The paleoecology and taphonomy of AMK (Bed I, Olduvai Gorge) and its contributions to the understanding of the “Zinj” paleolandscape

Accepted Manuscript The paleoecology and taphonomy of AMK (Bed I, Olduvai Gorge) and its contributions to the understanding of the “Zinj” paleolandscape

Julia Aramendi, David Uribelarrea, Mari Carmen Arriaza, Héctor Arráiz, Doris Barboni, José Yravedra, María Cruz Ortega, Agness Gidna, Audax Mabulla, Enrique Baquedano, Manuel DomínguezRodrigo PII: DOI: Reference:

S0031-0182(16)30811-2 doi: 10.1016/j.palaeo.2017.02.036 PALAEO 8220

To appear in:

Palaeogeography, Palaeoclimatology, Palaeoecology

Received date: Revised date: Accepted date:

2 December 2016 24 February 2017 24 February 2017

Please cite this article as: Julia Aramendi, David Uribelarrea, Mari Carmen Arriaza, Héctor Arráiz, Doris Barboni, José Yravedra, María Cruz Ortega, Agness Gidna, Audax Mabulla, Enrique Baquedano, Manuel Domínguez-Rodrigo , The paleoecology and taphonomy of AMK (Bed I, Olduvai Gorge) and its contributions to the understanding of the “Zinj” paleolandscape. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Palaeo(2017), doi: 10.1016/ j.palaeo.2017.02.036

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT The paleoecology and taphonomy of AMK (Bed I, Olduvai Gorge) and its contributions to the understanding of the “Zinj” paleolandscape. Julia Aramendia,b*, David Uribelarreac, Mari Carmen Arriazab,d, Héctor Arráiza,e, Doris Barbonie, José Yravedraa,b, María Cruz Ortegaf, Agness Gidnab,g, Audax Mabullag, Enrique Baquedanoh, Manuel Domínguez-Rodrigoa,b

MA

NU

SC RI PT

a. Department of Prehistory, Complutense University, Prof. Aranguren s/n, 28040 Madrid, Spain. b. IDEA (Institute of Evolution in Africa), Covarrubias 36, 28010 Madrid, Spain. c. Departaemnto de Geodinámica Universiadad Complutense, Avda. Complutense, Madrid. d. School of Geography, Archaeology and Environmental Studies, Evolutionary Studies Institute, University of the Witwatersrand, Private Bag 3, 2050 South Africa. e. CEREGE (UMR6635 CNRS/Université Aix-Marseille), BP80, F-13545 Aix-en-Provence cedex 4, France f. Universidad Complutense de Madrid-Instituto Carlos III (UCM- ISCIII), Centro de Investigacion de la Evolucion y Comportamiento Humanos, Madrid, Spain. g. Archaeology Unit, University of Dar es Salaam, P.O. Box 35050, Dar es Salaam, United Republic of Tanzania. h. Museo Arqueológico Regional. Plaza de las Bernardas s/n, 28801 Alcalá de Henares, Madrid, Spain.

AC

CE

PT

ED

* Corresponding author: [email protected]

1

ACCEPTED MANUSCRIPT

Abstract AMK (Amin Mturi Korongo) is a newly discovered site situated under Tuff IC (Bed I, 1.84 Ma). It contains several fossiliferous levels and the top one is situated on the same paleosurface as FLK-Zinj. For the first time this allows sampling the “Zinj” paleoenvironment well into the Secondary Gorge and expands the known area of this

SC RI PT

paleolandscape. Fossils found at this site show exceptional preservation. Several articulated units have been discovered, indicating minimal postdepositional disturbance and rapid sedimentation. This assemblage allows a general estimation of time span (the most elusive variable in archaeological analyses) for the formation of AMK. Phytolith analyses have discovered a dense palm forest at the site, expanding the forested area

NU

known on the slightly elevated platform that contains the FLK-Zinj – FLK-NN – PTK sites. Although a few artifacts have been discovered in the vicinity of AMK, the site was mostly naturally (i.e., non-anthropogenically) formed. This is of major relevance to

MA

determine that factors other than forested habitats must have influenced the formation of

ED

anthropogenic sites on the same platform as AMK in the Olduvai lacustrine basin.

AC

CE

PT

Keywords: natural background scatter; carnivore; FLK-Zinj, PTK; site formation

2

ACCEPTED MANUSCRIPT

1. Introduction Since the discovery of the FLK-Zinj site (Bed I, Olduvai Gorge, Tanzania), several interpretations of it have been presented to understand hominin behavior around 2 Myr ago. This site was first interpreted as a living floor (Leakey, 1971), which according to Isaac (1978) corresponded to the selection of a specific location for

SC RI PT

toolmaking, butchery, and collective food consumption activities by Early Pleistocene hominins. A subsequent analysis of the site by Bunn (1982) led Isaac (1983) to propose a new model (“central-place” foraging) for this type of sites. In this model, hominins appeared as primary agents in the acquisition of animal carcasses. In addition to raw materials, such carcasses were repeatedly brought to selected specific spots on the

NU

landscape to be collectively exploited. This main idea presented in the “central-place” foraging model about carcass transportation has been supported by several studies during the last decades (Bunn and Kroll, 1986, 1988; Potts, 1988; Bunn and Ezzo, 1993;

MA

Oliver, 1994; Blumenschine, 1995; Capaldo, 1995, 1997; Rose and Marshall, 1996; Domínguez-Rodrigo, 1997; Domínguez-Rodrigo and Pickering, 2003). However, the

ED

kind of access that hominins might have had to those transported carcasses has been a source of debate for several years. The accumulation at FLK-Zinj has been interpreted as the result of hominins acting as primary agents hunting and selectively transporting

PT

carcasses (Isaac, 1978, 1983; Bunn, 1982, 1983, 1991; Bunn and Kroll, 1986, 1988; Bunn and Ezzo, 1993; Oliver, 1994; Rose and Marshall, 1996; Domínguez-Rodrigo,

CE

1997; Domínguez-Rodrigo and Pickering, 2003; Bunn and Pickering, 2010), but also as the result of hominins acting as secondary agents, either transporting previously

AC

partially defleshed carcasses (Capaldo, 1995, 1997) or scavenging bones from defleshed felid kills (Blumenschine, 1986, 1991). Most recent taphonomic analyses have provided compelling evidence that hominins must have had primary access to fully-fleshed small and medium-sized carcasses (Domínguez-Rodrigo et al., 2007, 2010b, 2014). In addition to taphonomic evidence, geological and paleoecological studies have been essential for the understanding of the site and the related hominin carcass consumption behavior. The geological sediments around FLK have long been known to have been dominated by lake-margin facies, corresponding to a shallow lake with predominantly

fine-grained

sedimentation

(Hay,

1976).

According

to 3

ACCEPTED MANUSCRIPT paleoenvironmental reconstructions, FLK-Zinj occupied a location on the southern edge of an elevated platform by the lake floodplain characterized by the presence of a mixed landscape with densely wooded areas (Ashley et al., 2010a; Leakey, 1965; Bonnefille, 1984; Plummer and Bishop, 1994; Sikes, 1994; Spencer, 1997; Fernández-Jalvo et al., 1998; Domínguez-Rodrigo et al., 2010). Phytoliths analyses showed that around FLK the vegetation was dominated by palm trees, which would have provided shelter from predation and easy access to water (Ashley et al., 2010a). Such a landscape concurs

SC RI PT

with the taphonomic work carried out on the paleolandscape surrounding FLK-Zinj. The excavation of several trenches as well as the excavation of nearby sites such as FLK-NN, showed that bone deposition in the vicinity of FLK-Zinj was significantly less dense and, further, that little to no evidence for carcass butchering was preserved in the immediate surrounding environment outside the space containing the FLK

NU

archaeofaunal assemblage. This suggests that butchery was repeatedly carried out on the same spot, namely where FLK-Zinj was formed (Domínguez-Rodrigo et al., 2007, 2010b; Uribelarrea et al., 2014). Thus, the site would have provided shelter from

stays by hominins at the site.

MA

predation and easy access to water, both of which would have encouraged prolonged

ED

Only two newly discovered sites (PTK and DS) overlaid by Tuff IC and found on the same paleosurface show similar bone and tool concentration rates to FLK-Zinj.

PT

DS (David’s Site) was discovered in 2014 and is currently under study (see DomínguezRodrigo et al., submitted for further information). PTK (Phillip Tobias Korongo)

CE

preserves a clay stratum similar to that reported at FLK-Zinj, which contains two differentiated archaeological levels (Domínguez-Rodrigo et al., 2010b; Uribelarrea et

AC

al., 2014). The discovery of PTK proves two points: (1) there was no river channel crossing the area where the site is located as proposed by Blumenschine et al., (2012), and (2) there are particular and clearly distinguishable spots on the landscape that were selected by hominins to transport and utilize raw materials and exploit carcasses. Even though the research of the PTK and DS archaeofaunal assemblages is in progress, preliminary data (i.e., abundance of hominin-modified bones) indicate that those sites were formed intentionally by hominins. In the new paleoenvironmental reconstruction carried out by TOPPP (Uribelarrea et al., 2014), the Zinj paleolandscape appeared as a large area subdivided

4

ACCEPTED MANUSCRIPT into two subareas. To the West, there was a raised platform, where the main Zinj complex sites are found (FLK-NN, FLK-N, PTK and FLK-Zinj) and to the East a shallow depression that extended towards the KK fault. In the model presented by Uribelarrea et al. (2014), the Zinj paleolandscape was further subdivided into four units based on depositional energy, topography, and facies. One of these areas was interpreted as a mudflat where the paleolake Olduvai was located and where several rivers discharged forming a well-established drainage network upstream. The extension

SC RI PT

of the lake varied according to the rain and wet seasons. The archaeological sites are located nearby the paleolake in a low energy depositional environment that corresponds to a lacustrine terrace; a relatively flat and elevated area. Higher in elevation than this terrace is the area where the site presented here (AMK) is located. This area corresponds to the supralittoral environmental belt, i.e. the external border of the

NU

lacustrine basin.

AMK (Amin Mturi Korongo) is a newly discovered site situated under Tuff IC

MA

(Bed I, 1.84 Ma) that contains several levels, the top clay level being equivalent to the FLK-Zinj paleosurface (Fig. 1). AMK is situated in the Secondary Gorge, at the eastern Olduvai paleolake alluvial plain (~ 500 m from FLK-Zinj and 200 m away from PTK).

ED

The site was named AMK after the prominent Tanzanian archaeologist, Amin Mturi. A total of 16 m2 were excavated in 2013 and 2014, exposing several geological units and

PT

three fossiliferous levels.

Here we present the taphonomic study of AMK and a more detailed geological

CE

study of the site, including the results of new paleobotanical analyses, which together

AC

will contribute to a better understanding of the Zinj paleolandscape.

2. Methods and Samples 2.1. Geological and paleobotanical reconstruction AMK and its surroundings formed part of a previous paleoenvironmental study focused on reconstructing the northern, eastern, and southern portions of the FLK-Zinj paleolandscape (Uribelarrea et al., 2014). The geological study was based on three different techniques: a) stratigraphy and facies analysis, b) geomorphology, and c) structural tectonics (see Uribelarrea et al., 2014 for methodological issues). In this study 5

ACCEPTED MANUSCRIPT we add a more detailed local geological description based on the analysis of a stratigraphic cross-section under Tuff ID at AMK. The paleobotanical study rests on the analysis of phytoliths. To reconstruct the paleovegetation, several samples were taken along the Zinj landscape below Tuff IC (Arráiz et al., submitted). Two samples were collected in stratigraphic sections from AMK, right below Tuff IC, and were then prepared for phytolith analyses as explained

SC RI PT

in Barboni et al. (2010). Here we show the paleobotanical results.

