Journal of Human Evolution 53 (2007) 146e175
Vegetation and plant food reconstruction of lowermost Bed II,
Olduvai Gorge, using modern analogs
Sandi R. Copeland a,b,*
b
a
Department of Anthropology, University of Colorado at Boulder, Boulder, CO 80309, USA
Department of Anthropology, Rutgers, the State University of New Jersey, New Brunswick, NJ 08901, USA
Received 6 October 2005; accepted 2 March 2007
Abstract
Vegetation and plant foods for hominins of lowermost Bed II, Olduvai Gorge were modeled by examining vegetation in modern habitats in
northern Tanzania (Lake Manyara, Ngorongoro, Serengeti) that are analogous to the paleolandscape in terms of climate, land forms, and soil
types, as indicated by previous paleoenvironmental studies of Olduvai. Plant species in the modern habitats were identified in a series of sample
plots, and those known to be eaten by modern humans, chimpanzees, or baboons were considered potentially edible for early hominins. Within
the 50e80 kyr deposition of lowermost Bed II, periods of drier climate were characterized by low lake stands and a broad eastern lacustrine plain
containing a mosaic of springs, marsh, woodland, and edaphic grassland. Based on results of this study, plant food diversity in each of those
habitats was relatively low, but the mosaic nature of the area meant that hominins could reach several different habitat types within short distances, with access to potential plant foods including marsh plants, grass grains, roots, shrub fruits, edible parts from palms, leafy herbaceous
plants, and Acacia pods, flowers, and gum. Based on Manyara analogs, a greater variety of plant foods, such as tree fruits (e.g., Ficus, Trichilia)
and the roots and fruits of shrubs (e.g., Cordia, Salvadora) would be expected further east along the rivers in the lacustrine terrace and alluvial
fans. Interfluves of the alluvial fans were probably less wooded and offered relatively fewer varieties of plant foods, but there is sparse paleoenvironmental evidence for the character of Olduvai’s alluvial fans, making the choice of appropriate modern analogs difficult. In the western
side of the basin, based on modern analogs in the Serengeti, riverine habitats provided the greatest variety of edible plant food species (e.g.,
Acacia, Grewia, Justicia). If the interfluves were grassland, then a large variety of potentially edible grasses and forbs were present seasonally.
Periods of wetter climate resulted in a much expanded paleolake and a shrinking of the eastern lacustrine plain mosaic into a narrow zone. The
dominant landscape features were then forest-lined rivers in the eastern alluvial fans, and rivers in the western side of the paleo-Olduvai basin
were also better watered at these times, supporting denser woody vegetation with large varieties of edible fruits, leaves, and underground parts.
Ó 2007 Elsevier Ltd. All rights reserved.
Keywords: Hominin; Diet; Plant foods; Paleoenvironment; Serengeti; Ngorongoro; Manyara
Introduction
The role of plants and plant foods in the ecology of early
hominins is not well understood because of poor preservation
of plants in the fossil record and few traces of plant use in
the archaeological record. Nonetheless, the structure and
* 233 UCB, Department of Anthropology, University of Colorado at Boulder, Boulder, CO 80309-0233.
E-mail address: sandi.copeland@colorado.edu
0047-2484/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jhevol.2007.03.002
composition of the vegetation in which early hominins lived
must have played a significant role in their ecology and evolution, influencing the types of animals they encountered (competitors, predators), the types of plant foods available, the nature of
refuge from predators, and the availability of plant-based raw
material for tools, to name some examples. Plants were likely
to have been the major food source for early hominins in tropical environments, even among groups for which animal foods
provided a significant component of the diet. Unfortunately,
the nature of most paleoenvironmental evidence lends itself
to generalized reconstructions of vegetation, but in order to
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
understand hominin paleoecology we need to reconstruct local
habitats.
In this study, the rich vegetative details available in modern
habitats of East Africa are used to flesh out vegetation reconstructions of a fossil landscape, specifically to model the vegetation of local habitats at the Plio-Pleistocene site of Olduvai
Gorge, Tanzania, during lowermost Bed II times. The modern
analog habitats are in northern Tanzania in portions of Lake
Manyara National Park, Serengeti National Park, and Ngorongoro Crater Conservation Area. They were chosen because of
their similarity to Olduvai’s paleolandscape in terms of climate, land forms, and soil types, as outlined in detail below.
This study builds on the earlier work of Peters and Blumenschine (1995, 1996) that developed a paleolandscape model
for lowermost Bed II and predicted resources for hominins
based on reconnaissance surveys in eastern and southern
Africa. To expand upon their preliminary surveys, this study
included more detailed and formal vegetation surveys, with
systematic documentation of plant species composition, structure, and potential plant foods for hominins (Copeland, 2004).
Background
It is inherent in the nature of many techniques for paleoecological reconstruction to consider environments as though they
were homogeneous over areas larger than local habitats, in
which case local species composition and its variation within
a mosaic of habitats are lost in the broader perspective. For
example, the relatively large home range of medium-large
mammals, which often encompasses a variety of local habitats,
precludes their ability to indicate a specific habitat type.
Faunal-based reconstructions often indicate that a mosaic of
habitats was present, but alone they do not speak to the way
in which those habitats were distributed across the landscape.
While fossil pollen are extremely valuable because they pertain to the taxonomic composition of the ancient flora, they
too contain a large component of allochthonous elements
that indicate regional as opposed to local vegetation (e.g.,
Bonnefille, 1984).
Modern analogs can be studied at various spatial scales for
the purpose of reconstructing details of ancient vegetation,
such as plant foods, that are not preserved in the fossil/geological record. Peters and O’Brien (1981) pioneered the systematic study of modern vegetation as a way to understand the
nature of plant foods that might have been available to early
hominins. They have been documenting wild plant species
in sub-Saharan Africa that are edible to humans, chimpanzees,
or baboons (Peters and O’Brien, 1981; Peters et al., 1992) in
part to understand the fundamental plant food niche for
early hominins. At regional scales they have investigated the
predictive power of climate on edible plant species diversity
(O’Brien, 1988, 1993; O’Brien and Peters, 1991; Peters and
O’Brien, 1994) and potential keystone plant food species
(Peters, 1987; O’Brien and Peters, 1991). At finer spatial
scales, Peters and Maguire (1981) used plant foods of the
modern Makapansgat Valley, South Africa to model potential
plant foods for Pliocene Australopithecus africanus from the
147
Makapansgat fossil-bearing deposits. Sept (1986, 2001) documented the distribution and abundance of potential plant foods
in different sedimentary zones of modern riparian habitats in
Kenya that were analogous to riparian settings of Pliocene hominin sites at Koobi Fora. Sept (1990, 1994) also related edible
fruit abundance to vegetation structural categories along the
Semliki River in the Democratic Republic of Congo (then Zaire).
The current study benefited from the techniques and lessons
learned from these pioneering researchers, and combined
a vegetation sampling strategy with the hierarchical landscape
classification system reconstructed for the lowermost Bed II
paleoenvironment by Peters and Blumenschine (1995). They
synthesized the geological work of Hay (1976) and numerous
other paleoenvironmental studies from Olduvai and defined
the paleolandscape in terms of landscape facets, which in
turn are grouped into the broader categories of landscape associations and regions. Landscape facets are roughly equivalent
to local habitats and are defined by geographic relief, soil
type, water regime, and climate (Webster and Beckett, 1970;
Gerresheim, 1974). Not only are these factors potentially
recognizable in the fossil record, but they are also the factors
that to a large degree control the nature of the vegetation and
the ecology of each landscape facet.
The identification of a potential plant food for a hominin
would ideally incorporate a variety of characteristics of the
plant’s nutritional and chemical composition, as well as the
particular tolerance of the hominin in question, based on characteristics common to its species as well as its age and nursing
status. The degree to which plant parts can be digested varies
greatly with plants and consumers. Since the digestive capabilities of early hominins are unknown, it is necessary to broadly
define what was potentially edible. For the purpose of this
study, wild plant species that have been observed to be eaten
by humans, chimpanzees, or baboons will be considered as
potential hominin plant foods following Peters and O’Brien
(1981) and Peters et al. (1992). The case for particular plant
species is discussed in the text.
In general, baboons have a higher tolerance to many plant
toxins than modern humans and probably chimpanzees, as
suggested by the fact that they are relatively less frugivorous
and more folivorous than hominoids (Chivers and Hladik,
1984). For example, in a study at Kibale, hominoid diets
were found to differ from those of cercopithecoids in that
they are less toxic and of higher quality (Conklin-Brittain
et al., 1998; Wrangham et al., 1998). Nonetheless, there are
some compelling reasons to examine baboon plants foods as
potentially edible to early hominins. The modern habitats examined in this study are all inhabited by baboons but not by
chimpanzees or modern human foragers, so baboons are the
only relatively large-bodied primates that can serve as an immediate guide to potential wild plant foods in these particular
habitats. It is also relevant that at least some early hominins
appear to have lived in more open and/or arid habitats of the
type that are inhabited by modern baboons but not by other
(non-human) apes. In addition, fossil baboons are often found
at the same localities as fossil hominids. Thus, consideration
of baboon plant foods may provide insights into early hominin
148
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
diets, especially in terms of what was potentially available.
Ultimately, determining the realized niches of hominins
requires independent means of testing such as isotopic analyses
of fossil material and studies of dental wear.
Aims
This study has four specific aims: 1) to identify modern localities that serve as analogs to portions of the lowermost Bed
II paleolandscape based on previously documented paleoenvironmental evidence for climate, general vegetation type, soils,
and specific landscape units; 2) to document the vegetation
(species composition, structure) of each of these modern landscape facets; 3) to identify potential plant foods for hominins
in each of these modern landscape facets; and 4) to predict
the vegetation and plant food characteristics of the lowermost
Bed II paleolandscape, assuming that the modern analogs
provide accurate representations of the past. Consideration
of possible differences between the modern analog facets
and paleo-Olduvai facets and variations in climate throughout
the deposition of lowermost Bed II provides insights into the
degree of accuracy that can be expected from various aspects
of the modeling process and how vegetation and plant foods
may have changed over time. This study does not specifically
document the relative abundance, seasonal availability, or potential competition for plant foods in the modern habitats because
those ecological details cannot be reconstructed with any reasonable degree of certainty for a Plio-Pleistocene paleolandscape,
given the current state of paleoenvironmental evidence. The
focus here is on broadly useful fundamental issues, such as the
general structure and composition of the vegetation and types
of plant foods available.
The fossil study area: lowermost Bed II, Olduvai Gorge
Olduvai Gorge bisects the easternmost Serengeti Plains, exposing layers of sediment that range from almost two million
years old at the base to Holocene age near the top (Hay, 1976).
The gorge is about 22 km west to east, and the plains in the
vicinity of the gorge have an elevation of 1,350 to 1,520 meters. To the east and south are the Crater Highlands, a volcanic
range with elevations from 2,100 to 2,400 meters at the rim of
Ngorongoro Crater (Hay, 1976).
During the Plio-Pleistocene, the area that is now Olduvai
Gorge was a paleolake basin containing a shallow, salinealkaline lake. The geomorphology of the southern and eastern
portions of the basin was dominated by the processes of mountain erosion and volcanic activity. The northwestern sides of
Crater Highland Mountains Ngorongoro, Olmoti, Sadiman,
and Lemagrut drained directly into an Olduvai paleolake basin
via an extensive area of alluvial fans, and there were periodic
eruptions of volcanoes, particularly the now extinct Olmoti
(Hay, 1976). The modern Olbalbal Depressiondthe drainage
sump between the Crater Highlands and Olduvai Gorgeddid
not exist during Bed II times, as it was created with riftassociated down-faulting about 400,000 years ago (Hay, 1976).
Lowermost Bed II is bracketed above by Tuff IIA (1.7 Ma;
Hay, 1976) and below by Tuff IF. Although the isotopic age
for Tuff IF was reported as 1.75 Ma by Walter et al. (1991),
it may be closer to 1.78 Ma based on its provisional stratigraphic occurrence coincident with the geomagnetic polarity
reversal at the top of the Olduvai subchron (Blumenschine
et al., 2003). Therefore, the estimated duration of lowermost
Bed II is on the order of 50e85 thousand years.
At least two hominin species were present at Olduvai during lowermost Bed II times, Paranthropus boisei and Homo
habilis (Leakey, 1971; Tobias, 1991). The larger-bodied
Homo ergaster appears in northern Kenya by 1.78 Ma (Feibel
et al., 1989; Wood, 1992). Although its remains have not been
discovered in lowermost Bed II, it is present in the upper parts
of Bed II and could have been present earlier as well.
The hypothetical landscape associations and landscape
facets of the lowermost Bed II paleo-Olduvai basin are
summarized in Table 1 and are depicted in the map in Fig. 1.
Throughout lowermost Bed II, the paleolake fluctuated in size
seasonally, with drought years and with longer orbital cycles.
Figure 1 shows a period of relatively dry climate in which the
lake was about 10e15 km long and 5e20 km wide (Hay,
1976). The concentric zones shown around the lake are based
on the original descriptions by Hay’s (1976) and Peters and Blumeschine’s (1995; 1996) landscape unit divisions. The innermost zone was a perennial, saline-alkaline lake, surrounded
by an intermittently dry zone, followed by an intermittently
flooded lake margin. The lacustrine terrace marks the most extreme high lake levels during the wettest of climates, but most
often existed as a perennially dry zone.
The lacustrine plain adjacent to the lake itself was a low
gradient surface of seasonally inundated mudflats and marshlands. The exact location of the lacustrine plain shifted over
time with changes in climate, being closer to the mountains
in wetter times when the lake expanded. Streams originated
in the steep montane areas where rainfall was higher than at
the lower altitude lake-shore, and they crossed the alluvial
fans, lacustrine terrace, and lacustrine plain. Groundwater
flow probably occurred as well, perhaps emerging as freshwater springs around the eastern lake margin (Hay, 1996).
Evidence for the paleogeography of the western side of the
paleo-Olduvai basin is sparse compared to the well-studied
eastern lake margin area. The topographic relief in the western
basin was more gradual than in the east, and rivers draining the
eastern Serengeti probably emptied into the western side of the
lake. There is evidence for ephemeral rivers with fluvial flooding conditions in the western basin from Bed I times where the
Homo habilis maxilla OH 65 was found (Blumenschine et al.,
2003). The water catchment area on the western side of the
basin was much smaller than that from the highland-dominated
east and south. This may help to explain why the western side
of the paleolake was more saline-alkaline, while the eastern
lake waters were fresher (Hay, 1976).
The paleogeography of the northern side of the paleoOlduvai basin is also poorly understood, since the only exposures of lowermost Bed II age are limited outcrops along the
Fifth Fault (Hay, 1976). The Gol Mountains begin about
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
149
Table 1
Hypothetical landscape units of the Olduvai lowermost Bed II paleolandscape1
Landscape association
Landscape facet
Description
Serengeti Woodlands
Non-riverine
Riverine
Hilly area; soil of basement rock-derived material (granite, quartzite) mixed with volcanic
ash, rivers
Serengeti peneplain
Non-riverine
Riverine
Gently rolling landscape; calcareous loams and sandy clay loams overlying weathered tuff;
ephemeral rivers
Western lacustrine plain
Non-riverine
Riverine
Low gradient intermittently flooded to intermittently dry zone on the west side of the lake. Possible
ephemeral streams but with less volume of discharge than rivers on the east/southeastern lake side
Eastern/southeastern
lacustrine plain
Stream-fed wetland
Stream-fed ‘‘dry land’’
Small spring wetland
Small spring ‘‘dry land’’
Large spring wetland
Large spring ‘‘dry land’’
Upper non-riverine
Riverine, perennial
Riverine, ephemeral
Intermittently flooded zone of the lake. Low gradient, clay-dominated landform with some sand and
pebbles. Lower portions flooded for extended periods of time with saline alkaline lake water; upper
portions flooded rarely (except during wet climate phases). Low-sinuosity streams create stream-fed
wetlands. In places small or large springs emerge, supporting marsh and localized shrubland or
woodland adjacent to the springs. Ephemeral and/or perennial streams cross the upper lacustrine plain
Lacustrine terrace
Non-riverine
Riverine
Narrow transition zone corresponding to an old high lacustrine plain; soils well-drained, somewhat
alkaline and non-saline, and coarser than the lacustrine plain. Rivers cross the lacustrine terrace.
Major rock outcrops
Isolated inselbergs with some vegetation unique from the surrounding areas
Alluvial fans
Non-riverine
Riverine
Soils coarser-grained with reworked tuff, deeper and better drained than soils in the lacustrine
plain and terrace, but still somewhat alkaline. Ephemeral and possibly perennial streams
Mountainsides
Non-riverine
Riverine
Coarse, shallow volcanic soils with gravel and sand at high elevations; steep slopes and
freshwater streams
1
Based on Peters and Blumenschine (1995, 1996), Copeland (2004), and references therein.
Fig. 1. Hypothetical paleolandscape facets of lowermost Bed II, Olduvai Gorge, with outline of modern Olduvai Gorge in the background. Modified from Hay
(1976) and Peters and Blumenschine (1995).
150
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
15 km north of the modern Olduvai Gorge and are a series of
roughly east-west running ridges of metamorphic Early Paleozoic basement rocks, dominantly quartzite and gneiss/schist
(Hay, 1976). The soils in this vicinity were heavily influenced
by the eroding Gol Mountains.
Climate and vegetation of the lowermost
Bed II, Olduvai basin
Carbon and oxygen stable isotope studies of soil carbonates
suggest that the climate during lowermost Bed II times was
similar to today in terms of having distinctive wet and dry seasons, but that the mean annual temperature was 15e17 C,
compared to a mean of 22 C at Olduvai today (Cerling and
Hay, 1986). Soil carbonates and calcretes are relatively scarce
in Bed I and lower Bed II, which is consistent with an annual
rainfall exceeding 750 to 850 mm (Cerling and Hay, 1986), as
opposed to 400e600 mm in the region today (Norton-Griffiths
et al., 1975). Sikes (1994) interpreted stable carbon isotopes
from paleosol organic matter and co-existing pedogenic carbonates from lowermost Bed II within one square kilometer
of the eastern lake margin to reflect vegetation with 20 to
45% C4 plants, which is consistent with grassy woodland to
riparian forest environments.
The fossil pollen record corroborates the evidence of distinctive wet and dry seasons, a semi-arid climate, and an environment moister than the present during the Plio-Pleistocene at
Olduvai. Five successful pollen samples were derived from
lowermost Bed II, all from the eastern lake margin area (Bonnefille and Riollet, 1980; Bonnefille, 1984). The Afromontane
elements of the arboreal pollen were probably transported by
wind and rivers into the paleolake basin from the nearby Crater
Highlands (Bonnefille and Riollet, 1980; Bonnefille, 1984).
Much of the Sudano-Zambezian elements of the arboreal pollen derived from plants that grew near the paleolake basin,
such as Capparidaceae shrubs, which are insect-pollinated
and, therefore, the pollen grains are not easily transported by
wind (Bonnefille, 1984). One sample from HWK-EE included
a relatively large proportion of Acacia (Bonnefille and Riollet,
1980). When compared to modern pollen rain samples from
Olduvai, the fossil pollen suggests that during lower Bed I
and lowermost Bed II times, the Afromontane forest of the
nearby Crater Highlands was two to three times its present
size, which may reflect greater rainfall in the past (Bonnefille,
1984). However, recent deforestation due to modern human
activities may have artificially decreased the recent size of
the Afromontane forest (Bonnefille, 1984).
Hay (1976) recognized the presence of marsh environments
in the eastern paleolake margin, and several lines of paleoenvironmental evidence support this conclusion. Clay geochemistry suggests that within the exposed eastern lake margin
during lowermost Bed II times, a marshy wetland existed in
the area of HWK-E and freshwater conditions persisted long
enough to alter the clay chemistry from its original saline/alkaline signature when the clays were deposited under a high
lake stand (Deocampo et al., 2002). Macroplant fossil remains
and phytoliths from a two kilometer area of the lowermost Bed
II eastern lake margin indicate the presence of sedges, grasses,
and dicotyledonous plants at FLK-N, suggesting marshland
and grassland close to the paleolake after the deposition of
Tuff IF (Bamford et al., 2006). Macroplant fossils and phytoliths in the area of HWK-EE, also in the eastern lake margin,
preserve evidence of sedges and indicate that for a time the
area was characterized by grassland with clumps of palms
(Albert et al., 2006). The macroplant fossil and phytolith
data show evidence that vegetation varied over small distances
(100e200 m) and over relatively short periods of time (Albert
et al., 2006), highlighting the importance of understanding the
nature of variation in these local habitats/landscape facets in
terms of hominin resources.
Fossil wood from lowermost Bed II was determined by
Bamford (2005) to most likely represent the species Guibourtia
coleosperma, a tree species that grows today on Kalahari sands
south of Tanzania. The presence of such a tree at Olduvai suggests the presence of some large trees and that some periods of
lowermost Bed II were wetter than today (Bamford, 2005).
