Review of chronological data from the Rabat-Témara caves (Morocco)

Implications for understanding human occupation in north-west Africa during the Late Pleistocene

Synthèse des données chronologiques des grottes de Rabat-Témara (Maroc) : implications pour la compréhension des occupations humaines en Afrique du Nord-Ouest au cours du Pléistocène récent

Eslem Ben Arous, Christophe Falguères, Roland Nespoulet et Mohamed Abdeljalil El Hajraoui

p. 177-201

Résumés

During the Late Pleistocene, a major technical and cultural shift occurred in North Africa, corresponding to the transition from the Middle Stone Age (MSA) to the Later Stone Age. So far, the cultural significance of this key period has been poorly understood. Its duration is still debated due to the lack of reliable dating evidence and well-stratified records, particularly across the period 60,000-20,000 yrs (MIS 4-2). Hence, models of modern human migrations in the Late Pleistocene based on genetic data have proven difficult to compare spatially and temporally with regional archaeological data.
This work focuses on the timing of human occupation during the Late Pleistocene in north-west Africa, using the example of the Témara region that is characterised by an important concentration of coastal archaeological caves. The main purposes of this paper are
(a) to review the current dating framework in Rabat-Témara while integrating the new dating, thus making a contribution to refining the chronological resolution for the final MSA, (b) to assess the relevance of this current chronological data and to discuss the possibility of human occupation continuity and (c) to compare this with other data from different sites during the end of the Late Pleistocene and to discuss implications for environmental and human conditions.

Au cours du Pléistocène récent, un changement technique et culturel majeur s'est produit en Afrique du Nord, correspondant au passage du Middle Stone Age (MSA) au Later Stone Age. Jusqu'à présent, l'importance culturelle de cette période-clé a été difficilement appréhendée. Sa durée fait encore l’objet de débats en raison de l’absence de données chrono-stratigraphiques fiables, particulièrement sur la période de 60 000-20 000 ans (MIS 4-2). De ce fait, les modèles de migrations humaines des hommes modernes au Pléistocène supérieur fondés sur des données génétiques se sont révélés difficiles à comparer spatialement et temporellement avec des données archéologiques régionales.
Ce travail porte sur la chronologie des occupations humaines pendant le Pléistocène supérieur en Afrique du Nord-Ouest, en prenant l'exemple de la région de Témara, caractérisée par une concentration importante de grottes archéologiques côtières. Il vise à :
(a) réviser le cadre chronologique à Rabat-Témara, en intégrant de nouvelles datations, et ainsi contribuer à affiner la résolution chronologique de la fin du MSA,
(b) évaluer la pertinence de ces nouvelles données chronologiques et ainsi examiner la possibilité d’une continuité des occupations humaines,
(c) confronter ces nouvelles données à celles de différents sites de la fin du Pléistocène supérieur et examiner les implications sur les conditions environnementales des occupations humaines de la région à cette période.

Entrées d’index

Mots-clés : Afrique du Nord-Ouest, Middle Stone Age, Later Stone Age, méthodes de datation, transition

Keywords : north-west Africa, Middle Stone Age, Later Stone Age, dating methods, transition


Texte intégral

Introduction

1During the last decades, research conducted in north-west Africa has emphasised the role of the region in the improvement of our understanding of human evolution, especially for the origin of H. sapiens and the emergence of associated modern behaviours (McBrearty & Brooks 2000; d’Errico & Stringer 2011; Scerri 2017; Hublin et al. 2017). From 60,000 to 20,000 yrs (MIS 4‑3), North Africa is marked by climate fluctuations that had an impact on the environmental context (Blome et al. 2012; Drake et al. 2013). During this time, a cultural shift occurred from the Middle Stone Age (MSA) to the Later Stone Age (LSA), attested by a change in cultural and symbolic markers (e.g. disappearance of the use of Nassarius shells (Barton & D’Errico 2012), use of cavities for sepulchral purposes (Ferembach 1976b; Aoudia-Chouakri 2013)), and in lithic assemblages (e.g. use of backed bladelet technologies (Bouzouggar et al. 2008)). Recent genetic studies (Kefi et al. 2016; Loosdrecht et al. 2018) from the LSA Moroccan sites show that many human migrations took place during this period, suggesting that several dispersals of modern humans occurred within and out of Africa. Unfortunately, this evidence for migration is more difficult to assess for the MSA due to the lack of genetic data. Considering the large size of the territory covered by the MSA and its significant regional lithic variability (Van Peer et al. 2003; Garcea & Giraudi 2006; Dibble et al. 2012; Douka et al. 2014), its disappearance in North Africa is hotly debated, and this issue remains complicated. Although the archaeological data indicate a cultural disconnect between the MSA and the LSA, chronological data raise the question of a potential continuity between the MSA and the LSA. However, is essential to establish an accurate temporal framework for occupation as it relates to climate and environmental changes.

2Many reviews discuss this chronological framework and point out the problems of its resolution (Debénath 2000; Linstädter et al. 2012; Doerschner et al. 2016). Starting in the 2000s, new excavations that took place mainly in the Moroccan Rif and on the Moroccan Atlantic coast unearthed new contextualised and usable material (El Hajraoui et al. 2012; Linstädter et al. 2012). New dating applications using OSL (Optically Stimulated Luminescence), TL (thermoluminescence) and combined US‑ESR (uranium series and electron spin resonance) have greatly improved the chronology of the MSA, particularly with regard to the emergence of modern behaviours and resolving the Mousterian/Aterian contemporaneity issues. These new dating applications have attempted, successfully, to date the MSA, for instance, by combining US‑ESR, TL and OSL, and have pushed back in time the first occurrence of the MSA (Richter et al. 2010; Jacobs et al. 2012; Richter et al. 2017), whereas new radiocarbon dates have provided an older estimate for the LSA (Barton et al. 2015; Hogue & Barton 2016). The transition from the MSA to the LSA is characterised by a cultural discontinuity. This discontinuity, apart from the technological ‘rupture’ of industries, must be questioned because, for example, an occupation gap has been hypothesised between the Aterian (MSA) and the Iberaromaurusian (LSA) in the Témara region (Debénath et al. 1986), whereas sedimentary sequences with lithic industries are present in mountainous and contemporaneous areas of this hiatus, as in Taforalt (Barton et al. 2015). In the Sahara area, other authors have highlighted sedimentary gaps that are reflected by a lack of archaeological documentation during the transition phase, highlighting a possible lack of sedimentary continuity (Cremaschi et al. 1998).

3Current data in North Africa place the MSA disappearance between 70,000 and 25,000 yrs. It is difficult to reduce this time range due to (a) a low number of MIS 4‑2 sites recording this period across North Africa and (b) difficulties in using dating techniques in certain contexts (e.g. in caves, bioturbations/human activities causing sediments to mix) (Barton et al. 2009; Schwenninger et al. 2010; Jacobs et al. 2012; Janati-Idrissi et al. 2012; Dibble et al. 2013; Ben Arous et al. 2019). In this paper, we will focus on the second aspect.

