5. Tectonics, Volcanism, Landscape Structure
and Human Evolution in the African Rift
Geoff Bailey, Geoffrey King and Isabelle Manighetti
Tectonic movements and volcanism in the African Rift have usually been considered of relevance to
human evolution only at very large geographical and chronological scales, principally in relation to longterm topographic and climatic variation at the continental scale. At the more loco1 scale of catchment
basins and individual sites, tectonicfeatures are generally considered to be at worst disruptive and at best
incidental features enhancing the preservation and exposure of early sites. We demonstrate that recent
lava flows and fault scarps in a tectonically active region create a distinctive landscape structure with a
complex and highly differentiated topography of enclosures, barriers and fertile basins. This landscape
structure has an important potential impact on the CO-evolutionof prey-predator interactions and on
interspecific relationships more generally. In particular, we suggest that it would have oflered unique
opportunities for the development of a hominid niche characterised by bipedalism, meat-eating and stone
tool use. These landscape features are best appreciated by looking at areas which today have rapid rates
of tectonic movement and frequent volcanic activity, as in eastern Afar and Djibouti. These provide a
better analogy fur the Plio-Pleistocene environments occupied by early hominids than the present-day
landscapes where their fossil remains and artefacts have been discovered. The latter areas are now less
active than was the case when the sites were fumed. l l e y have also been radically transfomed by
ongoing geomorphological processes in the intervening millennia. Thus, previous attempts to reconstruct
the local landscape setting adjacent to these early howinid sites necessarily rely on limited geological
windows into the ancient land surface and thus tend to filter out small-scale topographic detail because
it cannot be reliably identwed. It is precisely this local detail that we consider to be of importance in
understanding the environmental contribution to CO-evolutionaryd e ~ e ~ o ~ ~ e n ~ s .
Keywords: NORMALFAULTING;
LAVAFLOWS;
AFAR;AFRICAN
RIFT; HOMINIDS.
INTRODUCTION
Our aim in this paper is to bring together two bodies of
knowledge that have, for the most part, been pursued in
isolation from each other. On the one hand is the
geological investigation of the dynamics of rift formation
using the new techniques of tectonic geomorphology.
Considerable advances have been made during the past
decade in our understanding of African tectonics both in
terms of large-scale dynamics and, of particular relevance
to this paper, their influence on local and regional changes
of the physical environment (Stein et al. 1991;Manighetti
1993: Manighetti et al. 1997, 1998). These studies have,
however, been largeIy pursued without reference to their
potential impact on the course of human development.
On the other hand the palaeoanthropological and
archaeological investigation of human evolution has
focused on such issues as changes in the biological and
cultural potential of early horninids, their intra-specific
social interactions, and their inter-specific ecological
interactions with prey and predator organisms. Discussion
ofthe physical environment in relation to early hominids
has mainly emphasised large-scale changes of climate,
32
GeofS Bailey, GeofSrey K ring and Isabelle Manighetti
vegetation and tectonics, and interactions between them
(e.g. Foley 1994, in press; Partridge et al. 1995a; Vrba
1996; Vrba et al. 1995), or small-scale reconstructions
of sedimentary environments, food and raw materials
available within the vicinity of archaeological sites (e.g.
Blumenschine & Peters 1998; Brown & Feibel 1991;
Harris & Herbich 1978; Rapp & Vondra I98 1). Tectonic
factors have in general played very little role in interpretation except in indirect terms: as an ultimate cause of
global climatic change (Ruddiman & Raymo 1988); as
an indirect forcing agent on mammalian evolution through
the impact on regional climatic variation (Partridge et
al. 1995b); as a source of ecological diversity (Coppens
1994; Foky 1987: Gamble 1993); or simply as a
mechanism for accelerating the protection and discovery
of finds by rapid sedimentation and subsequent exposure
by erosion.
Thus, the landscapes studied by geomorphologists,
geologists and geophysicists are typically dominated by
physical dynamics, and the human occupants are essentially out of sight or at best passive spectators. Conversely,
an archaeological or palaeoanthropological perspective
is one dominated by a foreground of biological and cultural
dynamics with hominids as the centre of focus and an
essentially passive and distant, albeit variable and
changing, physical environment. The artist's reconstruction of an early hominid scene (Fig. 5.1) offers a
graphic if somewhat exaggerated ilIustration of this point,
with a foreground of active and indeed violent socid
interactions, and an environmentalbackground composed,
appropfiately enough, of a volcanic mountain largely
obscured by cloud.
Here we focus on the dynamic interactions that occur
at the interface between the physical environment and
hliman behaviour at the local scale. In particular we aim
to show that the tectonics of the African Rift create a
distinctive and complex topographic structure characterised by varying combinations of changing lake basins
and river vaIleys, fauIt scarps and lava flows. We argue
that a landscape structured in this way was highly
attractive to early hominids, and may have exerted
selective pressures favouring bipedalism, the exploitation
of animaI foods, and evolutionary divergence.
Discussion of interactions between humans and the
physical environment tends to veer towards one of two
extremes. Either humans are seen as passive tools of
environmental change, or the environment is treated as
essentially inert until acted on by human agency. Both
are equally deterministic in their own way and both imply
an essentially one-way relationship - either the physical
environment is seen as determining behaviour, or
behaviour is seen as determining what is significant in
the physical environment. Intermediate interactions of
varying strength can, however, be envisaged. h the
hypothesis that we advance below, we do not imply that
the course of human evolution was determined by the
structbre of tectonic landscapes. Our point is rather that
the interaction between hominids and tectonically active
environments resulted in new configurations of hominid
behaviour that wouId not otherwise have occuhed. Early
hominids selected certain sorts of environments, and these
in turn selected for certain sorts of hominFd behaviours
in a process of reciprocal interaction that amplified some
patterns of behaviour at the expense o& others. This
process is simiIar in some respects to that of a coevolutionary relationship, commonly defined in biology
as a situation in which two or more taxa undergo
evolutionary change as a result of reciprocal selective
pressures that each imposes on the other through their
mutual ecoIogica1 interaction (Pianka 1980). Recent
examples of land use, where human activity is having a
dramatic impact on the physical landscape and the
changed physical landscape in its turn is further affecting
human activity, could properly, in our view, be described
as an example of a CO-evolutionaryprocess invoIving
reciprocal interactions between physical, biological and
cultural variables. The example that we describe below
is not strictly a case of CO-evolutionarydevelopment in
that sense because the physical landscape was not (so
far as we know) affected by the presence of hominid or
other large-mammal activity. On the other hand, the
distinctive landscapes that we describe below could have
significantly altered or accelerated the pattern of coevolutionary relationships between biological species,
and cannot be treated as an essentially passive or uniform
tabula rasa awaiting the imprint of ecologicaI and
evolutionary processes. We suggest that concepts of coevolutionary behaviour and environmental selection
provide a fruitful framework for examining interactions
between variations of the physical environment and its
biotic occupants including humans, and one that avoids
the charge of determinism and the consequent dismissal
of relevant factors - environmental, behavioural or
cultural as the case may be.
