. 5
( 11)


variations, technological developments followed broadly parallel courses
Regional diversification and specialisation 143

throughout Africa -- as, indeed, in other parts of the Old World. The general
course of this development was outlined in the introductory section of this
chapter and need not be repeated here. Recent research south of the Sahara
has tended to blur the distinction formerly proposed between the stone-tool
industries of modes 3--5 in the Congo Basin and adjacent regions on the
one hand and, on the other, those in more easterly and southerly parts of
the continent. Likewise, the mode-3 Aterian industry of the Sahara is now
seen as having much stronger af¬nities with those of more southerly lati-
tudes than was previously appreciated. As chapter 3 has shown, the origins
of the prepared/radial-core tradition which is one of the de¬ning charac-
teristics of mode 3 can be traced far back into Acheulean times. Diversity
through parallel evolution, already apparent in the ¬nal Acheulean, seems
to have accelerated around 250,000--200,000 years ago and to have continued
throughout the time-span to which this chapter is devoted.
More truly innovative, perhaps, was the eventually almost ubiquitous
mode-5 backed-microlith technology which involved far more economical
use of raw material, and the facility to repair or modify tools without
resorting to their total replacement. Although direct evidence is lamentably
sparse, invention of the bow and arrow may have taken place around this
time, allowing much greater force and precision in the use of projectiles to
which, on occasion, poison may have been applied. In parts of North and
Northeast Africa, as in the Levant and elsewhere, this technology is best
and conventionally seen as a development from mode-4 blade industries
in which techniques for making backed retouch were frequently practised.
Further to the south, however, backed microliths appear in the Howieson™s
Poort industries at dates which are signi¬cantly earlier than those known
for comparable technology elsewhere in the world, and the roots of their
development may be traced to even earlier times in south-central Africa.
Archaeologists are still far from being able to explain why these parallel
technological developments should have taken place; simple diffusion from
a common source is not a likely explanation. In both southern and south-
central Africa, recent research has demonstrated the length and complexity
of the archaeological sequence of mode-3 and mode-5 industries. The very
early dates proposed for Border Cave and Twin Rivers strongly suggest that
the post-Acheulean/Sangoan technological traditions began at an approxi-
mate time-depth between 250,000 and 200,000 years ago. Their emergence
may have taken place regionally at different times, as was demonstrably the
case with the later development of mode-5 technology.
In recent years, particular emphasis has been attached to seeking archaeo-
logical evidence for trends in the development of human culture and
behaviour which might parallel the evolution of anatomically modern
people illustrated by current osteological and genetic studies (Crow 2002).
144 afric an archaeolog y

There is no clear agreement among archaeologists as to what might com-
prise a list of ˜modern™ behavioural elements, or how certain factors might be
selected for inclusion on such a list. Elements often proposed include: tech-
nological specialisation; increased selectivity and economy in the use of raw
materials; symbolism as exempli¬ed, for example, in art, pigment use and
the application of personal adornment; and, for reasons never adequately
explained, the working of bone and the exploitation of aquatic resources
(Stringer 2002a). The list was initially based on the comparatively well-known
archaeological record of Europe, taking elements which were represented
in Upper Palaeolithic contexts associated with the remains of ˜anatomically
modern™ humans but absent from earlier contexts (Mellars 1993, 2002). Sub-
sequent research has tended to blur this distinction, and certain criteria
(notably the working of bone and the use of pigment) have been tacitly
modi¬ed. Attempts to apply this schema in other parts of the world have
cast further doubt on its validity (cf. Wadley 2001a). Some researchers (e.g.
Willoughby 1993; Dickson and Gang 2002) have uncritically correlated the
development of microlithic technology per se with ˜modern™ behaviour. There
is wide acceptance that language development must have played a crucial
role in the onset of cultural modernity, but little agreement as to how its
stages are to be de¬ned or how (if circular argument be eschewed) they
might be recognised in the archaeological and palaeontological records. In
view of the uncertain validity of the criteria employed for its recognition,
it is hardly surprising that there is controversy over whether the advent of
cultural modernity was a single event, perhaps linked to a genetic mutation
(Klein 2000a), or whether -- as here advocated -- it was a more gradual process
(Stringer 2002b).
The objection has been made that the list of criteria is based primarily on
evidence from Europe, which is now seen as a region peripheral to the devel-
opments under investigation. It is also questionable whether the ill-de¬ned
concept of cultural modernity is appropriately tracked by the simple expe-
dient of checking individual items on a list which has itself been subject to
gradual uncritical modi¬cation. Be that as it may, the African archaeological
record surveyed in this chapter provides crucial evidence both that these
developments took place signi¬cantly earlier in Africa than in Europe and
also that their appearances were results of gradual processes. The view that
it was speci¬cally in southern (H. J. Deacon and Wurz 2001), or southern
and eastern, Africa that cultural modernity ¬rst arose is less well founded,
and may be due to the fact that archaeological research in these areas has
been much more intensive than has been the case elsewhere. A broader
pan-African view is taken by McBrearty and Brooks (2000), who also argue
convincingly that this was a long, slow development, as was in fact foreseen
in the ¬rst edition of this book (D. W. Phillipson 1985b). Such a scenario is in
Regional diversification and specialisation 145

accord with the increasing body of genetic evidence for multiple movements
of anatomically modern or near-modern people from Africa over a substan-
tial period (Lahr and Foley 1994; Quintana-Murci et al. 1999; Milliken 2002;
Templeton 2002) which, it is now generally agreed, better ¬ts the available
fossil evidence than either the old multi-regional model (Wolpoff et al. 1994;
cf. Lahr 1994) or the single-migration view (Stringer 2001). It also accom-
modates the geneticists™ requirement for relatively small source-populations
without resort to dramatic bottlenecks of the type hypothesised by Ambrose
(1998a). This problem remains a major focus of current research.
The cultural signi¬cance of the blade-based technology which charac-
terises the mode-4 industries of the European and Levantine Upper Palae-
olithic has been evaluated by Bar Yosef and Kuhn (1999), while Ambrose
(2002) has undertaken a similar exercise for mode-5 microlithic technology.
The signi¬cance of bone tools is less clear: the original distinction, based on
European materials, emphasised the ¬rst appearance of elaborate carving of
bone in the Upper Palaeolithic, but this distinction has been lost or forgot-
ten when the criteria have been applied to much older African materials, as
at Blombos (p. 102 above). The remarkably close similarity of African barbed
bone harpoons over a very wide area and, it now seems likely, through a
long period of time, has been noted by Yellen (1998). Such artefacts have
been subject to varied interpretations: to evaluate technological abilities, to
reconstruct activities (often in conjunction with the study of non-artefactual
residues), to trace typological similarities or contrasts, and as hallmarks of
˜cultural modernity™.
The emphasis of much recent research has moved away from purely tech-
nological studies to a consideration of overall life-style in both the economic
and socio-religious spheres. Such studies have until recently focused primar-
ily on Holocene populations and have made greater advances in southern
Africa than in other parts of the continent. As a result, complex reconstruc-
tions are now possible which view past settlement patterns as seasonally
shifting foci of resource-exploitation, sometimes with alternating aggrega-
tion and dispersal of population. Similar patterns emerge from studies of
the ethnography of some recent San peoples. The relevance of San ethnog-
raphy for suggesting interpretations of the southern African past is further
emphasised by study of the region™s rock art, much of which is now seen
as intimately connected with a socio-religious system based on trance and
shamanism, which may have been a feature of life-styles in the region for
many thousands of years.
It is now clear that Africa, and more speci¬cally the sub-Saharan parts
of that continent, occupied a key rˆle in the cultural and physical devel-
opment of the world™s recent human populations. Archaeological, palaeon-
tological and genetic studies all point to this conclusion. To study these
146 afric an archaeolog y

processes from a Eurocentric perspective can only yield a biased, incomplete
and misleading picture. Human fossils and other archaeological materials
from widely distributed African regions demonstrate both the gradual evo-
lution of anatomically modern humans and the equally gradual inception
of cultural practices that are accepted as representing signi¬cant advances
over those of earlier times. In Africa, these processes are now seen as hav-
ing begun possibly as long ago as 250,000 years and to have continued
into the period, some 50,000--35,000 years ago, when their end-results made
a relatively sudden appearance in the insigni¬cant peninsula now known
as Europe. Both humanity itself and cultural/anatomical modernity were
African developments. No longer can Africa be seen as peripheral to, or
laggard in, the cultural innovations which distinguish modern people from
their predecessors.
The beginnings of permanent settlement
This chapter describes and considers the signi¬cance of trends towards
increased sedentism that occurred in ¬nal Pleistocene and early Holocene
times over wide areas of northern and parts of eastern Africa. Rooted in the
culture of their hunter-gatherer predecessors, some societies of this period
are now seen as having provided the basis for the later adoption of herding
and cultivation. The relevant literature is particularly confusing because of
the various terminologies that have been applied, similar adaptations being
designated Mesolithic, Epipalaeolithic, Late Stone Age or even Neolithic by
different authors (cf. Vermeersch 1992). More than any other in this book,
this chapter cuts across conventional time-divisions in its treatment of an
important but localised process of African cultural development, but it pro-
vides a vital conceptual link between the predominantly foraging societies
and those whose base in cultivation and/or herding has provided the foun-
dation for much of the continent™s more recent development.

