. 2
( 11)


robust forms subsequently appeared; opinion is again divided as to whether
these are suf¬ciently distinct to merit placement in a separate genus, Paran-
thropus, rather than being subsumed within Australopithecus. Of these robust
hominids, Paranthropus (Australopithecus) boisei appeared in East Africa around
2.5 million years ago, possibly related to or descended from P. (A.) aethiopicus,
while P. (A.) robustus appeared in South Africa about 0.5 million years later,
by which time the earliest representatives of the genus Homo are attested in
both areas.
There can now be little reasonable doubt that, by about 2.5 million years
ago if not before, several distinct types of hominid co-existed in broadly
similar environments. The exact number of parallel hominid lineages and
their relationship to each other are subjects of controversy. Until recently,
the most widely accepted classi¬cation of these early hominids placed them
in two distinct groups (Tobias 1980). This broad dichotomy is still accepted,
although some authorities now feel that one of these categories, the aus-
tralopithecines, may comprise more than one genus, being represented by
the so-called gracile species A. africanus in South Africa and by a somewhat
more robust counterpart further to the north. Although all types of australo-
pithecine displayed the characteristics noted above, there are none the less
signi¬cant differences (Klein 1999). The front teeth (incisors and canines)
of A. africanus were appreciably larger than those of its robust counterparts;
this contrasts to the generally more substantial build of the latter species. (It
should be emphasised that the distinction between gracile and robust in an
australopithecine context refers primarily to the dentition and associated
musculature of the skull: in overall body size the two categories were very
While A. africanus was essentially omnivorous, Paranthropus (Australopithe-
cus) robustus may have evolved a specialised predominantly vegetarian diet,
to which large grinding molar teeth were well suited (Grine 1988; see also
Lee-Thorp et al. 1994). To this feature may be linked also the massive muscu-
lature, especially that of the jaw, which in turn gave rise to the large ridges
of bone to which the muscles were attached. This was probably the sole func-
tion of the sagittal crest noted above, a feature which both P. (A.) robustus and
P. (A.) boisei shared with the modern male gorilla. There was also signi¬cant
postcranial variation, as yet imperfectly understood, between skeletons of
the different australopithecine species.
The emergence of humankind in Africa 29

The Bouri area of the Middle Awash Basin has yielded remains of an
australopithecine, designated A. garhi and dated around 2.5 million years
ago, which, it has been suggested, shows features transitional to those of
Homo (Asfaw et al. 1999; de Heinzelin et al. 1999). Subsequently, further fossils
have been recovered and attributed to H. erectus/ergaster (Asfaw et al. 2002).
This period, again one when arid conditions prevailed (de Menocal 1995),
is of particular signi¬cance because of the archaeological evidence, to be
discussed below, for the earliest use of tools.
The ¬rst representatives of the genus Homo, to a single species of which
all types of modern humans belong, may now be dated as early as 2.5--2.4
million years ago, although most specimens older than 2.0 million years are
very fragmentary. Homo may have evolved from an australopithecine such as
A. afarensis (Tobias 1980; A. C. Walker and Leakey 1993; Klein 1999). The ear-
liest specimens are of the type designated Homo habilis, known from several
sites in eastern Africa and possibly from South Africa also. It was similar in
overall body size to the australopithecines, but had a signi¬cantly larger
brain averaging about 640 cubic centimetres, which is 45 per cent greater
than the equivalent ¬gure for A. africanus (Fig. 9). The teeth were smaller
and more closely resemble those of modern people, as do the bones of
the hand. The posture of H. habilis seems to have been completely upright
and there is no evidence for such massive muscles, with bony ridges for
their attachment, as were characteristic of the contemporary Paranthropus (A.)
boisei and the earlier individuals now classed as P. (A.) aethiopicus. H. habilis
none the less shows substantial variability and some authorities consider
that its more massive representatives (such as the famous ˜1470™ skull from
Koobi Fora, Kenya) should be regarded as a distinct species, for which the
name H. rudolfensis has been proposed (Lieberman et al. 1996). Virtually com-
plete, the 1470 skull (Wood 1991) comes from a context which is securely
dated to about 2.0 million years ago. The rounded skull-vault with a well-
developed forehead housed a brain which, at about 800 cubic centimetres,
was some 70 per cent larger than those of the contemporary P. (A.) boisei.
The sagittal crest and massive muscle attachments of the latter species were
not present in 1470.
By 1.75 million years ago, if not before, a second species of Homo may
be recognised in East Africa. It was at one time given the designation
H. erectus to emphasise its perceived similarity to certain East and Southeast
Asian fossils, but some authorities now consider that the single designation
is inappropriate and prefer to class the East African material as H. ergaster
(Klein 1999: 287--95). As more fossils are discovered, it is becoming apparent
that the Homo population at this time was extremely diverse (Wood 1985);
its later development is described in more detail in chapter 3. One way in
which these early hominid types may have been related to one another is
30 afric an archaeolog y

Fig. 9: Skulls of
1, Australopithecus
africanus from
2, A. boisei from
3, large-brained
Homo, cf. H.
rudolfensis (1470)
from Koobi Fora

shown above in Figure 6. It should be stressed that there is particular contro-
versy concerning the relationship between Australopithecus and the earliest
members of the genus Homo; some authorities (Wood and Collard 1999) deny
that two genera are represented, regarding H. habilis as a gracile australo-
pithecine. The dispute serves to emphasise the dif¬culty, noted above, of
describing evolutionary processes in Linnaean terms.
To conclude this survey of early hominid evolution, it may be instruc-
tive brie¬‚y to compare the physical features of Australopithecus africanus both
with a modern person and with a modern great ape, in this instance a
gorilla (Fig. 10). The ¬rst point that one notices is the small size of the
australopithecine: adults of both sexes stood less than 1.5 metres high and
weighed between 33 and 67 kilogrammes (McHenry 1988). Comparison of
the skulls shows that the jaws and teeth of the australopithecine, despite
the creature™s small overall size, were actually larger than those of a modern
person. The brain, on the other hand, was only about one-third as large, at
The emergence of humankind in Africa 31

Fig. 10:
Skeletons, to the
same scale, of 1,
2, Australopithecus
africanus; 3, Homo

about 450 cubic centimetres, which is approximately the same size as the
modern gorilla™s brain. In the gorilla the brain is placed behind the face,
and the neck-muscle attachment is at the back of the skull. In modern peo-
ple, the brain extends above the face, with the development of a true fore-
head, and the muscle is attached to the base of the skull. In both respects,
australopithecines occupy a position intermediate between the gorilla and
modern humans. In two important ways, Australopithecus africanus was much
closer to a person than to a gorilla: the posture was completely upright,
and the canine teeth were much reduced in size. A gorilla™s molar teeth
serve essentially a crushing function: in people and the australopithecines
they are primarily grinders. Postcranially, the differences in all cases are less
marked: arms became shorter and legs longer, and there are also changes to
the pelvis accompanying more upright posture. The ribcage is progressively
32 afric an archaeolog y

reduced in australopithecines and in modern people, presumably mark-
ing a reduction in gut-size which may have accompanied dietary emphasis
on concentrated foodstuffs including meat (Aiello and Wheeler 1995; Foley

