectile technology developed in situ from the MP (e.g., Bar-Yosef and Kuhn, 1999;  Copeland, 2003; Marks, 2003; Skrdla, 2003). In the central Zagros, the link between the Middle and the Upper Paleolithic is based mostly on typology and the economy of the retouched tools (e.g., Olszewski and Dibble, 2006). Therefore, taking into account that only a few sites were excavated and well documented in the Zagros, we suggest that there is no clear local origin for the UP blade and bladelets industries. Recently, while sorting through the bone assemblage from Warwasi in the Field Museum of Natural History in Chicago, an unpublished human tooth was discovered from the lower Baradostian levels H-I-J (Braidwood et al., 1961). It was previously described in 1960 by Dr. Albert Dahlberg, dental anthropologist from the University of Chicago (Personal communication, F. Hole). The tooth is a lower left second molar, undoubtedly from a modern human (Personal communication, J.-J Hublin and R. Martin). The tooth is in preparation for future analysis. The absence of radiocarbon dates for Warwasi and the lack of antecedent blade technology indicate, for the moment, that the Iranian Zagros cannot be argued to be the source of bladelet technology and cultural modernity. However, if we suppose that the Early Baradostian assemblage of Warwasi is from the boundary of the Middle to Upper Paleolithic (based on its stratigraphic position and typology), it appears to be more connected to the new socioeconomic relation typical for the UP and consequently representing an acceleration phase of cultural intermixing or acculturation of the Mousterian. This Mousterian character of the Early Baradostian of Warwasi also can be seen as due to the nature of the activities on the site and the consumed fauna. The makers of the UP Baradostian

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

were almost certainly moderns. The Early Baradostian phase at Yafteh, with full UP development, appears for the moment to be the oldest dated UP associated with AMH in the Zagros at around 35 ky 14 C BP (Otte et al., 2007). Synchronous cultural development to the Early Baradostian tradition and the bladelet projectile technology spread by AHM are the Ahmarian in the Levant, the Early Kozarnikian from the Eastern Balkans, the Protoaurignacian, the Fumanian from Southwest Europe, and Krems-Hundssteig in Central Europe (Tsanova et al., 2012). These techno-complexes, which represent relatively homogeneous lithic assemblages and technology, have a high potential to establish the most accurate picture of the appearance of the early modern human populations in Europe, their behavioral variability and cultural contacts. Acknowledgements First of all, I would like to thank the Fyssen Foundation for financing this project. I am grateful to H. Dibble, F. Hole, D. Olszewski, J.-G. Bordes, O. Bar-Yosef, N. Zwyns, E. Spinapolice, R. Iovita, S. Shidrang, M. Otte, Y. Demidenko and V. Aldeias. I would like to acknowledge S.P. McPherron for patiently reviewing the manuscript many times, for discussions and for advice. My special thanks to the editor and the reviewers. They have offered suggestions that improved considerably some interpretations and the entire paper. I am very grateful to J. Phillips from the Anthropology Department of the Field Museum in Chicago for the final review and discussion on this paper. Thank you to R. Martin, W. Parkinson and W. Simpson from The Field Museum in Chicago. I am pleasured to thank the America for Bulgaria Foundation and the Anthropology Department at the Field Museum of Natural History in Chicago, and D. Ivanova. Finally, I would like to express my gratitude to J.