Introduction

The use of the term Neolithic in relation to the Sudanese Nile valley has been strongly influenced by the Nabta Playa sequence in the Egyptian Western Desert (Wendorf et al. 2001; Nelson 2002). In the Nabta-Khiseiba area (Fig. 1), this term has been applied—based on the erroneous link between the appearance of pottery and the Neolithic (Smith 2013), and on the largely disputed domesticated nature of the African aurochs at this date (see below)—to archaeological phases (Early Neolithic of El Adam, El Ghorab and El Nabta) that are a better fit with pottery-bearing hunter-gatherers from the valley. This has produced aberrant descriptions of archaeological materials, for example, the labelling as Neolithic of most of the Mesolithic pottery production in Lower and Upper Nubia (e.g., Wolf 2004; Smith and Herbst 2005; Edwards 2004; Osman and Edwards 2012). This confusion has since spread to colleagues working in Egypt (e.g., Tristant 2012b).

Fig. 1
figure 1

Map of the Nile valley with Nubian and central Sudan Neolithic sites mentioned in the text

Despite the development of a wider scientific approach to the study of regional and micro-regional environments in northern Africa, central Sahara, the Sahel, and the Egyptian Western and Eastern deserts (see online Appendix A), there remain large areas where the limited nature of the archaeological data prevents the measuring of local variability in socio-cultural processes. Upper Nubia and central Sudan are among the less understood areas in this respect, at least for the Early and Middle Holocene, mainly due to the still very scant data on Mesolithic and Neolithic sites and cemeteries (Usai 2016).

No systematic work has been planned or conducted on Mesolithic sites in Upper Nubia, with the exception of the Kerma area (Honegger 2012). Limited test excavations at an apparently unstratified Mesolithic settlement area and an Early Neolithic site have been carried out at the Fourth Cataract (Dittrich et al. 2007, 2015; Dittrich and Gessner 2014; Petrick 2012; Wotzka et al. 2012). Only a Middle Neolithic (5th millennium cal BC) cemetery from this area has been fully published (Salvatori and Usai 2008a). Excavations at other Neolithic sites and cemeteries located in the Kerma area have only been published in preliminary form (Honegger 2003, 2004b, 2005a, 2005b; Reinold 1994, 2001, 2004a, b).

The situation is not much better in central and south-central Sudan. Nearly all the Mesolithic sites excavated or tested experienced heavy post-depositional disturbance (cf. Salvatori 2012, Appendix 1; Usai 2014), except for some sites located along the western bank of the White Nile, south of Khartoum, that recently produced a stratigraphic sequence for the 7th millennium cal BC Khartoum Mesolithic (Salvatori 2012; Salvatori et al. 2011, 2014; Usai 2014. All dates in this paper have been calibrated using OxCal 4.3: Ramsey and Lee 2013; Reimer et al. 2013). At present, a Czech archaeological mission at Sabaloka, at the Sixth Cataract, just to the north of Khartoum, is verifying the stratigraphic integrity of Mesolithic deposits in the area (Suková and Varadzin 2012; Suková et al. 2014). Post-depositional disturbance also affects the excavated Neolithic deposits in central Sudan (cf. Salvatori and Usai 2008b, 2016). Only some of the sites, such as the Early Neolithic cemeteries at Kadero (Chłodnicki et al. 2011) and Ghaba (Salvatori et al. 2016) and a Late Neolithic cemetery at Kadada (A & B) (Reinold 2007), have been fully published.

New data and a more scientific and holistic approach help in tracing a new history of the Mesolithic and Neolithic periods in Sudan (Salvatori and Usai 2019), finally offering a new way to look at and interpret mainly ideological behaviours (funerary ideology) that have wrongly been seen to endorse a narrative depicting local societies as pastoral, hiding the true economic structure of the Neolithic communities as well as the complexity, the variability and the variation along their chrono-cultural trajectories.

Pastoralism

Often when researchers apply the term ‘pastoralism’ to the transition from foraging to food production and to the social organisation of the Neolithic communities of northern Africa, the Sahara and the Nile Valley, they do not specify the socio-cultural system or the prevailing ‘mode of production’ (for discussions of this problem, see Khazanov 1984; Bar-Yosef and Khazanov 1992; Linseele 2010, 2013; di Lernia 2013a; Arbuckle and Atici 2013; Hammer 2014; Linseele et al. 2016; Hammer and Arbuckle 2017; Arbuckle and Hammer 2018). It also seems that in the archaeological literature pertaining to northeastern Africa, the term pastoralism is used in an attempt to differentiate it from earlier and contemporaneous cultural processes of the Near East. In the literature on the Near East, in contrast, some researchers (e.g. Makarewicz 2013) have recently denounced the under-estimation of the contribution of the pastoral sector of the economy in the construction of local production systems and socio-cultural organisation during the Neolithic.

If, in the archaeology of the Near East and the Levant, we can observe a growing interest in detecting the many facets of the ‘agriculture–pastoralism’ relationship (Makarewicz 2013), in the archaeology of northern Africa it is still rare to find pastoral trajectories studied within well-defined regional, environmental and chronological contexts, and in a manner that stresses both diachronic transformation and synchronic relationships with other economic activities, such as hunting, fishing, collecting, cultivation and agricultural practices. Only recently has a new line of research resulted in detailed regional and sub-regional studies of mixed economies in northern Africa and the Sahara, providing complex socio-cultural and economic trajectories that cannot be subsumed under the summary and equivocal label of ‘pastoralism’ (Biagetti and di Lernia 2003, 2007; Cancellieri and di Lernia 2014; Cremaschi and di Lernia 1996, 1999; Cremaschi and Zerboni 2009; Cremaschi et al. 2014; di Lernia 2001, 2002, 2006; di Lernia and Tafuri 2013; Mercuri 2008a, b; Tafuri et al. 2006; Dunne et al. 2012, 2013; Lucarini 2013; Linseele 2010, 2013; Mulazzani et al. 2016; Lucarini et al. 2016).

Research in the Tadrart Acacus area and the high variability in cultural trajectories in the Egyptian Western Desert (Gehlen et al. 2002) show, as assumed by Human Niche Construction theory, that human communities exhibit individual ways of reacting to and interacting with environmental and climatic conditions and variations, as well as geomorphological peculiarities (Laland and O’Brien 2010; Riede 2011; Rowley-Conwy and Layton 2011; Odling-Smee et al. 2003; Smith 2015). Because local communities were co-actors in cultural and economic transformations, any valid approach to the study of the transition to food production requires a well-delimited analytic and geographic viewpoint. For the Nile valley, we suggest that the correct analytic viewpoint would be to recognize the spatial, temporal, climatic, environmental, geomorphological and demographic diversity along both the longitudinal axis (Western Desert and oases vs. Eastern Desert) and the latitudinal axis (the Delta, Upper, Middle and Lower sectors of the Nile river).