The oldest discoveries in eastern Africa
It must be emphasised that the distribution of fossil discoveries is controlled
not only by the former geographical extent of the relevant species, but also
by the presence or absence of conditions suitable for their preservation,
survival and eventual recovery. These conditions have occurred in both East
and South Africa, but in very different situations (Fig. 11). In East Africa
the lake basins of the Rift Valley provided habitats favourable for the early
hominids and their associated faunas. Rapid sedimentation rates ensured
the preservation of bones and their artefactual associations in comparatively
undisturbed contexts and the volcanic activity of the area provided geologi-
cal materials that can be dated, notably by potassium/argon analysis. Lastly,
more recent developments have often led to the erosion of the deposits,
thus exposing their fossil and artefact contents for collection or excavation
and subsequent study. These factors have combined to stimulate very inten-
sive ¬eld research on Plio-Pleistocene sites in eastern Africa from the 1960s
onwards. In South Africa, on the other hand, bones accumulated -- often
through non-hominid agency -- in dolomite or limestone caves, in deposits
that were consolidated by minerals carried down by water seepage. They
have subsequently been exposed, for the most part, in the course of mining
operations. Research has been intermittent and, until recently, on a small
scale. In the absence of volcanic material suitable for potassium/argon analy-
sis, direct dating of these cave deposits has rarely proved possible, although
study of the faunal remains has enabled them to be set in sequence and
tentatively correlated with the dated East African succession (Howell 1982;
Klein 1999).
Australopithecine remains dated between 5.8 and 3.0 million years ago
(Walter and Aronson 1982) come mainly from three parts of the Rift Valley in
eastern Africa: Hadar and the Middle Awash areas of Ethiopia some 500 kilo-
metres northnortheast of Addis Ababa, the Lake Turkana Basin astride the
Ethiopia/Kenya border (discussed in a separate section below), and Laetoli in
northern Tanzania. The sparse specimens dated before 4.0 million years ago
have been described above. More comprehensive and informative material
comes from Hadar and the Middle Awash where lacustrine and river-delta
deposits are separated by a series of volcanic tuffs. The arid conditions cur-
rently prevailing at these now low-lying sites have been brought about by
subsequent Rift Valley earth movements. At one Hadar locality the remains
The emergence of humankind in Africa 33


R . N ile
Middle Awash

West Turkana Koobi Fora
Lothagam L. Turkana
R. Co

L. Victoria

L. Tanganyika


L. Malawi

R. Zambezi


ge R. al
Oran Va

Rift Valley

0 1000 km

0 500 miles

Fig. 11: The principal sites at which Australopithecus and Homo habilis fossils and/or well-dated
Oldowan artefacts have been discovered

of thirteen early hominids, including four juveniles, were found together.
The most informative discovery, however, is a 40 per cent complete skele-
ton of a female, popularly known as Lucy. Her pelvis and leg bones indi-
cate a well-developed upright posture. Her original height was probably
about 1.2 metres (Johanson and Edey 1981). At Laetoli, aeolian tuffs have
34 afric an archaeolog y

yielded remains which consist mainly of jaws and teeth (M. D. Leakey
and Harris 1987), allowing detailed comparisons to be made. Because of
their physical similarity and near contemporaneity, the Laetoli and Hadar
hominid fossils are generally regarded as representatives of the same species,
Australopithecus afarensis. The Hadar discoveries are additionally important as
providing the best sample of postcranial remains belonging to any eastern
African australopithecine population. The same deposits at Laetoli included
layers of hardened ash-covered mud in which were preserved a remarkable
series of footprints of hominids and other creatures (Fig. 12), supplement-
ing the evidence of the fossil bones that A. afarensis had a fully bipedal
No artefacts have been recovered from the Laetoli deposits, but at Hadar
and in the Middle Awash Basin small numbers of apparently arti¬cially
¬‚aked cobbles (Fig. 13) have been reported in contexts provisionally dated to
some 2.6 million years ago (H. Roche and Tiercelin 1980; J. W. K. Harris 1983;
Kimbel et al. 1996; Semaw 2000). Not far to the south, the Bouri area of the
Middle Awash has yielded stone ¬‚akes and worked bone fragments from the
same locality as remains attributed to A. garhi and dated around 2.5 million
years ago (Asfaw et al. 1999; de Heinzelin et al. 1999). These specimens and
their associations await full investigation; should preliminary accounts be
con¬rmed, these occurrences are by a substantial margin the earliest known
incidence of hominid-made artefacts. They are generally attributed to the
mode-1 Oldowan industry, discussed below, although they are signi¬cantly
older than the occurrence after which that industry is named (Ludwig and
Harris 1998).

The Lake Turkana Basin and Olduvai Gorge
Discoveries of great richness and importance come from parts of the Lake
Turkana Basin in both Ethiopia and Kenya (Coppens et al. 1976; F. H. Brown
1994). The area around the lower Omo River, north of the lake in southern
Ethiopia, was the ¬rst to be investigated (Howell 1976), followed by inten-
sive research in the Koobi Fora area on the northeast shore (M. G. and R. E.
Leakey 1978; Wood 1991; Isaac 1997). The focus of research then shifted to
the exceptionally signi¬cant sites in the western part of the basin (J. M.
Harris et al. 1988). At times between 4 and 2 million years ago the area
drained to the Indian Ocean, but both before and afterwards it has com-
prised a closed basin with no outlet except an over¬‚ow channel to the
Nile which functions only when the Lake Turkana waters reach a very high
level (Butzer 1980; Harvey and Grove 1982). The height and size of the lake
have thus ¬‚uctuated considerably; it and its feeder rivers have laid down
complex series of sediments, up to 1000 metres thick in places, in which
hominid and other fossils are exceptionally well preserved. The principal
The emergence of humankind in Africa 35

Fig. 12: The trail
of hominid
discovered at
Laetoli and dated
to about
3.8 million years
ago. The
footprints are
interpreted as
those of two or
three upright-
hominids, one
smaller than the

sediments with which we are here concerned are designated the Shungura
and Usno Formations in the north, the Koobi Fora Formation in the east and
the Nachukui Formation in the west; they are each separated into various
members by horizons of consolidated volcanic debris known as tuffs, from
which numerous potassium/argon age-determinations have been obtained,
36 afric an archaeolog y

Fig. 13: Mode-1
artefacts: top
row, from Hadar
(after Roche and
Tiercelin 1980);
bottom row, from
the KBS site at
Koobi Fora (after
Isaac et al. 1976);
remainder, from
the Omo Valley
(after Merrick
and Merrick

permitting the establishment of a remarkably detailed chronology (Fig. 14;
see Klein 1999).
The long sequence of fossil-bearing deposits in the Omo Valley began
about 4.0 million years ago (Howell 1976; Howell et al. 1987). The Omo River,
¬‚owing into the northern end of Lake Turkana, has exposed a complex
series of deposits that were laid down in a variety of lakeside and riverine
The emergence of humankind in Africa 37

Fig. 14:
stratigraphy of
the Koobi Fora
Formation and
discoveries; note
that each
member is
named after the
tuff which marks
its base

environments. These deposits have yielded abundant fossil material, mostly
very fragmentary. The hominid remains consist for the most part of isolated
teeth. The oldest specimens are probably almost as ancient as those from
Laetoli, but they are of relatively little value in illustrating the evolution of
the various hominid species.
Although the fragmentary hominid fossils from the Omo are in them-
selves less spectacular and informative than those from elsewhere in the
basin, their associations and the stratigraphy are exceptionally informative;
most belong to the period between 3.0 and 2.0 million years ago. The earlier
hominid fossils are attributed to Australopithecus afarensis, the later ones to
38 afric an archaeolog y

A. aethiopicus and Paranthropus (A.) boisei. Stone artefacts began to be made
about 2.4 or 2.3 million years ago (Howell et al. 1987), by which time Homo cf.
habilis was present. These specimens, with those from the Nachukui Forma-
tion the earliest artefacts known from the Lake Turkana Basin, are simple
¬‚akes struck from small nodules of quartz, which was the only suitable
material available in the area (Fig. 13). They are attributed to the mode-1
Oldowan industry; their characteristics and signi¬cance are discussed below.
The fossil-bearing deposits at Koobi Fora appear to span the period from
rather more than 2.0 million until about 1.2 million years ago. Detailed
studies have demonstrated the local circumstances in which the deposits
were laid down, and thus the immediate environments in which the var-
ious species represented in the fossil assemblages lived and died. At least
two, and almost certainly three, hominid lines are represented at Koobi Fora
(M. G. and R. E. Leakey 1978; Wood 1991); some interpretations yield num-
bers as high as six. There is a robust australopithecine and possibly a more
gracile variety resembling A. africanus. The genus Homo also occurs through-
out the sequence. The early form, which some would attribute to a large-
brained form of H. habilis or to H. rudolfensis, is best represented by the ˜1470™
skull discussed above. There may have been a tendency at Koobi Fora for
remains of Homo to occur predominantly in former lakeside environments,
and those of Paranthropus (A.) boisei in riverine situations (Behrensmeyer
Concentrations of artefacts have been excavated at Koobi Fora, but unfor-
tunately they are rarely in direct association with the hominid fossils.
Although artefacts have occasionally been found in earlier contexts, the
most informative occurrence, attributed to the Oldowan industry, is at the
˜KBS site™, dated to about 1.8 million years ago (Isaac et al. 1976; Isaac and
Harris 1978; Isaac 1997). The KBS site appears to have been originally located
in the sandy bed of a seasonal stream and its area was perhaps restricted
by the availability of shade. Shortly after the site™s apparently brief occu-
pation, volcanic activity covered the area with a thick deposit of ¬ne ash,
thus ensuring its preservation in a virtually unmodi¬ed state. The traces of
hominid activity cover an area 12--15 metres across. It has been calculated
that 400--500 stone artefacts were originally abandoned on the site. The
presence of very tiny chips and splinters shows that the tools were made on
the spot rather than brought to the site from elsewhere, although the lava
from which they were made must have been carried from about 5 kilometres
away. Broken animal bones, including those of porcupine, pig, waterbuck,
gazelle, hippopotamus and crocodile, were also preserved. Only 1 kilometre
to the south, in a similar ash-¬lled stream channel, were found many bones
representing the remains of a single hippopotamus. These were mixed with
The emergence of humankind in Africa 39

