Introduction

1Despite the acknowledged wealth of evidence for the Middle Stone Age (MSA) and the Later Stone Age (LSA) in Northwest Africa, surprisingly little is known about the cultural transitions within and between these periods. To redress this imbalance several projects in Algeria, Tunisia and especially Morocco have been collecting chronological and environmental data from known sites as well as surveying promising new areas with high potential for further research. These works have focused primarily on cave sites (Morocco and Algeria) and more rarely on open air sites (Tunisia) (figure 7.1) because they currently offer the best opportunities for studying long stratified records of human activity and are locations in which datable organic materials and other cultural residues are often well preserved. This is particularly so in Morocco where sites contain MSA through to LSA sequences. Unlike Europe, these cultural phases appear to have occurred exclusively within Homo sapiens populations (Hublin 2001; Hublin et al. 2017) and are often characterised by an abundance of symbolic artefacts and other indicators of cognitive complexity (Bouzouggar et al. 2007b; Nespoulet et al. 2008; D’Errico et al. 2009).

figure 7.1

Definition, origin and chronology of the Northwest African MSA

2Traditionally, the Middle Stone Age of Northwest Africa has been sub-divided into two facies: the Mousterian and the Aterian (table 7.1). Based primarily on lithic artefact typo-technological features, the Aterian is assumed to have succeeded the Mousterian chronologically (Bordes 1976‑1977). The differences between these two assemblage-types, however, remains controversial and is not helped by the fact that there are very few stratigraphic sequences where Aterian and Mousterian assemblages are represented and have been described in detail.

3The most widely accepted classification of the Aterian is still that of J. Tixier which describes the Aterian as being made up of Levallois blades (Tixier 1959), endscrapers and a high proportion of tanged/pedunculate pieces (around 25%). Other authors would include as Aterian a variant with pebble tools (Debenath 1994). In contrast, the definition of the Mousterian is mainly based on the presence or absence of certain tool-types. The most significant of these is the absence of tanged or pedunculate points. While accepting these definitions at face value, it is clear that much variation exists within the broadly defined MSA, some of which may be linked to function while others to chronology or to geography (Scerri 2013a, 2017). A study supporting the latter was published by Roche for Grotte des Pigeons at Taforalt (Roche 1969). In it he described the following succession (from the base up): Aterian with Mousterian tradition without pedunculate pieces in layer H, a Middle Aterian in layer F and an Upper Aterian in layer D. Our long-term excavations at the same cave have failed to confirm this sequence, although we have noted many more subdivisions than described by Roche, including levels with and without pedunculate tools (Bouzouggar et al. 2006). Roche’s scheme was adopted at other sites like Rhafas (Wengler 1993) but, here again, more recent excavations have been unable to corroborate a similar succession (Dorschner et al. 2016).

4A further hindrance to understanding the development of the MSA in this region has been a scarcity of reliable dating evidence. Originally estimated as spanning the period 40,000 BP to 20,000 BP (Debenath et al. 1986), the first serious challenge to this model came from the application of luminescence dating to sites in the eastern Sahara (Cremaschi et al. 1998; Garcea 2001, 2004). Here the Aterian was dated to c. 75,000 years, while in Cyrenaica there is an estimated age of c. 70,000 years for the “Levalloiso-Mousterian” at Haua Fteah (Douka et al. 2014), suggesting a much longer chronology than earlier suspected. In Morocco, Electron Spin Resonance dating methods subsequently revealed that the Aterian in El ‘Aliya Cave in the Northwestern peninsula was dated to between 42,000 years (EU) and 56,000 years (LU) (Wrinn & Rink 2003), though uncertainties persist due to the fact that the samples derive from ‘‘old’’ excavations. At Rhafas Cave, luminescence ages of between c. 135,000 years and c. 56,000 years have now been recorded for the Aterian (Dörschner et al. 2016). At Grotte des Pigeons, Taforalt, a range of proxies (Optically Stimulated Luminescence, Thermoluminescence, and Uranium Series dating) indicate that the earlier Aterian levels lie between 73,000 years and 91,000 years (Bouzouggar et al. 2007) and possibly earlier (Clark Balzan et al. 2012). Even older ages have been claimed for Ifri n’Ammar, eastern Morocco (Mikdad et al. 2004; Richter et al. 2012), while new research in Dar es-Soltan 1 near Rabat demonstrates that the Aterian with pedunculates can be dated by OSL to 110,000 years and is of considerably greater antiquity than previously suspected with a chronology extending back into MIS 5 (Barton et al. 2009). In other caves along the Atlantic coast the Aterian can be extended as far back as c. 116,000 years at Contrebandiers, El Harhoura and El Mnsara Caves. Based on all of these new data, it is clear that the known chronology has at least tripled in length since the application of modern scientific dating techniques. The same pattern seen in Morocco has been noted in Tunisia, where an Aterian with pedunculates can be documented in lake shoreline deposits at Nefta and dated by OSL to between 72,000 and 98,000 years (Schwenninger perscomm). The often quoted TL dates of 70,000 years for Oued El Akarit (Roset & Harbi-Riahi 2007) are less reliable but may provide a broad indication of the age of an Aterian-like industry at that site.

