pante-michael-c-the-carnivorous-feeding-behavior-of

Journal of Human Evolution xxx (2017) 1e21
Contents lists available at ScienceDirect
Journal of Human Evolution

journal homepage: www.elsevier.com/locate/jhevol
The carnivorous feeding behavior of early Homo at HWK EE, Bed II,
Olduvai Gorge, Tanzania
Michael C. Pante a, *, Jackson K. Njau b, Blaire Hensley-Marschand b, c, Trevor L. Keevil a,
Carmen Martín-Ramos d, Renata Franco Peters d, Ignacio de la Torre d
a
Department of Anthropology, Colorado State University, 1787 Campus Delivery, Fort Collins, CO, 80523, USA
b
Department of Earth and Atmospheric Sciences, Indiana University, 1001 E Tenth Street, Bloomington, IN, 47405, USA
c
Department of Anthropology, Indiana University, 701 E. Kirkwood Avenue, Bloomington, IN, 47405, USA
d
Institute of Archaeology, University College London, 31-34 Gordon Square, WC1H 0PY, London, UK
a r t i c l e i n f o a b s t r a c t
Article history: The regular consumption of large mammal carcasses, as evidenced by butchery marks on fossils
Received 28 December 2016 recovered from Early Stone Age archaeological sites, roughly coincides with the appearance of Homo
Accepted 20 June 2017 habilis. However, the significance of this niche expansion cannot be appreciated without an under-
Available online xxx
standing of hominin feeding behavior and their ecological interactions with mammalian carnivores. The
Olduvai Geochronology and Archaeology Project (OGAP) has recovered a large and well-preserved fossil
Keywords:
assemblage from the HWK EE site, which was deposited just prior to the first appearance of Acheulean
Homo habilis
technology at Olduvai Gorge and likely represents one of the last H. habilis sites at Olduvai. This taph-
Megafauna
Bone surface modifications
onomic analysis of the larger mammal fossil assemblage excavated from HWK EE shows evidence of
Taphonomy multiple occupations over a long period of time, suggesting the site offered resources that were attractive
Scavenging to hominins. There was a water source indicated by the presence of fish, crocodiles, and hippos, and there
Oldowan was possible tree cover in an otherwise open habitat. The site preserves several stratigraphic intervals
with large fossil and artifact assemblages within two of these intervals. Feeding traces on bone surfaces
suggest hominins at the site obtained substantial amounts of flesh and marrow, particularly from smaller
size group 1e2 carcasses, and exploited a wide range of taxa, including megafauna. A strong carnivore
signal suggests hominins scavenged much of their animal foods during the two main stratigraphic in-
tervals. In the later interval, lower carnivore tooth mark and hammerstone percussion mark frequencies,
in addition to high epiphyseal to shaft fragment ratios, suggest hominins and carnivores did not fully
exploit bone marrow and grease, which may have been acquired from nutritionally-stressed animals that
died during a dry period at Olduvai. The diversity of fauna that preserve evidence of butchery suggests
that the HWK EE hominins were opportunistic in their acquisition of carcass foods.
© 2017 Elsevier Ltd. All rights reserved.
1. Introduction morphological changes observed in the genus Homo (Milton, 1987;

Shipman and Walker, 1989; Speth, 1989; Ruff and Walker, 1993;
The initial encroachment of our ancestors upon the larger Aiello and Wheeler, 1995; Foley, 2001, 2002; Anto n et al., 2002;
carnivore guild has important implications for the feeding ecology Anton, 2003; Anto n and Swisher, 2004). However, whether this
and adaptive capabilities of our ancestors (Blumenschine and niche expansion began in the form of hunting or scavenging is hotly
Pobiner, 2007). The regular consumption of large mammal car- contested, particularly for the FLK 22 (Zinjanthropus) level assem-
casses, as evidenced by butchery marks on fossils recovered from blage from Olduvai Gorge (Binford, 1981, 1986, 1988; Bunn and
Early Stone Age archaeological sites, roughly coincides with the Kroll, 1986, 1988; Blumenschine, 1988, 1995; Bunn and Ezzo,
appearance of Homo habilis and has traditionally been linked to the 1993; Oliver, 1994; Selvaggio, 1994, 1998; Capaldo, 1995, 1997,
1998; Domínguez-Rodrigo, 1997; Dominguez-Rodrigo and Barba,
2006; Blumenschine et al., 2006; Pobiner et al., 2008; Pante et al.,
2012, 2015; Ferraro et al., 2013; Dominguez-Rodrigo et al., 2014a;
* Corresponding author.
Parkinson et al., 2015). Despite the apparent lack of consensus
E-mail address: Michael.Pante@colostate.edu (M.C. Pante).
http://dx.doi.org/10.1016/j.jhevol.2017.06.005
0047-2484/© 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Pante, M.C., et al., The carnivorous feeding behavior of early Homo at HWK EE, Bed II, Olduvai Gorge, Tanzania,
Journal of Human Evolution (2017), http://dx.doi.org/10.1016/j.jhevol.2017.06.005
2 M.C. Pante et al. / Journal of Human Evolution xxx (2017) 1e21
concerning the mode of carcass acquisition regularly practiced by carcasses were fed upon by vultures, and in some cases minimally
H. habilis, researchers are largely in agreement that flesh and defleshed by carnivores prior to disarticulation and marrow
marrow were important resources to the species. extraction by humans, followed by grease removal by carnivores.
Contrary to those of H. habilis, studies of assemblages associated One additional scavenging model, the carnivore-to-hominin-to-
with Homo erectus all suggest that the species regularly obtained carnivore model (CeHeC) developed by Selvaggio (1994, 1998), is
early access to carcasses, possibly through hunting (Monahan, considered but not directly compared with the archaeological as-
1996; Pickering et al., 2004; Pobiner et al., 2008; Dominguez- semblages. This model was not included with those statistically
Rodrigo et al., 2009a, 2014b; 2014c; Pante, 2013). However, the reanalyzed by Pante et al. (2012), but it represents a simulation of
small number of assemblages on which these interpretations are passive scavenging from carnivores and is relevant to the in-
based limits our ability to infer the precise timing and nature of terpretations of the archaeological assemblages.
what may signal a pivotal progression in the predatory behavior of When considered together on the basis of all three mark types
hominins. This limitation has historically been the result of a lack of (tooth, cut and percussion), these models become powerful tools in
well-preserved fossil assemblages on which the feeding traces of interpretations of hominin and carnivore feeding ecology (Pante
hominins and carnivores can be observed. et al., 2012). While alternative methods of interpretation based
Here we report on a new and well-preserved fossil assemblage on frequencies of feeding traces in fossil assemblages have recently
that dates to around 1.7 Ma, which was recovered from the HWK EE been proposed (Dominguez-Rodrigo et al., 2014a), their effective-
site, Olduvai Gorge, Tanzania (McHenry, submitted). The site ness has been questioned based on multivariate analyses that
comprises specimens from four separate archaeological trenches emphasize cut mark frequencies over all other feeding traces (see
(Trench 1-Main Trench, T27, T28 and T29) and three stratigraphic Pante et al., 2015). We maintain the models used here are more
intervals (Lemuta, Lower Augitic Sandstone [LAS] and Tuff IIB), two effective because they give equal attention to tooth, cut and per-
of which (Lemuta and LAS) are associated with Oldowan technol- cussion marks, and they also consider specimens that preserve both
ogy (Fig. 1; see also de la Torre et al., submitted-a). Given the tooth and butchery marks (Pante et al., 2015).
technology and stratigraphic position of the site, the archaeological
material from HWK EE is more likely attributed to H. habilis than 2.2. The HWK EE assemblage
the presumed maker of Acheulean technology, H. erectus, and may
capture the behavior of the species near the end of its existence at The HWK EE fossil assemblage was analyzed throughout mul-
Olduvai. However, neither species, nor even Paranthropus boisei, tiple field seasons at Olduvai Gorge between 2009 and 2016. All
can be ruled out as the maker and user of stone tools at the site. fossils were catalogued, and both taphonomic and taxonomic data
The goal for this study is to provide interpretations of the were collected from the larger mammal assemblage. The teeth have
carnivorous feeding behavior of the HWK EE hominins along with been subjected to additional isotopic (Rivals et al., this volume; Uno
the ecological interactions they had with carnivores. A more gen- et al., submitted) and meso/microwear (Rivals et al., submitted)
eral description of the assemblage taphonomy and paleoecology analyses, while both fossils and teeth have undergone detailed
can be found elsewhere (Bibi et al., submitted; de la Torre et al., taxonomic identification (Bibi et al., submitted) beyond what is
submitted-a). It is hypothesized that when assessed with the presented here. A taxonomic analysis of the collection of bird fossils
same methods, the HWK EE hominins will be found to have recovered from the site has also been conducted (Prassack et al.,
exhibited feeding strategies similar to those of the FLK Zinjan- submitted).
thropus hominins, who likely regularly obtained access to carcasses Great care was taken during and after excavation to preserve the
that had been partially defleshed by carnivores (Pante et al., 2012, HWK EE fossil assemblage. While some fossils were found in almost
2015). Here we test this hypothesis by examining the frequency pristine condition, others were brittle, misshapen, fractured, and/or
and location of hominin butchery marks and carnivore tooth marks. crushed. Large fossils or those in poor condition often needed to be
stabilized by temporary consolidation before block lifting was
2. Methods carried out to remove them from the site. Cyclododecane (C12H24, a
cyclic alkane hydrocarbon; CDD) was used on the large and un-
2.1. Experimental controls stable fossils because of its sublimation properties, which elimi-
nated the need for solvent removal after excavation (Rowe and
The HWK EE assemblage is assessed through comparison with Rozeik, 2008; Peters et al., 2017). Other consolidants and adhe-
experimental samples that were first developed by Blumenschine sives were used when CDD was not appropriate or when it was
(1988, 1995), expanded upon by Capaldo (1995), and then refined insufficient to ensure that fragile or fragmented material did not
and described in detail by Pante et al. (2012). Five feeding scenarios suffer unnecessary damage during the excavation process. Daily
were modeled and characterized based on the proportions of tooth, temperature oscillations between 8 and 38 C limited choices to
cut and percussion marks in the assemblages. The models include those with higher glass transition temperatures (Tg). Thus, paraloid
two control scenarios modified by a single actor, two simulations of B-72 (ethyl methacrylate/methyl acrylate co-polymer) or B-44
primary access to carcasses by hominins followed by carnivore (methylmethacrylate/ethylacrylate) were used, with a preference
ravaging, and one scenario that simulates passive scavenging of for B-44 due to its higher Tg.
completely defleshed carcasses by hominins. The two control sce- Specimens that needed further conservation before being
narios are the hammerstone only (HO) and carnivore only (CO) studied were singled out during the excavation and cataloguing
models, both of which simulate complete consumption of carcasses processes. Treatments included mechanical removal of sedimen-
by humans or carnivores, respectively. The simulations of primary tary accretions with a variety of dental tools and scalpels, or with
access to carcasses by hominins followed by carnivore ravaging rotary tools when accretions were impenetrable by hand tools.
include the hammerstone-to-carnivore (HeC) model, in which Great care was taken to avoid contact with bone surfaces during the
hominins would have had primary access to both flesh and marrow, removal of accretions. Fossils that had been consolidated or
and the whole bone-to-carnivore (WBeC) model, which simulates repaired in situ required further treatment when the consolidant or
primary access by hominins to flesh and subsequent carnivore adhesive used in the trenches enclosed soil or when alignment of
consumption of grease and marrow. The simulation of scavenging is broken pieces needed to be corrected. After the resins were reduced
the vulture-to-hominin-to-carnivore (VeHeC) model in which or removed, the finds were repaired with B-44 applied with glass
M.C. Pante et al. / Journal of Human Evolution xxx (2017) 1e21 3
Figure 1. (A) Location of HWK EE in Olduvai Gorge (map template after Jorayev et al., 2016). (B) Aerial view of HWK EE trenches. (C) Stratigraphic correlation of HWK EE trenches
(simplified from de la Torre et al., submitted-a, Figure 20).
capillary tubes, micro spatulas or brushes, while gap-fills were Table 2

