ArticlePDF Available

An ancient continuous human presence in the Balkans and the beginnings of human settlement in western Eurasia: A Lower Pleistocene example of the Lower Palaeolithic levels in Kozarnika cave (North-western Bulgaria)


Abstract and Figures

The lower levels of the archaeological sequence of Kozarnika cave (north-western Bulgaria) provide levels with non-Acheulian core-and-flake (as opposed to pebble-core) industries. Associations with numerous taxa of large mammals indicate that the date of the lower layers of the cave falls between MNQ17 and MNQ19. These layers produced several bones showing anthropic traces. Its date and geographical position place Kozarnika cave as a landmark site in the primary phase of the settlement of Europe.
Content may be subject to copyright.
An ancient continuous human presence in the Balkans and the beginnings of
human settlement in western Eurasia: A Lower Pleistocene example of the
Lower Palaeolithic levels in Kozarnika cave (North-western Bulgaria)
N. Sirakov
, J.-L. Guadelli
, S. Ivanova
, S. Sirakova
, M. Boudadi-Maligne
, I. Dimitrova
Fernandez Ph
, C. Ferrier
, A. Guadelli
, D. Iordanova
, N. Iordanova
, M. Kovatcheva
I. Krumov
, J.-Cl. Leblanc
, V. Miteva
, V. Popov
, R. Spassov
, S. Taneva
, T. Tsanova
National Institute of Archaeology and Museum of Bulgarian Academy of Sciences, 2, Saborna Street, 1000 Soa, Bulgaria
PACEA/IPGQ-UMR5199 CNRS, Université Bordeaux I, Avenue des facultés, Bâtiment B18, 33405 Talence cedex, France
LAMPEA-UMR6636 CNRS, MMSH, 5 Rue du château de lHorloge, BP 647, 13094 Aix en Provence cedex 2, France
Institute of Geophysic of Bulgarian Academy of Sciences, Acad. G. Bontchev Street, bat.3, 1113 Soa, Bulgaria
New Bulgarian University, Department of Archaeology, Building 2, Room 219, Soa, Bulgaria
TRACES, UMR5608 CNRS, Université de Toulouse le Mirail, Maison de la Recherche, 5, allées Antonio Machado, 31048 Toulouse Cedex 1, Bulgaria
Institute of Zoologie of Bulgarian Academy of Sciences, 1, bd. Tsar Osvoboditel, 1000 Soa, Bulgaria
Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Deutscher Platz 6, D-04103 Leipzig, Germany
article info
Article history:
Available online 10 March 2010
The lower levels of the archaeological sequence of Kozarnika cave (north-western Bulgaria) provide levels
with non-Acheulian core-and-ake (as opposed to pebble-core) industries. Associations with numerous
taxa of large mammals indicate that the date of the lower layers of the cave falls between MNQ17 and
MNQ19. These layers produced several bones showing anthropic traces. Its date and geographical position
place Kozarnika cave as a landmark site in the primary phase of the settlement of Europe.
Ó2010 Elsevier Ltd and INQUA. All rights reserved.
1. Introduction
Investigations on the commencement of human settlement in
western Eurasia show that there are about ten sites dated before or
approximately at 1 Ma. Dates such as 1.7 Ma for Dmanisi (Caucasus),
1.6 e1.4 Ma for Kozarnika in the Balkans, between 1.6 and approxi-
mately 1 Ma for Piro Nord,Monte Poggiolo and Cerpano in the Italian
peninsula, around the million years for Vallonet and Lunery in
France, around 1.4 Ma for the sites of Orce and 1.2e1.1 Ma f or Sima
del Elefante (Spain) show clearly an East-West aligned expansion
which was limited to the Southern zone of the continent (Fig.1). The
lithic assemblages of all these sites are non-Acheulian core-and-ake
industries, most with more or less frequent exploitation of pebbles as
cores (Peretto et al.,1998).
The idea that this primary population arrived via Gibraltar
seems less probable (Aguirre and Carbonell, 2001), being replaced
by passage from the east (Carbonell et al., 2008) or directly through
the Asian steppes. This is more than likely, given the passage along
the north coast of Black Sea revealed by the recent data of Rodniki
and Bogatyri for the Tamanian Peninsula (Shchelinsky et al., 2010).
However it is also necessary to consider the most direct route via
the Dardanelles and Bosporus. Also, it should not be forgotten that
the passage by Gibraltar is not strictly excluded, but that is another
debate which is not the subject of this paper.
2. The site
First mentioned in 1931 by Popov (1933), Kozarnika cave is
situated in the north-western part of the lower Balkans near the
Danube plain, approximately 3 km from the village of Oreshets
(district of Belogradchik) and some 30 km from the Serbian border
(Fig. 2). Regrettably too poor to be able to be utilized, the only
information previously available on the Palaeolithic occupation of
the region is that of Popov (1933, 1936) from Mirizlivka cave, and of
Djambazov (1958) from Magura cave.
*Corresponding author. Tel.: þ33 5 40 00 65 62; fax: þ33540008451.
E-mail addresses: (N. Sirakov), jeanluc.guadelli@ (J.-L. Guadelli), (S. Ivanova), bodysirakova@ (S. Sirakova), (M. Boudadi-
Maligne), (I. Dimitrova), (F. Ph), (C. Ferrier), (A. Guadelli), (I. Krumov), (J.-Cl. Leblanc), viviana_ (V. Miteva), (V. Popov), stanimira.taneva@ (S. Taneva), (T. Tsanova).
Contents lists available at ScienceDirect
Quaternary International
journal homepage:
1040-6182/$ esee front matter Ó2010 Elsevier Ltd and INQUA. All rights reserved.
Quaternary International 223-224 (2010) 94e106
Kozarnika cave opens to the south at an altitude of 481m on
a northern hillside of the valley of a Skomlia River tributary (itself
a tributary of the Danube River into which it ows 30 km to the
northeast). The hillside is composed of steep Upper Jurassic
calcareous slopes of the Belogradchik anticline lying above
conglomeratic formations. The valley, about 185 m deep, cuts
through three geological formations; from the top, the Upper
Jurassic gray limestones in which Kozarnika cave lies (85 m above
the valley oor), the Dogger yellow gritty limestones (60 m thick)
and 10 m of Lias red conglomerates at the bottom.
The strike of the calcareous beds (13
NE) and their fractures
have governed the development and morphology of the cave, its
unique gallery 210 m long directed N9
, probably developing from
a diaclase. A porch, set back about 10 m from the outside ledge, is
18 m deep and 11 m wide (Fig. 3). The archaeological layers are
6e7 m in depth.
From the time of the Eneolithic, the inll of Kozarnika (Goat
shed cave), also known as Cuwj Tey(Dry Cave) or Jivanova Dupka,
was disturbed by the digging of storage pits which contain frag-
ments of ceramic, ints and faunal remains reworked from the
Fig. 1. More possible directions of the beginning of the colonization of Western Eurasia, considering the sites dated before 1 Ma. (N. Sirakov, J.-L. Guadelli, A. Guadelli).
Fig. 2. Location map (J.-L. Guadelli from Geoatlas).
N. Sirakov et al. / Quaternary International 223-224 (2010) 94e106 95
Pleistocene and Holocene levels. Later, the cave was used as a goat
shed, mushroom cellar, and more recently in the 1960s, a fallout
shelter when a wall was constructed to block the entrance.
This cave is one many karst features of north-western Bulgaria
prospected in 1994, when two test-trenches dug by an archaeo-
logical team managed by N. Sirakov, H. Laville and S. Ivanova
revealed Holocene and Pleistocene archaeological levels. Since
1996, more than 60 m
have been excavated in the vestibule as well
as in another location 72 m in from the entrance (Fig. 3).
The Pleistocene sequence is more or less well preserved, but
almost everywhere its upper part containing the most recent
Gravettian levels (classically corresponding chronologically to
evolved epi-Gravettian), was truncated. Furthermore, between 12 m
from the entrance and the bottom of the vestibule, Pleistocene layers
are perturbed down to the depth of layer 11a, and even as far as layer
11c, whichis the upper part of the Lower Palaeolithic sequence. In this
disturbed zone, Mousterian and Upper Palaeolithic levels are
replaced either by mixed Holocene and Pleistocene reworked
material or by deposits resulting from the use of the cave as a goat
3. Stratigraphy of the Pleistocene sequence
3.1. Layers 13e11a
At a thickness of about 2.5 m, layers 13e11a currently form the
lower partof the deposits (Fig. 4).These layers arecomposed of rather
compact orange loamy sediments, blotched with dark manganese
and more or less rich in coarse sediment fractions (limestone, int
from the wall of the cave and more rarely rounded pebbles of quartz).
In the upper parts (layer 11a) and especially towards the eastern cave
wall, these layers are rich in heterometric and angular blocks and
pebbles organized into lenses with diffuse outlines. More central in
these layers (layers 11b and 11c) and towards the lower parts (layers
12 and 13) blocks and calcareous pebbles show more alteration, the
darker silty fraction containing int and quartz gravels and coprolite
fragments(more abundant in layer 12). The lenses,including those in
layer 13, represent the last indication of runoff of endokarstic origin.
Above, the homogeneous aspect of the deposits, their silty texture,
absence of organization in the coarse fraction and in some places in
the ne fraction, suggest that the deposits were formed, at least
partially, as a succession of small mud ows (Ferrier et al., 2009). All
these layers contain Lower Palaeolithic industries.