TECTONIC ENVIRONMENTS
AND PALAEOLITHIC SUBSISTENCE
We begin with a brief example from the Middle and
Upper Palaeolithic of north-west Greece, an area which
is subject to very high rates of tectonic activity as a
consequence of its position at the boundary between the
African and European plates (BaiIey et al. 1993; King et
al. 1994). It is also an area that has been the focus of
detailed studies of Palaeolithic environment, economy
and archaeology (Bailey 1997). We emphasise and
elaborate on the following four points:
I. Tectonic activity accelerates processes of landscape
change both directly by uplift and subsidence, and
indirectly by amplifying or moderating the effect of
climatic change and human land use, and it does so
at a variety of chronological and geographical scales.
Tectonics, Volcanism, Landscape Structure and Human Evolution in the African R@
/
33
d
Figure 5.I Artist's reconstruction of early hominid social interactions in a lake-edge setting.
2. Tectonic change can create and renew locaI landscape
features that are attractive to human settlement. These
features concentratewater supplies,plants and animal
foods, or make them more easily accessible, and thus
sustain local conditions of dynamic equilibrium and
settlement stability for long periods.
3. The impact of underlying tectonic processes is not
uniform across a regional landscape. The same forces
that produce stabiIity in one part of a region may be
disruptive in other parts.
4. The overall effect of these tectonic processes is to
greatIy increase the patchiness of the environment,
both spatially and chrono1ogically. Such patchiness
can have a significant impact on ecological and
evolutionary processes.
Tectonic landscapes are highIy dynamic and are liable
to undergo relativeIy rapid and dramatic re-mouIding of
the physical surface. They are characterised by a compIex
topography with multiple series of uplifting mountain
rages and intervening valleys, rivers that cut across this
ridge-and-valley pattern often carving deep-cut gorges,
and more localised basins of subsidence which act as
sediment and water traps and are often fiIIed with lakes.
They produce active erosion and sedimentation alternating over quite short distances and complex hydrologicaI regimes, and these patterns may be further acted
on by changes in climate and vegetation or intensive
human land-use practices.
Areas of the earth's surface primarily subjected to
compression by convergent plate motions show a general
trend towards regional uplift and mountain building, as is
the case in north-west Greece. In extensional areas where
the earth's crust is being stretched by subduction or plate
separation, the general trend is towards subsidence, as in
the Aegean basin, It is important to appreciate that this
alternation of upIift and subsidence on sub-continental
and larger geogaphical scales is also mirrored at a smaller
scale. Reverse faulting in compressiona1 environments
and normal faulting in extensional ones both result in
adjacent zones of uplift and subsidence at the Iocal scale,
such as fault-bounded lake basins, and similar patterns
recur at the regionaI scale.
These changes can create and renew local environments that are attractive to human settlement. This
attractiveness takes two forms. Localised subsidence
focuses sediment accumulation and water supplies and
thus creates a highly fertile environment that concentrates
plant and animal life. At the same time the associated
uplift creates a series of barriers that greatly faciIitates
the human prediction and manipulation of the movements
of mobile animal resources. This makes potentialIy fastmoving or eIusive prey species more easily accessible to
an intelligent predator without the need for biological
34
GsoffBailey, GeoffreyKin:g and Isabdle Manighetti
abiIities of rapid movement or elaborate technological
means of killing at a distance such as spear throwers and
guns. Repeated fault movements on the same axis
continuously rejuvenate these features, sustaining a fertile
environment and maintaining sharp topographic barriers,
and can create a climatically insensitive Iocal area that
sustains attractive conditions for human existence
regardIess of external cIimatic changes.
The region of north-west Greece has been undergoing
compression throughout the Tertiary and Quaternary
periods, and continues to show rates of seismicity and
tectonic activity which are amongst some of the highest
in the world (King et al. 1993). The archaeological record
extends back at least over the past 100,000 years, with
long sequences of Middle and Upper PalaeoIithic material
in open air sites or limestone rockshelters associated
with a subsistence economy in which large game animals
(red deer, horse, cattle and ibex) were a major resource
(Bailey 19971,
The archaeological sites are associated with tectonically created features such as fault-bounded lake basins
and limestone gorges, and these appear to have favoured
human habitation in a variety of ways and at a variety of
geographical scales. At the regional scale, faulting and
instability has created impressive NW-SE trending
limestone mountain ridges, with softer younger flysch
rocks on their lower flanks that create large-scaIe
badlands erosion. These pose major barriers to animal
movement and demarcate virtually enclosed large-scale
grazing basins with limited entry and exit points that
facilitate the control and prediction of seasonal animal
migrations (Fig. 5.2).
Figuce 5.2 Map of Epirus showing basins and barriers
at the regional scale,
At a more local scale, in the vicinity of individual
sites, cumulative uplift on individual faults and @sociated
subsidence has produced and maintained local lakes and
sediment traps. At the local scale, as at the regional
scale, active tectonics also creates physicd barriers or
natural fences and local enclosures. These can be used
to control and predict the local movemnts of herd
animals, and to trap or corral them. Such a local
topography also provides secluded and protected locations from which people can observe animals without
disturbing them (Fig. 5.3). These same locations abo
provide protection for the human group from predators
or human competitors. Site sequences in these locations
show rich and repeated human occupation over many
millennia. Typically they span major episodes of climatic
change from virtually fully glacial to fully interglacial
conditions and persist throughout conditions of late
glacial aridity.
Tectonic activity does not provide uniformly favourable
conditions for human settlement throughout the region.
In some areas continuing activity has transformed local
Figure 5.3 Map of local territory of the Kastritsa
rockshelter, showing relationship to lake-edge environments, main routes of animal movement and local
dopographic barriers. The side was occupied frona
c. 21,000 to 11.000 BP.
Tectonics, Volcanism, Landscape Structure and Human Evolution in [he African Rift
environments which' were once attractive basins and
sediment traps into uplifted and eroded badlands landscapes. This is evidently so in the case of the Kokkinopilos
red beds, which contain some of the earliest Palaeolithic
artefacts of the region (Bailey et al. 1992;Runnels & Van
Andel 1993). Hare a once-fertile basin of sediment
accumulation has been transformed into a zone of erosion.
A similar transformation has also taken place in the
Mazaraki basin in the north of the region. Conversely the
Ioannina lake basin has remained in broadly its present
form for at least the duration of the Pleistocene and
probably much longer (Brousoulis et al. 1999). Ongoing
tectonic activity can, therefore, have disruptive as well as
stabilising consequences for human settlement and
subsistence, depending on the time span of interest, the
rates of tectonic activity, and the nature of the locaI fault
motions.
From the point of view of human hunters, the complex
topography is something of a two-edged sword. For the
same features that appear to facilitate human access to
mobile or elusive prey can also provide means of escape
or refuge for the prey species. In the long-run this may
be of over-riding benefit to the human population by
maintaining resilient relationships which reduce the risk
of extinction to both animal-prey and human-predator
populations alike (Winder 1997). By the time of the
Upper Palaeolithic period, if not earlier, human population~were clearly skilful hunters who had established
effective relationships with a variety bf prey animals
including both fast-moving (red deer and horse), defensive (cattle) and eIusive (ibex and chamois) animals.