The Onset of Food Production: Pastoralism as a Monolithic Explanatory Concept?

The term pastoralism is used as a catch-all in most of the archaeological literature, both old and recent, on northeastern Africa and in historically untenable ethnographic analogy and ethnoarchaeological research (Lane 2014; Gosselain 2016). Only a few scholars consider pastoralism to be a highly variable economic system that (including during the recent prehistory of northern Africa) has produced socio-ideological and behavioural constructs differing both synchronically and diachronically. The worldwide literature on pastoralism is impressively large, but in only a relatively few cases has it moved to disentangle differences within the general pattern of pastoralism (see online Appendix B).

To properly assess regional variability, research in northeastern Africa needs a methodological shift, such as that accomplished through the multidisciplinary research conducted in the Tadrart Acacus (southwestern Libya). This wide-spectrum project allowed the identification of specific features of the local pastoralism that, in its earliest phase, was characterized by reduced vertical mobility (Cremaschi and di Lernia 2001; Cremaschi and Zerboni 2009; Tafuri et al. 2006), similar to the model described for a Bronze Age community in Central Asia (Frachetti 2008). Indeed, defining different regional forms of pastoralism and their relationship to other kinds of food production is of great importance to understanding economic and social structures in relation to the environment. The recent publication of Neolithic cemeteries excavated in Nubia and central Sudan (Kobusiewicz et al. 2010; Honegger 2004b; Salvatori and Usai 2008a; Salvatori et al. 2016; Chłodnicki et al. 2011; Reinold 2007) dated to between the 6th and the 4th millennium cal BC, despite limitations imposed by the scarcity of excavated contemporary settlements, can assist with constructing a new perspective on the economic and ideological structures of the local communities that have been commonly defined as pastoralist (e.g., Caneva 1991; Gautier and Van Neer 2011; Garcea 2016).

Moreover, a new perspective could help to avoid the construction of a fixed dichotomy between agriculturalists and pastoralists as suggested by Fiona Marshall and Elisabeth Hildebrand (2002), which can only lead to an emphasis on the role of one to the detriment of the other, as was observed for the Neolithic of the Near East (Makarewicz 2013).

Critical to understanding the onset of food production in the Sudanese Nile valley are the data and related chronology concerning the first evidence for domestic animals and plants.

Cattle, Sheep and Goats in the Archaeology of Northeastern Africa

The African domestication of Bos primigenius (advanced for the Nabta Playa enclave by Gautier 2001; Bradley et al. 1996; Caramelli 2006; Edwards et al. 2010) has been widely debated (A. B. Smith 1986, 1992; A. C. Smith 2013; Wendorf and Schild 1994; Riemer 2007; Pöllath 2009; Linseele 2013; Brass 2013, 2018). Nevertheless, even if early attempts to control and exploit Bos primigenius occurred at Nabta Playa, as suggested by control of Barbary sheep in the central Sahara (di Lernia 2001), this does not imply that full domestication was attained (Zeder 2008; Stock and Gifford-Gonzalez 2013; Linseele 2013). Genetic research (Ho et al. 2008; Pérez-Pardal et al. 2010; Decker et al. 2014; Olivieri et al. 2015) now points to a single cattle domestication event in the Near East and suggests that the small stock of domestic cattle brought from the Near East to northeastern Africa went through hybridisation with the local aurochs, as has been argued to have occurred in parts of Europe, such as Italy (Götherström et al. 2005; Svensson and Götherström 2008; Achilli et al. 2009; Decker et al. 2014).

It seems worth noting that in northeastern Africa, the first indisputably domestic cattle are found in association with sheep and goats (hereafter referred to as ovicaprines) derived from the Near East (Linseele 2013). It seems possible, therefore, to hypothesise the arrival of a wider Neolithic package along one or more of the three suggested routes of penetration, namely, through the northern Sinai Desert corridor, through the southwestern Sinai Desert and across the Red Sea, or following a sea route along the Mediterranean coast (Bar-Yosef 2013; Shirai 2016). At present, the archaeological evidence suggests that domestic animals, and probably domesticated cereals too, arrived from the southern Levant, either via the southern Sinai or via the western shore of the Red Sea to the area of Sodmein Cave. Both routes had been known to Pre-Pottery Neolithic (PPN) B people since the end of the 9th millennium cal BC, as suggested by the presence of Helwan arrowheads in the Fayum (Shirai 2002; contra Holdaway et al. 2017) and, above all, in the Eastern Desert of Egypt (Tristant 2010, 2012a). An ingression of small, southern Levantine Neolithic groups from the Sinai corridor to the Nile delta and then south along the Nile to southern Lower Nubia and northern Upper Nubia cannot be ruled out (Bar-Yosef 2013; Shirai 2016).

Different kinds of contacts occurred between the Near East and northeastern Africa at least since the Late Pleistocene (Shirai 2006; Tristant 2010, 2012a; Bar-Yosef 2013). So it is no surprise that available data suggest that domesticated ovicaprines arrived in the Egyptian Eastern Desert and possibly northern Sudan from the Near East by the end of the 7th millennium cal BC (Vermeersch et al. 1996, 2015; Close 2002; Linseele 2013). Domestic cattle is also present in the same general area from that time onwards (Pöllath 2009; Linseele 2013; Linseele et al. 2014), a period that in the Levant corresponds to the PPNC. A different, maritime path for the introduction of domestic ovicaprines in northwestern Africa—that is, the northeastern and southeastern coast of the Mediterranean Sea (Cardium pottery tradition developed along the Marmarican and Cyrenaican littoral)—has been suggested on the basis of Y-chromosome and mtDNA analyses (Pereira et al. 2009 and references therein; Muigai and Hanotte 2013) and new archaeological data (Linstädter 2008, 2011; Linstädter and Müller-Sigmund 2012; Linstädter et al. 2012; Ballouche et al. 2012; Hachid 2015; De Faucamberge 2016).

While it seems too soon to arrive at decisive conclusions about the routes of domestic livestock ingression into northeastern Africa, it is clear to us, based on new archaeological data, that a shift in focus is required, away from ‘domestication processes’ toward ‘the introduction of food production processes’, and the acceptance of these processes by communities who had long been involved in the cultural control of wild animals and in intensive collecting. At the moment, the northeast African evidence points to a domestic stock ingression from the Sinai corridor and then south to the area of Sodmein Cave, or across the Red Sea from southwestern Sinai, at the end of the 7th millennium cal BC, possibly driven by the rapid climate change around 8200 cal BP (Morrill and Jacobsen 2005; Overpeck and Cole 2006; Alley 2007; Lal et al. 2007; Weninger et al. 2009; Berger and Guilaine 2009). Ovicaprines and cattle are attested to from the beginning of the 6th millennium BC onwards at Nabta Playa (Gautier 2001), and a little later (from c. 5900 cal BC) in the Dakhla Oasis (Linseele et al. 2014; Riemer 2007; Brass 2018). Later still, they reach the Fayum (Linseele et al. 2014) and the Merimde site (Close 2002).

