FOUR THOUSAND YEARS IN THE BLUE NILE:

Paths to inequality and ways of resistance

Víctor M. Fernández

Departamento de Prehistoria. Universidad Complutense. 28040 Madrid, Spain.

 

 

 

 

      Abstract

  1. On the origin and meaning of the Mesolithic cultures.

  2. The Mesolithic period in the Blue Nile region.

  3. The transition to a Neolithic economy.

  4. The gap in the archaeological record: retreat or resistance?

     Bibliography

 

 

 

ABSTRACT

As general conclusions of the papers of the Blue Nile Project dossier herein, some ideas are suggested about the prehistoric societies investigated by Spanish archaeologists in Central Sudan during the 1990’s. The project included a survey of the Wadi Soba-El Hasib region east of Khartoum and excavations of two Mesolithic sites and one Neolithic site in the Wadi Soba area. Data from different sources are combined in an attempt to construct a historical narrative. Vestiges of some cultural hiatuses were noticed in the region, namely at the beginning and the end of the Mesolithic period, the latter involving the emergence of social stratification and the decline of women status. The archaeological gap at the end of the Neolithic period is interpreted as being a consequence of hindrance to social division.

 

Early cultures of resistance and population movements towards the Ethiopian escarpment as a refuge area are proposed as longue duréeprocesses among Nilo-Saharans of the Eastern Sahel.

 

 

1. On the origin and meaning of the Mesolithic cultures.

The Mesolithic sites surveyed and excavated in the Blue Nile area by the Spanish archaeological project, whose study results have been presented in this volume (Fernández et al. 2003a, 2003b), belong to a Holocene Saharan-Sahelian “technocomplex” called in several different ways. An “archaeographic” successful denomination has been “Khartoum-Horizon Style” (Hays 1971: 134) which refers to the uniform pottery decoration that was first discovered at the Khartoum site (Arkell 1949a). The idea is that the same pottery style was diffused to many different groups in the Sahara and Sahel, whose previous existence before the pottery diffusion is induced from their different lithic technologies (Hays 1971: 136). Recent archaeological data reveal the existence of a short “Pre-ceramic” or “Epi-palaeolithic” phase in the Sahara, chronologically placed in the interval from the onset of humid conditions and the spread of pottery techniques. Evidence has been collected at northern Mali and Niger (Camps 1974: 214-6), the Egyptian eastern desert (Gabriel 1977), the Eastern Sudan (Marks 1987; Elamin 1987) or the West Lybian Sahara (Cremaschi and Di Lernia 1998; Garcea 2001). In the latter region the cultural phase, called “Early Acacus”, is dated to c. 9800-9000 bp. The lithic industries of the period are characterized by a high frequency of backed bladelets as it is also the case in contemporary industries of the Maghreb (Capsian) (Tixier 1963; Camps 1974) and the Egyptian Nile valley (Qaruniam, Shamarkian) (Wendorf 1968; Wendorf and Schild 1976).

 

Around 9000 bp, however, human groups all along the Sahara adopted the pottery production with a strikingly analogous decoration style that shows hardly any differences over so a wide region. The same rocker impression technique (RK) was used from the Atlantic areas, e.g. in the Western Sahara (Almagro 1946: 200-1) or Mauritania (Commelin et al. 1992: fig. 3, a-f), to near the Red Sea coast, as in the Saroba phase of the Khashm el Girba area (Fattovich et al. 1984: fig. 3; cf. Arkell 1949a: 116; Garcea 1993: 189). The technique of alternately pivoting double stamp which is a variant of the general rocker impression system, is also found throughout the Sahara and the Nile (Garcea 1998: 93). Dotted wavy line (DWL) was also widely spread over the desert area, but it appeared in the Nile valley at a later date (Ibid.; Caneva 1983; Caneva and Marks 1990; Caneva et al. 1993).

