Disclosing Archaeological Complexity of the Khartoum Mesolithic: New Data at the Site and Regional Level
- First Online:
- Cite this article as:
- Salvatori, S. Afr Archaeol Rev (2012) 29: 399. doi:10.1007/s10437-012-9119-7
In the last decade, prehistoric archaeology in central Sudan and Nubia has been characterised by a regional approach and the use of proper stratigraphic methods in excavation strategies. This has also led to the discovery of well-preserved stratified Mesolithic deposits at sites affected by heavy post-depositional anthropogenic disturbances. For the first time, 65 years after the excavation of the Khartoum Hospital site, it is possible to perceive and describe material production variability, settlement pattern discontinuity and/or continuity. It has now become possible to face the problem of social complexity of hunter–gatherer–fisher groups along the middle Nile Valley, a cultural phase which lasted for at least 3,000 years. The new data suggest a reworking of the static picture of this culture, as emerging from the scientific literature, in order to move the debate in a new and more productive direction. This contribution will only be a first step, based mainly on freshly collected pottery assemblages, towards a new approach to the Khartoum Mesolithic pottery culture. It also begins a critical appraisal of the methodological and theoretical faults that hampered a correct evaluation of the data collected from previous surveys and excavations in central Sudan. Incidentally, it will help to revitalise the study of pottery–bearing hunter–gatherer–fisher societies, and supply fresh data to the worldwide anthropological debate on this complex and yet unresolved topic.
KeywordsCentral SudanKhartoum MesolithicMesolithic potterySettlement patternChronology
Durant ces dix dernières années, l’archéologie préhistorique au Soudan central et en Nubie a été caractérisée par une approche régionale de la recherche et, en même temps, par l’emploi d’une méthode stratigraphique dans les stratégies d’exploration des sites archéologiques. Ces nouvelles stratégies ont conduit à la découverte de dépôts archéologiques stratifiés et bien conservés, même dans des sites qui subissent de fortes transformations post-dépositionnelles dues à l’activité humaine. Pour la première fois, soixante-cinq ans après la fouille du site de «Khartoum Hospital», il est possible de percevoir et de décrire aussi bien la variabilité dans les productions matérielles, que la discontinuité et/ou la continuité dans les modalités d’usage et de fréquentation des sites archéologiques préhistoriques. Il est désormais possible d’aborder le problème de la complexité sociale des groupes de chasseurs– cueilleurs–pêcheurs le long de la vallée du Nil moyen, une phase culturelle qui a duré pendant au moins trois mille ans. Les nouvelles données recueillies dans des sites au sud de Khartoum suggèrent une révision radicale de l’image du Mésolithique de Khartoum, telle qu’elle est véhiculée dans la littérature scientifique, pour transférer le débat dans une direction nouvelle et plus productive. Cet article ne sera qu’une première étape, principalement fondée sur les données récentes apportées par la céramique, en abordant le Mésolithique de Khartoum d’une façon nouvelle. De plus, sera ici fournie une évaluation critique de l’approche méthodologique et théorétique suivie dans l’archéologie du Mésolithique au Soudan central, et son incapacité à décrire et à expliquer convenablement les changements observés. Incidemment, nous espérons aider à réactualiser l’étude des chasseurs–cueilleurs–pêcheurs producteurs de céramique et à fournir des données nouvelles pour le débat en anthropologie sur ce problème complexe et toujours entier.
The Khartoum Mesolithic culture was discovered and initially described by A.J. Arkell with the publication of the Khartoum Hospital settlement excavation (Arkell 1947, 1949). The site, although not stratified (Arkell 1949: 4), furnished a wide inventory of the material culture of a previously unknown early Holocene pottery-bearing Mesolithic culture. These hunter–gatherer–fisher groups produced elaborate ceramics together with bone instruments, a geometric chipped–stone industry, sandstone grinders, grinding stones and many other stone and bone instruments. The faunal inventory from the site, which also included large mammals and gastropods, pointed mainly to a fishing-based subsistence strategy (Peters 1986).
After the pioneering work of Arkell from the 1970s onwards, several other Mesolithic sites in central Sudan have been excavated or superficially examined both to the north and south of the country’s capital city, and many other sites with Early Khartoum pottery scattered on their surface have been located, providing a picture of an impressively wide distribution of this culture (Ali Hakem and Khabir 1989; Caneva 1983a, 1988; Caneva and Santucci 2006; Clark 1973, 1989; Donner 1967–1968; Elamin and Mohammed-Ali 2004; El-Anwar 1981; Fernández et al. 1997, 2003a, b; Fernández 2003b; Fuller 2004; Fuller and Smith 2004; Garcea and Hildebrand 2009; Gatto 2006a, b; Geus 1998–2002; Haaland and Magid 1995; Honegger 2005, 2010; Jesse 2003; Khabir 1985, 1987; Krzyzaniak 2002; Marks and Mohammed-Ali 1991; Nelson 2001; Reinold 2001; Riemer and Jesse 2006; Salvatori and Usai 2006, 2008; Salvatori et al. 2011; Schuck 1988, 1993; Usai 2001, 2006; Usai and Salvatori 2002, 2005; Usai et al. 2010). Unfortunately, most, if not all, of the excavated sites were badly affected by post-depositional disturbances of various origins (e.g., erosion, burrows, pedogenetic transforms), but mainly due to their almost systematic re-use as Neolithic, Meroitic, Post-Meroitic and later graveyards. Such circumstances are at the base of a profound misunderstanding of the evolution of the Khartoum Mesolithic culture (Usai in press). From a comprehensive survey of the excavated Mesolithic sites provided in Appendix 1, it is evident that our present knowledge of this cultural phenomenon suffers a great deal from the lack of any stratigraphically meaningful information and from the unreliability of radiocarbon determinations—in the sense that it was almost never possible to correlate 14C dates to material assemblages from discrete stratigraphic units (see also Usai in press).