A Mediterranean route seems suggested by the evidence of Farafra (Hidden Valley) where goats are present starting from the very end of the 7th millennium cal BC (Barich 2014, p. 25; Barich and Lucarini 2014, p. 473).

Archaeozoological data on the timing of the spread of ovicaprines and cattle were first presented by Angela Close (2002, Fig. 9) and subsequently refined by Heiko Riemer (2007, Fig. 5.9). New evidence from the Kerma area in Upper Nubia is used here to contribute to this picture. Remains of domestic cattle (in the form of a frontal with horn cores) have been found in an Early Neolithic (early 6th millennium cal BC) grave at El-Barga (Honegger 2005b; Linseele 2013, p. 101), and domestic sheep/goat bones dated to 5700 cal BC have been recovered at Boni Island (Petrick 2012; Wotzka et al. 2012). Claims of domestic cattle from the older site of Wadi el Arab (Chaix and Honegger 2014) have subsequently been rejected (Linseele 2012; Chaix and Honegger 2014, p. 213). It can therefore be concluded that domestic animals (cattle, sheep and goat) appear in southern Egypt and northern Sudan at the very end of the 7th millennium or in the early 6th millennium cal BC.

If, indeed, the data concerning the introduction of domestic ovicaprines in the Egyptian Eastern Desert make it possible to push the date for the beginning of the neolithisation process back into the end of the 7th millennium, and to hypothesize a contemporaneous introduction of domestic cattle, attested to at that date in the Sinai Peninsula (Zeder 2008), a similar date for the arrival of domestic cattle in southern Yemen (Crassard 2009; McCorriston 2013; Henton et al. 2014, p. 121; Scerri et al. 2018) lends support to the hypothesis of a nearly concurrent migratory process from the southern Levant to northern Arabia. This region was, at least partially, exploited by Levantine groups since the PPNA and B (Crassard et al. 2013), possibly as an outcome of the 8200 BP climate crisis. From this point in time the Neolithic in the peninsula took different paths with several forms of pastoralism in the interior, and coastal fishers with some domestic animals (Uerpmann et al. 2000; Crassard and Drechsler 2013; Uerpmann et al. 2013).

Domesticated Plants in the Nile Valley

New data on Early and Middle Holocene cultural evolution in the Sudanese part of the Nile valley have substantially changed the perception of the neolithisation of this region, and especially its timing (Table 1). The analytical study of some cemeteries in the northern part of the country, south of Kerma (Fig. 1), namely, in the Kadruka area (Reinold 2000, 2001, 2005, 2006) and the Seleim Basin (R12: Salvatori and Usai 2008a), and Kadero (Chłodnicki et al. 2011) and Ghaba (Salvatori et al. 2016), in central Sudan, has enhanced the level of detail of the chronological framing of the period (Table 1; Salvatori and Usai 2016; Usai 2016). Some of these new data have not been given sufficient attention in the scientific literature, even though they are essential to any consideration of the process of neolithisation that hypothesises the arrival, together with domestic animals, of a wider Neolithic package, including domestic plants of the Triticae family, winter rainfall crops that are not native to Sudan in any shape or form.

Table 1 Neolithic chronology of Nubian and central Sudanese cemeteries with details of the number of graves and presence/absence of bucrania and vegetal pillows

Archaeological excavations at the Middle Neolithic B cemetery of Kadruka 1 (second half of the 5th millennium cal BC), revealed graves with barley ears positioned, like a pillow, under the skulls of the dead (Reinold 2000, 2006; Yves Lecointe, pers. comm.). Recently, analyses of a sample of the white deposit recovered under the skull of an inhumed individual (as well as under the skulls of at least five other individuals) in an early Middle Neolithic A period grave (Grave 46 inferior) at the R12 cemetery, excavated in the Northern Dongola Reach (Fig. 2), and two of the 39 burials from the Early Neolithic of central Sudan cemetery of Ghaba (Shendi Reach between the Fifth and Sixth Cataracts), show the presence of domestic cereals at a very early date (Madella et al. 2014; Out et al. 2016; Ryan et al. 2016). Whilst Triticeae are not present in the white deposit, the analysis of dental calculus from the individuals buried at Ghaba has shown that this community was also consuming Triticeae cereals (Madella et al. 2014; Ryan et al. 2016). The ‘pillows’ turned out to be a mass of phytoliths pertaining almost exclusively to wheat and barley, in the case of R12, and mainly Paniceae, with a lesser presence of Panicum/Setaria, Sorghum and Digitaria, in the case of Ghaba (Madella et al. 2014; Out et al. 2016; Ryan et al. 2016). Of further interest is the fact that a high proportion of the phytolith skeletons of both Ghaba samples show straight and curved edges, sometimes stepped, that cannot be explained by natural tissue fragmentation and indicate plant processing (Out et al. 2016, pp. 115–116).

Fig. 2
figure 2

Photo and artwork of Grave 129 at R12 Nubian Middle Neolithic cemetery; the white area behind the skull indicates the ‘pillow’ made of phytoliths

Phytoliths from Grave 46 Inferior (one of the oldest graves according to the cemetery seriation presented in Salvatori 2008) at R12 have been AMS dated to the end of the 6th millennium cal BC (5311–5066 cal BC, Madella et al. 2014). The phytolith samples from Ghaba Grave 233 and Grave 295 have been AMS dated to the second half of the 6th millennium cal BC and the first half of the 5th millennium cal BC (Grave 233: 5620–5480 cal BC; Grave 295: 4730–4540 cal BC, Madella et al. 2014) respectively, which accords with their position in the graves’ seriation (Salvatori 2016).

These new data demonstrate that cultigens originating in the Near East were present in both Upper Nubia and central Sudan from at least the second half of the 6th millennium cal BC. For Ghaba, these new data also show intensive exploitation of millet and other grasses, together with wheat and barley, as part of the diet of this population. These new data drastically push back the date for the presence of domestic wheat and barley in Upper Nubia and central Sudan, and thereby force us to reconsider the neolithisation process of the region, especially because the presence of domestic cereals at this early date makes it reasonable to envisage the possibility that all domesticates arrived as a coherent Neolithic package (sheep, goat, cattle, wheat and barley) in the Sudanese Nile valley between the end of the 7th and the beginning of the 6th millennium cal BC.