North Africa, the Sahara and the Nile Valley
It was noted in chapter 4 how the greater part of the Sahara had little or
no human settlement during the arid period which broadly corresponded
with the coldest part of the last northern-hemisphere glaciation. There are
clear geomorphological traces, such as now-consolidated sand dunes (Roset
1987; Breunig and Neumann 2002), which indicate that the desert at this
time extended even further to the south than it does today. However, the
relatively favourable environment afforded by the Nile Valley at this time,
between 20,000 and 11,000 years ago, allowed its population to increase
within a tightly circumscribed area. Competition for resources and socio-
political demarcation were the largely inevitable results of such processes.
The relevant archaeological evidence may now be discussed (Fig. 69).
By this time, as shown in chapter 4, a number of communities in North
Africa and the Nile Valley were engaged in the intensive and specialised
exploitation of selected plants and animals. In the latter region, there
is evidence for the presence of distinct localised communities. The best-
documented and also one of the earliest incidences of this development is
at Wadi Kubbaniya near Aswan (Wendorf et al. 1989; Barakat 2002), where
tubers of wild nut-grass (Cyperus rotundus) were a major element of the diet
around 18,000--17,000 years ago. It was evidently for the processing of these
0 1000 km

0 500 miles


R. N
Ti-n-Torha, Kharga Esna
Uan Afuda Kom Ombo
Wadi Kubbaniya
Catfish Cave
Jebel Sahaba
Adrar Bous
Tamaya Mellet
Saggai Shaqudud

Kourounkorokale Early Khartoum


ig e

L. Chad



Ele Bor
Present vegetation Lopoy
Desert Ishango
?L. Nakuru

Fig. 69: Location of settlement sites discussed in chapter 5
The beginnings of permanent settlement 149

tubers, rather than cereals, that the heavily worn grindstones at the site
had been used. The unusual preservation of charred tubers thus provides a
salutary corrective to the tendency to emphasise ancient use of cereals at
the expense of other plant foods (Hillman 1989).
During the period of high Nile levels, which lasted from about 15,000 until
about 11,000 years ago linked in part to rising temperatures and the melting
of glaciers in the mountains of Ethiopia, the inhabitants of the Nubian Nile
Valley produced artefact assemblages whose variety demonstrates the estab-
lishment of distinct localised societies. One of these industries, the Qadan,
also shows considerable inter-site variation in the comparative frequencies
of the various microlithic tool types. This evidently re¬‚ects the varied activi-
ties carried out by the populations who ¬shed, hunted wild cattle and other
large ungulates, and also made considerable use of wild plant foods, includ-
ing cereals. This last food source is indicated by the presence on Qadan
sites of large numbers of grindstones and also by the fact that many of
the microliths bear on their edges the characteristic polish known as sickle-
sheen which, it has been demonstrated, may have resulted from their use to
cut grasses (Unger-Hamilton 1988). The Qadan people buried their dead in
cemeteries, at one of which, Jebel Sahaba, a substantial proportion of those
interred could be shown to have met violent deaths (Wendorf 1968). This is
perhaps a further indication of territoriality and inter-group con¬‚ict.
To the north, further downstream, other industries of this time, such as
the Sebilian with its trapeziform microliths, and the Sebekian, present a
similar picture of pronounced variability. On the plain of Kom Ombo, for
example, at least three distinct groups are attested, and the available food
resources were comprehensively exploited: those of the plain itself, the sur-
rounding desert, the river, and its wooded fringe (P. E. L. Smith 1967). Some
Kom Ombo sites seem to have been occupied on a year-round basis, and
here too wild grasses were harvested. At Esna in Upper Egypt pollen analy-
sis suggests that wild barley was, by 12,000 years ago, one of the varieties
gathered on the ¬‚ood plain (Wendorf and Schild 1976). Several of the ani-
mal species whose bones are represented in the food debris from sites of
this period would have had narrowly circumscribed habitats and this may
have enabled people to experiment with the development of management
techniques over wild or semi-wild herds. These important innovations may
be seen, at least in part, as responses to the concentration of population in
the Nile Valley brought about by desiccation of the surrounding deserts.
Shortly after 12,000 years ago there was a remarkably rapid return to
better-watered conditions in most of what is now the Sahara. Increased
run-off from the highlands coupled, presumably, with higher rainfall and
decreased evaporation resulted in the return of a regular ¬‚ow of water
to the long-dry wadis, the great enlargement of existing swamps and
150 afric an archaeolog y

lakes -- notably Chad -- and the formation of many new ones (Petit-Maire
1988; Grove 1993). There were corresponding changes in vegetation and in
the distribution of wild animals. The reasons for these large-scale trends are
not fully understood. One of their most puzzling features is the rapidity
with which they took place: the lakes appear to have reached their maxi-
mum heights as early as about 11,000 years ago (Hassan 1997).
Little research has yet been undertaken on archaeological sites which may
con¬dently be attributed to the ¬rst human presence in the Sahara under
the ameliorated climatic conditions of the early Holocene. Settlement more
than 10,000 years ago, if it took place, must have been extremely rare. There
were major short-term climatic ¬‚uctuations (Hassan 2002) and it seems likely
that human settlement extended only very slowly into the newly habitable
regions (Close 1992; Garcea 1993). From Mauritania to the Western Desert of
Egypt, the earliest post-Aterian assemblages include characteristic pointed
bladelets with basal retouch, known as Ounanian points (A. B. Smith 1993b).
Assemblages of this type from the Western Desert, as at Kunkur Oasis, also
contain many roughly trimmed ¬‚akes but no backed microliths, although
the latter artefacts are important components of contemporary assemblages
at Nabta Playa and Kharga Oasis (Schild and Wendorf 1977). At all these
sites the presence of grindstones suggests that here, as at many Nile Valley
sites of the same period, cereal grains -- presumably wild -- may have been
harvested and prepared for use as food. A similar industry occurs at Adrar
Bous in the Tenere Desert of Niger (A. B. Smith 1976). The degree of similarity
between these Saharan industries and broadly contemporary material from
southern Tunisia, such as the Upper Capsian, supports the view that part
of the initial repopulation of the Sahara may have taken place from the
north. Whatever the population source, it is clear that, although numerous
and widespread, the Saharan sites of this time were individually of limited
extent and brie¬‚y occupied, indicating a population of small mobile groups
(J. D. Clark 1980). It is possible that the earliest Saharan rock art, consisting
primarily of engravings of wild animals (Fig. 70), may date from this period
(Mori 1974; but see also Muzzolini 1991).
By 9000 years ago signi¬cant changes had taken place, although their
form varied signi¬cantly, as did local environmental conditions. Excavations
in the Acacus highlands of southwestern Libya (Barich 1987) have revealed
stone structures in Ti-n-Torha rockshelter (Fig. 71) which suggest some per-
manency of settlement. In the same area there is evidence at Uan Afuda (di
Lernia 1999; see also di Lernia 1998) for intensive exploitation of large herds
of wild Barbary sheep, as has been suggested for earlier times at Tamar
Hat (p. 136, above). The clearest illustration of the development of seden-
tism at this time comes from research at Dakhleh Oasis in the Egyptian
Western Desert (McDonald 1991). Between about 9000 and 8500 years ago,
The beginnings of permanent settlement 151

Fig. 70: Rock
engraving of
Bubalus antiquus,

the area™s inhabitants seem to have adopted a more settled life-style, with
circular stone-based huts, exploiting a signi¬cantly smaller territory than
their predecessors had done. Further to the south at this time, local cattle
were probably already being herded, as will be discussed below in chapter 6
(p. 175).
In well-watered situations, ¬shing and the exploitation of other aquatic
food resources now played a large part in the economy of a vast area of the
central and southern Sahara from the Nile Valley at least as far to the west as
Mali. Sites were concentrated on the shores of rivers and lakes which were
signi¬cantly higher and more extensive than those of today. Although at
most sites hunting and grain-collecting were continued on a reduced scale,
the pre-eminence of ¬shing now allowed larger populations to remain for
longer periods of time at individual sites. Bone harpoon heads were the
characteristic artefacts indicative of this new development. Pottery is also
152 afric an archaeolog y