4Most of the MIS 4‑2 data in North Africa come from coastal or inland Moroccan sites (figure 8.1) that provide well-documented data on human occupations/cultures over a long period of time. This is the case for some caves in the Rabat-Témara region in Morocco (figure 8.1) that contain a sub-continuous sedimentary record covering the last 120 millennia (Schwenninger et al. 2010; Jacobs et al. 2012). Because of its strategic location near marine resources, this coastal region may have played a role in the cognitive development of H. sapiens and their dispersal during the Late Pleistocene (Campmas et al. 2018). Rabat-Témara archaeological sequences provided by MSA and LSA occupations offer the opportunity to study the transition from the MSA to the LSA at one regional scale. For example, El Harhoura 2 cave is one of the rare sequences to present both periods in a good state of conservation without taphonomic disturbances of archaeological remains (Stoetzel et al. 2011, 2012).

5The aims of this paper are (a) to overview the current dating framework in Rabat-Témara, and (b) to assess the relevance of this current chronological data compared with other data from different sites in order to discuss a possible human occupation continuity during the end of Late Pleistocene.

figure 8.1

Image

Spatial distribution of the sites with dates from MIS 4 to MIS 2 in north-western Africa. Right: Rabat-Témara region (map modified from Nespoulet et al. 2012). Mentioned in the text, Doukala 1 and 2 have not been represented on this map because no dates are available for these two sites. Left: MSA and LSA sites in north-western Africa and the dated sites (modified from Aumassip 2004).

Available stratigraphies in the Témara region: an overview

6The Rabat-Témara region (figure 8.1) is located c. 4 km to the south of Rabat city and delimited by Wadi Bou Regreg in the north-west and Wadi Ykem in the south-west. Several coastal caves have been known since the 1950s for well-preserved human fossils (Debénath 2000; Nespoulet et al. 2008b). Some of them have been explored by successive excavations conducted by international teams between 2001 and 2015 (Nespoulet et al. 2008b; Barton et al. 2009; Dibble et al. 2012; El Hajraoui et al. 2012). Recent work in this region (Chahid et al. 2016, 2017) has clarified the morphology of the coastline: there is a succession of 12 Plio-Quaternary sand dunes parallel to the current coastline. All the archaeological caves, which are located in the second sand dune, were formed before MIS 5 (Barton et al. 2009; Jacobs et al. 2011), probably by karstic and/or marine erosion of the calcarenites (Chahid 2017). Six archaeological caves (Contrebandiers, El Mnasra, El Harhoura 1 and 2, Dar es Soltan 1 and 2), one open-air site (Chaperon-Rouge) and two caves in quarries (Doukala 1 and 2) are known. The sandy sedimentary sequences are similar for these caves: they range from c. 4 m to sometimes more than 7 m in thickness and cover the Upper Pleistocene (c. 125,000 yrs) to the Holocene (c. 5000 yrs BP). Five caves (Dar es Soltan 2, El Harhoura 1 and 2, El Mnasra and Contrebandiers) have delivered H. sapiens (Hublin 2001) remains associated with Aterian industries (MSA) (Debénath et al. 1986; Debénath 2000) (figure 8.2, table 8.1). Except for Dar es Soltan 1 and 2, additional LSA and Neolithic human remains have been found in sepulchral contexts in the other caves. Meanwhile, dating of the fossils discovered during previous excavations often remains unfinished or controversial (table 8.1). Archaeological discoveries from these sites have made this region a landmark for documenting prehistoric human occupations from the MSA to the Neolithic, in particular for the dispersal and cultural evolution of H. sapiens.

figure 8.2

Image

Stratigraphical log of the Témara caves, with cultural attribution per layer and associated human remains. Cultural correlation is also represented (MSA, LSA and Neolithic). The reference of the archaeostratigraphy is given below each sequence. For Contrebandiers cave, the three excavated sectors have been represented separately, but only the CEA and sector V are considered in the discussion in the section entitled ‘Chronological comparison’ in the text. The layers considered as sterile are in light grey. For correlations between the different stratigraphy available for each site, refer to tables 8.2 to 8.7.

table 8.1

SiteDescriptionAge (years)MethodsCulture/PeriodReferences
ContrebandiersMandibleUndated but associated with Ouljian breccias (MIS 5)GeologyMousterian (MSA)Vallois & Roche 1958
Occipital and frontal fragmentsUndatedMousterian (MSA)Ferembach 1976
Juvenile skull and partial skeletonc. 110,000OSLMousterian (MSA)Balter 2013; Dibble et al. 2012
Dar es Soltan 2Partial adult skull and calvaria; hemi-mandible; immature calvaria; mandible; isolated teeth associated1. c. 35,00014CAterian (MSA)1. Debénath 1980
2. c. 120,000OSLAterian (MSA)2. Barton et al. 2009
El Harhoura 1Isolated canine and mandiblec. 40,000‑30,000TLAterian (MSA)Debénath et al. 1986
El Harhoura 2Cervical vertebrae; metatarsus; right metacarpal; right medial wedge1. c. 55,000‑75,000OSLAterian (MSA)El Hajraoui et al. 2012; Jacobs et al. 2012
3. c. 40,000‑50,000Combined US‑ESRAterian (MSA)Janati-Idrissi et al. 2012; Ben Arous et al. 2019
Lumbar vertebraec. 110,000OSLAterian (MSA)Jacobs et al. 2012
El MnasraPhalanx; skull fragment; 4 teeth, one of which is incisor1. c. 70,000‑80,000Combined US‑ESRAterian (MSA)El Hajraoui et al. 2012; Janati-Idrissi et al. 2012
2. c. 110,000OSLAterian (MSA)El Hajraoui et al. 2012; Jacobs et al. 2012

Human remains from Rabat-Témara’s sites with their cultural and chronological attribution (modified from Scerri et al. 2017).

7In order to contextualise the dates and to discuss their relevance, a synthetic view of the stratigraphy of each cave is presented in the next section.

Témara caves: Stratigraphy and dating

Dar es Soltan 1

8This cave was excavated in both 1937 and 1938 by A. Ruhlmann, and 12 layers (A to M from top to bottom) for a total thickness of 7.5 m were initially described (Ruhlmann 1951). The sequence lies on bedrock and is capped by Neolithic layers (shell midden) at the top. A total of four archaeological layers, separated by sterile layers, have been defined on the basis of the industries (from the oldest to the youngest layer): layers I, C2 and C1 are attributed to the Aterian (MSA) and layer B to the Iberomaurusian (LSA). Fallen blocks from the cave are located between layers C and B. J. Roche emphasised the difficulty of characterising the C2 layer because of heterogeneity and post-depositional disturbances (Roche 1956).

9In 2005, a new geomorphological and lithostratigraphical study has been refined into five main stratigraphical units (Group 1 to Group 5), based on sedimentary differentiations and erosion surfaces (Barton et al. 2009; Schwenninger et al. 2010; Bouzouggar et al. 2018). This work allowed for the identification of more archaeological units in cross-section than in Ruhlmann’s first study and for the application of the OSL dating method (Schwenninger et al. 2010; Bouzouggar et al. 2018) (table 8.2). The Early Aterian (MSA) in this cave is dated at between 112,000 and 50,000 yrs.