At present, domestic cereals evidence from Egypt is later than that recovered from R12 and Ghaba: domestic cereals evidence at Merimde-Benisalama (von den Driesch and Boessneck 1985; Eiwanger 1988; Wetterstrom 1993), the Fayum (Wendrich et al. 2010) and Mostagedda (Wengrow et al. 2014) provided more recent dates.

The hypothesis of two separated ingressions, the first at the end of the 7th millennium cal BC, of domestic animals, and the second at the end of the 6th millennium cal BC, of domesticated cereals, as proposed by Ofer Bar-Yosef (2013; see also Linseele et al. 2014), cannot be ruled out. Indeed, different ingression trajectories can at present be described for northwestern Mediterranean Africa, which necessitates the consideration of a more complex, chrono-culturally and geographically manifold process (Ballouche and Marinval 2003; Linseele 2010; Linstädter and Kehl 2012; Lucarini 2013; Morales et al. 2013, 2016; Peña-Chocarro et al. 2013; De Faucamberge 2016; Mulazzani et al. 2016; Martínez-Sánchez et al. 2018).

Early and Middle Neolithic (6th and 5th Millennia cal BC): The Data Set for Upper Nubia

Post-depositional processes have reduced most of the settlements to a vast spread of archaeological material, transforming them into undecipherable palimpsests. Therefore, at present, funerary contexts provide the only available evidence with which to reconstruct the ideological underpinnings as well as the social and economic structures of Early and Middle Neolithic communities in Nubia and the central Sudan.

As stated before, the oldest evidence that can be linked to a food-producing economy in Sudan has been recorded from the cemetery excavated at El Barga (Honegger 2004b), in the Kerma district, dating to the first half of the 6th millennium cal BC. A bucranium—the (upper) frontal with horn cores—from domestic stock (Figs. 2, 3), representing the earliest evidence for the Neolithic in the Sudanese Nile valley (early 6th millennium cal BC), was found in the El Barga cemetery. In the later Neolithic phases, in Nubia and in central Sudan, bucrania become one of the most common funerary attributes.

Fig. 3
figure 3

Photo and artwork of Grave 150 at R12 Nubian Middle Neolithic cemetery; the skeleton of the individual is flanked by two bucrania

Only a few of the 103 graves excavated there have been published to date (Honegger 2004b, 2005a, b), but a preliminary anthropological report is available (Crèvecoeur 2012). Grave goods were recorded in two-thirds of the graves; these consist of stone, bone, shell (Conus sp.) and ostrich-eggshell beads; ivory bracelets; lip-plugs; sandstone palettes; and freshwater (Chambardia sp.) and marine (Glycymeris pectunculus, Cypraea sp.) shells that are typologically related to grave goods from 5th millennium cal BC Neolithic cemeteries excavated in Lower and Upper Nubia (Kobusiewicz et al. 2010; Salvatori and Usai 2008a). However, ceramic vessels, which are rarely present in el-Barga graves, clearly pertain, with modification, to the previous Mesolithic tradition (Honegger 2004b, 2005a, b). According to the excavator,

ces dates ne correspondent pas à la première phase du Néolithique de la région. En effet, ce cimetière montre déjà des transformations techniques et sociales parfaitement accomplies et on peut se douter que l’introduction de l’élevage s’est produite à une étape antérieure [‘these dates do not correspond to the first phase of the Neolithic of the region. Indeed, this cemetery already shows perfectly accomplished technical and social transformations, and one could suspect that the introduction of stock breeding occurred during a previous stage’ transl. D. Usai] (Honegger 2005b, p. 247).

No other data pertaining to subsistence are available.

After this Early Neolithic phase, the sequence shows a gap of about three to four hundred years, until the second half of the 6th millennium cal BC (Table 1).

The earliest date for the Middle Neolithic phase (5311–5066 cal BC) was produced by the remains of the Triticeae phytolith pillow recorded from Grave 46 at the R12 cemetery (Madella et al. 2014), located farther south of Kerma, in the Seleim Basin. Underneath Grave 46 is an even older but undated grave, Grave 46 Inferior, also furnished with a bucranium. At R12 cemetery (Salvatori and Usai 2008a), 40 graves out of a total of 166 (24.1%) were furnished with up to six bucrania, and additional faunal remains of domestic and wild species were present (Pöllath 2008). The distribution of bucrania within the cemetery does not show any specific pattern; where the sex could be determined, they are equally distributed between the sexes (14 M, 14 F). Bucrania also occur in the graves of children and infants (11 out of 16). Sex is undetermined for only one adult individual (Salvatori 2008: Tab. 11.1). Other evidence linked to the diet of the population comes from one of the six graves (3.7%) where vegetal pillows (Fig. 2) had been preserved.

Bucrania and vegetal pillows are consistently present throughout the cemetery sequence, whereas during the second half of the 5th millennium cal BC (Middle Neolithic B or Multaga phase) the sequence shows a radical change in the pottery grave furnishings, from a large variability in shapes and decorations to a very standardized vessel typology and decoration limited to a burnishing of the external surfaces and an impressed decoration on the rims (Salvatori 2008; Usai 2016; Salvatori and Usai 2019). There are few other grave goods, namely white sandstone armlets and some bead types typical of this phase.

Little is known about the numerous cemeteries located immediately south of Kerma, in the Kadruka area. A radiocarbon date from the Kadruka 33 cemetery (6520 ± 80 bp; 56185338 2 σ cal BC; Reinold 2004a), older than that obtained on R12 Grave 46, would place this one within the gap documented in the El Barga sequence. This cemetery was, unfortunately, destroyed by gravel quarrying (Reinold 2004a), and no detailed descriptions of the graves are available.

Better known is Kadruka 1 cemetery, dating to the Middle Neolithic B phase (c. 4500–4000 cal BC), in which several individuals were found with ears of barley under their skulls (Reinold 2000: Fig. on pg. 57).

To the north of Kerma, in the 29 graves excavated at Sedeinga which, according to radiometric determinations and pottery types, pertain to the Middle Neolithic B of the region, neither bucrania nor vegetal pillows were recorded (Reinold 2004a, 2005), nor were any other elements that can be linked to the diet of the population. No information at all is available for the six Neolithic graves of uncertain chronology excavated at Ashkan, also north of Kerma (Bonnet 1986). No vegetal pillows, or bucrania or other faunal remains, have been recorded from the 19 Neolithic graves excavated at Umm Melyekta Island, upstream of the Fourth Cataract (Fuller 2004; Edwards and Fuller 2005), or from the 53 Middle Neolithic B graves excavated by the Section française de la direction des antiquités du Soudan (SFDAS) in the El Multaga area (Geus and Lecointe 2003a, b; Peressinotto et al. 2004). Intensive collection of wild or domestic plants is possibly suggested by the presence at Multaga of three bone sickle handles (?) of a type known from R12 (Cenci 2008, Pl. 7.2: Grave 83/375), Kadruka 1 (Reinold 1994, fig. on p. 9 and fig. on p. 10) and Umm Melyekta (Edwards and Fuller 2005, p. 24).