Fig. 71: Stone
structures at

present at most sites of this type (Close 1995): it often bears the distinctive
wavy-line decoration discussed in greater detail below (p. 156). In the Sahara,
such pottery probably appeared at a slightly earlier date than in the Nile
Valley or East Africa, being attested between 9500 and 8500 years ago in the
Acacus in southern Libya, at Tamaya Mellet west of the Aïr Mountains of
Niger (A. B. Smith 1980a), and at Amekni (Fig. 72) in the extreme south of
Algeria (Camps 1969). Further to the west, in Mali, wavy-line pottery occurs
also at Outeidat near Timbuktu (Gallay 1966), and barbed bone harpoons at
Kourounkorokale near Bamako (Szumowski 1956; MacDonald 1997).
The Nile Valley industries between 12,000 and 8000 years ago are poorly
illustrated by the research that has so far been undertaken. Thereafter the
Nubian stone industry is of the microlithic type known as Shamarkian
(Fig. 73), which includes small numbers of Ounanian points similar to those
which also occur in Saharan industries of this time, as noted above. Some-
what later and far downstream, the Fayum Depression between 8000 and
7000 years ago supported lakeside camps of people who made microlithic
artefacts, mounted ¬sh jaws as points for arrows, and made their livelihood
by a combination of hunting and ¬shing (Caton-Thompson and Gardner
1934; Wendorf and Schild 1976; Brewer 1989). This Qarunian occupation
beside the extensive lake which formerly occupied the Fayum Depression
provides examples of arrow-manufacturing techniques which continued in
Egypt into dynastic times.
The beginnings of permanent settlement 153

Fig. 72: Pottery
from Amekni
(after Camps

It is, however, from the central Sudan that we have our most detailed
knowledge of the inhabitants of the Nile Valley some 8000 years ago. The
area was occupied by hunters and ¬shers who made use of substantial base-
camps, probably occupied for much of the year, such as that which is known
as Early Khartoum (Arkell 1949). (The local antecedents of its population
remain unknown.) At Early Khartoum the stone industry included edge-
retouched ¬‚akes, backed microliths, and larger tools for which the name
˜crescent adzes™ has been proposed; they are thought to have been used for
the shaping of spear- or harpoon-shafts or similar wooden objects. Bone
harpoon heads, barbed on one side only, were also a characteristic part of
the assemblage. Stone rings and other objects, best interpreted as weights
154 afric an archaeolog y

Fig. 73: Artefacts
from Nubia (after
Wendorf 1968): 1,
pottery; 2--4,
microliths (4 is
an Ounanian

for nets, suggest that harpooning was not the only method by which ¬sh
were taken. Pottery was common, generally decorated with multiple-grooved
wavy lines, which may be duplicated experimentally by dragging a cat¬sh
spine over a surface of wet clay. During the later phases of the site™s occu-
pation, these designs were elaborated by jabbing the clay with a pointed
object to produce a series of impressed dots (Fig. 74). Traces of sun-dried
daub were recovered at Early Khartoum, suggesting that structures of some
sort were erected there. Fishing and hunting were both important subsis-
tence activities; and the presence of several swamp species in the faunal
assemblage shows that the Khartoum area was signi¬cantly wetter than it
is at the present time. The dead were buried, in contracted positions, in
graves within the settlement. The Early Khartoum site was excavated before
the development of radiocarbon dating techniques, so its age remains uncer-
tain, although probably falling between 8000 and 7000 years ago. A related
but presumably earlier site with harpoons but no pottery, at Tagra on the
White Nile some 200 kilometres to the south, is dated to about 8300 years
ago (Adamson et al. 1974).
The information derived from Early Khartoum has received signi¬cant
con¬rmation from investigations by J. D. Clark (1989) and his colleagues at
Shabona, some 140 kilometres upstream of the former site. Beside an embay-
ment of the high-level White Nile an extensive area saw repeated settlement
between about 7500 and 7000 years ago. Baked ¬sh formed an important
element in the diet of people whose material culture, including pottery, was
essentially the same as that known from Early Khartoum. Wild grasses --
probably Digitaria -- were abundant, being used as temper for pottery; the
excavator has suggested that the seeds may also have been collected and
The beginnings of permanent settlement 155

Fig. 74: Artefacts
from Early
Khartoum (after
Arkell 1949): 1,
backed microlith;
2, edge-retouched
¬‚ake; 3, barbed
bone harpoon
head; 4, potsherd
with wavy-line
5, grooved stone
interpreted as
weight for a ¬sh
net; 6, stone ring;
7, grindstone

˜ground into ¬‚our to make porridge™ (J. D. Clark 1989: 407). Flat depressions
up to 2.5 metres in diameter may represent dwelling areas.
At Shaqadud in the Butana Plain, 150 kilometres northeast of Khartoum
and 50 kilometres distant from the Nile (Marks and Mohammed-Ali 1991),
a long sequence of prehistoric occupation began with settlement of Early
Khartoum type which lasted from about 8000 until 6000 years ago when cli-
matic conditions were somewhat wetter than those which prevail today.
At Saggai (Caneva 1983), another site a short distance east of the Nile
30 kilometres north of Khartoum, which was more brie¬‚y occupied rather
less than 7500 years ago, there is evidence for harpoon ¬shing, shell¬sh col-
lecting and hunting, suggesting human presence at more than one season
of the year.
Sites similar in general terms to Early Khartoum have been investigated
in the area of the Nile/Atbara con¬‚uence (Haaland and Abdul Magid 1995);
they date between 8600 and 6800 years ago with pottery in use throughout.
However, in Nubia to the north, pottery seems to have remained unknown
until late Shamarkian times, around 6500 years ago. There can thus be no
156 afric an archaeolog y

reasonable doubt that, whatever the origin of the Early Khartoum pottery
(and this will be discussed below), it does not represent a technology that
spread up the Nile from Egypt. Comparable pottery does occur further to
the south, despite a substantial gap in its known distribution between the
Sudanese Nile Valley and Lake Turkana in northern Kenya.
By 7000 years ago, at Cat¬sh Cave near Abu Simbel in Egyptian Nubia,
a Shamarkian-like stone industry without pottery was associated with spe-
cialised ¬shing equipment in the form of bone harpoon heads barbed along
one edge (Wendt 1966). These sites do not mark the beginning of inten-
sive ¬shing in this region; at least in the Nile Valley, this had already been
practised for several thousands of years (Van Neer 1989). As noted above,
analogous harpoons are widely distributed in the Sahara; Nubia lay on the
northeastern fringe of the area occupied by the harpoon-¬shers.
˜Wavy-line pottery™ (Mohammed-Ali and Khabir 2003) has long been recog-
nised as a signi¬cant and widespread innovation in Saharan Africa. First
recognised in the Nile Valley, it was formerly believed to represent an expan-
sion westwards into the desert (Hays 1974). More recent research has, how-
ever, demonstrated that some of its Saharan occurrences are signi¬cantly
older than those beside the Nile (Close 1995); in fact they precede all other
pottery known either in the Nile Valley or in the Mediterranean coastal
regions of Africa. The available radiocarbon dates indicate a strong possi-
bility that this early Saharan pottery was a local invention at least as early
as 9500 years ago, there being no reliable evidence for any older material
in adjacent areas from which the necessary technology could have been
derived. Such an interpretation is fully plausible, since the accompanying
relatively settled life-style, also an independent development, would have
made possible the adoption of heavy fragile receptacles. These were very
useful in a semi-permanent settlement but would have been ill-suited to
the more mobile life-styles of earlier times.
Haaland (1992) has provided a stimulating overview of the processes lead-
ing from ¬shing/hunting/gathering to full mixed farming, although she con-
centrates on sites in the Sudanese Nile Valley and underestimates the sig-
ni¬cance of developments in more westerly areas. There can be little doubt
that semi-permanent settlement preceded the practice of herding (Caneva
et al. 1993) and that wild cereals were cultivated for many centuries, if not
millennia, before the plants concerned display any of the morphological
changes associated with domestication.

East Africa
With one very signi¬cant exception, the East African sites relevant to
this chapter are located around Lake Turkana and may be linked with its
The beginnings of permanent settlement 157

Fig. 75: Old
beach deposits at
80 metres above
the present water
level of Lake
Turkana. The
volcano, Mount
Kulal, is visible in
the background.