The basis of that skill, according to our Greek case study,
lies less in the use of technology, than in the intelligent
manipulation of topographic features to monitor and
controI Iarge areas of the landscape. We cannot of course
assume that earlier hominids had the same sorts of
technological and cognitive skills. But we see no reason
why tectonic features such as we have described above
should not have offered significant CO-evolutionary
opportunities arising out of a dynamic and 'patchy'
topography at any period or in any biological context.
Indeed, it is part of our argument that these features in
the African context may have actively selected for
emergent cognitive skills that we see in a more fully
developed form in the later Pleistocene.
THE EAST AFRICAN RIFT
The East African Rift is a much larger and longer-lived
structure than those we have described for Greece, and
the archaeological record is much longer, thereby
requiring us to think about the dynamic implications at a
lager geographical and chronologica1scale. NevertheIess
the initial focus here as in Greece is the local Iandscnpe
structure.
In general the development of the East African Rift
35
involves processes of plate separation that have been
underway for at least 12 million years, and exhibit
extensional features, with normal fault scarps and
volcanic activity. At present, much of the East African
Rift is relatively inactive (extension rate of less than 5
mmlyr). However, although not yet well documented,
earIier rates appear to have been greater, with the features
we discuss below more widespread. Furthermore, the
most: intensive voIcanic activity seems to have been
associated with the inception of rifting and has become
more subdued with time (Manighetti 1993; Tapponnier
et al. 19901, a feature also observed for other continental
rifts such as the Baikal Rift in Siberia. It is not surprising,
then, that the features we emphasise have in general
been overlooked or discounted in previous environmental
reconstructions. In order to appreciate what local
environments would have looked like to their early
hominid occupants, we need to examine currently active
areas of the Rift such as the Afar depression.
The tectonics of Afar
The Afar depression is a complex system of active
features (faults, fissures and volcanoes) resulting from
the interaction between the Red Sea and the Gulf of
Aden rifts (Fig. 5.4) (De Chabalier & Avouac 1994;
Deniel et al. 1994; Manighetti et al. 1997; Stein et al.
1991; Tapponier et al. 1990). For many parts of the East
African Rift, due south of the Afar depression, the
geometry of opening seems simpler, with only a single,
major active rift. Nonetheless similar processes recur.
Activity causes the centraI part of the rift to subside and
one or both of the adjacent sides to uplift and tilt away
from the active axis. This typically means that earlier
rift axes that form these flanks become perched at a
higher Ievel, from tens to hundreds of metres above the
new rift.
Active volcanoes appear both within the active rift
and on the rift flanks. Similar patterns to those found
throughout the African Rift system are seen for other
extensional regions: the Basin and Range of the USA;
the Aegean region; North Island New Zealand; and
Iceland. By the standards of some other continental rift
systems (North Sea, Rhine, Rhone system, the Aegean
system or the southern Basin and Range system),
however, the volcanic activity of the African Rifts is
high compared to the rate of extension (e.g. Ellis & King
1991).
The active central grabens commonly form internaIIy
draining basins, dotted with volcanoes and in many places
covered by lava flows. Smaller or larger lakes are found
everywhere. Contemporary sedimentation consists of
slope wash, river and Iake-deposits consisting of fine
silt, reworked volcanic ash, and evaporites. The regions
around the rivers and lakes have supported many African
savannah animals in the past although hunting with
automatic riffes has now greatly reduced their numbers.
36
GeoffBailey, GeofSrey Kii~g and Isabelle Manighetti
Almost everywhere recent lava flows disrupt the useful
land. They are not traversed by roads and represent
formidable barriers to movement of any sort, with steep
sides and jagged broken surfaces (Fig. 5.5). In time,
especially in wet climates, they may become eroded and
smoothed or reduced to bouIder fields. In drier climates,
lava fields with ages of many thousands of years remain
impassable for domestic animals, or any large quadrupeds
for that matter. Modern humans, in contrast, can cross
lava flows but rarely do so unless there are good reasons.
Regions of volcanic activity are also associated with
vertical faulting, resulting in impressive vertical barriers
(Fig. 5.6). In contrast, the fault scarps in non-volcanic
regions commonly have slope angles of 45" or less, which
form less impressive obstacles to the movement of large
mammals.
I
Simplified Tectonic Map of
Afar
di8&@
>>..-*
1
:o~tkm
Volcanoes and lava fields
The lava fIows and fault scarps associated with the areas
around the Manda volcano (belonging to the MandaHararo rift) and the Gablaytu and Loma volcanoes (Fig.
5.4) create a complex and patchy local mosaic of barriers
and small basins, with more open and extensive savannah
regions beyond. The Manda volcano lies on the active
rift axis and is cut by SSE-NNW oriented normal faults,
the two largest of which have throws of some tens of
metres. SmaIler ones with throws of a few metres are not
marked in Fig. 5.4. The lava flows are typically 3.5 m
thick with steep sides (see Fig. 5.5), and the fault scarps
are steep and very hard to climb, as is the case throughout
the region (Fig. 5.6). An annotated aerial view of the
Manda area showing the combined effect of these features
"onlocal landscape structure is shown in Fig. 5.7, and an
oblique general view in Fig. 5.8.
The age of the visible pare of the Manda voIcanic
system is thought to be between 20,000 and 40,000.
This is partly derived from direct dating of the Invas [Y.
Gillot pers, comm.) and partly deduced from the observed
fault offsets and a rough knowledge of the vertical slip
rates in the region (1-3 mmlyear maximum on a fault).
Within the lava flows are numerous small sedimentary
ponds completely isolated from their surroundings.
Around the Manda volcano they are associated both with
the volcanic cone and with the active faults. The present
climate is arid, but both vegetation and water are present,
and in a slightly wetter climate these would be quite
fertile enclosures.
AIthough they are completely unused now (except
for a camel track occasionally employed by smugglers
en route to Djibouti) there is ample evidence of earlier
human activity within the Manda flows. ParticularIy near
their eastern and western edges d o s e to large open
(savannah-like) spaces, numerous worked artefacts were
observed. These have not yet been subject to systematic
study, nor are they dated. However, footprints of an
adult and of an infant were found in water re-deposited
Figure 5.4 A simpl$ed tectonic map of the Afar region,
showing the distribution of main active faults and recent
lava flows, and areas where we have undertaken more
detailed examination of local landscapes. Thin lines murk
active faults and solid lines indicate the centres of the
active r#s. The faults and lava flows shown are
suflciently young to exhibit the features shown in Figs.
5.5-5,8.The location of Hadar where 'hey' was found
is shown and today lies oa the smoothed uplifted flank
of the contemporary active East African rift in a setting
quite different from 4 Ma.. The inset shows two crosssections of the rift at 4 Ma ago and as it is today.