5In terms of the origins of the MSA in Northwest Africa, the currently oldest known dates come from Djebel Irhoud (Richter et al. 2017) and Grotte du Rhinocéros (Rhodes et al. 2006), both around 300,000 years, with further examples from Benzu Cave dating to around 255,000 years (Ramos Munoz et al. 2008). The assemblages contain evidence of Levallois technology and at Djebel Irhoud this is accompanied by a wide range of side scrapers and a few foliate points have also been reported (Richter et al. 2017). Elsewhere, MSA assemblages without pedunculate points have been recovered at Ifri n’Ammar (Morocco) in levels putatively dated to 171,000 ± 12,000 years (Nami & Moser 2010: 35), and small tanged points occur at this site from around 145,000 ± 9000 years (Nami & Moser 2010: 35). Isolated finds of small pedunculate points have been noted at various locations in the Western Desert in the Sahara and seem to suggest widespread occupation across Northwest Africa during the more humid phases of MIS 5 (Drake & Breeze 2016). After about 116,000 years the number of dated Aterian sites can be shown to increase in the Maghreb with assemblages alternating between those ‘with’ and ‘without’ pedunculate points (Dibble et al. 2013). Another important indication of the Aterian includes the appearance of blade manufacture at sites such as Taforalt, Rhafas, El Mnasra, and Contrebandiers. In some cases there is also a persistence of some late Acheulean “archaic tools” as at Chaperon Rouge II (Texier et al. 1986) and Bizmoune Cave (Bouzouggar et al. 2017). On the basis of this information we would suggest that the MSA in Northwest Africa probably represents an indigenous development rather than being introduced from the east from places such as Egypt or Nubia (Caton-Thompson 1946). Leaving aside the relevance of typological fossils directeurs in the lithic inventories, the Maghrebian MSA shows the widespread use of a hafted technology, in which lithic artefacts were part of composite tools set in organic shafts and handles.

The Iberomaurusian

6The Iberomaurusian is a microlithic bladelet industry recorded at sites across the Maghreb (present-day Morocco, Algeria, Tunisia and parts of Libya) (Barton et al. 2013). It represents part of the early Later Stone Age in North Africa and marks a cultural break with the flake and blade technologies associated with the Middle Stone Age in this region (Close & Wendorf 1990). The term Iberomaurusian was first adopted by Pallary (Pallary 1909) who noted similarities between lithic assemblages in Spain and Morocco that contained a profusion of tiny backed blades and acute points (‘une profusion de très petites lames à dos retouché et à pointe très aiguë’). This definition was refined by Camps who described it as a microlithic tradition dominated by backed bladelets, which account for between 40‑80% of the retouched tool assemblages (Camps 1974). The use of the microburin technique for producing microlithic backed points is a central feature of this technology.

7In addition to the Iberomaurusian, LSA backed bladelet assemblages in North Africa have been referred to by several other terms, including the ‘Oranian’ (Gobert & Vaufrey 1932) and ‘Mouillian’ (Goetz 1941) for sites in Algeria, while the Eastern Oranian for sites in Cyrenaican Libya (McBurney 1967). Since these microlithic industries are in practice difficult to distinguish from one another, we refer to them generically as Later Stone Age whilst maintaining the common local name.

