![Frequencies of fracture outlines, fracture angles, fracture edges and shaft circumferences for long bone remains (≥2 cm) from the Gran Dolina and Bolomor Cave sites according to the criteria established by Villa and Mahieu [37].](https://www.researchgate.net/profile/Sergi-Lozano/publication/235727138/figure/fig3/AS:340782306349071@1458260241735/Frequencies-of-fracture-outlines-fracture-angles-fracture-edges-and-shaft_Q320.jpg)
Content uploaded by Sergi Lozano
Author content
All content in this area was uploaded by Sergi Lozano
Content may be subject to copyright.
Learning by Heart: Cultural Patterns in the Faunal
Processing Sequence during the Middle Pleistocene
Ruth Blasco
1,2,3
*, Jordi Rosell
1,2
, Manuel Domı
´nguez-Rodrigo
4,5
, Sergi Lozano
1,2
, Ignasi Pasto
´
1,2
,
David Riba
1,2
, Manuel Vaquero
1,2
, Josep Ferna
´ndez Peris
6
, Juan Luis Arsuaga
7,8
, Jose
´Marı
´a Bermu
´dez de
Castro
9
, Eudald Carbonell
1,2,10
1IPHES, Institut Catala
`de Paleoecologia Humana i Evolucio
´Social, Tarragona, Spain, 2A
`rea de Prehisto
`ria, Universitat Rovira i Virgili (URV), Tarragona, Spain, 3The
Gibraltar Museum, Gibraltar, 4Department of Prehistory, Complutense University, Madrid, Spain, 5IDEA (Instituto de Evolucio
´nenA
´frica), Museo de los Orı
´genes, Madrid,
Spain, 6SIP (Servei d’Investigacio
´Prehisto
`rica), Museo de Prehistoria, Diputacio
´n de Valencia, Valencia, Spain, 7Departamento de Paleontologı
´a, Facultad de Ciencias
Geolo
´gicas, Complutense University, Madrid, Spain, 8Centro de Investigacio
´n (UCM-ISCIII) de Evolucio
´n y Comportamiento Humanos, Madrid, Spain, 9CENIEH (Centro
Nacional de Investigacio
´n sobre Evolucio
´n Humana), Burgos, Spain, 10 Institute of Vertebrate Paleontology and Paleoanthropology (IVPP), Beijing, China
Abstract
Social learning, as an information acquisition process, enables intergenerational transmission and the stabilisation of cultural
forms, generating and sustaining behavioural traditions within human groups. Archaeologically, such social processes might
become observable by identifying repetitions in the record that result from the execution of standardised actions. From
a zooarchaeological perspective, the processing and consumption of carcasses may be used to identify these types of
phenomena at the sites. To investigate this idea, several faunal assemblages from Bolomor Cave (Valencia, Spain, MIS 9-5e)
and Gran Dolina TD10-1 (Burgos, Spain, MIS 9) were analysed. The data show that some butchery activities exhibit variability
as a result of multiple conditioning factors and, therefore, the identification of cultural patterns through the resulting cut-
marks presents additional difficulties. However, other activities, such as marrow removal by means of intentional breakage,
seem to reflect standardised actions unrelated to the physical characteristics of the bones. The statistical tests we applied
show no correlation between the less dense areas of the bones and the location of impacts. Comparison of our
experimental series with the archaeological samples indicates a counter-intuitive selection of the preferred locus of impact,
especially marked in the case of Bolomor IV. This fact supports the view that bone breakage was executed counter-
intuitively and repetitively on specific sections because it may have been part of an acquired behavioural repertoire. These
reiterations differ between levels and sites, suggesting the possible existence of cultural identities or behavioural
predispositions dependant on groups. On this basis, the study of patterns could significantly contribute to the identification
of occupational strategies and organisation of the hominids in a territory. In this study, we use faunal data in identifying the
mechanics of intergenerational information transmission within Middle Pleistocene human communities and provide new
ideas for the investigation of occupational dynamics from a zooarchaeological approach.
Citation: Blasco R, Rosell J, Domı
´nguez-Rodrigo M, Lozano S, Pasto
´I, et al. (2013) Learning by Heart: Cultural Patterns in the Faunal Processing Sequence during
the Middle Pleistocene. PLoS ONE 8(2): e55863. doi:10.1371/journal.pone.0055863
Editor: Michael D. Petraglia, University of Oxford, United Kingdom
Received May 28, 2012; Accepted January 4, 2013; Published February 2 , 2013
Copyright: ß2013 Blasco et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The Bolomor excavation is part of the program of archaeological excavations conducted by the SIP (Prehistoric Investigation Service) of the Prehistory
Museum of Valencia under the authority of the Provincial Council of Valencia, Spain. The field excavation work in the Atapuerca sites is supported by Junta de
Castilla y Leo
´n and Atapuerca Foundation. This research was supported with funding from the Spanish Ministry of Science and Innovation, project nos. CGL2012-
38434-C03-01, CGL2012-38434-C03-02, CGL2012-38434-C03-03, CGL2012-38358 and HAR2010-18952-C02-01, and from Generalitat de Catalunya, 2009 SGR 188. R.
Blasco and S. Lozano are Beatriu de Pino
´s post-doctoral research fellowships (Generalitat de Catalunya and COFUND Marie Curie Actions, EU-FP7) and D. Riba is
a pre-doctoral research fellowship from Atapuerca Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or
preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: rblascolopez@gmail.com
Introduction
Archaeological records are composed of multiple individual
actions, which lead to the creation of variability in the
assemblages, either as idiosyncratic features or traits submerged
in repetitive patterns [1]. From this perspective, Binford [2]
provides a scientific frame of reference, championing the system
and the group as units of analysis. Therefore, the processing of
animal resources and storage decisions are analysed as outcomes of
group behaviour [3,4].
Social learning and imitation are information-gathering me-
chanisms based on the experience or behaviour of other
individuals. They allow the existence of an intergenerational
transmission of social type and a stabilisation of cultural forms,
which, in turn, strengthens the behavioural traditions of human
groups [5–12]. Imitation can be considered as an alternative social
learning mechanism, which reflects the characteristics of a specific
demand for different types of cultural products and enables
a highly reliable transmission of the information [13,14]. During
imitation, active communication between the user and the
observational learner is not always necessary, as the reproduction
of an event can be based on passive observation. Observation leads
to the reproduction of an action, the perpetrator of which does not
have to understand its aim a priori; this is known as cognitive
‘‘opacity’’. Understanding the process and its aim is a subsequent
PLOS ONE | www.plosone.org 1 February 2013 | Volume 8 | Issue 2 | e55863
0
result, which culminates in an assimilation of relevant concepts,
causes and consequences that are accepted and eventually learned
by the observer. Cognitive ‘‘opacity’’, implicit in imitation
contexts, slows the process of comprehension and represents
a structural learning ability problem in early forms of socio-
cognitive transmission mechanics, thereby impeding the cultural
reproducibility of new practices that were reliant on it. For
Gergely and Csibra [13], this circumstance meant an evolutionary
pressure on early hominids culminating in the selection of a new
type of socio-cognitive learning mechanism that would ensure the
fast and efficient transmission of information. This new system,
human pedagogy, favoured a new mechanism, so that learning via
imitation would be guided [15,16]. Within this process, commu-
nication through language could favour the transfer of culturally
relevant knowledge in a highly effective way. The use of pedagogy
implies the existence of cognitive resources on the part of both
participants in the communicative process, which guarantees the
selective efficiency of the cultural knowledge.
The anatomical adaptations related to the presence of a highly
efficient oral communication system (anatomical structure of the
basicranium, morphological and metric variation in the hyoid
bone and auditory capacities) seem to be present in the human
fossil material from Middle Pleistocene sites in Africa and Europe,
as well as from Neanderthal specimens [17–19]. In this way, the
hominids of this period would have had the capacities for
repeating actions, assimilating concepts with a particular objective
and for learning from other members of the group who ‘‘knew’’
certain processes or activities. Consequently, it is reasonable to
hypothesise that within a group certain traditions would culminate
at the archaeological level in standardised patterns. These patterns
may be different from those developed by other groups and could
foster the existence of certain group or territorial entities.
According to Enloe [20], present-day hunter-gatherer subsistence
varies in organisational strategy, which should be manifested
culturally in the patterning of faunal remains. From ethnograph-
ical observations, Yellen [21] states that there are culturally
prescribed rules, which have no cross-cultural type of logic behind
them and which might be expected to vary from group to group.
To evaluate this idea, we have studied several faunal assemblages
from Bolomor Cave (Valencia, Spain) and a sample from the
TD10-1 sublevel of Gran Dolina (Sierra de Atapuerca, Burgos,
Spain). Some of these sets show well standardised long bone
breakage patterns, which express the concept of the individual as
a knowledgeable actor, able to influence outcomes through
involvement in a social context. Our endeavour, therefore, is to
use faunal data to place the individual as a member of a group in
the landscape and at their evolutionary stage, showing how bones
carry more information than just dietary knowledge. This fact
could open up another perspective on the Middle Pleistocene
faunal data, which all too often are regarded as too meagre to
answer social questions, and help to interpret occupational
dynamics during the formation of the archaeological sites.
Methodology
In order to study the faunal assemblages from Gran Dolina
TD10-1 and Bolomor Cave, we have followed a methodological
approach developed in the zooarchaeology discipline, with
a special focus on anthropogenic damage produced during the
nutritional phase of carcass use. To assess completeness of the
sample, NR (Number of remains) or NISP (Number of Identified
Specimens), MNE (Minimum Number of Elements), MNI
(Minimum Number of Individuals) and MAU (Minimal Anatomic
Units) with their respective percentages have been calculated
[22,23]. Both Gran Dolina TD10-1 and Bolomor Cave contain
a significant volume of remains that are not taxonomically
identifiable. To include the remains with the identified specimens,
weight categories were established following the criteria in-
troduced by Bunn [24], which were translated into the metric
system and grouped into five animal body size classes.
Surface alterations generated by the hominids are treated at
both macroscopic and microscopic levels. For microscopic study,
an Olympus Europe SZ11 (magnification up to 110) and
Environmental Scanning Electron Microscope -ESEM- (FEI
QUANTA 600) were used. The anthropogenic damage observed
on the faunal remains includes cut-marks, intentional bone
breakage, burning and human tooth-marks. In order to identify
possible standardised processes on bone remains, the location and
distribution of modifications in terms of anatomical area and
region (portion and side) were registered. These bone landmarks
are appropriately standardised for ease of consideration using
a devised numerical system. This system divides the long bones
into five portions, with portion 1 being the most proximal part of
the skeletal element (proximal epiphysis) and portion 5 the most
distal (distal epiphysis). Portions 2, 3 and 4 are located on the
diaphysis, in which portion 2 is the most proximal part (proximal
metadiaphysis), portion 3 is the medial region (mid-shaft), and
portion 4 is the most distal part (distal metadiaphysis).
Cut-marks have been identified according to the criteria
established by Binford [4], Potts and Shipman [25], Shipman
and Rose [26], Bromage and Boyde [27], Shipman et al. [28] and
Noe-Nygaard [29]. Three types of cut-marks have been identified
and grouped into incisions, scrapes and chop-marks. These three
types differ according to the manner in which human groups used
stone tools [23,26,30]. The analysis of cut-marks took into account
the number of striations, location on the anatomical element,
distribution over the surface (isolated, clustered, crossed), orienta-
tion with respect to longitudinal axis of the bone (oblique,
longitudinal, transverse) and delineation (straight or curved).
Measurements of each striation (maximum and minimum length)
were taken in millimeters using a digital calliper and in some cases,
a stereoscopic microscope. The different types of cut-marks have
been associated with specific butchering activities following the
observations carried out by Binford [4,31], Fisher [32], Dom-
ı
´nguez-Rodrigo [33,34], Nilssen [35] and Lyman [23].
Bone breakage is classified following the criteria established by
Bunn [36] and modified by Villa and Mahieu [37]. The outline
(transverse, curved/V-shaped, longitudinal), fracture angle (ob-
lique, right, mixed) and surface edge (smooth, jagged) is recorded.
Bone breakage on the small animal bones was analysed and
classified as old (occurring at or near the time of deposition) or new
(occurring during or after excavation) [38]. This last type was well
defined by colour changes in the section of bone and the outline
and fracture angle.
Surface damage caused during bone breakage was also analysed
and the diagnostic elements of anthropic breakage were
documented on faunal remains. These modifications included
percussion pits or percussion marks, percussion notches or
conchoidal scars, impact flakes, adhering flakes and peeling.
Percussion notches are semicircular shaped indentations on
fracture edges with corresponding negative flake scars [39,40].
Impact flakes refer to positive flakes of the percussion notches and
display the same basic technical attributes as stone flakes (mainly
ventral face with point of detachment and bulb). Percussion pits or
percussion marks are often closely associated with patches of striae
that result from slippage of stone against bone during impact
events [40,41]. Peeling defines a roughened surface with parallel
grooves or fibrous texture produced when fresh bone is fractured
Learning by Heart
PLOS ONE | www.plosone.org 2 February 2013 | Volume 8 | Issue 2 | e55863
and peeled apart, similar to bending a small fresh twig with two
hands [42]. These collected data were compared with published
arguments regarding the revision of criteria for distinguishing
between percussion and carnivore tooth-marks and geochemical
etching and bioerosion [43–46]. In order to observe possible
reiterations, descriptions include the location of the damage on the
anatomical element (portion and side) and the distribution over the
surface. To check the existence of a possible correlation between
location of percussion impacts and bone mineral density, different
bivariant tests (Spearman’s rho and Kendall’s tau) were employed,
taking into account each skeletal element portion (epiphysis,
proximal and distal metadiaphysis and mid-shaft). For calculation
of this, bone density data estimated by Lyman [22] and Lam et al.