Sites are mostly destroyed by Meroitic, Post-Meroitic, Christian and Muslim graves. The building method of Post-Meroitic tumulus-like graves consisted of digging a large chamber along the border (mainly facing towards the Nile) and sometimes just on top of the small mounds that were relicts of the Mesolithic sites. After burying the dead, the deposit of the prehistoric sites was collected and used to erect an earthen tumulus above the refilled burial chamber. The tumulus was subsequently partly flattened or at best lowered by colluviation, resulting in the prehistoric material being redistributed on a larger area around the tumulus. Only a few of these structures were needed to completely destroy a Mesolithic site. Pedogenesis, animal activities, Meroitic, Christian, Mediaeval, modern and contemporary graves, wind and water erosion were also active agents of post-depositional disturbances.
Excavated sites have no stratigraphy and the cultural material is mixed. This can be more easily detected when the original stratigraphy was the result of Mesolithic and Neolithic frequentation. When the site is monocultural (e.g., Mesolithic), it is impossible to detect the mixing effect and disentangle it. No chronologically sequenced assemblages are yet available to be considered as a control tool of the archaeological material recovered.
A side effect of this phenomenon is that the kind of data (the lack of discrete assemblages) coming from excavated sites cannot be used to measure diachronic and/or synchronic variability, similarity/dissimilarity of dichotomic traits at regional and supra-regional levels, and to select hypothetical diagnostic cultural traits (e.g., Dotted Wavy Line contra Incised Wavy Line; Alternately Pivoting Stamp contra Dotted Wavy Line or similar) to construct any middle-range theory (Kosso 1991; Raab and Goodyear 1984; Tschauner 1996). The situation is such that not only studies of pottery, polished stone tools, chipped lithics and worked bone, but also of other kinds of materials, such as faunal and botanical remains are marred by the mixture produced by re-deposition processes described under point (a). We are dealing with something very different from post-depositional mixing between discrete stratigraphic units due to bio- or cryoturbation and/or anthropogenic factors such as everyday activities at an ancient settlement with postholes, rubbish pits, foundation trenches, etc., and thus the best constructed models cannot help to reconstruct meaningful correlations among material assemblages (e.g., Brantingham et al. 2007).
In such a situation, radiocarbon dates are never associated with discrete assemblages so that we have to acknowledge that they date nothing else than themselves, and therefore must be approached according to a “chronometric hygiene” principle (sensu Spriggs 1989).
Nonetheless, those radiocarbon dates can possibly be used to obtain at least a preliminary and approximate idea about the timespan covered by the central Sudan Mesolithic. From this point of view, two dates from Sarurab II (see Appendix 1: Table 12) could provisionally suggest a starting point of the cultural period at the beginning of the ninth millennium cal. BC. A number of dates from the Atbara sites (Aneibis, Abu Darbein and El Damer) help to cover the eighth millennium cal. BC (see Appendix 1: Tables 3, 4 and 5), while the bulk of the dates from central Sudan Mesolithic sites covers the seventh and sixth millennium cal. BC. If this is correct, the “Khartoum Mesolithic” would span four millennia and, consequently, it would likely have experienced, variation at the level of material production, social organisation, settlement patterning, subsistence strategies, etc.
If we cannot approach middle-range theory explanations from the data collected from central Sudan Mesolithic sites, we have to go back to the low-range theory level and approach the problem of protocols of data collecting in terms of excavation procedures, data processing and publication (see also Usai in press). Whichever explanatory theory is favoured, data collection methods need to conform to well established and declared protocols. Indeed, it is disappointing to see that strict stratigraphic procedures are not routinely applied in Sudanese archaeology. The consequence is very low stratigraphic control of the data collected, a lack of concern regarding data inequality according to the different situations encountered in cultural deposits and, finally, the production of explanatory models based on idiosyncratic perceptions of the cultural and chronological meaning of unstratified pottery types and stylistic traits.
As it emerges from Appendix 1, all the sites reviewed here have been excavated using artificial horizontal cuts because of the lack of discrete features and distinct deposition layers but, with the exception of an attempt made by Caneva et al. (1993), no one sought to define a strategic approach to understand the negative post-depositional effects due to Meroitic and Post-Meroitic practices of using prehistoric sites as tumulus-like grave fields.
Comments on Data Use and Interpretations
Many researchers have speculated on the highly biased data from central Sudan Mesolithic sites producing a picture of the Khartoum Mesolithic which is static, undifferentiated and occasionally reliant only on ethnographic analogy. Many clichéd statements occur in old and more recent syntheses of the Khartoum Mesolithic (e.g., Caneva and Santucci 2006: 45; Caneva and Zarattini 1982: 6; Garcea 2006a: 209; Haaland 1987: 50–51; Haaland and Magid 1995; Midant-Reynes 2000: 91–99; Wengrow 2006: 20–21). Regarding site dimensions, for example, it has been written that “The main features of the Early Khartoum include a high density and large extension of the sites, covering between 1 and 4.5 ha, which did not occur either before or after the Early Khartoum expansion” (Garcea 2006a: 209, italics mine). However, site dimensions are usually calculated according to the spread of archaeological material on the surface, but we have to recall that a recurrent and very peculiar type of post-depositional disturbance brought about by Meroitic and Post-Meroitic tumuli erection and subsequent colluviation processes, produced a distribution of the prehistoric material that makes the actual size of the material spread two to three times larger than the original site area (see below for the discussion about sites 10-X-6 and 16-D-5; Salvatori et al. 2011; Usai in press; Zerboni 2011).