8The chronology of the Iberomaurusian has been the subject of intensive study over the last decade. A long sequence of over 50 AMS radiocarbon dates for Iberomaurusian cultural deposits at Grotte des Pigeons, Taforalt, has revealed a time-span covering 23,459 to 12,548 cal BP (Barton et al. 2013, 2016; Bouzouggar et al. 2008). Further dates for the Iberomaurusian have been published from Afalou Bou Rhummel in Algeria (Hachi 1996), Ifri n’Ammar (Linstadter et al. 2012), Ifri el Baroud (Linstadter et al. 2012) and Kehf el Hammar in Morocco (Barton et al. 2005; Bouzouggar et al. 2006), but none of these are older than Taforalt.

9A slightly earlier age determination of 25,845‑25,270 cal BP has recently been published for the oldest Iberomaurusian deposits at Tamar Hat in Algeria (Hogue & Barton 2016). Importantly, the earliest date for the Eastern Oranian at the Haua Fteah appears to be no older than c. 17‑19 ka BP (Douka et al. 2014), implying that the first appearance of the Iberomaurusian is more recent in Cyrenaica than further to the west.

10There is currently no agreement concerning the origins of the LSA/Iberomaurusian and three main theories are suggested:

Indigenous development

11The Iberomaurusian developed in Northwest Africa with no presently known cultural antecedents. The epicentre may have been in Algeria, with a subsequent spread both westwards into Morocco and eastwards into Libya and Cyrenaica. This scenario is consistent with the dating evidence, which reveals that the earliest dates are for Tamar Hat with slightly younger ages from Grotte des Pigeons, Taforalt and Kehf el Hammar (Barton et al. 2005), and much younger dates from Libya and Cyrenaica. These dates point towards a cultural break at around 25,000 cal BP.

Exogenous development

12Over the years a number of theories have been advanced concerning the possible introduction of the Iberomaurusian from Europe either from Sicily or across the Gibraltar Strait (Ferembach 1985). There is little direct evidence for this idea despite the fact that Camps referred to the Iberomaurusian as ‘African Epigravettians’ (Camps 1974) based on similarities between the Iberomaurusian and the Italian Epigravettian. A substantial reduction in sea levels during the Last Glacial Maximum would have reduced the distance between Tunisia Tunisia and Sicily to possibly as little as of nine kilometres (Capart & Capart 1986), making the shorelines easily inter-visible. A weakness of this argument is that there are no known early Iberomaurusian sites in northern Tunisia and few Epigravettian sites in either Sicily or southern Italy that can be dated to the LGM (Mannino et al. 2011, 2012). However, this may reflect a lack of fieldwork in some areas and it is possible that there are relevant sites submerged in areas of the continental shelf that are now flooded.

13An alternative scenario is that populations using microlithic bladelet technologies spread westwards from the Near East to the Maghreb before 15 ka BP. This model would necessitate a relationship between the Iberomaurusian and the Natufian and earlier Epipalaeolithic technologies of the Near East. This is inconsistent with current evidence, which suggests that the oldest Iberomaurusian microlithic bladelet technologies are mostly earlier than cultural equivalents in the intervening geographic areas of Egypt and Libya. Nevertheless it might be possible to accommodate an arrival of people from the east who admixed with local populations and replaced the MSA industries with ones dominated by microlithic bladelets.

