Technological change in Iberomaurusian culture: The case of Tamar Hat, Rassel and Columnata lithic assemblages (Algeria)

Abstract

This paper highlights the results of a technological analysis carried out for the first time on nine lithic assemblages belonging to three Iberomaurusian sites in Algeria (Tamar Hat rockshelter, Rassel cave and Columnata open-air site). The purpose of this paper is to identify the technical system adopted for the production of lithic artifacts in each of the studied assemblages in order to search for similarities and discrepancies in the technical behaviours between coastal and inland Iberomaurusian populations. This requires recognition of the reduction core strategies and transformation of the produced blanks, as well as research into debitage methods and techniques employed. In all three sites, local raw materials were preferred and the lithic production was primarily geared towards obtaining relatively standardized lamellar blanks with different debitage schemes. In lower occupations at Tamar Hat, the bladelets were mainly produced by implementing a single and complex chaîne opératoire of lamellar production which integrated variable blanks according to numerous and different schemes. The dominant scheme was an elaborate debitage oriented to the production of short and narrow bladelet blanks from reduced prismatic
and sub-pyramidal cores, while by-products were exploited as bladelet cores to produce microbladelets
and burin spalls. A change in the core reduction strategies appeared in the upper occupations
of Tamar Hat where elongated bladelets were produced according to a common simple debitage
widespread in the latest occupations. The implementation of a single chaîne opératoire is preserved in the lower layer of Rassel, although with a less elaborate roughing out processing, as for the upper occupations of Tamar Hat. In contrast, at Columnata three independent chaînes opératoires were implemented to produce robust blanks, which represents a stark contrast with previous methods known in the other sites. The technological analysis has provided strong arguments for a different know-how in technical behaviours between Iberomaurusian populations living in coastal rockshelters and those in hinterlands open-air sites. Thus, the different geographical areas seem to cover variable economic entities which would suggest a new adaptation of the same populations to different ecological niches.

Full text

Technological change in Iberomaurusian culture: The case of Tamar
Hat, Rassel and Columnata lithic assemblages (Algeria)
Latifa Sari
Centre National de Recherches Préhistoriques, Anthropologiques et Historiques, Prehistory, 3 rue Franklin Roosevelt, Alger, Algeria
article info
Article history:
Available online 25 April 2013
abstract
This paper highlights the results of a technological analysis carried out for the first time on nine lithic
assemblages belonging to three Iberomaurusian sites in Algeria (Tamar Hat rockshelter, Rassel cave and
Columnata open-air site). The purpose of this paper is to identify the technical system adopted for the
production of lithic artifacts in each of the studied assemblages in order to search for similarities and
discrepancies in the technical behaviours between coastal and inland Iberomaurusian populations. This
requires recognition of the reduction core strategies and transformation of the produced blanks, as well
as research into debitage methods and techniques employed. In all three sites, local raw materials were
preferred and the lithic production was primarily geared towards obtaining relatively standardized
lamellar blanks with different debitage schemes. In lower occupations at Tamar Hat, the bladelets were
mainly produced by implementing a single and complex chaîne opératoire of lamellar production which
integrated variable blanks according to numerous and different schemes. The dominant scheme was an
elaborate debitage oriented to the production of short and narrow bladelet blanks from reduced pris-
matic and sub-pyramidal cores, while by-products were exploited as bladelet cores to produce micro-
bladelets and burin spalls. A change in the core reduction strategies appeared in the upper occupa-
tions of Tamar Hat where elongated bladelets were produced according to a common simple debitage
widespread in the latest occupations. The implementation of a single chaîne opératoire is preserved in the
lower layer of Rassel, although with a less elaborate roughing out processing, as for the upper occupa-
tions of Tamar Hat. In contrast, at Columnata three independent chaînes opératoires were implemented to
produce robust blanks, which represents a stark contrast with previous methods known in the other
sites. The technological analysis has provided strong arguments for a different know-how in technical
behaviours between Iberomaurusian populations living in coastal rockshelters and those in hinterlands
open-air sites. Thus, the different geographical areas seem to cover variable economic entities which
would suggest a new adaptation of the same populations to different ecological niches.
Ó2013 Elsevier Ltd and INQUA. All rights reserved.
1. Introduction
Various typological frameworks for the Iberomaurusian culture
of the Maghreb were developed in several regional typological
syntheses based essentially on the lithic assemblages. Thus, in
reference to both typological analysis elaborated successfully by
Tixier (1963) who classified by types the various retouched tools,
and the first absolute dates, Camps (1974) saw implicitly several
regional facies which maintain a linear chronological relationship
in three successive phases. Conversely, Brahimi (1969a,1972) and
Aumassip (1979) designed a model leading to the identification of
regional groups independent from chronological phases, whereas
Close (1977) found a style to be a key factor in the variability of
backed bladelets. Since the 1980s, the linear conception of Camps is
increasingly rejected by many authors, who see the Iberomaurusian
as a set of cultural facies with regional particularities. These authors
refer to systematic excavations conducted using a multidisciplinary
approach in which the environment is considered as a major factor
determining the economic and technical choices of Iberomaurusian
populations (Lubell, 1984;Hachi, 1987;Hachi et al., 2002;Moser,
2003;Barton et al., 2007,2008;Nami, 2007).
Referring to chrono-stratigraphic and typological arguments,
most of these authors agree to divide the Iberomausian of the
Maghreb into two main evolutionary phases related to paleo-
environmental changes. In Grotte des Pigeons at Taforalt, the
transition of the Iberomaurusian from the yellow to the grey series
of deposits is dated around 13,000 BP. This transition coincides with
E-mail address: sari.latifa@cnrpah.org.
Contents lists available at SciVerse ScienceDirect
Quaternary International
journal homepage: www.elsevier.com/locate/quaint
1040-6182/$ esee front matter Ó2013 Elsevier Ltd and INQUA. All rights reserved.
http://dx.doi.org/10.1016/j.quaint.2013.04.014
Quaternary International 320 (2014) 131e142

