Introduction

1Human groups associated with Aterian techno-complexes were thought to have almost never extended to the Nile Valley (e.g. Kleindienst 1998). Wadi Kubbaniya in Egypt (figure 4.1) seemed to be the only site in the valley representing this industry (Wendorf & Schild 1992). If this may still hold true with regard to the Lower, Egyptian, Nile Valley, the Sudanese stretch of the Middle and Upper Nile calls for a reconsideration of this assumption.

2The recent extensive publication by Carlson (Carlson 2015) of the collections from Khor Abu Anga and Magendohli, excavated in the 1960s by the University of Colorado, provides consistent evidence on Aterian occurrences in the Sudanese Nile Valley. Other finds and personal research in the Amara West district (Garcea 2015)1 and at El-Multaga-Abu Dom (Garcea 2003)2 corroborate the presence of Aterian materials in the Sudanese Nile Valley. Furthermore, a multitude of Aterian sites from the Wadi Shaw, the Laqiya valley and the Wadi Howar, in western Sudan, support a link between the Eastern Sahara and the Sudanese Nile Valley.

3Another highly debated topic regards the northern vs. southern route of the Out of Africa dispersals of anatomically modern humans (AMH). Scholars focused on either North Africa or East Africa tend to suggest alternative, dichotomous views and see their area of interest as “the” privileged launchpad of the Out of Africa migrations (cf. among others, Lahr & Foley 1994; Van Peer 1998; Quintana-Murci et al. 1999; Petraglia & Alsharekh 2003; Drake et al. 2008, 2011; Osborne et al. 2008; Van Peer et al. 2010; Armitage et al. 2011; Beyin 2011; Soares et al. 2012; Groucutt et al. 2015b; Pagani et al. 2015). The Sudanese Nile Valley happens to be in a geographical position that rules out this antithesis and its Middle Stone Age (MSA) evidence offers substantial contributions to both the northern and the southern routes. It demonstrates that early anatomically modern human populations in Sudan likely brought their cultural influences in both directions, to the Levant, through the northern route, and to the Arabian Peninsula, through the southern route.

4This paper aims at offering a new perspective on the Sudanese Nile Valley by discussing these two extensively debated topics: the eastern geographical extent of the Aterian and Sudan’s contributions to the northern and southern Out of Africa routes of early modern humans.

figure 4.1

The ultimate frontiers of the Aterian

5Aterian assemblages occur over a vast area of North Africa, comprised between the Mediterranean Sea, the southern fringes of the Sahara, the Atlantic coast and the Nile Valley. They are mostly spread in desert areas, but also reach the African Mediterranean coast and the Middle and Upper Nile Valley. Chronologically, they span over more than 100,000 years, being dated from 145,000 BP at Ifri n’Ammar in Morocco (Richter et al. 1998) until about 30‑40,000 BP. Their youngest dates are > 40,000 BP at Kharga Oasis in the Egyptian Eastern Sahara (Kleindienst 2006), 43-44,000 BP in the Jebel Gharbi in Libya (Garcea 2004), and in Morocco at Mugharet el ‘Aliya (39,000 BP) (Wrinn & Rink 2003), Taforalt (37,000 BP) (Barton et al. 2013), and Wadi Noun (30,000 BP) (Weisrock et al. 2006). Different chronological, technological and typological Aterian horizons have been identified even within the same country, such as Libya (Garcea 2006). Poor temporal resolution and leopard skin site distribution do not allow delineating a clear chrono-geographical variability beside general north-western vs. north-eastern features (Scerri 2013a; Garcea 2016). Nevertheless, general common basic traits may be still outlined.

6Apart from tanged tools, a combination of distinctive features characterise the Aterian, including butt thinning and ventral retouching (Caton-Thompson 1946a), bulbar basal thinning and bifacial retouching (Kleindienst 1998), ovoid flakes, small centripetal Levallois and discoidal cores, and bifacial foliate or lanceolate points (Bouzouggar 1997a; Bouzouggar & Barton 2012; Spinapolice & Garcea 2013, 2014). In addition, notched tools and “pseudo-tanged” tools should be considered as equally typical as real tanged tools (Garcea in press). Pseudo-tangs had not been distinguished by J. Tixier in his classification of tanged tools, but he recognised that not all tangs are made by retouching their four sides —left and right margins, ventral and dorsal faces (Tixier 1958‑1959: 133‑135)— and that some of them may have been only retouched on three sides (Tixier 1958‑1959: figs 6.2, 9.2) or two sides, ventrally (Tixier 1958‑1959: fig. 2.1), dorsally (Tixier 1958‑1959: figs 2.2, 15.1) or alternately (Tixier 1958‑1959: fig. 4.1).

