Neanderthal settlement of the Central Balkans during MIS 5: Evidence from Pešturina Cave, Serbia

Quaternary International 610 (2022) 1–19 Contents lists available at ScienceDirect Quaternary International journal homepage: www.elsevier.com/locate/quaint Neanderthal settlement of the Central Balkans during MIS 5: Evidence from Pešturina Cave, Serbia Dušan Mihailović a, *, Stefan Milošević a, Bonnie A.B. Blackwell b, c, 1, Norbert Mercier d, Susan M. Mentzer e, Christopher E. Miller e, f, g, Mike W. Morley h, Katarina Bogićević i, Dragana Đurić j, Jelena Marković k, Bojana Mihailović k, Sofija Dragosavac a, Senka Plavšić a, Anne R. Skinner b, c, Iffath I.C. Chaity c, Yiwen E.W. Huang c, Seimi Chu c, Draženko Nenadić i, Predrag Radović a, l, Joshua Lindal m, Mirjana Roksandic m, n a Department of Archaeology, Faculty of Philosophy, University of Belgrade, 18-20 Čika Ljubina, 11000, Belgrade, Serbia Department of Chemistry, Williams College, Williamstown, MA, 01267-2692, USA c RFK Science Research Institute, Glenwood Landing, NY, 11547-0866, USA d Institut de Recherche sur les Archéomatériaux, UMR 5060 CNRS - Université Bordeaux Montaigne, Centre de Recherche en Physique Appliquée à l’Archéologie (CRP2A), Maison de l’archéologie, 33607, PESSAC Cedex, France e Senckenberg Centre for Human Evolution and Palaeoenvironment, University of Tübingen, 72070, Tübingen, Germany f Institute for Archaeological Sciences, University of Tübingen, 72070, Tübingen, Germany g SFF Centre for Early Sapiens Behaviour (SapienCE), University of Bergen, 5007, Bergen, Norway h College of Humanities, Arts & Social Sciences, Flinders University, GPO Box 2100, Adelaide, Australia i Department of Palaeontology, Faculty of Mining and Geology, University of Belgrade, Kamenička 6, P.O. Box 227, 11 000, Belgrade, Serbia j Natural History Museum, 51 Njegoševa, 11000, Belgrade, Serbia k National Museum of Belgrade, 1a Republic Square, 11000, Belgrade, Serbia l National Museum Kraljevo, 2 Trg Svetog Save, 36000, Kraljevo, Serbia m Department of Anthropology, University of Manitoba, 432 Fletcher Argue Building, 15 Chancellor Circle, Winnipeg, Manitoba, R3T 2N2, Canada n Department of Anthropology, University of Winnipeg, 515 Portage Avenue, Winnipeg, Manitoba, R3B 2E9, Canada b A R T I C L E I N F O A B S T R A C T Keywords: Central Balkans Middle Paleolithic Charentian MIS 5 Neanderthal behavior Recent research in the southern Central Balkans has resulted in the discovery of the first Middle Paleolithic sites in this region. Systematic excavations of Velika and Mala Balanica, and Pešturina (southern Serbia) revealed assemblages of Middle Paleolithic artifacts associated with hominin fossils and animal bones. This paper focuses on Pešturina Layer 4, radiometrically and biostratigraphically dated to Marine Isotope Stage (MIS) 5, which yielded traces of temporary hunting camps. The remains of large ungulate prey are associated with predominantly Quina-type artifacts made of quartz. Artifacts from Pešturina Cave have no parallels at Mousterian sites in the Balkans but are rather similar to the Central European Charentian, which demonstrates that this cultural unit was widespread during MIS 5, not only in the southern Pannonian Basin but also in the Central Balkans. The position of the site – on the outskirts of the known spread of the Quina model of techno-economic behavior during MIS 5 – raises several questions related to population movements, residential mobility, and technological variability in the early Middle Paleolithic of Central and Southeast Europe. 1. Introduction 5e and the somewhat cooler periods of MIS 5d-5a. Several rich open-air sites, caves, and rockshelters in the region suggest that during MIS 5 Neanderthals were engaged in systematic hunting and (likely) scavenging of megafauna (Patou-Mathis, 2000, 2006, 2006; Gaudzinski, 2004; Dusseldorp, 2009). Numerous non-standardized tools of Research at Middle Paleolithic sites in Central Europe has demonstrated that this area provided favorable conditions for settlement, both during the Last Interglacial Marine (Oxygen) Isotope Stage (MIS, or OIS) * Corresponding author. E-mail address: dmihailo@f.bg.ac.rs (D. Mihailović). 1 Deceased https://doi.org/10.1016/j.quaint.2021.09.003 Received 30 May 2021; Received in revised form 10 August 2021; Accepted 6 September 2021 Available online 10 September 2021 1040-6182/© 2021 Elsevier Ltd and INQUA. All rights reserved. D. Mihailović et al. Quaternary International 610 (2022) 1–19 microlithic dimensions made on pebbles were collected at the Taubachian sites to the north, in the Pannonian Basin (Borel et al., 2017), while the Charentian sites in the southern parts of the basin are characterized by Quina sidescrapers and artifacts produced via Quina method (Simek and Smith, 1997; Mester and Moncel, 2006; Banda and Karavanić, 2019). Neanderthal fossils are known from several Central European sites dated to this period, especially Krapina, with its large hominin collection (Janković et al., 2016). In contrast to Central Europe, only a few pre-MIS 4 sites were recorded south of the Sava and Danube Rivers. The site of Zobište in northern Bosnia was dated to MIS 5a to MIS 4 (Montet-White et al., 1986; Baumler, 1988). In Bulgaria, Kozarnika Cave contained materials correlated with MIS 6 (Guadelli et al., 2005; Tillier et al., 2017). Meanwhile, some coastal sites, including Crvena Stijena in Montenegro, and Theopetra in Greece, record thick combustion deposits testifying to intensive settlement (Karkanas et al., 2015; Whallon, 2017). Until recently, however, no sites dated to MIS 5 were known from the interior Balkans, and the explanations for the perceived absence of MIS 5 sites ranged from insufficient research, a lack of mineral resources suitable for knapping, to competition with carnivores (Churchill, 2014; Milošević, 2020). Recent field surveys and excavations of Paleolithic sites in Serbia have changed this picture significantly. Hominin remains have been recovered from the deepest stratum in Mala Balanica Cave in Sićevo, dated to older than 400 ka (Roksandic et al., 2011; Rink et al., 2013; Skinner et al., 2016). The upper layers (2a-2c) of Mala Balanica and the lower layers of the neighboring Velika Balanica (3a-3c) contained numerous artifacts, bones, and fireplaces dated to MIS 9–7 (Mihailović and Bogićević, 2016), and provided initial insight into the different aspects of the Lower to Middle Paleolithic transition in the Balkans. The findings from Velika and Mala Balanica could not be linked to the much later materials from the other Middle Paleolithic sites in Serbia (Mihailović, 2014) until that gap was filled (to an extent) by the excavations and systematic dating of the nearby Pešturina Cave, which revealed MIS 5 materials. Dated to MIS 5c (Blackwell et al., 2014), Pešturina Layer 4b contained Neanderthal fossils (Radović et al., 2019; Lindal et al., 2020), lithic artifacts (Mihailović and Milošević, 2012), and numerous remains of Pleistocene fauna (Milošević, 2020), including a cave bear bone with sequential incisions likely made intentionally by humans (Majkić et al., 2018). Research at Pešturina has enabled us to examine in more detail the technological and subsistence behavior of the Central Balkans Neanderthal communities during MIS 5. In this paper, we present the results of excavations, dating, and analyses of material culture and faunal remains from Pešturina Layer 4 to understand the stratigraphy and chronology of the site, and to reconstruct Neanderthal activities within the cave. 2. Material and methods 2.1. The cave and the context Pešturina Cave is located on the eastern edge of the Niš Basin, on the western slopes of Suva Mountain (Fig. 1A). The Niš Basin is bordered on the north by the slopes of the Svrljig Mountains, and on the south by the slopes of the Suva Mountain. The main communication routes linking the southern Balkan Peninsula and the Pannonian Basin cross this area. The cave was discovered in 2006 during a systematic archaeological survey of the Niš Basin which preserves many Pleistocene terraces, with numerous caves on its periphery and bordering cliffs. The 22-m-deep cave is located near the top of a limestone hill, overlooking the eastern Niš Basin and the short valley cut through by the Jelašnica River. The mouth of the cave (15 m wide and 3.5 m high) is located 330 m above sea level and faces west, towards the Nišava River. The initial excavations of the cave did not indicate that it contained rich remains, as a 2.5 × 1.5 m test pit, excavated to a depth of 1.3 m yielded only a dozen Paleolithic artifacts. In 2010, when the test pit was expanded, Pešturina proved to have substantial deposits. Systematic excavations revealed that the upper layers contained Upper and Middle Paleolithic artifacts. Layer 2, located below the Holocene sediment (Layer 1), yielded about a hundred artifacts attributed to the Gravettian, while the assemblage from Layer 3 was roughly attributed to the Denticulate Mousterian (Mihailović and Milošević, 2012). Unfortunately, it soon became apparent that Layer 3 was significantly disturbed by erosion, bioturbation, and recent anthropogenic activities, causing inconsistent dates ranging from 70 ka to 38 ka (Blackwell et al., 2014; Alex et al., 2019). Therefore, in this paper, we focus on the assemblage which originated from the secure stratigraphic context of Layer 4, which is presented here in detail. Excavation proceeded in squares and quadrants (50 × 50 cm), with the recording of all in situ findings. All sediment was dry screened using a 3 mm mesh, while samples from specific contexts and subsamples from individual squares were wet-screened on the bank of the Nišava River, using a 2 mm mesh. From 2010 to 2018, 20 m2 were excavated to a depth of about 4 m. 2.2. Methods of micromorphological analysis In order to further document the depositional and post-depositional Fig. 1. The geographical position of Pešturina in the Central Balkans (A); the ground plan and excavated area showing marked positions of samples for OSL and ESR dating (B). 2 D. Mihailović et al. Quaternary International 610 (2022) 1–19 processes, field observations of the sedimentary sequence were coupled with the archaeological micromorphology. Micromorphological samples were collected in 2019 from the laminar carbonate, Layers 4a-4c, and weathered limestone boulders located near the cave mouth. The oriented blocks of sediment were indurated with a mixture of polyester resin and styrene catalyzed with MEKP. The blocks were then sliced and processed into 6 × 9 cm petrographic thin sections. The thin sections were observed under low and high magnification with a petrographic microscope, and the description followed the guidelines listed in Stoops (2003). Major element distribution maps were produced from thin sections and cut block faces under vacuum using a Bruker M4 Tornado micro-x-ray fluorescence instrument. associated sediment samples were analyzed with (NAA for U, Th, and K (e.g., Blackwell et al., 2016). Volumetrically and time-averaged sedimentary dose rates and time-averaged cosmic dose rates were calculated for each tooth, based upon total station data, in situ measurements, photographs, profiles, local geological and geomorphological analyses, and ramped box models for the time-averaging (e.g., see Deely et al., 2011; Blackwell et al., 2019). Mean ages were calculated by inversely weighting them by their errors with Isoplot v. 4.2 for each tooth, and then for the whole layer, if appropriate. 2.4. Methods of archaeological investigations Several sediment samples were taken from the cave (at least one from every layer and sublayer) and the remains of small vertebrates were extracted from these. A total of more than 150 remains of herpetofauna and small mammals were found in Layer 4 and preliminarily determined. All species of these tiny vertebrates still live today, often near the site. In order to examine Neanderthal subsistence at the site, we examined the taphonomy of the large mammalian remains, lithic production sequence and technological characteristics and the spatial distribution of artifacts and fauna. Taphonomy of large mammal remains was studied following Domínguez-Rodrigo (1999), Domínguez-Rodrigo and Piqueras (2003), Esteban-Nadal et al. (2010) and Fosse et al. (2012). Haynes (1983) was followed for carnivores; Binford (1981), Pickering and Egeland (2006), and Domínguez-Rodrigo and Yravedra (2009) were followed for cut and impact marks; Stiner et al. (1995) was followed for burning; Villa and Mahieu (1991) was followed for bone breakage patterns. Abiotic traces important to this material, such as water dissolution and surface weathering, were studied according to Behrensmeyer (1978), Hedges (2002), and Hedges and Millard (1995). The lithic analysis includes the material from previous excavations (2006–2017), a total of 464 artifacts. The raw materials were identified from macroscopic features (color, texture, inclusions) and technological categories were differentiated based on the criteria for distinguishing the Quina, discoid, and Levallois methods (Turq, 1989; Bourguignon, 1996, 1997, 1997; Boëda, 2013). The typological analysis was carried out on the basis of standard classification criteria (Debenath and Dibble, 1994). Identification of Quina retouch followed the criteria by L. Bourgignon (Bourgignon, 1997, 2001, 2001; Lemorini et al., 2016). The stratigraphic and spatial integrity of the remains was tested by the distribution of lithic artifacts and archaeozoological material containing traces of human processing and carnivore teeth. Since the geoarchaeological analysis proposed that there was some waterflow in the cave, we used two analyses in order to test if the spatial distribution of artifacts was affected. Artifact orientation analysis (Lenoble and Bertran, 2004; McPherron, 2005; Benito-Calvo and de la Torre, 2011), as well as size sorting analysis (Alperson-Afil and Goren-Inbar, 2010: 20; Alperson-Afil, 2017), were applied. 