over a hundred stone artefacts, mostly ¬‚akes, essentially similar to, but less
varied than, those from the KBS site. Whether or not the hippopotamus
was killed by hominids, there can be little doubt that it was butchered, and
that stone tools were manufactured on the site for this purpose. The signi¬-
cance of these Oldowan artefacts and of the sites from which they have been
recovered is discussed in greater detail below (pp. 49--50). Somewhat later
contexts at Koobi Fora have yielded a distinct series of artefacts for which
the name ˜Karari industry™ has been proposed; this material is discussed in
chapter 3.
The succession provided by the Nachukui Formation is the longest in the
Lake Turkana Basin, extending from about 4.3 until 0.7 million years ago,
the members from between 3.0 and 2.0 million years ago being particularly
informative. Hominid fossils, although not particularly numerous, are some-
times exceptionally complete. The relatively robust Paranthropus (Australop-
ithecus) aethiopicus (pp. 27--8 above) was present as long ago as 2.5 million years
(A. C. Walker et al. 1986), subsequently becoming signi¬cantly more com-
mon and representing about half of the total hominid sample. Specimens
of P. (A.) boisei and one of Homo habilis occur in poorly dated contexts within
the period 2.3--1.6 million years ago; from the beginning of this time-span,
Oldowan artefacts are also attested (H. Roche et al. 1999). By about 1.6 million
years ago, a more advanced hominid is attested in the fossil record by a par-
tial male skeleton from Nariokotome which may con¬dently be attributed
to H. ergaster (A. C. Walker and Leakey 1993). This juvenile individual, fur-
ther discussed in chapter 3, had an even larger cranial capacity than his
predecessors, and presented a striking contrast with P. (A.) boisei, which may
have been his only hominid contemporary. Oldowan artefacts derive from
Nachukui contexts some 2.35 million years old, almost exactly contemporary
with their Omo counterparts (Kibunjia 1994), while Acheulean-type artefacts
(see chapter 3) were in use around 1.6 million years ago (Klein 1999, citing
H. Roche), when H. ergaster was also present in the area.
Further con¬rmation of the co-existence of early Homo and a robust aus-
tralopithecine comes from Olduvai Gorge in northern Tanzania (Fig. 15).
Here, natural erosion has exposed a deep series of superimposed beds which
contain abundant artefact and fossil assemblages covering the greater part
of the last 1.8 million years (L. S. B. Leakey 1965; M. D. Leakey 1971; Hay
1976; Johanson et al. 1987; Blumenschine and Masao 1991). In the lowest
horizons, Bed I and the lower part of Bed II, laid down in lakeside condi-
tions between 1.86 and 1.75 million years ago, remains of both hominids
are found in association with concentrations of Oldowan stone artefacts.
It is considered unlikely that members of two hominid genera would have
occupied the same environmental niche at the same time and been engaged
40 afric an archaeolog y

Fig. 15: Olduvai

in the manufacture and use of apparently identical artefacts; although this
argument is not conclusive, it is probably safe to assume that the Oldowan
artefacts were the work of H. habilis, who was physically, and presumably
also intellectually, the more advanced of the two.
As at Koobi Fora, concentrations of the earliest stone artefacts have been
investigated in situ at Olduvai. Although there is some doubt as to the
extent to which non-human forces may have contributed to their dispo-
sition, several sites in Bed I and the lower part of Bed II provide possible
evidence for the former existence of some simple type of shelter, notably a
setting of stones which may have served as the foundation of some kind of
windbreak and which enclosed the densest part of the artefact scatter on
one particular site (Fig. 16). There is, however, controversy over the extent
to which these sites are truly unaffected by post-depositional disturbance.
Faunal material from these Olduvai sites includes numerous remains of
small creatures and ¬sh, the collecting of which may have been a major
subsistence activity. However, the previously held belief that such small
The emergence of humankind in Africa 41

Fig. 16: Plan of a
stone circle on
an occupation
horizon at site
DK in Bed I at
Olduvai Gorge
(after M. D.
Leakey 1971).
This may
represent the
base of a shelter
constructed of
42 afric an archaeolog y

creatures were more abundantly represented in the oldest sites at Olduvai,
being then gradually supplanted by larger species, is not supported by recent

Central and south-central Africa
Although early hominid fossils have mainly been recovered at sites in eastern
and (as will be shown below) South Africa, artefacts of demonstrably Plio-
Pleistocene age are also known from two places in the intervening regions.
These are at Senga on the Semliki River of easternmost D. R. Congo (J. W.
K. Harris et al. 1987) and in the Chiwondo Beds south of Karonga on the
northwestern shore of Lake Malawi (Kaufulu and Stern 1987; J. D. Clark et al.
1995). An age as great as 2.3 million years has been suggested for the Senga
artefacts, but the evidence is not compelling (Boaz et al. 1992). The Plio-
Pleistocene Chiwondo Beds have yielded abundant faunal material; hominid
remains are scarce although, at Uraha, they have yielded a mandible best
attributed to Homo rudolfensis, the heavily built variety of H. habilis (Bromage
et al. 1995). If the age of this occurrence, apparently in the order of 2.5--2.4
million years, is con¬rmed, the Uraha hominid may be one of the oldest
known representatives of the genus Homo. Far to the west, in Chad, tooth
and jaw fragments resembling those of Australopithecus afarensis have been
published as A. bahrelghazali (Brunet et al. 1996), although doubts have sub-
sequently been expressed about both the identi¬cation as a separate species
and the dating (cf. Klein 1999). These discoveries in Chad, D. R. Congo and
Malawi serve to emphasise the large part played by chance in securing the
preservation of the earliest archaeological remains, and also the extent to
which future research may radically alter our present knowledge. However,
the absence of hominids in very large fossil assemblages both in the extreme
south and in the far north of Africa (Hendey 1981; Raynal et al. 1990) strongly
suggests that, prior to 2.0 million years ago, these creatures may have been
restricted to the equatorial latitudes of the continent.