No bucrania or vegetal pillows were recorded from the small Neolithic cemeteries, contemporary with the Upper Nubian Middle Neolithic B (c. 4500–4000 cal BC), that were excavated at Gebel Ramlah (Schild et al. 2002; Kobusiewicz et al. 2004, 2009, 2010), near Nabta Playa, in the Lower Nubia Western Desert. Only two single cattle horns, possibly used as cosmetics containers, were present at this site (Kobusiewicz et al. 2010). In Upper Nubia, we do not yet have evidence of Neolithic cemeteries dating to the 4th millennium cal BC, although some settlement areas pertaining to a Pre-Kerma period (end of the 4th millennium cal BC) have been excavated there (Honegger 2004a, 2011).

To complete the picture, a Neolithic ‘facies’ known as the Abkan, recognised in the area of the Second Cataract during the Aswan Salvage campaign (Shiner 1968b; Nordström 1972; Myers 1958), is documented mainly from surface sites dating from the mid 6th millennium cal BC onwards (Table 2; Gatto 2002). Settlements dating to this Neolithic facies discovered in the area of the Second Cataract (Shiner 1968b) are listed in Table 2, which shows that indicators of subsistence system are quite scanty. There are many more in the Seleim Basin (Welsby 2001), but most of these have been reduced to vast surface scatters of human artefacts; they rarely yield ecofacts (Welsby 2001; Grant 2001). Of the many Neolithic settlements, only Site 8 in the Kerma area has revealed preserved features like oval-shaped huts and fireplaces (Honegger 1999, pp. 77–79; 2004b, p. 38). About the nature of the occupation of the site, the excavator writes,

Table 2 Data illustrating in brief the contexts available from Nubian Neolithic sites

Les arguments en faveur d’un établissement temporaire sont cependant fragiles. L’exposition du site aux crues du Nil ne signifie pas qu’il ait été inondé annuellement, mais peut-être seulement lors d’épisodes exceptionnels. La construction d’une architecture nécessitant l’enfoncement de poteaux de 10 cm de diamètre n’est pas forcément compatible avec l’idée d’un campement saisonnier, où l’on s’attendrait à trouver des installations plus légères. Enfin, le fait qu’il n’y ait pas de structures de stockage ne permet pas de conclure à l’absence de collecte ou d’activités agricoles à proximité. [‘The arguments in favour of a temporary settlement are, however, weak. The exposure of the site to the Nile floods does not mean that it was flooded annually; perhaps it was flooded only during exceptional episodes. The construction of architecture requiring the sinking of posts of 10 cm in diameter is not necessarily compatible with the idea of a seasonal camp, where one would expect to find lighter constructions. Finally, the fact that there are no storage structures does not allow one to conclude that there were no collecting or agricultural activities nearby.’ Transl.: D. Usai] (Honegger 2006, p. 8).

However, thus far, only the pottery has been published from Site 8 (Honegger 2004a).

Four Neolithic settlement areas were recorded at Saï Island (8-B-10A, 8-B-10B, 8-B-10E, and 8-B-76: Garcea 2007), two of which were explored. They have been attributed to the Abkan Neolithic (Garcea et al. 2016: Tab. 1 site 8-B-76 Level 1). The dates from the site (second half of the 6th to the 5th millennium cal BC) not only confirm continuity of the neolithisation process throughout the 6th millennium cal BC, but also give us realistic hope that future research in the area will completely fill the chronological gap that currently exists. However, they produced faunal data (Table 2)—some shells and a single, fragmentary cattle lower molar (Garcea et al. 2016)—that seem too meagre to allow any kind of inference on subsistence strategies.

Altogether the settlement data for Lower Nubia appear currently very poor, and until this is rectified, the differential transformative effects of funerary rites on the living will be extremely difficult to disentangle.

Early Neolithic (5th Millennium cal BC): The Data Set for Central Sudan

Our knowledge of the Neolithic period in central Sudan differs from that for Nubia, in that some sites have been partially but extensively excavated, namely, the settlements of Shaheinab (Arkell 1953), Geili (Caneva 1988) and Shaqadud (Marks and Mohammed-Ali 1991) and the two 5th millennium cal BC cemeteries of El Ghaba (Lecointe 1987; Salvatori et al. 2016) and Kadero (Chłodnicki et al. 2011). Unfortunately, the deposits of the settlements have suffered the most from post-depositional disturbance (see Salvatori 2012: Appendix 1, for an overall review of the situation). The faunal remains have been studied in detail (Table 3), and they do provide an overview—albeit problematic because of the difficulty of linking the material to specific features—framed within a specific chronological sequence. In contrast, the two cemeteries provided rich information regarding the cultural material as well as the socio-economic background, within a more detailed chronological framework. Ghaba cemetery, located in the Shendi area, produced 265 Early Neolithic graves. Radiometric determinations help to date the cemetery’s use between 5200 and 3800 cal BC (Salvatori 2016: Tab. 4.1). The analysis of the El Ghaba graves indicates two periods of use, a first period with clear connections with the Middle Neolithic A of Upper Nubia and a second period, beginning around 4600 cal BC, fully pertaining to the central Sudan Early Neolithic phase of El Shaheinab (Arkell 1953; Caneva 1988).

Table 3 Data illustrating in brief the contexts available from central Sudan Neolithic sites

At Ghaba, only eight graves (3.02%) have bucrania (one in each grave). Seven of these graves were clustered in the northwestern corner of the excavation’s main trench. A single, isolated grave was located in the southwestern corner of the main trench, and the presence of another, similar cluster to the west of this grave cannot be ruled out. White areas, or vegetal pillows, were recorded in 39 (14.7%) of the Ghaba graves. Unlike the bucrania, which were present only in first period graves, vegetal pillows occurred throughout the sequence, without any evidence of discrete spatial or chronological clustering. A co-occurrence of vegetal pillows and bucrania is confined to only three cases.

Pottery material (Chłodnicki 2011), as well as radiometric determinations (Kabaciński 2011), from the Kadero cemetery (Chłodnicki et al. 2011), where 218 graves have been excavated (Krzyżaniak 2011), place most of the graves in the second phase of the central Sudan Early Neolithic (Shaheinab phase). A handful of graves were dated to the early Late Neolithic (beginning of the 4th millennium cal BC). No bucrania or vegetal pillows were recorded at Kadero, but in a single grave (Gr. 114), a series of 11 quartz segments have been interpreted as the remains of a sickle, the bone or wooden handle of which has not survived (Kobusiewicz 2011, p. 294, Figs. 17–19).