10The LSA was initially attributed to layer B (Ruhlmann 1951). Recent lithostratigraphical studies have shown that the Group 5 unit associated with layer B is disturbed by burrows, with rare LSA, Neolithic and protohistoric artefacts (Bouzouggar et al. 2018). The dates for this layer range from 33,000 yrs to c. 8000‑7000 yrs and should be considered with great care given the disturbed sedimentary context.

table 8.2

Stratigraphy (Ruhlmann 1951)Lithostratigraphy (Barton et al. 2009)Sample codeAges (ka)±MethodMaterialArchaeologyPeriod
/Culture
References
AG5NeolithicBarton et al. 2009
BG5.2OSL 157.60.5OSL (MG-SAR)QPALSA/IBMBarton et al. 2009
OSL 166.80.4OSL (MG-SAR)Q
G5.1OSL 14332.3OSL (MG-SAR)QPABarton et al. 2009
CC1G4.8PAMSA/ATBarton et al. 2009
G4.7OSL 1352.83.2OSL (MG-SAR)Q
G4.6
G4.5PAMSA/ATBarton et al. 2009
C2G4.4Barton et al. 2009
G4.3Barton et al. 2009
G4.2Barton et al. 2009
G4.1OSL 1261.74.4OSL (MG-SAR)Q
D-EG3.13OSL 1172.86OSL (MG-SAR)Q
G3.12PAMSA/ATBarton et al. 2009
G3.11OSL 1867.75.3OSL (MG-SAR)QBarton et al. 2009
G3.10Barton et al. 2009
G3.9Barton et al. 2009
G3.8Barton et al. 2009
G3.7Barton et al. 2009
G3.6Barton et al. 2009
G3.5Barton et al. 2009
G3.4OSL 1088.75.9OSL (MG-SAR)QBarton et al. 2009
G3.3OSL 989.56.4OSL (MG-SAR)Q
G3.2PAMSA/ATBarton et al. 2009
FG3.1OSL 878.45.7OSL (MG-AR)QPAMSA/ATBarton et al. 2009
GG2.8
G&HG2.7OSL 7111.910.1OSL (MG-SAR)QBarton et al. 2009
IG2.6OSL 6122.99.1OSL (MG-SAR)QPAMSA/ATBarton et al. 2009
G2.5
G2.4
G2.3
G2.2
JOSL 5A107.17.3OSL (MG-SAR)Q
G2.1OSL 5B109.99.8OSL (MG-SAR)QPAMSA/ATBarton et al. 2009
KG1.19PAMSA/ATBarton et al. 2009
G1.18
G1.17
G1.16OSL 4157.512.8OSL (MG-SAR)Q
G1.15
G1.14
G1.13
G1.12
G1.11
G1.10OSL 20139.810.5OSL (MG-SAR)Q
G1.9
G1.8OSL 3113.99.7OSL (MG-SAR)Q
LG1.7OSL 2112.18.1OSL (MG-SAR)QPAMSA/ATBarton et al. 2009
G1.6
G1.5
MG1.4OSL 1151.49.1OSL (MG-SAR)QBarton et al. 2009
G1.3OSL 17131.417.7OSL (MG-SAR)Q
G1.2
G1.1
BEDROCKOSL 19155.514.2OSL (MG-SAR)QBarton et al. 2009
OSL 21209.229.5OSL (MG-SAR)Q

Synthetical stratigraphy and dates for Dar es Soltan 1 cave. Q, quartz; HB, human burial; PA, presence of artefact in section; AT, Aterian; IBM, Iberomaurusian.

Dar es Soltan 2

11Close to Dar es Soltan 1, this cave was excavated from 1969 to 1976 (Debénath 1976). Excavations were concentrated in the front of the cave because some parts of the inside were obstructed by large blocks. The stratigraphy, with a total thickness of 4.5 m up to the bedrock, is described by a succession of seven layers. Among these, four layers have delivered archaeological material: layer 1 assigned to historical periods; layer 2 attributed to the Neolithic; layer 3 attributed to the Iberomaurusian (subdivided into 3a, 3b and 3c); and layer 6 assigned to the Aterian. Neolithic and LSA layers (respectively layer 1 and 2) including burial (Debénath 1976; 2000). Layer 4 was attributed to a stone slab, layer 5 was considered sterile, as well as layer 7 (characterized by beach sand) As in Dar es Soltan 1, the stratigraphy (table 8.3), compacted in the lower part, includes large blocks of the cave and there is a visible chronological gap between levels (Schwenninger et al. 2010). The new lithostratigraphical study (Schwenninger et al. 2010) has refined the description of the sequence performed by A. Debénath, and shown the presence of burrows, which strongly modify material from other layers, in layers 3b1, 3b2 and 6. Cranial remains have been found in layer 7 (Debénath et al. 1982), which is composed of marine sands. These human remains are not directly associated with Aterian lithic industries or other artefacts in this level. A radiocarbon date initially obtained on Helix shells suggested an age of about 37,000 yrs BP for these Aterian remains (Debénath 1994). However, it was already known that the stratigraphy of this cavity was not well established, so these ages should be considered with great caution (Debénath et al. 1986). New OSL ages obtained using the multi-aliquot quartz grain SAR protocol (MG-SAR) (Schwenninger et al. 2010) placed layer 3a (end of the LSA) at 13,400 ± 700 yrs and layer 7 around 120,000 yrs, a considerably older age.

table 8.3

Stratigraphy (Debénath 1978)Sample codeAges (ka)±MethodMaterialArchaeologyPeriodReferences
1Disturbed + recent
2Neolithic
3aOSL SOL2‑05‑04 X240213.40.7OSL (MG-SAR)QHBLSA/IBMSchwenninger et al. 2010
3b1
3b2OSL SOL2‑05‑03 X240131.12.1OSL (MG-SAR)QSchwenninger et al. 2010
4
5OSL SOL2‑05‑02 X2400101.411.1OSL (MG-SAR)QSchwenninger et al. 2010
6MSA/AT
7OSL SOL2‑05‑01 X2399121.78.2OSL (MG-SAR)QHRSchwenninger et al. 2010
> 27.014CCHBiberson et al. 1977

Synthetical stratigraphy and dates for Dar es Sotlan 2 cave. Q, quartz; CH, charcoal; HB, human burial; HR, human remains; AT, Aterian; IBM, Iberomaurusian.