[47] have been used. For discrete multivariate data, we have
applied the implementation of the Exact Multinomial Test in
EMT library of the statistical package R and the function
‘‘multinomial.test’’. For testing differences of categorical groups
with factors in contingency tables, we have used the Fisher’s exact
test (FET) (see Text S1 for more details). These tests were
graphically complemented with a correspondence analysis. An
asymmetric graph was used with rows in the principal coordinates
and columns providing the standardized residuals. The ‘‘ca’’ R
library was selected for this purpose.
Burning damage is used here in terms of present/absent and
based mainly on colour changes (mainly from brown, black, and
grey to white). Bones exposed to fire exhibit modifications
differentially, including colour changes, cracking, fractures and
shrinkage (e.g.,[48–53]). The most obvious alterations are changes
in the natural coloration of bone material. As in the case of cut-
marks and percussion marks, the anatomical area and region of
alteration by burning are also registered. We designate degrees of
alteration by burning according to six categories of intensity,
degree 0 being the unburned bones and degree 5 the calcined
remains [54].
Archaeological Approach: TD10-1 of Gran Dolina and
Bolomor Cave
Gran dolina TD10-1. Gran Dolina is a large cavity located
in the Sierra de Atapuerca (Burgos, Spain). Its stratigraphic
succession of up to 18 m high was initially divided into 11
stratigraphic units called TD1 to TD11 from bottom to top, which
was slightly revised in subsequent studies [55–57]. Palaeomagnetic
data placed the Matuyama-Brunhes boundary at the top of level
TD7, which divides the stratigraphic sequence into an Early
Pleistocene section (TD1-2 to TD7) and a Middle Pleistocene
section (TD8 to TD11). The TD10 level is the most recent deposit
at the site with archaeo-paleontological remains. Its sediments are
composed of sands with gravel and limestone clasts [55]. TD10 is
divided into four litho-stratigraphic sub-units: TD10-1 (it includes
TD10-sup), TD10-2, TD10-3 and TD10-4 in the base. A variety
of dating methods (U/Th, ESR, TL, IRSL) have been applied at
the site. Available geochronological studies in TD10 provided a TL
date of 379657 ky for the bottom of TD10-1, and a mean date of
337629 for its top [57–60]. From a technological point of view,
TD10-1 is classified as a transitional moment between Mode 2 or
Acheulean and Mode 3 or Mousterian [60–62]. Flakes, denticu-
lates and side-scrapers are the most common elements. Lithic
refitting related to production sequences indicates mainly short
and incomplete knapping activities at the site [63]. All the used
raw materials (two types of chert, quartzite, quartz, sandstone and
limestone) are found within a 5 km radius of the site [64]. Bone is
occasionally exploited to make artefacts, both directly (bone
hammer) and previously configured (side-scrapers) [65].
Regarding the faunal remains, the sample from TD10-1
analysed here (2000–2001 excavation season) is composed of 22
taxa, with Cervus elaphus,Equus ferus and Oryctolagus sp. as dominant
species. Adult specimens determine the profile based on the age at
death (MNI = 40 adults of 60 individuals) (Table 1). The
proportion of long bone fragments (NR = 5728 of 11081 remains)
is slightly higher than flat bones (NR = 4066/11081). Among the
long bones, the assemblage consisted of mainly long bone shaft
fragments (NR = 3950 of 5728 long bone fragments), which are
not always identifiable taxonomically. From %MAU data, the
skeletal representation of ungulates is characterised by the
abundance of stylopodials (femur and humerus) (61.8), zeugopo-
dials (radius and tibia) (51.8) and mandibles (62.5) and by a low
representation of the axial skeleton (vertebrae and ribs) (8.5) [66].
The elements with the greatest marrow value are those with the
greatest representation. This phenomenon could be interpreted as
the product of anthropogenic selective transport [4,67].
Analysis of bone breakage shows that the curved/V-shaped
predominates, along with oblique angles and smooth edges [66].
The degree of bone breakage at TD10-1 is related to green bone
breakage, according to the criteria established by Bunn [36] and
Villa and Mahieu [37] (Table 2). Diagnostic elements of
intentional anthropogenic breakage are recognised on 1329 bone
fragments. These diagnostic elements are mainly percussion
notches (121) and impact flakes (1166). Percussions notches on
bone fragments identified anatomically show a high diversity in
location (in terms of portion and overall surface) (Figure 1;
Table 3). Actually, the assumption of uniformly distributed notches
could not be discarded for any of the six bones under
consideration after applying Exact Multinomial Tests to the
observed distributions. Metacarpus was the worst-fitted case (p-
value = 0.2045) (Table S1). In cases of small prey, fragmentation
registers as transverse and curved-shaped fracture outlines close to
the ends of limb bones and at oblique and mixed angles. For
several researchers, this type of breakage is usually found in
anthropogenic contexts and corresponds to green-bone fractures
[68,69,70,71,72]. In addition, shaft cylinders (limb bone-shaft
fragments with their full circumference) have been recovered
(NISP = 35 of 113 belonging to stylopodials and zeugopodials)
together with a significant proportion of extremities (NISP = 67/
113).
Cut-marks are documented on 584 bone fragments. These are
mainly located on the remains of large and medium-sized animals.
Although incisions show variability, it is possible to distinguish two
groups based on their type and location (Table S2). Oblique and
longitudinal incisions are mainly situated on limb bone diaphyses,
and transversal sawing marks on areas that present difficulties for
the extraction of soft tissues, such as attachment areas for muscles
or tendons, insertions, crests or tubercles. Short and deep
striations, related to dismemberment and disarticulation of the
anatomical portions, are also identified on some epiphyses. With
regard to skinning, animals are often skinned from the skull to the
metapodials and in some cases up to the second phalanx. Scraping
marks are often related to periosteum removal, although they may
also arise from the excision of meat remnants from bones during
surface preparations for subsequent breakage events. From the
EMT, the p-values (,0.05, so the results differ significantly from
the ab-initio model) indicate patterning for the humerus, femur and
tibia, with cut-marks clustered preferentially on mid-shafts (Text
S2).
Several carnivore remains are also recovered in the analysed
sample from TD10-1: Ursus arctos,Canis lupus,Vulpes vulpes,Panthera
leo fossilis and Lynx sp. Some of these predators were processed by
human groups (a lion and a fox) while others may have been
Learning by Heart
PLOS ONE | www.plosone.org 3 February 2013 | Volume 8 | Issue 2 | e55863
introduced naturally into the cave, or even been brought in by
other carnivores. The overall record of tooth-marks on bones
supports the view that predatory animals frequented the cave in
order to scavenge remains abandoned by human groups and
probably sought refuge or made dens between intervals of human
occupation [66,73,74].
Bolomor cave. Bolomor Cave is located on the southern
slope of the Valldigna valley (Valencia, Spain). The site is
comprised of an elevated rock-shelter approximately 100 m above
sea level. The stratigraphic sequence is divided into 17 levels, with
numbering commencing from the top of the deposit, and with
a maximum thickness of 14 m [75]. The karstic deposition has
been dated by AAR and TL to between MIS 9 and MIS 5e [76].
The lithic industry from Bolomor Cave is classified as a Middle
Palaeolithic techno-complex. This techno-complex is older than
the regional Classic Mousterian age and has its beginning at some
point during the Middle Pleistocene, under the consideration of an
Ancient Middle Palaeolithic, although it is not related to the
Acheulian period [77]. Used raw materials mainly consist of flint
and, to a lesser extent, of limestone and quartzite. As a distinctive
feature, the level IV industry mainly includes small tools,
predominantly scrapers, denticulates and various retouched pieces,
which are characterised by intensive re-use and recycling [78].
Several combustion structures have been documented at levels
II, IV, XI and XIII. The hearths are morphologically simple and
are not superimposed, they have a lenticular appearance with
diameters between 30–120 cm and an average thickness of 5–
10 cm [79].
The Bolomor faunal record includes more than 30 species
belonging to the categories of Cercopithecinae, Carnivora,
Ungulata and small prey (Leporidae, Aves, Testudinidae,
Amphibia and Salmonidae). Cervus elaphus and Oryctolagus cuniculus
are the most represented taxa, along with Aythya sp. at level XI and
Testudo hermanni at level IV. In the age at death profile for the
individuals recovered (MNI), adults clearly predominate
(XVIIc = 23 adults of 30 individuals; XVIIa = 35/38; XI = 24/
30; IV = 83/99) (Table 4, Table 5). The proportion of long bone
fragments (NR) (XVIIc = 550 of 1307 remains; XVIIa = 677/
1732; XI = 409/1047; IV = 16657/25323) is higher than flat
bones (XVIIc = 214 of 1307 remains; XVIIa = 317/1732;
Table 1. NR, NISP, MNE and MNI by ages from the TD10-1 faunal assemblage.
Taxa NR NISP MNE MNI neo inf juv ad sen
Ursus arctos 3321 1
Canis lupus 10 10 6 2 1 1
Vulpes vulpes 16 16 13 2 1 1
Panthera leo fossilis 17 17 15 1 1
Lynx sp. 1 1 1 1 1
Hystrix sp. 2 2 1 1 1
Stephanorhinus cf. hemitoechus 52 52 9 2 1 1
Equus ferus 260 260 62 9 2 3 3 1
Equus cf. hydruntinus 12 12 5 2 1 1
Sus scrofa 1111 1
Cervidae indet. 121 121 24 2 1 1
Megaloceros giganteus ?1111 1
Dama dama clactoniana 2221 1
Cervus elaphus 762 762 232 9 1 1 6 1
Bison sp. 144 144 55 5 1 1 2 1
Hemitragus bonali 5551 1
Capreolus capreolus 3332 1 1
Erinaceidae 11 11 8 1 1
Oryctolagus sp. 329 329 167 12 1 11
Aves, unident. 7 7 2
Passeriformes 25 25 18 1 1
Phasianidae 9 9 8 1 1
Corvidae 17 17 16 1 1
Pisces 1 1 1 1 1
Very large size 101 11
Large size 1432 97
Medium size 4726 202
Small size 2342 159
Very small size 32 13
Unident. 637
Total 11081 1811 1139 60 1 9 7 40 3
doi:10.1371/journal.pone.0055863.t001
Learning by Heart
PLOS ONE | www.plosone.org 4 February 2013 | Volume 8 | Issue 2 | e55863
Learning by Heart
PLOS ONE | www.plosone.org 5 February 2013 | Volume 8 | Issue 2 | e55863
XI = 231/1047; IV = 6286/25323). The Bolomor assemblages
consisted of mainly long bone shaft fragments (XVIIc =369 of 550
long bone fragments; XVIIa = 406/677; XI = 325/409;
IV = 14984/16657). According to %MAU, the assemblages are
mainly composed of cranial elements (mandibles and maxillaries)
(XVIIc = 56.1; XVIIa = 51.6; XI = 51.8; IV = 71.8) and proximal
appendicular bones (XVIIc = 60.4; XVIIa = 60.5; XI = 39.5;
IV = 75.9) in ungulates. The axial skeleton is underrepresented
in all levels (XVIIc = 1.8; XVIIa = 4.8; XI = 2.2; IV = 6). For small
prey, such as lagomorphs, all the skeletal elements are represented,
although hind and fore limbs predominate over all other skeletal
parts at level IV (XVIIc = 44.8; XVIIa = 64.2; XI = 57.1;
IV = 40.6), in addition to scapular and pelvic girdles at level XI
and pelvic girdles at level XVII (scapula: XI = 57.1/pelvis:
XI = 42.9; XVIIc = 47.2; XVIIa = 73.3). Regarding birds, the
anatomical elements with the highest survival are coracoid and
tibiotarsus at level XI and IV (coracoids: XI = 93.7; IV = 63.3/
tibiotarsus: XI = 87.5; IV = 46.7). In the case of tortoises, the
anatomical elements most abundant are humeri (IV = 94.7),
femuri (XI = 50; IV = 73.7) and plastron (XI = 100; IV = 68.4)
[66].
Bone breakage indicates that curved/V-shaped fractures,
oblique angles and smooth edges are predominant in the Bolomor
faunal record (Table 2). According to Bunn [36] and Villa and
Mahieu [37], these patterns indicate that the bone was mainly
fresh when broken. Evidence for anthropogenic bone breakage is
documented on 82 ungulate remains at sublevel XVIIc, 117 at
sublevel XVIIa, 57 at level XI and 839 at level IV. The most
abundant diagnostic elements are percussion notches (XVIIc = 31;
XVIIa = 28; XI = 6; IV = 112) and impact flakes (XVIIc = 48;
XVIIa = 79; XI = 47; IV = 676). Percussions notches on the
identified skeletal elements show a high standardisation in location
(in terms of portion and side) depending on the archaeological
level (Figure 1; Figure 2; Table 3). This phenomenon is especially
pronounced at level IV, which presents the highest number of
remains with diagnostic elements of anthropogenic breakage and
shows distributions statistically different from a uniformly distrib-
uted case (i.e., when Exact Multinomial Tests were applied to each
bone, we obtained a p-value ,0.05 for the radius and p-values
,0.001 for the rest of cases) (Table S1). For instance, the ungulate
humeri (both Artiodactyla and Perissodactyla) from level IV
contain percussion notches on the posterior side of the distal
metadiaphysis in 19 bone fragments of the 21 recovered. Similarly,
the ungulate tibiae show impact points on the posterior side of the
distal metadiaphysis in 14 of the 16 registered remains.