South Africa
Remains of Australopithecus have been found at sites in three parts of South
Africa; in one of these regions remains of Paranthropus and, less certainly,
early Homo also occur. As noted above, the sites comprise dolomite or lime-
stone caves where fossil bones became incorporated in earthy deposits which
have since hardened to produce the rock-like material known as breccia. The
¬rst discovery was made during quarrying operations at Taung, near the
Harts River 130 kilometres north of Kimberley, in 1924 (Fig. 17). It consisted
The emergence of humankind in Africa 43

Fig. 17: Calci¬ed
breccia and
deposits at
Taung. The
cairn in the
the discovery of
africanus in 1924.
44 afric an archaeolog y

of a magni¬cently preserved complete skull of a juvenile hominid who had
probably been about three years old at the time of death. Publication of
the specimen (Dart 1925) stressed the view that it belonged to a previously
unknown creature intermediate between apes and modern humans. It was
on the basis of this single, immature specimen that Australopithecus africanus
was named. Its discovery made surprisingly little impact upon archaeolog-
ical thinking at that time, because Southeast Asia and Europe were then
believed to have been the main areas where the early stages of human
evolution had taken place. A further problem was, and remains, that of
establishing the absolute age of the Taung hominid which seems, like the
other animal bones with which it was associated, to have been brought to
the site by a leopard, eagle or other predator (Berger and Clarke 1995). The
site has yielded no artefacts. Estimates of its age have varied between 2.8
and 1.0 million years, with the consensus now favouring the earlier part of
this range. Doubt has even been expressed whether this juvenile specimen
is best attributed to A. africanus at all (Peabody 1954; Butzer 1974; Tobias
1978a; McKee 1993).
From 1936 onwards further discoveries began to be made in South Africa,
¬rst at the sites of Sterkfontein, Swartkrans and Kromdraai in the Blaaubank
Valley near Krugersdorp, and then at Makapansgat some 300 kilometres
to the northeast near Mokopane (formerly Potgietersrus) (Brain 1993; Klein
1999 and references; H. J. and J. Deacon 1999; Kuman and Clarke 2000).
More recently, further sites have been discovered in the Sterkfontein vicin-
ity (Mitchell 2002 and references). Between them these sites have yielded
the remains of several hundred australopithecines. Although over the years
these specimens have been attributed to a bewildering variety of species and
genera, it is now widely believed that most of them belong to two species:
A. africanus and Paranthropus (A.) robustus. What may be the oldest specimen,
a remarkably complete skeleton from Member 2 at Sterkfontein for which
an age around 3.3 million years has been proposed (Partridge et al. 1999),
shows features which suggest that it may be morphologically as close to
A. afarensis as to later A. africanus (Clarke 1998, 1999).
There has been considerable controversy over the precise nature of these
sites and how the bone concentrations amongst which the hominid remains
occurred came to be accumulated. It was at one time thought that many of
the animal bones at Makapansgat had been selected by hominids and taken
to the site for use as tools (Dart 1957), an hypothesis based upon the uneven
representation of different body parts and upon the seemingly standardised
fractures on many of the bones. This view, which involved acceptance of
a tool-making status for Australopithecus, is not generally held today; the
breakage patterns are paralleled at other sites where there is no possibility
of hominid agency (Brain 1967). The differential representation of body parts
The emergence of humankind in Africa 45

Fig. 18: Chrono-
logical chart
showing the
probable ages of
the South
African australo-
pithecine sites

is also like that which occurs at leopard lairs where carcasses are often
deposited in trees, out of reach of scavengers, for consumption at leisure. It is
now believed that, over prolonged periods, hominid carcasses were brought
to such trees by leopards and that their bones thus gradually accumulated
in the caves and rock crevices below (Brain 1981). The Makapansgat hominid-
bearing deposits are best dated around 3.0 million years ago, when the area
enjoyed a high-rainfall sub-tropical environment (Rayner et al. 1993). All the
hominid fossils from this site are now attributed to Australopithecus africanus.
Objects generally accepted as artefacts do not occur at Makapansgat until
signi¬cantly later.
Despite various attempts, no wholly satisfactory method has yet been
devised for directly dating the South African australopithecine sites, which
lack volcanic deposits such as have yielded potassium/argon dates for the
early hominid sites in East Africa. On the basis of comparisons of the faunal
assemblages with those of the relatively well-dated East African sequences,
supplemented by measurements of palaeomagnetism and the provisional
results of other newly developed techniques, it has however proved possi-
ble to show that the South African sites probably fall within the period
between 3.4 and 1.4 million years ago (Fig. 18). The earliest occurrences are
those at Makapansgat and the earliest part of the Sterkfontein sequence.
Signi¬cantly, in both these deposits Australopithecus africanus is the only
hominid represented, and there are no stone tools. After what may have
46 afric an archaeolog y

Fig. 19: Horn and
bone tools from
1.0--1.5 million
years old,
believed to have
been used for
digging up edible
roots and tubers,
or for opening
termite mounds.
The larger (horn)
tool is 14 cm

been a substantial gap in the local sequence, there are further occurrences
at Sterkfontein and in the Older Breccia at Swartkrans. Here, Paranthropus
(A.) robustus is the only australopithecine species present. There are also frag-
ments of burnt bone which may represent an early use of ¬re, whether or
not this was fully controlled. Stone tools attributed to a mode-1 Oldowan
industry are present only in the Sterkfontein area, in contexts for which
faunal associations suggest a maximum age of around 2.0 million years
(Kuman 1994, 1998). Heavily utilised bone fragments (Fig. 19) also occur but
it is not yet clear whether their use was contemporary with the Oldowan;
their polish suggests possible use to obtain underground plant foods, grubs
or termites (Backwell and d™Errico 2000). Remains of a creature attributed
to the genus Homo, possibly H. habilis, have also been recorded in somewhat
later horizons at Sterkfontein and at Swartkrans, associated with stone arte-
facts (Clarke 1988; J. D. Clark in Brain 1993; Kuman and Clarke 2000) which,
as noted below in chapter 3, are accepted by most archaeologists as akin
to those from upper Bed II at Olduvai. The ¬nds from Kromdraai also prob-
ably belong to this same general period -- perhaps about 2.0 million years
ago or shortly thereafter, by which time A. africanus was probably no longer
present in South Africa (Kuman et al. 1997); it should be noted that the Krom-
draai artefacts were not found in direct association with fossil hominids,
but came from an adjacent cave, the link being demonstrated by faunal
association. As will be further discussed in chapter 3, fossil fragments from
The emergence of humankind in Africa 47

Swartkrans, Sterkfontein and the recently discovered nearby site of Drimolen
have been attributed to H. ergaster, with ages subsequent to 1.8 million years

The earliest tool-makers
As has been shown, incontrovertible archaeological evidence for the earliest
recognisable stages of human material culture comes almost exclusively
from Ethiopia and from Kenya, and dates from between 2.5 and 2.0 million
years ago. This particular time-span is not well represented at South African
sites where, although earlier contexts have yielded hominid fossils, there is
as yet no conclusive evidence for tool-making older than 2.0 million years.
In both regions, the earliest recognised artefacts are of stone and belong
to J. G. D. Clark™s ˜mode 1™ of stone-tool technology (p. 18, above). They are
attributed to the Oldowan industry named after Olduvai Gorge where Bed
I and the lower part of Bed II have yielded what is now seen to be a rather
late occurrence of this industry. Despite the very long period of time over
which these artefacts were produced, little development has been discerned
(Ludwig and Harris 1998); occurrences from the Ethiopian Rift Valley, the
Lake Turkana Basin and Sterkfontein are all suf¬ciently similar to justify
their inclusion in the Oldowan taxon.
The most important characteristic of the Oldowan, in contrast with all
later stone industries, is the absence of standardisation (Toth 1985a; Toth
and Schick 1986; Kuman 1996; Isaac 1997). It appears that the basic process
employed was banging together two stones or pounding one with another
so as to detach sharp ¬‚akes from which could be selected artefacts suited
for a particular need; on occasion a ¬‚ake might receive simple retouch, or
the core itself might be employed. Most of the artefacts are ¬‚akes, up to 6--7
centimetres long, very few of which show any signs of deliberate shaping or
trimming. The artefacts from the lower part of the Olduvai sequence clearly
bear a close technological resemblance to those from the KBS and contem-
porary sites at Koobi Fora. There are, however, some signi¬cant differences.
Whereas unretouched ¬‚akes predominate at Koobi Fora, at Olduvai there is
a higher proportion of cores or core-tools, including cobbles from which a
few ¬‚akes have been removed as if to produce a cutting or chopping edge.
Flake tools, some of remarkably small size, are also a feature of assemblages
from Olduvai (Fig. 20). The variations in artefact size and morphology need
not necessarily re¬‚ect cultural traditions, but may be explained at least in
part by reference to the nodules and ¬‚aking qualities of the different raw
materials from which the artefacts were made (Ludwig and Harris 1998;
Stiles 1998).
48 afric an archaeolog y