3.2. Layers 10c and 10b
These are found in some sectors excavated between 7 and 12 m
from the entrance. Varying in thickness from 15 to 50 cm, the
sediments which make up these layers consist of non-compacted
dark-brown loamy sand containing numerous gravels and calcar-
eous pebbles that are more or less weathered. The brown coloration
may be the result of human activity (charcoals) but could also be
due to an accumulation of organic matter of pedological origin,
namely the resumption of a humic horizon. These layers indicate
amelioration in climatic conditions that could also explain the
inclusion of the weathered calcareous elements. These sediments
have been removed by soliuction after deposition (Ferrier et al.,
2009). These layers contain the lower extremity of the Mouste-
rian sequence.
3.3. Layers 10ae5a
The average thickness of this set is about 1.2e1.5 m. These are
characterized by a nely powdered beige silty fraction, probably of
aeolian origin. The coarse fraction (calcareous gravels and hetero-
metric pebbles), varies in abundance in different layers, as a result of
the fragmentation of the walls and ceiling. Allochthonous elements
appear absent. Excavation of layer 8 showed that it was a den inll
and not a geological layer in the strict sense. However, to avoid
confusion, the lower layers were not remained to exclude this one,
Fig. 3. Kozarnika. Map of the cave (drawing: C. Ferrier, J.-C. Leblanc).
Fig. 4. Kozarnika. Sector II. South prole. Lower layers (photo C. Ferrier).
N. Sirakov et al. / Quaternary International 223-224 (2010) 94e10696
and so the sequence passes directly from 7 to 9. The characteristics of
layers 6 and 7, similar and indeed almost identical, were difcult to
differentiate in most areas,so 6/7isused as the label for this level. It
is a silty sediment (3e15 cm thick), light brown(layer 7) to light gray
(layer 6), containing rare calcareous gravels as well as charcoal. In
some squares, layer 6/7 corresponds to a thin layer of charcoal, but
this is not always present (Ferrier et al., 2009).
As regards the post-sedimentary processes, the weathered
calcareous elements are present in layer 10a but their intensity
decreases in the overlying layers. Layers 10a to 6/7 are affected by
soliuction and 5a by cryoturbation. Thisgroup of layers, especially in
the upper parts, corresponds to a time span in which aeolian sedi-
mentation occurred and low temperatures allowed the sediment of
Some volcanic ashes recognized in layer 5b (26,490þ/-270 BP
(GifA-99044); 31,237þ/-389 cal BP CalPal Online) by C. Ferrier in
2004 (Ferrier et al.,2009) have to be correlated to some tephra layers
recovered from the northeastern part of Lake Ohrid, (Republics of
Macedonia and Albania) (Vogel et al., 2010). Furthermore, in summer
2009, D. White took several samples in top part of the middle Pleis-
tocene sequence and in the upper Pleistocene sequence of Kozarnika
to nd and identify the origin of the tephra (RESET project).
Layers 10ae9a contain the archaeological levels XIII eIX of the
Mousterian sequence, layer 6/7 includes level VIII, which corre-
sponds to an industry (or industries?) with two components that is
characteristic of the Middle Palaeolithic and the initial Upper
Palaeolithic, still poorly dened. With layer 5, levels VII, VI and V
the Kozarnikien begins, which is a local tradition of blade industries
containing backed pieces.
3.4. Layers 4e3a
The total thickness of this series is 1e1.4 m. The layers, the limits
of which are often strongly deformed, are composed of calcareous
gravel in a light brown to whitish powdered silty matrix with
a lamellar structure. The origin of the sediment is the same as that of
the underlying layers, but the fraction coming from the walls of the
cave is more abundant than the aeolian component, the ne fraction.
The lamellar structure and cryoturbation provides evidence for
alternating frost/thaw post-sedimentary processes in the sediment.
This series indicates, as does the previous series, the rigorous
climatic conditions in which the site existed. These climatic nuances
explain the development of the freezing structures as well as the
reduction in deposition rates of the aeolian contributions (Ferrier
et al., 2009). These layers contain archaeological levels IVbe0I
which are attributed to middle and recent stages of the Kozarnikien.
3.4.1. Stratigraphy of Locus IV
As this locus was only recognized 3 years ago, only a preliminary
interpretation of the sedimentary ensemble is provided, based on
a synthesised study of geological data (Fig. 5). Three ensembles can
be recognized. Ensemble 1. Mainly visible on thewest prole making up the
top of the sequence, it includes units1, 2, 3, 4, 5,6, a, b, and c. These are
characterized by the presence of numerous limestone pebbles and
gravels contained in a brown silty matrix of supporting clasts with
a closed structure. Units 5 and b are distinguished by their orange
colouring due to the presence of small bone splinters or coprolite
fragments. The units slope gently to the south, towards the cave
mouth. These deposits correspond to rockslides moved by gravity
and derived from the weathering of the gallery walls and ceiling. Ensemble 2. The middle part of the sequence includes units
d, e, f, g, h, i and j. These deposits, localized mainly on the south, east
and north proles, are characterized by the scarcity of a coarse frac-
tion, the quasi-absence of calcareous fragments, a silty texture and
more or lessdarkened brown orgreyish tints. Theseunits are affected
by various deformations including one visible in square K72 on the
south prole. The absence of calcareous elements in the coarse
fraction, and the presence of int, suggests weathering by dissolu-
tion. This may correspond to phosphate neogenesis.
Here, it is necessary to envisage a reduction in the initial thick-
ness of the layers. Furthermore, the limits between the units would
not reect a change in the dynamics of accumulation but could be
the result of post-depositional processes. Micro-morphological
observations, particularly of unit d, identied microstructures linked
to frost and thaw alternations conrming the deformations of this
ensemble were produced by cryoturbation. Ensemble 3. These deposits are presently situated at the
bottom of the prole including units k, l and a sandy-rocky lens
Fig. 5. Kozarnika. Sector IV. West prole (photo S. Ivanova).
N. Sirakov et al. / Quaternary International 223-224 (2010) 94e106 97
Table 1
Kozarnika. Sectors I and II. Distribution of the identied taxa according to geological layers (Arabic numerals) and archaeological levels (Roman numerals).
N. Sirakov et al. / Quaternary International 223-224 (2010) 94e10 698
situated in the corner of the north and east prole. The texture
of the sediments, as well as the presence of lenses with sorted
lling, suggests an origin attributable to diffuse or concentrated
3.4.2. Correlation with the deposits of the vestibule
Several hypotheses have been advanced. The fauna species list
and their evolutionary level seems to indicate a period falling
between 35 and 40 or perhaps 45 ka, which suggeststhat these levels
could be equivalent to layers 5c and 6/7 in the vestibule. However,
from the point of view of the lithic industries in units a, b, c, g, h, i, j, k,
assemblages correspond to the Mousterian sequence of the vestibule
where the basal part (layers 10b and 10c) dates to the end of the
Middle Pleistocene (MIS 6 or 7).
3.5. Magnetostratigraphy
In winter 1998, the south prole of sector II in Kozarnika was
chosen for a paleomagnetic study. In this prole, between depths of
589 and 677 cm, 8 single block samples were taken (h¼10 cm). From
paleomagnetic analysis, the sedimentation oflayers 10ato the middle
part of 11b took place in a period of normal magnetic polarity
(Brunhes), as the deposits are younger than 780,000 years. However
the low values of the paleo-latitude in the lower part of layer 11b has
been affected by a magnetic transition of inverted polarity
(Matuyama) towards normal polarity. Unfortunately, some problems
of consolidation of the blocks from the bottom part did not allow
continuation of this study.
4. Palaeontology and biochronology
On the basis of the study of about 163,017 elements, approxi-
mately 50% of the collected large mammals, (NRT ¼163,017;
NISP ¼10,939), a minimum of 69 taxa were identied (Table 1). This
is a minimum number because Cervidae, some Bovinae and some
small Bovidae from the lower levels of the Locus II are still not
precisely determined. There are a large number of petried bones
which cannot be attributed to a known taxon despite the size of the
comparative collection.
Amongst the carnivores of layers 11be13 of sector II were several
Panther-sized felines (Guadelli, 2009b), morphologically different
which can only be attributed to Panthera schaubi. This rare feline,
sometimes described as Puma pardoides or Viretailurus schaubi,is
known only in Middle Villafranchian (biozone MNQ17) or between
1.6 a nd 1.4 Ma (Viret, 1954; Hemmer, 1964; Ficarelli and Torre, 1968;
Kurten,1977; Argantand Ballesio, 1996; Argant, 2004). Layer 13 also
contained a tooth P/4 of Chasmaporthetes lunensis (Fig. 6). This
Hyaenidae, the lower cheek tooth of which has the look of feline
(Viret, 1954), appears during the Ruscinien (biozone MNQ15) and
disappears at the end of the Upper Villafranchien (biozone MNQ18)
(Crégut-Bonnoure, 1996). Also, the genus Xenocyon, a taxon known
in Europe during the second half of the Lower Pleistocene, appears in
layer 12 (Boudadi-Maligne, 2009).
