Reassessing Hominin Skills at an Early Middle Pleistocene Hippo Butchery Site: Gombore II-2 (Melka Kunture, Upper Awash valley, Ethiopia)
Single-carcass sites of Lower and Middle Pleistocene age have attracted much attention since they were first recorded. They have been the focus both of science and of museum displays, with reconstructions of “hominins-feasting-on-a-carcass” purposefully illustrating a major step in human evolution. Here we report the Acheulean site Gombore II-2 in the upper Awash Valley of Ethiopia, dating to 0.7 Myr. In the 1970s, due to the presence of hippo remains, the site was published as a single-carcass butchering site. New excavations revealed an ichnosurface displaying animal and human footprints associated with bones and lithics. Subsequent studies of lithic and faunal remains of recent and past excavations as well as archive studies show that Gombore II-2 represents one of the earliest sites with hominin-hippo interaction. The hippo remains belong to a minimum of three carcasses, at least one of them butchered by hominins and subsequently ravaged by hyenas. However, instead of single carcasses exploited on the spot, evidence suggests the existence of a living floor where butchering episodes were performed through time, possibly transporting portions from scavenging sites at a distance. Gombore II-2 thus provides unique insight into planning capacities and control over the environment probably by early representatives of Homo heidelbergensis.
KeywordAcheulean Butchering site Hippopotamus Melka Kunture Gombore II-2
Animal butchery can be traced back as early as 2.6–2.5 million years ago (Ma) (de Heinzelin et al. 1999; Domínguez-Rodrigo et al. 2005) or even earlier (McPherron et al. 2010; Domínguez-Rodrigo et al. 2011). At around 2 Ma, hominins exploited small- and middle-sized terrestrial and aquatic fauna (Braun et al. 2010). During these early stages of becoming human, we also witness the intentional butchering of megafauna (e.g. Leakey 1971; Isaac and Crader 1981; Isaac and Isaac 1989; Bunn 1997; Pobiner et al. 2008; Domínguez-Rodrigo et al. 2010; Sanhouni et al. 2013; Merritt 2017; Pante et al. 2018). At several Lower and Middle Pleistocene sites, mostly discovered during the 1970s, complete or articulated parts of megafauna carcasses were discovered in association with lithic tools. As even parts of megafauna carcasses are difficult to transport, it was suggested that human groups exploited these single carcasses on the spot. These single carcass sites even became formally classified by Isaac and Crader as “B type sites,” i.e. “sites in which artifacts are found with bones of the carcass of a single large animal” (Isaac and Crader 1981: 43). Reconstructions depicting hominin groups during the exploitation of a carcass acquired iconic status, which is retained today in the media and in countless museum reconstructions.
Over the last 50 years, single carcass sites have been the focus of much scientific attention. Several discoveries were made at African Lower Pleistocene sites, mostly deposited in a low-energy event: Mwanganda’s Village (Clark and Haynes 1970; Thompson et al. 2013), WK Hippo Cliff (found in 1970; Leakey and Roe 1994), FLK N6 and FLKN Deinotherium (Leakey 1971; Domínguez-Rodrigo et al. 2007, 2012), FXJj3 at Koobi Fora (found in 1971-1972; Isaac and Crader 1981; Isaac and Isaac 1989; Bunn 1997), Gadeb 8F (found in 1975-1977; Clark 1987), Olorgesailie Hippo Banda Site (Isaac 1977) and Barogali (found in 1985-1987; Chavaillon et al. 1987; Berthelet et al. 1992; Chavaillon and Berthelet 2001; Berthelet 2001). Single carcass sites were also documented during the Middle and early Upper Pleistocene in Europe, as for instance at Venosa Notarchirico, Aridos 1 and 2, La Polledrara, Cava Campitello, Lehringen, Gröbern and Pagnano d’Asolo, to mention just some of them (Santonja et al. 1980; Weber 2000; Mussi et al. 2005; Santonja and Péres-Gonzàles 2005; Mussi and Villa 2008; Yravedra et al. 2010; Santucci et al. 2015).
The Lower to Middle Pleistocene African record lacks evidence for active hunting of megafauna. Due to a research paradigm assuming that early hominins were equipped with only limited cognitive and technological capacities, the absence of evidence has been interpreted in terms of hominins taking advantage of naturally deceased large animals (Domínguez-Rodrigo 2002; Domínguez-Rodrigo and Pickering 2003; Surovell et al. 2005; Domínguez-Rodrigo and Barba 2006; Bunn and Pickering 2010; Domínguez-Rodrigo et al. 2010, 2014a; Ben-Dor et al. 2011; Lupo and Schmitt 2016; Agam and Barkai 2018). Megafauna exploitation, be it by hunting or scavenging, implies technological skills, has social implications as well as dietary consequences (Isaac and Crader 1981; Ben-Dor et al. 2011; Domínguez-Rodrigo et al. 2014a; Domínguez-Rodrigo and Pickering 2017; Agam and Barkai 2018) and demonstrates successful competition over carnivores (e.g. Espigares et al. 2013). Accordingly, single carcass sites are visible proof that a major step in human evolution was made. We will focus here on a case study demonstrating hominin interaction with hippos. Today, the hippopotamus is a highly aggressive species (Eltringham 1999), and we assume that this was also the case in the past, though we know that aggression level in hippopotamus corresponds to population density (Attwell 1963; Eltringham 1999). Weighing an average 1.4 tons, including 5% of fat, a hippo provides more than 900,000 calories (Eltringham 1999; Ben-Dor et al. 2011).
Gombore II-2, dated to ~0.7 Ma, is one of the earliest evidences of hominin/hippo interaction. It is mentioned in the literature as an example of a single carcass butchering site (e.g. Delagnes et al. 2006). The site was first investigated in 1974. Detailed information on stratigraphy, site formation, find distribution, taphonomy and archaeozoology are however lacking. The lithic remains were described in a brief, partial and, to some extent, contradictory report by Chavaillon and Berthelet (2004). An updated overview of the site is provided by Mussi et al. (2016) and Altamura et al. (2018) based on results of new excavations. Here we report the complete archeological and zoorchaeological evidence for Gombore II-2 and evaluate its interpretation as a “single-carcass butchery site.” Moreover, we evaluate the implications of single-carcass butchery sites for human evolution.
History of Research
Earlier Research at Gombore II-2 (1974–1995)
The Re-excavation of Gombore II-2 (2012–2015)
In 2012, the team of the Italian Archeological Mission at Melka Kunture and Balchit, directed by one of us (MM), started cleaning the area, bringing to light dozens of finds on the previously excavated surface and along the standing walls. A trench was also opened to record the stratigraphic sequence between Gombore II-2 and Gombore II-1, a middle Acheulean site located 6 m downslope (Fig. 3) and dated to 0.85 Ma (Gallotti et al. 2010; Mussi et al. 2016; Altamura and Mussi 2017).
Consecutive excavations headed by the first author (FA) were undertaken between 2013 and 2015, reestablishing and expanding the Chavaillon-excavation grid southwards towards still intact deposits. A further 35 m2 were investigated, doubling the previously researched exposure (Fig. 2).
Materials and Methods
Excavations and Recording
The lithostratigraphic sequence was reinvestigated between 2012 and 2015. Fieldwork at the site included a systematic description of the exposed section of previous excavations to reconstruct the lithostratigraphic succession and paleoenvironment. Sedimentological features were additionally documented for the re-evaluation of the site formation processes. Undisturbed samples of layer 2 and 3 were collected for micromorphological analyses, while for grain size and petrographic data we follow Raynal and Kieffer (2004) and Raynal et al. (2004).
For the determination of raw materials we followed Kieffer et al. (2002). A variety of raw materials of volcanic origin were classified and described in the area of Melka Kunture, exposed both at primary outcrops and in secondary deposits, mostly as pebbles in alluvia. A dozen types of rock were exploited at Melka Kunture in varying intensity. Obsidian in particular was abundantly used throughout the Pleistocene (Piperno et al. 2008; Gallotti and Mussi 2017).