¬‚uctuating high levels during the early Holocene (Butzer 1971). The waters
of the lake rose rapidly around 10,000 years ago to a level 80 metres above
its modern surface, which gave the lake nearly twice its present area and
maintained an over¬‚ow channel to the northwest which eventually con-
nected with the Nile. Deposits in the Omo Valley of southern Ethiopia,
which represent a major northerly extension of Lake Turkana at this time,
have yielded bone harpoons from levels dated to between 10,000 and 8000
years ago (F. H. Brown 1975). More detailed information concerning the
makers of these artefacts comes from several sites located near the 80-metre
beach lines on the eastern and southwestern sides of the lake in Kenya
(Robbins 1974; Barthelme 1985; Stewart 1989). At Lowasera (Fig. 75) to the
southeast, harpoons were probably in use by 9000 years ago; there is evi-
dence elsewhere for a brief period of signi¬cantly lower water levels between
about 7500 and 6800 years ago (D. W. Phillipson 1977c). Before this inter-
val, pottery -- where present -- was rare and of wavy-line type; later sherds
are much more abundant and generally undecorated. In the Lake Turkana
Basin the earlier harpoons have a nick at the base for the attachment of the
line; the later ones used several concentric carved grooves for this purpose
(Fig. 76). The associated stone industry was, throughout, a mixture of backed
microlithic elements and of large retouched ¬‚akes and core-tools, with a few
grindstones. Food remains consisted almost exclusively of the bones of ¬sh,
crocodile and hippopotamus. The second harpoon-¬shing phase continued
until after 4500 years ago, by which time the waters of Lake Turkana were
again falling; there are, indeed, indications that a similar life-style continued
158 afric an archaeolog y

Fig. 76: Artefacts
from Lowasera
(after D. W.
Phillipson 1977c):
1, edge-retouched
¬‚ake; 2, backed
microlith; 3,
barbed bone
harpoon head of
early type with
basal nick for
attachment of
the line; 4,
barbed bone
harpoon head of
later type with
basal grooves;
5, early potsherd
with wavy-line
6, later
7, hammerstone

at Lopoy, west of the lake, until less than 1000 years ago, when the lake™s
water-surface still stood about 18 metres above its present level (Robbins
et al. 1980). Today, Elmolo people of the Lake Turkana littoral ¬sh by means
of barbed harpoons, now with iron heads.
Although there are inconclusive hints from the Lake Nakuru Basin in
southern Kenya, the only other part of East Africa to have yielded evidence
for harpoon ¬shermen is that around Lake Edward/Rutanzige on the border
between Uganda and D. R. Congo. Here, we are concerned with a single site,
Ishango on the northwestern shore, which has preserved a long sequence
of occupation, although its dating remains uncertain (de Heinzelin 1957;
Brooks and Smith 1987). Bone harpoon heads occurred throughout, at ¬rst
barbed on both sides, latterly on one side only. The crude stone industry, of
quartz, included edge-retouched ¬‚akes and some backed microliths. There
was no pottery. An age of at least 7000 years is demonstrated, but there are
The beginnings of permanent settlement 159

good reasons for accepting that the site is very considerably older, perhaps
as much as 18,000--16,000 years. Conditions wetter than those of the present
are once again indicated by the animal bones that were preserved in the
deposits; because of its outlet Lake Edward is less susceptible to changes in
water level than is Lake Turkana, and its height at this time was only about
12 metres above that which prevails today. As noted in chapter 4 (p. 120
above), there are indications that ¬shing with bone-headed harpoons may
have an even greater antiquity in this area. Be that as it may, it may be con-
cluded that the Lake Edward Basin, which is the most southerly location of
the East African harpoon-¬shing adaptation, is also its oldest known manifes-
tation. Here and around Lake Turkana, it may be shown that this adaptation
developed signi¬cantly before the local beginning of pottery manufacture.
When pottery did appear, its earliest East African manifestation (as repre-
sented at Lowasera) showed strong similarities with those of the Sahara and
the Sudanese Nile Valley. These ¬shing settlements represent, of course, only
one of several economic adaptations in East Africa at this time.

These settled communities represent economic adaptations over a very wide
area that parallel similar processes of environmental change (Barich and
Gatto 1997; Cremaschi and di Lernia 1999). At least in North Africa and
the Nile Valley, there had been for some millennia previously a tradition of
intensive exploitation of various food resources, including aquatic ones, in
circumstances which sometimes encouraged prolonged settlement of a par-
ticular site. Later semi-permanent settlements, beginning around 8500--8000
years ago, were characterised by the use for ¬shing of harpoons with barbed
bone heads, and by pottery with wavy-line decoration. Contrary to previous
opinion, the Nile Valley does not appear to have been the original home-
land of the harpoon-¬shing adaptation. On the evidence currently available
the East African material of this type seems to be the oldest, a view which
is strengthened by the very early date now proposed for the Ishango site,
where biserially barbed harpoons were used by people who were ignorant
of pottery. When pottery was developed by later hunter/¬shers, it seems to
have been an independent invention, perhaps originating in what is now
the Sahara (di Lernia and Manzi 1998). The two characteristic elements of
these settlements™ archaeological remains, although often found together,
thus appear to have originated separately.
The bone harpoon heads, which are widely distributed at sites in the
Sudanese Nile Valley, in the southern Sahara and in parts of East Africa,
at one time received undue emphasis from some archaeologists (e.g. Sutton
1974) who suggested that they may indicate a uni¬ed ˜civilisation™ based
160 afric an archaeolog y

upon the exploitation of aquatic resources. Many of these objects were
indeed used for ¬shing, but it cannot be assumed that this was invariably
the case. At Daima, in northeastern Nigeria, a bone harpoon head was found
embedded between the bones of a human skeleton (Connah 1981), demon-
strating that such weapons were, at least in later times, occasionally used
against human targets. Despite the basic typological similarity over a very
wide area of both the bone harpoons and the pottery, these two elements
can be traced to widely separated origins, and the chipped-stone assemblages
with which they are associated show considerable variation, seeming gen-
erally to be rooted in distinct local traditions which may be traced back
into earlier times. For these reasons it seems most satisfactory to regard the
¬shing settlements as representing a common adaptation (or parallel adap-
tations, with many shared features) to a common economic opportunity,
rather than to consider them as belonging to a single uniform culture.
The ¬shing settlements, being concentrated in location, are the most read-
ily recognised archaeological sites of this time. Their inhabitants were able
to exploit enormously rich food resources that were obtainable with very lit-
tle effort. Other means of livelihood were, however, pursued both alongside
¬shing and on different sites, away from permanent water. Hunting and,
probably, grain collection continued in the Sahara. Elsewhere, as at Ele Bor
in northern Kenya, hunters maintained their traditional life-style without
pottery long after this had been adopted in the lakeside ¬shing settlements
(D. W. Phillipson 1984; Gifford-Gonzalez 2003).
Various speculations have been expressed concerning the linguistic af¬ni-
ties of the early settled communities considered in this chapter. Sutton
(1974) raised the question of whether their members might have spoken
an ancestral Nilo-Saharan language, the recent discontinuous distribution
of these languages having been brought about by subsequent movement
and dispersal. Ehret™s (1993) argument that early Nilo-Saharan speakers were
farmers, discussed further below (pp. 211--12), is on much ¬rmer ground,
although not necessarily incompatible with Sutton™s view.
The full importance of the innovations represented by the harpoon-¬shing
adaptation and of the settled life-style that accompanied it can only be fully
appreciated when we consider the subsequent adoption of farming. This will
be discussed in detail in chapter 6.

African peoples 10,000 years ago
This is an appropriate point to view the overall distribution of African pre-
historic communities at a single time, before effective farming is attested
in any part of the continent. We will also examine the evidence for the
distribution of human physical types at this general period and enquire to
The beginnings of permanent settlement 161