El Harhoura 1

12This cave, discovered in 1976, was the subject of a rescue excavation in 1976 and 1977 (Debénath 1978; 1979a; Debénath & Sbihi-Alaoui 1979). The sequence is c. 4.5 m thick and the bedrock has been reached by the excavation. In total, three layers (1 to 3 from the base to the top, table 8.4) with archaeological artefacts were identified: layer 3, in the top part of the filling, is attributed to the Neolithic, and the other two layers (2 and 1) are attributed to the Aterian. No Iberomaurusian layer has been recognised in this cave. Human remains represented by a mandible and a canine tooth (Debénath 1979a) were discovered in layer 2. Radiocarbon analysis dated the Neolithic layer to around 5000 yrs BP (Daugas et al. 1989). TL dates (Debénath et al. 1986) obtained on burnt calcarenite placed the age of layer 1 at about 41,000 yrs and 32,000 yrs, respectively. However, because the use of TL was not yet fully developed at this time, A. Debénath (Debénath et al. 1986) recognised that these dates must be considered with caution. Moreover, a radiocarbon age obtained on Helix shells from the same layer gave a younger result of c. 26,000 yrs BP (Occhietti et al. 1993). Although there is a difference between these ages, the end of the Aterian in this cave is placed between MIS 3 and MIS 2. Dates and stratigraphy for this cave are summarised in table 8.4.

table 8.4

Stratigraphy
(Debénath 1978)
Ages (ka)±Ages (ka)
cal BP
(95.4%)
MethodMaterialArchaeologyPeriodReferences
35.40.34.9‑3.614CHRNeolithicDaugas et al. 1989
2HRMSA/AT
141.23.5TLGCMSA/ATDebénath et al. 1986
132.24.8TLGCMSA/ATDebénath et al. 1986
125.60.128.3‑27.414CCHMSA/ATBiberson et al. 1977

Synthetical stratigraphy and dates for El Harhoura 1 cave. SH, sandstone heated; CH, charcoal; HR, human remains; AT, Aterian; IBM, Iberomaurusian. Radiocarbon ages are calibrated with InterCal 13 curves (Reimer et al. 2013).

El Harhoura 2

13This cave, discovered in 1977, was excavated between 1977 and 1996 (Debénath 1979b; Debénath & Sbihi-Alaoui 1979) and again from 2001 to 2015 (Nespoulet et al. 2008b, 2008a). The stratigraphy of c. 6.5 m includes 11 levels (table 8.5), defined by the archaeological material content: layers 1, 2, 3, 4 and 8 have delivered sufficient lithic material for cultural diagnosis and are attributed to the Aterian; layers 9, 10 and 11 did not contain artefactual lithic material, have been excavated in a reduced area and are considered sterile according to the state of the current excavation. Unlike other caves in the region, the bedrock has not been reached. Layer 1 is attributed to the Neolithic. Layer 2 is Iberomaurusian and has delivered many well-preserved human burials (El Hajraoui et al. 2012). The other layers were attributed to the Aterian, even if no tanged tools have been found. These levels contain postcranial human bones. In contrast to other caves in the region, the undisturbed archaeostratigraphical record in El Harhoura 2 makes the site, in theory, a good candidate for chronologically constraining the disappearance of the MSA in the region and discussing the MSA-LSA transition.

table 8.5

StratigraphyLithostratigraphySample codeAges (ka)±Ages (ka) cal BP (95.4%)MethodMaterialArchaeologyPeriodReferences
El Hajraoui et al. 2012
1ISacA 289385.50.044.5‑4.3AMS-14CCHNeolithicEl Hajraoui & Nespoulet 2012 - excavation report
SacA 289346.00.045.0‑4.8AMS-14CCHEl Hajraoui & Nespoulet 2012 - excavation report
SacA 289370.050.03-AMS-14CCHEl Hajraoui & Nespoulet 2012 - excavation report
1FEl Hajraoui & Nespoulet 2012 - excavation report
2IISacA 289351.60.040.4‑0.6AMS-14CCHLSA/IBMEl Hajraoui & Nespoulet 2012- excavation report
2F?8.00.047.1‑6.8AMS-14CCHEl Hajraoui & Nespoulet 2012 - excavation report
3EH09‑1057.34.2OSL (SG-SAR)QHRMSA/ATJacobs et al. 2012
EH09‑0452.63.3OSL (SG-SAR)QJacobs et al. 2012
EH09‑0357.74.2OSL (SG-SAR)QJacobs et al. 2012
EH09‑0251.63.6OSL (SG-SAR)QJacobs et al. 2012
EH09‑0161.94.1OSL (SG-SAR)QJacobs et al. 2012
EH08‑1061.93.5OSL (SG-SAR)QJacobs et al. 2012
EH2‑2993393Combined US‑ESRHBen Arous et al. 2019
EH2‑3986447Combined US‑ESRHBen Arous et al. 2019
EH2‑2730484Combined US‑ESRHBen Arous et al. 2019
3F
4AEH08‑973.74.1OSL (SG-SAR)QJacobs et al. 2012
EH0605443Combined US‑ESRHJanati-Idrissi et al. 2012
EH2‑10221495Combined US‑ESRHBen Arous et al. 2019
EH2‑10984463Combined US‑ESRHBen Arous et al. 2019
SacA 2894322.20.1425.0‑24.1AMS-14COMSEl Hajraoui & Nespoulet 2012 - excavation report
4BIIIEH08‑899.95.8OSL (SG-SAR)QJacobs et al. 2012; Janati-Idrissi et al. 2012
SacA 2895921.60.1AMS-14COMSEl Hajraoui & Nespoulet 2012 - excavation report
5EH08‑7102.65.7OSL (SG-SAR)QJacobs et al. 2012; Janati-Idrissi et al. 2012
6EH08‑6116.46.6OSL (SG-SAR)QJacobs et al. 2012; Janati-Idrissi et al. 2012
EH08‑5108.16.3OSL (SG-SAR)QJacobs et al. 2012; Janati-Idrissi et al. 2012
7EH0603624Combined US‑ESRHJanati-Idrissi et al. 2012
8IVEH08‑4106.76.6OSL (SG-SAR)QJacobs et al. 2012
EH06019211Combined US‑ESRHJanati-Idrissi et al. 2012
9EH08‑3108.15.9OSL (SG-SAR)QHRJacobs et al. 2012
10EH08‑2118.37.1OSL (SG-SAR)QSterileJacobs et al. 2012
11EH08‑1123.77.4OSL (SG-SAR)QJacobs et al. 2012

Synthetical stratigraphy and dates for El Harhoura 2 cave. CH, charcoal; OMS, organic material from sediment; Q, quartz; H, hydroxyapatite; HR, human remains; AT, Aterian; IBM, Iberomaurusian. Radiocarbon ages are calibrated with InterCal 13 curves (Reimer et al. 2013). The dashes in the column ‘Age cal BP’ indicate that several ‘platter’ did not allow for giving a 14C age with 95.6% probability.

14OSL dates obtained using the single grain aliquot SAR protocol (SG-SAR) (Jacobs et al. 2012) placed the Aterian layers between c. 110,000 yrs and c. 55,000 yrs. This cave was the subject of new chronological applications to test the possibility of maximally constraining the end of the MSA. For this, the combined US‑ESR method was applied to date five herbivorous teeth (hydroxyapatite) located in layers 3 and 4A (figure 8.3, table 8.5), covering the most recent MSA occupations. Combined US‑ESR dates show a discrepancy of up to 15,000 yrs between OSL ages (Janati-Idrissi et al. 2012; Ben Arous et al. 2019) for the end of the MSA. In addition, the application of the AMS-C14 method to the organic material from sediments of the same layers (excavation report (Nespoulet & El Hajraoui 2012)) have provided younger ages; these ages cannot be accepted because the origin of the organic phase is not clearly identified in the sediment. However, a unique AMS-14C age in layer 2 (LSA) from the back of the cave that was obtained on the charcoal has provided an age of 8010 yrs ± 40 BP (Nespoulet & El Hajraoui 2012), but it is possible that the LSA begins about 6000 yrs earlier. Thus, by combining all of these dates (US-ESR and AMS-14C), we highlight a chronological gap of between about 30,000 and 25,000 yrs between the end of the MSA and the LSA in El Harhoura 2 cave (MIS 3-MIS 2), although no layer of transition in the stratigraphy has been identified by the excavators on the site. This chronological gap does exist, whatever the method used to date the end of the MSA (SG-SAR or combined US-ESR).