As regards techno-typological and use-wear analysis of lithic implements, stone tool analysis identified the chaine opératoires from raw material provisioning to production, use and discard. Studies followed terminology and technological concepts by Inizan et al. (1999). Pseudo-retouch was documented as a possible indicator for trampling. Ashes produced by volcanic activities in the immediate vicinity of Melka Kunture contained aggressive chemicals that affected the surfaces of tools differently depending on the raw-material used. Obsidian, which is a volcanic glass, is rarely if ever affected, while the surface of some lavas and basalts rapidly underwent modification. Thus we refrained from documenting abrasion stages, which, in a different geological context, can be used to discriminate between lithics deposited in situ and those that were transported/moved. Artifacts made from obsidian were often dehydrated, displaying patinated surfaces, while other volcanic rocks used for stone tool manufacture were abraded. Furthermore the brittle edges of the artifacts would have been heavily damaged by the production of casts using silicone or similar material. Accordingly, we refrained from traceological studies.
Zooarchaeological and taphonomic analysis at Gombore II-2 aimed at reconstructing the taphonomic history of the faunal assemblage and at evaluating the role of hominins in assemblage formation. Of primary importance was therefore the identification of hominin induced modifications in the form of cut-marks and traces of marrow extraction in addition to the documentation of skeletal element representation. Traces of carnivore modification and bone weathering were additionally documented to identify further variables within the taphonomic chain before burial. Analysis of skeletal element representation with reference to hydrodynamic sorting, the documentation of fragment size, breakage patterns, abrasion and the presence of carbonates together served to evaluate the formation process of the faunal assemblage.
Taxonomic determination for Melka Kunture was mainly undertaken by D. Geraads (Geraads et al. 2004). His primary, generally unpublished data on Gombore II-2 were made available to the authors and form the basis of the current study. Taking into account the highly fragmented state of the bone assemblage, the taxonomically indeterminable material was sorted according to size classes, based on bone-thickness and taking into account the variety of species identified by Geraads et al. over the years (Geraads et al. 2004). Size class sorting goes back to taphonomic studies undertaken by Fiore and Tagliacozzo in 2004 for Garba IV, another site of Melka Kunture (Fiore and Tagliacozzo 2004). Five size classes were defined. Size class 5 covers large sized taxa such as elephants, hippopotamus and rhinoceros. With size class 4, the larger bovids (Connochaetes and Pelorovis) are covered. Size class 3 represents medium sized bovids (such as Damaliscus and Parmularius) and equids. With size class 2, the small bovids (such as Gazella and Antidorcas) are represented. Finally, size class 1 equates to mammals below the size of Gazella. Bone epiphyseal fusion stages as well as tooth eruption and wear patterns were additionally documented, aiming to collate population data in order to describe the section of the biocoenose, which survived at Gombore II-2. However, due to poor bone preservation, the overwhelming majority of bones could not be included in this part of the analysis.
Skeletal Element Representation
The zooarchaeological study included the attribution of bones to species and skeletal element. Both variables were recorded, using the number of identified specimen per taxon (NISP), the minimum number of elements (MNE) and the minimum number of individuals (MNI) (Binford 1981; Lyman 1994). For calculating the taxonomic frequency, all elements assigned to a particular taxon were counted (NISP values), including epiphyses and long bone shaft fragments. The most frequent anatomical part of a particular taxon served for MNI calculation, also taking into account body side, age and sex. MNE values were obtained by counting the epiphyses of a bone element per body side and those long bone shaft fragments whose anatomical position could be exactly determined per bone-landmark-analysis (Lam et al. 1998; Morlan 1994). Discrete patterns of skeletal element survival, such as representation of different anatomical units and intra-elemental survivorship for each taxon should provide data concerning hydrodynamic sorting, density-related bone loss and the agent of bone destruction.
Possible hydrodynamic sorting was investigated through an analysis of skeletal element representation. Fluvial transport can lead to a selective loss of skeletal elements more easily transportable than others. It results in density mediated bone preservation, and thus the evaluation of skeletal element selection by hydrodynamic sorting should start by asking whether the skeletal element representation displays density-mediated bone survival. Bone density has been shown to influence the survival potential of bones during certain pre- and post-depositional taphonomic processes (Lyman 1994). Bones with low mineral density are more easily affected by density mediated attrition than high-density bones. Density values have been recorded for species of different size groups. It could be demonstrated that inter-taxonomical variability in bone density is generally low (Lam et al. 1999). This means that the available data can be applied to the analysis of a variety of species, as long as the analyzed species are from the same size group. This analytical step could not be undertaken at Gombore II-2 as density values for hippo, the dominant species, are not available for all skeletal elements and aquatic mammals show an increased deposition of compact bone compared with terrestrial mammals (Wall 1983) for which density values are available. Consequently, the evaluation of hydrodynamic sorting was restricted to the evaluation of skeletal element representation according to Voorhies Groups. The transport potential of different bones (density, size and shape) is known from experiments in flume channels and from experimental studies (Voorhies 1969; Behrensmeyer 1975). Depending on their susceptibility to fluvial transport, bones are grouped together representing selected carcass parts characteristic of lag, winnowed and transported faunal assemblages. The susceptibility of a skeletal element to fluvial transport also depends on taxon and the degree of fragmentation (Fernandez-Jalvo and Andrews 2016).
Size Distribution and Breakage Patterns
Regarding the degree of bone fragmentation, length, width and breadth were documented for all bone fragments. The documentation of breakage patterns followed Villa and Mahieu (1991). In addition to green and dry bone breaking, bending breaks and occasional step breaks were also documented.
Bone Surface Modifications
In order to estimate exposure time between the death of an animal and its final burial, the bone weathering stage was documented as defined by Behrensmeyer (1978). The classification and interpretation of bone-weathering stages is relevant for eastern African as well as for a number of other environments (Tappen 1994) and thus can also be used for the interpretation of weathering stages in Ethiopian assemblages.
Stages of Abrasion
Differing stages of abrasion were defined specifically for the current study. Bones in mint condition without evidence of alteration visible with the bare eye were assigned to Abrasional Stage 0 (SI Fig. 14). Abrasional Stage 1 describes skeletal elements with slight abrasion visible on edges and ridges, affecting bone proximal and/or distal epiphyses. Abraded skeletal elements with protuberances still well visible were attributed to Abrasional Stage 2 (SI Fig. 14). Abrasional Stage 3 describes bones characterized by heavy abrasion, with bone protuberances hardly visible (SI Fig. 14). Bones with very heavy abrasion but retaining their general shape are attributed to Abrasional Stage 4 (SI Fig. 14), and, finally, completely rolled bones account for Abrasional Stage 5.
For each bone, the presence/absence of carbonate concretions was registered. The documentation distinguished between the presence/absence of carbonate concretions and noted when the entire bone surface was completely covered.
The activity of carnivores results in a variety of bone modifications among them: scratches on the bone surface, gnaw marks, puncture marks and traces of digestion. Binford (1981), Haynes (1983), Blumenschine and Selvaggio (1988, 1991) and Blumenschine (1995) identified criteria relevant for the identification of carnivore damage.
Anthropic Bone Modification
Bone surfaces were studied using 10–20× hand lenses and a Dino-Lite Edge 5MP. All traces were registered per bone and recorded according to their anatomical position. In the process of the taphonomic chain of events, bone surface modifications are continuously altered, implying that interpretations of biotic and abiotic induced modifications, according to diagnostic templates but without their integration into the complete biostratinomic sequence, are highly ambiguous. Depending on the taxon, bone morphology and structure, bone modifications that pass through the same biostratinomic sequence can display completely different patterns of bone preservation and bone modification. As a consequence for the diagnosis and interpretation of bone modifications, a strict determination of the time of cessation of the biostratinomic process is mandatory (Gaudzinski-Windheuser et al. 2010; Gaudzinski-Windheuser and Kindler 2012). Against the background of the lithostratigraphic facies preserved at Gombore, the variables employed to disentangle the faunal record helped the reconstruction of part of the biostratinomic sequence. And against this background, a prioritized two-stage procedure for the identification of cut-marks (according to criteria defined by Potts and Shipman 1981; Shipman 1986; Domínguez-Rodrigo et al. 2009; Fernandez-Jalvo and Andrews 2016) was employed that also takes into account the complete absence of (micro-)striations on the bone surfaces of the Gombore fauna: (1) v-shaped cross-section and straight orientation and (2) absence of accompanying/overlapping striations and microstriations and the anatomical location. Semi-circular notches, sometimes accompanied by intentional or unintentional retouch, were defined as hominin induced impact marks, according to the criteria defined by Blumenschine and Selvaggio (1988, 1991) and Domínguez-Rodrigo and Barba (2006).