a broader cultural change characterized by a scarcity of Ouchtata
bladelets, microburins and Mouillah points in the grey series where
the pointed and straight backed bladelets are well represented
(Barton et al., 2008;Bouzouggar et al., 2008;Taylor et al., 2011).
This transition which occurred during the climatic amelioration of
the Bolling-Allerod coincides also with a perceived change in sub-
sistence and mortuary practices (Taylor et al., 2011). The decrease of
Ouchtata bladelets for acute backed bladelets and segments is also
reported in the upper layers of Ifri n’Ammar cave, while in Ifri el
Baroud cave this transition is reflected by a general trend towards
reduction of blank sizes as well as decreasing frequencies of backed
laminar blanks and notched/denticulated pieces (Nami, 2007,p.
191). In Afalou rockshelter, the changes that begin around
12,020 170 BP and 13,120 370 BP (layer IV) are perceived by a
decrease in the number of backed bladelets for notched/denticu-
lated pieces as well as the development of sophisticated weapons
such as Ain Keda and Chacal points with many triangles and rare
trapezoids. The triangles are associated with fish remains and
malacofauna and are related to a life-style oriented towards fishing
(Hachi, 2003, p. 230).
Although the current chronostratigraphic data support the
periodization assumption, the variability in the Iberomaurusian
lithic assemblages still remains essentially guided by the variation
of typological tool groups of which backed bladelets are dominant.
However, the quantitative and/or qualitative variation of the tool
groups can depend on the function of the site, the excavated area
and for the backed bladelets on their rate of fragmentation.
Consequently, a question emerges: is the typological variability
observed in the Iberomaurusian lithic assemblages diachronic or
rather linked to regional particularities varying from coastal and
inland sites?
Here we outline some avenues for reflection on this question by
using a technological approach which allows recognizing the
choices made by the Iberomaurusian knappers throughout the core
reduction processing and the transformation of produced blanks as
well as the research of debitage methods and techniques employed.
Technology is also dedicated to the study of relationships between
the technical system and socio-economic phenomena. “This is one
of the most fruitful and rapidly developing means of approaching
prehistoric life-styles”(Inizan et al., 1995, p. 16).
This approach has been applied for the first time at three
Algerian Iberomaurusian sites: Tamar Hat rockshelter, Rassel cave
and the Columnata open-air site, in the framework of a PhD
dissertation (Sari, 2012). The lithic assemblages rich in backed
bladelets were chosen for study according to the variability of their
paleogeographic, chronological and cultural conditions (Fig. 1).
They were entirely attributed to the Iberomaurusian based on tool
group variation (Arambourg et al., 1934;Cadenat, 1948;Brahimi,
1969b,1970,1972;Close, 1977,1978,1980e81). In this paper, we
report results of the technological analysis carried out on these
three sites focusing on lithic production modalities.
2. Archaeological setting
Tamar Hat rockshelter was first excavated in 1928e1930 by
Arambourg et al. (1934). In 1967, the site was re-excavated by
Brahimi (1969b) who dug only a small sounding in the upper part of
the deposits. In 1973, Saxon et al. (1974) excavated in front of the
rockshelter a surface area of 4.50 m
2
which yielded abundant
archaeological material rich in faunal and lithic remains. The stra-
tigraphy exposed is composed of 85 layers grouped into six strati-
graphic zones belonging to the Iberomaurusian. Our study is
concerned with lithic material collected during the 1973 excavation
only. Rassel cave was excavated by Camps et al. (1959), then by
Brahimi (1970); this last identified a disturbed upper layer and an
Iberomaurusian lower layer in situ. Columnata was first noted by
Cadenat (1948,1966 ) who carried out four excavation campaigns
until 1961 in front of the sandstone cliff and discovered rich
archaeological material relating to Iberomaurusian, Columnatian,
Upper Capsian and Neolithic, as well as a necropolis belonging to all
of these. The Iberomaurusian deposits were not overlapped by
those of more recent cultures which prolonged them horizontally
towards the sandstone cliff. Later, Brahimi (1972) has resumed
archaeological excavations in the Iberomaurusian deposits. He
excavated two undisturbed areas in 1969 (south sector: 5 m
2
) and
1971 (north sector: 4 m
2
) ten meters north of the Iberomaurusian
area discovered by Cadenat. All of the archaeological material from
south and north sectors excavated by Brahimi belongs to the Iber-
omaurusian and all lithic material analyzed in our study came from
these two sectors.
The occupations of the three sites cover the entire duration of
the Iberomaurusian over a period of some 10,000 years from the
Last Glacial Maximum to the early Holocene (Table 1), although the
single date for the lower layer of Rassel obtained on snail shell must
be taken with caution (Rahmouni et al., 1972a, p. 354). By the same
token so should be the sole date from Columnata (Rahmouni et al.,
Fig. 1. General location of the main Iberomaurusian sites. Triangles indicate the three sites discussed in the text. (map modified after ASTER GDEM Product of METI and NASA).
L. Sari / Quaternary International 320 (2014) 131e142132

1972b, p. 1). Tamar Hat and Rassel are both in a mixed landscape of
mountain, forest and coastal plains, while Columnata is an open-air
site on the high inland semi-arid plateau, about 200 km from the
sea and protected by a high sandstone cliff. According to zooarch-
aeological analysis, Tamar Hat is a butchering and consumption site
and this function does not seem to have changed during 4500 years.
The Tamar Hat populations practiced selective hunting directed
primarily towards Barbary sheep in late autumn and early winter.
The other taxa such as Algerian megaceroides (Megaceroïdes
algericus), red fox (Vulpes vulpes), wild boar (Sus scrofa) and few
antelope are less well represented (Merzoug, 2005;Merzoug and
Sari, 2008). Rassel cave is in topographic conditions similar to
those of Tamar Hat. In the absence of zooarchaeological analysis,
the seasonality and function of this site cannot be confirmed;
however, the mammalian assemblage studied by Coppens (cited by
Brahimi, 1970, p. 80) includes Barbary sheep, the most dominant
taxa at Tamar Hat. Moreover, specialized activities including those
related to hunting are attested (Sari, 2008,2012). In contrast, in late
summer and early autumn Columnata populations practiced
specialized hunting of Alcelaphus which were butchered and eaten
at the site; other ungulate taxa, less well represented, include
aurochs and gazelles (Merzoug, 2005,2008).
3. Methods and materials
3.1. Methods
Technological analysis begins with the processing of raw ma-
terial data acquired in fieldwork and takes into account the raw
material procurement modalities according to their textural, geo-
metric and volumetric properties (e.g. Demars, 1982;Féblot-
Augustin, 1997;Bressy, 2003). This is essential for the
delimitation of the territory and the mobility of human groups
(Perlès, 1980,1991;Féblot-Augustin, 2008).
Techniques are inferred by comparison with numerous experi-
mental replication reference collections carried out by Pelegrin on
European Upper Paleolithic and Mesolithic assemblages (Pelegrin,
2000, p. 77). It appears that when considering Iberomaurusian
butts of end-products submitted to Pelegrin, we find the stigmata
of mineral soft-hammer percussion (Sari, 2012). This approach
yielded for the first time an assessment of the variability in
percussion techniques within analyzed Iberomaurusian lithic
assemblages.
The identification of debitage methods entails a “technical
reading”of each artifact in order to assess its position in the chaîne
opératoire (Tixier, 1978;Tixier et al., 1980, p. 35). It was not possible
to refit stone artifacts of the three studied sites; however, “mental
refitting”(Pelegrin, 1995, p. 23) is involved. This takes into account
many criteria recorded on lithic pieces such as morphology of un-
modified blocks, the presence and the elimination of cortex, as well
as the ordering and the shape of removals.
Sub-pyramidal and prismatic terms describe bladelet cores with
a slightly curved to straight debitage surface running parallel to the
longitudinal axis. Debitage proceeded according to two principles
of core volume management: a semi-tournant and a frontal strategy
both of which employed the thickness of blocks, large flakes or
block fragments (Fig. 2). The terms burin-core and endscraper-core
are attributed to some thick cortical flakes deriving from the
roughing out sequence and typologically considered as “carinated”
burins or endscrapers. Indeed, in recent years study of the lamellar
chaîne opératoire of such artifacts has experienced an increase
among researchers working on various European Upper Paleolithic
assemblages (e.g. Aubry et al., 1995;Lucas, 1999;Le Brun-Ricalens
Fig. 2. Schematic diagram showing expansion of core volume management.
Table 1
Radiocarbon dates of studied Iberomaurusian sites
14
C data are reported as conventional and calibrated BP.
14
C ages are reported with 2
d
(95% probability). Calibration was
performed with CALIB 5 (Stuiver M. and Reimer P.J., 2005).
Site Author/origin Layer Lab. Code
14
CBP
14
C calBP Material
Tamar Hat Brahimi sounding Upper Alg-4
Alg-5
10,350 350 11,078e12,980 Charcoal
Tamar Hat Brahimi sounding Lower Alg-5 12,450 480 13,376e15,880 Charcoal
Tamar Hat Saxon excavation 9 (zone I) MC-817 16,100 360 18,728e19,955 Charcoal
Tamar Hat Saxon excavation 15 (zone I) MC-812 17,040 400 19,409e21,184 Charcoal
Tamar Hat Saxon excavation 44 (zone IV) MC-818 18,750 500 20,929e23,694 Charcoal
Tamar Hat Saxon excavation 50 (zone V) MC-820 19,800 500 22,321e24,980 Charcoal
Tamar Hat Saxon excavation 85 (zone VI) MC-822 20,600 500 23,594e26,000 Charcoal
Rassel Brahimi excavation Lower Alg-3 14,270 600 16,065e18,501 Snail-shell
Columnata Brahimi excavation South sector Alg-97 10,850 425 11,321e13,461 Lacustrine shell
L. Sari / Quaternary International 320 (2014) 131e142 133