In 2009 a left upper canine of a Primate was found in layer 12. At
rst sight, it was attributed to Cercopithecidae Colobinae, and most
likely Dolichopithecus sp. (Fig. 7). This discovery (that is still under
study) is unique because in Piro Nord (Italia), Cueva Victoria (Spain),
Ahl Oughlam (Maroc) and Ubeidiya (Israel) the animal present was
Theropithecus (De Guili et al., 1987; Gibert I Clols, 1994; Alemseged
and Geraads, 1998; Rook et al., 2004; Belmaker, 2010). In Tenevo
(Bulgaria), some elements were found of a new species of Dolicho-
pithea, Dolichopithecus balcanicus, but these occurred in a level dated
from the beginning of the Pliocene (Spassov and Geraads, 2007).
Among Cervidae, layers 11be13 contained 4 taxa typical of the
Lower Pleistocene, Croizetoceros ramosus (MNQ16e19), Cervus philisi
Fig. 6. Kozarnika. Chasmaporthetes lunensi s. Right LP4, layer 13 (labial view) (photo J.-L.
Fig. 7. Kozarnika. Dolichopithecus sp.(?) Left Upper Canine, layer 12 (labial, anterior,
antero-lingual views) (photo J.-L. Guadelli).
Fig. 8. Kozarnika. Procamptoceras cf. brivatense. Left Metatarsal, layer 13 (proximal
view) (photo J.-L. Guadelli).
N. Sirakov et al. / Quaternary International 223-224 (2010) 94e106 99
(MNQ16e20), Eucladoceros sp. (MNQ17e20), and Cervalces sp.
(MNQ18 in 23) (Simon, 2008).
In layers 12 and 13, Procamptoceras brivatense (Fig. 8)(Fernandez
and Crégut-Bonnoure, 2007) was identied, a form recognized from
Middle Villafranchian (biozone MNQ17) to Upper Villafranchian
(biozone MNQ18) (Duvernois and Guérin, 1989). Gazellospira (Fig. 9),
the temporal extension of which is included between MNQ16 and
MNQ19 (Crégut-Bonnoure, 2007), as identied in layer 12. This tooth
is still under study, and for now cannot be attributed to one of the
two species already described: G. torticornis Aymard, 1854 present
throughout Europe and Asia (Duvernois and Guérin, 1989)andG.
gromovae Dmitrieva, 1975 recognized in Russia and in Tadjikistan
(Sotnikova et al.,1997; Vislobokova, 2005; Takai et al., 2008).
In layers 13, 12, 11c and possibly 11bthere are also the remnants
of a stenonoid Equidae, large and gracile and probably undescribed
(Fig. 10).
The large mammals coming from the layers and/or from the
biostratigraphical units of Kozarnika, are probably associated with
the biozones dened by Guérin (1980, 1982), Guadelli et al. (2005),
Fernandez and Crégut-Bonnoure (2007), Fernandez (2009).
Following some criticism (Azzaroli, 1977; Daams and Freudenthal,
1981), the concept of the biozone was re-developed by Guérin
(1988). Thus, if the principle that every taxon is associated with one
or several biozones is accepted, then the shorter the chronological
extension of a taxon, the more certain the rate of denition of the
biozone with regard to the same taxon.
The following ratio is obtained:
Dt [k=Xt
Dt:rateofdenition of the biozone with regard to one single
k¼1 (constant of the presence of the considered taxon)
t: nbr of biozone(s) in which the considered taxon occurs
The rate of denition of a biozone in relation to all taxa presents
inside (Db) can be then reduced to a percentage by the following
Db [XDt 3100=Totalnumberoftaxa
For lower levels, up to layer 11b, and examining the occurrence
of the taxa of layers 13, 12, 11c and 11b in their relationship to
biozones (Tables 2e5), layers 13 and 12, can be correlated
respectively to biozones MNQ17 and 18, whereas layers 11c and
11b are correlated with biozone MNQ19. On the other hand, in
Table 6, in which layers 13e11c are gathered into the same bio-
zone eKozarnika B2-2, on the basis of associations between
micromammals (Guadelli et al., 2005; Popov and Marinska, 2007),
the result correlates with MNQ18 or possibly to MNQ17 pro parte.
This table shows a certain inadequacy in grouping layers bio-
chronologically, or at least it must be acknowledged that inac-
curacies are sometimes introduced and these raise more problems
than they solve.
5. Lithic industry of the lower Palaeolithic
Staying within the framework of the question of the earliest
settlements of Europe, discussion is limited to the industries of the
Lower Palaeolithic. Information concerning other parts of the
archaeological sequence has been presented (Sirakova and Tsanova,
2002; Tsanova, 2003, 2008; Guadelli et al., 2005; Sirakov et al.,
2007; Guadelli, 2009a,b; Taneva, 2009).
The Lower Palaeolithic sequence of Kozarnika delivered a total
of about 10,000 artefacts (Ivanova, 2009) but about 1/3 of these
artefacts have difcult provenances resulting from disrupted,
deformed contexts, where the recognition of the primary strati-
graphic position is imprecise. Only 2/3 of the artefacts are derived
from well recognized contexts. In this group, between 40 and 50%
are small pieces recovered from sieving, and these are mostly
aking waste and other debris. 15e25% of the artefacts collected
from clearly stratied sediments are tiny (generally between
2 and 5 cm, rarely more than 6e7 cm in greatest extent) and are
also waste, cassonsor other undiagnostic or uncharacteristic
This high rate of fragments from primary aking and the other
undenable residue derived from the following steps of the
Fig. 9. Kozarnika. Gazellospira sp. Right M/3 (E12-3658, c.12). Lingual, vestibular, mesial, occlusal views. (Photo Ph. Fernandez).
Fig. 10. Kozarnika. Equus cf. stenonis. Left LP4-M3, layer 12 (occlusal view) (photo J.-L. Guadelli).
N. Sirakov et al. / Quaternary International 223-224 (2010) 94e10 6100
reduction sequence (chaîne opératoire) is characteristic of the
industries of the Lower Palaeolithic of Kozarnika, although it
continues in the more recent levels until the end of the Upper
Palaeolithic sequence. As previously indicated (Guadelli et al.,
2005) this characteristic is explained by the peculiarities of the
local int which makes up at least 98 % of the raw material and is
very often cracked and fragile. The knapping of this int was gov-
erned by some cleavage planes or surfaces of preferential fracture,
both external and especially internal, which proved difcult for the
knapper to control.
5.1. Lower and middle levels of the Lower Palaeolithic sequence
In the lower and middle levels (layers 13, 12, 11c and the basal
part of 11b), several technological approaches were recognized
which t into common methods of aking and produce recogniz-
able patterns of debitage.
5.1.1. Simple unipolar to bipolar debitage
These are akes produced by direct knapping without prepara-
tion, or sometimes with limited preparation of the striking platform,
but where the surface of the core that will be removed is considered
suitable. For the latter, akes from the primary deliberate aking
action were used, and include plate-shaped, massive akes, natural
fragments and occasionally more-or-less whole nodules. These
methods present variations according to the orientation of the
knapping in relation to the longitudinal axis of the core:
Knapping on the shortest axis (Fig. 11: 2a, 2b; Fig. 12: 1a, 1b),
Knapping on the longest axis (Fig. 11: 3a, 3b).
5.1.2. Debitage from successive rotation of striking and removal
Comparable methods of knapping have been described from
the Clactonian assemblages by Forestier (1993), who grouped
them in the system of knapping surface alternation.Itis
a matter of the succession of the rotations of the knapping
platform, rather than on nearby surfaces of the core. The debitage
from each surface is mostly unipolar, but in Kozarnika, the same
platforms show some bipolar, converging, and multidirectional
knapping (Fig. 11: 1a and 1b). Generally the polyhedral cores are
the result of shorter aking sequences. On the other hand, after
several rotations the knapping sequence seems almost chaotic
and cores become spherical.
Table 2
Kozarnika. Probability of denition of the biozone of the layer 13 (identications M. Boudadi-Maligne, Ph. Fernandez and J.-L. Guadelli).
Layer 13 MNQ 15 16 17 18 19 20 21 22 23 24 25 26 27
Panthera schaubi 0,333 0,333 0,333
Panthera gombaszoegensis 0,143 0,143 0,143 0,143 0,143 0,143 0,143
Homotherium cf. crenatidens 0,200 0,200 0,200 0,200 0,200
Felis cf. lunensis 0,143 0,143 0,143 0,143 0,143 0,143 0,143
Lynx issiodorensis 0,250 0,250 0,250 0,250
Chasmaporthetes lunensis 0,250 0,250 0,250 0,250
Pachycrocuta perrieri 0,200 0,200 0,200 0,200 0,200
Canis etruscus 0,250 0,250 0,250 0,250
Ursus etruscus ? 0,250 0,250 0,250 0,250
Ursus deningeri 0,167 0,167 0,167 0,167 0,167 0,167
Croizetoceros ramosus 0,250 0,250 0,250 0,250
Cervus philisi 0,200 0,200 0,200 0,200 0,200
Hemitragus orientalis 0,500 0,500
Soergelia sp. 0,500 0,500
Procamptoceras brivatense 0,333 0,333 0,333
Equus cf. stenonis 0,200 0,200 0,200 0,200 0,200
Sum Dt 0,250 3,026 3,860 3,502 2,586 1,252 0,452 0,452 0,310 0,310 0,000 0,000 0,000
Db ¼S.Dt * 100 / total Nber taxa 1,563 18,914 24,122 21,890 16,161 7,827 2,827 2,827 1,935 1,935 0,000 0,000 0,000
Table 3
Kozarnika. Probability of denition of the biozone of the layer 12 (identications M. Boudadi-Maligne, Ph. Fernandez and J.-L. Guadelli).