Essentially the same cross-section can be drawn for
other African Rift hominid sites.
volcanic ash, which lies within the period of circa 40,000
to 20,000 years ago according to our preliminary dating
of the associated lava flows (Fig. 5.9 and inset of Fig.
5.7).
Two other volcanic systems were examined, Gablaytu
and Loma, together with other basalt cliffs created by
recent faulting. A11 of these volcanic systems and cliffs
were associated with extensive stone-artefact scatters.
Both the GabIaytu and Loma volcanoes have internal safe
'
I J ~ I T ~ ~ . YSC~~~IU~ I~. IJ I IaJ L~' Jlunacl~~
d
Emlurion b~ rhe Africun Riff
~ ' L ~ C I U I I IJo/~~~~.utti.lnl,
~I~S,
................
I ' ' i # u ~5..5
~ I V ~
O
P
................
37
.
IOW
in ilrr M(~nrlcrVolcntlio sysfrrrl ,sho\t.r.ing the tyj>ic:c11 hrigIti und jcrpgrdnesr c $ ~y o u r fenlurr..
~~
?.l~rhc~
pri,sP rr hr~rrir,.
I O t t ~ o v r ~ hur
~ ~ (IT(,
~ ~ tioi
~ a tills-r~rmounrublebv brpeds.
38
Geoff Bailey, GeofSrey King and Isabelle Manighetti
Figure 5.7 An annotated satellhe photograph of faulted lava flows associated with the Manda Volcanic system.
Part offhe region is shown in an oblique photograph, Fig. 5.8, token in the direction indicated by an arrow, from
the summit of the southern Mundu volcanic cone. The locations of extensive artefact scatters are shown, together
with the location of the footprints described in the text. Internul safe oreas are highlighfed by a light shading.
-Figure 5.8 Oblique phofograph of rhe Manda region shown in Fig. 5.7. A lovopow with ~ p i c a features
l
is in the
foreground and additional extensive lava flows appear as dark areas in the middle distance.
,
;
Tectonics, Volcunism, Landscape Srructure and Human Evolution in the African Rift
39
Figure 5.9 20,000 year-old footprints in water re-deposited volcanic ask deposits in the Manda region
(see Fig. 5.7for location).
increase in height by a metre.There is clearly a succession
areas. The Loma volcano has a crater lake with an
of Iocal areas partially or totally enclosed by lava flows,
obviously reliable water source and Gablaytu probably
The smaller areas are nested within Iarger enclosures
SO. Some correlation between the location of artefact
defined by fault barriers and more extensive lavas, and
scatters and blind canyons appeared evident for all of the
these give way to more extensive areas of open terrain
volcanoes, suggesting a hunting strategy involving
that would provide suitable habitat for larger game
entrapment.
animals. It is this combination of enclosures at varying
There is evidence that the now contracted Lake Abbe
spatid scales that combines security with access to food
(Fig. 5.3) has in the past reached and surrounded a nearby
supplies. And it is these sorts of features that are typically
fault scarp. The enclosed spaces so created may aIso have
associated with Iake basins in active areas of the African
played a role similar to those that we attribute to faulted
Rift, and which we would expect to have characterised
volcanoes. Such lake-side environments would provide
many of the l&e-edge environments inhabited by early
local barriers and enclosures that could be used in a
hominids.
simiIar way to fauIted Iavn flows, and similar environments
Figure 5.12 is a cartoon summarising many of the
could also be produced by down-cutting rivers.
features
that are actually observed in an active rift. To
The region of Lake Asal in Djibouti provides another
allow
them
to be shown in one picture the flat savannah
example of the way in which faults and Iava flows may
areas
are
greatly
reduced in size, as are the distances
produce a complex series of barriers and partially enclosed
between
volcanic
centres.
Important features shown are
areas of varying size adjacent to alake-edge environment
enclosed
areas
within
lavaflows,
blind valleys created by
(Figs. 5.10 and 5.11). Today the climate here is very arid,
faulted volcanic cones, and fault scarps that can extend
and the IocaI environments are fairly barren, being used
for many tens of kilomet~sacross otherwise featureless
only for brief seasonal stopovers by mobile pastoralists.
Only smaIl climatic changes, however, would C O ~ V . E I ~ alluvial regions, and which potentialIy provide some
security for travelling greater distances across the
this area into a more fertile region with a variety of plant
landscape.
and animal food supplies. The maze of lava flows and
fault scarps visible today is not a transient feature of the
landscape but a persistent feature that is constantly
Active tectonics and inward draiaage
renewedby repeated vobanic activity and fault movement.
Although the centre of discussion in this paper is the
Considering the 100 km length of nft in Djibouti as an
role
of lava flows and the vertical or near vertical faulting
example, several eruptions occur per century creating
associated
with active tectonics, it is worth noting that
lava flows of 50 km2or more each time. At the same time,
active
extension
also creates internal drainage systems.
vertical fault scarps severaI kilometres in length can
GeoJSPBailey, Geoflrey King and Isabelle Manighetti
40
1
.........
.........
.* ....
- *,.
..A
l
/
....
.........
.........
.....7
5
vallevs with steep
"terraced sides
ll-d-dl-dl--l
1
-- -.
flows younger than 50 years
m flows
....
..........
.........
......
....
younger than 200 years
perched lake beds
+
0
fissure eruption
v ~ l c a n ~ ccraters
maior (>30
\ inkrneible
visible
centres
or edif~ces
metres) vertical fault escarpments
smaller escarpments are not
I
1
Figure 5.10 Map of Lake Asal region: ( a ) satellite photograph showing general features; ( b ) detailed mapping of
faults and topography within the white box of (a); (c) simpl8ed map of landscape features showing major faults
escarpments, volcanoes and lavaflows.
Tectonics, Volcanim, Landscape Sfructure and Human Evolution in the African Rij?
41
Figure 5.11 Oblique photograph of Lake Asal region, taken from the top of one ofthe highest fault escarpments
shown in Fig. 5.10 looking westwards towards Lake Asal. L.uva j h w s and smaller faults create U series of partial
and successively smaller enclosures as one movesfrom lefr to right of the picture and towards the lake shore in the
right-hand corner.
For example all of the lakes between Lake Asal in
Djibouti and Lake Turkana in Kenya currently have no
outlet to the sea. Such inward draining systems are very
sensitive to changes in precipitation and consequently to
relatively rapid changes of water level. Some 6000 years
ago Lake Asal was 200 metres higher and similar changes
have occurred in other lakes. We have already noted the
importance of interactions between lake shores and other
features to create complex and partially enclosed environments. Rapidly varying lake levels, by controlling the
erosional power of streams, also create steep-sided
vaIleys and terraces which can exhibit similar features
of partial enclosure and barriers to movement that we
attribute to faults and lava flows.
EVOLUTIONARY OPPORTUNITIES
AND CONSEQUENCES
Meat eating and bipedalisrn
As Davis (1987,94) has noted, "The beginning of meateating, like the adoption of bipedalism, is shrouded in
mystery", although both are considered to be critical
factors in the development of human characteristics.