 

The reasons for this quite rapid process of expansion of one single cultural item over such a vast area (more than 6000 km wide) still await elucidation. Since in most of the known sites the pottery horizon is the first cultural context after the Late Pleistocene demographic hiatus, the problem of pottery origins is actually the problem of the human re-occupation of the Saharan area. Comparable cases of rapid ceramic diffusion, such as the Linear Danubian pottery in Early Neolithic Europe (Whittle 1985) or the Bantu Urewe Ware in the Early Iron Age of Eastern Africa (Phillipson 1977a) have been interpreted as the result of the expansion of human groups throughout a wide geographical area.

 

In two widely circulated and classical articles, J.E.G. Sutton (1974, 1977) interpreted the rocker pottery techno-complex as the first migration of Negroid peoples, speaking ancient variants of the current Nilo-Saharan language phylum, outside their homeland in Central Africa (the second would be the Bantu expansion, several millennia later). The expansion was propelled by the onset of newly humid conditions in the Sahara, and the migrant groups consequently adopted an economic orientation based on river and lake resources, hence the name “aqualithic” proposed for the cultural complex. Sutton’s belief that the movement was also pushed by the expansion of tropical forest in Central Africa (Sutton 1977: 27), seems to be confirmed by more recent data. After being confined to a few “sanctuaries” during the maximum glacial conditions at the end of the Pleistocene, the rainforest expanded to reach its highest extension about the mid-Holocene period (Lieth and Werger 1989). According to a well-known yet controversial theory (Bailey and Headland 1991), tropical forest was an inaccessible niche for humans before the establishment of a Neolithic technology. Therefore, foraging groups would have reacted to the forest growth by migrating towards more open areas and in the band immediately north of “Middle Africa” this direction would be northwards.

 

As for the physical anthropological features, the idea of a fundamentally Negroid stock spreading through the Sahara during the Holocene, which was common when Sutton wrote his articles after data on single skeletal variables from Early Khartoum and other sites, (cf. Derry in Arkell 1949a: 30-3), was later rejected on the basis of multivariate data supporting the presence of northern Mechtoid populations (Petit-Maire and Dutour 1987). Recent investigations based on genetic variables have shown, however, that both in the Nile and the desert areas there is currently an almost continuous gene variation between Negroid features prevailing in the south and Caucasoid in the north (Tay and Saha 1988; Fox 1997). Irrespective of this fact being interpreted as coming from gene flow or local selection, the scenario is one of gradual variation throughout the area. The earliest rock art images of the Central Sahara, the “Round Heads” style roughly contemporaneous to the Mesolithic period, show human figures usually interpreted as “Negroid” (Sansoni 1994; 1998). In the following, Bovidian period dated after the beginning of animal husbandry in the VII millennium bp, the presence is attested of different racial types in the region, Negroid being allegedly earlier than the Europoid figures (Muzzolini 1986; Gallay 1987).

 

From a linguistic point of view, the relations between the current Nilo-Saharan languages and the Saharan-Sahelian expansion have been but strengthened by several authors in the recent years (Bender 1982; Blench 1993: 136; Ehret 2000: 281-9). Archaeologically, the southern origin seems also reinforced, as the great antiquity of bone harpoons (the other fossil directeur of the complex, besides pottery) and generally of the aquatic economic adaptation has been recently reported for Central Africa, with the early dates for the R.D. Congo sites such as Ishango and Katanda (McBrearty and Brooks 2000: 503-6, 510-13). This greater age appears to match and even supersede the alleged significance of the Nile valley or the Eastern African lakes as other areas of origin for the Saharan Holocene cultural events (Stewart 1989).