Layer 12 MNQ 15 16 17 18 19 20 21 22 23 24 25 26 27
Panthera schaubi 0,333 0,333 0,333
Panthera gombaszoegensis 0,143 0,143 0,143 0,143 0,143 0,143 0,143
Homotherium cf. crenatidens 0,200 0,200 0,200 0,200 0,200
Felis cf. lunensis 0,143 0,143 0,143 0,143 0,143 0,143 0,143
Lynx issiodorensis 0,250 0,250 0,250 0,250
Pachycrocuta perrieri 0,200 0,200 0,200 0,200 0,200
Canis etruscus 0,250 0,250 0,250 0,250
Xenocyon sp. 0,333 0,333 0,333
Ursus etruscus ? 0,250 0,250 0,250 0,250
Ursus deningeri 0,167 0,167 0,167 0,167 0,167 0,167
Cervalces (Alces cf. latifrons) 0,167 0,167 0,167 0,167 0,167 0,167
Cervus philisi 0,200 0,200 0,200 0,200 0,200
Eucladoceros sp. 0,000 0,250 0,250 0,250 0,250
Hemitragus orientalis 0,500 0,500
Soergelia sp. 0,500 0,500
Gazellospira sp. 0,250 0,250 0,250 0,250
Procamptoceras brivatense 0,333 0,333 0,333
Equus cf. stenonis 0,200 0,200 0,200 0,200 0,200
Sum Dt 0,000 2,776 3,860 4,002 3,336 2,002 0,619 0,619 0,476 0,310 0,000 0,000 0,000
Db ¼S.Dt * 100 / total Nber taxa 0,000 15,423 21,442 22,235 18,532 11,124 3,439 3,439 2,646 1,720 0,000 0,000 0,000
N. Sirakov et al. / Quaternary International 223-224 (2010) 94e106 101
Table 4
Kozarnika. Probability of denition of the biozone of the layer 11c (identications M. Boudadi-Maligne, Ph. Fernandez and J.-L. Guadelli).
Layer 11c MNQ 15 16 17 18 19 20 21 22 23 24 25 26 27
Panthera schaubi 0,333 0,333 0,333
0,143 0,143 0,143 0,143 0,143 0,143 0,143
Felis cf. lunensis 0,143 0,143 0,143 0,143 0,143 0,143 0,143
Lynx issiodorensis 0,250 0,250 0,250 0,250
Canis etruscus 0,250 0,250 0,250 0,250
Ursus deningeri 0,167 0,167 0,167 0,167 0,167 0,167
Eucladoceros sp. 0,250 0,250 0,250 0,250
Equus cf. stenonis 0,200 0,200 0,200 0,200 0,200
Sum Dt 0,000 0,843 1,426 1,569 1,736 0,902 0,452 0,452 0,310 0,310 0,000 0,000 0,000
Db ¼S.Dt * 100 /
total Nber taxa
0,000 10,536 17,827 19,613 21,696 11,280 5,655 5,655 3,869 3,869 0,000 0,000 0,000
Table 5
Probability of denition of the biozone of the layer 11b (identications Ph. Fernandez and J.-L. Guadelli).
Layer 11b MNQ 15 16 17 18 19 20 21 22 23 24 25 26 27
Panthera schaubi 0,333 0,333 0,333
0,143 0,143 0,143 0,143 0,143 0,143 0,143
Felis cf. lunensis 0,143 0,143 0,143 0,143 0,143 0,143 0,143
Lynx issiodorensis 0,250 0,250 0,250 0,250
Ursus deningeri 0,167 0,167 0,167 0,167 0,167 0,167
Croizetoceros ramosus 0,250 0,250 0,250 0,250
Cervus philisi 0,200 0,200 0,200 0,200 0,200
Eucladoceros sp. 0,250 0,250 0,250 0,250
Equus cf. stenonis 0,200 0,200 0,200 0,200 0,200
Sum Dt 0,000 1,043 1,626 1,769 1,936 1,102 0,452 0,452 0,310 0,310 0,000 0,000 0,000
Db ¼S.Dt * 100 /
total Nber taxa
0,000 11,587 18,069 19,656 21,508 12,249 5,026 5,026 3,439 3,439 0,000 0,000 0,000
Table 6
Kozarnika. Probability of denition of the biozone of the biostratigraphic entity B2-2 (identications M. Boudadi-Maligne, Ph. Fernandez and J.-L. Guadelli).
Biozone B2-2 MNQ 15 16 17 18 19 20 21 22 23 24 25 26 27
Panthera schaubi 0,333 0,333 0,333
0,143 0,143 0,143 0,143 0,143 0,143 0,143
Homotherium cf.
0,200 0,200 0,200 0,200 0,200
Felis cf. lunensis 0,143 0,143 0,143 0,143 0,143 0,143 0,143
Lynx issiodorensis 0,250 0,250 0,250 0,250
0,250 0,250 0,250 0,250
Pachycrocuta perrieri 0,200 0,200 0,200 0,200 0,200
Canis etruscus 0,250 0,250 0,250 0,250
Xenocyon sp. 0,333 0,333 0,333
Ursus etruscus 0,250 0,250 0,250 0,250 0,000
Ursus deningeri 0,167 0,167 0,167 0,167 0,167 0,167
Croizetoceros ramosus 0,250 0,250 0,250 0,250
Cervalces (Alces cf.
0,167 0,167 0,167 0,167 0,167 0,167
Cervus philisi 0,200 0,200 0,200 0,200 0,200
Eucladoceros sp. 0,250 0,250 0,250 0,250
Hemitragus orientalis 0,500 0,500
Soergelia sp. 0,500 0,500
Gazellospira sp. 0,250 0,250 0,250 0,250
0,333 0,333 0,333
Equus cf. stenonis 0,200 0,200 0,200 0,200 0,200
Sum Dt 0,250 3,276 4,360 4,502 3,586 2,002 0,619 0,619 0,476 0,310 0,000 0,000 0,000
Db ¼S.Dt * 100 /
total Nber taxa
1,250 16,381 21,798 22,512 17,929 10,012 3,095 3,095 2,381 1,548 0,000 0,000 0,000
N. Sirakov et al. / Quaternary International 223-224 (2010) 94e10 6102
It is necessary to emphasise that, statistically, knapping on the
shorter axis of the core seems to occur most frequently, based on the
analysis of the limited lithic series described above. This preference
seems to suggest the likely existence of a desire for more-or-less
predetermined products, namely short akes. Moreover, this
approach could minimize the échisand other knapping acci-
dents, which occur much more frequently in the case of knapping
on the long axis.
Generally, the debitage in the lower and middle levels consists
of shorter aking sequences of short removals (mean 2e5 cm). This
is not only a reection of an elementary technological level, but
probably also a way to avoid mis-strikes and thus to increase the
efciency of production by a deliberate simplication of the
reduction sequence.
5.2. Terminal levels of the Lower Palaeolithic sequence
In these levels (upper part of the layer 11b, layer 11a) the ancient
knapping methods used from the beginning of the sequence
continued. At the same time new technological tendencies
emerged. These manifest as isolated cores in layer 11b and in
a small series in the most recent levels of the layer 11a which show
elements of debitage tending towards Levallois and Discoid
concepts. The best expressed and most numerous cores exhibit
knapping and\or core preparations which approach Levallois
methods, among which:
- knapping is reminiscent of the method for producing prefer-
ential akes (Fig. 12: 2),
Fig. 11. Kozarnika. Knapping methods in the lower and the middle layers of the Lower Palaeolithic. Simple unipolar to bipolar debitage: 2a, b eon the short axis of the support; 3a,
beon the long axis of the support. 1a, b: debitage by successive rotation and change of striking and removal platforms on polyhedral to spherical nuclei (drawing: S. Taneva, Ts.
Tsanova, N. Sirakov).
N. Sirakov et al. / Quaternary International 223-224 (2010) 94e106 103
- attempts at lengthened removals parallel to the long axis after
a partial preparation on the perimeter of the core (Fig. 12:3)
- initial preparation (although interpretable also as an endscraper)
on the perimeter of a cortical primary ake, a very characteristic
occurrence for the series of Levallois cores with debitage on the
ventral side, found from the lower levels of the Mousterian
sequence (Fig. 12:4).
Flaking techniques directed towards the Discoid concept are less
obvious. The most decipherable tendency is debitage derived from
Fig. 12. Kozarnika. Knapping methods in the lower and the middle levels of the Lower Palaeolithic: 1a, b ebipolar simple debitage on the short axis of a lengthened surface.
Emergence of new technological tendencies in the terminal assemblages of Lower Palaeolithic: 2, 3, 4 enuclei with preparation and debitage which aim towards the Levallois
concept (drawing: S. Taneva, Ts. Tsanova, N. Sirakov).
Fig. 13. Kozarnika. Emergence of new technological tendencies in the terminal assemblages of Lower Palaeolithic. 1a, b, 2: elements directed to the Discoid concept edebitage of
the nuclei on which the striking planes and debitage surfaces with uneven convexity were probably only partially, or not exchangeable (drawing: S. Taneva, Ts. Tsanova, N. Sirakov).