2.3. Dating methods Layers 4a, 4b, and 4c were dated using OSL and ESR methods. The OSL sample was taken during the early stages of excavation, and it originated from the southern profile of M11, from the top of Layer 4 (Fig. 1B). The sediment was sampled for luminescence dating using a light-tight pipe (3 cm in diameter) inserted into the cleaned profile. The resulting hole was then enlarged to determine the current gamma doserate using a calibrated gamma probe (Mercier and Falguères, 2007). The depth below the surface was used to calculate the current cosmic dose-rate according to Prescott and Hutton (1994); this value was not corrected for sediment accumulation over time and was used for the age calculation. In the laboratory, a portion of the sample was prepared under dim orange light to avoid bleaching of the OSL signal: the dominant granulometric fraction (20–40 μm) was extracted from the sediment by wet sieving and treated with H2O2 and HCl to remove organic matter and carbonates, respectively. Additional treatment with H2SiF6 dissolved the feldspar grains. The resulting quartz grains were mounted on stainlesssteel discs with silicone oil and analyzed with a Risoe DA-15 luminescence reader. The SAR protocol (Murray and Wintle, 2000) was applied and a series of tests were carried out to optimize the choice of the measurement parameters: the pre-heat plateau test indicated no dependance with temperature and equivalent doses (Des) were then determined with a pre-heat temperature of 260 ◦ C (applied to natural and regenerated doses) and a cut-heat at 160 ◦ C (applied to the test dose used for normalization). Sixteen individual De values were measured, resulting in a mean De value of 156.0 ± 4.6 Gy. In addition to the gamma and cosmic dose rates, and due to the chosen granulometric fraction, the alpha and beta dose rates also had to be taken into account. The U, Th, K contents of the sample were then determined by analyzing approximately 100 g of dried and homogenized sediment with a high-purity Ge gamma-ray detector. No significant disequilibrium was detected in the U-series chain. The contributions of the dose rate components are given in Table 1. Ten teeth from Layer 4 were sampled for the ESR dating (Fig. 1B). Except for one subsample for each tissue in each tooth, all dental tissues were analyzed only for U by neutron activation analysis (NAA). To assess sedimentary dose rates we performed >120 associated sediment sample analyses. Samples were prepared with protocols for standard enamel ESR dating with contaminant containment procedures for a Class 10,000 clean laboratory (Blackwell, 1989). After powdering to ≤100 mesh, all 3. Results 3.1. Sedimentary sequence – field observations The sequence at Pešturina is dominated by fine-grained sediment and is very different from the often coarse, limestone clast-dominated Table 1 Alpha and beta dose-rates were calculated using the conversion factors of Adamiec and Aitken (1998) and assuming the alpha-efficiency for quartz determined by Tribolo et al. (2001). Water content (defined as mass of water/mass of dry sediment) used in the age calculation was the current water content of the sediment sample (15 ± 3%). The OSL age includes systematic errors related to beta source calibration (3%) and U, Th, K standards (10% for each element). Sample OSL 4 Layer 4 U Th K alpha (ppm) (ppm) (%) dose rate (μGy/a) beta 3.85 5.43 1.36 171 1472 gamma cosmic Annual ± De ± (Gy) 525 3 80 2248 675 156 AGE ± (ka) 4.6 69.4 6.3 D. Mihailović et al. Quaternary International 610 (2022) 1–19 sediment usually found in caves in the region. The sequence contains several sub-horizontal truncations which are most likely caused by water flowing across and eroding material from within the cave. Remnant breccias and vertically eroded deposits are also present along the cave walls. As such, it is probable that there are at least three discontinuities in the sequence where portions of the deposits have been removed, two of which possibly occurred through the action of water. The sequence can be broadly divided into an upper and lower zone, separated by a cemented calcareous layer which was also likely formed through water action. Above the calcareous layer, which served as a marker horizon during excavation, the sediment is generally loose, and sub-horizontally bedded, with localized pedogenesis. Below the calcareous layer, the uppermost sediment in the lower zone is light-colored, clean, compact and horizontally bedded, while the reddish basal deposits are more homogeneous. Evidence for water activity is present towards the base of the sequence where strongly etched and weathered blocks of éboulis (rockfall) have been locally coated in flowstone. Flowstone is also present on top of the cave floor. The full stratigraphic sequence was divided during excavation into five layers based on archaeological materials (Fig. 2). From the surface downwards, the dark Holocene layer (Layer 1) is relatively thin (30–50 cm) and contains disturbed late prehistoric remains. Below it lies a series of lighter silty deposits containing the Upper Paleolithic (Layer 2) and the Middle Paleolithic assemblages (Layers 3 and 4, separated by the carbonate horizon). The cave floor descends sharply from a high point in the rear of the cave to its mouth, where the floor lies almost 4 m below the current ground surface. Layer 5 is a thin, archaeologically sterile unit located just above the cave floor near the cave mouth. From a lithological perspective, the four main sedimentary packages from the bottom up are: a) the sandy Layer 5, b) the reddish, clay-rich Layers 4c and 4b, c) the silt-rich Layers 4a, 3, and 2, and d) the dark, and much disturbed anthropogenic Layer 1. Major erosional surfaces separate Layers 4b from 4a, 4a from 3, and 2 from 1. The most recent unconformity has a possible anthropogenic origin. Despite the unconformity between Layers 4a and 3 and its chronological gap of ~30 ky (Blackwell et al., 2014) these two units appear very similar lithologically. Within Layers 4a and 4b, at least three very thin horizons occur within Squares L8-P10. Each averaging 2–3 cm thick, these horizons contain very high Th concentrations and somewhat elevated K concentrations that produce high sedimentary dose rates. In Golema Pešt, North Macedonia, such highly elevated elemental concentrations mask the presence of crypto- and microtephra (Lowe et al., 2012; Blackwell et al., 2019). Layer 5 contains greyish sandy sediment with very few finds. Layer 5 was resting on impenetrable flowstone which probably lies directly above the cave floor. Layer 4 can be divided into three separate sublayers, with a total thickness of almost 2 m. Layer 4c contains rock fragments and dark brown loose sediment and exhibits a strong granular structure, except where it is protected by large blocks of rockfall. Both Layer 4b and 4c contain fragments of bone and limestone that are stained by oxides. Layer 4b contains the majority of the faunal and lithic remains, as well as more abundant limestone clasts. Remnant breccias located at higher elevations along the cave walls exhibit some similarities with the lithological characteristics of Layer 4b. Large burrows within Layer 4b contain grey silty sediment with inclusions of charcoal. The combination of the remnant breccia and the “missing” grey sediment that is only present as burrow infills suggests that erosion may have removed part of Layer 4b and possibly, an overlying grey unit, in the past. The upper part, Layer 4a, is lithologically distinct from Layers 4b and 4c, and more similar to the overlying Layers 2 and 3. Layer 4a is only 15–30 cm thick. Compared to the upper sequence, Layer 4a is slightly redder and unequally distributed, and – like the overlying deposits – it is locally bedded. Large coprolites occur about 10 cm below the top of Layer 4a. Much of Layer 4a contains large blocks of bedrock having strongly weathered karren on their surfaces (Lundberg, 2019), and Fig. 2. The northern profile in Pešturina, within squares L9, L10, and L11. The yellow horizontal line is set at 303.522 m above sea level. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) which are sporadically coated with flowstone. These flowstones do not extend into the surrounding sediment (4b) which is loose, darker reddish, more clay-rich and lacks the laminations present in 4a. Pleistocene deposits in the upper stratigraphic complex consist of Layers 2 and 3, which are composed of light brown silts whose thickness varies between 30 and 70 cm. Due to the vague nature of the contact between these two layers, namely that Layer 3 has been more compacted, their mutual boundary was better observed during excavations than in the excavated profiles. Layer 3 also contains small lenses of gravel. At the base of Layer 3 is a discontinuous carbonate that is expressed as cementation in some places and in others is thicker, with a laminar fabric. The internal morphology of the cave played a key role in the deposition and erosion of the sediments. It is very likely that water draining from the plateau above the cave ceiling periodically scoured out the sediments in the northern part of the cave mouth. Éboulis caused internal topographic variations, and in places trapped sediment and prevented 4 D. Mihailović et al. Quaternary International 610 (2022) 1–19 disturbance from erosion and bioturbation. 4a and 3 itself exhibits a laminated fabric. The expression of the carbonate is variable around the site. Where the carbonate is thick, at its base is a sediment consistent with 4a that has been cemented by carbonate. Within this are concentrations of secondary phosphates and secondary calcification of guano and coprolite fragments. At the top of the carbonate are areas with vegetal fabric with calcified plant material. In other areas of the site, the carbonate is very thin, evidenced only by localized cementation of the underlying 4a sediment, and mm-scale discontinuous lenses of microcrystalline calcite. The overall laminar nature observed in the field may be due to a combination of cementation of already laminated sediment (i.e., the top of 4a), and secondary carbonate that locally forms micrite lenses and masses of calcified plants. Petrographic analysis of the large, weathered limestone roof blocks revealed that sand- and silt-sized grains of quartz are present as inclusions. Therefore, some of the quartz observed in thin section could be derived from within the cave itself. Likewise, the papules observed in 4c could either derive from the argillic horizons of soils located outside the cave, or from ponded clay deposits in the rear of the karstic system (cf. Goldberg et al., 2019). The other components of the deposits are endogenous to the cave. The only potential anthropogenic materials observed were fragments of charcoal at the base of the sequence in Layer 4c (Fig. 3A). The overall package is consistent with the following: 3.2. Sedimentary micromorphology In thin section, Layer 4c exhibits the granular microstructure that was also visible in the field (Fig. 3A). The reddish, clay-rich sediment forms sand-sized and gravel-sized rounded aggregates, and also coats fragments of limestone and quartz. The darker color that distinguished this sublayer in the field occurs likely due to the abundance of red clay in the matrix. Papules composed of limpid and laminated red clay are also present in the sand fraction. This sublayer contains sand-sized charcoal fragments (Fig. 3B). Layer 4b (Fig. 3B–E) is of particular importance for this study because of its rich artifact assemblage. Although red clay, quartz silt, and sand are dominant components, the sediment in this unit is notably rich in sand- and gravel-sized fragments of phosphatic grains, including carnivore coprolites (cf. Horwitz and Goldberg, 1989). The gravel-sized inclusions of limestone exhibit evidence of etching of their outer edges. These dissolution features are overlain by coatings of secondary manganese oxides. Manganese oxides are also present around fragments of bone. The microstructure of the 4b samples is dominated by channels. Larger chambers filled with granules and crumbs are also present. In the northern part of the excavated area, Layer 4b exhibits some weak laminations, and one discontinuous, fine-grained and phosphatic crust was observed. Here carbonate hypocoatings are also present in some channel voids. Strongly laminated sediment was documented in 4a in multiple sampling locations. The sublayer is composed of mm-scale lenses of sorted sediment ranging from brownish-yellow clay to fine sand and siltsized quartz grains, to coarser materials dominated by rounded coprolite fragments (Fig. 3F–H). The carbonate that defines the contact between 1. A bioturbated cave deposit (Layers 4c and 4b) that was strongly weathered in place, along with blocks of limestone, under acidic and wet conditions. 2. Erosion that partially removed the previously deposited sediment, and in places left gravel lags. This erosion appears to be coupled with a change in the drip water chemistry towards more alkaline Fig. 3. Photomicrographs of key features in the Pešturina sediment: A) Layer 4c, showing the strong granular structure with a large fragment of bone (b), and a smaller, rounded fragment of charcoal (ch) (planepolarized light = PPL); B) a papule (p) in Layer 4c provides evidence of either transport of material from local soils into the cave or reworking of ponded cave deposits (crossed-polarized light XPL); C) features in Layer 4b include a bone with manganese oxide staining around its edges (b), and a fragment of a coprolite (co) (PPL); D) a fragment of limestone (ls) in Layer 4b with a partially dissolved outer edge, which indicates acidic conditions (XPL); E) phosphatic crust (ph) in Layer 4b likely formed in place, yellow grains of fragmented coprolite or guano are also visible (PPL); F) Layer 4a, reworked sediment with numerous fragments of coprolites, rounded fragments of bone and limestone, and channel voids (PPL); G) strongly laminated sediment just above Fig. 3F in Layer 4a provides evidence for reworking by water (PPL); H) carbonate at the contact between Layer 4a and 3 in an area where it is expressed as a thin, microcrystalline, discontinuous lens as indicated here with an arrow (XPL); I) here, the carbonate deposit is thicker at the contact than in Fig. 3H, seen along with calcified plant remains (ca) (XPL). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) 5 D. Mihailović et al. Quaternary International 610 (2022) 1–19 conditions, which caused the formation of thin flowstones on top of the weathered limestone roof blocks. 3. Sedimentation resumed with deposition of a laminated package of waterlain sediment ranging in texture from clay to gravel. The presence of sand-sized coprolite fragments suggests that some, if not all, material in Layer 4a could have derived from internal sources within the cave (Fig. 3G), although an aeolian origin for some of the quartz cannot be excluded. The range of textures indicates fluctuating depositional energies. 4. Finally, the sequence was capped with carbonate, which cemented the laminated deposits and initialized localized phosphatic diagenesis and plant growth under higher humidity. reworked from a (missing) layer, originally deposited between Layers 2 and 4, and subsequently eroded before the deposition of Layer 3. Assuming that the sediment in that missing layer did not differ geochemically from Layer 4a, ET8 would date at 74 ± 6 ka and AT36 to 68 ± 7 ka. Both ages would correlate with MIS 4. Therefore, based on AT24’s age alone, Layer 4b likely dates at 111 ± 11 ka, which correlates with MIS 5e-5c. In Layer 4c, samples AT47 and AT63 gave ages that are in agreement, given their associated errors, while their mean age, 107.4 ± 2.8 ka, correlates with MIS 5d. Both ages agree well stratigraphically with that for AT24 in Layer 4b. Without more ages for teeth from Layer 4b and 4c, the two deeper teeth may have been reworked from Layer 4b. Alternatively, Layers 4b and 4c could simply be slightly different facies of one geochemically uniform layer, which differ only in their éboulis concentrations. All three teeth found in situ within Layers 4b and 4c correlate well with MIS 5e-5c. 3.3. Chronometric dating The site’s lower layers (4a, 4b, and 4c) were dated using OSL and ESR methods. 