Lastly, a group of 38 graves, dated by means of grave goods and 14C determinations to the Shaheinab phase (4600–4000 cal BC), were excavated at the site of al-Khiday 2, south of Omdurman and Khartoum, on the left bank of the White Nile (Jakob 2014; Salvatori et al. 2014). No bucrania or vegetal pillows were found, but dental calculus analysis revealed that the group was eating barley and/or wheat (Buckley et al. 2014, S2).

Late Neolithic (4th Millennium cal BC): The Data Set for Central Sudan

No Late Neolithic cemeteries have been excavated in Upper Nubia, while only three such cemeteries have been excavated in central Sudan. The cemeteries at Kadada A and B, which together contained 92 graves (Reinold 1982, 2007), have been fully published. That at Kadada C (Reinold 1982, 2005), where at least 221 graves were excavated, has been only partly published (56 graves) (Reinold 1982). Kadada C is slightly older than Kadada A and B. The third cemetery, Geili, produced only 11 Late Neolithic graves, which have been published (Caneva 1988). Four child pot-graves were excavated at Es-Sour (Sadig 2014).

At Kadada C, 9 of the 56 published graves (16.1%) were furnished with bucrania (Reinold 1982), possibly attesting to the re-emergence of this practice in the Shendi region at the beginning of the 4th millennium cal BC. While no bucrania were present at Kadada A and B, we suggest, based on the description of the grave furnishings by Reinold (1982), that traces of vegetal pillows were present in at least three of the graves (5.3%, Gr. 85/11, 86/2, 86/3). Vegetal pillows were apparently more frequent at Kadada A and B. They are described as ‘fragments d’origine végétale et des anomalies dans le remplissage près des squelettes, relevées sous l’appellation de zone à coloration blanchâtre (vraisemblablement les traces d’un témoin disparu)’ [‘fragments of vegetal origin and anomalies in the fill near the skeletons, identified under the term “whitish-coloured zone” (presumably traces of vanished remains)’ Transl.: D. Usai] (Reinold 2007, p. 128). It is not known whether samples for analysis were taken from these graves; however, this description is reminiscent of the vegetal pillows found at Ghaba. Unfortunately, there is no analytic evidence, nor any spatial distribution, available for the other 156 graves excavated at this cemetery. Neither bucrania nor vegetal pillows are mentioned for the Geili graves (Caneva 1988).

The Ideological Message of the Funerary Symbolism and its Relationship to Socio-Economic Organisation

The decidedly inadequate volume of data and research activities on socio-economic organisation mean that only a provisional picture of the Neolithic economy along the Nile in Sudan can be surmised. Current data point to economic activities based on hunting, gathering and fishing, with an additional contribution provided by the integration of wild plant collection and/or cultivation and agricultural activities based on domestic cereals and the herding of domestic animals. These domestic cereals and animals originated from the Near East and were introduced into the valley between the end of the 7th and the beginning of the 6th millennium cal BC.

As discussed earlier, the frontier area between present-day Egypt and Sudan seems a plausible jumping-off point for the spread of the Neolithic package to central and northern Egypt, as well as to the Egyptian oases in the Western Desert (Close 2002; Riemer 2007). Aside from the evidence gained through the study of domestic cereals and animals, genetic and physical anthropological studies could make an important contribution to unravelling this issue. While these are still unable to produce a detailed human history for the period covering the late Pleistocene to early and middle Holocene in the area, a recent preliminary study of Mesolithic and Early Neolithic human remains (Crèvecoeur 2012) describes strong and significant differences (anatomical discontinuity) between the two populations from el-Barga (in the Kerma area). The Mesolithic group is more similar in terms of body size and robustness to the groups at Jebel Sahaba, Taforalt and Wadi Halfa (Crèvecoeur 2012, p. 28). Moreover, the genetic or anatomic discontinuity between the late Pleistocene population of Jebel Sahaba and that of the Gebel Ramlah Final Neolithic (following the Wendorf terminology) implies that ‘replacement or genetic swamping of an existing gene pool by an outside group, or groups, occurred after the Pleistocene’ (Irish 2005, p. 520). If this suggestion is correct, we anticipate that this discontinuity occurred near the end of the 7th millennium cal BC and that it is linked to the arrival of small agro-pastoral groups from the Levant (Bar-Yosef 2013, p. 244), apparently in connection with the so-called 8200 BP climate crisis, as suggested on genetic grounds (Smith, A.C. 2013). Whether or not the economic transition to food production along the Nile river was driven by external factors, we know it took place in a fertile environment inhabited by hunter-gatherer-fisher populations with a nearly sedentary lifestyle, and that their sedentism was probably related to the fact that fishing was the dominant subsistence activity (Van Neer 2004; Linseele and Zerboni 2017). Pastoralism has been considered by far the earliest and primary food-production strategy in northern Africa and the Nile valley (di Lernia 2013b; Garcea 2016). The spread of agricultural practices may have been constrained by the uneven distribution of agricultural land, as pointed out by many researchers (most recently by Gautier and Van Neer 2011; Williams et al. 2015). Nevertheless, the emphasis placed on pastoralism by Gautier and van Neer (2011) may be an overestimation of the role of animal breeding compared with agricultural and/or cultivation activities. The problem also seems to be a taphonomic one: not only do faunal remains preserve better than plant remains, but there may well be a pervasive, but misleading, tendency to equate the presence of animal remains in funerary contexts (e.g. bucrania) with the economic base of the Neolithic communities. The contribution of archaeobotanical data, enhanced by phytolith and starch studies, makes both the economic and the ideological picture more varied and complex.

Various authors have argued for a diversified contribution of the different economic sectors (Gautier 1983; Peters 1986, 1991; Caneva et al. 1993; Caneva and Gautier 1994; Chaix 2003). We argue that the observed variability in economic activity is strongly linked to site location, to the carrying capacity of the different catchment areas, and to the seasonality of the available food resources. The wide spectrum of exploited resources (see also Table 3) does not justify the univocal and sometimes naive narratives linked to the pastoral hypothesis, which seem the output of an ideologically driven approach that perversely decides to ignore some data and privilege other data. In fact, given the presence of complete grass spikes at Kadruka 1 cemetery and a complete grass inflorescence at R12 (Madella et al. 2014), it seems untenable to state that domestic cereals from the R12, Ghaba and Kadruka 1 cemeteries, which cover the end of the 6th and the entire 5th millennium cal BC (Madella et al. 2014; Ryan et al. 2016), are the result of trade activities by people in Upper Nubia and central Sudan with the Levant (as argued by Hildebrand and Schilling 2016, p. 85).