made with a mixture of B-44, acetone and glass microballoons Number of identified specimens (NISP) for limb bones in the analyzed and
comparative samples.a
(microscopic glass spheres used as fillers due to their low mass). All
treatments were documented, photographed and entered in the Stratigraphic Analyzed Comparative
conservation database. interval
All limb Midshaft All limb Midshaft
A total of 29,808 mammal fossils were recovered in stratigraphic bones fragments bones fragments
position from HWK EE (de la Torre et al., submitted-a), of which Tuff IIB 21 10 0 0
29,115 are included in Table 1. A subset of 6017 were given unique LAS 992 699 334 217
coordinates, of which 5324 were from large mammals and are Lemuta 843 645 412 311
Total 1856 1354 746 528
analyzed in detail here. The remaining 24,484 fragments were
a
predominantly recovered during screening. These specimens were See text for definition of samples. LAS, Lower Augitic Sandstone.
counted, and those that were found to be identifiable were
removed and included in the detailed taphonomic and taxonomic
analyses. None of these totals includes surface material, which was Refits were attempted within and between archaeological levels
mapped and collected but excluded from the analyses presented prior to any calculations, with a total of 28 sets of refits discovered.
here. Most refits were the result of recent breaks, and two refits were
The HWK EE assemblage is separated into three different found between levels (L1/L4 and L8/LCHA). In cases of ambiguity
stratigraphic intervals from oldest to youngest: Lemuta, LAS and concerning any specimen overlap, the faunal specimens were
Tuff IIB. Lemuta and LAS are likely contemporaneous with Mary directly compared to one another. These comparisons increased
Leakey's clay and sandy conglomerate, respectively (Pante and de la MNE totals as even small areas of overlap identified the presence of
Torre, submitted). Most analyses focus on the Lemuta and LAS in- multiple elements that would have been missed if previous iden-
tervals due to the limited number of specimens recovered from Tuff tifications alone were consulted. These calculations were done
IIB. separately for each archaeological level, and the levels within each
The sample of limb bones used for statistical comparisons with stratigraphic interval were subsequently added together to provide
feeding trace models (comparative sample throughout) is detailed the MNE estimates presented here.
in Table 2. A total of 746 limb bone fragments from the LAS and Long bone portions were divided into epiphyses (fragments
Lemuta intervals were included in this sample. The comparative with an articular surface, including the proximal ends of meta-
sample was extracted from the larger analyzed sample of 1856 limb podials), near-epiphyses (fragments with cancellous bone on the
bone fragments to enhance comparability with feeding trace medullary surface but no articular surface), and midshafts (frag-
models that are limited to bovids in size groups 1e4 (see Bunn, ments that did not have an articular surface or cancellous bone).
1982 for description of size groups), and have not been subjected Specimens that had multiple portions were identified by the most
to the post-depositional processes experienced by the fossil diagnostic portion; epiphyses were the most diagnostic and mid-
assemblage, which could affect assemblage-wide frequencies of shafts the least. Therefore, specimens identified as an epiphysis or
bone surface modifications (see Blumenschine, 1995; Pante et al., near-epiphysis may also include a midshaft portion.
2012; Pante, 2013). Excluded specimens were 1) less than 20 mm Metric measurements collected from each specimen included
in maximum dimension; 2) had been recently broken with more length, width, thickness, and cortical thickness (for limb bones
than 10% of the fragment missing, estimated based on the size of only). Measurements were recorded to the nearest millimeter using
the bone fragment and the length of the broken edge; 3) had poor digital calipers. Length was measured as the longest dimension of
surface visibility due to mechanical rounding, exfoliation, and/or the specimen, width was taken perpendicular to length at the
adhering matrix covering more than 10% of the fragment; 4) were widest point and thickness was taken perpendicular to width at the
not green broken, determined by the presence of other fracture thickest point. Cortical thickness was taken from the thickest
types, such as step or transverse; or 5) were known to be from an portion of cortical bone exposed on limb bone fragments. Length,
animal larger than size 4 or a taxonomic family other than bovid. preserved circumference, and cortical thickness were used to help
assess animal size groups from midshaft specimens.
Taphonomic analyses were conducted following Blumenschine
2.3. Assemblage analysis et al. (1996) and published morphological and contextual criteria
for distinguishing surface marks (e.g., Bunn, 1981; Potts and
Taxonomic and skeletal part and portion identifications were Shipman, 1981; Blumenschine and Selvaggio, 1988; Njau and
carried out using comparative collections located on-site at Olduvai Blumenschine, 2006; Blumenschine et al., 2006; Dominguez-
Gorge and in the Zooarchaeology and Paleoanthropology Lab at Rodrigo et al., 2009b). Bone surfaces were inspected macroscopi-
Colorado State University. Minimum number of element (MNE) cally under a 60w table lamp followed by closer inspection with a
calculations were conducted following the taxonomic and tapho- 10 hand lens. The orientation of the bone in relation to the light
nomic analyses and were based on the taxonomic, body size, source was systematically altered to allow perception of depth for
skeletal element, and element portion descriptions found therein. each mark following Blumenschine (1995). Tooth, cut and percus-
sion marks were also observed using a portable Dino-lite micro-
scope at magnifications up to 60. Notable marks were
Table 1
Number of identified specimens (NISP) of mammal fossils for all HWK EE trenches.
photographed with the portable microscope and/or a Nikon D7100
camera equipped with a 40 mm macro lens and Helicon focus
Stratigraphic NISP NISP NISP Non-individualized Total stacking software. In some cases, marks were also molded or
intervala bones teeth specimensb
scanned directly with a Nanovea ST400 white-light confocal pro-
Tuff IIB 35 34 491 560 filometer following Pante et al. (2017) to produce measurable three
LAS 2614 486 21,319 24,419
dimensional (3D) models.
Lemuta 1847 308 1981 4136
Total 4496 828 23,791 29,115 Percussion and tooth notches were identified and recorded
a
following Capaldo and Blumenschine (1994). Hammerstone per-
LAS, Lower Augitic Sandstone.
b
May include non-mammal fossils.
cussion notches were identified on the basis of having a broad and
arcuate plan form, while carnivore tooth notches were identified on assemblages from the LAS and Lemuta intervals are compared with
the basis of having a semi-circular plan form. Some notches could feeding trace models.
not be assigned to either of these categories using these criteria and
were considered indeterminate. 3.1. Length of limb bone midshaft fragments in the comparative
sample
2.4. Statistical analyses
The length of midshaft specimens in the LAS and Lemuta
The results for the HWK EE fossil assemblage were compared comparative assemblages from HWK EE are both strongly corre-
with the 95% interquantile ranges used by Pante et al. (2012) to lated with the HeC, CO and HO models (Table 3, Fig. 2). The LAS
characterize the feeding trace models; the 95% interquantile ranges assemblage is more strongly correlated with the models than the
were based on the 2.5% and 97.5% percentiles for the distributions Lemuta assemblage, which has a minor deficiency in specimens
of means that were generated using a bootstrap algorithm that was between 20 and 30 mm in length in comparison with the models.
designed to accurately represent the behavioral processes that
contributed to the feeding trace models (see Pante et al., 2012 for 3.2. Skeletal part profiles
further details). The 95% interquantile ranges were used to assess
differences between the HWK EE assemblage and the feeding trace The HWK EE assemblage consists of a minimum of 941 skeletal
models at the 0.05 level of probability. elements with 591 coming from the LAS interval, 334 from the
Lemuta interval, and 35 from the Tuff IIB interval (Table 4). These
3. Results results are presented by trench and level in Supplementary Online
Material (SOM) S1. The most abundant elements are limb bones,
The HWK EE assemblage preserves abundant traces of both followed by axial bones, compact bones, cranial bones, and pelves/
hominin and carnivore feeding on large mammal carcasses. There is scapulae. A closer look at the limb bones (Table 5) shows that their
also evidence that hominins and carnivores occasionally fed from distribution is relatively even in both the LAS and Lemuta intervals
the same carcasses at the site. Below, the results for the when separated into animal size groups 1e2 and 3e4. The excep-
tions are a deficiency in the abundance of the ulnae for all but the
Table 3 size group 1e2 sub-sample of the LAS and a relatively low number
Spearman's statistics for comparisons between length of midshaft fragments in the of metatarsals for size group 3e4 of the Lemuta. The femur and
comparative samples and feeding trace models.a
humerus are also slightly underrepresented in the size group 3e4
Statistics HeC CO HO sub-samples of both intervals.
Spearman's p-value Spearman's p-value Spearman's p-value
rs rs rs 3.3. Comparisons with feeding trace models
LAS 0.98 <.001 0.93 <.001 0.97 <.001
Lemuta 0.75 0.02 0.69 0.04 0.74 0.03 Fossils in the HWK EE collection exhibit percussion marks
a
Data shown in Figure 1 and described in the text. CO, Carnivore only; HO, Ham- (Fig. 3), mammalian carnivore and crocodile tooth marks (Fig. 4),
merstone only; HeC, Hammerstone-to-Carnivore; LAS, Lower Augitic Sandstone. and cut marks (Fig. 5). Results for the proportions of bones bearing
Figure 2. Length of limb bone midshaft fragments from the comparative samples of the Lower Augitic Sandstone (LAS) and Lemuta intervals compared with those from the feeding
trace models: HeC, Hammerstone-to-Carnivore; CO, Carnivore only; HO, Hammerstone only.
6
Table 4
Number of identified specimens (NISP) and minimum number of elements (MNE) for HWK EE.a
Skeletal parts Tuff IIB Lower Augitic Sandstone Lemuta HWK EE site
Size 1 Size 2 Size 3 Size 4 Size 5 Size 6 Total Size 1 Size 2 Size 3 Size 4 Size 5 Size 6 Total Size 1 Size 2 Size 3 Size 4 Size 5 Size 6 Total Grand total
NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE NISP MNE
Cranial 1 e 0 e 0 e 0 e 0 e 0 e 1 e 22 e 22 e 14 e 0 e 0 e 0 e 58 e 15 e 41 e 17 e 0 e 0 e 0 e 73 e 132 e
fragment
Frontal 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1
Horn core 0 0 0 0 1 1 0 0 0 0 0 0 1 1 10 7 15 9 11 7 0 0 0 0 0 0 36 23 7 4 6 4 3 2 1 1 0 0 0 0 17 11 54 35
Hyoid 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Maxilla 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 2 4 2 5 3 1 1 0 0 0 0 13 8 1 1 4 2 2 2 0 0 0 0 0 0 7 5 20 13
Mandible 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 4 18 7 12 8 0 0 2 2 0 0 39 21 8 4 9 5 9 7 1 1 0 0 0 0 27 17 66 38
Nasal 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 1 1 1
Occipital 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 2 2 1 1 0 0 1 1 0 0 0 0 0 0 2 2 4 4
Premaxilla 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 2 2 0 0 0 0 0 0 0 0 3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 3
Sphenoid 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2
Temporal 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 5 0 0 0 0 0 0 5 5 1 1 3 2 2 1 0 0 0 0 0 0 6 4 11 9
Zygomatic 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 0 0 1 1 0 0 0 0 0 0 0 0 1 1 2 2
Cranial total 1 0 0 0 1 1 0 0 0 0 0 0 2 1 43 14 63 22 51 27 1 1 2 2 0 0 160 66 33 11 65 15 34 13 2 2 0 0 0 0 134 41 296 108
Atlas 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 2 2 1 1 0 0 0 0 5 5 2 1 0 0 1 1 0 0 0 0 0 0 3 2 8 7
Axis 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 2 2 0 0 0 0 0 0 3 3 0 0 1 1 1 1 0 0 0 0 0 0 2 2 5 5
Caudal vertebra 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 3 4 4 1 1 0 0 0 0 0 0 8 8 1 1 1 1 1 1 0 0 0 0 0 0 3 3 11 11
Cervical vertebra 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 8 6 9 9 2 2 0 0 0 0 20 18 1 1 2 2 17 7 0 0 1 1 0 0 21 11 41 29
Clavicle 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Lumbar vertebra 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 3 17 13 5 4 1 1 0 0 0 0 26 21 2 2 9 4 4 0 0 0 0 0 0 0 15 6 41 27
Rib 2 1 0 0 0 0 0 0 0 0 0 0 2 1 39 10 67 8 59 11 9 4 2 2 1 1 177 36 43 6 92 12 73 13 16 4 5 3 1 1 230 39 409 76
Rib 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 1 1
Sacrum 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 6 5 1 1 0 0 0 0 0 0 8 7 0 0 1 1 4 2 0 0 1 1 0 0 6 4 14 11
Sternum 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 0 0 0 0 1 1 1 1 0 0 0 0 2 2 3 3
Thoracic vertebra 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 6 26 17 13 8 1 1 0 0 0 0 47 32 1 1 12 8 12 7 0 0 0 0 0 0 25 16 72 48
Vertebra 0 e 0 e 0 e 0 e 0 e 0 e 0 e 3 e 17 e 4 e 0 e 0 e 0 e 24 e 9 e 12 e 6 e 0 e 0 e 0 e 27 e 51 e
fragment
Axial total 2 1 0 0 0 0 0 0 0 0 0 0 2 1 59 26 146 54 97 39 14 9 2 2 1 1 319 131 59 12 130 29 121 34 17 5 7 5 1 1 335 86 656 218
Femur 0 0 0 0 1 1 0 0 0 0 0 0 1 1 10 7 13 8 21 11 0 0 0 0 1 1 45 27 10 4 20 4 12 7 1 1 0 0 1 1 44 17 90 45
Fibula 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 3 3 0 0 0 0 0 0 5 5 0 0 1 1 0 0 0 0 0 0 0 0 1 1 6 6
Humerus 0 0 2 1 1 1 0 0 0 0 0 0 3 2 6 4 31 11 14 8 1 1 1 1 0 0 53 25 16 5 21 6 22 6 3 3 0 0 1 1 63 21 119 48
Limb bone 1 e 2 e 2 e 0 e 0 e 0 e 5 e 176 e 286 e 126 e 7 e 6 e 0 e 601 e 161 e 201 e 98 e 2 e 1 e 0 e 463 e 1069 e
fragments
Metacarpal 1 1 0 0 3 3 0 0 1 1 0 0 5 5 4 4 17 8 20 11 2 2 3 3 0 0 46 28 8 5 21 5 21 10 0 0 1 1 0 0 51 21 102 54