 

There have been quite a number of theories on the causes of pottery invention and its first functions. Most authors have insisted on the new possibility of processing food by boiling and steaming that render meal more digestible and palatable (e.g. Haaland 1992: 48). Among the Mesolithic groups of the Sahara and Nile Valley the meal could have consisted of gathered plants (Ibid.), mostly cereals whose seeds have been found in some archaeological deposits (Barakat and Fahmy 1999) and impressions were recorded on sherds from several excavated Mesolithic and Neolithic sites (Magid 1989, 1995, 1999, 2003; Stemler 1990). Also the fish has been naturally proposed as being processed in the form of stews or soups in the pottery vessels (Sutton 1974; Stewart 1989; Haaland 1992, among others). Plant gathering and domestic pottery making have been usually interpreted as female activities, on the basis of widespread ethnological data (Murdock and Provost 1973: table 1). About fishing, information exists that today is mostly, and thus was probably in the past, a quasi-masculine activity up to 82% of the African study cases (Ibid.: table 3). Men perform most of the current fishing activities in the Sahel and the Lakes areas, though some cases are known of female participation as among the Nuer (Murdock 1967: 188). Nonetheless, in some of the more traditional groups of the Ethio-Sudan border formerly called “pre-nilotes” (see map in Fernández et al. 2003a: fig. 1), such as the Gumuz, Meban or Koma, fishing is a predominantly female task (Grottanelli 1948: 300; Cerulli 1956: 18-9).

 

The contribution of women to fishing being only probable (Barich 1998: 108), their nearly certain association with pottery and food-plants makes a good case for an important female role in the Mesolithic expansion. A broad system of women exchange marriage has been advanced as a possible explanation for the similarities of Saharan pottery decoration (Caneva 1988b: 369). If decoration and generally stylistic behaviour may be considered as a system for information display, aimed at a target population group that both need and can decode the messages (Wobst 1977), then the symbols embedded in the pots could be “read” from the Atlantic Ocean to the Red Sea. One is tempted to imagine one single shared ideology for all that vast area. A distant glimpse of this ideological domain can be caught in the strange scenes of the Round Heads style of rock art from the Central Sahara. Here male figures outnumber women but these appear often playing a prominent role (Sansoni 1994: 208, 1998: 149). Sometimes women are represented next to hemispherical containers with probable seeds, or dancing in “worship” scenes, their bodies richly decorated with motifs (scars?) that remind the ceramic decoration repertoire (Barich 1998: 112-3). Regular association of motifs on pots and on the human body has been reported in Africa, as also the possibility that scarification was one of the primary arts of the continent (Barley 1994: 128-32; Rubin 1988: 15). Moreover, early pottery excavated at Nabta in the Egyptian western desert has been ascribed a social and symbolic function because of its scarcity (Close 1992: 162-3), and the long tradition of Sudanese finely decorated pottery that continues until the Meroitic and even the Christian period, has been also interpreted as evidence of its probable ritual purpose (Edwards 1996: 74-5).

 

Women and children are clearly over-represented in the meager funerary evidence known from the Mesolithic period. At the Sudanese site of Saggai, four out of five excavated burials belong to women (Coppa and Macchiarelli 1983: 118-22), and the few graves excavated in the Sahara, for instance at Uan Muhuggiag (Lybia) and Amekni (Algeria) belong also to women and children (Barich 1998: 111). At the Egyptian Neolithic site of Merimda Beni Salama, the exclusive internment of women and children has been interpreted as a probable indication of matrilineality (Hassan 1988: 169). In Nubia, women also played an important role in Early Neolithic society, as it is evidenced by the almost exclusive female and children burials in El-Barga cemetery near Kerma (Honegger 2003: 289; pers. comm.) and the prominent location of some female graves in the Kadruka 18 cemetery (Reinold 2000: 80-1, 2001: 6). Sometime later, data from the small Kerma cemetery of Abri in Northern Sudan (Fernández 1982: 289-302), where women graves are more richly furnished than men burials, suggest the probable persistence of female status in the rural areas far from the power centers such as the Kerma capital itself.

 

2. The Mesolithic period in the Blue Nile region.  

The survey has revealed very few data on the Palaeolithic period in the region. The few Middle Palaeolithic remains found suggest that the area was not totally uninhabited, yet they do not add much to what was already known from other few localities, such as Singa or Abu Hugar (Arkell 1949b: 45-7, pl. 27: 5-7). No Upper Palaeolithic sites were recognized, though lithic tools considered as typical of that period are more abundant in the Mesolithic sites of the wadi at east than in the river ones, maybe as a result of cultural influences from Eastern Sudan where Late Palaeolithic industries have been recorded (Elamin 1987; Marks 1987).