3.4. Paleoecological indicators 3.3.1. OSL dating With a total external dose rate of 2.25 mGy/a, the sample OSL 4 has an age of 69.4 ± 6.3 ka (Table 1). Therefore, sediment in the upper portion of Layer 4, bordering with Layer 3, was potentially deposited during MIS 4 (boundaries: 59–74 ka; Martinson et al., 1987), so during a relatively cold climatic event. However, changes over time in the cosmic dose rate and in the water content could have made the mean dose rate lower, leading to an older age estimate. Consequently, a MIS 5a age can not be totally ruled out. Easily traversed longitudinally along the Morava Corridor, the Central Balkans is a mostly mountainous region, with average altitudes at 300–500 m, and mountains up to 2,600 m high. The high-altitude Dinaric Alps, which separate the Central Balkans from the Mediterranean to the south and southwest, were never heavily glaciated during the Upper Pleistocene (Djurović, 2012), enabling movements of human and animal populations from the Mediterranean towards the interior of the Balkans during warm events of the Upper Pleistocene. The steppe belt shifted south during the cold stages, allowing steppe-dwelling species to penetrate all the way to the Velika (Great) and Južna (Southern) Morava valleys. However, forest vegetation persisted in the interior of the Balkans during the cold intervals (Van Andel and Tzedakis, 1996). The southern Pannonian Basin and the interior of the Balkans represented a suitable ecological refugium at the beginning of Last Glacial Maximum (Tzedakis and Bennett, 1995; Tzedakis, 2004), and the Late Pleistocene glacial refugia were perhaps not exclusively tied to the Mediterranean climatic belt. So far, sequences of Late Pleistocene climatic indicators in the Balkans are fragmentary. These include paleoecological studies of loess in the Middle and Lower Danube River basin (Fitzsimmons et al., 2012), including benthic ostracods from core samples and sediment samples from the Ohrid and Pamvotis Lakes (Frogley et al., 2001; Belmecheri et al., 2010), and the Stalać loess profile (Obreht et al., 2016; Bösken et al., 2017), showing that during MIS 5, the southern Balkans were characterized by a warm, wet Mediterranean climate close to subtropical, with one cooling event identified at around 106 ka BP (Frogley et al., 2001; Belmecheri et al., 2010). Paleoecological indicators generally correspond to the dating evidence. Among the faunal remains, thermophilic and forest species are much more numerous than steppe forms in the lower parts of Layer 4, indicating that Pešturina was formed when rather warm conditions prevailed. The remains of porcupine (Hystrix vinogradovi), fallow deer (Dama dama), roe deer (Capreolus capreolus), and pig (Sus scrofa) indicate warm, forested climate conditions for Layer 4. Remains of herpetofauna are also the most numerous in this layer. At least three species of lizard, three snake species, and one species of turtle were found. Of these species, Zamenis longissimus, Vipera аmmodytes, and Lacerta viridis are characteristic of the Mediterranean and European geographical area, while Rana temporaria and Coronella austriaca are representatives of the European and Eurosiberian regions (Jovanović et al., 2020). The turtle remains belong to the genus Testudo, most likely to the species T. hermanni; this turtle inhabits open and semi-open areas of the Mediterranean type, avoiding wetlands and dense forests, and it is still common in these habitats throughout the Balkans (Tomović et al., 2014). The appearance of typical steppe forms in the upper levels of Layer 4 suggests that this layer may have been deposited during the gradual cooling period after the last interglacial or at the beginning of the MIS 4. 3.3.2. ESR dating In all the teeth except AT47, the enamel U concentrations averaged ≤0.2 ppm, while the dentinal U averaged ≤2.4 ppm. Due to these very low U concentrations, none of the calculated ages depended on the assumed U uptake model, except for AT47. Table 2 reports all three standard U uptake models for completeness. By using many sedimentary geochemical analyses, volumetric analyses produced sedimentary dose rates with much higher precision and accuracy, permitting more precise and accurate age estimates. In three thin horizons within Layers 4a and 4b, at 153–156 cm, 173–176 cm, and 194–197 cm below datum within the large central excavation bounded by Squares L9-M8-N10-O9 (Fig. 1B), high Th concentrations and somewhat higher K concentrations have caused higher than average sedimentary dose rates, compared to the typical cave sediment elsewhere in Pešturina. In Layer 4a, which had relatively few éboulis, the volumetrically and time-averaged external dose rates ranged from 462 ± 40 to 499 ± 41 μGy/y (Table 2). In Layers 4b-4c, the concentrations of both small and large éboulis increased with depth, causing the volumetrically and time-averaged external dose rates to drop to as low as 298 ± 22 μGy/y in Layer 4c. In Layer 4a, because all five teeth had very similar accumulated doses ranging from 45.30 ± 1.90 to 49.85 ± 0.98 Gy, their ages showed no significant difference from the top to the bottom of the layer. With a mean age of 93.1 ± 1.4 ka, all the individual ages agree well and correlate well with MIS 5b. Aberrant accumulated doses, ages, U enamel, and dentinal concentrations, all provide evidence that a tooth was likely reworked (Blackwell, 1994). At Pešturina, however, the very low U concentrations removes two pieces of possible evidence for reworking, leaving only their abnormally low accumulated doses and the ages which depend strongly on their accumulated doses. Based on its accumulated dose and its age at 110.5 ± 11.1 ka, AT24 likely was found in situ within Layer 4b. For both ET8 and AT36, however, their significantly lower accumulated doses consequently gave significantly younger ages compared to those in either Layer 4a or 4b. Both would be outliers among the teeth in Layer 4a. Accordingly, ET8 and AT36 have most likely been reworked from higher in the cave by burrowing animals. Since ET8 sat only 2–5 cm below the Layer 2–4b unconformity in Square J14a, ET8 was likely 6 D. Mihailović et al. Quaternary International 610 (2022) 1–19 Table 2 ESR ages for Layer 4, Pešturina Cave, Serbia. Elevation Concentrations1 Accumulated Depth, z [Uen ] [Uden ] Dose, AΣ Dext (t) EU (m) (ppm) (ppm) (Grey) (μGy/y) (ka) (ka) (ka) 302.85 1.73 ± 0.05 0.03 1.64 0.30 46.12 0.71 488.7 42.0 84.5 2.6 91.4 3.0 95.6 3.3 302.82 1.76 ± 0.04 0.04 1.39 0.34 49.85 0.98 499.3 48.0 84.2 3.5 91.0 3.9 96.6 4.3 4a/ 4b 302.753 1.827 ± 0.05 0.03 1.00 0.26 47.60 0.35 473.6 44.8 89.2 1.6 94.4 1.9 98.8 2.4 4a 302.70 1.88 ± 0.07 0.02 0.51 0.02 45.30 1.90 461.8 39.9 88.2 7.6 93.5 8.6 97.2 9.3 302.68 1.90 ± 0.04 0.04 1.63 0.02 45.56 1.37 472.0 41.0 86.3 5.4 91.3 5.8 95.1 6.2 87.5 1.2 93.1 1.4 97.4 1.7 Sample Mean Square Layer AT22 (8) 2012PES50a M10d 4a ET5 (6) 2012PES47a N9a AT32 (8) 2012PES51a M10d AT66 (1) 2014PES114 O10c AT65 (3) 2014PES124 N9d (n) 4a 4a/ 4b Mean (n = 5 teeth; 28 subsamples) 4a AT24 (1) 2012PES28a I14d 4b ET8a (1) 2012PES60 J14d AT36a (1) 2012PES48a M10a AT47 (4) 2014PES134 O10b AT63 (4) 2014PES103 M10a Mean (with AT63) (n = 2 teeth; 8 subsamples) Dose Rate 3 ± ESR Ages1,2 LU RU 303.394 1.186 ± 0.20 0.02 1.94 0.02 50.44 2.82 381.7 38.1 91.5 8.4 110.5 11.1 129.4 14.7 4b reworked 303.36 1.22 ± <0.01 0.02 1.50 0.41 31.94 1.80 399.8a 24.7a 67.7 5.3 73.7 5.9 79.0 6.6 4b reworked 302.674 1.906 ± 0.07 0.02 2.40 0.18 26.62 1.89 335.6 26.9 58.1 5.4 67.6 6.5 76.7 7.9 4c 301.83 2.75 ± 0.89 0.19 6.19 1.57 70.87 1.19 324.6 47.2 78.9 2.6 111.6 3.9 159.5 6.8 301.57 3.01 ± 0.10 0.04 0.64 0.01 30.92 0.38 297.6 22.4 93.5 3.5 102.0 4.0 108.3 4.5 85.3 2.1 107.4 2.8 132.9 4.0 4c reworked? 4c ± 1 Abbreviations. [Uen ] = the mean U concentration in the enamel. [Uden ] = the mean U concentration in the dentine. Dsed (t) = the volumetrically and time-averaged sedimentary dose rate. Dcos (t) = the time-averaged cosmic dose rate. Dext (t) = the volumetrically and time-averaged external dose rate. Dsed (t) = the volumetrically and time-averaged sedimentary dose rate. Dcos (t) = the time-averaged cosmic dose rate. EU = assuming early U uptake, p = −1. LU = assuming linear U uptake, p = 0. RU = assuming recent U uptake, p = 10. 2 Ages calculated using: α/γ factor, κα = 0.15 ± 0.02. enamel density, ρen = 2.95 ± 0.01 g/cm.3. dentinal density, ρden = 2.75 ± 0.01 g/cm.3. cementum density, ρcem = 2.85 ± 0.01 g/cm.3. bone density, ρbone = 2.85 ± 0.01 g/cm.3. carbonate sediment density, ρcct = 2.96 ± 0.01 g/cm.3. quartose sediment density, ρqtz = 2.66 ± 0.01 g/cm.3. initial activity ratio, (234U/238U)0 = 1.2 ± 0.20. tooth radon loss, Rntooth = 0.0 ± 0.0 vol%. All uncertainties are 2 σ. 3 .Dext (t) = Dsed (t) + Dcos (t) a This tooth has been reworked, likely from a layer now eroded between Layer 3 and 4a. This age was calculated assuming that the missing layer’s sedimentary dose rates equaled those in Layer 4a. 7 D. Mihailović et al. Quaternary International 610 (2022) 1–19 effects of which are well-explained in sedimentary micromorphology, is corrosive and weakens the bone structure, leading to delamination and longitudinal cracking in bones and detachment of enamel from dentine in teeth. Because of this process, it is possible that a number of cutmarks were obliterated, thus lowering the original cutmarked NISP and MNE counts. Dissolution has also led to a certain amount of post-depositional in situ breakage, which largely affected low-density axial skeletal elements (Lam et al., 1999, 2003, 2003; Stiner, 2004). For this reason, inferences about human processing patterns should be considered with caution. The presence of anthropogenic marks is generally low (3% in Layer 4; >1% in Layer 3), and processing patterns can be observed only in Layer 4 (Table 4). The majority of cutmarks appear on horse remains. They include dismemberment, filleting, and long bone breakage. However, the remains of the European ass (Equus hydruntinus) show no butchery traces. Processing marks are less numerous on red/fallow deer remains, but the specific pattern of their appearance suggests filleting and extensive bone breakage at the site, including the metapodials and phalanges. Other taxa containing processing marks are Proboscidea, bison, wild boar, and roe deer. Skinning marks were also recorded on cave bear remains. Gnawing and digestion marks were observed on 6% of total NISP in Layer 4. Diameters of punctures and scores are consistent with a hyena in size, while the absence of long bone ends and heavily nibbled margins of long bone fragments are also suggestive for this carnivore. Carnivore damage appears on remains of at least 16 taxa in Layer 4. Gnawing marks appear across all skeletal portions of affected ungulates, and are suggestive for both primary access and scavenging that is typical for hyena foraging behavior. This is also supported by specimens containing both cut and tooth marks. In Layer 4, skeletal profiles vary between two main groups of large ungulates: medium-sized cervids and equids. Equids present more complete carcasses, while cervid carcasses are more ravaged (Fig. 4A). Medium-sized cervids have underrepresented upper limb portions, which are somewhat better represented in equids. Over the course of the skeletal element identification, special attention was given to long bone shaft fragments (LBSF). Since it is now established that there is a direct connection between the difference in bone mineral densities (BMD) of different elements and their representation in zooarchaeological materials (Lam et al., 1999, 2003, 2003; Stiner, 2004) it is exactly the teeth and lower limb bones that have the highest mineral densities and thus the chances of preservation as well. Many authors (Stiner, 1991; Marean and Kim, 1998; Bartram and Marean, 1999; Rogers, 2000; Marean and Cleghorn, 2003; Bar-Oz and Munro, 2004; Faith and Gordon, 2007; Marín-Arroyo, 2009) have stressed that LBSF represents a very important factor in long bone element quantification, especially the minimum number of elements (MNE) and the number of taphonomic traces per element, since long bone epiphyses are significantly more prone to destruction due to lower mineral density and feeding choice of predators. Quantification of anatomical regions through minimum animal units (MAU) is instructive when interpreting butchery and transport strategies, or identifying the predator responsible for further processing after humans, or accumulation of bone material. Humans tend to leave unprocessed smaller bones (such as carpals, tarsals, phalanges) that Table 3 Identified avian taxa from Layer 4 of Pešturina. Class Taxa NISP Anseriformes Galliformes Anas crecca Tetrao tetrix Perdix perdix Coturnix coturnix Perdicinae indet. Falco tinnunculus Gyps fulvus Crex crex Otis tarda Scolopax rusticola Aegolius funereus Picus canus Alauda arvensis Anthus trivialis Ptionoprogne rupestris Petronia petronia Sitta europaea Fringilla coelebs Pyrrhula pyrrhula Oriolus oriolus Garrulus glandarius Pyrrhocorax graculus Corvus monedula 1 2 6 2 Falconiformes Accipitriformes Gruiformes Charadriiformes Strigiformes Piciiformes Passeriformes Aves indet. Total 1 1 1 1 1 1 1 1 1 2 3 2 25 Most of the avifaunal species (Boev and Milošević, 2020) occurring in the Last Interglacial context of Layer 4 at Pešturina (Table 3) inhabited steppe environments (Perdix perdix, Coturnix coturnix, Crex crex, Otis tarda/Tetrax tetrax) and temperate forests (Tetrao tetrix, Scolopax rusticola, Aegolius funereus, Sitta europaea, Fringilla coelebs, Garrulus glandarius). The presence of rock sparrow (Petronia petronia) deserves special attention. Its present-day summer range lies far to the south of the location of Pešturina Cave (Mingozzi and Onrubia, 1997), and the site represents one of the northernmost occurrences of rock sparrow in the Pleistocene of Europe. Also, the presence, in the same layer, of crag martin (Ptyonoprogne rupestris) – which preferably feeds on wetland flying insects – suggests relatively high summer temperatures (Blondel et al., 2010). 