The diversity of fossil bovids from Bed I and lower Bed II
indicate the presence of a mosaic of physiognomic vegetation
types, including open grassland, bushland, and woodland, with
an increase in more open and drier conditions throughout the
Bed I and lower Bed II sequence (Kappelman, 1984; Potts,
1988; Shipman and Harris, 1988; Plummer and Bishop,
1994; Marean and Ehrhardt, 1995; Kappelman et al., 1997;
Fernandez-Jalvo et al., 1998).
Unfortunately, there is little evidence for the type of vegetation that existed on the alluvial fans that dominate the eastern and southeastern portions of the paleo-Olduvai basin, due
in part to the fact that much of these sediments are buried
under the Olbalbal Depression. The alluvial fan area was predicted by Peters and Blumenschine (1995) to be a focus of
hominin activity due to the hypothetical presence of fruitladen trees along drainage and the presence of trees for refuge
and sleeping. Evidence for the types of vegetation that grew in
the northern and western portions of the basin is also rare due
to limited exposures of lowermost Bed II in these areas.
At the time of lowermost Bed II deposition, the most important climatic cycle was a 41,000 year period as opposed to the
100,000 year ‘‘ice age’’ cycle that dominates the current climate (deMenocal, 1995). Climatic changes related to the
earth’s orbit were also of a much lower amplitude prior to
the late Pleistocene (deMenocal, 1995), so the climate and vegetation may not have changed as dramatically over the period
of orbital precessions during lowermost Bed II times as it did
in the glacial-interglacial cycles of late Pleistocene and Holocene times. Nonetheless, models of the paleovegetation need
to account for the variation that did occur, including climatic
changes ranging from those that are seasonal, to those that fluctuate within tens to thousands of years, up to an entire 41,000year climatic cycle resulting from changes in the earth’s orbit.
In sum, the paleoenvironmental evidence paints a picture of
lowermost Bed II as slightly moister than the present situation,
with a wet/dry seasonal climate and longer wet and dry phases
in accordance with orbital cycles. The eastern and southeastern sides of the basin were a mosaic of landscape facets that
151
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
included marshy wetlands and seasonally inundated lake flats
near the lake, open grassland, grassland with patches of palms,
and areas with trees such as riparian forests, woodland, and/or
bushland. The paleoenvironments of the northern and western
sides of the basin and of the alluvial fan are as yet poorly indicated other than by their general paleogeographical setting
combined with overall rainfall estimates and geological evidence for soil types.
Modern analog study areas
Since the Plio-Pleistocene Olduvai environment was topographically different than the modern setting, one must venture away from Olduvai Gorge to find analogous modern
landforms. The modern analog study areas were chosen by
three criteria. First, all have an overall climatic setting that
is semi-arid with seasonal rainfall similar to the annual amount
reconstructed for paleo-Olduvai. Second, a large component of
their soil derives from eruptions of the Crater Highlands volcanoes, as was the case for paleo-Olduvai. Third, at a finer
scale, particular modern landscape facets were chosen because
their topographic, hydrological, and ecological circumstances
have been reconstructed for the paleolandscape at Olduvai
based on paleoenvironmental information. The modern study
areas are listed in Table 2 and shown in the maps in Fig. 2.
As described in more detail below, Lake Manyara National
Park includes a Rift Valley lake with an associated lacustrine
plain, lacustrine terrace, and alluvial fan, similar at gross
levels to landscape facets reconstructed for the southeastern
side of paleolake Olduvai. The floor of Ngorongoro Crater is
another lacustrine plain with a shallow fluctuating lake, and
it serves as an analog for the lacustrine plains of paleolake
Olduvai. Serengeti National Park is immediately west of Olduvai Gorge and is presented as an analog for the western side
of the paleo-Olduvai basin.
Lake Manyara National Park
Lake Manyara is situated in a down-dropped portion of
a half-graben, with the rift escarpment rising steeply near
the western edge of the lake. The strip of land between the
northwestern edge of the lake and the escarpment, shown
boxed in the map in Fig. 2, is protected as a national park.
The average elevation of the lake is about 960 meters, while
the adjacent rift escarpment rises to approximately 1200e
1300 meters (Loth and Prins, 1986). The lake itself is shallow,
alkaline, brackish, and fluctuates in size seasonally.
The climate of Manyara follows the typical wet and dry
seasons, and annual rainfall is about 650 mm per year (based
on an average over 25 years; Loth and Prins, 1986). At the
northern edge of Lake Manyara, rivers and perennial springs
spill out into the valley at the base of the rift escarpment creating natural irrigation. The well-drained alluvial fans in the
northern area of the park support a lush, evergreen groundwater forest that would not otherwise be possible under the existing climatic regime (Greenway and Vesey-Fitzgerald, 1969).
Further south, the lacustrine terrace between the lake and
Table 2
Modern analog study areas in northern Tanzania
Region
Study area
Lake Manyara Mkindu interfluve
Mkindu on alluvial fan
Ndala lake flat
Msasa lake flat
Mkindu on lacustrine plain
Msasa on lacustrine plain
Ndala-Chemchem interfluve
Ndilana-Msasa interfluve
Msasa on lacustrine terrace
Ndilana on lacustrine terrace
Vegetation
structure1
Map
location2
forest
forest
bush grassland
bush grassland
forest
bushland
bushland
bushland
bushland
bushland
1
2
4
5
6
7
8
9
10
11
Serengeti
Barafu Plain
grassland
Barafu Valley
bush grassland
Mbalageti River
bushland
Sangare River
bushland
Nyamara interfluve
grassland
Seronera-Wandamu interfluve grassland
Nyamara River
bushland
Seronera River
bushland
12
13
14
15
16
17
18
19
Ngorongoro
Crater
Ngoitokitok North
20
Ngoitokitok South
Engitati
Kidogo Spring
Mti Moja
Mystery Spring
Seneto
Vernonia
Gorigor Midwest
Gorigor North
Gorigor West
Munge Marsh
Munge River
1
2
marsh, grassland,
woodland
marsh, woodland
marsh
marsh, grassland
marsh, grassland
marsh, shrubland
marsh, grassland,
woodland
marsh, shrubland
marsh, grassland
marsh, grassland
marsh
marsh, grassland
marsh, grassland
21
22
23
24
25
26
27
28
29
30
31
32
Vegetation structure categories follow Pratt and Gwynne (1977).
Numbers refer to map locations in Fig. 2.
the rift escarpment is characterized by Acacia bushland and
woodland, interrupted by ephemeral streams draining into
the lake. The lacustrine plain contains patches of edaphic
grasslands and open mudflats near the lake edge.
Baboons (Papio anubis) live throughout Lake Manyara National Park, and vervet monkeys (Cercopithecus aethiops) are
common in the Acacia woodlands. Blue monkeys (Cercopithecus mitis) are present but rarely venture outside of the groundwater forest.
It is unclear what the natural role of fire would be in this environment, but fire may have been absent from Lake Manyara
National Park since at least 1958, and possibly since 1934 (Prins
and Van Der Jeugd, 1993, who cite a Warden’s Report and personal communication from A. Seif, a professional hunters’
guide in Manyara between 1934 and 1958). The vegetation history of Manyara prior to the 20th century is obscure, but fires set
by humans were likely to have been a factor. Ancient stromatolites about 20 m above the modern level of Lake Manyara
show high lake stands around 10,000e12,000 BP; 25,000 BP;
and around 90,000 BP (Casanova and Hillaire-Marcel, 1992).
The fluctuating lake levels over several thousand years mimic
152
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Fig. 2. Maps showing the modern study areas in Tanzania. Numbers refer to the study areas as listed in Table 2. Maps modified from Sinclair (1979b), Loth and
Prins (1986), and the Frankfurt Zoological Society (1971).
the conditions of paleolake Olduvai, which fluctuated over similar time scales (Hay, 1976; Deocampo, 2004).
Ngorongoro Crater
Ngorongoro Crater is a caldera within the Crater Highlands
immediately southeast of Serengeti and Olduvai and northwest
of Lake Manyara. The caldera is an oval bowl with a relatively
flat crater floor about 250 km2 in size (Herlocker and Dirschl,
1972), at an elevation of about 1,737 m (Hay, 1976). Ngorongoro
Crater formed almost 2 Ma in association with rift faulting (Hay,
1976), and is a closed basin fed by springs and streams. Colluvial
deposits of basalts, tuffs, and scoria cover most of the southern,
eastern, and northern parts of the caldera, and lacustrine deposits
dominate around the area of Lake Makat (Hay, 1976). The crater
floor receives an annual rainfall of 628e797 mm according to
Anderson and Herlocker (1973), but the figure varies greatly
from year to year and may be as low as 300 mm in the drier, western portion (Herlocker and Dirschl, 1972).
The vegetation of the crater floor is dominated by open
grassland, interrupted most conspicuously by the Lerai woodland in the southwest, the Ngoitokitok/Gorigor marsh in the
southeast with the Ngoitokitok pool and woodland fringing
its eastern edge, the Munge River lined by a sparse woodland
in the north, and the seasonally fluctuating, shallow Lake
Makat near the center of the crater (Fig. 2). In addition to
the large Ngoitokitok/Gorigor marsh, many smaller wetlands
associated with spring-fed seeps occur within the Crater.
Animals inhabiting the crater include elephants, hippos,
buffalo, lions, a year-round population of zebra and wildebeest,
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
baboons (Papio anubis), and vervet monkeys (Cercopithecus
aethiops).
Ngorongoro Crater has had protected status since 1921.
Since 1975, the Ngorongoro Conservation Area Authority
has administered the crater as part of a larger, 8,292 squarekilometer multiple land use area that allows for protection
from hunting for the wildlife, but also allows for use of lands
by the Masai for grazing their goats and cattle (Runyoro et al.,
1995). Although climate, grazing, burning, and soils all influence the vegetation of the crater, soil factors may be of the
greatest importance in determining the distribution and nature
of the vegetation types (Anderson and Herlocker, 1973). The
grasslands of Ngorongoro Crater are associated with poorly
drained soils and anaerobic conditions (Belsky, 1990). There
were periods during the late Pleistocene when the majority
of the crater floor was submerged by a lake, as evidenced by
high lake-level marks in the eastern portion of the crater
(Hay, 1976). Like Manyara, the expanding and contracting
lake at Ngorongoro may replicate the scenario at Olduvai of
episodically high and low lake levels over thousands of years.
For this study, plant samples were taken on the crater floor
in ten small wetlands, two large wetlands, and the ‘‘dry
lands’’, or non-saturated lands directly adjacent to all of those
wetlands. These serve as analogs for stream-fed and spring-fed
wetlands on Olduvai’s eastern/southeastern lacustrine plain.
Serengeti National Park
Serengeti National Park is immediately west of Olduvai
Gorge and the Crater Highlands. The gently rolling hills and
open grasslands of the Serengeti Plains constitute the southeastern portion of the park, while the rockier tree- and
shrub-dominated northern and western portions of the park
are known as the Serengeti Woodlands (Fig. 2).
The Serengeti Plains are comprised of very old (2.5 Ga)
rocks of the Tanganyika Shield overlain by layers of volcanic
ash (Hay, 1976). The plains are dotted with kopjes, protrusions
of granitic gneisses and quartzite that jut out from the volcanic
soil forming rocky islands. The soils, climate, and height and
composition of grasses and herbaceous species vary across
a roughly southeast-northwest gradient (Anderson and Talbot,
1965). Prevailing winds blow in a northwesterly direction,
which largely accounts for the rainfall and soil patterns. The
Crater Highlands draw moisture out of the air, which creates
a rain shadow in the vicinity of Olduvai Gorge and the Serengeti Plain. The strong rainfall gradient ranges from approximately 500 mm annual rainfall in the semi-arid southeastern
plains near Olduvai Gorge to > 800 mm rainfall per year in
the northernmost part of Serengeti National Park (NortonGriffiths et al., 1975). Volcanoes in the Crater Highlands
have been spewing volcanic dust into the atmosphere periodically for the past four million years (Hay, 1976), and the
winds consistently transport this volcanic dust to the northwest, dumping most of the volcanic ash in the southeastern
portion of the Serengeti. The short grasslands in the southeastern Serengeti Plain are edaphic and result from the shallow,
alkaline soils derived from sodic, carbonatite ash, underlain by
153
calcrete layers less than one meter below the surface (Anderson and Talbot, 1965; Hay and Reeder, 1978). The western
edge of the Serengeti Plains has a brown calcareous soil that
is lighter-textured, better-drained than the eastern soils, and
supports longer grasses (Anderson and Talbot, 1965). Rivers
in the plains have scattered trees and shrubs in contrast to
the surrounding grasslands. Most of the rivers in the Serengeti
drain west toward Lake Victoria, so there is also fluvial transport of volcanic materials westward.
The area known as the Serengeti Woodlands is formed by
Late Precambrian sedimentary rocks that unconformably overlie the Tanganyika shield (Hay, 1976). The hills and valleys of
this area have soils composed of basement rock-derived particles (granite and quartzite) mixed with volcanic ash (Sinclair,
1979a). The Serengeti Woodlands are comprised of a patchwork of mainly woodland and bushland, but also some grassland landscape facets. Wooded landscape facets in this area are
differentiated by topographic position and tree species composition, but most are dominated by different species of Acacia
trees (Herlocker, 1975).
Primates in Serengeti include baboons (Papio anubis) and
vervet monkeys (Cercopithecus aethiops), which I commonly
saw in the study areas. Black and white colobus monkeys (Colobus abyssinicus) occur in the riverine forests of western Serengeti and patas monkeys (Erythrocebus patas) are rarely
found in the northern corridor, but neither of these species
was observed in the vicinity of the facets studied here.
Portions of the Serengeti were established as a game reserve in 1929, as a protected area in 1940, and as a national
park in 1951 (Sinclair, 1995). Pastoralist tribes such the Masai
historically occupied the area, where they practiced little agriculture but tended cows and goats. Pastoralists often burn large
swaths of land, as they still do in places surrounding the park
today. Areas inside Serengeti National Park are burned in the
form of controlled fires set by park authorities. Pellew (1983)
estimated that 10% of the grassland areas of the park are
burned each year. Natural fires are a feature of the Serengeti
land region as well, given that lightning strikes are frequent
during thunderstorms of the wet seasons. It is not known
how long humans have set fires in the Serengeti, but they
may have done so since before the invention of agriculture
or the domestication of animals.
Within Serengeti National Park the vegetation was sampled
along two rivers in the Serengeti Woodlands, and at three paired
riverine/non-riverine sites in the Serengeti Plains (Fig. 2).
Sampling methods
The spatial extent of each individual modern landscape
facet was defined based on published landscape classification
systems and maps as well as field reconnaissance trips. In Loth
and Prins’ (1986) vegetation map of Manyara, the smallest
units are equivalent to landscape facets as they are ‘‘delineated
and described on the basis of landscape-forming factors such
as climate, geology, geomorphology, and soil characteristicsd
in addition to vegetation’’ (Loth and Prins, 1986:115).
Herlocker’s vegetation classification and map focusing on
154
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Serengeti’s woody vegetation uses mapping units that are
‘‘essentially the land facet of Gerresheim (1971)’’ (Herlocker,
1975:15). The published vegetation descriptions of the Ngorongoro Conservation Area by Herlocker and Dirschl (1972)
and Anderson and Herlocker (1973) were based on physiognomic categories and particularly the distribution of soil types.
The vegetation sampling strategy at Manyara and Serengeti
consisted of 50 50 m sampling plots in a stratified random distribution in each facet (Greig-Smith, 1983). For example, along
rivers in the Serengeti plots were located at a random point
within each 500 m stretch of the river. Riverine plots had one
side immediately adjacent to the channel bed itself. In larger, interfluvial landscape facets plots were placed in a similar stratified random manner along a series of parallel transects.
The 50 50 m square plot was used to sample trees
(>6 m), with nested plots 50 10 m for shrubs and 5 2 m
for herbaceous plants. For woody plants, the height, crown diameter, and count of individuals were recorded, and top cover
area was estimated for each plot. For herbaceous species, presence/absence data were recorded in each plot. The proportions
of actual vegetation that were sampled within each chosen
landscape facet ranged from about one to eight percent, based
upon the sum of the plot sizes divided by the estimated size of
the landscape facet.
The sampling strategy at Ngorongoro Crater was necessarily different from that of Serengeti and Manyara because many
of the landscape facets were much smaller or were wetlands
not amenable to transect sampling. At small wetlands, all plant
species were identified and percent cover area of each was visually estimated. For larger wetlands like Noitokitok/Gorigor,
percent cover was estimated for plants within the area in close
proximity, which was generally about 50 50 m.
In open grassland areas at Ngorongoro, 5 1 m plots were
used to identify ‘‘dry land’’ herbaceous species. The shrubland
areas near some of the small springs were also small enough to
identify all woody species present within the entire facet. In the
larger ‘‘dry land’’ woodlands associated with the large springs
at Ngoitokitok North and Ngoitokitok South, 50 50 m plots
with nested sub-plots were used as described for the Serengeti
and Manyara vegetation. At Ngoitokitok North and Ngoitokitok South, a similar proportion of the vegetation was sampled
as that of Serengeti and Manyara, between about one and eight
percent. Other Ngorongoro facets were sampled by 25e100%
of their total area because they were very small.
Three assistants from the National Herbarium in Arusha
were employed to help in the identification of plant species:
Vetes Kalema, Emanuel Mboya, and Daniel Sitoni. During
the field work at Manyara, one of them was present at any given
time and the small herbarium at Ndala Research Camp (Lake
Manyara) provided the first resource for plant comparison
and identification. Plant samples were assigned a sample number using the three letters of the study area followed by a number (e.g., BAR-1 for a plant specimen collected at Barafu in
Serengeti). The author collected specimens at Serengeti and
Ngorongoro, and those were brought to the National Herbarium
in Arusha for identification. In some cases plants were keyed
out with reference books (Haines and Lye, 1983; Blundell,
1987; Ibrahim and Kabuye, 1987; Coates Palgrave, 1993;
Beentje, 1994). Representative samples of the plant species
are stored in the Human Origins Laboratory at the National
Museum of Natural History in Arusha, Tanzania.
The modern analog vegetation studies were conducted in
Ngorongoro Crater during JulyeAugust 1995, which is the
dry season. In Manyara and Serengeti field work was conducted
from September 1997 through July 1998. Most study areas were
only sampled once, so seasonal comparisons are limited to general observations as opposed to quantitative measures.
The documentation and references for plants being edible
to humans, baboons, or chimpanzees can be found in Peters
et al. (1992), with additional references noted in the text.
Grasses with edible grains are recorded in the seed/pod category rather than as a ‘‘fruit’’.
Results
Table 3 summarizes which modern study areas were used to
model particular parts of the hypothetical lowermost Bed II Olduvai paleolandscape, and lists the main plant species that were
found in each of the modern study areas. A complete list of plant
species, the locations in which they were found, and the parts
that are known to be edible by humans, chimpanzees, or baboons
(based on Peters et al., 1992) are given in Appendix 1.
Vegetation and plant foods in the modern study areas by
hypothetical Olduvai landscape facet
1. Western Serengeti, riverine. Modern study areas: along
the eastern Mbalageti and Sangare Rivers in Serengeti National Park. These modern riverine facets are within the Serengeti Woodlands.
The dominant tree along both rivers is Acacia xanthophloea, which ranges in height up to 18 meters. The second
most common trees are Acacia kirkii (up to 12 m) along the
Mbalageti and Acacia robusta (up to 9 m) along the Sangare.
In the non-riverine facets away from the bushland fringing the
river, the vegetation is dominated by smaller trees, which are
mainly Acacia clavigera (Herlocker, 1975), but the non-riverine facets were not systematically sampled here, so a description will be limited to riverine facets.
The Somalia-Masai phytochoria, which includes northern
Tanzania, Kenya, and Somalia, has 30 endemic species of Acacia (White, 1983). Since they are legumes, mature Acacia trees
produce ‘‘beans’’ whose pods and seeds have varying degrees
of edibility, and sometimes the leaves and flowers are eaten.
Almost all Acacia species tend to ooze gum from wounds in
their trunks, and Acacia gums can be a valuable resource, as
most are non-poisonous (Story, 1958).
