. 3
( 11)


Fig. 30: Handaxes
and other
preserved as
discovered at

and Harris 1998). This evidence, indicating settlement of terrain more than
2300 metres above sea level now frequently subject to night frost, comes
from several areas. At Melka Kunture near Addis Ababa (Chavaillon 1979),
a long series of occurrences, including some which have been described as
˜Oldowan™, has been investigated but remains incompletely published. One
of these Acheulean-type occurrences has yielded a jaw fragment attributed
to Homo ergaster. On the Gadeb Plain to the southeast, early Acheulean-like
material is of particular interest as it provides evidence that raw material for
tool-making was carried over a substantial distance ( J. D. Clark and Williams
1978; J. D. Clark 1987).
Some of the most informative occurrences of Acheulean-type artefacts
in Africa are at Olorgesailie in the Kenyan Rift Valley, some 50 kilometres
southwest of Nairobi (Isaac 1977; Potts 1989). Here, several major concentra-
tions of handaxes, cleavers and other stone artefacts were deposited over a
long period in a succession of lake-shore environments (Fig. 30). It is evident
that the places most favoured for settlement were either on a low rocky
promontory, or on patches of sand which generally occurred in the chan-
nels of seasonal streams draining into the lake. Acheulean people seem to
have preferred to camp in these dry watercourses rather than on the shore
of the lake itself. Comparison with modern Rift Valley lake basins suggests
reasons for these preferences: the stream beds provided open sandy ground
in an area that was elsewhere covered with hummocks of coarse grass, and
The consolidation of basic human culture 67

the trees lining the channel supplied welcome shade that was absent on
the lake shore. Water from holes dug in the bed of the channel was likely
to be less saline than that of the lake. The rocky promontory might have
been used in the wet season, when the streams were ¬‚owing, and when the
extra height provided relief from mosquitoes.
Most of the Olorgesailie Acheulean-type occurrences are thought to have
suffered some degree of sorting or movement as a result of water action
since the sites were abandoned. However, even if the artefacts were not
recovered absolutely in their original positions, it was possible to estimate
from the extent of the artefact scatters that the camp areas were generally
between 5 and 20 metres across. The larger scatters may represent areas
of successive shifting use, or the presence of groups larger than a single
nuclear family.
Bone was not always well preserved at Olorgesailie, and no hominid fossils
have been found there. At one locality remains of many giant baboons were
discovered, and these may represent a troop that was attacked whilst sleep-
ing in the trees overgrowing the site. The presence of many stones, which
must have been carried to the site from elsewhere and which were of a
suitable size and shape for use as missiles, lends support to this suggestion.
This may be an early instance where hunting, as opposed to scavenging, is
Many of the Olorgesailie artefacts are made of types of stone which must
have been brought there from some distance away. The scale of this transport
makes it seem highly likely that, by this stage in their development, people
had invented some sort of bag for carrying their possessions from one place
to another. While it seems highly probable that the makers of Acheulean-
type artefacts at Olorgesailie, like their counterparts elsewhere, collected
and made much use of vegetable foods, no ¬rm evidence for this has been
A particularly noteworthy feature of these lithic assemblages is the range
of their variation, both in composition and in artefact style. In some, han-
daxes and cleavers were the most frequent tool-types; in others, they were vir-
tually absent. While in some instances size-sorting of the artefacts by water
action may have contributed to this variation, it is very unlikely to have
been the prime cause. Also, in the same area and at the same time there is a
great range in quality of workmanship and other stylistic features within the
same artefact categories. It seems reasonable to suppose that this variation
was primarily due to differences in the skill and preferences of individual
tool-makers or social groups, as well as to the intended uses of the arte-
facts. If this were the case, it might imply that tool-making was essentially
a taught skill (I am indebted to Dr L. Phillipson for this suggestion). Further
68 afric an archaeolog y

examples of this variation may be seen at the Kenyan sites of Kilombe and
Isenya (Gowlett 1978; H. Roche et al. 1988).
The various occurrences of Acheulean-type artefacts at Olorgesailie are all
contained within a single geological formation, although they are scattered
through about 40 metres of its vertical thickness. It is therefore probable
that they were deposited at intervals over a substantial period of time, per-
haps extending over thousands of generations. Unfortunately, it has not
proved possible to date the site very precisely. The best interpretation of
potassium/argon dates for pumice contained in the Olorgesailie Formation
suggests an age between 1.0 and 0.5 million years ago (Bye et al. 1987; Potts
1989). This is not contradicted by other lines of investigation, such as faunal
comparisons with dated sites elsewhere in East Africa.
A later, perhaps near-¬nal, stage of the East African Acheulean may be
represented by a poorly understood site on the Kinangop Plateau of central
Kenya which has yielded a potassium/argon date in the order of 400,000
years for an industry of small handaxes, ¬‚ake points and side-scrapers
(Evernden and Curtis 1965). Similar developments may be discerned else-
where in East Africa, as at Kapthurin near Lake Baringo, Kenya, where a
hominid mandible is claimed to be associated with an Acheulean-type lithic
assemblage showing use of a well-developed prepared-core technique (D. M.
Leakey et al. 1969; Cornelissen et al. 1990). An age of about 200,000 years has
been suggested, but comparison with evidence from other regions indicates
that this may be too young.

Acheulean in south-central Africa
Some relatively late occurrences of Acheulean-type artefacts occur at a site
at Kalambo Falls near the southern end of Lake Tanganyika (Fig. 31); they
are particularly important because of the preservation of wood and other
plant remains. The Kalambo River, which here forms the border between
Tanzania and Zambia, ¬‚ows over the edge of the Rift escarpment as a
spectacular 220-metre single-drop waterfall. Immediately above the water-
fall, the Kalambo Valley is very narrow; successive blocking and unblocking
of this stretch has on several occasions caused the formation of a small
lake in the wider section of the valley further upstream. In the silts and
shore deposits of these ancient lakes abundant traces of prehistoric occupa-
tion have been preserved, both in near-primary and in secondary contexts
( J. D. Clark 1969, 1974, 2001a). The earliest such traces so far recovered repre-
sent two phases associated with late Acheulean-type handaxes and cleavers,
but it is not impossible that even older material may lie beneath the modern
water level.
The consolidation of basic human culture 69

Fig. 31: The
Kalambo Falls.
The Acheulean
and later
sites are in the
small lake basin
through which
the Kalambo
River ¬‚ows above
the Falls.

At an early stage in the research at Kalambo Falls, radiocarbon dates were
obtained which indicated for these Acheulean-type artefacts an age of about
55,000 to 50,000 years. These results were at the very furthest end of the
radiocarbon dating method™s time-range; and more recent work has shown
that they should be regarded as representing minimum, rather than ¬nite,
ages. New uranium-series analyses and comparison with dated successions
in other regions suggest that the true age of this material may be between
400,000 and 300,000 years ( J. D. Clark 2001a: 25--8, 665--74). Broadly similar
70 afric an archaeolog y

artefacts preserved in a riverside situation at Isimila in southern Tanzania,
about 300 kilometres east of Kalambo, are probably of comparable antiquity
(Howell et al. 1962; Hansen and Keller 1971), and the same may be true of
the Kamoa site in D. R. Congo some 500 kilometres to the southwest (Cahen
The artefact-bearing horizons at Kalambo cover substantial areas which
probably represent shifting foci of small-scale activity. On several horizons
at the bottom of the investigated sequence, abundant Acheulean-type arte-
facts were recovered, apparently only minimally disturbed since they were
originally abandoned by their makers on banks of sand and gravel beside
the river, presumably during the dry season. The handaxes and cleavers
are exceptionally large and uniformly ¬nely worked, owing probably in
large part to the excellent ¬ne-grained silici¬ed sandstone that was plen-
tifully available in the immediate vicinity. They show little if any evidence
for the development of a prepared-core technique. Bone was not preserved,
but wood did survive; several pieces show signs of having been cut or shaped
deliberately (Fig. 32). Traces of burning on some of the wood indicate the
presence of ¬re, but there is no way of knowing whether this ¬re was used
and controlled by the site™s inhabitants, or merely derived from naturally
generated ˜bush™ ¬res. In this context it may be signi¬cant that none of the
areas investigated presented features which could plausibly be interpreted
as hearths. Particularly interesting is an arc of stones enclosing an area of
about two square metres which, it has been suggested, may represent the
base of a simple shelter or windbreak. Two hollows ¬lled with compressed
grass may have been sleeping places. Seed pods and the remains of fruit
show that a range of vegetable foods was exploited.
The silts which incorporated the Acheulean materials at Kalambo Falls
contained pollen grains which have been identi¬ed to provide a reconstruc-
tion of the local vegetation. Conditions warmer and drier than those of today
are initially attested, with swamps and riverine gallery forest in the valley,
but dry woodland beyond. The ¬nal Acheulean-type artefacts at Kalambo
appear to have been deposited during a time with a cooler, damper climate,
associated with a type of forest which today only survives in this part of cen-
tral Africa at considerably higher altitudes; these assemblages show trends
towards some of the typological features characteristic of later periods ( J. D.
Clark 1964). Further climatic change may also provide the background for
subsequent developments in the Kalambo Falls sequence; those designated
Sangoan are described below (pp. 81--2) and the later ones in chapter 4.
Although long-since destroyed, a cave at the Broken Hill mine near the
town of Kabwe on the central plateau of Zambia is of great signi¬cance
because of the discovery there, during mining operations in 1921, of human
The consolidation of basic human culture 71