Metapodial 0 e 1 e 0 e 0 e 0 e 0 e 1 e 4 e 24 e 24 e 0 e 1 e 0 e 53 e 4 e 11 e 12 e 0 e 0 e 0 e 27 e 81 e
Metatarsal 0 0 1 1 1 1 0 0 0 0 0 0 2 2 6 4 26 10 15 13 2 2 0 0 0 0 49 29 11 6 27 7 10 6 0 0 0 0 0 0 48 19 99 50
Radius 0 0 1 1 1 1 0 0 0 0 0 0 2 2 6 4 18 7 22 13 1 1 0 0 1 1 48 26 11 5 25 6 20 10 1 1 0 0 0 0 57 22 107 50
Tibia 0 0 1 1 1 1 0 0 0 0 0 0 2 2 4 3 36 9 23 10 3 3 5 2 0 0 71 27 13 5 22 9 31 12 0 0 1 1 0 0 67 27 140 56
Ulna 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 4 10 7 5 5 1 1 0 0 0 0 21 17 2 2 9 4 11 5 0 0 0 0 0 0 22 11 43 28
Appendicular 2 1 8 4 10 8 0 0 1 1 0 0 21 14 222 31 462 61 273 74 17 10 16 6 2 2 992 184 236 32 357 42 237 56 7 5 3 3 2 2 843 139 1856 337
total
M.C. Pante et al. / Journal of Human Evolution xxx (2017) 1e21
Astragalus 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 3 7 7 3 3 0 0 1 1 1 1 15 15 4 4 3 3 0 0 0 0 0 0 0 0 7 7 22 22
Calcaneus 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 4 3 3 3 3 1 1 0 0 0 0 11 11 0 0 1 1 0 0 0 0 0 0 0 0 1 1 12 12
Carpal or tarsal 0 e 0 e 0 e 0 e 0 e 0 e 0 e 1 e 3 e 8 e 0 e 1 e 0 e 13 e 0 e 0 e 1 e 0 e 1 e 0 e 2 e 15 e
indet.
Cuneiform 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 1 1 0 0 0 0 0 0 3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 3
Fibula 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1
External-medial 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 5 5 3 3 0 0 1 1 0 0 10 10 1 1 1 1 2 2 0 0 0 0 0 0 4 4 14 14
cuneiform
Lunate 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 2 2 0 0 0 0 0 0 3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 3
Magnum 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 3 3 4 4 1 1 0 0 0 0 10 10 1 1 1 1 2 1 0 0 0 0 0 0 4 3 14 13
Navicular-cuboid 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 4 4 6 6 0 0 0 0 0 0 11 11 1 1 0 0 1 1 0 0 0 0 0 0 2 2 13 13
Patella 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 2 2 0 0 0 0 0 0 3 3 0 0 3 3 0 0 0 0 0 0 0 0 3 3 6 6
Phalange 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 4 12 9 15 14 0 0 1 1 0 0 32 28 3 3 5 5 0 0 0 0 1 1 0 0 9 9 41 37
proximal
Phalange 0 0 0 0 1 1 0 0 0 0 0 0 1 1 7 7 6 6 13 13 0 0 0 0 0 0 26 26 2 2 2 2 3 3 0 0 0 0 0 0 7 7 34 34
intermediate
Phalange distal 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 9 8 7 7 0 0 0 0 0 0 18 17 0 0 1 1 0 0 0 0 0 0 0 0 1 1 19 18
Phalange 0 e 1 e 0 e 0 e 0 e 0 e 1 e 5 e 9 e 4 e 0 e 0 e 0 e 18 e 0 e 3 e 0 e 0 e 0 e 0 e 3 e 22 e
fragment
Pisiform 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 4 0 0 0 0 0 0 4 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 4
Sesamoid 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 3 3 3 0 0 0 0 0 0 6 6 1 1 1 1 0 0 0 0 0 0 0 0 2 2 8 8
Scaphoid 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 4 4 2 2 0 0 0 0 0 0 8 8 0 0 0 0 2 2 0 0 0 0 0 0 2 2 10 10
Trapezoid 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1
Table 5
e
2
201
941
941
34
43
77 Summary of limb bones by animal size groups.a
74
2
3228
1268
4496
244
102
176
Skeletal part LAS Lemuta

41
11
16
27
e
0
1445 334
1847 334
Size 1e2 Size 3e4 Size 1e2 Size 3e4
0
47
39
47
86
Humerus 15 9 11 9
Femur 15 11 8 8
Radius 11 14 11 11
0
0
0
0
0
3
e
Ulna 11 6 6 5
0
0
0
0
0
3
e
Tibia 12 13 14 12
Metacarpal 12 13 10 10
10
e
10
0
2
0
0
0
Metatarsal 14 15 13 6
0
2
0
0
0
12
e
12
a
Values represent minimum number of elements (MNE) for the Lower Augitic
Sandstone (LAS) and Lemuta intervals.
e
0
0
1
2
3
15
15
29
e
29
0
0
1
2
3
bone surface modifications at HWK EE are based on the compara-

e
0
9
6
6
124
124
12
tive samples unless otherwise specified. The ‘all limb bone frag-
11
26
18
44
e
0
447
447
ments’ category includes epiphyses, near-epiphyses, and midshaft

fragments from the humerus, femur, radius, ulna, tibia, and meta-
0
3
6
9
113
113
18
podials. The proportions of bones with modifications in the LAS and

21
24
31
e
0
604
604
Lemuta intervals can be found in Table 6, and Figures 6e9, and they
are summarized separately below when they differ and together
13
71
e
71
0
1
2
3
when they do not.

e
0
5
3
8
349
349
13
35 23
55 27
90 50
e
2
195 160
1756 591
2614 591
3.4. Percussion marks

2
858
The proportion of percussion-marked bones is higher in the

e
0
1
0
0
0
Lemuta interval than the LAS interval (Table 6, Fig. 6). However, the
0
1
0
0
0
4
e
assemblages compare similarly to the interquantile ranges of the

feeding trace models. The incidence of percussion marking for the
14
e
14
0
3
1
0
1
comparative sub-samples from the LAS interval falls within the 95%
0
4
1
0
1
25
e
25
interquantile ranges of all sub-samples of the VeHeC model,

e
0
2
2
0
2
24
24
within both size group 3e4 sub-samples of the HeC model, and
within the size group 3e4 sub-sample for midshaft fragments of
37
e
37
0
2
3
0
3
the HO model. The incidences of percussion marking for the

e
1
227
227
69
12
18
comparative sub-samples from the Lemuta interval fall within the

81
15
24
e
1
526
526
95% interquantile ranges of all sub-samples of the VeHeC model,

within the size group 3e4 sub-sample for all limb bone fragments
1
9
7
211
211
58
16
of the HeC model, and within the size group 3e4 sub-sample for
74
15
22
37
e
1
782
782
midshaft fragments of the HO model. It falls just outside the upper

27
13
e
0
5
8
111
111
range of the size group 3e4 sub-sample for midshaft fragments of

the HeC model and just below the lower range of the size group 12
e
0
382
382
33
18
25
sub-sample for midshaft fragments of the HO model.

e
0
1
0
0
0
17
17
27
35
0
2
0
0
0
3.5. Mammalian carnivore tooth marks

0
0
0
0
0
0
e
The proportion of tooth-marked bones is higher in the Lemuta

e
0
0
0
0
0
interval than the LAS interval (Table 6, Fig. 7). However, both as-
semblages fall within the same interquantile ranges of the feeding
e
0
0
0
0
0
trace models and are summarized together. The incidences of tooth

0
0
0
0
0
1
e
marking for the comparative samples from the LAS and Lemuta
e
0
0
0
0
0
intervals fall within the 95% interquantile ranges for both size
group 1e2 sub-samples of the VeHeC model and within the large
e
0
0
0
0
0
interquantile range of the size group 3e4 sub-sample for midshaft

e
0
1
0
0
0
10
10
fragments of the WBeC model. In general, the proportions of tooth-

12
e
12
0
1
0
0
0
Size groups based on Bunn, 1982.
marked bone in the assemblages from the LAS and Lemuta intervals
fall intermediate to the HeC and CO models.
0
0
0
0
0
4
e
e
0
1
0
0
0
3.6. Cut marks

e
0
0
0
0
0
The proportion of cut-marked bones is higher in the Lemuta

0
0
0
0
0
5
e
interval than the LAS interval for all but size group 3e4 midshaft
mammal frags
Total identified
Indeterminate
Compact total
Pelvis/scapula
fragments (Table 6, Fig. 8). However, the assemblages again fall