3.5. Large mammal assemblages Large mammals are taxonomically diverse throughout Layer 4. Large and medium-sized equids (Equus ferus germanicus, Equus hydruntinus), medium and small-sized cervids (Cervus elaphus, Dama dama, and Capreolus capreolus), and large bovids (Bison priscus) dominated the ungulate assemblage, while the remains of hyena (Crocuta spelaea) are the most numerous among the carnivores (Milošević, 2020). Animal remains from Pešturina are highly fragmented, with the majority of specimens represented by longitudinal splinters, 30–40 mm in length. Water dissolution strongly accounts for the post-depositional alteration of material due to chemical weathering, affecting a quarter of total specimen counts in Layer 4, while surface weathering and trampling are present on less than 1% of specimens. Water dissolution, the Table 4 Representation of different taphonomic marks on ungulate remains from Layer 4 of Pešturina. Ungulate taxa Bos s. Bison Mammal indet. size II Equidae Mammal indet. size II/III Cervidae size III Mammal indet. size III Layer 4a 4b 4c 4a 4b 4c 4a 4b 4c 4a 4b 4c 4a 4b 4c 4a 4b 4c Dismembering cutmarks Filleting cutmarks Cone fractures mNISP/tNISP Gnaw marks mNISP/tNISP – – – – 1 33% 2 – 1 10% 14 66% – – – – 1 100% – – – – 3 75% – – 2 20% 5 55% – – – – – – 1 – – 11% 3 33% 3 3 2 13% 35 58% – – – – – – – 1 1 14% 4 57% – – 1 8% 10 83% – – – – – 26% – 1 – 8% 6 50% – 5 10 22% 22 33% – – – – 1 20% – 1 – 8% 7 58% – 3 2 13% 20 50% – – – – – – 8 D. Mihailović et al. Quaternary International 610 (2022) 1–19 Fig. 4. Wild horse (a) and medium-sized cervids (b) skeletal profiles in terms od minimum number of elements (MNI) and percent of minimum animal units (%MAU) from Layer 4 (A); Equid (a) and cervid (b) MAU to FUI layer 4 (B); ratio of minimum of animal units (MAU) to bone mineral density (BMD) for horse (a) and medium sized cervids (b) from Layer 4 (C). Fig. 5. Time and distance to mass. Ungulates inhabiting steppe and forest are faster to acquire, but larger quantities of bulk yield for the same handling time is apparent in the steppe, because of larger ungulates that inhabit it – like wild horse and bison. Here we observe that return rates of search and handling times are considerably overlapping in steppe and forest biomes, but neither of them significantly overlap with montane biome. Ratios between different ungulates underline the Neanderthal choice of biome exploitation, which is also reflected in their prey focus on animals inhabiting steppe and forest, and its intermediaries. 9 D. Mihailović et al. Quaternary International 610 (2022) 1–19 have little or no nutritional value, and they are often transported as a side (schlepp) effect. MAU can explain different skeletal patterns and tie them to human processing strategies or carnivore consumption. Food utility indices (FUI) (Jones and Metcalfe, 1988; Metcalfe and Jones, 1988) help to understand the reason behind the presence of different butchery and teeth marks on different elements, and processing intensity of different elements with a nutritional yield of the elements, enabling us to understand if the processing occurred as to their utility ranks. The ratio between MAU and FUI for horse and deer taxa is negative (Fig. 4B), suggesting relatively extensive processing, but not the extensive consumption of the prey at the site location, apart from bone marrow. On the other hand, MAU to BMD (Fig. 4C) for cervids and horses are closely correlated, indicative of skeletal element attrition by specific bone mineral density. Hyena elements are all fairly well represented but biased towards teeth. All better-represented taxa have low preservation of axial elements, and overrepresented teeth, owing to their complex taphonomic history: pre-burial carnivore ravaging and post-depositional attrition by water dissolution. Pray carcass mass and the relative distance from the site at which it was acquired is presented in Fig. 5. Table 6 General structure of lithic assemblages from Layer 4 of Pešturina. Cores Blades Flakes Chunks Tools Total Chips and small fragments High quality flint Chalcedony Limestone Jasper Quartz/Quartzite Silicious rock Indeterminate Total 4a 4b 4c 4 4-TOTAL 15 10.8% 18 12.9% 5 3.6% 1 0.7% 1 0.7% 99 71.2% 0 0.0% 0 0.0% 139 99.9% 1 16.7% 0 0.0% 1 16.7% 0 0.0% 0 0.0% 4 66.7% 0 0.0% 0 0.0% 6 100.1% 6 3.4% 28 16.0% 21 12.0% 1 0.6% 0 0.0% 115 65.7% 3 1.7% 1 0.6% 175 100.0% 26 6.9% 49 13.1% 32 8.5% 2 0.5% 1 0.3% 259 69.1% 4 1.1% 2 0.5% 375 100.0% 4 4TOTAL 2 3.7% 1 1.8% 27 50.0% 6 11.1% 18 33.3% 54 99.9% 2 6 4.4% 2 1.5% 74 54.4% 17 12.5% 37 27.2% 136 100% 27 0 0.0% 0 0.0% 2 33.3% 1 16.7% 3 50% 6 100.0% 5 11 6.2% 4 2.3% 90 51.1% 20 11.4% 51 29.0% 176 100.0% 39 19 5.1% 7 1.9% 193 51.9% 44 11.8% 109 29.3% 372 100% 95 4a 4b 4 4-TOTAL 1 0 0 0 0 1 0 2 3 0 1 0 1 0 1 6 2 2 1 1 0 5 0 11 6 2 2 1 1 6 1 19 The same core types are recorded in Layers 4a and 4b (Table 7). The most numerous are those made on quartz pebbles (Fig. 6: 1), knapped via the Quina method (Turq, 1989; Hiscock et al., 2009). One chert core with traces of alternate flaking (Fig. 6: 2) along two axes of the pebble (Bourgignon, 1996) can also be associated with this method. In addition, we identified cores made on quartz pebbles (Fig. 6: 3) and cortical flakes subjected to the discoid unifacial method (Terradas, 2003), preferential (non-Levallois) cores, one Levallois core with traces of centripetal recurrent preparation (Boëda, 2013), as well as one burin-like core for the production of blades. The technological process could be partially reconstructed only for the quartz artifacts. In both layers, elongated pebbles (broken into two, rarely into four parts), and sometimes thick flakes were used as core blanks. Salami slice flakes – characterized by cortical edges on both lateral margins – were not recorded, but numerous flakes related to other Quina strategies (cortical backed, cortical covered, recurrent alternative) have been found (Hiscock et al., 2009). The process of core knapping was mainly focused on the production of flakes – asymmetric in cross-section (Fig. 7: 1–3) – with cortical or flaked (a dos de débitage) backs (Turq, 1989; Bourgignon, 1996). These account for 52.3% of the total number of complete flakes. Flakes with knapped back are the most numerous in Layer 4b (19 pcs.) in which no flakes with cortical back were found, while in Layer 4a flakes with cortical back were more numerous (4 out of 7). Judging by the macroscopically observable edge damage (present on 53.1% of pieces), most of the naturally backed flakes were likely used immediately after production, while a smaller proportion (10.9%) was used as blanks for tool manufacture. The use of the discoid and Levallois methods is testified by the presence of Levallois (9 pcs.), pseudo-Levallois (8 pcs.), and débordant flakes (1 pc.) (Fig. 7: 4, 5). Six Levallois flakes were struck from unipolar recurrent Levallois cores, while three pieces were produced from preferential cores and cores with centripetal preparation (Boëda, 2013). Levallois artifacts from both layers were not retouched. The application of non-Levallois blade technology is confirmed on a small number of blades only (1.9% of the total sample). With the Table 5 Raw material structure (excluding chips) of Pešturina assemblages from Layer 4. 4 7.3% 3 5.4% 5 9.1% 0 0.0% 0 0.0% 41 74.5% 1 1.8% 1 1.8% 55 99.9% 4c Quina-type cores Discoid unifacial cores Preferential cores (non-Levallois) Levallois cores Unipolar blade cores Irregular cores Core fragments Total The structures of the lithic assemblages from Layers 4a and 4b are surprisingly similar considering the chronological distance. They are dominated by quartz/quartzite artifacts (71–75%), with the raw material originating in nearby alluvial or colluvial deposits, probably along the Studena and Nišava Rivers, to the east and north of the site. Flint and chalcedony are also abundant in the Pešturina collection (Table 5). The absence of detailed geological maps and previous elaborate research of flint deposits limit the potential for studying raw material procurement in depth. Sampling of potential sources in the area of 5 km radius has been carried out: as expected, abundant deposits of optimal quality flint and chalcedony were not found. Approximately half of the total flint and chalcedony found in Pešturina shows a strong similarity (including a characteristic white patina) with the raw material from Kremenac, a primary deposit of flint situated around 17 km to the northwest. However, since many of the archaeological specimens preserve pebble neocortex, it is more likely that they were also collected in secondary deposits, presumably closer to the cave. The general structure of artifact assemblages (Table 6) is almost identical in both layers (4a, 4b): cores represent 5–6%, flakes 50–55%, chunks 11–12%, and tools between 27% and 33%. About 50% of quality material (flint and chalcedony) from Layers 4a and 4b are represented by tools, indicating that a number of artifacts were brought from other locations. n % n % n % n % n % n % n % n % n % 4b Table 7 Core types from Layer 4 of Pešturina. 3.6. Lithic assemblages Low qualtity flint n % n % n % n % n % n % n 4a 10 D. Mihailović et al. Quaternary International 610 (2022) 1–19 Fig. 6. Cores on quartz and chert pebbles from Layer 4. exception of a single quartz piece, all of the blades were made of flint and knapped from flat or slightly convex surfaces with unidirectional parallel scars. There is no observable standardization of morphology or dimensions. Attributes of the platforms and ventral sides of these artifacts are not informative with regards to the reconstruction of the knapping techniques used. The high-intensity core reduction in Layer 4 is attested by the relatively low number of cortical artifacts (Table 8) and the small dimensions of the artifacts. The average lengths of unretouched and retouched artifacts in Layer 4b are 29.3 mm and 34.1 mm, respectively. In Layer 4a, unretouched artifacts are 27 mm long on average, while the retouched ones have an average length of 39.1 mm. Sidescrapers and denticulated pieces dominate the artifact assemblages (Fig. 7: 6–10, Table 9), with sidescrapers less well represented in Layer 4a (16.7%) than in Layer 4b (27%). Simple retouched flakes are represented by 16.7% in Layer 4a and 24.3% in Layer 4b. In contrast, combined denticulated and notched pieces are more common in Layer 4a (33.3%) than in Layer 4b (21.6%), indicating their increased usage over time. Other tool types found in both layers represent less than 10% of the total tool number. Approximately 50% of the sidescrapers were retouched via Quina and demi-Quina retouch (Fig. 7: 6–7, 9–10), poorly visible on the quartz pieces (Table 10). Most pieces display two series of retouch negatives – indicating a short cycle of working edge rejuvenation – while a pair of specimens with three series of negatives could be attributed to a long cycle (Lemorini et al., 2016). Resharpening flakes could not be reliably identified among quartz pieces. 3.7. Spatial analyses of bones and artifacts Zones of human activities on site are partially overlapping, as shown by the spatial distribution of lithic artifacts and animal bones with traces of burning and butchering. Carnivores almost certainly contributed to the secondary deposition of remains of human activities, but smaller Table 8 Cortex coverage on complete artifacts from Layer 4 of Pešturina. 0% Fig. 7. Lithic artifacts from Layer 4 (1–8, 10–14) and from the contact zone between Layers 3 and 4 (9): cortically backed flakes (1–3), unretouched Levallois blade (4), pseudo-Levallois point (5), bilateral convergent sidescraper (6), transversal sidescrapers (7–10). <50% >50% 100% Total 11 n % n % n % n % n % 4a 4b 4c 4 4-TOTAL 17 68.0% 6 24.0% 1 4.0% 1 4.0% 25 100.0% 49 73.1% 7 10.4% 10 14.9% 1 1.5% 67 99.9% 1 50.0% 0 0.0% 0 0.0% 1 50.0% 2 100.0% 52 56.5% 29 31.5% 5 5.4% 6 6.5% 92 99.9% 119 63.9% 42 22.6% 16 8.6% 9 4.8% 186 99.9% D. Mihailović et al. Quaternary International 610 (2022) 1–19 Table 9 General classification of retouched tool types from Layer 4 of Pešturina. Notched pieces Denticulated pieces Sidescrapers Retouched flakes Retouched Levallois Burins Endscrapers Raclettes Truncations Perforators Fragments of tools Total n % n % n % n % n % n % n % n % n % n % n % n % 4a 4b 4c 4 4-TOTAL 2 11.1% 4 22.2% 3 16.7% 3 16.7% 0 0.0% 0 0.0% 2 11.1% 1 5.6% 2 11.1% 1 5.6% 0 0.0% 18 100.1% 3 8.1% 5 13.5% 10 27.0% 9 24.3% 0 0.0% 0 0.0% 0 0.0% 2 5.4% 3 8.1% 1 2.7% 4 10.8% 37 99.9% 2 66.7% 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0.0% 1 33.3% 0 0.0% 3 100.0% 6 11.8% 11 21.6% 10 19.6% 7 13.7% 1 2.0% 1 2.0% 3 5.9% 3 5.9% 5 9.8% 4 7.8% 0 0.0% 51 100.1% 13 11.9% 20 18.3% 23 21.1% 19 17.4% 1 0.9% 1 0.9% 5 4.6% 6 5.5% 10 9.2% 7 6.4% 4 3.7% 109 99.9% Table 10 Sidescrapers from layer 4 of pešturina. Lateral Transversal Quina Demi-Quina Non-Quina 4a 4b 4 4-TOTAL 2 1 1 0 2 7 3 3 1 6 6 4 4 0 6 15 8 8 1 14 Fig. 8. Bone accumulation in Layer 4b. and the dripline. This is consistent with other Middle Paleolithic sites such as Riparo Bombrini (Riel-Salvatore et al., 2013) where butchering was performed in the areas near the dripline or completely outside of the cave, while artifacts are grouped in the central zones of activities, mostly associated with very small fragments of fauna. Neanderthal groups who visited Pešturina during the formation of Layer 4a used mainly the central area of the site with potential differentiation between butchering and domestic activities. In contrast, the distribution of artifacts in Layer 4b indicates two zones where artifacts group (Fig. 9B). The first zone occupies Squares O10, O11, P10, P11 in the southwestern part of the excavated area, while the second zone can be seen in the northeast part of the site, in Squares I13, I14, J13, and J14. This corresponds well with the grouping of cut/impact marked and burnt large mammal remains in the same areas. It is especially instructive to notice impact marked and burnt specimens in the same zone, as this is indicative of marrow and grease rendering (Costamagno et al., 2006). Unlike Layer 4a, in Layer 4b these zones are located on the periphery of the site, while the central part of the site contains very small amounts of finds. Considering that Layer 4b has not been fully excavated in the central part of the site, this may be considered as the cause of such distributions in Layer 4b. However, by careful analysis of the distribution map of Layer 4b, it is noticeable that squares adjacent to square K, in which Layer 4b was excavated, contain few or no artifacts. This situation points to the conclusion that the distribution map of Layer 4b does reflect the realistic distribution of artifacts within the layer. fragments with the indication of burning were less susceptible to this process (Camarós et al., 2013). The largest number of chewed specimens and coprolites come from the central part of the trench (Fig. 8). Regarding large mammal remains, specimens with butchering marks are distributed randomly, while specimens with traces of burning are confined into two zones: one in the L/M squares and another in the O squares. The overlapping spatial distribution of human-processed and gnawed specimens shows that hyenas largely disturbed the primary zones of human