and Brou, 2003) and Near Eastern techno-complexes (eg. Tixier and
Inizan, 1981;Bar-Yosef, 1991;Belfer-Cohen, 2003). Such results
have now reached a new level of meaning which owes much to
lithic refitting. It should be noted that, bladelets are parallel-sided
blanks measuring less than 12 mm in width and twice as long as
they are wide (Tixier, 1963,p.36e39).
The examination of damage patterns characteristic of hunting
with chipped stone projectile points is taken into account. Their
recognition has significance for reconstructing subsistence be-
haviours and the understanding of site function. Although
experimental studies are not yet undertaken and we know of no
such tests using Iberomaurusian replicas, it is possible to find
many references in the literature (e.g. Fischer et al., 1984;
Geneste and Plisson, 1990;Pelegrin and O’Farell, 2005). The
diagnostic patterns were sought in all examined Iberomaurusian
lithic assemblages (Sari, 2012). Such patterns were already re-
ported with the identification protocol for the material from
zone I in Tamar Hat rockshelter (Merzoug and Sari, 2008;Sari,
2011).
3.2. Materials
The studied Iberomaurusian lithic assemblages belong to the six
stratigraphic zones of Tamar Hat rockshelter (Saxon excavation),
the lower layer of Rassel cave (Brahimi excavation) and the south
and north sectors of Columnata open-air site (Brahimi excavation).
The abundance of the artifacts is a major asset for the reconstruc-
tion of the technical behaviour of human groups during this period.
The total number of lithic artifacts that we examined is 37,721
pieces of which 35,468 are blanks and 1759 cores.
All phases of lithic core reduction are well represented in the
assemblages which suggest they functioned as a workshop. More-
over, the transformation of blanks took place at the sites as sug-
gested by the presence of microburins and burin spalls (Table 2).
The debitage is dedicated to the production of lamellar blanks and
the dominance of unretouched and retouched lamellar pieces does
not seem to have varied over time (Table 3). Moreover, the high
percentage of unretouched bladelets suggests the existence of a
stock in which end-products are reserved for further use.
Table 2
Lithic composition of studied assemblages.
Technological category Tamar Hat Rassel Columnata Total
Zone VI Zone V Zone IV Zone III Zone II Zone I Lower South North
Cores 97 161 179 69 104 120 224 364 441 1759
Non retouched pieces 3076 3423 4841 1218 2262 3504 697 3369 5183 25,573
Retouched pieces 806 943 1135 308 776 728 1051 855 1305 7907
Burin spalls 1 9 9 3 4 7 5 14 19 71
Debris 2 2 ee 29 26 30 e36 125
Microburins 9 21 19 5 9 30 100 30 63 286
Total 3991 4559 6183 1603 3184 4415 2107 4632 7047 2253
Table 3
Lithic composition of studied assemblages by blank types.
Blank type Tamar Hat Rassel Columnata
Zone VI Zone V Zone IV Zone III Zone II Zone I Lower South North
Rt. Nr. Rt. Nr. Rt. Nr. Rt. Nr. Rt. Nr. Rt. Nr. Rt. Nr. Rt. Nr. Rt. Nr.
Total flakes 48 846 102 1308 89 1528 15 429 89 570 90 1492 125 223 153 1574 273 2305
First cortical 7 53 14 99 12 204 3 56 17 34 17 220 11 25 10 28 10 27
Tablette 1 5 1 24 2 18 e5 1 12 3 23 8 15 1 16 2 30
Rejuvenation 5 13 11 23 18 22 1 15 9 15 5 142 19 55 14 56 31 96
Semi-cortical 8 277 33 532 36 616 5 143 19 331 25 677 30 51 30 364 66 733
Non-cortical 27 498 43 630 21 668 6 210 43 178 40 430 57 77 98 1110 164 1419
Total blades 12 159 19 206 24 343 2 98 21 111 23 194 27 92 27 274 56 334
First cortical e4e10 e15 e3e3e23 eee2ee
Crest e6e425e3e6e4e8 1 11 4 32
Neocrest e5ee e 14 e10 1 14 e1 1 11 4 18 2 12
Undercrest ee ee 2e1e2ee1eee231
Sided blade ee ee e 2e5ee ee 14e326
Semi-cortical 1 70 4 126 7 141 e42 2 76 9 117 1 24 4 22 15 57
Non-cortical 11 74 15 66 13 166 1 35 16 12 14 48 24 45 18 216 30 226
Total bladelets 745 2071 822 1909 1022 2970 291 691 666 1581 615 1818 896 374 675 1521 976 2544
First cortical bladelets e2e9e12 e4e3e5e3e10 e9
Crest 10 73 4 66 6 86 5 26 3 46 5 47 18 56 4 68 4 79
Neocrest 95013e21 147 12 37 16 74 15 25 40 83 12 70 17 90
Undercrest 4126133 1211 37 312e3e7e9
Sided bladelet 16 27 2 22 62 6 16 23 7 16 6 17 32 11 29 13 24
Semi-cortical 38 273 52 188 67 342 5 99 17 490 40 344 38 30 35 79 44 254
Non-cortical 668 1634 745 1633 903 2309 262 508 604 954 536 1379 783 167 613 1258 898 2079
Indeterminate blanks 1 eeeeeeeeeee38eeee
Total blanks 806 3076 943 3423 1135 4841 308 1218 776 2262 728 3504 1051 697 855 3369 1305 5183
Rt. Retouched pieces; Nr. Non retouched pieces.
L. Sari / Quaternary International 320 (2014) 131e142134

4. Results
4.1. Raw material procurement strategies
The occupants of Tamar Hat exploitedfor the most part siliceous
rocks abundantly available as pebbles in the local alluvial and
marine deposits. The exploitation of pebbles reflects a choice made
for their shapes which are mostly spherical or have rounded edges.
Unworked pebbles are rare which suggests the absence of a stock
due to the proximity of outcrops. A first sorting of these pebbles
was made at the outcrop as evidenced by the absence of tested
pebbles on the site, except for exogenous rocks. The same raw
material procurement strategies are preserved throughout all the
human occupations (Fig. 3a); however, beginning from the occu-
pation of zone IV, Liasic flint was exploited in the same proportions
as the better quality Eocene flint, which is a break with the econ-
omy of Eocene flint seen in the previous occupation (zones VI and
V). The raw material procurement strategy which exploited local
and regional mineral resources was maintained by the occupants of
the lower layer at Rassel cave who had a preference for local alluvial
flint (Fig. 3b). The constitution of a reserve was unnecessary given
the proximity of these alluvial flint outcrops. This is corroborated
by the absence of unmodified blocks and tested pebbles. In
contrast, occupants of Columnata favored local voluminous blocks
of two chalcedony varieties (cryptocrystalline and microcrystalline)
compared to other better siliceous rocks coming from farther away
(Fig. 3c). The collected rocks were knapped at the site as evidenced
by the presence of various sequences of chaîne opératoire, with the
exception of Eocene flint transported to the site as preforms. There
is no great variation in the representation of raw material exploi-
tation rate in the south and north sectors.
Local raw materials were preferred at all three sites even if they
were not always satisfactory, as is the case for chalcedony blocks of
Columnata. The estimation of displacement shows that, for the raw
material, they did not exceed 30 km, except for exogenous siliceous
rocks of which the origin is still unknown. These last are very few
and were introduced to the site as end-products. The maintenance
of the same raw material procurement low-cost process is possibly
related to a recurrent territorial organization similar in each
occupation and the dominance of local siliceous rocks is rather
associated with a reduced mobility of these populations. Collection
of blocks may have occurred while traveling for hunting, fishing
and gathering plants.
4.2. Core reduction strategies
The use of mineral soft-hammer percussion, a technique well
suited to the production of regular blanks during plein débitage,is
observed from the earliest occupation of Tamar Hat (Sari, 2012).
There is a long survival of mineral soft-hammer percussion in the
Iberomaurusian tradition with variability in technical procedure
within the same technique. Knappers of Tamar Hat and Rassel
removed blanks of plein débitage from the cores near edge of the
striking platform according to a tangential oblique percussion,
while those of Columnata removed blanks in set back from edge of
the striking platform in order to obtain thick butts. A description
and occurrence of stigmata deriving from this technique will be
reported and discussed in a forthcoming paper (Sari, 2013).
Fig. 3. Frequencies of raw material exploitation.
L. Sari / Quaternary International 320 (2014) 131e142 135