Furthermore, any measure of site distribution density along the Nile can be meaningful only if calculated according to a well-defined correlation between chronology and discrete material assemblages, that is to say, after establishing strict criteria of site contemporaneity. The present distribution and thus settlement density is the cumulative effect of not necessarily synchronous archaeological sites during a period of three or four millennia. Different strategies of residential mobility, reduced residential mobility and logistical mobility over such a wide time span would have produced, at a local as well as at a regional scale, a multiplying effect on the spatial distribution of settlement, frequentation evidence and, consequently, on site density.
The thickness of the archaeological deposits, frequency of burials and density of potsherds and animal bones have been used as indications that sites represent permanent or semi-permanent occupations (e.g., Caneva and Santucci 2006; Garcea 2006a: 209; Haaland and Magid 1995). However, thickness of archaeological deposits at sites that have been so dramatically disturbed is a dubious parameter by which to assess degrees of sedentism. Moreover, this parameter needs to meet a number of other preconditions (Varien and Mills 1997; Gallivan 2002 and quoted literature). In most cases, thickness of the deposit can only be related to redeposition phenomena and not to a coherent stratification of more or less continuous site frequentation (in Appendix 1, deposit thickness, when available, has been reported and the data show extremely high variability). Besides, a number of sites had Neolithic habitations that surely contributed to the original thickness of the deposit (before the disturbances produced by natural, animal and pedogenetic transformations and by Neolithic, Meroitic, Post-Meroitic and later graves). It is also highly possible, and must be taken into account, that varied settlement behaviour at different times could have contributed differently to the cultural deposits. Where a proper stratigraphic control of superimposed, discrete layers, possibly supported by radiometric determinations is not at hand, it would be awkward to speculate about degrees of sedentism or mobility only on the basis of mixed deposits’ thickness. Moreover, the claim of the existence of “Deeply stratified sites and remains of hut floors” (Garcea and Hildebrand 2009: 3; see also Garcea 1996: 321) at Khartoum Mesolithic sites in the Khartoum region would appear not supportable on the basis of the published evidence reviewed in Appendix 1.
“....the degree of occupational redundancy in areas with resources suited to mass collecting and boiling, especially small seeds, should be better correlated with pottery use. While occupational redundancy may promote decreased mobility and/or sedentism in the long run…, quite mobile peoples can still be tethered to certain locations by making consistent and repeated use of them.” (Eerkens 2003: 736)
Nevertheless, permanent or semi-permanent occupations, possibly linked to intensive or delayed-return economies (Kusimba 2005: 340), are the result of a process and have to be explained within the local culture history. In the case of the Khartoum Mesolithic, the solution of this critical problem is constrained by the fact that we cannot attribute pottery production and faunal remains to a given phase of the Khartoum Mesolithic. It means that we transfer an uncontrolled dataset to the entire chronological span of this culture, to affirm that no change occurred in 3,000 or 4,000 years in settlement, subsistence, material production strategies and, moreover, in demographic, environmental and climatic variables.
In the context of prehistoric studies, frequency and distribution of burials could be considered a crucial indicator of a degree of sedentism; but in the case of the Khartoum Mesolithic, it becomes a priority to critically re-examine chronological attribution of graves against stratigraphic and cultural information (see also Usai in press). Graves of different periods have been found at the sites listed in Appendix 1. Leaving aside easily recognisable Neolithic (when provided with grave goods or when displaying typical grave structure), Meroitic, Post-Meroitic, Christian and Islamic graves, a restricted number of unfurnished graves have been thought to be Mesolithic in date. First, we have to remember that Neolithic graves are not always furnished, as ascertained at a number of excavated cemeteries (e.g., Salvatori 2008a: 132–134). Thus, a different kind of information, if possible, is necessary to disentangle the problem. Doubtless, strongly disturbed deposits cannot help to establish stratigraphic relationships between graves and a largely non-existent stratigraphy (note the inconsistency of chronological attribution and the drawing of the deposit section at Saggai 1 in Caneva (1983b): Fig. 6). In the end, the number of possible Mesolithic graves in central Sudan is almost zero; but to support some degree of sedentism, we would need formal cemeteries or structured behaviours such as graves repeatedly placed inside or in strict spatial relation with habitation structures (e.g., Honegger 2005: 100, Fig. 5). Moreover, we have to remember that the Khartoum Mesolithic spanned a very long time and mortuary practices are sensitive to population mobility patterns. Relationships among people, social and ideological perceptions of territoriality, food resource concentration and critical resource availability (sensu Keeley 1988: 376) are important correlates of mortuary practices and can produce great variation in the archaeological record (Littleton and Allen 2007).
At the moment, only one skeleton can be positively attributed to the Mesolithic period of central Sudan thanks to a radiocarbon date (see Appendix 1) obtained directly from human bones at Shabona, a site briefly investigated by Desmond Clark in 1973 (Clark 1989: 395).
Another oft-repeated statement in Mesolithic Khartoum studies concerns the chronological divide between incised wavy line, and impressed dotted wavy line and rocker stamp dotted zigzag decoration motifs (Caneva et al. 1993; Caneva and Santucci 2006; Garcea 2006a: 209).
In fact, chronological/cultural development in the Khartoum Mesolithic is generally restricted to a barely verifiable transition from incised wavy line (IWL) to impressed dotted wavy line (DWL) decoration patterns, but this approach to cultural change is, I would argue, not justifiable based on the archaeology. Cultural change (or “development”) cannot be restricted to a single trait in pottery decoration to be considered as a profitable heuristic tool. The appearance and disappearance of single traits are firstly to be understood in the context of pottery making practices, inside a wider production and symbolic system. Consequently, variability patterns and cultural change have to be considered and measured in well stratified and dated contexts (e.g., Shennan 2000) and this is not a condition satisfied in the field of Khartoum Mesolithic archaeology. In conclusion, the debate about IWL and DWL decoration motifs (Caneva 1996; Caneva et al. 1993; Caneva and Zarattini 1982; Garcea 1993, 2004, but a different and contrasting position in Garcea 2006b: 99–100; Jesse 2000, 2001, 2003, 2004) would not seem particularly useful and is based on, at best, controversial data sets both at the chronological and cultural level (Mohammed-Ali and Khabir 2003: 50–53; and see below).