Lithostratigraphic, Archaeological and Paleoenvironmental Reconstruction
Layer 4b and 4a: Layer 4b is a fluvial silty-sandy layer and is covered by a fluvial sand deposit (layer 4a) and by the tuff layer 3. The layer 4a is up to 0.1 m thick and preserved over 20 m2 but absent in the westernmost part of the excavated area. Accordingly, over the remaining western 15 m2, the layer 4b is sealed directly by the tuff layer 3. The surface of layer 4b was heavily trampled by animals (hippos, bovids and other small mammals, birds and hominins). Hundreds of fossil footprints, often superimposed, were exposed, especially around a small depression, 0.1 m deep, filled by the sands of layer 4a (Fig. 5). We suggest that this was originally a small shallow water swamp (backswamp) (Altamura et al. 2018). The ichnological evidence is directly associated with an abundant archaeological record (Fig. 6). Layer 4a is a sand layer, which yielded 160 finds: 60 lithic implements (3 half-buried ones were left in situ, see §2.2) and 100 faunal remains. The finds were mostly recovered within the sand or at its base, i.e. almost in contact with the underlying layer 4b. Only few specimens were found included within the sandy infill of the fossil footprints or at the surface of layer 4a outcropping from the sand, which partially buried them. The latter ones were also directly covered by the overlying tuff (i.e. layer 3). As mentioned earlier, layer 4b was left in situ to preserve the ichnosurface, but 12 lithic tools and 16 bone remains were recovered from the very top of the layer. The orientation pattern of elongated finds is isotropic (Fig. 6), indicating no major displacement of finds (e.g. by water) before their final burial. Preservation and the spatial orientation patterns of the archaeological assemblage, as well as the formation and preservation of footprints, point to a floodplain paleo-environment characterized by slight floods and constant humidity (Altamura et al. 2018).
Layer 3 and 2: The two layers have both been described by Raynal et al. (2004) as dacitic tuff resulting from distal volcanic ash fall in water. These clayey-silty deposits, 0.9 m thick, are massive and light-grey in color. The two layers are separated by a thin black-colored lens, rich in diatoms (Mussi et al. 2016). Archaeological remains were mainly documented at the bottom of layer 3, which directly sealed the floodplain paleo-landscape of layer 4.
Layer 1: This layer is up to 0.6 m thick of alternating silts, sands and pumice with planar cross-stratification, which was deposited in a floodplain environment. Layer 1 was dated by Morgan et al. (2012) to 0.709 ± 0.013 Ma and yielded an eroded and partially preserved archaeological deposit.
We traced 29 faunal remains and five lithics from the 1974 trench. We also relocated and re-analyzed 114 of the 133 finds, including 31 lithic implements, found in 1993 and 1995 and re-attributed to layer 4. The missing pieces were characterized making use of preliminary descriptions available in the contemporary inventories. The 1974 inventory further describes two “retouched” long bone fragments as well as a possible “utilized” rib fragment (Chavaillon 1975b). Unfortunately, these remains were not relocated and are not available for revision.
Typological characterization and average dimensions of the lithic industry attributed to layer 4 (all excavations)
2013–2015 layer 4a
2013–2015 layer 4b
Pebble with fractures and/or percussion marks
Other tools on flake
Unmodified material (small pebbles)
Average length for lithic artifacts (cm)
The 1974 and 1993–1995 campaigns
A total of 74 artifacts was unearthed from layer 4 (n = 26 from the 1974 trench and n = 48 from the 1993–1995 excavations). Thirty-six of these were available for re-analysis. Some missing pieces described in the original documentation as well as in Chavaillon and Berthelet (2004) could also be included. Thus, only four finds from the 1974 campaign remained unconsidered.
The debitage comprises nina cores, 14 flakes, six retouched flakes and seven formal tools, i.e. denticulates and side-scrapers. The cores have been exploited in an unsystematic way: a maximum of two series of short flakes were extracted from each knapping surface. They mostly attest either a multidirectional or centripetal exploitation method or a single knapping surface method, initialized on the scar left by a transversal opening flake. Once the knapping surface provided no more favorable angles and volumetric features, the cores were discarded without any further rejuvenation or shaping out attempt, suggesting an opportunistic behavior. Two rabots were also discovered.
Two small roundish stone fragments, less than 4 cm in length, mentioned in the documentation probably represent manuports.
The 2013, 2014 and 2015 Campaigns
The debitage includes 38 flakes, mostly fragments or flaking waste. Even though not confirmed by refitting studies, this suggests in situ knapping. Obsidian flakes (n = 23) have an average length of 4.3 cm, while flakes made from basalt, ignimbrite and other volcanic rocks are generally bigger, with 6.5 cm in average length. The butts are mostly simple (flat, punctiform and cortical); only four flakes have a facetted or dihedral butt. The dorsal negative scars suggest simple knapping methods: the direct percussion of blocks and pebbles from multiple striking platforms, mostly showing unipolar recurrent exploitation, or a centripetal method. A pebble flake is probably the result of bipolar flaking; another specimen possibly proceeds from façonnage, as from the shaping of a handaxe. Three basalt cores were found. Two are small residual flake cores, exploited multidirectionally. The third and largest one (15 × 13 × 7.5 cm; Fig. 10: 10) has two convex and tangential surfaces with evidence of unsystematic centripetal exploitation. Retouch is rare. Three obsidian and flint flakes were retouched into side-scrapers (two déjetés and a convex one, Fig. 10: 5–7). Two more retouched flakes, a scaled piece and a notch (Fig. 10: 9) are also documented.
Eleven pieces from layer 4b were available for study. Ten are of basalt and only one is an obsidian artifact. They have an average length of 7.2 cm. Five pebbles have fractures or percussion marks. Two basalt cores are further evidence of multidirectional debitage or debitage from opposite striking platforms. A third core was re-used as a rabot. The debitage includes a retouched flake made of obsidian and a basalt flake produced by centripetal flaking. Another basalt flake (Fig. 10: 8), with a large and flat butt, was modified by direct retouch into a double convex side-scraper.
Three small heavily-rolled gravel pieces, less than 2 cm in length, were found in the fine-grained layer 4a. They are not artificially modified and are possible manuports.
Generally speaking the lithic collection is well preserved and does not show substantial signs of fluvial transport over distance: the obsidian implements are lightly patinated and show relatively ‘fresh’ edges. Evidence of trampling, indicated by pseudo-retouch and edge-modifications, were not documented.
Composition of the faunal assemblage from Gombore II-2, layer 4 according to the number of individual specimen per taxon (NISP) and the minimum number of individuals (MNI)
Hippopotamus cf. amphibius
Size class 5
Size class 4
Size class 3
Size class 2
Not attributable to size class
Composition of size classes 4–2 of the faunal assemblage from Gombore II-2
Size class 4
Size class 3
Size class 2
n = 5
n = 7
n = 25
In order to assess whether the character of the skeletal element representation documented for Gombore II-2 results from hydrodynamic sorting, assemblage composition was considered according to its susceptibility to fluvial transport, expressed in the so-called Voorhies groups. Ribs, vertebrae, sternum and sacrum are subsumed under Group I as these bones are immediately moved by water. Scapulae, phalanges and ulnae belong to the intermediate Group I/II. Group II summarizes bones that are gradually removed, while the skull and the mandible represent Group III, considered to represent a lag deposit (Behrensmeyer 1975). For Gombore II-2 skeletal elements of Groups I–III have been recorded; most of the bones present belong to dispersal Groups I and I/II. Behrensmeyer (1975) notes that the presence of Group I bones indicates a non-fluvially winnowed assemblage.
Hydrodynamic sorting is not evident but there is obviously a selection according to shape. Of the teeth, rod shaped elements dominate, and, for the bones, it is the blade shaped elements such as scapulae and bodies of vertebrae that dominate. Frostick and Reid (1983) point out that skeletal elements whose shape approximates rods and spheres are more easily transported than skeletal elements equivalent to blades and discs. This suggests that the selection according to shape outlined for Gombore II-2 is not a result of hydrodynamic transport, as was already indicated by the presence of bones representing all Voorhies Groups.