activities, possibly by scavenging remains of human prey, since the primary zones of butchery activities cannot be recognized. Additionally, bad preservation of old breakages made long bone refits questionable. Several complete limb joint sections, as well as teeth rows, that were both recorded in situ or from close spatial contexts refitted during the analysis of archaeozoological material, point to the lack of significant post-deposition vertical movement of the material. Layer 4 in Pešturina has 3 sublayers, of which only 4a and 4b have yielded adequate amounts of lithic artifacts for spatial analysis. The artifact distribution patterns of Layers 4a and 4b were tested using the Kolmogorov–Smirnov (KS) test. Empirical distribution was tested against the Poisson distribution and the KS test showed a statistically significant difference between empirical and Poisson distribution for the total number of artifacts in Layers 4a (p < 0.05) and 4b (p < 0.05). The test results, therefore, showed that the artifact distributions recorded in both Layers 4a and 4b vary significantly from the Poisson distribution. Based on this, we can conclude that distributions of lithic artifacts in Layers 4a and 4b are not random. The distribution map of Layer 4a shows artifacts grouping in the central area of the site, mainly in squares K-L/10–11 (Fig. 9A). No finer differentiation of activity areas was observed within this main zone of artifact accumulation. Distribution of faunal remains indicates accumulation in squares L-M/9, which are closer to the entrance of the cave 3.8. Hominin fossils Three hominin fossils recovered (Radović et al., 2019; Lindal et al., 2020) are briefly summarized below. An isolated hominin upper first 12 D. Mihailović et al. Quaternary International 610 (2022) 1–19 Fig. 9. Spatial distributions of lithic artifacts in Layer 4a (A) and 4b (B), and spatial distribution of animal remains with cut marks, impact marks, and traces of burning in Layer 4a (C) and 4b (D), quantified as NISP with marks per sub-square unit. molar (Pes-3) was recovered from Layer 4b at Pešturina and dates to MIS 5c. This permanent molar displays an excellent state of preservation and minimal wear allowing a taxonomic assessment (Radović et al., 2019). The crown shows internally compressed cusps as reflected in the relatively small occlusal polygon area, a buccolingually skewed crown shape reflected in the angles of the occlusal polygon, and a relatively large hypocone, which represents the second largest cusp by crown base area. These features are considered typical for Neanderthal M1s (Bailey, 2002, 2004, 2006, Martinón-Torres et al., 2013; Quam et al., 2009). The results of the volumetric analysis of crown tissues also support a Neanderthal assessment, particularly with regards to measurements related to enamel thickness. In both 3D volume proportions and 2D cross-sections, Pes-3 exhibits similar enamel thickness to Neanderthals, but differs significantly from modern humans, at least in the 2D analyses. Pes-3 also expresses several non-metric traits that are potentially taxonomically informative: at the enamel-dentine junction, Pes-3 displays a “twinned” paracone dentine horn, a post-paracone tubercle, and a metacone dentine horn which is positioned centrally with regards to the marginal ridge. These traits have been observed in relatively high frequencies in Neanderthals and in lower frequencies in other taxa (Martin et al., 2017; Ortiz et al., 2017). Dental calculus from this tooth produced the oldest microbiome recovered so far, consistent with a Neanderthal microbiome (Yates et al., 2021). Two additional hominin fossils recovered from the Pleistocene strata of Pešturina are from the younger strata. A left lateral mass of an atlas (Pes-1) was recovered from the lower portion of Layer 2. The layer is dated to 31–29 ka cal BP (i.e., MIS 3) and contains an Upper Paleolithic (Gravettian) lithic assemblage. Based on metric and non-metric morphological traits, (i.e, weakly developed tubercle for the insertion of the transverse atlantal ligament), Pes-1 was classified as anatomically modern Homo sapiens (Lindal et al., 2020). A juvenile radius shaft (Pes-2) originated from the contact zone between Layers 3 and 4, giving it a very wide chronological range of 38.9–92 ka (i.e., MIS 3–5b). The specimen was tentatively assessed as Neanderthal based on the traits such as the teardrop-shaped cross-section and the apparently strong lateral curvature of the diaphysis which probably represents a primitive trait retained in Neanderthals (Lindal et al., 2020). The fossil remains conform to the expected attribution of Middle Paleolithic industries in the region to the Neanderthal and Upper Paleolithic to the modern human remains. 4. Discussion 4.1. Behavioral inferences Large mammal remains at Pešturina were mostly accumulated due to predation, and to a small extent due to natural deaths of animals that used the cave as a den or shelter. Since hyenas prey from all available biomes, and Neanderthals only from some of them, this resulted in a broad taxonomic representation of large mammals. Hence, the site represents an excellent opportunity to assess the human and hyena prey choices in MIS 5 from available biomes. Based on taxonomic composition, the immediate cave surroundings are characterized by a mosaic ecosystem throughout MIS 5, consisting of steppe, forest with broadleaf component, and montane foothold. In addition to typical cold steppe species, there are several species that are more confined to the presentday Mediterranean climate. Throughout MIS 5 in the interior of Europe, the Neanderthals preferred to inhabit these mosaic ecosystems, with a wide array of large ungulate choices (Gaudzinski, 2004). In general, Neanderthal subsistence focused primarily on size II ungulates, and less extensively on ungulate sizes I and III. Human processing patterns can be partially observed only in Layer 4, on remains of horse and medium-sized deer. Horse processing for the acquisition of bulk meat and marrow included dismemberment, filleting, and marrow extraction. In contrast, red/fallow deer remains show that mostly middle and lower leg portions were processed at the site, with cutmarks including only filleting, and extensive bone breakage for marrow specifically targeting the lower limb portions. For the other taxa, it was not 13 D. Mihailović et al. Quaternary International 610 (2022) 1–19 possible to establish processing patterns due to the smaller sample size. Of species that show human processing marks, only the horse remains in Layer 4 are clearly characterized by a dominance of prime adults, which is the preference characteristic for Middle Paleolithic large mammal hunting (Mellars, 1996; Fernández and Legendre, 2003; Boscato and Crezzini, 2006; Patou-Mathis, 2006; Arceredillo and Fernández-Lomana, 2009; Morin, 2012). The carnivore to ungulate ratio in terms of total MNI from Layer 4 was 50%, pointing to the extensive use of the cave as a den. The high prevalence of hyena remains, and a hyena mortality pattern which includes both juveniles and adults indicate a communal den during MIS 5. Judging by the distribution of gnaw marks across different taxa (Table 4), hyenas put pressure on the same ungulate size classes as Neanderthals, but here taphonomic traces point to both being relatively successful in terms of primary access to the carcass. Although the human processing marks are not numerous, processing patterns in Layer 4 suggest logistical visits to the cave and wellstructured subsistence behavior reflected in structured large mammal processing for the bulk of caloric yield – from filleting to marrow extraction. This is observed in the negative MAU to FUI ratio (Fig. 4B), indicating that the carcasses were intensely processed. The evidence suggests that the Neanderthals used Pešturina as a temporary butchery camp to process large ungulates hunted in the vicinity, mainly in the open steppe and to a lesser extent in forest habitat, while avoiding caprines confined to broken/mountainous habitat. Horses were more intensely processed at the site than red deer, possibly because their carcasses are larger and require more processing time (Fig. 4D), as carcass mass and the relative distance from the site at which the prey is acquired influences processing patterns (Fig. 5). This processing strategy focused on the bulk of marrow and meat is consistent with previous conclusions so far witnessed in the Central Balkans: red deer processing at Velika Balanica Cave (Marín-Arroyo, 2014) and Crvena Stijena rockshelter (Morin and Soulier, 2017), and the bison processing at Šalitrena Cave (Marín-Arroyo and Mihailović, 2017). The structure of the artifact assemblage is entirely consistent with the activities identified by the faunal analysis, especially since it indicates a short-term settlement of communities that did bring artifacts to the site from somewhere else, but which mostly used intensively knapped quartz pebbles acquired from the immediate vicinity. The artifact production in the cave was focused on the cortical and non-cortical backed flakes, which were either used directly or as blanks for making sidescrapers and other tools. Artifact accumulation in Layer 4a indicates that the main activities were organized in the central part of the site, while faunal remains group near the dripline and the entrance to the cave, where butchering most likely took place. In comparison, two main zones of artifact grouping are observed in Layer 4b, and they coincide with the accumulation of faunal remains. The two zones of artifact grouping in Layer 4b do not differ in technological or typological content, suggesting two possibilities: both zones could have belonged to the same occupation, or these zones could have resulted from two different occupations where human groups chose slightly different places in the cave to perform their activities. While we consider the latter to be a more likely scenario, further research will be necessary to demonstrate the nature of these two areas. (Patou-Mathis, 2000; Moncel, 2001; Dusseldorp, 2009). Lithic industries have an expedient character (Binford, 1979) in the sense that they are dominated by small and non-standardized flakes made from low-quality raw materials, collected in the immediate vicinity of settlements (Valoch, 1984). A different situation was observed at the sites located in the southern parts of the Pannonian Basin, which were attributed to the Central European (CE) Charentian (Gábori-Csánk, 1968; Gábori, 1976; Kozłowski, 2016) or to the Charentian sensu lato (Simek and Smith, 1997; Banda and Karavanić, 2019). Most of these sites were correlated with MIS 5, but only Krapina has a reliable date of 130 ka (Rink et al., 1995). These are mostly cave sites, from which relatively few artifacts have been recovered, with significant presence of carnivores and their prey among the faunal remains (Brajković and Miracle, 2008; Miracle et al., 2010; Daschek and Mester, 2020). The industries show Quina elements, present both in the knapping technology and the inventories of retouched tools (Simek and Smith, 1997; Mester and Moncel, 2006; Banda and Karavanić, 2019). It is known that the classification of Middle Paleolithic industries in Central and Southeast Europe does not conform to the Bordes classification (Bordes, 1953), which is also true for the Charentian. Unlike the Western European Charentian, in which the Quina method predominates (Bourguignon, 1997), different methods occur within the Central Europena (CE) Charentian (including Quina), which is mainly conditioned by the nature of the raw materials used (Simek and Smith, 1997; Mester and Moncel, 2006; Banda and Karavanić, 2019). Furthermore, compared to the Western European Charentian, the typical Quina sidescrapers are far less frequent. All this, however, should not call into question the general nature of this type of industry, which differs significantly from the Micoquian industries in the north of the Pannonian Basin (Kozłowski, 2014), but also from the Mousterian industries of Southeast Europe. In the north of the Balkans, only non-Quina industries have so far been recorded from MIS 5 (Fig. 10). The Levallois component is particularly pronounced in the artifact assemblages from sites in northern Bosnia, including the site Zobište, dated to c. 80–65 ka (Montet-White et al., 1986; Baumler, 1988). In the Eastern Balkans, Layers 10c and 10b in Kozarnika represent the only confirmed occurrence of the early Middle Paleolithic, with the presence of Levallois artifacts and leaf-shaped points. Based on the biostratigraphic data, the layers were attributed (Guadelli et al., 2005) and recently dated to MIS 6 (Tillier et al., 2017). Flint artifacts predominate at sites in the interior of the Balkans, while quartz and quartzite artifacts are poorly represented. Several sites dated to MIS 5 were recorded in the coastal zone: Crvena Stijena (Whallon, 2017), Asprochaliko (Huxtable et al., 1992), Theopetra (Panagopoulou, 1994; Valladas et al., 2007; Karkanas et al., 2015), and Kalamakia (Darlas and Psathi, 2016). These sites showed traces of intensive inhabitation, manifested not only in the sheer amount of the recovered materials but also in the thick recorded deposits with charcoal and ash (Karkanas et al., 2015; Whallon, 2017). The lithic industries have a weak Quina character and have been attributed to different cultural units – the Typical Mousterian, the Crvena Stijena Mousterian, etc. (Mihailović, 2017). The only possible exception is the Theopetra Cave in Thessaly, originally attributed to Quina Mousterian (Panagopoulou, 1994), although this was later called into question (Darlas, 2007). These records document various biomes, where the remains of carnivores were poorly represented (Morin and Soulier, 2017) in contrast to the sites in the interior of the Balkans. All this evidence indicates an increased presence of Neanderthal groups in the coastal zone during MIS 5, although the number of recorded sites remains relatively small. The artifacts from Layer 4 of Pešturina cannot be linked to sites in the Eastern and Western Balkans. Rather, they can be associated with those in the southern part of the Pannonian Basin. Similar to Pešturina, different methods of pebble knapping (Simek and Smith, 1997; Moncel, 2003; Mester and Moncel, 2006) were applied at these southern Pannonian sites. In addition to Quina artifacts, the majority of these sites 4.2. Regional context The regional picture of population density, settlement patterns, and modes of resource acquisition in MIS 5 in the Balkans is quite different from the one seen in Central Europe. Taubachian sites with numerous artifacts and faunal remains have been recorded in Central and Northwestern Europe, documenting the intensive settlement of locations near water sources (Richter, 2008). Some of these sites were reliably dated to the last interglacial (Rink et al., 1995; Richter, 