Continuity with the Middle Stone Age

14This model is based on the premise that the Iberomaurusian emerged directly from the MSA or from one or more ‘transitional technocomplexes’ in Africa. At Taforalt, the layers directly underlying the Iberomaurusian are rich in adzes (Barton et al. 2013). Further west in Cyrenaica the transition to the Iberomaurusian /Oranian is marked by a gradual evolution from the Dabban (McBurney 1967). The Dabban industry incorporates both typical North African MSA artifacts including small Levallois cores and ‘Upper Palaeolithic’ tools such as backed blades and bladelets, chamfered tools, burins and end-scrapers. Continuity with the overlying layers is shown by increasing numbers of backed bladelets and a diminution of some ‘UP’ types (such as awls and chamfered pieces). At Haua Fteah the Early Dabban is dated between c. 39,000 and 32,000 BP and the Late Dabban is dated between c. 22,000 and 19,000 BP (Douka et al. 2014) thus bracketing the period in which the Iberomaurusian appears to have originated. A small number of Dabban-like artefacts have been identified in early Iberomaurusian assemblages at Tamar Hat (laterally retouched blades) and at Taforalt (laterally retouched blades and adzes) but other tools that are typical of this industry are lacking. The adzes at Taforalt also bear little resemblance to the small core axes from Taforalt. It is plausible that environmental conditions in North Africa during MIS-3 led to intermittent isolation and bottle necking of human groups giving rise to a diversification in lithic traditions and the staged appearance of microlithic bladelet technologies from essentially MSA origins.

15An argument for continuity can also be made in West Africa where the MSA persists for much longer than in North Africa (Scerri 2017). Lithic assemblages from a number of sites in West Africa show similarities with the adzes at Taforalt. For example, adzes and bifacial pieces at Ounjougou, Mali (Rasse et al. 2004) have been dated to 35,000 ± 1 ka years ago and several sites in the Falémé Valley in Eastern Senegal have yielded bladelets and segments (Chevrier et al. 2016). However, the sites in both Mali and Senegal are currently poorly dated and it is notable that no Iberomaurusian sites have been identified further north until the area of Agadir in southern Morocco.

Potential transitions within the Northwest African Middle Stone Age

16At present it can be fairly confidently stated that the Northwest African MSA extended from at least 315,000 years until 30,000 years. Current information suggests that it was associated exclusively with Homo sapiens populations. This is supported by finds from Jebel Irhoud of robustly built archaic Homo sapiens, some of the earliest known for the continent. There is still much uncertainty concerning whether the onset of the MSA and other cultural indicators which appear within this period occurred as a result of population replacement or a cultural transition. The latest palaeontological evidence supports the view of multiregional origins of Homo sapiens in Africa (Scerri et al. 2018), which might also apply to multiple appearances of the MSA. Although more analysis is needed to better understand the spatial signature and temporal dimensions of these distinctive markers of the Northwest African MSA, from the available data we can divide this cultural technocomplex into several chronological periods:

300,000 years-145,000 years

17The new discoveries at Jebel Irhoud demonstrate the presence of MSA humans in Morocco from MIS 9 (Hublin et al. 2017; Richter et al. 2017). For much of this early period, the Maghreb was isolated from the rest of Africa by the Sahara and there were no obvious interactions with Europe via the Mediterranean. The technology of the MSA groups displayed a significant Levallois component including mainly flakes, points and various scrapers (Hublin et al. 1987). Although Northwest Africa may have been less accessible to other regions, based on technological similarities a potential dispersal route from Eastern Africa (Brooks et al. 2018) has been suggested (figure 7.2). While there are less convincing connections with the Nile Valley, a possible alternative is that the early Northwest African MSA originated from the local Upper Acheulean.

figure 7.2

145,000 years-60,000 years

18The early part of this period (Marine Isotope Stage 6: 160‑129,000 years) would have been relatively arid in Northwest Africa followed by more humid conditions in MIS5 from 129,000 years to 92,000 years (Drake & Breeze 2016). It is plausible that during the drier stage an important shift began in the MSA with more common use of foliates and small tanged points as well as tiny preferential radial cores similar to the ones from the Asinipodian in southwest France (Dibble & Mcpherron 2006) and dated to around 145,000 or even earlier. The clearest examples of the Aterian with tanged points can be found at Ifri n’Ammar (see above), while post-129,000 years similar occurrences can be documented at many sites including Dar es-Soltan 1, El Harhoura 2, El Mnasra, Contrebandiers and Taforalt. Although we regard pedunculate and foliate points as linked with new hafting techniques we do not consider them to have functioned exclusively as projectile points (Marchand & Aymé 1935). According to a microwear study of pedunculate pieces from Aterian sites in Morocco (Bouzouggar et al. 2007; Bouzouggar & Barton 2012), observations showed few signs of impact fractures. Instead preliminary results at high and low magnification revealed uses in cutting and working a range of organic materials, including both soft and hard materials such as bone.