State of Fossil Preservation
The bones from Gombore II-2 showed different stages of preservation. Of some bones, only the trabecular parts survived, though this was only documented for small fragments. For the trabecular bones, fragment sizes remained undocumented. Of other bones, only the outer surface was present, but they were otherwise completely disintegrated. These elements could not be considered for further analysis. The faunal assemblage was affected by coatings of carbonate concretion (n = 79), which covered the bone surfaces to different degrees for all size classes (size class 2: n = 12, size class 3: n = 3, size class 4: n = 1, size class 5: n = 9, size class indet.: n = 27). This strongly supports the assumption that the burial milieu was comparable for bones of all size classes. It appeared that concretions had been partly removed due to post-excavation bone treatment. Therefore, only the presence/absence of these surface coverings was documented. In cases where size evaluation was distorted by concretion cover, sizes remained undocumented. The material unearthed from 2013 to 2015 was particularly strongly affected by surface coatings. However, the majority of these fragments belonged to the smallest fragment size.
The faunal assemblage from Gombore II-2 is highly fragmented. Spiral fractures are the prevailing fracture pattern documented for all size classes (size class 5: n = 13; size class 4: n = 3; size class 3: n = 5; size class 2: n = 10). Analysis of fragmentation patterns for elements belonging to size class 5 (n = 71) identified elements displaying bending breaks (n = 9) indicating fresh bone breakage. Bending breaks cause the cortical peeling of ribs (White 1992; Pickering et al. 2013; Domínguez-Rodrigo et al. 2014b). These modifications have been discussed as a result from human activities (Pickering et al. 2013). As the bending breaks at Gombore II-2 are restricted to the large size classes, it seems apt to suggest damage by lions or spotted hyenas as documented by Haynes and reported in Pickering et al. (2013). Two impact flakes, which could be attributed to long bone fracturing by hominins, were additionally documented, though these flakes can also be produced by hyena scavenging (Domínguez-Rodrigo and Martinez-Navarro 2012).
The results strongly indicate that the overwhelming majority of the bones were fragmented when still fresh and that biotic agents were responsible for most of the bone damage. Fracturing due to diagenetic or burial and/or post-burial processes to a considerable degree is not evident. These results underline the above suggestion that the faunal assemblage shared a relatively homogeneous burial history.
Weathering stages could be determined for 157 faunal remains, not all of which could be attributed to size classes. The preservation of bone fragments of all size classes is mainly characterized by weathering stages 1 and 2 (n = 138). Weathering stages 3 and 4 were recorded for a further eight specimens, representing size classes 2, 4 and 5. The documented weathering stages indicate that bones from differing size classes were buried relatively soon after the death of the animals.
Stages of Abrasion
Stages of abrasion were documented for a limited amount of faunal material (n = 181). Differences in the degree of abrasion were observed for different size classes, suggesting a connection between abrasion and bone fragment size. This is well illustrated by a comparison of bones belonging to size class 5 (n = 68) that mainly show slight to moderate abrasion according to stages 1 and 2 whereas bones from size class 2 (n = 16) primarily display abrasion according to stage 4, i.e. bones affected by heavy abrasion with the original shape of the bone still recognizable. Moreover, a correlation between the degree of bone mineralization and the degree of abrasion was noted, in that bones showing heavy abrasion were often also heavily mineralized. Due to the regular presence of concretions which affected bone surface preservation, this observation could not be further quantified.
Bone Modification due to Biotic Agents
Hominins also impacted the assemblage. Microscopic analysis of two gnawed bone specimens, a left scapula and a right tibia, revealed cut-marks made by stone-tools indicating defleshing and disarticulation of hippo carcasses by hominins. Among these specimens is the heavily gnawed scapula A-35, with cut-marks on its lateral surface. Defleshing of the scapula can result in longitudinal cuts on the medial and lateral face of the bone. These traces were subsequently overlain by a carnivore tooth mark, confirming that carnivores gained access to the carcass only after it was discarded by hominins (Fig. 11). The tibia cut-marks were observed on the medial face of the bone, probably resulting from the removal of the musculus semitendinosus.
Since its publication in 1978 (Chavaillon 1978; Chavaillon et al. 1978), Gombore II-2 has been accepted as a single-carcass butchering site with Acheulean artifacts. When interpreting it as a butchering site, Chavaillon was probably strongly influenced by contemporary research at other Pleistocene African sites where similar evidence was beginning to be understood in this way (SI: Supplementary discussion). This straightforward interpretation suggested by Chavaillon and Berthelet (2004) and again subsequently was made without any zooarchaeological analysis and by merging together layers 3, 4a and 4b, which are of very different origin, into a “single event” (SI: Reassessing earlier excavations).
Isaac and Crader’s (1981): 43) Type-B sites are “sites in which artifacts are found with bones of the carcass of a single large animal.” We assume that a number of parameters are prerequisite for identifying a “single-carcass butchery site”: (1) an explored area large enough to be representative of the archaeology and of hominin activities; (2) a definite paleo-surface in a low-energy context; (1) a distinct covering or englobing layer preserving the paleo-surface (as opposed to Isaac and Crader’s Type-D sites); (4) direct association of large herbivore bones and lithics on the paleo-surface; (5) bones of large herbivores still in articulation; and (6) anthropic modifications on bones, i.e. cut- and percussion-marks and/or the presence of lithic implements with any identified use-wear indicating carcass exploitation and, ideally, organic residues. This last requirement was not fully available in the early 1980s but was at least foreseen by Isaac and Crader (1981).
New excavations and re-analysis of the record allow for a new and detailed evaluation of Gombore II-2. The deposits making up the geo-archaeological sequence are fine-grained. Lithic resources suitable for knapping were not locally available. The lithic implements were made on raw materials from coarse-grained alluvial deposits occurring at some distance. They were both introduced as finished or semi-finished implements and produced on the spot. At Melka Kunture, procurement changed over time to be not only local but also from increasingly distant sources. At the end of the Lower Pleistocene hominins travelled up to 15–20 km for the procurement of high quality rocks used to produce bifaces and cleavers (Gallotti and Mussi 2017). The typo-technological analysis confirms that the lithic industry of layer 4 can be classified as Acheulean. Bifaces, as at Gombore II-2, are usually rare at butchering sites (Clark and Haynes 1970; Isaac 1971, 1977; Leakey 1971; Bunn 1981; Isaac and Crader 1981; Clark 1987; Berthelet and Chavaillon 1996; Berthelet 2001; Delagnes et al. 2006; De la Torre 2011). Both heavy duty tools and debitage products were recovered, presumably used respectively for bone-breakage and meat-cutting. Accordingly, the lithic assemblage is consistent with those found at butchering sites, as known from the abovementioned literature and elsewhere, where butchering and meat-cutting processes were mostly performed with small cutting tools (e.g. Venditti et al. 2019).
The skeletal element representation is mainly characterized by cancellous bones such as scapulae, vertebrae and rib fragments from hippo. The under-representation of compact bones from the autopodium is remarkable, especially in view of the presence of hippo teeth, mainly canines and incisors. The samples documented for the smaller faunal size classes are small, though they equally show a dominance of the cancellous bones of the axial skeleton. Our analysis showed that this selection according to shape does not result from hydrodynamic transport. As the selective process impacted on almost all size classes, we rather suggest that hydraulic processes during site-formation possibly led to this particular sample composition. Accordingly, we are dealing here with selective processes that sorted the bones available in the immediate surroundings of the spot of final burial.
Furthermore, bones in various stages of abrasion most probably formed part of the same hippo carcass (SI Fig. 14). Considering the low energy environment in which the carcass was deposited, also reflected in the overall isotropic distribution pattern of the archaeological material (Fig. 6), it can be excluded that the degree of bone abrasion as well as the degree of mineralization are related to bone transport by non-biotic agents or duration of burial and thus the time that has elapsed between the death of the animal and their final burial. This strongly suggests that diagenetic processes connected to the deposition of the volcanic ash acted in situ upon an autochthonous or para-autochthonous faunal assemblage.