Ideological constructs, as the output of social negotiations aimed at stabilising community social order and identity, vary according to the different roles played by distinct economic sections within a community.

Animal sacrifice is widespread in antiquity and also in modern and contemporary societies, and it is often associated with funerary contexts (as summarized by Russell 2012). The symbolic use of bucrania is attested to in many Neolithic and post-Neolithic societies that can hardly be defined as pastoral. Painted and three-dimensional bucrania are used at Çatalhöyük, in central Turkey (Mellaart 1967), in a context based ‘on domesticated cereals and pulses, as well as domestic sheep and goat, but [also] on wild cattle, boar, deer and equid’ (Hodder and Meskell 2010). At this and other Neolithic Anatolian sites, the bucrania are from wild aurochs (Henton 2013). Paintings and low-relief carvings of bucrania are common in the Neolithic cultures of Sardinia (agro-pastoral context) (Robin 2014; Tanda 2015), as well as in Neolithic and post-Neolithic contexts in central Europe (Bradač 2005) and Greece (Souvatzi 2008) and on the Minoan sites (Loughlin 2000). In Early Dynastic Egypt, bucrania were used to adorn the external platforms of First Dynasty graves (van Dijk 2013; Wengrow 2006).

Moreover, new archaeobotanical data and a different approach to the ideological and symbolic materiality embedded in the funerary program of Neolithic communities of Upper Nubia and central Sudan rule out pastoralism as a descriptive economic category and suggests agro-pastoralism as a more workable and fitting description of the period. Moreover, hunting and fishing practices are well attested from 5th millennium cal BC Neolithic sites, both in Upper Nubia (Honegger 2006) and in central Sudan, not only by archaeozoological remains (Table 3; Peters 1986; Linseele and Zerboni 2017), but also by fish-hook production, as attested to at Shaheinab (Arkell 1953) and Nofalab (El-Anwar 1981).

The argument more often used to define the Neolithic communities of Upper Nubia and central Sudan as pastoral is that these settlements are rare and small in Sudan (Wengrow 2001, pp. 95–96, 2006, p. 64). A survey of the literature, however, paints a very different picture (see online Appendix C). In spite of the large number of Neolithic settlements and possibly camp-sites that have been located both in Upper Nubia and in central Sudan, data are still scarce, and they do not allow us to understand the pattern of their occupation (whether seasonal or permanent) or their internal organisation and structure. What can be said is that they are mostly distributed along the Nile river, in areas suitable for both agriculture and animal breeding, together with other activities, such as hunting, gathering and fishing.

The Symbolic Use of Bucrania

The above analytic survey of the Upper Nubian and central Sudan Neolithic evidence aimed to indicate that the use of the label ‘pastoralism’ as a proxy for the economic foundation of the Neolithic populations living along the Middle Nile might need to be reconsidered. While the new evidence from R12 and Ghaba points more to an agro-pastoral agenda, additional data is required to unravel the complexity of the adoption of this food-production strategy, by looking at local differences and their determinants after the introduction of a complete or partial Neolithic package (Halstead 2011). The funerary practice of placing bucrania in graves cannot be interpreted in the same way from north to south and from one time period to another, because, as shown in Table 1, variability is very high. In Upper Nubia, the social meaning of the practice may be understood as a way to reaffirm cattle as a marker of wealth and/or status, rather than as a manifestation of a wider socio-economic strategy. Likewise, the funerary practice of placing vegetal pillows made of barley and wheat under the deceased’s skull suggests a similar reaffirmation.

This is clearly shown in the case of the Ghaba cemetery, where funerary practices place a major emphasis on cultigens rather than cattle, and where there is a clear segregation between the two in the graves. Only a minority of the deceased were accompanied by bucrania. This variability is also evident in the later Shaheinab phase at Ghaba, where bucrania are no longer attested to in graves, as well as at Kadero and the smaller contemporary cemetery at al-Khiday, where both ideological markers also disappear. However, the importance of cereals, at least at al-Khiday, is known from isotopic and dental calculus analyses that prove that people were eating C3 and C4 cereals and tubers (Iacumin 2008; Buckley et al. 2014). While the story of the economic systems adopted by the Neolithic societies of the Sudanese Nile valley has yet to be fully reconstructed, funerary archaeology illustrates, at a more general level, how life in the second half of the 5th millennium, both in Upper Nubia and in central Sudan, was the outcome of a consolidation period producing a regional, non-verbal communication system (mainly based on pottery and decoration types) that brought together diverse local communities (Salvatori and Usai 2016). This consolidation process, as suggested by the homogeneous distribution of the Multaga phase pottery in Upper Nubia and the Shaheinab pottery and decoration types of the Shaheinab phase in central Sudan, marks two distinctive regional and cultural spheres, maintaining noticeable differences between the individual communities of the respective areas.

For later periods, we can note that in central Sudan (the Shendi area), bucrania and possibly vegetal pillows have been found in graves at Kadada cemetery C (at the beginning of the 4th millennium cal BC). At the later Kadada A and B cemetery, not a single bucranium was found in the 92 graves, while there is (according to the excavator’s descriptions) conspicuous evidence of grinding stones and the mention of possible vegetal pillows. The much smaller, contemporary cemetery of Geili provides no evidence of bucrania or vegetal pillows. This again highlights a high degree of variability in the use of bucrania in graves, and is further evidence that the presence or absence of bucrania should not be used as an indicator of subsistence strategy. As noted above, there is a gap in our knowledge concerning 4th millennium cal BC Neolithic cemeteries from Upper Nubia.

The interpretation of the symbolic meaning of grave goods would be best approached through regional studies, as suggested by the interdisciplinary study provided for the Holocene central Sahara (Tafuri et al. 2006; di Lernia and Tafuri 2013). The many different sides of cattle symbolism are well evidenced, for instance, at the Neolithic site of Çatalhöyük (Hodder 2010), where wild cattle symbolism is used only until the appearance of domestic cattle, as also happens at the older, PPNA, sites of Göbekli Tepe and Nevali Çori (Peters and Schmidt 2004). To project the same meaning and symbolic value onto burials of partial or complete domestic cattle, cattle representations, and bucrania in burials would be inconsistent with the high and conflicting variability in time and space that emerges upon closer examination of the evidence at a regional level. Several groups can be labelled as pastoral in northern Africa during the Holocene, but the social construct of the ideological and symbolic use of cattle in graves or in other contexts is not—barring a few exceptions (as noted by di Lernia 2006)—a marker of a pastoral socio-economic organisation.