2.2. Orientation analyses

Orientation patterns of bones were taken into consideration to reinforce the geological and paleobotanical reconstruction of the site and to assess the impact of water flows on the documented faunal arrangement. Vertical and horizontal orientations

NU

were directly taken from each item using compasses and clinometers (Voorhies, 1969; Fiorillo, 1991; Howard, 2007). These measurements were taken along the longitudinal axis that divided the item more or less symmetrically and only in specimens that were at

MA

least twice as long as they are wide. This symmetry axis has been proved to be taphonomically informative since long objects tend to orient according to their longitudinal axes (Toots, 1965; Voorhies, 1969). Orientation data were statistically

ED

treated using the “circular” package in R software (http://www.r-project.org). Three different tests were used to assess orientation patterns. A Kuiper’s test and a Reyleigh’s

PT

test were used to detect isotropy (Fisher, 1995), while a Watson test was performed to detect anisotropy. Statistical tests were supplemented with graphical procedures.

CE

Orientation data were displayed graphically using rose diagrams (ORIANA free software). Orientation analyses were only conducted for Level 1 and 2, since sample

AC

size for Level 3 was too small. Level 1 and 2 samples were bootstrapped 100 times each to ensure reliable results.

2.3. Taphonomic analysis A total of 198 remains (including hominin and faunal fossils and lithics) were recovered at AMK. Stone tools and hominin bones were not included in our analysis, so in the end 183 bone remains were analyzed from a taphonomic perspective. Most fossils (>58%) presented a good cortical surface, some of them being exceptionally well

6

ACCEPTED MANUSCRIPT preserved. Only around 17% of the remains were too deteriorated to be able to analyze bone surface modifications, but were included when possible to complete the skeletal part profiles and the study of taxa. Bad preservation was mainly due to alterations related to soil humidity. Carbonates, mineralization and biochemical alterations related to bacteria are some of the most common non-biotic processes detected at AMK. A significant portion of the fossils discovered at AMK underwent some conservation treatments in order to ensure the preservation and enable this study. Intervention

SC RI PT

strategies on the fossils undergo different stages and always follow a criterion of minimum intervention. Our main objective for treatments applied at sites during excavation is the stabilization of the fossils. The treatments applied in the field include preventative gauze coatings and/or the application of adhesive to surface areas; isolated adhesion of fragments to avoid their dispersion and/or loss; consolidation to maintain

NU

cohesion and mechanical resistance; and the extraction with a packaging composed of tissue paper and aluminum foil as a kind of frame. Treatments carried out in the laboratory include unpacking, cleaning (removal of sediments and carbonates

MA

mechanically at the tip of scalpel and lathe and surface dirt with water and acetone in cases that are adhered and / or consolidated), consolidation, adhesion of fragments

ED

and/or reconstruction and packaging (Ortega et al, 2016). After having properly cleaned and treated the fossils, the AMK Levels were studied independently. Level 1 includes 98 remains, Level 2 has 70 and Level 3 contains 15 bone remains. Due to sample size,

PT

Level 3 was excluded from the taphonomic analysis. We did not include the

CE

microfaunal remains in this study.

2.3.1. Skeletal part representation

AC

All specimens were identified to element and carcass size whenever possible. In order to analyze skeletal part profiles, carcasses were divided into three anatomical regions: skull (horn, cranium, mandible and teeth), axial (vertebrae, ribs, pelves, and scapulae) and appendicular (limb bones), with limbs further divided into upper (humeri and femora), intermediate (radii and tibiae), and lower (metapodials) units (DomínguezRodrigo, 1997). Pelves and scapulae were grouped with vertebrae and ribs due to their similarity in bone texture and taphonomic properties (Domínguez-Rodrigo et al., 2007). Skeletal profiles are based on the number of identified specimens (NISP) and an estimate of the minimum number of elements (MNE). MNE estimates were based on an

7

ACCEPTED MANUSCRIPT integrative approach that includes the analysis of shafts according to their bone section, muscle insertion sites, cortical bone thickness, cross-sectional shape, and medullary surface (Patou-Mathis, 1985; Munzel, 1988; Barba and Domínguez-Rodrigo, 2005; Yravedra and Domínguez-Rodrigo, 2009). Bone fragments were also identified according to orientation: cranial, caudal, lateral or medial. Carcass size was classified using Bunn’s (1982) methodology: small carcasses include sizes 1 (< 25 kg, e.g. gazelle) and 2 (25-125 kg, e.g. impala), medium-sized

SC RI PT

carcasses are size 3 (125-350 kg, e.g. lesser kudu and zebra) and large carcasses are sizes 4 to 6 (350-1000 kg, e.g. buffalo; 1000-3000 kg, e.g. giraffe; > 3000 kg, e.g. elephant). MNE estimation was then used to calculate the minimum number of individuals (MNI), where criteria such as size, side, age and biometrics, usually used in a comprehensive analysis (Lyman, 1994), were taken into consideration. Age profiles

NU

were established based on unfused elements and dental specimens.

2.3.2. Breakage patterns

MA

Breakage planes were identified as green or dry following Villa and Mahieu (1991) and were analyzed following the methods outlined by Domínguez-Rodrigo et al.

ED

(2007). The appearance of dry breakage patterns is associated with non-biotic processes such as geological forces (i.e., sediment compaction), while green fractures are related to the action of biotic agents (Pickering et al., 2005; Alcántara et al., 2006). Green

PT

breakage planes show smoother surfaces and are more likely to be oriented obliquely to the long axis of the bone. Green breakage can also be subdivided into patterns created

CE

by dynamic (impact using hammer stone) and static (pressure exerted by carnivore gnawing) loading according to the angle between the cortical and the medullary surface

AC

(Pickering et al., 2005; Alcántara et al., 2006). Dry breakage planes have a rough texture characterized by the presence of micro-step fractures and tend to be longitudinal and/or transverse to the long axis of the bone, with a nearly 90º angle between the cortical and the medullary surface. A complementary approach that allows the differentiation between bonebreakage processes (i.e., dynamic versus static loading) involves the study of notches. Experimental studies have shown that different types of notch morphologies can be identified and that their frequencies respond to the action of different agents (Capaldo and Blumenschine, 1994; De Juana and Domínguez-Rodrigo, 2011). The identification

8

ACCEPTED MANUSCRIPT of notches was performed as a proxy in the present study, following De Juana and Domínguez-Rodrigo’s (2011) classification; but no statistical analysis was performed since the sample was too small to obtain reliable results. Bone impact flakes were also examined. Bone impact flakes show a similar morphology to lithic flakes and their frequency varies according to the modifying agent. Bone flakes are much more common in anthropogenic sites, whereas carnivores do not usually produce these kind of fragments, except for the hyenas, which do so in small

SC RI PT

numbers (Pickering and Egeland, 2006; Domínguez-Rodrigo and Martínez-Navarro, 2012).

Shaft breakage patterns were analyzed following Bunn’s (1982) circumference classification, where Type 1 refers to shafts that preserve <50% of the circumference, Type 2 includes shafts with >50% of the section left, and Type 3 are complete

NU

cylinders. AMK circumference types among prey size 3 were compared with those found in the Syokimau hyena den (Egeland et al., 2008) and the Olduvai carnivore site (OCS), which has been proposed as a bone accumulation generated by lions and

MA

scavenged by spotted hyenas (Arriaza et al., 2016). The three sites are dominated by

ED

medium-sized bovids.

2.3.3. Bone surface modifications

Surface macroscopic and microscopic marks on bone surfaces were studied and

PT

analyzed for each level independently. Marks were analyzed with the aid of hand lenses (15 X – 20 X) under strong light (60 W) as outlined by several authors (Blumenschine,

CE

1988, 1995; Blumenschine and Selvaggio 1988, 1991; Bunn, 1981; DomínguezRodrigo et al., 2009). The identification of the different surface modifications (e.g. cut

AC

marks, percussion marks, tooth marks, biochemical marks) was carried out following previous works (Binford, 1981; Bunn, 1981; Shipman and Rose, 1983; Shipman et al., 1984; Blumenschine, 1988; Blumenschine and Selvaggio, 1988; Blumenschine et al., 1996). In the case of AMK, special attention was paid to tooth marks since they were the most frequent modifications. Tooth mark morphology, size, frequency and anatomical location were recorded and analyzed for each level independently. Modifications of the axial skeleton were also carefully observed considering that felids and hyaenids alter vertebrae differently (Domínguez-Rodrigo, 1999; Parkinson et al.,

9

ACCEPTED MANUSCRIPT 2015). Among the four different classes of modifications usually left by carnivores – pits, scores, punctures and furrowing (Bunn, 1981; Binford, 1981; Blumenschine and Marean, 1993; Blumenschine, 1988, 1995) – pits (Andrews and Fernández-Jalvo, 1997; Selvaggio and Wilder, 2001; Domínguez-Rodrigo and Piqueras, 2003; Andrés et al., 2012) and furrowing (Domínguez-Rodrigo et al., 2015) provide useful information in order to determine the kind of carnivore involved. Due to sample size, we only conducted a statistical analysis of pits (breadth and length). Pit dimensions documented

SC RI PT

at AMK were compared to the most complete carnivore reference collection available (Andrés et al., 2012). Only the Maasai Mara hyena den and the lion sample were used for the comparison since from the carnivores analyzed in Andrés et al. (2012) these two species are the most likely agents at AMK due to chronology and location of the site. Pits on medium-sized shafts were selected in order to make AMK sample comparable to

NU

the data available. Tooth pit dimensions were calculated using a two tailed confidence alpha value of 0.025, so that the variation of most of the sample is represented as

MA

explained in Andrés et al. (2012).

2.3.4. Disarticulation patterns and site formation

ED

Due to the fossil spatial pattern detected during excavation, special attention was paid to disarticulation patterns (Hill, 1979; Hill and Behrensmeyer, 1984), bone type and shape for potential transportability by physical processes (Voorhies, 1969; Boaz

PT

and Behrensmeyer, 1976).

Disarticulation patterns provide information about timeline exposure, especially

CE

if coupled with weathering stages (Hill, 1979; Hill and Behrensmeyer, 1984). The presence and/or absence of certain intact joints in any given ungulate skeleton are

AC

indicative of the dismembering process. AMK Level 2 was investigated for intact joints and weathering stages since several complete elements appeared in anatomical connection.

Classic studies in vertebrate taphonomy (Voorhies, 1969) have shown that there are some types of bones more prone to be transported than others. That means, that the predominance of certain bones must be taken into account in order to establish the autochthony or allochthony of a site (Boaz and Behrensmeyer, 1976).

10

ACCEPTED MANUSCRIPT The maximum length of each fragment was recorded to examine bone breakage patterns, size-sorting and possible preservation bias due to postdepositional events, such as the presence of water or secondary agents. Likewise, other macroscopic modifications such as polishing or abrasion (Thompson et al., 2011), biochemical modification (i.e., fungi) (Marchiafava et al., 1974) and weathering (Behrensmeyer, 1978) were also analyzed, since they can contain information about biostratinomic processes and paleoenvironmental conditions. Furthermore, they can also contain

SC RI PT

information of subaerial exposure times.