Fig. 20: Oldowan
artefacts from
site DK at
Olduvai Gorge
(after M. D.
Leakey 1971)

Isaac (1997: 296) has described the Oldowan as involving ˜little more than
least-effort solutions for needs for sharp edges™ and noted that ˜this classi¬-
cation need not imply participation in a common culture tradition system.
However, it is entirely credible that across East Africa there existed a loose
network of stone-tool-making social groups of early hominids, which over
time were subject to the diffusion of cultural information.™ Since there is
no widely accepted evidence that particular artefact types were reserved for
special purposes, it is only through experiment and the study of traces of
wear on the ¬‚akes that we can discover the uses to which they were put.
Such studies on Oldowan stone ¬‚akes from Koobi Fora (Keeley and Toth 1981;
Toth in Isaac 1997) have yielded evidence for the probable cutting of wood,
meat and bone. The last of these uses is also demonstrated by cut-marks on
bones recovered at several sites (Bunn and Kroll 1986). Among many other
uses which may be surmised are the grubbing up of roots and tubers, open-
ing nuts and seed pods, and extracting grubs from the bark of trees. It is
also perfectly possible that a major function of the Oldowan artefacts may
have been the production of wooden tools which have not survived in the
archaeological record but which could also have served some of the above
The emergence of humankind in Africa 49

Several early hominid sites have yielded traces of burning (J. D. Clark and
Harris 1985), but there is much uncertainty whether this was purposefully
controlled. Fire does, of course, occur naturally, and may have contributed
signi¬cantly to changes in vegetation at various times in the past, as well
as to hominid dietary practices, security and ability to survive in hostile
environments. It may be supposed that human control over ¬re came about
in at least two stages: transfer and maintenance of naturally occurring ¬re,
followed by ability to make it. So far, archaeologists have made little agreed
progress with resolving the chronology of these developments.
It is clearly important to ascertain, if possible, which hominid or hominids
were the makers of the Oldowan artefacts. The picture is made more complex
by the presence in both eastern and southern Africa at this time of hominids
which most authorities believe to represent more than one genus: Australo-
pithecus and Homo. It should be noted that the earliest lithic artefacts occur
only in areas where the presence of early Homo is also indicated; Australo-
pithecus, on the other hand, is frequently found without such associations.
The earliest artefacts at Hadar, Middle Awash and in the Omo Valley are all
contemporary with the ¬rst local attestation of Homo. The same is true in
South Africa. It is noteworthy that, on osteological grounds, the hand of
Homo habilis is believed to have been better adapted to precise manipulation
of tools than that of at least some australopithecines (Susman 1998). It may
thus be argued that Australopithecus is unlikely to have been the maker of
Oldowan stone artefacts.
It must be stressed that those recognisable artefacts attributed to the
Oldowan probably represent a late stage of what is likely to have been a
prolonged process of development. We have noted the absence of standard-
ised tool-types and suggested that sharp ¬‚akes were selected for immedi-
ate purposes. Such selection may previously have been applied to stones
which had not been purposefully fractured and which it would be virtually
impossible for the archaeologist to recognise as artefacts. Others, long-since
perished, may have been made of wood. The development from tool-use to
tool-making may have been long and hesitant. We know that tool-use today
is not a prerogative of the hominids, so this development may be one which
stretched far back into Pliocene or even earlier times where no archaeolog-
ical record has yet been recovered. The initial stages may have involved the
use of random hammerstones, then the use of chance-struck ¬‚akes, both
processes preceding the intentional bashing of stones in order to produce
sharp edges. The earliest artefacts will only be properly understood when
we can suggest the uses to which they were put, and research to this end is
only just beginning.
Although our knowledge is still very incomplete, it is possible from the
evidence so far available to draw some tentative conclusions about the
50 afric an archaeolog y

life of early Homo at sites where Oldowan artefacts have been recovered.
The KBS site at Koobi Fora has been noted above; an interesting feature
of the faunal remains is that animals are represented from several differ-
ent environments: hippopotamus from the lake and gazelle from the drier
inland plains, for example. The different source-areas of the species repre-
sented suggest that the early hominids may have used the site as a home-base
and brought back to it carcasses or joints of meat that they obtained else-
where from more than one source, whether as prey or as carrion. This obser-
vation suggested to Isaac (1978) that one of the most basic features of human
behaviour, the transport of food in order to share it, had already been devel-
oped. This interpretation has been the subject of much controversy; some
prehistorians consider that the site may represent a succession of brief activ-
ities rather than a single episode, that the bone accumulations owe little
to hominid agency, and that their association with stone tools merely indi-
cates that several species, including hominids, frequented the same place
for some purpose such as getting water (Binford 1981; Potts 1984, 1986; but
see also Blumenschine and Masao 1991; Rose and Marshall 1996). A second
type of site, known both at Koobi Fora and at Olduvai, shows less controver-
sially that the early hominids on occasion made use of temporary butchery
sites for the dismemberment or consumption of single carcasses too large
to be transported entire, such as elephant and hippopotamus.
Some prehistorians have tended to assume that the bones recovered on
Oldowan (or later) sites are necessarily those of animals that were hunted
for food. It must be stressed that this may be an unjusti¬ed assumption. Sev-
eral studies of the various possible mechanics of bone accumulation have
shown that hominid activities need not always have been responsible. Fur-
thermore, animal food was not necessarily obtained by hunting. There is
an increasing body of evidence that the early hominids frequently obtained
their meat by scavenging (Shipman and Rose 1983; Shipman 1986; Bunn and
Kroll 1986). Animals killed by other carnivores seem to have been obtained
by the early hominids and either butchered on the spot or removed for con-
sumption elsewhere. The range of species eaten is thus more indicative of
wild carnivores™ predation abilities than of hominid hunting skills. Other
potential sources of animal protein include capturing small animals, rep-
tiles or insects. It is also important not to underestimate the signi¬cance of
vegetable foods in early hominid diets, as often indicated by fossil dentition:
unfortunately traces of foods in these categories are hardly ever preserved
in the very early archaeological record. However, as Foley (2002) has pointed
out, there is biological evidence that Plio-Pleistocene hominids may have
become progressively more dependent than their predecessors on concen-
trated foodstuffs such as meat.
The emergence of humankind in Africa 51