-J. Hublin, the Max Planck Society, and the Department of Human Evolution in the Max Planck Institute for Evolutionary Anthropology (Leipzig). Appendix A. Supplementary material Supplementary material related to this article can be found at http://dx.doi.org/10.1016/j.jhevol.2013.04.005. References Adler, D.S., Bar-Oz, G., Belfer-Cohen, A., Bar-Yosef, O., 2006. Ahead of the game: Middle and Upper Paleolithic hunting behaviours in the southern Caucasus. Curr. Anthropol. 47, 89e118. Adler, D.S., Bar-Yosef, O., Belfer-Cohen, A., Tushabramashvili, N., Boaretto, E., Mercier, N., Vallada, H., Rink, W.J., 2008. Dating the demise: Neandertal extinction and the establishment of modern humans in the southern Caucasus. J. Hum. Evol. 55, 817e833. Bakken, D., 2000. Hunting strategies of Late Pleistocene Zarzian population from Palegawra Cave, Iraq and Warwasi rock shelter, Iran. In: Mashkour, M., Choyke, A.M., Buitenhuis, H., Poplin, F. (Eds.), 2000. Archaeozoology of the Near East IVA, vol. 32. ARC-Publicate, Groningen, pp. 11e17. Bar-Yosef, O., Kuhn, S.L., 1999. The big deal about blades: laminar technologies and human evolution. Am. Anthropol. 101, 1e17. Bar-Yosef, O., Belfer-Cohen, A., Adler, D.S., 2006. The implications of the Middlee Upper Paleolithic chronological boundary in the Caucasus to Eurasian Prehistory. Anthropologie XLIV, 49e60. Bar-Yosef, O., Belfer-Cohen, A., Mesheviliani, T., Jakeli, N., Bar-Oz, G., Boaretto, E., Golberg, P., Kvavadze, E., Matskevich, Z., 2011. Dzudzuana: an Upper Paleolithic cave site in the Caucasus foothills (Georgia). Antiquity 85, 331e349. Belfer-Cohen, A., Goring-Morris, A.N., 2007. From the beginning: Levantine Upper Palaeolithic cultural change and continuity. In: Mellars, P., Boyle, K., BarYosef, O., Stringer, C. (Eds.), Rethinking the Human Revolution. McDonald Institute for Archaeological Research, Cambridge, pp. 199e206. Belfer-Cohen, A., Goring-Morris, A.N., 2012. The Earlier Upper Palaeolithic: a view from the southern Levant. In: Otte, M., Shidrang, S., Flas, D. (Eds.), 2012. The Aurignacian of Yafteh Cave and its Context (2005e2008 Excavations), vol. 132. ERAUL, Liège, pp. 127e132. Berillon, G., Asgari, A., Khaneghah, A.A., Antoine, P., Bahain, J.-J., Chevrier, B., Zeitoun, V., Aminzadeh, N., Beheshti, M., Chanzanagh, H.E., Nochadi, S., 2007. Discovery of new open-air Paleolithic localities in Central Alborz, Northern Iran. J. Hum. Evol. 52, 380e387.

63

Berillon, G., Asgari, A., Khaneghah, A.A., Antoine, P., Bahain, J.-J., Chevrier, B., Zeitoun, V., Aminzadeh, N., Beheshti, M., Chanzanagh, H.E., Nochadi, S., 2009. Late Pleistocene prehistory in central Alborz: preliminary results of the French and Iranian paleoanthropological program 2006 on the excavation of Garm Roud 2 (Amol, Mazandaran). In: Otte, M., Biglari, F., Jaubert, J. (Eds.), Iran Paleolithic. BAR International Series, vol. 1968, pp. 101e109. Oxford. Biglari, F., 2004. The preliminary observations on Middle Paleolithic raw material procurement and usage in the Kermanshah plain, the case of Do-Ashkalt cave. In: Stollner, T., Slotta, R., Vatandoust, A. (Eds.), Persian Antiques Splendor, Mining Crafts and Archeology in Ancient Iran. Deutsches Bergbau-Museum Bochum, Bochum, pp. 130e138. Biglari, F., 2007. Approvisionnement et utilisation des matières premières au Paléolithique moyen dans la plaine de Kermanshah (Iran): le cas de la Grotte DoAshkaft. In: Moncel, M.H., Moigne, A., Arzarello, M., Peretto, C. (Eds.), Aires d’Approvisionnement en Matières Premières et Aires d’Approvisionnement en Ressources Alimentaires Approche Intégrée des Comportements. Workshop 23, XV Congrès UISPP, Lisbonne, vol. 5, pp. 227e239. BAR International Series 1725. Boëda, E., Bonilauri, S., 2006. The intermediate Paleolithic: the first bladelet production 40,000 years ago. Anthropologie XLIV, 75e92. Bon, F., 2002. L’Aurignacien entre Mer et Océan. Réflexion sur l’Unité des Phases Anciennes de l’Aurignacien dans le Sud de la France. In: Mémoire de la Société Préhistorique Française, vol. 29. Paris. Bordes, J.