Potential hominin plant foods in northern Tanzania: semi-arid savannas versus
savanna chimpanzee sitesq
Sandi R. Copelanda,b
aDepartment of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Deutscher Platz 6, D-04103 Leipzig, Germany
bDepartment of Anthropology, University of Colorado at Boulder, UCB 233, Boulder, CO 80309, USA
a r t i c l e i n f o
Received 1 December 2007
Accepted 29 June 2009
a b s t r a c t
Savanna chimpanzees are useful as referential models for early hominins, and here potential differences
between chimpanzee and early hominin ecology is the focus. Whereas chimpanzees inhabit only
a handful of modern African savannas, there is evidence that early hominins occupied relatively more
open and arid savannas than those in which chimpanzees live. In order to help expand potential models
of early hominin palaeoecology beyond savanna chimpanzee-like scenarios, and to provide a basis for
future modeling and testing of actual hominin diets, this study compares the types of plant foods
available in modern semi-arid savannas of northern Tanzania to plant foods at savanna chimpanzee sites.
The semi-arid savannas are not occupied by modern chimpanzees, but are potentially similar to envi-
ronments occupied by some early hominins. Compared to savanna chimpanzee habitats, the northern
Tanzania semi-arid savanna has a lower density and fewer species of trees that produce fleshy fruits.
Additionally, the most abundant potential hominin plant foods are seasonally available Acacia seeds/pods
and flowers, grass seeds, and the underground parts of marsh plants, as evidenced by vegetation surveys
and by studies of the diets of baboons that forage in similar areas. The information from this study should
be useful for framing hypotheses about hominin diets for sites with palaeoenvironmental contexts
similar to those of the northern Tanzania semi-arid savannas and for contextualising tests of actual
hominin diets (e.g., those based on dental microwear or isotopes).
? 2009 Elsevier Ltd. All rights reserved.
Savanna chimpanzees are useful as referential models for early
hominins and have helped to create testable hypotheses about
hominin foraging behaviour based largely on proposed similarities
in chimpanzee and early hominin physiology, behaviour, and
environments (Moore, 1996). It can also be productive to investi-
gate the potential differences between ecological aspects of early
hominins and modern chimpanzees, as these differences may
include features that make hominins unique among the apes.
Throughout this paper the term savanna is used in its most
general sense: as a juxtaposition to the other major biomes of
Africa, tropical forest and desert. Savanna is defined as an area with
seasonal rainfall and a grassy understory ranging from treeless
grassland to woodland. Chimpanzees inhabit forests and only
a handful of relatively moist, modern African savannas, and seem to
be excluded from more arid savannas possibly due to a lack of year-
round food sources, trees in which to nest, and/or drinking water
(Kano,1971; McGrew et al.,1988). In contrast, there is evidence that
by the late Pliocene at least some early hominins were able to
survive in relatively more open and arid savannas (e.g., Plummer
et al., 1999; Zazzo et al., 2000; Avery, 2001). Relatively open and
arid savannas are hypothesised to have different plant foods
available than do the moister savanna habitats of chimpanzee.
Thus, the nature of these differences is of particular interest for
distinguishing the ecological adaptations of modern chimpanzees
from those of any early hominins that many have lived in relatively
more open and arid savannas.
relatively open and arid African savannas that serve as analogues for
some hominin palaeoenvironments, and to compare that with food
availability at modern savanna chimpanzee sites. This information
if indeed they were able to survive in the types of open and arid
savannas being modeled, and how similar or different potential
hominin diets may have been from that of modern savanna chim-
panzees. The goal of this study is not to make conclusions about
qThis article is part of the ‘Palaeoanthropology Meets Primatology’ Special Issue.
E-mail address: firstname.lastname@example.org
Contents lists available at ScienceDirect
Journal of Human Evolution
journal homepage: www.elsevier.com/locate/jhevol
0047-2484/$ – see front matter ? 2009 Elsevier Ltd. All rights reserved.
Journal of Human Evolution 57 (2009) 365–378
Modern vegetation and plant foods have been used to model
palaeohabitats in different African contexts over the past several
decades (e.g., Carr, 1976; Peters and O’Brien, 1981; Sept, 1986;
Peters and Blumenschine, 1995; Copeland, 2007; Andrews and
Bamford, 2008). In each case, we face the problem that no modern
analogue will ever be exactly like an ancient setting in terms of the
plant species present. However, without the use of modern
analogues our understanding of ancient vegetation contexts will
remain rudimentary and virtually useless for interpreting early
hominin palaeoecology (Copeland, 2007). One potential solution is
to differentiate between the details and generalisations that arise
from modern analogue studies of vegetation. While data collection
in modern habitats is necessarily detailed and meticulous
(e.g., plant species lists, individual tree counts, and other abun-
dance measures), the results that are relevant to reconstructing
palaeoenvironments are the generalisations that follow. For
example, while the exact species list from a modern setting is
bound to differ from that of an ancient setting, it is more likely that
the plant families or genera will be consistent between the two.
While the list of fleshy fruits present at the species level probably
differs between the modern analogue and the ancient setting, it is
still likely that a trend such as an abundance of edible fleshy fruits
will hold when environmental characteristics such as climate and
soil type are similar in the present and the past.
Another relevant theoretical consideration when using modern
analogues to model potential hominin plant foods was outlined by
Peters and O’Brien (1981) as the difference between the ‘funda-
mental’ as opposed to the ‘realised’ plant food niche for hominins.
The ‘fundamental’ plant food niche refers to the ‘‘maximum
abstract plant-food niche of the early hominids constrained only by
physiological limitations’’ versus the ‘realised’ niche of plant foods
‘‘actually exploited under biotic constraints such as competition’’
(Peters and O’Brien, 1981: 127–128). The current study is an
attempt to document the ‘fundamental’ plant food niche for early
hominins in semi-arid East African savannas. This study does not
test the ‘realised’ niche or what hominins actually ate, as that must
be done by more direct methods such as dental microwear and
stable isotopes. The combination of contextual information
provided by a study of plant foods available (such as this study),
with direct dietary information provided by other studies, should
result in a more comprehensive understanding of likely hominin
plant food use at any given site.
In this study, plant foods that are available in six relatively moist
savanna habitats occupied by savanna chimpanzees are compared
to plant foods that are available in the relatively dry and open semi-
arid savannas of northern Tanzania. The types of plant foods
available in the northern Tanzania semi-arid savanna habitats are
described based on the results of vegetation surveys in Serengeti
National Park, Ngorongoro Crater, and Lake Manyara National Park.
The plant foods in savanna chimpanzee habitats are described
based on published results of primatological studies at savanna
chimpanzee sites. Some comparisons are also made with plant
foods available in forested habitats occupied by chimpanzees, and
with plant foods available in other (non-northern Tanzania) semi-
arid savannas inhabited by baboons. The comparison can be used to
determine how similar early hominin diets could have been to
those of extant savanna chimpanzees, and should help to pinpoint
the nature of the most likely differences.
Background to hominin environmental preferences
Most modern chimpanzees in the wild live in the dense
woodlands or tropical forests of Africa. While some modern
chimpanzees live in savannas, they generally do not live in areas
with less than about 900 mm annual rainfall, with the exception of
Ishasha with 750 mm, at which the chimpanzees inhabit mainly
the riverine forest (Sept, 1992). Chimpanzee fossils come from the
middle Pleistocene Kapthurin Formation in Kenya (McBrearty and
Jablonski, 2005). Although that location is outside the range of
modern chimpanzee distribution, the fauna found with the fossil
chimpanzee remains suggest that the habitat was a closed, semi-
arid environment (McBrearty and Jablonski, 2005). There is also no
historical evidence for chimpanzee occupation of more open
Fossil hominins are generally found in the context of other
savanna fauna, as opposed to in the context of forest or desert
fauna. Fossil hominins in Africa are often found in association with
baboons (Papio spp.), but never with other African apes. It may be
that a taphonomic bias prevents animals in forest habitats from
being preserved, as they have very little chance of becoming fos-
silised, while animals dying along a savanna lakeshore have a much
greater chance of quick burial and subsequent fossilisation. Thus,
we cannot rule out the possibility that some hominins lived in
tropical forests but that their remains have not been preserved. The
aim here, however, is not to determine the entire range, nor the
preferred habitat of any hominin species. Rather, it is to document
the plant food context of semi-arid East African savannas such that
if hominins can indeed be shown to have survived in such habitats,
then realistic assessments of their plant food diet can be made and
ultimately tested by direct means.
There is indisputable evidence that at least some early hominins
lived in savanna environments as opposed to tropical forests, and
there is very compelling evidence that at least some early hominins
lived in savannas that were more arid and open (less tree or shrub
cover) than the savannas inhabited by modern savanna chimpan-
zees. Stable carbon isotopes indicate that early hominins from
South Africa (Paranthropus robustus, Australopithecus africanus, and
Homo habilis) ate an average of 20–40% C4(grass-derived) foods,
which must have derived from savanna environments (Lee-Thorp
et al.,1994, 2000; Sponheimerand Lee-Thorp,1999; Van der Merwe
et al., 2003; Sponheimer et al., 2005). East African hominins
(Paranthropus boisei, Homo habilis) consumed from 23–81% C4
foods (Van der Merwe et al., 2008), indicating a dominance of
grass- or sedge-derivedfoods in theirdiet. In contrast, chimpanzees
feed on little to no C4foods even when they live in the savanna
environments of Fongoli (Senegal), Ishasha (Democratic Republic of
Congo), and Ugalla (Tanzania; Schoeninger et al.,1999; Sponheimer
et al., 2006). This evidence alone is not enough to determine
whether hominins lived in moist or more arid savannas, but it does
indicate that there was plentiful grass present, that they must have
lived in savannas, and that their diet was substantially different
than that of savanna chimpanzees.
The palaeoenvironmental evidence associated with some Plio-
cene and early Pleistocene hominins, including Ardipithecus
ramidus, Australopithecus bahrelghazali, Paranthropus boisei, Homo
habilis, and Homo erectus, indicates that they lived in semi-arid,
relatively open savanna habitats in Africa. Examples of fossil hom-
inin sites at the most arid/open end of the palaeoenvironmental
spectrum are described below.