what extent it is possible to recognise populations that were ancestral to
more recent ones.
The modern indigenous population of Africa comprises several varied
physical types, the distinctions between which are not always readily appar-
ent. It is appropriate in this context brie¬‚y to investigate the extent to which
this diversity may be traced back in the archaeological record. Several writers
(e.g. Hiernaux 1968; Weiner and Huizinga 1972) have stressed the dif¬culty
of de¬ning discrete recent populations even on the basis of a complete range
of physical characteristics, including such features as ¬nger prints and blood
groups. In recent years, genetic research has begun to reinforce this picture
of ill-de¬ned heterogeneity (Watson et al. 1997; Kittles and Keita 1999), as
will be explained below. Since African populations have lived in close con-
tact with one another for many thousands of years, it is inherently highly
improbable that any population will have retained in unmixed form the
characteristics of any prototype which may formerly have existed. Yet it is
such non-homogeneous populations that have, perforce, to form the basis
for our recognition of the groupings concerned.
There are thus major problems not only of recognition but also of de¬ni-
tion. These problems have led some authorities to ask whether it is reason-
able for archaeologists to attempt to identify discrete human physical types
ancestral or related to those of the present day. It is nevertheless widely
recognised that there are metrical skeletal features, notably of the skull
and teeth (Rightmire 1972; Irish and Turner 1990; Irish 1994), which are
generally more characteristic of one population than of another. Although
observed ranges of variation frequently overlap, the ranges noted in a pre-
historic population may show a close degree of ¬t with those of a particular
recent group; the af¬nities of isolated individuals are correspondingly dif-
¬cult to determine (cf. A. W. F. Edwards 2003). Bearing these many hazards
in mind, it is worthwhile to survey the views that have been proposed for
the af¬nities of ¬nal Pleistocene/early Holocene African populations.
In the Maghreb the Oranian hunter-gatherers were replaced by the mak-
ers of the varied microlithic industries which eventually gave rise to the
Capsian. It has been suggested that people of proto-Mediterranean type ¬rst
appeared in North Africa at this time, joining the robust ˜Mechta-Afalou™
population who seem to have been the sole inhabitants of the region dur-
ing Oranian times (Dutour 1989). These proto-Mediterranean people may
have been at least partly responsible for the resettlement of the northern
Sahara following the post-Aterian period of desiccation.
To the south, the harpoon-using ¬shermen of the central and southern
Sahara, the Sudanese Nile Valley and parts of East Africa are represented
by a few fragmentary skeletons from Ishango, Lothagam on Lake Turkana
(Robbins et al. 1980), Lowasera (Rightmire in D. W. Phillipson 1977c), Early
162 afric an archaeolog y

Khartoum and elsewhere which are said to show negroid physical features.
Similar characteristics occur in skeletons from the Qadan cemetery at Jebel
Sahaba in Nubia (Wendorf 1968), where Mechta-Afalou features have also
been recognised. These remains may be from a population ancestral to
present-day Nilotic negroids. Other authorities, emphasising the presence of
features which are also seen in KhoiSan and Northeast African ˜caucasoid™
populations, prefer to interpret this material as representing a more gener-
alised ˜ancestral African™ physical type, which may be regarded as akin to a
common ancestor of several more recent populations. This explanation also
seems plausible for the varied human remains that have been recovered
in association with broadly contemporary and rather later industries from
southern Kenya (Rightmire 1975; see also Morris 2003).
The archaeology of this period in West Africa remains poorly understood,
being covered by only two dated sequences, at Iwo Eleru in southwest-
ern Nigeria and Shum Laka in Cameroon. In this area the hunter-gatherer
people of the forest margin had adopted a backed-microlith technology
akin to that of the neighbouring savanna. Elsewhere in the West African
forest regions non-microlithic stone-artefact manufacture continued. A sin-
gle human skeleton some 12,000 years old from the lowest level at Iwo
Eleru has been described as already showing speci¬cally negroid features
(Brothwell and Shaw 1971), but the broadly contemporaneous burials at
Shum Laka (Ribot et al. 2001) were too fragmentary to permit detailed eval-
uation. In other parts of West Africa, skeletal material has not been pre-
served. The same is unfortunately true of the whole of the Congo Basin and
adjacent forested regions where the Tshitolian industries had by this time
On the southern savanna many areas had seen the practice of backed-
microlith production for many thousands of years. By 10,000 years ago
related technology had been adopted by hunter-gatherers in most of the
region except, it appears, in areas of the arid southern African interior
and perhaps in parts of the upper Zambezi Valley to the north. Substan-
tial numbers of human skeletons have been recovered in association with
these industries, particularly noteworthy (although late) being the group
of thirty-three individuals of about 5000--3000 years ago from Gwisho hot-
springs in southern Zambia (Brothwell 1971) and a somewhat earlier series
from graves in cave sites on the south Cape coast. There can be little doubt
of the KhoiSan af¬nities of most of this material. However, some skeletons,
particularly from north of the Zambezi, are also stated to show features char-
acteristic of recent southern African negroid populations (e.g. de Villiers in
D. W. Phillipson 1976). This observation serves to emphasise the dif¬culties
in a continent where for many millennia few barriers to genetic mixing
have existed. Such links do little more than reinforce the view that both
The beginnings of permanent settlement 163

KhoiSan and negroid stocks are derived from a single generalised African
ancestral population (Tobias 1978b).
By the early Holocene, it thus appears possible to make the ¬rst tentative
correlations between communities represented in the African archaeolog-
ical record and the ancestors of some modern populations. The principal
physical types seen in recent indigenous populations may also be recog-
nised in the archaeological material, where the distinctions between them
are no more clear than they are between the modern groups. What we see
in the incomplete archaeological picture so far available is the gradual pro-
cess of differentiation of modern human types from ancestral populations
that were genetically as mixed as those of the present day.
In recent years, considerable advances have been made over the earlier
treatments of this topic, brought about by the development of genetic
research and by further re¬nement of linguistic methodologies. In both
cases, primary research is effectively restricted to modern populations, but
the resultant data can yield important implications about the past. Study
of the DNA in modern African populations (e.g. Passarino et al. 1998; Salas
et al. 2002) permits the mapping of genetic variation independently of phys-
ical characteristics. These plots may be interpreted to generate historical
hypotheses and the time-scales involved may be estimated through refer-
ence to what is known about rates of mutation (Watson et al. 1997; Pereira
et al. 2001). For example, it is becoming apparent that the populations of
sub-Saharan Africa are genetically far more diverse than are modern people
elsewhere, and that some of these characteristics diminish in more northerly
and northeasterly parts of the continent (Lalueza Fox 1997; Krings et al. 1999).
As noted in chapter 1, a somewhat similar methodology may be followed
in the manipulation of linguistic data. It is now recognised that the several
language families of Africa can probably be traced back for 7000--10,000
years or, in some cases, even longer (Blench 1999; Ehret 2003). It has been
suggested that, in Africa as in other parts of the world, language change and
the adoption of farming techniques may have proceeded together (Bellwood
and Renfrew 2003). Some aspects of genetic continuity and variation are
better understood than are processes of linguistic change: transmission of
mitochondrial DNA, for example, takes place exclusively through the female
line. This means that reconstruction of historical processes based on genetic
evidence may sometimes have a ¬rmer foundation than those proposed
exclusively on linguistic grounds.
Attempts to correlate conclusions based on different methodologies have,
in theory, considerable potential. Unfortunately, in practice, most such
attempts have foundered because few scholars are adequately familiar
with the methodologies and constraints of more than one discipline. For
example, an attempt to use genetic divergence between groups of modern
164 afric an archaeolog y

KhoiSan-speakers to estimate the antiquity of that language family (Knight
et al. 2003) is marred by false assumptions concerning language transmis-
sion and change. Such studies are, however, in their infancy and offer much
opportunity for re¬nement. It is already clear that, as the nature of human
variation becomes better understood and self-centred stereotypes (whatever
their basis) become increasingly untenable, genetic studies -- especially when
used in conjunction with other approaches -- offer great potential for aug-
menting our understanding of the history of modern human variation.
Early farmers
Cultivation and herding
Previous chapters of this book have discussed the stages of human devel-
opment during which people relied for their livelihood on the plants and
animals that were present in their natural environment, feeding on wild
vegetable foods, as well as on the meat of wild animals, birds, ¬sh and
insects. In Africa, as in other parts of the world, people have been exclu-
sively foragers for more than 99 per cent of their existence. Here, attention
will be drawn to the processes by which greater control over animals and
plants gradually gave rise to domestic forms (Bower 1995). In the northern
half of Africa (Fig. 77), these developments were achieved by people who had
not yet learned metallurgical skills; in many more southerly areas, the ¬rst
use of domestic plants and animals was made by people who also worked
metals, as will be discussed in chapter 7.
It has been shown in chapter 5 how, from as early as 18,000 years ago,
some Nile Valley communities in Upper Egypt were making intensive use
of vegetable foods in the form of tubers. It is likely that this practice has,
in fact, a far greater antiquity, although its earlier manifestations have not
yet been revealed by archaeology. By 15,000--11,000 years ago, people in this
area had begun to utilise wild cereals in a similar way, as their successors in
many Saharan regions have continued to do into recent times (Harlan 1989;
Wasylikowa 1992). These practices may have been accompanied by care of
wild grasses through such means as control of weeds, clearance of ground
and, perhaps, occasional provision of water. Under these circumstances sow-
ing might also have been employed to increase the density of growth in the
tended places and to extend the areas colonised by the wild plants. With or
without conscious selection for desired qualities, such practices may even-
tually have led to the development of crops morphologically distinct from
their fully wild prototypes (cf. Zohary 1969; Chikwendu and Okezie 1989).
By such means, some of the late Pleistocene peoples of the Nile Valley seem
on occasion to have exploited wild cereals and other plants, but there is
no evidence that true cultivation was developed at this time. These are
among the earliest instances of intensive cereal utilisation that have yet
been demonstrated anywhere in the world, and there can be no reasonable
doubt that they were indigenous African achievements. But they did not
lead to the widespread adoption of such practices at this early date. A later
El Khril
Oued Guettara
Haua Fteah