3. Results

NU

3.1. Paleoenvironmental description

3.1.1. Geology and Stratigraphy

AMK is a Bed I site located in the Olduvai Secondary Gorge, 400 m upstream

MA

from the Main Gorge (Fig. 1). The site is characterized by lake-margin facies and consists of three fossiliferous levels. The upper fossiliferous level, Level 1, lies right

ED

below Tuff IC, dated around 1.832 + 0.003 Ma (Deino, 2012). According to previous paleoecological reconstructions of the FLK-Zinj paleolandscape (Uribelarrea et al., 2014), AMK is situated in the supralittoral area of the lake margin, very close to the

PT

southern limit of the basin. In comparison to other contemporaneous and nearby sites like FLK-Zinj, FLK-N, PTK and DS, the area occupied by AMK is less affected by

CE

lacustrine sedimentation processes and significantly influenced by external sediment

sites.

AC

inputs. As a result, the local stratigraphic sequence is more complex than at the other

The geological contextualization of the three fossiliferous levels excavated at AMK was conducted based on a 1.5 m cross-section under Tuff ID (Fig. 2). From bottom to top, the stratigraphic sequence begins with a tuffaceous waxy-clay layer that corresponds to a lacustrine sedimentation process over a volcanic substratum of very fine-grained sediment (ash). Above this unit, a 50 cm layer of very fine volcanic reworked sands was deposited. These fine sands are pale yellow (2.5Y-8/2, according to the Munsell Soil Color Charts, 1994; revised edition), and are interrupted by a) lenticular layers of coarse sands and some 3-5 cm basalt fragments, and b) by a

11

ACCEPTED MANUSCRIPT discontinuous and channel-like level of coarse grey sands (5Y 6/1). In this unit, load cast, flame structures and bioturbation caused by roots are abundant, and their presence has notably modified the internal structure of the deposit. In the uppermost 10 cm of this unit lies the oldest fossiliferous level exposed at AMK, Level 3. As a whole, this lower unit corresponds to a mid-distal alluvial fan, mainly formed by massive fine sediments, where pebbles have been sporadically detected and surface runoffs would have contributed with coarser sands. The presence of root and load casts indicates high

SC RI PT

standing of the water table at least during certain periods. In this environment, there are also some local depressions that allow the decantation of silt and clay, as is the case of the second geological unit identified at AMK. This second unit consists of a 5 cm layer of overlaying silts and clays that decreases its section to the East of the site and defines the paleosurface of AMK Level 2.

NU

The second fossiliferous level was covered by a fast and low-energy event, clearly not strong enough to transport the fossil remains. Above AMK Level 2, five parallel stacked and fining upwards sequences (3-12 cm) were deposited. The first of

MA

them is characterized by a fine-grained sedimentation, with very fine sands and clayey silt; whereas the upper sequences are formed by coarser sediments. Each sequence

ED

corresponds to a different flooding event that generated diffuse runoffs over the alluvial system. The increase in thickness and grain size along the sequence might be indicating a progradation of the alluvial system to the basin center in the North. According to

PT

previous studies, this stratigraphic position is occupied by a laminar tuffaceous level (the Chapati Tuff) in the other lowermost Bed I sites located between the 4th and the

CE

FLK Fault (Uribelarrea el al. 2014). Above these 5 sequences is the geological unit equivalent to Zinj Level 22. In

AC

the case of AMK, this clay level contains a greater percentage of tuffaceous silt that is intensively laminated, especially in the lower half (Fig. 3). Although the sedimentation process corresponds to a high standing of the lake level, the influence of the setting of AMK in the supralittoral belt is evident in the form of coarser sediment inputs. In contrast to other pene-contemporaneous sites, where the Zinj clay is clearly divided into two levels (Levels 22a and 22b), here several thin clay levels of millimetric thickness have been observed. Thus, while the rest of the sites are characterized by two clearly distinguishable subaerial exposure phases (two paleosurfaces), AMK shows a more continuous sedimentation process, though with a low sedimentation rate. The

12

ACCEPTED MANUSCRIPT fossiliferous Level 1 was formed on top of the clay unit and was directly buried by Tuff IC.

3.1.2 Phytolith results The AMK vegetation was largely controlled by water table oscillations, as previously hypothesized by Uribelarrea et al. (2014). Environmental fluctuations are also reflected in phytolith analyses. The AMK site shows strong differences between

SC RI PT

samples. In one of the samples (DB14-27C), palm phytoliths represent 56%, whereas in the other sample (DB14-29) they represent a mere 2%. However, both samples (Arráiz et al., submitted) indicate a densely wooded area, in which palm trees would have been a regular component, but also other elements (e.g. grasses and sedges) would have been

NU

present forming a mixed vegetation cover (Table 1).

3.2. Orientation analysis

MA

Orientation analyses of fossils for Levels 1 and 2 show a uniform distribution. Both rose diagrams (Fig. 4) indicate that most of the azimuth confidence interval shows

ED

no clearly identifiable anisotropic orientation. Statistical tests support these results, as p values >0.05 indicate that the null hypothesis of uniform distribution cannot be rejected. A Rayleigh test of a general unimodal alternative (Level 1: p = 0.9156; Level 2: p =

PT

0.4706), a Kuiper test of uniformity (Level 1 and Level 2: p > 0.15), and a Watson goodness-of-fit test of circular uniformity (Level 1 and Level 2: p > 010) suggest that

CE

the assemblage is largely autochthonous. Thus, the hydraulic flows detected in the geology of the site were not sufficiently intense to preferentially orient materials and

AC

cause a strong post-depositional disturbance.

3.3. Faunal analysis 3.3.1. Skeletal part profiles The analyzed bone assemblage (Levels 1 and 2) is composed of 168 specimens, representing a significant number of taxa. The majority of bone remains recovered could be attributed to carcass size, especially in Level 2 where fossils are less damaged. Most bones could be also assigned to animal families and in few cases to the species

13

ACCEPTED MANUSCRIPT (Table 2). Level 1 bone assemblage is composed of 98 bone remains. Of these fragments, 56 specimens could be identified to element, representing at least 12 individuals. Carcass size attribution was more difficult due to high fragmentation. However, 18 elements could be attributed to medium-sized bovids, 3 are representing small bovids and only 1 element corresponds to a large carcass individual. The rest of the elements could be attributed to smaller not Bovidae species. The Level 2 bone assemblage contains more than 57 identified specimens, representing different taxa and

SC RI PT

at least 13 individuals. A total of 34 elements (excluding bird and fish remains) could be attributed to carcass size: 30 elements belong to medium-sized carcasses and 4 to smaller individuals.

Among herbivores, most of the identified specimens are bovids, equids being limited to one bone. Medium-sized bovids are the most represented type of carcasses in

NU

both levels. In Level 1, bovid representation is more varied, with some remains that belong to small bovids (sizes 1 and 2). Only few dental remains have enabled a more

MA

precise identification, certifying the presence of a sub-adult Antilopini in Level 1 and a sub-adult Reduncini in Level 2. In both levels small carnivores are represented, but it was not possible to determine the species. Several primates have also been identified in

ED

both levels, including hominins, a small-sized primate and a Theropithecus. Nonmammals do also appear in both levels; at least one bird is represented in each level and

PT

two catfishes were recovered from Level 2, certifying the presence of water at AMK. Regarding anatomical sections, differential preservation could be observed in

CE

Levels 1 and 2 (Fig. 5). Long bones are the best represented elements in Level 1; they make up half of the MNE (56%). Axial elements and compact bones seem to be highly

AC

underrepresented as measured by MNE according to MNI. Axial elements are the most abundant bones in Level 2, but they are still underrepresented according to MNI and MNE estimations. Long bones represent 33% of the MNE. In both levels, cranial bones are poorly represented. The high representation of the appendicular bones (64% in Level 1, 53% in Level 2) suggests that the accumulation is the result of the transport of partial carcasses or of in locus death (kill) with subsequent postdepositional carnivore ravaging (deleting a substantial amount of axial elements). It is interesting that, overall, upper limb bones are better represented than metapodials in Level 2, even though the latter are denser and, thus, tend to preserve better than other long bones. Both humeri and femora are best represented by their distal epiphyses. Axial bones (including pelves, 14

ACCEPTED MANUSCRIPT scapulae, vertebrae and ribs) are more prominent than cranial specimens in both levels. In theory, axial and compact elements tend to disappear in carnivore scavenged assemblages, creating skeletal profiles dominated by cranial and long limb bones that are generally better represented by shaft fragments.

3.3.2. Bone breakage Fragmentation patterns vary at AMK, with higher degrees of breakage in Level

SC RI PT

1 and the preservation of entire elements in Level 2. In Level 1, a total of 44 specimens could be identified as bearing green breakage and 31 showed dry breakage patterns. The assemblage is highly fragmented, and specimens > 4 cm predominate (70%). Breakage is also evident when looking at long bone fracture patterns. Most long bones (>69%) have less than 50% of the circumference left. The rest are complete and no long bone

NU

shows an incomplete shaft with more than 50% of the circumference preserved (Fig. 6). Level 2 includes 48 bones bearing green or/and dry breakage. Of these, 36 specimens exhibit dry breakage and only 12 specimens show green breakage. Breakage is less

MA

intense in Level 2. More than 72% of the specimens are >4cm and the majority of long bones remain almost complete. Only 5 long bones preserve less than 50% of the

ED

circumference and 1 more than 50% (Fig. 6).

Theoretically, in both human and carnivore assemblages, and in the accumulations created by the action of both agents, shaft circumference Type 1 (<50%

PT

of the shaft is preserved) is predominant. However, carnivore assemblages show a greater proportion of Types 2 (>50% of the shaft) and 3 (complete shaft) with a higher

CE

frequency of complete sections (Bunn, 1982, 1983b; Marean et al., 2004). However, such patterns vary according to the carnivore involved in the modification of carcasses.

AC

Felids are known to be less destructive than hyenas. The comparison of the circumference types found at AMK with a hyena den and a lion accumulation (Fig. 6) suggest that each AMK Level is representing a different pattern. While Level 1 results are more similar (though not exactly the same) to those found at Syokimau hyena den (Egeland et al., 2008), with a preponderance of Type 1 followed by complete cylinders; Level 2 shows almost the same pattern observed in the OCS, with most long bones remaining complete, almost a third being highly fractured and a small percentage maintaining more than 50% of the circumference left. According to Arriaza et al. (2016) the OCS assemblage is the result of a large felid (most likely lion) primary access

15

ACCEPTED MANUSCRIPT followed by minimal hyena scavenging. The presence of complete appendicular bones, as observed in many Olduvai Bed I sites (Potts, 1988), has been interpreted as a result of a meat surplus (Egeland and Byerly, 2005). That is, Type 3 would be the result of an environment with little competition among predators with small impact of scavenging strategies. On the other hand, predominance of Type 1 sections would be indicative of intensive marrow exploitation and, therefore, higher competition. When Type 1 is as abundant as 2 and 3 together, or even smaller, it is assumed that the site represents a low

SC RI PT

competition scenario where bone destruction was reduced. This is the case in AMK Level 2 (Fig. 6). Meanwhile, the presence of complete bones along with others that are very fragmented in Level 1 (Fig. 6) suggests different depositional events with different carnivores exploiting carcasses.