Humans and baboons have been recorded to eat the gum of
Acacia robusta, while baboons also eat the seeds, shoots, and
flowers. The recorded uses of Acacia kirkii are by humans who
make tea and rope from the bark, medicines from the gum, and
use the wood for fuel (Marcan, 1998). Acacia xanthophloea
pods and seeds are eaten by baboons, but only rarely, and Acacia xanthophloea trees full of fruits are often ignored by
155
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Table 3
Vegetation and potential plant foods at modern localities used as analogs for the Olduvai lowermost Bed II paleolandscape
Olduvai facets Modern analog
being modeled study areas1
Serengeti
Woodlands
riverine
Serengeti Woodlands:
Mbalageti River (14),
Sangare River (15)
Serengeti Plain:
Serengeti
peneplain
Barafu Plain (12),
non-riverine Seronera-Wandamu
interfluve (17), Nyamara
interfluve (16)
Description of modern vegetation, noting potentially edible species2
Bushland. Main tree species: Acacia xanthophloeaSP,BC, Acacia kirkiiBC, Acacia robustaFl,SP,L,BC
Other woody species: Albizia gummifera, Albizia harveyiFl,SP, Cordia monoicaFr, Ficus sp.Fr, Phyllanthus
fischeri, Hibiscus ovalifoliusSP,L, Abutilon mauritianumFl,SP, Grewia forbesiiFr,Fl,L, Grewia tembensisFr
Common herbaceous species: Justicia matammensisW, Justicia striataFl,L, Hypoestes forskaleiL, Achyranthes
asperaL, Sporobolus consimilisSP,L,U, Pennisetum mezianum, Panicum maximumSP,L,St
Grassland. Main grass species: Sporobolus iocladosSP and Sporobolus fimbriatusSP in short-intermediate
grasslands (Barafu); Pennisetum mezianum and Themeda triandraFl,SP,L,St in long grasslands (SeroneraWandamu, Nyamara)
Woody species are rare, but include Balanites aegyptiacaFr,SP,BC, Commiphora schimperi, Acacia
tortilisFl,SP,L,BC, Hibiscus micranthusFr, Dichrostachys cinereaL,BC
Other common herbaceous species: Harpachne schimperi, Leucas neuflizeana, Indigofera volkensii, Melhania
ovata?, Crotalaria spinosa, Crotalaria polyspermaL, Rhynchosia minimaSP, Solanum incanumFr,L,St, Tephrosia
pumilaSP, Sida ovataL, Craterostigma plantagineum
Riverine
Serengeti Plain:
Barafu Valley (13),
Seronera River (19),
Nyamara River (18)
Bush grassland. Main tree species: Acacia xanthophloeaSP,BC, Acacia tortilisFl,SP,L,BC
Other woody species: Commiphora africanaFr,L,St,BC,U, Commiphora merkeri, Commiphora schimperi,
Balanites glabra, Phyllanthus fischeri, Ficus sycomorusFr,L (rare), Aspilia mossambicensisL, Hibiscus
micranthusFr, Acalypha volkensii, Aerva lanataL, Acacia drepanolobiumSP
Common herbaceous species: Indigofera coluteaSP, Achyranthes asperaL, Hirpicium diffusum, Monechma
debileFl,L, Themeda triandraFl,SP,L,St, Sporobolus fimbriatusSP, Pennisetum mezianum
Western
lacustrine
plain
No analogs studied
NA
Eastern
lacustrine
plain
stream-fed
wetland
Ngorongoro:
Gorigor Midwest (28),
Gorigor North (29),
Gorigor West (30),
Munge Marsh (31),
Munge River (32)
Marsh. Main species: Cyperus immensusU, Cyperus laevigatusU
Other species: Cyperus laxus, Diplachne fusca, Typha latifoliaFl,U, Scirpus inclinatus, Aeschynomene
schimperi, Hydrocotyle mannii, Hydrocotyle sibthorpioides, Ludwigia sp., Panicum cf. repens
Stream-fed
‘‘dry land’’
Ngorongoro:
Gorigor Midwest (28),
Gorigor North (29),
Munge Marsh (31),
Munge River (32)
Short grassland. Dominated variously by Chloris gayana, Cynodon dactylonSP,W, Sporobolus spicatusSP,
Cyperus laevigatusU, Cyperus rotundusU, Pennisetum clandestinum, Persicaria decipiensL, Themeda
triandraFl,SP,L,St
Small spring Ngorongoro:
Engitati (22),
wetland
Kidogo Spring (23),
Mti Moja (24),
Mystery Spring (25),
Seneto (26),
Vernonia (27)
Small marshes. Main species: Cyperus immensusU, Scirpus inclinatus, Cyperus laevigatusU
Other species in the marshes: Cynodon dactylonSP,W , Cyperus laxus, Cyperus sesquiflorus, Digitaria
macroblephara, Diplachne fusca, Hydrocotyle sibthorpioides, Sphaeranthus suaveolens,
Themeda triandraFl,SP,L,St, Typha latifoliaFl,U
Small spring Ngorongoro:
‘‘dry land’’ Kidogo Spring (23),
Mti Moja (24),
Mystery Spring (25),
Seneto (26),
Vernonia (27)
Small areas of woodland, shrubland, and grassland. Main woody species: Acacia xanthophloeaSP,BC
Other woody species: Euphorbia candelabrumFl,L, Justicia betonica, Lippia ukambensisSP, Ocimum suave,
Senna didymobotrya, Vangueria madagascariensisFr, Vernonia myrianthaFl,L,St.
Main grasses: Cynodon dactylonSP,WP, Pennisetum clandestinum
Other herbaceous species: Achyranthes asperaL, Aristida adscensionis, Cassia angustifolia, Chloris gayana,
Chloris pycnothrix, Crotalaria sp., Digitaria milanjianaSP, Eragrostis arenicola, Launaea cornutaL, Leonotis
nepetifoliaFl,L, Lippia javanicaL, Lotus goetzei, Solanum incanumFr,L,St, Solanum sp., Sphaeranthus suaveolens,
Sporobolus spicatusSP
Large spring Ngorongoro:
wetland
Ngoitokitok North (20),
Ngoitokitok South (21)
Large marsh. Main species: Cyperus immensusU, Typha latifoliaFl,U, Phragmites mauritianusL,St, Cyperus
papyrusSt
Other species along marsh edges: Crassocephalum picridifolium, Crassocephalum vitellinumL, Cynodon
dactylonSP,W, Cynodon nlemfuensis, Cyperus laevigatusU, Cyperus rotundusU, Cyperus sesquiflorus,
Diplachne fusca, Eragrostis tenuifolia, Hydrocotyle sibthorpioides, Lotus arabicus, Ludwigia stolonifera,
Persicaria senegalensisL,U, Scirpus inclinatus, Sphaeranthus suaveolens, Stegnogramma pozoi, Vigna
schimperi, Vigna vexillataFl,SP,L,U
(continued on next page)
156
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Table 3 (continued )
Olduvai facets Modern analog
being modeled study areas1
Description of modern vegetation, noting potentially edible species2
Large spring Ngorongoro:
‘‘dry land’’ Ngoitokitok North (20),
Ngoitokitok South (21)
Woodland. Main tree species: Acacia xanthophloeaSP,BC
Other woody species: Justicia betonica, Capparis tomentosaFr, Vangueria madagascariensisFr, Croton
macrostachyusFr, Abutilon mauritianumFl,SP, Albizia gummifera, Calpurnia aurea, Conyza newii, Cordia
monoicaFr, Helinus mystacinus, Lippia ukambensisSP, Ocimum suave, Senna obtusifoliaSP, Tagetes minuta,
Vernonia galamensis
Common herbaceous species: Achyranthes asperaL, Lippia javanicaL, Setaria verticillata, Cynodon
dactylonSP,W, Bidens pilosaL, Vigna schimperi
Upper,
Manyara lacustrine plain:
non-riverine Msasa lake flat (5),
Ndala lake flat (4)
Bush grassland. Main tree species: Acacia tortilisFl,SP,L,BC
Other woody species: Acacia xanthophloeaSP,BC, Hyphaene petersianaFl,Fr,SP,St,BC, Salvadora persicaFr,L,BC,
Capparis tomentosaFr, Maerua triphyllaFr,U, Acalypha fruticosaFr,St, Justicia cordata, Aeschynomene indica
Common herbaceous species: Peristrophe bicalyculata, Gutenbergia polytrichotoma, Justicia flavaL,
Commelina africanaW, Achyranthes asperaL, Sida albaFl,SP,L, Tephrosia villosa, Indigofera arrectaU, Ocimum
basilicumL, Indigofera coluteaSP, Sporobolus cordofanus, Cynodon dactylonSP,W, Dactylotenium aegyptiumSP,U
Riverine,
perennial
Manyara lacustrine plain:
Mkindu River on
lacustrine plain (6)
Riverine forest. Main tree species: Acacia xanthophloeaSP,BC
Other woody species: Senna bicapsularis, Rauvolfia caffra, Tabernaemontana ventricosa, Acalypha
fruticosaFr,St, Phoenix reclinataFr,L,BC, Cordia goetzei, Trichilia emeticaFr
Herbaceous species (rare): Achyranthes asperaL, Eclipta prostrataW
Riverine,
ephemeral
Manyara lacustrine plain:
Msasa River on
lacustrine plain (7)
Bushland. Main tree species: Trichilia emeticaFr, Acacia tortilisFl,SP,L,BC
Other woody species: Acalypha fruticosaFr,St, Cordia sinensisFr,U, Capparis tomentosaFr, Maerua triphyllaFr,U,
Cadaba farinosaL,St, Salvadora persicaFr,L,BC, Thilachium africanumFl,U, Hibiscus ovalifoliusFl,SP,L
Common herbaceous species: Sida albaFl,SP,L, Abutilon bidentatum, Bidens schimperiL, Achyranthes asperaL
Lacustrine
Manyara lacustrine
terrace
terrace:
non-riverine Ndala-Chemchem
interfluve (8),
Ndilana-Msasa
interfluve (9)
Riverine
Major rock
outcrops
Manyara lacustrine
terrace: Msasa River on
lacustrine terrace (10),
Ndilana River on
lacustrine terrace (11)
Bushland. Main tree species: Acacia tortilisFl,SP,L,BC
Other woody species: Acacia robusta subs. usambarensisFl,SP,L,BC, Balanites aegyptiacaFr,SP,BC, Ficus
sycomorusFr,L (rare), Acalypha fruticosaFr,St, Cordia sinensisFr,U, Maerua triphyllaFr,U, Salvadora persicaFr,L,BC,
Cadaba farinosaL,St, Cordia monoicaFr, Acalypha indicaL, Vepris uguenensis, Gardenia ternifoliaFr, Hibiscus
ovalifoliusFl,SP,L
Common herbaceous species: Monechma debileFl,L, Peristrophe bicalyculata, Achyranthes asperaL, Indigofera
arrectaU, Gutenbergia cordifolia, Bidens pilosaL, Commelina africanaW, Digitaria velutinaSP,W, Setaria
sagittifolia
No analogs studied
NA
Alluvial fans Manyara alluvial fan:
non-riverine Mkindu interfluve (1)
Riverine
Manyara alluvial fan:
Mkindu River on
alluvial fan (2)
Mountainsides No analogs studied
1
Bushland. Main tree species: Acacia tortilisFl,SP,L,BC
Other woody species: Maerua triphyllaFr,U, Salvadora persicaFr,L,BC, Acalypha fruticosaFr,St, Cadaba
farinosaL,St, Justicia stachytarphetoides, Cordia monoicaFr, Vepris uguenensis, Pavonia patens, Hibiscus
ovalifoliusFl,SP,L, Acalypha indicaL
Common herbaceous species: Achyranthes asperaL, Gutenbergia polytrichotoma, Monechma debileFl,L,
Commelina africanaW, Indigofera arrectaU, Peristrophe bicalyculata, Cyathula orthacantha
Forest. Main tree species: Trichilia emeticaFr, Tabernaemontana ventricosa
Other woody species: Cordia goetzei, Croton macrostachyusFr, Ficus sycomorusFr,L, Acalypha fruticosaFr,St,
Acalypha ornataFr,L, Senna bicapsularis, Hibiscus ovalifoliusFl,SP,L, Celtis zenkeri, Acalypha indicaL
Common herbaceous species: Malvastrum coromandelianum, Hypoestes forskaleiL, Achyranthes asperaL
Forest. Main tree species: Tabernaemontana ventricosa, Rauvolfia caffra, Ficus sycomorusFr,L, Trichilia
emeticaFr
Other woody species: Hibiscus ovalifoliusFl,SP,L, Senna bicapsularis, Acalypha ornataFr,L, Acalypha fruticosaFr,St,
Phoenix reclinataFr,L,BC, Hippocratea paniculata, Celtis zenkeri, Ocimum suave
Common herbaceous species: Malvastrum coromandelianum, Justicia glabra, Achyranthes asperaL
NA
Numbers refer to map localities in Fig 2.
2
Species that are potentially edible to hominins because they are documented as having been eaten by humans, chimpanzees, or baboons (Peters et al., 1992) are
labeled with superscripts indicating the edible plant parts: Fl ¼ flowers/inflorescenses; Fr ¼ fruits; SP ¼ seeds, pods, grass grains; L ¼ leaves, shoots; St ¼ stems;
BC
¼ bark, cambium, gum; U ¼ underground part; W ¼ whole plant; ? ¼ unspecified.
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
baboons (Wrangham and Waterman, 1981). Acacia xanthophloea flowers, which appear in the late dry season, are not
consumed by baboons and only rarely by vervet monkeys,
and, therefore, are not considered as potentially edible to hominins. Vervet monkeys and baboons are particularly fond of
eating Acacia xanthophloea gum, which contains over 50%
carbohydrates (sugars), and less than 3% (dry weight) total
phenolics and condensed tannins (Wrangham and Waterman,
1981; Hausfater and Bearce, 1976). In sum, the main resource
provided by Acacia xanthophloea is an edible gum, though its
pod, seed, and bark have been known to be eaten rarely by
baboons.
Other woody plants along the Mbalageti and Sangare
Rivers include Ficus sp., Cordia monoica, Grewia forbesii,
and Grewia tembensis, all of which have edible fruits. Several
of the herbaceous species are known to be edible to humans,
including two species of Justicia and Hypoestes forskalei
with edible leaves. Three of the grass species present are edible to baboons, including Panicum maximum, the grain of
which is also eaten by humans and chimpanzees.
2. Serengeti peneplain, riverine and non-riverine. Modern
study areas: along and near the the Barafu Valley, eastern Seronera River, Wandamu River, and Nyamara River in Serengeti National Park.
These modern study areas are in the Serengeti Plains, in
which non-riverine areas are open grassland and along the
ephemeral rivers the physiognomic structure can be classified
as bush grassland (Pratt and Gwynne, 1977). The vegetation
differs somewhat between the easternmost sampling area in
the vicintiy of the Barafu Valley, which supports short-intermediate
grasslands, and the western sampling areas near the Seronera,
Wandamu, and Nyamara rivers, which support long grasslands.
Anderson and Talbot (1965) recognized these differences as resulting from soil variations due to the combined factors of the
degree of wind erodability, soil depth, texture, and salt concentration, which in turn determine the water storage capacity of the
soils and the availability of moisture to plants.
The non-riverine samples at Barafu are open grasslands 1e
2 km away from the Barafu Valley and are heavily grazed and
dominated by Sporobolis ioclados and Sporobolus fimbriatus,
both of which have potentially edible grains and seed heads
(Peters et al., 1992). Other grasses occur in less abundance,
along with numerous small forbs, most commonly the inedible
species Leucas neuflizeana and Indigofera volkensii.
In the Barafu Valley itself, a small spring seep emerges and
the ephemeral stream is lined with scatted Acacia tortilis and
Commiphora africana trees and shrubs. Humans have been observed to eat the pods (not the seeds) of Acacia tortilis, for example, by pounding them into flour and mixing that with milk
for porridge (Birnie, 1997). Most humans living around Acacia
tortilis trees today do not consider them edible. Rather, the
trees provide firewood, food for domestic stock, and shade
in otherwise relatively open habitats. The flowers of Acacia
tortilis, which appear in the late dry season, are eagerly consumed by primates during this time of general resource scarcity (Wrangham and Waterman, 1981). Baboons and vervet
monkeys only rarely eat Acacia tortilis gum, which contains
157
less than 1% carbohydrates, and greater than 27% total phenolics and condensed tannins (Hausfater and Bearce, 1976;
Wrangham and Waterman, 1981). Thus, the main potential
hominin foods provided by Acacia tortilis are pods and
flowers, although the leaves, gum, bark, and seeds have been
observed to be consumed by baboons and/or humans in rare
instances.
Shrubs of the genus Commiphora are common in the Somalia-Masai phytochoria (White, 1983), are typically thorny, and
most produce a pungent resin. Commiphora africana is the African myrrh tree, referring to the valuable spice that derives
from its resin (Marcan, 1998). Commiphora africana is one
of the most common shrubs around Olduvai Gorge today.
Humans eat the fruit, leaf, and stem, the gum of Commiphora
africana is a chewing gum, and the roots can be eaten like
a cassava substitute (Marcan, 1998). Baboons eat the fruit as
well.
Other, less common woody plants in the Barafu Valley include Commiphora merkeri (inedible), Hibiscus micranthus,
and Acacia drepanolobium. Like many Hibiscus species in
Africa, Hibiscus micranthus provides some edible parts for
primates as humans have been observed to eat the fruits. Humans also eat the young green pods of Acacia drepanolobium.
Ground-level forbs and small woody plants in the Barafu Valley were dominated by Indigofera colutea, of which baboons
eat the edible pods, and inedible Hirpicium diffusum.
In the western Serengeti plains, the long grasslands are
composed mainly by Pennisetum mezianum and Themeda triandra. These rolling open plains are interrupted by the ephemeral
Seronera, Wandamu, and Nyamara rivers. Acacia xanthophloea
was by far the most common tree lining the Seronera and
Nyamara rivers, along which the riverine plots were located.
Shrubs were also present along these rivers, and included the
inedible Phyllanthus fischeri as well as Aspilia mossambicensis, the leaves of which are eaten by chimpanzees, but are
more likely consumed for medicinal rather than nutritional
purposes (Wrangham and Nishida, 1983).
A few of the grasses in the plains are eaten by humans and/
or baboons, such as the common Themeda triandra, and the
less common Sporobolus fimbriatus, Sporobolus ioclados,
and Panicum maximum. In these open grassland habitats, a variety of small forbs and small woody plants are intermixed
with the grasses. In the Nyamara Interfluve, common species
were Craterostigma plantagineum and Solanum incanum. Humans have been reported to eat the leaf of Solanum incanum
and baboons eat the yellow tomato-like fruits, but the modern
Masai people living around Olduvai Gorge regard the fruits as
poisonous (personal observation). Another common non-grass
plant in the Seronera-Wandamu interfluve was Crotalaria spinosa, which is not known to be edible. The very few shrubs
and trees encountered in the long-grass plains included Balanites aegyptiaca (the desert date) with edible fruits, Acacia
tortilis, Commiphora schimperi, and Hibiscus micranthus.
3. Eastern lacustrine plain, stream-fed wetlands and adjacent ‘‘dry land’’. Modern study areas: Ngorongoro Crater,
small stream-fed wetlands at Munge Marsh, Munge River,
Gorigor Midwest, Gorigor North, and Gorigor West.
158
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
These modern habitats in Ngorongoro Crater include relatively small, localized wetlands and fringes of the large Gorigor wetland, all of which are fed by surface waters. North of
Lake Makat, the Munge River becomes a wide pool and marsh
in which hippos live. The area is surrounded by grassland. The
hippos maintain an area of open water through bioturbation.
This is the Munge Marsh, dominated by Cyperus immensus
with large patches of Scirpus inclinatus and Typha latifolia.
Herbaceous species growing around the pool include Pennisetum clandestinum (inedible), Polygonum salicifolium with
leaves that are edible to humans, and the C4 dwarf sedge Cyperus rotundus (Peters and Vogel, 2005). Cyperus rotundus
has an edible underground bulb (humans and baboons), but
this part may become woody and inedible when it is mature
(Peters, 1994).
The Munge River plots were located near the Munge River’s entrance into Lake Makat, where small patches of marsh
plants occurred within the ephemeral stream, mainly Cyperus
immensus and Cyperus laevigatus. The surrounding short
grassland is Pennisetum clandestinum and Cynodon dactylon.
In the western extremities of Gorigor much of the standing water evaporates during the dry season and the water is
more saline and alkaline. pH ranged from 8.5 to 10.1 at the
time the plant samples were taken (Deocampo and Ashley,
1999). At Gorigor West the small sedge Cyperus laevigatus
(height less than 0.5 m) is most common. There are also occurrences of Cyperus immensus and the grasses Diplachne
fusca, Cynodon dactylon, and Sporobolus spicatus. At Gorigor Midwest the marsh contains Cyperus immensus and Cyperus laevigatus. At Gorigor North the marsh plants are
Cyperus immensus, Cyperus laevigatus, and Typha latifolia.
In all cases, the ‘‘dry lands’’ adjacent to these stream-fed
wetlands are well-grazed short grassland with grasses such as
Chloris gayana, Cynodon dactylon, and Sporobolus spicatus.
Cyperus immensus and Cyperus laevigatus are the two most
common sedges in the smaller wetlands of Ngorongoro Crater.