19It is during humid MIS 5 or probably earlier that evidence for personal ornaments first occurs in Aterian assemblages with pedunculates. Marine shell beads (Tritia (Nassarius) gibbosula) are documented at Contrebandiers as early as around 116,000 years (Steele & Esteban Alvares 2019), while slightly younger examples are known from El Mnasra (El Hajraoui et al. 2012) and at Taforalt where they can be shown to date to around 82,000 years (Bouzouggar et al. 2007). Given the longevity of this material culture we would not be surprised if the history of bead use were to extend further back in time within MIS 6. An interesting point concerning the occurrence of personal ornaments in this part of Africa is that those found in sites located less than 50 km from the coast contain both perforated and unperforated shells, while those located beyond this distance display only perforated examples. This would imply that the shells were mainly collected by people occupying sites close to the coast and were transported to locations further inland as part of an exchange network. The MSA groups stopped making and probably exchanging the beads around 70,000 to 60,000 years (D’Errico et al. 2009) during the drier MIS 4 climatic stage.

20Another relevant and intriguing aspect of the Northwest African MSA during MIS 5 (c. 90,000 years) concerns the appearance of blades and bladelets and this mainly relates to the production of elongated laminar artefacts from bipolar cores (Bouzouggar & Barton 2012). Although few of the MSA lithic collections contain genuine prismatic blade cores of the types described for the Upper Palaeolithic (Roche 1963; Tixier 1963), the MSA groups clearly had the technical skill to make them. Furthermore, in contrast to the widespread existence of small Levallois cores, the presence of blades and laminar flakes seems to vary much more from site to site and may show some degree of regional patterning (figure 7.3). For example, these types are well-represented in sites in eastern and northern Morocco, at Rhafas Cave (Wengler 1993), at El Aliya (Bouzouggar et al. 2002), and probably at Taforalt (Bouzouggar et al. 2007a), but are much less common in the Atlantic coastal zone, at sites like Dar es-Soltan 1 (Ruhlmann 1951; Roche 1956) and at Contrebandiers from layers 10, 9 to 8 (Bouzouggar 1997a, b). It is possible that this may be chronologically or functionally related, since good quality raw materials were abundantly available in both inland and coastal zones.

figure 7.3

21The production of bladelets is a somewhat more complex issue. This debitage type has been recognized in Northwest African MSA and especially the Aterian contexts such as Contrebandiers (Roche 1963, 1976; Bouzouggar 1997), but without true bladelet cores. Nevertheless, at Contrebandiers it was possible to reconstruct part of the chaîne opératoire, showing bladelets with negatives that appeared to cut the posterior crest and initiate a change in debitage direction (Bouzouggar 1997a, b). Examples of bladelet technology have also been recorded at Dar es-Soltan 1 in layer C1 (Ruhlmann 1951) and in eastern Morocco (Wengler 1993), and also now at Taforalt.

22In addition to the stone artefacts, another potentially important signature of the MSA Aterian may be found in the appearance of worked bone and ivory. At Dar es-Soltan 1, two ivory objects, one pointed, the other a small plaquette, were reported from layer I (Ruhlmann 1951). Subsequent examination has shown that the point was probably deliberately shaped and modified (Kaouane 2002, 2013). Another bone tool probably used as a knife was found in a securely context during modern excavations at the cave (Barton et al. 2009) and dated to 90,000 years (Bouzouggar et al. 2018). More examples of bone knives and other bone tools have been described in El Mnasra (Hajraoui 1994; El Hajraoui et al. 2012) (figure 7.4).

figure 7.4

23Present information suggests a continuity of population in Northwest Africa after 300,000 years with further examples of robustly built archaic Homo sapiens occurring at Aterian sites. A point that needs emphasizing, however, is that up until now the majority of such remains have come from western Morocco in a restricted area of the coastal zone between Rabat (Dar es-Soltan1&2) and Temara (Contrebandiers and El Harhoura I caves). This is almost certainly connected with a focus of fieldwork in this area over the past fifty years. It does not detract from the view that human remains in Aterian contexts are relatively scarce and that the understanding of evolutionary relationships between human fossils in the MSA is still problematic and needs to be clarified.

