Journal of Human Evolution 61 (2011) 425e446

Contents lists available at ScienceDirect

Journal of Human Evolution

journal homepage: www.elsevier.com/locate/jhevol

Carcass transport decisions in subsistence strategies

Palmira Saladié a,b,*, Rosa Huguet a,b, Carlos Díez c, Antonio Rodríguez-Hidalgo a,b,d, Isabel Cáceres a,b, Josep Vallverdú a,b, Jordi Rosell a,b, José María Bermúdez de Castro e, Eudald Carbonell a,b,f a IPHES, Institut Català de Paleoecologia Humana i Evolució Social, C/Escorxador s/n, 43003 Tarragona, Spain b Area de Prehistoria, Universitat Rovira i Virgili (URV), Avinguda de Catalunya 35, 43002 Tarragona, Spain c Laboratorio de Prehistoria, IþDþi, Universidad de Burgos, Plaza Misael Bañuelos s/n, 09001 Burgos, Spain d Equipo Primeros Pobladores de Extremadura, Casa de la Cultura Rodríguez Moñino, Avda, Cervantes s/n, 10003 Cáceres, Spain e Centro Nacional de Investigación sobre Evolución Humana (CENIEH), Paseo Sierra de Atapuerca s/n, 09002 Burgos, Spain f Institute of Vertebrate and Paleoanthropology of Beijing, article info abstract

Article history: foragers used several prey acquisition and processing strategies. These strategies and their Received 23 November 2009 associated decisions are elucidated by taphonomic studies that cover transport, modifications by Accepted 26 May 2011 different agents and archaeological remains. Interpretative models of archaeological sites are by necessity based on natural and experimental observations. Ethno-archaeological data shows that several Keywords: factors influenced decisions about carcass transport from the kill site to the home site. These factors often Differential transport have little archaeological visibility. Díez et al. (1999) has previously interpreted the general character- Anatomical profiles istics of the macro- remains from Gran Dolina Level TD6-2 (Sierra de Atapuerca, Burgos, Spain) Archaic Homo fi Zooarchaeology as the result of anthropic accumulation, in which the anatomical pro les appeared to be the result of ’ Hunting selective transport based on the weight. Recent taphonomic analysis has shown that carcasses Pleistocene with different weights may be subject to similar transport strategies, suggesting that other factors influenced these choices. The hominins that occupied TD6-2 (the TD6-2 hominin group), at least sometimes, transported large carcasses to the cave in their entirety, implying participation by groups of individuals in hunting parties. These individuals delayed their consumption of large amounts of food, instead moving it to Gran Dolina, where it was shared with other group members. These decisions are evidence of social cooperation and food sharing amongst early European hominins. Ó 2011 Elsevier Ltd. All rights reserved.

Introduction 1994; Milo, 1998, 2005), c) carnivore gnawing and ravaging (e.g., Binford, 1981; Marean and Spencer, 1991; Marean et al., 1992; It is uncommon to find whole carcasses in archaeological Blumenschine and Marean, 1993; Madrigal and Holt, 2002; assemblages. Anatomical connections are usually scarce, with Munson and Garniewicz, 2003; Marean and Cleghorn, 2003; Faith a high level of broken remains and a bias in the animals’ anatomical and Behrensmeyer, 2006), and/or d) differential destruction of profiles. This may be due to: a) hominins selecting certain elements some bones or their portions by post-depositional processes (e.g., for transport (e.g., Yellen, 1977; Binford, 1978; Bunn and Kroll, 1986; Brain, 1981; Klein and Cruz-Uribe, 1984; Marean, 1991; Lyman, O’Connell et al., 1988, 1990; Gifford-Gonzalez, 1993; Oliver, 1993; 1994). The aforementioned causes can hinder the reconstruction of Monahan, 1998; Faith and Gordon, 2007), b) activities that occur at the original anatomical composition. the home base, e.g., bone cooking, cremation, bone breakage and/or Skeletal-part profiles can help to interpret the type of access to food sharing (e.g., Todd and Rapson, 1988; Blumenschine and animals (Bunn and Ezzo, 1993) and the transport strategy patterns Selvaggio, 1988; Gifford-Gonzalez, 1993; Oliver, 1993; Marshall, used by Pleistocene hominins. Unfortunately, strict behavioral conclusions based on anatomical profiles may only be valid at sites with a simple taphonomic history (Lupo, 1999). Biological and non- * Corresponding author. biological processes subjected the majority of fauna assemblages, E-mail addresses: [email protected] (P. Saladié), [email protected] (R. many of which consisted of multiple events, to destruction and Huguet), [email protected] (C. Díez), [email protected] (A. Rodríguez-Hidalgo), modification after hominin activity (Marean and Cleghorn, 2003). [email protected] (I. Cáceres), [email protected] (J. Vallverdú), jordi.rosell@ urv.cat (J. Rosell), [email protected] (J.M. Bermúdez de Castro), ecarbonell@iphes. Many other circumstances can also affect transport decisions cat (E. Carbonell). and therefore influence different anatomical profiles. The results

0047-2484/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jhevol.2011.05.012 426 P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 obtained by (O’Connell et al., 1988, 1990 and Bunn and colleagues Decisions that are the basis of human behavior may be highly Bunn and Kroll, 1986, 1988; Bunn, 1993), who both presented variable and subject to different determining factors. These opposing interpretations of Hadza carcass transport after studying factors can lead to various results in the archeological record, similar samples obtained in relatively similar time frames, which are sometimes superimposed. These superimpositions can demonstrates this problem. These contradictions are partially due hinder the legibility of archaeological assemblages from specific to their use of different study methods. They did agree on issues events and when combined with the possible involvement of such as the relationship between transport type and animal weight, other agents, makes inferences about subsistence strategies by and the dependence of the final decision about bone transportation prehistoric hominins very difficult. This paper presents on fat and marrow content. Bunn (1993) also found a link between a zooarchaeological perspective on an updated analysis of the several decisions and the anatomy of specific species regardless of remains from Level TD6-2 at Gran Dolina (Sierra de Atapuerca, the animal’s weight. Monahan (1998) reassessed the data pub- Burgos, Spain). Our aim is to interpret Homo antecessor subsis- lished by Bunn and Kroll (1986); O’Connell et al. (1988,1990),ashe tence strategies (access to and transport of prey), and to high- believed that the two data sets were not sufficiently different to light several social factors that may have been involved. A justify contradictory interpretations. Monahan (1998) concluded previous study of TD6-2 (Díez et al., 1999) proposed that the that the Hadza do not transport the carcasses on the basis of the transport decisions made by H. antecessor were related to prey schlepp effect (i.e., tendency to only transport limbs, Perkins and weight. Current data make this hypothesis more complex, and Daly, 1968). Instead, he determined that the transport and pro- enables us to consider whether the selection decisions of early cessing costs, as well as the levels of usefulness of various items, hominins regarding carcass transport may have involved other helped to explain the transport behavior of these groups and factors. variations in their decisions. According to Monahan (1998: 422), the anatomical profiles of the items transported based on animal weight: “.do not reflect a single, uni-modal pattern of carcass The Gran Dolina site and Level TD6-2 transport, but the sum total of different transport modes”. Addi- tionally, the distance between the kill/butchering site and the home Gran Dolina is a cave located in the Railway Trench (Sierra de base (Bunn and Kroll, 1988; O’Connell et al., 1990; Monahan, 1998), Atapuerca, Burgos, Spain). Eleven lithostratigraphic units have been the number of animals to be processed, the number of participants documented in the sedimentary deposits, defined as TD1-TD11 in the expeditions, the location and time of day of carcass acqui- from the base to the top (Parés and Pérez-González, 1999; Pérez- sition (Yellen, 1977; Binford, 1978, 1981, 1984; Bunn and Kroll, 1986, González et al., 2001). See Parés and Pérez-González (1999) and O’Connell et al., 1988, 1990; Bunn, 1993; Gifford-Gonzalez, 1993; Rodríguez et al. (2010) for detailed descriptions of the site Oliver, 1993; Monahan, 1998; Faith et al., 2009) and the risk of lithostratigraphy. predation by other carnivores during the initial processing Magnetic polarity dating has situated Level TD6 in the (Monahan, 1998) are other variables that may have strongly influ- Matuyama Chron, more than 780 thousand years ago (ka) (Parés enced transport decisions. and Pérez-González, 1995, 1999). Combined with ESR and Researchers must consider the relative presence of various uranium series, these results show an age of between 780 and elements and preserved portions in the archaeological assemblages 857 ka (Falguères et al., 1999). The most recent thermo- in order to assess the anatomical profiles. For this purpose, luminescence and infrared stimulated luminescence data suggest researchers mainly use the repetition of bone portions to estimate an age of 960 120 ka (TL) (Berger et al., 2008). the Minimum Number of Elements (MNE) and Minimum Animal The excavation of Level TD6 occurred in two stages. The first Units (MAU) (Binford, 1984). These estimates may sometimes be stage occurred between 1994 and 1996, and the second stage began somewhat imprecise, as some bones (e.g., artiodactyl tibiae and in 2003 and has since continued. The first stage involved the metapodials) have very clear features that facilitate their identifi- excavation of a biostratigraphic section. The removal of the over- cation, while other items such as femora and humeri can lead to hangs above the infilled section, also involving TD6, began in 2003. doubts due to the absence of landmarks on the shafts of many TD6 has a two 2.5 m stratigraphic profiles and includes Layer TD6- portions. Furthermore, anatomical identification does not always 2. The matrix is composed of very fine sands and shale. The lower entail taxonomic identi fication. The correct construction of part of the sequence has a chaotic appearance while the top (the anatomical and taxonomic profiles is an essential tool in all last 50 cm) is well stratified. On top of the latter lies a layer about zooarchaeological analyses, as it is used as a basis for all inferences 20e25 cm thick. This layer consists of red shale with an abundant concerning assemblages (Pickering et al., 2003). Incorrect or number of massive limestone blocks, known as ‘Aurora stratum’. deficient identification can lead to erroneous or partial conclusions. The aforementioned layer now forms part of TD6-2 (Parés and To solve this problem, size-weight categories are established to Pérez-González, 1999; Pérez-González et al., 2001; Vallverdú group species of similar sizes, since in many cases, attribution to et al., 2001; Bérmudez de Castro et al., 2008). a specific animal species is impossible (Bunn et al., 1980; Brain, Abundant faunal and hominin remains have been discovered on 1981). Pleistocene hominins may have used various transport Level TD6-2, as well as more than 800 lithic artifacts made from strategies and variability in decision-making. The influence of chert, quartzite, sandstone, quartz and limestone. The hominin events and possible destructive processes show that skeletal data in belong to H. antecessor (Bermúdez de Castro et al., 1997). The themselves are insufficient to ascertain which of these strategies Journal of Human Evolution published a monograph on Level TD6 Pleistocene hominins employed (Marean and Spencer, 1991; in 1999. This volume included a taphonomic and zooarchaeological Marean et al., 1992; Monahan, 1998; Marean and Cleghorn, 2003; study of the fauna and hominin remains from the ‘Aurora stratum’ Faith and Behrensmeyer, 2006; Faith et al., 2009). This conclusion (Díez et al., 1999; Fernández-Jalvo et al., 1999). This study docu- suggests that the study of bone modifications to assess the role and mented the differential transport of carcasses. The absence of axial sequence of access by different agents (mainly carnivores versus skeletons of larger animals led to the conclusion that the main hominins) is important (e.g., Bunn, 1981; Blumenschine and limitation on transport was weight. In this paper, we have studied Selvaggio, 1988; Selvaggio, 1994a; Blumenschine, 1995; Capaldo, a considerably more extensive collection of bones than in previous 1995, 1997, 1998; Domínguez-Rodrigo and Barba, 2006; works. This increase in bone number has enabled us to broaden the Domínguez-Rodrigo et al., 2009). scope of the data and interpretations. P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 427

Methods in conjunction with a conceptual numerical system. In this system, we divided the bones into five fixed portions. In the long bones, For this study, we constructed and analyzed anatomical profiles portions 1 and 5 were the proximal and distal epiphysis, respec- and bone surfaces of the carcasses, with a special focus on modi- tively. Portions 2, 3 and 4 were in the diaphysis, in which portion 2 fications during the nutritional phase (Capaldo, 1997). We only was the most proximal part, portion 3 was the medial region, and included remains larger than 1 cm (since smaller sized bones portion 4 was the distal diaphysis. For flat bones, which have cannot be identified). We relied on modification analysis to include a different anatomical structure, the division in portions was breakage of long bone fragments measuring more than 2 cm. similar with portion1 being the most proximal part of the bone, and TD6-2 contains a significant volume of remains that are not portion 5 the most distal. We considered the preserved side(s) of taxonomically identifiable. To include the remains with the iden- the bone as well as the portions. When a portion or side retained tified specimens, we established four weight categories, following half or less than half of the area, we displayed them in brackets. Huguet et al. (1999) and the weights defined by Rodríguez (1997) When there was more than one portion or side, we included a plus for the various species found at the Sierra de Atapuerca sites. We sign in the expression (Fig. 1). analyzed this distribution using prior knowledge of the taxonomic The repetition of regions and sides (considering the presence/ representation of the studied assemblages. We also took into absence of landmarks) yielded the MNE. We included instances of account the animals’ age of death. The four weight categories are isolated teeth when calculating the MNE of maxillas and mandibles. shown in Table 1, and include: a) Very large-sized animals weighing We show the age of the items whenever possible, based on the more than 1000 kg, roughly equivalent to Size 6 in Africanist fusion of epiphyses and dental eruption/replacement patterns and methodologies (Bunn, 1982; Bunn and Ezzo, 1993), b) large-sized wear. We established four age groups: infant, sub-adults, adults and animals weighing between 300 and 1000 kg, equivalent to Sizes old. Whenever possible, we used tooth characteristics in calcula- 4 and 5, c) medium-sized animals weighing between 100 and tions of the MNI (Minimum Number of Individuals), as teeth 300 kg, equivalent to size 3, and d) small-sized animals weighed provide the most information of MNI. Infants are animals with between 10 and 100 kg, size 1B. A significant number of the fauna deciduous teeth. Animals classified as sub-adults had lightly to remains could not be included in any of the groups (see Results). moderately worn deciduous teeth and erupted M1s and M2s. We described the anatomically unidentifiable fragments according Permanent teeth had no or very slight wear. These animals were to their morphology as: 1) long bones, including those with two close to adult body size. Adults had moderate wear on all perma- epiphyses, a diaphysis and a marrow cavity (humeri, radii/ulnae, nent teeth. Lastly, we classified teeth as old in cases of very femora, fibulae, tibiae, and phalanges), 2) flat bones, in which the advanced wear on most of the crown (less than half the crown marrow cavity was small or non-existent and the distance between remained). For H. antecessor, we considered the results provided by the two cortical surfaces was narrow (cranial fragments, mandibles, (Bermúdez de Castro et al., 2006, 2008, 2010), who also used scapulae, clavicles, ribs, coxae and vertebral laminae and apoph- information from remains of jaws and isolated teeth. yses), and 3) articular bones with abundant spongy tissue (carpals, We utilized the %MAU to standardize anatomical profiles and tarsals and sesamoids). Additionally, we also included some compare frequencies of the various elements (Binford, 1984; epiphyseal fragments and unidentified vertebral bodies in category Lyman, 2008). Here, we produced a correlation coefficient 3 because of their morphological characteristics. between %MAU and bone mineral density, taking into account the Data described for each specimen were: element, taxon, size, MNE of each portion of the elements. We used the following data position, age, portion, side and landmarks. Following Bunn (1983) for correlation: sheep for small animals (Lyman, 1994), caribou for and Villa and Mahieu (1991), we also listed shaft circumference medium-sized animals, blue wildebeest for large animals (Lam and shaft length along with fracture outline, angle and edge. et al., 1999), and data from Suby (2006) for H. antecessor.We We established the MNE by interpreting the frequency of each subsequently correlated the %MAU with the Standard Food Utility bone within the assemblage using portions, sides and position of Index (SFUI) proposed by Metcalfe and Jones (1988), the Utility one of the elements. Several methods provide estimates of MNE Index from Emerson (1993), and the Unsaturated Marrow Index (see e.g., Bunn et al., 1988; Marean and Spencer, 1991; Stiner, 1994), (UMI) from Morin (2007). We employed the methods of Marean although they are all based on the most common repetition of et al. (1992; Marean and Cleghorn, 2003; Cleghorn and Marean, specific bone regions. In this study, we considered bone landmarks 2004) to produce the various correlation coefficients. Finally, to complete this data set, we calculated the Shannon evenness index Table 1 on the basis of proposals by Faith and Gordon (2007). Tests con- Distribution of taxa of TD6-2 level by size weight categories. ducted by these researchers show that correlation analysis may be highly sensitive to interactions between sample size and the Weight size Taxa fi e uniformity of the distribution of skeletal items. Non-signi cant Small-sized (10 100 kg) H. antecessor fl Dama “nestii” vallonetensis correlations may re ect artifacts even where sample sizes are Infantile Cervidae indet. small (Faith and Gordon, 2007). They suggested that a more or less Sus scrofa uniform representation of the skeletal items in a series reflected mosbachensis transport and butchering strategies. These studies suggest that the Vulpes praeglacialis Shannon evenness index can be successfully applied as a quantita- Lynx sp. Cercopithecidae tive means of distinguishing assemblages with different transport Medium-sized (100e300 kg) Adult and juvenile Cervus elaphus strategies (Faith and Gordon, 2007). Adult Crocuta crocuta The interpretation of abundance or absence of elements may not Adult and juvenile Ursus dolinensis reflect the assemblages resulting from discard by hominins. Result Large-sized (300e1000 kg) Juvenile and infantile Stephanorinus etruscus Equidae (Stenionan) interpretations may thus be subject to taphonomic analysis of the cf. Bison voigstedtensis assemblages (Lyman, 1993, 1994). We utilized bone surface modi- Eucladoceros giulii fications to assess the significance of different actors (as per the Infantile Mammuthus sp. nomenclature in Gifford-Gonzalez, 1991) in archaeological assem- Very large-sized (>1000 kg) Adult S. etruscus blages. In the TD6-2 assemblage, we inspected the complete surface Adult Mammuthus sp. of all bones, teeth and antlers macroscopically and microscopically 428 P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446