The frequencies of dry breakage identified in both levels (see above) indicate

NU

important non-biotic postdepositional modifications. The ratio of green breakage is also noteworthy, especially in Level 1. Green fractures might be interpreted in relation to

MA

other surface modifications such as tooth marks (see below) and notches. The calculation of the ratios of notch Types C/A and D/A has been used to discern between anthropic and carnivore assemblages (Domínguez-Rodrigo et al., 2007). Carnivores

ED

tend to generate more Type C notches, since they have more than one contact point (using the cusps of their teeth). Opposing notches appear when carnivores apply

PT

pressure with the upper and lower dentition at two opposing points (Egeland, 2010; De Juana and Domínguez-Rodrigo, 2011). Six specimens in Level 1 have been identified as

CE

bearing Type C and D notches (Table 3), whereas no notches have been observed in Level 2. Some of the few notches identified at AMK appear associated with pits and

AC

scores, indicating a possible relation. Micro-notches (<1 cm) have also been observed in Level 1. This typology is also characteristic of carnivores (Capaldo and Blumenschine, 1994). A Type C notch that is broader than deep (morphology more typical of dynamic loading) was observed on a metapodial shaft in association with a percussion mark. Usually Type C notches appear associated with percussion marks on denser bones, as a result of multiple impacts made in order to fragment the bone and extract the marrow (De Juana and Domínguez-Rodrigo, 2011). Although the sample of notches is not large enough to conduct reliable statistical analyses (Table 3), the predominance of Type C (double overlapping) and D (double opposing) notches underscores that the breakage

16

ACCEPTED MANUSCRIPT patterns identified in both levels could be associated with static loadings, such as the ones carried out by carnivores. Finally, at AMK no bone impact flakes have been documented. This observation indicates hominids were not a significant bone modifying/accumulating agent.

3.3.3. Bone surface modifications Most bone remains had a good preservation, with cortical surfaces that enabled

SC RI PT

the identification of conspicuous marks. However, the state of preservation varied within each assemblage: some of the bones had biochemical alterations or different stages of weathering and/or trampling and carbonate concretions (Fig. 7). This differential alteration of the cortical surfaces might indicate multiple deposition episodes of variable duration.

NU

Trampling is the most common non-biotic alteration observed in both levels. A total of 31 specimens (31.6%) in Level 1 and 20 in Level 2 (28.6%) show trampling marks. This might be related to the higher amount of coarser sediments at the site as

MA

explained above. Biochemical alterations are also significant in the form of shallow and irregular exfoliations (Fig. 7D). More than 30% (30 specimens) of the remains

ED

recovered from Level 1 and around 27% of the bone fragments (19) from Level 2 show some kind of biochemical alteration related to the presence of root etching of associated fungi and bacteria (Fig. 7B), typical of humid environments (Marchiafava et al., 1974;

PT

Sharmin et al, 2003; Domínguez-Rodrigo et al., 2007). The precipitation of carbonates is important in the assemblage and is also linked to the presence of water. More than

CE

27% (n = 27) of the assemblage in Level 1 and more than 17% (n = 12) in Level 2 are affected by carbonates (Fig. 7A). Manganese alterations (Fig. 7C) have also been

AC

observed, but in a much lower frequency (n = 6 in Level 1, n = 9 in Level 2). Early stages of chemical weathering have been identified among some remains. Traces of subaerial weathering are marginal and refer to stage 1; most of the well-preserved assemblage exhibit weathering stage 0 (Behrensmeyer, 1978). In sum, 59% (99 specimens) of the studied assemblage was well preserved, 24% (40) had a regular preservation and 18% (29) showed bad cortical surfaces. Only specimens that enable the identification of conspicuous microscopic marks underwent a full scrutiny in order to determine the action of biotic agents.

17

ACCEPTED MANUSCRIPT Our results indicate that anthropogenic activity is marginal at AMK. Only one percussion mark on a long bone shaft was identified in Level 1. No other taphonomic evidence for human activity was documented in any of the three fossiliferous levels. Thus, no functional link between the few lithic tools recovered at AMK (one basalt and one gneiss artefacts in Level 1, one quartzite artefact in Level 2) and the bone assemblage could be established. In contrast, tooth marks appear in the whole assemblage. The type of tooth

SC RI PT

marks and their frequencies differ between levels (Table 4). Scores (Fig. 8B) have been found on several long bones and axial elements in Level 1: 34% of bovid elements bear at least one tooth score. Most tooth scores are on the diaphyses and do not appear in large numbers on each element. There is only one rib where more than 10 tooth scores were recorded. By contrast, specimens with tooth scores are very scarce in Level 2

NU

(8%), appearing only on a long bone shaft and on a rib.

Pits are also more abundant in Level 1 (45%) than in Level 2 (27%). In Level 1 more specimens bearing pits have been found, and also the number of pits per specimen

MA

is higher than in Level 2. Bones with carnivore surface modifications are mainly long limb bones and axial elements such as ribs, scapulae or coxals (Table 3). Among all the

ED

pits recovered, a few triangular pits on a tibia and a metapodial from Level 1 stand out (Fig. 8A). The size of the pits was statistically analyzed and compared to experimental analogues (see Andrews and Fernández-Jalvo, 1997; Domínguez-Rodrigo and Piqueras,

PT

2003; Andrés et al., 2012). Pit sizes of AMK were compared to the Maasai Mara spotted hyena den and the lion samples published in Andrés et al. (2012). Pits on shafts

CE

from medium-sized carcasses were selected in order to make the AMK sample comparable to the referential frameworks. Such selection largely reduced our sample, so

AC

that Level 2 yielded mean estimates since sample size was too small to derive confidence intervals (Table 5). Pits are scarce in Level 2 (n = 10) and most of them are located on axial bones (ribs, scapula, coxal) and on long bone ends. Level 1 pit dimensions are more similar to the ones observed in bones modified by spotted hyenas. The range of pit breadth and length in Level 1 fall within the confidence interval for the Maasai Mara hyena den (Fig. 9), although the AMK pits are clearly smaller (see maximum values in Table 5). Level 2 results point to an increase in pit size but data are too scarce to draw any conclusions (Fig. 9, Table 5).

18

ACCEPTED MANUSCRIPT Furrowing has been only detected on a few long bones (Table 4). Though in low frequencies, furrowing is more significant in Level 2 than in Level 1. This might seem contradictory as carnivore modification is much more prominent in Level 1 than in Level 2, but this can be explained by differential fragmentation patterns. While in Level 1 only few long bones remain sufficiently undamaged to properly identify furrowing, several complete and unmodified long bones (Fig. 10) have been recovered from Level 2. Typical hyena modifications (see Domínguez-Rodrigo et al., 2012a) have been

SC RI PT

recorded in a humerus from Level 1, where the proximal and distal epiphyses are missing (humerus on the right in Fig. 11). Equally, two other humerii in Level 2 (left and middle humeri in Fig. 11) might have been furrowed by hyenas, as they lack either the whole proximal epiphysis, or both ends. A femur (size 3a) in Level 2 does also show typical hyena furrowing as both ends have been deleted. In addition to epiphyseal

NU

deletion, some vertebral apophyses are also missing in Level 2. The disappearance of vertebral apophyses along with the absence of other kind of tooth marks is associated with felid consumption (Domínguez-Rodrigo, 1999). A lumbar vertebra in Level 2

MA

exhibits exactly the patterns observed after felid action, that is, modification of the apophyses and intact body (Fig. 8C).

ED

No bone with both human and carnivore modifications was found in any of the three levels. Thus, no clear evidence of their interaction on the assemblage could be found

PT

taphonomically.

CE

3.3.4. Disarticulation patterns, specimen size and site formation As explained above, specimens < 4 cm form a significant part of the assemblage

AC

(30% in Level 1, and 28% in Level 2). The presence of small fragments might indicate that hydraulic forces were not strong enough to generate a significant bias towards larger specimens. Along with these small specimens, the presence of less dense and easily transportable elements has been documented (Voorhies, 1969). Axial and compact bones, corresponding to Voorhies Group I and I&II, have been found in both AMK Levels. Cranial elements are poorly represented in both levels (Fig. 5). Cranial bones belong to Voorhies Group III, the least transportable group. The identification of elements of all Voorhies groups (Voorhies, 1969) indicates that the AMK accumulations represent neither a hydraulically-transported nor a lag assemblage in any of the two levels, despite the limitations due to small sample sizes. 19

ACCEPTED MANUSCRIPT The presence of epiphyses varies between levels at AMK. While Level 1 shows a lower frequency of long bone ends, in Level 2 a significant representation of epiphyses was recorded. The transportation of epiphyses in relation to hydraulic forces has been discussed by several authors (see Behrensmeyer, 1975; Pante and Blumenschine, 2010; Domínguez-Rodrigo et al., 2012b). The fact that at AMK epiphyses might have been subjected to different scavenging pressure and the different breakage patterns in Levels 1 and 2, might explain why epiphyses are much more

SC RI PT

abundant in the latter level. Taking all these results into account, postdepositional distortion, though present, must have been low or moderate, and in any case it was not constant during the formation of the site. The absence of bones with abrasion or polishing marks, as well as the orientation analyses presented above, also support this hypothesis. Thus, the accumulation at AMK seems to be autochthonous.

NU

Carnivore ravaging of bones in Level 1 does not facilitate an estimate of the time of formation of the assemblage other than by subaerial weathering stages, which do not exhibit large variation from 0-1, although few of specimens display higher

MA

weathering stages. Chemical weathering caused during diagenesis is documented in some specimens too. However, most of the assemblage seems to have a recent origin

ED

prior to sedimentation. The discovery of a partial primate hand with several carpal, metacarpal and phalangeal specimens in connection indicates a short exposure period. Phalanges are disarticulated very early (in the first 5% of time exposure until complete

PT

carcass disarticulation) (Hill, 1979; Hill and Behrensmeyer, 1984). This is why elements in which phalanges are still in connection to the limb most commonly display

CE

weathering stages 0-1 (Hill, 1979; Hill and Behrensmeyer, 1984). The presence of an almost complete articulated front limb of an artiodactyl in

AC

Level 2 also suggests short exposure time. Front limbs get disarticulated before hindlimbs. In this case, the humerus-radius-ulna, metacarpal and distal phalanx appear in anatomical connection. The distal phalanx was found a few cm away from the articulated limb. None of these elements display subaerial weathering stages beyond 01. This indicates a short exposure and rather quick sedimentation process (Hill, 1979; Hill and Behrensmeyer, 1984).

4. Discussion

20

ACCEPTED MANUSCRIPT AMK is one of the densest concentrations of remains caused by nonanthropogenic agencies on the FLK-Zinj paleolandscape. Its mostly paleontological nature shows a contrast in the density of remains per space unit (bones/m2) when compared to the core area of FLK-Zinj. At the latter site, an area similar in extension to the excavated area in AMK contained thousands of bones, in contrast with the 98 specimens found in AMK Level 1 and the 70 specimens found in Level 2. In addition, the taphonomic signatures of both assemblages also differ. Bones at FLK-Zinj display

SC RI PT

high frequencies of green-broken bones bearing abundant percussion and cut marks, whereas at AMK, the faunal assemblage is either fragmented with only tooth marks (Level 1) or moderately fragmented with a high survival of complete bones (Level 2). These results suggest that AMK was a palimpsest created by the action of several agents (with a marginal hominin input) superimposed over time on the landscape. To support

NU

this, only one hominin-made mark and lithics and the recovery of some hominin remains provide evidence for hominin presence at AMK. According to Bunn (1986), certain bone accumulations without lithics or with

MA

few artifacts like AMK should be considered natural background scatters. The main difference between natural and archaeological sites is the density of bone remains and

ED

the functional association of lithics and bones in the latter. Natural background scatters show significantly lower NISP, MNE and MNI estimations, as they are the result of the natural (i.e., as opposed to anthropogenic) deposition of carcasses in the landscape,

PT

instead of a repeated transport of carcasses to the same spot in the landscape by either humans or any other bone-accumulating agent such as hyenas (Potts, 1988).