2016; Borel et al., 2017). The recovered materials (artifacts and remains of fauna) indicate, among other things, the systematic exploitation of megafauna 14 D. Mihailović et al. Quaternary International 610 (2022) 1–19 cannot be established, since no sites dated to MIS 6 are currently known in the area, with the exception of Kozarnika, where Levallois artifacts and leaf-points have been confirmed in layers of that age (Guadelli et al., 2005; Tillier et al., 2017). After MIS 6, the CE Charentian, along with the Taubachian and Micoquian in the northern Pannonian Basin, represented the main Middle Paleolithic cultural unit in Central and Southeast Europe (Kozłowski, 2016). In addition to the Quina scrapers and characteristic products of debitage, this complex is characterized by a significant quartz component, especially when it comes to the combined curated-expedient model of techno-economic behavior (Vaquero and Romagnoli, 2018) and the use of quartz pebbles for producing ad hoc tools (Péntek, 2019). This flexible concept in the Balkans can be traced back to the settlement of Mala Balanica, and had continued to be practiced even after the MIS 5, with the Levallois blanks replacing the Quina scrapers within the mobile tool-kit (Mihailović, 2014). Given the current state of research in the Balkans, and contrary to the situation in Western Europe (Bourguignon, 1997; Hiscock et al., 2009; Faivre et al., 2014) it seems that the Quina aspect becomes less and less prominent in the lithic industries over time and that only the industries which are (more or less) tied to the so-called Typical Mousterian occur during MIS 5. The chronologically latest Quina elements come from the sites in the coastal zone (e.g., Crvena Stijena XXII, De Nadale, Klissoura), which are dated to MIS 4 or the beginning of MIS 3 (Sitlivy et al., 2008; Jéquier et al., 2015; Mihailović, 2017). The question of whether the CE Charentian occurs after MIS 4 remains unresolved. The Charentian layer from Petrovaradin Fortress (Mihailović, 2009) was recently dated by OSL to 47–36 ka MIS 3, although there is a possibility that some part of the material was in fact redeposited from the lower layer (dated to 90.2 ± 9.4 ka) due to rodent activity (Marković et al., 2020), while the ages of other sites attributed to this unit: Érd (Mester and Moncel, 2006), Veternica (Banda and Karavanić, 2019), Vindija (Ivanova, 1979), Betalov Spodmol (Osole, 1991) and Borosteni (Cârciumaru et al., 2002) have not been determined yet. 5. Conclusions Pešturina Cave represents the first systematically excavated site dated to MIS 5d-5b in the Balkan interior, and the only site within this area where Neanderthal fossil remains have been discovered. In the Balkans, MIS 5d-5b provided favorable conditions for Neanderthal settlement, despite the intense competition with carnivores. Neanderthals seem to have favored grasslands around the Nišava River Basin for subsistence, probably due to a large ungulate diversity and density in this biome. The focus on one open grassland and one large intermediate/ forest ungulate species (horse and red/fallow deer, respectively) reflects the most successful strategy with the shortest search times, which minimized handling costs while maximizing return rates. The cave itself might have served as a temporary hunting camp in close proximity to the hunting grounds, or even a kill spot – as suggested by nearly complete skeletal patterns and bulk butchery strategies. In cultural, technological, and behavioral terms, Pešturina corresponds fully to the occurrences observed in the southern edge of the Pannonian Basin. Both hominin and carnivore occupations occur at all these sites, which are likely temporary hunting camps. Neanderthals probably used the spring near or in Pešturina Cave for ambush hunting, which generally fits the MIS 5 settlement model (Richter, 2008; Dusseldorp, 2009). Whether the Nišava Valley was inhabited to the same extent as the Zapadna (West) Morava Valley, and whether the communities practiced a radiating settlement model (Richter, 2008), for which some indications exist, remains to be determined. The presence of CE Charentian in the Central Balkans, with pebble tools made mainly on quartz, could be explained in several ways. Although at first glance it seems that the structure of the assemblages could be fully explained by resource distribution and settlement models, it is increasingly evident that the CE Charentian is a distinctive and Fig. 10. MIS 5 sites in the Pannonian Basin and the Balkans mentioned in the text. The dashed line separates the area where the Charentian industries occur in the north from the area with industries attributed to the (non-Quina) Mousterian in the south. Question marks indicate the sites where the status is still debatable. also contain Levallois products. Sidescrapers and denticulate tools predominate among the retouched artifacts, while naturally backed knives are presented in significant percentages. Although the southern position of Pešturina within the distribution zone of the CE Charentian may come as a surprise, it should be kept in mind that the Južna Morava Valley is geographically and ecologically connected to the Velika Morava Valley, which represents a part of the Peri-Pannonian area (Ćalić et al., 2012). Therefore, it is quite understandable why Neanderthal communities that inhabited these territories shared similar sets of cultural traits and practiced similar models of techno-economic behavior. However, the question arises as to how much this model of behavior can be related to the social connections and cultural traditions of Neanderthal communities in Southeast Europe, and how much to the availability of resources and the settlement systems that they practiced. Regarding the evidence that could indicate technological and potentially demographic continuity, the nearby Mala and Velika Balanica Caves (dated to MIS 9–7) show a non-Levallois Quina-type industry (Mihailović and Bogićević, 2016). The technological process directed towards the production of cortically backed tools, with the most commonly used Quina cortical backed and cortical covered methods (Bourguignon, 1997; Hiscock et al., 2009), parallels Pešturina. However, a direct connection between Pešturina and Velika and Mala Balanica 15 D. Mihailović et al. Quaternary International 610 (2022) 1–19 fossil hominin record. Joshua Lindal: interpreted the fossil hominin record. Mirjana Roksandic: Formal analysis, Writing – original draft, designed the study, wrote the study, interpreted the fossil hominin record. geographically limited cultural phenomenon. Although it is not clear to what extent the appearance of these industries can be related to Quina industries confirmed in Velika and Mala Balanica as early as MIS 9–7, or whether this type of industry occurs in later periods (in MIS 4 and 3) for which there are indications but not reliable evidence, Pešturina’s research has definitely shown that in MIS 5 this complex was spread not only on the southern edge of the Pannonian Basin, but also in the Central Balkans. The question remains whether the Morava Valley had been inhabited during the entire duration of MIS 5 or only during the period of climate instability that characterized the MIS 5d-MIS 5a interval. Recent research has demonstrated that the Upper Pleistocene is marked by frequent turnovers, population shifts and regional extincitions in the rhythm of glacial/interglacial cycles (Hublin and Roebroeks, 2009; Müller et al., 2011; Bermúdez de Castro and Martinón-Torres, 2013; Hajdinjak et al., 2018). Regional extinctions are less likley in the Central and Southeast Europe, given the milder climate during glaciations and the lack of geographic barriers to the south. Expansions from the southeast northward could have happened in interglacial phases, and in the opposite direction during glacial periods, with the expansion of steppe elements and the search for refugial areas. New dates for Pešturina indicate that the settlement of this locality occurred somewhat later than culturally similar sites in Western-Central Europe (Richter, 2016), so it is possible that at the onset of cooling (at the beginning of MIS 5d, 5b and 4) there were movements of Neanderthal populations from Central Europe towards refugial areas in the south, similar to what was suggested for the Caucasus (Picin et al., 2020). We hope that the analysis of DNA of Neanderthal remains from Pešturina, the results of which are expected soon, will add further insights into the problem of population movements and interactions between Neanderthal groups in this part of Europe (Peyrégne et al., 2019). Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements Funding was provided by the Ministry of Culture and Information and the Ministry of Education, Science and Technological Development of the Republic of Serbia (project 177023) to DM and NSERC RGPIN2017-04702 and 499 RGPIN-2019-04113 to MR. The ESR Foundation, McMaster Nuclear Reactor (MUNR), Williams College, and the RFK Science Research Institute funded the ESR dating. We thank the RFK staff, RFK and WC students for their assistance in the lab. Alice Pidruczny (MUNR) performed the NAA analyses. SMM and CEM thank Panagiotis Kritikakis (HEP-Tübingen) for preparing the micromorphological blocks and thin sections. NM thanks the LaScArBx program (ANR-10-LABX-52) for financial support. We would like to thank S. Kuhn for his comments on an earlier version of the paper. References Adamiec, G., Aitken, M.J., 1998. Dose-rate conversion factors: Update. Ancient TL 16, 37–50. Alex, B., Mihailović, D., Milošević, S., Boaretto, E., 2019. Radiocarbon chronology of Middle and Upper Paleolithic sites in Serbia, Central Balkans. J. Archaeol. Sci. 25, 266–279. https://doi.org/10.1016/j.jasrep.2019.04.010. Alperson-Afil, N., 2017. Spatial analysis of fire: archaeological approach to recognizing early fire. Curr. Anthropol. 58 (Suppl. 16), S258–S266. Alperson-Afil, N., Goren-Inbar, N., 2010. Ancient flames and controlled use of fire. Vertebrate paleobiology and paleoanthropology. Springer, dordrecht. The Acheulian Site of Gesher Benot Ya’aqov II. Arceredillo, D.A., Fernández-Lomana, J.C.D., 2009. Age of death and seasonality based on ungulate tooth remains from the Upper Pleistocene site of Valdegoba (Burgos, Spain). Journal of Taphonomy 7 (2), 73–89. Bailey, S.E., 2002. Neandertal Dental Morphology: Implications for Modern Human Origins. Arizona State University. Ph.D. Dissertation. Bailey, S.E., 2004. A morphometric analysis of maxillary molar crowns of Middle-Late Pleistocene hominins. J. Hum. Evol. 47, 183–198. Bailey, S.E., 2006. The evolution of non-metric dental variation in Europe. MGFU Mitteilungen 15, 9–29. Banda, M., Karavanić, I., 2019. Mustjerska industrija špilje Veternice. Prilozi Instituta za arheologiju u Zagrebu 36, 5–40. Bar-Oz, G., Munro, N.D., 2004. Beyond cautionary tales: a multivariate taphonomic approach for resolving equifinality in zooarchaeological studies. Journal of Taphonomy 2, 201–220. Bartram Jr., L.E., Marean, C.W., 1999. Explaining the ‘“Klasies Pattern”’: Kua Ethnoarchaeology, the Die Kelders Middle Stone Age Archaeofauna, Long Bone Fragmentation and Carnivore Ravaging. J. Archaeol. Sci. 26, 9–29. Baumler, M., 1988. Core reduction, flake production, and the Middle Paleolithic industry of. Zobiste (Yugoslavia). In: Dibble, H., Mellars, P. (Eds.), Upper Pleistocene Prehistory of Western Eurasia. University of Pennsylvania Press, Philadelphia, pp. 255–274. Behrensmeyer, A.K., 1978. Taphonomic and ecologic information from bone weathering. Paleobiology 4 (2), 150–162. Belmecheri, S., von Grafenstein, U., Andersen, N., Eymard-Bordon, A., Régnier, D., Grenier, C., Lézine, A.-M., 2010. Ostracod-based isotope record from lake Ohrid (Balkan Peninsula) over the last 140 ka. Quat. Sci. Rev. 27–28, 3894–3904. Benito-Calvo, A., de la Torre, I., 2011. Analysis of orientation patterns in Olduvai bed I assemblages using GIS techniques: implications for site formation processes. J. Hum. Evol. 61, 50–60. Bermúdez de Castro, J.M., Martinón-Torres, M., 2013. A new model for the evolution of the human Pleistocene populations of Europe. Quat. Bar Int. 295, 102–112. https:// doi.org/10.1016/j.quaint.2012.02.036. Binford, L.R., 1979. Organization and formation processes: looking at curated technologies. J. Anthropol. Res. 35, 255–273. Binford, L.R., 1981. Bones: Ancient Men and Modern Myths. Academic Press, New York. Blackwell, B.A., 1989. Laboratory Procedures for ESR Dating of Tooth Enamel. McMaster University Department of Geology Technical Memo 89.2., p. 234 Blackwell, B.A., 1994. Problems associated with reworked teeth in Electron Spin Resonance dating. Quat. Sci. Rev. 13 (5–7), 651–660. Data availability The datasets used or analyzed during this study are available from the corresponding author on reasonable request. CRediT authorship contribution statement Dušan Mihailović: Formal analysis, Writing – original draft, designed the study, performed the lithic analysis, wrote the study. Stefan Milošević: Formal analysis, Writing – original draft, performed the analyses of large fauna, performed spatial analysis, wrote the study. Bonnie A.B. Blackwell: provided the ESR dating, Writing – original draft, wrote the study. Norbert Mercier: Writing – original draft, provided the OSL dating, wrote the study. Susan M. Mentzer: Formal analysis, Writing – original draft, provided sedimentological descriptions, performed micromorphological analysis of sediment samples, wrote the study. Christopher E. Miller: Formal analysis, Writing – original draft, provided sedimentological descriptions, performed micromorphological analysis of sediment samples, wrote the study. Mike W. Morley: Writing – original draft, provided sedimentological descriptions, wrote the study. Katarina Bogićević: Formal analysis, Writing – original draft, performed the analysis of small mammals, wrote the study. Dragana Đurić: Formal analysis, Writing – original draft, performed the analysis of herpetofauna, wrote the study. Jelena Marković: contributed data on raw-material sources, wrote the study. Bojana Mihailović: Formal analysis, Writing – original draft, performed the lithic analysis, wrote the study. Sofija Dragosavac: Formal analysis, Writing – original draft, performed the lithic analysis, wrote the study. Senka Plavšić: Formal analysis, Writing – original draft, performed spatial analysis, wrote the study. Anne R. Skinner: SC provided the ESR dating. Iffath I.C. Chaity: SC provided the ESR dating. Yiwen E.W. Huang: SC provided the ESR dating. Seimi Chu: SC provided the ESR dating. Draženko Nenadić: Writing – original draft, provided paleoenvironemtal data, wrote the study. Predrag Radović: interpreted the 16 D. Mihailović et al. Quaternary International 610 (2022) 1–19 Esteban-Nadal, M., Cáceres, I., Fosse, P., 2010. Characterization of a current coprogenic sample originated by Canis lupus as a tool for identifying a taphonomic agent. J. Archaeol. Sci. 37, 2959–2970. Faith, J.T., Gordon, A.D., 2007. Skeletal element abundances in archaeofaunal assemblages: economic utility, sample size, and assessment of carcass transport strategies. J. Archaeol. Sci. 34, 872–882. Faivre, J.