Figure 1. Example of nomenclature used in the description of portions and sides present in specimens. Portion 1: proximal epiphysis; Portion 2: proximal diaphysis; Portion 3: mid diaphysis; Portion 4: distal diaphysis; Portion 5: distal epiphysis. Brackets indicate that half or less than half of the portion is preserved. þ indicates the presence of more than one portion or side. (Bone drawings from Pales and Lambert, 1971).

(OPTHEC 120 Hz model), recording distribution and incidence of TD6-2: skinning, dismembering and/or disarticulation, defleshing, modifications. We then recorded hominin- and carnivore-induced evisceration, periosteum removal, and possible tendon removal. damage in terms of the anatomical area and the region (portion We also analyzed surface modifications during anthropic and side) of the modifications. breakage of bones. We recorded these results in terms of the We detected three types of cut marks: slicing or incisions, presence or absence of each modification in the analyzed remains. scraping and chop marks. The three types differ according to the These modifications included: percussion pits (Blumenschine and manner in which hominins used stone tools (Shipman and Rose, Selvaggio, 1988), conchoidal scars and flakes (Díez et al., 1999; 1983; Blumenschine et al., 1996; Lyman, 2008). Tools used to Fernández-Jalvo et al., 1999) and peeling (White, 1992). Descrip- apply force parallel to the long axis of the tool edge caused tions included the location of the damage on the remains. We slicing marks. Perpendicular application created scraping marks, assumed that conchoidal scars and flakes were of anthropic origin resulting in various marks in a single direction. This type of when associated with percussion pits. marking is often related to periosteum removal, although it may We established the following types of damage for tooth marks: also arise from the extraction of meat remnants attached to the scores, pits, punctures and perforations (Maguire et al., 1980; bone (Domínguez-Rodrigo, 2002), as surface preparation for Binford, 1981). We took measurements of pits, punctures and breakage is only strictly necessary in metapodials (Domínguez- perforations (length and width) into account, as well as tissue Rodrigo, 1997a). Finally, application of a dynamic and/or location, the most accurate method for establishing actor size percussive force with an edged tool caused chop marks (Lyman, (Selvaggio, 1994b; Selvaggio and Wilder, 2001; Domínguez-Rodrigo 2008). and Piqueras, 2003), and the type of tissue, bone and weight range of By taking their location, position and morphology into account, the element where these manipulations took place (Delaney-Rivera we associated the different types of cut marks with specific et al., 2009; Saladié, 2009). We noted tooth mark distribution on butchering activities. We applied observations by Binford (1981, bone portions, following Egeland et al. (2008). We then calculated 1984), Fisher (1995) and Nilssen (2000, from Pobiner et al., 2008) a NISP portion by counting the portions present on each specimen. in our designations. We complemented these contributions with Additionally, we included the presence of other modifications such observations from a series of experiments by the IPHES team (e.g., as licking, pitting, scooping-out and digestion (Binford, 1981). Ollé et al., 2010) using , wild goats, lions and sheep. Additionally, we employed White’s (1992) contributions for Results H. antecessor, and observations by the same IPHES team during the butchering of a chimpanzee. We conducted all experiments using In this study, we analyzed 4411 faunal remains. This assemblage stone tools. We detected the following butchering activities in included Level TD6-2 remains from the 1994e1996 and the P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 429

Table 2 of taxa that are less represented by the NISP (Lyman, 2008). Bearing Taxonomic determination, NISP, MNI and age category of the fauna from the TD6-2 this in mind, there is no doubt that if we used both the NISP and level of Gran Dolina. MNI as the basis, there would be more herbivores and hominins in Taxon NISP MNI Infantile Sub-adult Adult Old TD6-2 than carnivores. H. antecessor 156 11 8 2 1 e The estimated age of death of individuals suggested a presence Eucladoceros giulii 72 e 11e of all the age groups (Table 2). Immature individuals and adults Dama nestii vallonetensis 20 2 e 2 ee e were more common, while there were few old individuals. It Cervus elaphus cf. acoronatus 293 4 121 fi Cervidae indet. 271 2 2 eee should be noted that the old age group may be arti cially small, as cf. Bison voigtstedtensis 119 4 e 121 it is generally only visible by means of particular dental items. Equus (Stenonian) 59 3 e 12e There is a greater presence of sub-adult and infantile animals Stephanorhinus etruscus 45 2 1 1 ee amongst the hominins and herbivores, which is accentuated by Sus scrofa 11 ee1 e the presence of juveniles among H. antecessor. The age distribution Cercopithecidae 2 1 ee1 e Mammuthus sp. 1 1 1 eee of carnivores significantly increased the presence of adults in the Canis mosbachensis 17 1 e 1 ee assemblage. Vulpes praeglacialis 71 e 1 ee The identification frequency in TD6-2 was 23.1%. For an overall ee e ee indet. 1 assessment of the assemblage, we grouped the unidentified Ursus dolinensis 91 e 1 ee Crocuta crocuta 31 e 1 ee remains into weight categories (Table 3). The results show an Lynx sp. 4 1 e 1 ee abundance of small and medium-sized animals, including homi- indet. 15 ee e ee nins, while the MNI results show there may have been more small- size animals. Total 1030 38 12 14 10 2 Most of the remains analyzed were less than 40 mm long (Fig. 2). The presence of flat (n ¼ 1221) and long (n ¼ 1321) bone fragments was similar. Amongst the long bones, the assemblage 2003e2007 excavations. The taxonomic representation of the TD6- consisted of mainly bone shaft fragments (n ¼ 1176), which were 2 assemblage was high (see Table 2), with 16 identified species from not always identifiable taxonomically (Table 4). 207 teeth were not taxa at different levels of the food chain (Bermúdez de Castro et al., attached to bones. 1997; García and Arsuaga, 1999, 2001; Made, 2001; García, 2003). Most of the diaphysis studied retained less than a quarter of Deer, followed by H. antecessor, were the dominant groups their length and a third of their original circumference (71.9%). according to the NISP and MNI (Table 2). Species identified as deer These measurements were consistent with the results yielded by were in three different weight ranges. We identified five carnivore the fragmentation index. The breakage analysis showed a preva- species from the few remains, with a MNI equal to one individual lence of green breakages. Curved/V-shaped delineations were more for each species. The MNI may generally exaggerate the importance numerous than transversal delineations (70.3% versus 29.7%,

Table 3 NISP and MNE (Minimal Number of Elements) grouped by size weight category and H. antecessor remains.

NISP (MNE) Very large-sized Large-sized (300 Medium-sized (100 Small-sized H. antecessor Indeterminate Total (þ1000 kg) e1000 kg) e300 kg) (100 kg) (100 kg) Antler/Horn e (e)4(e)55(e)2(e) e (e) 160 (e) 221 (e) Cranium 2 (2) 20 (2) 58 (5) 62 (5) 21 (3) 12 (e) 175 (17) Maxilla e (1) 3 (6) 4 (3) e (e) 4 (5) e (e) 11 (15) Mandible 6 (2) 8 (6) 11 (7) 10 (2) 5 (5) 1 (e) 41 (22) Maxilla/ e (e)1(e)2(e) e (e) e (e) e (e)3(e) Mandible Isolated Tooth 13(e)71(e)52(e)15(e)24(e)31(e) 206 (e) Hyoid e (e) 1 (1) e (e) 1 (1) e (e) e (e) 2 (2) Vertebra 2(2) 17 (7) 49 (7) 63 (11) 19 (14) 9 (e) 159 (41) Clavicle e (e) e (e) e (e) e (e) 3 (3) e (e) 3 (3) Rib 4 (1) 72 (11) 89 (8) 129 (18) 31 (14) 16 (e) 341 (52) Coxa 1 (1) 2 (1) 7 (1) 5 (5) 2 (1) 1 (e) 18 (9) Scapula e (e) 4 (2) 6 (2) 9 (6) 3 (3) 1 (e) 23 (13) Humerus 1 (1) 12 (6) 34 (7) 9 (2) 3 (3) e (e) 59 (19) Radius e (e) 19 (4) 26 (7) 14 (7) 2 (2) e (e) 61 (20) Ulna 2 (1) 6 (1) 8 (3) 4 (4) 2 (2) e (e) 22 (11) Carpal 1 (1) 9 (9) 1 (1) 6 (6) 5 (5) e (e) 22 (22) Femur 1 (1) 15 (4) 30 (5) 25 (12) 4 (2) e (e) 75 (24) Patella e (e) e (e) e (e) 2 (2) 2 (2) e (e) 4 (4) Tibia e (e) 31 (11) 26 (7) 10 (6) 2 (2) e (e) 69 (26) Fibula e (e) e (e) 2 (2) e (e) 2 (2) e (e) 4 (4) Talus e (e) 2 (2) 2 (2) 2 (2) e (e) e (e) 6 (6) Calcaneus e (e) 2 (2) e (e) e (e) 1 (1) e (e) 3 (3) Tarsal 1 (1) 1 (1) 4 (4) 3 (3) e (e) e (e) 9 (9) Metapodial e (e) 44 (12) 42 (12) 27 (10) 6 (5) 3 (e) 122 (39) Phalanx 1 (1) 15 (10) 13 (13) 19 (17) 24 (18) e (e) 72 (59) Long bone 1 (e) 181 (e) 461 (e) 186 (e) e (e)9(e) 838 (e) Flat bone 4 (e) 118 (e) 198 (e)95(e) e (e)9(e) 424 (e) Articular bone 2 (e)14(e)10(e)10(e) e (e)3(e)39(e) Indeterminate 1 (e)58(e)3(e)5(e) e (e) 1312 (e) 1379 (e)

Total 43 (15) 730 (98) 1,193 (96) 713 (119) 165 (92) 1,567 (e) 4,411 (419) 430 P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446

0 4 0

5 3 5

3 0

5 2 5

0 2 0

5 1 5 Percentage

0 1 0

5

0 2 < 2 0 2 0 -3 0 3 0 -4 0 4 0 -5 0 5 0 -6 0 6 0 -7 0 7 0 -8 0 8 0 -9 0 9 0 -1 0 0 1 0 0 -1 1 0 1 1 0 -1 2 0 1 2 0 -1 3 0 > 1 3 0 Measurements

Figure 2. Bar diagram showing sizes of specimens in TD6-2, grouped in 10 mm intervals. respectively). Likewise, most breaks presented an oblique angle possible reasons: 1) the result of various decisions by hominins (71.8%), followed by right angles (19.4%) and mixed angles (8.8%). when capturing and transporting carcasses, 2) carnivore ravaging of scraps left behind by hominins, 3) the involvement of post- Anatomical representation depositional processes in the destruction of portions with a lower mineral density, and/or 4) the combination of several of these In the three main weight categories (large, medium and small), factors. In order to provide an explanation, it is important to consider the long bones and crania were more common and the presence of the biological and non-biological modifications of the bone surfaces. the post-cranial axial skeleton was more limited, as in the case of Post-depositional modifications were generally scarce in TD6-2, phalanges and carpal/tarsal bones (Table 5 and Fig. 3). Broadly except for black stains from manganese oxide deposits found in speaking, based on the criteria of Marean et al. (1992; Marean and 33.7% of the remains and the cemented sediment attached to the Cleghorn, 2003; Cleghorn and Marean, 2004) the anatomical specimen surface, found in 28.7% of the remains. These modifica- profile of TD6-2 showed a greater presence of high-survival tions indicate that the cave was relatively damp, followed by dry elements (long bone shaft fragments, skulls and mandibles) than periods, although the minimal presence of surface cracking (1.3%) low-survival elements. However, the results show that not all low- suggests that this dampness was more or less constant with no survival elements (post-cranial axial elements, scapulae, phalanges sudden changes. and compact bones) were under-represented, with high propor- Sub-aerial weathering (stage 1 according to Behrensmeyer, tions of scapulae and coxae amongst small animals, while the % 1978) was present in 5.6% of the remains. Open-air modifications MAU of the talus for medium-sized animals was 40% (Table 5). may indicate some connections with the area outside the cave. A The correlation between bone mineral density and %MAU very specific area of the excavation accounted for the location of fi (large-sized, rs ¼ 0.318, p ¼ 0.002; medium-sized, rs ¼0.188, these modi cations (north of the currently excavated area). We also p ¼ 0.237; small-sized, rs ¼ 0.142, p ¼ 0.230; H. antecessor, found specimens with root marks (1% of the total) in this area. The fi rs ¼ 0.365, p ¼ 0.148) showed a positive and statistically significant combination of both modi cations in the same area supports the relationship, albeit slight, for the large weight category. The results hypothesis of the entry of light into the cave and its possible link to for the medium, small weight categories and H. antecessor were not the outside. We found evidence of trampling (8.2%) in the form of significant. This suggests a slight differential attrition of some random striations. Rounding and polishing (6.3%) was mild in most portions of the elements. This inference holds true provided that cases, and only affected bone edges. Finally, surface soil corrosion we assume the original assemblage consisted of entire skeletons, affected 3.1% of specimens. which is difficult to prove in most archaeological assemblages (Lam These percentages indicate that the assemblage was well- and Pearson, 2005). preserved. Although the formation of the final assemblage As a result, although the anatomical profile of TD6-2 was not involved post-depositional processes, these were not significant entirely precise, there was a lower presence of the low-survival enough to have played a determinant role in the construction of the elements (Marean et al., 1992; Marean and Cleghorn, 2003; resulting anatomical profile. Cleghorn and Marean, 2004; Faith and Gordon, 2007). High The following sections describe both the human and carnivore- survival elements were most common in this assemblage for several induced modifications to assess their role in the formation of TD6-2.