It has for long been believed that, for the period considered in this chapter,
hominids were restricted to the African continent. The date at which they
¬rst expanded into Eurasia remains a matter of controversy (Turner 1999).
Some stone-artefact assemblages from southern Spain have been compared
with those of the African Oldowan (Gibert et al. 1998) and an age in excess
of 1.0 million years has been proposed. The argument is not wholly convinc-
ing, however, and the presence of similar assemblages apparently preceding
the Acheulean in more recent contexts on the Atlantic coast of Morocco
(Biberson 1961; Raynal et al. 1995) provides no support for attributing such
an early date to the Spanish material. Intensive research in Israel (Bar Yosef
1987; Goren-Inbar and Saragusti 1996; Gu©rin et al. 1996) has provided a
clearer indication that the earliest hominid presence in this key area link-
ing Africa with Eurasia belonged to Acheulean times, perhaps slightly before
1.0 million years ago. This conclusion is not so far convincingly contradicted
by ongoing research at Dmanisi in Georgia (Gabunia et al. 2000) or by recent
age determinations in Indonesia (Swisher et al. 1994). Lowered sea levels in
the Late Pliocene may have left a land-bridge between the Horn of Africa and
southern Arabia at Bab el Mandeb (Milliken 2002). The age and nature of
the earliest hominid occupation of Eurasia remains very poorly understood,
but there is little reason to suppose that it is of any relevance to the early
periods of hominid development discussed in this chapter.
Our knowledge of the early hominids is, and will presumably always
remain, very incomplete. We can learn something of these creatures™ appear-
ance and physical abilities from their fossil remains. We can learn about
some of the places they frequented and some of the foods that they ate.
Their artefacts, when made of imperishable materials, can tell us something
about their technological abilities. Taken together, these factors allow us to
reconstruct a very incomplete view of their life. The list of what we do not
know is far longer. What was the social basis for the groups of hominids
that left remains such as those at the KBS site? Was there any socio-political
unit larger than such a group? Were such associations permanent? How did
they exploit the seasonally shifting and changing resources of their African
homeland? All these questions, like the all-important ones of intellect and
communication, are ones that research is only now beginning to address.
A consensus is none the less emerging that the levels of group size and
co-operative behaviour indicated at this period need not imply abilities in
vocal communication very much more advanced than those exhibited by
modern great apes.
The consolidation of basic human culture
Acheulean and Sangoan in Africa
This chapter deals primarily with one of the most remarkable and least
understood phenomena of world prehistory: the enormously wide distri-
bution, both in time and space, of people who made stone artefacts of
the type conventionally known to archaeologists as ˜Acheulean™. (Problems
of de¬nition are considered below.) These artefacts ¬rst appeared in the
archaeological record, in eastern Africa, about 1.6 million years ago; and
they seem to have survived in most areas until the period between 350,000
and 250,000 years ago: a time-span of at least 1.25 million years. In addi-
tion to Africa, Acheulean-type artefacts are found in Europe and Asia, from
Spain and Britain in the west and north, to India and perhaps China in
the east. In some parts of the Old World they are the oldest known mani-
festation of human settlement (but see Carbonell et al. 1999). The name is
taken from Saint Acheul in the Somme Valley of northern France, one of the
places where these characteristic stone artefacts were ¬rst recognised during
the mid-nineteenth century. Also discussed here, because they are seen as a
¬nal manifestation of the Acheulean tradition, are the African stone-artefact
assemblages designated Sangoan after Sango Bay on the western shore of
Lake Victoria (cf. G. H. Cole 1967).
Although the Acheulean has often been regarded as an entity, there are
good reasons to believe that this long period was one during which impor-
tant human behavioural developments occurred in the conceptual, social
and organisational ¬elds, as well as in the more readily discernible technol-
ogy ( J. D. Clark 1996). There are some indications, as will be shown below, for
group activities that may have required linguistic communication for their
organisation. The evidence for these processes is essentially circumstantial
and dif¬cult to interpret (Gibson and Ingold 1993; Mellars and Gibson 1996;
Noble and Davidson 1996; McPherron 2000). Signi¬cant physical and bio-
logical evolution of hominids also took place at this time (Klein 1999), and
the geographical distribution of tool-makers was greatly expanded. This,
together with the larger numbers of known later sites, must indicate a sub-
stantial increase in human population. Adaptation to varied environmental
conditions is also attested, with the ability to obtain food and safety in cir-
cumstances markedly different from those of the restricted areas known to
have been occupied by the earliest hominids. As environments ¬‚uctuated,
The consolidation of basic human culture 53

so did the areas subject to hominid settlement. Stone-knapping techniques
were adapted with notable skill to utilise different raw materials, which now
had rather less effect on the form of the ¬nished product than had been the
case in earlier times. The picture remains incomplete and open to differing
interpretations, but there can be little doubt that, by the time of the lat-
est Acheulean, humankind had developed many abilities and characteristics
that earlier hominids had lacked.
The hominids responsible for the Acheulean and Sangoan were all mem-
bers of the genus Homo. For much of the relevant period, until about 700,000
years ago, robust australopithecines -- Paranthropus (Australopithecus) boisei and
P. (A.) robustus -- were also sometimes present in the same general areas, but
it seems unlikely that they shared responsibility for the Acheulean with the
hominid here termed H. ergaster. (As noted in chapter 2, some scholars pre-
fer to retain the designation ˜African H. erectus™ in place of ˜H. ergaster™.) By
late Acheulean times, it appears that Homo had evolved considerably in the
general direction of modern people; while some authorities informally des-
ignate fossils of this type as archaic H. sapiens, others (e.g. Rightmire 1996,
1998) prefer to attribute them to H. heidelbergensis. Although no hominid
fossils are known which are incontrovertibly associated with the Sangoan,
it may be reasonable to assume that those responsible were of archaic
H. sapiens type. Further discussion of these hominids is placed at the end of
this chapter.
The Acheulean has conventionally been de¬ned by the presence, with
variable frequency, of the stone implements known as handaxes. These are
characteristically pear-shaped, pointed or ovate in outline and biconvex in
cross-section, usually 12 to 20 centimetres in length and often ¬‚aked over at
least part of both surfaces (Fig. 21). A fairly sharp edge may be restricted to
the area of the more pointed end or may extend around all or the greater
part of the implement™s periphery. The earliest handaxes tend to be crudely
shaped, fewer than a dozen ¬‚akes having been removed from them: the scars
which mark the position of these ¬‚akes are deep, suggesting the use of a
simple stone hammer. These handaxes are usually fairly thick in relation to
their breadth. In later assemblages, although crude examples may continue,
there are also much more ¬nely ¬‚aked specimens with shallow scars such as
may have been produced by use of a softer (wood or bone) hammer or by indi-
rect percussion; these implements are generally thinner and more symmet-
rical in cross-section. In many Acheulean assemblages, particularly in Africa,
there are also found implements known as cleavers (Ranov 2001). Similar in
size and manner of production to the handaxes, these were often made on
large ¬‚akes and have a straight or transverse cutting edge in place of the
point. Acheulean stone-knapping corresponds to J. G. D. Clark™s ˜mode 2™
lithic technology (cf. p. 18, above). Despite uncertainties as to the uses to
54 afric an archaeolog y

Fig. 21:
artefacts from
Montagu Cave,
South Africa
(after Keller
1973): 1,
2, cleaver; 3--4,
retouched ¬‚akes

which these artefacts were originally put, the terms ˜handaxe™ and ˜cleaver™
are so widely used and understood that they have been retained in this
So ubiquitous and numerous are these handaxes and cleavers in some
Acheulean assemblages that it seems they were used for several different
purposes. Despite the implication of the conventional designations, these
implements were probably used for purposes different from those that these
names suggest. Patterns of damage of the type caused by heavy digging are
rarely discernible, but generally we have little de¬nite knowledge of the
uses to which these artefacts were put (cf. L. Phillipson 1997), although
microscopic studies of edge-wear (e.g. Keeley 1980; Beyries and Roche 1982;
Binneman and Beaumont 1992) and plant residues (Dominguez-Rodrigo et al.
2001) are beginning to produce signi¬cant results. Experiments have shown
that the cleaver in particular is remarkably effective as a butchering instru-
ment (P. R. Jones 1980) and for skinning large game. Handaxes were probably
particularly versatile, and may have been used for butchery, as weapons, and
for numerous other purposes (Kohn and Mithen 1999; Gamble and Marshall
The consolidation of basic human culture 55

2001). In some instances tools have been found in association with other evi-
dence for a particular activity, such as butchering a carcass, and cut-marks
on the bones may provide additional evidence for the use of these artefacts.
Although handaxes and cleavers are generally regarded as the most char-
acteristic features of Acheulean assemblages and, indeed, are often seen as
their de¬ning feature, they are not the only types of implement which occur.
Numerous ¬‚akes -- both those removed in the course of handaxe-production
and others -- were used untrimmed as cutting and scraping tools, or were
retouched into a variety of relatively unstandardised forms. Some artefacts
are indistinguishable from individual specimens seen in mode-1 industries.
Cores, which sometimes include types resembling Oldowan ones, may have
been discarded as waste or used as choppers.
Stone-artefact assemblages that have been designated Acheulean occur in
most parts of Africa (Fig. 22). That they have not been recorded from the
densely forested regions of West Africa and the Congo Basin probably re¬‚ects
a genuine discontinuity in their distribution rather than lack of research.
It thus seems probable that it was the makers of these artefacts who were
responsible for the ¬rst human settlement of much of the continent away
from the eastern savanna, to which the earliest such phenomena had appar-
ently been restricted ( J. D. Clark 1967, 2001b). Just when this major expan-
sion took place cannot yet be ascertained, but such evidence as is available
suggests a date of about a million years ago. This was probably also the time
when some tool-makers moved from Africa into adjacent parts of Southwest
Asia; their extension into Europe may have been signi¬cantly later (Milliken
Throughout Africa and, indeed, in many other regions of the Old World,
Acheulean-type stone tools are remarkably standardised. Detailed investiga-
tion is hindered by the scarcity of demonstrably primary-context, single-
event assemblages. Comparative studies suggest, however, that much of the
variation that does occur is due to the ¬‚aking qualities and nodule sizes of
the available raw materials (Gamble and Marshall 2001). At ¬rst sight, the
artefacts also show relatively little consistent change through time although,
as noted above, the Acheulean remained the dominant tool-making tradi-
tion for some 1.3 million years. Despite this apparent standardisation, it
would not be prudent to propose that there was a single Acheulean indus-
try, using the term ˜industry™ in the same sense as has been applied for more
recent periods. The artefact assemblages so designated are chronologically
and geographically very widely distributed; and they sometimes vary consid-
erably in composition (as will be shown below), with their main diagnostic
components (the handaxes and cleavers) differing so much in frequency,
and presumably in function, that they are best regarded merely as belong-
ing to the same mode-2 technological taxon. In other words, the Acheulean
56 afric an archaeolog y