Innominate
Grand total
Unciform
within the same interquantile ranges of the feeding trace models

total
Scapula
and are summarized together. The incidences of cut marking for the
a
all limb bone fragment sub-samples of the LAS and Lemuta
Figure 3. Percussion marks on limb bones from HWK EE. (A) metatarsal midshaft from a size 3 bovid (T1_LCHA_179, LAS); (B) radius midshaft fragment from a size 3 bovid
(T1_L2_2109, LAS); (C) radius midshaft fragment from a size 4 bovid (T1_L6_174, LAS).
intervals fall within the 95% interquantile ranges of the size group end of the ranges reported for the feeding trace models, as all
1e2 sub-sample of the WBeC model and the size group 3e4 sub- sub-samples have values at or below 6.1% (Table 6, Fig. 9). The
samples of the HeC and VeHeC models. The incidences of cut intervals are similar with the exception of the size group 3e4
marking for sub-samples of midshaft fragments for the LAS and sub-sample of midshaft fragments for which the LAS interval
Lemuta intervals fall within the 95% interquantile ranges of the size lacks any specimens that have the co-occurrence of tooth and
group 1e2 sub-samples of the HO and WBeC models and the size butchery marks.
group 3e4 sub-samples of the HO and HeC models. The proportion of tooth- and butchery-marked specimens in the
LAS interval falls within all sub-samples of the VeHeC model and
3.7. Tooth and butchery marks the size group 1e2 sub-sample for midshaft fragments of the
WBeC model. The values for all sub-samples fall below 3% with the
The proportion of specimens that are both tooth- and exception of the size group 3e4 sub-sample for all limb bone
butchery-marked in the LAS and Lemuta intervals is on the lower fragments, which has a value of 5%.
Figure 4. Tooth marks on fossils from HWK EE. (A) Large furrows on a size 4 ilium (T1_LCHA_86, LAS); (B) Crocodile tooth mark on a complete size 3 bovid radius (T1_L2_4176, LAS);
(C) Cut mark overlying a tooth mark on a size 1 bovid femur fragment (T1_L10_1383, Lemuta).
The proportion of tooth- and butchery-marked specimens in the attachment, such as the deltoid tuberosity of the humerus or the
Lemuta interval falls within all sub-samples of the VeHeC model, lesser trochanter of the femur, and cut marks appear to be
both size group sub-samples for midshaft fragments of the WBeC concentrated in several areas like the distal tibia, distal femur and
model, and the size group 3e4 sub-sample for all limb bone frag- proximal humerus.
ments of the HeC model. The size group 3e4 sub-samples have Cut marks are found on taxa in all size groups and on many parts
more tooth- and butchery-marked specimens than the size group of the skeleton (Table 7, Fig. 5). Taxa that are cut-marked in the
1e2 sub-samples. assemblage include bovids in size groups 1e4, equids, suids, hip-
popotamids, giraffids, a proboscidean, and a hyaenid. Skeletal parts
3.8. Cut mark patterns that are cut-marked include mandibles, vertebrae, ribs, all limb
bones, innominates, scapulae and some podials, although the ma-
Cut marks on bone fragments that could be precisely located jority of cut marks are found on limb bones. Those found on limb
on the appendicular skeleton (Fig. 10) show most cuts were bones are mostly evenly distributed across the flesh bearing ele-
inflicted on limb bone near-epiphyses or midshafts rather than ments when sample sizes are large enough to make this assessment
epiphyses. Many cut marks occur near places of muscle (Table 8).
Figure 5. (continued).
Figure 5. Cut marks on fossils from HWK EE. (A) Proboscidean astragalus; (T1_L2_433, LAS); (B) Crocuta scapula (T1_L4_294, LAS); (C) Proximal humeri of a size 2 bovid (T1_L4_330,
LAS); and shaft/near epiphyses of a size 1 juvenile bovid; (T1_L4_1567, LAS); (D) Tibia shaft of giraffid (T1_L6_1233, LAS); (E) Complete rib from a size 3 bovid (T1_L10_1105,
Lemuta); (F) Mandible of a hippopotamid (T1_L6_1621, LAS).
3.9. Notches
the 2.5% and 97.5% quantiles. Shaded boxes and bold text indicate where the comparative samples from the LAS and Lemuta interval, respectively, are within the 95% interquantile ranges and therefore, statistically indis-
Data for feeding trace models from Pante et al. (2012). Mean % is the proportion of specimens bearing at least one mark of the specified type. The tooth and butchery mark category is the proportion of specimens bearing at
least one tooth and one percussion and/or cut mark. CO, Carnivore only; HO, Hammerstone only; HeC, Hammerstone-to-Carnivore; VeHeC, Vulture-to-Hominin-to-Carnivore. 95% I.Q.R. (Interquantile range), values between
Lemuta
e
12.6
23.5
37.4
45.7
The limb bones from both intervals of HWK EE preserve tooth
9.9
2.2
6.1
10
16.7 and percussion notches (Table 9). The incidence of tooth notching
33.7
10.4
falls well below the CO model and above the HeC model. The in-
LAS
6.5
4.7
1.7
e
e
25
0
cidences of percussion marking falls below both the HO and HeC
models.
11.8e58.8
5.9e47.1
3.7e38.3
VHC
0e29.4
0e10.0
0e14.4
4. Discussion
23.6
19.2
35.5
11.6
5.6
2.5
0
0
0
0
The large mammal fossil assemblage from HWK EE is tapho-
Midshaft fragments
23.6e35.8
10.5e18.9
10.9e20.6
7.9e16.9
8.1e22.4
7.3e23.5
nomically complex due to a fluvial component to its deposition (de

3.2e9.4
1.3e5.8
HeC
la Torre et al., submitted-a), but it is also rich in the feeding traces of

29.4
14.6
14.5
14.8
15.4
12.2
5.9
3.2 both hominins and carnivores. The assemblage offers evidence of

multiple occupations at the site despite indications, such as the
18.1e89.8
56.1e84.7
13.0e80.1
2.5e36.0
2.6e24.4
WBeC
presence of carnivore remains and high carnivore tooth mark fre-

0e19.7
e
e
e
e
quencies, that the environment was highly competitive. Beyond

70.5
57.3
16.4
40.0
12.3
8.2
this competition for faunal resources, hominins likely had a high

risk of predation from both mammalian carnivores and crocodiles.
11.3e60.0
14.0e37.8
3.1e17.6
The behavior of hominins at the site appears to have varied be-

0e40.0
HO
e
e
e
e
e
e
e
e
tween the two main stratigraphic intervals represented, and this

26.7
33.5
10.3
20.1
will be the focus of the interpretations presented below.

54.8e81.0
75.7e95.8
4.1. Effect of fluvial processes

CO
e
e
e
e
e
e
e
e
e
e
e
e
69.1
86.5
Evidence for fluvial activity at HWK EE is clear in the stratigraphy

of the site, particularly during the LAS interval. However, proxies
Lemuta
such as rounding of lithic specimens, specimen length distributions,

e
e
11.6
39.4
45.9
10.3
14.1
2.1
5.8
and orientation of elongated specimens show only minimal evi-

20
dence of fluvial transport and abrasion of the material (de la Torre

11.9
31.3
35.6
12.9
et al., submitted-a). The length distributions of midshaft fragments

LAS
7.3
2.6
e
e
6
5
Bone surface modifications for the comparative samples of the LAS and Lemuta intervals and feeding trace modelsa.
in the comparative sample (Fig. 2) also suggest hydraulic processes

had only a minimal effect on fossil assemblages from both the LAS
21.7e65.2
4.4e34.8
7.7e49.1
1.6e22.3
9.1e28.7
and Lemuta intervals. Both intervals are strongly correlated with the
VHC
0e18.4
0e21.7
length distributions of the feeding trace models, none of which was

17.5
26.9
43.4
18.5
8.5
8.7
9.9
exposed to fluvial processes. Low-energy fluvial processes can

0
0
All limb bone fragments
transport bone fragments differentially, with smaller fragments

11.5e25.5
12.1e21.9
24.3e36.6
14.9e23.7
19.6e30.3
14.4e24.4
5.5e10.9
5.4e12.5
more likely to be transported than their larger counterparts (Pante

HeC
and Blumenschine, 2010). The similarity in length distributions

30.3
18.0
19.1
24.9
18.9
16.8
8.6
8.0
between HWK EE and the feeding trace models suggests the effect of
fluvial processes is minimal enough to not alter interpretations
55.7e84.3
61.9e95.0
22.1e69.5
16.7e57.5
5.5e34.8
3.3e22.4
WBeC
based on comparisons with the models.

e
e
e
e
70.6
78.9
18.4
41.9
11.6
33.7
4.2. Transport of skeletal parts to HWK EE

26.3e46.2
29.1e70.8
20.3e33.7
25.0e66.7
Skeletal part profiles from HWK EE show that high-density limb

HO
e
e
e
e
e
e
e
e
bones are the most common elements in the assemblage, but axial,
36.6
49.9
27.2
45.9
cranial, and compact bones (podials) are also well represented.

This suggests that low-density elements were not transported
56.7e82.5
77.9e95.5
away from the site by fluvial processes (see Voorhies, 1969;

CO
e
e
e
e
e
e
e
e
e
e
e
e
Behrensmeyer, 1975) or completely consumed by carnivores. The

70.9
87.6
size group 1e2 sub-sample of the LAS interval has a relatively even
distribution of limb bones, with the exception of the ulna, which is
95% I.Q.R
95% I.Q.R
95% I.Q.R
95% I.Q.R
95% I.Q.R
95% I.Q.R
95% I.Q.R
95% I.Q.R
Statistics
Mean %
Mean %
Mean %
Mean %
Mean %
Mean %
Mean %
Mean %
usually only represented by part of its proximal end or a portion of

the midshaft fused to a radius fragment. This suggests an unbiased
strategy of transport of smaller animals to the site by hominins. This
tinguishable from the models.
is also true of the size group 1e2 sub-sample of the Lemuta interval,
Animal size
except that the femur is slightly underrepresented. The size-group

Size 1e2
Size 3e4
Size 1e2
Size 3e4
Size 1e2
Size 3e4
Size 1e2
Size 3e4
3e4 sub-sample of the LAS and Lemuta intervals are both deficient
in relatively high utility femora and humeri (see Metcalfe and Jones,
1988 for discussion of food utility indices), suggesting they were
butchery
Percussion
not transported to the site as frequently as lower utility elements.