 

Geochemical evidence from the basal levels of El Mahalab (EM) site suggests that climate was wetter before 8000 bp than in the following years (Lario et al. 1997). It seems that the whole research area was almost empty before the eighth millennium, probably because of frequent Blue Nile floods and the formation of swampy areas close to the riverbed and in the wadis (Wickens 1982: fig. 6). Significantly, most dates from other Mesolithic sites in the Khartoum region are also younger than that date, the few earlier ones (from one single site, Sarurab, cf. Khabir 1987) being not very reliable as probably not associated with the cultural remains (Caneva 1999: 33). Six dates from Abu Darbein near Atbara in Eastern Sudan are between 8640 and 8330 bp (Haaland and Magid 1995: 49), this being perhaps another indication of the pre-eminence of eastern influences on the area. Dates from the scarce sites known south of Khartoum hint at an even later date for the Mesolithic adaptation: e.g. 7470-7050 bp at Shabona (Clark 1989: 389). Furthermore, not one single important Mesolithic site was found during our exploration of the Blue Nile area from Wad Medani to Singa (Fernández et al. 2003a: sites nos. 85-6, 92). Several Mesolithic-like sherds were found by the author in a recent excavation in Ethiopia near the Sudanese border (Bel Kurkumu rock shelter in Assosa, Benishangul), in a level dated to c. 5000-4500 bp. A slightly earlier chronology has been proposed for the few wavy line sherds discovered in the Lake Turkana basin (Phillipson 1977a: fig. 16, 10, 1977b: fig. 19, 3). All this evidence, albeit scanty, hint at a Saharan rather than a Central African origin for the Early Khartoum culture, which thus probably begun at a later date than in the Mesolithic core area (cf. Close 1995: table 3.1).

 

What clearly distinguishes the Khartoum region, when compared with neighboring areas, is the abundance of incised wavy line (WL) over other decoration types, namely rocker impression which is characteristic of other regions. WL is also found in other Nilotic areas, but quite less frequently: 4% at Shabona (Clark 1989: fig. 12), 11% at Abu Darbein (Haaland 1995: 113) and around 16% in the Dongola reach (Shiner 1971: 141). Some of our surveyed and excavated sites, such as Karnus or Sheikh Mustafa (SM), have percentages of WL amounting to more than 60% of the pottery sherds (see Fernández et al. 2003a: table 5). A high frequency has also been recorded at the Early Khartoum site (Mohammed-Ali 1982: 76). Our seriation results further suggest that even earlier sites producing only WL pottery without the rocker variety could exist and be found in the future (Fernández et al. 2003a: fig. 46). That WL technique was invented in this region can be confidently postulated and perhaps it is no coincidence that the undulating lines originated in the best-watered region of all the aqualithic complex extension. Later, the same symbol was used to represent water in the Egyptian hieroglyph script (Wilkinson 1992). The ensuing gradual substitution of RK for WL as the main decoration technique, which is evidenced by seriation and stratigraphical data both in the Blue and the main Nile, may be a reflection of progressive cultural influences from the Saharan area. Also the early arrival of DWL pottery to the Nile, attested at some of our sites, suggests that Saharan connections existed during most of the Mesolithic period. Culturally, the Central Sudanese area ended up by loosing its originality and integrating itself in the larger desert region. A few of the sites discovered in the survey can be associated with a later phase of the Mesolithic period in the region, characterized by the vanishing of WL pottery and the abundance of RK and DWL types (Caneva and Marks 1990: 21-2; Caneva et al. 1993: 247-8).