In all three sites, lithic production was primarilygeared towards
obtaining relatively standardized lamellar blanks and the knappers
adopted different operative debitage schemes. This variability in-
creases inter-site diversity because some schemes are present and
others are more or less represented or simply absent.
4.2.1. Tamar Hat
During the lower occupations of Tamar Hat (zones VI and V), the
knappers prepared cores from medium-sized nodules selected
from the available local raw material. Core reduction involved a
maximal number of technical gestures especially in the preparation
Fig. 4. Core reduction strategies of Iberomaurusian lithic assemblages.
L. Sari / Quaternary International 320 (2014) 131e142136

of volumes and surfaces to be knapped. This was facilitated by
preferential selection and exploitation of the blocks. The bladelets
were mainly produced by implementing a single and complex
chaîne opératoire of lamellar production which integrated blanks
deriving from the same chaîne opératoire according to different
schemes (Fig. 4A).
The recurrent and dominant scheme is an elaborate debitage
primarily oriented to produce bladelet blanks from reduced
prismatic and sub-pyramidal cores with a broad debitage surface
and single striking platform which is usually facetted. The initial
phase of debitage was performed by striking a first-cortical blade
to initiate the debitage from one of the two core sides. A few
cortical and semi-cortical blades and laminar flakes were suc-
cessively extracted as by-products. The rejuvenation of the
striking platform consisted in removing a core tablet (often
partial), while the maintenance of core convexities was achieved
by striking several thinner rejuvenation products and crested
pieces; these last were often more posterior than lateral. A sec-
ond striking platform was used, when needed, to correct knap-
ping errors mainly to the main debitage surface caused by
plunging or hinging. Bladelet blanks were extracted as end-
products by unidirectional convergent removals that progres-
sively expanded to adjacent faces, thus giving the core a semi
tournant front which indicates a desire to maximize the debitage
surface (Fig. 4A, n

1). The end-products were shorter and regular
blanks obtained from reduced cores proceeded by removing a
series of core tablets and rejuvenation flakes. A second scheme,
using the same roughing out sequences allowed production of
rectilinear bladelets intercalated with the production of longer
bladelets (Fig. 4A, n

2).
Other schemes, restricted to Eocene flint, consisted of the
exploitation of robust by-products deriving from the same lamellar
chaîne opératoire. Some of these by-products were exploited as
bladelet cores either by using the burin-blow technique on trun-
cation to produce thick narrow bladelets with triangular sections,
or by extracting curved bladelets from the lower surface (Fig. 4A,
n

3). Moreover, a few thick by-products considered typologically as
carinated endscrapers were exploited as bladelet cores to produce
micro-bladelets with a curved profile. In this case, the carinated
debitage surface was bounded by two lateral crests obtained by
small orthogonal splintered removals (Fig. 4A, n

4).
The implementation of the single and complex chaîne opératoire
of lamellar production is preserved in the upper occupations of
Tamar Hat beginning from stratigraphic zone IV; however, changes
began from this zone. The elaborate semi tournant debitage strategy
known in the lower occupations was brutally interrupted in favor of
a simple debitage limited to one side and which exploited the
natural convexities of the block from a single and preferentially
plain striking platform. In this case, the debitage process was
stopped at a certain length, after the extraction of one or two
lamellar blank generations of medium size. Other changes also
occur from zone IV: the alternate debitage becomes increasingly
scarce, while the production of micro-bladelets derived from
endscraper-cores disappears. Burin-cores become scarcer and no
longer have a truncation. Moreover, a new scheme appears from
zone II. It is the exploitation of dihedral cores or so called “orange
slices”knapped on Liassic flint pebbles. This scheme is related to
the production of bladelets with a triangular to sub-trapezoidal
section and irregular shape (Fig. 4A, n

5) and has already been
reported for Afalou and is documented by a refitting sequence
(Hachi, 1987,1996).
4.2.2. Rassel
As with the occupants of Tamar Hat, those of the lower level at
Rassel produced different blanks within a single chaîne opératoire.
The debitage is dedicated to the production of lamellar blanks.
However, the core reduction strategy is limited to three schemes
as is the case in upper occupations of Tamar Hat. The common
scheme is the debitage of prismatic and sub-pyramidal cores
performed on broad debitage surface with a single and plain
striking platform. Shorter and regular lamellar blanks are
Fig. 5. Distribution of fragmented and unbroken backed bladelets.
Fig. 6. Backed bladelets metrical data.
L. Sari / Quaternary International 320 (2014) 131e142 137

obtained from the cores of which the volume is reduced by
removal of successive rejuvenation products (Fig. 4A, n

1). In this
scheme, the natural convexities are often exploited thus avoiding
an elaborate roughing out operation. The alternate lamellar deb-
itage is rarely attested and is expected to be devoted to yellow
pinkish flint coming from farther away (Fig. 4A, n

2). Moreover,
the debitage conducted on burin-cores, performed along the edge
of the blank-core, is systematically prepared on a plain striking
platform (Fig. 4A, n

3).
4.2.3. Columnata
Unlike those at Tamar Hat and Rassel, the occupants of Col-
umnata adopted different reduction core strategies for specific
blanks. For each objective, an independent chaîne opératoire was
implemented. The presence of an autonomous debitage of blades
and flakes is in stark contrast with methods known at the other
sites.
Lamellar debitage was applied according to various schemes,
some of which are reserved for a particular type of raw material.
The exploitation of the prismatic and sub-pyramidal cores on a
broad surface debitage was done as at Tamar Hat and Rassel
(Fig. 4B, n

1). The burin-cores are mostly dihedral (Fig. 4B, n

2),
while alternated lamellar debitage is absent. Carinated and globular
debitage are two new schemes of debitage, without roughing out
processing, appearing for the first time in this site. The former is
only applied to bladelet cores on flint pebbles to produce short and
wide lamellar blanks from a carinated debitage surface with a
single striking platform in direction of the greater width of the
block (Fig. 4B, n

3), while the latter is devoted to globular bladelet
cores knapped without visible organization on large blocks of
chalcedony. The striking platform position is in direction of the
greater length of the block (Fig. 4B, n

4).
Autonomous debitage of blades is shown by the presence of a
few cores on chalcedony blocks. The blades were obtained with
hard-hammer percussion, indifferently from a broad or narrow
debitage surface bounded by a lateral or posterior crest (Fig. 4B,
n

5). As well, the autonomous debitage of flakes was frequent and
restricted to large blocks of chalcedony knapped in three main
schemes. The facial scheme provided flakes from large blocks in
direction of the greater length of the block (Fig. 4B, n

6). Blanks
were removed with hard-hammer percussion from one or two
opposed debitage surface used alternately. The use of roughing
out is not attested and the blanks obtained were relatively short.
The multipolar scheme allowed removal of thicker flakes from
large blocks without visible organization of debitage (Fig. 4B,
n

7), while the centripetal scheme provided short and thin flakes.
In the latter case, the back opposite to the debitage surface had a
rounded and shaped appearance trimmed by small splintered
removals (Fig. 4B, n