In the palaeoenvironmental reconstructions of the early Plio-
cene at Gona, there is evidence that Ardipithecus ramidus (Sagantole
Formation, w5.2–3.9 Ma) lived in a habitat with extensive C4
grasses, as the carbon isotope values of the fossil fauna are similar
to extant herbivores living in bushland, thicket, and grassland
habitats (Levin et al., 2008). Although this may have been a moist as
opposed to a semi-arid or arid savanna, it indicates that grasses
were an important component of at least some hominin environ-
ments very early in the hominin lineage.
Stable isotope data from fauna associated with the late Pliocene
Australopithecus bahrelghazali in Chad (w3.5–3.0 Ma) indicates
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378 366
a dominantly grassland palaeoenvironment, as nearly all of the
fauna analysed show ‘‘a quasi-exclusive reliance on C4foodstuffs,’’
even among taxa that are not necessarily always grazers, such as
the Rhinocerotidae, Proboscidea, Suidae, and Hippopotamidae
(Zazzo et al., 2000: 305). Of the fauna analysed from the hominin-
associated unit, 22 of 25 animals had pure C4diets, 3 of 25 were
mixed feeders, and none consumed a pure C3diet (Zazzo et al.,
2000), indicating that the environment was mainly open savanna
and unlikely to be a moist savanna.
At Kanjera South, Kenya, in sites up to 2.2 Ma (Ditchfield et al.,
1999), soil carbon isotopes and grassland-adapted fauna indicate
that the habitat comprised more than 75% C4grasswith no evidence
of closed forest (Plummer et al., 1999). Within the KBS member of
the Koobi Fora Formation in northern Kenya, pollen and faunal
remains suggest that Homo erectus and Paranthropus boisei inhabi-
ted a dry, open area dominated by grassland, shrubland, or sub-
desertic steppe (Bonnefille,1984; Harris,1991; Reed,1997). A pollen
study from 1.5 Ma sediments in the Peninj site of northern Tanzania
suggests a very open, grassy environment (Dominguez-Rodrigo
et al., 2001). Paranthropus boisei from the Okote Member of the
Koobi Fora Formation is associated with wetlands and grasslands,
but not much woody cover as indicated by the lack of fauna with
arboreal locomotion and a high abundance of grazers (Reed, 1997).
At Olduvai Gorge, palaeoenvironmental evidence from bovids,
microfauna, pollen, and soil carbon isotopes in the Bed I and lower
Bed II sequence indicate the presence of a mosaic of habitats from
open grassland to woodland (e.g., Kappelman, 1984; Shipman and
Harris, 1988; Plummer and Bishop, 1994; Sikes, 1994; Fernandez-
Jalvo et al.,1998), all clearly within the savanna as opposed to forest
biome. While some of those time periods appear to be more
consistent with a moist as opposed to a semi-arid climate, during
the deposition of the Lemuta Member of Bed II (1.62–1.67 Ma) and
stratigraphically equivalent geological facies such as the lower
augitic sandstone containing OH 13 (Homo habilis), the climate was
warmer and drier than during Bed I times. Soil carbonate isotopes
suggest that during this period the vegetation was up to 90% C4
grass cover and that the mean annual temperature maximum was
22–25?C (Hay, 1976; Kappelman, 1984; Cerling and Hay, 1986).
Early Homo in Sterkfontein Member 5 also appears to have
lived in an ‘‘open or wooded grassland or plains region’’ where the
ecological traits of the fossil fauna include 0% arboreality and 0%
frugivory (Reed, 1997), indicating very little woody cover. Based
on a study of micromammals, Avery (2001) also concluded that
the palaeoenvironment of the Sterkfontein Valley between about
2 and 1 Ma was quite open, even more arid than present, with
310–550 mm annual precipitation, and with higher temperatures
than at present.
Materials and methods
Definition of edible
Because the digestive capabilitiesofearly homininsareunknown
and may have differed between hominin species, a broad definition
for potentially edible plants for hominins is used in this paper. Plant
species that are known to be consumed in the wild bychimpanzees,
baboons, or humans are considered edible for hominins (Peters and
O’Brien, 1981; Peters et al., 1992). Baboons appear to have broader
digestive capabilities than African apes (Chivers and Hladik, 1984),
and therefore likely hominins, but since baboons are the only large-
bodied primates that live in the northern Tanzanian study areas, as
well as many other savannas in Africa, they serve as an important
guide to the potential edible wild plant foods available in those
habitats. It is possible that some plants consumed by chimpanzees
may not have been edible to some hominins; for example hominins
may have been incapable of digesting the large amounts of fibre
typically consumed by chimpanzees (e.g., Conklin-Brittain et al.,
2002). Thus, the set of plants defined here as ‘‘potentially edible’’ for
hominins is relatively generous and probably overestimates the
number of plant species that were actually edible to any given
For each of the chimpanzee sites discussed here, the list of
potential plant foods for hominins is based on those plant foods
recorded to be eaten by the chimpanzees that live there. These
recordings are based either on direct observations of consumption
or indirect observation via food/processing remains or faecal
samples (see references in Table 5). At Assirik and Gombe, data are
also available about the diets of local baboons, which allows for an
expanded list of potential plant foods for hominins at those two
sites. Comprehensive data are not available at any of the chim-
panzee sites for plant species edible to humans, aside from anec-
dotal mention of humans consuming a few of the same items as
chimpanzees. The list of potential plant foods for hominins at the
chimpanzee sites would almost certainly contain more items if
a vegetation survey of plant foods edible to chimpanzees, baboons,
and humans were conducted.
The best documented and published savanna chimpanzee sites
are summarised in Table 1 and described in more detail below. For
comparison, Table 1 also includes the wetter and more forested
chimpanzee localities of Gombe, Kibale, and Mahale, and the semi-
arid northern Tanzania habitats that were examined specifically for
potential hominin plant foods in this study.
The Mount Assirik and Fongoli savanna chimpanzee study areas
are both in eastern Senegal and are only 45 km apart. They receive
from 900–1100 mm annual rainfall, which falls in a four to five
month wet season leaving a long, hot dry season of seven to eight
months (McGrew et al., 1981; Pruetz, 2006). At Assirik, the vege-
tation is 3% gallery forest, 37% drought-deciduous lowland wood-
land, 27% grassland, 5% bamboothicket, and 28% plateau, an area on
which little vegetation can thrive (McGrew et al., 1981; Hunt and
McGrew, 2002). Common trees in the woodland at Assirik are
Hexolobus monopetalus, Pterocarpus erinaceus, and Afzelia africana,
and common species in the gallery forest are Saba senegalinsis,
Combretum tormentosum, and Oncoba spinosa (Hunt and McGrew,
2002). The gallery forest provides essential food and water
resources to the chimpanzees in that 29% of plant foods derive from
it even though it constitutes only 3% of the area (McGrew et al.,
1988). A study of chimpanzee nests showed that 32% occurred in
gallery forest, 56% in woodland, and 12% in grassland (Baldwin
et al., 1982). Thus, although the chimpanzees make dispropor-
tionate use of the riverine forests, they utilise the woodland and
grassland habitats as well.
The vegetation at Fongoli is 2% gallery forest, 46% woodland,16%
grassland, 20% plateau,12% bamboowoodland, and 4% horticultural
fields (Pruetz, 2006). Common trees are in the genera Zizyphus,
Combretum, Pterocarpus, and Piliostigma (Pruetz, 2006). Although
the gallery forest provides important food and shelter for Fongoli
chimpanzees, virtually all fruits eaten by the chimpanzees can be
found in the woodland, grassland, or plateau habitats (Pruetz,
2006). Despite some use of the relatively open portions of their
environments, however, analysis of the stable carbon isotopes of
hair samples indicates that chimpanzees at Fongoli consume few or
no C4resources (tropical grasses, sedges, or animals that eat those
foods; Sponheimer et al., 2006).
The savanna chimpanzee localities in western Tanzania are
different in vegetation structure and composition from the West
African savanna chimpanzee sites, as the westernTanzania sites are
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378 367
dominated by miombo woodlands, characterised by evenly-spaced
Brachystegia, Julbernardia, and Isoberlinia trees with a grassy
herbaceous layer (Suzuki, 1969; Kano, 1971; Moore, 1992). The
western Tanzania savanna chimpanzee localities discussed here,
Kasakati and Ugalla, both receive between 950–1050 mm annual
rainfall and have a distinct dry season that lasts for 4–6 months
(Suzuki,1969; McGrew et al.,1981; Moore,1992,1994; Hernandez-
Aguilar et al., 2007). Kasakati and Ugalla differ from one another in
terms of the size and importance of gallery forests as described
below (Suzuki,1969; Moore,1992; Hernandez-Aguilar et al., 2007).
According to Suzuki (1969), the landscape of Kasakati is 59%
woodland, 16% ‘‘savanna’’ (he uses the term ‘‘savanna’’ to mean
structurally intermediate between woodland and grassland), 16%
grassland, and 10% gallery forest lining a perennial river. The gallery
forest includes trees of the genera Albizzia and Cordia and provides
71% of the plant species eaten by the chimpanzees in the study area
Summary climate and vegetation characteristics of 1) savanna chimpanzee sites, 2) some of the well-documented wetter/more forested chimpanzee sites, and 3) the
comparative semi-arid northern Tanzanian savannas which are not occupied by chimpanzees.