7Recently, experimental and functional analysis on use-wear of Aterian tanged and notched tools from the Jebel Gharbi, in north-western Libya has demonstrated that notching or shouldering was not always intended for hafting, but could also be related to longitudinal or transversal actions practiced for either cutting or scraping (Massussi & Lemorini 2004‑2005; Falzetti et al. 2017). In this case, two facing notches on one end may form pseudo-tangs and could be used as the active part of the tool. Pseudo-tangs are produced by two juxtaposed notches on one end of the tool, which create an apparently similar, but technologically and functionally distinct morphology with respect to real tangs (Falzetti et al. 2017). The latter are made by invasive or covering bifacial retouching with a soft hammer and are aimed at hafting (Tomasso & Rots 2018). Their blanks often show burin-like and transverse snap fractures. By contrast, pseudo-tangs are usually made by stepped or, more rarely, scalar retouching to be hand-held and show clear cutting and scraping usewear (Falzetti et al. 2017). Both types of tools were part of Aterian tool-kits. However, as pseudo-tanged tools may look like tanged tools at first sight, it is likely that they have often been classified as the latter. Although the Aterian should not be identified with tanged tools, these pieces, when they occur and whether they have true or pseudo-tangs, are unavoidably associated with Aterian complexes (e.g. Iovita 2011; Scerri 2013a). In the Jebel Gharbi, notches and pseudo-tangs are a prominent component of the Aterian tool-kits and were more frequent than real tangs actually made for hafting (Garcea in press). These considerations call for a revision of the potential biases in classifying tools as Aterian tangs, when they come from North Africa and the Sahara and vice-versa as notched tools, when they come from the (Sudanese) Nile Valley. Although these two types of tools are clearly different, geographical biases affected their attributions.

8Currently numerous sites and assemblages with tanged and pseudo-tanged tools have been recovered in various parts of Sudan, following the first tanged tool recorded in the country at Arkin 5, near Wadi Halfa (figure 4.2) (Chmielewski 1968: figs 3a-3b).

figure 4.2

Khor Abu Anga

9The site of Khor Abu Anga (figure 4.1) was found along a dry gully on the western Nile bank that flowed into the main river in the vicinity of the present city of Omdurman, 1 km below the confluence of the Blue and White Nile. It was known since the 1940s and over the years attracted the attention of various scholars, who collected, mostly on the surface, thousands of artefacts, but rarely documented them. A.J. Arkell (Arkell 1949) recovered over 1,100 artefacts, of which only 185 from the stratigraphic deposit; F. Hibben excavated several test pits, but could not study the assemblages due to complicated organisational reasons (Carlson 2015); the Guichards (Guichard & Guichard 1965) also collected surface artefacts, but did not publish them in detail. The area must have been extraordinarily rich in artefacts because, years later, R.L. Carlson (Carlson 2015) was still able to find surface material and, most importantly, preserved stratigraphic sequences in the trenches that he excavated, including Late Acheulean, Sangoan, and Lupemban complexes.

10The stone assemblages from Khor Abu Anga suggested intriguing speculations on the origin of tanged tools and their relations with the Aterian. Arkell (Arkell 1949) considered that handaxes with thinned and narrowed butts could foreshadow tanged tools and that they could possibly be the lowest common denominator of the Aterian and the Stillbay techno-complexes. Although it is now clear that there is no direct relation between Acheulean handaxes and Aterian and Stillbay tools, butt thinning and narrowing are still considered as distinctive Aterian features and were correctly related to the Aterian. In their preliminary publication, also Carlson and colleagues (Carlson et al. 1967‑1968) mentioned tanged tools in the Abu Anga tool-kit and indicated an affiliation with the Aterian. In his final publication, Carlson (Carlson 2015) described tanged tools (figure 4.3) in what he assigned to Late Sangoan and Lupemban levels. These tools are usually made on blunt blanks, with unilateral or, mostly, bilateral notches that form a tang on one end of the tool. Whether they were hafted or not, their morphology shows that they could hardly be projectile tips.