Mark type
Tooth and
This may suggest that hominins processed these elements off-site

Tooth
Table 6
or that hominins did not have regular access to the humerus and
Cut
femur when acquiring the carcasses of large animals.
Figure 6. Incidence of percussion-marked bone for (A) animal size group 1e2, (B) animal size group 3e4. HO, Hammerstone only; HeC, Hammerstone-to-Carnivore; VeHeC,
Vulture-to-Hominin-to-Carnivore. Data shown for feeding trace models include the mean and 95% interquantile range. Data shown for LAS and Lemuta represent the proportion of
percussion-marked bone in the comparative samples.
The pattern of less even representation of limb bones in larger even representation of limb bones in larger animals, it is not in the
animals observed at HWK EE has also been reported for other way that has been observed for other archaeological sites, specif-
archaeological sites, most notably for the Klasies River Mouth ically a dominance of foot bones and head bones when the com-
assemblage (Klein, 1989). It has been argued that these patterns are plete skeleton is considered (Klein, 1989), also known as the
likely the result of the methods used to reconstruct skeletal part Schlepp Effect (Perkins and Daly, 1968). The tibia and radius are
profiles at archaeological sites (Marean and Frey, 1997; Marean and well represented in both intervals and the metatarsals are actually
Kim, 1998; Bartram and Marean, 1999). Marrow extraction by underrepresented in the Lemuta interval. This suggests that the
hominins using a hammerstone-on-anvil technique fragments the HWK EE skeletal part profile is not the result of the methods
bones of larger animals two to four times more than those of employed, which included attempting to refit fossils (see de la Torre
smaller animals due to the greater amount of force required et al., submitted-a) and considered midshaft fragments in the
(Bartram and Marean, 1999). When these fragmented bones are construction of MNEs.
subsequently ravaged by bone crunching carnivores, those with
low-density epiphyses (i.e., femora and humeri) appear to be un- 4.3. The feeding behavior and ecology of hominins at HWK EE
derrepresented when midshaft fragments are not used in the
reconstruction of skeletal part profiles (see also Marean and Two of the stratigraphic intervals contained at the HWK EE site
Spencer, 1991). While the HWK EE assemblage does show less preserve substantial traces of hominin and carnivore feeding. The
Figure 7. Incidence of tooth-marked bone for (A) animal size group 1e2, (B) animal size group 3e4. CO, Carnivore only; WBeC; Whole Bone-to-Carnivore; HeC, Hammerstone-to-
Carnivore; VeHeC, Vulture-to-Hominin-to-Carnivore. Data shown for feeding trace models include the mean and 95% interquantile range. Data shown for LAS and Lemuta
represent the proportion of tooth-marked bone in the comparative samples.
older Lemuta interval shows intense use of carcasses by both hom- less likely for the Lemuta member because the majority of the
inins and carnivores, as evidenced by high frequencies of bone sur- assemblage came from a single archaeological level (L10 in T1-Main
face modifications. Limb bone fragments in the younger LAS interval Trench) and some of the Lemuta layers were finer grained, sug-
bear fewer traces of hominin and carnivore feeding. There are several gesting a lower energy depositional environment. However, the
possibilities to explain the overall lower frequencies of bone surface frequencies of tooth- and butchery-marked bone in the LAS assem-
modifications in the LAS interval. Lower tooth mark frequencies blage suggests that contributions of carcasses not associated with the
could be the result of reduced competition for carcass foods among hominin occupation of the site did not have a large impact on the
carnivores or earlier access to carcasses by hominins. However, assemblage, as they fall within ranges predicted by dual-patterned
neither explains the lower rates of percussion marking in the LAS. feeding trace models. The final possibility is that the Lemuta and
Mark frequencies could also have been depressed by the addition of LAS represent seasonal variations in the accumulation of the as-
carcasses that were not associated with the hominin occupation of semblages. Below we discuss data that support this last hypothesis.
the site, either through fluvial transport or through time averaging, Seasonal or climatic variations in the accumulation of the LAS
as multiple archaeological levels comprise the LAS interval. This is and Lemuta assemblages are indicated by isotope, mesowear and
Figure 8. Incidence of cut-marked bone for (A) animal size group 1e2, (B) animal size group 3e4. HO, Hammerstone only; WBeC; Whole Bone-to-Carnivore; HeC, Hammerstone-
to-Carnivore; VeHeC, Vulture-to-Hominin-to-Carnivore. Data shown for feeding trace models include the mean and 95% interquantile range. Data shown for LAS and Lemuta
represent the proportion of cut-marked bone in the comparative samples.
microwear data collected from bovid teeth (Rivals et al., (Bibi et al., submitted), while the substantially greater number of
submitted). Isotopes and mesowear are long-term measures of crocodile teeth in the LAS (de la Torre et al., submitted-a) suggests
diet and at HWK EE they indicate that antilopin and alcelaphin that crocodiles were drawn to the site possibly due to the presence
bovids were predominantly grazing in both the Lemuta and LAS of a pool that would have afforded a water source in an otherwise
(Rivals et al., submitted). However, dental microwear, a measure of dry landscape. Active pools are known to attract crocodiles during
what an animal was eating near the time of its death, suggests dry periods and result in concentrations of shed crocodile teeth
antilopin and alcelaphin bovids were browsing more during the (Njau, 2012).
LAS (Rivals et al., submitted). This could be the result of carcass Seasonality can affect the nutritional condition of animals at the
deposition in the LAS during a dry period or season when grasses time of their death, and this can alter the utility of carcass parts and
were in lower abundance. The high abundance of fish remains the feeding behavior of hominins and carnivores. Herbivores
within HWK EE indicates they died at the same time and in large deplete fat stores in the dry season when they are more
numbers and suggests the mortality pattern was the result of nutritionally-stressed (McNaughton and Georgiadis, 1986) and this
recurring seasonal events that caused fluctuations in water levels is known to directly affect the fat content of their bone marrow
Figure 9. Incidence of tooth- and butchery-marked bone for (A) animal size group 1e2, (B) animal size group 3e4. WBeC; Whole Bone-to-Carnivore; HeC, Hammerstone-to-
Carnivore; VeHeC, Vulture-to-Hominin-to-Carnivore. Data shown for feeding trace models include the mean and 95% interquantile range. Data shown for LAS and Lemuta
represent the proportion of tooth- and butchery-marked bone in the comparative samples.
(Sinclair and Duncan, 1972). Carnivores have been shown to ignore result of hominin and carnivore behavioral changes in response to
fat-depleted bones resulting in higher epiphyseal to shaft fragment the consumption of fat-depleted carcasses.
ratios and lower tooth mark frequencies (Blumenschine and Comparison with the feeding trace models shows that the as-
Marean, 1993). A chi-square test shows that the LAS assemblage semblages from both intervals do not fit any one model for all sub-
has a significantly higher epiphysis to shaft fragment ratio (.23) samples. However, nearly all size and portion sub-samples from
than the Lemuta (.15) assemblage (Table 10, p < 0.001, x2 ¼ 11.17, both intervals fall within the VeHeC model. This is, in part, the
d.f. ¼ 1) indicating a lower degree of carnivore ravaging in the LAS. result of the larger intervals for the VeHeC model, which occur
The value for the LAS falls within the range of modern carnivore due to smaller sample sizes of assemblages on which the model is
consumption of fat-depleted carcasses in the Serengeti based (see Pante et al., 2012), but it also reflects similarities in the
(Blumenschine and Marean, 1993) and suggests that many of the actors that inflicted modifications on bone surfaces through
carcasses in the LAS assemblage were fat-depleted at the time of feeding on carcasses. The co-occurrence of butchery and tooth
death. This would likely have also affected hominin consumption of marks on individual specimens allows elimination of both the
marrow, resulting in fewer percussion broken bones. Therefore, carnivore only and hammerstone only model as the sole scenarios
lower tooth and percussion mark frequencies in the LAS may be the for the behaviors that led to the accumulation of the assemblages
Figure 10. Cut mark locations and orientations on fossil bones from HWK EE. Cut marks on size 1e2 animals are represented by a blue line and regular font specimen number; cut
marks on size 3e4 animals are represented by a red line and bold font specimen number; cut marks on size 5e6 animals are represented by a green line and bold italic font
specimen number. Specimen numbers that are starred represent bones from the Lower Augitic Sandstone interval, while all others are from the Lemuta interval. Locations of cut
marks are estimated based on morphological features found on the fragments that they were discovered on.
in both intervals. Nonetheless, components of each in the assem- microstriations, were obscured by fluvial abrasion. It is also
blage cannot be ruled out. possible that hominins were breaking the limb bones of smaller
The overall incidence of percussion marking at HWK EE sug- animals more efficiently or in a way not replicated by the models,
gests hominins were breaking the majority of limb bones for thereby leaving fewer impact marks. Lastly, it cannot be ruled out
marrow extraction at the site. However, the incidences of per- that at least some of the carcasses, particularly in the LAS inter-
cussion marking for the size group 1e2 sub-samples of both in- val, were contributed only by carnivores, which would depress
tervals are relatively low and fall within only the lower end of the frequencies of percussion and cut marks while inflating tooth
VeHeC model. This does not necessarily suggest hominins were mark frequencies (see Pante, 2013). None of these scenarios, nor
not exploiting these smaller carcasses to the same extent as the effect of seasonal variations in carcass consumption hy-
their larger counterparts. It is possible that some of the less pothesized above, can be eliminated based on the percussion
conspicuous percussion marks, such as isolated patches of mark data alone.
Table 7 the mean values for the HeC model. The Lemuta assemblage only
Cut marks by taxon and skeletal parts.a falls within the ranges of models where either carnivores were
Taxa LAS Lemuta involved in the initial defleshing of carcasses (the VeHeC model
simulates carnivore consumption of some flesh for size 1e2 car-
Bovidae Y Y
Equidae Y Y
casses, but not size 3e4, see Pante et al., 2012 for discussion), or
Giraffidae Y Y where carnivores had sole access to marrow from defleshed limb
Hyaenidae Y N bones (i.e., the WBeC model). The results for tooth marking in the
Hippopotamidae Y N Lemuta interval also resemble those reported by Selvaggio (1994,
Proboscidea Y N
1998) for her three-stage carnivore-hominin-carnivore model
Suidae N Y
Skeletal parts (CeHeC), which simulates secondary access to carcasses by homi-
Mandible Y Y nins and reports an incidence of tooth marking of 65% for all limb
Lumbar vertebrae Y N bones and 47% for midshaft fragments. When considered with the
Rib Y Y
incidences of percussion marking in the assemblage, tooth mark
Thoracic vertebrae N Y
Femur Y Y
frequencies suggest the possibility for a CO component in the
Humerus Y Y accumulation of size 1e2 carcasses but not for the size group 3e4
Metacarpal Y Y sub-sample, which has a percussion mark frequency at the high end
Metatarsal Y Y of the HeC model. This suggests that the Lemuta hominins may have
Radius Y Y
had earlier access to size 1e2 carcasses than the tooth mark fre-
Tibia Y Y
Ulna Y Y quencies indicate, but likely later access to larger animal carcasses.
Astragalus Y N The incidence of tooth marking for sub-samples of the LAS in-
Calcaneus Y N terval fall within the same models as those for the Lemuta interval,
Phalange Y Y
but in all cases, the absolute values are lower than those in the
Innominate Y N
Scapula Y Y
Lemuta. In fact, the incidence of tooth marking for the size group
a
3e4 sub-sample of midshaft fragments in the LAS interval falls just
Taxa and skeletal parts that preserve at least one cut mark indicated by Y when
above the HeC model, suggesting that hominins may have had
present and N when not present for the Lower Augitic Sandstone (LAS) and Lemuta
intervals. relatively early access to these carcasses. It is also possible that the
lower incidence of tooth marking for the LAS interval, which fol-
Table 8 lows that of percussion marking, may be attributed to other taph-
Incidence of cut marks on midshaft fragments from flesh-bearing limb bones.a onomic processes such as time averaging or mixing of assemblages
Size group Skeletal part LAS Lemuta due to fluvial processes. Alternatively, the lower degree of carnivore
ravaging in the LAS, indicated by the high epiphysis to shaft frag-
Size 1e2 Femur 0 (4) 12.5 (16)
Humerus 10 (10) 13.3 (15)
ment ratios described above, would also result in lower tooth mark
Tibia 6.3 (16) 20 (15) frequencies. In fact, carnivore ravaging of fat-depleted carcasses
Radius 0 (8) 11.8 (17) often results in a complete absence of carnivore tooth marks
Size 3e4 Femur 25 (4) 25 (4) (Blumenschine and Marean, 1993), so the low degree of ravaging
Humerus 0 (2) 16.7 (6)
indicated for the LAS suggests that the tooth marks present in the
Tibia 28.6 (7) 10 (10)
Radius 50 (2) 9 (11) assemblage may have been inflicted during the initial defleshing of
a
carcasses by carnivores, refuting a hypothesis of earlier access to
Number of specimens on which percent cut are based given in parentheses.
LAS ¼ Lower Augitic Sandstone.
carcasses by hominins in the LAS interval. The incidence of tooth
and percussion marking for size 1e2 carcasses in the LAS interval
suggests the possibility of a substantial CO component much like
Table 9
Frequency of notches in the feeding trace models and for HWK EE.a
those for the Lemuta interval, while the lower values of both
relative to the Lemuta could again be the result of seasonal varia-
Carnivore tooth notch Hammerstone percussion notch tions in the accumulation of the assemblages.