Table 4 Conserved portions for the main long bones (humerus, radius/ulna, femur, tibia and metapodials) divided by weight sizes.

Proximal Epiph. Prox. þ shaft Shaft Shaft þ dist. Epiph. Distal Epiph. Epiph. Whole Total Very large-sized ee5 (0.4%) eeee5 (0.4%) Large-sized 5 (0.4%) 3 (0.2%) 285 (22.9%) 5 (0.4%) 4 (0.3%) 5 (0.4%) e 307 (24.6%) Medium-sized 8 (0.6%) 3 (0.2%) 606 (48.6%) 2 (0.2%) 6 (0.6%) 2 (0.2%) e 627 (50.3%) Small-sized 10 (0.8) 3 (0.2%) 254 (20.4%) 2 (0.2%) 5 (0.4%) e 1 (0.1%) 275 (22.1%) Indeterminate 1 (0.1%) 3 (0.2%) 14 (1.1%) 1 (0.1%) 3 (0.2%) e 22 (1.7%) H. antecessor ee12 (1%) eeee12 (1%)

Total 24 (1.9%) 12 (0.9%) 1176 (94.3%) 10 (0.8%) 18 (1.4%) 7 (0.6%) 1 (0.1%) 1247

Prox. ¼ proximal; Dist. ¼ distal; Epiph. ¼ epiphyses. P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 431

Table 5 We based the detailed description of human induced modifi- %MAU values according to weight sizes groups and H. antecessor. cations on the segment divisions mentioned in the Methods %MAU Large-sized Medium-sized Small-sized H. antecessor section. Cranium 100.0 100.0 83.3 100.0 Table 6 provides a summary of skeletal items and weight Mandible 50.0 70.0 16.7 50.0 categories. Vertebra 3.5 4.2 5.6 8.5 Cranium Twenty-two cranial remains preserving cut marks were Rib 7.6 6.7 12.5 11.7 from animals of all sizes, as well as H. antecessor. These remains Coxa 8.3 10.0 41.7 10.0 Scapula 16.7 20.0 50.0 30.0 showed mainly slice marks, although there was also a scrape Humerus 50.0 70.0 16.7 30.0 mark on a H. antecessor maxilla. Carcass skinning accounted for Radius 33.3 70.0 58.3 20.0 the associated marks on the zygomatics, occipitals, parietals and Ulna 8.3 30.0 33.3 20.0 nasal bones. Nilssen (2000, from Pobiner et al., 2008) suggests Carpals 12.5 1.7 8.3 8.3 that cut marks on occipitals may be related to cranium Femur 33.3 50.0 100.0 20.0 Patella 0.0 0.0 16.7 20.0 extraction. In TD6-2, there was an occipital from a medium- Tibia 91.7 70.0 50.0 20.0 sized animal with incisions that related to skin removal. The Fibula 0.0 20.0 0.0 20.0 longitudinal direction of these marks indicates that they were Talus 16.7 40.0 0.0 0.0 not produced to separate the trunk and cranium. The TD6-2 Calcaneus 16.7 20.0 16.7 10.0 Tarsal 1.7 8.0 5.0 0.0 hominin group made incisions in the temporal bone and Metapodial 50.0 60.0 41.7 5.0 zygomatic arch of H. antecessor. These marks may indicate the Phalanx 6.9 10.8 11.8 6.7 scalp and ear removal (White, 1992) although the location of the marks on muscle insertion has related mostly with defleshing. In the maxillae, the TD6-2 hominin group made fl Human induced modifications these marks during skinning and de eshing. Marks made during flesh removal from the mandible of Modifications by hominins included cut marks and bone medium-size animals and H. antecessor were apparent. On the breakage. 13.2% of the remains preserved some type of butchering ramus, the TD6-2 hominin group caused marks during skinning, fl marks, either cut marks, bone breakage, or both. These modifica- de eshing and disarticulation of the jaw and cranium. We located fl tions affected many of the taxa documented on the site, and all the de eshing marks on the ventral edge of a mandible of weight categories. The taxonomic groups containing the largest a medium-size animal. Disarticulation of a deer jaw showed number of specimens with anthropic modifications were deer transversal incisions on the external side beside the mandibular fl (n ¼ 106) and H. antecessor (n ¼ 80). notch. We located de eshing marks on the lingual side of the We found cut marks on 10% of the recovered remains. The taxa mandibular body and their arrangement, oblique to the long axis of fl affected were H. antecessor, Eucladoceros giulii, Dama nestii vallone- the bone, suggests tongue removal. Evidence of de eshing is seen tensis, Cervus elaphus, cf. Bison voigtstedtensis, Equus sp., Stephanor- on two mandibles of H. antecessor. A left hemi-mandible displays inus etruscus, Cercopithecidae, Ursus dolinensis and Vulpes a single incision near the oblique line and a group of incisions in the praeglacialis. All anatomical segments presented cut marks, ascending ramus. A second mandible displays an oblique incision including the axial segment. Anatomically unidentified long bone on the ventral edge, by the M1. fragments were amongst the items with cut marks. We found slicing Trunk We found cut marks on the vertebrae, clavicles, ribs and os marks on 96.6% of remains with cut marks (Fig. 4). We also observed coxae. On the vertebrae, we mainly located cut marks on the scraping marks and chop marks, albeit less frequently. Occasional surface and at the base of the spinous process, which likely fl isolated specimens presented several types of cut marks. occurred during esh removal from the carcasses. Marks present on We identified several butchering activities in TD6-2 including: cranial and caudal articular facets in other cases suggests that they skinning, defleshing, disarticulation, dismembering, evisceration, occurred during disarticulation. In one case, we located cut marks periosteum removal, and possible tendon removal. transversely on the dorsal side of a rhinoceros atlas fragment. There

Figure 3. %MAU distribution by element and weight categories, and H. antecessor. 432 P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446

Figure 4. Cut marks on specimens from TD6-2 level. a) Slice marks on a H. antecessor tibial shaft fragment. b) Chop marks on the external face of rib of small-sized animal. c) Slice marks on a long bone shaft from a large-sized animal. d) Cut marks on an Equus phalanx. These marks were associated with carcass skinning. e) Cut marks on a H. antecessor phalanx. Cause unknown. is no doubt in this case that the TD6-2 hominin group separated the marks occurred during the dismembering of the upper extremity cranium from the neck. and trunk of H. antecessor. Three H. antecessor clavicles showed cut marks in the central Ribs were the trunk elements with the largest number of part of the diaphysis, on the inferior side, suggesting processing of specimens bearing cut marks (n ¼ 62). We located cut marks in the the pectoralis major muscle, according to White (1992). In one case, neck area and in the diaphysis. It was not always possible to relate there were scrape marks. Their location on a bone lacking the the marks to specific activities. Nevertheless, we identified disar- marrow cavity and the absence of breakage in this element shows ticulation, defleshing and viscera removal. We located incisions that scraping took place during the defleshing of the bone, and not made during disarticulation in the neck area. These incisions were during the removal of the periosteum. The presence of transverse transverse in four cases. We identified flesh removal as the most incisions on the acromial extremity suggests that perhaps the frequent butchering activity on ribs (Fig. 4b), located in various P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 433

Table 6 Number of remains with cut marks, location, morphology and activity deduced by elements and weight size category and H. antecessor.

Element Size Weight Category N Cm Location Cm-m Activity Cranium Large-sized 1 Zygomatic Sl Sk Medium-sized 4 Occipital and parietals Sl Sk Small-sized 3 Nassal and endocranial bones Sl Sk H. antecessor 6 Temporal and zygomatic Sl Sk þ Df Maxilla Medium-sized 1 Above molars Sl Sk H. antecessor 1 Below orbital bone Sl/Sc Df Mandible Very large-sized 2 Below molars Sl Df Large-sized 1 Below molars/premolars Sl Df Medium-sized 2 Ramus, lateral side Sl Sk þ Da Small-sized 1 Ramus and notch, lateral side Sl Da H. antecessor 2 Below molars and ramus Sl Df Vertebra Very large-sized 2 Above articular facets and transverse processes Sl Df þ Da Large-sized 2 Spinous process Sl Df Medium-sized 6 Laminae, spinous process and above articular facets Sl Df þ Da Small-sized 1 Laminae Sl Df H. antecessor 5 Laminae, spinous process and above articular facets Sl Df þ Da Clavicle H. antecessor 3 Shaft and ends Sl/Sc Df þ Da Rib Very large-sized 2 Angle and shaft, external side Sl/Sc Df þ Da Large-sized 24 Neck and shaft, ventral and external sides Sl Df þ Da þ Ev Medium-sized 20 Neck and shaft, ventral and external sides Sl Df þ Da þ Ev Small-sized 12 Shaft, ventral and external sides Sl Df þ Ev H. antecessor 12 Neck and shaft, ventral and external sides Sl Df þ Da þ Ev Coxa Medium-sized 1 Inferior edge of ilium Sl/Sc Df Small-sized 1 Inferior edge of ilium Sl Df H. antecessor 1 Edge of acetabular fossa Sl Da Scapula Large-sized 2 Neck, borders and ventral surface Sl Df þ Dm Small-sized 2 Neck, and dorsal and ventral surfaces Sl Df þ Dm H. antecessor 1 Neck Sl Dm Humerus Large-sized 7 Medial and distal shaft, anterior, lateral and medial sides Sl Df Medium-sized 15 Proximal, medial and distal shaft, all sides Sl/Sc Df þ Da þ Pr Small-sized 5 Medial and distal shaft, posterior, lateral and medial sides Sl Df þ Da H. antecessor 1 Medial and distal shaft, anterior and posterior sides Sl Df Radius Large-sized 7 Medial and distal shaft, some sides Sl/Sc Df Medium-sized 9 Proximal, medial and distal shaft, all sides Sl Df þ Da Small-sized 5 Medial and distal shaft, anterior, lateral and medial sides Sl/Sc Df þ Da H. antecessor 2 Distal shaft, anterior and posterior sides Sl/S/Ch Df þ Da þ Pr Ulna Very large-sized 2 Trochlear notch Sl/Sc Da Large-sized 1 Medial shaft, posterior side Sl Df Medium-sized 2 Medial shaft and distal end, posterior side Sl Da or Sk H. antecessor 2 Medial and distal shaft, posterior side Sl/Sc Df þ Pr Femur Very large-sized 1 Distal shaft, posterior side Sl/Sc Df Large-sized 6 Proximal, medial and distal shaft, and proximal end Sl Df þ Da Posterior, lateral and medial sides Medium-sized 9 Proximal, medial and distal shaft, all sides Sl Df Small-sized 6 Proximal, medial and distal shaft, and proximal end Sl Df Posterior, lateral and medial sides H. antecessor 4 Proximal, medial and distal shaft, all sides Sl/Sc Df þ Pr Tibia Large-sized 13 Proximal, medial and distal shaft, all sides Sl Df Medium-sized 9 Proximal, medial and distal shaft, all sides Sl Df/Da or Sk Small-sized 3 Proximal and medial shaft and proximal end Sl Df Posterior, lateral and medial sides H. antecessor 2 Distal side, anterior side Sl Df Fibula Medium-sized 1 Medial shaft, lateral side Sl Df H. antecessor 1 Proximal shaft, lateral side Sl Df Hamate Medium-sized 1 Lateral side Sl Sk Sesamoid Medium-sized 1 Lateral side Ch Sk Metapodial Large-sized 12 Proximal, medial and distal shaft, all sides Sl Sk Medium-sized 6 Medial and distal shaft, all sides Sl/Ch Sk Small-sized 3 Proximal and medial, lateral and medial sides Sl Sk H. antecessor 3 Medial shaft, lateral side and proximal end, anterior side Sl Df þ Da Phalanx Large-sized 3 Proximal shaft and end, anterior side Sl Sk Medium-sized 2 Shaft medial, anterior side Sl Sk Small-sized 3 Proximal, medial and distal shaft, anterior and palmar sides Sl Sk or Da H. antecessor 4 Proximal, medial and distal shaft, anterior and palmar sides Sl Tr or df cm-m ¼ Cut marks morphology; Sl ¼ Slice marks; Sc ¼ Scrape marks; Ch ¼ Chop marks. Activity: Sk ¼ Skinning; Df ¼ Defleshing; Da ¼ Disarticulation; Dm ¼ Dismembering; Ev ¼ Evisceration; Pr ¼ Periosteum removal; Tr ¼ Tendon removal. regions of the diaphysis but in all cases on the dorsal side. We animals made during flesh removal. Cut marks on the acetabular identified evisceration on the ventral side of 12 specimens. Note edge of an os coxa from H. antecessor undoubtedly occurred during that we documented this part of the butchering process in ribs from the separation of the trunk from the lower extremity. large, medium and small sized animals, as well as H. antecessor ribs. Forelimbs In the forelimbs (upper limbs for hominins), we located In the pelvis, the TD6-2 hominin group probably made incisions the cut marks on the scapulae, humeri, and radii/ulnae. In the on the lower edge of the iliac crest in medium and small sized scapulae, we located the marks transversely on the neck in three 434 P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446

Figure 5. a) Maxilla of H. antecessor showing a percussion pit. b) Cervid radius with extractions caused by percussion breakage. c) Conchoidal flake of a long bone from a large-sized animal. d) Adhered flake in a long bone caused by anthropic breakage. e) Peeling on a rib of a large-sized animal. f) Peeling on a rib of H. antecessor.

cases, and near the glenoid fossa, indicating the separation of the from two different activities. As mentioned above, scraping occurs humerus and scapula. The location of cut marks in other areas of when a blade is dragged perpendicular to the bone surface. This the scapular body suggested that they occurred during defleshing. activity may occur during flesh removal, or when the periosteum is In the humeri, radii and ulnae, we generally located the cut removed during surface preparation for breakage. The two activi- marks in the medial areas of the diaphysis, which were in most ties could not be distinguished in a majority of the remains. cases oblique with the occasional lengthwise mark. Generally, these Scraping related to the removal of flesh remains the most likely cut marks occurred during flesh removal from carcasses, and explanation in cases where the scraping marks occurred on a broad involved incisions. The scraping located in these areas resulted surface, with no signs of breakage in the adjacent area. Where

Table 7 Elements with anthropic breakage.