Cyperus immensus is a C4 sedge, and is described as providing
‘‘morsels’’ to eat, including the base of the young leaf and the
base of the young culm (Peters and Vogel, 2005). Cyperus laevigatus is common on saline ephemerally wet soils throughout
sub-Saharan Africa (Smith, 1976; Ellery et al., 1993), and its
stem base (bulb) is reportedly eaten by baboons (Peters et al.,
1992). Typha latifolia, or cattail, grows in brackish water and
fresh water environments. It has large, pleasant-tasting edible
rootstocks and nutritious edible pollen (Prendergast et al.,
2000). Typha latifolia is less tolerant to salinity than Cyperus
immensus, as the former occurs in only the freshest of the
small wetlands around the crater, but in abundance in the large
freshwater Ngoitokitok spring (see below). Scirpus inclinatus
is a common sedge in the crater, but it is not reported to be
edible.
The grass Diplachne fusca can grow in waterlogged or dry
soils and can tolerate high soil salinity levels, up to
400 mol m3 NaCl (Warwick and Halloran, 1991). Cynodon
dactylon is a very common grass enjoyed by grazers that often
grows on subsaharan African floodplains and lake flats. Sporobolus spicatus is a short, very salt-tolerant grass species
which is often associated with the sedge Cyperus laevigatus
(Ellery et al., 1993).
4. Eastern lacustrine plain, small spring wetlands and adjacent ‘‘dry land’’. Modern study areas: Ngorongoro Crater
small springs at Seneto, Vernonia, Mti Moja, Kidogo Spring,
Engitati, and Mystery Spring. These modern study areas are
all individual small springs that emerge in the crater, some
near the crater wall and some closer to the center of the crater,
near Lake Makat.
Seneto spring is adjacent to the western Crater wall. The central and main portion of the marsh (100 150 m) is dominated
by Typha latifolia and Cyperus immensus. This dense wetland
area is surrounded by open water and sparse clumps of Scirpus
inclinatus and Cyperus laevigatus. The surrounding short grassland is mainly Cynodon dactylon. The marsh edge, especially
near the crater wall, is scattered with Acacia xanthophloea trees.
On the lower slopes of the crater wall immediately adjacent to
this spring are some cactus-like Euphorbia candelabrum trees,
of which baboons eat the shoots and flower buds.
Adjacent to the southern crater wall, Vernonia Spring supports a small marsh about 25 50 m in size, which is dominated by two sedges, Scirpus inclinatus (<1 m) and Cyperus
sesquiflorus (<0.5 m), as well as the salt-tolerant grass Diplachne fusca. Overhanging the spring head wall is a small
patch of shrubs, mainly Vernonia myriantha, of which humans
eat the leaf and chimpanzees eat the flower and the pith, and
Vangueria madagascariensis, which produces sweet edible
fruits that are fig-like in size and taste.
The spring known as Mti Moja emerges on a small rise
directly adjacent to the northeastern lacustrine plain of Lake
Makat. The groundwater flows about 700 meters out onto the
playa flats (Deocampo and Ashley, 1999), and the marsh is concentrated within the proximal 25 300 m area. Closest to the
spring head, the marsh plants are mainly the sedge Cyperus immensus, and further downstream the marsh is dominated by the
shorter and more saline-tolerant Cyperus laevigatus. The spring
is surrounded by short grassland, including Sporobolus spicatus
and Cynodon dactylon. There is also a dense patch of the tall
grass Sporobolus consimilis nearby, of which baboons eat the
grain, leaf, and rhizome (Peters et al., 1992). According to
park rangers, in recent years there were two Acacia xanthophloea trees growing on the spring head, a claim which is supported by sections of weathering tree trunk in the vicinity.
Kidogo is a series of spring seepages near the western edge
of Lake Makat. The small marshes (about 10 10 m each)
contain almost exclusively the dwarf sedge Cyperus laevigatus, incidating saline soils (Smith, 1976; Ellery et al., 1993).
On the surface there are basalt rocks (some of which are artifacts) and the surrounding vegetation is short grassland,
mainly Cynodon dactylon.
In the northern part of the crater floor, a spring-fed marsh
(1 2 km) near Engitati hill is dominated by Cyperus immensus, Scirpus inclinatus, Cyperus laxus, and Cyperus laevigatus.
The surrounding ‘‘dry land’’ is heavily grazed short grassland
with Cynodon dactylon and Themeda triandra.
Mystery Spring is on the western edge of the crater floor
adjacent to the crater wall. The central marsh (15 30 m) is
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Cyperus immensus with Scirpus inclinatus at the edges, and
footprinted mud surrounding the wetland. The surrounding
dry land is mainly short grassland, but there are three patches
of shrubs, including Senna didymobotrya (inedible), Solanum
incanum, Lippia javanica (humans use the leaf to make tea),
and five Acacia xanthophloea trees.
5. Eastern lacustrine plain, large spring wetlands and adjacent ‘‘dry land’’. Modern study areas: Ngorongoro Crater large
spring at Ngoitokitok North and Ngoitokitok South. The Ngoitokitok-Gorigor system is a large wetland in the southern crater
floor covering a total area of about ten square kilometers, although the area of waterlogged soil varies greatly with seasons
and from year to year. The water emerges in a series of springs
along its eastern margin creating a perennial freshwater pool.
This eastern margin is the locality of Ngoitokitok North and
Ngoitokitok South, where the pH of the water ranges from
6.9 to 7.5 (Deocampo and Ashley, 1999), and supports a tall, extensive marsh and a large woodland on adjacent dry land.
The central area of the perennial Ngoitokitok marsh, which
can be up to three meters in height, is characterized by stands
of four marsh plants: Typha latifolia, Cyperus immensus,
Phragmites mauritiana, and Cyperus papyrus. Phragmites
mauritiana, a large grass, is a fresh water-loving marsh plant
with pith eaten by chimpanzees and baboons, and shoots eaten
by chimpanzees and humans. Cyperus papyrus can tolerate
a relatively wide range of hydrological regimes, but requires
perennial water and in the Okavango Delta it is restricted to
low-conductivity waters not exceeding 45e60 micromhos/
cm1 (Smith, 1976). Cyperus papyrus has edible parts that include rhizomes, the base of new aerial shoots, and the mature
culm base (Peters, 1999; Peters and Vogel, 2005). Peters and
Vogel (2005) considered papyrus to be potentially the most
important C4 plant food for African early hominins, as it can
provide relatively substantial amounts of food compared to
other C4 plants. In this study, Cyperus papyrus was found
only in the Ngoitokitok marsh at Ngorongoro Crater.
Further south and west, and corresponding to increasing
levels of pH, conductivity, and salinity away from the fresh water spring source, the marsh plants in the Ngoitokitok-Gorigor
system change to the more salt- and alkaline-tolerant sedges
and smaller growth forms of Typha latifolia and Cyperus immensus, and eventually to Scirpus inclinatus and Cyperus laevigatus.
The woodland along the eastern edge of the NgoitokitokGorigor system was sampled in two areas: Ngoitokitok North
and Ngoitokitok South. Ngoitokitok North is a relatively low,
flat area with an Acacia xanthophloea woodland and a very
sparse understory of scattered shrubs, mainly Justicia betonica
and Tagetes minuta, neither of which are edible. Herbaceous
species include Vigna schimperi (inedible) and Cynodon dactylon. The woodland at Ngoitokitok South abuts the crater wall
and most of the land surface there is two to four meters higher
in elevation than the woodland at Ngoitokitok North. At Ngoitokitok South, Acacia xanthophloea trees grow next to the
marsh edge, while several other tree and shrub species grow
on the higher ground. These include Croton macrostachyus,
the broad-leaved cotton tree with fruits eaten by chimpanzees,
Capparis tomentosa shrubs, the fruits of which are eaten by
159
baboons and cooked and eaten by humans as a famine food
(Peters et al., 1992), and Vangueria madagascariensis and
Cordia monoica, both shrubs with pleasant-tasting fruits (personal observation).
6. Upper eastern lacustrine plain non-riverine areas, along
perennial rivers, and along ephemeral rivers. Modern study
areas: Lake Manyara lacustrine plain along and near the
Mkindu, Msasa, and Ndala Rivers. On Manyara’s upper lacustrine plain, the non-riverine facets of Msasa lake flat
and Ndala lake flat are each about 0.2 2 km2 in size.
The vegetation of the non-riverine facets can be classified as
bush grassland (Pratt and Gwynne, 1977) with sparse tree and
shrub cover and many herbaceous species. The scattered trees
are mainly Acacia tortilis, but there are also a few Acacia xanthophloea and Hyphaene petersiana palm trees. Of palm trees
native to sub-Saharan Africa, at least 17 species in six genera
produce edible fruits (Peters et al., 1992). Hyphaene petersiana
has fruits that are known to be eaten by foragers in southern
Africa (Story, 1958) and by the Gwembe Tonga of Zimbabwe
(Scudder, 1962). The fruits are slightly smaller than a tennis
ball, and have a two to three cm thick edible mealy layer between the skin and the inner kernel (Story, 1958). Humans
also eat the seed (raw), the flower nectar, and the palm heart,
and baboons eat the fruits (Peters et al., 1992).
Shrubs in non-riverine facets of Manyara’s upper lacustrine
plain included Capparis tomentosa, Salvadora persica, Maerua
triphylla, and Acalypha fruticosa, all of which are also common
in Manyara’s lacustrine terrace (see below). Salvadora persica
shrubs produce clusters of tiny, juicy, sweet edible fruits during
the dry season. Salvadora persica is one of the most desired
supplement foods of the foraging/agriculturalist Sandawe of
Tanzania, being eaten in significant quantities by individuals
who are in pursuit of other activities such as herding livestock
or collecting firewood (Newman, 1975). Stems of Salvadora
persica can be broken off and used to clean the teeth, hence
the shrub’s common name, the toothbrush tree (Beentje,
1994). Several species of the genus Maerua have fruits that
are edible to humans or baboons (Peters et al., 1992), and
Maerua triphylla has long (2e7 cm), cylindrical edible fruits
and edible roots. However, both parts are labeled ‘‘may be poisonous’’ from East African herbarium records (Peters et al.,
1992), and, therefore, may need treatment such as cooking or
soaking before they are edible to modern humans. Acalypha
fruticosa is a small shrub, typically less than two meters tall,
of which baboons are known to eat the small (2 3 mm), dehiscent fruits and the pith (Peters et al., 1992).
The non-riverine area of Manyara’s upper lacustrine plain
was particularly rich in herbaceous species, in part because
it was sampled in the late rainy season. A total of 69 non-grass
herbaceous species were recorded in the two study areas, 35 of
which have at least one part that is edible to humans, chimpanzees, or baboons. Some of the more common forbs were the
inedible species Peristrophe bicalyculata, Gutenbergia polytrichotoma, and Tephrosia villosa, and edible Indigofera arrecta,
Ocimum basilicum (basil, which may be exotic; Marcan,
1998), and Cyperus usitatus, a dwarf C4 sedge with an edible
bulb (Peters and Vogel, 2005).
160
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
The two riverine facets that were studied on Manyara’s upper lacustrine plain differed in vegetation structure and composition because the Mkindu River is a small, perennial
stream, whereas the Msasa River is a large channel that only
flows seasonally. The perennial lower Mkindu is about two
meters across and is fed by a spring further upstream in the
alluvial fans. On the upper lacustrine plain it is lined with
a dense, narrow strip of Acacia xanthophloea trees up to
20 m in height. A few inedible Rauvolfia caffra trees were
also present. Smaller woody plants included the inedible
Senna bicapsularis and Tabernaemontana ventricosa, and the
edible shrub Acalypha fruticosa. The Mkindu River takes its
name from the Swahili word for the Phoenix reclinata palms
that occur there. The fruit, palm heart, and sap of these palms
are eaten by humans, and the fruit and soft new shoots are
eaten by baboons (Peters et al., 1992). Herbaceous species
were relatively sparse and included Achyranthes aspera,
with leaves that are edible to humans and baboons, and Eclipta
prostrata, the whole plant of which is eaten by humans as a
potherb (Peters et al., 1992).
The ephemeral lower Msasa River is as wide as 25 m where
it crosses Manyara’s upper lacustrine plain. The vegetation
along its banks can be classified as bushland (Pratt and
Gwynne, 1977), where the trees are mainly Trichilia emetica
and Acacia tortilis. Trichilia emetica has seed arils known to
be eaten by humans and baboons (Peters et al., 1992). The
bright red arils can be peeled or bitten off the black seeds,
but the seeds themselves may be poisonous to humans (Mbuya
et al., 1994). Most of the common shrubs along the lower
Msasa River had edible parts, such as Cordia sinesis, Capparis
tomentosa, Maerua triphylla, Salvadora persica, and Thilachium africanum. In the Capparaceae family, Thilachium
africanum has roots that can be boiled and eaten, and this species was used as a famine food by the Sandawe in years of low
crop production (Newman, 1975). Baboons are also known to
eat the flowers and flower buds of Thilachium africanum
(Peters et al., 1992). Common herbaceous species were the
inedible Abutilon bidentatum, Sida alba (with leaves eaten
by humans and seeds and flower buds eaten by baboons),
and Bidens schimperi, of which humans eat the bitter leaves
and young shoot (Peters et al., 1992).
7. Lacustrine terrace, riverine and non-riverine. Modern
study areas: Lake Manyara lacustrine terrace along and near
the Ndala, Chemchem, Msasa, and Ndilana Rivers.
The lacustrine terrace of Lake Manyara is characterized by
bushland (Pratt and Gwynne, 1977) in non-riverine areas and
along the ephemeral rivers. Two non-riverine facets were
studied: one between the Ndala and Chemchem Rivers
(2 0.8 km2) and one between the Ndilana and Msasa Rivers
(3 1 km2). The riverine facets were located along the Msasa
River (2 km 50 m on each side of the river) and the Ndilana
River (3 km 50 m on each side of the river).
In the non-riverine facets of the lacustrine terrace the main
tree species was Acacia tortilis, which grew as much as 14 m,
and the most common shrubs included the inedible Vepris
uguenensis, and the edibles Acalypha fruticosa, Maerua triphylla, Salvadora persica, and Cadaba farinosa, of which
humans eat the leaf and stem (Peters et al., 1992). Herbaceous
species were numerous in these bushland habitats. Edible species included Achyranthes aspera, Monechma debile, of which
humans eat the leaf and baboons eat the leaf and flower, and
Commelina africana, of which humans eat the whole plant
as a potherb (Peters et al., 1992). Common inedible herbaceous species included Gutenbergia polytrichotoma and Peristrophe bicalyculata.
In the riverine facets of Manyara’s lacustrine terrace, the
dominant tree species was again Acacia tortilis, although
a few Balanites aegyptiaca and Acacia robusta trees were encountered as well, both of which provide edible parts. Most of
the common shrubs also had edible parts, including Acalypha
fruticosa, Maerua triphylla, Salvadora persica, Cadaba farinosa, Acalypha indica, and Cordia sinensis. Acalypha indica
has leaves that are edible to humans. Cordia sinensis is one
of the three most important fruit species that provide food
for the Hadza foragers of Tanzania (Vincent, 1985), who
live within 200 km of Olduvai Gorge. Cordia sinensis grows
as a shrub or small tree and produces a tasty, ovoid yellow/orange fleshy fruit that is about 1 cm in diameter. The root is also
reported to be edible raw by humans and baboons eat the fruits
as well (Peters et al., 1992).
Common herbaceous species in the riverine lacustrine terrace at Manyara were very similar to those in the non-riverine
facets. The inedibles included Peristrophe bicalyculata and
Gutenbergia cordifolia, and the edibles Achyranthes aspera,
Monechma debile, and Commelina africana.
8. Alluvial fans, riverine and non-riverine. Modern study
areas: alluvial fans at Lake Manyara along and near the Mkindu
River. The small, spring-fed, perennial Mkindu River crosses
the alluvial fans in the northern portion of the Lake Manyara
National Park. The vegetation was virtually the same in riverine and non-riverine facets of this area, as the ubiquitous
groundwater clearly had a more dominant effect on most of
the plants than the presence of localized surface water. Therefore, the vegetation of riverine and non-riverine facets in Manyara’s alluvial fan/groundwater forest will be discussed together.
The most common large tree is Trichilia emetica, the Cape
mahogany, that grows to heights of up to 20 meters. Although
the fruits of Trichilia emetica, with their edible seed arils, typically grow at great heights in the canopy, many fall to the forest floor where they can be eaten on the spot or collected for
later consumption. Other trees included the inedibles Cordia
goetzii and Tabernaemontana ventricosa, Croton macrostachyus, which has fruits that are eaten by chimpanzees in western Tanzania (Kano, 1972), and Ficus sycomorus with its tasty
figs and edible leaves (Peters et al., 1992). Common shrubs
were Acalypha fruticosa and Acalypha ornata, of which chimpanzees in western Tanzania eat the fruits and leaves (Van
Lawick-Goodall, 1968; Nishida and Uehara, 1983). Also common were the small shrubs Hibiscus ovalifolius and Senna
bicapsularis. Herbaceous species were relatively rare on the
shaded forest floor but included Achyranthes aspera and
Malvastrum coromandelianum. One of the rare grasses that occurred in the groundwater forest was the shade-loving, inedible
Oplismenus hirtellus (Ibrahim and Kabuye, 1987).
161
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Table 4
Number of plant species providing potentially edible parts for hominins (Appendix 1) in the modern study areas
Modern study area
Fruit
w
Serengeti
Serengeti Woodlands
riverine
Serengeti Plain
non-riverine
riverine
Ngorongoro Crater
Stream-fed
wetland
‘‘dry land’’
Small spring
wetland
‘‘dry land’’
Large spring
wetland
‘‘dry land’’
Manyara
Lacustrine plain
non-riverine
riverine, perennial
riverine, ephemeral
Lacustrine terrace
non-riverine
riverine
Alluvial fan
non-riverine
riverine
h
Flower/
inflorescence
Seed/pod
Leaf/shoot
Stem
Bark/
cambium/gum
w
h
w
h
w
h
w
h
w
Underground
part
unspecified
h
w
h
w
h
17
6
6
4
7
5
13
9
3
4
6
0
3
3
1
0
4
17
3
7
2
4
3
9
3
8
13
15
3
14
11
21
0
2
5
9
4
6
0
1
0
2
5
10
0
0
1
1
0
0
1
0
0
0
1
1
0
0
1
3
0
0
0
2
0
0
0
1
0
0
0
0
0
0
3
3
0
0
0
0
0
1
1
1
0
2
2
1
0
2
2
3
0
2
1
5
0
1
1
1
0
1
0
0
0
0
4
1
0
0
0
0
0
4
1
0
0
1
2
1
0
4
2
2
0
1
4
5
0
0
2
2
0
1
0
0
0
0
7
2
0
0
0
0
7
6
7
10
1
0
4
0
4
10
1
3
5
1
3
16
1
5
4
3
6
32
2
6
2
1
2
6
1
2
4
2
3
2
0
0
2
0
3
8
1
3
0
0
0
0
0
0
8
16
7
8
4
5
8
7
5
7
12
11
5
8
24
25
2
2
6
6
3
4
0
0
3
3
8
9
0
0
0
0
9
8
1
0
3
1
0
0
3
1
0
0
7
4
4
3
1
1
0
0
2
1
0
0
1
0
0
0
0
0
0
0
Note: w ¼ woody plants, h ¼ herbaceous plants.
Comparison of edible plants by landscape facet
Table 4 summarizes the number of plant species with parts
potentially edible by hominins in each of the modern landscape
facet types. The large variation in numbers of potentially edible
species reflects major differences in the overall species richness, or alpha diversity, of each facet. This variation is expected
given the size area differences of facets. For example, some of
the small springs are only 10 10 m2 each, whereas Manyara’s
non-riverine lacustrine terrace covers an area of 3 1 km2. The
wetlands have particularly low total and edible species richness, even in the large Ngoitokitok-Gorigor wetland. Species
richness is not a measure of actual plant food abundance, although it might be important in terms of the value of a facet
to particular hominins. For example, the greater variety of plant
food types along most riverine facets could support a variety of
specialized foragers or generalists, whereas only hominins who
were particularly interested in marsh plants would likely forage
in the wetlands of the lacustrine plain.
Figure 3 shows the proportion of all species in each facet
type that contain particular edible plant parts. In general, woody
species tend to provide higher proportions of potentially edible
fruits. For example, in the bushland habitats along the rivers of
the Serengeti Woodlands, Manyara’s lacustrine terrace, and
Manyara’s lacustrine plain, edible fruit-producing shrubs included Grewia, Salvadora, Cordia, Vangueria, Cadaba, and
Maerua. In Manyara’s alluvial fans (the groundwater forest),
large, edible fruit-producing trees included Trichilia, Croton,
and Ficus. Only woody plants that grew in the dry lands adjacent to small springs provided a lower proportion of edible
fruits than other parts. Sept (1990, 1994) found at Semliki
that woody, fruit-bearing species were frequent in wooded
grasslands and extremely frequent in bushland.