11The increase of tanged scrapers observed in the Lupemban assemblage from Khor Abu Anga, combined with an increase of notched tools, may be associated with comparable functional purposes of scraping of certain “tanged” tools. These tools represent “pseudo-tangs” and should not be confused with real hafting tangs. As a matter of fact, a parallel increase of notches together with tanged tools commonly anticipates a typical phenomenon of Aterian techno-complexes and has been clearly documented in the Jebel Gharbi, in north-western Libya (Falzetti et al. 2017).

figure 4.3

Magendohli

12Beside the tanged tools from Khor Abu Anga, truly Aterian techno-complexes have been discovered in Sudan. A well-representative site is Magendohli in North Saras, south of the Second Cataract (figure 4.4: 1). It is located on the top of an outcrop about 70 meters above the present western Nile bank. Tanged points (figure 4.4: 1‑2) and tanged sidescrapers (figure 4.4: 3‑9) are the most common tools and were associated with a prevalence of unfinished pieces and unstruck cores indicative of a quarry and workshop site. The striking abundance of tanged tools persuaded Carlson to assign this assemblage to the Aterian (Carlson et al. 1967‑1968; Carlson 2015). Among other tools, notches and “single- and double-notch gravers” are frequent. The bifacial technology is extremely rare, whereas Levallois products are frequent.

figure 4.4

Western Sudan

13Aterian assemblages particularly abound in western Sudan. Gabriel and Kröpelin (Gabriel & Kröpelin 1997: 344) recounted that, when Sandford visited western Sudan in the 1930s, he reported “countless thousand implements” of “exclusively later Middle Palaeolithic type” around Laqiya Arbain and even supposed human population “overcrowding” this area in the late MSA (Sandford 1933: 220).These tool-kits consist of numerous points (figure 4.5: 1‑3), bifacial foliates (figure 4.5: 4‑10) and tanged tools (figure 4.6) that were later collected at several sites in the Wadi Howar, Wadi Shaw, and Laqiya valley (Idris 1994; Gabriel & Kröpelin 1997; Garcea personal observations).

14During the Pleistocene the Wadi Howar was a very important hydrological system between Jebel Marra and the Nile, corroborating human interactions between the Nile and the Sahara (Gabriel & Kröpelin 1997). Also during the Holocene, the Wadi Howar and the Wadi el Melik, the two major western Nile tributaries, linked western Sudan to the Nile Valley (Jesse 2004), creating a west-east interconnected region. Numerous palaeolakes, namely Selima, Oyo, the Western Nubian Palaeolake, El Atrun, and Gureinat, are well-documented with regard to the early Holocene and were probably also intermittently active in the Upper Pleistocene being fed by groundwater and local monsoonal rainfall (Ritchie & Haynes 1987; Ritchie 1994; Hoelzmann et al. 2001, 2010; Kröpelin et al. 2008). Westerly air masses from the Atlantic could also be responsible for precipitation that replenished groundwater and formed spring deposits (Williams 2019: 243). The considerable population attested to by the “thousand” MSA, including Aterian, lithics could not have certainly survived without important water outcrops and drainage systems.

figure 4.5

figure 4.6

Additional occurrences

Northern Sudan

15The district of Amara West is located on the left (northern) side of the present Nile Valley in front of the modern village of Abri, in northern Sudan (figure 4.1). It revealed numerous prehistoric sites, which have been surveyed and test excavated on various occasions. The area was first surveyed by A. Vila (Vila 1977), who described five “Late Palaeolithic” sites (2-R-15, 2-R-20, 2-R-21, 2-R-71, 2-R-74) and mapped two other ones (2-S-17 and 2-S-22) in the easternmost part of the district (Vila 1977: 14).