Study area Rainfall and seasonality Vegetation descriptionReferences
Savanna chimpanzee sites:
954 mm/yr (1976–1979);
7 month dry season
Heterogeneous Sudanian savanna
with gallery forest (3%),
woodland (37%) grassland (27%),
bamboo thicket (5%),
and plateau (28%)
McGrew et al., 1981, 1988
8 month dry season
savanna with gallery forest
(2%), woodland (46%),
bamboo woodland (12%),
plateau (20%), and horticultural
977 mm/yr (at Kigoma,
avg. over 29 yrs);
4 month dry season
Miombo woodland (59%)
with a mosaic of riverine
forests (10%), ‘‘savanna’’
(16%; intermediate between
woodland and grassland),
and grassland (16%)
Suzuki, 1969; McGrew et al., 1981; Moore, 1992
5 month dry season
Miombo woodland dominated
(86%) with 1–2% riverine forests
Moore, 1992, 1994; Hernandez-Aguilar et al., 2007
1206 mm/yr; two dry
seasons per year:
2 months and 3 months
Dry Combretum savanna and
Borassus palm savanna; rivers
lined with woodland grading
to riverine forest
Hunt and McGrew, 2002
Dem. Rep. Congo
7 month dry season
Gallery forest about 1 km wide
surrounded by grassland
Wetter/more forested chimpanzee sites:
1495 mm/yr (1968–1974)Mosaic of riverine forest,
miombo woodland, and
grassland with scattered trees
Wrangham, 1977; McGrew et al., 1981; Goodall, 1986;
Collins and McGrew, 1988
1400 mm/yr (Bilenge);
1750 mm/yr (Kasoje)
Mostly primary forest, also areas
of miombo woodland, secondary
bamboo-dominated areas, and
Nishida and Uehara, 1983; Collins and McGrew, 1988
1570–1700 mm/yrMosaic of mid-altitude
moist forest, secondary forest,
tall grassland, swamps,
and softwood plantations
Isabirye-Basuta, 1989; Wrangham et al., 1991, 1993
1780–1900 mm/yr Moist semi-deciduous
Reynolds et al., 1998; Newton-Fisher, 1999
Northern Tanzania semi-arid savannas:
? 5 month dry season
Short grassland with underlying
calcrete, long grassland with
deeper soils, Acacia dominated
bushland and bush grassland.
Anderson and Talbot, 1965; Norton-Griffiths et al., 1975;
? 5 month dry season
Mosaic of bush grassland,
Acacia tortilis bushland,
and Trichilia groundwater forest.
Loth and Prins, 1986; Copeland, 2007
? 5 month dry season
Marshy wetlands and
Deocampo, 2004; Copeland, 2007
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378 368
(Suzuki, 1969). Twenty-four percent of the plant species eaten by
chimpanzees come from miombo woodland, which is dominated
by Julbernardia and Isoberlinia trees, and the remaining 4% come
from ‘‘savanna,’’ which is characterised by the trees Combretum,
Diplorhynchus, and Brachystegia longifolia (Suzuki, 1969). Although
chimpanzees at Kasakati depended on fruits from the riverine
forest for foods during the wet season, theyalso depended on fruits
or ‘‘hard seeds’’ of Caesalpiniaceae and Papilionaceae from the
miombo woodland during the dry season (Suzuki, 1969; Kano,
1971). Suzuki (1969) found that 53% of nests were in miombo
woodland and 47% were in riverine forest, confirming that the
chimpanzees at Kasakati heavily utilise both of these habitats.
The Ugalla study area is approximately 60 km east of Kasakati
and is composed of about 86% miombo woodland dominated by
Brachystegia and Julbernardia trees with open woodland canopy
and a grassy understory (Moore, 1992, 1994; Hernandez-Aguilar
et al., 2007). This woodland is disrupted bywidelyscattered, mostly
ephemeral rivers with thin strips of evergreen forest comprising
only 1–2% of the area (Moore, 1994; Hernandez-Aguilar et al.,
2007). The remaining 12% of the area includes Combretum/Termi-
nalia communities on plateau tops and local concentrations of
Acacia scrub in valley bottoms (Moore, 1994). Most nests at Ugalla
are found in woodland rather than forests (Itani, 1979; Nishida,
1989; Hernandez-Aguilar, 2009). An isotopic study of carbon and
nitrogen stable isotope ratios in the hair of Ugalla chimpanzees
suggests that they feed mainly in the miombo woodland as
opposed to in forest and ingest considerable quantities of legumi-
nous plants, a category that includes the dominant trees in the
miombo woodland (Schoeninger et al., 1999).
The Semliki chimpanzee study area in western Uganda is
dominated by an undulating plain, flanked to the east by a tall rift
escarpment (all descriptions based on Hunt and McGrew, 2002).
Semliki is hot and humid and receives 1206 mm annual rainfall,
which comes in a long rainy season (August–December) and
a short rainy season (March–April). The vegetation is predomi-
nantly dry Combretum savanna and Borassus palm savanna,
crosscut by water courses with 50–250 m wide strips of riverine
forest. The home ranges of what appear to be four distinctive
groups of chimpanzees each focus along a major river and all nests
at Semliki were found in riverine forest. The Mugiri River, which
has been a focus of chimpanzee research at Semliki, abuts the
eastern escarpment and supports flora grading from woodland to
tall riverine forest with trees such as Celtis, Albizia, and Cynometra.
Fruiting trees used heavily by the Mugiri chimpanzees are
distinctly distributed. For example, Beilschmiedia, Tamarindus, and
Phoenix are located on the rift floor in or within metres of the
Mugiri riverine forest, Securinega trees tend to grow on the lower
escarpment slopes, and edible fruit-bearing Grewia trees grow only
on the escarpment in open woodland and bushland. In general, the
chimpanzees at Semliki, like those at Assirik, spend much of their
time in gallery forests despite the fact that forest is only a fraction
of the local habitat.
A small population of chimpanzees lives along the Ishasha River
in the Democratic Republic of Congo (formerly Zaire). This small,
perennial river is lined by an approximately 1 km wide lush gallery
forest dominated by trees such as Cynometra alexandrii, Pterygota
mildebraedii, and Chrysophyllum albidum (Rowell,1966; Sept,1992).
There is a sharp boundary between the gallery forest and the open
grasslands and thickets of the surrounding floodplain (Sept, 1992).
This area receives annual rainfall of only 750 mm per year, with an
average of seven dry months a year (Sept, 1992). At Ishasha,
chimpanzee nests are all located in the riverine forest as opposed to
the surrounding grasslands and thickets (Sept,1992). Although the
chimpanzees are not habituated for direct observation, a study
of carbon stable isotope ratios in chimpanzee hair from Ishasha
indicates that the chimpanzees feed primarily within the forest as
opposed to in more open areas (Schoeninger et al., 1999).
Savanna chimpanzee localities are clearly variable in terms of
their vegetation structure, composition, climate, and the ways in
which the habitats are exploited. In some places, such as Ishasha,
chimpanzees are in effect living in a forested environment that is
surrounded byopenwoodland or grassland. Atother places, such as
Fongoli and Ugalla, the chimpanzees seem to thrive in woodland
habitats and obtain most of their resources there. In Assirik, Kasa-
kati, and Semliki, they rely on the combination of resources from
both riverine forest and woodland or other open canopy habitats.
At each of these sites, however, the chimpanzees depend on fruits
from trees and year-round access to water and nesting trees.
Vegetation sampling methods
This study focuses on the relatively open, semi-arid savannas of
chimpanzees,but it fallswithin
oenvironmental parameters, such as temperature range, vegetation
cover, and plant taxa (as indicated by fossil pollen at the family and
sites; e.g., Bonnefille, 1995). This area is characterised by young
volcanic soils, seasonal rainfall in the range of 500–800 mm per year,
a dominance of grassland and thorn trees such as Acacia and Com-
miphora, and mostly ephemeral but occasional perennial streams.
Potentially edible plant foods for hominins in the northern
Tanzania study areas were determined on the basis of vegetation
surveys (Copeland, 2004, 2007). The sampling strategy, described
in more detail below, involved: 1) identifying habitats of interest
within the northern Tanzania region, 2) defining the boundaries of
local structurally-defined habitats, 3) obtaining plant species lists
and, in some cases, relative abundance using a series of stratified
random plots within each habitat, and 4) identifying potentially
edible plant foods for hominins using mainly the compilation of
Edible wild plants of sub-Saharan Africa by Peters et al. (1992).
Field studies of vegetation and plant food availability were con-
ducted in the semi-arid savannas of northern Tanzania in Serengeti
National Park, Lake Manyara National Park, and the Ngorongoro
Crater Conservation Area (Fig. 1). These areas are not inhabited by
chimpanzees nor are they known to have been historically, but
baboons (Papio anubis) and vervet monkeys (Cercopithecus aethiops)
are common. Habitats that are similar to the modern study areas in
terms of soils,rainfall,seasonality, and vegetation extendthroughout
much of northern Tanzania and Kenya.
The semi-arid savannas of northern Tanzania typically consist of
a mosaic of habitats due to the uneven distribution on a regional
scale, but localised similarity in topography, soil type, hydrology,
and climate (Webster and Beckett, 1970; Gerresheim, 1974; Belsky,
1990; Coughenour and Ellis, 1993). In this study, each habitat
within the mosaic was defined on the ground by its relatively
homogeneous vegetation composition and structure. In the case of
Manyara and Serengeti, this was guided by previously published
local habitat maps (Herlocker,1975; Loth and Prins,1986). In order
tomaintain consistency,habitat structural
assigned following the relatively quantitative definitions set out by
Pratt and Gwynne (1977, Table 2).
The local habitats of interest are of inherently different sizes.
Riverine woodlands ranged in width from about 50–200 m and the
length along which each river was sampled varied from0.8–5.0 km.
The grassland, bush grassland, and bushland interfluves of Seren-
geti and Manyara that were sampled are relatively large, varying in
size from 1?1 km to 2?5 km. The wetlands of Ngorongoro ranged
from the smallest at about 15?30 m to the massive Ngoitokitok/
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378369
Gorigor Marsh, which can cover up to 7?10 km in its entire extent,
but which was sampled around the accessible edges in areas of
about 50?50 m each.
At the very small marshes, all species could be identified and
relative cover abundance of species was estimated visually. In other
habitats, plants were sampled with a series of plots. A 50 ?50 m2
plot was used to sample trees (taller than 5 m) with nested plots of
50?10 m for shrubs and 5?2 m for herbaceous plants. For woody
plants, the height, crown diametre, and count of individuals was
recorded, and top cover area of all woody plants was estimated for
each plot. For herbaceous species, presence/absence of species
were recorded in each plot. With the exception of the small
wetlands, at which all plants were measured, the proportion of
vegetation that was sampled within each habitat of interest in this
study ranged from one to eight percent, based upon the sum of the
plot sizes divided by the estimated size of the habitat.