-P., Discamps, E., Gravina, B., Turq, A., Guadelli, J.-L., Lenoir, M., 2014. The contribution of lithic production systems to the interpretation of Mousterian industrial variability in south-western France: the example of Combe-Grenal (Dordogne, France). Quat. Bar Int. 350, 227–240. https://doi.org/10.1016/j. quaint.2014.05.048. Fernández, P., Legendre, S., 2003. Mortality curves for horses from the Middle Palaeolithic site of Bau de l’Aubesier (Vaucluse, France): methodological, palaeoethnological and palaeoecological approaches. J. Archaeol. Sci. 30, 1577–1598. Fitzsimmons, K.E., Marković, S.B., Hambach, U., 2012. Pleistocene environmental dynamics recorded in the loess of the middle and lower Danube Basin. Quat. Sci. Rev. 41, 104–118. Fosse, P., Selva, N., Smietana, W., Okarma, H., Wajrak, A., Fourvel, J.B., Madelaine, S., Esteban-Nadal, M., Cáceres, I., Yravedra, J., Brugal, J.P., Prucca, A., Haynes, G., 2012. Bone modification by modern wolf (Canis lupus): a taphonomic study from their natural feeding places. Journal of Taphonomy 10 (3–4), 197–217. Frogley, M.R., Griffiths, H.I., Heaton, T.H.E., 2001. Historical biogeography and Late Quaternary environmental change of Lake Pamvotis, Ioannina (north-western Greece): evidence from ostracods. J. Biogeogr. 28, 745–756. Gábori, M., 1976. Les civilisations du paléolithique moyen entre les Alpes et l’Oural. Akadémiai Kiadó, Budapest. Gábori-Csánk, V., 1968. La station du paléolithique moyén d’Erd, Hongrie. Akadémiai Kiadó, Budapest. Gaudzinski, S., 2004. A matter of high resolution? The Eemian interglacial (OIS 5e) in North Central Europe and Middle Palaeolithic subsistence. Int. J. Osteoarchaeol 14, 201–211. Goldberg, P., Conard, N.J., Miller, C.E., 2019. Geißenklösterle stratigraphy and micromorphology. In: Conard, N.J., Bolus, M., Münzel, S. (Eds.), Geißenklösterle: Chronostratigraphie, Paläoumwelt und Subsistenz im Mittel- und Jungpaläolithikum der Schwäbischen Alb. Tübinger Monographien zur Urgeschichte, Kerns Verlag, Tübingen, pp. 11–47. Guadelli, J.L., Sirakov, N., Ivanova, St, Sirakova, S., Anastassova, E., Courtaud, P., Dimitrova, I., Djabarska, N., Fernandez, P., Ferrier, C., Fontugne, M., Gambier, D., Guadelli, A., Iordanova, D., Iordanova, N., Kovatcheva, M., Krumov, I., Leblanc, J.C., Mallye, B., Marinska, M., Miteva, V., Popov, V., Spassov, R., Taneva, S., TisteratLaborde, N., Tsanova, T., 2005. Une séquence du Paléolithique inférieur au Paléolithique récent dans les Balkans: La grotte Kozarnika à Orechets (Nord-Ouest de la Bulgarie). In: Molines, N., Moncel, M.H., Monnier, J.L. (Eds.), Les premiers peuplements en Europe: Colloque international: données récentes sur les modalités de peuplement et sur le cadre chronostratigraphique, géologique et paléogéographique des industries du Paléolithique ancien et moyen en Europe (Rennes, 22-25 septembre 2003). BAR International Series 1364, Oxford, pp. 87–103. Hajdinjak, M., Fu, Q., Hübner, A., Petr, M., Mafessoni, F., Grote, S., Skoglund, P., Narasimham, V., Rougier, H., Crevecoeur, I., Semal, P., Soressi, M.A., Talamo, S., Hublin, J.J., Gušić, I., Rudan, P., Kućan, Ž., Doronichev, V.B., Golovanova, L.V., Krause, J., Posth, C., Nagel, S., Korlević, P., Slatkin, M., Nickel, B., Reich, D., Patterson, N., Meyer, M., Prüfer, K., Kelso, J., Pääbo, S., 2018. Reconstructing the genetic history of late Neanderthals. Nature 555, 652–656. Haynes, G., 1983. A guide to differentiating mammalian carnivore taxa responsible for gnaw damage to herbivore limb bones. Paleobiology 9, 164–172. Hedges, R.E.M., 2002. Bone diagenesis: an overview of processes. Archaeometry 44, 319–328. Hedges, R.E.M., Millard, A.R., 1995. Measurements and relationships of diagenetic alteration of bone from three archaeological sites. J. Archaeol. Sci. 22, 201–211. Hiscock, P., Turq, A., Faivre, J.P., Bourguignon, L., 2009. Quina procurement and tools production. In: Adams, B., Blades, B.S. (Eds.), Lithic Materials and Paleolithic Societies. Wiley–Blackwell, Singapore, pp. 232–246. Horwitz, L.K., Goldberg, P., 1989. A study of Pleistocene and Holocene hyaena coprolites. J. Archaeol. Sci. 16 (1), 71–94. Hublin, J.-J., Roebroeks, W., 2009. Ebb and flow or regional extinctions? On the character of Neandertal occupation of northern environments. C R Palevol 8 (5), 503–509. Huxtable, J., Gowlett, J.A.J., Bailey, G.N., Carter, P.L., Papaconstantinou, V., 1992. Thermoluminescence Dates and a New Analysis of the Early Mousterian From Asprochaliko. Anthropol. 33 (1), 109–114. Ivanova, S., 1979. Cultural differentiation in the Middle Palaeolithic of Balkan Peninsula. In: Kozłowski, J.K. (Ed.), Middle and Early Upper Palaeolithic in Balkans. Zeszyty Naukowe Uniwersytetu Jagiellonskiego, Warszawa–Krakow, pp. 13–33. Janković, I., Ahern, J.C.M., Karavanić, I., Smith, F.H., 2016. The importance of Croatian Pleistocene hominin finds in the study of human evolution. In: Harvati, K., Roksandic, M. (Eds.), Paleoanthropology of the Balkans and Anatolia. Vertebrate Paleobiology and Paleoanthropology, Springer, Dordrecht, pp. 35–50. Jéquier, C.A., Peresani, M., Romandini, M., Delpiano, D., Joannes-Boyau, R., Lembo, G., Livraghi, A., López-García, J.M., Obradović, M., Nicosia, C., 2015. The de Nadale cave, a single layered Quina Mousterian site in the North Italy. Quartar 62, 7–21. Jones, K.T., Metcalfe, D., 1988. Bare bones archaeology: bone marrow indices and efficiency. J. Archaeol. Sci. 15 (4), 415–423. Jovanović, M., Bisbal-Chinesta, J.F., Đurić, D., Bogićević, K., Nenadić, D., Agusti, J., Blain, H.-A., 2020. Pleistocene herpetofaunal studies in Serbia (Balkan Peninsula, SE Blackwell, B.A.B., Chu, S., Chaity, I., Huang, Y.E.W., Mihailović, D., Roksandic, M., Dimitrijević, V., Blickstein, J.I.B.B., Huang, A., Skinner, A.R., 2014. ESR dating ungulate tooth enamel from the Mousterian layers at Pešturina, Serbia. In: Mihailović, D. (Ed.), Palaeolithic and Mesolithic Research in the Central Balkans. Serbian Archaeological Society, Belgrade, pp. 21–38. Blackwell, B.A.B., Skinner, A.R., Blickstein, J.I.B., Montoya, A.C., Florentin, J.A., Baboumian, S.M., Ahmed, I.J., Deely, A.E., 2016. ESR in the 21st Century: from buried valleys and deserts to the deep ocean and tectonic uplift. Earth Sci. Rev. 158, 125–159. Blackwell, B.A.B., Šalamanov-Korobar, L., Huang, C.L.C., Zhuo, J.L., Kitanovski, B., Blickstein, J.I.B., Florentin, J.A., Vasilevski, S., 2019. Sedimentary Radioactivity in an Upper Paleolithic-Middle Paleolithic (MP-UP) Transition Site: Increasing ESR Tooth Dating Accuracy at Golema Pešt, North Macedonia. Radiat. Protect. Dosim. 186, 94–112. Blondel, J., Aronson, J., Bodiou, J.-Y., Boeuf, G., 2010. The Mediterranean Region Biological Diversity in Space and Time. Oxford University Press, Oxford. Boëda, E., 2013. Techno-logique & Technologie: une paléo-histoire des objets lithiques tranchants. @rchéo-éditions.com. Boev, N.Z., Milošević, S., 2020. Late Pleistocene avifauna of the Pešturina cave (Nišava district, SE Serbia) and its implications for late Pleistocene refugia on the Central Balkans. Bull. Nat. Hist. Mus. Plovdiv 4, 1–14. Bordes, F., 1953. Essai de Classification des industries “moustériennes”. Bull. Soc. préhist. fr. 50, 457–466. Borel, A., Dobosi, V., Moncel, M.H., 2017. Neanderthal’s microlithic tool production and use, the case of Tata (Hungary). Quat. Int. 435, 5–20. Boscato, P., Crezzini, J., 2006. Homo neanderthalensis e Homo sapiens: lo sfruttamento delle parti scheletriche degli ungulati. In: Guerci, A., Consigliere, S., Castagno, S. (Eds.), Atti del XVI Congresso degli Antropologi Italiani (Genova, 29–31 ottobre 2005). Edicolors Publishing, Milano, pp. 231–246. Bösken, J., Klasen, N., Zeeden, C., Obreht, I., Marković, S.B., Hambach, U., Lehmkuhl, F., 2017. New luminescence-based geochronology framing the last two glacial cycles at the southern limit of European Pleistocene loess in Stalać (Serbia). Geochronometria 44 (1), 150–161. https://doi.org/10.1515/geochr-2015-0062. Bourguignon, L., 1996. La conception de debitage Quina. Quat. nova VI 149–166. Bourguignon, L., 1997. Le Moustérien de type Quina: Nouvelle définition d’une entité technique. Ph.D. Dissertation. Université de Paris X. Bourguignon, L., 2001. L’apport de l’expérimentation et de l’analyse technomorphofonctionnelle à la reconnaissance du processus d’aménagement de la retouche Quina. In: Bourguignon, L., Ortega, I., Frère-Soutot, M.-Ch (Eds.), Préhistoire et approche expérimentale, Préhistoires 5. Editions M.M. Montagnac, Paris, pp. 35–63. Brajković, D., Miracle, P.T., 2008. Middle Palaeolithic and Early Upper Palaeolithic subsistence practices at Vindija cave, Croatia. In: Darlas, A., Mihailović, D. (Eds.), The Palaeolithic of the Balkans. BAR International Series 1819, Oxford, pp. 107–116. Ćalić, J., Milošević, M., Gaudenyi, T., Štrbac, D., Milivojević, M., 2012. Pannonian Plain as a morphostructural unit of Serbia. Bulletin of the Serbian Geographical Society 92 (1), 47–70. https://doi.org/10.2298/GSGD1201047C. Camarós, E., Cueto, M., Teira, L.C., Tapia, J., Cubas, M., Blasco, R., Rosell, J., Rivals, F., 2013. Large carnivores as taphonomic agents of space modification: an experimental approach with archaeological implications. J. Archaeol. Sci. 40, 1361–1368. Cârciumaru, M., Moncel, M., Anghelinu, M., Cârciumaru, R., 2002. The Cioarei-Borosteni Cave (Carpathian Mountains, Romania): Middle Palaeolithic finds and technological analysis of the lithic assemblages. Antiquity 76 (293), 681–690. Churchill, S.E., 2014. Thin on the Ground: Neandertal Biology, Archeology, and Ecology. Iowa, Wiley Blackwell, Ames. Costamagno, S., Meignen, L., Beauval, C., Vandermeersch, B., Maureille, B., 2006. Les Pradelles (Marillac-le-Franc, France): a Mousterian reindeer hunting camp? J. Anthropol. Archaeol. 25, 466–484. Darlas, A., 2007. The Mousterian of Greece under the light of recent research. L’Anthropologie 111, 346–366. https://doi.org/10.1016/j.anthro.2007.04.001. Darlas, A., Psathi, E., 2016. The Middle and Upper Paleolithic on the Western Coast of the Mani Peninsula (Southern Greece). In: Harvati, K., Roksandic, M. (Eds.), Paleoanthropology of the Balkans and Anatolia. Vertebrate Paleobiology and Paleoanthropology. Springer, Dordrecht, pp. 95–118. Daschek, É., Mester, Z., 2020. A site with mixed occupation: Neanderthals and carnivores at Érd (Hungary). J. Archaeol. Sci. 29, 102–116. https://doi.org/10.1016/j. jasrep.2019.102116. Debénath, A., Dibble, H., 1994. Handbook of Palaeolithic Typology, Volume I. Lower and Middle Palaeolithic of Europe. University of Pennsylvania Press, Philadelphia. Deely, A.E., Blackwell, B.A.B., Mylroie, J.E., Carew, J.L., Blickstein, J.I.B., Skinner, A.R., 2011. Testing cosmic dose rate models for ESR: dating corals and molluscs on San Salvador. Bahamas. Radiat. Meas 46, 853–859. Djurović, P., 2012. The Debeli Namet glacier (Durmitor, Montenegro): from the second half of the 20th century to the present. Acta Geogr. Slovo. 52 (2), 277–301. Domínguez-Rodrigo, M., 1999. Flesh availability and bone modifications in carcasses consumed by lions: palaeoecological relevance in hominid foraging patterns. Palaeogeogr. Palaeoclimatol. Palaeoecol. 149 (1–4), 373–388. Domínguez-Rodrigo, M., Yravedra, J., 2009. Why are cut mark frequencies in archaeofaunal assemblages so variable? A multivariate analysis. J. Archaeol. Sci. 36 (3), 884–894. Domı́nguez-Rodrigo, M., Piqueras, A., 2003. The use of tooth pits to identify carnivore taxa in tooth-marked archaeofaunas and their relevance to reconstruct hominid carcass processing behaviours. J. Archaeol. Sci. 30 (11), 1385–1391. Dusseldorp, G., 2009. A View to Kill. Investigating Middle Palaeolithic Subsistence Using an Optimal Foraging Strategy. Sidestone Press, Leiden. 17 D. Mihailović et al. Quaternary International 610 (2022) 1–19 Europe): state of art and perspectives. Quat. Sci. Rev. 233 https://doi.org/10.1016/j. quascirev.2020.106235, 106236. Karkanas, P., White, D., Lane, C.S., Stringer, C., Davies, W., Cullen, V.L., Smith, V.C., Ntinou, M., Tsartsidou, G., Kyparissi-Apostolika, N., 2015. Tephra correlations and climatic events between the MIS6/5 transition and the beginning of MIS 3 in Theopetra Cave, central Greece. Quat. Sci. Rev. 118, 170–181. Kozłowski, J.K., 2014. Middle Palaeolithic variability in Central Europe: Mousterian vs Micoquian. Quat. Bar Int. 326–327, 344–363. Kozłowski, J.K., 2016. Taxonomy of the Early Middle Palaeolithic in Central Europe. Litikum - J. Lithic Res. Roundtable 4, 19–27. Lam, Y.M., Xingbin, C., Pearson, O.M., 1999. Intertaxonomic variability in patterns of bone density and the differential representation of bovid, cervid, and equid elements in the archaeological record. Am. Antiq. 64 (2), 343–362. Lam, Y.M., Pearson, O.M., Marean, C.W., Chen, X., 2003. Bone density studies in zooarchaeology. J. Archaeol. Sci. 30, 1701–1708. Lemorini, C., Bourguignon, L., Zupancich, A., Gopher, A., Barkai, R., 2016. A scraper’s life history: morpho-techno-functional and use-wear analysis of Quina and demiQuina scrapers from Qesem Cave, Israel. Quat. Bar Int. 398, 86–93. Lenoble, A., Bertran, P., 2004. Fabric of Palaeolithic levels: methods and implications for site formation processes. J. Archaeol. Sci. 31, 457–469. Lindal, J., Radović, P., Mihailović, D., Roksandic, M., 2020. Postcranial hominin remains from the late Pleistocene of Pešturina cave (Serbia). Quat. Int. 542, 9–14. Lowe, J., Barton, N., Blockley, S., Bronk Ramsey, C., Cullen, V.L., Davies, W., Gamble, C., Grant, K., Hardiman, M., Housley, R., Lane, C.S., Lee, S., Lewis, M., MacLeod, A., Menzies, M., Müller, W., Pollard, M., Price, C., Roberts, A.P., Rohling, E.J., Satow, C., Smith, V.C., Stringer, C.B., Tomlinson, E.L., White, D., Albert, P., Arienzo, I., Barker, G., Borić, D., Carandente, A., Civetta, L., Ferrier, C., Guadelli, J.