Preliminary Data for a New Starting Point in the Study of the Khartoum Mesolithic
The “El Salha Project”: Survey and Excavations in Central Sudan
From October 2000, a team of the Italian Institute for Africa and the Orient, Rome, under the direction of Donatella Usai started an extended project of survey and excavations named the “El Salha project” from the name of a village 15 km south of Omdurman, along the western bank of the White Nile. The area intensively surveyed was 5 × 35 km from the bank of the river to the Jebels to the west (Cremaschi et al. 2007; Salvatori and Usai 2006, 2008; Usai 2003, 2005, 2006; Usai and Salvatori 2002, 2005; Usai et al. 2010), later extended further south towards the Jebel Aulia dam.
During the early and perhaps middle Holocene, large areas west of the White Nile were dotted by marsh and/or palaeolake formations of different size (Cremaschi et al. 2007; Williams and Adamson 1980: 299), with the one in the surveyed area arriving at its maximum extent at the foot of Jebel Baroka, some 30–35 km from the river (Cremaschi et al. 2007).
More than 160 sites of different periods (from the Palaeolithic to Islamic period) were located and named according to Hinkel’s (1977) system. Among them, a number of sites with Khartoum Mesolithic pottery were found along the bank of the White Nile, in the interior along the southern shore of the aforementioned palaeolake, on the terraces of the Jebels, and along a 3-km-wide strip along the White Nile to the Jebel Aulia dam.
Along the western bank of the White Nile, 10 sites with Mesolithic pottery were located almost regularly spaced on the contour line of 382 m. All the sites are low mound-like structures. They were later settled or frequented in the Neolithic period and used as burial grounds in Late Meroitic, Post-Meroitic, Christian, Early Islamic and present times. Four more Mesolithic sites have been located at a distance ranging from 1.5 to 5 km from the White Nile.
At the same time, the investigations at a restricted number of sites show the limits posited by previous researches and the new perspectives opened on the understanding of the Khartoum Mesolithic culture evolutionary trends.
The Lesson from the 10-X-6 Excavation
This means that from the present dataset, we can only say that the Khartoum Mesolithic is a pottery producing hunter–gatherer–fisher cultural horizon. No material, cultural, behavioural, economic, social etc. variability can be described. Only a kind of aseptic (typological) descriptive list of artefacts and faunal remains can be provided, but no coherent assemblages can be described and positively correlated to radiocarbon determinations or at least to a relative chronology.
The Al Khiday Sites
About 10 km south of 10-X-6, a number of Mesolithic, Neolithic, Late and/or Post-Meroitic sites were located in 2004. They were labelled 16-D-3 (or Al Khiday 3: a large yet unexcavated mound-like site with Mesolithic and Neolithic pottery on the surface); 16-D-4 (or Al Khiday 2: a low mound-like site with Mesolithic and few Neolithic evidence on the surface. The site turned out to be a pluristratified graveyard with a Mesolithic occupation phase; Usai et al. 2010); 16-D-4B (or Al Khiday 2b: a very low morphologic relief where a number of Mesolithic pits have been excavated); 16-D-5 (or Al Khiday 1: a relatively large mound-like site with Mesolithic and Neolithic pottery on the surface and evidence of Late or Post-Meroitic tumulus-like graves on its eastern half); 16-D-6 (or Al Khiday 4: a flat Neolithic site with discrete clusters of pottery and other artefacts; only a small excavation test trench has been dug in 2009 and a radiometric determination gave a date of 5360 ± 80 BP or 4325–4064 1σ cal. BC); and 16-C-2, a large Late or Post-Meroitic cemetery with more than a hundred tumuli. Sites of interest in the context of this paper are 16-D-5, 16-D-4 and 16-D-4B. At the first site, a test trench was opened in 2005 and later widened to cover 200 m2.
The 16-D-5 (Al Khiday 1) Settlement
In the lower, well-preserved section (Salvatori et al. 2011; Zerboni 2011), two different cultural phases were identified. The lower and older one is characterised by several micro-layers of ashes and one fireplace containing amounts of shell fragments (mainly Pila wernei), fish and mammal bones, lithic and bone tools, pottery sherds and a grinding stone with its working surface covered with red and yellow ochre powder. These layers were separated from the virgin substrate (the sandy bar of the Nile) by a thin and very hard layer containing pottery sherds and faunal remains.
14C dates from Al Khiday 1 settlement (16-D-5)
13C/12C ratio (‰)
Age 14C BP
Age 1σ Cal BC
Age 2σ Cal BC
7980 ± 40
7960 ± 40
16D5 SU 455a
7940 ± 40
16D5 SU 455b
7830 ± 40
16D5 SU 48
7870 ± 40
16D5 SU 00
7740 ± 50
16D5 SU 37
7710 ± 40
The Sites 16-D-4 and 16-D-4B (Al Khiday 2 and 2b)
16-D-4 is situated on a low relief about 50 m to the north of 16-D-5, where almost 800 m2 were excavated. The area was mainly used as a cemetery in at least three different periods: (1) Pre-Mesolithic, (2) Neolithic and (3) Late Meroitic (Usai et al. 2010).