CO 19.1 0 The incidence of cut marking in the HWK EE assemblage is high,
HO 0 26 especially for size group 3e4 carcasses where the values for both
HeC 2.2 16.4
intervals fall within the lower ranges of the HeC model, suggesting
LAS 5.4 7.8
Lemuta 6.5 6.5 relatively early access to flesh by hominins. Following the values for
a
percussion marks, the frequency of cut marking for the size group
Data for the feeding trace models taken from Blumenschine (1995). CO, Carnivore
only; HO, Hammerstone only; HeC, Hammerstone-to-Carnivore.
1e2 sub-samples of both intervals is lower overall. This trend of
fewer hominin-induced modifications on the limb bone fragments
of smaller carcasses again suggests that hominins did not have
Table 10
Limb bone portion frequencies for the HWK EE assemblage.a access to at least some of these animals. However, abrasive pro-
cesses cannot be completely ruled out for this pattern. Our ongoing
Limb bone portion LAS Lemuta
research shows that abrasion of bone surface modifications in a
Epiphyses 185 112 rock tumbler has a greater effect on cut marks than tooth marks,
Shafts 790 727
which is consistent with the pattern of modification observed for
Total 975 839
the HWK EE site, and could be depressing the hominin signal for
a
LAS, Lower Augitic Sandstone. size group 1e2 animals at the site.
Overall, the incidences of tooth- and butchery-marked bone in
The incidence of tooth marking in the HWK EE assemblage the HWK EE assemblage is low but not inconsistent with the
suggests a substantial mammalian carnivore component to the models, some of which allow for the complete absence of a co-
consumption of carcasses at the site. This is especially true of the occurrence of tooth and butchery marks on single specimens. The
assemblage from the older Lemuta interval, where the incidences of presence of specimens that exhibit modifications from both hom-
tooth marking for the sub-samples double, and in some cases triple, inins and carnivores demonstrates conclusively that hominins and
carnivores fed from the same animals and were in direct compe- Cut marks also show hominins exploited a wide range of taxa at
tition for carcass foods at the site. In one case from the Lemuta the site, including large fauna such as elephants (Fig. 5A), giraffes
interval, a cut mark overlies a tooth mark, showing secondary ac- (5D), and hippos (Fig. 5F), and at least one carnivore (Fig. 5B). The
cess by hominins to this element, a size group 1e2 femur (Fig. 4C). diversification of hominin diets and exploitation of megafauna has
Notably, this is one of the elements consumed first by carnivores been previously observed at sites within Olduvai Gorge but, until
due to its high utility (Blumenschine, 1986). While no other spec- now, had been observed only at sites that are younger and associ-
imens exhibit marks that directly overlie one another, there are ated with H. erectus (Domínguez-Rodrigo et al., 2014b,c). Our
examples of tooth- and butchery-marked specimens from each analysis of the fossil assemblage from the Acheulean EF-HR site also
limb bone element in the Lemuta and from the humerus, radius, indicated H. erectus exploited megafauna there (de la Torre et al.,
tibia and metatarsal in the LAS. submitted-b). The results for HWK EE show hominins were get-
Cut mark patterns on limb bones in the HWK EE assemblage ting more regular access to a variety of fauna before the Acheulean
suggest that hominins were probably skilled butchers and may have appears at Olduvai. With regards to megafauna, our interpretation
had a somewhat standardized method of butchery. Cut marks are of this evidence is consistent with those of previous studies
located around muscle attachments on limb bone near-epiphyses (Domínguez-Rodrigo et al., 2014b,c), suggesting that acquisition
and midshafts and, in some cases, overlap in location on specimens and consumption of these massive animals was likely opportunistic
from separate animals (Fig. 5C). Clusters of cut marks are often sub- and through scavenging.
parallel and obliquely oriented, suggesting control over the tool and The incidences of tooth and percussion notching for the HWK EE
perhaps defleshing skill. These butchery patterns are inconsistent assemblage correspond with the bone surface modification results.
with what we have observed from inexperienced students, who often The incidence of tooth notching is higher than predicted by the
leave randomly oriented and widely dispersed cut marks, but we HeC model for both intervals, suggesting carnivores were breaking
have yet to subject the pattern observed for HWK EE to a detailed more limb bones than was simulated by the models. This is also
spatial comparison to the cut mark patterns left by expert and inex- supported by tooth mark frequencies that exceed the HeC model.
perienced butchers. Others have also noted that butcher experience The incidence of percussion notching for both intervals falls below
may affect cut mark morphology and frequency (Haynes, 1991; that predicted by the HeC model, indicating hominins were not the
Domínguez-Rodrigo, 1997a,b), but further actualistic work is neces- only agents responsible for breakage of the limb bones at the site.
sary to quantify these differences. As such, these interpretations are This is supported by the relatively low percussion mark frequencies
hypothetical and tentative, but nonetheless intriguing. Skilled for some sub-samples. These patterns suggest the possibility that
butchery behavior at 1.7 Ma should not be surprising given that cut- the HWK EE assemblage includes CO and HeC components, or they
marked bones appear in the archaeological record possibly as early as could suggest hominins may not have broken all of the limb bones
3.3 Ma (McPherron et al., 2010), with hominins regularly butchering that they defleshed, implicating a WBeC component where limb
animals by 2.0 Ma (Ferraro et al., 2013). bones were defleshed by hominins but not broken for marrow
The distribution of cut marks on flesh-bearing limb bones consumption. As noted above, hominins may have been leaving the
suggests hominins usually had access to at least some flesh on the fat-depleted bones of nutritionally-stressed animals unbroken.
humerus, femur, tibia and radius. It was shown that the FLK Zin-
janthropus hominins more often had access to flesh on the hu- 5. Conclusions
merus than the femur, as indicated by a significantly higher
percentage of cut-marked humerus fragments when compared The HWK EE site includes large and well-preserved fossil as-
with femur fragments (Pante et al., 2012, 2015). This was inter- semblages from two stratigraphic intervals deposited just prior to
preted as an indication that hominins acquired access to carcasses the first appearance of Acheulean technology at Olduvai Gorge. The
after the femur had been defleshed by carnivores, as it ranks site was attractive to both hominins and carnivores, with croco-
higher in Blumenschine's (1986) carcass consumption sequence. diles, hyaenids and felids represented in the fossil assemblage (Bibi
However, this is not the pattern observed for HWK EE, despite et al., submitted; de la Torre et al., submitted-a) in addition to an
other indicators of scavenging such as high tooth mark fre- abundance of feeding traces from both hominins and carnivores.
quencies on limb bones. The relatively even distribution of cut Deposited just prior to the appearance of H. erectus at Olduvai and
marks across flesh-bearing limb bones suggests hominins ob- preserving Oldowan technology, HWK EE may represent one of the
tained earlier access to carcasses than at the earlier FLK Zinjan- last H. habilis sites in the region. This implicates the site as an
thropus site. Tooth mark frequencies that are lower than those invaluable reference point for our understanding of the feeding
reported for the FLK Zinjanthropus site (60.7% of all limb bone behavior and ecology of the species as it continued to encroach
fragments, see Pante et al., 2012 for further details) support this upon the larger carnivore guild by consuming flesh and marrow
interpretation. However, sample sizes of midshaft fragments that from large mammals.
were attributable to a particular skeletal part are low for all but The HWK EE hominins regularly obtained access to flesh and
the size group 1e2 sub-sample of the Lemuta interval, suggesting marrow from all limb bones and occasionally from axial elements.
these results should be considered with caution. Further, the They exploited a wide range of taxa, including carnivores and fauna
slight deficiency of femora and humeri observed in skeletal part much larger than themselves, such as elephants, hippos and gi-
profiles from the assemblage suggests the possibility that homi- raffes, but they more regularly obtained access to bovids and less so
nins did not transport these flesh- and marrow-rich limb bones to to equids and suids. Hominin and carnivore feeding behaviors
the site if they had previously been defleshed by carnivores. This appear to have varied with animal size and by stratigraphic interval.
would result in a more even distribution of cut marks across low- It is likely that hominins had earlier access to size 1e2 carcasses
and high-ranking limb bones. than size 3e4 carcasses in both intervals given the relatively even
Cut marks outside of the appendicular skeleton show that distribution of cut marks on size 1e2 limb bones and tooth mark
hominins at least occasionally had access to flesh from all regions of frequencies that may have been inflated by CO components of the
carcasses (Fig. 5E), but this was not necessarily a regular occurrence assemblages. Lower tooth and percussion mark frequencies and
as the majority of cut marks are found on limb bones. Still, it shows higher epiphysis to shaft fragment ratios in the LAS suggest that
that hominins could have had access to substantial amounts of flesh hominins and carnivores did not fully exploit bone marrow and
and marrow from carcasses at the site. grease that may have been acquired from nutritionally-stressed
animals that died during a dry period at Olduvai. The diversity of References
the fauna that hominins exploited at the site suggests opportunistic
acquisition of carcasses without specialization for a particular prey Aiello, L.C., Wheeler, P., 1995. The expensive-tissue hypothesis: the brain and the
digestive system in human and primate evolution. Curr. Anthropol. 36,
size or type, a behavior that is inconsistent with what we know of 199e221.
modern predators that often specialize in a particular prey size, Anto n, S.C., 2003. Natural history of Homo erectus. Yearb. Phys. Anthropol. 46,
likely due to biomechanical constraints (see Schaller, 1972; Eaton, 126e170.
Anto n, S.C., Swisher, C.C., 2004. Early dispersals of Homo from Africa. A. Rev.
1974; Bailey, 1993; Slater et al., 2009). Whether this suggests Anthropol. 33, 271e296.
hominins scavenged much of this food or that Oldowan tools Anto n, S.C., Leonard, W.R., Robertson, M., 2002. An ecomorphological model of the
allowed them to kill prey much larger than themselves is debatable. initial hominid dispersal from Africa. J. Hum. Evol. 43, 773e785.
Anto n, S.C., Potts, R., Aiello, L.C., 2014. Evolution of early Homo: An integrated
The HWK EE hominins appear to have obtained earlier access to biological perspective. Science 345, 45e58.
high-ranking carcass parts, such as axial elements and the flesh and Bailey, T.N., 1993. The African Leopard: Ecology and Behavior of a Solitary Felid.
marrow rich femur, than their FLK Zinjanthropus counterparts. Columbia University Press, New York.
Bartram, L.E., Marean, C.W., 1999. Explaining the “Klasies pattern”: Kua ethno-
Whether this was made possible through occasional hunting or a
archaeology, the Die Kelders Middle Stone Age archaeofauna, long bone frag-
more aggressive form of scavenging (see O'Connell et al., 2002) mentation and carnivore ravaging. J. Archaeol. Sci. 26, 9e29.
cannot be ascertained from the available evidence. However, as Behrensmeyer, A.K., 1975. The taphonomy and paleoecology of Plio-Pleistocene
vertebrate assemblages east of Lake Rudolf, Kenya. Bull. Mus. Comp. Zool.
noted above, the broader patterns of feeding traces at HWK EE are
146, 473e578.
most consistent with hominin scavenging. The stone tools associ- Bibi, F., Pante, M.C., Boisserie, J., Werdelin, L., Fortelius, M., Hlusko, L., Njau, J., Stewart,
ated with the HWK EE hominins do not reveal evidence for K., Souron, A., Varela, S., de la Torre, I., submitted. Large mammals and fish from
advanced weaponry, and even if they possessed wooden spears Middle Bed II, Olduvai, and the Paleoecology of the Serengeti. J. Hum. Evol.
Binford, L.R., 1981. Bones Ancient Men and Modern Myths. New York Academic
that are not typically preserved in the archaeological record, Press, New York.
hunting would likely have required them to be dangerously close to Binford, L.R., 1986. Comment on “Systematic butchery by Plio-Pleistocene hominids
kill their prey. Hunting of smaller size group 1e2 prey was likely at Olduvai Gorge.” Curr. Anthropol. 27, 444e446.
Binford, L.R., 1988. Fact and fiction about the Zinjanthropus floor: data, arguments,
less of a risk for hominins, although the rewards would not have and interpretations. Curr. Anthropol. 29, 123e135.
been as great. Feeding traces at HWK EE suggest hominins obtained Blumenschine, R.J., 1986. Carcass consumption sequences and the archaeological
earlier access to small prey than large prey during both intervals. distinction of scavenging and hunting. J. Hum. Evol. 15, 639e659.
Blumenschine, R.J., 1988. An experimental model of the timing of hominid and
Regardless of whether they hunted or scavenged their food, the carnivore influence on archaeological bone assemblages. J. Archaeol. Sci. 15,
evidence here indicates that the hominins at HWK EE were likely 483e502.
skilled butchers. This behavior could have been selected for as Blumenschine, R.J., 1995. Percussion marks, tooth marks, and experimental de-
terminations of the timing of hominid and carnivore access to long bones at FLK
climate became increasingly unstable, resulting in unpredictable
Zinjanthropus, Olduvai Gorge, Tanzania. J. Hum. Evol. 29, 21e51.
n et al., 2014).
resource availability (Anto Blumenschine, R.J., Marean, C.W., 1993. A carnivore's view of archaeological bone
As zooarchaeologists continue to expand upon the number of assemblages. In: Hudson, J. (Ed.), From Bones to Behavior: Ethno-archaeological
and Experimental Contributions to the Interpretation of Zooarchaeological
fossil assemblages that characterize the feeding behavior of our