Plant species were identified in the field via reference collec-
tions at Ndala Research Camp in Manyara and at Seronera Research
Institute in Serengeti, keyed out with reference books (Haines and
Lye, 1983; Blundell, 1987; Ibrahim and Kabuye, 1987; Coates
Palgrave, 1993; Beentje, 1994), or brought to the National
Herbarium in Arusha for identification. With the addition of a few
other references as noted throughout the text, the majority of
potentially edible plant species for hominins were identified by
their listing in Peters et al. (1992).
Description of the northern Tanzania study areas
and annually variable rainfall that totals about 500 to 800 mm per
year in the study areas. In general, the rainy season lasts from about
November to May, sometimes interrupted by a short dry season in
January and February, and sometimes the ‘‘short rains’’ of November
and December fail altogether (Norton-Griffiths et al.,1975). The long
dry season from June to October (five months) is consistent and at
Fig. 1. Map showing the semi-arid savanna study localities in northern Tanzania: the central/eastern portion of Serengeti National Park, Ngorongoro Crater, and the northern
portion of Lake Manyara National Park. Shaded areas represent higher elevation and dashed lines are rift escarpments. Numbers represent sampling sites (see Copeland, 2007).
Maps modified from Sinclair (1979), Loth and Prins (1986), and the Frankfurt Zoological Society (1971).
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378370
times extends for longer when the short rains fail (Prins and Loth,
Serengeti. The area of focus within Serengeti National Park was
the central/eastern portion as shown in Figure 1. The short grass
plains, which are in the easternmost portion of the park where the
elevation is about 1730 m, were characterised by two study areas
representing two different habitats near the Barafu Valley. One
habitat studied was the open grassland itself with virtually no trees
or shrubs, while the other habitat was the bush grassland along an
ephemeral creek with widely scattered Acacia tortilis and Commi-
phora trees and shrubs.
The vegetation of the western area of the Serengeti Plains was
characterised by plots in four different study areas representing
two habitat types. Two study area habitats were the open ‘‘long
grassland’’ interfluves, one between the Seronera and Wandamu
rivers and the other west of the Nyamara River. In these areas the
elevation is about 1550 m, and tall grass grows on deep soils that
lack the underlying calcrete that is present in the easternmost
Serengeti Plains (Anderson and Talbot,1965). Two other study area
habitats were the riverine belts along the ephemeral streams that
cross the long grasslands, one along the Seronera River and one
along the Nyamara River, which are conspicuously lined with tall
Acacia xanthophloea trees and are classified as bushland (Pratt and
Additional study area habitats in the Serengeti were those along
two ephemeral streams within the Serengeti Woodlands (Sinclair,
1979), the Mbalageti River and the Sangare River. These bushland
habitats, also at elevations around 1550 m, were lined with
a slightly denser stand of trees (predominantly Acacia xanthophloea
and Acacia kirkii) than the Serengeti Plain rivers. These tall-tree-
lined rivers were surrounded by interfluves of open woodland.
Lake Manyara. The vegetation of the northern portion of Lake
Manyara National Park (Fig. 1) includes a large groundwater forest
in the north, an Acaciatortilisbushland covering the largelacustrine
terrace further south below and adjacent to the rift escarpment,
open grasslands and bush grasslands along the lacustrine plain, and
rivers which crosscut all of these habitats, forming belts of riverine
vegetationwith various compositions and structures. Lake Manyara
receives approximately 650 mm annual rainfall, and like Serengeti,
the dry season lasts for a minimum of five months from June to
October, and the amount and duration of rainfall varies widely from
year to year (Loth and Prins, 1986; Prins and Loth, 1988). The
elevation of the lake is about 960 m, and the habitats studied here
are within 50 m above that elevation (Loth and Prins, 1986). The
adjacent rift escarpment rises to approximately 1200–1300 m, but
areas on top of the escarpment were not examined in this study.
Ten local habitats were sampled within Lake Manyara National
Park. The groundwater forest was sampled in interfluve and riverine
habitats, both of which are classified as forest with a tall closed
canopy of trees (including Trichilia emetica, Tabernaemontana
ventricosa, and Ficus sycomorus) and a sparse herbaceous layer with
verylittle grass.Manyara’s Acaciatortilis-dominatedareaadjacentto
the rift escarpment was sampled in fourdifferentlocal habitats. Two
were riverine habitats, along the Msasa River and along the Ndilana
River, and two were interfluves, the Ndilana-Msasa interfluve and
the Ndala-Chemchem interfluve. All of these areas can be classified
as bushland due to slightly more than 20% woody cover composed
of a mixture of trees and shrubs and a grassy understory (Pratt and
Four habitats near the shore of Lake Manyara were sampled and
some differed from one another in vegetation structure. The Ndala
Lake flat and the Msasa Lake flat were interfluvial lacustrine plain
areas characterised by bush grassland (grassland with scattered
treesand shrubs comprising 2–20%cover;Pratt and Gwynne,1977).
Two riverine habitats were also measured near the lakeshore. One
was the small but perennial Mkindu River crossing the lake flat that
was lined with a tall, dense stand of Acacia xanthophloea trees,
forming a very narrow band (about 50 m wide) of forest. The other
riverine habitat sampled near the lakeshore was along the wide but
ephemeral Msasa River, which crosses the lake flat and was lined
with bushland composed of a broad assortment of tree and shrub
Ngorongoro Crater. Situated within the Crater Highlands of
northern Tanzania, Ngorongoro Crater is a grassland-dominated
caldera with scattered wetlands, small rivers, and a few wooded
areas. The floor of the caldera has an altitude of 1737 m (Hay,1976).
Rainfall averages 541 mm per year (Deocampo, 2004) with
a minimum five-month dry season, but the amount of rain and the
length of the dry season are variable from year to year.
This study focused on 13 wetland habitats around the Crater
floor. Most were localised, natural, saline-alkaline marshes char-
acterised by sedges such as Cyperus immensus, Cyperus laevigatus,
and Scirpus inclinata. The large Ngoitokitok/Gorigor wetland was
sampled in two localities along its eastern, perennially wet side,
and this relatively fresh water area was characterised by a large
stand of cattail (Typha latifolia) and papyrus (Cyperus papyrus),
although it also supported many other marsh plants. This large
Ngoitokitok/Gorigor wetland was also sampled in three localities
along its northern border, and those distal areas supported patchy
marshes similar in structure and composition to many of the small
localised marshes elsewhere in the Crater.
The first part of this section describes the potential plant foods
for hominins that were found in the vegetation survey conducted
for this study in the semi-arid savannas of northern Tanzania. The
second part of the section provides a comparison of the northern
Tanzania plant foods with plant foods in previously documented
savanna chimpanzee habitats.
Potential hominin plant foods in the northern Tanzania semi-arid
As described above, for the purpose of this analysis the three
major study areas in northern TanzaniadSerengeti, Manyara, and
Ngorongorodwere subdivided into local structurally defined
Definitions of vegetation structural categories (habitats) following Pratt and
ForestClosed stand of trees of one or more stories,
with an interlaced canopy. 7 to 40 m height. Ground cover
dominated by herbs and shrubs.
Trees up to 20 m in height have an open or continuous,
but not thickly interlaced canopy, and canopy cover >20%. Shrubs,
if present, constitute less than 1/10th of the canopy cover. Grasses
and other herbs dominate the ground cover.
Shrubs, usually not more than 6 m in height,
with a canopy cover >20%. Trees, if present, constitute
less than 1/10th of the canopy cover.
Trees and shrubs both present. May be dominated by shrubs,
but trees are always conspicuous. Tree þshrub cover >20%.
Grassland with scattered or grouped trees, the trees
always conspicuous, with cover area from 2 to 20%.
Grassland with scattered or grouped shrubs, the shrubs
always conspicuous, with cover area from 2 to 20%.
Grassland with scattered or grouped trees and shrubs, both
always conspicuous, but tree þshrub cover from 2 to 20%.
Dominated by grasses and other herbs. Treeþshrub cover <2%.
Herbaceous swamp with permanent or ephemeral wetland.
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378 371
habitats (Pratt and Gwynne, 1977). Serengeti included grassland,
bush grassland, and bushland habitats; Manyara included bush
grassland, bushland, and forest habitats; and Ngorongoro included
grassland and marsh habitats.
Plant species that are defined as potentially edible to hominins
(because they are known to be eaten by chimpanzees, baboons, or
humans) within each structurally-defined habitat type of the
northern Tanzania study areas comprise between 41–53% of all
plant species identified in this vegetation survey. Specifically, the
mean proportion of edible plant species relative to all plant species
in marsh, grassland, bush grassland, bushland, and foresthabitats is
42%, 41%, 50%, 53%, and 49%, respectively.
In order to help clarify major differences in the types of plant
foods available in the specific northern Tanzania habitats, the data
are further broken down by plant life form: trees, shrubs, grasses,
and forbs (non-grass herbaceous plants, including marsh plants) in
Figure 2, which summarises the data presented in Tables 3 and 4.
Of all the trees encountered in the vegetation surveystudy plots,
those with potentially edible parts comprised 98% of total tree
cover in bush grassland, 90% in bushland, and 88% in forests,
whereas the marsh and grassland habitats that were sampled had
no trees. Trees in the forest have a higher proportion of edible fruits
than do trees in bushland or bush grassland habitats (see the
‘‘Trees’’ column of Fig. 2). The dominant tree species in Manyara’s
groundwater forest was Trichilia emetica, which along with other
common species such as Croton macrostachyus and Ficus sycomorus,
has potentially edible fruits (Peters et al.,1992). The most common
trees in the bushland and bush grassland habitats were species of
Acacia. Acacia tortilis was dominant in bushland and bush grassland
habitats at Manyara and the bush grassland-lined Barafu River in
the eastern short grass plains of the Serengeti. Acacia xanthophloea
was the most common tree along other bushland-lined rivers in the
Serengeti. Compared to the forest trees, those in bushland have
a higher proportion of edible seeds/pods and flowers rather than
edible fruits, due in large part to the Acacias. Other less common
bushland trees were Acacia robusta and Albizia harveyi.