14C dates from Al Khiday 2 and 2B (16-D-4 and 16-D-4B)
13C/12C ratio (‰)
AGE 14C BP
Age 1σ Cal BC
Age 2σ Cal BC
16D4 Pit 29
7770 ± 40
16D4 Pit 6a
7760 ± 90
16D4 Pit 75
7640 ± 110
16D4 Pit 52
7620 ± 50
16D4 Pit 74
7600 ± 90
16D4B Pit 6
7540 ± 50
16D4 Pit 73
7530 ± 100
The Pottery from the Al Khiday Sites
The following section will principally use the generally accepted typology of Mesolithic and Neolithic Sudanese pottery established by Isabella Caneva (1983c; Caneva and Marks 1990). Caneva’s typology is based on a tree-like structure organised according to techniques, tools, elements, motifs and structures. It does not consider stylistic variation in motif execution and in pottery fabric (e.g., motif performance and different temper recipes, sensu Gosselain 1992, 2000). Co-occurrence of sherd decoration, thickness and temper are analysed in detail in Appendix 2.
The Pottery Sequence at 16-D-5
Pottery from 16-D-5 and the other Al Khiday sites, as well as that from all the other Sudanese Mesolithic sites (e.g., Haaland 1993: 68; Caneva 1991: 265; Elamin and Mohammed-Ali 2004: 104), is formed by hand with the coiling technique and the edge of each clay ring is bevelled to provide better joining as proved by a large number of observations on breaks occurring at the coil joints. Vessels bases, from where the coiling formation starts, are usually very thick (from 1.75 to 3 cm).
Shapes are various but it is very difficult to trace meaningful statistics because sherds with rims and large section of the body are relatively few. Nevertheless, we can say that, in the older phase, restricted, vertical and open bowls form the majority of the sample with mouth diameters varying from 16 to 28, 20 to 38 and 16 to 32 cm (clustering around 20 cm), respectively. Jars of this phase are mainly globular with a mouth diameter between 20 and 22 cm.
Shapes of the second phase are not much different. The majority of the fragments can be referred to as restricted to vertical and open bowls, with mouth diameters varying from 16 to 28, 25 to 29 and 18 to 30 cm, respectively. Jars are mainly globular or sub-cylindrical with convex walls and their mouth diameters vary from 20 to 28 cm.
Temper variability is impressively large (30 different qualitative recipes). It has been observed through fresh cuts with the help of a magnifier lens and described according to a code list (Appendix 2: Table 21). When temper recipes are associated with decoration motifs it is possible to detect differential behaviours that we can summarise here (for a more detailed analysis, see Appendix 2, Tables 22 and 23). IWL pottery is usually tempered with feldspar and quartz grains with angular edges (>2 mm), while in the second phase of the sequence it is more often tempered with fine feldspar and quartz grains with angular edges (<1 mm). Other recipes are only rarely used in the first phase IWL pottery, while sand tempers progressively increase in the second phase reaching a maximum (12 %) in the last stratigraphic unit (SU 29). A very similar trend is present in the temper variability of pottery decorated with a rocker stamp dotted zigzag (RS d zz) motif. Feldspar tempers are the main types associated with this decoration motif (71.16 %, cumulative percent value) with an increase of fine feldspar grains with angular edges (<1 mm) and sandy tempers during the second phase of the sequence. Lunula-type decoration, rocker stamp drops (RS drops), rocker stamp plain zigzag (RS pl zz) and scraped sherds are mostly characterised by sandy tempers in various admixtures with ochre, calcareous grains and vegetal additions. APS sherds are, on the contrary, characterised by a fine feldspar and quartz grains with angular edges (<1 mm) temper. DWL sherds are mainly tempered with sandy recipes (73.34 %) and to a lesser extent with feldspars and quartz grains with angular edges (26.66 %). A principal component biplot and correlation matrix (Appendix 2, Fig. 35 and Table 24) of decoration types contra temper recipes confirm the strong link between Q and Qf tempers on one side and among the sandy tempers on the other.
Sherd thickness ranges from 0.43 to 3.2 cm, but we can detect five size classes according to specific decoration types (see Appendix 2, Table 25). In this case, it is worth mentioning here that APS pottery is much thinner since its first appearance in the sequence (mean 0.57 cm), than all the other types (means from 0.78 to 1.07). This kind of pottery is scarcely represented and it appears only towards the end of the site sequence and we can anticipate that it is better represented (but never exceeding 4 %) at the late Mesolithic site 10-W-4 (late sixth millennium cal. BC) (Salvatori and Usai 2008; Salvatori et al. 2011).
The Pottery from 16-D-4 Pits
As reported above, about 48 Mesolithic pits have been excavated in 16-D-4. The pottery collected from the fill is directly linked with the pottery assemblages from the second phase of 16-D-5. The pits’ inner stratigraphy shows that they were used several times. Ashy deposits are in some case separated by one sandy layer that can be interpreted as a discontinuity in use. Each ashy layer inside the pits shows fired stone pebble beds at different depths pointing to a repeated and continuous use during a relatively long period of time. Pottery from lower levels in some of the pits (no. 44, 64, 70, 71, 73 and 75) is characterised by the presence of deep dot or drop decoration which helps to make a strong correlation with the first sub-phase of the second period of 16-D-5. The bulk of the pottery from the pits is well dated to the second sub-phase, while during the 2011 campaign at least two pits can be assigned to the first sub-phase of the local sequence, based on the pottery assemblage.
Rs drop decoration types are the most frequent in the pits and we can notice, as at 16-D-5, a decreasing presence of WL types. Tempering materials are exactly the same as in the nearby settlement; namely Q and Qf types prevail with WL, RS d zz and WL + RS drop types. For the other decoration types, a sandy temper prevails, alone or mixed with ochre, calcite and/or vegetal elements (Appendix 2, Fig. 36 and Tables 26, 27 and 28).