8).
4.3. Blank transformation: the example of backed bladelets
The intention of the core reduction processing can be under-
stood when examining the different tool groups (Table 4). These are
largely dominated by the lamellar blanks often transformed into
backed bladelets. In this case, the sought lamellar blanks mainly
belong to plein debitage, while bladelets from the rejuvenation
process as well as burin spalls were rarely selected. The total
number of backed bladelets is 6142 of which 3988 pieces are
fragments (Fig. 5).
Metrical data for backed bladelets provide valuable information
for comparison: all complete and regular bladelets from plein
debitage (n¼1214), except for those from roughing out were
measured (Fig. 6,Table 5). The mean lengths, widths, thicknesses of
unbroken backed bladelets from plein debitage have been compared
by the means of t-distribution with a level of significance equal to
0.05 (Table 6).
The mean lengths are significantly different with the test mean
(25.64 mm), except for the two sectors of Columnata; moreover,
samples from zones IV to I of Tamar Hat are much longer than
those from previous zones. The mean widths are significantly
different with the test mean (7.08 mm), except for samples from
Zones IV to I of Tamar Hat; however, samples from the two sectors
of Columnata are larger than the ones of the other assemblages.
Finally, the mean thicknesses are also significantly different with
the test mean (2.85 mm), except for zones IV to I of Tamar Hat
where backed bladelets are thicker than those from the underly-
ing zones, but not as much as those from the two sectors of
Columnata.
Table 4
Tool groups by blank type.
Blank-type/tool-group Es Pf Br Bb Ct Bbl Nd Tr Gm Ms Total
Tamar Hat. Zone VI
Flakes 12 1 2 6 e26e19 48
Blades ee23 e3ee 412
Bladelets e21e717 5 10 1 9 745
Others eeee eeee 44
Total 12 3 5 9 717 10 16 1 37 810
Tamar Hat. Zone V
Flakes 69 1 6 1 e58e12 102
Blades 7 e14 e41e219
Bladelets 6 1 7 e749 12 24 3 20 822
Others eeee eeee 44
Total 82 2 14 5 749 21 33 3 38 947
Tamar Hat. Zone IV
Flakes 69 e1e73e989
Blades 10 1 5 e23e324
Bladelets 16 ee 950 26 17 3 10 1022
Others e ee eeee 11
Total 95 1 6 950 35 23 3 23 1136
Tamar Hat. Zone III
Flakes 9 1 ee12e215
Blades eeeee1e12
Bladelets 2 ee 284 1 1 e4 292
Total 11 1 284 2 4 7 309
Tamar Hat. Zone II
Flakes 58 ee e e 210e19 89
Blades 6 1 e2e33e621
Bladelets 9 2 1 e629 7 6 e12 666
Total 73 3 1 2 629 12 19 37 776
Tamar Hat. Zone I
Flakes 64 2 2 1 e44e13 90
Blades 15 ee 3e22e123
Bladelets 27 ee e 559 8 8 5 8 615
Total 106 2 2 4 559 14 14 5 22 728
Rassel. Lower layer
Flakes 35 e1ee15 13 e61 125
Blades 9 ee 3e92e427
Bladelets 9 5 ee 830 19 5 5 23 896
Others eeee eeee 88
Total 53 5 1 3 830 43 20 5 96 1056
Columnata. South sector
Flakes 49 1 e10 1 e34 12 e45 152
Blades 3 e18ee 53e727
Bladelets 3 1 eee564 15 7 56 30 676
Others eeeeeeeee e
Total 55 2 1 18 1 564 54 22 56 82 855
Columnata. North sector
Flakes 107 e212 e69 13 e70 273
Blades 9 1 5 15 e15 3 e856
Bladelets 1 1 1 e860 33 13 30 37 976
Others eeee eeee 33
Total 117 2 8 27 860 117 29 30 118 1308
Es: end-scrapers; Pf: perforators; Br: burins; Bb: backed blades; Ct: composite tools;
Bbl: backed bladelets; Nd: notches and denticulates; Tr: truncated pieces; Gm:
geometric; Ms: Miscellaneous.
L. Sari / Quaternary International 320 (2014) 131e142138

4.3.1. Tamar Hat
In the lower occupations of Tamar Hat (zones VI and V), the
desired lamellar blanks are shorter and narrower. The metrical data
indicate a concentration at 20e25 mm length, 4e7 width and 2 mm
thickness and the length range corresponds exactly to the mini-
mum size of rejected prismatic and sub-pyramidal cores. Besides
these blanks, micro-bladelets (length: 8e20, width: 4e6, thickness:
1e2 mm) and burin spalls (length: 12e22, width: 3e5, thickness:
3e4 mm) produced from minority schemes are seldom chosen.
It is only from the occupation of zone IV and above that new
objectives appear leading to the production of robust and elongated
lamellar blanks with a slightly curved profile (concentration data at
26e35 length, 7e8 width and 2e3 mm thickness). Micro-bladelets
become scarcer and have a straight to slightly curved profile and
result from extreme exploitation of prismatic cores. These changes
are not accompanied by the appearance of a particular type of
backed bladelet; however, backed acute bladelets with retouched
base are more common than in lower occupations. This is may be
due to paring down the base for handling constraints. The backed
edge is often formed either by Ouchtata or semi abrupt retouch in
the lower occupations and microburin blow technique is well
attested.
4.3.2. Lower layer of Rassel cave
During the occupation of the lower layer at Rassel, blanks
selected for backed bladelets were short and narrow. The metrical
data are concentrated at 20e25 mm length, 5e8 width and 1e
2 mm thickness. They were extracted from a broad debitage surface
of reduced bladelet cores. The partially backed acute bladelets us-
ing the microburin blow technique are numerous and Ouchtata
retouch is well represented.
4.3.3. Columnata
The occupants of Columnata site looked for, without discrimi-
nation of raw material, large and thick blanks up to 8 mm wide and
greater than 3 mm thickness for backed bladelets. These robust
blanks were thick bladelets and flakes obtained by hard-hammer
percussion. When the thickness equals or exceeds 4 mm, the back
is achieved by on anvil retouch applied to one third of the backed
bladelet population. It could be related to constraints associated
with hafting and the function of projectile points. Arch-backed
bladelets obtained by the microburin blow technique are well rep-
resented. The scarcity of crossed abrupt retouch at Tamar Hat and
Rassel is probably related to the selection of blanks that were nar-
rower and thinner, thus requiring a slight reduction in width.
4.4. Backed bladelets: fragmentation and abandonment hypotheses
The large number of backed bladelet fragments reflects intense
activity in the site and it includes breakages resulting from post-
depositional alteration as well as accidents that arose during
backing or repairing. The latter are attested by the presence of cone
fractures (with a bulb of percussion) and Krukowski microburins.
Moreover, the damage patterns characteristic of hunting with
chipped stone projectile points are well attested in all three sites
which validates the hypothesis of the return from the hunt. This is
in agreement with the results of zooarchaeological analysis carried
out on Tamar Hat and Columnata faunal assemblages. The game
acquisition mode is similar to these sites in which whole carcasses
are brought to be butchered and eaten (Merzoug, 2005,2008;
Merzoug and Sari, 2008).
There are only a few pieces with bending fractures resulting
from an impact characterizing hunting projectile points and they
are equitably distributed among the different Iberomaurusian oc-
cupations (Table 7). A large number of them appear to be on the
ventral side and they are often more or less perpendicular to the
morphological axis of the blank, suggesting a mechanical constraint
parallel or sub-parallel to this axis. Facial step-terminated fracture
(Fig. 7) is more frequent than feather-terminated and lateral step-
terminated. This last is likely related to a violent axial impact and
is absent from the south sector (Fischer et al., 1984;Odell and
Cowan, 1986;Shea, 2006).
Table 5
Descriptive statistics for metrical data of backed bladelets.
Samples Length Width Thickness
N Mean SD SEM Mean SD SEM Mean SD SEM
Tamar Hat Zone VI 179 24.30168 7.063 0.52791 6.44693 2.21365 0.16546 2.41899 0.89811 0.06713
Tamar Hat Zone V 184 23.46739 6.32706 0.46644 5.86413 1.70759 0.12589 2.46196 0.73824 0.05442
Tamar Hat Zone IV 228 29.85526 7.83312 0.518761 7.19283 1.90626 0.12765 2.88158 0.87501 0.05795
Tamar Hat Zone III 63 28.66666 6.77757 0.853893 7.24194 1.41056 0.17914 2.87544 0.8549 0.051
Tamar Hat Zone II 62 28.67742 6.83235 0.867710 7.25397 1.50234 0.18928 2.87302 0.75117 0.09464
Tamar Hat Zone I 159 28.08805 7.30908 0.57965 7.16561 1.59264 0.12711 2.83019 1.01388 0.08041
Rassel. Lower layer 94 23.05319 4.91731 0.50718 6.31915 1.57407 0.16235 2.29787 0.88997 0.09179
Columnata. South sector 66 26.34848 6.58318 0.81033 8.0303 1.66382 0.2048 3.77273 1.03471 0.12736
Columnata. North sector 179 25.60335 5.97976 0.44695 8.3388 2.03134 0.15016 3.55866 1.19016 0.08896
Table 6
Test statistics for metrical data of backed bladelets.
Samples Length Width Thickness
tStatistic DF Probability Results tStatistic DF Probability Results tStatistic DF Probability Results
Tamar Hat Zone VI 2.53511 178 0.01210 Different 3.82623 178 1.79E-4 Different 6.42066 178 1.19E-9 Different
Tamar Hat Zone V 4.65788 183 6.13E-6 Different 9.65856 183 4.27E-18 Different 7.13007 183 2.25E-11 Different
Tamar Hat Zone IV 8.12563 227 2.85E-14 Different 0.88385 227 0.37774 Same 0.54494 227 0.58633 Same
Tamar Hat Zone III 3.54454 62 7.55E-4 Different 0.90395 62 0.36958 Same 0.49893 62 0.61822 Same
Tamar Hat Zone II 3.50049 61 8.74E-4 Different 0.91912 61 0.3616 Same 0.2432 61 0.80866 Same
Tamar Hat Zone I 4.22334 158 0.5407 Different 0.67349 158 0.50163 Same 0.24639 158 0.8057 Same
Rassel. Lower layer 5.10036 93 1.78E-06 Different 4.6864 93 9.48E-6 Different 6.54514 93 3.21E-09 Different
Columnata. South sector 0.87431 65 0.38517 Same 4.6401 65 1.73E-5 Different 7.24479 65 6.39E-10 Different
Columnata. North sector 0.082 178 0.93474 Same 8.38299 178 1.37E-14 Different 7.96633 178 1.87E-13 Different
L. Sari / Quaternary International 320 (2014) 131e142 139