Overall, there are more shrub species (n¼129) than tree species
(n¼25) in the northern Tanzania savanna habitats (excluding
marsh and grassland habitats which have no shrubs or trees). The
shrubs in bushland habitats have a higher proportion of edible
fruits and edible roots than shrubs in the other habitats (see the
‘‘Shrubs’’ column of Fig. 2). In the bushland habitats at Lake
Manyara, for example, shrubs with potentially edible fruits include
Salvadora persica, Maerua triphylla, Cordia monoica, Cordia sinensis,
Acalypha fruticosa, and Capparis tomentosa. Shrubs with roots in the
bushland habitats that are known to be eaten raw by humans
include Cordia sinensis and Lannea triphylla (Peters et al., 1992).
Forbs in the marsh habitats are clearly distinguished from the
other growth forms in that a greater proportion of species have
edible underground parts (see the ‘‘Forbs’’ column of Fig. 2). Marsh
plants with potentially edible underground parts include the
sedges Cyperus immensus, C. papyrus, C. rotundus, and C. laevigatus,
which are known to be C4 plants (Peters and Vogel, 2005),
C. usitatus, and the cattail Typha latifolia, a C3plant. Forbs in habitats
other than marsh tend to have a higher proportion of species with
edible leaves compared to the other growth forms.
Like forbs, grass species in marsh habitats have a higher
proportion of edible underground parts (see the ‘‘Grass’’ column of
Fig. 2). Also, thepith of the large marsh grass Phragmites mauritianus
is edible. Grasses in other habitats have relatively higher propor-
tions of edible seeds compared to plants of other growth forms,
except in theforest, which hasvery little grassin anycase because of
the shaded understory.
Comparisons between savanna chimpanzee and semi-arid savanna
plant food availability
Table 5 summarises the number of plant parts recorded to be
consumedbythe fourdifferent populationsofsavanna chimpanzees
for which data are available (Assirik, Fongoli, Kasakati, and Semliki),
aswellasforchimpanzees in thewetterstudyareas(Kibale, Mahale,
and Gombe; see references in Table 5). Figure 3 provides a graphic
representation comparing the number of species with fruits, seeds/
pods, or underground parts known to be edible to chimpanzees,
baboons, or humans at the northern Tanzania semi-arid savannas,
the savanna chimpanzee sites, and the forest chimpanzee sites. Data
Fig 2. Summary of data on potentially edible plants for hominins gathered from the vegetation survey in the semi-arid northern Tanzania study areas. Tree and shrub graphs show
the proportion of cover area of plants with edible parts. Forb and grass graphs show the proportion of the number of species with edible parts.
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378 372
on plant foods recorded to be eaten by chimpanzees were available
for all of these sites (the top portion of the figure labeled ‘‘Chim-
panzee foods’’), but of the sites of interest here, data on baboon
foods was only available for the northern Tanzania study areas,
Assirik and Gombe. No survey of wild plant foods for humans has
been published for any of the chimpanzee study areas, hence
northern Tanzania is the only source of comparative data for the
‘‘Human Foods’’ portion of Figure 3.
The number of plant species with fruits eaten by baboons at
three sites, and by humans in the northern Tanzanian habitats
taken together as a whole, is comparable to the number of plant
species with fruits edible to savanna chimpanzees at Assirik and
Fongoli, whereas Kasakati chimpanzees eat more species of fruits
and Semliki chimpanzees eat fewer species of fruits (see left half of
Whereas savanna chimpanzees tend to eat mostly tree fruits,
fruits available in the northern Tanzania habitats were mostly from
shrubs and forbs. Of the 34 fruits and seeds/pods consumed by the
Fongoli chimpanzees, 23 are from trees, nine are from species
identified as shrubs or tree/shrub, and two are from climbers
(Pruetz, 2006). At Assirik, of the 34 fruits consumed by chimpan-
zees, 23 are from trees, seven are from shrubs, and four are from
vines or lianas (McGrew et al., 1988). At Mahale, at least 55 tree
species had fruits eaten by chimpanzees (Nishida and Uehara,
1983). Of all of the fruits in the northern Tanzania vegetation plots
edible to chimpanzees, baboons, or humans, nine are from trees, 30
are from shrubs, and 24 are from forbs.
In general, there are fewer trees bearing fleshy fruits in the
northern Tanzania semi-arid savannas compared to the savanna
chimpanzee sites. In the relatively forested habitat occupied by the
chimpanzees at Kibale, the density of large, fleshy fruit producing
trees ranged from approximately 26 to 195 per hectare, comprised
of 34 different fleshy fruit producing tree species (Balcomb et al.,
2000). At the savanna chimpanzee site of Fongoli, there are 27
large, fleshy fruit producing trees species and the chimpanzees feed
on at least 17 of them, although the density of these trees is not
reported (Pruetz, 2006). As a comparison, I categorised tree fruits
from the northernTanzanian studyareas as dry or fleshy, regardless
of whether they have been shown to be edible by primates,
following Balcomb et al. (2000). Of the 25 tree species represented
in the Serengeti and Lake Manyara vegetation plots, 15 produce
fleshy fruits and 10 produce dry fruits including seeds and pods
(Table 6). In the Manyara groundwater forest, the average density is
32 fleshy fruit trees per hectare, which is near the low end of the
density of fleshy fruit producing trees at Kibale. One might expect
a similar density in gallery forests lining perennial rivers in semi-
arid northern Tanzania savannas, as they are similar in structure
and species composition to the Manyara groundwater forest
(Hughes, 1988; Copeland, 2007). However, those areas make up
a very small portion, probably less than 1%, of the semi-arid
savanna region in northern Tanzania and Kenya in general. The
majority of the area is more similar to the bushland and bush
grassland habitats of the northern Tanzania vegetation plots, in
which the density of large, fleshy fruit producing trees was two and
Calculated cover area expressed as percent (based on crown diametre) of trees and shrubs with edible parts in the Serengeti and Lake Manyara vegetation plots.a
HabitatGrowth form FruitsSeeds/podsFlowers Leaves, shootsStems Bark, cambium, gumUnderground part
239 1121370 112
aSince crowns can overlap, some calculated cover areas are >100%.
Mean number of species per study area with edible plant parts from the northern Tanzania study areas, grouped by habitat and growth form.
HabitatGrowth formFruitsSeeds/podsFlowers Leaves, shootsStemsBark, cambium, gumUnderground part
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378 373
less than one per hectare, respectively. Thus, overall there are
relatively few fleshy fruit producing trees in the northern Tanzania
semi-arid savannas, and those that do exist are concentrated in
groundwater or riverine forest. In contrast, the dry fruit producing
trees, including those producing potentially edible seeds and pods,
are far more abundant in bushland and bush grassland habitats.
The number of species with seeds and pods eaten by chim-
panzees varies greatly and there is no clear pattern that differ-
entiates savanna chimpanzee versus wetter chimpanzee sites
(see the ‘‘Seeds’’ column of Fig. 3). The number of species with
seeds eaten by chimpanzees ranges from one at Kibale to 16 at
Kasakati. Not including seeds from grasses, the northern Tanza-
nian vegetation plots had four species for which chimpanzees are
known to eat the seeds/pods, and 21 and 22 plants with seeds/
pods eaten by baboons and humans, respectively, for a total of 28
potentially edible seeds/pods for hominins. The most abundant
trees with edible seeds/pods at the northern Tanzanian sites are
Acacia spp. Other plants with edible seeds include the tree Bal-
anites aegyptiacus and the palm Hyphaene petersiana, shrubs or
woody forbs (Abutilon spp. and Hibiscus spp.), and five herba-
ceous legumes (Copeland, 2007). Some of the most important
seed/pod foods for savanna chimpanzees in the miombo wood-
land site of Kasakati come from the dominant leguminous trees
Brachystegia and Julbernardia (Suzuki, 1969), but those miombo
tree species do not grow in the semi-arid habitats of northern
The number of edible underground parts (USOs) eaten by
chimpanzees in the wild is relatively low, with only one species
recorded at Kasakati (Suzuki, 1969), two at Assirik (McGrew et al.,
1988), none in the other chimpanzee study areas depicted in
Figure 3, but seven reported to be eaten at Ugalla (Hernandez-
Aguilar et al., 2007). Baboons eat USOs of two species at Gombe
(Ransom, 1981) and two at Assirik (McGrew et al., 1982). In the
northern Tanzania study areas, no plant species were encountered
that are known to have USOs consumed by chimpanzees. However,
a total of 28 species were found with USOs known to be consumed
by baboons and/or humans. Of those, eight are grasses, four are
sedges (Cyperus immensus, C. rotundus, C. usitatus, C. laevigatus), one
is the cattail Typha latifolia, and seven are the roots of woody
species (Copeland, 2007). Since studies of human plant foods are
not available for the savanna chimpanzee sites, it is not clear
whether USOs edible to humans exist there or not.
Fig. 3. The number of plant species with fruits, seeds/pods (‘‘Seeds’’), and underground parts (‘‘USOs’’) recorded to be eaten by chimpanzees, baboons, and humans in the semi-arid
savanna vegetation study in northern Tanzania and in several savanna chimpanzee and forest chimpanzee sites. References for chimpanzee foods at savanna and forest chimpanzee
sites are in Table 6; baboon food references: Gombe (Ransom, 1981), Assirik (McGrew et al., 1982). *The counts for Mahale and Fongoli include fruits and seeds/pods. **Assirik
baboon food data is number of genera rather than species.
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378 374
The northern Tanzanian semi-arid savannas include fewer
fleshy fruit bearing tree species, and fewer fruiting trees in general,
than the savanna chimpanzee sites. Chimpanzees appear to be
dedicated, though not exclusive, frugivores and many of their foods
are the large fleshy fruits of trees (Wrangham et al.,1998; Newton-
Fisher, 1999; Balcomb et al., 2000). Indeed, fruits account for more
than 60% of food for most chimpanzee populations (Pruetz, 2006).
In the savanna chimpanzee studies for which we have data,
chimpanzees eat morefruit species than anyothercategoryof plant
part (Table 5). Thus, in terms of diet, savanna chimpanzees may not
differ very much from forest chimpanzees in that even in ecosys-
tems that are dominated by grasses, they depend on tree fruits as
their mainstay. Chimpanzees need larger home ranges in wood-
lands or savannas with lower availability of fruits than they do in
forests (Yamagiwa and Basabose, 2006).