The Pottery from 16-D-4B Pits
The eight pits excavated at the 16-D-4B site yielded a small amount of pottery, all pertaining to the second sub-phase of the second phase of 16-D-5, apart from one sherd with a Deep Dot or Drop decoration (Appendix 2, Fig. 37 and Tables 29, 30 and 31). The excavated pits represent only a small sample of the many pits dotting this area. The single radiometric date from pit 6 (see Table 2) confirms a later date of this group of pits, falling in the second sub-phase of 16-D-5 phase two. No variation has been noticed regarding temper recipes when compared with contemporary assemblages from 16-D-4 to 16-D-5.
The Late Mesolithic Site 10-W-4
The 10-W-4 Late Mesolithic site (second half of the sixth millennium cal. BC) is located 5 km to the west of 10-X-6 and about 9 km from the actual course of the White Nile. It is a flat site, about 2 ha large, placed on a stabilised Pleistocene dune at the border of a large wadi which is nowadays used as a sand quarry for building purposes.
In 2004 and 2005, excavations have been carried out at the site opening an area of 350 m2. The stratigraphy at the site is very complex because of recurrent frequentation of the area which can be assumed to have been seasonal. No later disturbances were noted, except for the prehistoric huts cutting older structures of the same type and cultural affiliation. Water and wind erosion processes acting on a substantially flat surface produced no measurable effect. No water gully was noticed on the surface of the site. Major damages at the northern edge of the archaeological area were produced by the sand quarry. Some compression effect on the surface was produced by trucks travelling to the quarry. After our second excavation season, a planned urbanization of the area started, which now covers large parts of the prehistoric site with new houses.
In the excavated area, we cleaned one complete semi-subterranean hut, 8 m long and 4 m wide, with the longer axis oriented north–south. A second hut of the same type has been excavated only partially along the southern limit of the excavation area. Some pottery and lithic clusters were investigated just outside the huts, while in the second partially excavated hut, a garbage pit was excavated containing large amounts of animal bones, lithic debitage and tools, and pottery sherds (Linseele 2010; Salvatori et al. 2011).
The single 14C determination on charcoal available from the site can be considered only approximately suggestive of the site frequentation chronology (Beta-201726: 6490 ± 40 BP/1σ 5488–5379 cal BC; 2σ 5529–5367 cal BC). Stratigraphic evidence from the excavated area (repeated use with partial cutting into previous huts), faunal remains from the garbage pit (faunal remains are mainly from hunting activities; Linseele 2010; Salvatori et al. 2011, in press) and the presence of discrete, separated clusters of cultural material spread on such a wide area suggests a seasonal frequentation of the site.
10-W-4: The Pottery Assemblage
All the pottery was handmade with a coiling technique as seen at the Al Khiday sites. The low number of sherds from which it is possible to calculate the mouth diameter hampers statistical analysis of different vessel shape occurrences. Nonetheless, we can say that bowls of different size and shape (restricted and open varieties) are well represented. Much rarer is the occurrence of dishes. The base is usually progressively thickened in the case of bowls and dishes, while it is of a thicker and more convex shape in the case of small and larger jars.
External surfaces, in closed forms, are usually polished and only rarely burnished. On sherds from open shapes, we noticed polishing and decoration also on the internal surface. Sherd thickness does not vary too much across the different decoration types except for APS sherds which are clearly separated from the rest (Appendix 2, Table 35), as at 16-D-5.
The data from the El Salha and Al Khiday sites cast light on a number of basic questions about the cultural history of the Khartoum Mesolithic. Moreover, the same data can provide a firm basis to describe a changing path in subsistence strategies, settlement patterns, and material production in the light of possibly climatic and environmental changes during the Early to Middle Holocene. The archaeological sequence reconstructed at these sites covers the first quarter of the seventh millennium cal. BC (first phase at 16-D-5) and the second and third quarters of the same millennium (second phase at 16-D-5 and features from 16-D-4 and 16-D-4B) or at best 700 years. A later segment of the sequence (the third quarter of the sixth millennium cal. BC) is represented at the Late Mesolithic site of 10-W-4. This means that there are still large gaps to fill in the future and that new excavations may modify, expand and better detail the processes we sketch out in the following pages.
On Pottery Variability
As suggested by our sequence, variability in pottery production is a multivariate phenomenon whose complexity should be more deeply investigated. Only segments of a much longer and wider trajectory of the Khartoum Mesolithic pottery production can actually be observed and we can hypothesise that we still miss the beginning of the process (the eighth millennium cal. BC?), together with later but equally important segments (from the fourth quarter of the seventh to the middle of the sixth millennium cal. BC and from the last quarter of the sixth millennium BC to the rise of a food production economy in the area). Equally important limitations on understanding and reconstructing the regional dimension of pottery variability are set by the lack of other stratified Mesolithic deposits in central Sudan. In fact, it seems highly possible, that future information from other sites, from the Atbara River to south of Khartoum, will result in an altered picture of pottery type and style distribution and thus of pottery variability in central Sudan.
The evidence from the Al Khiday sites suggests, at the rise of the seventh millennium cal. BC, the presence of a well developed ceramic production with range of decoration types, some of which persisted and developed in the following periods and others that disappeared abruptly at the end of the first phase or early in the second phase. Together with the disappearance of incised Lunula-like decoration motifs, there is a consistent decrease of IWL pottery and the parallel increase of RS drops and dots and RS d zz decorations.