Fig. 32: Wooden
objects, with
probable traces
of working, from
Acheulean levels
at Kalambo Falls
(after J. D. Clark

remains representing at least four individuals, including a superbly pre-
served cranium originally attributed to Homo rhodesiensis (Pycraft et al. 1928).
Unfortunately little was recorded concerning the precise circumstances of
the discovery, and its true archaeological associations are consequently
uncertain ( J. D. Clark et al. 1950; J. D. Clark 1959). However, as will be shown
below, there is some evidence that the Broken Hill cranium may belong to
the same time-span as the late Acheulean-type assemblages of about 400,000
years ago; the site is therefore discussed here.
Most of the artefacts recovered from the Broken Hill cave, although not
necessarily in direct association with the cranium, are ¬‚akes and retouched
¬‚akes, some of which were struck from prepared cores. In the absence of
72 afric an archaeolog y

Fig. 33: Skull
from Broken Hill
Mine, Kabwe,
attributed to
heidelbergensis or
to an archaic
form of H. sapiens

handaxes, cleavers or other core-tools, they were originally attributed to
the post-Acheulean industry known as Charaman and now regarded as a
facies of the Sangoan (p. 81). An occurrence with handaxes of Acheulean
type has subsequently been located in the vicinity, and other assemblages
lacking handaxes are now known from the same time-span. Of the large
animal species represented in the abundant faunal collection from the
cave, a quarter are now extinct, suggesting a Middle-Pleistocene age (Klein
The Broken Hill cranium (Fig. 33), with its pronounced brow ridges and
receding forehead, has a capacity of 1280 cubic centimetres, which is well
within the range shown by modern populations. Its closest parallel is the
skull from Bodo in the Middle Awash Basin of Ethiopia, noted above. The
evolutionary position of these hominids is discussed below; here it needs
merely to be noted that some authorities attribute them to Homo heidelber-
gensis, others to an archaic sub-species of H. sapiens intermediate between
H. ergaster and H. s. sapiens. Despite the uncertainties about its age and
archaeological associations, the specimen is of major signi¬cance as one
of the few hominid fossils from sub-Saharan Africa which may represent
the makers of the later Acheulean-type artefacts (Klein 1999).
The consolidation of basic human culture 73

Acheulean in southern Africa
Despite the large mass of Acheulean-type artefacts that has been recovered
from south of the Zambezi, the amount of reliable information which it
yields is disappointingly meagre (Klein 2000b). Although hundreds of occur-
rences are known, less than a score come from sealed contexts and even
fewer relate to single events (cf. Shackley 1985). The absence of volcanic activ-
ity during the Pliocene and Pleistocene epochs in southern Africa means
that none of this material has been dated by the potassium/argon method;
estimates of its age depend very largely on correlations of faunal assem-
blages, climatic ¬‚uctuations and geomagnetic reversals with those known
and dated in East Africa. The general picture that emerges is that the oldest
artefacts of Acheulean type so far recognised in southern Africa are from
Sterkfontein near Krugersdorp, dated less than 1.7 million years ago, prob-
ably between 1.5 and 1.0 million years ago. Occurrences of artefacts and/or
Homo ergaster fossils here and at the nearby sites of Swartkrans, Kromdraai
A and Drimolen (Kuman et al. 1997; Kuman and Clarke 2000) may be broadly
Hominids may not have been present in cooler (southerly) and drier (west-
erly) parts of South Africa until after 1.0 million years ago, when Acheulean-
type artefacts are ¬rst attested there. On faunal grounds it may be suggested
that one of the oldest such occurrences in these regions is that at Cornelia
in northern Free State Province (H. B. S. Cooke 1974), and one of the most
recent at Duinefontein 2, near the Atlantic coast not far to the north of
Cape Town (Klein et al. 1999).
Among other South African sites yielding artefacts of Acheulean type, the
few which are thought on geological or technological grounds to be early are
all in river gravels or other disturbed contexts: these include Klipplaatdrif
and Three Rivers near Pretoria (Mason 1962). At the nearby suburb of
Wonderboompoort, large quantities of Acheulean-type artefacts accumu-
lated in a natural de¬le which could have been used for trapping game.
Unusually, some of the most informative occurrences of such materials in
South Africa, all probably late, have been found in caves: the Cave of Hearths
(Mason 1962) near Mokopane, Wonderwerk near Kuruman (Beaumont 1990;
Fig. 34), and Montagu Cave (Keller 1973) in Western Cape Province.
Further sites of note are those in spring deposits at Amanzi near
Uitenhage (H. J. Deacon 1970), where vegetable remains including worked
wood fragments are also preserved, on an 18-metre raised beach at Cape
Hangklip (Sampson 1974), and in eroding sands at Hope¬eld (also known
as Elandsfontein) north of Cape Town. The Hope¬eld deposits (Singer and
Wymer 1968) have also yielded a hominid skull-vault with pronounced brow
74 afric an archaeolog y

Fig. 34:

ridges, variously attributed by earlier writers to Homo erectus or to an early
form of H. sapiens, but now generally regarded as a further example of
H. heidelbergensis which is, on the evidence of the associated fauna, some-
what earlier in date than the specimen from Broken Hill mine. There is no
reason to believe that any of these Cape occurrences, or others in the arid
western parts of southern Africa, are signi¬cantly older than about 600,000
years. A partial skull from Florisbad in Free State Province (Clarke 1985) is
probably even more recent, as is discussed below in chapter 4.
It has long been recognised that use of a prepared-core technique has a
high antiquity in South Africa, where it has been conventionally designated
the ˜Victoria West technique™, used to produce large ¬‚akes as blanks for
handaxes and cleavers (Fig. 35; McNabb 2001). Recent investigation of a site
on the Harts River near Taung (Kuman 2001) has clearly demonstrated that
this technique arose in a local Acheulean context.
In parts of the South African interior, the ¬nal Acheulean assemblages
contain small well-¬nished handaxes with rounded butts and concave sides,
formerly regarded as representing a distinct ˜Fauresmith industry™. It is now
realised that these artefacts are almost always found in those areas where
lydianite (hornfels) was the preferred raw material for tool manufacture
(Humphreys 1970); their technological distinctiveness, however, remains
incompletely understood. One of the most informative assemblages is that
from Rooidam near Kimberley (Fock 1968). Sampson (2001) has argued that
southern Africa at this time supported a denser human population than
other parts of the continent or, indeed, of the world.
The consolidation of basic human culture 75

Fig. 35: Artefacts
from a site on
the Harts River
near Taung,
showing the
˜Victoria West
technique™ of
(after Kuman
2001): 1, 4,
prepared cores;
2, 3, cleaver and
produced from
such ¬‚akes

Acheulean in West Africa and the Sahara
The archaeological evidence for the ¬rst human settlement in West Africa
is still poorly known. It seems probable that this area was not inhabited by
the very earliest hominids, such as are known from the eastern part of the
continent; it is nevertheless possible that such populations were originally
present but that their remains have not been preserved or are yet to be
discovered. For example, archaeological traces of this period may perhaps
be preserved in the Chad area, where lake deposits are known to extend to
great depths (Tillet 1985).
There can be no doubt that true handaxes of Acheulean type were some-
times made in West Africa, although there is as yet no convincing demon-
stration of their age. The most notable occurrences are along the Volta River
in Ghana, and on the Jos Plateau in Nigeria (Fig. 36). In the latter area (Soper
1965), river gravels mined for tin extraction have been found to contain
numerous handaxes, cleavers and other Acheulean-type artefacts, as well as
material which is clearly of much more recent date. None, however, has
been recovered from primary or datable contexts.
Surveys of other West African occurrences of early artefacts, mostly in
disturbed or poorly documented contexts, have been reported by several
archaeologists (Davies 1964, 1967; Shaw 1985; Allsworth-Jones 1987, 2001). A
number of places in West Africa, notably Beli in northeastern Nigeria, have
yielded artefacts which would not be out of place in either an Oldowan or
an Acheulean assemblage. However, since these discoveries all come from
disturbed and undated contexts, there is no convincing evidence for an
Oldowan presence in West Africa.
76 afric an archaeolog y

Fig. 36: West
artefacts from
the Jos Plateau,
Nigeria (after
Shaw 1978):
1, 2, handaxes;
3, cleaver

Similarly, in the Sahara, although artefacts of Acheulean type have been
found in many places, there are very few assemblages that have been care-
fully excavated from undisturbed situations ( J. D. Clark 1980, 1992). The
occasional collections that appear on typological grounds to be early are
all from secondary contexts, most notably in southern Morocco and central
Mauritania. Later Acheulean material seems to be more widely distributed.
Where its original geological associations can be ascertained, these indicate
conditions wetter than those of the present, and this is con¬rmed both by
study of the faunal remains and by pollen analysis. It appears that there were
major ¬‚uctuations in the Saharan climate during the time when Acheulean-
type artefacts were produced. Only local studies of these have so far been
undertaken and no overall picture has yet emerged. As in other areas, water-
side locations were clearly preferred, and it may be that human habitation
was far more extensive during wetter phases. A typical Saharan site is at
Erg Tihodaïne, in southern Algeria between the Tassili and Hoggar massifs,
where presumably late artefacts of Acheulean type occur in the clays of a
former swamp, together with bones of extinct types of elephant and buffalo,
as well as other creatures which cannot survive in the area today (Reygasse
1935; Arambourg and Balout 1955; H. Thomas 1979).
Several regional technological and stylistic variants may be recognised
in the Saharan Acheulean. For instance, there is an incidence of devel-
oped prepared-core technique in the Wadi Saoura of central Algeria, as at
Tachengit: cores were prepared to yield large ¬‚akes requiring only minimal
trimming to produce effective cleavers (Balout 1967; Alimen 1978; Dauvois
The consolidation of basic human culture 77

1981). There is, however, little chronological information, so we cannot ascer-
tain whether these variants were in fact contemporary with one another.