The uniformity of width/thickness values of backed bladelet
fragments at Tamar Hat and Rassel is very marked, compared to
Columnata where robust pieces were preferred (Fig. 8). This is likely
due to the constraints of hafting. The existence of proximal and
distal fragments involved elements mounted axially at the tip of the
foreshaft. The proximal fragments can be introduced intothe site as
points still in their shafts, whereas, the distal fragments are con-
tained in the game brought back to this site. However, the presence
of mesial fragments suggests that pieces were mounted in a row.
Moreover, the massive blanks of Columnata were probably inten-
ded to be mounted axially at the tip of foreshaft. Unfortunately, no
example of organic components of hunting implements has been
found in all three sites, nor in other Iberomaurusian sites.
5. Discussion
The choice of Iberomaurusian settlements is closely related to
proximity to raw material outcrops, especially in the case of coastal
sites on marine and alluvial terraces, and is consistent with the
literature data (e.g. Tixier, 1954;Roche, 1963;Hachi, 1987). From a
diachronic point of view, maintenance of the same raw material
procurement strategy might reflect territorial populations and the
seasonality attested at Tamar Hat and Columnata might be seen as
periods of semi nomadic life, since the occupants returned peri-
odically to the same places.
The use of mineral soft-hammer percussion to extract lamellar
end-products is well attested in the studied lithic assemblages,
despite the variability observed in the technical procedure
observed at Columnata and which may refer to different know-how
of Iberomaurusian populations. Nevertheless, the persistence of the
use of this technique reveals a common technical tradition pre-
served by the Iberomaurusian populations. In fact, this technique is
widespread among Late Glacial populations of Europe who favored
it at the expense of soft organic percussion (Pelegrin, 2000;Sari,
2012).
The study of debitage methods gives an idea on the progress of
operating schemes. Until then, the previous studies reported that
for Iberomaurusian populations, debitage was simple and often
performed on small pebbles (e.g. Brahimi, 1969b,1970;Close,
1980e81;Hachi, 1987;Hellal, 2005). However, in this paper, the
technological analysis demonstrates that at Tamar Hat there is a
change in core reduction strategy beginning at the end of the
occupation of zone V. It is interesting to note that this change does
not continue past the last numerous snail deposit (layer 44 of zone
IV) dated to 18,750 500 BP, and comes just before the so called
“faunal break”previously reported by Saxon et al. (1974) who
associated it to a change in stratigraphy at the end of the Last
Glacial Maximum. Would it have any relation to the change in the
core reduction strategies identified from zone IV? This may reflect a
new adaptation of Iberomaurusian populations at the end of this
last cold cycle. Unfortunately, no change in subsistence behaviour
was previously reported for this site (Merzoug, 2005) and paleo-
environmental data including those linked to micro-fauna are
deficient, which justifies the urgency to undertake fieldwork to
reassess stratigraphic and paleoenvironmental data, as well as
research into other occupations synchronous with this ancient
phase of Iberomaurusian culture.
Most of the operating schemes observed in the upper occupa-
tions of Tamar Hat (zones IV to I) persist in the lower layer of Rassel.
This could refer to the possible existence of similar ecological
conditions. However, the absence at Rassel of the dihedral core
scheme that was adapted perfectly to flint pebbles and existed in
the upper occupations of Tamar Hat, suggests that there may be
regional particularities in the Iberomaurusian technical traditions
beyond chronology. This is even more likely given that the dihedral
core method is also reported in the upper layers of Afalou which are
dated between 13,000 and 11,000 BP and which contain numerous
clay figurines (Hachi, 1987,1996,2006). It is barely 1 km from
Tamar Hat rockshelter and belongs to the same paleogeographic
unit.
Another argument corroborating the regional particularities
assumption is the fact that Columnata occupants adopted a
completely different and innovative production strategy. Indeed,
they produced different blanks from differentiated chaînes
Table 7
Occurrence of backed bladelets with impact fragmentation.
Tamar Hat Rassel Columnata
Fragmentation degree Zone VI Zone V Zone IV Zone III Zone II Zone I Lower South North
Little fragmented 5 5 ee34212
Proximal 6 7 7 7 3 7 2 2 6
Mesial 1 10 3 4 4 4 1 4 7
Distal 2 6 10 2 3 6 1 2 7
Total 14 28 20 13 13 21 6 9 22
Fig. 7. Fragments with diagnostic impact damage. a, b. Facial step-terminating bending fractures; c, d. Lateral step-terminating bending fractures; e. Feather-terminating bending
fracture. Provenance: a. Tamar Hat; b, c, e. Lower level of Rassel; d. Columnata.
L. Sari / Quaternary International 320 (2014) 131e142140