Percussion breakage Peeling

Large-sized Medium-sized Small-sized H. antecessor Large-sized Medium-sized Small-sized H. antecessor Crania ee e5 (23.8%) ee 2 (3.2%) 2 (9.5%) Mandibles ee e1 (20%) ee ee Vertebrae ee 1 (1.6%) 1 (5.3%) e 5 (10.2%) 4 (6.3%) 5 (26.3%) Ribs 4 (5.6%) 1 (1.1%) 1 (0.8%) 1 (3.2%) 9 (12.5%) 13 (14.6%) 23 (17.8%) 9 (29%) Coxae ee ee ee ee Scapulae ee 3 (33.3%) eee ee Humeri 3 (25%) 3 (8.8%) 2 (22.2%) eee ee Radii 1 (5.3%) 3 (11.5%) 1 (7.1%) eee e2 (100%) Ulnae ee e2 (100%) ee 3 (75%) e Femora 1 (6.7%) e 4 (16%) 2 (50%) ee ee Tibiae 10 (32.3%) 3 (11.5%) 1 (10%) 1 (50%) ee ee Metapodials 5 (11.4%) 2 (4.8%) 3 (11.1%) eee e1 (16.7%) Phalanges 1 (6.7%) 1 (7.7%) e 1 (4.2%) ee e2 (8.3%)

The percentage shown refers to the appearance of each element in the series. P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 435 scraping was near the edges of breaks that showed evidence of Anthropic bone breakage The assemblage preserved anthropic impact and percussion marks, we attributed it to periosteum breakage (n ¼ 219) in long and flat bones (Fig. 5). The bone surface removal. We associated scraping located on two specimens of damage related to anthropic breakage included stone tool inflicted H. antecessor (radius and ulna) and one deer humerus with percussion pits and peeling (Johnson, 1985; Blumenschine and anthropic breakage of the bones. Selvaggio, 1988; White, 1992). As well as percussion pits, the bones We documented forelimb disarticulation in elements from also occasionally presented conchoidal flakes and adhered flakes. medium and small-sized animals and one rhinoceros ulna. The ulna Variations in the breakage techniques, depending on the thick- presented incisions in the trochlear notch. The TD6-2 hominin ness of the elements, included hammerstone percussion and group completed the disarticulation of the segment with incisions peeling (Table 7). We found breakage by percussion in 112 speci- around the olecranon fossa of several humeri from medium-sized mens, mainly in long bones, although it was also present in and small animals. Despite the differences in size among the mandibles and crania. Eight axial elements (six ribs, one vertebra affected animals, the location of cut marks in both elements was the and one os coxa) completed this group. The possible aim of result of the same disarticulation process. This finding suggests that percussion on these items was to reduce the bones to smaller the disarticulation pattern might not differ among weight cate- segments during the disarticulation of the trunk elements. gories. Finally, a H. antecessor radius showed a chop mark on the The cranial fragments of H. antecessor also presented anthropic distal metaphysis, which possibly occurred during disarticulation breakage. We found percussion pits on two zygomatics, near the from the distal part of the extremity. orbit, on the zygomatic arch and in the pteric region in other cases. One deer ulna bore transverse cut marks in the medial area of These marks provide clear evidence of brain removal. One the diaphysis. It was difficult to determine in which part of the mandible also showed percussion pits on the body, inferior to M2. butchering process they occurred, although the location and The TD6-2 hominin group broke this element for marrow extrac- orientation suggest that it was during carcass skinning. tion, similar to long bones. Hindlimbs In this segment, the bones with cut marks were femora There were 97 long bones with percussion pits and impact and tibiae of all weight categories, and included H. antecessor. The marks (Table 8 and Fig. 5). Forty of these remains were diaphysis features and layout of the cut marks were similar to those found on fragments that we did not identify anatomically. All long bones the forelimbs. The cut marks were generally slice marks on the from the extremities and all weight categories were represented diaphysis, located oblique to the longitudinal axis of the bones. In amongst the remainder (n ¼ 57). We found percussion pits mainly these cases, there is no doubt that actions taken during defleshing in the central areas of the diaphysis. caused the cuts. We associated scraping on the diaphysis fragment Breakage by bending and twisting played an especially impor- of a H. antecessor femur with bone breakage, probably related to tant role in the sample of flat bones (Fig. 5e and f). The assemblage periosteum removal. contained 108 specimens with this type of breakage, primarily in Two large mammal femora preserved disarticulation-related bones from small-sized animals and H. antecessor. In the vertebrae, marks. In one case, we located the slice marks on the femoral head we found peeling in transverse apophysis and articular facets, made during the dismembering of the extremity and trunk. In the generated during carcass disarticulation. Mandibles showed other case, we located the slice marks in the distal part, as a result of peeling on the coronoid process, as a result of jaw and cranium the disarticulation of the femur and tibia. In both cases, the marks disarticulation. Ribs evidenced peeling at various heights of the were transverse to the sagittal axis of the bone. A tibia with slice diaphysis, and the TD6-2 hominin group probably broke them into marks on the distal part of the anterior aspect completed this fragments for easier handling during consumption. We attributed grouping. The TD6-2 hominin group made these incisions when the the presence of peeling at the neck of the ribs to disarticulation, as Achilles tendon was cut. However, we cannot be sure of the purpose was the case with the H. antecessor metapodials and phalanges, of the cuts (skinning, defleshing or disarticulation). where peeling was always near the metaphyses and epiphyses. Feet Defleshing was the most common activity in this segment. Most cut marks were transverse incisions (Fig. 4d and e). We Carnivore-induced modifications recorded chop marks in two cases. The location of the cut marks varied, as we found them in various elements including The TD6-2 assemblage preserves evidence of carnivore activity, sesamoids, metapodials and phalanges. The position also varied affecting 4.8% of the total sample. The types of bone surface damage within each element. caused by carnivores were, in order of frequency: tooth marks (pits, H. antecessor metapodials and phalanges also displayed cut punctures and scores) (30.1%), corrosion by gastric acid (26.8%), marks, usually located in the central part of the diaphysis, in furrowing (26.8%), crenulated edges (9.4%), scooping-out (8.9%), transverse groupings. In the experiment using a chimpanzee, these and/or pitting (5.6%). marks occurred while the experimenters were cutting tendons. We mainly found this type of damage in bones that could not be Nonetheless, the purpose of these marks could not be defined. identified anatomically (n ¼ 115). The most abundant identifiable

Table 8 Location of tooth marks caused by carnivores on long bones and incidence compared to portions recovered in TD6-2.

Size Prox. Epiph. Prox. Epiph. þ Shaft Shaft Distal Epiph. þ Shaft Distal Epiph. Epiph. Whole Total Very large-sized ee2/5 (40%) eeee2/5 (40%) Large-sized 1/5 (20%) 0/3 (0%) 19/285 (6.6%) 0/5 (0%) 1/4 (25%) 0/5 (0%) e 21/307 (6.8%) Medium-sized 2/8 (25%) 0/3 (0%) 31/606 (5.1%) 0/2 (0%) 2/6 (33.3%) 1/2 (50%) e 36/627 (5.7%) Small-sized e 0/3 (0%) 19/254 (7.4%) 0/2 (0%) 0/5 (0%) e 0/1 (0%) 19/265 (7.1%) H. antecessor e 0/2 (0%) 0/14 (0%) 0/1 (0%) 0/3 (0%) ee0/20 (0%) Indeterminate ee0/12 (0%) eeee0/12 (0%)

Total 3/13 (21.4%) 0/11 (0%) 70/1176 (5.9%) 0/10 (0%) 3/18 (16.6%) 1/7 (14.3%) 0/1 (0%) 78/1237 (6.3%)

The first figure in the cells refers to the number of specimens affected, and the second (after the/) refers to the total of each category in the series. The percentage refers to specimens with tooth marks in each category (Prox. ¼ proximal; Dist. ¼ distal; Epiph. ¼ epiphyses). 436 P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 elements included fragments of antlers (n ¼ 16), followed by ribs (n ¼ 15) and metapodials (n ¼ 10), although bones from the entire skeleton showed damage. According to the NISP of each weight category, items from

medium-sized animals contributed the largest number of remains 0/123 (0%) with tooth marks (n ¼ 64). According to the relative frequency of each weight category, the most affected category was very large, and the least affected was small (very large: 9.3%; large ¼ 7.8%; medium-sized ¼ 5.3%; and small ¼ 5% with carnivore-induced modifications). Epiphyseal fragments showed marks more frequently than shaft fragments (20.5% vs 6%). Of the diaphyseal fragments, nine speci- mens were fragments near the epiphyses and the rest were mid- shaft. None of the epiphyses attached to diaphyseal fragments 0/9 (0%) 42/1312 (3.2%) 50/1524 (3.3%) showed carnivore chewing marks (Table 8). In the other elements (non-long bones), there was a greater prevalence of os coxae with ee carnivore-induced modifications (Table 9). Blumenshine (1986, 1995) and Capaldo (1998), warn that in human-to-carnivore

experimental situations, tooth marks are mainly located in or 0/1 (0%) 1/4 (25%) 0/1 (0%) 2/25 (8%) e 2/15 (13.3%) 12/118 (10.2%)1/13 (7.7%)2/19 (10.5%) 8/198 4/58 (4%)0/24 (6.9%) (0%) 4/95 (4.2%) 36/350 (10.3%) 0/3 (0%) 0/5 (0%) 28/512 (5.5%) 17/420 (4%) near the epiphyses and the axial elements tend to completely -or 5/72 (6.9%) 25/424 (5.9%) 46/1379 (3.3%) 133/2954 (4.5%) almost completely- disappear. The suggested model is similar to the one obtained from TD6-2, with only one difference: the under- representation of near epiphyses portions, while in the experi- mental model, carnivores tend to generate fragments belonging to these bone regions. We identified carnivore breakage on the basis of tooth marks

associated with notches on fracture edges. However, this type of 0/4 (0%) breakage was very scarce in the assemblage (n ¼ 34) and mainly affected long bones, most of which were more than 30 mm long. Table 10 shows tooth mark measurements (pits, punctures and

perforations). Average measurements appear to be those caused by e small or medium-sized carnivores (Selvaggio and Wilder, 2001; Domínguez-Rodrigo and Piqueras, 2003; Delaney-Rivera et al., 2009). The maximum values report on the effects of a large carni- vores, such as bears, lions or hyenas (Domínguez-Rodrigo and Piqueras, 2003; Delaney-Rivera et al., 2009), thus making these data ambiguous for interpretative purposes. The imprecise nature of these data could be due to various factors, including 1) the 1/4 (25%) 0/1 (0%) 9/72 (12.5%) 1/2 (50%) 1/4 (25%) 3/28 (10.7%) involvement of more than one type of carnivore, 2) the fact that 5/89 (5.6%) 1/7 (14.3%) 0/60/16 (0%) (0%) 2/17 (11.8%) 0/1 (0%) 0/1 (0%) 0/3 (0%) 0/129 (0%) 2/5 (40%) 0/9 (0%) 4/23 (17.4%) large carnivores also produce many small marks, which could cause the resulting measurements to beinaccurate (Delaney-Rivera et al., 2009), and/or 3) the overlapping values of various actors e e e e (Domínguez-Rodrigo and Piqueras, 2003; Delaney-Rivera et al., e 2009). Punctures and perforations can provide the most information of all tooth marks, as there may be a link between their dimensions and those of the teeth that generated them. However, this type of mark is less abundant than others caused by carnivores, and the information provided may contain a bias. We observed punctures on six elements and perforations on three in TD6-2. Four elements preserved punctures less than 5 mm long and 4 mm wide. Another group of remains contains several larger marks. Taking into account the resistance of different tissues (cancellous bone, cortical and thin cortical bone), the location of the punctures, the weight category of 0/2 (0%) 0/6 (0%) 0/2 (0%) 0/62 (0%)0/25 (0%) 3/10 (30%) 2/63 0/5 (3.2%) (0%) 0/19 (0%) 0/3 (0%) 0/31 (0%) 0/5 (0%) 0/3 (0%) 0/8 (0%) the elements and the type of bones surveyed, we considered that activity by two actors of different sizes produced these punctures. This consideration is in agreement with the findings of Selvaggio and Wilder (2001) in regards to the weight category of the elements and Delaney-Rivera et al. (2009) in regards to bone type 8/160 (5%) 0/12 (0%) 0/1 (0%) 0/9 (0%) 16/221 (7.2%) 4/189 (2.1%) 4/41 (9.8%) 4/159 (2.5%) 0/3 (0%) 15/341 (4.4%) 4/18 (22.2%) 1/23 (4.3%) 9/83 (10.8%) e e e surveyed. (Table 11) Finally, modification by stomach acid corrosion was a significant

factor amongst the remains that were affected by carnivores in gure in the cells refers to the number of specimens affected and the second (after the/) refers to the total of each category in the series. The percentage refers to specimens with tooth marks in each category. fi TD6-2 (n ¼ 59) (Fig. 6a). This corrosive stomach acid modification rst mainly affected deer carpals/tarsals, phalanges and metapodials. fi Indet. Total Large-sized 0/4 (0%) 2/24 (8.3%) 0/8 (0%) 2/17 (11.8%) Medium-sizedSmall-sized H. 8/55 antecessor (14.5%) 2/64 (3.1%) 1/11 (9.1%) 0/49 (0%) SizeVery large-sized Antlers Crania Mandibles Vertebrae Clavicles Ribs Coxae Scapulae Carpals/Tarsals/Sesamoids Phalanges Flat bones Indet. Total The Pickering (2001) confirmed that leopards and hyenas consume Table 9 Non-long bone elements with tooth marks. P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 437