60,000 to 30,000 years

24While the MSA technology still displays Levallois component around 60,000 years, the situation is unclear between 50,000 to 30,000 years. This is a major gap in our knowledge which we are hoping to rectify with new stratigraphic studies that are now in progress.

30,000 to 15,000 years

25The transition between the MSA and LSA occurs between 30,000 and 20,000 years in Northwest Africa, but may not involve a simple and uniform process. At Grotte des Pigeons at Taforalt in Morocco, the youngest age estimates for the latest MSA levels are in the range of c. 29,000 cal BP based on several independent dating techniques (Barton et al. 2016). The layers that overlie the typically MSA sequence are rich in adzes and adze flakes and a simple flake technology (Barton et al. 2013, 2016). This distinctive “adze industry” lacks any obvious techno-typological or functional similarities with the overlying the Iberomaurusian (Barton et al. 2016). The dating of the adze layers is still preliminary but suggests an age range of c. 26,000 and 24,000 cal BP (Barton et al. 2016). Until now, the presence of such technology seems to be restricted to Taforalt (figure 7.5), although a similar example has been found amongst unstratified material at Ghar Cahal in the Tingitane peninsula of Northern Morocco (unpublished data).

figure 7.5

26The Iberomaurusian is a Later Stone Age industry dominated by microlithic backed bladelets recorded at numerous sites from inland and coastal areas of the Maghreb (modern Morocco, Algeria, Tunisia and parts of Libya). The chronology of the Iberomaurusian has been the subject of intensive study in recent years and the dating evidence now supports the emergence of microlithic LSA technology in the Maghreb by at least c. 25,000 cal BP (Hogue & Barton 2016). The oldest date is from Tamar Hat in Algeria, which has a sequence of age determinations beginning at 25,845‑25,270 cal BP and spanning at least 5000 years of Iberomaurusian occupation (Hogue & Barton 2016). The oldest dates for this period in Morocco are at Taforalt, where Iberomaurusian deposits have been dated by a sequence of more than 50 radiocarbon dates on bone and wood charcoal covering a time-span from 23,459 to 12,548 cal BP (Bouzouggar et al. 2008; Barton et al. 2013; Barton et al. in press).

27The Iberomaurusian marks a major break with preceding flake and blade technologies in this region raising the question of whether the appearance of this industry coincided with the arrival of new human group in NW Africa or whether it emerged and diffused independently within the region (Bouzouggar et al. 2008). Genomic evidence can provide an independent perspective on the chronology and origins of the Iberomaurusian. Mitochondrial consensus sequences of seven individuals buried at Taforalt belong to the U6a (n = 6) and M1 (n = 1) haplogroups and the diversification of both haplogroups is estimated at c. 24,000 BP (Van de Loosdrecht et al. 2018). This coincides remarkably well with the age of the earliest known Iberomaurusian and may imply dispersal into the region by people with a distinctive microlithic backed bladelet tradition at that time.

28The evidence for human presence at sites such as Taforalt in Morocco and Tamar Hat in the early Iberomaurusian is relatively ephemeral and this together with the absence of grindstones and other non-portable components in the toolkit and a lack identified burials, suggests a relatively mobile lifestyle during this phase. With the possible exception of an isolated skull from Taza in Algeria (Meier et al. 2003) there are no human remains unambiguously associated with this period.

15,000‑12,700 years

29Several sites in eastern Morocco document a marked sedimentary transition at c. 16‑15kcal BP marking a separation between the early and later phases of the Iberomaurusian. At Taforalt there is a major stratigraphic division between gradually forming Yellow Series deposits and the overlying Grey Series deposits which comprise a rapidly forming ashy midden, rich in occupational debris including charred plant remains. The earliest calibrated age for the Grey Series is 15,204‑14,261 cal BP. Similar sedimentary transitions are recorded in archaeological deposits at Ifri El Baroud and at Ifri n’Ammar at approximately 16,000 cal BP (Moser 2003; Nami 2007; Linstadter 2012).