 

In the light of the distribution of Mesolithic settlements over the Blue Nile landscape, a model of seasonal movements between the river and wadi areas has been inferred. Probably as the Nuer in recent times (Evans-Pritchard 1940), the groups moved towards the river and split up into small parts at the beginning of the dry season, gradually concentrating on the last available water sources at the end of the season. Small and big archaeological sites recorded in the riverine area could correspond to camps at the beginning and end of that period. More permanent villages seem to have been erected during the rainy season and the river flooding, when people would leave the alluvial plain and move to the wadi area where elevated land made settlements more feasible. Analysis of fish remains (Chaix 2003) and pollen (López and López 2003) indicate the proximity of deep waters at the wadi site (El Mahalab). Also the ceramic seriation and settlement patterns (Fernández et al. 2003a: section 6) agree with the model, which had been already proposed on the basis of ethnographic analogy (Clark 1989: fig. 14).

 

The copious material inventory found at most of the sites suggests that groups came every year to the same spot, where they probably kept safe part of their material paraphernalia, for example the heavy stone grinders, when they moved to the new camps and villages. Perhaps they also abandoned some of their pots, and that is why we now find so many sherds on the sites. Some 150 sherds per cubic meter of archaeological deposit were recorded at Sheikh Mustafa, and even more, 275 sherds per cubic meter at El Mahalab. The mean size of the sherds at Sheikh Mustafa site is 7.3 cm2, which represents a mere 0.4% of the total area of a hemispherical bowl with 35 cm in mouth diameter (approx. mean value, see Fernández et al. 2003b: table 4). As an average value, then, each pot broke in 250 fragments. Could it have been the result of a deliberate process? As Nigel Barley puts it, “in Africa death involves the breaking of pots while marriage involves making them” and it is the very friability of pots which makes them “a source of ritual power” (Barley 1994: 92, 112).

 

Besides the aforementioned synchronic differences, historical trends can also be inferred. Multivariate statistical analysis of in-site artifacts distribution has allowed to compensate partly for the deflation processes operating since Prehistoric times, and some of the original stratigraphic array could be reconstructed linking data from the central and peripheral areas of each site (Fernández et al. 2003b: fig. 11, passim). Results of ceramic seriation offer some evidence of a gradual shift of settlements from the wadi to the river towards the end of the Mesolithic period, possibly influenced by the climatic deterioration at the time. Faunal remains also give some clues about a general humidity reduction during the Mesolithic period. Thus, fish bones are more abundant in the lower levels excavated at Sheikh Mustafa, dated to c. 7930 bp, than in its upper levels and in El Mahalab, dated to 7705-6940 bp (Chaix 2003). The tendency is further confirmed Four thousand years in the Blue Nile: Paths to inequality and ways of resistance by a sandy level in the upper levels of El Mahalab site indicating arid conditions sometime between c. 7400 and 6900 bp (Fernández et al. 2003: section 1).

 

The lithic material analysis in Sheikh Mustafa and El Mahalab also shows a change from many backed points and few lunates to the opposite, many lunates, especially broad types, and few backed points. The trend continued during the Neolithic period when broad lunates are predominant. The change has been interpreted as related to climatic change and consequent game availability. Many of the narrow backed bipointed bladelets from the earliest times, particularly frequent in the lower levels of Sheikh Mustafa, could have served as fishhooks, since bone harpoons have not been found. Both the faunal analysis (Chaix 2003) and the palaeodietary analysis of human bones (Trancho and Robledo 2003) from the same level indicate abundant fish consumption. Some points and narrow lunates were probably used as sharpened arrowheads, especially effective to kill big animals, while broad lunates were more efficient as chisel-ended arrowheads to hunt smaller and faster game (Clark et al. 1974; Nuzhnyi 1989). Faunal data from the three sites show a constant reduction in game size from the earlier to the later Mesolithic sites (SM to EM) and then in the Neolithic site of Sheikh el Amin (SA) (Chaix 2003; Fernández et al. 2003b: fig. 66).

 

3. The transition to a Neolithic economy.  

While it has been commonly held that the Khartoum Neolithic developed out of the Khartoum Mesolithic, there is a paucity of radiocarbon dates and archaeological information from the period of transition that has puzzled researchers for many years (Marks et al. 1985: 262-3). Later inquiry came to fill the gap in a certain way, with a few sites dated to the second half of the seventh millennium bp: El Qala’a and Kabbashi on the main Nile north of Khartoum (6620-6150 bp; Caneva et al. 1993: table 1) or the middle levels of Shaqadud cave in the northern Butana (Caneva and Marks 1990). One of the radiocarbon dates from the Sheikh Mustafa site falls within the referred period (6295 bp), but it contradicts the evidence coming from the rest of the accepted dates and the general earlier appearance of the material culture at the site.