CE

Natural deposits can be also distinguished by the absence of clear human traces and because they present high frequencies of cranial bones and axial elements

AC

(Bunn, 1986). This pattern is not completely observed at AMK, where cranial elements are rather scarce. Low energy runoffs could explain the lack of some other elements that are expected at such sites and that at the same time are easily transported by water. However, although present at AMK, flowing water seems to have played a small role in the assemblage configuration, not being the main cause of the bone accumulation and not altering the orientation pattern in a significant way. This is also supported by the presence of easily transported bones such as axial and compact bones, and a high percentage of fragments < 4 cm in every level. At AMK, we did not find any signs of polishing or abrasion that could potentially support the transportation of part of the

21

ACCEPTED MANUSCRIPT assemblage. The presence of animals of various sizes (including micromammals) has been documented. These facts suggest a local origin for the faunal assemblage, despite the potential effect of low-energy hydraulic rearrangements. It could be argued that the influence of water could have caused a small bias towards less dense elements; however, this could also be the result of carnivore post-depositional damage. Taphonomic analyses reveal a clear predominance of carnivore behavior at AMK, showing different behavioral patterns reflected in different fragmentation

SC RI PT

degrees and tooth mark frequencies. The frequency of tooth marks, including pits and scores, varies substantially between Level 1 and Level 2. Level 1 shows a higher frequency of tooth marks per specimen and according to MNE. The location of tooth marks is mainly concentrated on the long bone diaphyses. Fragmentation is reflected in long bone circumference types, which show similar patterns to those found in the

NU

Syokimau hyena den (Egeland et al., 2008). In contrast, fossils found in Level 2 are barely modified, many of them being complete or almost complete. This last pattern fits the skeletal and bone modification patterns documented in non-scavenged or only

MA

slightly scavenged felid-made bone accumulations, like the OCS (Olduvai Carnivore Site) (Arriaza et al., 2016, 2017). In the case of large and medium-sized carnivores,

ED

felids are the most specialized flesh-eaters, with teeth almost exclusively developed for meat-slicing and that barely leave marks on the cortical surfaces (Turner and Anton, 1997). Thus, it is understood that felids would produce fewer tooth marks than

PT

durophagous carnivores like hyenas (Brain, 1981; Selvaggio, 1994). Almost 99% of all complete long bones from the carcasses consumed by lions in the wild bear less than ten

CE

marks on the same shaft specimen (Gidna et al., 2014). This feature was also found among carcasses consumed by leopards: complete long bones rarely show more than 3–

AC

5 tooth marks (Domínguez-Rodrigo et al., 2007). The low frequency of tooth marks on midshafts of complete long bones is characteristic of felid primary access to carcasses. No specimen, including the complete long bones from Level 2 show more than 2 tooth marks on the same shaft. In addition, a lumbar vertebra with an intact centrum and damaged apophyses (Cavallo, 1998; Domínguez-Rodrigo, 1999), such as the one found in Level 2 supports our interpretation of felid intervention at the site. We conclude that differences between AMK Level 1 and 2 are most likely the result of the action of different carnivores.

22

ACCEPTED MANUSCRIPT The frequency of tooth marks recorded at Level 1 could be the result of the overlapping activity of felids and hyaenids. It seems plausible that a felid acted as the main accumulating agent leaving few marks on bone surfaces, and, secondly, hyenas partially ravaged the assemblage increasing the frequency of tooth marks, the degree of bone fragmentation and leaving typical hyena furrowing patterns. We propose this scenario because the bone destruction at AMK is lower than the one expected in an exclusively hyena-made bone accumulation. In Level 1, only 30.4% of the identified

SC RI PT

specimens bear some kind of carnivore modification, while in Syokimau the percentage reaches the 34.8% (Egeland et al., 2008) and in Kisima Ngeda hyena den 44% of the assemblage (Prendergast and Domínguez-Rodrigo, 2008). Although these percentages are not extremely different, it is noteworthy that Syokimau den is considered a low competition setting, with considerably lower destruction patterns than other open-air

NU

actualistic sites in East Africa (Egeland et al., 2008). Carnivore modification in Level 2 is even lower. In addition, some of the complete elements appear articulated. This may indicate the transport of complete animals or limbs by a carnivore. The discovery of

MA

other specimens that might belong to the same individual (axial and long bones from front and hind limbs) could be indicating that the complete skeleton was at some point

ED

in time at the site. In that case, the question is whether it represents a natural death or whether it was transported or hunted in locus. Felids have been documented to occasionally transport complete carcasses at a short distance from kills (Brain, 1981;

PT

Ruiter and Berger, 2000; Domínguez-Rodrigo and Pickering, 2010; Arriaza et al., 2016).

CE

Several Bed I sites in Olduvai Gorge have been generated by carnivores, while hominin intervention on faunas is documented only sporadically. This is the case for the

AC

FLK-N, FLK-NN and DK (Domínguez-Rodrigo, 2001; Domínguez-Rodrigo et al., 2007; Egeland 2007). The high specialization pattern of prey, low breakage, low tooth mark frequencies, skeletal part representation and typical bone modification of felids on the axial skeleton and long bone ends, suggests that the accumulating agent in some of these sites was a felid (Domínguez-Rodrigo et al., 2007; Arriaza and DomínguezRodrigo, 2016). In some of these sites (e.g. DK or FLK-NN) subsequent scavenging by hyenas has been also detected (Domínguez-Rodrigo et al., 2007; Egeland, 2007), showing similar patterns to those documented at AMK Level 1.

23

ACCEPTED MANUSCRIPT The choice of a specific spot in the landscape used repeatedly by carnivores is usually associated with locations that might have served as shelters (Cavallo, 1998). That means that the immediate environment would have provided sufficient safety for the consumption of carcasses. According to paleoenvironmental analyses, the habitat surrounding AMK would have been characterized by an even more dense vegetation than the one be found in the FLK-Zinj area (Arráiz et al., submitted). Such a dense arboreal environment would have encouraged predators to transport their prey (if

SC RI PT

obtained in nearby more open spaces) to the site, since it would have been an area of safe consumption. This type of interpretation was provided for understanding the formation of FLK-N (Domínguez-Rodrigo et al., 2010a). Similar places are currently spots of very low competition in modern African landscapes where the accumulation of appendicular elements is very high due to repeated transport and low postdepositional

NU

disturbance (Blumenschine, 1986; Cavallo, 1998; Dominguez-Rodrigo, 2001). In most of the FLK sites, biochemical modifications caused by fungi or bacteria have been detected. Such modifications were also observed at AMK. The presence of these marks

MA

suggests that the remains were deposited in moist and shaded areas (Marchiafava et al, 1974; Hackett, 1981; Piepenbrink, 1984; Child, 1995; Sharmin et al, 2003). These

ED

conditions rarely occur in open floodplains, but are more widely documented on bones deposited in wooded environments near water sources.

PT

Recent paleovegetation reconstruction of the FLK-Zinj paleolandscape (Arráiz et al., submitted) contradicts previous paleoecological interpretations of the area.

CE

According to geological and paleobotanical analyses, Blumenschine et al. (2012) suggested the existence of grasses and sedges in the wetlands close to a river channel 50

AC

to 200 m southeast of FLK-Zinj site (between PTK and FLK-Zinj). Palm phytoliths, like those revealed by our analyses, are incompatible with the river channel proposed by Blumenschine et al. (2012), since in that case there would have been no mature soils enabling the growth of palm tree roots. The recent paleovegetation description matches the one published for FLK-NN (Ashley et al., 2010b) and is consistent with previous studies that suggest that the vegetation of the area was densely wooded during middle and uppermost Bed I times. The presence of ferns in some of the phytoliths samples suggests shaded and humid habitats, supporting the existence of a river input southeast of PTK site, as indicated by Uribelarrea et al. (2014). The somewhat elevated platform in the Zinj paleolandscape seems to have been a mosaic dominated by palm trees, but 24

ACCEPTED MANUSCRIPT also including wetland areas and some more open sedges and grasslands (Barboni at al., 2010). Such a mixed environment acted as focal point for a wide range of herbivores and, therefore, also for carnivores and hominins (Domínguez-Rodrigo et al., 2010b). The discovery and the interpretation of AMK presented here is also of major relevance to show that factors other than forested habitats must have influenced the formation of anthropogenic sites on the FLK-Zinj platform, since the dense arboreal

SC RI PT

vegetation was not limited to the locations where those anthropogenic sites occurred.

5. Conclusions

AMK is a recently discovered Olduvai Bed I site consisting of three fossiliferous levels where human presence has been detected in form of human fossil remains and

NU

very few lithic artefacts. Due to its location at the Eastern Olduvai paleolake side, an area characterized by dense vegetation, AMK appears within a close and lowcompetition environment, which predators may have used to consume their prey, and

MA

resulting in the deposition of carcass remains.

The superimposition of depositional events at AMK would have created a

ED

palimpsest reflecting the independent actions of carnivores (possibly felids and hyaenids) and the marginal though perceptible presence of hominins. The discovery of

PT

such sites should make archaeologists reflect on the functional relationship between artifacts and bones found together in the same fossiliferous levels. For instance, AMK

CE

represents an accidental association between very few lithic tools and a substantial number of bone remains.

AC

According to the taphonomic results presented here, natural processes would have been responsible for the formation of AMK. This site could be described as a natural background scatter in a space intensively covered with vegetation. Assemblages like AMK Level 1 and 2 and FLK-NN1 and 3 show the contrast in density of remains between anthropogenic (FLK-Zinj, PTK, DS) and non-anthropogenic sites and also shows that despite the occasional availability of marrow-bearing bones on the Olduvai paleolandscape, hominins disregarded them most of the time, as more clearly documented slightly later at 1.8 Ma at FLK-N.

25

ACCEPTED MANUSCRIPT

Acknowledgments We wish to thank The Ngorongoro Conservation Area Authorities, COSTECH and the Antiquities unit for permits to conduct research at Olduvai, and Museo Arqueológico Regional de la Comunidad de Madrid. The authors greatly appreciate the major funding provided by the Spanish Ministry of Science and Innovation through the

SC RI PT

European project I + D HAR2013-45246-C3-1P and the Spanish Ministry of Culture through the Heritage Institute and the Program of Funding for Archaeological Projects Abroad. We would like to thank the entire TOPPP team, especially the local workers and Ainara Sistiaga for the help provided in the conduction of the research at AMK. We highly appreciated the field work conducted by the field school students during 2013

NU

and 2014 campaings. The corresponding author would like to thank Fundación La Caixa and the Spanish Education, Culture and Sports Ministry for funding her postgraduate education programs. Finally, we would like to thank two anonymous reviewers for their

MA

comments and suggestions. This paper highly benefited from their feedback. References

PT

ED

Alcántara, V., Barba, R., Barral del Pino, J.M., Crespo, A.B., Eiriz, A.I., Falquina A., Herrero, S., Ibarra, A., Megías, M., Pérez Gil, M., Pérez Tello, V., Rolland, J., Yravedra, J., Vidal, A. & Domínguez-Rodrigo, M., 2006: Determinación de procesos de fractura sobre huesos frescos: un sistema de análisis de los ángulos de los planos de fracturación como discriminador de agentes bióticos. Trabajos de Prehistoria 63 (1), 37-45.