N. Sirakov et al. / Quaternary International 223-224 (2010) 94e10 6104
rather conical cores with more or less uneven convexity on both
knapping and striking surfaces, which were probably somewhat
inter-changeable (Boëda, 1993) in their function (Fig. 13).
6. Conclusions
According to the most recent chronological results the
sequence is dated between 1.6 and 1.4 Ma and 0.5 Ma, namely
MIS 53-45 to 15-13 (Jian et al., 2001). The long-term permanent
occupation in Kozarnika does not support the hypotheses of long
breaks and discontinuities in the advance of settlement, at least in
the Balkans.
Recent excavations have brought new light to human activities.
Several bone fragments with cuts that are not the result of butch-
ering were discovered. Among them is the shaft part of a bovine
shinbone from layer 12 with 4 regular series of 4 cuts (Fig. 14),
interpreted as the oldest example of symbolism extant (Guadelli and
Guadelli, 2004; Guadelli et al., 2005). Avoiding here any debate on
this fragment, the term symbolismis used in its strictest meaning.
No semiotic function for the symbol (its meaningfulness) is implied,
but the symbol is a visible and tangible reality suggesting other
invisible realities. Layer 12 contained the rst phalanx of a Marmot
with oblique cut-marks in its disto-lateral part (Fig. 15)indicativeof
skinning, which makes it the oldest evidence for skin working, not
previously known until this discovery.
From the typological aspect, there is one more important direct
use of the akes and the un-retouched pieces from aking, which
correlates with the analysis of the non-Acheulian lithic industries of
Italy (Peretto et al., 1998). However, the retouched forms from
Kozarnika, akes being most numerous, core-tools, a few typical
side scrapers, rare end-scrapers and borers, are generally better
represented than in the Italian sites.
The total absence of pebble-tools and bifaces throughout the
sequence is notable. Although the beds of the rivers around Kozar-
nika have accumulations rich in pebbles, these were used only
sporadically. The absence of pebble-tools can be explained by the
absence of techno-economic traditions of pebble exploitation, rather
than by difculties of access to the raw material or by other peculiar
palaeo-environmental conditions. From this point of view the
notions of core-chopperand core-and-ake industriesare not
identical (see for example Bar-Yosef and Belfer-Cohen, 2001)and
their use cannot be univocal. If it was necessary to classify the Lower
Palaeolithic of Kozarnika as a Mode, it would be a Mode 1 which:
- was a core-and-akeindustry, withoutpebble-cores and pebble-
- lasted approximately 1 Ma,
- was not replaced by Acheulian
- and ends directly in the Mousterian/Levallois with leaf shaped
Regarding the classication scheme of Mode 1, 2 etcfor the
early lithic industries and cultural traditions, the case of Kozarnika
brings further argument and lack of precision, indicating the
inadequacy of this terminology, which is still à la mode, as already
noted elsewhere (Bar-Yosef and Belfer-Cohen, 2001).
As regards the cultural attribution in the whole sequence, there
is a very slow evolution of a type of non-Acheulian core-and-ake
industry (not based on pebble-cores). Dated about 1.6e1.4 Ma, the
lowest layers of Kozarnika Cave bring new data about the situation
in the Balkans and their role in the rst settlements of Europe.
This research was undertaken within the framework of an inter-
national cooperation between the National Institute of Archaeology
of the Bulgarian Academy of Sciences (N. Sirakov) and the UMR5199
CNRS PACEA/IPGQ (J.-L. Guadelli), the works of the Franco-Bulgarian
Prehistoric Mission in Northern Bulgaria are nancially supported by
the Advisory Committee of the Archaeological Researches abroad
(MAEE, France) eDGRCST, by the CNRS (Centre National de la
Recherche Scientique), by the Region Aquitaine, by the University
Bordeaux 1 (exceptional funding in 2004), by the Max Planck Insti-
tute for Evolutionary Anthropology, Department of Human Evolution
(2004), by the Bulgarian Academy of Sciences and by the foundation
Stichting Horizon. To all of them we send our deepest thanks.
Aguirre, E., Carbonell, E., 2001. Early human expansions into Eurasia: the Atapuerca
evidence. Quaternary International 75, 11e18.
Alemseged, Z., Geraads, D., 1998. Theropithecus atlanticus (Thomas, 1884)
(Primates: Cercopithecidae) from the late Pliocene of Ahl al Oughlam, Casa-
blanca, Morocco. Journal of Human Evolution 34, 609e621.
Argant, A., 2004. Les Carnivores du gisement Pliocène nal de Saint-Vallier (Drôme,
France). The Carnivores of the late Pliocene site of Saint-Vallier (Drôme, France).
Geobios 37, S133eS182.
Argant, A., Ballesio, R., 1996. La famille des Felidae. In: Guérin, C., Patou-Mathis, M.
(Eds.), Les Grands Mammifères Plio-Pléistocènes dEurope. Masson Paris éd.,
Collection Préhistoire, pp. 195e215.
Fig. 14. Kozarnika. Layer 12. Symbolism. Shaft part of a shin of bovine with 4 regular
series of 4 cuttings (photo A. Guadelli).
Fig. 15. Kozarnika. Layer 12. Skinning activity. First phalanx of Marmot with oblique
cut-marks (photo J.-L. Guadelli).
N. Sirakov et al. / Quaternary International 223-224 (2010) 94e106 105
Azzaroli, A., 1977. Mammal units versus biozones. In: Alberdi,M.T., Aguirre, E. (Eds.),
Round Table of the W. Mediterranean Neogene, Trabajos Neógeno-Cuaternario,
vol. 7, pp. 25e27.
Bar-Yosef, O., Belfer-Cohen, A., 2001. From Africa to Eurasia early dispersals.
Quaternary International 75, 19e28.
Belmaker, M., 2010. The presence of a large cercopithecine (cf. Theropithecus sp.) in
the Ubeidiya formation (Early Pleistocene, Israel). Journal of Human Evolution
58, 79e89.
Boëda, E., 1993. Le débitage discoïde et le débitage Levallois récurrent centripète.
Bulletin de la Société préhistorique française 90, 392e404.
Boudadi-Maligne, M., 2009. Les Loups (Canis lupus) des niveaux paléolithiques
supérieurs des grottes de Kozarnika (Gara Orechets) et de Redaka II. Approche
paléontologique et taphonomique. In: Gatsov, I., Guadelli, J.-L. (Eds.), Saxa
Loquuntur! The Stone Will Speak!. Avalon, Soa, pp. 109e117.
Carbonell, E., Bermudez de Castro, J.M., Pares, J.P., Perez-Gonzalez, A., Cuenca-
Bescos, G., Olle, A., Mosquera, M., Huguet, R., Van der Made, J., Rosas, A., Sala, R.,
Valverde, J., Garcia, N., Granger, D.E., Martinon-Torres, M., Rodriguez, X.P.,
Stock, G.M., Vergès, A., Benito, A., Diez, C., Lozano, M., Mateos, A., Navazo, M.,
Rodriguez, J., Rosell, J., Arsuaga, J.L., 2008. The rst hominin of Europe. Nature
452, 465e469.
Crégut-Bonnoure, E., 1996. La famille des Hyaenidae. In: Guérin, C., Patou-
Mathis, M. (Eds.), Les Grands Mammifères Plio-Pléistocènes dEurope, Masson
Paris éd., Collection Préhistoire, pp. 215e223.
Crégut-Bonnoure, E., 2007. Apport des Caprinae et Antilopinae (Mammalia, Bovi-
dae) à la biostratigraphie du Pliocène terminal et du Pléistocène dEurope.
Quaternaire 18/1, 73e97.
Daams, R., Freudenthal, M., 1981. Aragonian: the stage concept versus Neogene
Mammal zones. Scripta Geology 62, 1e17.
De Guili, C., Masini, F., Torre, D.,1987. The latest Villafranchian faunas of Italy. The
Pirro Nord fauna (Apricena, Gargano). Palaeontographia Italica 74, 51e62.
Djambazov, N.,1958. Risunkite v peshterata Magura. Soa. In: Priroda i znanie, vol.
4, pp. 21e22.
Duvernois, M.-P., Guérin, Cl., 1989. Les Bovidae (Mammalia, Artiodactyla) du Villa-
franchien supérieur dEurope occidentale. Géobios 22, 339e379.
Fernandez, Ph, 2009. Mammalian dynamics and palaeoecological analysis during
the Pleistocene in Kozarnika Cave (Bulgaria). In: Gatsov, I., Guadelli, J.-L. (Eds.),
Saxa Loquuntur! The Stone Will Speak!. Avalon, Soa, pp. 59e73.
Fernandez, Ph., Crégut-Bonnoure, E., 2007. Les Caprinae (Rupicaprini, Ovibovini,
Ovini et Caprini) de la séquence pléistocène de Kozarnika (Bulgarie du Nord):
morphométrie, biochronologie et implications phylogéniques. Revue de Palé-
obiologie 26, 425e503.
Ferrier, C., Leblanc, J.-Cl., Berthet, A.-L., Delfour, G., 2009. Kozarnika. Étude géo-
logique et sédimentologique. In: Gatsov, I., Guadelli, J.-L. (Eds.), Saxa Loquuntur!
The Stone Will Speak!. Avalon, Soa, pp. 13e27.
Ficarelli, G., Torre, D., 1968. Upper Villafranchian panthers of Tuscany. Palae-
ontographia Italica XXXIV, 173e184.