A different case is observed on small prey. No diagnostic
elements of fracturing by active or passive percussion are identified
on their remains. Nevertheless, the fragmentation is present both
on the fore and hind limb in the form of transverse and curved/V-
shaped fracture outlines close to the ends and oblique and mixed
angles. The breakage near the ends is not common in non-
anthropogenic or post-burial contexts [68–72]. Following the
study carried out by Cochard [69], Cochard et al. [70] and
Sanchis Serra [80], the curved/V-shaped fractures are features of
breakage on fresh avian and leporid bones. In order to fracture the
small prey bones, the human groups of Bolomor probably
combined the actions of their hands and teeth. As a result of
such combinatory actions, well-established patterns can be
observed on different skeletal elements in the form of shaft
cylinders (NISP XVIIc = 31 of 132 belonging to stylopodials and
zeugopodials; XI = 15/138; IV = 71/182) and isolated ends (NISP
XVIIc = 95/132; XI = 50/138; IV = 51/182). Following the cri-
teria described by Cochard [69], Laroulandie [71], Pe´rez Ripoll
[72], Sanchis Serra [80], Landt [81] and Lloveras et al. [82],
human tooth-marks can be identified on rabbit, bird and tortoise
remains at Bolomor. These marks are associated with fracture
edges and, in some cases, they form crenulated edges or peeling
[66,83–86].
Figure 1. Graphical representation showing the relationship between the total of recovered bone portions and bone portions with
impact notches for the TD10-1 faunal assemblage and level IV of Bolomor Cave by skeletal elements belonging to the appendicular
skeleton of ungulates. P = Proximal; M = Mid; D = Distal; lat = lateral; med = medial; post = posterior; ant = anterior.
doi:10.1371/journal.pone.0055863.g001
Table 2. Frequencies of fracture outlines, fracture angles, fracture edges and shaft circumferences for long bone remains ($2 cm)
from the Gran Dolina and Bolomor Cave sites according to the criteria established by Villa and Mahieu [37].
Gran Dolina Bolomor
TD10-1 XVIIc XVIIa XI IV
No. Fractures 10081 3391 1756 492 7758
Fracture outline Transverse (%) 1619 (16.1) 659 (19.4) 365 (20.8) 114 (23.2) 1684 (21.7)
Curved/V-shaped (%) 5608 (55.6) 1606 (47.4) 822 (46.8) 253 (51.4) 3681 (47.4)
Longitudinal (%) 2854 (28.3) 1126 (33.2) 569 (32.4) 125 (25.4) 2393 (30.8)
Fracture angle Oblique (%) 5548 (55) 1557 (45.9) 779 (44.4) 235 (47.8) 1934 (24.9)
Right (%) 1985 (19.7) 1062 (31.3) 563 (32.1) 194 (39.4) 1932 (24.9)
Mixed (%) 2548 (25.3) 772 (22.8) 414 (23.6) 63 (12.8) 3892 (50.2)
Fracture edge Smoothed (%) 9079 (90.1) 2910 (85.8) 1501 (85.5) 422 (85.8) 5835 (75.2)
Jagged (%) 1002 (9.9) 481 (14.2) 255 (14.5) 70 (14.2) 1923 (24.8)
Shaft circumference #1/4 2364 299 378 81 1854
1/4–1/2 409 97 94 37 360
1/2–3/4 85 8 7 7 22
$3/4 9
doi:10.1371/journal.pone.0055863.t002
Learning by Heart
PLOS ONE | www.plosone.org 6 February 2013 | Volume 8 | Issue 2 | e55863
Cut-marks are documented on 119 faunal remains at sublevel
XVIIc, 117 at sublevel XVIIa, 79 at level XI and 1817 at level IV.
Striations are mainly identified on the long bones of medium and
large-sized animals, with the majority of the incisions on the
diaphysis, and the sawing marks on the metaphysis (Table S3).
The EMT p-values show no indication of patterning for most long
bones from Bolomor XVIIc, XVIIa and IX (p-value .0.05), in
contrast with the cut-mark distribution on long bones from
Bolomor IV (p-value = 0.000) (Text S2). Whereas this may be
suggestive of random placement of cut-marks on the Bolomor
levels (other than Bolomor IV) it actually has more to do with
sample size. The number of cut-marks reported for Bolomor
XVIIc,a and IX is substantially smaller than those reported for
Bolomor IV. Given that the EMT tests are very conservative, it is
not surprising to detect no pattern when the sample size is small.
In contrast, the larger sample of cut-marks for Bolomor IV
Table 3. Notch distribution patterns on the main long bone (humerus, radius, femur, tibia and metapodials) belonging to
ungulates from the Gran Dolina TD10-1 and Bolomor Cave faunal assemblages.
Gran Dolina Bolomor
TD10-1 XVIIc XVIIa XI IV
NR Portion Side NR Portion Side NR Portion Side NR Portion Side NR Portion Side
Humerus S.cf. hemitoechus 1 2 lat
Equus ferus 2 4 med, lat 1 4 post
M. giganteus 1 4(3) lat 1 3 lat
Dama sp. 1 4 lat 2 4 post
Cervus elaphus 10 2,3,4 ant,post,med,lat 4 4 lat 3 4 med 13 3,4 lat,post
Bison sp. 1 4 lat
Bos primigenius 2 4 post
H. cedrensis 1 4 post
Very large size 1 3 lat
Large size 1 4 post
Radius Cervus elaphus 6 2,3,4 post,med,lat 2 3 ant 1 3 med 1 3 lat 1 3 lat
Bison sp. 3 2,3 post,ant,med
Bos primigenius 1 3 lat
Medium size 1 3 ant 1 4 ant
Femur Equus ferus 1 3 lat 1 3 ant
Cervus elaphus 8 2,3,4 ant,post,med,lat 2 3 med 1 3 ant 1 3 med 2 3 ant,med
Cervidae unident. 1 4 lat
Bison sp. 1 3 med
Bos primigenius 1 3 ant
Small size 1 3 lat 1 3 ant
Tibia Equus ferus 1 2 med 2 3 post,med 4 3,4 post
Cervus elaphus 12 2,3,4 ant,post,med,lat 3 3 med 3 3 lat 1 3 med 9 4 post
Bison sp. 1 2 post
Bos primigenius 2 4 post
H. cedrensis 1 3 post
Small size 1 3 lat
Mtc Equus ferus 13 med
Cervus elaphus 9 2,3 lat,med 2 3 lat 1 3 lat 3 3 lat
Bos primigenius 3 3 ant,med
H. bonali 1 3 lat
Mtt Equus ferus 1 3 lat 1 3 lat 1 3 med/lat
M.giganteus 1 3 lat/med
Dama sp. 1 3 lat
Cervus elaphus 9 3,4 ant,post,med,lat 2 3 lat,post 5 3 lat,med
Bos primigenius 13 med
H.bonali 1 3 lat
Mtc = Metacarpal; Mtt = Metatarsal; ant = anterior or cranial; post = posterior or caudal; med = medial; lat = lateral. Each long bone was divided into fiv e different
portions: proximal epiphysis (1), proximal metadiaphysis (2), mid-shaft (3), distal metadiaphysis (4) and distal epiphysis (5).
doi:10.1371/journal.pone.0055863.t003
Learning by Heart
PLOS ONE | www.plosone.org 7 February 2013 | Volume 8 | Issue 2 | e55863
indicates that patterning exists, since cut-marks are cluster mainly
on mid-shafts instead of on metadiaphyseal sections.
Burning is identified on 645 faunal remains at level XI and
15585 at level IV. This modification is observed on every type of
skeletal element, with a definite predominance of long bones of
medium and small-sized animals at level IV and of vertebrae and
tibiae at level XI. In the case of lagomorphs and birds, the highest
grades of burning on bones with double colouration coincide with
the areas of the skeleton with less muscle (mainly joints of limb
bones). In tortoises, the carapace is the most affected element, and
degree 2 is the most abundant modification type. On the contrary,
degrees 4 and 5 are practically non-existent in the assemblage.
Table 4. NR, NISP, MNE and MNI by ages from the Bolomor faunal assemblages: levels IV and XI.
IV XI
Taxa* NR NISP MNE MNI imm ad sen NR NISP MNE MNI imm ad
Macaca sylvana 1111 1
Carnivora unident. 5 5 4
Ursus arctos 1111 1
Canis cf. lupus 2221 1
Vulpes vulpes 2221 1
Panthera leo 3 32211
Lynx pardina 2221 1
Castor fiber 2211 1
Palaeoloxodon antiquus 44211
Stephanorhinus hemitoechus 332211
Equus ferus 65 6525413 2221 1
Equus hydruntinus 16 16 9 1 1
Hippopotamus amphibius 46 465211
Sus scrofa 115 115 55 5 2 2 1
Megaloceros giganteus 2221 1
Dama sp. 91 9141312 4441 1
Cervus elaphus 647 647 193 12 2 10 55 55 35 4 2 2
Bos primigenius 213 213 63 4 1 3 2 2 2 1 1
Hemitragus bonali 16 16 13 2 1 1
Hemitragus cedrensis 121 121 47 3 1 2
Oryctolagus cuniculus 789 789 440 20 4 16 262 262 150 7 2 5
Passeriforme 25 25 21 2 2
Galliformes 19 19 16 1 1
Phasianidae 24 24 16 2 2
Anas sp. 29 29 25 2 2
Aythya sp. 34 34 28 3 3 202 202 167 8 8
Corvidae 20 20 13 1 1
Pyrrhocorax sp. 6 6 6 1 1
Columba sp. 34 34 25 2 2
Strigidae 1 1 1 1 1
Aves, unident. 17 17 2
Testudo hermanni 526 526 131 19 19 4 4 3 1 1
Bufo sp. 4 4 2 2 2
Pisces 2 2 2 1 1 1 1 1 1 1
Very large size 37 6
Large size 1975 49 16 4
Medium size 10274 116 128 14
Small size 9053 275 247 20
Very small size 304 61 92 8
Unident. 816 9
Total 25323 2864 1689 99 15 83 1 1047 555 428 30 6 24
*Human remains have not been included in this study.
doi:10.1371/journal.pone.0055863.t004
Learning by Heart
PLOS ONE | www.plosone.org 8 February 2013 | Volume 8 | Issue 2 | e55863
Double colouration on different sides is observed mainly on
tortoise bones and, specifically, on the shell. The carapaces show
a greater degree of burning on the dorsal surface than on the
ventral surface in 84.61% of the cases [83].
The presence of carnivores at Bolomor Cave is rare. Neverthe-
less, fossil remains of Ursus arctos,Ursus tibetanus,Canis lupus, Panthera
leo, Lynx pardina,Vulpes vulpes and Meles meles have been recovered.
The tooth pit sizes on bone fragments according to bone type
(cancellous and dense cortical) might be related mainly to small
and medium-sized carnivores. The presence of non-human
predators and their activities may suggest, with some exceptions
(evidence for human processing on fox, lynx and lion at level IV),
sporadic events of non-human predators that act as scavengers of
hominid refuse [66,85,86].
Discussion
Patterns in the Faunal Processing Sequence
Following the time sequence that organises faunal processing
into a hierarchy, four main actions can be distinguished:
procurement, method of transporting the carcass, the processing
and consumption techniques and the subsequent disposal of the
remains. The different segments of the sequence are highly
interrelated, and therefore the first step (the method and type of
acquisition) greatly determines the subsequent exploitation
sequence, thereby influencing the presence of marks on the bones.
Taking into account that this circumstance may introduce
variations or elements of distortion when standardised action
sequences are being analysed in the faunal record, the pro-
curement method should be tackled as a general dynamic within
the assemblage using different elements. In the archaeological
cases presented here, 1) the systematic proportion of skeletal
elements with a high nutritional value, 2) the predominance of
adult animals, 3) cut-marks related to the removal of the viscera
and 4) oblique and longitudinal incisions on the diaphysis of the
limb bones -which are associated with defleshing of large muscle
masses- suggest that the access hominids had to the animals was
mainly primary and immediate (see extensive discussion in [66]).
These practices, nevertheless, might coexist with sporadic events of
secondary access to the carcasses in the case of the TD10-1
sublevel and level IV [66,86].
On this basis, Roebroeks [87] states the view that large-
mammal hunting in the Middle Pleistocene of Europe must have
been a co-operative activity only made possible by verbal
communication and social interaction between individuals. In his
opinion, such co-operation would have involved the exchange of
information, most likely including language, between older and
younger individuals. Thus, if it is possible to establish a standar-
dised behaviour as a result of the existence of communication and
learning for some techniques of hunting, why can this principle not
be extended to the subsequent processing sequence?
The extraction of both the external (skin, tendons and meat) and
internal (fat and marrow) resources of the bones could reflect the
existence of standardisation during the processing of the carcasses.
Table 5. NR, NISP, MNE and MNI by ages from the Bolomor faunal assemblages: sublevels XVIIa and XVIIc.
XVIIa XVIIc
Taxa NR NISP MNE MNI imm ad NR NISP MNE MNI imm ad
Canis cf. lupus 4 441 1
Palaeoloxodon antiquus 2 211 12 2111
Stephanorhinus hemitoechus 8 832111 111 1
Equus ferus 77 77 30 2 2 56 56 22 1 1
Megaloceros giganteus 10 10 9 1 1 8 8 5 1 1
Dama sp. 27 27201 113 13101 1
Cervus elaphus 177 177 58 4 1 3 132 132 47 4 1 3
Bos primigenius 24 24 13 1 1 22 22 13 1 1
Hemitragus bonali 28 28 20 1 1 6 6 6 2 2
Oryctolagus cuniculus 620 620 346 15 1 14 457 457 234 12 5 7
Lepus sp. 3 331 1
Passeriforme 5 5 5 1 1 9 9 9 2 2
Galliformes 8 8 7 2 2
Phasianidae 18 18 14 3 3 10 10 9 2 2
Anatidae 4 4 4 1 1
Anas sp. 16 16 14 2 2
Bufo sp. 1 111 1
Very large size 81
Large size 186 11 219 11
Medium size 364 22 235 16
Small size 160 24 95 11
Very small size 81
Unident. 6 10
Total 1732 1016 595 38 3 35 1307 732 411 30 7 23
doi:10.1371/journal.pone.0055863.t005
Learning by Heart
PLOS ONE | www.plosone.org 9 February 2013 | Volume 8 | Issue 2 | e55863
Learning by Heart
PLOS ONE | www.plosone.org 10 February 2013 | Volume 8 | Issue 2 | e55863
In that case, we could observe at the archaeological level what
Yellen [21] terms ‘‘style’’ in the butchery and consumption of
mammals in the !Kung Bushmen, or what Bunn calls ‘‘illogical
ways’’ (in [88,24]). For Yellen [21], the analysis of faunal remains
can be used as a valuable tool to approach archaeological
questions related to cultural relationships through time and space,
i.e., ‘‘traditions’’.