Current evidence suggests that both scavenging and
hunting are likely to have played a significant role in
hominid diets from at least as early as the fist appearance
of the Homo lineage (Bum & Ezzo 1993). One puzzle,
however, is how relatively defenceless and unspecidised
e d y horninids descended from vegetarian tree dweIIers
could have developed a meat-eating niche in the open
savannah in competition with a number and variety of
highly specialised carnivores and scavengers.
Scavenging offers one potential solution to the
problem of gaining access to animaI foods in the open
savannah: that is reliance on specialised carnivores to
do the hard work of running down the animal, and then
moving in to take what is left. But in solving one problem,
human scavengers are exposed to another, and that is the
risk of themselves becoming prey victims. Studies of
modern situations show that even today human groups,
and especidly women and children, are vulnerable to
lethal attacks by carnivores (Treves & Naughton-Treves
1999). The problem of protection for hominid groups
with dependent offspring has been discussed by a number
of authors (e.g. Foley 1987, 183), and the usual solution
has been to suggest the use of trees, which act as sources
of food as we11 as shelter. Trees, however, also have
disadvantages. Young chimpanzees, for example, are
exposed to the risk of falls and require continuous
protection (Goodall 1968). There is also the risk of
exposure to tree-climbing predators.
Daytime foraging has been suggested by Wheeler
(1984, 1985) as another tactic for avoiding carnivores,
and one which might also have increased the selection
pressure for bipedalism as a means of maximising heat
loss. This hypothesis, however, leaves open the issue of
protecting the young from predators, especially where
there is an extended period of dependence on adult
protection and food provisioning beyond weaning.
We suggest that a topography of fault scarps, lava
42
GeoffBailey, Geoffrey King and Isabelle Manighetti
Cartoon Rift Landscape
i-
I
Types of enclosed w a s are shown that could have been exploited by early Elorninids. Such fentures rtre*sepslfiltedby
substantial Savnnnah redons in a reat active lanscnp
Old Fscnrpmcnl
Arcacnclowl bg faults nith unler
,ab
at
old c s ~ r p m c l ~
Figure 5.12 A cartoon repreientation of features within an active r$t. An older smoothed escarpment is shown to
the lefr and could have an age of 4 Mu.A river is shown cutting a narrow gorge into the escarpment. The same
hwn-cutting is responsible for revealing fossils in earlier rift floor sediments. The view shows regions enclosed by
lava flows (kupuka) or by near vertical faults or a combination of both. A lake is shown at the base of the older
eroded escarpment in a similar position to the current Lake Gamarri. Distances between volcanic features have
been contracted. Khey are commonly separated by 15 km or more. These regions have typical savannah characteristics
although faults can extend all or part of the way between the regions and offer some temporary refuge for hominids
moving between secure environments.
flows and steep sided valleys could have offered solutions
to both issues, that of security and protection, and that of
access to mobile and potentially eIusive prey, while aIso
accentuating selective pressures towards bipedalism.
The unique feature of lava flows is the creation of totally
secure environments. Basalt lava flows are remarkably
difficult to cross. A traverse of 100 metres involves
repeated cIirnbing up and down jagged fragments. Such
manoeuvres are possible for humans, as they would be
for apes, but they are extremely difficult for quadruped$
mammals. Areas enclosed by Iava flows provide protection from attack whether by speed or stealth without
the need to depend on trees or the disadvantages of
reliance on them. They aIso facilitate the protection of
the vulnerable young while adult members of the group
are engaged in subsistence practices elsewhere, and
would have created opportunities for extendin6the period
of juvenile dependence.
Some modern ground-dwellingprimates living in more
open savannah are known to make sleeping nests on cliff
faces or in caves, and the hamadryas baboons of Ethiopia
frequently make use of fault scarps for protection
(Kummer 1968). The idea of early hominids making use
of lava flows and fault scarps as protective devices is
thus entirely plausible.
Access to animal foods
Complex interlocking patterns of barriers and blind
canyons composed of fauIt scarps and lava flows would
have offered opportunities for diverting and trapping
mobile animal species without the need for biological
weapons of speed and attack. We cannot be sure that
early hominids had the ability to behave like predictive
hunters in the way we have described for Greece. Indeed
Tectonics, Volcanism, Landscape Structure and Human Evolution in the African R@
.
:
there is considerabIe controversy over the extent to which
hunting was carried out at all in the earlier periods of
human evolution. But the presence of topographic
opportunities would certainly have created selective
pressures for the development of such abilities, by
offering important competitive advantages to a vulrterable
horninid in otherwise relatively open savannah environments.
Food supplies
Unlike trees, lava flows, or small areas isolated within
them, provide little or no food, which must be brought in
from elsewhere. Animals brought down by hunting might
conceivably have been diverted into topographic traps
formed by Iava flows before being killed, so as to
minimise the distance over which the carcass had to be
carried to a safe location. Even if hunting were not
practised, meat acquired from scavenging would most
probably have required transportation as would plant
foods, if they were to be eaten at leisure in secure
locations. As noted by numerous other authors bipedalism
favours the transport of food by largely freeing the upper
limbs {Lovejoy 1981). Limited modification leaves the
same animal with the ability to negotiate cliffs and lava
flows. We emphasise that food transportation in such a
situation does not presuppose food sharing or a division
of labour as envisaged by Isaac's (1978b) original foodsharing hypothesis, though such behaviours may be an
outcome. It presupposes only the need to remove food
from the point of capture or collection to a safe location
for consumption.
Tools
Lava flows provide simple stone tools, sharp rock
fragments are to hand everywhere and are usable without
modification. The materials to create more sophisticated
tools such as obsidian are also available.
Fire
Very early evidence for the use of fire remains controversial, but the association of early hominid activity with
volcanically active areas would certainly have enhanced
the possibilities for observing and making use of the
benefits and effects of fireand heat (Gowlett et al. 1981).
Furnaroles might have encouraged experiments with
cooking.
Pressure towards change
Forested environments are essentially uniform in terms
of the physical selection pressures they impose on
species adapted to a forested habitat, except to the
extent that areas of forest may expand or contract with
climatic change. Such large-scale variations can alter
43
the patchiness of a landscape and thus the balance of
predator-prey relationships, and the general contraction
of forest in the late Tertiary is, of course, generally
considered to be a key large-scale factor in opening up
the hominid niche. At the smalIer scales that are our
focus here, however, landscapes characterised by faulting and volcanic lava flows may show greater variability
than forests in the degree to which they provide enclosure and protection. Lava flows are not all the same.
Although numerous completely enclosed areas exist,
many more are partially enclosed or are in Iava flows
that are eroded sufficiently to be more readily traversed.
Furthermore as climate changes, lava flows stay fixed.
This offers a challenging environment to a species that
inhabits them.
The existence of a range of niches that were similar
in their general characteristics but different in detail
would have provided an added incentive to evolutionary
change, either through niche separation or by selecting
for intra-species adaptability.