Remains. Center for Archaeological Investigations, University of Southern Illi-
ancestors, our understanding of the importance of carnivory to the nois, Carbondale, pp. 273e300.
evolution of the genus Homo has become more apparent. It is clear Blumenschine, R.J., Pobiner, B.L., 2007. Zooarchaeology and the ecology of Oldowan
that early Homo relied on animal foods for what was likely a sub- hominin carnivory. In: Ungar, P. (Ed.), Early Hominin Diets: The Known, the
Unknown, and the Unknowable. Oxford University Press, Oxford, pp. 167e190.
stantial portion of its caloric intake, and that this behavior may Blumenschine, R.J., Selvaggio, M.M., 1988. Percussion marks on bone surfaces as a
have allowed it to thrive during the climatic instability of the new diagnostic of hominid behaviour. Nature 333, 763e765.
Pleistocene. HWK EE may represent the beginnings of a diversifying Blumenschine, R.J., Marean, C.W., Capaldo, S.D., 1996. Blind tests of inter-analyst
correspondence and accuracy in the identification of cut marks, percussion
carnivorous diet and a continuation of an increased reliance on marks, and carnivore tooth marks on bone surfaces. J. Archaeol. Sci. 23,
flesh and marrow that persisted with H. erectus at Olduvai Gorge. 493e507.
Blumenschine, R.J., Prassack, K.A., Kreger, D., Pante, M.C., 2007. Carnivore tooth marks,
microbial bioerosion and the invalidation of Dominguez-Rodrigo and Barba's
Acknowledgements (2006) test of Oldowan hominin scavenging behavior. J. Hum. Evol. 53, 420e426.
Bunn, H.T., 1981. Archaeological evidence for meat-eating by Plio-Pleistocene
We would like to thank the Tanzanian Antiquities Department, hominids from Koobi Fora and Olduvai Gorge. Nature 291, 574e577.
Bunn, H.T., 1982. Meat-eating and human evolution: studies of the diet and sub-
Commission for Science and Technology and Ngorongoro Conser-
sistence patterns of Plio-Pleistocene hominids. Ph.D. Dissertation, University of
vation Area Authority for research permits, and Mr. John Paresso California, Berkeley.
and Olduvai staff for their cooperation. We also thank Briana Bunn, H.T., Ezzo, J.A., 1993. Hunting and scavenging by Plio-Pleistocene hominids:
nutritional constraints, archaeological patterns, and behavioural implications.
Pobiner, Jason Lewis, and an anonymous reviewer for their com-
J. Archaeol. Sci. 20, 365e398.
ments, which served to improve this manuscript. The research was Bunn, H.T., Kroll, E.M., 1986. Systematic butchery by Plio-Pleistocene hominids at
funded by the European Research Council (European Union's Sev- Olduvai Gorge, Tanzania. Curr. Anthropol. 27, 431e452.
enth Framework Programme (FP7/2007e2013)/ERC grant agree- Bunn, H.T., Kroll, E.M., 1988. Fact and fiction about the Zinjanthropus floor: data,
arguments, and interpretations. Curr. Anthropol. 29, 135e149.
ment n 283366 (ORACEAF), National Science Foundation Capaldo, S.D., 1995. Inferring hominid and carnivore behavior from dual-patterned
Archaeology (reference number BCS-0852292), the British Acad- archaeofaunal assemblages. Ph.D. Dissertation, Rutgers University.
emy Small Research Grant (reference number SG121630) and the Capaldo, S.D., 1997. Experimental determinations of carcass processing by Plio-
Pleistocene hominids and carnivores at FLK 22 (Zinjanthropus), Olduvai Gorge,
Wenner-Gren Foundation (Gr. 9245). We wish to acknowledge the Tanzania. J. Hum. Evol. 33, 555e597.
work done by archaeologist Adrian Arroyo and conservators Eli Capaldo, S.D., 1998. Simulating the formation of dual-patterned archaeofaunal as-
Diaz Pila and Rebecca Bennett, and by numerous UCL conservation semblages with experimental control samples. J. Archaeol. Sci. 25, 311e330.
Capaldo, S.D., Blumenschine, R.J., 1994. A quantitative diagnosis of notches made by
students in the last four years, especially Eri Ohara-Anderson for hammerstone percussion and carnivore gnawing on bovid long bones. Am.
her dedication to the project. MCP would like to thank Connie Antiq. 59, 724e748.
Fellmann for comments on a draft of this manuscript. de la Torre, I., Albert, R.M., Arroyo, A., Macphail, R., McHenry, L., Mora, R., Njau, J.K.,
Pante, M.C., Rivera Rondon, C., Rodriguez-Cintas, A., Stanistreet, I., Stollhofen, H.,
Wehr, K. New excavations at the HWK EE site: archaeology, palaeo-
Supplementary Online Material environment and site formation processes in the basal part of Middle Bed II
(Olduvai Gorge, Tanzania). J. Hum. Evol. submitted-a.
de la Torre, I., Albert, R.M., Macphail, R., McHenry, L., Pante, M.C., Rodriguez-
Supplementary Online Material related to this article can be
Cintas, A., Stanistreet, I., Stollhofen, H. The contexts and early Acheulean
found at http://dx.doi.org/10.1016/j.jhevol.2017.06.005.
archaeology of the EF-HR paleo-landscape (Olduvai Gorge, Tanzania). J. Hum. Njau, J.K., Blumenschine, R.J., 2006. A diagnosis of crocodile feeding traces on larger
Evol. submitted-b. mammal bone, with fossil examples from the Plio-Pleistocene Olduvai Basin,
Dominguez-Rodrigo, M., 1997a. Testing meat-eating in early hominids: an analysis Tanzania. J. Hum. Evol. 50, 142e162.
of butchery marks on defleshed carcasses. Hum. Evol. 12, 169e182. O'Connell, J.F., Hawkes, K., Lupo, K.D., Blurton Jones, N.G., 2002. Male strategies and
Dominguez-Rodrigo, M., 1997b. Meat-eating by early hominids at the FLK 22 Zin- Plio-Pleistocene archaeology. J. Hum. Evol. 43, 831e872.
janthropus site, Olduvai Gorge (Tanzania): an experimental approach using cut- Oliver, J.S., 1994. Estimates of hominid and carnivore involvement in the FLK Zin-
mark data. J. Hum. Evol. 33, 669e690. janthropus fossil assemblage: some socioecological implications. J. Hum. Evol.
Dominguez-Rodrigo, M., Barba, R., 2006. New estimates of tooth mark and per- 27, 267e294.
cussion mark frequencies at the FLK Zinj site: the carnivore-hominid-carnivore Pante, M.C., 2013. The larger mammal fossil assemblage from JK2, Bed III, Olduvai
hypothesis falsified. J. Hum. Evol. 50, 170e194. Gorge, Tanzania: Implications for the feeding behavior of Homo erectus. J. Hum.
Dominguez-Rodrigo, M., Mabulla, A., Bunn, H.T., Barba, R., Diez-Martin, F., Evol. 64, 68e82.
Egeland, C.P., Esilez, E., Egeland, A., Yravedra, J., Sachez, P., 2009a. Unraveling Pante, M.C., Blumenschine, R.J., 2010. Fluvial transport of bovid long bones fragmented
hominin behavior at another anthropogenic site from Olduvai Gorge by the feeding activities of hominins and carnivores. J. Archaeol. Sci. 3, 846e854.
(Tanzania): new archaeological and taphonomic research at BK, Upper Bed II. Pante, M., de la Torre, I., submitted. A hidden treasure of the Lower Pleistocene: the
J. Hum. Evol. 57, 260e283. Leakey HWK-EE assemblage. J. Hum. Evol.
Dominguez-Rodrigo, M., de Juana, S., Gal an, A.B., 2009b. A new protocol to differ- Pante, M.C., Blumenschine, R.J., Capaldo, S.D., Scott, R.S., 2012. Validation of bone
entiate trampling marks from butchery cut marks. J. Archaeol. Sci. 36, 2643e2654. surface modification models for inferring fossil hominin and carnivore feeding
Dominguez-Rodrigo, M., Bunn, H.T., Yravedra, J., 2014a. A critical re-evaluation of bone interactions, with reapplication to FLK 22, Olduvai Gorge, Tanzania. J. Hum.
surface modification models for inferring fossil hominin and carnivore in- Evol. 63, 395e407.
teractions through a multivariate approach: application to the FLK Zinj archae- Pante, M.C., Scott, R.S., Blumenschine, R.J., Capaldo, S.D., 2015. Revalidation of bone
ofaunal assemblage (Olduvai Gorge, Tanzania). Quatern. Int. 322e323, 32e43. surface modification models for inferring fossil hominin and carnivore feeding
Domínguez-Rodrigo, M., Bunn, H.T., Mabulla, A., Baquedano, E., Uribelarrea, D., Pe rez- interactions. Quatern. Int. 355, 164e168.
Gonz alez, A., Gidna, A., Yravedra, J., Diez-Martin, F., Egeland, C.P., Barba, R., Pante, M.C., Muttart, M., Keevil, T., Blumenschine, R.J., Njau, J.K., Merritt, S.M., 2017.
Arriaza, M.C., Organista, E., 2014b. On meat eating and human evolution: a A new high-resolution 3-D quantitative method for identifying bone surface
taphonomic analysis of BK4b, (Upper Bed II, Olduvai Gorge, Tanzania) and its modifications with implications for the Early Stone Age archaeological record.
bearing on hominin megafaunal consumption. Quatern. Int. 322e323, 129e152. J. Hum. Evol. 102, 1e11.
Domínguez-Rodrigo, M., Diez-Martín, F., Yravedra, J., Barba, R., Bunn, H.T., Parkinson, J.A., Plummer, T., Hartstone-Rose, A., 2015. Characterizing felid tooth
Mabulla, A., Baquedano, E., Uribelarrea, D., Sanchez, P., 2014c. A taphonomic and marking and gross bone damage patterns using GIS image analysis: An
́
paleoecological study of the faunal assemblage of the main site at SHK (Bed II, experimental feeding study with large felids. J. Hum. Evol. 80, 114e134.
Olduvai Gorge, Tanzania). Quatern. Int. 322e323, 107e128. Perkins, D., Daly, P., 1968. A hunter's village in Neolithic Turkey. Sci. Amer. 219,
́
Eaton, R.L., 1974. The Cheetah: The Biology, Ecology, and Behavior of an Endangered 96e106.
Species. Van Nostrand Reinhold, New York. Peters, R.F., Diaz Pila, E., Ohara-Anderson, E., 2017. Bridge over Olduvai Gorge: ca-
Ferraro, J.V., Plummer, T.W., Pobiner, B.L., Oliver, J.S., Bishop, L.C., Braun, D.R., pacity building and conservation of lithics and fossils. In: Brigdland, J. (Ed.), 18th
́
Ditchfield, P.W., Seaman III, J.W., Binetti, K.M., Seaman Jr., J.W., Hertel, F., ICOM-CC Triennial Conference, Linking Past and Future, 4e8 September 2017,
Potts, R., 2013. Earliest archaeological evidence of persistent hominin carnivory. Copenhagen. International Council of Museums, Paris.
PLOS ONE 8, e62174. Pickering, T.R., Dominguez-Rodrigo, M., Egeland, C., Brain, C.K., 2004. New data and
́ ́
Foley, R.A.F., 2001. The evolutionary consequences of increased carnivory in ideas on the foraging behaviour of Early Stone Age hominins at Swartkrans
́ ́
humans. In: Stanford, C.B., Bunn, H.T. (Eds.), Meat-eating and Human Evolution. Cave, South Africa. S. Afr. J. Sci. 100, 215e219.
Oxford University Press, Oxford, pp. 305e331. Pobiner, B.L., Roger, M.J., Monahan, C.M., Harris, J.W.K., 2008. New evidence for
Foley, R.A.F., 2002. Adaptive radiations and dispersals in hominin evolutionary hominin carcass processing strategies at 1.5 Ma, Koobi Fora, Kenya. J. Hum. Evol.
ecology. Evol. Anthropol. Suppl. 1, 32e37. 55, 103e130.
́
Haynes, G., 1991. Mammoths, Mastodonts and Elephants. Cambridge University Potts, R., Shipman, P., 1981. Cutmarks made by stone tools on bones from Olduvai
Press, Cambridge. Gorge, Tanzania. Nature 291, 577e580.
́
Jorayev, G., Wehr, K., Benito-Calvo, A., Njau, J.K., de la Torre, I., 2016. Imaging and Prassack, K., Pante, M.C., Njau, J.K., de la Torre, I., submitted. The paleoecology of
photogrammetry models of Olduvai Gorge (Tanzania) by unmanned aerial ve- fossil birds from Middle Bed II, at Olduvai Gorge, Tanzania. J. Hum. Evol.
hicles: A high-resolution digital database for research and conservation of Early Rivals, F. Uno, K., Bibi, F., Pante, M.C., Njau, J.K., de la Torre, I., submitted. Dietary
Stone Age sites. J. Archaeol. Sci. 75, 40e56. traits of the ungulates from the HWK EE site at Olduvai Gorge (Tanzania):
Klein, R.G., 1989. Why does skeletal part representation differ between smaller and Diachronic changes and seasonality. J. Hum. Evol.
larger bovids at Klasies River Mouth and other archaeological sites? J. Archaeol. Rowe, S., Rozeik, C., 2008. The uses of cyclododecane in conservation. Rev. Cons. 9,
Sci. 6, 363e381. 17e31.
Marean, C.W., Frey, C., 1997. Animal bones from caves to cities: reverse utility curves Ruff, C.B., Walker, A., 1993. Body size and body shape. In: Walker, A., Leakey, R.
as methodological artifacts. Am. Antiq. 62, 698e716. (Eds.), The Nariokotome Homo erectus Skeleton. Harvard University Press,
Marean, C.W., Kim, S.Y., 1998. Mousterian large-mammal remains from Kobeh Cave: Cambridge, pp. 234e265.
Behavioral implications for Neanderthals and early modern humans. Curr. Schaller, G.B., 1972. Predators of the Serengeti: the social carnivore. Nat. Hist. 81,
Anthropol. 39, S79eS114. 38e49.
Marean, C.W., Spencer, L., 1991. Impact of carnivore ravaging on zooarchaeological Selvaggio, M.M., 1994. Evidence from carnivore tooth marks and stone-tool-
measures of element abundance. Am. Antiq. 56, 645e658. butchery marks for scavenging by hominids at FLK Zinjanthropus, Olduvai
McHenry, L., submitted. Tephrochemistry of Bed II and above, Olduvai Gorge, Gorge, Tanzania. Ph.D. Dissertation, Rutgers University.
Tanzania, and the chronology of the Oldowan-Acheulean transition. J. Hum. Selvaggio, M.M., 1998. Evidence for a three-stage sequence of hominid and carni-
Evol. vore involvement with long bones at FLK Zinjanthropus, Olduvai Gorge,
McNaughton, S., Georgiadis, N., 1986. Ecology of African grazing and browsing Tanzania. J. Archaeol. Sci. 25, 191e202.
́
mammals. Annu. Rev. Ecol. Syst. 17, 39e66. Shipman, P., Walker, A., 1989. The costs of becoming of a predator. J. Hum. Evol. 18,
McPherron, S.P., Alemseged, Z., Marean, C.W., Wynn, J.G., Reed, D., Geraads, D., 373e392.
Bobe, R., Bearat, H., 2010. Evidence for stone-tool-assisted consumption of animal Sinclair, A.R.E., Duncan, P., 1972. Indices of condition in tropical ruminants. E. Afr.
tissues before 3.39 million years ago at Dikika. Ethiopia. Nature 466, 857e860. Wildl. J. 10, 143e149.
Metcalfe, D., Jones, K.T., 1988. A reconsideration of animal body-part utility indices. Slater, G.J., Dumont, E.R., Van Valkenburgh, B., 2009. Implications of predatory
Am. Antiq. 53, 486e504. specialization for cranial form and function in canids. J. Zool. 278, 181e188.
Milton, K., 1987. Primate diets and gut morphology: implications for hominid Speth, J., 1989. Early hominid hunting and scavenging: the role of meat as an energy
evolution. In: Harris, M., Ross, E.B. (Eds.), Food and Evolution Toward a Theory source. J. Hum. Evol. 18, 329e343.
of Human Food Habits. Temple University Press, Philadelphia, pp. 93e115. Uno, K., Rivals, F., Bibi, F. Pante, M., Njau, J., de la Torre, I., Large mammal diets and
Monahan, C.M., 1996. New zooarchaeological data from Bed II, Olduvai Gorge, paleoecology across the Oldowan-Acheulean transition at Olduvai Gorge,
Tanzania: implications for hominid behavior in the Early Pleistocene. J. Hum. Tanzania from stable isotope and tooth wear analyses. J. Hum. Evol. submitted.
Evol. 31, 93e128. Voorhies, M., 1969. Taphonomy and Population Dynamics of an Early Pliocene
Njau, J.K., 2012. Crocodile Predation and Hominin Evolution: Landscape Paleoan- Vertebrate Fauna, Knox County, Nebraska. Contributions to Geology, Special
thropology at Olduvai Gorge. LAP GmbH and Co. KG, Saarbrucken, Germany. Paper No 1. University of Wyoming.