Acheulean in North Africa
Parts of North Africa have yielded evidence for a long sequence of Acheulean-
type industries; it has also been suggested that there may be indications
of earlier human occupation akin to the Oldowan. Unfortunately, direct
dating of these North African sites has not proved possible. The tentative
chronology originally proposed was based upon faunal correlations with
the European sequence and upon connexions with high sea levels which,
it was believed, could be attributed to phases of the northern-hemisphere
glacial succession. Both methods, especially the latter, have been shown to be
tenuous and now merit little reliance; comparisons with the dated material
from eastern Africa are thus dif¬cult to establish. The current consensus is
that hominid penetration of large areas of the Sahara and of northwestern
Africa did not take place until, at the earliest, 1.0 million years ago, perhaps
as late as the beginning of the Middle Pleistocene ( J. D. Clark 1992).
The most complete picture comes from the Casablanca area on the
Atlantic coast of Morocco (McBurney 1960; Biberson 1961, 1967, 1971; Deb©-
nath 2000). The detailed subdivision of this sequence into technological
phases, as originally proposed, is now felt to lack adequate support since
in relatively few instances were assemblages recovered from primary con-
texts. The earliest traces of human activity in this region are associated with
beach deposits about 100 metres above the modern sea level. It is generally
believed that this beach was formed during a period of warm climate at a
relatively early phase of the Middle Pleistocene. Two series of stone artefacts
have been recognised in these deposits, distinguished mainly on the basis
of their physical condition. The rolled series, evidently derived from still
earlier deposits, consists of pebbles from which ¬‚akes had been removed
in one direction only. The second series, which is in fresher condition, has
had ¬‚akes removed in two directions so as to produce a jagged cutting or
chopping edge. Unfortunately, there are very few artefacts in either of these
series, and it must be emphasised that they have not been recovered from
their original positions, being disturbed by wave action, soil slip and other
natural agencies.
The ¬rst appearance of handaxes in the Casablanca sequence is in deposits
which, being between 60 and 70 metres above the present sea level, must
be signi¬cantly later than those described in the preceding paragraph. The
most generally accepted dating for this episode is approximately 500,000
years ago. If this order of magnitude is supported by future research (and it
78 afric an archaeolog y

must be admitted that the chronology at present remains extremely tenta-
tive), then the development of the Acheulean in Morocco must have lagged
far behind the corresponding events in eastern Africa. On the basis of the
same correlation, even the earliest occurrence of humanly chipped stone at
Casablanca may be less than 1.0 million years old. It is clear that human
occupation of North Africa and Europe covers a signi¬cantly shorter time-
span than the corresponding sequence in eastern Africa (Isaac 1972; Klein
1999; Rolland 2001). The relationship with analogous occupation of the
Iberian peninsula remains to be satisfactorily demonstrated.
Some degree of con¬rmation for the presence in northwestern Africa of
early stone industries lacking handaxes comes from the site of Ain Hanech
near El Eulma in Algeria, where the faunal remains, including many species
now extinct, suggest an age at the beginning of the Middle Pleistocene, about
700,000 years ago, if not earlier (Arambourg and Balout 1952; McBurney
1960; Sahnouni 1998). The Ain Hanech artefacts (Fig. 37) are ¬‚aked to a
roughly spherical shape, with jagged edges; they resemble in some respects
objects which are frequent in the East African assemblages designated
˜Developed Oldowan™. A few handaxe-like objects from Ain Hanech may
be attributable to a subsequent phase represented in a higher layer. The
Terni¬ne site (now Tighenif), not far from Oran in northwestern Algeria,
lies beside a spring whose deposits contained abundant faunal remains of
early Middle Pleistocene types, demonstrably later in date than those from
Ain Hanech. The Terni¬ne artefacts appear to belong to a fairly early stage of
the North African Acheulean (Balout et al. 1967; Geraads et al. 1986). Three
human mandibles and a skull fragment, all exceptionally well preserved,
are of Homo ergaster type (Fig. 38).
The Casablanca sequence of Acheulean-type industries is best exposed at
the Sidi Abderrahman and Thomas quarries (Fig. 39; see also Raynal and
Texier 1989). While the sea was retreating from a level of undetermined max-
imum height, handaxes, cleavers and other artefacts were laid down upon
the beach and subsequently buried beneath a massive sand dune which,
as the sea level dropped still lower, accumulated on the old shoreline. This
beach, consolidated into sandstone, is exploited by quarrymen whose activi-
ties have been responsible for unearthing much of the Casablanca sequence.
Later, the sea rose again, cutting back the dune to form a cliff, in which
several caves came into being. Beach deposits in these caves are between 27
and 30 metres above modern sea level, which enables their formation to be
correlated provisionally with some part of a prolonged interglacial period
around 300,000 years ago, to which many important European Acheulean
occurrences also belong. Two of these caves exposed at Sidi Abderrahman,
the Grotte de l™Ours and the Grotte des Littorines, are of particular signif-
icance because they were occupied by people soon after the formation of
The consolidation of basic human culture 79

Fig. 37: North
African artefacts
(after McBurney
1960): 1, 2, ¬‚aked
spheroids from
Ain Hanech;
3, 4, cleaver and
handaxe from
Sidi Zin

the 30 metre beach. Abundant well-made handaxes, cleavers, ¬‚ake tools and
other artefacts were recovered; many of both the handaxes and the cleavers
were made on ¬‚akes struck from remarkably massive cores. Several succes-
sive handaxe-bearing horizons occur in these caves and, in one of the upper
ones at the Grotte des Littorines, part of a human lower jaw was recovered,
80 afric an archaeolog y

Fig. 38: Mandible
of Homo ergaster
from Terni¬ne
(now Tighenif)

Fig. 39:
section of
deposits exposed
in the walls of
the Sidi
quarry near
Casablanca (after
McBurney 1960).
The ¬gures on
the left indicate
heights above
mean sea level.

closely similar to somewhat earlier examples from Terni¬ne in Algeria. Orig-
inally attributed to Homo erectus/ergaster, these fossils are not now considered
to be more than 300,000 years old and are widely regarded as an early mod-
ern form retaining some H. erectus features (Hublin 1985). Probably some-
what more ancient is a similar mandible from the nearby Thomas quarry
The consolidation of basic human culture 81

(Sausse 1975), also associated with artefacts of Acheulean type, albeit a sparse
occurrence. The ¬nal Acheulean assemblages at the Sidi Abderrahman caves
include superbly ¬nished handaxes with a twisted, S-shaped pro¬le and also
¬‚akes struck from carefully prepared cores. Although none of these coastal
sites has yielded any evidence for the exploitation of marine foodstuffs, this
does not necessarily mean that such were not used.
To the east, the Libyan coastland has yielded no informative occurrences
of Acheulean-type materials; such information as we possess about the
Acheulean of Northeast Africa comes from the Nile Valley and its immedi-
ate environs. Here there are numerous occurrences in the Wadi Halfa area
which may pre-date the Nile™s capture of its southern tributaries. Where
this material is in an undisturbed context, it appears to represent quarry or
tool-making sites rather than settlements (Chmielewski 1968). There is no
evidence for the absolute age of this material, nor of that recovered from the
alluvial terraces of the Egyptian Nile Valley ( J. D. Clark 1980, 1992; Hassan