Cut-marks show the anthropogenic extraction of the external
soft tissues and allow us to deduce the preparation of the carcass
for subsequent consumption. Several studies have sought to
identify butchery patterns among modern hunter-gatherer groups
in order to explain behavioural questions (e.g, [89–93]). There
remains, however, a long-standing debate on these matters,
especially when results of the discussion are also invoked for
interpretations of the early Pleistocene [94,95]. Some authors use
the same data on butchery to infer differentiated conclusions about
anthropic access to the carcasses, which range from passive
scavenging to primary access, passing through mixed strategies of
early, intermediate and late access [96–102]. Some researchers
question cut-mark frequency and distribution as a good way to
establish the pattern of activity on carcasses [100] and others argue
that opposite butchery behaviours may generate similar cut-mark
patterns [103–105]. Additionally, this interpretative difficulty may
be also be increased by the nature of faunal assemblages; i.e. the
anthropogenic accumulations from Pleistocene archaeological sites
are often the product of overlapped activities and/or occupations,
which can involve disruptive processes, such as cleaning, transport
or trampling. The results are palimpsests composed of multiple
singular events that could make the archaeological interpretations
difficult [106]. From this perspective, animals that compose the
assemblages could reflect more than one method of acquisition
and therefore, more than one technique of processing.
Cut-marks must be understood as accidents produced during
the extraction of the external resources [93]. These accidents are
generated when the cutting edge of the lithic tool comes into
contact with the bone surface. Padilla [107] shows by means of
butchery experiments that even highly skilled butchers intention-
ally deflesh carcasses with the aim of minimizing the number of
cut-marks on bones leaving diagnostic traces in specific anatomical
areas. Therefore, this evidence is not intentional and is subject to
different conditioning factors, which appear to have a significant
effect on the defleshing and disarticulation activities. We must take
into account that the animals’ physical, morphological and
physiological characteristics could condition the presence and
reiteration of cut-marks on certain areas of bone. In this way, the
hominids may focus more on areas where there are muscular
insertions and tendons, or where bone morphology prevents the
easy extraction of meat. Such factors could give rise to un-
intentional patterns on the bones during the defleshing process
that seem to be guided not by an individual intention but, as
Binford [3] argued, by animal anatomy. In addition, other
multiple factors may influence butchery processes, such as the
experience of the butcher, prey size, site functionality, seasonality,
ground characteristics and/or available human technology -
including boiling and metal tools- [89,90,92,100,108–110]. These
variables affect behaviours and, in consequence, may produce
a high variability in the resultant cut-marks. In spite of this,
a patterning related to location of cut-marks is detected from the
exact multinomial tests in those archaeological assemblages with
larger samples -Gran Dolina TD10-1 and Bolomor IV- (Text S2).
The reason for this is that, despite the bias in preservation of
various sections (properly taken into account in the ab-initio model),
cut-marks cluster preferably on mid-shafts instead of on metadia-
physeal sections. This pattern may be conditioned by the main
type of access to the carcasses identified at both TD10-1 and
Bolomor IV (primary and early). For Domı
´nguez-Rodrigo and
Pickering [101], these kinds of cut-marks could have resulted from
hominids butchering fully fleshed carcasses, being inconsistent
with other procurement modalities, such as passive scavenging.
Nevertheless, the Fishers exact test (FET) used to compare the cut-
mark distribution in TD10-1 and in Bolomor IV shows significant
differences between them (e.g., p-value = 0.040 for the tibia and p-
value = 0.020 for the humerus). This phenomenon may be related
to uncontrolled conditioning elements or variables involved in
a complex interplay of cultural and non-cultural factors. With
these ideas, we do not rule out the existence of a possible
transmission of information for the development of defleshing,
disarticulation or skinning, rather we note the difficulty involved in
its archaeological identification. The movements made during
these activities could be standardised as a result of learning;
however, the cut-marks could be only registered on those areas
that, due to above-mentioned conditioning factors, facilitate the
contact of the tool with the bone. An example of this phenomenon
is provided by the Nunamiut ‘‘butchery school’’, where the pupil
imitates as precisely as possible the master butcher in order to
learn how to disarticulate the foot of a caribou [3]. During this
process, knowledge is transferred as part of the relationship
between master and pupil and overall, as a set of body techniques
and movements, which are not always registered on bones. For
Lyman [93], understanding the variability in cut-mark frequencies
presents a great deal of difficulty if multivariate interpretative
models are not tackled. According to this author, multiple and
diverse elements can comprise the processing sequence, including
fortuitous variables.
Burning alterations identified on the faunal remains may suggest
the presence of several processes, such as thermal treatment of the
meat, the preparation of bones to facilitate their breakage or the
development of possible cleaning activities. Double colouration on
the same bone surface has been recorded at levels XI and IV of
Bolomor. The presence of these alterations suggests a differential
preservation of the meat at the moment it is exposed to fire. The
least affected areas are those that present a greater quantity of
muscle and a lower degree of cremation, while the most affected
are those that hardly have any tissue attached and therefore reach
the highest degrees of colouration. This phenomenon allows us to
infer that the meat was roasted prior to the removal of the bone.
Despite this, double colouration on different sides is also
documented. Their presence allows us to infer the existence of
other processes, which may not be related to the cooking of the
meat, such as cleaning activities or simply unintended actions that
lead to the burning of the bones once they have been broken. This
situation could hide the existence of patterns in a set and alter their
possible standardisation. In spite of this, at times, it is possible to
establish systematisation in roasting based on different degrees of
coloration according to the sides. This is the case for shells of
Testudo hermanni recovered at level IV of Bolomor [83]. Although
there is a certain degree of variability, these alterations seem to
describe a pattern based on the differential burning of the bone
Figure 2. Graphical representation comparing the relationship between the total of recovered bone portions and bone portions
with impact notches for the XVIIc and XVIIa faunal assemblages from Bolomor Cave by skeletal elements belonging to
appendicular skeleton of ungulates. P = Proximal; M = Mid; D = Distal; lat = lateral; med = medial; post = posterior; ant = anterior.
doi:10.1371/journal.pone.0055863.g002
Learning by Heart
PLOS ONE | www.plosone.org 11 February 2013 | Volume 8 | Issue 2 | e55863
surface (with a greater degree of burning on the dorsal surface than
on the ventral surface). This phenomenon has been interpreted as
a result of cooking these animals before consumption. The
characteristics of double colouration indicate that the tortoises
could have been placed into the fire upside down. This pattern has
been described ethnographically by Werner [111] for the
preparation of tortoises among the Kayapo´ of Central Brazil.
For several authors, these modifications represent elements that
are diagnostic for the human consumption of tortoises [112–115].
After consideration of the different actions related to processing
and consumption of external resources, bone breakage is perhaps
the activity that best allows us to assess the presence of patterns in
a Middle Pleistocene assemblage. The development of this process
at the habitat places and the relative abundance of bone remains
with elements that are diagnostic for this activity allow us to tackle
this issue with guarantees. The ethnographic observations of
Yellen [21,88,116] confirm that random is not the rule and that
each limb bone is treated following a standardised process.
However, and despite that the use of actualistic analogies can play
a significant role in any attempt to understand archaeological
evidence, comparative analyses between communities of modern
hunter-gatherers and Pleistocene assemblages must be carried out
with caution, given the evident differences related to technology
and/or social systems. Gifford-Gonzalez [108] notices that the
recent incorporation of cook-pots and boiling technology in some
groups of current hunter-gatherers may substantially change the
bone breakage patterns, which seem to adapt to items such as the
size of container. Nevertheless, some activities related to the
extraction of internal resources, such as the release of bone grease
contained within spongy tissues, may be considered [117].
According to Outram [118], the systematic exploitation of marrow
and grease by hitting, crushing or grinding, generates assemblages
characterized by an almost total absence of elements with
cancellous tissue (mainly epiphyses and vertebrae) and the
presence of some diaphyseal cylinders, which could be turned
into splinters by subsequent taphonomical processes. In our case
study assemblages, appendicular epiphyses (NR TD10-1 = 592
whole and/or fragmented epiphyses of 5728 long bones;
XVIIc = 62/550; XVIIa = 97/677; XI = 31/409; IV = 247/
16657) and vertebrae (NR TD10-1 = 381; XVIIc = 26;
XVIIa = 41; XI = 24; IV = 198) are recovered. In spite of their
presence, axial elements show an underrepresentation in some
animal body size classes (large and medium-sized ungulates at
TD10-1 and large, medium and small-sized at Bolomor). This
phenomenon has been interpreted as the product of human
selective transport [4,66,67,86]. But, to assess correctly the
presence of epiphyses and axial bones in the anthropogenic
Pleistocene assemblages, we must take into account the existence
of post-depositional processes, the use of bone as fuel and,
especially, the activities generated by carnivores during the
secondary accesses. The smell of human refuse can attract many
predators, which seem to act as scavengers in search of potentially
consumable resources [4]. Several observations and experimental
reproductions focused on this matter document a predilection for
epiphyses of limb bones and elements of the axial skeleton (e.g.,
[119,120]). The carnivore damage on these anatomical portions is
so intense that in many cases they may make them disappear. In
contrast, the diaphyses, which are generally highly fractured
during the human processing and consumption, show little
alteration. In the case of TD10-1 and Bolomor, carnivore damage
is focused mainly on epiphyses, vertebrae and basipodials
belonging to the most abundant animals, especially at sublevel
XVIIc with a percentage of 72% (TD10-1 = 78 of 454 tooth-
marked bones; XVIIc = 21/29; XVIIa = 12/28; IV = 41/142).
This phenomenon may have altered the initial composition of the
faunal assemblages generating a bias on specific anatomical
portions. Taking into account the above-mentioned factors, bone
breakage at Gran Dolina TD10-1 and Bolomor seems to
correspond more to a strategy focused on marrow removal than
to a repetitive exploitation of grease by crushing and/or grinding.
The intentional extraction of marrow tends to follow standard-
ized patterns on ungulate long bones in some levels of Bolomor.
Unlike other processing activities, this systematisation does not
appear to respond to physical or morphological conditioning
factors of the bone. According to our application of the
Spearman’s rank correlation test, the bone density of the
anatomical portions by species and the location of percussion
notches on identified skeletal elements are positively correlated
(TD10-1: r= 0,42587, p-value = 0.01895; IV: r= 0,46444, p-
value = 0.009721), i.e. the bones do not seem to be intentionally
broken in the less dense zones (Table S4). In spite of this, there is
a significant contrast of the marrow with fat availability in equids
and cervids which may entail different techniques for the
extraction of internal resources regardless of bone density. Binford
[3] noted that the mechanical extraction of caribou marrow from
metapodials was determined by the need to open up the
diaphyseal tube without imbedding the marrow with bone
fragments. This can be contrasted with the relatively small
marrow cavity of equid metapodials, which are difficult to break
and stingy in yield compared to the amount of grease held in the
cancellous tissue of the thin cortex proximal and distal ends [121].
The fracture of these two taxa may be different while achieving
nutritional goals, and thus, the impacts on dense vs. less dense
zones may be fundamentally different. However, if we observe the
impact distributions by taxa at level IV of Bolomor, the notch
distribution pattern by skeletal element tends to be the same
regardless of species. Both artiodactyls and perissodactyls show, for
example, the same standardized damage on the posterior side of
distal metadiaphysis in tibiae. Thus, a guide for breaking based on
the bone density differences by taxa cannot be suggested at least in
the cases presented here. Besides, if there existed a guide for every
particular species, a high diversity in the notch distribution among
the remains of the same taxa would not be registered in the case of
TD10-1 (serve as an example humerus of Cervus elaphus) (see
Table 3).
To check other morphological conditioning factors, two
experimental series were conducted under conditions of isolation,
in which none of the individuals could see how bones were broken
by other members participating in the experiment. Our attempt
was to reproduce a context in which there was no knowledge
transmission in between non-trained experimenters (Text S3). In
cases where humeri were broken by hammerstone percussion
(experimental series 1), the selections made by each individual
experimenter concerning the same impact location may have
arisen from their implementation of intuitive parameters, i.e. the
medial side of humerus is flatter than the curvy lateral side and is
more apt to stabilize the shaft prior to impact. Likewise, the
proximal lateral shaft exposes a wider area for impact (and thinner
than the distal shaft) which is ideal for bone breakage. However,
humeri from level IV of Bolomor show a different preferred
selection for impact, which is located on the distal posterior
metadiaphysis (Text S3; Figure 3). This notch distribution pattern
seems to follow counter-intuitive parameters not related to bone
morphology, since humeri show a thicker cortex anterior-poster-
iorly [122]. But at this point, the question arises whether the use of
another technique could generate similar patterns to those
observed at Bolomor IV and in turn, different from those caused
by hammerstone percussion. The EMT results from experimental
Learning by Heart
PLOS ONE | www.plosone.org 12 February 2013 | Volume 8 | Issue 2 | e55863
series 2 show that notches produced when the bone is hit directly
against a stone object or an anvil are patterned (p-values = 0 for all
long bones). This type of bone-breaking process makes experi-
menters select the impact area by judging the ergonomics of the
bone; i.e. the subjects tend to hold the bones by the narrower zone
and hit by the wider area depending on the skeletal element.
Similar reiterations in the spot location were experimentally
documented by Peretto et al [123] during the processes of
percussion by batting, which reassert the importance of the
morphological factors when this technique is used. However, the
FET outcomes display that these damage distribution patterns do
not coincide with those from Bolomor IV, discarding the
functional convergence in this case and suggesting that such
patterns may have been behaviourally-induced (Text S3; Figure 4).