-G., 2000. La séquence aurignacienne de Caminade revisitée: l’apport des raccords d’intérêt stratigraphique. Paléo 12, 387e412. Bordes, J.-G., 2006. News from the West: a reevaluation of the classical Aurignacian sequence of the Perigord. In: Bar-Yosef, O., Zilhão, J. (Eds.), Towards a Definition of the Aurignacian. Trabalhos de Arqueologia, vol. 45, pp. 147e171. Lisboa. Bordes, J.-G., Shidrang, S., 2009. La séquence Baradostienne de Yafteh (Khorramabad, Lorestan, Iran). In: Otte, M., Biglari, F., Jaubert, J. (Eds.), Iran Paleolithic. BAR International Series, vol. 1968, pp. 85e100. Oxford. Braidwood, R.J., Howe, B., 1960. Prehistoric Investigation in Iraqi Kurdestan. In: Oriental Institute Studies in Ancient Oriental Civilisation, vol. 31. University of Chicago Press, Chicago. Braidwood, R.J., Howe, B., Reed, C.A., 1961. The Iranian prehistoric project. Science 133, 2008e2010. Broglio, A., Bertola, S., De Stefani, M., Marini, D., Lemorini, C., Rossetti, P., 2005. La production lamellaire et les armatures lamellaires de l’Aurignacien ancien de la grotte de Fumane (Monts Lessini, Vénétie). In: Le Brun-Ricalens, F., Bordes, J.-G., Bon, F. (Eds.), Productions Lamellaires Attribuées à l’Aurignacien: Chaînes Opératoires et Perspectives Technoculturelles. XIVe Congrès de l’UISPP, Liège 2e8 Septembre 2001. Imprimerie Fr. Faber, Luxembourg, pp. 415e436. Cahen, D., Karlin, C., Keeley, L.H., Van Noten, F., 1980. Méthode d’analyse technique, spatiale et fonctionnelle d’ensembles lithiques. Heliniul XX, 209e259. Conard, N., Ghasidian, E., Heydary, S., Zeidee, M., 2006. Report on the 2005 Survey of the Tübingen-Iranian Stone Age Research Project in the Provinces of Esfahan, Fars and Kohgiluyeh-Boyerahmad. University of Tübingen, Tübingen. Coon, C.S., 1951. Cave Exploration in Iran, 1949. The University Museum, Philadelphia. Copeland, L., 2003. The Levantine Upper Palaeolithic: a commentary on contributions to the Philadelphia symposium. In: Goring-Morris, A.N., Belfer-Cohen, A. (Eds.), More Than Meets the Eye. Studies on Upper Palaeolithic Diversity in the Near East. Oxbow Books, Oxford, pp. 242e248. Debénath, A., Dibble, H.L., 1994. Handbook of Paleolithic Typology. In: Lower and Middle Paleolithic of Europe, vol. 1. The University Museum, Philadelphia. d’Errico, F., 2003. The invisible frontier. A multiple species model for the origin of behavioral modernity. Evol. Anthropol. 12, 188e202. Dibble, H.L., Holdaway, S.J., 1993. The Middle Paleolithic industries of Warwasi. In: Olszewski, D.I., Dibble, H.L. (Eds.), The Paleolithic Prehistory of the Zagrose Taurus. The University Museum, Philadelphia, pp. 75e99. Dibble, H.L., McPherron, S.P., 2007. Truncated-faceted pieces: hafting modification, retouch, or cores? In: McPherron, S.P. (Ed.), Tools versus Cores. Alternative Approaches to Stone Tool Analysis. Cambridge Scholars Publishing, Cambridge, pp. 75e90. Fischer, F., Veming Hansen, P., Rasmussen, P., 1984. Macro and micro wear traces on lithic projectile points: experimental results and prehistoric examples. J. Danish Archaeol. 