Sangoan assemblages
As noted above, the stone-tool assemblages immediately following the
Acheulean over a huge area extending from Botswana to Ethiopia, includ-
ing also Sudan and West Africa (Fig. 40), are of the type generally desig-
nated Sangoan. They are characterised by the crude, pointed, triangular-
sectioned core-tools often referred to as picks, thick handaxes or core-axes
and a variety of small retouched or utilised ¬‚akes (Fig. 41). These components
occur in very variable frequencies. The heavy core-tool element is generally
dominant at sites which occupy river-valley situations or thickly wooded
environments; it is assemblages of this type that have conventionally been
designated Sangoan. In more open-plateau situations, notably between the
Zambezi and the Limpopo, the light-duty ¬‚ake-tool element dominates the
assemblages, which have been regarded as belonging to a Charaman industry
taking its name from an open site in northwestern Zimbabwe. The Sangoan
and Charaman are here regarded together as linked facies in a continuum
of variation which began in the Acheulean ( J. D. Clark 1982a).
Some of the most informative Sangoan occurrences overlie the Acheulean
at Kalambo Falls (Sheppard and Kleindienst 1996; J. D. Clark 2001a). Particu-
larly abundant remains of this period also occur in the valley of the Kagera
River, west of Lake Victoria, on the modern border between Uganda and
Tanzania. All the Kagera material, however, occurs in disturbed river-gravels;
East Africa has so far yielded very few undisturbed Sangoan horizons com-
parable with those at Kalambo Falls. The principal exception is at Muguruk,
near the Winam Gulf of southwestern Kenya, where a near-pristine industry
82 afric an archaeolog y

200 400 1000 km
Fig. 40: The
distribution of
100 500 miles
Sangoan sites

investigated by McBrearty (1988) includes Sangoan-type artefacts in an
undated context without associated fauna.
Detailed study of this Sangoan/Charaman material has been hampered
by the fact that the only unmixed unsorted assemblages from strati¬ed
contexts are either extremely small, like those designated Charaman from
the basal horizons of Bambata and Pomongwe caves in the Matopo Hills of
Zimbabwe (C. K. Cooke 1963, 1969), or lacking in associated fauna, like the
Sangoan from Kalambo Falls ( J. D. Clark 1964, 1969, 2001a). Much of this
material has an antiquity in excess of 250,000 years, although the possibility
that the Zimbabwe occurrences may be more recent cannot be ruled
At least the more southerly parts of West Africa appear to fall within
the distribution of Sangoan-type artefacts (see p. 81, above) similar to those
The consolidation of basic human culture 83

Fig. 41: Sangoan
artefacts from
the Luangwa
Valley, Zambia
(1--3, after J. D.
Clark 1950) and
from Kalambo
Falls (after J. D.
Clark 2001a):
1, core or
2, edge-retouched
¬‚ake; 3, pick;
4, 5, core-axes

found further to the south and east, although certain features serve to
differentiate the West African ¬nds from their neighbours. Sangoan arte-
facts occur at numerous localities in the gravels and terraces of the major
rivers of Nigeria and Ghana (especially the Volta) and in the coastal regions
of Ghana, where drier conditions would have caused contraction of the
forest. With the exception of a few ¬nds near the headwaters of the Gambia
River in southern Senegal and Guinea, artefacts of Sangoan type have not
been convincingly reported from west of the Ivory Coast; this absence
may to some extent re¬‚ect the vagaries of research emphasis (Davies 1964,
1967; Wai-Ogusu 1973; Paradis 1980; Chenorkian 1983; Allsworth-Jones 1986,
It has been argued that some of the typical Sangoan-style tools were used
for woodworking, and that the development of the Sangoan was a response
to a wetter climate and a resultant spread of forest cover ( J. D. Clark 1970).
84 afric an archaeolog y

Although plausible, such an argument rests upon very slender foundations,
for we have no direct information about the uses to which individual arte-
facts were put, and the evidence for a shift to wetter conditions at this time
is less convincing than was once believed. The scarcity of single-episode,
well-studied assemblages precludes a more detailed consideration of this
problem, although related phenomena in several regions of Africa will be
discussed below. Unfortunately no comprehensive faunal assemblages come
from demonstrably Sangoan or Charaman contexts in this part of Africa,
nor are there any associated hominid fossils.

Acheulean/Sangoan artefacts and their makers
The initial appearance of Acheulean-type artefacts in eastern Africa is, at
¬rst sight, puzzling. The former view, based on the sequence at Olduvai
Gorge, that there was a demonstrable linear development from Oldowan,
into Developed Oldowan, into Early Acheulean, is now disproved stratigraph-
ically and has few adherents. In contrast, some archaeologists have suggested
that the ¬rst appearance of Acheulean materials at Olduvai may represent an
intrusion into the sequence of a different stone-tool-making tradition whose
earlier stages had taken place elsewhere. This hypothesis is now supported
by the discovery of Acheulean-type artefacts in earlier contexts further to
the north, at Konso and West Turkana. Perhaps the most reasonable inter-
pretation of the currently available facts is that the Acheulean, beginning
about 1.7 million years ago in what is now Ethiopia, represents the domi-
nant surviving technological tradition which prevailed after a long period
which saw accelerated diversi¬cation in the physical and cultural evolution
of hominids.
In eastern Africa, hominid fossils of this period include at least three
species and, according to some classi¬cations, as many as six; the arte-
facts are similarly varied and include the types which have been named
Karari, Developed Oldowan and Early Acheulean (see Fig. 42). It is relevant
to consider the extent to which these assemblages represent distinct cul-
tural developments, perhaps being the work of different groups of hominids,
rather than variations of a common tradition. Archaeologists are far from
agreement concerning the signi¬cance of this diversity, but there is increas-
ing support for the view that it may represent diverse facies of a single
tradition. The point may be understood by considering the status of the
Karari industry (p. 60 above). Isaac (1997) regarded the Karari industry as
a variant of the Oldowan; here, by contrast, it is viewed as part of the
diverse development which ultimately gave rise to the Acheulean (see also
Cachel and Harris 1998). The Karari industry represented a major advance
The consolidation of basic human culture 85

Fig. 42: Early
traditions of
lithic technology
in Africa

in tool-making ability as illustrated both by the manual dexterity and con-
trol of stone-¬‚aking technique which the artefacts demonstrate, and by its
makers™ ability to visualise and then to produce a standardised artefact type
whose ¬nal shape bore little or no resemblance to that of the initial piece
of raw material. The Karari industry was but one of several parallel devel-
opments which appear in the archaeological record at approximately this
time, re¬‚ecting a probable period of experimentation and the development
of tools speci¬cally intended for particular purposes. This in turn led to the
teaching and learning of technological tradition and to the beginnings of
86 afric an archaeolog y

standardisation in place of the more random tool-making of earlier times.
Such periods of accelerated evolution separated by consolidation and conti-
nuity have been recognised as recurrent features of hominid development
(cf. Isaac 1972) and, indeed, of most forms of animal evolution. Not infre-
quently, they may be linked with periods of environmental stress (cf. Vrba
et al. 1992; Foley 1994; Reed 1997). So far in this book, we have recognised
such periods approximately 5.5, 2.8 and 1.7 million years ago (cf. pp. 24 and
27--8, above); it is noteworthy that all appear to have been periods when com-
paratively cool, dry conditions prevailed (Williams et al. 1993; de Menocal
1995). Such conditions seem to have been particularly conducive to acceler-
ated evolution and diversi¬cation.
From the middle of Bed II onwards, the Olduvai sequence contains varied
assemblages which some archaeologists (e.g. M. D. Leakey 1971) have inter-
preted as representing two parallel traditions -- Developed Oldowan and
Acheulean. It is, however, hard to envisage how they could have maintained
their separate identities for the enormously long period -- about 500,000
years -- which this view requires, even if (as some have suggested) they were
the work of different hominid species existing at the same time and in the
same place. A more economical hypothesis, therefore, is that these assem-
blages represent different aspects of a single tradition, such as continued
also into far more recent times when it was characterised more exclusively
by artefacts of Acheulean type (Gowlett 1986). The in¬‚uence of raw-material
availability on this early diversity has not yet been evaluated conclusively.
While it has been demonstrated that different types and sizes of rock were
used as raw materials for the various Olduvai assemblages, further research
is needed to ascertain whether these were deliberately selected or imposed
by availability at a particular time and place (cf. Stiles 1991, 1998). A simi-
lar situation has been observed in the broadly contemporary Middle Awash
sequence, where it has been suggested that mode-1 technology may have
been employed for activities undertaken in stable ¬‚oodplain environments,
mode-2 assemblages corresponding to situations beside shifting streams
( J. D. Clark et al. 1994; J. D. Clark and Schick in de Heinzelin et al. 2000).
It is only very rarely that the archaeological record yields convincing evi-
dence for the complete replacement of an established technology by an
In due course, the African Acheulean provided the context for major cul-
tural developments. It is useful to view these against the background of the
considerable physical evolution which took place among the hominids that
were responsible for the Acheulean artefacts. There can be little reasonable
doubt that the earlier African Acheulean, in the wide sense of that term here
proposed, was the work of hominids varyingly attributed to African Homo
erectus or to H. ergaster (cf. Wood 1992). As we saw, fossil remains of this type
The consolidation of basic human culture 87