Sidi Zin
Casablanca Hanech


Erg Tihodaine

Wadi Halfa

L. Chad
Jos Middle Awash sites
R. Melka Kunture

Beli Konso

Nariokotome Koobi Fora

Kilombe, Kinangop
L. Ndutu
Olduvai Peninj
Kalambo Falls

Broken Hill Mine
Victoria Falls

Cave of Hearths
Sterkfontein sites
Klipplaatdrif Wonderboompoort
Wonderwerk aal Cornelia,Three Rivers
R Rooidam, Harts River

Hopefield Cave
Cape Hangklip
Fig. 22: Principal 0 2000 km
Acheulean and
0 1000 miles
related sites in

should not be regarded as an industrial or cultural entity on grounds other
than the very general technological and/or morphological similarity of its
stone artefacts. It would be premature, in the present state of knowledge,
to assume any stronger af¬nity among the life-styles of its makers.
In most areas there is a marked variation in the proportions in which
handaxes and other tool types are represented even in the few assemblages
which can be dated and which have been preserved for study in their
The consolidation of basic human culture 57

Fig. 23: Core
preparation. The
illustrations in
the top row are
of idealised
examples (after
Schick and Toth
1995); the other
specimens are
from Nubia (after
Wendorf 1968).

entirety. There are also contemporaneous assemblages from which handaxes
and cleavers are largely or completely absent, and there has been consider-
able controversy as to whether or not these should be regarded as belonging
to the Acheulean. The causes for this variation are not yet properly under-
stood. Most likely a number of cultural and economic factors were all partly
responsible, such as style, personal or group preference, raw material avail-
ability, and the range of activities that were undertaken at the different sites.
These activities may in some cases have been linked with season, environ-
ment or subdivision within the social group. It is therefore not surprising
that archaeologists have not been able to recognise signi¬cant patterning
in this variation, either between different geographical areas or through
In some areas, the later phases also saw the adoption of techniques
whereby core-preparation could be used to pre-determine the form of a ¬‚ake;
this may be regarded as a development linked to increased expertise in
biface production. While not ubiquitous, locally distinct techniques of this
type have been recognised in southern, eastern and northern Africa, as well
as in Europe (Dibble and Bar Yosef 1995; Schlanger 1996; see also Fig. 23).
The designation ˜Levallois™, established in the latter area, has also sometimes
been applied to some local African equivalents. Such loose usage is not
58 afric an archaeolog y

followed in this book, where the general term ˜prepared-core™ is preferred,
and particular local manifestations are discussed individually. These varied
techniques had a common purpose: to ensure that standardised ¬‚ake tools
could be produced without extensive marginal trimming. Although closely
linked with handaxe-production, they became particularly common in later
Acheulean industries of certain areas and survived, often in modi¬ed form
or highly developed, into subsequent periods, as discussed in chapter 4.
Artefacts of Acheulean type are extremely common in certain areas of
Africa, but sites where they have been found undisturbed in their original
contexts are relatively few (e.g. Sampson 2001). As with the earlier periods,
a high proportion of our evidence of this time comes from eastern Africa.
Proximity to water appears to have been an important factor in determin-
ing areas suitable for settlement; and some of Africa™s great river valleys are
thus particularly rich in Acheulean-type remains. Noteworthy in this respect
are the Vaal (Sohnge et al. 1937; H. J. and J. Deacon 1999: 81--3), the Zambezi
( J. D. Clark 1950, 1990), and the Nile valleys (Sandford and Arkell 1933;
Sandford 1934). As these rivers have developed and their rates of ¬‚ow ¬‚uctu-
ated, most of the early sites on their banks have been disturbed or destroyed,
and the artefacts which they contained have been moved, sorted and selec-
tively incorporated in ¬‚uviatile deposits (Fig. 24). Such movement may not
only alter the con¬guration of larger artefacts, it may destroy or remove
smaller ones altogether, leaving a greatly distorted picture of the original
assemblage. Although in certain circumstances the artefact contents of suc-
cessive river deposits may give us some information about the local archae-
ological sequence, this disturbed material is now recognised as having very
limited value; research emphasis has therefore shifted away from the great
river valleys to the location and excavation of undisturbed sites.
In much of central Africa (and indeed over a much wider area extending
from Botswana to Ethiopia and Sudan), the stone-tool assemblages immedi-
ately following those designated Acheulean are of the type known as San-
goan (see p. 52 above). In many parts of the Congo Basin, Sangoan-type
assemblages are the earliest trace of human settlement yet found (Mortel-
mans 1962; Van Noten 1982). They comprise massive pointed core-tools, but
also a signi¬cant proportion of lighter-duty retouched ¬‚akes (Fig. 41, below).
Unfortunately, as with their earlier counterparts, virtually no research has
yet been undertaken to try and establish the uses to which these Sangoan
artefacts were put. Much of the regional variation that has been recognised
may be attributed to the varying quality of the raw materials that were avail-
able for stone-knapping. Many Sangoan sites appear to belong to a time when
semi-arid conditions prevailed and when the equatorial forest was signi¬-
cantly less widespread and dense than it had been previously (McBrearty
1987). The suggestion previously made (e.g. J. D. Clark 1970) that the
The consolidation of basic human culture 59

Fig. 24: Air view
of the Victoria
Falls on the
Zambezi River.
The Falls are
2 kilometres
wide and almost
100 metres high.
Downstream of
the present Falls
a series of zig-zag
gorges marks the
positions of
waterfalls. Gravel
deposits now
preserved near
the edges of the
gorges were laid
down on the bed
of the river and
have been
exposed by its
Study of stone
preserved in
these gravels
( J. D. Clark 1950,
1990) enabled a
sequence of
industries to be
proposed and
correlated with
the back-cutting
processes of the
Victoria Falls.

Sangoan represents an adaptation to heavily forested conditions and that
some of its characteristic artefacts were used for woodworking is thus no
longer tenable.
These general points having been made, it is now appropriate to survey
the archaeological evidence for the Acheulean and Sangoan in the various
regions of Africa, before returning to a more detailed consideration of over-
all issues. Organisation of the regional evidence is not easy in view of the
enormously long time-span involved; a basically geographical arrangement
has been adopted.
60 afric an archaeolog y