3, 19e46. Fox, J.R., 2003. The Tor Sadaf lithic assemblages: a technological study of the Early Upper Palaeolithic in the Wadi al-Hasa. In: Goring-Morris, A.N., Belfer-Cohen, A. (Eds.), More Than Meets the Eye. Studies on Upper Palaeolithic Diversity in the Near East. Oxbow Books, Oxford, pp. 80e94. Garrod, D., 1930. The Palaeolithic of southern Kurdistan: excavations in the caves of Zarzi and Hazar Merd. Bull. Am. Sch. Prehist. Res. 6, 8e43. Garrod, D., 1937. The Near East as a gateway of prehistoric migration. In: MacCurdy, G.G. (Ed.), Early Man. JB Lippincott Company, Philadelphia, pp. 33e40. Garrod, D.A., 1953. The relations between southwest Asia and Europe in the later Palaeolithic age. J. World Hist. 1, 13e38. Golovanova, L.V., Cleghorn, N., Doronichev, V.B., Hoffecker, J.F., Burr, G.S., Sulergizkiy, L.D., 2006. The early Upper Paleolithic in the northern Caucasus (new data from Mezmanskaya cave, 1997 excavation). Euras. Prehist. 4, 43e78. Golovanova, L.V., Doronichev, V., Cleghorn, N., 2009. Getting back to basics: a response to Otte ‘Comments on Mezmaiskaya’. Euras. Prehist. 5, 131e136. Goring-Morris, A.N., Belfer-Cohen, A. (Eds.), 2003. More Than Meets the Eye. Studies on Upper Palaeolithic Diversity in the Near East. Oxbow Books, Oxford. Henshilwood, C.S., Marean, C.W., 2003. The origin of modern human behaviour. Curr. Anthropol. 44, 627e651.

64

T. Tsanova / Journal of Human Evolution 65 (2013) 39e64

Hofman, J.L., Enloe, J.G.E., 1992. Piecing Together the Past: Applications of Refitting Studies in Archaeology. In: BAR International Series, vol. 578. Oxford. Hole, F., Flannery, V., 1967. The prehistory of southwest Iran: a preliminary report. Proc. Prehist. Soc. 33, 147e206. Hole, F., 2012. The 1960s excavations at Yafteh Cave. In: Otte, M., Shidrang, S., Flas, D. (Eds.), 2012. The Aurignacian of Yafteh Cave and its Context (2005e2008 Excavations), vol. 132. ERAUL, Liège, pp. 11e25. Inizan, M.-L., Reduron-Balliger, M., Roche, H., Tixier, J., 1995. Technologie de la Pierre Taillée. CREP, Meudon. Inizan, M.-L., Reduron-Balliger, M., Roche, H., Tixier, J., 1999. Technology and Terminology of Knapped Stone. CREP, Nanterre. Jaubert, J., Biglari, F., Bordes, J.-G., Bruxelles, L., Mourre, V., Shidrang, S., Naderi, R., Alipour, S., 2004. New Research on Paleolithic of Iran: Preliminary Report of 2004 IranianeFrench Joint Mission, vol. 1. Université Bordeaux, Bordeaux. Klein, R.G., 1999. The Human Career: Human Biological and Cultural Origins. University of Chicago Press, Chicago. Klein, R.G., 2000. Archeology and the evolution of human behaviour. Evol. Anthropol. 9, 17e36. Kozlowski, J.K., Otte, M., 2000. The formation of the Aurignacian in Europe. J. Anthropol. Res. 56, 513e534. Kuhn, S.L., Stiner, M.C., Güleç, E., 1999. Initial Upper Palaeolithic in south-central Turkey and its regional context: a preliminary report. Antiquity 73, 505e 517. Le Brun-Ricalens, F., Bordes, J.-G., Bon, F. (Eds.), 2005. Productions Lamellaires Attribuées à l’Aurignacien: Chaînes Opératoires et Perspectives Technoculturelles. XIVe Congrès de l’UISPP, Liège 2e8 Septembre 2001. Imprimerie Fr. Faber, Luxembourg. Marks, A., Kaufman, D., 1983. Boker Tachtit: the artefacts. In: Marks, A.E. (Ed.), Prehistory and Paleoenvironments in the Central Nequev, Israel. The Avdat/Aqev Area, Part 3, vol. III. Southern Methodist University Press, Dallas, pp. 69e126. Marks, A.E., 2003. Reflections on the Levantine Upper Palaeolithic studies: past and present. In: Goring-Morris, A.N., Belfer-Cohen, A. (Eds.), More Than Meets the Eye. Studies on Upper Palaeolithic Diversity in the Near East. Oxbow Books, Oxford, pp. 249e264. Mashkour, M., Radu, V., Mohaseb, A., Hashemi, N., Otte, M., Shidrang, S., 