Among herbaceous plants, relatively more species with edible underground parts occur in wetland facets, particularly
marsh plants with edible rootstocks, bulbs, and corms. Leaves
are the most common edible plant part from herbaceous species in non-wetland habitats, with the exception of the two
grassland habitats (non-riverine Serengeti Plains and the Ngorongoro ‘‘dry lands’’ adjacent to stream-fed wetlands). Grasslands have more edible parts in the seeds/pods category due
to the higher proportion of potentially edible grass species.
Just 22 of the plant species recorded are edible to baboons
only. If plant species that are recorded as being eaten only by baboons are removed from the data (since they may not be ‘‘edible’’
to hominoids), the pattern remains almost unchanged (Fig. 3).
Plant food reconstruction of lowermost Bed II, Olduvai
The following reconstruction of vegetation and plant foods
assumes that the modern analogs described above provide an
accurate representation of the landscape facets and vegetation
of the lowermost Bed II paleolandscape. The degree to which
this claim can be substantiated is considered later in the paper.
162
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Woody
Woodland
Plain
small
spring
Ngorongoro
20%
40%
60%
80%
100%
riverine
streamfed
Serengeti
0%
non-riverine
riverine
wetland
Fr
dry land
Fl
wetland
SP
dry land
large
spring
L
wetland
St
dry land
BC
lacustrine
plain
non-riverine
U
riverine, perennial
?
lacustrine
terrace
non-riverine
alluvial
fan
Manyara
riverine, ephemeral
non-riverine
riverine
riverine
Herbaceous
Woodland
Plain
small
spring
large
spring
Ngorongoro
20%
40%
60%
80%
100%
riverine
streamfed
Serengeti
0%
non-riverine
riverine
wetland
dry land
wetland
dry land
wetland
Fr
Fl
SP
L
St
dry land
BC
lacustrine
plain
non-riverine
U
riverine, perennial
?
lacustrine
terrace
alluvial
fan
Manyara
riverine, ephemeral
non-riverine
riverine
non-riverine
riverine
Fig. 3. Relative proportions of edible plant parts in the modern landscape facets from woody and herbaceaous species. Species counts are shown in Table 4. Edible
plant part abbreviations in the legend are: Fr ¼ fruits; Fl ¼ flowers/inflorescenses; SP ¼ seeds, pods, grass grains; L ¼ leaves, shoots; St ¼ stems; BC ¼ bark, cambium, gum; U ¼ underground part; ? ¼ unspecified.
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
The seasonally inundated lake flats on the lowest portions
of Olduvai’s paleolake margin were virtually devoid of vegetation when the lake level dropped temporarily during dry seasons. Exceptions to this were localized marshes where small
springs emerged near the lake shore, or where spring-fed perennial rivers drained into the lake. The marsh plants in these
wetlands likely had small amounts of edible rootstocks, bulbs,
and corms, such as those from Cyperus immensus and Cyperus
laevigatus. If the water was fresh enough, or the spring or
stream large enough, stands of Typha latifolia might have
occurred, providing juicy rootstocks and nutritious pollen at
these localized wetlands.
The upper lacustrine plain, at least on the eastern/southeastern side of the paleolake, contained a mosaic of marsh, patches
of bushland or woodland, heavily grazed edaphic grasslands,
and tree-lined streams. Marshes occurred at springs and along
the lower reaches of rivers. Large spring-fed pools similar to
that of Ngoitokitok might have provided drinking water that
was fresher than that in the lake, and if stands of fresh water
were perennial they may have sustained a Cyperus papyrus
marsh from which hominins could obtain edible rhizomes,
shoots, and culms. Such a large spring-fed system would likely
support a localized woodland on adjacent higher ground with
Acacia xanthophloea trees and perhaps fruiting shrubs and
other large seed/pod- or fruit-bearing trees (e.g., Balanites,
Croton, Phoenix, Justicia, Vangueria). The open grasslands
on the upper lacustrine plains included grass species with
edible seeds and stem bases, such as Cynodon dactylon and
Sporobolus spicatus. After the rains a flush of forbs appeared
in the undergrowth, many of which had edible leaves and underground parts (e.g., Commelina, Achyranthes, Sida, Indigofera). Perennial streams crossing the upper lacustrine plains
were probably lined with a narrow Acacia xanthophloea woodland, perhaps including other woody species, like the palms
Phoenix reclinata and Hyphaene petersiana, which contain edible fruits, palm hearts, and sap (Peters et al., 1992). Such
palms also may have existed in clumps away from rivers, as
occurs on Manyara’s upper lacustrine plain and as indicated
by the phytolith evidence from lowermost Bed II (Albert
et al., 2006; Bamford et al., 2006).
A lacustrine terrace landform might have existed around
some portions of the paleo-Olduvai basin reflecting old high
lake levels. A lacustrine terrace with well-drained soils similar
to the one at Manyara would have supported a bushland or woodland with Acacia tortilis trees (edible pods), and a high diversity
of edible fruit-bearing shrubs such as Cordia, Maerua, Salvadora, Capparis, Cadaba, Acalypha, and Hibiscus. Ephemeral
rivers crossing the lacustrine terrace had a greater density of
woody plants and potential edibles than the non-riverine areas,
but both were richer in edible plant species than the small facets
on the upper lacustrine plain. In the wet season, a flush of herbaceous growth in the lacustrine terrace would have provided
potentially edible leaves, stems, and small underground parts.
Alluvial fans occurred along the eastern and southern portions of the paleo-Olduvai basin in the lower reaches of the
Crater Highlands. If groundwater from the highlands made
its way close to the surface in at least some portions of the
163
lower alluvial fans, then those areas of Olduvai could have
supported a groundwater forest like the one at Lake Manyara.
In such areas one would expect tall trees (>20 m) with a closed
canopy, including trees with edible fruits, such as Trichilia and
Ficus. The undergrowth would be sparse due to the shady environment. Groundwater forest would also be expected along
perennial streams crossing the alluvial fans.
Let us assume for now that some of the areas presently to
the west of Olduvaidthe Serengetidare good analogs for
those areas in the past. Both short grasslands with underlying
calcretes, such as those near Barafu, and longer grasslands in
deeper soils, like those near Seronera would have been challenging places for a hominin searching for plant foods unless
it was interested in eating grasses. Ephemeral streams crossing
the grasslands, however, would have provided a greater diversity of potential plant foods, including seasonally available
fruits (e.g., Commiphora, Hibiscus), seeds/pods (e.g., Acacia
tortilis, Acacia drepanolobium), and leaves from woody
plants, gum from Acacia xanthophloea trees, and some herbaceous foods. Further west in the Serengeti Woodlands, where
the outcropping basement rock contributed to the soil in addition to recent volcanic material, the rivers were lined with a diversity of edible part-bearing woody plants such as Acacia
robusta and Grewia shrubs. Non-riverine habitats in this
area were also wooded with various species of Acacia.
Discussion
Climate change within lowermost Bed II
Seasonal variations in plant food availability would have
been of central importance to ranging patterns of foraging hominins since most plant parts are available for very limited times
of the year. Many Acacia trees produce flowers and pods during
the dry season, providing edible resources that were probably
concentrated in the lacustrine plain and lacustrine terrace zones
of the paleo-Olduvai basin. In fact, most shrubs within Manyara’s lacustrine terrace bushland are reported to vegetate and
flower during the dry season (Greenway and Vesey-Fitzgerald,
1969), and this was observed in the current study for Salvadora
persica. Edible roots from shrubs such as Maerua triphylla and
Cordia sinensis, and from the herb Indigofera arrecta, also common in the lacustrine terrace, are probably available at any time
of year. The rootstocks and edible underground parts of sedges
and Typha are available year-round as well, but might be best
during the early dry season when the water levels recede but
the plants still retain firm underground parts that have not dried
out. The most common tree in Manyara’s alluvial fan, Trichilia
emetica, produces its edible fruits during the wet season, and the
tree Ekebergia capensis also has edible fruits during the wet season. Individual trees of Ficus sycomorus produce figs at different
times throughout the year (Coates Palgrave, 1993). In grassland
facets, the edible parts of grasses and other herbaceous plants
appear almost exclusively during the wet season.
Over the longer time period of approximately 50e85 ka
during which lowermost Bed II was deposited, the distribution
of landscape facets would have altered due to climatic changes
164
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
associated with the Earth’s orbit (Hay, 1976; deMenocal,
1995). The stratigraphy within lowermost Bed II records multiple expansions and contractions of the lake reflecting periods
of wetter and drier climate. The lake level fluctuations are represented in the eastern lake margin by alternating strata of
waxy and earthy claystones (Hay, 1976, 1996; Deocampo
et al., 2002), in the lowest alluvial fans by the presence of
clay lake deposits near the base and near the top of two sections
(Ashley and Driese, 2000), and in the central basin of paleolake
Olduvai by changes in ultrafine clay chemistry as represented
by the octahedral cation index (Deocampo, 2004). At a regional
scale, the faulting that occurred with rift formation also caused
shifts in lake levels during the deposition of lowermost Bed II
(Hay, 1996).
How would the landscape facets have changed from extended periods of low lake level to extended periods of high
lake level? The distribution of paleolandscape facets as shown
in Fig. 1 represents the landscape during a relatively dry period in which the upper lacustrine plain is rarely flooded. During wet phases, the lake extended to its maximum size, near
the boundary of the upper lacustrine plain or lacustrine terrace
for several hundred years or more. The eastern lake margin
was only exposed for use by hominins and accumulation of artifacts for a portion of lowermost Bed II deposition.
During a wetter climate with an expanded lake, the catena
of landscape facets did not simply shift together to higher elevations. Different lithologies across the basin may have
prevented this shift. Lake deposits from Bed II are about
90% claystone, 5% sandstone, and 5% tuff, and eastern lake
margin deposits are 64% claystone, 18% siliceous earthy claystone and sandstone, and 15% tuff (Hay, 1976:67). The alluvial fan deposits in the east, however, are 89% tephra
deposits such as reworked tuffs and lapilli tuffs, only 6% claystone, and 4% conglomerate (Hay, 1976:73). In times of lake
expansion, the tuffaceous soils of the lower alluvial fans became lake margin areas, and the type of vegetation supported
there differed from that which flourishes in a clayey lake margin zone. The lacustrine plain and its vegetation as described
here probably shrank to a very narrow zone during periods
of lake expansion. Within the alluvial fans, only the vegetation
of the lowermost fans was affected directly by the periodic
inundation of saline-alkaline lake waters. However, higher
groundwater tables throughout the alluvial fans during wet periods favored vegetation that thrives in moister habitats. The
rivers crossing the alluvial fans became larger and less ephemeral or perennial in a wet climate due to higher rainfall in the
highlands, with subsequent expansion of riverine forest extent
and density. Non-riverine areas that depended directly on rainfall for moisture might have supported more densely wooded
environments as well.
How would plant resources for hominins have changed
from times of dry climate to times of wetter climate? During
wetter climates, the shrinking and/or disappearing of the mosaic of eastern/southeastern lacustrine plain facets resulted in
fewer edible rootstocks of marsh plants, fewer edible seeds/
pods from Acacia trees, less edible lacustrine plain grasses
during the wet season, and fewer edible underground parts
from shrub roots of the lacustrine terrace. For example, the localized marsh/wetland that existed in the area of HWK-E and
its associated marsh plant foods and woodland were completely inundated by the lake. The larger variety of plant foods
in higher parts of the basin, such as tree fruits of riverine
forests, were still available in the alluvial fans during wet climates, perhaps in greater abundance than during dry climates.
Similarly, riverine and interfluvial plant foods in the western
basin may have increased in variety during wetter climates.
How accurate are the modern analogs?
All three of the modern regions used as analogs in this study
have annual rainfall amounts that are slightly lower than the
800 mm suggested by isotope and pollen evidence for lowermost
Bed II (Bonnefille and Riollet, 1980; Cerling and Hay, 1986).
For Lake Manyara and Ngorongoro Crater this does not necessarily mean that the modern habitats are more ‘‘open’’ than
they would be if there was higher rainfall, because their grasslands are edaphic. The grasslands on these lacustrine plains
are present because of the saline-sodic nature of the soils
(Anderson and Herlocker, 1973; Belsky, 1990). Edaphic grasslands in East Africa are also strongly associated with poorly
drained soils, such as clay-rich soils at the bottoms of soil catenas
(Belsky, 1990). Soils on the ancient lacustrine plain at Olduvai
were both saline-sodic and clay-rich, at least close to the lake,
so lacustrine edaphic grasslands were probably present. The
presence of reduncine antelopes in lowermost Bed II supports
this conclusion (Marean and Ehrhardt, 1995; Cushing, 2002).
Ample paleoenvironmental evidence suggests that during
Bed I and lowermost Bed II times, Olduvai’s eastern lake margin contained a mosaic of habitat types in addition to edaphic
grassland, including wetlands and areas with trees and/or
shrubs (e.g., Kappelman, 1984; Potts, 1988; Shipman and
Harris, 1988; Plummer and Bishop, 1994; Peters and Blumenschine, 1995, 1996; Kappelman et al., 1997; Fernandez-Jalvo
et al., 1998; Albert et al., 2006; Bamford et al., 2006). The modern lacustrine plains of Ngorongoro Crater and Lake Manyara
contain a similar mosaic of habitat types, and the freshwater
wetlands in Ngorongoro Crater may be particularly good analogs for Olduvai’s paleowetlands. Wetlands in both the crater
and the paleo-Olduvai basin are characterized by concentrations of biogenic silica surrounded by clays with high carbonate
and zeolite content reflecting the dominant saline-alkaline lake
deposits (Deocampo and Ashley, 1999).
The grasslands of the eastern Serengeti Plains exist because of the soils, which are shallow (30e100 cm deep),
highly alkaline (pH 8.0e10.0), and are underlain by calcrete
(Belsky, 1990). The calcrete, which helps to prevent the
growth of trees, formed intermittently during the deposition
of the Ndutu Beds over a period of perhaps a few hundred
thousand years, beginning 0.05e0.04 Ma (Hay and Reeder,
1978). The calcium carbonate that formed this calcrete derived from the carbonatite ash of Ol Doinyo Lengai, which
is the currently active volcano but which was inactive during
lowermost Bed II times (Hay, 1976; Hay and Reeder, 1978).
Olmoti, the volcano active during lowermost Bed II times, is
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
not thought to have produced carbonatite ash (Hay, 1976). Older
calcretes do exist in the Olduvai Beds. In the Lemuta Member of
Bed II (stratigraphically above lowermost Bed II) there are discontinuous massive calcretes, and in Bed I massive calcretes
0.6e1.0 m thick are indicated for the western part of the paleo-Olduvai basin, but neither are as thick or well-developed
as the recent Ndutu calcrete (Hay and Reeder, 1978).
Thus, it is likely that during lowermost Bed II times, the
area that is now the eastern Serengeti Plains was not underlain
by calcrete and, therefore, was potentially amenable to tree
growth. Consequently, the modern vegetation study sites in
the eastern plains near the Barafu Valley may be poor analogs
for the lowermost Bed II paleolandscape. However, the grasslands near the Seronera and Nyamara Rivers at the western
edge of the Serengeti Plains are in brown soils with calcareous
concretions but do not overlie a hardpan (Anderson and
Talbot, 1965), and, therefore, may still serve as useful analogs
for areas west or north of the paleo-Olduvai basin.
Although they were not included in this study, areas in the
Serengeti’s northern extension might serve as good analogs for
ancient Serengeti habitats because there is no underlying calcrete (Belsky, 1984) and the rainfall is about 700e800 mm
(Norton-Griffiths et al., 1975). The character of the vegetation
in this area depends on rainfall, grazing/browsing pressure,
fire, and soils, and the vegetation has fluctuated between grassland and woodland/bushland within the past century (Sinclair,
1979b; Belsky, 1990). The two Serengeti Woodland sites from
this study along the eastern Mbalageti River and along the
Sangare River may also be appropriate analogs for ancient
Serengeti habitats in terms of soil type and climate.
An important control on the ratio of herbaceous to woody
plant biomass in semi-arid landscapes of East Africa is the
amount and spatial distribution of soil moisture, which locally
derives either from rainfall or groundwater. In riverine habitats, the size, volume of discharge carried, and degree to which
the river is ephemeral all help control vegetation by influencing the amount of water available to plants (Hughes, 1990;
Coughener and Ellis, 1993). In areas away from rivers,
groundwater or the redistribution of regional rainfall is an
important factor controlling plant-available water and, therefore, the structure of the vegetation (Greenway and VeseyFitzgerald, 1969; Belsky, 1990; Coughener and Ellis, 1993).
The particularly high groundwater table in the northern area
of Lake Manyara supports the groundwater forest on the alluvial
fan, and a moderately high groundwater table may support the
Acacia tortilis-dominated bushland on the well-drained soils of
the lacustrine terrace (Greenway and Vesey-Fitzgerald, 1969).
Loth and Prins (1986) claim that the Acacia tortilis trees depend
on rainfall, but these interpretations are not at odds because the
level of the groundwater also depends on rainfall. The particular
redistribution of rain water at Lake Manyara is tied to its geological setting in which the rift escarpment is in close proximity to
the lake, a situation which is not paralleled in the Olduvai
paleolandscape. However, Lake Manyara’s example serves to
illustrate that the particular nature of the re-distribution of Crater
Highlands rainfall into the paleo-Olduvai basin (i.e., through
rivers and/or groundwater flow) was probably the single most
165
important factor influencing the character of the vegetation and
the distribution facets across the basin’s paleolandscape.
Beyond Olduvai’s eastern paleolake margin, the paleoenvironmental evidence diminishes rapidly along with the degree
of confidence with which modern analogs can be specified
and ancient vegetation reconstructed. Paleoenvironmental evidence from lowermost Bed II is consistent with the presence of
at least some patches of forest, but it is very unlikely that the
entire alluvial fan area was a Manyara-type groundwater forest, because then we would expect to see more forest-adapted
fauna among the fossils. In an analysis of faunal samples from
a landscape sample of lowermost Bed II, Cushing (2002:143)
found no strictly forest taxa and an overwhelming bias toward
open-habitat dwellers for terrestrial taxa overall. Reddishbrown, root-marked claystones with analcime suggest saline,
alkaline soils in the lowermost portions of the alluvial fans
(Hay, 1976), which might have deterred woody plant growth
in non-riverine areas. These lines of evidence suggest that
away from channels the vegetation graded into bushland,
bush grassland, or grassland habitats. Nonetheless, Manyara’s
alluvial fan groundwater forest may serve as an appropriate
analog for localized groundwater forests and/or riverine forests in the lowermost Bed II paleolandscape. Manyara’s
groundwater forest is similar in structure and species composition to large riverine forests of East Africa (Hughes, 1988). If
the rivers that crossed Olduvai’s ancient alluvial fan, lacustrine
terrace, or the northern and western sides of the basin were perennial, for example, then they were almost certainly lined
with riverine forest. Until there is more precise geological information concerning the size and nature of paleorivers in lowermost Bed II, however, the appropriateness of the particular
modern analog rivers studied here is difficult to judge.
In all modern analog habitats, the way in which humaninduced fire may have changed the vegetation is extremely complex, because fire has such an interdependent relationship with
factors like browsers (particularly elephants), grazers, soils, local perturbations in climate, the time of season in which the fires
occur, and the fact that ‘‘unnatural’’ fires may have been ongoing
for greater than 10,000 years (Norton-Griffiths, 1979; Frost and
Robertson, 1987; Dublin, 1995). If major changes in floristic
composition have occurred because of fire (or other factors),
then the nature of those changes would best be deciphered
from ancient flora of the type found in pollen cores that record
vegetation changes in the recent past over thousands of years.
Most pollen cores from East Africa are from mountain lakes
(e.g., Coetzee, 1967; Livingstone 1967; Hamilton, 1972), and
therefore record vegetation changes in the highlands. The
ways in which lowland vegetation of East Africa (e.g., nonAfromontane areas) changed during the late Quaternary and
Holocene are still poorly understood.
Conclusion
The current study attempts to reconstruct vegetation and plant
foods at a particular locality, Olduvai Gorge, using modern ecosystems as models. Although the modern analogs were chosen
carefully, no analog will ever have the same combination of plant
166
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
species that occurred in a given paleolandscape. On the other
hand, without the use of modern analogs our understanding of
paleovegetation will remain rudimentary and virtually useless
in terms of interpreting early hominin paleoecology. Details available only in modern settings, such as the nature of the edible
plants in terms of seasonality, nutritional value, abundance, and
requirements for processing and consumption, are our only guide
for such features of ancient landscapes.
A realistic hope is that if modern analogs can be identified
that are similar enough to a fossil setting in terms of the factors
that control vegetation, such as climate, topography, and soil
types, then the vegetation of the modern analog will be similar
in fundamental ways to the past vegetation, even if the actual
species composition is different. The proposed similarities include the structure of the vegetation and the types of plant
foods that are available. The best way to confirm or deny
such similarities is through plant fossils, such as pollen, phytoliths, and macrofossils. Plant fossils are also invaluable
because of their potential to identify elements not present in
modern analogs; for example, the identification of Guibourtia
coleosperma at Olduvai (Bamford, 2005).