Savanna habitats are generally assumed to have less fruit avail-
able than do forests (e.g., Isbell and Young, 1996), and the species
richness of edible fruits does seem to be generally higher at forested
versus savanna sites (Balcomb et al., 2000; Hunt and McGrew, 2002;
Pruetz, 2006). Although the abundance of fruit rather than richness
of fruiting tree species may be a more important measure of food
availability (Pruetz, 2006), it is clear that the northern Tanzanian
savannas have a lower abundance of fleshy fruits from trees than
many of the moister savanna chimpanzee sites. This information is
essential for framing hypotheses of how early hominins might have
survived in semi-arid savannas in the past.
A second important difference in the composition of the
vegetation in semi-arid savannas versus the savanna chimpanzee
sites is the dominance of Acacia trees in semi-arid savannas.
Acacias are both species rich and have the highest density of
woody plants in nearly all of the northern Tanzania habitats except
for the groundwater forest. These Acacia dominated habitats also
tend to have a high richness in potentially edible plant food for
hominins of seed/pods (of Acacia spp., as well as other taxa),
grasses, and underground parts, although it is not possible to
compare the relative abundance of these in northern Tanzania
versus savanna chimpanzee sites due to a lack of information
regarding human and baboon foods in most of the chimpanzee
Since baboons are the only large bodied, nonhuman primates
currently living in the northern Tanzania semi-arid savannas,
studies of theirdiets in similarhabitats can help toshed light on the
actual abundance (as opposed to species richness) of various plant
foods in these habitats. Baboons appear to have a more flexible diet
than chimpanzees, and probably have broader digestive capabil-
ities, which enable them to live in a wide variety of habitats across
Africa. In semi-arid savanna habitats in Kenya that are similar tothe
northern Tanzanian study areas, baboons feed on large amounts of
grass and some Acacia parts.
Barton et al. (1992) conducted a study of baboon food resource
availability at Laikipia, Kenya, an area of wooded and bushed
grassland with Acacia trees and 550 mm annual rainfall. Only
a small fraction of plant food biomass was available from trees and
shrubs, whereas most could be found in the herbaceous layer
including grass and in sedge corms. At Gilgil, Kenya, an area of open
grassland interspersed with patches of scrub that receives 732 mm
rainfall annually, baboons spent 80% of their feeding time on grass
(Harding,1976). At Amboseli, Kenya, an area of open grassland with
scattered Acacia and Commiphorawoodland and bushland, baboons
spent 40% of their feeding time eating grass (Altmann and Altmann,
1970; Stacey, 1986).
Whether early hominins considered various Acacia species and
grass species edible is questionable, but one with important
implications regarding if or how hominins might have survived in
northern Tanzania-like savannas. Acacia xanthophloea seeds and
pods are often ignored by baboons (Wrangham and Waterman,
1981), whose digestive capabilities are almost certainly more
tolerant than those of apes and modern humans. While baboons,
vervet monkeys, and other animals eat Acacia tortilis pods, it is their
flowers, which appear during the dry season, that are most eagerly
consumed by primates (Wrangham and Waterman,1981). Modern
humans do not generally consider Acacia tortilis edible, although
they can make edible porridge out of the pods with cooking tech-
nology (Birnie, 1997).
Although Acacia trees and shrubs do occur in some of the
savanna chimpanzee habitats, important edible seeds and pods for
chimpanzees were those from Brachystegia and Julbernaria trees,
which are common in miombo woodlands but not present in semi-
arid northern Tanzania. Like some of the western Tanzanian
chimpanzees, baboons also live in miombo woodlands such as at
Mikumi, Tanzania. There they consume parts from more than 180
plant species, far more than in the semi-arid savannas of Kenya, and
their ‘‘staples’’ included grasses, tree fruits, herbs, and vines at
various times of year (Norton et al., 1987).
Number of plant food parts eaten by chimpanzees, also shown as percentage of the total number of plant food parts in that locality.
FruitsSeed/podLeaf/shootStem/stalkFlower/inflor. Bark/camb./sapUnderground partTotal # spec. References
43McGrew et al., 1988
Semliki–Hunt and McGrew, 2002
––Wrangham et al., 1991
2922–198Nishida and Uehara, 1983
Gombe1415 1911–147Wrangham, 1977
aFruits and seeds/pods were counted in one category.
bLeaf/shoot and stem/stalk were counted in one category.
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378 375
Thus, whereas chimpanzees do not live in semi-arid savannas of
northern Tanzania, perhaps due to a lack of fruits, nesting trees,
and/or water, baboons are able to survive in those and similar
habitats by heavily exploiting grasses, and also by consuming
Acacia parts, sedge corms, and other plants. If some hominins lived
in these types of savannas, then they almost certainly were not as
dependent as chimpanzees on fleshy fruits for daily survival. It is
unlikely that they ate a baboon-like diet with grasses or Acacias as
a mainstay, but this remains a possibility.
A third potential difference in plant food availability between
northern Tanzania’s semi-arid savannas and savanna chimpanzee
sites may lie in the abundance of edible USOs. However, unlike the
abundance of fleshy fruits and presence of Acacia habitats, this
difference is largely hypothetical due to a lack of data regarding the
potential abundance of edible USOs in the savanna chimpanzee
sites. On average there are more edible USOs available in savannas
than in forests (Vincent, 1985; Laden and Wrangham, 2005).
Chimpanzees eat very few USOs in both savanna and forest envi-
ronments, although Hernandez-Aguilar et al. (2007) have shown
that chimpanzees do exploit USOs at the savanna chimpanzee site
Underground parts have long been predicted as potentially
important to early hominin diets (e.g., Hatley and Kappelman,1980;
Vincent,1985; Laden and Wrangham, 2005). Modern human Hadza
foragers rely heavilyontubersforfood andlivewithin about 100 km
of the northernTanzaniavegetationplots,inhabitatsthat are similar
in terms of woody cover and annual rainfall (Vincent, 1985).
However, none of the tubers on which the Hadza depend were
encountered in the northern Tanzania savanna study areas. In
a study of tuber use among the Hadza, Vincent (1985) brought
several Hadzawomen to Lake Manyara to search for tuberous plants
there, but none were found.
Wetland habitats may provide an important source of USOs that
are potentially edible to hominins (Conklin-Brittain et al., 2002;
Sponheimer et al., 2005; Wrangham, 2005). Plant composition in
the marsh habitats in the semi-arid northern Tanzania savannas of
this study confirms that many of the marsh plants are edible, and
that the most common edible part is the corm, stem base, bulb, or
root of the marsh plant. These marsh plants can occur in vast stands,
depending on the wetland habitat and size, and therefore should be
considered as a potential foraging resource for hominins. However,
data are not currently available to determine whether marshy
wetlands are more abundant in semi-arid savannas versus other
savanna types, including those inhabited by savanna chimpanzees.
Such information will be helpful for future modeling of early hom-
inin foraging behaviour in various savanna environments.
Given the information gathered above, what ecological charac-
teristics might have been required for early hominins to survive in
relatively open semi-arid savannas? It seems unlikely that early
hominins living in semi-arid savannas could have been as depen-
denton fleshy fruits as chimpanzees, but if so, then theymight have
survived by having greatly increased home range sizes to follow
relatively sparse but long riverine forests where fleshy fruits from
trees are more likely to be found. Alternatively, a hominin who
could survive on either grass or Acacia plant foods would not need
to have a very large home range in semi-arid northern Tanzania
savannas, although the seasonal nature of these resources would
mean that other staples must be depended upon at other times of
year. The exploitation of wetland marsh plant USOs might also
provide seasonal resources, becoming more accessible in the early
dry season when marsh water levels have receded. Digging and
chopping tools would facilitate the exploitation of these resources.
Finally, a highly increased focus on animal foods would be another
option for hominin survival in semi-arid savannas and could help
account for the C4isotopic signature found in hominin teeth.
This study presents the context of hominin plant food avail-
ability for relatively open semi-arid African savannas that serve as
analogs for some hominin palaeoenvironments, and compares that
with food availability at the relatively moist savanna chimpanzee
sites. The semi-arid savannas have fewer fleshy tree fruits, are
dominated by Acacia and grass plant foods, and have wetland
habitats with abundant edible underground parts of marsh plants.
According to these modern analogs, early hominins that lived in
semi-arid savannas were not likely to be as frugivorous as chim-
panzees and must have had a substantially different ecological
Tree species from Serengeti and Lake Manyara categorised by fruit type following Balcomb et al. (2000).
FamilySpecies Fruit typeBush grasslandBush landForest
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378 376
focus. It is clearly important to continue in the endeavor to refine
palaeoenvironmental reconstructions of early hominin environ-
ments, as well as to study a variety of modern analog environments
in order to model the range of potential differences in plant food
availability between different savanna types. When combined with
direct tests of hominin diets, this combination of information on
the ‘fundamental’ and the ‘realised’ plant food niche of early
hominins may help us understand some of the early factors that
differentiated hominins from other African apes.
I would like tothank Bill McGrew for organising the Primatology
meets Palaeoanthropology conference at Cambridge University,
and the many generous participants of that meeting. For useful
discussions on this topic, I would like to thank Matt Sponheimer,
William McGrew, Ryne Palombit, Adriana Hernandez-Aguilar, Rob
Blumenschine, Craig Feibel, Jack Harris, and Colleen Hatfield. For
help with field work I thank Gail Ashley, Rob Blumenschine, Amy
Cushing, Dan Deocampo, Jim Ebert, Grant Hopcraft, Sammy Lauo,
Fidelis Masao, Joseph Masoy, John Mosi, Jackson Njau, Godfrey Ole-
Moita, Raphael Ole-Moita, Goodluck Peter, Charles Peters, Nancy
Sikes, Joanne Tactikos, Amiyo Tlaa, Augustino Venance, and Peyton
West. I would also like to acknowledge the very helpful comments
of Susan Anto ´n, Jim Moore, and the anonymous reviewers. I thank
the Commission on Science and Technology, Tanzania, the Ngor-
ongoro Conservation Area Authority, and the Tanzania National
Parks Association for permission to conduct the fieldwork, NSF
doctoral dissertation grant #9728984 and the Center for Human
Evolutionary Studies at Rutgers University for funding to conduct
the fieldwork, and the Max PlanckSociety for funding whilewriting
up the results.