The question is what was the original composition of the recovered faunal sample from which this deposit was formed? At first sight, the heavy dominance by at least three hippos stands out in this very small faunal assemblage. Even though remains of different class sizes are also documented, one might propose that hippo remains provided a substantial and characteristic component in the landscape. The straightforward association of hippo bones and lithics is beyond dispute and is pervasive in the record of Melka Kunture (Chavaillon and Berthelet 2004) as at other African sites (e.g. Leakey 1994). Overall, their frequent association in the archaeological record suggested that hippo paleobiology is correlated with the ecology of early humans (Boisserie and Gilbert 2008). Since adult hippos can weigh up to 3 tons (Eltringham 1999; Kingdon 2015), their behavior and body mass have a great impact on the hydrological network morphology. Hippos affect vegetation composition both by selectively grazing so that the so-called ‘hippo lawns’ develop and by producing dung in such a way that it is widely spread around (McCarthy et al. 1998; Deocampo 2002; Mosepele et al. 2009; Altamura et al. 2017). The changes are overall positive for birds and herbivores (Boisserie and Gilbert 2008) and by implication for hominins.
However, the association of stone tools and hippo carcasses does not necessarily mean in itself that the components are interrelated. At the 1.4 Ma of ‘Ubeidiya, Israel, complete hippo carcasses have been interpreted as reflecting part of the natural background fauna due to the high number of juvenile individuals and the absence of anthropogenic signals on the bone surfaces (Gaudzinski 2004; Gaudzinski-Windheuser 2005). For Olduvai WK Hippo Cliff, dating to around 0.7 Ma, a comparable scenario was outlined by Leakey (1994). Partially preserved carcasses of Hippopotamus gorgops have been unearthed together with crocodile and fish remains. The bones were uncovered together with Acheulean tools in a conglomerate filling a former river or stream channel (Leakey 1994: 36). As the animals are an adult and a juvenile individual, the river might have been a hippo habitat and the association with the stone tools merely accidental (Leakey 1994). Clark et al. similarly report on Middle Pleistocene sites of Ethiopia (Clark et al. 1984). At Hargufia A-2 the record survived in a para-autochtonous context. Remains from two hippos were uncovered from fluvial sands together with Acheulean stone artifacts. Direct traces of human interaction with the carcasses were not observed. Still, the anthropic exploitation of hippo carcasses is well assessed in Africa, starting before 1 Ma. In Bed II of Olduvai, at 1.7 Ma HWK EE site and at 1.5 Ma SHK site, hippo cut-marked bones and hippo bones with possible impact scars have been reported (Diez-Martín et al. 2014; Domínguez-Rodrigo et al. 2014b; Pante et al. 2018). More remains of broadly similar age with scars due to butchering are described at Koobi Fora, Kenya (Bunn 1994; Pobiner et al. 2008; Merritt 2017). Later evidence comes from Buia, Eritrea, where bones of Hippopotamus gorgops dated to around 1 Ma were collected on the surface of fluvio-lacustrine sediments (Fiore et al. 2004). The disarticulation of a carcass is evident from cut-marks on a distal tibia and a calcaneus. In addition, the femur of a Hippopotamus sp. shows traces of defleshing (Fiore et al. 2004). In the Daka Member of Bouri (Middle Awash, Ethiopia), of similar age, hippo remains with butchering scars are mentioned together with Acheulean tools (Asfaw et al. 2002).
At Gombore II-2, the faunal collection is more complex than stated in publications by Chavaillon and collaborators. This is confirmed by the footprints on the surface of layer 4b, which show that this particular spot was visited by small- and medium-sized ungulates, whereas hippo presence was rare (Altamura et al. 2018). The remains of at least three hippos did not survive at the site in anatomical articulation. There is some bone concentration in the north-central part of the area but no certainty of any single carcass. The scattering of an exposed carcass is a frequently occurring natural process, since a dead animal can be disjointed and dislocated in few hours or days by carnivores and necrophages (Haynes 2005, 2015). However, in the case of Gombore II-2, hominin interaction with hippos is also proven by the observed cut-marks. Hominins could have taken advantage of a dead animal located exactly at that spot or could have transported selected portions of a carcass there from scavenging- or kill-sites at a distance (see examples in Agam and Barkai 2018). The reconstruction of the timing of events on the hippo scapula A-35 indicates that carnivores subsequently scavenged defleshed and disarticulated carcasses left by hominins.
Summing up past and recent excavations, an overall area of ca. 70 m2 was investigated at Gombore II-2, large enough to reconstruct the local environment at a micro-scale. The stratigraphic sequence was re-assessed in detail. During a short period of time, before a volcanic event buried the area, a paleo-surface was trampled by animals and hominins walking close to a body of still water. From an archeological perspective, the lithic implements and bone remains littering the surface are closely associated. This satisfies headings 1 to 4 of the expanded version of Isaac and Crader’s (1981) definition mentioned above: an explored area large enough to be representative of the archaeology and of hominin activities; a definite paleo-surface in a low-energy environment; a distinct covering or englobing layer preserving the paleo-surface; direct association of large herbivore bones and lithics on the paleo-surface. No anatomical connections were preserved, but hippos were certainly butchered on the evidence of the zooarcheological analysis. Altogether, Gombore II-2 does not fit into Isaac and Crader’s (1981) “type B site” category. After our own reassessment (SI: Reassessing earlier excavations), the “butchering site” is not a single event, but the outcome of a brief series of episodes over a short length of time.
The chronology of Gombore II-2 is firmly grounded, post-dating 0.78 Ma (Tamrat et al. 2013) and close to 0.7 Ma (Morgan et al. 2012). The site formed during a favorable climatic episode of the Middle Pleistocene Transition, such as MIS 19 or within an oscillation of MIS 18 (Mussi et al. 2016). The evidence from new fieldwork, the archives and the altitudinal profiles allow us to define rather finely the stratigraphic position of a substantial part of the material produced by previous research. This, in turn, allows a re-evaluation of the butchering activity that we refer to layer 4 only. The recent excavations produced lithics and fossils further substantiating this interpretation.
In 1974, a preliminary determination of faunal MNIs was provided based only on the finds from a small trench. It was never updated, even when the explored area was significantly extended and many more bones were unearthed (e.g. Berthelet and Chavaillon 1996; Berthelet et al. 2001; Chavaillon and Berthelet 2004). The interpretation as a butchering site was consistent with the state of contemporary research but obviously lacked the traceological and taphonomic analyses, which became standard in the following decades. After our zooarchaeological and geoarchaeological analysis, the cut-marked bones, their close association to lithic tools in a low-energy environment and the ichnological traces, all prove that there was indeed proximity and interaction between hominins and hippos. The ichnological surface at the top of layer 4b is positive evidence for a living floor that, given the preservation of footprints, formed over a short time, such as a single season, before being sealed by the deposits of volcanic origin of layer 3 (Altamura et al. 2017, 2018). This allows a different interpretation of the site: layer 4 as a whole registered episodes of human occupation, including butchery events, over a rather short time span, but it was not a “single-carcass butchering site.” Stone implements and scattered remains from various species are documented on the spot.
The case-study of Gombore II-2 challenges the validity of surviving research paradigms established decades ago. The “single-carcass” sites, which apparently conform to a straightforward interpretation, instead deserve careful re-investigation in light of advances in the archaeological sciences. Using in a convincing way the expanded version of Isaac and Crader’s classification, probably few supposed single carcass sites of the African record could today be fully validated. A few seem promising, like FLK N6 of Olduvai (1.82 Ma) (Leakey 1971; Domínguez-Rodrigo et al. 2012) and Barogali in Djibouti (1.6-1.3 Ma) (Chavaillon et al. 1987; Berthelet et al. 1992; Berthelet and Chavaillon 1996; Berthelet 2001). At both, there were extensive excavations, with well-defined single-episode archaeology. Elephant skeletons in anatomical connection are definitely surrounded by lithic implements. At FLK N6 cut-marks have also been detected (Bunn 1986; Bunn and Pickering 2010). However, even this evidence would need tight scrutiny and cannot be accepted any more at face value.