2009. The Upper Paleolithic faunal remains from Yafteh cave (Central Zagros), 2005 campaign. A preliminary study. In: Otte, M., Biglari, F., Jaubert, J. (Eds.), Iran Paleolithic. BAR International Series, vol. 1968, pp. 73e84. Oxford. McBrearty, S., Brooks, S.A., 2000. The revolution that wasn’t: a new interpretation of the origin of modern human behavior. J. Hum. Evol. 39, 453e563. Melé, M., Javed, A., Pybus, M., Zalloua, P., Haber, M., Comas, D., Netea, M.G., Balanovsky, O., Balanovska, E., Jin, L., Yang, Y., Pitchappan, R.M., Arunkumar, G., Parida, L., Calafell, F., Bertranpetit, J., Genographic Consortium, 2012. Recombination gives a new insight in the effective population size and the history of the Old World human population. Mol. Biol. Evol. 29, 25e30. Mellars, P., 2006. Archeology and the dispersal of modern humans in Europe: Deconstructing the “Aurignacian”. Evol. Anthropol. 15, 167e182. Normand, C., O’Farrell, M., Rios Garaizar, J., 2008. Quelle(s) utilisation(s) pour les productions lamellaires de l’Aurignacien archaïque? Quelques données et réflexions à partir des exemplaires de la grotte d’Isturitz (Pyrénées-Atlantiques, France). In: Pétillon, J.-M., Dias-Meirinho, M.-H., Cattelain, P., Honegger, M., Normand, C., Valdeyronn, N. (Eds.), Recherches sur les Armatures de Projectiles du Paléolithique Supérieur au Néolithique. Actes du Colloque C83, XVe Congrès l’UISPP, Lisbonne, 4e9 sept. 2006. Palethnologie, vol. 1, pp. 7e46. O’Farrell, M., 1996. Ph.D. Dissertation. Approche technologique et fonctionnelle des pointes de La Gravette, vol. 1. Université de Bordeaux. Ohnuma, K., 1988. Ksar ’Akil, Lebanon. A Technological Study of the Earlier Upper Palaeolithic Levels of Ksar ’Akil, Levels XXV-XIV. In: BAR International Series, vol. 426. Oxford. Olszewski, D.I., Dibble, H.L., 1994. The Zagros Aurignacian. Curr. Anthropol. 35, 68e75. Olszewski, D.I., 1993. The Late Baradostian occupation at Warwasi rockshelter, Iran. In: Olszewski, D.I., Dibble, H.L. (Eds.), The Paleolithic Prehistory of the Zagrose Taurus. The University Museum, Philadelphia, pp. 187e206. Olszewski, D.I., 1999. The Early Upper Palaeolithic in the Zagros Mountains. In: Davies, W., Charles, R. (Eds.), Doroty Garrod and the Progress of the Palaeolithic. Studies in the Prehistoric Archaeology of the Near East and Europe. Oxbow Book, Oxford, pp. 167e180. Olszewski, D.I., 2001. Rumination on the Early Upper Paleolithic in the Zagros Mountains. In: Hays, A.M., Thacker, T.P. (Eds.), Questioning the Answers: Resolving Fundamental Problems of the Early Upper Paleolithic. BAR International Series, vol. 1005, pp. 79e89. Oxford. Olszewski, D.I., Dibble, H.L., 2006. To be or not to be Aurignacian: the Zagros Upper Paleolithic. In: Bar-Yosef, O., Zilhao, J. (Eds.), Towards a Definition of the Aurignacian. Trabalhos de Arqueologia, vol. 45. Instituto Português de Arqueologia, Lisboa, pp. 355e373. Otte, M., 2012. Les vestiges techniques en pierre taillée. In: Otte, M., Shidrang, S., Flas, D. (Eds.), 2012. The Aurignacian of Yafteh Cave and its Context (2005e2008 Excavations), vol. 132. ERAUL, Liège, pp. 51e115. Otte, M., Kozlowski, J.K., 2004. La place du Baradostien dans l’origine du Paléolithique supérieur d’Eurasie. L’Anthropologie 79, 395e406. Otte, M., Kozlowski, J.K., 2007. L’Aurignacien de Zagros, vol. 118. ERAUL, Liège. Otte, M., Biglari, F., Flas, D., Shidrang, S., Zwyns, N., Mashkour, M., Naderi, R., Mohaseb, A., Hashemi, N., Darvish, J., Radu, V., 2007. New research on the Aurignacian in the Zagros Region: test excavation at Yafteh Cave, Lorestan, Iran. Antiquity 81, 82e96.