15How, then, can we explain this discrepancy between the OSL and US-ESR ages? The latter are systematically younger than the former. In El Harhoura 2, the stratigraphy is not disturbed by human or biological activities for the end of the MSA (Stoetzel et al. 2014) and the teeth and other artefacts would always be embedded in a sediment deposited thousands of years before; teeth are considered to be contemporaneous with human occupation. Methodologically, first, the US-ESR ages could be younger because of an underestimation of the De (equivalent dose) values obtained for the fossil teeth due to the presence of unstable radicals in the hydroxyapatite, as observed in Joannes-Boyau and Grün (Joannes-Boyau & Grün 2011) and Richter et al. (Richter et al. 2017) on an enamel fragment. However, ESR dose estimates were carried out on enamel powder, resulting in a random spatial distribution of the enamel crystals that renders such a comparison difficult. Nevertheless, the impact on these unstable components remains unclear and needs to be further investigated. It is, indeed, sample dependent, and some teeth are simply not affected by this issue (Dirks et al. 2017) Second, the OSL ages could have been overestimated. When using SG-SAR, many quartz grain populations with common De are identified on the basis of a statistical model (Galbraith & Green 1990; Galbraith et al. 1999). However, it is difficult to understand the meaning of these different population grains (unclear mixing of grains from underlying/overlying levels or the local variations effect of the beta dose rate on individual grains (Jacobs et al. 2012)) in terms of geology: the sedimentary processes in karstic coastal caves are not well understood in terms of chronology and origin because of the impossibility of differentiating between quartz grains from different origins (aeolian, marine, exfoliation of the cave).

El Mnasra

16El Mnasra cave was discovered by J. Roche in the 1950s and excavated in the 1960s (El Hajraoui et al. 2012). Excavations were then carried out again in the 1990s (El Hajraoui 2004). The resumption of excavations in 2004 (Nespoulet et al. 2008a) allowed the definition of 13 levels (table 8.6), including 9 archaeological levels. The bedrock has been reached. Layer 2 is assigned to the Neolithic and layers 4 to 10 are assigned to the Aterian. The cultural attribution of layer 3 is unclear (Neolithic (Nespoulet et al. 2008a) or Iberomaurusian (Jacobs et al. 2012)). In addition, the presence of the Iberomaurusian in El Mnasra is questionable because no associated material was found during recent excavations. Dating by the OSL method (SG-SAR) has placed this uncertain layer 3 between about 50,000 to 60,000 yrs (Jacobs et al. 2012), which seems incompatible with the cultural context for the LSA in the other local caves. Other dates, which have been obtained by combined US‑ESR (Janati-Idrissi et al. 2012), show a discrepancy of c. 10,000 yrs compared with OSL ages. The recent lithostratigraphical reading of A. Lenoble (Lenoble 2010, unpublished) on the cave has shown that layer 3 mainly corresponds to the reworked layer 4, and may also contain reworked sediments from layers 5, 6 and 7. It appears, therefore, that several biological disturbances in El Mnasra have mixed the layers. The radiocarbon ages (excavation report (Nespoulet & El Hajraoui 2012), table 8.6) of layer 3 are incompatible with the cultural context (LSA? Neolithic?) and illustrate this problem of sedimentary disturbances.

table 8.6

Archeostratigraphy (El Hajraoui 1993)

Archeostratigraphy (Debénath 2006)

Sample code

Ages (ka)

±

Ages (ka) cal BP (95.4%)

Method

Material

Archaeology

Period

References

1

1

 

 

 

 

 

 

Recent

 

2

2

SacA 28931

7.5

0.04

6.6‑6.5

AMS-14C

CH

Neolithic

El Hajraoui & Nespoulet 2012 - excavation report

SacA 28957

4.2

0.04

-

AMS-14C

OMS

El Hajraoui & Nespoulet 2012 - excavation report

3

3

SacA 28930

6.3

0.05

-

AMS-14C

CH

Neolithic? LSA/IBM?

El Hajraoui & Nespoulet 2012 - excavation report

SacA 28933

6.3

0.04

-

AMS-14C

CH

El Hajraoui & Nespoulet 2012 - excavation report

SacA 28954

8.9

0.05

-

AMS-14C

OMS

El Hajraoui & Nespoulet 2012 - excavation report

SacA 28955

15.6

0.08

16.9‑16.5

AMS-14C

OMS

El Hajraoui & Nespoulet 2012 - excavation report

EM10‑1

75.3

5.6

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

EM10‑4

ind.

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

EM10‑2

71.6

5.3

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

4

4

SacA 28956

18.1

0.09

20.3‑19.8

AMS-14C

OMS

MSA/AT

El Hajraoui & Nespoulet 2012 - excavation report

EM10‑5

95.4

9.3

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

EM10‑3

94.6

9.7

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

EM10‑6

106.7

9.6

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

EM08‑12

103.5

7.5

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

5

5

SacA 28958

35.0

0.5

38.9‑36.5

AMS-14C

OMS

S

El Hajraoui & Nespoulet 2012- excavation report

X2416

105.5

12

 

OSL (MG-SAR)

Q

Schwenninger et al. 2010

EM08‑11

106.5

6.5

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

EM08‑10

107.5

6.6

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

6

6a

EM0601

66

2

 

OSL (SG-SAR)

Q

Janati-Idrissi et al. 2012

EM08‑9

107.4

5.8

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

6b

EM0603

67

2

 

Combined US‑ESR

H

Janati-Idrissi et al. 2012

EM08‑6

116.7

6.4

 

Combined US‑ESR

H

Janati-Idrissi et al. 2012

7

7

EM0604

89

6

 

Combined US‑ESR

H

Janati-Idrissi et al. 2012

EM08‑7

108.8

6.6

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

X2415

107.6

9.6

 

OSL (MG-SAR)

Q

Schwenninger et al. 2010

8

8

EM08‑6

116.7

6.4

OSL (SG-SAR)

Q

S

Jacobs et al. 2012

OSL (SG-SAR)

Q

Jacobs et al. 2012

9

9

EM08‑5

108.9

6.2

 

OSL (SG-SAR)

Q

S

Jacobs et al. 2012

10

10

EM08‑4

108.5

6.3

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

11

EM08‑3

108.3

6.6

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

11

X2414

111.5

6.9

 

OSL (MG-SAR)

Q

S

Schwenninger et al. 2010

12

12

EM08‑2

121

6.9

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

X2413

118.6

20.1

 

OSL (MG-SAR)

Q

Schwenninger et al. 2010

13

EM08‑1

133.2

7

 

OSL (SG-SAR)

Q

Jacobs et al. 2012

Synthetical stratigraphy and dates for El Mnasra cave. CH, charcoal; OMS, organic material from sediment; Q, quartz; H, hydroxyapatite; HR, human remains; S, sterile; AT, Aterian; IBM, Iberomaurusian. Radiocarbon ages are calibrated with InterCal 13 curves (Reimer et al. 2013). The dashes in the column ‘Age cal BP’ indicate that several ‘platter’ did not allow for giving a 14C age with 95.6% probability.