Finally, it seems misleading to look at cattle symbolism as a unifying trait among the Neolithic populations along the Nile valley from Egypt to Sudan (Wengrow 2001). And notwithstanding the role of cattle in the Egyptian dynastic world, it would be hazardous to define that society as pastoralist. Similarly, the exhibition of bucrania in Nubian Neolithic graves cannot be equated with the exhibition of bucrania in élite graves at Kerma since the Middle Kerma. The different scale of the phenomenon and the rather diverse presence in the Neolithic graves points to separate social and ideological processes.

The Kerma evidence points to a shift in emphasis, from a society based on power and wealth—possibly within a strongly hierarchical chiefdom structure—to a state organisation, which reached its mature form during the Middle Kerma period. Elite burials from this period were marked by multiple bucrania encircling the tumulus, in one case c. 5000, accompanying Grave 253 (Chaix et al. 2012). Whereas the bucrania in the Neolithic graves surely represent cattle sacrifice restricted to the funerary programme, the bucrania placed around the Kerma tumuli can be better interpreted as the evidence of food sharing engaged in by leading individuals on different occasions throughout their lives. This interpretation offers a better explanation for the problem, recently highlighted by Emberling (2014), of the provenance of the enormous number of bucrania placed around elite graves at Kerma. Indeed, the conservation of bucrania of past feasting occasions provided by elite families is reminiscent of the practice recorded by Webb Keane (1997, 2010) on West Sumba.

Within this general pattern, the funerary practice of placing bucrania in élite graves changed significantly along the Kerma sequence (Chaix et al. 2012).

A large number of bucrania were placed outside the largest Middle Kerma graves, and a much smaller, though still impressive, number were placed outside the Classic Kerma elite graves, and special funerary buildings highlighted the status of the most prominent individuals.

The bucrania in the Classic Kerma graves were almost certainly a marker of the very high status of the buried individuals and of the relationship with socially and ideologically constructed correlates of wealth and power. A similar evolution, from the Middle to the Classic Kerma period, has been reported for the great Kerma cemetery at Saï Island (Gratien 1975, 1981). It seems interesting to underline that this practice has not been observed for the Kerma period cemeteries in the Northern Dongola Reach (Welsby 2012), or at the contemporary cemeteries of Wadi Um Rahau (Kołosowska et al. 2003), El-Gamamiya 19 and 55 (Osypiński 2010) and Al Widay I (Emberling and Williams 2008, 2009), in the Fourth Cataract region, northeast of Karima. These differences may suggest that this practice was restricted to the ruling class of the Kushite kingdom at Kerma itself.

But whereas the symbolism attached to bucrania in the Kerma graves appears almost straightforward, the relationship with the economic base of Kerma society is much less obvious. The data from the settlement, in fact, show that the meat diet of the capital’s population changed from predominantly cattle to ovicaprines, while ‘the consumption of cereals is also testified by the numerous carbonised barley seeds and by very abundant implements for grinding’ (Chaix 2001, p. 364). Furthermore, the importance of agriculture in the Kushite state is attested to by the presence of villages mainly involved in agricultural practices (Gratien 1997; Bracco and Gratien 2002).

Conclusions

The analysis of the archaeological evidence pertaining to the neolithisation of Upper Nubia, central Sudan and adjacent regions illustrates the presence of domestic cattle, sheep and goats since the very end of the 7th or the beginning of the 6th millennium cal BC in the area from Sodmein Cave and Kerma and northeastern Africa and the presence of domestic cereals in the same area at least from the second half of the 6th millennium cal BC, if not before. This defines the period when the local communities made the transition from a foraging to a food-producing economy. An accurate assessment of faunal-remains data from Neolithic settlement sites (Tables 2, 3) describes a meagre picture that limits our capacity to distinguish specific cultural trajectories—pastoralism, agro-pastoralism, agriculture—resulting from diverse ways human communities react to and interact with environmental and climatic conditions and variations, as well as geomorphological peculiarities. This picture forces us to rely on the more eloquent data provided by the Neolithic funerary contexts. Highlighting the frequently neglected wide variability of the funerary ideology, as evident from the 5th and 4th millennia cal BC, not only changes the perception of the Sudanese Neolithic from pastoral to agricultural, but also warns us about the danger of generalizing and reading ideological elements (including bucrania and vegetal pillows) as economic markers, disregarding the variability embedded in the meaning of the symbols used in the historical process of identity building.

The presence of bucrania in the graves (although their frequency varies considerably from one cemetery to the next), when analysed in the Neolithic and Bronze Age contexts in the Nubian region (the 3rd millennium cal BC being attested, at present, only in the Kerma area), displays a clear funerary ideological meaning, with differences among the individual local communities both during the 6th–5th and 4th millennia cal BC and during the 3rd millennium cal BC. In the Kerma State we observe a drastic shift in the use of bucrania, which are now exhibited in large number around the tumuli of dominant individuals, rather than being confined inside the graves, as in the Neolithic.

If we consider funeral rites as a complex, multifaceted social act with a strong symbolic meaning at different levels of individual, family, group and community social identity, and as the place where ‘practices, concepts, beliefs and orientations’ (Hodder 2006, p. 242) are reaffirmed and social role and rank are renegotiated, then we have to pay attention to the symbolic meaning of the different material categories and, therefore, to animal bones placed in graves. It is important to note that the bucrania deposited in the graves of the Neolithic cemeteries of Upper Nubia and central Sudan would have been visible only during the funerary event, and that such a practice is very different from the placing of bucrania outside the grave, as on the benches of Early Dynastic Egypt elite graves (van Dijk 2013; Wengrow 2006) or around the tumulus, as in the Early and Middle Kerma period graves at Kerma and Saï (Chaix 2001; Chaix et al. 2012; Gratien 1975).

The bucrania in the Neolithic graves of R12 and Kadruka 1, which would have been invisible once the grave was covered over, may have conveyed a meaning, strictly confined to the deceased’s household, of differential spending capacity for the funerary program, in the form of the killing of one or more economically valuable animals. In addition, the household derived ideological and social betterment from the redistribution of meat during the funeral. The presence of vegetal pillows containing, in the case of R12 and Kadruka 1, almost exclusively wheat and/or barley urges a more complex reading than the purely symbolic value that this practice might suggest. The introduction into the symbolic narrative of domestic cereals, which directly recall agricultural activities as a valuable practice embedded in the production system of the local community, did not necessarily result in competition with the symbolic narrative of animal breeding, as demonstrated by the co-occurrence of bucrania and vegetal pillows in some graves. Moreover, the absence of bucrania does not mean that cattle were absent from the economic realm; in this strongly symbolic context, the absence of bucrania can only mark an ideological shift to cereal-growing practices as a substantial economic base for the community. It must also be underlined that symbols are subject to transformation and can combine to transform the ideological and social structure (Hodder 2010). Nevertheless, any local community, although sharing common, well-defined traits at the higher level of supra-community identity, develops or maintains an ideological construct that differentiates that community from others.