It is still difficult to produce explanatory models for the total variability of pottery decoration and temper recipe types. It can be reasonably stated that Khartoum Mesolithic hand-made pottery production was accomplished at the household level and for household consumption. If we consider the wide range of motifs with a long, though quantitatively uneven, permanence in our sequence we could easily explain the temper recipe variability and the number of stylistic variants as the effect of individual pot-makers’ habitus (sensu Gosselain 2000: 189). In turn, seen within the frame of a socially accepted and shared meta-linguistic medium, this can be interpreted as an inter-household as well as an infra- and inter-site communication device (Braun 1991; David et al. 1988; Evers et al. 1988) in a specific local historical setting. Different values of decoration type frequencies along the sequence should be linked with variability in households or affiliation groups’ demographic trends. Total variability can be better linked to the social organisation of pottery production at the level of the household (transmission, stability, variation, fluctuations in group and population size) (Eerkens and Lipo 2005, 2007; Hart and Brumbach 2009; Shennan 2000). Furthermore, archaeologists should not forget that in the case of well stratified sites: “The site components …. represent palimpsests created over many years of periodic visits by different subpopulations and/or different generations of the same subpopulation” (Hart and Brumbach 2009: 377). The relatively high level of variation in the 16-D-5 pottery assemblages concerning decoration motifs, rendering styles and temper recipes is well explained in the case of relatively mobile populations with a household-based pottery production by a one-to-one vertical transmission (Richerson and Boyd 2005) or better in the frame of the “descent with modification” perspective (Jordan and Shennan 2009; Shennan 2000).
Only apparently different appears to be the problem concerning IWL pottery. Its highly conservative temper recipe would point to a specific use of this kind of vessels. This pottery is considerably stronger than others and possibly more resistant to mechanical than thermal shocks (for a survey of the many dimensions of the problem see for example: Bronitsky and Hamer 1986; Schiffer and Skibo 1987; Schiffer et al. 1994; Sterba et al. 2009). The decrease of the type frequency in the sequence would be hardly understandable if we would not assume that the need for such a durable pottery type was decreasing as well, remaining at a very low level for the rest of the Mesolithic period. On the contrary, a different explanation that would bring IWL back to the general trend of pottery production is at hand, considering that the decrease of this kind of pottery is equally balanced by the increase of RS d zz types that are also characterised by feldspar and quartz grains with angular edges tempers. This last pottery type has the same temper composition as the IWL and the same characteristics regarding the parameter of robustness. The need for a pottery apt to resist mechanical shock would have remained more or less unchanged along the sequence while satisfied by the differential co-occurrence of differently decorated pottery with the same physical prerequisite.
The frequent recourse, in the literature on the Khartoum Mesolithic, to pottery “fossil types” as chronological markers needs to be re-considered if not entirely abandoned. The IWL/DWL chronologic-cultural dichotomy has been, since Arkell’s excavations at Khartoum Hospital (Arkell 1949, 1972: 221) and Qoz (Arkell 1953: 101), generally assumed as proved (Caneva 1996; Caneva et al. 1993; Caneva and Zarattini 1982; Garcea 1993, 2004), apart from some exceptions (Jesse 2000, 2003, 2004; Mohammed-Ali and Khabir 2003). The Al Khiday sites attest the presence of the DWL type since the beginning of the local sequence with an early contextual date in the first quarter of the seventh millennium cal. BC. DWL, in different stylistic variants, is present along the entire sequence. Its assumed introduction from the Sahara during the Late Mesolithic is to be abandoned as well as its use as a chronological and cultural marker of the Khartoum Mesolithic sequence. Varieties of this kind of decoration, with a fixed chronological position in the Al Khiday sequence, can on the contrary be useful to place some sherds from the Wadi Howar (Jesse 2000, Fig. 2.1–2), west of the southern Dongola Reach, in their correct chronological perspective.
Unfortunately, the lack of stratified assemblages from other Mesolithic sites prevents us from measuring inter-site variability in the entire central Sudan macro-region. Only well established and contextually dated pottery assemblages would allow us to enter into the social dimension of pottery production, and its possible relevance, together with other classes of data, on detecting regional changes, different economic trends and social boundaries (Gosselain 2000: 188) both at micro- and macroscale (Pauketat 2001) inside this widespread cultural horizon.
Between Mobility and Sedentism
On the contrary, the second phase of the settlement, has revealed more consistent structural remains such as mud (pisé) walls, packed mud floors and fireplaces possibly used for long periods (Salvatori et al. 2011; Zerboni 2011). Moreover, contemporary separated activity areas have been excavated at the close by sites 16-D-4 and 16-D-4B (see above). Forty-eight deep fireplace pits that may have been used for freshwater gastropods, meat and fish processing (possibly also smoking) where investigated at 16-D-4 and 14C dated within the second phase of 16-D-5 settlement; moreover, eight garbage pits, among many others, were excavated in the nearby site 16-D-4B. The relevance of refuse disposal behaviours in understanding complexity levels among hunter–gatherer communities does not need any further comment (Hardy-Smith and Edwards 2004 and quoted literature). The complexity of specialised feature distributions in the area around the settlement, the architectural mud remains at 16-D-5, the presence of large pottery vessels with mouth diameters of about 50 cm, possibly used as storage facilities (e.g., Ingold 1983; Kent 1999), and the mound-like structure of the deposit at the main site, allow to envisage a shifting towards a more restricted residential mobility during the seventh millennium BC.
In a household mode of pottery production, an increase in the number of decoration motifs, stylistic varieties, and temper recipes, as attested in the second phase at 16-D-5, can also be explained by a residential mobility shift, as the side effect of group fission processes in response to a population increase due to an enhanced residential stability (Cohen 1985; Chatters 1987; Keeley 1988; Marshall 2006). A residential mobility shift may be suggested by the presence, along the White Nile western bank, of more or less regularly spaced similar mound-like sites (10-S-3, 10-X-6, 10-X-4, 10-X-5, 10-X-8, 16-D-1, 16-D-5, 16-D-3, 16-I-3; Salvatori and Usai 2005) with Khartoum Mesolithic pottery which can be dated to the second phase at 16-D-5 on typological and stylistic grounds.