which have been recovered from eastern and South Africa illustrate a rel-
atively large-brained (average around 900 cubic centimetres) hominid with
heavy brow ridges and a low cranial vault (Rightmire 1990, 1996; Klein 1999).
At Olduvai, it appears that both H. ergaster and a robust australopithecine
were present at this time, represented by the middle and upper levels of
Bed II. If only one hominid type was responsible for the stone industries, this
was probably H. ergaster. This would accord with evidence for the hominids
responsible for the earlier Acheulean in other parts of Africa.
The partial skeleton of Homo ergaster from Nariokotome in Kenya suggests
that this taxon had developed considerably in build and stature from the
presumably ancestral H. habilis; only comparatively minor differences in the
postcranial skeleton are apparent between H. ergaster and modern humans.
Few anthropologists would now disagree that H. sapiens evolved from
H. ergaster/erectus during the long Middle Pleistocene time-span represented
in the archaeological record by the later Acheulean industries. Prior to about
700,000 years ago, no hominid more advanced than H. ergaster/erectus is
attested anywhere in the world. Later Acheulean industries in several parts
of Africa and Europe are associated with a distinct grade of hominid which
some specialists have designated H. heidelbergensis, while others prefer the
less formal ˜archaic H. sapiens™. Particularly informative African examples of
this taxon come from Bodo in Ethiopia, Lake Ndutu in Tanzania, Broken Hill
mine in Zambia, and Hope¬eld in South Africa; specimens from Morocco
should probably also be included.
It is tempting to see this division in hominid associations re¬‚ected in the
artefact typology. Early attempts to recognise numerous successive stages of
increasing re¬nement of handaxe production, as at Olduvai (L. S. B. Leakey
1951) and in the Vaal River gravels (van Riet Lowe 1952), are not supported by
recent evidence; in fact, it is becoming increasingly apparent that only two
basic stages may be recognised: an earlier Acheulean with thick, minimally
trimmed bifaces, and a later Acheulean with more varied, thin, elaborately
¬‚aked examples (Roe in J. D. Clark 2001a). Much of the typological varia-
tion observed between Acheulean assemblages is perhaps better attributed
to differences in raw-material availability than to cultural tradition. These
considerations lead to the proposition that, in addition to the burst of accel-
erated evolution 1.7--1.5 million years ago which resulted in the emergence
and subsequent expansion of Homo ergaster and the Acheulean, there may
have been an analogous if less visible event around 0.8 million years ago
which saw the relatively rapid transition to H. heidelbergensis and the later
The extent of such few undisturbed sites as have been preserved and fully
investigated suggests that population groups may have numbered between
twenty and ¬fty individuals. As in earlier times, these sites, which seem
88 afric an archaeolog y

often to have had more intensive or long-lasting use than had those of their
Oldowan counterparts, were supplemented by sites used for speci¬c activ-
ities such as butchery or tool manufacture. It is noteworthy that remains
of the not-readily-transportable elephant are more frequently encountered
on butchery sites than at base-camps, including those in caves (Shackley
1985). The preservation of evidence for hunting and/or scavenging almost
certainly tends to overemphasise the importance of meat in the Acheulean
diet at the expense of grubs, insects or vegetable foods, traces of which
have rarely survived. (It is necessary once again to emphasise that animal
bones from archaeological sites do not necessarily represent the remains of
food.) When meat was demonstrably eaten, it is often impossible to distin-
guish between kills and carrion although, where many carcasses of a single
species occur together, as at Olorgesailie and at Olduvai, several individuals
may have hunted together in a group. Such a level of organisation implies
that the hominids responsible had developed some capacity for planning
and communication, as is further discussed below (pp. 89--90). Thereafter,
the individuals who were most successful, in the evolutionary sense, were
those with the greater communicative and co-operative skills. It is in these
cultural and cognitive ¬elds, rather than in the basic stone-tool-making tech-
nology, that we can most readily appreciate the progress which took place
during the Acheulean time-span.
Indications of the way handaxes were made, held and used show a level
of dexterity analogous to that of modern people (L. Phillipson 1997). The
hominids™ slim, elongated build was adapted to an upright, fully bipedal
life-style. The comparatively narrow hips suggest that the trend towards dif-
¬cult childbirth and long post-natal maturation had already begun; this
may have resulted in greater male involvement in caring for mother and
child. Taken as a group, fossils attributed to H. ergaster show a much smaller
degree of sexual dimorphism than do those of any earlier hominid taxon;
this has been interpreted as indicating that males™ ongoing competition for
mates was being replaced by long-term pairing. Prolonged infantile depen-
dence would have been accompanied by increased educational and inter-
generational socialising and teaching opportunities, resulting in strength-
ened cultural and behavioural tradition. If these interpretations are correct,
then the arrival of H. ergaster and the Acheulean represents a major advance
in human physical and cultural evolution. It has been argued above that
these events in North Africa took place very much later than they did in
the east and south.
The progressive standardisation and elaboration of tool types and pro-
cesses of manufacture observed in later Acheulean assemblages may also
be interpreted as a further indication that their makers were becoming
increasingly able to talk to one another. This development may be linked
The consolidation of basic human culture 89

neurologically to preferential right-handedness (Isaac 1976; Ludwig and
Harris 1998), which is demonstrable by this time (Toth 1985b; L. Phillipson
1997). Cranial asymmetry, also discernible in Homo heidelbergensis but only
controversially in earlier hominids, is an osteological manifestation of these
developments (Tobias 1991). Much discussion of this topic has been marred
by the tendency to regard language development as a single event rather
than a prolonged multi-stage process (but see Aiello 1996; Jablonski and
Aiello 1998). The extraordinary and, in functional terms, probably unnec-
essarily ¬ne ¬nish of some handaxes suggests the existence of basically
aesthetic standards, in turn implying that cultural values now extended
into non-utilitarian spheres (L. Phillipson pers. comm.; Kohn and Mithen
1999). Later Acheulean people may thus be seen as having taken substantial
steps towards modern human culture (Butzer and Isaac 1975; Gowlett 1986;
Mellars and Gibson 1996).
By later Acheulean times, people eventually spread to most parts of Africa
except those most densely forested and the highest mountain ranges; their
sites were almost invariably located beside or within easy reach of water;
there is now evidence that caves were occasionally occupied. Although the
evidence is inconclusive, controlled use of ¬re may have been achieved by
late Acheulean tool-makers (Gowlett et al. 1981; J. D. Clark and Harris 1985).
As noted above, Acheulean sites seem generally to have been located in
places with easy access to open water; this correlation is probably not due
solely to factors of preservation. Studies of the few South African sites where
faunal remains have been preserved in association with Acheulean artefacts
and which are, by implication, comparatively late suggest that the climate
was generally wetter than that which has prevailed in more recent times
(Klein 1988) or indeed that indicated at the time the tradition initially
appeared. Taken together, these observations raise the possibility of ¬‚uc-
tuating population dispersal, with more restricted settlement during drier
periods. In areas such as the Sahara and southwesternmost Africa, this could
have had a major effect on the extent of Acheulean penetration at certain
Despite this restriction, it is clear that, in comparison with their pre-
decessors, Acheulean hominids possessed a progressively improved capac-
ity for planning as they learned, perhaps from waterside bases, to exploit
their environment over a wider area and in a more comprehensive
fashion (Gowlett 1996). Hominid settlement extended to higher altitudes,
some 2300 metres above sea level, as in the Ethiopian and Lesotho highlands;
this may be an indication that clothing was coming into use. The probable
concomitant invention of receptacles for carrying, for which independent
evidence has been cited above, would have had far-reaching implications
for hominids™ mobility, adaptiveness, range of material culture and, most
90 afric an archaeolog y

important, forward-planning ability so as to keep things for use at some
future time.
Although much change may still remain undetected by archaeologists,
there is enough to suggest that the conservative nature and apparent uni-
formity suggested by initial impressions of Acheulean-type artefacts may
be misleading. None the less, in purely technological terms, the artefacts
provide testimony for the strength of a cultural tradition which developed
about 1.6 million years ago and was maintained for about 1.25 million years,
even if we may question the conclusion that their makers were ˜not only
extremely conservative . . . but they appear to have been relatively rare and
inconspicuous members of the ancient large mammal community™ (Klein
2000b: 107).
Regional diversi¬cation and specialisation
The ˜Middle Stone Age™ and the ˜Late Stone Age™
Developments throughout Africa from rather more than 250,000 until about
10,000 years ago show an accelerated shift away from broad cultural uni-
formities towards the establishment of increasingly distinct regional tra-
ditions. These trends may most readily be traced in the archaeological
record by study of stone-artefact typology; but it is sometimes possible to
go beyond such investigations in an attempt to illuminate the nature of the
ancient societies which, to an increasing extent, may be seen as ancestral
to recent African populations. Despite this growing diversity, developments
seem to have followed roughly parallel courses with inherent continuity
in different parts of Africa and, indeed, in other regions of the Old World
(J. G. D. Clark 1977). The reasons for this are not yet fully understood, and
the various stages were not necessarily reached at the same time in differ-
ent areas. In recent years, the study of the earlier part of this period has
received increased international attention because of the realisation that it
may have been particularly important in the physical and cultural develop-
ment of modern people.
The period with which this chapter is concerned has conventionally been
divided by archaeologists of sub-Saharan Africa into the ˜Middle Stone Age™
and the ˜Late Stone Age™. As research has progressed, it has become appar-
ent that there was no sharp divide between these, any more than there was
between industries of the so-called Middle Stone Age and their predecessors;
the distinctions have become increasingly hard to de¬ne on other than arbi-
trary grounds. It is now recognised that there was a much stronger degree of
continuity between them than was previously believed. It should be empha-
sised that, as with the Acheulean and the Sangoan, the named stone-tool
industries described here are not always clearly de¬ned. Prehistorians have
often attempted to recognise compartments in what were really continuous
ranges of variation through both time and space, as has been emphasised
above (pp. 15--20). A related problem has arisen from the tendency of some
archaeologists, particularly those working in northern parts of Africa, to
present their discoveries in terms of concepts previously adopted beyond
the Mediterranean.
Recent research has also demonstrated that this period of African prehis-
tory, like many others, covered a far longer time-span than was formerly
92 afric an archaeolog y