AC

CE

Andrés, M., Gidna, A., Yravedra, J. & Domínguez-Rodrigo, M., 2012: A study of dimensional differences of Toth marks (pits and scores) on bones modified by small and large carnivores. Archaeological and Anthropological Sciences 4, 209-219. Andrews, P. & Fernández-Jalvo, Y., 1997: Surface modification of the Sima de los Huesos fossil humans. Journal of Human Evolution 33, 191-217. Arráiz, H., Barboni, D., Baquedano, E., Mabulla, A., Uribelarrea, D., DomínguezRodrigo, M., submitted. The FLK Zinj paleolandscape: reconstruction of a 1.84 Ma wooded habitat in the FLK Zinj-AMK-PTK-DS archaeological complex, Middle Bed I (Olduvai Gorge, Tanzania). Arriaza, M.C. & Domínguez-Rodrigo, M. 2016: When felids and hominins ruled at Olduvai Gorge: A machine leraning analysis of the skeletal profiles of the nonanthropogenic Bed I sites. Quaternary Science Reviews 139, 43-52. Arriaza, M.C., Domínguez-Rodrigo, M., Yravedra & J., Baquedano, E., 2016: Lions as Bone Accumulators? Paleontological and Ecological Implications of a Modern Bone 26

ACCEPTED MANUSCRIPT Assemblage from Olduvai doi:10.1371/journal.pone.0153797

Gorge.

PLoS

ONE

11(5):

e0153797.

Arriaza, M.C., Yravedra, J., Domínguez-Rodrigo, M., Maté-González, M.A., García Vargas, E., PAlomeque-González, J.F., Aramendi, J., González-Aguilera, D., Baquedano, E., 2017. On applications of micro-photogrammetry and geometric morphometrics to sutides of tooth mark morphology: The modern Olduvai Carnivore Site (Tanzania). Palaeography, Palaeoclimatology, Palaeocology (in this Vol.). doi: 10.1016/j.palaeo.2017.01.036.

SC RI PT

Ashley, G.M., Barboni, D., Dominguez-Rodrigo, M., Bunn, H.T., Mabulla, A.Z.P., Diez- Martin, F., Barba, R. & Baquedano, E. 2010a: A spring and woodland: new paleo-environmental reconstruction of FLK-Zinj, Olduvai Gorge, Tanzania. Quaternary International 74, 304-314. Ashley, G.M., Barboni, D., Domínguez-Rodrigo, M., Bunn, H.T., Mabulla, A., DíezMartin, F., Barba, R. & Baquedano, E. 2010b: Paleoenvironmental and paleoecological reconstruction of a freshwater oasis in savannah grassland at FLK North, Olduvai Gorge, Tanzania. Quaternary Research 74, 333-343.

MA

NU

Barba, R. & Domínguez-Rodrigo, M. 2005: The taphonomic relevance of the analysis of bovid long limb bone shaft features and their application to element iden- tification: study of bone thickness and morphology of the medullary cavity. Journal of Taphonomy 3, 17-42.

ED

Barboni, D., Ashley, G.M., Domínguez-Rodrigo, M., Bunn, H.T., Mabulla, A.Z.P. & Baquedano, E. 2010: Phytoliths infer locally dense and heterogeneous paleovegetation at FLK North and surrounding localities during upper Bed I time, Olduvai Gorge, Tanzania. Quaternary Research 74, 344-354.

PT

Behrensmeyer, A.K. 1975: The taphonomy and paleoecology of Plio-Pleistocene vertebrate assemblages east of Lake Rudolf, Kenya. Bulletin of the Museum of Comparative Zoology 145, 473–574.

CE

Behrensmeyer, A.K. 1978: Taphonomic and ecologic information from bone weathering. Paleobiology 4 (2), 150-162

AC

Binford, L. R. 1981: Bones: Ancient Men, Modern Myths. New York: New York Academic Press. Blumenschine, R.J. 1986: Early Hominid Scavenging Opportunities. Implications of Carcass Availability in the Serengeti and Ngorongoro Ecosystems. Oxford: B.A.R. International Series 283. Blumenschine, R.J. 1988: An experimental model of the timing of hominid and carnivore influence on archaeological bone assemblages. Journal of Archaeological Science 15, 483–502. Blumenschine, R.J., 1991: Hominid carnivory and foraging strategies, and the socioeconomic function of early archaeological sites. Philosophical Transactions of the Royal Society London 334, 211-221.

27

ACCEPTED MANUSCRIPT Blumenschine, R.J. 1995: Percussion marks, tooth marks and the experimental determinations of the timing of hominin and carnivore access to long bones at FLK Zinjanthropus, Olduvai Gorge, Tanzania. Journal of Human Evolution 29, 21-51. Blumenschine, R.J. & Marean, C.W. 1993: A carnivore’s view of archaeological bone assemblages. In: From Bones to Behavior: Ethnoarchaeological and Experimental Contributions to the Interpretations of Faunal Remains. Southern Illinois University, 271–300.

SC RI PT

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 marks, and carnivore tooth marks on bone surfaces. Journal of Archaeological Science 23, 493– 507. Blumenschine, R.J., & Selvaggio, M.M., 1988: Percussion marks on bone surfaces as a new diagnostic of hominin behavior. Nature 333, 763-765.

NU

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: Cultural Beginnings: Approaches to Under- standing Early Hominin Life Ways in the African Savanna. R. Habelt. Bonn, 17-32.

MA

Blumenschine, R.J., Stanistreet, I.G., Njau, J.K., Bamford, M.K., Masao, F.T., Albert, R.M., Stollhofen, H., Andrews, P., Prassack, K.A., McHenry, L.J., Fernández- Jalvo, Y., Camilli, E.L. & Ebert, J.I. 2012: Environments and hominin activities across the FLK Peninsula during Zinjanthropus times (1.84 Ma), Olduvai Gorge, Tanzania. Journal of Human Evolution 63, 364-383.

ED

Boaz, N. & Behrensmeyer, A.K. 1976: Hominid taphonomy: transport of human skeletal parts in an artificial fluviatile environment. American Journal of Physical Anthropology 45 (l), 53-60.

PT

Bonnefille, R. 1984: Palynological Research at Olduvai Gorge. Research Reports 17 National Geographic Society, 227-243.

CE

Brain, C. K. 1981: The Hunters or the Hunted? Chicago: Chicago University Press.

AC

Bunn, H.T. 1981: Archaeological evidence for meat-eating by Plio-Pleistocene hominins from Koobi Fora, Kenya. Nature 291, 574-577. Bunn, H.T., 1982: Meat-eating and Human Evolution: Studies on the Diet and Subsistence Patterns of Plio-pleistocene Hominids in East Africa. Ph.D. Dissertation, University of California, Berkeley. Bunn, H.T. 1983: Comparative analysis of modern bone assemblages from a San Hunter gatherer camp in the Kalahari desert, Botswana, and from a spotted hyena den near Nairobi, Kenya. In: Animals and archaeology: Hunters and their Prey. B.A.R. International Series 163, 143-148. Bunn, H.T. 1986: Patterns on skeletal representation and hominid subsistence activities at Olduvai Gorge, Tanzania, and Koobi Fora, Kenya. Journal of Human Evolution 15, 673–690.

28

ACCEPTED MANUSCRIPT Bunn, H.T. 1991: A taphonomic perspective on the archaeology of human origins. Annual Review of Anthopology 20, 433-467. Bunn, H.T. & Ezzo, J.A. 1993: Hunting and scavenging by Plio-Pleistocene hominids: nutritional constraints, archaeological patterns, and behavioural implications. Journal of Archaeological Science 20, 365-398. Bunn, H.T. & Kroll, E.M. 1986: Systematic butchery by Plio-Pleistocene hominids at Olduvai Gorge, Tanzania. Current Anthropology 27, 431-452.

SC RI PT

Bunn, H.T. & Kroll, E.M. 1988: Fact and fiction about the Zinjanthropus floor: data, arguments, and interpretations (a reply to Binford). Current Anthropology 29, 135-149. Bunn, H.T. & Pickering, T.R. 2010: Bovid mortality profiles in paleoecological context falsify hypotheses of endurance runningehunting and passive scavenging by early Pleistocene hominins. Quaternary Research 74, 395-404. Capaldo, S.D., 1995: Inferring Hominid and Carnivore Behavior from Dual-patterned Archaeological Assemblages. Ph.D. Dissertation Rutgers University, New Brunswick.

NU

Capaldo, S.D. 1997: Experimental determinations of carcass processing by PlioPleistocene hominids and carnivores at FLK 22 (Zinjanthropus), Olduvai Gorge, Tanzania. Journal of Human Evolution 33, 555-597.

MA

Capaldo, S.D. & Blumenschine, R.J. 1994: A quantitative diagnosis of notches made by hammerstone percussion and carnivore gnawing in bovid long bones. American Antiquity 59, 724-748.

ED

Cavallo, J.A. 1998: A re-examination of Isaac’s centralplace foraging hypothesis. Ph.D. Dissertation Rutgers University, New Brunswick.

PT

Child, A.M. 1995: Towards an understanding of the microbial decomposition of archaeological bone in the burial environment. Journal of Archaeological Science 22, 165–174.

CE

De Juana, S. & Domínguez-Rodrigo, M. 2011: Testing analogical taphonomic signatures in bone breaking: a comparison between hammerstone-broken equid and bovid bones. Archaeometry 53, 996-1011.

AC

Deino, A.L. 2012: 40Ar/39Ar dating of Bed I, Olduvai Gorge, Tanzania, and the chronology of early Pleistocene climate change. Journal of Human Evolution 63, 251– 273. Domínguez-Rodrigo, M. 1997: Meat-eating by early hominids at the FLK 22 Zinjanthropus site, Olduvai Gorge, Tanzania: an experimental approach using cut mark data. Journal of Human Evolution 33, 669-690. Domínguez-Rodrigo, M. 1999: Flesh availability and bone modification in carcasses consumed by lions. Palaeogeography, Palaeoclimatology and Palaeoecology 149, 373388.

29

ACCEPTED MANUSCRIPT Domínguez-Rodrigo, M. 2001: A study of carnivore competition in riparian and open habitats of modern savannas and its implications for hominid behavioral modelling. Journal of Human Evolution 40, 77–98. Domínguez-Rodrigo, M & Martínez-Navarro, B. 2012: Taphonomic analysis of the early Pleistocene (2.4 Ma) faunal assemblage from A.L. 894 (Hadar, Ethiopia). Journal of Human Evolution 62, 315-327. Domínguez-Rodrigo, M. & Pickering, T.R. 2003: Early hominids, hunting and scavenging: a summary of the discussion. Evolutionary Anthropology 12, 275-282.

SC RI PT

Domínguez-Rodrigo, M. & Pickering, T.R. 2010: Un estudio tafonómico multivariante de las acumulaciones de fauna de hiénidos (Crocuta crocuta) y félidos (Panthera pardus). Zona Arqueológica 13, 50–66. Domínguez-Rodrigo, M. & Piqueras, A. 2003: The use of tooth pits to identify carnivore taxa in tooth-marked archaeofaunas and their relevance to reconstruct hominid carcass processing behaviors. Journal of Archaeological Science 30, 13851391.

NU

Domínguez-Rodrigo, M., Barba, R. & Egeland, C.P. 2007: Deconstructing Olduvai. New York: Springer.

MA

Domínguez-Rodrigo, M., de Juana, S., Galán, A.B. & Rodríguez, M. 2009: A new protocol to differentiate trampling marks from butchery cut marks. Journal of Archaeological Science 36, 2643-2654.