This idea does not exclude the possibility that mechanically
efficient patterns can be the result of acquired knowledge in bone
breakage methods. Different factors may condition the human
reasons that guide the processing of faunal resources, amongst
which, bone morphology cannot be discarded. The archaeological
problem in such cases lies in differentiating when a functional
convergence phenomenon exists or when standardized damage is
the result of information-transmission mechanisms. In the case of
Bolomor IV, this question is cleared up, since the reiterations of
impacts do not coincide with intuitive, functionally effective or
potentially convergent ones.
Different forms of processing can be observed ethnographically
depending on the group. Examples of standardised actions to
break the bones have been observed in present-day hunter-
gatherers, such as the !Kung [21] or the Nunamiut [3,20], for
example. In all cases, there are previous traditions that are fixed in
the memory of the group, which are repeated through time
without the need to be checked. This phenomenon could be
thought of as ‘‘learning by heart’’, i.e., learning something so well
that it can be repeated mechanically without thinking. In the bone
breakage process, these behaviours seem to cause the majority of
the patterns. A noteworthy case is observed by Yellen [21] in the
breakage of the kudu radius among the !Kung. The young
Bushmen have learned from older individuals that the heads of the
kudu radius do not contain enough marrow and therefore, this
anatomical segment is treated in a different way. This conception
is a subjective and unchecked judgement (inherited concept), but,
nevertheless, is learned and repeated.
The consequences of a ‘‘learning by heart’’ procedure limit the
acquisition of innovative skills without breaking customs, and
demands further empirical and theoretical insights. For instance,
the phenomenon of apprenticeship has been observed in the lithic
industry both ethnographically [124] and at the archaeological
level in several Upper Palaeolithic sites, such as Geissenklo¨ sterle
[125], Gough’s Cave [126], Marsangy [127,128] and Old-
eholtwolde [129], or even in the Chatelperronian site of the
Grotte du Renne at Arcy-sur-Cure [130]. However, these learning
processes can be tracked almost as far back as the Middle
Pleistocene. Significant examples are documented at site K of
Maastricht-Belve´de`re [131] or at several localities of Rhenen
[131,132], where some failed lithic artefacts have been thought to
be pieces created by children during their learning processes.
Bone Breakage Patterns as an Element to Identify
Occupational Dynamics
Understanding what type of occupation existed at Pleistocene
sites is a complex issue, especially if we take into account that the
formation processes of the archaeological levels can mix up
occupational events or make them indistinguishable. Nevertheless,
some elements can help us to define the main types of settlement
using the faunal approach, such as the volume of archaeological
material in relation to the rate of sedimentation, the degree of
intervention of carnivores, species diversity and variety of pro-
curement methods, and the hearths and spatial distribution,
among others. Many of these elements cannot be individualised,
since they could characterise both short and long-term occupa-
tions. For instance, continuity in specific human occupations could
culminate in greater control over the territory and, therefore,
greater amplitude in the exploitation of faunal resources, which
would result in high taxa diversity in the archaeological record.
However, we must take into account that short-term settlements
can also record a wide variety of taxa. This is because they often
represent stops along the way, at which the animals of the
immediate surroundings are exploited without any selection
criteria [133]. The accumulation of different short-term events
in archaeological sites with low sedimentation rates could lead to
a palimpsest characterised by species diversity. This diversity in the
taxonomic profile could be similar to that observed in occupations
of a longer duration. Another ambiguous element is the presence
of anthropic combustion structures. For some authors, the
presence of hearths is evidence of the residential character of the
settlements, regardless of their duration [4,134–138]. At the
ethnographical level, some groups of hunter-gatherers build
hearths not only at base camps, but also in places that are briefly
occupied [3,116]. In line with this evidence, the combustion
structures documented at archaeological sites could represent
either important occupational continuance or a short-term
settlement. For this reason, their relationship with other elements
becomes essential in order to determine the general dynamics of
human occupations during the formation of an archaeological
level. In this regard, the presence or absence of bone breakage
patterns within an assemblage represents an important exception.
The presence of this factor can help us to infer the existence of
a cultural tradition in the area and, therefore, a pattern of
relatively stable territorial occupation. In this sense, a search for
breakage patterning should be considered a valid addition to
discussions about human occupational dynamics.
Since members of a group tend to repeat actions, either through
learning or due to guided imitation of observed actions in others
[3,13,21,88], the systematisation of faunal processing enables
inferences to be drawn concerning the prolonged presence of
a particular group in a site, or the persistence of a shared cultural
tradition in the locality, which may be related indirectly to
a territorial stability. On this basis, the low standardisation
documented in bone breakage damage at TD10-1 could imply not
only the existence of various groups in the area but also the brief
character of their occupations. If a group inhabited Gran Dolina
for a long time period, a higher number of repeated actions should
be identifiable in the faunal record. On the other hand, a lack of
standardisation could represent the intercalated occupations of
different groups with different ways of processing carcasses during
the formation of the assemblage (repeated short-term occupations).
From this zooarchaeological perspective, it is not yet possible to
establish a cultural tradition among the hominids that inhabited
Gran Dolina. With this assumption, we do not state that the
human groups of TD10-1 were not able to develop patterns as
a result of learning, but that the general dynamic of the human
occupations during the assemblage formation prevents the
observation of such phenomena.
The opposite case can be observed at level IV of Bolomor,
where a reiteration has been identified as sufficiently high at the
quantitative level to consider systematisation when bone-marrow is
intentionally extracted. This fact allows us to propose the existence
of groups with the same way of proceeding, making it possible to
Learning by Heart
PLOS ONE | www.plosone.org 13 February 2013 | Volume 8 | Issue 2 | e55863
establish a cultural tradition during the formation of this
stratigraphic deposit. But, unlike TD10-1, the existence of highly
standardised bone breakage patterns does not solve the question
related to the durability of the settlements. In the level IV case,
which can also be extended to the XVIIc sublevel, the presence of
patterns raises at least two possibilities: 1) the cave was inhabited
by human groups with shared cultural traditions during repeated
but short-time periods or; 2) the cave was used by human groups
with shared cultural traditions during long-time periods. On this
basis, the archaeological line that differentiates the types of
settlement is diffuse and therefore, we must resort to other
elements or disciplines that can complement the data provided by
bone breakage patterns. From this approach, many different
processes related to both human activity and the action of
carnivores or raptors occur during the formation of the
archaeological assemblages. The actions that these non-human
predators generate on the faunal set, either by modifying or adding
elements (e.g., coprolites, transporting specific skeletal elements),
are an essential tool for inferring the existence of periods of human
abandonment of the site. The shortage of carnivore damage at
level IV (0.5%; NISP = 142; Total NISP = 25323), together with
the presence of standardised bone breakage patterns, may be used
to back up the hypothesis related to several long-term human
occupations during the formation of this level. In contrast, at
TD10-1, the presence of damage is higher (4%; NISP =454; Total
NISP = 11081). In addition, some characteristic elements of
carnivore dens, such as digested bones, pitting or diaphyseal
cylinders, have been recovered. These data allow us to suggest the
Figure 3. Multiple correspondence analysis showing the distances between location of notch types and the cases comprising the
experimental series 1 (hammerstone percussion) and archaeological samples (data from Text S3; see Statistical methods in Text S1
for more details). Individual characters (a–h) correspond to the 8 individuals involved in the experimental series 1 (each subject is represented by
an alphabetic letter). Coupled characters indicate bone region -side and portion- (e.g., ap = anterior, proximal metadiaphysis; mm = medial mid-shaft;
pd = posterior, distal metadiaphysis).
doi:10.1371/journal.pone.0055863.g003
Learning by Heart
PLOS ONE | www.plosone.org 14 February 2013 | Volume 8 | Issue 2 | e55863
existence of intermittent periods of human occupation alternating
with brief intrusions of carnivores during the formation of TD10-
1. On this basis, the low sedimentation rates at the bottom of the
sublevel may have helped the overlapping of several types of
occupations, producing accumulations that are apparently uni-
form at the archaeological level [139].
Both TD10-1 and Bolomor level IV seem to correspond to
extremes within the occupational dynamics. However, between
a short-term human occupation and a long one, there exists a wide
range of potential settlement types [140–146]. An example of
intermediate occupations could be suggested at sublevel XVIIa.
Although in this assemblage some repetitions have been registered,
variability is one of the main characteristics in the extraction of the
internal resources from the bones. This variability does not reach
the degree of diversity identified at TD10-1 or the standardisation
documented at level IV or sublevel XVIIc. Therefore, it is possible
that XVIIa shows an intermediate dynamic or that it responds to
an accumulation mainly generated by short-term human occupa-
tions with the odd event of certain stability, although not
excessively long. At this point, we must note that our interpreta-
tions attempt a general theoretical explanation of the nature of the
assemblages and it is possible that within the dynamics outlined,
other sporadic events may exist, causing exceptions and distor-
tions. In addition, both the XVII (a/c) level and especially the XI
level contain relatively few bones with percussion notches and,
therefore, the breakage patterns should be interpreted with
caution. It must be noted that not all fragments can be attributed
to a skeletal element, and this is a very common problem when
Figure 4. Multiple correspondence analysis showing the distances between location of notch types and the cases comprising the
experimental series 2 (percussion by batting) and archaeological samples (data from Text S3; see Statistical methods in Text S1 for
more details). Individual characters (a–f) correspond to the 6 individuals involved in the experimental series 2 (each subject is represented by an
alphabetic letter). Coupled characters indicate bone region -side and portion- (e.g., ap = anterior, proximal metadiaphysis; mm = medial mid-shaft;
pd = posterior, distal metadiaphysis).
doi:10.1371/journal.pone.0055863.g004
Learning by Heart
PLOS ONE | www.plosone.org 15 February 2013 | Volume 8 | Issue 2 | e55863
dealing with bone splinters. Unidentified fragments most likely
derive from fore and hind limb shafts, but their recognisability is
a critical problem [20].
The case of small prey also needs mentioning, because its
breakage patterns appear to constitute an exception. We consider
small prey here to be species of less than 10 kg (rabbits, hares,
birds, etc.). Within this rubric, small game at Gran Dolina TD10-1
and Bolomor does not present the same diagnostic elements of
breakage as those observed on ungulate remains; i.e. small prey
bones show no diagnostic features of fracturing by active or passive
percussion such as percussion pits, percussion notches, impact
flakes or counterblows. Nevertheless, standardized morphotypes
are observed on their bones, overall on the fore and hind limbs of
rabbits. These patterns are the result of the isolation of the
epiphyses or epiphyses with part of the metadiaphyses in
stylopodials and zeugopodials. This almost systematic process of
separation gives rise to a large number of diaphyseal fragments
that include from one to all four sides of the bone (diaphyseal
cylinders).These well-established morphotypes are generated when
the limb bones are broken using the teeth, hands or stone
hammers on epiphyses or metadiaphyses. At both sites, shaft
cylinders have been recovered together with a high proportion of
extremities, which could be related to the use of the bite and/or
flexion or the combination of both actions, rather than hammer-
ing. For this reason, the zone methods used to analyse impacts on
large mammal bones are not appropriate for small prey remains.
In Bolomor, the same morphotypes are repeated in a more or less
standardised way throughout the whole sequence, irrespective of
the type and intensity of the occupation. Based on this, tackling the
occupational dynamics from the search for patterns in these
animals is not possible. This phenomenon could represent a priori
a theoretical contradiction with our proposition that standardised
processing sequences may form a characteristic feature of a group.
However, we must take into account that significant differences
exist between ungulates and small prey that enable the
individualisation of their sequences of exploitation and distinctions
to be made in their handling [88]. Therefore, the same criteria
cannot be applied or assumed when establishing patterns of action
on these animals. Distinguishing characteristics, such as the size
and the use of hands or teeth as the main tool for the immediate
consumption seem to condition the appearance of these
standardised morphotypes on the bones, regardless of the site
location or chronology. The conditioning factors of these animals
allow the systematization to be recorded in numerous periods and
sites, such as at level 4 of the French Mousterian site of Canalettes
[69,70], at different sites of the European Upper Palaeolithic
(e.g.,[72,147–155]) and even among some present-day hunter-
gatherer groups, such as the Ache´ of Paraguay [156].
The transmission of knowledge as an explanation for the
standardization of the archaeological record can be identified in
other behavioural domains. In fact, it is worth analyzing whether
a correlation exists between standardization in bone breakage
patterns and, for example, knapping strategies and tool
manufacturing. In lithic assemblages, standardized morphological
patterns are well-known since the appearance of handaxes in the
Early Pleistocene sites. It has been suggested that Acheulean
assemblages indicate the existence of well-structured strategies for
knowledge transfer between generations [157–159], which would
explain the wide spatial and temporal distribution of handaxes.
Some ethnographic references suggest that manufacturing bifacial
morphologies similar to Acheulean handaxes require long and
complex learning processes [160]. The chronological span that
includes the assemblages analyzed in this paper coincides with the
emergence in Europe of the Levallois method [161], which is
based on a well-defined volumetric conception of the core oriented
to predetermine blank shape and/or size. This volumetric design
results in a great standardization of core morphology, which would
be difficult to explain if knowledge transfer processes were absent
from groups. The preconceived character of this knapping method
clearly distinguishes the Levallois assemblages from those associ-
ated with expedient strategies, which are simply aimed at reducing
the core in a recurrent way, without a predetermination of the
products. Therefore, the ability to adjust the technical behaviour
to transmitted normative patterns seems to be fully acquired in the
late Middle Pleistocene. However, when we talk about handaxes
and Levallois cores we are referring to major technological
categories that are represented in the archaeological record for
hundreds of thousands of years throughout much of the Old
World. We need a more fine-grained analysis of technical
variability to identify more discrete spatial and temporal units,
showing standardized patterns specific to certain regions or sites,
to approach social entities such as those inferred from level IV of
Bolomor Cave. Identification of small-scale standardized patterns
in both subsistence and technology opens an interesting avenue for
future researches.