For the Late Khartoum Mesolithic period (ca. sixth millennium cal. BC), we only have few and very preliminary data. Of the four sites located between El Salha and the Jebel Aulia dam, at a distance of 1.5–5 km from the western edge of the White Nile floodplain, only 10-W-4 has been partly excavated. All four sites share similar characteristics: they are flat sites more than 2 ha large and show a discrete concentration of cultural material with pottery assemblages that are roughly comparable. These sites are hardly comparable with the mound-like sites of the previous period, and may be seasonal camps. If this hypothesis is confirmed, we would have to consider another shift in the residential pattern, or rather an increasing mobility of the local Mesolithic population. While this remains, at this point, a working hypothesis, it is worth noting that an increasing amount of environmental, climatic and hydrologic studies suggest that around the beginning of the sixth millennium cal. BC, the entire area from the western Egyptian desert to the regions south of Khartoum underwent, with a gradual trend from north to south, a consistent climate deterioration (decreasing annual rainfall, desertification, aeolian deflation, etc.; Bubenzer et al. 2007; Nicoll 2001, 2004). The dramatic effects of such a phenomenon (the progressive abandonment of previously settled areas from the Middle Holocene onward: Nicoll 2004) are thought to be at the base of the transition from a collecting to a food producing economy in southern Egypt and northern Sudan (Close 2002), a shift that in the north did not affect pottery styles for about one millennium (Gatto 2006a, b). Central Sudan was only later experiencing the major effects of the desertification process and, obviously, not as an abrupt phenomenon as documented by a number of well-dated alternating periods of dry and wet climate (Williams 2009; Williams et al. 2010). But the process was in action as well, and we can suppose it had repercussions on population size, subsistence strategies, mobility and possibly relocation (e.g., Hassan 2002). Some evidence of the process of desertification in the Al Khiday area is furnished by the desiccation of early Holocene lakes dated by a series of radiocarbon determinations to the sixth millennium cal BC (Williams and Adamson 1980: 299–300; Adamson et al. 1982).
This work offers fresh data on pottery assemblages of the Khartoum Mesolithic culture with the documentation of an early horizon characterised by new decoration types (Lunula and deep drop types) and the stylistic variability of other types in the course of time. A number of radiocarbon determinations help to place change and variability of the pottery assemblages of the Khartoum Mesolithic into a definite chronological frame, that is to say during the seventh millennium cal. BC. Frequency of pottery decoration types as well as technical aspects, like temper used in the process of ceramic production, have been analysed to understand patterns of variability in the relative frequency of decoration types and temper recipes linked with pottery types of different utilisation.
Moreover, a reappraisal of the archaeological literature on the Khartoum Mesolithic makes it evident that after 70 years of field research, there are still no feasible data about the where, when, how and why a ceramic technology and production was appearing, or in other words when, how and why “…a sophisticated merging of previously separated domains of human knowledge and experience: resources, technological processes, and needs” (Rice 1999: 3) appeared in central Sudan. Following the pragmatic approach used in this article, it becomes clear that this important set of questions can only be answered by future investigations of other stratified sites, hopefully producing the data needed to disentangle the problem and to describe the formative period of the Khartoum Mesolithic culture.
At the present stage of the research, it seems to be fortunate that a first sequence, though still incomplete, has been finally excavated. This helps to approach the complex history of this cultural horizon on the basis of discrete pottery assemblages, material culture variability, varieties of formal and technical choices in pottery production (decorative motifs in a contextual setting, temper recipes, stylistic variability in decoration motif execution), possible changes in economic strategies and degrees of mobility with consistent effects on site location, distribution and organisation.
Demographic fluctuation suggested by site differential distribution and shape could have been one of the many factors shaping the material culture and the social structure, the last being more elusive and needing more and more field work. Ongoing studies on faunal remains from 16-D-5, 16-D-4a and b, and 10-W-4 sites could finally give way to exploring hypotheses of differential hunting strategies in the different Khartoum Mesolithic phases according to possible resource and environmental stress and human hunting pressure (Allué et al. 2010; Dusseldorp 2010; Munro 2003, 2004, 2009a, b; Munro et al. 2009; Munro and Atici 2009; Peters and Pöllath 2004; Starkovich and Stiner 2009; Stiner and Munro 2002; Stutz et al. 2009).
Other problems, foreseen during this excurse, will remain difficult to address. An example is the topic of infra-cultural (i.e., to detect discrete sub-regional clusters and the dynamics of cultural and social exchange) and inter-cultural boundaries (cultural and social exchange among eastern, southern and northern areas such as the Kassala/Jebel Moya region on one side and Nubia and the western Egyptian and Sudanese desert on the other).
However, signifying that it is useless and misleading to base hypotheses of cultural transmission and contacts with other regions on an unpatterned set of data coming from unstratified and mixed deposits, we offer a new path and new data on the Khartoum Mesolithic as the primary goal that we wanted to pursue with the present paper.
Finally, we have to stress that a new beginning in the study of the Khartoum Mesolithic is needed, not only to better understand its cultural, social and economic complexity, but also to contribute to a more general anthropological cognition of late Pleistocene and early Holocene complex hunter–gatherer–fisher communities.
The El Salha project, directed by Donatella Usai, has been supported by Ministero degli Affari Esteri (2000–2011), Istituto Italiano per l’Africa e l’Oriente (2000–2011), Università degli Studi di Parma (2005–2011), Michela Schiff Giorgini Foundation (2002–2003, 2005, 2007), Università degli Studi di Padova (2010–2011) and GASID of Torino (2000–2009).
I am much indebted to Tina Jakob, the physical anthropologist of the project, for editing the English of this paper and for her invaluable comments on the text. I wish to thank Donatella Usai, Andrea Zerboni, Friederike Jesse, and the three anonymous reviewers for their valuable comments on earlier drafts of the paper. I thank Adria LaViolette for her consultations on this paper and her editorial contributions. Of course, shortcomings and inaccuracies in the paper are mine.