16All these sites have been relocated in the 2010s and assigned to the MSA (Garcea 2015). Furthermore, another previously unknown site was discovered on a higher terrace and, according to the Archaeological Map of Sudan followed by Vila, was given number 2-R-77. This site appeared as the best preserved MSA settlement in the area, possibly spanning the entire MSA (Smith, pers. comm.). Quartz and chert are the most frequent raw materials and quartzite and sandstone were also used. The assemblage includes Levallois cores of large sizes and centripetal cores with a predominant flake component, including elongated flakes. The tool-kit consists of simple and denticulated Levallois flakes, cortical bipolar flakes, notched flakes and flakes with alternate backed retouches. Among sidescrapers, one was tanged (figure 4.7: 1). Another tanged, bifacially retouched sidescraper on a Levallois blade (figure 4.7: 2) was located nearby. In the same district, Vila (Vila 1977) recorded other tanged pieces at Site S5, together with Levallois products. This site was interpreted as a quarry and workshop.

figure 4.7

Central Sudan

17The Multaga-Abu Dom area (figure 4.1) is located along the left bank of the Nile Valley, at the second bend of the river, between the modern towns of Ed-Debba to the west and Korti to the east in central Sudan. It is located on the opposite Nile bank, in front of the MSA sites of the Affad Basin in the Southern Dongola Reach (Osypiński & Osypińska 2016; Osypiński et al. 2016). Salvage explorations were part of the programme of the Merowe Dam Archaeological Salvage Project at the Fourth Cataract of the River Nile carried out in the 1990s. MSA sites were identified in the Pleistocene deposit between two wadis perpendicular to the current Nile that cut the southern terrace related to an ancient Nile course and located above the present river, about 5 km south of the present Nile bank (Garcea 2003). They exhibit MSA artefacts eroded out on the surface and partly preserved on the remnants of the Pleistocene deposit.

18After a review of the sites and an assessment of their archaeological potential, Sites 93, 97, 106, and 108 were selected for 2-sq.-m test excavations, three of them being located in the south-eastern corner of the explored Multaga-Abu Dom area (figure 4.8). They were all located near outcrops of chert. Sites 93 (extension: 17 x 15 m) and 97 (extension: 28 x 38 m) were the best preserved ones. Site 93 was located at the south-western margin of the Pleistocene deposit, whereas Site 97 was in the middle and was the richest in terms of artefactual material. Thicknesses of the archaeological deposits at Sites 93 and 97 were 66 cm and 68 cm, respectively. Both sites revealed an earlier deposit below level 3, which was separated from the upper deposit in Level 2 by erosion marked by a thin aeolian sandy level (figures 4.9, 4.10).

figure 4.8

figure 4.9

figure 4.10

19Various types of chert are the most frequent raw materials; fossil wood is the second most common rock, whereas quartz, quartzite and sandstone are rare. Artefacts show fresh preservation condition and come from in situ occupation levels (Moeyersons 2002). The stratigraphic deposit and artefactual material indicate at least two distinct human occupations. Single platform cores prevail in the upper deposit (Levels 1-2), where the Nubian technology is also represented, but only with a single atypical Nubian core at only one site (Site 93) (figure 4.11). By contrast, Levallois cores (figure 4.12: 1), which are generally rare, only occur below Level 2; Like Levallois cores, also Levallois flakes are few and are more common below Level 2. Retouched tools are common (figures 4.13, 4.14) with a prevalence of sidescrapers (figure 4.12: 2) and notched and denticulated flakes (figure 4.12: 3). Endscrapers are more common in the upper levels (figure 4.12: 4‑5). Pseudo-tanged tools (figure 4.12: 6‑7) were recorded at Site 97 (table 4.1). Although these assemblages cannot be assigned to a particular cultural unit, Aterian influences are likely, possibly originating from the Sahara through western Sudan. In fact, the Wadi Howar joined the Nile not far from the Multaga-Abu Dom and may have been a communication route.