Table 10 most low-cost bones are high survival bones, while some high- Statistical values of pits and punctures analyzed in TD6-2. survival elements are also high cost, and vice versa. For this n Mean C.I. C.I. þ95% Minimum Maximum S.D. reason, we tested the %MAU of herbivore animals with SFUI 95% value value according to their survival capacity and their processing costs. In Length Cort. 520 2.23 2.13 2.34 0.45 9.0 1.23 TD6-2, there was no correlation between the %MAU of herbivores Th. Cort. 37 3.67 3.16 4.17 1.52 7.1 1.52 and the SFUI (r2 ¼ 0.076, p ¼ 0.591). Considering the low-survival Can. 39 2.64 2.07 3.21 0.96 10.4 1.75 elements and the high survival elements in isolation, based on Wide Cort 520 1.54 1.47 1.62 0.26 6.4 0.88 Th. Cort. 37 2.70 2.28 3.12 1.22 5.7 1.26 the proposals of Marean and Cleghorn (2003), there were no Can. 39 1.99 1.48 2.51 0.69 9.3 1.58 significant changes from the overall group trend, but when the fi Data on length and width given (Cort. ¼ Cortical; Th. Cort. ¼ Thin Cortical, remains are grouped by cost, the trend appeared to be signi cant 2 Can ¼ Cancellous bone/C.I. ¼ Confidence Interval and S.D ¼ Standard deviation). amongst the low cost elements (r ¼ 0.738, p ¼ 0.03) (Fig. 7). In TD6-2, there seems to be no relationship between the frequency of high survival elements and their economic usefulness. whole toes without chewing them, leading to a skeletal profile According to the Shannon evenness index, based on proposals by similar to that found in the digested bones from TD6-2. The Faith and Gordon (2007), the groups have very uneven anatomical majority of digested specimens in this assemblage were less than representations in all weight categories (Table 12). This includes the 40 mm long. Horwitz (1990) states that bones digested by large skeletal distribution of TD6-2 as the result of a gourmet strategy carnivores (striped hyenas, and dogs) never exceed these (Binford, 1984; Faith and Gordon, 2007), which would maximize dimensions. We noted that some items from TD6-2 exceeded these the quality of the transported elements. We must consider these measurements, most of which were from an antler fragment data with caution due to the small sample size and the fact that we approximately 60 mm in length. According to Tappen and only took the high survival elements into account. Wrangham (2000), the dimensions of bones that have passed Faith and Gordon (2007) showed that the smaller the sample, through the intestinal tract are not usually a good predictor of the the less likelihood of finding a correlation between the abundance consuming predator’s body size. Hyenas, for example, can digest of elements and SFUI, despite its existence. This type of error does and regurgitate pieces that are larger than those they excrete not usually affect assemblages that are the result of a gourmet (Tappen and Wrangham, 2000). Participation by hyenas in the strategy. Furthermore, in this test Faith and Gordon (2007) advise assemblages recovered from TD6-2 is thus possible. against testing low-survival elements, as Marean and Cleghorn The TD6-2 collection contains a number of remains (n ¼ 22) (2003) point out the marginal possibility of their representation with cut marks and tooth marks on their surface, although these being similar to the original. This trend seems clear in the TD6-2 marks explicitly overlap in only two cases (Fig. 6b). Defleshing assemblage, as there are few ribs, vertebrae or compact bones. caused the cut marks in both cases, while we located the tooth The presence of scapulae and os coxae was nonetheless slightly marks above the cut marks, indicating that the carnivores acted higher, mainly amongst the small animals. after the hominins. Lastly, we correlated the %MAU of the TD6-2 remains and Emerson’s Utility Index (Emerson, 1993), which revealed a close relationship between the fat content of bones and their presence in Food utility index all weight categories groups and the hominins. The analysis of marrow content showed a close and statistically significant rela- fi Marean and Cleghorn (2003) classi ed skeletal elements on the tionship for large and medium-sized animals (Table 13). We found basis of two variables: their survival capacity (low and high survival negative correlations between the proportions of skeletal parts in elements) and the processing cost from SFUI (low and high cost the TD6-2 assemblage and the UMI from Morin (2007) (large-sized, elements). There is a major overlap between the two categories, as r2 ¼0.579, p ¼ 0.227 medium-sized, r2 ¼0.678, p ¼ 0.138; small- sized, r2 ¼ 0.115, p ¼ 0.826; H. antecessor, r2 ¼0.925, p ¼ 0.008). Table 11 Dimensions (mm) of the punctures and perforations located in TD6-2 assemblage. Thus, the TD6-2 hominin group showed no predilection for marrow with high proportions of unsaturated fatty acids. Element Tooth mark Tissue Length Wide Actor In their study of the remains recovered from TD6-2, Díez et al. Cervical vertebra Puncture Th. Cort. 3.59 3.04 Small Cr (1999) pointed out the differential transport of animals depend- small-sized animal Puncture Th. Cort. 2.44 2.2 Small Cr ing on the their weight. In the analyzed sample, Díez et al. (1999) Puncture Th. Cort. 2.46 2.32 Small Cr Puncture Th. Cort. 4.34 3.42 Small Cr drew these conclusions in the absence of axial elements, which Puncture Th. Cort. 3.86 3.19 Small Cr become more pronounced as the animal’s weight increases. Our Puncture Th. Cort. 4.51 3.47 Small Cr interest lies in carcass transport and processing stages and whether Rib large-sized animal Puncture Cort. 5 2.06 Cr these decisions were closely related to the weight category of an Metacarp medium-sized deer Puncture Cort. 1.56 1.34 Cr Puncture Cort. 1.76 1.18 Cr animal, which our research suggests. In the present assemblage, the Femur medium-sized deer Puncture Can. 2.30 1.67 Small Cr presence of vertebrae and ribs was similar in all weight categories Puncture Can. 1.60 1.5 Small Cr according to the %MAU. Long bone medium-sized Puncture Cort 4.99 4.61 Large Cr To make the skeletal profiles more visible, we distributed the animal skeletal elements by anatomical segments in accordance with the Coxae C. crocuta Puncture Th. Cort. 7.12 5.72 Large Cr Puncture Th. Cort. 4.6 4.2 Large Cr systematic dismembering of carcasses, based on our own experi- Puncture Th. Cort. 3.9 3.68 Large Cr mental observations and in a similar way to Stiner (2002). The Medial phalanx Perforation Th. Cort. 3.96 3.85 Cr segments used included the cranium (cranial bones and jaws), MEDIUM-sized deer trunk (vertebrae, ribs and os coxae), forelimbs (scapulae, humeri, Medial phalanx Perforation Th. Cort. 5.58 4.8 Large Cr fi small-sized deer and radii/ulnae), hindlimbs (femora, tibiae and bulae), meta- Talus medium-sized deer Perforation Can. 5.97 5.18 Large Cr podials and phalanges. In this grouping, we interpreted the pres- ence of an element as the possible presence of an entire segment in Details are given of the elements in which they were found, the type of tissue, the length, width and deduced size of the carnivore (Cort. ¼ Cortical, Th. Cort. ¼ Thin the cave. After calculating the Minimum Number of Segments Cortical, Can. ¼ Cancellous bone; Cr ¼ Carnivore). presented in the assemblage, we applied the Minimum Animal 438 P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446

Figure 6. a) Antler and bone fragments with modification by stomach acid corrosion. b) Explicit overlap (ESEM detail on bottom right) of tooth marks and cut marks, showing that the carnivore acted after the hominids.

Segments (MASg) with the corresponding Minimal Animal provided arguments for the role of TD6-2 as a residential site. All of Segments as per Binford (1984) to calculate MAU and %MAU. the exploitation sequences for the major types of stone tools were Fig. 8 shows the application of this formula to the TD6-2 present in the assemblage, there was a wide range of available raw assemblage. Comparison of the results from the %MAU and the % materials, and cores in different stages of reduction and re-fittings MASg identified some nuances, which included a decrease in the were present (Carbonell et al., 1999). number of crania in large and small animals, fewer trunks in all of Díez et al. (1999) suggested two hypotheses for hominin strate- the weight categories, including H. antecessor, but with a similar gies for the acquisition and transport of various animals. First, that presence in all weight categories, and a pronounced spike in the the hominins had immediate, primary access to small and medium- representation of the hindlimbs from large and small animals, with sized animals, while they had secondary access to the carcasses of both limbs of medium-sized herbivores being well represented. large animals. According to this proposal, the hominins could have Additionally, the appearance of H. antecessor trunks was similar to had access to the prey of large felids, which tended to leave the bones the forelimb and hindlimb pattern, possibly due to the fact that we of carcasses almost intact with some meat left (Blumenschine et al., only included taxonomically identified elements in this group. The 1994; Domínguez-Rodrigo, 1997a, 1999). However, no diagnostic large number of unidentified diaphyseal fragments may also elements indicated felid activity in TD6-2. Second, Díez et al. (1999) include hominin remains. proposed the differential transport of carcasses according to the animals’ weight. In other words, the TD6-2 hominin group trans- Discussion ported small animals whole, and only moved selected parts (extremities and crania) of large ungulates to the cave. Level TD6-2 is currently the richest deposit of Lower Pleistocene Díez et al. (1999) based both hypotheses on the relative abundance remains in Sierra de Atapuerca. This sub-unit provides an oppor- of various skeletal items in the assemblage, and particularly the poor tunity to consider several economic issues regarding the forager representation of post-cranial axial skeletons from larger animals. groups that inhabited this area around one million years BP, a time Analysis of the present assemblage, which now contains more period of European prehistory about which we know little. specimens, shows very poor representation of the post-cranial axial The first scientific analysis of the assemblage date back to exca- skeleton in all weight categories, although it is true that this bias is vations between 1994 and 1996 (Carbonell et al.,1999). Based on the slightly more marked in large animals. Current opinion suggests study of the lithic and faunal remains, as well as the taphonomic and that low-survival items such as ribs, vertebrae, os coxae, phalanges sedimentological analysis, it was concluded that Gran Dolina was and compact bones should not be considered in research of original a ‘residential site’ during the formation of this level. In other words, skeletal profiles, and therefore in the conclusions regarding deci- the TD6-2 hominin group frequently used this space for several sions on butchering and transport strategies by hominins (Marean, activities, some of which the inhabitants carried out from the 1991; Marean et al., 1992; Marean and Cleghorn, 2003; Cleghorn opening to the back of the cave (Bermúdez de Castro et al., 1999). and Marean, 2004; Faith and Gordon, 2007). Due to this factor, Studies have interpreted the accumulation of remains as the we must seek other empirical foundations apart from anatomical result of anthropic activity, in which the role of carnivores was representation to investigate the types of access and transport irrelevant (Díez et al., 1999; Rosell, 2001). Stone tools have also decisions by the hominins who occupied TD6-2. P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 439

Figure 7. Correlation between the %MAU of low and high survival elements and according to processing cost and standard food utility index. a) Correlation between %MAU of herbivores and SFUI of low-survival elements (rs ¼0.230, p ¼ 0.390). b) Correlation between %MAU of herbivores and SFUI of high survival elements (rs ¼ 0.006, p ¼ 0.971). c) Correlation between %MAU and high cost elements (rs ¼0.542, p ¼ 0.131). d) Correlation between %MAU and low cost elements (rs ¼ 0.738 p ¼ 0.03). SFUI data according Metcalfe and Jones (1988). Division of elements based on survival and processing cost, according to Marean and Cleghorn (2003).

It is difficult to distinguish simple events in palimpsests like TD6- by hominins involves establishing the sequence in which carni- 2, yet there are specific processes that leave explicit traces in the vores and hominins consumed and modified the carcasses. fossil record and permit conclusions regarding specific behaviors. Possible scenarios for consideration are carnivoreehominin, When we find a specific pattern based on repetitions of these traces, homininecarnivore, and carnivoreehomininecarnivore it is possible to establish the existence of general processes or (Blumenschine and Selvaggio, 1988; Selvaggio, 1994a, 1998; behavior. In the present study, we believe that the fossil record Blumenschine, 1995; Capaldo, 1995, 1997, 1998; Domínguez- clearly shows that hominins had early access to whole carcasses and Rodrigo, 1997a; Capaldo, 1998; Selvaggio, 1998; Domínguez- that carcass size did not determine selective transport decisions, Rodrigo, 1999; Selvaggio and Wilder, 2001; Domínguez-Rodrigo whenever taken. Both points are considered in more detail below. and Barba, 2006), ruling out those scenarios in which only one actor was involved (hominin only, carnivore only). The coincidence of tooth marks and cut marks in some specimens and their explicit Type of access overlap (tooth marks on top of cut marks) suggests a scenario in which hominins participated prior to carnivores. Bone surface modifications indicate participation by at least two actors (carnivores and hominins) in the TD6-2 assemblages during the nutritional phase of the carcasses. Inferring the type of access Table 13 General utility rate grouping by size weight.

Table 12 Utility rate Large-sized Medium-sized Small-sized H. antecessor Minimum number of elements (MNE) of high survival elements, evenness and General utility rs 0.013 0.105 0.139 0.272 ’ e Spearman s rank order correlation (rs) between the high survival elements and the p 0.968 0.756 0.682 0.417 standard food utility. Food utility rs 0.091 0.036 0.189 0.397 p 0.789 0.914 0.575 0.226 Size MNE Evenness rs Bone fat rs 0.826 0.775 0.774 0.716 ¼ Large-sized 46 0.233 0.374 (p 0.408) p 0.003 0.008 0.008 0.19 Medium-sized 53 0.251 0.807 (p ¼ 0.02) Bone marrow rs 0.794 0.740 0.521 0.111 ¼ Small-sized 48 0.240 0.21 (p 0.641) p 0.010 0.022 0.150 0.775 H. antecessor 26 0.248 0.741 (p ¼ 0.05) data taken from Emerson (1993). 440 P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446

Figure 8. %MASg distribution for weight categories and H. antecessor.

One of the methods for reconstructing hominin exploitation of intermediate limb elements, we regarded ulnae and fibulae as carcasses and the way they acquired them is the distribution of cut significant elements, which significantly increased the relative marks on the animal skeletons (Bunn and Kroll,1986, 1988). For this frequency of this portion. Metapodials had the lowest purpose, Domínguez-Rodrigo (Domínguez-Rodrigo, 1997a, 1999; frequency of cut marks associated with flesh removal. These Domínguez-Rodrigo and Pickering, 2003; Domínguez-Rodrigo data were also consistent with those from FLK Zinj, where and Barba, 2006) proposed that the distribution patterns of cut Domínguez-Rodrigo (1997a) noted a clear difference in the marks according to the skeletal section and the specific elements distribution of elements with cut marks, as upper and inter- with their location in specific areas, are indicative of the ways in mediate limb bones accounted for most of the cut-marked which hominins exploited carcass resources, as well as their access specimens (88%) in comparison with metapodials (12%). to them. These data also consider the sequence in which carnivores 3. Finally, there is a differential distribution of cut marks in consume large ungulate carcasses and the distribution of tooth primary access according to bone section. In assemblages with marks on the various elements and portions (Blumenschine, 1986, no scavenging after hominin activity, cut marks in the central 1995; Selvaggio, 1994a; Capaldo, 1995, 1998; Selvaggio, 1998; areas of the diaphysis are roughly 43%, and are more abundant Domínguez-Rodrigo, 1999; Domínguez-Rodrigo and Pickering, (50 and 70%) in assemblages with the complete dual process 2003; Domínguez-Rodrigo and Barba, 2006). (hominin to carnivore). Tests on secondary access (Domínguez- Domínguez-Rodrigo (1997a) proposed the following series of Rodrigo, 1997a) show that neither proximal nor intermediate features for long bones in assemblages, depending on whether bones presented cut marks in the central area of the diaphysis. hominins had primary or secondary access: In TD6-2, we found 94.7% of cut marks (not including meta- podials) generated during flesh removal from long bones in 1. In primary access assemblages, the long bones with cut marks diaphysis in all weight categories and Homo antecessor. In TD6- made during flesh removal account for between 30% and 60%. 2, the scarcity of epiphyses influenced this result. In the three-stage scenario (carnivore to hominin to carnivore), the percentage is between seven and 19.5%. At TD6-2, we found We performed a correspondence analysis (Fig. 9) to assess the cut marks in 30.7% of the long bones (humeri, radii, femurs, frequency and distribution of cut marks on long bones from TD6-2 tibiae and metapodials), suggesting early access. in relation to other actualistic assemblages (Domínguez-Rodrigo, 2. In primary access, there is an overlap of differential cut marks, 1997a,b; Lupo and O’Connell, 2002). The most discriminant vari- depending on the elements. Proximal appendicular elements able in Axis 1 (eigenvalue ¼ 74.6) was cut mark frequency on lower (femora and humeri) have a higher percentage of cut marks limb bones. This variable clearly separates assemblages with early (60%) than intermediate elements (radii/ulnae and tibias) access from those that were the product of secondary access. The (30%), while intermediate elements have a higher percentage analysis shows a clear relationship between the TD6-2 assemblage than distal elements (metapodials) (10%). Secondary access and assemblages in which hominins had early access to carcasses. generally has a higher percentage of cut marks in the inter- mediate and distal bones of the extremities than in the prox- imal elements, which may not include any specimen with cut Table 14 marks. The distribution of cut marks in TD6-2 was similar to the Frequency of bones with cut marks on proximal limbs (humeri and femora), inter- distribution of primary access, although not strictly as mediate limbs (radii/ulnae, tibiae and hominid fibulae), and distal limbs proposed by Domínguez-Rodrigo (1997a) (Table 14). Small- (metapodials). sized animals matched this distribution best. Medium-sized Location of cut Very large- Large- Medium- Small- H. antecessor Total animals presented a descending pattern in frequencies, with mark sized sized sized sized a larger number of specimens bearing cut marks on proximal Proximal limb 40.0 31.0 47.8 52.4 28.6 40.6 appendicular elements. Large-sized animals and H. antecessor Intermediate 40.0 45.2 41.3 38.1 50.0 43.0 presented more specimens with cut marks on the intermediate limb Distal limb 20.0 23.8 10.9 9.5 21.4 16.4 appendicular elements. In the case of H. antecessor P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 441