Conclusions
Bone damage generated during the extraction of external and
internal resources can be used to assess the existence of systematic
activities or standardised processes among human groups. In the
case of the TD10-1 sublevel from Gran Dolina and the XVII, XI
and IV levels from Bolomor Cave, the location, disposition and
distribution of modifications in terms of anatomical area and
region (portion and side) have been used to observe possible
human reiterations on faunal assemblages. The development of
some butchery activities, such as defleshing, appears to be
conditioned by several factors (e.g., method of acquisition, the
animal’s anatomy, prey size, the experience of the butcher, site
functionality, seasonality, ground features and/or available human
technology) and therefore, hinder the archaeological identification
of certain cultural processes within the processing sequence. A
different case appears to arise in bone-marrow extraction through
the intentional breakage of the bones that make up the analysed
sets. Statistical tests used in this study show no correlation between
the less dense areas and the localisation of the impacts, i.e., there
appear to be no guidelines based on the density of certain areas of
the skeletal element when fracturing the bones. Our EMT results
from experimental series show that the morphological factors do
not seem to condition the repeated selection of impacts for
breaking open the shafts. Level IV of Bolomor gives an example of
a counter-intuitive preferred selection for impacts. The repetition
of impact points appears to be the result of more complex
processes related to the transmission of intergenerational in-
formation within each group (cultural patterns). On this basis, the
hominids of the European Middle Pleistocene appear to be
capable of developing learning mechanisms that culminated in
their own cultural traditions, which are different to those
developed by other communities, suggesting the existence of
a certain group or territorial entity. The existence of these
traditions within the groups and their reflection at the archaeo-
logical level in standardised patterns can significantly contribute to
interpreting the occupational dynamics in the territory during the
formation of the assemblages. Thus, the low standardisation
documented at TD10-1 could be the result of the presence of
various different groups with different ways of processing the
carcasses. The opposite case can be observed at level IV of
Bolomor, where a reiteration has been identified that is high
enough at the quantitative level to consider systematisation in
Learning by Heart
PLOS ONE | www.plosone.org 16 February 2013 | Volume 8 | Issue 2 | e55863
marrow removal. This fact could suggest the presence of relatively
lengthy occupations that would share this behaviour, making it
possible to establish a cultural tradition during the formation of the
set. Thus, a different debate about the significance of patterning is
possible from the faunal assemblage. The identification of these
intra-group information transmission processes allows us to suggest
the existence of a high social component and perhaps specific
socialisation processes in which learning represents an important
element for the continuity over time of the human groups of these
chronologies. This social cohesion would also be essential for the
successful development of several hunting strategies, particularly
those involving large ungulates or those entailing the practice of
complex techniques. Finally, and as a future prospect, the
methodology developed here to assess occupational dynamics
should be set out within a broader general context, at both
geographical and chronological level. In the same way, the
interpretations of this paper must be viewed as a starting point.
These ideas should be contrasted at other sites and with other
disciplines with the aim of determining different dynamics and,
therefore, being able to assess the different subsistence strategies
developed by the human groups in other spaces and environments,
contributing to the knowledge of the ways of life of the hominids of
the European Middle Pleistocene.
Supporting Information
Table S1 Results from the Exact Multinomial Test
(EMT) for sublevel TD10-1 of Gran Dolina and level IV
of Bolomor Cave. Starting from the ab-initio hypothesis that
impacts should be uniformly distributed across bone parts,
probabilities in the tested null-model have been adjusted. See
Text S1 or http://rgm2.lab.nig.ac.jp/RGM2/func.
php?rd_id = EMT:multinomial.test for more details.
(XLS)
Table S2 NISP with cut-marks, location, morphology
and performed activity by skeletal elements, taxa and
weight size category from TD10-1. Abbreviations:
Cm = Cut-marks; Inc = incisions; Saw = sawing marks;
Scr = Scrape marks; Sk = Skinning; Df = Defleshing; Da = Disarti-
culation; Dm = Dismembering; Vr = viscera removal; Pr = Perios-
teum removal; Tr = Tendon removal.
(XLS)
Table S3 NISP with cut-marks, location, morphology
and performed activity by skeletal elements, taxa and
weight size category from the archaeological. Abbrevia-
tions: Cm = Cut-marks; Inc = incisions; Saw = sawing marks;
Scr = Scrape marks; Sk = Skinning; Df = Defleshing; Da = Disarti-
culation; Dm = Dismembering; Vr = viscera removal; Pr = Perios-
teum removal; Tr = Tendon removal.
(XLS)
Table S4 Bone density data estimated by Lyman [22]
and Lam et al. [47], NISP with percussion notches and
correlation coefficient by means Spearman’s rho and
Kendall’s tau from TD10-1 and Level IV.
(XLS)
Text S1 Statistical methods used to approach pattering
in the archaeological and experimental samples.
(DOC)
Text S2 Results from the Exact Multinomial Test
(EMT) and Fishers exact test (FET) applied to each site
and level to approach the cut-mark distribution.
(DOC)
Text S3 Experimental series and its comparison with
the archaeological cases of Gran Dolina TD10-1 and
Bolomor Cave.
(DOC)
Acknowledgments
We acknowledge all of the members of the Bolomor and Atapuerca
research teams involved in the recovery and study of the archaeological,
geological, and paleontological record. We are also deeply grateful to John
Yellen for helping us with the ethnographical references. We would like to
thank the Editor, Michael D. Petraglia, and three anonymous reviewers for
their comments on the manuscript that have greatly improved the final
version. We thank the Tritons’ team and the colleagues for conducting the
bone breakage experiments reported on here, especially Maite Arilla and
Jordi Fa`bregas for their comments and very useful help during the series.
Thanks to Elena Petrova for editing the text and Mick Vernon for his in-
depth reading of the manuscript.
Author Contributions
Social learning mechanisms: IP DR. Statistical analysis: MDR SL. Lithic
technology: MV JFP EC. Excavation project directors: JFP JLA JMBC EC.
Conceived and designed the experiments: RB JR MDR. Performed the
experiments: RB JR. Analyzed the data: RB. Contributed reagents/
materials/analysis tools: RB JR MDR SL. Wrote the paper: RB JR.
References
1. Gamble C, Gaudzinski S (2005) Bones and powerful individuals Faunal case
studies from the Arctic and the European Middle Palaeolithic. In: Gamble C,
Porr M, editors. The hominid individual in context: archaeological investiga-
tions of lower and middle Palaeolithic landscapes, locales, and artefacts. New
York: Routledge. 154–175.
2. Binford LR (2001) Constructing Frames of Reference: An Analytical Method
for Archaeological Theory Building Using Ethnographic and Environmental
Datasets. Berkeley: University of California Press. 563 p.
3. Binford LR (1978) Nunamiut Ethnoarchaeology. New York: Academic Press.
509 p.
4. Binford LR (1981) Bones: Ancient Men and Modern Myths. New York:
Academic Press. 320 p.
5. Tomasello M, Kruger AC, Ratner HH (1993) Cultural learning. Behav Brain
Sci 16: 495–552.
6. Tomasello M (1999) The cultural origins of human cognition. Boston: Harvard
University Press. 248 p.
7. Tomasello M, Carpenter M, Call J, Behne T, Moll H (2005) Understanding
and sharing intentions: The origins of cultural cognition. Behav Brain Sci 28:
675–691.
8. Dennett D (1995) Darwin’s Dangerous Idea: Evolution and the Meanings of
Life. New York: Simon and Schuster. 592 p.
9. Meltzoff AN (1996) The human infant as imitative generalist: A 20-year
progress report on inf ant imitation with impl ications for comparative
psychology. In: Heyes CM, Galef BG, editors. Social learning in animals:
The roots of culture. New York: Academic Press. 347–370.
10. Byrne RW, Russon AE (1998) Learning by imitation: A hierarchical approach.
Behav Brain Sci 21: 667–721.
11. Blackmore S (2000) The power of memes. Sci Am 283: 52–61.
12. Whiten A (2000) Primate culture and social learning. Cogn Sci 24: 477–508.
13. Gergely G, Csibra G (2006) Sylvia’s recipe: The role of imitation and pedagogy
in the transmission of cultural knowledge. In: Enfield NJ, Levenson SC, editors.
Roots of Human Sociality: Culture, Cognition, and Human Interaction.
Oxford: Berg Publishers. 229–255.
14. Tennie C, Call J, Tomasello M (2009) Ratcheting up the ratchet: on the
evolution of cumulative culture. Phil Trans R Soc B 364: 2405–241.
15. Csibra G, Gergely G (2006) Social learning and social cognition: The case of
pedagogy. In: Munakata Y, Johnson MH, editors. Progress of Change in Brain
and Cognitive Development. Attention and Performance, vol. XXI. Oxford:
Oxford University Press. 249–274.
16. Csibra G, Gergely G (2011) Natural pedagogy as evolutionary adaptation. Phil
Trans R Soc Lond B 366: 1149–1157.
Learning by Heart
PLOS ONE | www.plosone.org 17 February 2013 | Volume 8 | Issue 2 | e55863
17. Martı
´nez I, Arsuaga JL, Quam R, Carretero JM, Gracia A, et al. (2008)
Human hyoid bones from the middle Pleistocene site of the Sima de los Huesos
(Sierra de Atapuerca, Spain). J Hum Evol 54: 118–124.
18. Martı
´nez I, Arsuaga JL (2009) El origen del lenguaje: la evidencia
paleontolo´gica. Munibe 60: 1–16.
19. Martı
´nez I, Quam R, Arsuaga JL, Lorenzo C, Gracia A, et al. (2009) Approche
pale´ontologique de l’e´volution du langage : un e´tat des lieux. L’anthropologie
113: 255–264.
20. Enloe JG (1993) Ethnoarchaeology of marrow cracking: implications for the
recognition of prehistoric subsistence organization. In: Hudson J, editor. From
Bones to Behavior: Ethnoarchaeological and Experimental Contributions to
the Interpretation of Faunal Remains. Carbondale Southern Illinois: Center for
Archaeological Investigations Occasional Paper No 21. 82–97.
21. Yellen JE (1977) Cultural patterning in faunal remains: evidence from the
Kung!Bushmen.In:IngersollD,YellenJE,MacDonaldW,editors.
Experimental Archaeology. New York: Columbia University Press. 271–331.
22. Lyman RL (1994) Vertebrate Taphonomy. Cambridge: Cambridge University
Press. 552 p.
23. Lyman RL (2008) Quantitative Paleozoology. Cambridge: Cambridge
University Press. 348 p.
24. Bunn HT, Bartram LE, Kroll EM (1988) Variability in bone assemblage
formation from Hadza hunting, scavenging, and carcass processing. J Anthropol
Archaeol 7: 412–457.
25. Potts R, Shipman P (1981) Cut-marks made by stone tools on bones from
Olduvai Gorge, Tanzania. Nature 291: 577–580.
26. Shipman P, Rose J (1983) Early hominid hunting, butchering and carcass-
processing behaviors: approches to the fossil record. J Anthropol Archaeol 2:
57–98.
27. Bromage TG, Boyde A (1984) Microscopic criteria for the determination of
directionality of cutmarks on bone. Am J Phys Anthropol 65: 339–366.
28. Shipman P, Fisher DC, Rose J (1984a) Mastodo n butchery: microscopic
evidence of carcass processing and bone tool use. Paleobiology 10: 358–365.
29. Noe-Nygaard N (1989) Man-made trace fossils on bones. Hum Evol 4: 461–
491.
30. Blumenschine RJ, Marean CW, Capaldo SD (1996) Blind tests of inter-analyst
correspondence and accuracy in the identification of cut marks, percussion
marks, and carnivore toothmarks on bone surfaces. J Archaeol Sci 23: 493–507.
31. Binford LR (1984) Faunal Remains from Klasies River Mouth. New York:
Academic Press. 283 p.
32. Fisher JW (1995) Bone surface modifications in zooarchaeology. J Archaeol
Method Th 2: 7–68.
33. Domı
´nguez-Rodrigo M (1997) Meat-eating by early hominids at the FLK 22
Zinjanthropus site, Olduvai Gorge (Tanzania): an experimental approach using
cutmark data. J Hum Evol 33: 669–690.
34. Domı
´nguez-Rodrigo M (2002) Hunting and scavenging by early humans: the
state of the debate. J World Prehist 16: 1–54.
35. Nilssen PJ (2000) An actualistic butchery study in South Africa and its
implications for reconstructing hominid strategies of carcass acquisition and
butchery in the Upper Pleistocene and Plio-Pleistocene. Ph.D. Dissertation,
University of Cape Town.
36. Bunn HT (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, Kenia. In: Clutton-Brock J, Grigson C, editors.
Animals and Archaeology: 1. Hunters and Their Prey. Oxford: British
Archaeological Reports International Series 163. 143–148.
37. Villa P, Mahieu E (1991) Breakage patterns of human long bones. J Hum Evol
21: 27–48.
38. Steadman DW, Plourde A, Burley DW (2002) Prehistoric butchery and
consumption of birds in the kingdom of Tonga, South Pacific. J Archaeol Sci
29: 571–584.
39. Capaldo SD, Blumenschine RJ (1994) A quantitative diagnosis of notches made
by hammerstone percussion and carnivore gnawing in bovid long bones. Am
Antiquity 59: 724–748.
40. Pickering T, Egeland CP (2006) Experimental patterns of hammerstone
percussion damage on bones: implications for inferences of carcass processing
by humans. J Hum Evol 33: 459–469.
41. Blumenschine RJ, Sel vaggio M (1988) Percussion marks on bone surfaces as
a new diagnostic of hominid behavior. Nature 333: 763–765.
42. White TD (1992) Prehistoric Cannibalism at Ma ncos 5MTUMR-2346.
Princeton: Princeton University Press. 462 p.