20The nearby site of Affad 23, on the opposite bank of the Nile River, shows several similarities with Multaga-Abu Dom. They include the preference for chert as raw material, the prevalence of denticulates and sidescrapers, the lack of bifacial tools and the rarity of Levallois flakes and blade products, even though Levallois cores are more common at Affad 23 (Osypiński & Osypińska 2016; Osypiński et al. 2016). This site was previously dated to 16,000 and 15,000 BP (Osypiński & Osypińska 2016), but revised calculation methods have provided ages closer to 50,000 BP (Osypiński & Osypińska in press).

figure 4.11

figure 4.12

figure 4.13

figure 4.14

Sudan’s contributions to the northern and southern Out of Africa routes of early anatomically modern humans

21At least two major chronologically distinct events of AMH migration Out of Africa had been distinguished, one occurring during MIS 5 (130,000‑80,000 BP), the other taking place at the beginning of MIS 3, after 50,000 BP (Garcea 2010b, 2012; Shea 2010). The MIS-5 episode was defined “Out of Africa 2a” and the MIS-3 event “Out of Africa 2b” (Garcea 2010b, 2012). The population expansion that occurred in North Africa at the beginning of MIS 5 with increased humidity may have contributed to the Out of Africa 2a event. During MIS 3, climatic conditions were not as supportive for humans in North Africa, whereas they were more favourable in the Levant and may have triggered the Out of Africa 2b migration.

22Various interpretations exist on the cultural units associated with human migrations. Beside the Aterian, several other Upper Pleistocene cultural units contemporaneously existed in Sudan: Nubian Middle Palaeolithic (Guichard & Guichard 1965, 1968), Denticulate Mousterian (Marks 1968a), two types of Nubian Mousterian (Marks 1968a), Nubian Middle Stone Age (Wendorf & Schild 1992), and Nubian Complex (Van Peer 1998, 2016), establishing overly distinguished techno-complexes. In addition to groups carrying Aterian traditions and moving across the Sahara and easterly along the African Mediterranean coast (Garcea 2010b, 2012; Spinapolice & Garcea 2013, 2014), populations associated with the Nubian Complex are also likely to be among those who participated in dispersals along the Nile corridor (Van Peer 1998, 2016).

23At present there are no radiometric dates from the Aterian sites in Sudan, but dates from other regions have indicated that Aterian tool-makers, spanning from MIS 6 (190,000‑125,000 BP) to MIS 3 (58,000‑25,000 BP) could have taken part in both Out of Africa dispersals, 2a and 2b (Garcea 2010b, 2012). The Nubian Complex is slightly later and lasted less time than the Aterian: it emerged in the beginning of the Last Interglacial (MIS 5e, 125,000‑110,000 BP) and reached its best representation in the final MIS 5 (5a, 80,000‑70,000 BP). In Africa, the Nubian Complex extends from the Eastern Sahara to the Red Sea Hills and the Sinai. Nubian assemblages have been recently also found on the Negev highlands, in the southern Levant (Goder-Goldberg et al. 2016), where human dispersal could have been supported by the last Negev Humid Period, dated to 142,000‑109,000 BP (Goder-Goldberg et al. 2016) and is anthropologically documented by the Skhul and Qafzeh fossil remains, dated to 120,000‑90,000 BP (Grün et al. 2005; Grün 2006). Considering that no Nubian Complex sites are younger than MIS 5, its diffusion to the southern Levant must be associated to the Out of Africa-2a event. For the time being, the southern Levant is the northernmost frontier of the Nubian Complex. The most representative technological links between the southern Levant and the Nubian Complex come from two Sudanese sites near Dibeira East (1035 and 1038) that had been previously assigned to the Nubian Mousterian (Marks 1968a) and then to the Late Nubian Complex (Van Peer 1998). Their striking similarities include a lower frequency of Nubian cores over Levallois cores and no true Levallois points (Goder-Goldberg et al. 2016).

24The Nubian Complex does not only extend out of Africa beyond the northern route, but also occurs beyond the southern route, reaching the Arabian Peninsula. Hundreds of Nubian Complex occurrences have been recorded in Saudi Arabia, Oman, and Yemen (Rose & Marks 2014; Hilbert et al. 2017). Rose and Marks (Rose & Marks 2014: 66) describe this industrial tradition as “the most abundant and distinct type of Middle Palaeolithic on the Arabian Peninsula”. They also suggest that this techno-complex spread from the Arabian Peninsula along the Asian eastern coast of the Red Sea and eventually reached the southern Levant. If this is confirmed, a separate wave of AMH —in this case of Asian origin— should be envisaged following the Out of Africa-2a dispersal attested to by the Skhul-Qafzeh hominins.