 

A comparison of the ceramic seriation models proposed for the Mesolithic and Neolithic sites in our survey (Fernández et al. 2003a: figs. 46 & 56) shows that both at the end of the first period and the beginning of the second, the pottery decoration was predominantly based on the same technique, namely rocker impression. Notwithstanding, most of the other variables are different: the Neolithic pots are more finely made, with thinner walls, burnished and often slipped outer surfaces, and new smaller vessel types and decoration types appear just from the beginning, such as incision or black topped. The overall impression about pottery at hand is that of being of a quite different kind. Even the old rocker impression looks now mutated, combining different and finer comb-tools.

 

A significant shift also occurs in the settlement pattern: not one single important Mesolithic site was inhabited during the following period. Most, if not all, Neolithic settlements were occupied then for the first time. Although Mesolithic sherds have been found in some Neolithic sites (Arkell 1953: 68; Krzyzaniak 1978: 171), they come from small sites that were probably short-term camps such as those we have found during our research (Fernández et al. 2003a, fig. 48). Even if we take into consideration the case of the Sheikh el Amin Neolithic site, where our excavations have revealed a certain amount of Mesolithic DWL sherds all over the site, evidence of local transition between both periods does not seem to be at hand. My general impression is, then, that an important change took place in the region with the arrival of new groups with a different, livestock herding economy.

 

Climatic changes going on around 6000 bp represent the “great mid-Holocene arid phase” in the Sahara (Muzzolini 1995: fig. 30), which corresponds to the lower sea surface temperatures recorded after 5900 bp (Hassan 2002: 322) and the “Post-Late Neolithic arid phase” of the Nabta Playa record in the South Western desert of Egypt (Schild and Wendorf 2002: 24). This deterioration was perhaps the origin of a contraction of the aquatic economy that had been predominant in Central Sudan throughout the precedent millennia. In fact, only the riverside Neolithic settlements such as Shaheinab or Geili have significant fish remains in their faunal collection (Krzyzaniak 1978: 165; Gautier 1988). In Sheikh el Amin, though a shell fishhook was found during the excavation, fish remains are insignificant (Chaix 2003: table 12). When facing climatic deterioration, Africans are forced to choose “between their homes and their environment” (David 1982: 50), and the possibility that some of the Mesolithic people migrated further south where humidity was still high, cannot be ruled out.

 

The Lokabulo tradition, named after a rock shelter in Eastern Equatoria some 1000 km south of Khartoum and dated to 3800 bp, even though poorly known because of the research disruption caused by the recent war, presents some characteristics reminiscent of the Central Sudan Mesolithic (David et al. 1981). The pottery is quartz tempered and decorated mostly by rocker impression (though excavators only recognised the spaced zigzag pattern as such), including some DWL sherds (Ibid.: fig. 6, pl. 1). Peter Robertshaw warns against paralleling the same decoration technique (generally, comb impression) from distant areas, arguing that Lokabulo sherds differ from Jebel Moya ones and thus rejecting relations of southern and central Sudanese prehistoric cultures (Robertshaw 1982: 92). Nonetheless, central Sudan Mesolithic rocker pottery is also clearly distinct and earlier to the simple impressed Jebel Moya pots, and Robertshaw’s comparison of Lokabulo and Kenyan Kansyore pottery, from which a part appears to be also made by simple impression, seems neither substantiated (Ibid.: fig. 2). The faunal remains of Lokabulo consisted only of hunted wild fauna and the excavated deposits yielded quite a good number of mollusc shells, though they were devoid of fish remains (David et al. 1981: 11-19; David 1982: 52-3). Yet, linguistic data suggest the presence of a food-producing economy in Southern Sudan since the third millennium bc (Ehret 1982: 28) and thus contradict the scenario resulting from Lokabulo and other sites in the Eastern Equatoria, which implies a Later Stone Age hunting economy well into the first millennium AD (David 1982: 53). The more western site of Jebel Tukyi, with a younger date (2130 bp), produced large domestic cattle (Ibid.: 51) and rocker impressed pottery which Randi Haaland has identified as similar to the Khartoum Neolithic tradition (Haaland 1992: fig. 11, 61-2).