opératoires with a selection of raw materials, while those at Tamar
Hat and Rassel produced different blanks with the same chaîne
opératoire, regardless of raw materials. Moreover, at Columnata the
calibration of the blanks which were transformed to backed bla-
delets included bladelets as well as flakes and often used an abrupt
to crossed retouch.
The presence of stone hunting implements in all three sites is an
argument for hunting activity as well as the return of hunters to the
site. This is supported by the results of the zooarchaeological ana-
lyses (Merzoug, 2005,2008;Merzoug and Sari, 2008). Moreover,
the coexistence of different types of hafting is assumed. Do these
supposed types of hafting coexist within the same technical tra-
ditions? Or should we rather consider the question in terms of
variation relating to a suitable mounting adapted to the size of the
game? This is a fundamental difference that can address the drivers
of change in terms of hafting. Experimentation is required to vali-
date this hypothesis. However, one significant factor may be the
ecological conditions in which Columnata populations lived. This
was an open steppe landscape that required a suitable armament
different from the forest and mountain environment in which
Tamar Hat and Rassel occupants were living.
6. Conclusion
The technological analysis has provided strong arguments for a
different know-how in technical behaviours between populations
living at coastal shelter sites and those at inland open-air sites.
Indeed, the identical organization of technical projects for blank
production observed in the Tamar Hat and Rassel lithic assemblages
gets lost over time in Columnata. In this open air-site, the search for
robust blanks adapted to a new type of weapon due to a renewal of
both landscape and hunted species is evident. Diachronically, the
different geographical areas seem to cover variable economic en-
tities which would suggest rather a new adaptation of the same
populations to different ecological niches. This is in agreement with
the claims of previous authors that the Iberomaurusian is a set of
cultural facies with regional particularities. Nevertheless, the
continuous use of microburin blow technique at all three sites
shows diachronic continuity in the method of producing backed
bladelets and is a strong argument for cultural continuity.
This study gives a new way to understand variability within
Iberomaurusian assemblages. It is important to extend this study to
other Iberomaurusian sites of the Maghreb. Particular attention will
be given to look for technical features of diachronic and/or regional
patterns such as endscraper-core and dihedral core debitage, so as
to have a better understanding of their chronological position and
territorial expansion. Finally, it is hoped to establish experimental
tests using Iberomaurusian replicas of projectile points in order to
support the results observed and to confirm initial hypotheses.
Acknowledgments
I would express my thanks to the organizer of the workshop, Pr.
David Lubell for the opportunity to present this research. I am also
very grateful to him for revising and improving the previous En-
glish text and for his useful comments. I also address my thanks to
Linda Touchi-Benmansour who corrected my English translation of
this paper. I am solely responsible for the translation of French
quotations. I gratefully acknowledge one anonymous reviewer for
his helpful comments on previous versions of this manuscript.
References
Arambourg, C., Boule, M., Vallois, H., Verneau, R., 1934. Les grottes paléolithiques
des Béni Seghoual (Algérie). In: Archives de l’Institut de Paléontologie Humaine,
Mémoire 13. Masson, Paris.
Aubry, T., Detrain, L., Kervazo, B., 1995. Les niveaux intermédiaires entre le Grave-
ttien et le Solutréen de l’abri Casserole (Les Eyzies-de-Tayac): mise en évidence
d’un mode de production original de microlithes. Bulletin de la Société Pré-
historique Française 92 (3), 296e301.
Aumassip, G., 1979. Le bas Sahara dans la Préhistoire. Thèse de Doctorat ès sciences,
Aix-en-Provence.
Barton, R.N.E., Bouzouggar, A., Ramsey, C.-B., Collcutt, S.-N., Higham, T.-F.-G.,
Humphrey, L.-T., Parfitt, S.-A., Rhodes, E.-J., Schwenninger, J.-L., Stringer, C.-B.,
Turner, E., Ward, S.-C., 2007. Abrupt climatic change and chronology of the
Upper Palaeolithic in northern and eastern Morocco. In: Bar Yosef, O.,
Mellars, P.-A., Stringer, C.-B. (Eds.), Rethinking the Human Revolution: New
Behavioural and Biological Perspectives on the Origins and Dispersal of
Modern Humans. McDonald Institute for Archaeological Research, Cambridge,
pp. 177e186 .
Barton, R.-N.-E., Bouzouggar, A., Humphrey, L.-T., Berridge, P., Collcutt, S.-N.,
Gale, R., 2008. Human bu rial evidence from Hattab II Cave (Oued Laou-Tétuoan,
Morocco) and the question of continuity in Late Pleistocene-Holocene mor-
tuary practices in Northwest Africa. Cambridge Archaeological Journal 18 (2),
195e214.
Bar-Yosef, O., 1991. The search for lithic variability among Levantine Epi-Paleolithic
industries. In: 25 ans d’études technologiques en préhistoire. Actes des Ren-
contres 18e20 octobre 1990, XI
e
Rencontres Internationales d’Archéologie et
d’Histoire d’Antibes. APDCA, Juan-les-Pins, pp. 319e335.
Belfer-Cohen, A., 2003. Why microliths? Microlithization in the Levant. In:
Elston, R.G., Kuhn, S.L. (Eds.), Thinking Small: Global Perspectives on Micro-
lithization, pp. 57e68. Archaeological Papers of the American Anthropological
Association, Arlington 12.
Fig. 8. Scatter plot for width/thickness values of backed bladelets.
L. Sari / Quaternary International 320 (2014) 131e142 141