There were no significant differences between the weight groups in 8.6% are associated with flesh removal. We also found cut marks on the assemblage. Axis 2 (eigenvalue ¼ 11.9) creates two groups that the ventral side of ribs in all weight categories. These marks are seem well associated between the early access assemblages. The undoubtedly related to carcass evisceration (Nilssen, 2000 from first group (bottom left graph) represents sets in which the Pobiner et al., 2008). We found evisceration-related cut marks on butchery process was intensive (Domínguez-Rodrigo et al., 2009). 27.3% of ribs, a relatively high percentage bearing in mind their The experimental assemblage H1S1/2d (modeling by Dominguez- scarcity in the assemblage, and the fact that our observations Rodrigo, 1997a) and the Hadza (Lupo and O’Connell, 2002) repre- showed that this part of the butchering process usually leaves the sents the second group (top left). Less intensive defleshing related fewest marks on the bone surface. It is difficult to imagine evis- marks occur in these assemblages, during which meat scraps are ceration of the thoracic cage in situations of passive secondary left behind. TD6-2 was part of the latter group. The difference access to carcasses. between the two groups seems marked by the different frequency According to Blumenschine (1986, 1995), the sequence of of cut marks on the upper limb bones and on the intermediate limb carnivore access to carcasses segments is: 1) hindquarter flesh, 2) bones, as indicated Lupo and O’Connell (2002). These results ribcage and forequarter flesh, 3) head flesh, 4) hindlimb marrow, 5) support primary access by TD6-2 hominin group to carcasses. forelimb marrow, and 6) head content. The TD6-2 hominin group However Lupo and O'Connell (2002), have questioned cut mark often accessed entrails and substantial amounts of meat located in frequency and distribution as a good way of establishing the the proximal parts of the extremities. This is not consistent with pattern of activity on carnivore and hominin carcasses. For this a general model of secondary access for any of the weight reason, we judged the frequency of tooth marks and percussion categories. marks on mid-shaft long bones to be very informative (Marean The presence of cut marks in all elements and all weight cate- et al., 2000; Thompson, 2010). We calculated these frequencies gories shows that the TD6-2 hominin group processed crania, limbs for the TD6-2 assemblage to supplement the information yielded and the axial skeleton, including abdominal contents. The butch- by cut mark distribution and frequency. Fig. 10 shows tooth and ering pattern was similar in all cases, with the most significant percussion mark frequencies on mid-shafts in three weight groups. differences found in H. antecessor. These discrepancies may be TD6-2 is related to ‘human first’, supporting the hypothesis of explained by the anatomical differences between hominins primary access in the TD6-2 assemblage. (bipeds) and ungulates (quadrupeds). All of the above-mentioned considerations refer to long bones. The current TD6-2 assemblage shows early and immediate To complete the descriptive model and permit inferences about the access to at least a large proportion of the animals and hominins. type of access by hominins, we must also consider flat bones, This model is obvious in small and medium-sized animals, and in primarily vertebrae and ribs. Despite usually being scarcer in this H. antecessor. Acquisition of very large- and large-sized animals was type of reconstruction, these elements should not be ignored in possibly subject to more than one type of strategy. Unfortunately, it interpretations, as they can provide highly informative data. is impossible to interpret the fossil record at this level. Cut marks Although it is true that they are relatively scarce in TD6-2, the related to evisceration and flesh removal from the humeri and modifications presented by these remains allow certain conclu- femora show that the TD6-2 hominin group acquired at least some sions to be drawn on the types of hominin access. First, the pres- of these large animals in the same way as those weighing less and ence of cut marks in scapulae is significant (21%), although only that they were responsible for the entire processing of the

Figure 9. Multiple Correspondence Analysis of experimental, ethnoarchaeological and TD6 cut mark distribution. Experimental and ethnoarchaeological data from Domínguez- Rodrigo (1997a* 1997bþ) and Lupo and O’Connell (20020). Data summarized in Dominguez-Rodrigo et al. (2009: Table 14). Variables are: Cm-ULB ¼ cut marks on upper limb bones; Cm-ILB ¼ cut marks on intermediate limb bones; Cm- LLB ¼ cut marks on lower limb bones; Cm-Midshafts ¼ cut marks on mid-shafts; CmMBMSH:cm ¼ is the proportion of cut-marked upper limb bones compared with total cut-marked long bones (Domínguez-Rodrigo et al., 2009). Plot shows distribution of actualistic/TD6 assemblage and dimensional distribution of variables. Legend: H1S1 ¼ Hominid Only; H1S1/2a, H1S1/2b, H1S1/2c, H1S1/2d, Hadza ¼ Homind to Carnivores; H2S1/2a, H2S1/2b ¼ Carnivore to Hominid; H2S1/2/ 3a, H2S1/2/3c, H2S1/2/3d ¼ Carnivore to Hominid to Carnivore; TD6-LS ¼ TD6 large sized animals, TD6-MS ¼ TD6 medium sized animals; TD6eSS ¼ TD6 small sized animals; TD6 ¼ total TD6. Explanation yielded by 86.5%, taking into account the two dimensions (eigenvalue Axis 1 ¼ 74.6, eigenvalue Axis 2 ¼ 11.9). 442 P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446