Contrebandiers

17This cave was discovered in 1955 by J. Roche. Surveys were carried out between 1955 and 1957 and excavations from 1967 to 1975 (Roche 1976). The cave is known for its large number of MSA human remains, including a mandible and the occipital of separate adults discovered by J. Roche in 1956 and 1975 (Vallois & Roche 1958; Roche 1976). The precise context of these remains, however, is not well known. In 2009, a new immature fossil was added to this ensemble (Balter 2011).

18The sequence (table 8.7, figure 8.3), c. 5.5 m thick, was divided into 16 layers (Roche 1976) in the central excavation area (CEA). Among these, there are seven archaeological layers. Layers 2, 4 and 5 are attributed to the Neolithic and layers 7, 8, 9 and 11 are associated with the LSA. Further excavations were undertaken in 1994 (Bouzouggar 1997b). The excavations in 2006 (Dibble et al. 2012) allowed the opening of new sectors (I to V). Although J. Roche initially recognised only seven archaeological layers, subsequent studies (Schwenninger et al. 2010; Dibble et al. 2012) have shown more accurately the lateral variations over the entire site, both in terms of unit composition and stratigraphical subdivision. The stratigraphy and taphonomic of the site are further complicated by the presence of numerous Iberomaurusian pits, also present in El Harhoura 2 (El Hajraoui et al. 2012), that extend into the underlying Aterian layer. Layers 1 to 5 (Dibble et al. 2012) are associated with Neolithic occupation of the site and are separated from the underlying LSA context (limited to layer 7) by a sterile layer (layer 6). The well-preserved levels are in layers 8 to 14 and cover two sterile sandy units (layers 15 and 16). The latter unit contains abundant marine shell debris, and is assumed to represent an interglacial beach deposit. Radiocarbon dating on bone, shell and sediment samples was initiated in the 1970s (Daugas 2002). However, these results are considered to be problematic (Delibrias et al. 1982): the reliability of dates is highly questionable, given the difficulties surrounding the dating of these types of materials (Occhietti et al. 1993) and there is a lack of modern pre-treatment and control methods (Occhietti et al. 2002).

table 8.7

Sector IV
ArchaeostratigraphyLithostratigraphy Sample codeAges (ka)±MethodMaterialPeriodReferences
IV1aH2LSA/IBM
IV1bH1PT47A40.64.7EU-ESRHDibble et al. 2012
PT47A54.17.2LU-ESRHDibble et al. 2012
PT47A64.99.4RU-ESRHDibble et al. 2012
PT47B38.12.6EU-ESRHDibble et al. 2012
PT47B48.23.4LU-ESRHDibble et al. 2012
PT47B59.74.6RU-ESRHDibble et al. 2012
PT4644.32.8EU-ESRHDibble et al. 2012
PT4656.33.9LU-ESRHDibble et al. 2012
PT4670.65.3RU-ESRHDibble et al. 2012
IV-2G2CONT-338711TLHSMSA/ATDibble et al. 2012
CONT-348710TLHSDibble et al. 2012
CONT-3611511TLHSDibble et al. 2012
CONT-378011TLHSDibble et al. 2012
G1CONT-398511TLHSDibble et al. 2012
CONT-5317914TLHSDibble et al. 2012
SC30Jacobs et al. 2011
SC39968OSL (SG-SAR)QJacobs et al. 2011
Central excavation area
ArchaeostratigraphyLithostratigraphyStratigraphy (Roche, 1976b) Sample codeAges (ka)±MethodMaterialPeriodReferences
RT-4SC31926QMSA/ATJacobs et al. 2011
RT-4SC32977 QJacobs et al. 2011
4aF28?
X240959.410.2OSL (MG-SAR)QSchwenninger et al. 2010
4cSC211036OSL (SG-SAR)QJacobs et al. 2011
SC351059OSL (SG-SAR)QJacobs et al. 2011
4dF19/ 10SC201047OSL (SG-SAR)QJacobs et al. 2011
SC81089OSL (SG-SAR)QJacobs et al. 2011
SC71179OSL (SG-SAR)QJacobs et al. 2011
X241196.114OSL (MG-SAR)QJacobs et al. 2011
X2408100.18.4OSL (MG-SAR)QJacobs et al. 2011
X241099.310.6OSL (MG-SAR)QJacobs et al. 2011
5aD311aCONT-58916TLHSMSA/MOUSTDibble et al. 2012
CONT-289214TLHSDibble et al. 2012
FT93614.8EU-ESRHDibble et al. 2012
FT9383.67.9LU-ESRHDibble et al. 2012
FT93117.417.9RU-ESRHDibble et al. 2012
PT4443.70.7EU-ESRHDibble et al. 2012
LU-ESR65.31.1LU-ESRHDibble et al. 2012
PT4489.71.9RU-ESRHDibble et al. 2012
FT8364.57.7EU-ESRHDibble et al. 2012
FT8388.76.2LU-ESRHDibble et al. 2012
FT83123.29.7EU-ESRHDibble et al. 2012
SC161188OSL (SG-SAR)QJacobs et al. 2011
5bD211bX2407105.810OSL (MG-SAR)QSchwenninger et al. 2010
CONT-528914TLHSDibble et al. 2012
FT8447.12.3EU-ESRHDibble et al. 2012
FT8470.43.5LU-ESRHDibble et al. 2012
FT84114.87.4RU-ESRHDibble et al. 2012
SC151168OSL (SG-SAR)QJacobs et al. 2011
5cD111cX2406104.57.8OSL (MG-SAR)QSchwenninger et al. 2010
SC191249OSL (SG-SAR)QJacobs et al. 2011
SC141137OSL (SG-SAR)QJacobs et al. 2011
SC61148OSL (SG-SAR)QJacobs et al. 2011
SC131127OSL (SG-SAR)QJacobs et al. 2011
SC11158OSL (SG-SAR)QJacobs et al. 2011
SC281158OSL (SG-SAR)QJacobs et al. 2011
SC21148OSL (SG-SAR)QJacobs et al. 2011
CONT-5011613TLHSDibble et al. 2012
5dSterile?
6a/bC2 crust12SC31078OSL (SG-SAR)QJacobs et al. 2011
SC121197OSL (SG-SAR)QJacobs et al. 2011
C113a/13b/13cSC181146OSL (SG-SAR)QJacobs et al. 2011
SC291097OSL (SG-SAR)QJacobs et al. 2011
X2405100.410.3OSL (MG-SAR)QSchwenninger et al. 2010
6cB13d/14/15SC111229OSL (SG-SAR)QMSA/MOUSTJacobs et al. 2011
SC171197OSL (SG-SAR)QJacobs et al. 2011
SC41208OSL (SG-SAR)QJacobs et al. 2011
SC101309OSL (SG-SAR)QJacobs et al. 2011
X2404121.813.8OSL (MG-SAR)QSchwenninger et al. 2010
X24031296.6OSL (MG-SAR)QSchwenninger et al. 2010
7A16X241270.13.8OSL (MG-SAR)QSterileSchwenninger et al. 2010
BEDROCKSC2234127OSL (SG-SAR)QJacobs et al. 2011
Sector V
ArchaeostratigraphyLithostratigraphy Sample codeAges (ka)±MethodMaterialPeriodReferences
V-B2V1aAT1804.2EU-ESRHMSA/ATDibble et al. 2012
AT1109.26.6LU-ESRHDibble et al. 2012
AT1132.29.1RU-ESRHDibble et al. 2012
PT4582.87.2EU-ESRHDibble et al. 2012
PT45105.311LU-ESRHDibble et al. 2012
PT45120.914.4RU-ESRHDibble et al. 2012
FT9580.12.4EU-ESRHDibble et al. 2012
FT9593.93LU-ESRHDibble et al. 2012
FT95107.73.9RU-ESRHDibble et al. 2012
FT9467.91.6EU-ESRHDibble et al. 2012
FT9486.32.1LU-ESRHDibble et al. 2012
FT94107.83.1RU-ESRHDibble et al. 2012
V-B1V1bSC23
V-AV2SC34