believed. Much of it lies beyond the range of radiocarbon dating and it is
now recognised that many such age determinations of 40,000 years or more
are in fact minimum rather than absolute dates. Several absolute dating
methods are becoming available for earlier times; most remain experimen-
tal and their preliminary results should be regarded as tentative (Klein 1999;
Schwarcz 2001).
Broadly speaking, the industries formerly designated ˜Middle Stone Age™
display a stone-tool-making technology derived from that illustrated by ¬nal
Acheulean and/or Sangoan artefacts, being often based upon elaborations
(eventually with reduced size) of a prepared-core technique (pp. 57--8 above).
The division between chapters 3 and 4 of this book is thus to some extent
arbitrary. In due course, small radially ¬‚aked plano-convex cores became
the norm, providing more economical use of raw material but losing some
bene¬ts of intentional preparation: these artefacts are here referred to as
radial cores. This is the mode-3 technology of J. G. D. Clark (1969), and
was marked in several regions of Africa by the production of parallel-sided
¬‚akes, or blades, some of which were subsequently trimmed into a variety
of standardised forms. So-called ˜Late Stone Age™ industries generally show
a further reduction in artefact size, and the resultant tiny tools (microliths)
were often ¬tted into handles, several sometimes being used together to
form a composite tool. This innovation, which characterises Clark™s mode-5
technology, arose in sub-Saharan Africa in a mode-3 context; it involved
devising a new way of steeply trimming the edges of ¬‚akes or blades in
order to blunt them, so that they did not split their hafts or cut their users™
¬ngers. This blunting retouch, known as backing, also served to provide a key
for the mastic that was used to hold the stone inserts in place in their hafts.
Studies of edge-wear and mastic remains have provided some indications of
the various ways in which such artefacts may have been used (L. and D. W.
Phillipson 1970; D. W. Phillipson 1976; Wadley and Binneman 1995); and
in rare instances complete hafted specimens have survived, as in certain
South African caves (J. D. Clark 1958; H. J. Deacon 1966) and at Columnata
in Algeria (Cad©nat and Tixier 1960). It is highly probable that some backed
microliths were used as tips and/or barbs for arrows which are ¬rst indicated
in the archaeological record around this time. Ancient Egyptian specimens
(J. D. Clark et al. 1974) are also informative in this connexion (Fig. 43). Use of
the bow and arrow, especially with poison applied to the latter, would have
revolutionised hunting.
In North Africa, including the Sahara and the Nile Valley, archaeologists
have conventionally referred to lithic industries using terminology akin to
those used in Europe or the Levant (see Kleindienst 2001). Mode-3 indus-
tries have here been frequently designated ˜Middle Palaeolithic™, and mode-5
ones ˜Epipalaeolithic™ or ˜Mesolithic™. In some of these areas there also occur
Regional diversification and specialisation 93

Fig. 43: Mounted
1, 2, composite
arrow-point of
microliths set in
3, retouched
¬‚ake set in
mastic, from
Plettenberg Bay,
South Africa
(nos. 1--3 after
Sampson 1974);
4, wooden handle
with mounted
backed bladelets,
from Fayum,
Egypt (after
and Gardner

industries which belong to J. G. D. Clark™s mode 4, based upon the produc-
tion of long parallel-sided blades struck from prismatic cores by means of
punches; the term ˜Upper Palaeolithic™ has sometimes been applied to these
North African industries, as it usually is to their European counterparts.
The apparent idiosyncrasy of these industries in an African context may be
due largely to the ¬ne isotropic materials from which they were made, stone
with similar qualities being comparatively rare in more southerly parts of
the continent.
Throughout the period covered by this chapter, people in Africa remained
hunters and gatherers, obtaining their food from wild sources. Their mate-
rial culture became more elaborate, with regional specialisations increas-
ingly apparent -- tendencies which may be attributed, at least in part, to their
94 afric an archaeolog y

progressive mastery of the natural environment and to their development
of more ef¬cient methods of exploiting the wild sources of food which it
supplied. Different situations offered varying potential, and distinct tech-
nologies were developed to utilise these, with parallel trends in conceptual-
isation and communication abilities. Despite increasing inter-regional diver-
si¬cation, it will be seen that technological development followed broadly
parallel (but not necessarily contemporaneous) courses in different parts of
Africa. Chapters 5 and 6 will discuss how, in regions north of the equator,
the later makers of mode-5 microlithic industries eventually adopted the
food-producing economies which formed the basis for future developments
throughout the continent.
Changing environments at this time are being increasingly studied both
on a long-term, continental scale (Hamilton and Taylor 1991) and in greater
local detail (cf. Gamble and Soffer 1990; Hassan 2002). Useful summaries are
available for North Africa and the Sahara (Vernet 1995), West Africa (Dupont
et al. 2000), eastern and equatorial regions (Elenga et al. 1994; Bonne¬lle et al.
1995; de Busk 1998), and southern Africa (J. Deacon and Lancaster 1988;
Stokes et al. 1998; Tyson 1999; D. S. G. Thomas et al. 2000). The general
picture that emerges is one of numerous minor ¬‚uctuations around a gen-
eral trend of reduced temperature and increased aridity corresponding to
the last glacial period, beginning around 60,000 years ago and reaching its
maximum about 18,000 years ago, with comparatively rapid amelioration
thereafter. Such periods of cool aridity were marked by increased severity
and extent of deserti¬cation in both northern and southern Africa, and
by marked reductions in the areas covered by equatorial forest, with corre-
sponding shifts in the disposition of intervening isohyets and vegetational
zones (Bonne¬lle 1999). It is probable, although detailed evidence remains
incomplete, that broadly comparable changes took place in earlier times
also. Coastal regions saw major changes in sea level, which dropped during
glacial periods, exposing extensive coastal plains in some areas, notably the
southernmost coast of South Africa.
The precise identity of the hominids responsible for the ¬rst post-
Acheulean industries remains obscure. It is clear, in general terms, that
they belonged to archaic representatives of Homo sapiens more advanced
than those designated H. heidelbergensis discussed in chapter 3. It has been
estimated that this differentiation may have occurred within the period
between 350,000 and 250,000 years ago (Br¤uer et al. 1997). There are,
perhaps surprisingly, far fewer well-preserved and adequately documented
human fossils from African contexts of this period than there are from late
Pliocene and early Pleistocene times. Particular interest thus attaches to
specimens (discussed below) from the Middle Awash, Omo-Kibish in south-
ern Ethiopia, Singa in eastern Sudan, and Florisbad in the Free State Province
Regional diversification and specialisation 95

of South Africa, despite uncertainties concerning their archaeological asso-
ciations and, in some cases, their dating. Sites in southernmost Africa are
more comprehensively understood but their hominid fossils are mostly frag-
mentary. For more recent periods the picture is somewhat clearer: mode-5
industries throughout Africa appear to have been the work of people who
were fully modern anatomically. Far more human skeletons have been recov-
ered from sites of the last 20,000 years than are available for earlier periods;
it is sometimes possible to make very tentative attempts at recognising
the ancestors of some of Africa™s recent populations, as is explained in
chapter 5.
This general period has become one of major interest and controversy
because studies of both mitochondrial and nuclear DNA in modern human
populations world-wide seem to indicate that all of them are descended
from a single ancestor or a small ancestral population that may have lived
in Africa about 500,000--100,000 years ago (Cann et al. 1987; Mountain et al.
1993; Stringer and McKie 1996). This genetic research is still at a relatively
early stage; the data are, however, increasing rapidly. Much controversy has
arisen from uncertainty over the speed of change in mitochondrial DNA
and, in particular, whether this takes place at a more-or-less constant rate.
It is, however, indisputable that all modern human populations, despite
their geographical and cultural diversity, are extremely similar genetically --
far more so, indeed, than other primates whose distributions are in every
case far more restricted (Ruvulo et al. 1993). It was originally suggested by
Cann and her colleagues (1987) that the African common ancestor of modern
humans had lived in the general period 280,000--140,000 years ago which,
according to current interpretations, would correspond with immediately
post-Acheulean developments. This estimate was later expanded to 500,000--
50,000 years ago (Stoneking and Cann 1989), while others (e.g. Wolpoff
1989) argued for an age closer to 1.0 million years, which would imply
that the common ancestor was Homo ergaster or H. erectus, not H. sapiens at
More recently, the controversy has focussed particularly on its impli-
cations in the prehistory of Europe for the relationship between Nean-
derthals and the modern people who replaced them there rather less than
50,000 years ago (Mellars 1993). This has led to the assumption that mod-
ern behaviour originated world-wide at about that time (e.g. Klein 1999,
2000a). This proposal, of breathtaking Eurocentricity, effectively ignores the
African evidence which it is one of the aims of this chapter to evaluate.
Some degree of support for the original hypothesis, that modern people
originated in Africa some 280,000--140,000 years ago, is provided by skeletal
material discovered at several African sites, notably but not exclusively in
South Africa, which -- as will be shown below -- suggests that such persons
96 afric an archaeolog y