Otte, M., Shidrang, S., Zwyns, N., Flas, D., 2011. New radiocarbon dates for the Zagros Aurignacian from Yafteh cave, Iran. J. Hum. Evol. 61, 340e346. Pelegrin, J., 1995. Technologie Lithique: Le Châtelperronien de Roc-de-Combe (Lot) et de La Côte (Dordogne). CNRS, Paris. Pelegrin, J., 2000. Les techniques de débitage laminaire au Tardiglaciaire: critères de diagnose et quelques réflexions. In: Valentin, B., Bodu, P., Christensen, M. (Eds.), L’Europe Centrale et Septentrionale au Tardiglaciaire. Actes de la Table-Ronde de Nemours,1997. Mémoires du Musée de Préhistoire d’Ile de France, vol. 7, pp. 73e86. Petraglia, M.D., 1992. Stone artifact refitting and formation process at the abri Dufaure, an upper Palaeolithic site in south-west France. In: Hofman, J.L., Enloe, J.G. (Eds.), Piecing Together the Past: Application of Refitting Studies in Archaeology. BAR International Series, vol. 578, pp. 163e178. Oxford. Phillips, J.L., Saca, I.N., 2003. Variability and change in the Early Upper Palaeolithic of the Levant. In: Goring-Morris, A.N., Belfer-Cohen, A. (Eds.), More Than Meets the Eye. Studies on Upper Palaeolithic Diversity in the Near East. Oxbow Books, Oxford, pp. 95e105. Rosenberg, M., 1985. Iran. Report on the 1978 sondage at Eshkaft-e Gavi, vol. 23, pp. 51e62. Roustaei, H., Vahdati Nasab, H., Biglarie, F., Heydari, S., Clark, G.A., Lindly, J.M., 2004. Recent Paleolithic surveys in Luristan. Curr. Anthropol. 45, 692e707. Schroeder, B., 2007. Truncated-faceted pieces from Jerf al-Ajla. In: McPherron, S.P. (Ed.), Tools versus Cores. Alternative Approaches to Stone Tool Analysis. Cambridge Scholars Publishing, Cambridge, pp. 17e41. Scott, J.E., Marean, C.W., 2009. Paleolithic hominin remains from Eshkaft-e Gavi (southern Zagros Mountains, Iran): description, affinities, and evidence for butchery. J. Hum. Evol. 57, 248e259. Shea, J.J., Sisk, M.L., 2010. Projectile technology and Homo sapiens dispersal from Africa to Western Eurasia. Paleoanthropology, 100e122. Shidrang, S., 2007. The Early Upper Paleolithic Lithic Assemblages from F15 Test Pit (2005), Yafteh Cave, Iran. A Typo-technological Study. Ph.D. Dissertation. Universita degli Studi di Ferrara. Sirakov, N., Tsanova, T., Sirakova, S., Taneva, S., Krumov, I., Dimitrova, I., Kovatcheva, N., 2007. Un nouveau faciès lamellaire du début du Paléolithique supérieur dans les Balkans. Paléo 19, 131e144. Skinner, J., 1965. The Flake Industries of Southest Asia: A Typological Study. Ph.D. Dissertation. Columbia University.  