ED

Domínguez-Rodrigo, M., Mabulla, A.Z.P., Bunn, H.T., Diez-Martin, F., Baquedano, E., Barboni, Barba, R., Domínguez-Solera, S., Sánchez, P., Ashley, G.M., Yravedra, J. 2010a: Disentangling hominin and carnivore activities near a spring at FLK North (Olduvai Gorge, Tanzania). Quaternary Research 74, 363-375.

CE

PT

Domínguez-Rodrigo, M., Bunn, H.T., Mabulla, A.Z.P., Ashley, G.M., Diez-Martin, F., Barboni, D., Prendergast, M.E., Yravedra, J., Barba, R., Sánchez, A., Baquedano, E., Pickering, T.R., 2010b. New excavations at the FLK Zinjanthropus site and its surrounding landscape and its behavioral implications. Quaternary Research 74, 315332.

AC

Domínguez-Rodrigo, M., Gidna, A.O., Yravedra, J. & Musiba, C. 2012a: A Comparative Neo-Taphonomic Study of Felids, Hyaenids and Canids: an Analogical Framework Based on Long Bone Modification Patterns. Journal of Taphonomy 10, 151-169. 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., Barboni, D., Arriaza, C., Gidna, A. 2012b: 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. Domínguez-Rodrigo, M., Bunn, H.T. & Yravedra, J. 2014: A critical re-evaluation of bone surface modification models for inferring fossil hominin and carnivore interactions

30

ACCEPTED MANUSCRIPT through a multivariate approach: application to the FLK Zinj archaeofaunal assemblage (Olduvai Gorge, Tanzania). Quaternary International 322-323, 32-43. Domínguez-Rodrigo, M., Yravedra, J., Organista, E., Gidna, A., Fourvel, J-B. & Baquedado, E. 2015: A new methodological approach to the taphonomic study of paleontological and archaeological faunal assemblages: a preliminary case study from Olduvai Gorge (Tanzania). Journal of Archeological Science 59, 35-53.

SC RI PT

Domínguez-Rodrigo, M., Cobo, L., Uribelarrea, D., Arriaza M.C., Yravedra, J., Gidna, A., Organista, E., Sistiaga, A., Baquedano, E., Martín, D, Aramendi, J. & Mabulla, A. (submitted). Spatial simulation and modeling of the early Pleistocene site of DS (Bed I, Olduvai Gorge, Tanzania): a powerful tool for predicting potential archaeological information from unexcavated areas. Egeland, C.P. 2007: Zooarchaeology and taphonomy of the DK site. In: Deconstructing Olduvai. New York: Springer, 253-268.

NU

Egeland, C.P. 2010: Zoorarchaeological and Taphonomic perspectives on hominis and carnivore interactions at Oduvai Gorge, Tanzania. Ph.D. Dissertation, Indiana University.

MA

Egeland, C.P. & Byerly, R. 2005: Application of return rates to large mammal butchery and transport among hunter-gatherers and its implications for Plio-Pleistocene hominid carcass forgaring and site use. Journal of Taphonomy 3 (3), 135–157. Egeland A., Egeland C.P. & Bunn H.T. 2008: Taphonomic analysis of a modern spotted hyena (Crocuta crocuta) den from Nairobi, Kenya. Journal of Taphonomy 6, 275–299.

ED

Fernández-Jalvo, Y., Denys, C., Andrews, P., Williams, T., Dauphin, Y. & Humphrey, L., 1998: Taphonomy and palaeocology of Olduvai bed I (Pleistocene, Tanzania). Journal of Human Evolution 34, 137-172.

CE

PT

Fiorillo, A.R. 1991: Taphonomy and depositional setting of Careless Creek Quarry (Judith River formation), Wheatland County, Montana, U.S.A. Palaeogeography, Palaeoclimatology, Palaeoecology 81, 281-311. Fisher, N.I. 1995: Statistical Analysis of Circular Data. Cambridge: Cambridge University Press.

AC

Gidna, A.O., Kisui B., Musiba, C., Mabulla, A. & Domínguez-Rodrigo, M. 2014: An ecological neo-taphonomic study of carcass consumption by lions in Tarangire National Park (Tanzania) and its evidence for human evolutionary biology. Quaternary International 322–323, 167–180. Hackett, C.J. 1981: Microscopical focal destruction (tunnels) in exhumed human bones. Medicine, Science and the Law 21, 243–265. Hay, R. 1976: Geology of the Olduvai Gorge. Berkeley: University of California Press. Hill, A. 1979: Disarticulation and Scattering of Mammal Skeletons. Pelobiology 5 (3), 261-274. Hill, A. 1989: Bone Modification by modern Spotted Hyenas. In: Bone Modification. University of Maine: Center for the Study of the First Americans, 169-178. 31

ACCEPTED MANUSCRIPT Hill, A. & Behrensmeyer, A.K. 1984: Natural Disarticulation and Bison Butchery. American Antiquity 50 (1), 141-145. Howard, P. 2007: Archaeological Survey and Mapping. London: Taylor and Francis. Isaac, G.L. 1978: The food-sharing behavior of protohuman hominids. Scientific American 238, 90-108.

SC RI PT

Isaac, G.L. 1983: Bones in contention: competing explanations for the juxtaposition of Early Pleistocene artifacts and faunal remains. In: Animals and Archaeology 1. Hunters and Their Prey, B.A.R. International Series, Vol. 163, 3-19. Leakey, L. 1965: Olduvai Gorge 1951-1961. A Preliminary Report on the Geology and Fauna. Cambridge: Cambridge University Press. Leakey, M. 1971: Olduvai Gorge, Excavations in Bed I and II, Vol. 3. Cambridge: Cambridge University Press, 1960-1963. Lyman, R. 1994: Vertebrate Taphonomy. Cambridge: Cambridge University Press.

NU

Marchiafava, V., Bonucci, E. & Ascenzi, A. 1974: Fungal osteoclasia: a model of dead bone resorption. Calcified Tissue Research 14, 195–210.

MA

Marean, C.W., Domínguez-Rodrigo, M. & Pickering, T.R. 2004: Skeletal element equifinality in zooarchaeology begins with method: the evolution and status of the “shaft critique”. Journal of Taphonomy 2, 69–98.

ED

Munsell Soil Color Charts, 1994. Revised Edition. New York: Macbeth Division of Kollmorgen Instruments Corporation, New Windsor, 10. Munzel, S.C. 1988: Quantitative analysis and archaeological site interpretation. Archaeozoologia 2, 93-110.

CE

PT

Oliver, J.S. 1994: Estimates of hominid and carnivore involvement in the FLK Zinjanthropus fossil assemblage: some socioecological implications. Journal of Human Evolution 27, 267-294.

AC

Ortega, M.C., Canseco, O.T, Pastor, A., Arsuaga, J.L., Pérez-González, A., Laplana, C., Márquez, B. & Baquedano, E. 2016: In situ conservation strategies at the Pleistocene sites of Pinilla del Valle, Madrid (Spain). Journal of Paleontological Techniques 15, 84-111. Pante, M.C. & Blumenschine, R.J. 2010: Fluvial transport of bovid long bone fragmented by the feeding activities of hominins and carnivores. Journal of Archaeological Science 37, 846-854. Parkinson, J., Plummer, T. & Hartstone-Rose, A. 2015: Characterizing felid tooth marking and gross bone damage patterns using GIS image analysis: An experimental feeding study with large felids. Journal of Human Evolution 80, 114–134. doi: 10.1016/j.jhevol.2014.10.011. pmid:25467112 Patou-Mathis, M.E. 1985: La fracturation des os longs de grands mammifères: élaboration d’un lexique et d’une fiche type. Outillage peu élabore en os et en bois de cervidés. Artefacts 1, 11-22. 32

ACCEPTED MANUSCRIPT Pickering, T.R., Domínguez-Rodrigo, M., Egeland, C.P. & Brain, C.K. 2005: The contribution of limb bone fracture patterns to reconstructing early hominid behaviour at Swartkrans cave (South Africa): archaeological application of a new analytical method. International Journal of Osteoarchaeology 15 (4), 247-260. Pickering, T.R. & Egeland, C.P. 2006: Experimental patterns of hammerstone percussion damage on bones: implications for inferences of carcass processing by humans. Journal of Archaeological Sicience 33, 459-469.

SC RI PT

Piepenbrink, H. 1984: Examples of signs of biogenic decomposition in bones long buried. Anthropologischer Anzeiger 42, 241–251 Plummer, T.W. & Bishop, L.C. 1994: Hominid paleocology at Olduvai Gorge, Tanzania, as indicated by antelope remains. Journal of Human Evolution 27, 47-76. Potts, R. 1988: Early Hominid Activities at Olduvai. New York: Aldine.

NU

Prendergast, M. & Domínguez-Rodrigo. M. 2008: Taphonomic analyses of a hyena den and a natural-death assemblage near Lake Eyasi (Tanzania). Journal of Taphonomy 6, 301–335. Rose, L. & Marshall, F. 1996: Meat eating, hominid sociality, and home bases revisited. Current Anthropology 37, 307-338.

MA

Ruiter, D. & Berger, L. 2000: Leopards as taphonomic agents in dolomitic caves– implications for Bone accumulations in the hominid-bearing deposits of South Africa. Journal of Archaelogical Science 27, 665–684. doi: 10.1006/jasc.1999.0470

PT

ED

Selvaggio, M.M. 1994: Identifying the timing and sequence of hominid and carnivore involvement with Plio-Pleistocene bone assemblages from carnivore tooth marks and stone-tool butchery marks on bone surfaces, Ph.D. Dissertation. Rutgers University, New Brunswick

CE

Selvaggio, M.M.& Wilder, J. 2001: Identifying the Involvement of Multiple Carnivore Taxa with Archaeological Bone Assemblages. Journal of Archaeological Science 28, 465-470

AC

Sharmin, S., Kishi, F., Sano, A., Kamei, K., Nishimura, K. & Miyaji, M. 2003: Direct invasion of bones by highly pathogenic fungi in an in vitro model and its ecological significance. Japanese Journal of Medical Mycology 44, 17–23. Shipman, P. & Rose, J. 1983: Early hominids hunting, butchering and carcass processing behaviors: approach to the fósil record. Journal of Anthropological Archaeology 2, 57- 98. Shipman, P., Fisher, D.C. & Rose, J. 1984: Mastodon butchery: microscopic evidence of carcass processing and bone tool use. Paleobiology 10, 358–365. Sikes, N. 1994: Early hominid habitat preferences in East Africa: paleosol carbonate isotopic evidence. Journal of Human Evolution 27, 25-45. Spencer, L.M. 1997: Dietary adaptations of Plio-Pleistocene Bovidae: implications for hominin habitat use. Journal of Human Evolution 32, 201-228.

33

ACCEPTED MANUSCRIPT Thompson, C.E., Ball, S., Thompson, T.J.U. & Gowland, R. 2011: The abrasion of modern and archaeological bones by mobile sediments: the importance of transport modes. Journal of Archaeological Science 38, 784-793. Toots, H. 1965: Orientation and distribution of fossils as environmental indicators. In: Nineteenth Field Conference of the Wyoming Geological Association, 219-229. Turner, A. & Anton, M. 1997: The Big Cats and their Fossil Relatives. New York: Columbia University Press.