Acheulean in eastern Africa
The earliest appearance of Acheulean-type industries in eastern Africa is
dif¬cult to pinpoint. The archaeological record at this time, about 1.7--1.5
million years ago, presents a confusing complexity of artefact assemblages;
the situation is further obscured by the terminology that is conventionally
applied. It is helpful ¬rst to discuss the archaeological evidence from the
sites of Koobi Fora and Olduvai Gorge which, although probably not the
earliest, are most comprehensively known.
In the Koobi Fora area east of Lake Turkana, several sites have yielded the
distinctive Karari industry ( J. W. K. Harris and Isaac 1976; Isaac 1997). So far
known only from Koobi Fora, the Karari industry occurs on a number of
sites strati¬ed between the Okote Tuff of 1.65 million years ago and the 1.39
million-year-old Chari Tuff (F. H. Brown 1994). By contrast, all artefact assem-
blages from below the Okote Tuff are attributed to the simpler Oldowan
KBS industry (p. 38 and Fig. 13, above). Like the KBS ones, Karari sites are
mostly located near ancient stream-courses. Karari artefacts include numer-
ous cores which have often been regarded as heavy-duty tools, some of them
bifacially worked. Edge-trimmed ¬‚akes are also represented. A good level of
stone-working skill is evidenced and some types closely resemble artefacts
once regarded as characteristic of much later periods (Fig. 25). At other
broadly contemporary sites, such as that code-named FxJj 50 (Bunn et al.
1980), the simpler artefacts of KBS type continued to be made.
The Karari industry appears largely, but perhaps not completely, to have
replaced the KBS industry of the Oldowan tradition at Koobi Fora by about
1.5 million years ago. These advances were broadly contemporary with the
appearance in the Lake Turkana Basin of a new hominid, Homo ergaster
(Fig. 26), both at Koobi Fora around 1.8--1.7 million years ago and, somewhat
later, at Nariokotome on the west side of the lake.
The Nariokotome hominid remains (A. C. Walker and Leakey 1993), dated
to about 1.6 million years ago, comprise the near-complete skeleton of an
immature male whose physical appearance would have been far closer to
that of modern people than that of Homo habilis or the australopithecines
had been. He was probably about 10 years old at death. He had an elon-
gated physique, with narrow hips and a barrel-shaped chest which contrasted
with those of his predecessors. His cranial capacity was approximately 880
cubic centimetres, representing a 40 per cent increase over the average for
H. habilis. Had he reached maturity, it is estimated that he would have been
about 1.8 metres tall and 68 kilogrammes in weight -- comfortably within the
ranges displayed by modern human populations. It should be emphasised
that, although Acheulean-type artefacts have been recovered from deposits
The consolidation of basic human culture 61

Fig. 25: Artefacts
of the Karari
industry from
Koobi Fora (after
Isaac and Harris

Fig. 26: Skull of
Homo ergaster,
Lake Turkana

of the Nachukui Formation west of Lake Turkana which are contemporary
with -- or slightly older than -- the Nariokotome skeleton (H. Roche, cited
by Klein 1999), none was found in direct association with these hominid
At Olduvai Gorge, various artefact assemblages of this age, ¬rst occurring
in the middle part of Bed II, have been divided between ˜Developed Oldowan™
and ˜Acheulean™ categories (M. D. Leakey 1971, 1975, 1976). They show both
62 afric an archaeolog y

a greater variety of tool types and more standardisation than do the Bed-I
industries (pp. 40--2 above). Developed Oldowan assemblages contain a few
simply ¬‚aked bifacial tools which recall the more numerous, and gener-
ally more competently ¬‚aked, handaxes and cleavers of the assemblages
described as Acheulean (Fig. 27). However, both are here regarded as part of
the overall range of variation within a loosely de¬ned Acheulean (pp. 84--6
In the middle and upper parts of Bed II at Olduvai, stone tools are more
numerous relative to the faunal remains than was the case in the lower
levels. The same is true for Beds III and IV (M. D. Leakey and Roe 1994),
although in no case can one be certain that the site is undisturbed and
that no non-human agencies have contributed to its disposition. It is also
noteworthy that large animals, such as rhinoceros and giraffe, are more
commonly represented on the later sites: the two observations may well be
interconnected. By this time the Bed-I lake had largely dried up and an open
grassland environment is indicated. It has been argued that the majority of
the hominid sites at Olduvai represent dry-season encampments (Speth and
Davis 1976; see also Blumenschine and Masao 1991). The Olduvai sequence
is summarised in Figure 28.
The earliest hominid fossils from Olduvai have been noted in chapter 2.
Specimens attributed to Paranthropus (Australopithecus) boisei continue into
the upper part of Bed II, but Homo habilis appears to have been replaced by
H. ergaster in upper Bed II and in Beds III and IV (Rightmire 1979b, 1990;
M. D. Leakey and Hay 1982). Later forms of Homo are poorly represented at
Olduvai, but a fragmentary skull from Lake Ndutu, some 20 kilometres to
the west (Mturi 1976; Clarke 1976, 1990), is attributed to H. heidelbergensis and
may be broadly contemporary with the latest occurrences of Acheulean-type
artefacts at Olduvai.
Artefacts of Acheulean type are well documented by 1.5--1.4 million years
ago in deep water-lain deposits with interbedded volcanic tuffs at Konso
(formerly known as Konso-Gardula) in southern Ethiopia (Asfaw et al. 1992;
Yonas Beyene et al. 1996, 1997), but have not yet been published in detail.
Hominid fossils include several that are attributed to a robust australop-
ithecine and two Homo ergaster specimens (Suwa et al. 1997; Klein 1999).
In the Middle Awash area of the Ethiopian Rift Valley, establishment of an
absolute chronology for the Acheulean-type artefact assemblages has proved
problematic (Renne in de Heinzelin et al. 2000); mode-2 industries were
almost certainly present there by 1.0 million years ago and continued until
well after 400,000 years ago. In some instances, in-situ occurrences were asso-
ciated with the remains of a single large animal, usually a hippopotamus or
elephant; these are interpreted as butchery sites. An age determination of
about 640,000 years, relating to volcanic deposits in the Bodo and Dawaitoli
The consolidation of basic human culture 63

Fig. 27: Stone
artefacts from
Bed II at Olduvai
Gorge (after M. D.
Leakey 1971):
1, 6, handaxes;
2, burin or
graver; 3, 8, edge-
retouched ¬‚akes;
4, cleaver;
5, scaled piece;
7, ¬‚ake retouched
to a point,
perhaps for use
as an awl
64 afric an archaeolog y

Fig. 28:
Summary of the
sequence at
Olduvai Gorge.
The Acheulean
and ˜Developed
Oldowan™ are
believed by many
to be facies of the
same tradition.

areas where a hominid skull (discussed below) and artefacts of types des-
ignated ˜Middle Acheulean™ and ˜Oldowan™ have been recovered ( J. D. Clark
et al. 1994), is more problematic. If the age of the ˜Oldowan™ assemblages,
including one that was found in probable association with hippopotamus
bones, is con¬rmed, it will be tempting to compare it with those lacking
bifaces in Bed III at Olduvai, noted above. Sites with very small handaxes and
prepared radially ¬‚aked cores (cf. p. 92 below) are also present in the Middle
Awash area: their position in the sequence cannot yet be demonstrated, but
they may belong to a very late phase.
In the Middle Awash Valley, numerous hominid fossils have been recov-
ered from contexts associated with Acheulean-type artefacts (White in
The consolidation of basic human culture 65

Fig. 29: The Bodo

de Heinzelin et al. 2000), but only in one case are signi¬cant details yet
available. The Bodo skull (Fig. 29), found in 1976 in deposits subsequently
dated to 600,000 years ago, displays features generally regarded as interme-
diate between those of Homo ergaster and those of H. sapiens (Conroy et al.
1978; Rightmire 1996). It is strikingly similar to the skull from Broken Hill
mine at Kabwe in Zambia, discussed below, designated ˜archaic H. sapiens™
or ˜H. heidelbergensis™. White (1986) describes cut-marks which suggest that
the Bodo skull may have been purposefully de¬‚eshed. A skull from Buia,
across the border in what is now Eritrea, is probably somewhat earlier in
date (Abbate et al. 1998).
A further early East African occurrence of Acheulean-type artefacts is
at Peninj, beside Lake Natron, near the Kenya/Tanzania border northeast
of Olduvai (Isaac 1967; see also Dominguez-Rodrigo et al. 2001). Two in-
situ Acheulean horizons have been located in deep deposits that indicate
a lake-shore environment similar to that of the present time. The ¬nely
preserved lower jaw of a robust australopithecine was found in a context
somewhat earlier than the Acheulean-associated horizons, which are dated
by potassium/argon analysis to about 1.4 million years ago (Isaac and Curtis
Probably somewhat later, but at least one million years old, is the ¬rst
evidence for Acheulean-related occupation of the Ethiopian plateau (Cachel
66 afric an archaeolog y


. 2
( 11)