Merimde Helwan
Badari Hamamiya
This Nakada
Deir el Bahari Sodmein
Uan Muhuggiag Hierakonpolis
Bir Kiseiba Aniba Afyeh
Abu Simbel
Jebel Uweinat
Dhar Tichitt Semna Buhen
St Louis Adrar Bous Kerma
Arlit Borkou
Esh Shaheinab El Geili Agordat
Casamance Kassala
Aksum, Gobedra
Dufuna Daima Cave
Kamabai Ntereso
L. Besaka
Iwo Eleru
Bosumpra Shum Laka Bouar
Cape Three Points Ileret Ele Bor
Bioko North Horr
Principal rock art areas
(paintings and engravings)

Ogoou© o Falls

0 1000 km

0 500 miles
Ngovo Gombe Point

Fig. 77: Location of principal sites with rock art and/or evidence for early cultivation or herding
Early farmers 167

section of this chapter will offer some speculation as to why this may have
been so.
Similarly, it now appears that the late Pleistocene/early Holocene peoples
in several regions of northern Africa came to concentrate their exploita-
tion of wild animal resources on single or very few species (Barich 1987).
The example of the Barbary sheep (Ammotragus lervia) has been cited above
(pp. 136, 150--1). The increased knowledge and awareness that accompanied
this aspect of specialisation may have led in the direction of herding or,
conversely, to conditions where animals previously herded elsewhere could
readily be adopted. It will be argued below that domestic animals did not
become widespread in the northern half of Africa until about the sixth mil-
lennium bc. By that time plant cultivation and animal herding had both
been long established in the Levant (Bar Yosef 2003). Although it is not easy
to show the extent to which some early farming practices in northern Africa
may have been derived from the Levant, it is clear that many developments
both here and further to the south were autochthonous (Harlan 1992). In
this context it is particularly illuminating to consider where the wild forms
of the various plant and animal species that were eventually domesticated
may have originated. It has for long been assumed that cultivated wheat,
barley and ¬‚ax were all introduced into Africa from the Levant, although
wild barley and, perhaps, einkorn wheat occur in several areas of Egypt
where their food value was recognised from very early times (cf. p. 149,
above). If we exclude crops such as maize and cassava (manioc), which are
known to have been introduced from the New World within the last ¬ve hun-
dred years, and those such as bananas brought signi¬cantly earlier across
the Indian Ocean, most of the crops which are or have been cultivated in
Africa belong to species that are indigenous to that continent and which
must presumably have been ¬rst cultivated there. Examples are the more
important types of yam, African rice, certain pulses and the cereals sorghum,
¬nger millet (Eleusine), bulrush millet (Pennisetum) and teff (Eragrostis tef),
together with the Ethiopian plants enset and noog (Fig. 78), all of which are
derived from plants which grow wild in the sub-Saharan latitudes. It will
be noted from this list that indigenous African food crops fall into three
primary categories: rice, other cereals, and vegetatively propagated plants
such as yams and enset. Each of these categories requires a distinct method
of cultivation, and there is no reason to suppose that the development of
these horticultural modes was in any way interconnected. Much discussion
of early African food-crops centres on the cultivation of cereals, because
of the nature of the primary evidence and its greater chances of preserva-
tion, but cereals were not necessarily the earliest such crops nor, in many
areas, are they the ones that have made the greatest contribution to human
168 afric an archaeolog y

0 1000 km

2 2
7 1 5
2 2
7 2
Fig. 78: The 8 9 94
probable areas of 5
the initial
1 Yam 6 Bulrush millet 4
domestication of
2 Guinea rice 7 Fonio
indigenous 3 Sorghum 8 Groundnuts
African crops 4 Finger millet 9 Ensete and noog
(after Harlan 5 Teff

In the case of domestic animals the position is somewhat different. Sheep
(Ovis) and goats (Capra) have no African wild prototypes, for it has been
shown that the Barbary sheep (Ammotragus) is most unlikely to have given
rise to any known domestic forms (Epstein 1971; Cassinello 1998). It is
often dif¬cult to differentiate between sheep and goats on the basis of
the fragmentary bones that are recovered from archaeological sites, and
references to ˜ovicaprids™ or to ˜small stock™ in later sections of this book
will indicate remains which cannot be attributed with con¬dence to the
correct species. Cattle present more dif¬culties: although it was formerly
believed that domestic cattle originated in western Asia or southeastern
Europe, recent genetic studies illustrating the antecedents of particular cat-
tle breeds strongly support the view that there were in fact two centres of
origin, one being in northern Africa (Marshall 1989; Grigson 1991; Blench
and MacDonald 2000). Cattle of North African derivation were of paramount
importance in subsequent developments throughout the continent (Grigson
2000; Marshall 2000). Although much more circumscribed in distribution,
the donkey also appears to have been an indigenous African species that
was domesticated in one or more of the northeastern regions (Blench 2000).
Domestic fowls remain poorly understood: the earliest in Africa was prob-
ably the guinea fowl in West and North Africa. Chickens, of Asian origin,
were known in Egypt from Ptolemaic times if not before, but not south
of the Sahara until the ¬rst millennium ad. It is, however, abundantly
clear that none of the major domestic animals of sub-Saharan Africa, other
than cattle and donkeys, is derived from a species that is indigenous to
that part of the continent; they must therefore have been introduced from
Early farmers 169

Cultivation and herding have often been considered together by archae-
ologists as ˜food-production™. Domestic animals and cultivated plants are,
however, far more than sources of food. They provide important raw materi-
als for clothing and many other purposes, as well as stimulants and drugs.
They sometimes play an important role in religious affairs and, as will be
shown below, provide the basis for far-reaching social, political and eco-
nomic developments. In several African societies domestic animals serve as
embodiments of wealth and indicators of status. Food-production is thus
just one aspect of practices which have become central to the whole life-
styles and belief-systems of many peoples. In view of this wide importance,
in the following pages the term ˜farming™ will be employed in place of the
more usual ˜food-production™. Likewise, the term ˜agriculture™ is best avoided
since it is often used ambiguously either to include or to exclude manage-
ment of domestic animals. Since the cultivation of plants and the herding
of animals are two quite distinct cultural activities, it is often preferable
to treat them separately, without the implication that the practice of one
necessarily implies the other.
A further confusion may arise from imprecise application of the term
˜pastoralist™ to any person or community possessing domestic animals, irre-
spective of the importance which these animals may have had in the overall
life-style of the people concerned. Used correctly, the term applies only to
societies or individuals who rely upon domestic animals for a very large
component of their subsistence, and whose lives are largely controlled by
the need to care for their herds. In this book, unless there is very clear
evidence for true pastoralism, the more neutral term ˜herder™ is used to
designate someone who owns or controls domestic livestock.
Here, it is appropriate brie¬‚y to evaluate the types of archaeological evi-
dence that may be accepted as proof of ancient farming. By far the most
convincing are the actual remains of cultivated plants or domestic animals,
or unequivocal artistic representations of them. Pollen grains or seeds may
survive under appropriate conditions, or impressions of seeds may be pre-
served on pottery. Some food-plants, such as yams and bananas, will by their
very nature hardly ever be represented in the archaeological record, and this
has led to undue emphasis on the better-preserved evidence for cereal cul-
tivation. There are indications that study of phytoliths may ¬ll this gap
(Piperno 1988), but the relevant techniques are still under development and
it is only occasionally that detailed identi¬cations can be made (e.g. Mbida
et al. 2000). Only recently, and still infrequently, have excavators in Africa
routinely employed techniques such as ¬‚otation which are generally neces-
sary if plant material is to be recovered comprehensively. Since, by contrast
with plant remains, bones are relatively indestructible and easy to identify,
170 afric an archaeolog y