were indeed living in that part of the continent more than 100,000 years
ago, as was ¬rst pointed out by Beaumont et al. in 1978 (see pp. 98--102 below).
With increasing emphasis on the anatomical and genetic homogeneity
of modern human populations, many authorities have abandoned their
attempts to retain the Linnaean classi¬catory system, and refer simply to
˜anatomically modern people™. This has given rise to the search for ele-
ments of behaviour and culture which may be deemed equally ˜modern™
or, at least, may be correlated exclusively with ˜anatomically modern peo-
ple™ (Klein 1995). If, as most now agree, anatomical modernity ¬rst arose in
Africa (albeit at a date signi¬cantly earlier than that for which many have
recently argued), it is reasonable to seek the beginnings of cultural moder-
nity in that continent also. The regional evidence for these long, gradual
and complex processes will be evaluated below, with an attempted synthesis
at the end of the chapter (pp. 141--6).
With the passage of time, human culture became more complex, and its
archaeological remains are consequently more varied as well as more abun-
dant. Furthermore, as people developed life-styles closer in many ways to our
own, the surviving traces of their activities are easier for us to interpret. The
thought-processes and beliefs that lie behind these activities may, for these
more recent periods, occasionally be illustrated in the archaeological record
through the investigation of such cultural manifestations as graves, settle-
ment layout or rock art. Such clues will be touched upon at appropriate
places in the regional survey which follows.

Southern Africa
It is convenient to begin this survey in southern Africa where there has
been a greater amount of recent high-quality archaeological research on this
general period than in most other parts of the continent. Not only is the
prehistoric sequence there relatively well known (although many striking
gaps in our knowledge remain), but there are indications that the major
cultural developments of this period seem to have taken place there at least
as early as did their counterparts elsewhere in the continent or, indeed, in
other parts of the world. Much of the research that will be described has
been undertaken very recently, and most has been published only in a series
of preliminary reports. Environmental circumstances have been summarised
by J. Deacon and Lancaster (1988).
Most of southern Africa lies outside the area where Sangoan industries
are generally recognised, and the end of the local Acheulean is particularly
ill-de¬ned. There is, regrettably, no site which provides evidence for the
change from the Acheulean to its immediate successors, such as is available
further to the north at Kalambo Falls (p. 120 below). At those rare places,
Regional diversification and specialisation 97

such as the Cave of Hearths near Mokopane (Mason 1962), Wonderwerk
near Kuruman (Beaumont 1990) or Montagu Cave in Western Cape Province
(Keller 1973), where Acheulean deposits are overlain by later material, a
long gap is apparent between the two phases. The date of the ¬rst post-
Acheulean assemblages is also uncertain, although it is now clear that such
material has a far greater antiquity in South Africa than was, until the
1980s, considered likely.
Some of the earliest relevant material comes from springs at Florisbad,
40 kilometres northwest of Bloemfontein in Free State Province, which were
frequented over a long period, resulting in a succession of open-air occu-
pation horizons on some of which distinct activity-areas can be recognised
(Kuman and Clarke 1986; Kuman et al. 1999; Bamford and Henderson 2003). A
human skull discovered in the early 1930s cannot, unfortunately, be located
precisely within this sequence but direct dating of the specimen suggests
that it may be about 250,000 years old (Clarke 1985; Grün et al. 1996).
Although its brow ridges are signi¬cantly less pronounced than those of
the Broken Hill specimen (see p. 72), it retains a number of archaic fea-
tures and is best regarded as intermediate between Homo heidelbergensis and
H. sapiens (Foley and Lahr 1997), perhaps following the trend claimed for the
Hope¬eld/Elandsfontein hominid noted above (pp. 73--4). If this intermediate
form requires attribution to a separate species, the designation H. helmei has
priority (Dreyer 1935; cf. Lahr and Foley 2001). Faunal remains suggest that
the Florisbad sequence may have continued, presumably intermittently, for
more than 100,000 years, but it has not proved possible to demonstrate sig-
ni¬cant change in the associated stone industries, which are of mode-3 type
throughout. Worked wood is also preserved.
At Border Cave in northernmost KwaZulu-Natal, on the frontier with
Swaziland, the post-Acheulean sequence extends back more than 200,000
years (Beaumont 1973; Butzer et al. 1978; Grün and Beaumont 2001). No
comprehensive account of the Border Cave investigations has yet been pub-
lished, but preliminary accounts imply that the oldest industry at this site
was based on the production of parallel-sided ¬‚ake-blades from prepared
cores. It was succeeded by an occupation represented by artefacts, originally
described as ˜Epi-Pietersburg™, which appear to be analogous to those of the
broadly contemporary Howieson™s Poort industry, discussed below (p. 100). A
later mode-3 industry is tentatively dated between 60,000 and 40,000 years
ago. Published accounts indicate that, by about 38,000 years ago, an industry
originally designated ˜Early Late Stone Age™ was being produced at Border
Cave. It is characterised by the frequent use of small bipolar cores and by
the rarity of formally retouched tools; backed microliths are absent. Partic-
ular interest attaches to the fossil human remains (Beaumont et al. 1978;
Rightmire 1979a; Beaumont 1980; Grün and Beaumont 2001), which attest
98 afric an archaeolog y

Fig. 44: The
Klasies River
Mouth sites, seen
from the east
prior to
excavation in

the presence of ˜anatomically modern people™ since pre-Howieson™s Poort
At Klasies River Mouth on the south coast, 40 kilometres west of Cape
St Francis, several dating methods combine to indicate that the earliest
occupation took place at least 120,000 years ago (Singer and Wymer 1982;
H. J. Deacon and Geleijnse 1988; Thackeray 1989; H. J. Deacon 1989, 1995),
during the Last Interglacial period when the sea stood approximately 6--8
metres above its present level. The remarkable sequence of industries at
this site, supplemented, where necessary, with evidence from other sites
on the south Cape coast, may be used as a framework for an account of
the stone-tool industries of southernmost Africa during the past 120,000 or
more years.
Investigation of the cave complex at Klasies River Mouth (Fig. 44) has
revealed archaeological deposits more than 18 metres in total depth, rang-
ing in age from about 120,000 years to 1000 years ago. The earliest industry,
retaining no trace of Acheulean technology, was of mode-3 type; this tradi-
tion continued through the greater part of the sequence, parallel-sided and
pointed ¬‚ake-blades being produced from the local quartzite beach-cobbles
(Thackeray and Kelly 1988). A small proportion of these pointed ¬‚ake-blades
was retouched unifacially and may, along with unretouched specimens, have
been hafted for use as spearpoints. On a number of examples the bulb of
percussion was ¬‚aked away, as if to make hafting easier; other retouched
Regional diversification and specialisation 99

Fig. 45: Artefacts
from Klasies
River Mouth
(after Singer and
Wymer 1982):
1--4, from the
lowest horizon;
5--8, from the
Howieson™s Poort
Artefacts of the
Bambata and
industries (after
Sampson 1974): 9,
10, Bambata-type
industry from
the Cave of
Hearths; 11--13,
industry from
14--18, artefacts
of the Robberg
industry from
(after Mitchell

tools included the edge-retouched pieces conventionally known as ˜scrapers™
(Fig. 45). This stone industry continued, with one notable interruption which
will be described below, for perhaps as much as 80,000 years. Throughout
this period, both terrestrial and marine creatures were exploited for food,
although there is little if any evidence for ¬shing as opposed to shell¬sh-
collecting. Initially, large land animals such as eland and buffalo dominate
the associated faunal remains, suggesting a mixed forest--grassland environ-
ment. There has been controversy over the extent to which these remains
represent the prey of hunters rather than scavengers (Binford 1984; Milo
1998); opinion now favours the former, supported by the discovery of the
100 afric an archaeolog y