Whereas initially paleoenvironmental proxies direct us to
potentially relevant modern analogs, the modern analogs can
in turn direct paleoenvironmental inquiries if important factors
that control relevant aspects of the modern vegetation (e.g.,
plant food type) are identified. Ideally, modern analog studies
of vegetation should seek to identify such ‘‘controlling factors’’
in modern environments, although it is likely that many interdependent factors, such as soil nutrient availability, pH, rainfall,
and so on, are involved. O’Brien (1988, 1993) and O’Brien
and Peters (1991) have accomplished this on a regional scale
by using climate to predict woody edible species richness
across southern Africa. At a landscape scale, those predictive
capabilities would help to interpret archaeological remains
and paleoecological contexts of specific hominin groups or individuals. Although that lofty goal will be difficult to achieve, it
is hoped that the information documented here helps to make
reconstructions of vegetation and plant foods at Olduvai more
realistic and more tangible for non-plant specialists.
Acknowledgements
This research was supported by NSF doctoral dissertation
grant #9728984 and the Center for Human Evolutionary Studies at Rutgers University. Field work was carried out with the
kind permission of the Commission on Science and Technology, Tanzania; the Ngorongoro Conservation Area Authority;
and the Tanzania National Parks Association. This project
was inspired and supported by the Olduvai Landscape Paleoanthropology Project (OLAPP) and directors Rob Blumenschine and Fidelis Masao. Vetes Kalema, Emanuel Mboya,
and Daniel Sitoni of the National Herbarium in Arusha were
essential in plant identification, both in the herbarium and in
the field. Thanks for fieldwork help and companionship from
Amy Cushing, Joanne Tactikos, Goodluck Peter, Jim Ebert,
Dan Deocampo, Joseph Masoy, and many others. Thanks to
Bill McGrew and Matt Sponheimer for comments on the original draft. Peter Andrews and three anonymous reviewers
made thorough critiques that significantly changed and improved the manuscript. Special thanks also to the pioneers of
actualistic plant research on whose foundations this work is
based: Charles Peters, Eileen O’Brien, and Jeanne Sept.
Appendix 1. Plant species encountered in the modern study areas in Serengeti, Lake Manyara, and Ngorongoro Crater
Family or
subfamily
Species
G1
Modern landscape facets2
Malvaceae
Malvaceae
Malvaceae
Malvaceae
Malvaceae
Malvaceae
Malvaceae
Mimosoideae
Mimosoideae
Mimosoideae
Mimosoideae
Mimosoideae
Mimosoideae
Mimosoideae
Mimosoideae
Mimosoideae
Mimosoideae
Mimosoideae
Euphorbiaceae
Euphorbiaceae
Euphorbiaceae
Euphorbiaceae
Euphorbiaceae
Abutilon angulatum (Guill. and Perr.) Mast
Abutilon bidentatum Hochst. ex A. Rich.
Abutilon grandiflorum G. Don
Abutilon hirtum (Lam.) Sweet
Abutilon longicuspe Hochst. ex A. Rich.
Abutilon mauritianum (Jacq.) Medik.
Abutilon sp. Mill.
Acacia albida Delile
Acacia brevispica Harms
Acacia drepanolobium Harms ex Sjostedt
Acacia kirkii Oliv.
Acacia lahai Steudel and Hochst. ex Benth.
Acacia robusta Burch.
Acacia robusta subsp. usambarensis (Taub.) Brenan
Acacia schweinfurthii Brenan and Exell
Acacia senegal (L.) Willd.
Acacia tortilis (Forssk.) Hayne
Acacia xanthophloea Benth.
Acalypha crenata Hochst. Ex A. Rich.
Acalypha fruticosa Forssk.
Acalypha indica L.
Acalypha ornata Hochst. ex A. Rich.
Acalypha sp. L.
w
h
w
w
w
w
w
h
w
w
w
w
w
w
w
w
w
w
h
w
w
w
w
14
5e7, 9, 11
11
9, 11
11, 15
8e11, 13e15, 21
1, 5, 9, 14
5
15
13
14, 18
5
1, 14, 15
7, 10
11
16
4, 5, 7e11, 13, 14, 16, 18
5, 6, 13e15, 18e21, 25, 26
15
1, 2, 4e11, 14, 15
1, 4, 5, 9e11
1, 2, 11, 15, 18
14
Edible parts (eaten by)3
SP (H)
SP (H), SP, Fl (B)
SP (H), SP, BC (B)
? (B)
SP (H)
BC (H)
BC (H), Fl, SP, L, BC (B)
BC (H), Fl, SP, L, BC (B)
BC (H), Fl, SP, L, BC (B)
SP, L, BC (H), Fl, SP, L (B)
SP, BC (B)
Fr, St (B)
L (H)
Fr, L (C)
167
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Appendix 1 (continued )
Family or
subfamily
Species
G1
Modern landscape facets2
Euphorbiaceae
Compositae
Compositae
Amaranthaceae
Acalypha volkensii Pax
Acanthospermum hispidum DC.
Acanthospermum sp. Schrank
Achyranthes aspera L.
w
h
h
h
Amaranthaceae
Papilionoideae
Papilionoideae
Compositae
Mimosoideae
Mimosoideae
Scrophulariaceae
Aloaceae
Poaceae
Commelinaceae
Orchidaceae
Moraceae
Poaceae
Compositae
Acanthaceae
Acanthaceae
Malvaceae
Balanitaceae
Balanitaceae
Acanthaceae
Acanthaceae
Labiatae
Labiatae
Compositae
Compositae
Acanthaceae
Acanthaceae
Compositae
Nyctaginaceae
Nyctaginaceae
Capparaceae
Capparaceae
Poaceae
Aerva lanata (L.) Juss. ex Schult.
Aeschynomene indica L.
Aeschynomene schimperi Hochst. ex A. Rich.
Ageratum conyzoides L.
Albizia gummifera (J.F. Gmel.) C.A. Sm.
Albizia harveyi E. Fourn.
Alectra vogelii Benth.
Aloe volkensii Engl.
Andropogon greenwayi Napper
Aneilema petersii (Hassk.) C.B. Clarke
Ansellia sp. Lindl.
Antiaris toxicaria Lesch.
Aristida adscensionis L.
Aspilia mossambicensis (Oliv.) Wild
Asystasia schimperi T. Anders.
Asystasia sp.
Azanza sp. Alef.
Balanites aegyptiaca (L.) Delile
Balanites glabra Mildbr. and Schltr.
Barleria eranthemoides R. Br.
Barleria submollis Lindau
Becium capitatum (Baker) Agnew
Becium sp. Lindl.
Bidens pilosa L.
Bidens schimperi Sch. Bip.
Blepharis maderaspatensis (L.) Heyne ex Roth
Blepharis panduriformis Lindau
Blumea mollis (D. Don) Merr.
Boerhavia diffusa L.
Boerhavia erecta L.
Boscia mossambicensis Klotzsch
Boscia salicifolia Oliv.
Brachiaria deflexa (Schumach.) C.E. Hubb.
ex Robyns
Brachiaria scalaris Pilg.
Brachiaria serrifolia (Hochst.) Stapf
Brachyachne patentiflora (Stent and Rattray)
C.E. Hubb.
Buglossoides arvensis (L.) I.M. Johnst.
Cadaba farinosa Forssk.
Caesalpinia trothae Harms
Calotropis procera (Aiton) W.T. Aiton
Calpurnia aurea Baker
Capparis fascicularis DC.
Capparis sepiaria L.
Capparis sp. L.
Capparis tomentosa Lam.
Cardiospermum halicacabum L.
Carissa edulis (Forssk.) Vahl
Cassia angustifolia Vahl
Celtis africana Burm. f.
Celtis zenkeri Engl.
Cenchrus ciliaris L.
Cenchrus sp. L.
Centella asiatica (L.) Urb.
Chamaecrista mimosoides (L.) Greene
Chloris gayana Kunth
Chloris pycnothrix Trin.
Chloris virgata Sw.
w
w
h
h
w
w
h
w
h
h
w
h
h
w
h
h
h
w
w
h
h
h
h
h
h
h
h
h
h
h
w
w
h
14, 15, 18
4, 5, 7e11
8
1, 2, 4e11, 13e15, 18,
19, 21, 25, 26
13, 14, 18
4
32
4
14, 15, 21
14, 15, 18
13
13
20
6
14
1
4, 9, 10, 12, 13, 19, 24
14, 15, 18, 19
12, 13, 18
14
10
9, 11, 13, 16, 18
13, 14
7, 9e11
7, 13, 21
12, 13
12, 16
1, 4, 5, 8e11, 21
7, 10, 11, 19
9, 11
18
10
8e11
4, 5, 9, 10
14
14
4, 5, 9e11
h
h
h
4
9
13
h
w
w
w
w
w
w
w
w
h
w
h
w
w
h
h
h
h
h
h
h
13
7e11, 14, 18
14
4
21
10, 18
14
18
5, 7, 8, 10, 14, 16, 18, 21
4
18
26
7
1, 2
10, 13
16, 18
6
17
13, 15e20, 24, 28, 29
12, 13, 24
4, 5, 8e13
Poaceae
Poaceae
Poaceae
Boraginaceae
Capparaceae
Caesalpinioideae
Asclepiadaceae
Papilionoideae
Capparaceae
Capparaceae
Capparaceae
Capparaceae
Sapindaceae
Apocynaceae
Caesalpinioideae
Ulmaceae
Ulmaceae
Poaceae
Poaceae
Umbelliferae
Caesalpinioideae
Poaceae
Poaceae
Poaceae
Edible parts (eaten by)3
L (H, B)
L (H)
Fl, L (H)
Fl, SP (B)
Fr (C)
L (C)
L (H)
Fr, SP (H), Fr, SP, BC (B)
L (H)
L (H)
SP, L, U (H)
Fr (H)
L, BC, U (H)
SP (H), SP, L, U (B)
SP (H)
L, St (H)
Fl (H)
Fr (H)
Fr (H, B)
L (H)
Fr, U (H)
L (H, C)
L (H), SP (B)
(continued on next page)
168
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Appendix 1 (continued )
Family or
subfamily
Species
G1
Modern landscape facets2
Menispermaceae
Orobanchaceae
Rutaceae
Capparaceae
Cucurbitaceae
Combretaceae
Commelinaceae
Commelinaceae
Commelinaceae
Commelinaceae
Commelinaceae
Passifloraceae
Burseraceae
Burseraceae
Burseraceae
Burseraceae
Burseraceae
Compositae
Compositae
Compositae
Cucurbitaceae
Tiliaceae
Tiliaceae
Boraginaceae
Boraginaceae
Boraginaceae
Compositae
Compositae
Scrophulariaceae
Papilionoideae
Papilionoideae
Papilionoideae
Papilionoideae
Papilionoideae
Papilionoideae
Euphorbiaceae
Cucurbitaceae
Cucurbitaceae
Cucurbitaceae
Cucurbitaceae
Amaranthaceae
Asclepiadaceae
Poaceae
Cissampelos mucronata A. Rich.
Cistanche tubulosa (Schenk) Hook. f.
Clausena anisata (Willd.) Hook. f. ex Benth.
Cleome gynandra L.
Coccinia grandis (L.) Voigt
Combretum sp. Loefl.
Commelina africana L.
Commelina benghalensis L.
Commelina erecta L.
Commelina foliacea Chiov.
Commelina sp. L.
Commicarpus plumbagineus (Cav.) Standl.
Commiphora africana (A. Rich.) Engl.
Commiphora eminii Engl.
Commiphora merkeri Engl.
Commiphora schimperi (O. Berg) Engl.
Commiphora sp. Jacq.
Conyza bonariensis (L.) Cronquist
Conyza newii Oliv. and Hiern
Conyza stricta Willd.
Corallocarpus epigaeus (Rottler) C.B. Clarke
Corchorus tridens L.
Corchorus trilocularis L.
Cordia goetzei Gürke
Cordia monoica Roxb.
Cordia sinensis Lam.
Crassocephalum picridifolium (DC.) S. Moore
Crassocephalum vitellinum (Benth.) S. Moore
Craterostigma plantagineum Hochst.
Crotalaria deflersii Schweinf.
Crotalaria incana L.
Crotalaria kirkii Baker
Crotalaria polysperma Kotschy
Crotalaria sp. L.
Crotalaria spinosa Hochst. ex Benth.
Croton macrostachyus Hochst. ex Delile
Ctenolepis cerasiformis (Stocks) Hook. f.
Cucumis aculeatus Cogn.
Cucumis dipsaceus Ehrenb.
Cucumis figarei Naudin
Cyathula orthacantha (Hochst. ex Asch.) Schinz
Cynanchum hastifolium N.E. Br.
Cynodon dactylon (L.) Pers.
w
h
w
h
h
w
h
h
h
h
h
h
w
w
w
w
w
h
w
h
h
h
h
w
w
w
h
h
h
h
h
h
h
h
h
w
h
h
h
w
h
h
h
Poaceae
Poaceae
Cyperaceae
Cyperaceae
Cyperaceae
Cyperaceae
Cyperaceae
Cyperaceae
Cyperaceae
Cyperaceae
Cyperaceae
Cyperaceae
Cyperaceae
Vitaceae
Cynodon nlemfuensis Vanderyst
Cynodon plectostachyus (K. Schum.) Pilg.
Cyperus immensus C.B. Clarke
Cyperus involucratus Rottb.
Cyperus laevigatus L.
Cyperus laxus Lam.
Cyperus oblongus (C.B. Clarke) Kük.
Cyperus papyrus L.
Cyperus pulchellus R. Br.
Cyperus rotundus L.
Cyperus sesquiflorus (Torr.) Mattf. and Kük.
Cyperus sp. L.
Cyperus usitatus Burch.
Cyphostemma kerkrooderi (Dewit) Desc. ex
Wild and R.B. Drumm.
Cyphostemma kilimandscharicum (Gilg) Desc.
ex Wild and R.B. Drumm.
Cyphostemma serpens (Hochst. ex A. Rich.) Desc.
Dactyloctenium aegyptium (L.) Willd.
h
h
h
h
h
h
h
h
h
h
h
h
h
h
6
7
14, 18
11
4, 5, 8e11
18
4, 5, 8e11
10, 13, 15, 16
13
11
9
8, 14
13
14
13
13, 14, 16, 18
15
10
21
12
13
4, 5
10, 19
1, 2, 6
6e11, 14e16, 18, 19, 21
4, 5, 7, 9e11
20
20
16
13
5, 9, 19
17
4, 12
26
13, 17, 19
1, 2, 21
4, 9, 11
11
4, 5, 9, 11
14, 18
8e10
15
4e6, 12, 13, 19e24,
26, 28, 29, 32
10, 20
7, 9e11
20e22, 24e27, 30e32
2, 6, 14
20, 22e24, 26, 28e30, 32
22, 24, 26, 27, 30
4
20, 21
4
16, 19, 20, 31
20, 21, 24, 26, 27
5
4
16
h
13
h
h
16
4, 5, 8e11, 13
Vitaceae
Vitaceae
Poaceae
Edible parts (eaten by)3
Fr (H)
Fl, L, St (H)
Fr, L (H)
W (H)
Fl, Fr, L, St (H)
Fl, St (B)
Fr, L, St, BC, U (H), Fr (B)
L (H)
L (H)
Fr (H, B)
Fr, U (H), Fr (B)
L (C)
L (H)
Fr (C)
SP (H)
Fr (H)
SP (H), W (B)
SP, L, St, W (B)
U (B)
U (B)
St (C), L, St, U (H)5
U (H, B)
U (H, B)
SP, U (H)
169
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Appendix 1 (continued )
Family or
subfamily
Species
G1
Modern landscape facets2
Solanaceae
Mimosoideae
Acanthaceae
Acanthaceae
Amaranthaceae
Poaceae
Poaceae
Poaceae
Poaceae
Poaceae
Sterculiaceae
Flacourtiaceae
Flacourtiaceae
Acanthaceae
Poaceae
Poaceae
Poaceae
Compositae
Meliaceae
Poaceae
Compositae
Gentianaceae
Poaceae
Poaceae
Datura stramonium L.
Dichrostachys cinerea (L.) Wight and Arn.
Dicliptera sp.
Dicliptera verticillata (Forssk.) C. Chr.
Digera muricata (L.) Mart.
Digitaria ciliaris (Retz.) Koeler
Digitaria macroblephara (Hack.) Paoli
Digitaria milanjiana (Rendle) Stapf
Digitaria velutina (Forssk.) P. Beauv.
Diplachne fusca (L.) P. Beauv. ex Roem. and Schult.
Dombeya cincinnata K. Schum.
Dovyalis sp. E. Mey. ex Arn.
Dovyalis xanthocarpa Bullock
Dyschoriste sp. Nees
Echinochloa colona (L.) Link
Echinochloa haploclada (Stapf) Stapf
Echinochloa sp. P. Beauv.
Eclipta prostrata (L.) L.
Ekebergia capensis Sparrm.
Eleusine indica (L.) Gaertn.
Emilia coccinea (Sims) G. Don
Enicostema axillare (Lam.) A. Raynal
Enneapogon cenchroides (Licht.) C.E. Hubb.
Enteropogon macrostachyus (Hochst. ex A. Rich.)
Munro ex Benth.
Eragrostis arenicola C.E. Hubb.
Eragrostis aspera (Jacq.) Nees
Eragrostis cilianensis (All.) Vignolo ex Janch.
Eragrostis cylindriflora Hochst.
Eragrostis heteromera Stapf
Eragrostis racemosa (Thunb.) Steud.
Eragrostis superba Peyr.
Eragrostis tenuifolia (A. Rich.) Hochst. ex Steud.
Eriochloa fatmensis (Hochst. and Steud.) Clayton
Eriochloa meyeriana (Nees) Pilg.
Euclea divinorum Hiern
Euclea racemosa L.
Euclea sp. L.
Euphorbia candelabrum Tremaux ex Kotschy
Euphorbia heterophylla L.
Euphorbia mossambicensis (Klotzsch and
Garcke) Boiss.
Euphorbia systyloides Pax
Ficus sp. L.
Ficus sycomorus L.
Flueggea virosa (Roxb. ex Willd.) Voigt
Galinsoga parviflora Cav.
Gardenia ternifolia Schumach. and Thonn.
Gisekia pharnaceoides L.
Glycine sp. Willd.
Gomphocarpus integer (N.E. Br.) Bullock
Gomphocarpus physocarpus E. Mey.
Gomphocarpus semilunatus A. Rich.
Grewia bicolor Juss.
Grewia forbesii Harv. ex Mast.
Grewia stolzii Ulbr.
Grewia tembensis Fresen.
Gutenbergia cordifolia Benth. ex Oliv.
Gutenbergia petersii Steetz
Gutenbergia polycephala Oliv. and Hiern
Gutenbergia polytrichotoma Wech.
Harpachne schimperi A. Rich.
Helichrysum forskahlii (J.F. Gmel.) Hilliard and
B.L. Burtt
h
w
h
h
h
h
h
h
h
h
w
w
w
h
h
h
h
h
w
h
h
h
h
h
4, 5, 11, 21
15, 16, 18
5
9e11
11
15
12, 16, 22
13, 16, 26
4, 5, 7e13
20, 21, 24, 26, 27, 30
18
1, 2
2, 10
17
10
4, 5, 8, 19
11
4, 6, 8
2, 6
5
17
12
4, 9e11, 13
9
h
h
h
h
h
h
h
h
h
h
w
w
w
w
h
h
24
9, 13
12
19
12, 13
13
10
20
5
5, 10
18
15
18
26
10
12, 13
h
w
w
w
h
w
h
w
w
w
w
w
w
w
w
h
H
w
h
h
h
13
15
1, 2, 6, 10, 18
11, 14, 15, 18
8, 10
1, 8e11
4
18
19
13
14
14, 15, 18
15
15
14, 15
9e11, 15, 17, 19, 21
13, 15, 17
14
4, 5, 8e11
12, 13
12
Poaceae
Poaceae
Poaceae
Poaceae
Poaceae
Poaceae
Poaceae
Poaceae
Poaceae
Poaceae
Ebenaceae
Ebenaceae
Ebenaceae
Euphorbiaceae
Euphorbiaceae
Euphorbiaceae
Euphorbiaceae
Moraceae
Moraceae
Euphorbiaceae
Compositae
Rubiaceae
Aizoaceae
Papilionoideae
Asclepiadaceae
Asclepiadaceae
Asclepiadaceae
Tiliaceae
Tiliaceae
Tiliaceae
Tiliaceae
Compositae
Compositae
Compositae
Compositae
Poaceae
Compositae
Edible parts (eaten by)3
L, BC (H)
L (H)
SP (H)
SP (B)
W (H), SP (B)
Fr (H)
SP (H, B)
SP (H, B)
W (H)
Fr (H)
SP (H)
L (H)
SP (H, B)
Fr, L (H)
Fr (H, B)
Fl, L (B)
Fr (B)
Fr (H)
Fr, L (H), Fr, L (B)
Fr (H, B)
Fr (H)
L (H)
Fr
Fr
Fr
Fr
(H, B)
(H), Fl, Fr, L (C)
(H)
(H)
(continued on next page)
170
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Appendix 1 (continued )
Family or
subfamily
Species
G1
Modern landscape facets2
Edible parts (eaten by)3
Compositae
Rhamnaceae
Boraginaceae
Boraginaceae
Poaceae
Helichrysum glumaceum DC.