Bouzouggar, A., Barton, R.-N.-E., Blockley, S., Bronk-Ramsey, C., Collcutt, S.-N.,
Gale, R., Higham, T.-F.-G., Humphrey, L.-T., Parfitt, S., Turner, E., Ward, S., 2008.
Reevaluating the age of the Iberomaurusian in Morocco. African Archaeological
Review 25 (1e2), 3e19.
Brahimi, C., 1969a. Corrélations entre quelques groupes d’outils des cultures ibér-
omaururisiennes. Libyca 17, 93e99.
Brahimi, C., 1969b. L’industrie lithique de l’abri de Tamar Hat (Kabylie). Libyca 17,
35e48.
Brahimi, C., 1970. L’Ibéromaurusien du littoral de la région d’Alger. Mémoire du
C.R.A.P.E. 23, Alger.
Brahimi, C., 1972. Deux campagnes de fouilles à Columnata (1969 et 1971). Libyca
20, 49e101.
Bressy, C., 2003. Caractérisation et gestion du silex des sites mésolithiques et
néolithiques du Nord-Ouest de l’arc alpin: une approche pétrographique et
géochimique. In: British Archaeological Reports, International Series 1114. Ox-
ford Archaeopress.
Cadenat, P., 1948. La station préhistorique de Columnata. Bulletin de la Société
Géographique et d’Archéologie d’Oran 70, 3e66.
Cadenat, P.,1966. Vues nouvelles sur les industries de Columnata. Libyca 14, 189e207.
Camps, G., 1974. Les civilisations préhistoriques de l’Afrique du Nord et du Sahara.
Doin, Paris.
Camps, G., Curtes, A., Tixier, J., 1959. Gisement Epipaléolithique (vraisemblablement
Ibéromaurusien) au Cap Chanoua en avant d’une grotte à remplissage bré-
choïde quaternaire. Bulletin de la Société Préhistorique française 56, 297. Prise
de date.
Close, A.-E., 1977. Identification of Style in Lithic Artefacts from North East Africa. In:
Mémoires de l’Institut d’Égypte 61. S.O.P. Press, Cairo.
Close, A.-E., 1978. The identification of style in lithic artefacts. World Archaeology 10
(2), 223e237.
Close, A.-E., 1980e81. The Iberomaurusian sequence at Tamar Hat. Libyca 28e29,
69e104.
Demars, P.Y., 1982. L’utilisation du silex au Paléolithique supérieur: choix, appro-
visionnement, circulation. L’exemple du Bassin de Brive. In: Cahier du Qua-
ternaire 5. CNRS, Paris.
Féblot-Augustin, J., 1997. La circulation des matières premières au Paléolithique.
ERAUL 75, Liège.
Féblot-Augustin, J., 2008. Paleolithic raw material provenance studies. In:
Pearsall, D.M. (Ed.). Encyclopedia of Archaeology, vol. 2. Academic Press, New
York, pp. 1187e1198.
Fischer, A., Vemming-Hansen, P., Rasmussen, P., 1984. Macro and microwear traces
on lithic projectile points. Journal of Danish Archaeology 3, 19e46.
Geneste, J.-M., Plisson, H., 1990. Technologie fonctionnelle des pointes à cran sol-
utréennes: l’apport des nouvelles données de la grotte de Combe Saunière
(Dordogne). In: Kozlowski, J.-K. (Ed.), Feuilles de pierre: les industries à pointes
foliacées du Paléolithique supérieur européen. Actes du Colloque de Cracovie,
vol. 42. Études et Recherches Archéologiques de l’Université de Liège, Liège,
pp. 293e332.
Hachi, S., 1987. Les cultures de l’Homme de Mechta El Arbi. Le gisement d’Afalou
Bou Rhummel (massif des Babors, Algérie): les niveaux supérieurs e130 0 0e
11000 BP. Thèse de Doctorat 3
ème
cycle. Université d’Aix-en-Provence.
Hachi, S., 1996. L’Ibéromaurusien, découverte des fouilles d’Afalou. L’Anthropologie
100 (1), 77e87.
Hachi, S., 2003. Les cultures de l’Homme de Mechta-Afalou. Le gisement d’Afalou-
Bou- Rhummel (massif des Babors, Algérie), les niveaux supérieurs 13000e
11000 BP. Alger. In: Mémoire du Centre National de Recherche Préhistoriques
Anthropologiques et Historiques, Nouvelle série 2.
Hachi, S., 2006. Du comportement symbolique des derniers chasseurs Mechta-
Afalou d’Afrique du Nord. Comptes Rendus Palevol 5 (1/2), 439e440.
Hachi, S., Fröhlich, F., Gendron-Badou, A., de Lumley, H., Roubet, C., Abdessadok, S.,
2002. Figurines du Paléolithique supérieur en matière minérale plastique cuite
d’Afalou Bou Rhummel (Babors, Algérie): premières analyses par spectroscopie
d’absorption infra-rouge. L’Anthropologie 106 (1), 57e97.
Hellal, H., 2005. L’Ibéromaurusien de l’abri Alain (Algérie occidentale). Approche
techno-typologique des industries lithiques (collection de l’IPH). Mémoire de
DEA, Paris I.
Inizan, M.-L., Reduron, M., Roche, H., Tixier, J.,1995. Technology and Terminology of
Knapped Stone, vol. 5. CREP, Paris. http://www.arkeotek.org/ebooks/
TerminologyKnappedStone.pdf.
Le Brun-Ricalens, F., Brou, L., 2003. Burins carénés-nucléus à lamelles: identification
d’une chaîne opératoire particulière à Thèmes (Yonne) et implications. Bulletin
de la Société Préhistorique Française 100 (1), 67e83.
Lubell, D., 1984. Paleoenvironment and Epi-Paleolithic economies in the Maghreb
(ca. 20,000 to 5000 BP). In: Clark, J.-D., Brandt, S.-A. (Eds.), From Hunters to
Farmers: the Causes and Consequences of Food Production in Africa. University
of California Press, Berkeley, pp. 41e56.
Lucas, G., 1999. Production expérimentale de lamelles torses: approche prélimin-
aire. Bulletin de la Société Préhistorique Française 96 (2), 145e151.
Merzoug, S., 2005. Comportements de subsistance des Ibéromaurusiens d’après
l’analyse archéozoologique des mammifères des sites de Tamar Hat, Taza 1 et
Columnata (Algérie). Thèse de Doctorat. Muséum National d’Histoire Naturelle
de Paris.
Merzoug, S., 2008. Etude archéozoologique des mammifères ibéromaurusiens de la
collection Cadenat du site de Columnata (Tiaret, Algérie). In: Aouraghe, H.,
Haddoumi, H., El Hammouti, K. (Eds.), Le Quaternaire Marocain dans son
Contexte Méditerranéen. Actes de la Quatrième Rencontre des Quaternaristes
Marocains, 2007. Publications de la Faculté des Sciences d’Oujda, pp. 255e264.
Merzoug, S., Sari, L., 2008. Re-examination of the Zone I material from Tamar Hat
(Algeria): zooarchaeological and technofunctional analyses. African Archaeo-
logical Review 25 (1e2), 57e73.
Moser, J., 2003. La Grotte d’Ifri n’Ammar. Tome 1. L’Ibéromaurusien Band 8. Allge-
meine und Vergleichende Archäologie-Forschungen, Köln.
Nami, M., 2007. Les techno-complexes Ibéromaurusiens d’Ifri El Baroud (Rif
Oriental, Maroc). Zeitschrift für Archäologie Außereuropäischer Kulturen 2,
183 e239.
Odell, G., Cowan, F., 1986. Experiments with spears and arrows on animal targets.
Journal of Field Archaeology 13, 195e212.
Pelegrin, J., 1995. Technologie lithique: le châtelperronien de Roc-de-Combe (Lot) et
de la Côte (Dordogne). In: Cahiers du Quaternaire, 20. C.N.R.S., Paris.
Pelegrin, J., 2000. Les techniques de débitage laminaire au Tardiglaciaire: critères
de diagnose et quelques réflexions. In: Valentin, B., Bodu, P., Christensen, M.
(Eds.), L’Europe centrale et septentrionale au Tardiglaciaire. Table-ronde de
Nemours (13e16 mai 1997), Mémoires du Musée de Préhistoire d’Ile-de-
France 7, pp. 73e86.
Pelegrin, J., O’Farell, M., 2005. Les lamelles retouchées ou utilisées de Castanet. In:
Le Brun-Ricalens, F. (Ed.), Productions lamellaires attribuées à l’Aurignacien,
chaînes opératoires et perspectives, technoculturelles. Actes du XIV
ème
Congrès
de l’UISPP. Collection Internationale d’Archéologie Préhistorique Eurasienne,
Luxembourg, pp. 103e121.
Perlès, C., 1980. Économie de la matière première et économie du débitage: deux
exemples grecs. In: Tixier, J. (Ed.), Préhistoire et technologie lithique. Journées
du 11e13 mai 1979. URA 28, Valbonne, pp. 37e41. Cahier 1 du CNRS.
Perlès, C., 1991. Économie des matières premières et économie de débitage: deux
conceptions opposées. In: 25 ans d’études technologiques en Préhistoire: bilan
et perspectives. Actes des XI
ème
Rencontres Internationales d’Archéologie et
d’Histoire, Antibes. APDCA, Juan-les-pins, pp. 35e45.
Rahmouni, O., Roussillot, C., Armanet, F., 1972a. Algiers radiocarbon measurements
II. Radiocabon 12 (2), 353e357.
Rahmouni, O., Roussillot, C., Armanet, F., 1972b. Algiers radiocarbon measurements
II. Radiocabon 14 (1), 1e5.
Roche, J., 1963. L’Epipaléolithique marocain. Livraria Bertrand, Lisboa.
Sari, L., 2008. Production lamellaire en contexte ibéromaurusien: variabilité et
implication technoéconomique. Cas du niveau inférieur de Rassel (littoral Ouest
algérois). In: Aouraghe, H., Haddoumi, H., El Hammouti, K. (Eds.), Le Quaternaire
Marocain dans son Contexte Méditerranéen. Actes de la quatrième rencontre
des quaternaristes marocains, 2007. publications de la faculté des sciences
d’Oujda, pp. 376e388.
Sari, L., 2011. Variabilité et modalités de production des supports lamellaires: cas de
l’Ibéromaurusien de la zone I de l’abri de Tamar Hat (Algérie nord-orientale). In:
Actes du Colloque de Tamanrasset, 2007. Editions du CNRPAH, pp. 267e281.
Sari, L., 2012. L’Ibéromaurusien. Culture du Paléolithique supérieur tardif. Approche
technologique des productions lithiques taillées de Tamar Hat, Rassel et Col-
umnata (Algérie). Thèse de Doctorat (non publiée). Université de Paris X-
Nanterre.
Sari, L., 2013. Mise en évidence de l’usage de la percussion tendre minérale durant
l’Ibéromaurusien de l’Algérie (in preparation).
Saxon, E.-C., Close, A.-E., Cluzel, C., Morse, V., Shackleton, N.-J., 1974. Results of
recent investigations at Tamar Hat. Libyca 22, 49e91.
Shea, J.-J., 2006. The origins of lithic projectile point technology: evidence from
Africa, the Levant, and Europe. Journal of Archaeological Science 33, 823e846.
Taylor, V.-K., Barton, R.-N.-E., Bell, M., Bouzouggar, A., Collcutt, S., Black, S., Hogue, J.-
T., 2011. The Epipalaeolithic (Iberomaurusian) at Grotte des Pigeons (Taforalt),
Morocco: a preliminary study of the land Mollusca. Quaternary International
30, 1e10.
Tixier, J., 1954. Le gisement préhistorique d’El-Hamel. Libyca 2, 78e120.
Tixier, J., 1963. Typologie de l’Épipaléolithique du Maghreb. In: Mémoire du
C.R.A.P.E. 2. Arts et Métiers Graphiques, Paris.
Tixier, J., 1978. Notice sur les travaux scientifiques. Thèse de doctorat d’Etat. Uni-
versité de Paris X, Nanterre.
Tixier, J., Inizan, M.-L., Roche, H.,1980. Préhistoire de la pierre taillée 1: terminologie
et technologie. C.R.E.P., Paris.
Tixier, J., Inizan, M.-L., 1981. Ksar ‘Aqil, stratigraphie et ensembles lithiques dans le
Paléolithique Supérieur: fouilles 1971e1975. In: Préhistoire du Levant, Chro-
nologie et Organisation de l’Espace depuis les Origines jusqu’au V
ème
Millénaire.
Maison de l’Orient, CNRS, Lyon, pp. 353e367.
L. Sari / Quaternary International 320 (2014) 131e142142