Forestier, H., 1993. Le Clactonien: mise en application dune nouvelle méthode de
débitage sinscrivant dans la variabilité des systèmes de production lithique du
Paléolithique ancien. PALEO 5, 53e82.
Gibert I Clols, J., 1994. Signicado de la fauna de Cueva Victoria. Memorias de
Arqueologia, 26e32.
Guadelli, A., 2009a. Les artefacts en matière dure animale du paléolithique en
Bulgarie. Thèse de doctorat de lInstitut National dArchéologie, Académie
Bulgare des Sciences, Soa. 2 volumes: vol. 1, texte, illustrations 376 pp, vol. 2,
catalogue, illustrations, 173 pp.
Guadelli, J.-L., 2009b. Les Félins de la séquence pléistocène de Kozarnika. In:
Gatsov, I., Guadelli, J.-L. (Eds.), Saxa Loquuntur! The Stone Will Speak!.
Avalon, Soa, pp. 75e100 .
Guadelli, A., Guadelli, J.-L., 2004. Une expression symboliquesur os dans le
Paléolithique inférieur de la grotte Kozarnika, Bulgarie du Nord-Ouest. Colloque
international La spiritualité. U.I.S.P.P. VIIIème Commission: Paléolithique
supérieur. Liège, 10-11-12 décembre 2003, vol. 106. ERAUL, pp. 87e95.
Guadelli, J.-L., Sirakov, N., Ivanova, St., Sirakova, Sv., Anastassova, E., Courtaud, P.,
Dimitrova, I., Djabarska, N., Fernandez, Ph., Ferrier, C., Fontugne, M., Gambier, D.,
Guadelli, A., Iordanova, D., Iordanova, N., Kovatcheva, M., Krumov, I., Leblanc,
J.-Cl., Mallye, J.-B., Marinska, M., Miteva, V., Popov, V., Spassov, R., Taneva, St.,
Tisterat-Laborde, N., Tsanova, Ts., 2005. Une séquence du Paléolithique inférieur
au Paléolithique récent dans les Balkans: La grotte Kozarnika à Oreshets (Nord-
Ouest de la Bulgarie). In: Molines, N., Moncel, M.-H., Monnier, J.-L. (Eds.), Les
premiers peuplements en Europe. Actes du colloque international Données
récentes sur les modalités de peuplement et sur le cadre chronostratigraphique,
géologique et paléogéographique des industries du Paléolithique inférieur et
moyenenEurope, Rennes, 22e25 septembre 2003, vol. S1364. British Archae-
ological Research, pp. 87e103.
Guérin, Cl, 1980. Les rhinocéros (Mammalia, Perissodactyla) du Miocène terminal au
Pléistocène supérieur en Europe occidentale: Comparaison avec les espèces
actuelles. Documents des laboratoires de Géologie de Lyon, 1185 pp.
Guérin, Cl, 1982. Première biozonation du Pléistocène européen, principal résultat
biostratigraphique de létude des Rhinocerotidae (Mammalia, Perissodactyla)
du Miocène terminal au Pléistocène supérieur dEurope occidentale. Géobios 15
(4), 593e598. Lyon.
Guérin, Cl, May 1988. Biozones or mammals units? Methods and limits in Bio-
chronology. In: Lindsay, E.H., Fahlbusch, V., Mein, P. (Eds.), European Neogene
Mammal Chronology, NATO Advanced Research Workshop European Neogene
Mammal Chronology, Munich. Plenum Press, New York, pp. 119e130.
Hemmer, H., 1964. Studien an Pantheraschaubi Viret aus dem Villafranchien von
Saint-Vallier (Drôme). Neues Jahrbuch für Mineralogie, Geologic und Pal-
äontolologie. Abhandlungen 122, 324e336.
Ivanova, S., 2009. Ranno paleolitni ensembli ot Peshtera Kozarnika. In: Gatsov, I.,
Guadelli, J.-L. (Eds.), Saxa Loquuntur! The Stone Will Speak!. Avalon, Soa,
pp. 29e57. 22 g.
Jian, Z., Cheng, X., Zhao, Q., Wang, J., Wang, P., 2001. Oxygen isotope stratigraphy
and events in the northern South China Sea during the last 6 million years.
Sciences in China, (D) 44, 952e960.
Kurten, B., 1977. Bären und Hyänenreste aus dem Pleistozän von Taubach. Quar-
tärpaläont 2, 361e378.
Peretto, C., Amore, F.O., Antoniazzi, A., Bahain, J.-J., Cattani, L., Cavallini, E.,
Esposito, P., Falgueres, C., Gagnepain, J., Heddley, I., Laurent, M., Lebreton, V.,
Longo, L., Milliken, S., Monegatti, P., Olle, A., Pugliese, N., Renault-Miskovsky, J.,
Sozzi, M., Ungaro, S., Vannucci, S., Verges, J.M., Wagner, J.-J., Yokoyama, Y., 1998.
Lindustrie lithique de CaBelvedere di Monte Poggiolo: Stratigraphie, matière
première, typologie, remontages et traces dutilisation. LAnthropologie 102,
Popov, R., 1933. La grotte Mirizlivka. Contribution à létude de la faune diluvienne et
de la culture de lhomme quaternaire en Bulgarie, vol. 26. Cahier du musée
national dArchéologie, Soa, pp. 5e69, (résumé en français).
Popov, R., 1936. Fosilni i subfosilni jivotinski ostanki b izledvanite do cega pechteri
na Bulgaria. Isvestia na Bulgarskoto Pechterno Drujestvo, Kniga 1, 1e12.
Popov, V.V., Marinska, M., 2007. An almost one million year long (Early to Late
Pleistocene) small mammal succession from the archaeological layers of
Kozarnika Cave in Northern Bulgaria. Courier Forschungsinstitut Senckenberg
259, 79e92.
Rook, L., Martinez-Navarro, B., Howell, C., 2004. Occurrence of Theropithecus sp. in
the late Villafranchian of Southern Italy and implication for early Pleistocene
‘‘out of Africa’’ dispersals. Journal of Human Evolution 47, 267e277.
Shchelinsky, V.E., Dodonov, A.E., Boigusheva, V.S., Kukalov, S.A., Simokova, A.N.,
Tesakov, A.S., Titov, V.V., 2010. Early Paleolithic sites on the Taman peninsula
(Southern Azov Sea region): Bogatyri/sinyaya Balka and Rodniki. Quaternary
International, 223e224. 28e35.
Simon, F., 2008. Contribution à létude des Cervidae de la séquence pléistocène de
Kozarnika (Bulgarie du Nord). Paléontologie, Biochronologie et Paléoenvir-
onnement. Mémoire de Master 2 Sciences et Technologies, Mention Anthro-
pologie Biologique ePréhistoire Spécialité: Préhistoire. Université Bordeaux I,
PACEA équipe IPGQ, UMR5199 CNRS, 73 pp.
Sirakov, N., Tsanova, Ts., Sirakova, S., Taneva, S., Krumov, I., Dimitrova, I.,
Kovatcheva, N., 2007. Un nouveau faciès lamellaire du début du paléolithique
supérieur dans les Balkans. PALEO 19, 129e142.
Sirakova, S., Tsanova, T., 2002. Lensemble lithique de la zone gravettienne de
contact IV a et IV b de la grotte Kozarnika. Reconstruction et phases du proc-
essus technologique. Annuaire de lAcadémie Bulgare des Sciences, Soa, II (in
Sotnikova, M.V., Dodonov, A.E., Penkov, A.V., 1997. Upper Cenozoic bio-magnetic
stratigraphy of Central Asian mammalian localities. Palaeogeography, Palae-
oclimatology, Palaeoecology 133, 243e258.
Spassov, N., Geraads, D., 2007. Dolichopithecus balcanicus sp. nov., a new Colobinae
(Primates, Cercopithecidae) from the early Pliocene of southeastern Europe,
with a discussion on the taxonomy of the genus. Journal of Human Evolution
52, 434e442.
Takai, M., Maschenkob, E.N., Nishimurac, T.D., Anezakid, T., Suzuki, T., 2008.
Phylogenetic relationships and biogeographic history of Paradolichopithecus
sushkini Tromov 1977, a large-bodied cercopithecine monkey from the Plio-
cene of Eurasia. Quaternary International 179, 108e119.
Taneva, S., 2009. Étude des ensembles à pièces à dos du Paléolithique supérieur
en Bulgarie dans le contexte du techno-complexe gravettien en Europe.
Thèse de doctorat de lInstitut National dArchéologie, Académie Bulgare des
Sciences, Soa, 2 volumes: vol. 1, 293 pp.; vol. 2, catalogue, illustrations,
211 pp.
Tsanova, T., 2003. Le Gravettien en Bulgarie du Nord: niveau IVb de la grotte
Kozarnika. In: Tsonev, T., Kokelj, E.M. (Eds.), Proceedings of the ESF Workshop,
Soa, 3e6 Sept. 2003. The Humanized Mineral World: towards Social and
Symbolic Evaluation of Prehistoric Technologies in South Eastern Europe, vol.
103. ERAUL, pp. 33e39.
Tsanova, T., 2008. Les débuts du Paléolithique supérieur dans lEst des Balkans.
exion à partir de létude taphonomique et techno-économique des
ensembles lithiques de Bacho Kiro (couche 11), Temnata Dupka (couches VI
et 4) et Kozarnika (niveau VII), vol. 1752. BAR International Series, Oxford,
326 pp.