Figure 10. Scatter plot of frequencies of percussion marks and tooth marks on long bone mid-shafts. Data from Blumenschine(1988), Blumenschine and Marean (1993), Capaldo (1997, 1998), Marean and Kim (1998), Selvaggio (1998) summarized by Marean et al. (2000). Legend: 1B ¼ hominid only, Blumenshine; 1M ¼ hominid only, Marean; 2B ¼ hominid to carnivore, Blumenshine; 2C ¼ hominid to carnivore, Capaldo; 2M ¼ hominid to carnivore, Marean; 3B ¼ carnivore only, Blumenshine; 3C ¼ carnivore only, Capaldo; 3M ¼ carnivore only, Marean; 4S ¼ carnivore to hominid, Selvaggio; 5S ¼ carnivore to hominid to carnivore, Selvaggio; 6a ¼ TD6 large sized animals; 6b ¼ TD6 medium sized animals; 6a ¼ TD6 small sized animals. carcasses, from defleshing to breakage of bones from animals in all According to this hypothesis, only small animals and hominins weight categories, including hominins. would have been transported whole. A more variable pattern The disappearance of skeletal portions (post-cranial axial would have existed for medium-sized animals, which would have elements, phalanges, compact bones and epiphysis) is consistent been subject to other factors. Finally, larger animals would have with post-ravaging by carnivores (Blumenschine and Selvaggio, been subject to initial processing and consumption at the kill site, 1988; Bunn et al., 1988; Marean et al., 1992; Blumenschine and where the segments of least nutritional interest would have been Marean, 1993; Blumenschine, 1995; Capaldo, 1998; Marean and discarded. The data now available suggest that most archaeofaunal Cleghorn, 2003). Tooth marks were less frequent than cut marks assemblages are characterized by a more or less significant absence and percussion marks in TD6-2, and we located them more of axial skeletons (Marean et al., 1992, 2004; Blumenschine and frequently on the epiphysis of long bones and os coxae recovered Marean, 1993; Domínguez-Rodrigo, 1999; Marean and Cleghorn, from the site. Explicit superposition of tooth marks on cut marks 2003; Pickering et al., 2003; Cleghorn and Marean, 2004; Faith supports the secondary role of carnivores in TD6-2. Furthermore, and Gordon, 2007). Hyena activity on the remains discarded by we assigned most bone breakage in TD6-2 to hominin activity. hominins left a similar pattern to the schlepp effect (Perkins and Primary and secondary access may have included different Daly, 1968; Blumenschine, 1986, 1995; Marean et al., 1992; strategies for the acquisition of animal resources (Bunn and Ezzo, Blumenschine and Marean, 1993; Capaldo, 1998). This uniformity 1993), such as hunting and confrontational scavenging. means that skeletal profiles are insufficient for robust conclusions Domínguez-Rodrigo and Barba (2006) formulated the hypothesis on the transport decisions made by Pleistocene hominins. that African hominins in the Plio-Pleistocene acquired carcasses by Studies of transport strategies in modern hunter-gatherer hunting, as they believe that confrontational scavenging was groups show that they must consider many issues prior to trans- a highly dangerous strategy. We believe that this assumption is port, which are sometimes complex in inferential terms for Pleis- valid for the interpretation of TD6-2, although it is difficult to tocene contexts (Yellen, 1977; Binford, 1978, 1981; Bunn and Kroll, generalize this practice to all of the carcasses acquired, processed 1986, 1988; Bunn et al., 1988, 1990; Bunn, 1993; Monahan, 1998). and consumed in TD6-2. We cannot rule out the possibility of Despite the weight of the animals, which was undoubtedly an scavenging (active or passive) at least on an occasional basis, important issue for transport between the kill site and the home although the clearest empirical evidence (anatomical distribution base, the number of individuals taking part in a hunting party, the of cut marks, anatomical profiles and overlapping marks) suggests number of prey and therefore the number of animals to be pro- that hunting was the most common strategy for acquiring animal cessed, were all important factors. Gifford-Gonzalez (1993) resources. considers that in addition to the animals’ weight, their anatom- ical structure and nutritional composition may have influenced Carcass transport decisions processing and transport strategies. Monahan (1998) mentions the importance of transport and processing costs in decision-making. Díez et al. (1999) argued that the scarcity of axial elements in Another consideration is the danger posed by other predators TD6-2 was the result of differential transport of prey from the kill/ nearby, which may have limited the processing time at the kill/ butchering site to the home base, depending on its weight and size. butchering site, as well as the kill/butchering site’s distance from P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 443 the home base (Monahan, 1998; Faith et al., 2009). The last factor essential for protection of foragers and their prey against other that may have an important influence on the final decision relates dangerous predators operating in the same environment. to the way the TD6-2 hominin group may have cooked the Marshall (1994) suggested that food distribution amongst carcasses (Gifford-Gonzalez, 1993; Oliver, 1993). TD6-2 contains no members of a group can play an important role in transport deci- evidence pointing to the use of fire or cooking of food, preventing sions. Transporting the carcass to the home base and delaying us from considering this factor in the interpretation of the TD6-2 consumption of at least some parts implies that the hunters shared assemblage. part of this meat with other members of the group (Stiner et al., Correlation rates in TD6-2 between relative abundance of high- 2009). During their occupancy of TD6-2, the TD6-2 hominin survival skeletal items and the food utility of different skeletal group transported large quantities of animal resources, evidenced elements are ambiguous, and therefore of little use in resolving by cut-mark patterns and related butchering activities. These issues of transport decisions. Anatomical profiles do not in them- findings suggests that future studies should give food distribution selves appear to indicate the existence of selective transport in serious consideration as one of the decisive factors in transport terms of weight categories. There is a great deal of similarity choices. In modern hunter-gatherer groups, food sharing is linked between the frequencies of the skeletal segments in the three main to strong coding systems, which are beneficial for the group (e.g., weight categories. Cut marks and breakage of long bones are also Gould, 1981; Hayden, 1981; Lee, 1981; Hawkes, 1993; Marshall, part of a very similar pattern in all weight groups and H. antecessor. 1998), thereby endorsing the social links amongst its various The uniformity of the site in terms of skeletal profiles and bone members. modifications in the different sizes means that the TD6-2 hominin group did not necessarily, or at least not solely, link transport Conclusions decisions to prey weight. In TD6-2, we documented the presence of thoracic and lumbar Hominin groups occupied Level TD6-2 as a home base during vertebrae and ribs amongst the elements from large animals. We the Lower Pleistocene. These hominins were the main agents that may thus assume the presence in the cave of some trunks from accumulated large in the cave. The archaeological record large-sized animal carcasses, bearing in mind the logical carcass suggests that the subsistence strategy used by these hominin processing order. If the TD6-2 hominin group only transported groups was opportunistic hunting. Skeletal profiles and modifica- extremities and crania after initial processing, it would be impos- tions to the bone surfaces provide information about foraging sible to correctly infer skinning and evisceration of large animals. strategies and carcasses transport by these Pleistocene groups. In Contrary to this assumption, we documented the complete pro- most cases, they acquired intact carcasses, as shown by the cut cessing of all weight categories and H. antecessor. marks and breakage patterns, which were not related to the weight This model leads to the conclusion that the TD6-2 hominin of the animals. This model suggests early and immediate access to group transported at least some of the animal segments, or most, to prey on most occasions. the home base, and that they occasionally segmented and partially The transport model used by the TD6-2 hominin group was the processed other animals at the kill site. This discrepancy shows that result of several choices, which various factors helped to determine. they must have based transport decisions to more than weight, We can conclude that the TD6-2 hominin group transported although this issue influenced them. carcasses whole, at least sometimes, to the home base. The absence Factors relating to decision making included the nutritional of some items from the TD6-2 fossil record is partially due to value of bones, the distance between the kill site and the home carnivore activity. The lack of axial skeleton elements is not entirely base, the number of participants in the hunt and number of prey attributable to the action of these animals, although it is clear that caught, as well as the possible distribution of food to other they may have been responsible for the removal of entire vertebrae, members who may not have been direct participants in animal ribs and os coxae from carcasses processed beforehand by the TD6- acquisition. The latter factors are indicative of the level of social 2 hominin group. cooperation within the group. According to observations by Bunn This study demonstrates that animal weight was not the only and Kroll (1988) and Monahan (1998), the number of trans- factor taken into consideration in decisions concerning prey porters can vary, reaching a dozen individuals or more. The variable transportation at the TD6-2 site. Transport strategies related to nature of carcass transport strategies, taking into account the various factors, some of which have a low or even zero potential of animals’ weight, can point to a variation in the number of hunters archaeological visibility. However, we can deduce that several and therefore the number of individuals able to transport intact or individuals participated in hunting parties and/or carcass transport. portioned prey. This variable number of individuals also implies The potential variation in the number of participants is a complex food distribution to other members of the group. issue, but it is nevertheless a clear sign of social cooperation within The phrase ‘social or communal hunting’ is used to describe a group, food sharing and possible division of subsistence tasks, hunting parties involving participation by more than one or two conceivably to ensure the group’s survival. These conclusions individuals (Hayden, 1981; Steele and Baker, 1993). In view of the support the initial hypotheses formulated by Leakey (1971); Isaac variable weight of the prey, the possible existence of small scale co- (1978), and as such should be reconsidered in socioeconomic operative hunting (involving two to four hunters) and/or large- interpretations of Lower Pleistocene hominin groups. scale hunting involving five or more hunters (Hayden, 1981), must be considered. Social hunting enabled large prey, such as large Acknowledgments bovids and rhinoceroses, to be caught successfully. Steele and Baker (1993) believe that this strategy was possibly one of the most We are deeply grateful to the Atapuerca research team, the important factors that enabled hominins to catch these animals. participants in the fieldwork and J. L. Arsuaga, co-director of the More than one individual must have participated in this type of Atapuerca Project. We are profoundly grateful to our colleague A. transport, which inevitably involved cooperation among individ- Ollé for his in-depth reading of the manuscript, his comments and uals and thus supports the use of the term with regard to strategies our discussions, and to C. Lorenzo and J. I. Morales for their assis- for acquiring animal resources. This suggests the existence of social tance in the statistical interpretations. We would like to thank the cooperation based on the assumptions of Isaac (Isaac, 1978; Isaac Editor, D. Begun, and three anonymous reviewers for their and Crader, 1981). Union amongst individuals was perhaps also comments on the manuscript that have greatly improved the final 444 P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 version. The Ministerio de Ciencia y Investigación (project N Bunn, H.T., 1982. Meat-Eating and human evolution: studies on the diet and CGL2009-12703-C03-02/BTE) and Generalitat de Catalunya (SGR subsistence patterns of Plio-Pleistocene hominins in East Africa. Ph.D. Disser- tation, University of California, Berkeley. 2009-188 and 2009 SGR 324) supported this research. Funding for Bunn, H.T., 1983. Comparative analysis of modern bone assemblages from a San the fieldwork came from Consejería de Cultura y Turismo, Junta de hunter-gatherer Camp in the Kalahari Desert, Botswana, and from a spotted Castilla y León and the Atapuerca Foundation. P. Saladié received hyena den near Nairobi, Kenia. In: Clutton-Brock, J., Grigson, C. (Eds.), Animals and Archaeology: 1 e Hunters and Their Prey. British Archaeological Reports. a grant from the Duques de Soria Foundation-Atapuerca Founda- International Series 163. Oxford. pp. 143e148. tion, and R. Huguet received a grant from the Atapuerca Founda- Bunn, H.T., 1993. Bone assemblages at base camps: a further consideration of tion. A. Rodríguez-Hidalgo has a predoctoral research fellowship carcass transport and bone destruction by the Hadza. In: Hudson, J. (Ed.), From Bones to Behavior. Ethnoarchaelogical and Experimental Contributions to the (FPI) from the Ministry of Science and Innovation of the Spanish Interpretation of Faunal Remains. Center for Archaeological Investigations. Government (CGL2009-12703-C03-02). Southern Illinois University at Carbondale, pp. 156e168. Capaldo, S.D., 1995. Inferring hominid and carnivore behavior from dual patterned archaeofaunal assemblages. Ph.D. Dissertation, Rutgers University. Capaldo, S.D., 1997. Experimental determinations of carcass processing by Plio- References Pleistocene hominids and carnivores at FLK 22 (Zinjanthropus), Olduvai Gorge, Tanzania. J. Hum. Evol. 33, 555e597. Behrensmeyer, A.K., 1978. Taphonomic and ecologic information from bone Capaldo, S.D., 1998. Simulating the formation of dual-patterned archaeofaunal weathering. Paleobiol. 4, 150e162. assemblages with experimental control samples. J. Archaeol. Sci. 25, 311e330. Berger, G.W., Pérez-González, A., Carbonell, E., Arsuaga, J.L., Bermúdez de Carbonell, E., García -Antón, D., Mallol, C., Mosquera Martínez, M., Ollé, A., Castro, J.M., Ku, T.L., 2008. Luminescence chronology of cave sediments at the Rodríguez, X.P., Sahnouni, M., Sala, R., Vergès, J.M., 1999. The TD6 level lithic Atapuerca paleoanthropological site, Spain. J. Hum. Evol. 55, 300e311. industry from Gran Dolina, Atapuerca (Burgos, Spain): production and use. Bermúdez de Castro, J.M., Arsuaga, J.L., Carbonell, E., Rosas, A., Martínez, I., J. Hum. Evol. 37, 653e694. Mosquera, M., 1997. A hominid from the lower Pleistocene of Atapuerca, Spain: Cleghorn, N., Marean, C.W., 2004. Distinguishing selective transport and in situ possible ancestor to Neandertals and modern humans. Science 276, 1392e1395. attrition: a critical review of analytical approaches. J. Taphonomy 2, 43e67. Bermúdez de Castro, J.M., Carbonell, E., Cáceres, I., Díez, J.C., Fernández-Jalvo, Y., Delaney-Rivera, C., Plummer, T.W., Hodgson, J.A., Forrest, F., Hertel, F., Oliver, J.S., Mosquera, M., Ollé, A., Rodríguez, J., Rodríguez, X.P., Rosas, A., Rosell, J., Sala, R., 2009. Pits and pitfalls: taxonomic variability and patterning in tooth mark Vergés, J.M., van der Made, J., 1999. The TD6 (Aurora stratum) hominid site. Final dimensions. J. Archaeol. Sci. 36, 2597e2608. remarks and new questions. J. Hum. Evol. 37, 695e700. Díez, J.C., Fernández-Jalvo, Y., Cáceres, I., Rosell, J., 1999. Zooarchaeology and Bermúdez de Castro, J.M., Carbonell, E., Gómez, A., Mateos, A., Martinón-Torres, M., taphonomy of Aurora stratum (Gran Dolina, Sierra de Atapuerca, Spain). J. Hum. Muela, A., Rodríguez, J., Sarmiento, S., Varela, S., 2006. Paleodemografía del Evol. 37, 623e657. hipodigma de fósiles de homininos del nivel TD6 de Gran Dolina (Sierra de Domínguez-Rodrigo, M., Barba, R., 2006. New estimates of tooth mark and Atapuerca, Burgos): estudio preliminar. Estudios Geológicos 62, 145e154. percussion mark frequencies at the FLK Zinj site: the carnivore-hominid- Bermúdez de Castro, J.M., Pérez-González, A., Martinón-Torres, M., Gómez- carnivore hypothesis failed. J. Hum. Evol. 50, 170e194. Robles, A., Rosell, J., Prado, L., Sarmiento, S., Carbonell, E., 2008. A new early Domínguez-Rodrigo, M., Pickering, T.R., 2003. Early hominid hunting and scav- Pleistocene hominin mandible from Atapuerca-TD6, Spain. J. Hum. Evol. 55, enging: a zooarcheological review. J. Taphonomy 12, 275e282. 729e735. Domínguez-Rodrigo, M., Piqueras, A., 2003. The use of tooth pits to identify Bermúdez de Castro, J.M., Martinón-Torres, M., Prado, L., Gómez-Robles, A., Rosell, J., carnivore taxa in tooth-marked archaeofaunas and their relevance to recon- López-Polín, L., Arsuaga, J.L., Carbonell, E., 2010. New immature hominin fossil struct hominid carcass processing behaviours. J. Archaeol. Sci. 30, 1385e1391. from European Lower Pleistocene shows the earliest evidence of a modern Domínguez-Rodrigo, M., Mabulla, A., Bunn, H.T., Barba, R., Díez-Martín, F., human dental development pattern. Proc. Natl. Acad. Sci. 102, 5674e5678. Egeland, C.P., Espílez, E., Egeland, A., Yravedra, J., Sánchez, P., 2009. Unraveling Binford, L.R., 1978. Nunamiut Ethnoarchaeology. Academic Press, New York. hominin behavior at another anthropogenic site from Olduvai Gorge Binford, L.R., 1981. Bones: Ancient Men and Modern Myths. Academic Press, New (Tanzania): new archaeological and taphonomic research at BK, Upper Bed. York. J. Hum. Evol. 57, 260e283. Binford, L.R., 1984. Faunal Remains from Klasies Rives Mouth. Academic Press, New Domínguez-Rodrigo, M., 1997a. Meat-eating by early hominids at the FLK 22 Zin- York. janthropus site, Olduvai Gorge (Tanzania): an experimental approach using cut- Blumenschine, R.J., Marean, C.W., 1993. A carnivore’s view of archaeological bone mark data. J. Hum. Evol. 33, 669e690. assemblages. In: Hudson, J. (Ed.), From Bones to Behavior. Center for Archaeo- Domínguez-Rodrigo, M., 1997b. A reassessment of the study of cut mark patterns to logical Investigations, Carbondale, pp. 273e301. infer hominin manipulation of fleshed carcasses at the FLK Zinj 22 site, Olduvai Blumenschine, R.J., Selvaggio, M.M., 1988. Percussion marks on bone surfaces as Gorge, Tanzania. Trabajos Prehist. 54, 29e42. a new diagnostic of hominid behavior. Nature 333, 763e765. Domínguez-Rodrigo, M., 1999. Flesh availability and bone modifications in carcasses Blumenschine, R.J., Cavallo, J.A., Capaldo, S.D., 1994. Competition for carcasses and consumed by lions: palaecological relevance in hominid foraging patterns. early hominid behavioral ecology: a case study and conceptual framework. Palaeogeogr. Palaeoclimatol. Palaeoecol. 149, 373e388. J. Hum. Evol. 27, 197e213. Domínguez-Rodrigo, M., 2002. Hunting and scavenging by early humans: the state Blumenschine, R.J., Marean, C.W., Capaldo, S.D., 1996. Blind tests of inter-analyst of the debate. J. World Prehist. 16, 1e54. correspondence and accuracy in the identification of cut marks, percussion Egeland, A., Egeland, C., Bunn, H., 2008. Taphonomic analysis of a modern spotted marks, and carnivore toothmarks on bone surfaces. J. Archaeol. Sci. 23, hyena (Crocuta crocuta) den from Nairobi, Kenya. J. Taphonomy 6, 275e299. 493e507. Emerson, A.M., 1993. The role of body part utility in small-scale hunting under two Blumenschine, R.J., 1986. Carcass consumption sequences and the archaeological strategies of carcass recovery. In: Hudson, J. (Ed.), From Bones to Behavior: distinction of scavenging and hunting. J. Hum. Evol. 15, 639e659. Ethnoarchaeological and Experimental Contributions to the Interpretation of Blumenschine, R.J., 1988. An experimental model of the timing of hominid and Faunal Remains. Center for Archaeological Investigation. Southern Illinois carnivore influence on archaeological bone assemblages. J. Archaeol. Sci. 15, University at Carbondale, pp. 138e155. 483e502. Faith, J.T., Behrensmeyer, A.K., 2006. Changing patterns of carnivore modification in Blumenschine, R.J., 1995. Percussion marks, toothmarks, and experimental deter- a landscape bone assemblage, Amboseli Park, Kenya. J. Archaeol. Sci. 33, minations of the timing of hominid and carnivore access to long bones at FLK 1718e1733. Zinjanthropous, Olduvai Gorge, Tanzania. J. Hum. Evol. 29, 21e51. Faith, J.T., Gordon, A.D., 2007. Skeletal element abundances in archaeofaunal Brain, C.K., 1981. The Hunters or the Hunted? an Introduction to African Cave assemblages: economic utility, sample size, and assessment of carcass transport Taphonomy. University of Chicago Press, Chicago. strategies. J. Archaeol. Sci. 34, 872e882. Bunn, H.T., Ezzo, J.A., 1993. Hunting and scavenging by Plio-Pleistocene hominids: Faith, J.T., Domínguez-Rodrigo, M., Gordon, A.D., 2009. Long-distance carcass nutritional constraints, archaeological patterns, and behavioural implications. transport at Olduvai Gorge? A quantitative examination of Bed I skeletal J. Archaeol. Sci. 20, 365e398. element abundances. J. Hum. Evol. 56, 247e256. Bunn, H.T., Kroll, E.M., 1986. Systematic butchery by Plio/Pleistocene hominids at Falguères, C., Bahain, J.J., Yokoyama, Y., Arsuaga, J.L., Bermúdez, J.M., Carbonell, E., Olduvai Gorge, Tanzania. Curr. Anthropol. 27, 431e452. Bischoff, J.L., Dolo, J.M., 1999. Earliest humans in Europe: the age of TD6 Gran Bunn, H.T., Kroll, E.M., 1988. Reply to Binford: fact and fiction about the Zinjan- Dolina, Atapuerca, Spain. J. Hum. Evol. 37, 343e352. thropus Floor: data, arguments, and interpretations. Curr. Anthropol. 29, Fernández -Jalvo, Y., Díez, J.C., Cáceres, I., Rosell, J., 1999. Human cannibalism in the 123e149. Early Pleistocene of Europe (Gran Dolina, Sierra de Atapuerca, Spain). J. Hum. Bunn, H.T., Harris, J.W.K., Isaac, G., Kaufulu, Z., Kroll, E.M., Schick, K., Toht, N., Evol. 37, 591e622. Behrensmeyer, A.K., 1980. FxJj 50: an early Pleistocene site in northern Kenia. Fisher, J.W., 1995. Bone surface modifications in zooarchaeology. J. Archaeol. World Archaeol. 12, 109e136. Method Theory 2, 7e68. Bunn, H.T., Bartram, L.E., Kroll, E.M., 1988. Variability in bone assemblage formation García, N., Arsuaga, J.L., 1999. Carnivores from the Early Pleistocene hominid- from Hadza hunting, scavenging, and carcass processing. J. Anthropol. Archaeol. bearing Trinchera Dolina 6 (Sierra de Atapuerca, Spain). J. Hum. Evol. 37, 7, 412e457. 415e430. Bunn, H.T., 1981. Archaeological evidence for meat-eating by Plio-Pleistocene García, N., Arsuaga, J.L., 2001. Les carnivores (Mammalia) des sites du Pléistocène hominins from Koobi Fora, Kenya. Nature 291, 574e577. ancien et moyen d’Atapuerca (Espagne). L’Anthropologie 105, 83e93. P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446 445