Skrdla, P., 2003. Comparison of Boker Tachtit and Stránská Skála MP/UP transitional industries. Mitekufat Haeven. J. Isr. Prehist. Soc. 33, 37e73. Smith, P.E.L., 1986. Paleolithic Archaeology in Iran. The University Museum, Philadelphia. Solecki, R.L., 1998. Archaeological survey of caves in northern Iraq. Int. J. Kurdish Stud. 12, 1e70. Solecki, R.S., 1955. Shanidar Cave, a Paleolithic Site in Northern Iraq. Smithsonian Institutional Annual Report 1954, pp. 603e635. Solecki, R.S., 1958. The 1956e1957 season at Shanidar, Iraq: a preliminary statement. Sumer 14, 1104e1108. Solecki, R.S., Solecki, R.L., 1993. The pointed tools from the Mousterian occupations of Shanidar Cave, northern Iraq. In: Olszewski, D., Dibble, H.L. (Eds.), The Paleolithic Prehistory of the ZagroseTaurus. University Museum Press, Philadelphia, pp. 119e146. Soriano, S., 2007. Recensions. Paléorient 33, 187e190. Teyssandier, N., 2007. En Route vers l’Ouest. Les Débuts de l’Aurignacien en Europe. In: BAR International Series, vol. 1638. Oxford. Teyssandier, N., Bon, F., Bordes, J.-G., 2010. Within projectile range. Some thoughts on the appearance of the Aurignacian in Europe. J. Anthropol. Res. 66, 209e229. Tixier, J., 1978. Méthodes pour l’étude des outillages lithiques. Notice sur les travaux scientifiques, vol. X. Ph.D. Dissertation, Université de Paris. Tsanova, T., 2008. Les Débuts du Paléolithique Supérieur dans l’Est des Balkans. Réflexion à Partir de l’Étude Taphonomique et Techno-Économique des Ensembles Lithiques de Bacho Kiro (Couche 11), Temnata (Couches VI et 4) et Kozarnika (Niveau VII). In: BAR International Series, vol. 1752. Oxford. Tsanova, T., Bordes, J.-G., 2003. Contribution au débat sur l’origine de l’Aurignacien: principaux résultats d’une étude technologique de l’industrie lithique de la couche 11 de Bacho Kiro. In: Tsonev, T., Kokelj, E.M. (Eds.), The Humanized Mineral World: Towards Social and Symbolic Evaluation of Prehistoric Technologies in South Eastern Europe. ESF Workshop Sofia, 3e6 September, 2003, vol. 103. ERAUL, Liège, pp. 41e50. Tsanova, T., Zwyns, N., Eizenberg, L., Teyssandier, N., Le Brun-Ricalens, F., Otte, M., 2012. Le plus petit dénominateur commun: réflexion sur la variabilité des ensembles lamellaires du Paléolithique supérieur ancien d’Eurasie. Un bilan autour des exemples de Kozarnika (Est des Balkans) et Yafteh (Zagros). L’Anthropologie 116, 469e509. Tsanova, T., 2010. Perspectives pour le début de Paléolithique supérieur dans le Zagros iranien et sur l’apparition de la modernité culturelle. Approche taphonomique et caractérisation techno-économique de deux ensembles lithiques de Warwasi et Yafteh (Iran), Annales Fyssen 26, Paris, pp. 108e125. Turnbull, F.P., 1975. The mammalian fauna of Warwasi rock shelter, west-central Iran. Fieldiana Geol. 33, 141e155. Willams, J.K., 1997. Lithic Use-wear Analysis of Early Ahmarian El-Wad Points: An Example from Southern Jordan. Ph.D. Dissertation. University of Tulsa. Wright Jr., H.R., 1961. Pleistocene glaciation in Kurdistan. Eiszeitalt. Gegenw. 12, 131e164. Zwyns, N., Flas, D., Shidrang, S., Otte, M., 2012. Les fouilles 2005e2008 à Yafteh et la chronologie radiocarbon. In: Otte, M., Shidrang, S., Flas, D. (Eds.), 2012. The Aurignacian of Yafteh Cave and its Context (2005e2008 Excavations), vol. 132. ERAUL, Liège, pp. 41e49.