Viret, J., 1954. Le lœss à bancs durcis de Saint-Vallier (Drôme) et sa faune de
mammifères villafranchiens, vol. 4. Nouvelles archives du Museum dHistoire
naturelle de Lyon, 200 pp.
Vislobokova, I., 2005. On Pliocene faunas with Proboscideans in the territory of the
former Soviet Union. Quaternary International 126e128, 93e105.
N. Sirakov et al. / Quaternary International 223-224 (2010) 94e10 6106
... The geostrategic position of the Balkans at the southeastern gateway of Europe has reasonably been emphasized and defined as ... "one of only two pathways into and out of Europe during the Pleistocene" (Ivanova et alii 2012, p. 2; see also Sirakov et alii 2010). Being heuristic abstractions, the migratory routes continuously serve as a conceptual framework for many discussions and systematic research. ...
... It has been suggested that several African and Asian mammalian lineages (e.g., Canis, Equus, Panthera, Mammuthus and several species of bovids) entered Europe through the Balkans during the colder and drier glacial stages of the Last Ice Age, when land bridges between Southeastern Europe and Asia Minor became periodically exposed (Spassov 2003. It is reasonable to hypothesize that Palaeolithic human dispersal into and out of Europe was at least in part driven by the environmental conditions in the Balkan Peninsula. ...
... The site is situated near to two more emblematic settlements in the Belogradchik karst area -Kozarnika and Magura caves, each of which has had an important impact on prehistoric research a broader scale. Kozarnika Cave offers an impressive sequence of faunal fossils and human activity from the Early Palaeolithic to the Epigravettian (Guadelli et alii 2005;Sirakov et alii 2010;. The recently published Middle Pleistocene human infant remains from Kozarnika shed light and arguments on the hypothesis of early Neanderthal presence in the eastern Balkans (Tillier et alii 2017). ...
Full-text available
Bulgaria encompasses most of the Balkan migratory corridor (part of the so-called Danubian corridor) which facilitated the dispersal of mammals and Homo populations into and out of Europe. In spite of the scarcity of human fossil remains, the rich evidence of more than 20 Palaeolithic sites in Bulgaria suggests long-term human presence and various activities. Northwestern Bulgaria is known for the massive karstic formations in the area around the present-day town of Belogradchik. Three important Palaeolithic settlements have been excavated there, namely the caves of Kozarnika, Magura and Mishin Kamik. This paper presents the results of five short but intensive excavation campaigns in Mishin Kamik Cave. No human fossil remains or lithic artefacts have been recovered. However, the excavations produced several bone artefacts and a number of intriguing features, including an apparently intentional arrangement of bear skulls, which suggest the cave was visited by humans at the end of the Middle Pleistocene.
... The cave gained interest for its human occupation potential after the discovery of potential bone artefacts and an intruiguing accumulation of bear skulls and bones (Gurova et al., 2015Gurova and Ivanova, in press) [Supporting Information (SI) 1]. Mishin Kamik cave site is located a few kilometres to the west of known human occupation sites such as Kozarnika (Guadelli et al., 2005;Sirakov et al., 2010), Temnata (Ginter et al., 2011) and Bacho Kiro (Hublin et al., 2020) and therefore provides an opportunity to fill chronological and spatial gaps regarding patterns of human occupation and dispersal along the Balkan route and its relationship to palaeoenvironmental conditions. ...
The Balkan Peninsula represents one of the most important human pathways into and out of Europe during the Pleistocene. Mishin Kamik cave, located in the karst region of Western Stara Planina, has a rich faunal content and shows promising features indicating a human occupation site with the discovery of potential bone artefacts and an intriguing accumulation of bear skulls and bones. Petrographic study and U-series dating of a stalagmite and other calcite deposits in the cave provide an absolute chronological frame for the detrital infillings and their archaeological content and inform the environmental and climatic context of the cave evolution. Most detrital deposits in the cave were probably deposited before Marine Isotope Stage (MIS) 5 and the cave morphology and sedimentary deposits display current morphologies since~135 ka. Consequently, the palaeontological and archaeological findings are older than~135 ka. Calcite dated on and under the accumulation of bear skulls and bones suggests deposition during MIS 7. A first depositional contextualization of the bone accumulation does not allow us to discriminate between a natural or anthropogenic origin. The study emphasizes the added value of speleothem studies in archaeological sites and particularly in bringing a well-constrained chronological and environmental framework.
... 12: Happisburgh 3 and Pakefield, United Kingdom (∼0.8 Ma; Lewis et al., 2019), which tool assemblages comprise flint flakes, flake tools, cores, and a handaxe in the case of Happisburgh 1 (∼0.5 Ma, MIS 13). 13: Kozarnika Cave, Layers 13a-c, Bulgaria (∼0.75 Ma;Muttoni et al., 2017), which has provided an abundant industry (∼10,000 artifacts, associated to a rich faunal assemblage) that shows a predominance of fragments from primary flaking (the local flint is very fragile), with the flakes obtained by simple unipolar to bipolar debitage (Sirakov et al., 2010). 14: Denizli, Turquey (1.2-1.1 Ma;Lebatard et al., 2014), which has provided the Kocabaş skull fragment, the only known Turkish fossil of Homo erectus. ...
Full-text available
The chronology and environmental context of the first hominin dispersal in Europe have been subject to debate and controversy. The oldest settlements in Eurasia (e.g., Dmanisi, ∼1.8 Ma) suggest a scenario in which the Caucasus and southern Asia were occupied ∼0.4 Ma before the first peopling of Europe. Barranco León (BL) and Fuente Nueva 3 (FN3), two Early Pleistocene archeological localities dated to ∼1.4 Ma in Orce (Guadix-Baza Depression, SE Spain), provide the oldest evidence of hominin presence in Western Europe. At these sites, huge assemblages of large mammals with evidence of butchery and marrow processing have been unearthed associated to abundant Oldowan tools and a deciduous tooth of Homo sp. in the case of BL. Here, we: (i) review the Early Pleistocene archeological sites of Europe; (ii) discuss on the subsistence strategies of these hominins, including new estimates of resource abundance for the populations of Atapuerca and Orce; (iii) use cartographic data of the sedimentary deposits for reconstructing the landscape habitable in Guadix-Baza; and (iv) calculate the size of the hominin population using an estimate of population density based on resource abundance. Our results indicate that Guadix-Baza could be home for a small hominin population of 350–280 individuals. This basin is surrounded by the highest mountainous reliefs of the Alpine-Betic orogen and shows a limited number of connecting corridors with the surrounding areas, which could have limited gene flow with other hominin populations. Isolation would eventually lead to bottlenecks, genetic drift and inbreeding depression, conditions documented in the wild dog population of the basin, which probably compromised the viability of the hominin population in the medium to long term. This explains the discontinuous nature of the archeological record in Guadix-Baza, a situation that can also be extrapolated to the scarcity of hominin settlements for these ancient chronologies in Europe.
... Avec les quelques nucléus et éclats mis au jour dans la couche 13 de Kozarnika située dans le Nord-Ouest la Bulgarie, il s'agirait des plus anciennes traces d'une présence humaine dans cette partie de l'Europe centrale. Les données biostratigraphiques avaient situé l'ensemble lithique de Kozarnika entre 1,4 et 1,7 Ma (Sirakov et alii 2010) mais une analyse paléomagnétique indique que la couche 13 contenant les pièces taillées remonterait au début de la période de Bruhnes soit un âge compris entre 600 et 750 ka (Muttoni et alii 2017). ...
Full-text available
The oldest human remains unearthed in Europe were discovered in the Iberian Peninsula. Their age is between 1.4 and 1.2 Ma. Lithic assemblages older than 1 Ma have been unearthed in southwestern Europe (Italy, southern France, Spain). They have often been attributed to Mode 1 but are contemporary with the African Acheulean. In North-West Europe, north of the Loire, the colonization of the high latitude regions seems to have taken place more recently, from 740 ka as shown by the flint artefacts discovered in East Anglia and the first manifestations of the Acheulean. The few data, both lithic and biological, indicate that in central Europe the oldest settlements are much later, of the order of 500 ka which would seem to exclude a settlement of Europe by an eastern route. The lithic industries of Central Europe are characterized by an absence of bifacial pieces. We have little information for the eastern part of the Carpathians, but some lithic series are older than 600 ka.
Kozarnika cave is a renowned prehistoric site in the Balkans, which contributes significantly to our understanding of the human past due to its rich assemblages associated with the Lower to Upper Palaeolithic. Various dating methods have been employed to unravel the timing of human occupations in Kozarnika. Radiocarbon dating was used to unfold the time frame for the Kozarnikian lithic tradition uncovered in the Upper Palaeolithic sequence of the cave, and palaeomagnetic dating assigned the Brunhes–Matuyama reversal to the layer beneath the Lower Palaeolithic assemblages. In this study, we employed luminescence-dating methods, including a body of different signals to date sediment layers from the top to the bottom of the sequence covering the period of ca 30 to 700 ka. Our investigations revealed that the Kozarnikian tradition in layers 5a-c falls between 30 and 35 ka. Following that, we suggest that the Middle Palaeolithic period initiated between 250 and 309 ka and lasted until 40–53 ka. More importantly, we have updated the age of the Neanderthal radius discovered in the Mousterian assemblages to 201 ± 17 ka. Our dating resulted in a period of ca 300–700 ka for the Lower Palaeolithic assemblages in the cave. Although this age range fits perfectly with the palaeomagnetic boundary <780 ka established for Kozarnika, the possibility of reaching the threshold of luminescence dating cannot be ruled out. Thus, at this stage, this age range may represent the minimum age for the Lower Palaeolithic.