-L., Karkanas, P., Koumouzelis, M., Müller, U.C., Orsi, G., Pross, J., Rosi, M., Shalamanov-Korobar, Lj, Sirakov, N., Tzedakis, P.C., 2012. Volcanic ash layers illuminate the resilience of Neanderthals and early modern humans to natural hazards. Proc. Natl. Acad. Sci. Unit. States Am. 109 (34), 13532–13537. Lundberg, J., 2019. Karren, cave. In: White, W., Culver, D., Pipan, T. (Eds.), The Encyclopedia of Caves, third ed. Academic Press, pp. 558–599. https://doi.org/ 10.1016/C2017-0-01162-X. Majkić, A., d’Errico, F., Milošević, S., Mihailović, D., Dimitrijević, V., 2018. Sequential incisions on a cave bear bone from the Middle Paleolithic of Pešturina cave, Serbia. J. Archaeol. Method Theor 25, 69–116. https://doi.org/10.1007/s10816-017-93315. Marean, C.W., Cleghorn, N., 2003. Large mammal skeletal element transport: applying foraging theory in a complex taphonomic system. Journal of Taphonomy 1, 15–42. Marean, C.W., Kim, S.Y., 1998. Mousterian large-mammal remains from Kobeh cave: behavioral implications for Neanderthals and early modern humans. Curr. Anthropol 39, S79–S113. Marín-Arroyo, A.B., 2009. A comparative study of analytic techniques for skeletal part profile interpretation at El Mirón Cave (Cantabria, Spain). ARCHAEOFAUNA 18, 79–98. Marín-Arroyo, A.B., 2014. Middle Pleistocene subsistence in Velika Balanica, Serbia: preliminary results. In: Mihailović, D. (Ed.), Palaeolithic and Mesolithic Research in the Central Balkans. Serbian Archaeological Society, Commission for the Palaeolithic and Mesolithic, Belgrade, pp. 121–129. Marín-Arroyo, A.B., Mihailović, B., 2017. The Chronometric dating and subsistence of late Neanderthals and early anatomically modern humans in the Central Balkans. Insights from Šalitrena Pećina (Mionica, Serbia). J. Anthropol. Res. 73 (3), 413–447. Marković, S.B., Vandenberghe, J., Stevens, T., Mihailović, D., Gavrilov, M.B., Radaković, M.G., Zeeden, C., Obreht, I., Perić, Z.M., Nett, J.J., Lehmkuhl, F., 2020. Geomorphological evolution of the Petrovaradin Fortress Palaeolithic site (Novi Sad, Serbia). Quat. Res. 1–14. https://doi.org/10.1017/qua.2020.88. Martin, R.M.G., Hublin, J.-J., Gunz, P., Skinner, M.M., 2017. The morphology of the enamel-dentine junction in Neanderthal molars: gross morphology, non-metric traits, and temporal trends. J. Hum. Evol. 103, 20–44. Martinón-Torres, M., Spěváčková, P., Gracia-Téllez, A., Martínez, I., Bruner, E., Arsuaga, J.L., Bermúdez de Castro, J.M., 2013. Morphometric analysis of molars in a Middle Pleistocene population shows a mosaic of ’modern’ and Neanderthal features. J. Anat. 223, 353–363. Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore, T.C., Shackleton, N.J., 1987. Age dating and the orbital theory of the ice ages: Development of a high-resolution 0 to 300 000 year chronostratigraphy. Quat. Res. 27, 1–29. McPherron, S., 2005. Artifact orientations and site formation processes from total station proveniences. J. Archaeol. Sci. 32 (7), 1003–1014. https://doi.org/10.1016/j. jas.2005.01.015. Mellars, P., 1996. The Neanderthal Legacy: an Archaeological Perspective from Western Europe. Princeton University Press, Princeton, New Jersey. Mercier, N., Falguères, C., 2007. Field gamma dose-rate measurement with a NaI(Tl) detector: Re-evaluation of the "threshold" technique. Ancient TL 25 (1), 1–4. Mester, Z., Moncel, M.H., 2006. Le site paléolithique moyen d’Erd (Hongrie): nouvelles données sur les chaînes opératoires et résultats morpho-fonctionnels de la production. Anthropologie 44 (3), 221–240. Metcalfe, D., Jones, K.T., 1988. A reconsideration of animal body-part utility indices. Am. Antiq. 53 (3), 486–504. Mihailović, D., 2009. Middle Palaeolithic Settlement at Petrovaradin Fortress. Petrovaradin edition II (The City Museum of Novi Sad, Novi Sad). Mihailović, D., 2014. Paleolit na centralnom Balkanu - kulturne promene i populaciona kretanja. Srpsko arheološko društvo, Beograd. Mihailović, D., 2017. Paleolithic-Mesolithic Crvena Stijena in relation to other sites. In: Whallon, R. (Ed.), Crvena Stijena in Cultural and Ecological Context – Multidisciplinary Archaeological Research in Montenegro. Montenegrin Academy of Sciences and Art. National Museum of Montenegro, Podgorica, pp. 205–229. Mihailović, D., Bogićević, K., 2016. Technological changes and population movements in the late Lower and early Middle Palaeolithic of the Central Balkans. In: Harvati, K., Roksandić, M. (Eds.), Paleoanthropology of the Balkans and Anatolia. Vertebrate Paleobiology and Paleoanthropology. Springer, Dordrecht, pp. 139–151. Mihailović, D., Milošević, S., 2012. Istraživanja paleolitskog nalazišta Pešturina kod Niša. Glasnik Srpskog arheološkog društva 28, 87–106. Milošević, S., 2020. Competition between Humans and Large Carnivores: Case Studies from the Late Middle and Upper Palaeolithic of the Central Balkans. BAR International Series 2961. Oxford. Mingozzi, T., Onrubia, A., 1997. Petronia petronia rock sparrow. In: Hagemeijer, E.J.M., Blair, M. (Eds.), The EBCC Atlas of European Breeding Birds: Their Distribution and Abundance. T. Poyser & A.D. Poyser, p. 699. London. Miracle, P.T., Mauch Lenardić, J., Brajković, D., 2010. Last glacial climates, "refugia", and faunal change in southeastern Europe: mammalian assemblages from Veternica, Velika pećina, and Vindija caves (Croatia). Quat. Int. 212, 137–148. Moncel, M.H., 2001. Microlithic Middle Palaeolithic assemblages in central Europe and elephant remains. In: Cavarretta, G., Giola, P., Mussi, M., Palombo, M.R. (Eds.), La Terra Delgli Elefanti: the World of Elephants, Proceedings of the 1st International Congress. Consiglio nazionale delle ricerche, Rome, pp. 314–317. Moncel, M.-H., 2003. Tata (Hongrie). Un assemblage microlithique du début du pléistocène supérieur en Europe Centrale. L’Anthropologie 107, 117–151. Montet-White, A., Laville, H., Lezine, A.-M., 1986. Le Paléolithique du Bosnie du Nord. Chronologie, environment et préhistoire. L’Anthropologie 90 (1), 29–88. Morin, E., 2012. Reassessing Paleolithic Subsistence: Neandertal and Modern Human Foragers of Saint-Césaire. Cambridge University Press, Cambridge. Morin, E., Soulier, M., 2017. New criteria for the archaeological identification of bone grease processing. Am. Antiq. 82, 96–122. https://doi.org/10.1017/aaq.2016.16. Müller, U.C., Pross, J., Tzedakis, P.C., Gamble, C., Kotthoff, U., Schmiedl, G., Wulf, S., Christanis, K., 2011. The role of climate in the spread of modern humans into Europe. Quat. Sci. Rev. 30, 273–279. https://doi.org/10.1016/j. quascirev.2010.11.016. Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Radiat. Meas. 32, 57–73. https://doi.org/ 10.1016/S1350-4487(99)00253-X. Obreht, I., Zeeden, C., Hambach, U., Veres, D., Marković, S.B., Bösken, J., Svirčev, Z., Bačević, N., Gavrilov, M.B., Lehmkuhl, F., 2016. Tracing the influence of Mediterranean climate on Southeastern Europe during the past 350,000 years. Sci. Rep. 6 https://doi.org/10.1038/srep36334, 36334. Ortiz, A., Bailey, S.E., Hublin, J.-J., Skinner, M.M., 2017. Homology, homoplasy and cusp variability at the enamel-dentine junction of hominoid molars. J. Anat. 231, 585–599. https://doi.org/10.1111/joa.12649. Osole, F., 1991. Betalov Spodmol, rezultati paleolitskih iskopavanj S. Brodarja – II del. Poročilo o raziskovanju paleolita, neolita in eneolita v Sloveniji XIX 7–129. Panagopoulou, E., 1999. The Theopetra Middle Palaeolithic assemblages: their relevance to the Middle Palaeolithic of Greece and adjacent areas. In: Bailey, G.N., Adam, E., Panagopoulou, E., Perlès, C., Zachos, K. (Eds.), The Palaeolithic Archaeology of Greece and Adjacent Areas: Proceedings of the ICOPAG Conference, Ioannina, September 1994. British School at Athens, London, pp. 252–265. Patou-Mathis, M., 2000. Neanderthal subsistence behaviours in Europe. Int. J. Osteoarchaeol. 10 (5), 379–395. Patou-Mathis, M., 2006. Comportements de subsistance des Néandertaliens d’Europe. In: Demarsin, B., Otte, M. (Eds.), Neanderthals in Europe. ERAUL 117, Liège, pp. 67–76. Péntek, A., 2019. Quartz and quartzite as lithic raw materials in the Hungarian Palaeolithic. Archeometriai Műhely XVI/2 65–84. Peyrégne, S., Slon, V., Mafessoni, F., de Filippo, C., Hajdinjak, M., Nagel, S., Nickel, B., Essel, E., Le Cabec, A., Wehrberger, K., Conard, N.J., Kind, C.J., Posth, C., Krause, J., Abrams, G., Bonjean, D., Di Modica, K., Toussaint, M., Kelso, J., Meyer, M., Pääbo, S., Prüfer, K., 2019. Nuclear DNA from two early Neandertals reveals 80,000 years of genetic continuity in Europe. Sci. Adv 5 (6), eaaw5873. https://doi.org/ 10.1126/sciadv.aaw5873. Picin, A., Hajdinjak, M., Nowaczewska, W., Benazzi, S., Urbanowski, M., Marciszak, A., Fewlass, H., Socha, P., Stefaniak, K., Żarski, M., Wiśniewski, A., Hublin, J.-J., Nadachowski, A., Talamo, S., 2020. New perspectives on Neanderthal dispersal and turnover from Stajnia Cave (Poland). Sci. Rep. 10 https://doi.org/10.1038/s41598020-71504-x, 14778. Pickering, T.R., Egeland, C.P., 2006. Experimental patterns of hammerstone percussion damage on bones: implications for inferences of carcass processing by humans. J. Archaeol. Sci. 33, 459–469. Prescott, J.R., Hutton, J.T., 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: large depths and long-term time variations. Radiat. Meas. 23, 497–500. Quam, R., Bailey, S.E., Wood, B., 2009. Evolution of M1 crown size and cusp proportions in the genus Homo. J. Anat. 214, 655–670. https://doi.org/10.1111/j.14697580.2009.01064.x. Radović, P., Lindal, J., Mihailović, D., Roksandic, M., 2019. The first Neanderthal specimen from Serbia: maxillary first molar from the late Pleistocene of Pešturina cave. J. Hum. Evol. 131, 139–151. https://doi.org/10.1016/j.jhevol.2019.03.018. Richter, J., 2008. Neanderthals in their landscape. In: Demarsin, B., Otte, M. (Eds.), Neanderthals in Europe. Gallo-Roman Museum, Tongeren,, pp. 17–32. Richter, J., 2016. Leave at the height of the party: a critical review of the Middle Paleolithic in Western Central Europe from its beginnings to its rapid decline. Quat. Bar Int. 411 (Part A), 107–128. 18 D. Mihailović et al. Quaternary International 610 (2022) 1–19 Riel-Salvatore, J., Ludeke, I.C., Negrino, F., Holt, B.M., 2013. A Spatial Analysis of the Late Mousterian Levels of Riparo Bombrini (Balzi Rossi, Italy). Canadian Journal of Archaeology/Journal Canadien d’Archéologie 37 (1), 70–92. Rink, W.J., Schwarcz, H.P., Smith, F.H., Radovčić, J., 1995. ESR dates for Krapina hominids. Nature 378(6552) 24. https://doi.org/10.1038/378024a0. Rink, W.J., Mercier, N., Mihailović, D., Morley, M.W., Thompson, J.W., Roksandic, M., 2013. New radiometric ages for the BH-1 hominin from Balanica (Serbia): implications for understanding the role of the Balkans in Middle Pleistocene human evolution. PloS One 8, e54608. Rogers, A.R., 2000. On equifinality in faunal analysis. Am. Antiq. 65 (4), 709–723. Roksandic, M., Mihailović, D., Mercier, N., Dimitrijević, V., Morley, M.W., Rakočević, Z., Mihailović, B., Guibert, P., Babb, J., 2011. A human mandible (BH-1) from the Pleistocene deposits of Mala Balanica cave (Sićevo gorge, Niš, Serbia). J. Hum. Evol. 61, 186–196. Simek, J.F., Smith, F.H., 1997. Chronological changes in stone tool assemblages from Krapina (Croatia). J. Hum. Evol. 32, 561–575. Sitlivy, V., Sobczyk, K., Karkanas, P., Koumouzelis, M., 2008. Middle Palaeolithic industries of Klissoura cave, Greece. In: Darlas, A., Mihailović, D. (Eds.), The Palaeolithic of the Balkans. BAR International Series 1891, Oxford, pp. 39–50. Skinner, M.M., de Vries, D., Gunz, P., Kupczik, K., Klassen, R.P., Hublin, J.-J., Roksandic, M., 2016. A dental perspective on the taxonomic affinity of the Balanica mandible (BH-1). J. Hum. Evol. 93, 63–81. https://doi.org/10.1016/j. jhevol.2016.01.010. Stiner, M.C., 1991. Food procurement and transport by human and non-human predators. J. Archaeol. Sci. 18, 455–482. Stiner, M.C., 2004. A comparison of photon densitometry and computed tomography parameters of bone density in ungulate body part profiles. J. Taphonomy 2 (3), 117–145. Stiner, M.C., Kuhn, S.L., Weiner, S., Bar-Yosef, O., 1995. Differential burning, recrystallization, and fragmentation of archaeological bones. J. Archaeol. Sci. 22, 223–237. Stoops, G., 2003. Guidelines for analysis and description of soil and regolith thin sections. Soil Science Society of America Inc. Terradas X. Discoid flaking method: conception and technological variability. In: Peresani, M. (Eds.), Discoid Lithic Technology: Advances and Implications. Oxford. BAR International Series 1120, Oxford, pp. 19–31. Tillier, A.-M., Sirakov, N., Guadelli, A., Fernandez, P., Sirakova, S., Dimitrova, I., Ferrier, C., Guerin, G., Heidari, M., Krumov, I., Leblanc, J.-C., Miteva, V., Popov, V., Taneva, S., Guadelli, J.-L., 2017. Evidence of Neanderthals in the Balkans: the infant radius from Kozarnika cave (Bulgaria). J. Hum. Evol. 111, 54–62. https://doi.org/ 10.1016/j.jhevol.2017.06.002. Tomović, Lj, Ajtić, R., Ljubisavljević, K., Urošević, A., Jović, D., Krizmanić, I., Labus, N., Đorđević, S., Kalezić, M.L., Vukov, T., Džukić, G., 2014. Reptiles in Serbia: distribution and diversity patterns. Bulletin of the National History Museum in Belgrade 7, 129–158. Tribolo, C., Mercier, N., Valladas, H., 2001. Alpha sensitivity determination in quartzite using an OSL single aliquot procedure. Ancient TL 19 (2), 47–50. Turq, A., 1989. Approche technologique et économique du facies Moustérien de type Quina: etude préliminaire. Bull. Soc. préhist. fr. 86, 244–256. Tzedakis, P.C., 2004. The Balkans as prime glacial refugial territory of European temperate trees. In: Griffiths, H.I., Kryštufek, B., Reed, J.M. (Eds.), Balkan Biodiversity. Springer, Dordrecht, pp. 49–68. Tzedakis, P.C., Bennett, K.D., 1995. Interglacial vegetation succession: a view from southern Europe. Quat. Sci. Rev. 14 (10), 967–982. Valladas, H., Mercier, N., Frogeta, L., Joron, J.-L., Reyssa, J.-L., Karkanas, P., Panagopoulou, E., Kyparissi-Apostolika, N., 2007. TL age-estimates for the Middle Palaeolithic layers at Theopetra cave (Greece). Quat. Geochronol. 2 (1–4), 303–308. https://doi.org/10.1016/j.quageo.2006.03.014. Valoch, K., 1984. Le Taubachien, sa géochronologie, paléoclimatologie et paleoethnologie. L’Anthropologie 88 (2), 193–208. Van Andel, T.H., Tzedakis, P.C., 1996. Palaeolithic landscapes of Europe and environs, 150,000-25,000 years ago: an overview. Quat. Sci. Rev. 15 (5–6), 481–500. Vaquero, M., Romagnoli, F., 2018. Searching for Lazy People: the Significance of Expedient Behavior in the Interpretation of Paleolithic Assemblages. J. Archaeol. Method Theor 25, 334–367. https://doi.org/10.1007/s10816-017-9339-x. Villa, P., Mahieu, E., 1991. Breakage patterns of human long bones. J. Hum. Evol. 21 (1), 27–48. Whallon, R. (Ed.), 2017. Crvena Stijena in Cultural and Ecological Context: Multidisciplinary Archaeological Research in Montenegro. Montenegrin Academy of Sciences and Art. National Museum of Montenegro, Podgorica. Yates, J.A., Velsko, I.M., Aron, F., Posth, C., Hofman, C.A., Austin, R.M., Parker, C.E., Mann, A.E., Nägele, K., Weedman Arthur, K., Arthur, J.W., Bauer, C.C., Crevecoeur, I., Cupillard, C., Curtis, M.C., Dalén, L., Díaz-Zorita Bonilla, M., Díez Fernández-Lomana, J.C., Drucker, D.G., Escribano Escrivá, E., Francken, M., Gibbon, V.E., González Morales, M.R., Grande Mateu, A., Harvati, K., Henry, A.G., Humphrey, L., Menéndez, M., Mihailović, D., Peresani, M., Rodríguez Moroder, S., Roksandic, M., Rougier, H., Sázelová, S., Stock, J.T., Straus, L.G., Svoboda, J., Teßmann, B., Walker, M., Power, R.C., Lewis, C.M., Sankaranarayanan, K., Guschanski, K., Wrangham, R.W., Dewhirst, F.E., Salazar-García, D.C., Krause, J., Herbig, A., Warinner, C., 2021. The evolution and changing ecology of the African hominid oral microbiome. Proc. Natl. Acad. Sci. Unit. States Am. 118 (20) https:// doi.org/10.1073/pnas.2021655118 e2021655118. 19