archaeologists have often tended to overemphasise animals at the expense
of plants. As a result, many reconstructions of ancient diets -- of all periods --
place excessive stress on the meat that was consumed and on other animal
products, tending to minimise the vegetable component (but see van der
Veen 1999). It must be admitted that this prejudice may mirror the views
of some ancient African societies. Today, meat is often regarded as the most
important food -- a view that may be connected with the traditional division
of tasks between men and women, where the provision of meat and other
animal products is seen as a male task, with women obtaining most of the
plant foods.
A further complication involves dating. Objects -- whether artefacts, bones
or plant remains -- found together in a primary archaeological context may
generally be assumed to be contemporaneous and, in appropriate circum-
stances, datable. However, in the case of small items like seeds, great care
must be exercised to ascertain whether they may be intrusive from a younger
or older context. Sometimes, particularly when only a few specimens are
preserved, it is appropriate to pay particular attention to obtaining direct
radiometric age determinations on the actual seeds or bones themselves
(e.g. D. W. Phillipson 2000), as is now practicable through the Accelerator
Mass Spectrometry radiocarbon method. The evidence of rock art, in which
domestic animals are not infrequently depicted, may also be hard to inter-
pret since the absolute age of such images can only rarely be determined.
Farming of both plants and animals eventually gives rise in most cases
to physical differences which serve to distinguish the domestic forms from
their wild prototypes. The domestication process involves deliberate selec-
tion and control of breeding. For example, preference may be given to the
largest yams or to cereals which do not shed their seeds as soon as they
are ripe, but retain them when harvested. Animals of docile temperament,
perhaps of small size, will more readily be incorporated in controlled herds.
Thus, after a number of generations, signi¬cant physical differences may be
established. It follows from this that the initial stages of domestication,
whether of plants or animals, are correspondingly dif¬cult to recognise
in the archaeological record (Haaland 1992). Furthermore, it now appears
highly probable that the emergence of fully developed cultivating and herd-
ing economies was the result of a far longer period of intensive exploitation
and experimentation than was previously realised. The distinction between
gathering and hunting on the one hand and farming on the other is thus
far from clear. While both plants and animals usually undergo some degree
of morphological change under domestication, it can take very many gener-
ations before these become readily apparent. The absence of such changes
thus cannot be taken as proof that practices leading to cultivation and herd-
ing had not begun.
Early farmers 171

Archaeological evidence for cultivation and herding, other than actual
plant or animal remains, is usually less convincing. The implications of
artefacts are often ambiguous: sickles and grindstones for instance provide
no indication of the socio-cultural context of their use, and could have been
employed for gathering and processing wild plant foods, or for other mate-
rials altogether. The nature of an ancient settlement may provide indirect
evidence for its economic basis: permanence or seasonality of occupation,
for example, facilities for the storage of produce or for the watering or pro-
tection of livestock may all be indicative of farming practice but often fall
short of conclusive proof.
Despite the elusiveness of the evidence, the importance of the adoption of
farming techniques should not be underemphasised. Practice of such tech-
niques gave people greater control than they had generally exercised over
their own supplies of food and other commodities. Although concentrated
natural resources had in earlier times occasionally allowed maintenance
of semi-permanent settlements, these were usually small. In several areas
farming seems to have been adopted at least partly in response to environ-
mental deterioration and population pressure, as later sections of this chap-
ter will argue; and it in turn enabled populations to increase still further.
The relatively settled life which is inherent both to cultivators and, less
markedly, to herders provided a stimulus for the accumulation of material
possessions beyond those which could readily be transported. A sedentary
life-style could also have facilitated increased child-bearing, as pregnancy,
nursing and closely spaced births are all hindrances to mobility. Commu-
nities could now more readily afford to maintain the old, the disabled, or
members who specialised in activities other than herding or cultivation.
The increased sizes of these communities and the frequency with which
they came into contact with their neighbours must often have necessitated
the development of political structures more complex than those which
had existed among the simpler societies of earlier times (Ingold 1988). A far-
reaching corollary may have been the replacement of communal reciprocity
by forms of individual ownership.
Before detailing the archaeological evidence on a regional basis, it is use-
ful to consider some general points relating to its interpretation. Farming
depends on the availability of suitable plants to cultivate or animals to herd.
Such plants and animals may be obtained in one of two ways: by exercis-
ing control over wild varieties already present in the area concerned, or
by acquisition of plants or animals previously controlled elsewhere. Recent
research, as this chapter demonstrates, is showing that the former trend has
played a much greater part in the development of African farming than was
at one time believed. The latter trend has none the less been important, not
only in bringing to Africa plants and animals that had been domesticated
172 afric an archaeolog y

elsewhere, but also within Africa itself. It has been subject to the constraints
of numerous factors, including cultural preferences, environmental condi-
tions and the prevalence of plant and animal diseases (Dahl and Hjort 1976;
Gifford-Gonzalez 2000; Hassan 2002). Means of transmission need also to
be considered, as is also true of many other cultural elements. Domestic
plants and animals are, by de¬nition, under human control; they require
human intervention to move from one place to another. However, contrary
to the implication of much earlier writing on the subject, this does not
necessarily mean that whole human populations migrated along with their
herds or crops. A comparatively small number of individuals could have been
responsible for their introduction and, indeed, evidence of continuity in
artefact typology often indicates that this was probably so (D. W. Phillipson
1979; Karega-Munene 1996). Against this must be set the increasing body of
evidence from many parts of the world that the adoption of farming was
often accompanied by language-change (Holden 2002; Bellwood and Renfrew
Such developments were not, of course, automatic. On the one hand they
help to explain how such a complex civilisation as that of ancient Egypt
arose apparently less than 2500 years after farming was ¬rst adopted on a
signi¬cant scale in the Nile Valley. On the other hand, in many areas of
Africa, peasant communities have been able to maintain themselves with-
out centralised state systems into recent times (Bohannan et al. 1958). Nor
must it be thought that farming, once adopted, necessarily led to the rapid
abandonment of foraging; on the other hand, farmers seem rarely to have
reverted to hunting and gathering. Both activities continued to play an
important part in most precolonial African economies, while a few com-
munities maintained an almost exclusively hunting-gathering life-style into
the twentieth century. With low population densities, such as prevailed in
many parts of Africa until recent times, the natural resources are such that
hunting and gathering provided a level of nutrition as high as, or higher
than, that achieved by farming peoples (Lee 1968; see also Brooks et al. 1984).
It is nevertheless true that cultivation and/or herding have provided the eco-
nomic basis for most of the major technological, artistic and socio-political
achievements of African culture during the past 7000 years.

The Sahara and North Africa
In much of northern Africa, study of early farming has been hampered by
imprecise usage of the term ˜Neolithic™, varyingly to designate the practice
of cultivation and/or herding, the presence of pottery and/or ground-stone
tools, or merely attribution to an ill-de¬ned time-period. Because of this
confusion over its implications, the term is not used in this book.
Early farmers 173

The oldest plausible evidence for any form of farming in Africa comes
from the Egyptian Western Desert. The ¬rst farming communities in this
region appear to have been direct descendants of their late Pleistocene/early
Holocene predecessors, described in chapters 4 and 5. The microlithic stone-
artefact assemblages of the two phases are very similar and seem to have
more in common with contemporary materials to the north than with
those of the Nile Valley to the east. Indeed, the evidence currently avail-
able strongly suggests that farming began in the Western Desert at a date
signi¬cantly earlier than that of the corresponding development in the Nile
Valley (Hassan 1988b; Barker 2003).
The early and mid-Holocene in northeastern Africa, and almost certainly
elsewhere, was a period which saw much rapid climatic ¬‚uctuation. Some
episodes of major environmental change were remarkably brief (Hassan
2002). Although there is broad agreement about this general picture, the
detailed sequence and the dating of some individual episodes remain con-
troversial (Wendorf and Schild 2003); it cannot be assumed that ¬‚uctuations
in neighbouring areas were necessarily contemporaneous. The scale of some
of these changes was dramatic, with rainfall in certain areas up to ¬fteen
times their modern levels, and extensive lakes in areas that are now totally
arid (Petit-Maire 1991; Grove 1993). It is none the less important not to
assume that lush environments were ubiquitous in the Western Desert, or
elsewhere in the Sahara, during this period. There was undoubtedly more
surface water than in earlier or more recent times, but this was concen-
trated in and around the highland areas, or in localised lakes or ponds.
On the plains between these well-watered areas the vegetation remained
very sparse, and the fauna consisted of such creatures as ostrich and gazelle
which can survive in arid conditions. More varied faunas, like the human
population, were concentrated around the ponds and other better-watered
places (cf. Roset 1987; Vernet 2002).
During mid-Holocene times in the Western Desert, as in the Nile Valley,
people seem to have experimented with the control of wild animals, includ-
ing antelopes and giraffes. The evidence for this comes from rock art, notably
that at Jebel Uweinat in southeastern Libya, which unfortunately cannot be
dated (Fig. 79). Giraffes are shown tethered and being led by halters (Van
Noten 1978). With the concentrations of human and animal populations
which are attested, it is easy to see how such experimentation could have
taken place, perhaps following the exploitation of wild Barbary sheep dis-
cussed in chapter 5. Animals thus controlled may have been taken to the
Nile Valley, where they are known to have been in demand.
The ¬rst evidence for farming in the Western Desert comes from Bir
Kiseiba and from nearby Nabta Playa, a pond-basin of 100 square kilome-
tres near the Egyptian/Sudanese border (Wendorf et al. 2001). The sites
Fig. 79: Rock
engravings at
Jebel Uweinat,
showing giraffes
tethered or
restrained by a
halter (after Van
Noten 1978)
Early farmers 175