pante-michael-c-the-carnivorous-feeding-behavior-of

Uploaded by

Copyright:

Available Formats

pante-michael-c-the-carnivorous-feeding-behavior-of

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

pante-michael-c-the-carnivorous-feeding-behavior-of

Uploaded by

Copyright:

Available Formats

Journal of Human Evolution xxx (2017) 1e21

Contents lists available at ScienceDirect

Journal of Human Evolution

1. Introduction morphological changes observed in the genus Homo (Milton, 1987;

capillary tubes, micro spatulas or brushes, while gap-ﬁlls were Table 2

Journal of Human Evolution (2017), http://dx.doi.org/10.1016/j.jhevol.2017.06.005

Skeletal part LAS Lemuta

bone surface modiﬁcations at HWK EE are based on the compara-

ments’ category includes epiphyses, near-epiphyses, and midshaft

podials. The proportions of bones with modiﬁcations in the LAS and

when they do not.

3.4. Percussion marks

The proportion of percussion-marked bones is higher in the

assemblages compare similarly to the interquantile ranges of the

interquantile ranges of all sub-samples of the VeHeC model,

the HO model. The incidences of percussion marking for the

comparative sub-samples from the Lemuta interval fall within the

95% interquantile ranges of all sub-samples of the VeHeC model,

midshaft fragments of the HO model. It falls just outside the upper

range of the size group 3e4 sub-sample for midshaft fragments of

sub-sample for midshaft fragments of the HO model.

3.5. Mammalian carnivore tooth marks

The proportion of tooth-marked bones is higher in the Lemuta

trace models and are summarized together. The incidences of tooth

interquantile range of the size group 3e4 sub-sample for midshaft

fragments of the WBeC model. In general, the proportions of tooth-

Size groups based on Bunn, 1982.

3.6. Cut marks

The proportion of cut-marked bones is higher in the Lemuta

fragments (Table 6, Fig. 8). However, the assemblages again fall

within the same interquantile ranges of the feeding trace models

all limb bone fragment sub-samples of the LAS and Lemuta

nomically complex due to a ﬂuvial component to its deposition (de

la Torre et al., submitted-a), but it is also rich in the feeding traces of

3.2 both hominins and carnivores. The assemblage offers evidence of

presence of carnivore remains and high carnivore tooth mark fre-

quencies, that the environment was highly competitive. Beyond

this competition for faunal resources, hominins likely had a high

The behavior of hominins at the site appears to have varied be-

tween the two main stratigraphic intervals represented, and this

will be the focus of the interpretations presented below.

4.1. Effect of ﬂuvial processes

Evidence for ﬂuvial activity at HWK EE is clear in the stratigraphy

such as rounding of lithic specimens, specimen length distributions,

and orientation of elongated specimens show only minimal evi-

dence of ﬂuvial transport and abrasion of the material (de la Torre

et al., submitted-a). The length distributions of midshaft fragments

in the comparative sample (Fig. 2) also suggest hydraulic processes

length distributions of the feeding trace models, none of which was

exposed to ﬂuvial processes. Low-energy ﬂuvial processes can

transport bone fragments differentially, with smaller fragments

more likely to be transported than their larger counterparts (Pante

and Blumenschine, 2010). The similarity in length distributions

based on comparisons with the models.

4.2. Transport of skeletal parts to HWK EE

Skeletal part proﬁles from HWK EE show that high-density limb

cranial, and compact bones (podials) are also well represented.

away from the site by ﬂuvial processes (see Voorhies, 1969;

Behrensmeyer, 1975) or completely consumed by carnivores. The

usually only represented by part of its proximal end or a portion of