Synthetical stratigraphy from the three excavated sectors and dates for Contrebandiers cave. CH, charcoal; OMS, organic material from sediment; Q, quartz; H, hydroxyapatite; HR, human remains; S, sterile; AT, Aterian; IBM, Iberomaurusian.

figure 8.3

Image

Chronological synthesis of the Témara caves. The ages of the same layer are located on the same vertical line, below the corresponding layer number. Dates on the same vertical line belong to the same layer. Some published dates are clearly associated with 2 or 1 sigma error (when this information is not clear, the error is represented by a dashed line).

19However, this cave provided an important set of data using several dating techniques (Jacobs et al. 2011; Dibble et al. 2012) spread over the three sectors (sector IV, CEA and sector V) (table 8.7, figure 8.3): OSL (SG-SAR and MG-SAR protocols), TL and ESR. Only the CEA is clearly related to Roche’s stratigraphy. There appears to be less age dispersion in the CEA (all methods combined). Some discrepancies (figure 8.4, table 8.7) concern the ESR ages and the TL/OSL or between OSL ages obtained using SG-SAR and MG-SAR. In sector V, discussing the timing of the MSA is premature in the absence of ages obtained by other methods.

figure 8.4

Image

Chronological overview comparing the Rabat-Témara sites and the eastern Morocco sites (Taforalt, Rhafas and Ifri n’Ammar caves). Isotopic curves from Lisieky and Raymo (Lisiecki & Raymo 2005).

20We mentioned three excavation sectors in Contrebandiers, but the dating of sector IV is not considered in the discussion in this paper (section entitled ‘Chronological comparison’), because the stratigraphy of this sector is not well established and some perturbation has taken place. In fact, as pointed out by H. Dibble (Dibble et al. 2012), sector IV provided the only layers that contained LSA along with numerous bone fragments and marine shells, but these layers were characterised by erosional action, and the contact with underlying layers (MSA) is unclear. ESR ages were applied to tooth enamel to date LSA/MSA assemblages using conventional models (early uptake (EU); linear uptake (LU); recent uptake (RU)). Due to the uranium content variation over time, this parameter was estimated according to the US model, as recommended by many authors (Grün 1989). However, EU-ESR can also be used, but as minimal ages. Further, while waiting for additional dates from this sector, it is impossible to use ESR ages, the latter not being compatible with the LSA context. This highlights problems with mixing of archaeological materials.

21Globally, as shown in figure 8.2, dating indicates a short sequence in the CEA, and layers 4 through 7 were all deposited during early to mid MIS 5.

Chaperon-Rouge open-air site

22Located 7 km south-east of Rabat, this stratified open-air site was discovered in 1974 by M. de Bazac. The site was the subject of excavations from 1976 to 1979 by J. Texier and J. Roche. Sand sequence (Texier et al. 1988) (table 8.8) is c. 1‑3 m thick and based on calcarenite. Several ‘horizons’ were described, each of them containing archaeological artefacts (from the top to the base): A1 (Neolithic), B and BCt (both LSA) and B’ (at the top, Aterian MSA industries) (Texier 1986). TL was applied in order to date burnt flint from the MSA layer (horizon B'). OSL was applied to quartz grains (10 cm above the layer with MSA artefacts). Two close dates were obtained, 28,200 ± 3300 yrs and 24,000 ± 3050 yrs, respectively. In addition to the paucity of chronological data for Chaperon-Rouge, the reliability of the OSL/TL date is limited due to the non-adapted protocols at the time (aliquot size, absence of sensitivity change control during luminescence measurements) as already pointed out by Doerschner et al. (Doerschner et al. 2016).

table 8.8

ArchaeostratigraphySample codeAges (ka)±MethodMaterialPeriodReferences
A1Texier et al. 1988
BTexier et al. 1988
BCtTexier et al. 1988
B'OX TL 724 g224.03.05OSLQMSA/ATTexier et al. 1988
B'OX TL 724 g128.23.3TLBFMSA/ATTexier et al. 1988

Synthetical stratigraphy and available dates for the Chaperon-Rouge open-air site. Q, quartz; BF, burnt flint.

Same fillings, same mixing problems: A limitation with regard to OSL and ESR dating

23A synthetical stratigraphical view for each cave is provided in figure 8.2. The majority of the Témara caves have similar sedimentary fillings that cover the entire Upper Pleistocene. Some layers, particularly for the end of the MSA and the LSA, have undergone post-depositional processes in relation to bioturbation that have affected the layers. These processes, whether bioturbation by humans or animals, or by chemical/physical processes, are still poorly understood. Such effects, unidentified in the layers (lateral sedimentary variations may exist inside the caves), have led to mixing of the layers and/or archaeological materials, thus limiting the application of dating methods such as OSL and US‑ESR.

Chronological comparison

Chronological data from the Témara caves: A heterogeneous framework for the end of the MSA

24Before discussing the chronological data, it should be mentioned that human occupations can vary from one cave to another over time, even if the caves are very close. It is, therefore, expected that discrepancies will exist between the end of MSA occupation dates. It should be noted that the dating results for the end of the MSA are quite heterogeneous and vary from one cave to another.