tip of a stone point embedded in the vertebra of a giant buffalo. Later,
somewhat colder conditions are attested both by the changed representa-
tion of shell¬sh species and by the increased numbers of bones from the
smaller antelope. Eventually, the sea fell below its present level, so that the
caves overlooked broad coastal plains, now again inundated. Subsequently
there was a prolonged period when the Klasies River Mouth sites were only
rarely visited, lasting until they were re-occupied about 3000 years ago.
One archaeological phase at Klasies River Mouth which marks a pro-
nounced discontinuity in the development of the mode-3 industry has been
the focus of much research, speculation and controversy. A markedly dis-
tinct stone industry was then produced, of a type known at several other
southern African localities -- including Border Cave -- and named after the
site of Howieson™s Poort near Grahamstown (J. Deacon 1995). Its age has
been disputed, but most scholars now place it around 70,000 years ago (H. J.
Deacon 1995, 1998). This corresponds with a period of falling sea level at the
end of the Last Interglacial, when the exposure of extensive coastal plains
may have brought signi¬cantly different environmental conditions and food
resources within the range of the site™s inhabitants. The Howieson™s Poort
industry has large numbers of small blades, mostly 2--4 centimetres long
and often trimmed by steep backing into crescentic or trapezoidal forms
anticipating, by tens of thousands of years, the signi¬cantly smaller mode-5
implements of more recent millennia. Small edge-retouched and notched
pieces are also represented. The local quartzite was mostly unsuitable for
making such small delicate tools, and ¬ner-grained materials were care-
fully sought and collected. Not surprisingly, the makers of the Howieson™s
Poort industry appear to have hunted more of the smaller antelope, such
as the steenbok, than had their heavily armed predecessors. Seafood was
still collected, although the shoreline was probably some distance away.
H. J. Deacon (1989) regarded the Howieson™s Poort as ˜a distinct substage of
the Middle Stone Age . . . con¬ned to areas south of the Zambezi™, although
a similar phenomenon has subsequently been indicated at Mumba in north-
ern Tanzania (p. 127 below) and may be present elsewhere in the intervening
regions. This interlude was clearly an integral element in the economic and
technological diversi¬cation of the period, but much controversy surrounds
its detailed interpretation (e.g. Wurz 1999); many of the proposed inter-
pretations fail to address the clear similarity in technology if not in size
with much later mode-5 industries. The available dating evidence is as yet
insuf¬cient to support the widespread assumption (e.g. H. J. and J. Deacon
1999) that all Howieson™s Poort occurrences are contemporaneous. After the
Howieson™s Poort interlude, Klasies River Mouth was re-occupied, albeit less
intensively, by people whose mode-3 stone industry was remarkably similar
to that of the initial occupation.
Regional diversification and specialisation 101

Fig. 46: The
archaeological Langebaan
sites in southern- Poort
Nelson Bay
most Africa Montagu
which have Blombos 0 100 200 300 km
yielded mode-3 Klasies
Die Kelders River Mouth
and related 0 100 200 miles

The South African Howieson™s Poort is probably the earliest industry yet
known anywhere in the world to be dominated by mode-5 technology.
Parallel technological developments took place in several regions of Africa
in later times when, in most cases, the artefacts were signi¬cantly smaller.
Possible explanations for these phenomena are discussed in a later section
of this chapter (pp. 141--6).
Fragmentary human fossils recovered from Klasies River Mouth in asso-
ciation with the mode-3 industries (Rightmire 1978; Rightmire and Deacon
1991; H. J. Deacon and Schuurman 1992) are attributed to ˜anatomically
modern people™. All the specimens are small fragments, and it has been
suggested that this may be due to the practice of cannibalism (White 1987).
Unfortunately, a few isolated teeth are the only human remains that have
been recovered from the Howieson™s Poort horizon. The excavations at Klasies
River Mouth have provided important information on hunting strategy, with
its emphasis on the young of certain species, notably buffalo (Klein 1978).
The extent to which vegetable foods were exploited at this time remains to
be demonstrated.
Further sites in southernmost Africa (Fig. 46) add to the data provided
by the Klasies River Mouth research (Thackeray 1992; H. J. and J. Deacon
1999; Mitchell 2002). One of the most informative is Die Kelders (Thackeray
2000; Grine 2000; Feathers and Bush 2000). Human footprints at Langebaan,
on the west coast north of Cape Town, are broadly contemporary with the
initial occupation of Klasies River Mouth (Roberts and Berger 1997). At Mos-
sel Bay in Western Cape Province (Keller 1969), the evidence of early exca-
vations demonstrates a sequence parallel to that at Klasies River Mouth
some 200 kilometres to the east. Between them, Nelson Bay Cave at Pletten-
berg Bay (Klein 1974; J. Deacon 1978) preserves a mode-3 industry overlain,
after a gap in occupation, by successive occurrences of the industries desig-
nated Robberg and Albany (dated respectively 19,000--12,000 and 12,000--8000
years ago) which are further described below. More recently, excavations in
102 afric an archaeolog y

Fig. 47: Ochre
with engraved
lines, from
Blombos Cave

a coastal cave at Blombos near Mossel Bay have attracted much attention
(Henshilwood et al. 2001a): a sequence of mode-3 industries extends back
before 100,000 years ago, associated with fragmentary human remains
which are described as anatomically modern (Grine et al. 2000; Grine and
Henshilwood 2002). The signi¬cance of simple bone tools and of worked
ochre fragments from this site (Fig. 47), dated before 70,000 years ago
(Henshilwood and Sealy 1997; Henshilwood et al. 2001b, 2002; d™Errico et al.
2001), has been exaggerated, as will be discussed below (pp. 143--5). Inland,
at Boomplaas Cave in the Cape Folded Mountains, a similar sequence shows
Robberg and Albany industries overlying a long succession of blade-based
occurrences which have not yet been described in detail but which include
a Howieson™s Poort horizon towards the base of the ¬nely strati¬ed deposits
(H. J. Deacon 1979, 1989, 1995).
It seems that mode-3 technology based upon the production of ¬‚ake-
blades, such as is present in the earliest levels at Klasies River Mouth, was
a fairly early post-Acheulean development in several parts of South Africa
(Volman 1984). In the north, such industries, known as Pietersburg after the
town now renamed Polokwane, are best seen in the long sequence at the
Cave of Hearths (Fig. 45). Here, Pietersburg material overlies that attributed
to the ¬nal Acheulean but, as noted above, the two assemblages are sep-
arated by a thick sterile deposit representing a period of unknown dura-
tion when the site was unoccupied (Mason 1962). The Pietersburg industry
is typi¬ed by large numbers of long, parallel-sided ¬‚ake-blades, often with
trimming or use-wear along their edges. Disc-cores and triangular points
are both relatively rare. Similar assemblages come from a number of sites
in the Free State and Northern Cape Provinces, where the artefacts™ mor-
phology was in¬‚uenced by the preferred raw material -- lydianite -- that
was obtained in large quantities at several outcrops in the middle Gariep
Basin and transported over a wide area. It is noteworthy that the latest
Regional diversification and specialisation 103

Fig. 48: Sibudu

Acheulean assemblages from the Cave of Hearths show a tendency towards
the production of elongated ¬‚ake-blades; and the Pietersburg may thus rep-
resent a local development from an Acheulean ancestor between the Vaal
and Limpopo rivers. Industries of Pietersburg type seem to have continued
in use for tens of thousands of years; the greater part of this material lies
beyond the range of radiocarbon dating (Thackeray 1992).
Rose Cottage Cave near Ladybrand in Free State Province provides the best
view of later developments in this part of the South African interior (Wadley
1996, 1997, 2001a; Harper 1997; Ouzman and Wadley 1997; A. M. B. Clark
1997, 1999). Strati¬ed deposits, containing abundant organic materials, are
preserved to a depth of 6 metres and are thought to have accumulated over a
period in excess of 100,000 years. The Rose Cottage sequence may, in general
terms, be seen as a continuation of that at Florisbad. A de¬nitive account is
awaited, but it is clear that Rose Cottage Cave contains a long succession of
mode-3 industries ending about 28,000 years ago, with a Howieson™s Poort
interlude not sharply de¬ned and as yet undated. Stylistic change in the
lithic artefacts seems to have accelerated markedly after about 30,000 years
In KwaZulu-Natal, Sibudu Cave (Fig. 48) on the Tongati River has revealed a
sequence contemporary with the mode-3 succession at Rose Cottage (Wadley
104 afric an archaeolog y

2001b). There was no later occupation characterised by mode-5 artefacts,
and human use of the site seems to have ceased around 35,000 years ago.
Radiocarbon dates extend through the last 10,000 years of occupation, but
the lower levels are as yet undated. Unifacial and bifacial points, some with
concave bases, are characteristic. Although there are occasional backed-blade
forms akin to those of the Howieson™s Poort, no horizon with concentrations
of such artefacts has yet been recognised at Sibudu.
The archaeology of this period in the now-arid Kalahari regions of the
southern African interior remains poorly known (Lane et al. 1998; but see
Helgren and Brooks 1983). Little is known about the archaeology of Namibia
between the ¬nal Acheulean and the appearance of mode-5 industries, other
than in the southwest (Vogelsang 1996, 1998). Surface occurrences suggest
the presence of industries of Pietersburg and Lupemban (pp. 117--20) types
(MacCalman 1963; MacCalman and Viereck 1967). The only long excavated
succession is in the extreme south of the country at the ˜Apollo 11 Cave™,
where a mode-3 sequence underlies an industry of about 12,000 years ago
with many large edge-trimmed ¬‚akes (Wendt 1972, 1976). As will be shown
below, this site is of particular importance because it has evidence for rock
art at a very early date.
In several areas of South Africa it appears that industries based on ¬‚ake-
blades were followed by assemblages with sub-triangular points and other


. 3
( 11)