that were occupied beside the pond margin around the eighth millen-
nium bc were more extensive than their precursors, featuring house ¬‚oors,
hearths and rows of storage pits. It seems that the sites™ inhabitants were
indigenous to the area rather than new arrivals from elsewhere, and that
they continued the hunting-and-gathering economy of their predecessors.
The stone industry was essentially the same as it had been in earlier times,
but with the addition of concave-based points which may have served as
arrowheads (Wendorf and Hassan 1980; Banks 1984). Pottery bowls stamp-
decorated with a rocked comb were locally made. Hunting of the wild fauna
evidently continued, but cattle were increasingly represented from about
8000 bc; arguments that they were domesticated are now widely accepted
although the case is not wholly conclusive, being based primarily on the
size of the beasts and indications that human effort was devoted to water-
ing them (Gautier 1987, 2002; Close and Wendorf 1992; Wendorf and Schild
2003). It has been suggested, on not wholly persuasive grounds, that the
cattle may have been brought to this part of the Western Desert during a
period when the local vegetation was comparatively lush, having been ini-
tially domesticated elsewhere, conceivably in what is now the northwestern
Sudan. There can be little doubt that the presence of domestic bovids in
the eastern Sahara signi¬cantly preceded the local appearance of sheep or
goat, which are not attested at Nabta before 5800 bc. The abundant and
varied plant remains at Nabta include seeds of sorghum and millet, but
none was demonstrably domesticated; it is possible that incipient cultiva-
tion was practised but had not yet given rise to morphological change in
the grains concerned (Wendorf et al. 1998; Barakat 2002). It may have been
during the sixth millennium bc, when several periods of severe drought
are attested, that cattle herding extended into both the Nile Valley and the
central Sahara.
On the North African coast, to the west of the Nile, the beginning of farm-
ing is best illustrated at Haua Fteah in Cyrenaica (McBurney 1967). Here, by
the early sixth millennium bc or shortly thereafter, the local Libyco-Capsian
stone industry (p. 138, above) was followed by a period of occupation during
which the economy of the site™s inhabitants was based upon herds of domes-
tic ovicaprids. There is no evidence for the presence of domestic bovids,
which appear to have been restricted to more southerly regions at this time.
The artefact assemblages show signs of continuity from the Libyco-Capsian,
with pottery in evidence from about 5000 bc, but it is clear that the live-
stock must have been introduced to the area from elsewhere -- presumably
from the Levant. Not only does Haua Fteah provide no evidence for the
gradual development of herding, but the sheep/goats that were kept there
were clearly not descended from the wild Barbary sheep which had been
intensively hunted by the Libyco-Capsians and their contemporaries in the
176 afric an archaeolog y

Maghreb and the Fezzan. Interestingly, ovicaprids also made their appear-
ance at Sodmein Cave in the Egyptian Red Sea Hills at this same time -- about
5800 bc -- in generally similar circumstances (Vermeersch et al. 1996). The
evidence for domestic animals in Cyrenaica and the Red Sea Hills was thus
at least as early as that for the Nile Delta region (pp. 187--8 below). There is
no evidence that cereal agriculture, which was practised in the Delta, was
known further to the west at this time.
In the Maghreb (Sheppard and Lubell 1990), pottery appeared in the con-
text of the Capsian-related stone industries by the sixth millennium bc.
From this time onwards the somewhat diverse industries north of the Atlas
Mountains and extending eastwards to Tunisia have generally been classed
by archaeologists as ˜Neolithic of Capsian tradition™. In fact, several distinct
traditions may be recognised. In northernmost Morocco, adjacent to the
Straits of Gibraltar, the earliest pottery, as at El Khril near Tangier, is dec-
orated with impressions of cardium shells, in a manner widespread in the
western Mediterranean coastland (Jodin 1959). To the east, in coastal Algeria,
the pottery at Oued Guettara is impressed at the rim with sticks or plant
stems (Camps 1974). Throughout this area it seems likely that the introduc-
tion of domestic ovicaprids was broadly contemporary with the beginning of
pottery manufacture: bones of such animals were recovered from the lowest
levels of El Khril and may represent the ¬rst type of farming to be practised
in this part of North Africa. At Capeletti Cave in the Aures Mountains of east-
ern Algeria, transhumant herding of cattle and small stock was practised
from the mid-¬fth millennium bc onwards by people who do not seem to
have had any knowledge of cereal cultivation (Roubet 1979). The human
remains found associated with these ˜neolithic™ industries are described as
being of predominantly Mediterranean type, and it has been suggested that
some of these people were directly ancestral to the more recent Berbers
(McBurney 1975; Camps 1982).
In several areas of the central Sahara there is similar evidence for cultural
continuity between the earliest farmers and their predecessors -- who in
some instances included harpoon-using ¬shers. The sequence in southwest-
ern Libya is particularly informative (Garcea 2001; Cremaschi 2002). Atten-
tion was drawn in chapter 5 to the extensive former lakes in the Fezzan;
these shrank or disappeared during a brief arid interlude between 6500 and
6300 bc but then expanded again. The lakes of this second phase had camps
of cattle herders along their shores, with ovicaprids in evidence from about
5700 bc. Further evidence comes from the adjacent Acacus massif where the
skull of a shorthorn ox at Uan Muhuggiag appears to date from about 4900
bc. At the same site, rock paintings of cattle, buried in the archaeological
deposits, are argued to be earlier than 3400 bc (Mori 1965; Shaw 1977; Barich
1998; Cremaschi and di Lernia 1999). There were numerous grindstones but
no positive evidence that cereals -- or any other plants -- were cultivated;
Early farmers 177

Fig. 80: Artefacts
from Adrar Bous
(after Camps
1, bifacial
triangular knife;
2, bifacial
projectile point;
3, 4, backed

intensive use of wild cereals seems probable. The associated mode-5 stone
industry and stamp-decorated pottery are characteristic of those found on
sites of this time over an enormous area of the Sahara, although regional
variants may readily be recognised (e.g. Barich 1987, 1992, 1998; Muzzolini
1993; Close 1995; Barker 2003).
One such variant is the Tenerean, best known from Adrar Bous and Arlit
in Aïr but extending eastwards to Borkou in Chad (Bailloud 1969; J. D. Clark
et al. 1973). The industry (Fig. 80) dates between 5000 and 3300 bc, the skele-
ton of a domestic shorthorn ox from Adrar Bous (Carter and Clark 1976; Van
Neer 2002) being dated to the mid-¬fth millennium. Domestic small stock
were also herded but the sole evidence for plant exploitation, other than
the ubiquitous grindstones, consists of a single impression on a potsherd
of a grain that is thought to be sorghum. Hunting was also important,
and the prey included warthog, antelope, hippopotamus and rhinoceros.
Pottery and ground-stone axes show some resemblance to those from
the broadly contemporary Nile Valley sites of Esh Shaheinab and Kadero
(pp. 181--3 below). Backed microliths were abundant. Projectile points and
disc-shaped knives were bifacially ¬‚aked.
Our knowledge of the early Saharan herders may be ampli¬ed by study
of the rock paintings and petroglyphs which are widely distributed in the
highlands (Le Quellec 1987; Caligari 1993; Muzzolini 1995). Only rarely
has it proved possible to date individual paintings precisely, but several
attempts have been made to distinguish stylistic sequences, notably in the
Hoggar, Acacus and Tibesti highlands. These sequences may then tenta-
tively be linked by their subject-matter with the archaeological succession
to provide a provisional chronology (e.g. Muzzolini 1991, 1993). There may
have been an initial phase with engravings, rather than paintings, which
depict exclusively wild animals (p. 150); since, however, the same techniques
were sometimes used in representations of cattle, it cannot be regarded as
178 afric an archaeolog y

Fig. 81:
painted ¬gure,

conclusively proven that there is a ˜pre-pastoral™ phase in the Saharan rock
art sequence (Mori 1974). The most numerous paintings are those portray-
ing pastoral scenes in which long-horned cattle are associated with human
¬gures, some of which have distinctive rounded heads (Figs. 81, 82). Later
styles, believed to date from the late second millennium bc onwards, are
marked by the successive appearance of horses and camels (P. E. L. Smith
1968; Muzzolini 1986; A. B. Smith 1993a). The art shows many details of
Early farmers 179

Fig. 82: Rock
painting of a
pastoral scene,

clothing and illustrates the domestic, social and ritual life of the Saharan
herders (Dupuy 1993; Le Quellec 1993; Holl 1995). Breeds of cattle may also


. 5
( 11)