We underline that hippos are known to be a serious threat and even a deadly one for modern humans (Kanga et al. 2012). Furthermore, they are territorial, living along watercourses in often large groups (Eltringham 1999). Exploiting a hippo carcass at the place of death of the animal probably meant confrontation not only with carnivores but also with the quite aggressive and dangerous members of the group of hippos. The re-investigation of Gombore II-2 suggests that, by the time of the Middle Pleistocene transition, a rather sophisticated hominin behavior for exploiting hippo carcasses had already developed, allowing them to locate them, probably at some distance, then butchering them repeatedly, taking advantage of the available meat before competing carnivores. This is in line with the complex pattern of fallow deer exploitation at the nearly contemporaneous site of Gesher Benot Ya‘aqov in the Levant (Rabinovich et al. 2008). The hominin tracks show that, at Gombore II-2, a mixed age group was busy on the spot where butchering activities were also performed. Women were definitely there, while children potentially as young as 12 months old were wandering around. Careful monitoring was not limited to food resources at Melka Kunture. It is further evidenced by raw material provisioning at increasing distances for specific types of lithic production (Gallotti and Mussi 2017). At this time, evaluating resources and foreseeing future needs were capacities fully engrained in human behavior. Circa 850,000 years ago, a probable direct candidate for the origin of Homo heidelbergensisis is recorded in the area (Chavaillon et al. 1974; Chavaillon and Coppens 1986; Profico et al. 2016). Therefore, it seems probable that, at Gombore II-2, 100,000 to 200,000 years later, the evidence reported here documents the behavior and capacities of Homo heidelbergensis.
The Ethiopian Authority for Research and Conservation of Cultural Heritage and the Oromia Region authorized fieldwork and laboratory and facilitated the research in many ways. We are grateful to Denis Geraads who shared with us the paleontological determination of the fossil fauna, to Laura Pioli who commented on volcanic products and to Eduardo Mendez-Quintaz for his comments on our spatial analysis, but any misinterpretation is obviously ours. The reassessment of the “butchery site” was the focus of the PhD research and part of the post-doc project ‘Il contributo dell’icnologia ai siti geo-archeologici del Pleistocene: casi di studio in Italia ed Etiopia’ (AR 008/2018) of the first author (FA) at the Sapienza University of Rome. We wish to thank the colleagues involved in the field research: Elisa Brunelli, Eliana Catelli, Alessandro Cecili, Luca Di Bianco, Flavia Piarulli, Lorena Russo, Noemi Tomei. The archive documents are due to the generous contribution of Jean Chavaillon’s relatives. We are also grateful to Martin Street, who kindly revised our English manuscript. Last but not least, we thank the two unknown reviewers and the editor Shannon McPherron for their constructive comments and suggestions.
The Italian Archaeological Mission at Melka Kunture and Balchit is funded by grants of Grandi Scavi di Ateneo of Sapienza University and of Ministero degli Affari Esteri e della Cooperazione Internazionale, awarded to the Director M.M. Johannes Gutenberg University Mainz and the Römisch-Germanisches Zentralmuseum, Leibniz-Research Institute for Archaeology (Germany) supported the zooarchaeological studies.
Compliance with Ethical Standards
Conflict of Interest
The authors declare that they have no conflict of interest.
- Altamura, F. (2017). Il sito acheuleano di Gombore II-2: revisione dei dati e nuove ricerche. University of Roma Sapienza, Ph.D. Dissertation.Google Scholar
- Altamura, F., & Mussi, M. (2017). Archeologia e impronte fossili nel sito acheuleano di Gombore II (0,85 Ma), Melka Kunture, Etiopia. Scienze dell’Antichità, 23(1), 21–35.Google Scholar
- Altamura, F., Bennett, M. R., D’Août, K., Gaudzinski-Windheuser, S., Melis, R. T., Reynolds, S. C., & Mussi, M. (2018). Archaeology and ichnology at Gombore II-2, Melka Kunture, Ethiopia: Everyday life of a mixed-age hominin group 700,000 years ago. Scientific Reports, 8, 2815.CrossRefGoogle Scholar
- Attwell, R. I. G. (1963). Surveying Luangwa hippo. Puku, 1, 29–49.Google Scholar
- Bailloud, G. (1965). Le gisement paléolithique de Melka-Kontouré. Addis Abeba: Institut Èthiopien d’Archéologie, Cahier n°1.Google Scholar
- Behrensmeyer, A. K. (1975). The taphonomy and paleoecology of Plio-Pleistocene vertebrate assemblages east of Lake Rodolf, Kenya. Bulletin of the Museum of Comparative Zoology at Harvard College, 146, 473–578.Google Scholar
- Berthelet, A. (2001). L’outillage lithique du site de dépeçage à Elephas recki ileretensis de Barogali (République de Djibouti). Comptes Rendus de l’Académie des Sciences Paris, 332, 411–416.Google Scholar
- Berthelet, A., & Chavaillon, J. (1996). Sites de boucherie à elephants et hippopotamus au Paléolithique inférieur en Afrique orientale. Les dossiers de l’Archéologie, Revivre le passé grâce à l’Archéologie, 216, 6–11.Google Scholar
- Berthelet, A., Bulgarelli, G. M., Chavaillon, J., & Piperno, M. (2001). Melka Kunture. La Guida-The Guide. Lavello: Finiguerra Arti grafiche.Google Scholar
- Binford, L. R. (1981). Bones: ancient men and modern myths. New York: Academic Press.Google Scholar
- Blumenschine, R. J., & Selvaggio, M. M. (1991). On the marks of marrow bone processing by hammerstones and hyenas: their anatomical patterning and archaeological implications. In J. Desmond Clark (Ed.), Cultural beginnings: Approaches to understanding early hominid life-ways in the African savanna (pp. 17–32). Bonn: Rudolf Habelt.Google Scholar
- Boisserie, J.-R., & Gilbert, W. H. (2008). Hippopotamidae. In W. H. Gilbert & B. Asfaw (Eds.), Homo erectus: Pleistocene evidence from the Middle Awash, Ethiopia (pp. 179–191). Berkeley: UC Press.Google Scholar
- Braun, D. R., Harris, J. W. K., Levin, N. E., McCoy, J. T., Herries, A. I. R., Bamford, M. K., Bishop, L. C., Richmond, B. G., & Kibunjia, M. (2010). Early hominin diet included diverse terrestrial and aquatic animals 1.95 Ma in East Turkana, Kenya. Proceedings of the National Academy of Sciences., 107(22), 10002–10007.CrossRefGoogle Scholar
- Bunn, H. T. (1997). The bone assemblages from the excavated sites. In G. Isaac & B. Isaac (Eds.), Koobi Fora Research Project, Plio-Pleistocene Archaeology, vol. 5 (pp. 402–458). Oxford: Clarendon.Google Scholar
- Chavaillon, J. (1975a). Rapport sur l’activite de la Mission franco-ethiopienne de Melka Kunture. Campagne 1974. ARCCH archives, Addis Ababa (Ethiopia), unpublished.Google Scholar
- Chavaillon, J. (1975b). Letter of J. Chavaillon to the Director of the Archaeological Institute of the Ministry of Culture of Addis Abeba, dated February 4, 1975, with excavation inventory in attachment. ARCCH archives, Addis Ababa (Ethiopia), unpublished.Google Scholar