Full-text available
Little is known about the subsistence practices of the first European settlers, mainly due to the shortage of archaeological sites in Europe older than a million years. This article contributes to the knowledge of the subsistence of the first Europeans with new zooarchaeology and taphonomic data from the Palaeolithic site of Barranco León (Orce, Granada, Spain). We present the results of the analysis of the faunal assemblages retrieved in the context of new excavations undertaken between 2016 and 2020. We have followed a standard methodology for the identification and quantification of species, mortality profiles, skeletal representation and taphonomic analysis. With regard to the taphonomic evidence, we have documented the extent of rounding, abrasion and other alterations. Finally, we examined traces from the activities of carnivores and hominins that led to the accumulation and alteration of the bone assemblages. Results indicate that the archaeo-paleontological deposits from Barranco León present a dual-patterned mixed taphonomic origin. The first phase primarily involved waterborne processes (BL-D1), which led to the accumulation of lithic raw materials, a few archaeological stone tools, and some faunal remains with percussion and cutmarks. The second phase (BL-D2) contains several stone tools associated with faunal remains with more anthropogenic alterations, such as cutmarks and percussion marks. After analysing the Barranco León zooarchaeological assemblage, the present study concludes that hominins had access to the meat and within-bone nutrients of animals of diverse sizes. However, the specific carcass acquisition mechanisms that hominins followed are less certain because the presence of tooth marks suggests that carnivores also played a role in the accumulation and modification of the Barranco León faunal assemblage.
The expansion of Homo sapiens and our interaction with local environments, including the replacement or absorption of local populations, is a key component in understanding the evolution of our species. Of special interest are artifacts made from hard animal tissues from layers at Bacho Kiro Cave (Bulgaria) that have been attributed to the Initial Upper Paleolithic. The Initial Upper Paleolithic is characterized by Levallois-like blade technologies that can co-occur with bone tools and ornaments and likely represents the dispersal of H. sapiens into several regions throughout Eurasia starting by 45 ka or possibly earlier. Osseous artifacts from the Initial Upper Paleolithic are important components of this record and have the potential to contribute to our understanding of group interactions and population movements. Here, we present a zooarchaeological, technological, and functional analysis of the diverse and sizable osseous artifact collection from Bacho Kiro Cave. Animal raw material sources are consistent with taxa found within the faunal assemblage including cervids, large bovids, and cave bears. A variety of bone tool morphologies, both formal and informal, indicate a diverse technological approach for conducting various on-site activities, many of which were focused on the processing of animal skins, likely for cold weather clothing. Technological flexibility is also evident in the manufacture of personal ornaments, which were made primarily from carnivore teeth, especially cave bear, though herbivore teeth and small beads are also represented. The osseous artifacts from Bacho Kiro Cave provide a series of insights into the bone technology and indirectly on the social aspects of these humans in southeast Europe, and when placed within the broader Initial Upper Paleolithic context, both regional and shared behaviors are evidently indicating widespread innovation and complexity. This is especially significant given the location and chronology of the site in the context of H. sapiens dispersals.
Full-text available
The main focus of this paper is to present the stone assemblage of one of the Lower Palaeolithic layers of Medzhibozh 1. Detailed typological and technological characteristics of the stone industry of layer III of the locality Medzhibozh 1 are given against the background of geomorphological characteristics and features of the Quaternary structure of the area. Medzhibozh 1 represents one of several multi-layered sites of Lower Palaeolithic age on the Upper Bug, among which are also Medzhibozh A, Golovchintsy 1 and Golovchintsy 2. Two layers containing Lower Palaeolithic evidence are distinguished in the profile of Medzhibozh to date. Medzhibozh 1 layer III is more recent, the age of culture-bearing sediments is currently correlated to MIS 11. This dating is fundamentally consistent with diverse biostratigraphic indications as well as several pilot ESR dates ranging between 373–399 thousand years. Geomorphological features of the structure of the Medzhibozh area allow for the age estimation of culture-bearing sediments as older and presumably attribute them to MIS 15–13. Despite its relatively young age, the Medzhibozh 1 layer III industry demonstrates quite archaic technology and typology. The knapping technology is almost totally characterised by the application of the bipolar on anvil technique. The basic range of lithic products includes bipolar cores, segmented supports, choppers, segments, flakes, and debris. Despite the abundance of items with signs of utilisation, pieces with secondary worked edges are isolated and atypical. Well-defined types of flake tools are practically not present. There are no freehand cores. There is no indication of use of bifacial technology. Currently, there is absolutely no formal ground to attribute the industry to the Acheulean. Instead, the Medzhibozh 1 layer III industry can currently be compared to a “core-and-flake” industry, what is traditionally defined as Oldowan or mode 1. Perhaps the reason for the observed archaism can to some extent be explained by the properties of the available raw materials (granitoides of crystalline shield and mostly small-sized siliceous supports of Upper Cretaceous age) and the economic profile of occupation or recovered area (recurrent butchering site).
This paper contains the results of the archaeological campaigns carried out at the El Pino Site (Campos del Paraíso, Carrascosa del Campo, Cuenca, Spain). The stratigraphy belongs to the Valdejudíos River floodplain and is rich in Mode 1 prehistoric industry made from quartzite rolling stones. The site has been dated by Optically Stimulated Luminescence (OSL) and Electron Spin Resonance (ESR), giving the last technique a terminus ante quem of 1 Ma for this technological horizon. The excellent preservation of the lithic surfaces has allowed us to discover the use given to the instruments through traceological methods. An effective usage of retouched and unretouched flakes (not the cores) for butchery, and rabots for processing wood is verified. The manufacturing operational systems of Mode 1 pieces have been studied and subsequently evaluated as being similar to the oldest ones of the Oldowan industries in Africa (2.6 Ma); however, they showed more “archaism” than other contemporary worldwide examples. The ineffectiveness of the idea of a linear evolution of material culture “from simplicity to complexity”, to assess the lithic technology of the European Lower Paleolithic, is revealed.
The current state of research on Lower Palaeolithic sites in Ukraine within its 1991 borders is the focus of this paper. Over the last 10–15 years, many new sites have been discovered in different parts of the eastern European area of the country, reassessed some old materials. In the central European region of the country, in the Ukrainian Transcarpathia, important new stratified Lower Palaeolithic sites have also been found. The current Ukrainian Lower Palaeolithic records demonstrate hominin presence in mountainous areas (Carpathians, Crimea) and the valleys of all major rivers, namely the Dniester, Southern Buh, Dnieper and Severskiy Donets. The article presents a brief review of the main currently known Lower Palaeolithic assemblages. Available geological, geomorphological, biostratigraphical data and ESR dates allow defining their age between 1.2 and 0.4 Myг; sites correlate with few warm phases between MIS 35 and MIS 11. Earlier sites, very tentatively dated at around 2 Myг, gravitate towards the seashore and mountainous areas. Later sites witness steady, though not continuous, colonisation of East European plain fringe areas. The main regularities of geographical setting, chronology, morphological and technological characteristics of assemblages of the Lower Palaeolithic sites of the Western segment of the East European plain are characterised. Typologically, industries are mainly characterised as belonging to Mode I. Core-and-flake industries survives to the Holsteinian. Essential difficulties in lithic raw materials supply could probably be a reason for the rise of a peculiar pattern of technological behaviour that involved mainly bipolar knapping and widely applied trimming technique of shaping the working edges of tools. Some signals of probable population movements penetrated the territory of Ukraine by the Asia Minor “western” trajectory and by Caucasian “eastern” way are revealed.
A paleontológical study of Caprinae (Rupicaprini, Ovibovini, Ovini and Caprini) was undertaken on the anthropic site of Kozarnika (Bulgaria). The sequence covers the whole Pleistocene and the base could correspond to the MNQ 18. The morphometric data of Caprinae were studied in a systematic way and the biochronological and phylogenic aspects of some taxa of this subfamily were considered in comparison with the main sites of western and eastern Europe. Thanks to a new method founded on the chronological extension of the large mammals from the biostratigraphic units B2-2, B2-1, B1, A2 and A1 of Kozarnika, we estimated their probability of being associated to the mammalian biozones defined by GUERIN (1980, 1982, 1988). Thus, the sequence of Kozarnika constitutes one of the fundamental biostratigraphic stakes for the large mammals of the Balkans. The older levels are correlated to the beginning of the Lower Pleistocene that is to say the MNQ 18.
The first attempt to construct a biozonation scheme for European mammal-bearing deposits was made by Thaler (1964). Neogene and lower Pliocene sites of southern France and of Spain were concerned, being situated in a series of rodent range zones each designated after the name of its most typical site. Some years later Heintz (1970) proposed a zonation based on Cervids for the Villafranchian of France and Spain; that biozonation was generalized with the whole Mammal class, excepting the Rodents, by Heintz et al. in 1974. The names given to the zones were those of the most important concerned fossiliferous places. Besides, some studies were carried out concerning the zonation of eastern Europe, notably by Fejfar (1976) and Fejfar and Heinrich (1981), utilizing some rodent families.