25Finally, the new discovery, again in the southern Levant, of a hemimaxilla of an anatomically modern human earlier than the Skhul and Qafzeh fossils at Misliya Cave, dated between 194,000 and 177,000 BP (Hershkovitz et al. 2018), pushes back the timing of the earliest migration of members of the modern human clade out of Africa. Hershkovitz et al. have deduced that this earliest AMH dispersal may have occurred before 200,000 BP (Hershkovitz et al. 2018). Stringer and Galway-Witham have ruled out direct local evolution of the Skhul-Qafzeh group from the Misliya population because the time span elapsed between the two interglacials was too long (Stringer & Galway-Witham 2018). They have advocated another independent Out of Africa wave, in addition to the two successive dispersals of MIS 5 and MIS 3, which also exhibit very distinct features.

26Considering the age of the Misliya fossil, it is possible that the makers of Sangoan industries may have also had a role in the earliest Out of Africa movement. In fact, according to Van Peer (Van Peer 2016), the Sangoan in Sudan results from the arrival from East Africa of new —anatomically modern— populations during MIS 7 (245,000‑190,000 BP) who did not replace the local Acheulean groups and coexisted with them. Sudan yielded one of the rare African sites, 8-B-11 on Sai Island in the Nile River, about 20 km south of Amara West, with evidence for an Early Stone Age/Middle Stone Age interstratification (Van Peer et al. 2003). The Late Acheulean and Early MSA, Sangoan, industries from this site were presumably produced by biologically different human groups who intermittently and alternatively occupied the site: Homo heidelbergensis and early Homo sapiens. Excavations brought to light two stratified Sangoan occupational levels, Lower and Middle, the former dating from 220,000 BP, which is synchronous with the Misliya hominin, and the latter being dated c. 180,000 BP. The Sangoan industries were distinguished from the Acheulean productions for the frequency of hafted core-axes and the lack of hand-held handaxes. In general, the groups associated with Sangoan industries at 8-B-11 show complex behavioural features, including functional specialisation (Van Peer et al. 2004). Furthermore, a Late MSA, Lupemban, level was dated to 152,000 BP (Van Peer et al. 2003; Rots et al. 2011). Nubian cores occur in both the Sangoan and Lupemban levels, where they increase. This site was considered as a specialised location for the production of quartz core-axes and retooling of core-axes of other raw materials, such as quartzite. Finally, the evidence for activities of exploitation and processing of mineral pigments and vegetal materials in the Sangoan levels further supports the complex behaviour of the occupants of the site (Van Peer et al. 2003).

Discussion

27The long colonisation process of Homo sapiens corresponds to the favourable paleoecological conditions of MIS 7-Misliya (245,000‑190,000 BP), MIS 5-Out of Africa 2a (130,000‑70,000 BP), and MIS 3-Out of Africa 2b (58,000‑25,000 BP), which were separated by abrupt and harsh climatic intervals. Nevertheless, these groupings by no means refer to single events: dispersals out of Africa are more likely to have occurred numerous times by small human communities.

28Wetter conditions created vegetated, resource-rich corridors that enabled migration paths (Drake et al. 2013; Garcea 2016; Hublin et al. 2017). During MIS 7, AMH spread from their refugia in East Africa (Zipkin et al. 2017; Brooks et al. 2018) into less favourable areas such as North Africa. Connections with East Africa continued to be supported by the White Nile Palaeolake, which existed during the Last Interglacial period where the Lower White Nile currently flows across the border between Sudan and South Sudan (Barrows et al. 2014; Williams et al. 2010; Williams 2019). The White Nile Palaeolake reached its maximum extent by about 110,000 BP and was 650-km long and 80-km wide, extending over a surface of 45,000 km², which would be like the present largest freshwater lakes in the world. This megalake was created during very high peak floods in the Blue Nile with strong river flow that caused water in the White Nile to back up for hundreds of kilometres and a narrowing of the White Nile channel towards the northern margin of the lake (north of 14°N) (Williams 2019).

29Finally, south-eastward connections were recorded in the southern part of the Sudanese Red Sea coast, around the Tagdara hills, inland from the Gulf of Agig, and on the slopes of the Grar basin Sea (figure 4.1), where Mode 3 assemblages made with the Levallois and Nubian techniques were found (Beyin et al. 2019b). Furthermore, the site by the Grar basin also included a foliate point featuring an unfinished tang (Beyin et al. 2019b: fig. 10.F).