30The sedimentary transition between the Yellow and Grey Series at Taforalt coincides with a change in subsistence behaviour including increased exploitation of wild plant foods and land snails, and other behaviours that point to increased sedentism (Taylor et al. 2011; Humphrey et al. 2014) (figure 7.6). The presence of grinding tools used to process pigments and edible plants, a diversity of bone tools and likely production of basketry provide further evidence for an expansion of subsistence activities. This onset of formation of the Grey Series deposits broadly coincided with the first use of the site for human burials in a collective burial area that was spatially demarcated from areas of the cave used for occupational activities (Humphrey et al. 2014).

figure 7.6

31The presence of human burials in archaeological deposits from the later phase of the Iberomaurusian is relatively common at sites in Morocco and Algeria including important collections from Taforalt (Ferembach 1962), Afalou (Arambourg et al. 1934; Hachi 1996) and Columnata (Chamla 1970). Funerary treatment is complex and highly variable within and between sites (Humphrey et al. 2012; Aoudia-Chouakri 2013). Recurrent features include seated burials, the use of ochre, inclusion of horn cores and other objects, closure of the burial with a stone and the representation of all age groups including infants (Humphrey et al. 2019).

32One of the distinctive characteristics of the later Iberomaurusian population is modification of the body during life, including deliberate removal (evulsion) of one or both upper central incisors (Humphrey & Bocaege 2008; De Groote & Humphrey 2016) and possible trepanation of the cranial vault (Dastugue 1962). The Iberomaurusians were also characterised by poor oral health, including a high prevalence of tooth decay, periodontal disease and rapid tooth wear (Humphrey et al. 2014; De Groote et al. 2018).

Concluding discussion

33In Northwest Africa, connections between the MSA/Aterian and a local Mousterian have been suggested (Hahn 1984) but are difficult to substantiate. Another idea is that the Aterian originated in Nubia (Caton-Thompson 1946b) but this hypothesis has been challenged by recent and updated technological and stratigraphic studies in the same region (Van Peer 1998) and it now appears that lithic technologies with tanged points are absent in Nubia. In other areas such as Egypt the exact relationship between the Khargan and the Aterian is even now unclear (Kleindienst 1999) since the archaeological finds derive largely from unstratified surface collections.

34A related theme in our project concerns the origin of the LSA Iberomaurusian. Various source areas have been proposed for this cultural grouping. Based upon technological similarities a potential dispersal route from Northwest Africa was put forward many years ago. However, at an early stage this was challenged because the ¹⁴C dating evidence from layer XII at Haua Fteah (McBurney 1967) showed that the Iberomaurusian lithic technology was younger at this site than at locations in Algeria and Morocco. There are even less convincing connections with the Nile Valley (Lubell 1974).

35In Northwest Africa, the most important collections of Iberomaurusian human remains are from Taforalt (Ferembach 1962), Afalou (Chamla 1968) and Columnata (Chamla 1970) and some similarities have been noted with human fossils in Egypt and Libya (Anderson 1968). While it seems that the human types associated with the Iberomaurusian are all modern Homo sapiens, there are no obvious reasons for seeking links with Egypt and Sudan because the lithic technologies from Wadi Halfa, Gebel Silsila and Deir el Fakhouri are quite different from those of Morocco and Algeria (Camps 1974; Vermeersch 1992). No less controversial are the proposed cultural links with Southern Europe during the Last Maximum Glaciation (Vallois 1969; Ferembach 1976a; Debénath et al. 1986; Debénath 2003) particularly since the palaeoenvironmental and chronological sources on which these are based have not yet been adequately established (Dutour 1991).

Acknowledgements

36We thank Alice Leplongeon, Mae Goder-Goldberger and David Pleurdeau for inviting us to the workshop “Not just a corridor. Human occupations in the Nile Valley and the neighbouring regions between 75,000 and 15,000 years ago”, the National Institute of Archaeological Sciences and Heritage (INSAP, Morocco). Special thanks to Simon Collcutt, Jacob Morales, Alison Roberts, Peter Berridge, Paul Berridge, Elaine Turner, Ali Freyne, Joshua Hogue and Victoria Taylor. We also thank Calleva Foundation, UK. The authors thank two anonymous reviewers for their contribution to the peer review of this work.