25We have compared all the published chronological data for the Rabat‑Témara region. The data used for this synthesis are summarised in tables 3 to 8. We consider archaeological layers dated by absolute dating and their cultural attribution. All dating methods are indicated in the legend of figure 8.3: AMS-14C, OSL (SG-SAR and MG-SAR), TL and ESR (RU, LU and EU models, and combined US‑ESR). Dates have been distributed according to the simplified stratigraphy of the Témara sites and the associated layers are indicated by the upper number. On the basis of this chronological comparison, we can make four main remarks:

  • The MSA of El Harhoura 1 cave and the Chaperon-Rouge open-air site overlaps with the LSA of Dar es Soltan 1 and 2 caves. Moreover, the context of samples from El Harhoura 1 is poorly understood, and there is limited reliability with regard to the OSL/TL dates for Chaperon-Rouge because MSA dates cannot be used with certainty in this discussion.
  • At Contrebandiers cave, only sector IV has delivered MSA (layer IV-2) and LSA layers (layers IV-1b and IV-1a). In this sector, the age of the MSA layer is contemporaneous with the oldest LSA elsewhere on the site. As seen above (section entitled ‘Contrebandiers’), LSA ages are unreliable (in view of the sedimentary context, there is incompatibility between ESR ages and the archaeological context).
  • In El Mnasra cave, the OSL ages for layer 3 (c. 70,000 yrs) are far too old considering the cultural association (LSA); in addition, there are reservations about cultural attribution if sedimentary disturbances are taken into account.
  • The MSA‑LSA transition is situated between 50,000 and 40,000 yrs in Dar es Soltan 1 (Barton et al. 2009), the gap between these two periods being greater at Dar es Soltan 2 (Schwenninger et al. 2010) (between 90,000 and 40,000 yrs). In addition, the OSL age of layer 3b2 at Dar es Soltan 2 is associated with unclear sedimentary context, as is layer B at Dar es Soltan 1, so it is doubtful whether they should be included in the discussion.

Upland sequences

26The comparison with other well-established sequences (figure 8.4) from caves in upland areas, such as Taforalt, underlines two new problems:

  • For the end of the MSA, there is a chronological gap between the upland and coastal regions. Indeed, the end of the MSA in Taforalt cave is dated to around 29,000 yrs (Barton et al. 2015), whereas it is around 40,000 yrs in Témara. This represents a 10,000 yrs difference between the two geographical areas for the end of the MSA.
  • In addition, the chronology in the Rabat‑Témara caves between the MSA and the LSA shows a human occupation gap of between about 30,000 to 25,000 yrs, whereas at Taforalt, during this gap, there is the presence of a transition layer (characterised by an intermediate adze industry that is still difficult to interpret) (Barton et al. 2015) and an overlying early Iberomaurusian dating from around 23,000 yrs cal BP. This intermediate industry in Taforalt is not found at the other sites dated (Ifri n’Ammar and Rhafas) in the same region.

Climate and human mobility between coastal and inland areas

27The dates from the Rabat‑Témara caves, particularly the combined US-ESR ages from El Harhoura 2 cave, imply an absence of human occupation between c. 40,000 and c. 8000 yrs. This human absence period could be correlated with a humid event until c. 30,000 yrs but, then, an aridification took place at the end of the Upper Pleistocene (Stoetzel et al. 2011; Blome et al. 2012) The climatic interpretation becomes more complicated if we include all the data from Témara that put the end of the MSA between about 70,000 to 50,000 yrs.

28In the caves of this region, alternating phases of human occupation and phases of human absence/presence of carnivores (Campmas et al. 2017) are linked to climatic fluctuations and variations in sea level over the past 120 millennia (El Hajraoui et al. 2012). At 40,000 yrs ago, the Témara caves were located at least about 20 km inland from the current coastline (Chahid 2017); these caves may have been isolated by the extension of the Sahara. Two hypotheses can, therefore, be proposed to explain this MSA human absence on the coastal area starting at about 40,000 yrs ago:

  • Human adaptation to a coastal environment during the MSA (Campmas 2017), as shown by the exploitation of marine resources, would probably indicate that the people followed the coastline during this aridification phase; these occupation sites are currently below sea level.
  • The possible long-distance inland mobility of groups exploiting coastal and upland/mountain resources during the MSA as suggested by the use of Nassarius in Taforalt. However, this marker disappears around 60,000 yrs, suggesting that mobility from the coast to the mountains ceased after this period. Other studies have shown the high mobility (up to distances of more than 50‑100 km) of human groups living in the coastal zone (because they needed to collect raw materials) (El Amrani El Hassani & Morala 2012) until the end of the MSA. This result suggests a two-way human mobility between coastal and inland areas.

29In order to explore the possibility of a coastal/inland connection at the end of the MSA, future research will focus on comparing other upland sites, such as M’Tsogatin cave in the Oulmes region discovered in 2012 (El Amrani El Hassani et al. 2013). It would then be possible to highlight the supposed existence of MSA occupations in mountains further inland during MIS 3 and MIS 2.

Conclusion

30The synthesis of chronological data from the Rabat‑Témara caves has revealed heterogeneous data for the end of the MSA. The sedimentary context for the end of the MSA is largely disrupted for some caves. The difficulty of interpreting the chronologies of the Rabat‑Témara caves is due mainly to the sedimentary context and the problematic interpretation of sediment fillings (human presence, faunal bioturbation, poorly known origin of quartz).

31Dating comparisons using different methods (combined US-ESR, TL, SA-SAR and AMS-14C) highlight a discrepancy of up to several thousand years old (as shown in El Harhoura 2, with combined US-ESR yielding more recent dates than those obtained by OSL). Further investigation is required to fully understand the reason for such discrepancies between methods; new excavations and the dating of more samples with different methods, such as ESR and TL, would have to be carried out to date human occupation in the Rabat‑Témara caves directly. Particular attention should be given to the stratigraphical complexity, and post-depositional processes, such as the actions of burrowing animals, which have in some cases disturbed the deposits (Campmas et al. 2016). This will allow the use of a multi-method approach to identify the upper limits of the MSA in these caves and refine the chronology of the region. Chronological data from the Rabat-Témara caves, especially from El Harhoura 2 and Dar es 2 Soltan, point to an absence of humans between about 40,000 yrs and 10,000 yrs. With regard to the LSA, dating and other studies are in progress but they do not appear to have closed the chronological gap identified.

32In order to infer a potential human presence between MIS 3 and MIS 2 at inland sites contemporaneous with the gap highlighted at Rabat‑Témara (MIS 3‑2), more sequences (e.g. M’Tsogatin) should be dated in both caves and at open-air sites. The correlation with available climatic data raises the question of the continuity of human occupation on the coast (caves vs open-air sites) and in the hinterland, as suggested by analogous archaeological sequences such as Taforalt cave and the so-called ‘MSA-LSA transitional’ layer. In order to highlight the possible presence of MSA populations at sites in the hinterland of Rabat-Témara between MIS 3 and MIS 2, much more dating in these areas will likely allow us to confirm the hypotheses with regard to continued human occupation here (M’Tsogatin 1, Oulmes area) and the connection with the coastal area.

Acknowledgements

33The authors thank A. Akerraz, the director of INSAP, for access to the faunal collections from El Harhoura 2. This study has not benefited from a grant or funding, but the authors would like to thank the workshop organisers, A. Leplongeon, M. Goder-Goldberger and D. Pleurdeau for their participation. We would also like to thank A. Havé and L. Legendre for their constructive and helpful discussion, advice and proofreading. We are grateful to Nick Barton and the anonymous reviewer for constructive and useful comments on an earlier version of this article.


Le texte seul est utilisable sous licence Licence OpenEdition Books. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.