DISCUSSION
The modern equivalents of the sites of early hominid
finds lie among the active faults and volcanoes of the
currently active parts of the East African Rift. No reports
of studies in these regions are to be found and they are
generally regarded as inhospitable and inaccessible. Yet
within then1secure areas exist that could provide a refuge
for an ape-like creature deprived of trees and an
environment where bipedalism would confer advantages.
Even a brief visit suggests that evidence for prehistoric
occupation is strong, and that these seeminglyinhospitable
landscapes provided attractions for human settlement.
Although not many sites have been identified, those that
have are associated with the contemporaneousrift activity.
Conversely, it is notable that scarceIy any trace of stone
tools was observed in the more open terrain on theManda
Hararo rift flank, despite the fact that more than 60
locations similar in many respects to those we refer to in
the central rift were visited forpalaeomagnetic dating. By
contrast, pastoral activity is now well developed on the
rift flanks and rock paintings of domesticated camels
suggest that this may have been so for some time. The
implication is that, until the advent of animal domestication, human habitation of such regions was not practical.
Although this is the savannah in which early hominids
have commonly been placed, it appears that human
occupation has only recently extended into it.
It also appears that at least some of the tribal people
of the Kenyan Rift Valley were familiar with lava-flow
environments. When deprived of firearms by the British
in Kenya, the local people notoriously disappeared into
a region of volcanoes and lava flows into which they
could not be followed. Presumably they knew how to
exploit the environment to survive. Thus it may be at
44
Geoff Bailey, Geoflrey King and Isabelle Manighetti
least as useful to examine how modem people have
exploited the active regions of the East African Rift
prior to the appearance of guns, as a source of insight
and andogy for the sort of ecological niche that we have
identified for early horninids, as to examine the behaviour
of our nearest living relatives amongst the Great Apes,
or the behaviour of carnivores with supposed functional
similarities to hominid scavengers or hunters.
The environments that we have described for the
present Manda Bararo rift are the exact conditions that
will in due course create a future geological environment
like that in which many early hominid remains have
been found. Interbedded lacustrine and fine terrestrial
sediments are associated with ash falls, water-reworked
ash and basalt lava flows at Hadar (Taieb er al. 1976;
Taieb & Tiercelin 1979), and at other classic Rift Valley
sites such as Omo, Turkana and Olduvai (Feibel et al.
1989; Hay 1976; Rapp & Vondra 1981). Examining a
contemporary example explains why geologists have such
problems correlating strata between individual exposures
in separate valleys. Lava flows and sediment traps simply
do not correlate except over very short distances, and a
layer cake stratigraphy misses an important insight into
the nature of the original environment.
Indeed this geological issue is one of several factors
that have obscured the role of tectonics at the local scale.
Reconstructions of the landscape around early hominid
sites tend to produce a picture-of relatively smooth
landscapes lacking in physical barriers and topographical
detail (e.g. at Olorgesailie (lsaac 1978a; Shackleton
1978), atche Bed I Olduvai sites (Isaac 1981) and at
Koobi Fora (Bunn et al. 1980; Isaac & Behrensmeyer
1997). This is an almost inevitable consequence of
attempting reconstructions from a Iimited number of
geoiogical. 'windows' into a landscape that has undergone
radical alteration through ongoing rifting, tectonic activity
and erosion since the time of hominid occupation. Such
reconstructions inevitably arrive at a lowest common
denominator in which locd topographic detail is largely
eliminated for the simple reason that it cannot be
reconstructed with any confidence - or else has been
smoothed away by erosion.
There are two other reasons, in our view, why the
tectonic factors we have cited have been overlooked.
Firstly, the role of the natural environment in evolutionary
trends tends to be dedt with in very general terms, in
relation to large-scale phenomena such as regional and
global climaticand biotic changes (e.g.Vrba et al. 1995).
Numerous studies of the more recent archaeological and
pdaeoenvironmental record, however, demonstrate that
broad climatic and environmental changes can be
significantly modified or moderated by local topographic
features (e.g. Bailey 1997).Landscape structure, especially at the local scale, thus becomes a key focus for
:mall-scale interactionsbetween populations and environmental factors, and these small-scale interactions can
have a significant if poorly understood impact on both
short-term ecological interactions and, longer-term
v
evolutionary trajectories.
Secondly, the African Rift where the best known early
hominid sites occur was most probably more active
tectonically and volcanically in the late Tertiary and early
Quaternary than is the case today. The hithest rates of
activity at the present day are to be foufld in eastern Afar
and Djibouti, and highly active areas such as these are
rarely visited by modern observers and usually regarded
as arid, inhospitable and inaccessible areas of lava flows.
There is consequently a general lack of awareness about
the physical structure of such active landscapes.However,
as we have shown above, there is good evidence that
these landscapes can be very attractive for human
settlement under appropriate climatic conditions.
I
CONCLUSION
The hypothesis that we advance here is that lava flows and
the young normal faults of tectonically active areas of the
African Rift created a local landscape structure that was
uniquely appropriate as an agent of environmental selection in the early stages of hominid evolution. We suggest
that this process operated both directly, in selecting for
and amplifying specjfic feattires such as bipedalism, meateating and cognitive development, and indirectly by
controlling the pattern of interactions between species and
their CO-evolutionarydeVelopment.In particular we emphasise the following features of tectonics in the African Rift:
Unusually high rates of tectonic activity and volcanism, associated with a complex pattern of often
impassable lava flows and vertical fault scarps.
Complex patterns of enclosure at a variety of spatial
scales that would have provided opportunities to
hominids for the development of new niches as
unspecialised predators, as well as protection to
hominids as potential prey victims.
The creation and maintenance of a rich and diverse
mosaic of resources subject to varying degrees of
patchiness.
Topographic and environmental conditions that were
locally variable in time as well as space, offering
varying degrees of local isolation and/or regional
mixing, and thus acting as an environmental pacemaker for evolutionary developments.
We suggest that this hypothesis has been overlooked both
because of the difficulties of reconstructing PlioPleistocene topography and locd environmentsin ahighly
active tectonic region, and because the classic sites of
early hominid discoveries are in areas of the East African
Rift which are now probably Jess active than when the
sites were formed. Existing reconstructions which show
early hominid sites surrounded by smooth alluvial plains
are highly misleading because they are uninformed by a
knowledge of topography in modem active environments,
,
Tectonics, Volcaaisq, Landscape Structure and Haman Evolution in the African Rift
and because the complex small-scale patterning of local
faults and lava barriers that is so critical to an understanding of these environments cannot easily be recovered
from the ancient landscape and has in consequence been
erased from the reconstructions.