Helinus mystacinus (Aiton) E. Mey. ex Steud.
Heliotropium steudneri Vatke
Heliotropium zeylanicum (Burm. f.) Lam.
Heteropogon contortus (L.) P. Beauv. ex Roem.
and Schult.
Hibiscus aponeurus Sprague and Hutch.
Hibiscus cannabinus L.
h
w
h
h
h
12, 13
21
13, 16, 18
4
10, 13
U (B)
w
h
15
5, 19
Hibiscus flavifolius Ulbr.
Hibiscus micranthus L. f.
Hibiscus ovalifolius (Forssk.) Vahl
Hibiscus sp. L.
Hippocratea africana (Willd.) Loes. ex Engl.
Hippocratea paniculata Vahl
Hirpicium diffusum (O. Hoffm.) Roessler
Hoslundia opposita Vahl
Hoslundia sp. Vahl
Hydrocotyle mannii Hook. f.
Hydrocotyle ranunculoides L. f.
Hydrocotyle sibthorpioides Lam.
Hyparrhenia filipendula (Hochst.) Stapf
Hyparrhenia hirta (L.) Stapf
Hyphaene petersiana Mart
Hypoestes forskalei (Vahl) Sol. ex Roem.
and Schult.
Hypoestes sp. Sol. ex R. Br.
Hypoxis sp. L.
Indigofera arrecta Hochst. ex A. Rich.
Indigofera basiflora J.B. Gillett
Indigofera colutea (Burm. f.) Merr.
Indigofera hochstetteri Baker
Indigofera sp. L.
Indigofera spicata Forssk.
Indigofera tenuis Milne-Redh.
Indigofera volkensii Taub.
Ipomoea arachnosperma Welw.
Ipomoea cairica (L.) Sweet
Ipomoea coptica (L.) Roem. and Schult.
Ipomoea hochstetteri House
Ipomoea jaegeri Pilg.
Ipomoea mombassana Vatke
Ipomoea obscura (L.) Ker Gawl.
Ipomoea sinensis (Desr.) Choisy
Ipomoea sp. L.
Jasminum fluminense Vell.
Justicia betonica L.
Justicia caerulea Forssk.
Justicia cordata (Nees) T. Anderson
Justicia diclipteroides Lindau
Justicia exigua S. Moore
Justicia flava (Forssk.) Vahl
Justicia glabra K. Koenig ex Roxb.
Justicia matammensis (Schweinf.) Oliv.
Justicia palustris (Hochst.) T. Anderson
Justicia sp. L.
Justicia stachytarphetoides (Lindau) C.B. Clarke
Justicia striata (Klotzsch) Bullock
Kalanchoe densiflora Rolfe
Kedrostis foetidissima (Jacq.) Cogn.
Kedrostis hirtella (Naud.) Cogn.
Kohautia caespitosa Schnizl.
Lagenaria siceraria (Molina) Standl.
w
w
w
w
w
w
h
w
w
h
h
h
h
h
w
h
13
4, 10, 13, 15, 16, 18
1, 2, 4, 5, 7e11, 14, 15, 18
9, 18
14
2
12, 13
10
18
29
6
20, 24, 26, 31
13, 17
20
5
1, 2, 10, 11, 14, 15, 19
h
h
h
h
h
h
w
h
h
h
h
w
h
h
h
h
h
h
h
w
w
h
w
h
h
h
h
h
w
h
h
h
h
h
h
h
h
9
16
4, 5, 8e11, 13, 14
16, 18
4, 12, 13
12, 13
9, 11, 13, 18
11, 12
4
12, 13
11, 15
6, 14
4
4
13
15
4, 8e11
15, 17, 19
11, 18
13, 18
14, 15, 18, 20, 21, 26
7, 9
4, 8, 11
5
5, 8
4, 5, 8e11
1, 2, 5e8, 11
15, 19
14
12e14, 16
4, 8e11
9, 15
13e15
4, 8e11
11
4, 13
5, 8e10
Malvaceae
Malvaceae
Malvaceae
Malvaceae
Malvaceae
Malvaceae
Celastraceae
Celastraceae
Compositae
Labiatae
Labiatae
Umbelliferae
Umbelliferae
Umbelliferae
Poaceae
Poaceae
Arecaceae
Acanthaceae
Acanthaceae
Hypoxidaceae
Papilionoideae
Papilionoideae
Papilionoideae
Papilionoideae
Papilionoideae
Papilionoideae
Papilionoideae
Papilionoideae
Convolvulaceae
Convolvulaceae
Convolvulaceae
Convolvulaceae
Convolvulaceae
Convolvulaceae
Convolvulaceae
Convolvulaceae
Convolvulaceae
Apocynaceae
Acanthaceae
Acanthaceae
Acanthaceae
Acanthaceae
Acanthaceae
Acanthaceae
Acanthaceae
Acanthaceae
Acanthaceae
Acanthaceae
Acanthaceae
Acanthaceae
Crassulaceae
Cucurbitaceae
Cucurbitaceae
Rubiaceae
Cucurbitaceae
L (C)
Fl, Fr, L, BC (H), SP, L, Fl
(B), Fr, L, Fl (C)
Fr (H)
Fl, SP (B), L(C)
Fr (H, B, C)
SP, L, St (B)
St (B)
Fl, Fr, SP, St, BC (H), Fr (B)
L (H)
U (H)
SP (B)
L (C)
Fl, Fr (B)
Fr (B)
L (H), Fr (B)
L, U (H) Fl, Fr (B)
Fl (H)
L (H)
W (H)
L (H), Fl, L (B)
L (H)
Fr (H)
Fr, L (H)
171
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Appendix 1 (continued )
Family or
subfamily
Species
G1
Modern landscape facets2
Edible parts (eaten by)3
Anacardiaceae
Anacardiaceae
Verbenaceae
Verbenaceae
Compositae
Lannea schweinfurthii (Engl.) Engl.
Lannea triphylla (A. Rich.) Engl.
Lantana rhodesiensis Moldenke
Lantana trifolia L.
Launaea cornuta (Hochst. ex Oliv. and Hiern)
C. Jeffrey
Lemna gibba L.
Lemna sp. L.
Leonotis nepetifolia (L.) R. Br.
Leonotis sp. (Pers.) R. Br.
Lepidagathis scabra C.B. Clarke
Leucas glabrata (Vahl) R. Br.
Leucas martinicensis (Jacq.) R. Br.
Leucas neuflizeana Courbon
Leucas sp. R. Br.
Lippia javanica (Burm f.) Spreng.
Lippia ukambensis Vatke
Lotus arabicus L.
Lotus goetzei Harms
Ludwigia sp. L.
Ludwigia stolonifera (Guill. and Perr.) P.H.
Raven
Lycium europaeum L.
Maerua triphylla A. Rich.
Malvastrum coromandelianum (L.) Garcke
Malvastrum sp. A. Gray
Mansonia sp. J. R. Drummond ex D. Prain
Maytenus heterophylla (Eckl. and Zeyh.) N. Robson
Medicago laciniata Mill.
Melhania ovata (Cav.) Spreng.
Melhania velutina Forssk.
Microglossa pyrrhopappa (Sch. Bip. ex A. Rich.)
Agnew
Monechma debile (Forssk.) Nees
Monsonia angustifolia E. Mey. ex A. Rich.
Nicandra physalodes (L.) Gaertn.
Ocimum basilicum L.
Ocimum sp. L.
Ocimum suave Willd.
Oldenlandia fastigiata Bremek.
Opilia amentacea Roxb.
Oplismenus hirtellus (L.) P. Beauv.
Oplismenus undulatifolius (Ard.) Roem. and Schult.
Ormocarpum kirkii S. Moore
Orthosiphon pallidus Royle ex Benth.
Orthosiphon parvifolius Vatke
Osteospermum vaillantii (Decne.) Norl.
Oxygonum sinuatum (Hochst. and Steud. ex Meisn.)
Dammer
Panicum atrosanguineum Hochst. ex A. Rich.
Panicum cf. repens L.
Panicum cf. subalbidum Kunth
Panicum deustum Thunb.
Panicum maximum Jacq.
Pappea capensis Eckl. and Zeyh.
Pavetta dolichantha Bremek.
Pavetta sepium K. Schum.
Pavetta sp. L.
Pavonia patens (Andrews) Chiov.
Pennisetum clandestinum Hochst. ex Chiov.
Pennisetum mezianum Leeke
Pennisetum sp. Rich.
Pennisetum subangustum (Schumach.) Stapf and
C.E. Hubb.
w
w
h
w
h
14
14
4
13
5, 24
Fr, St (H), Fr, St (B)
U (H)
Fr (H)
Fr (H)
L (H, B)
h
h
h
h
h
h
h
h
h
h
w
h
h
h
h
6
22, 24
4, 8, 10, 11, 19, 24
11
9
6, 17
11, 13, 15
5, 12, 13, 17
5
13, 18, 20, 21, 25
13, 14, 18, 21, 27
20
24
32
20
w
w
h
h
h
w
h
h
h
h
9, 11, 13, 16
5, 7e11, 14
1, 2, 6, 7, 10
2
21
13, 18
12, 13
12, 13, 16
4, 10
16
h
h
h
h
w
w
h
w
h
h
w
w
h
h
h
4, 5, 7e11, 19
12, 13
10
4, 5, 8e11
14
2, 10, 14, 18, 20, 21, 27
4
7, 10
1, 6
1
18
10
9, 12
13
4, 8
h
h
h
h
h
w
w
w
w
w
h
h
h
h
12, 19
28, 30
32
7, 11
4, 5, 8, 9, 14, 15, 19
13
11, 13
1, 7e11
14, 18
1, 4, 5, 8e11
20, 24, 26, 31, 32
12e19
13
12e14, 17e19
Lemnaceae
Lemnaceae
Labiatae
Labiatae
Acanthaceae
Labiatae
Labiatae
Labiatae
Labiatae
Verbenaceae
Verbenaceae
Papilionoideae
Papilionoideae
Onagraceae
Onagraceae
Solanaceae
Capparaceae
Malvaceae
Malvaceae
Sterculiaceae
Celastraceae
Papilionoideae
Sterculiaceae
Sterculiaceae
Compositae
Acanthaceae
Geraniaceae
Solanaceae
Labiatae
Labiatae
Labiatae
Rubiaceae
Opiliaceae
Poaceae
Poaceae
Papilionoideae
Labiatae
Labiatae
Compositae
Polygonaceae
Poaceae
Poaceae
Poaceae
Poaceae
Poaceae
Sapindaceae
Rubiaceae
Rubiaceae
Rubiaceae
Malvaceae
Poaceae
Poaceae
Poaceae
Poaceae
Fl, L (H)
L (H)
SP (H)
Fr, U (H)
Fr (H)
? (H)
L (H), Fl, L (B)
L (H)4
L (H)
Fr, L (H)
SP
SP
SP
Fr,
(H)
(H, B)
(H), SP (C), SP, L, St (B)
L (H), Fr (B)
(continued on next page)
172
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Appendix 1 (continued )
Family or
subfamily
Species
G1
Modern landscape facets2
Edible parts (eaten by)3
Asclepiadaceae
Acanthaceae
Polygonaceae
Polygonaceae
Acanthaceae
Arecaceae
Poaceae
Euphorbiaceae
Euphorbiaceae
Euphorbiaceae
Euphorbiaceae
Euphorbiaceae
Phytolaccaceae
Labiatae
Loranthaceae
Compositae
Compositae
Plumbaginaceae
Plumbaginaceae
Illecebraceae
Polygalaceae
Portulacaceae
Portulacaceae
Verbenaceae
Verbenaceae
Asparagaceae
Poaceae
Amaranthaceae
Apocynaceae
Anacardiaceae
Poaceae
Papilionoideae
Acanthaceae
Acanthaceae
Salvadoraceae
Cyperaceae
Caesalpinioideae
Caesalpinioideae
Pentarrhinum insipidum E. Mey.
Peristrophe bicalyculata Nees
Persicaria decipiens (R. Br.) K.L. Wilson
Persicaria senegalensis (Meisn.) Soják
Phaulopsis imbricata (Forssk.) Sweet
Phoenix reclinata Jacq.
Phragmites mauritianus Kunth
Phyllanthus amarus Schumach. and Thonn.
Phyllanthus fischeri Pax
Phyllanthus nummulariifolius Poir.
Phyllanthus odontadenius Müll. Arg.
Phyllanthus sepialis Müll. Arg.
Phytolacca dodecandra L’Hér.
Plectranthus flaccidus (Vatke) Gürke
Plicosepalus curviflorus (Benth. ex Oliv.) Tiegh.
Pluchea dioscoridis (L.) DC.
Pluchea sordida (Vatke) Oliv. and Harms
Plumbago dawei Rolfe
Plumbago zeylanica L.
Pollichia campestris Aiton
Polygala sphenoptera Fresen.
Portulaca foliosa Ker Gawl.
Portulaca oleracea L.
Priva adhaerens (Forssk.) Chiov.
Priva curtisiae Kobuski
Protasparagus africanus (Lam.) Oberm.
Psilolemma jaegeri (Pilg.) S.M. Phillips
Pupalia lappacea (L.) Juss.
Rauvolfia caffra Sond.
Rhus quartiniana A. Rich.
Rhynchelytrum repens (Willd.) C.E. Hubb.
Rhynchosia minima (L.) DC.
Ruellia megachlamys S. Moore
Ruellia patula Jacq.
Salvadora persica L.
Scirpus inclinatus (Delile) Asch. and Schweinf.
Senna bicapsularis (L.) Roxb.
Senna didymobotrya (Fresen.) H.S. Irwin and
Barneby
Senna obtusifolia (L.) H.S. Irwin and Barneby
Senna occidentalis (L.) Link
Senna sp. Mill.
Sesbania bispinosa (Jacq.) W. Wight
Sesuvium portulacastrum (L.) L.
Setaria homonyma (Steud.) Chiov.
Setaria incrassata (Hochst.) Hack.
Setaria sagittifolia (Hochst. ex A. Rich.) Walp.
Setaria sp. P. Beauv.
Setaria verticillata (L.) P. Beauv.
Sida acuta Burm. f.
Sida alba L.
Sida ovata Forssk.
Solanum incanum L.
Solanum nigrum L.
Solanum setaceum Dammer
Solanum sp. L.
Sphaeranthus cyathuloides O. Hoffm.
Sphaeranthus steetzii Oliv. and Hiern
Sphaeranthus suaveolens (Forssk.) DC.
Sphaeranthus ukambensis Vatke and O. Hoffm.
Sporobolus africanus (Poir.) Robyns and Tournay
Sporobolus consimilis Fresen.
Sporobolus cordofanus (Hochst.) Herincq Cosson
Sporobolus fimbriatus (Trin.) Nees
h
h
h
h
h
w
h
h
w
h
h
w
w
h
w
w
w
w
w
h
h
h
h
h
h
h
h
h
w
w
h
h
h
h
w
h
w
w
11
4, 5, 7e11
31
21
1, 2
1, 2, 6
20
12, 13
4, 13e15, 18, 19
4
11
10, 11
18, 21
5
10
14
14
18
14
13
12
4
4
7, 9
13
13
23
9e11, 14
2, 6
18
10
12, 13, 15, 18
9e11
4, 5
4, 5, 7e11
20, 22, 24e27, 31
1, 2, 6
1, 25, 26
Fr, L (H)
w
w
w
h
h
h
h
h
h
h
h
h
h
h
h
h
h
h
h
h
h
h
h
h
h
21
1, 4, 5, 11
20, 21
15
5
8, 10, 11, 15, 21
17
4, 7e11
10
4, 5, 8e11, 13e15, 21
10
5, 7, 9
4, 5, 8e13, 17
4, 5, 8e19, 25, 26
6, 10, 11, 13
4, 8e10, 13, 17
16, 24
4
15
20, 21, 24, 27
5,10
20
4, 5, 13, 14, 18, 19
4, 5, 9e11
12, 13, 16, 17, 19
Caesalpinioideae
Caesalpinioideae
Caesalpinioideae
Caesalpinioideae
Aizoaceae
Poaceae
Poaceae
Poaceae
Poaceae
Poaceae
Malvaceae
Malvaceae
Malvaceae
Solanaceae
Solanaceae
Solanaceae
Solanaceae
Compositae
Compositae
Compositae
Compositae
Poaceae
Poaceae
Poaceae
Poaceae
L (H)
L, U (H)
L (H)
Fr, L, BC (H), Fr, L (B)
L (H), L, St (C), St (B)
L (H)
W, Fr, Fl (H), Fl (B)
L, St (H)
L, U (H)
SP (H)
L (H)
Fr, L, BC (H), Fr (B)
SP (H)
SP (H)
L (H)
SP (B)
Fl, SP (B)
L (H), Fl, SP (B)
L (H)
L (H), Fr, St (B)
L (H, B)
SP, L, U (B)
SP (H, B)
173
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Appendix 1 (continued )
Family or
subfamily
Species
G1
Modern landscape facets2
Edible parts (eaten by)3
Poaceae
Poaceae
Poaceae
Asclepiadaceae
Thelypteridaceae
Scrophulariaceae
Apocynaceae
Compositae
Papilionoideae
Papilionoideae
Papilionoideae
Papilionoideae
Poaceae
Poaceae
Capparaceae
Euphorbiaceae
Poaceae
Zygophyllaceae
Meliaceae
Boraginaceae
Tiliaceae
Typhaceae
Poaceae
Poaceae
Rubiaceae
Rubiaceae
Rutaceae
Rutaceae
Compositae
Compositae
Compositae
Papilionoideae
Papilionoideae
Compositae
Aizoaceae
Rhamnaceae
Rhamnaceae
Rhamnaceae
Sporobolus ioclados (Trin.) Nees
Sporobolus spicatus (Vahl) Kunth
Sporobolus tenuissimus (Mart. ex Schrank) Kuntze
Stathmostelma pedunculatum (Decne.) K. Schum.
Stegnogramma pozoi (Lag.) K. Iwats.
Striga gesnerioides (Willd.) Vatke ex Engl.
Tabernaemontana ventricosa Hochst. ex A. DC.
Tagetes minuta L.
Tephrosia pumila (Lam.) Pers.
Tephrosia purpurea (L.) Pers.
Tephrosia rhodesica Baker f.
Tephrosia villosa (L.) Pers.
Tetrapogon cenchriformis (A. Rich.) Clayton
Themeda triandra Forssk.
Thilachium africanum Lour.
Tragia benthamii Baker
Tragus berteronianus Schult.
Tribulus terrestris L.
Trichilia emetica Vahl
Trichodesma zeylanicum (Burm. f.) R. Br.
Triumfetta flavescens Hochst. ex A. Rich.
Typha latifolia L.
Urochloa mosambicensis (Hack.) Dandy
Urochloa panicoides P. Beauv.
Vangueria madagascariensis J.F. Gmel.
Vangueria sp. Juss.
Vepris lanceolata (Lam.) G. Don
Vepris uguenensis Engl.
Vernonia galamensis (Cass.) Less.
Vernonia myriantha Hook. f.
Vernonia sp. Schreb.
Vigna schimperi Baker
Vigna vexillata (L.) A. Rich.
Xanthium pungens Wallr.
Zaleya pentandra (L.) C. Jeffrey
Ziziphus mucronata Willd.
Ziziphus pubescens Oliv.
Ziziphus sp. Mill.
h
h
h
h
h
h
w
w
h
h
w
h
h
h
w
h
h
h
w
h
h
h
h
h
w
w
w
w
w
w
w
h
h
h
h
w
w
w
4, 5, 8, 12, 13, 15, 18, 19
4, 13, 19, 24, 29
11
16
20
4
1, 2, 5, 6
5, 8, 10, 14, 20, 21
17
19
11
4, 10
9, 10
12, 13, 16e19, 21, 22, 29
1, 7, 9, 10
1
4, 10, 11
4, 9
1, 2, 6, 7, 10, 11
4, 11
5
20, 26, 29, 31
7e11
9e11
1, 21, 27
14, 18
18
7e11
21
27
13, 14
20
20
5
9
14, 18
1
1
SP (B)
SP (H)
U (H, B)
SP (B)
U (H)
L (H), Fl, SP, L, St (B)
Fl (B), U (H)
Fl, L (H), Fl, L (B)
Fr (H, B)
SP, L (H), Fl, L (B)
Fl, U (H)
SP (H), SP, U (B)
Fr (H)
L (H), Fl, St (C)
SP, L, U (H), SP, Fl, U (B)
L (H)
Fr (H), Fr (C), Fr, L (B)
Fr, L (H)
1
G ¼ growth form; h ¼ herbaceous and very small (<1 m) woody species, including grasses; w ¼ woody species (shrubs and trees).