Altmann, S.A., Altmann, J., 1970. Baboon Ecology: African Field Research. University
of Chicago Press, Chicago.
Anderson, G.D., Talbot, L.M., 1965. Soil factors affecting the distribution of the
grassland types and their utilization by wild animals on the Serengeti Plains,
Tanganyika. J. Ecol. 53, 33–56.
Andrews, P., Bamford, M., 2008. Past and present vegetation ecology of Laetoli,
Tanzania. J. Hum. Evol. 54, 78–98.
Avery, D.M., 2001. The Plio-Pleistocene vegetation and climate of Sterkfontein and
Swartkrans, South Africa, based on micromammals. J. Hum. Evol. 41, 113–132.
Balcomb, S.R., Chapman, C.A., Wrangham, R.W., 2000. Relationship between
chimpanzee (Pan troglodytes) density and large, fleshy-fruit density: conser-
vation implications. Am. J. Primatol. 51, 197–203.
Baldwin, P.J., McGrew, W.C., Tutin, C.E.G., 1982. Wide-ranging chimpanzees at Mt.
Assirik, Senegal. Int. J. Primatol. 3, 367–383.
Barton, R.A., Whiten, A., Strum, S.C., Byrne, R.W., Simpson, A.J.,1992. Habitat use and
resource availability in baboons. Anim. Behav. 43, 831–844.
Beentje, H.J., 1994. Kenya Trees, Shrubs, and Lianas. National Museums of Kenya,
Belsky, A.J., 1990. Tree/grass ratios in East African savannas: a comparison of
existing models. J. Biogeogr. 17, 483–489.
Birnie, A., 1997. What Tree is That? A Beginner’s Guide to 40 Trees in Kenya. Jaca-
randa Designs, Ltd., Nairobi.
Blundell, M., 1987. Collins Photo Guide to the Wild Flowers of East Africa. Harper
Collins Publishers, Hong Kong.
Bonnefille, R., 1984. Cenozoic vegetation and environments of early hominids in
east Africa. In: Whyte, R.O. (Ed.), The Evolution of East Asian Environments.
Paleobotany, Paleozoology. Paleoanthropology vol. II Center of Asian Studies,
University of Hong Kong.
Bonnefille, R., 1995. A reassessment of the Plio-Pleistocene pollen record of east
Africa. In: Vrba, E.S., Denton, G.H., Partridge, T.C., Burkle, L.H. (Eds.), Paleo-
climate and Evolutions with Emphasis on Human Origins. Yale University Press,
New Haven, pp. 299–310.
Carr, C.J., 1976. Plant ecological variation and pattern in the Lower Omo Basin.
In: Coppens, Y., Clark, H.F., Isaac, G.L., Leakey, R.E.F. (Eds.), Earliest Man and
Environments in the Lake Rudolf Basin. University of Chicago Press, Chicago
and London, pp. 432–467.
Cerling, T.E., Hay, R.L., 1986. An isotopic study of paleosol carbonates from Olduvai
Gorge. Quatern. Res. 25, 63–78.
Chivers, D.J., Hladik, C.M., 1984. Diet and gut morphology in primates. In:
Chivers, D.J., Wood, B.A., Bilsborough, A. (Eds.), Food Aquisition and Processing
in Primates. Plenum Press, New York, pp. 213–230.
Coates Palgrave, K., 1993. Trees of Southern Africa. Struik Publishers, Cape Town.
Collins, D.A., McGrew, W.C., 1988. Habitats of three groups of chimpanzees
(Pan troglodytes) in western Tanzania compared. J. Hum. Evol. 17,
Conklin-Brittain, N.L., Wrangham, R.W., Smith, C.C., 2002. A two-stage model of
increased dietary quality in early hominid evolution: the role of fiber. In:
Unger, P.S., Teaford, M.F. (Eds.), Human Diet: its Origin and Evolution. Bergin
and Garvey, Westport, pp. 61–76.
Copeland, S.R., 2004. Paleoanthropological implications of vegetation and wild
plant resources in modern savanna landscapes, with applications to Plio-
Pleistocene Olduvai Gorge, Tanzania. Ph.D. Dissertation, Rutgers, the State
University of New Jersey.
Copeland, S.R., 2007. Vegetation and plant food reconstruction of lowermost Bed II,
Olduvai Gorge, using modern analogs. J. Hum. Evol. 53, 146–175.
Coughenour, M.B., Ellis, J.E., 1993. Landscape and climatic control of woody vege-
tation in a dry tropical ecosystem: Turkana District, Kenya. J. Biogeogr. 20,
Deocampo, D., 2004. Hydrogeochemistry in the Ngorongoro Crater, Tanzania,
and implications for land use in a World heritage site. Appl. Geochem. 19,
Ditchfield, P., Hicks, J., Plummer, T., Bishop, L.C., 1999. Current research on the Late
Pliocene and Pleistocene deposits north of Homa Mountain, southwestern
Kenya. J. Hum. Evol. 36, 123–150.
Dominguez-Rodrigo, M., Lopez-Saez, J.A., Vincens, A., Alcala, L., Luque, L.,
Serrallonga, J., 2001. Fossil pollen from the Upper Humbu Formation of Peninj
(Tanzania): hominid adaptation to a dry open Plio-Pleistocene savanna envi-
ronment. J. Hum. Evol. 40, 151–157.
Fernandez-Jalvo, Y., Denys, C., Andrews, P., Williams, T., Dauphin, Y., Humphrey, L.,
1998. Taphonomy and palaeoecology of Olduvai Bed-I (Pleistocene, Tanzania).
J. Hum. Evol. 34, 137–172.
Frankfurt Zoological Society, 1971. Map of the Serengeti National Park and the
Surrounding Area. Compiled by T.M. Caro and drawn by Hunting Technical
Gerresheim, K., 1974. The Serengeti Landscape Classification (with Accompanying
1:250,000 Scale Map). Africa Wildlife Leadership Foundation, Nairobi.
Goodall, J., 1986. The Chimpanzees of Gombe: Patterns of Behavior. Harvard
University Press, Cambridge, MA.
Haines, R.W., Lye, K.A., 1983. The Sedges and Rushes of East Africada Flora of the
Families Juncaceae and Cyperaceae in East Africa with a Particular Reference to
Uganda. East African Natural History Society, Nairobi.
Harding, R.S.O., 1976. Ranging Patterns of a Troup of Baboons (Papio anubis) in
Kenya. Folia Primatol. 25, 143–185.
Harris, J.M.,1991. Koobi Fora Research Project, Vol. 3: The Fossil Ungulates: Geology,
Fossil Artiodactyls, and Paleoenvironments. Clarendon Press, Oxford.
Hatley, T., Kappelman, J., 1980. Bears, pigs, and Pio-Pleistocene hominids: a case for
the exploitation of belowground food resources. Hum. Ecol. 8, 371–387.
Hay, R.L., 1976. Geology of the Olduvai Gorge. University of California Press.
Herlocker, D., 1975. Woody Vegetation of the Serengeti National Park. Caeser
Kleberg Research Program. Texas A and M University, College Station, Texas.
Hernandez-Aguilar, A., 2009. Chimpanzee nest distribution and site reuse in a dry
habitat: implications for early hominin ranging. J. Hum. Evol. 57, 350–364.
Hernandez-Aguilar, A., Moore, J., Pickering, T.R., 2007. Savanna chimpanzees use
tools to harvest the underground storage organs of plants. Proc. Natl. Acad. Sci.
Hughes, F.M.R.,1988. The ecology of African floodplain forests in semi-arid and arid
zones: a review. J. Biogeogr. 15, 127–140.
Hunt, K., McGrew, W.C., 2002. Chimpanzees in the dry habitats of Assirik, Senegal
Marchant, L.F. (Eds.), Behavioural Diversity in Chimpanzees and Bonobos.
Cambridge University Press, Cambridge, pp. 35–51.
Ibrahim,K.M.,Kabuye,C.H.S.,1987. AnIllustratedManualofKenya Grasses. FAO,Rome.
Isabirye-Basuta, G., 1989. Feeding ecology of chimpanzees in the Kibale Forest,
Uganda. In: Heltne, P.G., Marquardt, L.A. (Eds.), Understanding Chimpanzees.
Harvard University Press, Cambridge, MA, pp. 116–127.
Isbell, L.A., Young, T.P., 1996. The evolution of bipedalism in hominids and reduced
group size in chimpanzees: alternative responses to decreasing resource
availability. J. Hum. Evol. 30, 389–397.
Itani, J., 1979. Distribution and adaptation of chimpanzees in an arid area. In:
Hamburg, D.A., McCown, E.R. (Eds.), The Great Apes. Benjamin/Cummings,
Menlo Park, CA, pp. 54–71.
Kano, T., 1971. Distribution and adaptation of the chimpanzee on the eastern shore
of Lake Tanganyika. Kyoto Univ. Afri. Stud. 7, 37–129.
Kappelman, J., 1984. Plio-Pleistocene environments of Bed I and Lower Bed II,
Olduvai Gorge, Tanzania. Palaeogeogr. Palaeoecol. 48, 171–196.
Laden, G., Wrangham, R., 2005. The rise of the hominids as an adaptive shift in
fallback foods: plant underground storage organs (USOs) and australopith
origins. J. Hum. Evol. 49, 482–498.
Lee-Thorp, J.A., Thackeray, J.F., van der Merwe, N.J., 2000. The hunters and the
hunted revisited. J. Hum. Evol. 39, 565–576.
Lee-Thorp, J.A., van der Merwe, N.J., Brain, C.K., 1994. Diet of Australopithecus
robustus at Swartkrans from stable carbon isotopic analysis. J. Hum. Evol. 27,
In: Boesch,C., Hohmann,G.,
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378377
Levin, N.E., Simpson, S.W., Quade, J., Cerling, T.E., Frost, S.R., 2008. Herbivore Download full-text
enamel carbon isotopic composition and the environmental context of Ardi-
pithecus at Gona, Ethiopia. In: Quade, J., Wynn, J.G. (Eds.), The Geology of Early
Humans in the Horn of Africa: Geological Society of America Special Paper 446,
pp. 215–234,. doi:10.1130/2008.2446(10).