Finally, we acknowledge that the ideas presented above
pose two sorts of challenges to future investigation, one
practical, the other theoretical. On the practical side,
systematic testing of our basic hypothesis will require
two sorts of observations to be undertaken: the systematic
archaeological investigation of those modern environments which provide contemporary analogies for the
landscapes occupied by early hominids; and the reconstruction of past landscape structure with an eye to barriers
and 'sharp' topographic features especially at the local
scale. The need for the former arises from precisely the
same set of factors which makes the latter so difficult,
namely the large-scale transformation of locd environmental features that has been effected by continued
tectonic activity and environmentd processes of erosion
and sedimentation over hundreds of thousands of rniIlennia. We do not underestimate the diff~cultiesposed by
both approaches, the dangers of extrapolation from modem
analogues in the former case, and the uncertainties of
reconstruction in the latter. But we suggest that both need
to be attempted, and that in combination they should lead
to a better understanding of the environmental context in
which early hominid evolution took place, and a fuller
integration of environmental factors into the understanding
of ~ ~ - e ~ ~ l ~ processes.
ti~nary
On the theoretical side, we observe that differences of
scale continue to provide one of the most common sources
of confusion and misunderstanding the field human
evolution and human history, especially between specialworking in different disciplines or in different time
periods. At the same time differences of scale also lie at
the very heart of ecodynamic theory and present one of
the most difficult challenges to theoretical understanding.
Not the least part of that
is the forging of a
common language that will allow communication across
disciplinary and sub-disciplinary boundaries. However, a
common conceptual framework that accommodates
differences of scale is unlikely to emerge without a degree
of CO-evolutionary intellectual development, in which
specialists emerge from behind barriers of isolation, old
ideas are abandoned or modified, and new ones experimented with. That process of CO-evolutionaryintellectual
development is not an easy or comfoflable one, as anyone
who has engaged in a large-scde multi-disciplinaty project
will know. Without it, however, there can be no prospect
of understanding the CO-evolutionarynature of the past.
ACKNOWLEDGEMENTS
We are grateful to John Coles for identifying the source
of Fig. 5.1 and to Topham Picturepoint for permission
45
to reproduce it, to Nick Winder for highlighting the coevolutionary significance of tectonic landscapes, to John
Price for comments on fumaroles, and to Rob Foley and
two anonymous assessors for critical readings of an
earlier draft. The fieldwork which stimulated the observations and ideas reported here was undertaken during
mapping and palaeomagnetic determination of deformation processes in Afar and Djibouti and field excursions and discussions associated with the Addis Ababa
Conference of 1997 on Flood Basalts, Rifting and
Palaeoclimates in the Ethiopian Rift and Afar depression.
We thank C.N.R.S. and I.N.S.U., Paris, and the University of Newcastle-upon-Tyne for financial support.
REFERENCES
Bailey, G. N. (ed). 1997. Klithi: Paiaeolithic Ssttlemer~r and
Qlraremaly landscapes in Northwest Greece. Vol. 1: Excavation
and Intra-Site Analysis at Klizhi. Vol. 2: Klithi in its Local and
Regional Setting. Cambridge: McDonald Institute for Archaeological Research.
Bailey, G. N., King, G. C. P. and Sturdy, D. A. 1993. Active tectonics
and land use strategies: a Palaeolithic example from Northwest
Greece. Antiqlrio 67, 292-3 12.
Bailey, G. N., Papaconstantinou, V, and Sturdy, D. A. 1992.
Asprochaliko and Kokkinopilos: TL dating and reinterpretation
of Middle Palaeolithic sites in Epirus, North-west Greece.
Cambridge Archaeological Journal 2, 136-44.
Blumenschine,R.J. and Peters, C.R. 1998. Archaeological predictions
for hominid land use in the paleo-Olduvai Basin, Tanzania, during
ofHu,ann Evolution 34, 565407.
lowermost Bed II times.
Brousoulis, J., Ioakim, C., Kolovos, G. and Papanikos, D. 1999. The
loannina basin: geologicaI and palaeoenvironmcntaI evolution in
Quaternaty and historical times, pp. ?? in G.N.Bailey, E. Adam,
E.PanagopouIou, C. Perlts & K.Zachos (eds.), The Paiaeolithic
Archueology of Greece and Adjacent Areas. London: British
school at
B
~ F. H,
~ and ~Feibel,~ C. S ,. 1991. Stratigraphy, depositional
environments and palaeogeography of the Koobi Fora formation,
pp. 1-30 in Harris, J.M.(ed), Koobi Fora Research Project,
Volume 3, The Fossil Ungulates: Geology, Fossil Artiodactyis,
and Pdueoenvironmsnts. Oxford: Clarendon Press.
Bun", H. T. and Ezzo, J. A. 1993. Hunting and scavenging by PlioPleistocene hominids: nutritional constraints, archaeological
patterns, and behavioural irnplications.~~ournal
of Archaeological
Science 20, 365-98.
Bunn, H. T., Hams, J. W. K., Isaac, G. L., Kaufulu, Z., KroI1, E.,
Schick, K., Toth, N. and Behrensmeyer, A. K. 1980. FxJj 50: an
early Pleistocene site in northern Kenya. World Archaeology 12,
109-36.
Coppens, Y, 1994. East side story: the origin of humankind. Scientific
American 270, 62-9.
Davis, S.1. M. 1987. TIte Archaeology of Animals. London: Batsford.
Dc Chabalier, J.-B.and Avouac, LP. 1994. Kinematics of the AsaI
Rift Djibouti determined from the deformation of Fieale VoIcano.
Science 265, 1677-81.
Deniel C., Vidal, Ph., Coulon. C., Vellutini, P. l. and Piguet, P.
1994.TempomI evolution of mantle sources during continentd
rifting: the volcanism of Djibouti Afar. Journol of Geophysical
Restarch 99(B2), 2853-69.
Ellis, M. and King, G. C. P. 1991. Structural control of flank
volcanism in Continental Rifts. Science 254, 839-42.
Feibel, C. S.. Brown, F. H. and McDougall, I. 1989. Stratigraphic
context of fossil hominids from the Orno group deposits: Northern
,
46
Geoff Bailey, Geoffrey King and lsubelle Manigketti
Turkana basin, Kenya and Ethiopia, American Journal of Physical
Anthropology 78, 595-622.
Foley, R. A. 1987. Another Unique Species: Patterns in Human
Evolutionaly Ecology. Harlow: Longman.
Foley, R. A. 1994. Speciation, extinction and climatic change in
hominid evolution. Journal of Human Evohtion 26, 27-89.
Foley, R. A. in press. Evolutionary geography of Pliocene African
hominids, in Bromage, T, and Schrenk, P. (eds.), African
Biogeography, Climaric Change and Early Horninid Evolution.
Oxford: Oxford trr~iversityPress.
Gamble, C. S. 1993. Timewalkers: the prshisroty of global colonization. Stroud Alan Sutton.
Goodall, J. 1968. The Bshaviour of Free-living Chimpanzees in the
Gombe S!ream Reserve. (Animal behaviour Monographs; vol.
1(3)), London: Bailliere, Tindall and Cassell.
Gowlett, J. A. J., Harris, J. W. K., Walton, D. and Wood, B. A. 1981.