Numbers refer to the map locations defined in Table 2 and depicted in Fig. 2.
3
Fl ¼ flowers/inflorescenses, Fr ¼ fruits; SP ¼ seeds, pods, or grass grains; L ¼ leaves or shoots; St ¼ stems; BC ¼ bark, cambium, or gum; U ¼ underground
part; W ¼ whole plant; ? ¼ unspecified; H ¼ human; B ¼ baboon; C ¼ chimpanzee. Reference: Peters et al. (1992).
4
Reference: Marcan, 1998; possible exotic from Asia.
5
Additional references: Peters, 1999; Peters and Vogel, 2005.
2
References
Albert, R.M., Bamford, M.K., Cabanes, D., 2006. Taphonomy of phytoliths
and macroplants in different soils from Olduvai Gorge (Tanzania) and
the application to Plio-Pleistocene palaeoanthropological samples. Quatern. Int. 148, 78e94.
Anderson, G.D., Herlocker, D.J., 1973. Soil factors affecting the distribution of
the vegetation types and their utilization by wild animals in Ngorongoro
Crater, Tanzania. J. Ecol. 61, 627e651.
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, 33e56.
Ashley, G.M., Driese, S.G., 2000. Paleopedology and paleohydrology of a volcaniclastic paleosol interval: implications for early Pleistocene stratigraphy
and paleoclimate record, Olduvai Gorge, Tanzania. J. Sediment. Res. 70,
1065e1080.
Bamford, M.K., 2005. Early Pleistocene fossil wood from Olduvai Gorge,
Tanzania. Quatern. Int. 129, 15e22.
Bamford, M.K., Albert, R.M., Cabanes, D., 2006. Plio-Pleistocene macroplant
fossil remains and phytoliths from Lowermost Bed II in the eastern
palaeolake margin of Olduvai Gorge, Tanzania. Quatern. Int. 148, 95e112.
Beentje, H.J., 1994. Kenya Trees, Shrubs, and Lianas. National Museums of
Kenya, Nairobi.
Belsky, A.J., 1984. The role of small browsing mammals in preventing woodland regeneration in the Serengeti National Park, Tanzania. Afr. J. Ecol. 22,
271e279.
Belsky, A.J., 1990. Tree/grass ratios in East African savannas: a comparison of
existing models. J. Biogeogr. 17, 483e489.
Birnie, A., 1997. What Tree Is That? A Beginner’s Guide to 40 Trees in
Kenya. Jacaranda Designs, Ltd., Nairobi.
Blumenschine, R.J., Peters, C.R., Masao, F.T., Clarke, R.J., Deino, A.L.,
Hay, R.L., Swisher, C.C., Stanistreet, I.G., Ashley, G.M., McHenry, L.J.,
174
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Sikes, N.E., van der Merwe, N.J., Tactikos, J.C., Cushing, A.E.,
Deocampo, D.M., Njau, J.K., Ebert, J.I., 2003. Late Pliocene Homo and
hominid land use from western Olduvai Gorge, Tanzania. Science 299,
1217e1221.
Blundell, M., 1987. Collins Photo Guide to the Wild Flowers of East Africa.
Harper Collins Publishers, Hong Kong.
Bonnefille, R., 1984. Palynological research at Olduvai Gorge. Nat. Geogr.
Soc. Res. Rep. 17, 227e243.
Bonnefille, R., Riollet, G., 1980. Palynologie, végétation et climates de Bed
I et Bed II à Olduvai, Tanzanie. In: Proceedings of the Eighth PanAfrican Congress of Prehistoric and Quaternary Studies, Nairobi. pp.
123e127.
Casanova, J., Hillaire-Marcel, C., 1992. Chronology and paleohydrology of
late Quaternary high lake levels in the Manyara basin (Tanzania) from isotopic data (18O, 13C, 14C, Th/U) on fossil stromatolites. Quatern. Res. 38,
205e226.
Cerling, T.E., Hay, R.L., 1986. An isotopic study of paleosol carbonates from
Olduvai Gorge. Quatern. Res. 25, 63e78.
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. 213e230.
Coates Palgrave, K., 1993. Trees of Southern Africa. Struik Publishers, Cape
Town.
Coetzee, J.A., 1967. Pollen analytical studies in East and Southern Africa.
Palaeoecol. Afr. 3, 1e146.
Conklin-Brittain, N.L., Wrangham, R.W., Hunt, K.D., 1998. Dietary response
of chimpanzees and cercopithecines to seasonal variation in fruit abundance. II. Macronutrients. Int. J. Primatol. 19, 971e998.
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.
Coughener, M.B., Ellis, J.E., 1993. Landscape and climatic control of woody
vegetation in a dry tropical ecosystem: Turkana District, Kenya. J. Biogeogr. 20, 383e398.
Cushing, A.E., 2002. The landscape zooarchaeology and paleontology of
Plio-Pleistocene Olduvai, Tanzania and their implications for early hominid
ecology. Ph.D. Dissertation, Rutgers, the State University of New Jersey.
deMenocal, P.B., 1995. Plio-pleistocene African climate. Science 270, 53e59.
Deocampo, D.M., 2004. Authigenic clays in East Africa: regional trends and
paleolimnology at the Plio-Pleistocene boundary, Olduvai Gorge, Tanzania. J. Paleolimnol. 31, 1e9.
Deocampo, D.M., Ashley, G.M., 1999. Siliceous islands in a carbonate sea:
modern and Pleistocene spring-fed wetlands in Ngorongoro Crater and
Olduvai Gorge, Tanzania. J. Sediment. Res. 69, 974e979.
Deocampo, D.M., Blumenschine, R.J., Ashley, G.M., 2002. Wetland diagenesis and traces of early hominids, Olduvai Gorge, Tanzania. Quatern. Res.
57, 271e281.
Dublin, H.T., 1995. Vegetation dynamics in the Serengeti-Mara ecosystem: the
role of elephants, fire, and other factors. In: Sinclair, A.R.E., Arcese, P.
(Eds.), Serengeti II: Dynamics, Management, and Conservation of an Ecosystem. University of Chicago Press, Chicago, pp. 71e90.
Ellery, W.N., Ellery, K., McCarthy, T.S., 1993. Plant distribution in islands of
the Okavango Delta, Botswana: determinants and feedback interactions.
Afr. J. Ecol. 31, 118e134.
Feibel, C.S., Brown, F.H., McDougall, I., 1989. Stratigraphic context of fossil
hominids from the Omo Group deposits: Northern Turkana basin, Kenya
and Ethiopia. Am. J. Phys. Anthropol. 78, 595e622.
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, 137e172.
Frankfurt Zoological Society, 1971. Map of the Serengeti National Park and
the Surrounding Area. Compiled by T.M. Caro and drawn by Hunting
Technical Services, Ltd.
Frost, P.G.H., Robertson, F., 1987. The ecological effects of fire in savannas.
In: Walker, B.H. (Ed.), Determinants of Tropical Savannas. Determinants
of Tropical SavannasIUBS Monograph Series, No. 3. IRL Press, Oxford,
pp. 93e140.
Gerresheim, K., 1971. Landscape classification and the storage of ecological
data. Second Conference of Land Use in Tanzania, The University of
Dar es Salaam, Paper No. 19, Mimeo.
Gerresheim, K., 1974. The Serengeti Landscape Classification (with accompanying 1:250,000 scale map). Africa Wildlife Leadership Foundation,
Nairobi.
Greenway, P.J., Vesey-Fitzgerald, D.F., 1969. The vegetation of Lake Manyara
National Park. J. Ecol. 57, 127e149.
Greig-Smith, P., 1983. Quantitative Plant Ecology. University of California
Press, Berkeley.
Haines, R.W., Lye, K.A., 1983. The Sedges and Rushes of East Africa e
a Flora of the Families Juncaceae and Cyperaceae in East Africa with a
Particular Reference to Uganda. East African Natural History Society,
Nairobi.
Hamilton, A.C., 1972. The interpretation of pollen diagrams from Highland
Uganda. Palaeoecol. Afr. 7, 45e149.
Hausfater, G., Bearce, W.H., 1976. Acacia tree exudates: their composition and
use as a food source by baboons. E. Afr. Wildl. J. 14, 241e243.
Hay, R.L., 1976. Geology of the Olduvai Gorge. University of California
Press, Berkeley.
Hay, R.L., 1996. Stratigraphy and lake-margin paleoenvironments of Lowermost Bed II in Olduvai Gorge. In: Magori, C., Saanane, C.B.,
Schrenk, F. (Eds.), Four Million Years of Hominid Evolution in Africa: Papers in Honour of Dr. Mary Douglas Leakey’s Outstanding Contribution in
Paleoanthropology. Kaupia 6. Darmstadter Beitrage zur Naturgeschichte,
Darmstadt.
Hay, R.L., Reeder, R.J., 1978. Calcretes of Olduvai Gorge and the Ndolanya
Beds of northern Tanzania. Sedimentology 25, 649e673.
Herlocker, D., 1975. Woody Vegetation of the Serengeti National Park. Caeser
Kleberg Research Program, Texas A&M University, College Station,
Texas.
Herlocker, D., Dirschl, H.J., 1972. Vegetation of the Ngorongoro Conservation
Area, Tanzania. Can. Wildl. Ser. Rep. Ser. 19, 5e39.
Hughes, F.M.R., 1988. The ecology of African floodplain forests in semi-arid
and arid zones: a review. J. Biogeogr. 15, 127e140.
Hughes, F.M.R., 1990. The influence of flooding regimes on forest distribution
and composition in the Tana River floodplain, Kenya. J. Appl. Ecol. 27,
475e491.
Ibrahim, K.M., Kabuye, C.H.S., 1987. An Illustrated Manual of Kenya
Grasses. FAO, Rome.
Kano, T., 1972. Distribution and adaptation of the chimpanzee on the eastern
shore of Lake Tanganyika. Kyoto Univ. Afr. Stud. 7, 37e129.
Kappelman, J., 1984. Plio-Pleistocene environments of Bed I and Lower
Bed II, Olduvai Gorge, Tanzania. Palaeogeogr. Palaeoclimatol. Palaeoecol.
48, 171e196.
Kappelman, J., Plummer, T., Bishop, L., Duncan, A., Appleton, S., 1997.
Bovids as indicators of Plio-Pleistocene paleoenvironments in East Africa.
J. Hum. Evol. 32, 229e256.
Leakey, M.D., 1971. Olduvai Gorge, vol. 3. Excavations in Beds I and II,
1960e1963. Cambridge University Press, Cambridge.
Livingstone, D.A., 1967. Postglacial vegetation of the Ruwenzori Mountains
in equatorial Africa. Ecol. Monogr. 37, 25e52.
Loth, P.E., Prins, H.H.T., 1986. Spatial patterns of the landscape and vegetation of Lake Manyara National Park. ITC J. 2, 115e130.
Marcan, H., 1998. Economic Uses of Central African Plants, with Emphasis
upon Ethnobotanical and Traditional Culture. Kallkwik, Banbury, U.K.
Marean, C.W., Ehrhardt, C.L., 1995. Paleoanthropological and paleoecological
implications of the taphonomy of a sabertooth’s den. J. Hum. Evol. 29,
515e547.
Mbuya, L.P., Msanga, H.P., Ruffo, C.K., Birnie, A., Tengnas, B., 1994. Useful
Trees and Shrubs for Tanzania. Regional Soil Conservation Unit/Swedish
International Development Agency, Nairobi.
Newman, J.L., 1975. Dimensions of Sandawe diet. Ecol. Food Nutr. 4, 33e39.
Nishida, T., Uehara, S., 1983. Natural diet of chimpanzees (Pan troglodytes
schweinfurthii): long-term record from the Mahale Mountains, Tanzania.
Afr. Stud. Monogr. 3, 109e130.
Norton-Griffiths, M., 1979. The influence of grazing, browsing, and fire on the
vegetation dynamics of the Serengeti. In: Sinclair, A.R.E., Norton-
S.R. Copeland / Journal of Human Evolution 53 (2007) 146e175
Griffiths, M. (Eds.), Serengeti: Dynamics of an Ecosystem. University of
Chicago Press, Chicago, pp. 310e352.
Norton-Griffiths, M., Herlocker, D., Pennycuick, L., 1975. The patterns of rainfall in the Serengeti Ecosystem, Tanzania. E. Afr. Wildl. J. 13, 347e374.
O’Brien, E.M., 1988. Climatic correlates of species richness for woody ‘‘edible’’ plants across southern Africa. Monogr. Syst. Bot. Mo. Bot. Gard. 25,
385e402.
O’Brien, E.M., 1993. Climatic gradients in woody plant species richness:
towards and explanation based on an analysis of southern Africa’s woody
flora. J. Biogeogr. 20, 181e198.
O’Brien, E.M., Peters, C.R., 1991. Ecobotanical contexts for African hominids. In: Clark, J.D. (Ed.), Cultural Beginnings: Approaches to Understanding Early Hominid Life-Ways in the African Savanna. R. Habelt,
Bonn, pp. 1e15.
Pellew, R.A.P., 1983. The impacts of elephant, giraffe and fire upon the Acacia
tortilis woodlands of the Serengeti. Afr. J. Ecol. 21, 41e74.
Peters, C.R., 1987. Nut-like oil seeds: food for monkeys, chimpanzees,
humans, and probably ape-men. Am. J. Phys. Anthropol. 73, 333e363.
Peters, C.R., 1994. African wild plants with rootstocks reported to be eaten
raw: the Monocotyledons, Part II. In: Seyani, J.H., Chikuni, A.C. (Eds.),
Proceedings of the XIIIth Plenary Meeting of AETFAT, Zomba, Malawi,
2e11 April 1991. National Herbarium and Botanic Gardens of Malawi,
Zomba, pp. 25e38.
Peters, C.R., 1999. African wild plants with rootstocks reported to be eaten
raw: the Monoctyledons, Part IV. In: Timberlake, J., Kativu, S. (Eds.), African Plants: Biodiversity, Taxonomy and Uses. Royal Botanic Gardens,
Kew, pp. 483e503.
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, 321e362.
Peters, C.R., Blumenschine, R.J., 1996. Landscape perspectives on possible
land use patterns for early Pleistocene hominids in the Olduvai Basin,
Tanzania: part II, expanding the landscape models. In: Magori, C.,
Saanane, C.B., Schrenk, F. (Eds.), Four Million Years of Hominid Evolution in Africa: Papers in Honour of Dr. Mary Douglas Leakey’s Outstanding Contribution in Paleoanthropology, Kaupia 6. Darmstadter Beitrage
zur Naturgeschichte, Darmstadt, pp. 175e221.
Peters, C.R., Maguire, B., 1981. Wild plant foods of the Makapansgat area:
a modern ecosystems analogue for Australopithecus africanus adaptations.
J. Hum. Evol. 10, 565e583.
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, 127e140.
Peters, C.R., O’Brien, E.M., 1994. Potential hominid plant foods from woody
species in semi-arid versus sub-humid subtropical Africa. In: Chivers, D.J.,
Langer, P. (Eds.), The Digestive System in Mammals: Food, Form and
Function. Cambridge University Press, Cambridge, pp. 166e192.
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 C4 plant foods and possible early
hominid diets. J. Hum. Evol. 48, 219e236.
Plummer, T.W., Bishop, L.C., 1994. Hominid paleoecology at Olduvai Gorge,
Tanzania as indicated by antelope remains. J. Hum. Evol. 27, 47e75.
Potts, R., 1988. Early Hominid Activities at Olduvai Gorge. Aldine de Gruyter,
Hawthorne, NY.
Pratt, D.J., Gwynne, M.D., 1977. Rangeland Management and Ecology in East
Africa. Hodder and Stoughton, London.
Prendergast, H.D.V., Kennedy, M.J., Webby, R.F., Markham, K.R., 2000.
Pollen cakes of Typha spp. [Typhaceae] - ‘lost’ and living food. Econ.
Bot. 54, 254e255.
Prins, H.H.T., Van Der Jeugd, H.P., 1993. Herbivore population crashes and
woodland structure in East Africa. J. Ecol. 81, 305e314.
Runyoro, V., Hofer, H., Chausi, E.B., Moehlman, P.D., 1995. Long-term trends in the
herbivore populations of the Ngorongoro Crater, Tanzania. In: Sinclair, A.R.E.,
175
Arcese, P. (Eds.), Serengeti II: Dynamics, Management, and Conservation of an
Ecosystem. University of Chicago Press, Chicago, pp. 146e168.
Scudder, T., 1962. The Ecology of the Gwembe Tonga. Kariba Studies, vol. II.
Manchester University Press, Manchester.
Sept, J.M., 1986. Plant foods and early hominids at site FxJj 50, Koobi Fora,
Kenya. J. Hum. Evol. 15, 751e770.
Sept, J.M., 1990. Vegetation studies in the Semliki Valley, Zaire as a guide to
paleoanthropological research. In: Boaz, N. (Ed.), Virginia Museum of
Natural History Memorial 1, pp. 95e121.
Sept, J.M., 1994. Beyond bones: archaeological sites, early hominid subsistence, and the costs and benefits of exploiting wild plant foods in east
African riverine landscapes. J. Hum. Evol. 27, 295e320.
Sept, J.M., 2001. Modeling the edible landscape. In: Stanford, C.B.,
Bunn, H.T. (Eds.), Meat Eating and Human Evolution. Oxford University
Press, Oxford, pp. 73e98.
Shipman, P., Harris, J.M., 1988. Habitat preference and paleoecology of
Australopithecus boisei in eastern Africa. In: Grine, F. (Ed.), Evolutionary
History of the ‘‘Robust’’ Australopithecines. Aldine de Gruyter, New York,
pp. 343e382.
Sikes, N., 1994. Early hominid habitat preferences in East Africa: paleosol
carbon isotope evidence. J. Hum. Evol. 27, 25e45.
Sinclair, A.R.E., 1979a. The Serengeti Environment. In: Sinclair, A.R.E., Norton-Griffiths, M. (Eds.), Serengeti: Dynamics of an Ecosystem. University
of Chicago Press, Chicago, pp. 31e45.
Sinclair, A.R.E., 1979b. Dynamics of the Serengeti ecosystem: process and pattern. In: Sinclair, A.R.E., Norton-Griffiths, M. (Eds.), Serengeti: Dynamics
of an Ecosystem. University of Chicago Press, Chicago, pp. 1e30.
Sinclair, A.R.E., 1995. Serengeti Past and Present. In: Sinclair, A.R.E.,
Arcese, P. (Eds.), Serengeti II: Dynamics, Management, and Conservation
of an Ecosystem. University of Chicago Press, Chicago, pp. 3e30.
Smith, P.A., 1976. An outline of the vegetation of the Okavango Drainage
System. In: Proceedings of the Symposium on the Okavango Delta and
Its Future Utilization. Botswana Society, Gaborone, pp. 93e112.
Story, R., 1958. Some Plants Used by the Bushmen in Obtaining Food and
Water. Botanical Survey of South Africa Memoir 30. Department of Agriculture, Division of Botany, Union of South Africa.
Tobias, P.V., 1991. Olduvai Gorge, vols. 4a and 4b. The Skulls, Endocasts and
Teeth of Homo habilis. Cambridge University Press, Cambridge.
Van Lawick-Goodall, J., 1968. The behaviour of free-living chimpanzees in
the Gombe Stream Reserve. Anim. Behav. Monogr. 1, 161e311.
Vincent, A.S., 1985. Wild tubers as a harvestable sesource in the East African
savannas: ecological and ethnographic studies. Ph.D. Dissertation, University of California, Berkeley.
Walter, R.C., Manega, P.C., Hay, R.L., Drake, R.E., Curtis, G.H., 1991. Laser
fusion 40Ar/39Ar dating of Bed I, Olduvai Gorge, Tanzania. Nature 354,
145e149.
Warwick, N.W.M., Halloran, G.M., 1991. Variation in salinity tolerance and
ion uptake in accessions of brown beetle grass [Diplachne fusca (L.)
Beauv.]. New Phytol. 119, 161e168.
Webster, R., Beckett, P.H.T., 1970. Terrain classification and evaluation using
air photography: a review of recent work at Oxford. Photogrammetria 26,
51e75.
White, F., 1983. The Vegetation of Africa. UNESCO, Paris.
Wood, B., 1992. Origin and evolution of the genus Homo. Nature 355,
783e790.
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, 949e970.
Wrangham, R.W., Nishida, T., 1983. Aspilia spp. leaves: a puzzle in the
feeding behavior of wild chimpanzees. Primates 24, 276e282.
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,
715e731.