García, N., 2003. Osos y otros carnívoros de la Sierra de Atapuerca. Fundación Oso Marshall, F., 1994. Food sharing and body part representation in Okiek faunal de Asturias, Oviedo. assemblages. J. Archaeol. Sci. 21, 65e77. Gifford-Gonzalez, D., 1991. Bones are not enough: analogues, knowledge, and Marshall, L., 1998. Sharing, talking, and giving: relief of social tensions among the ! interpretative strategies in zooarchaeology. J. Anthropol. Archaeol. 10, 215e254. Kung. In: Lee, R.B., DeVore, I. (Eds.), Kalahari Hunter-gatherers. Studies of the ! Gifford-Gonzalez, D., 1993. Gaps in the zooarchaeological analyses of butchery: is Kung San and Their Neighbors. Harvard University Press, London, pp. 349e371. gender an issue? In: Hudson, J. (Ed.), From Bones to Behavior. Ethno- Metcalfe, D., Jones, K.T., 1988. A reconsideration of animal body part utility indices. archaelogical and Experimental Contributions to the Interpretation of Faunal Am. Antiq. 53, 486e504. Remains. Center for Archaeological Investigations, Southern Illinois University Milo, R.G., 1998. Evidence for hominid predation at Klassies River Mouth, South at Carbondale, pp. 181e199. Africa, and its implications for the behaviour of early modern humans. Gould, R.A., 1981. Comparative ecology of food-sharing in Australia and Northwest J. Archaeol. Sci. 25, 99e133. California. In: Harding, R.S.O., Teleki, G. (Eds.), Omnivorous Primates. Columbia Monahan, C.M., 1998. The Hadza carcass transport debate revisited and its University Press, New York, pp. 422e454. archaeological implications. J. Archaeol. Sci. 25, 405e424. Hawkes, K., 1993. Why hunter-gatherers work: an ancient version of the problem of Morin, E., 2007. Fat composition and Nunamiut decision-making: a new look at the public goods. Curr. Anthropol. 34, 341e361. marrow and bone grease indices. J. Archaeol. Sci. 34, 69e82. Hayden, B., 1981. Subsistence and ecological adaptations of modern hunter/gath- Munson, P.J., Garniewicz, R.C., 2003. Age-mediated survivorship of ungulate erers. In: Harding, R.S.O., Teleki, G. (Eds.), Omnivorous Primates. Columbia mandibles and teeth in canid-ravaged faunal assemblages. J. Archaeol. Sci. 30, University Press, N.Y, pp. 344e421. 405e416. Horwitz, K.L., 1990. The origin of partially digested bones recovered from archae- Nilssen, P.J., 2000. An actualistic butchery study in South Africa and its implications ological contexts in Israel. Paléoríent 16, 97e106. for reconstructing hominid strategies of carcass acquisition and butchery in the Huguet, R., Cáceres, I., Díez, J.C., Rosell, J., 1999. Estudio tafonómico y zooarqueo- Upper Pleistocene and Plio-Pleistocene. Ph.D. Dissertation, University of Cape lógico de los restos óseos de macromamíferos de la unidad GII de Galería (Sierra Town. de Atapuerca). In: Carbonell, E., Rosas, A., Díez, J.C. (Eds.), Atapuerca: Ocupa- O’Connell, J.F., Hawkes, K., Jones, N.B., 1988. Hadza hunting, butchering, and bone ciones humanas y paleoecología del yacimiento de Galería. Junta de Castilla y transport and their archaeological implications. J. Anthropol. Res. 44, León. Consejería de Educación y Cultura, Zamora, pp. 245e264. 113e161. Isaac, G., Crader, D., 1981. To what extent were early hominids carnivorous? An O’Connell, J.F., Hawkes, K., Jones, N.B., 1990. Reanalysis of large mammal body part Archaeological Perspective. In: Harding, R.S.O., Teleki, G. (Eds.), Omnivorous transport among the Hadza. J. Archaeol. Sci. 17, 301e316. Primates. Columbia University Press, New York, pp. 37e104. Oliver, J.S., 1993. Carcass processing by the Hadza: bone breakage from butchery to Isaac, G., 1978. The food-sharing behaviour of proto-human hominids. Sci. Am. 238, consumption. In: Hudson, J. (Ed.), From Bones to Behavior. Ethnoarchaelogical 90e108. and Experimental Contributions to the Interpretation of Faunal Remains. Center Johnson, E., 1985. Current developments in bone technology. In: Schiffer, M. (Ed.), for Archaeological Investigations, Southern Illinois University at Carbondale, pp. Advances in Archaeological Method and Theory. Academic Press, New York, pp. 200e227. 157e235. Ollé, A., García, P., Vergès, J.M., Guardiola, M., Guiu, J., Geribàs, N., Cáceres, I., Klein, R.G., Cruz-Uribe, K., 1984. The Analysis of Animal Bones from Archaeological Esteban, M., Saladié, P., Rosell, J., Mestre, J., Roberts, M.B., 2010. Experimental Sites. University of Chicago Press, Chicago. Knapping and Butchery: Replicating Boxgrove. In: Ancient human occupation of Lam, Y.M., Pearson, O.M., 2005. Bone density studies and the interpretation of the Britain conference British Museum, April 8the10th 2010, Londres. faunal record. Evol. Anthropol. 14, 99e108. Pales, L., Lambert, P., 1971. Atlas ostéologique des mammifères. Paris. Lam, Y.M., Chen, X., Pearson, O.M., 1999. Intertaxonomic Variability in patterns of Parés, J.M., Pérez-González, A., 1995. Paleomagnetic age for hominid fossils at bone density and the differential representation of bovid, cervid, and equid Atapuerca Archaeological site, Spain. Science 269, 830e832. elements in the archaeological record. Am. Antiq. 64, 343e362. Parés, J.M., Pérez-González, A., 1999. Magnetochronology and stratigraphy at Gran Leakey, M.D., 1971. Olduvai Gorge: Excavations in Beds I and II, 1960e1963. Cam- Dolina section, Atapuerca (Burgos, Spain). J. Hum. Evol. 37, 325e342. bridge University Press, Cambridge. Pérez-González, A., Parés, J.M., Carbonell, E., Aleixandre, T., Ortega, A.I., Benito, M.A., Lee, R.B., 1981. La subsistencia de los bosquimanos !Kung: un análisis de input- Martín-Merino, M.A., 2001. Géologie de la Sierra de Atapuerca et stratigraphie output. In: Llobera, J.R. (Ed.), Antropologia Económica. Editorial Anagrama, des remplissages kasrtiques de Galería et Dolina (Burgos, Espagne). L’An- Barcelona, pp. 35e64. thropologie 105, 27e44. Lupo, K.D., O’Connell, J.F., 2002. Cut and tooth marks distributions on large animal Perkins, D., Daly, P., 1968. A hunter’s village in Neolithic Turkey. Sci. Am. 219, bones: ethnoarchaeological data from the Hadza and their implications for 96e106. current ideas about early human carnivory. J. Archaeol. Sci. 29, 85e109. Pickering, T.R., Marean, C.W., Domínguez-Rodrigo, M., 2003. Importance of limb Lupo, K.D., 1999. Archaeological skeletal part profiles and differential transport: an bone shaft fragments in zooarchaeology: a response to “On in situ attrition and ethnoarchaeological example from Hadza bone assemblages. J. Anthropol. vertebrate body part profiles” (2002), by M.C. Stiner. J. Archaeol. Sci. 30, Archaeol. 20, 361e378. 1469e1482. Lyman, R.L., 1993. Density-mediated attrition of bone assemblages: new insights. Pickering, M., Domínguez-Rodrigo, M., Egeland, C.P., Brain, C.K., 2005. The contri- In: Hudson, J. (Ed.), From Bones to Behavior. Ethnoarchaelogical and Experi- bution of limb bone fracture patterns to reconstructing early hominid behavior mental Contributions to the Interpretation of Faunal Remains. Center for at Swartkrans caves (South Africa): archaeological application of a new Archaeological Investigations, Southern Illinois University at Carbondale, pp. analytical method. Int. J. Osteoarchaeol. 15, 247e260. 325e341. Pickering, T.R., 2001. Taphonomy of the Swartkrans hominid postcrania and its Lyman, R.L., 1994. Vertebrate Taphonomy. Cambridge University Press, Cambridge. bearing on issues of meat-eating and fire management. In: Stanford, C.B., Lyman, R.L., 2008. Quantitative Paleozoology. Cambridge University Press, New Bunn, H. (Eds.), Meat-eating and Human Evolution. Oxford University Press, York. New York, pp. 33e51. Made van der, J., 2001. Les ongulés d’Atapuerca. Stratigraphie et biogéographie. Pobiner, B.L., Rogers, M.J., Monahan, C.M., Harris, J.W.K., 2008. New evidence for L’Anthropologie 105, 95e113. hominin carcass processing strategies at 1.5 Ma, Koobi Fora, Kenya. J. Hum. Evol. Madrigal, T.G., Holt, J.Z., 2002. White-Tailed deer meat and marrow return rates 55, 103e130. and their application to Eastern Woodlands archaeology. Am. Antiq. 67, Rodríguez, J., Burjachs, F., Cuenca-Bescós, G., García, N., Made van der, J., Pérez- 745e759. González, A., Blain, H.A., Expósito, I., López-García, J.M., García -Antón, D., Maguire, D.J., Pemberton, D., Collett, M.H., 1980. The Makapansgat limeworks Allué, E., Cáceres, I., Huguet, R., Mosquera, M., Ollé, A., Rosell, J., Parés, J.M., grey breccia: hominids, hyaenas, hystricids or hillwash? Palaeont. Afr. 23, Rodríguez, X.P., Díez, J.C., Rofes, J., Sala, R., Saladié, P., Vallverdú, J., Bennasar, M., 75e98. Blasco, R., Bermúdez de Castro, J.M., Carbonell, E., 2010. One million years of Marean, C.W., Cleghorn, N., 2003. Large mammal skeletal element transport: cultural evolution in a stable environment at Atapuerca (Burgos, Spain). Qua- applying foraging theory in a complex taphonomic system. J. Taphonomy, tern. Sci. Rev. 30, 1396e1412. 15e42. Rodríguez, J., 1997. Análisis de la estructura de las comunidades de mamíferos del Marean, C.W., Kim, S.Y., 1998. Mousterian large-mammal remains from Kobeh cave. Pleistoceno de la Sierra de Atapuerca. Revisión de metodologías. Ph. Disserta- Behavioral implications for Neanderthals and early modern Humans. Curr. tion. Universidad Autónoma, Madrid. Anthropol. 39, S79eS113. Rosell, J., 2001. Patrons d’aprofitament de les biomasses animals durant el Pleistocè Marean, C.W., Spencer, L.M., 1991. Impact of carnivore ravaging on zooarchaeo- Inferior i Mig (Sierra de Atapuerca, Burgos) i Superior (Abric Romaní, Barce- logical measures of element abundance. Am. Antiq. 56, 645e658. lona). Ph.D.Dissertation, Universitat Rovira i Virgili, Tarragona. Marean, C.W., Spencer, L.M., Blumenschine, R.J., Capaldo, S.D., 1992. Captive hyaena Saladié, P., 2009. Mossegades d’omnívors. Aproximació experimental i aplicació bone choice and destruction, the schlepp effect and Olduvai archaeofaunas. zooarqueològica als jaciments de la Sierra de Atapuerca. Ph.D. Dissertation, J. Archaeol. Sci. 19, 101e121. Universitat Rovira i Virgili, Tarragona. Marean, C.W., Abe, Y., Frey, C.J., Randall, R.C., 2000. Zooarchaeological and tapho- Selvaggio, M.M., Wilder, J., 2001. Identifying the involvement of multiple carni- nomic analysis of the Die Kelders cave 1 layers 10 and 11 Middle Stone Age vore taxa with archaeological bone assemblages. J. Archaeol. Sci. 28, larger mammal fauna. J. Hum. Evol. 38, 197e233. 465e470. Marean, C.W., Domínguez-Rodrigo, M., Pickering, T.R., 2004. Skeletal element Selvaggio, M.M., 1994a. Carnivore tooth marks and stone tool butchery marks on equifinality in zooarchaeology begins with method: the evolution and status of scavenged bones: archaeological implications. J. Hum. Evol. 27, 215e228. the ‘‘shaft critique’’. J. Taphonomy 2, 69e98. Selvaggio, M.M., 1994b. Evidence from carnivore tooth marks and stone-Tool- Marean, C.W., 1991. Measuring the post-depositional destruction of bone in Butchery marks for scavenging by hominids at FLK Zinjanthropus Olduvai archaeological assemblages. J. Archaeol. Sci. 18, 677e694. Gorge, Tanzania. Ph.D. Dissertation, Rutgers University. 446 P. Saladié et al. / Journal of Human Evolution 61 (2011) 425e446

Selvaggio, M.M., 1998. Concerning the three stage model of carcass processing at Tappen, M., Wrangham, R., 2000. Recognizing hominoid-modified bones: the FLK Zinjanthropus: a reply to Capaldo. J. Hum. Evol. 35, 319e321. taphonomy of Colobus bones partially digested by free-ranging chimpanzees in Shipman, P., Rose, J., 1983. Early hominid hunting, butchering, and carcass- the Kibale Forest, Uganda. Am. J. Phys. Anthropol. 113, 217e234. processing behaviors: approaches to the fossil record. J. Anthropol. Archaeol. Thompson, J.C., 2010. Taphonomic analysis of the Middle stone age faunal assem- 2, 57e98. blage from Pinnacle point cave 13B, Western Cape, South Africa. J. Hum. Evol. Steele, D.G., Baker, B.W., 1993. Multiple predation: a definitive human hunting 59, 321e339. strategy. In: Hudson, J. (Ed.), From Bones to Behavior. Ethnoarchaelogical and Todd, L.C., Rapson, D.J., 1988. Long bone fragmentation and interpretation of faunal Experimental Contributions to the Interpretation of Faunal Remains. Center for assemblages: approaches to comparative analysis. J. Archaeol. Sci. 15, 307e325. Archaeological Investigations, Southern Illinois University at Carbondale, pp. Vallverdú, J., Courty, M.A., Carbonell, E., Canals, A., Burjachs, F., 2001. Les sédiments 9e37. d’Homo antecessor de Gran Dolina (Sierra de Atapuerca, Burgos, Espagne). Stiner, M.C., Barkai, R., Gopher, A., 2009. Cooperative hunting and meat sharing Interprétation micromorphologique des processus de formation et enregistre- 400e200 ky at Qesem Cave. Israel. Proc. Natl. Acad. Sci. 106, 13207e13212. ment paléoenvironnemental des sédiments. L’Anthropologie 105, 45e71. Stiner, M.C., 1994. Honor among thieves: a zooarchaeological study of Neandertal Villa, P., Mahieu, E., 1991. Breakage patterns of human long bones. J. Hum. Evol. 21, ecology. Princeton University Press, Princeton. 27e48. Stiner, M.C., 2002. On in situ attrition and vertebrate body part profiles. J. Archaeol. White, T.D., 1992. Prehistoric Cannibalism at Mancos 5MTUMR-2346. Princeton Sci. 29, 979e991. University Press, Princeton. Suby, J.A., 2006. Estudio metodológico-comparativo de densidad mineral ósea de Yellen, J.E., 1977. Cultural patterning in faunal remains: evidence from the !Kung restos humanos por Absorciometría Fotónica (DXA). Intersecciones en antro- bushmen. In: Ingersoll, D., Yellen, J.E., Macdonald, W. (Eds.), Experimental pología 7, 277e285. Archeology. Columbia University Press, New York, pp. 271e331.