43. Domı
´nguez-Rodrigo M, Barba R (2006) New estimates of tooth mark and
percussion mark frequencies at the FLK Zinj site: the carnivore-hominid-
carnivore hypothesis falsified. J Hum Evol 50: 170–194.
44. Domı
´nguez-Rodrigo M, Barba R (2007) Five more arguments to invalidate the
passive scavenging version of the carnivore-hominid-carnivore model: a reply
to Blumenschine et al. 2007. J Hum Evol 53: 427–433.
45. Blumenschine RJ, Prassack KA, Kreger CD, Pante MC (2007) Carnivore
toothmarks, microbial bioerosion, and the invalidation of Dominguez-Rodrigo
and Barba’s (2006) test of Oldowan hominin scavenging behavior. J Hum Evol
53: 420–426.
46. Gala´n AB, Rodrı
´guez M, de Juana S, Domı
´nguez-Rodrigo M (2009) A new
experimental study on percussion marks and notches and their bearing on the
interpretation of hammerstone-broken faunal assemblages. J Archaeol Sci 36:
776–784.
47. Lam YM, Chen X, Pearson OM (1999) Intertaxonomic Variability in patterns
of bone density and the differential representation of bovid, cervid, and equid
elements in the archaeological record. Am Antiquity 64: 343–362.
48. Shipman P, Foster G, Schoeninger M (1984) Burnt Bones and Teeth: an
Experimental Study of Color, Morphology, Crystal Structure and Shrinkage.
J Archaeol Sci 11: 307–25.
49. Guillon F (1986). Os Brules secs ou fraiches? Paris: Anthropologie Physique et
Archeologie CNRS. 191–194.
50. Buikstra JE, Swegle M (1989) Bone modification due to burning: experimental
evidence. In: Bonnichsen R, Sorg MH, editors. Bone Modification. Orono:
University of Maine Center for the Study of the First Americans. 247–258.
51. Nicholson R (1993) A Morph ological Investigation of Burnt Animal Bone and
an Evaluation of its Utility in Archaeology. J Archaeol Sci 20: 411–28.
52. Mayne Correia P (1997) Fire modification of bone: a review of the literature.
In: Haglund W, Sorg M, editors. Forensic Taphonomy: the Postmortem Fate
of Human Remains. Boca Raton: CRC Press. 275–293.
53. Hanson M, Cain CHR (2007) Examining histology to identify burned bone.
J Archaeol Sci 34: 1902–1913.
54. Stiner MC, Weiner S, Bar-Yosef O, Kuhn SL (1995) Differential Burning,
Recrystallization and Fragmentation of Archaeological Bone. J Archaeol Sci
22: 223–37.
55. Pare´s JM, Pe´rez-Gonza´ lez A (1999) Magnetochronology and stratigraphy at
Gran Dolina section, Atapuerca (Burgos, Spain). J Hum Evol 37: 325–342.
56. Pe´rez-Gonza´ lez A, Pare´s JM, Carbonell E, Aleixandre T, Ortega AI, et al.
(2001) Ge´ologie de la Sierra de Atapuerca et stratigraphie des remplissages
karstiques de Galerı
´a et Dolina (Burgos, Espagne). L’Anthropologie 105: 27–
43.
57. Rodrı
´guez J, Burjachs F, Cuenca-Besco´s G, Garcı
´a N, Made Jvd, et al. (2011)
One million years of cultural evolution in a stable environment at Atapuerca
(Burgos, Spain). Quaternary Sci Rev 30: 1396–1412.
58. Falgue`res C, Bahain JJ, Yokoyama Y, Arsuaga JL, Bermu´ dez de Castro JM, et
al. (1999) Earliest humans in Europe: the age of TD6 Gran Dolina, Atapuerca,
Spain. J Hum Evol 37: 343–352.
59. Berger GW, Pe´rez-Gonza´ lez A, Carbonell E, Arsuaga JL, Bermu´ dez de Castro
JM, et al. (2008) Luminescence chronology of cave sediments at the Atapuerca
paleoanthropological site, Spain. J Hum Evol 55: 300–311.
60. Olle´ A, Mosquera M, Rodrı
´guez XP, Lombera-Hermida A, Garcı
´a-Anto´n
MD, et al. (2011) The Early and Middle Pleistocene technological record from
Sierra de Atapuer ca (Burgos, Spain). Quatern Int: in pr ess. 10.1016/
j.quaint.2011.11.009.
61. Rodrı
´guez XP (2004) Technical systems of lithic production in the Lower and
Middle Pleistocene of the Iberian peninsule. Technological variability between
North-Eastern sites and Sierra de Atapuerca sites. Oxford: BAR International
Series 1323. 184 p.
62. Mene´ndez L (2009) La Transicio´n del Modo 2 al Modo 3 vista a trave´s de la
industria lı
´tica de Gran Dolina TD10 (Atapuerca, Burgos) y Orgnac 3
(Arde´che, Francia). Desar rollo tecnolo´gico y posibles implicaciones ocupacio-
nales de los conjuntos, Ph. Dissertation, Universitat Rovira i Virgili, Tarragona.
63. Lo´pez-Ortega E, Rodrı
´guez XP, Vaquero M (2011) Lithic refitting and
movement connections: the NW area of level TD10–1 at the Gran Dolina site
(Sierra de Atapuerca, Burgos, north-western Spain). J Archaeol Sci 38: 3112–
3121.
64. Garcı
´a-Anto´n MD, Mosquera M (2007) Done´es pre´liminaires sur des aires
d’approvisionement et de selection des matie` res premie` res lithiques dans les
occupations du Ple´istocene Moyen du niveu TD10–1. In: Moncel MH, Moigne
AM, Arzarello M, Peretto C, editors. Aires d’approvisionement en matie`res
premie`res et aires d’approvisionement en resources alimentaires. Approche
inte´gre´e des comportements. Oxford: BAR International Series 1725. 171–185.
65. Rosell J, Blasco R, Campeny G, Dı
´ez JC, Alonso R, et al. (2011) Bone as
a technological raw material at the Gran Dolina site (Sierra de Atapuerca,
Burgos, Spain). J Hum Evol 61: 125–131.
66. Blasco R (2011) La amplitud de la dieta ca´rnica en el Pleistoceno medio
peninsular: una aproximacio´n a partir de la Cova del Bolomor (Tavernes de la
Valldigna, Valencia) y del subnivel TD10–1 de Gran Dolina (Sierra de
Atapuerca, Burgos). Ph. Dissertation, Universitat Rovira i Virgili, Tarragona.
67. Emerson AM (1993) The role of body part utility in small-scale hunting under
two strategies of carcass recovery. In: Hudson J, editors. From Bones to
Behavior: Ethnoarchaeological and Experimental Contributions to the In-
terpretation of Faunal Remains. Carbondale Southern Illinois: Center for
Archaeological Investigations Occasional Paper No 21. 138–155.
68. Louchart A, Soave R (2002) Changement d’ampleur de l’explotation des
oiseaux entre le Magdale´nien et l’Azilien: l’exemple du Taı
¨2(Droˆme).
Quaternaire 3–4: 297–312.
69. Cochard D (2004) Les Le´ poride´s dans la subsistance Pale´ olithique du sud de la
France. Ph D. thesis. Universite´ Bordeaux I-France.
70. Cochard D, Brugal J-P, Morin E, Meignen L (2012) Evidence of small fast
game exploitation in the Middle Paleolithic of Les Canalettes (Aveyron,
France). Quatern Int 264: 32–51.
71. Laroulandie V (2005) Anthropogenic versus non-anthropogenic bird bone
assemblages: new criteria for their distinction. In: O’Connor T, editor.
Biosphere to Lithosphere: New Studies in Vertebrate Taphonomy. Oxford:
Oxbow Books. 25–30.
Learning by Heart
PLOS ONE | www.plosone.org 18 February 2013 | Volume 8 | Issue 2 | e55863
72. Pe´rez Ripoll M (2005) Caracterizacio´ n de las fracturas antro´ picas y sus
tipologı
´as en huesos de conejo procedentes de los niveles gravetienses de la
Cova de les Cendres (Alicante). Munibe 57: 239–254.
73. Rosell J, Blasco R (2009) Home sharing: carnivores in anthropogenic
assemblages of the Middle Pleistocene. J Taphonomy 7: 305–324.
74. Blasco R, Rosell J, Arsuaga JL, Bermu´dez de Castro JM, Carbonell E (2010)
The hunted hunter: the capture of a lion (Panthera leo fossilis) at the Gran Dolina
site, Sierra de Atapuerca, Spain. J Archaeol Sci 37: 2051–2060.
75. Fumanal MP (1995) Los depo´ sitos cuaternarios en cuevas y abrigos rocosos.
Implicaciones sedimentolo´ gicas. El Cuaternario del Paı
´s Valenciano: 115–124.
76. Ferna´ndez Peris J (2007) La Cova del Bolomor (Tavernes de la Valldigna,
Valencia). Las industrias lı
´ticas del Pleistoceno medio en el a´ mbito del
Mediterra´neo peninsular. Valencia: Trabajos Varios del SIP 108. 462 p.
77. Ferna´ndez Peris J, Barciela Gonza´ lez V, Blasco R, Cuartero F, San˜udo P
(2008) El Paleolı
´tico Medio en el territorio valenciano y la variabilidad tecno-
econo´mica de la Cova del Bolomor. Treballs d’Arqueologia 14: 141–169.
78. Cuartero F (2008) Tecnologı
´alı
´tica en la Cova del Bolomor IV: ¿una economı
´a
de reciclado? Saguntum 39: 27–44.
79. Ferna´ndez Peris J, Barciela Gonza´ lez V, Blasco R, Cuartero F, Fluck H, et al.
(2012) The earliest evidence of hearths in Southern Europe: the case of
Bolomor Cave (Valencia, Spain). Quatern Int 247: 267–277.
80. Sanchis Serra A (2010) Los lagomorfos del Paleolı
´tico medio de la regio´n
central y sudoriental del Mediterra´ neo Ibe´rico. Caracterizacio´n tafono´mica y
taxono´mica. Ph.D. thesis, Universitat de Vale`ncia.
81. Landt MJ (2007) Tooth marks and human consumption: ethnoarchaeological
mastication research among foragers of the Central African Republic.
J Archaeol Sci 34: 1629–1640.
82. Lloveras L, Moreno-Garcı
´a M, Nadal J (2009) Butchery, cooking and human
consumption marks on rabbit (Oryctolagus cuniculus) bones: An experimental
study. J Taphonomy 7: 179–201.
83. Blasco R (2008) Human consumption of tortoises at Level IV of Bolomor Cave
(Valencia, Spain). J Archaeol Sci 35: 2839–2848.
84. Blasco R, Ferna´ndez Peris J (2009) Middle Pleistocene bird consumption at
Level XI of Bolomor Cave (Valencia, Spain). J Archaeol Sci 36: 2213–2223.
85. Blasco R, Ferna´ ndez Peris J (2012) A uniquely broad spectrum diet during the
Middle Pleistocene at Bolomor Cave (Valencia, Spain). Quatern Int 252: 16–
31.
86. Blasco R, Ferna´ndez Peris J (2012) Small and large game: human use of diverse
faunal resources at Level IV of Bolomor Cave (Valencia, Spain). C R Palevol
11: 265–282.
87. Roebroeks W (2001) Hominid Behaviour and the Earliest Occupation of
Europe: An Exploration. J Hum Evol 41: 437–61.
88. Yellen JE (1991) Small Mammals: !Kung San Utilization and the Production of
Faunal Assemblages. J Anthropol Archaeol 10: 1–26.
89. Frison GC (1970) The Glenrock Buffalo Jump, 48CO304: Late Prehistoric
Period Buffalo Procurement and Butchering on the Northwestern Plains.
Lincoln: Plains Anthropologist Memoir 7. 66p.
90. Frison GC (1978) Prehistoric Hunters of the High Plains. New York: Academic
Press. 457 p.
91. Gifford-Gonzalez D (1991) Bones are not enough: analogues, knowledge, and
interpretative strategies in zooarchaeology. J Anthropol Archaeol 10: 215–254.
92. Lyman RL (1995) A study of variation in the Prehistoric butchery of large
artiodactyls. In: Johnson E, editor. Ancient Peoples and Landscapes. Lubbock:
Museum of Texas Tech University. 233–253.
93. Lyman RL (2005) Analyzin g cut marks: lessons from artiodactyl remains in the
northwestern United States. J Archaeol Sci 32: 1722–1732.
94. Domı
´nguez-Rodrigo M (2008) Conceptual premises in experimental design
and their bearing on the use of analogy: an example from experiments on cut
marks. World Archaeol 40: 67–82.
95. Domı
´nguez-Rodrigo M, Yravedra J (2009) Why are cut mark frequencies in
archaeofaunal assemblages so variable? A multivariate analysis. J Archaeol Sci
36: 884–894.
96. Selvaggio MM (1994) Carnivore tooth marks and stone tool butchery marks on
scavenged bones: archaeological implications. J Hum Evol 27: 215–228.
97. Domı
´nguez-Rodrigo M (1997) Meat-eating by early hominids at the FLK 22
Zinjanthropus site, Olduvai Gorge (Tanzania): an experimental approach using
cut-mark data. J Hum Evol 33: 669–690.
98. Domı
´nguez-Rodrigo M (2002) Hunting and scavenging by the early hominids:
the state of the debate. J World Prehist 16: 1–53.
99. Domı
´nguez-Rodrigo M (2003) On cut marks and statistical inferences:
methodological comments on Lupo and O’Connell (2002). J Archaeol Sci
30: 381–386.
100. Lupo KD, O’Connell JF (2002) Cut and tooth marks distributions on large
animal bones: ethnoarchaeological data from the Hadza and their implications
for current ideas about early human carnivory. J Archaeol Sci 29: 85–109.
101. Domı
´nguez-Rodrigo M, Pickering TR (2003) Early hominid hunting and
scavenging: a zooarcheological review. Evol Anthropol 12: 275–282.