 

Some unpublished data from the recent Spanish research directed by the author in Benishangul, West Ethiopia, at half way between the Khartoum and Eastern Equatoria regions, are of possible relevance here. Excavations at several rock shelters near Assosa town yielded abundant quartz tempered, Mesolithic-like sherds with WL, DWL and especially RK decoration. As in the Lokabulo site, at the Bel Kurkumu rock shelter pottery appears in the upper part (radiocarbon dated to 4965-4470 bp) of a Late Stone Age level with an un-diagnostic flake and bladelet quartz industry with end scrapers and rare microliths. The same pottery types continue in the upper level, together with a similar industry yet with less formal tools, dated to 2020-875 bp. In another excavated shelter in the nearby, RK and DWL sherds appear together with different pottery types, dated to the beginning of the second millennium AD. The economy of these groups is not yet known since bones were not preserved in the shelters acid soils, neither plant remains were found in the deposits and the pottery sherds. A fragmented “net-sinker” in pottery (Haaland 1992: fig. 3) found in the shelter could indicate some fishing practices. Anyway, the persistence of old pottery types up to the first millennium AD recalls the Eastern Equatoria evidence, as also the archaic features of the “pre-Nilotic” peoples in the border region between Sudan and Ethiopia (Grottanelli 1948). Some of these traits, such as the relevance of plant gathering and fishing, lack of big livestock, matrilineal kinship remnants, incisor teeth extraction (Murdock 1959: 170-80; Bender 1975: 9-19) and even the racial morphological characteristics (Arkell 1949a: 114) remind to some extent of the Khartoum Mesolithic features.

 

The site of Sheikh el Amin shows important differences when compared with other known Neolithic sites in the region. First of all it is located in the Butana plain far from the Nile, and this means a savanna economic orientation with very little fish exploitation. Livestock also appears to have been of reduced relevance to its inhabitants, since faunal remains are mostly of hunted wild fauna (Chaix 2003: table 12). After the aforementioned crisis at the beginning of the Neolithic period, the climate became humid again as the faunal (e.g. Phacochoerus) and the vegetal remains (e.g. Carex, Celtis and Sorghum, see Magid 2003: table 1) suggest. Food plants exploitation seems to have been intensive in this site, where 30 plant impressions on pottery have been recorded. The proportion of sherds with plant impressions, nonetheless, is not higher but lower than in the Mesolithic sites, with respect to the total number of recovered and examined sherds (0.071% in SM, 0.085% in EM, 0.053% in SA). If the greater variation in plant species at SA is not an effect of the bigger pottery sample analysed (56761 sherds, as contrasted with 7001 in SM and 4680 in EM), or of any other factor concerning pottery making and plant impression processes, it would indicate a stronger orientation to gathering activities during the Neolithic times in the Butana plain. The large quantity of stone grinders excavated at the site, one of the biggest figures for the entire region, also hints at the same explanation. Their lower number at the later parts of the site possibly indicates the decreasing importance of plant exploitation when climate changed towards the current arid conditions in the Late Neolithic period. As regards the much debated issue of early plant cultivation during the Mesolithic and Neolithic (see recent arguments in Haaland 1996, 1999; Magid and Caneva 1998), the wide variation in species of our data, with 10 different plants identified in 39 pottery impressions, suggest a broad-spectrum exploitation of the environment, oriented to seeds and fruits, rather than a concentrated strategy on a particular cereal plant, even if sorghum is the most represented species as it also happens in other Sudanese sites (Magid 1989, 2003; Magid and Caneva 1998).