SC RI PT

Uribelarrea, D., Domínguez-Rodrigo, M., Pérez-Gonzalez, A., Vegas Salamanca, J., Baquedano, E., Mabulla, A., Musiba, C., Barboni, D. & Cobo-Sánchez, L. 2014: Geoarchaeological and geometrically corrected reconstruction of the 1.84Ma FLK Zinj paleolandscape at Olduvai Gorge, Tanzania. Quaternary International 322-323, 7-31. Villa, P. & Mahieu, E. 1991: Breakage patterns of human long bones. Journal of Human Evolution 21, 27-48. Voorhies, M.R. 1969: Taphonomy and population dynamics of the early Pliocene vertebrate fauna, Knox County, Nebraska. Contributions to Geology Special 1, 1-69.

AC

CE

PT

ED

MA

NU

Yravedra, J. & Domínguez-Rodrigo, M. 2009: The shaft-based methodological approach to the quantification of long limb bones and its relevance to understanding hominin subsistence in the Pleistocene: application to four Paleolithic sites. Journal of Quaternary Science 24, 85-96.

34

ACCEPTED MANUSCRIPT Figure caption Figure 1. A: Location of Olduvai Gorge in Tanzania and the contemporary lowermost Bed I sites in the junction of the Gorge. Upper-left corner: Digital Globe Foundation image. Upper-Right corner: location of AMK in the Secondary Gorge and the closest anthropogenic sites FLK Zinj, PTK and DS; the image was created using photogrammetric techniques using a dron. Below: Panoramic view of the four sites

SC RI PT

Figure 2. Stratigraphic cross-section of AMK, with the position of the three fossiliferous levels. Right: Views of the 2014 excavations at the site with the exact location of the three fossiliferous levels. The small box in the inferior picture corresponds to figure 3. Figure 3. Detail picture of the clay unit under Tuff IC, where the thin lamination in the lower half of the deposit can be appreciated. Figure 4. Rose diagrams for azimuth orientation in AMK Levels 1 and 2; and statistical tests and p values for orientation distribution, where p>0.05 indicates significant anisotropy.

NU

Figure 5. Skeletal part profiles according to MNE for AMK Levels 1 and 2.

MA

Figure 6. Circumference type frequencies (Bunn, 1982) identified for medium sized bovid long bones in AMK Levels 1 and 2, compared with Syokimau (Egeland et al., 2008) and the OCS (Arriaza et al., 2016) samples. The different types refer to 1: <50%, 2: >50% and 3: complete shafts.

ED

Figure 7. Examples of abiotic bone modifications documented at AMK: (a) carbonate concretions, (b) biochemical alterations related to the presence of root etching, (c) manganese alterations, (d) exfoliation.

CE

PT

Figure 8. Examples of carnivore surface modifications observed in AMK. A: Triangular pits on matapod shaft; B: Score on a shaft fragment; C: Lumbar vertebra lacking apophyses and any other carnivore modification.

AC

Figure 9. Mean values and 95% confidence intervals of tooth pit breadth and length on shafts in the AMK sample and the data published for hyenas and lions in Andrés et al. (2012). AMK pit dimensions were calculated following Andrés et al. (2012), using a two tailed confidence alpha value of 0.025. Level 2 is only represented by its mean dimensions since sample size is too small to obtain reliable statistical results. Figure 10. Complete medium sized bovid long bones recovered from AMK Level 2. Left to right: tibia, radius and metacarpus. Figure 11. Cylinders of humeri (size 3a) found at AMK showing intense deletion of ends typical of durophagous carnivores.

35

AC

CE

PT

ED

MA

NU

SC RI PT

ACCEPTED MANUSCRIPT

36

AC

CE

PT

ED

MA

NU

SC RI PT

ACCEPTED MANUSCRIPT

37

AC

CE

PT

ED

MA

NU

SC RI PT

ACCEPTED MANUSCRIPT

38

MA

NU

SC RI PT

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Fig. 4

39

MA

NU

SC RI PT

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Fig. 5

40

AC

CE

PT

ED

MA

NU

SC RI PT

ACCEPTED MANUSCRIPT

41

AC

CE

PT

ED

MA

NU

SC RI PT

ACCEPTED MANUSCRIPT

42

AC

CE

PT

ED

MA

NU

SC RI PT

ACCEPTED MANUSCRIPT

43

NU

SC RI PT

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA

Fig. 9

44

AC

CE

PT

ED

MA

NU

SC RI PT

ACCEPTED MANUSCRIPT

45

AC

CE

PT

ED

MA

NU

SC RI PT

ACCEPTED MANUSCRIPT

46

ACCEPTED MANUSCRIPT

Table 1. Detailed phytolith counts in AMK site samples

Taxonomy

Grass/Sedge FI*** FI*** Grass Fern FI***

ED

PT

CE

AC

Samples DB14-27C DB14-29 (2.993333/35.347583)* (2.993361/35.3475)* 182 5 23 35 1 5 3 0

SC RI PT

FI***

FI***

MA

Globular echinate body with distinct spines Faceted, smooth, granulate and irregular bodies Acicular body (L**16-20) Parallepiped body, bulliform-type texture (L**20-40) Parallepiped/cubic body, sinuous edges, psilate surface (L**10-40) Parallelepiped/cubic/ovate body, granulate surface (L**15-20) Blocky body cuneiform, likely silicified bulliform cell (L**20-30) Paralleleipedal. Faceted, sharp edges (L**20-80) Elongate tabular, irregular edges, rough surface (L**25-60) Elongate cylindrical body, surface psilate to slightly lacunate (L**20-30) Elongate tabular or slightly cylindrical, surface perforated-granulate (L**30) Elongate quadrangular section, body curved, surface spilate to slightly rugose(L**22-70) Elongate smooth glass-rod type straight body (L**15-30) Elongate cylindrical/subcylindrical body granulate (L**20-120) Dubious elongates, altered. Grass Silica ShortCells (GSSC) Mesophyl dicot cell Plate, thin, irregular contour, wavy surface, finely rugose (L**20-50) Thin plate, edge or surface lacunate (L**30) Tabular thick smooth, usually contorted (L**2040) Dubious phytoliths

NU

Description (size in microns)

Grass/Sedge

5 6 1 3 0

16

0

4

FI***

12 4

61 0

36 25 0 2 13 12

FI***

Σ phytoliths

0 0

5

12

FI***

56

1 323

89 294

*Latitude S and Longitude E Coordinates **L: length ***FI: Forest Indicator phytoliths

47

ACCEPTED MANUSCRIPT

Table 2. Taxonomic representation according to NISP, MNE and MNI at AMK

MNE

MNI

3 1 30 -

2 1 18 -

2 1 3 -

35 3

26 3

5 1

-

-

-

1

1

1

1

1

1

-

-

-

8

1

1

-

-

-

1 2

2 2

2 1

1 1

2 5

MA

3 5

SC RI PT

NISP

1

1

1

2

1

1

3 -

2 -

1 -

2 >10

2 >10

1 2

56

33

12

>57

>46

13

AC

CE

PT

Total

MNI

ED

Bovidae Bovid size 1-2 Antilopini Bovid size 3a/b Reduncini Equidae Equidae indet Suidae Suidae indet Herbivore indet Indet size 4-5 Primates Primate indet Theropithecus Hominidae Carnivora Carnivora indet Others Aves indet Siluriformes

Level 2

MNE

NU

Level 1 NISP

48

ACCEPTED MANUSCRIPT

Table 3. Types of notches identified at AMK Levels 1 Notch Type*

Nº

Type A

1

Type B

4

Location

Level 1

SC RI PT

LB shaft

Metapod shaft Tibia shaft LB shaft

Axial indet

Type C

3

LB shaft

NU

Tibia shaft

Type D

3

Radius shaft

MA

Tibia shaft

Micro-notch

4

Femur shaft

AC

CE

PT

ED

*Typologies are based on De Juana and Domínguez-Rodrigo (2011)

49

ACCEPTED MANUSCRIPT

Table 4. Distribution of surface marks identified on bovid bones at AMK

Carnivore

1/1 1/2 0/3 2/5 3/4 0/1 2/7 1/1 0/2 3/12

Anthropic Location Fur.

***

1/1 0/2 0/3 0/5 0/4 0/1 0/7 0/1 0/2 0/12

CE

AC

7/26(27%)

PM**** 0/1 0/2 0/3 0/5 0/4 0/1 0/7 0/1 0/2 0/12

DPa/ EPb dist. + prox. EPb + MEPc prox. /DPa DPa DPa/ MEPc prox. DPa

DPa

0/38(0%)

0/2 0/1 0/2 0/1 0/3 0/3 0/4 1/5 0/1 0/2 0/1 1/1

2/2 0/1 1/2 0/1 0/3 1/3 0/4 0/5 0/1 0/2 0/1 0/1

0/2 0/1 0/2 0/1 1/3 0/3 0/4 0/5 0/1 0/2 0/1 0/1

2/26(8%)

4/26(15%)

1/26(4%)

ED

1/2 1/1 1/2 0/1 0/3 0/3 1/4 3/5 0/1 0/2 0/1 0/1

MA

17/38(45%) 13/38(34%) 1/38(3%)

Hum. Rad./Ul. Fem. Tib. Met. Vert. Coxal Rib Mand. Astra. Phal. Indet. Total

*

1/1 2/2 1/3 3/5 3/4 0/1 3/7 1/1 0/2 3/12

Hum. Rad./Ul. Fem. Tib. Met. Vert. Rib Scap. Cran. Indet. Total

Level 2

S

PT

Level 1

**

SC RI PT

P

*

NU

Element

MEPc dist./ EPb dist. + prox./ EPb prox. EPb prox. DPa/ EPb dist. + prox. DPa APd

DPa

P: Pits S: Scores *** Fur.: Furrowing **** PM: Percussion Mark -The numerator indicates the number of elements showing such modifications, and the denominator expresses the MNE in each level -Location of the marks is indicated on the right: a DP: Diaphysis b EP: Epiphysis c MEP: Metaepiphysis d AP: Apophysis **

50

ACCEPTED MANUSCRIPT

n

Mea n

46

1.55

1.2

46

2.71

1.81

28 28

1.7 2.87

1.49 2.2

22

1.2

22

1.6

2

1.4

2

2.05

95% confidence Min interval upper SDa b Maxc 1.2

3.41

3.11 0.33 9.1

1.91 3.54

1.41 0.6 7.25 1.8 1.29 9.2

1.01

1.39

0.44 0.3

2.0

1.41

1.79

0.44 1.0

2.5

-

-

-

0.5

2.3

-

-

-

1.0

3.1

MA

NU

1.9

ED

PT

Hyena shaft breadth* Hyena shaft length* Lion shaft breadth** Lion shaft length** Level 1 shaft breadth Level 1 shaft length Level 2 shaft breadth Level 2 shaft length

95% confidence interval lower

SC RI PT

Table 5. Length and breadth of tooth pits documented at AMK compared to a hyena and a lion samples.

0.21 8.7

AC

CE

* Maasai Mara hyena den in Andrés et al. (2012) ** Lion sample in Andrés et al. (2012) Data include mean values, 95% confidence interval, standard deviation(a) and minimum(b) and maximum(c) values documented in each sample, except for Level 2 where due to sample size only simple observations could be done.

51

ACCEPTED MANUSCRIPT Highlights

The FLK area was a meeting point of herbivores, carnivores and hominins 2 Myr ago. AMK, a contemporary site to the anthropogenic FLK sites was examined.

SC RI PT

Results suggest that AMK might be a natural background scatter created by carnivores.

AC

CE

PT

ED

MA

NU

Factors other than forested habitats favored the formation of anthropogenic sites.

52