- Chavaillon, J. (1976). Mission archeologique franco-ethiopienne del Melka-Kunture. Rapport préliminaire 1972-1975. L’Ethiopie avant l’histoire, cahier 1 (pp. 1-11). CNRS.Google Scholar
- Chavaillon, J. (1980). Chronologie archéologique de Melka-Kunturé (Ethiopie). In: R.E. Leakey, B.A. Ogot (Eds.), Proceedings VIII Panafrican Congress of Prehistory and Quaternary Studies, Nairobi (pp. 200-201).Google Scholar
- Chavaillon, J. (1981-1982). Requests for authorization for archaeological research, Work programs 1981-1982 and 1982-1986. ARCCH archives, Addis Ababa (Ethiopia), unpublished.Google Scholar
- Chavaillon, J., & Berthelet, A. (2001). The Elephas recki site of Haïdalo (Republic of Djibouti). In G. Cavarretta, P. Gioia, M. Mussi, & M. R. Palombo (Eds.), The World of Elephants. Proceedings of the 1st International Congress, Rome, October 16-20 2001 (pp. 191–193). Consiglio Nazionale delle Ricerche: Rome.Google Scholar
- Chavaillon, J., & Berthelet, A. (2004). The archaeological sites of Melka Kunture. In J. Chavaillon & M. Piperno (Eds.), Studies on the Early Paleolithic Site of Melka Kunture, Ethiopia (pp. 25–80). Florence: Origines, Istituto Italiano di Preistoria e Protostoria.Google Scholar
- Chavaillon, J., & Chavaillon, N. (1980). Evolution de l'Acheuléen à Melka-Kunturé (Ethiopie). Anthropologie, XVIII(2/3), 153–159.Google Scholar
- Chavaillon, J., & Chavaillon, N. (1982). Les habitats paléolithiques de Melka-Kunturé (Ethiopie). Evolution des techniques et des structures. Abbay, XI, 23–44.Google Scholar
- Chavaillon, J., & Coppens, Y. (1986). Nouvelle découverte d'Homo erectus à Melka-Kunturé. Comptes Rendus de l’Académie des Sciences Paris, 303(1), 99–104.Google Scholar
- Chavaillon, J., & Piperno, M. (2004). Studies on the Early Paleolithic site of Melka Kunture, Ethiopia. Florence: Origines, Istituto Italiano di Preistoria e Protostoria.Google Scholar
- Chavaillon, J., Brahimi, C., & Coppens, Y. (1974). Première découverte d'Hominidé dans l'un des sites acheuléens de Melka-Kunturé (Ethiopie). Comptes Rendus de l’Académie des Sciences Paris, 278, 3299–3302.Google Scholar
- Chavaillon, J., Chavaillon, N., Hours, F., & Piperno, M. (1979). From the Oldowan to the Middle Stone Age at Melka-Kunture (Ethiopia). Understanding cultural changes. Quaternaria, XXI, 87–114.Google Scholar
- Chavaillon, J., Boisaubert, J.-L., Faure, M., Guérin, C., Ma, J. L., Nickel, B., Poupeau, G., Rey, P., & Warsama, S. A. (1987). Le site de dépeçage pléistocène à Elephas recki de Barogali (République de Djibouti): nouveaux résultats et datation. Comptes Rendus de l’Académie des Sciences Paris, 305, 1259–1266.Google Scholar
- Delagnes, A., Lenoble, A., Harmand, S., Brugal, J.-P., Prat, S., Tiercelin, J.-J., & Roche, H. (2006). Interpreting pachyderm single carcass sites in the African Lower and Early Middle Pleistocene record: A multidisciplinary approach to the site of Nadung’a 4 (Kenya). Journal of Anthropological Archaeology, 25, 448–465.CrossRefGoogle Scholar
- Diez-Martín, F., Sánchez-Yustos, P., Uribelarrea, D., Domínguez-Rodrigo, M., Fraile-Márquez, C., Obregón, R. A., Dìaz-Muñoz, I., Mabulla, A., Baquedano, E., Pérez-González, A., & Bunn, H. T. (2014). New archaeological and geological research at SHK main site (Bed II, Olduvai Gorge, Tanzania). Quaternary international, 322, 107–128.CrossRefGoogle Scholar
- Domínguez-Rodrigo, M., Mabulla, A. Z. P., Bunn, H. T., Diez-Martin, F., Baquedano, E., Barboni, D., Barba, R., Domínguez-Solera, S., Sanchez, P., Ashley, G. M., & Yravedra, J. (2010). Disentangling hominin and carnivore activities near a spring at FLK North (Olduvai Gorge, Tanzania). Quaternary Research, 74, 363–375.CrossRefGoogle Scholar
- Domínguez-Rodrigo, M., Bunn, H. T., Pickering, T. R., Mabulla, A. Z. P., Musiba, C. M., Baquedano, E., Ashley, G. M., Diez-Martin, F., Santonja, M., Uribelarrea, D., Barba, R., Yravedra, J., Barbonim, D., Arriaza, C., & Gidna, A. (2012). Autochthony and orientation patterns in Olduvai Bed I: A re-examination of the status of post-depositional biasing of archaeological assemblages from FLK North (FLKN). Journal of Archaeological Science, 39, 2116–2127.CrossRefGoogle Scholar
- Domínguez-Rodrigo, M., Bunn, H. T., Mabulla, A. Z. P., Baquedano, E., Uribelarrea, D., Pérez-González, A., Gidna, A., Yravedra, J., Diez-Martin, F., Egeland, C. P., Barba, R., Arriaza, M. C., Organista, E., & Ansón, M. (2014a). On meat eating and human evolution: A taphonomic analysis of BK4b (Upper Bed II, Olduvai Gorge, Tanzania), and its bearing on hominin megafaunal consumption. Quaternary International, 322, 129–152.CrossRefGoogle Scholar
- Domínguez-Rodrigo, M., Diez-Martín, F., Barba, R., Yravedra, J., Bunn, H. T., Mabulla, A., Baquedano, E., & Uribelarrea, D. (2014b). A taphonomic study of the faunal assemblage of the main site at SHK (Bed II, Olduvai Gorge, Tanzania). Quaternary International, 322-323, 153–166.CrossRefGoogle Scholar
- Egels, Y. (1971). Sites archéologiques de Melka-Konture: Gombore et Garba. Echelle:1/2000. Paris: Institut Géographique National.Google Scholar
- Eltringham, S. K. (1999). The Hippos. London: Academic Press.Google Scholar
- Espigares, M. P., Martínez-Navarro, B., Palmqvist, P., Ros-Montoya, S., Toro, I., Agustí, J., & Sala, R. (2013). Homo vs. Pachycrocuta: Earliest evidence of competition for an elephant carcass between scavengers at Fuente Nueva-3 (Orce, Spain). Quaternary International, 295, 113–125.CrossRefGoogle Scholar
- Fernandez-Jalvo, Y., Andrews, P. (2016). Atlas of taphonomic identifications: 1001+ images of fossil and recent mammal bone modification. Springer.Google Scholar
- Fiore, I., & Tagliacozzo, A. (2004). Taphonomic analysis of the bone remains from the Oldowan site of Garba IV. In J. Chavaillon & M. Piperno (Eds.), Studies on the Early Paleolithic Site of Melka Kunture, Ethiopia (pp. 639–682). Florence: Origines, Istituto Italiano di Preistoria e Protostoria.Google Scholar
- Fiore, I., Bondioli, L., Coppa, A., Macchiarelli, R., Russom, R., Kashay, H., Solomon, T., Rook, L., & Libsekal, Y. (2004). Taphonomic analysis of the late Early Pleistocene bone remains from Buia (Dandiero Basin, Danakil Depression, Eritrea): Evidence for large mammal and reptile butchering. Rivista Italiana di Paleontologia e Stratigrafia, 110, 89–97.Google Scholar
- Gallotti, R., & Mussi, M. (2017). Two Acheuleans, two humankinds: From 1.5 to 0.85 Ma at Melka Kunture (Upper Awash, Ethiopian highlands). Journal of Anthropological Sciences, 95, 1–46.Google Scholar
- Gaudzinski-Windheuser, S. (2005). Subsistenzstrategien frühpleistozäner Hominiden in Eurasien: taphonomische Faunenbetrachtungen der Fundstellen der ʿUbeidiya Formation (Israel). Habelt.Google Scholar
- Gaudzinski-Windheuser, S., & Kindler, L. (2012). Reconstructing site formation processes at GBY—The experiments. In R. Rabinovich, S. Gaudzinski-Windheuser, L. Kindler, & N. Goren-Inbar (Eds.), The Acheulian Site of Gesher Benot Ya ‘aqov Volume III (pp. 93–222). Dordrecht: Springer.Google Scholar