Loth, P.E., Prins, H.H.T., 1986. Spatial patterns of the landscape and vegetation of
Lake Manyara National Park. ITC J. 2, 115–130.
McBrearty, S., Jablonski, N.G., 2005. First fossil chimpanzee. Nature 437, 105–108.
McGrew, W.C., Baldwin, P.J., Tutin, C.E.J., 1981. Chimpanzees in a hot, dry, and open
habitat: Mt. Assirik, Senegal, West Africa. J. Hum. Evol. 10, 227–244.
McGrew, W.C., Baldwin, P.J., Tutin, C.E.G., 1988. Diet of wild chimpanzees
(Pan troglodytes verus) at Mt. Assirik, Senegal: I. Composition. Am. J. Primatol.
McGrew, W.C., Sharman, M.J., Baldwin, P.J., Tutin, C.E.G., 1982. On early hominid
plant-food niches. Curr. Anthropol. 23, 213–214.
Moore, J., 1992. ‘‘Savanna’’ chimpanzees. In: Nishida, T., McGrew, W.C., Marler, P.,
Pickford, M., de Wall, F.B.M. (Eds.), Topics in Primatology. Vol. I: Human Origins.
University of Tokyo Press, Japan, pp. 99–118.
Moore, J., 1994. Plants of the Tongwe East Forest Reserve (Ugalla), Tanzania. Tropics
Moore, J., 1996. Savanna chimpanzees, referential models and the last common
ancestor. In: McGrew, W.C., Marchant, L., Nishida, T. (Eds.), Great Ape Societies.
Cambridge University Press, Cambridge, pp. 275–292.
Newton-Fisher, N.E., 1999. The diet of chimpanzees in the Budongo Forest Reserve,
Uganda. Afr. J. Ecol. 37, 344–354.
Nishida, T., 1989. A note on the chimpanzee ecology of the Ugalla area, Tanzania.
Primates 30, 129–138.
Nishida, T., Uehara, S., 1983. Natural diet of chimpanzes (Pan troglodytes schwein-
furthii): long-term record from the Mahale Mountains, Tanzania. Afr. Stud.
Monogr. 3, 109–130.
Norton, G.W., Rhine, R.J., Wynn, G.W., Wynn, R.D., 1987. Baboon diet: a five-year
study of stability and variability in the plant feeding and habitat of the yellow
baboons (Papio cynocephalus) of Mikumi National Park, Tanzania. Folia Primatol.
Norton-Griffiths, M., Herlocker, D., Pennycuick, L., 1975. The patterns of rainfall in
the Serengeti ecosystem, Tanzania. E. Afr. Wildlife J. 13, 347–374.
Peters, C.R., Blumenschine, R.J., 1995. Landscape perspectives on possible land use
patterns for early hominids in the Olduvai Basin. J. Hum. Evol. 29, 321–362.
Peters, C.R., O’Brien, E.M., 1981. The early hominid plant-food niche: insights from
an analysis of plant exploitation by Homo, Pan, and Papio in Eastern and
Southern Africa. Curr. Anthropol. 22, 127–140.
Peters, C.R., O’Brien, E.M., Drummond, R.B., 1992. Edible Wild Plants of Sub-Saharan
Africa. Royal Botanical Gardens, Kew.
Peters, C.R., Vogel, J.C., 2005. Africa’s wild C4plant foods and possible early hominid
diets. J. Hum. Evol. 48, 219–236.
Plummer, T.W., Bishop, L.C.,1994. Hominid paleoecology at Olduvai Gorge, Tanzania
as indicated by antelope remains. J. Hum. Evol. 27, 47–75.
Plummer, T., Bishop, L.C., Ditchfield, P., Hicks, J., 1999. Research on late Pliocene
Oldowan sites at Kanjera South, Kenya. J. Hum. Evol. 36, 151–170.
Pratt, D.J., Gwynne, M.D., 1977. Rangeland Management and Ecology in East Africa.
Hodder and Stoughton, London.
Prins, H.H.T., Loth, P.E., 1988. Rainfall patterns as background to plant phenology in
northern Tanzania. J. Biogeogr. 15, 451–463.
Pruetz, J., 2006. Feeding ecology of savanna chimpanzees (Pan troglodytes verus) at
Fongoli, Senegal. In: Hohmann, G., Robbins, M.M., Boesch, C. (Eds.), Feeding
Ecology in Apes and Other Primates. Cambridge University Press, Cambridge,
Ransom, T.W., 1981. Beach Troop of the Gombe. Bucknell University Press,
Reed, K.E.,1997. Early hominid evolution and ecological change through the African
Plio-Pleistocene. J. Hum. Evol. 32, 289–322.
Reynolds, V., Plumptre, A.J., Greenham, J., Harborne, J.,1998. Condensed tannins and
sugars in the diet of chimpanzees (Pan troglodytes schweinfurthii) in the
Budongo Forest, Uganda. Oecologia 115, 331–336.
Rowell, T.E., 1966. Forest living baboons in Uganda. J. Zool. 149, 344–364.
Schoeninger, M.J., Moore, J., Sept, J.M., 1999. Subsistence strategies of two savanna
chimpanzee populations: the stable isotope evidence. Am. J. Primatol. 49,
Sept, J.M., 1986. Plant foods and early hominids at site FxJj 50, Koobi Fora, Kenya. J.
Hum. Evol. 15, 751–770.
Sept, J.M.,1992. Was there no place like home? a new perspective on early hominid
archaeological sites from the mapping of chimpanzee nests. Curr. Anthropol. 33,
Shipman, P., Harris, J.M., 1988. Habitat preference and paleoecology of Austral-
opithecus boisei in eastern Africa. Evolutionary history of the ‘‘Robust’’
Australopithecines. In: Grine, F. (Ed.). Aldine de Gruyter, New York, pp.
Sikes, N., 1994. Early hominid habitat preferences in East Africa: paleosol carbon
isotope evidence. J. Hum. Evol. 27, 25–45.
Sinclair, A.R.E., 1979. Dynamics of the Serengeti ecosystem: process and pattern. In:
Sinclair, A.R.E., Norton-Griffiths, M. (Eds.), Serengeti: Dynamics of an Ecosystem.
The University of Chicago Press, Chicago, pp. 1–30.
Sponheimer, M., Lee-Thorp, J.A., 1999. Isotopic evidence for the diet of an early
hominid Australopithecus africanus. Science 283, 368–370.
Sponheimer, M., Lee-Thorp, J., de Ruiter, D., Codron, D., Codron, J., Baugh, A.T.,
Thackeray, F., 2005. Hominins, sedges, and termites: new carbon isotope data
from the Sterkfontein valleyand Kruger National Park. J. Hum. Evol. 48, 301–312.
Sponheimer, M., Loudon, J.E., Codron, D., Howells, M.E., Pruetz, J.D., Codron, J., de
Ruiter, D.J., Lee-Thorp, J.A., 2006. Do ‘‘savanna’’ chimpanzees consume C4
resources? J. Hum. Evol. 51, 128–133.
Stacey, P.B., 1986. Group size and foraging efficiency in yellow baboons. Behav. Ecol.
Sociobiol. 18, 175–187.
Suzuki, A., 1969. An ecological study of chimpanzees in a savanna woodland.
Primates 10, 103–148.
Van der Merwe, N.J., Masao, F.T., Bamford, M.K., 2008. Isotopic evidence for con-
trasting diets of early hominins Homo habilis and Australopithecus boisei of
Tanzania. S. Afr. J. Sci. 104, 153–155.
Van der Merwe, N.J., Thackeray, J.F., Lee-Thorp, J.A., Luyt, J., 2003. The carbon
isotope ecology and diet of Australopithecus africanus at Sterkfontein, South
Africa. J. Hum. Evol. 44, 581–597.
Vincent, A.S., 1985. Wild Tubers as a Harvestable Resource in the East African
Savannas: Ecological and Ethnographic Studies. Ph.D. Dissertation, University of
Webster, R., Beckett, P.H.T., 1970. Terrain classification and evaluation using air
photography: a review of recent work at Oxford. Photogrammetria 26, 51–75.
Wrangham, R.W., 1977. Feeding behaviour of chimpanzees in Gombe National
Park, Tanzania. In: Clutton-Brock, T.H. (Ed.), Primate Ecology: Studies of
Feeding and Ranging Behaviour in Lemurs, Monkeys, and Apes. Academic
Press, London, pp. 503–538.
Wrangham, R.W., 2005. The delta hypothesis. In: Ungar, P.S. (Ed.), Evolution of the
Human Diet: The Known, The Unknown, and The Unknowable. Oxford
University Press, Oxford, pp. 308–323.
Wrangham, R.W., Conklin, N.L., Chapman, C.A., Hunt, K.D., 1991. The significance of
fibrous foods for Kibale forest chimpanzees. Phil. Trans. R. Soc. 334, 171–178.
Wrangham, R.W., Conklin-Brittain, N.L., Hunt, K.D., 1998. Dietary response of
chimpanzees and cercopithecines to seasonal variation in fruit abundance. I.
Antifeedants. Int. J. Primatol. 19, 949–970.
Wrangham, R.W., Rogers, M.E., I-Basuta, G., 1993. Ape food density in the ground
layer in Kibale Forest, Uganda. Afr. J. Ecol. 31, 49–57.
Wrangham, R.W., Waterman, P.G., 1981. Feeding behaviour of vervet monkeys on
Acacia tortilis and Acacia xanthophloea: with special reference to reproductive
strategies and tannin production. J. Anim. Ecol. 50, 715–731.
Yamagiwa, J., Basabose, A.K., 2006. Effects of fruit scarcity on foraging strategies of
sympatric gorillas and chimpanzees. In: Hohmann, G., Robbins, M.M., Boesch, C.
(Eds.), Feeding Ecology in Apes and Other Primates. Cambridge University Press,
Cambridge, pp. 73–96.
Zazzo, A., Bocherens, H., Brunet, M., Beauvilain, A., Billiou, D., Mackaye, H.T.,
Vignaud, P., Mariotti, A., 2000. Herbivore paleodiet and paleoenvironmental
changes in Chad during the Pliocene using stable isotope ratios of tooth enamel
carbonate. Paleobiology 26, 294–309.
S.R. Copeland / Journal of Human Evolution 57 (2009) 365–378 378