102. Pickering TR, Domı
´nguez-Rodrigo M, Egeland CP, Brainc CK (2004) Beyond
leopards: tooth marks and the contribution of multiple carnivore taxa to the
accumulation of the Swartkrans Member 3 fossil assemblage. J Hum Evol 46:
595–604.
103. Capaldo SD (1995) Inferring hominid and carnivore behavior from dual
patterned archaeofaunal assemblages. Ph.D. thesis, Rutgers University, New
Brunswick.
104. Capaldo SD (1997) Experimental determinations of carcass processing by Plio-
Pleistocene hominids and carnivores at FLK 22 (Zinjanthropus), Olduvai
Gorge, Tanzania. J Hum Evol 33: 555–597.
105. Capaldo SD (1998) Methods, marks and models for inferring hominid and
carnivore behavior. J Hum Evol 35: 317–320.
106. Rosell J, Blasco R, Ferna´ ndez-Laso C, Vaquero M, Carbonell E (2012)
Connecting areas: Faunal refits as a diagnostic element to identify synchronicity
in the Abric Romanı
´archaeological assemblages. Quatern Int 252: 56–67.
107. Padilla M (2008) Transmisio´n cultural y procesos de aprendizaje de carnicerı
´a:
un estudio experimental comparando expertos y novicios en el descarnado y
desarticulacio´n. Arqueo web 9 (2): 1–76.
108. Gifford-Gonzalez D (1989) Ethnographic analogues for interpreting modified
bones: some cases from East Africa. In: Bonnichsen R, Sorg M, editors. Bone
Modification. Orono: Center for the Studi of the First Americans. 61–71.
109. Oliver JS (1993) Carcass processing by the Hadza: Bone breakage from
butchery to consumtion. In: Hudson J., editor. From Bones to Behavior
Ethnoarchaelogical and Experimental Contributions to the Interpretation of
Faunal Remains. Carbondale Southern Illinois: Center for Archaeological
Investigations Occasional Paper No 21. 200–227.
110. Bunn HT, Stanford CB (2001) Conclusions: research trajectories on hominid
meat-eating. In: Stanford CB, Bunn HT, editors. Meat-eating and Human
Evolution. Oxford: Oxford University Press. 350–359.
111. Werner D (1990) Amazon Journey. An Anthropologist’s Year among Brazil’s
Mekranoti Indians. Englewood Cliffs, New Jersey: Prentice Hall. 206 p.
112. Flannery KV, Wheeler JC (1986) Animal food remains from preceramic Guila´
Naquitz. In: Flannery KV, editor. Guila´ Naquitz: Archaic Foraging and Early
Agriculture in Oaxaca, Mexico. Orlando: Academic Press. 285–295.
113. Sampson CG (1998) Tortoise remains from a later Stone Age rock shelter in
the Upper Karoo, South Africa. J Archaeol Sci 25: 985–1000.
114. Speth JD, Tchernov E (2002) Middle Paleolithic tortoise use at Kebara Cave
(Israel). J Archaeol Sci 29: 471–483.
115. Stahl PW, Oyuela-Caycedo A (2007) Early prehistoric sedentism and seasonal
animal exploitation in the Caribbean lowlands of Colombia. J Anthropol
Archaeol 26: 329–349.
116. Yellen JE (1977) Archaeological approaches to the present: models for
reconstructing the past. New York: Academic Press. 259 p.
117. Speth JD (2012) Middle Palaeolithic subsistence in the Near East: zooarch-
aeological perspectives -past, present and future. Before Farming 2: 1–45.
118. Outram AK (2001) A new approach to identifying bone marrow and grease
exploitation: Why the ‘‘indeterminate’’ fragments should not be ignored.
J Archaeol Sci 28: 401–410.
119. Blumenschine RJ (1988) An experimental model of the timing of hominid and
carnivore influence on archaeological bone assemblages. J Archaeol Sci 15:
483–502.
120. Marean CW, Spencer LM, Blumenschine RJ, Capaldo SD (1992) Captive
hyena bone choice and destruction, the schlepp effect, and Olduvai Gorge
archaeofaunas. Journal of Archaeological Science 18: 101–121.
121. Outram A, Rowley-Con wey P (1998) Meat and marrow utility indices for horse
(Equus). J Archaeol Sci 25: 839–849.
122. Domı
´nguez-Rodrigo M, Barba R (2005) The taphonomic relevance of the
analysis of bovid long limb bone shaft features and their application to element
identification: study of bone thickness and morphology of the medullary cavity.
J Taphonomy 3: 29–42.
123. Peretto C, Anconetani P, Crovetto C, Evangelista L, Ferrari M, et al. (1996)
Aproccio sperimentale alla comprensione delle attivita` di sussistenza condotte
nel sito di Isernia La Pineta (Molise-Italia). La fratturazione intenzionale. In:
Peretto C, editor. I Reperti Paleontologici del Giacimento Paleolitico di Isernia
La Pineta. Isernia: Istituto Regionale per gli Studi Storici del Molise ‘‘V.
Cuoco’’. 187–452.
124. Murdock GP, Provost C (1973) Factors in the division of labor by sex: a cross-
cultural analysis. Ethnology 12: 203–225.
125. Hahn J (1987) Zur Funktion einer Aurignacien-Feuerstelle aus dem
Geissenklo¨sterle bei Blaubeuren. Fundberichte aus Baden. Wu¨ rttemberg 14:
1–22.
126. Jacobi R (2004) The Late Upper Palaeolithic lithic collection from Gough’s
Cave, Cheddar, Somerset and human use of the cave. P Prehist Soc 70: 1–92.
127. Schmider B, Bodu P (1992) Marsangy. Un campement des derniers chasseurs
magdale´niens, sur les bords de l’Yonne. Lie`ge: E
´tudes et Recherches
Arche´ologiques de l’Universi te´deLie`ge Iniversite´deLie` ge, Volumen 55 de
Eraul. 275 p.
128. Zubrow E, Audouze F, Enloe JG (2010) The Magdalenian Household.
Unraveling Domesticity. The Institute for European and Mediterranean
Archaeology Distinguished Monograph Series. New York: State University of
New York Press. 335 p.
129. Johansen L, Stapert D (2004) Oldeholtwolde. A Hamburgian family
encampment around a hearth. Lisse: Balkema Publishers. 229 p.
130. Bodu P (1990) L’Application de la me´thode des remontages a` l’e´tude du
mate´riel lithique des premiers niveaux chaˆ telperroniens de la Grotte du Renne
a` Arcy-s ur-Cure (Yonne). In: Farizy C, editor. Pale´olithique moyen et
Pale´olithique supe´rieur ancien en Europe. Nemours: Me´moires du Muse´e de
Pre´histoire d’Ile-de-France 3. 309–312.
131. Stapert D (2007) Neanderthal children and their flints. PalArch’s Journal of
Archaeology of Northwest Europe 1, 2: 16–39.
Learning by Heart
PLOS ONE | www.plosone.org 19 February 2013 | Volume 8 | Issue 2 | e55863
132. Stapert D, Niekus M, Johansen L (2006) Curieuze vuistbijlachtigen van
Rhenen (Utr.). Ook eens iets voor kinderen? Paleo-Aktueel 17: 18–26.
133. David N, Kramer C (2001) Ethnoarchaeology in Action. Cambridge:
Cambridge University Press. 428 p.
134. Stevenson MG (1985) The Formation of Artifact Assemblages at Workshop/
Habitation Sites: models from Peace Point in Northern Alberta. Am Antiquity
50: 63–81.
135. Nadel D, Weiss E, Simchoni O, Tsatskin A, Danin A, et al. (2004) Stone Age
Hut in Israel Yields World’s Oldest Evidence of Bedding. PNAS 101: 6821–
6826.
136. Vaquero M, Pasto´ I (2001) The definition of spatial units in Middle Palaeolithic
sites: the hearth-related assemblages. J Archaeol Sci 28: 1209–1220.
137. Vaquero M, Vallverdu´ J, Rosell J, Pasto´ I, Allue´ E (2001) Neandertal behavior
at the Middle Palaeolithic site of Abric Romanı
´, Capellades, Spain. J Field
Archaeol 28: 93–114.
138. Vaquero M, Rando JM, Chaco´n MG (200 4) Neanderthal Spatial Behaviour
and Social Structure: hearth related assemblages from the Abric Romanı
´
Middle Palaeolithic site. In: Conard N, editor. Settlement Dynamics of the
Middle Paleolithic and Middle Stone Age. Tu¨ bingen: Kerns Verlag 2. 367–
392.
139. Mallol C, Carbonell E (2008) The collapse of Gran Dolina Cave, Sierra de
Atapuerca, Spain: site formation processes of layer TD10–1. Geoarchaeology
23: 13–41.
140. Isaac GL (1967) Toward the interpretation of occupation debris: some
experiments and observations. Kroeber Anthropological Papers 37: 31–57.
141. Binford LR (1987) Searching for camps and missing the evidence?: another
look at the Lower Paleolithic. In: Soffer O, editor. The Pleistocene of World.
New York: Plenum Press. 17–31.
142. Butzer KW (1982) Archaeology as human ecology: Method and theory for
a contextual approach. Cambridge: Cambridge University Press. 364 p.
143. O’Connell JF (1997) On Plio-Pleistocene archaeological sites and central
places. Curr Anthropol 38: 86–88.
144. Lyman RL (2003) The influence of time averagi ng and space averaging on the
application of foraging theory in zooarchaeology. J Archaeol Sci 30: 596–610.
145. Carbonell E, Rosell J (2004) Ocupaciones de homı
´nidos en el Pleistoceno de la
Sierra de Atapuerca. In: Baquedano E, Rubio S, editors. Miscelania en
Homenaje a Emiliano Aguirre. Zona Arqueolo´ gica 4: 102–115.
146. Banks WE, d’Errico F, Peterson AT, Vanhaeren M, Kageyama M, et al. (2008)
Human ecological niches and ranges during the LGM in Europe derived from
an application of eco-cultural niche modeling. J Archaeol Sci 35: 481–491.
147. Haws JA (2003) An investigation of Late Upper Paleoli thic and Epipaleolithic
hunter-gatherer subsistence and settlement patterns in central Portugal. Ph.
Dissertation, University of Wisconsin-Madison.
148. Pe´rez Ripoll M (2004) La consommation humaine des lapins pendant le
Pale´olithique dans la re´gion de Vale`ncia (Espagne) et l’e´tude des niveaux
grave´tiens de la Cova de les Cendres (Alicante). In: Brugal JP, Desse J, editors.
Petits animaux et socie´te´s humaines. Du comple´ment alimentaire aux
ressources utilitaires. Antibes: E
´ditions APDCA. 191–206.
149. Aura JE, Villaverde V, Pe´rez Ripoll M, Martı
´nez Valle R, Guillem PM (2002)
Big game and small prey: Paleolithic and Epipaleolithic economy from
Valencia (Spain). J Archaeol Method Th 9: 215–267.
150. Aura JE, Jorda´ JF, Pe´ rez Ripoll M, Morales JV, Garcı
´a O, et al. (2009)
Epipaleolı
´tico y mesolı
´tico en Andalucı
´a Oriental. Primeras notas a partir de
los datos de Cueva de Nerja (Ma´laga, Espan˜ a). Monografı
´as Arqueolo´gicas 44:
343–360.
151. Hockett BS, Haws JA (2002) Taphonomic and methodological perspective s of
leporid hunting during the upper Paleolithic of the western Mediterranean
basin. J Archaeol Method Th 9: 269–301.
152. Hockett BS (2006) Climate, dietary choice, and the Paleolithic hunting of
rabbits in Portugal. In: Bicho NF, editor. Animais na pre´-historia e arqueologı
´a
da penı
´nsula Ibe´rica. Faro: Universidade do Algarve. 137–144.
153. Allue´ E, Iba´n˜ez N, Saladie´ P, Vaquero M (2010) Small preys and plant
exploitation by Late Pleistocene hunter–gatherers. A case study from the
Northeast of the Iberian Peninsula. Archaeol Anthropol Sci. 2: 11–24.
154. Bicho NF, Hockett B, Haws J, Belcher W (2000) Hunter-gatherer subsistence at
the end of the Pleistocene: preliminary results from Picareiro cave, Central
Portugal. Antiquity 74: 500–506.
155. Bicho NF, Haws J, Hockett B (2006) Two sides of the same coin-rocks, bones
and site function of Picareiro Cave, Central Portugal. J Anthropol Archaeol 25:
485–499.
156. Jones K (1983) Forager archaeology: The Ache´ of eastern Paraguay. In:
Lemoine GM, MacEachern AS, editors. Carnivores, human scavengers and
predators: A question of bone technology. Calgary: The University of Calgary,
Archaeological Association. 171–191.
157. Mithen S (1994) Technology and Society during the Middle Pleistocene:
Hominid Group Size, Social Learning and Industrial Variability. Camb
Archaeol J 4: 3–32.
158. Gowlett JAJ (1996) Me ntal Abilities of Early Homo: Elements of Constraint and
Choice in Rule Systems. In: Mellars P, Gibson K, editors. Modelling the early
human mind. Ca mbridge: The Mc Donald Institute for Archaeological
Research. 191–215.
159. Goren-Inbar N (2011) Culture and cognition in the Acheulian industry: a case
study from Gesher Benot Ya’aqov. Phil Trans R Soc Lond B 366: 1038–1049.
160. Stout D (2002) Skill and Cognition in Stone Tool Production. An Ethnographic
Case Study from Irian Jaya. Curr Anthropol 43: 693–722.
161. White M, Ashton N (2003) Lower Palaeolithic Core Technology and the
Origins of the Levallois Method in North-Western Europe. Curr Anthropol 44:
598–609.
Learning by Heart
PLOS ONE | www.plosone.org 20 February 2013 | Volume 8 | Issue 2 | e55863