- Gaudzinski-Windheuser, S., Kindler, L., Rabinovich, R., & Goren-Inbar, N. (2010). Testing heterogeneity in faunal assemblages from archaeological sites. Tumbling and trampling experiments at the early-Middle Pleistocene site of Gesher Benot Ya’aqov (Israel). Journal of Archaeological Science, 37(12), 3170–3190.CrossRefGoogle Scholar
- Geraads, D., Eisenmann, V., & Petter, G. (2004). Paleontology. The large mammal fauna of the Oldowan sites of Melka Kunture. In J. Chavaillon & M. Piperno (Eds.), Studies on the Early Paleolithic Site of Melka Kunture, Ethiopia (pp. 169–192). Florence: Origines, Istituto Italiano di Preistoria e Protostoria.Google Scholar
- Haynes, G. (2005). Modern elephant-bone assemblages as models for interpreting Ambrona and other fossil waterside deathsites. In M. Santonja Gomez & A. Pérez-González (Eds.), Los Yacimientos Paleoliticos de Ambrona y Torralba, Zona Arqueológica, vol. 5 (pp. 154–174). Madrid: Museo Arqueológico Regional.Google Scholar
- Inizan, M.-L., Reduron-Ballinger, M., Roche, H., & Tizier, J. (1999). Technology and Terminology of Knapped Stone. Nanterre: CREP.Google Scholar
- Isaac, G. L. (1977). Olorgesailie: Archaeological studies of a Middle Pleistocene lake basin in Kenya. Chicago and London: The University of Chicago Press.Google Scholar
- Isaac, G. L., & Crader, D. C. (1981). To what extent were early hominids carnivorous? An archaeological perspective. In R. S. O. Hardinger & G. Teleki (Eds.), Omnivorous Primates (pp. 37–103). New York: Columbia University Press.Google Scholar
- Isaac, G. L., & Isaac, B. (Eds.). (1989). The archaeology of human origins. Cambridge: Cambridge University Press.Google Scholar
- Kieffer, G., Raynal, J.-P., Bardin, G. (2002). Cadre structural et volcanologiques des sites du Paléolithique ancien de Melka Kunture (Awash, Ethiopie): premiers résultats. In: J.-P. Raynal, C. Albore-Livadie, M. Piperno (Eds.), Hommes Et Volcans. De l’éruption à l’objet (pp. 77-92). Les Dossiers de l’Archéologie 2, Proceedings Symposium 15.2 , XIV UISPP Congress, Liège, 2001.Google Scholar
- Kingdon, J. (2015). The Kingdon field guide to African mammals. Bloomsbury Publishing.Google Scholar
- Leakey, M. D. (1971). Olduvai Gorge: Excavations in Beds I and II, 1960–1963, V.3. Cambridge: Cambridge University Press.Google Scholar
- Leakey, L. (1994). Olduvai Gorge. Excavations in the Bed II, IV and Masek Beds. Cambridge: Cambridge University Press.Google Scholar
- Leakey, M. D., & Roe, D. A. (1994). Olduvai Gorge: Excavations in Beds III, IV and the Masek Beds, 1968-1971. Cambridge: Cambridge University Press.Google Scholar
- Mussi, M., Anzidei, A. P., Boschian, G., Martini, F., Mazza, P., & Piperno, M. (2005). Hombres y Elefantes en las latitudes medias: una larga convivencia. In M. Santonja Gomez & A. Pérez-González (Eds.), Los Yacimientos Paleoliticos de Ambrona y Torralba, Zona Arqueológica, vol. 5 (pp. 396–417). Madrid: Museo Arqueológico Regional.Google Scholar
- Mussi, M., Altamura, F., Bonnefille, R., De Rita, D., & Melis, R. T. (2016). The environment of the Ethiopian highlands at the Mid Pleistocene Transition: Fauna, flora and hominins in the 850-700ka sequence of Gombore II (Melka Kunture). Quaternary Science Review, 149, 259–268.CrossRefGoogle Scholar
- Pickering, T. R., Domínguez-Rodrigo, M., Heaton, J. L., Yravedra, J., Barba, R., Bunn, H. T., Musiba, C., Baquedano, E., Diez-Martín, F., Mabulla, A., & Brain, C. K. (2013). Taphonomy of ungulate ribs and the consumption of meat and bone by 1.2-million-year-old hominins at Olduvai Gorge, Tanzania. Journal of Archaeological Science, 40(2), 1295–1309.CrossRefGoogle Scholar
- Piperno, M. (2001). La préhistoire de Melka Kunture (Éthiopie). Bulletin du Centre de recherche français à Jérusalem, 8, 25–35.Google Scholar
- Piperno, M., Gallotti, R. (2003). Recent activities of the Italian Archaeological Mission at Melka Kunture (Ethiopia): The Open Air Museum Project and the GIS application to the study of the Oldowan sites. In: J. Martínez Moreno, R. Mora Torcal, I. de la Torre Sainz (Eds.), Oldowan: Rather more than smashing stones (pp. 37-75). First Hominid Technology Workshop. Bellaterra, December 2001. Bellaterra: Treballs d’Arqueologia, 9.Google Scholar
- Piperno, M., Collina, C., Gallotti, R., Raynal, J.-P., Kieffer, G., Le Bourdonnec, F.-X., Poupeau, G., Geraads, D. (2008). Obsidian exploitation and utilization during the Oldowan at Melka Kunture, (Ethiopia). In: E. Hovers, D.R. Braun, E. Delson, R. MacPhee (Eds.), Interdisciplinary Approaches to the Oldowan (pp. 111-128). Springer.Google Scholar
- Profico, A., Di Vincenzo, F., Gagliardi, L., Piperno, M., & Manzi, G. (2016). Filling the gap. Human cranial remains from Gombore II (Melka Kunture, Ethiopia; ca. 850 ka) and the origin of Homo heidelbergensis. Journal of Anthropological Sciences, 94, 24.Google Scholar
- Raynal, J.-P., & Kieffer, G. (2004). Lithology, dynamism and volcanic successions at Melka Kunture (Upper Awash, Ethiopia). In J. Chavaillon & M. Piperno (Eds.), Studies on the Early Paleolithic Site of Melka Kunture, Ethiopia (pp. 111–135). Florence: Origines, Istituto Italiano di Preistoria e Protostoria.Google Scholar
- Raynal, J.-P., Kieffer, G., & Bardin, G. (2004). Garba IV and the Melka Kunture Formation. A preliminary lithostratigraphic approach. In J. Chavaillon & M. Piperno (Eds.), Studies on the Early Paleolithic Site of Melka Kunture, Ethiopia (pp. 137–166). Florence: Origines, Istituto Italiano di Preistoria e Protostoria.Google Scholar
- Sanhouni, M., Rosell, J., van der Made, J., Vergés, J., Ollé, A., Kandi, N., Harichane, Z., Derradji, A., & Medig, M. (2013). The first evidence of cut marks and usewear traces from the Plio-Pleistocene locality of El-Kherba (Ain Hanech), Algeria: Implications for early hominin subsistence activities circa 1.8 Ma. Journal of Human Evolution,64(2), 137–150.CrossRefGoogle Scholar
- Santonja, M., & Péres-Gonzàles, A. (2005). Arquelogia y elefantes en el Pleistocene Medio de la Peninsula Ibérica. In M. Santonja Gomez & A. Pérez-González (Eds.), Los Yacimientos Paleoliticos de Ambrona y Torralba, Zona Arqueológica, vol. 5 (pp. 382–395). Madrid: Museo Arqueológico Regional.Google Scholar
- Santonja, M., Martínez, N.L., González, A.P. (1980). Ocupaciones Achelenses en el valle del Jarama (Arganda, Madrid) (Vol. 1). Madrid: Diputación Provincial, Servicios de Extensión Cultural y Divulgación.Google Scholar
- Taieb, M. (1974). Evolution quaternaire du bassin de l'Awash. Ph.D. Dissertation, Paris: Université Pierre et Marie Curie.Google Scholar
- Voorhies, M. R. (1969). Taphonomy and population dynamics of an early Pliocene vertebrate fauna, Knox County, Nebraska. Vol. 1 (pp. 1–69). Laramie: University of Wyoming.Google Scholar
- Wall, W. P. (1983). The correlation between high limb-bone density and aquatic habits in recent mammals. Journal of Paleontology, 57(2), 197–207.Google Scholar
- Weber, T. (2000). The Eemian Elephas antiquus finds with artefacts from Lehringen and Gröbern. E.R.A.U.L., 51, 177-185.Google Scholar