Early archnenlogical sites, hominid remains and traces of fire
from Chesowanja, Kenya. Nulure 294, 12.5-9.
Harris, J. K. and Berbich, 2. 1978. Aspects of early Pleistocenc
hominid behaviour of east Lake Turkana, pp. 529-47 in Bishop,
W.W.(ed.), Geological Background to Fossil Man. Edinburgh:
Scottish Academic Press.
Hay, R.L. 1976. Geology ofthe Olduvai Gorge. Berkelcy: University
of California Press.
Issac, G. L. 1978a. The Olorgesailie formation: stratigraphy, tectonics
and thc palaeogeographic context of the Middle Pleistocene
archaeological sites, pp. 173-206 in Bishop, W.W. (ed.),
Geological Backgr~undto Fossil Man. Edinburgh: Scottish
Academic Press.
Issac, G. L. 1978b. The food-sharing behavior of protohuman
hominids. Scieniific American 238, 90-108.
Isaac, G. L. 1981. Stone Age visiting cards: approaches to the study
of early land-usepattcms, pp. 131-55 in Bodder, I., Ismc, G. and
Bammond, N. (eds.), Pattern of the Past. Cambridge: Cambridge
University Press.
Isaac, G. L. and Behrensmeycr, A. K. 1997. Geological context and
palaeoenvironments, pp. 12-19 in Isaac, G. L and Isaac, B. (eds.),
Koobi Fora Research Project, Vol. S. Plio-Pleistocene Archaeology. Oxford: Clarendon Press.
King, G. C. P., Sturdy, D. A. and Bailey, G. N. 1994. Active tectonics,
complex topography and human survival strategics. Journal of
Geophysical Research 99(310), 20063-78.
Kummcr, H . 1968. Social Organization of the HamQdryas Baboon.
Chicago: University of Chicago Press.
Lovejoy, C. 0.1981. The origin of man. Science 211, 341-50.
Manighetti, I. 1993. Dynamigue des sysikmes extensifs en Afar.
Unpublished PhD thesis, Institut de Physique du Globe de Paris.
Manighetti, L, Tapponnier, P,,Courtillot, V., Gruszow, S. and Gillot
P. Y. 1997. Propagation of rifting along the Arabia-Somalia plate
boundary: the gulfs of Aden and Tadjoura Journal ofGeopirysico1
Research 102, 268 1-7 10.
Manigheni, L, Tapponnier, P-, Gillot, P. Y, Jacques, E., Courtiliot,
V., Armijo, R., Ruegg, J. C. and King, G. 1998. hpagalion of
rifting along the Arabia-Somalia plate boundary into Afar. Journal
of Geophysicnl Research 103, 4947-74.
. deMenocal, P. B.
Partridge, T. C., Bond, G. C., flartnady, C. .lH.,
and Ruddiman, W. F. 1995a. Climatic effects of late Neogene
tectonisrn and volcanism, pp. 8-23 in Vrba, E. S., Denton, G. H.,
Partridge, T. C. and Burckle, L. H. (eds.), Paleoclimate and
Evolution, with emphasis on Human Origins. New Haven, CT:
Yale University Press.
Parhidge, T. C., Wood, B. A. and deMenoca1, P,',B. 1995b. The
influence of global climatic change and regional uplift on largemammalian evolution in east and southern Africa, pp. 331-55 in
Vrba, E. S., Denton, G. H., Partridge, T. C. and Burckle, L. B.
(eds.), Paleoclimata and Evoluiton, with*~mphasison Human
Origins. Near Haven, CT:Yale University Press.
Pianka, E. R. 1978. Evoluiionaty Ecology. 2nd ed$ion. New York:
Harper & Row.
?3
Rapp, G. and Vondra, C. F., (eds.) 1981. Hominid Siies: their geologic
settings. American Association for the Advancement of Science,
Selected symposium 63. Boulder, Colorado: Westview Press.
Ruddiman, W. F. and Raymo, M. E. 1988. Northern Hemisphere
climate regimes during the past 3Ma: possible tectonic connections, pp. 1-20 in Shackleton, N. J., West, R. W. and Bowen, D.
Q. (eds.), The Past Three Million Years: evolution of climntic
variability in the North Atlantic region. London: Royal Society.
Runnels, C. and Van Andel, Tj. H. 1893. A handue from Kokkinopilos, Epirus, and its implications for the Paleolithic of
Greece. Jortrnnl of Field Archaeology 20, 91-103.
Shackleton, R, S. 1978. in Isaac, G. L. The Olorgesailie formation:
stratigraphy, tectonics and the palaeogeographic context of the
Middie Pleistocene archaeological sites, pp. 173-206 in Bishop,
W. W. (ed.), Geological Background to Fossil Man. Edinburgh:
Scottish Academic Press.
Stein, R. S., Briole, P,,Ruegg, I. C., Tapponnier, P. and Gasse, F.
1991. Contemporary, Holocene and Quaternary deformation of
the Asal Rift, Djibouti: implications for the mechanics of slow
spreading ridges. JournaI of Geophysical Research 96, 21789806.
Taicb, M., Johanson, D. C., Coppcns, 5'. and Aronson, J. L. 1976.
Geological md palaeontological background of the Hadar hominid
site, Afar, Ethiopia. Natiire 260, 289-93.
Taieb, M. and Tierceiin, J.-J. 1979. Sddimentation Plioc&ne et
PaltSaenvironments de Rift: exemple dc la Formation BHorninidBs
d'Hadar 5, Afar, Ethiopie. Bulletin de la Sociitk Gblogique dc
Fmnce 21, 243-53.
Tapponnier P,, Armijo R., Manighetti I. and Courtillot, V. 1990.
Bookshelf faulting and horizontal block rotation between
overlapping rifts in southern Afar. Geophysical Research Letters
17(1), 1-4.
Treves, A. and L. Naughton-Treves, 1999, Risk and opportunity for
humans coexisting with large carnivores. Journal of Human
Evolution 36, 275-82,
Vrba. E. 1996. Paleoclimate and Neogene Evolution. New Haven,
CT: Yale University Press.
Vrba, E. S., Denton, G. B., Partridge, T. C. and Burckle, L. H. (eds)
1995. Paleoclimate and Evolution, with emphasis on Human
Origins. New Haven, C T Yaie University Press.
Wheeler, P .E.1984. The evolution of bipedality and loss of functional
body hair in hominids. Journal of Human Evolurion 13, 91-8.
Wheeler, P. E. 1985. The loss of functional body hair in man: the
influence OF thermal environment, body form and bipedality.
Journal of Human Evolulion 14, 23-8.
Winder, N. 1997. Dynamic modelling of an extinct ecosystem: refugia,
resilience and the overkill hypothesis in Pnlacolithic Epirus, pp.
625-36 in Bailey, G. N. (d.)
Klithi:
,
Palaeoli!hic Settlernenr
and Qwaiernary Landscapes in Northwest Greece. Vol. 2: KIithi
in its Local and Regional Setting. Cambridge: McDonald Institute
for Archaeological Research.