30In consideration of the current remarkable significance of Aterian assemblages in Sudan from Magendohli, the Wadi Shaw, the Laqiya valley and the Wadi Howar, the occurrences of tanged tools that previously seemed to represent single, insignificant finds, such as the specimen from Arkin 5, should be viewed differently. Beside Amara West and Multaga-Abu Dom, various other finds came to light, including a few tanged pieces together with Levallois and other MSA products recovered at El Hamra in the Ga’ab depression, northwest of Dongola (Tahir & Nassr 2015), and a broken tanged Aterian point found in Tergis (Osypiński 2003: fig. MTC II/4:e; Osypiński 2012). The combination of all these, previously neglected, Aterian sites and occurrences in Sudan calls for a reconsideration of the geographical distribution of this techno-complex, which evidently extended as far as the Middle and Upper Nile Valley.

Final remarks

31Thanks to Sudan’s geographical position, its early anatomically modern human populations offer substantial contributions to the Out of Africa dispersals since their earliest representation attested to by the Misliya hominin in the southern Levant, possibly related to the Sangoan. Later, during the Out of Africa 2a, groups settled in the Sudanese region seem to have participated in both the northern and the southern routes. The groups associated with the Nubian Complex in particular brought their cultural influences to the Levant as well as the Arabian Peninsula. Technological links have been pointed out between the northernmost frontier of the Nubian Complex in the southern Levant the Sudanese sites 1035 and 1038 near Dibeira East, suggesting a shared cultural background over such an extended area. Furthermore, Sudan does not only show northern cultural links, but also south-eastern ones along the Red Sea coast and inland, suggesting another viable route that possibly led across the Bab el-Mandab to the Arabian Peninsula.

32Whether the Aterian is part of the Nubian Complex (Van Peer & Vermeersch 2007), or is a separate, independent techno-complex and the two industries were produced by different populations (Scerri 2013a), the Aterian finds in Sudan demonstrate that this techno-complex was undeniably present in the Middle and Upper Nile Valley as well as its hinterland. Admittedly, tanged tools have often been overvalued and the number of Aterian sites may be overestimated in the Sahara. Yet, the opposing attitude of disregarding them in Sudan appears to be unreasonable. The recurrent number of tanged tools, combined by other indicators (i.e., thinned and narrowed ends, elongated flakes, bifacial tools, notched tools, pseudo-tanged tools) in northern, central, and western Sudan, substantiates west-east movements from the Sahara to the Middle and Upper Nile Valley by Aterian groups. Saharan groups may have been attracted to the Sudanese territory because it was able to offer favourable conditions not only along the Nile Valley, but also in the hinterland, where the Nile River basin is more extended than in Egypt and was recurrently able to benefit from groundwater and monsoonal rainfall.

Acknowledgments

33I sincerely thank Alice Leplongeon, Mae Goder-Goldberger and David Pleurdeau for the impeccable organisation and exquisite hospitality during the workshop “Not just a corridor, human occupation of the Nile Valley and neighbouring regions between 75,000 and 15,000 years ago” at the Institute of Human Palaeontology in Paris. I am most grateful to Francis Geus, then director of the French Archaeological Section of the National Museum in Khartoum, for inviting me to join his team of salvage survey and excavation in the Multaga-Abu Dom area in 2002, and to Neal Spencer of the British Museum, director of the Amara West Research Project, for inviting me to conduct fieldwork on the prehistoric sites in the area in 2011 and 2013. I am equally indebted to the National Corporation for Antiquities and Museums for granting me permission to work in both areas. I owe a big debt to Jennifer R. Smith, who joined my team in Amara West and provided essential interpretations of the geoarchaeology of the area we surveyed together. I also give warm thanks to Martin Williams for his wise and helpful comments on an earlier version of this paper. My gratitude finally goes to Bruna M. Andreoni of the University of Cassino and Southern Latium for skilfully drawing and editing the illustrations. This work benefited from the review by two anonymous readers whose comments strengthened and produced a more effective text. Funds for fieldwork were granted by the National Geographic Society —Committee for Research and Exploration (Grant #8715‑09), the Department of Letters and Philosophy of the University of Cassino and Southern Latium, Italy and the French Archaeological Section of the National Museum in Khartoum.